Easy Data, same old platforms?

- A Systematic Review of Digital Activism Methodologies

Suay Melisa Özkula suaymelisa.ozkula@unitn.it
Paul J. Reilly paul.reilly@glasgow.ac.uk
Jenny Hayes jhayes4@sheffield.ac.uk
 Easy Data, same old platforms?
A Systematic Review of Digital Activism Methodologies

Suay M. Özkula, Paul J. Reilly, and Jenny Hayes

 Twitter fatigue?

Digital bias? (Marres, 2017)

How much do we really know about digital activism?

 Bruns & Burgess (2015) on easy data & hard data

H#1: increasing prevalence of digital, digitised & hybrid methodologies

(esp. computational methods)

H#2: → preponderance of ‘easy data’ from ‘easy’ platforms

(e.g. Twitter)

H#3: → reductionism: “the low-hanging fruit of Twitter data”

= a glut of moderately sized studies of hashtag/keyword & @reply networks
 Hypotheses & question

What methodological questions are common in D.A. research?

H1: Following the computational turn, research methodologies draw

predominantly on digital, digitised, and hybrid methods in their data
collection.

H2: Empirical digital activism research predominantly focuses on

single-platform Twitter data (H2a) that is relatively easy to access and
monitor through APIs such as hashtags and reply networks (H2b).

RQ: To what extent are Twitter studies of activism prone to digital bias or
issues of representativeness?
The corpus

Total N: 315 articles

Trad. sources (only): 117 articles

Digital sources: 198 articles

Timespan: 2011-2018

Coded on methods of data collection & analysis,

case study & region, data size & timeframe,
SP/MP/holistic, platform, focus, API access,
software tool, ...
Timeline of articles using traditional methods (only) and studies using digital data sources (with or without trad. methods)
Comparison of popular methods by corpus in a distribution out of 100%
Timeline of common digital data sources (N=198)
Single-platform Twitter studies vs. single-platform Facebook studies timeline
Methodological archetypes
 H1: Following the computational turn, research methodologies draw
predominantly on digital, digitised, and hybrid methods in their data
collection

→ traditional methods were used in a majority of articles in the corpus.

→ less than half of the articles drew exclusively on trad. data sources, but many
studies using social media data also drew on them.

→ more studies drew on traditional methods than did on social media data.
 H2: Empirical digital activism research predominantly focuses on
single-platform Twitter data (H2a) that is relatively easy to access and
monitor through APIs such as hashtags and reply networks (H2b)

→ Twitter, closely followed by FB, was the most researched platform in those
articles drawing on social media data

→ Twitter was increasingly dominant in both single and multi-platform studies

→ SP computational Twitter studies (focussing on hashtags & reply networks)

were the most common methodological archetype
Conclusions...

→ computational methods do not outweigh trad. methods

...but:
→ digital/computational methods are on the rise
→ there is representation bias, particularly towards single-platform Twitter
hashtag studies (popular, Western, fewer API restrictions)

→ digital activism research disproportionately produces knowledge of

particular social groups as well as very specific dynamics of activism
Where do we go from here?
Needed from researchers:
...more transparency in research methodologies
...more contextualised, creative, and situational approaches

Needed from platform providers or policy-makers:

...the opening-up of platforms as research spaces
...a conversation about how access barriers allow platforms to govern research

Obstacles:
...limited funding (tools & access)
...limited research time
...changing landscapes
...platform interests
Thank you.

suaymelisa.ozkula@unitn.it

paul.reilly@glasgow.ac.uk