ContentslistsavailableatScienceDirect
Social
Networks
j ou rn a l h o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / s o c n e t
Social
influence
and
discourse
similarity
networks
in
workgroups
Johanne
Saint-Charles
a,∗,
Pierre
Mongeau
baDépartementdecommunicationsocialeetpublique,Centrederechercheinterdisciplinairesurlebien-être,lasanté,lasociétéetl’environnement (CINBIOSE),UniversitéduQuébecàMontréal,CasePostale8888,succ.Centre-Ville,Montréal,Québec,H3C3P8,Canada
bDépartementdecommunicationsocialeetpublique,UniversitéduQuébecàMontréal,CasePostale8888,succ.Centre-Ville,Montréal,Québec,H3C3P8, Canada
a
r
t
i
c
l
e
i
n
f
o
Articlehistory: Availableonlinexxx Keywords: Discoursesimilarity Network Similaritythreshold Socialinfluence Workgroupstages Socio-semantica
b
s
t
r
a
c
t
Adoptingasocio-semanticperspective,thisstudyaimstoverifytherelationbetweensocialinfluence anddiscoursesimilaritynetworksinworkgroupsandexploreitsmodificationovertime.Dataconsistof videotranscriptsof453-hgroupmeetingsandweeklysociometricquestionnaires.Relationbetweentie strength,actorcentralitywithintheinfluencenetwork,andsharedelementsofdiscoursebetweengroup membersareexaminedovertime.Observedcorrelationssupportthehypothesisofarelationbetween socialinfluenceanddiscoursesimilarity.Changesovertimesuggestasimilaritythresholdabovewhich therelationbetweensimilarityandinfluenceisreversed.
©2017ElsevierB.V.Allrightsreserved.
1. Introduction
Socialinfluenceandrelatedphenomena,suchasleadershipand socialpower,areintrinsictoanyhumanorganisingprocess.Social influencehasbeenconceivedasarelationalprocessforwhich inter-personalinfluencehappeningatthelevelofthedyadconstitutes abuildingblock(Carteretal.,2015;FriedkinandJohnsen,2011; Simpsonetal.,2015).Inworkgroups,socialinfluenceaffectshow groupsmakedecision,conducttheirwork,orshapetheir under-standingofthetaskandtheenvironment,andmore(Johnsonetal., 2015;Melamedand Savage,2016; Pavitt,2014;Westaby etal., 2016).Interactionsandcommunicationarecentraltosocial influ-ence(Guastello,2007; Moscovici, 1988), and discourse isa key componentofboth. Here,discourse isunderstoodas“ageneral termthatappliestoeitherwrittenorspokenlanguagethatisused forsomecommunicativepurpose”(Ellis,1999,p.81).
Thestudypresentedinthispaperoffersanovelperspectiveon theinterplaybetweensocialinfluenceanddiscourseinworkgroups byexploringtherelationbetweensocialinfluencenetworksand discoursesimilaritynetworks,a typeof networkin which rela-tionshipsbetweenpeopleareestablishedonsharedelementsof discourse(words,expressions,concepts,topics,etc.).Basedon ver-balinteractionsbetweenworkgroupmembersovertime,thestudy offersempiricalsupportfortherelationbetweensocial relation-shipsanddiscoursesimilarityatthedyadicandgrouplevelsand
∗ Correspondingauthor.
E-mailaddress:saint-charles.johanne@uqam.ca(J.Saint-Charles).
confirmstherelevanceofperformingthattype ofjointanalysis tobetterunderstandcommunication phenomenasuchas inter-personal influence. Our resultsalso highlightthepertinence to consider strength and reciprocity of social ties in thestudy of discoursesimilarity.Finally,thelongitudinalaspectofthestudy suggestsnon-linearco-evolutionbetweenthetwonetworksand thepresenceofa“similaritythreshold”abovewhichgreater sim-ilarityisassociatedwithlessinfluence,aresultthatcouldexplain someaspectsoftheevolutionofworkgroupsthroughtime.
2. Structuralanddiscursiveapproachestosocialinfluence
Thestudyofsocialinfluencecanbegroupedintotwobroadand generallydistinctapproaches:(a)astructuralapproachthat under-standssocialinfluencethroughrelationshipsbetweenindividuals and(b)adiscursiveapproachthatidentifiesinfluencewithinthe discourse of social actors. In what follows, we present a brief overviewofbothandarguethateachapproachmustbetakeninto accounttobetterunderstandsocialinfluenceinworkgroups.
2.1. Astructuralapproach
Thestructuralapproachfocusesontherelationalaspectofsocial influenceandrelatedconcepts,suchasleadershipandsocialpower. Atthedyadiclevel,therelationalproximityperspectivetheorises thatthemerepresenceofatieallowsforinformationexchange and sharedinfluence (Rice,1993).However, numerous studies, startingwithColemanetal.(1957),havefoundevidencethattie strengthmustbeconsideredinthestudyofinfluence,particularly http://dx.doi.org/10.1016/j.socnet.2017.09.001
giventherelationalintensityandreciprocityitimplies(Araland Walker,2014;Bondetal.,2012).Atthegroupororganisationlevel, socialinfluenceisseenasintertwinedwithmembers’relationship networksand,mostimportantly,thestructuresandpatternsthey create(Carteretal.,2015;Friedkin,1993;FriedkinandJohnsen, 2011;Saint-CharlesandMongeau,2009).AccordingtoBrassand Krackhardt(2012,p.355),“thestructureofsocialnetworksstrongly affects the extent to which personal attributes, cognition, and behaviour result in power in organisations”. Methodologically speaking,moststudiesusing this approach employsociometric questionnairestogatherdataoninfluencerelationshipsorother typesofrelationships,includingfriendship,advice,andsupport.
Networkcentralityseekstomeasure“theprominenceor impor-tanceoftheactorsina socialnetwork”(Wassermanand Faust, 1994,p.170).Degreecentralitybasedonthenumberofan indi-vidual’sincomingand outgoingties is thoughttoberelated to influencebecausecentralpeoplehaveeasieraccesstoresources (BrassandLabianca,1999).Manystudieshaveshownthatactors’ degreecentralitywithinnetworksislinkedtotheiroverall influ-encewithinthegrouportheorganisation,suchinfluencebeing measured through actors’ perceptions of the most influential individualsintheirgroupororganisationorusingavarietyof per-formancemeasures.Beginningwiththeearliestlaboratorystudies oncommunicationpatternsintask-orientedgroups(Bavelas,1950; Leavitt,1951),thislinkhasbeendemonstratedinavarietyof con-texts,includingissueresolutioninsmallorganisations(Friedkin, 1993); multiculturalteams (Salk and Brannen, 2000); the role of CEOs in mergers and acquisitions (El-Khatib et al., 2015); decision-making in Initial Public Offerings (Owen-Smith et al., 2015);innovation(Ibarra,1993a);organisationalstrategies(Boje andWhetten,1981);onlinediscussiongroups (Huffaker,2010); andmore.Nonetheless,theprocessthroughwhichanindividual assumesacentralpositioninanetworkisnotyetfullyunderstood. Researchershavealsoexploredhowcentralitycombineswith otherfactorsalsocorrelatedwithsocialinfluence,includingformal status(AstleyandSachdeva,1984);genderandethnicity(Ibarra, 1993b);politicalskills(Treadwayetal.,2013);trust(Sparroweand Liden,2005);anduseofstrategies,cognitiveandcommunicative abilities,emotionalabilities,andperformance(Brass,1985;Brass andBurkhardt,1993;Emery,2012;Krackhardt,1990;Zuchowski, 1987).
Thesekindsofstudiesarefrequentlycriticisedforfocusingon themetaphor of “ties as conduits”(Borgatti and Foster, 2003), whichneglectswhatcirculatesthroughtheconduits−notably, dis-course(LabiancaandBrass,2006;MongeandContractor,2003).In lightofthiscritique,thenextsectionexaminesstudiesthathave exploredtherelationbetweendiscourseandinfluence.
2.2. Adiscursiveapproach
Beginningwiththerhetoricaltradition(Craig,1999)and con-tinuingthroughagrowingcontemporaryinterestindiscoursein organisationalstudies(BarberioandMonti, 2014),discoursehas alwaysbeenseenasanimportantmeansofinfluenceandpower. Whathasbeencoined “thelinguisticturnof the20thcentury” (Rorty,1967)hasspurredtheemergenceofmanydistinctive,but complementary,fieldsofstudyfocusedon“discourse”(Barberio andMonti,2014;Oswicketal.,2010).Thevarioususagesofthe term“discourse” resultingfromthe field’sexpansion(Alvesson andKarreman,2000;Fairclough,2003;Jian,2008)haveproduced abroadconceptwhosecoremeaningisthatoflanguageusedby humansforcommunication.
Theinterplaybetweendiscourse andinfluencehasbeenthe subjectofasignificantnumberofstudies,manyofwhichfocused onhowdiscourse affectsorsupportstheinfluenceofdominant groupsorindividuals.Discourseisseenbothasinextricablylinked
withpowerandanessentialtoolofmanipulation(Aman,2009; Fairclough,1989;Van Dijk,1989; vanDijk, 2006;Wang,2006). Thesestudiesfavourpredominantlyqualitativemethods, includ-ingvariantsofdiscourseanalysis(critical,historical,Foucauldian); narrativeandmetaphoranalysis;linguisticanalysis,rhetoric,and interaction;andconversationanalysis.Qmethodology,a combina-tionofdiscourseandcontentanalysis,andothermixedmethods havealsobeenused(Balogunetal.,2014;Clareetal.,2013;Dijk, 2012;Holmes,2009;KotwalandPower,2015;McKenna,2016).
Thediversityof topicsand settingsin thisliteratureis itself testimonytothescientificandsocialimportanceoftherelation betweendiscourseandinfluence.Forexample,HardyandMaguire (2016)havequestionedthewaydiscourseonriskisintertwined withpowerissueswithinorganisations.Duvaletal.(2015)have shedlight ontheframing influenceofforms imposed ongrant applicants (primarily NGOs) by a granting agency for interna-tionaldevelopment.Othershaveexplored(dominant)publicpolicy discourseanddebatesoverissuessuchaswetlandmanagement (Clareetal.,2013).Anotherresearchtrendisconcernedwiththe “discourseofstrategy”,whichexaminesdiscoursefromthe per-spectiveof “strategyas practice”(Balogun et al.,2014,p. 176). Otherscholars haveexplored howdiscourse contributes tothe creationorreproductionofstatus,gender,andothersocial inequal-ities,aswellashowresistancetodominantdiscoursesisexpressed intheworkplace,education,publicpolicy,medicine,themedia, andsports(BergvallandRemlinger,1996;Codo,2011;Edleyand Wetherell,1997;EzeifekaandOsakwe,2013;KilbyandHorowitz, 2013;Mayes,2010;MenzandAl-Roubaie,2008;Prego-Vazquez, 2007;Toft,2014;ZanoniandJanssens,2015).
Organisational studies addressing managers’ influence and leadershipabound(DayandAntonakis,2012),andmorerecent the-orisationshavealsoconsideredleaders’discourses(Fairhurstand Connaughton,2014;FairhurstandCooren,2009).Transformational and neocharismatictheories(Jordan, 2005;Meda,2005), which arguethataleader’sinfluencedependsontheabilitytoframethe situationsoastoinspiremembers,arerepresentativeofthis litera-ture(Antonakisetal.,2004).Thisresearchtrendismoreconcerned with“whatmakesagoodleader”thanwithexposingpowerissues hiddeninthediscourse.
Insmallgroupsettings,the“dominantdiscourse”focusisalso lessprevalent.Forexample,researchershaveexaminedhow dis-courseinfluencesinteractionaldynamics duringmeetings; how chairpersonsguidemeetingstowardspecifictopics,andhow deci-sionsaremade(AsmußandSvennevig,2009;Barske,2009;Clifton, 2009;Holmes,2009;Lazzaro-Salazaretal.,2015).Micro-analytical approachesbasedondiscourse,linguistic,orconversationanalysis havefrequentlybeenusedtoexploretheseissues.
Theexplorationofconflicts,disagreements,andclique forma-tion as loci of influence hasalso revealed sequential discourse phenomenaintheconstructionofalliances(Kangasharju,2002). Nielsen (2009, p. 23) investigated “how interpretational work supportsorganisationalgoalsand values”inbusinessmeetings. Specifically,she shows how this interpretative work is accom-plishedthroughemployees’acquisitionoforganisationallanguage. Inequalityandpowerissuesbetweenstakeholdersinthecontext ofdeliberativesmallgroupsor“democracymeetings”havealso attractedattention(Lazzaro-Salazaretal.,2015;Vargasetal.,2016). With regard toleadership, Choiand Schnurr (2014) have used discourseanalysistoexploredifferentmembers’performanceof leadershipinaleaderlessteam.
Thisoverviewleaveslittledoubtastotheimportanceofthe interplaybetweensocialinfluenceanddiscoursebut,asKrinsky (2010,p.627)observes,thisresearchtrendhastendedtoneglect relationalcontexts,including “therelations betweenthethings speakerssay”,andtreatdiscourseasstaticratherthandynamic. In thesamevein, Brummans etal. (2008)notethat studies on
conflictframingare“unabletoexplaintheintricatecommunication processesthroughwhichpeoplemakesenseofasituation”.
Approachesfocusedontheco-constructionof meaninghave developedamoredialogicalinterpretationoftheinfluence pro-cess(ChoiandSchnurr,2014;FairhurstandConnaughton,2014) thatshedslightontherelationalandconstructedaspectsof mean-ing through discourse in cases, for instance, where influential individualsareseenasactingasintermediariesbetweenvarious representations of the situation (Barge, 1989, 1996; Barge and Hirokawa,1989;Thayer,1988).FairhurstandConnaughton(2014) haveshownintheirliteraturereviewanumberofstudieslookingat relationshipsandinteractionswithregardsmostlytoleadershipas aninfluence-orientedlanguage.Theirreviewhighlightsthe impor-tantrolediscourseandsocialinteractionsplayininfluenceandalso showthatthestructuralaspectremainsneglectedinthosestudies.
3. Linkingsocialnetworksanddiscoursesimilarity
A socio-semantic perspective for the joint study of social networksand discourse similarity appearsas a promising path tobringtogetherthestructuralandthediscursive.Scholarshave travelledthispath,althoughnotwithregardtoinfluence.
Homophily,thetendencyforpeopletoformrelationshipswith similarothers,isa well-establishedsocialphenomenon (Carley, 1991;CurryandDunbar,2013;McPhersonandSmith-Lovin,1987; McPhersonet al.,2001).Asstated byCurry and Dunbar(2013, p.337):“Individualswhoaresimilarwithregardtorace, ethnic-ity,sex,age,religion,education,occupation,socialclass,attitudes, opinions,andbeliefsaremorelikelytoassociatewithoneanother than would be expected by chance”. In view of this, semantic homophilyistobeexpected(RothandCointet,2010).Aswell,using similaritytohighlightcommonelementsof differentdiscourses hasbeensuggestedbyresearchersinterestedinidentifyinggroups ofindividualssharingsemanticuniverses(RoussiauandBonardi, 2001).Thereby,researchershavedevelopedmodelsandmeasures ofsimilarityofdiscoursebetweenindividualinterconnectedina socialnetwork toexplainvarious individual,socialand cultural phenomena.
Forinstance,Carley (1986a,1991) hasmodelledtherelation betweensocialinteractionsandcognitivestructureswithinsmall organisations.Hermodelfocusesonhowinformationexchanged throughthestructureofsocialinteractionsaffectsindividuals’ cog-nitivestructureandhowbothstructurescombinetocreategroup cohesion,consensus,and stability. In short,themore members sharethesameinformationthemorethegroupbecomes cohe-sive,capableofconsensusandstablethroughtime.Intheirstudyof riskperception,SchererandCho(2003)confirmedtheimportance social networks in building “groups or communities of like-minded”individuals.CointetandRothhavestudiedthecoevolution ofthesocialandsemanticnetworksofscientists(Roth,2008)and bloggers(CointetandRoth,2009)andhighlightedtheirsimilarities anddifferences(RothandCointet,2010).Forexample,although degreecentralityinthesocialnetworkiscorrelatedwith central-ityintheactors-conceptsnetworkbothnetworksexhibitdifferent patternsinthewaybothmeasuresarelinked.Intheirstudyofa community-basedprotest,TeoandLoosemore(2011)notedthat socialnetworksplay acritical rolein thetransmissionof ideas, information,beliefs,andperceptions.Theyobservedthatshared experiencesthroughrelationshipnetworkshelpedcreateshared culturalmeaningsandinterpretations.Danowski(2013)hasshown thatbetween2007and2011,“semantic-baseddivisionnetworks arebecomingmoredecoupledfromtheformalmembership-based networks”inhisstudyofthedivisionsoftheInternational Commu-nicationAssociation.Basovetal.(2017)andBasovandBrennecke (inpress)haveexploredinstrumental,expressiveandinteraction
networksofsmallself-runcreativeorganisationsandtheirlinksto sharedmeaningasawaytobringtogetherthesocialandthe cul-tural.Thecomplementaritybetweensocialnetworksanddiscourse similarityhavealsogivenrisetoseveralstudiesofsocialmediaand networksofco-citationandco-publication−trendsthatboth ben-efitfromtheincreasingavailabilityoflargedatasets(Roth,2013).
4. Workgroupsovertime
AworkgroupisunderstoodhereinasimilarwayasCohenand Bailey(1997):itisanensembleofindividualsmutuallyinteracting intheirtasksandsharingresponsibilityfortheoutcomes.Group membersconsiderthemselvesandarebeingconsideredbyothers asasocialentityembeddedinlargersocialsystems.Workgroups likesocialnetworksaredynamicandchangeovertime(Arrowetal., 2004;Faust,2011;MongeauandSaint-Charles,2011).
Manyofthenumerousgroupdevelopmentmodelsproposed thatgroupsevolvethroughvariousstages(Braaten,1974;Dunphy, 1968;Mills,1967;Schutz,1958;St-Arnaud,1978;Tuckman,1965; TuckmanandJensen,1977;Wheelan,1994).Althoughlinear mod-elshavebeencriticisedforneglectingtherolesoforganisational time, task types, and environmental uncertainty (Chang,2006; Hare, 2003; McGrath, 1991; Poole and DeSanctis, 1990), their validityissupportedbyresearch,particularlyforzero-history, lead-erless,andfixed-durationgroups(Braaten,1974;Dunphy,1968; Mills,1967;Wheelan,2009).
Linearprogressivemodelshaveidentifiedfromthree tonine stages leading to group maturity, at which point group mem-bersbecomemorefocusedontaskaccomplishment.Threemain stages are commonto all models(Chidambaram and Bostrom, 1996;Wheelan,2009).Borrowing fromtheterminologyinitially proposedbyTuckman(1965),wecallthesecommonstages form-ing,storming,andnorming.Theformingstage,asitsnameimplies, referstotheperiodduringwhichthegroupisformedandmembers areinobservationandtesting mode.Thisisfollowed by storm-ing,duringwhichdisagreementsarevoicedandconflictsappear. Resolvingtheseconflictsleadstotheestablishmentofnormsand groupstabilisation,ornorming.Thesestageshavefuzzyboundaries (McGrathetal.,2000);groupsmaylingerinonestagewithout nec-essarilymovingontothenext.Forexample,theformingstagemay takelongerinzero-historygroupsbecausemembersdonotalready knowoneanotherandmayneedmoretimeunderstandthegroup’s goals(ChidambaramandBostrom,1996).
With regard to the content of exchanges between group members,studieshaveshownasignificantincreaseinexchange prototypicalityandhomogeneityovertime(Postmesetal.,2005); thisisconsistentwithalinearprogressivemodel.Johnsonetal. (2015)observedthattheemergenceofalternativeopinionsearly ingroupdevelopmentmayleadtotaskconflicts,whichsuggests an explanation for the group moving into the storming stage. Theseresultsmayimply co-evolutionbetweengroupmembers’ exchangesandtherelationshipstheydevelop,thusaddingmore incentivetostudybothrelationshipsanddiscoursebetween mem-bers.
5. Hypothesis
Thevariousresearchtrendsidentifiedabovehaveestablished the importance of discourse and relationships for social influ-encein groups andorganisations, aswellas theneedto study themtogether.Wehavealsopresentedtherelevanceofdiscourse similaritynetworkstorevealrelationshipsbetweenindividuals. Therefore, our general hypothesis posits a positive association between influence networks and discourse similarity networks inworkgroups.Totestthisassociation,weproposefourspecific
hypotheses,oneatthedyadiclevel(interpersonalinfluence)and theothersatthegrouplevel.
Aswehaveseen,interpersonal(dyadic)influenceisconsidered oneofthebuildingblocksofsocialinfluenceandtheintensityand reciprocityofthedyadictieaffectthisinfluence.Thus,we hypothe-sisethatthestrongerthedyadictie,themoresimilarthediscourse.
Hypothesis1. Discourse similarity and influence tie strength
betweentwoindividualsarepositivelyrelated.
Atthegrouplevel,ourhypothesesarebased ondegree cen-tralityintheinfluencenetworkandareconsistentwithwhatis presentedintheliteraturecitedabove.Inordertotakeintoaccount thestrength ofties, weusedboth valueddegreecentrality and dichotomisedmatrices.Itisexpectedthatthemostcentral indi-vidualsintheinfluencenetworkwillbethosesharingthelargest numberofelementsofdiscoursewiththelargestnumberof mem-bers;theseindividualswillthereforebecentralinthediscourse similaritynetworks.
Hypothesis2. Thesimilarityofagroupmember’sdiscourseto
thatofallothergroupmembersispositivelyrelatedtohisorher valueddegreecentralityintheinfluencenetwork.
Becauseitaffectsdyadicinterpersonalinfluence,we hypothe-sisethattiestrengthwillalsobeofimportanceatthegrouplevel.
Hypothesis3. Thepositiverelationbetweendegreecentralityand
discoursesimilarityforstronginfluencetiesisstrongerthanthe relationbetweenweakinfluencetiesanddiscoursesimilarity.
Ourfourth hypothesisis basedonthethree stages ofgroup developmentpresentedabove.Weexpectthatthesestageswill bereflectedintheinterplaybetweentheinfluencenetworkand discoursesimilarity networks.Althoughwe expecta consistent positivecorrelationbetweendiscoursesimilarityandvalued cen-tralityintheinfluencenetwork,weassumethatthiscorrelation willbeweakerduringthesecondstage(storming)giventhatthis stageischaracterisedbydisagreementsandconflicts.The assump-tionhereisthatthewordsusedbymembersadvocatingdifferent opinionsshouldbedistinct.
Hypothesis4. Thepositiverelationbetweenvalueddegree
cen-tralityanddiscoursesimilaritywillbeweakerduringthesecond stageofgroupdevelopment.
6. Method
Theoverallmethodconsistsofcorrelatingdatafromsociometric questionnairesoninfluencetieswithinworkgroupswithdiscourse similaritynetworkscreatedfromgroupmembers’discussion dur-ingmeetings.
6.1. Sample
Dataforthisstudywerecollectedfromasampleof34 French-speakingstudents(29womenand5menintheirearlytwenties) enrolledin an undergraduate communication programme. Stu-dents were randomly divided into five groups of six to eight relatively homogeneous individuals in terms of ethnic origin, gender,educationalbackground,andage.Thesegroupscanbe con-sideredzero-history, leaderless,and of fixedduration. Students wereatthebeginningoftheirundergraduateprogramme,hadnot metoneanotherbefore,andnoformalstructureexistedwithin theirgroup.Thesegroupswereaskedtoworktogetheronatask worth40%oftheirtotalcoursemark;themarkwascollective.The taskwastoproduceanacademicpaperandpresentittotheir class-matesattheendofthesession.Thetaskwasthereforerealand impliedsignificantconsequencesforgroupmembers(thecourse
Table1
Categorisationofreciprocalties.
Tiestrength Reciprocity Strengthvaluecombination 1 Weakreciprocal(co-presence
influenceties)
1-1 2 Non-reciprocal(allinfluence
ties)
1-2,2-1,1-31-3,3−1 3 Reciprocal(stronginfluence
ties)
2-3,2-2,3-2,3-3
ismandatoryintheprogramme).Groupsmetforthreehoursper
weekninetimesovertenconsecutiveweeks(withtheexception
ofoneweek,whentherewasnomeeting).Allstudentswere
vol-unteers.Forethicalreasons,aprocedurewassetupallowingany
studenttoanonymouslyrefusetoparticipate.Thiswaspossible
becauseonlyhalfofthegroupsformedforthecoursewereretained
forthestudy.ThestudyreceivedResearchEthicsBoardapproval
fromtheauthors’institution.
6.2. Datacollection
Data oninfluence ties werecollected througha sociometric
questionnairefilledoutbygroupmembersthedayaftereach
meet-ing(n=306 withoutmissing values, 34 members×9 meetings;
consideringtheinevitableabsenceofmembersatmeetingsover
aperiodofnineweeks,theobservedminimumisn=280).
Datarelatedtodiscoursesimilaritywerecollectedfromvideo
transcriptsof453-hgroupmeetings(9meetings×5groups×3h)
fora totalof135hofvideo.Duetofinancialconstraints,
trans-cripts were limited to 30min per meeting. In order toensure
thattranscriptsbestreflectedmembers’contributions,theywere
dividedintothree10-minuteperiodspermeeting:beginning,
mid-meeting, and end-of-meeting.As a result, 22.5hof video were
transcribedforatotalof33,262“texts”,eachofwhichcorresponds
toaspeakingturn.
6.3. Influencetiestrength
Self-declaredtiestrengthwascapturedaftereachmeetingby
thefollowingquestion:“Inthelastmeeting,youwereinfluenced
by....”Acompletelistofgroupmemberswasprovidedand
fol-lowedbya3-pointLikert-typescale.Eachmembercouldusethis
scaletoevaluatehowmuchorhowoftenhe/shehadbeen
influ-encedbyeveryothermemberduringthemeeting:1=verylittle
orrarelyinfluenced;2=somewhatorsometimesinfluenced,and
3=greatlyorofteninfluenced.
We calculated the mean declared tie strength over the 9
meetingsforeachdyadwithineachgroup.Thecalculationwas
per-formedonorientedties(n=1454valid),i.e.theAline/Alexiadyad
wasconsidereddistinctfromtheAlexia/Alinedyad,forexample.
Strengthratingswerecategorisedaccordingtowhetherthey
werereciprocalornot.Giventhatthelowestrating“1”(weak
influ-enceties)canbeconsideredanexpressionofmereco-presence,
reciprocalratingsof“1”havebeenseparatedfromtheothers.When
membersofadyadusedtheratings2or3(influenceandstrong
influenceties),thetiewasconsideredreciprocal(seeTable1).
6.4. Centralityininfluencenetwork
Aswehaveseeninthepresentationofthestructuralapproach, degreecentralityhasbeenconsistentlyshownascorrelatedwith variousmeasuresof influence.As such,we canconfidently use centralityintheinfluencenetworkasaproxyforinfluence. Val-ueddegreecentralitiesininfluencenetworkswithineachgroup werecalculated.Centralitywasalsocalculatedfordichotomised matricesforeachstrengthratingandforcombinationsof2and
3(influence2).Becauseofthevariationingroupsize,meansof individualinfluencecentralitieswerenormalised.
6.5. Discoursesimilarity
Discoursesimilarityisoftenbasedontheuseofsimilarwords orexpressions(Carley,1993;MongeauandSaint-Charles,2014; Roth and Cointet, 2010) although it has been variously mea-sured.Forexample,similaritycanbederivedfromatopicanalysis (Blei and Lafferty, 2009; Danowski, 2011; Maier et al., 2015), automaticandmanualthematiccoding(Cohenetal.,2005; Saint-Charlesetal.,2014),semanticmapping(HolmbergandHellsten, 2016;LeydesdorffandHellsten,2006),cognitivemapping(Carley, 1997),QAPcorrelationofindividualsemanticnetworks(Basovand Brennecke,inpress;Danowski,2013)ordistributionalsemantics (FabreandLenci2015;Harris,1954).
Inthisstudy,discoursesimilarityisbasedonanalysisofvideo transcriptsofeachgroupmeeting.Textanalysisandcalculations usedtocreateadiscoursesimilarityindexwereperformedusing Sématosemanticanalysissoftware(Planteetal.,2005).
Adiscoursesimilarityindexwascalculatedbasedonan exami-nationofthesimilarityofwordsandnounphrasespresentwithin the transcribed text.Similarity was established using a French semanticthesaurusofsynonymsandsemanticproximities(Plante etal.,2005)togroupwordsandnounphrasesinthedatasetinto semanticfields.Asemanticfieldisacollectionofwordsornoun phrases with semantic affinities within a given dataset. Carley (1993)andRothandCointet(2010)suggestasimilartechnique togroupwordsandnounphrasesrepresentingthesameideato form“concept”.Forexample,inagivensetoftexts,thesemantic field“fight”mayincludethewords“confrontation”,“antagonism”, “battle”,“combat”,“conflict”,“opposition”,and“rivalry”ornoun phrasessuchas“siblingrivalry”.
A two-modematrix wascreatedby theties that linkgroup memberswiththesemanticfieldstheyused.Thesetiesarevalued accordingtothenumberoftimes anindividualemploysa par-ticularsemanticfield.Usingtheminimumvalueofallsemantic fieldssharedbyadyad,thismatrixisthenconvertedtoanactorx actormatrix.Forexample,ifAlineusesthesemanticfield“fight” 10timesduringthefirstmeetingwhileAlexiausesit20times,the dyadisattributed10“points”forthissemanticfield.This opera-tionisrepeatedforeachsharedsemanticfieldandthetotalisthen dividedbythetotal“points”foreachdyad.Giventhatsomegroup memberstalkmoreormoreoftenthanothers,frequency(number ofutterancesforeachmember)wasnormalisedandtotallengthof speechweightedusingthenumberofbitsinthetextfile.
Valuedcentralitywasthencalculatedforeachmeetingto pro-duceanindividualdiscoursesimilarityindex.Wealsocalculated themeansimilarityindexforeachindividualoverthe9meetings, aswellastheaveragediscoursesimilarityfordyads.
6.6. Groupstages
In order toexplore group development stages, we grouped observationsintothreeperiods.Becausetimingandmovingfrom onestagetoanothermayvaryfromgrouptogroup,wecreated threeequalperiodscontainingthreemeetingseach(thefirstthree, themiddlethree,andthelastthree).
7. Results
Hypothesis1whichstatesthatdiscoursesimilarityand influ-encetiestrengtharerelatedatthedyadiclevel,issupportedbythe observedcorrelationbetweenmeantiestrengthandaverage dis-coursesimilarityfordyadsoverthe9meetings.Thecorrelation,
Fig.1.Meandiscoursesimilarityaccordingtoreciprocityofinfluencetiestrength.
Table2
Correlationcoefficientsbetweeninfluencecentralityandsimilarityindex.
Influencecentrality Correlationwith individualmean similarityindex(n=34)
Correlationwith similarityindex (n=306) Allties(valued
centrality)
0.49** 0.30**
Stronginfluence(value 3)
0.57*** 0.35**
Influence(values2and 3)
0.40** 0.28**
Influence(value2) 0.33 0.11 Weakinfluence(value
1;co-presence)
−0.58*** −0.34***
**p<0.01;***p<0.001(two-tailedtest).
which is significant but moderated(r=0.42; p=0.000; n=200),
explains17.8%oftheobservedvariance.
Thishypothesisisalsosupportedbyasignificantdifferencein
averagediscoursesimilaritydependingonwhethertheinfluence
isreciprocal,non-reciprocal,orexpressesco-presence(dl=1451;
f=58.677;p<0.000).Fig.1showsthemeanlevelofsimilarityfor
eachcategory.
Hypothesis2whichstatesthatagroupmember’sdiscourse sim-ilaritywiththatofallothergroupmembersispositivelycorrelated with his or her valued degree centrality in the influence net-work,issupportedbyapositivebutmoderatecorrelationbetween meanvalueddegreecentralityandmeansimilarityindex(r=0.49; p<0.003;n=34)(Table2).
Totestwhetherthepositivecorrelationbetweenvalueddegree centrality anddiscourse similarity wasgreaterfor strong influ-encetiesthanforweakinfluenceties(Hypothesis3),wecalculated meandegreecentralityforthedichotomisedinfluencematricesand correlatedtheseresultswithmeandiscoursesimilarity(Table2). Resultsshowthatmeaninfluencecentralityismorestrongly corre-latedwithsimilarityindexforstrongerties(r=0.57;p<0.001)than fortiesofstrength2(r=0.33;p=n.s.).Moreover,moderate nega-tivecorrelationswereobservedbetweenmeandiscoursesimilarity andmeaninfluencecentrality(r=−0.58;p<0.001)forco-presence (weakinfluence)ties.Table2alsoshowscorrelationscalculatedfor all306observationsofall34peopleduring9meetings(ratherthan forthemean).Duetoweeklyvariability,correlationsarelowerbut similartowhatisobservedforthemeans.Strictlyspeaking,these resultsdonotsupportHypothesis3,whichhypothesisesaweaker butstillpositivecorrelation.Nonetheless,althoughthisresultis notentirelyconsistentwiththehypothesis,itisalignedwithitand providesanunexpectednuance.
Table3
Regressioncoefficientsforsimilarityasdependentvariable.
Influencevalue Beta(stderror) Model1 Stronginfluence(value3) 0.077(0.012)*** Model2 Stronginfluence(value3) 0.053(0.014)***
Weakinfluence(value1;co-presence)−0.0413(0.012)*** Unstandardisedcoefficients(stderror).**p<0.01;***p<0.001;(two-tailedtest).
Regression analysis shows that the best predictors of
dis-coursesimilarityresideattheextremesofmeasuredtiestrengths
(Table3).Itmeansthatthemoreonehasstrong(3)tieswithother members(model1)orthemoreheorsheonehasstrongtiesand thelessheorshehasco-presenceties(model2),themoreheor sheshareselementsofdiscoursewithothergroupmembers(or vice-versa).
We explored the existence of differences between low and highsimilarityindexes for valued influence centrality (without dichotomising).Todoso,wedividedvalidobservationsintotwo equal groups of low and high similarity indexes. Correlations betweendiscoursesimilarityand valued centralityinthe influ-encenetworkwithineachofthesegroupsarerelativelyweak,but significant.Inthesubgroupofthosewhosediscourseshowslittle similaritytothatofallgroupmembers,thereisapositive corre-lationwithvaluedcentrality(r=0.27;p<0.001;n=139),whilea negativecorrelation(r=−0.20;p<0.01;n=139)isobservedinthe subgroupofthosewhosediscourseishighlysimilartothatofall members.Inotherwords,withinthesubgroupoflowsimilarity, themoresimilaramemberistoallgroupmembers,thehigher hisorherinfluencecentrality,whiletheoppositeistrueforthe high-similaritysubgroup.Thisresultsuggeststhattheremaybea thresholdatwhichtherelationbetweendiscoursesimilarityand influencecentralityisreversed.
Totest whether thecorrelation betweencentrality and dis-course similarity weakened during the second stage of group development,asstatedby Hypothesis4, wecalculated correla-tionsbetweensimilarityindexesandcentralitiesforeachofthe threeperiods:P1–firstthreemeetings(n=92);P2–middlethree meetings(n=98);andP3–lastthreemeetings(n=88).Resultsshow partialsupportforHypothesis 4:forP1, similarityand central-itygrowsimultaneously(r=0.49;p<0.001);forP2, there isno significantcorrelationand,forP3,thecorrelationispositiveand significant,butlow(r=0.29;p<0.005).
Inordertobetterunderstandthis resultand consideringthe possibilityofathresholdatwhichtherelationbetween similar-ityandinfluencereversesdirection,wecreatedfiveequalgroups ofpercentile-basedobservationsforeachperiod.While maintain-ingasufficient number ofobservationsbygroups, thisdivision makesitpossibletoexaminethepotentialemergenceofasimilarity threshold.
Asexpected,therelationislinearduringP1(Fig.2a).ForP2, thenon-linearrelationexplainstheabsenceofasignificant corre-lation.Observationsabovethe80thsimilaritypercentileindicate lessinfluence(Fig.2b).Thecorrelationislinearbelowthis per-centile(r=0.41;p<0.001;n=60)andsimilartothatofP1(r=0.49; p<0.001; n=92), but negative above this threshold (r=−0.45; p<0.005;n=38).Thedifferencebetweenthemeansimilarityof thesetwogroupsissignificant(p<0.001;f=4.338;dl=37).ForP3, thislowcorrelationcanbeattributedtothepresenceoftwomain groups,onewithlowsimilarityandinfluence,theotherwithhigher similarityandinfluence.Themeandifferencebetweenthesegroups issignificant(p<0.001;dl=87;f=108.09)(Fig.2c).
8. Discussion
Oneofourstudy’sprimarycontributionsistobringempirical supporttotherelationbetweensocialnetworksanddiscourse sim-ilarity.Inthatwecomfortotherstudiesthathaveobtainedsimilar resultsinothercontextsandwithdifferentwaysofmeasuring dis-coursesimilarity(BasovandBrennecke,in press;Carley,1986a, 1991;CointetandRoth,2009;Danowski,2013;Roth,2008;Roth and Cointet,2010; Scherer andCho, 2003;Teoand Loosemore, 2011).Suchencouragingresultsareaninvitationtopursuethis lineofresearch.
Toourknowledge,thisisthefirsttimethatthistypeof longitudi-nalanalysishasbeenrealisedinworkgroupstobetterunderstanda communicationphenomenonsuchasinterpersonalinfluence.Also, fromamethodologicalpointofview,weareawareofnootherstudy usinga socio-semanticapproach toanalysediscoursesimilarity networksbasedonverbalinteractionsbetweengroupmembers1−
acommondatasourcefordiscourseanalysisinqualitativemethods (Holmes,2009).Ourstudydemonstratesthepertinenceof quan-titativeanalysisofverbalinteractionsasawaytouncoverother aspectsofthelinksbetweendiscourseandrelationships.
Asecondcontributionofthisstudyliesintheimportancegiven totiestrength(AralandWalker,2014;Bondetal.,2012).Under a moderatecorrelationbetweenvaluedcentralities ininfluence networksandsimilarityindexes,distinctphenomenaarehidden: thedirectionofthecorrelation isreversed betweenthose with numerous strong influence ties (r=0.57) and those with many co-presence(weakinfluence)ties(r=−0.58).Combinedwithour resultatthedyadiclevel,whichshowstheimportanceofstrength andreciprocityfordiscoursesimilarity,thissupportstheneedto considertiestrengthinthestudyofsocialinfluence.More impor-tantly,itopensnewavenuestodeepenourunderstandingofthe factorsinplaywhenonereachesacentralpositioninaworkgroup influencenetwork.Becauseourstudywasconductedwith zero-history,leaderless,highlyhomogeneousgroupswithregardtoage, genderandethnicity,weknowthatneitherpre-existingstructure norstatusdifferenceswerepresent.Thismayimplythat seman-tichomophily(RothandCointet,2010)playedaroleearlyinthe group’slifeandcontributedtothedevelopmentofstrong influ-encerelationshipsbetweensemanticallysimilargroupmembers, thereby“excluding”othersintheprocess.Usingthefrequencyof contactsasaproxyfortiestrength,Basovetal.(2017)haveobtained slightlydifferentresults.Indeed,theyfoundthatindividualswith thehighestfrequency ofcontactshave a relativelylowlevel of conceptssharingwhilethosewithlessfrequencyofcontactshave atendencytobridgecontainbetweensubgroups.Withregardto socialinfluence,thismaysuggesttheonesweinteractwith regu-larlyarenotnecessarilytheoneshavingthebiggestinfluenceon theevolutionofourdiscourse.
Alternativelytothestrengthoftieexplanation,individual char-acteristicsassociatedwithinfluenceintheliterature,suchhasa largerepertoire of communication strategies (Bass,1990), may havefavouredthecentralityofgroupmemberspossessingsuch arepertoireandrenderedtheirdiscoursemore“adoptable”. Con-versely, given that “perceived listening effectiveness” hasbeen showntoaffectemergingleadershipinworkgroups(Johnsonand Bechler,1998),itmaybethatgroupmembersbecamemore cen-tralintheinfluencenetworkbecausetheywereusingthewordsof otherswho,asaresult,feltlistenedto.
1Basovetal.(2017)dataincludesverbalexpressionsofgroupmembersextracted fromamixofinterviewsandethnographicobservationsofdyadicconversations combinedwithvarious“publications”madebymembers(posts,journalarticles, proseandpoetry),adifferentdatasetthanverbatimtranscriptofexchangesbetween membersduringworksessions.
Fig.2. a,b,c.Influencevaluedcentralityandfivepercentilegroupsofsimilarityindexesforthreeperiods.
Ourthirdandfourthcontributionsarebasedonalongitudinal exploratoryanalysis.Firstly,theresultsofthis studyshowthat therelationbetweeninfluencenetwork centralitiesand similar-ityindexesinworkgroupschangesthroughtimeinanon-linear mannerthat appears tofollow thethree main stages of group development.Duringthefirststage,thepositivecorrelationmay beexplainedbythefactthatthosewiththehighestsimilaritymay beperceivedasmore“prototypical”(Brownetal.,2004;Hoggetal., 2004)ortransformational(Jordan,2005;Meda,2005)becausethe wordsandexpressionsusedaresimilartothoseemployedbyother groupmembers.Atthesecondstage,theanalysissuggestsa “sim-ilaritythreshold”aroundthe80thpercentile,abovewhichgreater similarityis associated withlessinfluence. Thisconstitutes our fourthcontribution.ThisthresholdcouldbeparalleledtoCointet andRoth(2009)studyofbloggersinwhichthemost“influential” blogsseemed tobethosewiththelargerreadership − uptoa pointas“aboveacertainthreshold,theincreaseininfluenceis flat-ter,althoughstillrelativelyincreasing”(p.6).Withregardtofixed durationworkgroups,thesupposedthresholdappearsduringthe midtermmeetings,hereunderstoodasthesecondstageofgroup developmentmarked by debates, confrontations, and decision-makingregardingthetaskathand.Influenceand similarityare positivelycorrelatedbelow thisthresholdand negatively corre-latedaboveit.Weproposethatsimilaritymustbehighenoughfor peopletobe“in”thegroup,butlowenoughfortheirdifferenceto standout;originalelementsintheirdiscoursedistinguishthem. Ina debate,differencemaybeattractive(Robertand Mongeau, 2014).Thisassumptionisalsoconsistentwiththeexpectancy vio-lationtheoryproposedbyBurgoonandBacue(2009)inthestudyof non-verbalcommunication.Thistheorystatesthatbehavioursthat deviateslightlyfromthenormattractattention.Ourresultssuggest thepotentialgeneralisationofthistheorytoverbalcommunication. Alternatively,inhertheoryofstabilityinsocialgroup(Carley,1991) postulatesthataperfectlystablegroupwouldbecome unproduc-tiveforlackofnewinformationtoexchange.Inviewofthis,the similaritythresholdcouldbeinterpretedasamanifestationofa groupautoregulationmechanism:abovethesimilaritythreshold members’attentionshouldshiftfromsimilarelementstomore dissimilaronesastoomuchsimilarityreducesthemotivationto exchangeinformation.
Finally,thedivisionintotwogroupsatthelaststagemaybe interpretedasaphenomenonofexclusionofthelesssimilarand lesscentralmembersintheinfluencenetwork.Indeed,onecould postulatethat duringstage2,themostcentralmembersinthe influencenetworkwerethosemoreactivelyparticipatinginthe debates−continuingtheconversationuntiltheyreachacertain levelofconvergencethatexcludesothersfromthenew,negotiated, discourse.
Letus reiteratethat therelationbetweeninfluenceand dis-courseisnot seen hereinterms oflinear causality.Adifferent discourseintheearlystagesofagroup’slifecouldleadtohavingless influence,whichcouldinturnhelpmaintainadistinctdiscourse.
Conversely,holdingasimilardiscoursecouldsupportthecreation ofstronginfluenceties,whichmaythenhelpmaintaindiscourse similarity.
9. Conclusion
Thegoalofourstudywastoverifythehypothesisofa rela-tionbetweenworkgroupmembers’influenceandtheirdiscourse similaritywiththatofothergroupmembers.Todoso,we exam-inedtiestrength,actorcentralityininfluencenetwork,andshared discourseelementsinvideotranscriptsof453-hgroupmeetings andweeklysociometricquestionnaires.Wealsoexploredthe evo-lutionoftherelationbetweeninfluencecentralityanddiscourse similarityovertime.
Adoptingasocio-semanticperspective,ourstudycontributes inseveralwaystothejointapplicationofstructuralanddiscursive approaches.Amongstthestudiesbringingtogethersocialnetworks anddiscoursesimilarityitis,toourknowledge,theonlystudyto dothisforverbalexchangesinworkgroups,andwithafocuson socialinfluencefromalongitudinalperspective.
Ourresultsshowthatinfluencenetworkcentralityiscorrelated withdiscoursesimilaritynetworkcentrality(i.e.havingadiscourse thatmostresemblesthatofallgroupmembers).Atboththedyadic andgrouplevels,discoursesimilarityishigherforstrong(and recip-rocal)ties.Theseresultsalsosuggestthatinfluenceanddiscourse sharingco-evolvethroughtimeandgroupmembers’interactions. Atthebeginningofthegroup’slife, therelationbetween influ-enceandsimilarityislinearandpositive,butatthemidtermthis relationbecomesnegativeforthosewiththehighestcentralityin thediscoursesimilaritynetwork.Inthefinalmeetings,wenoted thepresenceoftwosubgroups(onewithhighsimilarityandhigh centralitymembersandtheotherwithlowsimilarityandlow influ-ence);althoughdifferencebetweenthesesubgroupsissignificant, differencewithinthemisnot.Thesechangesthroughtimehaveled ustopostulatetheexistenceofasimilaritythresholdbelowwhich similaritybringsgreaterinfluenceandabovewhichoriginalityand differencearerequired.
These findingsare theresult ofcapturing and analysingthe verbal interactions of members in homogeneous, zero-history, leaderless,fixed-duration,bonafidegroupsoverasufficientlylong period.Thestrengthofthisdatasetisalsoa limitation:because thesearestudentgroups,therearenotmanygroupsinother orga-nisationswhosharetheircharacteristics2orwhowouldbewilling
tohave allof theirmeetingsrecorded.Because observationsin ourdatavaryconsiderablyfromoneweektothenextandfrom onegrouptoanother,longerperiodsandmoregroupswouldbe required, butcapturing and transcribingverbalinteractions are timeconsumingandcostly.Othermeansarethereforecalledfor.
2Groupsof“founders”maywellhavesuchcharacteristics,atleastintermsof homogeneityand“leaderless”.
Althoughoneadvantageofthesocio-demographic homogene-ityofourgroupsisthatitlimitstheimpactofvariablessuchas statusandgender,thesearenon-negligibleintheprocessofsocial influence.Forinstance,studiestakinggenderintoaccountshow differencesbetweenmenandwomeninterms ofinfluenceand accordingtocommunicationstyle(Carli,2001;Eaglyetal.,2003). Gendermayalsoinfluencetheuseofcertainsemanticfields.Carley (1986b)hasshownthat,overtime,menadoptconceptsfirst pro-posedbywomen.Thegroupsstudiedherewerepredominantly composedofwomen;acomparativeanalysisbasedongenderwas thereforenotpossible.Aswehaveseen,studieslinkingcentrality toinfluencehavealsodemonstratedtheimpactofothervariables, includingcommunicationstrategiesandpoliticalskills, personal-ity,emotionalabilities,andethnicity.Thevarianceexplainedbyour results,whichfluctuatesbetween16%and34%,isnotable consid-eringthemultitudeoffactorspotentiallyatplay.Thisconfirmsthe relevanceofstudyingtheinterplaybetweendiscourseand influ-encenetworksandalsoraisesquestionsaboutthecombinedeffect ofmultiplefactors.
Regardingsemanticdata,ourstudyisbasedonsemanticfields automaticallyextractedbyatextanalysissoftwareusinga seman-ticthesaurus.Differentwaystomeasuresimilaritycouldleadto differentresults.Forinstance,analysesmorefocusedonmeaning, suchasthematicortopiccategorisation,maynarrowtheresults andmakeitpossibletodistinguishbetween,forexample,usingthe samewordstoexpressagreementordisagreement.Suchan anal-ysiswouldalsopermittheexplorationofsharedcognitiveframes. Anotherpotentiallineofinquiryistheanalysisofverbalstrategies usedbygroupmembers,giventhatleadershipemergencehasbeen linkedtostrategieslikesummarising,orientinggroupprocesses, andseekingevaluation(Barge,1989).
Toconclude,ourresultsreveala newavenue forclosingthe gapbetweentiesasconduitsandtheflowofcontent(Borgattiand Foster,2003)throughthestudyofgroupsimilarityandopennew perspectivesfor studyingtheemergenceofsocial influenceand leadershipinworkgroups.
Acknowledgement
ThisworkwassupportedbytheSocialSciencesandHumanities ResearchCouncilofCanada[grantnumbers410-2007-1438& 820-2008-1054].
References
Alvesson,M.,Karreman,D.,2000.Varietiesofdiscourse:onthestudyof organizationsthroughdiscourseanalysis.Hum.Relat.53,1125–1149. Aman,I.,2009.Discourseandstrivingforpower:ananalysisofBarisanNasional’s
2004Malaysiangeneralelectionmanifesto.DiscourseSoc.20,659–684. Antonakis,J.,Cianciolo,A.T.,Sternberg,R.J.,2004.Leadership:past,present,and
future.In:Antonakis,J.,Cianciolo,A.T.,Sternberg,R.J.(Eds.),TheNatureof Leadership.ThousandOaks:SagePublications,Californie,pp.125–170. Aral,S.,Walker,D.,2014.Tiestrength,embeddedness,andsocialinfluence:a
large-Scalenetworkedexperiment.Manage.Sci.60,1352–1370.
Arrow,H.,Poole,M.S.,Henry,K.B.,Whellan,S.,Moreland,R.,2004.Time,change, anddevelopment:thetemporalperspectiveongroups.SmallGroupRes.35, 73–105.
Asmuß,B.,Svennevig,J.,2009.Meetingtalk.J.Bus.Commun.46,3–22.
Astley,W.G.,Sachdeva,P.S.,1984.Structuralsourcesofintraorganizationalpower: atheoreticalsynthesis.Acad.Manage.Rev.9,104–113.
Balogun,J.,Jacobs,C.,Jarzabkowski,P.,Mantere,S.,Vaara,E.,2014.Placingstrategy discourseincontext:sociomateriality,sensemaking,andpower:placing strategydiscourseincontext.J.Manage.Stud.51,175–201.
Barberio,V.,Monti,A.,2014.Réseauxsémantiquesetlégitimédudiscours organisationnel:uneillustrationempirique.Communiquer.Revue internationaledecommunicationsocialeetpublique,7–25.
Barge,J.K.,Hirokawa,R.Y.,1989.Towardacommunicationcompetencymodelof groupleadership.SmallGroupBehav.20,167–189.
Barge,J.K.,1989.Leadershipasmedium:aleaderlessgroupdiscussionmodel. Commun.Q.37,237–247.
Barge,J.K.,1996.Leadershipskillsanddialecticsofleadershipingroupdecision making.In:Hirokawa,R.Y.,Poole,M.S.(Eds.),CommunicationandGroup DecisionMaking.Sage,London,pp.301–342.
Barske,T.,2009.Sametoken,differentactions:aconversationanalyticstudyof socialroles,embodiedactions,andokingermanbusinessmeetings.J.Bus. Commun.46,120–149.
Basov,N.Brennecke,J.DualitybeyondDyads:multiplexpatterningofsocialties andculturalmeaningsRes.Sociol.Organ.53:(inpress).
Basov,N.,Lee,J.-S.,Antoniuk,A.,2017.Socialnetworksandconstructionofculture: asocio-Semanticanalysisofartgroups.In:Cherifi,H.,Gaito,S.,Quattrociocchi, W.,Sala,A.(Eds.),ComplexNetworks&TheirApplicationsV:Proceedingsof the5thInternationalWorkshoponComplexNetworksandTheirApplications (COMPLEXNETWORKS2016).SpringerInternationalPublishing,Cham,pp. 785–796.
Bass,B.M.,1990.Bass&Stogdill’sHandbookofLeadership:Theory,Research,and ManagerialApplications.FreePress,NewYork.
Bavelas,A.,1950.Communicationpatternsintask-orientedgroups.J.Acoust.Soc. Am.22,725–730.
Bergvall,V.L.,Remlinger,K.A.,1996.Reproduction,resistanceandgenderin educationaldiscourse:theroleofcriticaldiscourseanalysis.DiscourseSoc.7, 453–479.
Blei,D.,Lafferty,J.D.,2009.Topicmodels.Textmining:classification,clustering, andapplications10,34.
Boje,D.M.,Whetten,D.A.,1981.Effectsoforganizationalstrategiesandcontextual constraintsoncentralityandattributionsofinfluenceininterorganizational networks.Adm.Sci.Q.26,378.
Bond,R.M.,Fariss,C.J.,Jones,J.J.,Kramer,A.D.I.,Marlow,C.,Settle,J.E.,Fowler,J.H., 2012.A61-million-personexperimentinsocialinfluenceandpolitical mobilization.Nature489,295–298.
Borgatti,S.P.,Foster,P.C.,2003.Thenetworkparadigminorganizationalresearch: areviewandtypology.J.Manage.29,991–1013.
Braaten,L.J.,1974.Developmentalphasesofencountergroupsandrelated intensivegroups.InterpersonalDev.,112–129.
Brass,D.J.,Burkhardt,M.E.,1993.Potentialpowerandpoweruse:aninvestigation ofstructureandbehavior.Acad.Manage.J.36,441–470.
Brass,D.j.,Krackhardt,D.M.,2012.Politicsinorganizations:theoryandresearch considerations.In:Ferris,G.R.,Treadway,D.C.(Eds.),SIOPOrganizational FrontiersSeries.Routledge,Taylor&FrancisGroup,NewYork,pp.355–375. Brass,D.J.,Labianca,G.,1999.Socialcapital,thesocialledger,andsocialressources
management.In:Leenders,R.,Gabbay,S.(Eds.),CorporateSocialCapitaland Liability.KluwerAcademicPublishers,Norwell(MA),pp.323–338.
Brass,D.J.,1985.Men’sandwomen’snetworks:astudyoninteractionpatternsand influenceinanorganization.Acad.Manage.J.28,327–343.
Brown,D.J.,Scott,K.A.,Lewis,H.,2004.Informationprocessingandleadership.In: Antonakis,J.,Cianciolo,A.T.,Sternberg,R.J.(Eds.),TheNatureofLeadership. ThousandOaks:SagePublications,Californie,pp.125–170.
Brummans,B.H.J.M.,Putnam,L.L.,Gray,B.,Hanke,R.,Lewicki,R.J.,Wiethoff,C., 2008.Makingsenseofintractablemultipartyconflict:astudyofframingin fourenvironmentaldisputes.Commun.Monogr.75,25–51.
Burgoon,J.K.,Bacue,A.E.,2009.Nonverbalcommunicationskills.In:Greene,J.O., Burleson,B.R.(Eds.),HandbookofCommunicationandSocialInteractionSkills. Routledge,NewYork.
Carley,K.M.,1986a.Anapproachforrelatingsocialstructuretocognitive structure.J.Math.Sociol.12,137–189.
Carley,K.M.,1986b.Knowledgeacquisitionasasocialphenomenon.Inst.Sci.14, 381–438.
Carley,K.M.,1991.Atheoryofgroupstability.Am.Sociol.Rev.56,331–354. Carley,K.,1993.Codingchoicesfortextualanalysis:acomparisonofcontent
analysisandmapanalysis.Sociol.Methodol.23,75.
Carley,K.M.,1997.Anapproachforrelatingsocialstructuretocognitivestructure. J.OrganizationalBehav.18,533–558.
Carli,L.L.,2001.Genderandsocialinfluence.J.Soc.Issues57,725–741. Carter,D.R.,DeChurch,L.A.,Braun,M.T.,Contractor,N.S.,2015.Socialnetwork
approachestoleadership:anintegrativeconceptualreview.J.Appl.Psychol. 100,597–622.
Chang,A.,2006.Understandingthemultidimensionalityofgroupdevelopment. SmallGroupRes.37,327–350.
Chidambaram,L.,Bostrom,R.,1996.Groupdevelopment(I):areviewand synthesisofdevelopmentmodels.GroupDecis.Negotiation6,159–187. Choi,S.,Schnurr,S.,2014.Exploringdistributedleadership:solvingdisagreements
andnegotiatingconsensusina‘leaderless’team.DiscourseStud.16,3–24. Clare,S.,Krogman,N.,Caine,K.J.,2013.Thebalancediscourse:acasestudyof
powerandwetlandmanagement.Geoforum49,40–49.
Clifton,J.,2009.Beyondtaxonomiesofinfluence:doinginfluenceandmaking decisionsinmanagementteammeetings.J.Bus.Commun.46,57–79. Codo,E.,2011.Regimentingdiscourse,controllingbodies:disinformation,
evaluationandmoralcategorizationinastatebureaucraticagency.Discourse Soc.22,723–742.
Cohen,S.G.,Bailey,D.E.,1997.Whatmakesteamswork:groupeffectiveness researchfromtheshopFIoortotheexecutivesuite.J.Manage.23,239–290. Cohen,H.,Lefebvre,C.,Meunier,J.-G.,Forest,D.,Biskri,I.,2005.Classificationand
categorizationincomputerassistedreadingandanalysisoftexts.In: HandbookofCategorizationinCognitiveScience.Elsevier,Amsterdam(pp. 955–978%@0080446124).
Cointet,J.-P.,Roth,C.,2009.Socio-semanticDynamicsinaBlogNetwork.,pp. 114–121.
Coleman,J.,Katz,E.,Menzel,H.,1957.Thediffusionofaninnovationamong physicians.Sociometry20,253–270.
Craig,R.T.,1999.CommunicationTheoryasaField.CommunicationTheory9, 119–161.2009(versionfranc¸aise)Lacommunicationentantquechamp d’études,Revueinternationaledecommunicationsocialeetpublique.2001 (2001),pp.2001–2042.
Curry,O.,Dunbar,R.I.M.,2013.Dobirdsofafeatherflocktogether?:the relationshipbetweensimilarityandaltruisminsocialnetworks.Hum.Nat.24, 336–347.
Danowski,J.A.,2011.SocialNetworksBasedonSemanticNetworkSimilarity: AssessmentofLatentDirichletApproximation(LDA)TopicModelingtoIndex Persons’Similarity.INSNA—SunbeltXXXI,St-Pete,Florida.
Danowski,J.A.,2013.ChangeOverTimeinICADivisionNetworksBasedon SemanticSimilarityofPapersPresentedVersusCo-Memberships.In:ICA(Ed.). InternationalCommunicationAssociation,London,England.
Day,D.,Antonakis,J.,2012.Leadership:past,present,andfuture.In:TheNatureof Leadership.SagePublications,pp.3–25.
Dijk,T.A.V.,2012.Knowledge,discourseanddomination.In:Meeuwis,M., ¨Ostman, J.-O.(Eds.),PragmaticizingUnderstanding.StudiesforJefVerschueren.John BenjaminsPublishingCompany,pp.151–196.
Dunphy,D.C.,1968.Phases,roles,andmythsinself-analyticgroups.J.Appl.Behav. Sci.4,195–225.
Duval,A.-M.,Gendron,Y.,Roux-Dufort,C.,2015.Exhibitingnongovernmental organizations:reifyingtheperformancediscoursethroughframingpower. Crit.Perspect.Acounting29,31–53.
Eagly,A.H.,Johannesen-Schmidt,M.C.,vanEngen,M.L.,2003.Transformational, transactional,andlaissez-Faireleadershipstyles:ameta-Analysiscomparing womenandmen.Psychol.Bull.129,569–591.
Edley,N.,Wetherell,M.,1997.Jockeyingforposition:theconstructionof masculineidentities.DiscourseSoc.8,203–217.
El-Khatib,R.,Fogel,K.,Jandik,T.,2015.CEOnetworkcentralityandmerger performance.J.FinancialEcon.116,349–382.
Ellis,D.G.,1999.FromLanguagetoCommunication,2nded.LawrenceErlbaum Associates,Mahwah,N.J.
Emery,C.,2012.Uncoveringtheroleofemotionalabilitiesinleadership emergence:alongitudinalanalysisofleadershipnetworks.Soc.Networks34, 429–437.
Ezeifeka,C.R.,Osakwe,N.N.,2013.Genderrepresentationinthe1999Nigerian constitution:acriticaldiscourseanalysisforsocio-politicalequity.Discourse Soc.24,687–700.
Fabre,C.,Lenci,A.,2015.DistributionalSemanticsToday,vol.56.TAL,pp.7–20. Fairclough,N.,1989.LanguageandPower.Longman,London;NewYork. Fairclough,N.,2003.AnalysingDiscourse:TextualAnalysisforSocialResearch.
Routledge,London;NewYork.
Fairhurst,G.T.,Connaughton,S.L.,2014.Leadership:acommunicativeperspective. Leadership10,7–35.
Fairhurst,G.T.,Cooren,F.,2009.Leadershipasthehybridproductionof presence(s).Leadership5,496–490.
Faust,K.,2011.Animalsocialnetworks.In:Scott,J.,Carrington,P.J.(Eds.),TheSAGE HandbookofSocialNetworkAnalysis.SAGE,London;ThousandOaks,Calif(p. 622).
Friedkin,N.E.,Johnsen,E.C.,2011.SocialInfluenceNetworkTheory:ASociological ExaminationofSmallGroupDynamics.CambridgeUniversity,PressNewYork. Friedkin,N.E.,1993.Structuralbasesofinterpersonalinfluenceingroups:a
longitudinalstudy.Am.Sociol.Rev.58,861–872.
Guastello,S.J.,2007.Non-lineardynamicsandleadershipemergence.Leadership Quarterly18,357–369.
Hardy,C.,Maguire,S.,2016.Organizingrisk:discourse,power,andriskification. Acad.Manage.Rev.41,80–108.
Hare,A.P.,2003.Roles,relationships,andgroupsinorganizations:some conclusionsandrecommendations.SmallGroupRes.34,123–154. Harris,Z.S.,1954.DistributionalStruct.10,146–162.
Hogg,M.A.,Abrams,D.,Otten,S.,Hinkle,S.,2004.Thesocialidentityperspective: intergrouprelations,self-Conception,andsmallgroups.SmallGroupRes.35, 246–276.
Holmberg,K.,Hellsten,I.,2016.Integratinganddifferentiatingmeaningsin tweetingaboutthefifthIntergovernmentalPanelonClimateChange(IPCC) report.FirstMonday21.
Holmes,J.,2009.Languageintheworkplace.In:LiteratureReview.Victoria UniversityofWellington,VictoriaNewZealand.
Huffaker,D.,2010.Dimensionsofleadershipandsocialinfluenceinonline communities.Hum.Commun.Res.36,593–617.
Ibarra,H.,1993a.Networkcentrality,powerandinnovationinvolvement: determinantsoftechnicalandadministrativeroles.Acad.Manage.J.38, 471–501.
Ibarra,H.,1993b.Personalnetworksofwomenandminoritiesinmanagement:a conceptualframework.Acad.Manage.Rev.18,56–87.
Jian,G.,2008.Fixingthemeaningofchange:analysisofchangemanagement meetingexcerpts.In:The58thAnnualInternationalCommunication AssociationConference,Montréal.
Johnson,S.D.,Bechler,C.,1998.Examiningtherelationshipbetweenlistening effectivenessandleadershipemergence:perceptions,behaviors,andrecall. SmallGroupRes.29,452–471.
Johnson,A.R.,vandeSchoot,R.,Delmar,F.,Crano,W.D.,2015.Socialinfluence interpretationofinterpersonalprocessesandteamperformanceovertime usingbayesianmodelselection.J.Manage.41,574–606.
Jordan,D.A.,2005.APhenomenologicalStudyofTranscendentLeadersin Healthcare.MedicalUniversityofSouthCarolina−CollegeofHealth Professions,SouthCarolina(p.395).
Kangasharju,H.,2002.Alignmentindisagreement:formingoppositionalalliances incommitteemeetings.J.Pragmatics34,1447–1471.
Kilby,L.,Horowitz,A.D.,2013.Openingupterrorismtalk:thesequentialand categoricalproductionofdiscursivepowerwithinthecallopeningsofatalk radiobroadcast.DiscourseSoc.24,725–742.
Kotwal,A.,Power,K.,2015.Eatingwords:adiscoursehistoricalanalysisofthe publicdebateoverIndia’s2013NationalFoodSecurityAct.OnHorizon23, 174–189.
Krackhardt,D.,1990.Assessingthepoliticallandscape:structure,cognitionand powerinorganizations.Adm.Sci.Q.,342–369.
Krinsky,J.,2010.Dynamicsofhegemony:mappingmechanismsofculturaland politicalpowerinthedebatesoverworkfareinNewYorkCity,1993–1999. Poetics38,625–648.
Labianca,G.,Brass,D.J.,2006.Exploringthesocialledger:negativerelationships andnegativeasymmetryinsocialnetworksinorganizations.Acad.Manage. Rev.31,596–614.
Lazzaro-Salazar,M.V.,Marra,M.,Holmes,J.,Vine,B.,2015.Doingpowerand negotiatingthroughdisagreementinpublicmeetings.PragmaticsSoc.6, 444–464.
Leavitt,H.J.,1951.Someeffectsofcertaincommunicationpatternsongroup performance.J.AbnormalSoc.Psychol.,38–50.
Leydesdorff,L.,Hellsten,I.,2006.Measuringthemeaningofwordsincontexts:an automatedanalysisofcontroversiesabout‘Monarchbutterflies,’
‘Frankenfoods,’and‘stemcells’.Scientometrics67,231–258.
Maier,D.,Waldherr,A.,Miltner,P.,Jaehnichen,P.,Pfetsch,B.,2015.Exploringissues inanetworkedpublicsphere.Combininghyperlinknetworkanalysisandtopic modeling.In:Atteveldt,W.V.(Ed.),BringingTogetherSocialandSemantic NetworksinCommunicationResearch.CACommunication&Technology, PoliticalCommunication&MassCommunicationDivisions,PuertoRico(p.36). Mayes,P.,2010.Thediscursiveconstructionofidentityandpowerinthecritical
classroom:implicationsforappliedcriticaltheories.DiscourseSoc.21, 189–210.
McGrath,J.E.,Arrow,H.,Berdahl,J.L.,2000.Thestudyofgroups:past,present,and future.Pers.Soc.Psychol.Rev.4,95–105.
McGrath,J.E.,1991.Time,interaction,andperformance(TIP):Atheoryofgroups. SmallGroupRes.22,147–174.
McKenna,E.F.,2016.Thediscourseofdeferenceanditsimpactontourist-Host powerrelations.J.TravelRes.55,555–565.
McPherson,J.M.,Smith-Lovin,L.,1987.Homophilyinvoluntaryorganizations: statusdistanceandthecompositionofface-to-Facegroups.Am.Sociol.Rev. 52,370–379.
McPherson,M.,Smith-Lovin,L.,Cook,J.M.,2001.Birdsofafeather.Homophilyin SocialNetworks27,415–444.
Meda,A.K.,2005.TheSocialConstructionofEthicalLeadership.Benedictine University%Uinternal-pdf://Meda,2005–1142137344/Meda,2005.pdf,Illinois. Melamed,D.,Savage,S.V.,2016.Status,factionsizes,andsocialinfluence:testing
thetheoreticalmechanism.A.J.S122(July).
Menz,F.,Al-Roubaie,A.,2008.Interruptions,statusandgenderinmedical interviews:theharderyoubrake:thelongerittakes.DiscourseSoc.19, 645–666.
Mills,T.M.,1967.TheSociologyofSmallGroups.Prentice-Hall,EnglewoodCliffs, NJ.
Monge,P.R.,Contractor,N.S.,2003.TheoriesofCommunicationNetworks.Oxford UniversityPress,NewYork.
Mongeau,P.,Saint-Charles,J.,2011.Approchescommunicationnellesdesgroupes danslesorganisations.In:Grosjean,S.,Bonneville,L.(Eds.),Communication Organisationnelle:Approches,ProcessusEtEnjeux.Chenelière,Montréal(p. 346).
Mongeau,P.,Saint-Charles,J.,2014.Centralitéderéseauxetsimilitudedediscours: uneapprochesociosémantiqueduleadershipémergentdanslesgroupesde travail.Communiquer.Revuedecommunicationsocialeetpublique,121–139. Moscovici,S.,1988.Lamachineàfairedesdieux:sociologieetpsychologie.Fayard,
Paris.
Nielsen,M.F.,2009.Interpretativemanagementinbusinessmeetings: understandingmanagers’interactionalstrategiesthroughconversation analysis.J.Bus.Commun.46,23–56.
Oswick,C.,Grant,D.,Marshak,R.J.,WolframCox,J.,2010.Organizationaldiscourse andchange:positions,perspectives,progressandprospects.J.Appl.Behav.Sci. 46,8–15.
Owen-Smith,J.,Cotton-Nessler,N.C.,Buhr,H.,2015.Networkeffectson organizationaldecision-making:blendedsocialmechanismsandIPO withdrawal.Soc.Networks41,1–17.
Pavitt,C.,2014.Aninteractiveinput-process-outputmodelofsocialinfluencein decision-makinggroups.SmallGroupRes.45,704–730.
Plante,P.,Dumas,L.,Plante,A.2005.Sémato.Logicield’analysesémantiquedes documentstextuels.UQAM,p.Logicield’analysesémantiquedesdocuments textuels.
Poole,M.S.,DeSanctis,G.,1990.Understandingtheuseofdecisionsupport systems:thetheoryofadaptivestructuration.In:Ful,J.,Steinfield,C.(Eds.), OrganizationsandCommunicationTechnology.Sage,NewburyParkCA,pp. 175–195.
Postmes,T.,Haslam,S.A.,Swaab,R.I.,2005.Socialinfluenceinsmallgroups:an interactivemodelofsocialidentityformation.Eur.Rev.Soc.Psychol.16,1–42.
Prego-Vazquez,G.,2007.Frameconflictandsocialinequalityintheworkplace: professionalandlocaldiscoursestrugglesinemployee/customerinteractions. DiscourseSoc.18,295–335.
Rice,R.E.,1993.Usingnetworkconceptstoclarifysourcesandmechanismsof socialinfluence,in:W.Richards,J.,&G.Barnett(Ed.),Progressin
communicationsciences:Advancesincommunicationnetworkanalysis.Ablex Norwood,NJ,pp.43–52.
Robert,P.F.,Mongeau,P.,2014.Caractéristiquessociosémantiquesdesméthodes conventionnellesetalternativesdedélibération:communiquer.Revuede communicationsocialeetpublique,101–120.
Rorty,R.,1967.TheLinguisticTurn:RecentEssaysinPhilosophicalMethod. UniversityofChicagoPress,Chicago.
Roth,C.,Cointet,J.-P.,2010.Socialandsemanticcoevolutioninknowledge networks.Soc.Networks32,16–29.
Roth,C.,2008.Coévolutiondesauteursetdesconceptsdanslesréseaux épistémiques:lecasdelacommunauté«zebrafish».Revuefranc¸aisede sociologie49,523.
Roth,C.,2013.Socio-Semanticframeworks.Adv.ComplexSyst.WorldSci.,16. Roussiau,N.,Bonardi,C.,2001.LesReprésentationsSociales—ÉtatDesLieuxEt
Perspectives.Mardaga,Sprimont.
Saint-Charles,J.,Mongeau,P.,2009.Differentrelationshipsforcopingwith ambiguityanduncertaintyinorganizations.SocialNetworks,33–39. Saint-Charles,J.,Surette,C.,Parkes,M.W.,Morrison,K.E.,2014.Connectionsfor
health,ecosystemsandsocietyleadingtoactionandchange.Ecohealth11, 279–280.
Salk,J.E.,Brannen,M.Y.,2000.Researchnotes.nationalculture,networks,and individualinfluenceinamultinationalmanagementteam.Acad.Manage.J.43, 191–202.
Scherer,C.W.,Cho,H.,2003.Asocialnetworkcontagiontheoryofriskperception. RiskAnal.23,261–267.
Schutz,W.C.,1958.FIRO:AThree-DimensionalTheoryofInterpersonalBehavior. HoltRinehart,NewYork.
Simpson,J.A.,Farrell,A.K.,Ori ˜na,M.M.,Rothman,A.J.,2015.Powerandsocia.In: luencein,relationships.,in:Mikulincer,M.,Shaver,P.R.,Simpson,J.A.,Dovidio, J.F.(Eds.),APAHandbookofPersonalityandSocialPsychology,Volume3: InterpersonalRelations.AmericanPsychologicalAssociation,Washington,pp. 393–420.
Sparrowe,R.T.,Liden,R.C.,2005.Tworoutestoinfluence:integrating leader-Memberexchangeandsocialnetworkperspectives.Adm.Sci.Q.50, 505–535.
St-Arnaud,Y.,1978.In:Morin,G.(Ed.),Lespetitsgroupes:participationet communication,2002.
Teo,M.M.,Loosemore,M.,2011.Community-basedprotestagainstconstruction projects:acasestudyofmovementcontinuity.Constr.Manage.Econ.29, 131–144.
Thayer,L.,1988.Leadership/communication:acriticalreviewandamodest proposal.In:Goldhaber,G.M.,Barnett,G.A.(Eds.),HandbookofOrganizational Communication.AblexNorwood,NJ,pp.231–263.
Toft,A.,2014.Contestingthedeviantother:discursivestrategiesforthe productionofhomelesssubjectivities.DiscourseSoc.25,783–809. Treadway,D.C.,Breland,J.W.,Williams,L.M.,Cho,J.,Yang,J.,Ferris,G.R.,2013.
Socialinfluenceandinterpersonalpowerinorganizations:rolesof performanceandpoliticalskillintwostudies.J.Manage.39,1529–1553. Tuckman,B.W.,Jensen,M.A.C.,1977.Stagesofsmall-groupdevelopmentrevisited.
GroupOrganizationalStud.,419–427.
Tuckman,B.W.,1965.Developmentalsequencesinsmallgroups.Psychol.Bull., 384–399.
VanDijk,T.A.,1989.Structuresofdiscourseandstructuresofpower.In:Anderson, J.A.(Ed.),CommunicationYearbook12.Sage,NewburyParkCA,pp.18–59. Vargas,A.,Lo,A.Y.,Rohde,N.,Howes,M.,2016.Backgroundinequalityand
differentialparticipationindeliberativevaluation:lessonsfromsmall-group discussionsonforestconservationinColombia.Ecol.Econ.129,104–111. Wang,J.,2006.Questionsandtheexerciseofpower.DiscourseSoc.17,529–548. Wasserman,S.,Faust,K.,1994.SocialNetworkAnalysisMethodsandApplications.
CambridgeUniversityPress.
Westaby,J.D.,Woods,N.,Pfaff,D.L.,2016.Extendingdynamicnetworktheoryto groupandsocialinteractionanalysis:uncoveringkeybehavioralelements, cycles,andemergentstates.OrganizationalPsychol.Rev.6,34–62. Wheelan,S.A.,1994.GroupProcesses:ADevelopmentalPerspective.Allyn&
Bacon,Boston.
Wheelan,S.A.,2009.Groupsize,groupdevelopment,andgroupproductivity. SmallGroupRes.40,247–262.
Zanoni,P.,Janssens,M.,2015.Thepowerofdiversitydiscoursesatwork:onthe interlockingnatureofdiversitiesandoccupations.OrganizationStud.36, 1463–1483.
Zuchowski,I.E.,1987.CommunicationCompetenceandtheLiaisonRole.Central MichiganUniversity:DepartmentofSpeechCommunicationandDramatic arts.