Scott Eldridge II - Bob Franklin - The Routledge Handbook of Developments in Digital Journalism Studies-Routledge (2018)

THE ROUTLEDGE HANDBOOK OF
DEVELOPMENTS IN DIGITAL
JOURNALISM STUDIES
The Routledge Handbook of Developments in Digital Journalism Studies offers a unique and
authoritative collection of essays that report on and address the significant issues and focal debates
shaping the innovative field of digital journalism studies. In the short time this field has grown,
aspects of journalism have moved from the digital niche to the digital mainstay, and digital
innovations have been ‘normalized’ into everyday journalistic practice. These cycles of disruption
and normalization support this book’s central claim that we are witnessing the emergence of
digital journalism studies as a discrete academic field.
Essays bring together the research and reflections of internationally distinguished academics,
journalists, teachers, and researchers to help make sense of a reconceptualized journalism and its
effects on journalism’s products, processes, resources, and the relationship between journalists and
their audiences. The handbook also discusses the complexities and challenges in studying digital
journalism and shines light on previously unexplored areas of inquiry such as aspects of digital
resistance, protest, and minority voices.
The Routledge Handbook of Developments in Digital Journalism Studies is a carefully curated
overview of the range of diverse but interrelated original research that is helping to define this
emerging discipline. It will be of particular interest to undergraduate and postgraduate students
studying digital, online, computational, and multimedia journalism.
Scott A. Eldridge II is an assistant professor at the University of Groningen, the Netherlands.

He is the author of Online Journalism from the Periphery: Interloper Media and the Journalistic Field
(2018), an associate editor of Digital Journalism, and co-editor with Bob Franklin of The Routledge
Companion to Digital Journalism Studies (2017).
Bob Franklin held the foundation Chair in Journalism Studies at Cardiff University from
2005–2018, is founding editor of the journals Digital Journalism, Journalism Practice, and Journalism
Studies, and edits the new book series Disruptions: Studies in Digital Journalism. Recent publications
include The Future of Journalism: In an Age of Digital Media and Economic Uncertainty (2016).
THE ROUTLEDGE HANDBOOK
OF DEVELOPMENTS IN
DIGITAL JOURNALISM
STUDIES
Edited by Scott A. Eldridge II and Bob Franklin

First published 2019
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
711 Third Avenue, New York, NY 10017
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2019 selection and editorial matter, Scott A. Eldridge II and Bob Franklin; individual chapters,
the contributors
The right of Scott A. Eldridge II and Bob Franklin to be identified as the authors of the editorial
material, and of the authors for their individual chapters, has been asserted in accordance with
sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised in any
form or by any electronic, mechanical, or other means, now known or hereafter invented,
including photocopying and recording, or in any information storage or retrieval system, without
permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation without intent to infringe.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book has been requested
ISBN: 978-1-138-28305-3 (hbk)
ISBN: 978-1-315-27044-9 (ebk)
Typeset in Bembo
by Apex CoVantage, LLC
CONTENTS
List of figures x
List of tables xii
List of contributors xiii
Acknowledgments xxi
Introduction: introducing the complexities of developments

in Digital Journalism Studies 1
Scott A. Eldridge II and Bob Franklin
PART I
The digital journalist: making news 13
1 Law defining journalists: who’s who in the age of digital media? 15

Jane Johnston and Anne Wallace
2 Studying role conceptions in the digital age: a critical appraisal 28

Folker Hanusch and Sandra Banjac
3 Who am I? Perceptions of digital journalists’ professional identity 40

Tim P. Vos and Patrick Ferrucci
4 The death of the author, the rise of the robo-journalist: authorship,

bylines, and full disclosure in automated journalism 53
Tal Montal and Zvi Reich
5 The entrepreneurial journalist 64

Tamara Witschge and Frank Harbers
v
Contents
PART II
Digital Journalism Studies: research design 77
6 Content analysis of Twitter: big data, big studies 79

Cornelia Brantner and Jürgen Pfeffer
7 Innovation in content analysis: freezing the flow of liquid news 93

Rodrigo Zamith
8 An approach to assessing the robustness of local news provision 105

Philip M. Napoli, Matthew Weber, and Kathleen McCollough
9 Reconstructing the dynamics of the digital news ecosystem: a case

study on news diffusion processes 118
Elisabeth Günther, Florian Buhl, and Thorsten Quandt
10 Testing the myth of enclaves: a discussion of research designs

for assessing algorithmic curation 132
Jacob Ørmen
11 Digital news users . . . and how to find them: theoretical and

methodological innovations in news use studies 143
Ike Picone
PART III
The political economy of digital journalism 155
12 What if the future is not all digital? trends in U.S. newspapers’

multiplatform readership 157
Hsiang Iris Chyi and Ori Tenenboim
13 On digital distribution’s failure to solve newspapers’ existential crisis:

symptoms, causes, consequences, and remedies 172
Neil Thurman, Robert G. Picard, Merja Myllylahti, and Arne H. Krumsvik
14 Precarious e-lancers: freelance journalists’ rights, contracts, labor

organizing, and digital resistance 186
Errol Salamon
15 What can nonprofit journalists actually do for democracy? 198

Magda Konieczna and Elia Powers
16 Digital journalism and regulation: ownership and control 211

Victor Pickard
vi
Contents
PART IV
Developing digital journalism practice 223
17 Defining and mapping data journalism and computational journalism:

a review of typologies and themes 225
Mark Coddington
18 Algorithms are a reporter’s new best friend: news automation and the
case for augmented journalism 237
Carl-Gustav Linden
19 Disclose, decode, and demystify: an empirical guide to algorithmic

transparency 251
Michael Koliska and Nicholas Diakopoulos
20 Visual network exploration for data journalists 265

Tommaso Venturini, Mathieu Jacomy, Liliana Bounegru, and
Jonathan Gray
21 Data journalism as a platform: architecture, agents, protocols 284

Eddy Borges-Rey
22 Social media livestreaming 296

Claudette G. Artwick
PART V
Digital Journalism Studies: dialogues 311
23 Ethical approaches to computational journalism 313

Konstantin Dörr
24 Who owns the news? The ‘right to be forgotten’ and journalists’

conflicting principles 324
Ivor Shapiro and Brian MacLeod Rogers
25 Defamation in unbounded spaces: journalism and social media 336

Diana Bossio and Vittoria Sacco
26 Hacks, hackers, and the expansive boundaries of journalism 348

Nikki Usher
27 Journalistic freedom and the surveillance of journalists

post-Snowden 360
Paul Lashmar
vii
Contents
PART VI
Minority voices and protest: narratives of freedom and resistance 373
28 How and why pop-up news ecologies come into being 375
Melissa Wall
29 The movement and its mobile journalism: a phenomenology

of Black Lives Matter journalist-activists 387
Allissa V. Richardson
30 Nature as knowledge: the politics of science, open data, and

environmental media platforms 401
Inka Salovaara
31 Opting in and opting out of media 412

Bonnie Brennen
32 Silencing the female voice: the cyber abuse of women on the internet 425
Pamela Hill Nettleton
PART VII
Digital limits: new debates and challenges for the future 439
33 Social media and journalistic branding: explication, enactment,

and impact 441
Avery E. Holton and Logan Molyneux
34 Reconsidering the intersection between digital journalism and games:

sketching a critical perspective 450
Igor Vobič
35 Native advertising and the appropriation of journalistic clout 463

Raul Ferrer-Conill and Michael Karlsson
36 User comments in digital journalism: current research and future

directions 475
Thomas B. Ksiazek and Nina Springer
37 Theorizing digital journalism: the limits of linearity and the rise

of relationships 487
Jane B. Singer
viii
Contents
38 Outsourcing censorship and surveillance: the privatization of governance

as an information control strategy in the case of Turkey 501
Aras Coskuntuncel
Epilogue: situating journalism in the digital: a plea for studying

news flows, users, and materiality 515
Marcel Broersma
Index 527
ix
FIGURES
6.1 Word clouds with top 100 words from news-related tweets and
non-news-related tweets. 83
6.2 Network of top 20 word occurrences in news-related tweets and other
tweets, as well as their 30 strongest connections in terms of co-occurrence
within tweets 86
9.1 Stepwise procedure to identify relevant events and related online news
reports 119
9.2 Curves of the digital diffusion processes of three exemplary events 125
9.3 Diffusion curves for 95 events as reported by German online news sites,
grouped by dynamics-based cluster membership 126
9.4 Aggregated curves for three clusters of digital diffusion processes featuring
similar dynamics, from a total of 95 diffusion processes among German
online news sites 128
9.5 Dynamics of aggregated diffusion processes, grouped by the time of the
day they were first reported 129
12.1 Composition of in-market combined readership, 2007–2015 164
12.2 In-market print and online reach by age, 2015 167
13.1 Print and digital revenues for the New York Times Co., 2011–2016 174
13.2 Print and digital revenues for Fairfax Media, 2011–2016 175
20.1 Sociogram representing friendship among school pupils (original title and
image accompanying Moreno’s 1933 New York Times interview) in the
original version and in the modern force-directed spatialization 267
20.2 The Décodex network spatialized by ForceAtlas2 271
20.3 Distribution of TDL in the Décodex network 272
20.4 Zoom on the French regional press 273
20.5 The ‘satirical’ websites according to the original Décodex classification 274
20.6 Highlight of the ‘reliable’ websites and ‘unreliable’ and ‘imprecise’ websites 275
20.7 Highlight of the ‘conspiratorial’ websites and ‘right’ and ‘extreme right’
websites 276
20.8 The heterogenous territories of the Décodex network 278
x
Figures
20.9 Distribution of the number of nodes per category 279

20.10 Connectivity between the categories of final classification 279
20.11 Hierarchical structure in the corpus, based on final categories 280
20.12 Simplified version of the statistical analysis presented in Figure 20.11 281
xi
TABLES
6.1 Sentiment analysis categories comparing news-related tweets with

non-news-related tweets 84
6.2 Top 10 words associated with five topics from news-related as well as
other tweets 87
6.3 Accuracy of the Support Vector Machine for predicting whether a user
account is news-related or not 88
8.1 List of media directories/databases consulted 109
10.1 Matrix of prototypical algorithmic curation designs 135
12.1 In-market print and online reach, 2007–2015 162
12.2 In-market online and print reach by age, 2015 165
13.1 Comparison of the time spent reading per day by online and print readers
of 11 UK national newspaper brands 176
13.2 Proportions of the annual time spent with each of 11 UK national newspaper
brands by their aggregated British adult audiences that come via their print
editions, via PCs and via mobile devices, April 2015–March 2016 (inclusive) 177
19.1 Summary of transparency factors across four layers of algorithmic systems 255
23.1 Table showing levels and grades of agency 317
38.1 Demiroren Group’s investments 506
xii
CONTRIBUTORS
Claudette Guzan Artwick is Associate Professor of Journalism and Mass Communications at

Washington and Lee University, US. Her research explores the role of digital technology in jour-
nalism, citizen livestreaming and eyewitness video, and digital media use and issue perception.
She is the author of a forthcoming book on social media livestreaming in the Routledge series
Disruptions: Studies in Digital Journalism, and of Reporting and Producing for Digital Media.
Sandra Banjac is a research associate and Ph.D. candidate in the Department of Communica-
tion at the University of Vienna. Her research explores journalistic role conceptions and audi-
ence expectations through the lens of intersectionality, with special interest in journalism cultures
beyond the West.
Eddy Borges-Rey is Senior Lecturer in Journalism Studies and Associate Dean for Research
and Knowledge Exchange in the Faculty and Arts and Humanities at the University of Stirling,
UK. He is also the program director of the MSc. Media and Communications Management at
Stirling-Vietnam National University. His research looks at the interplay between media, tech-
nology, and power, particularly around data journalism, critical data, code and algorithm studies,
AI and automation, and data literacy.
Diana Bossio is a senior lecturer and researcher in media and communication at Swinburne
University in Melbourne, Australia. Dr. Bossio’s book Journalism and Social Media: Practitioners,
Organisations, Institutions will be published in 2018.
Liliana Bounegru is a doctoral candidate at Ghent University and the University of Groningen,
Netherlands. She is also co-founder of the Public Data Lab, member of the Digital Methods Ini-
tiative (University of Amsterdam) and research associate at the Sciences Po Paris médialab. More
about her can be found at lilianabounegru.org.
Cornelia Brantner is (temporary) Professor of Media and Communication Studies at the Cen-
tre for Media, Communication and Information Sciences (ZeMKI) at Bremen University. Her
research focuses on digital communication (social media, geomedia, online journalism), visual
xiii
Contributors
communication, social movements, political communication, and qualitative and quantitative

methods.
Bonnie Brennen is the Nieman Professor of Journalism in the Diederich College of Com-
munication at Marquette University. Her research focuses on the intersection between labor and
journalism history as well as on relationships between media, technology, culture, and society.
Marcel Broersma is Professor and Director of the Centre for Media and Journalism Studies at
the University of Groningen, Netherlands. His research examines the use of social media in jour-
nalism and politics, shifting patterns of news use, emerging forms and styles of journalism, and
the epistemology of journalism. He is the author of numerous journal articles and book chapters.
His books include Form and Style in Journalism (2007), Rethinking Journalism (Routledge, 2013),
Rethinking Journalism Again (2016) – the latter two co-edited with Chris Peters – and Redefining
Journalism in the Era of the Mass Press 1880–1920 (Routledge, 2016, co-edited with John Steel).
Florian Buhl, M.A., is a research assistant at the Department of Communication, University of

Münster, Germany. His research interests in digital journalism focus on gatekeeping processes and
the dynamics of news flows and issue attention.
Hsiang Iris Chyi, author of Trial and Error: U.S. Newspapers’ Digital Struggles toward Inferiority,
is an Associate Professor in the School of Journalism at the University of Texas at Austin. Her
“Ramen Noodles Theory” suggests that online news, like ramen noodles, is an inferior good.
Mark Coddington is an assistant professor in the Department of Journalism and Mass Commu-
nications at Washington and Lee University, where he studies the sociology of digital journalism.
He has published research in journals such as Journalism Studies, Journalism & Mass Communica-
tion Quarterly, and Mass Communication and Society. He is currently working on a book on news
aggregation with Columbia University Press. He is a former newspaper reporter and contributor
to the Nieman Journalism Lab at Harvard University and holds a Ph.D. from the University of
Texas at Austin.
Aras Coskuntuncel is a Ph.D. candidate and instructor in communication at American Uni-

versity. He focuses on the political economy of media, internet governance, and surveillance. He
was a news editor at the Hurriyet Daily News in Istanbul, Turkey, before completing his master’s
degree in media studies at the University of Wisconsin-Milwaukee.
Nicholas Diakopoulos is an assistant professor at the Northwestern University School of

Communication, where he directs the Computational Journalism Lab. His research is in compu-
tational and data journalism, with emphases on algorithmic accountability and social computing
in the news.
Konstantin Dörr Before joining Deloitte Consulting, Dr. Konstantin Dörr was a Research
and Teaching Associate in the Media Change & Innovation Division at the IPMZ, University of
Zurich. His research is centered on algorithmic selection, big data, ethics, media innovations, and
media change in journalism.
Scott A. Eldridge II is an assistant professor at the Centre for Media and Journalism Studies at
the University of Groningen, Netherlands, where he studies changing conceptions of journalistic
xiv
Contributors
identity and the journalistic field. He is the author of Online Journalism from the Periphery: Interloper
Media and the Journalistic Field (2018) and co-editor with Bob Franklin of The Routledge Companion
to Digital Journalism Studies (2017). He is also an associate editor of the journal Digital Journalism.
Raul Ferrer-Conill is a doctoral candidate in the department of Media and Communication

Studies at Karlstad University, Sweden. He has published his work in television and new media,
journalism studies, and computer-supported cooperative work (CSCW) among others. His cur-
rent research covers gamification and digital journalism.
Patrick Ferrucci (Ph.D., University of Missouri) is an assistant professor in the Department

of Journalism in the College of Media, Communication and Information at University of
Colorado-Boulder. His research examines different aspects of media sociology, primarily eco-
nomic and technological influences on digital news production.
Bob Franklin held the foundation Chair in Journalism Studies at Cardiff University from
2005–2018, is founding editor of the journals Digital Journalism, Journalism Practice and Journalism
Studies and edits the new book series Disruptions: Studies in Digital Journalism. Recent publications
include The Routledge Companion to Digital Journalism Studies (2017, with Scott Eldridge II) and
The Future of Journalism: In an Age of Digital Media and Economic Uncertainty (2016).
Jonathan Gray is Lecturer in Critical Infrastructure Studies at the Department of Digital

Humanities, King’s College London. He is Cofounder of the Public Data Lab and Research
Associate at the Digital Methods Initiative (University of Amsterdam) and the médialab (Sciences
Po, Paris). More about him can be found at jonathangray.org and he tweets at @jwyg.
Elisabeth Günther is a Ph.D. candidate at the Department of Communication, University of

Münster, Germany. Her research interests are in online journalism, technological change, and
computational research methods.
Frank Harbers is an assistant professor at the Centre for Media and Journalism Studies at the
University of Groningen, Netherlands. His Ph.D. (2014) focused on the historical development
of the European newspaper; he now focuses on current developments and innovative initiatives,
particularly journalistic start-ups, in the journalistic field and has published several articles on
this subject.
Folker Hanusch, Ph.D., is Professor of Journalism in the Department of Communication at the

University of Vienna, as well as Adjunct Professor at Queensland University of Technology. His
research focuses on journalism cultures, comparative journalism studies, lifestyle journalism, and
digital transformations in journalism.
Avery E. Holton is an associate professor and Humanities Scholar in the Department of Com-
munication at the University of Utah. His research emphasizes the ways in which social and
digital media alter flows of information and constructions of identity.
Mathieu Jacomy is a research engineer at the Sciences Po Paris médialab since 2010. He
develops digital tools for the social sciences, like the web crawler Hyphe or Gephi. His current
research focuses on visual network analysis, digital methods, and issue mapping. He tweets at @
jacomyma.
xv
Contributors
Jane Johnston is Associate Professor of Communication at The University of Queensland. She

has been researching the interface between the media, legal systems, and the public for many
years. Most recently, this has focused on the impact brought about by changed media systems
and practices
Michael Karlsson is a professor in Media and Communication at Karlstad University, Sweden.

His research area is primarily digital journalism and his work has been published in journals
such as the Journal of Computer-Mediated Communication, New Media & Society, and Communication
Theory.
Michael Koliska is an assistant professor for journalism at Georgetown University’s Communi-

cation, Culture and Technology Program. His research focuses on issues of authenticity, transpar-
ency, and trust in traditional and computational journalism.
Magda Konieczna is an assistant professor in the journalism department at Temple University.

Her book about nonprofit news, titled Journalism Without Profit: Making News When the Market
Fails, will be published in 2018.
Arne H. Krumsvik, Ph.D., is considered to be one of the founders of media innovations studies.
He is Head of Department at Department of Media and Communications, University of Oslo,
Norway, and Professor II at Westerdals Oslo School of Arts, Communication and Technology.
Thomas B. Ksiazek holds a Ph.D. from Northwestern University, US, and is an associate pro-
fessor in the Department of Communication at Villanova University, US. His research interests
include patterns of cross-platform media use, new forms of user engagement with the news,
implications of audience behavior for society and the field of journalism, and the application of
network analysis to the consumption and production of media.
Paul Lashmar is a senior lecturer at City, University of London and conducts research on the
relationship between the media and intelligence agencies. He is an academic-practitioner and
has been a well-regarded investigative journalist working in television, radio, online, and print
for four decades.
Carl-Gustav Linden is adjunct professor at the Swedish school of social science, University of
Helsinki. His research focuses on media and journalism innovation, especially news automation.
He is also studying business models for media in the European Union and the United States.
Dr. Linden has a background in business journalism for television and print and also worked as
a communications consultant to the United Nations University, among other clients. He is an
affiliated lecturer at the Södertörn University in Stockholm.
Kathleen McCollough is a Visiting Assistant Professor of Media Studies at Augustana Univer-

sity in Sioux Falls, US. She recently completed her Ph.D. in the School of Communication &
Information at Rutgers University, where she served as the lead graduate research assistant on the
News Measures Research Project.
Logan Molyneux is an assistant professor of journalism at Temple University. His research

focuses on how journalism practice and new media technology, especially how journalists use
mobile and social media in fulfilling their democratic role.
xvi
Contributors
Tal Montal is a data engineer in the Israeli digital analytics industry. He completed his master’s
degree in the Department of Communication Studies at the Ben Gurion University of the
Negev, Israel. His research interests focus on robo-journalism and data journalism
Merja Myllylahti is a post-doctoral fellow at Auckland University of Technology (AUT), New

Zealand. She is also author and researcher at the AUT research center for Journalism, Media and
Democracy (JMAD). Her work has been published in international academic journals and books
including Journalism Studies, Digital Journalism and The Routledge Companion to Digital Journalism
Studies.
Philip M. Napoli is the James R. Shepley Professor of Public Policy in the Sanford School of
Public Policy at Duke University, where he is also a faculty affiliate with the DeWitt Wallace
Center for Media & Democracy. His research focuses on media institutions and policy
Pamela Hill Nettleton holds a BS in journalism, an MA in mass communication, and a

Ph.D. in communication studies from the University of Minnesota. She is an award-winning
researcher, teacher, and academic and popular press writer. She is an assistant professor at Mar-
quette University.
Jacob Ørmen is an Assistant Professor at the University of Copenhagen. He studies audiences

in a digital world and has published articles and book chapters on algorithmic curation, news
consumption, and sharing as well as empirical studies of media use on an international scale.
Jürgen Pfeffer is Professor of Computational Social Science & Big Data at Technical University of
Munich. His research deals with the analysis of large and dynamic social, political, and economic sys-
tems as well as with methodological and algorithmic challenges arising from analyzing such systems.
Robert G. Picard is North American Representative of the Reuters Institute at the University
of Oxford, a fellow of the Royal Society of Arts, and an affiliated fellow of the Information Soci-
ety Project at Yale University. He has authored and edited 30 books.
Victor Pickard is an associate professor of political economy and media policy at the University
of Pennsylvania’s Annenberg School for Communication. He is the author of America’s Battle for
Media Democracy and co-editor of Will the Last Reporter Please Turn Out the Lights.
Ike Picone is Assistant Professor at the Department of Communication Sciences at the Vrije
Universiteit Brussel and Senior Researcher within the research institute on Studies on Media,
Innovation and Technology (SMIT). His research is situated on the crossroad of journalism,
technology, and democracy.
Elia Powers is an assistant professor of journalism and new media at Towson University, Mary-
land. He is proud to have worked at a nonprofit news outlet that made an impact during his
reporting career. His current research focuses on, among other things, audience engagement and
the impact of journalism.
Thorsten Quandt is professor of online communication at the University of Münster, Germany.

His work is focusing on dysfunctional forms of online use, online journalism, cyber propaganda/
populism, and VR/AR.
xvii
Contributors
Zvi Reich is an associate professor in the Department of Communication Studies at the Ben
Gurion University of the Negev, Israel, and a former journalist. His research interests focus on
journalism studies, new technology, authorship, and journalistic knowledge and expertise.
Allissa V. Richardson, Ph.D., is Assistant Professor of Journalism at the University of Southern

California’s Annenberg School for Communication and Journalism. She is a Nieman Founda-
tion Visiting Journalism Fellow at Harvard University (’14), a two-time Apple Distinguished
Educator, and the 2012 Journalism Educator of the Year for the National Association of Black
Journalists.
Brian MacLeod Rogers is a media lawyer and adjunct professor at Ryerson University and was
founding President of Ad IDEM/Canadian Media Lawyers Association. He has argued several
cases before the Supreme Court of Canada, co-authors annual surveys on Canadian law for the
Media Law Resource Center, and wrote the Canadian chapter in the International Libel & Privacy
Handbook (2016).
Vittoria Sacco works at the statistical office of the canton of Fribourg as a scientific advisor.
She held a post-doctoral position at the Academy of Journalism and Media at the University of
Neuchâtel, Switzerland.
Errol Salamon is a Postdoctoral Research Fellow and Visiting Scholar at the University of Penn-
sylvania in the Annenberg School for Communication. He is co-editor of Journalism in Crisis:
Bridging Theory and Practice for Democratic Media Strategies in Canada (2016).
Inka Salovaara is Professor of Media Studies and Research Leader of the Digital Ecologies
Research Group (DERG) at the Department of Culture and Communication, University of
Southern Denmark. She studies post-human life and artificial intelligence in everyday contexts,
asking ‘What are new socio-technological nonhuman actors in digital media spaces?’
Ivor Shapiro teaches journalism ethics at Ryerson University in Toronto. His research in jour-
nalists’ professional identity, values, and practices has been published in Journalism Studies, Digital
Journalism, Canadian Journal of Communication, Newspaper Research Journal, and several books.
Jane B. Singer is Professor of Journalism Innovation at City, University of London, where she
also serves as research lead for the Department of Journalism. Co-author of Participatory Jour-
nalism and Online Journalism Ethics, her research explores digital journalism, including changing
roles, perceptions, norms, and practices.
Nina Springer is a senior lecturer at Södertörn University, Stockholm. Her research interests
include the ‘uses and gratifications’ of commentary features, that is, why people comment on
news or enjoy reading these contributions. Linked to this, she works on viewpoint diversity in the
news and the contribution of user comments to it but also on the negative impact of problematic
forms of audience feedback, for instance on female journalists.
Ori Tenenboim is a doctoral candidate in the School of Journalism and a Research Associate
with the Center for Media Engagement at The University of Texas at Austin. His research inter-
ests center on digital journalism, political communication, and media economics. Tenenboim
examines how journalists and news organizations hybridize older and newer behaviors, norms
xviii
Contributors
and forms in different cultural and political contexts. His work has been published in Journalism:
Theory, Practice & Criticism, International Journal of Communication, Digital Journalism and Journalism
Practice.
Neil Thurman is Professor of Communication in the Department of Communication Studies

and Media Research, LMU Munich. He holds a Volkswagen Foundation Freigeist Fellowship. Dr.
Thurman’s research focuses on the changes taking place in news production and consumption
as a result of the internet.
Nikki Usher is an associate professor at George Washington University’s School of Media and
Public Affairs. She is author of Interactive Journalism: Hackers, Data, and Code (2016), a look at the
evolving role of hacker journalism, and Making News at The New York Times (2014), an in-depth,
ethnographic portrait of the newspaper in the digital age.
Tommaso Venturini is researcher at the Institut des Systèmes Complexes and the École Nor-
male Supérieure of Lyon. He is recipient of the INRIA “Advanced Research” fellowship. In
2016, he was lecturer at the Digital Humanities Department of King’s College London and,
from 2009 to 2015, research coordinator of the médialab of Sciences Po Paris, with which he is
still associated.
Igor Vobič is Associate Professor at the Faculty of Social Sciences, University of Ljubljana, Slo-
venia. His research interests lie in the material and discursive aspects of technological innovations
in journalism, with a particular focus on transformations of newswork, societal roles of news, and
journalistic ideology.
Tim P. Vos is Chair and Associate Professor of Journalism Studies and Coordinator of Global
Research Initiatives at the University of Missouri School of Journalism. He is co-author (with
Pamela Shoemaker) of Gatekeeping Theory (2009) and co-editor (with François Heinderyckx) of
Gatekeeping in Transition (2015).
Melissa Wall is Professor of Journalism at California State University Northridge, where she
studies participatory media. She is the editor of Citizen Journalism: Valuable, Useless or Danger-
ous and the author of the forthcoming volume Citizen Journalism: Practices, Propaganda and
Pedagogy.
Anne Wallace is a professor and judicial administration scholar at La Trobe Law School in
Melbourne, Australia. Her research includes a focus on the impact of technology on court opera-
tions, and her most recent work investigates the impact of social media on jury trials and on the
practice of court reporting.
Matthew Weber is an associate professor of communication in the School of Communication &

Information at Rutgers University, where he also co-directs Rutgers’ NetSCI Network Science Lab.
His research focuses on organizational adaptation and change in response to new communications
technologies.
Tamara Witschge (Ph.D., University of Amsterdam) is Associate Professor and Rosalind Frank-
lin Fellow at the University of Groningen, Netherlands. Her research explores the ways in which
cultural, technological, economic, and social changes are reconfiguring journalism and other
xix
Contributors
cultural industries, with a particular focus on innovation and entrepreneurship. She is co-editor
of the Sage Handbook of Digital Journalism (2016) and co-author of Changing Journalism (2012).
Rodrigo Zamith is Assistant Professor in the Journalism Department at the University of Mas-
sachusetts Amherst. His research interests lie at the intersection of journalism and technology,
with a focus on the reconfiguration of journalism in a changing media environment and the
development of digital research methods.
xx
ACKNOWLEDGMENTS
This Handbook is the second of our editorial collaborations, and in each of these we have been
fortunate to enjoy the support of colleagues and friends whose feedback has helped keep us on
track as we have worked on these volumes. For their encouragement, we are both grateful. More-
over, we would like to acknowledge the contributions of the authors who have come together in
the pages that follow and thank them for joining us in this project. Without their appreciation
of what we have sought to achieve here, we would not be able to present such a wide-ranging
and fascinating collection of excellent scholarship. We are also thankful to the support of those
at Routledge, but especially Kitty Imbert, who helped guide this project from conception to
publication, and to the anonymous reviewers whose feedback helped strengthen the shape of the
work within.
xxi
INTRODUCTION
Introducing the complexities of developments
in Digital Journalism Studies
Development is a tricky thing; it can be positive, but there is no guarantee that it will be so.
Moreover, while it can be linear, it often meanders. And, while the term can signal the expansion
of a field, it can also refer to a growing body of ideas within an unruly domain with uncertain
boundaries. These dynamics are particularly in focus when the conversation concerns develop-
ments in a field of scholarly research like Digital Journalism Studies that is made up in equal
parts of developing research and research agendas and is inherently concerned with developments
in a rapidly changing field of journalism as it unfolds around us: it is perhaps unsurprising, then,
in this context that Peter Dahlgren notably dubbed digital journalism a “sprawling domain”
(Dahlgren, 2013: 160). In this Handbook, we identify developments in various components of
Digital Journalism Studies, including shifts in what we understand as digital journalism and
developments in how we research digital journalism as well as developments in how we think
about digital journalism and its place in our broader societies. The Routledge Handbook of Develop-
ments of Digital Journalism Studies aims to show how a field that itself has emerged only recently
(Franklin and Eldridge, 2017) has developed rapidly to present a more complex and richer set of
academic discussions.
In addressing a range of concerns journalism scholars have raised in recent years while mak-
ing sense of journalism within our increasingly (but not exclusively) digital environment, this
Handbook also incorporates a series of answers to those concerns and presents new modes of
research inquiry. In doing so, the ambition is to invigorate research agendas that can brighten
the corners of digital journalism research. This collection brings together scholars who are pro-
ducing leading-edge research in digital journalism, including emerging scholars and work that
focuses on aspects of change in journalism that have been underexplored as the field continues
to take shape. Contributors illustrate how to approach the shifting technologies of a digital age
and explore ways to establish these while regarding their implications for journalism and its place
in our societies. Chapters also show where digital developments have at times been more uneven
than we might have initially anticipated, examining where early enthusiasm for digital oppor-
tunities left important debates and unforeseen consequences underexplored. In bringing these
voices to the forefront, we also highlight research that explores those circumstances which make
it hard to see developments in digital technologies as wholly positive. In this introduction, we
guide a discussion through the 39 contributions in the Handbook, situating their analyses within
1
extant but also emerging areas of Digital Journalism Studies to introduce the themes, structure,
and topics addressed in this collection.
New complexities, new debates, and fundamental concerns

Digital Journalism Studies could easily be cast as a field defined by change, given that its focus
is on the most recent shifts in journalism, and could equally be seen in large part through its
inextricable link to the advancement of digital technologies. Yet this would paint an incomplete
picture of the breadth of scholarship involved (as argued in Eldridge and Franklin, 2017), since
while change remains a salient consideration and technology a key focus, we emphasize in this
Handbook where these themes are interwoven into the broader social dynamics of journalism.
Analysis also needs to consider where ‘digital’ does not necessarily offer an explanation of all
that occurs within our societies and where other factors temper any narratives of digital tech-
nology revolutionizing journalism or its place in society (Broersma and Peters, 2016: 9). This is
reflected in the work of Bonnie Brennen (Chapter 31), who helpfully reminds scholars of those
people who prefer to ‘opt out’ of digital media for a variety of reasons, and when we speak of the
scope and scale of change in increasingly but not exclusively digital societies, we need to bear in
mind that digital prominence is not the same as ubiquity. To begin our introduction, it is help-
ful to keep this consideration in mind for what it says about the interaction between societies
and technologies and publics and journalism. Such measured reflection draws our attention to
the ways Digital Journalism Studies is at once unpacking how news and information are being
communicated using new technologies but also to where journalism’s current practices and our
thinking about these resonate with journalistic roles that predate our digital age. Indeed, we see
it as an encouraging sign that within the discussions of digital journalism explored here, we join
researchers who have advanced research into digital journalism that considers the breadth of
change within societies, of technologies, and for individuals who through their own activities
help us better understand digital journalism.
We can explore this breadth, for instance, by seeing where perceptions of journalism acting
as a watchdog, a ‘fourth estate’, and as a source of news for audiences within societies have not
been abandoned in a digital age but reimagined (Benkler, 2011). Embedded within these terms
are an implied set of values and ideals that shape the ways we speak of journalism, often wrapped
in its normative aspirations (Steel, 2016). As a launching point for discussing roles taking shape
in more digital contexts, we can look at the ways values and ideals shape journalists’ role concep-
tions and where these could be understood with greater complexity. This is the focus of Folker
Hanusch and Sandra Banjac (Chapter 2), who identify role conceptions as central to journalists’
positioning of themselves in societies, though these are underexplored in all their richness. In
particular they highlight where audiences’ perceptions of journalists have been missing in extant
role conceptions research. In this discussion, they draw a critical link between journalists, their
products, and who they produce them for, which sees audiences’ own expectations as potential
shapers of journalists’ approach to their work. It is a link that is often at the center of how we talk
about journalism research but warrants being considered anew as research agendas move ahead.
Within such agendas, we can see for instance where developing research could embrace new
modes of understanding journalism as a societal field that is shaped by a range of internal and
external influences (Eldridge, 2018; Deuze and Witschge, 2017). These aspects of journalism
affected are linked to but not limited to the ways journalists imagine their roles or to audiences’
contributions to shaping these in a digital age (Hanitzsch and Vos, 2016). Instead, they are also
tied to journalism in transition, particularly as journalists move from identifying with the tradi-
tional institutional locus to new dimensions of practice which take various guises. Tim P. Vos and
2
Introduction
Patrick Ferrucci (Chapter 3) show this shift in their interviews with digital journalists, consider-
ing stasis and change together as this new classification of digital journalists incorporates both
new and traditional roles. These are arrived at by journalists who remain committed to making
meaningful contributions to society, despite shifts in the ways these are made (Baack, 2017). This
points to one of the interesting threads within journalist-centered research in this Handbook,
where the construction of journalistic identities is built around both emerging and tradition-
ally established narratives of journalistic importance. This is found among journalists working
in start-ups, and as Tamara Witschge and Frank Harbers (Chapter 5) outline, their identities are
shaped in part by aligning with and in part by contrasting against traditional notions of journal-
ism. This revolves in part around critical perspectives of journalism as a business (Hardy, 2017)
that eschew traditional commercial priorities while embracing a more familiar allegiance to
journalism’s importance for society. Elsewhere, Magda Konieczna and Elia Powers (Chapter 15)
reinforce this narrative, and when we contend with new types of journalism and where journal-
ists are actively reimagining the ways journalism can operate, as with nonprofit newsrooms, we
see traditional concepts of a journalistic fourth estate and journalism’s democratic roles promi-
nently within the innovative nature of digital work.
These discussions have led us some distance toward situating new types of actors within the
journalistic field. However, changes more widely in society have not been as a response to the
digital changes journalism has embraced and which Digital Journalism Studies’ scholars consider
(Schudson, 2017: 268–269). Within any discussion of journalists being newly able to capture
the ideals of a journalistic field in a digital era, it is worth being reminded that as a societal
field, journalism still contends with other institutions of power that have their own priorities
and demands (Benson, 2006). Errol Salamon (Chapter 14) shows this in particular, as shifting
practices of journalism among freelancers place them in precarious labor positions, particularly
when they are working at the whim of larger publishers. Further, Victor Pickard (Chapter 16)
shows from a political economy perspective how, higher up among fields of power in society,
ownership, regulatory frameworks, policy, and politics further pose risks to a more optimistic
future for digital journalism being realized. In our opening chapter, by Jane Johnston and Anne
Wallace (Chapter 1), the struggle in understanding journalism as recognizable, yet reflective of its
many digital forms, has also exposed how the rapid change journalism has undergone has out-
paced institutions’ ability to respond to change, particularly within judicial and legal frameworks.
Where Johnston and Wallace show these institutions are moving less rapidly than the changes
journalism itself has undergone, Ivor Shapiro and Brian MacLeod Rogers (Chapter 24) show how
within legal protections for the public – for those who would rather be ‘forgotten’, for various
reasons – institutions have been able to regulate privacy from digital search, even as the long-term
implications of such protections for journalism’s informative roles remain unclear. This exposes
a tension between journalists’ ability to inform and investigate those in power when legislation
favors one societal group over the other (Thorsen, 2017). This is a central theme among those
working in investigative journalism whom Paul Lashmar (Chapter 27) interviewed, as they are
forced to contend with the very surveillance apparatuses they also report on.
Revisiting journalism studies’ tripartite approach

This initial sketch of some of the challenges explored within this Handbook starts to outline the
need for researching journalism not on its own but within an increasingly complex society where
oppositions between different groups of actors and fields of powers regularly clash and brings us to
a particular struggle for understanding digital journalism in a digital age. While always a simplified
construction of the breadth of research in journalism studies, the attention of journalism research
3
has focused on some combination of studying the producers (journalists), the products (news),
and the people (audiences) involved one way or another in journalism (Conboy, 2013: 2–3). This
tripartite approach assumed, reasonably in many instances, certain things of each of these foci. Of
journalists, research could focus on those individuals in society who gathered, verified, and shared
information with a public in their interest, even as this group of individuals grew past familiar
boundaries (Eldridge, 2017). Of news, it assumed a certain facticity within information-based
content, communicated from one to many and later from many to many. Of audiences, it explored
these in a variety of ways, and while these become increasingly more sophisticated, moving from
public to publics and passive to active audiences, the focus was on their reception (Costera Meijer
and Groot Kormelink, 2015). For contemporary research to be situated within any of these three
areas has, in all of its forms, become a more complicated demand (Bruns, 2017).
First, we can see this within research understanding journalists. When we spoke of journalists’
ambitions or lauded their contributions or explored their roles, research has focused on journal-
ists acting on their journalistic agency (Hermida and Young, 2017). Such research was primarily
focused on human journalists, yet we see here – in the work of Tal Montal and Zvi Reich (Chap-
ter 4) – this is no longer sufficient, as so-called robot journalism has entered the discussion in
ways that further complicate our understanding of what it is to be a journalist, to exercise inde-
pendence, to have a reporter’s initiative, and to situate the agency of journalistic endeavor. Kon-
stantin Dörr (Chapter 23) analyzes this from an ethical perspective, where the use of algorithms
reorients computational approaches to journalism toward more dispersed agency across more
interconnected digital contexts. The implications for these changes have become clearer as both
technologies and research into these have advanced. Of course, with greater integration of com-
putational approaches into news, further problems for understanding these aspects of journalism
around specific societal understandings of journalism also emerge (Broussard, 2015). Namely,
these see journalism as driven by data and algorithms and caught within a crisis narrative that
suggests rather than improving the availability of information for society, the death of journalism
will instead soon follow (Franklin, 2016). A worrying prognosis, if it was to unravel in this way.
For such fears, we are well-served by considering more reflective discussions on the changes
journalism is undergoing, such as those provided by Mark Coddington (Chapter 17), and Carl-
Gustav Linden (Chapter 18), who each puts the role of computational programming in context
and integrates into the discussion of changes for news producers the discussion of news products.
Coddington offers a typology that demystifies the various types of computational journalism from
algorithmic to data-driven, while Linden emphasizes the advantages of automation where it com-
plements the work of journalists themselves. We see in their treatment of computational journal-
ism where work in digital journalism studies is at its best when it bridges the place of journalism
in society alongside the way digital technologies have enabled new types of newswork, enriching
new types of content that channel a familiar journalistic authority (Carlson, 2017). Such solutions
are often developed in the hands of computer-savvy journalists and coders – hacks and hackers, as
Nikki Usher (Chapter 26) describes – who advance the opportunities of digital journalism to bring
data, coding, and journalism’s societal contributions together. Such a story is also being told by Inka
Salovaara (Chapter 30), as media platforms use open data to reshape discussions of nature and the
environment and engender new engagement with the changes our world is facing. Digital story-
telling and using data to paint richer pictures within news is shown as a vibrant possibility by Tom-
maso Venturini, Mathieu Jacomy, Liliana Bounegru, and Jonathan Gray (Chapter 20), who show
how joining the technologies of data journalism and network visualizations offer journalists new
ways to engage with visual storytelling and open new ways for data journalists to do their work.
In these discussions, we are reminded that while ‘data’ is a prominent focus across digital jour-
nalism, it is not a singular source of information, nor is it a discrete aspect of news storytelling;
4
Introduction
rather, it is complex, and the way we think about data within digital journalism studies needs
to be equally engaged with that complexity (Lewis, 2014). For what remains a relatively small
aspect of overall journalism practice, Eddy Borges-Rey (Chapter 21) places data journalism in the
context of a diminished local news provision and outlines where it remains a point of tension
in newsrooms where data journalism is being integrated. And while greater computer-mediated
options have also meant new ways of telling journalistic stories, including through newsgames
and the ‘gamification’ of journalism, as Igor Vobič (Chapter 34) writes, that playfulness has impli-
cations for the materiality of journalism and its commercial imperatives – including blurring lines
between commercial incentives and information provision. Such caution alongside optimism
can also be found when looking at the advantages and risks of native advertising, unpacked by
Raul Ferrer-Conill and Michael Karlsson (Chapter 35), which have blurred the division between
editorial and advertising, even as they open new revenue streams for struggling news organiza-
tions. Similarly, journalists straddle the demands of self-promotion and journalistic norms in
the way they themselves have become ‘brands’, and as Avery E. Holton and Logan Molyneaux
(Chapter 33) point out, this demands journalists confront a range of decisions as they weave per-
sonal and professional aspects of themselves into their social media presence (Lasorsa et al., 2012).
These discussions are part of the challenges that journalism as an industry has also faced in the
digital era, including how journalists and news organizations make contact with the third aspect
of journalism research – their audiences. This includes both how they are measured and how they
are informed (Schrøder, 2017). While there may be (seeming) solutions in new forms of commu-
nicating news and information, for a large portion of journalism these are seated within financial
concerns, and a ‘crisis narrative’ of newspapers in particular trying to resolve their balance sheets
in a digital age (Gasher et al., 2016). Hsiang Iris Chyi and Ori Tenenboim (Chapter 12) look at
how this has taken shape across 20 years of multiplatform reading – comparing newspaper and
online readership figures alongside one another – to put newspapers’ struggles into perspective.
They argue from this we may need to revisit measures of readership, as these practices have grown
increasingly complex. In doing so, however, they note that where there is enthusiasm for online
approaches, digital content should not be seen as a panacea to newspapers’ existential crises;
instead, they contend ways we make sense of these challenges warrant refinement.
Neil Thurman, Robert G. Picard, Merja Myllylahti, and Arne H. Krumsvik (Chapter 13)
argue for such refinement, indicating new ways of considering audience engagement with jour-
nalism in light of newspapers’ financial interests. These, they contend, would consider more than
just the reach of newspapers and, by exploring journalism through engagement, would see a way
forward from ominous narratives of journalism in crisis. Newspapers may fade, they note, but the
journalism traditionally found in newspapers can live on through innovative approaches. Where
in instances this shift is demonstrable, new approaches are still needed in order to make sense
of such transitions that weigh where journalism’s move from paper to page has not been uni-
form (Williams et al., 2015). As Philip M. Napoli, Matthew Weber, and Kathleen McCollough
(Chapter 8) identify, for local news the shuttering of local news outlets and the move towards
online information provision has led to news deserts, at worst, and uneven information provision
elsewhere. Napoli and his colleagues pose a methodological approach that can help identify gaps
between journalism’s work in society and society’s need for information while making sense of
the many ways this may emerge online and where this is happening unevenly.
Caution and enthusiasm: critical perspectives

Stretching between caution and enthusiasm, we find not only in the ways journalists see their
work but also in how they conduct their work that recent technologies have foregrounded new
5
types of journalistic endeavor. Claudette G. Artwick (Chapter 22), for example, demonstrates
this in an examination of social media livestreaming, highlighting both its advantages for cut-
ting newswork and the risks of an open and instantaneous mode of broadcasting for journalism
and beyond in a series of case studies. One could draw similar tensions between enthusiasm and
consternation with news sites’ user comments (Ksiazek, 2016). These were imagined first as a
welcome space for interaction between newsrooms and their digital audiences but realized later
by Thomas B. Ksiazek and Nina Springer (Chapter 36) as uneven spaces for engagement, where
norms of civility are sometimes lacking and where the demands of maintaining a participatory
space for all mean that early enthusiasm often goes unrealized. Pamela Hill Nettleton (Chapter
32) takes caution a sobering step further. In a crucially important look at the violent language
and harassment directed at women journalists online, she brings into sharp focus that for all
the directions digital spaces could have developed, they remain spaces of risk for many working
in journalism. These aspects of trolling, harassment, and verbal assault, while deplorable, also
underscore the complex challenges in the digital environment where journalists, audiences,
and other members of society more generally all come together (Eckert, 2017). As much as
the boundaries of journalism have shifted in recent years (Carlson and Lewis, 2015), and as
we show throughout this volume, Diana Bossio and Vittoria Sacco (Chapter 25) illustrate that
when it comes to resolving the conflicts that emerge between people, including journalists, the
unbounded spaces of the internet pose particular challenges. Bossio and Sacco show in par-
ticular where the prevailing norms of journalism no longer control public discourse and how
within the ambient flows of news, defamatory speech sits alongside journalism in sometimes
uncomfortable ways.
These discussions shift our attention towards seeing the spaces of digital journalism as uneven
terrain and reveal where experiences for different actors vary dramatically. Take the activist-
journalists Allissa V. Richardson (Chapter 29) interviewed, who were involved in reporting
on Black Lives Matter but also in mobilizing communities. Here, digital technologies proved
crucial not only for reaching audiences and telling stories but in leveraging mobile technologies
they were able to bring hitherto unheard voices to the public. These opportunities are seized
upon in part out of necessity, as other avenues to be heard in society were not available for those
communicating and engaging with the Black Lives Matter movement (cf. Barnard, 2017; Smit
et al., 2017). In these cases where a lack of journalistic access or opportunity mean critical voices
may go unheard, one-way digital journalism has developed a path forward in the ‘pop-up’ news
ecologies of citizen and activist voices in journalism. Melissa Wall (Chapter 28) describes these
as emerging in contrast to authorities’ own accounts of ‘news’, including under authoritarian
regimes in places like Syria. Placing these cases in dialogue alongside one another paints a pic-
ture of a complex digital landscape of opportunities and challenges but also solutions. Yet there
are limits to the ways voices can subvert oppressive controls. Aras Coskuntuncel (Chapter 38)
has shown that reaching publics, particularly in Turkey, is a path controlled at least in part by
those in power, both commercially and politically, who have ‘re-geared’ their levers of power to
contend with digital forms of protest, activism, and journalism. Certainly, these restrictions are
worrying signs for journalism’s potential, and as Coskuntuncel points out, they are not only in
force in Turkey.
Development in methods
From these discussions, it may seem there is reason for pause in enthusiastic discussions of digi-
tal journalism as reflecting an all-encompassing change in our societies, and, to be certain, some
reflection to weigh the approach research is taking to make sense of digital journalism in light of
6
Introduction
these disparities is warranted (Carlson, 2016: 58). In some cases, we too advocate restraint; however,
within the frameworks advanced in this Handbook we also see where critical research agendas have
been developed in concert with refined methodologies that place these complexities in journal-
ism’s development in a broader societal context. Take, for instance, how ‘post-truth’ and ‘fake news’
ran through public and academic discourses in the past years. While the role of truth and of verac-
ity is evidently critical to journalism research, there seems a risk of research that favors immediacy
over reflection. Within this Handbook we join scholars who consider instead how within the cru-
cibles of innovation, the zeitgeist can be better understood by researching the new ways journal-
ism functions in new spaces within a grounded discussion of journalism’s broader role. From this
perspective, we can see the advantage of innovative research approaches that make sense of digital
journalism in all its complexities and that take into account its ever-changing nature – including
those that focus on the individuals involved in implementing these technologically driven changes.
This comes to our attention when considering news that comes to us through algorithmically tai-
lored presentations. Michael Koliska and Nicholas Diakopoulos (Chapter 19) explore this practice
not through analysis of the software itself but rather through focus groups with those involved in
employing algorithms for news organizations. Here we begin to see where attention paid to the
technology and in particular the way news organizations implement algorithms can bear in mind
the perspectives of those tasked with the implementation of these.
This brings this introduction into a discussion of methods, and in this volume we address
a range of research designs which tackle digital journalism research by deploying multiple
approaches. For as much as our discussion in this introduction has steeped critical engagement
within Digital Journalism Studies around journalism’s traditional contributions, it would be
nearsighted to suggest that the scale and pace of developing technologies have not posed par-
ticular methodological challenges for scholars that cannot be addressed without technological
approaches (Malik and Pfeffer, 2016). As Cornelia Brantner and Jürgen Pfeffer (Chapter 6)
illustrate, when we speak of analyzing content in this environment, for instance, we are often
considering billions of social media posts, and when we speak of audiences, we may be trying to
consider the perspectives and activities of millions of users. Indeed, the ways we assess journalists,
journalism, and their audiences have all been complicated by the ways each of these approach
digital technologies in different ways (Hasebrink, 2016).
These challenges, however, are not insurmountable, and in the work here we are shown paths
to understand journalism in these contexts. The contribution by Rodrigo Zamith (Chapter 7),
for instance, illustrates how computational approaches can ‘freeze’ the flow of news but also
where challenges still exist when designing research around content on regularly changing and
interactive websites. Grappling with changing content is a challenge that stretches across the
news ecosystem online and is the specific focus of Elisabeth Günther, Florian Buhl, and Thorsten
Quandt (Chapter 9), who are able to show how beyond the sheer scale of news available online,
tracking its diffusion can be achieved. These methods may seem blearily complex for those
unfamiliar with computational approaches to research, but we see in the discussions offered by
Zamith and by Günther, Buhl, and Quandt that situated within these approaches are fundamental
concerns of understanding journalism and reasonable ways of addressing them. Indeed, these two
chapters speak to similar concerns of understanding how information changes and the way this
is presented within news stories being told, and further how these changes echo across the larger
digital news environment (Van Hout and Van Leuven, 2016). They also bring our attention to
how specific challenges of research, when the content and technologies are beyond the investiga-
tive lenses of most researchers, can be overcome.
Ike Picone (Chapter 11) poses this as a challenge of confronting digital devices, platforms, and
their affordances, which more often than not are closed off to researchers (Diakopoulos, 2014). In
7
exploring these obstacles, Picone maps the ways we can take advantage of technologies to break
through and address the publics using news technologies. He poses a range of approaches that
may help better understand audiences and publics engaging with news devices. This can help
us critically explore common assumptions about algorithms and devices, including seeing users
among enclaves of similar users and restrained by filter bubbles that limit their awareness of the
world around them (Haim et al., 2017). Jacob Ørmen (Chapter 10) describes this as the ‘myth
of enclaves’, a challenge not to the existence of filtered repertoires of information but rather a
challenge for digital journalism scholars to confront, as there remains so much uncertainty about
what they are and what they look like.
Conclusion: further brightening the corners of Digital Journalism Studies

Across this Handbook we explore new ways of considering change – asking what defines a digi-
tal journalist and what makes his or her work unique, while focusing on the contradictions
that have emerged in the ways these questions have been posed and who they consider. At the
outset, we suggested that development can take shape in different ways and that the coherence
between traditional and new ways of understanding journalism needs to contend with each of
these. Fittingly, Jane B. Singer (Chapter 37) brings this all into focus in her contribution near the
end of this volume by revisiting this particular challenge. Seeing digital journalism as more than
merely the adjectival appendage of ‘digital’ to extant forms of journalism (Eldridge and Franklin,
2017: 1), Singer shows where digital journalism research insists on making sense of a complex set
of new relationships. She calls for a new approach called ‘relationship affects’ that moves beyond
trying to wrap so-called analog approaches of media effects around digital journalism and instead
toward grappling with the challenges of digital journalism and all of its developments. Singer’s is
a call for new approaches, informed by all we know and have known about journalism up until
now, but with a watchful eye toward the greater complexities we have come to realize and may
yet experience. One response to such a call is inscribed in Marcel Broersma’s epilogue, examining
whether journalism scholarship and its focus on journalism in a digital age should be resituated
not to examine the digital in journalism but rather how journalism takes root throughout a digi-
tal media ecology. Inasmuch as the breadth of research in this volume explores digital journalism,
this epilogue suggests that as we go forward, a new appraisal of the ways in which journalism
is woven throughout our experiences with digital technology is needed. Within this Handbook
we see such complexities as both the challenges and the opportunity facing the field of Digital
Journalism Studies as it continues to grow.
In this volume, contributors’ ambitions and our own have been to provide a grounded dis-
cussion for better understanding how any particular aspect of digital journalism that may seem
prominent in moments – ranging from fake news to online trolling to information deserts
and oppressive controls – can be more effectively appreciated by contextualizing such moments
within critical, reflective research. Authors contribute to this discourse by outlining methodolog-
ical approaches to journalism’s modern demands and introducing new agendas for understanding
journalism’s place in society. They also show where the nascent demands for exploring this field
already evident two years ago, when we were finishing the Routledge Companion to Digital Journal-
ism Studies (Franklin and Eldridge, 2017), have developed into a more complex set of dynamics
and where our intimation then that there was something unique occurring around Digital Jour-
nalism Studies has been reconfirmed. The collection here also demonstrates that when making
sense of a quickly moving object of study like digital journalism, the reflective space and pace of
academic scholarship becomes a help, not a hindrance, to offering measured reflections of news
and media in a digital era. Where in the Routledge Companion to Digital Journalism Studies we
8
Introduction
mapped the fundamentals of an emerging field, here we challenge the boundaries of that field by
asking whose voices have been left behind and unpack not only enthusiastic changes in journal-
ism that emerged in the interceding years but also the problems that have arisen. We ask as well
where narratives of digital journalism and technological change have favored enthusiasm over
reflection and in the work of contributors gathered here have introduced that necessary pause for
considering the broad implications of digital change.
Structure
The digital journalist: making news

Part I poses and addresses fundamental questions about journalistic identity, exploring what
shapes journalists’ self-perceptions of their changing roles, values, and relationships with other
actors and actants. Legal identity is particularly in focus to establish who is deserving of ‘addi-
tional’ legal protections reflecting their status as a journalist, which has emerged as a key debate
in Digital Journalism Studies. Further, but no less important, debates over who deserves a byline
in the context of automated robot journalism, how employers perceive the identity of journalists,
and, finally, the precise meaning and implications of ‘entrepreneurial’ in the phrase ‘entrepreneur-
ial journalist’ are examined.
Digital Journalism Studies: research design

Part II presents a group of chapters focusing on the changing methodological requirements for
effective research design in an age of digital media and new tools for digital research, including
data mining software and automated content analysis of journalism texts, that facilitate research
design and inquiry. Chapters here address the challenges of analyzing constantly changing digital
and news content, innovative methods for quantitative analysis of big data, news ‘dissemination’
studies being applied to research of online news networks, methodological approaches for assess-
ing local news provision, and new ways to consider the provision of news via search, as well as
methodological advances for studying news users.
The political economy of digital journalism

Part III considers prominent financial and political concerns present in the variety of approaches
to resourcing, sustaining, and establishing a viable and democratic environment for digital jour-
nalism. Chapters explore the complexities of the debate around liberation and regulation for
digital journalism, including ownership and content regulation of digital media, journalists’ resis-
tance and opposition to imposed precariat status, the commitments and contribution of non-
profit organizations to democratic ambitions for digital journalism, and a reassessment of trends
in readerships, sales, and resources for multiplatform news.
Developing digital journalism practice

Part IV addresses key debates in the development of digital journalism practice, including the
conceptualizing of journalistic uses of big data, the impact of the sustained use of automation in
the newsroom, the significant issue of transparency, rather than objectivity, as a central commit-
ment in journalism in the specific context of algorithms, the central role claimed for networks
in creating news narratives, and critical assessments of the expansive uses of big data. This section
9
introduces measured discussions of each of these aspects to situate their developments within the
broader context of a changing field.
Digital Journalism Studies: dialogues

Part V examines and rehearses significant (sometimes public) dialogues that emerge with height-
ened salience in the context of digital journalism and media – dialogues, for example, about
journalists’ freedom and rights to privacy in the context of the surveillance of journalists post-
Snowden; the right to be forgotten when social media/internet memory seems elephantine; the
apparent right to defame in online settings as well-established legal redress that applies in offline
settings becomes less available; the growth of grassroots movements like the Hacks/Hackers
Global Network, which seeks to create dialogues between journalism and technology to promote
news innovation; and, finally, dialogues about broad ethical concerns triggered by the advent of
algorithmic journalism for both journalistic practice and Digital Journalism Studies.
Minority voices and protest: narratives of freedom and resistance

Part VI offers a platform for voices of protest and resistance, including those rejecting aspects
of the digital world and digital journalism as well as those seizing its technologies as new plat-
forms of resistance. Contributions here include consideration of the impact of mobile technol-
ogy on citizen witnessing for African-American communities in the United States, the role of
computer-assisted cartography and data journalism platforms in producing environmental news,
and challenges faced by women working online in news media, alongside the empowered voices
of oppressed groups using the affordances of digital media and journalism to engage in journalism.
Digital limits: new debates and challenges for the future

Part VII explores what we have called the ‘Digital Limits’ to consider new debates in Digital
Journalism Studies and the challenges for the future. We begin by examining concerns about
personal branding by journalists and media corporations in their uses of Twitter and other social
media and go on to look at the blurred lines between commercial and informative roles in games
journalism. We also explore the increasing uses of native advertising in digital journalism and the
inevitable clash with commitments to transparency, as well as the growing furor about ‘civility’
(including abuse, bullying, and misogyny) in comments sections. The section concludes with
two essays exploring the reconceptualizing of long-established concepts in journalism studies
which have accompanied the development of its digital successor; the first considers and revital-
izes established understandings of the media effects literature, the second offers a national case
study reworking of Herman and Chomsky’s ‘propaganda model’ for the era of digital journalism.
Finally, the Handbook concludes with an epilogue that encourages Digital Journalism Studies
scholars to consider the priority that has been placed on the way digital technologies have been
integrated into journalism, calling for a renewed emphasis on the overall digital ecology, in which
journalism is woven throughout a range of media and networks.
References
Baack, S. (2017) “Practically engaged: The entanglements between data journalism and civic tech.” Digital
Journalism. doi:10.1080/21670811.2017.1375382
Barnard, S. (2017) “Tweeting #Ferguson: Mediatized fields and the new activist journalist.” New Media &
Society. doi:10.1177/1461444817712723
10
Introduction
Benkler, Y. (2011) “A free irresponsible press: Wikileaks and the battle over the soul of the networked fourth
estate.” Harvard Civil Rights-Civil Liberties Law Review, 46, 311–397.
Benson, R. (2006) “News media as a ‘journalistic field’: What Bourdieu adds to new institutionalism, and
vice versa.” Political Communication, 23(2), 187–202.
Broersma, M. and Peters, C. (2016) “Introduction: Towards a functional perspective on journalism’s role
and relevance.” In C. Peters and M. Broersma (eds.), Rethinking Journalism Again: Societal Role and Public
Relevance in a Digital Age. Abingdon: Routledge (pp. 1–18).
Broussard, M. (2015) “Big data in practice: Enabling computational journalism through code-sharing and
reproducible research methods.” Digital Journalism, 4(2), 266–279.
Bruns, A. (2017) “Making audience engagement visible: Publics for journalism on social media platforms.”
In B. Franklin and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon:
Routledge (pp. 325–334).
Carlson, M. (2016) “Establishing the boundaries of journalism’s public mandate.” In C. Peters and M.
Broersma (eds.), Rethinking Journalism Again: Societal Role and Public Relevance in a Digital Age. Abingdon:
Carlson, M. (2017) Journalistic Authority: Legitimating News in the Digital Era. New York, NY: Columbia
University Press.
Carlson, M. and Lewis, S. (2015) Boundaries of Journalism. Abingdon: Routledge.
Conboy, M. (2013) Journalism Studies: The Basics. Abingdon: Routledge.
Costera Meijer, I. and Groot Kormelink, T. (2015) “Checking, sharing, clicking and linking: Changing pat-
terns of news use between 2004 and 2014.” Digital Journalism, 3(5), 664–679.
Dahlgren, P. (2013) “Online journalism and civic cosmopolitanism: Professional vs. participatory ideals.”
Journalism Studies, 14(2), 155–171.
Deuze, M. and Witschge, T. (2017) “Beyond journalism: Theorizing the transformation of journalism.”
Journalism, 19(2), 165–181.
Diakopoulos, N. (2014) Algorithmic Accountability: On the Investigation of Black Boxes. Retrieved from https://
towcenter.org/research/algorithmic-accountability-on-the-investigation-of-black-boxes-2/
Eckert, S. (2017) “Fighting for recognition: Online abuse of women bloggers in Germany, Switzerland, the
United Kingdom, and the United States.” New Media & Society, 20(4), 1282–1302.
Eldridge, S. (2017) “The digital journalist: The journalistic field, boundaries, and disquieting change.” In
B. Franklin, and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon:
Eldridge, S. (2018) Online Journalism From the Periphery: Interloper Media and the Journalistic Field. Abingdon:
Routledge.
Eldridge, S. and Franklin, B. (2017) “Introduction: Defining digital journalism studies.” In B. Franklin and
S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon: Routledge (pp. 1–12).
Franklin, B. (2016) “The future of journalism: Risks, threats and opportunities.” Journalism Studies, 17(7), 798–800.
Franklin, B. and Eldridge, S. (2017) The Routledge Companion to Digital Journalism Studies. Abingdon:
Routledge.
Gasher, M., Brin, C., Crowther, C., King, G., Salamon, E. and Thibault, S. (2016) Journalism in Crisis: Bridging
Theory and Practice for Democratic Media Strategies in Canada. Toronto: University of Toronto Press.
Haim, M., Graefe, A. and Brosius, H. (2017) “Burst of the filter bubble?: Effects of personalization on the
diversity of Google News.” Digital Journalism, 6(3), 330–343.
Hanitzsch, T. and Vos, T. (2016) “Journalism beyond democracy: A new look into journalistic roles in
political and everyday life.” Journalism, 19(2), 146–164.
Hardy, J. (2017) “Money, (Co)production and power: The contribution of critical political economy to
digital journalism studies.” Digital Journalism, 5(1), 1–25.
Hasebrink, U. (2016) “Audiences and information repertoires.” In B. Franklin and S. Eldridge (eds.), The
Routledge Companion to Digital Journalism Studies. Abingdon: Routledge (pp. 364–374).
Hermida, A. and Young, M. (2017) “Finding the data unicorn: A hierarchy of hybridity in data and com-
putational journalism.” Digital Journalism, 5(2), 159–176.
Ksiazek, T. (2016) “Commenting on the news: Explaining the degree and quality of user comments on news
websites.” Journalism Studies, 19(5), 650–673.
Lasorsa, D., Lewis, S. and Holton, A. (2012) “Normalizing Twitter: Journalism practice in an emerging
communication space.” Journalism Studies, 13(1), 19–36.
Lewis, S. (2014) “Journalism in an era of big data: Cases, concepts, and critiques.” Digital Journalism, 3(3),
321–330.
11
Malik, M. and Pfeffer, J. (2016) “A macroscopic analysis of news content in Twitter.” Digital Journalism,
4(8), 955–979.
Schrøder, K. (2017) “Towards the ‘audiencization’ of mediatization research?: Audience dynamics as
co-constitutive of mediatization processes.” In O. Driessens, G. Bolin, A. Hepp and S. Hjarvard (eds.),
Dynamics of Mediatization. Basingstoke: Palgrave Macmillan (pp. 85–115).
Schudson, M. (2017) “Comment: The journalism studies tree.” In P. Boczkowski and C. W. Anderson (eds.),
Remaking the News: Essays on the Future of Journalism Scholarship in the Digital Age. Cambridge, MA: The
MIT Press (pp. 267–272).
Smit, P., Heinrich, A. and Broersma, M. (2017) “Activating the past in the Ferguson protests: Memory work,
digital activism and the politics of platforms.” New Media & Society. doi:10.1177/1461444817741849.
Steel, J. (2016) “Reappraising journalism’s normative foundations.” In C. Peters and M. Broersma (eds.),
Rethinking Journalism Again: Societal Role and Public Relevance in a Digital Age. Abingdon: Routledge (pp.
35–49).
Thorsen, E. (2017) “Whistleblowing in a digital age: Journalism after Manning and Snowden.” In B. Frank-
lin and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon: Routledge
(pp. 569–578).
Van Hout, T. and Van Leuven, S. (2016) “Investigating ‘churnalism’ in real-time news.” In B. Franklin
and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon: Routledge (pp.
117–125).
Willams, A., Harte, D. and Turner, J. (2015) “The value of UK hyperlocal community news: Findings from
a content analysis, an online survey and interviews with producers.” Digital Journalism, 3(5), 680–703.
12
PART I
The digital journalist

Making news
1
LAW DEFINING JOURNALISTS
Who’s who in the age of digital media?
Like journalism, the law is a profession built around language. This chapter focuses on one criti-
cal example of how legal institutions and lawmakers globally are grappling with language as they
quite literally redefine the terms ‘journalist’, ‘journalism’ and ‘news media’. Where, not so long
ago, journalists, and undoubtedly members of the public, could confidently assert “we all know
what a journalist is, and it’s silliness to argue about it” (cited in Ugland and Henderson, 2007:
242), few journalists would argue this today. Certainly, the legal profession has moved beyond
any simple understanding. Instead, as the New Zealand Law Commission (NZLC) has pointed
out, law and policymakers (including judges, politicians, and public administrators) are being
challenged to adequately and fairly summarize journalism and news media positioned within a
“digital ecosystem [in which] there is a growing symbiosis between new and old media” (Law
Commission, 2013: 54). As it explains (2013: 29):
In the pre-digital era identifying the target of media regulation and determining the
boundaries of intervention were relatively straightforward matters. However, deter-
mining what to regulate and how to calibrate, target, and enforce that regulation has
now become far more complex as bright line distinctions . . . become increasingly
blurred.
Within its remit, the NZLC was tasked with determining how to define ‘news media’ for the
purposes of the law. Other countries, including Australia, Canada, the United States, and Britain,
have also been compelled to face the questions of who and what should be afforded privilege and
responsibilities once provided to the more predictable institutions in newspaper, radio, television,
and other ‘mainstream’ news media and their employees. The complexity of determining clearly
worded definitions and the rifts that have resulted are highlighted in the debate that surrounded
the Free Flow of Information Act (2013) in the United States, which sought to ensure journal-
ism’s claim to a free flow of information within federal law. This debate provides an example of
the deep division surrounding the definition and its impact on the legal process. Central to the
division is whether bloggers and new media users – as one scholar writes, “bloggers, dilettantes,
and do-it-yourselfers” (Ugland and Henderson, 2007: 241) – should be part of the definition;
whether “A-list” (Singer, 2007: 80) bloggers might be included; or, whether definitions should
15
be limited to those who are in employed or paid work within a recognized media organization,
have completed media training, and adhere to a professional code of ethics.
This chapter will examine the key arguments raised in relation to this vexed issue, drawing
from legal and media scholarship as well as legal practice, political and policy environments, and
case law. It will provide examples of new definitions and, importantly, the contexts in which they
have been framed. For journalism practice, the issues and challenges faced by lawmakers and,
ultimately, the definitions that emerge from their determinations will provide important sign-
posting and some clarity for the future direction of the news media and its workers as definitions
are adopted and enshrined in law and policy.
Approaches to deciding ‘Who is a journalist?’

The questions of ‘Who is a journalist?’ and ‘What is the news media?’ have been closely examined
from a variety of perspectives by media and legal scholars who have sought to better understand
how changing practices and media ecosystems translate into new definitions and clearer under-
standing (see, for example, Alonzo, 2005; Gleason, 2015; Johnson and Kaye, 2004; Johnston and
Graham, 2013; Johnston and Wallace, 2017; Reese et al., 2007; Shennan, 2011; Singer, 2007;
Ugland and Henderson, 2007; West, 2014). Many of these arguments will be examined in other
chapters in this book, so we will focus our attention on the development of legal contexts and
frameworks. Shennan (2011) has criticized a so-called divide in scholarship between the sociol-
ogy of journalism on the one hand and the law on the other; a ‘dis-united’ approach that arguably
has undermined a deeper understanding of the changing shape of journalism and the media. We
see the inclusion of this chapter in a journalism collection as evidence of a more united approach,
incorporating a diverse range of perspectives and literature.
Peters and Tandoc (2013: 39) identify the following indicators in determining who is and is
not a journalist: medium, activities, output, employment, social roles, and intent. Ultimately, their analy-
sis of journalism gives rise to the following definition:
A journalist is someone employed to regularly engage in gathering, processing, and dissem-

inating (activities) news and information (output) to serve the public interest (social role).
(Peters and Tandoc, 2013: 61)
We agree that such a definition provides a strong conceptual base; however, it also presents sig-
nificant inherent problems by incorporating elements that are themselves in need of definition –
notably ‘news’ and ‘the public interest’. So, while the definition adds to the discursive examination
of who is a journalist and what is media, it also raises questions and may be of limited application
in practice. Alternately, the NZLC’s analysis provides a more pragmatic response to the question,
proposing the following criteria for defining news media:
• a significant element of their publishing activities involves the generation and/or aggrega-
tion of news, information, and opinion of current value;
• they disseminate this information to a public audience;
• publication is regular and not occasional; and
• the publisher must be accountable to a code of ethics and to the NMSA.1
(2013: 16)
In calling for a uniform approach to all New Zealand statutes that provide privileges or exemp-
tions relating to the news media, they note that most of the existing legislation of this type was
drafted in the pre-digital era, when the inclusion of such definitions was not considered necessary
16
Law defining journalists
(2013: 34). This observation speaks to the essence of why legal definitions have emerged as neces-
sary in the past few decades and why they deserve close consideration. While journalism and the
news media have, by necessity and due to their daily practice, become both highly informed and
articulate about the changed media landscape, changes to legal statutes and policy and the language
needed to drive this have followed more slowly, in part because it is not central to the day-to-day
practice of the law. As a result, attention to definitions has only recently gained a foothold in legal
literature, policy, and legislative development as an issue for policy-makers, regulators, and courts.
Historically, as Shennan notes, the rise of mass communication resulted in “a concept of
journalism, however defined” that was to claim distinction from other forms of communication
(2011: 134). However, in time, “[t]he gradual accumulation of legal rights attaching to journal-
ists [. . .] left unanswered the question of who can call himself or herself a journalist” (Shennan,
2011: 135). And, as the necessity to answer that question became more pressing, with digital
media channels increasingly enabling most anyone to publish online, lawmakers have been forced
to consider more closely who and what practices should be protected and privileged. As Shennan
further observed (2011: 134), the evolving definition of journalist in Europe is therefore a recent
trend, emerging in cases where legal protection has been in question. We now move to some of
the specific legal domains in which this has occurred, both in Europe and elsewhere.
Legal domains and definitions

As we have previously noted (Johnston and Wallace, 2017), there are two principal ways in which
journalists, however defined, come into contact with the legal process. The first of these occurs
when a journalist is part of the process, most commonly as a defendant or other actor in a legal case –
for example, seeking shield law protection for confidential sources; the second type of contact
occurs when a journalist is observing or reporting on the process, performing a court-reporting role.
This chapter builds on these two intersections, expanding earlier work to consider other contexts
in the first category, where the law is considering how to define journalists in cases of national
security and privacy laws. We find that the question of defining journalism – and its inherent
challenges – has become a major focus for politicians, law and policymakers, judges, and legal
commentators in many countries. We examine examples, drawn from several countries, that are
intended to provide contextualized understanding of the issue. These are in no way intended to
suggest uniform debate or legislative change; rather, they are to provide insights into the complex
and varied discussion and debate that surround the issue of defining journalist and news media.
Journalists as participant
Shield laws
In the United States, where shield laws (or similar protections) exist in 49 states (Johnston and Wal-
lace, 2017), it has been said that they provide “almost absolute protection from prosecution for
contempt in situations where an individual refuses to disclose their confidential sources, documents
or other information that could identify those sources” (Shennan, 2011: 133). In fact, the degree of
protection varies; some states confer an absolute privilege, in some it is qualified, and other shield laws
contain exceptions that remove the shield in certain circumstances (Reporters Committee for Free-
dom of the Press, n.d.). Other countries – New Zealand, Britain, and some Australian jurisdictions –
also have shield protections. Our particular interest in shield laws rests with the intersection these
laws bring to definitions, and the way changing media practices have impacted within this space.
The debate around the U.S. Free Flow of Information Act (2013) is instructive for the breadth
of analysis on this issue. First proposed in 2009 as a federal shield law to provide journalists with
17
a right to refuse to testify in criminal proceedings about confidential sources (Peters and Tandoc,
2013), the bill was negotiated between a coalition of 60 media organizations and the U.S. fed-
eral government. In its original form, it defined a journalist as someone entitled to invoke legal
protection, with:
“primary intent to investigate events and procure material”, to inform the public by
“regularly” gathering information through interviews and observations, and then dis-
seminating that information to the public. In addition, the person must intend to dis-
seminate the information at the start of the newsgathering process.
(Peters and Tandoc, 2013: 37)
However, the legislation stalled at the definitional stage (Johnston and Wallace, 2017; Durity,
2006; Greene, 2013). On one hand, the draft form of the legislation was said to grant a special
privilege to people who were not “real reporters” and had no professional qualifications; on the
other hand, it was argued that a more inclusive definition was needed to encompass new media
workers, people “who do real journalism in different ways” (Peters and Tandoc, 2013: 38). A
compromise was sought, with an amendment defining a journalist as a person employed by, or
in contract with, a news organization for a designated period of time; who had substantially
contributed to a publication as an author, editor, photographer, or producer; or was a journalism
student (Free Flow of Information Act, 2013). However, this has yet to be enacted. Thus, the
question of definition was a fundamental component of the lengthy discussion and debate and
ultimately became a stumbling block for the legislation itself; a conceptual issue that called for a
pragmatic determination was not resolved.
Canada, possibly the most recent country to move to implement shield laws and journalistic
definitions, did not have the same difficult in determining a definition. Bill 231, the Journalist
Sources Protection Act, was introduced to Federal Parliament in late 2016 in the wake of rev-
elations that the Montreal police and the Sûreté du Québec had been conducting surveillance
operations on various journalists and media organizations (Bellavance, 2017). The Canadian
Senate adopted the amendment in April 2017 prior to the bill going to the House of Commons,
where it was passed in October 2017.
During the consultation process the news media strongly supported the proposed changes,
arguing that they were long overdue, listing those countries with shield laws in place, including
the United States at state level.
This privilege is [. . .] explicitly recognized by specific laws in Australia, Argentina,

Germany, Belgium, El Salvador, France, Great Britain, Mexico, Norway, New Zealand,
Sweden, and Switzerland, to name but a few. Privilege even exists in the legislation of
countries known to have a more tense relationship with the media, such as Russia.
(Cooke in Bellavance, 2017: 5)
A report tabled to the Senate Committee by the Canadian Media Coalition went to great lengths
to examine the definition of journalist for the purposes of this protection, opting for a broad
definition in the following terms, which was adopted by the drafters of the legislation:
journalist means a person who, in connection with his or her primary paid occupation,
contributes or contributed directly and regularly or occasionally, to the collection, writ-
ing or production of information, editorials or columns for dissemination to the public
by the media, or anyone who assists such a person.
(2017: 7)
18
The coalition rejected the idea that a broad definition of journalist could be used to shelter from
prosecution people who used social media or websites occasionally, noting that the “definition
limits its application to individuals who are unquestionably career journalists. Moreover, it must
be borne in mind that a judge could always refuse to apply the definition in borderline situa-
tions or clear cases of abuse” (2017: 7). Because it was difficult to accurately foresee all possible
scenarios, it recommended adopting a clause that would give a judge discretion to “recognize
that a person was acting as a journalist in a specific instance, notwithstanding that the person in
question may not qualify under the general definition” (2017: 7). In short, it was argued:
If a person does not qualify as a journalist under the foregoing definition, that person
may nevertheless be recognized as being a journalist within the meaning of the Act if it
is demonstrated that the person has the usual characteristics of a journalist.
(Canadian Media Coalition, 2017: 7)
However, the proposed act does not incorporate this broader provision.
Shield laws have also provided the New Zealand courts with a reason to consider how broadly
to define a journalist and media organization. In the case of Slater v Blomfield the New Zealand
High Court had to determine whether a blogger was entitled to the “journalist privilege” con-
tained in the New Zealand Evidence Act, which contained the following definitions:
journalist means a person who in the normal course of that person’s work may be given
information by an informant in the expectation that the information may be published
in a news medium;
news medium means a medium for the dissemination to the public or a section of the
public of news and observations on news.
(Evidence Act, 2006, NZ)
In deciding that Slater, the author of the blog Whale Oil, qualified as a journalist under this defini-
tion, on the basis that he worked for his blog, Justice Raynor Asher noted that the definition did
not impose quality requirements, nor did it require the dissemination of news to be in a particular
format: “Slater’s reports contain genuine new information of interest over a wide range of topics . . .
while criticisms can be made of Mr Slater’s style and modus operandi, [his blog] . . . is not of such
low quality that it is not reporting news” (Slater v Blomfield, 2014: par 62). In his finding, the judge
strongly emphasized the policy reason behind the journalist privilege: “It is to promote the free
flow of information, a vital component of any democracy” (Slater v. Blomfield, 2014: para 136).
Australian shield laws, in the Commonwealth and a number of States and Territories, gener-
ally contain broad definitions similar to section 68 of the New Zealand Evidence Act (Evidence
Act, 1995 (Cth) s 126J, Evidence Act, 1995 (NSW) s 126J, Evidence Act 1906 (WA) s 20G). Nota-
bly, the Victorian legislation also requires the court to consider, in determining whether someone
is a journalist, whether “the person or the publisher of the information, comment, opinion or
analysis is accountable to comply (through a complaints process) with recognized journalistic or
media professional standards or codes of practice” (Evidence Act, 2008 (VIC) section 126J). The
ethical obligations associated with journalism have been considered a relevant factor in at least
one case (Madafferi v Age Company Ltd [2015] VSC 687: para 127). However, the extent to which
that may be an influential factor in interpreting the definition of journalist has not yet been tested
by any individual not employed by a traditional media organization.
As Australian legislators were rethinking the definition of journalist in the context of shield
laws, they took a different approach when considering the field of national security and terror-
ism law.
19
National security and terrorism law

Journalistic freedoms have been challenged by restrictions imposed in the name of national
security following the terrorist attacks on the World Trade Center in New York in September
2001 (Johnston and Pearson, 2008). New legislation potentially impacting journalists has seen
challenges to:
chasing security-related stories, by exposing them to detention and questioning, bug-

ging their communication, seizing their notes and computer files, breaching the confi-
dence of their sources, banning them from covering some court cases, suppressing facts
in trials, banning them from some newsworthy locations, rendering discussions with
some sources illegal, and restricting their publication of quotes from sources deemed to
be encouraging terrorism.
(Johnston and Pearson, 2008: 80)
As time has moved on, legislators have needed to consider how journalists are defined for the
purposes of this legislation. For example, the Australian Parliament’s Review of Security and Counter
Terrorism Legislation in 2006, included terms such as ‘media interest’, ‘media coverage’, and ‘media
bias’, without defining them. In contrast, less than a decade later, in 2014, when the Australian
Government reviewed its National Security Legislation Amendment Bill (No. 1), it spent consider-
able time on the issue of definitions – ultimately deciding that the difficulty of defining jour-
nalism made it inadvisable to provide an exemption for journalists from the proposed offense
provisions. It stated:
The Committee considers that it would be all too easy for an individual, calling them-
selves a ‘journalist’, to publish material on a social media page or website that had seri-
ous consequences for a sensitive intelligence operation. It is important for the individual
who made such a disclosure to be subject to the same laws as any other individual.
(Parliament of Australia, 2014: 62)
The failure to grapple with the issue of definition was implicit in media criticism that the review
had failed to provide avenues for legitimate journalistic activity in reporting terrorism, effectively
relegating journalists to anyone status and providing no defined privilege. Notably, the journalists’
union, the Media Entertainment and Arts Alliance (MEAA), responded that the review did not
address issues raised about the role of journalists or whistle-blowers (Warren, 2014).
It also calls to mind West’s distinction between professional and amateur journalists, or what
she calls “press-like” others, and her caution that:
Pretending that the press is no different than an army of individual speakers with
megaphones is a dangerous road to travel [. . . ]. Treating the press like all other speak-
ers obstructs the public’s right to know and impedes an important check on the
government.
(West, 2014: 2445)
A comprehensive discussion of anti-terrorism laws is outside the scope of this chapter; however,
the Australian example is instructive of how legislators may choose to overtly avoid definitions,
even arguing strongly against attempting them. Ironically, then, the increasing attention given
by lawmakers to terms such as journalist, journalism, or news media, once assumed and under-
stood as not in need of definition, may not result in any actual clarification of how they are to
20
be understood for the purposes of legislation. We now consider this in another legal context,
in which the issue of journalistic definition has come under focus – that of privacy and data
protection law.
Privacy and data protection

Since the 1970s, concerns about infringement on personal privacy associated with the growth of
information technology, increased government powers, and the development of new and more
intrusive business practices, such as credit reporting, direct marketing, and electronic surveillance,
have prompted consideration of new forms of legislation designed to protect privacy rights in
many Western countries (ALRC, 1983: xli–xliii). An examination of some more recent legislative
initiatives directed to data protection illustrates a general lack of clarity as to how exemptions for
privacy legislation, designed to assist the legitimate reporting of news in the public interest, will
be applied in the age of ‘new media’.
For example, section 32(1) of the UK Data Protection Act1998 provides that the processing of
data will be exempt from the restrictions under the Act where several conditions are satisfied,
the first being:
(a) the processing is undertaken with a view to the publication by any person of any journalis-
tic, literary or artistic material.
‘Journalistic’ is not defined in the legislation and does not appear to have been the subject of any
court interpretation, so it is unclear whether, for example, the activities of a blogger would enjoy
exemption from the Act.
Canada has also legislated privacy rights at the federal level. The Canadian legislation exempts
information collected for ‘journalistic’ purposes (Personal Information Protection and Electronic Docu-
ments Act, 2000 (Canada); Privacy Act, 1985); however, like the UK, ‘journalistic’ is not defined.
In a recent 2017 Canadian case, the Federal Court endorsed a broad approach by the Office of
the Privacy Commissioner, adopting a 2012 definition by the Ethics Advisory Committee of the
Canadian Association of Journalists. The case of A.T. v. Globe24h.com set out three criteria to be
satisfied for an activity to be classified as journalism:
1 Purpose: An act of journalism sets out to combine evidence-based research and verification
with the creative act of storytelling. Its central purpose is to inform communities about top-
ics or issues that they value.
2 Creation: All journalistic work – whether words, photography or graphics – contains an
element of original production.
3 Methods: Journalistic work provides clear evidence of a self-conscious discipline calculated
to provide an accurate and fair description of facts, opinion, and debate at play within a
situation.
(Ethics Advisory Committee, 2012)
Australian privacy legislation also uses a functional definition, focusing on the activities car-
ried out. It provides, in effect, that the legislation does not apply to acts and practices of media
organizations undertaken ‘in the course of journalism’ at a time when that organization has a
public commitment to privacy standards (Privacy Act, 1988 (Cth)). Journalism is not defined in
the Act; however, ‘media organization’ is defined broadly as “an organisation (which includes an
individual) that collects, prepares or disseminates to the public, news, current affairs, information
or documentaries, or commentaries and opinions on, or analyses of, such material” (Privacy Act,
21
1988, section 6(1)). As Waters (2002) notes, the breadth of this definition appears to be inten-
tionally designed to include, for example, issues-based community groups.
The Australian Law Reform Commission noted concerns that the scope of the definition of
news activity, particularly when coupled with the absence of a definition of journalism, could
enable any individual to claim the exemption by setting up a “publishing enterprise” (ALRC,
2008: para 42.26; f34). It recommended that the legislation be amended by including a defini-
tion of journalism to more clearly limit the scope of the exemption to acts and practices that are
associated with a clear public interest in freedom of expression and focus on the character of the
publication rather than the nature of the organization carrying it out, and it proposed the follow-
ing criteria for journalistic activity:
The collection, preparation for dissemination or dissemination of the following material

for the purpose of making it available to the public:
(a) material having the character of news, current affairs or a documentary;

(b) material consisting of commentary or opinion on, or analysis of, news, current
affairs or a documentary; or
(c) material in respect of which the public interest in disclosure outweighs the public
interest in maintaining the level of privacy protection afforded by the model Uni-
fied Privacy Principles.
The criteria recognized the difficulty of confining ‘journalism’ to traditional types of news dis-
semination activities, while nevertheless defining “news organisation” as “an organisation whose
activities consist of or include journalism” (ALRC, 2008, Recommendations 42–1 and 42–2). To
date, those recommendations have not been adopted by government.
New Zealand has also taken a broad approach to interpreting news activity that is exempt
from restrictions in that country’s privacy legislation (Roth, 2010: 555–557; Law Commission,
2011). In a 2011 report, the New Zealand Law Commission decided against attempting a more
specific definition of that term, deciding that it was better to leave it to be decided on a case-
by-case basis (Law Commission, 2011: para 4.30.) However, concern about the proliferation of
unregulated forms of new media led it to recommend including a definition of ‘news medium’
so that media that are not subject to a code of ethics that deals expressly with privacy and to a
complaints procedure administered by an appropriate body would not enjoy the protection of
exemption from the Privacy Act (Law Commission, 2011). These recommendations, also, have not
been adopted by government.
A very different approach has been taken in Europe to the interpretation of the 1995 EU
Directive on data protection (European Parliament, 1995: Article 9), which exempts from its
provisions the processing of personal data that “is carried out solely for journalistic purposes
or the purpose of artistic or literary expression”, where the derogation is “necessary to recon-
cile the right to privacy with the rules governing freedom of expression”. In a 2008 case, the
European Court of Justice adopted a wide approach to defining “journalistic purposes”, hold-
ing that it depends on whether “the sole object of processing the personal data is to disclose to
the public information, opinion or ideas” (Satakunnan Markkinapörssi Oy and Satamedia, cited in
Roth, 2010: 553). The exemption can apply to individuals as well as organizations and does not
depend on the medium of publication (for example, newspaper, radio, internet) (cited in Roth,
2010: 553–554). In an earlier case, the court noted that the drafters of this provision deliberately
chose not to use wording that restricted its application to institutional media and declined to
adopt definitions based on the medium of publication or whether or not there was inappropriate
editorial oversight involved (Tietosuojavaltuutettu v. Satakunnan Markkinapörssi Oy and Satamedia,
22
Oy C-73/07, 2008). It further found that the essential question was whether the disclosure of
personal data aimed to impact information and ideas on matters of public interest.
The Swedish Supreme Court has adopted a similarly broad definition, holding that the
exemption could apply to material published by persons who were not professional journalists or
part of the established media. It focused on the purpose of journalism, that is: “to inform, exer-
cise criticism and provoke debate about societal questions that are of larger significance for the
general public” (Bygrave, 2001: para 8). So, for example, a member of the public who published
a website of material to further a campaign about alleged malpractice in the Swedish banking
system has been found to be entitled to the exemption (Bygrave, 2001).
We now turn to an entirely different context in which the law and journalism intersect: the
role of the journalist as observer – the court reporter.
Journalist as observer – court reporting

The court reporter has traditionally provided ‘a window’ into the courtroom; a space that is
(in most cases) open to the public, but that, in practice, few individuals attend. Introduction of
technology in courts and the development of digital media practice have caused the courts to
reevaluate how court reporters are recognized and privileged. In a world where, as the Chief
Justice of Canada has pointed out, “Anyone with a keyboard and access to a blog can now be a
reporter” (McLachlin, 2012: 33), courts have had to decide who should be permitted to tweet or
blog from court and in what circumstances. In dealing with this issue, courts are devising criteria
and processes to manage how journalists will be recognized (Wallace and Johnston, 2015).
Much attention to the issue has focused on the use of live text-based communication (LTBC)
from court, with courts accepting that tweeting and blogging were now part of reporting. But
who should have access? In the United States and United Kingdom, developments began in
2010. In that year the Massachusetts Supreme Court restricted the use of electronic devices in the
courtroom to journalists registered with the court’s public information office (PIO) (Davidow,
2010). This policy enabled the PIO to assess whether the organization or individuals “regularly
gather, prepare, photograph, record, write, edit, report or publish news or information about
matters of public interest for dissemination to the public in any medium, whether print or elec-
tronic” (Supreme Judicial Court Rule, 1.19, 2012: part 2). This broad approach focused on the
nature of the function performed rather than the employment relationship or the status of the
organization. It was also concerned with ensuring that those so accredited familiarize themselves
and comply with limitations imposed by the court (Supreme Judicial Court Rule, 1.19, 2012:
part 2). Other U.S. states that have allowed LTBC, including California, Colorado, Florida, Iowa,
Kansas, Michigan, Pennsylvania, and Washington, DC, also require journalists to seek access via
the PIO or judge. Requests should outline:
• The nature and public benefit of the blogging or tweeting;

• The public interest in the case;
• The reporter’s professional experience or credentials; and
• An explanation of how the technology works
(Digital Media Law Project, n.d.)
The United Kingdom followed a similar line. The Chief Justice of England and Wales opened a
consultation process on the issue in 2010, commenting that:
Non-accredited members of the media cannot be presumed to have the same apprecia-
tion of the legal framework surrounding court reporting, or the industry standards set
23
out by the Press Complaints Commission, as accredited media representatives. . . . The

combination of instant reporting without the self-restraint presumed to be exercised by
accredited members of the media might lead to a great likelihood of prejudicial report-
ing, and must be considered.
(Judge, 2010: 11)
The guidance that followed the consultation allowed lawyers and representatives of the media
the automatic right to LTBC from court, whereas members of the public are required to seek
permission (Judge, 2011). While ‘representative of the media’ is not defined, the presumption
“that a representative of the media or a legal commentator using live, text-based communica-
tions from court does not pose a danger of interference to the proper administration of justice in
the individual case” (Judge, 2011: para 10) suggests a restriction to representatives of traditional
media, who can be assumed to understand the laws of contempt and other restrictions on court
reporting. A Scottish review of this issue subsequently recommended a similar approach to Eng-
land and Wales (Judicial Office of Scotland, 2013: 3; Dorrian, 2015: 30–31).
Canadian courts, except for the province of Quebec, have generally followed the lead of Eng-
land and Wales in allowing the media, along with lawyers, to use LTBC from court (Canadian
Centre for Court Technology, 2013). There also appears to be a general trend to require media
to be accredited or approved by the court for this purpose (Canadian Centre for Court Technol-
ogy, 2013).
Some Australian courts have also adopted accreditation policies for court reporters. For example,
the Supreme Court of Victoria allows LTBC from court by regular court journalists accredited
by the court (Supreme Court of Victoria, 2016). Accreditation requires a form signed by the
court’s media team (journalists and the news director or chief of staff ), implying employment by a
traditional media organization. It signifies “that the journalist has an understanding of court pro-
ceedings, legislative requirements, suppression orders and the procedures and policies as set out in
the court’s Media Policies and Practices document” (Supreme Court of Victoria, 2016). A similar
approach has been adopted in Queensland. From 2014, the Queensland Courts allowed journal-
ists employed by media organizations and accredited by the courts to use LTBC in the courtroom
(de Jersey, 2014, Part A Schedule, para 8). Other individuals can apply for accreditation if they can
“identify any ethical code of conduct to which the applicant subscribes as a journalist” (para 66),
a requirement that, in effect, appears to restrict the policy to traditional media. In 2017, the policy
was tightened by introducing an accreditation pass with photo identification and accreditation
clearance, following issues that arose during several high-profile trials. The pass – or Queensland
Courts security access card – can be issued to applicants who are regular or occasional journalists in
the courts and, like Victoria, signed off by the employer organization. This move further consoli-
dates access to mainstream media – dedicated court reporters, or those trained in the field – who
understand the laws of contempt and other related laws and protocols (Queensland Court, 2017).
In contrast, courts that do not hear criminal cases or rely on juries are less concerned by
accreditation for LTBC usage. The Federal Court of Australia, for example, does not prohibit
any person from using electronic devices to transmit text from court, provided it does not create
a disturbance, cause concern to a witness or other court participant, or result in a person out-
side the courtroom receiving information about the proceedings to which they are not entitled
(Federal Court Rule, 6.11(3)). The UK’s highest appeal court, the Supreme Court, has taken a
similarly liberal approach because the concerns about reporting in ways that could prejudice a
trial do not arise (Luft, 2011). Canadian appeal courts in Nova Scotia and British Columbia
also allow anyone to tweet and blog from court (Canadian Centre for Court Technology, 2013).
Nova Scotia has extended this policy to all court levels (The Courts of Nova Scotia, 2014), as has
the Hong Kong Judiciary (Ma, 2014).
24
Conclusions
The past decade has seen unprecedented attention given to the questions of ‘Who is a journal-
ist?’ and ‘What is the news media?’ Far from being “silly” (in Ugland and Henderson, 2007: 242)
or taken for granted as they may have been in the past, these questions have become challenging
and vexing for courts, legislators, policymakers, and media organizations. While courts and other
authorities must get on with the business of doing and managing the law, and in doing so try to
keep abreast of technological change, the practical ends and the functional tests need to be con-
sidered within the deeper conceptual elements; that complex layer that has emerged within the
new digital media world. And so, while legal definitions may present pragmatic solutions, they also
need to be flexible and reflexive enough to respond to and reflect constantly changing contexts.
Ultimately, media coalitions and media organizations (old and new) as well as policymakers and
legislative developers all bring their own language, understanding, experience, and contexts to the
changing definitions of ‘Who is a journalist?’ and ‘news media’ in the digital age.
Further readings
The literature and documentation that has informed this chapter is a diverse mix of journalism,
media and legal scholarship, legislation, case law, and policy, drawn from many countries. We
find that not only are questions surrounding ‘who is a journalist’ and ‘what is the news media’
mercurial and complex, but the diversity of views from different disciplines and countries pro-
vides a dialectic that moves the discourse in various directions. For more on the issues raised in
this chapter, Tim Gleason’s 2015 article “If We Are All Journalists, Can Journalistic Privilege
Survive?” in Javnost – The Public, Jonathan Peters, and Edson C. Tandoc’s 2013 article “People
Who Aren’t Really Reporters at All, Who Have No Professional Qualifications,” in New York
Journal of Public Policy Quorum, and Sonja R. West’s 2014 article in the Harvard Law Review, “Press
Exceptionalism,” offer extended discussions of the debates in this chapter.
Note
1 The Law Commission proposed a News Media Standards Authority (NMSA) which would assume the
collective functions of the Press Council, the Broadcasting Standards Authority (BSA), and the Online
Media Standards Authority (OMSA) (Law Commission 2013: 15).
References
Alonzo, J. (2005) “Restoring the ideal marketplace: How recognizing bloggers as journalists can save the
press.” NYU Journal of Legislation and Public Policy, 9, 751–780.
A.T. v. Globe24h.com, 2017, FC 114 (CanLII). Retrieved from http://canlii.ca/t/gx6b
Australian Law Reform Commission. (1983) Privacy: Report No. 22. Canberra: Australian Government
Publishing Service.
Australian Law Reform Commission. (2008) For Your Information: Australian Privacy Law and Practice: Report
No. 108. Canberra: Australian Government Publishing Service.
Bellavance, J.D. (2017, February 16) “Major media want legislation to protect journalistic sources.” La Presse+.
Retrieved from http://plus.lapresse.ca/screens/295c150f-70b6-4d42-ba50-d51dd35a9898%7C_0.html
Bygrave, L. A. (2001) “Balancing data protection and freedom of expression in the context of website
publishing – Recent Swedish case law.”Privacy Law and Policy Reporter,8(4),83. Retrieved from www.austlii.
edu.au/cgi-bin/sinodisp/au/journals/PrivLawPRpr/2001/40.html?stem=0&synonyms=0&query=
cth%20consol_act%20pa1988108%20s7b#fn2
Canadian Media Coalition. (2017) Brief Submitted to the Special Consultations and Public Hearings on Bill S-231,
to Act to Amend the Canada Evidence Act and the Criminal Code (Protection of Journalistic Sources) Presented
to the Standing Committee on Legal and Constitutional Affairs, February 15, 1–12.
25
Canadian Centre for Court Technology. (2013) Canada-Wide Summary of Court Policies on Live Text-Based
Communications From the Courtroom. Retrieved from http://wiki.modern-ourts.ca/images/5/57/Policies_
on_Live_Text_Based_Communications.pdf
Data Protection Act 1998 (UK).
Davidow, J. (2010, December 10) “Massachusetts to allow live twittering, blogging in courts.” Mediashift.
Retrieved from www.pbs.org/idealab/2010/12/massachusetts-to-allow-live-twittering-blogging-in-
courts337/
de Jersey, P. (2014) Practice Direction No. 8 of 2014, Electronic Devices in Court. Supreme Court of Queensland.
Retrieved from www.courts.qld.gov.au/__data/assets/pdf_file/0004/225553/sc-pd-8of2014.pdf
Digital Media Law Project (n.d.) Live Blogging and Tweeting From Court. Retrieved from www.dmlp.org/
legal-guide/live-blogging-and-tweeting-from-court
Dorrian, L. (2015) Report of the Review of Policy on Recording and Broadcasting of Proceedings in Court, and the
use of Live Text Based Communications from Court. Available at: www.scotland-judiciary.org.uk/25/1369/
Report-of-the-Review-of-Policy-on-Recording-and-Broadcasting-of-Proceedings-in-Court-and-Use-
of-Live-Text-Based-Communications.
Durity, L. (2006) “Shielding journalist-‘Bloggers’: The need to protect newsgathering despite the distribu-
tion medium.” Duke Law & Technology Review, 5, 1–19. Retrieved from http://scholarship.law.duke.edu/
dltr/vol5/iss1/8/
Ethics Advisory Committee of the Canadian Association of Journalists. (2012) What Is Journalism? Retrieved
from http://caj.ca/images/downloads/Ethics/caj_whatisjournalism.doc_1.pdf
European Parliament. (1995) Directive 95/46/EC of the European Parliament and of the Council of 24 October
1995 on the Protection of Individuals With Regard to the Processing of Personal Data and on the Free Movement
of Such Data (OJ No L 281, 23 November 1995, 31).
Evidence Act 1906 (WA).
Evidence Act 1995 (Cth).
Evidence Act 1995 (NSW).
Evidence Act 2006 (N.Z.).
Evidence Act 2008 (VIC).
Federal Court of Australia, Federal Court Rules 2011 (Cth).
Free Flow of Information Act 2013 (USA) S.987 (1113th).
Gleason, T. (2015) “If we are all journalists can journalistic privilege survive?” Journal of the European Institute
for Communication and Culture, 22(4), 357–386.
Greene, D. (2013, September 20) “State revises media shield law for the better, but it’s still imperfect.”
Electronic Frontier Foundation. Retrieved from www.eff.org/deeplinks/2013/09/senate-revises-media-
shield-law-better-its-still-imperfect
Johnson, T. and Kaye, B. (2004) “Wag the blog: How reliance on traditional Media and the internet influ-
ence credibility perceptions of weblogs among blog users.” Journalism & Mass Communication Quarterly,
81(3), 622–642.
Johnston, J. and Graham, C. (2013) “Shifting patterns in Australian newspaper writing styles: Results of a
longitudinal study.” Australian Journalism Review, 35(2), 117–131.
Johnston, J. and Pearson, M. (2008) “Australia’s media climate: Time to renegotiate control.” Pacific Journal-
ism Review, 14(20), 72–88.
Johnston, J. and Wallace, A. (2017) “Who is a journalist? Changing legal definitions in a de-territorialised
media space.” Digital Journalism, 5(7), 850–867.
Judicial Office of Scotland. (2013) Cameras and Live Text-based Communication in the Scottish Courts: A Con-
sultation. Retrieved from www.scotland-judiciary.org.uk/24/1159/Consultation-on-cameras-in-court
Law Commission. (2011) Review of the Privacy Act 1993: Review of the Law of Privacy Stage 4. Report no. 123,
Law Commission, New Zealand, Wellington. Retrieved from www.lawcom.govt.nz/sites/default/files/
projectAvailableFormats/NZLC%20R123.pdf
Law Commission. (2013) The News Media Meets the ‘New Media’. Report no. 128, Law Commission, New
Zealand, Wellington. Retrieved from http://r128.publications.lawcom.govt.nz/
Lord Judge. (2010) A Consultation on the Use of Live Text-based Forms of Communications for Court for the
Purposes of Fair and Accurate Reporting. Retrieved from www.scribd.com/document/48347019/Consul
tation-on-the-use-of-live-text-based-forms-of-communications-from-court-for-the-purposes-of-fair-
and-accurate-reporting
Lord Judge. (2011, December 14) Practice Guidance: The Use of Live Text-based Forms of Communication (includ-
ing Twitter) From Court for the Purposes of Fair and Accurate Reporting. Retrieved from www.judiciary.gov.
uk/wp-content/uploads/JCO/Documents/Guidance/ltbc-guidance-dec-2011.pdf
26
Luft, O. (2011) “Supreme court allows reporters to use Twitter.” Press Gazette: Journalism Today. Retrieved
from www.pressgazette.co.uk/node/46644
Ma, G. (2014, January 21) The Use of Information Technology and Text-based Communications in Courtrooms.
Practice Direction no. 32. Retrieved from http://legalref.judiciary.gov.hk/doc/npd/eng/PD32.htm
Madafferi v Age Company Ltd [2015] VSC 687.
McLachlin, B. (2012) “The relationship between the courts and the news media.” In P. Keyzer, J. Johnston
and M. Pearson (eds.), The Courts and the Media. Sydney: Halstead (pp. 24–35).
Parliament of Australia. (2006, December) Review of Security and Counter Terrorism Legislation. Canberra:
Commonwealth of Australia.
Parliament of Australia. (2014) Review of National Security Legislation Amendment Bill (No. 1). Canberra: Com-
monwealth of Australia. Retrieved from www.aph.gov.au/Parliamentary_Business/Bills_Legislation/
Bills_Search_Results/Result?bId=s969
Parliament of Canada, Bill 231, The Journalist Sources Protection Act. Retrieved from https://openparliament.
ca/bills/42-1/S-231/
Personal Information Protection and Electronic Documents Act 2000 (Canada).
Peters, J. and Tandoc, E. C. (2013) “‘People who aren’t really reporters at all, who have no professional
qualifications’: Defining a journalist and deciding who may claim the privileges.” New York Journal of
Public Policy Quorum, 34, 34–63.
Privacy Act 1985 (Canada).
Privacy Act 1988 (Cth) (Australia).
Queensland Court. (2017) Queensland Court Security Access Card Application Form.
Reese, S., Rutigliano, L., Hyun, K. and Jeong, J. (2007) “Mapping the blogosphere: Professional and citizen-
based media in the global news arena.” Journalism, 8(3), 254–280. doi:10.1177/1464884907076459
Reporters Committee for Freedom of the Press. (n.d.) Shield Laws and Protection of Sources by State. Retrieved
from www.rcfp.org/browse-media-law-resources/guides/reporters-privilege/shield-laws
Roth, P. (2010) “Data protection meets Web 2.0: Two ships passing in the night.” UNSW Law Journal, 32(2),
532–561.
Shennan, F. (2011) “Who are you calling a journalist – can one form of communication command special
protection?” In N. Carpentier, H. Nieminen, P. Pruulmann-Venerfeldt, R. Kilborn, E. Sundin and T.
Olsson (eds.), Critical Perspectives on the European Mediasphere. Ljubljana: ECREA European Media and
Communication Doctoral Summer School (pp. 133–145).
Slater v Blomfield. 2014. NZHC 2221 (12 September 2014) (N.Z.)
Singer, J. (2007) “Contested autonomy.” Journalism Studies, 8(1), 79–95.
Supreme Court of Victoria. (2016) Media Policies and Practices. Retrieved from http://assets.justice.vic.gov.
au//supreme/resources/6ebfaf77-d0b8-4c35-bb9c-a78f0699d3bc/mediapoliciespractices2016_v1.pdf
Supreme Judicial Court Rule. (2012) Supreme Judicial Court Rule 1:19: Electronic Access to the Courts, Mas-
sachusetts Courts System. Retrieved from www.mass.gov/courts/case-legal-res/rules-of-court/sjc/sjc119.
html
The Courts of Nova Scotia. (2014) “The Use of Electronic Devices in Courthouses.” The Courts of Nova
Scotia. Available at: www.courts.ns.ca/Media_Information/electronic_devices_policy.htm.
Tietosuojavaltuutettu v. Satakunnan Markkinapörssi Oy And Satamedia Oy C-73/07. (2008) Retrieved from http://
curia.europa.eu/juris/document/document.jsf;jsessionid=9ea7d0f130d6713372345ad94fc5be5d911
840e7b9f2.e34KaxiLc3eQc40LaxqMbN4Pax4Me0?text=&docid=67007&pageIndex=0&doclang=
EN&mode=lst&dir=&occ=first&part=1&cid=301473
Ugland, E. and Henderson, J. (2007) “Who is a journalist and why does it matter: Disentangling the legal
and ethical arguments.” Journal of Mass Media Ethics, 22(4), 241–261.
Wallace, A. and Johnston, J. (2015) “Tweeting from court: New guidelines for modern media.” Media &
Arts Law Review, 20(1), 15–32.
Warren, C. (2014, September 23) “Interview With Ben Grubb.” Sydney Morning Herald. Retrieved from
https://soundcloud.com/en_grubb/ben-grubb-interviews-chris-warren
Waters, N. (2002) “Can the media and privacy ever get on?” Privacy Law and Policy Reporter, 8(8),
149. Retrieved from www.austlii.edu.au/cgi-bin/sinodisp/au/journals/PrivLawPRpr/2002/1.html?stem=
0&synonyms=0&query=cth%20consol_act%20pa1988108%20s7b#fnb15
West, S. R. (2014) “Press exceptionalism.” Harvard Law Review, 27, 2434–2462.
27
2
STUDYING ROLE CONCEPTIONS
IN THE DIGITAL AGE
A critical appraisal
Digital developments have been a central feature of journalism’s transformation over the past
two decades. However, while journalism scholarship has paid considerable attention to these
developments, this has not necessarily been the case across all areas of inquiry in the field. Where
research has been slightly less responsive to digital transformations is the area of journalistic role
conceptions, despite its centrality for much of the work that goes on in journalism research
more broadly. Accounting for how journalists think about their work and the role they see for
themselves in society is crucial because it is believed that such conceptions eventually affect what
journalists actually do. While recent studies have suggested this link is not as strong as sometimes
assumed (Tandoc et al., 2013; Mellado and van Dalen, 2014), there remains a strong belief in
the importance of studying role conceptions as discursive devices to articulate what values are
important in journalism (Hanitzsch and Vos, 2016).
Studies of journalists’ role conceptions have predominantly addressed institutionalized, main-
stream journalism, and in increasingly global ways. The digital age, however, has led to a fragmenta-
tion within the journalistic field and the arrival of new actors who now practice journalism in ways
that were not possible previously. Further, audiences play an increasingly important role in affecting
journalistic work, whether through direct contact with journalists or in aggregated form through
tools such as web analytics or social media (Hanusch and Tandoc, 2017). How these developments
affect established understandings of role conceptions has so far been less fully explored.
In this chapter, we therefore call for increased scholarly attention to the ways in which the
impact of audiences and new journalistic actors need to be taken into account in future studies of
journalistic roles. We do so particularly by drawing on role theory and its relevance for studying
role conceptions, in order to provide a theoretical framework for such studies. The aim, then, is to
highlight some fruitful avenues for research that may enable us to arrive at a more comprehensive
understanding of journalism cultures in the digital age.
The evolution of role conceptions research

Studies of how journalists conceive of their roles have a long history in scholarship, dating back
more than 50 years. Throughout this time, conceptual and methodological approaches have con-
tinually been modified and expanded upon, both in analytical and normative terms (Hanitzsch
and Vos, 2017). Early accounts of journalistic roles as ‘neutral’ or ‘participant’ (Cohen, 1963)
28
Role conceptions in the digital age
eventually evolved into large-scale, quasi-longitudinal studies, such as in the United States (John-
stone et al., 1976; Weaver and Wilhoit, 1986, 1996; Weaver et al., 2007; Weaver and Willnat,
2016). Over time, these studies have become increasingly complex, with scholars expanding their
curiosity across the globe in comparative studies (Hanitzsch et al., 2011; Weaver and Willnat,
2012). Throughout this process, typologies of journalistic role conceptions have undergone a
continual process of terminological expansion and adaptation, but often – a few exceptions aside
(see, for example, Christians et al., 2009; Donsbach and Patterson, 2004; Hanitzsch, 2007) – with
little “feed back into conceptual work” (Hanitzsch and Vos, 2016: 3).
In response, a recent overview of existing scholarship on journalistic roles identified several
shortcomings (see Hanitzsch and Vos, 2016, 2017), highlighting how research has: (1) tended
to conceptualize journalistic roles within a Western framework that emphasizes journalism as
having a central function in democracy; (2) relied on a variety of terms to refer to different role
concepts inconsistently and interchangeably; (3) primarily explored roles through inductive and
descriptive approaches, leaving this area of research short on theory; and (4) tended to “conflate”,
or treat as one and the same, distinct “attitudinal and performative” dimensions of journalistic
roles (Hanitzsch and Vos, 2017: 4).
While these are important and valid critiques of the literature in this field, we would argue
that a key development affecting journalistic production has so far also been neglected in the
context of evolving role conceptions. Technological developments, while always having been
a part of journalism, have had a particularly fundamental impact on journalistic work recently.
The rise in interactive technologies has accompanied a veritable transformation of journalism
rarely seen before, which is important for future studies to take account of. First, technological
innovations have led to increased interaction between journalists and audiences, requiring the
former to reassess their views of their readers, listeners, and viewers (Loosen and Schmidt, 2012).
Second, these innovations have assisted the arrival of new journalistic actors, whose role concep-
tions may differ substantially from those of established mainstream journalists because of different
journalistic missions. For journalism scholarship to better reflect developments on the ground,
it is therefore crucial to investigate how this growing role of audiences and the impact of new
journalistic actors may be affecting journalistic role conceptions.
This chapter aims to address these key interrelated issues by discussing (a) the way in which
the increasingly reciprocal relationship between journalists and audiences has led to a shift in
audience expectations and consequently a reconceptualization of journalists’ understanding of
their roles in society; and (b) how the emergence of new types of news producers challenges
existing role conceptions of mainstream or traditional journalists. Exploring these two develop-
ments allows us to reconceptualize our approach to understanding journalistic role conceptions
and how we study these from different perspectives in future research.
Role theory in the study of role conceptions

A review of the literature on journalistic role conceptions reveals that surprisingly few stud-
ies appear to engage with a theory from which role conceptions seem to take their name and
which we argue is very useful as a theoretical framework. Role theory, with its origins in the
dramaturgical or theatrical perspective (Simmel, 1920; Goffman, 1959), where people are under-
stood to play interactive parts according to scripts, has only recently started featuring in studies
exploring journalistic roles (Tandoc and Duffy, 2016; Mellado and Hellmueller, 2015; Vos, 2005).
Frequently applied to studying roles in the educational setting (Webb, 1962), role theory offers
multiple perspectives and concepts through which role conceptions can be studied in diverse
societal settings (Biddle, 1986), including journalism.
29
Mimicking some of the terminological confusion found in the literature on journalistic roles
(Hanitzsch and Vos, 2017), role theory has endured scholarly disagreement over the definition
of what comprises a role – a discrepancy largely understood as terminological rather than con-
ceptual (Biddle, 1986). Among the numerous definitions stemming from different sociological
perspectives within role theory, a role can be understood as a “collection of patterns of behavior
which are thought to constitute a meaningful unit and deemed appropriate to a person occupy-
ing a particular status in society” (Turner, 1956: 316). Despite the terminological divergences,
most perspectives and definitions agree on the assumption that roles exist in response to expecta-
tions which role-incumbents have of their own behavior, of other people’s behaviors, and what
they imagine others expect of them (Turner, 1956; Merton, 1957).
This explicit conceptualization of expectations as “generators of roles” (Biddle, 1986: 69)
and emerging in response to patterned behaviors (Turner, 1956), interaction between individuals
(Blumer, 1962), and relationship conduct (Merton, 1957) offers one theoretical framework for
the study of journalistic role conceptions in an age of increased audience interaction. Key here
is that such a framework should draw on multiple perspectives in role theory in an attempt to
bridge gaps between role analysis at the system or macro-level (structural, functional), the meso-
level (organizational) and at the individual or micro-level (symbolic interactionist, cognitive)
(Stryker, 1980). Relying exclusively on one would dismiss the idea that journalistic roles are on
the one hand discursively and symbolically constructed and their relationship to audiences is
one increasingly built on interaction and, on the other hand, that journalists operate within an
institution that is organized and hierarchical in nature (albeit increasingly less so), where roles are
generated in response to expectations directed at their journalistic position and status in society.
If journalistic roles are understood as a collection of behavioral (performance) and cognitive
(orientation) patterns held by individuals within a particular position, we can assume that these
patterns create an opportunity for audiences to develop expectations of a journalist’s role in soci-
ety. The conceptual claim that expectations generate or affect certain role conceptions among
journalists is particularly worthy of studying in light of recent work that challenges the norma-
tive and theoretical assumption that audiences necessarily consider participation valuable or even
beneficial to journalism (Karlsson et al., 2018).
Studying audience expectations of journalistic roles

Typically, roles have been studied by focusing on journalists’ views of their relationship with
power, the existence of their voice in the news, and how they imagine and relate to their audi-
ences. Switching the perspective from the journalists to that of audiences allows us to more
accurately “assess the changing functions which journalism fulfills in society and the roles it
enables the public to play in social life” (Mellado and Van Dalen, 2016: 214). Research on
journalistic roles has focused on three areas: role conception (how journalists understand their
roles), role performance or enactment (how journalists enact or perform their role through their
work, evident in journalistic content), and role expectation (what audiences expect or believe
is journalists’ role), as well as the relationships and gaps between these areas (Tandoc and Duffy,
2016). Research on the relationship between role conceptions and role performance has detected
gaps (Mellado and Van Dalen, 2014) and some correlation (Van Dalen et al., 2012). At the same
time, the relationship between journalistic role conceptions and audience expectations remains
relatively underexplored, with sporadic and geographically scattered case studies so far revealing
both congruence and gaps between the two (cf. Schmidt and Loosen, 2015; Tsfati et al., 2006;
van der Wurff and Schoenbach, 2014; Tandoc and Duffy, 2016; Nah and Chung, 2012; Heider
et al., 2005).
30
Audiences and journalists: a shift in relations

Alongside the adjustment to digital technologies in everyday newswork, “one of the major
changes journalists encounter today is the expectation that they will form relationships with
the audience” (Ferrucci and Vos, 2017: 13; Vos and Ferrucci, this volume, Chapter 3). Whereas
previously the view of the relationship between journalists and audiences was more or less unidi-
rectional, with journalists producing and audiences consuming news, the “development of digital
technologies now enables participation on a scale not previously possible” (Karlsson et al., 2018:
578). This has narrowed the distance and changed communication patterns between these two
groups. The greater input and influence that audiences have on journalistic routines has led to a
blurring of roles captured by terms such as “produsage” (Bruns, 2008), “prosumerism” (Toffler,
1980), and “prodience” (Villi, 2012). Accessing the journalistic field through digital means gives
audiences “a press critic’s role and, therefore, makes them potentially powerful shapers of journal-
ism’s standards of performance” (Craft et al., 2016: 678). With this seemingly greater influence,
audiences are arguably in a position to exercise what the functional perspective in role theory
refers to as ‘sanctions’ – which is to boycott actors or their actions if these contradict or deviate
from the norms expected from them (Popitz, 1972). Thus, audiences become an “institutional
force that can exert pressure on journalists” (Craft et al., 2016: 680). Having originally been
reluctant to interact with audiences, journalists now appear more open to such engagement and
participation, which is arguably leading to changes in how online journalists view the relevance
and importance of certain norms (Agarwal and Barthel, 2015). How journalists perceive their
relationship with audiences, how likely they are to interact with them, and whether they choose
to do so online or offline has also been linked to particular journalistic roles (Holton et al., 2016).
Further, journalists are now receiving extremely detailed information about audience behav-
iors through mechanisms such as web analytics, resulting in a loss of autonomy and increased
pressure to deliver on audience expectations (Hanusch, 20167). Recent studies have also found
that audiences engage in increasingly diverse and complex news consumption practices and con-
siderations that determine what may or may not attract their interest (Costera Meijer and Groot
Kormelink, 2014; Groot Kormelink and Costera Meijer, 2018). The increased opportunity for
audiences to interact with and criticize the press and journalists on social and moral grounds is
also an increasingly relevant indicator “of the flattening of the hierarchical relationship that pre-
viously existed between readers and journalists” (Craft et al., 2016: 687).
Changes to role conceptions

Such a shift in the way audiences interact with news and news producers affects both the expec-
tations audiences have of journalists and, in turn, how journalists attempt to meet those expecta-
tions by adjusting their role conception (Banton, 1996). By challenging journalistic practices,
audiences are essentially asking journalists to reflect on the way they continue to discursively
construct their role in society (Hanitzsch and Vos, 2017). The longer a role and its behavioral
norms have had to develop and crystalize over time (Popitz, 1972), the more disruptive any shift
in expectations will be to the role-incumbent. This means that journalistic role conceptions
grounded in long-held normative ideologies might be more difficult or slower to change. In an
ongoing effort to discursively reaffirm their authority and maintain their societal/professional
legitimacy, journalists engage in different processes. The first relates to “role normalization”,
where journalists contest and consolidate roles into norms. The second relates to “role nego-
tiation” through which journalists either assimilate into a journalism culture or idea of what is
expected of a ‘good’ journalist, or they appropriate, i.e., readjust their professional aspirations
31
(cognitive roles) to align more closely with actual practice. Finally, if these mechanisms fail to
alleviate the role conflict, journalists may exit the journalistic field entirely (Hanitzsch and Vos,
2017). Exiting the field would therefore require a journalist to relinquish his or her professional
role of being a journalist. This is often a painful process, as has recently been observed among
professional journalists who were made redundant as a result of the economic crisis in journalism
(Zion et al., 2016).
If norms (and roles) are discursively constructed, they are also “discursively reconstructed” in
response to changes in the field. However, Vos and Singer stress that Bourdieu’s field theory also falls
short in elaborating on what happens when norms change, except “to point to how new entrants
to a field can be a disruptive force” (Vos and Singer, 2016: 156). As traditional journalists continue
to embrace evolving technological innovations, the emergence of new types of news producers and
the proximity and involvement of audiences, their experiences of role negotiation and even the
relinquishment of roles are a source of disorganization and distress (see Zion et al., 2016).
Multiple audiences, multiple expectations: a framework

for diverse audience communities
Obviously, audiences are but one of many external influences and sources of expectations that
impact journalists’ conception of roles (Donsbach, 2008). In fact, journalists simultaneously have
multiple role-relationships and role-sets (Merton, 1957) with, for example, their sources, fellow
journalists, editors, owners, and advertisers (Shoemaker and Reese, 2013). All of these are sources
of influence and pressure which “can have a hand in wearing away at journalistic cultural capi-
tal” (Craft et al., 2016: 678). Each role-relationship is a source of expectations, many of which
may not be compatible, leaving journalists exposed to role conflict (Stryker and Macke, 1978).
Such inter-role conflict has, for example, been identified in studies of freelance journalists who
also work in public relations (Obermaier and Koch, 2014). As conflicting expectations, and
consequently role conflict increase over time, actors may experience stress, poor professional
performance, low organizational commitment, and higher rates of resignations (Biddle, 1986).
Aside from influences stemming from diverse external actors, such as audiences, further con-
tributing to role strain (Goode, 1960) is the fact that audiences are far from unitary but instead
highly fragmented. It is therefore important to speak of audience communities (Villi, 2012) who
hold diverse expectations. This diversity of audience communities and expectations exists across
different journalism cultures and is likely to become further evident when looking at audience
communities formed around the intersectionality of social structures such as class, race, gender,
age, sexual orientation, ethnicity, and so on. Here, a framework that captures the various ways
in which journalism approaches the audience (Hanitzsch, 2007; Mellado, 2015) combined with
Bourdieu’s social, cultural, and economic capital to account for social differences (and inequali-
ties) (Bourdieu, 1984, 1985) might be useful to studying diverse audience communities.
Research on the impact of social structures and differences on audiences’ interaction with news
is limited (Lindell and Sartoretto, 2017; Hovden and Moe, 2017). Relying on Bourdieu’s cultural
and economic capital to conceptualize social class, Lindell and Sartoretto (2017) found that class
influenced the value young people placed on news, their news preferences and consumption pat-
terns, and importantly the way in which the moral significance of these news practices was used
by audiences to construct and maintain societal boundaries. Those in class categories that con-
sume news infrequently (so-called “news avoiders”) were evaluated by the middle-class groups
and elites as “the same people who skip school” (Lindell and Sartoretto, 2017: 15). The authors
conclude that class, social position, and socialization through the family and school systems influ-
ence “the extent to which young people ‘buy into’ the normative order that regards news as
inherently ‘good’, valuable and worthwhile” (Lindell and Sartoretto, 2017: 16).
32
Some of the key literature on journalistic roles also makes reference to audiences in ways that
can be helpful for developing a framework. In his seminal work on journalism’s institutional
roles, for example, Hanitzsch (2007) proposed three dimensions: interventionism, power distance,
and, of interest here, market orientation. This latter role, which has received much attention from
scholars, goes “to the very heart of journalism studies” (Hanitzsch, 2007: 374). It is argued to be
high where journalists orient themselves toward the “logic of the market” and address audiences
as consumers or clients, while it is low in places where journalists “produce the news primarily
in the ‘public interest’” and address audiences as citizens (Hanitzsch, 2007: 372–375). The latter
focuses on what audiences “should know” over what they “want to know”, which may include
news and information that centers around their “personal fears, aspirations, attitudes, and emo-
tional experiences” and therefore “provides help, advice, guidance, and information about the
management of self and everyday life” (Hanitzsch, 2007: 374–375). While journalism scholar-
ship has examined these aspects of market orientation for considerable time, rarely have studies
examined how audiences may be more or less directly influencing journalists’ role conceptions in
this regard. One recent analysis, however, suggests that increased audience interaction may indeed
be affecting how journalists view their role. In a study of changes in Australian journalists’ role
conceptions related to the market dimension, Hanusch and Tandoc (2017) have suggested that
perceived effectiveness of web analytics and frequency of reading reader comments may be lead-
ing journalists to become more consumer-oriented in their role conceptions.
Similar to Hanitzsch’s conceptualization, Mellado’s (2015) work on how journalistic roles mani-
fest in news content posits three roles concerned with how journalists approach their audiences: the
civic, service, and infotainment-spectator roles. Each of these roles may be simultaneously present
(varying in extent) in the way journalists relate to audiences as found in news content (Mellado
and Van Dalen, 2016). Arguably, however, the same coexistence and simultaneous emergence of
role-expectations could be found among audiences, where certain expectations may at different
times and contexts emerge more than others but nevertheless exist concurrently. To support the
importance of an analytical distinction between the consumer and citizen orientation, Hanusch and
Tandoc’s (2017) recent study was able to empirically identify them as separate dimensions. Inter-
estingly, however, they also found that the frequency of reading reader comments was a predictor
for both an increase in the importance of a citizen and a consumer orientation among journalists.
Going forward, we believe that the discussion presented here may provide the basis for a
potential framework for future analyses that takes account of existing work on the relationship
between journalistic roles and audience conceptions on one hand and social-structural differ-
ences among audience communities on the other. Hence, it might be possible to measure (1)
the prominence of diverse audience communities; (2) their diverse expectations; and (3) any
interconnectedness across these audiences, as well as commonalities among their expectations of
journalistic roles. We would expect that the prominence of certain audience communities and
their expectations will differ across countries depending on their specific geographical, political,
cultural, and historical contexts. Particularly in light of increasingly complex and detailed audi-
ence feedback mechanisms available to journalists, a more complex and nuanced study of such
hybrid audiences could shed light on the way diverse expectations may be influencing develop-
ments in journalists’ role conceptions over time. We will now move on to the second part of our
discussion, to examine the extent to which the emergence of new journalistic actors, facilitated
by digital developments, may impact on how scholars need to take account of journalistic roles.
Diversification of actors in the journalistic field

Digital developments in journalism have also had an important impact on the composition of
actors in the journalistic field. This has occurred in two main ways. First, the arrival of the digital
33
age has been accompanied by dramatic economic downturns for established, mainstream media
organizations in many Western countries, resulting in unprecedented job losses across a number
of countries (O’Donnell et al., 2012; Starr, 2012). This alone has led to more precarious working
conditions for many journalists and an increasing reliance on freelancers in some countries (Bak-
ker, 2014) and the breaking-down of traditional ‘walls’ between editorial and advertising depart-
ments (Coddington, 2015; Ferrer-Conill and Karlsson, this volume, Chapter 35), as well as a fear
that the resultant loss of autonomy and reliance on dwindling revenues may seriously undermine
journalism’s watchdog role (Siles and Boczkowski, 2012).
Second, in addition to these impacts on traditional forms of journalism, the introduction of
participatory technologies and breaking down of the entry barriers to news publishing has led
to an impressive diversification of the journalistic field. Having developed beyond the early days
of citizen journalism (Bruns, 2008), there now exists a wide range of new, professionalized actors
who are producing content that can to varying degrees be classified as journalism. In fact, jour-
nalism increasingly appears to be taking place outside the traditional newsroom (Anderson, 2011;
Bakker, 2014). These used to be a blind spot in journalism scholarship, but increasingly, research-
ers are becoming more and more interested in exploring these new kinds of actors (Deuze and
Witschge, 2018). The arrival of these new actors therefore calls for new conceptual approaches to
better understand journalistic work in the digital age, especially with regard to how they conceive
of their societal role.
The need to understand the roles of the journalistic “in-betweeners” (Ahva, 2016), who
conceptually might be located somewhere between traditional journalists and audiences, comes
at a crucial time. The distinction between the news producer and news consumer is blurring,
with audiences increasingly acting as both (Carlson and Lewis, 2015; Bruns, 2008). The digital
environment has allowed audiences to join journalists as co-producers of information, which has
left journalists with “little to distinguish them from regular citizens” (Ferrucci and Vos, 2017: 3).
To exemplify this blurring, Ferrucci and Vos (2017) suggest that, nowadays, both citizens and
journalists are bloggers, both share information on social media, and both have audiences for
their work.
The blurring of boundaries in journalism has led to a growing cluster of diverse, so-called
atypical producers of journalistic content, who will need to be accounted for by journalism
scholarship. This is a challenge in particular for research on journalistic role conceptions, a field
that has traditionally focused not only on mainstream journalism but also predominantly on
journalism’s relationship with politics, at the expense of exploring journalism’s role in everyday
life (Hanitzsch and Vos, 2016). This latter field, which includes areas such as lifestyle journalism,
has, however, been increasingly important and relevant for audiences during a time of major
social and cultural transformations (Hanusch and Hanitzsch, 2013). But more broadly, actors
beyond the traditional newsroom now include regular citizen journalists, bloggers, news start-
ups, entrepreneurial journalists, students aspiring to become journalists, academics contributing
to public discourse, activists, and influencers on social media. They are blogging on personal or
popular social media platforms like YouTube, Instagram, Twitter; producing for traditional media
on an (ir-)regular basis; and creating content of public interest, informing audiences and media
alike, and thus can be discerned from regular users sharing private content (Nicey, 2016).
Challenges for scholarship

However, our conceptual understanding of (1) what defines these new journalistic actors and
(2) how they might conceive of their role in society remains very limited. The first point is a
complex undertaking at the heart of much scholarship that interrogates forms beyond traditional
34
journalism. While such traditional views may not suffice in the digital age, there is also a risk in
labeling everything as journalism, resulting in too broad an approach that may render the term
meaningless. Of interest here would be the trend toward increased brand journalism (Light, 2014),
long-held debates about whether citizen blogs may constitute journalism (Nah and Chung,
2012), or the mix of commercial and journalistic motives and interests evident in successful social
media influencers’ accounts (Abidin, 2017; Pang et al., 2016).
This aligns with a point most recently raised by Deuze and Witschge (2018), who argue that
scholarship needs to go beyond traditional views of journalism as a stable institution distinct
from other social systems and beyond an almost exclusive focus on its democratic relevance. At
the same time, they believe that long-held understandings of journalism defined by terms such as
public service, objectivity, autonomy, immediacy, and a sense of ethics (Deuze, 2005) are still use-
ful to studying journalism from the inside out (Deuze and Witschge, 2018). They argue that “it
would be a mistake to assume that the types of journalism emerging outside and alongside legacy
news organizations are necessarily different or oppositional to the core values, ideals, and practices
of the profession” (Deuze and Witschge, 2018: 168). One key criticism is that most approaches
have focused on journalism’s relationship with politics, having led to an undertheorization and
underappreciation for emerging and increasingly popular factuality formats and related journal-
istic practices (McNair, 2000; Harrington, 2013). Hence, an arguably more open approach that
conceptualizes journalism as something much broader than has so far occurred may also help us
better identify journalistic boundaries that may continue to exist, as well as those that may be
breaking down.
On the second point, how new actors may conceive of their roles, as well as how such role
conceptions may also indirectly or directly affect established journalistic role conceptions, recent
work on journalism beyond its relationship with political life may prove at least a useful starting
point. The traditional focus on journalism’s relationship with politics and democracy has largely
been at the expense of journalism’s relevance for everyday life in societies where the media have
increasingly replaced traditional institutions to provide guidance to people on how to live their
lives. Recent accounts of lifestyle journalists demonstrate that quite different role sets exist for
such beats (Hanusch and Hanitzsch, 2013), which, it has been claimed, necessitates broader con-
ceptualizations of journalistic roles based in realities on the ground (Hanitzsch and Vos, 2017).
Taking account, then, of new journalistic actors’ role conceptions may prove useful in further
broadening our understanding. Again, while it will be useful to measure brand journalists’ or
social media influencers’ role conceptions vis-à-vis established journalistic tenets, explorative
work may be particularly promising in its ability to tease out entirely new, yet important types
of roles. In line with Deuze and Witschge’s (2017: 8) call, this may help us further address “the
diversity in roles, functions, and people’s backgrounds that exists in media work generally and
newswork in particular”. In fact, as a recent study has shown, just comparing digital journalists
with traditional journalists can help us along this path. Examining digital journalists, Ferrucci and
Vos (2017) found that this group of newsworkers were considerably more interested in report-
ing, analyzing, and contextualizing and had a better awareness of community needs, leading to a
decline in importance of the objectivity norm.
Further, today’s news ecology needs also to be seen as a complex mix of what Lewis and West-
lund (2015) have described as actors, actants, audiences, and activities in cross-media newswork.
It is particularly the new, nonhuman technological actants who, through their impact on jour-
nalistic work, may further contribute to evolving and even new role conceptions for journalists.
As we have already argued, then, new entrants to the journalistic field may challenge not only
existing norms (Vos and Singer, 2016) but also journalists’ discourses on their roles in society. Yet,
as research on these new actors’ role conceptions is still in its infancy, it is difficult to identify the
35
directions in which these could develop and what this may mean for the journalistic field at large.
Hence, we believe that this area will provide quite a fruitful avenue for future research.
Conclusion
The development of a broad spectrum of digital tools has enabled new actors – ranging from
bloggers, digital start-ups, and social media influencers to audiences partaking in news processes –
to access the journalistic field in unprecedented ways. These shifts in the previously tightly
guarded journalistic field are challenging its practices and, as a result, also long-established role
conceptions of mainstream journalists. In this chapter, we have argued for increased scholarly
attention to the impact that audiences and new journalistic actors have on the ways in which
journalists are (re)conceptualizing their roles in society in the digital environment.
Journalists are increasingly expected to form relationships with audiences (Ferrucci and Vos,
2017), and they appear to be more open to such engagement and participation (Anderson, 2011).
However, little is known about how the expectations audiences may have of journalists are affect-
ing their role conceptions (Tandoc and Duffy, 2016). In light of a shortage of theory in this area
of scholarship (Hanitzsch and Vos, 2017), we suggest that role theory offers a rich and relatively
underutilized framework for studying journalistic roles in response to audience expectations. We
conceptualized the relationship between journalists and audiences primarily through Hanitzsch’s
(2007) set of institutional roles, which provides a useful starting point to, for example, examine
change in journalists’ role conceptions as a result of audience feedback (Hanusch and Tandoc,
2017). In addition, we argue that social structures, such as class and race, among others, may
further contribute to the diversification of audience communities and divergent expectations of
journalists, leading to experiences of role conflict among journalists (Stryker and Macke, 1978).
An additional source of strain for mainstream journalists is the growing pressure to integrate
digital technologies into their daily work, accompanied by increased job precarity (O’Donnell
et al., 2012; Starr, 2012). While the breaking down of entry barriers to the journalistic field has
diversified actors within the journalistic field, this also means much journalistic work now takes
place outside of the traditional newsroom (Anderson, 2011; Bakker, 2014). Here, we can also
observe that these new actors are simultaneously audiences and respond to journalistic content in
that capacity. These changes call for scholarly work to engage with questions on reconceptualiz-
ing the roles of journalistic ‘workers’ beyond the newsroom and traditional journalistic boundar-
ies (Deuze and Witschge, 2017). More than ever, research on role conceptions of new journalistic
actors would benefit from shifting its view away from journalism’s relationship with politics
(Hanitzsch and Vos, 2018; Hanusch and Hanitzsch, 2013). This shift in definitional approaches
would allow us to broaden our conceptualizations of the roles of both traditional and new jour-
nalistic actors, resulting in an arguably more comprehensive understanding of journalistic work
in the digital age.
Beyond this broadening of our conceptualizations, future research would ideally study these
across cultures. So far, much of the research on digital developments in journalism has still tended
to be focused, single-nation studies, yet to be able to test our theories and challenge phenomena
and universality, we need to do so across different cultural, social, historical, and political contexts
(Livingstone, 2003). In addition, longitudinal studies would allow us to map gradual shifts in
the role conceptions of traditional journalists as they continue to embrace digital technologies
and the evolving role conceptions of new journalistic actors as their prominence and relevance
in the journalistic field continues to flourish. Lastly, future research ought to address the short-
age of scholarship on audience expectations of journalistic roles, as well as its tendency to rely
on quantitative surveys, modeled on items extracted from previous journalistic role conception
36
surveys. Studies based on qualitative and grounded theory approaches (Tandoc and Duffy, 2016)
would allow us to reconceptualize journalistic role conceptions through more nuanced audi-
ence expectations that may emerge when unprompted by predetermined survey conceptions of
journalistic roles.
Further reading
This chapter has drawn on key scholarship in the study of journalistic role conceptions, specifi-
cally, Hanitzsch and Vos’ “Journalistic Roles and the Struggle Over Institutional Identity: The
Discursive Constitution of Journalism” (2017), Tandoc and Duffy’s research on audience expec-
tations in “Keeping Up with the Audiences: Journalistic Role Expectations in Singapore” (2016),
Bruce Biddle’s overview of role theory in “Recent Development in Role Theory” (1986) and
Deuze and Witschge’s call to look at new journalistic actors emerging outside of the traditional
newsroom in “Beyond Journalism: Theorizing the Transformation of Journalism” (2017).
References
Abidin, C. (2017) “Influencer extravaganza: Commercial ‘lifestyle’ microcelebrities in Singapore.” In L.
Hjorth, H. Horst, A. Galloway and G. Bell (eds.), The Routledge Companion to Digital Ethnography. New
York, NY: Routledge (pp. 158–168).
Agarwal, S. D. and Barthel, M. L. (2015) “The friendly barbarians: Professional norms and work routines of
online journalists in the United States.” Journalism, 16(3), 376–391.
Ahva, L. (2016) “How is participation practiced by ‘in-betweeners’ of journalism?” Journalism Practice, 11(2–3),
142–159.
Anderson, C. W. (2011) “Blowing up the newsroom: Ethnography in an age of distributed journalism.” In
D. Domingo and C. Paterson (eds.), Making Online News. New York, NY: Peter Lang (pp. 151–160).
Bakker, P. (2014) “Mr. Gates returns.” Journalism Studies, 15(5), 596–606.
Banton, M. (1996) “Role.” In A. Kuper and J. Kuper (eds.), The Social Science Encyclopedia (2nd ed.). London:
Biddle, B. J. (1986) “Recent development in role theory.” Annual Review of Sociology, 12, 67–92.
Blumer, H. (1962) “Society as symbolic interaction.” In A. M. Rose (ed.), Human Behavior and Social Pro-
cesses. Boston, MA: Houghton Mifflin Co. (pp. 179–192).
Bourdieu, P. (1984) Distinction: A Social Critique of the Judgment of Taste. New York, NY: Routledge.
Bourdieu, P. (1985) “The social space and the genesis of groups.” Theory and Society, 14(6), 723–744.
Bruns, A. (2008) Blogs, Wikipedia, Second Life, and Beyond: From Production to Produsage. Bern, Switzerland:
Peter Lang.
Carlson, M. and Lewis, S. C. (eds.). (2015) Boundaries of Journalism: Professionalism, Practices and Participation.
London: Routledge.
Christians, C. G., Glasser, T. L., McQuail, D., Nordenstreng, K. and White, R. A. (2009) Normative Theories
of the Media: Journalism in Democratic Societies. Urbana, IL: University of Illinois Press.
Coddington, M. (2015) “The wall becomes a curtain: Revisiting journalism’s news-business boundary.” In
M. Carlson and S. C. Lewis (eds.), Boundaries of Journalism: Professionalism, Practices and Participation. New
Cohen, B. C. (1963) The Press and Foreign Policy. Princeton, NJ: Princeton University Press.
Costera Meijer, I. and Groot Kormelink, T. (2014) “Checking, sharing, clicking and linking.” Digital Journal-
ism, 3(5), 664–679.
Craft, S. et al. (2016) “Reader comments as press criticism: Implications for the journalistic field.” Journalism,
17(6), 677–693.
Deuze, M. (2005) “What is journalism?: Professional identity and ideology of journalists reconsidered.”
Journalism, 6(4), 442–464.
Journalism, 19(2), 165–181.
Donsbach, W. (2008) “Journalists’ role perception.” In W. Donsbach (ed.), The International Encyclopedia of
Communication. London: Blackwell.
37
Donsbach, W. and Patterson, T. E. (2004) “Political news journalists: Partisanship, professionalism, and polit-
ical roles in five countries.” In F. Esser and B. Pfetsch (eds.), Comparing Political Communication: Theories,
Cases, and Challenges. New York, NY: Cambridge University Press (pp. 251–270).
Ferrucci, P. and Vos, T. (2017) “Who’s in, who’s out?” Digital Journalism, 5(7), 868–883.
Goffman, E. (1959) The Presentation of Self in Everyday Life. New York, NY: Anchor Books.
Goode, W. J. (1960) “A theory of role strain.” American Sociological Review, 25(4), 483–496.
Groot Kormelink, T. and Costera Meijer, I. (2018) “What clicks actually mean: Exploring digital news user
practices.” Journalism, 19(5), 668–683.
Hanitzsch, T. (2007) “Deconstructing journalism culture: Toward a universal theory.” Communication Theory,
17(4), 367–385.
Hanitzsch, T. and Vos, T. P. (2018) “Journalism beyond democracy: A new look into journalistic roles in
political and everyday life.” Journalism, 19(2), 146–164.
Hanitzsch, T. and Vos, T. P. (2017) “Journalistic roles and the struggle over institutional identity: The dis-
cursive constitution of journalism.” Communication Theory, 27(2), 115–135.
Hanitzsch, T., et al. (2011) “Mapping journalism cultures across nations.” Journalism Studies, 12(3), 273–293.
Hanusch, F. (2017) “Web analytics and the functional differentiation of journalism cultures: Individual, orga-
nizational and platform-specific influences on newswork.” Information, Communication & Society, 20(10),
1571–1586.
Hanusch, F. and Hanitzsch, T. (2013) “Mediating orientation and self-expression in the world of consump-
tion: Australian and German lifestyle journalists’ professional views.” Media, Culture & Society, 35(8),
943–959.
Hanusch, F. and Tandoc, E. (2017) “Comments, analytics, and social media: The impact of audience feed-
back on journalists’ market orientation.” Journalism. [online first] doi:10.1177/1464884917720305
Harrington, S. (2013) Australian TV News: New Forms, Functions, and Futures. Bristol: Intellect.
Heider, D. et al. (2005) “What the public expects of local news: Views on public and traditional journalism.”
Journalism & Mass Communication Quarterly, 82(4), 952–967.
Holton, A. E. et al. (2016) “Interacting with audiences.” Journalism Studies, 17(7), 849–859.
Hovden, J. F. and Moe, H. (2017) “A sociocultural approach to study public connection across and beyond
media.” Convergence, 23(4), 391–408.
Johnstone, J., Slawski, E. and Bowman, W. (1976) “The professional values of American newsmen.” Public
Opinion Quarterly, 36, 522–540.
Karlsson, M. et al. (2018) “The public doesn’t miss the public. Views from the people: Why news by the
people?” Journalism, 19(5), 577–594.
Lewis, S. C. and Westlund, O. (2015) “Actors, actants, audiences, and activities in cross-media news work:
A matrix and a research agenda.” Digital Journalism, 3(1), 19–37.
Light, L. (2014) “Brand journalism: How to engage successfully with consumers in an age of inclusive
individuality.” Journal of Brand Strategy, 3(2), 121–128.
Lindell, J. and Sartoretto, P. (2017) “Young people, class and the news.” Journalism Studies. [online first] doi:
10.1080/1461670X.2017.1310628
Livingstone, S. (2003) “On the challenges of cross-national comparative media research.” European Journal
of Communication, 8(4), 477–500.
Loosen, W. and Schmidt, J.-H. (2012) “(Re-)discovering the audience.” Information, Communication & Society,
15(6), 867–887.
McNair, B. (2000) Journalism and Democracy: An Evaluation of the Political Public Sphere. London: Routledge.
Mellado, C. (2015) “Professional roles in news content.” Journalism Studies, 16(4), 596–614.
Mellado, C. and Hellmueller, L. (2015) “Professional roles and news construction: A media sociology con-
ceptualization of journalists’ role conception and performance.” Communication & Society, 28(3), 1–11.
Mellado, C. and Van Dalen, A. (2014) “Between rhetoric and practice: Explaining the gap between role
conception and performance in journalism.” Journalism Studies, 15(6), 859–878.
Mellado, C. and Van Dalen, A. (2016) “Challenging the citizen – Consumer journalistic dichotomy.” Jour-
nalism & Mass Communication Quarterly, 94(1), 213–237.
Merton, R. K. (1957) “The role-set: Problems in sociological theory.” The British Journal of Sociology, 8(2),
106–120.
Nah, S. and Chung, D. S. (2012) “When citizens meet both professional and citizen journalists: Social trust,
media credibility, and perceived journalistic roles among online community news readers.” Journalism,
13(6), 714–730.
Nicey, J. (2016) “Semi-professional amateurs.” In T. Witschge et al. (eds.), The SAGE Handbook of Digital
Journalism. London: SAGE (pp. 222–235).
38
Obermaier, M. and Koch, T. (2014) “Mind the gap: Consequences of inter-role conflicts of freelance jour-
nalists with secondary employment in the field of public relations.” Journalism, 16(5), 615–629.
O’Donnell, P., McKnight, D. and Este, J. (2012) Journalism at the Speed of Bytes: Australian Newspapers in the
21st Century. Sydney: Walkley Foundation.
Pang, A., Yingzhi Tan, E., Song-Qi Lim, R., Yue-Ming Kwan, T. and Bhardwaj Lakhanpal, P. (2016) “Build-
ing effective relations with social media influencers in Singapore.” Media Asia, 43(1), 56–68.
Popitz, H. (1972) “The concept of social role as an element of sociological theory.” In J. A. Jackson (ed.),
Role. Cambridge: Cambridge University Press (pp. 11–41).
Schmidt, J.-H. and Loosen, W. (2015) “Both sides of the story.” Digital Journalism, 3(2), 259–278.
Shoemaker, P. J. and Reese, S. D. (2013) Mediating the Message in the 21st Century: A Media Sociology Perspec-
tive. New York, NY: Routledge.
Siles, I. and Boczkowski, P. J. (2012) “Making sense of the newspaper crisis: A critical assessment of existing
research and an agenda for future work.” New Media & Society, 14(8), 1375–1394.
Simmel, G. (1920) “Zur Philosophie des Schauspielers.” Logos, 1, 339–362.
Starr, P. (2012) “An unexpected crisis: The news media in postindustrial democracies.” The International
Journal of Press/Politics, 17(2), 234–242.
Stryker, S. (1980) Symbolic Interactionism: A Social Structural Version. Menlo Park, CA: Benjamin Cummings.
Stryker, S. and Macke, A. S. (1978) “Status inconsistency and role conflict.” Annual Review of Sociology, 4,
57–90.
Tandoc, Jr., E. C. and Duffy, A. (2016) “Keeping up with the audiences: Journalistic role expectations in
Singapore.” International Journal of Communication, 10, 3338–3358.
Tandoc Jr., E. C., Hellmueller, L. and Vos, T. P. (2013) “Mind the gap: Between journalistic role conception
and role enactment.” Journalism Practice, 7(5), 539–554.
Toffler, A. (1980) The Third Wave. New York, NY: William Morrow.
Tsfati, Y. et al. (2006) “What is good journalism? Comparing Israeli public and journalists’ perspectives.”
Journalism, 7(2), 152–173.
Turner, R. H. (1956) “Role-taking, role standpoint, and reference-group behavior.” American Journal of
Sociology, 61(4), 316–328.
Van Dalen, A., de Vreese, C. H. and Albæk, E. (2012) “Different roles, different content? A four-country
comparison of the role conceptions and reporting style of political journalists.” Journalism, 13(7), 903–922.
van der Wurff, R. and Schoenbach, K. (2014) “Civic and citizen demands of news media and journalists:
What does the audience expect from good journalism?” Journalism & Mass Communication Quarterly,
91(3), 433–451.
Villi, M. (2012) “Social curation in audience communities: UDC (User-Distributed Content) in the net-
worked media ecosystem.” Participations: The International Journal of Audience and Reception Studies, 9(2),
614–632.
Vos, T. P. (2005) “Journalistic role conception: A bridge between the reporter and the press.” Paper pre-
sented at International Communication Association (ICA) Conference, New York, NY, May 29.
Vos, T. P. and Singer, J. B. (2016) “Media discourse about entrepreneurial journalism.” Journalism Practice,
10(2), 143–159.
Weaver, D. H. et al. (2007) The American Journalist in the 21st Century: U.S. News People at the Dawn of a New
Millennium. Mahwah, NJ: Erlbaum.
Weaver, D. H. and Wilhoit, G. C. (1986) The American Journalist. Bloomington, IN: Indiana University Press.
Weaver, D. H. and Wilhoit, G. C. (1996) The American Journalist in the 1990s: U.S. News People at the End of
an Era. Mahwah, NJ: Erlbaum.
Weaver, D. H. and Willnat, L. (eds.). (2012) The Global Journalist in the 21st Century. New York, NY:
Routledge.
Weaver, D. H. and Willnat, L. (2016) “Changes in U.S. journalism.” Journalism Practice, 10(7), 844–855.
Webb, J. (1962) “The sociology of a school.” The British Journal of Sociology, 13(3), 264–272.
Zion, L., Dodd, A., Sherwood, M., O’Donnell, P., Marjoribanks, T. and Ricketson, M. (2016) “Working for
less: The aftermath for journalists made redundant in Australia between 2012 and 2014.” Communication
Research and Practice, 2(2), 117–136.
39
3
WHO AM I? PERCEPTIONS
OF DIGITAL JOURNALISTS’
PROFESSIONAL IDENTITY
What it means to be a journalist has become fluid in the digital era (Black, 2010; Ugland and
Henderson, 2007). Journalists have been asked to learn new skills, develop new routines, and
adopt new outlooks (Anderson, 2013). Meanwhile, these same new skills, routines, and outlooks
seemingly make journalists more like bloggers and social media stars, blurring the boundaries
between professional journalists and citizens performing “acts of journalism” (Holt and Karlsson,
2014: 1795). Whether or not citizen journalists and bloggers have actually significantly displaced
work done by traditional journalists is almost immaterial. The discursive challenge – all the talk
about the changing definition of who’s a journalist – has registered in real ways, such as in how
journalists have perceived their own professional identity (Ferrucci and Vos, 2016).
Against the backdrop of changing journalistic identity, a new identity of sorts has emerged for
a subset of newsworkers – digital journalists. As this chapter demonstrates, digital journalists see
themselves as both distinct from citizen journalists and bloggers on the one hand and from pro-
fessional journalists working in traditional media on the other hand. They can match the digital
production, distribution, and interaction skills of the savviest of digital entrepreneurs, while also
performing the journalistic social roles of a venerable, truth-based profession. This sets them apart,
as they see it, as a new breed, evolved to survive in an age of technological and economic disrup-
tion. It’s an identity that is reinforced – albeit only recently – by the broader journalism profession.
So, what is the identity of digital journalists? How has the journalistic field perceived digital
journalists and how has that played into their identity? This chapter explores how digital journal-
ists construct their professional identity. It draws on interviews with more than 50 self-identified
digital journalists and an analysis of digital journalists’ own public discourse and of the broader
journalism field’s discourse about digital journalism in the trade press, journalism blogs, and
popular press from 2000 to the present. The chapter indexes the legitimating norms and practices
of the field by listening to how journalists speak about the nature of their digital work and what
makes it distinctive.
Thinking about identity

Identity is a tricky, complex matter (Roccas and Brewer, 2002; see also Johnston and Wallace, this
volume, Chapter 1). On the one hand, identity speaks to a constancy of character, to a steadi-
ness, regardless of situation (Kopytowska and Kalyango, 2014). This is a cardinal precept of social
40
Who am I?
identity theory – membership in a group with shared values and routines is important in main-
taining a fixed identity (Turner, 1982). Here, journalists never lose sight of what it means to be a
journalist or to be a professional. Technologies may change, economic circumstances may shift,
but being a journalist is still being a journalist (Craft and Davis, 2016). This steady knowledge of
who one is allows the individual to maintain a sense of social identity and maintain productive
work routines. If a journalist knows she is not a politician, preacher, or public relations practitio-
ner, she can stick to the tasks at hand. She might even find self-worth in her fixed professional
identity – seeing what she does as superior to the work of politicians, preachers, or public relations
practitioners (Fisher, 2015; Kitch, 2014). Identity thus might also do important work in identity
management, particularly when journalists are generally held in low esteem by the broader popu-
lation (Kovach and Rosenstiel, 2014). A clear and stable identity can insulate the journalist from
external actors who would seek to attack, undermine, or just disparage journalism practice.
On the other hand, individuals can be quite adept at “compartmentalization”, thus sometimes
maintaining multiple, even contradictory, identities (Roccas and Brewer, 2002: 88). Once we
accept an adjective for a kind of journalist – broadcast journalist, business journalist, advocacy
journalist – then we and they are primed to think about differences. Broadcast journalists are
more performers than their print counterparts; business journalists have more specialized knowl-
edge about their subject than general assignment reports; and advocacy journalists are more
invested in the plight of their subjects than are traditional journalists. How does adding ‘digital’
as an adjective to ‘journalist’ reflect differences with other journalists? This, of course, requires
understanding what digital journalists do that is distinct from work done by fellow journalists.
Just as important as knowing what digital journalists do is gaining a sense of how this unique
set of skills and outlooks positions digital journalists within the broader journalistic field and
how their position is reflected in their identity. The work of journalists generally has been
legitimated as essential to democracy: savvy; modern, vital, and in touch with social currents;
ruggedly autonomous; and essential to the market success of news organizations (Fallows, 1996;
Kovach and Rosenstiel, 2014; Vos, 2012; Vos and Craft, 2016; Vos and Singer, 2016). It is likely
safe to conclude that these expressions of legitimacy have formed a hierarchy – contribution to
democracy would come first and contribution to financial success last in this short list. Granted,
the remainder of these legitimating expressions has been harder to rank. Nevertheless, we should
expect that standing – and with it, identity – in the journalistic field is ordered according to this
hierarchy. Thus, political reporting – given its connection to democracy – grants a journalist
more legitimacy than sports reporting. Likewise, professional journalists have historically marked
advocacy journalists as practicing a lesser form of journalism, even questioning whether advocacy
journalism is journalism at all (McMillian, 2011; see Montal and Reich, this volume, Chapter 4).
Another form of legitimating discourse within the field of journalism has been tied to how
journalists view their social role or roles (Hanitzsch and Vos, 2017; Hanusch and Banjac, this
volume, Chapter 2). How a journalist understands the role of journalism has been described
in terms of “professional identities” (Johnstone et al., 1976: 131). A role represents the outlook
journalists bring to their work – it is what they see as valuable or essential in being a journalist.
Journalists have constructed their journalistic identity, for example, as an adversarial role, dis-
seminator role, interpretative role, and or populist mobilizer role (Weaver et al., 2007). But, here
too, these roles map onto normative hierarchies – some roles have had greater legitimacy within
the field than others.
Likewise, digital journalists’ identities will be formed against the backdrop of journalism’s
legitimating discourses. Whether digital journalists are discursively constructed as essential to
democracy or essential to profitability, for example, would likely shape how the field looks at
digital journalists and how digital journalists think about themselves. Whether digital journalists
41
see themselves performing an adversarial role, a populist mobilizer role or other role will also
constitute their identity relative to the field (Agarwal and Barthel, 2015). This ties to another
cardinal precept of social identity theory – individuals will go to great lengths to see themselves
positively and will seek for others to see them positively, too (Turner, 1982).
Meanwhile, the legitimacy and standing of quasi-journalistic actors – particularly those who
share skills or outlooks with digital journalists – also come into play. Bloggers and social media
stars can have general social legitimacy, but their legitimacy and standing will also be judged by
journalists based on journalism’s legitimating discourses. So, if digital journalists are discursively
attached to these quasi-journalistic actors, digital journalists’ own identity will also likely come
into play and challenge their ability to maintain a positive self-identity. But, also following from
social identity theory (Tajfel, 1982), if digital journalists see themselves as superior to quasi-
journalists, constructing them as an out-group, this would be a basis for their own identity.
All of which is to reinforce that identity can be complex, made more so by a compartmental-
ization of identities. Digital journalists might identify with other professional journalists at one
moment and see themselves as wholly distinct at another moment (Eldridge, 2016). Indeed, social
identity theory (Tajfel, 1982) would support this notion. Thus, the challenge here is to account
for journalistic identity in all its complexity.
The field’s view of the digital journalist

One obvious source of anyone’s identity is what others think of him or her. If those close to you
think you are popular, clever, or smart, you would likely think that of yourself. Thus, the story
of how the journalistic field has talked about digital journalists is a part of the story of digital
journalists’ identity. The story is not straightforward.
In the early days, when ‘digital journalist’ was beginning to enter the journalism field’s vocab-
ulary, the job was far from the pinnacle of the profession. It was seen as a job for those new to the
profession; little more than marketing; a derivative of real journalism; and – in the least generous
of characterizations – a corruption of journalism’s social role and norms. In the boundary work
being done by veteran journalists, digital journalists barely qualified as practitioners of journal-
ism. Simply put, digital journalists were seen as the embodiment of journalism losing its way.
Long-time professional journalists had been closely guarding their authority and autonomy
in the face of efforts to bring user-generated content into the big tent of journalism (see Lewis,
2012). As this war was being waged, digital journalists appeared on the scene to facilitate audience
contribution and involvement. Digital journalists, with their digital-first mentalities, advocated
for new ways of doing journalism: “We need to build relationship and community connection
into the processes of newsgathering and into its starting points. This is key to making journal-
ism less insular and more outwardly focused” (Sill, 2011). For some veterans, this put the digital
journalists in league with the enemy. They were squandering journalists’ hard-fought and hard-
won professional autonomy. As one journalist lamented, “There was a time, not that long ago,
when a news organization’s credibility was boosted by [. . .] the distance it kept from its audience.
All that played into its status as a Respected Institution” (Benton, 2008; italics in the original).
The lamentation pinned the change on digital journalists and mourned the loss of journalism’s
standing – a loss that struck at the heart of journalists’ professional identity.
Digital journalists were also sometimes seen as pawns of malevolent powers, again being
depicted as naively trading in treasured values for uncertain benefits. Digital journalists seemed
all too eager to play by the rules of third-party powers, such as Facebook and Google. For exam-
ple, veteran journalists tended to “shudder” at the practice of search engine optimization (Rice,
2010) – writing online headlines to get a news story to appear prominently in search results or
to get readers to share or click on news stories. This was precisely the kind of practice digital
42
Who am I?
journalists had perfected. Traditional journalists worried that the giants could not be trusted with
journalism’s best interests, arguing Google’s end game remained largely “unknown in large part
because, like most big institutions, Google limits transparency and is defensive when it comes to
criticism” (Osnos, 2009: 28). Chasing short-term financial payoffs, veteran journalists believed,
ultimately shortchanged public service journalism.
Legacy journalists rarely defended their publications’ struggling business model; however, they
remained fairly confident that the promises of a rosy digital future – represented by digital
journalists at born digital news outlets – were either implausible or came at too high of a price.
“Writers and editors know that click-driven Internet economics tend to reward lowbrow gim-
mickry” (Rice, 2010).
By January of 2013, however, the picture was different. A veteran journalist, whose career
harkened back to the golden years of newspapers, hailed digital journalists as “a new wave of
talent” (Dvorkin, 2013). “They relate to and engage with the audience unlike a past generation
of reporters who could care less what readers thought (after all, what do they know?),” Dvorkin
(2013) wrote. He added:
Using the tools of social media, they follow their colleagues as competitive beat reporters
to gain insight from them. Most important, they banter with them in full public view,
a far more raw, if not real, version of any “news analysis” than shows up in newsprint.
(Ibid.)
For Dvorkin, this also involves new perspectives on the ‘scoop’, “rid[ing] the crest of a competi-
tor’s scoop”, and,
filtering it through their own eyes for different audiences. They produce their own vid-
eos, photos and galleries and podcasts to extend their reach. And they trust in Google,
angling stories (and a story’s headline) to give them the best chance of reaching the
world.
(Ibid.)
This is a loving portrait of a modern, vital, and current field. But it is also not without a sort
of paternal(istic) pride. The column concludes that the future will be ushered in when digital
journalists and old school reporters learn from each other. So, what should the field of journalism
learn from the new wave of digital journalists? Not only a new set of skills but a new orientation
to the audience. These are journalists who reach out to the world. Journalistic autonomy – the
kind that is unconcerned with the thoughts of readers – is disparaged as a part of the past. This
makes the digital journalist more real than the cloistered journalists of a past generation.
But, above all, the digital journalist is elevated for the goods she can deliver. The digital jour-
nalist knows the alchemy of audience engagement. It was a mystery to traditional journalists, but
audience engagement would, it was believed, transform journalism’s prospects moving forward.
Job ads for digital journalists began to catch and project the dynamism that digital journalists
seemingly brought to a moribund industry. “We need a candidate that is full of ideas and aware
of content which drives digital audience, as well as being fast, organized and accurate, with the
ability to communicate brilliantly via social media” (“Job”, 2017). The journalism field had
long seen the “nose for news” as a mysterious quality that made a journalist a journalist (Vos
and Finneman, 2017: 271). Here, and elsewhere, the nose for connecting to a digital audience is
elevated in similar terms. It is what made a digital journalist a digital journalist.
Digital journalists’ audience engagement abilities have come to be seen in messianic terms –
even being labeled as the “salvation of journalism” (Davis, 2013). As journalism’s profitability took
43
on greater salience, digital journalists’ importance also became greater. But, claims of journalism’s
salvation were not confined to retaining and monetizing audiences. Digital journalists’ audience
engagement abilities also spoke to journalism’s democratic responsibilities, at least in the field’s
discourse. An engaged audience was an audience engaged in the public debates of its community.
Digital journalists allowed their organizations to “use data to create ‘actionable intelligence from
the noise’ by making sense of the constant flood of information, and encouraging evidence based
debate” (Davis, 2013). A description of a digital journalist’s audience engagement tool connected
the two legitimating discourses: “It has been designed as a mission related investment which will
generate ongoing income through commercial activity to provide ongoing funding for its mission
of providing state of the art tools for media and democracy” (Senter, 2014).
In short order, digital journalists came to be seen as the vanguards of journalism’s new age. For
example, college and university journalism programs – long a source of legitimation for new forms
of journalism (Vos, 2012) – embraced digital journalists as the new normal for the profession. The
website for a college’s new degree in digital journalism conveyed the perceived dynamism of the
craft: “Now more than ever, we access breaking news as it happens – on news websites, blogs, and
social media, via our computers and our smart phones” (Pace, 2017). The school promised the
degree “reflects the rapid change in this industry and prepares students for a robust job market.”
Again, the legitimating discourse of vitality and modernity raised the status of digital journalism.
Likewise, industry-training courses have also portrayed digital journalists as harbingers of
a new digital age. But a “master class” by the Guardian also stressed how digital journalism
enhances traditional journalistic skills: “we explore how to tell a story from multiple angles
and present data and eyewitness testimony in compelling new formats” (Guardian, 2015). Like-
wise, digital journalists offered new tools and abilities for verification – a skill and responsibility
at the heart of professional journalism (Kovach and Rosenstiel, 2014). Digital journalists were
lauded for taking user-generated content and “work(ing) through a verification process quickly”
(Roper, 2015). Thus, the discourse about digital journalists tapped into the journalism field’s
broader legitimating traditions.
All of this is to say, if digital journalists were forming their identity based on what the broader
journalism field was saying about them, they would be forgiven for seeing themselves in the
most flattering terms. The more established digital journalists might also be forgiven if they
had not entirely forgotten the earlier days, when the title of digital journalist carried less savory
connotations.
The digital journalists’ view of their identity

Digital journalists do indeed show signs of forming their identity in light of interlocutors in
the broader journalistic field. Some still discursively position their identity in terms of earlier
criticisms – defending the journalistic value of their work relative to traditional journalism and
making a distinction between what they do and what bloggers and other interlopers do. But most
appear to have moved on – constructing their identity in ways that position digital journalists
as the vanguard of journalism. Digital journalists seem to embrace the notion that they are the
salvation of journalism. The skills, tools, abilities, and outlooks they bring to the field are what
will keep journalism astride with the times and thus keep the field alive and thriving.
Some of this identity construction reflects the recent praise of digital journalism from the
broader field, but digital journalists also have a more expansive view of their value and hence
their identity. Yes, they position themselves as wholly distinct from and superior to bloggers and
other quasi-journalistic actors, but they also promote the work of digital journalists as distinct
from and superior to the work of traditional journalists. It is traditional journalists – not digital
journalists – who will be journalism’s undoing.
44
Who am I?
Digital journalists ultimately construct their identity in terms of the roles they perform and
in terms of the skills, abilities, and outlooks that allow them to perform those roles. While digital
journalists tend to think of themselves in homogenous terms, what kind of positions the digital
journalists hold seems to introduce some variation in how their identity is constructed. However,
there is a surprising level of uniformity in how they talk about their work, their standing in the
field, and hence their professional identity.
Identity relative to the journalism field

As a whole, digital journalists see themselves as the future of the journalism industry, as the people
forging a path for new entrants into the field. They believe that what they are doing is strikingly
different from what happens at legacy media organizations. One digital journalist noted this dif-
ference by discussing economics. “Look, they’re failing and we’re winning. It seems like pride
is keeping some people from just changing to what works” (Interview, Digital Journalist 1).
Another digital journalist predicted the future of the industry, a future that did not include legacy
media. His description implicitly illustrates how he views what he does differently from what, for
example, a newspaper journalist does:
Right now, we use newspaper websites a lot. We find out a lot about what’s happening
in the city by reading their site and then through social media. So that gives us a lot
of value, you know? Those operations won’t last much longer, though, unless they find
some magic financial bullet. That won’t happen, though. When that happens, I’ll be
affected because where am I going to find out certain things to write about? Don’t get
me wrong, I dig up news a lot, but my job is really to take already-out-there news a
step further. At that point, when I can’t rely on those places, social media will take on a
larger role, and I’ll have to be more diligent in verifying what I find there.
(Interview, Digital Journalist 2)
Digital journalists saw themselves as evangelists for the gospel of audience engagement. Here,
the digital journalists found themselves forever fighting legacy media, who “come off as ‘above’
their readers” (Hare, 2017). Digital journalists see themselves as closer to the public than other
types of journalists. This is especially the case with journalists working at digitally native organi-
zations. They believe their reporting style connects them to the people they cover as opposed to
separates them. “We don’t decide what’s news on our own,” said a journalist working for a local
digital outlet. “That’s the old style of journalism. We seek out people and talk to them before we
decide what is quote ‘newsworthy’” (Interview, Digital Journalist 3). This distinction was made
even clearer by a reporter who spent a decade at his city’s main newspaper before leaving for a
digitally native news nonprofit:
Back at the paper, I had an editor who would oftentimes tell me what’s important for
the day. I would then go out and report on stories about the topic. I would ask people
in the community about the topic. Sometimes, actually more often than not, I would
come up with an idea from just being around and I’d pitch to my editor. And then
I would do the same thing with finding sources and everything else. I realized how
wrong that is when I got here. At the paper, and most places like that, the public are
just the people you talk to when you need them. Here, we stress that we always need
them. We’ll have get-togethers where we literally will bring together all types of people
and just talk to them. We’ll pick their brain. We also have them come to news meet-
ings. What I’m getting at is the public, the people we cover in (the city), we involve
45
throughout. They are part of the process as opposed to just being some people good for
quotes about what someone else deems important. Here we are part of the public, and,
importantly, they’re part of us.
Others also discussed how involving citizens in news production processes separates them
from traditional journalists. While many digital journalists interviewed stressed that legacy jour-
nalists with a proper beat talk to a lot of people, they tend to be the same people. “When I
covered education, I basically talked to the same seven or eight people. That’s who I quoted. [At
my current job], we have rules against that and basically policies that help us diversify how many
people we talk to all the time” (Interview, Digital Journalist 5).
Thus, closely related to digital journalists’ views of their relationship to the public is a strong
sense of the role that digital journalists perform – a role that made them better journalists. Digital
journalists argued that by focusing on interpretation, they are actually closer to the spirit of “true
journalism” than traditional media employees. “I turn on the television sometimes to watch local
news. Is that even news? Really? It’s just high school-level reporting. They don’t break anything.
And they certainly don’t explain anything,” one interviewee said (Interview, Digital Journalist 6).
He added, “We do the hard work. We do journalism. They just read a script that explains as
much as maybe a third-grader wants to know.” One journalist working in a legacy media news-
room in a digital capacity even criticized his own co-workers in a similar way:
We’ve been doing a lot of interesting things here lately, since we got a new editor a little
while ago. We’ve become a better news company. That’s because of us online though.
We’re publishing a lot of cool tools and interesting stories that you don’t get in the news-
paper. We won a big award in the city for one package I worked on. It did a lot of good
in the city. But the people still working primarily on the newspaper end? They just do the
same thing they’ve always done. It’s a lot of boring stories told in the same boring ways.
Finally, one digital journalist explicitly discussed what he considered the difference between
what he does and what legacy media journalists do. He said,
I would go as far as saying that what a lot of people considered journalists do isn’t even
journalism. They just repeat back what others tell them. No thinking. No context. No
explanation. [. . .] Without that, it’s not journalism to me.
Thus, digital journalists form an identity that placed them at the forefront of journalistic
excellence (see Witschge and Harbers, this volume, Chapter 5). Digital journalists have also been
asserting their deft ability in advancing the bedrock skills of journalism, such as verification.
They use sophisticated digital tools for “establishing the date or location of an image or video,
or simply corroborating written accounts of an event” (Bartlett, 2014). Digital journalists have
claimed their place as savvy, resourceful reporters and editors.
Identity relative to organization role

Digital journalists still found some old newsroom assumptions have died slowly – like the per-
sistent assumption that “digital has time to do tasks other departments ‘don’t have time to do’”
(Hare, 2017). Such attitudes were seen as throwbacks to a time when digital staff consisted of
46
Who am I?
newly hired young people who were all too eager to please veteran journalists. For the most part
however, modern digital journalists feel empowered in their organizations, and this is reflected
in their identity. Their organizational attachment also functions as a clear line of demarcation
between digital journalism and quasi-journalism, such as blogging and citizen journalism, and
thus validates their standing and identity.
Meanwhile, the jobs digital journalists hold within their organizations also influence how
they form their identity. Reporters, for example, define the concept of a digital journalist differ-
ently from editors. Journalists who work in both roles also define themselves slightly differently.
For reporters, a digital journalist is someone who gathers and then interprets information for
the public. This role of interpreter is vital and important. Digital journalists believe it sets them
apart from other types of journalists. “Just reporting what happened at some point is meaning-
less today,” said one reporter for a digitally native news nonprofit. “Breaking news, that’s what
newspapers and TV care about too much. It’s completely antiquated” (Interview, Digital Jour-
nalist 9). Numerous reporters interviewed communicated this mind-set and distinction. One
journalist who only published on digital platforms but worked for a legacy media organization
noted that her job required providing the public “with a way to understand news” (Interview,
Digital Journalist 10). She explained that citizens can “learn about what’s happening at that very
moment from peers on Twitter and the like,” but she “provide(s) a framework for understanding
what’s actually happening in the world.” Numerous digital journalists referenced social media as
something that not only changed how they work but did not change legacy journalists, to their
detriment. Said one long-time journalist working at a digitally native organization:
That’s a losing game. It’s a game journalists can never win. Yeah, maybe, every so often,
a journalist might beat the public breaking news, but it happens rarely. So many news-
paper journalists seem to spend more time trying to break news on Twitter than they do
writing interesting and important stories. Here, we don’t break anything, and we don’t
waste precious resources trying to break anything. That’s not what we do well. We
know what our audience needs, and that’s stories that beyond the ‘what happened’ and
try to figure out why it happened and how it happened and what we can learn from it.
Others echoed with similar comments, explaining that social media essentially let anyone
break news and, therefore, this effectively changes the role of a journalist. These digital reporters
believe that their identity is shaped by interpreting information, something legacy media still fail
to accomplish. One reporter did not explicitly discuss how she views her role, but implicitly did
by comparing herself to her legacy counterparts. “I only recently started here. I had to leave [the
newspaper I previously worked at] because I felt like they believed it was still 1999. All they cared
about was scoops. It did not make sense to me. A journalist [today] does not look for scoops”
(Interview, Digital Journalist 12).
But while digital journalists working as reporters form their identity around the idea of inter-
pretation and analysis, digital journalists who report and edit spoke of interpretation but also did
not disregard breaking news. “Traditionally, it’s an essential function of journalism,” one digital
journalist said. “But today, with technology the way it is, we can’t make that our journalistically
priority” (Interview, Digital Journalist 7). Another journalist, who estimated he spends 75% of
his time as an editor and 25% as a reporter, talked about covering breaking news as something
“that happens” but not as something “the people that come to the site look for.” He said:
I know what we do is a little different than what’s out there. But I’ve worked at a half-
dozen news places. I’ve been a journalist for a while. When I started in this field, a lot
47
of journalists based their reputation on muckraking, on getting that big story before
anyone else. Here and other news websites, that big story is always welcome. But it can’t
be the ‘be all, end all,’ you know? If that’s you, then you probably need to work for old-
fashioned news. Here and other places like here, that story is nice, but it’s small potatoes.
We want to give people information and tell them what to do with it. Does that make
sense? That’s a major shift I’ve seen in journalism.
Seemingly, when a digital journalist also works as an editor, there’s an understanding of how
consistently changing and updated content is essential for any website’s survival. This is poten-
tially why they see their role as journalists slightly differently from journalists only working as
reporters. This difference could be the reason that while their identity is still shaped around the
role of interpreter, primarily as a way to distance themselves from traditional journalists, they also
have a less dismissive view of dissemination. For example, one journalist explicitly discussed this,
rationalizing the performance of both functions:
I think we’re different than the New York Times or Channel 7 or whatever, ABC News
or something. All they care about, at least first and foremost, is breaking a story. They
want to own it. And we’ll do that sometimes because it generates real traffic. That does
not mean that kind of thing is our thing. Just like the news at 6 might occasionally have
a story similar to something we might run. But our stories should go beyond breaking
it. We’ll own a story, but probably from right when it’s born, from when it’s an infant,
I guess.
In general, the role that digital journalists occupy within their organization has an impact
on how they form their identity as a journalist, but only to a certain extent. The essence of
being a digital journalist concerns, in their minds, publishing a different type of news from news
published by legacy media. However, editors, who face the pressure of populating a website, still
identity – even if only slightly – with the main ethos of legacy media.
Identity relative to quasi-journalists

While digital technologies have allowed for subtle and not-so-subtle shifts in the journalism
industry and in the everyday routines of journalists, they have also let other actors enter the jour-
nalism field. For example, while media criticism was once relegated to the fringes and only pub-
lished in predominantly journalistic outlets, now blogs can perform this function (see: Vos et al.,
2012). Digital journalists define their identity partly by separating themselves from interloping
actors, people on the fringe of journalism who might be mistaken for a journalist by the general
public (see: Eldridge, 2016). “When I tell people I’m a digital journalist,” one interviewee said,
“people assume I have a blog. I don’t have a blog. A blogger is not a journalist” (Interview, Digital
Journalist 14). Another journalist noted that just because someone utilizes certain journalistic
practices, it does not make that person a journalist. “Social media lets people break news now.
They can comment on stories and even have places to publish their own stories. I’m a journalist
because I get paid by a reputable place to do journalism. Those people are not” (Interview, Digital
Journalist 15).
Many of the digital journalists interviewed expressed frustration with what they considered a
lack of the same credibility given to legacy journalism. They implicitly and explicitly argued that
48
Who am I?
this was due to the democratization of digital tools and other industry trends. “When I got my
first job in journalism,” said one interviewee with only three years’ experience,
I told my friend about it. He asked if I was a citizen journalist. I’ve read about citizen
journalism and still don’t know what it is. We shouldn’t use the word journalist to
describe something like that because it makes our jobs look trivial.
One journalist with decades of experience discussed how as a digital journalist, he could engage
with the public and make them a part of news production processes. However, later in the inter-
view, he dismissed the labeling of anyone as a journalist, especially the public.
We have this idea that anyone can be a journalist now because we can all reach people.
But that’s kind of crap. I do something very different than regular people do. I do some-
thing very different than a damn blogger. I don’t think just anyone could do what I do.
It takes experience and expertise to be a good journalist, not just access to a webpage.
Digital journalists also believe that reporting must be part of their job – something else that
sets them apart from bloggers and in league with legacy journalism. While they often stress
the importance of interpretation and analysis, it cannot be on its own. Fundamentally, digital
journalists believe there is a distinction between a journalist as someone who talks to the pub-
lic and engages with sources and someone who just writes his or her opinion. “Some people
who get called a reporter are just like Sean Hannity,” one digital journalist said (Interview,
Digital Journalist 2). “They just write an opinion. That’s not what I do or what I think my
site and many others do. We get our information from primary sources. We don’t wing it.”
Other digital journalists interviewed conveyed similar thoughts, stressing that writing about
news does not make someone a journalist: “If we really want to get down to it, I think most
people called a journalist isn’t one. And that’s because we all publish to websites. That sounds
similar. But the devil is in the details” (Interview, Digital Journalist 17). That digital journalist
went on to explain some of those “details.” She said, “Everybody has opinions. Is everyone
who mentions Hillary Clinton on Twitter a journalist? No. We do the hard work, the neces-
sary work” (Ibid.).
Journalists interviewed also made a point to distinguish themselves from what one interviewee
called “social media stars”. Those interviewed described actors on social media who aim to break
news or have created a following through targeted, curated tweets. These people are not journalists,
according to digital journalists. “There are people I follow on Twitter,” one journalist said, “and they
have hundreds of thousands of followers from basically tweeting other people’s stories. They have
ideology-based audiences. It’s no different than aggregation” (Interview, Digital Journalist 12). The
digital journalist went on to argue that like aggregation, people on social media who merely dissemi-
nate other people’s stories are not journalists. Another journalist echoed this with a similar comment.
We have well-connected people here in (my city). They have a huge audience that
follows them. And if I publish a story, it might reach x amount of people. If those
well-connected people tweet about my story, it might reach 10 times the amount. But
that doesn’t mean they’re doing the same thing as me. It doesn’t make them a journalist
because they’re spreading news to people.
49
Digital journalists clearly see bloggers and social media stars as an out-group to journalists’ in-
group. Interestingly, when digital journalists are casting these interlopers as non-journalists, they
are more apt to identity with the broader journalistic field. However, this is quickly forgotten
when they consciously focus on the place of digital journalists in the broader journalistic field.
Conclusions
“Who am I? I’m the salvation of journalism.” The digital journalists studied here are not hum-
ble. They seek to see their subgroup in positive terms, and they are supported by the broader
journalistic field in those efforts. Recall that the label of “salvation of journalism” was applied
by someone who is not a digital journalist to digital journalists. Digital journalists are, at least
recently, being socialized into a clear sense that they are vital and savvy. Job advertisements,
training programs, and professional forums and trade press constantly validate digital journalists’
elevated standing in the field.
By identifying their chief skill and outlook as audience engagement, digital journalists have
tapped into two forms of legitimating discourse. Audience engagement could be seen as engag-
ing citizens in democratic self-governance. It could also be seen as building audience and hence
aiding the profitability of the journalistic enterprise. While profitability has historically been
shaky ground for journalists’ claims of legitimacy (Vos and Finneman, 2017), its salience has been
significantly heightened, just in time for digital journalists to ride concerns about profitability to
new professional standing.
Digital journalists expressly reject journalistic autonomy in its old-fashioned sense. They also
reject the dissemination role as the normative heart of journalism. However, nearly all other
forms of traditional journalistic legitimacy are now invoked in praise of digital journalists. This
is a remarkable shift in a remarkably short period of time.
Further reading
For more on the identity of digital journalists see the 2016 article by Ferrucci and Vos. The arti-
cle theorizes the fluidity of digital journalists’ identity while also noting how a populist mobilizer
role anchors their role identity. Like this chapter, Eldridge (2016) has also examined how digital
journalists have challenged the boundaries of the journalistic field, fighting for a place alongside
traditional journalists. His 2014 article (Eldridge 2014) touches on similar themes, focusing on
the ways traditional journalists have marginalized digital upstarts such as WikiLeaks. Meanwhile,
a 2014 article by García-Avilés examines the ethical frameworks used by digital journalists –
frameworks that have relevance for digital journalistic identity.
References
Agarwal, S. D. and Barthel, M. L. (2015) “The friendly barbarians: Professional norms and work routines
of online journalists in the United States.” Journalism, 16(3), 376–391.
Anderson, C. W. (2013) Rebuilding the News: Metropolitan Journalism in the Digital Age. Philadelphia, PA:
Temple University Press.
Bartlett, R. (2014) 10 Key Skills for Digital Journalists to Hone in 2014. Retrieved from www.journalism.
co.uk/news/10-key-skills-for-digital-journalists-to-hone-in-2014/s2/a555503/
Benton, J. (2008) “ProPublica argues they’re open enough.” Nieman Journalism Lab. Retrieved from www.
niemanlab.org/2008/11/propublica-argues-theyre-open-enough/
Black, J. (2010) Who is a journalist? In C. Meyers (ed.), Journalism Ethics: A Philosophical Approach. New
York, NY: Oxford University Press (pp. 104–116).
50
Who am I?
Craft, S. and Davis, C. N. (2016) Principles of American Journalism: An Introduction (2nd ed.). New York, NY:
Routledge/Taylor & Francis Group.
Davis, K. (2013) How Data Is Transforming News. Retrieved from https://blog.wan-ifra.org/2013/06/19/
how-data-is-transforming-news
Dvorkin, L. (2013, January 25) A New Wave of Digital Journalist Is Showing a Profession the Way For-
ward. Retrieved from www.forbes.com/sites/lewisdvorkin/2013/01/25/inside-forbes-a-new-wave-of-
digital-journalist-is-showing-a-profession-the-way-forward/ – 20b8f9238919
Eldridge, S. A. (2014) “Boundary maintenance and interloper media reaction.” Journalism Studies, 15(1),
1–16. doi:10.1080/1461670X.2013.791077
Eldridge, S. A. (2016) “The digital journalist: The journalistic field, boundaries, and disquieting change.”
In B. Franklin and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon:
Routledge.
Fallows, J. M. (1996) Breaking the News: How the Media Undermine American Democracy (1st ed.). New York,
NY: Pantheon Books.
Ferrucci, P. and Vos, T. P. (2016) “Who’s in, who’s out? Constructing the identity of digital journalists.”
Digital Journalism, 5(7), 868–883. doi:10.1080/21670811.2016.1208054
Fisher, C. (2015) “The disclosure dilemma: Returning to journalism after political media advising.” Com-
munication Research and Practice, 1(1), 58–70. doi:10.1080/22041451.2015.1042425
García-Avilés, J. A. (2014) “Online newsrooms as communities of practice: Exploring digital journalists’
applied ethics.” Journal of Mass Media Ethics, 29(4), 258–272. doi:10.1080/08900523.2014.946600
The Guardian. (2015, May 7) How to be a Digital Journalist: A One-Day Bootcamp. Retrieved from www.
theguardian.com/guardian-masterclasses/2015/may/07/how-to-be-a-digital-journalist-a-one-day-
bootcamp-journalism-course
Hanitzsch, T. and Vos, T. P. (2017) “Journalistic roles and the struggle over institutional identity: The dis-
cursive constitution of journalism.” Communication Theory, 27(2), 115–135. doi:10.1111/comt.12112
Hare, K. (2017) Journalists Need to Unlearn ‘We Are the Authority.’ ‘We Are Not.’. Retrieved from www.poynter.
org/2017/local-edition-journalists-need-to-unlearn-we-are-the-authority-we-are-not/457032/
Holt, K. and Karlsson, M. (2014) “‘Random acts of journalism?’: How citizen journalists tell the news in
Sweden.” New Media & Society, 17(11), 1795–1810. doi:10.1177/1461444814535189
Job description. (2017) Retrieved from www.holdthefrontpage.co.uk/jobsboard/view/trainee-digital-
journalist-somerset-4/
Johnstone, J. W. C., Slawski, E. J. and Bowman, W. W. (1976) The News People: A Sociological Portrait of
American Journalists and Their Work. Urbana, IL: University of Illinois Press.
Kitch, C. (2014) “Historical authority and the ‘potent journalistic reputation’: A longer view of legacy-
making in American news media.” In B. Zelizer and K. Tenenboim-Weinblatt (eds.), Journalism and
Memory. London: Palgrave Macmillan (pp. 227–241).
Kopytowska, M. W. and Kalyango, Y. (2014) “Introduction: Discourse, identity, and the public sphere.” In
Y. Kalyango and M. W. Kopytowska (eds.), Why Discourse Matters: Negotiating Identity in the Mediatized
World. New York, NY: Peter Lang (pp. 1–16).
Kovach, B. and Rosenstiel, T. (2014) The Elements of Journalism: What Newspeople Should Know and the Public
Should Expect (Revised and updated 3rd ed.). New York, NY: Three Rivers Press.
Lewis, S. C. (2012) “The tension between professional control and open participation.” Information, Com-
munication & Society, 15(6), 836–866. doi:10.1080/1369118X.2012.674150
McMillian, J. C. (2011) Smoking Typewriters: The Sixties Underground Press and the Rise of Alternative Media in
AMERICA. New York, NY: Oxford University Press.
Osnos, P. (2009) “What’s a fair share in the age of Google?” Columbia Journalism Review, 25–28.
Pace. (2017) Journalism, Digital, BS. Retrieved from www.pace.edu/academics/undergraduate-students/
digital-journalism-bs
Rice, A. (2010, May 12) “Putting a price on words.” New York Times Magazine. Retrieved from www.
nytimes.com/2010/05/16/magazine/16Journalism-t.html
Roccas, S. and Brewer, M. B. (2002) “Social identity complexity.” Personality and Social Psychology Review,
6(2), 88–106. doi:10.1207/S15327957PSPR0602_01
Roper, D. (2015) In the Realm of Verification, Context Is King. Retrieved from https://blog.wan-ifra.org/
2015/11/20/in-the-realm-of-verification-context-is-king
Senter, A. (2014) 9 Innovation Projects to Present at Collab/Space New York. Retrieved from http://mediashift.
org/2014/06/9-innovation-projects-to-present-at-collabspace-new-york/
51
Sill, M. (2011) My National Press Club Talk on ‘The Case for Open Journalism Now’. Retrieved from www.ojr.
org/my-national-press-club-talk-on-the-case-for-open-journalism-now/–sthash.yZoAJcot.dpuf
Tajfel, H. (1982) Social Identity and Intergroup Relations. Cambridge and New York, NY: Cambridge Uni-
versity Press.
Turner, J. (1982) “Towards a cognitive redefinition of the social group.” In H. Tajfel (ed.), Social Identity and
Intergroup Relations. Cambridge and New York, NY: Cambridge University Press (pp. 15–40).
Ugland, E. and Henderson, J. (2007) “Who is a journalist and why does it matter? Disentangling the legal
and ethical arguments.” Journal of Mass Media Ethics, 22(4), 241–261.
Vos, T. P. (2012) “‘Homo journalisticus:’ Journalism education’s role in articulating the objectivity norm.”
Journalism: Theory, Practice & Criticism, 13(4), 435–449.
Vos, T. P. and Craft, S. (2016) “The discursive construction of journalistic transparency.” Journalism Studies,
18(12), 1505–1522. doi:10.1080/1461670X.2015.1135754
Vos, T. P., Craft, S. and Ashley, S. (2012) “New media, old criticism: Bloggers’ press criticism and the jour-
nalistic field.” Journalism, 13(7), 848–866.
Vos, T. P. and Finneman, T. (2017) “The early historical construction of journalism’s gatekeeping role.”
Journalism: Theory, Practice & Criticism, 18(3), 265–280. doi:10.1177/1464884916636126
Vos, T. P. and Singer, J. B. (2016) “Media discourse about entrepreneurial journalism: Implications for jour-
nalistic capital.” Journalism Practice, 10(2), 143–159.
Weaver, D. H., Beam, R. A., Brownlee, B. J., Voakes, P. S. and Wilhoit, G. C. (2007) The American Journalist
in the 21st Century: U.S. News People at the Dawn of a New Millennium. Mahwah, NJ: Lawrence Erlbaum
Associates.
52
4
THE DEATH OF THE
AUTHOR, THE RISE OF THE
ROBO-JOURNALIST
Authorship, bylines, and full disclosure
in automated journalism
Introduction
The rise in recent years of “automated journalism” (Carlson, 2015; Weeks, 2014) or “robot jour-
nalism” (Clerwall, 2014; Van Dalen, 2012) is based on a new and exceptional use of algorithms,
artificial intelligence software platforms, and natural language generation techniques. These algo-
rithms are able to generate textual and visual journalistic content automatically and (to some
extent) autonomously, and their output “can be fully customized to fit a customer’s voice, style
and tone” (Ghuman and Kumari, 2013: 205). Automated journalism allows media organizations to
address the areas of interest of their “long-tail” readers with virtually no additional marginal costs.
Presented as a promise to release human journalists from routine reporting, it allows them to focus
on more complex tasks, in light of the financial difficulties these organizations are experiencing.
Currently, few media organizations employ automated journalism in the way of publishing
new stories that are automatically generated by algorithms. Despite their scarce number, they
include prominent news players such as Associated Press, Forbes, ProPublica, and the Los Angeles
Times, along with a small number of prominent software companies such as Narrative Science
and Automated Insights. Typically, algorithms characterize stories that use numbers and statistics
(Carlson, 2015), such as sports recaps, real estate market analysis, weather, and earnings previews.
It is predicted to expand to additional domains while performing as a stopgap for understaffed
media organizations.
One of the most fascinating challenges of “robot journalism” that has scarcely been addressed
is the authorship of algorithmic news – who should be acknowledged as the author of automated
items? Contrary to first impressions, algorithmic authorship raises exceptionally broad theoreti-
cal issues while attempting to illuminate the dark side of the rising technology of algorithms and
its social and cultural roles. The few existing studies in this domain primarily explore the effect
of automated journalism on human journalists’ practices and relations with the robot journal-
ists (Carlson, 2015; Van Dalen, 2012; Young and Hermida, 2015) and the effect on readers’ and
journalists’ perceptions (Clerwall, 2014; Kaa and Krahmer, 2014).
53
Automated news has potential implications for both journalists and readers, highlighting the
importance of transparency in both following facets. One is disclosure transparency (Karlsson, 2010), or
the levels at which media organizations reveal how a specific news story was selected and produced.
More important is the algorithmic transparency – the methodology, construction, and limitations of
the algorithm (Coddington, 2015; Diakopoulos, 2014). Disclosing how information is produced
in terms of collection, interpretation, analysis, or the methodology used to produce the output is
considered an inherent component of information transparency and a crucial criterion for evaluat-
ing “whether such information is the product of an ethical practice” (Turilli and Floridi, 2009: 109).
There are five major justifications for thoroughly exploring the innovative domain of auto-
mated journalism: first, the practical ramifications automated journalism might have on read-
ers’ practical decisions in fields such as real estate, health, and securities; second, the sociopolitical
implications stemming from the algorithmic potential to affect the visibility of political and
social actors while maintaining a predetermined agenda (Diakopoulos, 2014; Gillespie, 2014;
Katz, 2012; Stavelin, 2013; Tufekci, 2014; Young and Hermida, 2015); third, the psychological effect
of the perceived objectivity and credibility of algorithms among news audiences (Carlson, 2015;
Gillespie, 2014; Kaa and Krahmer, 2014), as “stabilizers of trust, practical and symbolic assurances
that their evaluations are fair and accurate, free from subjectivity, error, or attempted influence”
(Gillespie, 2014: 179); fourth, the legal and ethical vicarious liability held by journalistic organi-
zations toward the readers, which is especially relevant when algorithms act as agents whose
actions are within the scope of “employment” by the journalistic organization (Vladeck, 2014);
and finally, the threatening occupational implications the technology has on the employment of
human journalists, perceived as “constrained authors” (Reich, 2010), and their autonomy over
their domain (Abbott, 1988; Baluja, 2013; Reich, 2013; Van Dalen, 2012).
Studying specifically the authorship of algorithmic news is consequential since behind the
allegedly technical attribution of this entity or another, authorship indicates deep sociocultural
perceptions of algorithms and their potential and actual roles in journalism, challenging the
traditional dominance of ethnocentric perceptions of journalistic authorship that led the rise
of bylines, mainly across the twentieth century (Reich and Boudana, 2014; Reich and Klein-
Avraham, 2014). Explorations of different attribution regimes showed their capacity to trace
historical trends and their adjustment to changes in the news environment.
This chapter will provide a multidisciplinary theoretical integration that is required in policy
research (Sola Pool, 1974). Then we will offer a new comprehensive and consistent policy – an initial
framework for understanding and treating algorithmic authorship in ways that satisfy public interest.
Literature review
Automated journalism algorithms

The increasing use of algorithms and software, and in particular artificial intelligence (hereafter:
AI), was recently dubbed the “algorithmic turn” (Napoli, 2014; Uricchio, 2011). Algorithms
are no longer “step-by-step procedures for calculations that consist of instructions and follow a
finite set of rules” (Bunz, 2013: 6) nor a “precise recipe” (MacCormick, 2012: 3). They are now
considered as a series of computational actions performed concurrently to decision-making at
various levels of uncertainty (Latar and Nordfors, 2009). AI algorithms can replace humans in a
large variety of cognitive tasks and present self-learning and autonomic decision-making abilities
(Diakopoulos, 2014; Latar and Nordfors, 2009).
Following Carlson (2015: 417), automated journalism is conceptualized here as “algorithmic
processes that convert data into narrative news texts with limited to no human intervention
beyond the initial programming”. It is another link in the chain of technological developments
54
The rise of the robo-journalist
that has accelerated changes in news production processes, organizational structures and jour-
nalistic practices (Gynnild, 2013; Primo and Zago, 2015). Algorithms are seen as the “future of
journalism” (Anderson, 2013: 1009), contributing to the different stages of content production
(Napoli, 2014) and performing most traditional journalistic functions (Dörr, 2015).
The automated journalism algorithms, rooted in TaleSpin software (Meehan, 1977), perform
tasks that were originally exclusive to humans, hence (ro)bots, on the basis of AI software plat-
forms and natural language generation (hereafter: NLG) techniques that allow the transformation
of raw data into intelligible language. The essence of the algorithmic process of robot journalism
is an automation of storytelling (Baluja, 2013; Bunz, 2013), adapted to the human understanding
of life as a sequence of continuous narratives (Ghuman and Kumari, 2013). Automated journal-
ism algorithms perform a series of steps: (a) locate and identify relevant data in databases and
other data sources; (b) “clean” and categorize the raw data; (c) identify key facts while prioritiz-
ing, comparing, and aggregating the data; (d) organize it in a semantic structure of a narrative; and
(e) distribute and publish journalistic output of textual (and occasionally visual) content, avail-
able in various styles, languages, and levels of grammatical complexity (Bunz, 2013; Dörr, 2015;
Ghuman and Kumari, 2013; Weiner, 2014). These algorithms make a substantial commercial
contribution, considering the financial difficulties of media organizations and the fact that they
help these organizations address their “long-tail” readers, interested in niche coverage domains,
with negligible additional marginal costs.
Bylines and full disclosure

Bylines, or credits, are relevant for almost any creative domain (Fisk, 2006), reflecting labor rela-
tions and employment considerations (Reich and Boudana, 2014), making them a highly interest-
ing lens through which to observe “robot journalism”. Their historical emergence in journalism
involved significant organizational, legal, political, and literary implications, reshaping hierarchy
inside news organizations and leading to journalistic stardom (Reich, 2010; Fisk, 2006). Their
growing proportion throughout the years functions as “a simple, measureable and comparable
indicator for changes in their [the journalists’] status, professionalism, creativity and authorship”
(Reich and Klein-Avraham, 2014: 619).
Bylines reflect the hominid and erratic nature of human journalistic content (Reich, 2010),
which is already suffering from diminished credibility (Maier, 2005). Studies show that the mere
presence of bylines has a measurable impact on readers’ perceptions regarding the meaning,
authority, and credibility of texts (Burkhart and Sigelman, 1990; Das and Pavlíčková, 2013).
However, prevailing byline policies tend to be anthropomorphic; because almost every news
story is bylined and attributed to a human journalist (or news agency), stories without bylines are
perceived as dubious and suspicious (Reich, 2010).
The emergence of bylines coincided with their similar though somewhat earlier develop-
ment in literature, where the author’s position is still debated. The mere idea of an author reflects
evolving trends of commercialization (Rose, 1993) and individualization (Foucault, 1980), based
on the romantic notion of the author as a genius capable of creating something new and original
(Woodmansee, 1991). The author is considered capable of “closing” the content in terms of giv-
ing it a final and conclusive meaning (Barthes, 1977; Hirsch, 1978), which may seem less relevant
in the quantitative domains of coverage for which robot journalism algorithms are used.
Against that backdrop, a full disclosure policy is defined as one type of the “information poli-
cies that require firms to provide information about their product [. . .] and service” (Winston,
2008: 705). Full disclosure of information is considered an integral part of modern public life
(Schudson, 2015), in which organizations are “conveying information [. . .] that is complete and
timely” (Kaufmann et al., 1994: 29). In our case, it is an additional way of providing details about
55
the characteristics of the automated content and its author, its data sources, and its methodology
and identifying the particular story as automated.
Full disclosure can also respond to ethical challenges such as transparency and accountability,
expressed in the different stages of automated journalistic content production (Dörr and Hol-
lnbuchner, 2015), in addition to “establish[ing] acceptable risk levels” (Fung et al., 2007: 38), as
for the previously mentioned vicarious liability.
Legal views on computer-generated works

Computer authorship was identified decades ago as a major problem by the U.S. Register of
Copyrights (1965), distinguishing between cases of using the computer merely as an assisting
tool and cases in which “traditional elements of authorship [. . .] were actually conceived and
executed not by man but by a machine” (Register of Copyrights, 1965: 5). Computer authorship
was recognized slowly, notably because of the “slow maturation of AI research” (Bridy, 2012: 22),
with an appropriate legal status (Davies, 2011).
In 1988, almost two decades after the Berne Convention, which recognized computer pro-
grams as literary works eligible for protection (Rajan, 2011), the Copyright, Designs and Patents
Act (hereafter: CDPA) was legislated in the UK, presenting, inter alia, a framework for copyright
protection of computer-generated works (hereafter: CGW) – “works that are created in total
absence of any human intervention at the time of the creation of the work” (Perry and Margoni,
2010: 3), hence protecting the output of software. These automated works are differentiated from
those in which the computer was merely a tool, from intermediate works, and from expert sys-
tems (Bainbridge, 1997; McCutcheon, 2013a; Stokes, 2009). While legislators “felt the necessity
to state [. . .] the author” (Perry and Margoni, 2010: 2) in various jurisdictions such as the UK or
New Zealand (McCutcheon, 2013a), the U.S. legislative framework “has yet to define the author
for computer-generated works” (Glasser, 2001: 25) and in case of automated journalism “will
have to choose between ruling on correspondent questions case-by-case or picking a conceptual
structure to follow” (Weeks, 2014: 94).
Nevertheless, the author of these works was defined by the CDPA as the “person by whom
the arrangements necessary for the creation of the work are undertaken” (Bainbridge, 1997: 222),
leaving room for different author candidates. Can authorship be delegated to the algorithm pro-
grammer, whose intuition and transitivity (being the author’s author) suggest he/she be recognized
as the author (Bridy, 2012), or to the user/operator (Perry and Margoni, 2010; Rajan, 2011)? Can
the algorithm itself be recognized as author (Weeks, 2014), or can both human and algorithmic
factors be considered co-authors? Is it the output designer, initiator, or the customer buying the
software (McCutcheon, 2013b)? Is it the data provider (Weeks, 2014)? Or should the work perhaps
be seen as authorless, belonging to the public domain (Davies, 2011; Perry and Margoni, 2010)?
Numerous criteria may provide a way to resolve the author question. The originality criterion,
referred to as “the overarching standard of authorship” (Ginsburg, 2002: 1077), was used to allow
rights protection for works that are not copied from previous works and made with minimal
effort and expertise. The author is, therefore, considered “the source of originality” (McCutch-
eon, 2013a: 935). However, the authorship-originality correlation may be inadequate in the case
of contemporary generative algorithmic methods such as “robot journalism”, where the minimal
effort is presented merely in the decision to generate content rather than a true intellectual effort.
Other criteria might involve the intention to generate the work, whether it is algorithm-initiated
or merely scheduled by the programmer (McCutcheon, 2013b), the ability to predict the output
(Bridy, 2012), or the functional proximity to the actual creation process (McCutcheon, 2013b).
One of the most controversial criteria is the creativity criterion, defined as the “ability to
generate novel and valuable ideas” (Boden, 2009: 24) and characterized by human action – which
56
is positioned “at the core of the creative process” (Davis, 1991: 4). The notion of creativity has
been discussed thoroughly by philosophers and historians who presented diverse views – from
an inspirational, imaginary, and free-of-rules process, to a rational goal-directed one (Paul and
Kaufman, 2014), and is considered an elementary aspect of modern journalism (Berglez, 2011;
Markham, 2012). Boden (2009) identified three types of creativity: combinational – the “unfa-
miliar combination of familiar ideas” (Boden, 2009: 24) such as painting creating a collage;
exploratory – the research of potential and limitations in a conceptual space, such as a inventing
a new cuisine; transformational – the transformation of a perceptual space by “altering (or drop-
ping) one or more of its defining dimensions” (Boden, 2009: 25), hence “giv[ing] rise to ideas
that are not only new but fundamentally different from any that went before” (Boden, 2009: 25).
However, some scholars, including Boden herself, are divided on the question of algorithmic
authorship over several aspects, such as the ability of AI algorithms to ever be truly creative or the
need for incentives such as rights protection to be given to machines (Clifford, 1996; Samuelson,
1985). Scholars argue that while technology is rapidly evolving, the legal domain requires a pro-
gression, since there is no universal formula for dealing with CGW (Rajan, 2011).
Discussion
Automated journalism, therefore, challenges the most fundamental questions underlying com-
puter authorship. The scholarly literature raises expectations for the acceptance of algorithmic
authorship in the light of the evolving autonomic capacities of AI algorithms, their growing
involvement in media content production, or the progress of attribution policies in journalism.
On the other hand, it also raises expectations for the acceptance of anthropomorphic authorship,
since bylines and their underlying principles regarding authorship, full disclosure, and transpar-
ency remain tailored to the contours of a human journalist. Both views are reflected in the ongo-
ing legal debate regarding computer-generated works and algorithmic authorship. We detect
numerous discrepancies between challenges and suggestions available in the scholarly literature,
as well as common journalistic practices (Montal and Reich, 2016).
The first discrepancy between the scholarly literature and common perceptions and byline
policies is the human-bot attribution gap. Boden (2004) expresses reservations regarding the agency
of algorithms, such as their lack of intention, while other scholars like O’Hear (1995) provide a
solid negative response to the debate on whether computers can really be creative. These scholars
suggest that it is the human factor, in terms of imagination (Baluja, 2013) or consciousness (Clif-
ford, 1996), that constitutes the core of the creative process. While there is some consistency of
these views with the anthropomorphic author perceptions reflected in practice by attributing
bylines to human entities – some news organizations and media organizations (e.g., Los Angeles
Times, Wikipedia) currently attribute the algorithm itself as the author, which might imply an
initial process of deviation from traditional perceptions of human authorship.
The second discrepancy is the transparency gap between the consensus on the importance of
transparency regarding the origins of content as viewed by scholars (Coddington, 2015; Dia-
kopoulos, 2014; Karlsson, 2010; see Coddington, Ferrer-Conill and Karlsson, and Koliska and
Diakopoulos, this volume, Chapter 19) and day-to-day journalistic practice, where organizations’
bylining policies tend to ignore the unique nature of the automated stories and employ a hap-
hazard and inconsistent crediting policy – in contrast with the unanimous support of the readers’
right to transparency expressed in the literature.
Substantial gaps also emerged between the different views in the scholarly literature, one of
which is the debate regarding the true ability of generative algorithms – the artificial intelligence
gap. Scholars from various disciplines insist that algorithms “have become performing entities:
actualities that select, evaluate, transform and produce data” (Parisi, 2013: IX), with self-learning
57
abilities. Automated journalism algorithms, in particular, are not only constantly evolving but
already considered “a form of artificial intelligence” (Weeks, 2014: 87). On the other hand, many
scholars still regard the human entity behind the algorithm as the true author (Bridy, 2012; Perry
and Margoni, 2010; Rajan, 2011) and therefore frame the algorithms’ autonomous capabilities in
a limited scope.
A similar debate can be found in the legal domain, where one can find suggestions to attri-
bute creations of AI systems to fictitious entities, such as a fictional human author (Wu, 1997),
corporate bodies (Davies, 2011), or to special legislative entities especially constituted to bear
authorship (McCutcheon, 2013b). Samuelson (1985), in contrast, suggests referring to the human
user of the software as the author, for being the closest to the position of traditional author.
Others suggest a special ‘delayed authorship’ received after the work’s completion (Bainbridge,
1997), adopting the work-made-for-hire doctrine and referring to the algorithm as an employee
(Bridy, 2012), and even to classify CGW as public domain (Davies, 2011; Perry and Margoni,
2010), a solution bearing little relevance for commercial organizations, and for news organiza-
tions in particular. This exposes the legal-journalistic gap between the repertoire of legal solutions
for addressing the uniqueness of automated content and the common journalistic approach not
to treat algorithmically-created content differently than other kinds of human authorship.
The attribution norm gap refers to another discrepancy between the common use of bylines
seen among organizations’ output that reflects the variety of byline and full disclosure policies
(Montal and Reich, 2016), and the almost universal anthropomorphic attribution norms that
prevail in Western news organizations today (Reich, 2010; Reich and Boudana, 2014). This may
indicate a conceptual vagueness probably due to the unprecedented and enigmatic nature of
algorithmic authorship, while ignoring the author’s “key role in media users’ ability to critically
use [. . .] and evaluate [. . .] media content” (Das and Pavlíčková, 2013: 381). Common journal-
istic perceptions that are consistent with the views of some scholars according to which creative
processes are collaborative by nature (i.e. Barthes, 1977; Foucault, 1980) are fixed on human
authorship and hardly applicable to algorithmic authorship.
One of the most significant discrepancies, the ‘unbiased bias’ gap, is evident between the critical
positions of scholars from various disciplines regarding the seemingly unbiased nature of algo-
rithms, and the naive positions known to be held by journalists who tend to perceive raw data as
more objective, accurate, and neutral. These scholars argue that not only algorithmic processes
but any “elaboration processes of information are not ethically neutral” (Turilli and Floridi, 2009:
109). Algorithms, and AI algorithms in particular, have the ability to change in accordance to
business decisions while hiding their underlying assumptions, ideologies, and premises (Diako-
poulos, 2014; Katz, 2012; Stavelin, 2013; Tufekci, 2014), position themselves as free of interest,
errors, or subjectivity (Chung-Yan, 2012; Gillespie, 2014; Shirky, 2009), and affect perceptions of
both journalists and readers.
This series of discrepancies emphasizes the need for a consistent and coherent policy for main-
taining authorship and full disclosure, due to the diversity of entities that are subject to attribu-
tion, such as a human programmer, software companies, news organizations, or even a bot, as well
as the incompatible perceptions presented by scholars, journalists, and the audience of readers.
Conclusion
Our research shows that algorithmic authorship is not only a novel and unexplored issue but is
also complex and enigmatic, involving crucial theoretical matters from various disciplines, such as
algorithmic creativity, computer-generated works’ legal views and attribution, and full disclosure
policies. Though only a minority of pioneering organizations already uses automated journal-
ism, it arouses enormous interest and seems to have great potential, especially considering the
58
mounting financial pressure on media organizations and their quest for rapid content generation
with lower marginal costs.
This chapter delineates the crucial need for consistent and comprehensive byline and full
disclosure policies, given the abundant and substantial discrepancies between how current users
of the technology approach algorithmic authorship and their actual attribution policies and
between our empirical findings and the scholarly literature.
A systematic policy for bylining algorithm-generated news consisting of full disclosure is
needed not only in order to adjust to the universal norms of attribution that have become an
integral feature of modern journalism (Reich, 2010; Reich and Boudana, 2014) but mainly to
minimize the dangers of ambiguous authorship and demystify the naïve perceptions shared by
audiences and journalists regarding the seemingly unbiased nature of algorithms (Chung-yan,
2012; Gillespie, 2014; Stavelin, 2013) and the possible decline in readers’ judgment and evaluation
processes (Carlson, 2015). The departure point for such policy is our assumption that full disclo-
sure containing information regarding methodology, data sources, etc., is a mandatory apparatus
to mitigate the seemingly ‘unbiased bias’ toward algorithm-generated news. Our proposed policy
identifies the uniqueness of the new technology and the questions it raises, based on the core
debate regarding algorithmic authorship and the need for a policy centered on sponsoring public
interest.
The proposed attribution policy for automated news demarcates two major levels of algorith-
mic involvement in journalistic content creation that cover the range of common possibilities in
automated journalism. First is algorithmic content generation, in which the textual (or visual) content
is produced without the involvement of a human journalist/editor. Second is integrative content
generation, in which the textual (or visual) content is produced through collaboration between
one or more human journalists and a generative algorithm. This demarcation is consistent with
views in the legal literature that highlight the importance of identifying the degree of algorith-
mic involvement in the creative process (Bainbridge, 1997; McCutcheon, 2013a; Stokes, 2009)
and the journalist-machine relations they form (Van Dalen, 2012).
To formulate the crediting policy we must consider factors such as data sources, preliminary
natural language patterns, and of course – the algorithm code, which are, after all, created and
maintained by humans (Bunz, 2013), hence leading to crediting a human entity, while disclosing
the unique characteristics of each automated story. Additional factors are the resulting liability
of the journalistic and the software organizations toward the readers and the fact that these algo-
rithms are still not fully autonomic, or still do not present truly AI capabilities, in the sense of
expressing a transformative creativity (Boden, 2009).
Our suggested attribution policy for algorithmic content generation is as follows:
• The byline should be attributed to the software vendor or to the programmer, in the case of
an individual in-house programmer.
• The full disclosure should clearly state the algorithmic nature of the content (while describing
the software vendor or the programmer’s role in the organization) and detail the data sources
of the particular story and the algorithm methodology.
In the case of integrative content generation, our suggested policy is as follows:
• The byline should be attributed to the human journalist(s) as the representative of the col-
laborative work done with the algorithm, in accordance with the anthropomorphic charac-
teristics of the modern journalistic credit.
• The full disclosure should declare the objects created by an algorithm in the particular story
(a chart, map, specific paragraph, etc.), as well as the content’s algorithmic nature (describing
59
the software vendor’s business domain or the programmer’s role), data sources of the story,
and the algorithm methodology.
The suggested attribution policy provides a comprehensive and consistent method for han-
dling automated news stories, contrary to the haphazard and inconsistent prevailing policies, tak-
ing into account public interest, journalistic values and practices, and the need for transparency
toward audiences (Karlsson, 2010) but without violating trade secrets or revealing the source code
of the algorithm(s).
Our policy accepts the notion of the human author of automated content, as argued by most
scholars, as the representative of the collaborative work for integrative content generation and as
the programming entity (programmer or software vendor) for algorithmic content generation.
The policy highlights the importance of identifying the author and crediting each piece of
automated journalistic content in accordance with the norms of modern journalism (Reich, 2010;
Reich and Boudana, 2014) and the author’s key role in the readers’ evaluation process (Das and
Pavlíčková, 2013). This combined disclosure is essential not only due to the inability of audiences
to distinguish between human and algorithmic news stories (Clerwall, 2014), but also due to the
expected decline of the capability of audiences to judge and evaluate algorithmic content (Carlson,
2015; Kaa and Krahmer, 2014) with the increasing sophistication of generative algorithms.
Our suggested policy is tailored to the current level of technological development of robot
journalism algorithms. Any significant development that enhances the autonomic capabilities
and expertise level of these algorithms, such as autonomously identifying and locating potential
data sources, finding patterns and trends or developing speech patterns, will affect the core ques-
tion of the algorithmic authorship. Legislative progress regarding computer-generated works
may invite respective adjustments of the policy as well.
The domain of automated journalism is constantly expanding in terms of organizations that
use the new technology, domains of coverage, and the creativity of the algorithms. As algorithmic
news becomes more prevalent and ‘normalized’, with broader access to the actors in the field, the
authorship question may become even more crucial. Hence the importance of implementing a
consistent and comprehensive byline and full disclosure policy such as our suggested one at the
earliest stage possible.
Further reading
This chapter has tremendously benefitted from Tal Montal and Zvi Reich’s (2016) article “I,
Robot. You, Journalist. Who Is the Author? Authorship, Bylines and Full Disclosure in Auto-
mated Journalism”, which provides an integration of multidisciplinary theoretical framework
of algorithmic creativity, bylines and full disclosure policies, legal views on computer-generated
works, and an empirical study of attribution regimes in automated journalism. Arjen van Dalen’s
(2012) “The Algorithms Behind the Headlines: How Machine Written News Redefines the
Core Skills of Human Journalists” and Matt Carlson’s (2015) “The Robotic Reporter: Automated
Journalism and the Redefinition of Labor, Compositional Forms, and Journalistic Authority”
enhance the discussion over the influence on the practices of human journalists. Christer Cler-
wall’s (2014) “Enter the Robot Journalist: Users’ Perceptions of Automated Content” and Hille
van der Kaa and Emiel Krahmer’s (2014) “Journalist Versus News Consumer: The Perceived
Credibility of Machine Written News” further examine the effect automated content has on
readers’ and journalists’ perceptions, and Michael Karlsson’s (2010) “Rituals of Transparency:
Evaluating Online News Outlets’ Uses of Transparency Rituals in the United States, United
Kingdom and Sweden” highlights the importance of transparency in contemporary journalism.
60
Regarding the legal aspect, Jani McCutcheon’s (2013) “The Vanishing Author in Computer-
Generated Works: A Critical Analysis of Recent Australian Case Law” and “Curing the Author-
less Void: Protecting Computer-Generated Works Following IceTV and Phone Directories”
provide an extensive image of the challenges and possible solutions in the legal domain for this
innovative technology. Finally, the chapter has benefitted from Zvi Reich’s (2010) “Constrained
Authors: Bylines and Authorship in News Reporting” and Zvi Reich and Sandrine Boudana’s
(2014) “The Fickle Forerunner: The Rise of Bylines and Authorship in the French Press” as well
as Zvi Reich and Inbal Klein-Avraham’s (2014) “Textual DNA: The Hindered Authorship of
Photojournalists in the Western Press”, all of which provide a comprehensive background on the
development of the journalistic byline and its relations to authorship as a concept.
References
Abbott, A. (1988) The System of Professions: An Essay on the Division of Labor. Chicago, IL: University of
Chicago Press.
Anderson, C. (2013) “Towards a sociology of computational and algorithmic journalism.” New Media and
Society, 15(7), 1005–1021.
Bainbridge, D. (1997) Software Copyright Law (3rd ed.). London: Butterworths.
Baluja, T. (2013) Robot Reporters: The New Frontier in Journalism? The Canadian Journalism Project. Retrieved
from http://j-source.ca/article/robot-reporters-new-frontier-journalism (accessed 25 September 2014).
Barthes, R. (1977) “The death of the author.” Image, Music, Text, 142–148.
Berglez, P. (2011) “Inside, outside, and beyond media logic: Journalistic creativity in climate reporting.”
Media, Culture & Society, 33(3), 449–465.
Boden, M. (2004) The Creative Mind: Myths and Mechanisms. London: Psychology Press.
Boden, M. (2009) “Computer models of creativity.” AI Magazine, 30(3), 23.
Bridy, A. (2012) “Coding creativity: Copyright and the artificially intelligent author.” Stanford Technology
Law Review, 5–10.
Bunz, M. (2013) The Silent Revolution: How Digitalization Transforms Knowledge, Work, Journalism and Politics
Without Making Too Much Noise. Basingstoke: Palgrave Macmillan.
Burkhart, F. and Sigelman, C. (1990) “Byline bias? Effects of gender on news article evaluations.” Journalism
and Mass Communication Quarterly, 67(3), 492–500.
Carlson, M. (2015) “The robotic reporter: Automated journalism and the redefinition of labor, composi-
tional forms, and journalistic authority.” Digital Journalism, 3(3), 416–431.
Chung-yan, C. (2012). “Rise of the robot journalist.” South China Morning Post. Retrieved from www.scmp.
com/article/1007279/rise-robot-journalist (accessed 24 May 2014).
Clerwall, C. (2014) “Enter the robot journalist: Users’ perceptions of automated content.” Journalism Practice,
8(5), 519–531.
Clifford, R. (1996) “Intellectual property in the era of the creative computer program: Will the true creator
please stand up.” Tulane Law Review, 71, 1675–1703.
Coddington, M. (2015) “Clarifying journalism’s quantitative turn: A typology for evaluating data journal-
ism, computational journalism, and computer-assisted reporting.” Digital Journalism, 3(3), 331–348.
Das, R. and Pavlíčková, T. (2013) “Is there an author behind this text? A literary aesthetic driven approach
to interactive media.” New Media and Society, 16(3), 381–397.
Davies, C. (2011) “An evolutionary step in intellectual property rights – Artificial intelligence and intel-
lectual property.” Computer Law and Security Review, 27(6), 601–619.
Davis, R. (1991). Intellectual Property and Software: The Assumptions are Broken (No. AI-M-1328). Cambridge,
MA: Massachusetts Institute of Technology, Artificial Intelligence Lab.
Diakopoulos, N. (2014) “Algorithmic accountability reporting: On the investigation of black boxes.”
Tow Center for Digital Journalism Brief, Columbia University.http://towcenter.org/wp-content/uploads/
2014/02/78524_Tow-Center-Report-WEB-1.pdf
Dörr, K. (2015) “Algorithmic journalism, concept and consequences.” Re-Inventing Journalism Conference
(ECREA and DGPuK), Winterthur, Switzerland, 5–6 February.
Dörr, K. and Hollnbuchner, K. (2015) “Ethical challenges of algorithmic journalism.” The Future of Journal-
ism Conference, Cardiff, Wales, 10–11 September.
Fisk, C. (2006) “Credit where it’s due: The law and norms of attribution.” Georgetown Law Journal, 95, 49.
61
Foucault, M. (1980) Language, Counter-memory, Practice: Selected Essays and Interviews. Ithaca: Cornell Uni-
versity Press.
Fung, A., Graham, M. and Weil, D. (2007) Full Disclosure: The Perils and Promise of Transparency. Cambridge:
Cambridge University Press.
Ghuman, R. and Kumari, R. (2013) “Narrative science: A review.” International Journal of Science and Research
(IJSR), 2(9), 205–207.
Gillespie, T. (2014) “The relevance of algorithms.” Media Technologies: Essays on Communication, Materiality,
and Society, 167.
Ginsburg, J. (2002) “The concept of authorship in comparative copyright law.” DePaul Law Review, 52,
1063.
Glasser, D. (2001) “iBRIEF/copyrights and trademarks: Copyrights in computer-generated works: Whom,
if anyone, do we reward?” Duke Law and Technology Review, 24–42.
Gynnild, A. (2013) “Journalism innovation leads to innovation journalism: The impact of computational
exploration on changing mindsets.” Journalism, 15(6), 713–730. doi:10.1177/1464884913486393
Hirsch, E. (1978) The Aims of Interpretation. Chicago, IL: University of Chicago Press.
Kaa, H. V. D. and Krahmer, E. (2014) “Journalist versus news consumer: The perceived credibility of
machine written news.” Proceedings of the Computation+Journalism Conference, New York, NY.
Karlsson, M. (2010) “Rituals of transparency: Evaluating online news outlets’ uses of transparency rituals in
the United States, United Kingdom and Sweden.” Journalism Studies, 11(4), 535–545.
Katz, J. (2012) “Communication perspectives on social networking and citizen journalism challenges to
traditional newspapers.” Social and Management Sciences, 19(2), 51–62.
Kaufmann, J., Kesner, I. and Lee Hazen, T. (1994) “The myth of full disclosure: A look at organizational
communications during crises.” Business Horizons, 37(4), 29–39.
Latar, N. and Nordfors, D. (2009) “Digital identities and journalism content: How artificial intelligence and
journalism may co-develop and why society should care.” Innovation Journalism, 6(7).
MacCormick, J. (2012) Nine Algorithms that Changed the Future: The Ingenious Ideas that Drive Today’s Comput-
ers. Oxford: Princeton University Press.
Maier, S. (2005) “Accuracy matters: A cross-market assessment of newspaper error and credibility.” Journal-
ism and Mass Communication Quarterly, 82(3), 533–551.
Markham, T. (2012) “The politics of journalistic creativity: Expressiveness, authenticity and de-authorization.”
Journalism Practice, 6(2), 187–200.
McCutcheon, J. (2013a). “The vanishing author in computer-generated works: A critical analysis of recent
Australian Case Law.” Melbourne University Law Review, 36; UWA Faculty of Law Research Paper No. 27.
McCutcheon, J. (2013b). “Curing the authorless void: Protecting computer-generated works follow-
ing IceTV and phone directories.” Melbourne University Law Review, 37, 46–102; UWA Faculty of Law
Research Paper No. 28.
Meehan, J. (1977) “TALE-SPIN, an interactive program that writes stories.” IJCAI, 77, 91–98.
Montal, T. and Reich, Z. (2016) “I, robot. You, journalist. Who is the author? Authorship, bylines and full
disclosure in automated journalism.” Digital Journalism, 1–21.
Napoli, P. M. (2014) “Automated media: An institutional theory perspective on algorithmic media produc-
tion and consumption.” Communication Theory, 24(3), 340–360.
O’Hear, A. (1995) “Art and technology: An old tension.” Royal Institute of Philosophy Supplement, 38, 143–158.
Parisi, L. (2013) Contagious Architecture: Computation, Aesthetics, and Space. Cambridge, MA: The MIT Press.
Paul, E. S. and Kaufman, S. B. (eds.). (2014) The Philosophy of Creativity: New Essays. Oxford: Oxford Uni-
versity Press.
Perry, M. and Margoni, T. (2010) “From music tracks to Google Maps: Who owns computer generated
works?” Law Publications, Paper 27.
Primo, A. and Zago, G. (2015) “Who and what do journalism? An actor-network perspective.” Digital
Journalism, 3(1), 38–52.
Rajan, M. T. S. (2011) Moral Rights: Principles, Practice and New Technology. Oxford: Oxford University Press.
Register of Copyrights. (1965) Sixty Eights Annual Report. Washington, DC: The Library of the Congress.
Reich, Z. (2010) “Constrained authors: Bylines and authorship in news reporting.” Journalism, 11(6),
707–725.
Reich, Z. (2013) “The impact of technology on news reporting a longitudinal perspective.” Journalism &
Mass Communication Quarterly, 90(3), 417–434.
Reich, Z. and Boudana, S. (2014) “The fickle forerunner: The rise of bylines and authorship in the French
press.” Journalism, 15(4), 407–426.
62
Reich, Z. and Klein-Avraham, I. (2014) “Textual DNA: The hindered authorship of photojournalists in the
Western press.” Journalism Practice, 8(5), 619–631.
Rose, M. (1993) Authors and Owners: The Invention of Copyright. Cambridge, MA: Harvard University Press.
Samuelson, P. (1985) “Allocating ownership rights in computer-generated works.” University of Pittsburgh
Law Review, 47, 1185.
Schudson, M. (2015) The Rise of the Right to Know: Politics and the Culture of Transparency, 1945–1975. Cam-
bridge, MA: Harvard University Press.
Shirky, C. (2009). “A speculative post on the idea of algorithmic authority.” Clay Shirkey. Retrieved
from www.shirky.com/weblog/2009/11/a-speculative-post-on-the-idea-of-algorithmic-authority/
(accessed 20 September 2014).
Sola Pool, I. (1974) “The rise of communications policy research.” Journal of Communication, 24(2), 31–42.
Stavelin, E. (2013) Computational Journalism: When Journalism Meets Programming. Unpublished doctoral dis-
sertation, University of Bergen, Norway.
Stokes, S. (2009) Digital Copyright: Law and Practice (3rd ed.). Portland, OR: Hart.
Tufekci, Z. (2014) “Engineering the public: Big data, surveillance and computational politics.” First Monday,
19(7).
Turilli, M. and Floridi, L. (2009) “The ethics of information transparency.” Ethics and Information Technology,
11(2), 105–112.
Uricchio, W. (2011) “The algorithmic turn: Photosynth, augmented reality and the changing implications
of the image.” Visual Studies, 26(1), 25–35.
Van Dalen, A. (2012) “The algorithms behind the headlines: How machine written news redefines the core
skills of human journalists.” Journalism Practice, 7(5–6), 648–658.
Vladeck, D. C. (2014) “Machines without principals: Liability rules and artificial intelligence.” Washington
Law Review, 89, 117.
Weeks, L. (2014) “Media law and copyright implications of automated journalism.” NYU Journal of Intel-
lectual Property & Entertainment Law, 4, 67–94.
Weiner, A. (2014) “Fantasy football and the cold future of robot journalism.” The Daily Dot. Retrieved from
http://kernelmag.dailydot.com/issue-sections/features-issue-sections/10097/fantasy-football-and-the-
cold-future-of-robot-journalism/ (accessed 11 November 2014).
Winston, C. (2008) “The efficacy of information policy: A review of Archon Fung, Mary Graham, and
David Weil’s full disclosure: The perils and promise of transparency.” Journal of Economic Literature, 46(3),
704–717.
Woodmansee, M. (1991) “On the author effect: Recovering collectivity.” Cardozo Arts & Entertainment Law
Journal, 10, 279.
Wu, A. J. (1997) “From video games to artificial intelligence: Assigning copyright ownership to works gen-
erated by increasingly sophisticated computer programs.” AIPLA QJ, 25, 131.
Young, M. L. and Hermida, A. (2015) “From Mr. and Mrs. Outlier to central tendencies: Computational
journalism and crime reporting at the Los Angeles Times.” Digital Journalism, 3(3), 381–397.
63
5
THE ENTREPRENEURIAL
JOURNALIST
Introduction
While journalism has for a long while been understood as a practice thoroughly embedded in
classic institutional structures, exemplified by the newsroom, we are currently witnessing a sig-
nificant shift in journalism education, critique, and practice in which the future of journalism is
envisaged to (also) lie in journalistic start-ups (see: Deuze and Witschge, 2017). In this chapter
we investigate the ways that journalistic innovators discuss their work, especially how they posi-
tion themselves in the field and vis-à-vis their ‘institutional’ counterparts. To gain insight into
the self-understanding of entrepreneurial journalists and how they differentiate their work from
traditional news outlets, we examine the discourses and practices of recent start-ups: Mediapart
(France), De Correspondent (the Netherlands), and Krautreporter (Germany). These three outlets
each argue they break with traditional ways of doing journalism. They are all online-only play-
ers that emphasize the importance of a digital platform and have strong discourses concerning
what is ‘good’ journalism. These three start-ups feature prominently in ongoing debates about
journalism in the digital age, and they are financially successful or at least sustainable. As such,
they are highly informative in exploring the practice and understanding of emerging forms of
journalism.
In this chapter, we first discuss the rise of the entrepreneur in the field of journalism, high-
lighting the main themes in the literature on entrepreneurial journalism. We then consider the
methodological implications of changes in the field and discuss methods of data collection and
analysis employed for this particular research, and in the final section we detail findings. After
considering issues relating to the start-ups’ economic context, we focus on three themes that
emerge from the data:
1 The self-understanding of these digital players: what do they consider journalism to be?
2 Journalism’s role in society: what do they deem journalism to be for?
3 Role of and relation to their audiences: how do they view and relate to their audiences?
Ultimately, this chapter discusses how these start-ups present technology as an integral tool
for innovation. These case studies show how different forms of innovation go hand-in-hand:
technological and economic innovations are not easily separated from cultural and normative
innovations.
64
The entrepreneurial journalist
Journalism and ‘the digital entrepreneur’

The ways in which journalism is produced, consumed, funded, and monetized have been rap-
idly expanding. Moreover, the places where journalistic content is produced, and by whom,
are increasingly dispersed. Developments in the field have also shaped academic, societal, and
professional understandings of journalism. Whereas journalism has long been understood as a
practice thoroughly embedded in institutional structures, increasingly the future of journalism is
anticipated to lie in journalistic start-ups, in line with a broader trend to see entrepreneurialism
as a solution to societal problems (Audretsch (2009), for example, speaks of the ‘entrepreneurial
society’). Start-ups abound (for a review of the landscape in Europe, see Bruno and Kleis Nielsen,
2012; for Australia, see Simons, 2013; for some U.S. cases, see Schaffer, 2010 and Coates Nee,
2014); courses and degrees in entrepreneurial journalism have been developed (Vázquez Schaich
and Klein, 2013; Mensing and Ryfe, 2013); and more and more academic attention is devoted to
this area (Anderson, 2014; Vos and Singer, 2016). Emphasizing individual traits, skills, and mind-
sets, the future of journalism is envisaged in the form of journalists who (alone or in collabora-
tion) are able to monetize content in innovative ways, connect to their publics in interactive new
formats, grasp opportunities, and respond to (and shape) their environment (see: Briggs, 2012).
As new ways of defining and producing journalism affect the type of news and current affairs
available in society, it is important to examine changing journalistic work practices on a micro-level
and map their implications on a wider scale. The current journalistic landscape is characterized by a
precariousness of work (Deuze, 2008), where flexible working, freelancing, and the merging of per-
sonal and professional spaces (and time) have become increasingly common. In The Netherlands,
for instance, at least half of all journalists – approximately 7,000 people – work this way (Vinken
and IJdens, 2013). Moreover, entrepreneurial journalism involves an increasingly hybrid set of roles.
Journalists do not ‘simply’ produce the news but have to conduct part or all of the other aspects of
the (economic) process as well, such as monetizing content; identifying target audiences; defining
niche markets; designing websites optimally; and maintaining networks of sources and funders.
With the blurring of roles in journalism, the combining of professions (such as journalis-
tic and marketing work), and the range of actors that work outside of traditional journalistic
institutions (as freelancers, in start-ups, or as nonpaid workers), we need to consider “how the
professional identity of journalists is discursively constructed” (Olausson, 2017: 1) and how the
activities and self-understanding inform each other. Caroline Fisher, for instance, researched how
journalists, after working as political media advisers, “perceived and managed issues of conflict
of interest” (2015: 376). Similarly, Lia-Paschalia Spyridou et al. (2013: 79), quoting Thomas
Hanitzsch’s definition of a journalist, highlight that journalistic culture “becomes manifest in the
way journalists think and act”. Journalistic self-understanding and practice are intricately related
and mutually constitutive (see: Witschge and Harbers, 2018).
We see that the professional status of journalism, which has been a long-standing issue, and
the questions concerning what is ‘proper’ journalism and what sets journalists apart have become
ever more pressing with the entry of new producers to the field – be they semi-professional ama-
teurs (Nicey, 2016), activists (Breindl, 2016), sources (Carlson, 2016), or citizen witnesses (Allan,
2016). Moreover, the many different work settings and the increasingly ‘liquid’ nature of work
practices (Deuze, 2008) also pose challenges for the conceptualization of journalistic work, not
just for who is a journalist. At the same time, we see that journalists self-identify strongly with
the ‘profession’, referring to journalism as their ‘calling’, ‘duty’ or ‘moral obligation’ (Witschge,
2013). We focus here on how start-up journalists operating in a digital environment talk about
how they conceive and practice journalism and what they judge to be the aims and functions of
journalism. The selected new forms of journalism show us in particular how they strongly invest
in the debate about what is ‘proper’ journalism and what sets them apart from their competitors.
65
Exploring digital entrepreneurial journalism as practice and discourse

We analyze three cases that express their professional ideology in opposition to their traditional
counterparts (Wagemans et al., 2016; Harbers, 2016). They present interesting cases of entrepre-
neurial journalism as they are sustainable and even successful in monetary terms, and they are, in
terms of cultural capital, acknowledged as important and innovative new players that push the
boundaries of journalism further (van der Valk, 2016; Wirtschaftsplan Krautreporter).1
Mediapart is the oldest of these, founded in 2008 by several experienced and influential jour-
nalists. At their inception, their business model – completely rejecting the sale of advertisement
space as a revenue stream – was unique and has proved successful (Wagemans et al., 2016). De
Correspondent (established in 2013) and Krautreporter (in 2014) were similarly founded by senior
and well-regarded journalists with the aim of reinvigorating quality journalism and funding it
through subscriptions.
The data used in this chapter stem from two different projects that consider start-up culture
in a comparative perspective: the research project “Beyond Journalism”, led by Mark Deuze
and Tamara Witschge, and the project “Entrepreneurship at Work”, led by Tamara Witschge.2
As part of these projects, interviews have been conducted with freelance journalists and start-up
journalists, as well as designers and web developers at the journalistic start-ups. Semi-structured
interviews were conducted, focusing on the everyday practices, routines, and experiences of these
journalists in relation to their understanding of journalism and its societal role.
For this chapter, we draw from interviews with 10 journalists, one designer and one web
developer at Mediapart, De Correspondent, and Krautreporter, conducted between 2014 and 2017.
Two interviews with journalists and editors at De Correspondent were conducted by Shermin
Chavoushi in 2015. One interview with a designer at Momkai (the digital design agency respon-
sible for the design and development of the digital platform of De Correspondent) was conducted
by Frank Harbers in 2017. Six interviews with journalists were conducted by Andrea Wagemans
at Mediapart in 2014. At Krautreporter, two journalists and one web developer were interviewed
by Frank Harbers in 2017.
To evaluate the view that proposes entrepreneurial journalism as the future of journalism,
we provide a comprehensive empirical examination of the shifting understanding of journalism
and how entrepreneurial journalists perceive their role in society and their relationship with the
audience. To this end, we adopt a practice theory approach (Couldry, 2012; Bräuchler and Postill,
2010; Schatzki, 2001) to investigate the emerging definitions of entrepreneurial journalism and
how these relate to traditional conceptualizations. With new entrants to the field, the focus both
in journalism practice and theory has been on the boundary work undertaken to protect and
gauge the conceptualization of who is a journalist (see: Carlson and Lewis, 2015). In this chapter,
we focus on the ways in which the selected start-up journalists define journalism, how they view
the role that journalism plays in society, and how they view their relationship with audiences. We
conducted a qualitative analysis of the material, discussed later.
The entrepreneurial reconceptualization of good journalism

In a chapter on journalistic entrepreneurship, one might anticipate the business element to be
core: one of the main questions that journalism as a whole is faced with is a crisis of the economic
model (Picard, 2014). In light of the decreasing revenues from advertising and decreasing audi-
ence numbers, this aspect seems to be one of the most pressing issues (see Phillips and Witschge,
2012). What is of interest in the cases we focus on in this paper is, though all of them have an
alternative business model, the economic drive is only a ‘background theme’, and the focus is
much more on the ideological rather than the commercial logic that is underlying their start-up.
66
They are critical of mainstream models of income generation, rooted in attracting a large audi-
ence and maximizing advertisement revenues. They argue that, especially in a digital age, this
business model affects the journalistic quality of legacy media and online journalistic platforms.
This concern is illustrated by a senior journalist at Mediapart:
This industrial revolution [. . .] promotes an economic model that destroys the value
of the information model and drives on entertainment. It is, so to say, a form of free
publicity aimed at the audience, for which one needs page views and clicks. [. . .]
The audience, meaning the anonymous masses, demand content that is faster, more
superficial, reaching more people, creating more buzz, and therefore [this is] a model of
entertainment for me.
(Senior journalist A, Mediapart, 2014)3
This resistance to the mainstream models of business generation translates into the business mod-
els adopted by the start-ups: they do not host advertising on their platforms. When the inter-
viewers talk about this, they use rather passionate, at times emotional language, stating they regard
advertisements as an infraction of their editorial independence and a perverse incentive for valu-
ing circulation above journalistic quality.
Denouncing the legacy business models does not, however, mean that income generation
is not an issue for these start-ups. The interviewees are clearly aware of the economic context
within which they operate and appreciate the need for financial sustainability. But they stress that
this is not a goal in itself and, as becomes clear in the earlier quote, they are critical of what they
consider acceptable models of income generation. They regard the economic interests of their
start-ups mostly as a precondition to their journalistic goals. The entrepreneurial context of the
start-up, then, is primarily seen as a necessary hurdle to be overcome, one that facilitates, but more
often restricts, their journalistic practice:
We can execute one tenth of all the ideas we have. [. . .] We are only a small club, we
don’t have expertise on every issue, I am super-busy and [name of colleague] has to do
ten things at the same time. A lot of it remains discussing it and then hoping something
will come of it, for we cannot do more yet. But the larger we become, the more we will
eventually be able to do.
(Senior editor A, De Correspondent, 2015)
As this quote illustrates, the economic context is certainly on the agenda of these journalists.
But the focus in their responses is almost invariably on the journalistic aspects of the endeavor
rather than the entrepreneurial. In contrast to other areas of entrepreneurship, journalistic start-
ups do not focus primarily on profitability. As with social entrepreneurship (cf. Dacin et al., 2011),
journalistic start-up owners and employees discuss first and foremost the importance of the social
cause. And if they do discuss economic interests, it is how they impede or facilitate reaching the
social objective of the start-up.
Such a discourse has a function: these start-ups differentiate themselves from established players
precisely on the level of the business model. As such, the remarks by journalists about economic
interests reflect how they position themselves in the field and create commercial value for these
actors (see: Wagemans et al., 2016). Though De Correspondent and Krautreporter can be considered
sustainable endeavors, they are not as financially successful yet as Mediapart. Krautreporter especially is
exploring its economic model, like many start-ups in journalism (Bruno and Kleis Nielsen, 2012).
Interview discourses on the types of journalistic aims they wish to serve through their start-
up provide us with insights into the ideological positioning, the definitions of journalism, and
67
the roles that they see for themselves in society and vis-à-vis their audiences. We discuss these
here in turn.
Defining ‘good’ journalism

As we outline earlier, entrepreneurial journalists that we interviewed emphasized the journalistic
aspects of their start-up rather than the entrepreneurial. Focusing on what journalism is or, even
more so, what it should be, the start-up journalists brand themselves as much-needed alternatives
to the established news media, claiming that they are safeguarding quality journalism within the
digital environment of twenty-first-century journalism (see also Singer, 2015). They not only
reject the business model of legacy media but also the speed and superficiality of coverage they
see as accompanying trends of that business model. In discussing what sets them apart from
traditional players, they highlight the following key elements: in-depth reporting, both in terms
of research and output, as well as editorial autonomy, measured in terms of financial ties and
independence of the news agenda.
The entrepreneurial journalists at all three start-ups stress the importance of thorough report-
ing. Interestingly, while technology has been argued to further certain types of news consump-
tion (Witschge, 2012), the interviewees explicitly reject this commonly held idea. They argue
that the affordances of the internet do not necessarily entail a mode of news consumption in
which people are only willing to read short articles that are free of charge. They proffer their
start-ups as showcases that prove the possibility of creating a sustainable form of online journal-
ism that provides in-depth reporting for which people are willing to pay. The journalists attach
great importance both to the actual time spent on researching as well as the length of the publica-
tion, as this quote by a journalist at Krautreporter illustrates:
The main problem was not what the others [online journalists] did, but what they did
not do. They didn’t spend time on big research, they didn’t spend time on long explana-
tory pieces to make sure that their readers understand the world, they didn’t spend
time to make sure that their protagonists see themselves in what they write. And they
didn’t spend time to level the playing field between themselves and their readers, and
this resulted in stories that were good, or that made sense in an economic way, because
they [generated] traffic and you could sell ads against these stories. But they were not
necessarily stories that are in the readers’ interest.
(Journalist, Krautreporter, 2017)
Not only do the journalists consider there is an interest and willingness to consume long-form
journalism, but they also believe that audiences are willing to pay for that. A journalist at Krautre-
porter (2017) said, “I just think that the journalism we do [in-depth stories] is something people
want and are ready to pay for”.
In this light, we see that start-ups do not want to rely on advertisements for their income.
They all stress the value of what they deem as independent reporting. Where this is often seen as
free from government input, the start-up journalists in this sample refer to both political ties and
financial dependencies. As a senior journalist at Mediapart puts it:
I think that the press should not be dependent on anything else than citizens’ right
to know. To do that properly, I think the ideal is to only exist for your readership.
[. . .] And therefore, we at Mediapart don’t have any advertising, we refuse all state
68
subsidization and we don’t have financial supporters that are interested in buying influ-
ence. So we have all the means to be genuinely independent.
(Senior journalist A, Mediapart, 2014)
The journalists of all three outlets emphasize that they are not as strongly tied to the institutional
news agenda and argue that they have much more freedom to choose what to write and to delve
into the issues they deem important. Journalists at Krautreporter and De Correspondent stress that
they resist the traditional news logic in terms of what is reported, as in their eyes this news logic
has resulted in an overemphasis on incidents and extraordinary events, which gives a distorted
image of the world. A senior editor (A) at De Correspondent claimed (2015), “It is more about
showing how the world works instead of being about the spectacular exceptions in the world”.
In this aim to understand how the world works, the audience is very much involved in the
journalists’ view: the start-ups envisage journalism as a joint search by the reporter and reader for
information and understanding about what is going on in the world (see the section ‘Engaging
Audiences’). As such, being transparent about the way they gather, connect, and interpret infor-
mation is one of the identifying elements of this form of journalism (see also Meier and Reimer,
2011). A journalist at Krautreporter (2017) describes journalism as a journey in which journalists
explore and gradually gain insight about a certain issue through input from and interaction with
their audiences.
This shows how these journalists do not necessarily uphold the professional values of the
objectivity regime (Harbers, 2014): the start-up journalists reject coverage that heralds detach-
ment and neutrality, saying they want to be open about the fact that they select and make choices
in the reporting process that depend on their views on the world:
I think . . . it is important to be open about the choices you make and not pretend to
show reality objectively, but you show that the way in which you look at the world is
decisive in what you convey and not.
(Senior editor B, De Correspondent, 2015)
We see the interviewees are actively redefining the values of journalism. At the same time, they
are rather reflective about this and do not identify their journalism practice as the sole new stan-
dard for quality journalism. The entrepreneurial journalists show they are aware of the phase of
exploration and testing they are in and do not claim to have definite answers – in terms of busi-
ness model or journalistic approach. They are attempting to outline an alternative but view this
as work-in-progress. Moreover, the freedom to experiment with new ways of doing journalism
seems central here: they are not so much trying to outline a definite new form as trying to break
away from traditional ideas about how journalism is understood and practiced:
At De Correspondent we try to leave behind the idea that our correspondents need to
live up to certain expectations. If we [do things differently], does that mean we still
are a legitimate medium? We need to let go of such conventions. We are exactly the
medium we can be. We don’t need to live up to a specific image. Every day we can
reinvent ourselves.
In general, the question remains concerning the extent to which these entrepreneurial journalists
are radically altering journalism practice or are rather reviving the traditional professional values
69
within an online environment. A closer look at what they consider as setting themselves apart
from legacy media can be seen as a confirmation of traditional values. The interviewees seem to
effortlessly marry time-honored views on journalism with new, alternative, professional values
and practices.
We can date back the notion of editorial independence that these entrepreneurial journal-
ists champion to the period that saw the waning of the partisan press in the nineteenth century
(Broersma, 2007), though at the same time they apply this norm in a different way. In the period
of high modernism in journalism (Hallin, 1992), independence was also a core value, but with a
diverging interpretation: in contrast to the views of the entrepreneurs in our sample, it was not
considered to conflict with a profit-oriented business model rooted in the sale of advertising.
It is important to note the importance of national context here: the investigative journalism of
Mediapart seems to fit with core conceptualizations of journalism, but this form has never been
common in the context of France (Harbers, 2014). So in this case returning to traditional values
is actually innovative and part of the niche that this start-up is carving out, diverging from what
is considered common practice for legacy media.
What is journalism for?

The blending of traditional values and new ideological conceptualizations of journalism becomes
apparent not only in the definition of what constitutes (good) journalism but also comes to
the fore in the discourse about what journalism is for. Again, journalists combine a seemingly
conservative understanding of the societal role of journalism – such as performing a watchdog
function – with those that can be said to be more unconventional – such as more activist aims
of impacting on public views more directly. We consider these here in turn, but let us empha-
size that though they may seem ‘contradictory’, these discourses are closely connected, as both
describe the way journalism has an impact on society (Bro, 2008). They both allow journalists to
give meaning to what they do and to draw boundaries around that work.
One of the traditional values entrepreneurial journalists hold dear is in-depth research and
verification of issues that are in the public interest. A senior journalist at Mediapart maintains that
Mediapart is defending “the tradition of modernity”, and echoes Kovach and Rosenstiel’s The Ele-
ments of Journalism (2001) when he describes what journalism is for:
To publish information in the public interest so that citizens are free and autonomous
in their choices. The first obligation is to the truth, the first loyalty is to public opinion
and the primary discipline is verification.
In a similar vein, entrepreneurial journalists all relate to the notion of journalism as society’s
watchdog when envisioning journalism’s function in society. The journalists at Mediapart present
this as the core task of journalism and, referencing their economic and political independence,
state:
The journalist has almost become the only person without an agenda. No political
agenda, no economic agenda. I won’t say [journalism] is the fourth estate, but [a jour-
nalist] is the person who will write down what hurts.
(Senior journalist B, Mediapart, 2017)
One of the journalists of Mediapart even refers to the American investigative reporters Wood-
ward and Bernstein who brought the Watergate scandal to light in explaining how he works.
70
This reinforces a certain type of conceptualization of journalism’s role in society, which squarely
couples journalism to democracy (McNair, 2000). Such an understanding of journalism’s role
has been dominant in Western democracies (Nerone, 2013) and can serve to limit the diversity of
both the discourses and practices of journalism (see Deuze and Witschge, 2017).
We do see, however, that the entrepreneurial journalists in our sample do not simply regurgi-
tate the discourse that ties journalism to democracy but also stretch and perhaps breach this ‘con-
ventional’ understanding. For instance, the journalists from Krautreporter and De Correspondent see
journalists as watchdogs but conceive this notion in a broader sense. They not only aim to hold
authorities in check but also to provide insight in broader societal structures (such as financial
systems) that affect our day-to-day lives. As such, their understanding goes beyond providing the
journalistic ‘what, who, where, when, and why’ to unveil wrongdoings and focus on underlying
principles that are not as manifest. They stress the importance of putting world events in context
and helping citizens understand how certain aspects of society work on a structural level.
Moreover, what these journalists aim for is different from the traditional conceptualization of
journalism. We further discuss how these journalists relate to their audience in the next section,
but it is notable that here, too, they diverge from conventional understanding in that they attribute
to the audience a more active role. In doing that, they also allocate themselves a different task: to
empower the audience by providing them with the necessarily tools to hold power to account.
So this is definitely a role of journalists, controlling the powerful, but the question is
how we do it? Do we do it by really smart, investigative, hardworking, muckraking
journalism? Or do we do it by making sure the public is informed? To give the public
the tools to control the powerful themselves? And I see myself more as being in the
second camp. So I do love good investigative stories, but I don’t think it’s the main role
journalism should be playing. I think it’s more important to make sure that the citizens
of a country are able to understand what is happening.
This is not only about carving out a niche or identity for themselves (this is often seen as the core
challenge for entrepreneurs in the field (Briggs, 2012)). Entrepreneurial journalists also aim to
change the wider understanding of journalism by their practice. Entrepreneurial journalists are
engaged professionals, convinced of the necessity of journalism as a critical voice in society. This
means keeping track of politics and business and making people aware that traditional journalism
is failing to provide what society needs. In this sense, they feel a responsibility to reflect critically
on journalism and endeavor to show an alternative. As a senior editor at Mediapart says:
So, at the core of Mediapart is investigative reporting and its subsequent editorial posi-
tions regarding the reporting, which are radically different from other French media.
This has indirectly demystified the persistent conformism of the French press.
(Senior editor, Mediapart, 2014)
Journalists do not only ascribe a different role to the audience in the constellation of journalist/
audience/society, but also to themselves. They call for a more involved type of journalism in a
number of ways, including calling for what has been dubbed by some ‘constructive journalism’
(McIntyre, 2015): journalism should not only report society’s problems but also possible solu-
tions. Especially at De Correspondent we see this understanding highlighted; as a senior editor puts
it: “So, what is our task? Maybe it is outlining new directions in society” (2015).
Extending the role of journalism, all of these entrepreneurs strive toward making an impact
in society, signaling an understanding of their relationship to their audience. Whether it is by
71
uncovering wrongs in society, offering insight in how society works by researching and contex-
tualizing events and issues, or discussing and promoting certain solutions, the interviewees strive
to make a positive impact on society. Journalism’s role in this view is ‘impacting’ rather than
‘informing’: “It is our task to breach the wall of indifference” (Senior journalist A, Mediapart,
2014). The interviewees explicitly challenge the view of journalism as confined to ‘informing’,
as this quote illustrates:
To inform lacks several things that can make journalism relevant. It lacks the interaction
with your audience, the assistance in situating issues – interpretation, involvement –
why it is important.
Finding new ways to connect to the audience is therefore a core challenge for these start-ups, and
their relation to audiences lies at the heart of their journalistic conceptions.
Engaging audiences
As we have shown, entrepreneurial journalists, in conceptualizing journalism’s role in society,
aspire to activate and transform audiences. If we look more closely at how they see their relation-
ship with audiences, we see engagement relates not only to the end state of journalism but is
central in the whole process. To some extent this is a commercially inspired aim, as journalists no
longer assume audience loyalty amidst an increasingly competitive field:
The time that it was self-evident that the public came to you, simply because they read
that particular medium is fading. [. . .] There is no longer an ‘I read everything this
particular medium serves me.’ No, for every article and for every topic and every fasci-
nation you have, you need to prove you deserve someone’s attention.
Allowing the audience an active role in the journalistic process is not simply about the lofty ideal
of audience empowerment, it also shines through in the journalists’ motivation to regain trust. At
a time when trust in journalism is not always self-evident (Broersma and Peters, 2013), the inter-
viewees stress the need to find new ways to win the trust of the public, and one of their answers
lies in reinventing their relation to the audience.
As part of the way in which they imagine the active audience, the interviewees state that audi-
ences now demand insight and a say in the journalistic process (including the beliefs and values
of the producer of a story) and a place for debate. To the journalist, trust is work-in-progress that
needs to be actively maintained, where individual journalists become much more important; they
need to actively build their own personal brand, and can no longer rely on the institution for
authority (cf. Meier and Reimer, 2011). The “large network of members” needs to be involved
and kept in the loop at all stages of the journalistic process, as a senior editor (B) at De Correspon-
dent argues (2015). Or as this Krautreporter journalist explains:
Journalists need to develop a new relationship with their readers. [. . .] It’s like a chain.
You start with that they first see you on social media, they see a picture of you in a Face-
book post by you and they say ‘Oh, who is this guy, this journalist’ and they read your
stuff. And then you take them with you when they read your articles and subscribe to
a newsletter and then they follow you how you develop your stories.
72
The conceptualization of the audience as active, however, stems not only from an instrumental
view (to attract and maintain audiences). The aspiration to let audiences participate also seems to
be inspired by a more idealistic view of what the contribution of the audience could be:
What’s also something special about De Correspondent is that we are in constant con-
versation with our readers. The moment that our healthcare correspondent put out
a request it yielded a lot of suggestions. Those suggestions do not come from other
journalists but from people out doing real things in the real world. Ultimately, that is
much more valuable than journalists sharing what they have done amongst themselves.
We could argue that crowdsourcing in the broadest sense of the term plays an important part
in the way these journalists envisage journalism. The audience is portrayed as a vast and still
untapped reservoir of ideas, expertise, experiences, and perspectives, which, if included properly,
can be engaged to improve their journalism at every step in the process.
The digital beyond technology

In this chapter we have explored the ways in which entrepreneurial journalists, who almost
exclusively publish via digital platforms, understand and practice journalism. Technology plays a
central role in start-up culture. What we would like to consider in this final reflection is where
the concept of the ‘digital’ extends beyond the technological aspect. We could argue that the
‘digital’ in the emerging forms of journalism has become a symbol of a range of economic,
cultural, and social changes at the heart of shifts in the field. Technology is part of the self-
understanding of start-ups in journalism in a way that surpasses the tools of production and
distribution that technologies denote at first glance.
In this final section we consider how entrepreneurial journalists stretch our understanding of
the kinds of journalism that are possible through the affordances of technology. In this, they can
be seen as true innovators, or pioneers, as they make visible other types of using technology than
the mainstream use in journalism. When we consider ‘affordances’ of technology not simply as
synonymous with features of technology but rather a “process-oriented, socio-technical defini-
tion of affordance” (Nagy and Neff, 2015: 1), we can start to understand these start-ups as instru-
mental in the way technologies and their affordances for journalism are imagined in the field.
To some extent, entrepreneurial journalists adopt the same type of ‘deterministic’ discourse
their traditional counterparts do (see: Witschge, 2012). The internet and digital technologies are
seen as important drivers behind the changes in journalism:
It is an ecosystem that is entirely different; in that sense, the digital revolution has radi-
cally changed the conditions for producing and disseminating information. From that
perspective, I believe it has an impact that for me has the same importance as the inven-
tion of the printing press, because the producers are not only professionals anymore and
the distribution is horizontal.
At the same time, and as much of this analysis shows, these start-ups actively use technology to
create new ways of doing journalism online. As pointed out, they do not concur with the popular
notion that online journalism is for short and fast output: they deliberately produce long-form,
investigative pieces. Moreover, where journalism has found it extremely difficult to counter the
idea that “news is for free” (see also Chyi and Yang, 2009; Chyi and Tenenboim, this volume,
73
Chapter 12), these players have developed innovative revenue models. In particular, Mediapart,
which was deemed an unrealistic endeavor at the time of launching the start-up, has been suc-
cessful in monetizing its journalistic output (Wagemans et al., 2016).
Last, and perhaps most important, the journalists seem to be true to their entrepreneurial
nature in viewing the digital environment as a dynamic place to experiment with new forms of
coverage. As a senior journalist at Mediapart claims: “We are, I think, the laboratory of what can be
the new press of the 21st century” (Senior journalist A, Mediapart, 2014). As we outlined earlier,
these journalists consider journalism in a state of flux, and collaborations with non-journalists,
such as web developers or designers, are at the core of their start-ups, which furthers innovation.
We could argue that technology becomes more of a discursive or even ideological tool rather
than a mere technical tool. Journalists interviewed actively contribute to defining ‘good’ journal-
ism, what journalism is for, and what the role of the audience is in this, and at the heart of this
reconceptualization of journalism we find technology. In this way they contribute to our under-
standing of the affordances of journalism, of technology, and of the relation to the audience. The
cases analyzed in this chapter illustrate that innovation cannot be reduced simply to technology
but rather transpires and affects a complex web of cultural, technological, and economic relations.
Further reading
Entrepreneurial journalism is a topic that has grown relatively quickly in terms of the atten-
tion it received, in terms of practice (see Briggs (2012) for a hands-on overview of the different
aspects of starting a new journalistic business), in terms of theory (see the special issue of Journal-
ism Practice from 2016 (Volume 10(2), edited by Kevin Rafter), and in terms of curriculum at
journalism schools (see Anderson, 2014; Vázquez Schaich and Klein, 2013; Mensing and Ryfe,
2013). Vos and Singer (2016) offer an insightful introduction into the current discussion about
entrepreneurial journalism, and Deuze and Witschge (2017) discuss how the trend of entrepre-
neurial journalism fits in with other changes in the journalistic field. There is a growing amount
of empirical work conducted exploring the economic sustainability of start-ups: for a review of
the landscape in Europe, see Bruno and Kleis Nielsen (2012); for Australia, see Simons (2013);
for U.S. cases, see Schaffer (2010) and Coates Nee (2014). Usher (2017) offers a fruitful analysis
of the way the innovation of start-ups from both Europe and the US is shaped and limited by
traditional conceptions and critiques of journalism. Wagemans et al. (2016) and Harbers (2016)
offer in-depth case studies of the discourse and practice of influential journalistic start-ups in
France and the Netherlands, respectively.
Notes
1 http://genossenschaft.krautreporter.de/finanzen
2 The project “Beyond Journalism” was funded by the Reynolds Journalism Institute and ran from 2015–
16; the project “Entrepreneurship at Work” runs from 2015–2020 and is funded by NWO (number
276-45-003).
3 The interviews with Mediapart were conducted in French, with Krautreporter in English, and with De Cor-
respondent in Dutch. Where relevant, we have translated the quotes used into English.
References
Allan, S. (2016) “Citizen witnesses.” In T. Witschge et al. (eds.), Sage Handbook of Digital Journalism. London:
Sage Publications (pp. 266–279).
Anderson, C. W. (2014) “The sociology of the professions and the problem of journalism education.” Radi-
cal Teacher, 99(Spring), 62–68.
74
Audretsch, D. B. (2009) “The entrepreneurial society.” The Journal of Technology Transfer, 34(3), 245–254.
Bräuchler, B. and Postill, J. (eds.). (2010) Theorising Media and Practice. Oxford: Berghahn.
Breindl,Y. (2016) “Activists as news producers.” In T. Witschge et al. (eds.), Sage Handbook of Digital Journal-
ism. London: Sage Publications (pp. 250–265).
Briggs, M. (2012) Entrepreneurial Journalism: How to Build What’s Next for News. London: Sage Publications.
Bro, P. (2008) “Normative navigation in the news media.” Journalism, 9(3), 309–329.
Broersma, M.J. (2007) “Form, style and journalistic strategies. An introduction.” In M. J. Broersma (ed.),
Form and Style in Journalism. European Newspapers and the Representation of News 1880–2005. Leuven:
Peeters (pp. ix–xxix).
Broersma, M.J. and Peters, C. (2013) “Introduction: Rethinking journalism: The structural transformation
of a public good.” In C. Peters and M.J. Broersma (eds.), Rethinking Journalism: Trust and Participation in
a Transformed News Landscape. London: Routledge (pp. 1–12).
Bruno, N. and Kleis Nielsen, R. (2012) Survival Is Success: Journalistic Online Start-ups in Western Europe.
Oxford: Reuters Institute for the Study of Journalism.
Carlson, M. (2016) “Sources as news producers.” In T. Witschge et al. (eds.), Sage Handbook of Digital Jour-
nalism. London: Sage Publications (pp. 236–249).
Carlson, M. and Lewis, S. C. (2015) Boundaries of Journalism: Professionalism, Practices and Participation. Abing-
don: Routledge.
Chyi, H. I. and Yang, M. J. (2009) “Is online news an inferior good?: Examining the economic nature of
online news among users.” Journalism and Mass Communication Quarterly, 86(3), 594–612.
Coates Nee, R. (2014) “Social responsibility theory and the digital nonprofits: Should the government aid
online news start-ups?” Journalism, 15(3), 326–343.
Couldry, N. (2012) Media, Society, World: Social Theory and Digital Media Practice. Cambridge: Polity Press.
Dacin, M. T., Dacin, P. A. and Tracey, P. (2011) “Social entrepreneurship: A critique and future directions.”
Organization Science, 22(5), 1203–1213.
Deuze, M. (2008) “The changing context of news work: Liquid journalism and monitorial citizenship.”
International Journal of Communication, 2, 848–865.
Journalism, 19(2), 165–181.
Fisher, C. (2015) “Managing conflict of interest.” Journalism Practice, 10(3), 373–386.
Hallin, D. (1992) “The passing of the ‘high modernism’ of American journalism.” Journal of Communication,
42(3), 14–25.
Harbers, F. (2014) Between Personal Experience and Detached Information: The Development of Reporting and the
Reportage in Great Britain, the Netherlands and France, 1880–2005. Groningen: s.l.
Harbers, F. (2016) “Time to engage. De Correspondent’s redefinition of journalistic quality.” Digital Journal-
ism, 4(4), 494–511.
Kovach, D. and Rosenstiel, T. (2001) The Elements of Journalism. New York, NY: Three Rivers Press.
McIntyre, K. (2015) Constructive Journalism: The Effects of Positive Emotions and Solution Information in News
Stories. Chapel Hill, NC: s.l.
McNair, B. (2000) Journalism and Democracy: An Evaluation of the Political Public Sphere. London: Routledge.
Meier, K. and Reimer, J. (2011) “Transparanz im Journalismus. Instrumente, Konfliktpotentiale, Wirkung.”
Publizistik, 56(2), 133–155.
Mensing, D. H. and Ryfe, D. M. (2013) “Blueprint for change: From the teaching hospital to the entrepre-
neurial model of journalism education.” #ISOJ The Official Research Journal of the International Symposium
on Online Journalism, 2(2), 144–161.
Nagy, P. and Neff, G. (2015) “Imagined affordance: Reconstructing a keyword for communication theory.”
Social Media + Society, 1(2), 1–9.
Nerone, J. (2013) “The historical roots of the normative model of journalism.” Journalism, 14(4), 446–458.
Nicey, J. (2016) “Semi-professional amateurs.” In T. Witschge et al. (eds.), Sage Handbook of Digital Journal-
ism. London: Sage Publications (pp. 222–235).
Olausson, U. (2017) “The reinvented journalist.” Digital Journalism, 5(1), 61–81.
Phillips, A. and Witschge, T. (2012) “The changing business of news: Sustainability of news journalism.” In
P. Lee-Wright, A. Phillips and T. Witschge (eds.), Changing Journalism. London: Routledge (pp. 3–20).
Picard, R. G. (2014) “Twilight of new dawn of journalism? Evidence from the changing news ecosystem.”
Schaffer, J. (2010) New Voices: What Works. J-Lab: The Institute for Interactive Journalism. Retrieved from
www.kcnn.org/WhatWorks/introduction/
75
Schatzki, T. R. (2001) “Introduction: Practice research.” In T. R. Schatzki, K. Knorr Cetina and E. von
Savigny (eds.), The Practice Turn in Contemporary Theory. London: Routledge (pp. 10–23).
Simons, M. (2013) What’s Next in Journalism? Brunswick: Scribe.
Singer, J. B. (2015) “Out of bounds: Professional norms as boundary makers.” In M. Carlson and S. C.
Lewis (eds.), Boundaries of Journalism: Professionalism, Practices and Participation. Abingdon: Routledge (pp.
21–36).
Spyridou, L. et al. (2013) “Journalism in a state of flux: Journalists as agents of technology innovation and
emerging news practices.” International Communication Gazette, 75(1), 76–98.
Usher, N. (2017). “Venture-backed news startups and the field of journalism. challenges, changes and con-
sistencies.” Digital Journalism, 5(9), 1116–1133.
van der Valk, L. (2016, September 27) “Het kasboek van De Correspondent.” Nieuwe Journalistiek.
Vázquez Schaich, M. J. and Klein, J. S. (2013) “Entrepreneurial journalism education: Where are we now?”
Observatorio (OBS*), 7(4), 185–211.
Vinken, H. and IJdens, T. (2013) Freelance Journalisten, Schrijvers en Fotografen. Tilburg: s.l.
Vos, T. P. and Singer, J. B. (2016) “Media discourse about entrepreneurial journalism.” Journalism Practice,
10(2), 143–159.
Wagemans, A., Witschge, T. and Deuze, M. (2016) “Ideology as resource in entrepreneurial journalism.”
Witschge, T. (2012) “The ‘tyranny’ of technology.” In P. Lee-Wright, A. Phillips and T. Witschge (eds.),
Changing Journalism. London: Routledge (pp. 99–114).
Witschge, T. (2013) “Transforming journalistic practice: A profession caught between change and tradi-
tion.” In C. Peters and M. Broersma (eds.), Rethinking Journalism: Trust and Participation in a Transformed
News Landscape. London: Routledge (pp. 160–172).
Witschge, T. and Harbers, F. (2018) “Journalism as practice.” In T. Vos (ed.), Journalism: Volume 19 of Hand-
books of Communication Science Series. Berlin: De Gruyter Mouton (pp. 105–123).
76
PART II
Digital Journalism Studies

Research design
6
CONTENT ANALYSIS OF TWITTER
Big data, big studies
With billions of users and hundreds of millions of posts and tweets per day, social media’s big data
have attracted the attention of the social sciences. This opens up unprecedented possibilities, but
also necessities, to the introduction of automated analyses in order to extract meaning and gain
orientation in this mass of communications. In recent years, Twitter has attracted growing atten-
tion from researchers in many disciplines. In their literature review of Twitter research, Williams
et al. (2013) found 575 academic papers published between 2007 and 2011 focused on Twitter.
In their literature review of journal articles on Twitter published between 2007 and 2012, Zim-
mer and Proferes (2014) found that of the total 382 papers analyzed, the majority (59%) stemmed
from computer science or information science scholars, but communications scholars came third,
accounting for 14% of the Twitter publications. And since 2012, the last year captured by the
study of Zimmer and Proferes (2014), communication research on Twitter has been further
refined, especially in the field of journalism research.
The relevance of Twitter for journalists and news media reflects its growing importance as
a source for political information and news, in particular for younger audiences (Bastos, 2015;
Newman et al., 2016; Nielsen and Schroder, 2014). According to the news use across social
media study by Pew in 2017, for example, 11% of U.S. adults used Twitter as a news source
(Shearer and Gottfried, 2017). Communications’ interest in Twitter, termed an “ambient jour-
nalism” network by Hermida (2010), also reflects the important role of Twitter for both the
dissemination of news and interaction with the audience, as well as for being a source in the
news production process. Of special interest are new agenda-setting dynamics (Russell Neu-
man et al., 2014).
With respect to the news production process, newsroom observations and interviews and
surveys with newsroom directors and journalists focus on how media and journalists use and
incorporate Twitter in their daily working routine (e.g., Cision, 2015; El Gody, 2014; Neuberger
et al., 2014a, 2014b; Thurman and Walters, 2013). For example, Neuberger et al. (2014a, 2014b)
interviewed German newsroom directors on their uses of Twitter. Thurman and Walters (2013)
conducted interviews with journalists from the British Guardian.co.uk news site, asking how
they use and link to Twitter and content-analyzed blogs on the news site. Verweij and Noort
(2014) included qualitative interviews with leading journalists and editors in their Twitter study.
Since 2001 Cision, a public relations company, has published international comparisons about
Twitter use in selected countries (Cision, 2015). According to the Cision’s (2015) global social
79
journalism study, most journalists from the surveyed six countries have included social media in
their daily working routine. The frequency of social media use is above 92% in all surveyed coun-
tries, with Facebook and Twitter leading. Twitter use among journalists was lowest in Germany
at 43% and highest in the UK, Australia, and US, where between 71% and 75% are on Twitter.
While for U.S. and UK journalists the publication and promotion of their own content is the
main reason for the use of social media, journalists in Australia, Finland, Germany, and Sweden
state that sourcing is the primary use (Cision, 2015). In their 2010 survey of German newsroom
directors, Neuberger et al. (2014a: 349) found almost all German news departments used Twitter
to attract readers (97%), for investigative purposes (94%), and for monitoring audience responses
(91%). Two-thirds of newsroom directors said that they used Twitter to interact with users (66%)
and for live coverage and breaking news (63%). In a follow-up study in 2014 (Neuberger et al.,
2014b: 48–67), online newsroom directors ascribed to Twitter the strength of being particularly
well suited to real-time interaction with their audience, for investigation, in particular for the
continuous observation of prominent sources, the search for experts and the maintenance of
expert networks, and for inquiry of facts. For short breaking news and for live reporting, they
also favored Twitter.
When Twitter is studied as a news source for journalists (Bennett, 2016; Broersma and Gra-
ham, 2013), scholars analyze how tweets are embedded in news reporting. Several studies tri-
angulate methods (Barnard, 2016; Deprez and Leuven, 2017; Revers, 2014; Verweij and Noort,
2014). Revers (2014), for example, combined an observation of reporting practices, interviews,
and analysis of tweets to study the adoption of Twitter in the everyday working practices of
reporters. In his study of journalistic practice and meta-discourse on Twitter, Barnard (2016)
applied a combination of digital ethnography and content analysis. Verweij and Noort (2014)
combined qualitative interviews with leading journalists and editors with a network analysis of
the top 500 South African journalists on Twitter. Deprez and Leuven (2017) have analyzed the
social media sourcing practices of professional health journalists. They combined in-depth inter-
views with health journalists and a content analysis using digital methods with manual coding.
To determine how news organizations and journalists perform on Twitter, studies use con-
tent analysis of Twitter accounts and tweets and manually code tweet content (e.g., Brems et al.,
2017; Canter and Brookes, 2016; Coddington et al., 2014; Cozma and Chen, 2013; Engesser and
Humprecht, 2015; Golan and Himelboim, 2016; Hanusch and Bruns, 2017; Lawrence et al., 2014;
Molyneux et al., 2017; Mourão et al., 2016; Nuernbergk, 2016). Some of these studies addition-
ally use digital methods and measure large amounts of link-, retweet-, like-, mention-, follower-,
top-hashtag-structures and/or network-structures of media’s and/or journalists’ Twitter accounts
(e.g., Chorley and Mottershead, 2016; Enli and Simonsen, 2017; Hahn et al., 2015; Larsson and
Hallvard, 2015; Majó-Vázquez et al., 2017; Nuernbergk, 2016; Vergeer, 2015).
Whereas the latter-mentioned studies harvest data from preselected media and/or journalists’
Twitter handles, other authors (e.g., Groshek and Tandoc, 2017; Kirilenko and Stepchenkova,
2014) harvest and automatically analyze the Twitter discourses on certain topics and/or hashtags
and also look at the contribution of news media to these debates. For example, Groshek and
Tandoc (2017) and Kirilenko and Stepchenkova (2014) manually code the most important users
that contribute to debates, noting whether they are media or journalists or other users in order to
analyze the role and importance of professional media and journalists in these debates.
Faris et al. (2016: 5855) utilized a mixed-methods approach combining link analysis with
qualitative content analysis in order to analyze the evolution of the net neutrality policy debate
and thereby assess the role, reach, and influence of different media sources. However, in their
qualitative content analysis, they did not analyze the tweet texts but rather the linked stories. In
order to assess the media contributions on Twitter they analyzed the shared links and found that
among the top 25 shared stories only three came from mainstream news media. Yet instead of
80
Content analysis of Twitter
using automated approaches for attributing whether an account or a URL link is from media or
journalists, these studies assign it manually.
Most of the studies mentioned analyze preselected media and journalists’ social media handles
and harvest their tweets, applying qualitative or quantitative content analysis or digital methods
including network analysis, but hardly any of these studies use other automated text-analysis
approaches, such as topic modeling, sentiment analysis, or machine learning. Malik and Pfeffer
(2016) point out that there have been only a few studies bringing computational analysis to the
study of news organizations and journalists’ use of social media. Among the exceptions are Zamith
and Lewis (2015), who address big data studies in communications and social science, and Flaounas
et al. (2013), who apply automated content analysis to digital journalism. Yet these authors do not
analyze Twitter text content. The purpose of a study by Guo et al. (2016: 332) was to “evaluate
the efficacy and validity of different computer-assisted methods for conducting journalism and
mass communication research.” They utilize and compare unsupervised topic modeling (LDA
analysis, see later in this article) and a dictionary-based analysis (search and annotate predefined
words in texts) on 77 million tweets related to the 2012 U.S. presidential elections. The authors
extracted topics and studied the association of the two candidates, Obama and Romney, to these
topics. They identify the advantages and disadvantages of both techniques and conclude that over-
all, LDA analysis performed better than the dictionary-based analysis. Yet their analysis focused on
the text content but did not examine the tweet authors; thus, no inferences could be made about
the contribution of journalists or media organizations to the Twitter debate. Using a list of more
than 6,000 pre-identified news media and journalists’ Twitter handles, Malik and Pfeffer (2016)
studied 1.8 billion tweets and found less than 1% of Twitter content is news-media related and that
news organizations mainly use Twitter as a professionalized, one-way communication medium
to promote their own reporting. However, as they themselves state, by using a predetermined list
they probably underestimate the proportion of news media. Raghuram et al. (2016) suggest an
automated solution for the endeavor of tweet author detection: they show how several machine
learning algorithms (including the support vectors machine, which will be discussed later) can be
deployed for classifying Twitter accounts and categorize them into six user groups, namely politics,
entertainment, entrepreneurship, journalism, science and technology, and health care.
In this chapter we showcase some of the automated content analysis approaches for analyzing
large-scale Twitter data, namely sentiment analysis, network text analysis, topic modeling, and
machine-learning-driven text classification. We also debate the strengths and weaknesses of these
methods. First, we discuss briefly the strategies that are used to collect Twitter data, followed by
the steps necessary to preprocess and prepare the data for automated content analysis.
Harvesting Twitter data

The way tweets are gathered in large numbers is by utilizing Twitter’s API. This acronym stands
for application programming interface, i.e., a direct connection to Twitter’s data that can be
accessed with programming code.
Tweets can be collected in real time or in retrospect. Two different real-time data collection
APIs are available. First, the Sample API provides a random 1% sample of all tweets worldwide –
at the time of writing this article, this was about 3.5 million tweets per day. Second, with the
Filter API we are able to collect more specialized data by defining search terms. The Filter API
can handle user accounts (collect all tweets from these accounts), words (collect all tweets that
include at least one of the selected words), and geographic-boundary boxes (collect all tweets
sent from within this geographic area). The REST API can collect historic tweets from specified
users or the follower/followee lists of users. Filtering tweets based on keywords is only possible
in real time.
81
The characteristics of these data access points drive the data collection strategies of researchers.
If the Filter API is used to collect tweets from a list of user accounts, researchers need to predefine
that list, such as a list of journalists’ Twitter handles. This allows for analysis of how journalists
act on Twitter and how their tweets are disseminated (see previous examples, e.g., Chorley and
Mottershead, 2016; Majó-Vázquez et al., 2017; Nuernbergk, 2016; Vergeer, 2015). On the other
hand, if we want to study the role of journalism in Twitter discourses related to specific topics,
tweets from all Twitter users are collected based on predefined lists of keywords and hashtags (e.g.,
Groshek and Tandoc, 2017; Kirilenko and Stepchenkova, 2014). Finally, the REST API’s capabil-
ity of collecting followers of any user accounts can be used to create follower-networks and to
analyze the position of journalists and media outlets in these networks.
All these data access points are free of charge and open to all researchers and practitioners
(though other fee-based options are available). The programming code for accessing Twitter data
is well developed in many programming languages and tools, so all of the previously described
approaches to collect tweets can be implemented with 10–20 lines of code, found easily in the
web. This easy and free access is the key reason why Twitter became the predominant data source
for social media studies.
All of these approaches of data collection also have limitations typical of social media data
(Ruths and Pfeffer, 2014). Morstatter et al. (2013) show that tweets collected with the Filter API
do not necessarily represent the overall activity on Twitter, and the proportion of tweets provided
is not stable. Searching for tweets in certain geographic areas is biased by the fact that only a frac-
tion of users allows adding geographic information to their tweets.
Data preprocessing
To illustrate different quantitative content analysis approaches, we utilize data from Malik and
Pfeffer’s (2016) Egypt case study. The dataset consists of about 105,000 tweets written in English
from March to June 2014, including the hashtag “#Egypt.” The authors assembled 6,103 news/
journalism-related Twitter handles from news media websites, as well as from Twitter-handle
white pages, and identified 8% of tweets in this dataset that were news related, i.e., were either
tweeted by news media and journalists, mentioned news media, or linked to websites from media
outlets. The authors applied LDA topic modeling (Blei et al., 2003) to identify topics and to show
media-related tweets were over-proportionally represented in topics related “to journalism and
press freedom” (Malik and Pfeffer, 2016: 16). In the following we describe the text preprocessing
steps that are applied for most automated content analysis procedures. Specific steps have differ-
ing importance depending on language; we focus on texts in English.
Strip case. A straightforward first step of text preparation is to make all letters lower case. This
is done to make sure that, for instance, “Egypt”, “egypt”, and “EGYPT” are handled as the same
word. Sometimes analysts are particularly interested in proper nouns. Some languages, e.g., Ger-
man, have many more words with an initial capital letter that might be of interest for the analysis
of the texts. In this case, stripping cases should be handled more deliberately.
Tokenization. The process of splitting up a given text into a set of words is called tokenization.
For short texts like tweets, this includes the removal of punctuation. For longer texts, a sentence
or a paragraph can be the entity of analysis. Then, these elements must be preserved. Specialized
tokenizers have been developed that handle Twitter data. We used TweetTokenizer from the
Natural Language Toolkit (NLTK)1 for Python (Loper and Bird, 2002) to tokenize tweets and
also remove users mentioned in the tweets.
Delete list. Twenty-five percent of English texts consist of only 17 words, e.g. “the,” “a,” “is,”
etc. When we analyze texts, we are looking for terms that distinguish different texts, words that
82
occur everywhere impede that effort. Stopword lists include these stopwords as well as discourse
markers. We utilize Ranks NL’s “default English stopwords list” and its “MySQL stopwords list,”2
resulting in a delete list of 555 words. We also removed web-links from the tweets and the case-
stripped token “#egypt,” since this was the search term for our tweet collection and can be found
in every tweet. We did not remove the hashtag symbol from any hashtags.
Term normalization is the process of unifying words that should be identical but are written
(slightly) differently. The most important procedure for normalizing words is called stemming
and refers to the removal of inflectional endings of words so that singular and plural as well as
different verb forms map to the same term (Porter, 1980). Stemming is not often used for text
preprocessing, as reducing words to their stems can result in hard-to-understand terms and can
sometimes change meaning. In inflection-rich languages, as in French, term normalization is
more important and more complicated.
Deduplication. Deduplication removes multiple occurrences of the same text. Deduplication
can be performed on texts that are “exactly” identical. While rarely the case for longer texts, it
is a reasonable approach for tweets. Deduplication of “almost” identical texts is computation-
ally very expensive and can take days for large numbers of texts, while exact deduplication can
be quickly done. For this case study, we apply exact deduplication after the already described
preprocessing steps. This removed 40% of news-related tweets and 37% of non-news tweets. In
contrast, if we were interested in analyzing the importance of certain users or topics, we would be
interested in tweets occurring multiple times resulting from retweets or from multiple references
to one online news article. Then, deduplication would be counterproductive.
After performing these preprocessing steps, the remaining tweet corpora consists of 5,101
news-related tweets with 460,000 words, and 60,168 non-news-related tweets with 6.2 million
words. A very common and easy-to-create first analysis for obtaining overview and orientation
in the text are word clouds. Figure 6.1 shows two word clouds3 of the top 100 most frequent
terms from the two sets. Word clouds resemble visual representations of frequency lists – a larger
font for terms occurring more often in the text – and the position of a word is purely to optimize
the visual aesthetics. Without discussing any details, the word clouds support Malik and Pfeffer’s
(2016) observation that news-related tweets were especially concerned with Al Jazeera journal-
ists who were arrested in late December 2013. Two more technical details related to the creation
of these word clouds should be mentioned. First, we did not remove numbers, which is a fre-
quently used option for word clouds. Consequently, the term “529”, representing the number of
Morsi supporters who were sentenced to death, is visible. Second, the tool that we used to create
these figures could not visualize Arabic text but instead showed special characters, e.g. “ØμØ”.
(a)
Figure 6.1 Word clouds with top 100 words from (a) 5,101 news-related tweets and (b) 60,168 non-news-
related tweets
83
(b)
Figure 6.1 (Continued)
Table 6.1 Sentiment analysis categories comparing news-related tweets with non-news-related tweets
LIWC category Positive emotions Negative emotions Anxiety Anger Sadness
Example words happy, good hate, enemy nervous, afraid hate, kill grief, cry, sad
News-related 3.73 4.91 0.88 2.79 0.70

Other tweets 3.97 3.88 0.68 2.16 0.57
Sentiment analysis
Sentiment analysis is frequently used to describe the importance of certain emotional and other
word categories in texts (Pang and Lee, 2008). At its core, sentiment analysis is counting the
frequency of words that have been annotated with categories. LIWC (pronounced like “Luke”),
which stands for linguistic inquiry and word count (Pennebaker et al., 2007), is the most com-
monly used sentiment dictionary and comes with an easy-to-use tool. We load our two groups
of tweets into the LIWC tool to classify the tweets’ words.
Table 6.1 shows a selection of the results and contrasts news-related tweets with their non-
news-related counterparts. Tweets discussing negative emotions seem to be more prevalent in
news-related tweets. Here, analysis was not at the tweet level. Instead two large documents were
analyzed, one consisting of all news-related tweets and one with all the other tweets.
Network text analysis

Social network analysis is interested in social actors (individuals, organizations, etc.) and their
relationships as well as topological structures emerging from these relationships. Key research
question ask for important (central) actors, or how the network can be fragmented into groups.
Network text analysis is concerned with similar questions, and analyzing text as networks pre-
dates the advent of social media (Roberts and Popping, 1996). Networks consist of nodes and
edges (Hennig et al., 2012). In text networks, nodes represent words and edges depict connec-
tions among words. A connection between two words is created when two words co-occur in
84
the same sentence, paragraph, or text. With tweets, all words of a single tweet are connected. The
advantage of treating words as nodes and co-occurrence as edge is that this allows us to analyze
text networks with network analysis tools and methods (Wasserman and Faust, 1994). In other
words, we can determine which words are central, which words form groups (topics?), and which
words connect those groups.
The 65,000 tweets for this case study create 2.4 million links among words. In general, net-
works extracted from large numbers of texts tend to be big (many nodes) and dense (many links).
These large numbers create their own challenges for handling and analyzing the networks, a
discussion of which is beyond the scope of this article. For the sake of presenting the method, we
focus on the most important words and connections of our corpora. Figure 6.2 shows the top 20
words and top 30 co-occurrence links of each group of tweets. The size of the nodes shows the
number of occurrences in each of the two groups of tweets.
In Figure 6.2, network text analysis provides us with (a) framing information for every term
by showing its connections and (b) a structural overview of the content by revealing global posi-
tions of terms (central vs. peripheral) as well as groups of terms, which increases the readability
of the analysis. For instance, three topics are clearly visible in the news-related figure. Different
topological structures of text networks can be used to study and interpret different narratives
(Bearman and Stovel, 2000) or reveal how stories from news outlets overlap with those of eyewit-
ness accounts (Martin et al., 2013). We could further analyze different terms, framing of terms, or
groups of terms by focusing on different or larger parts of the text network. Other text networks
can be extracted from tweets. For instance, a network connecting user accounts to every word
that they use would allow us to study user groups created by overlapping sets of words.
Topic modeling
The goal of topic modeling is to identify topics in large text corpora. Latent Dirichlet allocation
(LDA) (Blei et al., 2003) is the most commonly used topic modeling technique. In a nutshell,
LDA creates a predefined number of groups as well as a probability for every document and for
every word in the corpora to belong to these topics. The groups can be interpreted as topics, and
the result of an LDA calculation often shows the top words with highest probability for every
topic, which is helpful for interpreting the topics. The algorithm behind calculating LDA topics
is mathematically challenging, but several tools as well as packages for programming languages
exist, so that LDA can be performed without knowing all the technical details. Based on their
exemplary analysis of the U.S. presidential election tweets from 2012, Guo et al. (2016) conclude
that LDA topic modeling performs better than more traditional dictionary-based techniques.
However, they also stress the necessity of human interventions in LDA analysis to avoid topic
allocation errors.
For the topic modeling example and for the machine learning example in the following sec-
tion, we created a modified dataset. We concatenated all tweets per user account with at least
two tweets. This resulted in 922 texts from the news-related tweets, each representing all tweets
of a single user. To limit the complexity of these examples, we select a random subsample of the
non-news-related users to have two equally sized datasets. We ran two different LDA calculations,
one for each dataset of texts, with the tool MALLET,4 a machine learning toolkit for language.
Table 6.2 shows the results. Every topic is represented by words that are strongest in their associa-
tion with the topic. Interestingly, topics from non-news-related tweets have a larger number of
hashtag terms.
Our example can be used to discuss some of the major limitations. First, in traditional LDA,
the number of topics must be predefined as a parameter for the algorithms. Normally, we do not
know whether five or 20 or any other number of topics is a good representation for our texts.
85
#freeajstaff
day
#cairo
police military
#anticoup
today
presidential
cairo
#cairo court
brotherhood military
sisi death
#sisi al trial
president police
muslim
death
jazeera sentences
court #sisi
journalists people egyptian
#freeajstaff 2
#news
egyptian president
iran
jailed
sisi
journalist
#iran #syria
Figure 6.2 News-related tweets (left); other tweets (right). Network of top 20 word occurrences in both groups of tweets as well as their 30 strongest
connections in terms of co-occurrence within tweets.
Table 6.2 Top 10 words associated with five topics from news-related as well as other tweets
# Topics from news-related tweets Topics from other tweets
1 #freeajstaff #cairo journalists journalist #sisi egyptian sisi president elections

al trial #‫ ﻣﺼﺮ‬jailed jazeera #‫ﻓﻦ‬ brotherhood presidential #cairo killed muslim
2 military #libya cairo egyptian #tunisia death people military sentenced court years
#westernsahara #morocco #us aid killed #freeajstaff journalists today coup
3 journalists #syria #israel deadly torture #uae #kuwait #saudi #ksa #ff free #bahrain
#iraq military prison nations piece watch #qatar #syria
4 #sisi president presidential election sisi #travel #tourism #photography egypt #art
vote elections sexual #breakingnews al #design #discover_egypt_come #journey
#cairo #welcometoegypt
5 death brotherhood muslim sentences iran #iran #iraq #maryamrajavi #syria
court police sisi supporters mass coup #news #world #cnn #android #iphone
Second, LDA is a probabilistic algorithm. Different runs with the same dataset and settings will
result in (slightly) different results. Third, all words are assigned to all topics; the difference is the
order (probability) of words. Consequently, words often occur multiple times in the topics repre-
sented by the top words for each topic. Finally, humans are very good at identifying patterns and
Gestalt (Koffka, 1935), even if there are none.
Machine learning approaches to text classification

A large number of machine learning approaches can be applied to Twitter text (Raghuram
et al., 2016). In general, we distinguish between supervised and unsupervised machine learning
approaches. For supervised learning algorithms, the goal is primarily to predict a certain continu-
ous variable (regression) or class (classification). It is called “supervised” because data are available
for which the correct solution (“golden truth”) is known. Unsupervised learning algorithms
do not utilize target variables that should be predicted. Instead, inherent characteristics of the
data (e.g. groups, topics) should be revealed. For our data, we know for every tweet whether it
is news-related or not, based on Malik and Pfeffer’s (2016) definition. Consequently, we can use
this to train a machine learning model that we can then apply to new tweets to automatically
classify them into these two groups. An essential preprocessing step for every machine learning
algorithm is to extract features (variables) from the dataset that will serve as independent variables.
The most straightforward approach is to use every single word in a text corpus as a feature and
count how often every word occurs in every text. It is common for machine learning approaches
to have more variables than cases.
Support vector machines (SVM) are very popular algorithms for classifying elements because
they perform well for many different datasets (Marsland, 2009). Imagine a two-dimensional plot
with body weight and body height on the axes and the respective data points from 25 men and
25 women. In this plot, men and women would most likely form two (overlapping) clusters of
points. An SVM for this dataset is the best straight line to separate these two groups. The accuracy
of the SVM is an assessment about how many data points are on the wrong side of the separa-
tion line.
Table 6.3 shows a typical result presentation of a machine learning classification model. We
used the scikit-learn5 package in the Python programming language to perform the calculations,
and we aggregated all tweets of a single user to one case for our classification model. The accu-
racy of a prediction model is normally described with two values. Precision is the proportion of
87
Table 6.3 Accuracy of the Support Vector Machine for predicting whether a user account is news-related
or not
Precision Recall F1-Score
News-related 0.79 0.85 0.82

Other tweets 0.84 0.78 0.80
Average/total 0.81 0.81 0.81
classified elements for a group that are classified correctly. Tweets that were predicted as being
non-news-related are correctly predicted at a higher rate than news-related tweets. Recall defines
the proportion of elements of a group that can be classified correctly. These accuracy metrics
measure different aspects of the prediction and can be combined to the F1-score, which repre-
sents the harmonic mean of precision and recall.
In this example, we classified accounts for which we already knew the correct classification
to show that the content of these two different groups of accounts is different enough to cre-
ate these groups. In a next step, we could classify new tweets from new accounts based on the
previously trained classification model. Raghuram et al. (2016) use several machine learning
algorithms to classify Twitter accounts into six different groups, including journalism. While this
allows for classifying large numbers of accounts based on a rather small manually classified train-
ing dataset, critical analyses of machine learning for classifying user groups show the quality of
the results can vary heavily if the training and the testing dataset show different characteristics.
For instance, Cohen and Ruths (2013) demonstrated that classifying the political orientation of
Twitter users works well with accounts from politically active users but performs very poorly
when non-activists need to be classified.
Discussion
The dissemination of computational methods for studying news media-related content on Twit-
ter has so far been limited (Malik and Pfeffer, 2016). However, with the prevalence of social media
and the growing importance of platforms such as Twitter for the everyday work of journalists
and publishers, an increased interest in computational methods for digital journalism studies can
be expected. A major reason why analyzing tweets is popular in many scientific fields is that data
can be accessed easily and without cost via Twitter’s APIs. But while data access is easy, Twitter
data, as with data from other social media outlets, bring tremendous challenges and pitfalls that
can jeopardize the reliability of research relying on those data sources (Ruths and Pfeffer, 2014).
In this chapter, we showcased and discussed methods that are used to analyze text from mil-
lions of tweets. The application of most of these approaches is technically not very challenging.
For instance, once the data are preprocessed into the right format, programming the support vec-
tor machine for text classification and reporting the precision/recall matrix takes four (!) lines of
code. We used a machine learning toolkit that requires few technical skills. A sentiment analysis
can be wholly accomplished with an easy-to-use tool, without any coding necessary. A sufficient
number of user-friendly tools for social network analysis are also available for free. So, while these
methods are relatively easy to use, some of them are algorithmically very complex and almost
impossible to comprehend in detail for researchers from most fields. This leads to the biggest
issue related to computational methods – researchers deploying methods without considering
their limitations or preconditions for the data. For instance, individual tweets are often regarded
as too short for useful topic models (Hong and Davison, 2010); nevertheless, the literature is rich
in LDA studies based on individual tweets.
88
A major conceptual issue regarding automated content analysis is the mismatch between
what these methods are expected to do (analyze sentiment, find topics) and what they actu-
ally do (count adjectives, count word-co-occurrences). Automated content analysis methods
are far from understanding text. Instead, text is treated as a bag of words. Written text is loaded
with culture, context, and linguistic complexity, none of which can be studied with the meth-
ods described in this article. Additionally, text preprocessing and content analytical methods are
optimized for English language texts. Text in other languages with more complex inflections and
word composites can require intensive manual tasks, e.g., cleaning text and compiling a diction-
ary, in order to create meaningful results.
The reduction of messages to bags of words also ignores another important feature of social
media communication: images. The increasing trend toward “distributed content” (Newman
et al., 2016) in journalism is linked to the increasing importance of visual elements in digital
journalism and to increased space dedicated to visuals within journalistic content circulated via
social media, such as Twitter. Images can have many roles in messages: they can be rather decora-
tive elements, illustrate what is shown in the verbal text, or convey central arguments of the mes-
sage visually (Brantner et al., 2011). Automatically analysis of image content is developing but is
mainly concentrated on the motifs of visuals. It is a first step, but only partially helpful, because
images do not convey messages only by what is shown but also by how it is shown – features that
can contribute to the visual framing of issues and their evaluations (see: Coleman and Wu, 2016).
Approaches to automated analysis of image content can be found mainly in computer science,
especially in the fields of image processing, pattern recognition, and computer vision, but have
recently been used also in communication research (see: Etlinger, 2017; Peng, 2017). Another
methodological challenge is that visuals and verbal text are complexly entangled in multimodal
media messages and influence each other.
In accordance with Lewis and colleagues (2013; Zamith and Lewis, 2015), we encourage
journalism researchers dealing with big data from social media to combine computational and
manual methods. That is, computational methods have to complement established research tech-
niques instead of replacing them. For example, a deliberate triangulation of the computational
analysis of big data with manually coded smaller samples can bring in the advantages of both
methodologies. The combination of proven and innovative social sciences methods with compu-
tational methods is indispensable to ensure validity and reliability and to further enhance theo-
retically and empirically well-founded assessments of the current state of (online) journalism.
Further reading
For analysis of the advantages of LDA analysis compared to dictionary-based text analysis for
Twitter research see Guo et al. (2016); for the utilization of machine learning algorithms to clas-
sify Twitter accounts see Raghuram et al. (2016); see Lewis et al. (2013) and Zamith and Lewis
(2015) on triangulation of computational and manual methods. Malik and Pfeffer (2016) pro-
vide a case study of journalistic agenda setting on Twitter applying LDA topic modeling. Ruths
and Pfeffer (2014) summarize challenges arising from using social media data to study human
behavior.
Notes
1 www.nltk.org/
2 www.ranks.nl/stopwords
3 Created with www.wordle.net/
4 http://mallet.cs.umass.edu/
5 http://scikit-learn.org/stable/index.html
89
References
Barnard, S. R. (2016) “‘Tweet or be sacked’: Twitter and the new elements of journalistic practice.” Journal-
ism, 17(2), 190–207.
Bastos, M. T. (2015) “Shares, pins, and tweets: News readership from daily papers to social media.” Journalism
Studies, 16(3), 305–325.
Bearman, P. S. and Stovel, K. (2000) “Becoming a Nazi: A model for narrative networks.” Poetics, 27(2),
69–90.
Bennett, D. (2016) “Sourcing the BBC’s live online coverage of terror attacks.” Digital Journalism, 4(7),
861–874.
Blei, D. M., Ng, A. Y. and Jordan, M. I. (2003) “Latent dirichlet allocation.” The Journal of Machine Learning
Research, 3, 993–1022.
Brantner, C., Lobinger, K. and Wetzstein, I. (2011) “Effects of visual framing on emotional responses and
evaluations of news stories about the Gaza conflict 2009.” Journalism & Mass Communication Quarterly,
88(3), 523–540.
Brems, C., Temmerman, M., Graham, T. and Broersma, M. (2017) “Personal branding on Twitter.” Digital
Journalism, 5(4), 443–459.
Broersma, M. and Graham, T. (2013) “Twitter as a news source: How Dutch and British newspapers used
tweets in their news coverage, 2007–2011.” Journalism Practice, 7(4), 446–464.
Canter, L. and Brookes, D. (2016) “Twitter as a flexible tool.” Digital Journalism, 4(7), 875–885.
Chorley, M. J. and Mottershead, G. (2016) “Are you talking to me?” Journalism Practice, 10(7), 856–867.
Cision. (2015) 2015 Global Social Journalism Study. Retrieved from www.cision.com/us/resources/
white-papers/2015-global-social-journalism-study/
Coddington, M., Molyneux, L. and Lawrence, R. G. (2014) “Fact checking the campaign: How political
reporters use Twitter to set the record straight (or not).” The International Journal of Press/Politics, 19(4),
391–409.
Cohen, R. and Ruths, D. (2013) “Classifying political orientation on Twitter: It’s not easy!” Proceedings of
International AAAI Conference on Weblogs and Social Media (pp. 91–99).
Coleman, R. and Wu, D. (2016) Image and Emotion in Voter Decisions: The Affect Agenda. Lanham, MD: Lex-
ington Books.
Cozma, R. and Chen, K.-J. (2013) “What’s in a tweet? Foreign correspondents’ use of social media.” Journal-
ism Practice, 7(1), 33–46.
Deprez, A. and Leuven, S. V. (2017) “About pseudo quarrels and trustworthiness.” Journalism Studies, 1–18.
El Gody, A. (2014) “The use of information and communication technologies in three Egyptian news-
rooms.” Digital Journalism, 2(1), 77–97.
Engesser, S. and Humprecht, E. (2015) “Frequency or skillfulness.” Journalism Studies, 16(4), 513–529.
Enli, G. and Simonsen, C.-A. (2017) “‘Social media logic’ meets professional norms: Twitter hashtags usage
by journalists and politicians.” Information, Communication & Society, 1–16.
Etlinger, S. (2017) “Language of the eye: How computer vision is remaking social media.” Presented at the
Annual Meeting of the International Communication Association, San Diego, CA.
Faris, R., Roberts, H., Etling, B., Othman, D. and Benkler, Y. (2016) “The role of the networked public
sphere in the U.S. net neutrality policy debate.” International Journal of Communication, 10, 5839–5864.
Flaounas, I., Ali, O., Lansdall-Welfare, T., De Bie, T., Mosdell, N., Lewis, J. and Cristianini, N. (2013)
“Research methods in the age of digital journalism: Massive-scale automated analysis of news-content –
Topics, style and gender.” Digital Journalism, 1(1), 102–116.
Golan, G. J. and Himelboim, I. (2016) “Can world system theory predict news flow on Twitter? The case of
government-sponsored broadcasting.” Information, Communication & Society, 19(8), 1150–1170.
Groshek, J. and Tandoc, E. (2017) “The affordance effect: Gatekeeping and (non)reciprocal journalism on
Twitter.” Computers in Human Behavior, 66, 201–210.
Guo, L., Vargo, C. J., Pan, Z., Ding, W. and Ishwar, P. (2016) “Big social data analytics in journalism and mass
communication: Comparing dictionary-based text analysis and unsupervised topic modeling.”Journalism &
Mass Communication Quarterly, 93(2), 332–359.
Hahn, K. S., Ryu, S. and Park, S. (2015) “Fragmentation in the Twitter following of news outlets: The repre-
sentation of South Korean users’ ideological and generational cleavage.” Journalism & Mass Communication
Quarterly, 92(1), 56–76.
Hanusch, F. and Bruns, A. (2017) “Journalistic branding on Twitter.” Digital Journalism, 5(1), 26–43.
Hennig, M., Brandes, U., Pfeffer, J. and Mergel, I. (2012) Studying Social Networks: A Guide to Empirical
Research. Frankfurt: Campus Verlag.
90
Hermida, A. (2010) “Twittering the news: The emergence of ambient journalism.” Journalism Practice, 4(3),
297–308.
Hong, L. and Davison, B. D. (2010) “Empirical study of topic modeling in Twitter.” Proceedings of Workshop
on Social Media Analytics (pp. 80–88).
Kirilenko, A. P. and Stepchenkova, S. O. (2014) “Public microblogging on climate change: One year of
Twitter worldwide.” Global Environmental Change, 26, 171–182.
Koffka, K. (1935) Principles of Gestalt Psychology. New York, NY: Harcourt Brace & World.
Larsson, A. O. and Hallvard, M. (2015) “Bots or journalists? News sharing on Twitter.” Communications,
40(3), 361–370.
Lawrence, R. G., Molyneux, L., Coddington, M. and Holton, A. (2014) “Tweeting conventions: Political
journalists’ use of Twitter to cover the 2012 presidential campaign.” Journalism Studies, 15(6), 789–806.
Lewis, S. C., Zamith, R. and Hermida, A. (2013) “Content analysis in an era of big data: A hybrid approach
to computational and manual methods.” Journal of Broadcasting & Electronic Media, 57(1), 34–52.
Loper, E. and Bird, S. (2002) “NLTK: The Natural Language Toolkit.” Proceedings of Workshop on Effective Tools
and Methodologies for Teaching Natural Language Processing and Computational Linguistics – Volume 1 (pp. 63–70).
Majó-Vázquez, S., Zhao, J. and Nielsen, R. K. (2017) “The digital-born and legacy news media on Twitter
during the French presidential election.” Reuters Institute for the Study of Journalism. Retrieved
from http://reutersinstitute.politics.ox.ac.uk/our-research/digital-born-and-legacy-news-media-twitter-
during-french-presidential-elections
Malik, M. M. and Pfeffer, J. (2016) “A macroscopic analysis of news content in Twitter.” Digital Journalism,
4(8), 955–979.
Marsland, S. (2009) Machine Learning: An Algorithmic Perspective. Boca Raton, FL: Taylor & Francis Ltd.
Martin, M. K., Pfeffer, J. and Carley, K. M. (2013) “Network text analysis of conceptual overlap in inter-
views, newspaper articles and keywords.” Social Network Analysis and Mining, 3(4), 1165–1177.
Molyneux, L., Holton, A. and Lewis, S. C. (2017) “How journalists engage in branding on Twitter: Indi-
vidual, organizational, and institutional levels.” Information, Communication & Society, 1–16.
Morstatter, F., Pfeffer, J., Liu, H. and Carley, K. M. (2013) “Is the sample good enough? Comparing data
from Twitter’s streaming API with Twitter’s firehose.” Proceedings of International Conference on Weblogs and
Social Media (pp. 400–408).
Mourão, R., Diehl, T. and Vasudevan, K. (2016) “I love big bird. How journalists tweeted humor during the
2012 presidential debates.” Digital Journalism, 4(2), 211–228.
Neuberger, C., Jo vom Hofe, H. and Nuernbergk, C. (2014a) “The use of Twitter by professional journal-
ists. Results of a newsroom survey in Germany.” In K. Weller, A. Bruns, J. Burgess, M. Mahrt and C.
Puschmann (eds.), Twitter and Society. New York, NY: Peter Lang (pp. 345–357).
Neuberger, C., Langenohl, S. and Nuernbergk, C. (2014b) “Redaktionsbefragung.” In C. Neuberger, S.
Langenohl and C. Nuernbergk (eds.), Social Media und Journalismus. Düsseldorf: LfM (pp. 34–91).
Newman, N., Fletcher, R., Levy, D. A. L. and Nielsen, R. K. (2016) “Reuters Institute digital news report
2016.” Reuters Institute for the Study of Journalism. Retrieved from https://reutersinstitute.politics.ox.ac.
uk/sites/default/files/Digital-News-Report-2016.pdf
Nielsen, R. K. and Schroder, K. C. (2014) “The relative importance of social media for accessing, finding,
and engaging with news: An eight-country cross-media comparison.” Digital Journalism, 2(4), 472–489.
Nuernbergk, C. (2016) “Political journalists’ interaction networks.” Journalism Practice, 10(7), 868–879.
Pang, B. and Lee, L. (2008) “Opinion mining and sentiment analysis.” Foundations and Trends in Information
Retrieval, 2(1–2), 1–135.
Peng,Y. (2017) “When images meet codes: Applying computer vision methods in communication research.”
Presented at the Annual Meeting of the International Communication Association, San Diego, CA.
Pennebaker, J. W., Booth, R. J. and Francis, M. E. (2007) Linguistic Inquiry and Word Count: LIWC [Com-
puter software]. Austin, TX: Liwc. Net.
Porter, M. F. (1980) “An algorithm for suffix stripping.” Program, 14(3), 130–137.
Raghuram, M. A., Akshay, K. and Chandrasekaran, K. (2016) “Efficient user profiling in Twitter social net-
work using traditional classifiers.” In S. Berretti, S. M. Thampi and S. Dasgupta (eds.), Intelligent Systems
Technologies and Applications (Vol. 2). Berlin: Springer-Verlag Berlin (pp. 399–411).
Revers, M. (2014) “The twitterization of news making: Transparency and journalistic professionalism.”
Journal of Communication, 64(5), 806–826.
Roberts, C. W. and Popping, R. (1996) “Themes, syntax and other necessary steps in the network analysis
of texts: A research paper.” Social Science Information, 35(4), 657–665.
Russell Neuman, W., Guggenheim, L., Mo Jang, S. and Bae, S. Y. (2014) “The dynamics of public attention:
Agenda-setting theory meets big data.” Journal of Communication, 64(2), 193–214.
91
Ruths, D. and Pfeffer, J. (2014) “Social media for large studies of behavior.” Science, 346(6213), 1063–1064.
Shearer, E. and Gottfried, J. (2017, September 7) News Use Across Social Media Platforms 2017. Retrieved
from www.journalism.org/2017/09/07/news-use-across-social-media-platforms-2017/
Thurman, N. and Walters, A. (2013) “Live blogging? Digital journalisms pivotal platform? A case study
of the production, consumption, and form of Live Blogs at Guardian.co.uk.” Digital Journalism, 1(1),
82–101.
Vergeer, M. (2015) “Peers and sources as social capital in the production of news: Online social networks as
communities of journalists.” Social Science Computer Review, 33(3), 277–297.
Verweij, P. and Noort, E. van. (2014) “Journalists’ Twitter networks, public debates and relationships in
South Africa.” Digital Journalism, 2(1), 98–114.
Wasserman, S. and Faust, K. (1994) Social Network Analysis: Methods and Applications. Cambridge, MA: Cam-
bridge University Press.
Williams, S. A., Terras, M. M. and Warwick, C. (2013) “What do people study when they study Twitter?
Classifying Twitter related academic papers.” Journal of Documentation, 69(3), 384–410.
Zamith, R. and Lewis, S. C. (2015) “Content analysis and the algorithmic coder: What computational social
science means for traditional modes of media analysis.” The ANNALS of the American Academy of Political
and Social Science, 659(1), 307–318.
Zimmer, M. and Proferes, N. J. (2014) “A topology of Twitter research: Disciplines, methods, and ethics.”
Aslib Journal of Information Management, 66(3), 250–261.
92
7
INNOVATION IN CONTENT
ANALYSIS
Freezing the flow of liquid news
Rodrigo Zamith
As scholars of journalism take greater interest in analyzing online news content, they find them-
selves facing novel methodological challenges (Karlsson and Sjøvaag, 2016). One key challenge
involves dealing with the increasingly “liquid” nature of that content and its ability to rapidly
mutate and sometimes disappear (Karlsson, 2012a; Sjøvaag and Stavelin, 2012). Indeed, such
developments have already begun to complicate the notion that such content “can be considered
as a finished, static object of study” (Deuze, 2008: 861)—a key consideration for content analyses.
This challenge becomes even more pressing as an increasing amount of news content is produced
and distributed through networked, digital platforms, and developed for increasingly technically
savvy audiences and increasingly powerful devices. This results in larger volumes of progressively
more complex artifacts that leverage new technologies and can quickly change (Zamith, 2017).
That, in turn, demands the development of new methods and frameworks for effectively “freez-
ing” content into static snapshots that may be computationally analyzed in an efficient manner in
order to permit rigorous content analyses and further scholars’ understanding of the evolution of
news content in an increasingly liquid ecosystem (Karlsson and Sjøvaag, 2016).
This chapter focuses on evaluating a range of methods for dealing with this growing chal-
lenge. It begins by synthesizing the literature around the concept of “liquid news” and how
it complicates traditional approaches to content analysis. It then describes key considerations
about the technical aspects of web pages and reviews existing methods for freezing online news
content. Those methods are then employed and critiqued through a case study whereby link
information is extracted from different parts of the fluid and interactive New York Times homep-
age. Finally, recommendations involving common research scenarios are offered, and directions
for future research are presented.
Background
Drawing on the work of Polish social theorist Zygmunt Bauman, Deuze (2008) argues that
newswork must become “liquid” in order to adapt to the media ecology that began to emerge
at the turn of the century. By this, Deuze means that “uncertainty, flux, change, conflict, and
revolution are the permanent conditions of everyday life” (ibid.: 851) for the institution of
journalism – much as it is for modern society (Bauman, 2005). These conditions apply both to
newsworkers and the news content they produce.
93
Rodrigo Zamith
A key insight derived from this work is that online news content has become increasingly
mutable and ephemeral (Karlsson and Strömbäck, 2010; Zamith, 2018b). As Karlsson and Sjøvaag
(2016) argue, the boundaries of liquid journalism, if they exist at all, are distinctly permeable. For
example, the text of an online news story may be updated several times with no apparent indica-
tion that any changes have been made to the original document. Indeed, a document’s Uniform
Resource Locator (URL) often does not change in response to modifications to that document,
and it may be impossible to retrieve its previous iteration (Karlsson and Sjøvaag, 2016). Drafts
therefore begin to bleed into one another. Similarly, whereas just a few variations of a news prod-
uct were previously possible (e.g., a handful of location-specific editions of the morning newspa-
per), liquid journalism faces no such limitation. Homepages may be updated at will by humans,
and personalized and algorithmically generated variants can be produced in milliseconds.
Such affordances are not only theoretically possible but are becoming increasingly important
to news organizations. As scholars have observed, immediacy and interactivity are key character-
istics that help distinguish online news from its analog counterparts (Boczkowski, 2004; Karlsson,
2011). Immediacy refers both to the notion that consumers can view content immediately after it
is produced (Lim, 2012) and to the expectation among modern news consumers that content will
be updated the next time it is refreshed (García Avilés et al., 2004). Interactivity primarily refers
to individuals’ ability to shape, in real time, the form and content of the material and environ-
ment they are engaging with and within (Steuer, 1992), such as opening and closing modules on
a page or engaging in co-production through user comments (Martin, 2015). Scholars have also
observed that multimodality and the hyperlinked nature of online news serve as distinguishing
characteristics (Karlsson, 2012a).
With these considerations in mind, one may adopt Karlsson’s (2012a: 387) conceptualization
of liquid news as “an erratic, continuous, participatory, multimodal and interconnected process
that is producing content according to journalistic principles”. To help illustrate the broad nature
of the concept and begin charting empirical dimensions that may be explored, Karlsson proposes
eight theoretically grounded variables for examining news liquidity. These include the number
of versions (i.e., drafts) a given news item goes through, the extent of user involvement through
participatory features, and the “accuracy problem” that may emerge with fluid events. This chap-
ter focuses on assessing change as a response to modifications by the producer (immediacy) and
by the consumer (interactivity), as these appear to be two of the most common interests among
scholars of digital journalism.
As Deuze (2008: 860) argues, news organizations should “embrace the uncertainty and com-
plexity of the emerging new media ecology, and enjoy it for what it is: an endless resource for
the generation of content and experiences by a growing number of people all around the world.”
Scholars should similarly embrace this shift as it can help them peer into the “black box” of
journalism and better understand it as a process that may be witnessed in real time and in public
(Tumber, 2001; Karlsson, 2011). Indeed, examinations of liquidity have already pointed to the
time-dependent nature of immediacy and versionality on the websites of news organizations
(Widholm, 2016), how distinct pictures of crises are painted through rapidly changing media
framing (Karlsson, 2012b), and how the popularity of news items may affect their future promi-
nence on a homepage (Zamith, 2018b).
Liquid news and content analysis

Embracing the paradigm behind liquid journalism presents a number of challenges to content
analysis methodology. As Karlsson and Sjøvaag (2016) aptly note, the logic of content analysis
is largely embedded in the affordances of print media. Digital media – and online news in
particular – complicate notions of space and time, making it “difficult to pin down where the
94
Innovation in content analysis
news is produced, distributed, and consumed as news appears in so many places – and comes in so
many forms – at once” (Karlsson and Sjøvaag, 2016: 180). They observe that content analyses of
online news – both manual and computational – require a set of considerations that distinguishes
them from traditional content analyses of analog materials. These include the mode of analysis,
variable design, scope, sampling procedure, unit of analysis, storage, generalizability, key agent, and
the aim of the analysis (Karlsson and Sjøvaag, 2016).
The notion of time, in conjunction with the concepts of immediacy and interactivity, is par-
ticularly important in the present context. As Deuze observes:
the study of content has always rested on the premise that content actually exists, that it
genuinely can be considered as a finished, static object of study. In the current media ecol-
ogy of endless remixes, mashups, and continuous edits, that is a problematic assumption.
(Deuze, 2008: 861)
The researcher’s challenge, then, is to turn a dynamic object into a static one (or a sequence of
static ones). This process of “freezing” (Karlsson and Strömbäck, 2010) is necessary in a manual
content analysis in order to ensure that coders are viewing the same content when an assessment
of intercoder reliability is taking place. In a computational analysis, frozen content is necessary
for training, tweaking, and testing algorithms, and to ensure reproducibility (Zamith and Lewis,
2015). Perhaps most importantly, for any assessment of some aspect of a dynamic object’s liquid-
ity to occur, it must be frozen into a series of static snapshots that can be used as the basis of an
evaluation of that object’s fluidity.
As Karlsson and Sjøvaag (2016) observe, traditional content analyses are often performed ex
post facto, with some central repository archiving all of the content that will be studied. For exam-
ple, libraries may keep an archive of all the issues of a given magazine or all recorded episodes
of an evening newscast, and third-party vendors may maintain a database with all the articles
appearing in a particular newspaper. Unfortunately, there is no such record of the web – even
for its most popular sites. While digital archives like the Internet Archive’s ‘WayBack Machine’
may store occasional snapshots of what the content looked like at a certain point in time, those
snapshots are infrequent and unevenly spaced, presenting serious limitations to scientific analyses
(Karpf, 2012). Additionally, these snapshots may be incomplete; not only might portions of the
page be missing, but linked content is sometimes left unarchived.
In the absence of third-party options, researchers must often collect their own data (Karlsson
and Sjøvaag, 2016; Zamith, 2017). The upside is that such collection need not be difficult or
resource-intensive and can be tailored to the research inquiry. The downside is that ex post facto
research is often impossible. Researchers must either anticipate an event and plan ahead of time –
not always a possibility with breaking news – or already be collecting data from the desired
sources when an event takes place.
The structure of online web pages

The study of liquidity may focus on a variety of elements, but it often centers on change – such
as the addition of some elements and the removal of others – on digital documents like web
pages (Karlsson and Strömbäck, 2010; Karlsson, 2012a; Lim, 2012; Sjøvaag et al., 2016; Zamith,
2018b). By their very design, web pages are semi-structured documents, typically consisting
of a mixture of the HyperText Markup Language (HTML), Cascading Style Sheets (CSS), and
JavaScript, a dynamic and interpreted programming language that can add functionality to web
pages through scripts that are executed on the user’s computer. All of a page’s source code is
downloaded when a user requests the page from the server (i.e., by accessing the page’s URL). It
95
Rodrigo Zamith
is then interpreted by the user’s browser, which renders the code and any associated media into
the audio-visual objects the user sees and hears when the page is loaded.
The nature of the web therefore provides researchers with access to most, if not all, of the ele-
ments necessary for storing a copy of a web page locally. Additionally, all of this computer code
provides some structure to the data social scientists typically want to capture. For example, the
beginning of an unordered list is denoted in an HTML document by the tag <ul>, and its end
by </ul>. Each bullet point within that list is then denoted by <li> and </li>, respectively,
and the text of the bullet point is denoted by the text between those two tags. The position of the
bullet point within that list (i.e., if it is the first or third bullet point) is denoted by the sequence
of the code.
Specific elements within the page can sometimes be identified by a researcher through its
id attribute or some unique combination involving the class attribute, either of which may
be included by the page’s author. Researchers typically do this by writing selectors – pieces of
code that identify content based on the presence of specified attributes (see Sjøvaag and Stavelin,
2012). If those attributes are not set, the researcher can also use the element’s place in the struc-
ture of the document, relative to other elements, to identify and extract the desired data.
Furthermore, as scholars have noted, news organizations typically use content management
systems in order to organize information, automatically place it on a template, and publish it
online (Diakopoulos, 2017). This is notable because it means the structure of each news organi-
zation’s homepage will generally only contain minor variations over extended periods of time. As
Zamith (2017) found in his analysis of the homepages of 21 news organizations, each organiza-
tion had at most a handful of distinct layouts. Thus, although the homepage was liquid over the
course of the day – new stories were added and removed, promoted and demoted, and different
“related content” was deemed relevant – the layout and underlying code saw limited variation.
This allows the researcher to leverage the uniformity of a web page’s code and identify common
patterns that permit the extraction of particular data. That, in turn, allows for both the preserva-
tion of the original artifact and the extraction (and subsequent freezing) of interesting aspects of
that artifact.
Methods for freezing liquid news content

A growing body of scholarship has focused on developing approaches for freezing liquid con-
tent. Karlsson and Strömbäck (2010) point to three techniques for accomplishing that. The first
technique is to take a screenshot of the content, typically by using a computer program or simply
utilizing the operating system’s native “print screen” functionality. As they observe, screenshots
are helpful because they provide a visual replica of the content at the time of viewing. Image
files like JPEGs and PNGs should not change if the snapshot is opened with different software
(or versions of those software). However, they observed that screenshots typically only capture
the content appearing in a specific frame. That is, any content that requires the user to scroll up
or down to become visible would not be captured (unless the researcher goes through the messy
process of taking multiple screenshots and stitching them together). While this is true of most
software, Widholm (2016) and Zamith (2017) have pointed to solutions that are able to capture
the entire page in a single image. Nevertheless, as Zamith (2017) observes, valuable information
is lost when interactive objects are transformed into flat images (e.g., a JPEG), such as the URL
a headline links to.
The second technique noted by Karlsson and Strömbäck (2010) is to print a copy of the con-
tent, often to a PDF file. As they observe, this often triggers a “print version” of the page, which
may omit valuable data such as user comments, interactive visualizations, and “related content”.
When a print version is not available, the browser will typically attempt to resize the content to
96
fit different dimensions (e.g., letter-sized paper using a portrait orientation). This requires the
elements on a page to be automatically rearranged, often resulting in copies that look very dif-
ferent from the original. Additionally, as with screenshots, a great deal of valuable information is
lost in the conversion.
The third technique noted by Karlsson and Strömbäck (2010) is mirroring the page. This
involves downloading all the computer code necessary to render the page, including all of the
associated media (e.g., images and stylesheets). One may do this manually by using features in
modern web browsers (i.e., “save page as”) or computationally using software like Wget and
HTTrack (e.g., Sjøvaag and Stavelin, 2012; Lewis et al., 2013). The advantage of using this tech-
nique is that it preserves the greatest amount of information, such as the specific styling options
for a particular element and its structural position within the hierarchy of the page. However,
tools like Wget and HTTrack often fail to capture all of the information necessary to generate
an exact replica of a page at the time it was mirrored. Specifically, they often fail to capture and
relink (to a local copy) certain elements required to accurately display that page, such as Java-
Script files. Moreover, they are unable to process JavaScript actions for loading external content
(e.g., loading user comments that require the user to click on a button titled “view comments”).
Recently, Karlsson and Sjøvaag (2016) called for the development and assessment of com-
putational approaches for handling liquid content in response to the growing and rapid shift
toward digital media production and consumption. Such methods are typically preferred because
they enable large-scale analyses, offer perfect reliability, permit post hoc malleability, and can
improve transparency (Zamith and Lewis, 2015). Recent scholarship has responded to that call
with the development of more advanced methods capable of capturing both the potential for
rapid change as well as the more advanced, interactive features of websites (e.g., Widholm, 2016;
Zamith, 2017). For example, Zamith (2017) has pointed to the value of the Selenium framework
for automating traditional web browsers. This permits the processing of JavaScript and storing
a complete rendering of a web page through full-page screenshots as well as the source code
after it has been processed by the browser. Moreover, it allows the researcher to simulate user
behavior, such as scrolling down the page and clicking on particular content in order to expand it
(Zamith, 2018b). However, Selenium often comes at a considerable computational cost in terms
of the processing power and memory required, which may present a challenge to truly large-scale
analyses (Zamith, 2017). One underexplored alternative, Zamith (2017) suggests, is the use of web
browsing engines that don’t require a graphical user interface to be loaded (headless browsing).
The array of potential solutions and limitations therefore begs the question: Which solution is
best for freezing the flow of liquid news?
Case study
In order to illustrate the effectiveness and limitations of each of those methods, the aforemen-
tioned tools and approaches were used to freeze the homepage of the New York Times on April 19,
2017. This examination was designed to attempt to replicate, on a micro scale, the data collection
performed by Zamith (2018b). In that study, Zamith froze the homepages of 14 news organiza-
tions every 15 minutes over two months. Of particular importance were link data pertaining to
the news items that appeared in areas of the page deemed to be prominent, the news items that
appeared on the list of most-viewed items on the homepage, and the news items that appeared
elsewhere on the page. Those data were then used to assess the impact of an item’s popularity on
its subsequent prominence and likelihood of being removed from the homepage.
This examination was performed using free, open-source software on consumer-grade com-
puter hardware.1 A copy of the source code used as part of this case study may be downloaded
from the author’s website, www.rodrigozamith.com.
97
Rodrigo Zamith
Manually saving content

The first option for freezing content is also the simplest: fire up a browser, surf to the Times’
homepage, and save the content. As noted by Karlsson and Strömbäck (2010), there are two pri-
mary ways to manually save content: saving its source code and creating a visual replica.
Saving a page’s source code simply requires making use of the “Save Page As” feature of the
browser. That feature includes options to save an “HTML only” copy of the page as well as a
“complete” one. The first created a single file comprised of the page’s HTML source code, which
included information about the structure of the page and much of its text content. However, a
great deal of the content on the page was simply referenced by the HTML source code (and thus
not embedded in it). For example, the pictures associated with the top headlines were hosted on
the Times’ server – which may change or be removed. If the researcher is merely interested in the
headlines and blurbs, this feature may suffice. However, if a complete recreation of the page is
necessary, this approach may prove problematic.
The second option, storing a “complete” copy of a page, proves more versatile. It not only
creates a copy of the source code but also of all the assets loaded by the browser (e.g., image files
and stylesheets).2 Notably, the HTML source code is automatically modified to reference the
locally stored copies. Additionally, the “complete” copy of the page also stores any modifications
to the page as a result of the execution of JavaScript at the time of the snapshot. For example, the
“complete” copy was able to store the code necessary to recreate the JavaScript-powered list of
most-viewed items, if the user scrolled down to it. The “HTML only” copy did not. Neverthe-
less, some aspects of the page may not be fully replicated. Areas containing Flash video files, for
example, were left as blank spots on the “complete” copy. As one might expect, the “complete”
copy was considerably larger (16.4 megabytes) than the “HTML only” copy (0.21 megabytes) in
this case study, which could present issues in a large-scale analysis.
Creating a visual replica of the page is usually done through one of two options: saving a PDF
copy of the page or taking a screenshot (Karlsson and Strömbäck, 2010). While Mozilla’s Firefox
did not have built-in functionality for creating PDFs, most modern operating systems offer a
virtual printer by default. This operation typically requires the page to be reformatted to fit the
standard paper size (i.e., letter) and some websites employ custom stylesheets that remove certain
elements and reflow the text. While this may prove to be sufficient if the research inquiry focuses
solely on the text of the page – provided the reformatted version does not omit it – the resulting
PDF would not be an actual replica of what is observed on the screen. In the case of the New
York Times’ homepage, no such stylesheet was offered. Instead, the browser attempted to reformat
content itself, resulting in a 10-page document with multiple overlapping objects that looked
very different from the rendered web page (and was difficult to read).
The second option is to take a screenshot of the page. Prior work has referenced the use of
software, such as an operating system’s native “print screen” tool, to capture such images, and
lamented the fact that such software is typically only able to capture the portion of the page that
is visible on a screen at any given moment. However, Mozilla’s Firefox includes a “Developer
Toolbar” feature that allows the program to generate a single image that encompasses the entire
page. This is done by entering the command, screenshot –fullpage [filename],
which in turn generates a PNG image file. This effectively bypasses the limitations previously
observed in the literature, though the resulting images are quite sizable. A single image from this
case study was a hefty 6.3 megabytes.
While saving “complete” copies of the Times’ homepage would be theoretically sufficient for
freezing the data necessary to perform Zamith’s (2018b) analysis, it would be impractical. It would
require an individual to simultaneously load 14 pages every 15 minutes and perform a series
of actions on each page. Over the course of two months, this would add up to nearly 82,000
98
snapshots that would need to be stored by a human being around the clock – introducing great
potential for human error.
Content retrieval software

In order to automate the freezing process, many researchers have turned to content retrieval
software, such as Wget, HTTrack, and the Requests library for Python. The first of these tools,
Wget, will simply download and store a copy of the contents of the URL provided. For example,
it is possible to store the equivalent of an “HTML only” copy of the New York Times’ homepage
through the simple command wget -O ’nyt.html’ ’www.nytimes.com’. Research-
ers may run Wget on an automated timer (e.g., every 15 minutes) by using system-provided
tools, such as Linux’s Crontab. Because only the original HTML source code is downloaded, the
aforementioned limitations apply.3
The second tool, HTTrack, is designed to create a mirror (an exact copy) of a website.
HTTrack offers functionality that is similar to storing a “complete” copy of a page through
Firefox, as well as more advanced features. Like a “complete” copy, all of the elements linked to
from a starting point (i.e., the Times’ homepage) are automatically downloaded, and references
to that external content are rewritten to refer to the new, local copies. HTTrack also offers the
ability to crawl websites through recursive downloading. For example, the command, httrack
’www.nytimes.com’ -O ’nyt’ -mirrorlinks -r2 instructs HTTrack to not only
download the Times’ homepage but also anything linked to from that page, such as the news
stories that appear on it.4 Expanding the scope of the mirror quickly increases the size of the
contents, though HTTrack will reuse assets when possible to conserve space.
The third tool is a library for the popular Python programming language called Requests.
Python is a general-purpose programming language used by several computational social sci-
entists (e.g., Sjøvaag and Stavelin, 2012; Zamith, 2017) due to its blend of power and ease of
use. It also includes a number of libraries that make it easier for the researcher to perform
common functions, from accessing the Twitter API (Tweepy) to creating graphical plots (Mat-
plotlib). Requests allows the researcher to embed Wget-like functionality into a Python script.
For example, the source code from the Times’ homepage can be obtained with the following
code: page_source = requests.get(‘www.nytimes.com’).text. This is use-
ful when one wants to collect a particular subset of data from a broader document. For example,
the BeautifulSoup library for Python – which transforms HTML code into an object that can
be computationally navigated – can be paired with Requests to identify and download all of the
links appearing in a certain part of the website. One could then create a variable with the navi-
gable source code with: page_source_soup = BeautifulSoup(page_source,
”lxml”). Then one could extract all of the links appearing near the top portion of the page
(identified in the page’s source code by the id attribute top-news) with: [x[‘href’] for
x in page_source_soup.find(‘section’, {‘id’: ’top-news’}).find_
all(‘a’, href=True)]. Finally, the Requests library can be leveraged to download each
of those links, using a variation of the aforementioned code.
Although these tools range in difficulty, they all enable automation of some sort. This permits
the researcher to engage in unsupervised data collection, perhaps checking in periodically to
confirm the operations are running as expected. However, none of these tools are able to fully
process JavaScript code, which restricts their utility within the modern web. In the present case
study, none were able to load the list of most-viewed items because of that limitation – therefore
precluding access to a key set of data. Moreover, these tools are unable to generate visual replicas
of the content (i.e., screenshots).
99
Rodrigo Zamith
Automated browsing software

An emerging set of software that permits the freezing of content from more complex websites
is automated browsing software, such as CasperJS and the Selenium WebDriver. These software
allow computer code to simulate human actions using either a full web browser engine, like
Mozilla’s Gecko or KDE’s WebKit, or a full browser, like Mozilla’s Firefox or Google’s Chrome.
This, in turn, allows for the complete simulation of a regular browsing session, which includes
the processing of client-side actions in JavaScript code.
One particularly useful tool for automated browsing is CasperJS, which permits headless
browsing or scripted web browsing without a graphical user interface. It builds on the Phan-
tomJS software, which in turn employs the WebKit browser engine used by Apple’s Safari and
Google’s Chrome and thus tends to render pages as intended by the pages’ authors. Because it
is a headless solution, it requires relatively few resources to process a web page – though not as
little as content retrieval software. CasperJS natively provides a JavaScript API to allow computer
scripts to simulate user behavior and automate browsing. For example, the source code from the
New York Times’ homepage can be obtained with the following code: casper.start(‘www.
nytimes.com’). A screenshot of the entire page can be taken by appending the following
code: this.capture(‘nyt.png’).
Unlike the previous tools, however, CasperJS is able to process JavaScript code and interact
with elements on a page. This enables it to load the list of most-viewed items on the Times’
homepage, which requires the user to scroll down to the list of most-viewed items before it will
load any content, and then click on the desired list (i.e., most-viewed versus most-emailed). To do
this, one may use this.scrollToBottom() to scroll to the bottom of the page – thereby
ensuring all position-specific elements load – and this.click(‘.most-viewed-tab’)
to simulate a user clicking on the list of most-viewed items, which is distinguished in the Times’
homepage through the class attribute most-viewed-tab. Finally, CasperJS is also able
to simulate waiting through the .wait() and .waitForSelector() functions, which is
sometimes necessary in order to avoid interstitial ads that have become common on many news
sites. Notably, CasperJS is able to save a page’s HTML source code after it has been processed
by the browser. This is a key distinction because it permits subsequent analysis of areas of a page
that require JavaScript to be displayed with tools like Python’s BeautifulSoup library. While it
is possible to extract the links appearing in the list of most-viewed items using CasperJS, such a
process is more cumbersome than when using a solution like BeautifulSoup.
Finally, the most comprehensive tool is the Selenium WebDriver (hereafter called Selenium),
which permits the automation of complete web browsers like Firefox and Chrome.5 Selenium is
fully implemented in and supported by a number of programming languages, including Python,
Ruby, and Java, making it accessible to programmers with different expertise. Using Python, for
example, Firefox may be opened using driver = webdriver.Firefox(). Then, driver.
get(“www.nytimes.com”) may be used to open the Times’ homepage, driver.page_
source to access its source code, and driver.save_screenshot(“nyt.png”) to save
a screenshot. To select the list of most-viewed items, one can use driver.find_element_
by_class_name(“most-viewed-tab”).click().
Selenium permits the researcher to not only replicate anything that may be accomplished
with CasperJS but also to use plugins (e.g., Adobe Flash), add-ons (e.g., uBlock Origin), and
offline storage (e.g., for websites like the Times’ Today’s Paper web app). Additionally, in the
author’s experience, Selenium will typically do a better job than CasperJS at rendering very
complex websites that make extensive use of JavaScript and bleeding-edge HTML and CSS fea-
tures. Moreover, Selenium offers the advantage of easy, simultaneous integration of the browsing
100
session with libraries like BeautifulSoup. This allows the researcher to perform common actions
like extracting links from areas of interest on a page and comparing pages to previous snapshots
within a single script, thereby removing the need for storing multiple versions of the same docu-
ment or adding steps to a workflow.
These benefits, of course, come at the expense of considerably higher computational costs.
Using a full browser like Chrome consumes far more memory and CPU cycles than the headless
use of the WebKit engine. Additionally, browsers like Firefox and Chrome require a graphical
user environment, and thus display hardware and windowing software. There are framebuffer
emulators like Xvfb for Linux, which permit Firefox to run in a terminal session (i.e., command
line). While such emulators are considerably less computationally expensive than a full window-
ing system and desktop environment, they nevertheless require additional resources, which can
be problematic when working with a large number of websites.
For the purposes of replicating Zamith’s (2018b) study, either of these two solutions would be
suitable. However, given that CasperJS is considerably less resource-intensive, it should be favored
for that kind of work. Alternatively, Selenium may be paired with PhantomJS – as opposed to a
full browser – to reduce the resource load while simultaneously incorporating popular Python
libraries like BeautifulSoup.
Weighing options and future directions

As this chapter has indicated, there are several potential solutions for freezing the flow of liquid
news. The selection of a solution is often dependent on the researcher’s technical ability, the
nature of the research inquiry, and the complexity of the digital artifacts being studied. For simple
analyses, manually saving pages or taking screenshots may do the trick.
However, demand for computational solutions is likely to increase since the study of the
liquidity of news will increasingly require researchers to look at shorter periods of time as expec-
tations of immediacy grow (García Avilés et al., 2004; Lim, 2012). This is an important con-
sideration because research designs that involve short intervals tend to require computationally
efficient solutions and generate large datasets. As Karlsson (2012a) notes, the dearth of empirical
analyses of the liquidity of content is partly due to the difficulty of the work. Indeed, Zamith
(2018a) reported having to limit the sample size of that study due to the computational and
memory requirements of automating Firefox through Selenium, and Zamith’s (2017) analysis
required nearly 650 gigabytes of storage space.
Computational solutions can be accessible to those with limited technical backgrounds,
though. Content retrieval software like Wget and HTTrack are very accessible and are capable
of freezing enough aspects of liquid content to be usable in a range of applications (e.g., Sjøvaag
and Stavelin, 2012). Put differently, such software can deal with the immediacy of liquid news,
though it may fail to deal with its interactive elements. Moreover, tools like Wget are computa-
tionally inexpensive and thus advantageous for large-scale analyses. It is thus advisable to try such
software first before progressing to more powerful solutions.
However, for certain projects and more complex pages such as the ones in this case study
more advanced solutions are required. As Martin (2015) and others have observed, news orga-
nizations are increasingly turning to integrated third-party platforms in order to incorporate
important interactive features onto websites, such as reader comments and indicators of trending
content. Moreover, increasingly popular features like interactive data visualizations require the
use of JavaScript to simply appear on a page. Automated browsing tools are necessary when
such data—or more faithful replicas—are of import to the researcher. In the present case study,
101
Rodrigo Zamith
headless browsing software like CasperJS offered superior performance to a Firefox- or Chrome-
powered Selenium solution while having similar functionality for dealing with the interactivity
of modern websites. Researchers are thus encouraged to initially consider that option for navigat-
ing complex websites, though a full browser may be occasionally required.
An additional consideration researchers may face is whether they should store their snapshots
as screenshots or source code – or both. Full-page screenshots offer the best chance that the
content will look exactly the same over time and may be ideal for qualitative analyses of text
and graphical content, as well as visual analyses of the assemblage of content. However, a great
deal of important information is lost in that conversion (e.g., the ability to extract the text or
link information), which hampers subsequent computational analyses. Moreover, screenshots are
often several times larger than the page’s source code – though they are typically smaller than
storing a “complete” copy of a page that includes all the referenced media. In many instances,
researchers would be well-served by storing both screenshots and the page’s source code. Zamith
(2017) reported that screenshots served as a helpful aid for manually ensuring that content was
being stored correctly and that the algorithms employed for computationally analyzing content
were making valid decisions.
Methods for freezing content will need to continue to evolve in order to keep up with
emerging trends in media production, distribution, and consumption. For example, recent work
has found that individuals are doing more of their news consumption on mobile devices and
that news organizations are devoting more resources to mobile news apps and chat apps (Belair-
Gagnon et al., 2017). Future efforts should seek to explore efficient ways to freeze the content
appearing on those apps, whether through official application programming interfaces (APIs)
or by emulating the experience. Additionally, as some services’ APIs become more restrictive
(e.g., Twitter and Facebook), researchers must identify solutions that can yield quality data using
less-restrictive modes of access, such as by scraping the websites of those services. Finally, as
algorithms become more prominent in social life – leading to questions about the development
of “filter bubbles” around news consumption and political discourse – researchers must develop
methods for simulating distinct user experiences in order to gather data on those algorithms’
impact on content. While the automated browsing approaches described here may serve as a
starting point, additional work is required for identifying the most efficient frameworks for
capturing distinct experiences.
Conclusion
News is becoming increasingly “liquid” as more of it is produced, distributed, and consumed
through digital platforms. Researchers have responded by developing new approaches for ana-
lyzing news content and identifying multiple tools to “freeze” it. However, such tools vary in
efficiency and sophistication and thus range in their ability to capture the sorts of liquid objects
that are of greatest interest to scholars.
As has been shown in this chapter, there is no single “best” tool for freezing liquid con-
tent. Old tools like Wget remain useful for less complex pages and require few computational
resources. New tools like CasperJS are better able to capture interactive elements and more-
intricate features of a page, though they come with higher computational costs. The appropri-
ateness and effectiveness of a tool thus depends on the nature of the research inquiry and the
objects that will be studied. However, one thing is clear: As the web evolves and news products
become more complex, existing approaches to content analysis will need to be refined and new
tools developed.
102
Further reading
A number of works helped shape this chapter and are recommended for further reading. Deuze’s
(2008) “The Changing Context of News Work” is a great introduction to the concept of “liquid-
ity” and how it extends to newswork. Karlsson’s (2012a) “Charting the Liquidity of Online
News” and Karlsson and Strömbäck’s (2010) “Freezing the Flow of Online News” do a wonder-
ful job of applying that concept to news content and highlighting the methodological challenges
it introduces to content analysis. Karlsson and Sjøvaag’s (2016) “Content Analysis and Online
News” aptly explicates an approach they call “liquid content analysis” and how it departs from
traditional content analysis along the many steps of the process. Finally, Zamith’s (2017) “Cap-
turing and Analyzing Liquid Content” offers a detailed and practical description of a complex
computational analysis of liquid content that includes useful snippets of code and explains them.
Notes
1 The following software were used: BeautifulSoup (4.5.1), CasperJS (1.1.4), HTTrack (3.49), Mozilla
Firefox (50.0.2), Selenium (2.53.4), and Wget (1.15).
2 Mozilla also offers a downloadable add-on to Firefox called “Mozilla Archive Format” that aims to create
an even more faithful, system-independent, and compressed copy of a page.
3 Wget includes a number of useful options to augment its basic functionality. For example, the command
wget -E -H -k -K -p -P ‘nyt’ ‘www.nytimes.com’ would achieve a result similar to sav-
ing a “complete” copy with Firefox. By appending the arguments -r -l 2, it is possible to replicate the
recursive functionality described for HTTrack. However, HTTrack offers more functionality than Wget
for these kinds of operations.
4 The last argument in that command ensures it goes no further than the homepage. Additional arguments
may be used to restrict the crawling to a small set of relevant domains.
5 The Selenium WebDriver is part of a broader framework for testing web applications, which includes a
complete integrated development environment and its own programming language. It may also be used
to automate CasperJS and PhantomJS for those who do not wish to write their own JavaScript. It also
supports SlimerJS, which provides near-headless access to Mozilla’s Gecko engine.
References
Bauman, Z. (2005) Liquid Life. Cambridge: Polity.
Belair-Gagnon, V., Agur, C. and Frisch, N. (2017) “The changing physical and social environment of
newsgathering: A case study of foreign correspondents using chat apps during unrest.” Social Media +
Society, 3(1), 1–10.
Boczkowski, P. J. (2004) “The processes of adopting multimedia and interactivity in three online news-
rooms.” Journal of Communication, 54(2), 197–213.
Deuze, M. (2008) “The changing context of news work: Liquid journalism for a monitorial citizenry.”
International Journal of Communication Systems, 2, 848–865.
Diakopoulos, N. (2017) “Computational journalism and the emergence of news platforms.” In B. Franklin
and S. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. New York, NY: Routledge
(pp. 176–184).
García Avilés, J. A., León, B., Sanders, K. and Harrison, J. (2004) “Journalists at digital television newsrooms
in Britain and Spain: Workflow and multi-skilling in a competitive environment.” Journalism Studies, 5(1),
87–100.
Karlsson, M. (2011) “The immediacy of online news, the visibility of journalistic processes and a restructur-
ing of journalistic authority.” Journalism, 12(3), 279–295.
Karlsson, M. (2012a) “Charting the liquidity of online news: Moving towards a method for content analysis
of online news.” International Communication Gazette, 74(4), 385–402.
Karlsson, M. (2012b) “The online news cycle and the continuous alteration of crisis frames: A Swedish case
study on how the immediacy of online news affected the framing of the swine flu epidemic.” Journal of
Organisational Transformation & Social Change, 9(3), 247–259.
103
Rodrigo Zamith
Karlsson, M. and Sjøvaag, H. (2016) “Content analysis and online news.” Digital Journalism, 4(1), 177–192.
Karlsson, M. and Strömbäck, J. (2010) “Freezing the flow of online news: Exploring approaches to the study
of the liquidity of online news.” Journalism Studies, 11(1), 2–19.
Karpf, D. (2012) “Social science research methods in internet time.” Information, Communication and Society,
15(5), 639–661.
Lewis, S. C., Zamith, R. and Hermida, A. (2013) “Content analysis in an era of big data: A hybrid approach
to computational and manual methods.” Journal of Broadcasting & Electronic Media, 57(1), 34–52.
Lim, J. (2012) “The mythological status of the immediacy of the most important online news.” Journalism
Studies, 13(1), 71–89.
Martin, F. (2015) “Getting my two cents worth in: Access, interaction, participation and social inclusion in
online news commenting.” #ISOJ, 5(1), 80–105.
Sjøvaag, H. and Stavelin, E. (2012) “Web media and the quantitative content analysis: Methodological
challenges in measuring online news content.” Convergence: The International Journal of Research into New
Media Technologies, 18(2), 215–229.
Sjøvaag, H., Stavelin, E. and Moe, H. (2016) “Continuity and change in public service news online.” Journal-
ism Studies, 17(8), 952–970.
Steuer, J. (1992) “Defining virtual reality: Dimensions determining telepresence.” Journal of Communication,
42(4), 73–93.
Tumber, H. (2001) “Democracy in the information age: The role of the Fourth Estate in cyberspace.” Infor-
mation, Communication and Society, 4(1), 95–112.
Widholm, A. (2016) “Tracing online news in motion.” Digital Journalism, 4(1), 24–40.
Zamith, R. (2017) “Capturing and analyzing liquid content.” Journalism Studies, 18(12), 1489–1504.
Zamith, R. (2018a) “A computational approach for examining the comparability of ‘most-viewed lists’ on
online news sites.” Journalism & Mass Communication Quarterly, 95(1), 122–141.
Zamith, R. (2018b) “On metrics-driven homepages: Assessing the relationship between popularity and
prominence.” Journalism Studies, 19(8), 1116–1137.
Zamith, R. and Lewis, S. C. (2015) “Content analysis and the algorithmic coder: What computational social
science means for traditional modes of media analysis.” The ANNALS of the American Academy of Political
and Social Science, 659(1), 307–318.
104
8
AN APPROACH TO ASSESSING
THE ROBUSTNESS OF LOCAL
NEWS PROVISION
Philip M. Napoli, Matthew Weber, and Kathleen McCollough
As local journalism evolves in response to challenges posed by the technological changes that
have taken place in the media sector (Anderson et al., 2012; Downie and Schudson, 2009;
Picard, 2014), one growing concern is that significant differences exist across communities
in terms of the extent to which journalism sources are serving people’s information needs.
Research raises concerns that, in some communities, local journalism is essentially collapsing,
with the decline and (in many cases) disappearance of traditional news outlets leaving massive
unfilled gaps (what Stites (2011) has termed “news deserts”; see also Ferrier’s (2013) analysis
of “media deserts”). These gaps may create greater opportunities for political and corporate
corruption to flourish in ways that can undermine effective democratic participation (Starr,
2009).
The extent to which local news is collapsing may vary according to the particular charac-
teristics (demographic, economic, political, technological) of individual communities (e.g., Pew
Research Center, 2015). A 2015 report noted, for instance, that large U.S. cities such as New
York, Washington, DC, and Los Angeles are employing an increasing proportion of the country’s
professional journalists, with smaller cities experiencing dramatic declines (Tankersley, 2015).
These patterns suggest an emerging condition of “journalism haves and have-nots” across the
country. As the Knight Commission observed in 2009:
the nation’s vast information needs are met unequally, community by community.
Some populations have access to local news and other relevant information through
daily newspapers, radio, and television broadcasts, local cable news channels, hyper-local
websites, blogs, mobile alerts. . . . Others are woefully underserved.
(Knight Commission, 2009: 3)
However, anecdotal evidence like that presented in the Knight Commission report has yet to be
accompanied by systematic empirical data on the state of local journalism that allows for analyses
across multiple communities.
Other areas of public interest, such as economic development, the environment, political par-
ticipation, and community engagement, have reasonably well-developed methods and measures
for assessing the health of local communities (e.g., Community Health Status Indicators Project
105
Philip M. Napoli et al.
Working Group, 2009; Sustainable Jersey, 2013). The same level of method and measure develop-
ment has not been the case, however, for local journalism.
This chapter describes ongoing efforts to address this analytical challenge. In it, we present
a multi-level methodological framework for assessing local journalism and the extent to which
it addresses communities’ critical information needs. The development of such a method and
the accompanying measures provide a valuable analytical tool for news organizations, funders of
journalism initiatives, advocacy groups, citizens, and policymakers.
The first part of this chapter contextualizes this methodological approach within a sequence
of calls for methods and measures that could inform public and private efforts to strengthen
local journalism. The second section presents the methodological approach and describes the
rationales underlying its design. The third section describes the types of metrics that can be
constructed and analyses that can be conducted with the data. The concluding section considers
further steps in the development and application of this methodology.
Local journalism and the need for assessment tools

Concerns about the state of local journalism have been widespread, as technological changes
dramatically affect the economics of the commercial news business and the dynamics of how
consumers access, share, and even produce journalism (see Barnett and Townend, 2015; Nielsen,
2015). These concerns have been particularly pronounced in the United States, where there is
often a greater reliance on the commercial model of news production than in virtually any other
developed nation (Schudson, 2005; see also Pickard, this volume, Chapter 16).
Reflecting the intensity of these concerns in the US, the Knight Commission released a 2009
report identifying access to credible and relevant information as a key requisite for healthy com-
munities (Knight Commission, 2009). As the Knight Commission (2009: 39) emphasized, it was
particularly important to assess the quantity and quality of information available to communities:
“If activists, policy makers, and the general public had more concrete ways of describing, mea-
suring, and comparing the systems of community news and information flow, it would be much
easier to mobilize public interest around community information needs”.
The Knight Commission (2009) report led to policy-related inquiries across various branches
of the U.S. government (see Napoli and Stonbely, in press; Leibowitz, 2009; U.S. Senate, 2009).
Most relevant to this analysis is the attention that the Federal Communications Commission
(2009) devoted to the issue. Picking up on the Knight Commission’s work, in 2009 the FCC
initiated a comprehensive assessment of how community information needs are being met in the
broadband era (Waldman, 2011). Among the FCC inquiry’s conclusions was a call (echoing the
Knight Commission) for a thorough accounting of the journalism provided at the community
level.
In 2013, the FCC followed its report by commissioning a study that would provide a “Multi-
Market Study of Critical Information Needs” (Social Solutions International, 2013). The pro-
posed study included content analysis of journalism from a range of media, including television,
radio, newspaper, and internet, analyzed in terms of the extent to which it covered “critical infor-
mation needs”. However, the proposed research produced a firestorm of controversy, on the basis
of concerns that the research represented government intrusion into newsrooms (see Pai, 2014).
Congressional hearings ensued, as well as threatened legislation to kill the research (Eggerton,
2014). As a result of these pressures, the FCC first scaled back and then ultimately canceled the
entire study (Flint, 2014; for a more detailed discussion, see Napoli and Friedland, 2016).
More recently, the Pew Research Center (2015) has produced a thorough analysis of the local
news ecosystems in three U.S. communities of different sizes. Extending upon earlier ecosystem
research on the flow of news in Baltimore (Pew Research Center, 2010), this study included
106
Assessing the robustness of local news
inventories of local media outlets, surveys of news consumers, and analysis of social media data.
While incredibly useful, the scope and depth of this analytical approach comes with costs that
prohibit scaling up to analyze a much larger sample of communities, which would enable more
extensive comparisons and, potentially, more generalizable findings.
Indeed, case studies examining the state of local journalism in a single community, or in a very
limited number of communities, are more easily found (Durkin and Glaisyer, 2011; Durkin and
Hadge, 2010; Gloria and Hadge, 2010; Morgan, 2011, 2013; Pew Research, 2010; Ramos et al.,
2013; Ryfe et al., 2012). The lack of studies on a larger scale are most likely a reflection of the
challenges associated with developing a methodology that could be applied to a larger sample
of communities at a manageable cost. In research in which larger samples of communities are
analyzed, the analytical framework is typically limited in terms of focusing on a single platform
and/or focusing on a particular issue or type of news (see Becker and Yanich, 2015; Fico et al.,
2013; Holt and Karlsson, 2015; Karlsson and Holt, 2014; St. John et al., 2014; Williams et al.,
2015). The narrow focus often occurs because the primary unit of analysis in such research is
typically the individual media outlet or platform type or the individual news story rather than
the community as a whole.
As this review indicates, there remains a gap in terms of a robust but scalable analytical
approach to broadly assessing local journalism across communities, platforms, and issues that
could be utilized by foundations, policymakers, researchers, and industry professionals. The
methodological approach presented here represents an effort to fill this gap.
Assessing local journalism infrastructure, output, and performance

In order to outline a new analytical approach, the assessment of local journalism has been broken
down into three connected conceptual dimensions: (1) the journalistic infrastructure; (2) journal-
istic output; and (3) journalistic performance, each with an associated methodological component.
This analytical approach borrows from, and modifies, the well-known structure – conduct –
performance theoretical framework from industrial organizational economics (see Caves, 1992).
This framework presumes causal relationships between the structure of markets (as represented
by the number, size, and characteristics of market participants), which in turn affects the conduct
of firms in these markets, which in turn affects firms’ market performance.
For this analysis, the notion of infrastructure refers to the number, size, and characteristics of
media outlets within a community. Output refers to the volume of journalistic output produced
by these media outlets. Performance refers to some fundamental qualitative dimensions of this
journalistic output (see later). The goal here, though, is less about exploring relationships between
the three conceptual dimensions of local journalism ecosystems than it is about offering a rea-
sonably comprehensive empirical framework for assessing the state of local journalism in ways
that can facilitate comparative analyses across communities and/or over time, and for (ultimately)
facilitating research that identifies those community characteristics that are useful for explaining
the state of local journalism in individual communities.
Assessing journalistic infrastructure

A key dimension of any local journalism ecosystem is the basic footprint of local journal-
ism, in terms of the number of outlets capable of producing/disseminating local news and
information within a given community. A fundamental premise of democratic theory-based
approaches to journalism is the notion that the democratic process and an informed citizenry
are enhanced via the presence of a diversity or plurality of sources of news and information
(Baker, 2002). This perspective can encompass very basic indicators, such as the number of
107
sources serving a community, or, it can delve deeper into the nature of these sources, includ-
ing criteria such as market shares or ownership characteristics (see Napoli, 2001). Reflecting
this theoretical perspective, this methodological approach begins by examining the journalistic
infrastructure in a community.
A starting point for assessing a local journalism infrastructure involves identifying each source
of journalism within a particular community (see Lin and Song, 2006). Such an activity has
become more complicated, as keeping pace with the profile of any local journalism ecosystem is
much more challenging in this time in which various journalistic initiatives are rapidly started
and stopped.
Given the inadequacy of available data sources (such as the directories of local media outlets
that existed in the pre-internet age; see Barnett and Townend, 2015), any effort to create an inven-
tory of the sources of local journalism serving a community is, to some extent, an ad hoc endeavor.
When one looks at previous research in this vein, relatively little concrete methodological detail
is provided as to exactly how the inventory of local journalism sources was constructed, which
is a reflection of the somewhat improvisational, somewhat impressionistic nature of the process.
However, for this methodological approach, we established – and outline in detail here – a
concrete, multi-stage data-gathering protocol in order to provide as much clarity and transpar-
ency about the process as possible. The process draws from – and to some extent combines –
approaches employed in previous research. It involves consultation with the most authoritative
relevant directories available and supplements these consultations with a systematic search and
discovery process that involves both online searching and engagement with members of the com-
munities being studied.
Even something as simple as defining and identifying a source of local journalism is a more
complex and challenging process than it once was, and any efforts to define and identify sources
can be critiqued as being too narrow or too expansive. The definitional approach we employed
leaned toward being expansive. First, we did not employ any criteria based on the presence of
minimum levels of journalistic content; at this very local level, we wanted to include both dedi-
cated and what we might term tangential or potential sources of journalism. Thus, for instance, local
radio stations would be included regardless of their format as part of the journalistic infrastruc-
ture of a community; any journalistic distinction between a news and music station occurs in the
analysis of journalistic output and performance dimensions (see later). And so, all community-
based media outlets (defined as television stations, radio stations, print/online newspapers, and
community and hyperlocal news websites) were included in the infrastructure level of analysis.
Our search protocol then sought to identify other potential online journalism sources (such as
blogs) left out of more the formal organizational criteria of a media outlet. However, in an effort
to impose some boundaries on the abundance of content available online, such an online source
needs to exhibit evidence on its homepage of addressing one or more of the critical information
needs described later. Community news and information-focused blogs have failed to materialize
in a very significant way in applications of our search protocol thus far, perhaps due to limitations
in our protocol (and the associated databases) or to the dearth of such community-centric blogs
within the communities of our analysis.
The focus of this methodological approach is local journalism. We define local in terms of
the geographic boundaries of the communities being studied. Thus, this approach focuses on the
journalistic sources that reside within and serve the selected communities (Lin and Song, 2006).
Operationalizing these parameters means excluding larger regional, county, or state-level outlets
that may cover the local community (e.g. if analyzing Newark, NJ, the New York Times would
be excluded) but that are not geographically based in the community and primarily focused on
the community.
108
The search process for identifying relevant journalistic sources is conducted in three stages:
Stage 1: Consult relevant media directories

A number of print and online data sources are available to identify media outlets at the local level.
Because research has shown that many such directories (including those offered by commercial
providers or government agencies) tend to be incomplete, multiple directories were consulted.
These directories are listed in Table 8.1. This process involves locating those media outlets asso-
ciated with the set of communities being analyzed. It is important to note that our design and
application of this methodological approach have thus far focused on individual municipalities as
the unit of analysis but can be modified to accommodate other units of analysis such as individual
designated market areas (DMAs) or individual states.
Stage 2: Supplement directory data with manual search

In order to supplement the data gathered from the directories, the second stage of data-gathering
involves a manual search for relevant journalistic sources. Following the approach employed by
Ramos et al. (2013), this process involves keyword searches via search engine and then visiting
those sites produced by the search queries to identify links to other relevant sources. Further-
more, those sites that were linked to by the original site are subsequently examined to determine
whether they contain links to any additional relevant sites. Keyword searches employ the name of
the town, county, and region, along with associated media terms such as ‘news’, ‘blog’, ‘radio’, and
‘television’. In addition, when a community is known by a particular nickname (e.g., ‘Brick City’
for Newark), that terminology is also employed in the search process. If the analysis is expanded
to include a wider geographic domain (e.g., DMA), search parameters for identifying sources
simply need to be expanded according to the definition of the local level.
Stage 3: Targeted interviews with community members

In order to identify potential sources not identified by Stages 1 and 2, a final step involves con-
ducting targeted interviews with community members in positions to be well-informed about
the journalistic sources serving the local community (see Morgan, 2011). Specifically, for each
Table 8.1 Media directories/databases consulted
Medium Directory
Television Community Media Database

Association of Public Television Stations’ Station Directory
FCC Broadcast Television License Database
Radio NPR Labs Mapping and Population System
FCC AM and FM Broadcast License Database
Print Library of Congress Directory of Newspapers
Editor & Publisher International Data Book
Online Knight Foundation’s Directory of Community News Sites
Columbia Journalism Review’s Guide to Online News Startups
Online Newspaper Directory for the World
Multiplatform National Directory of Ethnic Media
Cision Media Database
109
community being analyzed, 3–5 interviews are conducted with individuals in the following
categories: (1) local government (2) local news media; (3) activist organizations; and (4) ethnic
community organizations.
Assessing journalistic output

The logical question that arises from the infrastructure assessment described earlier is how much
journalistic output does the infrastructure generate? Thus, the output component assesses the
aggregate journalistic output within a selected community within a specified period of time.
The concern is one of quantity (the qualitative dimension is addressed in the performance dimen-
sion), because a reasonable indicator of the health of a local journalism ecosystem is the amount
of journalism produced for and within the community. Ultimately, though, this assessment of
journalistic output serves as a vital intermediate step toward conducting the more substantive
assessment of journalistic performance (see later).
This stage in the process involves gathering and analyzing a sample of news content from all of
the sources identified in each of the communities being analyzed. For our pilot test of this meth-
odology (see Napoli et al., 2017), we analyzed three communities in New Jersey (US) through
a sample week of content. For this small-scale effort, it was relatively easy to analyze all of the
relevant content on a daily basis for each day of analysis. However, to conduct this research on
a larger scale, content archiving is necessary. So, for instance, we are currently content analyzing
a constructed week’s worth of local news output for 100 randomly sampled U.S. communities.
Sampled communities had a maximum population size of 300,000 and a minimum population
size of 20,000.
For this analysis, we partnered with the Internet Archive’s Archive-IT platform to conduct a
custom web crawl and archive for future analysis. Even an ambitious Internet archiving project,
such as the Internet Archive, does not comprehensively archive local news sources associated with
all communities (particularly smaller communities). Rather, much of the crawling is random in
nature. Thus, for this custom archive, we provided the Internet Archive with the URLs for the
nearly 700 media outlets associated with the 100 sampled communities. We also provided them
with seven randomly sampled days for our constructed week – constructed from the four upcom-
ing months. The Internet Archive then proceeded to crawl and archive all of the homepages – as
well as any pages linked by one step (one click) from those homepages – for each media outlet.
This process has produced roughly 2.7 terabytes of data and included approximately 12 million
archived documents (web pages, images, movies, etc.).
Each media outlet’s homepage serves as its indicator of local news provision. Using website
homepages as representative of an outlet’s news output builds on the tradition of sampling a
newspaper’s front page, which is at once the most likely to be seen by readers and also represents
the news outlet’s judgment as to the most important news it has to offer (e.g. Benson, 2013).
Homepage stories are counted/coded for each sample day only if they were posted on the sample
day. Thus, for instance, if the homepage on the analyzed day of July 27, 2016, contains stories
dated earlier than July 27, those stories are not included in the analysis. The goal here is to aggre-
gate and analyze the news output produced on each selected day.
The methodological approach employed for the output dimension – and the performance
dimension that follows – relies on the journalistic content available online, regardless of the out-
let’s “native” platform. Thus, the journalistic outputs of daily and weekly newspapers, magazines,
radio stations, television stations, and local cable channels all were assessed via their online content
offerings, in the same way that the outputs of online news sources such as community journalism
sites were assessed.
110
This approach (which is a reflection of the effort to create a realistically scalable methodologi-
cal approach) runs counter to the common assertion that certain types of legacy media (e.g., local
weekly print publications, ethnic media outlets) remain slow to utilize the internet as a means of
disseminating their content. We believe that we are at a point in the evolution of legacy media and
their place within the broader media ecosystem that this generalization likely no longer holds true.
The economic and strategic pressures and incentives to have an online presence, combined with the
inherent economic imperative to distribute content production costs across as broad an audience
base as possible (Hamilton, 2004), suggest that the content available online can serve as a reliable
indicator of the relative journalistic output across individual outlets, regardless of their “native” platform.
The key term here is indicator, as we are not seeking to produce a comprehensive inventory of
journalistic output, only a set of indicators that are conceptually and methodologically robust; that
can be employed in comparative analyses across communities or over time; and that can be scaled
to multiple communities at reasonable cost. We have found that, within the 100 communities cur-
rently being analyzed, incidences of a local media outlet lacking an online presence are rare.
Assessing journalistic performance

At the performance level, the goal is to provide measures of the extent to which the output serves
the information needs of local communities. Thus, we focused on three criteria: (1) whether
the content is original; (2) whether it is about the local community; and (3) whether it addresses
communities’ critical information needs (CINs).
These criteria are clearly rough and superficial indicators of the complex notion of the “qual-
ity” (Lacy and Rosenstiel, 2015) of the journalism being produced. However, these criteria pro-
vide a relatively simple, economical, straightforward, and replicable set of indicators of journalistic
performance. They also address the fundamental concern about whether journalistic sources are
addressing each communities’ critical information needs.
The notion of “critical information needs” has been central to the ongoing discourse about
the performance of local journalism (Knight Commission, 2009; Waldman, 2011). The approach
employed here draws upon this discourse and the research it has inspired (e.g., Becker and Yanich,
2015; Friedland et al., 2012). Specifically, each story is content analyzed to determine whether it
fits into one or more of the critical information needs categories identified in Friedland et al.’s
(2012) comprehensive literature review prepared for the FCC. Friedland et al. (2012) provide
eight categories of community critical information needs. These categories are as follows (see
Friedland et al., 2012, for detailed descriptions that provided the basis for each category’s coding
criteria):
1 Emergencies and risks

2 Health
3 Education
4 Transportation systems
5 Environment and planning
6 Economic development
7 Civic information
8 Political life
These categories provide a comprehensive and relatively straightforward schema for content
analyzing local news stories/posts in a way that could certainly be used to explore differences in
critical information needs across communities and differences in the extent to which local jour-
nalism sources are addressing specific critical information needs. These are intended as universal
111
categories and do not reflect the fact that communities certainly differ in terms of the relative
importance of each of these categories. For the purposes of our analysis, we are primarily con-
cerned with the basic binary of whether a story addresses a critical information need, in order to
compute metrics related to the proportion of journalistic output that addresses critical informa-
tion needs (see later).
Each story is also coded for whether it is original (i.e., produced by the journalism outlet rather
than reprinted or shared from elsewhere) and whether it is about the local community. The emphasis
on original stories serves to separate aggregation, linking, sharing, retweeting, and re-publication
activities, in an effort to determine the amount of original journalism output being produced.
This measure is intended to tap at the robustness of local journalism sources by determining how
active they are in producing news stories. A website story is considered ‘original’ if it has a byline
by an outlet’s reporter or if it lacks any indication that it is a reposting of content originally pro-
duced elsewhere.
The emphasis on locality seeks to analyze the extent to which the output of local journal-
ism sources focuses on the local community. This measure taps into the extent to which local
journalism is truly local, in the sense that stories provide community members with news and
information about, and directly relevant to, their communities. The narrow measure of local
content reflects the long-standing localism principle, which has featured prominently in demo-
cratic theory perspectives on media and in media policymaking (see Napoli, 2001). From this
perspective, the extent to which citizens are engaged with and capable of informed democratic
participation in their communities is a function of the availability of local news and information
about their communities (see George and Waldfogel, 2006). For this variable, we utilize a strict
definition of community, where we identified an item as about the community only if the subject
was an issue/event oriented around the specific town.
Together, these three variables of focus reflect some of the primary concerns about the state of
local journalism today: (1) that the economic pressures on local journalism create overwhelming
incentives to aggregate and repurpose existing content rather than engage in original report-
ing (see Anderson et al., 2012; Doctor, 2010); (2) that the changing technological dynamics for
news distribution and consumption are exacerbating the extent to which large-market or out-
of-market news can infiltrate local communities (George and Waldfogel, 2006); and (3) that the
increasing challenges associated with attracting and retaining an audience for news are compel-
ling news outlets away from substantive topics in favor of an emphasis on “soft” news, celebrity,
and sensationalism (see Patterson, 2013). For these reasons, these three variables represent a useful
set of top-level indicators of how well local journalism facilitates informed participation and
engagement in local community affairs.
Trained coders conduct the content analysis with the goal of achieving high levels of inter-
coder reliability. Google Translate facilitates coding of foreign language content. Subscriptions
were obtained for any sites whose current content is behind a paywall making them inaccessible
to the Internet Archive’s web crawler.
Metrics and analysis

The data gathered here can be employed to facilitate multivariate analysis and/or to facilitate
more basic comparative analyses across communities. From a multivariate standpoint, we are
primarily interested in whether it is possible to identify those characteristics of individual com-
munities that are significantly related to the robustness of local journalism. Essentially, we are
interested in identifying the characteristics of ‘at-risk’ communities. For this analytical approach,
the raw counts of outlets, story output, and number/proportion of stories that are original, are
local, and that address a critical information need serve as the relevant dependent variables.
112
Another indicator of the robustness of local journalism that we have constructed with these
data is a concentration index that indicates the extent to which the journalistic output in a com-
munity is emanating from many or few sources. This measure is derived from the well-known
Herfindahl-Hirschman Index (HHI). The traditional HHI is calculated by summing the squared
shares of each firm in a market to produce a measure of concentration. It is expressed as fol-
N
lows: H   si2 . In the case of this analysis, shares of total journalism output within a com-
i 1
munity are used in place of market shares. This HHI can also be calculated to focus exclusively
on specific categories of journalistic output, such as original news stories, local news stories, or
stories addressing a critical information need. These categories can be analyzed individually or in
combination (e.g., how concentrated is the output of news stories that are original and local and
that address a critical information need?).
Such output concentration measures reflect the historical significance of source diversity in
the journalistic sphere as well as the frequent use of the HHI as a tool for assessing diversity in
media (see Napoli, 2001). The extent to which journalistic output is concentrated within few
sources would seem to be a relevant indicator of the health of local journalism ecosystems. The
utility of the HHI in this context is purely comparative. While in economics variations in the
HHI have been found to be associated with variations in the behavior of firms (thus, an HHI of
1,300, for instance, has a specific meaning), here the HHI is being used as a comparative metric
with the individual values, having no inherent interpretation.
We are particularly interested in determining if, or to what extent, the robustness of local
journalism is a function of community characteristics such as average income, ethnic diversity,
and proximity to a large media market. Based on indications from the pilot research (see Napoli
et al., 2017), we are curious as to whether these factors may be negatively related to the robustness
of local journalism. In conducting such an analysis, it is also important to take into account rel-
evant controls such as population size and density as well the geographic size of the community.
We have gathered such data for each community being analyzed.
For more descriptive, comparative analyses across communities, obviously comparing raw
numbers in terms of infrastructure, output, and performance is problematic, given the large
degree of variance that is likely to exist across key variables and the challenges associated with
measuring those variables. We are currently conducting multivariate analyses in order to identify
the best approach for applying controls across communities in a way that facilitates meaningful
apples-to-apples comparisons.
In addition, social network analysis provides a means to measure the robustness and resil-
ience of a local news ecosystem. Prior work has, for instance, shown that network analysis can
be used to effectively map the connections between sources within a given locality (Weber and
Monge, 2011; Anderson, 2013). Moreover, connections between local news sources provides
a means for understanding connectivity and for mapping the sharing of information (Lowrey
and Erzikova, 2016). In addition, the sharing of content and the use of hyperlinking to con-
nect online local news sources indicates a level of competition and robustness that is generally
not present when websites are disconnected from one another. Indeed, some have suggested
that reciprocity is a hallmark of local news, and local news sources are reliant on the flow of
information from a given community (Harte et al., 2017); it is plausible to expect the same
principle to extend to connectivity amongst websites, which is key to building robustness in
online communities.
The aggregated web data from each community provides a basis for creating and measuring
social networks within local news communities. A social network can be created by capturing
the hyperlinks that exist between websites. This can be done manually or automatically using
computational methods (see Lin et al., 2014). In addition, variables capturing specifics of a given
113
local news outlet can be added to enhance the analysis (Weber, 2012). It is important when
using social network analysis to consider what counts as a “connection” between two websites;
often, it is useful to look beyond the existence of a single hyperlink and to consider more strin-
gent definitions of connection. Subsequently, a broad number of measures can be used to assess
the strength of a local news community – for example, considering the issue of robustness, the
network concept of embeddedness, which captures the degree to which an entity is embedded
within its broader community. In network terms, this can be measured in a number of ways,
including density of ties or clustering. Resilience is the degree to which a network of entities is
able to withstand disruptions and is closely connected to embeddedness. Resilience can further
be captured by higher level measures such as triads, or the degree to which entities are connected
to more than one other website.
Network analysis provides another lens through which local news can be examined and
assessed. Focusing on connections between local news sources provides a means for understand-
ing the degree to which a community of information exists. Moreover, by loosening the defini-
tion of local, it is also possible to examine the degree to which information is restricted to a single
locale or flowing across local boundaries.
Conclusion
The methodology outlined here is intended to facilitate both explanatory analyses of the factors
related to the robustness of local journalism and descriptive, longitudinal analyses that facilitate
the monitoring of trends over time. In terms of the former, expansion to a larger, more diverse
sample of communities (incorporating, for instance, communities larger than 300,000 residents
and smaller than 20,000 residents) would be beneficial in terms of enhancing generalizability. In
terms of the latter, expansion in terms of systematic, annual data-gathering would be essential for
longitudinal analysis.
Despite the fairly superficial nature of the content analysis, data-gathering at the scale pre-
sented here is resource intensive and time consuming. Going forward, one key question is
whether the content analysis process could be automated. Automating the process would cer-
tainly make expansion and/or longitudinal application of the methodology more feasible. Given
the fairly superficial nature of the content variables being analyzed and advances in machine
learning and textual analysis, this would seem to be an avenue worth pursuing. The data gath-
ered thus far could serve as the basis for developing, refining, and testing such an approach.
This research initiative has resonated with content archivists at the Internet Archive and the
Library of Congress, who recognize the importance of local journalism and understanding the
factors that affect its evolution, and who also recognize that local journalism, at the level of
granularity being analyzed here, largely slips through the cracks of most web archiving efforts.
For this reason, the Internet Archive has taken it upon itself, independently of our research proj-
ect, to archive the first day of every month for all of the URLs associated with our 100 sampled
communities. The Library of Congress is also exploring initiating a more systematic effort into
archiving local journalism.
The content archive upon which our research is based could support a wide range of addi-
tional research inquiries. Indeed, our intention is to make this archive publicly available to facili-
tate additional research. For instance, one researcher interested in crime coverage is exploring
the possibility of using the data to explore the relationship between local crime statistics and
local crime coverage. Such research, along with our ongoing research, focuses on the commu-
nity as the unit of analysis. Of course, these data could facilitate more granular units of analyses
including the individual media outlet or story. From such perspectives, it would be possible, for
114
example, to explore the relationship between outlet characteristics (size, medium, ownership, etc.)
and journalistic output as a way of enhancing our understanding of contemporary news orga-
nizations. These are just some of the possible ways that this archive of local journalistic output
could be utilized in future research. And, of course, were this archive to become longitudinal in
nature, the range of research possibilities may be enhanced further.
Further reading
Researchers are devoting increasing attention to local journalism. A key conceptual frame-
work for this research is the notion of media deserts, which has its origins in the work of
Michele Ferrier (2013; see also Ferrier et al., 2016). For an international perspective, Rasmus
Kleis Nielsen’s (2015) edited volume Local Journalism: The Decline of Newspapers and the Rise
of Digital Media is a wide-ranging and valuable collection. For a focus on strategic approaches
to preserving the economic viability of local journalism, Penelope Abernathy’s (2014) book
Saving Community Journalism: The Path to Profitability is a valuable resource. For those with an
interest in local news audiences and how they go about navigating their local news environ-
ments, the News Measures Research Project has recently published a qualitative analysis based
upon focus group research conducted in a number of New Jersey communities (McCollough
et al., 2017).
References
Abernathy, P. M. (2014) Saving Community Journalism: The Path to Profitability. Chapel Hill, NC: University
of North Carolina Press.
Anderson, C. W., Bell, E. and Shirky, C. (2012) Post-Industrial Journalism: Adapting to the Present. Retrieved
from http://towcenter.org/research/post-industrial-journalism/
Baker, C. E. (2002) Media, Markets, and Democracy. New York, NY: Cambridge University Press.
Barnett, S. and Townend, J. (2015) “Plurality, policy and the local: Can hyperlocals fill the gap?” Journalism
Practice, 9(3), 332–349.
Becker, A. and Yanich, D. (2015) Citizens’ Critical Information Needs and Local TV News. University of Dela-
ware Center for Community Research and Service Working Paper. Retrieved from www.ccrs.udel.edu/
sites/ccrs.udel.edu/files/BeckerYanich%20Broadcast%20News%20CIN%20Model%2030%20July%20
2015_0.pdf
Benson, R. (2013) Shaping Immigration News: A French-American Comparison. Cambridge, UK: Cambridge
University Press.
Caves, R. E. (1992) American Industry: Structure, Conduct, Performance. New York, NY: Prentice Hall.
Community Health Status Indicators Project Working Group. (2009) Data Sources, Definitions, and Notes
for CHSI2009. Department of Health and Human Services, Washington, DC. Retrieved from http://
communityhealth.hhs.gov
Doctor, K. (2010) Newsonomics: Twelve Trends that Will Shape the News You Get. New York, NY: St. Martin’s Press.
Downie, Jr., L. and Schudson, M. (2009, November/December) “The reconstruction of American journal-
ism.” Columbia Journalism Review. Retrieved from www.cjr.org/reconstruction/the_reconstruction_of_
american.php
Durkin, J. and Glaisyer, T. (2011) An Information Case Study: Scranton. Retrieved from http://oti.newamer
ica.net/publications/policy/Scranton
Durkin, J., Glaisyer, T. and Hadge, K. (2010) An Information Case Study: Seattle. Retrieved from http://oti.
newamerica.net/publications/policy/seattle
Eggerton, J. (2014, February 25) “House to hold hearing on FCC study.” Multichannel News. Retrieved from
www.multichannel.com/news/policy/house-hold-hearing-fcc-study/307382
Ferrier, M. (2013) “Interview with Debra Galant.” NJ News Commons. Retrieved from www.youtube.com/
watch?v=nK9OqINpyys
115
Ferrier, M., Sinha, G. and Ourich, M. (2016) “Media deserts: Monitoring the changing media ecosystem.” In
M. Lloyd and L. A. Friedland (eds.), The Communication Crisis in America, and How to Fix It. New York,
NY: Palgrave Macmillan (pp. 215–233).
Fico, F., Lacy, S., Baldwin, T., Bergan, D., Wildman, S. S. and Zube, P. (2013) “Newspapers devote far less
coverage to county government than to city governance.” Newspaper Research Journal, 34(1), 104–111.
Flint, J. (2014, February 28) “FCC cancels study after backlash from lawmakers.” LA Times. Retrieved from
www.latimes.com/entertainment/envelope/cotown/la-et-ct-fcc-cin-study-20140228-story.html
Friedland, L., Napoli, P. M., Ognyanova, K., Weil, C. and Wilson, E. J. (2012) Review of the Literature Regard-
ing Critical Information Needs of the American Public. Retrieved from http://transition.fcc.gov/bureaus/
ocbo/Final_Literature_Review.pdf
George, L. and Waldfogel, J. (2006) “The New York Times and the market for local newspapers.” American
Economic Review, 96(1), 435–447.
Gloria, K. and Hadge, K. (2010) An Information Community Case Study: Washington, D.C. Retrieved from http://
oti.newamerica.net/publications/policy/an_information_community_case_study_washington_dc
Hamilton, J. T. (2004) All the News That’s Fit to Sell: How the Market Transforms Information Into News. Prince-
ton, NJ: Princeton University Press.
Harte, D., Williams, A. and Turner, J. (2017) “Reciprocity and the hyperlocal journalist.” Journalism Practice,
11(2–3), 160–176.
Holt, K. and Karlsson, M. (2015) “‘Random acts of journalism?’: How citizen journalists tell the news in
Sweden.” New Media & Society, 17(11), 1795–1810.
Karlsson, M. and Holt, K. (2014) “Is anyone out there? Assessing Swedish citizen- generated community
journalism.” Journalism Practice, 8(2), 164–180.
Knight Commission. (2009) Informing Communities: Sustaining Democracy in the Digital Age. Report of the
Knight Commission on the Information Needs of Communities in a Democracy. Retrieved from www.
knightcomm.org/read-the-report-and-comment/
Lacy, S. and Rosenstiel, T. (2015) Defining and Measuring Quality Journalism. News Measures Research Proj-
ect Working Paper. Retrieved from http://mpii.rutgers.edu/wp-content/uploads/sites/129/2015/04/
Defining-and-Measuring-Quality-Journalism.pdf
Leibowitz, J. (2009) ‘Creative Destruction’ or Just ‘Destruction’: How Will Journalism Survive the Internet Age?
Opening Remarks for FTC News Media Workshop. Retrieved from www.ftc.gov/sites/default/files/
documents/public_statements/creative-destruction-or-just-destruction-how-will-journalism-survive-
internet-age/091201newsmedia.pdf
Lin, J., Gholami, M. and Rao, J. (2014, April) “Infrastructure for supporting exploration and discovery in
web archives.” Proceedings of the 23rd International Conference on World Wide Web, ACM (pp. 851–856).
Lin, W. and Song, H. (2006) “Geo-ethnic storytelling: An examination of ethnic media content in contem-
porary immigrant communities.” Journalism, 7(3), 362–388.
Lowrey, W. and Erzikova, E. (2016) “News media ecosystems and population dynamics: A cross-cultural
analysis.” In Advancing Media Production Research. Basingstoke, UK: Palgrave Macmillan (pp. 79–94).
McCollough, K., Crowell, J. K. and Napoli, P. M. (2017) “Portrait of the online local news audience.”
Digital Journalism, 5(1), 100–118.
Morgan, F. (2011) The Research Triangle, North Carolina: A Region of Locally Owned Media Outlets and Entrepre-
neurs on the Verge. New America Foundation. Retrieved from http://oti.newamerica.net/publications/
policy/the_research_triangle_north_carolina
Morgan, F. (2013) “The stories not told: A case study of the information needs of Siler City, North Caro-
lina.” I/S: A Journal of Law and Policy for the Information Society, 8(3), 481–529.
Napoli, P. M. (2001) Foundations of Communications Policy: Principles and Process in the Regulation of Electronic
Media. Cresskill, NJ: Hampton Press.
Napoli, P. M. and Friedland, L. (2016) “US Communications Policy Research and the Integration of the
Administrative and Critical Communication Research Traditions,” Journal of Information Policy, 6, 41–65.
Napoli, P. M. and Stonbely, S. (in press) “Policy issues surrounding journalism.” In J. Nussbaum (ed.), The
Oxford Research Encyclopedia of Communication. New York, NY: Oxford University Press.
Napoli, P. M., Stonbely, S., McCollough, K. and Renninger, B. (2017) “Local news and the information
needs of local communities.” Journalism Practice, 11(4), 373–395.
Nielsen, R. K. (ed.). (2015) Local Journalism: The Decline of Newspapers and the Rise of Digital Media. London:
I.B. Tauris & Co.
Pai, A. (2014, February 10) “The FCC wades into the newsroom.” Wall Street Journal. Retrieved from www.
wsj.com/articles/SB10001424052702304680904579366903828260732
116
Patterson, T. E. (2013) Informing the News: The Need for Knowledge-Based Journalism. New York, NY: Vintage.
Pew Research Center. (2010) How News Happens: A Study of the News Ecosystem of One American City.
Retrieved from www.journalism.org/2010/01/11/how-news-happens/
Pew Research Center. (2015) Local News in a Digital Age. Retrieved from www.journalism.org/files/2015/03/
PJ_MediaEcology_030515_pdf.pdf
Picard, R. (2014) “Twilight or new dawn of journalism? Evidence from the changing news ecosystem.”
Ramos, D., Gunes, M. H., Mensing, D. and Ryfe, D. (2013) “Mapping emerging news networks: A case
study of the San Francisco Bay Area.” Studies in Computational Intelligence, 424, 237–244.
Ryfe, D., Mensing, D., Ceker, H. and Gunes, M. (2012, April) “Popularity is not the same thing as influ-
ence: A study of the Bay Area news system.” Paper presented at the International Symposium on Online
Journalism, Austin, TX. Retrieved from https://online.journalism.utexas.edu/2012/papers/Mensing.pdf
Schudson, M. (2005) “The US model of journalism: Exception or exemplar.” In H. de Burgh (ed.), Making
Journalists: Diverse Models, Global Issues. New York, NY: Routledge.
Social Solutions International. (2013, April). “Research design for the multi-market study of critical
information needs.” Prepared for the Federal Communications Commission. Retrieved from www.fcc.gov/
encyclopedia/research-design-multi-market-study-critical-information-needs
Starr, P. (2009, March 4) “Goodbye to the age of newspapers (hello to a new era of corruption).” The New
Republic. Retrieved from www.newrepublic.com/article/goodbye-the-age-newspapers-hello-new-era-
corruption
Stites, T. (2011, December 8) “Layoffs and cutbacks lead to a new world of news deserts.” Nieman Lab. Retrieved
from www.niemanlab.org/2011/12/tom-stites-layoffs-and-cutbacks-lead-to-a-new-world-of-news-
deserts/
St. John III, B., Johnson, K. and Nah, S. (2014) “Patch.com: The challenge of connective community jour-
nalism in the digital sphere.” Journalism Practice, 8(2), 197–212.
Sustainable Jersey. (2013) A Better Tomorrow, One Community at a Time. Retrieved fromwww.sustainable
jersey.com/fileadmin/media/Grants_and_Resources/Green_Team_Resources/Program_Overview/
Sustainable_Jersey_Overview_Brochure.pdf
Tankersley, J. (2015, April 23) “Why the PR industry is sucking up Pulitzer winners.” The Washington Post.
Retrieved from www.washingtonpost.com/blogs/wonkblog/wp/2015/04/23/why-the-pr-industry-
is-sucking-up-pulitzer-winners/
U.S. Senate. (2009) “The future of journalism.” Hearing before the U.S. Senate Committee, on Commerce, Sci-
ence, and Transportation, Subcommittee on Communications, Technology, and the Internet. 111th Congress, first
session, May 6.
Waldman, S. (2011) The Information Needs of Communities: The Changing Media Landscape in a Broadband Age.
Federal Communications Commission, Washington, DC. Retrieved from www.fcc.gov/infoneedsreport
Weber, M. S. (2012) “Newspapers and the long-term implications of hyperlinking.” Journal of Computer-
Mediated Communication, 17(2), 187–201.
Weber, M. S. and Monge, P. (2011) “The flow of digital news in a network of sources, authorities, and bubs.”
Williams, A., Harte, D. and Turner, J. (2015) “Filling the news hole? UK community news and the crisis in
local journalism.” In R. K. Nielsen (ed.), Local Journalism: The Decline of Newspapers in the Rise of Digital
Media. London: I.B. Tauris (pp. 203–223).
117
9
RECONSTRUCTING THE
DYNAMICS OF THE DIGITAL
NEWS ECOSYSTEM
A case study on news diffusion processes
Elisabeth Günther, Florian Buhl, and Thorsten Quandt
Introduction: preconditions to observe news diffusion

processes in the digital news ecosystem
News diffusion processes as the conceptual frame

at the ecosystem level
The digitization of news production, distribution, and consumption has provided journalism
researchers with plenty of opportunity to explore the establishment of new structures and pro-
cesses within the developing networked media system (Franklin and Eldridge, 2017). Alongside
rather manifest distinctions of online journalism, such as the reorganization of newsrooms, jour-
nalism scholars have started to discover digital news phenomena whose accessibility requires
the advance of research designs in the first place. Observers might recognize these phenomena
without the help of innovative research methods. Readers might recall personal experiences
with online journalists’ publish-first-and-update-later routines (Karlsson and Strömbäck, 2010;
Saltzis, 2012; Widholm, 2016) or the acceleration of the news cycle (Rosenberg and Feldman,
2008). However, transferring such anecdotal evidence into a set of more systematic, generalizable
observations of such phenomena is a reconstructive endeavor that benefits from the interplay of
methodological advances and theoretically focused procedures of data analysis.
In this vein, Zamith (this volume, Chapter 7) presents innovations in content analysis that
enable journalism researchers to systematically trace the updates of online news items and web-
sites (see also Karlsson and Strömbäck, 2010; Saltzis, 2012; Widholm, 2016). While we share the
shift from a static to a dynamic perspective, our main focus is on the digital news ecosystem as
a whole – or subsections, at least – and on discovering the potential dynamics of news diffusion
processes among the variety of providers within. Our approach is inspired by the idea that online
journalists’ reaction times to emerging stories is getting shorter and shorter in the wake of van-
ishing publication deadlines (Risley, 2000) – due to technological opportunities of live reporting
(see Artwick, this volume, Chapter 22), the normalization of immediate coverage beyond major
news events (Lim, 2012; Rosenberg and Feldman, 2008), or the orientation toward the latest sto-
ries of competing news sites (Boczkowski, 2010). As the trend toward immediate event coverage
appears to be a global attribute of digital newswork, we aim to complement studies observing
118
The dynamics of the digital news ecosystem
the routines in single newsrooms (Boczkowski, 2010) with a research design that reconstructs the
flow of information in the digital news ecosystem (Weber and Monge, 2011) and especially its
dynamics. If a multitude of online newsrooms tend to cover newsworthy events instantaneously,
the routines of individual newsrooms will lead to close timing of event-related publication deci-
sions at the ecosystem level. Conceptually borrowing from the literature on the diffusion of news
in the public (Rogers, 2000), we developed a research design to capture resulting news diffusion
processes in the digital news ecosystem (Buhl et al., 2018). Within this analytical framework,
we can map the dynamics of these processes, as we relate the amount of time elapsed since the
first report on an event with the subsequently accumulating number of news sites covering the
story, too.
We aimed first to establish and systematically capture the phenomenon of digital news dif-
fusion processes and second to analyze the conditional factors of their dynamics (see Buhl et al.,
2016, 2018).
Computational methods enable the reconstruction

of news diffusion processes
Similar to tracking the velocity of online news items (see Zamith, this volume, Chapter 22; see
also Karlsson and Strömbäck, 2010; Widholm, 2016), capturing the dynamics of news diffu-
sion processes among online news sites relies heavily on recent advances in automated content
analysis. News diffusion processes can be uncovered only when shared routines of prompt news
dissemination in digital journalism converge with innovative methods of content analysis in
journalism research, which are both highly sensitive to time and scalable to the ecosystem level of
analysis. Real time data collection of news items published by the variety of providers within the
same digital news ecosystem is a precondition for the reconstruction of diffusion processes. We
believe it is fair to say this analytical prerequisite would push manual coding to its logistic limits
(see Figure 9.1). Additionally, the project’s ‘catch all – select later’ strategy requires data storage in
a searchable database, because researchers can identify relevant events only ex post facto.
Overall data collection 480,727 online news reports
Identification of relevant events:

Scan news for events meeting three criteria:
Attracts interest
Is unexpected 131 events
Is clearly identifiable with few keywords
61,132
Event-based identification of news articles: online news reports
Conduct database query with a regular expression of up
to four keywords for each event
Data management:
For each event, within 24 hours, manually select the 95 events
first report by each online news site
1,919
Data cleaning: online news reports
Exclude any event which cannot be clearly
distinguished from similar events
Exclude any event reported by < 10 news sites
Figure 9.1 Stepwise procedure to identify relevant events and related online news reports (Buhl et al., 2018)
119
Elisabeth Günther et al.
We compiled a research design meeting these requirements in our work on the dynamics of
diffusion processes among professional online news sites in Germany (Buhl et al., 2016, 2018),
which we will use as an example throughout this chapter. However, we will generalize from
this case study to provide readers with the basic principles of systematically capturing diffusion
processes within the digital news ecosystem. In the subsequent section, we will detail the require-
ments for data collection and provide readers with solutions for each step during the process.
In the third section, we will turn to the data analysis, first outlining how to reconstruct digital
news diffusion processes and how to systematically describe their dynamics. Thereafter, we will
introduce analytical procedures both to identify subgroups of diffusion processes that share simi-
lar dynamic patterns and to test potential relationships between event attributes and diffusion
dynamics. In the final section, we will discuss links of the news diffusion process-approach to
similar concepts as well as potential future advances of studying the dynamics of digital journal-
ism from an ecosystem perspective.
Tracking news diffusion processes online: basic research design

Over the past decades, both object of study and methods of analysis in digital journalism research
have been transformed by the digital revolution. Tracking news diffusion processes online is a
good example for the changes that need to be implemented on both sides of the research process –
concerning both what and how – and a challenging task at multiple levels. In the following, we first
outline challenges and possible solutions related to the blurry boundaries and fluidity of content
online (What is the digital news ecosystem? How do we perform data collection on relevant con-
tent?) and then turn to the specific unit of analysis (What are the markers by which to define rel-
evant events? How can online news diffusion processes on an event be reconstructed for analysis?).
Sampling unit: representing the online news ecosystem

Defining the boundaries of the online ecosystem is a Sisyphean task: online, legacy media are
embedded in a complex network of information providers, with multiple ties to and from social
media (Leskovec et al., 2009), internet-specific news formats (e.g., BuzzFeed, or news aggregators
such as Reddit), and of course traditional sources and distribution channels offline. Links are not
always made explicit and thus further obscure the shape and boundaries of the news ecosystem.
Consequently, the list of relevant websites for a study on news diffusion online is, to a certain
degree, open to discussion. Combined with the sheer size of available information online, it is at
this step inevitable to make a compromise of some sort. In terms of defining the sampling unit,
we identify two main strategies.
Researchers interested in the interplay between different types of news providers might prioritize
diversity over completeness. By drawing a sample of a few sources from many different formats,
this allows for anecdotal insights into their role and strategy. Researchers who aim to explore a
specific phenomenon, as we do in our case study, might follow the opposite strategy and focus on
reproducing a comprehensive model of one domain only. By targeting an internally consistent frac-
tion of the information cosmos (in our case, German print newspapers’ online outlets) we can
benefit from two main advantages. First, this approach provides a common baseline for assessing
in-group differences and identifying characteristic subgroups. Second, basing the sample on a
predefined number of websites makes it feasible to add meta-information on each of the provid-
ers, enabling us to control for variations in their professional background and working routines.
When working with a high variety of sources, as accomplished by the first approach, the work-
load associated with this step is likely to get out of hand and thus push the logistic boundaries of
a project (Waldherr et al., 2017).
120
For our case study, the second strategy provides an adequate starting point to explore the ways
traditional news providers have adapted to the challenges presented by the 24-hour news cycle.
Instead of taking on the impossible task of trying to cover all news websites online, we start with
the well-structured set of print newspapers and work our way forward from there: key data is
provided by Schütz (2012), listing all print publications for the German market as of 2012. For
each entry, we investigated whether an online news site with an RSS (Real Simple Syndication)
feed was provided. This feature provides considerable benefits with regard to a reliable data col-
lection (as explained in the following section) and, since RSS is a common online standard, did
not noticeably compromise our sample. This way, we constructed a sample of 28 websites, cover-
ing all major news outlets in Germany (for a detailed list, see Buhl et al., 2018).
Data collection: considerations

Now that we are equipped with a list of news websites to represent the online news ecosys-
tem, the next challenge is to develop a plan for the data collection. Given the fluidity of online
content, it is crucial to consider possible complications ahead of time and test the practicability
of one’s approach in a pretest. We follow a catch-all strategy, meaning that we collect all news
articles published by the given websites over the course of a given time period before selecting
the actual units of analysis, i.e., the events for which we then reconstruct digital news diffusion
processes. There are two main reasons behind this decision. First, we are interested in unexpected
events (our rationale for this criterion is explained in the following section), so we cannot predict
their beginning and/or end.
Second, pretests indicate that online news diffusion processes cannot reliably be recovered ex
post facto – even if only a few days or even just several hours had passed. This is mainly due to
online journalists’ routine of updating, rewriting, or deleting content during the initial time span
in which news on unexpected events unfolds, adapting their stories to new information in real
time (Karlsson and Strömbäck, 2010; Saltzis, 2012; Widholm, 2016). If not made transparent, such
live editing can leave researchers confused: Looking at published articles in hindsight, it might
appear as if one news outlet was hours ahead of others, with detailed coverage on an event at a
point in time when others only had a note along the lines of “something has happened, we will
elaborate shortly”. In the highly competitive environment that the online news ecosystem con-
stitutes, such a scenario is unlikely. Even in a case where one newsroom is given a head start due
to exclusive access to information, other websites will catch on quickly; online, co-orientation
among news providers has become both easier and, with immediacy being a crucial production
norm, more important (Boczkowski, 2010; Karlsson and Strömbäck, 2010). A more plausible
explanation is that the timestamp of the initial post has remained unchanged throughout later
edits. If the research interest is to reconstruct the first substantial report on a specific event for
each news website, as we plan to do, this situation makes continuous and (near) real-time data
collection necessary. For our case study, we therefore set up a comprehensive data collection over
the course of nine months (June 2013 to March 2014), making sure that seasonal effects, such as
variations in working routines over the holiday season in December or the so-called ‘silly season’
over the summer months, can be ruled out, and that a sufficient number of relevant events are
covered within the sample period.
Data collection: technical setup

Both the reliable collection and the efficient storage of online content are associated with certain
technical challenges. While web-scraping is surely not a trivial task, there is good introductory
material available to guide motivated non-IT researchers through the basic steps of this process.
121
Trilling (2017) provides a good example of a step-by-step guide, handholding readers through
the lines of a custom Python script. Collecting news articles from a couple of websites over
the course of a week is, with such introduction and a little practice, doable even for beginners.
Widening the scope both regarding the number of sources to be monitored simultaneously and
regarding the time frame of the data collection (over several weeks, months, or even years) will,
however, quickly lead to a data volume that requires a more advanced setup.
The task of reconstructing online news diffusion processes, as presented in our case study, can
consequently be characterized as a Big Data project. Setting up our data collection, several techni-
cal aspects had to be considered: (1) As described, we need to download new articles regularly,
aiming for at least an hourly interval, to ensure that quick developments are captured in near real
time. (2) One round of downloads, preferably including preprocessing and data cleaning, has to
be completed within this time frame to avoid “collusion” with the following rounds, otherwise
risking a delay. Due to the aforementioned journalistic routine of re-editing published content,
regular and continuous data collection is essential to guarantee the reliability of our project.
(3) Given the number of steps involved in each round (access RSS feeds, identify links to news
articles, extract article metadata, download raw HTML content, save to database, perform data
cleaning and preprocessing, etc.) and the repetitive nature of the overall process, performance
needs to be optimized to the level of split seconds per task, since they quickly add up to serious
delays. (4) In times of breaking news coverage, the article volume per hour is expected to go up.
On top of all previous considerations, this means the whole process needs to be implemented in
a way that does not already exhaust the available resources (time and processing power) during
standard operation. This way, we make sure that the data collection stays in running order and
does not slow – or even worse – break down during the high-intensity phases that are of special
interest to us.
The key concept here is scalability: for example, saving 1 second per article does not sound like
much, but means that we can save 5 minutes for the simultaneous download of 300 articles – a
plausible number for an hourly collection of 30 news websites. Double the number of articles
(which might happen during especially busy morning hours), and we save 10 minutes for just
one task in our data collection pipeline. One of the main technical levers to optimize perfor-
mance lies in parallel processing, meaning that independent tasks are efficiently distributed across
all available CPU cores of the system. Applying parallel processing to each step can thus quickly
cut working time to a fraction: before optimizing performance, our data collection (without pre-
processing and cleaning) took 30 minutes – a task that is now1 completed in less than 2 minutes.
Depending on the specifics of the project (number of sources, time frame for data collec-
tion, hardware specs, skills, and experience of the researchers involved), arising challenges might
move the task of the automated data collection from a social science pet project into the field of
expertise of an IT professional. In our case, we rely on the datamessie open-source project (avail-
able on GitHub under the GNU General Public License). In cases where a custom solution is
required, collaborations with technically skilled colleagues or even professional contracting is
recommended for beginners. For advice on database selection and setup, Günther, Trilling, and
van de Velde (2018) introduce relevant options and considerations. In any way, we would like to
stress the point that given the fluidity of online content, a regular and timely monitoring of the
data collection is key to avoid irreversible loss of information.
Unit of analysis: selecting ‘relevant’ events

Based on the described setup, our continuous hourly data collection of 28 news websites over
the course of nine months results in an overall sample of N = 480,727 news articles. Not all
are relevant to our analysis. The next step is to identify relevant events that occurred within the
122
observed time frame and which we will then trace in our sample to reconstruct their news diffu-
sion processes. The ‘relevance’ of an event is not easy to operationalize and is contingent on the
specific research question – for our case study, we defined three criteria to guide this process: (1)
The event is newsworthy enough to trigger news coverage from a sufficient number of websites
in our sample. This criterion establishes that there is broad enough interest within the online
news ecosystem, allowing us to observe and compare varying reaction times across the websites
in our sample. (2) The event is unexpected, meaning that news coverage has a clear but unknown
starting point. As an example, the Winter Olympics 2014 are certainly a remarkable event but
have been announced and discussed months ahead of the actual opening ceremony. This makes
it difficult to isolate reports on the event itself in the following analysis phase. In contrast, an
underdog winning a gold medal in one of the competitions might trigger the wide and easily
traceable news diffusion process we aim at. (3) The event can be described with a few distinc-
tive keywords. This is a technical prerequisite: Given the size of our sample, we need to keep the
manual workload in the following selection phase to a limit and will therefore filter reports on
each event by means of an automated keyword search.
By evaluating every downloaded news article from two online news sites under study against
all three criteria, we identified 131 possibly relevant events within the sampled time frame.
Putting it all together: reconstructing news diffusion processes

With all necessary information at hand, we can start processing and analyzing the collected data
(see Figure 9.1). First, we are equipped with a large collection of all news articles published by the
observed 28 websites in the sample period, stored in a database, and second, we have a tailored list
of 131 potentially relevant events that occurred within this time frame. The analysis builds upon
both sets of information by retrieving all news articles that make a reference to one of the selected
events, technically implemented by means of an automated keyword search. This requires basic
database skills, such as a working knowledge of SQL and a fundamental understanding of regu-
lar expressions (Friedl, 2006). For each event, we defined a regular expression with one or two
mandatory keywords and up to two optional keywords: to be identified as event related, an article
must mention both keyword1 and, if specified, keyword2, in combination with either keyword3
or, if specified, keyword4 (see Appendix A in Buhl et al., 2018 for the final list of keywords for
all events). While this sounds rather technical, it is not complicated when put into action: For
example, the birth of ‘Royal Baby’ Prince George is described by a combination of ‘William’ with
either ‘birth’ or ‘son’ – any news article that matches these criteria is filtered for further analysis.
Keywords typically referred to the who – including named entities for the main actors involved in
an event (names of persons, institutions, and organizations such as the Italian Senate), the where –
names of the place or scene (e.g. London Heathrow airport), and the what – dense descriptions of
the event if available (e.g., birth, train accident). Since our interest is constrained to highly news-
worthy and unexpected events, finding clear keywords for them was mostly a straightforward
process. To avoid overlooking important cues, we systematically double-checked and refined all
keywords in a manual pretest. Although manual reevaluation is costly, we recommend erring
on the side of caution and choosing keywords that rather yield too many false positives than to
miss important articles with different phrasing. In addition, certain information that perfectly
describes an event in hindsight might not have been available yet in the initial stages of reporting.
Finally, we joined the selected keywords to a regular expression and then conducted a database
query for matching news articles (see Figure 9.1). In our case study, this query resulted in 61,132
matches for 131 events – a considerable reduction, but still a lot of information to be processed.
For each event, we now manually selected the very first news article by each news site (within a
24-hour limit after the first report was published), which also yielded information on the overall
123
number of websites that reported on the event in question. Given our criteria for relevance as
described in the previous section, aimed to restrain the scope to events characterized by high
newsworthiness, we determine a minimum range of the diffusion rate and removed all events
that were picked up by less than 10 news outlets. This way, we constructed a final sample of
1,919 news articles regarding 95 events. As with the data collection itself, this massive reduction
of information brings with it a high workload for pretesting and double-checking to ensure a
reliable research process.
Analyzing the dynamics of news diffusion processes

While the concept of news flows might seem straightforward at first, there are different ways
to look at it (we will expand on this aspect later). In our case study, we conceptualized news
diffusion processes as the relationship between the amount of time elapsed since the particular
event has been reported for the first time (x-axis) and the accumulated number of online news
providers in the same news ecosystem having reported the event so far (y-axis, see Figure 9.2).
The diffusion of stories among news providers is calculated by continually adding news sites at
the time lags of their first reports on the specific story. As we were able to determine publication
times to the split second in our case study (Buhl et al., 2018), usually there was only one news
site joining the diffusion process per time frame. But we also found instances of several news sites
releasing their stories at the very same point of time (we assume because some local news sites
share national newsdesks), so that the accumulation grew by multiple newspapers in that particular
second.
Literally, you can best get a picture of the dynamics of digital news diffusion processes by
plotting them. Figure 9.2 displays the curves of diffusion processes of three exemplary events
from our case study on professional online news sites in Germany (cf. Buhl et al., 2018): the
birth of “Royal Baby” Prince George (circles), the release of allegations of child abuse against
Woody Allen (triangles), and a Texas state court decision declaring a ban on same-sex marriage as
unconstitutional (squares). At first glance, the three cases do not seem to be that different: Within
the 24-hour time frame observed for each diffusion process, it takes 13.7 hours for the last online
news site to report the birth of Prince George, 9.8 hours in the case of the allegations against
Woody Allen, and 14.0 hours in case of the ruling on same-sex marriage in Texas. Comparing
the range of the diffusion processes, i.e., the total number of online news sites having reported
each event after 24 hours, we find a slightly stronger hint at differentiated patterns of diffusion
processes. There were 26 online news sites reporting the birth of the ‘Royal Baby’, 21 sites cover-
ing the Woody Allen story, and only 11 sites reporting the Texas court decision.
From the plots of the three diffusion curves (see Figure 9.2), however, it appears obvious that
specifying length and range of complete digital news diffusion processes does not tell the full story
about their dynamics. Contrary to the process patterns of the diffusion of news about the allega-
tions against Woody Allen and the Texas court ruling, the diffusion curve for the ‘Royal Baby’
approaches saturation many hours prior to the final completion of the process. That is, during
diffusion processes of this type, the vast majority of online news sites issue their first reports on
the corresponding event in the direct aftermath of the incident – or its very first coverage in
the digital news ecosystem, respectively. During this short time frame, the number of news sites
joining the diffusion process grows at very fast pace, while it slows down significantly for the
remaining time span. We found this dynamic to be a common pattern in our case study on dif-
fusion processes within the ecosystem of German online news sites (cf. Buhl et al., 2018). For
the curves of 68 from a total of 95 diffusion processes, we could identify a shift from an early
burst to slower rates of accumulation. To describe news diffusion processes accurately, we conse-
quently need to differentiate the main diffusion phase from the process as a whole. To determine
124
Range
●
25 ●
●
●
● Maindiffusion
●
20 ● phase Event
●
● "Royal Baby": Birth
● ●
● of Prince George
15 ●
● Woody Allen:
●
●
Abuse allegations
●
10 ● TX: Ban on same-sex
● marriage unconstitutional
●
●
●
5 ●
●
●
Cumulative number of online news sites

●
● Length
0
0 3 6 9 12 15 18 21 24
Time lag (hours)
Figure 9.2 Curves of the digital diffusion processes of three exemplary events
the point of shifting accumulation pace for each curve, we had to rely on visual inspection in our
own study, because the number of data points per curve – the total number of news providers
reporting each event – was too small for the application of more sophisticated procedures. For
the ‘Royal Baby’ story, for example, we found the endpoint of the main diffusion phase to be
reached after 1.6 hours, including 23 online news sites (see Figure 9.2). For processes lacking a
clear transition from a surge at the beginning to slower accumulation rates, e.g., the diffusion of
the Woody Allen story and the Texas court ruling, one option is to define the main diffusion to
last for the full process (as we did), so that the endpoint of the complete process coincides with
the endpoint of the main diffusion phase.
Exploring dynamics of online news diffusion processes

in large data collections
The groundwork regarding data preprocessing and cleaning, combined with the set of compu-
tational methods that have been outlined in the previous section, now allows us to apply this
procedure to many more diffusion processes from the same digital news ecosystem. Therefore,
we can widen our analytical perspective beyond describing single case studies of digital news
diffusion processes in their own right to mapping their general temporal patterns across a
variety of cases (see Figure 9.3 for the sample of 95 digital diffusion processes from our own
research). From this global point of view, we consider each reconstructed diffusion process
as a case and the whole of them as the sample. Once we have extracted information about
length, range, and duration of the main diffusion phase from each individual process, we can
provide frequency tables and histograms for each measure and calculate their central tendency
and variability for the whole sample of processes – or for subsamples, if we wish. In our own
research, we made all time-related calculations exact to the second, the time format of our
time-lag data, but we also reported the results converted into hours for better readability (Buhl
et al., 2018).
Figure 9.3 Diffusion curves for 95 events as reported by German online news sites, grouped by dynamics-
based cluster membership (Buhl et al., 2018)
126
In our research of diffusion processes among German online news sites, we found the distribu-
tion of lengths of main diffusion phases to be strongly skewed to the right, with its mean at 5.1
hours (18,219 seconds) and its median at 2.2 hours (7,916 seconds), confirming the commonness
of bursts and surges at the beginning of diffusion curves. To make the specific diffusion dynamics
accessible beyond the endpoints of main diffusion phases, we statistically reconstructed the aver-
age diffusion curves, which was done by pooling the time-lag data from each diffusion process
belonging to the same subset into a virtual, aggregated diffusion process. The pooled data can be
rearranged from small to large time lags with the help of procedures of event history analysis, e.g.,
according to the Kaplan-Meier estimation method. The resulting aggregated diffusion curves
represent the typical diffusion-process dynamics in the sample. They provide the proportion of
online news sites having reported events at specific time lags, visualizing the fast accumulation
of reporting sites at the beginning of diffusion processes, for example. This analytical perspective
becomes especially informative as we start comparing the dynamics among subsamples of diffu-
sion processes (see examples in subsequent sections).
Structuring samples of digital news diffusion processes

by diffusion dynamics
One of our research questions was aimed at identifying recurring and common dynamics of
digital news diffusion processes. The three exemplary diffusion processes plotted in Figure 9.2
indicate that these dynamics are not restricted to a single pattern. To account for this variety of
dynamics, we first employed an inductive procedure based on process dynamics. We divided the
full sample of diffusion processes into subsamples featuring similar dynamics by clustering the
diffusion processes. The resulting process clusters serve as an indicator of common, recurring
diffusion dynamics in the digital news ecosystem.
In our research on diffusion processes among German online news sites (Buhl et al., 2018),
we conducted a hierarchical agglomerative cluster analysis. We treated each diffusion process as
a case and used the full sets of time-lag data from each process. We deliberately decided against
controlling for the length of the processes, as we consider this information a crucial differentiat-
ing factor among diffusion dynamics. The results of the cluster analysis suggests a three-cluster
solution, i.e., a differentiation among three types of common diffusion dynamics (see Figure
9.3): high-range diffusion processes characterized by bursts briefly after the very first reports
about the respective incident (Cluster 1, n = 43, e.g., the birth of Prince George); processes
featuring less distinct bursts at their beginning and lower range both in total and relative to
the amount of time elapsed (Cluster 2, n = 28, e.g., the publication of abuse allegations against
Woody Allen); and processes accumulating slowly during longer time frames (Cluster 3, n =
24, e.g., the Texas state court ruling in favor of same-sex marriage). To characterize the process
dynamics of each homogeneous subsample and to make comparisons among them, summary
statistics of range, length, and duration of main diffusion phases for each subsample provide
valuable insights. Additionally, we can reconstruct typical diffusion dynamics in each subsample
with the help of the data-pooling procedure outlined in the previous section. For each of the
three dynamics-based clusters of diffusion processes we found in our study, Figure 9.4 displays
the curves of the resulting aggregated diffusion processes. The display of differentiated diffusion
dynamics of subsamples offers a good starting point for discovering post hoc explanations for
each pattern. Beyond high newsworthiness, we found events that follow the dynamics of Clus-
ter 1 normally do not require journalists to add contextual information to be easily understood
by the audience, for example.
127
1.00
Proportion of online news sites
0.25 0.50
0.00 0.75
0 3 6 9 12 15 18 21 24
Time lag (hours)
Cluster 1 Cluster 2 Cluster 3
Figure 9.4 Aggregated curves for three clusters of digital diffusion processes featuring similar dynamics,
from a total of 95 diffusion processes among German online news sites (Buhl et al., 2018)
Comparing the dynamics of online news diffusion processes

by process attributes
If prior knowledge about the logics within the digital news ecosystem allows researchers to out-
line hypotheses about factors influencing the dynamics of diffusion processes, this information
needs to be collected systematically. Factors may be attributes of events (either necessitating or
inhibiting immediate coverage) or conditions of working routines in online newsrooms varying
between diffusion processes. To explain diffusion dynamics in our own work on German online
news sites (Buhl et al., 2016), we analyzed relationships with the news factors of events as well as
with starting points during daytime vs. during the night, when online newsrooms are typically
less staffed. To test the assumed variations, we again relied on event history analysis according to
the Kaplan-Meier estimation method. Aggregated diffusion curves were reconstructed for sub-
samples of diffusion processes sharing the same predefined attribute. As a result, we can describe
the typical diffusion patterns of processes sharing the same attribute and compare subsets of
processes with different attributes. Figure 9.5 displays the result of this analysis for the time of
day a diffusion process set off as a contingent condition of digital newswork. Diffusion processes
starting between 10 a.m. and 10 p.m. (Berlin time) were more likely to exhibit distinct bursts
at their beginning, on average having already involved 12.5 news sites after just one hour. For
processes starting during the night (10 p.m.–10 a.m.), the diffusion rate is slower during the first
couple of hours (7.2 news sites after one hour). As visible in Figure 9.5, they need more time to
catch up to the range of day processes.
Conclusion and directions for future research

Computational methods allow us to explore dynamic environments, such as online news web-
sites, and to take a snapshot of the ever-changing online news ecosystem. In this chapter, we
128
1.00
Proportion of online news sites
0.25 0.50
0.00 0.75
0 3 6 9 12 15 18 21 24
Time lag (hours)
Day (10 a.m.–10 p.m.) Night (10 p.m.–10 a.m.)
Figure 9.5 Dynamics of aggregated diffusion processes, grouped by the time of the day they were first
reported (Buhl et al., 2016)
presented our take on this complex task, as illustrated by a case study on diffusion processes
among German online news sites.
It is important to point out that the processes reconstructed in our case study might not
necessarily qualify as news flows, however tempting this assumption might be. Observing time
lags among various online news outlets reporting the same story, as we do, may indicate a co-
orientation among journalists in various newsrooms, resulting in the circulation of a story. But
time lags might also be caused by independent, parallel processing of information from the very
same source. Time lags then indicate varying speed, for example due to the availability of staffers
during different times of the day or variations in strategy and fact-checking routines (see also
Harder et al., 2017). Especially when focusing on the coverage of major news events, as we do,
we expect little uncertainty among journalists about whether or not to cover the story, so that
they are unlikely to back up their news decisions by co-orientation processes (Donsbach, 2004).
There are many other possible – and equally valid – ways to approach and analyze dynam-
ics in online news. Depending on the research perspective, dynamics might be understood as
the spreading of a particular data file. Hussain (2012), for example, explores online gatekeeping
mechanisms that mediate the distribution of the top 120 viral videos in the 2008 U.S. presidential
campaign and finds that relevant actors have expanded well beyond journalists. In a similar take,
Trilling et al. (2017) examine factors that contribute to a news article’s shareworthiness on social
media. Here, dynamics refer to the development of an article’s success based on its distribution
via social media. Based on less clear-cut pieces of information, Leskovec et al. (2009) explore
online news dynamics in terms of the handoff of memes – distinctive ideas or sentences – through
the global news cycle, with results indicating a common pattern of influence from mainstream
to social media. In yet another take on online news dynamics, Weber and Monge (2011) trace
the information flow by means of hyperlink networks, highlighting the critical role of a few key
websites in effectively controlling the overall conversation. Also based on hyperlinks, Messner
129
and Distaso (2008) analyzed the sources cited in weblogs in order to determine who sets their
agenda, and the other way around, examined blogs as a reference in mainstream media. The latter
two studies consequently define dynamics as an attribute of the online news ecosystem’s structure
rather than as the flow of particular data or information.
In light of these various approaches, it becomes apparent that analyzing the dynamics in
online environments is not an easy undertaking. Journalism researchers have quickly adapted to
the new challenges that the digital transformation has afforded. Various perspectives, based on
innovative methodological approaches, add to our understanding of digital news as a complex
living entity. Each approach has its own advantages. With our case study on online news diffu-
sion, we add a framework that focuses on the speed and shape with which news on real-world
events spreads through the online news ecosystem, and, by means of the described methodologi-
cal walkthrough, hope to provide inspiration for future research.
Further reading
For readers who are interested in research that – in line with our own work (Buhl et al., 2016,
2018) – maps the diffusion of stories in the digital news ecosystem, we recommend Welber’s
(2016) thesis on Gatekeeping in the Digital Age, as well as Harder et al. (2017) “news story”
approach to “Intermedia Agenda Setting in the Social Media Age”. With regard to data analysis,
there are many more ways to explore the content of large text collections. For an overview of
various automated content analysis techniques, Günther and Quandt’s (2016) “Word Counts and
Topic Models” might serve as a good starting point. Alvarez’s (2016) collected volume on the
Computational Social Sciences is recommended for readers with an overall interest in this field, with
excellent case studies that might serve as inspiration for future research.
Note
1 Unfortunately, performance optimization was only implemented when the data collection for our case
study was already up and running. With the current setup, a much shorter interval is well within reach,
e.g., collecting news articles every 15 minutes in future projects.
References
Alvarez, R. M. (ed.). (2016) Computational Social Science: Discovery and Prediction. New York, NY: Cambridge
University Press.
Boczkowski, P. (2010) News at Work: Imitation in an Age of Information Abundance. Chicago, IL: University of
Chicago Press.
Buhl, F., Günther, E. and Quandt, T. (2016) “Unambiguous burstiness: Towards explaining the dynamics
of digital news flows from opportunity structures, news factors, and topics.” Paper presented at the 66th
Annual Conference of the International Communication Association, Fukuoka, 9–13 June.
Buhl, F., Günther, E. and Quandt, T. (2018) “Observing the dynamics of the online news ecosystem: News
diffusion processes among German news sites.” Journalism Studies, 19(1), 79–104.
Donsbach, W. (2004) “Psychology of news decisions: Factors behind journalists’ professional behavior.”
Journalism, 5(2), 131–157.
Franklin, B. and Eldridge II, S. (eds.). (2017) The Routledge Companion to Digital Journalism Studies. London:
Routledge.
Friedl, J. (2006) Mastering Regular Expressions. Sebastopol, CA: O’Reilly Media.
Günther, E. and Quandt, T. (2016) “Word Counts and Topic Models: Automated Text Analysis Methods
for Digital Journalism Research.” Digital Journalism, 4(1), 75–88.
Günther, E., Trilling, D. and van de Velde, B. (2018) “But how do we store it? (Big) Data architecture in
the social-scientific research process.” In C. Stuetzer, M. Welker and M. Egger (eds.), Computational Social
130
Science in the Age of Big Data: Concepts, Methodologies, Tools, and Applications. Cologne: Herbert von Halem
Verlag (pp. 161–187).
Harder, R. A., Sevenans, J. and van Aelst, P. (2017) “Intermedia agenda setting in the social media age: How
traditional players dominate the news agenda in election times.” The International Journal of Press/Politics,
22(3), 275–293.
Hussain, M. (2012) “Journalism’s digital disconnect: The growth of campaign content and entertainment
gatekeepers in viral political information.” Journalism, 13(8), 1024–1040.
Karlsson, M. and Strömbäck, J. (2010) “Freezing the flow of online news: Exploring approaches to the study
of the liquidity of online news.” Journalism Studies, 11(1), 2–19.
Leskovec, J., Backstrom, L. and Kleinberg, J. (2009) “Meme-tracking and the dynamics of the news cycle.”
Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
(pp. 497–506).
Lim, J. (2012) “The mythological status of the immediacy of the most important online news: An analysis
of top news flows in diverse online media.” Journalism Studies, 13(1), 71–89.
Messner, M. and Distaso, M. W. (2008) “The source cycle: How traditional media and weblogs use each
other as sources.” Journalism Studies, 9(3), 447–463.
Risley, F. (2000) “Newspapers and timeliness: The impact of the telegraph and the internet.” American Jour-
nalism, 17(4), 97–103.
Rogers, E. M. (2000) “Reflections on news event diffusion research.” Journalism & Mass Communication
Quarterly, 77(3), 561–576.
Rosenberg, H. and Feldman, C. S. (2008) No Time to Think: The Menace of Media Speed and the 24-Hour News
Cycle. New York, NY: Continuum International.
Saltzis, K. (2012) “Breaking news online: How news stories are updated and maintained around-the-clock.”
Journalism Practice, 6(5/6), 702–710.
Schütz, W. (2012) “Redaktionelle und verlegerische Struktur der deutschen Tagespresse.” Media Perspektiven, 11.
Trilling, D. (2017) “Doing computational social science with python: An introduction.” Social Science
Research Network. Retrieved from http://papers.ssrn.com/abstract=2737682
Trilling, D., Tolochko, P. and Burscher, B. (2017) “From newsworthiness to shareworthiness: How to predict
news sharing based on article characteristics.” Journalism & Mass Communication Quarterly, 94(1), 38–60.
Waldherr, A., Maier, D., Miltner, P. and Günther, E. (2017) “Big data, big noise: The challenge of finding
issue networks on the web.” Social Science Computer Review, 35(4), 427–443.
Weber, M. S. and Monge, P. (2011) “The flow of digital news in a network of sources, authorities, and hubs.”
Welbers, K. (2016) Gatekeeping in the Digital Age. Doctoral thesis, Vrije Universiteit Amsterdam, Amsterdam,
the Netherlands. Retrieved from https://research.vu.nl/ws/portalfiles/portal/1587652
Widholm, A. (2016) “Tracing online news in motion: Time and duration in the study of liquid journalism.”
131
10
TESTING THE MYTH
OF ENCLAVES
A discussion of research designs for assessing
algorithmic curation
Jacob Ørmen
More and more of the information we receive in the world is curated by algorithms. Every
time people use digital intermediaries, such as social network sites or search engines, computer
programs guide us to the information that serves our needs (supposedly). As these services play
an increasingly important role as access points to news and information (Newman et al., 2016),
users rely more and more on algorithms to guide them to what they encounter on the internet.
News organizations also depend heavily on digital intermediaries to serve news to their readers.
In this relationship, search engines and social network sites act as algorithmic gatekeepers (Boz-
dag, 2013) in the intersection between content producers and users. It remains a crucial task for
journalism research and practitioners alike to understand how algorithmic curation affects the
type of information users are exposed to and interact with.
To study this, this chapter takes a widespread myth about algorithmic curation as a starting
point. The ‘myth of enclaves’ is the idea that digital intermediaries like search engines and social
network sites drive people further into ‘filter bubbles’ (Pariser, 2012) or ‘echo chambers’, where
they are only exposed to content and ideas in line with their own ideological beliefs. It is a myth
not because it is (necessarily) fictional, but because it remains contested among scholars. Some
research has found algorithmic curation to lead people into enclaves through personalization;
that is the mechanisms whereby algorithms adapt the presentation of content to the individual
user (Bakshy et al., 2015; Jacobson et al., 2016). In contrast, others have found algorithms to
concentrate, rather than fragment, attention on the already well-established actors online (Dutton
et al., 2017; Hindman, 2009). Moreover, a third group of scholars argue that digital intermediaries
expose people to a greater variety of information than they seek out themselves (Flaxman et al.,
2016; Vaccari et al., 2016). The ‘myth of enclaves’ lives on in part because the academic public is
so mixed about what to think about it.
The aim of this chapter is neither to confirm nor debunk the ‘myth of enclaves’ but instead to
discuss the various methods applied to investigate the role of algorithmic curation in fragment-
ing, concentrating, or polarizing the public’s exposure to content on the internet. By critically
assessing the methods and techniques used to test the myth of enclaves, we can achieve a firmer
ground on which to evaluate the credibility of the claims in favor or against the myth. To do this,
this chapter begins by outlining what ‘algorithmic curation’ covers. It then proceeds by analyzing
132
Testing the myth of enclaves
the methodological strengths and weaknesses of four prototypical research designs for studying
algorithmic curation. In the discussion, the chapter situates the research designs in context with
a fifth approach that moves beyond mere exposure.
Algorithmic curation
Algorithmic curation describes a process whereby a computer program selects pieces of content
to present to users based on a step of logical operations encoded by human programmers (Braun
and Gillespie, 2011). A wide range of services online make use of algorithmic curation to provide
users with the options that best match their presumed interests, be it TV and film (e.g., over-
the-top distributors like Netflix and HBO), music (e.g., streaming services like Spotify or Apple
Music), goods (e.g., marketplaces like Amazon and Taobao), information (e.g., search engines
like Google or Baidu), or social activity (e.g., social network sites likes Facebook and WeChat).
There are at least three ways algorithmic curation matters for how users meet content on the
internet. First, algorithms operate by selecting content to display to the users. The programs select
content by searching through a library (e.g. all indexable documents on the web in the case
of search engines or all posts by a person’s network in a given time frame in the case of social
network sites) from where a few pieces of content are chosen to be presented to the users (e.g.,
the search results list or the newsfeed). For the purpose of methodological discussions, we could
call the library for a population and the content selected to be displayed to individual users for a
sample. The key operation of curation algorithms is identifying the specific sample in the popu-
lation that best matches the needs of the user in the particular moment of usage. The greater the
size of the population and the smaller the sample needed, the larger is the role played by algorith-
mic curation. The growing number of websites, videos, songs, movies, goods, and social material
online, combined with the limited time available for users, makes curation (in algorithmic or
other forms) a necessary premise for finding content on the internet at all.
The second way algorithmic curation plays out is by ranking the selected content. To do this,
algorithms rely on relevance criteria encoded by programmers. In its early days, Google developed
a search algorithm for assessing the relevance of individual web pages that takes both the hyper-
link network between websites (structural level) and the content on web pages (semantic level)
into account (Brin and Page, 1998). In this way, each website would get a popularity score – its
structural importance in the link network, called PageRank – and a relevance score based on the
search string. These criteria would be universal for all searchers using the search engine. Similarly,
content providers often rank content on their websites based on general popularity (e.g., measured
by clicks or time spent with content). Sometimes this information is fed back to users as “most
popular” lists on news websites or trending stories on social network sites (Webster, 2014). Rank-
ing plays an important role in guiding users to particularly relevant content in the sample selected.
The third way curation algorithms operate is by personalizing the sample to match each indi-
vidual user. On top of the general relevance criteria discussed earlier, digital intermediaries inte-
grate context-specific signals in the curation process. Over the years, the complexity of signals
that feed the curation algorithms in Google Search has risen steeply and now includes factors
such as geographical location, language-setting, and search history (Granka, 2010). Facebook
originally used an algorithm (EdgeRank) to assess the relevancy of each post based on the affinity
of user connections, previous interactions with the content, and how recently the content was
posted. This has since been supplanted by a machine learning approach that takes “more than
100,000” factors into considering when ranking content in the Facebook newsfeed (McGee,
2013). Currently, personalization algorithms like those governing search results and newsfeeds
are so complex in structure and execution that few – if any – know exactly how they function.
133
Jacob Ørmen
For research purposes, personalization has made it difficult to study curation empirically, since
algorithms produce different samples of content for each user and each use situation. It makes a
difference who interacts with the algorithm, when, and where.
This raises particular issues for how to research algorithms. For recommendation mechanisms
such as those suggesting other goods to purchase (on Amazon), movies to watch (on Netflix), or
music to listen to (on Spotify), personalization is driven largely by previous interactions within the
service. However, social and search algorithms increasingly draw on information outside of the
service. Facebook, for instance, tracks user activity across the internet and uses that information
to sample content in the newsfeed (Debatin et al., 2009). Likewise, Google displays ads (through
the AdSense network) based on search and click patterns around the web, both on and off Google
services. This development has wide-ranging methodological consequences for studying algo-
rithmic curation. It entails that there are no universal scores determining the output of the algo-
rithm but rather a range of particular scores that depends in large part on signals picked up from
the individual user. After the introduction of personalization factors in algorithmic curation, it is
an inductive fallacy to generalize the specific results for one user (e.g., the search results list to a
given query) to all other users across time and space.
This chapter focuses on the challenge personalization presents to empirical research on algo-
rithmic curation. Therefore, by focusing on intermediaries like social network sites and search
engines that rely heavily on personalization, it leaves out curation that only selects and rank but
not personalizes, such as generic databases relying on keyword searches or popularity lists (e.g.,
most-read news stories on a news website) that are updated algorithmically but not adapted to
the individual user. Understanding ways to study personalization makes a difference for how
scholars can make sense of information curation in particular and the myth of enclaves in general.
Directed and undirected curation

Apart from the various ways algorithms work, there is also a fundamental difference between
types of curation. When a user browses the feed on a social network site, the algorithm displays
results without concrete input from the user. Curation is undirected. For this type of curation, the
algorithms produce a ranked sample of content based solely on what it estimates to be in our
greatest interest at that particular moment. The algorithm pushes information to the user. This
is reminiscent of the way human editors have selected and ranked information (to be displayed
in their publications) for centuries. In contrast, when a searcher enters keywords into a search
engine, the algorithms deliver results in response. Here, curation is directed. The input provided
by the searcher (a query or a search string) gives the algorithm further information to determine
the relevance of each piece of content in the population. A user pulls information through the
algorithm. This is similar to the work carried out by human librarians for millennia. In reality,
many services integrate both types of curation, as is the case when users search on a social net-
work site, or when search engines display information to individual users solely based on previous
behavior. However, the distinction between undirected and directed curation makes an important
difference for how we can analyze personalization mechanisms.
Directed curation is by far the easiest to track, since data are available from the services. To
query Google, the searcher does not have to create an account and log on (the same goes for
services like Twitter and YouTube, but not for Facebook). It is thus possible to access and retrieve
data in simple ways, as each query forms a unique URL. For instance, a search on google.com
with the string “barack obama” appears like this www.google.com/search?q=barack+obama
(cleared for other type of metadata, such as browser information that Google integrates as well).
In short, the search term is directly retrievable from the search URL. This makes it possible to
study which terms people search for and the results they click on solely by knowing the URLs
134
they have visited and in which sequence. Another benefit is that directed curation algorithms
operate on one grand population (such as all indexable documents on the web). In theory, every
searcher can receive exactly the same results to a query. Thereby, researchers can compare directly
the sample of results (e.g., top 10 results) presented to each user as a response to the same search
query. Accordingly, it is possible to study both the extend as well as the type of personalization
occurring.
In contrast, undirected curation presents a number of unique challenges to research. First, to
get access to undirected curation on social network sites a researcher has to create a profile or
log in through existing user accounts. It used to be possible to access the newsfeed through the
API (Application Programming Interface), but that service has been discontinued as of October
2015.1 Currently, scraping newsfeeds directly through the user interface is prohibited by Face-
book’s Terms of Service.2 Instead, it is necessary to access the newsfeed directly through each
individual user account. Second, the algorithms rely solely on the content shared by connections
in each user’s (as well as sponsored) content to compose the sample. It is difficult to compare
newsfeeds between users directly, since the samples of content are not drawn from the same
population. Instead, it is possible to study the relationship between sample and population for
each user – that is, which content the algorithms deem more or less relevant to the specific user.
Accordingly, there are substantial differences in how researchers can access and study algorith-
mic curation. In the next section, I go through widespread strategies for studying both indirect
and direct curation. To make the analysis more concrete, the search engine Google and the social
network site Facebook serve as cases of directed and undirected curation, respectively.
Four prototypical designs for assessing algorithmic curation

Now I look into ways research has assessed the effect of algorithmic curation on the type of
information users are exposed to and interact with. The studies I discuss here all rely on behav-
ioral data collected with/on users as they interact with algorithms. This is not to discount the
importance of theoretical or ethical discussions of algorithms – which remain a pertinent and
prolific activity in communication research (for discussion see Nahon, 2016; Dörr, this volume,
Chapter 23) – but merely to zoom in on the part of the literature that focus on methodological
issues of researching curation mechanisms.
To guide the discussion, I have constructed a matrix that encapsulates what can be considered
from four prototypical research designs for studying algorithmic curation (see Table 10.1). In the
columns, the model distinguishes between whether the research designs operate with artificial
or real users. Artificial user accounts are those constructed for the purpose of research. Typi-
cally, this is done by setting up specific profiles on a social network site or a search engine. It
also includes research profiles that are modeled on real user accounts but inscribed with artificial
agency for the purpose of research. In the rows, the matrix distinguishes between artificial and
real settings. The idea is to make a delineation between studies that operate in a preexisting and
live environment (real setting) and those that construct a specific environment for the purpose
of research (artificial setting). When the rows and columns are crossed, four prototypical designs
emerge: simulation (artificial setting and artificial user), experimentation (artificial setting and real
Table 10.1 Matrix of prototypical algorithmic curation designs
Artificial user Real user
Artificial setting Simulation Experimentation

Real setting Manipulation Observation
135
Jacob Ørmen
user), manipulation (real setting and artificial user), and observation (real setting and real user). These
designs configure the extent to which the researcher has access and control over the user experi-
ence as well as the research environment.
The four designs are prototypical in that each design is a rough characterization of typical
features shared by studies in general but not applicable to any study in particular. Many of the
projects working on algorithmic curation draw on elements from different designs, and for good
reason, as will be returned to in the discussion. Nonetheless, the designs are helpful as sketches
to think about the different methods and techniques researchers can apply to conduct empirical
studies.
Simulation
With simulation, researchers have the benefit of controlling both the setting and user input. This
approach is strictly not empirical but computational. The researcher does not study real users
or real settings but instead constructs a research environment where the interaction between
artificial user profiles and platform characteristics are simulated. This approach draws on game-
theoretical assumptions to model human behavior in socio-technical systems, termed agent-
based simulation.3 In recent years, it has been a popular approach to test and design information
systems (Ryczko et al., 2017). The goal is often to improve ways of displaying information, for
instance through personalization steps, based on assumptions about which outcomes serve the
individual user better or worse (Micarelli et al., 2007). Thus, a core element of simulation stud-
ies is to set up parameters, such as assumptions about agents’ rationality, as well as methods for
evaluation outcomes. Often the models draw on data from real-world users, but they are treated
as computational agents inscribed with certain preferences and behavioral tendencies.
Simulation studies have dealt with a range of issues in algorithmic curation. The spread of
information such as news stories through social network sites is estimated with simulations of
millions of user accounts (Del Vicario et al., 2016; Goel et al., 2016). Here, the goal is to pre-
dict information cascades (or virality) of various types of content. Other studies have sought
to find optimal levels of personalization for individual users by simulating interactions among
artificial profiles (Guo et al., 2013; Chung et al., 2016). Together, simulation studies are help-
ful in showing general tendencies across many or all users of a search engine or social network
site. This design provides an important insight into how signals affect algorithms, such as how
likely it is for content to show up in individual newsfeeds given previous patterns of informa-
tion diffusion.
The obvious downside of such a design is the loss of ecological validity. The simulated envi-
ronment can only approximate a real setting but never replicate the complexity of a Facebook
or Google ecology. Predicting future patterns of user behavior from logs of past activities might
function well for very general phenomena such as the spread of information cascade – although
scholars do debate whether simulation models can even do this reliably (Cheng et al., 2014) –
but it is not well-suited to show how algorithmic curation affects the individual user. A related
issue is the behavioral assumptions programmed into the simulations. Without an input from
real-world users, researchers have to rely on general heuristics such as rational choice theory to
simulate user agency. These assumptions will naturally be rather crude and in some cases mislead-
ing. Lastly, the need to operate with a ground-truth to evaluate the results of simulations is prob-
lematic in itself. Scholars should always be wary of attributing fixed preferences to individuals
from the outset (such as person x always preferring content y to content z). Nonetheless, social
research is well advised to draw inspiration from simulation studies to understand personalization
mechanisms at play.
136
Manipulation
The most widespread approach to study directed curation has been to set up artificial research
profiles and let them loose in real-world settings (notably on Google Search). This approach
makes it possible to control and manipulate some personalization signals – such as language
settings, geo-location of IP-address, anonymous browsing – while holding others constant. In
this way, the researcher can seek to mitigate (or work with) the effects of the algorithm on the
outcome presented to users (for an elaborated discussion see Ørmen, 2015). Whereas some have
advocated a strategy where the researcher keeps a clean research profile with as few personaliza-
tion signals as possible (Rogers, 2013), others have actively manipulated one or more signs such as
language settings or geographical location to study the causal effect on personalization (Ørmen,
2015; Kliman-Silver et al., 2015). A third line of research has sought to construct “extreme cases”
(Kuzel, 1999) to compare profiles that are as different as possible (Dutton et al., 2017; Feuz et al.,
2011). In the latter case, research teams have trained profiles to display ideological preferences
by feeding each profile with a predefined set of search strings belonging to either the left or
right side of the political spectrum. Subsequently, a set of common keywords have been queried
through each profile, and the differences in results have then been attributed largely to person-
alization processes. By constructing profiles bottom-up, from a blank slate to an ideological
extremist, these studies attempt to set up a quasi-experimental design in semi-controlled condi-
tions. In this way, researchers can attempt to ‘game’ the algorithm to maximize (or minimize)
personalization of results.
However, operating with artificial research profiles also creates issues of ecological validity. It is
difficult to mimic real human behavior on the internet solely by setting up a fake Google profile.
First, algorithms on search engines and social network sites take a plethora of signals into account,
including general browsing behavior (including beyond the boundaries of the website or app).
For instance, if one is an avid user of popular services online, these services are likely to show
up in the Google search results as well, in particular if they are Google related, such as YouTube
(Edelman and Lai, 2016). Second, real users interact with content in ways that can be hard to rep-
licate with artificial accounts (such as clicking patterns). Third, digital intermediaries experiment
with and randomize the samples produced, which introduce substantial – and unknown – noise
into analysis. There are simply too many unknown factors at play for artificial research profiles
to truly mimic real-world user behavior.
Experimentation
The opposite approach has been to study real-world users in artificial settings. A number of
studies have designed applications mimicking well-known social network sites to experiment
with real-world users. This approach follows the controlled experiment approach known from
laboratory research, where participants are recruited and randomly assigned to either a treatment
or a control group. One study extracted newsfeed data through the now-unavailable Facebook
API to test various personalization designs with participants in a lab setting (Eslami et al., 2016).
In a similar manner, studies have shown the influence of social endorsements on social network
sites by randomly assigning types of social interactions (kinship, tie-strength) to news stories on
an experimental platform designed for the purpose of research (Kulkarni and Chi, 2013; Messing
and Westwood, 2014).
The greatest advantage of an experimental approach is the ability to assess causality. By con-
trolling which algorithmic signals influence which users, it is possible to test hypotheses directly.
The downside is that external validity might be harmed. If the lab settings (artificial platforms
137
Jacob Ørmen
designed to look like well-known services) fail to reproduce the complexity of behavior online,
then the conclusions drawn do not hold. For instance, the importance of social cues for interac-
tions on social network sites might not be as clear or consistent when users are exposed to con-
tent in the messy real-world settings where there are many more variables at play. A further issue
is that the experimental setting itself might interfere with the validity of the results. A lab setting,
where a person is asked to focus on one or more specific tasks (and possibly rewarded to do so),
does not replicate a use situation in daily life, where people interact with content in the midst of
all the other tasks, thoughts, and duties that preoccupy their minds in daily life. Accordingly, the
controlled experiment has merits but suffers greatly from the artificial situation that lab settings
create, both offline and online.
Observation
The most desirable approach is to study real-world users in real time as they are exposed to and
interact with content in real settings. This is often done by tracking users with cookies and
through user accounts as they move across the internet. It is an observational approach in the
sense that it is reminiscent to a “fly on the wall” approach in ethnography, where researchers
seek to blend in with the environment and observe unnoticed. The major advantage of tracking
users online is that researchers can document behavioral patterns of individuals in an unobtrusive
manner. In this way, researchers can collect detailed levels of data on a large number of users.
Unfortunately, it is typically only industry researchers that have access to this kind of data. Most
of their research is for optimizing business and thus considered proprietary information not made
available to researchers.
Therefore, research on algorithmic curation using observational data primarily comes from
the digital intermediaries themselves. In the case of directed curation, researchers at Google have
assessed the impact of various personalization measures on the click-through patterns of Google
News users in real time (Liu et al., 2010; Das et al., 2007). For nondirected curation, Facebook’s
research team has carried out studies on personalization using observation data on millions of
users (Bakshy et al., 2015). Working with the raw data collected by Facebook, the researchers
are able to compare the total population of content available to be displayed to each user (all the
posts created or shared by their connections) and the sample of content presented to each user
(the personalized newsfeed) as well as the select few pieces of content that users interact with
(by clicking, liking, sharing, etc.). These studies provide crucial insight into algorithmic curation,
since they can cover the whole user population in real time as well as past time (historical data).
This allows for comparisons on an unprecedented scale and detail.
If researchers gain access to archived or live observational data, there is the possibility to move
beyond mere observation and conduct natural and field experiments. For a natural experiment,
researchers compare two or more preexisting groups that vary on key characteristics and then
investigate differences between groups in time. Thus, the researcher cannot manipulate the vari-
ables as in a controlled experiment but will have to attribute causality based on the differences
observed. For instance, one natural experiment showed how users tend to click on Google prod-
ucts more often when the algorithm ranks such products higher in search results lists (Edelman
and Lai, 2016). The major downside of a natural experiment is, naturally, the lack of possible
interference by the researcher. Causality can be assumed from observations but not tested in
practice. In contrast, field experiments offer the ability to work with real-world users in their real
settings. Field experiments in online settings can be done unobtrusively and are thus preferred
to controlled experiments, discussed earlier. A Facebook study showed how a get-out-to-vote
button significantly increased the propensity of users to actual get out and vote on election day
(Jones et al., 2017). The downside here is that field experiments require full access to live user
138
data, and it is challenging to do while abiding by academic ethics codes. In short, rich obser-
vational data allow for studies that attribute causality by comparison (natural experiments) or
interference (field experiments) in live environments.
There are challenges to the observation design as well. First of all, few researchers have access
to the scale and depth of data required to assess personalization mechanisms. Forming partner-
ships with or buying data directly from industry actors is typically the only way to gain access
to large-scale observational data. For field experiments – where one needs access not only to the
user base but also to back-end programming – this is even more difficult. The most promising
way to produce observational data from the outside has been to recruit human participants (e.g.,
through Mechanical Turk) and then instruct them to query particular keywords on Google (Dil-
lahunt et al., 2015; Hannak et al., 2013) or record their Facebook newsfeed (Bhargava et al., 2015).
However, it is not ethically unproblematic to hire people to participate in research in this way, in
particular if the project relies on precarious labor platforms like Mechanical Turk. Accordingly, in
academic practice large-scale observational studies will often be infeasible or unethical – or both.
Moving forward: an alternative design

So far, the analysis of methods has dealt solely with the type of content algorithms tend to expose
users to. Such a focus overlooks the part played by users in decoding curation processes. As
media research has shown in the past 50 years, users are not mere passive consumers but reflexive
and critical beings. We need to take user agency seriously in assessing how algorithmic cura-
tion affects their practices. Some studies have attempted to do so through interviews or surveys
with users (Bucher, 2017; Rader and Gray, 2015; Powers, 2017). Getting user discourses serves an
important purpose in outlining general attitudes of and affections toward algorithmic curation.
However, it tells us little about how users relate to the specific acts of curation, for instance the
particular constellation of search results or stories in the newsfeed. It remains an open question
to which extend users navigate, peruse, and contest the results displayed by the algorithm. To
do this, we need to integrate the focus on exposure as outlined in the four designs with user
interpretations.
Arguably, the best way to test the ‘myth of enclaves’ would be to document not what people
get exposed to but what they think about what they are exposed to. One way to do this would be
to recruit a small-n sample of participants, for instance following a maximum variation sampling
strategy (Kuzel, 1999), to explore differences across individuals. Researcher could install logging
software on their primary devices and interview them about their use experiences. Currently,
several applications exist that can log usage across devices (e.g. laptop, smartphone, and tablets)
and some that can collect detailed descriptions of web behavior (e.g. www.webhistorian.org).
Following the ideas of experience sampling methodology (Larson and Csikszentmihalyi, 1983),
users could be prompted with small questions right after their visits to specific services, such as
Facebook and Google, to get immediate reactions to and interpretations of the curation process.
Likewise, the logged behavioral data (or screenshots of a search results list or newsfeed) could
be used as prompts in follow-up interviews to trigger reflections on routinized and habitual
behavior (Ørmen and Thorhauge, 2015). In this way, researchers and participants can discuss
algorithmic curation on a detailed level with data on the participant’s own behavior as the basis.
Such an approach, an interaction design, would also enable researchers to situate ordinary people’s
practices with algorithmic curation in their broader practices of media use. We know from
other studies that most people keep a varied media diary, moving far beyond social media (for
an overview, see Helles et al., 2015). Including this contextual information about individuals in a
study of algorithmic curation would be a strong test of the ‘myth of enclaves’ and a supplement
to existing research designs.
139
Jacob Ørmen
Conclusion
In the introduction, we saw how scholars have come up with widely different empirical results
when they have investigated the ‘myth of enclaves’. Now that the chapter has discussed and eval-
uated methods for testing this myth, part of an explanation for these inconsistencies emerges. The
fundamental differences in how scholars study algorithmic curation and personalization mecha-
nisms naturally affect the type of conclusions that can be drawn. The empirical material used in
simulation, manipulation, experimentation, and observation studies is so different that it renders
direct comparisons between design types untenable. The findings from one study do not translate
well into results from a study using a fundamentally different design. Unfortunately, the research
design that tends to produce the most convincing empirical results (observational studies of real
users in real-world settings) requires special access to proprietary data owned by digital interme-
diaries. One way to remedy this is for academic researchers to seek partnerships with industry
researchers to get access to real-world observational data and to conduct field experiments with
real-world users in a live environment. More of this research would surely strengthen the empiri-
cal basis in the debate on ‘enclavization’. To move the discussion forward, I suggested an alterna-
tive route that combines observational data (logged by research applications) with people’s own
discourses about algorithmic curation. Such a design would shed further light on poorly covered
aspects of polarization and fragmentation – namely, the experiences of ordinary people in the
broader context of their media-saturated lives. The need to move beyond single services or plat-
forms in the study of personalization effects is the key takeaway from this chapter. Given all these
challenges to empirical research, it is likely that the ‘myth of enclaves’ will neither be confirmed
nor debunked in the near future. Instead, the myth will live on and, hopefully, inspire researchers
to design new and innovative ways to study algorithms at work in our daily lives.
Further reading
The most comprehensive overview of the relationship between algorithms and users is The Mar-
ketplace of Attention (2014) by James G. Webster. This book also offers a sobering view on the
polarization and fragmentation debate underlying the ‘myth of enclaves’. Matthew Hindman’s
2009 book The Myth of Digital Democracy continues to be a relevant reminder of the power of
algorithms to concentrate rather than disperse public attention. Solomon Messing and Sean
Westwood’s paper from 2014, “Selective Exposure in the Age of Social Media”, provides a good
overview of empirical studies on polarization and critically evaluates the experimental approach
in online settings.
Notes
1 https://developers.facebook.com/docs/graph-api/reference/v2.9/user/home, accessed May 4, 2017.
2 www.facebook.com/terms, accessed May 4, 2017. It is possible to apply for a scraping permit from
Facebook, but Facebook explicitly prohibits scraping for the purpose of “academic consumption” www.
facebook.com/apps/site_scraping_tos.php, accessed May 4, 2017.
3 Thanks to Andreas Gregersen for making me aware of this connection to game theory.
References
Bakshy, E., Messing, S. and Adamic, L. (2015) “Exposure to ideologically diverse news and opinion on
Facebook.” Science, 348(6239), 1130–1132. doi:10.1126/science.aaa1160
Bhargava, P., Brdiczka, O. and Roberts, M. (2015) “Unsupervised modeling of users’ interests from their
Facebook profiles and activities.” Proceedings of the 20th International Conference on Intelligent User Interfaces,
ACM, New York, NY (pp. 191–201).
140
Bozdag, E. (2013) “Bias in algorithmic filtering and personalization.” Ethics and Information Technology, 15,
209–227.
Braun, J. and Gillespie, T. (2011) “Hosting the public discourse, hosting the public: When online news and
social media converge.” Journalism Practice, 5(4), 383–398.
Brin, S. and Page, L. (1998) “The anatomy of a large-scale hypertextual web search engine.” WWW7: Pro-
ceedings of the Seventh International Conference on World Wide Web 7, Elsevier, Amsterdam (pp. 107–117).
Bucher, T. (2017) “The algorithmic imaginary: Exploring the ordinary affects of Facebook algorithms.”
Information, Communication & Society, 20(1), 30–44. doi:10.1080/1369118X.2016.1154086
Cheng, J., Adamic, L., Dow, P. A., Kleinberg, J. M. and Leskovec, J. (2014) “Can cascades be predicted?”
WWW ’14: Proceedings of the 23rd International Conference on World Wide Web, ACM, New York, NY (pp.
925–936).
Chung, T. S., Wedel, M. and Rust, R. T. (2016) “Adaptive personalization using social networks.” Journal of
the Academy of Marketing Science, 44(1), 66–87. doi:10.1007/s11747-015-0441-x
Das, A. S., Datar, M., Garg, A. and Rajaram, S. (2007) “Google news personalization: Scalable online col-
laborative filtering.” WWW ’07: Proceedings of the 16th International Conference on World Wide Web, ACM,
New York, NY (pp. 271–280).
Debatin, B., Lovejoy, J. P., Horn, A.-K. and Hughes, B. N. (2009) “Facebook and online privacy: Attitudes,
behaviors, and unintended consequences.” Journal of Computer-Mediated Communication, 15(1), 83–108.
doi:10.1111/j.1083-6101.2009.01494.x
Del Vicario, M., Bessi, A., Zollo, F., Petroni, F., Scala, A., Caldarelli, G., Stanley, H. E. and Quattrociocchi, W.
(2016) “The spreading of misinformation online.” Proceedings of the National Academy of Sciences, 113(3),
554–559. doi:10.1073/pnas.1517441113
Dillahunt, T. R., Brooks, C. A. and Gulati, S. (2015) “Detecting and visualizing filter bubbles in Google and
Bing.” CHI EA ’15: Proceedings of the 33rd Annual ACM Conference: Extended Abstracts on Human Factors
in Computing Systems, ACM, New York, NY (pp. 1851–1856).
Dutton, W. H., Reisdorf, B. C., Dubois, E. and Blank, G. (2017, January 5) Search and Politics: The Uses and
Impacts of Search in Britain, France, Germany, Italy, Poland, Spain, and the United States. Quello Center Work-
ing Paper. Retrieved from https://ssrn.com/abstract=2960697
Edelman, B. and Lai, Z. (2016) “Design of search engine services: Channel interdependence in search
engine results.” Journal of Marketing Research, 53(6), 881–900. doi:10.1509/jmr.14.0528
Eslami, M., Karahalios, K., Sandvig, C., Vaccaro, K., Rickman, A., Hamilton, K. and Kirlik, A. (2016) “First
I ‘like’ it, then I hide it: Folk theories of social feeds.” CHI ’16: Proceedings of the 2016 CHI Conference on
Human Factors in Computing Systems, ACM, New York, NY (pp. 2371–2382).
Feuz, M., Fuller, M. and Stalder, F. (2011) “Personal web searching in the age of semantic capitalism: Diag-
nosing the mechanisms of personalisation.” First Monday, 16(2).
Flaxman, S., Goel, S. and Rao, J. M. (2016) “Filter bubbles, echo chambers, and online news consumption.”
Public Opinion Quarterly, 80(S1), 298–320. doi:10.1093/poq/nfw006
Goel, S., Anderson, A., Hofman, J. and Watts, D. J. (2016) “The structural virality of online diffusion.”
Management Science, 62(1), 180–196. doi:10.1287/mnsc.2015.2158
Granka, L. A. (2010) “The politics of search: A decade retrospective.” The Information Society, 26(5), 364–374.
doi:10.1080/01972243.2010.511560
Guo, J., Zhang, P., Zhou, C., Cao, Y. and Guo, L. (2013) “Personalized influence maximization on social
networks.” CIKM ’13 Proceedings of the 22nd ACM International Conference on Information & Knowledge
Management, ACM, New York, NY (pp. 199–208).
Hannak, A., Sapiezynski, P., Molavi Kakhki, A., Krishnamurthy, B., Lazer, D., Mislove, A. and Wilson, C.
(2013, May 13–17) “Measuring personalization of web search.” WWW ’13 Proceedings of the 22nd Inter-
national Conference on World Wide Web, ACM, New York, NY (pp. 527–538).
Helles, R., Ørmen, J., Radil, C. and Jensen, K. B. (2015) “The media landscapes of European audiences.”
International Journal of Communication, 9, 299–320.
Hindman, M. (2009) The Myth of Digital Democracy. Princeton, NJ: Princeton University Press.
Jacobson, S., Myung, E. and Johnson, S. L. (2016) “Open media or echo chamber: The use of links in audi-
ence discussions on the Facebook pages of partisan news organizations.” Information, Communication &
Society, 19(7), 875–891. doi:10.1080/1369118X.2015.1064461
Jones, J. J., Bond, R. M., Bakshy, E., Eckles, D. and Fowler, J. H. (2017) “Social influence and political
mobilization: Further evidence from a randomized experiment in the 2012 U.S. presidential election.”
PLoS ONE, 12(4), e0173851. doi:10.1371/journal.pone.0173851
Kliman-Silver, C., Hannak, A., Lazer, D., Wilson, C. and Mislove, A. (2015) “Location, location, location:
The impact of geolocation on web search personalization.” IMC ’15: Proceedings of the 2015 Internet
Measurement Conference, ACM, New York, NY (pp. 121–127).
141
Jacob Ørmen
Kulkarni, C. and Chi, E. (2013) “All the news that’s fit to read: A study of social annotations for news read-
ing.” Chi’13: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, ACM, New
York, NY (pp. 2407–2416).
Kuzel, A. J. (1999) “Sampling in qualitative inquiry.” B. F. Crabtree and W. L. Miller (eds.), Doing Qualitative
Research. Thousand Oaks, CA: Sage Publications (pp. 31–44).
Larson, R. and Csikszentmihalyi, M. (1983) “The experience sampling method.” New Directions for Method-
ology of Social Behavioral Science, 15, 41–56.
Liu, J., Dolan, P. and Pedersen, E. R. (2010) “Personalized news recommendation based on click behavior.”
IUI ’10: Proceedings of the 15th International Conference on Intelligent User Interfaces, ACM, New York, NY
(pp. 31–40).
McGee, M. (2013) “EdgeRank is dead: Facebook’s news feed algorithm now has close to 100K weight
factors.” Markering Land. Retrieved from http://marketingland.com/edgerank-is-dead-facebooks-news-
feed-algorithm-now-has-close-to-100k-weight-factors-55908
Messing, S. and Westwood, S. J. (2014) “Selective exposure in the age of social media.” Communication
Research, 41(8), 1042–1063. doi:10.1177/0093650212466406
Micarelli, A., Gasparetti, F., Sciarrone, F. and Gauch, S. (2007) “Personalized search on the world wide web.”
In B. Peter, K. Alfred and N. Wolfgang (eds.), The Adaptive Web. Berlin: Springer-Verlag (pp. 195–230).
Nahon, K. (2016) “Where there is social media there is politics.” In A. Bruns, G. Enli, E. Skogerbo, A. O.
Larsson and C. Christensen (eds.), The Routledge Companion to Social Media and Politics. New York, NY
and London: Routledge.
Newman, N., Fletcher, R., Levy, D. A. L. and Nielsen, R. K. (2016) Digital News Report 2016. Reuters Insti-
tute for the Study of Journalism, University of Oxford. Retrieved from http://reutersinstitute.politics.
ox.ac.uk/sites/default/files/Digital-News-Report-2016.pdf?
Ørmen, J. (2015) “Googling the news: Opportunities and challenges in studying news events through
Google search.” Digital Journalism, 4(1), 107–124.
Ørmen, J. and Thorhauge, A. M. (2015) “Smartphone log data in a qualitative perspective.” Mobile Media &
Communication, 3(3), 335–350. doi:10.1177/2050157914565845
Pariser, E. (2012) The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We
Think. New York, NY: Penguin Books.
Powers, E. (2017) “My news feed is filtered?: Awareness of news personalization among college students.”
Digital Journalism, 1–21. doi:10.1080/21670811.2017.1286943
Rader, E. and Gray, R. (2015) “Understanding user beliefs about algorithmic curation in the Facebook news
feed.” Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems, ACM, New
York, NY (pp. 173–182).
Rogers, R. (2013) Digital Methods. Cambridge, MA: The MIT Press.
Ryczko, K., Domurad, A., Buhagiar, N. and Tamblyn, I. (2017) “Hashkat: Large-scale simulations of online
social networks.” Social Network Analysis and Mining, 7(1). doi:10.1007/s13278-017-0424-7
Vaccari, C., Valeriani, A., Barberá, P., Jost, J. T., Nagler, J. and Tucker, J. A. (2016) “Of echo chambers and
contrarian clubs: Exposure to political disagreement among German and Italian users of Twitter.” Social
Media + Society, 2(3). doi:10.1177/2056305116664221
Webster, J. G. (2014) The Marketplace of Attention: How Audiences Take Shape in a Digital Age. Cambridge,
MA and London: The MIT Press.
142
11
DIGITAL NEWS USERS . . . AND
HOW TO FIND THEM
Theoretical and methodological innovations
in news use studies
Ike Picone
When we look at media use practices today, we cannot but witness how these are increasingly
scattered across media, or, more precisely, across different kinds of media contents, different devices
and platforms, and different socio-spatial contexts (Picone, 2016; Picone et al., 2015). The same is
true for news use practices. News users are not limited to the newspaper or cable channel they are
subscribed to but can access a virtually unlimited range of news sources. They do so through an
ever-expanding array of connected devices. After smartphones’ steep rise to prominence as a gate-
way to news and VR headsets’ highly anticipated potential for innovative storytelling, smart speakers
such as the Amazon Echo or Google Home seem the next in line to contend for people’s attention.
Each of these devices comes with specific affordances that inspire novel ways for using news.
For example, ‘news feeds’ on social media incited scrolling and skimming through news, which in
turn led journalists to spice up their news titles to convince news users to click through. Short-form
video, native advertising, clickbait, interactives, infographics, etc., are all examples of how news pro-
ducers try to adapt journalistic products to evolving news use practices. As a result, understanding
news users and the various ways in which they engage with the news has gained importance in
newsrooms in recent years. In an environment where competition for people’s attention increases
while the share of advertising revenues decreases, the companies that best succeed in understanding
the news users are believed to have the competitive advantage over the others.
Of course, circulation numbers, television ratings, and click rates have since long informed edito-
rial decisions, but in the current media landscape, this has intensified due to the affordances of digital
media. Online, news users leave much more precise traces of their doings, which can be monitored in
real time through tools such as Google Analytics, Chartbeat, and Hotjar, to name but a few. Editorial
analytics, “systematic analysis of quantitative data on various aspects of audience behaviour aimed at
growing audiences, increasing engagement, and improving newsroom workflows” (Cherubini and
Nielsen, 2017: 7), are finding their way into newsrooms. As a result, journalists and editors can wit-
ness how their audiences react on news items as they happen, leading to situations where they alter
titles of badly performing articles or boost others amongst relevant target audiences, for example.
Hence, and without ignoring the importance of market logics behind news production, the
way in which people use the news forms an important driver for changes in the field of journal-
ism. Unsurprisingly then, within journalism studies we also find a renewed interest in the way
143
Ike Picone
news use practices are shaping journalism (Picone et al., 2015). This is not to say that the news
audience has not been addressed in journalism studies so far. The public has been central to vari-
ous scholarly debates since the cultural turn in the social sciences that advocated considering how
people give meaning to news in the study of journalism. Important debates revolved around the
distinction between what people ‘need to know’ and what they ‘want to know’ (touching upon
issues of tabloidization and dumbing down vs. sociocultural dimensions of popular audiences) or
on the involvement of the audience in the news-making process (touching upon the democrati-
zation of the news production process) (Picone et al., 2015).
All too often, however, these debates dealt with the question of how journalists envision their
public rather than with the question of how the public experiences journalism. Moreover, these
discussions were informed by what scholars think the audience to be rather than by what they
know it to be. Seldom were reflections on news audiences within journalism studies empirically
grounded. Rather, they were normative in nature, implying how an informed citizenry should
come to be. As Chris Peters (2012: 704) puts it, “we must certainly begin to speak with audi-
ences, as opposed to just about them”, which brings us to the core of this chapter.
The news users in journalism studies

Even more than was the case before, both journalists and journalism scholars have come to
acknowledge that journalism is being shaped by the way audiences use news. News use stud-
ies, both qualitative and quantitative, should then be a fundamental part of journalism scholars’
methodological toolkit, leading to a more empirically informed, less normative conception of
the news audience (Picone et al., 2015). But the current evolutions in people’s news use pose
challenges to the empirical study of it.
In this chapter, I explore the current conceptual and methodological advancements within
news use studies that aim to tackle these challenges. I will first introduce two main challenges
posed by evolving news use practices. Subsequently, I will elaborate on how these challenges
urge us to rethink the way we conceptualize news audiences on the one hand and to adopt new
research methodologies on the other hand. On a conceptual level, we draw from audience studies
and science and technology studies to propose an academically grounded conceptualization of
the news (media) user. On a methodological level, we look for inspiration amongst the (digital)
methods that are being explored in fields such as time use studies and technology adoption studies.
Grasping the news users: challenges ahead

Where journalists have a hard time keeping up with adapting their journalistic output to the
many channels and devices through which news users access the news, user researchers are
equally struggling to grasp the digital news user and the nuances of news use practices. The chal-
lenges can be summarized in this way: (1) how to follow news users across media and (2) how to
encompass the nuances in news use.
How to follow news users across media

Before the advent of digital media, there used to be a tight relationship between a medium and
its content. Audiovisual content, for instance, was related to television and cinema, while written
content was related to newspapers and books. This also implied a relatively concrete focus for
studying the users of these media: studying news viewers meant a focus on viewing news broad-
casts in the living room, studying newspapers meant a focus on someone reading news articles on
the sofa or while commuting (Courtois et al., 2013).
144
Digital news users and how to find them
Now, print and audiovisual content can be consumed through a wide range of other devices
such as laptops, tablet computers, or smartphones, introducing news use in different socio-spatial
contexts. As news practices are scattered across media outlets and devices, so is the researcher’s
focus. Where do you find news readers? Behind a computer on their desk, behind a tablet, a
smartphone, a smartwatch, and soon maybe even a smart refrigerator or table? And on those
devices, they might be reading a digital newspaper, but equally updates from a news app, links
shared on social media, books, etc. In turn, those apps and updates might contain video and audio
fragments as well, hence making news readers viewers and listeners at the same time. So, a first
challenge for news use researchers is following news users as readers, viewers, listeners, etc., across
all the media devices, outlets, and platforms through which they access the news.
How to encompass the nuances in news use

Because of the wide adoption of connected devices, media are now ubiquitous: news is available
24/7, wherever you are. Mark Deuze (2012) argues that media are increasingly so ubiquitous and
pervasive that they become transparent through our daily use of them. A good example is the
way in which many media users come across news stories while checking social media. They are
not purposefully looking for news at that time. News updates are another example: it might liter-
ally take but a glance at the notification center of your smartphone or smartwatch to see news.
Checking the news is then not always a well-defined activity in our daily routine but rather an
action that takes place in those in-between moments of free time when we reach for our mobile
devices in anticipation of our next activity (Dimmick et al., 2011: 25).
How can a researcher interpolate news users about such practices when they are gone before
one even realizes? Checking the news on the go, at a glance, or in between moments is so volatile
that it almost slips one’s consciousness and, indeed, becomes ‘transparent’. As a result, these actions
are not only difficult to observe but also difficult for respondents to recollect, both posing a chal-
lenge for user research. How then to encompass the many nuances in peoples’ everyday news use,
and what level of granularity is required when taking a closer look at news use practices?
In sum, news flows are more intricate than ever, intertwining professional and collaborative infor-
mation production; journalistic, social, and individual news selection; and private and public, seden-
tary and nomadic news use. As a result, grasping the digital news user as he or she flows through
different news settings grows more complex, especially from a methodological point of view.
If we want to empower news use researchers in the future to equip themselves with adequate
means to tackle these challenges, we will need to innovate along both theoretical and method-
ological lines. The methods we use to gather empirical data are intertwined with the conceptual
lens through which we look at new phenomena. This lens focuses on the questions we want to
find an answer to and, in turn, the methods that will allow us to find those answers.
Hence, I will first address these challenges on a theoretical level in the next section, developing
the notion of ‘news users’ in a way that allows us to conceptually grasp people using news across
media. Subsequently, I will explore promising avenues in user research that can inspire method-
ological advancements. Based on this exploration, I will propose a series of guiding principles for
methodological innovation in news use studies.
Theoretical advancement in news use studies: conceptualizing

the digital news user
So far in this chapter, I have been talking about media users, or more precisely news (media)
users. But who are we talking about exactly? In media studies, they have long been known as ‘the
145
Ike Picone
audience’, but “[i]t becomes clear that the range of practices and habits associated with the term
‘audience’ have diversified to the extent that questioning the usefulness of the term audience seems
justified (yet again)” (Madianou, 2009: 343). Both media researchers and professionals have come to
talk about ‘users’ when addressing people in relation to media. This seems to have happened almost
intuitively, as if ‘the user’ simply felt like a better match than ‘the audience’. Carpentier (2011: 113)
explains the popularity of the denominator, at least partially, “by its capacity to emphasize online
audience activity, where people use media technologies and content more actively”.
Of course, the influence of disciplines such as human–computer interaction, internet studies,
and science and technology studies in positioning the notion of user is undeniable. Closer to
media studies, the ‘uses and gratifications’ (U&G) tradition and later the work of Roger Silver-
stone have been essential in introducing the user to audience studies. Numerous studies have
traced everyday media use practices and how they have expanded across media, resulting in a
rich terminology [to] outline changes in patterns of audience practices [that originated]
in a strong urge to develop new conceptual frameworks in order to be able to map,
explain, and understand the complex audience practices which may all, in various ways,
be characterized as ‘cross-media’.
(Bjur et al., 2014: 16)
By intuitively, however, I mean that the adoption of the term ‘user’ within media studies was
not accompanied by an in-depth theoretical and conceptual exploration and argumentation
about why we should address audience members as media users. The adoption of the term ‘user’
and its suitability have been debated in audience studies, but so far without a decisive answer.
Audience researchers themselves remain divided about the matter (Picone, 2017).
I would however suggest that we can think of a theoretical rather than a pragmatic argu-
mentation in favor of ‘news user’ as a concept that allows us to better grasp the digital news user
moving across media. These argumentations revolve around three analytical arguments: the term
‘news user’ being (1) medium-agnostic, allowing users to be understood as they cross from one
device or source to another within their whole media repertoire; (2) scalable, allowing people to
be addressed simultaneously as individuals with very personal ways of using media and as mem-
bers of global audiences; and (3) nonlinear, considering audience members as more than merely
receivers of content. These arguments are strongly informed by and at the same time giving
direction to current approaches to cross-media use. Concepts like ‘media repertoires’, ‘worth-
whileness’, and ‘situational context’ will recur throughout the arguments being made.
‘Media user’ is medium-agnostic

First, ‘media user’ is a medium-agnostic term: It does not require defining people in relation to a
specific type of media device or media practice. Contrary to ‘the reader’ or ‘the viewer’, ‘media user’
considers people turning to a whole repertoire of media sources and devices (Hasebrink and Domeyer,
2012), often simultaneously. For example, data from the 2015 Reuters Institute Digital News Report
shows that already in 2015, a key global trend is that ever more people are using two or three devices to
access the news, with smartphones and tablets specifically “extending our access points [and] making
us more connected to the news at home and on the move” (Newman et al., 2015: 8). The notion of
media user – and not ‘medium’ user for that matter – lets us take this reality into account.
While the study of individual modes of media consumption remains relevant, it becomes
equally important to look at the cross-media combinations that people select and why they do
so (Hasebrink and Domeyer, 2012). At the same time, the notion of media use accommodates
146
under a single term the whole range of media practices that media users can undertake via their
media repertoire, from watching, reading, or listening to sharing, commenting, and posting con-
tent. Employing ‘media user’ permits researchers to address audience members in their capacity
as people performing a whole range of practices through a whole range of devices and sources.
At the same time, the notion can be narrowed down when necessary. ‘News (media) users’, for
example, refers to people consulting the news in different forms over different platforms and
devices. ‘Tablet users’ delineates the users of that specific artifact but still encompasses all the dif-
ferent forms of content people can use it for.
‘Media user’ is scalable

Second, in contrast to, for example, the audience, the user has a singular and plural form, which
makes the concept more scalable. Audience has, of course, also been used in its plural form, espe-
cially in relation to the encoding/decoding approach and how it leads to sub-publics that ‘read’
content in different ways (Hall, 1980); however, it cannot account for the truly individual and per-
sonalized forms that media use can take today. Over a decade ago, Livingstone (2003: 348) argued
that the notion of a user “tends to be overly individualistic and instrumental, losing the sense of a
collectivity which is central to ‘the audience’”. But in a digital media environment characterized
by information abundance and convergence, users increasingly spread their media consumption
across multiple platforms, thus creating their own “personal information space” (Deuze, 2007:
30–33). “[T]he particular constellation of media on which one individual draws may be quite
different than another’s” (Couldry et al., 2007: 190–191). One could even argue that there are
virtually as many media constellations as there are media users (Schrøder and Larsen, 2010).
By talking about the media user we can consider the very individual level of media use, while
the plural form still allows us to refer to the ‘sense of collectivity’ that is traditionally associ-
ated with the notion of audiences. Speaking of media users in plural still permits us to refer to
people as a group, sharing common features or experiences, be it in what sources they consult
(TV series, games, etc.), how (on TV, on a smartphone, etc.), when (in real time, postponed, etc.),
where (at home, commuting, etc.), and with whom (family, friends, social media, etc.), or in the
combination of these elements (e.g. the live television audience of “Britain’s Got Talent” or the
worldwide group of Apple aficionados).
‘Media user’ is nonlinear

Third, the notion of the media user deals with the linear relationship among production, text,
and audience that has been haunting the mass communication tradition for decades. The signi-
fier ‘audience’ both etymologically and historically implies an endpoint in the media infrastruc-
ture, the moment where the message is ‘heard’. ‘User’ does not have this connotation. People can
‘use’ media to consult information; comment on or share it; and produce their own or remix,
redistribute, or in any other form repurpose content, and by doing all this give meaning to the
content and devices adopted. Media technologies are shaped not before but because they are
used by people in their everyday lives (Silverstone, 2006). ‘Media user’ incorporates that idea
and acknowledges the active part that ‘ordinary’ people play in shaping media’s role in society.
Other terms like ‘produser’ (Bruns, 2008) have also been proposed in this regard, but the
advantage of ‘user’ is twofold. First, terms like ‘produser’ imply that media users produce informa-
tion ‘by definition’. From a theoretical perspective, we should surely encompass the ‘productive’
potential of media practices but not at the cost of eclipsing the ‘consumptive’ one. Most people
still mainly engage in merely consuming media. Speaking of the user allows us to include both
147
Ike Picone
dimensions, as somebody can use media to watch a film or to write a review about it. Second,
‘user’ incorporates consumptive and productive dimensions without implying an “analytical col-
lapse between producers and users” often suggested in the literature (Bechmann and Lomborg,
2013: 778). Individual users can engage in various acts of ‘self-publication’ in which they produce
(bits of) information, but looking at the media system from an economic perspective still requires
distinguishing end users from professional producers, as their roles are still distinct when it comes
to value creation.
In sum, I propose to conceptualize news users both as individuals as well as social groups,
actively engaging with news media through a variety of socio-spatial contexts of news use (see
also Picone, 2016), addressing ‘people in relation to media’ in a more encompassing way than
‘audience’ ever could. It gives audience researchers a term that, at its most abstract theoretical
level, is inclusive of the many emanations of their object of study. The term overcomes traditional
tensions in audience studies in a way that turns this intuitively useful word into an analytically
valuable concept. How then can this conceptual base inform the way in which we seek new
methods in news use studies?
Methodological advancements in news use studies: pathways

toward methodological innovation
Starting from the conceptual base already laid out, we can consider a set of guiding principles for
methodological innovation in news use studies that could help researchers follow users across media
and encompass the nuances of news use. We will explore current methodological innovations
beyond the field of journalism studies as an inspiration for future directions in news use studies.
First, if we want to study people’s news experience, we need to take the whole media repertoire
into account. The specific constellation of people’s news diet is constituted by what people per-
ceive to be worthwhile to pay attention to in a specific situation (Schrøder and Kobbernagel, 2010:
116). News competes with various other possible occupations, ranging from skimming through
social media to playing games. But also within one’s news repertoire, various types of news con-
tent and platforms compete for news users’ attention. Focusing on news repertoires forms a more
holistic approach, where the study of news use is embedded into the study of media repertoires.
Second, this variation in the constellation of media repertoires is amplified even further
because media content can be consumed ‘anywhere, anytime’, increasing the importance of situ-
ational factors in shaping people’s media experience. News use studies should consider the socio-
spatial context of news use to answer the question when and where specific news channels are
given a more prominent role within news repertoires. This requires methodologies that consider
where and when news is used and how it is rooted in people’s daily lives.
Adopting these principles offers the advantage of studying news as it is actually rather than
presumably or preferably used. News reaches us through various platforms throughout the day.
Every platform will have intrinsic affordances that appeal to specific persons. And every situa-
tion lends itself to certain practices more than others, affecting how people experience the media
they use in that situation. The more news becomes a cross-platform experience, the more these
affordances will become relative to the affordances of other news media and relative to the situa-
tion they are used in. Previous research on channel repertoires has shown that the availability and
access of media can be more significant predictors of media repertoires than individual demo-
graphics or content preferences (Taneja et al., 2012).
Looking at methodological advancements beyond journalism studies, where can we find
interesting methodologies that adhere to these principles?
148
Time use research

Time use research, although a well-established field of research, offers a first interesting quantita-
tive pathway for methodological innovation. At its core, time use research “describes the alloca-
tion of time among various circumstances and subjective states []. It can provide measures of the
extent, durations, and purposes of access to leisure activities, or of information technology use”
(Gershuny, 2011: 4).
Traditionally, time use research uses diary surveys to register the sequence of activities people
perform during the day: from a list of possible activities, people are asked to select which activi-
ties they did when and for how long during the day, usually complemented by where they did
it, with whom, what accompanying activities they were engaging in, and how they felt about it.
Media use is one of the categories of activity. Recently, researchers have stepped away from
device-focused definitions such as ‘watching TV’ or ‘listening to the radio’ to defining activities
on the level of the sensory experience, i.e., listening, watching, reading, communicating, gaming,
and surfing (see Sonck and Pennekamp, 2014). Each of these broad categories is subdivided into
more specific ones, like listening to a radio program or watching a television series, asking in more
detail about the format of media use. Every time a respondent registers a media activity, he or she
is prompted to answer two consecutive questions: what device was used for the activity and what
kind of content was consulted. As a result, an example of a possible registered activity would be
reading, more specifically reading news, more precisely news headlines (content) via tablet (device).
An important plus of time use research is it being media-agnostic by adopting the activity as
the object of analysis and probing for device and content on a secondary level. Decoupling activ-
ity, content, and device, each of which is registered independently but always in relation to time
and place, allows shifting our focus depending on the questions to be tackled. When it comes
to news, we can get a picture of when and where people consult news, irrespective of the device
they use to do so. Or we can focus on a specific news genre and look at the situations that genre
is consumed in most of the time.
Second, time use research enables researchers to follow individual respondents across the vari-
ous daily situations in which they turn to news media. News use is not bound to occur in a
specific setting. The news can be accessed at home, at work, on the go, etc. Time use research
takes this into consideration, as it does not only probe the time dimension but also simultaneous
activities and where and with whom these occur.
A related advantage is that time use research can be considered “non-media centric” (Krajina
et al., 2014). By looking at time spent, time use research does approach media use in relation
to media users’ other daily activities and routines. It does not only allow positioning news rep-
ertoires in the larger perspective of media repertoires but media use in turn as part of people’s
daily activities and routines. This enables researchers to go beyond the relationship between users
and media to find correlations between media use and other aspects of life like the availability of
leisure time, commuting patterns, etc.
An obvious caveat of time user research, and by extension all forms of survey research, is the
inevitable limitation of self-reporting. Media use research primarily uses data that are gathered, be
it in a quantitative or qualitative way, by asking people to report what they do, think, feel, or expe-
rience. The issues that have been associated with this form of data collection, like retention bias,
socially desirable responding, or flawed self-assessment, are regarded as a necessary evil. Research-
ers have traditionally turned to data triangulation techniques to minimize these shortcomings.
When faced with the challenge of how to encompass the nuances of news use in an information-
saturated, always-connected context, the issue of self-reporting becomes even more pressing. As
we already mentioned, a lot of interactions with the news through mobile or social media are
149
Ike Picone
very volatile, taking place in between our daily activities. Hence, it becomes difficult for news
users to consciously single out when they are interacting with the news. Imagine having to recall
precisely all the news articles, videos, updates, alerts, etc., you consumed over the past few hours,
where, and on which device. Here, a lot is expected from tracking or logging users through
mobile devices, another avenue for methodologic innovation.
Tracking and logging studies

As mentioned, newsrooms are increasingly reliant on editorial analytics, tools that provide them
with unfiltered, real-time data about their readers to inform their editorial decisions. But aca-
demics are also considering the use of tracking tools in various ways, developing specific apps to
enable this (see also Ørmen, this volume, Chapter 10).
One option is integrating tracking options in mobile diaries. Within time user research, for
example, researchers are developing smartphone apps that allow tracking and logging what peo-
ple do on their smartphone. Instead of participants having to report the news sites and apps they
have been using, such a tracking app can register which apps and sites were consulted, what
information was accessed on these apps and sites, when, and for how long, and even, using GPS
data, where participants used which piece of news information. Participants can then still review,
edit or complement that information in the diary app. Also, participants can be prompted with
extra questions once a specific action is registered. Imagine someone checking a news alert being
prompted with an on-screen question ‘Did you find this news alert useful?’.
A digital diary tracking tool offers many advantages: by tracking and logging people’s media
use, you tackle the issue of reporting the nuances of small acts of news consumption. These log
data can be analyzed as part of a quantitative analysis of a large sample of users or can be used to
elicit reflections from participants in a qualitative research setup (see later). Also, the activities reg-
istered are supplemented with accurate and reliable time (including frequency and duration) and
location (via GPS) information. This allows researchers to obtain a precise view on the context
in which news is consulted. These apps have the potential to integrate algorithmic functions that
could identify someone’s specific location, by connecting GPS information to existing databases
such as Google Maps, or translate moving patterns (e.g. moving at high speed) into activities (e.g.
commuting). Furthermore, as health monitoring apps like Apple Health have already shown, the
sensory data that smartphones and smart watches collect can also serve to assess one’s level of
physical activity, sleep, stress, etc. – all data that news use studies could potentially tap into.
These tracking tools hold a large potential for news use studies, but they are not sanctifying, either.
Following user across media remains a challenge, as these apps can track smartphone use but not tele-
vision use or radio use. This is a big concern for audience metric companies such as Nielsen, which
has developed a total audience metric aimed at measuring the viewership of programs combined
over cable TV, on demand services, connected TV devices, mobile, PC, and tablets. Such metrics
require companies to include tracking code in their various offline and online channels. When
Nielsen rolled out its total audience metric at the end of 2016, NBC Universal issued critical remarks
on the service, showing that issues of comparability and implementation of such systems remain.
As our media use continues to shift from analog to digital, we leave behind ever more traces
of our media use. These traces offer an opportunity for news use researchers to grasp the nuances
in people’s digital news use. Even if often still very costly, they can develop their own tools for
this purpose, a field where much room for improvement through interdisciplinary collaboration
is still available. A larger issue, however, is one of openness or academic researchers’ access to pri-
vate data. A lot of traces of digital (news) media consumption are already being tracked by news
companies on their news channels and by intermediaries such as internet/cable/connected TV/
150
social media providers on their platforms. Because of the commercial value of these data, players
are reluctant to share them with researchers – or with any other player. Hence, a big challenge
remains for news use researchers to collaborate with private companies for the analysis of these
data. However, these data are often so sizeable and complex that even if granted access, the ques-
tion for social scientists remains how to analyze these data in any meaningful way.
Connected and semi-experimental ethnography

When our news use becomes highly traceable, it is tempting to assume that we might obtain
a complete understanding of it by analyzing quantitative data. But an equally interesting path-
way for methodological innovation can be pursued along the lines of qualitative, ethnographic
approaches to the study of news use. Inductive, interpretative approaches remain valuable.
Media ethnography, too, is confronted with the issues already touched upon. Historically,
ethnography takes bounded spaces or delineated situations as its methodological starting point.
But how can media ethnographers define the locus of news use practices when these are scattered
over an ever-expanding and more mobile digital media repertoire accessible anytime, anywhere?
News use can take place on different crossroads between public and private, virtual, and physi-
cal spheres, meaning that the media ethnographer would need to be present at, or at least have
access to, these crossroads, which is not self-evident. Participant observation seems difficult, as it
would imply that participants would be virtually stalked by the researchers as they consult media
at home, on the train, at work and even in between times. Both in terms of the researchers’ effort
and of invasiveness into the participant’s life, this is obviously not feasible. How to proceed with
media ethnography, then?
Part of the answer is offered by danah boyd (2015) when she proposes finding different entry
points into a phenomenon by following relations between people and practices by envisioning
relations between persons, spheres, and objects instead of approaching those elements in isolation
from each other. She has further developed this approach for the study of technology-mediated
everyday practices in what she calls a ‘connected field sites’ protocol. Boyd’s protocol (2015)
includes participant-observations of online and offline practices, ‘deep hanging out’ in physical
spaces where the practices occur, and semi-structured face-to-face interviews. This offers a good
starting point for the ethnographic study of news use, scattered and volatile as it is today.
To further address the issue of news use being invisible to the observer’s eye, a semi-experimental
approach can be considered. For example, tracking tools and (online) diaries can be introduced in
an ethnographic setting to follow people moving through private spheres or to grasp ephemeral
moments of news consumption. Here, tracking tools serve not so much to gather data to be ana-
lyzed as such but rather data that can be used as part of elicitation techniques during in-depth inter-
views. Also, setting up a tracking procedure does not always require a full-fledged online diary app.
Thinking creatively, a lot can already be achieved by querying the APIs of social networks through
tools such as IFTTT (if this then that).
For example, if you would want to track the links that a participant shares on Facebook, it
would suffice to ask them to connect their account to IFTTT for you to obtain a continuously
updated Google sheet containing the links and their publication time. Researchers can tap into
these data to surface specific moments or actions during the interview for participants to reflect
upon. For example, confronting a participant’s view on her news use with her actual behavior as
reflected in track logs can foster interesting reflections from her side during the interview.
In certain circumstances, researchers can go even further in adopting an experimental
approach. As already mentioned, while journalism is still getting accustomed to the steep rise of
mobile devices, the next wave of voice-driven, home-based smart assistants is already on its way.
151
Ike Picone
Especially from a more applied research perspective, questions arise about the adoption of these
kinds of technologies for the use of news. Often, researchers turn to lead users to get a sense of
how new technologies might be adopted and how new use practices might develop. An impor-
tant disadvantage of such lead-user research in terms of news use is that these are often tech-savvy,
news-oriented persons whose ways of using news are hardy predictive for regular news users’
engagement with the news. Think of bloggers, for example, and how the large majority of news
users left blogging aside.
Here, living lab methodology (Schuurman et al., 2011) and proxy technology assessments (Lievens
et al., 2008) offer interesting ways forward. These methodologies introduce new technological
artifacts or services in real-life settings. Their uses are then studied through qualitative, digi-
tal data-gathering methods. Living labs can be described as “environments for innovation and
development where users are exposed to new ICT solutions in (semi)realistic contexts, as part
of medium- or long- term studies targeting evaluation of new ICT solutions and discovery of
innovation opportunities” (Følstad, 2008: 116). In proxy technology assessments, the technology
introduced is conceived as a proxy for a possible future technology. For example, you are won-
dering how news users would engage with an all-you-can-eat online news subscription. In the
absence of such service, you could present participants with an account on Blendle, a pay-per-
article service, which is then linked to a pay card from the research project.
Digital news users . . . and how to find them

News is ‘anytime, everywhere’, and so are news users. On mobile devices and social media
platforms, important parts of people’s news consumption become unintentional and volatile,
even if, ironically, it can be traced in detail. In such a context, finding and following news
users and understanding all the nuances of news use is challenging. In this chapter, I have
explored both conceptual and methodological ways to tackle this challenge. Rather than
pledging to offer ready-made solutions, I propose guiding principles and possible pathways
along which news use researchers can further develop suitable methodologies in a construc-
tive and creative way.
When putting people at the center and following them across different contexts and devices
and consulting various forms of content, it is difficult to hold on to the notion of audience. The
term ‘media user’ is more versatile and allows people to be addressed in their different capacities
as watchers, readers, listeners, communicators, commenters, posters, sharers, etc., simultaneously.
The successful new methods for studying active news use will be those that can grasp news users
in as many of their capacities as possible.
Further readings
This chapter draws on and synthesizes some of the ideas developed in previous articles. I address
conceptual and methodological advancements in news use studies in the article “Grasping the
Digital News User. Conceptual and Methodological Advances in News Use Studies” (2016).
When it comes to the specific strands of methodological innovation, a few works offer interest-
ing starting points for further readings. The article of Joeri Minnen et al. (2014) Modular Online
Time Use Survey (MOTUS) – Translating an Existing Method in the 21st Century provides insights
into time use research; Jakob Ohme et al.’s (2016) article “Exposure Research Going Mobile: A
Smartphone-Based Measurement of Media Exposure to Political Information in a Convergent
Media Environment” gives an overview of tracking and logging studies; danah boyd develops the
notion of connected ethnography in the book chapter “Making Sense of Teen Life: Strategies
152
for Capturing Ethnographic Data in a Networked Era”, whereas I explore the idea of semi-
experimental ethnography in the article “Situating Liquid Media Use: Challenges for Media
Ethnography” (2013).
References
Bechmann, A. and Lomborg, S. (2013) “Mapping actor roles in social media: Different perspectives on value
creation in theories of user participation.” New Media & Society, 15(5), 765–781.
Bjur, J., Schrøder, K., Hasebrink, U., Courtois, C., Adoni, H. and Nossek, H. (2014) “Cross-media use:
Unfolding complexities in contemporary audiencehood.” In N. Carpentier, K. Schrøder and L. Hal-
lett (eds.), Audience Transformations: Shifting Audience Positions in Late Modernity. London: Routledge (pp.
15–29).
boyd, d. (2015) “Making sense of teen life: Strategies for capturing ethnographic data in a networked era.”
In E. Hargittai and C. Sandvig (eds.), Digital Research Confidential: The Secrets of Studying Behavior Online.
Cambridge, MA: The MIT Press (pp. 79–102).
Bruns, A. (2008) Blogs, Wikipedia, Second Life, and Beyond: From Production to Produsage. New York, NY: Peter
Lang.
Carpentier, N. (2011) “New configurations of the audience? The challenges of user-generated content
for audience theory and media participation.” In V. Nightingale (ed.), The Handbook of Media Audiences.
Malden, MA: Blackwell Publishing (pp. 190–212).
Cherubini, F. and Nielsen, R. K. (2017) Editorial Analytics: How News Media Are Developing and Using Audi-
ence Data and Metrics. Research Report. Reuters Institute for the Study of Journalism, Oxford.
Couldry, N., Livingstone, S. and Markham, T. (2007) Media Consumption and Public Engagement: Beyond the
Presumption of Attention. Basingstoke: Palgrave Macmillan.
Courtois, C., Schrøder, K. C. and Kobbernagel, C. (2013) “Exploring landscapes of news consumption
cross-nationally.” In F. Zeller, C. Ponte and B. O’Neill (eds.), Revitalising Audience Research: Innovations in
European Audience Research. Abingdon: Routledge (pp. 123–142).
Deuze, M. (2007) Media Work. Cambridge: Polity Press.
Deuze, M. (2012) Media Life. Cambridge: Wiley.
Dimmick, J., Feaster, J. C. and Hoplamazian, G. J. (2011) “News in the interstices: The niches of mobile
media in space and time.” New Media & Society, 13(1), 23–39.
Følstad, A. (2008) “Living labs for innovation and development of information and communication tech-
nology: A literature review.” The Electronic Journal for Virtual Organizations and Networks, 10, 99–131.
Gershuny, J. (2011) Time-Use Surveys and the Measurement of National Well-Being (Research report). Oxford:
Centre for Time Use Research, University of Oxford.
Hall, S. (1980) “Encoding/decoding.” In S. Hall, D. Hobson, A. Lowe and P. Willis (eds.), Culture, Media,
Language. London: Hutchinson (pp. 128–138).
Hasebrink, U. and Domeyer, H. (2012) “Media repertoires as patterns of behaviour and as meaningful prac-
tices: A multimethod approach to media use in converging media environments.” Participations: Journal
of Audience and Reception Studies, 9(2), 757–779.
Krajina, Z., Moores, S. and Morley, D. (2014) “Non-media-centric media studies: A cross-generational
conversation.” European Journal of Cultural Studies, 17(6), 682–700.
Lievens, B., Torben Nielsen, K., Pierson, J. and Jacobs, A. (2008) “Proxy technology assessment as a multi-
method approach for identifying social requirements of co-creative media.” Presented at the NORDICHI
2008 Conference, 20–22 October.
Livingstone, S. (2003) “The changing nature of audiences: From the mass audience to the interactive media
user.” In A. Valdivia (ed.), Companion to Media Studies. Oxford: Blackwell Publishing (pp. 337–359).
Madianou, M. (2009) “Audience reception and news in everyday life.” In K. Wahl-Jorgensen and T.
Hanitzsch (eds.), The Handbook of Journalism Studies. New York, NY: Routledge (pp. 325–337).
Minnen, J., Glorieux, I., van Tienoven, T. P., Daniels, S., Weenas, D., Deyaert, J., . . . Rymenants, S. (2014)
“Modular Online Time Use Survey (MOTUS) – Translating an existing method in the 21st century.”
International Journal of Time Use Research, 11(1), 73–93.
Newman, N., Levy, D. A. L. and Nielsen, R. K. (2015) Reuters Institute Digital News Report 2015: Tracking the
Future of News (Research report). Oxford: Reuters Institute for the Study of Journalism.
Ohme, J., Albaek, E. and de Vreese, C. H. (2016) “Exposure research going mobile: A smartphone-based
measurement of media exposure to political information in a convergent media environment.” Commu-
nication Methods and Measures, 10(2–3), 135–148.
153
Ike Picone
Peters, C. (2012) “Journalism to go.” Journalism Studies, 13(5–6), 695–705.

Picone, I. (2013) “Situating liquid media use: Challenges for media ethnography.” Westminster Papers in Com-
munication and Culture, 9(3), 49–70.
Picone, I. (2016) “Grasping the digital news user. Conceptual and methodological advances in news use
studies.” Digital Journalism, 4(1), 125–141.
Picone, I. (2017) “Conceptualising media users across media. The case for ‘media user/use’ as analytical
concepts.” Convergence: The International Journal of Research Into New Media Technologies, 23(4), 378–390.
Picone, I., Courtois, C. and Paulussen, S. (2015) “When news is everywhere. Understanding participation,
cross-mediality and mobility from a radical user perspective.” Journalism Practice, 9(1), 35–49.
Schrøder, K. C. and Kobbernagel, C. (2010) “Towards a typology of cross-media news consumption: A
qualitative-quantitative synthesis.” In Northern Lights: Film & Media Studies Yearbook (Vol. 8). London:
Intellect (pp. 115–137).
Schrøder, K. C. and Larsen, B. S. (2010) “The shifting cross-media news landscape.” Journalism Studies, 11,
524–534.
Schuurman, D., De Moor, K., De Marez, L. and Evens, T. (2011) “A living lab research approach for mobile
TV.” Telematics and Informatics, 28(4), 271–282.
Silverstone, R. (2006) “Domesticating domestication. Reflections on the life of a concept.” In T. Berker, M.
Hartmann,Y. Punie and K. Ward (eds.), Domestication of Media and Technology. Berkshire: Open University
Press (pp. 229–248).
Sonck, N. and Pennekamp, S. (2014) Media:tijd 2014. Research Report. Sociaal Cultureel Planbureau, Den
Haag.
Taneja, H., Webster, J. G., Malthouse, E. C. and Ksiazek, T. B. (2012) “Media consumption across platforms:
Identifying user-defined repertoires.” New Media & Society, 14(6), 951–968.
154
PART III
The political economy

of digital journalism
12
WHAT IF THE FUTURE IS NOT
ALL DIGITAL?
Trends in U.S. newspapers’ multiplatform
readership
The web has been publicly available for more than 20 years, and U.S. newspapers’ online experi-
ment has been going on for nearly as long. Motivated initially by high hopes for the potential of
the internet, newspaper firms have expended substantial resources digitizing and distributing their
content through a number of channels. Underlying newspapers’ sustained enthusiasm for a digital
future is a long-time (but unchecked) assumption that print newspapers are dying (Picard and
Brody, 1997; Penenberg, 2004; Chyi et al., 2012). The recent recession, which accelerated declines
in print circulation and advertising revenue, reinforced such a belief. It is against this backdrop
that U.S. newspapers, especially national and metro dailies, became more determined than ever to
complete their transition from print to online. They slashed resources for the print edition, laid off
print staff, and reduced printing and delivery schedules while expanding digital operations, hiring
multimedia reporters and social media editors, and requiring journalists to reinvent themselves
digitally. “Digital first” (Paton, 2010) has become a mantra, a goal, and a path leading to the future.
However, amid early excitement and recent determination about an all-digital future for news,
a long-time problem facing a vast majority of newspaper firms persists – digital revenue remains
insufficient to cover the loss on print revenue (State of the News Media 2015, 2015). Industry
discourse often focuses on the loss on the print side without critically examining the perfor-
mance of the digital products by framing the problem as “newspapers have not yet figured out
how to monetize online news users,” implying that a solution will emerge in the future (Chyi,
2013). However, to date, the performance of newspapers’ digital products in terms of advertising
and subscription revenue has remained underwhelming. The result of newspaper firms’ transi-
tion from print to online is, in the business sense, “exchanging analog dollars for digital dimes”
(quoted in Dick, 2009: para. 1).
Economic reasoning suggests that at least part of the difficulty in monetizing digital content
lies in readership, which is the foundation for subscription and advertising revenue. Therefore,
this chapter presents a comparative analysis of 51 U.S. newspapers’ online and print readership
data collected in 2007, 2011, and 2015 (Chyi and Tenenboim, 2017). The readership trends
challenged conventional wisdom and triggered a debate over U.S. newspaper firms’ technology-
driven strategy. This chapter also addresses the implications of the debate.
157
Twenty years into their experiments with the web, it is time to scrutinize demand for newspa-
pers’ digital offerings, revisit newspaper firms’ technology-driven strategy, and critically examine
unchecked assumptions about the future of newspapers.
Growth of online news access

Online news use, like all things digital, has been growing. In 1995, roughly 2% of American adults
reported getting news online three days a week or more. By 2010, about half (46%) did so. That
was the year remembered as when “Web tops newspapers as news source for first time” (Olivarez-
Giles, 2011), or “Online news readership overtakes newspapers” (Choney, 2011), and the State of
the News Media announced that “[f]or the first time [. . .] more people said they got news from
the web than newspapers” (Rosenstiel and Mitchell, 2011: para. 11). Such statements are often
interpreted by newspaper managers as a sign of encouragement for their digital endeavors.
What is often neglected, however, is when users seek news online, many flock to online news
aggregators such as Yahoo News, as opposed to newspaper sites. For example, the Pew Center
reported that the most frequently visited news sites in 2006 were MSNBC (31%), Yahoo (23%),
CNN.com (23%), Google (9%), AOL (8%), and FoxNews.com (8%) – none of which was affili-
ated with a newspaper. The most visited newspaper sites were NYTimes.com and USAToday.
com, each mentioned by 5% of online news users (Pew Research Center for the People & the
Press, 2006). In 2008, Yahoo became the leading news site and has remained the top online news
destination at both national (Pew Research Center for the People & the Press, 2012; State of the
News Media 2015, 2015) and local levels (Chyi and Lewis, 2009).
In recent years, social networking sites have become major sources for news. In 2016, 44%
of U.S. adults got news from Facebook (Gottfried and Shearer, 2016). With its unprecedented
power for content distribution, Facebook has been accused of “swallowing journalism” (Bell,
2016: para. 1). But the dominance of news aggregators and social media has not stopped U.S.
newspapers from pursuing their digital dreams.
Transition from print to online

A major driver behind the industry-wide digital experiment is the ‘newspapers are dying’ assump-
tion. Indeed, newspaper penetration has declined steadily at 1–2% each year since 1950 (Picard
and Brody, 1997; Picard, 2008), and print circulation has been declining since 1987 (weekday)
and 1993 (Sunday) despite a growing population (Newspaper Association of America, 2012). In
2008, the recession quickened declines in print circulation and advertising revenue, providing
further support for the ‘newspapers are dying’ narrative. Driven primarily by fear and uncertain-
ties, newspaper firms addressed their financial woes by slashing resources for their print editions
and continued their incomplete transition online.
The number of full-time professional editorial employees declined from 55,000 in 2007
to 32,900 in 2015 (ASNE, 2016) – the reduction occurred mostly on the print side. Many
newspapers, including the New York Times, reduced pages or dropped sections from the paper
(Pérez-Peña, 2008); others, such as Detroit Free Press, the Detroit News, and the Times-Picayune
(New Orleans), eliminated home delivery on certain days of the week. To reduce reliance
on print advertising, many implemented significant price hikes on their print product (Case,
2009).
Despite dwindled resources, newspapers continued expanding their digital offering, managing a
cross-media product portfolio that includes content for print, web, e-readers, smartphones, tablets,
and social media platforms, and emphasizing multimedia content such as photos, audio, and video.
158
What if the future is not all digital?
Digital revenue: advertising and subscription

Despite all the efforts to make ‘digital’ work, most newspapers have not been able to generate
sufficient digital revenue to cover the loss on the print side. From 2007 to 2014, U.S. newspapers’
print advertising revenue dropped from $42.2 billion to $16.4 billion, while digital ad revenue
increased from $3.2 billion to $3.5 billion during the same period. Despite substantial declines,
the dead-tree edition still generates about 80% of total ad revenue (State of the News Media 2015,
2015). The Newspaper Association of America (NAA) stopped releasing industry-wide revenue
data after 2014, but the latest development is that three out of five publicly traded newspaper
firms reported declines in their digital ad revenue in 2015 – 6% for The Tribune Publishing, 5%
for Gannett, and 1% for A.H. Belo (Barthel, 2016).
Another revenue source is online subscription. After offering online news for free for over
a decade, the industry implemented ‘paywalls’ in an attempt to boost online revenue. In March
2011, The Dallas Morning News started charging $16.95 a month for a digital package that includes
web and apps, known as a “hard paywall” (Doctor, 2011). A few weeks later, the New York Times
implemented a metered model, or a “soft paywall”, requiring online users who view more than
20 articles (reduced to 10 in April 2012) per month to pay $15, $20, or $35 a month (Sulzberger,
2011). By 2016, 78% of U.S. newspapers with circulations over 50,000 have adopted the digital
subscription model (Williams, 2016).
Despite the prevalence of digital paywalls among newspaper sites, most publishers do not
reveal digital subscription revenue (Myllylahti, 2013). According to the NAA (2013), digital-
only circulation revenue accounted for only 1% of total circulation revenue in 2012. Anecdotal
information also suggests that the number of digital subscribers for non-national newspaper sites
is underwhelming. For example, merely 65,000 digital-only subscribers signed up for Gannett’s
81 local dailies as of June 2013 (Mutter, 2013). Some papers (e.g., The Dallas Morning News and
San Francisco Chronicle) have dropped their paywalls.
In sum, despite the industry’s focus on digital, U.S. newspapers’ online offerings generate only
limited advertising and subscription revenue. In contrast, their weakening print product remains
the primary revenue driver. This raises questions about the size and composition of newspapers’
online (and print) readership, because a healthy audience base should bring in healthy subscrip-
tion and advertising revenue. To tackle the mystery behind ‘analog dollars’ and ‘digital dimes’,
major newspapers’ online and print reader base warrants scrutiny.
Audience measurement for newspapers

Audience measurement is of paramount importance because it carries profound, sometimes
unappreciated, consequences for the industry and ultimately can provoke change to content itself
(Webster, 2014). For newspapers, circulation is the most-watched indicator of demand for the
print product, and circulation decline is often cited as evidence supporting ‘the internet is kill-
ing newspapers’ narrative. However, the Alliance for Audited Media (AAM), formerly the Audit
Bureau of Circulations (ABC), has in recent years changed how circulation is measured, making
year-to-year comparisons impossible (Mutter, 2010; State of the News Media 2015, 2015).
On the digital side, the newspaper industry often publicizes ‘unique visitors’ figures. Research
firms (e.g., Nielsen/NetRatings and comScore Media Metrix) collect such data through track-
ing software installed on a national panel of internet users who chose to participate. Publishers
also collect their own ‘unique visitors’ data through Google Analytics. Many factors influence the
accuracy of such data, even when major research providers are involved (Gordon, 2007; Thurman,
2017). The number of unique visitors for newspaper sites often seems sizable because data are usu-
ally reported monthly (whereas print circulation is reported daily) (see Thurman et al., this volume,
159
Chapter 13). In addition, this metric as an aggregate measure does not distinguish repeated users
from one-time visitors (e.g., those coming from search engines or social networking sites). The
2015 State of the News Media report used such data to generate the list of the top 50 news sites and
acknowledged that many of their unique visitors were flybys as opposed to regular visitors (State of
the News Media 2015, 2015). Nevertheless, inflated audience size serves as a great sales pitch, which
explains the popularity of this particular measure in industry discourse (Zheng et al., 2012).
Another indicator of newspaper demand is ‘readership’. Print readership is the number of
adults who have read or looked through a print newspaper during a specific time frame (e.g., one
week or one month); online readership is the number of adults who have visited a newspaper site
(Newspaper Association of America, 2015).
Scarborough Research (now Nielsen Scarborough), a long-time provider of media use and
consumer information in the US, collects readership data (for print and online editions) through
random-sample telephone interviews of adults age 18 or older1 in a large number of designated
market areas (DMAs).2 Although such data are collected in the local market and thus do not
include out-of-market readership, this metric allows for side-by-side comparisons between a
newspaper’s print and online demand in its home market. Therefore, through analyzing reader-
ship data, user demand for major U.S. newspapers’ print and digital offerings can be assessed.
Demand for multiplatform newspapers

For multiplatform media firms, understanding user demand for different product offerings helps
identify which product is the “cash cow” and which is the “problem child” (Picard, 2003, 2005).
Through enumerating demand for individual newspapers’ online and print products in a well-defined
geographic market, a number of studies have identified consistent patterns characterizing consumer
demand for online and print products under one newspaper brand (Chyi and Lasorsa, 1999, 2002;
Chyi, 2006; Chyi and Huang, 2011; Hargrove et al., 2011). These patterns were theorized as:
Online Edition Readers Print Edition Readers Hybrid Readers

   50%
GP or WU GP or WU Online Edition Readers
where GP = general public

where WU = web users
where hybrid readers are online edition readers who also read the print edition
These relationships suggest that, within the local market, (1) a local newspaper’s print penetra-
tion is higher than its online penetration – that is, print readers outnumber online readers;
(2) compared with the general public, readers of the online edition are more, not less, likely to
read the same newspaper’s print edition; and (3) the majority of a newspaper’s online readers also
read its print edition – that is, hybrid readers outnumber online-only readers.
Drawing on these research findings, this study (Chyi and Tenenboim, 2017) examines user
demand for major U.S. newspapers’ online and print editions. The first set of research questions
addresses the size of print and online readership in the local market over time.
RQ1: Which product – print or online – reaches more readers in a newspaper’s DMA market?
RQ2: To what extent has print readership changed over time?
RQ3: To what extent has online readership changed over time?
Within a local market, residents have access to the local newspaper in print and online for-
mats. Among local readers, one may further distinguish (1) print-only readers, (2) online-only
160
readers, and (3) hybrid readers, who access their local paper in both print and online formats.
Print-only readers are traditionalists. Online-only readers represent an audience segment con-
tributed uniquely by digital products, while hybrid readers are those who read the online edition
but still hang on to the legacy product.
RQ4: What is the composition of the net combined readership – print-only, hybrid, and
online-only? To what extent has the composition changed over time?
There used to be two schools of thought regarding how a newspaper’s digital edition relates to
its legacy product – cannibalization or complementation. Several studies (Chyi and Lasorsa, 1999,
2002; Chyi, 2006; Chyi and Huang, 2011; Hargrove et al., 2011) revealed that, compared with the
general public, readers of the online edition were more, not less, likely to read the same newspaper’s
print edition. In addition, when the newspaper was the leading one in the market, it was found
that the majority of readers of the online edition still hang on to its print counterpart – in other
words, hybrid readers outnumbered online-only readers.
RQ5: Compared with the general public, are readers of the online edition more likely to read
the same newspaper’s print edition? Do hybrid readers outnumber online-only readers?
Finally, a major driver behind newspaper firms’ digital transformation is pursuing young
readers (Graybeal, 2011). Twenty years into the experiment, this study seeks to examine the
effectiveness of U.S. newspapers’ digital strategy in attracting young readers. Compared with
print readership, to what extent is online readership younger? Which format – print or online –
reaches more readers within each age group?
RQ6: What is the online and print reach among different age groups?
Empirical analysis
Sample: According to the 2015 Editor & Publisher Newspaper Databook, 1,331 daily newspapers
currently operate in the United States (Editor & Publisher, 2015; Newspaper Databook, 2015). To
include major daily newspapers in the analysis, the researchers selected all newspapers with daily
(M–F) circulations of more than 120,000. A total of 64 newspapers reached the threshold, of
which 13 were excluded from the sample because their readership data were not completely
available in the AAM-Nielsen Local audience summary database. The final sample consisted of
51 metro daily newspapers, with daily circulation ranging from 120,473 to 793,582.
Data sources: Print and online newspaper readership data collected in 2007 and 2011 by Scar-
borough were retrieved from the AAM’s “Audience Snapshot+” tool.3 These data were released
on September 30 of the respective year, covering the six preceding months. The 2015 data were
retrieved from the AAM Media Intelligence Center,4 covering Q3 of 2015. Data retrieved were
seven-day in-market (1) print readership, (2) online readership, and (3) net combined readership,
as well as percentages of reach for each of the three.5 The earliest data available in the databases
were collected in 2007, and the latest were from 2015, with 2011 being the midpoint of the
two. Seven-day print and online readership data among five different age groups (18–24, 25–34,
35–44, 45–54, and 55+) were retrieved from the AAM-Nielsen Local audience summary data-
base (or “Audience Snapshot”).6 These data were released by Scarborough in 2015.
Data analysis: To address RQ1, comparative analysis was conducted using the 2015 online and
print readership data across all 51 newspapers under study. To address RQ2 and RQ3, the analysis
tracked readership data collected in 2007, 2011, and 2015 for print and online, respectively. To address
161
RQ4, print-only readership was calculated by subtracting online readership from combined reader-
ship. Online-only readership was calculated by subtracting print readership from combined reader-
ship. Hybrid (both print and online) readership was calculated by subtracting print-only readership
from print readership. The analysis tracked the changes in composition using data collected in 2007,
2011, and 2015. To address RQ5, online readers’ propensity of reading the print edition is calculated
by dividing the number of hybrid readers by the number of total online edition readers (hybrid plus
online-only). The general public’s propensity of reading the print edition is simply print reach. To
address RQ6, online and print readership data among five different age groups were compared.
Findings
RQ1 asked which product – print or online – reaches more readers in a newspaper’s DMA market.
Table 12.1 presents 2015 readership data of the 51 newspapers under study – in terms of percent
Table 12.1 In-market print and online reach, 2007–2015
State Newspaper Print reach Online reach
2007 2011 2015 2007 2011 2015
Arizona The Arizona Republic 52% 44% 33% 18% 17% 16%
Arkansas Arkansas Democrat-Gazette 63% 54% 45% 6% 6% 6%
California Los Angeles Times 34% 29% 24% 6% 8% 11%
California San Francisco Chronicle 32% 24% 19% 10% 14% 13%
California San Jose Mercury News n/a 39% 32% n/a 12% 11%
California The Orange County Register 11% 9% 8% 2% 4% 3%
California The Press-Enterprise 6% 5% 4% 1% 1% 1%
California The Sacramento Bee 36% 30% 24% 8% 9% 8%
California U-T San Diego 55% 44% 37% 17% 16% 11%
Colorado The Denver Post 51% n/a 31% 7% n/a 10%
Connecticut The Hartford Courant 41% 31% 28% 11% 8% 8%
Florida Orlando Sentinel 45% 37% 28% 9% 10% 9%
Florida South Florida Sun-Sentinel 29% 23% 19% 7% 8% 6%
Florida Tampa Bay Times 34% 34% 30% 6% 6% 5%
Florida The Miami Herald 39% 31% 26% 8% 9% 8%
Florida The News-Press 40% 32% 27% 8% 7% 6%
Florida The Tampa Tribune 30% 25% 20% 10% 11% 6%
Georgia Atlanta Journal-Constitution 47% 35% 25% 18% 17% 15%
Hawaii Honolulu Star-Advertiser 62% 53% 50% 11% 16% 11%
Illinois Chicago Sun-Times n/a 42% 23% n/a 7% 6%
Illinois Chicago Tribune 46% 34% 33% 10% 10% 12%
Indiana The Indianapolis Star 49% 40% 33% 13% 12% 11%
Kentucky The Courier-Journal 64% 56% 46% 10% 12% 9%
Massachusetts The Boston Globe 40% 29% 25% 16% 20% 16%
Michigan Detroit Free Press 49% 40% 36% 10% 12% 14%
Minnesota St. Paul Pioneer Press 28% 24% 20% 6% 5% 5%
Minnesota Star Tribune 53% 45% 42% 12% 13% 13%
Missouri St. Louis Post-Dispatch 57% 43% 38% 13% 14% 14%
Missouri The Kansas City Star 60% 51% 39% 13% 12% 11%
Nevada Las Vegas Review-Journal n/a 45% 36% n/a 11% 11%
New Jersey The Star-Ledger 12% 10% 7% 3% 5% 6%
New York New York Daily News 29% 26% 20% 3% 3% 5%
162
2007 2011 2015 2007 2011 2015
New York Newsday 15% 13% 10% 3% 4% 4%

North Carolina The Charlotte Observer 46% 39% 29% 10% 12% 10%
North Carolina The News & Observer 37% 32% 24% 10% 8% 8%
Ohio The Columbus Dispatch 60% 52% 38% 9% 12% 10%
Ohio The Plain Dealer 47% 40% 31% 8% 13% 11%
Oregon The Oregonian 54% 47% 32% 10% 13% 18%
Pennsylvania Pittsburgh Post-Gazette n/a 40% 28% n/a 8% 9%
Pennsylvania The Philadelphia Inquirer 36% 27% 22% 5% 7% 7%
Pennsylvania Tribune-Review 29% 31% 32% 5% 6% 8%
Texas Austin American-Statesman 59% 45% 40% 21% 18% 16%
Texas Fort Worth Star-Telegram 23% 18% 15% 4% 6% 7%
Texas Houston Chronicle 52% 43% 32% 10% 14% 10%
Texas San Antonio Express-News 63% 47% 40% 16% 15% 14%
Texas The Dallas Morning News 41% 31% 25% 8% 9% 8%
Utah Deseret News 20% n/a 16% 7% n/a 11%
Virginia The Virginian-Pilot 53% 44% 37% 15% 15% 14%
Washington The Seattle Times 44% 34% 31% 14% 16% 13%
Washington DC The Washington Post 61% 54% 41% 21% 21% 21%
Wisconsin Milwaukee Journal Sentinel 65% 56% 42% 15% 16% 15%
Average 42.4% 35.9% 28.8% 9.8% 10.7% 10.0%
reach. In each and every one of the 51 cases, print readership outnumbers online readership by a
wide margin. On average, the print edition reaches 28.8% of local adults; the online edition reaches
10%; and the print edition is the most read product in all these major newspapers’ home markets.
RQ2 asked to what extent print readership has changed over time. Table 12.1 presents the
percent reach of the print edition for each newspaper in 2007, 2011, and 2015. Print reader-
ship declined through the years, which is in line with the well-reported circulation trends. The
average reach dropped from 42.4% in 2007, to 35.9% in 2011, to 28.8% in 2015. The decline
amounts to 13.6 percentage points over eight years.
RQ3 asked to what extent online readership has changed over time. Table 12.1 presents the
percent reach of the online edition for each newspaper in 2007, 2011, and 2015. Surprisingly,
online readership has shown little to no growth during the period. The average reach increased
slightly from 9.8% in 2007 to 10.7% in 2011 and then decreased slightly to 10.0% in 2015.
Regarding more recent changes in online readership (2011–2015), as many as 32 newspapers
experienced a decrease (ranging from 0.1 to 5.6 percentage points); only 17 newspapers achieved
an increase (0.1 to 4.4 percentage points).
RQ4 asked about the composition of the net combined readership – print-only, hybrid, and
online-only. In 2015, the combined readership consisted of: 23.3% print-only, 5.5% hybrid
(both print and online), and 4.5% online-only. Over time, print-only readership declined (from
35.4% in 2007 to 23.3% in 2015), and so did hybrid readership (from 7.1% to 5.5%). Online-
only increased from 2.7% to 4.5%. Figure 12.1 illustrates the composition of combined read-
ership in 2007, 2011, and 2015. Despite substantial declines, print-only remains the primary
readership segment. Hybrid remains the secondary, and online-only remains a fraction of overall
readership.
163
Figure 12.1 Composition of in-market combined readership, 2007–2015
RQ5 asked whether readers of the online edition, compared with the general public, are more
likely to read the same newspaper’s print edition. Results showed that, across all 51 newspapers,
without an exception, online edition readers’ propensity of reading the print edition is higher
than the general public’s propensity of reading the print edition by a wide margin (54.7% vs.
28.8% in 2015). In other words, online edition readers are nearly twice as likely to read the same
paper’s print edition as the general public.
RQ5 also asked whether hybrid readers outnumber online-only readers. Among 36 out of the
51 newspapers, this is true. Overall, more than 50% of online edition readers (54.7% in 2015) read
the print edition; the rest are online-only readers.
RQ6 asked about print and online reach among different age groups. Table 12.2 presents
the results. In all 51 cases, the oldest group (55+) is the most likely to read the print edi-
tion; for online, results are mixed. In 24 out of 51 cases, the highest online reach falls on
35–44-year-olds.
When the comparison is made between print and online, print reach surpassed online reach by
a wide margin across all age groups. Figure 12.2 illustrates the gap. Among the youngest group
(ages 18–24), an average of 19.9% read the print edition of the newspaper during the past seven
days; only 7.8% accessed the paper digitally.
Reality and irrationality

These findings confirm that the examined newspapers are stuck between a shrinking market
for print and an unsuccessful experiment online. From 2007 to 2015, print readership declined
substantially, while their in-market online readership saw very little or no growth at all during
the same period.
It is alarming that none of these major newspapers’ websites – other than the Washington Post
and the Austin American-Statesman – has ever reached 20% of its local population. This suggests
164
Table 12.2 In-market online and print reach by age, 2015
18–24 25–34 35–44 45–54 55+ 18–24 25–34 35–44 45–54 55+
Arizona The Arizona Republic 21.8% 18.3% 24.7% 37.2% 45.2% 9.2% 17.8% 28.4% 16.6% 11.9%
Arkansas Arkansas Democrat-Gazette 26.9% 31.5% 52.0% 43.4% 55.0% 6.0% 6.3% 9.0% 3.5% 4.8%
California Los Angeles Times 15.9% 22.2% 18.2% 23.5% 31.5% 12.2% 14.7% 12.7% 11.3% 6.1%
California San Francisco Chronicle 9.6% 17.1% 12.6% 19.8% 27.1% 8.7% 14.7% 17.1% 15.6% 9.3%
California San Jose Mercury News 23.9% 22.2% 25.7% 35.7% 44.2% 10.4% 17.3% 14.5% 10.9% 6.5%
California The Orange County Register 6.4% 5.0% 5.7% 7.8% 11.4% 2.4% 4.5% 4.0% 3.0% 1.6%
California The Press-Enterprise 1.7% 2.1% 3.5% 4.4% 6.5% 1.3% 1.0% 2.0% 2.0% 1.0%
California The Sacramento Bee 12.9% 15.9% 18.1% 27.3% 32.2% 5.3% 10.5% 9.7% 10.3% 5.3%
California U-T San Diego 35.3% 22.1% 25.0% 38.0% 52.6% 7.2% 18.6% 8.6% 12.7% 7.9%
Colorado The Denver Post 24.3% 19.4% 24.8% 31.5% 43.6% 7.1% 16.5% 11.1% 9.8% 6.4%
Connecticut The Hartford Courant 19.0% 22.9% 28.7% 28.8% 32.9% 2.1% 10.1% 11.0% 12.4% 5.7%
Florida Orlando Sentinel 23.1% 29.0% 28.7% 28.6% 29.4% 10.5% 12.9% 15.1% 7.7% 5.1%
Florida South Florida Sun-Sentinel 7.1% 13.8% 9.7% 16.8% 21.6% 10.1% 20.9% 16.5% 20.9% 22.1%
Florida Tampa Bay Times 21.3% 31.7% 26.0% 30.4% 33.2% 6.4% 7.0% 6.0% 4.7% 3.4%
Florida The Miami Herald 21.5% 17.7% 22.6% 27.1% 32.9% 8.3% 6.5% 7.8% 11.1% 6.4%
Florida The News-Press 21.8% 19.7% 15.1% 23.2% 33.1% 1.4% 5.0% 8.9% 10.7% 4.1%
Florida The Tampa Tribune 14.8% 18.0% 17.5% 19.8% 21.5% 4.7% 7.5% 9.2% 6.1% 4.0%
Georgia Atlanta Journal-Constitution 11.3% 16.2% 23.6% 24.8% 37.7% 6.9% 19.8% 18.6% 20.4% 10.0%
Hawaii Honolulu Star-Advertiser 33.9% 49.1% 45.0% 45.2% 60.6% 7.9% 10.0% 12.4% 11.9% 11.6%
Illinois Chicago Sun-Times 14.7% 24.2% 23.0% 23.4% 25.0% 6.3% 7.1% 8.8% 6.9% 3.6%
Illinois Chicago Tribune 22.6% 24.8% 27.1% 31.0% 44.7% 12.1% 16.9% 13.6% 11.9% 8.4%
Indiana The Indianapolis Star 19.3% 22.1% 30.8% 31.6% 46.0% 6.8% 17.6% 16.3% 9.3% 7.0%
Kentucky The Courier-Journal 33.7% 38.9% 35.4% 47.0% 58.4% 11.5% 10.1% 8.8% 7.7% 7.8%
Massachusetts The Boston Globe 21.4% 18.7% 21.5% 23.9% 31.5% 14.2% 20.8% 28.8% 18.0% 8.6%
Michigan Detroit Free Press 19.9% 20.5% 28.3% 33.5% 41.8% 9.1% 12.1% 19.6% 16.5% 7.8%
Minnesota St. Paul Pioneer Press 12.3% 8.8% 20.7% 21.4% 26.4% 3.4% 3.2% 10.0% 9.0% 3.0%
Minnesota Star Tribune 20.9% 33.0% 47.3% 47.7% 47.7% 6.9% 20.6% 15.1% 15.8% 9.2%
Missouri St. Louis Post-Dispatch 32.9% 21.2% 34.8% 40.6% 47.0% 7.5% 15.2% 20.0% 18.5% 11.7%
(Continued)
Table 12.2 (Continued)
18–24 25–34 35–44 45–54 55+ 18–24 25–34 35–44 45–54 55+
Missouri The Kansas City Star 25.4% 26.1% 31.9% 40.4% 52.2% 2.8% 15.3% 15.4% 15.7% 8.1%
Nevada Las Vegas Review-Journal 25.1% 24.0% 20.7% 38.0% 54.3% 8.5% 7.1% 16.9% 13.6% 8.4%
New Jersey The Star-Ledger 6.8% 4.9% 4.9% 7.7% 9.5% 4.1% 8.2% 8.2% 6.8% 3.0%
New York New York Daily News 15.2% 18.5% 17.6% 21.1% 22.5% n/a n/a n/a n/a n/a
New York Newsday 7.1% 7.6% 8.3% 11.5% 13.4% 2.7% 4.9% 4.5% 4.3% 2.2%
North Carolina The Charlotte Observer 16.3% 21.8% 22.6% 30.8% 39.3% 4.8% 11.3% 13.8% 12.8% 7.2%
North Carolina The News & Observer 8.6% 24.3% 25.1% 23.9% 29.4% 6.9% 11.0% 9.1% 8.6% 7.2%
Ohio The Columbus Dispatch 28.9% 28.9% 30.4% 37.1% 49.8% 4.5% 9.9% 8.7% 14.5% 9.3%
Ohio The Plain Dealer 18.1% 23.1% 31.8% 32.9% 37.4% 10.3% 11.9% 17.0% 11.6% 8.9%
Oregon The Oregonian 28.6% 24.2% 31.3% 28.8% 39.6% 17.5% 25.4% 25.6% 17.3% 9.9%
Pennsylvania Pittsburgh Post-Gazette 13.2% 14.9% 27.7% 30.5% 35.5% 6.5% 10.6% 18.7% 8.7% 7.0%
Pennsylvania The Philadelphia Inquirer 20.1% 13.9% 19.1% 18.2% 30.1% n/a n/a n/a n/a n/a
Pennsylvania Tribune-Review 11.2% 13.8% 19.1% 19.4% 24.6% 13.0% 6.3% 10.4% 8.4% 5.1%
Texas Austin American-Statesman 34.4% 27.8% 31.0% 42.7% 56.0% 8.2% 18.8% 21.8% 19.8% 13.0%
Texas Fort Worth Star-Telegram 7.9% 8.8% 14.2% 14.8% 20.8% 6.9% 4.1% 9.9% 7.5% 6.8%
Texas Houston Chronicle 22.4% 19.2% 22.4% 32.2% 51.0% 4.7% 10.4% 11.5% 14.1% 9.0%
Texas San Antonio Express-News 23.4% 36.4% 24.8% 45.5% 54.5% 5.3% 23.7% 13.8% 18.6% 10.8%
Texas The Dallas Morning News 21.4% 14.3% 17.7% 27.5% 35.3% 6.8% 8.7% 6.8% 9.9% 6.2%
Utah Deseret News 17.1% 12.9% 10.8% 12.9% 22.9% 10.1% 12.6% 14.7% 10.3% 8.5%
Virginia The Virginian-Pilot 26.8% 26.1% 30.9% 44.2% 45.8% 13.2% 19.5% 11.9% 17.3% 9.5%
Washington The Seattle Times 29.2% 24.5% 22.6% 29.8% 38.6% 9.4% 17.0% 18.8% 15.4% 8.8%
Washington DC The Washington Post 39.0% 26.3% 34.1% 40.2% 54.3% 22.6% 22.4% 25.2% 22.2% 15.1%
Wisconsin Milwaukee Journal Sentinel 20.9% 31.9% 35.0% 42.7% 58.2% 9.1% 19.2% 26.0% 14.7% 11.4%
Average 19.9% 21.2% 24.1% 29.1% 37.2% 7.8% 12.7% 13.7% 12.0% 7.7%
Note: Bolded text indicates highest reach within the format

Figure 12.2 In-market print and online reach by age, 2015
that not a single local U.S. newspaper has ‘made it’ digitally. The average online reach at 10%
explains why they have failed to generate sufficient digital subscription or advertising revenue –
because they have never reached sufficient online readers. In contrast, readers’ attachment to the
print edition has remained strong, with as many as 28.8% of the local population still reading (and
paying $300 to $500 a year for) the ‘dead-tree’ edition.
In sum, after 20 years of trial and error, the performance gap between U.S. newspapers’ online
and digital products is deep and wide. However, despite the fact that the print edition remains the
core product that outperforms the same newspaper’s digital offerings by almost every standard,
industry discourse has remained overwhelmingly ‘pro-digital’, often emphasizing digital gain (by
focusing on growth as opposed to absolute volume) and ignoring the stronger-than-expected
appeal of the print product.
In October 2016, journalist and media critic Jack Shafer at Politico Magazine introduced some
of the findings presented in this chapter (originally published in Journalism Practice, Chyi and
Tenenboim, 2017). His column (2016), titled “What If the Newspaper Industry Made a Colossal
Mistake?”, was shared extensively and triggered an intense debate about the future of newspapers.
Many industry observers participated in the debate, and some digital enthusiasts responded to our
findings strongly and emotionally. They put forward many ‘pro-digital, anti-print’ arguments.
Some simply declared that there is no future for print newspapers – without providing any
evidence. Others blamed Shafer and the researchers for triggering an unnecessary debate. Some
rejected the idea that newspapers’ technology-driven approach is a losing proposition, arguing
that ‘digital first’ was never really tried in U.S. newsrooms. Others justified newspapers’ digital
failure by saying that digital media are still young – thus ‘it is too early to tell’. Others believe
digital news is simply better than print. Period.
Overall, these arguments reflect how deeply rooted the ‘digital mentality’ is and expose some
of the irrationalities that may have shaped U.S. newspapers’ digital strategy. But 20 years’ empiri-
cal research indicates that newspaper readers are by no means “platform-agnostic” (Chyi, 2017).
Users perceive online news as an inferior good, a less satisfying alternative to print newspapers
(Chyi and Yang, 2009; Chyi, 2013).
167
Considering these readership trends; the oversupply of news, information, and entertainment
on digital platforms; and Google, Facebook, and other tech giants’ domination in the digital
world, “it is hard not to conclude that there is no such thing as an all-digital future for the vast
majority of U.S. newspapers” (Chyi, 2017: 21).
Regarding managerial implications, given dwindled resources, the development of digital
products often comes at the expense of their print counterpart. That is why Chyi (2017) believes
that “most US newspapers can no longer afford pursuing unrealistic digital dreams. They have
no choice but to refocus on print, where their competitive advantage lies” (p. 21).
Looking into the future, newspaper firms must drop unrealistic expectations, acknowledge the
reality, value audience research, offer quality, noteworthy content, and deliver it through preferred
platforms. Albeit no longer “wildly profitable,” there may still be a future for newspapers.
Limitations and future studies

This study analyzed readership data collected by Scarborough Research. Several limitations war-
rant discussion. First, our analysis included only larger local daily newspapers with circulation of
120,000 or above, not nationally circulated newspapers (The Wall Street Journal, USA Today, and
The New York Times) or the large number of small community newspapers in the United States.
One should take the characteristics of this sample into consideration when interpreting the results.
In addition, online readership data employed in this study do not include readers visiting
a newspaper site from outside the newspaper’s local market (or DMA), thus underestimating
the total number of online readers. Research has shown that some newspapers, especially those
circulated nationally, can reach a substantial number of overseas users (Thurman, 2014). Local
newspapers may also attain long-distance users (Chyi, 2011). One study based on 28 U.S. local
newspapers’ online reader survey data reported that an average of 27.5% of those papers’ online
users resided outside the newspaper’s designated market (NDM)7 (Chyi and Sylvie, 2010). Since
most local newspapers heavily rely on local advertisers (Runett, 2005; Newspaper Association of
America, 2007), in-market online readers have been considered more valuable than long-distance
readers (Chyi and Sylvie, 2010).
Another limitation is that the Scarborough approach may not fully consider audience exposure
to newspaper content through social media, search engines, and news aggregators. For example,
when users encounter a newspaper story on Facebook or Twitter, some follow the link to the
newspaper site. Among them, not all would identify themselves as readers of the online edition in
Scarborough’s readership survey because they may not remember visiting a newspaper site – thus
Scarborough may underestimate the number of visitors to newspaper sites. Nevertheless, Scarbor-
ough’s data focus on self-identified readership, which is more important than flyby visits through
incidental exposure in terms of time spent on the site (Mitchell et al., 2014) and business prospect.
Additionally, while this study compared newspapers’ print and online readership, which is based
on reach, it did not consider the level of engagement, such as time spent or frequency of use. Research
shows that visits to newspaper sites do not last long (Thurman, 2014, 2017; State of the News Media
2015, 2015). Future studies may examine other dimensions of online engagement such as loyalty,
depth, and stickiness (Zheng et al., 2012) to fully understand the performance of digital news products.
To conclude, no audience measures generate completely accurate results, but our longitudinal
analysis compared readership trends over time using (arguably) the best data available to industry
practitioners. As multiplatform readership gets increasingly complicated, the definition of read-
ership may have to change accordingly. Nevertheless, it is important that newspaper firms con-
stantly monitor reader composition and use that intelligence to guide (or correct) their product
strategy so as to better serve readers on preferred platforms.
168
Further reading
This chapter is an extension to Iris Chyi’s Trial and Error: US Newspapers’ Digital Struggles Toward
Inferiority (2013), which presents 20 years’ research findings regarding online users’ lukewarm
responses to newspapers’ digital offerings. The book also explains Chyi’s “online news as an infe-
rior good” theory (a.k.a. “Ramen Noodles Theory”) in detail. Jack Shafer’s column “What If
the Newspaper Industry Made a Colossal Mistake?” passionately introduced the readership gap
presented in this chapter to the industry. In a WAN-IFRA (World Association of Newspapers and
News Publishers) report, Chyi (2017) addresses many “pro-digital, anti-print” arguments surfaced
during the debate.
Notes
1 The interviews are conducted through random digit dialing (RDD) and include cell-phone-only (CPO)
respondents in all local markets.
2 A DMA is a geographic area to which a county in the US is exclusively assigned on the basis of the televi-
sion viewing habits of the people residing in the county, defined by A.C. Nielsen.
3 Formerly known as ABC’s Audience-FAX* eTrends Tool. http://abcas3.auditedmedia.com/audience-
fax/default.aspx
4 http://auditedmedia.com/data/media-intelligence-center/reports-and-tools/
5 Respondents are asked whether they have “read or looked into” any part of each of the weekday printed
newspapers available in their local area in the past seven days. They are also asked whether they have visited
the newspaper websites in the past seven days. The 2015 question prompted respondents to include access by
laptop, tablet, smartphone, etc. ‘Reach’ is the percentage of adults ages 18+ in the market whom are reached
by the newspaper.
6 http://abcas3.auditedmedia.com/scarborough/login.aspx
7 Note that a NDM is usually smaller than the DMA in which Scarborough conducts surveys to collect
readership data.
References
ASNE. (2016) Minority Employment in Daily Newspapers. Retrieved from http://asne.org/content.asp?pl=14
0&sl=129&contentid=129 (accessed 19 June 2017).
Barthel, M. (2016) “Newspapers: Fact sheet.” Pew Research Center. Retrieved from www.journalism.
org/2016/06/15/newspapers-fact-sheet/ (accessed 17 December 2016).
Bell, E. (2016) “Facebook is eating the world.” Columbia Journalism Review. Retrieved from www.cjr.org/
analysis/facebook_and_media.php (accessed 18 December 2016).
Case, B. (2009, October 11) “Dallas Morning News takes premium value approach.” Dallas Morning News.
Retrieved from www.dallasnews.com/sharedcontent/dws/bus/stories/DN-DMN_11bus.ART0.State.
Edition1.3cf5397.html (accessed 30 March 2010).
Choney, S. (2011) “Online news readership overtakes newspapers.” NBC News.com. Retrieved from www.
nbcnews.com/technology/online-news-readership-overtakes-newspapers-124383 (accessed 21 Janu-
ary 2014).
Chyi, H. I. (2006) “RE-examining the market relation between online and print newspapers: The case of
Hong Kong.” In X. Li (ed.), Internet Newspapers. Mahwah, NJ: Lawrence Erlbaum (pp. 193–205).
Chyi, H. I. (2011) “Online readers geographically more dispersed than print readers.” Newspaper Research
Journal, 32(3), 97–111.
Chyi, H. I. (2013) Trial and Error: U.S. Newspapers’ Digital Struggles Toward Inferiority. Navarra, Spain: Univer-
sity of Navarra (Media markets monographs).
Chyi, H. I. (2017) “Reality and irrationality: U.S. metropolitan newspapers between print and digital.” In
Print-online Performance Gap: A U.S.-only Study Raises Debate. Frankfurt, Germany: WAN-IFRA (pp. 12–22).
Chyi, H. I. and Huang, J. S. (2011) “Demystifying the demand relationship between online and print prod-
ucts under one newspaper brand: The case of Taiwan and the emergence of a universal pattern.” Asian
169
Chyi, H. I. and Lasorsa, D. L. (1999) “Access, use and preference for online newspapers.” Newspaper Research
Journal, 20, 2–13.
Chyi, H. I. and Lasorsa, D. L. (2002) “An explorative study on the market relation between online and print
newspapers.” The Journal of Media Economics, 15(2), 91–106.
Chyi, H. I. and Lewis, S. C. (2009) “Use of online newspaper sites lags behind print editions.” Newspaper
Research Journal, 30(4), 38–53.
Chyi, H. I., Lewis, S. C. and Zheng, N. (2012) “A matter of life and death? Examining how newspapers
covered the newspaper ‘crisis’.” Journalism Studies, 13(3), 305–324.
Chyi, H. I. and Sylvie, G. (2010) “Are long-distance users an inconvenient truth? Profiling U.S. newspapers’
online readership in the dual-geographic market.” International Journal on Media Management, 12(2), 1–20.
Chyi, H. I. and Tenenboim, O. (2017) “Reality check: Multiplatform newspaper readership in the United
States, 2007–2015.” Journalism Practice, 11(7), 798–819. doi:10.1080/17512786.2016.1208056
Chyi, H. I. and Yang, M. J. (2009) “Is online news an inferior good? Examining the economic nature of
online news among users.” Journalism & Mass Communication Quarterly, 86(3), 594–612.
Dick, B. (2009) “Spending analog dollars to get digital pennies.” TV Technology. Retrieved from www.
tvtechnology.com/default.aspx?tabid=204&entryid=1097
Doctor, K. (2011) “Nine questions on the Dallas Morning News pay plan.” Newsonomics. Retrieved from http://
newsonomics.com/nine-questions-on-the-dallas-morning-news-pay-plan/ (accessed 28 March 2011).
Editor & Publisher 2015 Newspaper Databook. (2015) Fountain Valley, CA: Editor & Publisher.
Gordon, R. (2007) “Pondering panels: ComScore, Nielsen report dissimilar numbers due to methodology
differences.” Newspaper Association of America. Retrieved from www.naa.org/Resources/Articles/Digital-
Edge-Pondering-Panels/Digital-Edge-Pondering-Panels.aspx (accessed 25 December 2009).
Gottfried, J. and Shearer, E. (2016) “News use across social media platforms 2016.” Pew Research Cen-
ter. Retrieved from www.journalism.org/2016/05/26/news-use-across-social-media-platforms-2016/
(accessed 18 December 2016).
Graybeal, G. M. (2011) “Newspapers publish less in print: Focus on web to attract young readers.” Newspaper
Hargrove, T., Miller, J. and Stempel, G. H. (2011) “Study compares print, online use of metro, community
newspapers.” Newspaper Research Journal, 32(1), 84–89.
Mitchell, A., Jurkowitz, M. and Olmstead, K. (2014) “Social, search and direct: Pathways to digital news.”
Pew Research Center’s Journalism Project. Retrieved from www.journalism.org/2014/03/13/social-search-
direct/ (accessed 19 August 2014).
Mutter, A. D. (2010, October 4) “Faced with falling circ, publishers retool rules.” Reflections of a Newso-
saur. Retrieved from http://newsosaur.blogspot.com/2010/10/faced-with-falling-circ-publishers.html
(accessed 10 April 2016).
Mutter, A. D. (2013) “How many people really pay for digital news?” Reflections of a Newsosaur. Retrieved
from http://newsosaur.blogspot.ca/2013/09/how-many-people-really-pay-for-digital.html (accessed
23 December 2013).
Myllylahti, M. (2013) “Newspaper paywalls – The hype and the reality: A study of how paid news content
impacts on media corporation revenues.” Digital Journalism, 1–16. doi:10.1080/21670811.2013.813214
Newspaper Association of America. (2007) “Newspapers’ online operations: Performance report 2006.” News-
paper Association of America. Retrieved from www.naa.org/docs/Digital-Media/200710_NP_Online_Per_
Present.pdf
Newspaper Association of America. (2012) Newspaper CirculationVolume. Retrieved from www.naa.org/Trends-
and-Numbers/Circulation-Volume/Newspaper-Circulation-Volume.aspx (accessed 18 March 2013).
Newspaper Association of America. (2013) The American Newspaper Media Industry Revenue Profile 2012.
Retrieved from www.naa.org/Trends-and-Numbers/Newspaper-Revenue/Newspaper-Media-Industry-
Revenue-Profile-2012.aspx (accessed 3 November 2013).
Newspaper Association of America. (2015) Audience-Fax User Guide. Retrieved from http://online.audi
encefax.com/Audience-Fax%20User%20Guide%204.08.pdf
Olivarez-Giles, N. (2011) “Web tops newspapers as news source for first time.” Los Angeles Times. Retrieved
from http://articles.latimes.com/2011/mar/15/business/la-fi-web-newspapers-20110315 (accessed
21 January 2014).
Paton, J. (2010) John Paton’s Dec. 2 Presentation at INMA Transformation of News Summit in Cambridge, Mass.,
Digital First. Retrieved from http://jxpaton.wordpress.com/2010/12/02/presentation-by-john-paton-
at-inma-transformation-of-news-summit-in-cambridge-mass/ (accessed 23 October 2011).
170
Penenberg, A. L. (2004, November 24) “Newspapers should really worry.” Wired. Retrieved from www.
wired.com/culture/lifestyle/news/2004/11/65813?currentPage=all
Pérez-Peña, R. (2008, September 6) “Times plans to combine sections of the paper.” The New York Times.
New York, p. C3.
Pew Research Center for the People & the Press. (2006) Online Papers Modestly Boost Newspaper Readership:
Maturing Internet News Audience Broader than Deep. Retrieved from www.people-press.org/2006/07/30/
section-1-watching-reading-and-listening-to-the-news-2/
Pew Research Center for the People & the Press. (2012) Section 2: Online and Digital News. Retrieved from
www.people-press.org/2012/09/27/section-2-online-and-digital-news-2/ (accessed 30 March 2013).
Picard, R. G. (2003) “Cash cows or entrecôte: Publishing companies and disruptive technologies.” Trends in
Communication, 11(2), 127–136.
Picard, R. G. (2005) “The nature of media product portfolios.” In R. G. Picard (ed.), Media Product Portfolios:
Issues in Management of Multiple Products and Services. Hoboken, NJ: Routledge (pp. 1–22).
Picard, R. G. (2008) “News consumption and the business of journalism.” The 2008 Breaux Symposium: New
Models for News (pp. 112–118). Retrieved from http://uiswcmsweb.prod.lsu.edu/manship/ReillyCen
ter/JohnBreauxSymposium/PreviousBreauxSymposia/item16005.html
Picard, R. G. and Brody, J. H. (1997) The Newspaper Publishing Industry. Boston, MA: Allyn & Bacon.
Rosenstiel, T. and Mitchell, A. (2011) “Overview.” The State of the News Media. Retrieved from http://
stateofthemedia.org/2011/overview-2/ (accessed 10 April 2011).
Runett, R. (2005) “Newspapers’ online operations: Performance report 2004.” The Digital Edge. Retrieved
from www.digitaledge.org/download/Performance_2004.pdf
Shafer, J. (2016) “What if the newspaper industry made a colossal mistake?” POLITICO Magazine. Retrieved
from http://politi.co/2eb1yFo (accessed 20 December 2016).
State of the News Media 2015. (2015) Pew Research Center. Retrieved from www.journalism.org/2015/04/29/
newspapers-fact-sheet-2015/?utm_campaign=the%20latest&utm_source=hs_email&utm_
medium=email&_hsenc=p2anqtz-9wfmmtaymmp6lptdtuynnj874tlrygc7h9bc3rq7puhpfofyejcltfsjzk-
loydokundycn
Sulzberger, A. O. (2011, March 18) “A letter to our readers about digital subscriptions.” The New York Times.
Retrieved from www.nytimes.com/2011/03/18/opinion/l18times.html (accessed 1 April 2011).
Thurman, N. (2014) “Newspaper consumption in the digital age: Measuring multi-channel audience atten-
tion and brand popularity.” Digital Journalism, 2(2), 156–178. doi:10.1080/21670811.2013.818365
Thurman, N. (2017) ‘Newspaper consumption in the mobile age.” Journalism Studies, 1–21. doi:10.1080/1
461670X.2017.1279028
Webster, J. G. (2014) The Marketplace of Attention: How Audiences Take Shape in a Digital Age. Cambridge,
MA: The MIT Press.
Williams, A. T. (2016) “Paying for digital news: The rapid adoption and current landscape of digital sub-
scriptions at U.S. newspapers.” American Press Institute. Retrieved from wwwx.americanpressinstitute.
org/publications/reports/digital-subscriptions/ (accessed 17 December 2016).
Zheng, N., Chyi, H. I. and Kaufhold, K. (2012) “Capturing ‘human bandwidth’: A multidimensional model
for measuring attention on web sites.” International Journal on Media Management, 14(2), 157–179. doi:1
0.1080/14241277.2011.619153
171
13
ON DIGITAL DISTRIBUTION’S
FAILURE TO SOLVE NEWSPAPERS’
EXISTENTIAL CRISIS
Symptoms, causes, consequences, and remedies
Neil Thurman, Robert G. Picard, Merja Myllylahti,

and Arne H. Krumsvik
Newspapers’ online experiments have not reversed their failing fortunes. Despite the decreases in
print circulation suffered by newspapers in developed countries and two decades of investment
in digital distribution, many newspapers still have a larger number of readers for their print prod-
ucts than for their online editions via PCs (see, e.g., NRS, 2017a; see also Chyi and Tenenboim,
this volume, Chapter 12). The effects of these undersized online audiences are exacerbated by
the fact that readers of digital editions are an order of magnitude or two less attentive than their
print counterparts (Thurman, 2017). The result is that newspapers receive by far the greater part
of their audience attention from their print channels (ibid.). This distribution of attention helps
to explain print’s continued delivery of high proportions of newspaper revenue (Pew Research
Center, 2016: 14).
This chapter examines some of the symptoms and causes of the crisis facing newspapers via
analyses of their finances and audience measures. The consequences of the crisis, and whether
there are any realistic remedies, are also considered, both in relation to journalism as a product
and to the institutions, such as newspapers, that have traditionally produced it.
We start with an analysis of the financial performance of multiplatform news publishers in
Australia, Europe, and the US, which leads us to conclude that digital distribution is not revers-
ing newspapers’ decline, and raises questions about the support for journalism in the long term.
Next, some of the consequences of the declines that have already taken place are discussed.
Moving from consequences to possible remedies, the chapter focuses on two areas. First, media
policy, and second, journalism as a product: what news should be produced and how it should
be delivered. Another strand of the chapter concerns audience measures. They are used to help
explain newspapers’ continuing dependency on print revenues, and are understood, depending
on their constitution and use, as both a party to the crisis and as an able assistant in its alleviation.
Symptoms
News publishers are reporting strong increases in their digital revenues (see, e.g., Fairfax Media,
2017). We must, however, be careful how we interpret such claims and skeptical about the ability
172
On digital distribution’s failure
of ‘digital-first’ strategies to transform the fortunes of struggling news outlets. As this section
illustrates, increases in news publishers’ digital revenues must be seen in the context of their total
revenue and, if we are interested in the sustainability of the journalism that they produce, we must
disentangle the performance of their news and non-news businesses.
Although the circulations of newspapers in some Western economies have been in decline for
decades – they peaked in 1984 in the United States (Chittum, 2014) and around 1990 in Canada
(Communications Management Inc., 2011: 6) – revenues from print held up for longer. In the
US, for example, income from print advertising peaked around the turn of the millennium (Perry,
2012). Falls in the revenue of newspapers in the Anglosphere early in the twenty-first century
accelerated as a result of the global financial crisis of 2007–8 (Perry, 2012; Pew Research Center,
2015: 27), prompting some newspapers to start to engage in “major transformation[s]” (Sabbagh,
2011) involving a renewed focus on digital distribution.
Although there had been a long history of newspapers experimenting with online editions, dat-
ing back to the 1980s (see, e.g., Finberg, 2003), the moves made following the Great Recession were
more fundamental, justified as being essential to newspapers’ survival. For example, in 2011 Andrew
Miller, chief executive of the UK’s Guardian Media Group, committed his titles to a “‘digital-first’
strategy” because, he said, “doing nothing was not an option” (Sabbagh, 2011). The story was simi-
lar in Australia where, by 2012, with their print revenues in crisis, Fairfax Media and APN News &
Media also launched digital-first initiatives (Myllylahti, 2015: 194). In May 2016, Fairfax Media’s
chief executive, Greg Hywood, said that it was “inevitable” that the publisher would stop printing
weekday editions of the Age and the Sydney Morning Herald in the near future (Hayes, 2016).
The idea behind digital-first strategies is to lessen news publishers’ reliance on print as a
delivery platform, making digital reporting, in the words of the New York Times Co.’s Innovation
Report, “the top priority” (The New York Times, 2014: 82). The report advocated a rapid digital
transformation in order that the New York Times could compete in a crowded, well-funded and
innovative digital marketplace (3) and claimed that digital-first was more than a “buzz phrase”
and was taking “hold across the industry” (83).
Three years on, in February 2017, the New York Times Co.’s chief executive, Mark Thomp-
son, stated that the company’s digital growth was “unprecedented” (Ember, 2017) and, in another
report, claimed that its digital revenue “towers above that of any news competitor” (The New York
Times, 2017). Yet at the same time, the report admitted that “for all the progress we have made, we
still have not built a digital business large enough on its own to support a newsroom that can fulfil
our ambitions” (ibid.).
In the same month, Fairfax Media published its half-year financial results on the other side
of the world. Although the Australian company’s chief executive, Greg Hywood, announced a
strong increase in digital revenues, this was driven mainly by its digital real estate listing service,
Domain. Somewhat surprisingly, the company also announced an about-turn in its publishing
strategy, revealing that it was planning to continue to print its daily publications “for some years
yet” (Fairfax Media, 2017: 3).
Such statements by the New York Times Co. and Fairfax Media indicate that digital-first may
not have been as transformative as some had initially hoped. In this chapter, we analyze the New
York Times Co.’s and Fairfax Media’s print and digital revenues from 2011 to 2016 in order to
gauge the extent of their digital transformations in financial terms. There is no publicly avail-
able historical data about their digital revenues outside this period. The two companies operate
in different markets – in the United States and Australia – but are comparable as they are both
primarily news publishers and both have implemented digital-first strategies, including paywalls.
Our analysis of the New York Times Co.’s annual reports between 2011 and 2016 reveals that
the company’s total revenue declined by 33%, despite a doubling of digital subscription revenues
between 2012 and 2016. At the end of the period, digital contributed just 28% of total revenue
173
Neil Thurman, Robert G. Picard et al.
2,500
US$ millions Print revenues Digital revenues
2,000
1,500
1,000
500
0
2011 2012 2013 2014 2015 2016
Figure 13.1 Print and digital revenues for the New York Times Co., 2011–2016
Note: Only the total revenue is shown for 2011, as in that year the New York Times Co. did not fully
break down their revenue into its print and digital components.
Source: The New York Times Co.’s annual reports
(see Figure 13.1). Print revenues (including subscriptions) declined 11.5% between 2012 and
2016, while digital revenues grew 32% over the same period. Digital subscription revenues may
have grown 100%, but digital advertising revenues declined 2.7%. Based on these figures, it is
clear that the publisher is still heavily reliant on print in terms of its revenue.
Fairfax Media suffered a 21% decline in total revenue between 2011 and 2016. Though
digital subscription revenues grew 58% between 2014 and 2016, and digital advertising revenues
by 24%, at the end of 2016 digital made up just 21% of Fairfax’s total revenue (see Figure 13.2).
These data clearly demonstrate that Fairfax Media is, like the New York Times Co., still heavily
reliant on print revenues, which may well explain the 2017 change in its print publishing strategy.
Comparing the print and digital revenues of the New York Times Co. and Fairfax Media
to those of other newspaper publishers is difficult. For example, U.S. newspaper publisher Gan-
nett has only revealed the proportion of its total revenue coming from digital since 2013, and
the digital component of its advertising revenues since 2014. There is still no publicly available
information about Gannett’s digital subscription income. Gannett’s 2016 annual report revealed
that 26% of its total revenue came from digital (Gannett, 2017). Canada’s Postmedia also releases
incomplete data on its digital revenues. Its 2016 annual report shows that digital made up 13% of
its total revenue (Postmedia, 2016: 11).
Other companies, such as Germany’s Axel Springer and Norway’s Schibsted, report that digital
revenues account for a much higher proportion of their total revenue – 67% in 2016 for Axel
Springer (Axel Springer, 2017) and “over 70 percent” for Schibsted (Schibsted, 2017: 8) – but have
174
1,800
US$ millions Print & other revenues Digital revenues

1,600
1,400
1,200
1,000
800
600
400
200
0
2011 2012 2013 2014 2015 2016
Figure 13.2 Print and digital revenues for Fairfax Media, 2011–2016
Source: Fairfax Media’s annual reports
different corporate structures to those of the New York Times Co. and Fairfax Media. Although
journalism is a “foundation” (Axel Springer, 2016: 19) of Axel Springer’s business model, the
company has been diversifying its holdings, divesting itself of some newspapers, such as Hamburger
Abendblatt and Berliner Morgenpost (Eddy, 2013), making investments in established non-news digi-
tal businesses such as Airbnb (Axel Springer, 2012) and Uber (Hawkins, 2017) and founding a
digital incubator that invests in early stage, mostly digital companies such as ReachHero, which
arranges product placements in social media, and Massagio, which delivers “high-quality massages
on demand with just a few clicks” (Axel Springer, n.d.). Schibsted has a large holding of online
classified advertising sites – 36 sites in 30 countries (Schibsted, n.d.) – as well as owning a number
of other digital service businesses in areas such as health.
Explaining reach/revenue disparities

The relative contribution made by digital output to the total revenues of the publishers in our
sample that are reliant mainly on news businesses averages out at little more than a fifth: 21% at
Fairfax, 28% at the New York Times Co., 26% at Gannett, and 13% at Postmedia. As we have
mentioned, it is common for many newspaper brands to have a larger number of readers for their
print products than for their online editions via PCs (see, e.g., NRS, 2017a), but if we take mobile
visitors into account then newspaper brands’ digital visitors start to surpass their print readers in
number. For example, on a monthly basis UK newspaper brands average 4.62 million print read-
ers and 3.43 million readers via PCs, but 10.48 million mobile readers (NRS, 2017b).
175
Why, then, is the reach that digital strategies can deliver so much higher than the revenues they can
earn? One reason is the structure of the digital advertising business, with its many intermediaries –
the digital advertising networks and data firms – taking a slice of the pie. Another is the vast inven-
tory that digital offers, driving prices down through the law of supply and demand. Another is
the difficulty, particularly on mobile, of incorporating advertisements into online editorial content.
Yet another is consumers’ reluctance to pay for digital content. However, these reasons do not, we
believe, offer sufficient explanation for the revenue/reach disparity between print and digital.
A further explanation can be found in the levels of attention paid by print readers and online
visitors, which differ by one or two orders of magnitude. For example, while the print editions
of national newspapers in the UK are read for an average of 40 minutes per reader per day, those
same newspapers’ online editions are read for an average of less than 30 seconds per user per day,
an 80-fold difference (see Table 13.1). It is reasonable to assume that there is a correlation between
the attention an information product, like a newspaper, can attract from its consumers and the
amount they are willing to pay for it. Products that hold consumers’ attention for short periods
are likely to be valued less highly than those that hold their attention for longer. Although few
publishers sell advertising based on the length of time readers spend with their brand (the Financial
Times is one exception),1 the cost of exposing an advertisement to 1,000 potential customers (the
CPM cost) is much higher via printed newspapers than via websites. One estimate, by Morgan
Stanley (2010: 26), put the average print newspaper CPM at around $16.5 and the average internet
CPM at around $2.5. Such a difference would imply that the extra attention paid by print readers
is already priced into advertising costs. Some have argued for the costs of advertising to be more
explicitly linked to attention. Bassett and Green (2015) proposed the metric of “Gross Attention
Minutes”, which they defined as “Reach (Audience size) x Ad engagement (Av Dwell time per
exposure/impression)”.
Table 13.1 Comparison of the time spent reading per day by online and print readers of 11 UK national
newspaper brands
Minutes
Online usage per Print reading time per

visitor per day1 reader per day2
The Mail 2.00 43

The Guardian 0.68 39
The Sun 0.67 32
Mirror 0.35 37
Record 0.32 31
The Times 0.29 47
The Telegraph 0.29 53
Star 0.26 31
Express 0.20 39
The Scotsman 0.16 38
The Herald 0.15 50
Average 0.49 40
1
Includes PC and mobile usage. The monthly figure for March 2016 (provided by comScore) was divided
by 31 to arrive at an approximate daily figure. Data is for the UK audience. Numbers represent whole or
fractions of minutes, not minutes and seconds.
2
Source: NRS. Based on reading time for Monday–Friday editions (which is lower than for Saturday and
Sunday editions). Data is an average for the period April 2015–March 2016 and for Great Britain.
Source: Originally published in Thurman (2017)
176
So, looking at how audiences allocate their attention, rather than simply relying on reach, can
help to explain newspapers’ continuing financial dependency on print. But could audience measures
be more than a diagnostic tool? Might they also have been a contributory cause of the crisis? For
many years,2 newspapers reported their multiplatform audiences using different time periods, on
a daily or weekly basis for print and monthly for online. This favored their digital editions, giving
them a longer time period over which to accrue audiences – a benefit of particular import because
of the relative infrequency with which online readers consume newspaper brands. For example, UK
newspaper brands’ websites and apps are visited an average of just 5.3 times per visitor per month.3
By comparing their print and online audiences over different time periods, and by promoting
a false equivalence between print and online readers, newspapers’ accounts of their own audi-
ences have shown a preoccupation with audience size. The measures they favored had audiences
for online dwarfing audience for print, whereas, in fact, the attention captured by newspapers’
print editions has far exceeded that achieved by their online editions. For example, of the time
spent with UK national newspaper brands by their British audiences over a year, 88.5% is still in
print, with just 11.5% coming via their websites and apps (see Table 13.2).
A preoccupation with audience size can, says Leo Bogart, lead to “erroneous decisions in the
management of media content” (Bogart, 1966: 47). In the case of newspapers, that preoccupation
has resulted in a focus on building digital reach, including via low-quality ‘clickbait’ content, at
the expense of investments in journalism that engages audiences.
Audience measures may have been a contributory cause of the crisis, but they also have the
potential to assist its alleviation by providing news publishers with a better understanding of their
customers’ platform and content preferences, issues we will return to in this chapter’s last section.
The sustainability of newspaper brands

Our examination of the digital transformations and revenues of a range of publishers raises some
important questions about the sustainability of newspaper brands in the Anglosphere. At one end
of the spectrum, the New York Times Co. represents journalism as a business in a relatively pure
form, very heavily reliant on selling news content to audiences and selling audiences for that news
content to advertisers. Those foci, in particular on its subscribers, set it apart, we and the company
Table 13.2 Proportions of the annual time spent with each of 11 UK national newspaper brands by their
aggregated British adult audiences that comes via their print editions, via PCs and via mobile
devices, April 2015–March 2016 (inclusive)
Print PC Mobile
The Mail 78% 8% 14%

The Sun 98% 1% 2%
The Telegraph 90% 5% 5%
Mirror 92% 2% 6%
The Times 98% 0% 2%
The Guardian 75% 8% 18%
Express 96% 2% 2%
Star 96% 1% 3%
Record 95% 1% 4%
The Herald 98% 1% 1%
The Scotsman 93% 3% 4%
Source: Adapted from Thurman (2017)

177
believe, from many other media organizations (The New York Times, 2017). However, as we have
shown, growth in the New York Times Co.’s digital revenues has failed to halt falls in its revenue
overall, and, as of 2016, digital represented less than a third of its total income and was unable to
support a newsroom of the size that the company feels it needs (The New York Times, 2017).
At the other end of the spectrum are Axel Springer and Schibsted, oft-touted as having made
a ‘successful’ transition to digital (see, e.g., Burgelman, Siegel and Kissick, 2016 and Anand, 2007).
With more than two-thirds of their revenue now coming from digital, we can see why they are feted
by some in the business community. However, as we have shown, looking more closely at where their
digital revenues come from reveals many of the sources to be ventures unrelated to news.
There is nothing new about news publishers supporting themselves, or being supported, by
non-news enterprises. After all, journalism, especially of the serious kind, is not always a profit-
able product. For example, some ‘quality’ British newspaper brands – like the Times, Sunday Times,
the Guardian, and the Independent – lost money for decades (see, e.g., Turvill, 2016), carried by
owners with other sources of income. Such subsidies and support are no less necessary now, as
other sources of income for newspapers are in decline. For example, the unbundling of classified
advertising from the editorial content of newspapers following the emergence of cheap – or free –
online alternatives, such as Craigslist and eBay, has decimated a revenue stream newspapers used to
rely on – in the US for between a quarter and a third of revenue (Edmonds et al., 2013).
The issue, then, is not so much where the non-news revenues of the likes of Axel Springer and
Schibsted come from, but rather where they go to. Although Axel Springer recognizes that “quality
journalism [is] indispensable [. . .] for [. . .] society” (Axel Springer, 2016: 5) and says “journalism
[. . .] always will be the foundation of our business model” (19), the question is how much will
quality journalism, in the future, remain a visible monument to such companies’ activities or instead
become an artifact buried under new strata of digital businesses operating in different contexts.
The signs are decidedly mixed. Fairfax Media has announced plans to unbundle its successful real
estate listings portal, Domain, from the rest of the business in order to “maximise returns for [. . .]
shareholders” (Ward, 2017) and has said that it will not “prop up or cross-subsidize newspapers”,
running them “only as long as they are profitable” (Mathieson, 2013). Axel Springer’s CEO Mathias
Döpfner says that, as a publicly listed company, it cannot do journalism “out of nostalgia”. Unless
journalism can have a “real, profitable business model” and make a “contribution to other busi-
nesses” without subsidies, it has “no future” he says (Doctor, 2015a).
This focus on financial return is an almost inevitable consequence of these companies’ cor-
porate structures. It would be naïve to expect shareholders at Axel Springer or Fairfax to sup-
port loss-making news brands. It is also naïve, as we hope we have shown, to interpret these
companies’ financial results and the statements they make about their digital transformations, at
face value. In particular, we must eschew the assumptions that digital distribution is halting, or
reversing, newspapers’ decline and that corporate institutions will, over the long term, continue
to cross-subsidize journalistic activity. All this does not necessarily mean an end for journalism.
For Döpfner it is a challenge: publishers need to make journalism something consumers desire
so that they will pay for “great storytelling . . . information [and] entertainment” (ibid). We will
return to some of the ways this challenge might be met a little later. First, however, let us consider
briefly the wider consequences of newspapers’ decline.
Consequences
News is relatively resource-intensive to produce. But while the cost of journalism is high, the
lack of journalism comes at an even higher cost. In societies where the media are unfree or weak,
corruption is not revealed, people abuse power without being exposed, and people’s lives and
opportunities are adversely affected (Waldman, 2011: 246–7).
178
On the supply side, media shadows emerge as geographical areas and particular issues – such
as suburbs, local government, private enterprises, and church affairs – are given lower prior-
ity or are overlooked by news media due to commercial or resource considerations (Nord and
Nygren, 2002). There are further risks of decreases in the width, depth, and quality of reporting,
as resource-intensive investigative journalism seldom generates higher advertising revenues. In the
current climate, scarce resources are more likely to be allocated to the creation of content that
generates revenues quickly than to tasks such as fact-checking (Bakker, 2014).
Two further issues on the demand side exacerbate the consequences of the reduced resources
(see, e.g., Doctor, 2015b) in editorial departments. First, social media, search engines, and aggre-
gators have become an important source of news for many, especially younger users, taking a
key role as distribution channels for news media (Newman, 2016). Algorithmic selection and
prioritization of content may expose users to a narrower range of news content and give higher
priority to unreliable sources (Allcott and Gentzkow, 2017).
Second, an increasing proportion of the population rarely or never consumes news. Young
people and those with lower levels of educational attainment and interest in politics are over-
represented in this group. The proportion of non-newspaper users is increasing the most, but the
proportions of radio and television audiences who do not consume news on those platforms are
also increasing (Trilling and Schoenbach, 2013).
These two demand-side trends amplify the consequences of media shadows on the supply
side, leading to questions about whether intervention is required in order to support public-
interest journalism. This chapter now turns to explore two types of remedy: policy-based and
product-based.
Remedies
Policy-led innovation and entrepreneurship

Due to the current crisis facing news institutions, discussions about media policy have, increas-
ingly, focused on economic sustainability, motivated by concerns about the ability of legacy
news providers to fulfill their established roles. A key question is whether incumbents should
be subsidized or whether a different approach should be taken to supporting the development
of sustainable public-interest journalism. With newspaper revenues continuing to fall and com-
mercial cross-subsidies insecure, propping up legacy news providers from the public purse would
require a blank check, something governments are unlikely to countenance.
Beyond this budgetary impediment, there are other arguments against general subsidies. First,
government intervention conflicts with liberal notions of freedom of expression. Second, there
are competition and trade policy considerations, which may prohibit subsidies. Third, and per-
haps most importantly, subsidizing incumbents could, from a theoretical perspective, limit inno-
vation from newcomers. And because newspapers and radio and television broadcasters have
been found to be significantly less innovative in product, service, and process innovation than
other service firms (Krumsvik et al., 2017), subsidies directed their way may have an even stron-
ger anti-innovation effect.
Media policy has, traditionally, tended to focus on deterring unwanted outcomes – such as
monopolies or the dissemination of harmful content – rather than incentivizing innovation in the
media industries. Indeed, it has been suggested that it is often “out of kilter with business interests”,
impedes technology development, and typically attempts “to solve yesterday’s problems” (Lund,
2016). Policy-based interventions in the crisis facing public-interest journalism require, then, some-
thing of a culture change in policy circles. We would argue in favor of such interventions as long
as they enable innovation and entrepreneurship.
179
It is all very well talking about innovation in general terms, but how best can journalism pre-
pare for the future? In the next section, we take a product perspective on the problem, and suggest
that news organizations need a better understanding of the public’s needs and desires in order to
deliver value to their consumers.
The product perspective

The challenges presented by changing technology, audience behavior, and revenue require clear
thinking and level-headedness. In seeking solutions, it is critical to determine whether the focus
is on journalism or the institutions of journalism. Journalism represents practices designed to
ensure veracity and accuracy of news and to treat subjects of news coverage fairly. These prac-
tices developed first in print, then news agencies and television, and they are now employed by
professional journalists and by others working in a variety of digital forms. The institutions of
journalism are the organized workplaces and labor settings in which journalism is practiced.
These institutions have produced and distributed news and information for three centuries, with
commercial interests being central to them for the latter half of that history. Institutional deter-
mination of what is news and how it should be covered, presented, and distributed has been sig-
nificantly influenced by institutional interests, and these sometimes produce deviations from the
best practices of journalism and from the information desired by news consumers.
Although there is some concern about journalism, most sustainability concerns are focused
on the institutions of journalism. Journalism is evolving, finding effective ways to use digital
opportunities, and changing to embrace the interactivity and new functionality provided by
digital platforms. Journalism institutions and many journalists working in them have, however,
embraced the idea that the challenge in the digital world is primarily how to monetize digital
media and increase revenues. Only limited attention is being paid to the nature of the journal-
ism practiced and the kinds of content provided. The latter is thought of primarily as a means
of driving readership upward on various platforms in the wishful pursuit of advertising rev-
enues. This institutionally focused strategy is designed to serve institutional interests rather than
to improve the institution’s offerings. Pursuing solutions to the revenue challenge masks a far
greater challenge – the news product itself.
All products serve purposes, and their popularity and survival are affected by their usefulness,
available resources, changing technologies, lifestyle shifts, and social fashion. These factors have,
over time, affected products as fundamental as salt (Kurlansky, 2003), complex products such as
steam engines that endured for three centuries before being replaced by other forms of power
production (Hills, 1989), and indulgences such as cigarettes that were glamorous and then reviled
(Brandt, 2009). Products are reflections of the cultures in which they are produced, and products
are not equally successful across cultures (Rapaille, 2006) because they are not equally useful or
do not fit lifestyles, cultural values, or individual needs.
Applying the product perspective to news involves understanding its necessity, the functions
it serves for consumers and the requirements that news products must meet. These elements are
essential for all product design (Ulrich and Eppinger, 2015). It is easier to apply product design
principles to new products. It is difficult for existing firms with well-established products to
adapt them over time because institutional history, culture, and practices often conflict with
needed innovation.
Determining the nature of news products requires significant consideration today. News is
not organic. It is based on cultural and political values and the interests of those providing it.
Determining its nature was less of a challenge when there were few options for consumers and
monopolies or near monopolies existed in news and information in print and television. Even
in those conditions, journalism research showed public disagreement with journalists and editors
180
over what constituted news and how it was presented, a sentiment that continues (van der Wurff
and Schoenbach, 2014). The public is not well-served, for example, by procedural coverage
of politics and public agencies and constant commentary; it needs broader understanding of
issues and explanations of developments (Heider et al., 2005). Horserace coverage of elections is
criticized because it drowns out content that provides understanding about candidates’ views of
issues and their visions (Craig, 2014). News about crime focuses on salacious details of individual
crimes but rarely on its broader social implications (Graber, 1980). In nearly every category of
news that has been studied – foreign relations, domestic social policy, education, health, etc. –
research has revealed problems associated with an overemphasis on event-centered and breaking
news reporting and with needs and desires for more explanation in news reports (Heider et al.,
2005; Dagnes, 2010). Nevertheless, the event-centered style has continued because it fits institu-
tional needs by making planning easier, reducing costs, and reinforcing clear newsbeats. In today’s
high-choice environments for news and information, however, many readers and viewers have
already moved away from legacy news providers and continue to do so.
Journalism faces new economic conditions for production and distribution, new funding
methods, more participatory audiences, and more opportunities than in the past (Picard, 2014).
This environment is challenging because journalism has never been a viable commercial product
on its own. It has always depended on subsidies and financial streams from activities such as com-
mercial printing and advertising, as well as on monopolies and exclusivity arrangements, to make
it financially feasible (Picard, 2016).
Nevertheless, there are now more types of journalism operations than in the past. Never have
we had such an array of journalism practice, primarily because the economics of digital environ-
ments are friendlier to news and information start-ups, and digital platforms support a variety of
business models (Picard, 2011).
A fundamental challenge in the digital environment is that news consumers are shifting from
single-destination sources to distributed media sources, increasingly obtaining their news from a
variety of platforms and services. This is creating a shift away from traditional mass media mar-
kets to the serving of smaller audiences and individuals. Such a change requires news providers to
pay greater attention to individual motives, needs and expectations, personal interests and tastes,
individual lives and social contexts, and the locations and settings of individual media use. This
requires news providers to have a much better understanding of their consumers than in the past,
whether those providers wish it or not.
News providers today must understand what their platforms and services provide to customers
and what customers like and dislike about them. They must consider how they can deliver better
value, the functions they are performing for customers, the emotional and expressive needs they
are satisfying, the problems they are solving for customers, how they can make customers’ lives
easier or more enjoyable, and what customers’ expectations are and how they can be exceeded.
Accomplishing that requires deep consideration of value and value creation (Picard, 2010).
Fundamentally, this requires news organizations to shift from a goods logic to a service logic.
The focus cannot be on creating a newspaper, newscast, or online site that is available to custom-
ers but does not have customers at the heart of its thinking. There must, instead, be a customer-
centric approach in which the organization helps customers to solve their informational and
news needs. To be successful, news organizations must establish and maintain relationships with
their news customers so that they can anticipate and solve their needs through a variety of inter-
actions on various platforms.
These platforms and products should not be perceived as substitutes, however. Each must be
valuable independently of the others, and collectively they must provide greater value than the
sum of their individual values. The purpose of multiple platforms and products should not be
merely to find new customers but to increase and improve contact with existing customers.
181
There are no easy solutions to the challenges currently facing news providers. They must,
therefore, have a heightened awareness of those challenges, commit to better servicing the long-
term interests of news consumers rather than focusing on their own short-term needs, and invest
in improvements that meet those customers’ needs.
Conclusion
This chapter has considered the crisis facing newspapers in some highly developed economies,
describing the decades of dwindling circulations and the post-millennial retreat in revenues, a trend
that accelerated following the Great Recession. Newspapers’ consequent digital-first strategies have
failed to reverse their fortunes. The combination of undersized, inattentive online audiences and the
particular characteristics of digital media and advertising explain why, by 2016, news outlets remained
reliant on print for an average of around 80% of their revenues. With most of those outlets contract-
ing, it is no surprise to see owners diversifying out of the newspaper business and ruling out cross-
subsidies for journalism in the long-term. While the contractions that have affected newspapers in
the past – and will affect them in the future – have, undoubtedly, resulted in the emergence of media
shadows, to the disservice of democracy, we have some hope that this situation can be alleviated.
To come to this conclusion, we needed to separate the fate of newspapers as institutions from
the fate of the journalism that they have, traditionally, produced. While we hope, and expect, that
some newspapers will survive, it may be that other individuals and institutions are better adapted
to produce the news products demanded in the digital age. Policy and regulation need to change
to encourage innovation and entrepreneurship in the news business, whatever their source. And
that innovation needs to be based around a better understanding of what audiences want and
need from news. Better metrics will help – those that value attention, not just reach – and so will
an increased appreciation of the public’s desire for analysis, explanation, and personalization, and
of their platform and presentational preferences.
Newspapers are diminishing, unlikely to return to the levels of popularity and influence they
once enjoyed. Journalism, however, does not have to follow the same trajectory. Indeed, it should
not. These troubling times require an abundance of its best qualities.
Further reading
Neil Thurman’s (2017) article, “Newspaper Consumption in the Mobile Age: Re-Assessing Multi-
Platform Performance and Market Share Using ‘Time-Spent’”, argues that newspapers should
look beyond the number of readers they reach if they are to capture the character of multiplat-
form consumption, demonstrates how this can be done, and analyzes the results of doing so. The
progress being made by news publishers in building digital revenues, in particular via subscriptions,
is described in Myllylahti (2017). Robert G. Picard’s (2010) Value Creation and the Future of News
Organizations: Why and How Journalism Must Change to Remain Relevant in the Twenty-First Century
argues that the fundamental value of news products needs to be reconsidered in the digital age, that
traditional value creation is insufficient, and that new value must be sought.
Acknowledgments
This work was partly supported by a Volkswagen Foundation Freigeist Fellowship.
Notes
1 Since 2015 the Financial Times has sold advertising space on a “cost per hour” basis (FT, 2015).
2 For example, from at least 1996 until 2012 in the UK (Thurman, 2017).
182
3 Data is for June 2016 and covers the UK audience of the Guardian, Daily Mail, Telegraph, Mirror, Express,
and Sun. Source: comScore.
References
Allcott, H. and Gentzkow, M. (2017) “Social media and fake news in the 2016 election.” Journal of Economic
Perspectives, 31(2), 211–236.
Anand, B. N. (2007, April 16) “Schibsted.” Harvard Case Study. Retrieved from https://hbr.org/product/
schibsted/707474-PDF-ENG
Axel Springer. (2012, February 2) Axel Springer and Airbnb Join Forces. Retrieved from www.axelspringer.
de/en/presse/Axel-Springer-and-Airbnb-join-forces_2213956.html
Axel Springer. (2016) Annual Report 2016. Retrieved from www.axelspringer.de/dl/27011919/Annual_
Report_2016.pdf
Axel Springer. (2017, March 9) Axel Springer Further Expands Its Position as Europe’s Leading Digital Pub-
lisher. Retrieved from www.axelspringer.de/en/presse/Axel-Springer-further-expands-its-position-as-
Europe-s-leading-digital-publisher_31107938.html
Axel Springer. (n.d.) Startups. Retrieved from www.axelspringerplugandplay.com/startups
Bakker, P. (2014) “Mr. Gates returns: Curation, community management and other new roles for journal-
ists.” Journalism Studies, 15(5), 596–606.
Bassett, D. and Green, A. (2015) “Engagement as visual attention: A new story for publishers.” Paper pre-
sented at the Publishing and Data Research Forum, London, 17–20 October. Retrieved from www.pdrf.
net/wp-content/uploads/2015/10/35_Syn11Green-Bassett-Engagement.pdf
Bogart, L. (1966) “Is it time to discard the audience concept?” Journal of Marketing, 30(1), 47–54.
Brandt, A. (2009) The Cigarette Century: The Rise, Fall, and Deadly Persistence of the Product that Defined America.
New York, NY: Basic Books.
Burgelman, R. A., Siegel, R. and Kissick, R. (2016) “Axel Springer in 2016: From transformation to accel-
eration?” Stanford Case Study. Retrieved from www.gsb.stanford.edu/faculty-research/case-studies/
axel-springer-2016-transformation-acceleration
Chittum, R. (2014, May 28) “Reader revenue and the great newspaper ad bubble: Four decades of artifi-
cially high advertising goes pop.” Columbia Journalism Review. Retrieved from http://archives.cjr.org/
the_audit/newspaper_subscription_revenue.php
Communications Management Inc. (2011, May 6) Sixty Years of Daily Newspaper Circulation Trends. Retrieved
from http://media-cmi.com/downloads/Sixty_Years_Daily_Newspaper_Circulation_Trends_050611.pdf
Craig, R. (2014) Polls, Expectations, and Elections: TV News Making in U.S. Presidential Campaigns. Lanham,
MD: Lexington Books.
Dagnes, A. (2010) Politics on Demand: The Effects of 24-Hour News on American Politics. New York, NY: Praeger.
Doctor,K. (2015a,April 7) “What are they thinking? The eight principles for transforming Axel Springer.”Polit-
ico.com. Retrieved from www.politico.com/media/story/2015/04/what-are-they-thinking-the-eight-
principles-for-transforming-axel-springer-003649
Doctor, K. (2015b, July 28) “Newsonomics: The halving of America’s daily newsrooms.” NiemanLab. Retrieved
from www.niemanlab.org/2015/07/newsonomics-the-halving-of-americas-daily-newsrooms/
Eddy, M. (2013, August 7) “German publisher’s exit from print seen as betrayal.” The International Herald
Tribune, p. 15.
Edmonds,R.,Guskin,E.,Mitchell,A. and Jurkowitz,M. (2013,May 7) “Newspapers:By the numbers.”Stateofthe-
media.org. Retrieved from www.stateofthemedia.org/2013/newspapers-stabilizing-but-still-threatened/
newspapers-by-the-numbers/
Ember, S. (2017, February 2) “New York Times Co.’s decline in print advertising tempered by digital gains.”
The New York Times. Retrieved from www.nytimes.com/2017/02/02/business/media/new-york-times-
q4-earnings.html
Fairfax Media. (2017) Half-Year Results Announcement. Retrieved from www.fairfaxmedia.com.au/Article
Documents/193/2017-02-22_Media%20Announcement.pdf.aspx?Embed=Y
Finberg, H. (2003, October 27) “Before the web, there was Viewtron.” Poynter.org. Retrieved from www.
poynter.org/2003/before-the-web-there-was-viewtron/17738/
FT. (2015, May 18) “Financial Times rolls out ‘cost per hour’ advertising metric.” The Financial Times. Retrieved
from http://aboutus.ft.com/2015/05/18/financial-times-rolls-out-cost-per-hour-advertising-metric/
#axzz4R6hXhGrR
Gannett. (2017) 2016 Annual Report. Retrieved from www.annualreports.com/HostedData/AnnualRe
ports/PDF/NYSE_GCI_2016.pdf
183
Graber, D. A. (1980) Crime News and the Public. Westport, CT: Praeger.
Hawkins, A. J. (2017, April 19) “German media company Axel Springer buys a stake in Uber.” The Verge.
Retrieved from www.theverge.com/2017/4/19/15360180/uber-axel-springer-stake-investment
Hayes, A. (2016, May 7) “Fairfax boss Hywood says it is ‘inevitable’ SMH and The Age will close weekday print
editions.” Mumbrella. Retrieved from https://mumbrella.com.au/fairfax-boss-hywood-says-closing-its-
metro-media-business-would-only-cost-company-150m-365310
Heider, D., McCombs, M. and Poindexter, P. M. (2005) “What the public expects of local news: Views on
public and traditional journalism.” Journalism and Mass Communication Quarterly, 82(4), 952–967.
Hills, R. L. (1989) Power From Steam: A History of the Stationary Steam Engine. Cambridge: Cambridge Uni-
versity Press.
Krumsvik, A. H., Kvale, K. and Pedersen, P. E. (2017) “Market structure and media innovation policies in
Norway.” In H. Van Kranenburg (ed.), Innovation Policies in the European News Media Industry: A Compara-
tive Study. Cham, Switzerland: Springer (pp. 149–160).
Kurlansky, M. (2003) Salt: A World History. New York, NY: Penguin Books.
Lund, A. B. (2016) “A stakeholder approach to media governance.” In G. Ferrell Lowe and C. Brown (eds.),
Managing Media Firms and Industries: What’s So Special About Media Management? Cham, Switzerland:
Springer (pp. 103–120).
Mathieson, C. (2013,May 1) “Fairfax chief Greg Hywood sizes up the end of papers.”The Australian. Retrieved from
www.theaustralian.com.au/business/media/fairfax-chief-greg-hywood-sizes-up-the-end-of-papers/
news-story/ae1682538bf09004a3a4dfd0c29d16cf
Morgan Stanley. (2010) Internet Trends. Retrieved from http://kpcbweb2.s3.amazonaws.com/files/82/Inter
net_Trends__Jun_2010_(CM_Summit).pdf?1401210348
Myllylahti, M. (2015) No More Free News for All? A Study of How the Business Models of APN and Fairfax
Have Evolved towards Paywalls. PhD thesis, Auckland University of Technology, Auckland, New Zealand.
Myllylahti, M. (2017) “Newspaper paywalls and corporate revenues: A comparative study.” In B. Franklin and S.
A. Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. Oxon: Routledge (pp. 166–175).
Newman, N. (2016) Reuters Institute Digital News Report 2016. Reuters Institute for the Study of Journal-
ism, Oxford.
The New York Times. (2014) Innovation Report. Retrieved from www.presscouncil.org.au/uploads/52321/
ufiles/The_New_York_Times_Innovation_Report_-_March_2014.pdf
The New York Times. (2017) Journalism that Stands Apart: The Report of the 2020 Group. Retrieved from
www.nytimes.com/projects/2020-report/
Nord, L. and Nygren, G. (2002) Medieskugga [Media Shadows]. Stockholm: Atlas.
NRS. (2017a) “NRS Jan–Dec 16 fused with comScore Nov2016.” National Readership Survey. Retrieved
from www.nrs.co.uk/downloads/padd-files/pdf/nrs_padd_jan16_dec16_newsbrands.pdf
NRS. (2017b) “NRS PADD (Jan–Dec 16 integrated with comScore Nov2016) Monthly audience estimates
for mobile (smartphone and tablet).” National Readership Survey. Retrieved from www.nrs.co.uk/down-
loads/mobile-data/xlsx/nrs_padd_mobile_standard_tables_jan_dec16.xls
Perry, M. J. (2012, February 26) “Newspaper ad revenues fall to 60-yr. low in 2011.” Carpe Diem blog.
Retrieved from http://mjperry.blogspot.de/2012/02/newspaper-ad-revenues-fall-to-50-year.html
Pew Research Center. (2015, April 29) State of the News Media 2015. Retrieved from www.journalism.org/
files/2015/04/FINAL-STATE-OF-THE-NEWS-MEDIA1.pdf
Pew Research Center. (2016, June 15) State of the News Media 2016. Retrieved from http://assets.pewre
search.org/wp-content/uploads/sites/13/2016/06/30143308/state-of-the-news-media-report-
2016-final.pdf
Picard, R. G. (2010) Value Creation and the Future of News Organizations: Why and How Journalism Must
Change to Remain Relevant in the Twenty-First Century. Lisbon: Media XXI.
Picard, R. G. (2011) Mapping Digital Media: Digitization and Media Business Models, Reference Series No. 5.
London: Open Society Foundations Media Programme.
Picard, R. G. (2014) “Twilight or new dawn of journalism? Evidence from the changing news ecosystem.”
Picard, R. G. (2016) “Funding digital journalism: The challenges of consumers and the economic value
of news.” In B. Franklin and S. A. Eldridge II (eds.), Routledge Companion to Digital Journalism Studies.
London: Routledge (pp. 147–154).
Postmedia. (2016) 2016 Annual Report. Retrieved from www.postmedia.com/wp-content/uploads/2016/
11/2016-Annual-Report-FINAL.pdf
Rapaille, C. (2006) The Culture Code: An Ingenious Way to Understand Why People Around the World Live and
Buy as They Do. New York, NY: Crown Business.
184
Sabbagh, D. (2011, June 16) “Guardian and Observer to adopt ‘digital-first’ strategy.” The Guardian. Retrieved
from www.theguardian.com/media/2011/jun/16/guardian-observer-digital-first-strategy
Schibsted. (2017) Schibsted Media Group Annual Report 2016. Retrieved from http://hugin.info/131/R/
2096898/793519.pdf
Schibsted. (n.d.) The Schibsted Classified World. Retrieved from www.schibsted.com/en/Marketplaces/
Thurman, N. (2017) “Newspaper consumption in the mobile age: Re-assessing multi-platform perfor-
mance and market share using ‘time-spent’.” Journalism Studies. doi:10.1080/1461670X.2017.1279028
Trilling, D. and Schoenbach, K. (2013) “Skipping current affairs: The non-users of online and offline news.”
European Journal of Communication, 28(1), 35–51.
Turvill, W. (2016, October 19) “Independent profitable for first time in 23 years after print closure, says owner.”
Cityam.com. Retrieved from www.cityam.com/251773/independent-profitable-first-time-23-years-
after-print
Ulrich, K. and Eppinger, S. (2015) Product Design and Development (6th ed.). New York, NY: McGraw-Hill
Education.
van der Wurff, R. and Schoenbach, K. (2014) “Civic and citizen demands of news media and journalists:
What does the audience expect from good journalism?” Journalism & Mass Communication Quarterly,
91(3), 433–451.
Waldman, S. (2011) The Information Needs of Communities: The Changing Media Landscape in a Broadband
Age. Working Group on Information Needs of Communities. Federal Communications Commission.
Retrieved from https://transition.fcc.gov/osp/inc-report/The_Information_Needs_of_Communities.pdf
Ward, M. (2017, February 22) “Fairfax Media confirms domain separation with real estate business
to list separately on ASX.” Mumbrella. Retrieved from https://mumbrella.com.au/fairfax-media-
confirms-domain-separation-real-estate-business-list-separately-asx-428114
185
14
PRECARIOUS E-LANCERS
Freelance journalists’ rights, contracts, labor
organizing, and digital resistance
Errol Salamon
Freelance journalists are self-employed workers that print journalism companies have long con-
tracted to do short-term assignments or piece work (Bibby, 2014: 11–12; Cohen, 2016; Salamon,
2016a). Yet between 1994 and 1996, companies in the United States, the United Kingdom, and
Canada pioneered a new type of contract for freelance contributors. This contract has typically
demanded that freelancers waive their moral rights and assign their copyrights to the company –
rights that by default are theirs, as enshrined in copyright legislation (Salamon, 2016a: 986).
The International Federation of Journalists (IFJ, 2016) explains that in general, these contracts
demand that freelance contributors
assign to the publisher a worldwide, exclusive right to use, reproduce, display, modify, and
distribute [their] work in all types of platform, known or future [and] allow the publisher to
transfer [their] works to third parties without additional payment to the author and exploit
[their] works in any way the publisher deems necessary.
With these contracts, corporations have thus had the potential to increase revenues and profits while
decreasing the potential for freelance contributors to resell their work and maximize their income. In
response, freelance journalists have taken collective action, adopting three key tactics to resist rights-
grabbing contracts: class-action lawsuits, boycotts, and strikes. They have used digital communica-
tions as a tool to facilitate the two latter tactics. In this chapter, I outline a history of these contracts
and the digital resistance of freelancers’ labor organizations, situating them within a broader political
economy of print journalism, freelance employment, and labor organizing in a digital age.
Long underrepresented in studies on the political economy of journalism, scholars have built
a burgeoning body of literature on freelance journalism and labor organizing in recent years.
Researchers have considered the professional role perceptions, labor process, and working condi-
tions of freelance journalists (Cohen, 2016; McKercher, 2014; Salamon, 2016a). On the one hand,
journalists may willingly choose freelancing to gain relative control over the labor process, with
more autonomy, flexible employment relationships, and increased job satisfaction. On the other
hand, they may earn relatively low incomes, lack social security benefits and access to ongoing
labor protections, lose control over their work, and feel forced to accept freelance contracts if
they are unable to secure stable employment – “forced lancing” (Bittner, 2011: 13) by “reluctant
freelancers” (Bibby, 2014: 16).
186
Precarious e-lancers
Corresponding to these markers of precarity – nonstandard, insecure, and underpaid work –

Malone and Laubacher (1998) described the growth of a ‘new economy’ starting in the 1990s,
mediated by digital communications. They called it an ‘e-lance economy’, the fundamental unit
of which was the ‘individual’. In this e-lance economy, freelancers carry out tasks autonomously:
“These electronically connected freelancers – e-lancers – join together into fluid and temporary
networks to produce and sell goods and services. When the job is done . . . the network dis-
solves, and its members become independent agents again”. Conversely, Salamon (2016a: 995)
has developed the concept of ‘e-lancer resistance’, arguing electronically connected ‘e-lancers’
form temporary networks to not only sell goods and services but also to resist company demands.
Expanding on the concept of e-lancer resistance, this chapter considers the concept of alter-
native communication to examine how freelancers struggle to resist rights-grabbing contracts in
the US, Canada, and the UK. I focus on how digital communications tools can be understood as
a form of alternative communication that freelance journalists’ labor organizations use as cam-
paigning tools to communicate freelancers’ struggles to the public and pressure print media cor-
porations to rescind rights-grabbing contracts. By surveying freelance labor organizations’ email
lists, websites, and social media campaigns in the three countries, I demonstrate evidence of what
Bettig (1996: 235, 238) describes as freelance journalists’ “industrial struggles over copyrights”
with print media companies to (1) maintain control of their works and (2) protect employment
conditions in journalism by relying on digital communications to “resist the law of copyright”.
As alternative communication practices, freelance journalists’ digital communications campaigns
bring together “forms of communication” and “cultural phenomena with [freelance workers’]
experiences of struggle” (Mattelart, 1980: xviii).
This chapter adopts a radical political economy of communication approach to intellectual
property. This approach draws attention to three key tendencies: first, the economic structure of
the print media industries and the trend toward ownership concentration (Bettig, 1996: 33; Bettig,
1997: 139); second, the effects of the logic of capital on ownership and control of copyrights; and
third, the “contradictions and forms of resistance” within the print media industries (Bettig, 1996:
33). This chapter contributes primarily to research in the third category. To set the scene, I first
briefly describe the state of print journalism industries in a digital economy, foregrounding owner-
ship and control of electronic copyrights, and outline the employment conditions and labor orga-
nizations of freelance journalists. I then examine freelance contributors’ digital resistance campaigns,
situating them within their broader economic structures by linking them to the corporate profiles
of the print media companies with which freelance journalists have been in dispute. Accordingly,
I consider each company’s ‘political profile’ (e.g., corporate ownership and labor) and ‘economic
profile’ (e.g., financial data, market share, and corporate properties) (Birkinbine et al., 2016: 6–7). I
rely on print journalism industry reports, documentary sources from print journalism companies
and journalists’ labor organizations, and a critical review of the published body of literature on free-
lance journalism labor. Conducting a labor union standpoint analysis (Salamon, 2017b), I identify
common themes stemming from the digital resistance campaigns that help to articulate why and
how these campaigns communicate freelance journalists’ struggles.
Print journalism industries and employment conditions

in a digital economy
Newspaper and magazine companies have relied on advertising and circulation revenues since
the nineteenth century (Bibby, 2014: 2), but revenues decreased consistently in the twenty-first
century: from $215,331 million in 2008 to $129,873 in 2017 in the global newspaper industry
(IBSISWorld, 2017: 25) and from $99,710 million in 2007 to $83,679 million in 2016 in the
187
Errol Salamon
global magazine industry (IBSISWorld, 2016: 26). Some researchers have concluded that print
journalism is in crisis due to this downward economic trend and to the growth of a digital
economy since the World Wide Web was popularized after 1993 (Gasher et al., 2016; Siles and
Boczkowski, 2012). However, others have not been ready to make such “catastrophic claims”
(Winseck, 2011: 45) about the “the death of the newspaper” (OECD, 2010: 6), calling it a “myth”
because print media ownership groups are still profitable (Edge, 2014). The average profit mar-
gin was 7.8% in 2012 for the global newspaper industry and increased to an estimated 8.2% in
2017 (IBSISWorld, 2017: 16). The profit margin was 9.1% for the global magazine industry in
2011 and increased to 11.4% in 2016 (IBSISWorld, 2016: 17). Nevertheless, in efforts to reduce
organizational spending, print media companies have incorporated web and mobile platforms,
shortened print publication cycles, reduced in-house local, regional, or foreign reporting, or
closed publications altogether (Salamon, 2016a: 981). Coinciding with these trends, print media
companies have attempted to tighten their long-standing control over copyright of journalistic
content, while benefiting from freelance employment relationships.
By 1993, freelance journalists had discovered that print media publications had been reproduc-
ing their works en masse in electronic databases without providing them with additional com-
pensation (Salamon, 2016a: 985–986). Cvetkovski (2013: 22) views this “corporate ownership
of copyright” as an “entrepreneurial copyright”, the intent of which is to “preserve and protect
copyright embedded in a particular product for as long as possible if it is deemed commercially
viable”. Bettig (1996: 97) explains, “copyright owners benefit from long-term legal protection that
allows them to recycle their properties through existing as well as emergent media forms and to
continue earning royalties”.
In response, Jonathan Tasini, president of the U.S.-based National Writers Union (NWU), filed
a class-action lawsuit in 1993 on behalf of 10 freelance writers against five publishing companies,
including the New York Times Co. and Time Warner (Tasini v. New York Times) (Salamon, 2016a:
986). In 2005, Tasini won a settlement of $18 million. In Canada, Heather Robertson launched a
class-action lawsuit in 1996 with the support of the Professional Writers Association of Canada
(PWAC) against the Globe and Mail, a Thomson-owned daily newspaper (Robertson v. Thomson
Corp.). The company had published Robertson’s works in three electronic databases without her
consent (Cohen, 2016: 108). In 2009, Robertson won a settlement of $11 million. Robertson
launched a second class-action lawsuit in 2003 against five electronic publishing and print media
companies, obtaining a settlement of $7.9 million in 2011 (Robertson v. ProQuest, CEDROM, Toronto
Star Newspapers, Rogers and CanWest).
While these freelance journalists won, research suggests that their class-action lawsuits led media
companies to introduce rights-grabbing contracts (Salamon, 2016a: 986). By 1994 in the US, Condé
Nast, Time Inc., Readers Digest Association, and Hearst were demanding that their freelance con-
tributors sign written contracts that would assign their copyrights to the companies without pro-
viding the contributors with additional royalties. In the UK, EMAP (now Ascential) introduced a
contract in January 1995 that demanded that freelance contributors to assign their copyrights to
the publishing company. Canada’s Globe and Mail asked numerous freelancers sign written contracts
starting in February 1996, which the company revised in December 1996 to expand the scope of
its electronic rights clause: “for perpetual inclusion in the internal and commercially available data-
bases and other storage media (electronic and otherwise) of The Globe or its assignees and products
(electronic and otherwise) derived therefrom” (cited in D’Agostino, 2005: 189). Publishing compa-
nies arguably introduced these contracts to prevent further class-action lawsuits.
Since the turn of the twenty-first century, researchers have used the concept of ‘precarity’ to
describe employment conditions and workers’ labor organizing responses to broader economic
trends (Bibby, 2014: 8; Cohen, 2016; Walters et al., 2006: vi). Research suggests that there are three
key markers of precarity in print journalism industries. The first marker is a decrease in total
188
number of jobs. The second is a relative increase in ‘atypical work’ or nonstandard employment
relationships, which refers to employment that is “not permanent and/or full-time” (Walters et
al., 2006: 6). Freelancing is one subcategory of atypical work, which also includes subcontracted,
casual, or temporary work (Bibby, 2014: 11–12; Cushion, 2007: 122). The third marker is rela-
tively low pay and “earning low income” (LCO, 2012: 11, 27). These markers of precarity are
interrelated. As the OECD (2015: 135) puts it, “non-standard jobs tend to pay lower wages than
standard jobs, especially at the bottom of the earnings distribution”. Between 1995 and 2013,
more than half of all jobs that were created in OECD countries were nonstandard jobs (OECD,
2015: 29–30). By 2013, nonstandard work represented about one-third of total employment in
OECD countries. Between 1985 and 2013, income inequality increased by about 10% in 17 of
22 OECD countries (OECD, 2015: 23).
Under such precarious employment conditions, journalism faces what Cohen (2016: 22)
calls a “precarity penalty”: “The current organization of freelance work inserts individuals into
competitive and unequal social relations and infuses their work and their lives with insecurity”.
In effect, workers may feel pressured to produce journalistic works that can be not only “‘sold”
but also “produced quickly, making the small fees paid per word, per article, or per hour worth a
freelancer’s time” (Cohen, 2016: 22). According to a 2006 survey of 41 IFJ member organizations
in 38 countries, about 75% of journalists thought that their employment conditions impacted
journalistic content (Walters et al., 2006: 4). Bibby (2014: 30) suggests, then, that “more precari-
ous employment conditions will direct[ly] affect the quality of journalism” and “the cultural and
democratic life of the societies [journalists] serve”.
Freelance journalists’ labor union organizing

In addition to signifying the employment conditions mentioned earlier, precarity refers to the
“potentialities that spring from workers’ own refusal of labor” to exert control over their employ-
ment conditions (Neilson and Rossiter, 2005). In other words, precarity captures both corporate
“restructuring” and worker “recomposition” (Dyer-Witheford, 1999: 66). In this sense, precarious
work leads workers to adopt refusal tactics, disrupting the process of normalizing precarity with a
“process of recomposing work and life” (Lorey, 2010). To facilitate this process, journalistic workers
have long formed or joined labor organizations. Trade unions have advocated for higher wages as
well as stable and permanent employment, helping journalistic workers secure ongoing labor protec-
tion. However, unions have struggled to organize freelance journalists and bargain collectively due to
barriers in labor legislation (in the US and Canada) and because freelancers tend to work for mul-
tiple employers, dispersed across geographical space (Salamon, 2016a: 984). While trade unions have
experimented with ways to represent freelance workers, freelance workers have also established ‘alt-
labor’ organizations. According to Cohen (2016: 185), alt-labor organizations “mobilize and repre-
sent workers who under law cannot access trade unions for the purpose of collective bargaining”.
In the UK, the National Union of Journalists (NUJ) has been open to freelance journalists since
it was formed in 1907 (Bohère, 1984: 151; Cohen, 2016: 183–184) and established a special free-
lance branch in 1951. By 1973, the NUJ had secured agreements on freelance rates with publishers,
among them the Manchester Evening News. The NUJ included a special section in such agreements
for freelance journalists that outlined minimum pay rates at the national level, but the rate increases
were not linked to increases in the minimum pay rates of staff journalists. The NUJ’s freelance
members went on strike against IPC Media (now Time Inc. UK) over pay rates in 1981, becoming
the union’s only freelance journalists’ strike. IPC employees who were NUJ members refused to
take the freelance work, helping the freelance members win a pay rate increase of more than 20%.
In recent years, the NUJ has secured collective agreements that cover the use of freelance contribu-
tors’ works as casual workers on employers’ premises as well as an agreement with the Guardian
189
Errol Salamon
Media Group that stipulates minimum pay rates for freelance contributors (Bibby, 2014: 18). The
NUJ’s recent membership data provide further evidence of two of the key markers of precarity, as
mentioned earlier: a decrease in the total number of members and an increase in the proportion
of freelance members (Bibby, 2014: 13). Between 2005 and 2012, the NUJ’s total number of full
members decreased from 28,678 to 26,521. Over this period, total freelance members increased
from 6,985 to 7,334, accounting for 24.3% of members in 2005 and 27.6% of members in 2012.
Freelance workers in Canada and the US have also joined trade unions. For instance, the
Canadian Media Guild (CMG) represents print, broadcast, and online media workers and was
originally known as the Canadian Wire Service Guild when it was formed in 1949 (CMG,
2017). The CMG is the 6,000-member union Local 30213 of the Communications Workers
of America (CWA) Canada. The local established the CMG Freelance Branch in 1998 due to
the “increasing amount of freelance and temporary media employment” (CMG, 1999). The
CMG Freelance Branch’s aim is “to improve the terms and conditions of such work, not only
for the sake of freelancers, but also to protect regular employees from being undercut” (CMG,
1999). The Freelance Branch represents 600 freelance workers at the Canadian Broadcasting
Corporation, the country’s public broadcaster, and they are the only freelance journalism workers
in Canada to be covered by a collective agreement (Cohen, 2016: 217, 221–222). In 2014, the
Freelance Branch launched a “paid membership” program for non-CBC independent freelancer
members for whom the union cannot bargain collectively, and by July 2017, CMG Freelance had
183 “paid subscribers” (CMG Freelance, 2017). The CMG advocates for freelance workers’ rights
and offers members advice on negotiating contracts, holds workshops and social events, provides
access to online training and webinars, and offers group health insurance benefits.
In the US, there are two types of unions for freelance journalists. First, the NWU, an all-writers’
union, was formed in 1981. At the “Why a Union?” workshop during the Nation Institute’s
Writers’ Congress, 3,000 writers endorsed a proposal to “create a union for writers in all genres
to actively press for better pay and treatment” (NWU, 2017). Writers initially organized eight local
chapters across the US, and there are now 13 chapters (Cohen, 2016: 178; NWU, 2017). Later in
1981, the Washington, DC, local made agreements with Black Film Review and Musician magazine,
becoming the union’s first and second agreements on freelance terms. In 1983, the NWU was
officially chartered, with members ratifying a constitution that guaranteed local chapter autonomy.
In 1991, NWU joined the United Auto Workers (UAW) because it was “known for its progressive
history and its commitment to organizing non-traditional workers such as lawyers and graduate
students” (NWU, 2017). The NWU became UAW Local 1981. One year later, the union launched
health and libel insurance plans, which led to a 20% increase in membership (NWU, 2017). In 1995,
two years after launching Tasini’s copyright class-action lawsuit, the NWU published a Guide to
Electronic Rights. After winning the lawsuit in 2001, the union’s membership increased to 7,300. In
addition to regularly holding conferences and events, the NWU has offered its members webinars
since 2010 on topics such as negotiating contracts and protecting copyright.
Second, two traditional journalists’ unions in the US have bargaining units for freelance work-
ers: first, the Pacific Media Workers Guild (PMWG) – NewsGuild Local 39521 of the CWA,
which was chartered as the Newspaper Guild of Northern California in 1936 (PMWG, 2017),
and second, the Chicago News Guild (CNG) – NewsGuild Local 34071 of the CWA, which
was chartered as the Chicago Newspaper Guild in 1936 (Kritzberg, 1973: 412). In 2009, the
PMWG established a unit called Guild Freelancers (Cohen, 2016: 182; PMWG, 2011). The Guild
Freelancers’ aims are to provide freelance journalists with services and referrals and “build soli-
darity between staff journalists and freelancers” (PMWG, 2011). The unit advocates for freelance
workers, issues press credentials, provides contract advice, assists with grievances, and offers pro-
fessional development and benefits. In 2013, the CNG formed a bargaining unit called Working
Journalists to support freelance media professionals after the Sun-Times Media Group laid off
190
photojournalists (Cohen, 2016: 182; CNG, 2017). Working Journalists offers press credentials,
payment collection assistance, networking opportunities, and professional development.
E-lancers’ digital communications campaigns to resist

rights-grabbing contracts
Through their collective organizations, freelance journalists have used digital communications
to facilitate two tactics to resist rights-grabbing contracts: boycotts and strikes (Salamon, 2016a).
Freelance journalists’ organizations have exploited digital communications to recruit and organize
freelancers (Bibby, 2014: 22). These tools illuminate what Dyer-Witheford (1999: 72) describes as
the contradictory forces that shape digital technologies: the “interaction between business’s drive
to extend commodification and democratic aspirations for free and universal communication”.
As publishing companies have used digital technologies to exploit freelancers’ works and expand
capital, some freelance workers report that new technologies such as social media have made it
easier to find freelance work (Edelman Intelligence, 2016: 44). Yet the campaigns of journalists’
organizations also draw attention to the “process of deconstructing and reconstructing technolo-
gies as itself part of the movement of the struggle against capital” (Dyer-Witheford, 1999: 72).
While boycotts and strikes are not new tactics, digital communications have created opportu-
nities for workers to broaden the “scope and focus of collective action”, launching single-issue
campaigns and “cyber-picketing” (Diamond and Freeman, 2002: 592). Workers may conduct
this ‘e-disruption’ over issues such as pay rates via email campaigns and host these campaigns on a
website with a distinct organizational identity that is separate from an individual established trade
union. These distinct organizational identities are evidence of “alt-labor” (Cohen, 2016: 185) or
“alternative institutions of interest representation” (Saundry et al., 2007: 178–179, 181). In doing
so, workers form “virtual unions”: “minority unions that exist on the web but lack company
recognition” (Diamond and Freeman, 2002: 579). Virtual unions are examples of “temporary
labor convergence”, a short-term and micro-level campaigning tactic that trade unions adopt to
organize journalistic workers, mobilizing massive community support (Salamon, 2017b). How-
ever, these temporary virtual unions must struggle to translate temporary digital convergence into
“offline actions” (Fowler and Hagar, 2013: 224).
As print journalism companies have issued new contracts, attempting to secure tighter control
over freelance contributors’ digital rights and exploit them for additional profit, freelance journal-
ists have created online bulletin boards and email lists to discuss these contracts and resist them
(Salamon, 2016a: 989). These “informal channels of communication”, Cohen (2016: 177) explains,
“are a low-effort way to break the isolation of working alone and to gain a sense of community.
They are also useful for advice on stories, contracts, and careers”. In the UK, freelance journalists
pioneered a digital communications tool to organize workers even before the Mosaic web browser
helped to popularize the internet in 1993 (Salamon, 2016b). In 1992, the Freelance Branch of the
NUJ launched a digital communications network called NUJnet, which initially attracted hun-
dreds of users. According to NUJnet co-creator Mike Holderness, the union created NUJnet to
help alleviate freelance isolation and give union members direct and rapid access to union repre-
sentatives. NUJnet provided subscribers with email service and access to 15 private bulletin boards
for advice and discussion, as well as access to commercial bulletin boards and online databases. One
bulletin board provided news of labor conflicts, while another bulletin board became a platform for
NUJnet members to discuss the emerging issue of digital copyright and rights-grabbing contracts.
Members also participated in international cyber actions in support of press freedom. Freelance
workers could thus also use digital communications “to develop a collective critique about the
condition of freelance writing [and photography] as an occupation and for recognizing and naming
the power relations that structure the industry” (Cohen, 2016: 177).
191
Errol Salamon
In Canada and the US, freelance workers developed digital communications tools in the late
1990s. In Canada, the CMG launched an email newsletter in 1999 to communicate information
to its members, complementing the union’s website and print publications (Salamon, 2016b).
Around this time, Toronto-based freelance writers David Hayes, Alex Gillis, and Jess Ross created
the “Toronto Freelance Editors and Writers” email list to build a community support group.
The email list currently has more than 900 members. In the US, the NWU launched its first
online “Member Networking Directory” in 1999, connecting members across the US; 440
members have uploaded short biographies and contact information (NWU, 2017). Starting in
2005, the NWU committed to using the internet as the primary mode of communicating with
its members, issuing regular e-bulletins, and later the NWUsletter, the union’s current e-newsletter.
Accordingly, with the advent of digital communications, “[f]reelancers with common interests
and problems could communicate freely, cheaply and in far greater numbers than would ever
make it to a formal meeting” (Sutcliffe et al., 2000: 23).
Yet, as suggested earlier, freelance workers have formed digital networks to not only “spread
information rapidly” but also “inspire resistance” (Sutcliffe et al., 2000: 24). By 1999, freelance
journalists were using digital communications tools to facilitate contract boycotts and strikes in
the US and Scotland (Bittner, 2011: 16). In the US, freelance photojournalists in the San Francisco
Bay Area at the McGraw-Hill-owned Business Week (now Bloomberg Businessweek) launched an
alt-labor organization – an email list and online discussion group in April 1999 called “Editorial
Photographers” (EP) to boycott rights-grabbing contracts and what they considered to be under-
paid work (Salamon, 2016a: 989). While Business Week was imposing rights-grabbing contracts,
McGraw-Hill generated nearly $4 million in revenues and $0.8 million profits in 1999, increasing
from $3.5 million and $0.6 million, respectively, in 1997 to $3.7 million and $0.7 million, respec-
tively, in 1998 (McGraw-Hill Companies Inc., 2000: 3–4). At the time of the contract dispute,
McGraw-Hill had 16,376 employees. McGraw-Hill was the 12th-biggest magazine publisher
in the US in 1996, with 1.9% of total industry revenue share, but it became the seventh-biggest
publisher by 2001, obtaining 2.7% of total revenues (Noam, 2016a: 506–508). The e-disruption
helped the Business Week photojournalists to secure a satisfying pay increase and additional com-
pensation so that the company could reprint their photographs in foreign and online editions of
the publication and in advertisements (Salamon, 2016a: 989).
In Scotland, about 40 freelance photographers formed an alt-labor organization in November
2000 called the Scottish Newspapers’ Association of Photographers (SNAP) to organize a campaign
to resist a rights-grabbing contract at Business a.m., a Bonnier Group-owned daily newspaper (Sal-
amon, 2016a: 990). Founded in 1804, the Swedish-based Bonnier Group launched Business a.m.
in September 2000 with 70 staff journalists (Doward, 2000). Bonnier Group operated magazines,
books, television and radio stations, newspapers, and film production in 17 European countries.
Although Bonnier Group was imposing rights-grabbing contracts, as mentioned earlier (Salamon,
2016a: 990), the company’s global revenues, profits, and number of employees had increased: the
company generated 16.8 billion (SEK) in revenues in 2000, an increase from 14.8 billion (SEK)
in 1999; it made 1.1 billion (SEK) in profits in 2000, an increase from 807 million (SEK) in 1999;
and it had 10,700 employees in 2000, an increase from 9,000 in 1999 (Bonnier AB, 2000, 2001).
If contributors signed the contract, Bonnier Group would have been able to reuse their photo-
graphs without providing them with additional compensation (Salamon, 2016a: 990). In response,
the SNAP created an email list to organize this campaign, exchanging information and discussing
ways to improve their employment conditions. After campaigning for five months, Bonnier Group
amended the contract, protecting contributors’ copyrights and guaranteeing them a reprint fee.
Since the turn of the twenty-first century, freelance journalists have also been using social
media to resist rights-grabbing contracts in the UK and Canada. In the UK, 200 freelance writers
and photographers went on strike in April 2010 at three Bauer Media-owned music magazines:
192
Kerrang!, Q, and Mojo (Salamon, 2016a: 990–991). The German-based Bauer Media Group acquired
the magazines from EMAP in 2008 (Bauer Media Group, 2016; Noam, 2016b: 1116–1119). Bauer
Media also had interests in magazines, digital media outlets, and television and radio stations in the
UK, Germany, and Russia, among other countries. Bauer Media was the fourth-biggest magazine
publisher in the world in 2011, generating $2.9 billion in global revenues. Since then, Bauer Media’s
global revenues have totaled about $2.7 billion in 2013 and 2015, while the company’s global
workforce increased by 6.5% to 11,500 employees between 2014 and 2015. Bauer Media was the
second-biggest magazine publisher in the UK in 2008, obtaining 20% of the country’s total indus-
try revenues, and it became the country’s biggest magazine publisher in 2011, with 26% of revenues
(Iosifidis, 2016: 432). By signing the magazine contracts, the freelance contributors would waive
their moral rights and the company would be able to reuse the contributors’ works on all media
platforms without having to pay them additional compensation (Salamon, 2016a: 990–991). In
response, the freelance contributors temporarily converged, creating a Twitter account called Bauer
Contract Help. They used Twitter to communicate news and signs of support for the resistance and
called on Bauer Media to rescind the contract. However, Bauer Media reinforced the terms of the
contract in June 2010, so the freelance contributors terminated their strike.
In Canada, journalists’ labor organizations launched two separate social media campaigns in
2009 and 2013 to resist the freelance contributors’ contracts of publishing conglomerate TC
Transcontinental. Founded in 1976, by 2009, TC Transcontinental operated magazines, newspa-
pers, and printing facilities in Canada, the US, and Mexico with 12,500 employees, and about
9,000 employees in 2013 (TC Transcontinental, 2016). TC Media, the company’s newspaper
and magazine division, generated $607 million in revenues and a $93.3 million profit in 2009,
compared to $712 million and $40 million, respectively, in 2013 (TC Transcontinental, 2010: 49;
Salamon, 2015: 445). TC Media was the second-biggest magazine publisher in Canada, with
6.2% of total industry revenues in 2008 and 2010, while Rogers was the biggest company, with
6.3% of market share (Winseck, 2016).
In September 2009, the “Bad Writing Contracts” coalition of 14 writers’ organizations con-
verged to resist TC Transcontinental’s new author master agreement (Salamon, 2016a: 991–992;
Salamon, 2017a). The organizations included the PWAC and the Canadian Writers Group (a
writers’ agency that partnered with the CMG at the time). By signing the publishing contracts,
freelance contributors would assign their copyrights to the company and waive their moral rights
that give contributors the right of attribution and protect the integrity of their works. The writ-
ers’ coalition sought a contract that would give them a fair share of the revenue that the company
generated from their works. To raise awareness and communicate information about the cam-
paign, the coalition launched a website, Twitter account, and Facebook group.
To resist another TC Transcontinental agreement in February 2013, the CMG and L’association
des journalistes indépendants du Québec launched the #nesignezpas (#dontsign) hashtag cam-
paign and a Facebook group called “Back Off, Elle Canada and Canadian Living Publisher”
(Salamon, 2016a: 992; Salamon, 2017a). Campaign organizers also created a private and invitation-
only Facebook group to organize freelance journalists who feared company retribution. By Sep-
tember 2013, TC Media released a revised contract in the province of Quebec. While labor
organizers attributed the campaign’s success in large part to the use of digital communications
tools, they also recognized the limits of using social media as a broadcasting tool in a public digital
space. Still, the #nesignezpas, EP, and SNAP campaigns are noteworthy because they illuminate
the potential to link temporary labor convergence and online actions to “offline actions” and
organizational change (Fowler and Hagar, 2013: 224). Such digital communications campaigns
have been important for journalists’ unions, giving them the potential to “bypass official media
‘gatekeepers’” and making them “less dependent on mass media in building ties to the public”
(Drew, 2013: 115, 118).
193
Errol Salamon
Conclusion
In a digital economy, big print journalism conglomerates in the US, UK, and Canada have con-
tinued to generate revenues and remain profitable. Big corporations such as McGraw-Hill, Bon-
nier, Bauer Media, and TC Transcontinental have issued contracts to freelance contributors that
have served as an instrument of corporate control and wealth generation in the cycle of capital
accumulation. With these contracts, media companies have the potential to sustain their control
and ownership of freelance contributors’ rights and to make more profit. In adopting this cor-
porate strategy, these companies have devalued the labor of their freelance contributors, limiting
their potential to resell their works to other publishers and supplement their income.
As low-paid and nonstandard journalistic work has become commonplace, freelance journal-
ists have used digital communications tools to resist rights-grabbing contracts. Through their
websites, email lists, and social media campaigns, freelance journalists’ traditional and alt-labor
organizations have attempted to resist the law of copyright (Bettig, 1996). The research con-
ducted for this chapter suggests that the digital communications campaigns of journalists’ orga-
nizations are evidence of alternative communication practices (Mattelart, 1980). An alternative
communication practice helps us understand how journalists’ organizations use digital technolo-
gies to express and circulate e-lancers’ experiences of struggle. As Mattelart (1980: xviii) would express
it, these campaigns bring together “forms of communication” and “cultural phenomena with
[freelancers’] experience of struggle”. While there is only some evidence that digital commu-
nications campaigns can help e-lancers secure fair contracts, the research demonstrates these
campaigns could at least create the conditions for “publicization”, showing how e-lancers could
gain visibility in their struggles over rights (Salamon, 2017a: 993). In this context, journalists’
organizations may be more successful if they use digital technologies as one part of broader orga-
nizing initiatives rather than foregrounding them in efforts to resist rights-grabbing contracts.
As Cushion (2007: 127) puts it: “The role of unions, in different countries, is critical to ensur-
ing journalists are paid a fair and rewarding salary”. In a digital economy, there may be a greater
urgency for such “collective agency amongst journalists, as employers are acting unilaterally to
cut costs and undermine employment rights” (Cushion, 2007: 128).
This urgency, in turn, justifies the need for and value of conducting more labor union stand-
point research in digital journalism studies (Salamon, 2017b). More research on freelance journal-
ists’ rights, labor organizing, and resistance could help radical political economy scholars understand
how copyright ownership and control serve as a key corporate strategy in digital journalism. More
of this research could also help scholars uncover how freelance journalists and their labor orga-
nizations could gain relative control through collective action over the works that they produce
(Cohen, 2016; Salamon, 2016a: 994). In this way, the concept of the precarious e-lancer articulates
how journalists’ labor organizations not only form temporary networks to sell goods and services
(Malone and Laubacher, 1998) and resist company demands (Salamon, 2016a) but also develop long-
term labor convergence strategies to organize and protect freelance contributors (Salamon, 2017b). Yet
Cohen (2016: 22) reminds us that “collective organization in media industries is not just about
protecting individual workers, but also about ongoing efforts to democratize journalism”.
Further reading
This chapter has benefited significantly from Nicole S. Cohen’s (2016) Writers’ Rights: Freelance
Journalism in a Digital Age. Errol Salamon’s (2016a) “E-lancer Resistance: Precarious Freelance
Journalists Use Digital Communications to Refuse Rights-Grabbing Contracts” further explores
the devaluation of freelance journalism labor and freelance workers’ collective action to defend
their rights in digital journalism in North America and western Europe. Additionally, Cohen’s
194
(2016) book, Salamon’s (2016a) work, and Catherine McKercher’s (2014) “Precarious Times,
Precarious Work: A Feminist Political Economy of Freelance Journalists in Canada and the
United States” explore the gendered nature of freelance digital journalism. Andrew Bibby’s (2014)
Employment Relationships in the Media Industry offers an accessible report on atypical media employ-
ment more broadly for digital journalism scholars. Finally, Mark Spilsbury’s (2016) Report for the
National Council for the Training of Journalists explores the state of freelance journalism in the UK,
while Edelman Intelligence’s (2017) Freelancing in America and Contently’s (Baker, 2016) recent
study analyze the state of freelancing in the US for creative workers, including journalists.
References
Baker, D. (2016, July 5) “Contently study: The state of freelancing in 2016.” The Freelancer. Retrieved from
http://contently.net/2016/07/05/trends/contently-study-state-freelancing-2016/ (accessed 17 August
2017).
Bauer Media Group. (2016) Bauer Media Group Reports Increased Worldwide Sales in 2015. Retrieved from
www.bauermedia.co.uk/newsroom/press/bauer-media-group-reports-increased-worldwide-sales-
in-2015 (accessed 26 July 2017).
Bettig, R. V. (1996) Copyrighting Culture: The Political Economy of Intellectual Property. Boulder, CO: Westview
Press.
Bettig, R. V. (1997) “The enclosure of cyberspace.” Critical Studies in Mass Communication, 14(2), 138–157.
Bibby, A. (2014) Employment Relationships in the Media Industry. International Labour Office, Sector Activi-
ties Department, Geneva. Retrieved from www.ilo.org/wcmsp5/groups/public/–ed_dialogue/–sector/
documents/publication/wcms_249912.pdf (accessed 6 July 2017).
Birkinbine, B. J., Gómez, R. and Wasko, J. (2016) “Introduction.” In B. J. Birkinbine, R. Gómez and J.
Wasko (eds.), Global Media Giants. New York, NY: Routledge (pp. 1–7).
Bittner, A. K. (2011) Managing Change: Innovation and Trade Unionism in the News Industry. European Fed-
eration of Journalists, Brussels. Retrieved from www.ifj.org/fileadmin/images/EFJ/EFJ_documents/
Reports/Managing_Change_Draft_2.pdf (accessed 8 July 2017).
Bohère, G. (1984) Profession: Journalist. Geneva: International Labour Office.
Bonnier,A. B. (2000) Verksamhetsbeskrivning. Retrieved from https://web.archive.org/web/20000128195058fw_/
www.bonnier.com:80/bokstavsordning_foretag.asp?FID=275 (accessed 25 July 2017).
Bonnier, A. B. (2001) Business Profile. Retrieved from https://web.archive.org/web/20011122043401fw_/
www.bonnier.com:80/eng/bokstavsordning_foretag.asp?FID=275 (accessed 25 July 2017).
Canadian Media Guild (CMG) (1999) Freelance Branch. Retrieved from https://web.archive.org/web/
19991013053201/www.cmg.ca/freelance.html (accessed 10 July 2017).
Canadian Media Guild (CMG). (2017) About the Canadian Press Branch. Retrieved from www.cmg.ca/en/
cp/about/ (accessed 20 August 2017).
Canadian Media Guild (CMG) Freelance. (2017) How We Operate (updated Sept. 8, 2017). Unpublished.
Chicago News Guild (CNG) (2017) Working Journalists. Retrieved from www.chicagonewsguild.org/working-
journalists/ (accessed 11 July 2017).
Cohen, N. S. (2016) Writers’ Rights: Freelance Journalism in a Digital Age. Montreal and Kingston: McGill-
Queen’s University Press.
Cushion, S. (2007) “Rich media, poor journalists: Journalists’ salaries.” Journalism Practice, 1(1), 120–129.
Cvetkovski, T. (2013) Copyright and Popular Media: Liberal Villains and Technological Change. New York, NY:
Palgrave Macmillan.
D’Agostino, G. (2005) “Freelance authors for free: Globalisation of publishing, convergence of copyright
contracts and divergence of judicial reasoning.” In F. McMillan and K. Bowrey (eds.), New Directions in
Copyright Law (Vol. 1). Northampton: Edward Elgar (pp. 166–215).
Diamond, W. J. and Freeman, R. B. (2002) “Will unionism prosper in cyberspace? The promise of the
internet for employee organization.” British Journal of Industrial Relations, 40(3), 569–596.
Doward, J. (2000, September 10) “Business am: Scotland’s new daily newspaper.” The Guardian. Retrieved
from www.theguardian.com/business/2000/sep/10/theobserver.observerbusiness11 (accessed 25 July
2017).
Drew, J. (2013) A Social History of Contemporary Democratic Media. New York, NY: Routledge.
Dyer-Witheford, N. (1999) Cyber-Marx: Cycles and Circuits of Struggle in High Technology Capitalism. Chicago,
IL: University of Illinois Press.
195
Errol Salamon
Edelman Intelligence. (2016) Freelancing in America: 2016. Retrieved from www.slideshare.net/upwork/

freelancing-in-america-2016/1 (accessed 17 August 2017).
Edelman Intelligence. (2017) Freelancing in America: 2017. Retrieved from www.upwork.com/i/freelancing-
in-america/2017/ (accessed 23 October 2017).
Edge, M. (2014) Greatly Exaggerated: The Myth of the Death of Newspapers. Vancouver: New Star Books.
Fowler, T. and Hagar, D. (2013) “‘Liking’ your union: Unions and new social media during election cam-
paigns.” Labor Studies Journal, 38(3), 201–228.
Gasher, M., Brin, C., Crowther, C., King, G., Salamon, E. and Thibault, S. (eds.). (2016) Journalism in Crisis:
Bridging Theory and Practice for Democratic Media Strategies in Canada. Toronto: University of Toronto Press.
IBSISWorld. (2016) IBSISWorld Industry Report: Global Magazine Publishing. Retrieved from www.ibisworld.
com/industry-trends/global-industry-reports/manufacturing/magazine-publishing.html (accessed
18 August 2017).
IBSISWorld. (2017) IBSISWorld Industry Report: Global Newspaper Publishing. Retrieved from www.ibisworld.
com/industry-trends/global-industry-reports/manufacturing/newspaper-publishing.html (accessed
18 August 2017).
International Federation of Journalists (IFJ) (2016) Fair Contracts for Journalists Campaign. Retrieved from
www.ifj.org/campaigns/fair-contracts-for-journalists/ (accessed 30 May 2017).
Iosifidis, P. (2016) “United Kingdom.” In E. M. Noam (ed.), Who Owns the World’s Media?: Media Concentra-
tion and Ownership Around the World. New York, NY: Oxford University Press (pp. 425–454).
Kritzberg, B. (1973) “An unfinished chapter in white-collar unionism: The formative years of the Chicago
Newspaper Guild, local 71, American Newspaper Guild, A.F.L-C.I.O.” Labor History, 14(3), 397–413.
Law Commission of Ontario (LCO) (2012) Vulnerable Workers and Precarious Work: Final Report. Retrieved
from www.lco-cdo.org/en/vulnerable-workers-final-report (accessed 7 July 2017).
Lorey, I. (2010) “Becoming common: Precarization as political constituting.” E-Flux, vol. 17. Retrieved from
www.e-flux.com/journal/becoming-common-precarization-as-political-constituting/ (accessed 7 July
2017).
Malone, T. W. and Laubacher, R. (1998, September–October) “The dawn of the e-lance economy.” Harvard
Business Review. Retrieved from https://hbr.org/1998/09/the-dawn-of-the-e-lance-economy (accessed
30 May 2017).
Mattelart, A. (1980) Mass Media, Ideologies and the Revolutionary Movement. Atlantic Highlands, NJ: Humani-
ties Press.
McGraw-Hill Companies Inc. (2000) 1999 Annual Report. Retrieved from http://getfilings.com/o0000
950123-00-002550.html (accessed 25 July 2017).
McKercher, C. (2014) “Precarious times, precarious work: A feminist political economy of freelance jour-
nalists in Canada and the United States.” In C. Fuchs and M. Sandoval (eds.), Critique, Social Media and
the Information Society. New York, NY: Routledge (pp. 219–230).
National Writers Union (NWU) (2017) History and Achievements. Retrieved from https://nwu.org/about/
history-and-achievements/ (accessed 12 July 2017).
Neilson, B. and Rossiter, N. (2005) “From precarity to precariousness and back again: Labor, life and unsta-
ble networks.” Fibreculture, vol. 5. Retrieved from http://five.fibreculturejournal.org/fcj-022-from-pre
carity-to-precariousness-and-back-again-labour-life-and-unstable-networks/ (accessed 7 July 2017).
Noam, E. M. (2016a) “United States.” In E. M. Noam (ed.), Who Owns the World’s Media?: Media Concentra-
tion and Ownership Around the World. New York, NY: Oxford University Press (pp. 500–573).
Noam, E. M. (2016b) “The world’s major media companies.” In E. M. Noam (ed.), Who Owns the World’s
Media?: Media Concentration and Ownership Around the World. New York, NY: Oxford University Press
(pp. 1109–1179).
Organisation for Economic Co-operation and Development (OECD) (2010) The Evolution of News and the
Internet. Paris: OECD Publishing. Retrieved from www.oecd.org/sti/ieconomy/45559596.pdf (accessed
6 July 2017).
Organisation for Economic Co-operation and Development (OECD) (2015) In It Together: Why Less Inequal-
ity Benefits All. Paris: OECD Publishing. Retrieved from http://dx.doi.org/10.1787/9789264235120-
en (accessed 19 August 2017).
Pacific Media Workers Guild (PMWG) (2011) Guild Freelancers: About. Retrieved from https://web.archive.
org/web/20111021021644/http://guildfreelancers.org/gf/?page_id=11 (accessed 11 July 2017).
Pacific Media Workers Guild (PMWG) (2017) Who We Are. Retrieved from http://mediaworkers.org/who-
we-are/ (accessed 20 August 2017).
Salamon, E. (2015) “(De)valuing intern labour: Journalism internship pay rates and collective representation
in Canada.” tripleC: Communication, Capitalism & Critique, 13(2), 438–458.
196
Salamon, E. (2016a) “E-lancer resistance: Precarious freelance journalists use digital communications to
refuse rights-grabbing contracts.” Digital Journalism, 4(8), 980–1000.
Salamon, E. (2016b, July 5) “Freelancers helped invent digital communications tools to build online networks.”
Story Board. Retrieved from www.thestoryboard.ca/freelancers-helped-invent-digital-communications-
tools-build-online-networks/ (accessed 30 May 2017).
Salamon, E. (2017a, March 2) “New UK report highlights the challenges of freelancing.” Story Board. Retrieved
from www.thestoryboard.ca/new-uk-report-highlights-challenges-freelancing/ (accessed 9 July 2017).
Salamon, E. (2017b) “Temporary labor convergence: Newsworkers mobilize massive community support to
organize the newspaper chain, 1963–1966.” Journalism Studies. doi:10.1080/1461670X.2017.1301780.
Saundry, R., Stuart, M. and Antcliff, V. (2007) “Broadcasting discontent – Freelancers, trade unions and the
internet.” New Technology, Work & Employment, 22, 178–191.
Siles, I. and Boczkowski, P. J. (2012) “Making sense of the newspaper crisis: A critical assessment of existing
research and an agenda for future work.” New Media & Society, 14(8), 1375–1394.
Spilsbury, M. (2016) “Report for the National Council for the Training of Journalists.” Exploring Free-
lance Journalism. Retrieved from www.nctj.com/downloadlibrary/EXPLORING%20FREELANCE%20
JOURNALISM%20FINAL.pdf (accessed 7 July 2017).
Sutcliffe, P., Vaughan, J. and Macaskill, H. (2000) Battling for Copyright: Freelance Journalists Versus the Media
Conglomerates. London: National Union of Journalists.
TC Transcontinental. (2010) 2009 Annual Report. Retrieved from http://tctranscontinental.com/documents/
10180/37760/2009_Annual_Report.pdf (accessed 25 July 2017).
TC Transcontinental. (2016) History. Retrieved from http://tctranscontinental.com/en/about/history
(accessed 25 July 2017).
Walters, E., Warren, C. and Dobbie, M. (2006) “The changing nature of work: A global survey and case study
of atypical work in the media industry.” International Federation of Journalists. Retrieved from www.ilo.org/
wcmsp5/groups/public/–ed_dialogue/–sector/documents/publication/wcms_161547.pdf (accessed
6 July 2017).
Winseck, D. (2011) “The political economies of media and the transformation of the global media indus-
tries.” In D. Winseck and D. Y. Jin (eds.), The Political Economies of Media: The Transformation of the Global
Media Industries. London: Bloomsbury Academic (pp. 3–48).
Winseck, D. (2016) “Media ownership and concentration in Canada.” In E. M. Noam (ed.), Who Owns the
World’s Media? Media Concentration and Ownership Around the World. New York, NY: Oxford University
Press (pp. 455–499).
197
15
WHAT CAN NONPROFIT
JOURNALISTS ACTUALLY DO
FOR DEMOCRACY?
Journalism is considered the ‘fourth estate’, tasked with helping to uphold democracy. Despite
this weighty responsibility, there is little consensus about how its practitioners are expected to
perform this task. As Gans (1998) argued, the journalistic theory of democracy offers little clar-
ity on whether and how journalists contribute to democracy. Perhaps as a result, many U.S.
journalists have, at least until recently, felt uncomfortable discussing the impact of their work
and claimed to be agnostic about outcomes (Ettema and Glasser, 1998), often arguing that their
responsibility ends at publishing or broadcasting a story. As news organizations struggle to define
and promote their own relevance, some argue that journalists should make explicit their efforts
to contribute to democracy and the results of those efforts (Anderson et al., 2014; Keller and
Abelson, 2015).
Lately, we have begun to see journalism evolve on this front. Advancements in audience-
tracking tools help news organizations measure engagement metrics (Napoli, 2014; Stray, 2012)
of interest to advertisers and foundation funders seeking a return on their investments (Simons
et al., 2016; Tofel, 2013). This process of collecting data on audience analytics and the more
difficult-to-quantify impact – what happens when the story is published or broadcast – presents
journalists with an important opportunity to sort out which metrics matter.
With both journalism and democracy in trouble in the twenty-first century, news organiza-
tions have started to shed their veneer of impartiality, allowing and even encouraging reporters to
consider the impact of their work. National data show journalists overwhelmingly care about the
impact their stories might have (McIntyre et al., 2016). Of course, that may always have been the
case; it may be that journalists with this orientation now perceive that they have greater freedom
to openly discuss outcomes.
This chapter explores the recent evolution of American journalism’s orientation toward
impact. We focus on nonprofit news organizations, which have been more explicit than their
counterparts about their desire to affect change through journalism. Specifically, we explain our
case study, which examined how journalists at the Pulitzer Prize-winning nonprofit collaborative
International Consortium of Investigative Journalists (ICIJ) talked and thought about the impact
of their work.1 This helps us, as journalists and scholars, broaden our understanding of journal-
ism’s role in supporting democracy.
198
What can nonprofit journalists actually do?
Journalism’s (potential) impact on democracy

While journalism and democracy are inextricably linked, the precise nature of that link has not
been fully specified. That means that although journalists tend to be dedicated to strengthening
democracy, they cannot always articulate how that orientation translates into practice. Those
missing links threaten to erode journalism’s usefulness to society.
Renowned sociologist Herbert Gans (1998) argued that these problems are the result of short-
comings in the journalistic theory of democracy, which he described as consisting of three steps:
1 Journalism’s role is to inform citizens;

2 The more informed citizens are, the more likely they are to participate; and
3 The more they participate, the more democratic the country will be.
Gans argued that the theory fails on all three levels. It is not clear whether or how journal-
ists inform the public. How that information encourages participation is also foggy. Finally, it
is not obvious that more participation leads to more democracy. (Gans argued that in fact this
conceptualization ignores the imbalance of power in society – specifically the fact that powerful
individuals and interest groups can override the influence of citizens). Ultimately, Gans wrote,
the theory enables journalists to ignore consequences by suggesting that newly informed citizens
spontaneously choose to act, which somehow leads to a strengthened democracy – a chain of
vague events that doesn’t require journalists to take responsibility for the outcome of their work.
Indeed, in their extensive study of investigative journalists, Ettema and Glasser found that most
resisted discussing the impact of their work because they “underst[oo]d their power, as well as
their responsibility, to be limited to telling a story” (1998: 82). While public policy might change
or public opinion might shift as a result of their work, the investigative reporters Ettema and
Glasser spoke with in the 1990s professed not to care.
Merritt and McCombs (2004) argued, however, that it was not sufficient for journalists to be
agnostic concerning impact. Instead of clinging to objectivity or detachment, journalists should
redefine their role as fair-minded participants in public life, Merritt (1998) argued elsewhere.
While journalists claim to simply be reporting the news, they are in fact making decisions based
on values about how the world should work. If journalists were open about that orientation,
journalistic content and journalism’s relationship to the public would change, he argued: “We
could, for one, regain some lost standing with the public and, as a result, be more effective in our
role in the democratic process” (1998: 12).
This gap between journalism’s operations and its democratic impact has long been evident.
The public journalism movement of the 1990s was one attempt to remedy this disconnect,
with goals that included “deliberatively positioning ordinary people as capable of some action,”
(Schaffer, 2015: 1) establishing journalism as “democracy’s cultivator, as well as its chronicler,”
and restyling the press so that it “supported a healthier public climate” (Rosen, 1999: 4). Schaf-
fer (2015) argued that public journalism worked because it built in simple ways for people to
participate in democracy. Yet it ultimately failed, in part because the movement didn’t develop its
own theory of how journalism should function in a democracy, focusing instead on “creat[ing]
and sustain[ing] a conversation to help newsrooms break out of their disconnected relationship
with their audiences” (Glasser and Lee, 2002: 205).
Still, much has changed since then. The thinking around journalistic impact has evolved in
the last two decades in part through the improved ability to track audience metrics and, for jour-
nalistic nonprofits, pressure from funders to show a return on their investment. A report by the
199
Tow Center for Digital Journalism (Keller and Abelson, 2015) found that the field of measure-
ment and evaluation “has taken a strong foothold in journalism” (2015: 9). The report describes
a scenario that would make the reporter-subjects of Ettema and Glasser’s (1998) project cringe:
an editor at the New York Times who stuffs envelopes with responses to and mentions of particular
news stories, with the thickest envelopes – that is, the ones with the most public resonance –
more likely to earn the newspaper’s support as submissions to the Pulitzer Prize.
Impact-oriented investigative journalism

If, as Merritt and McCombs (2004) argue, investigative journalism is inherently more impact-
oriented than are other forms of journalism, investigative nonprofits are more impact-oriented
still. A report by the Knight Foundation, which funds and promotes nonprofit journalistic proj-
ects, referred to a group of news nonprofits it studied this way: “They are producing high-quality
journalism that is reaching hundreds of thousands of people, changing government policies and
laws, sending wrongdoers to jail, and protecting consumer interests” (Knight Foundation, 2013: 4).
Small, nonprofit investigative newsrooms are at the forefront of impact measurement (Keller and
Abelson, 2015), driven, at least in large part, by funders who want to contribute to journalism that
spurs change. Inundated with grant applications, foundations are seeking ways to fund promising
ventures (Lewis and Niles, 2013). “The livelihoods of nonprofit newsrooms have become increas-
ingly linked to their ability to collect and report meaningful metrics of impact” (Keller and Abelson,
2015: 20). Traditional web metrics provide limited value to nonprofit newsrooms (Keller and Abel-
son, 2015; Knight Foundation, 2013; Napoli, 2014). Such analytics may be able to hint at audience
engagement but are less useful in determining whether they have made the world better. “None
of our existing analytics tools measure impact – they don’t tell us how our reporting has influenced
the public to create a better society” (Kaiser, 2016: para 6). Responding to the lack of methods to
quantify impact for nonprofit news outlets, the writers of the Tow report (Keller and Abelson,
2015) developed an open-source analytics platform to enable news organizations to measure their
impact. The Center for Investigative Reporting also created an open-source impact tracker that lets
news organizations measure qualitative and quantitative metrics (Green-Barber, 2016). Still, neither
platform goes so far as to define impact itself, leaving that up to each news organization.
Few are as transparent about impact as the investigative nonprofit ProPublica, which each year
highlights the ways its reporting has produced change, such as prompting congressional hearings,
new legislation, and reform of public agencies and private companies (ProPublica, 2016). While
news organizations increasingly track audience metrics and user engagement and may define
what impact means in their own newsroom, they have until recently rarely shared this definition
publicly or engaged in a broader conversation about how journalists should define (or redefine)
impact in the digital age. This research examined one attempt to begin that redefinition in a way
that we feel hints at a need for new conceptualizations of the role of journalism in democracy.
An impact typology for investigative journalism

The lack of a clear definition of impact in journalism is a major reason the culture of measure-
ment and evaluation has yet to become widespread in the field (Knight Foundation, 2013; Keller
and Abelson, 2015). “We don’t have the same names for common phenomena of change that
flow from our work. So we – media makers, content producers, researchers, and foundations –
face obstacles to collectively increase our understanding of how and why impact occurs” (Green-
Barber, 2014: par. 2).
Is impact, for instance, synonymous with outcomes? In a report prepared for the Bill &
Melinda Gates Foundation and the John S. and James L. Knight Foundation, Green and Patel
200
defined impact as “changes among individuals, groups, organizations, systems, and social or phys-
ical conditions that your work helps to advance” and noted that “here we use impact in the
same way that conventional evaluation uses the term outcomes, referring to the desired change
among a target community” (2013: 29). The Gates Foundation, however, made a clear distinction:
Outcomes are “intermediate observable and measurable changes that may serve as steps toward
impact for a population, community, country, or other category of beneficiary”, and impacts are
“ultimate sustainable changes, sometimes attributable to action” (Gates Foundation, 2010: 8).
Napoli (2014) made a similar distinction: Outcomes are shorter-term effects that journalism
can have, such as informing, engaging, and mobilizing audiences, while impact refers to longer-
term changes in individual behavior or changes in public policy. Another demarcation is between
the micro-level orientation of fields such as media effects that typically focus on changes in an
individual’s attitudes, beliefs, and behaviors and the macro-level orientation of media impact
measurement that, in addition to focusing on individual change, includes broad systemic changes
in organizations and institutions (Napoli, 2014).
Impact, some argue, is a nebulous term by necessity. Simons et al. (2016) wrote that useful
definitions of impact should be “adaptive” given that it can be measured longitudinally or after
exposure to a single story. Tofel argued that there cannot be one universal definition of impact
because “different sorts of journalism have different objectives, and will therefore produce – seek
to produce – quite different sorts of impact” (2013: 3). Explanatory journalism, for instance, aims
to clarify complex topics, and so perhaps its desired impact is a readership with a clearer under-
standing of the issues. For investigative journalism, however, the goal is to reveal information that
someone wanted to keep secret. Thus, one major desired impact is altering the frame around
policy change – something that is, in a sense, outside of the control of the journalist. Protess et al.
(1987) made the case that measuring the impact of investigative journalism depends on the target
of the impact – on the public, on elites, or on policy.
Ettema and Glasser (1998) categorized the accomplishments of investigative reporting as pub-
licity, accountability, and solidarity. Publicity means trying to bring public attention to cases of
systemic break down that have largely gone unnoticed or have been concealed. Then-Philadelphia
Inquirer reporter Bill Marimow explained his criteria for selecting stories to Ettema and Glasser,
and his explanation was representative of the absence of the notion of impact among the investi-
gative journalists they studied: “Is the information important? Is it interesting?” (1998: 8). Even if
there is no measurable shift in public opinion, Ettema and Glasser argued that investigative report-
ing can change the way that public affairs issues are understood. Accountability is about demand-
ing responses – whether deliberative (hearings, public debates, etc.) or substantive (passing laws,
enacting reforms, taking punitive action against those who have committed maleficence) – from
those in positions of power. Solidarity is defined as helping the audience establish an ‘empathetic
link’ toward those who have suffered. Ettema and Glasser argued that investigative journalism does
not always have to result in action by politicians or voters to have an effect. It can simply test the
conscience of a community.
In our examination of the International Consortium of Investigative Journalists, though, we
noted a slightly different conception of impact.
ICIJ and its typology

ICIJ is one of a growing body of nonprofit news organizations founded by journalists concerned
about the struggling news industry. It was born within one of the oldest and largest news non-
profits, the Center for Public Integrity, in 1997, and was spun off from CPI in 2017. ICIJ has
roughly a dozen employees in its Washington, DC, office and works collaboratively, often with
dozens of news organizations around the world and its in-house data experts, to make sense of
201
some of the largest journalistic leaks in history. Some of ICIJ’s readers come to its site for stories
written by its journalists, summaries of reporting in other countries, and blog posts describing
the ICIJ’s work, but a much larger audience sees its reports when news organizations around the
world republish them or write their own versions (ICIJ, 2012).
We were particularly interested in ICIJ because its staffers are explicit about their orientation
to impact. Staff members explained that they consider three criteria before deciding whether to
pursue a project:
1 Is a system that’s designed to protect people broken or failing?

2 Is it of global concern?
3 Are they going to get a reaction? In other words, is the work likely to have impact?
(Boland-Rudder, personal communication,
12 February 2015)
We examined in detail four projects undertaken by ICIJ:
1 “Secrecy for Sale”, published in 2013, which examined offshore tax havens through an
anonymous leak of 2.5 million secret files (ICIJ, 2013);
2 “Luxembourg Leaks”, a 2014 investigation into 340 companies that channeled hundreds of
billions of dollars into Luxembourg and saved billions in taxes thanks to rulings secured by
PricewaterhouseCoopers;
3 “Swiss Leaks”, which, in 2015, detailed how HSBC provided services to clients connected
with the global arms trade, blood diamonds, and bribery; and
4 “The Panama Papers”, published in 2016, which involved 400 journalists from six continents
in what ICIJ called “the biggest investigation in journalism history” (Alecci, 2017: par. 1), and
won the ICIJ and partner organizations a Pulitzer Prize.
We looked at how ICIJ journalists discussed the impact of each project in all of the docu-
ments we could collect, including news articles by and about ICIJ reporters, blog posts, emails to
readers, a report to funders, and tweets. We supplemented that with two interviews with a senior
ICIJ staff member.
We observed ICIJ writers discussing impact in four ways:
1 Being cited or referenced by other news organizations;

2 Spurring deliberation that could lead to public policy change;
3 Changing public opinion; and
4 Causing substantive change, including suggestions that ICIJ’s work is changing policy or
triggering official forms of deliberation that could lead to policy change.
ICIJ staffers were adamant that they were not seeking a particular type of reaction or response
to their work. Regarding LuxLeaks, online editor Hamish Boland-Rudder told us the ICIJ didn’t
aim for particular results because “we’re journalists, we’re not activists.” Overall, the goal is “to
provoke public interest of an issue and provoke debate hopefully at the highest levels; the out-
come of those debates we leave in the hands of the people making them, the public, anyone
who cares about those issues” (Boland-Rudder, personal communication, 4 March 2015). These
thoughts were echoed by ICIJ’s director, Gerard Ryle, in a radio interview about the SwissLeaks
project. When prompted by the interviewer for his opinion, Ryle noted that “I don’t think it’s
our job to comment or to get involved. Once we’ve published we need to walk away and allow
authorities and the public to decide what should happen next” (Moss-Coane, 2015).
202
These statements offer insight into the type of impact ICIJ is not looking to have. Below, we
document how ICIJ staff members refer to the impact they are seeking. We use the four catego-
ries listed, which arose from our initial analysis of the texts: references by other news organiza-
tions, deliberation, changing public opinion, and substantive change. Our analysis revealed that
mentions of impact were never apparent in the initial investigative news stories produced by ICIJ.
That is, they appeared not in the work itself, but only when staffers talked about their work – not
just to us but also to audience members and the broader public, often in blog posts describing
their projects, rather than within the work itself.
Citations or references by other news organizations

In the projects we examined, ICIJ relied on other news organizations to distribute its content, and
the degree to which others published or mentioned its articles was a major way ICIJ conceptual-
ized impact. ICIJ assiduously tracked pickup of its stories, even though the resulting data were
challenging to analyze. ICIJ tracked page views on its own site, asked media partners to embed
code on their websites to count visitors to those pages, and hired a clippings service that flagged
stories that explicitly mentioned ICIJ, but also used keywords to capture second-wave stories –
reporting on the reporting (Boland-Rudder, personal communication, 4 March 2015).
ICIJ also frequently drew attention, through all available communication channels, to men-
tions of its work. The impact section of its report to a funder counts page views on ICIJ’s own
site, partner sites, social media shares, and a database where they made the documents available, as
well as the number of media partners.
On social media, ICIJ promoted partnerships with news organizations and shared content
produced by journalists who collaborated on investigations. Nearly all ICIJ tweets about the
projects linked to articles or videos produced by media partners or other news organizations
reporting on ICIJ’s initial investigation. ICIJ also tweeted quotations from editorials supportive
of holding accountable the individuals or institutions associated with offshore tax havens. One
tweet, for instance, quoted an editorial published in the EU Observer, making the case that “If
there is one thing we should learn from LuxLeaks, it is the power of transparency”.
Interestingly, when it came to Panama Papers, just one story tagged “impact” on the ICIJ website –
about journalists facing blowback for reporting on Panama Papers – noted references to the series by
other news organizations. Still, ICIJ frequently referenced the number of media partners, produced
a video focused largely on press coverage of the initial Panama Papers reporting (Alecci, 2017: par
2), and highlighted the fact that it shared the Pulitzer Prize with two partner news organizations.
ICIJ frequently referenced moments in which other news organizations highlighted wide-
spread attention being paid to ICIJ’s work. Regarding the LuxLeaks investigation, ICIJ noted in
blog posts that the New York Times reported that revelations had sparked a “rising furor,” that
Reuters called the response a “tax storm,” and that a Bloomberg editorial called for the European
Commission President’s resignation (Boland-Rudder et al., 2014).
Despite this assiduous tracking, it remained a challenge for ICIJ to understand how to analyze
the data, with each staffer using his or her own measure of success: “We all have our own personal
things. We all have the writers we admire, the publications we admire, so for each of us there’s
that moment of ‘oh my god, we’re on 60 Minutes or the front page of Financial Times’” (Boland-
Rudder, personal communication, 12 February 2015).
Deliberation
ICIJ carefully tracked and reported to readers and funders the conversation on social media
and other online platforms generated by its initial investigative reporting. That’s because,
203
Boland-Rudder explained, while policy change is a desirable outcome, public conversation is

just as valuable.
We’re looking at things like social media engagement, what sorts of conversations are
happening online around our work and even down to what comments are being left
on our website. [. . .] It’s those ways of having it injected into the public conversation
that’s also really valuable, but it’s a difficult thing to measure and it’s a difficult thing to
quantify exactly how that impact will play out.
4 March 2015)
ICIJ often highlighted comments made by public officials in response to revelations from the
leaks, such as “Responding to the new ‘Lux Leaks’ revelations, the Luxembourg finance ministry
conceded . . .” (Boland-Rudder, 2014a). ICIJ suggested its work had a long-term deliberative
impact by writing that tax avoidance schemes “have been in the spotlight in Europe since ICIJ
and its partners published more than 500 secret Luxembourg tax agreements” (Boland-Rudder,
2014b). It also used social media to highlight deliberative responses by policymakers, often start-
ing tweets with the word “IMPACT” in bold letters.
Ultimately, deliberative change – along with mention by news organizations – can help
amplify impact, Boland-Rudder explained in an interview.
Having a story on the front page of Financial Times or The New York Times, it definitely
plays into building that public conversation [emphasis added] and quite often that public
conversation finishes with questions pitched squarely at the decision makers, often poli-
ticians. So the more reach a story has, the more people are talking about it, it definitely
plays into the amount of impact. It becomes an issue people can’t ignore.
12 February 2015)
In the case of the Panama Papers project, a large number of ICIJ stories on its impact referenced
policy discussions, such as government proposals to increase financial transparency, proposals
from tax agencies and other experts about tax reform – specifically targeting offshore enablers –
and government hearings featuring testimony from journalists, world leaders, and bankers.
Public opinion
References to change in public opinion were rare and tended to be subtle – not quantified evi-
dence of changing opinion but rather statements bordering on oblique references to public senti-
ment. For instance, an ICIJ blog post referring to the Secrecy for Sale project quoted a report by
Offshore Incorporations Limited, a leading offshore services firm, saying that “The scandal has
created a crisis of confidence in the industry” and led to issues concerning the banking industry’s
public image (Chavkin, 2013).
On several occasions, ICIJ tweeted quotations that made the case that the LuxLeaks investi-
gation had changed public perception on tax havens. “Before #LuxLeaks, ‘nobody really cared
about this’ and now ‘the whole world has started to discuss rulings’”, ICIJ tweeted, citing a Bloom-
berg article that quoted Luxembourg’s new finance minister. Indeed, Boland-Rudder tweeted:
“cool that #LuxLeaks and #SwissLeaks may now be part of Europe’s lexicon”.
In an interview, Boland-Rudder explained why ICIJ had a vested interest in how the public
perceived its investigations:
204
We knew, particularly from receiving feedback from reporters, that there would be
strong interest from local audiences in this story. Wherever there’s a strong interest
there’s likely to be strong discussions. Strong discussions often lead to positive outcomes
or at least a better informed public.
(personal communication, 4 March 2015)
Substantive change
ICIJ journalists also frequently referenced substantive change occurring as a result of their work,
especially when reporting led to “any inquiries, any prosecutions, laws changed, systems fixed”
(Boland-Rudder, personal communication, 12 February 2015).
One dramatic result of the SwissLeaks investigation was the fact that Swiss authorities opened
a criminal investigation against HSBC and raided the bank’s offices in Geneva. “The raid comes
amid growing scrutiny of the bank following ICIJ’s Swiss Leaks investigation”, ICIJ told its read-
ers (Boland-Rudder, 2015). Many other references to this form of impact related to investigations
and inquiries.
Among the most prominent substantive impact of the LuxLeaks project was a no-confidence
vote over the European Commission’s president, who had been the prime minister of Luxem-
bourg during the time of the tax deals (Blenkinsop, 2014). ICIJ explained in a blog post that the
president “has been under pressure since ICIJ and its media partners published stories detailing
secret Luxembourg tax deals” (Boland-Rudder, 2014c).
The LuxLeaks investigation was followed by a call for an inquiry, which ultimately became
a “special committee.” ICIJ referred to the inquiry as “a LuxLeaks inquiry committee” (Schilis-
Gallego, 2015) and explained that the European Parliament instead opted for a special committee
to “look into tax avoidance in Luxembourg, following the LuxLeaks revelations published by
ICIJ and its media partners in November [2014]”.
Several articles on the ICIJ website about the Panama Papers project referenced public pol-
icy changes, including new financial reforms, legislation targeting tax avoidance, mandates that
banks disclose information, and a range of actions – lawsuits, fines, and arrests (of the law firm’s
founders) – taken against the firm at the center of Panama Papers. In its review of impact one year
after the publication of Panama Papers reports began, ICIJ noted that the investigation prompted
police raids, arrests and resignations of high-profile figures.[. . .] Reporting by ICIJ and
its partners has sparked more than 150 inquiries, audits, and investigations in 79 coun-
tries and has driven new legislation and financial rules in the U.S. and abroad.
(Hudson, 2017: par. 15)
Implications: broader impact

Investigative journalism has traditionally been more impact-oriented than other forms of jour-
nalism (Ettema and Glasser, 1998), and the ICIJ journalists we studied push that orientation fur-
ther toward explicitly discussing impact. ICIJ focused on potential outcomes from the outset of
its projects, then assiduously tracked the stories produced by themselves and others, noting when
they triggered deliberation, substantive impact, and changed public opinion. Tracking impact was
fueled by both internal interest and external requirements from funders who wanted to know
more than just how many stories were produced.
Still, we found that ICIJ journalists rarely suggested that change came from their work, instead
describing their journalism alongside its outcomes. For instance, ICIJ noted that tax avoidance has
been in the spotlight since the ICIJ revelations, stopping short of suggesting it was because of them.
205
And its employees did not articulate desire for a particular impact, saying instead that they do
not decide ahead of time on a desired outcome, which would look more like activism than like
journalism (Boland-Rudder, personal communication, 4 March 2015).
ICIJ journalists did repeat explicit connections between their work and its impact when those
connections were made by people outside the organization. For instance, ICIJ repeated an assertion
about its own impact made by Luxembourg Finance Minister Pierre Gramegna, noting in a blog post
that “Gramegna said the Lux Leaks revelations brought the world’s attention to the issue”(Boland-
Rudder, 2014d). Similarly, ICIJ noted that “The New York Times said the revelations have sparked a
‘rising furor’ in Europe. Reuters called the reaction a ‘tax storm’” (Boland-Rudder et al., 2014).
As the nonprofit news sector becomes increasingly metrics-driven (Epstein and Yuthas, 2014),
one might expect the foundations and individuals that fund ICIJ to look for impact measures
from the organization as well. Boland-Rudder acknowledged that people who had given them
money wanted to see results, noting the organization’s funders “don’t have any potential financial
gain from our stories but they still want to see value for the money they’re spending so we have
to present that more and more in terms of impact” (Boland-Rudder, personal communication,
4 March 2015). Still, the hesitation and resistance described earlier suggest a desire to maintain
some degree of traditional journalistic ethos – a distinction that has ICIJ walking a fine line.
Indeed, ICIJ journalists are not the only investigative reporters who feel a level of discomfort
in discussing the impact of their work. A manager of ProPublica, perhaps the most prominent
nonprofit investigative news center in the United States, had this to say:
Before moving on to the mechanics of charting impact, there is another key issue with
which we need to wrestle: Is it even appropriate for journalists to seek impact from their
work? Does such an objective cross the line from journalism into advocacy?
(Tofel, 2013)
Its connection to foundation and philanthropic funding is not the only thing that differenti-
ates ICIJ from mainstream journalism. ICIJ describes itself as intentionally distinct, an antidote to
traditional newsrooms that have cut costs and moved away from investigative reporting (Interna-
tional Consortium of Investigative Journalists, 2012). For ICIJ, solving these problems requires
leveraging the capacity of its small, U.S.-based staff through international collaborations and by
giving its content away. ICIJ’s success depends on other organizations publishing its work and
adding local context. In other words, ICIJ attempts to do something different than mainstream
journalism, but its symbiotic relationship with traditional news outlets means it cannot stray
significantly from norms and practices of mainstream journalism. This reliance limits how ICIJ
can conceive of its work, requiring the organization to balance its attempt to produce content
that will be picked up by journalists and mass audiences with an effort to improve journalism’s
place in a democracy and outwardly discuss impact – if not directly in its news stories, then at
least when addressing its readers.
Through this lens, we can make better sense of Ryle’s comments on public radio that: “I don’t
think it’s our job to comment or to get involved. Once we’ve published we need to walk away
and allow authorities and the public to decide what should happen next” (Moss-Coane, 2015)
and Boland-Rudder’s statement that “We don’t like to dictate what result we’d like to get because
we’re journalists, we’re not activists” (personal communication, 4 March 2015). These statements
seem to echo what a Philadelphia Inquirer reporter told Ettema and Glasser: “It’s not part of the
process to go out there and lobby for reform” (1998: 9) – even though in many other ways, ICIJ
staffers were much more open about discussing impact.
For both the Inquirer reporter and ICIJ staff, lobbying for a particular change is out of bounds.
Both organizations follow journalistic norms that prohibit such overt support for specific
206
outcomes. Indeed, an ICIJ employee told us that when there is conflict between orientation
toward impact and a push for being disinterested,
we will always fall back on our journalistic values and our journalist ethics whereby the
story comes first and we are there to explore the issue, to shine a light in a dark place
and open it up to public discussion, not to necessarily focus on any particular agenda.
(Boland-Rudder, personal communication, 4 March 2015)
This, then, is the line. ICIJ discusses its impact openly, while other investigative journalists may
hesitate to do so, but open advocacy or ignoring journalistic norms would mean going too far.
Toward a new theory of democracy

This evolution in the notion of journalistic impact presents an opportunity for scholars to reopen
a discussion about the role of journalism in democracy – an opportunity we should take seri-
ously given the troubles facing journalism and modern democracy. In the journalistic theory of
democracy described (and critiqued) by Gans, journalists’ responsibility ends with informing the
public – an approach that enables journalists to ignore the consequences of their work. In the new
theory being forged by ICIJ and others, journalists take part in every step along the way.
Our study identified the four aforementioned ways ICIJ understands the impact of its projects,
offering this path for how journalism can help build a stronger democracy:
1 Informing the public;

2 Spurring conversation in media outlets and among the general public;
3 Changing opinions; and, ultimately
4 Changing policy.
This new theory makes concrete the notion that the responsibility of journalists to uphold
and strengthen democracy does not end at publishing or broadcasting stories. Journalists should
instead work to ensure their reporting reaches a wide audience. Spurring conversation also
requires journalists to engage the public directly through social media and other platforms. While
journalistic norms understandably prevent many journalists from calling for specific changes in
public opinion or prescribing policy changes, tracking this type of change when it comes about
alongside their reporting should not raise ethical concerns. We suggest that carefully tracking
and informing the public about the impact of a story is less self-promotion than it is part of
journalists’ duties to report on the lifecycle of an investigation. This theory may promise to build
stronger and more citizen-oriented communities and to bring new direction to the journalistic
profession, which has become unmoored from its democratic mission.
At ICIJ it seems these steps are a means to measure success rather than to make recommenda-
tions about journalistic behavior on a broad scale – a place where scholars, rather than journalists,
can contribute. Still, while there is work to be done in fleshing out these steps (and we suggest
some possible research directions later), ICIJ’s approach to impact suggests an emergent new
theory of journalistic democracy – one that is slowly being embraced by other organizations (see,
e.g., Keller and Abelson, 2015).
Finally, we agree with Merritt’s argument that journalists may spur positive change by openly
acknowledging their values, perhaps regaining lost standing with the public and becoming “more
effective in our role in the democratic process” (1998: 12). Those suggestions are particularly per-
tinent given ICIJ’s unusual newsroom processes. The broad interest in ICIJ’s stories means that
traditional news organizations, on one hand eager to distance themselves from the impact of their
207
work and on the other eager to snap up the free, quality work offered by ICIJ, are publishing
stories that originated from a newsroom that does not necessarily share their orientation toward
impact. Given this tension, it may well be that journalistic transparency and outcomes would be
better served by traditional news organizations owning up to the impact-oriented nature of what
they do – or at least of the ICIJ stories they carry.
While we found ICIJ’s explanations of its work insightful, questions remain. Does this orien-
tation toward impact exist at other nonprofit news organizations and, indeed, at the mainstream,
commercial news organizations that reach the broadest audiences? How do foundations and
other funders of investigative journalism define impact and want it to be measured? Are these
entities pushing a revised understanding of the role of journalism in democracy?
As these empirical studies mount, we will also need serious normative arguments that attempt
to flesh out the journalistic theory of democracy. These will come not from observation of jour-
nalists at work but rather from an examination and understanding of what democracy needs and
what journalists can do to support it.
Further reading
Work about citizens’ roles in democracy can help make sense of why journalists need to think
more about the impact of their work. Harry Boyte writes about reinventing citizenship as “public
work,” done “by publics, for public purposes, in public.” This suggests citizenship is larger than
membership and instead sees citizens as creative agents and co-creators of the worlds they inhabit,
an orientation that helps us understand why it is important for journalists to focus on impact. That
piece, published in 2013, was titled “Reinventing Citizenship as Public Work: Citizen-Centered
Democracy and the Empowerment Gap” and was published by the Kettering Foundation.
Note
1 This study is described in greater detail in Konieczna and Powers (2017).
References
Alecci, S. (2017, April 3) “Video: Twelve months of investigations, impact and outrage.” International Con-
sortium of Investigative Journalists. Retrieved from https://panamapapers.icij.org/20170403-anniversary-
fundraising-campaign.html
Anderson, C., Bell, E. and Shirky, C. (2014) “Post-industrial journalism: Adapting to the present.” Tow Center for
Digital Journalism. Retrieved from http://towcenter.org/research/post-industrial-journalism-adapting-
to-the-present-2/
Bill & Melinda Gates Foundation. (2010) A Guide to Actionable Measurement. Retrieved from https://docs.
gatesfoundation.org/documents/guide-to-actionable-measurement.pdf
Blenkinsop, P. (2014, November 27) “EU’s juncker survives no-confidence vote over tax deals.” Reuters.
Retrieved from www.reuters.com/article/2014/11/27/us-eu-juncker-idUSKCN0JB13Q20141127
Boland-Rudder, H. (2014a, December 11) “Spotlight on ‘weakened’ juncker as officials pledge action.” ICIJ.
Retrieved from www.icij.org/blog/2014/12/spotlight-weakened-juncker-officials-pledge-action
Boland-Rudder, H. (2014b, December 3) “Europe to fast-track action against ‘sweetheart’ Deals.” ICIJ.
Retrieved from www.icij.org/blog/2014/12/europe-fast-track-action-against-sweetheart-deals
Boland-Rudder, H. (2014c, November 27) “Calls for reform and a resignation over LuxLeaks.” ICIJ.
Retrieved from www.icij.org/blog/2014/11/calls-reform-and-resignation-over-luxleaks
Boland-Rudder, H. (2014d, December 22) “LuxLeaks a ‘game-changer’ for Europe: Finance minister.” ICIJ.
Retrieved from www.icij.org/blog/2014/12/luxleaks-game-changer-europe-finance-minister
Boland-Rudder, H. (2015, February 18) “HSBC’s Geneva office raided as Swiss open investigation.” ICIJ.
Retrieved from www.icij.org/blog/2015/02/hsbcs-geneva-office-raided-swiss-open-investigation
208
Boland-Rudder, H., Wayne, L. and Carr, Kelly. (2014, November 19) “LuxLeaks causes ‘tax storm’ of govern-
ment, media response.” ICIJ. Retrieved from www.icij.org/blog/2014/11/lux-leaks-causes-tax-storm-
government-media-response
Chavkin, S. (2013, November 26) “Tax havens face crisis in wake of offshore leaks, report says.” ICIJ.
Retrieved from www.icij.org/blog/2013/11/tax-havens-face-crisis-wake-offshore-leaks-report-says
Epstein, M. J. and Yuthas, K. (2014) Measuring and Improving Social Impacts: A Guide for Nonprofits, Companies,
and Impact Investors. San Francisco, CA: Berrett-Koehler Publishers, Inc.
Ettema, J. S. and Glasser, T. L. (1998) Custodians of Conscience: Investigative Journalism and Public Virtue. New
York, NY: Columbia University Press.
Gans, H. J. (1998) “What can journalists actually do for American democracy?” Press/Politics, 3(4), 6–12.
Glasser, T. L. and Lee, F. F. (2002) “Repositioning the newsroom: The American experience with ‘public
journalism’.” In E. Neveu and E. Kuhn (eds.), Political Journalism: New Challenges, New Practices. London:
Green, D. and Patel, M. (2013) “Deepening engagement for lasting impact: A framework for measuring
media performance and results.” Bill & Melinda Gates Foundation and the John S. and James L. Knight
Foundation. Retrieved from www.namac.org/wp-content/uploads/2015/08/deepening-engagement-
for-lasting impact.pdf
Green-Barber,L. (2014,September 18) “The language of impact:Introducing a draft glossary.”The Center for Investiga-
tive Reporting. Retrieved from www.revealnews.org/article-legacy/the-language-of-impact-introducing-a-
draft-glossary/
Green-Barber, L. (2016, September 22) “CIR’s impact tracker: How to use it and why you need it.” The Center for
Investigative Reporting. Retrieved from www.revealnews.org/article/cirs-impact-tracker-how-to-use-it-
and-why-you-need-it/
Hudson, M. (2017, April 10) “Panama Papers win Pulitzer Prize.” International Consortium of Investigative
Journalists. Retrieved from https://panamapapers.icij.org/20170410-pulitzer-prize.htm
International Consortium of Investigative Journalists. (2012) About the ICIJ. Retrieved from www.icij.org/
about
International Consortium of investigative Journalists. (2013) Offshore. Retrieved from www.icij.org/offshore
Kaiser, J. E. G. (2016, September 26) “A quantitative basis for measuring media impact at the media consortium.”
MediaShift. Retrieved from http://mediashift.org/2016/09/quantitative-basis-measuring-media-impact/
Keller, M. and Abelson, B. (2015) “NewsLynx: A tool for newsroom impact measurement.” Tow Center
for Digital Journalism. Retrieved from http://towcenter.org/wp-content/uploads/2015/06/Tow_Cen-
ter_NewsLynx_Full_Report.pdf
Knight Foundation. (2013) Finding a Foothold: How Nonprofit News Ventures Seek Sustainability. Retrieved
from www.knightfoundation.org/features/nonprofitnews/
Konieczna, M. and Powers, E. (2017) “What can nonprofit journalists actually do for democracy?” Journal-
ism Studies, 18(12), 1542–1558.
Lewis, C. and Niles, H. (2013) “Measuring impact: The art, science and mystery of nonprofit news.”
Investigative Reporting Workshop. Retrieved from http://investigativereportingworkshop.org/ilab/story/
measuring-impact/
McIntyre, K., Dahmen, N. S. and Abdenour, J. (2016) “The contextualist function: US newspaper journal-
ists value social responsibility.” Journalism. doi: 10.1177/1464884916683553
Merritt, D. (1998) Public Journalism and Public Life: Why Telling the News Is Not Enough. New York, NY:
Routledge.
Merritt, D. and McCombs, M. E. (2004) The Two W’s of Journalism: The Why and What of Public Affairs Report-
ing. Mahwah, NJ: Lawrence Erlbaum Associates.
Moss-Coane, M. (2015, March 1) “HSBC’s accused of tax evasion and money laundering.” Radio Times, WHYY.
Retrieved from http://whyy.org/cms/radiotimes/2015/03/01/hsbcs-accused-of-tax-evation-and-money-
laundering/
Napoli, P. M. (2014) “Measuring media impact: An overview of the field.” Norman Lear Center, University of
Southern California. Retrieved from www.learcenter.org/pdf/measuringmedia.pdf
ProPublica. (2016) Impact From ProPublica. Retrieved from www.propublica.org/about/impact/
Protess, D. L., Cook, F. L., Curtin, T. R., Gordon, M. T., Leff, D. R., McCombs, M. E. and Miller, P. (1987)
“The impact of investigative reporting on public opinion and policymaking: Targeting toxic waste.”
Public Opinion Quarterly, 51(2), 166–185.
Rosen, J. (1999) What Are Journalists for? New Haven, CT: Yale University Press.
Schaffer, J. (2015, June 22) “If audience engagement is the goal, it’s time to look back at the suc-
cesses of civic journalism for answers.” Nieman Lab. Retrieved from www.niemanlab.org/2015/06/
209
if-audience-engagement-is-the-goal-its-time-to-look-back-at-the-successes-of-civic-journalism-for-
answers/
Schilis-Gallego, C. (2015, February 11) “Calls for EU LuxLeaks inquiry rejected despite support.” ICIJ.
Retrieved from www.icij.org/blog/2015/02/calls-eu-luxleaks-inquiry-rejected-despite-support
Simons, M., Tiffen, R., Hendrie, D., Carson, A., Sullivan, H., Muller, D. and McNair, B. (2016) “Understand-
ing the civic impact of journalism.” Journalism Studies, 18(11), 1400–1414.
Stray, J. (2012, August 30) “Metric, metrics everywhere: How do we measure the impact of journalism?”
Nieman Lab. Retrieved from www.niemanlab.org/2012/08/metrics-metrics-everywhere-how-do-
we-measure-the-impact-of-journalism/
Tofel, R. J. (2013) “Non-profit Journalism: Issues Around Impact.” ProPublica. Retrieved from http://s3.
amazonaws.com/propublica/assets/about/LFA_ProPublica-white-paper_2.1.pdf
210
16
DIGITAL JOURNALISM AND
REGULATION
Ownership and control
Victor Pickard
Critical questions about the ownership, control, and regulation of media institutions are often
given insufficient attention in digital journalism scholarship. Many factors contribute to this
neglect, including the marginalized position of political economic approaches to media within
the broader field of communications, as well as the often-invisible nature of media policies and
other structural factors that shape journalistic practices and institutions. This oversight is also
symptomatic of a broader technocentric discourse that imagines the digital media landscape as
a wild terrain of disruptive innovation with little governmental oversight. According to such
narratives, this landscape tends to create a more equitable and open space for news producers
and consumers, one that is left unfettered by the state. Despite such libertarian assumptions, this
chapter brings into focus structural questions – particularly issues connected to media policy –
that arise in the face of journalism’s digital future.
One might begin by asking why journalists and journalism scholars should care about media
policy? This question seems provocative because it is often assumed that journalism and policy are
two completely separate fields. There is no natural reason why “journalism policy” is an uncom-
mon phrase and concept (Pickard, 2017b). This strangeness speaks to the odd bifurcation between
the study of journalism’s content and practices and the study of the larger structures that shape its
institutions. This partially reflects the popular mythology that government has never been involved
in journalism. But history tells us otherwise. From early debates over the postal system to the
founding of each new medium from the telegraph to the internet, government policy has always
been central to journalism and media institutions. For issues as diverse as copyright, broadcast
spectrum management, public interest regulations, and the enforcement (or lack thereof) of media
ownership restrictions and antitrust laws, policy plays a key role in how media systems are designed,
owned, and operated. To pretend otherwise is deeply misleading. It discourages public involvement
in media policy debates and helps to reify media systems that disproportionately favor corporate
interests (Pickard, 2015a). Indeed, media policy’s relationship to journalism is arguably of greater
concern now than ever before. Establishing new laws and policies dealing with intellectual property,
privacy, and internet governance will lay the foundation for journalism’s digital future.
Given this history and ongoing policy debates we can conclude: the government is always
involved – often to the benefit of large media corporations. So the real question becomes how the
government should be involved. The analysis in this chapter assumes that numerous government
interventions directly or indirectly affect journalism – from shield laws to tax laws – but I draw
211
Victor Pickard
specific attention to several key areas. Ranging from questions of ownership to regulation of media
content and infrastructure, these policy debates show how the future of journalism is inextricably
bound up with questions of broadband provision, digital ‘red-lining’, copyright concerns, and a
proliferation of fake news and false advertising, to name a few regulatory concerns. Although
journalism’s digital transition has in many cases lowered barriers of entry and seemingly increased
user choice (for those with internet access), it has also enabled price discrimination, clickbait, fake
news, mass corporate and state surveillance, and many other potential hazards.
This chapter examines a number of often-overlooked structural questions that have profound
implications for digital journalism’s future. Given the focus of my research, much of the fol-
lowing draws from an ongoing case study of American journalism, but this analysis also holds
implications for the state of journalism and journalism studies internationally. With this in mind,
analysis falls into six sections, focused respectively on the new American media landscape, the
American journalism crisis, alternatives to the advertising revenue model, economic and regula-
tory discourse about journalism, new areas for regulatory concern, and why media ownership
matters for journalism. Before I turn to a discussion of media regulation and policy, I provide an
overview of issues related to ownership and control of the new American media landscape. The
chapter concludes with a brief discussion of why media ownership matters and what policies
could be deployed to help democratize the control of journalistic institutions.
The new American media landscape

Commercial media around the world share many attributes, but taken as a whole the U.S. media
system is in a category of its own. This is true for three general reasons. First, in many sectors
the U.S. media system is dominated by a handful of corporations. Second, the American media
system is only lightly regulated by public interest protections. Earlier protections such as the Fair-
ness Doctrine – which mandated that broadcasters cover important and controversial issues and
to do so in a balanced manner – have long been repealed (Pickard, 2015a). Third, the American
media system is largely commercial, with only weak public alternatives such as public radio and
television. Many countries face one or two of these challenges, but rarely all three at once. This
American media exceptionalism makes for an interesting case study where scholars can observe
the effects of largely unmitigated commercialism on journalistic practices.
To demonstrate the third point about the impoverished state of the American public media
system (the following numbers derive from Benson and Powers, 2011): The U.S. government
allocates about $1.35 per person per year to public media. If you add further local and state
subsidies, it comes out to less than $4 per citizen per year toward funding its public media.
Japan’s NHK receives approximately $54 per citizen every year, while the BBC receives over $90.
Northern European countries like Denmark and Finland spend well over $100/year per person.
In comparison, the US is a global outlier among leading democracies for how little it spends
on its public media. This may worsen with the Trump administration and many congressional
Republicans threatening to cut public media’s funding even further (Pickard, 2017a). Without
public media there is no safety net if the market fails to support journalism. To give one example
of where there is a safety net, the BBC has recently assigned 150 “democracy reporters” to British
news organizations across the country to focus on local journalism (Plunkett, 2016).
The structural crisis stemming from the lack of funding for public service journalism is often
masked by what seems like an explosion of digital media outlets. In recent years, basic ownership
patterns in the commercial media sector have shifted in visible ways, with a new crop of digital
start-ups taking root within the media landscape (see Witschge and Harbers, this volume, Chap-
ter 5). At the surface, this emergence seems to suggest a greater abundance of media choices than
ever before. However, this gloss of diversity masks an underlying uniformity, especially in terms
212
Digital journalism and regulation
of media ownership and control. A number of scholars and journalists have pointed out that
media are concentrating and replicating older patterns and power relationships. One report for
Bloomberg noted that many newcomers like Vox Media, Vice Media, and BuzzFeed are owned or
supported by the very legacy media companies these start-ups are purportedly displacing (Molla
and Ovide, 2016).
Increasingly, old media and telecommunication giants like AT&T, Verizon, Disney, Comcast,
Time Warner and others are investing hundreds of thousands of dollars into these new media
outlets, or buying them outright. Dean Starkman (2016) convincingly shows that, far from
decentralizing media power, in actuality big media companies are simply getting much bigger
in our digital age via continuous mergers and acquisitions. At this very moment, several mega-
mergers, like those between AT&T & Time Warner, are in the final stages of regulatory scrutiny.
Furthermore, scholarship has long shown that online traffic and audience attention is dominated
by incumbent media players, which challenges the notion that the digital media landscape has
opened up spaces for new voices and viewpoints (Hindman, 2008).
In a more general sense, the American newspaper industry has seen dramatic tumult in recent
decades, and especially in the last few years. While the overall number of daily newspapers in the
US remained relatively stable for decades, from 1,760 papers in 1955 to 1,676 in 1985 (Williams
and Pickard, 2016), the number of independent papers fell by nearly 50% during these years as
large chains went on a buying spree (Neiva, 1996). Since at least the late nineteenth century, most
major U.S. magazines and newspapers have been owned or controlled by wealthy individuals or
families. Between the 1970s and early 2000s, however, media companies increasingly became
publicly traded stock corporations that often expanded into large chains. As family-owned news-
papers became public companies with shareholders, their value increased, which in turn incen-
tivized owners to sell off their controlling shares to newspaper chains for high profits. One of
the largest chains, Gannett, owns USA Today and over 100 other daily newspapers (Benson and
Pickard, 2017).
In general, publicly traded newspapers face more severe commercial pressures. While a private
company can de-emphasize profits imperatives, a publicly traded company is legally obliged to
maximize shareholder value. Privileging profitability may encourage compromises in profes-
sional standards, democratic concerns, and a commitment to local communities. Some newspa-
pers have been able to lessen these commercial pressures through different ownership structures.
For example, at the New York Times there has been a two-tiered stock ownership structure where
the Ochs-Sulzberger family was able to maintain some degree of control, and at the Washington
Post the Grahams continued to control voting stock after going public in 1971. By building in
safeguards that buffer news organizations from direct commercial pressures, private ownership
can potentially liberate news organizations from Wall Street pressures. Ideally, these mitigating
structures might allow publishers to absorb short-term losses and stave off cost-cutting measures,
possibly for long-term gains or out of a commitment to public service journalism (Benson and
Pickard, 2017).
However, private ownership often succumbs to the same pressures as publicly traded com-
panies. Exemplifying this trend is one of the fastest-growing forms of private media ownership
today: the investment company. This ownership structure is often linked to aggressively profit-
driven private equity firms such as hedge funds. The largest investment groups include New
Media/Gatehouse, which owns 125 daily newspapers and is now larger than Gannett; Digital
First Media, which owns 62 daily papers; and Tronc/Tribune, which owns the Chicago Tribune,
Los Angeles Times and 17 other dailies (these numbers are drawn from Benson and Pickard, 2017).
As the newspaper industry became a profitable industry by the 1980s, it also became increas-
ingly concentrated as newspaper chains purchased one another in pursuit of higher profits
(Soloski, 2013; Williams and Pickard, 2016). By the 1980s and 1990s, most large newspaper
213
Victor Pickard
companies had profit margins exceeding 20% (Picard, 2008), and advertising revenue continued
to climb steadily into the 2000s (Picard, 2002). Until about 2005, newspaper companies were
among the most profitable businesses, maintaining 20–30% profit margins (Pickard, 2017c). Also
noteworthy was that, unlike any other newspaper industry in the world, American newspapers
relied on advertising revenue for 80% of their aggregate revenues (Benson, 2013). This left them
particularly vulnerable to specific kinds of market failure (Pickard, 2014b).
The American journalism crisis

Over the past decade this period of high profits came to a dramatic end as newspapers’ advertis-
ing revenue rapidly declined and their fundamental business model collapsed. More specifically,
as readers and advertisers continue to migrate to the web where digital ads pay a tiny fraction of
what traditional paper-based print ads yield, such high profits are unlikely to return. What digi-
tal advertising revenue is being generated is going primarily to monopolistic internet firms like
Google and Facebook, which are taking around 85% of every dollar spent on advertising. The
newspaper industry has lost tens of billions of dollars in annual advertising revenue since 2000.
The digital advertising revenue that is being generated is largely siphoned off to platforms and
search engines that host links to the news content and are increasingly becoming consumers’ first
point of call in accessing news media content.
Despite the growth in digital ad revenue in recent years, it has not been nearly enough to
offset enormous losses from traditional advertising. Recent annual reports by the Pew Research
Center cast these trends into stark relief. A 2012 Pew study found that since 2003, declines of up
to 50% in print advertising revenue far exceeded any gains in online digital revenue by more than
10 to 1 (Rosentiel and Mitchell, 2012). The newspaper industry’s decline is also exemplified by
an increase in newspaper closures, especially in the few remaining two-paper cities. For example,
the 150-year-old Rocky Mountain News shut down, and the Seattle Post-Intelligencer went online
only in 2009, cutting all but a handful of employees.
While high-profile closures have received the most attention, the journalism crisis has hurt all
papers, especially large metro papers that have difficulty providing unique news coverage and are
losing paid circulation faster than leading national papers, smaller circulation mid-sized cities, and
community papers (Edmonds et al., 2013). These declines are leading to a rise in newspaper bank-
ruptcies. Among the nation’s top 100 newspapers, 22 newspapers filed for bankruptcy between
2005 and 2015 (Williams and Pickard, 2016). In addition to restructuring through bankruptcy,
many newspapers have aggressively cut costs via mass layoffs. The American Society of News
Editors estimated that from 2005 to 2015 there had been roughly a 39% decline in news industry
workers. Other signs of the ongoing and worsening lack of investment in news production include
the reduction of home deliveries by leading metro dailies like the Cleveland Plain Dealer and New
Orleans Times-Picayune (Pickard, 2015a). In the coming years, revenue, jobs, and circulation will
likely continue to decline. Despite a post-election ‘Trump Bump’ in subscriptions for many pub-
lishers, the overall financial health of the American newspaper industry will likely worsen.
These structural factors – an over-reliance on advertising, few policies that can correct against
commercial excesses, and virtually no public safety net – all combine to create a news media
system in the US that is subject to unmitigated commercial pressures. This leads to specific
vulnerabilities and biases that created the perfect storm for a structural journalism crisis. The
150-year-old revenue model for commercial newspapers appears to be beyond repair. But because
this economic relationship has been around for so long, it is often assumed to be the natural order
of things, and alternative models are beyond imagination.
As news organizations cut costs and chase ever-diminishing revenues, news organizations are
also pressured to do more with less. The stresses associated with such austerity measures translate
214
to fewer and lower-paid jobs with fewer benefits. In addition to this growing casualization and
overall growing precarity in news labor, a concern that rises to the fore is the loss of particular
kinds of public service journalism. These are sometimes referred to as “news deserts” (Stites,
2011), especially at the state and local levels, where entire geographic areas and particular beats
go uncovered. In onging research, Alex Williams is finding that these deserts disproportionately
affect communities of color and lower socioeconomic neighborhoods, what he refers to as “news
redlining”, in which news gaps map onto economic and racial inequities.
However, print journalism, especially the newspaper industry, even in its beleaguered state,
remains the major source of original reporting in the US. The ongoing search for new profitable
business models has eluded even the most entrepreneurial start-ups. Yet even if it is no longer
profitable, democratic imperatives dictate that journalism must be supported, which necessarily
opens up a discussion for nonmarket alternatives.
Alternatives to the advertising revenue model

Most alternatives to the advertising-dependent model for journalism fall into four categories. As
advertising revenues plummet, the first and most common commercial model is digital paywalls
or online subscriptions (Pickard and Williams, 2014). A variation of this approach is the ‘mem-
bership model’, which relies on paying members to generate revenue. Thus far, this model has
seen a mixed record (Williams and Pickard, 2016). At the very least, revenue from paywalls has
not replaced the tremendous losses of advertising revenue. And one recent analysis has found that
most leading newspapers in the US continue to rely on other revenue streams outside of paywalls
(Stulberg, 2017).
A second model might be called the citizen journalism/crowd-sourcing/crowd-funding
model. The best articulation of this model so far is by Cage (2016), who argues that an institu-
tionalized form of crowdsourcing is a necessary means of support. However, there are few success-
ful models to note thus far. Some proponents of this model assume that professional institutions
are no longer necessary, but this view has faded along with some of the earlier utopian discourse
around the democratic promises of the internet. A more common model that has begun to gain
momentum in recent years is the benevolent billionaire/nonprofit/foundation-supported model.
Coming out of this approach are a number of exciting nonprofit ventures emerging as the com-
mercial model collapses. For example, ProPublica is often seen as the best exemplar of this model,
but another promising experiment in hard-hitting investigative reporting is The Intercept, launched
by Glenn Greenwald, Laura Poitras, and Jeremy Scahill. The Intercept is the first project of First
Look Media, a 501(c)3 nonprofit funded by eBay founder Pierre Omidyar, who has promised
hundreds of millions of dollars to ensure editorial independence and autonomy instead of relying
on the whims of advertisers and investors.
Another exciting venture was launched in early 2016 when the owner of the Philadelphia
Media Network (PMN), which consists of Philadelphia’s two newspapers and a news website,
donated PMN to the newly created Institute for Journalism in New Media, a nonprofit organi-
zation with a $20 million endowment. This unique structure, what is technically a ‘public benefit
corporation’, preserves editorial influence for PMN by allowing it to retain its own independent
board of directors while also allowing the institute to solicit grants to raise funds. As a hybrid
model (its ownership structure is nonprofit but its newsrooms will still be run as for-profits), it
may be shielded from some commercial pressures.
These experiments signal that journalism is something that might not be economically sus-
tainable if it relies solely on market forces. Nonetheless, democratic societies still require public
service journalism regardless of profitability. Unfortunately, although nonprofit ventures present
potential alternatives, they arguably are not systemic solutions to a structural crisis. Research
215
Victor Pickard
thus far shows increases in revenue streams such as philanthropy and capital investment still only
account for 1% of the financial support for news (Pew Research Center, 2014). Truly sustaining
public service media for the long term will require more sweeping measures that seek to rescue
journalism’s civic duty from commercial capture.
The least discussed model is the public media option, which relies on some form of public
subsidy. The argument for subsidizing public service journalism enjoyed a brief discursive oppor-
tunity from 2009 to 2011 when a number of journalism-related reports came out in support of
preserving public service journalism (Pickard et al., 2009; Downie and Schudson, 2009; Knight
Foundation, 2009; Federal Trade Commission, 2010). However, calls for state intervention soon
fell out of favor with the return of the “nervous liberals” (Pickard, 2015b), whose evasion of
addressing market failure (Pickard, 2014a) is exemplified by the Federal Communication Com-
mission’s major report “Information Needs of Communities” (Waldman, 2011). This report
accurately diagnosed the structural problems underlying the journalism crisis but dismissed out
of hand any meaningful role for government in finding structural solutions. Nonetheless, argu-
ments for subsidies are no less relevant now than they were then, and the debate around their
deployment was unfortunately closed down prematurely. There are many international and his-
torical examples of successful subsidy models.
Beginning with the U.S. postal system, the U.S. government has long subsidized news media.
In addition to postal and printing subsidies, news media were subsidized by broadcasting spec-
trum giveaways as well as a number of indirect subsidies. Even the birth of the internet was made
possible by a massive allocation of public subsidies. However, the most prominent subsidy, albeit
indirect and not with a public service goal, has always been advertising. And that subsidy is only
a fraction of what it once was. Outside of the American context, successful media subsidies are
far more common (Pickard, 2011; Kammer, 2016; Murschetz, 2014).
Although such interventions are not certain fixes for all that ails journalism, policy reforms aimed
specifically at reducing market pressures on media institutions could allow them to become more
focused on adversarial reporting and more accountable to diverse communities. Reinventing jour-
nalism might involve subsidies for an expanded public media system, tax incentives to encourage
struggling media institutions transition into low- and nonprofit status, and government-sponsored
research and development efforts for new digital models, including public/private hybrids (see:
Pickard, 2015a: 212–231). While not a silver bullet, subsidies can reduce market pressures to allow
media institutions to become more adversarial and accountable to diverse communities by reducing
distortionary market pressures and help restore journalism’s public service mission. In addition to
subsidizing public journalistic alternatives, policy interventions focusing on the infrastructure side
of news media dissemination could seek to rein in monopolistic internet service providers. This
would require enforcing public interest protections like net neutrality and encouraging competition
by allowing communities to offer their own broadband services.
Economic and regulatory discourse about journalism

Looking at the discourses that define our understanding of journalism can reveal a great deal.
Recent years have seen the prevalence of ‘crisis narratives’. The Pew Research Center’s 2016
“State of the News Media” report was strikingly dire. It stated that 2015 was the worst year for
newspapers since the Great Recession. After showing that daily circulation, advertising revenue,
and newsroom staffing had all significantly declined, the report concluded: “this accelerating
decline suggests the industry may be past its point of no return” (Barthel, 2016). Coming from
the venerable Pew Research Center, such a conclusion speaks volumes. If this imminent collapse
is true, it is a serious social problem worthy of a national conversation. But thus far, no such
discussion has occurred, and there has been virtually no public policy response.
216
Part of this inaction can be traced back to policy battles in the 1940s (Pickard, 2015a), which
resulted in a social contract between the state, the public, and media institutions that I call the
‘Postwar Settlement for American Media’. This arrangement was defined by three features:
self-regulation, industry-defined social responsibility, and a negative understanding of the First
Amendment – negative in the Isaiah Berlin “negative freedom” sense, which emphasizes a lib-
ertarian individualist ‘freedom from’, as in media owners’ freedom from government regulation
as opposed to the public’s positive ‘freedom for’ or ‘freedom to’ a diverse media system (Berlin,
1969). This kept in place a commercial media system with little public or governmental oversight
and few challenges from noncommercial media. This framework largely remains the dominant
paradigm for American media policy today – an ideological arrangement I refer to as “corpo-
rate libertarianism” (Pickard, 2015a). This assumes that government has little legitimate role in
intervening in media markets, which, as noted at the beginning of this chapter, is a libertarian
fantasy, since the government is always present – just usually in ways that benefit media owners.
In our current context, a lack of a policy response to the journalism crisis thus far can be
attributed to several discursive impediments. First and foremost is the market fundamentalism
that has seeped into the master categories of how we think and talk about journalism. This
framework has led to certain assumptions that continue to depoliticize the design of journalistic
institutions, practices, and policies. First, the state of journalism understood as supply and demand
treats news as a simple commodity bought and sold on the market, not as a vital public service.
The implication is that if journalism is unprofitable for publishers and media owners, then we
should let it wither. Of course supply and demand in the unfettered free market does not always
reflect accurate assessments of social value. And in many cases, a demand for news still exists, but
the market will not support it.
A second assumption is that the institutional collapse of journalism is beyond our control,
something that happens to us, like a natural disaster or an act of God. Analysts like Clay Shirky
take a Schumpeterian view that we are living in a revolutionary moment of creative destruc-
tion where new alternatives will soon organically emerge (2011). We simply have to wait for
something new to organically emerge. While this argument is seductive – seemingly radical in
its encouragement for replacing obsolete institutions with innovative models – it assumes that
the market should be the final arbiter of what kind of journalism we should receive. Moreover, it
ignores the deeply institutional and social nature of journalism that should not be left to rise or
fall according to market forces.
A third assumption is that market forces and new technologies will combine to guide us out
of this predicament: paradoxically, even though the internet is often seen as the sole cause of the
journalism crisis (instead of the overreliance on the advertising revenue model), digital technolo-
gies are also often assumed to be journalism’s sole saviors. To be sure, new technologies and their
affordances can help, but these efforts need to be guided by sound public policy and public input.
Furthermore, professional norms and other cultural factors can go a long way toward ensuring
that good journalism persists. But putting our faith entirely in the market to provide the news
media that our democracy requires has always been, and is especially now, a risky proposal.
As a kind of antidote for this market fundamentalism, a number of political economic ratio-
nales exist for nonmarket support for journalism. For example, news and information can be
treated as public goods. Because public goods are nonrivalrous (one person’s consumption
doesn’t detract from another’s) and nonexcludable (difficult to prevent ‘free riders’), they differ
from other commodities within a capitalistic economy. Many public goods such as clean air,
parks, and even knowledge also produce tremendous positive externalities (benefits that accrue to
parties outside of the direct economic transaction). Society requires these goods, but individuals
typically undervalue (are unable or unwilling to pay for) these goods, leading to their underpro-
duction and “systemic market failure” (Pickard, 2014a).
217
Victor Pickard
This category of market failure deserves closer attention. Many scholars conclude their analy-
sis with the invocation of public goods, but we should extend the argument to highlight struc-
tural flaws that justify – indeed, necessitate – policy intervention into media markets. ‘Market
failure” generally refers to the market’s inability to efficiently allocate important goods and ser-
vices. The traditional commercial model overlooked the fact that news was a kind of by-product
of the transaction between advertisers and media owners (Pickard, 2015a: 214). Advertisers have
never really cared about whether their revenues supported foreign bureaus or good local news;
they were after markets. Advertising itself was a kind of subsidy. Jay Hamilton’s (2016) recent
book Democracy’s Detectives shows how investigative reporting has always been very expensive for
news organizations, but it affords many benefits for society as a whole.
New areas for regulatory concern

Gaps in scholarly research often track with growing regulatory quandaries. A number of rising
problematic trends facing digital journalism are worthy of regulatory oversight, ranging from
broadband deployment to the rise of run-amok advertising. One example of the latter is the pro-
liferation of sponsored content or native advertising (see Ferrer-Conill and Karlsson, this volume,
Chapter 35). Often presented as news, this form of advertising ranges from the fairly innocuous
“infomercial” to a more problematic variety of corporate propaganda. Ethical concerns about
misinformation, public trust, and social responsibility are valid, especially as this form of advertis-
ing becomes more ubiquitous online. Greater transparency is always a good measure, but some
degree of deception is often inherent in this practice. Branded or native advertising blurs the
distinction between news and advertising, usually designated by small print that says sponsored
by a particular company. Studies consistently show that readers do not catch the distinction and
are essentially deceived (Sass, 2015). Determining what standards would be truly appropriate for
a democratic society should be open to public debate.
As struggling news organizations are increasingly reliant on deceptive and invasive forms of
advertising, another area of concern is online behavioral advertising, which relies on an ethically
dubious form of behavioral tracking. One study shows that news organizations are among the
worst culprits in exposing readers to third-party advertisers and data brokers online, with an aver-
age of 19 third parties compared to an overall average on non-news websites of 8 third parties.
On a good day the New York Times might subject readers to a whopping 44 third parties without
readers knowing it (Libert and Pickard, 2015). Many of these sites might be innocuous, but some
are often not, and there is very little oversight. The growing prevalence of ad blockers is chang-
ing this calculus somewhat, but it remains a persistent problem in need of public discussion and
possibly regulatory oversight.
Another area of concern worth noting is the growing power of internet monopolies. Media
ownership policy, which may sound quaint in today’s digital age, is still a major concern for jour-
nalists. Whether considering new digital media monopolies like Comcast and Verizon or legacy
newspapers, the structural policies that allow media companies to merge and acquire new prop-
erties can have dramatic consequences for journalism. For example, when media outlets combine
and conglomerate, journalism jobs are often lost. Major media mergers are currently pending in
the US, such as the mergers of AT&T and Time Warner and of Sinclair and Tribune, whose pas-
sage would have a potentially profound effect on American journalism. And platform monopo-
lies such as Facebook and Google also have profound power over the future of journalism.
A final area of concern involves key internet policies. While questions pertaining to media
conduits and content are often dealt with separately, internet policy questions, such as the lack
of quality internet access, are central to journalism’s growth. Poor internet access, for example, is
an infrastructure problem with serious implications for the future of news media, but it is rarely
218
addressed within journalism studies. The ‘digital divide; in the US means nearly a fifth of all US
households still lack broadband internet (Federal Communications Commission, 2015). Even
for those with access, services are often subpar and costly compared to other democracies. This
problem could be somewhat alleviated by allowing communities to offer their own broadband
services, and providing internet subsidies to low-income Americans would expand access and
facilitate the flow of online news media.
Problems related to accessing digital news media also come to the fore in the recent net neu-
trality debate, which reveals the artifice of divisions between journalism and policy. Furthermore,
since we know that the future of journalism is largely digital, we should be talking more about
internet policies that dictate the ownership and control of digital infrastructures. For example,
lacking net neutrality protections – a distinct possibility given recent changes by Donald Trump’s
FCC under Ajit Pai to weaken net neutrality – a “cable-ized” internet might transform into a
hostile environment for all media makers, ranging from professional print journalists to amateur
bloggers who cannot afford to pay exorbitant fees to internet service providers and who are
consequently relegated to digital ‘slow lanes’ and essentially ghettoized online. These inequi-
ties are neither accidental nor inevitable; they result from explicit policies that accommodate
oligopolistic markets and corporate interests more generally (McChesney, 2013). Because they
directly impact on the flow of online news media, these policy issues should be a central concern
for journalism studies.
Why media policy matters for journalism

Looking ahead, we need a clear-eyed view of the structural problems facing journalism. Not
only does this entail understanding that questions of media ownership and control are central to
journalism’s future, but it also necessitates an understanding that government will have to take a
more active role. This is especially challenging in the US, where for many years a “corporate lib-
ertarian” paradigm has dominated policy debates and First Amendment absolutism impeded gov-
ernment intervention on behalf of news institutions. Nonetheless, scholars can play an important
role in clarifying what is at stake and what policy interventions are necessary to guarantee a viable
system for public service journalism.
Returning to the idea of media subsidies, while it is not likely to happen at the federal level any-
time soon, the local level may hold promise. For example, recent campaigns have sought to allocate
money raised by public television stations from the recent spectrum auction toward a fund for local
journalism. The US could also leverage already-existing public infrastructure. One possibility is to
transform post offices into local community media centers. In addition to providing public internet
access, these spaces could help facilitate local reporting through various media – including print,
digital platforms, and low-power radio stations. The Urbana-Champaign Independent Media Cen-
ter launched a similar model over a decade ago when it raised money to buy the downtown post
office for community media-making and other progressive projects (Pickard, 2015a).
We can also look at historical models. One model worth more attention was an experiment
with subsidized municipal newspapers in the early 1900s (Pickard, 2011). Another idea is to
tweak tax laws to facilitate struggling commercial newspapers into low- and nonprofit status to
make it easier for them to receive charitable contributions from philanthropists and foundations.
Many observers have long argued that ultimately we need a large centralized public media trust
that can allocate resources with no strings attached to support the kinds of public service journal-
ism that the market will never sustain (Pickard, 2015a). British reformers are pushing for a bill
right now that would force Google and Facebook to devote 1% of their revenues toward such a
fund (Greenslade, 2016). Creative methods exist for generating this money; what is lacking is the
imagination and political will to implement them.
219
Victor Pickard
Structural reforms are extremely difficult until we reframe the debate about journalism and
wrench it from its market fundamentalist framework. This debate should involve everyone,
including journalists, but also technologists and public advocates. It should address the future of
journalism as a problem for all of society. We need to return to normative foundations of what
journalism is supposed to do in a democracy, with an emphasis on local journalism and openness
toward non-market-based means of support. This discussion should include a debate about the
social responsibilities of media platform monopolies like Facebook as it increasingly becomes a
leading news source. Mark Zuckerberg has refused to even acknowledge that Facebook is any-
thing more than a technology company (Ingram, 2016).
With so much hope being placed on civil society – especially the press – to serve as a kind
of countervailing force against governmental overreach, now could be a real opportunity for
strengthening public service journalism. Journalism scholars have an important role to play in
redefining journalism ethics and news media’s social responsibilities for the digital age. With this
in mind, media policy’s relationship to journalism should be of greater concern now than ever
before. Laws and policies dealing with intellectual property, privacy, and net neutrality will lay
the foundation for journalism’s digital future. Policy is a social contract for all of us – for the
press, for the government, and for the public. If journalists do not turn on to media policy, media
policy might turn on them.
Further reading
Researchers who are interested in reading more about the intersections of journalism studies and
media policy may wish to consult the following books: America’s Battle for Media Democracy by
Victor Pickard; The Death and Life of American Journalism by Robert McChesney and John Nich-
ols; The Politics of Media Policy by Des Freedman; and Saving the Media by Julia Cage.
References
Barthel, M. (2016, June 15) “5 key takeaways about the state of the news media in 2016.” Pew Research
Center. Retrieved from www.pewresearch.org/fact-tank/2016/06/15/state-of-the-news-media-2016-
key-takeaways/
Benson, R. (2013) Shaping Immigration News: A French-American Comparison. New York, NY: Cambridge
University Press.
Benson,R. and Pickard,V. (2017,August 11) “The slippery slope of the oligarchy media model.”The Conversation.
Retrieved from https://theconversation.com/the-slippery-slope-of-the-oligarchy-media-model-81931
Benson, R. and Powers, M. (2011) Public Media and Political Independence: Lessons for the Future of Journalism
From Around the World. Washington, DC: Free Press.
Berlin, I. (1969) “Two concepts of liberty.” In Four Essays on Liberty. Oxford: Oxford University Press.
Cage, J. (2016) Saving the Media. Cambridge, MA: Harvard University Press.
Downie, L. and Schudson, M. (2009) “The reconstruction of American journalism.” Columbia Journalism
Review. Available at: https://archives.cjr.org/reconstruction/the_reconstruction_of_american.php
Edmonds, R., Guskin, E., Mitchell, A. and Jurkowitz, M. (2013) “Newspapers: Stabilizing but still
threatened.”The State of the News Media 2013, Pew Research Center’s Project for Excellence in Journal-
ism. Available at: http://stateofthemedia.org/2013/newspapers-stabilizing-but-still-threatened/
Federal Communications Commission. (2015) “2015 Broadband Report.” Available at: https://www.fcc.
gov/reports-research/reports/broadband-progress-reports/2015-broadband-progress-report
Federal Trade Commission. (2010) The FTC in 2010: Federal Trade Commission Annual Report. Available
at: https://www.ftc.gov/sites/default/files/documents/reports_annual/annual-report-2010/2010chair
mansreport_0.pdf
Greenslade,R. (2016, November 8) “Make Google and Facebook pay for public service reporting.”The Guardian.
Retrieved from www.theguardian.com/media/greenslade/2016/nov/08/make-google-and-facebook-
pay-for-public-service-reporting
220
Hamilton, J. (2016) Democracy’s Detective: The Economics of Investigative Journalism. Cambridge, MA: Harvard
University Press.
Ingram, M. (2016, August 30) “Sorry Mark Zuckerberg, but Facebook is definitely a media company.” For-
tune. Retrieved from http://fortune.com/2016/08/30/facebook-media-company/
Kammer, A. (2016) “A welfare perspective on Nordic media subsidies.” Journal of Media Business Studies,
13(3), 140–152.
Knight Foundation. (2009) Knight Foundation Initiative seeks new models for investigative reporting. Available at:
www.knightfoundation.org/news/press_room/knight_press_releases/detail.dot?id=348319
Libert, T. and Pickard, V. (2015) “Think You’re Reading the News for Free? New Research Shows You’re Likely Pay-
ing with Your Privacy”, The Wire. Available at: https://thewire.in/media/think-youre-reading-the-news-
for-free-new-research-shows-youre-likely-paying-with-your-privacy
McChesney, R. (2013) Digital Disconnect. New York: The New Press.
McChesney, R. and Nichols, J. (2010) The Death and Life of American Journalism: The Media Revolution that
Will Begin the World Again. New York, NY: Nation Books.
Molla, R. and Ovide, S. (2016, May 23) “New media shares old media’s roof.” Bloomberg.
Murschetz, P. (2014) State Aid for Newspapers – Theories, Cases, Actions, Berlin: Springer-Verlag.
Neiva, E. M. (1996) “Chain building: The consolidation of the American newspaper industry, 1953–1980.”
Business History Review, 70(1), 1–42.
Pew Research Center’s Project for Excellence in Journalism. (2010, January 11) The Study of the News Eco-
system of One American City. Retrieved from www.journalism.org/analysis_report/how_news_happens
(accessed 2 February 2013).
Pew Research Center. (2014) State of the News Media. Retrieved from: http://assets.pewresearch.org/wp-
content/uploads/sites/13/2017/05/30142556/state-of-the-news-media-report-2014-final.pdf
Picard, R. (2002) “US newspapers ad revenue shows consistent growth.” Newspaper Research Journal, 23(4),
21–33.
Picard, R. (2008) “Shifts in newspaper advertising expenditures and their implications for the future of
newspapers.” Journalism Studies, 9(5), 704–716.
Pickard, V. (2011) “Can government support the press? Historicizing and internationalizing a policy
approach to the journalism crisis.” The Communication Review, 14(2), 73–95.
Pickard, V. (2014a) “The great evasion: Confronting market failure in American media policy.” Critical Stud-
ies in Media Communication, 31(2), 153–159.
Pickard, V. (2014b) “Wither(ing) journalism.” Public Culture. Retrieved from www.publicbooks.org/
nonfiction/withering-journalism
Pickard, V. (2015a) America’s Battle for Media Democracy: The Triumph of Corporate Libertarianism and the Future
of Media Reform. New York, NY: Cambridge University Press.
Pickard, V. (2015b) “The return of the nervous liberals: Market fundamentalism, policy failure, and recur-
ring journalism crises.” The Communication Review, 18(2), 82–97.
Pickard, V. (2017a) “A social democratic vision of media: Toward a radical pre-history of public broadcast-
ing.” Journal of Radio and Audio Media, 24(2), 200–212.
Pickard, V. (2017b) “Rediscovering the news: Journalism studies’ three blind spots.” In P. Boczkowski and
C. W. Anderson (eds.), Remaking Digital News. Cambridge, MA: The MIT Press.
Pickard, V. (2017c) “Structural collapse: The American journalism crisis and the search for a sustainable
future.” In P. Berglez, U. Olausson and M. Ots (eds.), What Is Sustainable Journalism? Integrating the Envi-
ronmental, Social, and Economic Challenges of Journalism. New York, NY: Peter Lang (pp. 351–366).
Pickard, V. Stearns, J. and Aaron, C. (2009) Saving the News: Toward a National Journalism Strategy. Wash-
ington, DC: Free Press.
Pickard, V. and Williams, A. (2014) “Salvation or folly? The perils and promises of digital paywalls.” Digital
Journalism, 2(2), 195–213.
Plunkett, J. (2016, May 11) “BBC to fund 150 local news journalists.” The Guardian. Retrieved from www.
theguardian.com/media/2016/may/11/bbc-to-fund-150-local-news-journalists
Rosentiel, T. and Mitchell, A. (2012) “State of the News Media,” Pew Research. Available at: http://
assets.pewresearch.org/wp-content/uploads/sites/13/2017/05/24141622/State-of-the-News-Media-
Report-2012-FINAL.pdf
Sass, E. (2015, December 31) “Consumers can’t tell native ads from editorial content.” MediaPost. Retrieved
from www.mediapost.com/publications/article/265789/consumers-cant-tell-native-ads-from-editorial-
con.html
221
Victor Pickard
Shirky, C. (2011) “Newspapers and thinking the unthinkable.” In Will the Last Reporter Please Turn Out the Lights:
The Collapse of Journalism and What Can Be Done to Fix It. New York, NY: The New Press (pp. 38–44).
Soloski, J. (2013) “Collapse of the US newspaper industry: Goodwill, leverage and bankruptcy.” Journalism,
14(3), 309–329.
Starkman, D. (2016, September 28) “The ever-expanding media giants.” Traffic Magazine.
Stites, T. (2011) “Layoffs and cutbacks lead to a new world of news deserts.” NiemanLab. Retrieved from
www.niemanlab.org/2011/12/tom-stites-layoffs-and-cutbacks-lead-to-a-new-world-of-news-deserts/
Stulberg, A. (2017, September 22) “In paywall age, free content remains king for newspaper sites.” Columbia
Journalism Review. Retrieved from www.cjr.org/business_of_news/newspaper-paywalls.php
Waldman, S. (2011) The Information Needs of Communities: The Changing Media Landscape in a Broadband Age.
Washington, DC: Federal Communications Commission. Retrieved from www.fcc.gov/infoneedsreport
Williams, A. and Pickard, V. (2016) “The costs of risky business: What happens when newspapers become
the playthings of billionaires?” Poster session presented at Association for Education in Journalism and Mass
Communication, 4–17 August, Minneapolis, MN.
222
PART IV
Developing digital journalism

practice
17
DEFINING AND MAPPING
DATA JOURNALISM AND
COMPUTATIONAL JOURNALISM
A review of typologies and themes
Mark Coddington
In every era, there is a subset of journalistic practice that draws substantial attention (perhaps
inordinately so) from journalists and especially scholars of journalism, outpacing the prevalence
of the practice itself or its interest among the public. During the 1990s, the discussion of public
journalism ran throughout both journalism scholarship and the news industry itself, while its
practice was largely limited to a few well-publicized examples. In the 2000s, citizen journalism
became an industry buzzword and the subject of much scholarly rumination, though success-
ful instances of citizen-led news operations and projects were difficult to come by. During this
decade, data journalism and its many related forms has become an object of scholarly fixation and
much excitement in the industry and in journalism education. Numerous news organizations
developed specialized data teams, global data journalism awards were instituted, and attendance
at the annual conference of the seminal U.S. group National Institute for Computer-Assisted
Reporting (NICAR) tripled between 2010 and 2014 (Stiles, 2017). In academia, universities
around the world developed data journalism courses to meet the industry’s demand for those
skills (Griffin and Dunwoody, 2016; Hewett, 2016), and scholarship on data journalism boomed.
Nevertheless, data journalism remains only a small subset of journalism, something that exists
only in precarious forms, if at all, in many smaller newspapers and news organizations (Fink
and Anderson, 2015). Likewise, scholarship on data journalism remains only a niche within
journalism studies, published by a couple of dozen scholars worldwide, largely in a handful of
journalism-centric journals. But that small constellation of scholars has produced a flurry of
research over the past decade, often overlapping and occasionally contradicting each other in its
rush to document and understand a rapidly emerging phenomenon.
I offered a typology for understanding data journalism, computational journalism, computer-
assisted reporting, and other forms of quantitative journalism in a 2015 journal article I opti-
mistically titled “Clarifying Journalism’s Quantitative Turn” (Coddington, 2015). Little clarity
has emerged since then, with academic studies continuing to churn out, and many of them
either developing another novel way to define and organize forms of quantitative journalism or
largely side-stepping the issue of definition and classification. But I hope to continue moving
toward clarity in this chapter by examining some of those typologies (both complementary and
competing) and outlining themes that have begun to congeal in scholars’ characterizations of
225
Mark Coddington
data journalism, computational journalism, and other similar practices while acknowledging the
breadth of scholarship and perspectives on this issue.
First, though, it is important to clarify why such an endeavor is important in the first place.
Why devote so much attention to the classification and definition of various slightly differenti-
ated specialized forms of journalism? Why not simply focus on the practices themselves, with-
out fixating on what they should be called or how they should be grouped? This is a common
sentiment among journalists themselves – including, famously, data journalism pioneer Adrian
Holovaty (2009) – but one that ultimately should be guarded against, at least among scholars. I
echo the arguments made by Nikki Usher (2016: 73–74; see also Usher, this volume, Chapter
26) in defense of such typological work by noting that as buzzwords within the industry, terms
like data journalism and computational journalism are in danger of being discarded as meaning-
lessly trendy (and later, passé) without careful reflection that can attach substantial meaning to
them. This type of categorization also reminds us that these terms and practices have institutional,
cultural, and epistemological roots; they did not emerge out of nowhere, nor are they merely
new names slapped onto an existing practice like computer-assisted reporting. As Usher argues,
this work “helps us understand what is novel, old, and the same about these groups and as they
compare to traditional journalism” (2016: 74). There may come a time when further work on
the classification of these concepts becomes a pedantic and picayune exercise without real value,
but the subfield is still too jumbled to have reached that point yet.
The state of data journalism scholarship

In the introduction to a special issue on data journalism edited for Digital Journalism that origi-
nally appeared online in 2014, Seth Lewis described the scholarship on quantitative journalism
as a “rapidly growing body of work” that is “seemingly ‘everywhere’ based on the industry buzz
and accelerating scholarly interest” (Lewis, 2015: 322–323). Lewis was right about acceleration;
research into data journalism has only increased since then. Typical of early research into a phe-
nomenon, much of that data journalism research has been exploratory and descriptive (Ausser-
hofer et al., 2017). Early on in the current wave of scholarship, C. W. Anderson voiced concern
that research on computational journalism was largely internalist, with “a tendency to consider
the problems of journalism scholarship from the point of view of the journalism profession”
(2013: 1007). Such early internalist work has remained at the forefront of scholarship, picking up
numerous citations, particularly in computational journalism (Ausserhofer et al., 2017;Young and
Hermida, 2015). But research since then has eased off this more utilitarian perspective, with sev-
eral studies taking a more conceptual or critical tack on journalism practice (Borges-Rey, 2017,
and this volume, Chapter 21; Bucher, 2017; De Maeyer et al., 2015; Hammond, 2017). Some of
the most useful work conceptually has centered on the epistemological approaches of various
forms of quantitative journalism, using them to trace shifts in the philosophical frameworks and
processes of knowledge production among journalism more broadly (Anderson, forthcoming;
Borges-Rey, 2016, 2017; Parasie, 2015; Parasie and Dagiral, 2013).
Throughout the development of scholarship in this field, researchers (including myself ) have
complained of the proliferation of terms used to describe quantitative forms of journalism, yet the
litany of labels continues to grow. In addition to the central terms of data journalism (and data-
driven journalism), computational journalism, and the much older computer-assisted reporting
(CAR), researchers have described these and related phenomena as “computational exploration
in journalism” (Gynnild, 2014), “interactive journalism” (Usher, 2016), “big data journalism”
(Tandoc and Oh, 2017), and others. In the last few years, automated forms of journalism have also
arisen as a key set of journalistic practices adjacent to these, and those have been given various
226
Defining and mapping data journalism
names like ‘automated journalism’, ‘algorithmic journalism’, and ‘robot journalism’, as well (Dörr,
2016; Splendore, 2016).
As others have noted (Bucher, 2017; Hammond, 2017), the fact that scholars have hurried
to come up with so many different terms for similar journalistic phenomena is notable in itself,
an indication of the rapid change in the journalistic field and haziness in the scholarly thinking
about it – and perhaps the need to distinguish oneself in a crowded academic marketplace as well.
Though the maze of terms and concepts is indeed a sign of confusion and a bit of franticness
in scholarship in this area, that’s not necessarily a cause for deep concern (yet). This is a natural
stage when scholars are encountering widespread new phenomena that are practiced differently
in a variety of contexts and that change quickly. In part, the confusion stems from the fact that
scholars are torn between borrowing terms from the profession and developing their own. The
terms from the profession (such as data journalism) are far more widely used and therefore seem
more ‘natural’, but since journalists themselves do not agree on what those terms mean, they
bring a variety of sometimes conflicting meanings into the academic context. But when scholars
try to counter this ambiguity by developing their own terms, they seem unnecessary and out of
step with the way practices are being talked about on the ground. When the actors being studied
have themselves developed widely used but imprecise terms about the work that is being studied,
scholars are left with the unenviable task of either imposing clarity on those terms or supplanting
them in the academic discourse.
And yet it is a task that must be undertaken. As data journalism scholarship transitions out
of its initial exploratory phase into deeper and more conceptually rich research, it also needs to
begin pinning down clear, stable definitions and conceptualizations of the practices it is studying.
This is a necessary stage in the maturation of scholarship in this area, and it is one that researchers
should be ready by now to enter.
Typologies of data journalism practices

A few scholars have attempted to move this clarifying work forward with typologies organizing
and distinguishing between various types of quantitative journalism. Astrid Gynnild’s (2014) was
the first significant attempt to formally parse multiple forms of this type of work, as she gave
the umbrella term “computational exploration in journalism” to work that encompassed use of
algorithms and social science methods to retrieve, analyze, and visualize data. She distinguished
between three approaches to computational exploration in journalism: a newsroom approach
built on journalistic traditions that included computer-assisted reporting (CAR) and data jour-
nalism; an entrepreneurial approach that involves creating databases for web or mobile apps and
included journalism as programming; and an academic approach that consisted predominantly of
computational journalism. Gynnild’s typology is notable for including data journalism and CAR
within the same category, and it relies heavily on the journalistic setting in which the practice
takes place. This places a valuable emphasis on professional values and influences, though it leaves
aside important epistemological elements that others have fruitfully used to distinguish between
practices. In some cases, the setting on which this typology relies can be fungible; a key aspect
Gynnild uses to characterize the entrepreneurial approach is that it often takes place outside
traditional newsrooms or on the margins of the profession. This can change quickly and lead to
blurring: if a data journalist leaves the New York Times to do similar work at a news start-up, does
she switch from the newsroom approach to the entrepreneurial approach by virtue of her new
place of employment?
My own typology (Coddington, 2015) outlines three forms of quantitative journalism: data
journalism, computational journalism, and CAR. It then introduces four dimensions with which
227
Mark Coddington
to evaluate them: a closed and professional vs. open and networked orientation, transparency
vs. opacity, epistemological orientations toward targeted sampling vs. big data, and an active
vs. passive vision of the public. This typology introduces both epistemological and professional
elements, as well as audience-related factors, to distinguish types of quantitative journalism. (In
contrast to Gynnild, I place CAR as a distinct practice from data journalism, which I classify as
closer to computational journalism.) But though its four factors can be used on any journalis-
tic case, it is not directly grounded in empirical research. Its characterizations of data journalism,
computational journalism, and CAR are Weberian ideal types based on a review of literature
and discourse rather than arising out of empirical data gathered from journalistic practice itself.
At least two other scholars have attempted to expand on this typology: Splendore (2016) adds
a three-part comparative dimension to the four originally laid out, looking at epistemological
implications in the access, process, and editing stages of news production. In the access stage,
he differentiates between manual and automated access to data; in the process stage, between
hypothesis-based investigation and data-driven analysis; and in the editing stage, between design
for a broader public or for individual personalization. Splendore analyzes data journalism, com-
putational journalism, and automated journalism through these new dimensions (he drops CAR
from his analysis), finding data journalism to be largely distinct from the more similar compu-
tational and algorithmic journalism. Splendore’s expansion to the typology is a useful process-
based addition, though like my original typology, it consists of ideal types of practices and thus
requires generalizations that may not match up with actual observed practice.
Michalski (2016) gives the typology a more empirical application with the development of
a content analysis scheme based largely off the four dimensions in the original typology (Cod-
dington, 2015) as well as elements drawn from numerous other studies. He uses this operational-
ization to compare projects by the Guardian and the Washington Post, placing the Guardian project
closer to data journalism and the Post project near CAR. Michalski’s 32-item coding scheme
could use some refinement in several spots, but he does valuable work in bridging the conceptual
framework with the kind of quantitative empirical analysis of journalistic output that has been
largely absent from work in this area (Ausserhofer et al., 2017). Both Splendore and Michalski’s
studies are useful extensions of the typology that expand its conceptual range and empirical
applicability, though both hold the same limitation as the original typology: it is a helpful way of
seeing dimensions by which these forms of journalism might be evaluated, though it only does
initial work in actually evaluating and distinguishing the journalistic forms themselves.
In her book on interactive journalism, Nikki Usher (2016) introduces a typology that diverges
from mine and Gynnild’s (2014) in several respects. First, Usher’s is not at root a typology of
journalistic practices but a typology of the people who engage in them. Second, Usher’s area of
concern only partly overlaps with the quantitative forms examined here; she is primarily focused
on classifying the types of people who work to produce news interactives, some but not all of
whom work with data or computation. Third and most helpfully, Usher’s typology is built on
field-based evidence, arising in a grounded manner from the way journalists themselves talk
about who they are. She distinguishes between hacker journalists, programmer journalists, inter-
active journalists, and data journalists, finding that data journalists have the closest connection of
those groups to traditional journalism, through CAR. Her categories are not mutually exclusive,
though; she describes many hacker and programmer journalists as doing data journalism, but
notes that some data journalists aren’t programmers and don’t code at all. Data journalists, she
says, are defined by their orientation toward data: “they are primarily working with data in the
service of stories, actively trying to tell stories with data, and spend most of their time working
specifically with data” (2016: 90). In addition to her bottom-up generation from journalistic
practice, Usher’s study is also valuable for its classification of all of these forms as a professional
228
subspecialty, which absorbs many of the norms and values of professional journalism but offers a
distinct knowledge claim to help expand the profession’s jurisdiction and reassert its relevance.
Though the professional subspecialty she describes is interactive journalism rather than data jour-
nalism, many of its characteristics transfer over smoothly and form an apt theoretical framework
for thinking about data journalism as a set of practices.
Borges-Rey (2017) synthesizes the types laid out by myself and Gynnild (2014) as well as
Parasie’s research (Parasie, 2015; Parasie and Dagiral, 2013) into two basic epistemological para-
digms: the “newshound”, which is subordinate to journalism’s professional norms and traditional
ways of dealing with data, and the “techie”, which switches between journalistic and computa-
tional mindsets in approaching data. Borges-Rey places CAR in the newshound paradigm and
computational journalism in the techie paradigm, though data journalism is split: as Gynnild
(2014) conceives of it, it is in the newshound approach, and as I conceive of it, it is in the techie
approach. Borges-Rey’s simple typology is useful as an overarching framework, and it addresses
both the professional setting and epistemology together. It doesn’t add to our ability to make
fine-grained distinctions between forms (for example, the distinction between data journalism
and computational journalism), but it does helpfully compress many of the themes Gynnild and
I identify into a single unified spectrum.
Though it is perhaps a bit excessive to have at least four different typologies of a subfield
emerge within three years, these frameworks are more complementary than competitive. Each
serves a different purpose and best applies to different conceptual tasks. Borges-Rey’s (2017) is
most valuable as a general orienting device to gauge the overall position of a journalistic prac-
tice, while mine (Coddington, 2015) offers a toolkit to make more specific distinctions between
various aspects of different practices. Usher’s (2016) is most helpful for analyzing the roles, back-
grounds, and orientations of the individual actors who are engaging in those practices, especially
given the diverse roles and approaches of the people who work in this area, even within the same
project. Scholars shouldn’t feel obligated to use all of them – that would probably be conceptu-
ally counterproductive – but any of them might prove the most useful tool for the analytical job
at hand.
Characterizing data journalism and computational journalism

Having examined the broader typologies by which these forms are organized, it’s important also
to look more closely at the forms themselves and the emerging scholarly consensus about their
chief characteristics, alongside the disagreements about what values and practices they constitute.
For both data journalism and computational journalism, some general themes have arisen that
can help to broadly characterize both forms, even as they continue to shift and our knowledge
of them expands.
Data journalism
The degree to which data journalism is tied to previous forms of journalism such as CAR is a key
point of contention among scholars, but the story of how it rose to its newfound place of promi-
nence has begun to solidify. Programmer-journalist Adrian Holovaty is often seen among both
journalists and scholars as its most influential early innovator, with his 2006 blog post outlining
a vision of journalism as structured data (Holovaty, 2006) and his late 2000s projects fleshing out
that vision, ChicagoCrime and EveryBlock (Anderson, forthcoming; Ausserhofer et al., 2017;
Gray et al., 2012). The Guardian also plays a crucial role in accounts of data journalism’s rise, with
its 2009 crowdsourced project analyzing UK MPs’ expenses and 2010 analysis and presentation
229
Mark Coddington
of WikiLeaks’ large-scale document leaks as key projects bringing data journalism into the public
consciousness. Its Datablog, initially edited by Simon Rogers, extended that work in highlighting
data-driven journalistic work on a regular basis (Gray et al., 2012; Howard, 2014; Knight, 2015).
Soon afterward, others have identified 2013 and 2014 as a time when discourse about data jour-
nalism boomed, and the term began serving as a major form of self-identification for journalists
(Howard, 2014; Royal and Blasingame, 2015).
Scholars have had a difficult time arriving at an agreed-upon definition for data journalism,
which reflects the amorphous nature of the practice itself. Howard’s (2014) is probably the most
robust and straightforward, but even this is presented in a few parts: He defines data journalism
as, “gathering, cleaning, organizing, analyzing, visualizing, and publishing data to support the
creation of acts of journalism” and also as the application of data science, or the extraction of
knowledge from data, to journalism, as well the combination of treating data as a source, applying
statistics to interrogate it, and using visualizations to present it (ibid.: 4–5). As Ausserhofer and
colleagues (2017) and Royal and Blasingame (2015) both note, scholars have differed on whether
data journalism is fundamentally about a process or product, and I echo their assertion that it is
both, though I think the process – an epistemological and professional approach to gathering,
analyzing, and presenting data – is more fundamental to data journalism than any product.
The best way to characterize data journalism may not be through a precise definition but
through a set of themes that emerge across numerous scholarly analyses of it. The most central
theme in definitions of data journalism is storytelling – the idea that data journalism, as Howard
(2014) puts it, “is telling stories with numbers, or finding stories in them” (5). In Royal and Blas-
ingame’s (2015) review of 63 definitions of data journalism by academics and professionals, sto-
rytelling emerges as the most common theme and in some cases is the sole or primary element of
the definition (Mair and Keeble, 2013; Splendore, 2016). Storytelling is a particularly focal point
in data journalists’ understandings of their own work, where they see telling stories as the end
and data simply as a means (Borges-Rey, 2016; Rogers, 2013). While this characterization of data
journalism as storytelling serves to link data journalism with the traditional journalistic paradigm,
other scholars have identified storytelling as a primary point of divergence between data journal-
ism and CAR. Whereas in CAR data primarily served an investigation, in data journalism, data’s
main end is as a story, potentially on its own and often through interactivity or personalization
(Felle, 2016; Usher, 2016).
Two other themes of data journalism are related – transparency and visualization. Rather than
allowing data to recede into the background of a story, data journalists tend to advocate trans-
parency that allows audiences to access the data on which the story is built, because of both the
influence of open-source philosophy and the increased demand for unfiltered information online
more generally (Coddington, 2015). This openness often involves publishing the full data online
as part of the project, though this is not always an element of data journalism, and the published
data has always been processed and formatted, in ways that are not always described to the audi-
ence (Lesage and Hackett, 2014; Usher, 2016). Transparency is also a key form of legitimation for
data journalists, providing some accountability for their work, though most audience members
do not have the skills to perform that kind of function (Borges-Rey, 2016; Lesage and Hackett,
2014). In many cases, the transparency work of the publication of data is done through a visual-
ization that may be interactive. In this way, visualization – a product that is a key but not neces-
sary element of data journalism – is tied not only to the idea of data as a story in itself but also
to transparency, by allowing audience members to drill into datasets and find elements relevant
to them (Gray et al., 2012; Tandoc and Oh, 2017). Not all data journalism involves visualization,
nor does it all involve interactivity. But the emphasis on using visualization and interactives to
aid transparency and allow audiences to act as coproducers of the truth claims of the journalistic
230
artifact is an important definitional characterization of data journalism as a whole (Splendore,

2016; Usher, 2016).
Crowdsourcing has also occasionally been a part of this transparency process of data jour-
nalism, particularly in high-profile examples such as the Guardian’s MPs’ expenses project or
ProPublica’s “Free the Files”, in which audiences analyzed massive troves of documents to help
journalists discern trends. As journalists have characterized data journalism, they have sometimes
included crowdsourcing as a key component (Gray et al., 2012; Rogers, 2013). But it’s simply not
widespread enough in practice to be classified as a central theme for data journalism. Especially
in more recent studies, scholars have found that while data journalists are receptive to crowd-
sourcing and see it as an important part of data journalism work, few have actually successfully or
consistently employed it (Borges-Rey, 2017; Felle, 2016). Crowdsourcing is an extension of data
journalism’s vision of an active public and transparent data analysis process, but only in rare cases
is it a significant part of actual data journalism practice.
Another element emphasized by some researchers but not fundamental to data journalism is
the size of datasets used. Ausserhofer and colleagues (2017) identify large datasets as a common
part of definitions of data journalism, and Uskali and Kuutti (2015) define data journalism as
exclusively working on large datasets. Indeed, data journalism tends to work with larger datasets
than CAR has, which entails an epistemological shift from sampling to extrapolate from smaller
datasets to mining and processing larger ones (Coddington, 2015). But while the number and
size of datasets freely available served as a catalyst for data journalism’s growth, the size of data
is not an inherent part of data journalism’s approach. Several studies have found that journalists
see a sharp distinction between the type of time-intensive, team-based, investigative data jour-
nalism that has attracted much of the scholarly and professional attention and the more quick-
turnaround, light, routine data journalism that happens on an everyday basis (Borges-Rey, 2016;
De Maeyer et al., 2015; Rogers, 2013). Uskali and Kuutti (2015) have given the most thorough
characterization of these two types (they call them “investigative data journalism” and “general
data journalism”). Although Uskali and Kuutti center their definition of data journalism on large
datasets, their conceptualization shows that a significant part of data journalism is smaller-scale
work, a finding confirmed by other researchers (Borges-Rey, 2016; De Maeyer et al., 2015; Fink
and Anderson, 2015).
The greatest disagreement among scholars regarding data journalism is over the degree to
which it represents a pattern of continuity with previous forms of quantitative journalism (par-
ticularly CAR) vs. substantial change. As Borges-Rey (2017) notes, this gets to the heart of
the question of what data journalism is: is it investigative journalism reborn with new tools at
its disposal? Or is it a new combination of the logics of computer science and journalism that
reshapes news production? Data journalists themselves tend to affirm the former, emphasizing
their connections to the normative aims of traditional investigative journalism and foreground-
ing the importance of the journalistic mission (Felle, 2016; Rogers, 2013; Royal and Blasingame,
2015). And epistemologically, data journalism is indeed built on the largely positivist traits of
objectivity that have governed journalism for decades, with its emphasis on ascertaining reality
through verified observation in the form of data (Lesage and Hackett, 2014). Data journalism
clearly has roots in CAR, whose key figures and institutions – notably the National Institute for
Computer-Assisted Reporting, or NICAR – continue to play a pivotal role in structuring and
developing data journalism as a subfield.
But some of data journalism’s defining traits also represent departures from previous forms of
quantitative journalism. Several of them have been outlined here: data journalism’s emphasis on
data itself as story, its embrace of transparency as a core norm, its focus on interactivity and per-
sonalization, and its orientation toward mining large datasets rather than sampling small ones are
231
Mark Coddington
all to varying degrees divergences from the paradigm of CAR. As Hammond (2017) notes, data
journalists have professional legitimacy to gain by emphasizing their continuity with traditional
journalism, which can lead us to overestimate that continuity based on their discourse. Anderson
(2015) describes the continuity-based perspective as “deeply and fundamentally flawed” (351)
and says quantitative journalism’s story in the United States is one of rupture. But he argues
that the fundamental changes that formed the foundation for data journalism took place not
within the past decade or two but during the 1960s and 1970s, with the development of pro-
cessing speed through computers, prestige for investigative journalism, a focus on patterns rather
than incidents, and a move beyond he-said, she-said journalism (Anderson, forthcoming). The
answer to the question of data journalism’s continuity with previous quantitative forms, then,
may depend largely on how far we zoom in or out. When we look closely at data journalism, we
see important fissures between it and CAR. But when we zoom out, we see that both are built
on the same substrate of positivistic values and methods fused with high-modern public service
journalism.
Computational journalism
Though computational journalism has drawn a substantial amount of scholarly attention, it is a
far smaller niche than data journalism. Early research on computational journalism overlapped
heavily with data journalism, with some of the same projects such as the Guardian’s MPs’ expenses
and EveryBlock serving as examples of both (Flew et al., 2012; Hamilton and Turner, 2009). But
in recent years it has been distinguished from data journalism as a primarily academic approach
to quantitative journalism (Gynnild, 2014; Royal and Blasingame, 2015; Usher, 2016). Compu-
tational journalism’s structural roots are in the American academy, with some early adoption in
Europe, starting with courses at the Georgia Institute of Technology and programs at Northwest-
ern University and Columbia University in the late 2000s (Anderson, 2013; Gynnild, 2014). Its
common definitions as the application of social science or computer science methods to journal-
ism (Flew et al., 2012; Hamilton and Turner, 2009; Usher, 2016) reinforce its deep academic ties.
But computational journalism has shifted into the professional sphere in recent years, largely
through the rise of automated and algorithmic journalism. Consider the shift in definitions
given to computational journalism by one of its pioneers, Nicholas Diakopoulos. In a brief but
influential paper on computational journalism published in 2011, Diakopoulos defined compu-
tational journalism in terms of “the application of computing and computational thinking to the
activities of journalism” (Diakopoulos, 2011). Five years later, he defined computational journal-
ism instead as “finding, telling, and disseminating news stories with, by, or about algorithms”
and referred to it interchangeably with algorithmic journalism (Diakopoulos and Koliska, 2017:
810; see also Koliska and Diakopoulos, this volume, Chapter 19). Computational journalism had
moved from being more generally the application of all forms of computing and computational
thinking to journalism to the more specific and practical application of algorithms to journal-
ism. This change reflected a fusion of the two ideas in the practice of computational journalism
itself, as the use of algorithms in journalism began to become, in many ways, the application of
computing to journalism. Diakopoulos’ practical merging of computational and algorithmic
journalism is a perceptive move, sharpening its distinction from data journalism and grounding
it more deeply in journalistic practice.
Beyond algorithms, the other core element of computational journalism is computational
thinking. As articulated by Wing (2008) and others, computational thinking is the process of
abstracting information or problems beyond their immediate material context and using funda-
mental concepts of computer science such as automation (this is where computational thinking is
232
especially closely tied to algorithms as an output) and recursivity. The thought process in compu-
tational thinking goes beyond mere abstraction, though – philosophers also specialize in abstract-
ing problems, although most of them could hardly be called computational thinkers. Specifically,
it involves approaching abstraction in computational terms – that is, binary code and the machine
processes that code triggers – and understanding information in terms of what can be encoded
and related between databases (Anderson, forthcoming). More broadly, it centers on the ques-
tion of what can and cannot be calculated, and how – which, Bucher (2017) notes, “is not just
a purely technical question; it is also a deeply social, cultural, political and economic one” (930).
These two elements – the cognitive process of computational thinking and the technical
output of automation or algorithms – complement each other to form the core components of
computational journalism, both as process and product. Where these elements had appeared in
relation to journalism almost entirely in an academic context several years ago, they are beginning
to make their way into professional practice through both processes of algorithmic accountability
(Diakopoulos and Koliska, 2017) and journalism as structured data (Anderson, forthcoming).
Finally, it’s important to note a couple of elements that are common among both data and
computational journalism, as well as adjacent forms such as CAR and algorithmic journalism. As
Splendore (2016: 345) notes, all of these forms are defined by their hybridity. They all represent
a convergence of several fields: journalism, computer science, social science, and data analysis and
visualization. This hybridity is not simply incidental; it is a constitutive and animating element
of these forms. They are not defined by the norms of any one of these fields but instead by the
generative tension created when these norms run up against each other and shape new articu-
lations and practices. These forms of quantitative journalism are continually shifting between
normative and epistemological poles drawn from these fields, and the tensions created by these
shifts have led to both the explosion of new forms and the overlap between them (Borges-Rey,
2017). These varying forms are ultimately different combinations of very similar epistemological,
social, and cultural elements.
More broadly, as Anderson (forthcoming) and Borges-Rey (2016) argue, all of these forms are
means by which journalists seek to present their knowledge claims as more certain and to solidify
their own professional and social legitimacy. The notions of ‘objective’ data, computational neu-
trality, social scientific rigor, and rituals of transparency are all at least in part strategies to garner
public trust and authority for journalism in an era when both are flagging throughout much of
the world and especially in the United States. All of these forms adhere to some type of objectiv-
ity and transparency but reinterpret them to incorporate norms from other converging fields and
to make them more palatable for a more skeptical public (Splendore, 2016).
Conclusion
I do not conclude this chapter by offering yet another typology of quantitative journalism, or
adding another dimension to an existing typology, or defining or redefining a form. As we have
seen, there are plenty of typologies and definitions already; we have many of the conceptual,
historical, and intellectual tools we need to think clearly and precisely about these journalistic
forms. We would do well to use those tools to continue to prune excess classifications and allow
a clearer and more focused conceptualization of quantitative journalism to bear fruit. But we
must also be responsive to developments in the fields we are studying and leave conceptual space
for forms to evolve and new practices to emerge. To that end, I offer a few final thoughts on the
conceptualization of quantitative forms of journalism.
I have spent little time addressing CAR directly in this chapter, and scholars have been divided
on whether CAR is still present in contemporary journalism in any meaningful way. I have
233
Mark Coddington
argued in the past that while CAR continues to be practiced, “it appears to be invoked more
often as a historical mode of quantitative journalism than a contemporary practice” (Coddington,
2015: 334). CAR does indeed still exist: to the extent that journalists are still working with data in
a way that is subordinated to the traditional professional, investigative journalistic paradigm and
incorporates little interactivity, substantial transparency, or data as story, they are practicing CAR.
This characterization generally tracks with statements about CAR made by Usher (2016) and
Anderson (forthcoming). This is becoming a smaller part of journalism, though, and one whose
significance is more in establishing the historical context for more widely practiced forms like
data journalism than in describing widespread everyday journalistic practice. Because of that,
dropping CAR from typologies of quantitative journalism (e.g., Splendore, 2016) is not necessar-
ily a problem. But it remains important to continue to understand what CAR has stood for and
how it relates to the other modes of quantitative journalism, as Anderson’s (forthcoming) valuable
history does. As we discuss the continual tension between continuity and change within data and
computational journalism, CAR represents the continuity pole in that dialectic and thus remains
a crucial area for continued study and conceptualization.
With CAR playing a more historical-contextual role and computational journalism more
clearly splitting off to characterize largely algorithmically based work, it appears that data jour-
nalism is emerging as the central term for this area of journalistic activity, and appropriately so.
It matches the dominant term used for this type of practice in many parts of the world, and in
broad strokes, at least, its themes seem to accurately characterize the most prominent mode of
quantitative journalism at this point – though more so at large or elite news organizations than
smaller or local ones (Fink and Anderson, 2015). Other terms like computational journalism,
interactive journalism, or hacker journalists have great value in clarifying different specializa-
tions within this area, and those should continue to find use and generate fruitful research. But if
researchers begin to settle on a carefully explicated conceptualization of data journalism as their
most common classification for this area of practice, that would likely be a healthy and welcome
development.
I outlined this chapter with the goal of helping to clear up some of the haziness surrounding
the classification of quantitative forms of journalism. But some haziness is inevitable when we are
examining a constellation of practices and norms that is drawn from a complex and shifting set
of disparate influences. Data journalism is both new and old, professional and marginal, partici-
patory and exclusive, static and shifting. Those tensions make it very difficult to understand and
convey fully, especially when we attempt to account for social, cultural, historical, epistemologi-
cal, and professional factors in addition to the more conspicuous technological ones. But it is a
worthwhile and necessary endeavor nonetheless, and even as we push toward a clearer picture of
quantitative journalism, some of the most beneficial research will be the work that continues to
trouble our neat distinctions with complex and nuanced perspectives arising from a fascinatingly
unruly object of study.
Further reading
With so much research pouring out on this subject, it can be difficult to single out just a few
pieces to commend for further reading. But there are a handful of recent works that have been
especially insightful in clarifying quantitative forms of journalism. Nikki Usher’s Interactive Jour-
nalism (2016) is especially valuable for its vivid and empirically grounded conceptualization of
several emerging areas of journalistic work. Her concept of the professional subspecialty is also
extremely useful in helping to clarify the reasons that data journalists adopt, transform, or reject
professional journalistic values. C. W. Anderson’s Apostles of Certainty ( forthcoming), on the history
234
of data journalism, is a crucial examination of the roots and social context of these phenomena,
and a useful reminder of their shared epistemological lineage. And Eddy Borges-Rey’s recent
work (2016, 2017) is some of the most nuanced conceptualization of data journalism and other
related forms that has been undertaken so far.
References
Anderson, C. W. (2013) “Towards a sociology of computational and algorithmic journalism.” New Media &
Society, 15(7), 1005–1021.
Anderson, C. W. (2015) “Between the unique and the pattern: Historical tensions in our understanding of
quantitative journalism.” Digital Journalism, 3(3), 349–363.
Anderson, C. W. (forthcoming) Apostles of Certainty: Data Journalism and the Politics of Doubt. Oxford:
Oxford University Press.
Ausserhofer, J., Gutounig, R., Oppermann, M., Matiasek, S. and Goldgruber, E. (2017) “The datafication
of data journalism scholarship: Focal points, methods, and research propositions for the investigation of
data-intensive newswork.” Journalism [online before print]. doi:10.1177/1464884917700667
Borges-Rey, E. (2016) “Unravelling data journalism: A study of data journalism practice in British news-
rooms.” Journalism Practice, 10(7), 833–843.
Borges-Rey, E. (2017) “Towards an epistemology of data journalism in the devolved nations of the United
Kingdom: Changes and continuities in materiality, performativity and reflexivity.” Journalism [online
before print]. doi:10.1177/1464884917693864
Bucher, T. (2017) “‘Machines don’t have instincts’: Articulating the computational in journalism.” New
Media & Society, 19(6), 918–933.
De Maeyer, J., Libert, M., Domingo, D., Heinderyckx, F. and Le Cam, F. (2015) “Waiting for data journal-
ism: A qualitative assessment of the anecdotal take-up of data journalism in French-speaking Belgium.”
Diakopoulos, N. (2011, April 22) A Functional Roadmap for Innovation in Computational Journalism. www.nick
diakopoulos.com/2011/04/22/a-functional-roadmap- for-innovation-in-computational-journalism/
Diakopoulos, N. and Koliska, M. (2017) “Algorithmic transparency in the news media.” Digital Journalism,
5(7), 809–828.
Dörr, K. N. (2016) “Mapping the field of algorithmic journalism.” Digital Journalism, 4(6), 700–722.
Felle, T. (2016) “Digital watchdogs? Data reporting and the news media’s traditional ‘fourth estate’ func-
tion.” Journalism, 17(1), 85–96.
Fink, K. and Anderson, C. W. (2015) “Data journalism in the United States: Beyond the ‘usual suspects’.”
Flew, T., Spurgeon, C., Daniel, A. and Swift, A. (2012) “The promise of computational journalism.” Journal-
ism Practice, 6(2), 157–171.
Gray, J., Bounegru, L. and Chambers, L. (2012) The Data Journalism Handbook. Sebastopol, CA: O’Reilly
Media.
Griffin, R. J. and Dunwoody, S. (2016) “Chair support, faculty entrepreneurship, and the teaching of statisti-
cal reasoning to journalism undergraduates in the United States.” Journalism, 17(1), 97–118.
exploration on changing mindsets.” Journalism, 15(6), 713–730.
Hamilton, J. T. and Turner, F. (2009) Accountability Through Algorithm: Developing the Field of Computa-
tional Journalism. Behavioral Sciences Summer Workshop, Stanford, CA. Retrieved from www.stanford.
edu/~fturner/HamiltonTurnerAccbyAlgFinal.pdf
Hammond, P. (2017) “From computer-assisted to data-driven: Journalism and big data.” Journalism, 18(4),
408–424.
Hewett, J. (2016) “Learning to teach data journalism: Innovation, influence and constraints.” Journalism,
17(1), 119–137.
Holovaty, A. (2006, September 6) A Fundamental Way Newspaper Sites Need to Chang. Retrieved from www.
holovaty.com/writing/fundamental-change/
Holovaty, A. (2009, May 21) The Definitive, Two-Part Answer to ‘Is Data Journalism?’. Retrieved from www.
holovaty.com/writing/data-is-journalism/
235
Mark Coddington
Howard, A. B. (2014) The Art and Science of Data-driven Journalism. New York, NY: Tow Center for Digital
Journalism. Retrieved from http://towcenter.org/wp-content/uploads/2014/05/Tow-Center-Data-
Driven-Journalism.pdf
Knight, M. (2015) “Data journalism in the UK: A preliminary analysis of form and content.” Journal of
Media Practice, 16(1), 55–72.
Lesage, F. and Hackett, R. A. (2014) “Between objectivity and openness – The mediality of data for journal-
ism.” Media and Communication, 2(2), 42–54.
Lewis, S. C. (2015) “Journalism in an era of big data: Cases, concepts, and critiques.” Digital Journalism,
3(3), 321–330.
Mair, J. and Keeble, R. L. (2013) Data Journalism: Mapping the Future. Suffolk: Abramis.
Michalski, D. (2016) Reader Engagement with Data Journalism: Comparing the Guardian and Washington Post’s
Coverage of People Killed by Police. Unpublished master’s thesis, Hank Greenspun School of Journalism &
Media Studies, University of Nevada, Las Vegas.
Parasie, S. (2015) “Data-driven revelation? Epistemological tensions in investigative journalism in the age of
‘big data’.” Digital Journalism, 3(3), 364–380.
Parasie, S. and Dagiral, E. (2013) “Data-driven journalism and the public good: ‘Computer-assisted report-
ers’ and ‘programmer-journalists’ in Chicago.” New Media & Society, 15(6), 853–871.
Rogers, S. (2013) Facts Are Sacred: The Power of Data. London: Faber & Faber.
Royal, C. and Blasingame, D. (2015) “Data journalism: An explication.” #ISOJ, 5(1). Retrieved from
https://isojjournal.wordpress.com/2015/04/15/data-journalism-an-explication/
Splendore, S. (2016) “Quantitatively oriented forms of journalism and their epistemology.” Sociology Com-
pass, 10(5), 343–352.
Stiles, M. (2017, February 24) “Charting NICAR attendance, over the years.” The Daily Viz [blog post].
Retrieved from http://thedailyviz.com/2017/02/24/charting-nicar-attendance-over-the-years/
Tandoc, E. C. and Oh, S.-K. (2017) “Small departures, big continuities? Norms, values, and routines in the
Guardian’s big data journalism.” Journalism Studies, 18(8), 997–1015.
Usher, N. (2016) Interactive Journalism: Hackers, Data, and Code. Urbana, IL: University of Illinois Press.
Uskali, T. I. and Kuutti, H. (2015) “Models and streams of data journalism.” The Journal of Media Innovations,
2(1), 77–88.
Wing, J. M. (2008) “Computational thinking and thinking about computing.” Philosophical Transactions:
Series A, Mathematical, Physical, and Engineering Sciences, 366(1881), 3717–3725.
236
18
ALGORITHMS ARE A REPORTER’S
NEW BEST FRIEND
News automation and the case for augmented
journalism
Carl-Gustav Linden
Automated routine news is gradually complementing manually produced journalistic texts, cre-
ating emotional stress in many newsrooms. This chapter provides an alternative consideration to
the conventional view that automation amounts to an existential threat to the work of journal-
ists. Focusing on opportunities, it discusses how journalists and editors could embrace new media
technology built on algorithms in ways that create value and meaning. Theoretically the study
is inspired by a social constructionist perspective, arguing that technical, economic, and social
factors determine the pace and extent of automation. This chapter turns the usual discussion
about job losses on its head with a claim that “algorithms are a journalist’s new best friend”.
The real risk is not too much automation but that journalists and the news media falls behind
other industries in developing new cognitive technologies because of, for instance, lack of skills,
strategy, and scarce financial resources.
In recent years, we have seen new job titles surfacing in newsrooms; suddenly there are news
robot trainers, API (Application Programming Interface) editors or algorithm editors working
side by side with more traditional journalists (see also Montal and Reich, this volume, Chapter 4).
There has been a revolution in the production of news by smart machines thanks to the advance-
ment of Natural Language Generation (NLG) technologies, thus extending human capabilities
and creating augmented journalism (Leppanen et al., 2017; Marconi and Siegma, 2017; Scherer,
2010). Journalists are facing the automation of news with mixed feelings, also as a threat to the
profession (Linden, 2017; Carlson, 2015; Van Dalen, 2012). “Should We Be Afraid or Excited
About Robot Journalism?” asks the Huffington Post (Taibi, 2014), while the advice from Edi-
tor & Publisher is “Learn to Stop Worrying and Love Robot Journalists” (Tornoe, 2014). The
economist David Autor, an expert on the history of automation, criticizes journalists and ana-
lysts who “overstate the extent of machine substitution for human labor and ignore the strong
complementarities that increase productivity, raise earnings, and augment demand for skilled
labor” (Autor, 2014: 130). The computer scientist Ben Shneiderman also believes the discussion
around automation has gone off track: “Robots and AI [artificial intelligence] make compelling
stories for journalists, but they are a false vision of the major economic changes”, he notes in a
survey on AI, robotics, and the future of jobs (Smith and Anderson, 2014: 6). In reality, there has
been more talk about news automation than there are actual practical examples (Dörr, 2015) and
237
Carl-Gustav Linden
one is right to suspect that the debate serves as PR for service providers looking for better-paying
customers in industries outside the financially restrained media sector.
Journalists have for decades, mostly unreflectively, been assisted by less controversial algo-
rithms in the newsroom for editing, publishing, and distributing content, especially in broadcast
production (Kolkey, 1999; Vigneaux, 1996). The term algorithm refers to a self-contained step-
by-step set of operations to be performed, such as calculation, data processing, and automated
reasoning – a set of rules that defines a sequence of precise instructions that can be understood by
a computer. Adoption of new technology is a key feature of journalism and media work (Pavlik,
2000). However, organizational structure and occupational practice shape the ways in which new
technology is adopted and with what effects (Boczkowski, 2005; Dickinson et al., 2013; Young
and Hermida, 2014). The discussion among journalists about “robot journalism” is here framed
as a recent example of “automation anxiety” (Akst, 2013), which has been a recurrent topic in
debates on work and technological development for centuries. Reading older research from the
1980s on computer anxiety among journalists (Shipman, 1986), one is tempted to ask: what is
new? The fear of technology seems to be a constant factor in the newsroom (Boczkowski, 2005).
Therefore, one can also inquire, if algorithms have been aiding journalists ever since the 1970s,
would it be appropriate to take a more nuanced view? The risk of “false alarmism” when discuss-
ing technological disruption and labor is imminent (Atkinson and Wu, 2017).
We could start by looking at the jobs issue. Surveys in the United States (Weaver, 2015) show
that the number of full-time journalists working for mainstream general-interest news media
have declined substantially from about 122,000 in 1992 to about 83,000 in 2013. In contrast,
according to the Bureau of Labor Statistics, in 2014, 54,400 people were working as “Reporters,
Correspondents, and Broadcast News Analysts” (Bureau of Labor Statistics, 2016), with a forecast
that 9%, or 4,400 jobs, will disappear by 2024. However, that comparison does not reflect the
changing nature of the occupation, since the full-time qualifier does not quantify the transition
from salary to contract work or the increasing share of freelancers and other contributors.
In the United Kingdom, the situation is different: 2016 was a record year, with 84,000 peo-
ple employed as “journalists, newspaper and periodical editors”, according to official estimates
(Office for National Statistics, 2016). This is a dramatic increase from the previous record year,
2013, with 70,000 people. Before the financial crisis struck in 2007 there were 57,000 people
officially working as journalists. The proportion of journalists working online is on the rise, with
the share of newspaper journalists declining. An estimated 30,000 journalists work wholly or
partly online; however, many bloggers are excluded from this count, “along with others whose
journalistic identity is complex” (Hargreaves, 2016: 4).
The history of automated storytelling is more than 40 years old (Meehan, 1977) and since
the 1960s automatic text summaries have been applied, for example, to weather forecasts (Glahn,
1970), and in the late 1990s to sports, medical, and financial reports. Despite these opportuni-
ties, technologists predict that creative occupations such as journalism are the least susceptible
occupations to workplace automation (Bakhshi et al., 2015). It seems that where journalism jobs
have been disappearing, the prime reasons were changing consumer behavior and media business
models, not automation.
Previous research
In recent years software-generated news has gradually become part, albeit a small part, of the news-
making ecosystem, thus blurring the boundaries between journalism, computer science and statistics.
Popularly labeled “robot journalism” (Clerwall, 2014; Latar, 2014; Levy, 2012; Matsumoto et al.,
2007; Rutkin, 2014) or “automated journalism” (Carlson, 2015; Lecompte, 2015; Napoli, 2014),
“algorithmic journalism” (Anderson, 2013; Diakopoulos, 2014; Dörr, 2016), “machine-written
238
Algorithms are a reporter’s new best friend
news” (Van Dalen, 2012; van der Kaa and Krahmer, 2014) or “computational journalism” (Anderson,
2011; Coddington, 2015, and this volume, Chapter 17; Flew et al., 2012; Gynnild, 2014; Karlsen and
Stavelin, 2014; Stavelin, 2014), these are algorithm-driven automated processes using structured sets
of data from sports, real estate, and stock markets as input to create news items as output. Platforms of
advanced natural generation language transform data into text indistinguishable from what a human
person would write (Clerwall, 2014; Dörr, 2016; Hammond, 2015; van der Kaa and Krahmer, 2014).
“Robot journalism” is a somewhat banal conceptualization, where illustrators often portray
robots writing on computer keyboards instead of visually less attractive software. In this study, the
unifying concept is computational journalism, which, according to Young and Hermida (2014:
381), refers to “forms of algorithmic, social scientific and mathematical processes and systems for
the production of news” or, with a more normative approach, “the combination of algorithms,
data, and knowledge from the social sciences to supplement the accountability function of jour-
nalism” (Hamilton and Turner, 2009: 2). We can simply conclude, like Coddington (2015), that
computational journalism is concerned with the application of the processes of abstraction and
automation to information. Here we are talking about intelligence augmentation (IA) rather than
artificial intelligence (AI), following the path of the computer science pioneer Douglas Engelbart,
who believed that computers should be used to “augment” or extend human capabilities rather
than to mimic or replace them (Markoff, 2016: xii).
Journalists as a relevant group

This study departs from a social constructionist perspective, arguing that social action is a central
element in technological development, that innovation in technology is a “multicentred, complex
process”, and that there is a “spectrum of possible technological choices, alternatives and branching
points” (Winner, 1993: 366). The theoretical focus is narrowed to social construction of technology
(Bijker et al., 1987), and the fundamental argument is that human action shapes technology. The
development and use of technology is embedded in its social context, in this case the newsroom.
Journalists are the main users of new technology in newsrooms, but following Winner (1993)
begs questioning if and how the experiential knowledge of journalists is useful when it comes
to the origin, character, and consequences of technical practices in a domain such as automated
news production. To what extent are journalists considered a relevant social group that should be
involved in development of automated news systems, or are they regarded as a brake on advance-
ment? These questions are motivated by deep divisions in the debate on journalism and technologi-
cal determinism. Studies of the introduction of new technology in newsrooms, from the computer
onwards, show many examples of resistance, mitigation, and adaptation among journalists. Norms
and conventions of journalism play an important role (Dickinson et al., 2013). Following Deuze
(2005), it is understood that the ideology of journalism is a collection of “values, strategies and
formal codes” that characterizes journalism and is shared by most members of the occupation as an
active practice that is continually negotiated (Reese, 1990). Journalism ideology can be categorized
in five different traits: journalists provide a public service and see themselves as watchdogs on behalf
of a public; they are impartial, neutral, objective, and fair; they must be autonomous and free from
bindings; they have a sense of immediacy, actuality, and speed; and journalists have a strong moral
and ethical foundation that gives them a legitimate and valid reason to act on behalf of the public
literature (Deuze, 2005; Golding and Elliott, 1999; Kovach and Rosenstiel, 2007).
From Goffman to artificial intelligence

Artificial intelligence is enjoying something of a renaissance after several false starts (Markoff,
2016). Increased computational resources, the explosive growth of digital data, a focus on specific
239
Carl-Gustav Linden
problems, knowledge engineering with machine learning, and alternative reasoning models have
together contributed to the emergence of truly “intelligent” machines (Hammond, 2015). Many
systems of artificial intelligence make use of computerized processes of human reasoning that
sound familiar to any student of journalism, psychology, or sociology: what is happening around
us (observing data), what does it mean (learning from it) and what is going to happen next?
(Hammond, 2015; Hayes, 2013). Cognitive system engineering theories concentrate on human-
like reasoning. In modeling human behavior, such as human practical reasoning, theories resonate
well with, for instance, framing theory, which, as Carolyn Merchant (1989: xvi) notes, helps us
to understand questions such as: ‘How do people “conceptualise” the world?’ ‘How do they
“behave” in relation to the world?’ ‘How do they “give meaning” to the world?’ With Erving
Goffman (1974), we navigate through these issues by asking two crucial questions that open the
doors to framing theory as a method for understanding social reality. The first – “Under what
circumstances do we think things are real?” – works to identify the conditions under which feel-
ings of reality are established. The second – “What is it that is going on here?” – identifies the
proper response to the current situation. In the same way computers are tuned to ask questions
when trying to make sense of the environment and perform tasks, this is one-way journalists
with limited understanding of computational thinking (Wing, 2011) can follow how reasoning
and problem-solving is turned into software. Astrid Gynnild (2014: 724) actually asks “to what
extent is training in logical, algorithmic, scientific thinking and in journalistic sorting and selec-
tion part of the same package?” Human-based and computer-driven approaches to sense-making
and knowledge logic should be seen as complementary or supplementary (Flew et al., 2012). In
other words, algorithmic and editorial logic should not compete in the newsroom bearing in
mind that tightly categorized and structured information, the prerequisite for computerization,
can at the same time have a damaging impact on human creativity (Kim and Zhong, 2017).
Human-machine advancement
The human-machine development in journalism began long before the introduction of the
computer into newsrooms in the 1970s, and no empirical evidence suggests that there will be a
break with the past, rather we can anticipate a continuous but rapidly accelerating technological
development. Journalists are already surrounded by algorithms taking care of everything from
web searches to photo and text editing, mostly using them unconsciously and seamlessly. Tak-
ing away these pieces of software would reveal to what extent journalists and editors already are
dependent on automated decision systems. Algorithms are the result of human thinking and
reasoning, but they are “neither entirely material, nor are they entirely human – they are hybrid,
composed of both human intentionality and material obduracy” (Anderson, 2013: 1016). Nicho-
las Diakopoulos (2015; see also Koliska and Diakopoulos, this volume, Chapter 19) notes that
human influences are embedded into algorithms, such as criteria choices, training data, semantics,
and interpretation.
The principles of human-agent teaming in automation states that humans are responsible for
outcomes in human-agent teams and therefore must be in command of the human-agent team
(Billings, 1997; Tweedale et al., 2007; Urlings, 2004). To be in command, the human must be
actively involved in the team process and, to remain involved, the human must be adequately
informed. The human must be able to monitor agent behavior, and the activities of the agents
must therefore be predictable. The agents must also be able to monitor performance of the
human. To achieve all this, each team member (humans and agents) must have knowledge of the
intent of the other members. In the computer science literature, agents are defined as computers
with software: an agent can be seen as a software and/or hardware component of system capable
of acting exactingly in order to accomplish tasks on behalf of its user (Nwana, 1996).
240
The co-creator of the automated news provider Narrative Science, Kris Hammond, also a
professor of computer science, underlines the importance of people in the workplace understand-
ing how these new technologies work: “Black boxes that give us answers without explanations,
or systems that fail to communicate with us, cannot become our trusted partners” (Hammond,
2015: 7). Therefore, people need to understand the basics of how these decision systems reason
and come up with answers; otherwise, people cannot “anticipate the actions and limits of the
automation” (Carr, 2015: 164).
Data and method

The empirical material consists of 35 qualitative research interviews with data journalists, news
managers, computer scientists, academics, and industry experts in eight countries (US, UK, Swe-
den, Denmark, Norway, Germany, Spain, and France) that inhabit and reproduce this new eco-
system of computation, journalism, and statistics. The concept of saturation (Glaser and Strauss,
1967) has been the guiding principle: when the gathering of new data passes a point of dimin-
ishing return, it should be stopped – i.e., because it is too time consuming and not resulting in
new information. Thus, this is not a representative randomized survey, but a study where the
interviewees can be seen as key informants with special knowledge in the field, therefore also
“participants” in the research. Rather than passively responding to questions, they were involved
in a dialogue and have actively shaped the course of the interview (King, 1994). The duration
of interviews was between 15 minutes and three hours, organized around a set of predetermined
open-ended questions, with other questions emerging from the dialogue between interviewer
and interviewee (DiCicco-Bloom and Crabtree, 2006). The choice of method was motivated
by a need to focus on the meaning and experiential perceptions of automated news as well as
historical accounts on an individual participant level. It is emphasized that qualitative research
is concerned with meaning and not making generalized hypothesis statements (Mason, 2010).
This study is exploratory work on an emerging field of automation and occupational effects
where there is no broad agreement on the future position of journalism. Interviews were audio-
recorded, transcribed (by transcription service Rev.com), and coded according to different aspects
on the future of computational journalism. The transcripts were reread and compared over again
until an interpretation of their meaning was obtained. As Steinar Kvale (2007: 109) notes, in
principle such a hermeneutic text interpretation is an infinite process, whereas in practice it ends
when a sensible coherent meaning has been reached.
Respondents were chosen based on their documented expertise in computational journalism
(for instance, speakers at conferences, widely quoted in articles, blogs, and research reports) and
were thought to be well placed to analyze the present and future of the industry. Snowball sam-
pling (Goodman, 1961) was also used to find persons outside this group. These people represented
a broad range of 25 different organizations such as new academic initiatives in digital journalism
(Northwestern University, UNC School of Media and Journalism, University of Maryland, and
University of Nebraska), local newspapers (Local Labs, Mittmedia), natural language genera-
tors of massive amounts of news texts (Automated Insights, Narrativa, AX-Semantics, Syllabs,
Yseop, Narrative Science, Associated Press, United Robots), those analyzing official documents
as a service (Nyhetsbyrån Sirén, Textual Relations), providing financial news (Thomson Reuters,
Nasdaq, and Ritzau), and data journalists (at the Guardian, Boston Globe, The Lens, ProPublica,
The Center for Investigative Reporting, and Journalism++ Stockholm).
The study addresses two research questions:
RQ1: What is the state of news automation in the United States and Europe?
RQ2: What are experts saying about constraints and opportunities for news automation?
241
Carl-Gustav Linden
Findings
Automated news was first introduced to newsrooms around 2009–2010 in the United States
by software companies Narrative Science and Automated Insights, which have now established
something of a duopoly. Their software already produces hundreds of thousands of texts for news
organizations, including respected media companies such as Associated Press and Forbes.
Their automated news is dependent on tightly structured datasets for its operation. A dataset
is a big table of numbers in which every row describes a single observation and every column
describes a single attribute for each observation represented by the rows (Conway and White,
2012). The availability of data determines what sorts of texts can be generated. Automated
Insights, for instance, relies on company earnings reports and analyst estimates to produce finan-
cial news for Associated Press, while Narrative Science uses data collected by team leaders with
the mobile phone application Game Changer for sports news.
In Europe, automated news have also entered the commercial stage: companies such as AX-
Semantics, Text-o-matic, Retresco, and Aexea in Germany, Narrativa in Spain, Syllabs in France,
and Arria in the UK are offering limited services to media customers, mainly news on the stock
market, weather forecasts, and football. In France, the interest in robot journalism has grown since
Syllabs produced 34,000 articles for Le Monde during the departmental election night in March
2015. After this successful experiment, newsrooms in France have woken up to the possibility
of producing articles on a massive scale and, according to a data journalist at Syllabs, “it is not
difficult to get newsrooms to see the possibility with personalised content”.
In the Nordic region, several Swedish newspaper companies and one in Finland are publish-
ing sports news from soccer games created by the service provider United Robots and based on
match data collected by Everysport.com. A development manager at of one of these companies,
Mittmedia, which publishes 28 newspapers, noted that scalability of products is a challenge: “It is
difficult to create generic solutions, we have to start from scratch for each new case, and relatively
little is reusable”.
United Robots is looking for opportunities to automate publishing in cases where speed
is important and content can be personalized. The company has started to extract news from
large sets of data such as real estate transactions and earnings reports. Most studies of innovation
in newsrooms, including automation, show that there is a “clear drive towards more efficient
production processes” and the introduction of new technology is motivated by increasing pro-
ductivity and saving costs (Dickinson et al., 2013). The fears of journalists are confirmed in the
interviews: automated news is introduced to make newsrooms more efficient, for instance, by
increasing the volumes. Automated news has been met with skepticism among the employees at
Mittmedia (Svenlin, 2015), but managers say the purpose of introducing the new technology is
not to get rid of employees and save costs but to create more local content and opportunities to
sell local niche advertisement.
The news agency Siren is, with software developed by the consultancy Textual Relations,
servicing media clients with automatically generated updates from tax authorities, administrative
courts and the National Board for Consumer Disputes that can be used to write news reports.
One of the developers at Textual Relations explains that for automated news either the input
or the output has to be well-structured, “otherwise it becomes suddenly really complicated”.
Another service provider, Journalism ++ Stockholm, creates automated texts for Sirens custom-
ers based on, for instance, monthly crime statistics that are relevant to the specific location of the
news agency’s customers.
In Denmark, the financial news agency Ritzau Finans has created templates for 70 listed com-
panies to produce auto-flashes for earnings reports. The management believes that it makes sense
investing in automation, not just because of speed but also for quality: the robot makes fewer
242
mistakes. They have also built into the system algorithms detecting potential errors, which means
that earnings reports that divert too far from expectations are not auto-published. “It has taken
some time to develop. We have to do a separate template for each company, but the job gets easier
the more we do them”, one Danish news manager claimed. After this first step, Ritzau has no
further plans for automation, and progress so far seems to be quite modest.
There are also other newcomers in the Nordic region. In Norway, the news agency NTB
has been producing soccer news automatically since 2016 with the help of software from NTB/
Bakken & Bæck. In Finland, three separate election bots were developed for the local elections in
April 2017, covering over 34,600 candidates and results from more than 300 different munici-
palities. One of these bots created news in three languages simultaneously, Finnish, Swedish, and
English.
Social friction as barrier to automation

Social action is a central element in technological development. There are elements of friction
that create barriers to increased automation in every form possible, ranging from culture and
social norms to technical barriers. In news making, such friction might be an audience that is
averse to automated news, another the ideology of journalism which has not lost its strength –
much to the despair of a computer scientist interviewed for this study: “They’re [journalists]
very skeptical to the point of being [. . .] in a technology community they would be considered
negative”.
The risks are high that one core value of journalism, to question critically, challenge, and dis-
pute, becomes a liability in the domain of automated news. Journalists are to an increasing extent
dependent on colleagues with specialist skills in programming, while they also have a reputation
for being bad at co-operating with colleagues with other skill sets. At the same time, journalists
have been surprisingly resilient in refusing to trade in ideology for new knowledge, learning, and
work opportunities (Boczkowski, 2004; Singer, 2004). However, recent examples of collaborative
data journalism projects shows that sharing knowledge, tricks, and work practices is one of the
core values of computational journalism. Basic technology literacy has become as important a
skill as basic editorial skills (Anderson et al., 2012).
What to automate?
Whenever structured data are available, speed is essential, costs for automation investment go
down, and the added value of the human touch in each particular piece of news is low, automa-
tion might take place. One perfect example would be financial news based on earnings reports,
where speed is crucial, since the users of this information are basically trading software: the
numeric data are stored inside SAP or other financial systems of the companies and are then
summarized and put into a press release. Trading machines take the text and turn it back into
numbers to make sense and turn the information into signals, buy, sell or hold. But the inter-
views show that neither data journalists nor news managers are sure if this will be a dominant
feature of journalism. There is agreement around the notion that basic news stories just as well
could be written by computers or, as a news manager said, “Writing a news story about what
happened in a game yesterday? It is probably the most banal commodity for journalism I can
think of.”. This view is supported by a data journalist: “There are definitely a lot of routine,
mundane items that computers can do quite competently, quite faster than humans, which free
people up to work on other stories that computers can’t”. But this data journalist also believed
news is too complex to be replaced by software in every part, and the same news manager
continued: “Show me a news story that they generated algorithmically that isn’t driven by a
243
Carl-Gustav Linden
highly structured, sequential, longitudinal dataset, like a sports story. Show me. And then I’ll
start worrying”.
There also seems to be some consensus around the notion that qualitative and experienced
journalists will not be replaced, but they should focus on “higher value-added activities” (con-
sultant). Asked under what circumstances the automation of journalism is worthwhile, a news
manager in Denmark answers: “When you know that something is going to happen and you
need to be fast. The next step is to anticipate unpredicted events”. Despite all the discussion
about automated news and “robot journalism”, it should be noted that it is still a small domain
within news making, with few signs of fast progress toward a future where automated news has
become mainstream. Even technical experts of automated news are a bit skeptical regarding the
added value. One Swedish developer says:
Personally, I have a hard time seeing value in texts written on a few data points. I think
the real value in the long term will be to work with unstructured data sources such
as alarm feeds and documents and use robots to identify news rather than to write it.
Distribution is one of the core issues. For a news agency like Associated Press or Ritzau, it
makes sense to automate news because of the sheer volume of news and speed of operations.
They also have an established ecosystem of services and customers: “If you want to create huge
volumes of text you also need an organization which can receive it and distribute it,” a Swedish
developer says.
Availability of data
Despite all the buzz about big data and the possibilities of mining huge sets of information, avail-
ability of data is actually one of the barriers to increased automation. One of the French journal-
ists involved in the Le Monde project explained: “The difficulty in the election project was not in
the writing of texts, it was more in collecting and cleaning data”, and continued, “the challenge
is to get useful data, which is a problem in France”.
Data within the news production process is always a matter of choice, whether thinking
about whom to interview or what set of statistics to use for a story. The integrity, quality,
and reliability of the data available is crucial to automated news, just as it is to other forms of
journalism. Therefore, the choice and evaluation of data should be a journalistic process, while
validation, standardization, and normalization are normally handed over to programmers. A
U.S. data journalist explained that getting interesting and useful data is often the result of long
struggles based on Freedom of Information Act (FOIA) requests. He has observed a clash within
government agencies between people “who really believe that the data belongs to the people”
and that data should be published openly, on the one hand, and other people “who feel very
proprietary and territorial about it, and want to keep everything in-house and worry about
releasing it to the world”. Even though government agencies are publishing more data online,
he still has to fight hard to get other data that the same agencies keep. Another U.S. data jour-
nalist confirmed that the information she uses is “almost always” the result of a FOIA request.
Open access is no guarantee, either: in Sweden, with its 250-year history of having the world’s
first FOIA, protocols from meetings in all 290 municipalities are available online. Data journal-
ism agency Journalism++ Stockholm has scraped them all, and the work reveals that “there
are probably more ways of writing protocols than there are municipalities”, according to one
of the project managers. Thus, the availability of data that are both structured and interesting
becomes a constraint.
244
Automation for predictive journalism

Automation of news making can address managerial concerns to develop more efficient editorial
processes and a turn toward an anticipatory journalism, focused more on predictive than reactive
routines to use resources more efficiently. News alarms like those based on automated searches
of police arrest records that Los Angeles Times crime reporters are using will be much more com-
mon. Local Labs, a company in Chicago, is one example of how automation for proactive or
predictive news making is used in a coherent way. Local Labs, according to its web page, provides
“community news production, advertising and advertorial content, print special sections and web
and mobile application management” (www.locallabs.com). The company is using a combina-
tion of a database with sources, a calendar, and scheduled email to automatically collect updates
on activities in local communities. Certain events tend to get repeated with some frequency,
like activities related to Christmas, and a news manager at the company explains: “There are
certain things that require reactive resources. The reactive resources should be not the norm”.
Preexisting information – events that the newsroom has knowledge on – is used to predict events
through a practical application of probability theory. This saves the journalists time and effort and
makes news making more efficient. A U.S. data journalist believed prediction is probably one of
the most useful applications of news automation:
There are a lot of things that are predictable, where you know likely what form that
information’s going to come in, so you could write a program that could interpret that
predictable feed and write a predictable story based on it. It’s harder for things that are
more nuanced, or less predictable.
Limited role of journalists

The data journalists interviewed for this study expressed feelings of both acceptance and criti-
cism of automated news. That is understandable for several reasons. All innovation involves
unintended and undesirable consequences, which are seldom discussed in the innovation research
literature (Sveiby et al., 2009). Journalists have a legitimate reason to think about new technology
critically and discuss how it affects their work. Automated news has been introduced by software
companies and sold as a service or plug-and-play software. The role of journalists in developing
automated news has been limited to “training” the algorithms to choose the proper wording for
natural language generation, and in some cases where journalists have been involved, the experi-
ence has not been a pleasant one.
Journalists, moreover, are not ready to accept machine-written texts based on a dozen differ-
ent templates except for basic reports on sports or the weather. A data journalist explained why
her brief experience with automation was interesting but not something she wanted to repeat.
She called it a “nightmare” because of the work involved in checking computer-generated lan-
guage: “When we write a story with numbers, we painfully pick the right words to say each
number and report about it”. This statement stands in stark contrast to another by a computer
scientist: “a lot of the training in journalism is really more about finding the story and finding
the information and less about the writing per se”.
Journalists are probably right to fear the risk that templates and formulas for automated news
are reductive in terms of news diversity and complexity. Computer intelligence is predictive and
entirely lacking in curiosity, imagination, and worldliness (Carr, 2015: 124). To be objective, fair,
and credible is a key element in the self-perception of journalists, and using appropriate words is
an expression of this: choosing the right vocabulary that does not misrepresent the information in
245
Carl-Gustav Linden
numbers is not a minor issue. There is also concern with getting facts right, as one U.S. data jour-
nalist said: “Fast is good. Accuracy is more important, so those two items are always competing”.
Only the beginning

NLG technology is only gradually being developed for more sophisticated storytelling (Caswell
and Dörr, 2017), and there is reason to believe that present news automation practices are just a
taste of the future. The first wave of news written by software will be followed by much more
advanced systems that not just journalists but also programmers will have difficulties to super-
vise: Machine learning means that “computer programs build their own algorithms through
training, feedback, and iteration, and are therefore much more difficult to understand and con-
trol” (Bogen, 2015). Even when relatively simple NLG techniques are implemented in editorial
environments, the complexity increases very quickly for most useful applications (Caswell and
Dörr, 2017).
In the existing model for automated news, algorithms are precise, perfect, meticulously pro-
grammed and severe in manner. Human engineers are giving instructions to the computer that
will handle certain inputs and generate certain outputs: the logic of the program is the code
written by humans. Machine learning is about building programs that build themselves and are,
for instance, used to predict users’ preferences by companies such as Netflix and Amazon. These
machine-generated programs are “trained” by their designers through giving feedback on the
results. “They are difficult (sometimes impossible) to understand, tricky to debug, and harder to
control. Yet it is precisely for these reasons that they offer the potential for far more ‘intelligent’
behavior than traditional approaches to algorithms and A.I.” (Auerbach, 2015).
These are learning algorithms that make other algorithms figure it out on their own, by mak-
ing inferences from data: “The more data they have, the better they get” (Domingos, 2015: xi).
This alternative model of programming and computation “sidesteps the limitations of the clas-
sic model, embracing uncertainty, variability, self-correction, and overall messiness” (Auerbach,
2015). Translated to journalism it means that computer programs will not need to be instructed
to understand the complex, often subconscious and contradictory processes of journalistic work
to produce news. Instead, using a corpus of news texts, they will do the job. “If we give a learner
[a learning algorithm] a sufficient number of examples of each, however, it will happily figure out
how to do them on its own” (Domingos, 2015: 6).
The risk is that media companies will not be able to face this technological challenge. One
European provider of NLG solutions says he negotiated for three months with a media company,
which ultimately said no to the offer because it was deemed too expensive: “Never again will we
create any more prototypes before we see the money”. A representative for an American service
provider expresses similar frustration. “Media companies are difficult customers to deal with and
they have no money”. Instead, this company is focusing on the financial industry. A Swedish
representative for a service provider says that the news business is good for testing systems, but his
company is planning to service customers outside the media industry instead. One of the most
promising customer segments are companies involved in global e-commerce, as they have a great
need for updating product information in multiple languages on their websites. Therefore, the
risk is obvious that media companies for a number of reasons will fall behind other industries
that use cognitive technologies.
Conclusion
This exploratory study turns the usual discussion about job losses on its head with a claim that
“algorithms are a journalist’s new best friend”. The real risk is not too much automation but that
246
journalists and the news media are unable to follow other industries in developing new cognitive
technologies because of, for instance, lack of skills, strategy, and scarce financial resources. In real-
ity, the latest development – popularly but somewhat misleadingly named “robot journalism” –
is just another step in human-computer advancement and cognitive technologies that simulate
human reasoning and perceptual skills. I argue that besides the general conclusion that creative
jobs such as journalism are at low or no risk to automation, journalism as ideology, understanding
how journalists give meaning to their work (Deuze, 2005), probably also will be a strong mitigat-
ing effect in the future. Journalists are working in a “complex and contradictory set of macro-
sociological influences” and make a good case study to understand “occupational discourses
in times of rapid economic and cultural change and widespread professional delegitimation”
(Anderson, 2014). The discussion on why there still are so many jobs in journalism could be rel-
evant to broader arguments about resilience in creative and artisanal jobs and also would deserve
more interest outside the field of journalism studies or media management research.
Further reading
This chapter has benefitted tremendously from ground-breaking work on sociology of journal-
ism by Mark Deuze, especially the 2005 article “What is Journalism? Professional Identity and
Ideology of Journalists Reconsidered”. Regarding emerging forms of newswork and profes-
sional journalism, C.W. Anderson’s work has been very helpful, including “Towards a Sociology
of Computational and Algorithmic Journalism” (2013). A great source of inspiration are the
writings of Nick Diakopoulos, especially the 2014 report Algorithmic Accountability Reporting: On
the Investigation of Black Boxes for the Tow Center for Digital Journalism, Columbia School of
Journalism.
References
Akst, D. (2013) “Automation anxiety.” The Wilson Quarterly Summer. Retrieved from http://archive.wilson
quarterly.com/sites/default/files/articles/AutomationAnxiety.pdf
Anderson, C. W. (2011) “Notes towards an analysis of computational journalism.” HIIG Discussion Paper
Series. Retrieved from http://papers.ssrn.com/sol3/papers.cfm?abstract id=2009292 (accessed 11 Janu-
ary 2016).
Society, 15(7), 1005–1021.
Anderson, C. W. (2014) “The sociology of the professions and the problem of journalism education.” Radi-
cal Teacher, 99, 62–68.
Anderson, C. W., Bell, E. and Shirky, C. (2012) Post-industrial Journalism: Adapting to the Present. New York,
NY: Tow Center for Digital Journalism, Columbia University.
Atkinson, R. D. and Wu, J. (2017) False Alarmism: Technological Disruption and the U.S. Labor Market, 1850–
2015. Washington: Information technology & Innovation foundation.
Auerbach, D. (2015) “The programs that become the programmers.” Slate, New America and ASU. Retrieved
from www.slate.com/articles/technology/bitwise/2015/09/pedro_domingos_master_algorithm_how_
machine_learning_is_reshap ing_how_we.html (accessed 11 December 2015).
Autor, D. (2014) Polanyi’s Paradox and the Shape of Employment Growth. Cambridge, MA: National Bureau
of Economic Research.
Bakhshi, H., Frey, C. B. and Osbourne, M. (2015) Creativity Vs. Robots: The Creative Economy and the Future
of Employment. London: Nesta.
Bijker, W., Hughes, T. and Pinch, T. (eds.). (1987) The Social Construction of Technological Systems: New Direc-
tions in the Sociology and History of Technology (Vol. 17). Cambridge, MA: The MIT Press.
Billings, C. E. (1997) Aviation Automation: The Search for a Human-Centered Approach. Mahwah, NJ: Law-
rence Erlbaum Associates.
Boczkowski, P. J. (2004) “The processes of adopting multimedia and interactivity in three online news-
rooms.” Journal of Communication, 54(2), 197–213.
247
Carl-Gustav Linden
Boczkowski, P. J. (2005) Digitizing the News: Innovation in Online Newspapers. Cambridge, MA: The MIT
Press.
Bogen, M. (2015) “Algorithms of war.” Slate, New America and ASU. Retrieved from www.slate.com/
articles/technology/future_tense/2015/12/the_dangers_of_enlisting_algorithms_in_statecraft.
html?wpsrc=sh_all_tab_tw_ru (accessed 11 December 2015).
Bureau of Labor Statistics. (2016–2017 Edition) Occupational Outlook Handbook: Reporters, Correspondents, and
Broadcast News Analysts. Washington, DC: U.S. Department of Labor.
Carr, N. (2015) The Glass Cage: Where Automation Is Taking Us. London: The Bodley Head.
Caswell, D. and Dörr, K. (2017) “Automated journalism 2.0: Event-driven narratives: From simple descrip-
tions to real stories.” Journalism Practice, 1–20.
Clerwall, C. (2014) “Enter the robot journalist.” Journalism Practice, 8(5), 519–531.
Conway, D. and White, J. M. (2012) Machine Learning for Hackers. Sebastopol, CA: O Reilly Media.
Deuze, M. (2005) “What is journalism? Professional identity and ideology of journalists reconsidered.”
Journalism, 6(4), 442–464.
Diakopoulos, N. (2014) Algorithmic Accountability Reporting: On the Investigation of Black Boxes. New York,
NY: Tow Center for Digital Journalism, Columbia School of Journalism.
Diakopoulos, N. (2015) “Algorithmic accountability: Journalistic investigation of computational power
structures.” Digital Journalism, 3(3), 398–415.
DiCicco-Bloom, B. and Crabtree, B. F. (2006) “The qualitative research interview.” Medical Education, 40(4),
314–321.
Dickinson, R., Matthews, J. and Saltzis, K. (2013) “Studying journalists in changing times understanding
news work as socially situated practice.” International Communication Gazette, 75(1), 3–18.
Domingos, P. (2015) The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our
World. St Ives: Allen Lane.
ism Practice, 6(2), 157–171.
Glahn, H. R. (1970) “Computer-produced worded forecasts.” Bulletin of the American Meteorological Society,
51(12), 1126–1131.
Glaser, B. and Strauss, A. L. (1967) The Discovery of Grounded Theory: Strategies for Qualitative Research. New
York, NY: Aldine Publishing Company.
Goffman, E. (1974) Frame Analysis: An Essay on the Organization of Experience. New York, NY: The Maple Press.
Golding, P. and Elliott, P. (1999) “Making the news.” In H. Tumber (ed.), News: A Reader. Oxford: Oxford
University Press (pp. 112–120).
Goodman, L. A. (1961) “Snowball sampling.” The Annals of Mathematical Statistics, 32(1), 148–170.
exploration on changing mindsets.” Journalism, 15(6), 713–730.
Hamilton, J. T. and Turner, T. (2009) “Accountability through algorithm: Developing the field of com-
putational journalism.” Center for Advanced Study in the Behavioral Sciences. Retrieved from http://web.
stanford.edu/~fturner/Hamilton%20Turner%20Acc% 20by%20Alg%20Final.pdf
Hammond, K. (2015) Practical Artificial Intelligence for Dummies. Hoboken, NJ: John Wiley & Sons.
Hargreaves, I. (2016) “Foreword.” In N. Thurman, A. Cornia and J. Kunert (eds.), Journalists in the UK.
Oxford: Reuters Institute for the Study of Journalism.
Hayes, A. F. (2013) Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based
Approach. New York, NY: Guilford Press.
Karlsen, J. and Stavelin, E. (2014) “Computational journalism in Norwegian newsrooms.” Journalism Practice,
8(1), 34–48.
Kim, Y. J. and Zhong, C.-B. (2017) “Ideas rise from chaos: Information structure and creativity.” Organiza-
tional Behavior and Human Decision Processes, 138, 15–27.
King, N. (1994) “The qualitative research interview.” In C. Cassel and G. Symon (eds.), Qualitative Methods
in Organizational Research: A Practical Guide. London: Sage Publications (pp. 1–35).
Kolkey, D. (1999) “Features-newsroom automation-new systems offer efficient solutions to newsroom
staffs.” Broadcast Engineering, 41(10), 94–99.
Kovach, B. and Rosenstiel, T. (2007) The Elements of Journalism: What Newspeople Should Know and What the
Public Should Expect. New York, NY: Three Rivers Press.
248
Kvale, S. (2007) Doing Interviews: The Sage Qualitative Research Kit. Los Angeles, CA: Sage Publications.
Latar, N. L. (2014) “Robot journalists: 2015. ‘Quakebot’ is just the beginning.” Wharton, University of Penn-
sylvania. Retrieved from http://knowledge. wharton.upenn.edu/article/will-robot-journalists-replace-
humanl-ones/ (accessed 15 January 2015).
Lecompte, C. (2015) “Automation in the newsroom.” Nieman Reports, 69(3), 32–45.
Leppanen, L., Munezero, M., Granroth-Wilding, M. and Toivonen, H. (2017) “Data-driven news genera-
tion for automated journalism.” International Conference on Natural Language Generation, INLG 2017, 4–7
September 2017, Special Interest Group on Natural Language Generation (SIGGEN).
Levy, S. (2012) “The rise of the robot reporter.” Wired, 20(5), 132–139.
Linden, C.-G. (2017) “Algorithms for journalism: The future of news work.” The Journal of Media Innova-
tions, 4(1), 60–76.
Marconi, F. and Siegma, A. (2017) The Future of Augmented Journalism: A Guide for Newsrooms in the Age of
Smart Machine. New York, NY: Associated Press.
Markoff, J. (2016) Machines of Loving Grace: The Quest for Common Ground Between Humans and Robots. New
York, NY: Harper Collins.
Mason, M. (2010) “Sample size and saturation in PhD studies using qualitative interviews.” Forum: Qualita-
tive Social Research, 11(3), 1–19.
Matsumoto, R., Nakayama, H., Harada, T. and Kuniyoshi, Y. (2007) “Journalist robot: Robot system making
news articles from real world.” IEEE, April 10–14, 2007.
Meehan, J. R. (1977) “Tale-spin, an interactive program that writes stories.” Paper presented at the IJCAI’77
Proceedings of the 5th International Joint Conference on Artificial Intelligence (pp. 91–98).
Merchant, C. (1989) The Death of Nature: Women, Ecology, and the Scientific Revolution. San Francisco, CA:
Harper and Row.
Nwana, H. S. (1996) “Software agents: An overview.” Knowledge Engineering Review, 11(3), 205–244.
Office for National Statistics. (2016) EMP04: Employment by Occupation. Retrieved from www.ons.gov.uk/
employmentandlabourmarket/peopleinwork/employmentandemployeetypes/datasets/employment-
byoccupationemp04 (accessed 6 May 2017).
Pavlik, J. (2000) “The impact of technology on journalism.” Journalism Studies, 1(2), 229–237.
Reese, S. D. (1990) “The news paradigm and the ideology of objectivity: A socialist at the wall street jour-
nal.” Critical Studies in Media Communication, 7(4), 390–409.
Rutkin, A. (2014) “Rise of robot reporters: When software writes the news.” New Scientist, 221(2962), 22.
Scherer, E. (2010) “Le ‘Journalisme Augumenté’ En 10 Points.” OWNI. http://owni.fr/2010/11/07/le-
%C2%AB-journalisme-augmente-%C2%BB-en-10-points/ (accessed 15 May 2017).
Shipman, J. M. (1986) “Computerization and job satisfaction in the newsroom: Four factors to consider.”
Newspaper Research Journal, 8(1), 69–78.
Singer, J. B. (2004) “Strange bedfellows? The diffusion of convergence in four news organizations.” Jour-
nalism Studies, 5(1), 3–18.
Smith, A. and Anderson, J. (2014) AI, Robotics, and the Future of Jobs. Washington, DC: Pew Research Center.
Stavelin, E. (2014) Computational Journalism: When Journalism Meets Programming. PhD thesis, The University
of Bergen.
Sveiby, K.-E., Gripenberg, P., Segercrantz, B., Eriksson, A. and Aminoff, A. (2009) “Unintended and unde-
sirable consequences of innovation.” ISPIM, 21–24 June 2009.
Svenlin, Johan. (2015) “Roboten Kommer!” Journalisti. Accessed December 16, 2015.www.journalisti.fi/
artikkelit/2015/15/roboten-kommer-/
Taibi, C. (2014). “Should we be afraid or excited about robot journalism?” Huffington Post. Retrieved
from www.huffingtonpost.com/2014/07/11/robotjournalism-good-technology-_n_5577651.html
(accessed 23 December 2015)
Tornoe, R. (2014) “Learn to stop worrying and love robot journalists.” Editor & Publisher, 147(9),
24–26.
Tweedale, J., Ichalkaranje, N., Sioutis, C., Jarvis, B., Consoli, A. and Phillips-Wren, G. E. (2007) “Innova-
tions in multi-agent systems.” Journal of Network and Computer Applications, 30(3), 1089–1115.
Urlings, P. (2004) Human – Machine Teaming. PhD thesis, University of South Australia.
Van Dalen, A. (2012) “The algorithms behind the headlines: How machine-written news redefines the core
van der Kaa, H. and Krahmer, E. (2014) “Journalist versus news consumer: The perceived credibility of
machine written news.” Columbia University, 24–25 October. http://compute-cuj.org/cj-2014/cj2014_
session4_paper2.pdf.
249
Carl-Gustav Linden
Vigneaux, S. (1996) “The integration of a newsroom computer system with a server-centered news produc-
tion system.” IET. Retrieved from https://ieeexplore.ieee.org/abstract/document/643225/
Weaver, D. H. (2015) “Studying journalists and journalism across four decades: A sociology of occupations
approach.” Mass Communication and Society, 18(1), 4–16.
Wing, J. (2011) “Research Notebook: Computational Thinking–What and Why?” Retrieved from https://
www.cs.cmu.edu/link/research-notebook-computational-thinking-what-and-why
Winner, L. (1993) “Upon opening the black box and finding it empty: Social constructivism and the phi-
losophy of technology.” Science, Technology, and Human Values, 18(3), 362–378.
Young, M. L. and Hermida, A. (2014) “From Mr. and Mrs. Outlier to central tendencies.” Digital Journal-
ism, 3(3), 381–397.
250
19
DISCLOSE, DECODE, AND
DEMYSTIFY
An empirical guide to algorithmic transparency
Scholars and educational and professional organizations of journalism have embraced transpar-
ency as a new critical norm guiding journalistic practice (Deuze, 2005; Allen, 2008; McBride and
Rosenstiel, 2014; SPJ, 2014; RTDNA, 2015). The rise of transparency to a new journalistic core
value has been fueled by digital technologies and the belief that transparency may counter the
loss in credibility and trust (Kovach and Rosenstiel, 2014; Hayes et al., 2007; Plaisance, 2007) the
news media has experienced in the recent past (Pew Center, 2012).
The primary tenet of transparency is that it enables audiences to learn more of how news
is produced but also about individual journalists and news outlets. Deuze (2005: 455) defines
transparency as the “ways in which people both inside and external to journalism are given a
chance to monitor, check, criticize and even intervene in the journalistic process.” However, the
increasing implementation of computational journalism1 and use of algorithms in news produc-
tion, curation, and dissemination (Diakopoulos, 2016) challenge the new norm of transparency
due to the opacity in algorithmic automated decision making (Diakopoulos, 2015).
Computational or algorithmic journalism is increasingly becoming part of the daily news
production and news consumption of crime and sports stories, earning reports, and more (Graefe,
2016; see also Montal and Reich, this volume, Chapter 4). Writing software employed by the likes
of the Associated Press (AP) and Yahoo can churn out thousands of stories. Currently the focus
is primarily on sports, financial, and political stories (AP, 2015; Yahoo, n.d.), but algorithms have
also been used to report on earthquakes (Oremus, 2014), generate event-driven narratives around
things like car chases (Caswell and Dörr, 2017), or write and post entries to The Los Angeles Times
homicide blog (Young and Hermida, 2015).
The application of computational journalism goes beyond the creation and writing of news
stories. In the news media algorithms are also being used to simulate and predict possible out-
comes based on vast amounts of data. News organizations such as the New York Times, BuzzFeed,
and Mashable use algorithms to select and distribute stories suitable for social media (Wang, 2015;
Nguyen et al., 2015; Albeanu, 2015). In order to cope with the issues of scale, due to ever grow-
ing amounts of data, news production processes make increasing use of algorithmic tools to find,
verify, and filter sources, while also disseminating content via news bots on social media (Park
et al., 2016; Thurman et al., 2016; Lokot and Diakopoulos, 2016).
The opaqueness that comes with algorithmic news media work finds its counterbalance in
algorithmic transparency, which we define as the disclosure of information about algorithms to enable
251
monitoring, checking, criticism, or intervention by interested parties. The values, practices, and standards
for algorithmic transparency are still developing, especially regarding questions of how and
when information concerning algorithms can and should be disclosed. Initial attempts focused
on explaining the methods of data-driven stories (see ProPublica; Grochowski Jones and Orn-
stein, 2016) or occasionally providing open-source code and data on repository sites such as
Github (see BuzzFeed, FiveThirtyEight). Other efforts to inform the public about the various
technologies and algorithms used in news production have taken on the forms of academic
papers, as in the case of the BBC (Shearer et al., 2014) or blogs (see “Open” on the New York
Times website).
While transparency is just one specific effort toward greater accountability of algorithms in
the news media (Dörr and Hollnbuchner, 2017), it’s a notion that has gained greater purchase
within journalism and society at large (Schudson, 2015). Hence this research explores the ques-
tion: How exactly might the news media become transparent about the algorithms it uses in
news production? The underlying goal of this study is to develop a transparency or disclosure
spectrum that allows professional journalists to swiftly and effectively consider options of infor-
mation disclosure concerning algorithms within a specific institutional context. For this purpose
we examined the discussions of nine focus groups that included members from national news
outlets and universities in the US.
Our findings suggest various information disclosure opportunities about algorithms across
four areas: data, model, inference, and interface. Additionally, results show that despite the auto-
mated nature of algorithms, the human factor and human decision making are critical factors in
the consideration of transparency.
Transparency
Transparency and the news media

The field of journalism increasingly acknowledges transparency as an important norm for today’s
journalism (Vos and Craft, 2016). The New York Times public editor Liz Spayd (2016) for instance,
suggested that audiences “want more transparency in stories that are shapeshifting before their eyes”.
This is because transparency is deemed to help to “discover social truth” and “the truth about the
manufacturing of news”, while also “increasing legitimacy with citizens” (Allen, 2008: 324).
In the research literature, journalistic transparency is frequently discussed as an ethical impera-
tive (Plaisance, 2007) and as a way to increase legitimacy (Allen, 2008; Karlsson, 2010) and
credibility in the news media (Hayes et al., 2007). This is because the opening up process that
transparency enables “gives the public a basis on which to judge whether a particular kind of
journalism is the kind they wish to encourage and trust” (Kovach and Rosenstiel, 2014: 291).
Supporters of the concept consider transparency “the new objectivity” (Weinberger, 2009), a
better form of truthtelling (Singer, 2007), and a means to showcase that professional journalism
is superior to nonprofessional content (Karlsson, 2011).
The notion of transparency is also related to corrections, as the show of vulnerability by
admitting failure is deemed to build trust (Silverman, 2013). Yet critics of this type of transpar-
ency such as Broersma (2013) suggest that the journalistic institution may lose legitimacy, as the
admittance of fallibility would undermine the authority of journalism. Some research shows that
transparency about corrections may actually have limited effects and does not necessarily lead to
more credibility or further journalistic authority (Karlsson et al., 2017). Overall, research still has
to find whether transparency will indeed affect news audiences’ credibility perceptions (Karlsson
et al., 2014; Roberts, 2007; Tandoc and Thomas, 2017).
252
Disclose, decode, and demystify
The academic and institutional discourse concerning transparency suggests that transparency
has become critical and a core value for journalism, yet research has also shown that journalists
are still struggling to adopt transparency in their daily work (Chadha and Koliska, 2014; Koliska
and Chadha, 2017) or favor objectivity as a core journalistic norm (Hellmueller et al., 2013).
Nevertheless, journalism is also being influenced by open-source culture, creating a fertile ground
for more transparency such as open data and public involvement in news production (Codding-
ton, 2015; Lewis and Usher, 2013).
Algorithmic transparency
The growing use of algorithms in all aspects of news production and distribution allows
for faster and more effective work processes, yet at the same time this new efficiency based
on difficult to parse automated processes runs counter to the notion of transparency and
automated technologies (Diakopoulos, 2015). A major issue regarding algorithmic systems
is the possibility of subtle biases (Friedman and Nissenbaum, 1996) which can influence
information analysis and the understanding of issues that various publics may rally around
(Gillespie, 2014). Algorithmic transparency in the news is then the idea to publicly show and
reveal the workings of computational systems in order for users to discern the embedded
values of specific algorithms, which would allow a better understanding of journalism with
a specific point of view (McBride and Rosenstiel, 2014). Research shows that some forms
of algorithmic transparency such as information disclosure and explanations of algorithmic
systems can positively affect user perceptions. Cramer et al. (2008) showed that explanations
about a specific recommender algorithm can positively influence the acceptance of particu-
lar recommendations. Kizilcec (2016) indicated that transparency can indeed increase trust
when used in an information seeking context. But additional information provided through
transparency can also diminish the user experience (Schaffer et al., 2015). Other studies have
also explored information disclosure in personalization (El-Arini et al., 2012) and ranking
(Diakopoulos et al., 2014).
Exploring algorithmic transparency in the media

This exploratory study is an attempt to move research and practice of algorithmic transparency
toward a more common ground that is applicable to a wide range of automated applications in
the news media. Fifty participants across nine focus groups discussed the scope of information
disclosure about algorithms including aspects such as feasibility, benefits, and limitations. In order
to inform transparency practice and guide research we asked:
RQ1: What elements of algorithms could be made public?

RQ2: What are the limitations of algorithmic transparency?;
RQ3: What are possible comprehensive disclosure mechanisms for algorithmic transparency?
Methodology
We conducted focus groups to collect a wide variety of perspectives, perceptions, and practices
(Puchta and Potter, 2004) and to stimulate interactivity to evoke group insights (Morgan, 1996).
The issue of algorithmic transparency can be considered a highly specialized subject matter, and
as such we invited participants with a similar professional background to elicit expert evaluations
(Kemper et al., 2003).
253
Participants
Focus group participants comprised of academic and industry specialists on algorithmic applica-
tions from the fields of journalism and computer sciences. Academic participants came from
North American universities such as Columbia, Stanford, NYU, Northeastern, Rutgers, CUNY,
University of Maryland, Harvard, Princeton, and others. Industry practitioners included editors
and reporters from the New York Times, Washington Post, Texas Tribune, NPR, Boston Globe, Associ-
ated Press, etc., but also data scientists and managers with companies such as CNN, Mashable,
Vocativ, SmartNews, Chartbeat, and Bloomberg.
Case studies – stimuli

The basis for focus group discussions were three case studies (CS) specifically developed to mir-
ror current use of algorithms in content creation (CS1), content curation (CS2), and simulation,
prediction and modeling in storytelling (CS3). CS1 addressed the implementation of natural
language generation software to write content (e.g., Automated Insights and Narrative Science).
CS2 dealt with filtering algorithms that curate, rank, and recommend or personalize content akin
to the Facebook newsfeed. The third case study’s (CS3) emphasis was on simulation in news sto-
ries such as political predictions and models for data visualization. All case studies were explained
(including relevant technological aspects) and shared before as well as during the discussions.
Procedure
The fifty participants were randomly assigned to three focus groups of 14–18 participants, as
larger groups are deemed to elicit a wider range of responses (Morgan, 1996). Three different
moderators led and facilitated each group discussion. Each focus group discussed the individual
case studies for about an hour. Groups were mixed randomly again after each session so that
everyone participated in each case study once.
As moderators guided and prompted the discussion about algorithmic transparency, par-
ticipants collectively gathered, categorized, and displayed possible avenues for information
disclosure. In a second step, pros and cons of disclosing information about algorithms such
as manipulation and costs were discussed. All participants were debriefed at the end as one
large group.
Stenographers transcribed each discussion in real time, excluding any identifying information
of the participants. We then analyzed all transcripts employing open iterative coding including
affinity diagramming, typologizing, and memoing (Lofland and Lofland, 1994). A codebook was
created and applied to all transcripts after coding five case study transcripts independently resolv-
ing any inconsistencies.
Study findings
The focus group discussion analysis of the three case studies provided us with an empirically
grounded typology covering a broad spectrum of information disclosure for various algorithms.
A central theme that emerged from our typology is the human factor that is deeply embedded in
algorithmic systems. Several focus group participants acknowledged that the underlying human
involvement in the creation, design, and decision making of algorithms needs to be parsed out
and disclosed. Specifically, a participant suggested that “when we’re talking about algorithmic
accountability, we also want to elucidate the human accountability that is going on inside news-
rooms in order to make the public trust newsrooms more” (CS2).
254
Table 19.1 Summary of transparency factors across four layers of algorithmic systems
Layer Factors
Data • Information quality

Accuracy

Uncertainty (e.g., error margins)


Timeliness

Completeness

Sampling method

• Definitions of variables
• Provenance (e.g., sources, public or private)
• Volume of training data used in machine learning
• Assumptions of data collection
• Inclusion of personally identifiable information
Model • Input variables and features
• Target variable(s) for optimization
• Feature weightings
• Name or type of model
• Software modeling tools used
• Source code or pseudo-code
• Ongoing human influence and updates
• Thresholds or other embedded rules
Inference • Existence and types of inferences made
• Benchmarks for accuracy
• Error analysis (including, e.g., remediation standards)
• Confidence values or other uncertainty information
Interface • Algorithmic presence signal
• On/off
• Tweakability of inputs, weights
We categorized findings according to the notion of a data pipeline, which organizes our typol-
ogy into four phases of an input-output framework. Following the flow of information toward
the end user, we distinguish between: data (inputs), model (transformation), inference (output),
and interface (output). In Table 19.1 we detail and discuss the relevant information types for
disclosure of human-algorithm systems that emerged during the focus group discussions.
Humans and algorithms

The disclosure of human involvement in algorithmic systems is critical to mapping the affor-
dances of algorithmic transparency, especially because human actions are like “a black box” and
much “harder to be objective with” than parameters of an algorithm (CS1). Focus group partici-
pants suggested that human actions could be disclosed through editorial explanations concerning
the selection or optimization of data and use of algorithms but also by identifying the authors,
managers, or designers of algorithmic systems. One critical aspect mentioned was the determi-
nation of credibility of automatically generated news content (CS1), which could be enhanced
by indicating the existence of a human editor through a byline. As algorithmic systems are the
result of collaborative work, such labeling efforts may be problematized (Nissenbaum, 1996) even
though they could contribute to normative and institutionalized standards that are still in flux
(Montal and Reich, 2016, and this volume, Chapter 4).
255
The acquisition of software can further complicate disclosure practices, yet participants sug-
gested making aspects like configuration and parameterization of such off-the-shelf software
transparent. Some aspects considered for disclosure were much more ephemeral and difficult to
parse out, like the social-cultural context or chronological history that influenced the develop-
ment of an algorithm. Other more concrete aspects that could shed light on the human involve-
ment in algorithmic use and production included naming the developers of an algorithm and
explaining their rationale and assumptions but also disclosing who supervises algorithmic perfor-
mances and how, or as a participant pithily stated, “who is behind it all” (CS2).
The multiple levels of human involvement in algorithmic systems reflect the obstacles of
parsing out human influences. Participants thus considered the notion of simply disclosing the
presence of an algorithm in the first place. In particular, discussions around the automatic gen-
eration of news content (CS1) triggered deeper considerations of making human involvement
transparent. One participant outlined the scope of the issue:
I think it is very interesting [. . .] why there is still this split between an algorithmically
generated piece and a human-generated piece. The human-generated piece ends up
having to be accountable. I could fire that person. There is an incentive to do a reason-
able job. A machine is engineered to be as good as it can be. When you scale it up to
hundreds of thousands or millions of pieces and there is some margin of error and at
some time it will screw up – it could potentially be a real problem [CS1].
Another participant summarized, “there is a worry you cannot correct for this systematic error”
(CS1), as the processes of implementing improvement can be much more complicated for algo-
rithmic systems.
Input-output pipeline
As algorithms can be part of various processes in the production, curation, and/or management
of information, we focus on the input-output pipeline as a model to describe critical phases of
automated systems. The pipeline starts with the ‘data’ or information that is fed into a system
then transformed by a ‘model’ and put out as an “inference” or classification, which is presented
to the user through an “interface.” Based on this pipeline, we describe in the following subsec-
tions the various avenues of potential information disclosure around algorithms.
Data
Data is the most basic and crucial element for any type of algorithmic system including such
processes as machine learning, simulation, personalization, and automatically generated content.
Focus group participants repeatedly discussed ways of disclosing information concerning the
quality and validity of data. The core aspects of data quality that emerged were accuracy, uncer-
tainty (e.g., error margins), and completeness or missing elements. With respect to validity, par-
ticipants suggested factors such as timeliness (i.e., when data was collected), provenance (e.g.
sources), sampling method, and inherent assumptions or limitations in the data collection but also
data transformation, vetting, cleaning, or editing. Participants suggested disclosing the data with
annotated descriptions including a reference if human involvement was present in these pro-
cesses. Metadata was identified as a vital element for such information, possibly shedding light on
private or public sources or identifying information. Focus group participants were specifically
interested in learning how data profiles based on user activity influence algorithmic performance:
“These would be aspects of the profile of users or type of personal information used and how
256
the algorithm is tested, how the algorithm ranks stories based on traffic, different inputs and
variables” (CS2). Such information, a participant suggested, could also be used to correct personal
data, ultimately improving algorithmic performance.
The underlying principles of algorithmic systems, even though they operate on a quantified
simplification of the world, are human-made definitions and rules. The human influences on the
selection and processing of data are thus vital elements of any transparency information around
algorithmic systems.
Model
The model of an algorithmic system, which can be understood as a simplified or abstracted ver-
sion of reality, was seen as a critical element for algorithmic transparency. Participants indicated
that understanding the methodology and the modeling process would be vital to discern an algo-
rithm’s process of prediction, classification, ranking, or association. Specifically, discussants were
interested in the type of model (e.g., linear, nonlinear, etc.), what is being optimized by the model,
how different variables are weighted, and what assumptions or limitations may be embedded in
software modeling tools. Similarly, transparency of data used to train a model could offer criti-
cal insights concerning the optimization and machine learning features of algorithmic systems.
Some participants interpreted these aspects as editorial factors since specific parameters in the
model not only determine the semantics but also the interpretation of the data. In automatically
generated news stories, for instance, it was pointed out that understanding thresholds could be
critical for end users to see how a value is interpreted, e.g., as “moderate” versus “extreme.”
Several participants said they want to see the source code because
if I have the data and the models I can put everything in a website, somebody can come
and say your regression model is wrong. So you can have discussion based on hard facts,
meaning you have the code and data. [CS3]
Disclosing the code and its various revised reiterations allows tracing the modeling process from
input to output, yet several participants also pointed out that the code may be too technical for
the average user.
Focus group discussions again acknowledged human involvement and rationales in algorith-
mic modeling and decision-making processes, pointing toward aspects like organizational culture
that influences aspects like weighting, exclusion-inclusion of variables, and also the evolution of
models. For instance, a participant asked: “How are election models learning from the previous
elections process?” (CS2). Several discussants saw parsing the human decision-making processes
from algorithmic optimizations as a challenge for algorithmic transparency. Yet two approaches
emerged: (1) disclosing human decisions or editorial choices influencing the modeling and (2)
disclosing how those choices are applied in the code. Such forms of transparency could allow end
users to better gauge a news organization’s culture and social context.
Inference
The issue of algorithmic output inferences was frequently addressed with questions concerning
uncertainty and accuracy. Discussants deemed that learning more about output inferences of
an algorithm such as classifications, predictions, or recommendations would provide significant
insights to better understand algorithmic systems. In particular, participants wanted to know
how outliers and ‘normal’ cases would be handled in the various inferences. Such information, it
was pointed out, would allow a better understanding of embedded assumptions in an algorithm:
257
“People get all fancy about their underlying statistical model and the mathematical assumptions
in some algorithm, and they forget to account for those assumptions when drawing conclusions”
(CS3). In the case of Facebook, participants suggested that algorithms often assume relations and
preferences of users, and thus they suggested that disclosure concerning inference should both
include accuracy but also what inferences are being made to begin with.
Study participants proposed to include not simply a summary of statistics but also specific
details such as confidence values or “standard deviation [to show] how precise these results are”
(CS3). Disclosing aspects of uncertainty and error including sources (human, data, algorithm)
would help users to understand the scope of inferences including possible underlying assump-
tions. Such forms of transparency, participants suggested, should be accompanied by contex-
tualizing descriptions of the various aspects of uncertainty and how those relate to the overall
inference process.
Interface
The last stage of the input-output pipeline – the interface – directly interacts with end users to
provide transparency information in a comprehensive form. Participants offered a wide range
of ways to present transparency information about data, model, and inferences, from FAQs to
periodic transparency reports, possibly by ombudspeople. At its most basic form it was suggested
to simply signal the user via an icon on the interface level that the content has been, for instance,
personalized by an algorithm. Or as a participant put it, “I wonder if any time you receive infor-
mation that had been curated by an algorithm, that there was some visual cue like ‘algorithm at
work’” (CS2).
Interactivity was seen by participants as an ideal way to explore transparency information but
also seen as a possible limitation depending on “whether or not the designer has exposed certain
things for you to play with” (CS2). Discussants suggested allowing users to access data for sports
stories such as box scores or providing an editorial explanation. Participants also asked “how can
the user control the algorithm” (CS2) and proposed that interactivity could help to learn more
about an automated system:
You could click on the forecast and see how uncertain it was and how it changes but
you could even expand this model by giving people the possibility to, for example, play
different economic indicators and see how they would affect the forecast [CS3].
In the same vein, some focus group members suggested an on/off switch for algorithms to
“know when I’m in a personalized space or not” (CS2). Such forms of algorithmic manipulation
including the alteration of parameters would allow for a direct comparison of the algorithm’s
impact on content.
An unintended side effect of transparency or disclosure of information on the interface layer
is the possibility of information overload. Some focus group participants voiced concerns that
transparency efforts may hamper the attempt at making algorithmic systems more discernable
for the average user. In that respect several discussants suggested that difficult to access techni-
cal information should be translated “for people to actually get it intuitively and easily” (CS2).
Nevertheless, other participants warned that such “translations” could lead to imprecisions and
wrong conclusions regarding the algorithmic systems and their underlying assumptions.
A number of focus group discussants saw algorithmic transparency in itself as impractical
and asked if users would indeed want such information. One participant even suggested that
describing an “extremely detailed model [could] actually turn off the public” (CS3). And some
wondered whether algorithmic transparency is a self-serving effort by media outlets: “The
258
organization is happy, but do nonspecialists care?” (CS1), while others argued that the audience
has changed and in fact “wants transparency now” (CS1). The concerns regarding algorithmic
transparency are indeed problematic, as primarily experts but not necessarily the general public
are able to comprehend intricate technical information. Thus balancing the various information
needs of users while at the same time avoiding information overload or diminishing the user
experience remain a practical and theoretical challenges.
Consequences and challenges to algorithmic transparency

As a final step of the focus group discussions about avenues for algorithmic transparency,
participants were asked to gauge the various disclosure options according to aspects such
as feasibility, manipulation, or general organizational costs. In fact, business considerations
featured as the most prevalent moderators for algorithmic transparency. Participants pointed
toward a higher workload to produce transparency information, which would result in higher
costs when preparing source code, editorial explanations, and data. On the other hand, it
was suggested that algorithmic transparency “might be technically very easy to do, but then
in terms of the news organization’s self-interest, particularly from management and from a
financial perspective, it might be costly” (CS2). At the same time, discussions around feasibil-
ity and costs showed that financial or organizational benefits for greater transparency were
difficult to assess, and, as such, participants struggled to propose concrete incentives to be
more transparent.
Focus group discussants also saw pitfalls of transparency with respect to legal and privacy
concerns when, for instance, data was not sufficiently anonymized or errors were made. “If we
published something inaccurate [. . .] to say so publicly might be admitting in some jurisdictions
an amount of guilt, like for something that we could be sued about” (CS2). Moreover, there was
also doubt whether disclosing uncertainty information such as error rates would undermine the
validity of an article, as users may come to see the entire content as faulty.
Another issue hampering algorithmic transparency involves proprietary concerns, since dis-
closing source code could mean losing a competitive advantage or possibly opening a system up
to manipulation via feedback loops into the algorithm. At the same time, some organizations
may not be able to be as transparent as they might like “because the news organizations that are
buying [software/algorithms] can’t produce the codes” (CS2).
Even though participants did not provide clear incentives for organizations to implement
transparency, several discussants suggested that disclosure practices would improve perceptions of
credibility, reputation, and legitimacy. One participant argued that disclosing information about
algorithms is not necessarily about evaluating information, but “you kind of evaluate trust”
(CS1). Nevertheless, such beliefs were also met with questions like: “Do you think it would seri-
ously impact your credibility?” (CS3).
Generally, participants considered algorithmic transparency as a way to hold news organiza-
tions accountable. But they also recognized the strategic incentives for organizations to signal
greater transparency in order to increase legitimacy, while possibly withholding information that
could undermine the reputation of an organization.
Discussion and conclusion

The findings of this focus group study offer a variety of avenues for transparency of algorithmic
systems along the input-output pipeline. As transparency is increasingly becoming a core concept
for journalistic accountability and journalistic truth telling, the discovery and recommendation
of feasible transparency practices is critical for improving journalism’s credibility and legitimacy.
259
Yet at the same time our results indicate two factors that may hinder the implementation of
transparency: (1) lack of business incentives and (2) information overload for users.
This empirical study echoes theoretical considerations around the ethics of algorithmic
journalism including data and code transparency (see Dörr, 2016; Dörr and Hollnbuchner,
2017; Dörr, this volume, Chapter 23). Our findings also expand recently proposed ethical
frameworks around algorithms (see Ananny, 2016) by explicitly stressing the role of human
involvement in the production of algorithmic systems. While Neyland (2015) made first
inroads through ethnographic techniques underlining the dynamic human design processes
in developing algorithmic systems, it remains a challenge for future studies to separate human
and algorithmic decision processes while also acknowledging the intrinsic hybridity of such
automated systems.
We argue that the various factors of disclosure this study proposes (see Table 19.1) are
critical for an emerging standard of algorithmic transparency, which may be quintessential for
increasing institutional legitimacy. An algorithmic transparency standard would then work
twofold. On the one hand it could inform use and design of algorithms within the news media
industry, and on the other it would allow institutional members and the public to hold orga-
nizations accountable. The institutional implications of furthering algorithmic transparency
may not become apparent in the short term, as adopting, practicing, and integrating transpar-
ency into everyday journalism should be considered a longitudinal process. Time is a critical
factor in the adoption and dissemination of innovations (Rogers, 2003) such as algorithmic
transparency.
The consideration of time may be especially critical when considering that many new
media platforms have been built on different foundations in comparison to traditional media,
because such new media platforms often focus or build their value systems around the end user
experience (Ananny and Crawford, 2015), frequently neglecting issues of greater public inter-
est such as diversity and accountability including transparency (Napoli, 2015). Our analysis
of focus group discussions suggest a similar trend, as participants frequently pointed out that
transparency information may diminish the user experience. The difference in institutional
orientation indicates that new media platforms may have a more pragmatic and customer-
oriented outlook rather than an ethical outlook, which in turn raises questions regarding
such organizations’ willingness to be increasingly transparent in the first place, especially as no
proven business incentives seem to currently exist (Fengler and Russ-Mohl, 2014; Saurwein
et al., 2015). Yet a changing audience that expects more transparency (McBride and Rosenstiel,
2014) could be a trigger for more disclosure in the future. Transparency may then be under-
stood not simply as an ethical principle but as a viable avenue to enhance the user experience
(Diakopoulos et al., 2014).
While the focus on user experience could ultimately lead to greater transparency, there is a
much more urgent need to bolster institutional values, perceptions, and practices of media work in
general and journalism in particular. The adoption of isomorphic practices and values can increase
institutional legitimacy (DiMaggio and Powell, 1983). In a primarily digital media environment
that allows, voluntarily and involuntarily, for more disclosure about the journalistic production and
producers, such isomorphic practices include transparency as a path toward greater media account-
ability. In order to strengthen the institutional standing of media organizations, institutional mem-
bers need to adopt and implement transparency practices in their daily work (Meyer and Rowan,
1977), creating an impetus to be more transparent despite concerns about the user experience. Such
disclosure practices may include ombudspeople’s periodic transparency reports, editorial explana-
tion, disclosure of source code, error analysis, or the methodology of data collection and processing.
Such practices do not only hold media organizations accountable but also form and uphold an
institutional culture of transparency and accountability. In fact, a survey among journalists found
260
that upholding organizational or editorial guidelines (59%) and institutional or professional codes of
ethics (50%) can impact journalistic behavior the most (Powell and Jempson, 2014).
Normatively speaking, transparency around algorithms is of vital importance even if there are
practical concerns such as information overload. Many users may not care on a daily basis about
the methodology of algorithmic systems or the quality of the data. But the news media serves a
broad public, of which some concerned users may value expert accounts and explanations (Tilly,
2006). Thus in order to satisfy the various information needs of users – from general users to
experts – we propose a multi-layered ‘pyramid’ model of transparency. This pyramid starts at
the tip with a hint or an opening to transparency information at the user-interface level. Users
could then work their way down to gradually more detailed levels of explanations of the system
and content production processes. Such an approach could decrease the danger of information
overload by offering increasingly more data from one level to the next.
While our framework for this study – based on three case studies – is not exhaustive in regard
to the use of algorithms in news systems, we believe that this study’s pragmatic approach will
offer a basis for an algorithmic transparency standard that can be built upon in future research.
In summary, this research makes clear that algorithmic transparency requires the consideration
of multiple stakeholder perspectives, including organizational incentives and costs, end user tasks
and utility, and ethical practices.
Further readings
To see various demonstrations of the practical application of the transparency model presented in
this chapter see “Enabling Accountability of Algorithmic Media: Transparency as a Constructive
and Critical Lens” by Nicholas Diakopoulos. A healthy critical examination of transparency and
its various shortcomings is presented by Mike Ananny and Kate Crawford in “Seeing Without
Knowing: Limitations of the Transparency Ideal and Its Application to Algorithmic Accountabil-
ity”. Finally, for a more general treatment of ethical implications regarding the use of algorithms
throughout various aspects of society, see “The Ethics of Algorithms: Mapping the Debate” by
Brent Mittelstadt and colleagues. To keep on top of the state-of-the-art in how to be transparent
about algorithms, the Fairness, Accountability, and Transparency in Machine Learning (FATML)
workshop is an excellent venue: www.fatml.org.
Note
1 The notion of computational journalism or “algorithmic journalism” (Dörr, 2016) refers to finding, tell-
ing, and disseminating news stories with or by algorithms.
References
Albeanu, C. (2015) “Inside velocity: Mashable’s predictive social analytics tool.” Journalism.co.uk. Retrieved
from www.journalism.co.uk/news/inside-velocity-how-mashable-s-predictive-social-analytics-tool-
fits-in-the-newsroom/s2/a565861/
Allen, D. S. (2008) “The trouble with transparency.” Journalism Studies, 9(3), 323–340.
Ananny, M. (2016) “Toward an ethics of algorithms: Convening, observation, probability, and timeliness.”
Science, Technology, & Human Values, 41(1), 93–117.
Ananny, M. and Crawford, K. (2015) “A liminal press: Situating news app designers within a field of net-
worked news production.” Digital Journalism, 3(2), 192–208.
Ananny, M. and Crawford, K. (2016) “Seeing without knowing: Limitations of the transparency ideal and
its application to algorithmic accountability.” New Media & Society, 20(3), 973–989.
Associated Press. (2015) “AP, NCAA to grow college sports coverage with automated game stories.” Asso-
ciated Press. Retrieved from www.ap.org/content/press-release/2015/ap-ncaa-to-grow-college-sports-
coverage-with-automated-game-stories
261
Broersma, M. J. (2013) “A refractured paradigm. Journalism, hoaxes and the challenge of trust.” In C.
Peters and M. J. Broersma (eds.), Trust and Participation in a Transformed News Landscape. New York, NY:
Caswell, D. and Dörr, K. (2017) “Automated journalism 2.0: Event-driven narratives.” Journalism Practice,
12(4), 477–496.
Chadha, K. and Koliska, M. (2014) “Newsrooms and transparency in the digital age.” Journalism Practice,
9(2), 215–229.
Coddington, M. (2015) “Clarifying journalism’s quantitative turn: A typology for evaluating data
journalism, computational journalism, and computer-assisted reporting.” Digital Journalism, 3(3),
331–348.
Cramer, H., Evers, V., Ramlal, S., van Someren, M., Rutledge, L., Stash, N., Aroyo, L. and Wielinga, B. (2008)
“The effects of transparency on trust in and acceptance of a content-based art recommender.” User Mod-
eling and User-Adapted Interaction, 18(5), 455–496.
Deuze, M. (2005) “What is journalism?: Professional identity and ideology of journalists reconsidered.”
Journalism, 6(4), 422–464.
Diakopoulos, N. (2016) “Computational journalism and the emergence of news platforms.” In B. Franklin
and S. Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon: Routledge (pp.
176–184).
Diakopoulos, N., Cass, S. and Romero, J. (2014) “Data-driven rankings: The design and development of
the IEEE top programming languages news app.” Proceedings of Symposium on Computation + Journalism.
Diakopoulos, N. (2017) “Enabling accountability of algorithmic media: Transparency as a constructive and
critical lens.” In. T. Cerquitelli, D. Quercia, and F. Pasquale (eds.), Transparent Data Mining for Big and Small
Data. Cham, Switzerland: Springer (pp. 25–44).
DiMaggio, P. and Powell, W. (1983) “The iron cage revisited: Institutional isomorphism and collective ratio-
nality in organizational fields.” American Sociological Review, 48(2), 147–160.
Dörr, K. (2016) “Mapping the field of algorithmic journalism.” Digital Journalism, 4(6), 700–722.
Dörr, K. and Hollnbuchner, K. (2017) “Ethical challenges of algorithmic journalism.” Digital Journalism,
5(4), 404–419.
El-Arini, K., Paquet, U., Herbrich, R., Van Gael, J. and Agüera y Arcas, B. (2012) “Transparent user models
for personalization.” Proceedings of International Conference on Knowledge Discovery and Data Mining (KDD)
(pp. 678–686).
Fengler, S. and Russ-Mohl, S. (2014) “The (behavioral) economics of media accountability.” In S. Fengler,
T. Eberwein, G. Mazzoleni, C. Porlezza and S. Russ-Mohl (eds.), Journalists and Media Accountability: An
International Study of News People in the Digital Age. New York, NY: Peter Lang Publishing (pp. 213–230).
Friedman, B. and Nissenbaum, H. (1996) “Bias in computer systems.” ACM Transactions on Information
Systems, 14(3), 330–347.
Gillespie, T. (2014) “The relevance of algorithms.” In T. Gillespie et al. (eds.), Media Technologies: Essays on
Communication, Materiality, and Society. Cambridge, MA: The MIT Press (pp. 167–194).
Graefe, A. (2016) “Guide to automated journalism.” Tow Center for Digital Journalism Report. Retrieved from
https://towcenter.org/research/guide-to-automated-journalism/
Grochowski Jones, R. and Ornstein, C. (2016) “Matching industry payments to Medicare prescribing
patterns: An analysis.” ProPublica Whitepaper. Retrieved from https://static.propublica.org/projects/
d4d/20160317-matching-industry-payments.pdf?22
Hayes, A. S., Singer, J. B. and Ceppos, J. (2007) “Shifting roles, enduring values: The credible journalist in
a digital age.” Journal of Mass Media Ethics, 22(4), 262–279.
Hellmueller, L., Vos, T. P. and Poepsel, M. A. (2013) “Shifting journalistic capital?” Journalism Studies, 14(3),
287–304.
Karlsson, M. (2010) “Rituals of transparency.” Journalism Studies, 11(4), 535–545.
Karlsson, M. (2011) “The immediacy of online news, the visibility of journalistic processes and a restructur-
ing of journalistic authority.” Journalism, 12(3), 279–295.
Karlsson, M., Clerwall, C. and Nord, L. (2014) “You ain’t seen nothing yet.” Journalism Studies, 15(5),
668–678.
Karlsson, M., Clerwall, C. and Nord, L. (2017) “Do not stand corrected: Transparency and user’s attitudes
to inaccurate news and corrections in online journalism.” Journalism & Mass Communication Quarterly,
94(1), 148–167.
262
Kemper, E., Stringfield, S. and Teddlie, C. (2003) “Mixed methods sampling strategies in social science
research.” In A. Tashakkori and C. Teddlie (eds.), Handbook of Mixed Methods in Social & Behavioral
Research. Thousand Oaks, CA: Sage Publications (pp. 273–296).
Kizilcec, R. F. (2016) “How much information?: Effects of transparency on trust in an algorithmic inter-
face.” Proceedings of Conference on Human Factors in Computing Systems (CHI).
Koliska, M. and Chadha, K. (2017). “Transparency in German newsrooms: Diffusion of a new journalistic
norm?” Journalism Studies. [online first] doi: 10.1080/1461670X.2017.1349549
Should Expect (3rd ed.). New York, NY: Three Rivers Press.
Lewis, S. C. and Usher, N. (2013) “Open source and journalism: Toward new frameworks for imagining
news innovation.” Media, Culture & Society, 35(5), 602–619.
Lofland, J. and Lofland, L. H. (1994) Analyzing Social Settings: A Guide to Qualitative Observation and Analysis
(3rd ed.). Belmont, CA: Wadsworth.
Lokot, T. and Diakopoulos, N. (2016) “News bots: Automating news and information dissemination on
Twitter.” Digital Journalism, 4(6), 682–699.
McBride, K. and Rosenstiel, T. (2014) The New Ethics of Journalism: Principles for the 21st Century. Thousand
Oaks, CA: CQ Press.
Meyer, J. W. and Rowan, B. (1977) “Institutionalized organizations: Formal structure as myth and cer-
emony.” American Journal of Sociology, 83(2), 340–363.
Mittlestadt, B. D., Allo, P., Taddeo, M., Wachter, S., and Floridi, L. (2016) “The ethics of algorithms: Mapping
the debate,” Big Data and Society, 3(2), 1–21.
Montal, T. and Reich, Z. (2016) “I, robot. you, journalist. Who is the author?” Digital Journalism, 5(7):
829–849.
Morgan, D. L. (1996) “Focus groups.” Annual Review of Sociology, 22, 129–152.
Napoli, P. M. (2015) “Social media and the public interest: Governance of news platforms in the realm of
individual and algorithmic gatekeepers.” Telecommunications Policy, 39(9), 751–760.
Neyland, D. (2015) “Bearing account-able witness to the ethical algorithmic system.” Science, Technology, and
Human Values, 41(1), 50–76.
Nguyen, D., Kelleher, A. and Kelleher, A. (2015) “Introducing pound: Process for optimizing and under-
standing network diffusion.” BuzzFeed Tech Blog. Retrieved from www.buzzfeed.com/daozers/
introducing-pound-process-for-optimizing-and-understanding-n
Nissenbaum, H. (1996) “Accountability in a computerized society.” Science and Engineering Ethics, 2(1),
25–42.
Oremus, W. (2014) “The first news report on the L.A. earthquake was written by a robot.” Slate. Retrieved
from www.slate.com/blogs/future_tense/2014/03/17/quakebot_los_angeles_times_robot_journal-
ist_writes_article_on_la_earthquake.html
Park, D., Singh, S., Diakopoulos, N. and Elmqvist, N. (2016) “Interactive identification of high quality
online news comments.” Proceedings of Conference on Human Factors in Computing Systems (CHI).
Pew Center. (2012) Further Decline in Credibility Ratings for Most News Organizations. Retrieved from www.
people-press.org/files/2012/08/8-16-2012-Media-Believability1.pdf
Plaisance, P. L. (2007) “Transparency: An assessment of the Kantian roots of a key element in media ethics
practice.” Journal of Mass Media Ethics, 22(2&3), 187–207.
Powell, W. and Jempson, M. (2014) “More accountability in the digital age? The influence of new tech-
nologies.” In S. Fengler, T. Eberwein, G. Mazzoleni, C. Porlezza and S. Russ-Mohl (eds.), Journalists
and Media Accountability: An International Study of News People in the Digital Age. New York: Peter Lang
Publishing (pp. 115–128).
Puchta, C. and Potter, J. (2004) Focus Group Practice. London: Sage.
Roberts, C. (2007) Measuring the Relationship Between Journalistic Transparency and Credibility. Doctoral thesis,
University of South Carolina.
Rogers, E. M. (2003) Diffusion of Innovations. New York, NY: Free Press.
RTDNA. (2015) “New RTDNA code of ethics.” Radio Television Digital News Association. Retrieved from
http://rtdna.org/content/rtdna_code_of_ethics#.VZRTgO1VhBc
Saurwein, F., Just, N. and Latzer, M. (2015) “Governance of algorithms: Options and limitations.” Info, 17(6),
35–49.
Schaffer, J., Giridhar, P., Jones, D., Höllerer, T., Abdelzaher, T. and O’Donovan, J. (2015) “Getting the mes-
sage?: A study of explanation interfaces for Microblog data analysis.” Proceedings of Conference on Intelligent
User Interfaces (IUI).
263
Schudson, M. (2015) The Rise of the Right to Know: Politics and the Culture of Transparency, 1945–1975. Cam-
bridge, MA: The BelkPress of Harvard University Press.
Shearer, M., Simon, B. and Geiger, C. (2014) “Datastringer: Easy dataset monitoring for journalists.” Proceed-
ings of Symposium on Computation + Journalism.
Silverman, C. (2013) “Corrections and ethics.” In K. McBride and T. Rosenstiel (eds.), The New Ethics of
Journalism – Principles for the 21st Century. Thousand Oaks, CA: Sage Publications (pp. 151–161).
Singer, J. B. (2007) “Contested autonomy.” Journalism Studies, 8(1), 79–95.
Spayd, L. (2016, September 24) “Taking the stealth out of editing.” The New York Times. Retrieved from
www.nytimes.com/2016/09/25/public-editor/liz-spayd-new-york-times-public-editor.html
SPJ. (2014) “Code of ethics.” Society of Professional Journalists. Retrieved from www.spj.org/ethicscode.asp
Tandoc, E. C. and Thomas, R. J. (2017) “Readers value objectivity over transparency.”Newspaper Research
Journal, 38(1), 32–45.
Thurman, N., Schifferes, S., Fletcher, R., Newman, N., Hunt, S. and Schapals, A. K. (2016) “Giving com-
puters a nose for news: Exploring the limits of story detection and verification.” Digital Journalism, 4(7),
838–848.
Tilly, C. (2006) Why? What Happens When People Give Reasons . . . and Why. Princeton, NJ: Princeton
University Press.
Vos, T. and Craft, S. (2016) “The discursive construction of journalistic transparency.” Journalism Studies,
18(12), 1505–1522.
Wang, S. (2015) “The New York Times built a slack bot to help decide which stories to post to social
media.” NiemanLab. Retrieved from www.niemanlab.org/2015/08/the-new-york-times-built-a-slack-
bot-to-help-decide-which-stories-to-post-to-social-media/
Weinberger, D. (2009) Transparency is the New Objectivity. Retrieved from www.hyperorg.com/
blogger/2009/07/19/transparency-is-the-new-objectivity
Yahoo (n.d.) “Matchup recaps.” Yahoo.com. Retrieved from https://help.yahoo.com/kb/SLN8766.html
264
20
VISUAL NETWORK EXPLORATION
FOR DATA JOURNALISTS
Tommaso Venturini, Mathieu Jacomy, Liliana Bounegru,
and Jonathan Gray
Networks are classic but under-acknowledged figures of journalistic storytelling. Who is con-
nected to whom and by which means? Which organizations receive support from which others?
What resources or information circulate through which channels, and which intermediaries
enable and regulate their flows? These are all customary stories and lines of inquiry in journal-
ism, and they all have to do with networks. Additionally, the recent spread of digital media has
increasingly confronted journalists with information coming not only in the traditional form
of statistic tables but also of relational databases. Yet journalists have so far made little use of the
analytical resources offered by networks.
To address this problem in this chapter, we examine how ‘visual network exploration’ may be
brought to bear in the context of data journalism to explore, narrate, and make sense of large and
complex relational datasets. We illustrate this technique through a network of French informa-
tion sources compiled by Le Monde’s The Décodex. We establish that visual exploration is an
iterative process where the demarcation of categories and territories are entangled and mutually
constitutive. To enrich investigation, we suggest ways in which the insights of the visual explora-
tion of networks can be supplemented with simple statistics on the distributions of nodes and
links. We conclude with reflection on the knowledge-making capacities of this technique and
how these compare to the insights that journalists have used in the Décodex project – suggesting
that visual network exploration is a fertile area for further exploration and collaborations between
data journalists and digital researchers.
Introduction
Few people know as well as journalists that the world is made of relations. Following alliances,
unveiling links, and unraveling threads are, and have long been, a central part of their investiga-
tions. If social scientists can speculate about long-standing structures and global arrangements,
journalists have no such leisure. Their work consists in tracing the specific associations that
connect individuals and institutions to uncover how lumps of money, influence, and knowledge
are exchanged through them and where unethical behavior, corruption, fraud, or unfair politi-
cal influence may occur. The advent of digital technologies has made such work both easier
and more difficult: easier, because it has increased the traceability of economic and political
265
Tommaso Venturini et al.
associations; more difficult, because it has submerged journalists under more information than
their investigative toolkit is used to handling.
When, for example, the reporters of the International Consortium for Investigative Jour-
nalism (ICIJ) received the 2.6 terabytes and 11.5 million documents composing the so-called
‘Panama Papers’, they obviously could not process them manually (Baruch and Vaudano, 2016).
Note that this is not just a ‘big data’ problem. The trouble with the leak was not only its size
but the fact that its interest came from the links it established between specific individuals and
particular tax havens. Extracting ‘key’ figures through statistical aggregation or abstracted com-
putational models would miss the point of many of the stories that journalists were most keen
to explore. The inquiry could not simplify the dataset but had to explore each and every one of
the connections it exposed. This was done, among other ways, through a tool called Linkurious
(http://linkurio.us), whose interest comes less from its computational power than from the way
in which it allows its users to see and follow the connections of a network.
The Panama Paper case is interesting but also interestingly isolated. Despite long-standing
interest, the use of networks in journalism remains comparatively marginal (cf. Bounegru et al.,
2016 for an overview of the emerging uses of networks in journalism). The reasons are not dif-
ficult to imagine. Graph mathematics is more demanding and less widely known than traditional
statistical approaches and does not come with the same readily accessible and publicly recognized
vocabulary of visual motifs. With all its computational power, graph mathematics does not fit
journalistic needs because it tends to be obscure for both reporters and their readers.
In this chapter, we address this difficulty by suggesting a technique for the visual exploration
of networks. As we will try to show, when performed correctly, the visual representation of net-
works translates some of the most important graph structures into graphical variables (thereby
supporting investigative work) and allowing the interpretation of networks with conventions
similar to those developed for geographical maps (thereby remaining legible for a large audience).
After having introduced the mathematical and historical bases of our approach, we will present
our technique for the visual exploration of networks. Using as an example the network of the
French information sphere, we will illustrate the recursive work of interpretation and categoriza-
tion that allows reading the network as an organized territory. Visual network exploration, which
is growing in prominence amongst digital methods researchers for social and cultural research,
may be useful not only for studying media landscapes but also for digital journalism practitioners
who are interested in exploring and telling stories with networks and relational data.
Understanding force-directed layouts

Far from being merely aesthetic, the graphical representation of networks has an intrinsic herme-
neutic value, which you will have experienced if you have ever used a public transportation map.
Such maps are distinctively different from road maps or city maps. It is not only that transportation
maps are simpler (the level of details depending only on the resolution of the map), it is that they
represent a network and not a geographical territory. An illustration of this difference can be found
in the famous map of the London tube as designed by Harry Beck in 1933. Before Beck’s redesign,
the diagram was a classic geographical map locating stations according to their coordinates. After the
redesign, it became the network of correspondences we are still using today. The gain in legibility
is evident, as the function of transportation maps is not to situate stations in urban space but relative
to each other, so as to help users to move from one to another (a type of orientation that resembles
strikingly to one used by traditional sea navigators; see, for example, Turnbull, 2000: 133–165).
Another example of such a mapping approach comes from early work in social network
analysis (Freeman, 2000). Jacob Moreno, founder of SNA, is explicit about the importance of
266
Visual network exploration
visualization: “A process of charting has been devised by the sociometrists, the sociogram, which
is more than merely a method of presentation. It is first of all a method of exploration” (1953:
95–96). In an interview released by Moreno to the New York Times in 1933, network analysis is
presented as a ‘new geography’. More important than the title, however, is the figure that accom-
panies that interview, depicting friendships among fourth-grade pupils (see Figure 20.1). The
sociogram presented by these figures powerfully reveals how friendship is not equally distributed
in the class. One only need to know that triangles represent boys and circles girls to see how inter-
gender relationships are discouraged at that specific age (or at least the declaration of such friend-
ships). The trick, of course, only works because the nodes are not positioned randomly in the space
but in a way that minimizes line crossing (in Moreno’s own words “the fewer the number of lines
crossing, the better the sociogram” (1953: 141)). It is because triangles are pushed on one side and
circles on another that it is easy to spot the existence of a single inter-gender connection.
Moreno’s rule of spatialization is easy to follow on a graph of a few dozen nodes and edges
but impracticable on larger networks. Graphs with thousands of nodes and edges are so intricate
(a)
(b)
Figure 20.1 Sociogram representing friendship among school pupils (original title and image accompanying
Moreno’s 1933 New York Times interview) (a) in the original version and (b) in the modern
force-directed spatialization
267
that the direct counting of line-crossings becomes prohibitively time consuming. An indirect
approach consists of drawing closer the connected nodes to minimize the length of the edges and
therefore the possibility of crossings. But even in this case, since each node may be connected to
several other nodes that are themselves connected to many other nodes, minimizing the length of
the edges is far from a trivial exercise.
Thus, we might explore the network using a technique called ‘force-directed spatialization’.
Such spatialization follows a physical analogy: nodes are charged with a repulsive force that drives
them apart, while edges act as springs binding the nodes that they connect. Once the algorithm
is launched it changes the disposition of nodes until it reaches a balance of such forces (Jacomy
et al., 2014). Such equilibrium reduces line-crossings and improves the legibility of the graph.
Früchterman and Reingold (1991), who proposed the first efficient force-directed algorithm, cite
line crossing as the second of their aesthetic criteria.
Yet scholars working with networks soon realized that avoiding line crossing is not the most
interesting effect of force-directed layouts. At equilibrium, the visual density of nodes and edges
becomes an approximate but reliable proxy of the mathematical structure of the graph (for a
detailed mathematical proof, see Venturini et al., forthcoming). Groups of nodes gathering in the
layout tend to correspond to the clusters identified by community-detection techniques (Noack,
2009); structural holes (Burt, 1995) tend to look like sparser zones; central nodes move toward
middle positions; and bridges are positioned some way between different regions (Jensen et al.,
2015).
The trick of force-directed algorithms is all the more remarkable, given that the space of net-
works is relative rather than absolute (it can be rotated or mirrored without distortion of infor-
mation) and that it is a consequence and not a condition of element positioning. In traditional
geographical representation, the space is defined a priori by the way the horizontal and vertical
axes are constructed. Points are projected on such preexisting space according to a set of rules that
assign a univocal position to a pair of coordinates. The same is true for any Cartesian diagram
(scatterplots, for instance) but not for networks, in which the space is defined by the position of
the nodes and not the other way around.
Despite such differences (which should not be forgotten), force-directed algorithms allow
reading networks as geographical maps, translating complicated mathematical concepts into more
conventional vocabulary of regions and margins, path and landmarks, centers and peripheries
(Lynch, 1960). This is a crucial advantage that explains why force-directed algorithms have
become the de facto standard of network visualization: they facilitate the exploration of networks
and relations by means of more familiar and intuitive spatial metaphors, as well as through less
familiar computational and statistical metrics.
The Décodex: a controversial case study

In the following pages, we will illustrate the technique of visual network exploration draw-
ing on a concrete example. Our case study is a network of websites extracted from a listing
compiled by the French journal Le Monde. Since 2009, a group of journalists gathered under
the name of Les Décodeurs (www.lemonde.fr/les-decodeurs/article/2014/02/12/l-equipe-des-
decodeurs_4365082_4355770.html) has verified the accuracy of thousands of stories circulating
in the French blogosphere and in social media. In January 2017 (at the beginning the French
presidential campaign), Les Décodeurs launched an online tool called the Décodex (www.lem
onde.fr/verification), allowing readers to search for the most important sources of online infor-
mation relevant to French public debates (though not necessarily in French). Each source is
accompanied by a short description and, more crucially, by an evaluation of its trustworthiness
according to the journalists of Le Monde.
268
Not surprisingly, the classification provided by Les Décodeurs has stirred considerable debate
in French media spheres. Several of the sources categorized as imprecise or unreliable, along with
other newspapers and blogs, have contested the Décodex, with critique spanning from chal-
lenging the way in which websites are oversimplistically classified to questioning the right of Le
Monde (which is itself a rival source of information) to note the reliability of other websites, to
disputing the legitimacy and interest of such classification in general (arguing that some of the
websites in the list mean to circulate opinions rather than information). Les Décodeurs themselves
admitted the difficulty of their exercise, the many ambiguities that they were obliged to decide
on, and the errors and inaccuracies that may have derived from them. At the same time, they
defended their work by pointing to the increasing quantity of false or partisan information cir-
culating online and by affirming their openness to discussing their classification and revising it
if necessary.
The controversy around the Décodex is a good example of difficulties connected to the
detection of fake news online (Bounegru et al., 2017), but also of the more general debates sur-
rounding all kind of classifications. Categorizing things is never a self-evident or innocent prac-
tice (Bowker and Star, 1999) and should always be carried out with the greatest caution. This is
true for the initial classification of the Décodex, but it is also true for the network extracted from
it. As we will see in the following pages, the visual exploration of network involves a constant
toing and froing of categorization and observation, typology, and topology.
To build our example network, we have extracted, in collaboration with Les Décodeurs, all the
websites contained in the Décodex and investigated the way in which they cite each other. To do
so, we employed Hyphe (http://hyphe.medialab.sciences-po.fr), a web crawler developed by the
médialab of Sciences Po that facilitates the exploration of websites and following the hyperlinks
present in their pages (Girard, 2011; Girard et al., 2012; Jacomy et al., 2016). All the websites
comprising the Décodex corpus have been crawled at a depth of one click, starting from the
homepage. We so obtained a network with 653 nodes and 5,943 edges. While Les Décodeurs focus
on editorial judgments about how to classify websites in the French media landscape, our network
exploration examines the relations between them and other websites by means of their linking
practices. While some researchers focus on how networks are held together through financial ties,
organizational affiliations, business relationships, and family and social relations, we consider their
relations according to the hyperlink, in accordance with a longer tradition of digital methods,
digital sociology, and new media studies research (see, e.g., Marres and Rogers, 2005; Rogers, 2013)
The treatment of social platforms (such as Facebook, Twitter, YouTube, etc.) in our crawl
requires some additional explanation. These platforms are both sources of information as a whole
and containers of multiple individual sources in the form of pages or accounts. Since extracting
all the hyperlinks from a site as large as Facebook would have been impossible, we only crawled
the accounts that were specifically mentioned in the Décodex. We have, however, kept a record
of all the links pointing toward the main social media platform to investigate how they are cited
by the other websites of our corpus.
A visual exploration of the Décodex network

The visual exploration of networks exploits three visual variables to graphically represent their
features: position, size, and hue (for a definition of these variables and their semiotic affordances,
see Bertin, 1967). For the reasons discussed earlier, position is crucial in translating the mathemat-
ical characteristics of the graphs. Force-directed layouts create regions where numerous nodes are
densely assembled and regions that are less crowded. These differences of density, determined by
the uneven distribution of links, reveal the uneven association among the entities of the network.
Everything may be connected in this world, but not everything is equally connected.
269
Discerning the spatial structure of networks, however, is not always straightforward. In the
easiest cases, the difference in the density of association is such that clusters appear as well-defined
knots of nodes and edges separated by empty (or almost empty) zones. These zones are called
“structural holes” (Burt, 1995), and, when they exist, they provide a crucial guidance for the
interpretation of the network. Thanks to the ruptures created by structural holes, the boundaries
of clusters can be easily detected, like cliffs separating a plateau from a valley. Most natural and
social networks, however, do not exhibit such a clear separation, and the borders of their cluster
tend to be gradual as the hillside slopes. The fuzziness of clusters’ frontiers is not necessarily an
obstacle to their recognition (one can point at a hill even when it is impossible to say exactly
where it starts and ends), but it certainly makes their identification more difficult. This is why
visual network analysis is often more like an exploratory expedition – where meanings and
findings are progressively and hermeneutically generated – than statistical confirmation of a set
of preexisting hypotheses (on the difference between exploratory and confirmatory analysis see
Tukey, 1977, and Behrens and Chong-Ho, 2003).
This is certainly the case for our Décodex network, which, at a first look, does not present any
manifest structural hole or any clear spatial structure. To visualize our network we used two main
tools: Gephi (https://gephi.org) for filtering and spatializing the network (using in particular
the force-driven algorithm ForceAtlas2) and Graph Recipes (http://tools.medialab.sciences-po.
fr/graph-recipes) to tweak the visual rendering of the network (see Figure 20.2). Though no
structural holes are evident in the Décodex network, looking closely at the layout makes it pos-
sible to notice that the network does not spatialize as a perfect circle but rather in an avocado-like
shape with a smaller top and a larger bottom. These irregularities (as weak and subtle as they can
be) often suggest the presence of polarizing effects that can be interesting to investigate further.
The first and most crucial way to explore our network is to look at the identity of the nodes
that occupy its different regions. This may seem trivial, but it is not. It is a distinct advantage of
visual exploration, compared to other forms of statistical analysis, that it does not aggregate the
individual entities that compose its corpus: each and every node is visible in the layout and can
be interrogated by the researcher. Even on a small network like the one in our example, however,
the quantity of nodes can make it difficult (and time consuming) to look at all of them.
This is where the second variable of our visual exploration, size, comes in handy. Since, in
networks, nodes are defined first and foremost by their connections, we have ranked the nodes
according to the number of edges pointing to them. In the jargon of network analysis, this num-
ber is called ‘in-degree’, and nodes with an elevated in-degree are called ‘authorities’, because
they are recognized and referred to by many others. In the previous figure and in all following,
we have sized the nodes according to their in-degree so that a greater authority literally translates
into increased visual prominence.
Reading the names of websites that occupy the two poles of our avocado, it seems natural to
suppose that their separation derives from a linguistic fracture. The websites in the lower part are
predominantly French, while those in the upper part are more international. A way to highlight
this is to show the uneven distribution of TLD (top-level domain) in the network (Figure 20.3).
The linguistic separation we just highlighted, however, is not particularly surprising or inter-
esting. This kind of division is regularly observed in networks of websites and hyperlinks. Detect-
ing it is important, but in a negative way – it makes us aware that in order to generate more
interesting findings, we will have to look beyond it.
Further exploring the network, we may notice the role of not just languages but also social
network platforms, such as YouTube, Facebook, Twitter, Instagram, and Dailymotion. With the
remarkable exception of Wikipedia, all the main social media platforms are located in the middle
right of the layout – somewhere in between the English and the French websites (as one would
270
Figure 20.2 The Décodex network spatialized by ForceAtlas2. The size of nodes is proportional to
in-degree.
expect given the multilinguality) but also separated from both by their distinctive nature (and
possibly by the different way in which they have been treated in the crawl).
Moreover, by focusing on the lower and larger part of the network, we can recognize two
different sub-poles, with national sources (such as Le Monde, Le Figaro, FranceInfo, Libération, etc.)
occupying most of the lower region and the regional press clustering at the bottom right of the
layout (see Figure 20.4).
The distinctive position of the platforms and the national/regional press are both interesting
and nontrivial findings, but we can push our analysis further. The way to do so is by playing with
the third visual variable exploited by visual exploration of network: the hue of the node. This is
a laborious but revealing part of our visual exploration. It consists of categorizing the nodes of
the network according to multiple classifications and visualizing these classes on the network as
different colors or (as in this chapter) as different shades of gray. It is important to notice that the
operation of classifying the nodes and of reading the disposition of classes are not separated but
271
Figure 20.3 Distribution of TDL in the Décodex network
performed at the same time. As will become clear in the following pages, our technique does not
consist simply in the projection of a set of preexisting categories on a connectivity-based layout
but on recursively using the categories to make sense of the layout and the layout to define the
categories. It is important to remember that the color is a ‘non-mixable’ visual variable. A node
can be red or blue, for example, but not the two at the same time. When categorizing nodes, it is
therefore necessary to employ exclusive categories. A website, for example, can be classed in the
category ‘news’ or ‘satire’, but not in both. In the (not uncommon) case of nodes resisting a unique
classification, researchers can introduce a residual category such as ‘multiple’ or ‘miscellaneous’.
As a first step in our combined exploration of topology and typology, we will color the nodes
of the network according to the original categories of the Décodex. These categories refer to
the trustworthiness of the sources, as manually assessed by the journalists of Le Monde in the four
categories ‘reliable’, ‘imprecise’, ‘unreliable’, and ‘satirical’. Precisely because these categories have
been defined before and independently from the extraction of the network, their disposition does
not follow the spatial articulation of the network. Rather, it is possible to find nodes of every
272
Figure 20.4 Zoom on the French regional press
category in most regions of the network. A remarkable exception is the satirical websites that
are to be found on the right side of the layout both in its upper and lower part (see Figure 20.5).
Arguably, this position is not due to the hyperlinks between the satirical websites (which do not
cite each other very much) but by their strong connection with social media platforms to which
all these sites extensively link.
The other classes are distributed more evenly but not randomly. The ‘reliable’ websites tend to
occupy the center of both in the international and French pole, while the ‘imprecise’ and ‘unreli-
able’ take a more marginal position (see Figure 20.6). More interestingly, looking at the lower part
of the network, we observe two groups of ‘imprecise’ and ‘unreliable’ sources – while a major-
ity of these nodes are positioned above the core of national and reliable websites (and hence in
between the French and the international website), a significant minority is located below them.
To account for this separation, we introduce an additional categorization based on the political
leaning of the websites (Figure 20.7). In particular, we distinguish the websites that disseminate
unreliable or imprecise information because they pursue a right-wing or extreme-right agenda
Figure 20.5 The ‘satirical’ websites according to the original Décodex classification (nodes have been
highlighted by the black color and by doubling their radius despite their low degree)
274
Figure 20.6 Highlight of the ‘reliable’ websites (left) and ‘unreliable’ and ‘imprecise’ websites (right)
Figure 20.7 Highlight of the ‘conspiratorial’ websites (left) and ‘right’ and ‘extreme right’ websites (right)
(which occupy the center of the network) and the websites exhibiting a more general conspirato-
rial attitude (which occupy the bottom of the network).
Through our iterative exploration of typology and topology, we have eventually revealed a
partitioning of the network that, while invisible at first glance, allows us to interpret some of the
main contours of the French media landscape. Though these territories are not separated by clear
structural holes, the nodes that they contain are fairly consistent. Interestingly, our final classifica-
tion produces a homogeneous partition of the layout not in spite of but because of its heterogeneity,
which mixes linguistic categories, trustworthiness classes, and political leanings. The fact that a
nonhomogenous categorization turns up to offer the best characterization of the structure of
our network should not come as a surprise. Networks are complex objects that articulate diverse
elements through disparate logics. In this, they remind us of a passage by Jorge Luis Borges cited
by Foucault (1970) as a perfect example of a heterogeneous classification that, while defying our
traditional categories, is nonetheless highly efficient to describe the culture in which it has been
elaborated (Figure 20.8). Borges refers to a “certain Chinese encyclopaedia” in which animals
are classified as
(a) belonging to the Emperor, (b) embalmed, (c) tame, (d) suckling pigs, (e) sirens, (f) fab-
ulous, (g) stray dogs, (h) included in the present classification, (i) frenzied, (j) innumer-
able, (k) drawn with a very fine camelhair brush, (l) et cetera, (m) having just broken the
water pitcher, (n) that from a long way off look like flies.
And Foucault concludes by observing that
In the wonderment of this taxonomy, the thing we apprehend in one great leap, the
thing that, by means of the fable, is demonstrated as the exotic charm of another system
of thought, is the limitation of our own.
(Foucault, 1970: XV)
Linking patterns in the Décodex network

Now that, by means of visual exploration, we have defined a heterogeneous but hermeneutically
robust partitioning of our network, we can use it as a basis for a statistical analysis. While prais-
ing the advantages of the visual interpretation, we are also aware that not all structural properties
can be rendered visually. The direction of edges or the connection between different classes, in
particular, is not easily read in network images. These questions, however, can be investigated by
other means once the partitioning of the network has been defined.
Figure 20.9 shows the distribution of nodes in the regions identified in our final classification,
to which we have added the ‘satirical’ websites (which we discussed earlier but are not included
in Figure 20.8 for the sake of legibility), as well as ‘other reliable’ and ‘other unreliable’. These two
residual categories comprise together about one-fifth of the nodes of the network. This relatively
high figure is not uncommon. Given the heterogeneity of the networks they work with, social
scientists and journalists should aim at classifications that are robust and insightful (capable of
delineating homogenous zones in the graph) rather than comprehensive.
Our empirical categories are powerful tools to unveil different linking strategies in the
network. Figure 20.10 presents the links in the corpus aggregated by categories. As we can
see, not all categories cite or are cited the same way. ‘French national media’ and ‘platforms’
are greatly cited and by various actors (their columns contain larger circles), while ‘satirical’
websites are scarcely cited (their column is almost empty). Platforms are not cited much, but
277
Figure 20.8 The heterogenous territories of the Décodex network
Figure 20.9 Distribution of the number of nodes per category
Figure 20.10 Connectivity between the categories of our final classification.

Rows convey how many times the nodes of a given category cite the nodes of other categories.
Columns convey how many times the nodes of a given category are cited by the nodes of other categories.
this is merely a consequence of our method since (as explained earlier) most of them had not
been not crawled. ‘Right-wing’, conspiracy theorist and other ‘unreliable’ websites are, on the
contrary, the origins of the highest number of citations and, very interestingly, they seem to
favor ‘reliable’ sources over ‘unreliable’ ones. As expected, the reliable websites do not link back
to them, and this asymmetry reveals an important hierarchy (see Figures 20.11 and 20.12). To
investigate this linking pattern, we will compare the incoming and outgoing links of some of
the most interesting categories.
This kind of hierarchical structure is common on the web and has been explained as a conse-
quence of preferential attachment (Barabási and Albert, 1999): actors tend to link to other web-
sites that they perceive as higher in the hierarchy and avoid linking to those that they perceive as
lower. This style of preferential attachment whereby smaller actors link to establishment actors
without reciprocation of the linking act has elsewhere been called “aspirational linking” (Rog-
ers, 2013). Links in a network do not always produce a hierarchy of categories, but this behavior
does. This linking pattern and the way it fits our empirical categories may suggest an alternative
way to characterize the trustworthiness being investigated by Le Décodeurs: reliable sources are
cited by all types of websites, while unreliable sources are only cited by few other types (if any).
This observation is in many ways at odds with what is often affirmed about the ‘post-truth
era’ in which we have supposedly landed. While fake news is said to leverage the horizontality
of digital media to blur the boundaries between true and false, the linking patterns of the French
information spheres suggest a different picture. Despite their different ideological leanings, all
websites agree on the overall hierarchy of reliability by citing in one sense and not in the other.
Figure 20.11 Hierarchical structure in the corpus, based on our final categories. Black arrows on the right
side summarize the links structure between these categories.
280
Figure 20.12 Simplified version of the statistical analysis presented in figure 20.11
The ‘right-wing’ websites, for example, try to blur the lines by citing both their peers and more
reliable sources, but they also try to draw a line between them and the even less reliable ‘con-
spiracy theorist’ websites. Whatever its position in the pyramid of hyperlinking, every actor tries
to improve its situation by linking upwards to authorities above and not linking to less reputable
websites below, thus reinforcing the hierarchy.
Conclusion
This chapter discussed the visual exploration of networks with the aim of improving the
understanding of one of the dominant visual-analytical forms of our digital age – the network
diagram – and its potential role in relation to the study and practice of digital journalism. Draw-
ing on graph semiotics and traditional cartography, this chapter proposed a model whereby the
interpretation of network topology, with its regions, paths, cores, and peripheries, is guided by
three visual variables: position, size, and hue. The process that we described is one that emphasizes
the exploratory and iterative character of the investigation. While it seemed counterintuitive at
first, we emphasized that in order to surface the multiple logics that play out in the structure of
a network graph, analysis should not limit itself to one classificatory principle. Multiple hetero-
geneous criteria of classification are often necessary to characterize the topology of a network
map. Finally, we advocated for mixing methods, complementing visual network exploration with
statistical analyses in order to further characterize network properties.
Through the case study of French media hyperlink map, we tried to show how the visual
exploration of networks reveals new angles that other analyses may leave unexplored. In this
case, the chapter illustrated an alternative way to assess websites’ reliability that complements the
traditional fact-checking approach of qualifying content with an examination of the linking
patterns between different regions of the network as reputational markers (Rogers, 2013). In this
analysis, we have thus combined the manual classification of reliability undertaken by Le Monde’s
journalists with the standing of a source according to the hyperlinks that it receives and gives.
This approach enabled us to bring fresh findings to current debates around fake news. In spite of
the proliferation of fabricated content of various shades, reputation hierarchies on the web seem
to be maintained (at least to some extent), as fake and hyper-partisan sites deploy aspirational
hyperlinking styles that favor, perhaps surprisingly, authoritative sources.
Further reading
Interested readers can find complementary discussions of the use of networks in journalism by
the same authors in “How to Tell Stories with Networks: Exploring the Narrative Affordances
of Graphs with the Iliad” (2016, in M. T. Schäfer & K. van Es eds., Datafied Society. Amsterdam:
University Press) and “Narrating Networks: Exploring the Affordances of Networks as Story-
telling Devices in Journalism” (2017, Digital Journalism, 5:6, pp. 699–730). A more advanced
discussion of the mathematics of graph visualization can be found in Freeman, L. C. (2009)
“Methods of Social Network Visualization.” In R. A. Myers (Ed.), Encyclopedia of Complexity and
Systems Science, Berlin: Springer and in Liotta, G. (Ed.). (2004) Graph Drawing 11th International
Symposium, Berlin: Springer.
References
Barabási, A. L. and Albert, R. (1999) “Emergence of scaling in random networks.” Science, 286(5439), 509.
Retrieved from www.sciencemag.org/cgi/content/abstract/sci;286/5439/509
282
Baruch, J. and Vaudano, M. (2016, April 8). “‘Panama papers’: un défi technique pour le journal-
isme de données.” Le Monde. Paris. Retrieved from http://data.blog.lemonde.fr/2016/04/08/
panama-papers-un-defi-technique-pour-le-journalisme-de-donnees
Behrens, J. T. and Chong-Ho, Y. (2003) “Exploratory data analysis.” In I. B. Weiner (ed.), Handbook of
Psychology. London: Wiley (pp. 33–64). doi:10.1002/0471264385.wei0202
Bounegru, L., Gray, J., Venturini, T. and Mauri, M. (2017) A Field Guide to Fake News. Retrieved from
fakenews.publicdatalab.org
Bounegru, L., Venturini, T., Gray, J. and Jacomy, M. (2016) “Narrating networks: Exploring the affordances
of networks as storytelling devices in journalism.” Digital Journalism, 5(6), 699–730.
Bowker, G. C. and Star, S. L. (1999) Sorting Things Out: Classification and Its Consequences (Inside Technology
S.). Cambridge, MA: The MIT Press.
Burt, R. S. (1995) Structural Holes: The Social Structure of Competition. Cambridge MA: Harvard University
Press. Retrieved from http://books.google.com/books?id=E6v0cVy8hVIC&pgis=1
Foucault, M. (1970) The Order of Things. New York, NY: Pantheon Books.
Freeman, L. C. (2000) “Visualizing social networks.” Journal of Social Structure, 1(1).
Früchterman, T. M. and Reingold, E. M. (1991) “Graph drawing by force-directed placement.” Software:
Practice and Experience, 21(November), 1129–1164. Retrieved from http://onlinelibrary.wiley.com/
doi/10.1002/spe.4380211102/abstract
Girard, P. (2011) “HyperText Corpus Initiative: How to help researchers sieving the web?” In Using Web
Archives (ed.), Out of the Box Conference. Den Haag, The Netherlands: IIPC.
Girard, P., Jacomy, M., Baneyx, A. and Venturini, T. (2012) “The Hypertext Corpus Initiative: methods and
tools for Social Sciences to build corpus from the web.” In Websites as Sources. Luxembourg, Luxembourg:
CECV.
Jacomy, M., Girard, P., Ooghe, B. and Venturini, T. (2016) “Hyphe, a curation-oriented approach to web
crawling for the social sciences.” International AAAI Conference on Web and Social Media (pp. 595–598).
Retrieved from https://hal.archives-ouvertes.fr/hal-01293078/
Jacomy, M., Venturini, T., Heymann, S. and Bastian, M. (2014) “ForceAtlas2, a continuous graph layout
algorithm for handy network visualization designed for the gephi software.” PLoS ONE, 9(6), e98679.
doi:10.1371/journal.pone.0098679
Jensen, P., Morini, M., Karsai, M., Venturini, T., Vespignani, A., Jacomy, M., . . . Fleury, E. (2015) “Detecting
global bridges in networks.” Journal of Complex Networks, cnv022. doi:10.1093/comnet/cnv022
Lynch, K. (1960) The Image of the City. Cambridge, MA: The MIT Press. Retrieved from http://books.
google.com/books?hl=it&lr=&id=_phRPWsSpAgC&pgis=1
Marres, N. and Rogers, R. (2005) “Recipe for tracing the fate of issues and their publics on the web.” In B.
Latour and P. Weibel (eds.), Making Things Public: Atmospheres of Democracy. Cambridge, MA: The MIT
Press (pp. 922–935).
Moreno, J. (1953) Who Shall Survive? (2nd ed.). New York, NY: Beacon House Inc.
The New York Times. (1933, April 3) “Emotions mapped by New Geography.” The New York Times.
Noack, A. (2009) “Modularity clustering is force-directed layout.” Physical Review E, 79(2). doi:10.1103/
PhysRevE.79.026102
Rogers, R. (2013) Digital Methods. Cambridge, MA: The MIT Press.
Tukey, J. W. (1977) Exploratory Data Analysis. Reading, MA: Addison-Wesley.
Turnbull, D. (2000) Masons, Tricksters and Cartographers. London: Routledge.
Venturini, T., Jacomy, M. and Jensen, P. (forthcoming) “What do we see, when we look at networks.
towards a positive measure of spatialisation quality for force-driven network Layouts.” Forthcoming.
283
21
DATA JOURNALISM AS A
PLATFORM
Architecture, agents, protocols
Eddy Borges-Rey
Over the last decade, British journalism has come under increased public scrutiny. First, the
phone hacking scandal (2011) and the subsequent Leveson report (2011–2012) evidenced the
inability of a number of tabloid publications to act ethically and in the public interest. Then
the expert-layman relationship, which granted journalism its authority, suffered severe damage
when experts were discredited as a knowledgeable source of information by populist politicians
in 2016. Finally, a widespread dissemination of misinformation exacerbated by social media algo-
rithms took media distrust to unprecedented levels by the end of 2016.
During this period, while journalism struggled with its image crisis, computerized data gained
prominence as “the new oil of the Digital Economy” (Yonego, 2014), and public and private
actors entered a race to control and nourish this commodity. Data were also perceived as an
opportunity by a small number of news organizations, who invested in data journalism (and its
methodological rigor) to spearhead their efforts in positioning journalism as society’s watchdog
once again.
Although a number of data/computational journalism scholars have conceptualized data
journalism in more sophisticated ways (cf. Hamilton and Turner, 2009: 2; Lewis and Westlund,
2015b: 449; see Coddington, this volume, Chapter 17), at this stage I will define data journal-
ism as the type of news reporting through which journalists engage with computerized data to
inform their publics. Research on the field (Appelgren and Nygren, 2014: 40; De Maeyer et al.,
2015: 441; Fink and Anderson, 2015: 475; Borges-Rey, 2016; Borges-Rey, 2017) has, nevertheless,
shown that data journalism represents, to date, only a small proportion of the overall coverage by
the mainstream news media, hence challenging early expectations that a methodical reporting of
figures and numbers would stimulate public trust again.
Under such unstable conditions the implications of Big Data for journalism remain to be
seen. In addition, the closure and amalgamation of a series of local news outlets has not only
exposed a worsening democratic deficit but has also shifted news managers’ attention to another
dimension of data: automation. While editorial boards are presented with an ever more con-
strained set of resources, automation is set to play an essential role in the professional newsroom
of the future (Van Dalen, 2012; Clerwall, 2014; Latar, 2015; Carlson, 2015; Broussard, 2015),
releasing journalists from time-consuming activities such as identifying potential tips or leads,
determining whether a story is newsworthy enough, or making decisions when it comes to
ethical dilemmas. While professional journalism is still confronting the complexities of data
284
Data journalism as a platform
journalism practice, scholars have made a start on what some regard as the second wave of data
journalism research.
As a way of consolidating my contribution to this field over the last five years, this chapter
examines the tensions and convergences that accompany the integration of data processes into
professional newswork and their impact on traditional organizational news structures and cul-
tures. It seeks to examine, in detail, the role of British contemporary data journalism to explore
journalism in an increasingly data-driven society. As scholars continue providing a plethora
of diverse and often competing conceptualizations of data journalism, this chapter focuses on
two main epistemologies of data journalism practice (see Borges-Rey, 2017): a dominant one,
whereby data are handled by journalists following axiomatic norms of news reporting; and an
emerging one, whereby journalists use the methods and logics of data scientists to interact with
data in innovative ways.
Drawing on a set of semi-structured interviews with data journalists, data editors, developers,
graphic designers, and news managers based at the Guardian, the BBC, the Financial Times, Chan-
nel 4, Trinity Mirror Group, the Times, CNN, Thomson Reuters, the Daily Telegraph, STV, the
Scotsman, the Herald, the Detail, Wales Online, Irish News, and the Belfast News Letter, I use Mont-
fort and Bogost’s (2009) platform studies methodology to examine data journalism as a platform
that exists within an architecture (virtual and non-virtual ecosystems) that deals with agents
(journalists, data, automated entities) through a series of protocols (news cultures, computational
logics/cognitions, power relations/struggles).
Platform studies for non-computational ecosystems

Since Montfort and Bogost proposed platform studies as a methodology and attempted to con-
solidate it through the launch of an MIT book series back in 2009, their intention was to
“promote the investigation of underlying computing systems and how they enable, constrain,
shape, and support the creative work that is done on them” (Montfort and Bogost, 2009: vii). In
this sense, as noted by Apperley and Parikka, scholars interested in undertaking platform stud-
ies were asked to focus on three dimensions: (a) “A single platform or a closely related family
of platforms”; (b) “Technical rigor and in-depth investigation of how computing technologies
work”; and (c) “An awareness of and discussion of how computing platforms exist in a context
of culture and society” (2015: 2).
With a manifest emphasis on computing technologies, Montfort and Bogost proposed a five-
layer model that analyzes platform, code, form/function, interface, and reception/operation.
Using the computational system as a means to interpret larger cultural frameworks, they remark
that: “A computational platform is not an alien machine, but a cultural artifact that is shaped by
values and forces and which expresses views about the world” (Montfort and Bogost, 2009: 148).
In this section, I will argue that this perspective of platform studies is rather limited, and, in order
to achieve its full potential, it should span beyond the technical specificity of the computational
architecture. Similar criticism is voiced by Apperley and Parikka, who claim:
“Platforms are platforms not necessarily because they allow code to be written or run, but
because they afford an opportunity to communicate, interact or sell” (Gillespie, 2010:
351, emphasis added). Platforms, then, are not just technologies but techniques that
sustain interactions as well as offer an epistemological framework.
(Apperley and Parikka, 2015: 5)
Therefore, I propose an approach whereby architecture transcends the boundaries of the hard-
ware and reaches both the virtual and material datafied ecosystems that shape our modern society.
285
Eddy Borges-Rey
Likewise, software, i.e., a series of agents governed by a protocol to fulfill a purpose in the system,
transcends the instructions, parameters, and variables set by the code to reach the power struggle
or negotiation that takes place between data journalists and data entities during the production
of news.
Given that the methodology of platform studies has been deliberately left open (Apperley
and Parikka, 2015), it offers the flexibility to go beyond the tensions between computational and
journalistic logics, recorded by extant research on the field, to instead approach data journalism
as a technoculture that engages with a variety of materialities, performativities, and reflexivities,
placing, “the platform at the center of a materially grounded discourse” (Apperley and Parikka,
2015: 5–6). Consequently, I will shift the emphasis from data journalists at the epicenter of the
analysis to data journalists coexisting with computerized agents in a larger ecosystem.
Interface: the illusion of control

In the introduction, I defined data journalism as a type of news reporting through which journal-
ists engage with computerized data to inform their publics. The intention to place the emphasis
on the interaction between data journalists and data was deliberate. The idea of data as an imma-
terial commodity used by journalists is both limited and constraining, in my view. This is why
perhaps the most effective way to outline the contours of data journalism as a platform is by
recognizing data as an active material entity that affects the agency of data journalists. As analyti-
cal frameworks tend to depict interactivity as a human enabler, they dismiss interactivity’s mate-
rial facet; that is, its capacity to enable material entities to interact as well. During my five years
of data journalism research, I have realized that Miller’s (2005: 3) distinction between a “vulgar
theory of mere things as artefacts” and “a theory that claims to entirely transcend the dualism of
subjects and objects” is essential to expand beyond the reductionist conceptualization of data as
objective evidence. This way of approaching data has been prevalent in data journalism practice
in the United Kingdom and elsewhere. In this regard, data are commonly assumed by data jour-
nalists to be one of the many ingredients of the recipe we call news, hence controllable, inactive,
and subordinate to human agency. Data tend to be handled as an inanimate good that functions
as either evidential input for stories or, to a minor extent, as the fuel that powers data-driven
outputs. Once data are found within the convoluted entanglement of datasets held by public or
private institutions, they are claimed by data journalists and subsequently placed in the formulaic
structure of the news item.
Yet if we understand data as the vital fluid of societal institutions and data-driven automa-
tions as “bodies without organs” (Deleuze and Guattari, 1987: 4), data transcend the boundaries
imposed by the limited mind-set of data as evidence and become “something more than ‘mere’
matter: an excess, force, vitality, relationality, or difference that renders matter active, self-creative,
productive, unpredictable” (Coole and Frost, 2010: 9). Virtually every sector of society currently
deals with both media and computing software on a regular basis, generating vast amounts of
data. A rising institutional dependence on database architecture and algorithmic automation
ensures that society’s machinery is well oiled and working smoothly. Therefore, I argue that
in order to understand how profoundly data are embedded in the structures of public bodies,
governments, corporations, and civic society as a whole, data journalists must actively negotiate
shares of performative power with data brokers and their datafied entities, as control shifts from
traditional spheres to organizations that nourish the wealth of data. In this sense, long-standing
attitudes toward discovering patterns and correlations in the data to unearth novel insights or
corroborate hypotheses should be complementary to an understanding of notions of media
numerical representation, modularity, automation, variability, and transcoding (Manovich, 2001)
that are essential for recognizing how effectively data governs modern social life.
286
The tension between these two epistemologies can be seen at a fundamental level when data
journalists acknowledge that data enable them to generate exclusives in a process whereby data
journalists, while recognizing data’s capacity to make their stories more accurate and rigorous,
simultaneously are capable of challenging the quality of that data by looking for flaws within
datasets. At a more advanced level, this power negotiation can be seen in contemporary news-
rooms that depend on automations, such as web metrics, to inform editorial decision making;
or scrapers, to gather and analyze large-scale datasets. These custom-made entities are normally
found only in a limited number of elite news organizations that can afford such technology,
or the professionals that model them. Smaller, local news organizations must normally rely on
generic third-party solutions, which, though freely accessible through open-source regimes, are
not always compatible with the existing architecture of these organizations’ newsrooms, thus
hindering the aptness of the algorithmic organism for the ecosystem in which it performs. Here,
the access to the insight obtained through automated agents dictates a different power dynamic
between those with access to the means to harvest data-driven insights and those who are unable
to access this wealth of information because of financial constraints. An inability to program
these automations in-house makes local news outlets dependent on collaborations with program-
mers or a few generic software solutions that are limited by the universality of their design.
This distinction between data journalists who can model the behavior of automated agents
(for example, a scraper) through the design of their systems and those who have to rely on generic
third-party software (say, Google Sheets) is key to understanding the importance of the interface
in the power struggle between data journalists and automated systems. For instance, the transfer-
ence of power from the data journalists to the automated agent becomes evident when the data
journalist has to navigate, as suggested by Lister et al. (2009: 21–22), an abundance of finely tuned
options offered by the software interface to maximize their perceived interaction and freedom of
choice, thus increasing their apparent capacity to negotiate with the automated agent and even
regain certain shares of performative power. Although the limitations of third-party software are
sometimes valued by data journalists, as they prompt creative ways to circumvent these limita-
tions through the use of other applications, it is ultimately the interface that governs the rules and
levels of interaction of reporters with no coding knowledge, further evidence of how data affects
the performativity of these professionals.
Understanding interactivity and interface through the prism of the computerized system can
help researchers and journalists to detach themselves from the prevailing journalistic humanness
that undermines data journalism’s computational possibilities. Since the interface does not distin-
guish between digital objects and users, human agency is parametrized as another variable gov-
erned by system protocols: another command line in the code. While automated systems pervade
organizational and professional cultures and logics, so do their interpretations of the protocols
that govern the processes and professional practices they engage with. If the data journalists that
form part of that equation wish to engage with automated agents to negotiate shares of power,
they need to abandon the illusion of control and come to terms with data-driven bodies as equals
to understand the ways they perceive and interpret the ecosystem. A broader understanding of
data as the vital fluid of society’s virtual and material infrastructures has equipped a small number
of data journalists with a set of unique performativities that enable them to bend the rules of the
interface and extract information from sources unavailable to data journalists anchored in the
belief that data are only evidence. I will expand on this later in the chapter.
Form: the rules of the game

Revisiting my introductory definition of data journalism, I have characterized the practice as a
type of reporting. With this assertion, I intend to oppose enthusiastic expectations, such as those
287
Eddy Borges-Rey
of Tim Berners-Lee, the founder of the World Wide Web, which position data journalism as the
future of journalism (Gray et al., 2012). In this section, then, I map out the form that contem-
porary data journalism practice takes in the United Kingdom and how it functions, taking as a
point of departure the source of data and how data are gathered.
If data journalism follows a set of rules, this protocol, in the first instance, is strongly mediated
by the rigid ethos of the Freedom of Information Act (FOI Act) and the Open Data movement.
Although alternative ways of obtaining data, such as subscription to data feeds, data licenses,
leaks, or in-house collation were mentioned, most of the data journalists I have interviewed in
recent years agree that, by and large, the data they use originated within the public sector and
were obtained by means of FOI requests. The act, which has been adopted with varying degrees
of success in local, regional, and national administrations in the United Kingdom, has suffered
from limited resources and infrastructure; differing interpretations of the legislation exhibited by
civil servants; and a general public disenfranchisement with the scheme. These limitations have
not only undermined the potential for openness of the scheme but also have compelled journal-
ists to break with their overreliance on public data and to produce their stories with the help of
web-scraping or collaboration with third-sector or private organizations willing to grant access
to their databases. While this transition takes place, I will argue that open data largely infuse the
material character of the datasets used as evidence by these professionals, at present.
In light of this tension, data journalism practice tends to be mediated by ideals of alleged open-
ness and transparency while it is simultaneously subject to the bureaucracy and politics of public
institutions. Data generated within the public domain tend to be perceived by data journalists
as influenced by the agendas of politicians, press officers, and other spin doctors. They are also
perceived to be restricted by the bureaucratic mechanisms of public bodies, which in some cases,
after prolonged negotiation, provided the data in a non-machine-readable format, making it dif-
ficult to use. FOI-driven data also provide a distinctive flavor to the reports of data journalists that
is generally restricted to topics limited to census, public governance, health, education, or crime,
affecting not only the style of reporting but also the scope of the story.
Despite the growing expansion of open data, many data journalists have recognized that data
released through this scheme is too politicized to be used effectively for journalistic purposes and
see the data scraped from websites or obtained through informants or similar conventional meth-
ods to be more appropriate. Paradoxically, access to alternative sources of data or, for instance,
corporate data held by data brokers does not depend on the advanced computational skills of data
journalists but on the will of insiders, whistle-blowers, or leaks and similar traditional means, as it
is proprietary data protected by copyright laws.
Once data are collated, data journalists sort, clean, visualize, and make sense of the data they have
gathered in order to tell a story. Early predictions suggested that data journalism would engage audi-
ences with innovative forms of storytelling, and the familiar linear storytelling that has character-
ized journalism over more than 100 years would be disrupted, to some extent, by more interactive
and engaging forms of informational user experience with multi-layered, multiplatform, gami-
fied, database-linked dynamic content. This type of presentation requires users to input personal
data and, after performing real-time data analysis, tells them a bespoke story about themselves.
This informational experience therefore appears to be heavily mediated by the ontologies of user-
interface design, user-experience design, and human-computer interaction, not only signaling the
pervasiveness of computational thinking in data journalists’ reflexivity but also the potential of data
journalism to manage audiences’ informational expectations to a granular level, offering finely tai-
lored personal stories that challenge benchmarks such as the public interest and news values.
The tension between FOI stories and non-FOI stories, linear storytelling and informational
user experience, among other factors, has led to the diversification of data journalism practice
into three distinct forms:
288
a. A daily, quick-turnaround, generally visualized, brief form of data journalism;

b. An extensive, thoroughly researched, investigative form of data journalism; and
c. A light, editorialized, entertaining, often humorous, gamified form of data journalism.
This classification not only refutes common beliefs that data journalism stories are presented by
means of visualizations but also highlights the flexibility of the format to accommodate opinion-
ated and editorialized forms of popular data journalism. For instance, the Trinity Mirror Group’s
discontinued data journalism experiment Ampp3d glamorized numbers, popularized them, and
made them appealing. It used the most popular newsbeats within the publication to sell, to a
very specific segment of the public, the whole data journalism venture. It used humor and satire
to act as a liberal champion that often criticized the conservative government, thus protecting
the disadvantaged and underrepresented. Although it performed a paradoxical form of data
journalism – one that blended op-ed journalism with scientific rigor – it also showed its audi-
ences, in very applied terms, how to correlate trends, perform data extrapolation, understand the
extremely necessary context of official stats and figures and issues of sampling and population,
and make sense of ordinary numbers that represent issues affecting them directly. It performed
the necessary function of humanizing numbers by highlighting their social relevance. It also
showed how numbers are misused and misrepresented by media and politicians alike. The story
“Hey DWP! We Just Fixed Your Chart for You” (Warnes, 2014) serves as an example of Ampp3d’s
fact-checking ethos. The original tweet released by the Department for Work and Pensions of
the United Kingdom tried to illustrate how the benefit cap helped thousands move into work.
However, the figures shown contradicted the proportions of the bars on the chart. By just chang-
ing the graphic proportion of the bars on the infographic, Ampp3d’s illustration correlates more
accurately with the numeric proportions conveyed by the actual statistics. Through this type of
accessible explanations, Ampp3d attempted to show its public how the government could spin
numbers to disseminate an optimistic, yet often inaccurate, depiction of the administration. This
example also illustrates the ways through which Ampp3d expanded its storytelling capabilities
beyond the confines of the data story to reach the social media conversation triggered by it.
Ampp3d thus became an innovative player within the British data journalism landscape. It
created a versatile transmediality that enabled its journalists to strategize (a) data infrastruc-
tures and the stories potentially concealed within them; (b) the tensions that arise between
two conflicting forces – popular liberal satirical tabloid journalism and the surgical precision
of numeric facts; (c) the efforts to humanize statistics and highlight its mundane facet; (d) the
development of a distinctive digital persona intermediated by the charisma of the newshound;
and (e) the capacity to perform the data journalist’s function beyond the limits of the platform
or the publication.
In sum, the specificity, the specialized knowledge base, and the disruptive nature of many of its
features suggest that data journalism seemingly works better as an alternative methodology or phi-
losophy that adapts to journalistic themes or beats – be that sports or investigative journalism – to
provide both a robust backbone to stories and tools to efficiently make use of web-based knowl-
edge infrastructures. This will be explored at a performative and reflexive level in the next section.
Code: what makes data journalism tick?

This level – code – is potentially the one with the most challenges, as it refers to the compu-
tational code of the platform. Focusing on the comments, variable names, choices made when
writing programs, and the discourses attached to them, here I explore the epistemological, political,
and discursive codes that shape the performativity and reflexivity of data journalists rather than the
computational. In order to achieve this, I have proposed a taxonomy of performative codes that
289
Eddy Borges-Rey
data journalists use to legitimize their knowledge claims. As data journalists model their journal-
istic identity, these codes outline the contours of their working practice. But, more importantly,
they act as the definers used in the constant power struggle in which data journalists engage with
traditional reporters to find their place in their newsrooms and journalism at large.
There is no doubt that the idiosyncratic aspects discussed earlier pervade the performativity
and reflexivity of data journalists, mediating how they perceive themselves and their working
procedures. In this sense, I found data journalists who felt comfortable performing more elemen-
tary forms of data analysis to produce FOI-driven stories. Equally, I found a smaller proportion
that felt the need to utilize computer science methods to provide structure to data obtained
through less conventional methods, such as web-scraping, thus escaping the constraints of open
data regimes and the limited thematic flavor that the FOI Act scheme granted their stories.
Regardless of the computational complexity of each approach, to a greater or lesser extent these
performative codes were present in their newswork.
a Journalistic authority: in order to persuade their audiences of the veracity of their accounts
(Broersma, 2010), journalistic authority has endured as the imperative requirement to per-
form within the trade, notwithstanding the importance of data and computation. Data
journalists unanimously stated that the most significant professional attribute required to
legitimize their accounts was journalistic skills.
b Numeric infallibility: this provided quantifiable evidence to their stories, which was then rein-
forced by the rigor of statistical methods used during the news production process. At least
one-third of the data journalists that I interviewed acknowledged having a background
in a mathematics-related discipline. The remainder showed competence in basic statistical
operations, such as correlation, percentage change, standard deviation, large increases and
decreases, and trends over time. Data journalists were well aware of recurrent statistical errors
in news reporting and how to avoid them. They were also extremely cautious with respect
to their calculations, corroborating their approaches with statisticians within public bodies
as to the types of equations and algorithms used to process data.
c Scientific rigor: in order to remain trustworthy in the public eye, soundness and transparency
in their methods was imperative.
d Computational neutrality: this refers to technologies that arguably circumvent human bias
while efficiently performing automated gathering, analysis, and presentation of unstruc-
tured information. Regardless of the complexity of the computing tasks performed – from
number crunching and pivot-table analysis in a spreadsheet to back-end programming of
computerized algorithms – involving computation throughout the news-reporting process
arguably improved the sense of flawlessness in their accounts.
e Crowdsourced cooperation: by providing audiences with easy access to raw data and filtering
tools to make sense of these data, data journalists attempted to democratize both the data-
bases that serve as inputs for stories and the news workflow that produces such stories.
f Extra- or intra-newsroom collaborations: the absence of certain advanced computational skills
and/or restricted access to certain information has compelled data journalists to embrace
open-source ideals and seek internal or external collaboration in their efforts to overcome
these limitations. As data journalism impregnates the news culture of professional news-
rooms, specialized correspondents seek to collaborate with data journalism units to pro-
vide soundness and robustness to their stories through data methods. Collaborative projects
where data journalists, developers, statisticians, and graphic designers interacted were deemed
by data journalists to be very effective, combining the expertise of various disciplines to
produce ground-breaking news experiences. Data journalists with rather limited technical
290
competences and infrastructure tended to collaborate with academics, public archivists,

records officers and/or statisticians, civic organizations and open-source initiatives (such as
Hacks/Hackers or scraperwiki), libraries, trusts or foundations, and third-sector organizations
or private companies that contributed to the project by opening up their databases for jour-
nalistic scrutiny.
g Hyperlocal empathy: this is the final and perhaps most interesting code of the taxonomy. As
data journalists engaged with hyperlocal data, they connected strongly with their commu-
nity and local issues. This code was shown across the board, as data journalists passionately
addressed issues that include the lack of scrutiny in council meetings and reports, gov-
ernmental data scarcity, power abuse, administrative bureaucracy, communal deprivation, or
civic disengagement. Local data also strengthened local or national identity, as I found in the
case of the devolved nations of the United Kingdom.
At a reflexive level, data journalists tend to advocate journalistic mind-sets over computational
ones. Notably, as web-scrapers and similar automations become regular aspects of British data
journalism, data gathered through these techniques are infused with the flexible philosophy of
computerized methods, fostering, as a result, problem-solving and creative ways of finding, com-
piling, and understanding unstructured informational data. The adoption of such methods also
provides a wider range of alternative sources of data, which allows data journalists to cover more
diverse topics, thus overcoming the topical saturation of open/FOI data. In addition, this wider
range of sources can help journalists expose corporate wrongdoing, placing private institutions
under similar degrees of scrutiny to those experienced by public power holders.
Hence, I would argue that the optimum reflexive and performative state of data journalism
practice emerges when a nose for news and computational cognition blend to generate a hybrid
form of journalism. While a group of data journalists decides to move only within the con-
stricted architecture of public institutions to follow the limited protocols of FOI Act regimes,
another group exhibits a capacity to find stories in the back end of the largest piece of infor-
mational architecture of our times: the World Wide Web. These data journalists understand that
within this architecture, protocols are highly flexible and demand a different approach from
data journalists. Information is hidden in plain sight and is inaccessible to the majority of users,
not because it is protected in a black box but because the public is not data literate enough to
identify and untangle this information. This small number of data journalists reverse-engineer
black boxes, find exploits in the system, efficiently overcome encoding errors, see structure in
unstructured information, turn metadata into the variables of a spreadsheet, parse text, exhibit
an awareness of the ethical and legal boundaries around database fusion or web-scraping, publish
their analysis in Jupyter notebooks to enhance their accountability, seek advice in programmer
communities such as Stack Overflow, and organize hackathons to team up with activists, coders,
programmers, and ethical hackers to solve intricate data journalism problems.
Operation: the invisible user

Before proceeding to the last section of this chapter and discussing the platform level of data
journalism, I will review, one last time, the definition of data journalism that I provided in the
introduction. Bearing in mind that data journalism is a type of news reporting through which
journalists engage with computational data to inform their publics, this section will focus on the
last segment of the definition, the informational experience facilitated by data journalism, and
it will question whether members of the public significantly interact when they engage in this
experience.
291
Eddy Borges-Rey
When it comes to data, the relationship between data reporters and their publics appears to
be an active one. Traditionally, journalists have been entrusted by the public to retrieve, interpret,
and translate complex data that otherwise would remain cryptically inaccessible to them. This is
probably the reason why many of the senior and most experienced journalists tend to perceive
data (statistics) as an integral part of their trade.
I have already argued that data journalists have embraced a model of collaboration that enables
them to cooperate, together with other actors, with members of their public. This type of collab-
oration arguably makes their news production process more transparent, open, and participatory.
It is also believed that, through openness and transparency, their audiences are more aware of data
journalists’ methods of gathering, cleaning, and analyzing data, and, as a consequence, audiences
are exposed to elementary knowledge on data processing and sorting.
Notably, regardless of data journalists’ efforts to decode mathematical abstractions to make
them understandable for audiences, to create visualizations in order to present figures and num-
bers, and to share their methods and databases for public scrutiny, empirical evidence seems to
suggest that participation and engagement are not as widespread as was first anticipated.
For example, most of the data journalists interviewed who had attempted crowdsourcing
expressed that in most cases calls for members of the public to participate failed to attract num-
bers significant enough to deem the practice successful. These journalists attributed responsibility
for this lack of engagement with crowdsourcing to the common denominator of fragmented
local audiences who normally prefer to consume news from newspapers, radio, or TV and tend
to be disengaged from online news.
Further, numeracy, statistical literacy, or data literacy are seen as an exclusive skillset of math-
ematicians, statisticians, or social, computing, or data scientists, and a broader and more critical
understanding of data as a cultural artifact is limited for citizens overall. As long as data literacy
remains relatively low among members of the public, data journalists will report to a vacuum.
Therefore, despite efforts to translate complex data and enlighten the public, data are likely to
remain meaningless in the eye of the citizens, thus preventing data insights from powering civic
deliberation.
Let us revisit Ampp3d to illustrate the effects of public disengagement from data journalism.
During its two years of existence, Ampp3d was highly respected by data journalists at other news
outlets. This proven success, able to attract some 7 million unique browsers a month (Pons-
ford, 2014), was discontinued by Trinity Mirror executives in 2015 (Jackson and Sweney, 2015).
This was met by outcry and criticism from data journalists (Bradshaw, 2015). A Trinity Mirror
spokesperson expressed that Ampp3d belonged to a niche area with very high costs for the vol-
ume of traffic that it produced; therefore, in order to guarantee the sustainability of their digital
infrastructure, they would invest in an integrated digital team that focused on areas with more
mass-market appeal (Jackson and Sweney, 2015).
Unfortunately, cases such as this call into question the actual audience appeal of data journal-
ism and contest the effectiveness of its interactive, transparent, and open facets. It suggests that
despite the expectations of cooperation, which would bring data journalists and members of the
public together as companions in the quest for uncovering data insight and exposing corporate
and governmental wrongdoing, in reality citizens are more likely to only come into contact with
data journalists as digits in a database underpinning data-driven news pieces.
Platform: concluding remarks

Data is unquestionably one of the key areas for productive development, as governments and
industry alike face constant challenges in relation to the management of big data on the one
hand and exploring creative and innovative ways of engaging with data for the development
292
of society on the other. As society continues exploring the complexities of computerized data,
there is growing uncertainty around data literacy and its insertion in educational policy, in cur-
riculum design, and life-long learning. Research has shown that data literacy levels amongst data
journalists also vary. In this regard, there is a tension between two data literacies: a more instru-
mental one that relies on arithmetic and statistics and a more computational one that employs
problem-solving and programmatic thinking. Drawing on this initial distinction, this section
examines how data cultures, data awareness, data power, and critical perspectives of data, code, and
algorithms mediate the data literacy of both data journalists and their publics.
It is clear that a healthy and wide-ranging data literacy will remain unreachable as long as the
traditional approach of using numbers as hard facts and computers as tools for data processing
persists. Following the discussion in the previous section, I argue that data literacy is essential not
only for journalists to perform data journalism but, more importantly, for the public to make
sense of an increasingly datafied world, where rules are dictated by automated agents and their
protocols and power is mediated by data brokers.
In order to facilitate the transition from instrumental approaches to data to a more critical
approach to data, it is essential that citizens are able to distinguish the main data holders and data
gatherers – those responsible for granting not only access to data but also to the technologies that
process it. At present, data is mediated and processed by different types of institutions with different
kinds of agendas. Within public sector organizations, formal gatekeepers have significant power,
able to grant or deny access to information (such as labeling a request vexatious). Furthermore, the
degree of citizen access to open data held by governments is small in certain contexts (García and
Gertrudix, 2011). For all these reasons, it is paramount to understand the gatekeepers who perform
this mediation, the institutions within which they function, the principles and policies that guide
their work, and the impact of these processes and interactions on citizen empowerment.
Making use of data, therefore, not only requires ever increasing rates of data literacy but also
granting citizens access to data to facilitate their interaction with data infrastructures. As individ-
uals actively engage with data, data pervade their data awareness and enhance their computational
cognition. Factors such as (a) emergent forms of machine-mediated communication with smart
assistants such as Siri and Alexa; (b) the prevalence of wearable tracking systems; (c) the implanta-
tion of sensors in our everyday activities, driven by a disproportionate emphasis on ‘Internet of
Things’ technologies; and (d) the fast and steady development of cognitive computing not only
disrupt the ways we perceive these pervasive technologies but also the ways we interact with these
technologies as everyday life companions.
For this reason, I tend to make no distinction between the reflexivity of both data professionals
and the intervening automations involved in the production of data outputs, because in my view
they display similar mind (and system) processes and similar biases and perform decision making
following similar patterns and stages. We have senses and a brain; they have sensors and cognitive
computing. In fact, by approaching data as a material force similar to that of evolution, we can
acknowledge that data and algorithms are strong determinants of our destiny as human beings.
For instance, web metrics have become a powerful aid for news editors to decide what becomes
news. Algorithms measure which news stories are most liked by the public, and that insight is
then used in editorial decision making to create the news agenda. It is clear then that matters of
public concern, discussed every day, are nowadays strongly mediated by automated technologies
unknown to the public. Therefore, it is important to understand the instrumental dynamics of
data; namely the statistical operations that govern algorithmic decision making or the irrelevance
of sampling for big data processing. However, by acknowledging data materiality beyond its
object status, we can better understand the role of embryonic forms of AI in the production of
knowledge, the negotiation of societal power, the parameterization of the world and social life,
the potential remediation of humans, and algorithmic accountability.
293
Eddy Borges-Rey
As the effectiveness of data organizations in monetizing the insights derived from citizens’ data
increases, so does the power they hold over not only the individuals but also over the institutions
of society. Corporations such as Google and Facebook, with a core focus on quantifying everyday
life, have programmed algorithms capable of profiling and predicting people’s hopes and dreams
in an environment free of public or institutional scrutiny. In the past, this watchdog function was
performed by news media as part of a healthy democratic society. Nowadays, news organiza-
tions seem to be unable to monitor the contemporary institutional negotiation of data power, as
it arises in a scenario only accessible to actors with an understanding of the computational and
legal loopholes in the system.
During my research, it was clear that data journalists were fully aware of the power dynamics
driven by emergent data brokers. However, most of them felt that they could only contribute
to uncovering wrongdoing within these domains if they were collaborating with beat/specialist
journalists, those traditionally commissioned to cover these areas: business, technology, and sci-
ence. Notably, beat reporters were perceived by most of the data journalists I interviewed as the
most suitable professionals to investigate the behavior of news subjects, sources, and news events
within the confines of data power arenas.
To conclude, I argue that there is an imbalance in the rhetoric, wherein big data is predomi-
nantly framed as the epicenter of contemporary innovation and the driving force of societal
progress. This prevalent and tacitly neoliberal discourse appears to safeguard corporate data com-
modities to the detriment of overly exploited public data, making the data generated within
public bodies and by citizens a profitable good to nourish. This dynamic emphasizes a major gap
between the data brokers who, through a series of furtive transactions, exploit and commoditize
citizens’ data, and citizens who display varying degrees of awareness regarding the amount of data
they generate on social media or that is generated by public bodies, and how that data is utilized
by increasingly powerful data brokers. Notably, a number of non-legacy entities and civic-driven
initiatives operating in platforms such as GitHub or Hacks/Hackers (see Usher, this volume,
Chapter 26) and start-ups such as scraperwiki are now essential spaces and proficient enablers of
data-driven public deliberation and civic engagement, thus fostering an ever-expanding rate of
data literacy and improving public participation. Platform studies offers a novel perspective on
the study of data journalism that sheds light on a series of underlying dynamics commonly unno-
ticed by scholarship on the field. It has the potential to make researchers aware of those entities
that play a vital role in data journalism practice but are generally disregarded by human-centric
approaches, with their emphasis on the attitudes and behaviors of data reporters.
Further reading
In order to further explore some of the ideas discussed in this chapter, it is probably a good idea
to review some of the titles included in Nick Montfort and Ian Bogost’s MIT Press book series
Platform Studies, as well as Thomas Apperley and Jussi Parikka’s (2015) epistemic examination
of the Platform Studies Methodology contained in their article “Platform Studies’ Epistemic
Threshold”. Two seminal journal special editions that examine the relationship between journal-
ism practice and data materialities are Seth C. Lewis and Oscar Westlund’s “Journalism in an
Era of Big Data: Cases, Concepts, and Critiques” (2015a) and CW Anderson and Juliette De
Maeyer’s “Objects of Journalism and the News” (2015).
References
Anderson, C. W., and De Maeyer, J. (2015) “Objects of journalism and the news.” Journalism, 16(1), 3–9.
Appelgren, E. and Nygren, G. (2014) “Data journalism in Sweden: Introducing new methods and genres of
journalism into ‘old’ organizations.” Digital Journalism, 2(3), 394–405.
294
Apperley, T. and Parikka, J. (2015) “Platform studies’ epistemic threshold.” Games and Culture, 1, 22.
Borges-Rey, E. (2016) “Unravelling data journalism: A study of data journalism practice in British news-
rooms.” Journalism Practice, 10(7), 833–843.
Borges-Rey, E. (2017) “Towards an epistemology of data journalism in the devolved nations of the United
Kingdom: Changes and continuities in materiality, performativity and reflexivity.” Journalism [online first
ahead of print]. Retrieved from http://journals.sagepub.com/doi/abs/10.1177/1464884917693864
Bradshaw,P. (2015) “The legacy of Ampp3d,UsVsTh3m and Row Zed.”The Online Journalism Blog (UK). Retrieved
from https://onlinejournalismblog.com/2015/05/13/the-legacy-of-ampp3d-usvsth3m-and-row-zed/
Broersma, M. (2010) “Journalism as performative discourse. The importance of form and style in journal-
ism.” In V. Rupar (ed.), Journalism and Meaning-making: Reading the Newspaper. New York, NY: Hampton
Press (pp. 15–35).
Broussard, M. (2015) “Artificial intelligence for investigative reporting: Using an expert system to enhance
journalists’ ability to discover original public affairs stories.” Digital Journalism, 3(6), 814–831.
Clerwall, C. (2014) “Enter the robot journalist: Users’ perceptions of automated content.” Journalism Practice,
8(5), 519–531.
Coole, D. and Frost, S. (eds.). (2010) New Materialisms. Durham, NC: Duke University Press Books.
Deleuze, G. and Guattari, F. (1987) A Thousand Plateaus. Translated by Brian Massumi. Minneapolis, MN:
The University of Minnesota Press.
García, F. and Gertrudix, M. (2011) “Naturaleza y características de los servicios y los contenidos digitales
abiertos.” CIC Cuadernos de Información y Comunicación, 16, 115–138.
Gray, J., Chambers, L. and Bounegru, L. (2012) The Data Journalism Handbook: How Journalists Can Use Data
to Improve the News. Sebastopol, CA: O’Reilly Media.
Hamilton, J. T. and Turner, F. (2009, July) “Accountability through algorithm: Developing the field of
computational journalism.” Report From the Center for Advanced Study in the Behavioral Sciences, Summer
Workshop (pp. 27–41).
Jackson, J. and Sweney, M. (2015) “Trinity Mirror’s UsVsTh3m and Ampp3d thought to be facing
axe as jobs set to go.” Guardian (UK). Retrieved from www.theguardian.com/media/2015/may/13/
trinity-mirrors-usvsth3m-and-ampp3d-thought-to-be-facing-axe-as-jobs-set-to-go
Latar, N. L. (2015) “The robot journalist in the age of social physics: The end of human journalism?” In G.
Einav (ed.), The New World of Transitioned Media: The Economics of Information, Communication, and Enter-
tainment (The Impacts of Digital Technology in the 21st Century). Cham: Springer International Publishing
(pp. 65–80).
Lewis, S. C. and Westlund, O. (2015a) “Journalism in an era of big data: Cases, concepts, and critiques.”
Lewis, S. C. and Westlund, O. (2015b) “Big data and journalism: Epistemology, expertise, economics, and
ethics.” Digital Journalism, 3(3), 447–466.
Lister, M., Dovey, J., Giddings, S., Grant, I. and Kelly, K. (2009) New Media: A Critical Introduction. Abingdon:
Routledge.
Manovich, L. (2001) The Language of New Media. Cambridge, MA: The MIT Press.
Miller, D. (2005) Materiality. Durham, NC: Duke University Press.
Montfort, N. and Bogost, I. (2009) Racing the Beam: The Atari Video Computer System. Cambridge, MA: The
MIT Press.
Ponsford, D. (2014) “Mirror rewards Coles and Belam with new jobs in wake of UsVsTh3m and
Ampp3d successes.” Press Gazette (UK). Retrieved from www.pressgazette.co.uk/mirror-rewards-coles-
and-belam-new-jobs-wake-usvsth3m-and-amppd3d-successes/
Van Dalen, A. (2012) “The algorithms behind the headlines: How machine-written news redefines the core
Warnes, S. (2014) “Hey DWP! We just fixed your chart for you.” Ampp3d (UK). Retrieved from www.mir
ror.co.uk/news/ampp3d/hey-dwp-just-fixed-your-4584836
Yonego, J. T. (2014) “Data is the new oil of the digital economy.” Wired. Retrieved from www.wired.
com/2014/07/data-new-oil-digitaleconomy/
295
22
SOCIAL MEDIA LIVESTREAMING
In tumultuous times, journalists turn to technology to tell stories as they unfold. With Facebook
Live and Periscope, reporters are taking viewers to the heart of the action, through their own
feeds and the livestreams of others. Going live was once a costly resource limited exclusively to
television news stations. Today, livestreaming on social media has opened up this visual storytell-
ing form to anyone with a smartphone and a Facebook or Twitter account.
What does social media livestreaming mean for journalism and the flow of information in
society? How are journalists adapting to this form of reporting and storytelling? What is the role
of the citizen livestreamer in this mix? And how do forces such as technological determinism and
commercial interests drive its adoption and shape its use?
This chapter addresses these questions by exploring livestreams on Periscope and Facebook
Live, interviewing journalists who have adopted livestreaming technology, and reviewing relevant
research and milestones in live broadcasting.
Facebook Live for everyone1

Wearing his signature T-shirt and an impish grin, Facebook’s Mark Zuckerberg leans against a
computer desk in the “live video launch room”. Flanked by screens and two dozen eager faces,
he looks directly into the camera and utters his inaugural word on Facebook Live: “Everyone”
(Zuckerberg, 2016). Intentional or not, unveiling this technology to everyone can be viewed as
emblematic of its nature and scope, unleashing livestreaming to nearly 2 billion Facebook users –
from protester to politician, reporter to rock star.
For journalists, it is a valuable tool with the promise of enhanced storytelling. But adding the
layer of user-generated content opens up a Pandora’s box of behaviors and effects, along with
ethical and legal issues. While streams from the public can take viewers to exciting events and
offer social witnessing, they can also be gruesome and violent. But Zuckerberg shows nothing
but ebullient optimism in his Facebook Live launch. “We’re psyched”, he says, as he introduces
the “core team” while the camera pans the room to smiles, waves and a few “whooos”. After he
chides, “That was pretty subdued”, they burst into cheers and applause.
“The reason why we’re so excited about this is that it’s this new, really raw, personal, sponta-
neous way that people can share”, he says. More than just posting a video, Facebook Live offers
“new kinds of interactive and social experiences” (2016). Zuckerberg calls it the beginning of a
296
Social media livestreaming
journey. And Facebook Live users have taken to its path with vigor. After just 12 months, one
in five Facebook videos was a live broadcast, with the average watch time quadrupling (Simo,
2017). In its first year, those “raw” and “personal” livestreamed moments have ranged from the
joy and silliness of “Chewbacca mom” (Lui, 2016) to the horror of a young man killing his baby
daughter (Mozur, 2017).
Enter the “new golden age of video online” (Zuckerberg, 2016).
Early adopters: journalists on Periscope

Facebook Live came a full year after Twitter’s Periscope, which launched in March 2015 (Weil,
2015). In its first 10 days, 1 million users joined Periscope (Rodriguez, 2015), among them jour-
nalists who soon incorporated the tool into their reporting. Paul Lewis of the Guardian was one
of its earliest adopters, using the app in April 2015 to livestream perhaps the first interview with
a presidential candidate on Periscope (Lewis, 2015a). Rand Paul abruptly ended their exchange
by walking off camera as Lewis tried to ask about polling numbers. “It reflected somewhat badly
on him”, said Lewis (2017). The lights went out as Lewis wrapped up the livestream: “He wasn’t
treating it in the way he would a more conventional broadcast”.
Later that month, when violence erupted in Baltimore following the death of Freddie Gray,
Lewis ‘Periscoped’ live from the city’s streets. As he walked along, Lewis narrated what he saw
unfolding in his path. His senses guided his reporting, providing context to the visuals. One
broadcast opens with music blaring in the street. “The Way You Make Me Feel”, is playing on a
loudspeaker, and Lewis names the singer as Michael Jackson. Bottles are smashing nearby as he
says, “It’s definitely not (pause) feeling very safe here right now”. But he stays calm and holds
steady during what sounds like a shot nearby (Lewis, 2015b).
Lewis talks with people he encounters along the way, and at one point, things get tense. As
he begins an interview, his subject looks decidedly nervous. “Oh, oh, oh, oh”, the man says as
he pivots to protect two young girls in pink coats standing behind him. “Bring the girls back
this way, bring the girls . . . come, come, come”, you can hear Lewis as they turn to run, and the
camera cuts out (Lewis, 2015b).
In another broadcast, amidst tear gas and sounds of bottles smashing, he senses it’s time to stop
livestreaming, and says, “I’m going to check out, this is getting really dangerous” (Lewis, 2015c).
Despite the hazards, Lewis says there’s an advantage to working alone without bulky equipment
and a crew. “You can kind of melt into the crowd” (2017). It allows you to give people an “unvar-
nished perspective on what’s happening”. This contrasts with the Baltimore crowd’s reaction to Fox
News correspondent Geraldo Rivera as he prepared to go live from the scene. One of the protesters
confronted Rivera for hyping the violence and minimizing Freddie Gray’s death (Bishop, 2015).
Core training and experience give journalists the foundation for livestreaming during a vola-
tile situation. “You’re applying the same journalistic principles to the new technology”, said
Lewis. As always, “the stories are about people” (2017).
But what’s not part of that training is the rapidly swelling number of followers and a constant
stream of unfiltered on-screen comments. Lewis said the “really inappropriate comments” were
his biggest challenge. “You have people being extremely racist, using the fact that you have this
audience to propagate fake news or false rumors. The bubbles of text were disappointing and
alarming” (Lewis, 2017). He responded in his narration by calling out what he saw as inappropri-
ate and warned viewers he would turn off the comments.
After Baltimore, Lewis relocated to San Francisco to become the Guardian’s west coast bureau
chief. He also moved away from Periscope: “I have not used it once since it happened”. Having
amassed more than 30,000 followers on the platform, he explains, “You need to be mindful of
297
what they may or may not want to see”. And for Lewis, few stories have been “Periscope wor-
thy” after the violence in Baltimore (2017).
But other journalists soon turned to the tool to cover a volatile situation emerging in Europe
and the Middle East. After the body of a Syrian toddler washed ashore on a Turkish beach in
September 2015, a burgeoning refugee crisis was becoming evident. As thousands fled Syria and
other Middle Eastern countries, journalists walked alongside the migrants as they attempted to
cross international borders. With their portable, relatively unobtrusive smartphones, correspondents
could show the struggle with video and, when feasible, livestreams via Periscope. Among them was
Paul Ronzheimer of the German newspaper Bild. As he traveled across Europe with a group of
Syrian refugees – from the Greek Island of Kos to Germany – he used his iPhone to Periscope inter-
views and key moments throughout the journey (Dredge, 2015). He later created a documentary
from the footage that he saved to his phone, which is posted on Bild’s website (2017). Al Jazeera
English (2015) livestreamed Q&As on Periscope with correspondent Mohammed Jamjoom. And
photographer Patrick Witty (2015) Periscoped refugees landing in Lesbos, Greece, for BuzzFeed
News. International video correspondent Adam Ellick (2015) of the New York Times finds the
migrant story to be a “very smart example” for livestreaming because of its urgent and trending
nature and its appeal to large audiences through personal, character-driven narratives.
New interactive storyforms

It’s the day after the 2017 presidential inauguration, and women around the world are marching
to protest against Donald Trump’s election. Nearly 26,000 people watch Gloria Steinem on the
New York Times’ Facebook Live broadcast as she revs up the crowd at the rally in Washington,
DC: “And remember the Constitution does not begin with ‘I, the president’. It begins with . . .”
and the crowd shouts, “We, the people!” But the Times is doing more than a one-way broadcast
from the stage on the National Mall. Its journalists are using livestreaming to interact with people
on the scene and with viewers worldwide. At 12:30 p.m., @nytimes tweets, “We’re live from
Women’s Marches around the U.S. Have questions? Ask our reporters here. nyti.ms/2ijLQUZO #
WomensMarch”.
The link takes you to a page on the Times’ website with the embedded livestream, where
video journalist Deborah Acosta is interviewing a young woman from Seattle wearing a hot pink
‘pussyhat’, wide-rimmed glasses, and a lip piercing. They’re on the perimeter of the D.C. rally,
where Acosta can access internet service without losing the signal. She asks, “What prompted
you to come all the way from Seattle to this march today in Washington, D.C.?”
“I’m a queer woman”, the Seattleite responds. “Trump getting elected was one of the most
terrifying things that’s happened in my life” (The New York Times, 2017).
Back in the newsroom, Louise Story manages the Times’ live interactive operation. She’s the
reporter “face” on the Times’ website, where her photo, name, and title – Reporter and Executive
Producer, Live Interactive Journalism – appear on the page. Story interacts with correspondents
in the field around the country by text, email, and cell phone. A team of four people works
in-house, in what Story likens to a scaled-back TV control room. But unlike TV news, the
journalists are also communicating with viewers around the world who post their questions on
the live comment section of the Times’ website and via Facebook. The team had been covering
the election for months, and is well-equipped to answer questions. “I think it’s a pretty smooth
experience”, said Story (2017).
One technique that helps reporters in the field is the “pinning” feature. The person in the
newsroom who is monitoring user comments can post a question at the top and keep it there
by pinning it. That way, when reporters on the ground glance at their phones, they can see the
question of interest right up top (Story, 2017). For example,
298
FROM A READER
How do these women feel about
the women who voted for Trump?
J. Kane Austin, TX
Story responds:
J. Kane, good question. Our reporter in NYC, Paul Moon, will ask that question to
the women there shortly. We are back in NYC now with the marchers there. We’re
asking them your questions, so ask away here and Paul Moon will ask some of them to
the marchers.
(The New York Times, 2017)
The video feed from New York City shows the crowd marching while Moon reads the ques-
tion. He then looks for someone to answer it while continuing to point his camera at the general
crowd. This technique affords privacy to people who may not want to appear in a close-up shot
or be interviewed on a livestream.
“Hey guys, how’s it going? I’m with the New York Times”, says Moon. “I’m doing a live video
right now. Do you mind if I ask you all some questions?” He has some takers, and only then does
Moon point the camera at the source. He hands his ear buds to a man in a black baseball cap (to
use as a microphone) and asks him to say and spell his name. You can barely hear Moon off mic
as he asks the viewer’s question. But before the man can answer, the woman he’s with jumps in:
I voted for Clinton, Hillary, and I’m very concerned and upset, and feel the women
need their rights, and I’m very worried for the future, and I’m happy to see everybody
out here marching and supporting women’s rights.
Moon continues talking with the two marchers and then speaks to the viewers, “All right, so
there you have an answer. Thank you for the question J. Kaine”, says Moon. He then reiterates
the exchange for people just joining in and references what he’s showing on screen, “the parade
is getting denser now”.
When he sees another question, he reads it, and then asks, “Emma, if you could pin that
question – OK, thanks”. He lets the camera roll, as hundreds chant, drum, and march by. He
keeps providing value to the livestream with informative bits – how the march has been set up
by alphabetical order of last names, and reading signs as they parade by. Unlike TV news, where
the reporter on the scene features prominently on camera, we can only hear Moon, not see him
(The New York Times, 2017).
Hybridizing the TV live shot model

Social media livestreaming appears to be hybridizing the TV news live shot, keeping some of its
characteristics, dropping others, and introducing new components. The metamorphosis started
well before Facebook Live or Periscope, going back to the early days of Occupy Wall Street, when
Tim Pool teamed up with Henry James Ferry to broadcast live on Ustream (Lenzner, 2014).
Ferry appeared in the shot as Pool ran the camera. Viewers who were tuning in from all over the
world were sending comments through a live chat function on the channel wearetheother99.
com. But they didn’t want or need to see Ferry on camera. Pool quipped in a published inter-
view “that just presents the same old same old. Might as well be on CNN, if you’re just going to
watch some guy talk” (Pool, in Lenzner, 2014: 253). To Pool, the broadcast model – with a tower
299
sending a unidirectional signal to the audience – is “archaic” (Pool, in Lenzner, 2014: 257). In his
streams, he says, “I’m taking direction, requests and corrections from people who are watching.
So there is a return from the community” (Pool, in Lenzner, 2014: 257).
As today’s social media livestreaming protocol develops, we see a mix of the traditional broad-
cast live shot and Pool’s interactive livestream. Visit Facebook Live’s map view during a news
conference or event on a stage, and you’ll often find major news organizations streaming the
same shot. And when the feed gets embedded prominently on the news website, the traditional,
unidirectional approach is transferred to the online product. For non-static events like festivals
and protests, opportunities to get inside the action are sometimes passed over for the security of
a static location with reliable internet service on the outskirts. It’s a trade-off that risks missing a
major development that might break in the thick of the crowd.
A modified mobile approach with reporter and camera team ‘chasing’ the action was used by
BuzzFeed News (2017) with Periscope during the March for Science on Earth Day 2017. The
stream opens with a shot of Bill Nye and a banner-carrying group leading the march. Off screen,
a male voice welcomes viewers and references Nye, then lets the action unfold. The crowd vol-
leys a chant, “Science . . . matters . . . science . . . matters”. Unlike a TV live shot, the journalist
does not identify himself on camera. Also unlike TV, the video is shot vertically.
About two minutes in, a female journalist enters the frame as they march, “Hello, I’m Nidhi
Subbaraman, I’m a science reporter at BuzzFeed News”. She narrates as they walk, providing
background and context, and then when the crowd noise takes over, she lets it, and stops talking.
A few minutes later a different reporter, Azeen Ghorayshi, enters the scene and introduces herself.
The camera keeps her framed in a head-and-shoulder vertical shot and occasionally moves to the
crowd and then back again to her as they march. The same male voice from the top of the stream
can be heard toward the end, reminding viewers to post questions.
The BuzzFeed example illustrates an emerging, hybridized TV live shot model that goes
beyond unidirectional broadcasting. It incorporates interactivity among the camera person/
reporter team, viewers, and people on the scene, offers ample time to let the action unfold, and
provides context through selective reporter narration.
Activists and citizen livestreams

After Donald Trump signed an Executive Order (2017) banning US entry from certain predomi-
nantly Muslim countries, documentary filmmaker Michael Moore headed to Twitter to spread
the word about a protest forming at JFK airport:
Everybody in NYC area – head to JFK Terminal 4 NOW! Big anti-Trump protest
forming out of nowhere! Ppl mobilizing against Trump’s Muslim ban.
(2017)
As he continues tweeting, livestreams from protesters, organizations, and reporters show up on

Periscope and Facebook Live. Nearly 3,000 viewers watch the ACLU’s Facebook Live out of
JFK, as the crowd chants, “No Trump, No KKK, no fascist USA, No Trump, No KKK, no fascist
USA . . . .” Two accounts with Arab and English lettering – Community Organization and Said
Abbasy – stream a similar scene (author notes; video clips and screenshots). CNN’s Brian Stelter
tweets, “Airport protest story getting bigger”, embedding a video from BuzzFeed editor Tomi
Obaro, and before that retweets Vox.com journalist Elizabeth Plank’s photo from JFK. Both
news organizations are livestreaming on Facebook (author notes via video clips and screenshots).
Meanwhile, a citizen Periscope with 35 viewers is coming out of Dulles International Airport
near Washington, DC. “There is free legal help available”, says the young woman on camera
300
(Buchalter, 2017a). “They are by the phone in the welcome center, so please spread the word”.
Behind her, people are chanting, “No hate, no fear, refugees are welcome here”. She continues,
There are families who are being separated right now. They were on the plane when
that order was put through, and they had no idea until they got here that they weren’t
going to be allowed off the plane – people who have lived here a really long time.
(Buchalter, 2017a)
Her viewers are responding with messages on screen, “glad to see people stream this” and “SOLI-
DARITY” (Buchalter, 2017a). Reflecting on her livestream that day, Erin Buchalter said she didn’t
feel like a journalist or protester, but rather “a person with an amazing responsibility” (2017b).
Her children, ages 7 and 10, had been there with her at the airport, and she said it allowed her to
teach them “why and how their voices matter”. After the first judge blocked the ban, she said,
“we were able to show our kids, ‘Look, you helped make that happen’” (2017b).
Buchalter notes that interacting with people who comment on her stream is important,
“Even if they are rude I will respond kindly” (2017b). In addition to handling offensive com-
ments, Buchalter says reliable internet service has been a challenge when livestreaming, and cites
the “massive amounts of people” trying to use the network at the Women’s March, which was
“down completely” (2017b).
Issues with data networks aren’t limited to major U.S. cities but can be an obstacle for
livestreamers worldwide. And in some countries, citizens and journalists alike even risk hav-
ing their phones stolen or confiscated by authorities (Linares, 2017). Online news source Efecto
Cocuyo combats the problem by using older model Motorola phones for its livestreams on Peri-
scope. The service formed on Twitter in early 2015 as an alternative to the mainstream news in
Venezuela, even before it had its own website. When Periscope became available several months
later, Efecto Cocuyo quickly adopted the tool to stream protests and political events (Linares, 2017.)
Two years later, the livestreams continue and have been a solid presence for its 26,000 Periscope
followers during months of anti-government protests (Efecto Cocuyo, 2017).
Innovative and resourceful activists have found ways to circumvent barriers and problem solve
even before Periscope and Facebook Live. Occupy Wall Street’s livestreaming operation began with
just one roving camcorder, a laptop, and a 4G wireless card (Captain, 2011). The day before police
raided its ad hoc media operation in Zuccotti Park, OWS moved to a rundown building in lower
Manhattan, with “cobbled-together gear and a hint of body odor” (Captain, 2011). There, the
team monitored hundreds of live feeds to select for broadcast on their globalrevolution.tv (Captain,
2011). From marches to arrests, the streams brought the movement to viewers worldwide.
But along with social witnessing, livestreaming can bring graphic and violent images.
Streaming violence and death

During its first year, Facebook Live streamed suicide, torture, rape, accidental death, and murder.
Several instances involved young people in Chicago – the kidnapping and torture of a disabled
teenager (Schmadeke, 2017) and a gang rape (Rosenberg-Douglas, 2017). Public outrage and
criticism followed each horrific episode. Then, on Easter Sunday, a murderer took to the plat-
form. He had just killed a man at random on a street in Cleveland and made a video of the
murder. His victim was a 74-year-old grandfather. Robert Godwin Jr., “just happened to walk
into the path of the deranged gunman”, who posted the video to his Facebook page and then
livestreamed himself talking about it (Morice, 2017). On the stream, the killer said, “I shamed
myself. I snapped . . . I just killed 13 motherf******, man . . . I’m about to keep killing until
they catch me, f*** it” (Morice, 2017).
301
Just over a week later, a young man in Thailand broadcast a livestream as he killed his
11-month-old baby daughter. More than a quarter million people watched the video, which
remained online for about 20 hours (Mozur, 2017). the New York Times reported: “The video
showed Mr. Wongtalay fixing a noose around his daughter’s neck and then dropping her off
the side of a building. After a burst of crying, he climbed over the side to retrieve her body. His
subsequent suicide was not shown online” (Mozur, 2017).
After the two killings Mark Zuckerberg posted to his Facebook page that seeing “people
hurting themselves and others on Facebook” was “heartbreaking” and that the company was
adding 3,000 people to its 4,500 already in place to monitor content. They were also working
on tools to simplify the process for reporting problems, identifying violations, and for contacting
police (2017).
Using the social media research tool twXplorer, a search for “Cleveland” conducted 24 hours
after the shooting, returned a random “snapshot” of 500 tweets. The most common terms used
in those tweets were “video” and “instead”. The text of those tweets urged people to retweet
photos of the victim and his family rather than the video showing Godwin’s death and horror
(2017).
Godwin’s grandson had reached out on Twitter, pleading that people honor his grandfather’s
life instead of perpetuating the horror of his death. “Please please please stop retweeting that
video and report anyone who has posted it! That is my grandfather show some respect #Cleve-
land” (Hurst, 2017). Among those who honored the request was Chris Hurst, a former anchor-
man, who several years earlier saw his girlfriend murdered on live television. He retweeted Ryan
Godwin’s message, adding, “I know that cry and hope more listen to you than they did to me”
(Hurst, 2017).
Hurst was referencing the videos of his girlfriend’s murder. One was from the live television
news broadcast, and the other was made by the killer himself. Back in late August 2015, a former
employee at WDBJ7 killed reporter Alison Parker and cameraman Adam Ward during an early
morning live interview at Smith Mountain Lake in Virginia (Shear et al., 2015).
The killer videotaped the shooting and then posted it to his social media accounts. “I filmed
the shooting see Facebook”, he tweeted (archived in a screenshot of shooter’s Twitter feed, 2015).
The shooter fled the scene and continued posting to social media. Reporters tweeted the killer’s
account name and noted that he was tweeting (screenshots of author’s Twitter feed, 2015). After
responses such as “Dear. Lord” and “I want to vomit”, reporters warned their Twitter followers
to avoid the posts, “Don’t watch it people, don’t watch it”. Some journalists posted visuals but
then quickly changed their minds. Andy Carvin, known for tweeting the Arab Spring revolution,
said the only tweet he ever deleted was a screenshot of that gunman. His response after tweeting
it: “oh shit” (2015).
While some blame these incidents on livestreaming tools and the social media platforms
that provide them, others would argue the news media are complicit in perpetuating these
public acts of violence. Sensationalist headlines, repeated use of video clips in broadcast cov-
erage, embedding images into web stories, and focusing on the perpetrator rather than the
victim can almost glorify the violence. A number of news organizations referred to the per-
petrator as the “Virginia Shooter” or the “WDBJ shooter”. And in the Cleveland shooting,
they used the moniker “Facebook Killer”, which persisted through the month following
the murder (Gingras, 2017; Remington, 2017). Media ethicist Kelly McBride at The Poynter
Institute encourages journalists to resist and minimize using nicknames for notorious crimi-
nals, because those monikers can “cause harm” (2017). McBride says the harm can take the
form of “contagion”, as nicknames can increase the possibility of copycats in shootings as
well as suicides (2017).
302
Meindl and Ivy (2017) also highlight practices used by media in covering mass shootings that
may lead to what they call generalized imitation. They include naming and showing visuals of
the killer and detailing the event and the life of the shooter, which may confer social status and
notoriety (Meindl and Ivy, 2017). Research on mass shootings from 2000 through 2012 shows
the time frame for contagion at about two weeks (Kissner, 2016). The researcher concluded that
perhaps, “the spectacle of recent active shootings contributes to the tragic crystallization of long-
simmering active shooting plans” (Kissner, 2016: 58). Another study found “significant evidence
that mass killings involving firearms are incented by similar events in the immediate past” (Towers
et al., 2015). The time period between incidents averaged 13 days.
One can imagine how livestreamed murders and other violent acts might complicate and
exacerbate these issues. Not only are journalists and their news organizations faced with report-
ing the killings, but they must grapple with the additional layer of the act being broadcast by the
killer on social media. How much of the video, if any, should news organizations use in their sto-
ries? Should they embed screenshots or clips onto their websites? What else should they report?
This statement from the SPJ Code of Ethics is often tweeted when sensational news is breaking:
“Avoid pandering to lurid curiosity, even if others do” (2014).
Some news organizations and even law enforcement agencies are moving away from naming
the perpetrator. After the mass shooting in an Orlando nightclub in summer 2016, then-FBI
director James Comey refused to speak the gunman’s name during a live news conference. “‘You
will notice that I am not using the killer’s name, and I will try not to do that,’ Comey said during
the live news conference”, reported the Associated Press (Gurman, 2016). “‘I don’t want to be
part of that for the sake of the victims and their families,’ Comey said, ‘and so that other twisted
minds don’t think that this is a path to fame and recognition’” (Gurman, 2016).
Potential effects of viewing nonfictional violent images

Choosing not to watch violent livestreams and video postings can be the first defense against
their negative effects. But that may not be an option for journalists. Studies, however, suggest
relationships between frequent and extensive viewing of graphic, violent nonfictional content
and increased likelihood of psychological distress. Feinstein et al. (2014) studied journalists who
worked with extremely violent user-generated images and found that frequency was more dis-
tressing than duration. The more frequently the journalists were exposed to the graphic images,
the more likely they were to suffer from anxiety, depression, PTSD, or alcohol consumption.
Earlier research on viewing television coverage of the 9/11 World Trade Center terrorist
attacks suggests a relationship between hours watched and greater likelihood of PTSD (Schlenger
et al., 2002). Another study found similar results. Levels of PTSD and depression were higher
among participants who repeatedly saw scenes of people falling or jumping from the towers
(Ahern et al., 2002).
Another concern with viewing video violence is desensitization and diminished empathy.
Exposure to violence in real life and knowing that the video violence is real appears to make a
difference in viewers’ reactions. College students with limited exposure to real-world violence
had higher empathy, but higher exposure to violence in the real world was linked to “higher
trauma symptoms, escape to fantasy, and reduced empathy” (Mrug et al., 2014: 1106). In another
study, participants saw scenes of actual violence or fictional violence. Those who knew they
were watching real people being harmed responded with higher empathy than those who saw
the fictional scenes (Ramos et al., 2013). Taken together, these studies suggest that frequency and
time spent viewing images of real-world violence can relate to empathy levels as well as negative
psychological effects, including PTSD, depression, and anxiety.
303
When working with graphic images, experts recommend being mindful of your exposure
“load,” reducing unnecessary viewing, and building distance by avoiding faces and directing
attention to other elements of the scene (Dart Center, 2014).
Special considerations on covering livestreamed suicide

When a man pulled his truck over on a Los Angeles freeway in 1998 and began shooting into
the air, police closed the road, traffic backed up, and TV news choppers went live with the story.
Six L.A. stations and MSNBC stayed live as the man set his truck on fire and shot himself in
the head (Rogers, 1998). Los Angeles Times media columnist Howard Rosenberg called the live
broadcasts a “total abrogation of journalistic responsibility” (1998). This was nearly two decades
before mobile livestreaming, but concerns about volatility during live broadcasts ending in grue-
some acts existed then and are magnified today. At least six suicides have been livestreamed on
social media between May 2016 and May 2017. They included adolescent girls who interacted
with viewers before killing themselves (Miller and Burch, 2017) and a man with bipolar disorder
who set himself on fire (McCausland, 2017). The interactivity of social media livestreaming adds
new considerations and complexity to an already difficult situation. While people watching the
stream could potentially help by alerting authorities or simply by being there to listen, they could
instead exacerbate the situation. Intentionally goading the person to “do it” or inadvertently say-
ing the wrong thing could push the person closer to ending her life.
Livestreaming suicides has become a newsworthy issue. In the past, policy at many news
organizations had been to simply refuse to cover suicides, for both privacy and concerns about
contagion. But today, with people making their own livestreams, the images and events are
thrust into the public view. And with each new suicide attempt on Periscope or Facebook
Live, reporters and news organizations are confronted with questions on how to tell the story.
The website reporting on suicide.org warns that stories that use graphic images, explicitly
describe the suicide method, or sensationalize or glamorize death increase risk for additional
suicides (2015).
But Facebook is unlikely to stop these livestreams. The Guardian reports suicide attempts are
permitted on the platform (Hopkins, 2017).
Facebook will allow users to livestream attempts to self-harm because it “doesn’t want
to censor or punish people in distress who are attempting suicide”, according to leaked
documents. However, the footage will be removed “once there’s no longer an opportu-
nity to help the person” – unless the incident is particularly newsworthy.
(Hopkins, 2017)
Periscope’s community guidelines tell users not to post self-harm or suicide but also say the orga-
nization may reach out to help people who do (2017).
It appears that neither Facebook nor Periscope will automatically shut down livestreams
while people are broadcasting suicidal thoughts or preparing to take their own lives. And the
Facebook policy of removing the stream after there’s “no longer an opportunity to help the
person” suggests that could mean the person was rescued, had a change of heart, or died during
the livestream.
If and when the next suicide is streamed live on social media, journalists could help minimize
harm by avoiding sensationalizing the suicide, using family photos rather than images of the
scene, and including resource information (reporting on suicide.org, 2015).
304
Privacy
With social media livestreaming, does anyone any longer have any right to privacy? Stewart and
Littau say that mobile streaming video technologies could potentially reshape how we think
about privacy and the right to record in public (2016). The pervasiveness of video recording
may have moved into the realm of what Nissenbaum calls “the tyranny of the normal” because
they are so commonplace, and he says that objections are difficult to pursue “against the force
of reasonable expectation” (2009: 161). In addition, collecting and disseminating information
had previously been separate actions that could potentially affect privacy. Today, social media
livestreaming allows these activities to take place concurrently, increasing “the potential for harm
that cannot be undone” (Stewart and Littau, 2016: 317).
There’s also another layer, the commenting function, which allows the livestreamer to interact
with viewers in real time. Section 230 of the U.S. Communications Decency Act has provided
protection against defamation for reader comments posted to news stories (Reporters Com-
mittee, undated). It’s unclear, however, if that protection would carry over to comments during
livestreamed content. Individual state laws regulating recording and eavesdropping may be used
against reporters livestreaming (Stewart and Littau, 2016). Corporations may also dictate their
own policies for privacy and use of service. For example, United Airlines states online, “Photo-
graphing or recording other customers or airline personnel without their express consent is pro-
hibited” (2017). However, it’s unclear whether permission had been granted to passengers who
made videos of a man being forcibly removed from a United plane for refusing to give up his
seat on an overbooked flight. The videos spread virally on social media and attracted mainstream
news attention. Following the incident, United settled with the injured man and has agreed to
pay passengers up to $10,000 for giving up their seats on overbooked flights (Zumbach and
Byrne, 2017).
Technological determinism and livestreaming: ‘a force for good’

or profit and material power?
Taken with the idea of progress, Thomas Jefferson and Benjamin Franklin embraced technology
as a form of human betterment. Franklin even refused to patent his inventions, seeing them as a
way to serve others rather than enriching himself (Smith and Marx, 1994). Recognizing technol-
ogy as a key driving force in society, critics, however, feared it would sacrifice moral progress for
material power. Some viewed technology as autonomous and “out of control”, following “its
own course, independent of human direction” (Winner, 1977, in Smith and Marx, 1994: 31).
Social media livestreaming might reveal a bit of truth in both views. Media columnist Mar-
garet Sullivan, while calling out Facebook for being “better at making money and capturing
eyeballs than at owning its equally huge power and responsibilities”, acknowledged livestreaming
as “a force for good, too” (2017). She was referring to Diamond Reynolds, who livestreamed
her boyfriend Philando Castile dying after police shot him near Minneapolis. Sullivan called it
an “important piece of bearing witness” (2017). The video went viral, protests against excessive
force by police – especially against black men – ensued, and the officer went to trial for man-
slaughter (Horner, 2017). The officer was acquitted 16 June 2017.
On the other side are material power and the profit motive in livestreaming. In its report,
“What is happening to television news?” the Reuters Institute for the Study of Journalism
notes that livestreaming presents not only editorial challenges but “questions around what kind
of business private news providers can build around them” (Nielsen and Sambrook, 2016: 20).
Social media companies are now selling ads in their livestreams. Pre-roll ads became available on
305
Periscope to “select advertisers” in late March of 2017 (Folgner, 2017), and Facebook is beta test-
ing ad breaks in livestreams (Boland and Angelidou-Smith, 2017).
But the desire for profit in broadcasting is not limited to corporate interests, and it’s not new.
A Microsoft Research report explores motivations for livestreaming among popular streamers
on Periscope and then-Meerkat, and personal branding was the most common response. This
appeared motivated by income through creating a fan base, integrating work-related goals with
their streams, and “driving traffic toward their monetized online resources” (Tang et al., 2016).
Reaching back in time, when live broadcasting was first developing nearly a century ago, the
profit motive drove the U.S. commercial radio model (Barnouw, 1966). And over the years, a
quest for the dollar has gone hand in hand with advances in technology – from its creation to
adoption.
Conclusion
Coming full circle to the Facebook Live launch, when Mark Zuckerberg exuded enthusiasm for
the technology, “It’s like having a TV camera in your pocket all the time”, he said. “And we’re
excited to bring this superpower to everyone in our community” (2016). This “superpower”
has shown itself to be a force for good, while at the same time presenting challenges and ethi-
cal dilemmas for those who create, watch, and report on the streams, as well as creating possible
harm. Being mindful of social media livestreaming’s potential impact, journalists and the public
can take steps toward using it wisely, avoiding, as Thoreau wrote in Walden, becoming “tools of
their tools” (Thoreau, 1854, cited in Smith and Marx, 1994: 26).
Further reading
The Online News Association has been at the forefront of digital journalism since its incep-
tion in 1999 and offers a wealth of resources and opportunities. For guidance on working with
graphic and distressing images, the Dart Center for Journalism & Trauma provides tip sheets and
other useful materials. And Reporting on Suicide.org offers guidelines to help minimize harm
when covering livestreamed suicides.
Note
1 While Periscope and Meerkat apps predate the Facebook release, its availability to nearly 2 billion users
makes this a notable historic marker (Facebook Newsroom 2017). And because YouTube limits mobile
livestreaming to accounts with more than 1,000 users, it is not included in this chapter (Google 2017).
References
Ahern, J., Galea, S., Resnick, H., Kilpatrick, D., Bucuvalas, M., Gold, J. and Vlahov, D. (2002) “Television
images and psychological symptoms after the September 11 terrorist attacks.” Psychiatry: Interpersonal
and Biological Processes, 65(4), 289–300. Retrieved from http://guilfordjournals.com/doi/abs/10.1521/
psyc.65.4.289.20240
Al Jazeera English. (2015, October 9) Facebook. Retrieved from www.facebook.com/aljazeera/photo
s/a.10150243828793690.369310.7382473689/10153781845613690/
Barnouw, E. (1966) A Tower in Babel. New York: Oxford University Press.
Bild. (2017) “Video diary of the Bild reporter: Live transfer of an escape from hell.” Bild. Retrieved from
www.bild.de/politik/ausland/zuwanderung/flucht-aus-der-hoelle-das-video-42352374.bild.html
Bishop, T. (2015, April 29) “Watch a Baltimore resident confront Geraldo Rivera over Fox News’ irrespon-
sible coverage.” Media Matters for America. Retrieved from www.mediamatters.org/blog/2015/04/29/
watch-a-baltimore-resident-confront-geraldo-riv/203475
306
Boland, B. and Angelidou-Smith, M. (2017, February 23) “An update on video monetization.” Facebook.
Retrieved from https://media.fb.com/2017/02/23/update-on-video-monetization/
Buchalter, E. (2017a) Periscope Livestream Recorded by Author.
Buchalter, E. (2017b) Email Interview.
BuzzFeed News. (2017, April 22) Periscope livestream recording on Twitter. Retrieved from https://twitter.com/
BuzzFeedNews/status/855849076985708544
Captain, S. (2011, November 16) “Inside Occupy Wall Street’s (kinda) secret media HQ.” Wired. Retrieved
from www.wired.com/2011/11/inside-ows-media-hq/
Carvin, A. (2015, September 26) “Online News Association Annual Conference.” Author’s Notes From Panel.
Dart Center. (2014, August 12) “Working with traumatic imagery.” Dart Center. Retrieved from https://
dartcenter.org/content/working-with-traumatic-imagery
Dredge, S. (2015, September 13) “How live video on Periscope helped ‘get inside’ the Syrian refugees
story.” The Guardian. Retrieved from www.theguardian.com/media/2015/sep/13/periscope-app-syrian-
refugees-bild
Efecto Cocuyo. (2017) Periscope. Retrieved from www.periscope.tv/EfectoCocuyo/1gqGvbXlQanGB
Ellick, A. (2015) “Online News Association Annual Conference Table Talk Session.” Author’s Notes.
Executive Order. (2017, January 27) Protecting the Nation From Foreign Terrorist Entry into the United States. Retrieved
from www.whitehouse.gov/the-press-office/2017/01/27/executive-order-protecting-nation-foreign-
terrorist-entry-united-states
Facebook Newsroom. (2017) Statistics. Retrieved from https://newsroom.fb.com/company-info/
Feinstein, A., Audet, B. and Waknine, E. (2014) “Witnessing images of extreme violence: A psycho-
logical study of journalists in the newsroom.” Journal of the Royal Society of Medicine Open, 5(8), 1–7.
doi:10.1177/2054270414533323
Folgner, M. (2017, March 28) “Expanding pre-roll ads to Periscope video.” Twitter Marketing. Retrieved from
https://blog.twitter.com/2017/expanding-pre-roll-ads-to-periscope-video
Gingras, B. (2017, May 12) “Pennsylvania police open up about catching the ‘Facebook killer.’” CNN. Retrieved
from www.cnn.com/2017/05/12/us/beyond-the-call-pennsylvania-police-catch-facebook-killer/
Google. (2017) “Introduction to live streaming.” YouTube Help. Retrieved from https://support.google.com/
youtube/answer/2474026?hl=en
Gurman, S. (2016, June 28) “Terror talk changes as Comey refuses to name gunman.” Associated Press. Retrieved
from https://apnews.com/67d4fe9a88884794a79ffd464a8edc0b/terror-talk-changes-comey-refuses-
name-gunman
Hopkins, N. (2017, May 21) “Facebook will let users livestream self-harm, leaked documents show.” The
Guardian. Retrieved from www.theguardian.com/news/2017/may/21/facebook-users-livestream-self-
harm-leaked-documents
Horner, S. (2017, June 6) “Jury sees squad car video in opening day of Philando Castile shooting trial.”
Twin Cities.com Pioneer Press. Retrieved from www.twincities.com/2017/06/05/latest-from-yanez-trial-
opening-arguments-and-officers-side-of-story-could-come-monday/
Hurst, C. (2017, April 16) “I know that cry and hope more listen to you than they did to me.” Twitter.
Retrieved from https://twitter.com/ChrisHurstVA/status/853795288305328128
Kissner, J. (2016) “Are active shootings temporally contagious? An empirical assessment.” Journal of Police
and Criminal Psychology, 31(1), 48–58. doi:10.1007/s11896-015-9163-8
Lenzner, B. (2014) “The emergence of Occupy Wall Street and digital video practices: Tim Pool, live
streaming and experimentations in citizen journalism.” Studies in Documentary Film, 8(3), 251–266.
Lewis, P. (2015a, April 9) Rand Paul: The Guardian Interview on Periscope. Retrieved from www.theguardian.
com/us-news/2015/apr/09/ask-rand-paul-periscope-guardian-conversation
Lewis, P. (2015b, April 28) “The Baltimore riots: The night on Periscope – Video.” The Guardian.
Retrieved from www.theguardian.com/us-news/video/2015/apr/28/the-baltimore-riots-the-night-
on-periscope-video
Lewis, P. (2015c, April 29) “Baltimore unrest: Periscope captures tense build-up to curfew – Video.” The Guard-
ian. Retrieved from www.theguardian.com/us-news/video/2015/apr/29/baltimore-riots-periscope-
captures-tense-build-up-to-curfew-video
Lewis, P. (2017, May 5) Telephone Interview With Author.
Linares, C. (2017, April 12) “Independent Venezuelan news site triumphs with live video on Periscope
while many move to Facebook Live.” Knight Center for Journalism in the Americas. Retrieved from https://
knightcenter.utexas.edu/blog/00-18222-independent-venezuelan-news-site-triumphs-live-video-peri
scope-while-many-move-faceboo
307
Lui, K. (2016, December 11) “This is Facebook’s most watched live video of 2016.” Fortune. Retrieved from
http://fortune.com/2016/12/12/facebook-most-watched-live-video-2016/
McBride, K. (2017, April 21) “Don’t call Steve Stephens ‘the Facebook killer.’” Poynter. Retrieved from
www.poynter.org/2017/dont-call-steve-stephens-the-facebook-killer/456868/
McCausland,P. (2017,May 14) “Memphis,Tennessee,man fatally sets himself on fire on Facebook Live.”NBC News.
Retrieved from www.nbcnews.com/news/us-news/memphis-tennessee-man-fatally-sets-himself-fire-
facebook-live-n759366
Meindl, J. and Ivy, J. (2017) “Mass shootings: The role of the media in promoting generalized imitation.”
American Journal of Public Health, 107(3), 368–370. doi:10.2105/AJPH.2016.303611
Miller, C. and Burch, A. (2017, January 24) “Another girl hangs herself while streaming it live – This time
in Miami.” Miami Herald. Retrieved from www.miamiherald.com/news/local/article128563889.html
Moore, M. (2017, January 28, 3:20 p.m) Twitter. Retrieved from https://twitter.com/MMFlint/
status/825438373003280384
Morice, J. (2017, April 18) “Facebook killer chooses victim at random, laughs about killing in videos.”
Cleveland.com. Retrieved from www.cleveland.com/metro/index.ssf/2017/04/accused_facebook_live_
killer_c.html
Mozur, P. (2017, April 25) “Father in Thailand kills 11-month-old daughter live on Facebook.” The New
York Times. Retrieved from www.nytimes.com/2017/04/25/world/asia/thailand-phuket-facebook-
killing-daughter.html
Mrug, S., Madan, A., Cook III, E. and Wright, R. (2014) “Emotional and physiological desensitization to
real-life and movie violence.” Journal of Youth and Adolescence, 44(5), 1092–1108. Retrieved from https://
link.springer.com/article/10.1007/s10964-014-0202-z
The New York Times. (2017, January 21) “Women’s March highlights: Viewer questions.” The New York
Times. Retrieved from www.nytimes.com/interactive/2017/01/21/us/questions-march-live.html?_r=0
Nielsen, R. and Sambrook, R. (2016) What Is Happening to Television News? Reuters Institute for the Study of
Journalism. Retrieved from http://reutersinstitute.politics.ox.ac.uk/sites/default/files/What%20is%20
Happening%20to%20Television%20News.pdf
Nissenbaum, H. (2009) Privacy in Context: Technology, Policy, and the Integrity of Social Life. Stanford, CA:
Stanford University Press.
Periscope. (2017) Community Guidelines. Retrieved from www.periscope.tv/content
Ramos, R., Ferguson, C., Frailing, K. and Romero-Ramirez, M. (2013) “Comfortably numb or just yet
another movie? Media violence exposure does not reduce viewer empathy for victims of real violence
among primarily Hispanic viewers.” Psychology of Popular Media Culture, 2(1), 2–10. Retrieved from
http://psycnet.apa.org/journals/ppm/2/1/2/
Remington, K. (2017, May 13) “Pennsylvania State Police talk about catching Cleveland Facebook killer.”
Cleveland.com. www.cleveland.com/metro/index.ssf/2017/05/pennsylvania_state_police_talk_1.html
Reporters Committee, undated. “Immunity for reader comments under Section 230 of the Communications
Decency Act.” Retrieved from www.rcfp.org/browse-media-law-resources/digital-journalists-legal-
guide/immunity-reader-comments-under-section-23
Reporting on suicide.org. (2015) Recommendations for Reporting on Suicide. Retrieved from http://reportin-
gonsuicide.org/recommendations/#important
Rodriguez, R. (2015, May 26) “Periscope: Four ways it’s shaking up media.” CNN. Retrieved from www.
cnn.com/2015/05/26/tech/periscope-android-media/
Rogers, P. (1998, June) “L.A.’s TV news: Pulling away from live shots?” AJR. Retrieved from http://ajrar
chive.org/article.asp?id=2318
Rosenberg, H. (1998, May 2) “The Russian Roulette of live news coverage.” Los Angeles Times. Retrieved
from http://articles.latimes.com/1998/may/02/entertainment/ca-45412
Rosenberg-Douglas, K. (2017, March 22) “Facebook Live assault of girl, 15: ‘No one human deserves what
happened to her’.” Chicago Tribune. Retrieved from www.chicagotribune.com/news/local/breaking/ct-
teen-girl-missing-from-lawndale-20170321-story.html
Schlenger, W., Caddell, J., Ebert, L., Jordan, B., Rourke, K., Wilson, D., Thalji, L., Dennis, J., Fairbank, J.
and Kulka, R. (2002) “Psychological reactions to terrorist attacks findings from the national study of
Americans’ reactions to September 11.” JAMA, 288(5), 581–588. Retrieved from http://jamanetwork.
com/journals/jama/fullarticle/195165
Schmadeke, S. (2017, January 27) “Teens in Facebook Live attack indicted, lawyers call case ‘sensational-
ized.’” Chicago Tribune. Retrieved from www.chicagotribune.com/news/local/breaking/ct-facebook-
live-attack-court-met-20170127-story.html
308
Shear, M., Perez-Pena, R. and Blinder, A. (2015, August 26) “Ex-broadcaster kills 2 on air in Virginia shoot-
ing: Takes own life.” The New York Times. Retrieved from www.nytimes.com/2015/08/27/us/wdbj7-
virginia-journalists-shot-during-live-broadcast.html
Simo, F. (2017, April 6) Facebook. Retrieved from www.facebook.com/fidji.simo/videos/vb.576027062/
10154258738027063/
Smith, M. and Marx, L. (1994) Does Technology Drive History? The Dilemma of Technological Determinism.
Cambridge, MA: The MIT Press.
SPJ. (2014, September 6) “Code of ethics.” Society of Professional Journalists. Retrieved from www.spj.org/
ethicscode.asp
Stewart, D. and Littau, J. (2016) “Up, periscope: Mobile streaming video technologies, privacy in public,
and the right to record.” Journalism & Mass Communication Quarterly, 93(2), 312–331. doi:10.1177/
1077699016637106
Story, L. (2017, February 6) Telephone Interview With Author.
Sullivan, M. (2017, April 18) “Facebook still denies what it is. The Cleveland murder shows how.” The
Washington Post. Retrieved from www.washingtonpost.com/lifestyle/style/mark-zuckerberg-has-
vowed-to-limit-live-streamed-depravity-so-far-its-an-empty-promise/2017/04/18/d44de158–243c-
11e7-b503-9d616bd5a305_story.html?tid=ss_tw&utm_term=.93e690747c01
Tang, J., Venolia, G. and Inkpen, K. (2016) “Meerkat and periscope: I stream, you stream, apps stream for
live streams.” Microsoft Research, CHI’16, 7–12 May, San Jose, CA. Retrieved from https://pdfs.semantic
scholar.org/4501/613e3be5b0f9b4a5bda22c1ead846b111957.pdf
Towers, S., Gomez-Lievano, A., Khan, M., Mubayi, A. and Castillo-Chavez, C. (2015) “Contagion in mass
killings and school shootings.” PLoS ONE, 10(7), e0117259. Retrieved from http://journals.plos.org/
plosone/article?id=10.1371/journal.pone.0117259
twXplorer. (2017) Knight Lab. Retrieved from https://twxplorer.knightlab.com/
United Airlines. (2017) Electronic Device Policies. Retrieved from www.united.com/web/en-US/content/
travel/inflight/devices.aspx
Weil, K. (2015, March 26) “Introducing Periscope.” Twitter. Retrieved from https://blog.twitter.com/2015/
introducing-periscope
Witty, P. (2015, September 26, 2:37 p.m) Twitter. Retrieved from https://twitter.com/patrickwitty/
status/647842356922871808
Zuckerberg, M. (2016, April 6) “Live video launch.” Facebook Live. Retrieved from www.facebook.com/
zuck/videos/vb.4/10102764535989511
Zuckerberg,M. (2017,May 3) Facebook. Retrieved from www.facebook.com/zuck/posts/10103695315624661
Zumbach, L. and Byrne, J. (2017, April 27) “Passenger dragged from United plane at O’Hare reaches
settlement with airline.” Chicago Tribune. Retrieved from www.chicagotribune.com/business/ct-united-
passenger-dragging-settlement-0428-biz-20170427-story.html
309
PART V
Digital Journalism Studies

Dialogues
23
ETHICAL APPROACHES TO
COMPUTATIONAL JOURNALISM
Konstantin Dörr
Since the implementation and evolution of the personal computer and the internet in the 1980s
and 1990s, the way that news is selected, produced, distributed, and consumed has changed
significantly (Domingo and Paterson, 2011; Mitchelstein and Boczkowski, 2009; Powers, 2012).
Today, algorithms are supporting and replacing work autonomously as operations, decisions, and
choices are increasingly delegated to software (Acemoglu and Restrepo, 2016; Arntz et al., 2016;
Autor, 2015; Just and Latzer, 2016). Algorithms decide which news is visible on Twitter and
Facebook. They determine the results from a Google search query; they also recommend which
news items are worth reading and how they are consumed (Anderson, 2012; Bucher, 2012;
Latzer et al., 2016; Resnick and Varian, 1997; Rieder, 2005). They predict the demand for specific
journalistic products as well as observe the audience. They are news producers, demand predic-
tors, and gatekeepers (Napoli, 2014; Vos and Heinderyckx, 2015). Thus, the border between
professional actors, as well as between the intersection of production and distribution tech-
nologies, is vanishing (Neuberger, 2009; Weischenberg and Hienzsch, 1994: 478). Algorithmic
applications, e.g., WordSmith for automated text creation, can select, contextualize, and assign
relevance to certain pieces of information, generate journalistic value, and consequently fulfill
the functions of professional journalism, e.g., with automated sports reports (Caswell and Dörr,
2017; Dörr, 2016b). As journalism has the task of selecting, processing, and publishing news in
accordance with professional norms and values (Brosda, 2010: 259), algorithmic applications are
to be evaluated not only in terms of design – to fulfill the task they are programmed to complete
or respectively to proceed toward a pre-defined goal – but also ethical norms (Moor, 2006: 17).
This increasing complexity in human-computer-interaction is challenging the ways in which
professional journalism, the public, and communication studies assess and analyze the normative
and ethical challenges of digital news production like objectivity, accountability, and transparency
from a technical as well as theoretical perspective (Dörr et al., 2017; Ward, 2017). Along with this
advent and variety of algorithmic applications and the datafication of news production, the values
and responsibilities assigned to journalism and journalists are changing accordingly and must be
addressed from an ethical perspective (Brosda, 2010; Culver, 2016; Ward, 2017).
To provide ethical guidance along with normative questions for journalism practice, this
chapter locates the ethical discussion within media ethics and focuses on journalism ethics as a
reference point for an ethical and normative analysis of computational journalism. The chapter
shows how an ethical analysis of the question of responsibility can benefit from the theoretical
313
Konstantin Dörr
approaches of the sociology of technology, with a particular focus on Rammert and Schulz-
Schaeffer’s (2002) concept of distributed agency. Thereby, it is possible to open an ethical debate
for normative questions without being solely limited to human agency and to integrate the role
of technology, as both developments are mutually dependent (Lewis, 2017).
Ethical challenges of computational journalism

It is important to clarify the state of the art of digital journalism to address normative questions.
The number of definitions and concepts like data or data-driven journalism (Gray et al., 2012),
computer-assisted reporting (Mayo and Leshner, 2000), and algorithmic, robotic, or automated
journalism (Anderson, 2012; van Dalen, 2012; Dörr, 2016b; Thurman et al., 2017) complicates
the theoretical demarcation and analysis (see Coddington, 2015, and this volume, Chapter 17;
Borges-Rey, this volume, Chapter21; and Dörr, 2016a, for a classification). Each of these con-
cepts focuses on different use cases of data and software, ranging from research methods to story
visualization and automated text generation. However, it is agreed that computational journalism
serves as an umbrella term for technical innovations in journalism (Diakopoulos, 2017; Flew
et al., 2012; Koliska and Diakopoulos, this volume, Chapter 19; Neuberger and Nuernbergk,
2015). This understanding encompasses the focus on tooling and designing “practices or services
built around computational tools in the service of journalistic ends” (Coddington, 2015: 6). With
reference to Diakopoulos (2010, 2017: 177), computational journalism is here understood as “the
application of computing and computational thinking to enable journalistic tasks such as infor-
mation gathering, organization and sensemaking, storytelling, and dissemination”.
While automation processes have always been part of journalism (Pavlik, 2017), the rapid
growth and use of digital data is not a new phenomenon, either (Buhl et al., 2013; Gandomi and
Haider, 2015; Zikopoulos et al., 2013). The term ‘big data’ summarizes the economic and soci-
etal risks and opportunities of mostly unstructured digital data and evokes normative challenges
itself, e.g., on transparency, objectivity, and accountability (Mayer-Schönberger and Cukier, 2013;
Richter, 2016).
This influence of algorithms and the increasing use of digital data raise new ethical questions
for professional journalism, which arise on various levels and are detected and discussed at the
following levels:
• data search, origin, and use (Bradshaw, 2014; Zion and Craig, 2014; Zwitter, 2014),
• algorithmic accountability (Diakopoulos, 2015, 2016; Dörr and Hollnbuchner, 2016; Dörr
et al., 2017),
• algorithmic authority (Carlson, 2014; Kraemer et al., 2011; Young and Hermida, 2015),
• algorithmic objectivity (e.g., Gillespie, 2014), and
• algorithmic transparency (Diakopoulos and Koliska, 2016; McBride and Rosenstiel, 2014;
Montal and Reich, 2016).
As journalism ethics draws on and has been influenced by other ethical fields of application,
this chapter briefly explains the essential theoretical approaches that deal with the impact of
technology.
The terms ‘ethics’ and ‘morality’ must be distinguished at the outset. In general, ethics is a
discipline of practical philosophy dealing with and reflecting the morality of human action. In
contrast to ethics, morality means the variety of values, norms, and rules of human action that is
valid in a particular group, community, or society (Köberer, 2014: 21). Morality is “a system of
rules for guiding human conduct, and principles for evaluating those rules” (Tavani, 2011: 36). It
is a social phenomenon based on the social acceptance of norms and values. Norms, on the one
314
Ethical computational journalism
hand, are generally applicable regulations for human action that follow social values. Values, on
the other hand, are desired social conceptions used to establish or legitimize norms.
Due to the development in information and communication technologies (ICT), the field
and cluster of ethical problems is dynamic. For example, in the case of technology design, norms
and values are either embedded at an organizational level – as part of the software development
in tech companies or media organizations – or are starting to be adapted in machine learning
processes autonomously due to the progress in artificial intelligence. Machine learning is “any
methodology and set of techniques that can employ data to come up with novel patterns and
knowledge, and generate models that can be used for effective predictions about the data” (Van
Otterlo, 2013). This latest development raises ethical questions whenever existing or future tech-
nology is causing normative uncertainty (Grunwald, 2016: 25). Therefore, ethical approaches to
new technologies “pose new versions of standard moral problems and moral dilemmas, exacer-
bating the old problems, and forcing us to apply ordinary moral norms in uncharted realms” (for
computer ethics, Johnson, 1985: 1, 2009). Until the late 1990s, technology was seen only as an
intermediary, a tool, without the possibility of acting morally (Grunwald, 2016; Kroes and Ver-
beek, 2014). But the complexity of software systems and recent developments in artificial intel-
ligence show that technology is more than just a simple tool. Human actors are no longer the sole
moral authority when it comes to the selection and the processing of data and data signals (Adam,
2005). Consequently, it is necessary to analyze the degree of agency of algorithmic applications
in addition to their human counterparts and within the ethical discourse in journalism (Heise,
2016: 205; Mittelstadt et al., 2016).
Not only linked to computational systems and more of an encompassing approach, the eth-
ics of technology focuses on the social context of technology, its prospects, its intended use, and
on possible implications for society. It serves as a guide in the social discourse and is part of the
political, economic, and scientific discussion on technology in general. Counseling in these areas,
for example, provides ethical reflection on political debates about the regulation of the internet,
on the impact of autonomous software (artificial intelligence), the relationship between technol-
ogy and life (e.g., biotechnology, nanotechnology), or on questions of distributive justice and
digital divide (Grunwald, 2016, 30–32; Ropohl, 1995). The ethics of technology serves as a guide
in the early stages of a technical development, providing exploratory considerations on possible
consequences. Today, the transition to other ethical fields is permeable, as they intertwine.
Within technology ethics, machine ethics – sometimes referred to as robot ethics or algorithm
ethics – is another subfield. Emerging from artificial intelligence research, it specifically focuses
on the task of ensuring the ethical behavior of artificial agents (Shulman et al., 2009; Wallach
et al., 2008), e.g., on the morality of machines, or autonomous systems such as agents, robots, or
self-driving cars (Anderson and Anderson, 2011).
With relation to the field of journalism, media ethics and journalism ethics in particular allow
the ethical discussion of algorithms to be framed. Media ethics – a form of applied ethics – has
the task of constituting moral concepts for media-related actions and combining theoretical and
practical views in relation to ethical and normative standards (Köberer, 2014: 25). In the literature,
media ethics is often placed on a par with journalism ethics, but journalism ethics is a subfield of
media ethics. In addition to journalism ethics, media ethics also focuses on ethical questions in
other fields of application, such as advertising and public relations.
Within media ethics, journalism ethics serves specifically as a guide to approved societal moral
values as well as helps to define journalists’ work as professionals (Ward, 2017). In the scien-
tific discourse, there are various systematic approaches to journalism ethics (for an overview, see
Brosda and Schicha, 2010). In this chapter, journalism ethics is conceptualized as a part of media
ethics and as normative ethics, related as a subfield to the ethics of philosophy. Within journal-
ism ethics, the differentiation into various levels of action allows the attribution of responsibility
315
Konstantin Dörr
at an individual and a social ethical level, referring to individual journalists as well as to actors at
the organizational and social levels (Debatin, 1998: 121). In media ethics, both individuals (media
maker and media user) as well as companies and media organizations (e.g., publisher) and edito-
rial offices are defined as actors (Dörr et al., 2017: 124). Ethical questions of responsibility were
and are still solely linked to these stakeholders, as only human actors are able to act reasonably
(e.g., Kant KpV, 1788; Wunden, 1999). This chapter emphasizes the inclusion of socio-technical
approaches of the role of technology into the ethical discussion of journalism (Kroes and Ver-
beek, 2014). Although not emphasizing ethics, research is already beginning to address this issue
by focusing on actor-network theory (Latour, 2005; Lewis and Westlund, 2015; Salovaara, this
volume, Chapter 30). However, the increasing differentiation of journalistic tasks leads to ques-
tions of whether and how human beings or technology is assigned with responsibility. This calls
for a graduated approach. From a theoretical perspective, the concept of distributed agency by
Rammert and Schulz-Schaeffer (2002) allows the interplay between human agency and software
outcomes to be analyzed, and the distribution, as well as attribution, of responsibility in journal-
ism to be framed for algorithmic applications. With the adoption of this concept, it is possible
to include technical questions for a broader normative analysis in journalism. This approach of
techno-social interaction helps to overcome the dichotomy of technological and social deter-
minism, as the concept of distributed agency focuses on the interplay between technology and
society (Latzer, 2013; Rip, 2002).
Socio-technical concepts of agency

While questions of responsibility were throughout the philosophical and ethical discourse only
assigned to human actors, such as journalists, programmers, or service providers, it is valuable to
discuss the concept of agency and the moral status of technology in the light of algorithms and
developments in artificial intelligence (Kroes and Verbeek, 2014). To do so, this chapter refers to
socio-technical concepts of agency that help to visualize and theoretically frame the interplay
between journalistic actors and algorithms (Matsuzaki, 2011; Rammert, 2003, 2007).
In general, technical innovations like algorithmic search, text production, or recommenda-
tion applications can be classified as organizational, political, cultural, and institutional change
(Werle, 2007: 31). There are three different approaches to analyzing the impact of technology
on society. Technology can be seen as an actor, institution, or a system (Dolata and Werle, 2007:
17). Concepts of agency – e.g., the actor-network theory of Bruno Latour (2005) or the previ-
ously mentioned concept of distributed agency (Rammert and Schulz-Schaeffer, 2002) – focus
on the micro-level. Institutional and system approaches focus on the impact of technology on
an organizational (meso-) and social (macro-) level (Lessig, 1999; Winner, 1986). Central to this
classification is the question of whether software can be analyzed as an independent entity with
autonomous agency (Dolata and Werle, 2007: 46).
Within this theoretical discussion of agency, there are four different positions (Matsuzaki,
2011: 301). The first is sociological concepts, which solely attribute agency to human actors
(Collins and Kusch, 1998). The second is the approach of Bruno Latour (1987), in which agency
is related to the state of change, whether caused by human or technology. Third, there are
approaches that directly criticize this operationalization of agency, focusing instead on a more
complex and graduated approach to agency (Rammert and Schulz-Schaeffer, 2002). Finally, there
are approaches that operationalize technology as a fictional actor (Werle, 2002).
In this chapter, I argue that rather than attributing to technology aspects of intention or
consciousness in its operation – the selection and the assignment of relevance to data – solely
from a techno-deterministic point of view (Häußling, 2014), an ethical discussion should ana-
lyze technology in interdependence with human actors and as a part of collaborative activities.
316
At the same time, this involves moving away from Gidden´s three-level model of action (1984)
and Latour´s (1987, 2005) concept of agency (Rammert, 2008: 10). For this reason, the chapter
focuses on the concept of “distributed agency” (Rammert and Schulz-Schaeffer, 2002). This
approach helps to frame the action independently from the nature of the agent – whether
human or technical – and to classify its action into various levels. It permits the ethical questions
of responsibility to be posed in empirical research. From a technological perspective, agents are
specific computer programs:
That means that actions are delegated to them. The agents divide and delegate the
action among other agents. They cooperate with one another, thereby moving, being
initiative and addressing others. They coordinate the cooperation themselves and com-
municate the result of their activities to the human user.
(Rammert, 2008: 4)
Agency is here conceptualized as a three-level process to theoretically analyze the interaction

levels between human actors and algorithms as advanced technologies show signs of increased
self-activity, as “they are written with the intention that software agents can execute actions like
human agents” (Ibid.). This concept enables the discussion about the level, the linking, and the
attribution of agency to journalistic actors and computational processes in relation to the techni-
cal development and the status of the social discourse (Rammert and Schulz-Schaeffer, 2002: 43;
Rammert, 2007, 2008). Within this concept of distributed agency, the simplest form of action is
to cause a difference (level 1: causality). This is similar to Latour’s term “actant” or Callon’s term
“translation” (Callon, 1986; Latour, 2005). On this level, it does not make any difference whether
humans, machines, or programs execute the action.
The second level focuses on the action in which the actor has the space and capacity to act
differently and to choose between options (level 2: contingency). When technologies reach this
level of contingency, they no longer follow the mathematical paradigm of command and execu-
tion. For example, dialogical interfaces and internal user modeling increase this level of action
(Rammert, 2008: 11). Within this concept, there are different degrees of freedom. On level 2, the
spectrum ranges from the selection of predefined routines of software to self-adapting learning
algorithms.
Rammert and Schulz-Schaeffer (2002) also identify an even more complex form of action, in
which the actors can reflect their actions and add meaningful explanation (level 3: intentional-
ity) (see Table 23.1). Therefore, the question of which level of agency is present depends on the
interactive relationship between human actors and the software (Rammert and Schulz-Schaeffer,
2002: 50).
Table 23.1 Levels and grades of agency (c.f. Rammert and Schulz-Schaeffer, 2002; Rammert, 2008: 11)
317
Konstantin Dörr
Implications for journalism

The theoretical concept of distributed agency permits the inclusion of the role of technology as
an equal actor in news production. But without specific technical knowledge about the research
object, it is difficult to determine the potential and ethical impact of algorithmic applications
(Rath, 2006). Algorithmic processes are often invisible – described as a black box (Diakopoulos,
2015). Therefore, it is necessary to deconstruct these processes for algorithmic applications in
journalism to provide ethical implications. Most of the computational processes include priori-
tization, classification, association, and filtering (Diakopoulos, 2016). This approach is a part of
studies in technology assessment (Guston and Sarewitz, 2002). On a technical level, the concept
of reverse engineering, the “process of extracting the knowledge or design blueprints from
anything man-made”, is valuable (Eilam, 2005: 3; Diakopoulos, 2015: 404). This deconstruction
into input-throughput-output processes of algorithmic applications can lead to decision trees
that help to visualize decision-making rules and to apply an ethical analysis. These decision trees
are prevalent in business administration, informatics, and artificial intelligence research but less
frequently used in journalism (Bendel, 2015). However, the complexity of “decision-making
structures can quickly exceed the human and organizational resources available for oversight”
(Mittelstadt et al., 2016: 7; Kitchin, 2016). Therefore, it is necessary to embed computational
thinking as an aspect of news production as well as to create awareness of the use of software and
computational processes and emerging ethical challenges.
Furthermore, the analysis of ethical challenges in computational journalism should not only
focus on professional journalists. IT specialists, project managers, information architects, product
developers, and programming technicians as well as systems designers are already an essential part
of news production and the ethical debate, as “a designer must envision the design, the algo-
rithms, and the interfaces in use so that technical decisions do not run at odds with moral values”
(Friedman and Nissenbaum, 1996: 344). For example, the “ethical protocols design-approach”
can help to fulfill ethical requirements in media production, as it implements normative con-
straints in autonomous systems (Turilli, 2007). But this complexity also increases the demands
placed on media makers and media owners, as they must guarantee journalistic services in line
with ethical and normative standards. Alongside journalistic actors, this responsibility is to be
extended to data suppliers, software companies, and programmers, as “[a]utomated systems must
be designed with transparency and accountability as their primary objectives, so as to prevent
inadvertent and procedurally defective rule-making” (Citron, 2007: 1308). It is the combination
of empirical approaches to deconstructing algorithmic applications with the concept of distrib-
uted agency that enables a discussion of different levels of agency and the weighting of ethical
challenges. It can help to clarify which actor – whether man or machine – has a certain degree of
agency at a specific level in media production. This can be the deconstruction of a search query
or the process for automated text generation. It can clarify the following:
• who is responsible for a specific action,

• for the consequences of the action,
• who is affected,
• and at which level.
(Debatin, 1998: 117)
This hierarchy is characterized and influenced by different and often contradictory structural
conditions of the media system, e.g., economic and technological pressure. With a theoretical
framework, it is possible to identify and reorder the levels of responsibility before, during, and
after production regarding the impact of technology on journalism (Dörr and Hollnbuchner,
318
2016). However, it is difficult to attribute responsibility to specific individual actors. News pro-
duction is a complex and often multi-level process including several actors. An example of
journalistic action is the gap between the decision-making editorial team, the journalist, and
the responsible editor. Now, data suppliers, programmers, and autonomous software complicate the
differentiation even more. Therefore, it is necessary to attribute responsibility depending on the
specific technical application, the level of agency, and the specific setting in news production.
Outlook
This chapter is an attempt to extend the ethical discussion of responsibility in computational jour-
nalism to approaches of the sociology of technology. The theoretical concept of distributed agency
allows the distinct interplay of and tension between humans and technology in news production to
be framed. It also helps to include different actors in the ethical analysis. As it is unclear what will
happen if normative demands are changing and programmers, software, or data provider – none
socialized in professional journalism – start to play an even more important role in news production,
it is necessary for the impact of algorithmic applications, new actors, and their increasing influence
to be embedded within an analysis of how to govern these powers adequately. Recently, Saurwein
et al. (2015) identified and framed several governance options for algorithmic applications rang-
ing from market mechanisms to command and control regulation by state authorities as well as
alternative modes of governance including self-organization of individual companies and collective
self-regulation by industry branches and co-regulation – a regulatory cooperation between state
authorities and the industry. The governance of algorithmic applications in journalism includes,
for example, questions of manipulation, bias, censorship, the violation of privacy rights, or social
discrimination (Saurwein et al., 2015: 38). With technical developments challenging journalism on
a daily basis, it is important to frame the ethical debate within these regulatory choices. First and
foremost, the main task is to restore the public’s declining trust in journalism (Henry, 2007). To
build trust in algorithmic applications, it is necessary to show how the data are collected, where they
come from, to reveal parts of the code, and to indicate which actors took part in the programming
(Diakopoulos, 2016). It is necessary to assess the technical innovations in journalism and their spe-
cific impact on society, and ethicists, media experts, technicians, and journalists as well as regulators
will play a significant part in this discussion. The variety of structural and ethical challenges facing
journalism can only be solved jointly. To name a few tools for confidence building in professional
journalism: media organizations should build up and strengthen organizational and institutional
committees of (self)control, for example, calling for a digital press council as a self-regulatory mea-
sure. They should adapt and renew their ethical guidelines for the new digital challenges and visu-
alize computational power structures to ensure transparency and consequently to build trust. They
should collect data on their own to guarantee the highest standards of journalistic quality. If possible,
computational tools with an impact on the content and distribution of news should be developed
in-house. They should integrate the audience in the ethical discussion, e.g., for crowd-based appli-
cation development and data collection (open code and open data).
Further reading
Although it is regrettably available only in German, I recommend Handbuch Medien- und Informa-
tionsethik (Stuttgart: J.B. Metzler, 2016), edited by Jessica Heesen, for a broad overview of various
ethical approaches and the present challenges of technology and data in journalism. To investi-
gate the influence of technology on society, I recommend The Social Construction of Technological
Systems: New Directions in the Sociology and History of Technology (Cambridge, MA, 2012) edited by
Bijker, W.E., Hughes, T P., Pinch, T., and Douglas, D.G.
319
Konstantin Dörr
Note: Here and elsewhere, this chapter draws on material published in Dörr and Hollnbuch-
ner (2016) and Dörr et al. (2017).
References
Acemoglu, D. and Restrepo, P. (2016) “The race between machine and man: Implications of technology
for growth, factor shares and employment.” National Bureau of Economic Research, 22252. Retrieved from
www.nber.org/papers/w22252
Adam, A. (2005) “Delegating and distributing morality: Can we inscribe privacy protection in a machine?”
Ethics and Information Technology, 7(4), 233–242.
Society, 15(7), 1005–1021.
Anderson, M. and Anderson, S. L. (eds.). (2011) Machine Ethics. Cambridge: Cambridge University Press.
Arntz, M., Gregory, T. and Zierahn, U. (2016) “The risk of automation for jobs in OECD countries: A
comparative analysis.” OECD Social, Employment and Migration Working Papers, 189. doi:10.1787/5jlz9h
56dvq7-en
Autor, D. H. (2015) “Why are there still so many jobs? The history and future of workplace automation.”
Journal of Economic Perspectives, 29(3), 3−30.
Bendel, O. (2015) “Einfache moralische Maschinen: Vom Design zur Konzeption.” In T. Barton et al. (eds.),
Prozesse, Technologie, Anwendungen, Systeme und Management. Angewandte Forschung in der Wirtschafts-
informatik: Hochschule Luzern (pp. 171–180).
Bradshaw, P. (2014) “Data journalism.” In L. Zion and D. Craig (eds.), Ethics for Digital Journalists: Emerging
Best Practices. New York, NY: Routledge (pp. 202–219).
Brosda, C. (2010) “Journalismus.” In C. Brosda and C. Schicha (eds.), Handbuch Medienethik. Wiesbaden:
VS Verlag für Sozialwissenschaften (pp. 257–277).
Brosda, C. and Schicha C. (2010) Handbuch Medienethik. Wiesbaden: VS Verlag für Sozialwissenschaften.
Bucher, T. (2012) “Want to be on the top? Algorithmic power and the threat of invisibility on Facebook.”
New Media & Society, 14(7), 1164–1180.
Buhl, H. U., Röglinger, M., Moser, F. and Heidemann, J. (2013) “Big data: A fashionable topic with(out)
sustainable relevance for research and practice?” Business & Information Systems Engineering, 5(2), 65–69.
Callon, M. (1986) “Some elements of a sociology of translation: Domestication of the scallops and the
fishermen of St. Brieux Bay.” In J. Law (ed.), Power, Action, and Belief. London: Routledge (pp. 196–229).
Carlson, M. (2014) “The robotic reporter. Automated journalism and the redefinition of labor, composi-
Caswell, D. and Dörr, K. (2017) “Automated journalism 2.0: Event-driven narratives. From simple descrip-
tions to real stories.” Journalism Practice, 1–21 [online first].
Citron, D. K. (2007) “Technological due process.” Washington University Law Review, 85, 1249–1313.
Coddington, M. (2015) “Clarifying journalism’s quantitative turn. A typology for evaluating data journal-
Collins, H. M. and Kusch, M. (1998) The Shape of Actions: What Humans and Machines Can Do. Cambridge,
MA: The MIT Press.
Culver, K. B. (2016) “Disengaged ethics. Code development and journalism’s relationship with ‘The Pub-
lic’.” Journalism Practice, 11(4), 477–492.
Debatin, B. (1998) “Verantwortung im Medienhandeln. Medienethische und handlungstheoretische Über-
legungen zum Verhältnis von Freiheit und Verantwortung in der Massenkommunikation.” In W. Wun-
den (ed.), Freiheit und Medien. Beiträge zur Medienethik. Frankfurt a.M.: Abt. Verlag (pp. 113–130).
Diakopoulos, N. (2010) A Functional Roadmap for Computational Journalism. Retrieved from www.nickdiako
poulos.com/wp-content/uploads/2007/05/CJ_Whitepaper_Diakopoulos.pdf
Diakopoulos, N. (2016) “Accountability in algorithmic decision making.” Communications of the ACM,
59(2), 56–62.
Diakopoulos, N. (2017) “Computational journalism and the emergence of news platforms.” In S. A.
Eldridge II and B. Franklin (eds.), The Routledge Companion to Digital Journalism Studies. London: Rout-
ledge (pp. 176–184).
320
Diakopoulos, N. and Koliska, M. (2016) “Algorithmic transparency in the news media.” Digital Journalism,
5(7), 809–828.
Dolata, U. and Werle, R. (2007) “Bringing technology back in.” In U. Dolata and R. Werle (eds.), Gesellschaft
und die Macht der Technik. Sozioökonomischer und institutioneller Wandel durch Technisierung. Frankfurt a.M.:
Campus (pp. 15–43).
Domingo, D. and Paterson, C. (2011) Making Online News: Newsroom Ethnographies in the Second Decade of
Internet Journalism. New York, NY: Peter Lang.
Dörr, K. (2016a) “Algorithmen, big data und ihre Rolle im computational journalism.” In K. Meier and
C. Neuberger (eds.), Journalismusforschung. Stand und Perspektiven. Baden-Baden: Nomos (pp. 245–264).
Dörr, K. (2016b) “Mapping the field of algorithmic journalism.” Digital Journalism, 4(6), 700–722.
Dörr, K. and Hollnbuchner, K. (2016) “Ethical challenges of algorithmic journalism.” Digital Journalism,
5(4), 404–419.
Dörr, K., Köberer, N. and Haim, M. (2017) “Normative Qualitätsansprüche an algorithmischen Journal-
ismus.” In I. Stapf, M. Prinzing and A. Filipovi (eds.), Gesellschaft ohne Diskurs? Digitaler Wandel und
Journalismus aus medienethischer Perspektive. Kommunikations- und Medienethik. Baden-Baden: Nomos (pp.
121–131).
Eilam, E. (2005) Reversing: Secrets of Reverse Engineering. Indianapolis, IN: Wiley.
ism Practice, 6(2), 157–171.
Friedman, B. and Nissenbaum, H. (1996) “Bias in computer systems.” ACM Transactions on Information
Systems, 14(39), 330–347.
Gandomi, A. and Haider, M. (2015) “Beyond the hype: Big data concepts, methods, and analytics.” Interna-
tional Journal of Information Management, 35(2), 137–144.
Giddens, A. (1984) The Constitution of Society: Outline of the Theory of Structuration. Cambridge: Polity Press.
Gillespie, T. (2014) “The relevance of algorithms.” In T. Gillespie, P. Boczkowski and K. Foot (eds.), Media
Technologies: Essays on Communication, Materiality, and Society. London: The MIT Press (pp. 167–194).
Gray, J., Bounegru, L. and Chambers, L. (eds.). (2012) The Data Journalism Handbook: How Journalists Can
Use Data to Improve the News. Sebastopol, CA: O’Reilly Media.
Grunwald, A. (2016) “Technikethik.” In J. Heesen (ed.), Handbuch Medien- und Informationsethik. Stuttgart:
J. B. Metzler (pp. 25–32).
Guston, D. H. and Sarewitz, D. (2002) “Real-time technology assessment.” Technology in Society, 24(1),
93–109.
Häußling, R. (2014) Techniksoziologie. Baden-Baden: Nomos.
Heise, N. (2016) “Algorithmen.” In J. Heesen (ed.), Handbuch Medien- und Informationsethik. Stuttgart: J. B.
Metzler (pp. 202–209).
Henry, N. (2007) American Carnival: Journalism Under Siege in an Age of New Media. Berkeley, CA: O’Reilly
Media.
Johnson, D. G. (1985, 2009) Computer Ethics. London: Pearson.
Just, N. and Latzer, M. (2016) “Governance by algorithms: Reality construction by algorithmic selection on
the Internet.” Media, Culture & Society, 39(2), 238–258.
Kant, I. (1788) Kritik der praktischen Vernunft. Riga: Hartknoch.
Kitchin, R. (2016) “Thinking critically about and researching algorithms.” Information, Communication &
Society, 20(1), 14–29.
Köberer, N. (2014) Advertorials in Jugendprintmedien. Ein medienethischer Zugang. Wiesbaden: VS Verlag für
Sozialwissenschaften.
Kraemer, F., Van Overveld, K. and Peterson, M. (2011) “Is there an ethics of algorithms?” Ethics and Informa-
tion Technology, 13(3), 251–260.
Kroes, P. and Verbeek, P. (2014) The Moral Status of Technical Artefacts. Wiesbaden: Springer.
Latour, B. (2005) Reassembling the Social: An Introduction to Actor-network-theory. Oxford: Oxford University
Press.
Latzer, M. (2013) “Medienwandel durch Innovation, Ko-Evolution und Komplexität.” Medien & Kommuni-
kationswissenschaft, 61(2), 235–252.
Latzer, M., Hollnbuchner, K., Just, N. and Saurwein, F. (2016) “The economics of algorithmic selection on
the Internet.” In J. Bauer and M. Latzer (eds.), Handbook on the Economics of the Internet. Cheltenham:
Edward Elgar (pp. 395–425).
Lessig, L. (1999) Code and Other Laws of Cyberspace. New York, NY: Basic Books.
321
Konstantin Dörr
Lewis, S. C. (2017) “Digital journalism and big data.” In S. A. Eldridge II and B. Franklin (eds.), The Rout-
ledge Companion to Digital Journalism Studies. London: Routledge (pp. 126–136).
Lewis, S. C. and Westlund, O. (2015) “Actors, actants, audiences, and activities in cross-media news work.”
Matsuzaki, H. (2011) “Die Frage nach der ‘Agency’ von Technik und die Normenvergessenheit der Tech-
niksoziologie.” In N. Lüdtke and H. Matsuzaki (eds.), Akteur – Individuum – Subjekt: Fragen zu ‘Personal-
ität’ und ‘Sozialität’. Wiesbaden: VS Verlag (pp. 301–325).
Mayer-Schönberger, V. and Cukier, K. (2013) Big Data: A Revolution that Will Transform How We Live, Work,
and Think. London: John Murray Publishers.
Mayo, J. and Leshner, G. (2000) “Assessing the credibility of computer-assisted reporting.” Newspaper
McBride, K. and Rosenstiel, T. (2014) “New guiding principles for a new era of journalism.” In K. McBride
and T. Rosenstiel (eds.), The New Ethics of Journalism. Thousand Oaks, CA: CQ Press (pp. 1–6).
Mitchelstein, E. and Boczkowski, P. J. (2009) “Between tradition and change: A review of recent research
on online news production.” Journalism, 10(5), 562–586.
Mittelstadt, B. D., Allo, P., Taddeo, M., Wachter, S. and Floridi, L. (2016) “The ethics of algorithms: Mapping
the debate.” Big Data & Society, 3(2), 1–21.
Montal, T. and Reich, Z. (2016) “I, Robot. You, Journalist. Who is the Author?” Digital Journalism, 5(7),
829–849.
Moor, J. H. (2006) “The nature, importance, and difficulty of machine ethics.” IEEE Intelligent Systems,
21(4), 18–21.
Neuberger, C. (2009) “Internet, Journalismus und Öffentlichkeit. Analyse des Medienumbruchs.” In C.
Neuberger, C. Nuernbergk and M. Rischke (eds.), Journalismus im Internet: Profession – Partizipation –
Technisierung. Wiesbaden: VS Verlag für Sozialwissenschaften (pp. 19–105).
Neuberger, C. and Nuernbergk, C. (2015) “Verdatete Selbstbeschreibung der Gesellschaft – Über den
Umgang des Journalismus mit Big Data und Algorithmen.” In F. Süssenguth (ed.), Die Gesellschaft der
Daten. Über die digitale Transformation der sozialen Ordnung. Bielefeld: Transcript (pp. 199–224).
Pavlik, J. (2017) “Data, algorithms, and code: Implications for journalism practice.” In S. A. Eldridge II
and B. Franklin (eds.), The Routledge Companion to Digital Journalism Studies. London: Routledge (pp.
265–273).
Powers, M. (2012) “In forms that are familiar and yet-to-be invented: American journalism and the dis-
course of technologically specific work.” Journal of Communication Inquiry, 36(1), 24–43.
Rammert, W. (2003) “Technik in Aktion: Verteiltes Handeln in soziotechnischen Konstellationen.” TUTS
Working Papers, 2, 1–23.
Rammert, W. (2007) Technik – Handeln – Wissen. Zu einer pragmatistischen Technik- und Sozialtheorie. Wies-
baden: VS Verlag für Sozialwissenschaften.
Rammert, W. (2008) “Where the action is: Distributed agency between humans, machines, and programs.”
TUTS Working Papers, 4, 1–29.
Rammert, W. and Schulz-Schaeffer, I. (2002) “Technik und Handeln. Wenn soziales Handeln sich auf
menschliches Handeln und technische Abläufe verteilt.” In W. Rammert and I. Schulz-Schaeffer (eds.),
Können Maschinen handeln? Soziologische Beiträge zum Verhältnis von Mensch und Technik. Frankfurt a.M.:
Campus (pp. 11–64).
Rath, M. (2006) “Wahrhaftigkeit des Journalismus: Moralanspruch oder Marktfaktor?” In H. Niesyto, M.
Rath and H. Sowa (eds.), Medienkritik heute. Grundlagen, Beispiele und Praxisfelder. München: Kopaed (pp.
117–128).
Resnick, P. and Varian, H. R. (1997) “Recommender systems.” Communications of the ACM, 40(3), 56–58.
Richter, P. (2016) “Big data.” In J. Heesen (ed.), Handbuch Medien- und Informationsethik. Stuttgart: J. B.
Metzler (pp. 210–216).
Rieder, B. (2005) “Networked control: Search engines and the symmetry of confidence.” International Review
of Information Ethics, 3, 26–32.
Rip, A. (2002) Co-evolution of Science, Technology and Society. Expert Review, Bundesministerium Bildung und
Forschung, Förderinitiative Politik, Wissenschaft und Gesellschaft (Science Policy Studies).
Ropohl, G. (1995) “Die Dynamik der Technik und die Trägheit der Vernunft.” In H. Lenk and H. Poser
(eds.), Neue Realitäten – Herausforderung der Philosophie. Berlin: Akademie Verlag (pp. 102–119).
Saurwein, F., Just, N. and Latzer, M. (2015) “Governance of algorithms: Options and limitations.” Info, 17(6),
35–49.
322
Shulman, C., Jonsson, H. and Tarleton, N. (2009) “Machine ethics and superintelligence.” In C. Reynolds
and A. Cassinelli (eds.), AP-CAP 2009: The Fifth Asia-Pacific Computing and Philosophy Conference, 1–2
October, University of Tokyo, Japan, Proceedings AP-CAP 2009 (pp. 95–97).
Tavani, H. H. (2011) Ethics and Technology. Controversies, Questions, and Strategies for Ethical Computing. Hobo-
ken, NJ: Wiley.
Thurman, N., Dörr, K. and Kuntert, J. (2017) “When reporters get hands-on with robo-writing. Pro-
fessionals consider automated journalism’s capabilities and consequences.” Digital Journalism, 5(10),
1240–1259.
Turilli, M. (2007) “Ethical protocols design.” Ethics and Information Technology, 9(1), 49–62.
Van Dalen, A. (2012) “The algorithms behind the headlines. How machine-written news redefines the core
Van Otterlo, M. (2013) “A machine learning view on profiling.” In M. Hildebrandt and K. de Vries (eds.),
Privacy, Due Process and the Computational Turn-Philosophers of Law Meet Philosophers of Technology. Abing-
don: Routledge (pp. 41–64).
Vos, T. P. and Heinderyckx, F. (2015) Gatekeeping in Transition. New York, NY: Routledge.
Wallach, W., Allen, C. and Smit, I. (2008) “Machine morality: Bottom-up and top-down approaches for
modelling human moral faculties.” AI & Society, 22(4), 565–582.
Ward, S. (2017) “Digital journalism ethics.” In S. A. Eldridge II and B. Franklin (eds.), The Routledge Com-
panion to Digital Journalism Studies. London: Routledge (pp. 35–43).
Weischenberg, S. and Hienzsch, U. (1994) “Die Entwicklung der Medientechnik.” In K. Merten, S. J.
Schmidt and S. Weischenberg (eds.), Die Wirklichkeit der Medien. Eine Einführung in die Kommunikation-
swissenschaft. Opladen: Westdeutscher Verlag (pp. 455–480).
Werle, R. (2002) “Technik als Akteurfiktion.” In W. Rammert and I. Schulz-Schaffer (eds.), Können Mas-
chinen handeln? Frankfurt: Campus (pp. 119–139).
Werle, R. (2007) “Zur Interdependenz von Innovationen.” In H. Hof and U. Wengenroth (eds.), Innova-
tionsforschung. Hamburg: LIT (pp. 31–40).
Winner, L. (1986) The Whale and the Reactor. Chicago, IL: University of Chicago Press.
Wunden, W. (1999) “Freiheitliche Medienmoral. Konzept einer systematischen Medienethik.” In R. Funiok,
U. Schmälzle and C. Werth (eds.), Medienethik – Die Frage der Verantwortung. Bonn: Bundeszentrale für
politische Bildung (pp. 35–55).
Young, M. L. and Hermida, A. (2015) “From Mr. and Mrs. Outlier to central tendencies.” Digital Journal-
ism, 3(3), 381–397.
Zikopoulos, P., de Roos, D., Parasuraman, K., Deutsch, T., Corrigan, D. and Giles, J. (eds.). (2013) Harness
the Power of Big Data: The IBM Big Data Platform. New York, NY: McGraw-Hill.
Zion, L. and Craig, D. (2014) Ethics for Digital Journalists: Emerging Best Practices. New York, NY: Routledge.
Zwitter, A. (2014) “Big data ethics.” Big Data & Society, 1(2), 1–6.
323
24
WHO OWNS THE NEWS? THE
‘RIGHT TO BE FORGOTTEN’ AND
JOURNALISTS’ CONFLICTING
PRINCIPLES
‘The right to be forgotten’ (RTBF) is a relatively new concept in human-rights law, but it deals
in root ethical issues familiar to news people and their sources. Editors must routinely weigh the
news’ long-term role as a ‘historical record’ against its potential negative impacts on individuals.
In the digital journalism era, publication is at the same time both more enduring and less static,
creating new parameters and possibilities for ethical decision-making. Because news content may
be seen by more people in more places for much longer, the potential to do lasting good or harm
is greater, but because digital publication is more retractable and redactible than legacy platforms,
the possibility of correction, clarification, and removal creates both new harm-reduction oppor-
tunities and new challenges to the historical record. Also known as a ‘right to erasure’ or ‘right
to oblivion’, the RTBF, now accepted in the European Union, recognizes that, even in the age of
Google, people should retain some degree of control over information about themselves and their
pasts. (Factsheet on the ‘Right to be Forgotten’ ruling (C131–12), n.d.; Manna, 2014; Rosen, 2012).
This chapter will explore both legal and ethical implications of the issue.
Birth of a legal right

Data-protection regimes in Europe stem from a tradition, rooted in the Napoleonic Civil Code
and bolstered by a visceral post-Fascist resistance to surveillance, of legally protecting autonomy
over information about “the events of an individual life, both private and public” (Mantelero,
2013; Toobin, 2014). The European approach emphasizes human individual dignity, with cor-
responding rights and duties, whereas the common law treats privacy as a form of property right
(Eltis, 2011). For example, the common law tort of ‘public disclosure of embarrassing private
facts’ by definition is not intended to protect a potential plaintiff against disclosure of facts that
are in the public domain, such as a past criminal charge (Prosser, 1960).
Among common law countries, historically only the United States has given broad recogni-
tion to the tort of invasion of privacy, but it is balanced there by strong First Amendment rights
for free speech and relaxed defamation laws. This tort in its various forms is becoming more
accepted elsewhere in the common law, but Continental Europe has long exhibited a much
greater degree of respect for private matters, even for those in public life. For example, it was a
324
Who owns the news?
breach of criminal law to publish that the French president had an illegitimate daughter by a
long-standing mistress, as was subsequently revealed about François Mitterand (Baume, 2012).
This restriction on a free press would have been a nonstarter in common law jurisdictions.
Article 12 of the European Community’s 1995 Data Protection Directive allowed a person to
require the “rectification, erasure or blocking” of unnecessary personal data (“Directive 95/46/
EC” 1995). As the European Community considered amendments to this directive – a process
begun in 2012 and concluded in 2016 (European Commission – Justice, 2016) – the contentious
notion of RTBF emerged as a more robust recognition of individuals’ desire to control their own
information. False and defamatory statements could already be challenged through defamation
laws, but not truthful facts – unless they crossed an unclear boundary into someone’s “private
and family life” (European Convention on Human Rights, Article 8). The new idea of extend-
ing control over personal information was relatively easy to accept for information posted by the
person on a public website, for example, especially for a youth. Similarly, there was little debate
about third parties being held responsible for using personal information other than in the man-
ner intended and agreed to by the individual. But proponents took the notion a step farther to
encompass information that may have been appropriately published by others about the person.
Added to this respect for privacy interests came the notion of ‘redemption’, allowing people to
put their mistakes, even criminal ones, behind them in order to start anew – a principle that has
been accepted in many jurisdictions, but most commonly for young offenders, who are allowed
to escape the stigma of youthful indiscretions and crimes.
The seams between European and English legal cultures came into focus when English courts
sought to apply the Convention on Human Rights to privacy concerns under common law,
going beyond traditional breach-of-confidence case law to create a new tort for “misuse of pri-
vate information” (Campbell v. MGN Limited [2004 UKHL 22], 2004; Google Inc v Vidal-Hall &
Ors [2015] EWCA Civ 311, 2015; PJS v News Group Newspapers Ltd [2016] UKSC 26, 2016).
The key issue for this tort is whether the disclosure amounts to a breach of one’s “reasonable
expectation of privacy”. While this is not a subjective test, it is certainly one over which people
differ, particularly as between national and cultural backgrounds. It directly calls into play bound-
aries for free expression, also protected under the convention (Article 10), and requires a balanc-
ing of these rights – particularly where issues of public interest are involved.
Such a balance would become even tougher to find as information became more accessible,
thanks to Google. “God forgives and forgets but the Web never does”, noted the European
commissioner when calling for rules “to better cope with privacy risks online” (Reding, 2010).
This created the basis for the European Commission’s proposal for a RTBF that went beyond a
“right to deletion or erasure” to what has been called “a right to oblivion” – that is, rather than
applying only to self-published content or imposing acceptable time or usage limits on private
data, access would now be blocked to information considered harmful and outdated even though
true (Ambrose, 2014; McGoldrick, 2013; Xanthoulis, 2013). In 2014, however, the European
Parliament trimmed the scope of the RTBF included in Article 17 of the General Data Protec-
tion Regulation to restrict its primary focus to the “right to erasure” (European Parliament, 2014).
It was against this background that the Grand Chamber of the Court of Justice of the Euro-
pean Union (CJEU) issued a dramatic ruling in 2014, applying the 1995 directive to Google
search results after the applicant had been unable to get a local newspaper to remove appropri-
ately published articles from its own database (InfoCuria, 2014).
The CJEU upheld a Spanish court’s order that Google Spain should remove links to 1998
reports in the Catalonian newspaper La Vanguardia about the forced sale of real estate in attach-
ment proceedings against a Spanish man, Mario Costeja González, to settle his social security debts.
The CJEU ruled that Google was a “controller” of personal data governed by the 1995 directive
325
and that the continued “processing” of personal data by search engine operators, and the resulting
ubiquity of search results, was “liable to affect significantly the fundamental rights to privacy and to
the protection of personal data” when the individual’s name was searched (Google Spain, para. 80).
The new right is by no means absolute; rather, the court mandated a three-way analysis, taking
into account the rights of the subject, the economic freedom of the search engine company, and
the general public’s legitimate interest in information, particularly where the subject may play a
role in public life (Google Spain, para. 99). The court further suggested that information might
have reduced significance as time goes on (para. 93), giving news, in effect, a ‘best-before’ date,
like yogurt.
Significantly, the court drew a clear distinction between search engines and website publishers
with “solely . . . journalistic purposes”. Without a search engine listing, the item on the original
news web page becomes practically obscure, less likely to be stumbled upon, and thus less harm-
ful. In 2016, courts in Belgium and Italy ordered RTBF alterations to news archives, but at time
of writing, these rulings apply only to their respective jurisdictions (Matthews, 2016; Tomlinson,
2016).
In the wake of the Grand Chamber’s decision, Google faced a deluge of requests for removal,
and, after consulting a panel of expert advisors, it moved quickly to develop principles and prac-
tices for response (Lomas, 2014; Report of the Advisory Council to Google on the Right to be
Forgotten, 2015; Toobin, 2014). By May 2017, Google had received 719,327 RTBF requests and
delisted 43.1% of the URLs evaluated as a result, although these URLs are, at time of writing,
hidden only for searches conducted “from the country of the person requesting the removal”
(Google Transparency Report, 2017).
In the United States, privacy concerns continued to be trumped by freedom of speech (Palm
Beach Newspapers v. State of Florida et al., 2016). However, with pressure growing for European
courts to impose worldwide limits that would apply to Google and other search engines subject
to the courts’ authority, Google preemptively decided to apply court orders in Europe to its
worldwide algorithms, though only for searches initiated within the affected country (Keller,
2016). This step failed to satisfy the French regulator, but Google, arguing that France was impos-
ing its rules on citizens of other countries, launched an appeal. This issue will clearly require
involvement of the Court of Justice of the European Union, which will occur sometime after
this writing (Conseil d’Etat, 2017; Dong, 2016; Roberts, 2015).
Some media organizations responded assertively to RTBF deletions either by publishing the
fact of these deletions or republishing items that had been deleted (Lee, 2015; “Pink News
Republishes Stories Removed” 2016). While further appeals to data protection agencies can be
undertaken, as occurred in the Costeja case, the volume of requests has meant that Google itself
has become the key arbiter in enforcing the law. Nor has Europe’s 2016 data-protection regula-
tion clarified whether social media platforms such as Facebook and Twitter will be subject to
future RTBF orders (European Commission – Justice, 2016; Keller, 2015).
Nevertheless, the RTBF seed has been firmly planted in European law, with seemingly inevi-
table international repercussions and implications for journalistic practice, as we will show in the
remainder of this chapter.
A clash of ethical principles

The roots of all law lie in ethics: legislation and the common law codify a society’s perceived
consensus on rights and wrongs, and courts then apply those principles to life-specific situations.
In other words, the law seeks to codify and enforce a society’s consensus on values and mediate
conflicting values. Señor Costeja, for example, was far from the first person to express a desire
to assume some control of information about his past life, and news people are familiar with the
326
Who owns the news?
ethical essence of the Costeja quandary – a conflict between the duty to provide a lasting record
and the expression of ordinary humanity in mitigating harm.
The “right to privacy” clearly implies a reciprocal duty for people not to know certain types
of information about one another, or, if they do, to resolve pragmatically to ‘forget’ it – a deci-
sion that is increasingly difficult to apply in the digital age (Matheson, 2013: 197, 202; see also
Mayer-Schönberger, 2009). While the rules of court proceedings routinely expect judges and
juries to ‘forget’ information not in evidence and that fiduciary duty can also apply to others (for
example, teachers who suspect a student of cheating but lack solid evidence), it is another thing
altogether to impose a duty of forgetfulness on society at large through selective erasure of its
digital memory bank.
The moral right at the heart of this quandary is more accurately expressed as a right to
obscurity (Hartzog and Selinger, 2015) or, as styled by some Italian courts, oblivion – “the right
of any individual to see himself or herself represented in a way that is not inconsistent with his/
her current personal and social identity” (Manna, 2014). Even if total control of one’s universal
online profile were desirable in public policy – allowing for digital oblivion – it would be an
unrealistic goal (Fazlioglu, 2013; Rosenzweig, 2012). More realistic is the idea of relative obscu-
rity, a circumstance-limited right that, like freedom to express oneself, derives from autonomy:
the right to make the choices that drive one’s life, including whether to draw attention to oneself
or to remain quiet (Ausloos, 2012; Baker, 2004; LaRue, 2011). Similarly, the jurisprudence on
both defamation and hate speech asserts the importance of people’s ability to retain control of
their own lives or to participate in democratic society. Individuals’ rights (such as the expression
of provocative opinions) are, famously, limited by the social contract itself: incitement to racial
violence, for example, is not protected. Autonomy rights are, therefore, limited by the conflicting
attributes of a situation; for example, someone who may normally enjoy private-citizen status
may later become a public figure, making information about his or her past important to the
public interest.
‘Unpublishing’ and the right to privacy

The relativistic nature of autonomy rights has important practical limitations with respect to
availability of information. All the courts that broke ground in Costeja were careful to distinguish
the continued availability of news articles in news organizations’ online archives (Glasser, 2014;
Tomlinson, 2015). Likewise, journalists have traditionally resisted the idea of ‘unpublishing’ – the
retrospective redaction of error-free news reports.
In recent years, however, news organizations have received increasing numbers of requests to
adjust their digital records in this way, and there are signs that, rather than dismiss these out of
hand, many are seeking a way to achieve consistency and fairness (Brock, 2016: 87–88; Edmonds,
2016; Pantic, 2014: 18; Watson, 2012). As with Costeja, negative and long-past involvements with
the law represent a significant proportion of news sources’ attempts to refine their online identity,
a concern also evidenced by the sporadic success of unscrupulous website operators who post,
and then offer to delete for a fee, accused criminals’ mug shots or court reports (A.T. v. Globe24h.
com, 2017; Segal, 2013). Nor is it surprising or unreasonable that previously accused people try
to clean up the reporting of criminal charges retrospectively; crime reporting is notoriously epi-
sodic and often left unfinished in the public record (Andrews, 2014; Kauth, 2015).
Editors have responded to such requests in a variety of ways, depending on both inclination
and circumstances – these forms of response include wholesale granting or refusal of the request,
removing a source’s name while leaving the article otherwise intact, inserting an addendum or
correction, generating a follow-up story, and removing stories from Google’s cache (Tenore,
2010). However, North American editors surveyed by English in 2009 mostly rebuffed the need
327
for consistent policies on unpublishing, stating that news organizations “do not rewrite history;
we report what happened”, and, “Sorry, life isn’t fair. Journalism’s job isn’t to clean up your
driving record so you can get a job, is it?” They conceded that content should be removed online
under “very rare circumstances”, including where required by libel or other legal requirements.
“Serious consideration to an unpublishing request should also be given when someone’s life may
be endangered”, English concluded (English, 2009).
Five years after this survey, relatively few news organizations appeared to have implemented
comprehensive policies as to how to respond to these requests, preferring case-by-case adjudica-
tion of requests (Pantic, 2014: 52, 85). A panel report for the Canadian Association of Journalists’
ethics advisory committee revisited the committee’s earlier stated position that journalists “are
in the publishing business and generally should not unpublish”. The panel reconstructed a case
where a newspaper granted an unpublishing request and suggested that it might be time for jour-
nalists to “reassess their attribution practices”, with growing openness to offering anonymity, in
order to prevent long-term harm to people who volunteer information about “profoundly per-
sonal” situations (Currie and Brethour, 2014; English et al., 2010). More recently, there have been
signs of a growing appetite to try to codify an approach to resolving these requests (Edmonds,
2016). Some of these requests have become quite nuanced; in one instance, a person asked that
Canada’s CBC network merely remove his name from an online headline; the request was denied,
but the network’s ombud, clearly troubled, wrote that the unpublishing issue “requires nuanced
and ongoing consideration” (Enkin, 2016).
To describe the discussions around these issues as a direct clash between the right of free
expression and that of privacy is to dramatically oversimplify the questions involved. Rather,
applying a right to obscurity would involve weighing several social values, including freedom
of information and the integrity of historical research (Manna, 2014). Viewed under this more
nuanced lens, the arguments surrounding unpublishing expose no fewer than six ethical princi-
ples. Three of these principles would tend to foster continuity of publication, even of potentially
harmful material, while an equal number tend in the opposite direction.
First among the three pro-continuity principles is that free expression should be defended
(Weber, 2011). To allow people in the news to exert influence over what should be (or remain)
published flies in the face of the essence of press freedom, which rests to some extent on a news
culture of being “unconstrained by the long view or deep understanding”, an independence of
mind that mitigates such clearly undesirable constraints as dependency on “conventional wis-
dom” and on official and professional sources (Schudson, 2005: 24–26).
Second, publication continuity is supported by the idea that information in its original form
ought to be protected for the sake of historical integrity. This is an uncomfortable principle, and
not only insofar as it touches on news reports: why should an individual’s divorce-court affidavit
or contractual dispute become someone else’s business in the future? But for historians, these
documents provide insight into trends and issues – plus, character-revealing details about today’s
obscure individual could be of considerable public interest when that person assumes a more
public role in the future.
The third principle is that of accountability. Journalism is made credible by the verifiable and fal-
sifiable details that back up reports; for this reason, journalists tend to prefer named sources to veiled
ones (Gladney et al., 2013: 36–37; Vultee, 2010). Shafer is far from alone in advising readers to “dis-
count anything a shadowy unknown source is allowed to say in a news story” (Shafer, 2014). This
principle is undermined when a person’s name is removed from the online version of a news article.
Turning to principles that tend to constrain publication, the first and most familiar is the idea
of harm reduction. In media ethics codes, this idea has classically been applied to children and other
vulnerable subjects and to the right to a fair trial. It has guided practice in preventively veiling the
identity of whistle-blowers and sexual-assault victims. As parents watch their children’s digital
328
Who owns the news?
lives causing more than mere embarrassment, it may become harder to resist efforts to provide
people with means of erasing the public record – efforts that, once acquiring momentum, are
likely to sweep routine news reports into the same unsavory box as sex videos and careless tweets
(Ohm, 2015).
A second such principle is respect for privacy, which Moore defines as “a right to control
access to and uses of places, bodies, and personal information” (Moore, 2008: 421). News people
do not normally delve into adoption records, sexual matters, and personal financial records with-
out some clear public-interest justification. Indeed,
without strong privacy safeguards, it becomes far more difficult [. . .] for people to
exercise their human right to free expression. It is an established fact that when people
believe they’re being watched, their behaviour changes in very significant ways.
(Christopher, 2015)
In the discussions surrounding ‘big data’, an emerging principle is that the collection of data is
unobjectionable so long as the information is not individualized (Francis and Francis, 2014).
Search engines and news reports, however, individualize data almost by definition.
Less commonly associated with media ethics is the aforementioned idea of redemption. The
idea that people should generally be able to put past transgressions behind them and move for-
ward is expressed in the idiomatic idea of ‘forgive and forget’, and the law in some countries
allows people’s criminal records to be expunged after sufficient time has passed (Bennett, 2012:
166–167). Laws and regulations in the United States, for example, afford individuals the opportu-
nity to “move on” and get a “second chance” following bankruptcies, juvenile criminal behavior,
and credit reporting (Blanchette and Johnson, 2002: 4; Jones et al., 2012). Pervasive search results
butt against this principle.
How may these conflicting principles be reconciled? Because this task – weighing conflict-
ing principles against each other – is substantially similar in ethical reflection as in human rights
jurisprudence, it may be helpful to look again at the European Court’s argumentation in Costeja.
Two important distinctions

Stripped of legal foundations and arguments, the Grand Chamber’s decision rested on two episte-
mological distinctions that, in our view, may also help journalists to begin resolving the previously
described conflict between principles that tend to foster continuity and those that challenge it.
First, the court drew a clear distinction between truthfulness and ‘relevance’. While there is no
stale-date on accuracy, relevance is situational and, therefore, temporary. Often, this principle is
expressed with reference mainly to the passage of time. The idea of newsworthiness itself – starting
with the very word ‘news’ – may be seen as suggesting that old stories (yesterday’s news) are less
relevant than current ones (Potter, 2009: 5). After the Amsterdam Court of Appeal ruled that
negative publicity caused by a criminal offense continued to be ‘relevant’ information, the Dutch
Supreme Court, on appeal, held that reports of crimes were subject to the principles established
in Costeja (van den Brink, 2017). Yet, relevance is not adequately measured in days or years; it may
or may not diminish with the passage of time. Were a 40-year-old news report to be discovered
quoting a current political leader as expressing vile racist views, that country’s citizens might well
consider the old report ‘relevant’. Indeed, while timeliness is often cited as an element of ‘news-
worthiness’, additional elements are always present, often including prominence, proximity, and
probable impact (Schultz, 2007). Where U.S. courts have weighed in on an item’s ‘newsworthi-
ness’, they have tended to probe factors such as public concern, undue offensiveness, and whether
or not journalists have pursued it with morbid fascination or recklessness (McNealy, 2012).
329
The second important distinction is, as mentioned earlier, that between the mere availability
of information and its being instantly ‘findable’. A functionally infinite search-facilitator, such as
Google, imposes a level of scrutiny over past acts that seems, to some, less preferable to a “balance
between the protection of the individual memory and the rights to information for citizens”
(Salarelli, 2015).
When these distinctions are applied to unpublishing requests, some of the three pro-continuity
principles might wane in force.
Free speech significantly predates the existence of Google, and a news organization’s freedom to
choose what to report and publish also allows it to choose what not to publish – and, by exten-
sion, when to hide certain text from search engines for good reason. Even if not easily ‘findable’,
the original report may remain available digitally, either in PDF archives or on stored web page
archives such as the Wayback Machine at archive.org (Lepore, 2015), or through other creative
workarounds to conventional delisting (Renner, 2016). Likewise, compassionate deletion from a
news website and from Google’s cache should not require material to be unpublished from news-
papers’ databases. Unpublishing, in other words, need not involve an all-or-nothing decision.
Turning to accountability, the persuasiveness and credibility of a news report is primarily a
consideration with a view to contemporaneous readers; as time goes by, the reader will have alter-
native and, perhaps, more reliable means to get at the truth. Indeed, if the original report, seen
in isolation, provides a potentially distorted or inaccurate lens on the whole truth, then the idea
of ‘relevance’ – that is, practical usefulness to the reader – would favor the provision of a cleaner,
contextualized record. The ‘relevance’ principle hardly requires the outright deletion of a news
item, but the primary journalistic responsibility of seeking accuracy is widely acknowledged as
obliging a news organization to correct the facts in an original item’s digital manifestation. So,
at the very least, reports of criminal charges that led to acquittal should be updated, where the
accused is identified, in the name of proportionality and context (Brock, 2016: 88).
The argument for historical integrity is harder to shrug off. When the Toronto Star unpublished
a discredited article about vaccine safety in February 2015, the article did not merely disappear
from Google searches (Braganza, 2015); it disappeared from full-text library databases as well.
The disappearance of a published work from libraries sends a Big Brother-esque chill up the free-
information spine, and the test of ‘relevance’ may be fickle: information about a 25-year-old law
student’s conduct might seem highly relevant to informed public debate (the most critical pur-
pose of free expression itself) when he is later up for appointment as judge or attorney general.
These arguments suggest strong justification for adopting a reasonably fluid approach to cor-
rections and clarifications to the published record, with a view to ongoing care for accuracy, con-
textualization, and completeness. But when it comes to unpublishing, as opposed to transparent
revision, the difference between availability and ‘findability’ may be tough to nail down. A cred-
ible argument may be made that by today’s standards, a piece of information that’s absent from
Google results may as well, for practical purposes, be under lock and key: it will only be found by
those who know enough to search for it. It is hard to accept this kind of informational vanishing
act without coming close to a generalized social ‘duty to forget’.
With or without a RTBF, then, the idea of unpublishing remains no less complex than the
prepublication balancing acts often involved in responsible journalism itself. Journalists do not
merely collect and disseminate information: they are routinely expected to add edificatory and
evaluative content encompassing independent discovery, verification, and interpretation (Shapiro,
2014). It does not seem a stretch to argue that these responsibilities continue after publication
in correcting the record. News content is, by nature, temporally conditioned – a “partial, hasty,
incomplete, inevitably somewhat flawed and inaccurate rendering of some of the things we have
heard”, in the famous words of Washington journalist David Broder (Lewis and Crick, 2014:
59). Any work of journalism is thus subject to later addition or correction by other journalistic
330
Who owns the news?
work that may or may not be assigned within or beyond the original news organization, accord-
ing to highly subjective, even random, decisions on what constitutes ‘news’ on any given day
(Carey, 1987: 151). If weighing harm is already an accepted part of journalistic considerations
under some circumstances, then it would be difficult to exempt from further consideration the
somewhat arbitrary post-publication routine of deciding what is or is not subject to correction
or amplification.
The autonomy of news subjects: signs of a shift

Moving beyond the adjudication of unpublishing requests, growing awareness of the so-called
long tail of news is subtly altering discourse about the exercise of journalists’ ethical discretion
before publication. In the United States, a 2014 update of the Society of Professional Journalists’
code called for journalists to “consider the long-term implications of the extended reach and
permanence of publication” (Society of Professional Journalists, 2014). One investigative journal-
ist described to a journalism class taught by us how he had carefully explained to a woman
the possible consequences of broadcasting her on-camera confession of having used cocaine
in the presence of her child. In another class for the same group of graduate students, a differ-
ent reporter said he and a colleague had decided to include a video of a criminal act, showing
the face of the alleged perpetrator, but decided against including that person’s full name in the
written report. Their grounds for doing so: a face on video will not show up in name-based
search results.
It will probably remain unusual for such consideration to be shown to a criminal caught
in the act, and the danger of a slippery slope is worth marking. Journalism’s independence is
well-served by a culture of favoring audiences’ desire for information over potential benefit or
harm to people in the news. The innate tension between this culture and the exercise of caution
with respect to potential harm is, perhaps, an argument for shifting the adjudication of difficult
questions from the subjective realm of journalists’ impulses to the public realm of guidelines and
debate (Ward, 2014). Guidelines based on categories of situation types, as proposed in a more
general context by Hartzog and Selinger (2015), may help resolve most matters, while unavoid-
ably leaving others in grayer moral zones. Even within those areas of uncertainty, journalists
would probably prefer to see themselves as better placed to make tough judgments than the
alternative deciders – courts and governments. In Canada, for example, it would take but a small
stroke of Parliament’s pen to amend the Personal Information Protection and Electronic Docu-
ments Act (S.C. 2000, c.5) to remove Section 7(1)(c), which exempts journalistic research from
the prohibition on unauthorized collection of private information.
One way to proactively avoid tough RTBF requests would be more scrupulous attention to
follow up on crime stories, especially where an acquittal follows a charge. It will be increasingly
difficult for journalists to make a convincing case for maintaining the integrity of a historical
record when the historical record is blatantly and damagingly incomplete (Andrews, 2014; Brock,
2016: 88; Segal, 2011). However, it seems inevitable that increased consciousness of news cover-
age’s longevity will foster a willingness by journalists to consider granting to their subjects, as
Kennamer and Gillespie have suggested, a similar level of autonomy as is enjoyed by the subjects
of scientific research, under guidelines that guard against needless deception and recognize sub-
jects’ vulnerability to exploitation (Gillespie, 2009; Kennamer, 2005). The idea of informed con-
sent is not traditionally part of journalists’ parlance, but why would a morally aware journalist not
show consideration toward ordinary citizens who might not be “equipped with the proper level
of media literacy skills in order to manage the responsibility for their own privacy” (Nina and
Boers, 2013)? The authors of a Canadian ethics panel report failed to reach agreement on jour-
nalists’ obligations when it comes to informed consent but did suggest that news organizations
331
offer more guidance to journalists regarding consent protocols (Enkin, 2014; Levine et al., 2014).
As one of the authors put it:
journalists do engage in discussions about consequences with potential subjects and

sources all the time, but we tend to keep those conversations focused on the potentially
positive things [. . .]. Perhaps the time has come to include acknowledgment, in the
limited circumstances outlined above, of [. . .] some of the potential risks.
(Levine et al., 2014: 14)
Indeed, it might even be argued that journalists have an implicit contractual duty to use infor-
mation only in ways that are consistent with the intent of the source (Walker, 2012). In effect,
journalists make a business arrangement with their sources for the disclosure and transmission of
information, and any such arrangement is subject to the conditions to which both parties explic-
itly, or implicitly, agree. Where the conditions of that arrangement are ambiguous, one party is
likely to be unfairly disadvantaged. Journalists might find such ambiguity convenient, but it is
hard to imagine a moral principle on the basis of which their convenience should trump others’
rights to autonomy over their own future. In some cases, that includes “resetting their digital
life” (Brock, 2016: 86). In an age of infinite information longevity, it seems reasonable to accom-
modate moderate limits on public memory.
Further reading
Jeffrey Toobin’s “The Solace of Oblivion,” in the New Yorker (2014), provides an engaging look
into the consequences of Europe’s Right to be Forgotten ruling, a complete (and equally read-
able) analysis of which may be found in George Brock’s The Right to Be Forgotten: Privacy and the
Media in the Digital Age (2016). Turning from the law to moral choices, the Canadian Association
of Journalists’ ethics advisory committee provided relevant guidance and context in two 2014
panel reports: by Currie and Brethour on unpublishing, and by Levine, English, and Enkin on
informed consent.
References
Ambrose, M. L. (2014) “Speaking of forgetting: Analysis of possible non-EU responses to the right to be
forgotten and speech exception.” Telecommunications Policy, 38(8–9), 800–811.
Andrews, P. (2014, February 22) “Publishing the start and finish of crime stories.” Guelph Mercury (Guelph,
Ontario).
A.T. v. Globe24h.com [2017], FC 114 (CanLII), 2017. Retrieved from www.canlii.org/en/ca/fct/doc/
2017/2017fc114/2017fc114.html
Ausloos, J. (2012) “The ‘right to be forgotten’ – Worth remembering?” Computer Law & Security Report,
28(2), 143.
Baker, C. E. (2004) “Autonomy and informational privacy, or Gossip: The central meaning of the First
Amendment.” Social Philosophy and Policy, 21(2), 215–268.
Baume, M. D. L. (2012, November 30) “Mazarine Pingeot, Mitterrand daughter, looks back.” The New York
Times. Retrieved from www.nytimes.com/2012/12/01/world/europe/mazarine-pingeot-mitterrand-
daughter-looks-back.html
Bennett, S. C. (2012) “The ‘right to be forgotten’: Reconciling EU and US perspectives.” Berkeley Journal
of International Law, 30(1), 161–195.
Blanchette, J.-F. and Johnson, D. G. (2002) “Data retention and the panoptic society: The social benefits of
forgetfulness.” The Information Society, 18(1), 33–45.
Braganza, C .(2015) “The Toronto Star’s Gardasil controversy: A timeline.” J-Source. Retrieved from http://
j-source.ca/article/toronto-stars-gardasil-controversy-timeline
332
Who owns the news?
Brock, G. (2016) The Right to Be Forgotten: Privacy and the Media in the Digital Age. London: I.B. Tauris.
Campbell v. MGN Limited [2004 UKHL 22] (2004) Retrieved from www.publications.parliament.uk/pa/
ld200304/ldjudgmt/jd040506/campbe-1.htm
Carey, J. W. (1987) “Why and how: The dark continent of American journalism.” In R. K. Manoff and M.
Schudson (eds.), Reading the News. New York, NY: Pantheon Books (pp. 146–196).
Christopher, D. (2015) “True free expression requires privacy.” Canadian Journalists for Free Expression.
Retrieved from https://cjfe.org/resources/features/true-free-expression-requires-privacy
Conseil d’Etat. (2017) Right to be Delisted. Retrieved from http://english.conseil-etat.fr/Activities/
Press-releases/Right-to-be-delisted
Currie, T. and Brethour, P. (2014) “Case study and analysis on unpublishing. [A report of the Ethics Advi-
sory Committee of the Canadian Association of Journalists].” J-Source. Retrieved from www.j-source.
ca/article/caj-ethics-committee-case-study-when-unpublish
Dong, K. (2016) “News: Google takes a new approach to the ‘right to be forgotten’ in response to regulator
pressure.” Inforrm’s Blog. Retrieved from https://inforrm.wordpress.com/2016/03/10/news-google-takes-
a-new-approach-to-the-right-to-be-forgotten-in-response-to-regulator-pressure/
Edmonds, R. (2016) “Newspapers hit with a wave of requests to take down embarrassing archived
stories.” Poynter. Retrieved from www.poynter.org/2016/newspapers-hit-with-a-wave-of-requests-to-
take-down-embarrassing-archived-stories/432585/
Eltis, K. (2011) “Breaking through the ‘Tower of Babel’: A ‘right to be forgotten’ and how trans-systemic
thinking can help reconceptualize privacy harm in the age of analytics.” Fordham Intellectual Property,
Media & Entertainment Law Journal, 22(1), 69.
English, K. (2009) “The longtail of news: To unpublish or not to unpublish.” Online Journalism Credibility
Projects, Associated Press Managing Editors. Retrieved from www.apme.com/?Unpublishing
English, K., Currie, T. and Link, R. (2010) “The ethics of unpublishing. [A report of the Ethics Advisory
Committee of the Canadian Association of Journalists].” J-Source. Retrieved from http://j-source.ca/
article/ethics-unpublishing
Enkin, E. (2014) “On the record: Is it really consent without discussion of consequences?” J-Source. Retrieved
from www.j-source.ca/article/record-it-really-consent-without-talk-consequences
Enkin, E. (2016) “Public record vs. right to be forgotten.” CBC Ombudsman. Retrieved from www.ombuds
man.cbc.radio-canada.ca/en/complaint-reviews/2016/public-record-vs-right-to-be-forgotten/
European Commission – Justice. (2016) Protection of Personal Data. Retrieved from http://ec.europa.eu/
justice/data-protection/
European Parliament. (2014) Protection of Individuals With Regard to the Processing of Personal Data. Retrieved
from www.europarl.europa.eu/sides/getDoc.do?pubRef=-//EP//TEXT+TA+P7-TA-2014-0212+0+
DOC+XML+V0//EN
Factsheet on the ‘Right to be Forgotten’ Ruling (C131–12) (n.d.) European Commission. Retrieved from http://
ec.europa.eu/justice/data-protection/files/factsheets/factsheet_data_protection_en.pdf
Fazlioglu, M. (2013) “Forget me not: The clash of the right to be forgotten and freedom of expression on
the Internet.” International Data Privacy Law, 3(3), 149.
Francis, J. G. and Francis, L. P. (2014) “Privacy, confidentiality, and justice.” Journal of Social Philosophy, 45(3),
408–431.
Gillespie, B. (2009) Journalists and their Sources: A Critical Examination of Ethics in Conventional Journalism
Methodology. MA thesis, Integrated Studies, Athabasca University. Retrieved from http://dtpr.lib.
athabascau.ca/action/download.php?filename=mais/BruceGillespieProject.pdf
Gladney, G. A., Shapiro, I. and Ray, R. (2013) “Reasons for veiled sources spike after 2004 scandals.” News-
paper Research Journal, 34(2), 36–49.
Glasser, C. J. (2014, December) “Christmas arrives early for newsrooms: EU Privacy Working Group says
newspapers’ search engines are not subject to RTBF.” MLRC Bulletin.
Google Inc v Vidal-Hall & Ors [2015] EWCA Civ 311. (2015) Retrieved from www.bailii.org/ew/cases/
EWCA/Civ/2015/311.html
Google Transparency Report. (2017) European Privacy Reports for Search Removals. Retrieved from www.
google.com/transparencyreport/removals/europeprivacy/?hl=en
Hartzog, W. and Selinger, E. (2015) “Google’s action on revenge porn opens the door on right to be forgot-
ten in US.” The Guardian. Retrieved from www.theguardian.com/technology/2015/jun/25/googles-
revenge-porn-opens-right-forgotten-us
InfoCuria. (2014) Google Spain SL, Google Inc. v. Agencia Española de Protección de Datos (AEPD), Mario Costeja
González ECLI:EU:C:2014:317. Retrieved from http://curia.europa.eu/juris/document/document.jsf;
333
jsessionid=9ea7d2dc30dd22f58fa149c7444d8a018188ebb3b432.e34KaxiLc3qMb40Rch0SaxuPah10?t
ext=&docid=152065&pageIndex=0&doclang=en&mode=lst&dir=&occ=first&part=1&cid=210401
Jones, A., Leta, M., Friess, N. et al. (2012) Seeking Digital Redemption: The Future of Forgiveness in the Internet
Age. SSRN Scholarly Paper, Rochester, NY: Social Science Research Network. Retrieved from http://
papers.ssrn.com/abstract=2154365
Kauth, G. (2015, January 5) “Lingering media coverage a growing issue for defence lawyers.” Law Times.
Retrieved from http://lawtimesnews.com/201501054390/focus-on/focus-lingering-media-coverage-
a-growing-issue-for-defence-lawyers
Keller, D. (2015) “The final draft of Europe’s ‘Right to Be Forgotten’ law.” The Center for Internet and Society.
Retrieved from http://cyberlaw.stanford.edu/blog/2015/12/final-draft-europes-right-be-forgotten-law
Keller, D. (2016) “Global right to be forgotten: Delisting, why CNIL is wrong.” Inforrm’s Blog. Retrieved
from https://inforrm.wordpress.com/2016/11/22/global-right-to-be-forgotten-delisting-why-cnil-is-
wrong-daphne-keller/
Kennamer, D. (2005) “What journalists and researchers have in common about ethics.” Journal of Mass Media
Ethics, 20(1), 77–89.
LaRue, F. (2011) Report of the Special Rapporteur on the Promotion and Protection of the Right to Freedom of Opin-
ion and Expression. Human Rights Council, United Nations. Retrieved from www2.ohchr.org/english/
bodies/hrcouncil/docs/17session/A.HRC.17.27_en.pdf
Lee, D. (2015) “BBC forgotten list ‘sets precedent’.” BBC News. Retrieved from www.bbc.com/news/
technology-33287758
Lepore, J. (2015) “The cobweb: What the web said yesterday.” The New Yorker. Retrieved from www.
newyorker.com/magazine/2015/01/26/cobweb
Levine, M., English, K., Enkin, E. et al. (2014) “On the Record: Is it really consent without discussion of
consequences? [A report of the Ethics Advisory Committee of the Canadian Association of Journal-
ists].” J-Source. Retrieved from http://j-source.ca/article/informed-consent
Lewis, J. and Crick, P. (2014) Media Law and Ethics in the 21st Century: Protecting Free Expression and Curbing
Abuses. London: Palgrave Macmillan.
Lomas, N. (2014) “Google starts granting ‘right to be forgotten’ requests in Europe.” Techcrunch. Retrieved
from http://techcrunch.com/2014/06/26/right-to-be-forgotten-is-go/
Manna, L. (2014) “The right to oblivion, right of information and the internet: An overview of Italy’s current
trends.” Lexology. Retrieved from www.lexology.com/library/detail.aspx?g=ffc522d4-6d26-4b82-9563-
ca633c9e9442
Mantelero, A. (2013) “The EU proposal for a General Data Protection Regulation and the roots of the ‘right
to be forgotten’.” Computer Law and Security Review: The International Journal of Technology and Practice,
29(3), 229–235.
Matheson, D. (2013) “A duty of ignorance.” Episteme, 10(2), 193.
Matthews, A. (2016, September 20) “How Italian courts used the right to be forgotten to put an expiry
date on news.” The Guardian. Retrieved from www.theguardian.com/media/2016/sep/20/how-italian-
courts-used-the-right-to-be-forgotten-to-put-an-expiry-date-on-news
Mayer-Schönberger, V. (2009) Delete: The Virtue of Forgetting in the Digital Age. Princeton, NJ: Princeton
University Press.
McGoldrick, D. (2013) “Developments in the right to be forgotten.” Human Rights Law Review, 13(4), 761.
McNealy, J. (2012) “The Emerging conflict between newsworthiness and the right to be forgotten.” North-
ern Kentucky Law Review, 39(2), 119–135.
Moore, A. (2008) “Defining privacy.” Journal of Social Philosophy, 39(3), 411–428.
Nina, Ñ. and Boers, R. (2013) “Disliking the like.” New Library World, 114(7/8), 319–325.
Official Journal L 281, 23/11/1995 P. 0031–0050. (1995) Directive 95/46/EC of the European Parliament
and of the Council of 24 October 1995 on the Protection of Individuals With Regard to the Processing of Personal
Data and on the Free Movement of Such Data. Retrieved from http://eur-lex.europa.eu/LexUriServ/
LexUriServ.do?uri=CELEX:31995L0046:en:HTML
Ohm, P. (2015) “Sensitive information.” Southern California Law Review, 88(5), 1125.
Palm Beach Newspapers v. State of Florida, et al. (2016) Retrieved from http://cjrarchive.org/img/posts/
Post%20DCA%20opinion%20(1).pdf
Pantic, N. (2014) Gatekeeping and Unpublishing: How Editors Make Publishing and Unpublishing Decisions. MA
thesis, University of Missouri, Columbia. Retrieved from http://gradworks.umi.com/10/18/10181057.
html
PJS v News Group Newspapers Ltd [2016] UKSC 26. (2016) Retrieved from www.supremecourt.uk/cases/
uksc-2016-0080.html
334
Who owns the news?
Potter, D. (2009) Handbook of Independent Journalism. Orange Grove Books. Retrieved from www.free-
ebooks.net/ebook/Handbook-of-Independent-Journalism
Press Gazette. (2016) Pink News Republishes Stories Removed From Google Search Results Under ‘Right to
be Forgotten’. Retrieved from www.pressgazette.co.uk/pink-news-republishes-stories-removed-from-
google-search-results-under-right-to-be-forgotten/
Prosser, W. (1960) “Privacy.” California Law Review, 48(3), 383.
Reding,V. (2010) “Privacy matters:Why the EU needs new personal data protection rules.”European Commission
Press Release Database. Retrieved from http://europa.eu/rapid/press-release_SPEECH-10-700_en.htm
Renner, N. (2016) “A loophole in the right to be forgotten.” Columbia Journalism Review. Retrieved from
www.cjr.org/tow_center/loophole_google_right_forgotten.php
“Report of the Advisory Council to Google on the Right to be Forgotten” (2015) Retrieved from https://drive.
google.com/file/d/0B1UgZshetMd4cEI3SjlvV0hNbDA/view?pref=2&pli=1&usp=embed_facebook
Roberts, J. J. (2015) “Google defies France over ‘right to be forgotten’.” Fortune. Retrieved from http://
fortune.com/2015/07/30/google-france-right-to-be-forgotten/
Rosen, J. (2012) “The right to be forgotten.” Stanford Law Review Online, 64, 88–92.
Rosenzweig, P. (2012) “Whither privacy?” Surveillance & Society, 10(3/4), 344–347.
Salarelli, A. (2015) “The right to be forgotten: What changes after the European Court of Justice judgment
against Google.” Italian Journal of Library, Archives, and Information Science (JLIS.it) 6(1), 147–160.
Schudson, M. (2005) “The virtues of an unlovable press.” The Political Quarterly, 76, 23–32.
Schultz, I. (2007) “The journalistic gut feeling.” Journalism Practice, 1(2), 190–207.
Segal, C. (2011) “Is follow-up required on campus police blotter reports? Scholars speak out.” iMediaEth-
ics. Retrieved from www.imediaethics.org/is-follow-up-required-on-campus-police-blotter-reports-
scholars-speak-out/
Segal, D. (2013, November 10) “Mug-Shot websites, retreating or adapting.” New York Times.
Shafer, J. (2014) “The source may be anonymous,but the shame is all yours.”Reuters. Retrieved from http://blogs.
reuters.com/jackshafer/2014/06/16/the-source-may-be-anonymous-but-the-shame-is-all-yours/
Shapiro, I. (2014) “Why democracies need a functional definition of journalism now more than ever.”
Society of Professional Journalists. (2014) SPJ Code of Ethics. Available from: www.spj.org/ethicscode.asp
Tenore, M. J. (2010) “5 ways news organizations respond to ‘unpublishing’ requests.” Poynter. Retrieved
from www.poynter.org/2010/5-ways-news-organizations-respond-to-unpublishing-requests/104414/
Tomlinson, H. (2015) “Case Law, Spain: A and B v Ediciones El Pais, Newspaper archive to be hidden from
internet searches but no ‘re-writing of history’.” Inforrm’s Blog. Retrieved from https://inforrm.word
press.com/2015/11/19/case-law-spain-a-and-b-v-ediciones-el-pais-newspaper-archive-to-be-hidden-
from-internet-searches-but-no-re-writing-of-history-hugh-tomlinson-qc/
Tomlinson, H. (2016) “Case Law, Belgium: Olivier G v Le Soir. ‘Right to be forgotten’ requires anonymisation of
online newspaper archive.” Inforrm’s Blog. Available from: https://inforrm.wordpress.com/2016/07/19/
case-law-belgium-olivier-g-v-le-soir-right-to-be-forgotten-requires-anonymisation-of-online-news
paper-archive-hugh-tomlinson-qc/
Toobin, J. (2014) “The solace of oblivion.” The New Yorker, 90(29), 26–32.
van den Brink, J. (2017) “Media report: Missed opportunity – Dutch Supreme Court copy-pastes Google
Spain judgment.” Media Report: Kennedy Van der Laan Attorneys. Retrieved from www.mediareport.nl/en/
press-law/25042017/missed-opportunity-dutch-supreme-court-copy-pastes-google-spain-judgment-2/
Vultee, F. (2010) “Credibility as a strategic ritual: The Times, the interrogator, and the duty of naming.”
Journal of Mass Media Ethics, 25(1), 3–18.
Walker, R. K. (2012) “The right to be forgotten.” Hastings Law Journal, 64(1), 257.
Ward, S. J. A. (2014) “Why new journalism ethics have to be public, not personal.” MediaShift. Retrieved
from http://mediashift.org/2014/10/why-new-journalism-ethics-have-to-be-public-not-personal/
Watson, D. (2012) “Unpublishing requests are on the rise.” Columbia Journalism Review. Retrieved from
www.cjr.org/the_news_frontier/unpublishing_requests_are_on_t.php
Weber, R. H. (2011) “The right to be forgotten.” JIPITEC : Journal of Intellectual Property, 2, 120–130.
Xanthoulis, N. (2013) “The right to oblivion in the information age: A human-rights based approach.”
US-China Law Review, 10, 84.
335
25
DEFAMATION IN UNBOUNDED
SPACES
Journalism and social media
Normally a new hairstyle wouldn’t warrant a court summons, but Ziggy Mosslmani was adamant
that his hair had made him the target of defamation in the mainstream media. Mosslmani’s choice
of a distinctive ‘mullet’ haircut, which was shaved at the sides to showcase long black hair at the
back, came to public attention in 2015 when photographer Jeremy Nool posted a photo of
Mosslmani at a birthday party on his Facebook page. Over the next months, the post gained some
notoriety, with thousands of comments and memes made about Mosslmani’s hairstyle. The memes
and comments were shared widely via a number of social media platforms, including through the
creation of a new Facebook page called “The Holy Mullet”. While the memes about the hairstyle
were mostly the creation of external Facebook users, it was the mainstream media’s reportage on
the social media phenomenon and subsequent nationwide attention that drew Mosslmani’s ire.
He proceeded with legal action against the mainstream media organizations the Daily Mail, The
Daily Telegraph, and radio station KIIS FM, arguing that the attention of journalists, rather than the
memes themselves, had made him the subject of ridicule. Underlying Mosslmani’s lawsuit is a par-
ticular understanding of journalism’s legitimacy and authority in society. Mockery of Mosslmani’s
hairstyle in the mainstream media ‘of record’ is thus much more likely to injure his reputation than
a meme on social media – even if that meme is seen by many more people. While Mosslmani
could reasonably expect ridicule, mockery, and even contempt as part of the flows of social media
communication about his hairstyle, his lawsuit suggests an expectation that journalism’s social
authority will mean adherence to more stringent ethical and legal standards.
Mosslmani’s court case is a contemporary example of the new dynamics of journalism and
defamation in online and social media-enabled environments. While traditional journalism con-
tinues to be framed by particular norms of ethical conduct, the modes of communication that
have framed these norms have fundamentally changed. Indeed, the emergence of social media
communication environments has changed the traditionally ‘one-way’ communication relation-
ship between journalists and their audiences (Singer et al., 2011). In the context of journalism,
social media has assisted in creating seemingly ‘unbounded spaces’ for user participation where
audiences are able to critique, re-present, dismiss, or even completely ignore the journalist.
Focusing on defamation, this chapter uses examples from the Australian news media to illustrate
how traditional norms of public and participatory communication are contested and negotiated
by journalists and their audiences in online spaces. We frame this analysis with conceptual-
izations of public dialogue in online spaces. Indeed, the ideal of participatory communication
336
Defamation in unbounded spaces
(Jenkins, 2006) is the creation of online publics that allow for dynamic civic engagement in
social and cultural issues (Lévy, 1997). This ideal has been framed by the representation of uni-
fied groups working collaboratively within networks to debate and prompt change (Livingstone,
2005). However, the reality of this participatory online culture has been much more complex. In
particular, traditional norms of public dialogue – framed by professional ideologies and govern-
mental and legal norms – have become sites of contestation in online spaces, as audiences resist
or recreate the power structures that frame public dialogue, sometimes in deliberately defama-
tory ways, such as trolling or flaming. Journalists themselves are often engaging with or being
targeted by deliberatively defamatory commentary and discussion, which are seemingly part of
online news flows and often posted without repercussion. Journalists however, are still bound by
professional, ethical, and legal norms of practice in these spaces. Journalists have thus had to find
ways to negotiate new forms of online engagement to safely and productively participate within
online cultures prioritizing authenticity and, sometimes, to deal with deliberately disruptive and
defamatory content. This chapter begins by describing the traditional legal and social frame-
works for understanding defamation and its importance in journalism practice. We then explore
how these norms have been contested in social media spaces that prioritize particular commu-
nication cultures, including disruptions to public dialogue in the form of trolling and flaming.
Finally, using examples from the news media in Australia, this chapter illustrates how journalists
are using new online practices to negotiate civil dialogue in ‘unbounded’ online spaces.
Journalism and defamation

The law of defamation is a traditional protection against public attacks on the character and rep-
utation of a person. Though laws of defamation are particularized according to country, its origin
in the Western liberal tradition is based on the damage that public slander or defamatory publica-
tion could inflict on individuals, their personal reputations and their business dealings. Defama-
tion is in fact one of the oldest injuries recognized in law; the ancient Romans first utilized a
kind of actionable defamation law to prevent aggrieved parties using violence to uphold their
reputation (Troiano, 2006: 1451). The law was adopted more broadly in seventeenth-century
England after the political and judicial decline of the English monarchy and its subsequent failure
to protect the reputation of the royal court from the burgeoning and unregulated popularity of
the printing press (Veder, 1903: 547). The theory underpinning the law of defamation is that a
person’s reputation is invaluable. Indeed, in the violent and hierarchical era of Roman rule and
even until the English monarchy, dishonor was most often an element in a civil action – and
a good reputation its best defense (ibid: 548). While much has changed politically and socially
since then, reputation and good standing in the minds of others is still seen as an important mea-
sure of a person’s personal and professional character, and thus, the publication of offensive and
untrue statements by journalists and others is still an actionable offense.
Defamation was not actually a concern for journalists until fairly recently; journalists were
in fact initially the unregulated purveyors of social and political gossip and intrigue. The indus-
trialization of news in the early 1900s was marked by a period of ‘yellow’ journalism, where
newspaper editors competed for audience attention with tabloid-style scandals and opinionated
reportage (Schudson, 1978). However, the increasing professionalization of journalism beginning
in the 1920s resulted in an ideological preoccupation with norms of legal and ethical professional
behavior. New journalism professional associations in the US and Britain created codes of ethics
for practice, with the aim both to engender societal trust through responsible reportage and to
cement journalists’ social status as the only providers of reliable news. Traditional norms of ethi-
cal journalistic practice emphasize both truth and objectivity above all else. Through adherence
337
to accurate and unbiased reportage, journalists maintain a social contract to uphold and animate
democracy by providing information that enables public discussion and debate. Just as dishonor
was the motivating factor in the initial definition of defamation laws, defaming a person or
organization in reportage would also mean dishonoring the ideological framework of truth and
objectivity inherent to journalistic practice.
This framework of journalism practice assumes a particular kind of communicative relation-
ship with the public. Traditional journalistic practice is framed by understanding of the public
as a stable and unified mass audience, reliant on journalism for reliable information. Benedict
Anderson’s (1983) early conception of an “imagined community” understood the public as
socially constructed through the shared representation of a public, an affinity created through
media representation of a national, unified public to which it ‘speaks’. The conception of a
socially constructed and unified public has been an important aspect of the understanding of
journalism practice and the reifying of journalism’s importance to society. Within this concep-
tion, audiences behave as part of an ordered social structure that conforms to the ‘correct’ versions
of both media engagement and unified public dialogue that contributes to democracy. This, of
course, creates a power dynamic that is exploited by journalists and their organizational structures
to ensure their dominance in public and social life (Ang, 1991: 23). However, these previously
stable and industrialized conceptions of audience have been disrupted somewhat by the social
actions and interactions afforded by social media technologies.
In this chapter, social media are defined as:
web-based services that allow individuals to: 1) construct a public or semi-public pro-
file within a bounded system; 2) articulate a list of other users with whom they share a
connection; and 3) view and traverse their list of connections and those made by others
within the system.
(boyd and Ellison, 2008: 211)
Digital technologies and the creation of new online spaces such as social media platforms inevi-
tably shift the form of communication between journalists and their audiences from a one-to-
many approach to a many-to-many conversation (Rosen, 2006). Similarly, the increasing use
of online and social media has enabled audience participation in the different stages of news
production, distribution, and response (Deuze et al., 2007). The emphasis on participation and
sharing of personal expression and content, especially on social media, has meant user comments,
Facebook and Instagram ‘likes’, tweets, and other forms of informal discussion have become
increasingly important aspects of engagement with news. Furthermore, traditionally ‘silent’ news
audiences have increasingly directed the content of online news and social issues through these
forms of engagement (Lewis et al., 2014).
The social possibilities of online communication were forged around the de-territorialized,
nonhierarchical and expressive nature of the medium, where individuals would be able to forge
multiple relationships and groups according to interest (Vedel, 2003: 246–247). Many schol-
ars have been enthusiastic about the positive social and political potential of online and social
media-enabled communication. While some have suggested online potential to create a virtual
public sphere (Loader, 1997), others have claimed that the web represents an open and egalitarian
space for citizenship (Lévy, 2002) as well as a place where individuals can interact without inter-
ference from governmental and market power (Morris and Delafon, 2002). These online and
social media spaces could thus bring greater engagement and participation in democratic society
(Bimber, 1998; Rheingold, 1991). Research about social media has been similarly positive about
the potential to foster civic engagement and collective action because of its connectedness, speed
of communication, relative cost-effectiveness, and power to engage feedback loops (Obar et al.,
338
2012). Indeed, positive reciprocity and exchange for mutual benefit has been shown to increase
the possibilities for more active discourse and participation and strengthening community links
(Sankaranarayanan and Vassileva, 2009).
On the other hand, other research has criticized an overly positive approach to online and
social media communication, suggesting they are simply additional channels for political debate
(Margolis and Resnick, 2000) or a tool that only empowers those who can access technologies,
platforms, and languages (Fraser, 1990). Others (Koch, 2005: 160–161) have suggested that online
and social media communication has fallen short of the ideal of public communication because
of its distributed nature and the poor quality of online discussion. Some research (Thelwall
and Sud, 2011) has shown that users usually make singular contributions to debate, access only
opinions they share, and often disrespect other users’ opinions if they differ from their own. It is
therefore relevant to keep examining how traditional norms of public dialogue are negotiated in
online spaces, in particular in the context of journalism.
Online spaces and challenges to civic dialogue

The differing views about the effect of online and social media on news and public dialogue
illustrates the difficulty of defining norms of communication in online public spheres. The global
links and distributed nature of communication on social media platforms has meant traditional
‘rules of engagement’ between journalists and audiences have fundamentally changed. Social
media enables discussion between larger groups of disparate users based on shared interest in a
topic, event, or commercial activity and linked through sharing of content, hashtags, or ‘likes’.
Bruns and Burgess (2011) define these connections as “ad hoc publics”, where short-term dis-
cussion networks might appear around particular events, posts, or content and some users might
become influential through their engagement with these topics or events. For example, an ad
hoc public can form around a Twitter hashtag, creating short-term discussion or influential social
commentary. Ad hoc publics are thus not static and may change or become part of established
online communities over time.
As different social media platforms curate and mediate content more actively, streams of online
engagement mimic news flows. Bruns and Burgess (2011) suggest, using Twitter as an example:
these days, a curated version of the hashtag stream may even be broadcast alongside
television news coverage or displayed on a public screen. Furthermore, the speed with
which individual users can initiate social media discussion about an event, which can
then create an ad hoc public, also suggests the new dominance of the audience over
professional journalists in the news cycle.
Within these social media spaces, news becomes a distributed conversation where a number of
individuals are responding simultaneously to events as a kind of “ambient journalism” (Hermida,
2011). Social media interactions can thus be seen as an aggregated form of journalistic content,
experienced as a constant ambient presence in social media spaces but becoming more explicit
when major news breaks. Shaw et al. (2013: 23) suggest this results in a “public and collec-
tive” expression of public affect about particular news events, with or without connection with
traditional news organizations and journalists. Social media platforms encourage interpersonal
connection through networks of followers; similar to Anderson’s idea of ‘imagined communities’,
most users imagine this community of followers when posting information about themselves or
events – and furthermore, that those followers are just like them (Bruns, 2011).
Thus, what has changed in this new constitution of public dialogue and civic engagement is
the roles of participants and the rules of engagement in participatory forms of communication.
339
Crawford (2009: 528) suggests online spaces create interconnection through “access to the details
of someone’s everyday life, as prosaic as they often are, which contributes to the sense of ‘ambient
intimacy’ in social media”. This constant access to ambient intimacy, as well as the dominance
of affective content based on personal opinion and experience and, finally, the centrality of the
user in the construction of ad hoc publics engaged with news, creates public spaces with very
different cultures, politics, and rules of civil engagement. Indeed, participants in such complex
and distributed online publics collectively negotiate the roles – and rules – of social relations
in these spaces, with these negotiations unfolding through both positive and negative feedback
(Shaw et al., 2013: 26). While the interpersonal connections forged in online spaces through
sharing personal emotion and experience on social media empowers audiences and creates com-
munities, they can also create negative interactions and experiences. As Crawford (2009: 528)
suggests, this online intimacy can “generate discomfort, confusion and claustrophobia, amongst a
range of negative effects. People often express anger, sadness, fear, and resentment. They may also
misrepresent themselves and lie”. Participation in the intimate relations of online communication
do not always adhere to the traditional rules of civic engagement and can thus result in aberrant
forms of communication.
In this chapter, we situate trolling and flaming as aberrant forms of online communication,
defined as deliberately defamatory acts aimed at disrupting the norms of public dialogue online.
Trolling can be defined as the act of posting inflammatory or off-topic material online to cre-
ate discord in online discussion. Trolling is somewhat complex in intent, though most online
communities understand trolling as disruption to online discussion. The most common type of
trolling is a form of dark humor, where trolls deceive users into an angry or emotional response
by convincing users that they believe their offensive statements. Other types of trolling include
attempting to thwart discussion of political or social issues as a form of activism or silencing par-
ticular online commentators. Flaming differs from trolling because it is a less complex act; flam-
ing is simply the act of posting offensive messages sent to a particular person through discussion
forums, social media, or messaging services. A group of users may direct ‘flames’ to one person,
or two or more users can essentially trade insults over time in a ‘flame war’ – the intent simply to
victimize other users. Both trolling and flaming are intentionally offensive and defamatory and
have the intent to disrupt online public dialogue.
Describing participation in online communities as a kind of “digital citizenship”, McCosker
(2014) suggests these kinds of aberrant behaviors are often seen as problematic because they do
not adhere to the ethical and legal boundaries of traditional public communication. Aberrant
online behaviors like trolling and flaming disrupt the democratic and participatory potential so
often celebrated in online spaces; however, McCosker (2014) argues
a truly pluralistic participatory experience includes not just being affected by new
forms and flows of networked media content and communication, but also the power
to affect with new forms of reciprocal capacity to act out and even ‘act up’.
Thus, while idealized or unitary modes of participatory online engagement represent this dis-
ruption as defamatory and undemocratic, these aberrant behaviors can actually serve to better
animate public engagement by allowing contested interactions.
While trolling and flaming are often meant to offend, some of these aberrant online behaviors
are also meant to express disagreement, opposing opinion, or even satire. For example, Bergstrom
(2011) suggests trolling is often intended to make another user the victim, but often as the result
of a joke. While this would often be considered defamation in traditional public spheres (and is
often considered defamation by the victim), trolling and flaming appear to be part of the com-
munication flows of online communities. However, trolls or flamers sometimes also intend to
340
deceive or manipulate online discussion for a particular political aim (Osnos, 2016). Similarly,
there have been examples of on- and offline stalking, particularly of females, on social media
emerging from misogynistic and homophobic attitudes (Sierra, 2014). Thus, many of these dis-
cursive forms of aberrance (quite apart from physically dangerous behaviors) are the affective
forms of participation enabled by the intimate and distributed character of social media com-
munication and might intensify or fall away depending on engagement and interest (McCo-
sker, 2014). For many users, these forms of communication are simply part of broader digital
citizenship, the participation and contribution to social and cultural life through the particular
affordances of social media. While these combative forms of behavior can be part of the flows
of information on social media, they differ exponentially from the tightly governed, structured
forms of communication created by traditional norms of professional journalism.
Journalistic participation in social media spaces

Journalists have used new forms of online and social media communication by engaging with
audiences online, using online content in reportage, and collaborating with audiences in the
investigation and distribution of news. This audience engagement is represented by participatory
or reciprocal forms of journalism, defined as the process by which journalists and their audi-
ences are involved in the news production cycle and community-building activities that gravitate
around news (Singer et al., 2011). However, online interactions between journalists and audiences
have not always fulfilled the characteristics of participation due to ongoing tension about just
how open professional journalism practice should be to audience interaction, especially poten-
tially defamatory commentary. While these can sometimes be productive challenges serving to
illustrate where innovation or reconstitution of journalism practice can improve public dialogue,
there have nonetheless been issues in the structure and governance of this public dialogue, par-
ticularly when users engage in aberrant online behavior.
While the ambient flows of social media discussion create cultures of ‘authentic’ communica-
tion that can be both civil and aberrant and that affect users without repercussion, professional
journalism is framed by its ethical and legal commitment to civil public dialogue. Therefore, even
in these unbounded spaces, journalistic reportage is still regarded as a product that adheres to pro-
fessional practices and represents ethical social dialogue. For example, the former Australian Trea-
surer Joe Hockey successfully sued one of Australia’s largest media organizations, Fairfax Media,
for defamation following the publication of two tweets. The treasurer was awarded $80,000 for
two tweets reading: “Treasurer Hockey for sale”, which were distributed on Twitter via Fairfax’s
Victorian newspaper The Age (Joe Hockey, 2015). In finding for Hockey, Justice Richard White
acknowledged the need to promote news stories; however, he said the “140 character limit on
tweets would still have permitted alternative forms of eye-catching promotion of the articles”
(ibid., 2015). Justice White’s assessment suggests journalists must continue to abide by traditional
modes of verification, accuracy, and objectivity, despite changes in communication technologies
and their associated dialogue cultures. Thus, it seems journalists and news organizations are still
held to professional, ethical, and legal boundaries even in seemingly unbounded online and social
media spaces.
As journalists have transitioned into online practice, they have increasingly had to delve into
the “contested publics” that form around particular news content (McCosker, 2014). As well as
discussion and engagement with news events, journalists are also exposed to and interact with
comments on news stories, which can often be defamatory, obscene, racist, and insulting to both
the journalist and subject of news (Noci et al., 2012: 52). Within social media and news website
comment sections especially, there is extensive scope for aberrant participation. Recent stud-
ies have shown news events, especially those mentioning politics and religion, are most likely
341
to garner confrontational and negative commentary by online users (Thelwall and Sud, 2011).
Compared to other forms of online engagement, comments on news events can be continually
defamatory because they are difficult to delete and, if shared, reach a large number of people
(Carter, 2015). Journalists have suggested the most difficult aspect of engaging with online audi-
ences is the negativity and abuse that anonymous comments can foster. Hiding their identity
empowers online users to voice the strongest opinions and theories, often exceeding the norms
of civil public dialogue (Santana, 2014). Moreover, research has shown comments by anonymous
authors have been divisive, rather than unifying online communities and discussion (Hlavach
and Freivogel, 2011: 31). Finally, it was shown that in online communities where people could
not participate anonymously, participants were more inclined to adhere to community rules
(Blanchard et al., 2011). Despite the aims of participatory and open communication in online
spaces, journalists and news organizations have found it impossible not to structure, govern, and
moderate public dialogue about news in some way.
In this context, many media organizations have implemented guidelines about how journalists
can and should participate in online discussion. The two most common strategies to moderate
online news comments are pre- and post-moderation. Post-moderation is preferred by orga-
nizations because it is less labor-intensive and costly than pre-moderation (Reich, 2011). Post-
moderation does not review comments prior to publication; thus, media organizations consider
themselves “not legally responsible for the content of those contributions the moment they
appear” (Singer et al., 2011: 124). However, there is a legal responsibility to respond in post-
publication moderation, and if organizations fail to do so, especially in the case of abusive com-
ments, they could be legally prosecuted (Canter, 2013). Thus, the repercussions of unbounded
online spaces do apply to professional journalists and their news organizations and thus, several
traditional practices are often implemented to ensure costly business or reputational damage.
When post-moderation is effectively applied, it has been shown to decrease abusive comments.
However, it has also led to fewer comments and less engagement with online news (Hille and
Bakker, 2014). For example, some media organizations have effectively ‘outsourced’ comment
moderation to social media platforms, rather than enable comments on their own websites. Sev-
eral media organizations and publications, including USA Today, Popular Science, and Reuters, have
ceased reader comments on their websites to encourage discussion on social media platforms (Ellis,
2015; see also Ksiazek and Springer, this volume, Chapter 36). This has worked to decrease the
number of anonymous, abusive comments, but there has also been a subsequent reduction in the
number of posted comments overall (Hille and Bakker, 2014). Furthermore, interaction has been
more prominent between different users rather than with journalists (ibid, 2014). This is probably
related both to journalists’ lack of time and fear of losing their institutional authority in online
discussion spaces. Journalists thus seem to replicate a system of one-way communication, failing
to utilize engagement capabilities on these platforms due to the legal and ethical ramifications of
‘uncivil’ dialogue (Domingo, 2008). While this reversion to traditional communication formats
has stifled some abusive commentary, other journalists have actively negotiated their own and
their audiences’ participation in online spaces to productively respond to defamation and abuse.
Journalistic negotiation of social media spaces

Journalists have varied many of their traditional practices to participate in online communica-
tion in an ethical and professional way. But can journalists truly participate in ‘unbounded spaces’
when this labor can result in potentially dangerous forms of incivility? While much research has
focused on why aberrant behaviors occur and how they impact on users, there has been little
focus on the way trolling, flaming, and other potentially defamatory commentary forms part
of online labor, particularly for journalists and other media content producers. Furthermore,
342
there is little information about the kinds of engagement professional communicators believe are
adequate to work within these complex and disruptive online spaces. The labor of online jour-
nalists and commentators, particularly female, LGBTI, and ethnically diverse journalists, has been
impacted by defamatory trolling and flaming directed at them in online spaces. According to
recent analysis by the Guardian of more than 70 million comments left on its site since 2006, they
found the highest level of trolling was addressed to women, ethnic and religious minorities, and
LGBTI people, regardless of the topic discussed and despite the majority of their opinion writers
being white men (Gardiner et al., 2016). Similarly, an Australian survey developed by national
steering committee Women in Media supported by the Media, Entertainment & Arts Alliance
(Australian Media, 2016) found female journalists were most likely to experience trolling, while
freelancers are most likely to be cyber-stalked (see also Nettleton, this volume, Chapter 32).
The Australian columnist Van Badham has experienced first-hand the defamatory trolling
that comes with online journalism. Badham writes mostly about feminist issues, which she
suggests makes her a particular target for defamatory comments and trolling. She has also been
subjected to offline stalking, threatening materials sent in the mail, and one online menace even
impersonated her partner and her friends (O’Brien, 2016). Badham (2014) suggests female jour-
nalists and columnists are particular targets for flaming and trolling, as this reflects the broader
misogyny that attempts to silence women within society: “The experience of being cyberstalked,
poison-penned, monitored, physically followed, threatened, vexatiously litigated and even subject
to bomb threats is a matter of record amongst [. . .] women”. These trolling behaviors are per-
petrated mainly by anonymous sexist and misogynist ‘keyboard warriors’, enabled by the culture
of online and social media platforms (O’Brien, 2016).
Female columnists have negotiated their own ‘rules of engagement’ for dealing with trolls
and defamatory comments in online spaces. Badham for example, receives up to 100 defamatory
and abusive comments per day, and while she does employ tactics like blocking and ignoring
commenters, occasionally she will engage with abusive or trolling commenters. She will employ
tactics like ‘kittening’ – replying with images of kittens to repetitious trolling – or using mockery
to “make an example” of some trolls (Case, 2014). This simulates the affective-expressive intent
of the original trolling; these engagements are not meant to be dialogical but rather express
the often-confrontational nature of discourse in online and social media spaces. Furthermore,
‘counter-trolling’ by using witty remarks or other playfully disruptive modes of response clev-
erly reflects the expressive and inflammatory intent of trolling without engaging in potentially
defamatory acts.
Other female columnists like feminist writer Clementine Ford have employed much more
vigorous governance of participation in ‘her’ online spaces. Ford uses her Facebook page to dis-
cuss her columns and feminist issues. It is also the site of the most virulent abusive commentary,
and thus Ford places limits on the kinds of discussion that take place on her page. For example,
the first post pinned on her Facebook page states:
This is a feminist page. It is my page. I post about things that I’m passionate about or
that interest me, or issues in society that I want to see change [. . .]. You may not share
all of my opinions, and that’s okay. But this is my page, and I will keep posting my
opinions. Not yours and not anybody else’s. If this doesn’t suit you, perhaps consider
whether or not this is the right place for you.
In this way Ford asserts her ‘ownership’ and control of both her Facebook page and the discourse
that appears on it – this is somewhat unusual for typically open and participatory online spaces.
Some online researchers (Crawford, 2009: 529) have suggested users often lose control over com-
munication in ambient social media spaces due to the ‘always on’ nature of online technologies.
343
However, in the negotiation of the roles and rules of online spaces, Ford has merged tradi-
tional broadcast controls over spaces attributed to her with other participatory forms of online
engagement.
Ford also actively responds to trolling or flaming she receives, using screenshots of abuse
with identifying information so perpetrators are publically shamed by her large number of
followers. In one such incident, Ford shared a screenshot of a sexually explicit comment she
received from a man whose Facebook profile listed his employer. Ford posted an image of the
comment, along with other posts from the man’s Facebook page. She then emailed the images
to his employer, asking if they were aware of his behavior. Some days later Ford shared a mes-
sage sent by the employer announcing the man’s employment was terminated after an investiga-
tion into his online behavior. Ford received a deluge of abuse accusing her of causing the man
to lose his job, but Ford was unrepentant in her use of a fairly traditional form of response to
defamatory content. She suggested in another post: “He is responsible for the things he writes
and the attitudes he holds”, and these attitudes would not be acceptable in professional practice
(Clementine Ford describes, 2015). In her online “Daily Life” column, published by Fairfax,
Ford said her justification for these responses was based on her own negotiation of the roles and
rules of online engagement: “There are basically no consequences for men who behave like this,
so we have to start making consequences for them” (ibid, 2015). Ford’s response suggests online
spaces do not have the ethical or legal boundaries around what constitutes civil public engage-
ment and, thus, no consequences for the personal and professional injury aberrant discursive
behavior might cause.
There is no doubt trolling and flaming does have material consequences, especially for female
journalists and media producers. Female journalists have reported ongoing abusive commentary
and stalking behaviors have resulted in lost professional labor, wages, and personal time due to
consistent logging and responding to trolls and flamers (Friedersdorf, 2014). Some have suggested
the impact of dealing with this behavior has a chilling effect on female journalists, evidenced by
the dominance of men in online spaces (Dumas, 2016). While there are methods for ignoring,
blocking, and reporting abusive online commentary, it is often very difficult for female journalists
to engage in online public dialogue without also needing to negotiate aberrant behaviors that
come with it. Journalists like Ford suggests the real solution is for
society to collectively challenge the kind of behavior that thinks it’s ok for people to
either publicly or privately message you things about violence that they want to do to
your body, violence that should be done to your family, all with the intent of scaring
you into silence.
(Dumas, 2016)
For Ford, this has meant engaging her large online following to reflect the same confrontational
discourse back to trolls and flamers. For example, a man sent a message to Ford’s Facebook page
threatening to rape and bash her. Ford then shared the message publicly as a screenshot on her
Facebook page, which has about 45,000 likes (Caggiano, 2015). The screenshot garnered hun-
dreds of comments from other users on the social media platforms of the identified troll. The
man who sent the abuse subsequently apologized, ironically after receiving a number of com-
ments on his social media. In publically apologizing, he said, “everybody should think before
they post trolling comments online, as they may receive a bigger response than they intended”
(ibid, 2015). While the lack of acknowledgment of the misogyny inherent in his apology shows
that Ford’s technique does not necessarily create social change, it does effectively reflect onto
the perpetrator the personal and professional injury trolling and flaming can inflict. Rather than
going through the impossibly costly and laborious process of suing every troll or flamer for
344
defamation, Badham and Ford have shown reflecting the characteristics of aberrant behavior back
onto the perpetrator still allows an authentic engagement with social media cultures of public
dialogue but is also professionally safe and legitimate.
Conclusion
In this chapter we have focused on forms of defamation to understand how social media plat-
forms simultaneously empower and limit the spaces for dialogue between journalists and news
audiences. While aberrant online behavior such as trolling and flaming is intentionally defama-
tory, it can be seen as part of the ambient news flows of communication on social media, where
ad hoc publics form around shared interests or issues but also attract divisive and confrontational
forms of dialogue. While the ideal of online communication is to be participatory and unified
around shared civic goals, sites of contestation in public spaces, even those that are intention-
ally uncivil, can also act to animate online discussion and action. This aberrant behavior has
been described as an unacknowledged aspect of the culture of social media interactions, much
maligned and yet often occurring without repercussion. Nonetheless, for professional journalists
engaging in this space, not only are there still repercussions for their own unprofessional behav-
ior in social media spaces, but they are also often the target of abusive dialogue. As defamation
suits against journalists and news organizations have shown, change of publication platforms and
their associated cultures has not exempted journalists from the very ethical and legal norms that
they strove to commit to in order to engender public trust and their own authority in the public
sphere. While that social authority has become much more diffuse, journalists’ product – the
news – is still framed by the norms of journalism practice that uphold civic engagement with
audiences. Therefore, this chapter has shown journalists, especially those who risk being victim-
ized and silenced by targeted abuse, have had to negotiate their own forms of social media ‘rules
of engagement’ in ‘unbounded’ online spaces. Without professional, ethical or legal norms to
fall back on, journalists are instead negotiating new forms of online engagement that prioritize
online cultures of collaboration and authenticity but also structure some boundaries around
deliberately defamatory or abusive content.
Further reading
This chapter has benefitted from prior research published by McCosker (2014) on trolling, as
well as Crawford’s (2009) earlier work on listening on social media. Understanding of journalists’
negotiation of online spaces was informed by Hermida’s (2011) “Fluid Spaces, Fluid Journalism:
Lessons in Participatory Journalism” and Hille and Bakker’s (2014) “Engaging the Social News
User: Comments on News Sites and Facebook”. An excellent contextual resource for this chap-
ter was the edited collection Participatory Journalism: Guarding Open Gates at Online Newspapers
(2011) produced by Singer et al. The incredibly misogynistic and damaging impacts of trolling
of female journalists continues to be underrepresented by critical work in academia but is bravely
reported by some of the female journalists and news organizations cited in this chapter.
References
Anderson, B. (1983) Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso.
Ang, I. (1991) Desperately Seeking the Audience. London: Routledge.
Australian media still a ‘Blokesworld’ in 2016. (2016, June 3) Women in Media for Media, Entertainment and
Arts Alliance. Retrieved from www.meaa.org/news/australian-media-still-a-blokesworld-in-2016/
Badham,V. (2014, July 2) “When they call us frightbats they make us fair game.”The Guardian. Retrieved from www.
theguardian.com/commentisfree/2014/jul/02/when-they-call-us-frightbats-they-make-us-fair-game
345
Bergstrom, K. (2011) “‘Don’t feed the troll’: Shutting down the debate about community expectations on
Reddit.com.” First Monday, 16(8). Retrieved from www.uic.edu/htbin/cgiwrap/bin/ojs/index.php/fm/
article/view/3498/3029
Bimber, B. (1998) “The internet and political transformation: Populism, community, and accelerated plural-
ism.” Polity, 31(1), 133–146.
Blanchard, A. L., Welbourne, J. L. and Boughton, M. D. (2011) “A model of online trust: The mediating
role of norms and sense of virtual community.” Information, Communication & Society, 14(1), 76–106.
boyd, D. M. and Ellison, N. B. (2008) “Social network sites: Definition, history and scholarship.” Journal of
Computer-Mediated Communication, 13(1), 210–230.
Bruns, A. (2011) “A first map of Australia.” Mapping Online Publics. Retrieved from www.mappingonline
publics.net/2011/08/04/a-first-map-of-australia/
Bruns, A. and Burgess, J. (2011) “The use of Twitter hashtags in the formation of ad hoc publics.” Proceedings
of the 6th European Consortium for Political Research (ECPR) General Conference 2011, University of Iceland,
Reykjavik. Retrieved from https://eprints.qut.edu.au/46515/1/The_Use_of_Twitter_Hashtags_in_
the_Formation_of_Ad_Hoc_Publics_%28final%29.pdf
Caggiano, A. (2015, June 26) “Man shamed for trolling Clementine Ford apologises for online attack.”
The Sydney Morning Herald. Retrieved from www.smh.com.au/national/man-shamed-for-trolling-
clementine-ford-apologises-for-online-attack-20150625-ghxz24.html
Canter, L. (2013) “The misconception of online comment threads: Content and control on local newspaper
websites.” Journalism Practice, 7(5), 604–619.
Carter, L. (2015, August 25) “Social media defamation cases on the rise, as lawyers tell people to think
before they tweet.” ABC News. Retrieved from www.abc.net.au/news/2015-08-25/social-media-
defamation-cases-on-the-rise/6723328
Case, J. (2014, January 28) “‘A Misery of Trolls’: Van Badham’s tips on dealing with internet nasties.” The
Wheeler Centre Notes. Retrieved from www.wheelercentre.com/notes/8b3cbb2926fa
Clementine Ford describes deluge of abuse over Facebook post which saw hotel worker Michael Nolan
sacked. (2015, December 2) ABC. Retrieved from www.abc.net.au/news/2015-12-01/ford-responds-to-
sacking-after-offensive-facebook-message/6992192
Crawford, K. (2009) “Following you: Disciplines of listening in social media.” Continuum, 23(4), 525–535.
Deuze, M., Bruns, A. and Neuberger, C. (2007) “Preparing for an age of participatory news.” Journalism
Practice, 1(3), 322–338.
Domingo, D. (2008) “Interactivity in the daily routines of online newsrooms: Dealing with an uncomfort-
able myth.” Journal of Computer-Mediated Communication, 13(3), 680–704.
Dumas, D. (2016, December 23) “Clementine Ford and Mariam Veiszadeh on being trolled – and how hate
only makes them fight back harder.” The Sydney Morning Herald. Retrieved from www.smh.com.au/
lifestyle/news-and-views/news-features/clementine-ford-and-mariam-veiszadeh-on-being-trolled –
and-how-hate-only-makes-them-fight-back-harder-20161223-gthcmo.html
Ellis, J. (2015,September 16) “What happened after 7 news sites got rid of reader comments.”NiemanLab. Retrieved
from www.niemanlab.org/2015/09/what-happened-after-7-news-sites-got-rid-of-reader-comments/
Fraser, N. (1990) “Rethinking the public sphere: A contribution to the critique of actually existing democ-
racy.” Social Text, 25/26, 56–80.
Friedersdorf, C. (2014, January 7) “When misogynist trolls make journalism miserable for women.” The
Atlantic. Retrieved from www.theatlantic.com/politics/archive/2014/01/when-misogynist-trolls-make-
journalism-miserable-for-women/282862/
Gardiner, B., Mansfield, M., Anderson, I., Holder, J., Louter, D. and Ulmanu, M. (2016, April 12) “The dark
side of Guardian comments.” The Guardian. Retrieved from www.theguardian.com/technology/2016/
apr/12/the-dark-side-of-guardian-comments
Hermida, A. (2011) “Fluid spaces, fluid journalism: Lessons in participatory journalism.” In J. Singer, D.
Domingo, A. Heinonen, A. Hermida, S. Paulussen, T. Quandt, Z. Reich and M. Vujnovic (eds.), Participa-
tory Journalism: Guarding Open Gates at Online Newspapers. Malden, MA: Wiley-Blackwell.
Hille, S. and Bakker, P. (2014) “Engaging the social news user: Comments on news sites and Facebook.”
Hlavach, L. and Freivogel, W. H. (2011) “Ethical implications of anonymous comments posted to online
news stories.” Journal of Mass Media Ethics, 26(1), 21–37.
Jenkins, H. (2006) Convergence Culture: Where Old and New Media Collide. New York, NY: New York Uni-
versity Press.
Joe Hockey awarded $200,000 over defamatory ‘Treasurer for sale’ headline, tweets by Fairfax Media. (2015,
July 22) ABC News. Retrieved from www.abc.net.au/news/2015-06-30/treasurer-joe-hockey-wins-
defamation-case-against-fairfax/6582900
346
Koch, A. (2005) “Cyber citizen or cyborg citizen: Baudrillard, political agency, and the commons in virtual
politics.” Journal of Mass Media Ethics, 20(2/3), 159–175.
Lévy, P. (1997) Collective Intelligence. Cambridge, MA: Perseus Books.
Lévy, P. (2002) Cyberdémocratie: essai de philosophie politique. Paris: Odile Jacob.
Lewis, S. C., Holton, A. E. and Coddington, M. (2014) “Reciprocal journalism: A concept of mutual
exchange between journalists and audiences.” Journalism Practice, 8(2), 229–241.
Livingstone, S. (2005) Audiences and Publics: When Cultural Engagement Matters for the Public Sphere. Bristol:
Intellect.
Loader, B. (1997) The Governance of Cyberspace: Politics, Technology and Global Restructuring. London: Routledge.
Margolis, M. and Resnick, D. (2000) Politics as Usual: The Cyberspace “Revolution”. London: Sage Publications.
McCosker, A. (2014) “Trolling as provocation: YouTube’s agonistic publics.” Convergence. Retrieved from
http://journals.sagepub.com/doi/abs/10.1177/1354856513501413
Morris, D. and Delafon, G. (2002) Vote.com ou comment internet va révolutionner la politique. Mesnil-sur-l’Estrée:
Ed. Plon.
Noci, J. D., Domingo, D., Masip, P., Micó, J. L. and Ruiz, C. (2012) “Comments in news, democracy
booster or journalistic nightmare: Assessing the quality and dynamics of citizen debates in Catalan online
newspapers.” International Symposium on Online Journalism, 2(1), 46–64.
Obar, J. A., Zube, P. and Lampe, C. (2012) “Advocacy 2.0: An analysis of how advocacy groups in the
United States perceive and use social media as tools for facilitating civic engagement and collective
action.” Journal of Information Policy, 2, 1–25.
O’Brien, N. (2016, March 6) “Social media trolling of female journalists is insidious, report shows.” Sydney
Morning Herald. Retrieved from www.smh.com.au/nsw/social-media-trolling-of-female-journalists-is-
insidious-report-shows-20160305-gnba81.html
Osnos, E. (2016, May 19) “The year of the political troll.” The New Yorker. Retrieved from www.newyorker.
com/news/daily-comment/the-year-of-the-political-troll
Reich, Z. (2011) “User comments. The transformation of participatory space.” In J. Singer, D. Domingo, A.
Heinonen, A. Hermida, S. Paulussen, T. Quandt, Z. Reich and M. Vujnovic (eds.), Participatory Journalism:
Guarding Open Gates at Online Newspapers. Chichester: Wiley Blackwell.
Rheingold, H. (1991) “The great equalizer.” Whole Earth Review, 71(6), 95–104.
Rosen, J. (2006, June 27) “The people formerly known as the audience.” PressThink. Retrieved from http://
journalism.nyu.edu/pubzone/weblogs/pressthink/2006/06/27/ppl_frmr.html
Sankaranarayanan, K. and Vassileva, J. (2009) “Reciprocal and non-reciprocal relationships in an online
community.” Workshop on Adaptation and Personalization for Web 2.0. Retrieved from http://ceur-ws.org/
Vol-485/paper11-F.pdf
Santana, A. D. (2014) “Virtuous or vitriolic. The effect of anonymity on civility in online newspaper reader
comment boards.” Journalism Practice, 8(1), 18–33.
Schudson, M. (1978) Discovering the News: A Social History of American Newspapers. New York, NY: Basic
Books.
Shaw, F., Burgess, J., Crawford, K. and Bruns, A. (2013) “Sharing news, making sense, saying thanks.” Aus-
tralian Journal of Communication, 40(1), 23–39.
Sierra, K. (2014, October 8) “Why the trolls will always win.” Wired. Retrieved from www.wired.
com/2014/10/trolls-will-always-win/
Singer, J. B., Hermida, A., Domingo, D., Heinonen, A., Paulussen, S., Quandt, T., Reich, Z. and Vujnovic, M.
(2011) Participatory Journalism: Guarding Open Gates at Online Newspapers. Malden, MA: Wiley-Blackwell.
Thelwall, M. and Sud, P. (2011) “Commenting on YouTube videos: From Guatemalan rock to el big bang.”
Journal of the American Society for Information Science and Technology. Retrieved from http://onlinelibrary.
wiley.com/doi/10.1002/asi.21679/pdf
Troiano, M. (2006) “The new journalism? Why traditional defamation laws should apply to internet blogs.”
American University Law Review, 55(5), 1448–1471.
Veder, V. V. (1903) “History and theory of the law of defamation.” Columbia Law Review, 3(8), 546–573.
Vedel, T. (2003) “L’idée de démocratie électronique: origines, visions, questions.” In P. Perrineau (ed.), Le
désenchantement démocratique. La Tour d’Aigues: Editions de l’Aube.
347
26
HACKS, HACKERS, AND THE
EXPANSIVE BOUNDARIES OF
JOURNALISM
Nikki Usher
The boundaries of journalism are increasingly blurred as new entrants from people to new forms
of news expand the field. Of increasing importance to the future of journalism is the intersection
between programming and journalism. Code is the backbone of every news website, but having
staff who can program as part of editorial efforts has not only become more commonplace, but
it is also absolutely critical for news organizations looking to do more sophisticated digital story-
telling. But code is just one part of the larger story about this fusion between hacks (journalists)
and hackers (coders). As journalism becomes increasingly quantified, thanks to the rise of data
and tools that make it easier to understand this data, the importance of those who can computa-
tionally analyze these new sources of news have also become highly visible in many prominent
newsrooms across the world.
The chapter begins with an examination of some of the complexities and opportunities
afforded by the rise of these computationally minded journalists, collectively discussed here as
‘interactive journalists’ – an umbrella term that delineates this fusion of programming, data, and
journalism from other points in journalism’s long entanglement with computation (Usher, 2016).
However, given the 2016 Brexit and Trump votes, it’s worth probing whether the “master nar-
rative” of interactive journalism may well be dead. If quantification and visualization were sup-
posed to make information more visible and understandable, how did so many people – readers,
journalists, politicians – make faulty predictions about electoral futures? Did the reification of the
fusion of journalism and coding contribute to possible misinformation?
As interactive journalism came into the popular vernacular between 2009 and the present,
interactive journalists took on an aura of superiority (or mastery) – their special knowledge of
programming or data analysis or both made them relevant and important to the news indus-
try. A New York Magazine article from 2009 on the New York Times interactive staff reflects the
early exuberance surrounding interactive journalists with the subhead, “What are these renegade
cybergeeks doing at The New York Times? Maybe saving it” (Nussbaum, 2009). Key figures in
the movement were able to state the case for their superiority within the journalistic profession.
Nate Silver, the FiveThirtyEight election-predicting wunderkind, began a site separate from The
New York Times. In his manifesto, Silver (2014) argued that “one of the potential advantages of
data journalism is that it generalizes better than traditional approaches, particularly as datasets
increase in scale to become larger and more complex”. Moreover, in an era of big data, one
of journalism’s preeminent challenges will be how to harness its potential for adding to public
348
Hacks, hackers and journalism’s boundaries
knowledge. As issues from automation to algorithmic transparency become more prominent

(Carlson, 2015; Dörr, 2016), journalism requiring computation and data analysis will take on
additional relevance.
Given this potential, it’s easy to see how interactive journalism might fall victim to overly
high expectations and an undefined mission. Here, existing literature will be referenced to help
provide a cohesive sense of what we know so far. However, scholarship is contested as far as even
a basic historical trajectory of the field and simple definitional claims, making it hard to evaluate
what has been accomplished by a heterogeneous subfield of journalists. Moreover, much of the
literature takes the rise of programming in journalism as a normative good whereby quantifica-
tion and information visualization leads in sum to better journalism and therefore better public
knowledge.
CONTESTED HISTORY
In order to understand fully where interactive journalism fits into the profession as a whole,
it’s helpful to have some sense of its origins. This historical perspective provides insight into
the subfield’s lack of coherence. Some scholars and academics trace the origins of interactive
journalism to some of the first news infographics. Others argue that interactive journalism has
its antecedents in the move to make journalism more like social science, an approach amplified
by computer-assisted reporters. However, it makes more sense to bridge these two origin stories
with the argument that specific computation requirements were required in order to distinguish
interactive journalism as a wholly new field (Usher, 2016).
Some scholars chart the beginning of infographics themselves as part of a need to better com-
municate scientific discovery and world events. Monmonier (1989), one of the most extensive
chroniclers of mapping in journalism, argues that infographic-like images in the news might
be traced to sixth-century block printing. Friendly1 traces infographic history from the pre-
seventeenth-century visualizations of planetary movements. Some Guardian journalists argue that
they were doing data journalism from the very first day of publication in 1821, with a then-
revolutionary tabular presentation of education statistics (Datablog, 2011).
Throughout the development of newspaper printing, advances in information visualization
have largely been the result of war or major international events. The Titanic provided significant
incentive for journalists across the world to visually depict the tale of the sinking ship for read-
ers. World War II made Time and Fortune leaders in information visualization and encouraged
experimentation elsewhere. Today, war, disaster, and politics continue to provide fertile ground
for interactive development, with more data ever than before available for journalists to splice,
dice, present, and even share with their audiences.
Anderson (2017) ties together the maturation of data analysis with the rise of data-visualization,
arguing that journalism began to incorporate social science methods into journalism. The incor-
poration of interactives into newsrooms has not been a forward march of progress, with this
“technologically specific work” inspiring discourses of contestation, aspirations, and normaliza-
tion in the newsroom (Powers, 2012: 25).
But the history of infographics leaves out the computational aspect of interactive journalism.
Cox (2000) argues that the 1952 UNIVAC computer’s prediction of Eisenhower’s success for
CBS was a landmark in using data to help make journalism more precise. And it was that term,
‘precision’, that would inspire Detroit journalist Philip Meyer’s approach to what would go on to
be called computer-assisted reporting in its earliest iteration. Meyer (2002) advocated for bring-
ing social science methods, including hypothesis testing, into journalism and using this approach
349
Nikki Usher
to make sense of data that was either collected from government agencies or complied by news
organizations. His early work led to significant recognition in the field.
However, specifically thinking about interactives as simply the most technically sophisticated
version of computer-assisted reporting is incomplete. Significantly, web and mobile-enabled
interactive visualizations can include myriad forms of storytelling that move beyond just maps,
charts, graphs, and other displays of quantitative information. The birth of this new subfield
really brings with it the ability to use programming in order to enhance not only the reporting
experience but the user experience of engagement with web-enabled storytelling. The early ’80s
brought the arrival of tools like Microsoft Chart and hardware like the Macintosh; USA Today
shook up journalism with its page design focused around colorful weather maps and personal
finance graphics.
Toward the end of the 1980s, as computers became a more regular part of journalistic work
processes, the potential for new forms of journalist work became clear to some early visionaries.
Monmonier in 1989 offered a key prediction that proved to be true:
Electronic display makes possible map symbols that blink and move. . . . Video dis-
play technology and modern telecommunications offer the potential for linking highly
dynamic cartographic display software with large geographic databases . . . should elec-
tronic news databases ever emerge as a mass-communications medium, dynamic news
maps would become commonplace.
(Monmonier, 1989: 15–16)
At the time, studies about the rise of visual journalists in newsrooms suggested the inability
of other professional journalists to see their value as journalists first rather than artists (Lowrey,
2002), revealing tensions about the incorporation of innovation that foreshadowed some of the
issues seen in the integration of online journalism into traditional news but provides a point of
contrast for the “master narrative” of interactive journalism.
But computers needed to get faster and the web had to become more technically sophisticated
in order to enable the type of interactivity we take for granted today. Three factors – the rise of
cloud computing, the rise of ‘social’ open-source, and programming advances such as AJAX and
Flash set in motion new forms of storytelling as well as the need for new kinds of journalists.
Cloud computing allowed newsrooms to take advantage of cheap, distributed servers so as to
avoid having interactive projects clog up the newsroom’s mainframe. ‘Social’ open-source enabled
programmers to better develop community ties in order to share and build upon each other’s
code. Flash was easy enough for more seasoned infographic creators to move into the interactive
realm and to begin thinking about the potential of this form of storytelling, while AJAX enabled
basic functions like scrolling within a Google map or toggling over content without leaving the
page.
By 1995, the web was at least fast enough that most home computer users could support
some form of early multimedia. It’s at this point that we actually start to see the development
of interactives that took graphics beyond still representations and made them into something
tactile – something a user could actually engage with – a call-and response experience (Stromer-
Galley, 2004). Notably, significant innovation in U.S. newsrooms came from the fringes – small
newspapers rather than large ones. Young people who were excited by programming were get-
ting their first jobs at small newspapers across the country – in coastal Florida, in Lawrence, Kan-
sas, or in big-city metros like the Atlanta Journal-Constitution or The Chicago Tribune. Elsewhere, in
Europe, large papers often led key advances; El Pais in Spain, the Times of London, and the Daily
Telegraph (UK) were early leaders.
350
THE BIRTH OF THE ‘HACKER JOURNALIST’
A few key constitutive moments help mark the later acceleration in the 2000s and 2010s that
we have seen in the field and the celebration of the journo-techie as a ‘master’ of the newsroom:
the mythology of Adrian Holovaty, the institutionalization of interactive journalism within the
industry, and the bridge-building between journalists and programmers exemplified by Hacks/
Hackers and other extra-industry-spanning collaborations.
In the early 2000s, Holovaty, a University of Missouri-trained programmer, led a team of
young journalists who could program at the Lawrence Journal-World. The work in Lawrence
encouraged an industry-wide obsession with finding journalists who had a ‘start-up’ mentality,
programmer ‘cool’ as well as chops, and editorial knowhow. Holovaty, now no longer involved
in journalism, was one of the first to combine data with interactive visualization: a crime map
using Chicago crime stats and the ‘new’ Google Maps open API (at the time, this was called a
‘mash up’). His blog was also a starting point for the flourishing community – he posted job ads
and outlined a variety of manifestos; with others he also guest wrote entries and added extensive
commentary about their experimentation in the burgeoning subfield. Holovaty’s presence on the
journalism industry conference circuit was unavoidable for almost 10 years; many articles written
about him also reflect the industry fetishizing of his skills, with news executives expressing both
awe and the need for these journalists.
Institutionalization quickly followed. One of the first signs that the news industry at large
was seeing the form’s potential was the 2008 introduction of Knight-backed scholarships at
Northwestern University’s Medill School for Journalism for programmers to learn about jour-
nalism. Other programs fusing computer science and journalism or substantial opportunities for
credit emerging at leading journalism schools such as Columbia, Stanford, and the University
of Missouri – opportunities that are now increasingly standard at universities across the world.
Major newspapers like the New York Times and the Wall Street Journal, watching a Washington Post
newsroom effort led by Holovaty, scrambled to institutionalize teams with dedicated staff. In 2010,
international attention to the WikiLeaks cables and the MP scandals in the UK underscored the
significance of interactive journalism in storytelling (Howard, 2014). At this point, since 2010,
other newsrooms across the globe have developed either specific teams or emphasized bringing
on journalists with new expertise in programming into existing teams, though large newsrooms
have the ability to pay for and to attract top talent that smaller newsrooms often struggle to retain.
Further signs of institutionalization can be reflected in organizational growth and institutional
recognition. The National Institute for Computer-Assisted Reporting, or NICAR, began add-
ing more programming-specific bootcamps that delved beyond Excel manipulation to include
data scraping as well as the creation of both front-end and back-end development. NICAR is an
international conference with strong U.S. ties, but regional European-based conferences have also
emerged, such as NODA, the Nordic data journalism conference, begun in 2013, and DataHa-
rvest: The European Investigative Journalism Conference, begun in 2012. The New York Times’
2012 interactive “Snow Fall: The Avalanche at Tunnel Creek” helped both solidify and popular-
ize the potential of interactive journalism. The project won a Pulitzer and garnered record traffic
to the site from new visitors to the New York Times; in following years, interactives were consis-
tently among the newspaper’s top traffic-getters, setting a precedent for further investment in this
form of journalism and its practitioners not just at the Times but within the industry.
However, the development that really captured the attention of many hoping that program-
ming might be a way to help save journalism was the creation of Hacks/Hackers, a grassroots
effort that attempts to bridge gaps between the journalism community and the programming
community. Primary among these efforts was the journalist-created Hacks/Hackers organization
351
Nikki Usher
in 2009 – a loosely coordinated group that aims to bring together “hacks” – the journalists, and
“hackers” – the programmers – in order to begin a fusion between skill sets and cultures (Lewis
and Usher, 2014). At its height in 2014, the grassroots movement had about 75 international
chapters on four continents and an estimated membership of 23,000; today it boasts just over
50 chapters on three continents (Lewis and Usher, 2014). However, this group has had mixed
success in encouraging the fusion of techies and journalists. In some cases, the groups were
unable to attract any technology experts other than those already working in journalism. In other
cases, the groups diverged from the original journalistic purpose and instead served as a con-
vening point for a variety of information and knowledge-management fields such as civic tech
and citizen science. Newsrooms have had internal and externally focused hackathons to drive
innovation, taking cues from the Hacks/Hackers group’s role as a conveyor and culture-bridger
for technologists and journalists.
Major tech companies are invested in developing journalism’s technical chops. Knight and
Mozilla began a major international partnership in 2011 that featured a worldwide contest for
programming aficionados – the five winners were awarded with a one-year fellowship in one of
five international news organizations, no previous journalism experience necessary. This partner-
ship led to Open News and its online community, Source, which aims at connecting “a network
of developers, designers, journalists, and editors to collaborate on open technologies and pro-
cesses within journalism” and now has its own conference, SRCCON: the Open News Source
conference.
In 2015, Google launched “Google News Lab”, which provides support for fellowships for
programmer journalists, organizes a contest for interactive journalism, offers training material
(particularly around Google trends), and supports various news innovation projects, such as First
Draft, the citizen journalism verification effort. As part of pressure from European newspaper
publishers, Google has also provided about $164 million to help simulate news innovation from
start-ups to business models to interactive journalism. Facebook has recently launched additional
partnerships with journalists, though not around programming. However, questions remain
about the incentives of large technology companies to provide support for journalism when
their business models are a huge factor in the economic disruption of journalism. Nonetheless,
the maturation of the fusion of programming, data analysis, digital storytelling, and visualization
in newsrooms can be told through a variety of historical perspectives, but much needed is a focus
on the development and integration of new programming technology in the service of editorial
content.
WHO ARE THESE PEOPLE?
As the divergent potential for tracing interactive journalism’s history might indicate, defining
the field with some coherence has been difficult – with some scholars wondering whether it
even makes sense to have a clear definition. Are these practitioners ‘computer-assisted report-
ers’, ‘hacker journalists’, ‘data journalists’, ‘programmer journalists’, ‘developers’, ‘designers’, ‘news
nerds’, ‘computational journalists’, or something else (like ‘interactive journalists’)? Are they
doing ‘hacker journalism’, ‘computational journalism’, ‘data journalism’, ‘multimedia storytell-
ing’, or ‘interactive storytelling’? (see also Coddington, this volume, Chapter 17). Definitions
matter, though of course they are situated within the cultural context of the journalism being
conducted and need to reflect how the journalists themselves articulate their backgrounds and
self-definition. Furthermore, definitions also have normative dimensions and, in turn, reflect how
these journalists bring to bear a particular cognitive framework for understanding and creating
journalism. And finally, being able to define and delineate makes critique and feedback easier.
352
Powers (2012) argues that a multiplicity of discourses emerges; as journalists technologi-

cally innovate, they define their work in ways that designate both novel technologies and align-
ment with journalism. However, this is uneven, as this can be framed as disruptive, continuous,
or opportunistic. Within the broad umbrella of the subfield, computational journalism and
computer-assisted reporting have fairly established definitions. However, the terms ‘data journal-
ist’, ‘programmer journalist’, and ‘hacker journalist’ are much blurrier. Computational journal-
ism has its origins in 2006: Georgia Tech computer scientist Ifran Essa and his then-doctoral
student, Nicholas Diakopolous, coined the term as a way to think about how algorithms might
solve journalistic problems. The term is now often used to discuss how computers can be used
to aid journalism in the creation of tools that might help data analysis and assist journalists in
investigative work via algorithms, data, and social science methods (Flew et al., 2012; Hamilton
and Turner, 2009). Drawing from engineering and computer science, computational journalism
is increasingly associated with a move toward “algorithmic accountability”, whereby journalists
are encouraged to try to reverse-engineer algorithms, either computationally or through more
systematic, albeit anecdotal reporting efforts (such as comparing prices online by ZIP codes)
(Diakopolous, 2014).
Computer-assisted reporting is also connected to computing power to help to do investiga-
tive work, particularly through the application of social science methods. Computer-assisted
reporters may build databases, but generally, they are not building stand-alone tools that might
scale and aid the rest of the industry, nor are they working on sophisticated algorithmic pro-
gramming in the service of news. Rather, CAR reporters might be thought of as experts in
data-gathering (particularly via freedom of information act requests) and statistics (Coddington,
2015). Parasie and Dagiral (2013) found that CAR reporters were focused on statistics and anec-
dotes to be used for reporting. CAR reflects a temporal orientation prior to today’s ‘big data’
world and interactive web technology and recalls data exploration that was focused more on
internal newsroom use and includes only minimal emphasis on the presentation and exploration
of that data for users. Bounegru (2012) argues that CAR is primarily focused on investigative
reporting, while data journalism is not only more integrated into new forms of storytelling but
also reflects an approach to journalism that prioritizes openness, community-building, shared
resources, and learning. Overall, all journalists are now ‘computer-assisted reporters’ because they
all use computers.
‘Data journalism’, though, is a particularly messy term. Academics who have studied data
journalists across Norway, Sweden, Belgium, Germany, the US, and the UK find some coherence
when thinking about skills, practices, and newsroom integration. Particularly within Europe, data
journalism appears to be the predominant professional self-description. Holovaty is credited with
coining the term by many journalists, and a number of ‘data journalism’ handbooks (Bradshaw,
2013) and conferences use the moniker.
But the definitions break down, in part because the field incorporates a subgroup of journal-
ists who can have varying backgrounds (such as in technical or non-journalist fields) and varying
skill sets (some might be ace programmers, while others might only focus on data acquisition),
an unpredictable pathway to the profession. Royal (2012) found that Times interactive journalists
come from a variety of journalism and non-journalism backgrounds, described themselves by a
variety of titles, and had a variety of expertise and duties within the newsroom. Fink and Ander-
son (2015) also find a variety of skills within data journalists, who have different backgrounds,
expertise, and goals. Weber and Rall (2013) argue that interactive journalists are comprised of
programmers, journalists, and statisticians. Some consider what mastery of skills is required in
defining the field (Karlsen and Stavelin, 2014; Parasie and Dagiral, 2013). Smit et al. (2014) focus
on journalists whose work consists of “information visualization”, a term they argue refers to
“all mental models of data” (p. 345), ultimately concluding that there is no perfect convergence
353
Nikki Usher
of backgrounds working in this field. Howard (2014), drawing from interviews with dozens of
journalists working at the intersection of data and programming, suggests that data journalism
can be thought of as “gathering, cleaning, organizing, analyzing, visualizing, and publishing data
to support the creation of acts of journalism” (p. 4). Appelgren and Nygren (2014) define data
journalism as a combination of data analysis, programming, and visualization.
However, others resist the demarcation. De Maeyer et al. (2015) argue for a discursive view
of the term given the variety of identities and backgrounds data journalists in French-speaking
Belgium possess. The scholars conclude that lack of coherence makes it difficult for these jour-
nalists to have a shared interpretive community. They find there is a divergence in attitudes
toward data – as to whether the data is “ordinary” or doable by one person and within the
normal temporal patterns of newswork, or whether the data is “thorough” or requiring the “col-
lective mobilization of a range of skills ( journalism, computer science, statistics, graphic design)”
(p. 441). Trédan (2014) argues that there is no perfect way nor any actual definition that articu-
lates a coherent set of practical mastery. Along this line of thinking, Fink and Anderson (2015)
have proposed thinking about data journalism as an “assemblage” – or suggest that it is defined
by the fluidity of material, cultural, and practice-based understandings.
One strategy has been to look for epistemological differences as ways to delineate the group.
Coddington (2015) maps these divisions around openness, epistemology, vision of public, and the
professional orientation of journalists (for a helpful break down see p. 337). Parasie (2014) sug-
gests that programming journalism offers an epistemological challenge to traditional journalism
because it pushes journalistic work toward hypothesis and data-driven research rather than news
that relies on normative assumptions. In Germany (Weinacht and Spiller, 2014), note that while
data journalism is difficult to define, the scholars find both practical and epistemological differ-
ences that differentiate these journalists: different skill sets and a tendency to view the audience as
able to assist in developing new trends and ideas. However, scholars in Norway share this finding;
Karlsen and Stavelin (2014) found that despite different skill sets, there were many epistemologi-
cal similarities between the professional norms of data journalists and traditional journalists.
But ‘data journalism’ is fundamentally unsatisfying even across a multi-level analysis of indi-
viduals, organizations, and institutions. Data journalism restricts the definition of interactive
web-storytelling to a narrow focus on numeracy – but the story of interactive journalism is more
than just journalists dealing with ‘big data’. Parasie and Darigal (2013) and Karlsen and Stavelin
(2014) are among the few to use the term “programmer journalists” to define their research
participants, but their programmers focus on making data transparent for users – thus data jour-
nalism is at the core of their practices. Even teaching data journalism is difficult because of the
lack of a coherent definition – promoting educators to ask how legal training for data acquisition,
ethical considerations, or experience working with data via computational tools can fit together
(Hewett, 2013, 2016).
Many journalists working in this space do not work with data at all – and in fact, may be cod-
ing in the service of an editorial project that has little to do with the visualization of quantitative
information. Many of the technologists new to newsrooms are not data analysts, nor do they
wish to be; similarly, many of the journalists working in this area are more interested in using
programming to building novel solutions for storytelling that move far beyond structured data.
Interactive quizzes and interactive storytelling that provides an immersive journalism experience
replete with multimedia elements, newsgames, and beyond reflect a diversity in subject matter
and in content that is not represented by ‘data journalism’ and, moreover, fails to be reflected in
scholarship that tends to focus on more serious forms of quantitative storytelling, particularly
investigative journalism as conducted through data analysis. As a result, what is really different
about this subfield is the ability to create news applications that layer interactive features around
data and storytelling in ways that prompt engagement from the user. Thus, in my comprehensive
354
study of the field, found in Interactive Journalism: Hackers, Data, and Code, I think more broadly
about the convergence of hacks and hackers through what they create: “a visual presentation of
storytelling through code for multilayers, tactile user control for the purpose of news and story-
telling” (Usher, 2016: 20). If the field can only be categorized discursively, while scholarly inquiry
is well-suited to explore the ‘assemblage’, if there is no center and no common set of practices,
critique is more difficult – and critique is warranted at this point.
A reassessment
However, a reassessment is necessary surrounding the excitement and interest in interactive jour-
nalism from the field and from scholarship. As noted, the field embraces interactive journalism
as a savior – a ‘master narrative’ that includes two forms of mastery: masters over big data com-
prehension and the certainty of knowledge provided via quantification and as masters within the
newsroom: individuals who have skills others do not. Much of the scholarship has indeed focused
on what tech culture can bring to journalism as an epistemological offering: from transparency to
a reliance on open-source data to a fundamental openness that engages in user-directed inquiry
to narrative (Lewis and Usher, 2013, 2014, 2016; Usher, 2016).
But it’s time to think about the problems that data journalism/interactive journalism presents
to knowledge acquisition. That’s not to say that scholars haven’t pushed back entirely on interac-
tive journalism: they have studied convergence difficulties within the newsroom, for example.
But there’s a strong normative element to this work: that interactive journalists should be incor-
porated into the newsroom, and moreover, journalism should be more open, more data-driven,
and more engaged with the user. Perhaps those assumptions are ones worth putting on hold.
The first key point is that interactive journalists are, for the most part, not statisticians. They
know a lot, they’re self-trained, but in only the rarest instances are these journalists formally
trained in regression, logarithmic, and multivariate analysis, and beyond. To understand patterns
in data, all the machine learning and visualization in the world can’t help if journalists lack funda-
mental data comprehension skills. Ample evidence exists from the NICAR email list of journal-
ists struggling with basic statistical questions; if categorizing variables is difficult, moving beyond
descriptive statistics to understand the modeling of a dataset is simply not possible. When these
problems with quantification are then visualized, real issues emerge, because the visualizations
themselves are incomplete stories that belie the complexities of the numbers they represent. To
some degree, this has always been a problem with visualizations, but in this case, newsrooms are
often asking journalists to work with data and present it to audiences when journalists lack the
ability to “investigate” or “interview the data” (Willis, 2011).
These forms of storytelling matter to the public’s understanding of data: Silver’s NowCast and
the New York Times’ Upshot interactives provided predictive data, neatly visualized, which did not
reflect issues with uncertainty in polling and failed to communicate to users quantitative, data-
based reasons why their candidate might fluctuate each day in her chance of winning. We know
now that presenting people with all the data can be harmful – consider the effect of WikiLeaks’
data dump of documents on right-wing conspiracy nuts. The documents enabled speculation
about a child pornography ring with connections to Hillary Clinton being run out of a pizza
restaurant’s basement in Washington, DC. ‘Pizzagate’ caused a mentally imbalanced man to walk
into the restaurant and start shooting with an assault rifle. Other data interactives can prompt less
serious mistakes in judgment: ProPublica’s Dollars for Docs invites users to see if their doctor
accepts money from drug companies. This may prompt people to make judgments about doctors
being in the pockets of drug companies and choose to disregard medical advice or simply not
to go to a ‘corrupt’ doctor – my doctor and even my father appear on this database, but neither’s
acceptance of funds reflects anything dubious or even slightly questionable (an honorarium for
355
Nikki Usher
an invited lecture and the acceptance of a nominal amount then spent on treating staff to lunch).
If data is not contextualized – whether it is quantitative like Dollars for Docs or qualitative like
WikiLeaks and Pizzagate – audiences can’t make good judgments. The problem is even worse
when the information is rendered in visually pleasing but reductionist visualizations.
Another issue that is emerging is whether journalists are creating interactives that have the
intention of being storytelling vehicles or clickbait, or somewhere in the middle. The economic
incentives to create content that promotes extensive engagement on a site are real. Some of the
journalism that tells interactive stories don’t have advertising, though, either because the UX is
intended to be so immersive so that it is not interrupted or because advertising is hard to integrate
within the interactive experience. Other interactives are simply clickbait – albeit well-produced
clickbait. These are interactives that do not have significant public civic information but are
intended for entertainment: an interactive on the Westminster Dog Show, an interactive Oscar
Ballot, listicles of tourist destinations, and beyond. The worst offenders are interactive quizzes –
and publications from the New York Times to the BBC to Slate make use of them. The New York
Times dialect quiz was a top traffic getter, while Slate’s news quiz of the week has consistently
been a traffic driver for years. Newsgames have for the most part been civic in orientation,
but at what point will these games end up as “snack food” content? There is no problem with
entertainment journalism; however, the incentive to create more clickbait interactives is present,
many are clickbait, and they do return traffic. Perhaps the emphasis on interactives and their
overproduction may lead to more presentation and quantification errors (FiveThirtyEight is a
leading example in this case).
The ‘master narrative’ of interactive journalism has another problem – one that should have
been taken into account by the scholarship thus far but has not. The coinage of ‘the master narra-
tive of interactive journalism’ also brings to bear a development in journalism that is profoundly
gendered – masters connotes male mastery. All of the studies that attempt to classify the emerging
subfield by skills, geographical orientation, backgrounds, epistemologies, and beyond fail to actu-
ally address what might be the elephant in the room, so to speak: distinct imbalances in gender
and other forms of diversity. NICAR, an association that prides itself on inclusivity, had a speaker
list that was two-thirds male. The structural disadvantages that women face in being successful
in STEM education and later in STEM fields are now also a barrier that will affect the future
of the news industry. If interactive journalism can still save the industry, or be one critical force,
will those doing this work be mostly male? This is a conversation that needs to happen among
scholars and within the industry but has seldom been a focal point.
Finally, the ‘master narrative’ is also fundamentally problematic because of questions endemic
to cosmopolitanism, globalization, and partisanship. Though there are some indications of inter-
active journalism happening in hyperlocal newsrooms (Radcliffe, 2013), Fink and Anderson
(2015) rightfully point out that small newsrooms do not have the ability to foster cultures of
data journalism. Keeping a skilled interactive journalist is hard for a world-class newsroom that
can pay top dollar and near impossible for a small one that can’t do so and lacks the branding to
motivate an interactive journalist to take a pay cut. Though some tools are democratizing inter-
active journalism, such as basic chart builders and Google fusion software, interactive storytelling
stands to be a more robust path for reader engagement at news organizations that already have
better prospects in the digital ecosystem.
Moreover, the rise of partisan media also threatens the future of interactive journalism’s reach.
Partisan news organizations on both sides are manipulating data to enhance political claims;
poor data visualization from the wrong organizations can lead to serious implications (consider
Trump’s retweet of a neo-Nazi’s rendering of false crime statistics) (Hogan, 2015). In other
instances, the move away from “facts” to news organizations that choose “ideology” over “objec-
tivity” (Hemmer, 2016) suggests that the people who may benefit the most from interactive
356
journalism may not ever see it: those who most need rigorous interactive storytelling to bring
them to places unlike where they live and those who could benefit from data visualization about
hot-button issues may never see any of it. Interactive journalism holds promise, but it requires
significant critique moving forward.
Further reading
Scholarship on this subject continues to emerge. For a more comprehensive account of the rise
of interactive journalism, its incorporation into contemporary newsrooms, and the normative
paradigm it advances, consider Usher (2016)’s book on the subject, Interactive Journalism: Hack-
ers, Data, and Code, so far the only book-length treatment of the subject. Other related work on
the ideology of interactive journalism includes Lewis and Usher (2013) on open-source values
in journalism, Anderson’s (2013) effort to create a sociology of computational journalism, and
Parasie’s (2014) discussions of the epistemological tensions of interactive journalism. For work
outside the newsroom, consider Lewis and Usher’s (2014, 2016) look at the global network of
Hacks/Hackers and the formation of the Knight Mozilla News Technology Partnership, now
called Open News. Howard’s (2014) work provides a cataloging of interviews from dozens of the
US and UK’s top interactive journalists. Additional work by Broussard (2016) considers the take-
up of big data practices as code sharing and applied research methods. But before starting along
any line of research involving interactive journalism, Meyer’s (2002) Precision Journalism should
be examined as a classic that sets into motion the long line of scholarship and the incorporation
of interactives into practice.
Note
1 www.datavis.ca/milestones/
References
Society, 15(7), 1005–1021.
Anderson, C. W. (2017) “Social survey reportage: Context, narrative, and information visualization in early
20th century American journalism.” Journalism, 1, 81–100.
Appelgren, E. and Nygren, G. (2014) “Data journalism in Sweden: Introducing new methods and genres of
journalism into ‘old’ organizations.” Digital Journalism, 2, 394–405.
Bounegru, L. (2012) The Data Journalism Handbook: How Journalists Can Use Data to Improve the News.
Sebastopol, CA: O’Reilly Media.
Bradshaw, P. (2013) The Data Journalism Heist. Medium: Leanpub.
Broussard, M. (2016) “Big data in practice: Enabling computational journalism through code-sharing and
reproducible research methods.” Digital Journalism, 4(2), 266–279.
Cox, M. (2000) “The development of computer-assisted reporting.” Paper presented at Association of Educa-
tion in Mass Communication and Journalism, Chapel Hill, NC.
Datablog. (2011) “The first Guardian data journalism: May 5, 1821.” The Guardian: Datablog. Retrieved from
www.theguardian.com/news/datablog/2011/sep/26/data-journalism-guardian
Diakopolous, N. (2014) Algorithmic Accountability: Journalistic Investigations of Computational Power Structures.
New York, NY: Columbia University School of Journalism: Tow Center for Digital Journalism.
357
Nikki Usher
Flew, T., Daniel, A. and Spurgeon, C. (2012) “The promise of computational journalism.” Journalism Practice,
6(4), 157–171.
Hamilton, J. T. and Turner, F. (2009) Accountability Through Algorithm: Developing the Field of Computational
Journalism. Stanford, CA: Center for Advanced Studies in the Behavioral Sciences Summer Workshop.
Duke University in Association with Stanford University.
Hemmer, N. (2016) Messengers of the Right: Conservative Media and the Transformation of American Politics.
Philadelphia, PA: University of Pennsylvania Press.
Hewett, J. (2016) “Learning to teach data journalism: Innovation, influence and constraints.” Journalism,
17(1), 119–137.
Hogan, P. (2015) “Donald Trump retweeted a racist graphic with a fake stats from a Neo-Nazi Twitter account.”
Fusion. Retrieved from http://fusion.net/story/237019/donald-trump-retweets-racist-graphic/
Howard, A. (2014) “The art and science of data-driven journalism.” Columbia University School of Journalism:
Tow Center for Digital Journalism. Retrieved from http://towcenter.org/the-art-and-science-of-data-
driven-journalism/
Karlsen, J. and Stavelin, E. (2014) “Computational journalism in Norwegian newsrooms.” Journalism Practice,
8, 34–48.
Lewis, S. C. and Usher, N. (2013) “Open source and journalism: Toward new frameworks for imagining
news innovation.” Media, Culture & Society, 35(5), 602–619.
Lewis, S. C. and Usher, N. (2014) “Code, collaboration, and the future of journalism: A case study of the
Hacks/Hackers global network.” Digital Journalism, 2(3), 383–393.
Lewis, S. C. and Usher, N. (2016) “Trading zones, boundary objects, and the pursuit of news innovation: A
case study of journalists and programmers.” Convergence, 22(5), 543–560.
Lowrey, W. (2002) “Word people v. picture people: Normative differences and strategies for control over
work among newsroom subgroups.” Journalism, 5(4), 477–500.
Meyer, P. (2002) Precision Journalism: A Reporter’s Introduction to Social Science Methods. Lanham, MD: Row-
man & Littlefield.
Monmonier, M. (1989) Maps With the News: The Development of American Journalistic Cartography. Chicago,
IL: University of Chicago Press.
Nussbaum, E. (2009) “The new journalism: Goosing the gray lady.” New York Magazine. Retrieved from
http://nymag.com/news/features/all-new/53344/
Parasie, S. (2014) “Data-driven revelation? Epistemological tensions in investigative journalism in the age of
‘big data’.” Digital Journalism, 3(3): 364–380.
Parasie, S. and Dagiral, E. (2013) “Data-driven journalism and the public good:‘Computer-assisted-reporters’
and ‘programmer-journalists’ in Chicago.” New Media & Society, 15(6), 853–871.
Powers, M. (2012) “‘In forms that are familiar and yet-to-be invented’ American journalism and the dis-
course of technologically specific work.” Journal of Communication Inquiry, 36(1), 24–43.
Radcliffe, D. (2013) “Hyperlocal media and data journalism.” In J. Mair and R. Keeble (eds.), Data Journal-
ism: Mapping the Future. Bury St Edmunds: Abramis Academic Publishing (pp. 120–132).
Royal, C. (2010) “The journalist as programmer: A case study of the New York Times interactive news
technology department.” International Symposium on Online Journalism, Austin.
Silver, N. (2014) “What the Fox knows.” Fivethirtyeight. Retrieved from https://fivethirtyeight.com/
features/what-the-fox-knows/
Smit, G., De Haan, Y. and Buijs, L. (2014) “Visualizing news: Make it work.” Digital Journalism, 2(3),
344–354.
Stray, J. and Chambers, L. (2012) The Data Journalism Handbook: How Journalists Can Use Data to Improve
the News. Sebastapol, CA: O’Reilly Media.
Stromer-Galley, J. (2004) “Interactivity as product and interactivity as process.” Information Society, 20,
391–394.
Trédan, O. (2014) “Quand Le Journalisme Se Saisit Du Web : L’exemple Du Datajournalism.” [When
journalism takes hold of the web: the example of data journalism]. In F. Le Cam and D. Ruellan (eds.),
Changements Et Permanence Du Journalisme [Change and Permanence of Journalism]. Paris: L’Harmattan
(pp. 199–214).
Usher, N. (2016) Interactive Journalism: Hackers, Data, and Code. Chicago, IL: University of Illinois Press.
358
Weber, W. and Rall, H. (2013) “‘We are journalists’: Production practices, attitudes and a case study of the
New York times newsroom.” In W. Weber, M. Burmester and R. Tille (eds.), Interaktive Infografiken.
Wiesbaden, Germany: Springer (pp. 161–172).
Weinacht, S. and Spiller, R. (2014) “Data journalism in Germany.” Journalism, 59(4), 411–433.
Willis, D. (2011) “Interviewing Data”, The Scoop. Retrieved from http://thescoop.org/archives/2011/05/01/
interviewing-data/
359
27
JOURNALISTIC FREEDOM
AND THE SURVEILLANCE OF
JOURNALISTS POST-SNOWDEN
Paul Lashmar
A paradigmatic shift is sometimes revealed by an unanticipated and extraordinary event, and so

it was with Edward Snowden in 2013. A National Security Agency (NSA) contractor, Snowden
was so appalled at the exponential expansion of covert digital surveillance that he decided it was
his moral duty to inform the public, indeed the world. This he did from a hotel room in Hong
Kong when he gave a small group of selected journalists access to 1.7 million classified docu-
ments taken from the NSA. These documents revealed the global snooping capabilities of the
NSA and its ‘Five Eyes’ intelligence agency partners (ASIO in Australia, CSE in Canada, GCSB in
New Zealand, and the GCHQ in United Kingdom). The Five Eyes can vacuum up just about all
digital communications anywhere, anytime, and much else besides if they are so minded. Many
who take a deep interest in signals intelligence thought these Anglo-Saxon agencies had probably
increased their capabilities since 9/11, but even they were shocked when Snowden revealed the
sheer scale – it far exceeded any estimate of capability.
From Snowden’s leaked documents, journalist Glenn Greenwald discovered the Five Eyes’
mantra is “Collect it all”. In one article he quotes from his favorite NSA document – a favorite
because of its clarity in terms of just how comprehensive collection is:
At the top of the document, it says “new collection posture”. This is the NSA describ-
ing its new collection position, and right underneath is a really ugly, though helpful,
circle with six points on it.
(Bell et al., 2017: 45)
Greenwald goes on to detail how “Each of the six points has a different phrase that elaborates on
the ‘Collect It All’ mandate”, adding:
So you go clockwise around the circle, and the top it says “Sniff It All” and then it says
“Know It All,” “Collect It All,” “Process It All,” “Exploit It All”; and then the last one
is “Partner It All.”
(ibid: 45)
He continued:
360
Journalistic freedom and surveillance
This then is the institutional mandate for the NSA – it is collecting billions and billions
of telephone calls and emails every single day from populations and nations all over the
world including our own.
(ibid: 45)
Five years on from Snowden’s revelations, there has now been time to assess the impact of Snowden’s
controversial leak. In the months immediately after he went public (September 2013 – February
2014), I interviewed journalists from across the Five Eyes countries for their assessment of Snowden
and his revelations (Lashmar, 2016). I chose investigative journalists with national security reporting
experience who are more likely to have a deep knowledge about what Snowden meant for the wider
public and for journalism. They are also likely to be the journalists most ‘at risk’ from the surveillance
capabilities of these agencies. As the general counselor for BuzzFeed, Nabiha Syed observed:
There has always been some information asymmetry between reporters acting in the
public interest and powerful organizations – like government agencies – that possess
critical information. Increasingly, that imbalance is tilting against the interest of two
critical groups: national security reporters and independent journalists.
(Bell et al., 2017: 142)
My research cohort included reporters from both groups. In spring 2017 I went back to those I had
originally interviewed and, where possible, interviewed them again. There had been some changes;
one interviewee, Gavin McFadyen of the UK’s Centre for Investigative Journalism, had died, a sad
loss. Others had since moved away from national security reporting and felt they had nothing new to
add. I approached roughly 20 journalists and was able to interview 12 between April and May 2017.
There were at least two reporters from each of the Five Eyes countries. The journalists were; Andrew
Fowler (formerly on Australia Broadcasting Corporation’s Four Corners program) and Dylan Welch
(ABC’s 7.30 show) from Australia; Jim Bronskill (Canada Press), Andrew Mitrovica (freelance), and
David Seglins (CBC) for Canada; David Fisher (New Zealand Herald) and Nick Hager (freelance
and New Zealand’s leading investigative reporter) for New Zealand; Duncan Campbell (intelligence
expert and freelance journalist), Meirion Jones (ex-BBC and Bureau of Investigative Journalism),
and Peter Taylor (BBC Panorama) in the UK; and Scott Shane (New York Times) and Jeff Richelson
(National Security Archive) in the US. All have reported on intelligence agency excesses. At least
three (Campbell, Hager, and Fowler) have been subject to security agency raids in their careers as a
result of their stories. All have reported on or used the Snowden documents. One had met Snowden
(Taylor), and others worked with the Greenwald team to some extent.
The semi-structured interviews I conducted with them form the basis of the research for this
chapter. In addition, I conducted a review of books, reports, chapters, and papers on the impact
of Snowden for journalists and their sources (see: Bell et al., 2017; Fowler, 2015; Kuehn, 2016;
Bauman et al., 2014; Moore, 2014).
Perceptions of Snowden
In the immediate aftermath of Edward Snowden going public, the former NSA and CIA employee
was applauded by some commentators as a hero, but others accused him of being a traitor and
worse. The former director of the NSA, General Keith B. Alexander, stated that his leaks had
resulted in “the greatest damage to our combined nations’ intelligence systems that we have ever
suffered” (MacAskill, 2014). British intelligence has spoken of areas of the world having “gone
361
Paul Lashmar
dark” and of disruption caused to intelligence gathering. Back in 2013, some commentators and
journalists posited that Snowden was working for Russian or Chinese intelligence. However, that
criticism seems to have receded, and whether critics are for or against him, his sincerity is rarely
now questioned. At the time of writing he remains in Russia and would face serious charges if
he returned to the US. In 2015, Peter Taylor, one of the BBC’s most experienced current affairs
journalists, made a Panorama program about surveillance.
I was fortunate enough to meet Edward Snowden in Moscow and spent about two
hours with him. Before I met him, I was never quite sure about him. When I met him
I was in no doubt about his sincerity, motivation, and fierce determination to out the
things he thought the public should know. He had a powerful feeling the public was
being kept in the dark.
Perhaps not surprisingly, as investigative journalists, all of the interviewees were in favor of what
Snowden had done and felt that releasing information to the public was important. Some inter-
viewees described Snowden as a hero (Mitrovica, Seglins). Andrew Mitrovica said, “We owe
Snowden a debt of gratitude for risking his safety and freedom”. Mitrovica said he was frustrated
in the years before Snowden that the public and editors were not taking surveillance seriously.
But Snowden’s leaks changed that: “The public imagination got caught up in what he was doing.
I thank him for trying to help make these things known to the public”.
The name Snowden, it seems, has also become shorthand for global mass surveillance.
Impact of Snowden
Most of the interviewees felt that Snowden’s revelations had reached a global audience and that, in
terms of considering privacy and surveillance, there is a before and after Snowden. David Fisher
said, “In the intelligence and the security space there is far greater awareness of surveillance issues
and privacy issues”. He believed that the public, at least in New Zealand, now has no expectation
of privacy. Referring to the Five Eyes eavesdropping agencies, Fisher said: “Snowden has contex-
tualized what we are dealing with now. The power they have, if they choose to use it, is awesome”.
Asked about the impact his revelations had, the responses from those interviewed were var-
ied. Peter Taylor felt that what Snowden had done was “hugely important”. Mitrovica felt the
releases had impacted hugely on the public and “made what was going on clear”. Scott Shane
said Snowden had raised awareness. The US interviewees reported a mixed reaction from the US
public. Shane noted that in response to the Snowden revelations:
About half of Americans and about half of Congress were unhappy with some of what
was exposed, primarily the phone call metadata, and the Obama administration and
Congress scaled things back and changed the procedure to increase privacy protection
for Americans and made it less possible for the government to collect and store data on
millions of Americans.
Shane emphasized that the NSA is so powerful that it needs to be closely monitored: “The
capabilities of NSA obviously are so consequential that everybody needs to keep a close track”.
Duncan Campbell, who is UK journalism’s foremost independent expert on signals intel-
ligence (SIGINT), was surprised by the scale of the surveillance capability revealed by the docu-
ments. The scholarly Jeff Richelson, who was one of the leading American independent experts
on SIGINT, said the “vastness” of the Five Eyes operation did surprise him. He said the documents:
362
dramatically shifted the understanding of the nature of Sigint by the US and the British
in terms both of the reach of it but also in terms of targeting digital networks and extract-
ing intelligence from digital networks, and the lengths they went to and had gone to, and
presumably are going, to get that information in terms of not simply basic hacking or
passive intercept but also implants or planting devices in computers they have diverted.
As to what the impact on the public had been, Richelson did not feel qualified to comment. The
UK’s Meirion Jones was the most skeptical of the interviewees concerning impact and felt the
revelations merely confirmed what the public and journalists had suspected. Canada’s Dave Seg-
lins, who is an experienced CBC broadcaster on the national security beat, said he was “shocked
at the initial stories”, but he felt that Canadians were less skeptical than Americans of national
authorities. Fellow Canadian Andrew Mitrovica felt Snowden had become a major cultural fig-
ure in the world. Jim Bronskill agreed that Snowden had reset the public debate: “It was useful
and still is in the sense people are more mindful of the fact there are agencies collecting intel-
ligence, and with modern tools it is infinitely easier to do, and it is happening”.
Nicky Hager, who has had a number of run-ins with intelligence agencies in New Zealand
over his investigations, stated that the Snowden revelations were “absolutely incredible” and felt
“there had been a high level of public support for Snowden”. He commented: “The New Zea-
land public at large had a much larger reaction to the overall world news than stories about New
Zealand”. David Fisher and Hager both said there was, initially, a big reaction in New Zealand,
with town hall meetings and public demonstrations in the months after Snowden’s leaks. Author
of The War on Journalism (2015), Andrew Fowler, said the Australian reaction was divided. The
public, he observed: “Have been, I would say quite supportive, in the sense that they have always
believed their communications were being interfered with and their data might not be safe; but
this provided absolute proof of it”. Fowler and Dylan Welch felt that the Australian public did
not react strongly, as they are very conservative in their views when it comes to intelligence issues.
In the UK, Meirion Jones said the public reaction was mixed, but believes that if anything, the
revelations resulted in sizeable part of the UK public having increased pride in the intelligence
services: “That British intelligence is still something important, that they are ranked up there
with the Americans, dirtier than the Americans, it appeals to a James Bond aura, for them it
wasn’t negative”.
One of the most interesting aspects of the UK reaction is that in June 2013 much of the UK
press, particularly the Telegraph, the Sun, and The Daily Mail, turned on the Guardian for printing
Snowden documents and sided with the government and intelligence communities’ condemna-
tion. Campbell observed that coverage of Snowden in the UK was: “Highly slanted and quite
significant in that the voice of Snowden was muted, so the message was really only conveyed by
The Guardian and yes, the BBC, but muted through the onerous processes of purported balance”.
In the other Five Eyes countries, there was much less of a tendency for the other news media
to turn on the news organizations that had published exclusive Snowden material.
I asked interviewees how they would measure Snowden as a paradigmatic event – for instance,
how they would compare it with Watergate, perhaps the most recognized news story where the
news media had clashed with the secret state. Canada’s Dave Seglins and Andrew Mitrovica both
felt the Snowden affair was of global significance. Seglins stated:
Snowden was more important than Watergate. Watergate pierced the veneer of moral
leadership in the US but had less of an impact on the citizens of the world. I think the
Snowden revelations instantly ripped the shroud of secrecy from activities of the Five
Eyes countries but also made the entire world aware, realize what was technologically
363
Paul Lashmar
capable, possible. So I think it has far reaching consequences for people around the
world and not just in the Five Eyes countries.
The collection obsession

Duncan Campbell was very concerned that Snowden coverage had been too focused on bulk
collection and too little had been said about what GCHQ did with the data it collected. Camp-
bell pointed out GCHQ has “customers”: “It’s a business, that’s its raison d-être”. He noted GCHQ
has customer relations teams, a sales force, and delivery drivers – and for all the reporting of
Snowden’s UK documents highlighting GCHQ’s role:
It only focuses on one aspect – collection – the systems that steal all our data. It doesn’t
look at the intelligence process in the round, because for the most part that is what
these documents see, and generally when they did, with some salient exceptions, that is
not what the journalists went for. It seemed sexy to describe massive scoops on internet
cables and the factors of scale, which is truly astonishing and so on. That criticism can
probably be made of me.
He continued:
The fact of the matter is, to understand in its context, the harm or good that may be
done by signal intelligence agencies, you have to look at the tasking, the collection man-
agement, the analysis process, and above all the consumer reporting channel because the
core interactions are not collection directed against the citizen, or the business or the tar-
get – they are the customer – who the government pays for – the customer gets spy data.
He then observed:
Then the second interaction that matters is – what is done with it? So if the Snowden
documents, which they do on some occasions, speak to all of those processes, they
clearly have more force and show more of the picture, and when they don’t, they cer-
tainly show collection capability and scale. But what is done with it?
Campbell made the point that it is important to understand who gets the raw intelligence from
GCHQ surveillance – whether it is the Defence Intelligence (DI), MI6, MI5, or the police – and
what they do with it and whether it infringes the target’s rights under Article 8. This is a point
that is equally relevant to other Five Eyes countries. Indeed, the intelligence lobby was frustrated
by this post-Snowden emphasis on collection, which they describe as bulk collection, and argue
that is not the same as mass surveillance, as they filter out most data to focus on targets. These
are indeed set by their customers and do not eavesdrop on the public at large. However, we have
little idea how collected data impacts on the civil liberties of ‘targets’.
Impact on journalists
As UNESCO researchers noted:
While the rapidly emerging digital environment offers great opportunities for journal-
ists to investigate and report information in the public interest, it also poses particular
challenges regarding the privacy and safety of journalistic sources
(UNESCO, 2017: 5)
364
The Snowden documents revealed that some journalists were the targets of intelligence agencies.
In 2013, Der Spiegel reported that the NSA had intercepted, read, and analyzed internal com-
munications at Al Jazeera that had been encrypted by the news organization (Der Spiegel, 2013).
There are many other examples. In early 2015, the Guardian published a Snowden document that
revealed that a GCHQ information security assessment listed “investigative journalists” in a
threat hierarchy (Ball, 2015). Such incursions on journalists’ digital communications compromise
the globally established ethical obligation upon journalists to avoid revealing the identity of their
confidential sources.
The issue of source protection has come to intersect with the issues of mass surveil-
lance, targeted surveillance, data retention, the spill-over effects of antiterrorism/
national security legislation, and the role of third party Internet companies known as
“intermediaries”.
(UNESCO, 2017: 18)
The Pew Center’s research in the United States found that 64% of investigative journalists sur-
veyed believed that the U.S. government collected data about their communications. The figure
rose to 71% among national political reporters and those who report foreign affairs and national
security issues. Ninety percent of U.S. investigative journalists who responded to the Pew survey
believed that their ISP would routinely share their data with the NSA, while more than 70%
reported that they had little confidence in ISPs’ ability to protect their data (UNESCO, 2017:
103). Nearly all my interviewees felt that Snowden’s revelations had a big impact on journalists
generally and had raised very serious questions about whether journalists could protect their
sources. Shane said: “There was more awareness amongst journalists”. Back in 2014, Duncan
Campbell counseled it was important to keep things in perspective, and only a relatively small
number of journalists are likely to be subject to surveillance by the NSA network: “The impact
of Snowden’s revelations should not really be overstated for journalism, because the most critical
aspect relates to the conduct of the intelligence” (quoted in Lashmar, 2016).
In 2016, Campbell maintained this position. New Zealand’s David Fisher observed much
the same for most investigations but stated it is a different story if you are investigating Five Eyes
agencies: “If you are fucking with them there is no way they are not going to find out”. Other-
wise sensible trade craft will do, he said:
If it’s the spies you are messing with – they are going to track every single bit of metadata
you’ve got. They are going to intercept every bit of commination you’ve got. When
you are out of house, they will break in and download everything on your computers.
Fisher thought that sources are more alert: “There has been a chilling effect”. Meirion Jones and
Taylor said the Snowden revelations had impacted on sources. Shane reported it had an impact
on sources, but this had been somewhat negated by Trump, where sources are queuing up to dish
the dirt on the White House. Fowler was concerned how cavalier some sources are and that he
still gets emails that could incriminate the sender. Mitrovica was bullish: “It’s not had a chilling
effect on me” – nor was he worried about the impact on sources – “I think some sources have
been emboldened by Snowden”.
Methods
When it came to protecting their sources, the interviewees’ reaction to changing procedures was
mixed. Some (Fowler, Welch, Seglins, Fisher, Shane) said they had tightened up their security
since Snowden to protect their sources. Shane said he had become more cautious but made
365
Paul Lashmar
the point that in the US it was not just Snowden’s revelations that influenced journalists but
the Obama administration’s prosecution of journalist sources that had impacted on journalists.
Indeed, one of Shane’s own contacts, John Kiriakou, had been prosecuted and jailed. David Seg-
lins said that working with and reading the Snowden documents had fundamentally changed his
understanding of operational security as a journalist:
Everything from storage of documents to the use of encryption, encrypted communi-

cation, encrypted data storage, to how our mobile devices are potential listening devices
and how that affects a journalist’s ability to travel to places, meet sources, have discus-
sions with absolute certainty we are not being recorded or monitored or tracked.
Hager and Jones said that they had always employed rigorous source protections methods, so had
no plan to change. “I would rather lose a story than a source”, said Jones. Some interviewees
have incorporated new counter-surveillance digital methods routinely into their work. Encryp-
tion has become a regular tool in a way it was not before Snowden. Shane said he uses encrypted
email. Some are using PGP technology (Fowler, Welch, Hager, Seglins, Fisher, Jones, Shane) as
necessary, and some use TOR for browsing (Seglins, Jones, Hager). Like Shane, some also use
encrypted phone apps:
One of the things that has changed since we last talked is the proliferation of encryp-
tion communication apps. Many of us have run through the various ones, Silent Circle,
WhatsApp, Signal, so there is an increasing availability of encrypted communications.
I’m certainly more aware of what I am putting into a storable electronic record.
Some interviewees now include the PGP key and other encryption contact information into
their email or social media addresses. Welch said this told potential sources that he is serious about
source protection:
I list it all. I tell people where they can find my PGP, my public key. I tell them I have every
single one of the encrypted apps on my phone. I use them a lot. I don’t try to hide it.
Some reported their news organizations had decided to set up Secure Drop (a secure and
encrypted inbox) facilities for potential sources to send material to (this includes Sydney Morning
Herald, Canadian Broadcasting Corporation, New York Times, the Bureau for Investigative Jour-
nalism). Other organizations have decided against it (ABC in Australia, New Zealand Herald, and
Canada News). Fisher pointed out that using encryption can be a ‘red flag’ to interested intel-
ligence agencies that you are communicating with someone they might be interested in. Most
journalists who use encryption said it was only a partial solution to be used with care. There was
clear concern that the Five Eyes may have found ways to break encryption. Taylor said that while
examining the Snowden documents:
One thing that really surprised me, and really it should not have done, was that he had
GCHQ material from a training manual. The intelligence service GCHQ could tap
into your phone by planting malware inside it and listen to your conversations and take
photos of you and whoever you were with, even though your phone appeared to be
off. That really shook me.
Scott Shane said the New York Times made a lot of effort to protect sources and had recently
appointed a newsroom security adviser, and journalists were given training and advice from
366
lawyers. Almost all interviewees emphasized the importance of ‘non-digital’ means of communi-
cation with sources, making sure there was no digital footprint of the meeting – leaving mobile
phones and laptops at home. Jones said it was important to tell sources that you cannot guarantee
to protect them, though you would do your best.
Damage to national security

When the Snowden material was published, politicians and intelligence chiefs attacked journalists
for publishing the classified documents, and it was not uncommon for these critics to accuse editors
of putting lives in danger and damaging the ability of national security agencies to monitor and
deter terrorists. Interviewees were varied in their responses about whether Snowden had damaged
national security. Campbell thought there might have been an impact on operational effectiveness.
Richelson thought some techniques might have been revealed. After British intelligence claimed to
him that Snowden had put lives in danger, Taylor asked them to identify an example. They failed to
do so, saying: “We can’t comment on such information”. He does feel that there was some general
damage, but Snowden also performed a public service. Often robust in his position, Fisher in New
Zealand took the view that claims of damage were “a load of bollocks”, adding: “If there had been
any real consequences of that occurring that would have been rammed down all our throats”.
Fowler in Australia did not believe there had been any damage. Neither did Mitrovica in Canada:
“No it’s a myth, it hasn’t damaged their effectiveness. They always trot this out all the time”.
Seglins noted that if Canadians were to have confidence in their national security:
Part of national security is confidence in democracy, confidence in judicial oversight,

confidence in our law enforcement and intelligence agencies. And if they were operat-
ing in the dark, and/or illegally, and/or counter to public trust, then I would say the
Snowden leaks have enhanced national security because we were growing for a long
time in the dark, not knowing what our law enforcement agencies were up to. And that
secrecy and that vacuum of public knowledge and oversight are where corruptions and
breakdowns occur, we know that.
Shane said there may have been some damage to specific operations, but that was the price of
democratic debate:
If you live inside those bureaucracies you begin to think that it’s the end of the world
when someone learns something about what you are up to. But these trade-offs exist
in any democracy. We would all be safer from terrorism if there was not restrictions on
these agencies and they recorded and stored all America’s conversations and emails on a
permanent basis. We would all be safe from terrorism. On the other hand, that’s not the
way we want to live, and I think these agencies sometimes forget that.
Another major complaint from politicians and intelligence chiefs is that due to Snowden’s rev-
elations, not only are journalists, sources, and the public much more likely to use encryption
but so are criminals and terrorists. Agencies complain that some of their key targets “have gone
dark” because of encryption. Taylor made the point: “Remember Snowden was in early days
of encryption. Encryption is now the big problem”. Indeed, Snowden has publicly supported
people using encryption.
It is worth noting none of those interviewed disputed that there is a role for intelligence
agencies in tackling terrorism. Duncan Campbell took the view that “from available evidence”,
British intelligence was doing a good job.
367
Paul Lashmar
Responsibility of the media

At the time of writing, 6,000 Snowden documents have been released into the public domain
through a set of rigid procedures by a team of journalists led by Glenn Greenwald and Laura
Poitras. No interviewee thought the media that dealt with the Snowden material had acted
irresponsibly. Taylor thought the Guardian, as the first news organization to publish Snowden
documents, had been professional. Those who had used documents had checked with formal
intelligence links to make sure that they were not going to do any inadvertent damage. In some
cases extensive and repeated checks were made, and occasionally certain aspects of stories were
dropped if the news organization thought the national security people had made a compelling
case. Rusbridger has said the Guardian had over 100 contacts with the authorities before publica-
tion (Ponsford, 2014). Shane, Taylor, and Seglins reported detailed conversations. Redaction was
also used. Journalists (Taylor, Shane, Seglins) who dealt with these negotiations were critical of
the initial position of intelligence chiefs, which was to say nothing should be published. Over
a period of time, the negotiations became more sensible, and the intelligence agencies realized
that the journalists were more likely to listen if they made a good case for an element not being
published. Hager was more skeptical about any opening up:
There wasn’t a discussion. The most frustrating thing about intelligence as a public
policy issue, or as part of the life of the country, which it plainly is, is that there is a sense
of entitlement on the side of the authorities not to engage in debate, and they know
perfectly well that while that has an operational element it is also highly convenient.
Hager did not confer with the intelligence agencies about the content of his Snowden documents
because, he says, the agencies were incapable of a sensible discussion. Instead he went to “a lot
of trouble to make sure we did not tell stories that would really harm something that really mat-
ters”. Hager said there were aspects of New Zealand’s surveillance operations in Bangladesh “that
were really dodgy” and, if revealed, could have “brought a dangerous backlash to New Zealand”,
so he withheld the details.
Taylor said that despite the hard line initially taken by the intelligence chiefs that the BBC
should not broadcast Snowden documents, that after the Panorama program went out they
seemed to think it was fair. Fowler stated he did not think that journalists should refer back to
the agencies:
I do think they acted responsibly. In fact, my argument is that I think the journalists
acted, what I would call, without being too cute, too responsibly. I would trust the judg-
ment of a journalist whether or not to publish the material rather than running it past
government, as seems to have been case with the Snowden documents. . . . I don’t think
a journalist should need to do that, make a call on that and live with that.
Laws
With the exception of the United States, interviewees in the other Five Eyes countries said new
laws have been passed to enhance the power and scope of the intelligence agencies. In some cases
the laws were on their way already at the time of Snowden releasing the documents, and in other
cases the Snowden affair was either part or all of the reason for new laws. Nearly all the inter-
viewees felt that the laws gave excessive power to the national security community. In some cases
they took the view the laws were draconian. In the UK as a result of the Snowden revelations, in
368
February 2015, the intelligence watchdog, the Investigatory Powers Tribunal (IPT), found GCHQ
had breached human rights conventions in relation to the UK’s access to the NSA’s bulk data
collection program. Nonetheless the Investigatory Powers Act (a tougher and revised successor to
the controversial ‘Snooper’s Charter’) passed into law in 2016. As the act became law, Snowden
tweeted: “The UK has just legalized the most extreme surveillance in the history of western
democracy” (Snowden, 2016). The IPA is just one example of national security bodies being given
more powers. In July 2014, the UK government fast-tracked a new Data Retention and Investiga-
tory Powers Act as ‘emergency legislation’ and rushed it through parliament in a single day. The
act was designed to revise UK data retention law in response to an April 2014 ruling by the Euro-
pean Court of Justice (ECJ) invalidating the 2009 Data Retention Directive. The law not only
provides for ongoing blanket retention of communications data of UK residents, in direct con-
tradiction with the ECJ ruling, it also extends the reach of UK interception powers by enabling
the government to require companies based outside of the United Kingdom to comply with the
UK’s warrants. In addition, the UK’s Law Commission has carried out for the government a con-
sultation to update the Official Secrets Act into an Espionage Act. Critics say the initial proposals
suggest that journalists, sources, and whistle-blowers will be vulnerable to imprisonment.
Some interviewees felt the Snowden leaks had given governments the justification to toughen
the laws. In the UK, Taylor said: “The difference that Snowden has made is that we now have leg-
islation that is far more embracing than its predecessor”. David Fisher said that new legislation was
proposed in New Zealand in 2014 and there were protests, but when a more “enhanced” legisla-
tion was passed in 2017, there were no protests. For Australia, Fowler said: “The government has
introduced tough new laws as a result of Snowden, citing Snowden as one of the reasons why they
had introduced them”. As critics have noted, the initial draft of Australia’s metadata legislation
arrived without a dataset or safeguards. A review by the Parliamentary Joint Committee on Intel-
ligence and Security (PJCIS) added 39 recommendations, including a request for a separate review
on the impacts on journalists, the inclusion of a dataset, and additional oversight provisions. A
mandatory two-year metadata-retention scheme was among the many anti-terrorism measures. In
both Australia and the UK there have been examples of police using existing legislation, the Regu-
lation of Investigatory Powers Act (RIPA) in the UK and the Telecommunications (Interception
and Access) Act 1979 in Australia, in secret to identify sources by accessing journalists’ metadata.
Andrew Mitrovica said the Canadian government had enacted “draconian pieces of anti-
terrorism legalization” including anti-terrorism law C51 – “C51 broadens state powers to surveil
individuals and broadens what is considered dissent”. He added: “It’s not surprising; governments
inevitably act this way”. Seglins said that the Trudeau government has conducted a review of
legislation, and it is likely that it will include enhanced judicial oversight of the Canadian intel-
ligence agencies. Despite all the new legislation, there has been little improvement for journalists
and their sources. In 2016 in Canada, the courts have ordered a Vice reporter to hand over docu-
ments or face jail. Vice journalist Ben Makuch has been fighting a police order to hand over his
correspondence with Farah Mohamed Shirdon – a man who left Calgary to allegedly fight with
ISIS. Makuch refuses and was still in the court process at the time of writing.
The USA Freedom Act, which was passed by the House of Representatives in May 2015,
reduces government bulk collection of U.S. phone records. Americans’ phone records will still
be hoovered up – but now by the telephone companies, not the NSA – and access to them will
require a warranting process. Elements of transparency around government surveillance and the
operations of the secret FISA court will be introduced.
Some interviewees noted that these laws made little or no provision for journalists undertak-
ing their fourth estate role. The UNESCO report noted that, across 121 countries, technologi-
cal developments and a change in operational methods of police and intelligence services are
369
Paul Lashmar
redefining the legal classification of privacy and journalistic privilege internationally. The report
also notes that alongside digital developments, in less than a decade, increasingly restrictive anti-
terrorism and national security legislation has been passed into law that has or will override the
existing legal protections, including those known as ‘shield laws’:
This arises from moves to broaden the scope of “classified” information and exceptions to
coverage, and to criminalize all disclosure of “secret” information (including in some cases,
the publication thereof) irrespective of public interest or whistle-blowing considerations.
(2017: 20)
UNESCO adds:
The result of the increasing risk to both journalists and their sources is a further constrain-
ing, or “chilling”, of public interest journalism dependent upon confidential sources.
(ibid: 20)
Raising an ontological dilemma, Nicky Hager observed that knowledge of global surveillance
may well have a profound effect on citizens’ behavior and that journalists publicizing the issues
might well not help. Hager felt that journalists have to report on intelligence and its excesses, but
he also worried about being part of the chilling effect:
I know, on the one hand, that unless we publicize and debate and kind of have real
information rather than just vague fears, there can be no real progress. But on the other
hand, to publicize is to add to this chilling effect, and I worry about that. I think it is a
really important issue.
In Australia Fowler thought Snowden’s revelations had a definite chilling effect on people’s
behavior: “If it has changed my behavior, it will have certainly changed the behavior of people
I would normally talk to”.
Conclusions
As Taylor, Hager, and others pointed out, this discussion over the surveillance state had largely
closed down after 9/11. Twelve years later, Snowden revealed that we had sleepwalked into a
world where total digital surveillance was not only possible but was happening. What the inter-
viewees just about all agreed about was that Snowden had reset the public discussion for better or
worse. Taylor thought Snowden’s act in 2013 was “hugely important”, adding, “It was of its time
because secrecy had always been an issue, but at that particular time it had reached a head, and
that again is one of the reasons why Snowden revealed it. He was genuinely shocked”.
Interviewees felt that the changes in intelligence agencies’ powers and capabilities were so great
they needed to be referred back to the public, even if only for affirmation. Questions remain con-
cerning whether the techno-optimistic focus of SIGINT agencies to ever expand the technology
and ‘collect it all’ is the best means of deterring terrorism. In the UK and Europe, terrorism plots
have been prevented, but there is a rise in the number of successful ‘home grown’ attacks.
The post-Snowden world is a troubling place for investigative journalists, whistle-blowers,
and sources. It is clear that all five governments, whatever they said in public, had little respect
for journalists’ fourth estate role and were unmoved that the national security agenda had a
by-product of making the job of journalists even harder. As the UNESCO report noted, the
problem has grown in many countries, with tougher anti-terrorism laws that allow for access to
370
journalists’ records and enforce assistance. State secrets acts are increasingly broadly defined and
criminalize journalists who publish leaked information. This “occurs where it is un-checked
by measures designed to preserve fundamental rights to freedom of expression and privacy”.
UNESCO adds: “In practice, this leads to what can be identified as a ‘trumping effect’, where
national security and antiterrorism legislation effectively take precedence over legal and norma-
tive protections for confidential journalistic sources” (2017: 19).
From 2013 onwards in the United States, there was at least a discussion about the tensions
between surveillance, privacy, and freedom of expression. By comparison, the UK government,
intelligence chiefs, and even some editors took the approach to those who were concerned about
Snowden’s revelations: ‘Move along please, there is nothing to see’. It was a monumental arrogance
to decide that the public and journalists should not have the right to discuss such a major political
change as the development of an infrastructure capable of total surveillance. Dressing it up as a nec-
essary response to terrorism is just not good enough. Far from there being ‘nothing to see’, we have
moved into a different type of society. The intelligence lobby are playing a ‘dead bat’ to the critical
audience, while behind the scenes they lobby hard for further powers, resources, and capabilities.
That the government and intelligence lobby do not even want to debate compelling concerns
that we have moved into a digital surveillance state is not evidence of a strong democratic gov-
ernment at work but, to the contrary, a victory for terrorism. To have so fundamentally changed
the nature of UK society, to have made it so fearful, is a win for the terrorist and a defeat for a
democratic nation. That governments have become more authoritarian is also demonstrated by
the failure to make provision for journalists doing their fourth estate job, supposedly a vital inde-
pendent oversight mechanism in a democracy. There is a clear drive to close down journalistic
public interest endeavors for investigating malfeasance by the state, and especially when it comes
to the excesses of the secret state. This is not just an issue for the Five Eyes countries but sets
the tone for other countries, many of which are following suit, as the UNESCO report clearly
shows. We can blame Snowden, but he did not set up the Five Eyes – he was the messenger, not
the architect.
Journalists face an existential crisis. In the past, they could offer confidential sources a reason-
able promise of anonymity provided that sensible precautions were taken. Now journalists have
no precise idea of the level of risk their sources will face. Duncan Campbell makes the propor-
tionality point that it is unlikely that journalists or their sources will be under digital surveillance
unless they are delving into very sensitive areas like national security. The fact remains that jour-
nalists do not know if they are subject to surveillance.
Campbell is also correct that whether you call it bulk collection or mass surveillance, we are
currently too focused on the sheer scale of surveillance, but we know nothing about those who
are being targeted and the impact on their civil rights. In each of the Five Eyes, there have been
recent examples of security services exceeding their remit. Perhaps the most telling comment
in all the interviews and one that demonstrates the new paradigm came from Nicky Hager. He
is profoundly concerned over the impact global surveillance may have on the citizen’s behavior,
and because of this “they grow differently as a person because they have a background sense of
the lack of privacy”. Privacy is a fundamental human condition, and we do not know yet the
existential consequences of undermining the public’s fundamental sense of privacy. This can be
laid not only at the feet of Five Eyes but also of the internet giants like Google and Facebook, but
the national security involvement brings a totalitarian element to the debate. If intelligence and
surveillance are impacting negatively on ordinary people’s lives, then we need to stop and debate
this, not as the UK government and other Five Eyes’ governments have done, ignoring dissent.
What follows from ignoring dissent is the suppression of dissenters, and the tools for repression
are now firmly in place to do exactly that.
371
Paul Lashmar
Further reading
The writing of this chapter coincided with the publication of two excellent companion publica-
tions, Bell et al.’s Journalism after Snowden (2017) and UNESCO’s Protecting Journalism Sources in
the Digital Age (2017). I would also recommend Townend and Danbury’s Protecting Sources and
Whistleblowers in the Digital Age, an Information Law and Policy Centre report in association with
Guardian News and Media. The issues in this chapter will be developed in my forthcoming book:
Spin, Spies and the Fourth Estate: British Intelligence and the Media. Edinburgh University Press.
Acknowledgment
Jeff Richelson, one of the finest historians of U.S. intelligence, died in November 2017. This
chapter is dedicated to him.
References
Ball, J. (2015, January 19) “GCHQ captured emails of journalists from top international media.” The Guard-
ian. Retrieved from www.theguardian.com/uknews/2015/jan/19/gchq-intercepted-emails-journalists-
ny-times-bbc-guardian-lemonde-reuters-nbc-washington-post (accessed 20 February).
Bauman, Z. et al. (2014) “After Snowden: Rethinking the impact of surveillance.” International Political
Sociology, 8(2), 121–140.
Bell, E., Owen, T., Korana, S. and Henrichsen, J. (2017) Journalism After Snowden: The Future of the Free Press
in the Surveillance State. New York, NY: Columbia University Press.
Der Spiegel. (2013, August 31) Snowden Document: NSA Spied on Al Jazeera Communications. Retrieved from
www.spiegel.de/international/world/nsa-spiedon-al-jazeera-communications-snowden-document-
a-919681.html (accessed 26 May 2017).
Fisher, D. (2014, December 10) “Why NZ spy chiefs can no longer get away with saying ‘we can neither
confirm nor deny’.” New Zealand Herald. Retrieved from www.nzherald.co.nz/nz/news/article.cfm?c_
id=1&objectid=11371394 (accessed 29 June 2017).
Fowler, A. (2015) The War on Journalism. Sydney: William Heinemann.
Kuehn, K. M. (2016) The Post-Snowden Era: Mass Surveillance and Privacy in New Zealand. Wellington: BWB
Texts.
Lashmar, P. (2016) “No more sources? The impact of Snowden’s revelations on journalists and their confi-
dential sources.” Journalism Practice, 11(6), 665–688.
MacAskill, E. (2014, June 24) “New NSA chief says ‘sky not falling down’ after Snowden revelations.” The
Guardian.
Moore, M. (2014) “RIP RIPA? Snowden, surveillance, and the inadequacies of our existing legal framework.”
The Political Quarterly, 85(2), 125–113.
Ponsford, D. (2014) “Rusbridger on how no journalist’s sources are safe, joining IPSO and why he would
have kept News of the World open.” Press Gazette.
Snowden, E. (2016, November 17) “The UK has just legalized the most extreme surveillance in the his-
tory of western democracy.” Twitter Feed. Retrieved from https://twitter.com/i/web/status/799371
508808302596 (accessed 29 June 2017).
Townend, J. and Danbury, R. (2017) Protecting Sources and Whistleblowers in the Digital Age. doi: 10.2139/
ssrn.2961911
UNESCO. (2017) “Protecting journalism sources in the digital age.” UNESCO Series on Internet Freedom.
Retrieved from http://en.unesco.org/news/unesco-releases-new-publication-protecting-journalism-
sources-digital-age (accessed 25 May 2017).
372
PART VI
Minority voices and protest

Narratives of freedom and resistance
28
HOW AND WHY POP-UP NEWS
ECOLOGIES COME INTO BEING
Melissa Wall
Introduction
Much has been written about the instability of journalism in the twenty-first century, with
observers calling for new understandings, new concepts, and new theories to try to better
understand the discipline (Franklin and Eldridge, 2016; Witschge et al., 2016). News practices are
in a state of dizzying flux as social media platforms enable collaboration and greater interactions
between professionals, amateurs, and those in between and, at the same time, disrupt and potentially
diminish long-standing expectations of what news can and should be. The borders of journalism
have become increasingly permeable, with a host of other actors now also contributing to the
increasingly unpredictable production of news and information around the world. This lack of
fixity for journalism and its traditional norms has meant that those who study journalism are also
reconsidering long-standing assumptions about news and its boundaries, frequently with the aim to
travel beyond the current “edge” of the discipline (Carlson, 2015; Eldridge and Franklin, 2016: 8).
This chapter responds to these calls by aiming to further develop the concept of the tem-
porary or pop-up news ecology, a term that seeks to take into account the rapid changes and
liquidity in journalism (see Zamith, this volume, Chapter 7) as well as the move to rethink the
ways research has tended to limit itself to examining news within individual organizational
structures. The pop-up news ecology was originally introduced to characterize the dramatic
rise of citizen, activist, and other oppositional forms of media in Syria, when an entirely new
network of previously unheard voices sprang to life to offer a vastly different alternative to the
information being distributed by the authoritarian news system controlled by the Syrian state
(Wall and El Zahed, 2015). How this concept might apply to other cases situated within quite
different political, geographic, and social contexts is the goal animating the essay that follows
as it seeks to respond to change, disruption, and “deconstruction” in the news field (Ahva and
Steenson, 2016: 25).
News ecology: growth of a concept

Research clearly shows that journalism is being carried out within more fluid and contingent
spaces in which a range of nontraditional participants such as NGOs, grassroots activists, and even
nonhuman actants such as algorithms, drones, etc., now collect and produce news (Blaagaard,
2015; Holton et al., 2013; Jensen et al., 2016; Lewis and Westlund, 2015; Usher, 2017). Taking
375
Melissa Wall
into account these changes, scholars have asked that we broaden our considerations of news
beyond the traditional investigations of single institutions such as examining how BBC or the
New York Times operate (Reese, 2016; Sjøvaag and Karlsson, 2016). The concept of an ecology is
one approach used to grapple with the move away from fixed containers for news production.
Wahl-Jorgensen (2016) argues that as a “sensitizing concept”, news ecology can help us clar-
ify the increasingly complex environment within which journalism takes place, taking our focus
away from technology’s particular affordances to consider broader processes at work (15). In
Wahl-Jorgensen’s (2016) genealogy of the news ecology concept, the origins begin with Robert
Park and the Chicago School’s Progressive era attempts to understand immigrant communities
and their communication practices in urban spaces. Many decades later, building on McLuhan’s
work, Postman highlighted the idea of media ecologies as environments, launching the media
ecologies subarea of research (Strate, 2016). Other researchers taking a different perspective of the
ecological metaphor, such as Nardi and O’Day (1999), deployed the term information ecology.
They defined an information ecology as a “system of people, practices, values and technologies
in particular local environments” (Nardi and O’Day, 1999: 55). They further identified four key
characteristics of an information ecology: (a) relationships among actors; (b) diversity, with a
range of different actors occupying niches; (c) coevolution, in which change does not occur in
isolation; (d) “keystone species” or central actors; (e) locality not necessarily bound by geography
but by “spheres of influence and commitment” (55–57).
Applying the concept to journalism specifically, researchers have more recently used ecol-
ogy as a lens to assess the complexities of twenty-first-century news practices. Anderson (2013)
applied the concept of news ecology to show how news organizations in a specific city experi-
enced new levels of connectedness as a result of the digitization of journalism. Anderson defined
a news ecosystem as the “entire assemblage of individuals, organizations and technologies within
a particular geographic community or around a particular issue, engaged in journalistic produc-
tion and [. . .] consumption” (Anderson, 2016: 412). Drawing on Latour’s “Actor-Network
Theory”, he argued that news ecology actors consisted of humans, objects, and technologies,
and together these formed assemblages enabled by technological and social environments. These
phenomena, of course, have now given rise to social and participatory forms of journalism, which
are particularly destabilizing journalism cultures. Anderson’s work revealed the ways news orga-
nizations that were able to benefit from new, digitally enabled relationships were more likely to
succeed in new networked news formations. Coleman et al. (2016) also used the news ecology
lens to study a single city’s news environment, arguing that traditional and nontraditional news
entities did not compete so much as carve out their individual informational niches within the
overarching ecology. In addition to studying processes and practices of news ecologies, other
researchers have emphasized that they may be productively examined at the time they are coming
into being. Graeff et al. (2014) recommend examining the emergence of a news ecology around
specific stories and topics that are their connective point. Similarly, Lowrey (2015), focusing
specifically on the concept of organizational ecology, argues that to understand the ways news
entities have responded to uncertainty and change, it is helpful to highlight innovations, which
tend to emerge when an organizational ecology is developing. He argues that over time news
routines become entrenched, leading the ecology toward stasis.
Pop-up news ecologies

Of particular interest here is the connection of the news ecology concept with the growing insta-
bility of news, which has been called “liquid” or “pop up” (Deuze, 2009: 15, 24). These terms
work as metaphors for the rapid ways in which news actors, networks, and technologies may
appear and disappear, oftentimes with no intention of permanency. The pop-up news ecology
376
How pop-up news ecologies come into being
concept was developed through an examination of Syrian citizen, activist, and professional jour-
nalism that sprang into being at the start of the Syrian civil war (Wall and El Zahed, 2015).
Spurred into existence by resistance to an authoritarian regime, actors often working indepen-
dently of each other contributed to the creation of a new, emergent news network that operated
in opposition to the traditional, strictly controlled media system in Syria.
While Lowrey (2015) focuses on stabilization of the ecology as a protective move to reinforce
boundaries and Anderson emphasizes connectedness, here, the pop-up concept highlights insta-
bility and emergence. Indeed, Wahl-Jorgensen argues that the concept of ecology as a lens for
examining news systems works because it is not tied to assessing a single entity and thus allows
for a changing cast of actors. The assessment of the pop-up news ecology in Syria focused on
what Lowrey (2015: 145) highlights as a key to understanding journalism ecologies: its process of
“becoming”. The Syrian case found a set of key factors that enabled a pop-up news ecology to
come into existence, including:
• Existence of a news vacuum. A pop-up ecology forms in part because of an information

void. In the case of Syria, this vacuum existed due to the long-standing control of domestic
news media by the Syrian regime as well as the dangers of reporting on the emerging civil
war, which left foreign journalists often unable to safely enter Syria and cover the conflict.
(This also meant that external news outlets, especially Western ones, were highly dependent
on internally collected reports.)
• An ability to rapidly coalesce. Emergent, often oppositional news ecologies rapidly come
into existence, enabled in part by the availability of prosumer communication devices and
especially by social media platforms taking advantage of networked communication forms.
In Syria, a heightened sense of urgency around sharing news and information existed as the
country descended into a vicious civil war.
• Amplification by external connectors. While the content may be produced within a limited
geographical space or topical area, if the stories being shared are of high news value to
those not directly in the network, other institutions and entities, including professional
news media, pick up these stories and amplify them to much larger, sometimes even global
audiences. In the case of Syria, some of these connectors included diaspora Syrians, but
others were mainstream news media, some of whom had no boots on the ground.
• Adaptation to professional journalism norms. While many of the content producers in a
pop-up ecology may begin as self-taught citizen journalists or content producers from other
spheres (such as NGOs, activist groups, etc.), as the pop-up ecology grows, they appear
to adopt at least some professional practices. They do so in order to make their content
more palatable and trustworthy to outside news distributors. In the case of Syria, some
of these amateurs received Western training and began to evolve their content based on
these practices (see: Yousuf and Taylor, 2017). At the same time, some alternative content
producers were able to create digital news outlets that sought to take on the appearance and
some functions of mainstream news while existing only on YouTube and Facebook.
• Reliance on existing media tools and platforms. While analog efforts may also exist within
a pop-up news ecology, these are primarily digital spaces, heavily dependent on social media
platforms. However, social media are not neutral actors, merely hosting content or offering
platforms for all to use as they please. They may regularly interfere with users of their
networks; for example, they may take down content they deem inappropriate or in violation
of their terms.
It should be noted that the Syrian pop-up ecology with which these characteristics are associ-
ated grew at the start of the conflict; thus, the pop-up that initially emerged and attained global
377
Melissa Wall
attention was one affiliated most strongly with the activists who were working nonviolently
against the government. That said, other ecologies appeared later and were connected with the
armed insurgent networks, some of which eventually became part of the Islamic State, which
functioned within a sometimes overlapping but different news ecology (Al-Ghazzi, 2014; Far-
well, 2014). One of several fault lines between the activists’ pop-up ecology and the initial armed
insurgent one was language, in which the use of English was a key dissemination strategy for the
activists and their allies. It enabled them to speak to liberal Western countries but also meant that
some attempted to package their information in ways that appealed to journalists and policymak-
ers in Western democracies in order to gain their support. The armed insurgents’ oppositional
ecology was mostly reported in Arabic and developed tighter connections with audiences and
intermediaries in Middle Eastern states. Likewise, while the insurgents also adopted outsiders’
expectations, those frequently came from the Gulf States, particularly Kuwait and Saudi Arabia,
where opponents of the Syrian regime generally favored content that showed and sometimes
celebrated acts of violence and adherence to sectarianism (Lynch, 2016; Phillips, 2016). These
more violence-oriented ecologies, particularly that of the Islamic State, eventually supplanted the
activists in reaching global audiences, including within the West, from which they adopted high
production standards but not liberal political values.
A pop-up news ecology emerges

This chapter seeks to apply this pop-up news ecology concept to a different case to refine the
concept and see how it applies outside of Syria. The focus here is on the ways various individuals,
organizations, and other networked actors operated in a loose configuration to distribute news
and information about a series of highly publicized U.S. police shootings of African Americans
that became the focus of widespread protests and attention (see Richardson, this volume, Chapter
29). The shootings were part of the fabric of the U.S. criminal justice system’s disregard for
African Americans, who even now, in the second decade of the twenty-first century, are 2.5 times
more likely to be killed by police than white Americans (Lowrey, 2016a). The topic was chosen
as the focus for this chapter because, like Syria, it centers around high conflict situations involving
life and death confrontations. Of course, the violence in these instances is not occurring in a civil
war, and its victims do not take up arms. It also differs significantly from Syria as well in that
it takes place in a media-saturated and technologically sophisticated information environment,
occurring in a country with a long history of freedom of speech rights and a strong civil society
and rule of law. Nevertheless, I posit that a digital pop-up news ecology emerged around these
acts of police violence and resulting citizen deaths. New and preexisting sources of information
seeking to document and sometimes challenge these events together spurred the emergence of a
new, alternative news ecology focused on police violence against African Americans.
As with the Syria example, many of this ecology’s dominant actors include individuals on
social media such as livestreamers, podcasters, bloggers, Tweeters, and others using YouTube,
Tumblr, Vine, etc. This ecology also encompasses the online sites of the traditional black press and
newer iterations of online-only black media (e.g., This Week in Blackness [TWiB!]), along with
websites specifically dedicated to the issues at hand such as blacklivesmatter.com and fergusonac-
tion.com. Periodically joining these are the social media presences of black celebrities (e.g., Spike
Lee, Wyclef Jean) or other well-known black figures and community and activist groups (e.g.,
Black Youth Project, the Rev. Al Sharpton) (Florini, 2017; Graeff et al., 2014). The presence of
this emergent digital news ecology does not preclude the existence of a larger mainstream news
ecology but is viewed as separate (while sometimes overlapping and interacting). Indeed, there
is a long history of ethnic media establishing separate communication spheres specific to their
communities in the United States (Squires, 2002).
378
While the original explanation of pop-up news ecologies was that they arise only within
authoritarian countries (Wall and el Zahed, 2015), the pop-up news ecology here arose within
the context of a particular space of authoritarianism created when African Americans interact
with the U.S. criminal justice system. As researchers have argued, the “unmasking” of power
within a state apparatus comes during moments when authoritarian actions are taken (Hall
et al., 1978[2013]: 214). The violence against African Americans represents such a space. Indeed,
critics argue that authoritarian practices have long been a part of the criminal justice system in
the United States, and as the national security and counter-terrorism apparatuses have become
embedded within local police operations, it has become more visible to outsiders who are not its
focus (Bieler, 2016). This is seen particularly in the use of military equipment and operational
practices against civilian populations in minority communities. Thus, the law enforcement agen-
cies here exhibit a sort of “subnational authoritarianism”, that can exist within an otherwise
functioning democracy (Behrend, 2011; Gervasoni, 2016). This pop-up ecology is a response to
these tendencies.
Applying the concept

The key characteristics identified earlier as necessary for the emergence of a pop-up news ecology
are applied to a secondary review of research as well as some key journalistic accounts of the
police violence against African Americans. Here, my goal is an effort to refine the concept of a
pop-up news ecology to determine the ways it may have broader applicability.
Filling an information vacuum

The pop-up news ecology in Syria provided alternative views from the discourses created by
the Syrian government through its control of the existent media system. Much of this pop-up
or temporary ecology’s content was produced at the grassroots and took an oppositional stance,
created by citizen journalists and activists documenting what they were seeing and experiencing.
Similarly, the racial justice ecology developed to fill an information vacuum in the United
States. For example, Graeff and colleagues’ (2014) network analysis of the media reaction to the
shooting of unarmed teenager Trayvon Martin by a neighborhood watch member in Florida
clearly outlines the emergence of a new news ecology (although they don’t distinguish it as
separate from the mainstream ecology). These researchers argue that initially this shooting was
a small local news story, generating an uptick of attention mainly from nearby Florida-based
news media. Then, a media publicist hired by the Martin family amplified the story, gaining the
attention of some national mainstream news media, but that attention was both short-lived and
not widespread. From there, a web of other grassroots actors began to highlight the story within
their own communication networks. This included students from Howard University (a histori-
cally black university or HBCU) and the activist group Change.Org along with media aimed at
ethnic minorities such as Global Grind, Color of Change, Black Youth Project, and This Week in
Blackness. Together, these and others brought Martin’s story to a much higher level of attention.
As with Syria, these new actors filling the vacuum also sometimes created their own news
entities or what has been called an improvised “ad hoc news wire” (Richardson, 2016: 2). This
news wire was not inconsequential in terms of audience reach. For example, “This Week in
Blackness Reporting” was on the ground when the results of the Florida trial of the man who
killed Martin were announced, as well as in the response to the shooting of another unarmed
youth, Michael Brown, in Ferguson, Missouri. This digital operation has had 1 million down-
loads of its main podcast. Perhaps most notable has been the use of Black Twitter as a keystone
actor in this pop-up ecology. Indeed, Twitter was a main host for individual citizen journalists
379
Melissa Wall
who became movement leaders. For example, activist Deray McKesson, a former public school
administrator from Minnesota who traveled to Ferguson during the height of the protests, live
tweeted news about the protests, and also shared commentary and news about police violence
and racial injustice more broadly. Researchers have shown that McKesson was the “most ref-
erenced participant” producing content about the Brown shooting and subsequent uprising,
becoming a “trusted source of protest information”, generating more than 1.1 million retweets
and mentions on Twitter during this time period (Freelon et al., 2016: 54).
It is important to note that the vacuum included not just the lack of sustained coverage but
also of narratives challenging the dominant frames for what was reported. Examinations of main-
stream news coverage of these deaths has found a tendency to support the status quo view that
police violence against minorities is justified (Willis and Painter, 2016). Thus, mainstream news
of these high-profile incidents of police violence is often described by researchers as producing a
pro-police point of view (Araiza et al., 2016; Florini, 2017; Willis and Painter, 2016; Hockin and
Brunson, 2016). As sociologist Michael Eric Dyson notes, “We’re in a culture that disbelieves
black truth” (quoted in Wagner, 2016: para 8), and this in particular appears to greatly contribute
to creating a space for this pop-up news ecology. In sum, this new ecology offered a different
lens for viewing the violence against African Americans, highlighting alternative perspectives,
particularly from within that community (Florini, 2017). While mainstream news sometimes
presented the victims of the police violence as threatening and criminally inclined, the pop-up
ecology’s narratives framed the story as one about police brutality against ordinary people who
were victims of systemic racial injustice.
Coming into existence at high speeds

Just as with the Syrian case, this new news ecology was activated quickly, which led to the racial
justice movement’s stories being rapidly and widely shared in ways that traditional vehicles such
as the black press, operating on slower schedules and employing less nimble practices, could
not have carried out. Although Syria’s pop-up ecology also experienced a high-speed birth,
there, the internet connections were less robust and security was such that those who collected
media content, particularly videos, were not always the ones who uploaded their own stories.
Instead, a chain of participants funneled content to safer harbors for global dissemination, thus
delaying their posting. Here, the high velocity in distributing information is evident in the widely
available live coverage of the police actions against citizens, the protests, and other on-the-ground
responses. Content was oftentimes either uploaded immediately or simply live broadcast using
tools such as Ustream and, after its April 2016 introduction, Facebook Live (the tool that carried
live the shooting of Philando Castile, who was pulled over during a traffic stop in Minnesota.
His girlfriend sitting in the passenger seat used Facebook Live to document him being shot to
death by a police officer).
Researchers tracking some of the key cases of police violence find that the stories were brought
to public attention first on Twitter, which is considered much faster than Facebook (Bonilla and
Rosa, 2015; Hitlin and Vogt, 2014; Pew Research Center, 2012). In the case of the shooting of
Michael Brown in Ferguson, the incident garnered thousands of tweets within 24 hours (more
than 10.6 million would be posted within the first eight days after the incident), while it took
another 24 hours for the first wave of mainstream news coverage to begin when cable television
news covered the story in their prime-time news hours (Hitlin and Vogt, 2014). Jackson and
Welles (2016: 397) call the posters of these first waves content “initiators”, who are the first to
cover the event but often fade away or even simply stop posting after a day of serving as a top
initiator of social media coverage.
380
The speeded-up production of content can also enable an oppositional ecology to launch
a new, shared counter-narrative before one is established within a dominant mainstream news
ecology. In this pop-up ecology, this was particularly effectively done using hashtags, which
enable a rapid alternative storyline to widely circulate and be reinforced by social media denizens
(Bonilla and Rosa, 2015). For example, three days after the shooting of Michael Brown, grassroots
social media users launched the hashtag #IfTheyGunnedMeDown in response to the unflatter-
ing photos of Brown being used by mainstream news media that implied he was an unsavory
character. African Americans, particularly youth and young adults, used the hashtag with photos
juxtaposing images of themselves. In one image, they would appear wholesome (e.g., photos of
them graduating or serving in the military), and in the companion one they might be construed
as engaged in socially unacceptable behavior (partying, etc.). In this way, they rapidly challenged
racist images of Brown and other African Americans (Everbach et al., 2017).
Enacting autonomous values

The Syrian pop-up news ecology saw some of its activist contributors adopting professional
journalism norms, taking note of ways to source and present material that would make it more
trustworthy to professional journalists from the Western news media (and likewise, the insurgent
ecologies highlighted content showing visceral violence and whipping up the sectarianism their
funders preferred). The ecology that built up around opposition to police violence did not
follow these patterns. While it was not uncommon for the content of some contributors to
this temporary news ecology to be noticed by the traditional mainstream news outlets, this
incorporation did not appear to be sought after (Jackson and Welles, 2016). Instead of a move
to train and incorporate those creating content for this oppositional ecology, a reverse pattern
took place: mainstream news media adopted some of its practices. For example, a key citizen
journalist, Ferguson alderman Antonio French, relied heavily on Vine videos as he reported
from the resistance in the streets, and soon after, the mainstream news reporters adopted this
platform as one of their means of producing content about the story (Araiza et al., 2016). Indeed,
as noted earlier, rejection of the underlying frames evident in mainstream news coverage was a
key characterization of this oppositional ecology’s content, which established and followed its
own values, including emphasizing emotional content; heavily relying on live coverage and street
reporting; and fostering connections among like-minded voices.
Many of the key actors in this ecology relied on emotional language in their content produc-
tion, which Richardson (2016: 17). Describes as “unapologetically subjective”. For example, Black
Lives Matter activist Johnetta Elzie, one of the key Twitter voices covering Ferguson, wrote with
“unchecked emotion” (Lowrey, 2016b: 39) Likewise, Poepsel and Painter (2016) argue that the par-
ticipants’ “constant, live broadcasts carry implied tones for audiences” (106), often providing sub-
jective frames for their content. The embrace of live, on-the-scene reporting and use of oppositional
narratives can be seen in the approach of many of the Livestreamers, some of the most popular of
whom had more than 1 million views of their footage (Poepsel and Painter, 2016). Many sought
legitimacy by describing themselves as embedded with protesters, providing “first person coverage
of the movement” (p. 100). Like a generation of bloggers from earlier eras, authority derived here
from connections with an audience that supported the resistance to police brutality and violence.
Some Livestreamers not only engaged in practices outside the norms of traditional jour-
nalism, they worked to draw attention to those differences. Rather than seeking to establish
himself as a fellow journalist, a Livestreamer interviewed on CNN is instead shown aggressively
confronting the CNN reporter to, as he puts it, “shine light on police and police behavior”
(101). The Livestreamers’ connections to the activists and oppositional groups is reflected in one’s
381
Melissa Wall
comment that he would not show video of the protesters “doing things that could get them
into trouble” (Poepsel and Painter, 2016: 106). Indeed, as coverage would continue of protests,
etc., the goals of the pop-up ecology’s actors often shifted from merely producing content to
supporting and engaging the community. This could be seen in practices such as creating new
uses for hashtags, such as when activists used #FergusonFriday as “space for reflection on the
movement” that could bring voices together beyond providing immediate responses to action in
the streets (Bonilla and Rosa, 2015: 10). All of these differences are important because studies of
mainstream news ecologies in the US suggest that such ecologies continue to be overwhelmingly
dominated by traditional news providers in the form of professional news outlets (Pew Research
Center, 2015).
Amplified by external connectors

In another notable difference from the Syrian case, the relationship with other sorts of connectors,
in particular mainstream news media, are not the same here. In the Syrian conflict, one of the
important connectors was external mainstream news media, which served as important amplifiers
of the stories being produced. This included both Middle Eastern outlets, particularly satellite
news channels, as well as Western news media. Able to reach larger audiences, they pushed
out content from a range of oppositional actors. Activist actors in particular (as opposed to
armed insurgents) incorporated English versions of reporting into their content to reach
Western audiences. Oftentimes, though, activists and insurgents downplayed content for Western
audiences that reveled in violent actions.
In comparison, while mainstream media prompted additional attention to the issues circulat-
ing through the racial justice news ecology, the high percentage of African Americans on Twitter
meant that they propelled attention themselves regardless of traditional news notice, and activists
did not change their content to attract mainstream coverage. Here, there were at least three cat-
egories of connectors: allies, opponents, and parasites.
• Allies consisted of sympathetic alternative left-liberal news media including the Huffington
Post, The Nation, Vice, AJ Stream, and Gothamist as well as local alternative news outlets
such as the Riverfront Times in St. Louis (Poepsel and Painter, 2016). Such sites boosted the
reach of stories of police violence through their distribution and amplification to their own
networks. While these news producers regularly cover police violence and broader African-
American issues, for most, such stories are but one area of many in their reportage. In addi-
tion, the hacktivist group Anonymous, which periodically provides news and information
about social and political issues from the left or libertarian angle, often by leaking additional
information, served as a connector.
• Opponents consisted of conservative news outlets such as the Drudge Report and the Daily
Caller and right-wing social media denizens that produced counter-narratives to the pop-up
ecology (Graeff et al., 2014). Despite being hostile toward the racial justice ecology’s entire
raison d’être, nevertheless, opponents pushed the story (told through their counter-frames) to
additional audiences and thus likely helped it place higher on the mainstream news agenda.
Ironically, they worked in opposition to what those producing content for the pop-up
ecology believed, further contributing to the need for the pop-up ecology to exist.
• Parasites were the final type of connector, here consisting of two forms: one was made up of
opportunists who used the attention generated by the racial justice ecology to gain their own
audiences (leading to potential advertising revenue). For example, according to Freelon and
colleagues (2016), a key disseminator of news within the ecology was the Bipartisan Report,
a site that generates millions of Facebook visits. It bills itself as “the Internet’s newspaper”
382
with a liberal perspective but is described by observers as hosting “clickbait” content that
is not actually true (Westneat, 2016). The second includes Russian government-owned
English-language outlets such as RT (formerly Russia Today), which has become increas-
ingly well-known for participating in the distribution of alternative news (and “facts”)
about high-profile news stories in the United States and other Western countries. Indeed,
the Russian state has subsequently been identified as playing an outsized role in supporting
oppositional voices throughout the Western world. These gray actors are important because
they represent what has increasingly become a concern in liberal democracies, with the abil-
ity of social media to foster (and make money from) propaganda that thrives on disinforma-
tion and lies. It could be argued that Russian state media themselves have become a sort of
pop-up news ecology that serves to support the aims of an authoritarian state by involving
itself in the internal political conflicts of other countries, particularly through news media
coverage and social media manipulations.
Dependent on social media tools

Pop-up ecologies in Syria and the US are extraordinarily dependent upon social media to fuel
their communication networks, suggesting that these new media companies are foundational
actors without which these ecologies would be severely weakened or perhaps not even exist. In
the Syrian case, the citizen and activist content on YouTube, a key social media platform they
depended on, was sometimes taken down based on the company’s rules (for example a video
being too violent or otherwise offensive) (Wall and el Zahed, 2014). This meant that content
was unstable, and even the producers of this content also were shown to take down their own
accounts, often for unspecified reasons. Takedowns of social media content can happen in a much
freer system as well (although these may be less frequent due to the ability of citizens to flag
the apparent censorship). For example, when the Facebook Live video shot by the girlfriend of
Philando Castile went viral, it was initially taken down by Facebook. A public outcry followed
and the takedown was reversed, with the company claiming it was a technical glitch. Thus,
reliance on social media tools for distribution comes with those risks no matter which country
they are being used in due to the enormous power and reach of the media companies providing
these platforms such as Facebook, Twitter, and YouTube.
Conclusion
The pop-up news ecology that developed around the U.S. racial justice movement suggests some
of the same factors found in the case of the Syrian civil society-fueled ecology were present,
while others were not. Just as with Syria, an information vacuum existed due to a lack of news
coverage by mainstream news outlets along with the presence of a limited range of status quo-
oriented narratives that tended to support the police or draw boundaries around the ways the
stories could be addressed. As in Syria, this ecology came to life rapidly, generally disseminating
its information before the mainstream news ecology did, in part by frequently choosing to post
content live. This follows a history of social media forms leveraging speed to race ahead of
slower moving professional news outlets, which exercise greater levels of professional judgment.
Likewise, the pop-up ecology here was heavily dependent on social media and other corporate
media tools as its platforms for creating and disseminating content. That said, there were some
important differences.
Here, the pop-up system did not seek to mimic or adopt the mainstream news ecology’s
values, instead creating their own news values. These actors were able to leverage their own large
audiences to the extent that they were apparently less concerned about shaping a message to
383
Melissa Wall
fit outsiders’ expectations. That difference meant that the racial justice ecology may have been
less influenced by relationships with external connectors. In further developing the pop-up ecology
concept, the existence of a range of differently motived external connectors was identified, sug-
gesting a sort of news equivalent of biological polymorphism. Overall, the differences found here
from the Syrian case may be attributed in part to the more highly developed American news
environment. It appears likely that a similar ethnic-related pop-up news ecology formed in the
US in connection with the opposition to the Dakota Access Pipeline at Standing Rock. Examin-
ing that pop-up news ecology as well as additional cases from nondemocratic countries would
help clarify this chapter’s arguments.
In conclusion, regardless of where they originate, pop-up news ecologies are marked by their
liquidity. Key actors may quickly disappear to be replaced by others, and the technologies fueling
the ecology may become unstable or even unavailable. While this chapter focused on the forma-
tion of pop-up news ecologies, it does not track them through their decline and demise, which
remain important directions for future research. Pop-up ecologies may lose public attention or
become too fraught with challenges to continue. (This is in part what happened in the case of
Syria, as the political situation changed and the violence intensified.) Other possible outcomes for
pop-up ecologies include becoming so connected to a dominant, mainstream news ecology that
they cease to be distinct, losing much of their value and perhaps even being replaced themselves
by some other oppositional news ecology.
Further reading
Essential readings on news ecologies include Chris Anderson’s (2013) Rebuilding the News: Met-
ropolitan Journalism in the Digital Age and his chapter (2016) on “News Ecosystems” in The Sage
Handbook of Digital Journalism. Karin Wahl-Jorgensen (2016) expertly excavates the foundations
and history of the news ecology concept in her article “The Chicago School and Ecology:
A Reappraisal for the Digital Era.” The idea of a pop-up news ecology was first explicated in
“Syrian Citizen Journalism: A Pop-Up News Ecology in an Authoritarian Space”, which I co-
authored with Sahar el Zahed (2015). Coleman et al. (2016) provide an excellent road map for
assessing a local news ecology in their study of the city of Leeds in The Mediated City: The News
in a Postindustrial Context.
References
Ahva, L. and Steenson, S. (2016) “Deconstructing digital journalism studies.” In B. Franklin and S. Eldridge
(eds.), The Routledge Companion to Digital Journalism Studies. New York, NY: Routledge (pp. 25–34).
Al-Ghazzi, O. (2014) “‘Citizen journalism’ in the Syrian uprising: Problematizing Western narratives in a
local context.” Communication Theory, 24(4), 435–454.
Anderson, C. W. (2016) “News ecosystems.” In T. Witschge, C. W. Anderson, D. Domingo and A. Hermida
(eds.), The Sage Handbook of Digital Journalism. Thousand Oaks, CA: Sage Publications (pp. 410–423).
Araiza, J. A., Sturm, H. A., Istek, P. and Bock, M. A. (2016) “Hands up, don’t shoot, whose side are you
on? Journalists Tweeting the Ferguson protests.” Cultural Studies – Critical Methodologies, 16(3), 305–312.
Behrend, J. (2011) “The unevenness of democracy at the subnational level: Provincial closed games in
Argentina.” Latin American Research Review, 46(1), 150–176.
Bieler, S. (2016) “Police militarization in the USA: The state of the field.” Policing: An International Journal
of Police Strategies & Management, 39(4), 586–600.
Blaagaard, B. (2015) “The aesthetics of posthuman experience: The presence of journalistic, citizen-generated
and drone imagery.” Westminster Papers in Communication and Culture, 10(1), 51–65. doi:10.16997/
wpcc.212.
384
Bonilla, Y. and Rosa, J. (2015) “# Ferguson: Digital protest, hashtag ethnography, and the racial politics of
social media in the United States.” American Ethnologist, 42(1), 4–17.
Carlson, M. (2015) “Introduction.” In M. Carlson and S. C. Lewis (eds.), The Boundaries of Journalism: Profes-
sionalism, Practices and Participation. New York, NY: Routledge (pp. 1–18).
Coleman, S., Thumim, N., Birchall, C., Firmstone, J., Moss, G., Parry, K., Stamper, J. and Blumler, J. G.
(2016) The Mediated City: The News in a Post-industrial Context. London: Zed Books.
Deuze, M. (2009) “Journalism, citizenship and digital culture.” In Z. Papacharissi (ed.), Journalism and Citi-
zenship: New Agendas in Communication. New York, NY: Routledge (pp. 15–28).
Eldridge, S. and Franklin, B. (2016) “Introduction: Defining digital journalism studies.” In B. Franklin and
S. Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. New York, NY: Routledge
(pp. 1–12).
Everbach, T., Clark, M. and Nisbett, G. S. (2017) “# IfTheyGunnedMeDown: An analysis of mainstream
and social media in the Ferguson, Missouri, shooting of Michael Brown.” Electronic News, 12(1), 23–41.
Farwell, J. P. (2014) “The media Strategy of ISIS.” Survival, 56(6), 49–55.
Florini, S. (2017) “This week in blackness, the George Zimmerman acquittal, and the production of a net-
worked collective identity.” New Media & Society, 19(3), 439–454.
Franklin, B. and Eldridge, S. (2016) The Routledge Companion to Digital Journalism Studies. New York, NY:
Routledge.
Freelon, D., McIlwain, C. D. and Clark, M. D. (2016) Beyond the Hashtags: #Ferguson, #Blacklivesmatter, and
the Online Struggle for Offline Justice. Washington, DC: Center for Media & Social Impact, American
University.
Gervasoni, C. (2016) “The dimensions of democratic and hybrid subnational regimes.” In Illiberal Practices:
Territorial Variance Within Large Federal Democracies. Baltimore, MD: Johns Hopkins University Press.
Graeff, E., Stempeck, M. and Zuckerman, E. (2014) “The battle for ‘Trayvon Martin’: Mapping a media
controversy online and off-line.” First Monday, 19(2). Retrieved from http://firstmonday.org/ojs/index.
php/fm/article/view/4947/3821
Hall, S., Critcher, C., Jefferson, T., Clarke, J. and Roberts, B. (1978/2013) Policing the Crisis: Mugging, the
State and Law and Order. New York, NY: Palgrave Macmillan.
Hitlin, P. and Vogt, N. (2014) “Cable, Twitter picked up Ferguson story at a similar clip.” Pew Research.
Retrieved from www.pewresearch.org/fact-tank/2014/08/20/cable-twitter-picked-up-ferguson-story-
at-a-similar-clip/
Hockin, S. M. and Brunson, R. K. (2016) “The revolution might not be televised (but it will be live
streamed): Future directions for research on police – minority relations.” Race and Justice [online first].
doi:10.1177/2153368716676320.
Holton, A. E., Coddington, M. and Gil de Zúñiga, H. (2013) “Whose news? Whose values? Citizen jour-
nalism and journalistic values through the lens of content creators and consumers.” Journalism Practice,
7(6), 720–737.
Jackson, S. J. and Welles, B. F. (2016) “#Ferguson is everywhere: Initiators in emerging counterpublic net-
works.” Information, Communication & Society, 19(3), 397–418.
Jensen, J. L., Mortensen, M. and Ørmen, J. (2016) News Across Media: Production, Distribution and Consump-
tion. New York, NY: Routledge.
Lewis, S. C. and Westlund, O. (2015) “Actors, actants, audiences, and activities in cross-media news work:
A matrix and a research agenda.” Digital Journalism, 3(1), 19–37.
Lowrey, W. (2015) “Ecologies and fields: Changes across time in organizational forms and boundaries.” In
T. Vos and F. Heinderyckx (eds.), Gatekeeping in Transition. New York, NY: Routledge (pp. 140–160).
Lowrey, W. (2016a) “Aren’t more white people than Black people killed by police? Yes, but no.” The
Washington Post. Retrieved from www.washingtonpost.com/news/post-nation/wp/2016/07/11/
arent-more-white-people-than-black-people-killed-by-police-yes-but-no/
Lowrey, W. (2016b) They Can’t Kill Us All: Ferguson, Baltimore and a New Era in America’s Racial Justice Move-
ment. New York, NY: Little Brown and Company.
Lynch, M. (2016) The New Arab Wars: Uprisings and Anarchy in the Middle East. New York, NY: PublicAffairs.
Nardi, B. A. and O’Day, V. (1999) Information Ecologies: Using Technology With Heart. Boston, MA: The MIT Press.
Pew Research Center. (2012) How blogs, Twitter and Mainstream Media Have Handled the Trayvon Martin Case.
Retrieved from www.journalism.org/2012/03/30/special-report-how-blogs-twitter-and-mainstream-
media-have-handled-trayvon-m/
Pew Research Center. (2015) Local News in a Digital Age. Retrieved from www.journalism.org/2015/03/05/
local-news-in-a-digital-age/
385
Melissa Wall
Phillips, C. (2016) The Battle for Syria: International Rivalry in the New Middle East. New Haven, CT: Yale
University Press.
Poepsel, M. A. and Painter, C. (2016) “Alternative media and normative theory: A case of Ferguson, Mis-
souri.” CM: Communication and Media, 11(36), 89–114.
Reese, S. D. (2016) “The new geography of journalism research: Levels and spaces.” Digital Journalism, 4(7),
816–826.
Richardson, A. V. (2016) “Bearing witness while Black. Theorizing African American mobile journalism
after Ferguson.” Digital Journalism, 5(6), 673–698.
Sjøvaag, H. and Karlsson, M. (2016) “Rethinking research methods for digital journalism studies.” In B.
Franklin and S. Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. New York, NY:
Squires, C. R. (2002) “Rethinking the Black public sphere: An alternative vocabulary for multiple public
spheres.” Communication Theory, 12(4), 446–468.
Strate, L. (2016) “Media ecology.” In K. B. Jensen, R. T. Craig, J. D. Pooley and E. W. Rothenbuhler (eds.),
The International Encyclopedia of Communication Theory and Philosophy. Oxford: Wiley Online Library.
doi:10.1002/9781118766804.wbiect137
Usher, N. (2017) “The appropriation/amplification model of citizen journalism: An account of structural
limitations and the political economy of participatory content creation.” Journalism Practice, 11(2/3),
247–265.
Wagner, A. (2016) “To live and die on Facebook.” The Atlantic. Retrieved from www.theatlantic.com/
politics/archive/2016/07/to-live-and-die-on-facebook/490637/
Wahl-Jorgensen, K. (2016) “The Chicago School and ecology: A reappraisal for the digital era.” American
Behavioral Scientist, 60(1), 8–23.
Wall, M. and el Zahed, S. (2014) “Embedding content from Syrian citizen journalists: The rise of the col-
laborative news clip.” Journalism, 16(2), 163–180.
Wall, M. and el Zahed, S. (2015) “Syrian citizen journalism: A pop-up news ecology in an authoritarian
Space.” Digital Journalism, 3(5), 720–736.
Westneat, D. (2016) “Seattle’s own ‘click-bait’ news site serves up red meat for liberals.” Seattle Times. Retrieved
from www.seattletimes.com/seattle-news/seattles-own-click-bait-news-site-serves-up-red-meat-for-
liberals/
Willis, E. and Painter, C. (2016) “Race prominent feature in coverage of Trayvon Martin.” Newspaper
Witschge, T., Anderson, C. W., Domingo, D. and Hermida, A. (2016) “Introduction.” In T. Witschge,
C. W. Anderson, D. Domingo and A. Hermida (eds.), The Sage Handbook of Digital Journalism. Thou-
sand Oaks, CA: Sage Publications (pp. 1–4).
Yousuf, M. and Taylor, M. (2017) “Helping Syrians tell their story to the world: Training Syrian citizen
journalists through connective journalism.” Journalism Practice, 11(2/3), 302–318.
386
29
THE MOVEMENT AND ITS
MOBILE JOURNALISM
A phenomenology of Black Lives Matter
journalist-activists
Marissa Johnson said Sen. Bernie Sanders supporters never saw her coming. Although organizers
of the now-infamous Seattle rally for the presidential hopeful knew that Johnson was a leader in
the local Black Lives Matter chapter, they expected a man to interrupt the rally that day in August
2015, she said. So, she flanked herself with male colleagues – one black and one white – and went
forth. As the black, male activist walked to one corner of the stage, “They put all the security
over where he was”, Johnson said, smiling. With the diversion in place, her white male colleague
separated the metal barricades to the platform. In a flash, she walked up the makeshift stairs to the
dais. The rest became protest history. A young, black woman had just preempted the presidential
stump speech of a sitting U.S. senator. Johnson demanded 4½ minutes of silence in memory of
Brown, to symbolize the 4½ hours his body lay on a Ferguson street. Some yelled profanities
throughout the moment of silence. As it ended, Johnson began a speech on Seattle’s legacy of
police brutality. She usurped Sanders’s platform for nearly 30 minutes. The clashing imagery of a
seemingly frustrated and forlorn Sanders vis-à-vis Johnson’s bellicosity looped on television news
networks. In one interview she conducted with the MSNBC cable news network, she seemed
assured and proud of her confrontation. Nearly one-and-a-half years later, however, she said that
she has mixed feelings as to whether she would do it again. No one could have prepared her,
she said in a February 2017 interview, for the immense personal toll that her activism cost her.
As one of the 15 anti-police brutality protestors who participated in this study, Johnson pro-
vided insight into the lived experience, or phenomenology, of bearing witness as a Black Lives
Matter activist. Phenomenology is a mode of inquiry that focuses on the lived experiences of the
study’s subjects by evaluating one’s: (1) intentionality, (2) intuition, (3) empathy, and (4) intersub-
jectivity (Husserl, 1970). In terms of this study, it meant investigating: (1) the Black Lives Matter
activists’ stated purpose for practicing sustained acts of media witnessing in and through the
media; (2) whether or not they intuited their bearing witness as a form of journalism; (3) how
they viewed their bodies in relation to the world (for example, whether they saw themselves
in the body of a slain victim of police brutality); and (4) how they believed their individual
work as witnesses impacted the broader Black Lives Matter movement. Nearly nine hours of
semi-structured interviews revealed that the participant activists leveraged mobile journalism as a
means to disrupt existing power structures between the police, the press, and African-American
387
protestors. The activists expressed a nearly unanimous desire to reframe the way legacy news
media portray African-American victims of police brutality and the Black Lives Matter move-
ment. The activists also reported that they believe in journalism’s power to redress the grievances
that their communities may have against its local police department. This chapter begins by
exploring how theories of power help us understand what may motivate the Black Lives Matter
activists’ desire to bear witness. Then, I offer an account of the lived experiences of Black Lives
Matter activists who decided to use their smartphones to report news long after a high-profile,
fatal police shooting rocked their hometowns. This chapter closes with a discussion about what
the trend toward mobile journalism as a tool for counter-narrative might mean for protest news
coverage of the future.
The rise of Black Lives Matter

The Black Lives Matter movement has become one of the most prominent, sustained, African-
American social protests of the new millennium. It began on July 13, 2013, when a jury found
George Zimmerman, a 28-year-old white man, not guilty of second-degree murder after he shot
and killed 17-year-old Trayvon Martin, an unarmed black teenager, in Sanford, Florida. Three
grassroots activists launched the #BlackLivesMatter Twitter hashtag to protest what they believed
to be a travesty of justice. Alicia Garza, Patrisse Cullors, and Opal Tometi circulated the hashtag
widely on Twitter in 2013. It began trending in earnest in 2014, when a white police officer shot
an unarmed black teenager, Michael Brown, in Ferguson, Missouri (Garza, 2014). Race riots in
Ferguson began, and #BlackLivesMatter evolved from a Twitter campaign to actual street protests
there and in other cities across the country. In 2015, Black Lives Matter protesters expanded fur-
ther, organizing into nationwide chapters. Chapter members then began to follow candidates for
the United States presidency on their campaign trails, hoping to start a dialogue about the string
of highly publicized killings (ThinkProgress, 2015).
At the time of this study, Black Lives Matter has scaled internationally, with chapters in
Berlin, Dublin, Israel, London, Palestine, Toronto, and multiple townships throughout South
Africa (Khan, 2015). Although the tipping points that led to Black Lives Matter’s establish-
ment were rooted in efforts to end police brutality, the organization now states a broader
purpose on its website. The group promotes its list of causes as such: (1) efforts to end mass
incarceration of black people, (2) pathways to immigration, and (3) systemic support for dually
marginalized subgroups of black people, such as women, girls, the disabled, and members of
the LGBTQIA community.1 Despite the momentum that the Black Lives Matter campaign
has achieved since its inception, to date, little research has been published about the protest
journalism created by black witnesses from within the movement. This study highlights these
key documentarians, which belong to three distinct groups. Some of the interviewees for this
project tweet using the #BlackLivesMatter hashtag and lead nonprofit chapters of the Black
Lives Matter organization in their hometowns. Others tweet using the #BlackLivesMatter
hashtag yet align with sister organizations in the movement. The third group of activists Tweet
using the #BlackLivesMatter hashtag yet work independently as activists that claim no formal
affiliation to an organization, yet exhibit strong ties to leaders of Black Lives Matter chapters
on Twitter. These three groups of anti-police brutality activists are unified in their dedication
to producing mobile videos and Twitter updates on movement-related news. Their evolving
genre of protest journalism pushes discourse about police brutality – and other human rights
violations against African Americans – into the global news cycle every day. Their reports are
thought to disrupt normative ideas about black victimhood, police brutality, and collective
African-American protest.
388
The movement and its mobile journalism
On power and narratives of control

Roni Jackson (2016) asserted:
Historically, most media representations of minorities have presented a one-dimensional

portrayal of individuals of color, painting certain races with a very wide brush,
eliminating individualism and nuance, and feeding a culture of prejudice. For victims
of violence and tragedy, this representation serves to undermine both their person and
their victimhood.
(p. 318)
Does video witnessing serve as a corrective to Jackson’s claim? Does filming a human rights
atrocity have the power to restore one’s humanity? Defining the nature of power itself sheds light
on these questions. It might be useful to think of theories of power on an “ideology-discourse
continuum” (Stoddart, 2007), which ranges from late nineteenth-century thought to modern-
day critical race theories. In this vein, Karl Marx (1848) introduced with Friedrich Engels the
earliest theories of power in his seminal text, The Communist Manifesto. In it, he argued that power
flowed in one direction, from the ruling class (bourgeoisie) to the working class (proletariat).
Describing the bourgeoisie control of the proletariat through wages, Marx wrote:
The capitalist class is constantly giving to the working class drafts, in the form of money,
on a portion of the product produced by the latter and appropriated by the former. The
workers give these drafts back just as constantly to the capitalists.
(Marx and Engels, 1848: 712–713)
Marx believed, therefore, that the working class usurped power from the elites only by altering
the economy – especially in terms of the labor market. Although Marx’s ideas were celebrated
throughout Europe, the working class failed to topple the growing forces of capitalism in his
lifetime. Philosophers from the Frankfurt School thus attempted to generate a more robust theo-
retical model of power. They were pessimistic about the proletariats’ potential to lead revolutions
and posited that resistance must come from the “cultural industry” instead (Horkheimer and
Adorno, 2002; Benjamin (1968 [1936]), which broadcasts and inculcates ideological representa-
tions of the world to the masses. Antonio Gramsci added to this layer of understanding the media
as powerful purveyors of values and ideas with his theories of hegemony. Hegemonic power
functions as “common sense” facts that are “inherited from the past and uncritically absorbed”
(Gramsci, 1971: 333). Gramsci believed that dismantling systems of inequality could be achieved
if intellectuals and subaltern groups worked together to create counter-hegemonic media mes-
sages. Mark C. Stoddart (2007: 202) explained:
Where a society is characterized primarily by the exercise of hegemonic power instead

of coercion, a prolonged cultural war of position is more important, where the hege-
mony of the ruling classes is dissembled and a new hegemony is crystallized. This
occurs as subaltern groups realize their own capacity to become philosophers.
It is here, in realizing that narratives are tools of control, that schools of thought on power
begin to diverge significantly. Michel Foucault’s post-Marxist writings asserted that power does
not flow unidirectionally, from the elites to the working class, in strictly economic terms. Fou-
cault argued that power operates at diffuse nodes in modern society – through science, schools,
389
prisons, and government, too. Foucault theorized that resistance, then, occurred through examin-
ing micro-social tensions between varying groups. He wrote:
If we speak of the power of laws, institutions, and ideologies, if we speak of structures

or mechanisms of power, it is only insofar as we suppose that certain persons exercise
power over others. The term “power” designates relationships between “partners”.
(Foucault, 2000 [1994]: 337)
Foucault believed that actual discourse – not merely ideology – amasses social power. Stoddart
(2007) explained, “The regulation of discourse deals with who is allowed to speak on a given
topic, as well as which forms of knowledge are subjugated in the production of truth” (p. 205).
In Foucault’s own words, he notes, “The delicate mechanisms of power cannot function unless
knowledge, or rather knowledge apparatuses, are formed, organized, and put into circulation, and
those apparatuses are not ideological trimmings or edifices” (2003: 33–34).
The other theories of power important here are intersectional (Collins, 1998; Crenshaw,
1991) in nature and fall under the critical race theory paradigm. Whereas the definition of
power for Marx lies in economic relations; for the Frankfurt School, ideology; for Gramsci,
hegemony; and for Foucault, discourse; the nexus of power for many critical race theorists is
at the crossroads of gender, race, and class. This is not to say that critical race theorists eschew
all earlier thinking of power structures. Frantz Fanon, for example, analyzes colonialism in The
Wretched of the Earth (Fanon and Sartre, 1963) with a Gramscian sensibility, suggesting that intel-
lectuals and marginalized populations work together to create sites of counter-hegemony dur-
ing times of resistance. Stoddart explained also that Stuart Hall explored a Marxist approaches
to race in his essay, “Race, Articulation and Societies Structured in Dominance.” Stoddart
wrote:
[Hall] describes . . . racism [as] rooted in economic structure. Here, racial inequality
is a cultural reflection of the economic base of society. Racializing ideologies work to
justify a system of economic inequality that is beneficial for capital, in terms of provid-
ing a supply of cheap, disposable labor.
(2007: 215)
It may serve us better to think of critical race theory, then, as a series of postulations on power
that highlights inequalities in economic institutions, political arenas, and in media simultaneously,
since narratives of white supremacy are entrenched in all of these major sectors of life. As bell
hooks (2014b: 155) explained:
Certainly in the space of popular media culture black people in the U.S. and black
people globally often look at ourselves through images, through eyes that are unable to
truly recognize us, so that we are not represented as ourselves but seen through the lens
of the oppressor.
Hooks argued further that black gazes at an oppressor were a site of rebellion historically. She
wrote,
Black slaves, and later manumitted servants, could be brutally punished for looking, for
appearing to observe the whites they were serving, as only a subject can observe, or see.
To be fully an object then was to lack the capacity to see or recognize reality.
(2014a: 168)
390
The rise of sousveillance

The description by hooks of power-in-looking proved an intriguing theoretical frame for this
study. I hypothesized that many Black Lives Matter activists doing the work of journalists assumed
this “oppositional gaze” (2014a: 116) to establish a degree of narrative power over how African
Americans are portrayed in the media – especially when human rights violations are in question.
This power does not draw its strength from above, however, in that marginalized groups suddenly
become the all-seeing eye on high, hacking signals from security cameras and other surveillance
devices. Studies suggest otherwise. The black activist gaze draws its strength from below. In a
technologically mediated model of power, which Steve Mann and Joseph Ferenbok have called
sousveillance, smartphones and wearable technology devices afford African Americans and other
potentially marginalized people the power to elevate themselves from object to subject. Mann
and Ferenbok (2013: 23–24) channeled Foucauldian theories of power when describing how
sousveillance functions:
Foucault’s panopticon is a power metaphor for the distribution of institutional power

that works through the fear of being watched. . . . We are entering an age where people
can and will not only look back, but in doing so potentially drive social and political
change.
Hans Toch (2012) suggested that this change will come about as a result of great displays of
political theater and even likened sousveillant cop-watchers to Greek choruses. He wrote:
The involvement of spectators in police-citizen confrontations invites comparison

with the role played by the chorus in classic Greek tragedies. The chorus has been called
the moral barometer of the play in classical Greek theater because chorus members
constantly offered opinions on wickedness, punishment, and righteousness.
Ben Brucato (2015) indeed offered evidence of smartphone-toting protestors behaving like
Greek choruses during the Occupy Wall Street movement in 2011, stating, “Their use of video
streaming apps to live-broadcast such events – while chanting ‘The whole world is watching!’ –
showed how protesters framed watching as intercession” (p. 1). Additionally, a team of researchers
reported in 2013:
A variety of practices were uncovered that link YouTube and Twitter together, includ-
ing sharing cell phone footage as eyewitness accounts of protest (and police) activity,
digging up news footage or movie clips posted months and sometimes years before the
movement began; and the sharing of music videos and other entertainment content in
the interest of promoting solidarity or sociability among publics created through shared
hashtags.
(Thorson et al., 2013)
Likewise, scholars have marveled at how citizens in Tunisia and Egypt practiced sousveillance in
2011’s so-called Arab Spring revolts to circumvent traditional media outlets, which were run by
oppressive political regimes, to publish video directly to Twitter (Howard et al., 2011; Khondker,
2011; Lotan et al., 2011). Those who put themselves in harm’s way to document political unrest
in their countries were heralded as “mediated martyrs” (Halverson et al., 2013) who dared to
“overthrow the protest paradigm” (Harlow and Johnson, 2011). Until now, few studies have
explored whether Black Lives Matter activists intuited that they are assuming a sousveillant,
391
mobile-mediated vantage point to anti-police brutality uprisings that have erupted across the
United States in recent years.
Method
This study relied on semi-structured interviews as its qualitative data-gathering method. I
integrated two established philosophies while developing and refining my interview map:
phenomenology and Jerome Bruner’s functional approach to narrative analysis. Phenomenology
focuses on the lived experiences of the study’s subjects by evaluating a subject’s: (1) intentionality,
(2) intuition, (3) empathy, and (4) intersubjectivity (Husserl, 1970). While the phenomenological
school of thought is more than a century old, many contemporary journalism studies employ
this philosophy to frame qualitative interview questions. Edson Tandoc and Bruno Takahashi
(2016) studied the lived experiences of journalists-cum-ad hoc disaster relief workers who covered
a typhoon that hit the Philippines in November 2013; Tim Markham (2011) investigated the
normative newsroom practices of war correspondents who work in combat zones; and Michael
Arnold (2003) described how the social implications of emergent mobile devices must be
considered in terms of one’s agency to use that device in ways that may not have been conceived
by its manufacturer. Phenomenology allowed me to identify common threads in how Black
Lives Matter activists see themselves and situate themselves as protest journalists who have mobile-
mediated agency. This “day-in-the-life” perspective allowed me to cluster themes that emerged
from their personal narratives into categories.
Interviews were coded by using Jerome Bruner’s (1991) functional approach to narrative analy-
sis. Bruner proposed that narratives are the chief way that people construct their realities and
meanings of events. For instance, although the initial claim that Michael Brown had his hands up
and screamed “Don’t shoot!” just before a white police officer killed him proved to be untrue
(Lee, 2015), that narrative arguably compelled distant black witnesses to protest in Ferguson in
2014. Online photo memes of African-American NFL football players (NBC Staff, 2014), college
students (Clutch Staff, 2014), congresspersons (Larson, 2014), and even media personalities (Marsh,
2015) – all with their hands up in the air – went viral. An eponymous organization, Hands Up
United, even sprang up.2 In this vein, people made meaning of current events through the dominant
narratives that circulated at that time. Some narratives continued to stick, even if they were not true,
because people needed to shape chaotic events into a coherent story that made it easier to process.
Selection of participants
CNN produced in August 2015 a list of 13 “Disruptors” who rose to national prominence in the
year after the Ferguson uprisings.3 This list was the foundation for developing a larger snowball
sample of potential study participants. Snowball sampling is a non-probability technique that
leverages a priori knowledge of a small group of potential study subjects to cultivate a larger pool
of study participants that may have remained hidden, if not for a referral from the original study
subjects. Snowball sampling can be leveraged to gain access to (and trust from) anti-establishment
actors, such as activists, who may be suspicious of legacy media or academic researchers initially
(Cohen and Arieli, 2011; Jun, 2013).
My snowball sample began with the 13 CNN Disruptors. I contacted all of the activists
via Twitter initially, either through direct message (if they authorized this feature) or on their
public walls. Two of the 13 original interview prospects granted me interviews. Nine of the 13
declined my requests for interviews yet referred me to an ally in the movement who was still
“doing press”. The final two Disruptors never responded to me at all, despite three Twitter direct
messages and three emails. I only contacted potential subjects three times on either medium,
392
unless they told me to submit my interview request via Facebook so that they could verify my
identity. Initially, I had hoped that I would make contact with all 13 Disruptors who would each
recommend three more people, for a total of 39 potential interviewees. The influencers in the
movement are tightly knit, however, and similar names began to overlap by the fifth interview. I
conducted semi-structured interviews with 20 participants between October 2016 and February
2017. The total corpus comprises 8.42 hours of video footage. The average interview duration
was 68 minutes. I conducted three interviews in person. The remaining 12 interviews took place
virtually – via Skype, FaceTime, or Google Hangout – according to the participant’s preference.
Results
The top two narrative themes that emerged from 652 coded interview segments were:
(1) revision, with 152 coded segments; and (2) responsibility, with 121 coded segments. I used
the “revision” code if the subject spoke explicitly about his or her efforts to correct or reframe an
existing news media narrative. I used the “responsibility” code if the subject talked about feeling
a sense of duty to bear witness.
Revision
The participants in this study said that they report news to revise existing news narratives. This
was the most popular code in the corpus. The activists aim to challenge sensationalism and/or
factual errors in legacy reports and to oppose black news myths.
Revising news sensationalism and/or errors. Many interviewees said that professional journalists
did not do a good job covering the protests in which they took part. Brittany Packnett protested
on the front lines in Ferguson after Michael Brown’s death. She was activated to bear witness on
Twitter after seeing news reports that conflicted with what she experienced. She said:
People would be watching a live feed on CNN, and CNN was sitting there saying
people are breaking into the McDonald’s – there’s more looting happening. Well, we
would go on the [live] streams, and what we’d be tweeting is that people are being tear-
gassed and they’re breaking into the McDonald’s because they had milk in the McDon-
ald’s and milk is what you have to use on tear-gas. Not water. Right? And so that is the
instantaneous correction that you’re allowed to have. . . . We challenged the mainstream
media who were outsiders to our community, to tell the truth.
Packnett said that she continued to use Twitter to provide updates on the movement after
Ferguson ceased to be headline news because “media needs to always be held accountable”, she
said, adding, “the same kind of relationship that we should have with the free press, it’s the same
kind of relationship we should have with democracy. We should engage with it and reserve our
right to criticize”.
For Ieshia Evans, bearing witness on Twitter was a more reluctant endeavor. When she par-
ticipated in her first demonstration in July 2016 – to protest Alton Sterling’s killing in Baton
Rouge – she said she was surprised to see it portrayed in the news as a riot. She said: “People
were boisterous, they were rowdy as far as being very vocal, but there was no violence”. Evans
said that the police on the scene became increasingly physical, however, which led to the iconic
photograph of her facing off against them in a sundress. She said:
What kind of enraged me at this point was that you’re giving us permission to protest
in the grass, but then you have police officers beating on their shields and bashing up
393
the protesters while they’re already in the grass. I don’t even know what came over me,
but I just decided to stand in the street, like what’s your goal here?
Evans said that after the photograph of her confrontation with the Baton Rouge police went
viral, she began to notice factual errors in official news reports, such as her age or occupation. She
said she stopped watching TV news. General distrust – even cynicism – of news media was a run-
ning theme amongst the activists. When pressed for specific examples of journalistic errors that
made them feel this way, many had anecdotes to share. Attorney Chris Stewart, for example, was
nonplussed by the news reports that stated Black Lives Matter organized and led a rally to raise
awareness about the police killing a man named Deandre Phillips in February 2017. Stewart said:
Seventy percent of that crowd was white, but when I watched the news that night,
they called it a Black Lives Matter rally, and they only showed the black people, which
I thought was just hilarious, because everybody out there was just shocked how many
white people were out there supporting it, but you didn’t see that.
Stewart tweeted a crowd shot that day to counter the legacy media reports he began to see
elsewhere, he said. “[When] social media covers stuff, it’s kind of like a snowball effect”, he said,
adding,
Once it starts rolling and it starts picking up and picking up and picking up, it’s just
really effective to let people know what’s going on. Other than that, you have to rely on
TV news, and that’s not the most effective way, because you have no control [emphasis
mine] over that.
Devin Allen is the famed amateur photographer whose pictures of the Baltimore uprisings in
2015 made it onto the cover of TIME magazine. His coverage of the city’s response to Freddie
Gray’s death while in police custody inspired a book in 2017. Allen spoke in terms of power in
his interview, also, especially when asked if he ever would work as a photographer for a legacy
media publication. He said:
Once you work for those places, you don’t control [emphasis mine] the narrative of your
photos. The thing is, with me and my work, I told the story first, so it was like when
they want to use my pictures, I’m already telling them what’s going on.
Brittany Ferrell was a frontline Ferguson protestor, more than 800 miles away from Allen’s
native Baltimore. Still, she speaks of her use of Twitter to counter-frame in the same tones. She
explained,
[Twitter] allows me a place to tell the truth without any bias or anybody policing the
things that I choose to say, the things that I choose to tell the world. . . . It’s like we are
so much better connected in this struggle via social media because we know where to
turn to when we need the truth.
Revising black news myths. Another area of revision that some witnesses spoke about was the
desires to shatter the tandem news myths of black criminality and black marginality. One witness
uses his knowledge of statistics to spar with professional journalists and police. Another witness
tweets what he finds in his own investigations of police brutality with the public, to show that
394
African Americans are telling the truth. Yet another witness said he photographs his community
to provide the black human interest stories that news media tend to overlook.
Samuel Sinyangwe is a data scientist. When Michael Brown was killed, he found frontline
protesters on Twitter and asked how he could help in Ferguson. Within days, he took a leave of
absence from his job at PolicyLink, a Bay Area think-tank. Once he arrived in Missouri, he said
he kept hearing the same two narratives: either that shootings like these were one-off events or
that all shootings of unarmed black men began because the slain men had resisted arrest. Sinyan-
gwe said he discovered that the federal government does not collect data on fatal police shootings,
but websites such as Fatal Encounters or KilledbyPolice.net did. He decided to merge the data from
both sites. He explained:
About 40% of the records were not identified by race, and so I went through social
media profiles – like Facebook, Twitter, Instagram – went through obituaries, criminal
records databases, and between those could actually identify more than 90% of the
people in the database. Then for armed or unarmed – nobody was keeping track of
that – I had to go through all of the [police] reports to identify that column.
Sinyangwe said that being armed with this data made him realize that he could tell different
stories about police brutality. He explained a triumph in Colorado, where a police department
challenged his data, only to find out their numbers were incorrect. A local reporter then wrote
a scathing investigative piece about the department’s underreporting of fatal police encounters.
Sinyangwe said he targeted his analyses on his home state of Florida also, when he realized that
Orlando was “off the charts in terms of every level of police violence, whether it was use of force
or stops and searches, arrests or killings”, he said. He explained that once he realized this trend, he
convened a meeting with the Orlando Police Department’s leadership. The officers claimed that
the high rates of excessive force were due to equally high rates of tourism. Sinyangwe crunched
more numbers. He told Orlando PD that their rates were higher than New Orleans and Las
Vegas, which have equal – if not more – annual visitors than Orlando. Additionally, Sinyangwe
said he audited the Orlando Police Department’s use of force policy to show that the agency did
not have a rule that restricted officers from using lethal force as a final resort. The police depart-
ment in Tampa, about 90 miles away from Orlando, did. Sinyangwe persuaded Orlando’s police
department to follow Tampa’s lead by providing valuable data on why people were stopped by
police in the first place. Sinyangwe said:
Those interactions were starting off with people who were suspected of, quote, “suspi-
cious activity” or drug possession – like small, minor things that then get escalated into
deadly force. That helped debunk this narrative [emphasis mine] that police were kill-
ing people because they were trying to apprehend violent criminals.
While Sinyangwe battles the myth of black criminality with his brand of data journalism,
Chris Stewart combats such stereotypes by conducting his own investigative reporting as a civil
rights lawyer. He said he was sitting in church one Sunday in April 2015 when he received an
urgent message through Facebook’s text messaging feature on his smartphone. It was Walter
Scott’s niece. Stewart said:
She messaged me and said it was an emergency, and said that her uncle had just gotten
killed the night before, and the family really wanted to talk to an attorney. She had seen
all of the work that I had done on other civil rights cases on Facebook, and she had gone
to our website and all that stuff.
395
He promised to contact her after the worship service concluded. When Stewart eventually
connected with Scott’s brothers and mother by phone, he said he felt the police department’s
official report sounded suspicious. He said:
The video wasn’t out. You know, we didn’t even know there was a video. They asked
me to look into it online through the articles, and tell me what I thought. I looked at
the articles that were out in the media, but they were all saying that Walter Scott had
tried to kill the officer . . . [and] you have to kind of go with your gut, and it just didn’t
sound right – a man that age fighting a cop – none of it made sense.
The family told Stewart that they would hire him as the attorney if he could drive up to
South Carolina by morning. When he arrived the next day, he learned of the infamous video.
Stewart said he began to share limited updates to the case on Twitter, with the family’s permis-
sion, although he did share that he prefers the slower pace of Facebook. He said it allows him to
answer questions from supporters more thoughtfully.
The final news myth that one of the interviewees works to overcome is that of black margin-
ality. Devin Allen said that he tries to show the beauty that he sees in black Baltimore, not just its
turmoil. Allen mentioned that his earlier work before the Freddie Gray uprisings captured black
women with “natural hair, no makeup, not models, just my friends”, he said. He endeavored
also to photograph the positive things police officers did in his community to provide balance
to black-versus-blue tropes. Like Sinyangwe, Allen said he feels his reportage creates entry points
for dialogue between African Americans and police. He explained, “I did an art show for my
youth, and he [the commissioner] came. I was like, ‘I want him on the panel. He needs to be on
the panel. He needs to see my kids’ work. Period’”.
“This is the area that police are constantly . . .” his voice trailed off as he held back tears.
“People are being . . .” his voice cracked. Allen paused to collect himself. He sighed deeply
and said, “This is where Freddie Gray is from, and . . . if you don’t smother and kill our kids –
this is what they can do.”
Responsibility
Despite the persistent side-stepping of the word “journalism”, the activists still expressed the
occupational ethos that many professional journalists likely would say that they have. For example,
the interviewees all spoke of being guided by a sense of responsibility to let people know the
facts. Moreover, the interviewees frequently said that their Black Lives Matter reportage also
stemmed from a duty to: (1) use their education or professional skills for social justice and (2)
open doors for future activists and storytellers.
Responsibility to use one’s craft for social justice. Gordon Parks (1966) once wrote “I chose my
camera as a weapon against all the things I dislike about America – poverty, racism, discrimina-
tion”. Devin Allen spoke of his photography in similar terms, 50 years later. He said:
I believe if you’re a writer, the pen is your weapon. That’s like if you’re a rapper or a
singer: Your voice is your weapon. That’s why I think it’s so important that as black art-
ists, definitely in hip hop, we need to focus on these issues, because your voice reaches
the masses. My pictures reach the masses.
Allen noted that he is self-taught. The road to bearing witness to the world has been filled with
on-the-job training, he said. He feels obligated, therefore, to share what he has learned with
396
other would-be black witnesses. “I didn’t know what the hell a grant was growing up”, he said,
laughing. “This stuff, they didn’t teach us in school. I’m working on two grants right now”.
Eve Ewing said a sense of duty drives her to use mobile and social media in a similar way.
She holds a Ph.D. in education from Harvard University. She studies systemic racial inequalities
in public education, which often leads her to the frontlines of protests in her native Chicago.
She said people often think of movements as needing only a single, charismatic male leader. She
shared that when she helped organize a demonstration in front of the Chicago Police Depart-
ment’s Homan Square facility in North Lawndale in July 2016, however, she realized that the
movement needed much more. The Homan Square facility is a so-called “black site” of illegal
interrogation, she said, where Chicago police allegedly take African Americans in without legal
representation, subject them to torture, and coerce them into making false confessions. Activ-
ists from Black Lives Matter, Black Youth Project 100, and Let Us Breathe worked as a coalition
to produce a sit-in that they dubbed the “Freedom Square Occupation”. For one month, they
chained themselves together in front of the facility, demanding answers about missing black sus-
pects. Ewing said she never stopped tweeting during her involvement. She recalled:
When [we] first set up, we spent a lot of time communicating and people were like,
“Eve can you make this flyer in an hour?” “Can you do this or this?” . . . After that day
I put out these tweets and was like the revolution needs graphic designers. The revolu-
tion needs people that can show up to meetings.4 The revolution needs people that can
paint, people that can sew, people that can set up security systems on people’s phones so
that their phones don’t get hacked, people that know how to pitch a tent, people that
own sound systems and know how to do audio/visual stuff.5
Since Ewing posted her series of tweets on how people can use their talents to fuel the movement,
the thread has been retweeted several thousand times. It has become a mantra of sorts for
organizers, who often ask her permission to recite it at meetings, she said.
Ewing’s Harvard classmate, Clint Smith, had personal responsibility at the forefront of his
mind too as he began his graduate-level studies at one of the world’s most elite institutions in
2014. Ferguson erupted during the first week of his doctoral program. It haunted him, he said:
I started graduate school the same week Mike Brown was killed, and so it was impossible
for me to disentangle what I saw happening in the world and what I was reading about
in my books. I started to use Twitter as a way to share snippets of what I was learning.
Responsibility to open doors for future activists and storytellers. The activists spoke at length about
what the brave new world would look like for their descendants. Devin Allen, for example, said
he is bearing witness now because, “My daughter loves photography. She can walk through those
doors now [because] I walked through these doors”. In a broader sense, Allen said he is bearing
witness for all black children in his Baltimore neighborhood too. “I’m teaching them how to
love their own situation through photography. . . . I give kids cameras. They go out. I might not
never see them again, but they have a piece of me with them”.
Brittany Ferrell is a parent, too. She said that she tweets about the movement despite facing
protest-related felony charges in St. Louis because:
I have a soon-to-be 9-year-old daughter . . . I think a lot of times I look at the great
sacrifice that it’s going to take for black people to get free, and for her and so many kids
like her, [and it is] so worth it.
397
Conclusion
In this essay, I offered an account of a day-in-the-life of Black Lives Matter activists who choose
to do the work of professional journalists. I have illustrated how bearing witness while black is
more than the act of an African-American person picking up a cell phone to record a human
rights violation. It is an act of protest, too, since the black gaze has been outlawed historically.
In their own words, the leading activists of the movement explained how they are motivated
to report news on a long-term basis because they feel a deep responsibility to revise existing
news narratives about black victims of police brutality and black protesters. They also explained
how they believe journalism has the power to highlight injustices and evoke change. Still, the
participants in this study are loath to be called “journalists”.
“Blackness is often times very, very misunderstood”, Brittany Ferrell shook her head as her
interview drew to a close. She added, “When you’re black and you’re angry or you’re black and
you’re fighting for something, people don’t really receive that very well. . . . People don’t believe
that we’re worth what we’re fighting for”. Brittany Packnett challenges this line of thinking. She
told an audience at Citizen University’s annual conference in March 2017, “I made the deci-
sion to love myself so radically that I am worth more than whatever I could lose”.6 Both Ferrell
and Packnett are writing books about their journeys to bear witness during Ferguson. It will be
interesting to follow the professional trajectories of black witnesses as media makers, to observe
whether they join the mainstream to work alongside professional journalists or remain indepen-
dent storytellers that work outside of traditional newsrooms.
As Devin Allen wrapped his interview, he waxed lyrical about notions of power. He said that
he is aware that professional journalists feel he “jumped the line” to land the cover of a major
magazine, but he is not sorry, he said. He explained that his work has empowered him to see
black people in Baltimore – and marginalized people struggling elsewhere – with a humanity
that he never found when watching the news. “I used to be so hard and cold, but through my
photography and the way I see the world now, I’m a crybaby,” he said with a smile. Allen said
by photographing everyday black life in Baltimore, he finds a capacity to gate keep that which
is newsworthy and worth fighting for in his community. He has a new book of photography
and essays that was published in June 2017. “That’s the thing with this new movement in our
generation. You can’t smother us, because if we take one leader, another one is going to just jump
right back up”, he explained.
The activists who participated in this study embodied this spirit of dogged determination.
With their smartphones in hand, they were armed with a new kind of weapon. It may not be
the Winchester rifle that Ida B. Wells (1892) urged blacks to carry for self-defense in the early
twentieth century, but to some people, like Diamond Reynolds, who watched her beloved Phi-
lando Castile die in front of her, it offered protection during crisis. Whether the growing video
evidence of police brutality against black people will amount to increased police prosecutions or
widespread policing reforms remains to be seen. The one thing that seems certain, however, is
that African-American activists will continue to leverage the technologies of their day to report
news of human rights injustices that are leveled against its most vulnerable communities.
Further reading
This chapter is an excerpt from my doctoral dissertation, Bearing Witness While Black: African
Americans, Smartphones and the New Protest #journalism (2017). The full text offers nine more
themes gleaned from the participant interviews, which include (but are not limited to) journal-
ism as a tool for collective grief and the surveillance that black witnesses face after they distribute
their news. This chapter benefitted tremendously from Mann and Ferenbok’s (2013) concept
398
of sousveillance, or gazing from below. The ethnocentric layer of inquiry – that is, examining
what it means to “look” while black – emerged after reading bell hooks’s Black Looks: Race and
Representation (2014a).
Notes
1 See:http://blacklivesmatter.com/guiding-principles/
2 See: www.handsupunited.org
3 See:www.cnn.com/interactive/2015/08/us/disruptors/
4 https://twitter.com/eveewing/status/758381013911343104
5 https://twitter.com/eveewing/status/758382111984349185
6 See:https://twitter.com/CitizenUniv/status/845790199208525825
References
Arnold, M. (2003) “On the phenomenology of technology: The ‘Janus-faces’ of mobile phones.” Information
and Organization, 13(4), 231–256.
Benjamin, W. (1968 [1936]) “The work of art in the age of mechanical reproduction.” In H. Arendt (ed.),
Walter Benjamin: Illuminations. New York, NY: Schocken Books (pp. 217–251).
Brucato, B. (2015) “Policing made visible: Mobile technologies and the importance of point of view.” Sur-
veillance & Society, 13(3/4), 455.
Bruner, J. (1991) “The narrative construction of reality.” Critical Inquiry, 18(1), 1–21.
Clutch Magazine Staff. (2014, August 14) “Photo: Howard University students stand in solidarity with
#Ferguson.” Clutch Magazine. Retrieved from www.clutchmagonline.com/2014/08/photo-howard-
university-students-stand-solidarity-ferguson/
Cohen, N. and Arieli, T. (2011) “Field research in conflict environments: Methodological challenges and
snowball sampling.” Journal of Peace Research, 48(4), 423–435.
Collins, P. H. (1998) “It’s all in the family: Intersections of gender, race, and nation.” Hypatia, 13(3), 62–82.
Crenshaw, K. (1991) “Mapping the margins: Intersectionality, identity politics, and violence against women
of color.” Stanford Law Review, 1241–1299.
Fanon, F. and Sartre, J. P. (1963) The Wretched of the Earth (Vol. 36). New York, NY: Grove Press.
Foucault, M. (2000 [1994]) “The subject and power.” In J. Faubion (ed.), Power: Essential Works of Foucault
1954–1984 (Vol. 3). New York, NY: The New Press (pp. 326–348).
Foucault, M. (2003) Society Must Be Defended: Lectures at the College de France, 1975–1976. Edited by M.
Bertani and A. Fontana, translated by D. Macey. New York, NY: Picador.
Garza, A. (2014) “A herstory of the# BlackLivesMatter movement.” In J. Hobson (ed.), Are all the Women
Still White?: Rethinking Race, Expanding Feminisms. Albany, NY: State University of New York Press (pp.
23–28).
Gramsci, A. (1971) Selections From the Prison Notebooks of Antonio Gramsci. Translated by Q. Hoare and G. N.
Smith. New York, NY: International Publishers.
Halverson, J. R., Ruston, S. W. and Trethewey, A. (2013) “Mediated martyrs of the Arab Spring: New media,
civil religion, and narrative in Tunisia and Egypt.” Journal of Communication, 63(2), 312–332.
Harlow, S. and Johnson, T. J. (2011) “Overthrowing the protest paradigm? How the I, global voices and
Twitter covered the Egyptian revolution.” International Journal of Communication, 5, 16.
hooks, b. (2014a) Black Looks: Race and Representation. Boston, MA: South End Press.
hooks, b. (2014b) Yearning: Race, Gender, and Cultural Politics. Boston, MA: South End Press.
Horkheimer, M. and Adorno, T. (2002) Dialectic of Enlightenment: Philosophical Fragments. Translated by E.
Jephcott. Stanford, CA: Stanford University Press.
Howard, P. N., Duffy, A., Freelon, D., Hussain, M. M., Mari, W. and Maziad, M. (2011) Opening Closed
Regimes: What Was the Role of Social Media During the Arab Spring? Retrieved from http://dx.doi.
org/10.2139/ssrn.2595096
Husserl, E. (1970) The Crisis of European Sciences and Transcendental Phenomenology: An Introduction to Phenom-
enological Philosophy. Evanston, IL: Northwestern University Press.
Jackson, R. (2016) “If they gunned me down and criming while white: An examination of Twitter cam-
paigns through the lens of citizens’ media.” Cultural Studies? Critical Methodologies, 16(3), 313–319.
Khan, J. (2015) “Black Lives Matter has become a global movement.” The Root. Retrieved from http://bit.
ly/2d4Qirr
399
Khondker, H. H. (2011) “Role of the new media in the Arab Spring.” Globalizations, 8(5), 675–679.
Larson, L. (2014, December 2) “Four members of Congress do ‘hands up’ gesture on House floor.” NY Daily
News. Retrieved from http://nydn.us/11OBRBt
Lee, M. Y. H. (2015, March 19) “‘Hands up, don’t shoot’ did not happen in Ferguson.” The Washington Post.
Retrieved from http://wapo.st/1P6ubQS
Lotan, G., Graeff, E., Ananny, M., Gaffney, D. and Pearce, I. (2011) “The Arab Spring – the revolutions were
tweeted: Information flows during the 2011 Tunisian and Egyptian revolutions.” International Journal of
Communication, 5, 31.
Mann, S. and Ferenbok, J. (2013) “New media and the power politics of sousveillance in a surveillance-
dominated world.” Surveillance & Society, 11(1/2), 18.
Markham, T. (2011) “The political phenomenology of war reporting.” Journalism, 12(5), 567–585.
Marsh, K. (2015, June 1) “‘Hands up, don’t shoot’ never happened, but networks keep using it.” NewsBusters.
Retrieved from http://bit.ly/2dA4pKB
Marx, K. and Engels, F. (2002 [1848]). The Communist Manifesto. Retrieved from http://socialistregister.
com/index.php/srv/article/viewFile/5709/2605
NBC Staff. (2014, November 30) “Rams players enter field in ‘Hands up, don’t shoot’ pose.” NBCNews.com.
Retrieved from http://nbcnews.to/1vyhzbe
Parks, G. (1977) A choice of weapons. New York: Harper & Row.
Richardson, A. V. (2017) Bearing Witness While Black: African Americans, Smartphones and the New Protest
#Journalism. Ph.D. dissertation.
Stoddart, M. C. (2007) “Ideology, hegemony, discourse: A critical review of theories of knowledge and
power.” Social Thought & Research, 191–225.
Tandoc, E. C. and Takahashi, B. (2016) “Journalists are humans, too: A phenomenology of covering the
strongest storm on earth.” Journalism, doi:10.1177/1464884916657518.
ThinkProgress. (2015) “Why disrupting political events is an essential tactic for Black Lives Matter.” Think-
Progress.org. Retrieved from http://bit.ly/2d5nGyV
Thorson, K., Driscoll, K., Ekdale, B., Edgerly, S., Thompson, L. G., Schrock, A. and Wells, C. (2013) “You-
Tube, Twitter and the Occupy movement: Connecting content and circulation practices.” Information,
Communication & Society, 16(3), 421–451.
Toch, H. (2012) Cop Watch: Spectators, Social Media, and Police Reform. Washington, DC: American Psycho-
logical Association.
Wells, I. B. (1892) Southern Horrors: Lynch Law in All Its Phases. Auckland, NZ: The Floating Press.
400
30
NATURE AS KNOWLEDGE
The politics of science, open data, and
environmental media platforms
Inka Salovaara
The year 2016 was the hottest on record, according to the National Aeronautics and Space
Administration (NASA). In September 2016, when the U.S. presidential election race was in full
swing, NASA’s environmental data showed that the human species was living in the warmest era
in modern temperature monitoring. This was followed by months of environmental anomalies.
Later in 2017, President-elect Donald Trump ordered the elimination of all federally funded
climate change research, as a part of his aim to eradicate “politicized science”. Shortly afterwards,
NASA’s Earth Science Division faced funding cuts as the political priority moved from climate
change to the exploration of deep space. Under the new U.S. political reality, climate change was
a ‘hoax’, and the president mandated that all data from scientists should go through administrative
review before it was released to the public.
With a rumor spreading that the Trump administration had started to delete climate files on
NASA’s platform, the global hacker community activated, working to save and restore the data.
In April 2017, the Guardian reported that the U.S. Environmental Protection Agency’s (EPA)
climate change website was “being updated” and “undergoing changes” to better reflect “the
agency’s new direction” under Donald Trump. This act of data censorship evoked widespread
political dismay and unrest in the US and beyond.
It is not surprising that Raymond Williams has defined ‘nature’ as perhaps the most complex
word in the language (Williams, 1980: 221). This is especially apparent when a particular political
or media allusion is made to nature that reflects the speaker’s general ideological purpose. Hence,
a reference to nature and environment often reflects a historical moment of human thought,
political economy, or the current ideological climate, but very seldom nature itself. In public
discussions, nature and environment are as much cultural concepts as they are a material reality.
Nature and environment are always socio-ideological categories, as well as instruments of
social and political power. Therefore, when the discourse on nature and environment focuses on
neoliberalization, privatization, marketization, and financialization, we can see this is not a novelty
in political and media discourses. However, the recent transformation of how these discourses are
presented to the public, with increased computerization and datafication in environmental media,
has radically changed our perception of ‘nature’ and the ‘environment’.
Pervasive computerization has improved technical and cognitive tools for displaying, manipu-
lating, calculating, and reading environmental data. Moreover, computerization of environmen-
tal geo-data has enhanced our ability to produce detailed, interactive, and three-dimensional
401
Inka Salovaara
visualizations – both for use in the sciences and in media. Computerized visualization has also
increasingly contributed to ‘mentalistic cognitivism’ that understands visualizations, such as envi-
ronmental mapping and data visualizations, as cognitive aids for the public, as well as for scientists.
This databased cognitivism has given transformative, embodied, and performativity roles to visu-
alization in knowledge production and mediatization of environmental knowledge. That is to say,
visualization in the sense of ‘making visible’ is always constructing knowledge and facts (Carusi et
al., 2015: 3), and this is now the case for knowledge and facts about the environment and nature.
From these developments, both environmental media and open data platforms, including the
EPA’s, the Guardian’s Environment platform, and the InfoAmazonia (2015) journalism network,
that produce data, maps, and visualizations on climate change have become a focal point for
environmental knowledge production. These governmental, commercial, and community-based
media platforms highlight a dataficational turn in environmental media, where data journalism
can be seen as an interactive process of mapping complex ecological developments globally and
within local communities.
This chapter explores environmental platforms, interfaces, and open data through their com-
putational architecture, interactive affordances, and environmental knowledge production. In
particular, the focus is on the use of digitally created open data, platform features, interfaces, and
visualizations and the related digital affordances (Gibson, 1977). Within these, computational
journalism can be seen in dynamic and performative interactions between journalists, scientists,
communities, governments, and ecosystems.
Empirically, the chapter looks into three types of new digital platforms as environmental
media: public governmental bodies, i.e., the Environmental Protection Agency (EPA) Earth Sci-
ences platform; the Guardian’s environment platform; and the InfoAmazonia platform with its
community-based participatory mapping. Analytically, the chapter applies Latour’s (2005) ana-
lytical method of reassembling environmental platforms by analyzing their digital affordances
and actor-networks, as well as their roles in creating socio-technological assemblage while con-
structing digitalized nature (Salovaara, 2015, 2016).
ENVIRONMENTAL MEDIA AS ASSEMBLED NETWORKS
Media-science-open data platforms have appeared not only as socio-technological phenomena but
also as sites of the emergent digital spaces of public deliberation on the environment and climate
change. The current digital media ecosystem has changed the epistemological basis of how the
new digitally connected environmental public sphere is constructed. Instead of simply drawing
from journalism, environmental knowledge networks include scientists, open data platforms,
government units, mathematicians, developers, geo-data storages, politicians, and journalists. The
governing metaphors of these activities have been those of assemblages and networks.
The concept of assemblage stems from Deleuze and Guattari’s work and refers to ‘multiplicity’
within a phenomenon, instead of seeing it as separate, bounded, stable, and structured (Deleuze
and Guattari, 1987: 158). This mode of materialization of nature leads to transformative readings
without definitive meaning, while different organizations produce knowledge on nature. Patton
defines assemblage as “consisting of a multiplicity of heterogeneous objects, whose unity comes
solely from the fact that these items work together” (Patton, 1994: 158). DeLanda outlines assem-
blage as “being wholes whose properties emerge from the interaction between parts” (DeLanda,
2006: 3). Therefore we can see nature as a constellation that depends on time, space, ideological
climate, and media.
The network metaphor is adept at describing digital media ecosystems, as it stresses a nonhi-
erarchical way of thinking how environmental knowledge and data are produced and offered to
402
Nature as knowledge
the publics. Digital networks are also capable of constituting seemingly fluid but nevertheless
complex “power geometries” (Massey, 1993), as networks create hubs of political and environ-
mental agencies.
According to the actor-network view, environmental knowledge, visualizations, and data
platforms have meaning or action specifically when they are part of an assemblage of people,
discursive processes, and material things. According to this view, environmental visualization,
participatory mapping, and multiple platforms have agency if they are deployed in the actor-
network of practices. (Perkins, 2006; Salovaara, 2016). As media platforms and interfaces have the
productive power to modify multiple publics’ information landscape, often they effectively form
socio-technological assemblages, such as actor-networks binding the three open data platforms
explored in this chapter.
Actor-network theorists (Latour, 2003, 2005; Serres and Latour, 1995) argue for politics of
articulation (in both senses, as speaking and as linking) that direct attention to the heterogeneous
practices of such platforms, where natures and cultures are continuously and mutually reconsti-
tuted. Here data visualizations and data journalism on climate change, temperature maps, and
satellite pictures become embodied knowledge by constructing tangible facts.
In this chapter, new environmental platforms are analyzed as open-access data journalism
platforms, as knowledge-producing actor-networks, and as territorial assemblages. The data ana-
lyzed consist of the data architecture of the platforms, their technical and semantic features, and
the background information related to launching and maintaining the platforms. The method-
ological approach is inspired by Bruno Latour’s (2005) analytical method of reassembling the
platform-space by analyzing (1) the technological web 2.0 affordances of the platform, (2) the
features of the digital, participatory actor-networks, and (3) the roles of platforms as social and
technological assemblages.
ENVIRONMENTAL OPEN DATA PLATFORMS
The U.S. EPA data platform is a top-down, official, public data platform hosted by NASA. The
EPA was established in December 1970 in the wake of elevated concerns about environmental
pollution and the first signs of climate change. Its mission was to consolidate into one agency
a variety of federal research, monitoring, standard-setting, and enforcement activities to ensure
environmental protection. The EPA’s functions include: “environmental protection contributes
to making American communities and ecosystems diverse, sustainable, and economically
productive” (EPA, 2017a).
The EPA monitors and observes Earth systems with instruments, such as satellites, remote
sensors, and data collection network nodes. The primary function of the agency is to widen sci-
entific understanding and descriptions of Earth systems through research and analysis of collected
data. The EPA’s current five “fundamental activities” are: evaluating and predicting the changes
of these systems over time; “Engaging, advising, and informing the public and partner organisa-
tions with valuable information”; and finally “managing resources for the betterment of society,
economy and environment” (EPA, 2017b).
The British daily newspaper, the Guardian, was established in 1821 as the Manchester Guardian.
Today, the Guardian newspaper and its online platform are a part of the Guardian Media Group,
owned by The Scott Trust Limited. The Trust’s ownership secures the financial and editorial
independence of the Guardian against commercial and political interference, and its historical
motto is “The Facts are sacred”. The Guardian is not known only for its investigative and data
journalism, but also for scoops related to multiple data leaks. The Guardian online platform
adopts an open and free approach to access that is supported by ads and crowdsourced funds
403
Inka Salovaara
from volunteer Guardian supporters. The platform has been among the most widely read news
providers in the world, reaching over 100 million readers daily. (Guardian, 2014).
The InfoAmazonia environmental platform was established in June 2012 as an environmental
journalism network reporting on the Amazon delta. The rationale behind the InfoAmazonia data
journalism platform was to activate environmental data journalism with the help of the local
communities within the Amazon delta. InfoAmazonia contributes specifically to the debate for
more sustainable development in this area of global importance. It provides news, reports, and
data maps from a network of journalists that cover the nine countries that share the Amazon
delta. The platform also provides open access to geo-referenced data and map templates for visu-
alization on environmental changes within the region, in addition to tools to use and visualize
these data for environmental reporting (Salovaara, 2016).
ENVIRONMENTAL MEDIA AND PLATFORMIZATION
The three environmental and data platforms exhibit, to a certain extent, similar technological
and user-specific characteristics. The first structural and common feature can be defined as
‘platformization’, which according to Helmond (2015: 5) refers to “the rise of the platform as the
dominant architectural, infrastructural, and economic model of the web 2.0” (Helmond, 2015: 5).
Technologically, this means the platforms analyzed provide mark-up language for creating
native applications that make them ideal platforms to attract user interaction. This type of com-
putational architecture explicitly opens up websites by enabling their programmability with soft-
ware interfaces for third parties.
Platformization entails the extension of networked affordances of media and data platforms
“reaching to the rest of the web 2.0 through making their external data platform ready” to
enhance platform content circulation and brand leverage with the help of users and developers
(Helmond, 2015: 5). The platform architecture, according to Gillespie (2011: 352), contributes
to the democratizing power of web 2.0, and breeds user-generated content (UGC), including
amateur contributions, produser’s creative input, mash-ups, peer-level social networking, online
community networks, and online expert feedback (Benkler; 2006; Bruns, 2008, 2012; Bruns and
Schmidt, 2011; Burgess, 2007; Jenkins, 2006).
Platformization is also a way to create more sustainable and robust business models, as media
platforms can offer diverse content (data, documents, interactive maps, and journalism) online,
mobile format, and paper format. As many environmental platforms grow larger and become
visible, they start to attract attention, not just from their users, but from broader and more diverse
publics, which makes them politically and ideologically vigorous actors (Helmond, 2015; Gil-
lespie, 2009).
However, the new and hybrid platforms are not considered platforms only because they
allow coding and data usage but also because they also allow communication, interactivity, and
commodification of user data (Gillespie, 2011: 351). This gives environmental media and data
platforms a voice and an agency in the global, networked public sphere. Hence Gillespie (2009)
argues that various online platforms are becoming more akin to traditional media than they
would want to acknowledge.
Environmental platforms as discursive interfaces

While platformization turns the chosen environmental platforms into web 3.0 actors, they
communicate through their visual interfaces with multiple geo-located publics. All three
platforms have a global reach with English speaking/reading publics (based on English as the
404
Nature as knowledge
international lingua franca). The platform interfaces refer to coded places of interaction and
intercommunication between publics and platform, where there is a shared space or juncture
between two parties or systems. That is an interface by definition; whenever two interactive
actors come into contact, human or nonhuman, and they interact, an interface exists. Discursive
interfaces reflect and help establish a common cultural sense about what users can, cannot, and
should do. Hence interfaces produce the possible and normative action frames rather than acting
on any particular individual. Environmental platform interfaces have the same logics as any
media. They guide and herd users to act according to the interface’s content priorities and
enhance interactivity with the interface while also collecting data on users.
Environmental platforms communicate and allow interaction through their interfaces.
Interfaces simultaneously assist and restrict and modify communication, and this can take both
concrete and abstract forms. Typically, interactive texts or data-mapping templates function as
sub-interfaces that promote certain forms of interaction with the platforms. Through these
forms of assistance and restriction, the interface defines the potential interactions that are avail-
able: reading, uploading, downloading, and navigating.
The navigational architecture of the interface itself leads and pushes users to interact with a
platform in the ways preferred by that platform (November and Camacho-Hubner, 2010; Holmes,
2002). Interactive documents, for example, are the points of contact with the narrative as an object,
with the content as ideas, and with the author as a “constructed ethos” (Carnegie, 2009: 167).
Regarding computer-mediated communication, the notion of interface as a constructed ethos cov-
ers a range of issues including the physical, social, political, cultural, technological, and ergonomic
configurations of computer systems, screens, software, and human-computer interaction.
For environmental interfaces, the constructed ethos also gives these interfaces productive
power, as their web interfaces both reflect and reinforce informational, political, and social log-
ics (Stanfill, 2015). Moreover, interfaces have productive power as they govern the hierarchies of
information and data, visibility, and navigational practices. Interfaces can be seen as discourses
in that they are “practices that systematically form the objects of which they speak” (Foucault,
1972: 49). They are places where thought and action are structured by sets of knowledge assump-
tions about what are right or correct environmental phenomena (populations, territories, and
resources) that have given environmental media a political agency of revealing processes.
Environmental platforms as open data providers

Open data, by definition, refer to data that are accessible to anyone for use, reuse, and distribution,
with the responsibility to attribute the source of the data (Pollock, 2006; Kitchin, 2014). Open
data also often include the absence of technological restrictions, though they sometimes includes
integrity clauses, often stipulating that the modified version of the datasets should be presented
with a different name than the original version (Kitchin, 2014: 15).
Open ‘big data’, or content data in media platforms, have four features. According to Aberna-
thy (2017: 7), they are volume, variety, velocity, and veracity. Volume in ‘big’ data simply concerns
the large datasets that are produced daily, often in real time, where it is almost impossible for many
to capture even a tiny portion of it. The data variety refers to the diversity and heterogeneity of
data from social media content, Google searches, websites, emails to satellites, and remote sensors
to smart Internet of Things appliances. Velocity as a big data feature reflects how computer users
not only produce data, but they amass it faster than ever. As Abernathy (2017: 30) notes, in an
“internet minute”, users produce a staggering number of yottabytes of data (1 yottabyte equals
one trillion terrabytes, or 1024 bytes). Finally, data veracity considers the accuracy, correctness, and
truthfulness of the data. In an age of disinformation, the validity and reliability of data increas-
ingly require criticality about its authenticity and preciseness (Abernathy, 2017: 7–9).
405
Inka Salovaara
All three examples of data platforms studied here provide Open Application Programming
Interfaces (APIs) for developers and users to explore platform data. From a user’s perspective,
an environmental platform such as InfoAmazonia enables communities, sharing of content,
and other socio-informational network features. InfoAmazonia also has an open-source Word-
Press theme, which has been developed specifically for using the MapBox API and allowing
journalists and the public to post and geo-locate stories directly on the platform Amazon maps
(Salovaara, 2016). Also, InfoAmazonia Content Management Systems allow integration with
the data layers hosted by MapBox, and the maps are now highlighted according to specific
demands (Salovaara, 2016).
The EPA offers its Developer Guide, and the “Linked Open Data” exposed through this
platform is available for developers to then build their applications. This guide outlines how the
data is accessed and shows some example coding. Therefore, their platform architecture is open
and facilitates egalitarian participation and being part of the participatory web. The EPA API is
available for people to build queries to retrieve the data for an area of interest. The data can be
tailored to, or even selectively to, a specific need, such as for mobile devices. The EPA also pro-
vides an interactive tool to make it easier to use the API and build the necessary parameters to
retrieve the data. The how-to guide is attached to the tool (EPA, 2017b).
The Guardian online provides ‘Open API’ for its journalistic content. Its climate platform
provides both a developer API for free, as well as a commercial API that enables the customer to
explore the platform affordances. As with some of the big news providers, such as the New York
Times, National Public Radio (NPR), USA Today, and the BBC, the Guardian has also embraced
the open and free web approach by providing Open API as part of their economic and develop-
mental approach in the engaging audience in new ways.
According to Aitamurto and Lewis (2013), Open API brings multiple benefits for media
organizations. Through a content API, both the media and public organizations can, for example,
offer content in multiple ways on all platforms with the innovative applications that wouldn’t
have been done in-house. Simultaneously, Open API gives a new life to content and serves niche
audiences. An additional benefit for the organizations is the quest for retooling in-house Content
Management Systems (CMS) that are both easier to use in-house and for external developers’
data mining. Additionally, the development teams in news organizations learn to focus on the
user experience. Opening gates to external developers accelerates development, cuts costs, and
enhances collaboration between tech and editorial teams on media platforms. Commercial Open
Data media platforms also gain revenue from the APIs, driving traffic to platforms while spread-
ing content across the web; they gain improved brand leverage (Aitamurto and Lewis, 2013).
Analyzing new environmental media as actor-networks

The aim of understanding environmental data and journalism as an actor-network is to explain how
‘ordering effects’ – such as devices (remote sensors, satellites, computers), organizations (laboratories,
newsrooms), agents (scientists, journalists), and even knowledge (climate change as empirical
phenomena) – are generated within the network. Analysis of actors-networks has its primary focus
on investigating how different actors (platforms, scientists, journalists) and entities (organizations)
are performed and kept stable (Perkins, 2006; Dodge et al., 2009; Hassard, 2013; Salovaara, 2015,
2016). The actor-networks are built around nodes and actors (both human and nonhuman) but also
composed of embodied knowledge or competences that keep precarious actor-networks together.
The first feature in actor-networks, such as new types of environmental media, is boundary
objects that refer to objects such as technical standards that enable the sharing of information
across networks. A boundary object is simply the organizational knowledge ecology of the plat-
forms. These technological knowledge-management techniques are relatively similar on all three
406
Nature as knowledge
platforms, i.e., on the EPA platform, Guardian Environment platform, and InfoAmazonia. All
three rely on the circulation of their content and push their content out for digital and public
circulation. They all have an open and free approach within the structures of web 2.0, whether
as a public service platform (the EPA), as open journalism (the Guardian), or in InfoAmazonia’s
community-based participatory data-mapping and journalism networks.
Centers of calculation within the platforms refer to locations, such as platform agencies, where
observations accumulate, are synthesized, and are analyzed. At the Guardian and InfoAmazonia,
the observations accumulate in the international journalism network nodes and newsrooms as
well as in participatory mapping templates with the interaction between data analysts and jour-
nalists. InfoAmazonia also relies on local communities in the Amazon delta that act as centers of
calculation and produce observations, collect data, and generate environmental mapping in real
time. In the EPA, scientists collect the data from remote sensing devices and publish knowledge
based on those observations, as well as the raw data.
Inscription devices and geo-data concern technical artifacts that record and translate informa-
tion, such as tables of coordinates or satellite imaginary. The EPA and InfoAmazonia use techno-
logical devices to accumulate data, such as satellites, remote sensors, and remote sensing devices,
as well as infrared space telescopes, spectrometers, and ambient pollution sensors.
Obligatory points of passage signify sites in a network that exert control and influence. They can
be such things as a U.S. governmental department or indeed the president, such as in the case of
the EPA and its Open Data platform. In news media, the point of passage refers, for example, to
an editorial unit, the editor-in-chief, media regulators, and financial supporters.
Programs of action (the resources required for actors to perform certain roles) refers to different sets
of practices and strategies that allow the networks to remain viable and robust in order to function
as a node in broader networks of business models, funding strategies, and resourcing workforces,
as well as engaging multiple audiences and other actors that are involved in the action programs.
Trials by strength concern a process where different visions and processes within the network
compete for the superiority of being the knowledge or data provider. As actor-networks, they
create newly connected and global networks of information sources and news producers. Again,
here the platforms are growing politically visible and can be censored or harnessed when com-
peting visions and processes arise on the platforms that are not supported by the controlling
actors within the network.
Translation refers to the work through which actors modify, displace, and translate their vari-
ous and contradictory interests. Translation is a mechanism by which things take form through
‘displacements’ and ‘transformations’ – such as when actors’ identities, the possibilities of their
interaction, and the limits of their maneuvering are negotiated and delimited (Hassard, 2013). In
this case, an example of ‘translation’ was when the EPA was translated as a political agency instead
of a scientific research unit and had to go under review before publicizing its research.
The affordances of the environmental data and knowledge platforms such as the EPA, Guard-
ian, and InfoAmazonia are relatively similar. They have an open and free web approach, enabling
communities, and sharing of content, as well as other social network features. They provide
platforms as they afford a (global) discursive place from which to deliver, to speak, to be heard,
and for their content to be mediatized. They provide computational affordances, such as Open
APIs, for the developers and users to explore the platform data. Their architecture is open and
facilitates egalitarian participation through being part of the participatory web.
In so doing, through actor-network-theory, one can identify different actors within the new
environmental media sphere, including official intergovernmental agencies, such as the Earth
System Governance Project; governmental agencies such as the European Environment Agency
and the EPA; nongovernmental organizations, such as Greenpeace and the World Wide Fund for
nature; and participatory mapping platforms, such as The Rainforest Foundation’s Participatory
407
Inka Salovaara
Mapping Programme in the Congo Basin, Forest Peoples’ Participatory Mapping Programme,
and Water Aid Participatory Mapping.
By extending agency to the nonhuman (such as technological and digital devices, data, and
embodied knowledge in the forms of data-created artifacts), nature inhibits the production net-
work of strong social constructivism. In this way, bringing together human and nonhuman
actors, the construction of nature and of the environment exist only through knowledge that is
produced continuously, made and remade, and repeatedly performed so that the complex and
precarious networks will not dissolve (Hassard, 2013: 5–6).
Environmental media and nature as assemblage

‘Your Planet Is Changing. We’re On It’.
NASA uses the vantage point of space to increase our understanding of our home planet,
improve lives, and safeguard our future. We monitor Earth’s vital signs from land, air, and space
with a fleet of satellites and ambitious airborne and ground-based observation campaigns. NASA
develops new ways to observe and study Earth’s interconnected natural systems with long-term
data records. The agency freely shares this unique knowledge and works with institutions around
the world (NASA, 2017).
While writing this chapter, on its platform NASA still promises to improve lives and safeguard
our future on this planet. However, the EPA is expecting budgets cuts of one-third, and the Trump
administration said in 2017 it wanted to eliminate a quarter of the agency’s employees. President
Trump has just finished his first visit to Europe, where he discussed the Paris climate agreement
with the G7 leaders, after which he announced he was pulling out of the Paris agreement on cli-
mate change. There is increased concern within the European Union, and his visit and subsequent
announcement have left many EU leaders shaken. As Trump pulls back from the Paris agree-
ment, the US still tops the list of carbon emissions producers in the first world, a situation that is
untenable, with even major energy companies, such as ExxonMobil and Shell, seeking to persuade
President Trump not to abandon the global climate agreement reached in Paris.
It is not surprising that considerable attention has been paid to environmental neoliberaliza-
tion in news media. Environmental and economic liberalization are still uneven, spatially dif-
ferentiated, and include contradictory processes. Global interests in processes of deregulation
and reregulation of environment enable the refiguring of nature regarding ‘services’ and ‘natural
capital’. The effects of such processes and policies on communities and ecologies are often fatal.
The notions of neoliberal readings of environmental information and the rise of nondeterminis-
tic understandings of nature have occurred in conjunction with the neoliberalization of climate
change governance (Braun, 2015; Bakker, 2005).
Contemporary environmental media and data platforms are becoming simultaneously more
than information and science. Environmental knowledge has expanded to include geographic
databases, mapping software, geo-coded cartographies, interactive interfaces, and algorithms for
automated map analysis. When reassembling ‘nature’ on environmental platforms, constellations
consist of things, technologies (software, tools, satellites, data structures), practices, competences,
networked flows of information, organizations, agencies, and digital ecologies created by ubiqui-
tous computing (interfaces, servers, platforms, data clouds, code).
Instone (2004) rightly argues that the co-construction of environmental knowledge (as well
as political decisions) blurs the boundary between politics, science, and technology and extends
agency to heterogenic collectives and networks – both natural and artificial. Haraway (1991: 297)
contends that our environment is a co-construction “among many actors, not all of them human,
not all of them organic, not all of them technological”. This becomes very clear when platforms
are recognized as actors within the public discussion of climate change. This view emphasizes
408
Nature as knowledge
multiplicities and articulation of information, heterogeneous collectives, networks, assemblages,

and material-semiotic entities – envisioning the environmental knowledge is co-constructed by
multiple agents.
Bruno Latour (2005) calls this the “Janus face” of science. When nature is produced once by
measuring, picturing, politicizing, observing, and constructing, it is considered to be always and
already true. Hence nature turns into “assemblage” (Delanda, 2006). In the environmental media
platforms, discursive interfaces create their own category of ‘nature’ depending on their political
bending. As an assemblage, nature becomes a temporary structure dependent on time, space, and
ideological climate (Thrift, 2007; Bennett, 2010; Harris and Hazen, 2009).
Though publics mostly agree that climate change is not a hoax, there remains a fast-growing
distance between the environmental space where climate change has an impact and the political
space where such issues are managed. These complex and common problems are the source of
the distinct political crises that affect the institutions of global and regional governance (Castells,
2008). Specifically, the crisis of efficiency is critical as it concerns complex issues, such as climate
change, where ethical issues are intertwined with corporate capitalism and international politics
and hence cannot efficiently be solved. The crisis of identity signifies the transferring of political
power to supra-national actors, such as the EU, UN, and World Bank, and leaving the ‘organic’
democracies feeling powerless, played by stronger agencies and securing the assets of the strongest
states (Castells, 2008).
The crisis of legitimacy hits nation-state democracies that are not built for dealing with supra-
national challenges, as territorial democratic systems that are too restricted to deal beyond their
borders. Finally, according to Castells (2008), there is the crisis of equity at the geopolitical level
in relation to climate change that relates to the global political system where there are strong
nations and groups, that will always be more powerful, more able to push other players with
economic, military, or political threats and sanctions.
Climate change and global warming are objects that cut through the field of these crises, sci-
ences, politics, geography, technology, security, defense, and political ecology. This sees environ-
mental changes as meta-objects of political discussion, as they exist everywhere, have an impact
on everyone and everything, but do not have a specific location.
Epistemologically, therefore, ‘nature’ and ‘environment’ are particularly difficult objects for
knowledge production, and especially for journalism. This is pronounced when there is no
society out there to which scientists respond as they build their theories, nor is there nature to
constrain them to a unique telling of their stories. Rather, both scientists and journalists stand
between nature and society, as well as between politics and technology.
Considering the digitalization of nature, we must further explore how human and nonhu-
man knowledge contributes to the production of knowledges, networks, and social assemblages
around environment. Computational knowledge delivery causes ontological changes about the
ways in which technologies have inspired the new and flourishing study of embodied experience
and technology that nest between boundaries of techno-science, activism, and communities.
As digitalized nature(s) can be seen as products of globalization produced through comput-
erization, datafication, and visualization, they are productive of their forms, gaining distributed,
post-human agencies and, as such, they exhibit emergent features of digital environmental glo-
balism. Hence new environmental media and digitalized nature create new moral geographies of
action where affect and engagement create new dynamics around nature(s).
Further reading
This chapter has benefited remarkably from Bruno Latour’s Reassembling the Social: An Introduction
to Actor-Network-Theory, Politics of Nature (2005), as well as his Science in Action (1987). Rethinking
409
Inka Salovaara
Maps: New Frontiers in Cartographic Theory (2009), edited by Dodge Martin, Kitchin Rob, and
Chris Perkins, is a fantastic book from human geographers on how nature and environment
are constructed as maps and visualizations. In addition, Manuel DeLanda’s A New Philosophy of
Society: Assemblage Theory and Social Complexity (2006) is a clear and systematic presentation on
assemblage theory.
References
Abernathy, D. (2017) Using Geodata & Geolocation in the Social Sciences: Mapping Our Connected World. Los
Angeles; Washington; New Delhi: Sage.
Aitamurto, T. and Lewis, S. (2013) “Open innovation in digital journalism: Examining the impact of Open
APIs at four news organizations.” New Media & Society, 15(2), 314–331.
Bakker, K. (2005) “Neoliberalizing nature? Market environmentalism in water supply in England and
Wales.” Annals of the Association of American Geographers, 3, 542–565.
Benkler, Y. (2006) The Wealth of Networks: How Social Production Transforms Markets and Freedom. New Haven,
CT: Yale University Press.
Bennett, J. (2010) Vibrant Matter: A Political Ecology of Things. Durham, NC: Duke University Press.
Braun, B. (2015) “New materialisms and neoliberal natures.” Antipode, 47(1), 1–14.
Bruns, A. S. (2008) Blogs, Wikipedia, Second Life, and Beyond: From Production to Produsage. New York, NY:
Peter Lang Publishing.
Bruns, A. S. (2012) “Reconciling community and commerce?” Information, Communication & Society, 15(8),
815–835.
Bruns, A. S. and Schmidt, J.-H. (2011) “Produsage: A closer look at continuing developments.” New Review
of Hypermedia and Multimedia, 17(1), 3–7.
Burgess, J. (2007) Vernacular Creativity and New Media. Queensland University of Technology, Brisbane,
Australia. Retrieved from http://eprints.qut.edu.au/archive/00010076/01/Burgess_PhD_FINAL.pdf
Carnegie, T. (2009) “Interface as exordium: The rhetoric of interactivity.” Computers and Composition, 26(1),
164–173.
Carusi, A., Sissel Hoel, A., Webmoor, T. and Woolgar, A. (2015) Visualization in the Age of Computerization.
New York, NY and London: Routledge.
Castells, M. (2008) “The new public sphere: Global civil society, communication networks, and global gov-
ernance.” The Annals of the American Academy of Political and Social Science, 6(16), 78–93.
DeLanda, M. (2006) A New Philosophy of Society: Assemblage Theory and Social Complexity. London:
Bloomsbury.
Deleuze, G. and Guattari, F. (1987) A Thousand Plateaus: Capitalism and Schizophrenia. Minneapolis, MN:
University of Minnesota Press.
Dodge, M., Kitchin, R. and Perkins, C. (2009) “Thinking about maps.” In D. Martin, K. Rob and C. Perkins
(eds.), Rethinking Maps: New Frontiers in Cartographic Theory. London and New York, NY: Routledge (pp.
1–25).
Dodge, M., Kitchin, R., and Perkins, C. (eds.) (2009) Rethinking Maps: New Frontiers in Cartographic Theory.
Abingdon: Routledge.
Foucault, M. (1972) The Archaeology of Knowledge & the Discourse on Language. New York, NY: Pantheon
Books.
Gibson, J. (1977) “The theory of affordances.” In R. Shaw and J. Bransford (eds.), Perceiving, Acting, and
Knowing: Toward an Ecological Psychology. Hillsdale, NJ: Lawrence Erlbaum (pp. 127–143).
Gillespie, T. (2009) Wired Shut: Copyright and the Shape of Digital Culture. Cambridge, MA: MIT Press.
Gillespie, T. (2011) “The politics of platforms.” New Media & Society, 12(3), 347–364.
The Guardian. (2014) The Guardian’s Road to 100 Million Unique Users. Retrieved from www.theguardian.
com/media/video/2014/apr/17/guardian-road-100-million-unique-browsers-video
Haraway, D. (1991) Simians, Cyborgs and Women: The Reinvention of Nature. New York, NY and London:
Routledge.
Harris, L. and Hazen, H. (2009) “Rethinking maps from more -than-human-perspective: Nature-society,
mapping and conservatory territories.” In M. Dodge, R. Kitchin and C. Perkins (eds.), Rethinking Maps:
New Frontiers in Cartographic Theory. London and New York, NY: Routledge (pp. 50–67).
Hassard, J. (2013) “Actor-network theory.” In E. H. Kessler (ed.), Encyclopedia of Management Theory. Thou-
sand Oaks, CA: SAGE Publications, Inc. (pp. 19–22).
410
Nature as knowledge
Helmond, A. (2015) “The platformization of the web: Making web data platform ready.” Social Media +
Society, 1, 1–11.
Holmes, D. (2002) “Transformations in the mediation of publicness: Communicative interaction in the
network society.” Journal of Computer-Mediated Communication, 7(2). doi:10.1111/j.1083-6101.2002.
tb00142.x
InfoAmazonia. (2015) Geojournalism, Maps and Data. Retrieved from http://infoamazonia.org/about/
Instone, L. (2004) “Situating nature: On doing cultural geographies of Australian nature.” Australian Geog-
rapher, 35(2), 131–140.
Jenkins, H. (2006) Convergence Culture: Where Old and New Media Collide. New York, NY: New York Uni-
versity Press.
Kitchin, R. (2014) The Data Revolution: Big Data, Open Data, Data Infrastructures & Their Consequences. London
and Los Angeles, CA: Sage Publications.
Latour, B. (1987) Science in Action. Cambridge, MA: Harvard University Press.
Latour, B. (2003) We Have Never Been Modern. Cambridge, MA: Harvard University Press.
Latour, B. (2005) Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford: Oxford University
Press.
Massey, D. (1993) “Power-geometry and a progressive sense of place.” In J. Bird, B. Curtis, T. Putnam, G.
Robertson and L. Tickner (eds.), Mapping the Futures: Local Cultures, Global Change. London: Routledge
(pp. 59–69).
NASA. (2017) Earth Right Now. Retrieved from www.nasa.gov/topics/earth/overview/index.html
November, V. and Camacho-Hubner, V. (2010) “Entering a risky territory: Space in the age of digital navi-
gation.” Environment and Planning D: Society and Space, 28(3), 581–599.
Patton, P. (1994) “Metamorpho logic: Bodies and powers in A Thousand Plateaus.” Journal of the British Society
for Phenomenology, 25(2), 157–169.
Perkins, C. (2006) “Mapping golf: A contextual study.” The Cartographic Journal, 43(3), 208–223.
Pollock, R. (2006) The Value of Public Domain. Retrieved from https://rufuspollock.com/papers/value_of_
public_domain.ippr.pdf
Salovaara, I. (2015) “#Je Suis Charlie: Networks, affects and distributed agency of media assemblage.” Con-
junctions: Transdisciplinary Journal of Cultural Participation, 2(1), 103–115.
Salovaara, I. (2016) “Participatory maps: Digital cartographies and new ecology of journalism.” Digital Jour-
nalism, 3(2). doi:10.1080/21670811.2016.1173519
Serres, M. and Latour, B. (1995) Conversations on Science, Culture, and Time. Ann Arbor, MI: University of
Michigan Press.
Stanfill, M. (2015) “The interface as discourse: The production of norms through web design.” New Media &
Society, 17(7), 1059–1074.
Thrift, N. (2007) Non-Representational Theory: Space, Politics, Affect. Abingdon: Routledge.
United States Environmental Protection Agency (EPA) (2017a) Our Missions and What We Do. Retrieved
from www.epa.gov/aboutepa/our-mission-and-what-we-do
United States Environmental Protection Agency (EPA) (2017b) Developer Central: Resources for Developers.
Retrieved from http://developer.epa.gov/nepassist-report-api/
Williams, R. (1980) Problems and Materialism and Culture. London: Verso.
411
31
OPTING IN AND OPTING OUT
OF MEDIA
Bonnie Brennen
A movement begins
On April 9, 2017, 60 Minutes broadcast a story on brain hacking, which is the notion that our
phones, social media, and apps are being designed to program us. During the report, Anderson
Cooper (2017) interviewed Silicon Valley insiders who suggested that the constant stream of
distractions have been engineered to create addictive behavior. These distractions are shaping our
feelings, thoughts, and actions, weakening our relationships and ruining our children’s ability to
concentrate and focus. The story highlighted researchers who suggested that our smartphones
now keep us in a continuous state of anxiety where we feel the need to constantly touch them,
check them, and interact with them. Habit-breaking apps are being designed, but ultimately the
program suggested that in order to break our digital technology habit that we need to limit our
use and take time away from new media. The 60 Minutes report represents a growing trend of
questioning the impact of digital technology. It is a trend that contradicts a prevailing norm in
the United States that everyone should embrace new technologies and that people must have full
access to digital media so that they can obtain information necessary to function in a democratic
society.
While we haven’t quite reached a tipping point yet, during the last few years, public calls
to unplug, slow down, reject, restrict, and/or opt out of new media have become more urgent,
sustained, and popular. Tech leaders Elon Musk and Bill Gates have cautioned that the develop-
ment of artificial intelligence (AI) may have catastrophic results for humanity. The late renowned
physicist Stephen Hawking has echoed their concerns about AI and has warned that the internet
could soon “become a command center of terrorists” (Griffin, 2014). Even Apple co-founder
Steve Woszniak has publicly wondered if human beings will one day become family pets for
“robot overlords” (Dowd, 2017).
Signs such as “There is no app to replace your lap! Read to your child,” “The use of WMDs
(wireless mobile devices) is not permitted”, “This is a screen-free zone”, and “Talk to me – not
the screen” increasingly pop up outside of coffee shops and restaurants. Digital detox retreats,
vacations, and programs, designed to help break internet addictions, are becoming increasingly
popular. Five star hotels, game reserves, golf resorts, spas, and camps throughout the world now
promote “disconnect to reconnect” holidays, with specially designed vacation packages to help
people unplug.
412
Opting in and opting out of media
The Sabbath Manifesto, created by Reboot, an organization dedicated to affirming Jewish

values and traditions, has developed a digital Sabbath, a weekly 24-hour period where people
are encouraged to avoid technology, connect with friends and family, and give back. Journalist
Andrew Sullivan (2016), a self-described early adopter of “living-in-the-web”, decided to quit
digital technologies after he realized that “Every hour I spent online was not spent in the physi-
cal world. Every minute I was engrossed in a virtual interaction I was not involved in a human
encounter” (ibid.). Sullivan now lives exclusively in the real world and supports a digital Sabbath
to help people rebalance their lives.
New York Times columnist Ross Douthat (2017) has called for a digital temperance movement
to help us take back control of new technology. Finding Americans “enslaved” to an online life
that “breeds narcissism, alienation and depression” (ibid.), Douthat recommends creating more
tech free zones where internet usage is discouraged, to help restrict and control our digital tech-
nology usage. In addition, tech executives at Twitter, Tumblr, Huffington Post, and Yahoo now
recommend regular unplugging to help maintain a healthy balance between our virtual and real
lives (Bilton, 2013).
Screen-free camps help children and teens break their dependence on digital technologies,
and Screen Free Week, organized by the Campaign for a Commercial-Free Childhood, holds
an annual celebration each May to encourage individuals, schools, and communities to take a
break from technology and reconnect with family and friends. In addition, there are mobile
apps designed to help people unplug from technology. For example, Digital Detach “imposes a
timed detox session” (Borchers, 2015), turning off all functions of smartphones except making
telephone calls and texting, while the Digital Detox app turns off every mobile device feature
except dialing 911 in an emergency.
Contemporary research indicates that the level of digital inclusion in the US continues to
increase and that it is becoming less dependent on education, income, race, location, age, and gen-
der. Yet, according to a September 25, 2013, Pew Research Center report, in the US 15% of adults
age 18 and above (or about 35 million people) do not use email or the internet at all. While some
of these individuals find it too expensive, too difficult, or too frightening to use, 35% of those
who do not use email or the internet reject it because they do not find it relevant, interesting, or
integral to their lives. If the Pew Center report is correct, more than 10 million adult Americans
are actively choosing not to engage with email and the internet at all (Zickuhr, 2013). In addition,
a growing body of research suggests that millions more U.S. citizens are taking an active role in
deciding which technologies they engage with and which ones they choose not to use.
Responding to the opting-out trend, this project focuses on academic research, news reports,
and popular culture discussions about individuals who are actively making decisions regard-
ing their use or nonuse of new technology. It also draws on insights from 79 in-depth semi-
structured qualitative interviews that I conducted with people who indicated that they made
choices about their technology usage. Potential interviewees responded to online requests and
notices posted in local coffee shops and bookstores and were recommended to me by friends,
colleagues, and other interviewees. The 41 men and 38 women interviewed were a culturally,
geographically, and ethnically diverse group whose ages ranged between 22 and 81. At the time
of the interviews, 70 of the interviewees were employed full time, four were students and five
were retired. The interviews ranged from 50 minutes to two and a half hours in length, and 40
people were interviewed in person, while 39 interviews were conducted over the telephone.
Ultimately, my goal with the interviews was to understand the reasons why people chose not to
engage with some new technologies and how individuals made decisions regarding the technolo-
gies with which they interacted.
413
Bonnie Brennen
Conceptual foundations
Considerable research across several disciplines has investigated the use of new technology from
a digital divide perspective, framing the issue from a dichotomy of the haves versus the have-
nots. The technology haves have been envisioned as living well and acquiring a vast array of
technological resources. The have-nots are thought to exist in unstable environments with
limited resources, unable to afford new technologies. Insisting that those who are left behind the
digital revolution will become a permanent part of the underclass, digital divide researchers have
maintained that the use of new technologies is integral to maintaining our economic basis and
standard of living. As Mack (2001: 42) explained:
While computers and the internet are certainly no panacea for all of society’s ills, these
technology resources assist people in developing and improving their skills, knowledge,
marketability, and income. Ironically, those who could most benefit from these resources
are often foreclosed from acquiring or accessing them.
Assuming that access to new technologies is fundamental to success in the twenty-first century
and that no one would voluntarily reject new technologies, digital divide researchers have
sought to understand the conditions, issues, and problems restricting individuals’ access to new
technologies through considerations of class, race, ethnicity, socioeconomic status, income,
education, geography, household type, and age.
Much of the digital divide research has been framed from a technological deterministic posi-
tion, which has maintained that new technologies have taken us from an industrial society to
an information society and are responsible for creating modern people and our modern culture.
More recently, as levels of digital inclusion have continued to grow, some researchers have ques-
tioned the continued relevance, basis, or existence of a digital divide. Yet Sparks (2013) main-
tained that the term digital divide continues to be used because it has been equated with social,
economic, and political inclusion and since its inception has had “a normative bias towards the
benefits of digital inclusion” (ibid.: 30). As a prevailing perspective on the role of technology
in social change, technological determinists envision the development of new technologies as
creating the necessary conditions for social change and the development of our modern world.
From this perspective, it is assumed that everyone should embrace digital technologies and that
people must have full access to them so that they can obtain information necessary to function
in a democratic society.
However, cultural theorist Raymond Williams (1974) considered technological determinism
an “untenable notion” because it substituted the random autonomy of intention or “an abstract
human essence” for individuals’ “real social, political and economic intention” (ibid.: 124).
Insisting that new communication technologies have been developed by people with specific
“purposes and practices already in mind” (ibid.: 8), Williams developed a theory of cultural mate-
rialism, which restored human intention and agency to the research and development of new
technologies. Cultural materialism addresses the specific creation and production of material
culture within its particular historical context. Privileging human experience as a fundamental
component of any cultural analysis, cultural materialism views people as active agents who help
create their own culture through their individual experiences (Williams, 1977). Framed from
a cultural materialist perspective, this research focuses directly on human intention and agency,
suggesting that people create and use new technologies with certain social needs, purposes, and
practices in mind. From this theoretical vantage point, I maintain that individuals make informed
choices as to which technologies they use and which ones they reject.
414
While the digital divide has figured prominently in scholarly research on the usage of new
media, in the last few years researchers have begun to consider some individuals’ deliberate choice
not to engage with new technologies. Although studies are still limited, most of this research has
either framed lack of engagement with new technologies as acts of resistance, refusal, or pushback
against the “evertime” (Morrison and Gomez, 2014) of constant connectivity or as a problem
or pathology.
Choosing the term “media refusal” in an effort to equate the actions of those who did not
use Facebook with a focused revolt of “conspicuous non-consumption” against consumer
culture, Portwood-Stacer (2013: 1047) suggested that Facebook media refusal was a way that
individuals crafted their own identities and remade their lives into sites of media resistance
against the powerful forces of contemporary society. Similarly, Woodstock (2014) invoked
the term “media resister” to represent an individual’s active and thoughtful rejection of new
technologies. Woodstock considered media resistance a complicated and sometimes contra-
dictory act:
As active, elective makers of meaning, media resisters may ignore particular types of
media content (such as news or popular culture), or they may refuse to adopt one
new media technology but not others. And like most of us, they are not necessarily
consistent.
(ibid.: 1987)
Similarly, Rauch (2014) has suggested that unplugging rituals to limit new media usage, like the
digital Sabbath, are active resistance strategies used by individuals to exert their agency about the
role of technology in their lives.
Researchers who focus on individuals who refuse, resist, or unplug consider the role of resist-
ers as actively making choices as to their use of new media. This approach differs significantly
from social science research on media resistance, which sees the use of new technology as normal,
appropriate, and expected and envisions media refusal as a weakness or a problem. For example,
Stieger et al. (2013) investigated the “phenomenon of virtual identity suicide” (ibid.: 629), which
was the term they used to describe individuals who decided to stop using social media like Face-
book or Twitter, while Karppi’s (2011) research referred to opting out of social media as “digital
suicide”.
At first glance, the media resistance research may seem to illustrate Williams’ theory of cultural
materialism, as it restores human intention to decisions regarding which technologies individuals
choose to engage with. Yet much of the resistance research focuses on individuals who base their
actions out of fears regarding the influence or effects of technology. For example, Woodstock
(2014) included commentary from individuals concerned about privacy issues related to new
technologies tracking personal information. One of Woodstock’s interviewees stated: “I became
more aware that I wasn’t fully in control or even fully aware how my information was being used
[. . .] I worry about what I don’t know” (ibid.: 1991).
This type of resistance to new technologies seems to actually be what Williams (1974)
has called a “peculiar doomsday brand of technological determinism” (ibid.: xxvii) because it
assumes that changes in our society are based solely on the development and introduction of
new technologies. From my reading of the resistance research, at least some of the experiences
of individuals included in the studies are not based on people making their own decisions but
instead are the result of individuals reacting to the perceived power of the technology to change
their lives.
415
Bonnie Brennen
Coexistence of diverse technologies

Morrison and Gomez (2014) have suggested a five-part typology to describe people’s motivations
for resisting new technologies: (1) emotional dissatisfaction with them; (2) a lack of use because
of religious, moral, or political reasons; (3) limiting use of technologies to save time and energy;
(4) pushback due to a fear of addiction to new media; and (5) online privacy concerns.
While I found several instances where my interviewees’ experiences related to this typology,
much of their commentary was not clearly ranked or differentiated. In addition, the majority
of the people I interviewed highlighted an additional motivational category: identity creation.
They explained that their incorporation or rejection of new technologies was an important way
of creating their own identities. Opting out provided them with guidance to help them actively
determine what issues and concerns were of importance to them and make technology fulfill
their needs and interests. Rather than viewing their actions solely as acts of resistance to tech-
nologies that are exerting power over them, these interviewees saw their actions as asserting their
independence, authority, and dominance over the technology.
The diversity of option about opting in and opting out of media raised issues regarding the
concept of new technology itself. When asked how they defined new technology, interview-
ees’ responses ranged from “all mobile-based technologies”, “everything using screens”, “smart-
phones”, “social media – especially Twitter and Facebook”, “everything online”, “recent
technologies that are better than the older ones”, and “just about everything new that’s not alive”.
In addition to being grounded in technological determinism, much of the digital divide
research has been framed around the “doctrine of supersession” (Eisenstein, 1997: 1054), a
perspective through which the development of each new technology, technique, or artifact is
thought to destroy the viability of the previous one. From this vantage point, it is imperative to
embrace new technologies so as not to be left behind. Yet insights gathered from this research
project did not support any type of a doctrine of supersession but instead illustrated what Eisen-
stein (1997) has referred to as a coexistence of diverse media technologies, styles, and artifacts.
The individuals I interviewed consciously and actively made choices about what technologies
they used. Some embraced the latest technological advances – at least some of the time – while
others were cautious about adding any new technologies to their lives.
Although people’s reasons for opting in and opting out were complex and multidimensional,
they all supported the idea that people are comfortable mixing new and old media technologies.
However, there was one older technology that all 79 people clearly preferred – reading printed
books. While some individuals used their Kindles, tablets, or iPads when traveling, and others
read PDFs on their computers when necessary, everyone said that they preferred reading books
on paper. Comments such as “printed books are indispensable”, “I love the touch, feel, and smell
of books”, and “to me it’s the physical weight of a book that makes me happy” reinforced the
findings of a 2016 study by the Pew Research Center, which found that 65% of adult Americans
had read a printed book during the last year (Victor, 2016).
The resurgence and growing popularity of vinyl also illustrates the notion of a coexistence
of diverse media technologies. Some people I interviewed preferred to stream music through
sites like Spotify and Pandora, while others listened to their favorite songs on their iPods, tablets,
phones, and computers. Some respondents spoke of their extensive CD collections, while others
listened to music on their favorite FM radio stations. However, several tech-savvy interviewees
spoke about rejecting digital music in favor of albums and turntables, while still others explained
how they had recently returned to vinyl because electronic music became too sanitized for them.
Mitch,1 a 22-year-old music lover from Denver, described his enjoyment of the physical aspects
of vinyl:
416
I felt a very strong physical and tactile sense with it that I really enjoyed. I loved the fact
that halfway through I would have to flip it, and it really transformed what had been
thoroughly background listening for me and into something that was very ritualistic. I
listen to a lot of music to try to relax. Listening to albums was almost meditative – even
simply watching the revolution of the album pleased me. I loved the sound of it as well.
Similarly, Rick, a recent graduate student in his twenties from Tennessee, began his vinyl collec-
tion because of the artwork and explained that he “loved the ephemera, the album artwork, the
knowledge. I liked the idea of looking at a complete idea of what it was”. Rick maintained that
buying an album is much different than purchasing an audio file or streaming music. “It’s the idea
of an album that you’re buying more than just the music itself. You’re buying an experience or look-
ing at the artwork, looking at the information and reading the lyrics. I find that more immersive”.
Cultural and religious beliefs

Some of the people I interviewed opted out of new technologies because of their ethical, political,
religious, or cultural beliefs. As a resident of an intentional community in the small town of
Viroqua, Wisconsin, 42-year-old Lars spoke of limiting his use of new technologies in order
to be more focused and present in his relationships with family and friends. While he used a
smartphone in his work as a professional salesman, in his home life Lars preferred to use a landline
“to initiate any kind of deep or personal conversation”. Throughout his interview he expressed
concerns about the numbers of people constantly staring at their cell phones, and he found that
the use of new technologies had “gone past utility into some addictive, weird, antisocial strange
thing”. Lars disliked all social media but considered Twitter the “silliest of them” because it forced
users to “squish your message down to like a fragment for no good reason”. Lars mentioned that
his wife does not own a cell phone and that his family never watched television. His 12-year-old
son was allowed to play video games on an old Mac laptop for one hour each day and to listen to
an iPod Touch, which he referred to as “the gateway drug of the technology thing”.
Lars explained that as part of a back-to-the-land cultural movement, Viroqua residents sought
to have more of an influence over their children’s education than was possible in public schools.
They developed a school system based on the Waldorf educational philosophy of Rudolf Steiner.
According to Lars, the foundation of the Waldorf system was built on the development of an
organic culture, integrating intellectual learning, emotional character building, and practical
physical labor in a natural setting. The school instructs all children on mindfulness; students
meditate each morning, use their hands to build things and play musical instruments, and explore
their environments daily. As a member of the school board and “the most anti-technology per-
son there”, Lars said that the school has attempted “to eliminate devices from all classrooms”.
While many of the families in Viroqua have not rejected the use of all new technologies, most
parents enforced the school’s request to curtail all family members’ screen time after 7:30 each
evening. He said that the rationale for limiting the use of new technology is that “the education
is so carefully designed and it should permeate their whole entire day, and you don’t want to give
them strong powerful images that will overwrite with their learning in school”.
Like Lars, who felt that his decision to restrict his family’s use of new technology was directly
related to his cultural and ethical values, other individuals also view technology as something
that comes second to their cultural and religious identities. For example, the Amish selectively
decide which technologies are good and which are bad for their community, and they think
carefully about the long-term effects that technology might have on their culture. The Amish’s
417
Bonnie Brennen
use of technologies is based on their beliefs in humility, obedience, simplicity, modesty, and their
submission to the will of God (Gelassenheit). Technologies that do not uphold the Gelassenheit
principles or which encourage sloth, luxury, or vanity are banned. So while automobiles are not
allowed because they are viewed as a status symbol that separates family, when necessary Amish are
not opposed to renting buses, traveling by train, or flying. And while they reject radio, television,
computers, and movies because of the outside values they disseminate, the Amish use generators,
flashlights, batteries, solar panels, and water wheels to generate electricity (Kraybill, 2001).
Many Orthodox Jews also do not watch television, play video games, use computers, or
engage with popular culture. Finding electronic media contrary to their way of life, Orthodox
Jews tend to shun outside mediated “distractions” in favor of spending quality time with their
friends and family. Sarah, an Orthodox Jew from New York, said that her religious beliefs and
values informed her family’s decisions about which new technologies they should use.
Our family is observant but not Hasidic, yet we also believe that electronic devices are
not conducive to our way of life. New technologies provide people with immedi-
ate gratification, and their use encourages them to consume more and more material
things. We do not want our children to be distracted by consuming things – we want
them to use their minds and their imaginations to learn and to explore the world.
All four of Sarah’s children attend a Jewish day school, which is technology-free, and parents
are strongly urged not to allow their children to access the internet or play video games at home.
Although Sarah’s husband uses the internet and a smartphone at work, the family does not have
a computer or television at home. All members of the family play musical instruments, and as
Sarah noted, “all of my children love to read and write”.
Private religious schools are increasingly developing technology policies in line with their
moral and ethical ideals. For example, the Milwaukee, Wisconsin, Yeshiva Elementary School
(2013) includes a technology policy in its Parent Handbook based on a belief that “internet use
and abuse have led to serious family and community concerns” (ibid.: 9). Yeshiva’s technology
policy recommends that the family’s home computers have filters to block social networking sites
and pornography, that children’s internet usage should be monitored, and that students should
not have email accounts unless parents oversee them. Students do not have access to the internet
at the school.
Fears of new technologies

While fears associated directly with the use of digital technologies have been addressed in the
digital divide and the resistance literature and have been a key issue in popular press articles about
new technology, none of the 79 individuals I interviewed explicitly said that they were afraid of
any new technologies. Instead, most of the interviewees said that opting out of some technologies
has helped them to stay healthy, and they applauded their decisions to stay proactive about their
technology usage. However, fears of technology may be inferred from some of the interviewees’
commentary. Some individuals addressed the potential “negative effects” of using new media
continuously, suggesting that without setting boundaries, new technologies could take over
people’s lives.
Rather than expressing explicit fears about what new media might be doing to them, some
interviewees spoke of their ambivalence toward using them. While several people named Google
Maps as their favorite app, and more than half of those interviewed reported loving their smart-
phones and said they would hate it if they had to give them up, several individuals felt that they
would get by just fine if they had to make do without any digital technologies.
418
Limiting the use of new media in order to save time and energy was another theme that
emerged in my interviews. For example, David, a 65-year-old retired physician from Northern
California, spoke forcefully about his desire to limit his use of new technologies. David did not
use social media or watch any entertainment programming on television, calling Facebook and
Twitter “a time suck” and television programs “boring and stupid”. He rarely watched movies,
and when he did, he went to a movie theater to watch rather than streaming them at home. He
owned a smartphone, using it for calls and sending brief texts. However, he used only one app
on his phone – a bicycle app that recorded his biking distance and times and allowed him to
share this information with his friends. David owned a laptop but rarely used it, preferring his
iPad because he used it “for consuming content, not creating it”. He reads medical journals, the
Washington Post and the New York Times each day on his iPad and read and responded to emails.
When traveling, David reads books on his iPad, but at home he only read hardback books. Over-
all, David felt that he managed his use of digital media appropriately in order to enhance his life
and has never felt the need to take a break from it. “It doesn’t overwhelm my existence or take
over my day, so I don’t feel the need to take a vacation from it”.
Privacy concerns
Recent advances in AI have resulted in smartphone listening devices that double as personal
assistants. Unless they are manually disabled, these devices are always on, waiting for a wake-up
direction like “Hey, Siri” or “Alexa”. Personal assistants have continuous access to us even in our
homes, a concern that has prompted technology analysts to wonder: “In a world in which these
personal assistants are always listening for our voices and recording our requests, have we given
up on any expectation of privacy within our own homes?” (Edwards, 2017: 28).
While the individuals I interviewed did not explicitly report that they were afraid of new
technologies, several expressed their concerns related to a growing lack of privacy with the use of
new technologies. Sean, a 53-year-old business owner from Seattle, said, “I am concerned that the
new smartphones can trace your movements, and that intrusion into my privacy terrifies me”. In
response, Sean and his wife have chosen to leave their smartphones on the kitchen counter rather
than bringing them into their bedroom at night. Sean explained that privacy concerns kept him
from purchasing a personal assistant like Amazon’s Echo for his home:
I may seem old school, but I consider my home my castle, and I don’t take my family’s
privacy for granted. We can live without a smart utility meter, and we can turn off the
lights by ourselves. The thought of anyone being able to watch my children playing or
even sleeping, makes me crazy.
Sean said that after reading several news articles about businesses getting hacked that he worried
about the vulnerability of his company’s online records. In response, he said that he had contracted
with a privacy security firm to monitor both his personal and business bank accounts.
John, a 30-year-old technology analyst working in Washington, DC, also expressed privacy
concerns, particularly as they related to online banking.
Most people take it for granted that when they access their bank accounts on their
smartphones that the information is safe and protected. This is my area, and let me tell
you, that’s crazy. I don’t trust mobile banking, and I don’t recommend using mobile
banking. It’s not all that safe. There are too many ways that your information can be
compromised. If you have to bank online, use your computer to access your accounts.
Trust me, you’ll be better off in the long run.
419
Bonnie Brennen
Several other interviewees, who described themselves as “tech savvy”, echoed John’s privacy
concerns. For example, Meg, a 38-year-old tech manager from San Francisco, also addressed the
risks associated with disclosing personal information online:
I work in a research lab, and in that environment, I have learned a lot about techniques
and technologies that are available for getting information off of the internet. This is
something that we talk about a lot. I have learned about different capabilities and about
the type of access that some people in our lab have. I think it is because of working in
this environment that I really started to think about what that means for me as a person.
And in my work environment I come into contact with lots of really brilliant engineers
who know a lot about security-related issues, and I’ve seen how some of them are
extremely skeptical about putting certain types of information of theirs online. I think
just realizing that these people who know a tremendous amount about what’s possible
and seeing them so skeptical and concerned that I realized online privacy is something
I should be more aware of.
Developing identity, asserting dominance

Throughout my interviews, the notion of opting in or opting out to develop a person’s identity
or to assert an individual’s independence, authority, and dominance over new technologies was a
continuous theme. For example, vinyl users described their rejection of digital music technologies
as an integral part of their identities. It was a significant way that they defined themselves, and
several interviewees said that their use of vinyl provided a “shorthand representation” of who
they were. As Alexis, a 24-year-old barista from Portland, Oregon, noted:
With vinyl I am very aware of the fact that it takes effort putting it on the turntable,
placing the needle on it. I feel a very strong physical and tactile sense with it that I really
enjoy. I love the fact that halfway through I have to flip it. It’s something I do because
that’s who I am. I collect vinyl – that’s how I identify myself.
The choice not to engage with social media was another way interviewees asserted their
agency over digital media. Although Bobkowski and Smith (2013) suggested that social media
“non-adopters” were less social, had fewer friends, and had “few meaningful friendships”, (ibid.:
777), Andrea, a designer at a tech company in Boston who is in her thirties, chose not to engage
with Facebook or Twitter because she felt that social media encouraged superficial relationships.
I started to see that for a lot of my peers there were certain types of relationships that
they would only maintain through Facebook. So there were all of these acquaintances
that they would be connected to, they would be friends with on Facebook, and
maybe they would never have any interaction with those people in person or even on
electronic media in any way. It seemed to me to be more about the idea of maintain-
ing a connection with these people as opposed to really getting anything meaningful
or giving anything meaningful to the relationship.
First exposed to a precursor of Facebook called “Plans” as an undergraduate, Andrea felt her
initial “resistance to Plans when I was in college was that it was a huge time suck”. She was
worried that if she signed up for Plans that she would spend too much time on it, and she didn’t
want to “waste that much time”. However, as Facebook took off, Andrea’s rationale for rejecting
social media evolved because of shallow relationships that Facebook and Twitter encouraged. For
420
Andrea, posting pictures of food and clothing wasn’t her idea of friendship, and she felt it was
important for people to put effort into their relationships:
I liked the idea that if you and I were going to maintain a relationship, I was going
to have to put effort in and you would have to put effort in and we both knew that,
and that was somehow reinforcement that the connection was important to both of
us. I liked that idea, and with Facebook it kind of undoes all of that. There’s no effort
required to maintain a friendship. I was just uncomfortable with that.
In the last few years, Andrea said she has become increasingly concerned with the amount
of personal information that is available on social media and felt that her decision not to engage
with social media helped her to retain control over the access to her personal information. Like
others I interviewed, throughout her interview Andrea spoke of her choice to reject social media
as a way of asserting her independence over the technology and developing her identity as an
independent thinker. While she understood that her rejection of social media might have nega-
tive ramifications for her career, she felt that as an individual she had to “create and manage her
personal image and identity”.
Balancing new media usage was an important aspect of interviewees asserting their domi-
nance over technology. For example, Destiny, a 25-year-old marketing manager from Philadelphia,
reported being immersed in social media and tied to her cell phone and other digital technology
at work. However, she said that she was careful to balance her new media use during her free time.
When my friends and I get together for dinner and drinks we always put our phones
away so that we can spend time together without distractions. We have a rule that the
first one who grabs their phone has to pay.
As a digital native, Destiny felt that she had no trouble managing her use of technology, and she
felt that since she had to use it at work all day that she preferred to be screen-free most evenings.
She suggested that her attitude toward new technology grew out of her parents’ efforts to balance
the family’s use of technology while they were growing up, and she noted that since she has had
access to digital media her entire life that she felt it was easier for her to opt out when she wanted
a break.
Similarly, Ruth, an 81-year-old from Southern California who has embraced many new tech-
nologies, refuses to text. She stated that she owned an “old-fashioned” flip phone so that she had
an excuse not to text, and during our conversation she explained her reasoning:
I want to see their faces when I talk with my family and friends. If I can’t be with them
in person, I want to hear their voices, connect with them, not merely see the top of their
heads and share texts with them. I want to know what they are thinking, what they are
feeling, if they are doing well, if they need anything from me. You can’t get any real
conversation going through texts. You just share superficial silly comments.
Recounting a recent trip to Chicago where everyone on the airplane seemed to be immersed
in technology, she observed: “The airplane was so quiet. No one was talking. Everyone was
looking at their phones, their iPads, their Kindles or their laptops. No one was interacting and
no one was communicating. It’s sad.” Ruth saw her refusal to text a way of being true to herself
and asserting her authority over technology. She explained that while she enjoyed keeping up
with friends on Facebook and email, she insisted on maintaining face-to-face communication
whenever possible with her family.
421
Bonnie Brennen
A status symbol
In addition to creating one’s identity and asserting dominance over technology, the choices
that people make regarding their use or nonuse of new technologies are also evidence of
using technology as a status symbol. Some people opt out because they have the money and/
or power to do so. They can hire people to communicate their views with others. This is a
strategy that has been used effectively in the U.S. political realm. In 2015, when Secretary of
State Hillary Clinton’s use of a private email account and server became a partisan controversy,
several politicians admitted that they did not use email. For example, Senator John McCain told
MSNBC’s Andrea Mitchell that he had “opted out of using email altogether” because he was
afraid that he might regret sending something or that his emails might be taken out of context,
while Senator Lindsey Graham admitted on NBC’s Meet the Press that he had never sent an
email. Apart from concerns that congressional leaders who were shaping new technology policy
weren’t using those technologies, their actions have been seen as “the ultimate status symbol –
second only to sending someone to fetch your lunch” (Parker, 2015).
Framed from a cultural materialist perspective, this research reinforced Raymond Williams’
understanding of the role of human agency in the development and usage of new technologies.
Rather than seeing individuals who opted out of new media as unfortunate have-nots who were
left behind in contemporary society, or as media resisters fighting social oppression created by
new technologies, this research focused on the intentions of individuals and the specific reasons
why they chose to engage with some technologies and reject others. Apart from a consistent
preference to read books on paper, my research indicates that there are many reasons why indi-
viduals use some technologies and forgo others. Some interviewees opted out because of their
ethical, political, religious, or cultural beliefs. Others found new technologies simple diversions
that did not meet their needs or interests, and by opting out they took back their time and
refocused their lives on family and community concerns. Some citizens worried about issues of
privacy, rejecting new technologies out of safety concerns, while others asserted their indepen-
dence, authority, and dominance over technologies while constructing their individual identities
through the choices they made regarding their use of new technologies.
Finally, this research considers a quote from the St. Louis poet and filmmaker PrinceEA
(2016):
I’m so tired of performing in the pageantry of vanity and conforming to this accepted
form of digital insanity. Call me crazy but, I imagine a world where we smile when we
have low batteries, cause that will mean we’ll be one bar closer to humanity.
In his inspirational videos and talks, PrinceEA encourages people to think about their personal
relationships and take control of new technology. Calling Facebook an “anti-social network” that
encourages superficial interactions rather than interpersonal relationships, he suggests that our reli-
ance on technology has made us “more selfish and separate”, and he counsels people not to let
technology exert power over them. While the full impact of his message is yet to be known, since
2015, PrinceEA’s videos have amassed more than 500 million views on Facebook and YouTube.
Ultimately, his growing influence may help to remind people that they have the ability to exert con-
trol over digital technology and that it’s possible and may even be desirable to unplug when needed.
Further reading
Raymond Williams’s classic text Television, Technology and Cultural Form (1974) provides a nuanced
understanding of the prevailing views of technology as well as an excellent discussion of the main
422
components of cultural materialism. Although Television was first published more than 40 years
ago, it remains a valuable resource for understanding the continued influence of technologi-
cal determinism on contemporary culture. Colin Sparks’s (2013) “What is the Digital Divide
and Why Is It Important?” is particularly helpful in ascertaining the continued relevance of the
notion of the digital divide, while Louise Woodstock’s (2014) “Media Resistance: Opportuni-
ties for Practice Theory and New Media Research” offers key insights into the media resistance
movement.
Note
1 When requested, interviewees’ names were changed to protect their privacy. In addition, the interview-
ees have approved all demographic information included.
References
Bilton, N. (2013, May 13) “Even the tech elites leave gadgets behind.” The New York Times. Retrieved from
http://bits/blogs.nytimes.com/2013/05/12/disruptions-even-the-tech-elites-leave-gadgets-behind/
Bobkowski, P. and Smith, J. (2013) “Social media divide: Characteristics of emerging adults who do not use
social network websites.” Media, Culture & Society, 35(6), 771–781.
Borchers, C. (2015, January 15) “Need to unplug? There’s an app for that.” The Boston Globe. Retrieved
from www.bostonglobe.com/business/2015/01/12/need-unplug-there-app-for-that/
Cooper, A. (2017, April 9) “What is ‘brainhacking’? Tech insiders on why you should care.” 60 Minutes
Transcript. Retrieved from www.cbsnews.com/news/brain-hacking-tech-insiders-60-minutes/
Douthat, R. (2017, March 11) “Resist the internet.” New York Times. Retrieved from www.nytimes.
com/2017/03/11/opinion/sunday/resist-the-internet.html?smprod=nytcore-iphone&smid=nytcore-
iphone-share
Dowd,M. (2017) “Elon Musk’s billion-dollar crusade to stop the A.I. apocalypse.”Vanity Fair Magazine. Retrieved
from www.vanityfair.com/news/2017/03/elon-musk-billion-dollar-crusade-to-stop-ai-space-x
Edwards, H. S. (2017, May 15) “Alexa takes the stand: Listening devices raise privacy issues.” Time, 189(18),
28–29.
Eisenstein, E. (1997) “From the printed word to the moving image.” Social Research, 64(3), 1049–1066.
Griffin, A. (2014, December 2) “Stephen Hawking: AI could be the end of humanity.” The Independent.
Retrieved from www.independent.co.uk/news/science/stephen-hawking-ai-could-be-the-end-of-
humanity-9898320.html
Karppi, T. (2011) “Digital suicide and the bio-politics of leaving Facebook.” Transformations, 20, 1–13.
Kraybill, D. (2001) The Riddle of Amish Culture (2nd ed.). Baltimore, MD: Johns Hopkins University Press.
Mack, R. L. (2001) The Digital Divide: Standing at the Intersection of Race and Technology. Durham, NC: Caro-
lina Academic Press.
Morrison, S. and Gomez, R. (2014) “Pushback: Expressions of resistance to the ‘evertime’ of constant online
connectivity.” First Monday, 19(8), 1–15.
Parker, A. (2015, March 11) “In era of email, some senators do just fine without it.” The New York Times, A1.
Portwood-Stacer, L. (2013) “Media refusal and conspicuous non-consumption: The performance and polit-
ical dimensions of Facebook abstention.” New Media & Society, 15(7), 1041–1057.
PrinceEa. (2016) Anti-Social Network. Retrieved from www.princeea.com
Rauch, J. (2014) “Constructive rituals of demediatization: Spiritual, corporeal and mixed metaphors in
popular discourse about unplugging.” Explorations in Media Ecology, 13(3&4), 237–252.
Sparks, C. (2013) “What is the ‘digital divide’ and why is it important?” Javnost – The Public, 20(2), 27–46.
Stieger, S., Burger, C., Bohn, M. and Voracek, M. (2013) “Who commits virtual identity suicide? Differ-
ences in privacy concerns, Internet addiction, and personality between Facebook users and quitters.”
Cyberpsychology, Behavior, and Social Networking, 16(9), 629–634.
Sullivan, A. (2016, September 19) “My distraction sickness – and yours.” New York Magazine. Retrieved from
http://nymag.com/selectall/2016/09/andrew-sullivan-technology
Victor, D. (2016, September 2) “No the internet has not killed the printed book. Most people still prefer them.”
The New York Times. Retrieved from www.nytimes.com/2016/09/03/business/no-the-internet-has-
not-killed-the-printed-book-most-people-still-prefer-them/
423
Bonnie Brennen
Williams, R. (1974[1992]) Television: Technology and Cultural Form. Hanover: Wesleyan University Press.
Williams, R. (1977[1988]) Marxism and Literature. Oxford: Oxford University Press.
Woodstock, L. (2014) “Media resistance: Opportunities for practice theory and new media research.” Inter-
national Journal of Communication, 8, 1983–2001.
Yeshiva Elementary School. (2013) Parent Handbook. 5115 West Keefe Ave., Milwaukee, WI 53216.
Zickuhr, K. (2013, September 25) “Who’s not online and why.” Pew Research Center Report. Retrieved from
http://pewinternet.org/Reports/2013/Non-internet-users.aspx
424
32
SILENCING THE FEMALE VOICE
The cyber abuse of women on the internet
For a woman journalist in 2017, working on Twitter entails opening oneself to attacks such as:
“I hope you get raped” (Just Not Sports, 2016).“You need to be hit in the head with
a hockey puck and killed” (Just Not Sports, 2016).“You are clearly retarded, i hope
someone shoots then rapes you”.
(Hess, 2014)
The internet is touted as a democratic space in which nationality, class, race, gender, and sexuality
are rendered neutral. However, receiving digital media threats of violence, rape, and murder are
daily occurrences for female journalists. Internet harassment of women marginalizes their profes-
sional presence online, impinges on their freedom of communication, and, in an echo of outdated
and retrograde domestic violence attitudes, is minimalized and dismissed by law enforcement
and media publishers. Stalking, bullying, and intimidation that would not be tolerated in brick-
and-mortar workplaces are commonplace in comments, emails, tweets, and social media related
to the online work of female journalists. Studies in this emerging field point to an ugly truth:
the anonymity and ubiquity of the internet works to shelter and protect harassers and to allow
the cyber sexual harassment of women and marginalized groups, such as LGBTQ (lesbian, gay,
bisexual, transgender, queer/questioning) persons.
This chapter gathers key literature in the growing area of research into cyber sexual harass-
ment, points to some promising directions, draws parallels between cultural responses to it and
to domestic violence, and recounts key instances of harassment, stalking, and threats of physical
violence online. This reveals that the digital world is the newest location for age-old male behav-
iors of intimidation and violence against women. A discussion follows about the significance of
online gender violence and recommendations for addressing and limiting it.
The following sections explore existing research on cyber gender harassment, how digital
space is conceptualized, and how domestic violence in the physical world is characterized.
The gendered nature of the internet

From its inception, the internet has had a reputation for misogyny. In practice, individu-
als engaging in cyber gender harassment effectively police online areas and attempt to render
425
them male-only spaces, disciplining women for transgressing into male territory. During the
mid-1990s, women who resisted demeaning comments from men on internet ‘list-servs’ and
recreational channels were blamed for causing the comments by their very presence (Herring,
1996). The internet evolved with built-in gender disparity, using language and interaction styles
that were familiar and advantageous to men and downright hostile to women (McCormick
and Leonard, 1996). Characteristics of digital spaces that allow and support harassing behavior
include anonymity (McGarth and Casey, 2002), lack of eye contact (Lapidot-Lefler and Barak,
2012), and an easy escape route (Barak, 2005), which work together to create user disinhibition
(Barak and Fisher, 2002). Online users employ bolder sexual behaviors online than they would
do in person (Cooper et al., 2002). And the plethora of pornographic sites contributes to a sexu-
alized atmosphere of male space (Döring, 2000). In some ways, the internet is the ‘Wild, Wild
West’, operating with limited legal sanctions (or at least the belief that there are few sanctions)
and an absence of authority figures (Barak, 2005). In the physical world, sexual harassment is
addressed through education of victims and harassers, organizational cultural changes, and laws
(Paludi and Paludi, 2003), but online, free speech, libel, and slander laws as they pertain to the
internet are unclear and enforcing them problematic (Hiller and Cohen, 2002). Legal challenges
arise from the fragmented nature of criminal and civil law, out-of-date laws that don’t consider
the internet, and the failure of law to address systemic gender violence (Henry and Powell, 2015).
In short, a harasser is free to be anti-social and aggressive without fear of being discovered or
punished, existing in a community of others behaving equally badly and egging each other on
(DeKeseredy and Olsson, 2011). The design, policies, and governance structures of certain sites
attract and shelter anti-feminist activity (Massanari, 2017), such as #Gamergate (a campaign of
harassment of women in the gaming industry) and ‘the Fappening’ (illegally distributed celebrity
nudes, such as Jennifer Lawrence) at Reddit (Massanari, 2017: 330).
Nomenclature
As the phenomena has evolved and scholarly attention to online violence against women
has increased, the nomenclature for it is in flux. Karla Mantilla draws distinctions between
cyberharassment, cyberbullying, cyberstalking, revenge pornography, blackmail videos, and the
like (2015); Henry and Powell refer to these and other harassments as technology-facilitated
sexual violence (2015). ‘Gendertrolling’ is a term well-explicated by Mantilla (2015) as brought
about by a woman stating her opinion online, resulting in attacks that are graphic, sexualized, and
gender-based, often including credible rape and death threats and physical stalking. Attacks are
intense, occur frequently, continue for months or years, cross online platforms, and may involve
many trolls working together in ways that are “exponentially more vicious, virulent, aggressive,
threatening, pervasive, and enduring than generic trolling” (Mantilla, 2015: 11). “Cyber gender
harassment” is Danielle Citron’s term for harassment invoking stereotypical gender roles,
suggesting female journalists return to the kitchen or go home and have babies, threatening rape
or murder, and particularly demeaning women of color and lesbians (2009). For the purposes of
this chapter, all variations are included in and referred to as ‘cyber gender harassment’.
Prevalence
Cyber gender harassment is a worldwide phenomenon – the internet has no national boundaries.
The harassment of women journalists is insidious, says a report from Australia: 41% of in-house
journalists and 18% of freelancers have been made the target of obscene, violent, and harassing
tweets, comments, and postings (Uther, 2015). Philips and Morrissey (2004) estimate one-third
of female internet users have experienced cyber gender harassment; Barak (2005) estimates
426
Silencing the female voice
40%. The Organization for Security and Cooperation in Europe calls cyber gender harassment
an attack on fundamental freedoms that “limits the sharing and receiving of information and
constitutes a serious obstacle to online plurality, thereby restricting the freedoms and rights of
society as a whole” (Mijatović, 2016: 2) and is “the most prevalent human rights violation in the
world” (Antonijevic, 2016: 9). It is prevalent but underreported due to shame, embarrassment,
and the tendency of law enforcement to dismiss it (Citron, 2009). Working to Halt Online Abuse
has tracked online harassment at the rate of 50–75 cases a week since 1997, though statistics are
limited to self-reports and estimated to be low. Most victims are single, white women between 18
and 40 years of age; most harassers are male, and about one-fourth of cases escalate into physical
violence (WHOA, 2015). A Pew Research Center study found that 75% of adult internet users
have witnessed someone being harassed online, 40% have been harassed themselves, and 18%
of those (including a disproportionate percentage of women under the age of 24) have been
digitally stalked, sexually harassed, and physically threatened (Duggan, 2014).
Marginalized populations like the LGBTQ community are also targeted (Barnes, 2001). The first
statistics about online harassment of LGBTQ youth were gathered by the Gay, Lesbian & Straight
Education Network in 2010–2011. LGBTQ youth are bullied or harassed online almost three times
as often as are straight youth and are twice as likely to be harassed via text. They felt equally unsafe
whether online, at school, or traveling to school. One in four LGBTQ young people reported being
bullied online and one in five via text during the past year. Harassed LGBTQ youth report higher
rates of depression and of negative self-esteem. When online harassment was combined with harass-
ment in the real world, LGBTQ youth reported lower self-esteem, lower grade point averages, and
higher levels of depression. Interestingly, LGBTQ youth also describe the internet as a site for peer
support and educational information about sexuality and gender identification and find informa-
tion there they cannot always readily access in person. LGBTQ youth say they were twice as likely
as straight youth to have used the internet for health and medical information and more than twice
as likely to have at least one close online friend (GLSEN, 2013).
As this discussion turns now to cyber gender harassment and journalists, we realize harassment
also occurs to non-journalists and acknowledge the intersectionality of identities that encom-
passes online women journalists and other marginalized groups to which they may also belong,
such as persons of color, non-Christians, and members of the LGBTQ community. Future
research in this field may begin to discern between identities and broaden the scope of knowl-
edge about online harassment.
Consequences of being harassed

After being harassed, women shut down blogs, close Twitter accounts, and leave Facebook
(Citron, 2009); harassment causes women to avoid chat rooms (Fallows, 2005). Women fear
that online threats of rape and murder will be carried out in real life (Biber et al., 2002). This
self-censorship silences women (Mijatović, 2016). One type of harassment, ‘doxxing’, publishes
personal information like home addresses and phone numbers. Women’s careers are sabotaged by
so many negative comments that potential employers avoid hiring ‘problem’ employees (Citron,
1999). Verbal and cyber abuse addresses professional women as sexual objects (Abrams, 1998),
compromising their reputations as professionals. Some supervisors tell harassed women that it’s
their private problem to handle (Hagen, 2015). To avoid attacks, some women use gender-neutral
names or assume male identities (Yoshino, 2006). Online accounts with feminine usernames receive
100 sexually explicit or threatening messages a day; accounts with masculine usernames
receive 3.7 (Meyer and Cukier, 2006). Like domestic violence and physical workplace sexual
harassment before it, cyber gender harassment is trivialized by law enforcement (despite the
inclusion of rape and death threats and ‘Photoshopped’ images of women being strangled and
427
tortured) and dismissed as ‘locker room talk’, as just the way the internet works, and as trolls
just having fun – all clear signals that violence toward women is acceptable (Citron, 2009.). As
more daily functions like socializing, purchasing, reading the news, applying for jobs, and talking
with colleagues move to the internet, the civil rights and gender politics of cyberspace become
highly significant (ibid.)
How cyberspace is conceptualized

Optimistic assumptions about the internet characterize it as a Habermasian public sphere, a
democratic space open to all where freedom of speech reigns and where virtual spaces might
sidestep the sexism and racism that face-to-face interactions allow (Oksman and Turtainen, 2004;
Beck and Beck-Gernsheim, 2002). But such assumptions may naïvely background pervasive,
unequal gender relations and overlook the internet’s capacity for “technology-facilitated sexual
violence and harassment”, argue Henry and Powell (2015: 762). In digital space, criminal and
harmful behaviors can occur in both familiar and new ways (Lash, 1994; Adkins, 1999). The
possibility for increasing, rather than decreasing, isolation and inequity exists (Castells, 1997) due
to the gender gap in technology and inequity of access (Youngs, 2005; Ono and Zavodny, 2003).
The nature of digital spaces has a bearing on how cyber gender harassment is addressed. If the
internet is a purely virtual reality, as some scholars argue (Brown, 2006; Youngs, 2005), then the
harms sustained there might be considered to be virtual, as well, and not “real” (Williams, 2006).
But if the internet is, as video game designer and subject of “Gamergate” online attacks Brianna
Wu argues, “the public squares of 2015, where we make professional contacts, hang out with
our friends and make meaningful connections” (Wu, 2015: 48), then the harms incurred there
are more than ephemeral. Even if the internet is not a geospatial place, it is a socio-spatial one
in which people relate, work, and play and one upon which social interactions are increasingly
dependent (Youngs, 2005). Victims of cyber gender harassment are often told to ignore the
harassment since it is “just words” – however, if those words become acted-upon rape threats, or
if nude photos are published or doxxing is part of the harassment, those so-called virtual harms
become real-life harms (Guardian, 2016).
Conceptions of digital space raise problematic questions about whether or not a realm that
does not host bodies can be a site where significant harm can be enacted upon bodies, or at
least upon real people living real lives (Henry and Powell, 2015). Socio-spatial locations can host
techno-social harms that impact on daily lives and that express very real patriarchal and power
relationships that exist in both cyberspace and the physical world (Brown, 2006). Henry and
Powell argue that digital harms are distinct harms, though they find only rare examples of such
harms being taken as seriously as criminal harms and point to the relative paucity of scholar-
ship of gendered harassment (2015). Technology-facilitated sexual violence needs to be further
researched in order to construct a useful foundation for future public policy regarding prevention
and punishment.
Domestic violence
In the physical world, most domestic violence victims are women, most perpetrators men, and
occurrence is underreported. One in four women are violently attacked in their lifetimes, and
29% of all women are raped, stalked or physically assaulted (Black et al., 2011). Stalking creates
long-ranging and costly health consequences, including post-traumatic stress disorder, headaches,
ulcers, and poor mental health (Black et al., 2011). Stalking is likely to lead to assault and murder,
and 76% of women killed violently were stalked by their murderers (McFarlane et al., 1999).
Domestic violence has long been cast as the fault of the woman, not the man; we ask “Why does
428
she stay?” rather than “Why does he hit?” (Nettleton, 2011). Its social and economic impact has
been minimized for more than 200 years, and the history of trivializing physical and emotional
damage done to women is even longer (West, 1997). Only relatively recently has research begun
to consider domestic violence as more than one woman’s failure to find a man who won’t hit her
and move to more broad examinations of the social forces that make gender violence possible.
Weak social and legal support contribute (Coates and Wade, 2007), as does no prominent public
attitude that domestic abuse will not be tolerated (Bostock and Plumpton, 2009). Increasingly,
domestic violence research points to cultural practices and structures that help naturalize male
violence and hold patriarchy in place and that foster implied acceptance of male violence toward
women (Bou-Franch and Blitvich, 2014).
Sexual harassment in the physical workplace wasn’t even named until the 1970s (West, 1997)
and was not linked to Title IX legislation by the U.S. Supreme Court until 1989. Since then, it
has been the subject of significant study (Fitzgerald et al., 1995), finding that harassment is about
power, not sex (Barak, 2005; Hoffspiegel, 2002); that most victims are women (Paludi and Paludi,
2003); and that the effects are devastating, including reduced work performance, lower produc-
tivity, psychological distress, and depression (Dansky and Kilpatrick, 1997), as well as nausea and
eating disorders (Harned and Fitzgerald, 2002). There is evidence of post-traumatic stress disor-
dered behavior, such as sleep disturbances, substance abuse, and thoughts of suicide (Avina and
O’Donohue, 2002; Davis et al., 2002). Martin positions male violence against women as a direct
result of inequities in social and cultural power, rooted in historical attitudes toward women and
formative of cultural structures that reinforce patriarchy (1976). Notably, there are two crimes
for which the victim is blamed and the attacker is nearly sidelined in responsibility: rape and
domestic violence (Cuklanz, 2000).
The ways in which journalism represents domestic violence offer insights into cultural and
social attitudes about gender and also assist in buttressing patriarchy and retrograde gender atti-
tudes (Caringella-MacDonald, 1998). Prushank finds that media are so complicit in natural-
izing patriarchy “that men find the domination and exploitation of women and other men to
be not only expected, but actually demanded” (2007: 161). Sociologist Nancy Berns surveyed
women’s (1999) and men’s (2001) magazine coverage of domestic violence, finding that women
were held responsible for the violence men do in ways that de-gender the problem and gender
the blame. The violence of men is repositioned as violence in general, with gender excised, but
female victims are blamed for victim-like behaviors that are particularly attached to their gender.
Ignoring the role of the attacker’s gender while focusing on the victim’s gender repositions the
issue of men attacking women as being a woman’s private problem, caused by her poor choices
(Berns, 1999). This assists in keeping patriarchy and male privilege out of discussions of domestic
violence (Berns, 2001). A 10-year study of 10 leading women’s and men’s magazines found that
magazines blame women for being with violent men and do not hold men responsible for being
violent, with men’s magazines treating domestic violence as a joke, using it as a source of humor
and mocking women’s concerns (Nettleton, 2011). Newspaper articles blame domestic violence
victims for their own deaths (Consalvo, 1998; Meyers, 1994). Women are counseled to create
escape plans, change their names, and move their children and jobs to avoid angry partners, and
men are not counseled on any ways in which they might temper rage or stop hitting women
(Nettleton, 2011). Such blaming tactics may serve to excuse or even encourage male violence
and contribute to an environment in which men feel free to behave violently (Lowney and Best,
1995; Maguigan, 1991).
As this review of the literature contributing to discussions of cyber gender harassment dem-
onstrates, research focused on the harassment of marginalized groups, such as women or LGBTQ
persons, is limited, but growing. Compelling evidence suggests that the internet is the newest
location for male intimidation of and violence against women, and unique harms are enabled
429
by new technologies. Legal boundaries about hate speech, slander, libel, and harassment online
are evolving with the technology, and scholarly investigation and argumentation about such
boundaries is to be expected in law, media ethics, free speech, journalism, and feminist media
studies journals, as well. It is promising that the literature on physical domestic violence points in
directions likely to be helpful for digital media scholars, and it is hopeful that the existing body of
work on patriarchy’s role in domestic violence provides a useful foundation for emerging cyber
gender harassment research.
Cyber gender harassment examples and effects

Several cases of cyber gender harassment in recent years provide specific information about
the nature and tenor of online harassment, the typical reactions of law enforcement (when it is
brought in), and the consequences for and effects of harassment for the women victims. A viral
YouTube video public service announcement, “#MoreThanMean – Women in Sports ‘Face’
Harassment”, shows men (not the harassers) reading tweets sent to sports reporters Sarah Spain
and Julie DiCaro (Just Not Sports, 2016). Spain was an accomplished high school and college
athlete and a radio and TV reporter and host in Chicago before joining ESPN in 2010; she has
written and spoken on air about sexual harassment she’s suffered on the job (Joseph, 2016). Julie
DiCaro was an attorney in Chicago before becoming a baseball blogger covering the Chicago
Cubs, a weekend radio host, and anchor and a columnist for CBSChicago.com; she wrote an
account of being raped for the Huffington Post in 2013 (DiCaro, 2013). Reading the ugly words
written by other men, the discomfort of these ‘regular guys’ is obvious. They squirm, swallow
hard, look at the floor, take deep breaths, apologize, and look off-camera for direction. They read:
“One of the players should beat you to death with their hockey stick like the whore you are”,
“This is why we don’t hire any females unless we need our cocks sucked or our food cooked”,
“fuck this dumb cunt”, “hopefully this skank Julie DiCaro is Bill Cosby’s next victim”, “I hope
your boyfriend beats you”, “I hope you get raped again”, and “you need to be hit in the head
with a hockey puck and killed” (Just Not Sports, 2016).
Another well-publicized case of cyber gender harassment attack is ‘#Gamergate’. Beginning
as an ex-boyfriend’s personal blog about video game developer Zoë Quinn, it evolved into a
campaign against feminism in the gaming industry and expanded to attack other women devel-
opers and designers and games journalists, using the hashtag #GamerGate in various social media
outlets and on websites. The organized, targeted attack was launched in August 2014, surfacing
on Reddit, 8chan, 4chan, and other platforms that allowed anonymous comments and postings.
Women were threatened with murder, rape, and exposure of personal information online. Game
developer Brianna Wu reported that people threatened her pets, called her phone while they
were masturbating, posted photographs of her friend’s children on a pedophile forum, and falsely
reported a friend for criminal activity that resulted in a SWAT team being sent to her home (a
cyberharassment technique called ‘swatting’). Wu wrote, “There is no free speech when speaking
about your experiences leads to death threats, doxing and having armed police sent to your house
[. . .] online spaces are not safe for women” (Wu, 2015: 48).
Journalist Caroline Criado-Perez campaigned for the Bank of England to add women to the
historical figures featured on banknotes. Then the death and rape threats began.
threats to mutilate my genitals, threats to slit my throat, to bomb my house, to pistol-

whip me and burn me alive. I was told I would have poles shoved up my vagina, dicks
shoved down my throat. I was told I would be begging to die, as a man would ejaculate
in my eyeballs.
(Criado-Perez, 2015: 13)
430
Since much of the attack imagery was to her mouth and throat, she concluded the attackers
had one goal: to get her to stop talking (ibid.). She did not. She retweeted her harassers’ tweets,
attracted media attention, reported to the police, and asked Twitter to step in. Twitter placed
responsibility on the police, who placed responsibility on Twitter (Hess, 2014). Eventually, Scot-
land Yard arrested three men, but the harassment continued (ibid.).
A senior contributor at The Federalist and columnist at The Forward, Bethany Mandel is a
conservative blogger who criticized Donald Trump during his presidential campaign. She also
often writes about Jewish issues. She has been deluged with harassing tweets, once so many so
quickly that they “had to have been sophisticated coordination, they arrived within minutes of
each other, a deluge” (Personal interview, 2017). Via Twitter and Facebook, she’s received anti-
Semitic messages and death threats, including a threat to behead her baby (ibid.). She reported
the harassment to her small-town police department, which “blew me off ” until she contacted
the Jewish Federation, which had connections to Homeland Security, which contacted her local
police. “Then the cops contacted me and took it seriously,” she says. “They did drive-bys for
a long time, and I was on their radar. We had a conversation about swatting, how they should
respond, so they were cognizant of it as being a thing that might be done to me” (ibid.). Eventu-
ally, Mandel bought a gun and wrote about it, alerting her harassers (Mandel, 2016). When the
Anti-Defamation League examined an upsurge in anti-Semitic hate language in social media,
Mandel and fellow journalists Julie Iofff and Dana Schwartz contributed their stories about
their cyber gender harassment. Mandel noticed that, among the larger group of journalists who
spoke with the ADL, men who had been harassed were powerful editors of their own publica-
tions, but women who had been harassed were freelancers, further down the publishing ladder.
“We were viewed as equally dangerous as those men, and that says something”, said Mandel.
“[Harassers] will push back against any woman who is a journalist, but they only targeted the
upper echelons of the men. They took the time to target us with the same velocity” (Personal
interview, 2017).
A Pacific Standard piece by Amanda Hess, who has also written for Slate, ESPN, and Wired, won
the 2014 American Society of Magazine Editors award in the public interest category for “Why
Women Aren’t Welcome on the Internet”. It detailed her own harassment: “Im [sic] looking
you up, and when I find you, im going to rape you and remove your head”, “I did 12 years for
‘manslaughter’, I killed a woman, like you, who decided to make fun of guys [sic] cocks”, “[I’ll
put you and another feminist] in a gimp mask and tied to each other 69 so the bitches can’t
talk or move and go round the world, any old port in a storm, any old hole” (2014). Hess also
reported examples of harassment visited on journalist colleagues: “you are clearly retarded, i hope
someone shoots then rapes you” . . . “i hope someone slits your throat and cums down your
gob” . . . “A BOMB HAS BEEN PLACED OUTSIDE YOUR HOME. IT WILL GO OFF AT
EXACTLY 10:47 PM ON A TIMER AND TRIGGER DESTROYING EVERYTHING” (ibid.,
errors in original). When Hess reported the harassment, a police officer asked, “What is Twitter?”
A study launched by the UK newspaper the Guardian (2016) examined 70 million comments
left on its news site since 2006. It found that the 10 most abused writers were women (four
white, four non-white, two gay, one Muslim, one Jew) and the other two were black men (one
gay). The 10 least-abused writers were all men. LGBTQ people and ethnic and religious minori-
ties were also targeted. Avalanches of harassment cascade across all social media: “To the person
targeted, it can feel like the perpetrator is everywhere: at home, in the office, on the bus, in the
street” (The Guardian, 2016).
Some women are targeted for expressing opinions about controversial matters like immigra-
tion or LGBTQ rights, but others are harassed for innocuous content. One food blogger posted a
pretzel recipe and received “I hope you choke on your own pretzels and die, you bitch” (Mantilla,
2015: 30).
431
Discussion
Although digital technology makes possible new ways of being violent toward women, online
harassment and domestic violence are similar subordinations. Their commonalities include:
• Women are held responsible for the violence that men do.
• Attacked women are patronized by law enforcement and others and told to ignore attacks
and dismiss threats.
• Proposed ‘solutions’ are the woman’s responsibility to carry out and are aimed at disciplining
and containing women rather than solving the problem. The chief way offered to ‘solve’
domestic violence is to raise money to fund shelters, which does nothing to halt violence.
Commonly suggested ways to ‘solve’ cyber gender violence include closing social media
accounts (yet this is the workplace for women journalists), changing their online names (the
professional brand of journalists), and treating harassment as harmless joking.
• Women under physical and cyber attacks are positioned as helpless victims, and male
attackers are positioned as able to escape with few consequences, reifying characterizations
of women as ineffectual and reinforcing a sense of male entitlement to freely harass women.
Certain conditions of new delivery modes of journalism – comments sections, Twitter feeds,
and Facebook interactions – offer harassers opportunity to elevate their sense of self-importance
by writing directly to ‘famous’ journalists and becoming published themselves. Reader comments
blur distinctions between those who produce media and those who consume it. The responses to
the message become the message; harassers become media producers, no matter how incompetent
or unintelligible, and commenters are held to no professional accountability to be accurate, fair,
civil, or nonviolent. Technological revolution is not necessary progressive. As the digital future of
journalism is imagined in terms of business models, organizational structures, and evolving tech-
nologies, considerations of how to correct inequities of identities and of social and cultural power
in representation, staffing, and ownership too often remain unaddressed. Destructive expressions of
power and subordination are currently – not permanently, hopefully – more readily published and
far less challenged in the digital realm than in the physical. Digital journalism promises brave new
possibilities but cannot afford to overlook the persistent and troubling issues of intimidation and
marginalization that have been problematic in free speech for the history of media. However, to
date, these critical issues are not front and center in most futuristic imaginings of journalism.
Like rape, cyber gender harassment is about power, not sex. Sexual imagery and language are
used to debase women and reify the patriarchal power of being able to attack, mock, shame, and
threaten with few consequences. Harassers are drawn to the ‘celebrity’ of seeing their harassment
published. The potential for setting a violent and misogynistic tone for a comment section offers a
kind of perverted type of power. A voyeuristic opportunity is offered to those who read but do not
participate in posting harassing comments. Audiences are invited to gaze on the sexually displayed
cyber bodies of female journalists and witness their sexual humiliation. Sexualizing and weakening
professional women moves them out of position as powerful, independent, self-actualized human
beings and places them where male partners are essential, in pairs of sexual or violent activity.
Narratives of the attacked female journalist as being unable to interest law enforcement and
internet platform administrators in her plight surely encourages additional online violence. Male
harassers feel free to threaten with impunity. Ironically, the victims’ stories of frustrating and
failed attempts to get meaningful responses from law enforcement or internet platforms bear wit-
ness to how readily women can be attacked online, serving as a teaching manual for how to harass
women and get away with it, and grant permission to continue harassing unabated.
432
A critical absence in the conversation

Although domestic violence is a “product of gender inequality and the lesser status of women
compared to men” (Jewkes, 2002: 253), this cause is rarely pointed out in discussions of how to
limit or contain cyber gender harassment. Decades have been spent blaming women for being
hit and killed by men, but little time has been spent addressing the root causes of male violence
and patriarchy. With cyber gender harassment, there is an opportunity to avoid repeating this
same mistake in logic and justice. Yet the word ‘patriarchy’ only rarely surfaces in discussions of
gendered online violence. Following Stuart Hall, absences in representation can reveal prevailing
and resistant cultural attitudes and assumptions (Hall, 1992). Looking for what has been rendered
invisible, listening for what has been silenced, can create ‘aha!’ moments of insight and clarity.
It is more than telling – it is a klaxon horn sounding – that patriarchy and the imposition of
male power remain under-discussed in issues of cyber gender harassment. Patriarchy’s role in
creating a culture that values and allows male violence and which persistently positions and
repositions women as subservient deserves unapologetic illumination. When the conversation
shifts from terrified women journalists to stories of criminal male harassers, the light will have
been turned on.
Recommendations from victims and researchers

Research about cyber gender harassment is still young; recommendations are still limited but do
include insights from sociological, feminist, and media studies perspectives, along with anecdotal
experience of harassment victims.
Cultural and social

• Deliver messages of and practice zero tolerance.
• Teach children not to create hostile digital communication before complacent attitudes become
entrenched.
• Increase the numbers of women in the public sphere to make them commonplace rather
than threatening.
• Socialize boys differently, defining masculinity in ways other than dominance over women.
• Stop overlooking and trivializing harms to women.
Legal
• Train law enforcement in cyber practices, legalities, and threats.
• Track down IP locations, knock on doors, and order harassers to cease and desist.
• Impose significant consequences on harassers: fines, injunctions, and criminal convictions.
• Create a cyber civil rights agenda employing criminal, tort, and antidiscrimination laws.
• Treat cyberharassment as serious discrimination so more women report it and law enforcement
acts on it.
• Treat with equal seriousness online and offline crimes.
Government
• Avoid censorship but recognize cyber gender harassment as a direct attack on freedom of
expression.
• Commission the collection and analysis of data related to online harassment.
433
Online journalists
• Proactively moderate comment sections.
• Report harassment to internet platform administration and law enforcement.
Online users
• Speak directly against harassment to abusers, especially men-to-men.
• Initiate constructive comments on articles to set a reasonable tone.
Media organizations and websites

• Create corporate cultures of gender equality and zero tolerance of harassment.
• Work with journalist unions to provide support systems, psychosocial and legal assistance,
and industry-wide guidelines.
• Implement consistent, transparent procedures of moderating comments.
• Use pop-up messages and banners to communicate and enforce anti-harassment policies.
• Investigate harassing messages – particularly when delivered in orchestrated deluges –
identify IP addresses, and report them to proper authorities.
• Make terms of service, community guidelines, and enforcement consequences proportionate,
understandable, and easily available to all users.
• Compile data and statistics on online abuse to help facilitate comprehensive research.
Technology and the internet hold the promise of potential freedoms from traditional, physical-
world limitations of identity, but the real-world attitudes toward identity are carried into
cyberspace by the humans who use it. While a woman’s work might exist without a physical
body on the internet, the harms visited upon her there have very real consequences to intimidate,
injure, and silence her.
Further reading
Danielle Citron’s (2014) Hate Crimes in Cyberspace (and her entire body of work on this topic)
is authoritative and more expansive than this chapter can include. Azy Barak’s Psychologi-
cal Aspects of Cyberspace (2008) explores the psychological aspects of cyberspace. For a detailed
analysis of media characterizations of domestic violence, see my own “Domestic Violence in
Men’s and Women’s Magazines: Women are Guilty of Choosing the Wrong Man, Men Are
Not Guilty of Hitting Women” at http://epublications.marquette.edu/cgi/viewcontent.cgi?
article=1061&context=comm_fac.
References
Abrams, K. (1998) “The new jurisprudence of sexual harassment.” Cornell Law Review, 83(1997), 1169–1230.
Adkins, L. (1999) “Community and economy: A retraditionalization of gender?” Theory, Culture & Society,
16, 117–137.
Antonijevic, Z. (2016) “New challenges to freedom of expression: Countering online abuse of female jour-
nalists.” Organization for Security and Cooperation in Europe. Retrieved from www.osce.org/fom/220606
Avina, C. and O’Donohue, W. (2002) “Sexual harassment and PTSD: Is sexual harassment diagnosable
trauma?” Journal of Traumatic Stress, 15, 69–75.
Barak, A. (2005) “Sexual harassment on the internet.” Social Science Computer Review, 23(1), 77–92.
Barak, A. (2008) Psychological Aspects of Cyberspace. Cambridge: Cambridge University Press.
434
Barak, A. and Fisher, W. A. (2002) “The future of internet sexuality.” In A. Cooper (ed.), Sex and the Internet:
A Guidebook for Clinicians. New York, NY: Brunner-Routledge (pp. 267–280).
Barnes, S. B. (2001) Online Connections: Internet Interpersonal Relationships. Cresskill, NJ: Hampton.
Beck, U. and Beck-Gernsheim, E. (2002) Individualization: Institutionalized Individualism and Its Social and
Political Consequences. London: Sage Publications.
Berns, N. (1999) “My problem and how I solved it: Domestic violence in women’s magazines.” Sociological
Quarterly, 40(1), 85–108.
Berns, N. (2001) “Degendering the problem and gendering the blame: Political discourse on women’s
violence.” Gender and Society, 15(2), 262–281.
Biber, J., Doverspike, D., Baznik, D., Cober, A. and Ritter, B. A. (2002) “Sexual harassment in online com-
munications: Effects of gender and discourse medium.” Cyber Psychology & Behavior, 5(1), 33–42.
Black, M. C., Basile, K. C., Breiding, M. J., Smith, S. G., Walters, M. L., Merrick, M. T., Chen, J. and
Stevens, M. R. (2011) The National Intimate Partner and Sexual Violence Survey (NISVS): 2010 Summary
Report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control
and Prevention.
Bostock, J. and Plumpton, M. (2009) “Domestic violence against women: Understanding social processes
and women’s experiences.” Journal of Community & Applied Social Psychology, 19, 95–110.
Bou-Franch, P. and Garcés-Conejos Blitvich, P. (2014) “Gender ideology and social identity processes in
online language aggression against women.” Journal of Language Aggression and Conflict, 2(2), 226–248.
Brown, S. (2006) “Integration by way of the criminology of hybrids.” In G. Barak (ed.), Criminology: An
Integrated Approach. Plymouth: Rowman & Littlefield (pp. 118–121).
Caringella-MacDonald, S. (1998) “The relative visibility of rape cases in national popular magazines.” Vio-
lence Against Women, 4(1), 62–80.
Castells, M. (1997) The Power of Identity, the Information Age: Economy, Society, and Culture Volume II. Malden,
MA: Blackwell.
Citron, D. K. (2009) “Law’s expressive value in combating cyber gender harassment.” Michigan Law Review,
108, 373–415.
Citron, D. K. (2014) Hate Crimes in Cyberspace. Cambridge, MA: Harvard University Press.
Coates, L. and Wade, A. (2007) “Language and violence: Analysis of four discursive operations.” Journal of
Family Violence, 22(7), 511–522.
Consalvo, M. (1998) “Three shot dead in courthouse: Examining news coverage of domestic violence and
mail-order brides.” Women’s Studies in Communication, 21(2), 188–238.
Cooper, A., McLoughlin, I., Reich, P. and Kent-Ferraro, J. (2002) “Virtual sexuality in the workplace: A
wake-up call for clinicians, employers and employees.” In A. Cooper (ed.), Sex and the Internet: A Guide-
book for Clinicians. New York, NY: Brunner-Routledge (pp. 109–128).
Criado-Perez, C. (2015) “‘Women that talk too much need to get raped’: What men are really saying when
they abuse women online.” OSCE.org. Retrieved from www.osce.org/fom/220411?download=true
Cuklanz, L. (2000) Rape on Prime Time: Television, Masculinity, and Sexual Violence. Philadelphia, PA: University
of Pennsylvania Press.
Dansky, B. S. and Kilpatrick, D. G. (1997) “Effects of sexual harassment.” In W. O’Donohue (ed.), Sexual
Harassment: Theory, Research, and Treatment. Boston, MA: Allyn and Bacon (pp. 152–174).
Davis, K. E., Coker, A. L. and Sanderson, M. (2002) “Physical and mental health effects of being stalked for
men and women.” Violence and Victims, 17, 429–443.
DeKeseredy, W. S. and Olsson, P. (2011) “Adult pornography, male peer support, and violence against
women: The contribution of the ‘Dark Side’ of the internet.” In M. Vargas Martin, M. Garcia-Ruiz and
A. Edwards (eds.), Technology for Facilitating Humanity and Combating Social Deviations: Interdisciplinary
Perspectives. Hershey, PA: Information Science Reference (pp. 34–50).
DiCaro, J. (2013) “My astoundingly typical rape.” HuffPost. Retrieved from www.huffingtonpost.com/
julie-dicaro/my-rape_b_4429420.html
Döring, N. (2000) Feminist views of cybersex: Victimization, liberation and empowerment. CyberPsychology
and Behavior, 3, 863–884.
Duggan, M. (2014) “Online harassment: Summary of findings.” Pew Research Center. Retrieved from www.
pewinternet.org/2014/10/22/online-harassment/
Fallows, D. (2005) “How women and men use the internet.” Pew Research Center. Retrieved from www.
pewinternet.org/2005/12/28/how-women-and-men-use-the-internet/
Fitzgerald, L. F., Gelfand, M. and Drasgow, F. (1995) “Measuring sexual harassment: Theoretical and psycho-
metric advances.” Basic and Applied Social Psychology, 17, 425–445.
435
GLSEN. (2013) “Out online: The experiences of LGBT youth on the internet.” GLSEN. Retrieved from
www.glsen.org/sites/default/files/Out%20Online%20FINAL.pdf
The Guardian. (2016) “The dark side of Guardian comments.” The Guardian. Retrieved from www.the
guardian.com/technology/2016/apr/12/the-dark-side-of-guardian-comments
Hagen, A. L. (2015) “Shame, shock and speech injuries: Online harassment against journalists in Norway.”
Organization for Security and Cooperation in Europe. Retrieved from www.osce.org/fom/220606
Hall, S. (1992) “Race, culture and communications: Looking backward and forward at cultural studies.”
Rethinking Marxism, 5(1), 10–18.
Harned, M. S. and Fitzgerald, L. F. (2002) “Understanding a link between sexual harassment and eating
disorder symptoms: A mediational analysis.” Journal of Consulting and Clinical Psychology, 70, 1170–1181.
Henry, N. and Powell, A. (2015) “Embodied harms: Gender, shame, and technology-facilitated sexual vio-
lence.” Violence Against Women, 21(6), 758–779.
Herring, S. (1996) “Posting a different voice: Gender and ethics in computer-mediated communication.” In
C. Ess (ed.), Philosophical Perspectives on Computer-Mediated Communication. Albany, NY: State University
of New York Press (pp. 115–146).
Hess, A. (2014) “Why women aren’t welcome on the internet.” Pacific Standard. Retrieved from www.
psmag.com/health-and-behavior/women-arent-welcome-internet-72170
Hiller, J. and Cohen, R. (2002) Internet Law and Policy. Upper Saddle River, NJ: Prentice Hall.
Hoffspiegel, L. (2002) “Abuse of power: Sexual misconduct in the legal workplace.” Sexual Addiction and
Compulsivity, 9, 113–126.
Jewkes, R. K. (2002) “Intimate partner violence: Causes and prevention.” Lancet, 359(9315), 1423–1429.
Joseph, A. (2016) “ESPN’s Sarah Spain details the jarring harassment she faced during a job interview.” USA
Today. Retrieved from //ftw.usatoday.com/2016/12/espn-sarah-spain-harassment-job-interview-podcast
Just Not Sports. (2016) “#MoreThanMean – Women in sports ‘face’ harassment.” YouTube. Retrieved from
www.youtube.com/watch?v=9tU-D-m2JY8
Lapidot-Lefler, N. and Barak, A. (2012) “Effects of anonymity, invisibility and lack of eye- contact on toxic
online disinhibition.” Computers in Human Behavior, 28, 434–443.
Lash, S. (1994) “Reflexivity and its doubles: Structure, aesthetics, community.” In U. Beck, A. Giddens and S.
Lash (eds.), Reflexive Modernization: Politics, Tradition, and Aesthetics in the Modern Social Order. Cambridge:
Polity Press (pp. 110–173).
Lowney, K. and Best, J. (1995) “Stalking strangers and lovers: Changing media typifications of a new crime
problem.” In J. Best (ed.), Images of Issues: Typifying Contemporary Social Problems. New York, NY: Walter
de Gruyter (pp. 33–35).
Maguigan, H. (1991) “Battered women and self-defense: Myths and misconceptions in current reform
proposals.” University of Pennsylvania Law Review, 140, 379–486.
Mandel, B. (2016) “My Trump tweets earned me so many anti-Semitic haters that I bought a gun.” The
Forward. Retrieved from http://forward.com/opinion/336159/my-trump-tweets-earned-me-so-many-
anti-semitic-haters-that-i-bought-a-gun/
Mandel, B. (2017, May 21) Personal Interview With Author by Phone.
Mantilla, K. (2015) Gendertrolling: How Misogyny Went Viral. Santa Barbara, CA: Praeger.
Martin, D. (1976) Battered Wives. Volcano, CA: Volcano Press.
Massanari, A. (2017) “#Gamergate and the Fappening: How Reddit’s algorithm, governance, and culture
support toxic technocultures.” New Media & Society, 19(3), 329–346.
McCormick, N. and Leonard, J. (1996) “Gender and sexuality in the cyberspace frontier.” Women & Therapy,
19(4), 109–119.
McFarlane, J., Campbell, J., Wilt, S., Sachs, C., Ulrich, Y. and Xiao, X. (1999) “Stalking and intimate partner
femicide.” Homicide Studies, 3(4), 300–316.
McGarth, M. G. and Casey, E. (2002) “Forensic psychiatry and the internet: Practical perspectives on sexual
predators and obsessional harassers in cyberspace.” Journal of the American Academy of Psychiatry and the
Law, 30, 81–94.
Meyer, R. and Cukier, M. (2006) “Assessing the attack threat due to IRC channels.” International Conference on
Dependable Systems and Networks, 467. Retrieved from www.enre.umd.edu/content/rmeyer-assessing.pdf
Meyers, M. (1994) “News of battering.” Journal of Communications, 44(2), 47–63.
Mijatović, D. (2016) “New challenges to freedom of expression: Countering online abuse of female jour-
nalists.” Organization for Security and Cooperation in Europe. Retrieved from www.osce.org/fom/220606
Nettleton, P. (2011) “Domestic violence in men’s and women’s magazines: Women are guilty of choosing
the wrong man, men are not guilty of hitting women. Women’s Studies in Communication, 34(2), 139–160.
436
Oksman, V. and Turtainen, J. (2004) “Mobile communication as a social stage: Meanings of mobile com-
munication in everyday life among teenagers in Finland.” New Media & Society, 6, 319–339.
Ono, H. and Zavodny, M. (2003) “Gender and the internet.” Social Science Quarterly, 84, 111–121.
Paludi, M. and Paludi, C. (2003) Academic and Workplace Sexual Harassment: A Handbook of Social Science, Legal,
Cultural, and Management Perspectives. Westport, CT: Praeger.
Philips, F. and Morrissey, G. (2004) “Cyberstalking and cyberpredators: A threat to safe sexuality on the
internet.” Convergence: International Journal of Research into New Media Technologies, 10, 66–79.
Prushank, D. (2007) “Masculinities in teen magazines: The good, the bad, and the ugly.” Journal of Men’s
Studies, 15(2), 160–177.
Uther, B. (2015) “Women in media report 2016.” iSentia and MEAA.org. Retrieved from http://women
inmedia.net
West, R. (1997) Caring for Justice. New York, NY: New York University Press.
WHOA. (2015) “Cyberstalking statistics.” Working to Halt Online Abuse. Retrieved from www.haltabuse.
org/resources/stats/2013Statistics.pdf
Williams, M. (2006) Virtually Criminal: Crime, Deviance and Regulation Online. New York, NY: Routledge.
Wu, B. (2015) “Doxxed: Impact of online threats on women including private details being exposed and
“swatting.” Index on Censorship, 44(3), 46–49.
Yoshino, K. (2006) Covering: The Hidden Assault on Our Civil Rights. New York, NY: Random House.
Youngs, G. (2005) “Ethics of access: Globalization, feminism and information society.” Journal of Global
Ethics, 1, 69–84.
437
PART VII
Digital limits
New debates and challenges for the future
33
SOCIAL MEDIA AND
JOURNALISTIC BRANDING
Explication, enactment, and impact
Not so long ago, journalists were content creators who were largely tucked away behind the
pseudo-anonymity of bylines and headshots that told audiences little more than which journalist
they were reading, watching, or listening to. The public could only readily identify those jour-
nalists with celebrity status or marked cultural relevance. In essence, journalists were the bodies
behind the curtain, rarely revealed or engaged with in any way that could be construed as per-
sonal. They were represented, at least in part, by the content they produced and the organizations
that employed them.
The proliferation of social media has created cultures of ambient sharing and engagement,
which in turn prioritizes personal, intimate, and seemingly ‘authentic’ details about people’s lives
on a public scale. As Marwick (2013) has noted, social media has prompted the nearly 4 billion
people who make use of its various platforms (e.g., Snapchat, Instagram, Facebook, LinkedIn,
etc.) to consider who they are and how they want to be perceived. As users decide what pictures,
videos, memes, or gifs to post to their profiles and how to best describe themselves, so too do
they consider what content to share, where to direct that content (e.g., at other users, at specific
topics with hashtags), and who to engage with. These choices are driven by complex layers of
consideration that broadly take into account, for better or for worse, the user’s audience, real
or imagined. After all, the decision to post a weekend’s worth of vacation videos to a personal
YouTube channel is much different from the process of selecting which previous employment to
include on a LinkedIn profile.
The point is this: social media have given rise to the necessity of impression management, or
branding (Marwick, 2013), wherein individuals differentiate themselves through the selection
and presentation or exclusion of particular information and the platforms used to convey that
information. Journalists in particular curate professional images on social media that, while often
peppered with personal information and insights, tend to reflect organizational and institutional
norms and expectations (Bossio and Sacco, 2017). By considering how they want to represent
themselves to particular audiences, journalists (and social media users) rely on profile photos
and videos, memes, GIFs, biographical statements, status updates, and a multitude of other con-
tent options to create a digital self. While the social construction of the digital self, particularly
through social media, is relatively new, the effects are already seen across a breadth of profes-
sional industries. Journalism in particular has felt the impact that branding on social media has,
as journalists and news organizations wrestle with how to best personalize content and engage
441
with people in ways that build sustained audiences. Several recent studies (Holton and Molyneux,
2017; Molyneux et al., 2017) have shown that journalists have begun to work beyond simple
back-and-forth exchanges with audiences via email, comment sections, and tweets, instead find-
ing ways to make their content more approachable while also casting themselves as relatable. In
many ways, they have taken on the notion of personal branding, using their profiles and related
content to demonstrate their value. More importantly, they have intermingled personal informa-
tion about their lives – who they are married to, what their children look like, what their hob-
bies are, which presidential candidate they prefer, their pets’ names – alongside their professional
content (i.e., the news), creating a journalistic brand that is uniquely positioned as both personal
and professional, relatable and reliable.
This chapter considers that while journalists work to represent themselves as individual pro-
fessionals serving an organization and/or the journalism profession, they combine some of the
traditional tenets of journalism (e.g., truth, accuracy, transparency) with newly established norms
(e.g., opinion, personality, life casting). These are expanding the domain of what may be called
‘journalistic branding’ from something once demonstrated only by organizations or celebrity
journalists to an everyday aspect of individual journalism practice. As such, journalistic branding
can occur at the individual, organizational, and/or institutional levels. Here it is important to note
that while organizations certainly take on their own branding strategies, the focus here is on the
practices of individual journalists. As such, this chapter begins by considering the recent turn
toward journalistic branding, offering an explication of journalistic branding before moving into
a brief examination of how such branding is enacted in practice and what opportunities and
challenges have arisen from its relatively recent adoption.
An explication of journalistic branding

The idea that an individual and his or her social network combine to develop the individual’s
identity has been around much longer than social media. Building on the work of Erikson (1968),
scholars have suggested that identity exists only to the extent that others recognize it, necessitat-
ing the presence of an audience. Individuals performs certain aspects of self for this audience in
accordance with the traits they wish to emphasize, essentially creating and putting on different
faces for different groups (Goffman, 1959). These ideas have garnered renewed interest in the
age of social media, wherein individuals and their multiple audiences are collapsed into a single
context (Marwick and boyd, 2010). Thus in many ways social media are now a hub for the cre-
ation and presentation of identity, often in very public fashion. As more journalists take up social
media, often working daily on multiple platforms, questions arise about self-presentation and the
mingling of the personal with the professional (Bossio and Sacco, 2017, and this volume, Chapter
25; Carpenter et al., 2016; Kelly, 2017). Collectively, these studies have indicated that journalists,
including those performing freelance work, often consider organizational and institutional policies
and norms as well as their personal and professional audiences as they form their digital selves.
That is to say, journalists make decisions about their profiles and biographical information as well
as the content they post based largely on the expectations, real or perceived, of those who employ
them, those working in their profession, and those who will see what they post.
Journalists have not always needed to establish such a complex and publicly recognizable
identity nor a collaborative/sharing relationship with audiences. With the exception of some
broadcast personalities or syndicated columnists, most working journalists had little more than
a byline to tell their audiences about themselves. This has changed in recent years for a number
of reasons, but it is no coincidence that this shift has been coupled with journalists’ adoption
of social media, especially Twitter. The groundwork was laid more than a decade before Twit-
ter launched, when an article in Fast Company magazine popularized the notion of a personal
442
Social media and journalistic branding
brand (Peters, 1997). This made immediate sense for freelance journalists, but most other jour-
nalists could still borrow from the strength of their news organization’s brand (“I’m so-and-so,
a reporter for the Washington Post” carried substantial clout.). It wasn’t until these organiza-
tions were weakened by economic and cultural crises and journalists had a means of interacting
directly with their audiences that they began pitching themselves as individual experts as well as
people with lives outside of their professions (“Law & Ethics reporter covering the statehouse
and city hall. Mother of two”).
Self-promotion and identity formation online, particularly in the era of social media, is not
something unique to journalism. Scholarly work building on Marwick and boyd (2010, 2011)
indicates that one of the defining features of social media is the personal profile, which is at its
core a billboard for users to display what they wish for an imagined online audience. In gen-
eral, social media communications are marked by phatic expression and are strongly influenced
by celebrity practices – not that everyone is a celebrity, but everyone senses the potential for
celebrity status and tends to view followers as fans. This orientation toward treating the self as
a marketable entity and friends as consumers is a common way of understanding social media
practice generally, but journalists have built upon and customized these practices in specific ways
in response to challenges unique to their profession.
While news audiences and profit margins had been sliding for decades, it was the Great
Recession of the early 2000s that caused upheaval in the news industry (Anderson et al., 2012).
As advertisers began spending their money elsewhere and news organizations saw their revenue
streams drop by two-thirds, every journalist’s job seemed to be in peril. A natural response for
journalists was to seek to shore up their own stock, whether by building an audience of their own
that could follow them into a freelance career or by building a reputation that created value in
the eyes of their employer. Journalists not only had the impetus to do this but the autonomy to
do it for the first time, as organizations were distracted with keeping the ship afloat. Journalists
have said the early days of social media use in the newsroom (in 2008–2010, the depths of the
recession) were like a Wild West of experimentation as everyone sought ways to save themselves
and the industry they worked in.
This economic crisis also brought to a head another crisis within the news industry that had
been brewing for some time – a crisis of trust. When revenues disappeared and calls to save jour-
nalism began to multiply, many began to wonder, “We are saving what, exactly?” (McChesney
and Pickard, 2011). Observers have noted that years of anonymous sourcing, tabloidization, and
general softening and dilution of the news product had caused audiences to lose trust in news
media generally. One of the ways journalists responded to this problem was to seek transpar-
ency, whereby journalists allowed audiences to see more clearly how news is reported, linking to
original documents and exposing their processes. This, again, is quite easily achieved on social
media compared to more traditional news formats. Journalists and news organizations, while
continually hesitant to adopt the latest trends in social media for ethical, financial, and other
reasons, nonetheless have used Twitter, Facebook Live, Snapchat, and a number of other channels
to broadcast content and to invite viewers into the news process either as a fairly passive audi-
ence or as hyper-engaged content contributors. Letting the audience see the person behind the
curtain – and, in fact, be behind the curtain on occasion – was billed as a way to increase trust
and thereby audience size, and this pushed journalists not only toward their audiences but toward
revealing themselves as individuals.
Journalists with local, regional, national, and international prominence suddenly found them-
selves working in shared spaces. The opening up of a larger and more recognizable presence is
part of the performances Goffman identified, wherein individuals show off aspects of themselves
to whomever may encounter the information (the classic example is posting pictures of what
you had for lunch). This being a norm on social media, it was natural then for journalists to
443
begin sharing and showing off their work in a similar fashion. In order to make such engagement
more relevant for audiences on social media spaces, journalists began reworking their norms to
include humor, opinion, and personality above objective accounts of the news (Lawrence et al.,
2014). In order to gain a following on social media and in order to fit into those online networks,
journalists have begun turning away from straight-ahead news reports and toward sharing more
of themselves within the limitations imposed by personal boundaries and organizational policy
(Bossio and Sacco, 2017). In fact, in some ways, audiences have been demanding this type of
interaction.
Scholars have identified a growing perception among both audiences and journalists that these
two groups are expected to be co-participants in creating and distributing news (Rosen, 2006;
Lewis, 2012). This is especially the case on social media, where journalists and their audiences
occupy the same space and compete for attention. Journalists have often sought to establish their
authority by separating themselves from their audiences (Singer, 2003). But in the social media
arena, the lines between personal and professional are blurred, and what it means to be a journalist
is harder to define. Thus, without strong, trusted news organizations to borrow credibility from,
and without a clearly defined professional membership to distinguish themselves from laypeople,
journalists embedded in the sea of voices on social media have turned toward branding as a way
to help themselves stand out from the crowd.
This evolution in professional work has been somewhat tempered by news organizations,
many of which retain some level of influence over how their employees represent themselves and
the company online. It’s worth noting that, when it comes to audience engagement and brand-
ing, organizations often have different goals from rank-and-file journalists. This includes loyalty
to the news organization itself, a commitment to coworkers and, in some cases, an assigned beat.
Thus, if audiences have been pressuring journalists to put on a more transparent, personal face,
employers have been pressuring journalists to put on a more professional face, one that enhances
credibility or, at the very least, does not reflect poorly on the organization.
This mingling of the personal and professional has come at a time when journalists are wres-
tling with how to maintain, if not increase, their market value (Brems et al., 2017). In order to
remain relevant, journalists have had to rethink the boundaries that traditionally separated them-
selves from their audiences, searching for ways to engage with and include the public in the news
process while also rising to meet expectations of personalization on social media. Journalists
have largely faced a ‘join or die’ state on social media and, once there, find themselves juggling
traditional professional norms and demands for more personal and approachable (e.g., humorous,
opinionated, transparent) content from the public. Not surprisingly, this has left some journalists
feeling anxious, worrying about the ethics of embedding personal information in news content
or struggling to understand exactly what news organizations and audiences expect of them
(Bossio and Sacco, 2017; Brems et al., 2017; Molyneux et al., 2017).
These consternations are coupled with what some journalists have described as a loss of per-
sonal identity, wherein they are exchanging intimate content for more professionally oriented
material to appease news organizations (Holton and Molyneux, 2017). While this seems to suggest
journalists continue to value organizational and institutional norms and policies over audiences,
other studies have indicated that journalists are increasingly finding value in marketing themselves
as engaged, reliable, and personal on social media platforms (Bossio and Sacco, 2017). The develop-
ment of journalistic branding is thus the result of a convergence of influences including economic
and cultural crises within journalism, the rise of social media, and the resultant collision between
journalists and their audiences. Such branding takes into consideration a multitude of factors that
extend beyond the personal or individual to include the organizational and the institutional.
Thus, journalistic branding may broadly be defined as ‘seeking to establish and promote a
public-facing journalistic identity’. This identity distinguishes the journalist from others in the
444
field and from other sources of information online and seeks to establish her or him as a person
worth attending to. This identity may have one or many constituent parts, whether personal or
professional in nature; if professional, it may serve individual, organizational, and/or institutional
goals. This is different from gathering sources, presenting news, and other journalistic activities
in that it is primarily focused on developing a journalistic identity (i.e., the brand), generally for
commercial purposes. Evidence of this practice is most readily found on social media, but brand
development may happen in various mediated and unmediated communications.
The following section pays close attention to enactments of journalistic branding. A small
but growing number of studies have interviewed journalists (Brems et al., 2017; Holton and
Molyneux, 2017; Molyneux et al., 2017) to better understand how they brand themselves, their
organizations, and their profession, while others have examined journalists’ social media con-
tent (Carpenter et al., 2017; Hanusch and Bruns, 2017; Ottovordemgentschenfelde, 2017). Col-
lectively, these studies indicate that journalistic branding is being incorporated and employed
by journalists despite some continued trepidations toward social media and the inclusion of
branding.
Branding enactment
As Tandoc (2013) and his colleagues noted, understanding changes in journalism practice
requires constant attention to journalists and the content they produce. By paying close attention
to journalists’ attitudes and beliefs as well as their resulting behaviors and content, or enactments,
a clearer picture of their identities can begin to emerge. In terms of journalistic branding, several
studies have looked at how and why journalists incorporate branding on social media and what
their profiles and shared content can tell us about the current state of journalistic branding. These
studies indicate that while journalists’ choices in branding are driven in part by news organiza-
tions or perceptions of what journalism ought to look like, they also contain personal informa-
tion and professional content that is individual focused (e.g., links that point to a journalist’s
content or profiles that identify a journalist’s beat or area of interest).
For example, Carpenter and her colleagues (2017) found that journalists do combine pro-
fessional content with personal information on Facebook, though freelance journalists tend to
present a more professional or “serious” front than journalists employed by news organizations.
As the study suggests, this may indicate that freelance journalists are engaged in brand mainte-
nance that takes into consideration potential employers viewing their content. They may also
be striving for validation from audiences, who can sometimes associate credibility with employ-
ment. Nevertheless, journalists working for news organizations as well as those with professional
autonomy wove personal content alongside the news, offering a representation of themselves that
aligns with what other scholars have found on Twitter and other social media platforms: journal-
ists are including more personal information than ever alongside their news content (Brems et al.,
2017; Hanusch and Bruns, 2016; Ottovordemgentschenfelde, 2017).
In their analysis of the Twitter profiles (384) and tweets (1,903) of journalists in the United
States, Molyneux and his colleagues (2017) determined that, in addition to whatever personal
content journalists include in their profiles (e.g.: Sports reporter by day, beer enthusiast by night;
Mother of two; Purveyor of antiquities) and their tweets (e.g.: Off to the Ozarks today; Feelin’ this
new Drake album), journalists’ professionally oriented tweets also exhibit journalistic branding
on individual, organizational, and institutional levels.
At the individual level, journalists distinguished themselves by their areas of specialty (e.g.,
health reporter, sports broadcaster) and shared information about their work as well as content
they had created. They relied heavily on links to direct people to their work, exhibiting what
Brems et al. (2017) labeled an “explicit form of self-promotion”. While freelance journalists have
445
long engaged in this practice, it’s relatively new for journalists, who, as Brems and her colleagues
found, continue to worry about offending their audiences with an overabundance of “look-at-
me” content. Yet journalists reported in at least one study (Holton and Molyneux, 2017) that
despite worries that working self-promotion into their news content could undermine audience
trust, the ability to do so afforded them a renewed sense of self-worth and a means to distinguish
themselves and their content. In fact, some felt like they had to.
Molyneux et al. (2017) also found journalists embrace their organizations in their profiles
and in their tweets, sharing organizational level information such as the name of their employer
or content from a coworker. Journalists mentioned their employer in 80% of profiles analyzed,
with more than a quarter going so far as to include their organization in their Twitter handle
(e.g., @AshleyG_KVUE). Nearly half of all tweets that contained branding referenced work from
the journalist’s news organization or co-workers, with links to that work appearing frequently.
Such efforts may help journalists demonstrate loyalty and an adherence to organizational policies
that encourage social media sharing of coworkers’ content, though some journalists have expressed
concern over the practice. In exchanging personal content or individual level branding for organi-
zational, some journalists have reported a loss of self-identity online (Holton and Molyneux, 2017).
Others have worried that by putting the organization first, they sacrifice the very sort of personal-
ization that audiences now demand. Recently, the New York Times made this sacrifice explicit, stating
in its published social media policy (2017) that giving up a personal social media presence in favor
of a fully professional one is “the price of our employment by a major media institution”.
These same concerns bleed over into considerations of institutional-level branding, where
journalists reference their profession, other news organizations, or other journalists. While jour-
nalists more frequently promote work from themselves and their employer (Brems et al., 2017;
Carpenter et al., 2017; Molyneux et al., 2017), they do share or cross-promote work from other
outlets and discuss journalism and its norms broadly. They explicitly point out examples of guid-
ing principles such as objectivity and truth and police other journalists who tweet erroneous
information (Holton and Molyneux, 2017). In a sense, they work to ensure social media is a space
where the institutionalized values of journalism practice remain intact. This work has taken on
increasing intensity recently as journalism’s credibility is doubted in the uproar over “fake news”.
When those in power question the very necessity of a free press, it falls to journalists to make the
case for their work as an institution – all of this while continuing to meet the needs and expecta-
tions of their organizations, their peers, and their audiences.
Overall, the evidence noted here points to a rise of journalistic branding from nearly zero
to an activity that now takes up a substantial portion of nearly all journalists’ activity on social
media. Developing a personal brand is something all journalists interviewed were familiar with
(Molyneux and Holton, 2015), and the idea is taught in journalism schools and industry publica-
tions. As a recent Columbia Journalism Review article pointed out, it appears that, indeed, “per-
sonal branding is key to building a career in journalism” (Bech Sillesen, 2015).
Branding’s impact
As journalists respond to their environment, in part by enacting journalistic branding on social
media, it’s worth considering what effects this might have on them, their industry, and their audi-
ences. A key narrative in this discussion is the selection of priorities as journalists, their employers,
and their audiences all seek attention and control in the digital news arena.
Journalists working to prioritize their time on social media and in their jobs in general must
balance promoting themselves, their employers, their co-workers, and to a lesser degree, the insti-
tution of journalism. The benefits of promoting oneself are obvious – by building a social media
following based around individual identity, a journalist is more valuable as an employee and as
446
an individual, simultaneously improving job security and opening up opportunities in freelance

work. Yet pushing self-interest too far could risk upsetting a journalist’s employer or employers,
which in many cases have by now developed policies encouraging or requiring certain behaviors
on social media, including organizational-level branding. That a balance must be struck is clear to
journalists, who have noticed when news organizations fire or reprimand employees for behavior
on social media that does not align with organizational policy (Holton and Molyneux, 2017).
However, where that balance lies is not quite clear given that journalists continue to lament a lack
of clear organizational policy.
Rather, journalists rely on what they believe their organization, and indeed their colleagues
and their audiences, may want. Those perceptions are often tested with an adherence to employer
mandates and journalistic norms in mind, restricting some journalists from extending their
branding practices beyond those of others around them. While freelance journalists have more
autonomy and thus would seem more likely to test the boundaries of acceptable branding behav-
ior, they may also be more cautious if they are actively seeking employment with news organiza-
tions. Regardless, journalists across the board are engaging in new forms of audience interaction
that call into question the inclusion of journalistic branding.
Audiences have made it clear that they do not like ads (judging by the widespread use of
online ad blockers), and so journalists might do more harm than good if their social media time
is spent exclusively in promoting themselves and their organization. For their part, journalists
have expressed this as a chief concern (Brems et al., 2017), worrying that promoting their orga-
nizations or journalism in general may strip away their individual identity and that marketing
themselves too much might cause audiences to question their motives. Thus, journalists must
seek to balance promotional activities with other forms of engagement that the audience may
enjoy more in their efforts to retain followers, such as reciprocal exchanges of information or
public praise of individuals who provide news tips and content.
That raises the question of where exactly news fits into the schedule. If a journalist spends
all of her social media time propagating news stories, would followers retain interest? Perhaps
assuming that the social media audience wants only personal interactions misses the mark, so
again it becomes a question of balance in which the target equilibrium is not clear. Of deeper
concern, perhaps, is the thought that journalists risk losing independence when they place any
other priority in competition with reporting the news. Independence has traditionally been
interpreted as freedom from the interests of government, advertisers, and those journalists cover.
But is it possible that journalists have become more concerned with their public image than with
their work? For instance, might a journalist skip a story if he knows the audience won’t love and
share particular content across social media, either because the analytics say so or because of some
internal intuition, essentially prioritizing the news judgment of the audience over his own? Many
feel journalism should operate in the public interest, but with these trends in branding it’s worth
asking whether audiences, or at the very least journalists’ perceptions of their audiences, could
have too much influence over journalists.
Conversely, organizations working to understand how best to reach and maintain audiences
must balance their desire to control their public image – directing audience attention toward
their brand – with their desire to allow their best journalists to do their work and form meaning-
ful relationships with audiences. Workers in general value autonomy, and journalists perhaps even
more so; but experience has led many employers to set boundaries and expectations for social
media behavior. Complicating matters for employers is the need to sustain a commercial enter-
prise. If a journalist sometimes develops a personal brand as a backup plan (becoming a freelancer
in case the news organization fails), the news organization develops a brand as a primary means of
success, and therefore has more at stake. Businesses of all stripes now have social media strategies
and social media managers, but these efforts don’t translate directly into revenue for all businesses.
447
What audiences are left with, then, is a combination of shouting match and tug-of-war as
journalists and news organizations compete among themselves and with each other for attention
and control in the social media space. Economic theory generally suggests that competition is
good for the consumer as it drives down prices and encourages innovation, but this logic may
not apply in the case of digital journalism. First, the price is already zero and cannot be lower.
For some audience members wishing to get above the fray, online subscriptions are rising as
they begin to value certain commitments to quality or specialization that set news organizations
apart. In this situation, competition may be good for audiences to the extent that it diversifies
news offerings rather than creating increasingly cheaper versions of the same product. Second,
news organizations have shown themselves to be startlingly resistant to innovation, ceding this
territory to technology companies that benefit from different structures, funding models and
positions within the market. Competition among news organizations and journalists, then, has in
many cases felt like fighting over what’s left rather than spurring a race toward the future. This is
echoed in the findings of many of the studies included in this chapter that note the tension jour-
nalists express (and demonstrate in their social media profiles and content) as they wrangle with
what personal information to include as part of their self-representations and what individual,
organizational, or institutional-level branding to include alongside their news.
This brings us back to the evolving personal and professional identities of journalists. As
journalists consider the changing norms of their profession and the evolving demands of their
organizations and audiences, they are constantly reshaping their content as well as their self-
representations in ways that are altering the institution of journalism and their roles within
it. Even as some news organizations work to provide clearer policies of social media engage-
ment and branding that encourage rather than punish experimentation, others continue to place
restrictions on such behavior. Such an approach puts journalists in the precarious position of
aligning their branding efforts, and indeed their social media practices, more with their organi-
zations than themselves, a decision that journalists have observed is diminishing their personal
identities as well as their professional individuality.
As this chapter has illustrated, journalistic branding is a complex technique journalists are
employing more frequently to promote and differentiate themselves. More pointedly, this chapter
explicated journalistic branding as being personal or professional in nature and occurring at the
individual, organizational, or institutional levels to help construct identities at each of those posi-
tions. This definition is put forth with the hope that future discourse surrounding evolving journal-
ism and identity will build upon its beginnings. Journalistic branding is, after all, but one evolving
practice journalists and news organizations have begun to incorporate and understand. It will be
important for journalists and researchers alike to examine the ethical, cultural, and informational
consequences of this emerging practice, with an eye toward the democratic mission of the press. At
a broader level, these findings hint at changes in how identity is formed and enacted in networked
society, where the personal and the professional increasingly interweave and blend together.
Further reading
Our work in the area of journalistic branding began several years ago, originally culminated in
Branding (Health) Journalism: Perceptions, Practices, and Emerging Norms (Molyneux and Holton,
2015), and has since benefitted greatly from other scholars with similar interests. In Journalis-
tic Branding on Twitter: A Representative Study of Australian Journalists’ Profile Descriptions, Folker
Hanusch and Axel Bruns (2016) offer an intensive examination of how personal journalists can
be on social media despite their wariness in sharing such information. Cara Brems and her col-
leagues (2017) provide a glimpse of the identity struggles journalists today face in Personal Brand-
ing on Twitter: How Employed and Freelance Journalists Stage Themselves on Social Media. Finally, the
448
work of Diana Bossio and Vittoria Sacco (2017) in From ‘Selfies’ to Breaking Tweets: How Journalists
Negotiate Personal and Professional Identity on Social Media has provides a much-needed critical lens
with which to view the professional, organizational, and institutional drivers behind journalists’
attitudes toward and engagement with social media.
References
Anderson, C. W., Bell, E. and Shirky, C. (2012) Post-Industrial Journalism: Adapting to the Present. New
York: Tow Center For Digital Journalism. Retrieved from http://towcenter.org/research/post-
industrial-journalism/
Bech Sillesen, L. (2015) “Analyzing journalists’ Twitter bios.” Columbia Journalism Review. Retrieved from
www.cjr.org/analysis/analyzing_journalists_twitter_bios.php
Bossio, D. and Sacco, V. (2017) “From ‘selfies’ to breaking tweets: How journalists negotiate personal and
professional identity.” Journalism Practice, 11(5), 527–543.
Brems, C., Temmerman, M., Graham, T. and Broersma, M. (2017) “Personal branding on Twitter: How
employed and freelance journalists stage themselves on social media.” Digital Journalism, 5(4), 443–459.
Carpenter, S., Kanver, D. and Timmons, R. (2017) “It’s about me: A study of journalists’ self-presentation of
their visual and verbal selves.” Journalism Practice, 11(10), 1246–1266.
Erikson, E. H. (1968) Identity: Youth and Crisis. New York, NY: International Universities Press.
Goffman, E. (1959) The Presentation of Self in Everyday Life. New York, NY: Doubleday.
Hanusch, F. and Bruns, A. (2017) “Journalistic branding on Twitter: A representative study Australian jour-
nalists’ profile descriptions.” Digital Journalism, 5(1), 26–43.
Holton, A. E. and Molyneux, L. (2017) “Identity lost? The personal impact of brand journalism.” Journalism,
18(2), 195–210.
Kelly, S. (2017) Personal Branding for Entrepreneurial Journalists and Creative Professionals. New York, NY:
Routledge.
Lawrence, R. G., Molyneux, L., Coddington, M. and Holton, A. E. (2014) “Tweeting conventions: Political
journalists’ use of Twitter to cover the 2012 Presidential Campaign.” Journalism Studies, 15(6), 789–806.
Lewis, S. C. (2012) “The tension between professional control and open participation.” Information, Com-
munication & Society, 15(6), 836–866.
Marwick, A. E. (2013) Status Update: Celebrity, Publicity, and Branding in the Social Media Age. New Haven,
Marwick, A. E. and boyd, d. (2010) “I tweet honestly, I tweet passionately: Twitter users, context collapse,
and the imagined audience.” New Media & Society, 13(1), 114–133.
Marwick, A. E. and boyd, d. (2011) “To see and be seen: Celebrity practice on Twitter.” Convergence: The
International Journal of Research into New Media Technologies, 17(2), 139–158.
McChesney, R. W. and Pickard, V. (2011) Will the Last Reporter Please Turn Out the Lights: The Collapse of
Journalism and What Can Be Done to Fix It. New York, NY: The New Press.
Molyneux, L. and Holton, A. E. (2015) “Branding (health) journalism: Perceptions, practices, and emerging
norms.” Digital Journalism, 3(2), 225–242.
Molyneux, L., Holton, A. E. and Lewis, S. C. (2017) “How journalists engage branding on Twitter: Indi-
vidual, organizational, and institutional levels.” Information, Communication, & Society, Ahead of print.
The New York Times. (2017) The Times Issues Social Media Guidelines for the Newsroom. Available at www.
nytimes.com/2017/10/13/reader-center/social-media-guidelines.html
Ottovordemgentschenfelde, S. (2017) “Organizational, professional, and personal: An exploratory study of
political journalists and their hybrid brand on Twitter.” Journalism, 18(1), 64–80.
Peters, T. (1997) “The brand called you.” Fast Company. Retrieved from https://my.mandelfoundation.org.
il/mli_pdf/kenes2010/the_brand_call_you.pdf
Rosen, J. (2006) “The people formerly known as the audience.” PressThink. Retrieved from http://archive.
pressthink.org/2006/06/27/ppl_frmr.html
Singer, J. B. (2003) “Who are these guys? The online challenge to the notion of journalistic professional-
ism.” Journalism, 4(2), 139–163.
Tandoc, E. C., Hellmueller, L. and Vos, T. P. (2013) “Mind the gap: Between journalistic role conception
and role enactment.” Journalism Practice, 7(5), 539–554.
449
34
RECONSIDERING THE
INTERSECTION BETWEEN
DIGITAL JOURNALISM AND
GAMES
Sketching a critical perspective
Igor Vobič
Throughout the last century the news industry has continuously introduced technological inno-
vations in its attempts to bridge the private interests of news media owners in making profits and
the public aims of journalists to meaningfully connect citizens with social life – often, however, at
the expense of the latter (cf. Hardt, 1998: 173–190). In contemporary capitalist societies, mean-
ingful journalism is in a “delicate moment” (cf. Schudson, 2017) as it tries to rearticulate itself as
an acceptable and trustworthy production of knowledge necessary for a reflective and inclusive
social life. At the same time, journalism’s institutional structures, practices, roles, and percep-
tions of the public are in flux, resulting in the emerging “new forms” that are “accompanied
by much controversy” (Splichal and Dahlgren, 2016: 5–6), while journalism’s public character
is being continuously disfigured by the pressures of private owners and financial investors (ibid.:
12). As the news industry aims to overcome the contemporary crises of economic viability and
civic adequacy (Blumler, 2010: 439) by seeking and implementing new technologies and ways
and forms of (re)connecting with people, paradoxically, as citizens, consumers, and workers, the
questions of what is journalism, how it is done, and why are becoming more complex in today’s
increasingly digitized environment. This calls for continuous critical attention from journalism
scholars and others – especially in terrains lying on the borders of digital journalism but that
could – if studied – shed new light on old problems of journalism and its research. One such
terrain extends to the intersection between digital journalism and games.
“Journalism is not only becoming digital, it is also becoming playful”, write Ian Bogost et al.
(2010: 178) in the final chapter of their seminal work Newsgames: Journalism at Play, where they
indicate that the convergence between digital journalism and games touches on larger struggles
behind the boundaries of knowledge production, social reality representations, and perceptions of
and participation in political life. In this context, discussing the intersection of digital journalism
and games is not an easy task. On one hand, in the digitized world it is “increasingly harder to
pin down” the idea of journalism and its boundaries (Malik and Shapiro, 2017: 15–17), while,
on the other, the growing variety of forms, functions, and domains of games makes the signifier
“increasingly slippery” (Jagoda, 2017: 7–8).
450
Digital journalism and games
As elaborated later, although scholarship at the intersection of digital journalism and games relates
to larger issues of journalism as a social institution, profession, cultural practice, and labor, journal-
ism’s coverage of games, employment of the game-form and its rhetoric in news, and the application
of game elements to established journalism practice remain under-discussed in digital journalism
studies, which largely focus elsewhere (cf. Ahva and Steensen, 2017). As a result, this chapter closely
reviews the various connections between digital journalism and games while aiming to develop a
critical lens by grounding our materialist view in broader social theory in order to better understand
certain historical problems of contemporary journalism as articulated through games.
Literature review: three marginalized areas

By reviewing studies that interrogate the intersection between digital journalism and games,
this section attempts to chart this relatively marginalized terrain of digital journalism research.
At least three areas of scholarship can be identified, each dealing with a specific phenomenon at
the intersection of: (1) “game journalism” as a branch of journalism dealing with the coverage
of novelties in digital games and gaming; (2) “newsgames” as a form designed to illuminate a
specific facet of news by means of digital games’ particular rhetoric, with a goal to engage people
in relevant social issues; and (3) the “gamification of journalism” as a strategy of applying game
elements to digital journalism’s forms, production, and consumption.
Game journalism – journalistic coverage of digital games

One area of scholarship at the intersection of digital journalism and games deals with what
appears as a special branch of journalism that covers digital games through previews, reviews, fea-
tures, and other forms. Although games have been in our lives for about four decades and gradu-
ally have become an integral part of popular culture and a growing multibillion-dollar industry
(cf. Jagoda, 2017), surprisingly little attention has been paid to game journalism by either game
scholars or journalism researchers. Yet in the last decade or so, an increased attention is observ-
able. Namely, scholars (e.g. Hall, 2003; Nieborg and Sihvonen, 2009; Carlson, 2009; Fisher, 2012;
Ribbens and Steegen, 2012; Foxman and Nieborg, 2016; Perreault and Vos, 2016) have explored
social, politico-economic, and cultural particularities of game journalism and argued that this
branch is importantly shaped by the difficult, ambivalent ties with the game industry and audi-
ences. At the same time, it continues to contribute to the development of game culture as well
as the collective consciousness and complex tensions deriving from it, which have worryingly
been manifested in the recent #GamerGate controversies and harassments (cf. Mortenstein, 2016;
Cote, 2015; Fisher, 2015; Chess and Shaw, 2015; Braithwaite, 2016).
Traditional journalism has predominantly framed games as a social threat since the 1970s
(cf. Williams, 2003). Only in the new millennium has this changed in legacy media, when the
artistic merits of games started to be celebrated with the simultaneous rise of their economic
and cultural relevance (cf. McKernan, 2013). Writing about games, however, has evolved from
“consumer press” produced by large game developers (Cote, 2015: 1) to “a flood” of specialist
game magazines in printed forms or as e-zines (Fischer, 2015: 2). By considering its dubious
relationship with the game industry and gaming communities, Carlson (2009: [4.1]) refers to
game journalism as “enthusiast press” that produces “consumer-oriented publications that focus
on publicizing specific categories of goods, often high-end technological products” and carries a
connotation among “mainstream journalists” of not being “real” journalism (ibid.). Other schol-
ars also find it difficult to position game journalism in terms of the journalistic field. Some see it
as part of “entertainment journalism” that is chiefly concerned with the “reveal/preview/review
cycle” associated with new products (Fisher, 2012: 224–225); others understand it as “lifestyle
451
Igor Vobič
journalism”, a market-driven form of journalism that blends information, advice, and guidance
for consumers (Perreault and Vos, 2016: 4). In this context, Foxman and Nieborg (2016: 4) report
on ambiguous self-understanding in the field of game coverage and critique. As they place game
journalism somewhere between “arts criticism” and journalism, their study reveals the “network
of ambivalences” in this sphere (ibid.: 35), where game journalism
not only is connected economically to the industry [as a] critical source of information
about upcoming releases, but also has been complicit in establishing and representing
the culture of games. The fallout of their coverage, although it may not have been their
intent, is an environment rife with ambivalence and tension.
(ibid.: 36)
Similar findings are provided by Ribbens and Steegen (2012), who explore the difficulties game
journalists encounter when writing game reviews. According to the authors, game reviewers face
uncertainty as they stress that issues of “subjectivity” are exacerbated in game journalism, bound-
aries of the editorial-business divide are being blurred, and the recent diversification of game audi-
ences brings a variety of conflicting factors into game journalists’ conduct (Ribbens and Steegen,
2012: 216). In this setting, the genre of game reviews has evolved into an online genre with great
variation, as gamers in large numbers discuss games in online forums and on social media plat-
forms, digital distribution sites that sell games, and game journalism websites (cf. Thominet, 2016).
This domain of a sort of gamer journalism mixes together facts and opinion, debates and gossip, news
and reviews, and the deceptive and the insightful while employing different semiological forms
and platforms. Although hardly discussed, this example of the heterogeneous “citizen journalism”
blurs the traditional boundaries between journalism and non-journalism (cf. Vobič and Dahlgren,
2013: 16), deepening tensions in the intertwined networks between game journalism, its audi-
ences, and the game industry, while calling for additional scholarly inquiry.
In scholarship exploring game journalism, two prevailing perspectives can be identified: politico-
economy and cultural approaches. An example of the former is Carlson’s (2009) article, which
argues that game journalists are “mediators of commodity value” by performing “immate-
rial labour” shaped by the conditions of capitalism. While the functioning of game journalism
depends on the game industry as its prime advertisers and those that give newsrooms privileged
access to novelties in the market, game journalists “stand between consumers and producers as
fulcrums, spinning and molding the knowledge that each has (access to) about the other, impact-
ing consumption habits as they simultaneously shape production practices” (ibid.). A study that
adopts the cultural approach is by Perreault and Vos (2016), who reveal that game journalists
undertook “paradigm maintenance” in the midst of #GamerGate, a viral campaign that has both
questioned the ethics of game journalism and badgered women involved in game production
and game criticism (ibid.: 1). The authors state that game journalists “decisively broke with the
old paradigm of an enthusiast-press and claimed their place within the paradigm of traditional,
public-minded journalism” (ibid.: 13).
As indicated earlier, game journalism is in a complex social milieu where digital games are
emerging as an increasingly complex contemporary form and where the game industry is grow-
ing globally (cf. Jagoda, 2017). Since this review indicates that this branch of journalism reflects
larger difficulties of the field, further scholarly attention would be valuable – with a particular
focus on the porous boundaries between game journalism professionals and gamer-generated
content on games. One possible route would be to tackle the phenomena, processes, and relations
interrogated in this area of research by trying to supplement the prevailing perspective focused
on subjectivist agency and identity with materialist epistemological stands. This would allow
scholars to comprehensively explore cultural tensions that – as the reviewed studies indicate – are
452
importantly based on the politico-economic logics of game journalism and to better understand
its ambivalences and implications.
Newsgames – bringing news into digital games

Another area of studies focuses on a form of digital games that is designed to illuminate a par-
ticular side of news through the specific rhetoric of rule-based involvement with an aim to
meaningfully connect people to relevant social events or issues. A review of discussions in the
last decade or so suggests that scholarship initially primarily focused on defining newsgames and
discussing their distinctive features. Nevertheless, as elaborated here, very early on scholars (e.g.
Burton, 2005; Sicart, 2008; Treanor and Mateas, 2009; Bogost et al., 2010) shifted away from the
“ludology vs. narratology” tension in game scholarship (Anderson, 2013: 295) and through case
illustrations explored – to put it simply – “how games act as journalism” (ibid.) in order to discuss
their social relevance in a transforming communication environment.
The term newsgame was reportedly introduced by the game developer and scholar Gonzalo
Frasca in the early 2000s (cf. Bogost et al., 2010: 13). On Newsgaming.com, Frasca and his col-
leagues described the notion as “simulation meets political cartoon” that can be “a great tool
for better understanding our world”. As argued in more detail elsewhere (cf. Vobič et al., 2014:
126–127), later studies have related to but also departed from Frasca’s narrow understanding of
newsgames by discussing the notion theoretically and trying to map the empirical terrain in the
news industry and outside it (e.g. Burton, 2005; Sicart, 2008; Treanor and Mateas, 2009; Bogost
et al., 2010). Although differences in understandings of newsgames can be identified in the litera-
ture, there is a common discussion on how this particular form of knowledge operates to create
meaning and spur engagement. “Procedural rhetoric” (Bogost, 2006) is profoundly examined as
“the way that a videogame embodies ideology in its computational structure” or, in other words,
how meaning is conveyed through the dissection of a rule-based system of newsgames.
It is argued that a “paradigm shift” is at play here, as newsgames offer journalists an opportu-
nity to shift from “stories” to “systems” in their conduct (Bogost et al., 2010: 179) – “to share
raw behaviours and dynamics that describe a situation” (ibid.). Others, however, are more hesitant
as they identify “issues with procedural rhetoric” (Treanor and Mateas, 2009: 6) and suggest that
a “hard design problem” of newsgames lies in the “unintended rhetoric” that might manifest
through gameplay (ibid.: 7). Yet little is known about how people relate to newsgames and their
knowledge claims. Teixeira and her colleagues (2015a, 2015b) only provide small insights in their
studies of “usability and gameplay” into the documentary and simulation genres by conducting
experiments involving users. They do not go further than saying that pursuing these features of
involvement in the development is crucial for a newsgame to illuminate a certain aspect of news
as well as to become “fun and rewarding for players” (Teixeira et al., 2015a: 6067). In this context,
the all-encompassing definition of a newsgame as “any intersection of journalism and gaming”,
introduced by Bogost et al. (2010: 13), leaves much ground to tackle in future studies – not only
with respect to the notion of procedural rhetoric but also regarding larger material and subjective
aspects of technological innovations in journalism.
What appears common in this area of scholarship is understandings of the notion of news in
line with the “hegemonic” Western model of journalism (cf. Nerone, 2013), while scholars take
it for granted and hardly consider its historical roots and ideological underpinnings. It is thereby
difficult to explore newsgames as “an empathetic and immersive experience or mimicking a real-
world system” (Foxman, 2016), and – as it seems – it is becoming even harder and more complex.
While there are observable shifts to “transmedia strategies in journalism” (cf. Gambarato and
Tárcia, 2016), the “most significant deterrents” in newsgames production remain “the skills, time,
and financial resources required to create, deploy, and maintain these products” (Foxman, 2015: 4).
453
Igor Vobič
The industry has thus been “hesitant” to integrate newsgame design into workflow and develop
firm modi operandi that would guide the production of knowledge (Treanor et al., 2012: 1).
Although newsgames lost favor with producers and financiers after the initial experiments in
the 2000s, there are traces of an “organic evolution” of design and editorial in multimedia teams
across the industry (Foxman, 2015: 25) and a “wide spectrum” of game-based products that can
be identified in media and journalism, such as puzzles, quizzes and “gameworlds” (ibid.: 16–19).
The recent employment of immersive technologies to journalism (cf. de la Peña et al., 2010),
automation in news (cf. Carlson, 2017), and game development (cf. Nelson and Mateas, 2007;
Treanor et al., 2012) as well as larger changes in relations between newsmakers and users both
inside and outside the emerging “playful newsrooms” that increasingly employ game-based
operations (Foxman, 2016) make it difficult to predict the future of newsgames. What is evident
is that newsgames as a form and production of knowledge as well as public acceptance of their
producers’ claims are being interrogated, similarly to other areas within journalism (cf. Ekström,
2002). To further interrogate the notion and phenomenon of newsgames that is also defined by
the creative boundaries of the profit-oriented news industry (cf. Nerone, 2013), critical scholar-
ship with historically and contextually informed studies are needed in order to explore the limits
of more inclusive mechanics and rhetoric of news provision, the constrained contestation of
conventional journalistic constructions of place and time, and the boundaries of experimenta-
tion in individual and collective public engagement. A close critical study of the materiality of
newsgames would allow us to not only better understand newsgames as a form and production
of knowledge but also the broader difficulties of and tensions in journalism as a social institution.
Gamified journalism – employing games in journalism

A distinct area of research explores the strategy of gamification that applies game elements, such
as points, badges, and leaderboards, to journalism production, news formats, and user engage-
ment. Although gamification has received substantial attention in other contexts – both cel-
ebratory appraisals and more reserved considerations (cf. Waltz and Deterding, 2014), a review
of studies dealing with gamified journalism appears quite underexplored. Only recently have
scholars, most notably Raul Ferrer-Conill (2016, 2017), started to explore how the “playful logic
of gamification” (Ferrer-Conill, 2016: 49) is articulated in journalistic representations of social
life, people’s connection with and participation through journalism, and the production process
by professionals and nonprofessionals.
While newsgames as specific game-forms of contemporary digital journalism rest on pro-
cedural rhetoric in their illustrations of news, gamification as an innovation strategy brings a
“game-like experience” (Ferrer-Conill and Karlsson, 2016: 367) not simply to established and
evolving news forms but also to the production and consumption of news. To simplify the dis-
tinction, Ferrer-Conill and Karlsson (ibid.) suggest that newsgames “bring news to games” and
that gamification “brings games to news”. Similar to other innovations in the century of modern
journalism, gamification spurred quite simplified initial celebrations that were later countered
by more sober, even critical assessments (cf. Ferrer-Conill, 2016: 49). Advocates of gamifica-
tion stress the potential of gaining engagement, loyalty, and productivity, while critical perspec-
tives translate that into surveillance, control, and exploitation (ibid.). For instance, newsgames
designer and researcher Ian Bogost (2011) provides a vivid and often cited critique of the strategy:
“[G]amification is marketing bullshit, invented by consultants as a means to capture the wild,
coveted beast that is videogames and to domesticate it for use in the grey, hopeless wasteland of
big business, where bullshit already reigns anyway”.
Although it appears that problems with implementing gamification mean it has at least slightly
fallen out of favor in the news industry after its stellar rise in popularity in the early 2010s (cf.
454
Foxman, 2015: 9, 16), explorations of gamified journalism remain valuable for digital journal-
ism studies, as this strategy touches on the core of the perpetual conflict in journalism – the one
between the private interests of news media owners and the public goals of journalists (cf. Vobič,
2015). According to Raul Ferrer-Conill (2016), gamification is implemented within digital jour-
nalism in four ways.
By employing gamification in news stories, the experience of news shifts as users are turned
into players who engage with playable multimedia narratives and their immersive elements (ibid.:
53–54). While capturing the nuances of meaningful journalism, such gamified news stories come
close to what Ian Bogost and colleagues (2010: 47) understand as “playable infographics” that, as a
special newsgames genre, “adopt infographics principles but add layers of gameplay around them”.
Nick Diakopoulos (2009) defines this approach as “process-oriented journalism”, which has later
been identified and explored in sophisticated multimedia and interactive packages of legacy news
media, such as the New York Times (e.g. Jacobson, 2012) and Al Jazeera (e.g. Ferrer-Conill, 2016).
Digital outlets as well as news aggregators, such as Google News, have also introduced gami-
fication to strengthen the loyalty of users by offering points, badges, leaderboards, and gifts to
“engage” them with the platform and “habitualize their visits” (Ferrer-Conill, 2016: 56). Ulti-
mately, gamified loyalty programs are one of the unprecedented ways “to track, measure, and
quantify audience activity” (Cohen, 2015: 108), with the ultimate goal to boost advertising rev-
enues and standardize the decision-making process by “accelerating the feedback loops between
media organizations and their audiences, and changing the value of the information that goes
into editorial decision making” (ibid.: 109). Several media outlets make use of these and similar
“playful products” to “further integrate and reward loyal readers”, but also by way of a “check
and balance” on journalistic practices (cf. Foxman, 2016).
Through gamification users are also brought closer to news production with an aim to engage
them in “crowdsourcing” as a free workforce that – at least that is the idea – has fun through a
game-like experience (Ferrer-Conill, 2016: 52). An example of such gamified engagement was
Huffington Post’s introduction of badges for “key areas of activity” in 2010 that included sharing
stories on social media, commenting on content, and moderating comments and was presented
as a “fun new way of recognizing and empowering our community” (Huffington, 2010). A year
earlier, the Guardian employed gamified crowdsourcing to speed up the analysis of a large set of
leaked documents and data concerning the UK’s parliamentary expenses (cf. Daniel and Flew,
2010) by using a “progress bar” to visualize the data so far covered and a “leaderboard” to show
the most dedicated workers among the users (Ferrer-Conill, 2016: 51).
There are also examples of employing gamification in the production of news by using vari-
ous types of digital tokens as incentives to motivate newsmakers according to their writing and
publication metrics based on their contribution to the overall digital news output (Ferrer-Conill,
2016: 54–55). A case study of the sports news website the Bleacher Report and its gamified pro-
duction system indicates that good performance in the game has little to do with journalism but
more with labor control and exploitation (cf. Ferrer-Conill, 2017). By considering professional
and expert debates, Ferrer-Conill discusses how “an open representation of journalistic perfor-
mance based on metrics can provide a reward that is not necessarily connected to journalistic
values” (ibid.: 5) and shows that respondents involved in gamification “do not make an assess-
ment about the quality of their pieces in terms of the metrics, but they do seem to do it for their
careers”. Some of the journalists involved were in precarious labor relations and engaged in gami-
fied production on the assumption that one day they will become employed (ibid.).
Within the larger “quantitative turn in journalism” (cf. Coddington, 2015; see also Cod-
dington, this volume, Chapter 17), gamification exemplifies the short-term dangers of favoring
immersive game dynamics rather than the larger public aims of journalism. Through particu-
lar features of some sort of instrumentalized gameplay, the four interrelated ways of gamifying
455
Igor Vobič
journalism indicate complexities in how contemporary journalism ideology is being rearticulated

and, as such, call for a broader study of how existing power relations and dominant business con-
cerns in the news industry are reflected through these specific game-based dynamics between
production and consumption in journalism.
Journalism – games: convergence and commodification

This review of the scholarly terrain found at the intersection of digital journalism and games
indicates contemporary articulations of historical contradictions within journalism – as a form
and production of knowledge. However, existing scholarship provides valuable empirical insights
into the multifaceted convergence of digital journalism and games but rarely tackles it critically
by reflecting on its materiality to reveal the limits of the emancipatory potential of these innova-
tions to provide inclusive and reflective accounts of journalism and news. Therefore, this part
attempts to provide a basis for a more comprehensive understanding of the complex phenomena
that emerge at the crossroads with journalism and by focusing on the theories of the audience
commodity and digital labor. The following paragraphs dissect the commodification processes
in the convergence between journalism and games, where the exploitation of digital labor con-
strains the noninstrumental, reflexive, and emancipatory potential of play within game journal-
ism, newsgames, and gamified journalism.
From Marx to Smythe: audience commodity

According to Karl Marx (1859/1904: 19), a “single commodity” is a crucial precondition for
the reproduction of capitalist societies. The commodity is “simultaneously embodying both
use value and exchange value” (ibid.: 54), where the former is the purpose of the entire system
of production, while it becomes a means of exchange “when their supply exceeds the measure
of consumption” (ibid.: 53). The exchange of commodities constitutes the “social metabolic
process” in which “the exchange of the special products of private individuals is the result of
certain social relations of production into which the individuals enter in this interchange of
matter” (ibid.: 55–56). Accordingly, commodities embody the dual character of work (cf. Marx,
1867/1887: 30–33). As discussed by Christian Fuchs and Sebastian Sevignani (2013: 240), this
special character can be conceptually and semantically separated into: (1) work, which is in contrast
“a more general concept” that encompasses processes of making use of technology to “transform
nature, culture and society in such a way that goods and services are created that satisfy human
needs” (ibid.); and (2) labor, which is “a necessarily alienated form of work” characteristic of class
societies where workers do not control and own the means and results of production (ibid.). In
capitalist societies, people often have their access to the means of production prevented and are
forced to sell their labor-power as a commodity to capitalists and to engage in an exchange that
is essentially unfair (cf. Prodnik, 2012: 279). The exploitative character of production rests on the
aim of the buyers of the labor-power commodity to extract value through the labor process – not
only exchange-, but surplus-value (ibid.).
By being concerned with the processes of commodification, Dallas Smythe (1977) discussed
the logics of mass media communication by conceptualizing the “audience commodity” cate-
gory. He argued that “the audience itself – its subjectivity and the results of its subjective creative
activity – is sold as a commodity” (Fuchs, 2012: 704). Namely, advertisers buy “the services of
audiences with predictable specifications who will pay attention in predictable numbers and at
particular times” to newspapers, radio, television, or other communication means (Smythe, 1977:
4). According to Smythe (ibid.), these audiences are commodities that are “dealt with” in markets
by “producers” (mass media) and “buyers” (advertisers), while establishing prices “in the familiar
456
mode of monopoly capitalism”. The materials produced by the media and transmitted to audi-
ences represent an “inducement” to “recruit” potential audience members and to maintain their
“loyal attention” through reading, listening or watching (ibid.: 5). When, in this manner,
workers under monopoly capitalist conditions serve advertisers to complete the pro-
duction process of consumer goods by performing the ultimate marketing service for
them, these workers are making decisive material decisions which will affect how they
will produce and reproduce their labour power.
(ibid.: 6)
From Smythe to Fuchs: digital labor

In recent years there has been a revival of the audience commodity debate through the discus-
sions on the exploitation of “digital labour” (cf. Fuchs and Sevignani, 2013). In this context,
the dominant capital accumulation model online is based on the exploitation of unpaid labor
through the creation of data from users’ content production, locations, relations, networks, and
activities, such as browsing and liking, which are sold as commodities to advertisers who then
target these very users through advertisements on the basis of their presumed interests and behav-
ior (ibid.: 237). In this setting, Christian Fuchs (2012: 707–708) reviews the digital labor debate
and identifies three elements involved in the exploitation of digital labor: (1) the “coercion” of
users to use commercial online platforms in order to engage in social life; (2) the “alienation”
of users since enterprises, not themselves, own the platforms and collect the profits; and (3) the
“appropriation” of users’ social relations, created content and other “prosumption” activities as
the “data commodity” is sold by online corporations to other parties, such as advertisers.
“Another layer of digital audience commodity”, argues Nicole Cohen (2015: 100), is entailed
by people’s engagement with and through digital journalism, which holds a “direct implication”
for the form and character of knowledge being produced. As the commodification of journalism
deepens, the “recasting of audiences as inputs for production signals” with the rise of crowd-
sourcing and the use of metadata, analytics, and algorithms goes hand in hand with the creative
demise of journalism through the precarization of labor. In a digital environment, the historical
reciprocity of production and consumption in capitalist news enterprises therefore becomes only
more apparent. Slavko Splichal and Peter Dahlgren (2016: 6) claim that it “not entirely correctly”
seemed that a clear boundary existed in mass media journalism between professional produc-
ers and passive consumers, while “today’s digital media inexorably blur the boundary between
mediated content production and consumption” (ibid.). Exploitation of digital labor in the news
industry – from the appropriation of user engagement on news websites or mobile applications
to more obvious ways of what is known as “participatory” journalism – has its roots in the
historical process of the commodification of modern journalism, discussed more than a century
ago by Carl Bücher (1901: 242–243), yet it also is being rearticulated and strengthened via what
appears as a “qualitatively novel type” of commodification (cf. Prodnik, 2014).
Digitization has helped to extend and intensify commodification “throughout places that
have hitherto been untouched by capitalist market”, argues Jernej Amon Prodnik (ibid.: 216),
proposing the concept of a “seeping commodification” that is able “to more or less successfully
mimic the activities that are distinctive of communication” and as such to “trickle down to all
the niches and activities of society and human lives” (ibid.). The exploitation of digital labor is
an example of commodification seeping through boundaries and into all pores of human exis-
tence – people’s subjectivity and the results of its creativeness. Christian Fuchs (2012: 734) argues
we are witnessing the “manifestation of a stage of capitalism, in which the boundaries between
play and labour have become fuzzy” and that the exploitation of “play labour”, also referred
457
Igor Vobič
to as “playbour”, has become “a new principle”. At this point, it might be useful to employ
the distinction between playing, working, gaming, and laboring that was introduced by Arwid
Lund (2014, 2015). He criticizes the “ideological concept” of playbour by drawing on the dual
character of labor and thereby argues that working and laboring are “distinguished into a trans-
historical and a historical, capitalist category”, and likewise “playing” and “gaming”, where the
former is the trans-historical category and the latter the historical one (Lund, 2014: 735). In this
context, Lund (2015: 67–68) conceptualizes playing as a noninstrumental, capricious, and impro-
vised activity with emancipatory potential, whereas gaming introduces rule-based competition
into playing and through instrumentalization deforms its very essence. By considering concep-
tual dynamics, Lund (cf. 2014: 772–798) distinguishes exploitive and instrumental gamebour from
emancipatory playwork as a possible “new form of anti-capitalist struggle” through communica-
tion, or journalism for that matter.
Playing, gaming, working, and laboring in digital journalism

The phenomena at the intersection between digital journalism and games touch on commodi-
fication processes and their manifestations through digital labor. Therefore, it seems a consider-
ation of the dynamics between playing, working, gaming, and laboring might be useful when
elaborating on commodification in digital journalism = games’ convergence and discussing how
digital labor is articulated through the phenomena of: (1) game journalism; (2) newsgames and
gamified news; and (3) the gamification of journalism practice.
Game journalism is a branch of journalism that for decades has been defined by commodities
through the “reveal/preview/review cycle” of new games (Fischer, 2012: 224–225), its digital
form presenting a prime example of deeply commodified journalism that relies on labor exploi-
tation. As game journalists stand between the game industry (advertisers) and game fans (audi-
ences), or as Foxman and Nieborg (2016) write, “between a rock and a hard place”, they perform
as “mediators” in the production of commodity value by blurring the lines between producers
and consumers but at the same time reconstructing these boundaries as clear lines (Carlson, 2009:
[1.4]). When such production of knowledge, or put better, commodity value is interrogated,
game journalism repairs its paradigm by breaking with the naïve traditions of the “enthusiast-
press” and relying on the traditional journalism ideology (Perreault and Vos, 2016) in order to
reproduce or reinforce existing social relations. The dubious practices and normative trade-offs
of game journalism are the work of “immaterial labourers” (Carlson, 2009: [6.6]) who – “as they
slip into, and help to widen, the notches created by the forces of late capitalism” (ibid.) – produce
desires and values and participate in the exploitation of their audiences – as digital laborers. The
mainstream game journalism sites aim, as Smythe would put it, to “recruit” game fans and try to
maintain their “loyal attention”, while data from their online reviews, comments, and metadata
of other online activities is on one hand used as inputs for future production decisions and on
the other sold as a commodity to the game industry, which are both prime advertisers and main
providers of information about game novelties. This circle of labor self-exploitation in game
journalism is most vividly embodied in the “gamebour” of those fans “who are willing to submit
game (p)reviews texts without any other compensation than perhaps the inspection copy of the
title they were reviewing” (Nieborg and Sihvonen, 2009).
It seems that newsgames and gamified news reflect recent attempts by the news industry to
respond to the “double-legged crisis” of journalism (Blumler, 2010), with the aim to both rein-
vent commodification models and reconnect with the public. Newsgames (cf. Bogost et al.,
2010) and long-form gamified journalistic formats (cf. Ferrer-Conill, 2016) indeed contest con-
ventional presentations of place and time and introduce more inclusive rhetoric and ways of
engagement. An idea of constructing meaning through newsgamers’ dissection of the “inner
458
workings” (Sicart, 2008) or through a more straightforward “highly playful narrative” of gamified
forms that captures “the nuances of a hard news story without needing to reduce the quality”
(Ferrer-Conill, 2016: 54) relates to the notion of “gaming” as rule-based and goal-aimed “seri-
ous leisure” (Lund, 2015: 66) that, by facilitating a certain aspect of news, meaningfully connects
people with public life. Newsgames and gamified news in this respect come close to the notion
of “gamework” with its “focus on the quality of use value and gaming for prestige” (Lund, 2014:
784) and reflect the perpetual conflict in journalism – between its public character on one hand
and the private interests of news media owners on the other (cf. Vobič, 2015). However, due to
production costs and newsroom identity issues, the industry has mostly been hesitant to experi-
ment in this manner, while some media players have opted to create forms that incorporate games
and entertainment, such as puzzles and quizzes (cf. Foxman, 2017). Such game-based products
embody what Lund (2015: 67) calls the “quantitative instrumentality” of “gamebour” – users are
engaged in game-based activities that have little to do with public-minded journalism and are
simultaneously commodified and exploited as digital labor.
Gamified journalism practice embraces commodified phenomena that connect production and
consumption by employing game elements to crowdsourcing or news production (cf. Ferrer-
Conill, 2016, 2017), where users or journalists’ data from their analytical, writing, and other
activities are commodified through a particular reward system with points, badges, and lead-
erboards. Gamification is a strategy to motivate users as providers of digital labor to further
engage in production as a free labor force that “economises the amount of paid occupational
labour in the journalistic production” (Splichal and Dahlgren, 2016: 10). Gamified produsage
or users’ gamebour as an articulation of “unpaid labour” has become “central to media outlets’
cost-saving strategies, making it increasingly difficult to earn a living as a journalist, particularly
for growing numbers of freelance writers” (Cohen, 2015: 113). At the same time, it has become
normalized in the digital media sphere that those who regularly produce as digital journalists are
often part of “a large network of unpaid contributors” (ibid.: 106). It is precisely in these highly
commodified contexts that the use of gamified interfaces that are “sustained by datafication”,
“turning behaviour into data and offering digital rewards” (Ferrer-Conill, 2017: 2), is strategically
employed to intensify the control over and exploitation of (un)paid labor. In order to disguise
the commodification of labor, “gamification adds to datafication the rhetoric of play”, while
the qualitative understanding of news production is being renegotiated (ibid.: 11–12). In the
wider context of digital labor, gamified journalism practice reproduces the understanding of
journalistic labor in terms of, as Hanno Hardt (1998: 211) discussed two decades ago, “routinized
technical tasks responding to specific commercial interests” such as the production of “non-idea”
centered narratives or “non-controversial contextual material” to help intensify the processes of
commodification.
Concluding remarks
This chapter has briefly discussed phenomena in the underexplored terrain of digital journalism
studies and elaborated on the existing scholarship at the intersection of digital journalism and
games – theoretical reconsiderations, analytical approaches, and empirical findings. Scholarship
on game journalism, newsgames, and gamification in journalism shows that studying these mul-
tifaceted notions and phenomena appears useful for exploring articulations of historical problems
of journalism in the digital age.
At the same time, the review of studies indicates that critical considerations of digital
journalism-games convergence are rare where, in particular, the material aspects of these com-
plexities are hardly considered. To supplement the scarce yet rich investigations into digital jour-
nalism and games, this chapter has sketched out a critical perspective by drawing on Marx’s
459
Igor Vobič
theory of capitalism, Smythe’s category of audience commodity, and Fuchs’s conceptualization

of digital labour. The introduction of the dynamics between playing, gaming, working, and
laboring introduced by Lund, however, calls for greater attention in future explorations of digital
journalism and games in terms of deepening the theoretical argument and with respect to using
it in original empirical research.
More theoretical and empirical explorations are required that would – at least try to – bridge
materialist approaches and those that interrogate subjectivist agency in order to comprehensively
gasp the peculiar but telling phenomena of journalism-games convergence. Attention needs to be
given not only to the areas discussed here but also to journalism and media education, where games
and gamification have become employed as new tools (e.g. Aayeshah, 2012) and strategies (e.g.
Leaning, 2015) to motivate students, future journalists in the deeply commodified digital world.
Further reading
For further reading in this terrain, the monograph Newsgames: Journalism at Play by Ian Bogost
et al. (2010) provides a valuable scholarly account of newsgames and related pressing issues found at
the crossroads of game and journalism studies. On gamification in journalism, Raul Ferrer-Conill
has written extensively; particularly his recent article in Television & New Media (2017) is insight-
ful. On game journalism, the contribution in the Journal of Games Criticism (2016) by Maxwell
Foxman and David B. Nieborg provides a thoughtful discussion of material and subjective aspects
of its network of ambivalences. In addition, those seeking to locate broader critical discussions on
digital journalism’s problems and their historical roots would enjoy Nicole Cohen’s article in The
Communication Review (2015) and the one by Slavko Splichal and Peter Dahlgren in the European
Journal of Communication (2016). Further, those in search for profound theoretical accounts and
critique in the larger field of communication to pinpoint pressing issues of digital journalism and
its boundaries should look at the open-access journal tripleC: Communication, Capitalism & Critique.
Particularly useful for building arguments in this chapter were the critical essays on seeping com-
modification by Jernej Amon Prodnik (2014), on digital labor by Christian Fuchs and Sebastian
Sevignani (2013), and on playing, gaming, working, and laboring by Arwid Lund (2014).
References
Aayeshah, W. (2012) “Playing with news.” Asia Pacific Educator, 22(1), 29–41.
Ahva, L. and Steensen, S. (2017) “Deconstructing digital journalism studies.” In B. Franklin and S. A.
Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. London: Routledge (pp. 25–34).
Amon Prodnik, J. (2014) “A seeping commodification.” tripleC: Communication, Capitalism & Critique, 12(1),
142–168.
Anderson, S. L. (2013) “Start, select, continue.” The Review of Communication, 13(4), 290–301.
Blumler, J. (2010) “Foreword: The two-legged crisis of journalism.” Journalism Studies, 11(4), 439–441.
Bogost, I. (2006) “Playing politics.” First Monday, Special issue number 7. Retrieved from http://firstmonday.
org/issues/special11_9/bogost/index.html
Bogost, I. (2011, August 9) “Gamification is bullshit.” The Atlantic. Retrieved from www.theatlantic.com/
technology/archive/2011/08/gamification-is-bullshit/243338/
Bogost, I., Ferrari, S. and Schweizer, B. (2010) Newsgames – Journalism at Play. Cambridge, MA: The MIT Press.
Braithwaite, A. (2016) “It’s about ethics in journalism?” Social Media + Society, 2(4), 1–10.
Bücher, C. (1901) Industrial Evolution, New York: Henry Holt and Company.
Burton, J. (2005) “News-game journalism.” Australian Journal of Emerging Technologies and Society, 3(2),
87–99.
Carlson, M. (2017) “Automated journalism.” In B. Franklin and S. A. Eldridge II (eds.), The Routledge Com-
panion to Digital Journalism Studies. London: Routledge (pp. 226–234).
Carlson, R. (2009). “Too Human versus the enthusiast press.” Transformative Works and Cultures, 2. Retrieved
from http://journal.transformativeworks.org/index.php/twc/article/view/98/93
460
Chess, S. and Shaw, A. (2015) “A conspiracy of fishes, or, how we learned to stop worrying about #Gamer-
Gate and embrace hegemonic masculinity.” Journal of Broadcasting & Electronic Media, 59(1), 208–220.
Coddington, M. (2015) “Clarifying journalism’s quantitative turn.” Digital Journalism, 3(3), 331–348.
Cohen, N. (2015) “From pink slips to pink slime: Transforming media labor in a digital age.” The Com-
munication Review, 18(2), 98–122.
Cote, A. (2015) “Writing ‘gamers’.” Games and Culture, 13(5), 479–503.
Daniel, A. and Flew, T. (2010, November) “The Guardian reportage of the UK MP expenses scandal.” Paper
presented at the Communications Policy and Research Forum, Sydney. Retrieved from http://eprints.qut.
edu.au/38701/2/38701.pdf
de la Peña, N., Weil, P., Llobera, J., Giannopoulos, E., Pomés, A., Spanlang, B., Friedman, D. Sanchez-Vives,
M. V. and Slater, M. (2010) “Immersive journalism.” Presence, 19(4), 291–301.
Diakopoulos, N. (2009, February 16) “Functional and cultural tensions and opportunities for games in jour-
nalism.” NewsGames. Retrieved from http://jag.lcc.gatech.edu/blog/2009/02/functional-andcultural-
tensions-and-opportunities-for-games-in-journalism.html
Ekström, M. (2002) “Epistemologies of TV journalism: A theoretical framework.” Journalism, 3(3), 259–282.
Ferrer-Conill, R. (2016) “Points, badges, and news.” Comunicació, 33(2), 45–63.
Ferrer-Conill, R. (2017) “Quantifying journalism?” Television & New Media, 18(8), 706–720.
Ferrer-Conill, R. and Karlsson, M. (2016) “The gamification of journalism.” In H. Gangadharbatla and D. Z.
Davis (eds.), Emerging Research and Trends in Gamification. Hershey: IGI Global (pp. 356–383).
Fisher, H. D. (2012) Don’t Let the Girls Play: Gender Representation in Videogame Journalism and the Influence of
Hegemonic Masculinity, Media Filters, and Message Mediation. Doctoral dissertation, Ohio University, Ath-
ens. Retrieved from https://etd.ohiolink.edu/rws_etd/document/get/ohiou1332372302/inline
Fisher, H. D. (2015) “Sexy, dangerous – and ignored.” Games and Culture, 10(6), 551–570.
Foxman, M. (2015, February 17) “Play the news.” Tow Center for Digital Journalism. Retrieved from https://
towcenter.org/wp-content/uploads/2015/02/PlayTheNews_Foxman_TowCenter.pdf
Foxman, M. (2016) “The playful newsroom.” First Monday, 22(1–2). Retrieved from http://firstmonday.
org/ojs/index.php/fm/article/view/7260
Foxman, M. and Nieborg, D. B. (2016) “Between a rock and a hard place.” Journal of Games Criticism, 3(1), 1–51.
Fuchs, C. (2012) “Dallas Smythe Today.” tripleC: Communication, Capitalism & Critique, 10(2), 692–740.
Fuchs, C. and Sevignani, S. (2013) “What is digital labour? What is digital work? What’s their difference?
And why do these questions matter for understanding social media?” tripleC: Communication, Capital-
ism & Critique, 11(2), 237–293.
Gambarato, R. R. and Tárcia, L. (2016) “Transmedia strategies in journalism.” Journalism Studies,18(11):
1381–1399.
Hall, J. (2003, April 10) “Ethics in video game journalism.” Online Journalism Review. Retrieved from
http://www.ojr.org/ojr/ethics/1049994303.php
Hardt, H. (1998) Interactions. Lanhman: Rowman & Littlefield Publishers.
Huang, L. (2016, October) “Meaningful gamification for journalism students to enhance their critical
thinking skills.” Paper presented at the 9th European Conference on Games Based Learning, Steinkjer.
Huffington, A. (2010, June 29) “Introducing HuffPost badges.” HuffingtonPost. Retrieved from www.
huffingtonpost.com/arianna-huffington/introducing-huffpost-badg_b_557168.html
Jacobson, S. (2012) “Transcoding the news.” New Media & Society, 15(5), 867–885.
Jagoda, P. (2017) “Videogame criticism and games in the twenty-first century.” American Literary History,
29(1), 205–218.
Leaning, M. (2015) “A study of the use of games and gamification to enhance student engagement, experi-
ence and achievement on a theory-based course of an undergraduate media degree.” Journal of Media
Practice, 16(2), 155–170.
Lund, A. (2015) “A contribution to a critique of the concept of playbour.” In E. Fischer and C. Fuchs (eds.),
Reconsidering Value and Labour in the Digital Age. Basingstoke: Palgrave Macmillan (pp. 63–80).
Lund, A. (2014) “Playing, gaming, working and labouring.” tripleC: Communication, Capitalism & Critique,
12(2), 735–801.
Malik, A. and Shapiro, I. (2017) “What’s digital? What’s journalism?” In B. Franklin and S. A. Eldridge II
(eds.), The Routledge Companion to Digital Journalism Studies. London: Routledge (pp. 15–24).
Marx, K. (1867/1887) Capital – Volume I. Moscow: Progress Publishers.
Marx, K. (1859/1904) A Contribution to the Critique of Political Economy. Chicago, IL: Charles H. Kerr &
Company.
McKernan, B. (2013) “The morality of play.” Games and Culture, 8(5), 307–329.
Mortenstein, T. E. (2016) “Anger, fear, and games.” Games and Culture, 1–20. doi:10.1177/1555412016640408
461
Igor Vobič
Nelson, M. and Mateas, M. (2007) “Towards automated game design.” In R. Basili and M. T. Pazienza (eds.),
AI*IA 2007: Artificial Intelligence and Human-Oriented Computing. Berlin: Springer (pp. 626–637).
Nerone, J. (2013) “The historical roots of the normative model of journalism.” Journalism, 14(4), 446–458.
Nicholson, S. (2015) “A recipe for meaningful gamification.” In T. Reiners and L. C. Wood (eds.), Gamifica-
tion in Education and Business. Heidelberg: Springer (pp. 1–20).
Nieborg, D. B. and Sihvonen, T. (2009, September) “The new gatekeepers.” Proceedings of the 2009
DiGRA International Conference. Retrieved from www.digra.org/wp-content/uploads/digital-library/
09287.29284.pdf
Perreault, G. and Vos, T. (2016) “The GamerGate controversy and journalistic paradigm maintenance.”
Journalism, 19(4), 553–569.
Prodnik, J. (2012) “A note on the ongoing processes of commodification.” tripleC: Communication, Capital-
ism & Critique, 10(2), 273–301.
Ribbens, W. and Steegen, R. (2012) “A qualitative inquiry and a quantitative exploration into the meaning
of game reviews.” Journal of Applied Journalism & Media Studies, 1(2), 209–229.
Schudson, M. (2017, February 23) “Here’s what non-fake news looks like.” Columbia Journalism Review.
Retrieved from www.cjr.org/analysis/fake-news-real-news-list.php
Sicart, M. (2008) “Newsgames.” In S. M. Stevens and S. J. Saldamarco (eds.), Entertainment Computing –
ICEC 2008. Berlin and Heidelberg: Springer (pp. 27–33).
Smythe, D. (1977) “Communications: Blindspot of Western Marxism.” Canadian Journal of Political and Social
Theory, 1(3), 1–27.
Splichal, S. and Dahlgren, P. (2016) “Journalism between de-professionalisation and democratisation.” Euro-
pean Journal of Communication, 31(1), 5–18.
St. Clair, J. (2015) “Doing it for real.” Coolabah, 16, 122–142.
Teixeira, C., Carcalho, B., Arga, J., Martins, V., Lins, A., Soares, M. and Neves, A. (2015b) “Newsgames.” In
A. Markus (ed.), DUXU 2015: Design, User Experience, and Usabilitxy: Interactive Experience Design. Berlin:
Springer (pp. 292–302).
Teixeira, C., Carcahlo, B., Soares, M., Neves, A., Lins, A., Arga, J., Martins, V. and Soares, G. (2015a) “Usabil-
ity and gameplay in newsgames.” Procedia Manufacturing, 3 (AHFE 2015) (pp. 6060–6067).
Therrien, C. (2014/2016) “Immersion.” In M. J. Wolf and B. Perron (eds.), The Routledge Companion to
Video Games Studies. New York, NY: Routledge (pp. 451–458).
Thominet, L. (2016) “10/10 would review again.” Technical Communication Quarterly, 25(3), 167–177.
Treanor, M., Blackford, B., Mateas, M. and Bogost, I. (2012, May) “Game-O-Matic.” Proceedings of the Third
workshop on Procedural Content Generation in Games, Raleigh.
Treanor, M. and Mateas, M. (2009, September) “Newsgames.” Paper presented at the Digital Games Research
Association Conference, London.
Vobič, I. (2015) “Osiromašenje novinarstva.” [Impoverishment of journalism]. Javnost – The Public, 22(sup-
plement), S28–S40.
Vobič, I. and Dahlgren, P. (2013) “Reconsidering participatory journalism in the internet age.” Medijska
istraživanja, 19(2), 9–30.
Vobič, I., Dvoršak, L. and Vtič, M. (2014) “Digital games and journalism.” Teorija in praksa, 51(1), 123–142.
Waltz, S. P. and Deterding, S. (eds.). (2014) The Gameful World. Cambridge, MA: The MIT Press.
Williams, D. (2003) “The video game lightning rod.” Information, Communication, & Society, 6(4), 523–550.
462
35
NATIVE ADVERTISING AND
THE APPROPRIATION OF
JOURNALISTIC CLOUT
The use of native advertising has sparked a heated debate within traditional news media. While
similar formats have a long history within journalism, this new iteration furthers the blurring
of boundaries between news and ads by producing ads that look and feel like news but that are
clearly labeled as advertising. The novelty of native advertising is that it advocates for openly
merging commercial and editorial content, aggravating an existing tension between the pro-
fessional and commercial logics of journalism. This open relationship between journalists and
marketers calls for revisiting the traditional narrative that sustains journalistic autonomy.
Of all the metaphors comprising the social imaginary of contemporary journalism, the figure
of the wall is by far the one that has contributed the most to establish the symbol of journalistic
autonomy. During the golden age of journalism, the wall exemplified the organizational division
between the editorial and commercial operations of news media (Coddington, 2015). A strong,
impenetrable wall meant that journalists could escape political and commercial pressures and
focus on objective, autonomous, and ethical reporting (McNair, 1998). The image of watchdog
reporting is based upon this premise. The so-called fourth state cemented its legitimacy by pledg-
ing allegiance to the public (Hampton, 2010; Lewis et al., 2008), an allegiance that could only
be sustained by an independent group of newsworkers that adhered to the norms and values of
journalism (Kovach and Rosenstiel, 2007).
Of course, news media organizations do not conform to a monolithic form of journalism, and
the metaphoric wall ranged from reinforced concrete and steel to paper-thin porous materials
(Bagdikian, 2004; Shoemaker and Reese, 2014). While there are several accounts of permeabil-
ity where political and commercial actors influenced editorial content, journalistic autonomy
remained relatively unproblematized during the second half of the twentieth century. However,
during the last two decades, journalists have experienced an accelerated challenge to their auton-
omy in the newsroom (Sjøvaag, 2013). This happened mostly on three fronts: first, on the tech-
nological front, where new publishing channels, new digital means of production, and technical
innovations often dictate how information is gathered, formatted, and disseminated, reducing
journalists’ control over their own material; second, on the professional front, where the demand
for new skills and the power increase of non-journalistic actors – technologists, business people,
and PR practitioners – within newsrooms have reduced the authority of journalists to control
what news is; and third, on the commercial front, where the decline of sales and advertising
463
revenue have spurred news organizations to find alternative sources of income, often deriving in
commercial pressures on the editorial front (Shoemaker and Reese, 2014).
Commercial news media are still trying to find a viable business model that can keep their
operations profitable: on the one hand, the revenue of traditional advertising decreasing, mostly
due to ineffectiveness of former ad types and advertisers migrating to bigger audience broker
platforms like Facebook and Google, and on the other hand, the digital business still developing,
trying to find a sustainable revenue portfolio. Thus, native advertising has gained notoriety across
various types of digital content (Couldry and Turow, 2014; Matteo and Dal Zotto, 2015) as well
as within journalism (Carlson, 2015). Within the context of news media, native advertising is
defined as “a form of paid content marketing, where the commercial content is delivered adopt-
ing the form and function of editorial content, with the attempt to recreate the user experience
of reading news instead of advertising content” (Ferrer-Conill, 2016: 905). The relevance of
native advertising stems from the fact that it (a) openly tries to make advertising look like edito-
rial content, (b) has spread rapidly across several digital publishers, including legacy media like
the New York Times, the Guardian, and the Washington Post, and (c) is championed by chief edito-
rial roles at those publications. In this sense, native advertising not only questions the insulating
capacities of the wall metaphor, but it effectively transforms the wall into a sliding door. For
news media, this signifies a potential challenge to the autonomy that contemporary journalism
is built upon. Our goal is to provide an overview of native advertising as an emerging practice
in the context of news media and to problematize its adoption both in terms of discursive and
potential outcomes.
With this chapter, we aim to critically address the phenomenon of native advertising from
three different angles. First, a historical and conceptual discussion on native advertising brings
to the fore earlier models of native advertising that were referred to with different terms, as well
the current understanding of what native ads are. Second, the current normative debate about
native advertising is teased out to shed light on the rhetorical shift that seems to be permeating
the news industry. To publishers, this kind of advertising creates important streams of revenue; to
advertisers, it appropriates journalistic reputation, transforming their coexistence from an ago-
nistic relationship to a symbiotic one. Finally, the potential challenges and benefits of adopting
native ads vis-à-vis digital journalism are presented, aiming to demystify a commercial initiative
that could lead to economic viability but also could question journalistic allegiance to the citizen
and threaten the legitimacy and autonomy of contemporary news media.
The many names of native advertising

Almost paradoxically, it was a stronger commercialization of journalism that proved to be vital for
the professionalization and independence of newspapers against political influence in the middle
of the nineteenth century (Conboy, 2004). A commercially robust press could afford to be free,
and the need to appeal to a wider population curtailed partisanship. This led to a long-standing
use of advertising as a revenue stream, mostly in the form advertorial in print and banners
in digital. Traditional advertising was clearly demarcated as advertising and attempted to con-
vey commercial content without aspiring to mimic editorial content or blurring its advertising
nature. Any covert transgression of these boundaries has normally been considered as unethical
(Hamilton, 2004). The current trend, however, seems to be moving from that position.
In itself, native advertising can be traced to the traditional advertorial, a form of advertising
appearing in the 1940s that loosely resembled editorial content (Kim, 1995). The advertorial
already brought to the fore issues of transparency and deceptiveness (Ellerbach, 2004; Palser,
2002). Advertising industry trade literature identifies the traditional advertorial clearly as an
advertisement that attempts to sell a product, whereas native advertising is not as blatant and
464
Native advertising
sales-oriented and instead takes an informative approach that looks even more like news (Joel,
2013; Sharma, 2015). Thus, native advertising seems to differ from older forms of advertising
because it conflates both editorial and commercial perspectives into one single narrative that is
not overtly commercial. Interestingly, even if native advertising is the newest expression of an old
trend of journalism commercialization, it is rapidly spreading across the digital journalism land-
scape (Nielsen, 2016). The explanation may lie behind the fact that native advertising has been –
if we understand successful in terms of profitable sources of revenue – successfully adopted by
new media actors, who openly welcome this business model, mixing advertising, entertainment,
and news (Campbell and Marks, 2015).
As it gains popularity in the current news media market, the term native advertising competes
with other terms that seem to be referring to similar practices: paid content (Herbert and Thur-
man, 2007), paid-for-news (Erjavec and Kovačič, 2010), look-alike-news (Schechter, 1997), content
integration (Li and Leckenby, 2004), editorial integration (Stamm and Underwood, 1993), embedded
marketing (Hackley and Tiwsakul, 2006), and branded content (Carpenter, 2010). However, these
practices seem to differ in terms of how transparent – or indeed how opaque – they are when
disclosing the origin of content. Even when labeled with the umbrella term sponsored content
(Cole and Greer, 2013), this content is aimed to suit the publisher’s editorial line while offer-
ing content that aims to capture the readers’ attention without openly attempting to promote
products or services. The signifying factor is not only the appearance but also the intent to create
cohesive, attractive content that conveys a message without directly attempting to sell a product;
such content informs the reader while also delivering branded content that can be shared and
spread to other networks. In other words, native advertising is not designed to sell items openly
but to raise awareness and draw attention to topics advertisers and readers are invested in. Thus,
it provides news that is appealing to both the reader and to the marketer.
The difficulty in distinguishing between formats again raises the question of autonomy and
legitimacy. In his extensive literature review and qualitative study with journalists, Macnamara
(2014) states that native advertising, along with any other commercial and PR transgression into
editorial content, is generally seen as a negative challenge to journalists’ autonomy, worsening
journalistic content as well as raising transparency questions. This is supported by Edström’s
(2015) study on Swedish journalists. In her study, journalists express concerns and ethical dilem-
mas in the use of native advertising, fearing loss of credibility and trust. Such an approach should
provoke skepticism among legacy news media outlets. In the next section, we discuss how native
advertising is being widely adopted in the current digital journalism environment and the rapid
evolution of discursive tactics.
Shifting discourse – from agonism to symbiosis

The pervasiveness of advertising in news media that led to the professionalization of journalism also
became one of its weaknesses, as commercial success became necessary for journalistic endeavor. In
digital journalism, this development became even more apparent when legacy media outlets made
the decision to offer their content online for free, in the hopes that advertising alone would become
a viable business model. However, as news media sites lost popularity in favor of social media plat-
forms, advertisers followed the users and advertised elsewhere. Furthermore, the irruption of new
media platforms such as the Huffington Post, Gawker, or BuzzFeed strengthened innovative forms
of advertising (Couldry and Turow, 2014). Native advertising became a viable advertising stream
for these new media outlets, and while they were not direct competitors of legacy news media, their
economic success quickly became of interest to traditional news organizations.
In early 2013, most legacy news media openly advocated against native advertising (Ferrer-
Conill, 2016). The agonistic discourse – in the Foucauldian sense – by which contemporary
465
news media used the metaphor of the wall to keep advertising in check was still brandished to
reinforce the moral authority of well-established news organizations. Advertising, however, was
as necessary to the enterprise as was the need to keeping it separate from the editorial content.
Advertising was supposed to fund the news operation by coexisting with news in print and
online, and the struggle to present both news and advertising in the same product was the key
factor legitimating journalism. Without the presence of external influences that could shape news,
there was no need for maintaining the claim that news is objective and therefore uninfluenced
by these pressures.
Eventually, the struggle narrative gave way to a collaborative – if problematic – relationship
with marketers. This shift can be exemplified by a statement by Gerard Baker, editor of the Wall
Street Journal, who claimed that “(t)he clear delineation between news and advertising is becom-
ing more and more blurred. We have to resist that” (Pompeo, 2013). He labeled native advertising
a: “Faustian pact”. Only six months after those remarks, the Wall Street Journal published its first
native ad, a story titled “Cocainenomics”1 detailing how Colombian drug-lords impose their
business globally. The native ad was sponsored by Netflix in support of the show Narcos (Sebas-
tian, 2014a). In an almost identical turn of events, Alexandra MacCallum, the New York Times’
audience development editor, when discussing the commercial practices of BuzzFeed, said, “(t)
he mission of The New York Times is about the best journalism in the world and giving people
accurate, timely information. I don’t think that BuzzFeed is competing in that space” (Moses,
2015). The attempt here was to establish a distance from the newcomer by adopting the role of
a superior, legitimate news producer. However, the New York Times circulated in 2014 an internal
innovation report in which the leadership stated that
“(t)he very first step should be a deliberate push to abandon our current metaphors of
choice – ‘The Wall’ and ‘Church and State’ – which project an enduring need for divi-
sion. Increased collaboration, done right, does not present any threat to our values of
journalistic independent.
(2014: 61)
The same year, the New York Times published its first native ad in support of a Netflix show, this
time Orange Is the New Black. The ad, titled “Women Inmates: Why the Male Model Doesn’t
Work”,2 focused on the raising population of female inmates in American penitentiaries (Sebas-
tian, 2014b).
New media actors – Huffington Post, BuzzFeed, and Gawker among others – employed native
advertising successfully for advertisers and the audience, gaining traction with younger audiences
without harming their reputation (Campbell and Marks, 2015). Since it worked for them with-
out any apparent loss of credibility, legacy news media tried to adopt the format. One by one,
almost all American mainstream traditional news outlets started adopting native ads. Since then, a
similar development has occurred in several other Western countries, including Spain, Germany,
Norway, Sweden, the UK, and Israel (Knudsen et al., 2017).
The speed with which native advertising was adopted is surprising. While changes to journal-
istic norms and practices have traditionally happened over an extended period of time (Rodgers,
2007; Stoker, 1995), such a rapid change in practice – and the transition from decrying native
advertising to adopting it – required revisiting discourse of change. This narrative sees an agonistic
discourse, separating advertising from editorial content, quickly evolved into a symbiotic discourse
that welcomed advertising in collaboration rather than as a necessary evil that ‘pollutes’ journal-
ism’s editorial purity. In order to justify the adoption of native advertising, high-ranking figures in
news media were outspoken in favor of native advertising, playing down the metaphoric wall and
the commercial and editorial divide (for more on the discourse change, see Ferrer-Conill, 2016).
466
Native advertising
The novelty here is not in news media and advertising coexisting in a symbiotic relationship.
This was long embedded in the commercial nature of newspapers, as marketers bought exposure
in news media to reach an audience that advertisers could target, allowing newspapers to sustain
their operations. This remains. However, the previous discourse around this relationship was
distinctly agonistic. There was a need to make the public believe these conflicting worlds were
separate. Over time, as advertisers found larger audiences in search engines and social media,
traditional news media could no longer compete in audience size and precision – leading to a
crisis within the newspaper industry. Native advertising, however, provided the mechanisms by
which news organizations could leverage their existing credibility in exchange for establishing a
more symbiotic editorial/advertising relationship. This required shifting the discourse from news
media toward marketers so that it would be apparent they could collaborate with news media –
the agonistic discourse giving way to a symbiotic one, thereby securing a way to maintain and
increase advertising revenues.
As is the case with most discursive shifts, the transition requires a well-planned balance for it
to work. Decades of cementing the wall cannot be undone overnight, and hence, adopting native
advertising requires negotiating certain boundaries that had differentiated ads from journalistic con-
tent. In the following section, we detail how native advertising materializes in journalistic contexts.
Appropriation of journalistic reputation

The discursive construction of native advertising is particularly interesting because it has
attempted to change a well-established norm of journalism – the metaphor of the wall – in a
very short time, allowing the public to witness the process. Just as Carlson (2015: 862) suggests,
it is also imperative to monitor and critique the work of norm entrepreneurs who
are developing not just new practices for news, but, more importantly, new normative
underpinnings justifying these practices to the point of redefining what constitutes the
field of journalism.
What are then the underpinnings of native advertising and its fast deployment? If there is in fact
a transition from an agonistic to a symbiotic relationship, it must then enforce the need to satisfy
both partners: advertisers and news organizations.
From the perspective of the advertisers, mixing advertising and editorial content is viewed as a
way to counter traditional advertising’s limited reach and effectiveness in an increasingly cluttered
and competitive environment (Goodlad et al., 1997; van Reijmersdal et al., 2005). Furthermore,
banner fatigue and the increased use of ad blockers render traditional online ads less effective at
reaching the audience. Since native advertising appears as “regular” content, it slips through the
filters of ad blockers (Dowling, 2016). Finally, and most importantly, it is the blurring of bound-
aries between ads and editorial content that makes native advertising more attractive. On the one
hand, simply appearing in an article format resembling editorial content increases readership,
exposure, and reach, making it more effective (Wojdynski, 2016; Wu et al., 2016). On the other
hand, the degree of trust associated with branded websites and editorial content such as newspa-
per articles is much higher than, for example, online banner ads or advertising on social networks
(Campbell and Marks, 2015). Thus, the value lies in the association of an advertiser with a trusted
publisher that has established legitimacy with an audience. Embedding their message in the form
of an article offered by a trusted newspaper, advertisers effectively appropriate its reputation and
influence and borrow credibility from its editorial content (Cameron and Curtin, 1995).
For news organizations, native advertising primarily represents a new form of revenue. The
business of native advertising is growing and generates billions of dollars (Cameron and Ju-Pak,
467
2000; Campbell and Marks, 2015; Wojdynski and Golan, 2016). Native advertising can be under-
stood as a way to retain advertisers and slow the exodus to social media platforms and to capital-
ize on advertising more effectively. According to Erjavec (2004), advertising rates for publishing
what she calls promotional news are about three times higher than regular advertising. Furthermore,
it can be argued that a well-designed and profitable native advertising strategy could also attract
readers – due to a more enticing form of writing and producing content – as well as curbing the
influence from older types of advertising and other pressures from other public relations actors
(Schauster et al., 2016). News organizations could also claim that native advertising can also be
good reporting and, at the same time, fund more investigative reporting that is not part of a com-
mercial scheme. Additionally, considering the current job market and the loss of many journalist
positions in the media sector, native advertising specialists could be unemployed or freelancing
journalists that have the necessary storytelling skills to dress up a commercial message in editorial
clothing with the necessary sensibilities of journalistic content (Hallahan, 2014; Wright, 2016).
At a simple glance, most enthusiastic newsworkers may be able to outline a win-win situa-
tion. However, the problem with the symbiotic discourse is that it only operates at the theoretical
level; the same win-win situation is much less balanced than what the positive narratives appear
to suggest. At a practical level, the majority of risk of native advertising is placed on the publisher
and its reputation and credibility, while the advertiser is barely exposed to potential drawbacks of
the relationship. If the public believes journalistic independence is sustained, then the legitimacy
of the democratic values of journalism remains intact, as does the publisher’s reputation (Soloski,
1989). However, if this sense of independence is in question, it becomes difficult to claim the
audience’s interests are central to the operation.
Past cases show there is simply too much room for potential damaging outcomes. Following
Carlson’s (2015) sense making of the case of the Atlantic, alongside its use of native advertising,
there were several other potential sources of controversy. The reputation of the client – the
Church of Scientology – seemed to draw a big part of the reaction, including claims that sell-
ing space to such a marketer contests the reputation and authority of the Atlantic. Additionally,
another source of critique was the quality of the text itself. The excessive praise and propagan-
distic tones of the ad seemed to betray the readers’ expectations of quality content within the
Atlantic’s pages, with negative reaction directed at the magazine.
Any native advertising published in news media risks exposure to similar criticism and out-
comes, while the advertiser barely faces criticism. This is because the advertiser fulfills what is
expected of it, which is to advertise its message. Conversely, the news outlet introduces a new
behavior that transgresses the norms and values of what journalism is supposed to be, despite
attempts to shift the discourse from agonistic to symbiotic. Legitimacy and trust may be ques-
tioned once autonomy is compromised (Amazeen and Muddiman, 2017). The risks also may
not materialize, and instead some news organizations (including new digital-only outlets) are
able to capitalize on this partnership. The symbiotic discourse builds on the notion of a win-win
relationship; however, we argue news media are the ones with higher stakes, as losing trust and
credibility might have a deeper impact on their legitimacy.
In the following section, we address how the native advertising discourse manifests in reality.
We turn to the contradictions within practice and the negotiations that news organization need
to go through in order not to trigger criticism.
Shiny camouflage
The proponents of native advertising argue that such forms of revenue carry no drawbacks as
long as readers understand they are reading commercial content; to advertisers, native is attractive
because it mimics editorial content in such a way that readers barely identify it as commercial
468
Native advertising
content, hence appropriating journalistic clout. The critics claim disguising ads as news not only
deceives the readers but also potentiates the loss of credibility news media claim to have. Native ads
camouflaged as news transfer some of the newspapers’ reputation to the advertiser, for good or ill.
What seems clear is that native advertising is by definition a contradictory format, at least in
the way proponents claim it should work. As Wu et al. (2016: 1504) suggest, “the effectiveness
of native advertising is connected to audiences’ unawareness of its advertising nature”. The same
study found that user evaluation of the content was more positive when it was published in a
high-credibility media and a high-credibility company. Thus, native advertising will be most
effective when it is incorporated in the most reputable and credible media. This is why native
advertising is appealing for advertisers, because it adopts the publisher’s format and readers associ-
ate the ads with the publisher, transferring their influence from the advertiser to the publisher.
To maintain the image of autonomy, publishers claim the disclosure of the ads is visible enough
for readers to see. As long as the disclosure is visible enough, native advertising, they claim, is not
a problem. The issue is that these claims are contradictory. Even though it is the most prevalent
practice, disclosing native advertising as such goes directly against its intended effect. On the
one hand, native ads try to look as similar as possible to regular news content, with the aim of
making readers believe they are news. On the other hand, publishers are expected to label the
pieces as native ads as a warning to readers. These two characteristics point toward one of the
main issues of native advertising, deceptiveness (Wojdynski, 2016). If publishers really wanted
to disclose native ads as ads, then the effort to mimic editorial content would not be necessary.
Furthermore, if native ads were clearly identified as ads, then their value would not be three times
that of regular ads.
Thus, it is in the interest of the publication to make sure that the balance between mimick-
ing and disclosure is as subtle as possible, just enough to be profitable and not lose the readers’
confidence. This seems to be a paradoxical shiny camouflage; inconspicuous enough that readers
mistake advertising for editorial content but shiny enough to claim that they were warned.
An added difficulty is the different approaches to implementation that each publication
adopts. While the theoretical discussion of how native advertising can influence journalism is
done at a general level, the empirical analysis depends entirely on the cases to study. News orga-
nizations like the New York Times, Washington Post, or Guardian have their own in-house dedi-
cated native advertising departments – T Brand Studio, WP Brand Connect, and Guardian Labs,
respectively – but many others rely on external marketers to create their native ad campaigns,
with even less autonomy in negotiating the final product. In the next section, we address how
researchers have tried to study native advertising and some of the difficulties that they have find.
Measuring native advertising

Scholars researching native advertising and journalism face a difficult object of study. Theo-
retically, native advertising becomes relevant through the discursive construction of journalism,
its norms and values, and what its metaphorical constructs – like the wall – are experiencing.
Methodologically, the toolset seems to focus on qualitative approaches to gauge the attitudes
and opinions of producers and newsworkers. The audience is often approached through experi-
mental methodologies that target their cognitive and affective reactions to distinguish between
editorial content and native advertising and the possible impact they may have on consumption.
Ultimately, the assessment becomes an empirical exercise that offers limited room for generaliza-
tion. The empirical studies done in native advertising follow three different camps: production,
content, and consumption.
With the production process, Artemas et al. (2016) found that the interactions between busi-
ness and editorial and staff members of news organizations in the US increased as the economic
469
situation worsened. The ‘wall’ weakened as the commercial pressures grew. Schauster et al. (2016)
found that journalists, advertising, and public relations executives agreed that there are eco-
nomic benefits behind native advertising while raising concerns about the deceptive nature of
the format, the lack of transparency, and the fact that readers are unaware of its promotional and
persuasive nature. Additionally, respondents attributed the responsibility of achieving a balanced
product to others than themselves, with journalists the group expressing more concerns. The
last finding is supported by similar studies (Cameron and Curtin, 1995 and Edström, 2015),
even though some producers believe the majority of readers could identify native ads among
the editorial content (Goodlad et al., 1997). An ethnographic study in the US (Eckman and
Lindlof, 2003) found resistance from the news department of a mid-sized daily newspaper when
the advertising department requested a stronger resemblance between editorial and commercial
content, claiming that journalistic tradition was stronger than economic hardship. In a similar
ethnographic study in Slovenia, Erjavec and Kovačič (2010) found a different outcome; it was
the advertisers who shaped all stages of journalistic production – gathering, selection, writing,
and editing – in the creation of advertorials. Consequently, journalistic ideals were not, in this
case, a major consideration during the process of production. In Estonia, Harro-Loit and Saks
(2006) suggest smaller media markets are more susceptible to blurred lines between the edito-
rial and advertising departments as the competition for limited resources increases. Interestingly,
when looking at ethical guidelines in various organizations in Finland and the US, Ikonen et al.
(2017) found that sponsored content or native advertising was rarely mentioned, and neither were
transparency and disclosure procedures.
When it comes to content, Cameron and Curtin (1995) found advertorials present in the US,
while Goodlad et al. (1997) found them in the Scottish press. Ferrer-Conill (2016) found uneven
patterns of introducing native advertising in the US, the UK, and Spain, while Sweden’s legacy
media had not then started using native ads. Edström (2015) found native advertising practices
in Sweden emerging later on. The content most commonly associated with native advertising
seems to cover travel, lifestyle, and health (Cohen, 2002; Erjavec, 2004; Harro-Loit and Saks, 2006;
van Reijmersdal et al., 2005; Wojdynski and Golan, 2016). Cameron and Ju-Pak (2000) con-
cluded that many advertorials were not following recommendations to signal their distinction as
advertising – different typestyle and typesize, design, content, context – thus making it unneces-
sarily difficult for readers to separate editorial and commercial content.
On the user dimension, several experimental studies (Hoofnagle and Meleshinsky, 2015;
Kim, 1995; Wojdynski, 2016; Wojdynski and Golan, 2016; Wu et al., 2016) have established
that native advertising is often difficult to identify. Readers generally fail to correctly identify
native advertising/advertorials as advertising. Furthermore, their ability to recognize articles as
advertising usually indicates a lower assessment of article quality and news credibility for the
publisher, decreased intent to share the news story, and negative attitude toward the advertiser.
Incidentally, Dix and Phau (2009) found that news consumers were more concerned with the
harmful effects on media credibility and confidence in advertising than agencies and advertis-
ers. However, Wojdynski and Evans (2016) showed that clearly labeling or wording the text
as advertising did not greatly increase respondents’ ability to correctly identify the piece as
advertising. When it comes to credibility, Cameron (1994) found that readers tend to experi-
ence higher credibility from editorial content in comparison with advertorials. In a similar
way, van Reijmersdal et al. (2005) found that the audience perceived advertorials to be more
misleading and less acceptable than editorial content. This is supported by Wu et al. (2016),
who claim that identifying advertising as such triggers users to decode the content more criti-
cally. However, van Reijmersdal et al. (2005: 51) conclude that “the more an advertising mes-
sage is intertwined with its carrier, the more readers are inclined to respond positively”. This
suggests that when readers manage to differentiate between editorial and commercial content,
470
Native advertising
they favor editorial content, but as both become more entwined, the readers favor the mes-
sage regardless of its commercial nature. This would explain why advertisers are interested in
influencing editorial formats and content, and why the wall is such an important element of
contemporary journalism.
Conclusions
In this chapter, we have argued that despite the well-established journalistic norm of keeping the
commercial and editorial endeavors of news organizations separate, the adoption and prolifera-
tion of native advertising has been out of the ordinary, both in depth and in reach, and native
advertising has become a controversial source of income for news media organizations. This
controversy, we argue, is due to well-established tensions between the commercial and profes-
sional logics of traditional journalism, often separated by the metaphor of the wall. Additionally,
the narrative behind native advertising carries an attempt to change the discourse around native
advertising. The discourse of separation – an agonistic discourse – is being replaced rapidly
by narratives of collaboration between marketers and news organizations (Drew and Thomas,
2017), dismissing the wall and proposing a mutual understanding between the two camps – a
symbiotic discourse. Nevertheless, the fact that news media carry a reputation of objectivity
and autonomy makes native advertising a more dangerous venture for news organizations, who
risk losing the legitimacy to decide what news is, a privilege they have acquired during the
professionalization processes of the last century. For that reason, the discourse within legacy
media is still aimed at making a case for incorporating native advertising and at the same time
at maintaining their legitimacy as high-standards journalism producers (Russ-Mohl and Nazh-
diminova, 2015).
The relevance of this phenomenon is not merely economic, as it potentially redefines the
tensions between journalists and publishers, questioning the allegiance to the citizen and threat-
ening the legitimacy and autonomy of contemporary news media. Its reach spans across borders
and media systems, spilling over several types of journalism, and amalgamates disparate actors
who can openly influence news content – from marketers to politicians (Iversen and Knudsen,
2017). While news organizations argue that native advertising is unproblematic as long as it is
recognized as advertising, research shows members of the public usually find it difficult to distin-
guish commercial and editorial content; when they do, their attitudes tend to be negative. This
highlights a key issue this type of advertising is only profitable precisely because it is difficult to
identify. The need for commercial content to mimic editorial content, while still using labels
to clarify its commercial nature, endows native advertising with a characteristic akin to ‘shiny
camouflage’ that is contradictory in nature. For the advertiser to successfully appropriate the
journalistic cachet of the publisher, the article needs to look like regular editorial content. For
the news media to be able to justify native advertising, they need to clearly label ads as ads. The
resolution of these challenges is far from even across the journalistic field. If news media manage
to make native advertising work, it could be a viable business model that funds the rest of news
operations. If they do not manage to do so, it could lead to the downfall of journalistic legitimacy
and further accentuate the crisis in which journalism is immersed.
Further reading
This chapter has drawn inspiration from Shoemaker and Reese’s Mediating the Message in the 21st
Century: A Media Sociology Perspective (2014) and Martin Conboy’s Journalism: A Critical History
(2004). To further explore the overt and covert erosion of the separation between of editorial
and commercial content, the work of Ben Bagdikian (2004) and Mark Coddington (2015) has
471
proven to be particularly insightful. Matt Carlson’s (2015) and Raul Ferrer-Conill’s (2016) articles
further contextualize the use and discursive changes of native advertising occurring in journal-
istic contexts.
Notes
1 www.wsj.com/ad/cocainenomics
2 http://paidpost.nytimes.com/netflix/women-inmates-separate-but-not-equal.html
References
Amazeen, M. A. and Muddiman, A. R. (2017) “Saving media or trading on trust? The effects of native
advertising on audience perceptions of legacy and online news publishers.” Digital Journalism, 1–20. doi:
10.1080/21670811.2017.1293488
Artemas, K., Vos, T. P. and Duffy, M. (2016) “Journalism hits a wall: Rhetorical construction of newspapers’
editorial and advertising relationship.” Journalism Studies, 1–17. doi:10.1080/1461670X.2016.1249006
Bagdikian, B. H. (2004) The New Media Monopoly. Boston, MA: Beacon Press.
Cameron, G. T. (1994) “Does publicity outperform advertising? An experimental test of the third-party
endorsement.” Journal of Public Relations Research, 6(3), 185–207.
Cameron, G. T. and Curtin, P. A. (1995) “Tracing sources of information pollution: A survey and experi-
mental test of print media’s labeling policy for feature advertising.” Journalism & Mass Communication
Quarterly, 72(1), 178–189.
Cameron, G. T. and Ju-Pak, K.-H. (2000) “Information pollution? Labeling and format of advertorials.”
Newspaper Research Journal, 21(1), 65–76.
Campbell, C. and Marks, L. J. (2015) “Good native advertising isn’t a secret.” Business Horizons, 58(6),
599–606.
Carlson, M. (2015) “When news sites go native: Redefining the advertising – editorial divide in response to
native advertising.” Journalism, 16(7), 849–865.
Carpenter, S. (2010) “A study of content diversity in online citizen journalism and online newspaper
articles.” New Media & Society, 12(7), 1064–1084.
Coddington, M. (2015) “The wall becomes the curtain: Revisiting journalism’s news-business boundary.” In
M. Carlson and S. C. Lewis (eds.), Boundaries of Journalism: Professionalism, Practices and Participation. New
Cohen, E. L. (2002) “Online journalism as market-driven journalism.” Journal of Broadcasting & Electronic
Media, 46(4), 532–548.
Cole, J. T. and Greer, J. D. (2013) “Audience response to brand journalism: The effect of frame, source, and
involvement.” Journalism & Mass Communication Quarterly, 90(4), 673–690.
Conboy, M. (2004) Journalism: A Critical History. London: Sage Publications.
Couldry, N. and Turow, J. (2014) “Advertising, big data, and the clearance of the public realm: Marketers’
new approaches to the content subsidy.” International Journal of Communication, 8, 1710–1726.
Dix, S. and Phau, I. (2009) “Spotting the disguises and masquerades: Revisiting the boundary between edito-
rial and advertising.” Marketing Intelligence & Planning, 27(3), 413–427.
Dowling, D. (2016) “The business of slow journalism: Deep storytelling’s alternative economies” Digital
Journalism, 4(4), 530–546. doi:10.1080/21670811.2015.1111769
Drew, K. K. and Thomas, R. J. (2017) “From separation to collaboration: Perspectives on editorial-business
collaboration at United States news organizations.” Digital Journalism, 1–20. doi:10.1080/21670811.20
17.1317217
Eckman, A. and Lindlof, T. (2003) “Negotiating the Gray Lines: An ethnographic case study of organiza-
tional conflict between advertorials and news.” Journalism Studies, 4(1), 65–77.
Edström, M. (2015) “Blurring the lines. Ethical dilemmas for journalists with native advertising and other
look-alike editorial content.” Presented at the Nordmedia, August 2015. Copenhagen, Denmark.
Ellerbach, J. (2004) “The advertorial as information pollution.” Journal of Information Ethics, 2004(1), 61–75.
Erjavec, K. (2004) “Beyond advertising and journalism: Hybrid promotional news discourse.” Discourse &
Society, 15(5), 553–578.
Erjavec, K. and Kovačič, M. (2010) “Relations with the media: Who are the main actors in an advertorial
production process in Slovenia?” Journalism, 11(1), 91–109.
472
Native advertising
Ferrer-Conill, R. (2016) “Camouflaging Church as State: An exploratory study of journalism’s native adver-
tising.” Journalism Studies, 17(7), 904–914.
Goodlad, N., Eadie, D. R., Kinnin, H., Raymond, M., Goodlad, N., Eadie, D. R., Kinnin, H., Raymond, M.,
Goodlada, N., Eadieb, D. R. and Kinninc, H. (1997) “Advertorial: Creative solution or last resort?” Inter-
national Journal of Advertising, 16(2), 73–84.
Hackley, C. and Tiwsakul, R. (2006) “Entertainment marketing and experiential consumption.” Journal of
Marketing Communications, 12(1), 63–75.
Hallahan, K. (2014) “Publicity under siege: A critique of content marketing, brand journalism, native
advertising and promoted user endorsements as challenges to professional practice and transparency.”
Presented at the Proceedings of the 17th International Public Relations Research Conference (pp. 391–437).
Hamilton, J. T. (2004) All the News That’s Fit to Sell: How the Market Transforms Information Into News.
Princeton, NJ: Princeton University Press.
Hampton, M. (2010) “The fourth estate ideal in journalism history.” In S. Allan (ed.), The Routledge Compan-
ion to News and Journalism, Routledge Companions. New York, NY: Routledge (pp. 3–13).
Harro-Loit, H. and Saks, K. (2006) “The diminishing border between advertising and journalism in Esto-
nia.” Journalism Studies, 7(2), 312–322.
Herbert, J. and Thurman, N. (2007) “Paid content strategies for news websites: An empirical study of Brit-
ish newspapers’ online business models.” Journalism Practice, 1(2), 208–226.
Hoofnagle, C. and Meleshinsky, E. (2015) “Native advertising and endorsement: Schema, source-based
misleadingness, and omission of material facts.” Technology Science, 1–23.
Ikonen, P., Luoma-aho, V. and Bowen, S. A. (2017) “Transparency for sponsored content: Analysing codes
of ethics in public relations, marketing, advertising and journalism.” International Journal of Strategic Com-
munication, 11(2), 165–178.
Iversen, M. H. and Knudsen, E. (2017) “When politicians go native: The consequences of political native
advertising for citizens’ trust in news.” Journalism, 1–18. doi:10.1177/1464884916688289
Joel, M. (2013) “We need a better definition of ‘native advertising’.” Harvard Business Review. Retrieved from
https://hbr.org/2013/02/we-need-a-better-definition-of
Kim, B. H. (1995) The Effectiveness of Camouflaged Advertising Format on Audiences: With Special Focus on Adver-
torial Advertising. Tuscaloosa, AL: University of Alabama.
Knudsen, E., Ferrer-Conill, R., Lauerer, C. and Barnoy, A. (2017) “The boundaries of native advertising. An
international comparison of 20 newspapers in five countries.” Presented at the International Communica-
tion Association, May 2017, San Diego.
Should Expect, Completely Updated and Revised. New York, NY: Three Rivers Press.
Lewis, J., Williams, A. and Franklin, B. (2008) “A compromised Fourth State? UK news journalism, public
relations and news sources.” Journalism Studies, 9(1), 1–20.
Li, H. and Leckenby, J. D. (2004) “Internet advertising formats and effectiveness.” Center for Interactive
Advertising, 1–31.
Macnamara, J. (2014) “Journalism – PR relations revisited: The good news, the bad news, and insights into
tomorrow’s news.” Public Relations Review, 40(5), 739–750.
Matteo, S. and Dal Zotto, C. (2015) “Native advertising or how to stretch editorial to sponsored content
within a transmedia branding era.” In G. Siegert, K. Förster, S. M. Chan-Olmsted and M. Ots (eds.),
Handbook of Media Branding. Cham: Springer International Publishing (pp. 169–185).
McNair, B. (1998) The Sociology of Journalism. Oxford: Oxford University Press.
Moses, L. (2015) “Inside the NY Times’ audience development strategy.” Digiday. Retrieved from https://
digiday.com/media/inside-ny-times-audience-development-strategy/
The New York Times. (2014) Innovation. Retrieved from www.niemanlab.org/2014/05/the-leaked-
new-york-times-innovation-report-is-one-of-the-key-documents-of-this-media-age/
Nielsen, R. K. (2016) “The business of news.” In T. Witschge, C. W. Anderson, D. Domingo and A. Her-
mida (eds.), Handbook of Digital Journalism. London: Sage Publications (pp. 51–67).
Palser, B. (2002) “Advertorial with a twist: The NYTimes.com Laramie Project Archives is a revenue model
worth talking about. American Journalism Review, 24(4), 66.
Pompeo, J. (2013) “Wall Street Journal” Editor Gerard Baker Decries Native Advertising as a “Faustian Pact”.
Retrieved from http://politi.co/1Px6QV7
Rodgers, R. R. (2007) “Journalism is a loose-jointed thing: A content analysis of editor & publisher’s discus-
sion of journalistic conduct prior to the canons of journalism, 1901–1922.” Journal of Mass Media Ethics,
22(1), 66–82.
473
Russ-Mohl, S. and Nazhdiminova, R. (2015) “Media branding and media marketing: Conflicts with jour-
nalistic norms, risks of trial and error.” In G. Siegert, K. Förster, S. M. Chan-Olmsted and M. Ots (eds.),
Handbook of Media Branding. Cham, Switzerland: Springer International Publishing (pp. 337–353).
Schauster, E. E., Ferrucci, P. and Neill, M. S. (2016) “Native advertising is the new journalism: How decep-
tion affects social responsibility.” American Behavioral Scientist, 60(12), 1408–1424.
Schechter, D. (1997) The More You Watch, the Less You Know: News Wars/(Sub)merged Hopes/Media Adventures.
New York, NY: Seven Stories Press.
Sebastian, M. (2014a) “Wall Street Journal adopts native ads, tactic its editor has said can lead to Faus-
tian Pacts.” Adage. Retrieved from http://adage.com/article/media/wall-street-journal-introducing-
native-ads-site/292044/
Sebastian, M. (2014b) “Native ad production values keep growing with “Orange is the New Black” promo.”
AdvertisingAge. Retrieved from http://adage.com/article/media/york-times-runs-native-ad-orange-
black/293713/
Sharma, N. (2015) “Native advertising vs. advertorials: Knowing the difference.” Business 2 Community.
Retrieved from www.business2community.com/native-advertising/native-advertising-vs-advertorials-
knowing-difference-01240432
Shoemaker, P. J. and Reese, S. D. (2014) Mediating the Message in the 21st Century: A Media Sociology Perspec-
tive. New York, NY: Routledge.
Sjøvaag, H. (2013) “Journalistic autonomy: Between structure, agency and institution.” Nordicom Review,
34, 155–166.
Soloski, J. (1989) “News reporting and professionalism: Some constraints on the reporting of the news.”
Media, Culture & Society, 11(2), 207–228.
Stamm, K. and Underwood, D. (1993) “The relationship of job satisfaction to newsroom policy changes.”
Journalism & Mass Communication Quarterly, 70(3), 528–541.
Stoker, K. (1995) “Existential objectivity: Freeing journalists to be ethical.” Journal of Mass Media Ethics,
10(1), 5–22.
van Reijmersdal, E., Neijens, P. and Smit, E. (2005) “Readers’ reactions to mixtures of advertising and edito-
rial content in magazines.” Journal of Current Issues & Research in Advertising, 27(2), 39–53.
Wojdynski, B. W. (2016) “The deceptiveness of sponsored news articles: How readers recognize and perceive
native advertising.” American Behavioral Scientist, 60, 1475–1491.
Wojdynski, B. W. and Evans, N. J. (2016) “Going native: Effects of disclosure position and language on the
recognition and evaluation of online native advertising.” Journal of Advertising, 45(2), 157–168.
Wojdynski, B. W. and Golan, G. J. (2016) “Native advertising and the future of mass communication.”
American Behavioral Scientist, 60(12), 1403–1407.
Wright, K. (2016) “‘These Grey Areas’: How and why freelance work blurs INGOs and news organiza-
tions.” Journalism Studies, 17(8), 989–1009.
Wu, M., Huang, Y., Li, R., Bortree, D. S., Yang, F., Xiao, A. and Wang, R. (2016) “A tale of two sources in
native advertising: Examining the effects of source credibility and priming on content, organizations, and
media evaluations.” American Behavioral Scientist, 60(12), 1492–1509.
474
36
USER COMMENTS IN DIGITAL
JOURNALISM
Current research and future directions
User comments offer a unique opportunity for the public to participate in the news production
process. Research suggests that in 2014 the vast majority – over 90% – of news websites offered
the opportunity for users to comment on the news (Stroud et al., 2014). Commentary features
allow for ‘annotative reporting’ by embedding users’ viewpoints, experiences, or expertise within
an article’s context, providing readers with additional information to form opinions (e.g., Bow-
man and Willis, 2003), which can potentially enhance deliberative processes. However, these
normative standards aren’t always met in practice: empirical research shows that user commen-
tary tends to be shallow and uncivil, which is why some online publishers have recently banned
user comments from their websites, either closing down comment sections altogether or shifting
comments to third-party social media platforms.
This chapter discusses the ambivalent nature of user comments. It provides an overview of
current research on user comments in digital journalism and directions for future research. The
chapter profiles commenters by discussing who reads and writes comments as well as motivations
and inhibitors for doing so. We then move to a review of existing theory and research on the act
of commenting, including thoughts and findings on commenting as an act of public deliberation
and political action. We also discuss prominent issues that come along with the bipolar nature of
user comments, such as comment in/civility, comment moderation, and organizational perspectives and
commenting policies, as well as commenting effects. The chapter concludes with directions for future
research.
Profiling commenters and lurkers

By reading random comments on the news, one could easily get the impression that commenters
often use the feature only to vent their anger and frustration. The aim of this contribution is to
shed light on users, both active and passive-receptive, and the motives that drive people to read
and write comments. Tracing drivers helps to understand whose standpoint is voiced and why.
There are surely narcissistic people who (ab)use the opportunity to express their anger with
elite actors, such as journalists or politicians, as well as other commenters in a rough or disrespect-
ful way; for instance, as a valve in difficult situations or when they feel superior to their discussion
counterpart (Springer, 2014; see also Wu and Atkin, 2017). However, we argue here that there
is more to commenting than that. In brief and in the best cases, commenting is an activity like
475
writing letters to the editor or debating in a club. Although it mainly serves to ‘express one’s
own opinion’, it can also be a creative, socially motivated and identity-related productive activity.1
Posting and reading comments: demography and traits

Commenting on news is a hobby in which only a few invest a great amount of time. Besides
these so-called heavy users, there are many who comment once, twice, five, or 10 times, and per-
haps never again (e.g., Ruiz et al., 2011: 476). The latest data for Germany suggests, for instance,
that 14% of the population over 16 years commented on news sites at least once (Köcher, 2016a).
This finding roughly echoes data from Sweden (Karlsson et al., 2015). Data for the US suggests
that 55% of Americans have commented online, and 15% of those commenters engaged on a
news site or app, specifically (Stroud et al., 2016). According to another study of the “Engaging
News Project” (Stroud et al., 2017), commenters are loyal – 78% of commenters are “exclusive
commenters” where they comment on one site only.2
Reading comments is somehow more attractive than writing, but shares vary from country
to country, also due to different operationalizations and sample populations (e.g., for Sweden, see
Karlsson et al., 2015: 303; for the US, see Stroud et al., 2016; for Germany, see Köcher, 2016a).
Comparative data in regard to the commenters’ socio-demography is, again, both coherent and
differing. In Sweden, commenting on news is rather popular among younger users (Bergström, 2008:
72). Similarly, age was negatively correlated with commenting frequency in a U.S. survey (Meyer and
Carey, 2014: 222). Studies in Germany, however, found that commenters are older than the average
users of a news site (Ziegele et al., 2013: 88; see also Springer et al., 2015). It has to be noted, though,
that the German studies are nonrepresentative and that the recent shifting to social networking sites
could also affect measured age levels in the future. Further, commenting in the US is, compared to
comment reading, associated with lower levels of education (Stroud et al., 2016), while in the already
mentioned news users’ study in Germany, findings of Ziegele and colleagues suggest a high level of
education among commenters (cf. Ziegele et al., 2013; Köcher, 2016b). In regard to gender, however,
available data suggests that men comment more frequently than women (e.g., Chung and Yoo, 2008;
Friemel and Dötsch, 2015; Köcher, 2016b; Stroud et al., 2016).
Significant predictors for the usage of interactive features, such as the commentary feature,
were found to be internet efficacy/skill and interest in new technologies (Bergström, 2008;
Chung and Yoo, 2008; Larsson, 2011), as well as political interest and engagement (Chung, 2008;
Köcher, 2016a; Larsson, 2011). Swiss data suggest that commenters are more conservative than
comment readers (Friemel and Dötsch, 2015: 161). Similarly, a recent German study finds that
commenters disproportionately support the right-wing populist party “Alternative for Ger-
many” (AfD), the liberal party (FDP), and the “Pirate Party” (Köcher, 2016b).
Comment readers, on the other hand, seem to be such a diverse group of users that a Swedish
study could not find any characteristics concerning these so-called lurkers’ socio-demography
or traits, such as civic or political engagement (Larsson, 2011: 1190). However, they are very
frequent and intense news site visitors (ibid.: 1191). The aforementioned Swiss study suggests
that, besides being more “left-leaning” than commenters, comment readers tend to be “younger
than the average website user” (Friemel and Dötsch, 2015: 161). Hence, the authors assume “that
reader and writer of comments on news sites differ significantly with respect to some central
attributes” (ibid: 162).
Motivations and inhibitors

Studies applying survey or content analytic techniques identified different factors that triggered
commenting on news but also sought to discover de-motivations that kept people from voicing
476
User comments in digital journalism
their opinions. Since commenting starts – at least theoretically – with the article’s input, the
topic should first of all influence the number of incoming comments as well as the manner in
which people voice their opinions. (We discuss the topic’s influence on civility in the next sec-
tion.) Available databases are solid enough to generalize that people all over the world love to
discuss politics (e.g., Stroud et al., 2016; Ziegele et al., 2013: 89). This finding holds true in multi-
country studies (Goodman, 2013), as well as studies of British (Richardson and Stanyer, 2011),
German (Springer, 2014), Israeli (Tenenboim and Cohen, 2015), Turkish (Ürper and Çevikel,
2014), and U.S. (Boczkowski and Mitchelstein, 2012) news websites. Richardson and Stanyer
(2011) even showed that different media outlets relate to different commenting behaviors. On
broadsheet sites, topics such as general domestic politics/party policies, U.S. politics/elections, or
foreign affairs attracted the most comments, while on tabloid sites, religion- and crime-related
topics generated the most comments (ibid.: 990–991).
Not only the article’s topic but also the journalist and the journalistic contextualization and
interpretations of events influenced the amount of incoming user feedback. Weber (2014), for
instance, showed that the news factors proximity, impact, and frequency had a significant posi-
tive effect on the number of comments, while the news factor facticity had a negative one. If
journalists describe an event factually, without providing further analysis or interpretation, the
news story is less likely to attract comments. If the journalistic contextualization or interpretation
diverges from the perceptions and attitude of the readers, this dissonance can result in emotional
pressure that some people try to ease by corrective commenting (Springer, 2014; see also Chung
et al., 2015). The same holds true for “outraging posts” of fellow commenters (Springer, 2014).
A fair share of commenters appreciate engaging in a discussion with others (Springer, 2014;
Springer et al., 2015; Ziegele et al., 2013: 87; Ziegele, 2016). Hence, a sense of a virtual commu-
nity was found to be a relevant predictor for commenting (e.g., Meyer and Carey, 2014: 221).
However – and this is supported by content analyses – this motivation is not necessarily the most
important driver of comments (e.g., Jakobs, 2014: 201–203; Ruiz et al., 2011). On some news sites,
comments mostly serve self-representation; they react to or echo ideological positions expressed
in the articles, producing threads with low levels of discussion and little diversity of opinion. Such
behavior results in what Ruiz and colleagues (2011: 480) label “dialogue of the deaf ”.
On other websites, the authors found vivid “communities of debate” (ibid.: 482), in which
people discuss diverging viewpoints in a (mostly) polite manner. Shifting user comments to social
network sites is likely to encourage interactivity (Wu and Atkin, 2017), but a forum structure can
also lead to a high share of interactive comments (e.g., 69% for the German online news maga-
zine “Spiegel Online”; see Jakobs, 2014: 203). Further, previous comments should also be able to
fuel discussions. Ziegele and colleagues (2014; Ziegele, 2016) drew on the news value tradition to
distill “discussion factors” of (interactive) user comments. The authors found that commenters
could “trigger response comments by including controversy, unexpectedness, personalization,
and uncertainty in their postings and by avoiding incomprehensibility and negativity” (Ziegele
et al., 2014: 1129).
When asked about their motivation to engage, commenters and comment readers mentioned
or agreed to drivers that can be subsumed under the well-known “Uses-and-Gratification” cat-
egories that scholars established for traditional media usage. Besides the aforementioned need for
integration and social interaction,3 users seek or provide information and express and work on
their personal identity; that is, they use arguments as a means of self-presentation or expressions of
others as a means for social comparison. Additionally, they comment or read comments for their
entertainment and relaxation (e.g., Chung and Yoo, 2008; Diakopoulos and Naaman, 2011b: 5;
Springer, 2014; Wu and Atkin, 2017; Yoo, 2011).
The social setting in which commenting is embedded also serves to gratify a need for recog-
nition (e.g., Barnes, 2015: 822). Yet data also suggest that commenters who score higher on the
477
social-integrative dimension cannot always satisfy their need to socialize to the desired extent
through the use of interactive features and become frustrated over time (Springer, 2014; Springer
et al., 2015; Yoo, 2011). This could be because discussion with others always implies the persua-
sion of dissenters, and that understanding is complicated, as commenters are often opinionated
and thus quite resistant (e.g., Springer, 2014; Springer et al., 2015).
Furthermore, social-interactive motivations were found to not only be directed at users but
also at journalists (e.g., Heise et al., 2014: 418; Springer, 2014; Springer et al., 2015). However,
studies show that interaction with journalists happens only rarely in public, since the authors
usually refrain from engaging in user discussions below their articles (Graham and Wright, 2015;
Jakobs, 2014: 201; Singer, 2009). This absence might be frustrating for users seeking to reach
journalists publicly, thus possibly having an effect on the comments’ civility. We discuss this in
the section dealing with comment in/civility.
Beyond the analysis of de-motivations, studies also investigated why some people would not
want to voice their opinions publicly. Besides restrictive time budgets (e.g. Barnes, 2015), people
are hesitant to register and concerned about data protection. Another reason is simply the lack
of personal involvement (Springer, 2014; Ziegele, 2016). A low sense of opinionation can be
caused by disinterest, but comment readers of an Australian alternative journalism website did
also express intimidation by the “textual authority portrayed by other commenters or the jour-
nalist” (Barnes, 2015: 817), indicating low levels of confidence (ibid.: 818) and the subjective
perception of lacking competence (cf. Springer, 2014).
Furthermore, the perceived “quality” of discussions was found to hamper participation
(Springer, 2014; Springer et al., 2015). For instance, Diakopoulos and Naaman (2011b: 4) show
that nonusers would not read comments on a Californian news site because they found com-
ments uncivil. It is reasonable to assume that people would not want to discuss with or relate
to such individuals, whereas some users maintain that freedom of expression should outweigh
any concern for civil discussion (Robinson, 2010). Thus, studies worldwide suggest that a fair
share of news audiences see comments as an “integral part of the overall story” (Barnes, 2015:
820; cf. Bergström and Wadbring, 2015: 147; Springer, 2014). This is because reading comments
can gratify, especially information-related needs, and can be very entertaining – for instance, to
observe and follow a heated public dispute (Springer, 2014).
Comments as public deliberation

Since the integration of commenting platforms directly below news stories offers a unique oppor-
tunity for ‘ordinary people’ to engage in immediate public discussion of the news – including
the observation of these discussions and the receiving of input beyond peer-group opinions – it is
no wonder that research on user comments often draws on philosophical notions of the public
sphere as an overarching theoretical framework. This understanding of comments adapts the
work of Habermas (1989) to describe a space where the public can engage in rational, civil
discussion about the important events and issues of the day. Unlike the in-person coffee houses
and salons that Habermas depicted, comments are seen as a virtual opportunity for engaging in
public deliberation.
While both journalists and users see the deliberative potential in commenting platforms, there
is widespread concern about the quality of discussion occurring in these spaces (Barnes, 2015;
Goodman, 2013; Lee, 2012; Meltzer, 2015; Nielsen, 2012; Santana, 2011, 2014; Springer et al.,
2015). Broadly echoing what Singer et al. (2011) found in a collaborative study, a 2013 study
from the World Association of Newspapers conducted interviews with news professionals from
over 100 organizations across 63 countries (Goodman, 2013). The study found that journal-
ists valued user comments, particularly for engaging users, promoting loyalty, and encouraging
478
community, which all indirectly contribute to increased revenue, but they also expressed concern
about the general quality of these discussions and the resources necessary to curate productive
discussion spaces. Similarly, Meltzer (2015) analyzed industry discourse about user comments,
finding widespread concern among news professionals about incivility in comments. Others have
found similar evidence that journalists and news organizations see a need to improve the quality
of user comments (e.g., Nielsen, 2012; Santana, 2011) and that journalists and their organizations
should be responsible for encouraging more civil discussion.
Comment in/civility
In response to these concerns, we have seen the emergence of a substantial body of research on
the quality of user comments, focusing particularly on the civil and/or hostile nature of user
comments. Hostility, or incivility, is commonly identified as comments “intentionally designed
to attack someone or something and, in doing so, incite anger or exasperation through the use
of name-calling, character assassination, offensive language, profanity, and/or insulting language”
(Ksiazek et al., 2015: 854). Civility is less easily defined. While Herbst (2010: 19) offers a useful
conceptualization of civility as “constructive engagement with others through argument, delib-
eration, and discourse”, Ksiazek and colleagues (2015) find that much of the literature simply
treats civility as the absence of hostility.
Research suggests the in/civility of comments can be explained by a variety of factors, includ-
ing: story topic, sources, journalist participation in comment threads, journalist demography, and
organizational commenting policies (Diakopoulos and Naaman, 2011a; Ksiazek, 2016). It seems
reasonable to assume that the nature of comments will vary across different topics, especially
considering that comments are an “emotional response to the story” (Braun, 2015: 818). This
is supported by interviews with news professionals, who regularly observe that the tone of user
discussions is linked to the story topic (Diakopoulos and Naaman, 2011b; Goodman, 2013).
Research also demonstrates that comment in/civility varies across story topics. Coe et al.
(2014) found that hard news stories were more likely than soft news stories to exhibit incivility
in comments (sports being the exception). Ksiazek (2016) found variations in in/civility across
political news topics, with controversial topics (e.g., gun control) predicting not only more com-
ments but also more hostility in those comments.
In addition to story topic, the sources used in a given story also explain variations in com-
ment in/civility. Research suggests the inclusion of sources as a journalistic practice is positively
related to more hostility in user comments (Ksiazek, 2016). However, it is quite possible that the
use of sources, as a general practice, is less important than the actual sources quoted. The deci-
sion to comment and the tone of that comment is likely to stem from a cognitive and emotional
reaction to the source itself. In fact, Coe et al. (2014) found variations in incivility depending
on the specific source quoted in a story (e.g., stories with quotes from Barack Obama had more
incivility in the user comments).
As discussed earlier, research has also considered the role of journalist participation in com-
ment sections. Despite concerns about intruding on “user” space (Diakopoulos and Naaman,
2011b; Robinson, 2010) and the increased demand on journalists’ time in order to participate
(Diakopoulos and Naaman, 2011b), when journalists contribute comments to these discussions
we see more comments and more civil comments overall (Ksiazek, 2016). In terms of general
engagement (i.e., the volume of comments), perhaps users see participation by journalists as an
indicator of the value of commenting on news stories. It is also possible that users see this as
an opportunity to directly interact with professional journalists, something that was previously
relegated to letters to the editor and news talk radio. Regarding the improved civility of dis-
cussions where journalists contribute, this finding seems to support a surveillance effect where
479
users are more likely to act in a civil manner if they are aware that journalists are monitoring
their comments.
Organizational perspectives and commenting policies

There is general disagreement among journalists about whether they should take part in user
discussions. Robinson (2010) finds contrasting philosophies on the role of journalists in com-
menting platforms. Based on ethnographic interviews, she found two camps: “Traditionalists”
wanted to maintain a hierarchical relationship with the audience, while “Convergers” pushed for
more interactivity with their users.
Meanwhile, at the organizational level, Goodman (2013) found widespread agreement among
news organizations across the world that active participation in comment forums improves the
quality of comments, and the World Association of Newspapers included this as one of their
“best practices”. Consistent with this, Diakopoulos and Naaman (2011b) note that both users
and journalists wanted to see more journalist participation, although some editors did express
concerns about intruding on a ‘user’ space and the increased demands on reporters’ time, as noted
earlier.
News organizations typically have specific policies governing their comment sections. For
the most part, these are designed to encourage civil, productive dialogue among users. These
policies often include required user registration and profanity filters as well as more active pre-
and post-moderation strategies, prohibiting anonymous usernames, and allowing users to rate
and rank each other’s comments (Domingo, 2011). Research has explored the implementation
of various combinations of these policies, as well as their effectiveness in promoting more civil
comments (Braun, 2015; Coe et al., 2014; Ksiazek, 2015; Ruiz et al., 2011; Santana, 2014).
Each of the aforementioned policies has demonstrated a positive relationship with civil com-
menting independently. However, when we account for these policies collectively and control
for other explanatory variables (story topic, use of sources, journalist participation), only pre-
moderation, post-moderation, reputation management (i.e., allowing users to police themselves
through rating/ranking systems), and prohibiting anonymity predicted more civil and/or less
hostile comments (Ksiazek, 2016).
There is a growing body of research on the role of anonymity in commenting (e.g., Ksiazek,
2015, 2016; Santana, 2014). Reflecting on these findings, two patterns emerge. First, prohibiting
anonymity can encourage more civil discussion. When commenters are accountable and identi-
fiable, they contribute in more productive ways and tend to eschew hostility/incivility. However,
it also seems that a policy of prohibiting anonymity discourages user engagement, in general;
stories on sites that prohibit anonymous usernames exhibit less comments, overall, than those on
sites that allow anonymity (Ksiazek, 2016). Santana (2014) finds that U.S. news organizations are
divided in their approach to anonymity: roughly half require identifiable usernames, while half
allow anonymity. While some news organizations prohibit anonymous comments in an effort
to improve the civility of user comments, others may allow anonymity to encourage overall user
engagement.
Comments as political action: against the media, against the elite

One reason why user discussions tend to become uncivil is because user comments often criticize.
Commenters challenge the journalistic interpretation or point out factual mistakes; they criticize
actors cited in the articles or fellow commenters (e.g., Baden and Springer, 2014; Springer, 2014).
It is reasonable to assume that some of them take a critical stance toward political institutions, the
‘mainstream society’, and the elite actors ruling it. For instance, in light of what was said about
480
the overrepresentation of conservative stances among commenters, it makes sense to presume that
they use this tool to actively “correct” or “counterbalance” the perspectives of the more left-
leaning journalists (e.g., Willnat and Weaver, 2014 for the US; Steindl et al., 2017 for Germany;
see also Chung et al., 2015). It might not be far-fetched to characterize some users who com-
ment on news as ‘media skeptics’; such people perceive journalists and products of their work as
selective and biased (Tsfati and Cappella, 2005).
The relationship between coverage and “follow-up” or “talk back” communication of the
audience has been investigated intensively (e.g., Baden and Springer, 2014; Douai and Nofal,
2012; Reader, 2012; Toepfl and Piwoni, 2015). Findings suggest that comments are not only
replications or reproductions of media frames but also offer modifications and additions: Users
challenge and render journalistic interpretations in a different light by proposing alternative
interpretations. However, it is certainly questionable whether some of the formulated standpoints
are socially desirable or beneficial for democratic discourse (e.g., Baden and Springer, 2014;
Toepfl and Piwoni, 2015).
As previously mentioned, it has been acknowledged that users’ opinions help readers to get
a feeling about the “pulse of the public debate” (Douai and Nofal, 2012: 269; Springer, 2014)
and give these posts at least short-term persuasive potential, which we will discuss in the next
section. Thus, measures are applied to use the tool to spread opinions strategically (be it manually
or automated) and make them appear more popular than such viewpoints might actually be. For
instance, the so-called “Russian troll army” as a means of “information war” was discussed in
the media recently (e.g. Higgins, 2016), and “below the line” discussions (Graham and Wright,
2015: 317) on the latest forced migration movements were reportedly regularly dominated by
right-wing populists’ opinions (e.g., Faiola, 2016).
Given the intensity of these public debates, one might reasonably assume that “powerful
counter (sub)public sphere[s]” emerge (Toepfl and Piwoni, 2015: 482). It is suspected that ano-
nymity can not only be responsible for incivility but also for the infectious spreading of discrimi-
nating opinions. First, since people do not need to fear sanctions in their “real lives”, they feel safe
to utter what they might fear to say in their “offline world” (cf. Springer, 2014). Second, specific
viewpoints may become more visible and thus “louder” simply because of the comments’ reach:
while personal communication has a natural border that ends with one’s contacts, the commen-
tary feature can facilitate communication among people who do not personally know each other,
living all over a country but sharing the same attitudes (ibid.). This could lead to commenters
identifying with similar others, conferring upon a minority confidence to speak out. Apart from
socially undesirable viewpoints, such a ‘spiral of silence’ effect could have positive effects, too, for
instance in light of the (de-)marginalization of minorities.
Commenting effects
Actors interested in influencing public opinion would not invest resources if they did not assume
that user comments had an effect on others. For instance, the concept of the ‘spiral of silence’
proposes that we observe our environment closely to assess climates of opinion on issues at stake.
User comments provide such social cues. Hence, Lee (2012) found that comments can impact
users’ perceptions of broad public opinion. In particular, when high-ego-involvement users expe-
rience comments that contradict their opinions, they believe the public (and the media) is against
their position.
Available studies drawing upon experimental settings repeatedly show, at the least, the short-
term persuasive potential of comments. For instance, there is support for a negative spillover
effect on perceptions and actions. Negative comments can diminish the persuasive influence
of a news article (Winter et al., 2015). Similarly, criticizing or uncivil comments can negatively
481
affect the perception of an article’s quality (Kümpel and Springer, 2016; Prochazka et al., 2016).
Another study indicated that people’s exposure to prejudiced comments led them to post more
prejudiced comments themselves (Hsueh et al., 2015).
However, studies also found positive spillover effects (e.g., Kümpel and Springer, 2016). For
instance, positive comments can lead to a more favorable assessment regarding the perceived public
opinion climate, responsibility attribution, and attitudinal evaluations of a financial manager scan-
dalized by the news (von Sikorski and Hänelt, 2016). However, this study also found a contrast
effect of comments. In comparison to balanced comment threads, the study’s participants who were
only exposed to threads with one-sided, negative comments assessed the article’s journalistic qual-
ity significantly higher (ibid.: 563). Such contrast effects could stem from audience resistance. The
processing of user comments seems not only to be influenced by the recipients’ predispositions and
the relevance of the topic (e.g., Anderson et al., 2014; Winter et al., 2015) but also by the comments’
substance. For instance, Chen and Ng (2016) showed that participants exposed to civil comments
presumed their influence on the opinions of others to be clearly higher than participants exposed
to uncivil comments.
To summarize, the literature on commenting effects shows that user comments serve the
audience as information for what others think about controversial topics and mainstream media
coverage in general and that they have the potential for at least short-term influence, such as
affecting the audience’s perception of journalistic quality. Thus, user comments’ relevance cannot
be ignored. However, there are still numerous questions we cannot answer yet. We will outline
some of the most pressing ones in the chapter’s final section.
Future research
Over the last several years, comments have been praised, analyzed, criticized, and, as a result,
sometimes banned from news sites. The quality of user comments is the most pressing issue, and
we see two solutions: either shut comments off entirely or take ownership. Let’s face it: media
outlets introduced commenting spaces to build user engagement and generate revenues, and
curating public discussions was never really a priority. Since they did not take great interest in
what happened ‘below the line’, comment threads tended to become shout boxes. If we thought
that shutting down comments was the right solution, we would not bother to write this last sec-
tion. Instead, since we think that there is still potential for these discussion spaces, we hope that
scholars will begin to determine how the decline of user comment sections can be averted. This
will require becoming more interdisciplinary and more collaborative.
The media, communication scholars, and computer scientists should talk to each other on a
more regular basis. Together they can investigate resources for more advanced and efficient ways
to moderate, a practice that was found to have a positive influence on the degree of civility in
comments (Ksiazek, 2015, 2016). For instance, Diakopoulos and Naaman (2011b) recommend
filtering comments based on sentiment to improve the quality of comments on news websites,
while Braun (2015) observes the increasing use of “collapsing algorithms” that effectively filter
out (or “collapse”) comments that reach a threshold of low ratings. Interestingly, this latter prac-
tice integrates two of the organizational commenting policies that seem to have a positive impact
on comment quality: automated moderation and socially driven reputation management systems,
where communities of users hold each other accountable through ratings and rankings. Both
practices warrant further investigation.
Further, we recommend taking the task of moderation more seriously. Moderation as it is
organized today predominantly consists of banning and blocking or publishing. But moder-
ation should strive to be much more than that; moderators of group discussions pose ques-
tions, they summarize or mediate results of discussions (in the case of online commentary, to
482
participants and the newsrooms that host them). If we model moderation as more dialogical, this
of course comes with the investment of resources. Thus, automated tools could be useful: chat
bots, for instance, are able to imitate human communication. Such chat bots could be trained
and implemented transparently. Experiments then need to show whether such chat bots would
be ‘accepted’ by the participants and thus be able to positively influence the discussions’ climate.
Comment moderation still seems to be a highly experimental field that definitely calls for a more
systematic analysis.
The media has the data and access to the field; communication scholars have the empirical
and theoretical backdrop; and computer scientists have the know-how of automation for cost
reduction. Collectively, we should start to investigate the effects of not only passive moderation
through automated filters but also active moderation strategies. At the same time, while filters,
collapsing algorithms, and chat bots might better the overall quality of these virtual discussions
of current events, we need to be wary of the potential for these tools to silence commenters. If
we truly value commenting spaces for their potential to facilitate public deliberation and politi-
cal action, then as we pursue more innovative moderation techniques and other strategies for
improving the quality of user comments, we should aim to ensure that these attempts to promote
civil discussion do not limit free speech.
Further reading
For additional reading, consider work by Singer and colleagues (2011) and Ziegele and colleagues
(2017). Singer et al.’s monograph is based on the impressive fieldwork of a consortium of pres-
tigious scholars from around the world. It reflects on “how journalists in Western democracies
are thinking about, and dealing with, the inclusion of content produced and published by the
public”. Ziegele and colleagues (2017) guest edited a recent special issue of the ICA-affiliated
open-access journal Studies in Communication and Media (SCM) that contains multidisciplinary
and up-to-date studies on user commentary.
Notes
1 For the “profiling”, “action” and “effects” sections, parts of the research reviews are adapted from
Springer & Kümpel 2018 [in German language].
2 For the frequency of commenting on news in social media, comparative data can also be found in the
annual Reuters Institute Digital News Report from 2014 on.
3 Even comment readers can experience a certain feeling of belonging: Barnes (2015: 822), for instance,
found for an alternative journalism website that readers wanted to learn what the community thinks
about issues at stake, to sympathize by reading others’ opinions, and to associate themselves with the views
of likeminded commenters (see also Diakopoulos and Naaman 2011b).
References
Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A. and Ladwig, P. (2014) “The ‘nasty effect’:
Online incivility and risk perceptions of emerging technologies: Crude comments and concern.” Journal
of Computer-Mediated Communication, 19(3), 373–387. doi:10.1111/jcc4.12009
Baden, C. and Springer, N. (2014) “Com(ple)menting the news on the financial crisis: The contribution of
news users’ commentary to the diversity of viewpoints in the public debate.” European Journal of Com-
munication, 29(5), 529–548. doi:10.1177/0267323114538724
Barnes, R. (2015) “Understanding the affective investment produced through commenting on Austra-
lian alternative journalism website New Matilda.” New Media & Society, 17(5), 810–826. doi:10.1177/
1461444813511039.
Bergström, A. (2008) “The reluctant audience: Online participation in the Swedish journalistic context.”
Westminster Papers in Communication & Culture, 5(2), 60–79.
483
Bergström, A. and Wadbring, I. (2015) “Beneficial yet crappy: Journalists and audiences on obstacles and
opportunities in reader comments.” European Journal of Communication, 30(2), 137–151. doi:10.1177/
0267323114559378
Boczkowski, P. J. and Mitchelstein, E. (2012) “Clicking, e-mailing, and commenting: How users take
advantage of different forms of interactivity on online news sites.” Human Communication Research, 38(1),
1–22. doi:10.1111/j.1468-2958.2011.01418.x.
Bowman, S. and Willis, C. 2003. We Media: How Audiences Are Shaping the Future of News and Information.
Retrieved from www.hypergene.net/wemedia/download/we_media.pdf (accessed 2 December 2017).
Braun, J. A. (2015) “News programs: Designing MSNBC.com’s online interfaces.” Journalism, 16(1), 27–43.
doi:10.1177/1464884914545730.
Chen, G. M. and Ng, Y. M. M. (2016) “Third-person perception of online comments: Civil ones persuade
you more than me.” Computers in Human Behavior, 55, 736–742. doi:10.1016/j.chb.2015.10.014
Chung, D. S. (2008) “Interactive features of online newspapers: Identifying patterns and predicting use
of engaged readers.” Journal of Computer-Mediated Communication, 13(3), 658–679. doi:10.1111/j.1083-
6101.2008.00414.x
Chung, D. S. and Yoo, C. Y. (2008) “Audience motivations for using interactive features: Distinguishing
use of different types of interactivity on an online newspaper.” Mass Communication and Society, 11(4),
375–397. doi:10.1080/15205430701791048
Chung, M., Munno, G. J. and Moritz, B. (2015) “Triggering participation: Exploring the effects of third-
person and hostile media perceptions on online participation.” Computers in Human Behavior, 53, 452–
461. doi:10.1016/j.chb.2015.06.037
Coe, K., Kenski, K. and Rains, S. A. (2014) “Online and uncivil? Patterns and determinants of incivility in
newspaper website comments.” Journal of Communication, 64, 658–679. doi:10.1111/jcom.12104.
Diakopoulos, N. and Naaman, M. (2011a) Topicality, time, and sentiment in online news comments. Pro-
ceedings of CHI 2011, Vancouver, BC, Canada, May 7–12.
Diakopoulos, N. and Naaman, M. (2011b) Towards quality discourse in online news comments. Proceedings
of CSCW 2011, Hangzhou, China, March 19–23. doi:10.1145/1958824.1958844.
Domingo, D. (2011) “Managing audience participation. Practices, workflows and strategies.” In J. B. Singer,
A. Hermida, D. Domingo, A. Heinonen, S. Paulussen, T. Quandt, Z. Reich and M. Vujnovic (eds.),
Participatory Journalism: Guarding Open Gates at Online Newspapers. Sussex: Wiley-Blackwell (pp. 76–95).
Douai, A. and Nofal, H. K. (2012) “Commenting in the online Arab public sphere: Debating the Swiss
minaret ban and the ‘ground zero mosque’ online.” Journal of Computer-Mediated Communication, 17(3),
266–282. doi:10.1111/j.1083-6101.2012.01573.x
Faiola, A. (2016, January 6) “Germany springs to action over hate speech against migrants.” The Washington
Post. Retrieved from www.washingtonpost.com/world/europe/germany-springs-to-action-over-hate-
speech-against-migrants/2016/01/06/6031218e-b315-11e5–8abc-d09392edc612_story.html?utm_
term=.8dcc3f5af156 (accessed 17 April 2017).
Friemel, T. N. and Dötsch, M. (2015) “Online reader comments as indicator for perceived public opinion.”
Kommunikationspolitik für die digitale Gesellschaft, 1(8), 151–172.
Graham, T. and Wright, S. (2015) “A tale of two stories from ‘below the line’: Comment fields at the Guard-
ian.” The International Journal of Press/Politics, 20(3), 317–338. doi:10.1177/1940161215581926
Goodman, E. (2013) Online Comment Moderation: Emerging Best Practices. Darmstadt, Germany: World Asso-
ciation of Newspapers (WAN-IFRA) [online]. Retrieved from www.wan-ifra.org/reports/2013/10/04/
online-comment-moderation-emerging-best-practices (accessed 17 April 2017).
Habermas, J. (1989) The Structural Transformation of the Public Sphere: An Inquiry Into a Category of Bourgeois
Society. Cambridge, MA: The MIT Press.
Heise, N., Loosen, W., Reimer, J. and Schmidt, J.-H. (2014) “Including the audience: Comparing the atti-
tudes and expectations of journalists and users towards participation in German TV news journalism.”
Journalism Studies, 15(4), 411–430. doi:10.1080/1461670X.2013.831232
Herbst, S. (2010) Rude Democracy: Civility and Incivility in American Politics. Philadelphia, PA: Temple Uni-
versity Press.
Higgins, A. (2016, May 31) “Effort to expose Russia’s ‘troll army’ draws vicious retaliation.” The New
York Times. Retrieved from www.nytimes.com/2016/05/31/world/europe/russia-finland-nato-trolls.
html?_r=1 (accessed 17 April 2017).
Hsueh, M., Yogeeswaran, K. and Malinen, S. (2015) “‘Leave your comment below’: Can biased online com-
ments influence our own prejudicial attitudes and behaviors?” Human Communication Research, 41(4),
557–576. doi:10.1111/hcre.12059
484
Jakobs, I. (2014) “Diskutieren für mehr Demokratie? Zum deliberativen Potenzial von Leser-Kommentaren
zu journalistischen Texten im Internet.” In W. Loosen and M. Dohle (eds.), Journalismus und (sein) Pub-
likum. Schnittstellen zwischen Journalismusforschung und Rezeptions- und Wirkungsforschung. Wiesbaden: VS
(pp. 191–210).
Karlsson, M., Bergström, A., Clerwall, C. and Fast, K. (2015) “Participatory journalism – The (r) evolution
that wasn’t. Content and user behavior in Sweden 2007–2013.” Journal of Computer-Mediated Communica-
tion, 20(3), 295–311. doi:10.1111/jcc4.12115
Köcher, R. (2016a, November) Vertrauenskrise der Medien? Presentation at the VDZ Publishers’ Summit,
Berlin, Germany. Retrieved from http://publishers-summit.vdz.de/fileadmin/zt_de/user_upload/
downloads/vortraege2016/VDZPS16_Renate-K%C3%B6cher_Allensbach-Studie_Vertrauenskrise-
der-Medien_20161108.pdf
Köcher, R. (2016b) “Flüchtlingszustrom: Auswirkungen eines gesellschaftlichen Aufregungszyklus auf poli-
tisches Interesse und Mediennutzung.” AWA 2016. Retrieved from www.ifd-allensbach.de/fileadmin/
AWA/AWA_Praesentationen/2016/AWA_2016_Koecher_Fluechtlingskrise_Medien.pdf
Ksiazek, T. B. (2015) “Civil interactivity: How news organizations’ commenting policies explain civility and
hostility in user comments.” Journal of Broadcasting and Electronic Media, 59(4), 556–573. doi:10.1080/08
838151.2015.1093487
Ksiazek, T. B. (2016) “Commenting on the news: Explaining the degree and quality of user comments on
news websites.” Journalism Studies [online]. doi:10.1080/1461670X.2016.1209977
Ksiazek, T. B., Peer, L. and Zivic, A. (2015) “Discussing the news: Civility and hostility in user comments.”
Digital Journalism, 3(6), 850–870. doi:10.1080/21670811.2014.972079
Kümpel, A. and Springer, N. (2016) “Commenting quality. Effects of user comments on perceptions of
journalistic quality.” Studies in Communication and Media, 5(3), 353–366. doi:10.5771/2192-4007-
2016-3-353
Larsson, A. O. (2011) “Interactive to me – interactive to you? A study of use and appreciation of interactivity on
Swedish newspaper websites.” New Media & Society, 13(7), 1180–1197. doi:10.1177/1461444811401254
Lee, E.-J. (2012) “That’s not the way it is: How user-generated comments on the news affect perceived media
bias.” Journal of Computer-Mediated Communication, 18, 32–45. doi:10.1111/j.1083-6101.2012.01597.x.
Meltzer, K. (2015) “Journalistic concern about uncivil political talk in digital news media: Responsi-
bility, credibility, and academic influence.” The International Journal of Press/Politics, 20(1), 85–107.
doi:10.1177/1940161214558748.
Meyer, H. K. and Carey, M. C. (2014) “In moderation: Examining how journalists’ attitudes toward online
comments affect the creation of community.” Journalism Practice, 8(2), 213–228. doi:10.1080/17512786
.2013.859838
Nielsen, C. (2012) “Newspaper journalists support online comments.” Newspaper Research Journal, 33(1),
86–100. www.newspaperresearchjournal.org/
Prochazka, F., Weber, P. and Schweiger, W. (2016) “Effects of civility and reasoning in user comments on
perceived journalistic quality.” Journalism Studies [online]. doi:10.1080/1461670X.2016.1161497
Reader, B. (2012) “Free press vs. free speech? The rhetoric of ‘civility’ in regard to anonymous online com-
ments.” Journalism & Mass Communication Quarterly, 89(3), 495–513. doi:10.1177/1077699012447923
Richardson, J. E. and Stanyer, J. (2011) “Reader opinion in the digital age: Tabloid and broadsheet newspaper
websites and the exercise of political voice.” Journalism, 12(8), 983–1003. doi:10.1177/1464884911415974.
Robinson, S. (2010) “Traditionalists vs. convergers: Textual privilege, boundary work, and the journalist-
audience relationship in the commenting policies of online news sites.” Convergence: The International
Journal of Research into New Media Technologies, 16(1), 125–143. doi:10.1177/1354856509347719.
Ruiz, C., Domingo, D., Lluís Micó, J., Díaz-Noci, J., Masip, P. and Meso, K. (2011) “Public sphere 2.0? The
democratic qualities of citizen debates in online newspapers.” The International Journal of Press/Politics,
16(4), 463–487. doi:10.1177/1940161211415849
Santana, A. D. (2011) “Online readers’ comments represent new opinion pipeline.” Newspaper Research Jour-
nal, 32(3), 66–81. www.newspaperresearchjournal.org/
Santana, A. D. (2014) “Virtuous or vitriolic: The effect of anonymity on civility in online newspaper reader
comment boards.” Journalism Practice, 8(1), 18–33. doi:10.1080/17512786.2013.813194.
Singer, J. B. (2009) “Separate spaces: Discourse about the 2007 Scottish elections on a national newspaper
web site.” The International Journal of Press/Politics, 14(4), 477–496. doi:10.1177/1940161209336659
Singer, J. B., Hermida, A., Domingo, D., Heinonen, A., Paulussen, S., Quandt, T., Reich, Z. and Vuj-
novic, M. (eds.). (2011) Participatory Journalism: Guarding Open Gates at Online Newspapers. Sussex:
Wiley-Blackwell.
485
Springer, N. (2014) Beschmutzte Öffentlichkeit? Warum Menschen die Kommentarfunktion auf Onlinenachrichten-
seiten als öffentliche Toilettenwand benutzen, warum Besucher ihre Hinterlassenschaften trotzdem lesen, und wie die
Wände im Anschluss aussehen. Berlin: LIT.
Springer, N., Engelmann, I. and Pfaffinger, C. (2015) “User comments: Motives and inhibitors to write
and read.” Information, Communication & Society, 18(7), 798–815. doi:10.1080/1369118X.2014.997268.
Springer N. and Kümpel A. S. (2018) User-Generated (Dis)Content. In C. Nuernbergk and Neuberger C.
(eds.) Journalismus im Internet. Wiesbaden: Springer VS (pp. 241–271).
Steindl, N., Lauerer, C. and Hanitzsch, T. (2017) “Journalismus in Deutschland. Aktuelle Befunde zu
Kontinuität und Wandel im deutschen Journalismus.” Publizistik, 62, 401–423. doi:10.1007/s11616-
017-0378-9
Stroud, N. J., Van Duyn, E., Alizor, A., Alibhai, A. and Lang, C. 2017. “Comment section survey across
20 news sites.” The Engaging News Project. Annette Strauss Institute for Civic Life, University of Texas
at Austin. Retrieved from https://engagingnewsproject.org/wp-content/uploads/2017/01/Comment-
Section-Survey-Across-20-News-Sites.pdf
Stroud, N. J., Van Duyn, E. and Peacock, C. 2016. “Survey of commenters and comment readers.” The
Engaging News Project. Annette Strauss Institute for Civic Life, University of Texas at Austin. Retrieved
from https://engagingnewsproject.org/research/survey-of-commenters-and-comment-readers/
Stroud, N. J., Scacco, J. M. and Curry, A. L. 2014. “Interactive features in online news.” The Engaging News
Project. Annette Strauss Institute for Civic Life, University of Texas at Austin. Retrieved from http://
engagingnewsproject.org/enp_prod/wp-content/uploads/2014/04/ENP_News_Site_Analysis.pdf
Tenenboim, O. and Cohen, A. A. (2015) “What prompts users to click and comment: A longitudinal study
of online news.” Journalism, 16(2), 198–217. doi:10.1177/1464884913513996.
Toepfl, F. and Piwoni, E. (2015) “Public spheres in interaction: Comment sections of news websites as coun-
terpublic spaces.” Journal of Communication, 65(3), 465–488. doi:10.1111/jcom.12156
Tsfati, Y. and Cappella, J. N. (2005) “Why do people watch news they do not trust? The need for cognition
as a moderator in the association between news media skepticism and exposure.” Media Psychology, 7(3),
251–271. doi:10.1207/S1532785XMEP0703_2
Ürper, D. C. and Çevikel, T. (2014) “Reader comments on mainstream online newspapers in Turkey: Percep-
tions of web editors and moderators.” Communications, 39(4), 483–503. doi:10.1515/commun-2014-0022
von Sikorski, C. and Hänelt, M. (2016) “Scandal 2.0: How valenced reader comments affect recipients’
perception of scandalized individuals and the journalistic quality of online news.” Journalism and Mass
Communication Quarterly, 93(3), 551–571. doi:10.1177/1077699016628822
Weber, P. (2014) “Discussions in the comments section: Factors influencing participation and interactiv-
ity in online newspapers’ reader comments.” New Media & Society, 16(6), 941–957. doi:10.1177/
1461444813495165
Willnat, L. and Weaver, D. H. 2014. The American Journalist in the Digital Age: Key Findings. Retrieved from
https://larswillnat.files.wordpress.com/2014/05/2013-american-journalist-key-findings.pdf
Winter, S., Brückner, C. and Krämer, N. C. (2015) “They came, they liked, they commented: Social
influence on Facebook news channels.” Cyberpsychology, Behavior, and Social Networking, 18(8), 431–436.
doi:10.1089/cyber.2015.0005.
Wu, T-Y. and Atkin, D. (2017) “Online news discussions. Exploring the role of user personality and moti-
vations for posting comments on news.” Journalism and Mass Communication Quarterly, 94(1), 61–80.
doi:10.1177/1077699016655754
Yoo, C. Y. (2011) “Modeling audience interactivity as the gratification-seeking process in online newspa-
pers.” Communication Theory, 21(1), 67–89. doi:10.1111/j.1468-2885.2010.01376.x
Ziegele, M. (2016) Nutzerkommentare als Anschlusskommunikation. Theorie und qualitative Analyse des Diskus-
sionswerts von Online-Nachrichten. Wiesbaden: VS.
Ziegele, M., Breiner, T. and Quiring, O. (2014) “What creates interactivity in online news discussions? An
exploratory analysis of discussion factors in user comments on news items.” Journal of Communication,
64(6), 1111–1138. doi:10.1111/jcom.12123
Ziegele, M., Johnen, M., Bickler, A., Jakobs, I., Setzer, T. and Schnauber, A. (2013) “Männlich, rüstig, kom-
mentiert? Einflussfaktoren auf die Aktivität kommentierender Nutzer von Online-Nachrichtenseiten.”
Studies in Communication and Media, 3(1), 67–114. doi:10.5771/2192-4007-2013-1-67
Ziegele, M., Springer, N. Jost, P. and Wright, S. (2017) “Online user comments across news and other con-
tent formats: Multidisciplinary perspectives, new directions.” Studies in Communication and Media (SCM),
6(4), 315–332. doi:10.5771/2192-4007-2017-4-315
486
37
THEORIZING DIGITAL
JOURNALISM
The limits of linearity and the rise of relationships
Jane B. Singer
Digital media pose conceptual and methodological challenges for journalism studies scholars.
Our effects-oriented theoretical approaches, which traditionally dominated attempts to concep-
tualize how journalism works, rest on an inherently linear view of the communication process,
positing that messages are delivered (mostly by journalists) to audiences who are affected by and
responsive to them. And our methods tend to require, or at least to work best with, a fixed object
of study.
But the digital environment is in some ways like Heraclitus’ river: It can never be stepped in
twice, for it is constantly changing, and so are those it touches. Unlike the river, though, it also
flows in all directions at once, along uncontrollable courses and with unpredictable ripples.
This chapter begins with a closer look at why the value of concepts that were so fruitful in an
analog past has diminished in our digital present.1 It then explores how journalism studies schol-
ars are moving the field forward in the 2010s, concluding with a call for new concepts around
‘relationship effects’.
Limited effects
Defining exactly what we mean when we talk about journalists or journalism has never been
easy and gets harder by the day. Even a definition of what constitutes ‘news’ is difficult to nail.
News could once be identified as part of an institutionally sanctioned information package, pro-
duced by people who worked within the occupational space of a newsroom to make it available
at regular intervals to readers, viewers, and listeners (Singer, 2016). Although such a definition
suggested considerable complexity, it did enable the constitutive elements to be isolated and
examined productively (Wahl-Jorgensen and Hanitzsch, 2009). Doing so has become far trickier
in a contemporary media environment that is dramatically less bounded – in time, in place, and
in the nature of communication itself – and more interlaced.
A closer look at one set of theories relied on by communication scholars for more than
50 years illustrates the challenges. Particularly in the United States, ‘effects theories’ dominated
attempts to explain how communication worked through much of the twentieth century, a
time when the media industry commanded considerable social, political, economic, and even
moral power. Scholars sought to understand how the messages produced by such influential and
487
Jane B. Singer
ubiquitous entities were affecting consumers of those messages, which at the time meant nearly
everyone (Singer, 2016).
They did so primarily through observation and measurement of the components of a com-
munication process conceptualized as rooted in the transmission of a message from a sender to
a receiver, who then assimilated and acted (or not) on that message in discernible ways (Westley
and MacLean, 1957). A look at four of these effects theories, each widely and productively
applied in the late twentieth century, helps demonstrate why their utility has diminished in a
digital news environment. The earlier use of each theory is summarized, followed by an overview
of recent applications.
Spiral of silence theory

Traditional environment: Spiral of silence theory posits that the media have a significant effect on
audience behavior. Noelle-Neumann (1993) proposed that we respond to coverage of a major
event or issue by assessing public opinion on the topic; that assessment then shapes our actions.
If we believe our own views are not widely held, we are likely to remain silent in order to avoid
social isolation. By doing so, we contribute to public opinion as others perceive it, resulting in a
spiraling effect in which seemingly dominant views gain even more ground, while alternatives
retreat further. The news media thus pack a one-two punch: their initial coverage creates aware-
ness, and they then are instrumental in shaping impressions about which views are dominant and
which in decline – impressions that in turn inform individuals’ decisions about what is safe to say
in public without being ostracized.
The theorized effect, though, proved difficult to isolate and measure; repeated tests yielded
mixed and culturally distinctive results ((Donsbach et al., 2014; Scheufele and Moy, 2000). The
effect seems to vary depending on mitigating circumstances such as attitude certainty (Matthes
et al., 2010), the nature of conflict around an issue, the interaction of majorities and minorities
over time, the role of reference groups (Price and Allen, 1990), and the interplay between local
and national opinion climates (Salmon and Neuwirth, 1990). A meta-analysis of 17 studies, pub-
lished in 1997, identified only a very small, though statistically significant, relationship between
the degree to which a person believes others hold similar opinions and the willingness to speak
out (Glynn et al., 1997).
Digital environment: Does the theory hold up in a world of unfettered information and unin-
hibited discourse? Not consistently. Studies investigating a potential spiral of silence effect in
online discussion around contentious issues such as abortion (McDevitt et al., 2003) and same-
sex marriage (Ho and McLeod, 2008) found that subjects holding a minority opinion did not
seem to feel inhibited in expressing their views. But other studies have found support for the
theory online, for instance in relation to discussion of genetically modified food (Kim, 2012) or
nuclear power (Miyata et al., 2015).
Additional work that tests the theory in relation to social media highlights the importance of
individual characteristics in willingness to speak out on such platforms as Facebook or Twitter.
For instance, the national opinion climate on gay bullying was found to be somewhat related
to willingness to speak out on social media, but individual characteristics such as willingness to
self-censor also were key (Gearhart and Zhang, 2014). Individual differences also were central to
LGBT individuals’ willingness to self-disclose on Facebook: those in the closet seemed silenced
by a perceived heteronormative majority, while those who were ‘out’ used the site’s affordances
for vocal empowerment (Fox and Warber, 2014).
In general, Schulz and Roessler (2012) suggest the hybridity of the internet so greatly expands
available information choices that the theory becomes of limited value. They also reference the
role of subjectivity in assessing the climate of opinion in such an environment, a suggestion in
488
Theorizing digital journalism
line with growing awareness of the ways in which fake news and filter bubbles misinform civic
decisions (Tsfati et al., 2014).
Cultivation theory
Traditional environment: In the 1960s, as television became a household fixture and television news
a dominant source of civic information as well as entertainment, scholars sought to measure its
effect. Cultivation theory posits that watching television affects our world view, typically lead-
ing to a heightened sense of risk and insecurity (Gerbner and Gross, 1976). It is among the most
widely cited – and widely challenged – of mass media theories. As early as the 1980s, an inten-
sive period of cultivation theory testing (and defending), numerous caveats were offered. The
notion of ‘mainstreaming’, for instance, suggests that heavy television viewing leads to converged
outlooks across social groups, yet ‘resonance’ suggests nearly the opposite, with attitudinal effects
varying among population subgroups (Gerbner et al., 1980). Effects of television viewing turned
out to be neither uniform nor universal: other variables can and do intervene, and controlling for
them significantly lessens any remaining effect (Hirsch, 1981).
Over the years, however, evidence has grown that while the effects of TV viewing in general
may be elusive or even negligible, particular types of televised content do seem to have an impact
on attitudes about such topics as racism, crime, violence, and victimization, as well as on feel-
ings of alienation and anomia (Potter, 1993). Bryant and his colleagues (1981) found that heavy
viewing of action-adventure programs increased fearfulness and anxiety levels, especially if justice
was not shown to prevail. Oliver and Armstrong (1995) found higher levels of racial prejudice,
as well as punitive attitudes about crime, associated with frequent viewing and greater enjoy-
ment of reality-based programming. More recently, Kahlor and Eastin (2011) found that large
amounts of rape-related content in soap operas and crime dramas cultivate perceptions related
to sexual assault, with viewers of those genres more likely to accept rape myths and overestimate
false accusations.
Digital environment: In some ways, cultivation theory is well-suited to digital media; indeed,
many challenges stemmed from its tough-to-test premise of immersion in television rather than
selective viewing (Morgan et al., 2009) and its view that messages are systems rather than discrete
variables (Morgan and Shanahan, 2010). Moreover, its underlying concern is with the cultural
effects of storytelling, certainly of ongoing relevance in a digital world (Morgan et al., 2015). Yet
it is hard to get around the fact that cultivation is inherently a theory developed to explain the
effects of the historically and culturally specific phenomenon of television (Livingstone, 2004).
Despite the ubiquity of digital and mobile media use in Western society, few have even tried to
apply cultivation theory directly to digital media other than video games (Chong et al., 2012; Wil-
liams, 2006). Most contemporary work remains focused on television and its effects on perceptions
of such matters as materialism (Shrum et al., 2011), immigration (McKay Semmler et al., 2014),
and, of course, violence (Jamieson and Romer, 2014). Arguably, Putnam (2000) and those who
build on his work take their cue from cultivation theory in claiming that increased use of digital
technologies decreases trust in social institutions and participation in society; however, empirical
testing suggests the presumed online impact on socializing with others is “limited or nonexisting”
(Vergeer and Pelzer, 2009: 202). Particularly in the early days of the internet, connections between
its use and social capital, social trust, and well-being were repeatedly tested. No cultivation effect
was identified; many findings suggested minimal and even mildly positive impacts.
Theories of behavioral and attitudinal effects of the media, then, served to instigate tests and
advance knowledge but ultimately showed that a seemingly simple linear process was neither
simple nor linear. Media messages do influence behaviors and attitudes – but so do lots of other
things, and attempts to isolate effects are rarely definitive (Singer, 2016).
489
Jane B. Singer
Cognitive effects theories, which tend to posit weaker or less direct effects, generally hold up
better. Two such theories directly tied to journalistic output have enjoyed considerable support
over the years, particularly as applied to political information: the knowledge gap hypothesis and
agenda-setting theory.
The knowledge gap hypothesis

Traditional environment: The knowledge gap hypothesis states that as more information enters
a social system, “segments of the population with higher socioeconomic status (SES) tend to
acquire this information at a faster rate than the lower status segments, so that the gap in knowl-
edge between these segments tends to increase rather than decrease” (Tichenor et al., 1970: 159–
160). This initial proposition suggested effects are most evident in areas of general knowledge,
such as public affairs and health, rather than areas of niche interest. Subsequent research found
support for the hypothesis though also contradictions of it; clear evidence of a health-related
knowledge gap has been particularly elusive (Hwang and Jeong, 2009).
Extended theory testing suggests that effects, which not infrequently go in a direction oppo-
site to the one predicted, are at least partially contingent on the impact of interpersonal commu-
nication and individual motivation. Early on, controversy over a local issue was found to decrease
the knowledge gap effect within the affected community (Donohue et al., 1975); subsequent
work illuminated the role of powerful local groups, community pluralism, and promotional
efforts (Viswanath and Finnegan, 1996). Personal motivation to gain knowledge, group member-
ship, and other factors play a part in health knowledge acquisition (Ettema et al., 1983; Viswanath
et al., 1993).
In short, as Gaziano (1983) detailed in an early overview of knowledge gap studies confirmed
by a second meta-analysis a quarter-century later (Hwang and Jeong, 2009), potential inter-
vening variables are numerous, and findings are all over the conceptual map. As early as 1980,
Dervin was on to one likely reason why. The knowledge gap hypothesis, she said, is based on
the paradigm of communication as transmission: a source sends a message to a receiver. Long
before the rise of the internet, she called for a conceptual shift to user-constructed and -defined
information, emphasizing an individual’s need for sense making (Dervin, 1980, as cited in Severin
and Tankard, 2001).
Digital environment: Applications of this theory at the start of the internet era explored
the “digital divide” between those with and without online access (Bucy, 2000; Hindman,
2000). As digital technologies have diffused within and across societies, attention has shifted
from access to use of digital information (van Deursen and van Dijk, 2014). Findings suggest
that gaps persist even when availability is not an issue (Jeffres et al., 2012; Tran, 2013). For
example, Wei and Hindman (2011) found Americans with higher SES were more likely to use
the internet for informational purposes than their lower SES counterparts, accentuating and
extending gaps in political knowledge; Bonfadelli’s (2002) earlier findings in Switzerland were
similar. Gaps seem to remain in connection with the kinds of activities people pursue online
to “improve their human, financial, political, social and cultural capital” (Hargittai and Hin-
nant, 2008: 615).
Social media use has been posited as one potential solution to the knowledge gap because
of its ability to provide incidental access and thus passive learning opportunities even to people
uninterested in civic information (Bode, 2016). Measurable effects, however, have been under-
whelming for those not already engaged in politics. In the experimental component of her
study, Bode found that social media users were not more likely to be politically informed than
nonusers. A meta-analysis by Boulianne (2015) suggested that causality is unclear; by itself, social
media use seems to have minimal impact on political participation. In fact, despite their potential
490
to narrow knowledge gaps, such sites as Facebook and Twitter actually may amplify or reinforce
inequality of political engagement (Yoo and Gil de Zúñiga, 2014).
Agenda-setting theory
Traditional environment: All the effects theories used as examples so far are relevant to journal-
ism studies scholars, but none is as intertwined with news production and consumption as
agenda-setting theory – along with its offspring, such as second-level agenda setting, and its
close or distant cousins, such as framing or priming (Singer, 2016). The theory evolved from
the observation that the press “may not be successful much of the time in telling people what
to think, but it is stunningly successful in telling its readers what to think about” (Cohen, 1963:
13). This role of providing salience cues, particularly about political issues, was systematically
explored by McCombs and Shaw (1972), who found that voters did indeed tend to share the
media’s overall definition of which issues mattered most. Over the next two decades, hundreds
of agenda-setting studies tackled everything from the conditions that might enhance or limit
the effect, to the standards used in making judgments, to just how “the media” formulate an
agenda at all.
Agenda setting and its conceptual kin put forward more modest claims about the impact of
media messages than do such powerful effects theories as spiral of silence or cultivation, and they
have fared relatively well under exhaustive testing, across time, and in diverse national contexts.
It seems the media do make certain issues or aspects of issues easier for people to recall, thus
affecting public attitudes about candidates and political matters (Scheufele and Tewksbury, 2007).
Agenda-setting effects have been identified in virtually every U.S. election since the 1970s.
But this is not to say that the effects have been either uniform or unambiguous. On the con-
trary, political agenda setting turns out to be contingent on a host of conditions, including the
kinds of issues covered; the types of media outlets involved and the sort of coverage they provide
(Walgrave and van Aelst, 2006); the salience of issues to a given audience (Erbring et al., 1980);
and the degree of individual motivation and engagement (McLeod et al., 1974). The effect seems
strongest for relatively unobtrusive issues that do not directly affect most people, such as foreign
policy (Weaver et al., 2008).
Digital environment: Popularity of the agenda-setting concept has not abated in the digital
era. By the mid-2000s, it had amassed more than half a million citations (Bennett and Iyengar,
2008). But some of those citations come from scholars demonstrating challenges to the theory
as initially proposed. For example, a study of political blogs suggested that the agenda-setting
power of traditional media was “no longer universal or singular” as those media became “just one
force among many competing influences” (Meraz, 2009: 701). Online citizens themselves can,
in theory, participate directly in setting the public agenda, both by producing their own content
and “by rendering the agenda-setting processes of established professional media outlets radically
provisional, malleable and susceptible to critical intervention” (Goode, 2009: 7).
Theory testing has continued with social media, notably in explorations of its role in setting
the legacy media agenda. In a study of social media activity around the 2015 Belgian elections,
Harder and his colleagues (2017) found that although Twitter political activity can be influential,
media actors on Twitter have far more agenda-setting influence than other actors do. In a U.S.
presidential primary context, Conway et al. (2015) found a reciprocal or symbiotic relation-
ship between political players and journalists at leading legacy outlets: although they followed
candidates, journalists continued to set the agenda on many issues. But social media may open
up other avenues of influence: a case study of cable news host Rachel Maddow’s Facebook page
indicated a positive correlation between stories discussed on the social networking platform and
the subsequent inclusion of similar stories on television (Jacobson, 2013).
491
Jane B. Singer
In short, ambiguity pervades even our most robust media effects theories. Of course, such is
the nature of theories: Their strength lies not only in their ability to provide answers but also,
perhaps primarily, in their ability to stimulate questions. That is how knowledge expands (Singer,
2016). Yet undeniably, mass media effects that already were difficult to extricate from other vari-
ables during a time when we could more or less pinpoint what we meant by “mass media” are
virtually impossible to isolate in the media world we now inhabit.
The “Five I’s”

Effects theories benefit not only from an effect that is measurable but also from the presence of
a distinct and identifiable communicator, communications act or product, channel, and recipient.
Although those components characterized twentieth-century news media, the pieces were always
harder to define precisely and assess than they sound – and they are far harder now. Contempo-
rary digital journalism is, among other things:
• Immersive. The “media” today constitute a communicative space in which we live con-
stantly rather than a separate thing that we use occasionally and whose impact we therefore
can reasonably hope to isolate and measure. “The uses and appropriations of media can be
seen as fused with everything people do, everywhere people are”, Deuze writes in his intro-
duction to Media Life. “We can only imagine a life outside of media” (2012: x, emphasis in
original).
• Interconnected. Linear models do posit a (limited) amount of communication that swims
upstream – from, rather than to, media audiences. But that conceptualization, typically with
a steadfast gatekeeper directing the flow of traffic at the model’s core, is laughably inadequate
in today’s interactive world (Singer, 2016). Producers and consumers of all manner of medi-
ated content, including ‘news’, are interchangeable, with any given individual filling both
roles all but simultaneously. Moreover, in a structure that seemed fantastical at mid-century
(Bush, 1945) and remained just barely conceivable a generation ago to anyone not, literally,
a rocket scientist (Berners-Lee, 1999), every one of those communicators is linked to every
other one – well over 3 billion and growing – in an incomprehensibly complex global
network.
• Individualized. Yet at the same time, our mediated environment is unique to each of us to
an unprecedented degree. Each of those 3 billion online users puts the even-larger billions
of available pieces together in a different way, with the number of possible combinations
stretching nearly to infinity. It’s true that ‘mass media’ outlets reach more people than ever
thanks to the internet. But in comparison with the pre-internet era, those brands are a
much-diminished fraction of our daily information diet; moreover, any particular brand may
fail to penetrate our personal filter bubbles (Flaxman et al., 2016).
• Iterative. Traditional media are finite and definitive. Once the newspaper is published, it
becomes a self-contained and unchangeable product, and tomorrow’s paper will be a wholly
new (self-contained and unchangeable) product. Once the news broadcast is over, it’s over
(Singer, 2016). Not so with online news, which even before the rise of social media was
characterized by ongoing conversation and contestation (Boczkowski, 2004). Online mes-
sages are perpetually fluid constructions in the contemporary “liquid” information environ-
ment (Widholm, 2016).
• Instantaneous. Immediacy has always been a core attribute of digital information technolo-
gies. This “speed fetishism” (Correia, 2012: 109) has created considerable angst for journalists
concerned about accuracy, as well as other less predictable effects such as increasing homo-
geneity of news products (Boczkowski, 2010; Phillips, 2012). More broadly, the continual
492
flood of new content makes it hard to assess the impact of more than the minuscule percent-
age of messages fished out of the current and caught, briefly, in the net of collective attention
(Singer, 2016).
These “Five I’s” do not constitute a comprehensive list of the characteristics of our contempo-
rary media world. But even a partial litany of challenges to effects theories in a digital age high-
lights the inadequacy of a set of theories premised on distinct communication actors producing
messages whose effects can be isolated, observed, and measured. How are we to understand the
impact of any news item when message senders and recipients are interchangeable; when mes-
sages in disparate forms continuously arrive and are then instantly reshaped and redistributed in
myriad ways by myriad people and programs; and when everyone’s information diet is wildly
diverse and uniquely personal (Singer, 2016)?
Alternative responses
Not surprisingly, calls for a rethink are increasing (Jeffres, 2015; Steensen and Ahva, 2015; Wei-
mann et al., 2014). How, theorists wonder, might researchers best respond to the shortcomings
of a traditional effects-based approach? This section offers preliminary empirical clues to how
contemporary journalism studies scholars are answering that question.
The data described here were gathered by sampling 10 articles focused on digital journalism
from each of a dozen leading journals that publish journalism studies scholarship. The selected
articles, published between January 2010 and June 2017, were assessed to identify their theoreti-
cal or conceptual framework, as well as the method used if the article was empirical in nature.2
Although dominated by work in Europe and the United States, the sample included scholarship
from more than two dozen countries on five continents. Obviously, the results of this exploratory
inquiry can only be indicative. But they offer intriguing insights into how journalism scholars
are moving the field well beyond a consideration of mediated effects in a social and digital age.
Two observations stand out. One is that most of the published work in our field today rests on
empirical data; of the 120 sampled articles, only 17 (14.2%) were nonempirical, and five of those
were published in Communication Theory, which of course is devoted to theoretical concerns.
The second is that explicit theory-testing is something of a rarity. Three journals – Communication
Research, the Journal of Computer-Mediated Communication, and Mass Communication and Society –
published a total of 21 sampled articles that explicitly tested a theory; the other nine journals
published only 14 such articles among them. Overall, 70.8% of the sample consisted of items
either descriptive or loosely conceptual in nature.
Methodologically, traditional quantitative approaches remain prominent. The 103 empirically
based articles, 55 of which drew on multiple methods, included 26 content analyses, 19 surveys,
and 15 experiments. Another 20 articles rested wholly or in part on interviews, while 15 were
categorized as case studies and 13 as textual or discourse analyses. A range of other methods,
from focus groups to semiotic analysis, were employed, but only a handful seemed tailored to
the nature of digital content, such as analysis of links or search data. The sample contained only
a single network analysis.
Authors were more creatively eclectic in applying theoretical and conceptual frameworks,
however. Among the 21 articles that set out to test a theory, two dealt with agenda setting and
two more with inter-media agenda setting; two with gatekeeping theory; and another three with
the knowledge gap hypothesis or digital divide. In each, the theorized effects were identified. For
example, agenda-setting effects were found to hold across generations (Lee and Coleman, 2014);
U.S. newspapers appeared to maintain their gatekeeping role in reporting on natural disasters
around the world (Yan and Bissell, 2015); and gender, race, and age are factors in a “usage gap”
493
Jane B. Singer
related to the likelihood of creating content online (Correa, 2010). Other sampled theories based
on measuring media effects that received empirical support included cognitive dissonance, hos-
tile media effects, and exemplification effects, which relate to perceptions about the typicality of
sources.
Aside from effects-based research, the sampled articles suggest additional attention to concepts
perhaps better suited to the fluid and interconnected nature of digital media. Theories related
to social capital, social presence, and strong/weak ties all turned up, as did others related to path
dependence, structuration, substitutability, and temporality. Most scholars found at least moderate
levels of support for these and other theories; more intriguing than the empirical results, however,
is evidence of the productive expansion of frames of reference for understanding the contempo-
rary media space.
Still more support is offered by the bulk of sampled items that were conceptual in nature
but did not rise to the level of theory testing. Space does not permit exploring the hundreds of
diverse ideas explored even within this limited sample – most articles encompassed at least two
concepts – but suffice it to say that substantive knowledge generation about digital journalism is
well underway. A few broad themes emerge:
• The shifting nature of media ‘audiences’ – including roles filled, content created, and inter-
actions with journalists – is an area of active conceptual exploration. Sampled articles dealt
with citizen reporters (Al-Ghazzi, 2014; Davis, 2015), user comments (Erjavec and Kovačič,
2012; Prochazka et al., 2016), and other forms of “participatory journalism” (Karlsson
et al., 2015; Scott et al., 2015), as well as the evolving relationships among diverse actors
in the contemporary journalism arena (Lewis et al., 2014; Marchionni, 2013). In a related
vein, scholars are revisiting notions of civic engagement (Kaufhold, Valenzuela, and Gil de
Zúniga), activism (Liu, 2016), and digitally enabled public discourse (Rinke, 2016).
• Normative theories and their relation to journalistic practice in the digital and social era also
are getting a fresh look. In addition to exploring shifting journalistic boundaries (Carlson,
2016; Shanahan, 2011), scholars seem particularly attentive to credibility (Kruikemeier and
Lecheler, 2016), transparency and accountability (Kampf and Daskal, 2014; Revers, 2014),
and authority (Burroughs and Burroughs, 2012).
• The interplay between technological affordances and journalistic products and practices also
was well represented. Topics ranged from news liquidity (Widholm, 2016) to the role of
digital intermediaries (Nielsen and Ganter, 2017) to the rise of digitally enabled fact check-
ers (Graves et al., 2016) and of other new or newly adapted storytelling forms (Hiippala,
2017; Norris, 2017).
A call for ‘relationship effects’

As the study of ‘journalism’ increasingly encompasses the study of ‘digital journalism’, the con-
strictions imposed by a linear conception of the communication process are more and more
discomfiting for theorists. But if that’s the bad news, the preliminary data just summarized offer
plenty of good news to offset it. Journalism studies scholars and colleagues in other communica-
tion disciplines are identifying and applying an impressive range of conceptual frameworks that
are better suited to a digital media environment and that promise to broaden and deepen our
understanding of how this world works.
That said, there is considerable work to be done. One inescapable finding from these pre-
liminary data is that few new theories are as yet emerging from the swirl of ideas surrounding
journalism in a digital space. Indeed, this little study identified exactly . . . none. Relatively few
494
of the sampled articles attempted theory testing; those that did sought to identify whether, and if
so how, existing theories work today. This concluding section offers one suggestion about where
we might fruitfully build.
What we may need is, in fact, a new effects theory – one that expressly encompasses ‘relation-
ship effects’.
Relationships between journalists and ‘audiences’ were included in the 1957 Westley-MacLean
model, though given a subsidiary, dotted-line notation as feedback loops. And they are funda-
mental in other disciplines in both the social sciences and the humanities, which intimately link
communication and culture (Carey, 1989). An immersive media universe invites melding the
linear effects tradition with the intellectually rich understanding of how humans interact and
create social and cultural connections (Singer, 2016)
Although much contemporary work already points in this direction, it typically falls into the
category of practitioner studies, focusing on the effects on journalists themselves of inherently
mutual formats. How do journalists use Twitter (Lasorsa et al., 2012; Vis, 2013)? How do they see
user contributions affecting what happens inside the newsroom (Lewis et al., 2010; Paulussen and
Ugille, 2008)? How do these contributions challenge journalists’ ethical practices and normative
constructs (Singer and Ashman, 2009)?
Other scholars already have persuasively argued for a more holistic and culturally situated
approach to journalism studies (Hanitzsch, 2007; Zelizer, 2004) and for consideration of “jour-
nalism as process” (Robinson, 2011) – that is, a view of news not as a discrete product but as a
shared, distributed action with multiple actors engaged in shifting interactions.
But the concept of relationships has not always been explicit, nor has the idea been adequately
connected to the particular characteristics of the digital media environment. These traits open
up new opportunities to apply our extensive understanding of media effects to journalism today.
What are the effects of immersion, interconnectedness, and role interchangeability on journalists
and on journalism – journalism understood as a fluid, iterative process in which ‘messages’ are
ubiquitous and multi-directional and the roles of ‘senders’ and ‘receivers’ are perpetually recipro-
cal? What are the effects of a wholly nonlinear media system in which everyone is a participant
in an unbounded and ongoing communicative endeavor (Hjarvard, 2013)? What might we learn
when we shift from seeing journalistic relationships as finite, discrete, and readily depicted by
unidirectional black arrows to seeing them as ubiquitous, multi-directional and continually in
dotted-line flux (Singer, 2016)?
Let’s find out.
Further reading
Numerous scholars have adeptly summarized attempts to theorize the effects of digital jour-
nalism. Excellent resources include “Mass Communication Theories in a Time of Changing
Technology” by Leo Jeffres (2015); an introduction to a special 2015 issue of Mass Communication
and Society; and “Theories of Journalism in a Digital Age”, by Steen Steensen and Laura Ahva
(2015), introducing a special issue of Digital Journalism. Their fine essays should encourage read-
ers to further explore the timely and topical articles that the editors highlight. An introduction
to the Routledge Handbook of Journalism Studies, by editors Karin Wahl-Jorgensen and Thomas
Hanitzsch (2009), considers methodological as well as theoretical matters and is another strong
entry point not only to the volume but also to the subject in general. Finally, Hanitzsch’s 2007
article in Communication Theory, “Deconstructing Communication Theory: Toward a Universal
Theory”, is a key reference for journalism studies scholars seeking ideas for conceptualizing their
complex and constantly changing field.
495
Jane B. Singer
Notes
1 Portions of this section of the chapter, as well as the sections on the “five I’s” and on “relationship effects”
at the end, are drawn from “Transmission Creep: Media Effects Theories and Journalism Studies in a
Digital Era”. A link to the article, published in Journalism Studies in 2016, is available from: www.tand-
fonline.com/doi/abs/10.1080/1461670X.2016.1186498.
2 Details about the method used to draw and analyze the sample are available from the author.
References
Al-Ghazzi, O. (2014) “Citizen journalism in the Syrian uprising: Problematizing Western narratives in a
local context.” Communication Theory, 24(4), 435–454.
Bennett, W. L. and Iyengar, S. (2008) “A new era of minimal effects? The changing foundations of political
communication.” Journal of Communication, 58(4), 707–731.
Berners-Lee, T. (1999) Weaving the Web: The Original Design and Ultimate Destiny of the World Wide Web by Its
Inventor. New York, NY: Harper Collins.
Boczkowski, P. J. (2004) Digitizing the News: Innovation in Online Newspapers. Cambridge, MA: The MIT Press.
Boczkowski, P. J. (2010) News at Work: Imitation in an Age of Information Abundance. Chicago, IL: University
of Chicago Press.
Bode, L. (2016) “Political news in the news feed: Learning politics from social media.” Mass Communication
and Society, 19(1), 24–48.
Bonfadelli, H. (2002) “The Internet and knowledge gaps: A theoretical and empirical investigation.” Euro-
pean Journal of Communication, 17(1), 65–84.
Boulianne, S. (2015) “Social media use and participation: A meta-analysis of current research.” Information,
Communication and Society, 18(5), 524–538.
Bryant, J., Carveth, R. A. and Brown, D. (1981) “Television viewing and anxiety: An experimental exami-
nation.” Journal of Communication, 31(1), 106–119.
Bucy, E. P. (2000) “Social access to the Internet.” Harvard International Journal of Press/Politics, 5(1), 50–61.
Burroughs, B. and Burroughs, W. J. (2012) “The Masal Bugduv hoax: Football blogging and journalistic
authority.” New Media and Society, 14(3), 476–491.
Bush, V. (1945, July) “As we may think.” Atlantic Monthly. Retrieved from www.theatlantic.com/magazine/
archive/1945/07/as-we-may-think/303881/
Carey, J. W. (1989) Communication as Culture: Essays on Media and Society. London: Unwin Hyman.
Carlson, M. (2016) “Metajournalistic discourse and the meanings of journalism: Definitional control,
boundary work, and legitimation.” Communication Theory, 26(4), 349–368.
Chong, Y. M. G., Teng, K. Z. S., Siew, S. C. A. and Skoric, M. M. (2012) “Cultivation effects of video games:
A longer-term experimental test of first-and second-order effects.” Journal of Social and Clinical Psychology,
31(9), 952–971.
Cohen, B. C. (1963) The Press and Foreign Policy. Princeton, NJ: Princeton University Press.
Conway, B. A., Kenski, K. and Wang, D. (2015) “The rise of Twitter in the political campaign: Searching for
intermedia agenda-setting effects in the presidential primary.” Journal of Computer-Mediated Communica-
tion, 20(4), 363–380.
Correa, T. (2010) “The participation divide among ‘online experts’: Experience, skills and psychological
factors as predictors of college students’ web content creation.” Journal of Computer-Mediated Communica-
tion, 16(1), 71–92.
Correia, J. C. (2012) “Online journalism and civic life.” In E. Siapera and A. Veglis (eds.), The Handbook of
Global Online Journalism. Malden, MA: Wiley-Blackwell (pp. 101–118).
Davis, S. (2015) “Citizens’ media in the favelas: Finding a place for community-based digital media produc-
tion in social change processes.” Communication Theory, 25(2), 230–243.
Dervin, B. (1980) “Communication gaps and inequities: Moving toward a reconceptualization.” Progress in
Communication Sciences, 2, 73–112.
Deuze, M. (2012) Media Life. Malden, MA: Polity.
Donohue, G. A., Tichenor, P. J. and Olien, C. N. (1975) “Mass media and the knowledge gap: A hypothesis
reconsidered.” Communication Research, 2(1), 3–23.
Donsbach, W., Tsfati, Y. and Salmon, C. T. (2014) “The legacy of spiral of silence theory: An introduction.”
In W. Donsbach, C. T. Salmon and Y. Tsfati (eds.), The Spiral of Silence: New Perspectives on Communication
and Public Opinion. New York, NY: Routledge (pp. 1–18).
496
Erbring, L., Goldenberg, E. N. and Miller, A H. (1980) “Front-page news and real-world cues: A new look
at agenda-setting by the media.” American Journal of Political Science, 24(1), 16–49.
Erjavec, K. and Kovačič, M. P. (2012) “‘You don’t understand, this is a new war!’ Analysis of hate speech in
news web sites’ comments.” Mass Communication and Society, 15(6), 899–920.
Ettema, J. S., Brown, J. W. and Luepker, R. V. (1983) “Knowledge gap effects in a health information cam-
paign.” Public Opinion Quarterly, 47(4), 516–527.
Flaxman, S., Goel, S. and Rao, J. M. (2016) “Filter bubbles, echo chambers, and online news consumption.”
Fox, J. and Warber, K. M. (2014) “Queer identity management and political self-expression on social net-
working sites: A co-cultural approach to the spiral of silence.” Journal of Communication, 65(1), 79–100.
Gaziano, C. (1983) “The knowledge gap: An analytical review of media effects.” Communication Research,
10(4), 447–486.
Gearhart, S. and Zhang, W. (2014) “Gay bullying and online opinion expression: Testing spiral of silence in
the social media environment.” Social Science Computer Review, 21(1), 18–36.
Gerbner, G. and Gross, L. (1976) “Living with television: The violence profile.” Journal of Communication,
26(2), 172–199.
Gerbner, G., Gross, L., Morgan, M. and Signorielli, N. (1980) “The ‘mainstreaming’ of America: Violence
profile no. 11.” Journal of Communication, 30(3), 10–29.
Glynn, C. J., Hayes, A. F. and Shanahan, J. (1997) “Perceived support for one’s opinions and willingness
to speak out: A meta-analysis of survey studies on the ‘spiral of silence.’” Public Opinion Quarterly, 61(3),
452–463.
Goode, L. (2009) “Social news, citizen journalism and democracy.” New Media and Society 11(8), 1–19.
Graves, L., Nyhan, B. and Reifler, J. (2016) “Understanding innovations in journalistic practice: A field
experiment examining motivations for fact-checking.” Journal of Communication, 66(1), 102–138.
Hanitzsch, T. (2007) “Deconstructing journalism culture: Toward a universal theory.” Communication Theory,
17(4), 367–385.
Harder, R. A., Sevenans, J. and van Aelst, P. (2017) “Intermedia agenda setting in the social media age: How
traditional players dominate the news agenda in election times.” The International Journal of Press/Politics.
Retrieved from http://journals.sagepub.com/doi/abs/10.1177/1940161217704969
Hargittai, E. and Hinnant, A. (2008) “Digital inequality: Differences in young adults’ use of the Internet.”
Communication Research, 35(5), 602–621.
Hindman, D. B. (2000) “The rural-urban digital divide.” Journalism and Mass Communication Quarterly, 77(3),
549–560.
Hiippala, T. (2017) “The multimodality of digital longform journalism.” Digital Journalism, 5(4), 420–442.
Hirsch, P. M. (1981) “Distinguishing good speculation from bad theory: Rejoinder to Gerbner et al.” Com-
munication Research, 8(1), 73–95.
Hjarvard, S. (2013) The Mediatization of Culture and Society. New York, NY: Routledge.
Ho, S. S. and McLeod, D. M. (2008) “Social-psychological influences on opinion expression in face-to-face
and computer-mediated communication.” Communication Research, 35(2), 190–207.
Hwang, Y. and Jeong, S.-H. (2009) “Revisiting the knowledge gap hypothesis: A meta-analysis of thirty-five
years of research.” Journalism and Mass Communication Quarterly, 86(3), 513–532.
Jacobson, S. (2013) “Does audience participation on Facebook influence the news agenda? A case study of
the Rachel Maddow Show.” Journal of Broadcasting and Electronic Media, 57(3), 338–355.
Jamieson, P. E. and Romer, D. (2014) “Violence in popular U.S. prime time TV dramas and the cultivation
of fear: A time series analysis.” Media and Communication, 2(2), 31–41.
Jeffres, L. W. (2015) “Mass communication theories in a time of changing technologies.” Mass Communica-
tion and Society, 16(2), 523–530.
Jeffres, L. W., Neuendorf, K. and Atkin, D. J. (2012) “Acquiring knowledge from the media in the Internet
age.” Communication Quarterly, 60(1), 59–79.
Kahlor, L. and Eastin, M. S. (2011) “Television’s role in the culture of violence toward women: A study of
television viewing and the cultivation of rape myth acceptance in the United States.” Journal of Broadcast-
ing and Electronic Media, 55(2), 215–231.
Kampf, Z. and Daskal, E. (2014) “Communicating imperfection: The ethical principles of news correc-
tions.” Communication Theory, 24(2), 165–185.
Karlsson, M., Bergström, A., Clerwall, C. and Fast, K. (2015) “Participatory journalism – the (r)evolution
that wasn’t: Content and user behavior in Sweden 2007–2013.” Journal of Computer-Mediated Communica-
tion, 20(3), 295–311.
497
Jane B. Singer
Kim, S.-H. (2012) “Testing fear of isolation as a causal mechanism: Spiral of silence and genetically modified
(GM) foods in South Korea.” International Journal of Public Opinion Research, 24(3), 306–324.
Kruikemeier, S. and Lecheler, S. (2016) “News consumer perceptions of new journalistic sourcing techniques.”
Journalism Studies. Retrieved from www.tandfonline.com/doi/abs/10.1080/1461670X.2016.1192956
Lasorsa, D. L., Lewis, S. C. and Holton, A E. (2012) “Normalizing Twitter: Journalism practice in an emerg-
ing communication space.” Journalism Studies, 13(1), 19–36.
Lee, J. K. and Coleman, R. (2014) “Testing generational, life cycle and period effects of age on agenda set-
ting.” Mass Communication and Society, 17(1), 3–25.
Lewis, S. C., Holton, A. E. and Coddington, M. (2014) “Reciprocal journalism: A concept of mutual
exchange between journalists and audiences.” Journalism Practice, 8(2), 229–241.
Lewis, S. C., Kaufhold, K. and Lasorsa, D. L. (2010) “Thinking about citizen journalism: The philosophical
and practical challenges of user-generated content for community newspapers.” Journalism Practice, 4(2),
163–179.
Liu, J. (2016) “Digital media, cycle of contention, and sustainability of environmental activism: The case of
anti-PX protests in China.” Mass Communication and Society, 19(5), 604–625.
Livingstone, S. (2004) “The challenge of changing audiences: Or, what is the audience researcher to do in
the age of the Internet?” European Journal of Communication, 19(1), 75–86.
Marchionni, D. M. (2013) “Journalism-as-a-conversation: A concept explication.” Communication Theory,
23(2), 131–147.
Matthes, J., Rios Morrison, K. and Schemer, C. (2010) “A spiral of silence for some: Attitude certainty and
the expression of political minority opinions.” Communication Research, 37(6), 774–800.
McCombs, M. E. and Shaw. D. L. (1972) “The agenda-setting function of mass media.” Public Opinion
Quarterly, 36(2), 176–187.
McDevitt, M., Kiousis, S. and Wahl-Jorgensen, K. (2003) “Spiral of moderation: Opinion expression in
computer-mediated discussion.” International Journal of Public Opinion Research, 15(4), 454–470.
McKay Semmler, K., Semmler, S. M. and Kim, Y. Y. (2014) “Local news media cultivation of host receptivity
in Plainstown.” Human Communication Research, 40(2), 188–208.
McLeod, J. M., Becker, L. B. and Byrnes, J. E. (1974) “Another look at the agenda-setting function of the
press.” Communication Research, 1(2), 131–166.
Meraz, S. (2009) “Is there an elite hold? Traditional media to social media agenda setting influence in blog
networks.” Journal of Computer-Mediated Communication, 14(3), 682–707.
Miyata, K., Yamamoto, H. and Ogawa, Y. (2015) “What affects the spiral of silence and the hard core on
Twitter? An analysis of the nuclear power issue in Japan.” American Behavioral Scientist, 59(9), 1129–1141.
Morgan, M. and Shanahan, J. (2010) “The state of cultivation.” Journal of Broadcasting and Electronic Media,
54(2), 337–355.
Morgan, M., Shanahan, J. and Signorielli, N. (2009) “Growing up with television: Cultivation processes.”
In J. Bryant and M. B. Oliver (eds.), Media Effects: Advances in Theory and Research (3rd ed.). New York,
NY: Routledge (pp. 34–49).
Morgan, M. Shanahan, J. and Signorielli, N. (2015) “Yesterday’s new cultivation, tomorrow.” Mass Com-
munication and Society, 18(5), 674–699.
Nielsen, R. K. and Ganter, S. A. (2017) “Dealing with digital intermediaries: A case study of the relations
between publishers and platforms.” New Media and Society. Retrieved from http://journals.sagepub.com/
doi/full/10.1177/1461444817701318
Noelle-Neumann, E. (1993) The Spiral of Silence: Public Opinion – Our Social Skin. Chicago, IL: University
of Chicago Press.
Norris, W. (2017) “Digital humanitarians: Citizen journalists on the virtual front line of natural and human-
caused disasters.” Journalism Practice, 11(2/3), 213–228.
Oliver, M. B. and Armstrong, G. B. (1995) “Predictors of viewing and enjoyment of reality- based and
fictional crime shows.” Journalism and Mass Communication Quarterly, 72(3), 559–570.
Paulussen, S. and Ugille, P. (2008) “User generated content in the newsroom: Professional and organisational
constraints on participatory journalism.” Westminster Papers in Communication and Culture, 5(2), 24–41.
Phillips, A. (2012) “Faster and shallower: Homogenisation, cannibalisation and the death of reporting.”
In P. Lee-Wright, A. Phillips and T. Witschge (eds.), Changing Journalism. London: Routledge (pp.
81–98).
Potter, W. J. (1993) “Cultivation theory and research: A conceptual critique.” Human Communication
Research, 19(4), 564–601.
Price, V. and Allen, S. (1990) “Opinion spirals, silent and otherwise: Applying small-group research to public
opinion phenomena.” Communication Research, 17(3), 369–392.
498
Prochazka, F., Weber, P. and Schweiger, W. (2016) “Effects of civility and reasoning in user comments on
perceived journalistic quality.” Journalism Studies. Retrieved from www.tandfonline.com/doi/abs/10.10
80/1461670X.2016.1161497
Putnam, R. D. (2000) Bowling Alone: The Collapse and Revival of American Community. New York, NY: Simon
and Schuster.
Revers, M. (2014) “The twitterization of news making: Transparency and journalistic professionalism.”
Rinke, E. M. (2016) “The impact of sound-bite journalism on public argument.” Journal of Communication,
66(4), 625–645.
Robinson, S. (2011) “‘Journalism as process’: The organizational implications of participatory online news.”
Journalism and Communication Monographs, 13(2), 137–210.
Salmon, C. T. and Neuwirth, K. (1990) “Perceptions of opinion ‘climates’ and willingness to discuss the
issue of abortion.” Journalism Quarterly, 67(3), 567–577.
Scheufele, D. A. and Moy, P. (2000) “Twenty-five years of the spiral of silence: A conceptual review and
empirical outlook.” International Journal of Public Opinion Research, 12(1), 3–28.
Scheufele, D. A. and Tewksbury, D. (2007) “Framing, agenda setting and priming: The evolution of three
media effects models.” Journal of Communication, 57(1), 9–20.
Schulz, A. and Roessler, P. (2012) “The spiral of silence and the internet: Selection of online content and
the perception of the public opinion climate in computer-mediated communication environments.”
International Journal of Public Opinion Research, 24(3), 346–367.
Scott, J., Millard, D. and Leonard, P. (2015) “Citizen participation in news: An analysis of the landscape of
online journalism.” Digital Journalism, 3(5), 737–758.
Severin, W. J. and Tankard J. W. Jr. (2001) Communication Theories: Origins, Methods and Uses in the Mass
Media (5th ed.). New York, NY: Longman.
Shanahan, M. C. (2011) “Science blogs as boundary layers: Creating and understanding new writer and
reader interactions through science blogging.” Journalism, 12(7), 903–919.
Shrum, L. J., Lee, J., Burroughs, J. E. and Rindfleisch, A. (2011) “An online process model of second-
order cultivation effects: How television cultivates materialism and its consequences for life satisfaction.”
Human Communication Research, 37(1), 34–57.
Singer, J. B. (2016) “Transmission creep: Media effects theories and journalism studies in a digital era.”
Journalism Studies. Retrieved from www.tandfonline.com/doi/abs/10.1080/1461670X.2016.1186498
Singer, J. B. and Ashman, I. (2009) “‘Comment is free, but facts are sacred’: User-generated content and
ethical constructs at the Guardian.” Journal of Mass Media Ethics, 24(1), 3–21.
Steensen, S. and Ahva, L. (2015) “Theories of journalism in a digital age: An exploration and introduction.”
Tichenor, P. J., Donahue, G. A. and Olien, C. N. (1970) “Mass media flow and differential growth in
knowledge.” Public Opinion Quarterly, 34(2), 159–170.
Tran, H. (2013) “Does exposure to online media matter? The knowledge gap and the mediating role of
news use.” International Journal of Communication, 7. Retrieved from http://ijoc.org/index.php/ijoc/
article/view/1550/886
Tsfati, Y., Stroud, N. J. and Chotiner, A. (2014) “Exposure to ideological news and Perceived opinion
climate: Testing the media effects component of spiral-of-silence in a fragmented media landscape.”
International Journal of Press/Politics, 19(1), 3–23.
van Deursen, A. J. and van Dijk, J. A. (2014) “The digital divide shifts to differences in usage.” New Media
and Society, 16(3), 507–526.
Vergeer, M. and Pelzer, B. (2009) “Consequences of media and Internet use for offline and online network
capital and well-being: A causal model approach.” Journal of Computer-Mediated Communication, 15(1),
189–210.
Vis, F. (2013) “Twitter as a reporting tool for breaking news: Journalists tweeting the 2011 UK riots.” Digital
Journalism, 1(1), 27–47.
Viswanath, K. and Finnegan, J. R. Jr. (1996) “The knowledge gap hypothesis: Twenty-five years later.” In
B. R. Burleson (ed.), Communication Yearbook 19. New York, NY: Routledge (pp. 187–227).
Viswanath, K., Kahn, E., Finnegan, J. R. Jr., Hertog, J. and Potter, J. D. (1993) “Motivation and the
knowledge gap: Effects of a campaign to reduce diet-related cancer risk.” Communication Research, 20(4),
546–563.
Wahl-Jorgensen, K. and Hanitzsch, T. (2009) “Introduction: On why and how we should do journalism
studies.” In K. Wahl-Jorgensen and T. Hanitzsch (eds.), The Handbook of Journalism Studies. New York,
NY: Routledge (pp. 3–16).
499
Jane B. Singer
Walgrave, S. and van Aelst, P. (2006) “The contingency of the mass media’s political agenda setting power:
Toward a preliminary theory.” Journal of Communication, 56(1), 88–109.
Weaver, D., McCombs, M. and Shaw, D. L. (2008) “Agenda-setting research: Issues, attributes, and influ-
ences.” In L. L. Kaid (ed.), Handbook of Political Communication Research. Mahwah, NJ: Lawrence Erlbaum
Associates (pp. 257–282).
Wei, L. and Hindman, D. B. (2011) “Does the digital divide matter more? Comparing the effects of new
media and old media use on the education-based knowledge gap.” Mass Communication and Society, 14(2),
216–235.
Weimann, G., Weiss-Blatt, N., Mengistu, G., Tregerman, M. M. and Oren, R. (2014) “Reevaluating ‘the end
of mass communication?’.” Mass Communication and Society, 17(6), 803–829.
Westley, B. H. and MacLean, M. S. Jr. (1957) “A conceptual model for communications research.” Journal-
ism Quarterly, 34(1), 31–38.
Widholm, A. (2016) “Tracing online news in motion: Time and duration in the study of liquid journalism.”
Williams, D. (2006) “Virtual cultivation: Online worlds, offline perceptions.” Journal of Communication, 56(1),
69–87.
Yan, Y. and Bissell, K. (2015) “The sky is falling: Predictors of news coverage of natural disasters world-
wide.” Communication Research. Retrieved from http://journals.sagepub.com/doi/abs/10.1177/009365
0215573861
Yoo, S. W. and Gil de Zúñiga, H. (2014) “Connecting blog, Twitter and Facebook use with gaps in knowl-
edge and participation.” Communication and Society, 27(4), 33–48.
Zelizer, B. (2004) Taking Journalism Seriously: News and the Academy. Thousand Oaks, CA: Sage Publications.
500
38
OUTSOURCING CENSORSHIP
AND SURVEILLANCE
The privatization of governance as an
information control strategy in the
case of Turkey1
Aras Coskuntuncel
When Turkey’s Justice and Development Party (AKP) attempted to purge almost all dissent
in the aftermath of the July 2016 coup attempt,2 the state-run news agency, Anadolu Agency
(2016), reported that police arrested 1,213 users because of their online activities, while Twitter
and Facebook provided real-time data to help Turkish authorities identify these citizens. This
was but one of the many examples of the privatization of governance online – outsourcing sur-
veillance, censorship, and law enforcement functions to information intermediaries – in which
the private intermediaries serve the political power. As part of the strategy to control the flow
of information, the government’s efforts to privatize governance, surveil, and organize state-
sponsored information campaigns intensified after the Gezi Park uprising in 2013. At that point,
the party was already enjoying a clientelist3 relationship with traditional media conglomerates,
and since the 2002 elections, the AKP has given particular significance to exerting control over
the flow of information and instrumentalizing media as a whole in favor of powerful interests
while consolidating its undemocratic grip on power. This chapter analyzes the privatization of
governance as one of the realigned strategies of information control in the digital era in the case
of Turkey, as the political and economic elites adjust to incorporate digital technologies into the
task of perpetuating power relations.
The AKP regime faced one of the biggest challenges in its 16-year tenure during the nation-
wide Gezi movement in 2013. What started as a sit-in to protest the government’s plans to priva-
tize and demolish a park in central Istanbul in order to build a shopping mall turned quickly into
widespread discontent against the AKP’s neoliberal and anti-democratic policies in particular and
the qualities and politics of urban daily life in general – from neoliberal urbanization projects
without any input from citizens and civil society organizations to undemocratic regulations and
practices affecting citizens’ daily lives. The protests widened as more people became frustrated
by the government’s response, its excessive use of police force, and the mainstream media’s per-
formance. For example, after days of clashes and police violence, on June 7, 2013, seven national
newspapers bannered almost identical headlines citing the prime minister’s reaction to the pro-
tests: “I would give my life for the demands of democracy”. These stories implied that the
government was the real victim and the real guarantor of democracy. When the police attacks
501
Aras Coskuntuncel
intensified on the third day of the protests, CNN’s Turkish franchise, CNN Turk, opted to air a
documentary about penguins. The penguin was then made a symbol of the protests by the pro-
testers, who turned to anti-government outlets and seized upon digital media, from livestreaming
to micro blogging, to create spaces to express themselves, report from the ground, coordinate, and
disseminate information and their narratives while both bypassing and demanding mainstream
media coverage. Networking sites, specifically Twitter, became a crucial component of the pro-
tests for all sides; the number of Twitter users increased by almost 9 million during the peak of
the protests between May 29 and June 10 (Kuzuoglu, 2013).
After the protests, the demonstrators discovered that authorities were monitoring and detain-
ing many of them based on their interactions on networking sites; for example, people were
posting their addresses so that those who were injured or fleeing police could find sanctuary.
Volunteers created mobile first aid stations and tweeted their locations. Protesters and residents
posted the locations of safe corners, which streets needed barricades, and where police were
gathering. Government agencies, meanwhile, tracked these interactions so that later they could
detain and/or investigate those who were involved, and they did. Twitter, YouTube, and Facebook
shuttered accounts and censored statuses per government request, including newspapers and
journalists’ accounts and statuses, while Twitter introduced a ‘country-withheld content’ tool,
which blocks content in a specific country, in Turkey (E. Sozeri, 2015). In addition to delegated
censorship and access restrictions, the AKP also utilized new technologies for surveillance and
state-sponsored propaganda campaigns. These efforts include forming online media teams and
creating and strengthening extralegal relationships with media companies. The telecommunica-
tion agency and the country’s biggest internet service provider (ISP), meanwhile, integrated their
servers, and the national police launched a smartphone app for citizens to snitch on each other’s
online posts and accounts.
An overall analysis of the struggle to control the flow of information during recent protest
movements in various countries reveals a similar pattern: digital media were crucial to the often
radicalized, oppressed, and surveilled groups from Ferguson to Istanbul and São Paulo as they
created a space to express themselves and utilized these technologies to coordinate. But this is
also exactly where they are vulnerable; the same digital technologies are the tools of surveillance,
control, and exploitation – again, predominantly of the same groups. The Turkish case within
the broader trend illustrates a pattern of increasingly sophisticated tactics and regeared strategies
of control that strongly counter the democratic potential of interactive technologies (Deibert
et al., 2010; Tufekci, 2014; Tsui, 2015; Coskuntuncel, 2016).
Ronald Deibert et al. (2010) identify “three generations” of gradually more sophisticated
controls by analyzing the “colonization of cyberspace” within the 56 states of the Organization
for Security and Co-operation in Europe (OSCE). The “first generation of internet control
techniques” includes filtering and blocking, while the second and third generations include cre-
ating legal and technical capabilities for content removal requests, network attacks, surveillance,
and counter-information campaigns. Turkey effectively uses all three generations of control. But
the internet is not a free-floating entity, and the issue of control is not solely about governmental
censorship; claiming otherwise and ignoring the role of the private companies results in the fail-
ure to analyze the process of the privatization of governance, which is progressing in Turkey and
which has already been realized to a much greater extent in the United States.
Private internet intermediaries are playing a central role in regulating content, governing
expression, and carrying out law enforcement functions – often through the arrangements of
technical infrastructure – on their own volition (DeNardis, 2012, 2014), and part of this priva-
tized governance is governance in service of the ruling elite. The ruling political, bureaucratic,
and economic elite establish a relationship similar to their relationship with mainstream media
and outsource censorship and surveillance to private information intermediaries using legal cases,
502
Outsourcing censorship and surveillance
access restrictions, and flak.4 As DeNardis points out (2015), the internet is a control network
as much as it is a communication network. In the context of recent protest movements around
the world, it is not only a tool for revolt but also a means of suppression. Just like with many
other technologies, these contradictory politics are perpetually reproduced by the social relations
within which the medium operates; technologies shape our relations to each other and to nature,
while those same technologies are being shaped by the same relations. In this chapter, I first
address the prevalent ‘digital democracy’ discourse and then analyze the political economy of the
Turkish media and the relationship between the government and media companies.
The ‘digital democracy’ discourse

The ‘digital democracy’ discourse defines democratization and empowerment as inherent within
the new communication and information technologies. The progressive potential of these tech-
nologies, just like previous ones, is by no means insignificant, but there is a difference between
analyzing how these technologies open up new possibilities for democratization and inclusion on
one hand and claiming that the potential of these technologies is already realized on the other. It
is crucial to analyze how these technologies are owned and operated and to explore the changes,
continuities, and contradictions in digital media.
Considering the continuing ownership concentration and commercialization, increasingly
sophisticated state-sponsored information campaigns, and privatization of governance and sur-
veillance online, the contemporary struggles to control the flow of information look quite
different than they were once envisioned in both popular and academic discourses. Nicholas
Negroponte (1996), for example, talked about the upcoming death of media barons; Time maga-
zine (2006) claimed a shift in power in cultural industries as the magazine celebrated Web 2.0
users “for seizing the reins of the global media, for founding and framing the new digital democ-
racy” (Grossman, 2006); and Henry Jenkins (2006), Yochai Benkler (2006), and Manuel Castells
(1996, 2008), among others, suggested a rigid distinction between “old” and “new” media based
on passivity vs. interactivity and control vs. emancipation. Castells, for example, defined emanci-
pation within the medium’s technological features and argued that the medium’s architecture will
resist commercialization (1996: 356). He also claimed the recent protest movements spontane-
ously emerged from and were based on the internet (2012: 106, 229). This popular tendency to
treat recent uprisings as isolated and sudden awakenings ignores the long-term struggles, turmoil,
and even wars in various contexts leading up to the protest movements. It also undervalues the
use of other media available to the protesters and the crucial role of face-to-face interaction.
These accounts are more heavily centered on the supposedly inherent qualities of technolo-
gies than the conditions and the relations within which they operate and are deployed. Jenkins
(2006) repeated the claims that, thanks to the internet, “new consumers” are now empowered,
creative, and rebellious, while Benkler (2006) argued that digital technologies are changing the
mode of production. Meanwhile, in the same year Benkler, Jenkins and Time magazine pub-
lished their celebratory accounts about digital technologies, media mogul Rupert Murdoch, then
the founder, chairman, and CEO of one of the world’s largest media conglomerates, purchased
MySpace for more than a half-billion dollars and also hailed the “new media” by stating: “Tech-
nology is shifting power away from the editors, the publishers, the establishment, the media elite.
Now it’s the people who are taking control” (Reiss, 2006).
The digital democracy discourse sees new technologies and the ways to appropriate existing
goods and services as a source of value instead of as the relationship between capital and labor.
According to celebratory accounts, the internet is inherently changing and democratizing power
relations as the audiences of ‘old media’ become active participants; such accounts are shared
across disciplines and the political spectrum. This ahistorical dichotomy between ‘old media’ and
503
Aras Coskuntuncel
‘new media’ and the rigid understanding of ‘top-down’ vs. ‘bottom-up’ media neglect the power/
resistance dialectic, ignore the reproduction of ‘old media’ in ‘new media’, and assume that the
digital media is inherently resistant to being controlled by elites. This logic also suggests that there
is no need to worry about the workings of power. As Darin Barney rightly emphasizes, “[I]f this
is true – if network technology is inherently revolutionary – it leads one to wonder why existing
bureaucratic and corporate elites are so enthusiastic about, and so heavily invested in, the success
of this technology” (2000: 19). Mark Andrejevic notes, despite the “revolutionary promise of
participatory media [. . .] power relations remain largely unaltered” (2009: 1); in fact, in many
contexts, they are strengthened.
What these variants of the ‘digital democracy’ discourse are missing is not only an analysis
of the changes and continuities but also the power/resistance dialectic that existed in the ‘old
media’ era. The sharp contrast between critical interactivity and passive consumption related to
the ‘old’ and ‘new’ media is ahistorical and mechanic. The commercial characteristics of active
participation may very well lead to a new kind of passivity that neglects critical political thinking
and engagement (Barney, 2000; Sterne, 2012). And, as Zachary Glass (2007) argues in his study
of product placement in video games, passively viewing can also stimulate critical engagement.
From Jenkins to Time magazine, the proponents of the digital democracy discourse essentially
see the new information and communication technologies as a source of value instead of social
relations and perceive social relations as relations between things, which is a form of commod-
ity fetishism. This fetishism of technology also breeds the emphasis on the appearance (signifier,
medium, form) over the underlying social relations.
“We endow technologies – mere things – with powers they do not have (e.g., the ability to
solve social problems, to keep the economy vibrant, or provide us with a superior life)” (Har-
vey, 2003: 3). Today, the internet might be one of the most fetishized technologies. Technology,
according to Marx,
reveals the active relation of man to nature, the direct process of the production of his
life, and thereby it also lays bare the process of the production of the social relations of
his life, and of the mental conceptions that flow from those relations.
(1990: 493)
In Marx’s analysis, technology itself neither determines any particular social outcome nor is it
“some free-floating deus ex machina”; all these elements are interconnected “in the construction
and reproduction of the social order” (Harvey, 2003: 4). Highlighting the embeddedness between
technologies and the conditions of their deployment, Marx states that “machinery in itself short-
ens the hours of labor, but when employed by capital, it lengthens them” (1990: 568). Instead
of a biased approach toward particular features of technologies, Sheila Jasanoff ’s (2004) concept
of the coproduction of technology and social order is also useful when considering how these
relations are dialectically intertwined. Based on Marx’s explanation of how superior technology
can be a source of excess profits for capitalists only for a while, Harvey points to the difference
between seeing the machines themselves as a source of value and understanding that “profit arises
out of the social relation between capital and labor” (2003: 7).
Although no approach should rule out the technologies at hand as contributors to the solution,
bold claims of transformation tend to ignore the power relations behind them. Matthew Hindman,
through various data sources, shows that the internet actually is not capable of inherently chal-
lenging the monopoly of elites, nor will it inherently amplify “the political voice of the ordinary
citizens” (2009: 6). John Goldsmith and Tim Wu (2006) and Laura DeNardis (2014) show how
cyberspace is already controlled by private intermediaries, often regulating the flow of information
through the arrangements of the infrastructure in favor of the bureaucratic and financial elites.
504
Joseph Turow (2012) provides a detailed explanation of how commercialization transforms cyber-
space into a mechanism to stimulate consumption and commodify desires. McChesney (2013)
analyzes the increasingly concentrated and commercialized corporate media system in relation to
media’s democratic potential (see also: Pickard, this volume, Chapter 16). Schiller (2000, 2007),
Andrejevic (2007, 2013), and Fuchs (2014) offer excellent critiques of digitalization and the new
dynamics of the centralization of power from various perspectives including surveillance, com-
modification, and labor. Rather than endowing powers such as social change to particular features
of a technology, it is still important to analyze how these technologies are owned and operated.
Media ownership
Capitalism has monopolistic tendencies. Cultural industries have been no exception, and owner-
ship concentration, including conglomeration, in the information and communications sector is
well documented (Bagdikian, 2004; McChesney, 2004; Noam, 2009). Concentrated ownership
is crucial as it is closely related to “the central problem of democratic media” (Baker, 2007: 205).
The immediately visible issues related to ownership and control are the concentration of the
right to hire and fire, of decisions over budgets and allocation of resources, and of lobbying power
(Hesmondhalgh, 2013. Moreover, especially as they relate to the news media, some of the social
and political implications of concentrated private control are the marginalization of the voices
and issues of disadvantaged groups (McChesney, 2012); the standardization of news choices and
the dominance of corporate logic in the public sphere (Allen, 2005); and a heavy reliance on
official, elite sources (Greider, 1992). As Edwin Baker points out, diversity in media ownership
“is most importantly a process value, not a commodity value” (2007: 16, emphasis in original). In
other words, the issue of ownership is not necessarily about diversity of content or viewpoints but
a “fairer, more democratic allocation of communicative power” based on the basic democratic
principle of one-person/one-vote. Also, in concentrated media systems, bottom-line investment
tends to be more extreme, and those few owners’ choices for profitability are standardized. And
finally, conglomerates are vulnerable to outside pressure within the market system (Baker, 2007).
Another issue that applies to both the so-called ‘old’ and ‘new’ media related to concentrated
ownership is the creation of “extralegal relationships” between the owners of media companies
and authorities, which Benkler (2011) and DeNardis (2012) discuss in the context of the internet
governance. The internet is controlled by information intermediary companies that perform
the role of gatekeeping and increasingly carry out law enforcement functions (DeNardis, 2014).
Internet service providers perform deep-packet inspections, search engines take down content,
networking sites block content and accounts, and credit companies terminate money transac-
tions; the crucial point is that these companies are increasingly acting on their own volition,
often in favor of the interests of the political power. A well-known example of the privatization
of governance is the WikiLeaks case (DeNardis, 2012, 2014). In 2010, WikiLeaks, a prominent
whistle-blowing site, published U.S. diplomatic cables that irked the U.S. government. During
the CableGate saga, MasterCard, PayPal, Visa, Amazon, and internet services firm EveryDNS.net
terminated their services to the whistle-blowing site, to the benefit of the U.S. government, with-
out any government or legal action. Moreover, internet intermediaries “perform governance
function” not only when enacting censorship – per government request or not – but also when
they decide to deny censorship requests (DeNardis, 2014: 159).
When governance is privatized, both the government and corporations appear unaccountable
because private intermediaries are neither under democratic scrutiny from voters nor subject to
constitutional and other institutional confinements. In March 2016, WikiLeaks published a data-
base of more than 30,000 emails and attachments sent from Hillary Clinton’s private server. One
of the communiques between the State Department and Google reveals how the digital giant
505
Aras Coskuntuncel
reached out to the State Department to promote a new tool it created to encourage “more to
defect” from the Syrian army, “giving confidence to the opposition”, in line with U.S. govern-
ment interests and regime change goals abroad; the email was then circulated among high-level
officials (WikiLeaks, 2016b). Concentrated media ownership and a conglomeration model, in
which most media are owned by corporations that own other businesses and which is rife with
extralegal ties, are central to the discussion of the strategies of information control in Turkey.
Turkish media environment

The Turkish media environment is neoliberal and fraught with censorship and self-censorship.5
Due to the (de)regulation periods of 2002 and 2011, today there are almost no legal restrictions
on media ownership in the country.6 Most media products and therefore most advertising revenue
are owned by a handful of corporations, and this trend is only exacerbated after the post-coup
attempt purges.7 These conglomerates also own many other businesses, which account for their
main sources of revenue (see Table 38.1). The Demiroren group, one of the country’s energy and
construction giants, for example, in addition to owning newspapers, television channels, radio
stations, and digital platforms, also owns many companies in different sectors, including energy,
industry, tourism, and construction. Other media conglomerates also own dozens of companies
from construction to banking, most of which are dependent on state contracts and privatization.
As a result, these media outlets are more vulnerable to pressure and likely to promote government
interests in order to secure the government’s favor in public tenders and accumulate more profits.
Table 38.1 Demiroren Group’s investments
Company Media Investments Other Investments
Demiroren Newspaper:6 national titles Energy: 3 companies

Group (Hurriyet, Posta, Fanatik, Vatan, (Milangaz, Milan Petrol, Total Oil)
Milliyet, Hurriyet Daily News) Industry: 4 companies
News Agency: 1 (DHA) (Demiroren Heavy Metal, Parsat Piston,
TV Broadcasting: 7 channels MS Motor Services, D-Mermer)
(Kanal D, Euro D, Kanal D Real Estate and Construction: 3
Romania, CNN Turk, teve2, companies
DreamTV, Boomerang Turkiye) (Lidya Flats, Lidya Construction,
Radio Broadcasting: 2 stations Seyhan Park Residences)
(Radyo D, CNN Turk Radyo) Tourism: 2 companies
Digital TV Platform: 1 (Kemer Country, Demiroren Istiklal
Palas)
(D-Smart)
Shopping Mall: 1
Online Platforms: Almost two
(Demiroren Istiklal)
dozen platforms including web and
news portals, e-commerce, and stock Port Management: 2 companies
quote services (Zeyport, Dolfen Dock)
Printing and Distribution: 2 Education: 1 private school
companies (includes kindergarten, elementary,
(Dogan Distribution, Dogan middle, and high school).
Printing Center)
Advertising: 1 company
(MedyaNet)
Source: Compiled from Demiroren.com.tr, mulksuzlestirme.org, and bianet.org

506
The Demiroren Group bought two high-circulation newspapers, Milliyet and Vatan, from
another conglomerate Dogan Group in 2011 and the rest of Dogan’s media investments in 2018.
After both acquisitions , Demiroren fired journalists it saw as critical of the government.8 The
company, meanwhile, went on to increase its energy and construction investments, including the
bid to build one of the biggest seaports in Istanbul. In another example, Zirve Holding’s media
outlets are known for their pro-government coverage; the conglomerate periodically wins many
infrastructure and road construction bids throughout the country in addition to being part of the
consortium that won the bid to construct Istanbul’s third airport. Zirve also controls hydroelectric
power plants, has investments in international water and gas pipelines in the region, and was the
contractor of the Taksim development plan, which eventually led to the Gezi Park protests. The
Albayrak group, which also owns pro-government media outlets, frequently bids for and wins
public tenders including defense and urban development projects. Dogus Holding, the owner of
the NTV news channel, which sparked anger during Gezi protests because of its pro-Erdogan
coverage, also owns other television and radio stations, magazines, and online portals in addition
to power plants and banking, tourism, and construction companies. As a result, even before the
Demiroren’s acquisition of Dogan Media, at least 7 out of 10 most-read print newspapers, 7 out
of 10 most-watched TV channels, and 5 out of 10 most-read online news portals were owned by
companies that support or are affiliated with the AKP government, according to a joint report by
the Reporters Without Borders and Bianet (2016).
Because of the clientelist relationship that arises from the ownership structure and neoliberal
policies, the ruling elites often do not have to directly enforce their views; media companies perform
the role of governing the information flows themselves on behalf of the bureaucratic, political, and
financial elites. In exchange, these conglomerates gain the government’s favor and accumulate more
influence and profits. This delegated governance is not unique to Turkey, nor is it changed by new
technologies. Overtly or covertly, the privatization of governance is the product of this narrow alli-
ance of bureaucracy and private capital globally, of profit and control, and the Turkish government
is establishing a similar privatization of governance model with digital media companies, using
access restrictions in addition to legal and economic means and attacks from government officials.
Herman and Chomsky (2002) identify five filters in their analysis of the performance of mass
media in the United States. These five filters are (1) concentrated ownership and profit orienta-
tion, (2) dependence on advertising as the primary income source, (3) heavy reliance on elite
sources and official debate, (4) the use of flak by government and business figures to discipline the
media, and (5) ideologies of anti-communism and anti-terror. Through these filters, money and
power influence the flow of information, “marginalize dissent, and allow government and domi-
nant private interests to get their message across the public” (2002: 2). In Turkey, under a heavily
concentrated media ownership structure, their model operates visibly. Moreover, it appears that
the information control model at work has incorporated interactivity, which digital democracy
proponents purport should have rendered these filters moot.
The AKP and digital media

In addition to the already strong presence of government supporters on various networking sites,
the Turkish government recruits digital media teams to propagate, harass, and deter critics and
journalists, surveil citizens, and drive the discussion online. After the Gezi protests, The Wall Street
Journal reported that the government was “recruiting a 6,000-member social media team to woo
citizens and fight critics”; a senior party official told the newspaper in 2013 that those teams
would focus on Twitter, Facebook, YouTube, and Instagram to “promote the party perspective
and monitor online discussions” (Albayrak and Parkinson, 2013). In 2016, WikiLeaks published
Energy Minister Berat Albayrak’s hacked emails in a searchable archive, which demonstrated not
507
Aras Coskuntuncel
only the extralegal relationship between the government and media but also revealed details of
the AKP’s social media strategies. According to the emails, the AKP setup teams consisting of
experts (coders, designers, script writers, psychological warfare experts), users, and Twitter bot
accounts to disseminate pro-government propaganda, intimidate dissidents, and create a network
of informants (WikiLeaks, 2016a). According to a member of the Parliament’s security and intel-
ligence committee, by the beginning of 2017, the authorities prepared summaries of proceed-
ings for 17,000 users because of their online activities, and another 45,000 accounts were being
monitored (Cumhuriyet, 2017). And after the contested presidential referendum,9 access to many
accounts and statutes were restricted, especially those who opposed the decisions of the country’s
election authority and those who posted from the demonstrations held against the outcome
(Birgun, 2017).
Almost 80% of the Turkish population lives in urban settings, and the country has Europe’s
largest young population (World Bank, 2015). Turkey’s internet penetration rate was 61%, and
82% of the individuals participated in social networks one way or another in 2016, according to
the Turkish Statistical Institute (TUIK, 2016). In addition to restrictive deployment of the penal
code, anti-terror laws, and intellectual property laws in terms of speech and media, Turkey’s
internet-specific legislation and regulations also have grown more control- and surveillance-
oriented since the first internet law in 2007. In 2008, the country’s telecommunication authority
and the ISPs synced their infrastructure against online crimes such as child pornography; in 2014,
the content of these integrated servers between the government and the private ISPs expanded to
include deep-packet inspections; and the March 2015 amendments allowed the telecommunica-
tion agency to ban sites based on government request unless a court intervenes. And according
to a leaked email from California-based tech firm Procera Networks, the Turkish government,
through the country’s largest ISP, could perform an “NSA-grade” deep-packet inspection (Fox-
Brewster, 2016). Moreover, the ISPs, the telecommunication authority, “and by proxy the gov-
ernment” developed and deployed more sophisticated control and censorship tools, including
“blocking individual URL addresses instead of banning the entire domain, wholesale banning of
news topics instead of banning news websites”, and bandwidth throttling, false routing, and cut-
ting internet services in specific regions instead of using the nationwide kill-switch option (Yesil
et al., 2017: 9).
The AKP often used legal, economic, and technical tools and tactics together to control the
flow of information, not because the party “hates” the internet or just because these actions “res-
onate” well within Erdogan’s voter base, as some scholars like to claim (see, for example, Tufekci,
2014, 2016) – as if AKP supporters constitute a homogenous body or as if the internet cannot be
deployed as a means to control and suppress. These tools and tactics of information control by the
AKP regime are of course in part attempts to reinforce the party base and block different opin-
ions, but Turkey’s political and economic elites love the internet as a control network, encourage
the use of digital technologies, and use and develop sophisticated digital surveillance and censor-
ship capabilities. Moreover, one of the core reasons for these censorship and access restrictions is
to establish a relationship with digital media companies based on the privatized governance the
AKP enjoys with traditional mainstream media companies.
Turkey made headlines when Twitter was blocked in March 2014 – several months after the
Gezi movement and a series of leaked wiretapped conversations of AKP officials strongly indicat-
ing corruption on all levels – under the pretext of content-related legal cases like defamation.
During talks between the Turkish government and Twitter representatives after the ban, the
Turkish Constitutional Court overruled the ban. But although the ban was lifted (and the num-
ber of users increased, accessed through Tor and VPNs), Twitter not only took down the accounts
and content subject to the court order but also started censoring accounts upon the government’s
508
requests (Bianet, 2014; E. Sozeri, 2015). Twitter at first encouraged Turkish users to continue
tweeting and instructed them how to tweet via SMS text message. After the talks, the company
agreed to strengthen communications with the Turkish government; the then-deputy prime
minister declared, “Twitter now toes the line” (Hurriyet Daily News, 2014a; 2014b). A year after
the first wave of bans, Turkey again in 2015 blocked access to Twitter, this time only for several
hours, along with YouTube and 168 websites, because of another court order banning content
related to the murder of a public prosecutor. Twitter and Google this time were quick to satisfy
Turkish authorities’ demands, and Facebook was quickest to remove the material and therefore
did not face a ban.
Although Twitter releases “transparency reports”, the company does not give details about the
accounts and statuses in question. A study in 2015 on Twitter’s censorship in Turkey, for example,
“discovered over a quarter million censored tweets – two orders of magnitude larger than what
Twitter itself reports” (Tanash et al., 2015: 1). In their efforts to monitor the banned accounts
and tweets, individual researchers identified many anti-government accounts and cases involving
journalists and newspapers’ tweets and accounts in addition to those implicated in criminal cases
(E. Sozeri, 2015; Silverman, 2015). After these and many other occasions, news outlets and blog-
gers started to ask, “Why is Twitter aiding Turkish censorship?” (Silverman, 2015). It is highly
possible that, at the end, these accounts are worthless to Twitter because the user information is
inaccurate or obscured due to the usage of Tor and other services, and the company does not
want to risk falling far behind other networking sites in Turkey.
Conclusion
The concentrated, commercialized media system and surveillance-based business model of the
digital media open up channels for the reproduction and strengthening of the existing power
relations and control. Surveillance is employed increasingly as an extension of workplace manage-
ment in addition to security and policing and reproduces asymmetrical and undemocratic power
relations (Lyon, 1998; Andrejevic, 2007); governance is privatized, through which censorship and
law enforcement are delegated, while both the governments and private intermediaries emerge
unaccountable (DeNardis, 2012, 2014); profits are increasingly derived from rentier activities
online and via offloading previously paid labor onto consumers through digital technologies –
which in addition to deepening old forms of exploitation and alienation, efficiently enable “dicey
work arrangements” (Scholz ed., 2013; Huws, 2014); and state-sponsored information campaigns
easily expand and create an “infoglut” online, in which everything, including facts, turns into just
another opinion and loses its “symbolic efficiency” (Andrejevic, 2013).
The corporate media’s power to mobilize an elite consensus, reaffirm its legitimacy and manu-
facture consent lies in its omnipresence and aggregative nature. Every day through various forms
of content from various media, with all those experts, pundits, and politicians, we consume dif-
ferent versions of the same elite agenda and framework; most of the time even the changes in
the dominant discourse are tied to tensions within elite circles. Hegemony, as Gitlin interprets
Antonio Gramsci’s writings, is “a process that is entered into by both dominators and dominated.
Both rulers and ruled derive psychological and material rewards in the course of confirming and
reconfirming their inequality” (2003). The condition of penetrating the dominant ideologies
into the masses’ common sense and everyday practices in capitalist societies is far from being a
mechanical, top-down process but rather is distilled through net-like structures, where not only
the struggle between the dominant and dominated but also the struggle within the dominant
classes and ideologies takes place. And cultural industries are not only “increasingly sources
of wealth and employment in many economies” but also “involved in making and circulating
509
Aras Coskuntuncel
products that [. . .] have an influence on our understanding and knowledge of the world” (Hes-
mondhalgh, 2013: 4–9) as we reproduce “the world by symbolic work and take up residence
in the work we have produced” (Carey, 2009: 13). At the nexus of these struggles, mainstream-
corporate media specialize in the “production, relaying, and regearing of hegemonic ideology”
(Gitlin, 2003: 254), and the internet, from the arrangements of technical architecture to the
content layer, is not an exception.
From cutting off services to WikiLeaks in accordance with the U.S. government’s interests to
Twitter and Facebook’s censorship to please the Turkish government, private corporations fulfill
the role of governance on behalf of the bureaucratic/financial elites. As a result of this privatiza-
tion of governance, the conditions of “delegated censorship, delegated surveillance, delegated
copyright enforcement and delegated law enforcement” (DeNardis, 2014) have no less significant
outcomes in terms of rights, freedoms, and hegemony than seven national newspapers using the
same front page headline based on the prime minister’s speech about the Gezi Park protests or
CNN’s Turkey franchise airing a documentary about penguins when the first clashes erupted in
Istanbul.
Because systemic change will not happen overnight, the struggle for democratization and
emancipation should start today with realistic goals under current conditions with the tools at
hand and extend into the future, even if it goes against the logic of capitalist accumulation (Foster,
2015). Today, as Andrejevic (2013) also points out, the core of the problem is the commodifica-
tion of previously nonproprietary information that results in the separation of the user from her/
his data, which is being tracked, stored, and sold, and the expansion of the restrictive legal regime
for the enforcement of property rights over this information. The appropriation of this informa-
tion breeds not only privatization of governance and hegemony but also new forms of surveil-
lance, exploitation, and alienation. Short and midterm goals should include the development of
public service digital media without detailed monitoring and tracking.
Further reading
For a critical overview of the history, political economy, and impact of the internet, Misunder-
standing the Internet by James Curran et al. (2016) is a great starting point. Laura DeNardis’s The
Global War for Internet Governance (2014) provides an excellent analysis of the internet governance
struggles and the workings of the layers of the internet from the perspective of power relations.
Matthew Hindman puts bold claims of empowerment and emancipation to the test in his The
Myth of Digital Democracy (2009). In iSpy: Surveillance and Power in the Interactive Era (2007) and
Infoglut: How too Much Information is Changing the Way We Think (2013), Mark Andrejevic power-
fully analyzes and critiques the strategies of control from surveillance and exploitation to data
overload and manipulation. And Edwin Baker, in Media Concentration: Why Ownership Matters
(2007), comprehensively explores the links between democracy and media ownership.
Notes
1 Parts of this chapter are a modified version of text excerpted from Aras Coskuntuncel, “Privatization of
Governance, Delegated Censorship, and Hegemony in the Digital Era: The Case of Turkey,” Journalism
Studies, June 2016.
2 A group of commanders (mostly affiliated with a religious and political group, which was previously an
ally of the government) with mostly clueless foot soldiers in Istanbul and Ankara attempted to take over the
country (during almost prime-time), but they did so without capturing a single government minister and
instead carried out spectacles like occupying the bridges over Bosporus that connect Asia and Europe. The
subsequent purge was nevertheless massive: In just the first week of November 2016, for example, Turkish
510
authorities sacked more than 9,000 public servants, arrested 11 elected members of the second-biggest
opposition party, detained nine journalists and managers of an anti-government daily newspaper, closed 15
media outlets, used throttling to restrict access to networking sites, and blocked popular VPN services – all
based on an extended state of emergency (Ilgun 2016; Nebil 2016).
3 The concept signifies a patron-client relationship in political processes, where public office is treated as
private property so that parties establish their support based on nepotism and reward their supporters with
private goods. For an analysis of clientelism in the context of Japan, see Ethan Scheiner, Democracy Without
Competition in Japan: Opposition Failure in a One-Party Dominant State (Cambridge: Cambridge University
Press, 2006).
4 Edward S. Herman and Noam Chomsky (2002) define “flak” as one of the “filters” of their “propaganda
model”; in this context, it refers to attacks and/or “negative responses to a media statement or program”
through various means, including communicative, legal, and economic, to discipline media.
5 The term “neoliberalism” often is used excessively and inconsistently. Here I refer to the political eco-
nomic practice of (de)regulation, privatization, and marketization, meaning both the withdrawal of the
state from many areas and the creation of the markets by state action if markets do not exist (water, social
security, telecommunications, etc.) in favor of capital accumulation that has been commonplace since the
early 1970s. For a detailed discussion of neoliberalism, see: David Harvey, A Brief History of Neoliberalism
(New York, NY: Oxford University Press, 2005).
6 In 2002, restrictions on cross-ownership were lifted, while in 2011 the cap on foreign capital was raised
to 50%, which combined with various bypass methods eventually resulted in a hyper-neoliberal market in
which both national and international mergers and acquisitions further concentrated the media ownership.
7 Before the controversial coup attempt on July 15, 2016, there were 2,813 local, 107 regional, and 180
national newspapers in Turkey, while there were 39 national, 15 regional, 209 local, 93 cable, and 193 sat-
ellite television channels, and 1,074 radio stations nationwide (TUIK 2014) according to official reports.
Between the coup attempt and the contested presidential referendum in May 2017, a total of 158 media
outlets were shut down, and 150 journalists were put behind the bars in the purges under the extended
state of emergency, according to an OSCE report (2017).
8 After the first acquisition in 2011, then-Prime Minister Recep Tayyip Erdogan casually explained that
the new owner asked for his recommendation for whom to hire as Milliyet’s editor-in-chief soon after
the buying the newspaper – confirming the extralegal relationship between them. For more information,
see Ceren Sozeri, “Hukumeti Destekleyene Butun Kapilar Aciliyor,” P24 (2015) and Bianet, “Dogan
Medya’da Cikarilanlar, Istifalar, ve Atananlar,” Bianet (2018).
9 The country’s election board (YSK) issued a decision deeming even the ballots without official seals valid
and therefore cast serious doubts on the integrity of the results. For more details, see Jennifer Amur, “Why
Turkish Opposition Parties are Contesting the Referendum Results,” The Washington Post, April 17, 2017,
www.washingtonpost.com/news/worldviews/wp/2017/04/16/heres-why-turkish-opposition-parties-
are-contesting-the-referendum-results/
References
Albayrak, A. and Parkinson, J. (2013, September 16) “Turkey’s government forms 6,000-member social
media team.” The Wall Street Journal. Retrieved from www.wsj.com/articles/SB1000142412788732352
7004579079151479634742 (accessed 12 November).
Allen, D. (2005) Democracy Inc. Chicago, IL: University of Illinois Press.
Amur, J. (2017, April 17) “Why Turkish opposition parties are contesting the referendum results.” The
Washington Post. Retrieved from www.washingtonpost.com/news/worldviews/wp/2017/04/16/
heres-why-turkish-opposition-parties-are-contesting-the-referendum-results/
Anadolu News Agency. (2016, December 1) “Feto’nun ‘trolleri’ siber polisten kacamadi.” Anadolu News
Agency. Retrieved from http://aa.com.tr/tr/15-temmuz-darbe-girisimi/fetonun-trolleri-siber-polisten-
kacamadi/696867
Andrejevic, M. (2007) “Surveillance in the digital enclosure.” The Communication Review, 10, 295–317.
Andrejevic, M. (2009) “Critical media studies 2.0: An interactive upgrade.” Studies in Communication and
Culture, 1, 35–51.
Andrejevic, M. (2013) Infoglut: How Too Much Information Is Changing the Way We Think and Know. New
York, NY: Routledge.
Bagdikian, B. H. (2004) The New Media Monopoly. Boston, MA: Beacon Press.
511
Aras Coskuntuncel
Baker, E. C. (2007) Media Concentration: Why Ownership Matters. Cambridge: Cambridge University Press.
Barney, D. (2000) Prometheus Wired: The Hope for Democracy in the Age of Network Technology. Chicago, IL:
University of Chicago Press.
Benkler,Y. (2006) The Wealth of Networks: How Social Production Transforms Markets and Freedom. New Heaven,
Benkler, Y. (2011) “A free irresponsible press: WikiLeaks and the battle over the soul of the networked
fourth estate.” Harvard Civil Rights-Civil Liberties Law Review, 46, 311–397.
Bianet. (2014, December 23) “84 Twitter Accounts Withheld During Kobane Protests,” Bianet. Retrieved
from http://bianet.org/english/politics/161033-84-twitter-accounts-withheld-during-kobane-protests
Bianet. (2018, May 21) “Dogan Medya’da Cikarilanlar, Istifalar, ve Atamalar,” Bianet. Retrieved from http://
bianet.org/bianet/siyaset/197354-dogan-medya-da-cikarilanlar-istifalar-ve-atamalar
Birgun. (2017, May 2) “Cok Sayida Sosyal Medya Hesabi ve Internet Subesine Erisim Engeli,” Birgun.
Retrieved from www.birgun.net/haber-detay/cok-sayida-sosyal-medya-hesabi-ve-internet-sitesine-
erisim-engeli-157852.html
Carey, J. (2009) “A cultural approach to communication.” In Communication as Culture: Essays on Media and
Society. New York, NY: Routledge (pp. 11–29).
Castells, M. (1996) The Rise of the Network Society. Oxford: Blackwell.
Castells, M. (2008) “The new public sphere: Global civil society, communication networks, and global gov-
ernance.” The ANNALS of the American Academy of Political and Social Science, 616(1), 78–93.
Castells, M. (2012) Networks of Outrage and Hope: Social Movements in the Internet Age. Cambridge: Polity Press.
Coskuntuncel, A. (2016) “Privatization of governance, delegated censorship, and hegemony in the digital
era.” Journalism Studies, 1–19.
Cumhuriyet. (2017, January 14) “Sosyal Medyaya Buyuk Operasyon: 17 Bin Kisi Hakkinda Fezleke
Hazirlandi, 45 Bin Kullanici Araniyor.” Cumhuriyet. Retrieved from www.cumhuriyet.com.tr/haber/
siyaset/660327/Sosyal_medyaya_buyuk_operasyon__17_bin_kisi_hakkinda_fezleke_hazirlandi__45_
bin_kullanici_araniyor.html#
Deibert R. et al. (2010) Access Controlled: The Shaping of Power, Rights, and Rule in Cyberspace. Cambridge,
MA: The MIT Press.
DeNardis, L. (2012) “Hidden levers of internet control: An
infrastructure-based theory of internet governance.” Information, Communication & Society, 1–19.
DeNardis, L. (2014) The Global War for Internet Governance. New Haven, CT: Yale University Press.
DeNardis, L. (2015, April–June) “The internet design tension between surveillance and security.” Annals of
the History of Computing, 72–83.
Foster, J. B. (2015) “Marxism and ecology.” Monthly Review, 67(7). Retrieved from http://monthlyreview.
org/2015/12/01/marxism-and-ecology/ (accessed 10 May 2016).
Fox-Brewster, T. (2016, October 25) Is an American Company’s Technology Helping Turkey Spy on its Citizens?
Retrieved from www.forbes.com/sites/thomasbrewster/2016/10/25/procera-francisco-partners-turkey-
surveillance-erdogan/#71c230644345
Fuchs, C. (2014) Social Media: A Critical Introduction. London: Sage Publications.
Gitlin, T. (2003) The Whole World Is Watching: Mass Media in the Making and Unmaking of the New Left. Berke-
ley, CA: University of California Press.
Glass, Z. (2007) “The effectiveness of product placement in video games.” Journal of Interactive Advertising,
8(1), 23–32.
Goldsmith, J. and Wu, T. (2006) Who Controls the Internet? Illusions of a Borderless World. Oxford: Oxford
University Press.
Greider, W. (1992) Who Will Tell the People: The Betrayal of American Democracy. New York, NY: Simon &
Schuster.
Grossman, L. (2006, December 25) “You – yes, you – are time’s person of the year.” TIME. Retrieved from
http://content.time.com/time/magazine/article/0,9171,1570810,00.html
Harvey, D. (2003) “The fetish of technology: Causes and consequences.” Macalester International, 13, 3–30.
Harvey, D. (2005) A Brief History of Neoliberalism. Oxford: Oxford University.
Herman, E. S. and Chomsky, N. (2002) Manufacturing Consent: The Political Economy of the Mass Media. New
York, NY: Pantheon Books.
Hesmondhalgh, D. (2013) The Cultural Industries. New York, NY: Sage Publications.
Hurriyet Daily News (2014a, April 14) “Turkey and Twitter Find ‘Common Ground’,” Hurriyet Daily News,
April 14, 2014. Retrieved from www.hurriyetdailynews.com/turkey-and-twitter-find-common-
ground.aspx?pageID=238&nID=65018&NewsCatID=338.
512
Hurriyet Daily News (2014b, April 20). “Twitter Now Toes the Line, Deputy PM Says after Two More
Accounts Witheld,” Hurriyet Daily News, April 20, 2014. Retrieved from www.hurriyetdailynews.com/twit
ter-now-toes-the-line-deputy-pm-says-after-two-more-accounts-withheld.aspx?pageID=238&nID=
65282&NewsCatID=339.
Huws, U. (2014) Labor in the Global Digital Economy: The Cybertariat Comes of Age. New York, NY: Monthley
Review Press.
Ilgun, S. (2016, November 7) “Siyasi Darbe Operasyonunda Son Donemecteyiz.” Evrensel. Retrieved from
www.evrensel.net/haber/295022/siyasi-darbe-operasyonunda-son-donemecteyiz
Jasanoff, S. (2004) “The idiom of co-production.” In S. Jasanof (ed.), States of Knowledge: The Co-Production
of Science and Social Order. London: Routledge.
Jenkins, H. (2006) Convergence Culture. New York, NY: New York University Press.
Kuzuoglu,S. M. (2013, June 19) “Gezi Parki Eylemlerinin Sosyal Medya Karnesi.”Radikal. Retrieved from www.
radikal.com.tr/yazarlar/m-serdar-kuzuloglu/gezi-parki-eylemlerinin-sosyal-medya-karnesi-1138146/
Lyon, D. (1998) “Situating surveillance.” In The Electronic Eye: The Rise of Surveillance Society. Minneapolis,
MN: University of Minnesota Press (pp. 3–79).
Marx, K. (1990) Capital Volume I. London: Penguin Classics.
McChesney, R. W. (2004) The Problem of the Media. New York, NY: Monthly Review Press.
McChesney, R. W. (2012) “Farewell to journalism? Time for a rethinking.” Journalism Studies, 13, 682–694.
McChesney, R. W. (2013) Digital Disconnect: How Capitalism Is Turning the Internet Against Democracy. New
York, NY: The New Press.
Nebil, F. S. (2016, November 4) “BTK, VPN servislerini engellemeye başladı; işte yasaklanan VPN
sağlayıcıları.” T24. Retrieved from http://t24.com.tr/haber/btk-vpn-servislerini-engellemeye-basladi-
iste-yasaklanan-vpn-saglayayicilari,369109
Negroponte, N. (1996) Being Digital. London: Coronet.
Noam, E. M. (2009) Media Ownership and Concentration in America. Oxford: Oxford University Press.
OSCE. (2017, April 16) “International referendum observation mission: Republic of Turkey – Constitutional
referendum: Statement of preliminary findings and conclusions.” OSCE. Retrieved from www.osce.org/
odihr/elections/turkey/311721?download=true
Reiss, S. (2006, July) “His space.” Wired. Retrieved from www.wired.com/2006/07/murdoch-2/
Reporters Without Borders and Bianet. (2016) Media Ownership Monitor: Turkey. Retrieved from http://
turkey.mom-rsf.org/en/
Scheiner, E. (2006) Democracy Without Competition in Japan: Opposition Failure in a One-Party Dominant State.
Cambridge: Cambridge University Press.
Schiller, Dan. (2000). Digital Capitalism: Networking the Global Market System. Cambridge and London: MIT.
Schiller, Dan. (2007.) How to Think about Information. Urbana and Chicago: University of Illinois Press.
Scholz, T. (ed.). (2013) Digital Labor: The Internet as Playground and Factory. New York, NY: Routledge.
Silverman, J. (2015, April 8) “Why is Twitter aiding Turkish censorship?” The Baffler. Retrieved from
http://thebaffler.com/blog/twitter-turkey (accessed 31 October 2015).
Sozeri, C. (2015) “Hukumeti Destekleyene Butun Kapilar Aciliyor.” P24. Retrieved from http://plat
form24.org/medya-izleme/813/-hukumeti-destekleyene-butun-kapilar-aciliyor (accessed October 30).
Sozeri, E. K. (2015, January 31) “Uncovering the Accounts that Trigger Turkey’s War on Twitter.” The
Daily Dot. Retrieved from www.dailydot.com/politics/twitter-transparency-report-turkey-censorship/
(accessed 13 October 2015).
Sterne, J. (2012, April 9) “What if interactivity is the new passivity?” FLOW. Retrieved from https://www.
flowjournal.org/2012/04/the-new-passivity/ (accessed 25 November 2012).
Tanash, R. S. et al. (2015) “Known unknowns: An analysis of Twitter censorship in Turkey.” Proceedings of
the 14th ACM Workshop on Privacy in the Electronic Society (pp. 11–20).
TIME. (2006, December 25) “You – yes, you – are Time’s Person of the Year,” TIME. Retrieved from
http://content.time.com/time/magazine/article/0,9171,1570810,00.html
Tsui, L. (2015) “The coming colonization of Hong Kong cyberspace: Government responses to the use of
new technologies by the Umbrella Movement.” Chinese Journal of Communication, 8, 1–9.
Tufekci, Z. (2014) “Social movements and governments in the digital age: Evaluating a complex landscape.”
Journal of International Affairs, 68(1), 1–18.
Tufekci, Z. (2016, July 18) “How the internet saved Turkey’s internet-hating president.” The New York
Times. Retrieved from www.nytimes.com/2016/07/20/opinion/how-the-internet-saved-turkeys-
internet-hating-president.html?rref=collection%2Fcolumn%2Fzeynep-tufekci&action=click&content
Collection=opinion&region=stream&module=stream_unit&version=latest&contentPlacement=8&pgt
ype=collection
513
Aras Coskuntuncel
TUIK. (2016, August 18) “Information and Communication Technology (ICT) usage survey on house-
holds and individuals, 2016.” Turkish Statistical Institute. Retrieved from www.turkstat.gov.tr/PreHaber-
Bultenleri.do?id=21779
Turow, J. (2011) The Daily You: How the New Advertising Industry Is Redefining Your Identity and Your Worth.
London: Yale University Press.
WikiLeaks. (2016a, December 5) “Berat’s box.” Wikileaks. Retrieved from https://wikileaks.org/berats-box/
WikiLeaks. (2016b, March 16) “Hillary Clinton email archive.” Wikileaks. Retrieved from https://wikileaks.
org/clinton-emails/
World Bank. (2015) “Urban population data.” World Bank. Retrieved from http://data.worldbank.org/
indicator/SP.URB.TOTL?end=2015&locations=TR&start=1960&view=chart
Yesil, B., Sozeri, E. K. and Khazraee, E. (2017) “Turkey’s internet policy after the coup attempt.” Internet
Policy Observatory. Retrieved from http://repository.upenn.edu/cgi/viewcontent.cgi?article=1021&con
text=internetpolicyobservatory
514
EPILOGUE: SITUATING
JOURNALISM IN THE DIGITAL
A plea for studying news flows, users,
and materiality
Marcel Broersma
Digital journalism seems somewhat of a pleonasm nowadays. Isn’t all journalism digital? It took
some time for news organizations to properly integrate new platforms, such as websites, mobile
apps, and social media, as well as related tools and technologies, into their production routines and
publication strategies. However, in the past decade this has changed tremendously to the extent
that the “digital” has now been firmly domesticized and normalized in the newsroom. Most
legacy media such as newspapers and television stations have shifted to working digital first in a
continuous iteration between online, social, and legacy platforms. Moreover, every step in news
production has been digitized. While reporters might still take notes with pen and notebook,
they write their subsequent stories on the computer. Typesetting and printing have been fully
digital for over a decade, as have radio and television production. Therefore, even when we read
a newspaper that is delivered on our doorsteps by a boy on a bike, all except the ‘analog’ product
itself is digital until the last push of the print button.
So, why do we still speak of ‘digital journalism’ as if there is still a fundamental divide between
the digital and journalism, and as if there is still a journalism that is not digital? The reason might
be more than mere semantics. The term first and foremost reflects the close dependence of jour-
nalism studies on its object of study. As a field that quickly emerged in the last 20 years or so,
journalism studies has been riding the wave of the structural change in news production and the
industry (cf. Peters and Broersma, 2018). It has closely followed the transition of journalism from
the analog age into the digital era. An abundance of studies has mapped and analyzed how news
organizations have adapted to digitalization, mobile communication, and the rise of social media.
Within this research there is a clear emphasis on how new technologies and digital tools have
been integrated and ‘normalized’ in professional frameworks and established routines.
The body of research on changes in the industry and the profession because of digitalization
has evolved in close connection to a discourse of crisis that underscored the fear that the ‘crisis (or
even: crises) of journalism’ might have a profound negative impact on democracy. This doomsday
scenario provided a solid argument for why journalism is important and why we should study it.
Ironically, the rise of journalism studies has thus been legitimized by the (perceived) downfall of
its object of study. The shift to digital and all that came with it, ranging from decreasing reader-
ship to convergence, participatory culture, tabloidization, fake news, and so on (the list could be
extensive), conveniently helped to make an argument concerning why journalism studies should
515
Marcel Broersma
be a field of its own. Likewise, it provided a jurisdiction for scholars to study journalism and
provide public commentary about the ‘state of the news media’ (Peters and Broersma, 2018).
As a result, an interdisciplinary field has emerged that has created its own niche on the edges
of broader fields like communication studies and cultural studies – focusing on this sole object of
study: journalism. Moreover, the emphasis on how journalism as an (important or even essential)
societal institution struggled to keep afloat has resulted in a rather narrow focus on studying
news production and news texts. No wonder Ahva and Steensen (2017) concluded in their
meta-analysis of digital journalism research that sociology is the key discipline within this field.
It underpins studies on the adaptation of new technologies in the newsroom and on the develop-
ment of professional role perceptions and routines. A flood of research has indeed analyzed how
the news industry and the profession have ‘adapted to’, ‘integrated’, or ‘normalized’ the digital.
Scholarship has claimed to ‘redefine’, ‘reconceptualize’, ‘remake’, or ‘rethink’ journalism in light
of digitalization to thus make an institution dealing with stormy weather more resilient.
I would not disagree with the statement that journalism is important in and for society, and
that we should study it – who would expect differently from a journalism scholar? However, I do
argue that our almost exclusive focus on journalism and, on top of that, the increasing subdivi-
sion of the field has its downsides. Allow me to illustrate that by linking up to the metaphors of
disease and decline that are omnipresent in the ‘crisis of journalism’ discourse. When you are that
unlucky to end up in hospital, you might be tossed around by various medical specialists who
are only looking at a certain part of your body and (presumed) illness. While they are all highly
competent in what they do and highly knowledgeable of the parts of your body they are dealing
with, an overall diagnosis – not to speak of a treatment and a cure – might be obscured by their
acute focus on the individual parts of your body. This might be reinforced by the fact that their
focus on the physical makes them neglect contextual factors that might also have an impact on
your health or are even the cause of your disease.
What does this tell us about journalism and journalism studies? Well, instead of ever narrow-
ing down our object of study into even more specialized subfields, we should broaden it up. It
is good to pause every now and then in journalism scholarship to realize that journalism is cur-
rently part of a broader digital ecology for news and information. This pertains to professional
news production as a cultural system that is closely intertwined with other fields in society. But
even more it concerns the news it produces and disseminates in search of an audience. News
is now subject to the principles of numerical representation, modularity, automation, variability,
and transcoding (Manovich, 2001), which makes it, in principal, interchangeable with all other
cultural forms. On the day-to-day level, it is hard to distinguish from entertainment, satire, or
fake news but also from other informational genres, advertisement, and public relations. We can
therefore only understand the changes in our object of study in the context of these broader
interdependencies. This implies we should open up the field beyond a focus on journalism and
beyond news production and news texts, whether these take place in legacy news organizations
or new digital start-ups, within or outside the newsroom.
Paradoxically, a way forward for journalism studies might not be to narrow it down to jour-
nalism but to extend it to study what feels like news and journalism to users, how they develop
tactics to navigate the digital news and information ecology, how they negotiate the information
that a wide range of agents has on offer, and how journalism fits in to the platformization of the
online news and information ecology. In other words, strange as it may sound, it might be highly
beneficial for journalism studies to focus less on journalism. I therefore suggest to move away
from how the digital has been integrated in journalism in terms of technologies, platforms, and
businesses and focus instead on how journalism has been integrated into the digital.
Here, I will first discuss the rise of a decentralized informational ecology in which journalism
is only one cultural form among many others. In this context, it makes increasingly less sense
516
Epilogue
to study journalism as an isolated object. I will then sketch three research venues that might be
helpful to situate journalism firmly in the digital but also to reconnect it to the social structures
it simultaneously thrives on and produces. First, I will turn to circulation as a concept that helps
us to make sense of this new embedding for news. I move on to suggest that a focus on news use
and the practices of those who engage with news and journalism offer novel perspectives on the
role and function journalism does (or does not) fulfill in society. Finally, I argue that parallel to
the ‘turn to digital’ in journalism, or even as a result of this, the medium and the material deserve
more attention in journalism studies. While it might sound paradoxical, this becomes increas-
ingly important now that all journalism is digital. To conclude, I suggest to shift our gaze to news
as a cultural form and how this is renegotiated by users, journalists, and other institutional agents
alike, instead of focusing research on journalism that produces journalism.
A decentralized informational ecology

The digital turn confronted the news industry with a Catch-22 situation. If publishers were sane,
they would embrace the new ‘renaissance’ and migrate their business to the digital realm. News
could now be brought to the audience “any time, any place, any how”, as a 2002 report of the
World Association of Newspapers happily proclaimed (WAN, 2002). However, publishers would
be insane to do so. Profits were still made in the analog world, while online quickly turned out to
be a sinkhole in which not just profits but complete companies could disappear. That being said,
wouldn’t it be sane to go where the audience moved, and wouldn’t it be insane to miss that boat?
So, “sure there’s a catch”, Doc Daneek would say (Heller, 1961), and this catch has not only been
perplexing for the news industry but is also central to much scholarship on digital journalism (see
Chyi and Tenenboim, this volume, Chapter 12).
In the era of high-modernity (Hallin, 1992), news organizations enjoyed a monopoly in the
distribution channels for information about current events. Making news was to a large extent
driven by an industrial logic focusing on economies of scale and standardized production. As
Chris Peters and I argued elsewhere: “Just as the creation of the mass-market automobile was
the result of industrial logic brought to transportation, the rise of the mass press was the result
of industrial logic brought to information” (Broersma and Peters, 2013: 4). Journalism reached
out to a mass audience and rooted its norms in a professional philosophy of depersonalization,
detachment, and quality control. It could thus even be considered “the primary sense-making
practice of modernity” (Hartley, 1996: 12), with all the democratic functions and power that were
ascribed to it as a central institution in society. In this linear thinking about journalism, news, and
communication, a focus in academic research on how journalism was produced, its political and
economic confinements, and the meaning-making processes rooted in news texts was more than
sensible. And studying journalism in isolation by focusing on how news organizations produce
texts remains valuable. After all, journalism is still an important institution in society and a vast
producer of knowledge. However, approaches that think across agents and cultural forms would
be worthwhile to explore.
The online informational ecology is first and foremost reigned by a deindustrialized logic
that works asynchronously. We need to acknowledge that the homepages of websites or the
gated communities of news apps are not the dominant gateways for accessing news. Online, and
increasingly on mobile phones, users often do not purposively visit news organizations. Search
engines and social platforms, or the mix of them, ensure the majority of traffic (see Ørmen, this
volume, Chapter 10). Although news organizations still have a strong presence on online, mobile,
and social platforms and are important “hubs” in the online news ecology, their publications get
unbundled and are embedded within a diverse stream of information. Nowadays journalism is
only one of many public information providers in a broader informational ecology. Of course,
517
Marcel Broersma
we should not underestimate how many people still read a newspaper, watch the television news,
or listen to the radio – and conversely, how many are not online (see Brennen, this volume,
Chapter 31). But survey research indicates a consistent shift to online media as the main source
for news (Newman et al., 2017). This means that we face a de-ritualization of news consump-
tion. Whereas in the mass media era fixed moments during the day were reserved for reading the
newspaper or watching television news, news use nowadays is increasingly folded into everyday
routines and scattered throughout the day. While new rituals might emerge, the high-modern
experience of simultaneously consuming the same information, even on a national scale, and
being exposed to shared frames of reference might become obsolete or at least incidental (Bro-
ersma and Peters, 2013).
The fact that the online informational ecology is decentralized and networked suggests that
we should study journalism relationally (cf. Singer, this volume, Chapter 37). Given the abun-
dance of (interconnected) information that citizens need to navigate, we should analyze differ-
ent media objects and outlets in relation to one another, as well as journalists to other agents
in the network. Online, it is often hard to distinguish between different cultural forms because
genre conventions get fuzzy. Something can look like news but might instead be satire, PR, or
propaganda – or the other way around. Elsewhere I have argued that, because of this, journalism
suffers from a refractured paradigm, a symptom of institutional osteoporosis (Broersma, 2013).
Leaving this somewhat provocative metaphor aside, citizens have definitely become more skepti-
cal and, to a certain extent, more media literate. They have an abundance of information at their
disposal to critically compare and assess. For every fact and argument they encounter, multiple
‘alternative facts’ and counterarguments can be found. Now that journalism does not control
the gates anymore, “every mouse click reassures news users that news does not convey the truth
but a truth” (Broersma, 2013: 43). This growing awareness erodes trust and journalism’s power as
an important institution in society, not just by those who subscribe to the ‘FAKE NEWS’ media
discourse but also by those who feel a moral responsibility to carefully make sense of contradict-
ing information online.
In addition, functions news organizations traditionally fulfilled in people’s everyday lives are
now also performed, or even taken over, by new players who are empowered by the networked
logic of social media. Wikipedia has become an important source for background information.
Opinion formation takes place at a range of, usually rather partisan, forums and websites. Adver-
tisements have moved to a vast number of specialized websites for housing, cars, secondhand
goods, and so on. In other words, journalism is not capable of controlling the full value chain
anymore. Both manifest and latent functions it used to fulfill are redistributed to (better) fit into
the lives of people (again). A functional approach that critically scrutinizes what journalism does
for people might offer a new lens to study how journalism’s societal role is renegotiated (Broersma
and Peters, 2016).
Whereas journalism in the ‘long twentieth century’ had a monopoly over the selection and
distribution of news and other current information, the power to control news flows is now
redistributed among news organizations, social platforms, algorithms, and users according to a
networked logic (Van Dijck and Poell, 2013; Klinger and Svensson, 2015; Broersma and Graham,
2016). Analyses that focus on the supply side or the content of news texts, therefore, have only
limited explanatory power if we want to understand why certain news items get read or viewed
more than others within certain communities (or subnetworks), why they generate traffic to
news sites, or why they even go viral.
News is now being pushed through networks by means of the aggregated practices through
which users on platforms engage with content. This results in different assemblages of posts and
messages in users’ timelines. It gives “them each an experience of the site that is individualized
yet overlapping with others” (Baym, 2010: 90). Moreover, in these timelines news is embedded
518
Epilogue
in broader information repertoires. It is mixed with personal updates, commercial messages, and
other information genres, and sometimes hard to recognize. Incidental news consumption, when
users encounter news while looking for other information or when doing something else, seems
to become increasingly prominent (Bergström and Belfrage, 2018; Boczkowski et al., 2018).
News becomes a by-product that pops up in the midst of other activities.
This implies that a range of assumptions that underlie our thinking about the importance of
journalism (and journalism studies) have suddenly become unsettled. Normative theories about
journalism’s role and function in democracy presuppose that news organizations reach a some-
how clearly defined public and that citizens read, watch, or listen to their carefully produced and
selected news content. These theories argue that people (should) use news to engage socially
and participate in democracy and that this legitimizes journalism’s role in society. But to what
extent do people live up to that expectation (if they have ever done so)? How do they actually
distinguish between journalism and all other cultural forms out there that compete for attention?
The circulation of news

The eternal questions of ‘Who is a journalist?’ and ‘What is journalism?’ have gained renewed
importance in scholarship concerning digital journalism (Eldridge and Franklin, 2017; Maalik
and Shapiro, 2017). Both from a professional and a legal point of view, new boundaries and defi-
nitions are developed to provide an answer. It is insightful to analyze these discursive struggles
and see what arguments legislators, politicians, and journalists bring to the table (see Johnston
and Wallace, this volume, Chapter 1). And it is equally important to see how journalists’ percep-
tions of their professional role(s) have changed due to digitization, or not so much (see Hanusch
and Banjac, Vos and Ferrucci, this volume, Chapters 2 and 3, respectively). However, throughout
history journalism has never been a stable object. What journalism is and should be to certain
groups and agents is being renegotiated constantly – like any process of identity formation.
Open-endedness is precisely the essence of this discursive struggle for power.
Several chapters in this volume point out that there should be more attention in journalism
studies for “new journalistic actors” or what Eldridge (2018) has called “interlopers”. These can
be new entrepreneurial start-ups, although these are mostly firmly rooted in ‘traditional’ journal-
ism (see Witschge and Harbers, this volume, Chapter 5). They are often started by professional
journalists who either feel that journalism needs to change, or needs to return to “what it used
to be”, or have an economic incentive. Others focus on new agents outside the journalistic field
who perform something that resembles journalism. This ranges from WikiLeaks and bloggers
(Eldridge, 2018) to civic tech (Baack, 2018; Usher, this volume, Chapter 26) and nonhuman
agents such as algorithmic bots (Larsson and Moe, 2015). Another strand in scholarship deals
with “the people formerly known as sources” such as politicians, NGOs, and companies who
provide news organizations with information subsidies but also increasingly reach out to the
public themselves, thus circumventing journalism and shifting the power (Van Leuven and Joye,
2014; Broersma and Graham, 2016).
However, a production perspective, let alone one that looks through the lens of what makes
someone a journalist and something journalistic, might only reveal one part of what is currently
evolving online. This might also lead to somewhat circular reasoning in which we set off to intel-
lectually integrate new players in the familiar frameworks of journalism. After all, ‘journalistic’
or ‘semi-journalistic’ actors are not the only agents who have an impact on the circulation of
news and information. Online we see fluid configurations of individuals and institutions who
gather around news events and public discussions. The majority of them passively engage with
information by just “consuming” it, but others might commit “random acts of journalism”, as
Lasica (2003: 71) aptly phrased it.
519
Marcel Broersma
Although early views about citizens getting involved in “the process of collecting, reporting,
sorting, analyzing and disseminating news and information” (Lasica, 2003: 71) en masse turned
out to be overly optimistic, we can still signal an important shift of agency from news organiza-
tions toward users. In the online information ecology, engagement is key. This takes place with
different levels of intensity, ranging from ‘passive’ reading or watching to ‘semi-active’ liking and
sharing and ‘active’ posting and commenting. All of these practices, however, contribute to the
circulation of news. The information citizens get online is dependent on their individualized
network and the people they are connected to. Individuals’ networks might overlap, but they are,
as Baym (2010: 90) has argued, “ego-centric and no two will be identical”. While this does not
necessarily diminish the importance of social groups such as family and friends or make people
more individualistic and can even enhance social relations (Miller et al., 2016), it does result in
very different patterns of news flows.
If we argue that providing news is an important function journalism fulfills in society, and maybe
even the only one we can realistically expect, as Nielsen (2017) states, it would be worthwhile
to analyze more closely how and why information circulates in the digital realm (cf. Günther et
all, this volume, Chapter 9). Anderson already noted in 2010 that research “has focused on news
production within institutions rather than the circulation of news in ecosystems” (Anderson, 2010:
291). Building upon the work of Stuart Hall, Henrik Bødker (2015) made a similar plea for research
on “cultures of circulation” as temporal and spatial structures in which meaning gets produced.
In a digital landscape, various versions of a constantly updated text are spread and stored
in a variety of settings. This is produced through circulation in a news landscape in
which established cultures of journalism increasingly are drawn into close contact with
other, complimentary and/or challenging, domains of circulation.
(Bødker, 2015: 110)
However, whereas Bødker, reasoning from the perspective of producers, points at how different
cultures of circulation intersect in the digital realm, I would argue that we observe a specific
form of context collapse here. For users who encounter news in the online information ecology,
it is often not clear in which context or culture of circulation a digital object is produced. It has
simply become part of a smorgasbord of information that is laid out for them. They need to
decide which news items to engage with, deduce what the source is, and assess how to make sense
of divergent and conflicting information. Ultimately, they are the ones who, actively or (semi-)
passively, determine the scale of circulation of news items and information.
A shift to news use

Now news users have gained more prominence in the circulation processes of news, it makes
more sense to integrate journalism into broader digital configurations than to study the digital
in journalism. As I have already outlined, we should therefore move slightly away from concep-
tualizing journalism from the perspective of those institutions and agents that produce it. How
they perceive their professional roles (even against the backdrop of what ‘audiences’ expect from
them) and how they have integrated new technologies and platforms in professional practice
might simply be less relevant in the online information ecology than it has been in the analog
age, given the weight other agents have.
It might be insightful to look at what users perceive as news and journalism instead, not
just based on their attitudes as measured in surveys and interviews but also on how they
engage with news through online practices. This is especially challenging for journalism
520
Epilogue
studies (and might actually be an important reason for its focus on production and content)
because, as Bird (2011: 490) argues, “when one moves away from definitions of news that are
producer oriented, and begins with the consumer, the very understanding of what constitutes
news begins to blur, thus making it harder to conceptualize the relationship between news
and audience”.
In their meta-analysis of research into digital journalism, Ahva and Steensen (2017: 30) con-
cluded that what distinguishes digital journalism studies is “the position given to audiences or
users in research”. Indeed, there has been more attention to the audience in journalism studies
recently. Predominantly the decline of newspaper circulation, the stronghold of quality journal-
ism, made journalists and scholars alike realize that they could not take the audience for granted
anymore, like they had done in the era of the mass press. More recently, journalism studies schol-
ars therefore pleaded for rediscovering the audience (Loosen and Schmidt, 2012), revitalizing
audience research (Zeller, 2015), and taking a user-perspective (Picone, this volume, Chapter
11; Picone et al., 2015; Costera Meijer and Groot Kormelink, 2016). The audience thus became
problematized against the backdrop of the “crisis of journalism” and integrated in frameworks
that center on journalism.
As Ahva and Steensen (2017) argue, research into audiences has mainly been conducted from
a theoretical perspective, in what they call “user-oriented cultural analysis”. This means first that
the audience is rather talked about or researched on an aggregated level, as an abstract and quanti-
fied whole. The longitudinal survey research that industry-related institutions like Pew and The
Reuters Institute have set up have, for example, provided valuable insights in shifting consump-
tion patterns. However, surveys do not provide much insight into the practices and experiences
of users. Moreover, respondents tend not to report their actual but their perceived news use. They
tend to overreport their news use and emphasize what they think they should do to be a good
citizen (Prior, 2009).
Second, I would argue that news consumption is merely analyzed from either a commercial
(as ‘audience’) or democratic stance (as ‘publics’ or ‘the public’). It is discussed mostly vis-à-vis
how journalism adapts to digitization to cater for perceived user preferences (Anderson, 2011) or
from a normative perspective: how people need to behave to function as well-informed citizens
(cf. Ryfe, 2017). From a cultural studies tradition, emphasis is on how audiences interpret and
‘decode’ texts and how this challenges or reproduces dominant ideologies. Here effects are rather
theorized and assumed than empirically studied. While these traditions have explanatory value in
a mass media context and helped to build theories about news production and texts, they might
be less useful when we want to study how ‘networked individuals’ navigate news.
More recently, scholars have therefore shifted their gaze toward studying ‘news use’ and ‘users’
to understand how users actually deal with news in an everyday context. They analyze novel
use practices, experiences, and consumption patterns, either from a qualitative (Schrøder and
Kobbernagel, 2010; Swart et al., 2016; Costera Meijer and Groot Kormelink, 2017) or a digitally
enhanced quantitative perspective (Bruns, 2017). Others have started to explore news use from
a cross-media perspective to explain how users move across media and platforms (Schrøder and
Larsen, 2010; Bjur et al., 2014; Jensen et al., 2016; Lomborg and Mortensen, 2017). A media
repertoire approach (Hasebrink and Popp, 2006; Hasebrink and Domeyer, 2012) is useful to ana-
lyze “how people combine different media technologies, brands, genres, and products to struc-
ture their everyday life and fulfill their needs for information, entertainment, opinion formation,
sociability, and engagement” (Swart et al., 2016: 1345).
This has resulted in a still small but promising body of work that tries to capture how citizens
deal with current transformations in the online information ecology. Moving our scope beyond
what scholars predefine as journalists and journalism helps to situate journalism in the digital.
521
Marcel Broersma
Medium and material culture

It might again sound paradoxical, but with the rise of the digital, material culture has become
increasingly important. Traditionally, in journalism studies the medium was almost absent. In key
texts in the field, such as Thomas Hanitzsch and Karin Wahl-Jorgensen’s Handbook of Journalism
(2009) or Barbie Zelizer’s Taking Journalism Seriously (2004), references to any medium, let alone
substantial theoretical discussions of it, can be counted on the fingers of one hand. Just like the
public/audience, the medium has long been taken for granted as a conduit for transmitting news
content. The main assumption here was that the message could be analyzed independently from
the data carrier. Work on news content, whether it’s rooted in content, discourse, or cultural
analysis, tends to focus on the signs without paying any attention to the medium that carries
them (cf. Meyrowitz, 1998). Similarly, research on news production had little attention for how
the character of the medium impacted how news was gathered, selected, and crafted.
When the news industry became digitized, scholars started to pay attention to the integra-
tion or ‘normalization’ of new technologies in the newsroom. Some even speak of a material
turn. As Pablo Boczkowski (2015: 66) noted, there has been a “proliferation of studies looking
at various aspects of how materiality matters in newswork, and an incorporation of a number of
conceptual approaches to do this, dominated by STS and in particular actor-network theory”.
However, the focus was mostly on how journalists perceived these new ‘tools’ and how they tried
to align them with traditional professional standards (Weiss and Domingo, 2010; Anderson and
De Maeyer, 2015). How were they appropriated and incorporated in production routines and
normative frameworks? Research paid less attention to how journalists exactly interacted with
the distinctive grammar of software, interfaces, and new media technologies such as social media,
websites, RSS feeds, or news apps. Conversely, it would be worthwhile to explore more broadly
and in depth how the materiality of new platforms, tools, and technologies impacted modes of
communication. Formal and stylistic features of news coverage and their impact on the content
of news so far have gotten only little consideration.
This would be a worthwhile venue for journalism scholarship, though, because it can shed
light on how “objects of journalism” (Anderson and De Maeyer, 2015) circulate in the informa-
tion ecology and why certain pieces get picked up by publics more broadly than others and thus
get more prominence in the public sphere. This has not only to do with their content but also
with how they are embedded in the techno-material infrastructure of and across platforms. On
YouTube, for example, strategic tagging of content so it aligns with the material design of the
platform is extremely important to render videos visible. This has an impact beyond YouTube
itself because it is not just a stand-alone platform but also a main source for video that is embed-
ded on news websites and other social media platforms (Smit et al., 2015). This example also
makes clear that news production has to a large extent become dependent on external platforms,
technologies, and agents that have their own political and economic interests. Now that journal-
ism no longer succeeds in controlling the value chain, as it did in the high-modern era, news is
to a greater or lesser extent coproduced through and by the techno-material infrastructure of the
broader informational ecology.
Partly as a result of actor-network theory’s valuable insights about the agency of nonhu-
man actors, there is increasing attention in scholarship on the techno-material characteristics of
platforms and how their design structures use practices. This offers opportunities for journalism
studies to relate to. ‘Affordances’ has become a key concept in research in media and internet
studies and research on social media in particular. It refers to the properties of ‘things’ and how
people interact with them and thus emphasizes how “technologies are socially constructed and
situated on the one hand, and materially constraining and enabling on the other hand” (Bucher
and Helmond, 2017). This has resulted in, broadly, two strands of research. The first focuses on
522
Epilogue
the use of platforms and technologies. It studies new forms of communication that emerge in a
decentralized media ecology and result in distinct patterns of social interaction between users. A
second strand focuses on the political economy of platformization and datafication. Scholars here
critically study how big tech companies have (tried to) monopolized channels that satisfy certain
informational and social needs, just like news organizations managed to do in the twentieth cen-
tury. In this emerging field, users are said to be ‘trained’ by algorithms, subdued by technology,
and reduced to mere producers of commercially valuable data (cf. Couldry and Van Dijck, 2015).
Both research strands mainly focus on what technology does to people and tend to downplay
the agency and power of users in their interaction with technology. This is understandable because
they are rooted in the material characteristics of platforms as well as the political and economic
considerations that are woven into them. However, given the relatively ‘open’ nature of journalism
– a field that is existentially dependent on (the knowledge of) others – it would be valuable to look
relationally at how the power to produce and disseminate news, and thus provide legitimate repre-
sentations of social reality, is redistributed among multiple agents. Until now it still largely remains
an open question what the interaction between news organizations, ‘new journalistic actors’ and
users with social platforms, algorithms, and other techno-material objects results in.
Moreover, the material turn points at the importance of an approach that situates news in the
everyday lives of people. If we accept the premises that journalism would be a rather meaningless
practice without a public, setting aside that it might be economically unfeasible and that users
nowadays are essential to reach an audience in a decentralized information ecology, it makes sense
to ask what they actually perceive as news. In a media-saturated world, news consumption is
often habitual and interwoven with other media practices. The material and tactile characteristics
of the devices people use to access news shape their experiences. Moreover, news use increasingly
takes places in the interstices of daily life and becomes incidental (Dimmick et al., 2010; Bocz-
kowski et al., 2018; Bergström and Belfrage, 2018). It should make us cautious to pay more atten-
tion to the functions news and journalism fulfill in people’s everyday life. In-depth examination
of how news is contextualized, localized, and situated within familiar, taken-for-granted contexts
is crucial for understanding journalism’s functions and social significance in this time of flux.
To conclude: news as cultural form

In this epilogue, I have offered a threefold argument concerning why we should broaden up jour-
nalism studies instead of narrowing it down by focusing on this sole object of study. With the
establishment and burgeoning of the field – a valuable endeavor given the importance of journal-
ism as a societal institution and the various “crises” that threaten its existence – it has also become
increasingly fragmented. Although this has resulted in a rich and intellectually provocative body of
research on almost any manifestation of (digital) journalism one can think of, it also runs the risk
of disconnecting journalism from the social and techno-material structures it is not only embedded
in but also emanates from. A focus on how the digital impacted news production and texts – and
essentially is integrated in ‘journalism-as-we-know-it’ – might therefore obscure the more struc-
tural transformations that are currently taking place in the information ecology. It would therefore
be worthwhile to take an opposite perspective and situate journalism in the digital realm instead.
I sketched three possible trails that might be useful to explore: first, studying how news actu-
ally circulates in the information ecology and what determines these dynamics; second, how
users engage (or not) with objects of journalism and what they consider news; and third, how
the material affordances of platforms and technologies impact how the power to produce news
is redistributed and how people integrate news in their everyday life. This implies that we should
shift our attention from studying journalism (and trying to incorporate new phenomena in our
523
Marcel Broersma
preexisting intellectual frameworks) to studying news as a cultural form. While the core aim of
journalism’s modern ‘professional project’ was to secure a fixed and stable position for itself in
society, and scholarship in journalism studies – critically or uncritically – adopted these profes-
sional frameworks as point of departure for studying journalism, I suggest that research would
profit from embracing the challenges of opening up. The key question for journalism stud-
ies might be how news as a cultural form that appeals to the existential human need of being
informed about current events gets reshaped in the digital realm.
References
Ahva, L. and Steensen, S. (2017) “Deconstructing digital journalism studies.” In B. Franklin and S. A. Eldridge
(eds.), The Routledge Companion to Digital Journalism Studies. Abingon: Routledge (pp. 25–34).
Anderson, C. W. (2010) “Journalistic networks and the diffusion of local news: The brief, happy news life
of the ‘Francisville Four’.” Political Communication, 27(3), 289–309.
Anderson, C. W. (2011) “Deliberative, agonistic, and algorithmic audiences: Journalism’s vision of its public
in an age of audience transparency.” International Journal of Communication, 5, 529–547.
Anderson, C. W. and De Maeyer, J.(2015) “Objects of journalism and the news.” Journalism, 16(1), 3–9.
Baack, S. (2018) “Practically engaged. The entanglements between data journalism and civic tech.” Digital
Journalism [online first]. doi:10.1080/21670811.2017.1375382.
Baym, N. (2010) Personal Connections in the Digital Age. Cambridge: Polity.
Bergström, A. and Belfrage, M. J. (2018) “News in social media. Incidental consumption and the role of
opinion leaders.” Digital Journalism [online first]. doi:10.1080/21670811.2018.1423625
Bird, E. (2011) “Seeking the audience for news: Response, news talk, and everyday practices.” In V. Night-
ingale (ed.), The Handbook of Media Audiences. Oxford: Wiley-Blackwell (pp. 489–508).
Bjur, J. et al. (2014) “Cross-media use: Unfolding complexities in contemporary audiencehood.” In N.
Carpentier, K. C. Schrøder and L. Hallet (eds.), Audience Transformations: Shifting Audience Positions in Late
Modernity. London: Routledge (pp. 15–29).
Boczkowski, P. (2015) “The material turn in the study of journalism: Some hopeful and cautionary remarks
from an early explorer.” Journalism, 16(1), 65–68.
Boczkowski, P., Mitchelstein, E. and Matassi, M. (2018) “News comes across when I’m in a moment of
leisure: Understanding the practices of incidental news consumption on social media.” Journalism [online
first]. doi:10.1177/1461444817750396
Bødker, H. (2015) “Journalism as cultures of circulation.” Digital Journalism, 3(1), 101–115.
Broersma, M. (2013) “A refractured paradigm. Journalism, hoaxes and the challenge of trust.” In M. Bro-
ersma and C. Peters (eds.), Rethinking Journalism: Trust and Participation in a Transformed News Landscape.
Abingdon: Routledge (pp. 28–44).
Broersma, M. and Graham, T. (2016) “Tipping the balance of power: Social media and the transformation
of political journalism.” In A. Bruns et al. (eds.), The Routledge Companion to Social Media and Politics.
Abingdon: Routledge (pp. 89–103).
Broersma, M. and Peters, C. (2013) “Rethinking journalism. The structural transformation of a public
good.” In M. Broersma and C. Peters (eds.), Rethinking Journalism: Trust and Participation in a Transformed
Media Landscape. Abingdon: Routledge (pp. 1–12).
Broersma, M. and Peters, C. (2016) “Towards a functional perspective on journalism’s role and relevance.”
In C. Peters and M. Broersma (eds.), Rethinking Journalism Again: Societal Role and Public Relevance in a
Digital Age. Abingdon: Routledge (pp. 1–17).
Bruns, A. (2017) “Making audience engagement visible: Publics for journalism on social media platforms.”
In B. Franklin and S. A. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingon,
Oxon: Routledge (pp. 325–334).
Bucher, T. and Helmond, A. (2017) “The affordances of social media platforms.” In J. Burgess, T. Poell and
A. Marwick (eds.), The SAGE Handbook of Social Media. London and New York, NY: Sage Publications
(pp. 233–253).
Costera Meijer, I. and Groot Kormelink, T. (2016) “Revisiting the audience turn in journalism.” In B.
Franklin and S. Eldridge II (eds.), The Routledge Companion to Digital Journalism Studies. Abingdon: Rout-
ledge (pp. 345–353).
524
Epilogue
Costera Meijer, I. and Groot Kormelink, T. (2017) “What clicks actually mean. Exploring digital news user
experiences.” Journalism [online first]. doi:10.1177/1464884916688290.
Couldry, N. and Van Dijck, J. (2015) “Researching social media as if the social mattered.” Social Media +
Society, 1(2), 1–7.
Dimmick, J., Feaster, J. and Hoplamazian, G. (2010) “News in the interstices. The niches of mobile media
in space and time.” New Media & Society, 13(1), 23–39.
Eldridge II, S. A. (2018) Online Journalism From the Periphery: Interloper Media and the Journalistic Field.
Abingdon: Routledge.
Eldridge II, S. A. and Franklin, B. (2017) “Introduction: Defining digital journalism studies.” In B. Franklin
and S. A. Eldridge (eds.), The Routledge Companion to Digital Journalism Studies. Abingon: Routledge (pp.
1–12).
Hallin, D. (1992) “The passing of the ‘high modernism’ of American journalism.” Journal of Communication,
42(3), 14–25.
Hartley, J. (1996) Popular Reality: Journalism, Modernity and Popular Culture. London: Arnold.
Hasebrink, U. and Domeyer, H. (2012) “Media repertoires as patterns of behaviour and as meaningful
practices. A multimethod approach to media use in converging media environments.” Participations, 9(2),
757–779.
Hasebrink, U. and Popp, J. (2006) “Media repertoires as a result of selective media use. A conceptual
approach to the analysis of patterns of exposure.” Communications, 31(3), 369–387.
Heller, J. (1961) Catch-22. New York, NY: Simon & Schuster.
Jensen, J. L., Mortensen, M. and Ørmen, J. (eds.). (2016) News Across Media: Production, Distribution and
Consumption. New York, NY: Routledge.
Klinger, U. and Svensson, J. (2015) “The emergence of network media logic in political communication. A
theoretical approach.” New Media & Society, 17(8), 1241–1257.
Larsson, A. O. and Moe, H. (2015) “Bots or journalists? News sharing on Twitter.” Communications, 40(3),
361–370.
Lasica, J. D. (2003, Fall) “Blogs and journalism need each other.” Nieman Reports, 70–74. Retrieved from
www.nieman.harvard.edu/reports/03-3NRfall/V57N3.pdf
Lomborg, S. and Mortensen, M. (2017) “Users across media. An introduction.” Convergence, 23(4), 343–351.
Loosen, W. and Schmidt, J. (2012) “(Re-)discovering the audience: The relationship between journalism
and audience in networked digital media.” Information, Communication & Society, 15(6), 867–887.
Maalik, A. and Shapiro, I. (2017) “What’s digital? What’s journalism.” In B. Franklin and S. A. Eldridge
(eds.), The Routledge Companion to Digital Journalism Studies. Abingon: Routledge (pp. 15–24).
Manovich, L. (2001) The Language of New Media. Cambridge, MA and London: The MIT Press.
Meyrowitz, J. (1998) “Multiple media literacies.” Journal of Communication, 48(1), 96–108.
Miller, D. et al. (2016) How the World Changed Social Media. London: UCL Press.
Newman, N. et al. (2017) Reuters Institute. Digital News Report 2017. Oxford: Reuters Institute for the Study
of Journalism.
Nielsen, R. K. (2017) “The one thing journalism just might do for democracy. Counterfactual idealism,
liberal optimism, democratic realism.” Journalism Studies, 18(10), 1251–1262.
Peters, C. and Broersma, M. (2018) “Fusion cuisine: A functional approach to interdisciplinary cooking in
journalism studies.” Journalism [online first]. doi: 10.1177/1464884918760671
Picone, I., Courtois, C. and Paulussen, S. (2015) “When news is everywhere: Understanding participation,
cross-mediality and mobility in journalism from a radical user perspective.” Journalism Practice, 9(1),
35–49.
Prior, M. (2009) “The immensely inflated news audience: Assessing bias in self-reported news exposure.”
Ryfe, D. (2017) Journalism and the Public. Cambridge: Polity.
Schrøder, K. and Kobbernagel, C. (2010) “Towards a typology of cross-media news consumption: A
qualitative-quantitative synthesis.” Northern Lights: Film & Media Studies Yearbook, 8(1), 115–137.
Schrøder, K. and Larsen, B. (2010) “The shifting cross-media news landscape. Challenges for News Produc-
ers.” Journalism Studies, 11(4), 524–534.
Smit, R., Heinrich, A. and Broersma, M. (2015) “Witnessing in the new memory ecology: Memory con-
struction of the Syrian conflict on YouTube.” New Media & Society, 19(2), 289–307.
Swart, J., Peters, C. and Broersma, M. (2016) “Navigating cross-media news use. Media repertoires and the
value of news in everyday life.” Journalism Studies [online first]. doi:10.1080/1461670X.2015.1129285
Van Dijck, J. and Poell, T. (2013) “Understanding social media logic.” Media and Communication, 1(1), 2–14.
525
Marcel Broersma
Van Leuven, S. and Joye, S. (2014) “Civil society organizations at the gates? A gatekeeping study of news
making efforts by NGOs and government institutions.” The International Journal of Press/Politics, 19(2),
160–180.
Wahl-Jorgensen, K. and Hanitzsch, T. (eds.). (2009) Handbook of Journalism Studies. New York, NY:
Routledge.
WAN. (2002) “The distribution revolution.” Shaping the Future of the Newspaper. Strategy Report, 2(1).
Weiss, A. and Domingo, D. (2010) “Innovation processes in online newsrooms as actor-networks and com-
munities of practice.” New Media & Society, 12(7), 1156–1171.
Zelizer, B. (2004) Taking Journalism Seriously: News and the Academy. Thousand Oaks, CA, and London: Sage
Publications.
Zeller, F., Ponte, C. and O’Neill, B. (2015) Revitalising Audience Research: Innovations in European Audience
Research. New York, NY and Milton Park: Routledge.
526
INDEX
Page numbers in italics indicate figures and in bold indicate tables on the corresponding pages.
Abernathy, D. 405 53–54; supervised learning 87–88; see also data

accountability and right to be forgotten 328, 330 journalism
Acosta, D. 298 Al Jazeera 298, 455
activist and citizen livestreams 300–301 Allen, D. 394, 396, 397, 398
actor-network theory 376, 403; environmental Alliance for Audited Media (AAM) 159
media and 406–407 Amazon 134
advertising and subscription revenue 159; Amazon Echo 143, 419; see also Alexa
alternatives to 215–216; see also native American Society of Magazine Editors 431
advertising American Society of News Editors 214
affordances of technology 73, 522 America’s Battle for Media Democracy 220
African Americans and pop-up news ecologies Ampp3d 289, 292
378–380, 381, 382; see also Black Lives Matter Anadolu Agency 501
movement Anderson, B. 338, 339
agency, socio-technical concepts of 316–317, 317 Anderson, C. W. 226, 232–235, 349, 354, 356, 376
agenda-setting theory 491–492 Andrejevic, M. 504, 505, 510
agonism 465–467 anti-Semitism 431
Ahva, L. 516, 521 AOL 158
Airbnb 175 Apostles of Certainty 234–235; see also Anderson, C. W.
Aitamurto, T. 406 Appelgren, E. 354
AJ Stream 382 Apperley, T. 285
Albayrak, B. 507 Apple 412
Alexa (smart assistant) 293 Apple Safari 100
Alexander, K. B. 361 Arab Spring 391
algorithmic curation see curation, algorithmic Armstrong, G. B. 489
algorithmic journalism 54–55, 227; see also Arnold, M. 392
automated journalism Artemas, K. 469–470
algorithmic transparency 54 artificial intelligence (AI) 57–58, 239–240, 284,
algorithms 4, 7–9; acceptance of 57; algorithmic 315, 412, 419
curation (see curation, algorithmic); authorship Asher, R. 19
and 55–57; automated journalism 54–55; assemblage systems 402–403, 408–409
computational analysis 95, 102; evolving Associated Press 53, 244
capacity of 57–60; as future of journalism Atlanta Journal-Constitution 350
55; Latent Dirichlet allocation (LDA) 85–87; AT&T 213, 218
rise of automated journalism based on attribution norm gap 58
527
Index
A.T. v Globe 24h.com 21 Bipartisan Report 382–383

audiences: as commodities 456–457; diverse Bird, E. 521
32–33; engaged through good journalism Bjur, J. 146
72–73; journalistic role expectations of 30; Black Film Review 190
measurement of newspaper 159–160; shift in Black Lives Matter movement 6, 381, 387–388;
relations between journalists and 31; see also conclusion on 398; further reading on 398–399;
users, digital news power and narratives of control and 389–390;
augmented journalism see automated journalism research method on 392; rise of 388; rise
Ausserhofer, J. 230, 231 of sousveillance and 391–392; selection of
authorship: algorithmic 57; anthropomorphic 57 participants for study on 392–393; study results
authorship, algorithmic 57 on 393–397
authorship, anthropomorphic 57 Blasingame, D. 230
automated browsing software 100–101 Bleacher Report 455
Automated Insights 53, 242 bloggers 15–16, 40, 42, 519
automated journalism 227, 237–238; algorithms for Bloomberg 213
54–55; availability of data and 244; bylines and Bobkowski, P. 420
full disclosure in 55–56; conclusions on 58–60, Boczkowski, P. 522
246–247; data and method on 241; discussion Boden, M. 57
on 57–58; findings on 242–246; further Bødker, H. 520
reading on 60–61, 247; introduction to 53–54; Bogart, L. 177
journalists as relevant group and 239–241; legal Bogost, I. 285, 450, 453, 454–455
views on computer-generated works in 56–57; Boland-Rudder, H. 202–205
literature review on 54–57; predictive journalism Bonfadelli, H. 490
and 245; previous research on 238–241; social Bonnier Group 192
friction as barrier to 243; technological changes Borges-Rey, E. 229, 231, 233
and 246; what to automate in 243–244 Boulianne, S. 490
automation anxiety 238 boundary objects 406–407
autonomous values and pop-up news ecologies Bounegru, L. 353
381–382 Bourdieu, P. 32
autonomy of news subjects 331–332 boyd, d. 151, 338, 443
Axel Springer 174–175, 178 branded content 465
branding, journalistic 441–442; enactment of
Badham, V. 343, 345 445–446; explication of 442–445; further
Baker, E. 505 reading on 448–449; impact of 446–448
Baker, G. 466 Branding (Health) Journalism: Perceptions, Practices, and
Barak, A. 426–427 Emerging Norms 448
Barnard, S. R. 80 brands, sustainability of newspaper 177, 177–178
Barney, D. 504 Braun, J. A. 482
Bassett, D. 176 Brems, C. 445
Bauer Media Group 192–193, 194 Brexit 348
Bauman, Z. 93 Broder, D. 330
Baym, N. 520 Broersma, M. J. 252
BBC 212, 356, 406 Bronskill, J. 361, 363
Bearing Witness While Black: African Americans, Brown, M. 380, 381, 392, 393, 395
Smartphones and the New Protest #journalism 398 Brucato, B. 391
Beck, H. 266 Bruner, J. 392
Bell, E. 360, 361 Bruns, A. 339
Benkler, Y. 503, 505 Buchalter, E. 301
Bergstrom, K. 340 Bücher, C. 457
Berliner Morgenpost 175 Bucher, T. 233
Berns, N. 429 Burgess, J. 339
Bettig, R. V. 187 Business a.m. 192
Bibby, A. 189 Business Week 192
big data 226, 284, 314, 329, 405–406 BuzzFeed 213, 251, 300, 361, 465–466
Bild 298 bylines and full disclosure in automated journalism
Bill & Melinda Gates Foundation 200–201 55–56, 59–60
528
Index
Cameron, G. T. 470 content analysis 93; automated browsing software

Campbell, D. 361, 364, 365, 367, 371 for 100–101; background on 93–97; case study
Canadian Association of Journalists 21, 328 97–101; conclusion on 102; content retrieval
Canadian Media Guild (CMG) 190, 192 software for 99; further reading on 103; liquid
Carlson, M. 54, 467, 468 news and 94–95; manually saving content for
Carlson, R. 451, 452 98–99; methods for freezing liquid news content
Carpenter, S. 445 for 96–97; structure of online web pages and
Carpentier, N. 146 95–96; weighing options and future directions
Carvin, A. 302 for 101–102
Cascading Style Sheets (CSS) 95–96 content generation, algorithmic 59
CasperJS 100–102 content integration 465
Castells, M. 409, 503 contested history 349–350
Castile, P. 305, 380, 383, 398 contested publics 341
CBSChicago.com 430 convergence and commodification with games
censorship, government see Turkey 456–459
Center for Public Integrity 201 Conway, B. A. 491
Centre for Investigative Journalism 361 Cooper, A. 412
Chartbeat 143 Copyright, Designs and Patents Act (CDPA) 56
Chen, G. M. 482 Costeja González, M. 325–327, 329
Chicago News Guild (CNG) 190 Court of Justice of the European Union (CJEU)
Chicago Tribune 213, 350 325–326
Chomsky, N. 507 court reporters 23–24
Christopher 329 Cox, M. 349
Chyi, I. 168, 169 Cramer, H. 253
circulation of news 519–520 Crawford, K. 340
Cision 79–80 creativity criterion 56–57
citizen and activist livestreams 300–301 Criado-Perez, C. 430–431
Citron, D. 426 critical perspectives on journalism 5–6
Cleveland Plain Dealer 214 crowdsourcing 231
climate change see environmental media Cullors, P. 388
Clinton, H. 49, 355, 422, 505 cultivation theory 489–490
CNN 299, 381, 392, 393 curation, algorithmic 132–133; alternative design
CNN.com 158 for studying 139; conclusion on 140; description
Coddington, M. 239, 354 of 133–134; directed and undirected 134–135;
code and data journalism 289–291 experimentation 137–138; four prototypical
Coe, K. 479 designs for assessing 135, 135–136; further
Cohen, N. S. 189, 191, 194, 457 reading on 140; manipulation 137; observation
Cohen, R. 88 138–139; simulation 136
Coleman, S. 376 Curtin, P. A. 470
Columbia Journalism Review 446 Cushion, S. 194
Comcast 213 Cvetkovski, T. 188
Comey, J. 303 cyber abuse of women 425; consequences of
comments, user see user comments 427–428; critical absence in the conversation
Communication Research 493 about 433; discussion of 432; domestic violence
Communication Review 460 and 428–430; examples and effects of 430–431;
Communication Theory 493, 495 further reading on 434; gendered nature of
Communication Workers of America (CWA) 190 the Internet and 425–426; how cyberspace is
Communist Manifesto, The 389 conceptualized and 428; nomenclature in 426;
community members, interviews of local 109–110 prevalence of 426–427; recommendations from
computational journalism 226, 232–233; ethical victims and researchers on 433–434
approaches to 313–319
computational thinking 232–233 Dagiral, E. 353–354
computer-assisted reporting (CAR) see data Dahlgren, P. 1, 457
journalism Daily Caller 382
connected and semi-experimental ethnography Daily Mail, The 336, 363
151–152 Daily Telegraph, The 336, 350, 363
529
Index
Dallas Morning News 159 digital democracy discourse 503–505; further

data collection, government see freedom, reading on 510
journalistic Digital Detach 413
data journalism 225–226; characterizing Digital Detox app 413
computational journalism and 229–233; code digital economy, print journalism industries and
and 289–291; conclusion on 233–234; form in employment conditions in 187–189
287–289; further reading on 234–235; interface digital entrepreneurs 65; practice and discourse of
in 286–287; limited role of journalists in digital entrepreneurial journalism by 66; see also
245–246; operation in 291–292; as a platform entrepreneurial journalists
284–294; programmer journalists in 354; state of Digital Journalism 226, 495
scholarship on 226–227; typologies of practices digital journalism practice: algorithmic
in 227–229; visual network exploration for (see transparency and 251–261; algorithms and
visual network exploration); see also interactive 237–247; data journalism as a platform and
journalism 284–294; defining and mapping data journalism
Data Protection Act 1998, UK 21 in 225–235; social media livestreaming and
Death and Life of American Journalism, The 220 296–306; visual networks and 265–282
death and violence, livestreaming of 301–303; Digital Journalism Studies 1–2; circulation of
potential effects of viewing 303–304 news and 519–520; critical perspectives on 5–6;
decentralized informational ecology 517–519 decentralized informational ecology in 517–519;
Décodex case study: controversy over 268–269; development in methods in 6–8; dialogues in
linking patterns in 277–282, 278–281; visual (see dialogues in Digital Journalism Studies);
exploration of network in 269–277, 271–276 further brightening the corners of 8–9; new
De Correspondent 66, 67, 69; engaging audiences complexities, new debates, and fundamental
72–73; purpose of journalism and 71–72 concerns in 2–3; news as cultural form and
defamation 336–337; conclusion on 345; further 523–524; research design (see research design);
reading on 345; journalism and 337–339; revisiting journalism studies’ tripartite approach
journalistic negotiation of social media spaces in 3–5; shift to news use and 520–521; situating
and 342–345; journalistic participation in social journalism in the digital 515–524; structure of
media spaces and 341–342; online spaces and book 9–10
challenges to civic dialog and 339–341 digital journalists: death of the author and rise of
Deibert, R. 502 automated 53–60; entrepreneurial 64–74; law
Deleuze, G. 402 defining 15–25; professional identity of 40–50;
De Maeyer, J. 354 role conceptions in the digital age and 28–37
democracy 198; further reading on 208; impact- digital labour 457–458
oriented investigative journalism and 200; digital limits: censorship and surveillance 501–510;
impact typology for investigative journalism games and 450–460; journalistic branding and
and 200–201; International Consortium of 441–449; media theory and 487–496; native
Investigative Journalists (ICIJ) and 201–208; advertising and 463–472; user comments and
journalism’s (potential) impact on 199–200 475–483
Democracy’s Detectives 218 digital news ecosystem: analyzing the dynamics
DeNardis, L. 503, 504, 505 of news diffusion processes in 124–127,
Deprez, A. 80 125–126, 128; basic research design for
Der Spiegel 365 tracking news diffusion processes online in
designated market areas (DMAs) 160 120–122; comparing the dynamics of online
Detroit Free Press 158 news diffusion processes by process attributes
Detroit News 158 in 128; computational methods enabling
Deuze, M. 35, 66, 93–95, 145, 239, 251, 492 reconstruction of news diffusion processes in
Diakopoulos, N. 232, 240, 314, 353, 455, 478, 480, 482 119, 119–120; conclusion and directions for
dialogues in Digital Journalism Studies: defamation future research on 128–130, 129; data collection
and social media 336–345; ethical approaches to on 121–122; exploring dynamics of online news
computational journalism 313–320; hacking and diffusion processes in large data collections in
expansive boundaries of journalism 348–357; 126–127; further reading on 130; introduction
journalistic freedom and surveillance of to 118–120; reconstructing news diffusion
journalists post-Snowden 360–371; right to be processes in 123–124; sampling unit 120–121;
forgotten 324–332 structuring samples of digital news diffusion
DiCaro, J. 430 processes by diffusion dynamics 127, 128; unit
diffusion, news see digital news ecosystem of analysis for 122–123
digital citizenship 340 “digital suicide” 415
530
Index
direct curation 134–135 Erjavec, K. 468, 470

disclosure transparency 54 ESPN 431
discursive interfaces, environmental platforms as Essa, I. 353
404–405 ethics 313–314; challenges, in computational
Disney 213 journalism 314–316; defamation and (see
Distaso, M. W. 129–130 defamation); further readings on 319;
diverse audiences 32–33 implications for journalism 318–319; outlook
diversification of actors in the journalistic field on 319; in right to be forgotten 326–327; socio-
33–34 technical concepts of agency and 316–317, 317
Dix, S. 470 ethnography, connected and semi-experimental
domestic violence 428–430; see also cyber abuse of 151–152
women Ettema, J. S. 199–201
Döpfner, M. 178 EU Observer 203
Douthat, R. 413 European Community 325
Drudge Report 382 European Court of Justice 22
Dumas, D. 344 European Journal of Communication 460
Dvorkin, L. 43 European Union 409; right to be forgotten in 324,
Dyer-Witheford, N. 191 325–326
Evans, I. 393–394
Eastin, M. S. 489 Everysport.com 241
eBay 215 Ewing, E. 397
ecosystem, digital news see digital news ecosystem expectations as generators of roles 30
editorial integration 465 experimentation 137–138
Editor & Publisher 237
Edström, M. 465 Facebook 42, 133–134, 168, 218, 294; accused of
Efecto Cocuyo 301 ‘swallowing journalism’ 158; agenda-setting
Eisenhower, D. 349 theory and 491; birth of the ‘hacker journalist’
Eisenstein, E. 416 and 352; defamation on 336, 343–344; directed
e-lancers see freelance journalists curation and 134; livestreaming on 296–297,
Eldridge, S. A. 50, 519 301–306, 380, 383; non-adopters of 420–421;
Elements of Journalism, The 70 opting out of 415; pop-up news ecologies and
Ellick, A. 298 377, 380; public service journalism and 219–220;
Ellison, N. B. 338 raw data collected by 138; in Turkey 510
El Pais 350 Fairfax Media 173–175, 175, 341, 344
Elzie, J. 381 Fairness Doctrine 212
embedded marketing 465 Faris, R. 80
Employment Relationships in the Media Industry 195 Fast Company 442
Encyclopedia of Complexity and Systems Science 282 Federal Communication Commission 216
Engelbart, D. 239 Federalist, The 431
Engels, F. 389 Feinstein, A. 303
entrepreneurial journalists: defining ‘good’ Ferenbok, J. 391
journalism and 68–70; digital beyond Ferrell, B. 394, 397, 398
technology and 73–74; digital entrepreneurial Ferrer-Conill, R. 454–455
journalism as practice and discourse by 66; Ferrier, M. 105
engaging audiences 72–73; further reading Ferrucci, P. 34
on 74; introduction to 64; journalism and ‘the Ferry, H. J. 299
digital entrepreneur’ 65; purpose of journalism Fink, K. 354, 356
and 70–72; reconceptualization of good First Amendment 219, 324
journalism by 66–73; as remedy to newspaper First Look Media 215
crisis 179–180 Fisher, C. 65
environmental media 401–402; as assembled Fisher, D. 361–363, 365, 367
networks 402–403; environmental open data Five Eyes 360, 362, 364, 366, 368, 371
platforms and 403–404; further reading on “Five I’s” 492–493
409–410; platformization and 404–409 Flaounas, I. 81
environmental open data platforms 403–404 Forbes 53
Environmental Protection Agency (EPA) see force-directed layouts 266–268, 267
environmental media Ford, C. 343–345
Erikson, E. H. 442 form and data journalism 287–289
531
Index
Fortune 349 Gawker 465–466

Forward, The 431 Gaziano, C. 490
Foucault, M. 389–390, 391 GCHQ 360, 364, 365, 369
‘fourth estate,’ journalism as the 198 gendered nature of the Internet 425–426
Fowler, A. 361, 363, 367, 368, 370 Ghorayshi, A. 300
Foxman, M. 452, 458 Giddens, A. 317
FoxNews.com 158 Gillespie, T. 331, 404
Franklin, B. 305 Gillis, A. 192
Frasca, G. 453 GitHub 294
freedom, journalistic 360–361; conclusions on Gitlin, T. 509
370–371; damage to national security and 367; Glass, Z. 504
Edward Snowden and 360–367; further reading Glasser, T. L. 199–201
on 372; impact of government data collection Global War for Internet Governance, The 510
on 364–365; laws regarding 368–370; research Globe and Mail 188
methods on 365–367; responsibility of the Godwin, R. Jr. 301, 302
media and 368 Godwin, R. 302
Freedom of Information Act (FOIA) 244, 288, Goffman, E. 240
290–291 Goldsmith, J. 504
free expression 328 Gomez, R. 416
Free Flow of Information Act, US 15, 17–18 good journalism: defining 68–70; engaging
freelance journalist: branding by 445–446; audiences 72–73; entrepreneurial
conclusions on 194; further reading on 194–195 reconceptualization of 66–73; purpose of 70–72
freelance journalists 186–187; digital Goodlad, N. 470
communications campaigns by, to resist rights- Goodman, E. 480
grabbing contracts 191–193; labor union Google 42–43, 158, 168, 218, 294; birth of the
organizing by 189–191; print journalism ‘hacker journalist’ and 352; directed curation
industries and employment conditions in digital and 134; display ads and 134; fake accounts on
economy and 187–189 137; public service journalism and 219; right to
Freelancing in America 195 be forgotten and 325, 326, 330
Freelon, D. 382 Google Analytics 143, 159
free speech 330 Google Chrome 100–102
French, A. 381 Google Home 143
Friedland, L. 111 Google Maps 150, 418
From ‘Selfies’ to Breaking Tweets: How Journalists Google News 455
Negotiate Personal and Professional Identity on Google Search 133, 137
Social Media 449 Google Translate 112
Früchterman, T. M. 268 Gothamist 382
Fuchs, C. 456, 457, 460, 505 government censorship see Turkey
full disclosure policy 55–56, 59–60 government data collection see freedom, journalistic
Graeff, E. 376, 379
game journalism 458 Graham, L. 422
#GamerGate 426, 428, 430, 452 Gramsci, A. 389–390, 509
games 450–451; bringing news into 453–454; Gray, F. 297, 396
concluding remarks on 459–460; convergence Great Recession 182, 443
and commodification with 456–459; employed Green, A. 176
in journalism 454–456; further reading on 460; Green, D. 200–201
#GamerGate and 426, 428, 430, 452; journalistic Greenpeace 407
coverage of digital 451–453; literature review on Greenwald, G. 215, 360, 368
451–456 Groshek, J. 80
gamified journalism practice 459 Guardian, The 44, 297, 304; on cyber abuse
gamified news see newsgames 431; data journalism and 228, 231–232, 349;
Gannett 174, 175, 213 defamation and 343; Edward Snowden and
Gans, H. 198, 199 363, 365, 368; on the EPA 401–402; open data
Garcia-Avilés, J. A. 50 platforms 403–404, 406
Garza, A. 388 Guattari, F. 402
Gatekeeping in the Digital Age 130 Guide to Electronic Rights 190
Gates, B. 412 Günther, E. 122
532
Index
Guo, L. 81, 85 journalists 48–50; relative to the journalism field

Gynnild, A. 227–229, 240 45–46; thinking about 40–42
Ikonen, P. 470
Habermas, J. 478 illusion of control 286–287
hacks/hackers 294, 348–349; birth of the ‘hacker impact-oriented investigative journalism 200
journalist’ and 351–352; contested history and InfoAmazonia 402, 404, 406
349–350; further reading on 357; reassessment Infoglut: How too Much Information is Changing the
of 355–357; who are 352–355 Way We Think 510
Hager, N. 361, 363, 366, 368, 370, 371 information vacuum filled by pop-up news
Hall, S. 390, 433, 520 ecologies 379–380
Hamburger Abendblatt 175 infrastructure, local journalism 107–110
Hammond, K. 241 input-output pipeline 256–259
Hammond, P. 232 Institute for Journalism in New Media 215
Handbook of Journalism 522 Instone, L. 408
Handbuch Medien- und Informationsethik 319 integrative content generation 59–60
Hanitzsch, T. 33, 36, 522 interactive journalism 226, 348–349; birth of the
Hannity, S. 49 ‘hacker journalist’ and 351–352; contested
Hanusch, F. 33 history and 349–350; further reading on 357;
Haraway, D. 408 livestreaming as 298–299; reassessment of
Harbers, Frank 66 355–357; who is working in 352–355; see also
harm reduction and right to be forgotten 328–329, data journalism
330–331 Interactive Journalism: Hackers, Data, and Code 234,
Harro-Loit, H. 470 355, 357
Hartzog, W. 331 Intercept, The 215
Hate Crimes in Cyberspace 434 interface in data journalism 286–287
Hayes, D. 192 International Consortium of Investigative
Helmond, A. 404 Journalists (ICIJ) 198, 266; broader impact of
Henry, N. 426, 428 205–208; and its typology 201–205
Herbst, S. 479 International Federation of Journalists 186
Herfindahl-Hirschman Index (HHI) 113 Internet Archive 95, 110
Herman, E. S. 507 investigative journalism: broader impact of
Hermida, A. 79 205–208; impact-oriented 200; impact typology
Hess, Amanda 425, 431 for ICIJ 200–201
Hindman, D. B. 490 Ioff, J. 431
Hindman, M. 504 iSpy: Surveillance and Power in the Interactive Era 510
historical integrity and right to be forgotten 328, 330 Ivy, J. 303
Hockey, J. 341
Holovaty, A. 226, 351 Jackson, M. 297
hooks, b. 390 Jackson, R. 389
Hotjar 143 Jackson, S. J. 380
Howard, A. 230, 354 Jasanoff, S. 504
HTTrack 97, 99, 101 JavaScript 95–96, 97, 99, 100
Huffington Post 382, 430, 455, 465–466 Jefferson, T. 305
human-bot attribution gap 57 Jenkins, H. 503, 504
human-machine advancement 240–241 John S. and James L. Knight Foundation 200
humans and algorithms 255–256 Johnson, J. 20
Hurst, C. 302 Johnson, M. 387
Hussain, M. 129 Jones, M. 361, 363, 365–367
hybridizing of the TV live shot model using Journalism: A Critical History 471
livestreaming 299–300 Journalism after Snowden 372
HyperText Markup Language (HTML) 95–96, 100 journalism crisis, American 214–215; alternatives
Hywood, G. 173 to the advertising revenue model and 215–216;
newsgames and 458–459; see also newspaper crisis
identity, professional 40; conclusions on 50; Journalism Practice 74, 167
digital journalists’ view of their 44–45; field’s Journalistic Branding on Twitter: A Representative
view of digital journalists and 42–44; relative Study of Australian Journalists’ Profile
to organization role 46–48; relative to quasi- Descriptions 448
533
Index
journalists 15–16, 25; and approaches to deciding legal-journalistic gap 58

‘who is a journalist?’ 16–17; automated (see Le Monde 242, 244, 265, 282
automated journalism); entrepreneurial (see Leskovec, J. 129
entrepreneurial journalists); freelance (see freelance Leuven, S. V. 80
journalists); legal domains and definitions of 17; Levine, M. 332
national security and terrorism law and 20–21; Lewis, P. 297
as observers 23–24; as participants 17–23; privacy Lewis, S. 35, 81, 88, 226, 406
and data protection and 21–23; as relevant group LGBTQ community, cyber abuse of 427, 431
239–241; robo- (see automated journalism); role Lindell, J. 32
conceptions of (see role conceptions) Linux 99, 101
journalists, digital: conclusions on identity liquid journalism see content analysis
questions for 50; field’s view of 42–44; identity Lister 287
relative to quasi-journalists 48–50; identity livestreaming: activist and citizen 300–301;
relative to the journalism field 45–46; thinking Facebook Live 296–297; further reading on 306;
about identity of 40–42; view of their own hybridizing the TV live shot model 299–300; as
identity 44–45 new interactive storyform 298–299; Periscope
Journal of Computer-Mediated Communication 493 297–298; pop-up news ecologies and 380, 381,
Journal of Games Criticism 460 383; potential effects of viewing nonfictional
Justice and Development Party (AKP), Turkey violent images via 303–304; privacy and 305;
501–502, 507–509 suicide 304; technological determinism and
305–306; violence and death 301–303
Kahlor, L. 489 live text-based communication (LTBC) 23–24
Kaine, J. 299 living lab methodology 152
Karlsen, J. 354 LIWC tool 84, 84
Karlsson, M. 94–98, 101 local journalism 105–106; conclusions on 114–115;
Karppi, T. 415 further reading on 115; infrastructure of
KDE WebKit 100–101 107–110; metrics and analysis of 112–114; need
Kennamer, D. 331 for assessment tools for 106–107; output of 107,
Kerrang! 193 110–111; performance of 107, 111–112
Kiriakou, J. 366 Local Journalism: The Decline of Newspapers and the
Kirilenko, A. P. 80 Rise of Digital Media 115
Kizilcec, R. F. 253 Local Labs 245
Knight Commission 105, 106 Los Angeles Times 53, 213, 245, 251; livestreaming
knowledge gap hypothesis 490–491 and 304
Koliska, M. 232 Lowrey, W. 376, 377
Kovach, D. 70 Lund, A. 458, 460
Kovačič, M. 470 Luxembourg Leaks 202–204
Krautreporter 66–69, 71; engaging audiences 72–73
Ksiazek, T. B. 479 MacCallum, A. 466
Kuutti, H. 231 machine learning approaches to text classification
Kvale, S. 241 87–88, 88
Mack, R. L. 414
labor union organizing by freelance journalists Macnamara, J. 465
189–191 Maddow, R. 491
Latent Dirichlet allocation (LDA) 85, 87, 87, 88 Makuch, B. 369
Latour, B. 316, 317, 376, 402, 403, 409 Malik, M. M. 81–83, 87
Laubacher, R. 187 Malone, T. W. 187
La Vanguardia 325–326 Manchester Evening News 189
Lawrence Journal-World 351 Mandel, B. 431
laws: anti-terrorism 20–21, 370–371; on computer- manipulation 137
generated works 56–57; defamation 336–345; Mann, S. 391
journalistic freedom 368–370; legal domains Mantilla, K. 426
and definitions of journalists 17; right to be Marimow, B. 201
forgotten 324–332; Title IX 429 Marketplace of Attention, The 140
Lee, E.-J. 481 Markham, T. 392
legacy journalism 45–46, 48–49 Martin, F. 101
legacy media 45–47, 48, 67, 68, 70, 111, 451, 464, Martin, T. 379, 388
465, 491, 515 Marwick, A. E. 441, 443
534
Index
Marx, K. 389–390, 456, 459–460, 504 Mitchell, A. 422

Mashable 251 Mitrovica, A. 361–365, 367, 369
Massagio 175 Mitterand, F. 325
Mass Communication and Society 493, 495 Mittmedia 242
materiality and news flows 515–517; circulation of Modular Online Time Use Survey (MOTUS) –
news and 519–520; decentralized informational Translating an Existing Method in the 21st Century
ecology and 517–519; and news as cultural form 152
523–524; news medium and 522–523; shift to Mojo 193
news use and 520–521 Molyneux, L. 445–446
Mattelart, A. 194 Monge, P. 129
McBride, K. 302 Monmonier, M. 349, 350
McCain, J. 422 Montfort, N. 285
McChesney, R. 505 Moon, P. 299
McCombs, M. E. 199, 200, 491 Moore, A. 329
McCosker, A. 340 Moore, M. 300
McFadyen, G. 361 Moreno, J. 266–268
McGraw-Hill 192, 194 Morgan Stanley 176
McKesson, D. 380 Morrison, S. 416
Mechanical Turk 139 Morrissey, G. 426
Media Concentration: Why Ownership Matters 510 Morstatter, F. 82
media deserts 105 Mosslmani, Z. 336
media directories 109, 110 Mozilla Firefox 98, 100–102
Media Entertainment and Arts Alliance (MEAA) 20 Mozilla Gecko 100
Media Life 492 MSNBC 158, 422
media ownership 505–506 multiplatform newspapers, demand for 160–161
Mediapart 66–70, 74; purpose of journalism and Murdoch, R. 503
71–72 Musician 190
media policy and regulation 211–212; alternatives Musk, E. 412
to the advertising revenue model and 215–216; Myth of Digital Democracy, The 140, 510
American journalism crisis and 214–215; myth of enclaves 132, 139; see also curation, algorithmic
economic and regulatory discourse about
journalism and 216–218; further reading on Naaman, M. 478, 480, 482
220; importance for journalism 219–220; new Napoleonic Civil Code 324
American media landscape and 212–214; new Napoli, P. M. 201
areas for regulatory concern and 218–219; Nardi, B. A. 376
policy-led innovation in media and 179–180; Narrative Science 53, 241, 242
relevance to journalism 219–220 Nation, The 382
Mediated City: The News in a Postindustrial Context, National Aeronautics and Space Administration
The 384 (NASA) 401, 408
Mediating the Message in the 21st Century: A Media National Board for Consumer Disputes 242
Sociology Perspective 471 National Institute for Computer-Assisted
medium-agnostic ‘media user’ 146–147 Reporting (NICAR) 225, 231; birth of the
medium and material culture 522–523 ‘hacker journalist’ and 351; interactive
Meet the Press 422 journalism and 351, 355, 356
Meindl, J. 303 National Public Radio 406
Mellado, C. 33 national security: Edward Snowden and 367;
Meltzer, K. 479 terrorism law and 20–21
Merchant, C. 240 National Security Agency (NSA) 360, 361, 369
Merritt, D. 199, 200, 207 National Security Legislation Amendment Bill,
Messner, M. 129–130 Australia 20
Meyer, P. 349 National Union of Journalists (NUJ) 189–190, 191
Michalski, D. 228 National Writers Union (NWU) 188, 192
Miller, A. 173 native advertising 463–464; appropriation of
Miller, D. 286 journalistic reputation and 467–468; conclusions
minority voices: Black Lives Matter 387–398; cyber on 471; further reading on 471–472; many
abuse of women 425–434; environmental media names of 464–465; measuring 469–471; shifting
401–410; opting in and opting out of media discourse from agonism to symbiosis and
412–423; pop-up news ecologies and 375–384 465–467; as shiny camouflage 468–469
535
Index
Natural Language Generation (NLG) technologies 237 nonprofit journalism 198; broader impact of
NBC 422 205–208; further reading on 208; impact-oriented
Negroponte, N. 503 investigative journalism 200; impact typology
Netflix 133, 134, 466 for investigative journalism and 200–201;
net neutrality debate 219 International Consortium of Investigative
networks, assembled 402–403 Journalists (ICIJ) and 201–205; journalism’s
network text analysis, Twitter 84–85, 86 (potential) impact for democracy and 199–200
Neuberger, C. 80 Nool, J. 336
new American media landscape 212–214 Noort, E. van 79–80
New Orleans Times-Picayune 158, 214 NTB news agency 243
New Philosophy of Society: Assemblage Theory and Nygren, G. 354
Social Complexity, The 410 NYTimes.com 158
news as cultural form 523–524
news deserts 105, 215 Obama, B. 362, 366
news ecology 375–376 observation 138–139
newsgames 453–454, 458–459 observers, journalists as 23–24
Newsgames: Journalism at Play 450, 460 Occupy Wall Street movement 391
Newspaper Association of America (NAA) 159 O’Day, V. 376
newspaper crisis 172, 214–215; conclusions on OECD 189
182; consequences of 178–179; explaining Offshore Incorporations Limited 204
reach/revenue disparities in 175–177, 176; O’Hear, A. 57
further reading on 182; policy-led innovation Oliver, M. B. 489
and entrepreneurship and 179–180; product Omidyar, P. 215
perspective and 180–182; remedies for 179–182; online web pages, structure of 95–96
sustainability of newspaper brands and 177, Open Data movement 288
177–178; symptoms of 172–178, 174–175, open data providers, environmental platforms as
176–177 405–406
newspaper readership 157–158; digital advertising operation in data journalism 291–292
and subscription revenue from 159; empirical oppositional gaze 391
analysis of 161–162; findings on 162–163, opting out trend: beginnings of movement in
162–164, 164, 165–166, 167; further reading 412–413; coexistence of diverse technologies
on 169; growth of online 158; limitations and and 416–417; conceptual foundations in
future studies of 168; measurement of audiences 414–415; cultural and religious beliefs and
in 159–160; multiplatform demand in 160–161; 417–418; developing identity, asserting
reality and irrationality in declines of 164–168; dominance and 420–421; fears of new
transition from print to online 158 technologies and 418–419; further reading
news production, traditional 45–46 on 422–423; privacy concerns and 419–420;
news users see users, digital news technology as status symbol and 422
New York Times 48, 93, 158, 173–175, 174, 200, organization role, identity relative to 46–48
203, 213, 218, 251, 348, 413, 455, 466; birth originality criterion 56
of the ‘hacker journalist’ and 351; brand output, local journalism 107, 110–111
sustainability of 177–178; content analysis of
97–101; data journalism and 227, 355–357; Pacific Media Workers Guild (PMWG) 190
journalistic freedom and 366–367; livestreaming Pacific Standard 431
and 298–299; on livestreaming of violence 302; Packnett, B. 393, 398
open data platform 406; social media policy PageRank 133
of 446; on transparency 252; visual network Pai, A. 219
exploration and 267 paid content 465
New Zealand Law Commission (NZLC) 15, 16 paid-for-news 465
Neyland, D. 260 Panama Papers 202–205, 266
Ng, Y. M. M. 482 Parasie, S. 353–354
Nieborg, D. B. 452, 458 Parikka, J. 285
Nielsen, R. K. 520 Paris climate agreement 408
Nielsen Co. 150 Park, R. 376
Nissenbaum, H. 305 Parker, A. 302
NLG technology 246 Parks, G. 396
Noelle-Neumann, E. 488 participants, journalists as 17–23
nonlinear ‘media user’ 147–148 participatory journalism 494
536
Index
Participatory Journalism: Guarding Open Gates at Powell, A. 426, 428

Online Newspapers 345 power: and narratives of control 389–390; user
partisan news organizations 356–357, 382–383 comments challenging 480–481
Patel, M. 200–201 Powers, M. 353
Paul, R. 297 Poynter Institute 302
paywalls 159 precarity penalty 189
Pearson, M. 20 Precision Journalism 357
performance, local journalism 107, 111–112 predictive journalism 245
Periscope 297–298, 300, 301, 306 PrinceEA 422
Perreault, G. 452 print journalism industries, employment conditions
Personal Branding on Twitter: How Employed and in 187–189
Freelance Journalists Stage Themselves on Social privacy: data protection and 21–23; livestreaming
Media 448 and 305; opting out over concerns about
Peters, C. 144 419–420; right to 327–329; see also right to be
Peters, J. 16 forgotten (RTBF)
Pew Research Center 79, 106, 158, 216, 365, 413, Privacy Act, New Zealand 22
416, 427 Prodnik, J. A. 457
Pfeffer, J. 81–83, 87 product perspective and the newspaper crisis
Phantom JS 101 180–182
Phau, I. 470 Proferes, N. J. 79
Philadelphia Inquirer 201, 206 programmer journalists 354
Philadelphia Media Network (PMN) 215 ProPublica 53, 200, 215, 231, 355
Philips, F. 426 Protecting Journalism Sources in the Digital Age 372
Phillips, D. 394 Protecting Sources and Whistleblowers in the Digital
phone hacking scandal 284 Age 372
Pizzagate 355–356 Protess, D. L. 201
Plank, E. 300 proxy technology assessments 152
platformization and environmental media 404–409 Prushank, D. 429
platform studies 284–285; code in 289–291; Psychological Aspects of Cyberspace 434
concluding remarks on 292–294; form in 287–289; public deliberation, user comments as 478–480
further reading on 294; interface and illusion Putnam, R. D. 489
of control in 286–287; for non-computational
ecosystems 285–286; operation in 291–292 Q magazine 193
Platform Studies 294 quasi-journalists 48–50
Poitras, L. 215, 368 Quinn, Z. 430
political economy of digital journalism: digital
distribution 172–182; freelance journalists Raghuram, M. A. 81, 88
and 186–195; multiplatform readership and Rainforest Foundation 407–408
157–169; nonprofit journalists and 198–208; Rall, H. 354
regulation and 211–220 Rammert, W. 314, 316, 317
Politico Magazine 167 Ramos, D. 109
Politics of Media Policy, The 220 Rauch, J. 415
Pool, T. 299–300 ReachHero 175
Popular Science 342 reach/revenue disparities 175–177, 176
pop-up news ecologies: African Americans and readership see newspaper readership
378–380, 381, 382; amplified by external Reassembling the Social: An Introduction to Actor-
connectors 382–383; applying the concept of Network-Theory, Politics of Nature 409
379–383; coming into existence at high speeds Reboot 413
380–381; conclusion on 383–384; dependent on Rebuilding the News: Metropolitan Journalism in the
social media tools 383; emergence of 378–379; Digital Age 384
enacting autonomous values 381–382; filling an Reingold, E. M. 268
information vacuum 379–380; further reading relationship effects 494–495
on 384; growth of concept of news ecology and Report for the National Council for the Training of
375–376; introduction to 375; Syrian civil war Journalists 195
and 376–378, 380, 382 research design: Big Data 79–89; enclaves
Portwood-Stacer, L. 415 132–140; liquid news 93–103; local news
Postmedia 175 105–115; news diffusion 118–130; news use
post-moderation 342 studies 143–153
537
Index
Rethinking Maps: New Frontiers in Cartographic Saving the Media 220

Theory 409–410 Scahill, J. 215
Reuters Institute 305 scalable ‘media user’ 147
Review of Security and Counter Terrorism Legislation, Scarborough Research 160
Australia 20 Schaffer, J. 199
Reynolds, D. 305, 398 Schauster, E. E. 470
Ribbens, W. 452 Schibsted 174–175
Richardson, A. V. 381 Schiller, D. 505
Richardson, J. E. 477 Schulz, A. 488
Richelson, J. 361, 362–363 Schulz-Schaeffer, I. 314, 316, 317
right to be forgotten (RTBF) 324; autonomy of Schütz, W. 121
news subjects and 331–332; birth as a legal Schwartz, D. 431
right 324–326; clash of ethical principles in Science in Action 409
326–327; important distinctions in 329–331; Scotsman 285
‘unpublishing’ and right to privacy in 327–329 Scott, W. 395–396
Right to Be Forgotten: Privacy and the Media in the Scottish Newspapers’ Association of Photographers
Digital Age, The 332 (SNAP) 192
Ritzau Finans 242–243, 244 Screen Free Week 413
Riverfront Times 382 Seattle Post-Intelligencer 214
Robertson, H. 188 Seglins, D. 361, 363–364, 366, 367, 369
Robinson, S. 480 Selenium WebDriver 100–102
robot journalism 53, 227, 239, 247; see also self-employed journalists see freelance journalists
automated journalism Selinger, E. 331
Rocky Mountain News 214 sentiment analysis, Twitter 84, 84
Roessler, P. 488 September 11, 2001, terrorist attacks 20, 370
Rogers, S. 230 Sevignani, S. 456
role conceptions: challenges for scholarship on Shafer, J. 167
34–36; changes to 31–32; conclusions on 36–37; Shane, S. 361, 365, 366–367
diversification of actors in the journalistic field Shaw, D. L. 491
and 33–34; evolution of research on 28–29; Shaw, F. 339
multiple audiences, multiple expectations and Shennan, F. 16–17
32–33; role negotiation and 31–32; role theory shield laws 17–19
in study of 29–30; shift in relations between Shirdon, F. M. 369
audiences and journalists and 31; studying Shirky, C. 217
audience expectations of journalistic roles and 30 Shneiderman, B. 237
role strain 32 signal intelligence (SIGINT) 362, 370
role theory in study of role conceptions 29–30 Silver, N. 348
Ronzheimer, P. 298 Silverstone, R. 146
Rosenberg, H. 304 Simons, M. 201
Rosenstiel, T. 70 simulation 136
Ross, J. 192 Sinclair company 218
Routledge Handbook of Journalism Studies 495 Singer, J. B. 32
Royal, C. 230 Sinyangwe, S. 395–396
RT (Russia Today) 383 Siri (smart assistant) 293
Ruths, D. 88 60 Minutes 412
Ryle, G. 202, 206 Sjøvaag, H. 94–95, 97
Slate 356, 431
Sabbath Manifesto 413 Slater v Blomfield 19
Sage Handbook of Digital Journalism, The 384 smartphones 143, 418–419
Saks, K. 470 Smit, G. 354
Salamon, E. 187 Smith, C. 397
salvation of journalism through digital journalists Smith, J. 420
43–44 Smythe, D. 456–457, 460
Samuelson, P. 58 Snowden, E. 360–361, 366, 370; on collection
Sanders, B. 387 obsession 364; damage to national security
Sartoretto, P. 32 and 367; impact of 362–364; perceptions of
Saurwein, F. 319 361–362; responsibility of the media and 368
538
Index
Social Construction of Technological Systems: New symbiosis 465–467

Directions in the Sociology and History of Technology, Syrian civil war 377–378, 380, 382
The 319
social entrepreneurship 67 Takahashi, B. 392
social friction as barrier to automation 243 Taking Journalism Seriously 522
social media 43; defamation and 338–345; directed TaleSpin software 55
and undirected curation 134–135; journalistic Tandoc, E. 16, 33, 80, 392, 445
branding and 441–449; journalistic negotiation Tasini, J. 188
of spaces of 342–345; journalistic participation Taylor, P. 361, 362, 365, 367
in spaces of 341–342; journalist use of 79–81; TC Transcontinental 194
knowledge gap hypothesis and 490–491; technological determinism 305–306, 415
livestreaming on (see livestreaming); news sites technology, digital 73–74; coexistence of diverse
and 158; non-adopters of 420; opting out of 416–417; fears of 418–419; opting out of using
(see opting out trend); pop-up news ecologies (see opting out trend); platformization in 404–409;
and 377; pop-up news ecologies dependent on as status symbol 422; see also hacks/hackers
383; stars of 42; see also Twitter technology affordances 73, 522
Society of Professional Journalists 331 Teixeira, C. 453
socio-technical concepts of agency 316–317, 317 Television, Technology and Cultural Form 422–423
sousveillance 391–392 Television & New Media 460
Spain, S. 430 terrorism law 20–21, 370–371
Sparks, C. 414 text classification, machine learning approaches to
Spayd, L. 252 87–88, 88
Spin, Spies and the Fourth Estate: British Intelligence Textual Relations 242
and the Media 372 theorizing digital journalism: agenda-setting theory
spiral of silence theory 488–489 in 491–492; alternative responses in 493–494;
Splendore, S. 228, 233 call for ‘relationship effects’ in 494–495;
Splichal, S. 457 cultivation theory in 489–490; “Five I’s” in
sponsored content 465 492–493; further reading on 495; knowledge
Spotify 133, 134 gap hypothesis in 490–491; limited effects of
Spyridou, L-P. 65 487–488; spiral of silence theory in 488–489
stalking 428–429 Thompson, M. 173
Stanyer, J. 477 Thurman, N. 79
Starkman, D. 213 Time magazine 349, 394, 503, 504
State of the News Media 158, 160 time use research 149–150
Stavelin, E. 354 Time Warner 213, 218
Steegen, R. 452 Title IX 429
Steenson, S. 516, 521 Toch, H. 391
Steinem, G. 298 Tometi, O. 388
Stepchenkova, S. O. 80 topic modeling, Twitter 85–87, 87
Sterling, A. 393 Toronto Star 330
Stewart, C. 394, 395–396 Tow Center for Digital Journalism 200
Stieger, S. 415 tracking and logging studies 150–151
Stoddart, M. C. 389–390 transparency 233, 251–252; algorithmic 54, 253;
Story, L. 298 consequences and challenges to algorithmic
storytelling 230, 238; interactive journalism and 259; data journalism and 230; discussion and
355; see also visual network exploration conclusion on 259–261; exploring algorithmic
Strömbäck, J. 96–97, 98 253–254; further readings on 261; input-output
Studies in Communication | Media (SCM) 483 pipeline 256–259; as ‘new objectivity’ 252; and
subnational authoritarianism 379 the news media 252–253; study findings on
suicide, livestreaming of 304 254–256, 255
Sullivan, A. 413 transparency gap 57
Sullivan, M. 305 Trédan, O. 354
Sun, The 363 Trial and Error: US Newspapers’ Digital Struggles
support vector machines (SVM) 87–88, 88 Toward Inferiority 169
sustainability of newspaper brands 177, 177–178 Tribune company 218
SwissLeaks 202–205 Trilling, D. 122, 129
Syed, N. 361 tripleC: Communication, Capitalism & Critique 460
539
Index
Trudeau, J. 369 Usher, N. 226, 228–229, 234

Trump, D. 212, 214, 219, 298, 300, 348, 365, 401, Uskali, T. I. 231
408, 431 U.S. Register of Copyrights 56
Turkey 501–503; conclusion on 509–510; Justice Ustream 380
and Development Party (AKP) 501–502,
507–509; media environment in 506, 506–507 Value Creation and the Future of News Organizations:
Turow, J. 505 Why and How Journalism Must Change to Remain
Twitter 47, 49, 168, 301; agenda-setting theory Relevant in the Twenty-First Century 182
and 491; Black Lives Matter movement and (see van de Velde, B. 122
Black Lives Matter movement); cyber abuse of van Reijmersdal, E. 470
women on 425; defamation on 339; directed Verizon 213
curation and 134; discussion on results from Verweij, P. 79–80
study of 88–89; e-lancers use of 193; embedded Vice Media 213
in news reporting 80; further reading on Vine 381
89; growth and popularity of 79; harvesting violence and death, livestreaming of 301–303;
data from 81–82; journalistic branding and potential effects of viewing 303–304
442–443; LDA analysis of 81; machine learning virtual unions 191
approaches to text classification 87–88, 88; visualization 230
network text analysis 84–85, 86; non-adopters visual network exploration 265; conclusions on
of 420–421; Periscope on 297–298; pop-up 282; Décodex case study 268–282, 271–276;
news ecologies and 379–380; preprocessing data further reading on 282; introduction to
from 82–83, 83–84; relevance for journalists 265–266; understanding force-directed layouts
and news media 79; sentiment analysis 84, 84; and 266–268, 267
topic modeling 85–87, 87; in Turkey 501–502, Vos, T. 32, 34, 452
507–509, 510; word clouds from 83–84 Vox.com 300
Vox Media 213
Uber 175
‘unbiased bias’ gap 58 Wagemans, A. 66
undirected curation 134–135 Wahl-Jorgensen, K. 376, 377, 522
UNESCO 364–365, 369–370, 371 Wall Street Journal 351, 466, 507
United Auto Workers (UAW) 190 Walters, A. 79
United Nations 409; Convention on Human Ward, Adam 302
Rights 325 War on Journalism, The 363
United Robots 241 Washington Post 213, 228, 351
‘unpublishing’ 327–329 Watergate scandal 70, 363
Urbana-Champaign Independent Media Center 219 Waters, N. 22
USA Freedom Act 369 Wayback Machine 330
USA Today 213, 342, 406 Weber, M. S. 129
USAToday.com 158 Weber, P. 477
user comments 475; demography and traits of Weber, W. 354
people leaving 476; effects of 481–482; further web pages, structure of 95–96
reading on 483; future research on 482–483; in/ Wei, L. 490
civility in 479–480; motivations and inhibitors Welch, D. 361, 363
in 476–478; organizational perspectives and Welles, B. F. 380
policies on 480; as political action 480–481; Wells, I. B. 398
profiling commenters and lurkers in 475–478; as Westley-MacLean model 495
public deliberation 478–480 Westlund, O. 35
users, digital news 143–144; challenges ahead Wget 97, 99, 101
with studying 144–145; connected and semi- Whale Oil 19
experimental ethnography on 151–152; further White, R. 341
readings on 152–153; how to find 152; in Widholm, A. 96
journalism studies 144; medium-agnostic ‘media WikiLeaks 50, 230, 355–356, 505, 507, 510, 519
user’ 146–147; methodological advancements Williams, A. 215
in studies of 148–152; nonlinear ‘media user’ Williams, R. 414, 415, 422
147–148; scalable ‘media user’ 147; theoretical Williams, S. A. 79
advancement in studies of 145–148; time use Wing, J. M. 232
research on 149–150; tracking and logging Winner, L. 239
studies on 150–151; see also audiences Wired 431
540
Index
Witschge, T. 35, 66 Wu, M. 469, 470

Witty, P. 298 Wu, T. 504
women, cyber abuse of see cyber abuse of women
Woodstock 415 Yahoo News 158
word clouds 83–84 ‘yellow’ journalism 337
World Association of Newspapers 478, 517 YouTube 134, 137, 377, 378, 383, 430, 522
World Bank 409
World Trade Center terrorist attacks 20 Zamith, R. 81, 96–102, 118
World Wide Fund for nature 407 Zelizer, B. 522
Woszniak, S. 412 Ziegele, M. 477
Wretched of the Earth, The 390 Zimmer, M. 79
Writers’ Rights: Freelance Journalism in a Digital Age 194 Zimmerman, G. 388
Wu, B. 430 Zuckerberg, M. 220, 296, 302, 306
541

Scott Eldridge II - Bob Franklin - The Routledge Handbook of Developments in Digital Journalism Studies-Routledge (2018)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Scott Eldridge II - Bob Franklin - The Routledge Handbook of Developments in Digital Journalism Studies-Routledge (2018)

Uploaded by

Copyright:

Available Formats

THE ROUTLEDGE HANDBOOK OF

Scott A. Eldridge II is an assistant professor at the University of Groningen, the Netherlands.

Edited by Scott A. Eldridge II and Bob Franklin

Introduction: introducing the complexities of developments

1 Law defining journalists: who’s who in the age of digital media? 15

2 Studying role conceptions in the digital age: a critical appraisal 28

3 Who am I? Perceptions of digital journalists’ professional identity 40

4 The death of the author, the rise of the robo-journalist: authorship,

5 The entrepreneurial journalist 64

6 Content analysis of Twitter: big data, big studies 79

7 Innovation in content analysis: freezing the flow of liquid news 93

8 An approach to assessing the robustness of local news provision 105

9 Reconstructing the dynamics of the digital news ecosystem: a case

10 Testing the myth of enclaves: a discussion of research designs

11 Digital news users . . . and how to find them: theoretical and

12 What if the future is not all digital? trends in U.S. newspapers’

13 On digital distribution’s failure to solve newspapers’ existential crisis:

14 Precarious e-lancers: freelance journalists’ rights, contracts, labor

15 What can nonprofit journalists actually do for democracy? 198

16 Digital journalism and regulation: ownership and control 211

17 Defining and mapping data journalism and computational journalism:

19 Disclose, decode, and demystify: an empirical guide to algorithmic

20 Visual network exploration for data journalists 265

21 Data journalism as a platform: architecture, agents, protocols 284

22 Social media livestreaming 296

23 Ethical approaches to computational journalism 313

24 Who owns the news? The ‘right to be forgotten’ and journalists’

25 Defamation in unbounded spaces: journalism and social media 336

26 Hacks, hackers, and the expansive boundaries of journalism 348

27 Journalistic freedom and the surveillance of journalists

29 The movement and its mobile journalism: a phenomenology

30 Nature as knowledge: the politics of science, open data, and

31 Opting in and opting out of media 412

33 Social media and journalistic branding: explication, enactment,

34 Reconsidering the intersection between digital journalism and games:

35 Native advertising and the appropriation of journalistic clout 463

36 User comments in digital journalism: current research and future

37 Theorizing digital journalism: the limits of linearity and the rise

38 Outsourcing censorship and surveillance: the privatization of governance

Epilogue: situating journalism in the digital: a plea for studying

20.9 Distribution of the number of nodes per category 279

6.1 Sentiment analysis categories comparing news-related tweets with

Claudette Guzan Artwick is Associate Professor of Journalism and Mass Communications at

communication, social movements, political communication, and qualitative and quantitative

Florian Buhl, M.A., is a research assistant at the Department of Communication, University of

Aras Coskuntuncel is a Ph.D. candidate and instructor in communication at American Uni-

Nicholas Diakopoulos is an assistant professor at the Northwestern University School of

Raul Ferrer-Conill is a doctoral candidate in the department of Media and Communication

Patrick Ferrucci (Ph.D., University of Missouri) is an assistant professor in the Department

Jonathan Gray is Lecturer in Critical Infrastructure Studies at the Department of Digital

Elisabeth Günther is a Ph.D. candidate at the Department of Communication, University of

Folker Hanusch, Ph.D., is Professor of Journalism in the Department of Communication at the

Jane Johnston is Associate Professor of Communication at The University of Queensland. She

Michael Karlsson is a professor in Media and Communication at Karlstad University, Sweden.

Michael Koliska is an assistant professor for journalism at Georgetown University’s Communi-

Magda Konieczna is an assistant professor in the journalism department at Temple University.

Kathleen McCollough is a Visiting Assistant Professor of Media Studies at Augustana Univer-

Logan Molyneux is an assistant professor of journalism at Temple University. His research

Merja Myllylahti is a post-doctoral fellow at Auckland University of Technology (AUT), New

Pamela Hill Nettleton holds a BS in journalism, an MA in mass communication, and a

Jacob Ørmen is an Assistant Professor at the University of Copenhagen. He studies audiences

Thorsten Quandt is professor of online communication at the University of Münster, Germany.

Allissa V. Richardson, Ph.D., is Assistant Professor of Journalism at the University of Southern