A Typology of Public Engagement Mechanisms

A Typology of Public Engagement Mechanisms
Author(s): Gene Rowe and Lynn J. Frewer

Source: Science, Technology, & Human Values, Vol. 30, No. 2 (Spring, 2005), pp. 251-290
Published by: Sage Publications, Inc.
Stable URL: http://www.jstor.org/stable/1558037
Accessed: 24/09/2009 10:03
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=sage.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Science,
Technology, & Human Values.
http://www.jstor.org
A Typologyof Public
EngagementMechanisms
Gene Rowe
Instituteof Food Research
Lynn J. Frewer
Universityof Wageningen
Imprecisedefinitionof keytermsin the "publicparticipation"domainhave hinderedthe

conductof good researchand militatedagainst the developmentand implementationof
effectiveparticipationpractices. In this article, we define key concepts in the domain:
public communication,public consultation,andpublicparticipation.Theseconceptsare
differentiatedaccordingto the natureandflow of informationbetweenexercisesponsors
andparticipants.Accordingto such an informationflowperspective,an exercise'seffec-
tivenessmaybe ascertainedby the efficiencywithwhichfull,relevantinformationis elic-
ited from all appropriatesources, transferredto (and processed by) all appropriate
recipients,and combined(whenrequired)to give an aggregate/consensualresponse.Key
variables thatmay theoreticallyaffect effectiveness-and on whichengagementmecha-
nismsdiffer-are identifiedand used to developa typologyof mechanisms.Theresultant
typology reveals four communication, six consultation, and four participation
mechanismclasses. Limitationsto the typologyare discussed,andfutureresearchneeds
identified.
Keywords: public participation; public engagement; participation mechanisms;

typology;mechanismvariables
The Concept and Enactment of Public Participation
In recent times, there has been an internationaltrend towardincreased

involvement of the public in the affairs and decisions of policy-setting
bodies-a concept that is frequentlyreferredto as public participation.In
the United Kingdom, for example, this trendhas become apparentin both
nationaland local governmentin domains as diverse as transportplanning,
the environment,andhealthcare (see, e.g., Robertset al. 1999; Owens 2000;
Martinand Boaz 2000; Bickerstaffand Walker2001). In parallelwith the
increased drive for public participation has come a growing number of
Science, Technology,& HumanValues,Vol. 30 No. 2, Spring2005 251-290

DOI: 10.1177/0162243904271724
? 2005 Sage Publications
251
252 Science, Technology,& HumanValues
processes/techniques/instruments-which shall be collectively termed

mechanisms-for enabling involvement.The very existence of a varietyof
mechanismsimplies uncertainty(at least, at the level of those promotingand
developing the different mechanisms) as to how one should best enact
involvement.Put anotherway, if involvementwere a simple, bounded,and
well-understoodprocess, then one particularmechanism might suffice to
enable it to be effectively achieved (and research would be best directed
towardfindingthis);butinvolvementas widely understood(andimprecisely
defined) can take many forms, in many differentsituations(contexts), with
manydifferenttypes of participants,requirements,and aims (and so on), for
which different mechanisms may be requiredto maximize effectiveness
(howsoeverthis is defined).One importantoutcomeof research,we suggest,
should thus be a theory or model that predicts or describes how to enable
effective involvement(i.e., which mechanismto use, andhow) in any partic-
ular situation.
There are, however,a numberof definitionalissues that need resolution
before researchcan meaningfully,and with any significant chance of suc-
cess, be directedtowardthe developmentof such a theory or model of the
contingent utility of participationmechanisms. Definitions are both the
objectiveof empiricalresearchactivityand a requirementfor such activityto
be effective. Researchis ideally a process throughwhich humans increase
theirunderstandingof the universeand its characteristics;throughresearch,
we seek to define the universe-its objects, forces, activities, and the rela-
tionships among these-with greaterprecision. Ironically,definitions are
also a necessaryforerunnerof research,yet at the startof the researchprocess
we exist in a stateof lack of knowledge. As such, researchinvariablybegins
with loosely defined concepts-essentially, untested assumptions-which
arethenrefined(or refuted)in the light of researchfindingsto be moremean-
ingful, precise, measurable,and so on. The more precise our definitions,the
better(morereliably,validly)we can conductresearch,the easierit is to inter-
pretfindings,andthe greaterthe confidencewe can have in ourconclusions.
For example, to develop a theory of "whatparticipationmechanismis most
effective in enablingpublic participation,in what circumstances,"and to be
ableto test it, one mustpossess definitionsof such importantconceptsaspar-
ticipationmechanism,effective,and circumstances.
In the publicparticipationdomain,unfortunately,the key conceptsarenot
generally well defined, even after several decades (or, some might argue,
centuriesor even longer) of sporadicresearchinterest.Even the concept of
publicparticipationis not well formulated,such thatsome researchersmight
disagree with the scope of activities implicitly or explicitly includedwithin
the concept by others, and synonyms of uncertainequivalence(e.g., public
Rowe,Frewer/ A Typologyof Mechanisms253
involvementandpublic engagement)may be used in place of thatterm.What

is meant by effectiveness of participation is another uncertain issue
(addressedby Rowe and Frewer2004). Similarly,mechanismsfor enacting
the participationconcept (instruments,techniques,methods,tools, etc.) also
tend to be loosely defined. These range from simple surveys to complex
deliberative approaches involving members of the public taking part in
groupsor conferences,which attemptto structurethe debateandprovidebal-
anced informationon the issue (e.g., citizens'juries). Not only does the lack
of cleardefinitionshinderresearchactivitiesinto the effectivenessof the dif-
ferent mechanisms, but also the sheer abundanceof mechanisms-often
highly similarto one another,differingonly in the orderin which a numberof
processes are implemented-creates research problems in the sense of
multiplyingpotentialobjects of research.
In this article,a numberof definitionsof the most importantparticipation
concepts will be forwardedto clarify what public engagemententails and
does not entail, and to clarify how the variousmechanismsare similar and
dissimilar.We suggest thatthis will help reduceconfusionin the domainand
enhance the prospects of conducting high-quality research. The article
begins by defining participation. The bulk of the article then focuses on
defining public participationmechanismsby means of a typology, in which
conceptuallysignificantvariables(variablesthatresearchor theorysuggests
will have potential impact on the appropriatenessof a mechanism,i.e., its
potentialeffectivenessin a given context)will be identifiedandused to asso-
ciate anddissociatethe mainmechanismsinto a smallerset of classes essen-
tially distinguished by structuralcharacteristics.The literatureat present
lacks a thoroughand systematic descriptionof the available mechanisms,
discussionof theirsimilaritiesanddifferences,or discussionof how suchdif-
ferences may affect theircontingentappropriateness(e.g., Webler1999; see
in particularp. 61 for a quote from the US National ResearchCouncil). A
main aim of this articleis to addressthis deficit.
Definition of Public Participation:

Three Concepts of "Engagement"
Before classifying public participationmechanisms, it is necessary to

define the concept thatsuch mechanismsareintendedto enable,thatis, pub-
lic participation.A generaldefinitionof publicparticipationwith which few
would argueis the practiceof involvingmembersof the publicin the agenda-
setting, decision-making, and policy-forming activities of organizations/
institutionsresponsiblefor policy development.This definitionenables the
distinctionof participationsituationsfrom nonparticipationsituationsasso-

ciated with the moretraditionalmodel of governancein which elected policy
makers,generally with the help of nominatedexperts, are left to set policy
without furtherpublic reference.
This definitionof participationis, however, arguablytoo broad,leaving
room for variableinterpretation,because the public may be involved(in pol-
icy formation,etc.) in a numberof differentways or at a numberof levels-as
has been noted by others (e.g., Arnstein 1969; Nelkin and Pollak 1979;
Wiedemannand Femers 1993; Smith, Nell, and Prystupa 1997). In some
cases, the public may "participate"by being the passive recipientsof infor-
mation from the regulatorsor governing bodies concerned;in other cases,
public input may be sought, as in the solicitationof public opinion through
questionnaires;and in still other cases, there may be active participationof
public representativesin the decision-makingprocess itself, such as through
lay representationon an advisorycommittee.Thereareimportantconceptual
differences among these different situationsthat renderit inappropriateto
describe them all using a single term-be that public participation, public
involvement,or whatever.Indeed, one distinctionthathas been made in the
past is between participation and communication(e.g., Rowe and Frewer
2000), the key dimension of differencebeing thatinformationof some sort
flows fromthepublicto the exercise sponsorsin the former,ratherthansolely
from the "sponsors"to the public in the latter.We believe, however,thatthis
distinctiondoesn't sufficientlycapturethe essence of the differencesamong
the variousinvolvementsituationsand that a furtherdivision of concepts is
required.Instead,we propose using threedifferentdescriptorsto differenti-
ate initiativesthat have in the past been referredto as public participation,
based on theflow of informationbetween participantsand sponsors.These
are public communication,public consultation, and public participation.
From here onward,these concepts in combinationare referredto as public
engagement,andthe methodsintendedto enablethis as engagementmecha-
nisms (generically) or engagement initiatives or exercises (specifically).
Mechanismsintendedto enableone of the threeformsof engagementwill be
labeled appropriately,that is, communication,consultation,and participa-
tion mechanisms.The three concepts are defined below and representedin
Figure 1.'
In public communication,informationis conveyed from the sponsorsof
the initiativeto the public. (Here, and throughoutthis article,the termspon-
sor is used to refer to the party commissioning the engagement initiative,
which will usually-but not always-be a governmental or regulatory
agency, althoughrepresentativesof the public may sometimes be the spon-
sors. Ouranalysisis not affectedby the identityof the sponsor,and although
Rowe, Frewer/ A Typology of Mechanisms 255
Flow of Information
Public Communication:
Sponsor Public Representatives
Public Consultation:
Sponsor * Public Representatives
Public Participation:
Sponsor t Public Representatives
Figure1. Thethreetypes of publicengagement.
throughoutwe use phrasingthatmightbe takento assumethatthe sponsoris

a policy-setting organization,this is for the sake of convenience only. The
organizeris takenas the partythatconductsthe engagementexercise, which
may or may not be the same as the sponsor.)Informationflow is one-way:
thereis no involvementof thepublicperse in the sense thatpublicfeedbackis
not requiredor specifically sought. When the public attemptsto provide
information,there are no mechanismsspecified a priorito deal with this at
any level beyond, perhaps,simply recordingthe information.
Inpublic consultation,informationis conveyedfrommembersof the pub-
lic to the sponsorsof the initiative,following a process initiatedby the spon-
sor. Significantly,noformal dialogue exists between individualmembersof
the public and the sponsors. The informationelicited from the public is
believed to representcurrentlyheld opinions on the topic in question.
In public participation, informationis exchanged between membersof
the public and the sponsors.Thatis, thereis some degree of dialogue in the
process thattakesplace (usually in a groupsetting),which may involve rep-
resentativesof bothpartiesin differentproportions(dependingon the mecha-
nism concerned)or, indeed, only representativesof the public who receive
additionalinformationfrom the sponsors prior to responding.Ratherthan
simple, rawopinionsbeing conveyedto the sponsors,the act of dialogue and
negotiation serves to transformopinions in the members of both parties

(sponsorsand public participants).
These threeforms of engagementare sufficientlydifferentboth structur-
ally andin termsof theiraims thatthe mechanismsused to enable themneed
to be evaluatedagainstdifferentcriteriafor effectiveness.
Public Engagement Mechanisms
The numberand varietyof engagementmechanismsare large and grow-

ing. Rosener (1975) listed thirty-ninedifferent "techniques"ranging from
structuredprocedures,suchas "taskforces,""workshops,"and"citizenrefer-
enda,"to broaderconcepts, such as "publicinformationprograms"and"citi-
zen employment."A recent book, called Participation Works!(New Eco-
nomics Foundation,1999), detailstwenty-one"techniques"(andbrieflylists
morethana dozen more)includingrelativelynovel mechanismssuch as "cit-
izen juries"and "actionplanning,"along with othermechanismsthatappear
to be uniquely applied by particularorganizations.Even combining these
two lists does not encompass all the mechanismsthat are presentlyextant.
Figure2 lists, alphabetically,varioustermsfor mechanismsdescribedin the
literatureandprovidesreferencesfor the interestedreaderwho wishes to dis-
cover more. Of the references associated with the different terms, some
merely detail the mechanism, whereas others reportactual case studies or
even experimentalstudies or evaluations.
Thereareseveralimportantpointsthatneedto be madewith regardto Fig-
ure2 andthe informationwithinit, andwith how it shouldbe interpreted.The
first concernsthe comprehensivenessof termsin the figure. Although there
are more than 100 mechanismslisted, the bias is on UK and US types that
appearin the literatureor in technicalreportsthatareknown to us: thereare
undoubtedlymore. In other countries,these particularmechanismsmay be
known by differentnames, or theremay exist still othermechanisms(much
discussion of participationoccurs in the vast "gray"literaturethatexists on
this topic). A second issue concernsthefunctional equivalenceof the terms,
and a thirdconcernstheir independence.Some of the mechanismsare com-
posite processes, some specific techniques, and others tools (that is, not
stand-aloneprocesses for enabling engagement), and as such, some of the
mechanismsmay actuallyincorporateotherseithercompletelyor partly.For
example,a citizenpanel is generallytakento be a standingandrepresentative
sample of a particularpopulation,which may be used to gain public views
when needed. One way the views of the panelmay be attainedis via a survey
(anothermechanismtype). Similarly,a planning cell mechanismmay use a
Act Create Experience Service Users; Shared Drop-InCenter (also Neighbor- * Meeting
(ACE)(1) Interest(2, 10) hood Office,One-Stop/First- (location
ActionPlanning(1, 2) CommunityIndicators(1) Stop Shop) (2, 3) * Meeting
Appraisal(Community, CommunityPlans/Needs EnspiritedEnvisioning(1) tional,'ge
Public),e.g., village/parish/ Analysis(10) * "'FindingHome"("Visualizing - Meeting
environmental.(Also "Monitor- CommunitySite our futureby making England

ing,"e.g., citizenmonitorsand ManagementPlans (1) maps")(1) * Meeting
scrutiny.)(1,2) CommunityStrategic FishbowlPlanning(3) - Negotia
Arbitration (Mediation)(3, 4) Planning(1) Focus Group(3, 6, 10) 25, 26)

Broad-BasedOrganizing(1) CommunityTechnical - FromVisionto Action(1) * Neighbo
Cable Television Assistance (3) FutureSearch (1,2) Council
(Not Interactive)(2) Complaints/Suggestion Game Simulation(3) * Ombud
Cable Television Schemes (10) Guided Visualization(1, 2) * Open D
(Interactive)(3) Computer-Based(IT) Hotline (3) * Open H
Charette(3) Techniques(2, 3) HumanScale DevelopmentIni- * Open Sp
Choice Methods(1,2) Conference (generic term, tiative (1) * Opinion
Citizens'AdvisoryCommittee often withqualifiere.g., "plan- * Initiatives(CitizenInitiatedPeti- * Opinion
(CAC)(3, 5, 6, 7, 8, 9) "visualiza-
ning,""deliberative," tion) (2, 22) * Particip
CitizenAdvocacy(1) tion")(3, 10, 17) Imagine!(1) * Particip
CitizenEmployment(3) Consensus Building(1, 2) * InteractiveWeb-Site (10) ning (1)
CitizenHonoraria(3) Consensus Conference (2, 6, "Issues, Aims, Expectations, * Particip
Citizens'Jury(1, 2, 6, 10, 11, 18, 19, 20) &
Challenges Dialogues in a * Plannin
12, 13) ConsultationDocument Day"(1) * Plannin
Citizens'Panel (Research) (2) (Consultation)(10) * LearningService Team (2) * Plannin
Citizens'Panel (Standing)e.g., ConsultativePanel (2) Local SustainabilityModel(1) * PolicyC
HealthPanel (2, 10, 14, 15) Coordinatoror Coordinator- Maps/Mapping(Village,Parish) * Policy D
CitizenReviewBoard(3) Catalyst(3) (1,2) * PriorityS
CitizenTraining(3) Co-option(CitizenRepresenta- * Media-Based Issue * PriorityS
CommunityDinners(16) tives on Policymaking Balloting(3) * PublicH
CommunityForum- of: Place Bodies) (3, 10) Meeting-CommunitySpon- * PublicIn
(e.g., Neighborhood);Issues; DeliberativeOpinionPoll(2, 21) sored (3) Program
Design-In(3)
Figure 2. Alphabetical listing of "participation" mechanisms (references in parenth

SOURCES:(1) New Economics Foundation(1999); (2) Democracy Network (1998); (3) Rosener (1975); (4) Baugh
(2000); (7) Plumlee, Starling,and Kramer(1985); (8) Hannah and Lewis (1982); (9) Pierce and Doerksen (1976)
Lenaghan (1997); (13) McIver(1998); (14) Dowswell et al. (1997); (15) Kathlene and Martin(1991); (16) Carr an
kj Einsiedel, Jelsoe, and Breck (2001); (19) Guston (1999); (20) Joss (1998); (21) Fishkin and Luskin (1999); (22
-- Coglianese (1997); (26) Susskind and McMahon(1985); (27) Dienel and Renn (1995); (28) Stewart, Dennis, and E
Carroll(1983).
numberof decision-aidingtools, such as a Delphi process, to ascertainthe

views of the participatinggroup, whereas a question-and-answersession is
often an adjunctto apublic meeting.Manycase studiesof participationin the
literaturedetaillengthy anduniqueprocesses thatuse a varietyof techniques
or tools, such as those listed in the table, and attemptto assess the participa-
tion process holistically ratherthan assess the specific parts (e.g., Ouellet,
Durand,andForget 1994; Moore 1996). Such "unique"mechanisms(which
generally have no name per se) have no place in a categorizationscheme,
which worksbest on indivisibleunits,andthisbias is reflectedin the figure.
A fourthpoint-and the most importantfor the purposesof this article-
concernsthe uncertainand contradictorynomenclatureof the mechanisms.
There are two major problems associated with this: first, that dissimilar
mechanismshave in the pastbeen writtenaboutor describedusing the same
term;and second, that essentially similarmechanismshave been described
using differentterms. Both these problemshighlight the necessity of clear
mechanismdefinitionsand an associatedtypology. The firstproblemis best
demonstratedby a numberof examples. Although Crosby and coworkers
developed the citizens' jury at the JeffersonCenterin the United States, in
one of their earliest articles the mechanism was termed a "citizen panel"
(Crosby,Kelly,andSchaefer1986), andin one of the most significantarticles
on public participationsince published,the mechanismwas termeda "citi-
zens' review panel" (Fiorino 1990). Unfortunately,the term citizens'panel
has, at least in the United Kingdom,come to be associatedwith an entirely
differentmechanism-not one thatinvolves a small, select groupof the pub-
lic butone thatinvolves a large,standingselectionof individualsthataregen-
erally polled via surveys. Similar confusion is evident in a report by
Dowswell et al. (1997), which set out to surveythe existence of "healthpan-
els" in the United Kingdom. Because healthpanel wasn't sufficiently well
defined for respondents, a wide variety of mechanisms were detailed in
response to the survey,including standingcitizens' panels, citizens' juries,
and other mechanisms, yet the authorscontinued to refer to the resultant
mechanisms by using the generic "healthpanels" term. Another example
concernsan articleby GundryandHeberlein(1984), in which it was claimed
thatthree"publicmeetings"were the object of study,yet these turnedout to
actuallycompriseone publicmeeting,one set of 50 publicmeetings,anda set
of two workshops. Such confusion is inimical to conductingresearchand
unhelpfulto practitioners.The validity of the nomenclaturein the figure is
difficultto ascertain,and we make no claim regardingthis: the referencesin
the figurearesimply those thatuse the associatedtermto describethe named
mechanism.
The second problem,namely,the uncertainequivalenceof terms,is even

greaterthan the misnaming of mechanisms and leads to furtherconfusion
andtermproliferation.Forexample,is a communityforumthe same as a com-
munitymeeting?Is a citizen advisory board the same as a citizen advisory
committee?Is an opinionpoll the sameas a survey?Is a citizens'jurythe same
as a planning cell? Is a public meetingthe same as a town meeting?In many
cases, authorsmay use termssynonymously,but in othercases, authorsmay
use one termto make a meaningfuland deliberatedistinctionof a particular
mechanismtype from another.Forexample,a publicationby the Democracy
Network (1998) suggests that "surveys"are used to elicit the views of spe-
cific groupsof participantsto general questions,whereas"opinionpolls"are
used to elicit the views of a general group of participantsto specific ques-
tions, implying that these are not the same mechanism (although many
researchersand practitionersmight disagree with this). In many cases, it is
completelyunclearwhatauthorsmeanwhen using a particularterm,because
precise mechanismdefinitionrarelyever takes place in publishedresearch,
and the issue of the generalizability of study results is not adequately
addressed.In general, Figure 2 representssomething of a compromise, in
which some terms are combined when it seems obvious to us that identical
mechanisms are meant by differentauthorsusing synonyms, and separate
mechanismsare noted when authorshave made specific distinctions(even
thoughthe distinctionsmightnot seem sensible to us)-hence, opinionpolls
are noted as differentfrom surveys(DemocracyNetwork 1998), and neigh-
bourhoodmeetingsare distinguishedfrompublic meetings(Rosener 1975).
Other mechanisms may be so similar (conceptually,structurally)as to be
meaningfullyequivalent but have specific nomenclature,largely because
they have been developedby particularteams for practicalpurposesand are
writtenaboutin an uncriticalway by advocates(e.g., consultativepanel and
participatoryappraisal). These often contain similar elements (tools, pro-
cesses), perhapsarrangedin differentorders,but are concise enough to be
considered as stand-alonemechanisms as opposed to unique initiatives.It
will be arguedthat meaningful equivalence of mechanisms is actuallyfar
higherthanwhatat firstappearsto be the case andthatsignificantreductions
in the objects for research may be made through the development of a
typology.
In summary,the intentof Figure 2 is to demonstratethe confusing pleth-
ora of termsused in the public engagementdomain.Detaileddescriptionsof
select mechanismswill be made later.
Given the sheernumberof mechanismsavailablefor engaging the public
andalso the confusion as to whateach does anddoes not entail,andhow each
differs from the others, it is unsurprisingthat no significant theory has

emergedas to whatmechanismto use in whatcircumstanceto enableeffective
engagement.One way in which the problemcan be made more tractable-
not only in a researchsense, but also in a way that will aid the sponsor and
practitioner-is to properlydefine the differentmechanismsand to catego-
rize themaccordingto theircommonsignificantfeatures.One way of catego-
rizing engagementmechanismshas alreadybeen identified,thatis, dividing
them accordingto theirinformationflow into communication,consultation,
andparticipationmechanisms.A moreextensivetypology will be developed
afterpast categorizationefforts have been reviewed.
Categorizing Engagement Mechanisms
In the academic literature,there have been few efforts at developing a

typology of mechanismsper se. A numberof authorshave, however,recog-
nized the multidimensionalnatureof the participationconcept,differentiated
participationsubtypes (generally along a single dimension), and associated
differentmechanismswith each subtype. Such a process could lead to the
developmentof a simple typology of mechanisms,althoughthe authorsof
these articles have generally had other intentions than specifying generic
mechanismtypes. For example, Arnstein(1969) developed a "typology"of
participationby identifyingeight participationformatsthatdifferedaccord-
ing to the degree to which publics are empowered(i.e., differing along the
single dimensionof "empowerment"),andArnsteinillustratedeach by refer-
ence to one or more examples of mechanismsor specific exercises. Webler
(1999) suggestedthata numberof otherauthorshaveproposed"typologies"
based on a similarprincipleof empowerment(these largelyappearin a vari-
ety of handbooksor instructionmanuals,so they will be discussedno further
here; see Webler 1999, 61, for details).
In the case of Glass (1979), five differentparticipationobjectives were
identified(informationexchange,education,supportbuilding,supplemental
decision making, and representationalinput) with which were associated
four"techniquecategories"basedon structuralcharacteristics(unstructured,
structured,activeprocess, andpassive process-one of which corresponded
to two objectives). Here, either"structure"or "objectives"might be consid-
ered the one organizingdimensionthroughwhich Glass categorizeda dozen
"techniques"(e.g., citizen advisory committee, drop-incenter, and citizen
survey).In contrast,Nelkin and Pollak (1979) categorizedpublic participa-
tion accordingto "threedefinitions of the problemof public acceptability"
(lack of confidence,alienation,andinadequateinformation)with which they
associatedcertain"models"(i.e., mechanisms),namely,advisorymodels (of

which they identified four examples, e.g., royal commissions), public con-
sultationmodels (five types, e.g., public inquiriesand referenda),and infor-
mationmodels (five types, e.g., studycircles andenvironmentalimpactstate-
ments).Nelkin andPollakalso consideredfive variablesthatmightbe used to
differentiatethe models (who participates,general intention,who conducts
the procedure,what is the distributionof technical expertise, and is there
really a choice), althoughthey did not applythese variablesin any structured
way. Rosener(1975) also identifieda numberof potentiallyuseful categoriz-
ing variables.Althoughshe set out no typology per se, Rosenerlisted a large
numberof "techniques"and then noted which of fourteenfunctional attrib-
utes they possessed. These functionsinclude"solicitimpactedgroups,""dis-
seminateinformation,""resolveconflict,"and "facilitateadvocacy,"among
others.The functionsarebriefly describedin the article,butthereis no justi-
fication for choosing these as opposed to other functions, and no typology
was developedon thebasisof similaritiesordifferencesamongthemechanisms,
althoughthis would have been possible (the article neverthelessprovides a
useful checklist of mechanismsand their characteristics).
All of these past articleshave some potentialmeritin the sense thatthey
breakup the engagementprobleminto a numberof types-on the basis of
objectives, structure,or function-and associate certain mechanisms with
each. Practitionersmight use these frameworksto identify their particular
engagementproblemandnarrowdown the mechanismsthatsensibly may be
used to address it. Researchers might also benefit, because they have a
reducednumberof mechanismsto compareand contrastwithin a particular
engagement type. Our previous deconstructionof public engagementmay
serve a similarfunction,althoughits intentis ratherto differentiatewhat we
consider to be public participationfrom nonparticipation(i.e., public com-
municationand public consultation).None of the frameworksin these arti-
cles, however,may be called a typology of mechanisms,because the mecha-
nisms alludedto aregenerallyexamplesthatpossess one particularfunction,
structure,or objectiveandthatdifferfromotherexamplesthatpossess differ-
ent functional,structural(andso on) attributes.Examplesgiven in eachbroad
category in each framework still vary on a considerable number of
dimensions.
The role of a typology of mechanismsis to reducethe plethoraof exam-
ples into a lesser numberof classes, within which each mechanism shares
certainkeyattributes,among which each varies.The firstimportantstep is to
identify what these key attributesare-which is difficult,given thatengage-
ment or participationmechanismsvary on a vast numberof attributes:they
involve differentnumbersof participants,takedifferingamountsof time, are
commissioned for different purposes, use different amounts and types of

resources,compose the process in differentways, and so on. It is likely, how-
ever,thatsome of these sourcesof variancewill be more importantthanoth-
ers in termsof affectingthe contingenteffectivenessof mechanismsand that
some sources of variancewill have no significantimpacton effectiveness at
all (or, at least, lesser impactcomparedto other sources). Whatis meantby
importanceis consideredin the next section. Once these sourcesof variance
are identified,they may be used to classify mechanismsin a typology.
The Issue of Effectiveness
Some sourcesof mechanismvariancearemore importantthanothers.By

this, we mean that some mechanismvariablesare more likely to affect the
effectivenessof engagementexercises than others. But what does effective-
ness mean?There are many differentdefinitionsof this concept (see Rowe
andFrewer[2004] for a review). These variousdefinitions,however,essen-
tially allude to two main concepts: the first concerns the fairness of the
mechanism/exercise,and the second concernsthe competence/efficiencyof
the mechanism/exercisein achieving its intendedpurpose-whether that is
educatingthe public, achieving a good consensus, eliciting views, or some
otheraspectof the process or outcome (e.g., Webler1995; Rowe andFrewer
2000).
Related to the concept of fairness are concepts of public acceptability,
equity, democracy,representativeness,transparency,and influence, among
others. This concept concerns the perceptions of those involved in the
engagementexercise and/orthe wider public, and whetherthey believe that
the exercise has been honestly conducted with serious intent to collect the
views of an appropriatesampleof the affectedpopulationandto act on those
views (thisrelatesto publicconsultationandparticipation,as conceptualized
in this article, but not necessarily to public communication).In terms of
devising a typology of mechanisms,it is arguablethatthe fairnessconceptof
effectiveness is irrelevant.Mechanismsand the way in which they are struc-
tured are not intrinsically "fair"or "unfair"-they become so throughthe
intentof those who sponsor,organize,or participatein them, and thence the
way they are enacted.As an example, a citizenjury may be conductedfairly
or unfairly:a poor exercisemightresultif the sponsorbiases the information
fed to the participantsor chooses to ignore the jury output.Even the much-
maligned public meeting cannot be considered intrinsicallyunfair:it may
well be tokenistic,as generallyenacted,but this can be attributedto sponsor
behavioron one hand or, on the other,to the mistakeninterpretationof the

mechanismas of the participation, as opposed to the communication,type.
In these cases, the variancethatexists is not relatedto the general natureof
differentmechanismsbuttheirspecific applications(the varianceis withinas
opposed to among mechanisms).
The conceptof competence/efficiencyessentiallyrefersto the appropriate
elicitation,transfer,and combinationof public and/orsponsorviews. Using
the languageof ourinformationflow model of public engagement,it refersto
maximizingthe relevantinformation(knowledgeand/or opinions)from the
maximumnumberof relevantsources and transferringthis efficientlyto the
appropriatereceivers.Withregardto the threeaspectsof publicengagement,
competence/efficiencymeans the following:
* Publiccommunication: maximizingtherelevantinformation fromthespon-

sorandefficientlytransferringit (withminimalinformation loss)to themaxi-
mumnumberof therelevantpopulation, withtheefficientprocessingof that
information by thereceivers(thepublic/participants)
* Publicconsultation: maximizing therelevantinformation fromthemaximum
numberof therelevantpopulation andefficientlytransferring
it (withminimal
information loss)to thesponsor,withtheefficientprocessingof thatinforma-
tionby thereceivers(thesponsors)
* Publicparticipation: maximizing therelevantinformation fromthemaximum
numberof all relevantsourcesandtransferring it (withminimalinformation
loss)to theotherparties,withtheefficientprocessingof thatinformation by
thereceivers(thesponsorsandparticipants) andthecombiningof it intoan
accuratecomposite
Fromthis perspective,efficiency may be compromisedwhen the informa-

tion from the sources is somehow suboptimal(informationis incomplete,
irrelevant,or simply incorrect),when informationis lost or distortedin the
transferprocess, and when the receiverinappropriatelyprocesses the infor-
mation (by misinterpretationor selective attention).
It should be emphasizedthat other interpretationsof the functionalpur-
poses of "participation"are not necessarily antitheticalto the definitions
above.Forexample,the "education"(Sinclair1977) and"learning"of partic-
ipants (Mayer, de Vries, and Geurts 1995) or sponsors (Guston 1999) are
otherways of statingthatparticipantsor sponsorshave effectivelyprocessed
information,and "obtainingpublic input"(BlahnaandYonts-Shepard1989;
CarrandHalvorsen2001) equatesto eliciting informationfromparticipants.
The concept of representativeness-a common evaluationcriterion(Rowe
andFrewer2004)-is also incorporatedin the definitionsby way of the term
relevantpopulation.
& HumanValues
264 Science,Technology,
Althoughit is truethatmuch of the success or failureof a particularexer-

cise will stem fromhow the particularexerciseis applied,we suggest thatit is
also true that structuralfeatures of the general mechanisms will limit or
enhancethe chances of effectiveness. For example, the presence or absence
of a facilitatorin a group process is a structuralfeature of mechanisms:a
facilitator may aid in eliciting participantknowledge and so potentially
increase the relevantinformationfrom participants,althoughthe effective-
ness of any one exercise may dependon whetherthe facilitatoris skilled and/
or unbiasedin doing thatjob.
In summary,the effectiveness of public engagementwill depend on the
particularmechanism chosen and the way in which this mechanism is
applied(in the specific exercise). Differencesamongmechanismsaredue to
between-mechanismvariables and in the application of mechanisms to
within-mechanismvariables. Not all between-mechanism variables will
have a significantimpacton the effectiveness of the mechanism(in termsof
the competence/efficiencyconcept), and so some mechanisms that appear
differentstructurallymay resultin an equally effective exercise. Identifying
significant between-mechanismvariableswill allow the developmentof a
typology of mechanisms,in which a smallernumberof mechanismclasses is
distilledfromthe unmanageable(in botha researchandpracticesense) pleth-
oraof mechanisms.These classes will comprisemechanismsthatdo not vary
on significantsources of variance,althoughthey may do so on insignificant
ones. In the next section, we identify a numberof mechanismvariablesthat
are liable to affect engagement exercise effectiveness according to our
informationflow model.
Between-Mechanism Variables
The significantbetween-mechanismvariablesare discussed with regard

to theirpotentialimpacton the differentcomponentsof the informationflow
model of engagement(i.e., maximizing:participants,informationelicitation,
informationtransfer,informationprocessing,and informationaggregation).
Table1 gives a summaryof the significantbetween-mechanismvariablesand
indicateshow they might affect effectiveness.
Variables Associated with

Maximizing Relevant Participants
Therearethreefigures thatneed to be consideredwith regardto the issue
of maximizingparticipants,which arerelevantto all communication,consul-
Table 1. Summary of Key Mechanism Variables
MechanismVariable Levels of Variable Aspect of Effectiveness Pote
Participant Controlled Maximizerelevantparticipants

selection Uncontrolled
method
Facilitation
of Yes Maximizerelevantinformationf
information No
elicitation
Response mode Unlimited/open Maximizerelevantinformationf
Limited/closed
Information
input Set information Maximizerelevantinformationf
Flexibleinformation
Mediumof Face-to-face Maximizetransferand processi
information Non face-to-face
transfer
Facilitation
of Structuredcombination Aggregationof participantinfor
aggregation Unstructuredcombination
tation,and participationexercises. The first "figure"dependsentirelyon the

contextof the exercise (i.e., not on the way a particularmechanismis struc-
turedor an exercise is run), and this is the population of interested/affected
individuals.As an example,a governmentalpolicy shouldin theorybe of sig-
nificance to the entirenationalpopulation,and a local governmentinitiative
should be of interestto the populationliving in thatregion. In practice,this
numberis often difficultto ascertain,becausepolicies may haveimpactsout-
side of theirgeographicalor demographicboundaries(considera dambeing
built in one countryon a river that flows throughanother)or, indeed, may
practicallybe of limitedinterestwithinthe populationin question(for exam-
ple, a local transportplan might only be of significance to motoristsrather
than to the whole local population).Theoretically,a particularengagement
exercise shouldattemptto communicateinformationto, or elicit information
from, all membersof this population.Of course, this is an idealized stateof
affairs: in practice, any engagement initiative aims to communicatewith/
elicit informationfrom a smallersample of the population.This population
numberis importantas a benchmarkto calculatetwo othernumbersthatare
relatedto the structureand administrationof any particularexercise rather
than the context.
The intendedsample size is the second figure of significanceto engage-
ment initiatives,and this is the numberactuallyapproachedduringthe exer-
cise (althoughnot necessarilythe actualnumberof participants-as will be
discussed shortly). The value of this figure will be at least in part due to
choice of mechanism,consequenton structuralcomponents,as well as due
to the enactmentof the mechanism.Withregardto maximizingthe relevant
population,the actualnumberof the intendedsample is less importantthan
the numberrelativeto the population figure:for example, a surveythatwas
sent to 100 people out of an interestedpopulationof 1,000 would potentially
be able to access 10 percentof thepopulation information,whereasone sent
to 1,000 people out of an interestedpopulation of 1 million would have
potential access to a much lower proportion.The former survey (all else
being equal) would be betterin the sense thatit accesses a higherpercentage
of the relevantpopulationinformation.A surveyis an exampleof a consulta-
tion mechanism,althoughthis principleis relevantto the otherengagement
types. Consider, for example, the use of a newsletter (a communication
mechanism): different exercises will communicate with different sample
sizes, andthose thatattemptto access moreof the relevantpopulationwill be
betterfrom an informationmodel perspective(and, indeed, it is difficult to
imagine any occasion when accessing less informationcould be interpreted
as "better").
There is a thirdfigure, however, that is perhapsof even greatersignifi-

cance for the effectiveness of exercises thanthe "intendedsample size,"and
this is theproportionof the sample that is actively engaged. In termsof con-
sultationandparticipationmechanisms,this representsthe numberof people
in the sample (those who have potentialaccess to the informationbeing dis-
pensed, or those who might potentiallyrespondto the informationrequest)
who process informationor respond,respectively.Generally,this will vary
across the administrationof exercises of any particularmechanism. For
example, consider two surveys, each of which is sent to 1,000 people: one
might obtain a response rate of 20 percent,and one 30 percent.The differ-
ences in response rate will have furtherimplicationsfor the representative-
ness of the sampleand,hence, the amountof information(of the whole thatis
pertinentto the issue) that is successfully elicited. With regardto group-
based mechanisms (mainly participationmechanisms), although it might
appear thatthe numberof activeparticipantsis the same as the sampleof the
population(e.g., ten people selected for a group, all being "active"),this is
not the case. Not only might memberswithdraw(from a committee,panel,
etc.), but also those in attendancemay not be "active,"such as when an indi-
vidual participantin a citizen jury remainsquiet and does not contributeto
discussion. This is particularlylikely in large groups and when time is lim-
ited. A large numberof aspects relatedto the conductof the particularexer-
cise (within-mechanismvariables),from the comprehensibilityof informa-
tion to trustin the sponsorsand to how groups are run, may affect this third
figure. In general, sponsors of all engagement exercises should consider
these numbersand attemptto maximize the size of the sample, andthe num-
ber of active participants,up to the populationnumber.Doing so maximizes
the amountof potentiallyrelevantinformationthat might be distributedor
attained.
Intendedsamplesize providesan interestingproblemwith regardto clas-
sifying engagementmechanisms.Although some mechanismsimplicitly or
explicitly stipulateparticipantnumberswithin certain narrowlimits (e.g.,
consensusconferencesandothergroup-basedmechanisms),most do not. Of
those thatdo not, some tendto stipulateprecisebutvariablenumbersin each
exercise (e.g., surveys and newsletters),but others have high variabilityin
respondentswith absolutelyno controlover numbersinvolved(e.g., hotlines
and Internet-basedconsultations).Generally,when numbersare stipulated,
they tend to be low (numberof people who can fit into a room, for example),
butit is conceivablethatlow numbersmightbe attainedby using any mecha-
nism, and,hence, intendedsamplesize is not a relevantbetween-mechanism
variable, although it is a highly important within-mechanism variable
(because,all else being equal, a mechanismthatinvolves "high"numbershas

access to more populationinformationthan one with "low"numbersand is
betterfrom an informationmodel perspective).Althoughit may make some
practicalsense to distinguishmechanismsaccordingto how precisely they
stipulatethe numbersinvolved (i.e., "precise"when defined by the mecha-
nism, "loose"when not definedby the mechanismbutdefinedin each exam-
ple exercise, and"veryloose" when neitherdefinedby the mechanismnorin
any example exercise), it is difficult to see what difference classification
accordingto this variablewill makewith regardto the efficiency of informa-
tion flow.
Thereis, however,one between-mechanismvariableof significanceto the
issue of maximizingrelevantparticipants,andthis is theparticipantselection
method. Mechanisms can be roughly divided into those that involve some
degreeof controlof participantselection (usuallyby the sponsorsor organiz-
ers, by targetingcommunicationsat, or attemptingto elicit informationfrom,
a certainsampleof the population)andthose thathaveno control,relinquish-
ing choice of involvementto the publicparticipantsthemselves.Examplesof
the former include publicity via newsletters (communication),referenda
(consultation), and citizen juries (participation);examples of the latter
include drop-incenters (communication),computer/Internet-based consul-
tations(consultation),andtownmeetings(participation).In controlledselec-
tion, both the numberand relevanceof those engagedmay be determined(in
theory),whereasin uncontrolledselection,this is not the case, andeven if the
actively engaged are higher in number,many of these may be inappropriate
(the samplemay be biased).As such,controlledselectionmay be morelikely
to maximize the relevantpopulationinvolved than uncontrolledselection,
andthereforethis would seem an appropriatevariablefor use in a typology of
mechanisms.
Variables Associated with Maximizing

Relevant Information from Public Participants
Each active participant in an engagementexercise can be consideredto

possess a quantityof relevant informationregardingthe problem in hand
(whetherthis is knowledgeor simplyan opinion)as well as otherinformation
of no relevance.An effective exercise needs to elicit all relevantinformation
from each active participantwhile not eliciting irrelevantor spuriousinfor-
mation.Should appropriateinformationremainunelicitedor be confounded
or confused by irrelevant information, effectiveness will be negatively
affected.2In this section, we considerelicitationof informationfrom public
participantsonly.
There are at least two structuralaspects of engagementmechanismsthat

areliable to affectthe likelihoodof maximizingrelevantinformationelicited
from public participantsin consultationand participationmechanisms(not
communication).The presenceor absenceof adaptivefacilitationis the first.
Thatis, some mechanismsspecify thatthereis a facilitatorpresentwho plays
a role in managingthe elicitationprocess and gaining inputfrom all (in this
sense, the facilitatormay also play a role in maximizingparticipantnumbers
throughensuringthat all participantsare active). This is a particularfeature
of a number of group-basedmechanisms (e.g., focus groups and citizen
juries), althoughnot all (e.g., co-option and public meetings), althoughit is
rarein mechanismsthatseek individualresponse(a one-to-oneconsultation,
as froman interviewprocess,could, however,be deemedof this type). Active
facilitationhas been shown to increase relevantinformationelicited when
compared to some identical process without facilitation (e.g., Offner,
Kramer,and Winter 1996; Anson, Bostrom, and Wynne 1995). One way in
which it appearsto work is to countera common trendin groupsthatresults
in the adoptionof a "satisficing"strategyin which a groupsettles for the first
decision that proves satisfactory(e.g., with which no one greatly objects),
ratherthan adoptingan "optimizing"strategyin which the groupgoes on to
considerbetteralternatives(Rowe 1992). In essence, the membersof such a
grouphave within them furtherinformationthatthey could bringto bearon
the problem,butthey do not. Facilitationcan theoreticallyaid furtherconsid-
eration, and thereforethis seems an apt variableto use in developing the
typology of mechanisms,althoughaspects to do with the qualityof facilita-
tion will also affect informationelicitationeffectiveness.
A second mechanismaspectthatis liable to affect informationelicitation
in consultationand participationexercises is the responsemode available,in
particular,whether it is "open" or "closed."Mechanisms that only allow
respondentsto choose amongtwo or moreoptions(e.g., referendaor a survey
requiringratingson a scale or set questions)are "closed,"whereasthose that
allow free responses (e.g., focus groups and conferences) are "open."It is
reasonableto theorizethat"open"mechanismsaremorelikely to elicit more
of the relevantinformationfromparticipantsthanclosed ones (afterall, using
open questionsin social science researchis predicatedon the assumptionthat
they will yield richerdatathanwill closed questions)-although, in practice,
they might also elicit more irrelevantinformation.Regardinginformation
loss, consider a referendum in which participantsare limited to yes/no
answersor choices amongspecifiedoptions.Here,the sponsorwill not know
if those who say "yes"(for example) all have the same reasonsfor doing so,
whetherthose who say "no"do so for reasonsthatarein some sense more or
less importantand should be given greateror lesser weight, and so on. One
way in which participantsmightrespondto exercisesthatoverlyrestricttheir

chance to give full informationis to withdrawfrom the process-by not
respondingin the firstplace or by respondingin a nonusefulway, as in spoil-
ing a ballot paperin a referendum(i.e., becoming a nonactiveparticipant).
The maindangerin such cases, however,is thatthe sponsormay derivea mis-
takenbelief thatit has canvassedthe appropriateviews, when this is not the
case. Again, there are various subtleties in terms of the exact nature of
response mode used that may affect informationelicitation (e.g., whether
responses are numerical or nonnumerical),but these tend to vary across
examples of any particularmechanism.
In summary, active participantsonly representpotential information
sources: they need to be engaged in such a manner that comprehensive,
appropriateinformation(and not incomplete or irrelevantinformation)is
elicited from them, and there are several variablesrelated to engagement
mechanismstructurethat may affect this.

Relevant Information from Sponsors
Informationsources may include the sponsorsand theirexperts(in com-

municationandparticipationexercises), as well as the publicrepresentatives
(in consultationand participationexercises). Indeed,accordingto the infor-
mationflow model, it is just as importantfor sponsorinformationto be full
andrelevantas for thatfrompublicparticipants.The sponsorsresponsiblefor
initiatingengagementexercises invariablyassume thatany informationpro-
vided by them is relevant,comprehensive,and appropriatefor public under-
standinganddecision making.Whetherthisis the case is difficultto ascertain
in all communicationcases: the informationis often set prior to the initiative
and therefore,at least practically,outside of the considerationof it. Some
communicationmechanismsand most participationmechanisms,however,
do speakto this issue, in the sense thattheirstructureallows for flexible, vari-
able, and responsiveinformationprovisionfrom sponsors(i.e., information
elicitationby the public). It may be hypothesizedthatmechanismsof the lat-
ter type may, in general(and all else being equal), be more likely to resultin
maximizedrelevantsponsorinformationthan set-informationmechanisms,
in the sense that they enable the public participantsto identify holes in the
informationand to clarify uncertainties(e.g., when the informationis full of
technicaljargon). A distinctionis thereforemade between communication
mechanismswith set informationinput and those with flexible information
input (from sponsors). Examples of the set type include newsletters and
leaflets, and examples of the flexible type include telephone hotlines and
public meetings. Participationmechanisms,however,invariablyallow flexi-
ble informationinput,because dialogue and interactionwould be difficultif
one of the partiesinvolved could only readfrom a set informationsheet. As
such, this variable (flexible or set input) is of less use for distinguishing
differentparticipation mechanism types because most (if not all) are of
flexible type.

the Effective Transfer of Information to,
and Its Processing by, Recipients
The aim of engagementis to acquireall relevantinformationfromall rele-

vantmembersof the population(sources)andtransferthis to relevantrecipi-
ents (be these the sponsorsor the participants).A markof the efficiency of
transferis whetherthe recipientsfully understandthatinformation(i.e., pro-
cess it). The most significantvariablein this respectis the mediumof infor-
mation transfer.In communicationmechanisms,for example, information
may be deliveredoverthe phone (hotline/helpline),via computertechnology
(e.g., teleconferencing),or face-to-face (e.g., at informationcenters). Each
mediumhas its own attributes,advantages,anddisadvantages.In the firsttwo
cases, lack of physicalcontactremovesvisual, nonverbalcues thatmakeup a
large partof humancommunication(see Sproulland Kiesler [1986] for dis-
cussion) and might lead either the communicatoror recipientto misunder-
stand information, so diminishing the relevant information transfer.As
anotherexample,informationdistributedby mail may be treatedasjunk mail
andmightnot be read(inadequatetransferhereactingto reduceactivepartic-
ipants).People may be morelikely to watchor listen to a broadcastthanread
leaflets and brochures-although the informationthat could be communi-
catedis likely to be less (puttingat risk the maximaltransferof information).
Medium of information transfer-both from sponsor to public and vice
versa-is also pertinentwith regardsto consultationandparticipationmech-
anisms, and might likewise reduce respondents/recipientsand the informa-
tion elicited fromthem.Forexample,use of cable TV,the Internet,or the tele-
phone, may disenfranchisethose who do not possess these media(Rowe and
Gammack2004). Therearea numberof aspectsof questionnaireadministra-
tion that are believed to help in obtaining increased responses-that will
apply in varying forms to the differentelicitation media. For example, if a
questionnaireis sent by mail, then returnpostage on a supplied envelope
shouldbe prepaid;or in the case of televoting,the respondentshouldbe able
to dial a "freephone"number.Response rates might also be enhanced by

offering incentives such as money or shopping vouchers-although one
shouldbewarethatthis might bias the sample.Offersto providefeedbackor
details of findings might also enhanceresponserates.Chipmanet al. (1996)
have consideredwhat is the best mediumfor transmittinginformationto a
"target"audience; Rowe and Gammack (2004) reviewed evidence that
suggests there are qualitativedifferences in the way that people respondto
face-to-face and electronicallymediatedinformation.
Because face-to-face transferof informationseems less likely to lead to
informationloss or misrepresentation(SproullandKiesler 1986), it mightbe
hypothesizedthat,all else being equal, an engagementmechanismentailing
this type of process is likely to lead to a moreeffective exercisethana similar
mechanismthatdoes not. The mediumof informationtransfermay thus be
defined accordingto whethera mechanismspecifies face-to-facetransferof
information.Althoughotheraspectsof the transfermedium,such as whether
informationis presentedgraphicallyor textually,are likely to affect effec-
tiveness, these tend not to be definitivelyspecified by mechanisms(i.e., dif-
ferentmedia may be used within one mechanismor in differentexamplesof
one mechanism)and are consideredsourcesof within-mechanismvariance.
Within a particularmechanismusing a particularmedium, variabilitywill
also occur in the operationof the exercise. Consider,for example, a local
newspaper:the positioningof the informationwithinit could affect its likeli-
hood of being seen by the public, as could the paper'scirculation(different
papers have different levels of circulation within different communities).
Likewise, informationcenters might have variable success at information
communicationdepending on their physical location and ease of access to
the population.In both cases, poor administrationmight reducethe number
of active participants.
Othervariablesrelatedto the transferandprocessingof informationseem
to be largely within-mechanismones. One example is comprehensibility.
Thatis, it is importantthatrecipientsfullyunderstandall of the information
theyreceive.This applieslargelyto communicationandparticipationmecha-
nisms, althoughit includesthe understandingof questionsandtasksrequired
in consultationmechanisms.(Whetherrecipientsagree with informationis
anothermatter.We do not hold thatcommunicationshould be evaluatedby
its ability to convince-because the best, most full, and most compelling
informationin the world may still fail to convince those with entrenched
beliefs-but it is simply to inform. Arguably,if the informationis full and
comprehensive,then there is a greaterchance that recipientsmight be per-
suaded by it.) Aspects to do with the presentationof informationand its
wordingmight affect this. Forexample, technicallycomprehensiveand cor-

rect informationthatis writtenusing difficulttermsandjargonis unlikely to
be fully understoodby recipients.In this case, even if the informationis cor-
rect and the entirerelevantpopulationreceives it, informationis effectively
lost in the transferprocess, and the exercise is suboptimal.

the Aggregation of Relevant Information
In participationexercises and, to a lesser extent, consultationexercises,

the problemarisesas to how to mergethe variousparticipants'knowledgeor
opinionsinto some compositeresponsethataccuratelycombinesall relevant
informationfrom those participants.Clearly,inefficiency here can severely
harmthe effectiveness of an exercise, even if all relevantparticipantshave
been involved and all relevantinformationhas been elicited from these and
then transferred(and been processed by) the intendedrecipients.One vari-
able of relevancehere is whetherthereisfacilitation of the aggregationpro-
cess. This bears some similarityto facilitationof informationelicitation,but
it is not identical.An actualfacilitatormay help elicit informationfromgroup
participantsas well as help to combinethatinformationeffectively;however,
facilitatorsarerarein individual-basedmechanisms,althoughfacilitation of
the aggregationprocess is usual in these, as will be explained shortly.Fur-
thermore,the absence of facilitation of informationelicitation in a group
does not meanthataggregationitself cannotbe facilitated(a case herewould
be a nonfacilitatedmeeting, followed by a vote by all involved).
When values are elicited from participantsbehaving as individuals,it is
possible and usual (for organizersand sponsors) to combine these in some
equitablemannerthattakesinto accountall inputs.Forexample,responsesto
a surveywill usually be aggregatedto revealwhatproportionof participants
holds certainviews (this is true of qualitativedata as well as quantitative).
The aggregationprocess is structuredfollowing certainrules (even if certain
dataarediscarded,thereis generallya needtojustify this). On the otherhand,
when values areelicited fromgroups,the outputitself representsan aggrega-
tion performedwithinandby the group.It is unstructuredin the sense thatno
clearrulesareset out andfollowed, andequity,or inputfromallparticipants,
is not guaranteed. Indeed, various difficulties have been documented
regardinggroup behavior and inequity of influence in terms of dogmatic
individuals dominatingproceedings over less confident individuals,group
polarizationof response,and so on (e.g., Rowe 1992). It is likely thatgroup-
based output does not reflect the group opinion with complete accuracy,
althoughgroup solutions have been found to generallybe betterthanaggre-

gated individualjudgmentsin termsof qualityofjudgment(e.g., Hill 1982).
Facilitatedgroup aggregationcan take place, however, throughthe use of
proceduressuch as the Delphi technique(for reviews, see Rowe 1992, 1998),
and these have been shown to provide even betterjudgment than regular
groups(Rowe 1998) as well as enablean accurateaccountingof the opinions
of the whole group.The context of the exercise-whether it is seeking accu-
ratejudgmentsor a fair representationof opinions-seems importantin the
choice of appropriatemechanism.
The particularmethodsused to controlinformationaggregationvary not
only among different classes of participationmechanism but also among
examples of a particularmechanism. For example, a variety of methods
areused in publicconferences(for elicitationas well as aggregation),includ-
ing brainstorming,causal impact diagrams,ranking,timelines, community
maps, Venndiagrams,and policy Delphis (e.g., DemocracyNetwork 1998).
Such aspectsneed to be consideredfor the best enactmentof public engage-
ment mechanisms.
The Typology
Thereareseveralways in which a typology mightbe developed,thatis, of

using the significantbetween-mechanismvariablesto classify mechanisms.
One way is to list all of the engagementmechanismsandindicatewhatvalue
they takefor each variable.(This is essentiallywhatRosener[1975] did. Her
variables,however,were notjustified in any way. Furthermore,she went no
furtherthanidentifyingcharacteristicsof the differentmechanisms-it was
not her intentto note similaritiesanddifferencesamongthese, andto use this
informationas the basis for developing a typology.) The mechanismslisted
in Figure 2 could be used for this purpose. As previously noted, however,
there is some uncertaintyabout what many of these entail structurally,
because few have been writtenaboutextensively in the academicliterature.
In the absenceof cleardefinitionsof theirnecessarystructuresandprocesses
in definitive sources, we do not feel confidentin classifying the majorityof
them (thatmight be left to those who know thembetter).Furthermore,some
may be betterregardedas concepts as opposed to actualmechanisms(e.g.,
arbitration,mediation,citizen training,citizen honoraria,andco-option onto
committees) or as specific tools or processes thatmay form partof broader,
defined mechanisms(e.g., communitytechnicalassistance,mapping,policy
Delphi, question-and-answersessions, and workshops; also, arbitration,
mediation,and co-option again). Nevertheless,a limited numberof mecha-
nisms could be used for the purposeof developinga typology, althoughat the
risk of missing some classes of extant mechanisms.
A second strategyinvolves settingout the totalnumberof potentialclasses
of mechanismsby establishingall possible variations,namingthese classes,
and then considering which of the currentmechanismsfit into each class.
This allows the interestingpossibility of identifying hypotheticalclasses in
which thereareno presentmechanisms.The mainproblemwith this strategy,
however,is thatthe numberof classes is potentiallyvery large,andone of the
key aims in developinga typology is to reducethe domain'sobjectsof study
ratherthan increase them. For example, if there were only four significant
between-mechanismvariables,and each of these had two forms, then the
total numberof classes/combinationsis two to the power of four,or sixteen;
if five, then this is thirty-two,and so on. Of these hypotheticalclasses, it is
also possible that a large numberwill be practicallydifficult or insensible,
and this may be the underlying reason for the absence of any existing
examples.
In Table2, we thereforeuse the practicalmethodof the first strategy:we
detail a numberof the most formalizedof the engagementmechanismsfrom
Figure 2, describingthem accordingto their similaritiesand differenceson
the main between-mechanismvariables.In a numberof cases, the mecha-
nisms in Figure2 aremergedorbrokendown,eitherwhen it appearsthatthey
representsynonymousconcepts (e.g., opinion polls/surveysor the different
types of meetings andreferenda)or when a label appearsto representdiffer-
ent mechanisms (e.g., publicity-which includes newsletters and exhibi-
tions), respectively.Mechanisms with high variabilityin structureare not
included, such as workshops (Lundgrenand McMakin [1998] suggested
these may either have select members or open invitation, and may either
be nonfacilitatedor facilitated)3and citizen advisory committees (which
Rosener [1975] describedas a generic term denoting several techniques).
The informationin Table2 may now be used to identifyclasses of mecha-
nisms. In Table3, the mechanismsthatshareidenticalfeaturesin termsof the
between-mechanismvariablesaregroupedtogetherand described.For sim-
plicity,the tablesimply labelsthe differentclasses as type 1, type2, andso on.
Anotherapproachwould be to generatenames that encapsulatethe signifi-
cant structuralfeaturesof each class or to name the classes after the most
notablepopularmechanismwithin it. Because Table3 gives fairly complete
descriptionsof the mechanismclasses, we will only allude to a numberof
general issues and trendsin the text.
Table3 identifiesfour classes of communicationmechanisms,six classes
of consultationmechanisms,and four classes of participationmechanisms.
Many of the traditionalcommunicationapproachesare designated type 1,
(text continues on page 283)
Table 2. Key Engagement Mechanisms Classified According to Stru
Vari
Selection
Method: Elicitation Response
Controlled- Facilitation: Mode:
EngagementType Mechanisms Uncontrolled Yes-No Open-Closed
Communication Cable TV (not Uncontrolled NA NA
interactive)
Drop-incenters Uncontrolled NA NA
(open-house,
drop-incenter,
one-stop shop,
first-stopshop,
exhibitions)
Hotline Uncontrolled NA NA
Information broadcasts Controlled NA NA
via
("publicity"
TV,newsletters,
and/orradio)
Internetinformation Uncontrolled NA NA
("computer-based")
Publichearings/ Uncontrolled NA NA
inquiries
Publicmeeting (with Uncontrolled NA NA
question-and-answer
session)
Consultation Citizens'panel- Controlled Yes Open
group-based (e.g.,
health panel)
Consultation Controlled No Open
document
Electronicconsultation Uncontrolled No Open
("interactiveWeb site")
Focus group Controlled Yes Open
Open space Uncontrolled Yes Open
Opinionpoll Controlled No Closed
Referendum Controlled No Closed
(varioustypes)
Study circle Uncontrolled Yes Open
Survey Controlled No Closed
Telepolling/ Controlled No Closed
Televoting
Participation Actionplanning Controlled Yes Open
workshop
Citizens'jury Controlled Yes Open
Consensus conference Controlled Yes Open
Deliberativeopinionpoll Controlled Yes Open
Negotiated rulemaking Controlled No Open
Planningcell Controlled Yes Open
Taskforce Controlled No Open
Townmeeting (New Uncontrolled No Open
Englandmodel)-
withvotina
NOTE:FTF= face-to-face.
fromthe sponsors, even ifonlyinstructionsas to ho
a. Thereis alwaysa degree of information
< as NAbecause the informationis set, specific, and minimal.
3 Table 3. Types of Engagement Mechanisms
Mechanism
Classes Examples Characteristics
Communication Information Controlled These are traditionalcomm
type 1 broadcasts selection publicinformationprogra
(traditional ("publicity"via Set information geted withset informatio
publicity) television, Non-FTF ularlyused by councils in
newspaper, tax is spent.
and/orradio)
Communication Publichearings Uncontrolled These mechanisms relyon
type 2 Publicmeetings selection vice versa. As such, the in
(withquestions Flexible in terms of those most pr
and answers) Information cated face-to-face by spo
FTF ing to some degree (often
are often requiredwhen s
implementedor priorto th
initiatedby a local author
Communication Drop-incenters Uncontrolled These mechanisms relyon
type 3 CableTV (not selection centers (frequentin most
interactive) Set information butionpointsat which citi
Internet Non-FTF or look at displays or exhi
information modernmethods supply i
a Web site) or cable TV (e
informationis set in that t
available,althoughit is va
Althoughthere may be FT
be representativesof dec
informationratherthan sig
Communication Hotline Uncontrolled As withtype 3 mechanisms
type 4 selection ible, however,and suppli
Flexible not providedFTFbutvia
information line allows citizens to pho
Non-FTF ceive either a directansw
Consultation Opinionpoll Controlled These mechanisms are ess

type 1 Referendum selection swers to specific questio
Survey No facilitated importantthan quality(th
Telepolling/voting elicitation sponses are closed/limit
Closed response sources of within-mecha
mode pact of elicited responses
Non-FTF and advisory referenda,o
Structured mission (i.e., whetherpos
aggregation
Consultation Consultation Controlled This class aims to attainop
type 2 document selection mechanism is the consul
No facilitated potentiallyinterested peo
elicitation otherorganizations)with
Open response mode tentially,nonselected oth
Non-FTF if they are outside of the
Unstructured deliberatelyaim forwider
aggregation
Table 3 (continued)
Mechanism
Consultation Electronic Uncontrolled As type 2, but withuncontr

type 3 consultation selection have intranetsites invitin
(interactive No facilitated cal issues or service mat
Web site) elicitation
Open response mode
Non-FTF
Unstructured
aggregation
Consultation Focus group Controlledselection This type of consultationem
type 4 Facilitatedelicitation witheffortexpended to fa
Open response mode tion.Itis typifiedby the fo
FTF dozen people facilitatedi
Unstructured no significantsponsor inf
aggregation ratherthan participationm
Consultation Studycircle Uncontrolled This type is similarto type 4
type 5 Open space selection (participantsself-selected
Facilitatedelicitation Sweden and the US). In t
Open response mode issue or study a series of
FTF sions witha volunteerfac
Unstructured laiddown forthe conduc
aggregation semblies of self-selected
cussed in smallerworksh
plenarysession.
Consultation Citizenpanel- Controlledselection The main example of this ty
type 6 groupbased Facilitatedelicitation panel).This is characteri
(e.g., health Open response mode who meet in a facilitated
panel) FTF may meet several times a
Structuredaggrega- be traced throughouttim
tion end of meetings, opinion
secret ballot.Consultatio
differentmechanismclas
Participation Actionplanning Controlledselection The mechanisms of this typ

type 1 workshop Facilitatedelicitation participants,facilitatedgr
Citizens'jury Open response mode responses, and flexiblein
Consensus Flexibleinformation formof "experts"who are
conference FTF pants throughouta numb
Unstructured such and may depend on
aggregation maticindividuals,and so
Participation Negotiatedrule Controlledselection This class of mechanisms i
type 2 making No facilitated ence that there is no faci
Taskforce elicitation manyways, they are simp
Open response mode of inputfromgroupmemb
Flexibleinformation noted here use small gro
FTF readyaccess to all pertin
Unstructured
aggregation
00
Fo
00
)0
Table 3 (continued)
Mechanism
Participation Deliberative Controlledselection This class is also similarto

type 3 opinionpoll Facilitatedelicitation turedaggregationtakes p
Planningcell Open response mode the selected participants
Flexibleinformation the issue (and questionin
FTF gregationof all participa
Structured cells (a Germanmechani
aggregation ensure structuredconside
of opinions.
Participation Townmeeting Uncontrolled This mechanism class diffe
type 4 (New England selection Importantly,selection is u
model)-with No facilitated mationelicitation,althoug
voting elicitation exam is the town meeting
Open response mode tion)takes place after deb
Flexibleinformation
FTF
Structured
aggregation
NOTE:FTF= face-to-face.
a. Doorstepor street surveys are FTFand mightbe considered a separate mechanism.
b.The ostensible aimof specific mechanisms may be to gain consensus (e.g., consensus co
althoughformalaggregationmechanisms are not specified.
c. This is not to say thatmeetings are not chaired and thatthere is no controlat all of the gr
emphasis on solvinga problemratherthan gaining fairconsiderationof all views.
d. The presence of public input differentiatesthis participationmechanism from the no
mechanism.
which is distinguishedfromthe othertypes by involvingthe controlledselec-

tion of participantsratherthanallowing participantsto select themselves.By
distributingset informationusing mass media approaches(i.e., non-face-to-
face), they have the potential to reach many people with a standardmes-
sage-and therebyseem to have advantagesin termsof maximizingquantity
of informationto the intendedcommunity.The other types, relying on the
self-selection of participantsand also flexible information(types 2 and 4),
would seem less optimal from the perspectiveof maximizing information
distribution(unless one considers that the sponsors may themselves be
unclearas to who arethe appropriateparticipantsandwhatis the most impor-
tantinformationto impart).Effectivenessaccordingto the informationflow
model does notjust dependon the quantityof informationandnumberof par-
ticipants, however, but also on whether the participantsunderstandand
correctly process the informationthey receive. From this perspective, the
opportunityof uncertainindividuals to gather informationthey feel they
require might counter the disadvantagesthat would seem apparentin the
more flexible mechanismtypes. Clearly,then, the differentclasses may be
differentiallyappropriate,and empiricalresearchis needed to establish the
differentcontexts in which this is so.
Type 1 consultationmechanisms(e.g., opinion polls and referenda)bear
manysimilaritiesto type 1 communicationmechanisms:they aretypifiedby
the controlledselection of participantsand use of mass media (non-face-to-
face) to ascertainspecific informationfrom as many of the relevantpopula-
tion as practicallypossible. Because responses are elicited in a set format,
they are able to aggregate information to ensure maximum input from
respondents.Their naturedoes, however,ensure that little effort is made to
facilitate the informationfrom participantsby checking individualunder-
standingon the issue on which the consultationis taking place. The other
mechanismclasses, to a greateror lesser extent,limitthe controloverthe pro-
cess by allowing participantsto select themselves (e.g., types 3 and 5); by
allowing flexible, open responsesfrom participants(all otherclasses); or by
omittingany standardizedprocedurefor aggregatingparticipantinformation
(types 2, 3, 4, and 5). Whereextracontrolis exercisedin these othermecha-
nism classes it is in the elicitation of informationusing a more intensive
mediumof informationcollection (i.e., face-to-face:types 4, 5, and6) andin
facilitatingthe process of eliciting knowledge from participants(types 4, 5,
and 6). A distinction among mechanisms that appearsto correlatewith a
numberof these differencesis whetherthey are individual-or group-based
(thereis more emphasison controllingthe process and gaining information
quantityin the former,and of loosening controland concentratingon infor-
mation quality in the latter). As with communication mechanisms, the
appropriatenessof the different classes of mechanisms will be context

dependent.
Although the number of participationmechanisms is apparentlyvery
large, with many differenttypes of meetings, workshops,conferences, and
fora, thereis a high degree of similarityamong these (indeed,the names are
often used interchangeably).As a consequence, Table3 identifiesjust four
distinct classes of participationmechanism.All types involve face-to-face
(essentially group-based)processes with flexible input from the sponsors,
and most (types 1, 2, and 3) have controlled selection of participants.The
maindifferencebetweentype 1 mechanisms(typifiedby the citizenjury) and
type 2 ones (typified by the task force) is the presence or absence of active
facilitation of participantviews. Generally,type 2 mechanisms tend to be
used with knowledgeablestakeholders(publicrepresentatives),who may be
anticipatedto have the knowledge and motivationto play a role in solving a
real problem(hence, being in less need of facilitation),ratherthan with the
kind of lay public membersgenerally used in practicalexamples of type 1
mechanisms.Such contextfactors-motivation, degreeof preexistingpublic
knowledge on the topic, and specific versus hypothetical/generalproblem
solving-may be importantfor determiningwhich mechanismclass to use to
gain maximal effectiveness. Type 3 mechanismsadd furthercontrol to the
process by structuringthe aggregationof participantviews, using mecha-
nisms such as repeatedpolling or decision aids. This is also true of type 4
mechanisms(e.g., the New Englandmodel of town meetings) in which the
voting process enables aggregation, although this mechanism class has
uncontrolledselection of participantsand hence greaterpotentialinforma-
tion loss from not maximizingappropriateparticipants.Once again, empiri-
cal researchis needed to establishthe contingentappropriatenessof the dif-
ferent mechanisms, because each has potential benefits and drawbacksin
termsof maximizinginformationelicitation,transfer,processing,andaggre-
gation, as detailed in the informationflow model.
Discussion
In this article, it has been argued that imprecise understandingof key

terms in the public participationdomain has hinderedthe conduct of good
researchand militatedagainsteffective participationpractices.To add some
clarityto this domain, an attempthas been made to define key concepts and
providea frameworkfor research.The disparateareaof public participation
has been rephrasedas public engagement,and three significantlydifferent
activities within this domain have been identified and defined-as public
communication,public consultation, and public participation. The three
concepts have been differentiatedaccordingto the natureand flow of infor-
mationbetween the exercise sponsorsand public participants.Accordingto
such an informationflow model, the effectiveness of an exercise may be
ascertainedaccordingto the efficiency with which full andrelevantinforma-
tion is elicited fromall appropriatesources,transferredto (andprocessedby)
all appropriaterecipients,and combined (when this is required).
Public engagement,in its differentforms andits manydifferentinstances
(exercises), is enacted througha variety of structuredmechanisms.These
mechanismsaregreatin numberandgenerallypoorly defined-two charac-
teristics that hinder effective researchand practice. A typology of mecha-
nisms, in which classes of mechanisms are succinctly and appropriately
defined, may counterboth of these difficultiesto a degree. In the typology
thathas been developedhere, mechanisms(of communication,consultation,
andparticipationtypes) have been classified on the basis of theirsimilarities
anddifferenceson a numberof key variablesrelatedto theirstructures.These
between-mechanismvariablesareones thatmighthypotheticallyaffectexer-
cise effectiveness according to the informationflow model. The resultant
typology revealsfour classes of communicationmechanisms,six of consul-
tation mechanisms,and four of participationmechanisms.
It is importantto emphasize that the typology presentedin this article
shouldbe regardedprimarilyas a workingmodel andan aidto researchrather
than as a definitive typology (in many ways, the typology itself should be
seen as of secondary importanceto the explication of the rationale for its
necessary developmentand the process of producingit). Thereare certainly
limitationsto the typology itself. For example, theremay be otherbetween-
mechanismvariablesof equalor greaterimportanceto those used in develop-
ing the typology, which ought to be used in preferenceto, or in additionto,
these. And there may also be otherbasic mechanismclasses thathave been
missed, because we have not taken into account all existing engagement
mechanisms. Underlying these potential difficulties is the fact that of the
plethora of engagement mechanisms that have been developed and used,
there are relatively few definitive accounts of their natures(and these are
often contradictory),and this has limited the number of mechanisms we
could classify with confidence.
The existence of such a typology is an importantstep towarddeveloping
a theory of "what works best when" (Rowe and Frewer 2004)-a theory
of the contingent effectiveness of engagement mechanisms (because one
mechanismis unlikely to be the most appropriate/effectivein all situations).
A furtherstep involves understandingand defining, perhapsvia a second

typology, the different types of context in which engagement takes place.
Empirically,once effectiveness is defined and instrumentsaredevelopedfor
measuringthis, mechanismsof the differentclasses can be comparedto the
different context classes to establish contingent effectiveness. Developing
andtesting such a theoryis likely to have greatacademicandpracticalimpli-
cations (Rowe and Frewer2004), and should be the focus of futureresearch
efforts. The typology itself also needs furtherelaboration:it is an initial
attempt only, and we would expect furtherconceptual considerationand
empiricalresearchto lead to its evolution,perhapsthroughidentifyingother
key variablesthatneed to be included,perhapsby redefiningor undermining
presentlyincludedvariables,and almost certainlyby consideringthe pleth-
oraof mechanismsthatarenot classified hereandclassifying themappropri-
ately. As such, we would welcome otherresearchersextendingand amend-
ing the typology.
Matchingan appropriateclass of engagementmechanismsto an appropri-
ate contextwill not, however,guaranteethatanengagementexercisewill be a
success. Thereare otherimportantvariablesrelatedto the actualapplication
of the particularexercise that will play an equal and perhapsgreaterrole in
this respect (e.g., Webler 1999). We have termed these within-mechanism
variables:they differ from between-mechanismvariablesin showing varia-
tion across the differentpracticalapplicationsof any specific mechanism.
Identifyingthese, andunderstandingtheirpotentialimpacton exerciseeffec-
tiveness, is anotherarearequiringfuturestudy.
NOTES
1. In this article,we deliberatelyadopta simplifiedmodel of the natureandpurposeof com-

municationandparticipationforthepurposeof developinga typologyof mechanisms.This is not
to deny thatthe rationalefor communicationbetweensponsorandpublicis a complex, multifac-
eted issue; for a good discussion on the wider context, see, for example, Gregoryand Miller
(1998).
2. When we talk of "relevant"and "irrelevant"information,we arereferringto a theoretical
relevanceto the issue at hand-for example, in a debate on the safety of genetically modified
food, informationabout health risks is relevant,but informationon what I had for dinnerlast
nightis not. The sponsorobviouslyhas his or herown interpretationaboutwhatis andis not rele-
vant, but this political interpretationis not our concernhere-indeed, it is possibly relatedto a
within-mechanismvariableregardingsponsorintegrity(or some such) that does not affect our
typology, which is based on between-mechanismsources of variance.
3. As such, workshopsareinsteadconsidereda tool thatis presentin severalof the morespe-
cific namedmechanismsthat are included in the table.
REFERENCES
Anson, R., R. Bostrom,andB. Wynne. 1995. An experimentassessing groupsupportsystem and

facilitatoreffects on meeting outcomes.ManagementScience 41 (2): 189-208.
Armstein,S. R. 1969. A ladderof citizen participation.JournalAmericanInstituteof Planners
35:215-24.
Barnes,M. 1999. Buildinga deliberativedemocracy:An evaluationof two citizens'juries.Lon-
don: Institutefor Public Policy Research.
Baughman,M. 1995. Mediation.In Fairnessand competencein citizenparticipation:Evaluat-
ing modelsfor environmentaldiscourse, edited by O. Renn, T. Webler,and P. Wiedemann,
253-65. Dordrecht,the Netherlands:KluwerAcademic.
Bickerstaff,K., and G. Walker.2001. Participatorylocal governanceand transportplanning.
Environmentand PlanningA 33:431-51.
Blahna, D. J., and S. Yonts-Shepard.1989. Public involvementin resourceplanning:Toward
bridgingthegapbetweenpolicy andimplementation.SocietyandNaturalResources2:209-27.
Carr,D. S., and K. Halvorsen. 2001. An evaluationof three democratic,community-based
approachesto citizen participation:Surveys, conversationswith community groups, and
communitydinners.Society and NaturalResources 14 (2): 107-26.
Chipman,H., P.Kendall,M. Slater,andG. Auld. 1996.Audienceresponsesto ariskcommunica-
tion message in 4 media formats.Journal of NutritionEducation28 (3): 133-39.
Coglianese, C. 1997. Assessing consensus: The promise and performance of negotiated
rulemaking.Duke Law Journal46 (6): 1255-349.
Coote, A., and J. Lenaghan.1997. Citizens'juries: Theoryinto practice, London:Institutefor
Public Policy Research.
Crosby,N., J. M. Kelly,andP.Schaefer.1986. Citizens'panels:A new approachto citizenpartic-
ipation.Public AdministrationReview46:170-78.
DemocracyNetwork. 1998.Democraticpractice:A guide.London:Local GovernmentAssocia-
tion Publications.
Dienel, P. C., and0. Renn. 1995. Planningcells: A gate to "fractal"mediation.In Fairnessand
competencein citizenparticipation:Evaluatingmodelsfor environmentaldiscourse,edited
by O. Renn, T. Webler,and P. Wiedemann, 117-40. Dordrecht,the Netherlands:Kluwer
Academic.
Dowswell, T., S. Harrison,M. Mort,and P. Lilford. 1997. Healthpanels: A survey.Leeds, UK:
Nuffield Institutefor Health.
Einsiedel, E. F., E. Jelsoe, and T. Breck. 2001. Publics at the technology table:The consensus
conference in Denmark,Canada,and Australia.Public Understandingof Science 10 (1):
83-98.
Fiorino,D. J. 1990. Citizenparticipationandenvironmentalrisk:A surveyof institutionalmech-
anisms. Science, Technology,& Human Values15 (2): 226-43.
Fishkin,J. S., and R. C. Luskin. 1999. Deliberativepolling andcitizen consultation.UK CEED
Bulletin 55:6-9.
Glass, J. J. 1979.Citizenparticipationin planning:The relationshipbetweenobjectivesandtech-
niques. Journalof the AmericanPlanning Association452:180-89.
Gregory,J., andS. Miller. 1998. Science inpublic: Communication,cultureand credibility.New
York:Plenum.
Gundry,K. G., and T. A. Heberlein. 1984. Do public meetings representthe public?American
PlanningAssociation Journal(Spring): 175-82.
Guston,D. H. 1999. Evaluatingthe first US consensus conference:The impactof the citizens'

panel on telecommunicationsand the futureof democracy.Science, Technology,& Human
Values24 (4): 451-82.
Hannah,S. B., andH. S. Lewis. 1982. Internalcitizen controlof locally initiatedcitizen advisory
committees:A case study.Journalof VoluntaryAction Research 11 (4): 39-52.
Hill, G. W. 1982. Groupversusindividualperformance:AreN+ 1 headsbetterthanone?Psycho-
logical Bulletin91 (3): 517-39.
Joss, S. 1998. Danishconsensusconferencesas a modelof participatorytechnologyassessment:
An impactstudyof consensus conferenceson DanishParliamentandDanishpublic debate.
Science and Public Policy 25 (1): 2-22.
Kathlene,L., and J. A. Martin.1991. Enhancingcitizen participation:Panel designs, perspec-
tives, and planning.Journal of PolicyAnalysis and Management10:46-63.
Lowndes,V., G. Stoker,D. Pratchett,D. Wilson, S. Leach, and M. Wingfield. 1998. Enhancing
public participation in local government:A research report. London: Departmentof the
Environment,Transportand the Regions.
Lundgren,R., and A. McMakin. 1998. Risk communication:A handbookfor communicating
environmental,safety,and health risks, 2nd ed. Columbus,OH: Batelle.
Lynn,F.M., andG. J. Busenberg.1995. Citizenadvisorycommitteesandenvironmentalpolicy:
What we know, what's left to discover.RiskAnalysis 15 (2): 147-62.
Martin,S., and A. Boaz. 2000. Public participationand citizen-centredlocal government:Les-
sons from the best value and bettergovernmentfor older people pilot programmes.Public
Moneyand Management20 (2): 47-53.
Mayer,I., J. de Vries, andJ. Geurts.1995. An evaluationof the effects of participationin a con-
sensus conference.In Publicparticipationin science: The role of consensusconferencesin
Europe,edited by S. Joss and J. Durant,109-24. London:Science Museum.
Mclver,S. 1998.Healthydebate?An independentevaluationof citizens'juriesin healthsettings.
London:King's Fund.
Moore, S. A. 1996. Defining 'successful' environmentaldispute resolution:Case studies from
public land planningin the United States and Australia.EnvironmentalImpactAssessment
Review 16:151-69.
Nelkin,D., andM. Pollak. 1979. Publicparticipationin technologicaldecisions:Realityor grand
illusion? TechnologyReview 81:55-64.
New EconomicsFoundation.1999.Participationworks!London:New EconomicsFoundation.
Offner,A. K., Kramer,T. J., and J. P. Winter.1996. The effects of facilitation,recording,and
pauses on groupbrainstorming.Small GroupResearch27 (2): 283-98.
Ouellet,F., D. Durand,andG. Forget.1994. Preliminary-results of an evaluationof 3 healthycit-
ies initiativesin the Montrealarea.Health PromotionInternational9 (3): 153-59.
Owens, S. 2000. Engaging the public: Informationand deliberationin environmentalpolicy.
Environmentand PlanningA 32:1141-48.
Pierce,J. C., andH. R. Doerksen. 1976. Citizenadvisorycommittees:The impactof recruitment
on representationand responsiveness. In Waterpolitics and public involvement,edited by
J. C. Pierce and H. R. Doerksen, 249-66. Ann Arbor,MI: Ann ArborScience Publishers.
Plumlee,J. P.,J. D. Starling,andK. W. Kramer.1985.Citizenparticipationin waterqualityplan-
ning. Administrationand Society 16 (4): 455-73.
Roberts,T., S. Bryan,C. Heginbotham,and A. McCallum. 1999. Public involvementin health
care prioritysetting:an economic perspective.Health Expectations2:235-44.
Rosener,J. 1975.A cafeteriaof techniquesandcritiques.PublicManagement(December):16-19.
Rosener,J. B. 1982. Makingbureaucratsresponsive:A studyof the impactof citizen participa-

tion and staff recommendationson regulatory decision making. Public Administration
Review (July): 339-45.
Rowe, G. 1992. Perspectiveson expertise in the aggregationof judgments. In Expertise and
Decision Support,edited by G. Wrightand F Bolger, 155-80. London:Plenum.
.1998. The use of structuredgroupsto improvejudgmentalforecasting.In Forecasting
withjudgment,edited by P. Goodwin and G. Wright,201-35. Chichester,UK: Wiley.
Rowe, G., andL. J. Frewer.2000. Publicparticipationmethods:A frameworkforevaluation.Sci-
ence, Technology,& Human Values25 (1): 3-29.
. 2004. Evaluatingpublic participationexercises: A researchagenda.Science, Technol-
ogy, & Human Values29 (4): 512-56.
Rowe, G., andJ. Gammack.2004. Promiseandperils of electronicpublic engagement.Science
and Public Policy 31 (1): 39-54.
Rowe, G., R. Marsh,and L. J. Frewer.2004. Evaluationof a deliberativeconference.Science,
Technology,& Human Values29 (1): 88-121.
Sinclair,M. 1977.The publichearingas a participatorydevice:Evaluationof the IJCexperience.
In Public participationin planning, edited by W. R. D. Sewell and J. T. Coppock, 105-22.
New York:John Wiley.
Smith,L. G., C. Y.Nell, andM. V.Prystupa.1997.Theconvergingdynamicsof interestrepresen-
tation in resourcesmanagement.EnvironmentalManagement21 (2): 139-46.
Sproull,L., andS. Kiesler.1986. Reducingsocial contextcues: Electronicmail in organizational
communication.ManagementScience 32 (11): 1492-512.
Stewart,T. R., R. L. Dennis, and D. W. Ely. 1984. Citizen participationandjudgmentin policy
analysis:A case study of urbanair qualitypolicy.Policy Sciences 17:67-87.
Susskind,L., and G. McMahon. 1985. The theoryandpracticeof negotiatedrule making.Yale
Journal on Regulation3:133-65.
Twight, B. W., and M. S. Carroll. 1983. Workshopsin public involvement:Do they help find
common ground?Journalof Forestry(November):732-35.
Webler,T. 1995. "Right"discoursein citizen participation:An evaluativeyardstick.In Fairness
and competence in citizen participation:Evaluatingmodelsfor environmentaldiscourse,
editedby O. Renn,T.Webler,andP.Wiedemann,35-86. Dordrecht,the Netherlands:Kluwer
Academic.
1999.The craftandtheoryof publicparticipation:A dialecticalprocess.Journalof Risk
Research2 (1): 55-71.
Wiedemann,P. M., and S. Femers. 1993. Public-participationin waste managementdecision-
making-analysis and managementof conflicts. Journal of HazardousMaterials 33 (3):
355-68.
GeneRowegraduatedfromthe Universityof Bristol inpsychologyand gained his Ph.D.

from the Bristol Business School at the Universityof the Westof England(UWE).He is
currentlya senior researcherat the Instituteof Food Research,Norwich, in the United
Kingdom.His researchinterestsincludejudgmentand decision making,expertsystems,
forecasting, the Delphi technique,and the influenceof the media on public riskpercep-
tions. His main interest at present lies in public participation methods and their
evaluation.
LynnJ. Frewer is Research Professor in Food Safety and ConsumerBehaviour at the

University of Wageningenin the Netherlands. She has a backgroundin psychology
(MSc and PhDfrom UCL, London,and the Universityof Leeds, both in the UK). She
has research interests in risk analysis, as well as societal aspects offood safety, and
emerging technology.

A Typology of Public Engagement Mechanisms

Uploaded by

Document Information

Original Description:

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

A Typology of Public Engagement Mechanisms

Uploaded by

Copyright:

Available Formats

A Typology of Public Engagement Mechanisms

Author(s): Gene Rowe and Lynn J. Frewer

Imprecisedefinitionof keytermsin the "publicparticipation"domainhave hinderedthe

Keywords: public participation; public engagement; participation mechanisms;

The Concept and Enactment of Public Participation

In recent times, there has been an internationaltrend towardincreased

Science, Technology,& HumanValues,Vol. 30 No. 2, Spring2005 251-290

processes/techniques/instruments-which shall be collectively termed

involvementandpublic engagement)may be used in place of thatterm.What

Definition of Public Participation:

Before classifying public participationmechanisms, it is necessary to

distinctionof participationsituationsfrom nonparticipationsituationsasso-

Sponsor Public Representatives

Sponsor * Public Representatives

Sponsor t Public Representatives

Figure1. Thethreetypes of publicengagement.

throughoutwe use phrasingthatmightbe takento assumethatthe sponsoris

negotiation serves to transformopinions in the members of both parties

Public Engagement Mechanisms

The numberand varietyof engagementmechanismsare large and grow-

environmental.(Also "Monitor- CommunitySite our futureby making England

Arbitration (Mediation)(3, 4) Planning(1) Focus Group(3, 6, 10) 25, 26)

Figure 2. Alphabetical listing of "participation" mechanisms (references in parenth

numberof decision-aidingtools, such as a Delphi process, to ascertainthe

The second problem,namely,the uncertainequivalenceof terms,is even

differs from the others, it is unsurprisingthat no significant theory has

Categorizing Engagement Mechanisms

In the academic literature,there have been few efforts at developing a

associatedcertain"models"(i.e., mechanisms),namely,advisorymodels (of

commissioned for different purposes, use different amounts and types of

The Issue of Effectiveness

Some sourcesof mechanismvariancearemore importantthanothers.By

behavioron one hand or, on the other,to the mistakeninterpretationof the

* Publiccommunication: maximizingtherelevantinformation fromthespon-

Fromthis perspective,efficiency may be compromisedwhen the informa-

Althoughit is truethatmuch of the success or failureof a particularexer-

The significantbetween-mechanismvariablesare discussed with regard

Variables Associated with

MechanismVariable Levels of Variable Aspect of Effectiveness Pote

Participant Controlled Maximizerelevantparticipants

tation,and participationexercises. The first "figure"dependsentirelyon the

There is a thirdfigure, however, that is perhapsof even greatersignifi-

(because,all else being equal, a mechanismthatinvolves "high"numbershas

Variables Associated with Maximizing

Each active participant in an engagementexercise can be consideredto

There are at least two structuralaspects of engagementmechanismsthat

way in which participantsmightrespondto exercisesthatoverlyrestricttheir

Variables Associated with Maximizing

Informationsources may include the sponsorsand theirexperts(in com-

Variables Associated with Maximizing

The aim of engagementis to acquireall relevantinformationfromall rele-

to dial a "freephone"number.Response rates might also be enhanced by

wordingmight affect this. Forexample, technicallycomprehensiveand cor-

Variables Associated with Maximizing

In participationexercises and, to a lesser extent, consultationexercises,

althoughgroup solutions have been found to generallybe betterthanaggre-

Thereareseveralways in which a typology mightbe developed,thatis, of

Consultation Opinionpoll Controlled These mechanisms are ess

Consultation Electronic Uncontrolled As type 2, but withuncontr

Participation Actionplanning Controlledselection The mechanisms of this typ

Participation Deliberative Controlledselection This class is also similarto

which is distinguishedfromthe othertypes by involvingthe controlledselec-

appropriatenessof the different classes of mechanisms will be context