Page 1
OrganizationalCrisisandBolstering:Underrated
StrategyorPuffery?AnexperimentintotheeffectoftheBolsteringstrategyinmitigatingnegative
reputationfalloutduringinitialcrisiscommunicationsresponsesonsocial
media.
StudentName: NoudLeeflang
StudentNumber: 359167
Supervisor: Dr.YijingWang
MediaStudies:Media&Business
ErasmusSchoolofHistory,CultureandCommunication
ErasmusUniversityRotterdam
MasterThesisMedia&Business
22June2017
Page 2
MAMediaStudies:Media&BusinessMasterThesis
2
Page 3
N.M.Leeflang--359167ErasmusUniversityRotterdam
3
ABSTRACT
Theuseof thebolsteringstrategywas identifiedbyKimetal. (2009)asthe most prevalent crisis response strategy amongst practitioners. This is incontrast with what the Situational Crisis Communication Theory states: thatbolsteringisaminimaleffectstrategyonlytobeusedincertainsituations.Thisstudy used an online experiment (N=146) to determine how effective theincorporationofbolsteringisincombinationwiththedenialorcorrectiveactioninitialcrisisresponsestrategiesforacompanyofgoodpriorreputationfacingapreventable crisis. Four different crisis response strategies (with/ withoutbolstering X denial/ corrective action)were tested following a fictional articleaccusing Sony of negligence, resulting in dangerous product malfunction.Conditionswere presented randomly to our respondents.Our crisis responseswere framed as Facebook posts, as social media has evolved to an importantinitialcrisiscommunicationtool.
Bolstering was found to have little effect on the three factors crisiscommunication attempts to minimize the negative effects on: Post-CrisisReputation (reputation held after the crisis), Secondary Crisis Communication(intentiontospreadthecrisis)&SecondaryCrisisReactions(negativeword-of-mouth andpurchase intentions). Response strategy (denial/ corrective action)wasalso found tohave little effecton these factors.We found that thebiggestasset inminimizing negative reputational fallout is Pre-CrisisReputation, as isconfirmed inexisting theory.Ourresults indicate themost importantaspectofinitialcrisismanagementisthereputationheldbythecompanybeforethecrisisand that a positive pre-crisis reputation helps ‘shield’ the organization fromnegativecrisisfalloutonallthreeofourfactors.KEYWORDS:SCCT,Crisis,Communication,Reputation,Bolstering
Page 4
MAMediaStudies:Media&BusinessMasterThesis
4
TableofContents
ABSTRACT 31.Introduction 6
2.TheoreticalFramework 132.1Pre-CrisisReputation 152.2.ReputationHistory,RelationshipsandThreats 172.3.CorporateCrisistypes&Responses 182.4.SecondaryCrisisResponses:BolsteringStrategy 222.5.Social-MediatedCrisisCommunication 242.6.SecondaryCrisisReactions 26
3.Methodology 293.1.ResearchDesign 293.2.ExperimentalScenarios 303.3.DataCollection 313.4.Operationalization 323.4.1.Pre-CrisisReputation 323.4.2.AttributionofGuilt,History&Relation 333.4.2.Post-CrisisReputation 333.4.3.SecondaryCrisisCommunication 343.4.4.SecondaryCrisisReactions 34
3.5.ManipulationCheck 343.6.ComputedVariables:Reliability&FactorAnalysis 353.6.1.Pre-CrisisReputation 353.6.2.Relation 353.6.3.Attribution 363.6.4.PostCrisisReputation 363.6.5.SecondaryCrisisCommunication 373.6.6.SecondaryCrisisReactions 38
3.7.Description&Demographics 38
4.Results 414.1.Hypotheses1:TheeffectofBolsteringonPost-CrisisReputation 414.2.Hypotheses2:TheeffectofBolsteringonSecondaryCrisisCommunication 424.3.Hypotheses3:TheeffectofBolsteringonSecondaryCrisisReactions 444.4.AdditionalResultsonPost-CrisisReputation 464.4.1.Pre-CrisisReputation 464.4.2.LevelofAttribution 474.4.3.Age 474.4.4.Gender,Bolstering,ResponseStrategy&Facebook 47
4.5.AdditionalResultsonSecondaryCrisisCommunications 484.5.1.Pre-CrisisReputation 484.5.2.BolsteringStrategy 484.5.2.Attribution,Age,Gender&ResponseStrategy 494.5.3.Facebook&SecondaryCrisisCommunications 49
4.6.AdditionalResultsonSecondaryCrisisReactions 504.6.1.Pre-CrisisReputation 504.6.1.Attribution 504.6.3.ResponseStrategy 514.6.4.FacebookUse 51
Page 5
N.M.Leeflang--359167ErasmusUniversityRotterdam
5
4.6.5.Age,Gender,Bolstering&FBNews 514.7.RobustnessCheck 524.7.1.H1:PostCrisisReputation 524.7.2H2:SecondaryCrisisCommunications 544.7.3H3:SecondaryCrisisReactions 554.7.3.AdditionalResultsfromtheRobustData 55
5.Discussion 585.1.H1:Bolstering&Post-CrisisReputation 595.2.H2:Bolstering&SecondaryCrisisCommunications 605.3.H3:Bolstering&SecondaryCrisisReactions 615.4.Pre-CrisisReputationShieldsOrganizationsfromNegativeReputationFallout 625.5.OtherFindings 635.6.Conclusion 635.7.Limitations&FutureResearch 64
References 66
AppendixA-experimentconditions 72A1-FictionalNewsPost 72A2-FictionalFacebookPosts 73A2.1-Denial: 73A2.2-Denial+Bolstering 73A2.3-CorrectiveAction 73A2.4-CorrectiveAction+Bolstering 73
AppendixB-ExperimentSurvey 74
Page 6
MAMediaStudies:Media&BusinessMasterThesis
6
1.IntroductionIthasbecomeincreasinglyimportantforbusinesses,andespeciallythose
ofconsiderablesize,toattend,manageandbuildtheircorporatereputations
amongststakeholdersintheglobalized,interconnectedworld.Socialmediahas
empoweredstakeholdergroupstocreate,findorshareinformationon
organizationsandtheirconductintheformofsecondarycrisiscommunications
onsocialmedia(Mangold&Faulds,2009).Whereinthepast,peoplewere
reliantonthemainstreammediaandinformationfromspokespeople,theynow
havetheabilitytoconnectandengagewithcompaniesonlinebutcanalsolearn
ofmisconductorfaultsquickerandreactmoreeffectively.Individualshave
becomethepolice,juryandjudgewhenitcomestocorrectcorporatebusiness
conduct.Thedreaded‘mediashitstorm’,traditionallyreservedtomajorcrises,
hasbecomeanincreasinglycommonoccurrencewithbusinessesbeing
scrutinizedandinvestigatedbytheirstakeholders.Thiscanresultinnegative
effects,thesecondarycrisisreactions,likestakeholdersspreadingnegative
messagesorevenadaptingtheirpurchaseintentions.
Powerdynamicshaveshiftedascompanieshaveseenthehighlevelof
controltheyhadinthetraditionalintegratedmarketingcommunication
hierarchydiminishedinfavorofconsumerempowerment(O’Brien,2011).
CorporateCommunicationisnolongeraone-waystreet,butamyriadof
overlappingroadsdispersinginalldirections.Thishasopenedan
unprecedentedlevelofdialoguebetweenorganizationsandstakeholdersthatis
uniquetoourtime.Assuch,manyorganizationshavesuccessfullyintegratedthis
onlinepresenceandreputationmanagementintotheirstrategiesandevenuse
socialmediaasafirstcommunicationpointduringcrises(Schultz,Utz&Goritz,
2011).
Thisconnectiontothestakeholdercomesatapriceasconsumersare
moreawareoftheirrights,wishesandopinionsandwillhappilypointoutany
mistakethecompanymakesorinformtheirpeersofmisconduct(Mangold&
Faulds,2009;O’Brien,2011).Thedigitalera,withallitssharingcapabilities,has
madecrisesmorenumerous,visibleanddisastrouswithalargergroupof
victimsandawiderimpact(Seeger,Sellnow&Ulmer,1998;2003)Small
mistakesareeasilyhandled,buttherealchallengeistomanagecorporate
Page 7
N.M.Leeflang--359167ErasmusUniversityRotterdam
7
reputationintimesoforganizationalcrisis.Corporatereputationisanintangible
assetoftheorganizationformedbyevaluationsofthestakeholdersaboutthe
organization(Coombs,2007a).Organizationalcrisesarespecific,unexpected
andnon-routineevents,oftenwithhighlevelsofuncertainty,whichthreatenan
organizationsreputationandgoals.Thesecrisesaffecttheinteractionwith
stakeholdersandoftendiminishtheorganizationsreputation,credibilityand
legitimacy(Schultz,Utz&Goritz,2011).Stakeholdersmightevenadapttheir
purchasinghabitsorthewaytheytalkaboutanorganization,thesecondary
crisisreactions,basedonitsreputationandcrises(Coombs&Holladay,2014;
Stockmyer,1996).
Oneofthemostwell-knowncorporatecriseseveristheJohnson&
JohnsonTylenolcrisisof1982.Itisacaseoftencitedasanexemplaryguideon
howorganizationshouldrespondtoacrisis.Thecrisisitselfwas,withouta
doubt,oneofthemoreseriousoneswithsevenpeopledyingofcyanidelaced
Tylenoltablets.AnextremelytroublingeventthatcouldhavecostJohnson&
Johnsontheirreputation,haditnotbeenfortheirswiftandcorrectresponse.
ThemomentitbecameclearthatTylenoltabletshadbeenlaced,Johnson&
Johnsonimmediatelypulledallstock,haltedproduction,gotinvolvedinthe
investigationandoffereda$100,000rewardforfindingthekiller.Afterthe
crisisJohnson&Johnsonintroducedtemper-resistantpackagingandoffered
discounts.Thecrisishadcosttheminexcessof$100millionbuttheirreaction,
cooperationandtransparencyhadputthecompanyinafavorablelightandthe
brandwasabletorecoverquickly(Kaplan,2005).Johnson&Johnsondisplayed
howeffectivecrisismanagementcannotonlyminimizeacorporation’snegative
reputationfalloutfollowingacrisis,buthoweffectivecrisismanagementcan
alsohelpimprovereputationpost-crisis.
WhenToyotafacedacrisisin2009pertainingtotheirfloor-matsgetting
stuckbehindtheaccelerator,aproductfaultthatresultedinafatalaccident,it
wasthecatalystforaseriesofcustomercomplaints.Asaresult,morethan8
millionvehicleswererecalledworldwideandToyotafacedthebiggestcrisisin
itshistory.However,Toyotadidnotrespondtothecrisisuntil6monthsafterthe
crash(Davey,2010;Verschoor,2014).Toyotahadfailedtotakeappropriate
actionintimeandonlywhenthetraditionalmediapickedthestoryupdidthey
Page 8
MAMediaStudies:Media&BusinessMasterThesis
8
gointo‘crisismode’,eventhoughtherewerealreadydefinitesignsofacrisis
online.Companiesshouldthusmonitortheirreputationonlineand‘assumethe
worst’whenitcomestocrisesandrespondbeforespeculationandrumortake
over(Davey,2010;Schultz,Utz&Goritz,2011).Badcrisismanagementresults
inworsecrises,andinthecaseofToyotacostthecompanymillionsofdollars
andasignificantblowtotheirreputation.Nonetheless,Toyota’spriorgood
reputationprotectedtheminpartandtheywereabletorebuildtheirreputation
inthefollowingyears.
Evenwhenmonitoring,itisstillpossibleforanorganizationtobe
oblivioustothecorporatecrisisensuing.TakeforexamplePepsi,amultinational
beverageconglomerate,whorecentlylaunchedanewadinwhichanInstagram
celebrityappearstobemockingpoliticalprotestandthegrowingpolitical
divisionanddiscontentfoundamongstmanypeopleinsociety.Bytryingtomake
averyserioustopic‘fun&whimsical’,theyeffectivelyangeredmany
stakeholdersandwereputtoheavyscrutinybycivilrightsgroupsandthe
media.Pepsihadtriedto‘hook-in’onacurrentaffairbutfailedmiserably,andin
theendappearedtobemostlymockingsomeveryserioussocietalissues.Sucha
corporateblunder,resultinginafull-blowncrisis,isnotuncommonandoften
stemsfrominternalissuesormismanagementinsaidcompany.
WhatmakesthiscaseinterestingisthatfollowingthePepsiAdCrisis,
internationalbeercompanyHeinekenlaunchedanadinwhichtheydidwhat
Pepsiwastryingtodo,butdiditright.Insteadofmockingpoliticaldifferences
andprotest,theyputtwopeoplewithwidelydifferentviewpointsacrossone
another.OnlywhentheyhadbuiltarapportdidHeinekenrevealthewidely
differentviewseachpersonhad.Wouldtheystayanddiscusstheirdifferences
overabeerorwalkaway?Obviouslytheystayedandshareda‘Heiny’.Whatthis
teachesusisthatcorporatecrisescanappearsuddenly,thatmanycompanies
areinternallyoblivioustothewaystheymightstartacorporatecrisis,thateven
thebestintentionsarevoidwhenexecutedwronglyandthatonecompanies’
crisisisanothercompaniesopportunity.Whatisalsointerestingtonoteisthat
therewasevenoutragetowardstheHeinekenadbythe‘progressive’media.Itis
hardtoappeaseallthestakeholdersasanorganization.
Page 9
N.M.Leeflang--359167ErasmusUniversityRotterdam
9
Appeasingeveryonemightbedifficult,buttherearealsocasesinwhich
companiesblatantlyplacetheirprioritiesinthewrongballpark.The2009
financialcrisiswasharshforthefinancialindustry,butconsiderablyharsherfor
thepeoplewhowereafraidtolosetheirhome,incomeorlivingasaresultofit.
WhenJPMorgandecidedtoinvest138milliondollarsintoafewprivatejetsand
anewhangarafterhavingjustreceiveda25billionbailout,itwasnotsurprising
thatcriticswerequicktohighlightthehypocrisyofthis(Sarstedt,2009).This
wasconsequentlyusedasanexampletocondemnthewholebankingindustry
foritsirresponsibleandlavishspending,effectivelyspreadingthecrisisto
involvepartieswhohadnodirectguiltinthematter.Asimilarfeatoccurred
duringtheVolkswagenemissionscrisis,afterwhichthewholeautomobile
industrywasscrutinizedandinvestigatedafteritwasdiscoveredVWhadplaced
‘tamperingsoftware’intheirdieselcars.Thismeansthatcrisesthatstartinone
company,canspreadtothewholeindustrywhenmismanaged.
Tohelpprofessionalsinovercomingcrises,andtobetterunderstand
them,theSituationalCrisisCommunicationTheorywasdeveloped(Coombs,
2007a).Thistheoreticalguidehelpsinassessingthereputationalthreat,
identifyingthetypeofcrisisandoffersadviceonthecorrecttypeofresponseto
begivenbytheorganizationtominimizenegativereputationalfalloutfromthe
crisis.Duetotheextensivetestingandresearchonwhichitwasbuilt,theSCCTis
consideredacornerstoneofcorporatereputationmanagementandwillactasa
theoreticalguidethroughoutthisthesis.
TheSCCToffersthreeoverarchingresponsestrategiestocorporate
reputationcrises(Coombs,2007a).Thesethreecrisisresponsesaredenialofthe
accusation,rebuildofthereputationanddiminishoftheattribution.These
strategiesareeachapplicabletoadifferenttypeofcrisisandcrisisguilt
attribution.OneoftheleaststudiedaspectsoftheSCCTisthepracticeof
bolstering-tostrengthenorsupportthecorporatereputationbyhighlightingthe
goodpracticesofanorganization,praisingstakeholdersorvictimization(Benoit,
1997;Coombs,2007a).Bolsteringisgenerallyconsideredanadditionalcrisis
responsestrategytobeusedincombinationwithanotherstrategy,butassuch
limitedattentionhasbeengiveninacademicstothisstrategythatdoesnotfocus
onattributingresponsibility(Ma&Zahn,2016).Nonetheless,researchhas
Page 10
MAMediaStudies:Media&BusinessMasterThesis
10
shownthatbolsteringisindeedausefultoolinminimizingreputationaldamage
(Brinson&Benoit,1999;Sheldon&Shalot,2009).
ThisbecomesapparentwhenlookingatthecrisismanagementCitibank
employedduringthe2009financialcrisis.Thebailoutofthefinancialindustries,
andresultinglossoftrust,wasaserioustarnishtoCitibank’sreputation.Unlike
JPMorgan,Citibankadaptedtothecrisisandbyusingcrisisresponsestrategies,
thebankwasabletorecoveralargepartoftheirreputationandre-instillfaith
fromtheconsumers.Intheseresponses,Citibankusedbolsteringmostoftenin
responsetothecrisisandgovernmentbailout(Weberetal,2011).Bolsteringis
bestusedincombinationwithanothercrisisresponsestrategyasprevious
researchhasproven,andwasalsothecasewithCitibank(Coombs,2007a;Kim
etal,2009;Weberetal,2011).
Duetothelackofexperimentalresearchdealingwithbolstering
specifically,thisresearchwillfocusonthevalueofbolsteringasasupplementto
crisisresponsesintimesofcorporatecrisisandhowthisaffectspost-crisis
reputation.Itwillalsostudythewayinwhichonlineaudiencesmightengagein
secondarycrisiscommunications(sharing/commenting)onsocialmediaor
performsecondarycrisisreactions(recommendingorcondemningtopeers),
whichalsoincludespurchasingbehavior.Theresearchquestionthatthisthesis
willaimtoanswerwillthusbe:
Towhatextent,ifany,doesthebolsteringstrategyaffectpostcrisis
reputation,secondarycrisiscommunicationsandsecondarycrisisreactionsduring
timesoforganizationalcrisis?
Toanswerthisquestion,anonlineexperimentwasdesignedinwhicha
fictionalcrisiswasintroducedandtherespondentswereconsequently
presentedwithdifferentconditionsatrandom.Theseconditionspresented
differentcrisisresponsesthateitherincorporatedbolsteringordidnot.Another
aimofthestudywastodistinguishtheinfluenceofbolsteringonsecondary
crisiscommunicationsandsecondarycrisisreactions,whichwere
operationalizedusingpreviousresearchmodelsandtheory.
Page 11
N.M.Leeflang--359167ErasmusUniversityRotterdam
11
Bolsteringhaspreviouslybeenstudiedmainlyinacase-studydesign,
fromwhichtheresultsindicatethatitisawidelyusedandeffectivecrisis
responsestrategy(Kim,Avery&Larissa,2009;Sheldon&Sallot,2008;Weberet
al,2011).TheSCCT,however,claimsbolsteringisa‘minimalopportunity’to
increasereputationalassetsintimeofcorporatecrisis(Coombs,2007a).
Limitedresearchhasdealtwiththequestionifthebolsteringstrategymakes
crisisresponsestrategiesmoreeffectiveornotinlimitingnegativepostcrisis
falloutcomparedtostrategiesthatdonotincorporatebolstering(Ma&Zhan,
2016).
Thisstudywillattempttodiscoverthetrueeffectofthebolstering
strategyoncorporatereputationandstakeholderreactionsusingan
experimentalresearchdesign.Sucharesearchmighthelpusbetterunderstand
howbolsteringworksoncorporatereputationintimesofcrisis,andoffer
academicsaswellaspractitionersimportantinsightintotherealvalueof
bolsteringasasecondarycrisisresponsestrategy.Hopefullyitwillfillan
importantgapinthetheoryofcrisiscommunication.
Theimplicationthatmanypractitionersofcorporatecrisis
communicationareindeedusingthebolsteringstrategytoprotecttheir
reputationsindicatesagapbetweenwhatisfoundinthe‘realworld’andsociety,
andwhatresearchershaveincorporatedintomodelsandtheory(Claeys&
Opgenhaffen,2016).Thisresearchwillattempttofillthisgapandprovidethe
corporateworld,societyandacademicswithananswertothequestionif
bolsteringisaneffectivesupplementtotheprimarycrisisresponsestrategies,as
indicatedintherealworld,orthatitisminimalassetasisindicatedinacademic
theory.Theresultsofthisstudycouldhaveanimpactonboththesocietal;the
waybolsteringisusedincorporatecommunication,asontheacademic,inthe
formofanupdatedtheoryoratleasttheinsinuationthatmoreresearchis
required.
Implicationsforcorporatecrisiscommunicationcouldbethattheyare
eithercorrectinusingthebolsteringstrategyandthatitisanunderratedtoolin
minimizingnegativereputationfalloutfollowingacrisis.Or,itismerelya‘fluff’
pieceusedbypractitionersbecausetheyarereluctanttoonlycommunicate
negativeinformationlikeadmittingguiltandapologizing.Thiswouldmeanthat
Page 12
MAMediaStudies:Media&BusinessMasterThesis
12
thebolsteringstrategymightnotbeanythingmorethanapsychologicalcrutch,
makingitslightlyeasiertoadmitguiltandtakeappropriateaction.Alternatively,
usingbolsteringwhiledenyingguiltmightbeastrategytodivertattentionfrom
thenegativecrisisunfolding.
Beforementionedpotentialimplicationsforcrisiscommunicators,that
bolsteringisindeedaveryeffectivestrategyinminimizingnegativereputation
falloutwouldmeanthatacademicshaveconsistentlyunderestimatedthe
effectivenessofthestrategy.Ifthiswerethecase,theacademicfieldofcrisis
communicationwouldbewisetore-investigatethematterofbolsteringandifit
wereindeedpossiblethatthebolsteringstrategyhasbeenunderestimatedinthe
SCCT,thetheorywouldneedtobeadapted.
Thisresearchwillattempttodissecttheeffectivenessofthebolstering
strategyonthethreemainfactorstheSCCTattemptstominimizethenegative
reputationalcrisisfallouton.Thesefactorsarethereputationafterthecrisis
(post-crisisreputation),willingnesstospreadthecrisis(secondarycrisis
communications)andbehavioralactionstakenafterthecrisis(secondarycrisis
communications)
Page 13
N.M.Leeflang--359167ErasmusUniversityRotterdam
13
2.TheoreticalFrameworkCorporatereputationistheaccumulationofopinions,expectationsand
evaluationsmadebystakeholders,anyonewhohasastakeorinterestinthe
company,aboutanorganization(Coombs&Holladay,2006;VanRiel&
Fombrun,2007,ch2.).Theseevaluationsmakeupthecorporatereputationofan
organizationandcanresultineitherfavorableoutcomes,suchasattractiveness,
credibilityandimprovedfinancialperformance,ornegativeoutcomessuchas
lossofbusiness,distrustandsometimesevenprotestsorboycottsofthe
organization(Coombs,2007a,VanRiel&Fombrun,2007,ch2.).These
reputationsareconstantlybeingrenegotiatedinsociety,arealwayschanging
andneedtobemanagedcontinuously(Fombrun&vanRiel,2004,ch1.).O
Duetothebenefitsofagoodreputationorganizationswillthusaimtomaintaina
positivereputationamongstitsstakeholders,butwilllosesaidgoodreputation
whentheyneglecttoproperlymanageit.
Positivereputationalcapital,ora‘reservoirofgoodwill’,canalsoactasa
defenseagainstnegativepublicity,meaningthatanorganizationwithaprior
goodreputationwillexperiencelessharmduringacrisisthananorganization
withanegativeorevenneutralreputation(Coombs&Holladay,2006;Decker,
2012).Thisissupportedbybothcasestudiesasexperimentalresearch(Decker,
2012;Fombrun&Foss,2004;Mahon&Wartick,2003;Tucker&Melewar,2005;
Turketal,2012).Thepublicwillbemorereceptiveofthecorporation’scrisis
responseandhandlingif,upuntilthen,thecorporationhasbeenheldinhigh
esteem.Furthermore,apositivereputationcanbeusedasadefensive‘weapon’,
forinstancebybolsteringaboutpastgooddeeds,tomaintainandprotectthe
marketpositionandrelationshipwithstakeholdersandrecovermorerapidly
(Coombs,2007a;Decker,2012;Mahon&Wartick,2003).Becauseofthis,
bolsteringisgenerallyonlyrecommendedintheSCCTwhenthereisagoodprior
reputation.Onewayofmeasuringacorporation’spriorreputationisthroughthe
RepTraksystem,whichannuallyrankscorporatereputationsbasedon
worldwidesurveys(Ponzi,Fombrun&Gardberg;2011).
AccordingtoFombrun&VanRiel(2004,ch5)thefoundationofastrong
reputationarefiveprinciplesthatseparatewinningreputationsfromtherest.
Theseprinciplesarevisibility,distinctiveness,authenticity,transparency,and
Page 14
MAMediaStudies:Media&BusinessMasterThesis
14
consistency.Visibilityreferstothewillingnessandcapabilityofanorganization
tocommunicatewithitsstakeholders,themediaandreleaseinformationabout
itself.Distinctivenessreferstobeingunique,oratleastbeingperceivedas
unique,andusingthisfeatincommunicationandmarketing.Authenticityishow
believableacompanyis.Fakingauthenticityishard,andsomethingstakeholders
canfindoutquicklyintodaysconnectedandinformedworld.Authenticityis
relatedtotrustandcredibility:youneedtobegood,notjustappeargood.
Transparencymeansbeingopenabouthowyouworkandwhatyourgoalsare.
Companiesthatdonotshyawayfromcommunicatingwiththepublicand
openingtheirdoorstothemhavehigherreputationsbecausetheycanshowthe
worldtheyaremorethanafacelesscorporation.Consistencyinactionsand
messagebuildswinningreputations.Choosewhatyourorganizationstandsfor
andstickwithit.Finally,VanRiel&Fombrun(2007)addasixthfactorinalater
publication:responsiveness.Companiesthatareabletoquicklyandadequately
respondtosituations,negativeorpositive,buildmoretrustandcanoften
preventacrisisinitsinfancy.
Maintainingapositivereputationbecomesconsiderablymoredifficult
whentheorganizationfacesacrisis.Corporatecrisesareeventsthatare
unexpected,disruptiveandhavethepotentialtonegativelyinfluencethe
stakeholderrelationship(Coombs,2007a;Bundy,Pfarrer,Short&Coombs,
2016;Kahn,Barton&Fellows,2013).Thismeansmostorganizationswillaimto
avoidcrises,andifthatisnotpossibleattempttominimizethenegativefallout
fromthecrisis.Tominimizethiscrisisfallout,theorganizationmusttakeprior
reputation,crisisresponsibility,crisishistoryandpriorrelationsintoaccount
whendetermininghowtorespondtoacrisis(Chiciudean&David,2013;
Coombs,2007a).Someofthemainnegativereputationalfalloutscrisis
communicationattemptstominimizearepost-crisisreputation,secondarycrisis
communicationsandsecondarycrisisreactions.Post-crisisreputationhereby
referstotheevaluationsgivenbystakeholdersabouttheorganizationafterthe
crisis.Secondarycrisiscommunicationsaretheactionstakenbythestakeholder
followingacrisis;doesheorsheshare,commentorinformpeersaboutthe
crisis?Finally,secondarycrisisreactionsrefertotheopinionsandactiona
stakeholdermighttakeafterthecrisis,likespreadingnegativeword-of-mouthor
Page 15
N.M.Leeflang--359167ErasmusUniversityRotterdam
15
evenadaptingpurchaseintentions.
OneofthecornerstonesofcrisismanagementtheoryistheSituational
CrisisCommunicationTheorydevelopedbyTimothyCoombs(2007a).This
theoryoffersaframeworkthatcanbeusedtoprotectreputationalassetsduring
acrisis.TheSCCTattemptstopredicthowstakeholderswillreactinacertain
crisissituationbyassessingthelevelofreputationalthreat(basedonattribution,
crisishistoryandpriorrelationship)andwhatresponsestrategyisbestsuitedto
reducelossofreputation.TheSCCTdrawsontwoprevioustheoriesofcorporate
reputationcrises,andincorporatesthesetocreateatheorythataimstoidentify
andrestorecorporatereputationaftercrises.
ThefirsttheoryistheAttributiontheory,whichexplainsthatpeoplewill
searchforthecauseofeventsandattributeresponsibility(positiveornegative)
ofsaideventtoapersonororganization(Weiner,1985).Themostcommon
attributionsintheSCCTareangerandsympathy(Coombs,2007a).Attribution
canalsoinfluencebehavioralresponses,negativeonesforangerattributionand
positiveonesforsympathyattribution(Weiner,1985;Coombs2007a,2007b).
ThesecondtheoryistheImageRestorationtheory,whichaimstooffer
guidelinestorestorecorporatereputationwhenfacedwitha(negative)crisis
(Benoit,1997).TheSituationalCrisisCommunicationTheoryoffers
communicationstrategiesfordenying,evadingandreducingreputational
threatsduringacrisis,manyofwhichwereadaptedorbasedonthetwo
previoustheories,andassuchwillbeusedastheoverarchingtheoryinthis
paper.
2.1Pre-CrisisReputation Acorporatecrisishasthepotentialtoinflictreputationaldamage,which
canresultinanincreasedunfavorableviewoftheorganization.Onewayof
protectinganorganizationagainstnegativecrisisfalloutisbybuildingand
maintainingapositivereputationbeforeacrisisoccurs(pre-crisisreputation)
(Coombs,2007a).Wehavealreadydiscussedthemeritsofapositivereputation
fororganizations:attractiveness,trust,credibilityandimprovedfinancialgains
(Coombs,2007a,VanRiel&Fombrun,2007,ch2.).Thismeansmaintaininga
positivereputationisanassettoanorganizationthatwillaiditinappeasing
Page 16
MAMediaStudies:Media&BusinessMasterThesis
16
stakeholdersandimprovingbusiness.Ontheotherhand,anegativereputation
will(generally)negativelyaffectbusinessandcausestakeholderstochoose
alternativeorganizationswhenpossible.
Apositivereputationcanhelpaidacompanyinminimizingnegative
reputationfalloutfollowingacrisis.Afavorablepre-crisisreputationactsasa
bufferof‘reputationalcapital’duringcrisesandallowstheorganizationtosuffer
lessdamageandreboundquicker(Coombs,2007a;Fombrun&vanRiel,2004;
Knight&Pretty,1999;Gregory,1999).Companiesthathaveahigherpre-crisis
reputationwillalsohaveahigherpost-crisisreputationthantheircounterparts
withalowerpre-crisisreputation(Decker,2012,Fombrun&Foss,2004;Tucker
&Melewar,2005)Researchintotheeffectivenessofpre-crisisreputationon
negativereputationfalloutfollowingacrisisconfirmsthisandindicatesthat
thosewhohaveabetterpre-crisisreputationwillhaveamorefavorablepost-
crisisreputationthanorganizationsthathadanunfavorablepre-crisis
reputation(Claeys&Cauberghe,2015;Coombs&Holladay,2001;2002;2006;
Decker,2012;Sheldon&Sallot,2009).Thiscouldbeexplainedbyareluctanceof
thestakeholderstochangetheirpre-existingattitudeaboutanorganization,
evenattributinglessresponsibility,andgoingasfarasthatapositivepre-crisis
reputationcanprotecttheorganizationagainstnegativepublicityandexternal
allegationsfollowingacrisis(Claeys&Cauberghe,2015).
CoombsandHolladay(2006)callthisthe‘Haloeffect’.Thisreferstothe
powerofpre-crisisreputationindeterringnegativereputationfallout:the
organizationisbasicallysogooditreachesalevelof‘sainthood’.Theystatethis
haloeffectactsasashieldratherthana‘benefit-of-the-doubt’factor,meaning
thestakeholderwillbelesslikelytochangetheirowncognitionorexpectation
biasduetoacrisis(deflectedofftheshield).Thisdoesnotmeanthatpre-crisis
reputationcancompletelyomitthenegativefalloutfromacrisis.Volkswagen
hadapositivereputationbeforetheiremissionscrisisbutduetoimproper
managementandresponsetothecrisis,lostalargeportionoftheirpositive
reputation.AstudybytheReputationInstitute(2016b)foundthatpeoplewere
initiallynotthatfazedbythecrisisanditdidnotdamageVW’sreputation
greatly,butatalaterpointreportedalossoftrustandwillingnesstopurchase
VW’sproductsduetothewayVWrespondedtothecrisis.
Page 17
N.M.Leeflang--359167ErasmusUniversityRotterdam
17
Strongreputationsmightthusprotectanorganizationduringinitialcrisis
responsesbutmismanagementatalaterlevelcanstillresultinreputational
damage.Wecanthereforesaythatastrongreputationprotectsanorganization
bygivingitareservoirofgoodwill,butthatthatreservoirdoesrunoutatsome
pointwhennoactionsaretakentoreducetheoffensivenessofthecrisis.Suchan
assetofpositivereputationisthereforeofprimeimportancefororganizations
who,besidestheusualbenefitsoutsideoftimesofcrisis,willalsobebetter
protectedwhenacrisisdoesarise.
2.2.ReputationHistory,RelationshipsandThreats Everycrisishasitsownlevelofreputationalthreat.Thelevelof
reputationalthreatposedbyacrisisisinfluencedbythreedifferentfactors.The
firstistheinitialcrisisresponsibility.Thisisrelatedtotheattributiontheory
thatstatesthatwhenaneventhappens,peoplewillattributeresponsibility
aboutthecauseofacrisisinvaryingdegreestoanorganization(Kelley&
Michela,1980;Coombs,2007b&2006).Moreattributionthusmeansthata
greaterreputationalthreatisformedbecausestakeholdersthinkthecompanyis
toblameandguiltyofthecrisis.Thesecondandthirdfactorsarethe
organization’scrisishistoryandpriorrelationalreputation.Crisishistoryrefers
topastcrisesthatmighthavebeensimilarinshape,formandattributionand
priorrelationalreputationreferstothepastinteractionsbetweenthe
stakeholdersandtheorganization.
Whenpositive,thesefactorscanhelpreducetheorganizations’crisisand
evenhelppreventthem,muchlikeapositivepre-crisisreputationcan(Bundyet
al,2016;Coombs,2007a;Kahnetal,2013;Ulmeretal,2011)Ontheotherhand,
whennegative,thesesituationfactorsactasintensifiersforreputationalthreats
andcrisisattributionandthuscanthoroughlyescalatethesituation(Coombs&
Holladay,2001;Coombs,2006;2007a).Everycrisisisanaccumulationof
differentsituationalfactors,actionsandresultsanditisthereforeimpossibleto
deviseauniformityof‘bestpractices’ortoevenpredictwithassuranceifacrisis
willbesevereornot(Coombs,2015).
Levelofreputationalthreatisanimportantaspectofcrisis
communicationand,asmentionedbefore,isstronglytiedtoWeiner’sattribution
Page 18
MAMediaStudies:Media&BusinessMasterThesis
18
theory(1985).Duringareputationalcrisisitisthusimportanttofirstassessthe
levelofattributionandreputationalthreatanorganizationmightbesubjectto
fromsaidcrisis.Humanresponsewhenfacedwithanegativeeventistoseekout
thecauseofsaidevent(Weiner,1985).Causalityisthusgenerallythefirst
questionraisedwhenacrisisarises(Kelley,1967).Highlevelsofattribution
indicateaninternalcausality(itistheorganizationsfaultfordeviatingbehavior)
whilelowlevelsindicateanexternalattribution(theorganizationbehaved
normallybutwasinfluencedbyoutsidesituationalfactors).Higherlevelsof
crisisattributionareplacedonorganizationsthatcausedthecrisisinternallyand
willthereforeexperienceamoreseverecrisisthreat(Bundy2016;Coombs,
2007a;2004;Kimetal,2009).
Researchhasshownthatcrisishistoryalsohasaninfluenceonthe
reputationalthreat,andthatpastcrisescanintensifythecurrentlevelof
reputationalthreat(Bundyetal,2016;Coombs,2004;Jeong,2009;Siscoetal,
2010).Anegativerelationshipcanalsomakestakeholderslessforgivingbecause
theywillalreadyhaveanegativeopinionoftheorganizationandthecrisismight
helpreinforcethisbelief(Bundyetal,2016;Coombs,2004;2007a).Ontheother
hand,apositivehistoryorrelationshipwillactasanassettotheorganizationin
timesofcrisis,causethepublictobemoreacceptingofthecrisisresponseand
minimizethelong-termfallout(Decker,2012).Anorganizationmusttherefore
firstunderstandhowthepublicperceivesthecrisisbasedontheirreputation
historyandthelevelofattributionbeforechoosingaresponsestrategy(Coombs,
2007a;Choi&Chung,2013).
2.3.CorporateCrisistypes&ResponsesTheSCCTaimstoofferguidelinesforpractitionersbasedonempirically
testedevidence.Coombs(2007a)identifiesthreedifferenttypesofreputational
crisisthatarebasedontheimagerestorationtheory(Benoit,1995)andwhich
aresupportedinempiricalresearch(Coombs&Holladay,2004).Inthevictim
cluster,theorganizationisalsothevictimofthecrisisandexperiencesaweak
attributionofresponsibilityandthusamildreputationalthreat.Intheaccidental
cluster,thecrisiswasunintentionallycausedbytheorganizationandthus
minimalattributionscanbemadeandamoderatereputationalthreatis
Page 19
N.M.Leeflang--359167ErasmusUniversityRotterdam
19
experienced.Thefinalcluster,andtheonethatposestheseverestreputational
threatandattributionofcrisisresponsibility,isthepreventablecluster.Inthis
typeofcrisis,theorganizationknowinglytookinappropriateactionsorcouldat
leasthaveavoidedthem.
Oncethelevelofreputationalthreatandtypeofcrisishasbeenidentified,
theSCCToffersdifferentcrisisresponsestrategiesaimedatreducingthe
negativeaffectandpreventnegativebehavioralintentions(Coombs,2007a).The
centralthemeinthisisresponsibility,whichinturnrequiresaccountabilityfor
thecrisis-attributionplacedontheorganizationbythestakeholders.Thetypeof
responsestrategytheorganizationusescangreatlyinfluencetheresulting
stakeholderevaluations(Bundy&Pfarrer,2015;Bundyetal,2016;Coombs,
2007a).
TheSCCT’scrisisresponsestrategiescanbedividedintoprimaryand
secondarycrisisresponsestrategies.Theprimarystrategiesarebasedondenial,
diminishandrebuild(Coombs,2006;2007a).Thesecondaryresponsestrategy
istheBolsteringstrategyandwillbediscussedlateron.Thedenialstrategy
attemptstoremoveanyconnectionbetweenthecrisisandtheorganization.This
strategyismostusefulinthevictimcluster,whentheconnectionbetweenthe
organizationandthecrisisisuntrueoratleastcannotbeproven.Thediminish
strategyaimstoconvincestakeholdersthatthecrisisis‘notthatbad’orthatit
occurredoutsideoftheorganizationscontrol.Finally,therebuildingstrategy
aimstominimizethereputationaldamagebyshowingsympathyoroffering
compensation.
Page 20
MAMediaStudies:Media&BusinessMasterThesis
20
CrisisType Victim Accidental Preventable
Attribution Low Medium High
Examples • Natural
disasters
• Workplace
violence
• Rumors
• Accidents
• Producterrors
• Disputeswith
stakeholders
• Humancaused
accidents
• Humancaused
producterrors
• Misdeeds
Crisis
Response
(Primary)
Denial
• Attackthe
accuser
• Deny
• Scapegoat
Diminish
• Excuse
• Justification
Rebuild
• Apology
• Corrective
action
Figure1:SCCTcrisistypes,attributionsandresponsesoverview
Researchintotheuseofthesedifferentcrisisresponseshaslargely
confirmedtheSCCT.Oneresearchattemptedtoempiricallyprovethevalidityof
theSCCTbystudyingtheeffectivenessofthethreedifferentcrisistypesmatched
withdifferentcrisisresponsesinanexperimentresearchdesign(Claeys,
Cauberghe&Vyncke,2010).Itwasfoundthatcorporatereputationismost
underthreatwhenfacedwithapreventablecrisisandthattherebuildstrategyis
mosteffectiveinmitigatingnegativereputationfalloutfollowingacrisis.
Surprisingly,itwasalsofoundthatcrisesintheaccidentalorvictimclusterwere
treatedthesamebytheirrespondentsandresultedinsimilarreputational
effects.Finally,theyfoundthatthedenystrategywasonlymarginallyless
effectivethantherebuildstrategyandthatpersonaltraitsofrespondentsmight
beanimportantfactorinthis.
OtherresearchhasalsodealtwiththeSCCT,anditwasalsofoundthatthe
rebuildstrategywasthemosteffectiveinmitigatingnegativereputationfallout,
regardlessofcrisistype,whilethedenialordiminishstrategieselicitedthemost
negativeresponses(McDonald,Sparks&Glendon,2010).Resultsalsoshowed
thatthecrisis-causeandcrisis-responsibilityappearedtobemoreimportant
thanresponse.Thismeansresponsestrategyislessimportantthanifthe
organization(willingly)causedthecrisisornot.Theresponsethatattributesthe
mostresponsibilitytotheorganization,therebuildstrategy,actuallyresultsin
thestakeholderthinkingtheorganizationislessresponsible,oratleastthe
Page 21
N.M.Leeflang--359167ErasmusUniversityRotterdam
21
stakeholderthinkstheyarelessguilty.Thevice-versaeffectcanbeobservedfor
thestrategiesthataimtoreducetheperceptionofresponsibilityandactually
increasetheattributionreportedbyrespondents.Itseemsthattaking
responsibility,displayingtransparencyandremorsearethemosteffectivecrisis-
responsestrategies.Organizationsthattrytocorrectmistakesmadenomatter
what,likehowJohnson&JohnsonreactedtotheTylenolcrisiswiththepublicin
mindandnottheirprofits,areconsiderablylessaffectedbycrises.
Similarresultswerefoundinotherresearch.Cultural,personalor
relationalattributeswerefoundtobeimportantfactorsincrisiscommunication
responseeffectiveness(Verhoevenetal,2014;Luoma-aho,Moreno&
Verhoeven,2017).Responseswithalowlevelofaccommodation(denial)are
generallyregardedaslessethicalthanresponseswithamoderate(diminish)or
high(rebuild)levelofaccommodationtowardsthestakeholder(Coombs,1999;
Coombs&Holladay,2005;Coombs,Holladay&Claeys,2016;Decker,2012).
AnotherstudyinTaiwanfoundthattimely,consistentandactiveresponsestoa
crisisaremoreimportantthantheactualresponsestrategyused(Huang,2008).
Thisisgenerallysupportedinotherresearchthatfoundthatoneofthemost
detrimentalresponsestoacrisisfororganizationalreputationisgivingno
responseatall(Coombsetal,2016;Park&Reber,2010;Coombs,2006).Decker
(2012)alsomentionsthatthedenialstrategyisbetterthannoresponse,butonly
whenthereisapositivepre-crisisreputationthatcanhelpshieldthe
organization.
Itisnotsurprisingthatwefinddiscrepanciesbetweenstudiesandthe
SCCT.TheSCCTwasdevelopedasaguidebasedonempiricalresearch,which
couldbeappliedintoeverycrisissituation.Thisdoesnotaccountforsituational
factorslikeculture,geographyorspecificcrisesorcrisishistories.Nonetheless,
theSCCTremainstodatethemostinclusiveandcompletetheoryforcrisis
communicationandremainsvalidduetoitsextensivetestingandinclusioninto
multipleempiricalstudiesforwhichitlaidthegroundwork.
Page 22
MAMediaStudies:Media&BusinessMasterThesis
22
2.4.SecondaryCrisisResponses:BolsteringStrategyThesecondarycrisisresponsestrategy,whichisgenerallyconsideredan
additiontotheprimarystrategies,consistsoftheBolsteringresponsestrategy
(Coombs,2007a;Weber,etal,2011).Bolsteringisunderstoodtobethepractice
ofstressinggoodpastdeedsandtraitsofanorganizationintimesofcrisisto(re)
buildtheimageandreputation(Benoit,1997;Coombs,2007a;Weberetal,
2011).Coombs(2007a)saysthebolsteringstrategyoffersa‘minimal
opportunitytodevelopreputationalassets’whileotherresearchhasindicated
thatbolsteringisnotjustoneofthemostusedcrisisresponsestrategies,but
possiblyalsooneofthemosteffective(Kimetal,2009;Sheldon&Sallot,2008;
Weberetal,2011).Thispastresearchandindicationsfromthereal-worldcrisis
managementpracticescouldbeanindicationofbolsteringbeingunderplayed
withintheSCCT,andthereforefurtherresearchisrequired.
Benoit’s(1995)imagerestorationtheory,onwhichtheSCCTisbuilt,
claimsthebolsteringstrategyisbestusedwhenanorganizationisaccusedof
wrongfulactionsandwantstoreducetheoffensivenessofthecrisis.Wecanfind
anexampleinthisintheresponsefromanExxonchairmanaftertheValdezoil
spill:‘Exxonhasmovedswiftlyandcompetentlytominimizetheeffectthisoil
willhaveontheenvironment,fishandotherwild-life’(Benoit,1997).Although
thiswasoneofthemostdisastrousoilspillsinhumanhistory,Exxonstill
thoughtthatthebolsteringstrategywouldprotecttheirreputation.
TheSCCTclaimsthebolsteringstrategyisbestusedtosupportoneofthe
threeprimarycrisisresponsestrategies(deny,diminish,rebuild).Itistherefore
notavalidresponsestrategybyitself,andshouldalwaysbeusedincombination
withoneofthethreeprimarystrategies.Bolsteringstrategiesarestrategiesofan
opportunisticnaturebecausetheycanonlybeusedwhenanorganizationhas
donepastgoodwork(Coombs&Holladay,p.41,2011).Thismeansthata
companythathasdonenogoodworksimplyhasnothinggenuinetobolster
aboutwhileacompanythat,forinstance,haskeptahighservicestandard,won
awardsoradaptedaCSRstrategy,canbolsteraboutanyorallofthesepastgood
deeds.Bolsteringworksbyoffsettingthenegativeinformationbyadding
positiveinformationandthereforerequiresthepresenceofpositiveinformation
(goodpre-crisisreputation)tobepresent(Coombs,2015).
Page 23
N.M.Leeflang--359167ErasmusUniversityRotterdam
23
Incrisiscommunicationpractice,itwasfoundthatbolsteringwasthe
mostoftenusedfirstcrisisresponseandthatitwasusedmostoftenin
combinationwiththedenialstrategy,followedbythecorrectiveactionstrategy.
Itwasalsofoundthatbolsteringwasmostoftenadoptedinthepreventable
cluster,meaninginsituationswithahighlevelofinternalattribution(Kimetal,
2009).Thisindicatesthatbolsteringisaprevalentstrategyinreal-lifecrisis
communication,asalsobecomesclearwhenlookingatothercasestudy
orientatedresearch,andcouldbeexplainedbytheorganizations’reluctanceto
share(only)negativeinformationaboutthemselves(Kim&Jung,2014;Sheldon
&Sallot,2008;Wan,2004;Weberetal,2011).Thisisunsurprisingasthe
incorporationofbolsteringcanreduceperceptionsofguiltandwrongdoingand
increasedtrustandtruthfulness(Kazoleas,2008)Furthermore,theimplication
thatbolsteringisusedmostoftenincombinationwiththedenialstrategy
directlycounterstherecommendationsmadeintheSCCTthatbolsteringworks
bestwiththediminishandrebuildstrategies,althoughitisnotdirectlystatedit
shouldnotbeusedwithdenial(Coombs,2007a;Kimetal,2009).
Astheprevalenceofthestrategyamongstpractitionersinsinuates,
bolsteringmightthusbeanimportanttoolincrisiscommunication.Itappears
bolsteringisanalmostnaturalreactionformanycorporationswhenfacedwitha
crisis.Ithas,despiteitsprevalenceamongstpractitioners,notbeenstudied
sufficientlyinacademicliterature,andthereforeitseffectivenessremainslargely
unproven(Ma&Zahn,2016).Otherresearchhasindicatedthatatwo-sided
responsethatcombinesanegativeaspect(thecrisis)withapositiveone
(bolstering)ismoresuccessfulthanone-sidedones(Kim&Sung,2014).
Furthermore,itisinterestingthatpractitionersusethestrategyofbolstering
mostincombinationwithdenial,whilethisisindirectcontrasttotheSCCT
(Coombs,2007a).Consideringtheuseofbolsteringinpracticethisresearchwill
aimtodiscoverbolstering’svalueincrisiscommunicationfurtherbylookingat
theuseofbolsteringincombinationwiththetwomostusedcrisisresponsesby
practitioners,thecorrectiveactionstrategywhichisrecommendedbytheSCCT
andthedenialstrategywhichistheoreticallyadvisedagainstintheSCCT.
Wehypothesizethatanorganizationthatincorporatesbolsteringinto
theircrisisresponsewillhaveahigherlevelofpostcrisiscorporatereputation
Page 24
MAMediaStudies:Media&BusinessMasterThesis
24
forbothstrategiesthanorganizationsthatdonotincorporateabolstering
strategy.
H1a:Organizationsthatincorporatethedenial+bolsteringstrategywill
haveahigherlevelofpositivepostcrisisreputationafterapreventablecrisis
scenariothanorganizationsthatonlyusethedenialstrategy.
H1b:Organizationsthatincorporatethecorrectiveaction+bolstering
strategywillhaveahigherlevelofpositivepostcrisisreputationaftera
preventablecrisisscenariothanorganizationsthatonlyusethecorrectiveaction
strategy.
2.5.Social-MediatedCrisisCommunication TheInternethasbecomeincreasinglyimportantforcrisis
communicators.Monitoringandreactingtocrisesunfoldingonlineisaneffective
strategyinidentifyingandmanagingcrises.TheSocial-MediatedCrisis
CommunicationModeldealswiththeinfluenceofmediatypeoncrisis
communication,whichwasnotincludedintheSCCT,andhighlightsthe
importanceofsocialmediaduringacrisis(Austin,Fisher&Jin,2012).Themodel
identifiesthewayinwhich(young)adultsseekoutsocialmediaafteracrisisfor
insiderinformationandusethemtocommunicatewithfriendsorfamily.The
modelwasoriginallynamedtheBlogMediatedCrisisCommunicationModel,but
wasrenamedafterresearchindicatedanincreaseofimportanceforplatforms
likeFacebookandTwitter,andadecreaseintheuseofblogs(Austinetal,2012;
Jinetal,2010;2011).
IntheSMMCAustin,Fisher&Jin(2012)positthatindividualswithlittle
stakeintheorganizationinitiateandfulfillPRactivitiesthroughonline
interactivity.Normally,gettingotherpeopletotalkaboutyourorganization
positivelyisamuchsought-afterPRtool.However,duringtimesofcrisisthis
alsomeansthatanyoneonsocialmediacanbecomeaninfluentialintimesof
crisisbyspreadingmessagesonline(secondarycrisiscommunications)totheir
followers.Researchhasindicatedthatonlinebehaviorinacrisisisoften
replicatedinofflinebehavior,meaningthatonlineopinionsarereproducedand
repeatedinofflineword-of-mouth,andvice-versa(Dutta-Bergman,2006).
Furthermore,duringcrisisthepublicwilllooktowardsorganizations’social
Page 25
N.M.Leeflang--359167ErasmusUniversityRotterdam
25
mediachannelsforinformation,incombinationwithtraditionalmediathatis
stillfavoredandtrustedoveranorganizationssocialmediachannel(Austinetal,
2012,Jinetal,2010;2011).Otherresultshaveindicatedthatsocialmediause
amongststakeholdersandtheorganizationincreasesduringacrisis,further
highlightingitsimportanceincrisiscommunication(Bietal,2014;Schultzetal,
2011;Macaisetal,2009;Veiletal,2011)
Thisstudywillfurtherexploretheuseofsocialmediaasatoolfor
reputationmanagementbyfocusingoncrisisresponsesonsocialmedia,in
specificFacebook.Incontrasttoface-to-faceinteraction,socialmediahas
changedthecommunicationactivityintoadynamicandreal-timeprocessin
whichoneusercansharetheiropinionswithmanywiththesingleclickofa
buttonandwithoutinterferenceofjournalistsorothergatekeepers(Bietal,
2014;Veiletal,2011).Bystudyingcrisisresponsesonsocialmedia,wewill
attempttofindoutwhattypeofonlinesecondarycrisiscommunication-
commenting,sharingorinforming-iselicitedbydifferentcrisisresponse
strategies.SecondaryCrisisCommunicationsarethusunderstoodasthe
willingnessofstakeholderstospreada‘messageofcrisis’.Ideally,an
organizationwouldprefertonotseemessagesabouttheircrisisspread.The
toneofthemessage(negativeorpositive)andfollowingbehavioralactionsare
discussedinthenextchapteraboutSecondaryCrisisReactions.
Wehypothesizethatorganizationsthatusethebolsteringstrategyin
combinationwithanothercrisisresponsestrategywillexperiencelessnegative
secondarycrisiscommunications.Incorporationofthebolsteringstrategywill
resultinlowerwillingnesstoengageinsecondarycrisiscommunicationon
socialmediatowardstheorganizationduetobolstering’scrisis-diminishing
attributes(Benoit,1995,Coombs,2007a;Sheldon&Sallot,2008;Weberetal,
2011).
H2a:Organizationsthatincorporatethedenial+bolsteringstrategywill
experiencelessnegativesecondarycrisiscommunicationafterapreventablecrisis
scenariothanorganizationsthatonlyusethedenialstrategy.
H2b:Organizationsthatincorporatethecorrectiveaction+bolstering
strategywillexperiencelessnegativesecondarycrisiscommunicationaftera
Page 26
MAMediaStudies:Media&BusinessMasterThesis
26
preventablecrisisscenariothanorganizationsthatonlyusethecorrectiveaction
strategy.
2.6.SecondaryCrisisReactionsTheSCCTisbuilttoofferguidelinesthatcouldbeusedtorepairthe
reputationandreducethenegativeeffectsforanorganizationafteracrisis.
Theseaspectshavebeendiscussedintheprevioussectionofthisessay,butthere
isonemoreaspecttheSCCTattemptstoreducethenegativefalloutof,namely
thesecondarycrisisreactionscausedbyacrisis(Coombs,2007a;Schultzetal,
2011;Weberetal.2011;Stockmyer,1996).Themainbehavioralintentionsthat
crisiscommunicationattemptstoinfluenceareminimizingreputationaldamage,
maintainingpurchasingintentionandpreventingnegativeword-of-mouth
(Coombs&Holladay,2014;Stockmyer,1996).
Researchintosecondarycrisisreactionsfollowingcrisesonsocialmedia
hasgainedinnumbersascrisesarepubliclydisplayedinanopendialogueform
onsocialmediasites(Coombs&Holladay,2014;Schultzetal.2011;Sweetzer&
Metzgr,2007).Theeaseofonlineword-of-mouthcommunicationmakessocial
mediaafacilitatoringrowingcrises,butalsoatoolforidentifying,managingand
communicatingaboutcrises(Coombs,2008;Coombs&Holladay,2014;Schultz
etal.2011).Organizationsmightreceivesupportorcondemnationfrompublics
onlineandfromthoseopinionsmighthypothesizeiftheircrisismanagement
strategyisrejectedoraccepted(Coombs,2007a&2008;Coombs&Holladay,
2014).
Indicatorsoftheacceptanceorrejectionofthecrisisaccountcouldbe
foundinonlineword-of-mouth(Secondarycrisiscommunication,asdiscussed
previously)andthetoneofsaidword-of-mouthorofflinepurchasingintentions
(Secondarycrisisreactions)(Coombs,2008).Oneexampleofasecondarycrisis
reactionisboycottingtheorganizationandpersuadingotherstodothesameor
byspreadingnegativeword-of-mouthabouttheorganization(Schultzetal,
2011).Anotheraspectofthesecondarycrisisreactionsispurchaseintention,
whichentailsthemotivationorintenttodobusinesswiththeorganizationinthe
future.Secondarycrisisreactionsarethusdistinctfromsecondarycrisis
Page 27
N.M.Leeflang--359167ErasmusUniversityRotterdam
27
communicationinthatthereactionsareintendedbehaviorsintheformof
purchaseintentandtoneofword-of-mouth(recommendingorcondemning)
whilethesecondarycrisiscommunicationsareabouthowthestakeholderuses
socialmediadirectlyfollowingacorporatecrisis.Secondarycrisis
communicationsarethustheactionsofcommenting,sharingorinformingpeers
onsocialmedia,howthemessageisspread,whilesecondarycrisisreactionsare
theactionstakenbythestakeholdertoeithersupportofcondemnthe
organizationintheformoftheirpurchasingintentionsandtoneofword-of-
mouth,whatissaidinthemessage.
Secondarycrisisreactionswillberesearchedinthispaperinpresenceor
absenceofthesecondarycrisisresponsestrategybolstering.Wehypothesize
thatincorporatingbolsteringwithoneoftheothercrisisresponsestrategieswill
resultinlessnegativesecondarycrisisreactions.Meaningthatrespondents
presentedwiththebolsteringresponseswillbelesslikelytotalknegatively
abouttheorganization.
H3a:Organizationsthatincorporatethedenial+bolsteringstrategywill
experiencelessnegativesecondarycrisisreactionafterapreventablecrisis
scenariothanorganizationsthatonlyusethedenialstrategy.
H3b:Organizationsthatincorporatethecorrectiveaction+bolstering
strategywillexperiencelessnegativesecondarycrisisreactionafterapreventable
crisisscenariothanorganizationsthatonlyusethecorrectiveactionstrategy.
Page 28
MAMediaStudies:Media&BusinessMasterThesis
28
Belowistheconceptualmodelforthispaper.
Figure2:Researchoverview
DenialStrategy+Bolstering
Correc3veAc3onStrategy+Bolstering
• SecondaryCrisisReac3ons
• SecondaryCrisisCommunica3on
• PostCrisisReputa3onH1a+
H1b+H2a-
H2b-
H3a-
H3b-
Page 29
N.M.Leeflang--359167ErasmusUniversityRotterdam
29
3.Methodology
ThisstudyconsistedofanonlineexperimentdesignedusingQualitrics
andaimedatinvestigatingtheeffectofthebolsteringstrategyonpostcrisis
reputation,secondarycrisiscommunicationandsecondarycrisisreactions.
Participantswerepresentedwithafictionalcrisisforanexistingcompany
(Sony)withagoodpriorreputation,asthisiswhenbolsteringismosteffective.
Respondents’werethenrandomlyassignedtooneofourfourcrisisresponse
conditions.Questionsrelatedtotheirpriorrelationshipandreputational
perceptionofSonywereincluded,aswellasanattributionofguiltcheck.Other
questionsrelatedtothemeasuresofthisstudy,likepost-crisisreputation,
secondarycrisiscommunication&reactionswereoperationalizedaswritten
below.
3.1.ResearchDesignAcademicresearchintocrisiscommunicationcangenerallybedivided
intotwomainmethods,thecasestudythatexaminesrealpastcrises,and
experiments(surveys)inwhichsubjectsarepresentedwithslightlydifferent
scenarios(conditions)fromwhichtheiropinionsandintentionsareasked
(Averyetal.2010).Becausethisstudywillexaminebehavioralintentions,under
theinfluenceofdifferentstimuli,ofstakeholdersonsocialmedia,wewillusean
experimentdesignintheformofanonlinesurvey.Thechoiceforan
experimentaldesignwasmadebecauseexperimentsofferustheopportunityto
studycausality,inthiscasetheinfluenceofa(secondary)crisiscommunication
strategy,inacontrolledandrelevantdesign(Averyetal,2010).
Thisresearchisdesignedbasedona1crisistype(Preventablecrisis)X2
primarycrisisresponsestrategies(Denial/Correctiveaction)X2theuseof
secondarycrisisresponsestrategies(usingbolstering/notusingbolstering)
between-subjectdesign,asshowninthefigurebelow.
Denial CorrectiveAction
NoBolstering Condition1 Condition3
WithBolstering Condition2 Condition4
Figure3:ResearchConditions
Page 30
MAMediaStudies:Media&BusinessMasterThesis
30
3.2.ExperimentalScenariosTheexperimentconsistedofonefictionalcrisisscenariointroducedina
shortnewsarticle,followedbyashortFacebookpostfromthecompany,the
initialcrisisresponse.Thecompany‘Sony’isaJapaneseproducerofconsumer
electronicsthatoperatesworldwide.The2016RepTrakreportplacesSonyas
number9intheworld’smostreputablecompanies(ReputationInstitute,
2016a).Weusedanexistingcompanybecausereputationisaconstructbuilt
overtimeandconstantlychanging,somethingwecouldnotadequatelysimulate
inourexperimentalsetting(Fombrun,Gardberg&Sever,2000).Sony’sgood
priorreputationandlackofanysimilarcrisishistoriesisanessentialaspectof
thebolsteringstrategy,andassuchwehavedecidedtouseSonyasourcompany
offocus.
Thefictionalcrisisthatweintroducedisbaseduponthecrisisrecently
experiencedbySony’scompetitor,Samsung,whohadissueswiththebattery
explodinginoneoftheirphonemodels.Forthisresearch,weintroduceda
similarcrisisinwhichSonylaunchedtheirnewlaptop,butafterwhichconsumer
reportedthebatterieswereexploding.ThereportalsomentionedthatSony
knowinglydidnottesttheirnewproductenoughbecausetheywantedtolaunch
ittogetherwiththenewApplecomputer,makingitapreventablecrisiswith
internalattribution(seeAppendixA1).
Afterreadingthereport,respondentswerepresentedwithoneofthefour
possibleresponsesbySonyonFacebook.Theseresponseswere‘denial’,‘denial+
bolstering’,‘correctiveaction’and‘correctiveaction+bolstering’(seeappendix
A2).Thedenialresponsesdeniedtheinternalattributionofthecrisis,whilethe
corrective-actionresponseattemptedtosolvetheissuebyofferingreplacement
partstostakeholdersaswellasanapology.Allresponsesincludedwordsof
sympathyandinformation,asrecommendedintheSCCTforcrisisresponses
(Coombs,2007a).
Thebolsteringaspectofthecrisisresponseswasconstructedbasedon
threefields.BolsteringtheproductbyhighlightingthatSonyproduces‘award
winning’laptops.BolsteringthecompanybyhighlightingthatSonyisdedicated
toitsCorporateSocialResponsibilityinitiativesandfinally,agenericcorporate
Page 31
N.M.Leeflang--359167ErasmusUniversityRotterdam
31
statementinwhichSonybolstersthattheyhavealwaysprovidedgoodservice
andwillcontinuetodosointhefuture.
Figure4:ExampleofFBpost-Denial+Bolstering
3.3.DataCollection Priortodistribution,aPre-testwasconductedtoensureclarityofthe
researchdesignandhighlightanypointsofconfusion.ThePre-testwas
conductedonasmallsampleof10respondentsandbasedontheirexperience
theexperimentwasadapted.Theseresultswerenotincludedinthefinaldataset.
Nomajorchangeswereneededandtheexperimentwasdistributed.
Thisexperimentwasdistributedonlineintheresearchersownsocial
medianetwork(Facebook)usinganon-probabilityconveniencesampling
techniqueinwhichrespondentsonsocialmediawereaskedtopartakeinashort
experiment(Fricker,2008;Heckathorn,2011).Thispartofthedatacollection
tookplacebetweenApril5th2017andApril21st2017andtherequestwas
postedonthreeseparateoccasionswithinthistimeperiod.Sincethisstudy
largelyconcernedsocialmedia,distributingourexperimentsonsocialmedia
ensuredthatwereceivedrespondentsthataresocialmediausers,thus
guaranteeingtheyarefamiliarwiththemedium.
ThesecondmethodofdatacollectionwasthroughtheAmazon
MechanicalTurkprogram.ThisonlytookplaceonApril21standalldatawas
collectedwithinafewhours.MechanicalTurkisanautomatizedserviceinwhich
researcherscaneasilycollectrespondentsfortheirsurveyorexperimentby
offeringasmallmonetarycompensationforevery‘hit’.MechanicalTurkhas
Page 32
MAMediaStudies:Media&BusinessMasterThesis
32
beenusedbymanyresearchersandisgenerallyconsideredavalidandreputable
sourcefordatacollection(Lowryetal,2016).MechanicalTurkparticipantsare
generallydemographicallymorediverseandthecollecteddatacanbetreatedas
atleastasreliableastraditionalmethods(Buhrmester,Kwang&Gosling,2011).
Theexperimentwasdesignedinsuchawaythattherespondentreceived
oneofthefourconditionsatrandom.Eachconditionrequiredanywhere
between30-40respondents,puttingthetotalrespondentsataround140,asis
recommendedforexperiments(Box,1980).
Atotalof121respondentswerecollectedthroughtheconvenience
samplingonsocialmediamethodovertheperiod5-21April2017.The
researcherpostedarequesttofillinthesurveyonthreeseparateoccasionson
hisownsocialmediaprofile.Aftercleaningthedataforincompleteresponses,
thisresultedinatotalof102validresponses.
60responseswerecollectedthroughtheAmazonMechanicalTurk
Programon21stofApril2017.Respondentswereaskedtofillinthesurveyin
exchangeforamonetaryrewardof0.25euro.Aftercorrectingforinvalid
responses,thisresultedin45validresponses.Thetotalnumberofvalid
responsesandourfinalsamplesizeisthus(N=)147.
3.4.Operationalization Theindependentvariablesinthisstudythatweremanipulatedinthe
experimentweretheprimaryresponsestrategiesandthesecondaryresponse
strategyofbolstering.
Thisstudyoperationalizedmultipleconceptsintodependentvariables,
namelypostcrisisreputationlevel,secondarycrisiscommunication,secondary
crisisreactions.Furthermore,thelevelofpriorreputationwillbedetermined
amongsttherespondentsaswellasatesttodeterminethelevelofguiltthey
attributedtotheorganizationandtheirownhistoryandrelationshipwiththe
company.
3.4.1.Pre-CrisisReputation
Pre-CrisisReputationwasmeasuredwithfourquestionsfromRepTrak.
Thesequestionsaimtoassessthepublic’sopinionofacompanyandmeasure
howmuchthepubliclikes,trustsandadmiresacompany(Fombrun;Ponzi&
Page 33
N.M.Leeflang--359167ErasmusUniversityRotterdam
33
Newburry,2015;Ponzi,Fombrun&Gardberg,2015).Theseweremeasured
usingaseven-pointLikertscaleandare:‘Ihaveagoodfeelingaboutthe
company’,‘Itrustthecompany’,‘Iadmireandrespectthecompany’and‘the
companyhasagoodoverallreputation’.Ponzi,FombrunandGardberg(2015)
reportedaCronbach’salphaof0.96forthesemeasuresbasedon4separate
studies.Thisindicatesahighreliabilityforthismeasure.
3.4.2.AttributionofGuilt,History&Relation
Respondentswereaskedthreequestionsrelatedtothelevelofperceived
organizationalresponsibilitybasedonthestudyperformedbyLee(2004).These
questionsweremeasuredona7-pointLikertscaleandwere:‘Towhatdegreedo
youthinktheorganizationistoblame’,Sonycouldhaveavoidedthecrisis’and
‘thecrisiswascausedbyaprobleminsidetheorganization’.InherresearchLee
(2004)reportedaCronbach’salphaof0.85,makingthisareliablescale.
Themeasureforcrisishistorywasonequestionrelatingtoifrespondents
knowofanypastsimilarcrises.Astherearenonewecouldfind,weexpect
respondentstonotknowofanysimilarcrises.
Relationwillbemeasuredthroughthreequestions,namely:‘Iown(ed)
Sonyproducts’,‘Sonyhasalwaysofferedgoodservicetome’and‘Iamsatisfied
withtheSonyproductsIown(ed)’.Thesewillbemeasuredona7-pointLikert
scale.
3.4.2.Post-CrisisReputation
Coombs&Holladay(2002)offerfivemeasuresonorganizational
reputation.Wewillusethesemeasurementstoassessthereputationafterthe
crisisresponseona7-pointLikertscaletodeterminethedifferentinfluencesof
thestimuli.Thesefivemeasuresare:‘Theorganizationisconcernedwiththe
well-beingofitspublics’,‘Theorganizationisbasicallydishonest’,‘Idonottrust
theorganizationtotellthetruthabouttheincident’,‘Undermostcircumstances,
Iwouldbelikelytobelievewhattheorganizationsays’and‘Theorganizationis
notconcernedwiththewell-beingofthepublics’.Thesefivemeasureswere
reportedtohaveCronbach’salphaof0.86byCoombs&Holladay(2002).
Page 34
MAMediaStudies:Media&BusinessMasterThesis
34
3.4.3.SecondaryCrisisCommunication
Secondarycrisiscommunicationwasmeasuredusingthreeindicators
developedbySchultzetal(2011)butwasadaptedslightlyforourfocuson
Facebook.Subjectswereaskedhowlikelytheyareonaseven-pointLikertscale
tosharethemessageonFacebook,informtheirpeersonFacebookandtoleavea
reactionorcommentonFacebook.Notethatthissectiondoesnotconcernthe
opinionofthestakeholderbutmerelytheirreactionandwillingnesstotake
certainactions.Sotheactionofinforming,reactingorsharingandnotthe
reflectionoftheirownopinions.Schultzetal(2011)didnotreportaCronbach’s
alphaforthismeasure.
3.4.4.SecondaryCrisisReactions
SecondarycrisisreactionsmeasurementswereadaptedfromCoombs&
Holladay(2008).Theyofferthreemeasurementsofcrisisreactions:‘Iwould
encourageothersnottobuyfromorganizationX’,‘Iwouldsaynegativethings
aboutOrganizationXtootherpeople’and‘IwouldrecommendOrganizationXto
someonewhoaskedmyadvice’.Thesequestionsthusaimtodistinguishthe
respondents’opiniononthecompanyfollowingthecrisis.Thesemeasureswere
alsomeasuredusingaseven-pointLikertscale.ReportedCronbach’salphafor
thesemeasureswas0.86.
Purchaseintentionsweremeasuredusingthreemeasures.These
measureshavebeenadaptedfromStockmyers’(1996)purchaseintention
measures.Stockmyers’measuresdealwithfuturebehavioralintent,andfocuson
‘wouldpurchaseagain’,‘likelihoodofpurchase’and‘continuationofpurchasein
thefuture’.Thesewerealsomeasuredonaseven-pointLikertscale.Stockmyer
(1996)reportedaCronbach’salphaof0.91
3.5.ManipulationCheckAsamanipulationcheckweaskedaquestionthatrelatestoifthe
participantnoticedthebolsteringactinthecrisisresponseoftheorganization.
Thisquestionwasaskedaftertherespondenthadbeenpresentedwiththe
conditionattheendofthesurveytoavoidprimingthesubjects(Kidd,1976)
Hereby,itwaslogicalifrespondentsthathadaconditionwithnobolstering
answeredthattheydidnotnoticeanybolstering,whiletheoneswhodidfillin
Page 35
N.M.Leeflang--359167ErasmusUniversityRotterdam
35
oneofthebolsteringconditionswilllikelyhavenoticedit.Basedonthis
manipulationcheck,weformeda‘robustdataset’(N=53)fromwhichwe
comparedtheresultwithourfulldataset.
3.6.ComputedVariables:Reliability&FactorAnalysis Onceourdatawascollectedweproceededtocomputeouroverarching
measuresbasedontheoperationalizedquestionsfoundintheprevioussection.
Areliabilityanalysiswasconductedoneachmeasuretocheckforinternal
consistencyofthescales.Herebyanalphaof>0.7couldbeconsidered
satisfactory(Bland&Altman,1997).
AfterourreliabilityanalysisweconductedaPrincipalComponentFactor
Analysis(Varimax)todeterminetheunderlyingdimensionsandhelpus
constructouroverarchingmeasures.
3.6.1.Pre-CrisisReputation
Themeasure‘Pre-CrisisReputation’wasbasedonthefourquestions
derivedfromtheRepTraksystemofmeasuringreputation(Cronbach’sα=
0.889).OurFactoranalysisonthesefourmeasuresfoundthatonecomponent
hadaneigenvalueof3.022andexplained75.6%ofthevariance.Theother
componentswerenegligible(eigenvalue<1).
Table3.6.1:Pre-CrisisReputationFactorLoadings
IhaveagoodfeelingaboutSonyasacompany 0.819
ItrustSony 0.881
IadmireandrespectSony 0.884
IthinkSonyhasagoodoverallreputation 0.890
Basedonthisanalysiswecomputedanewoverarchingmeasure:Pre-Crisis
Reputation.Fromourquestionsthatmakeupthismeasurewecantreat‘Pre-
CrisisReputation’asascaleinwhich1=mostnegative,4=neutraland7=most
positive.
3.6.2.Relation
Themeasure‘Relation’wasbasedonthethreeoperationalizedquestions
relatingtothepastrelationshipoftherespondentwithSony(Cronbach’sα=
Page 36
MAMediaStudies:Media&BusinessMasterThesis
36
0.861).Ourfactoranalysisofthethreemeasuresfoundthatonecomponenthad
aneigenvalueof2.349andexplained78.3%ofthevariance.Theother
componentswerefoundtobenegligible(eigenvalue<1).
Table3.6.2:RelationFactorLoadings
Sonyhasalwaysofferedgoodservicetome. 0.884
IamsatisfiedwiththeSonyproductsIown. 0.875
Sonyoffersgoodvalueformymoney 0.895
Basedonthisanalysiswecomputedanewoverarchingmeasure:Relation.From
ourquestionsthatmakeupthismeasurewecantreat‘Relation’asascalein
which1=mostnegative,4=neutraland7=mostpositive.
3.6.3.Attribution
Themeasure‘Attribution’wasbasedonthreequestionsderivedfrom
Lee’s(2004)questionstoassesstheperceivedlevelofguilt(attribution)
(Cronbach’sα=0.794).Ourfactoranalysisofthesethreemeasuresfoundthat
onecomponenthadaneigenvalueof2.128andexplained71%ofthevariance.
Theothercomponentswerefoundtobenegligible(eigenvalue<1).
Table3.6.3:AttributionFactorLoadings
TowhatdegreedoyouthinkSonyistoblameforthecrisis? 0.876
SonycouldhaveavoidedtheCrisis. 0.820
Thecrisiswascausedbyaprobleminsidetheorganization 0.830
Basedonthisanalysiswecomputedanewoverarchingmeasure:Attribution.
Fromourquestionsthatmakeupthismeasurewecantreat‘Attribution’asa
scaleinwhich1=internalguilty,4=neutraland7=notinternalguilty.
3.6.4.PostCrisisReputation
‘PostCrisisReputation’wasmeasuredaccordingto5questionsderived
fromresearchbyCoombs&Holladay(2002)(Cronbach’sα=0.815).Ourfactor
analysisofthesefivemeasuresfoundthatonecomponenthadaneigenvalueof
2.890andexplained58%ofthevariance.Theothercomponentswerefoundto
benegligible(eigenvalue<1).
Page 37
N.M.Leeflang--359167ErasmusUniversityRotterdam
37
Somequestionshadtheirscalereversedinourdatabecausetheywere
themselvesalreadyreversedquestions.Allfivequestionshadtofollowtherule
that1=mostnegativeand7=mostpositive.Questionsthatwerereversedhave
beenmarkedassuchinthebelowtableandwereincludedintheirreversedform
inthereliabilityandfactoranalyses.
Table3.6.4:PostCrisisReputationFactorLoadings
Sonyisconcernedwiththewell-beingofitsconsumers 0.759
Sonyisbasicallybeingdishonest.(REVERSED) 0.761
IdonottrustSonytotellthetruthaboutthesituation.(REVERSED) 0.832
UndermostcircumstancesIwouldbelikelytobelievewhatSonysays. 0.602
SonyisnotconcernedwiththeWell-beingofitsconsumers(REVERSED) 0.824
Basedonthisanalysiswecomputedanewoverarchingmeasure:PostCrisis
Reputation.Fromourquestionsthatmakeupthismeasurewecantreat‘Post-
CrisisReputation’asascaleinwhich1=mostnegative,4=neutraland7=most
positive.
3.6.5.SecondaryCrisisCommunication
ThemeasureSecondaryCrisisCommunicationwasbasedonthree
questionsderivedfromthestudybySchultz(etal,2011)aboutactionsonsocial
media(Cronbach’sα=0.862).OurFactoranalysisofthethreemeasuresfound
onecomponentwithaneigenvalueof2.352thatexplained78.41%ofthe
variance.Theothercomponentswerefoundtobenegligible(eigenvalue<1).
Table3.6.5:SecondaryCrisisCommunicationFactorLoadings
IwouldsharethepostonFacebook 0.910
IwouldinformmyfriendsorpeersonFB 0.897
IwouldcommentunderSony’sFBpost 0.849
Basedonthisanalysiswecomputedanewoverarchingmeasure:Secondary
CrisisCommunication.Fromourquestionsthatmakeupthismeasurewecan
treat‘SecondaryCrisisCommunication’asascaleinwhich1=unlikelyto
communicate,4=neutraland7=mostlikelytocommunicate.
Page 38
MAMediaStudies:Media&BusinessMasterThesis
38
3.6.6.SecondaryCrisisReactions
ThemeasureSecondaryCrisisReactionswasbasedonthreemeasures
introducedbyCoombs&Holladay(2008)andaimedtomeasurewhatpeople
mightcommunicateaboutthecrisis.Thereliabilityanalysisforthismeasure
resultedinaCronbach’salphaofα=0.674.Inaddition,deletinganyitemdoes
nothelpimprovethealphatothe0.7thresholds.Accordingtoourpreviously
statedcutoffofα=0.7forasatisfactorynumber,wecouldthereforesaythe
reliabilityofthismeasureisquestionable(Bland&Altman,1997).Considering
theproximitytoour‘satisfaction’numberandresearchthatindicatesanalphaof
0.6couldbeseenasacceptable,wewillhoweveracceptourlowerCronbach’s
alphabuthavetonotethereliabilityis‘questionable’(Loewenthal,p:60,2001;
Lance,Butt&Michels,2006).
Wecontinuedtoperformourfactoranalysisonthethreemeasuresand
foundonecomponentwithaneigenvalueof1.836thatexplained61.2%ofthe
variance.Theothercomponentswerefoundtobenegligible(eigenvalue<1).
Somequestionshadtheirscalereversedinourdatabecausetheywere
themselvesalreadyreversedquestions.Allthreequestionshadtofollowtherule
that1=mostnegativeand7=mostpositive.Questionsthatwerereversedhave
beenmarkedassuchinthebelowtableandwereincludedintheirreversedform
inthereliabilityandfactoranalyses.
Table3.6.6:SecondaryCrisisReactionsFactorLoadings
IwouldrecommendSonytosomeonewhoaskedmyadvice. 0.561
IwouldencourageothersnottobuySonyproducts.(REVERSED) 0.851
IwouldsaynegativethingstoothersaboutSony.(REVERSED) 0.892
Basedonthisanalysiswecomputedanewoverarchingmeasure:Secondary
CrisisReactions.Fromourquestionsthatmakeupthismeasurewecantreat
‘SecondaryCrisisReactions’asascaleinwhich1=negativereaction,4=neutral
and7=positivereaction.
3.7.Description&Demographics Thedatawecollectedconsistedof(N=)146respondentsofwhich52%
wasmaleand46.5%wasfemale(Gender:M=1.49,SD=0.528).31%of
Page 39
N.M.Leeflang--359167ErasmusUniversityRotterdam
39
respondentswerefromtheNetherlandsfollowedby19%fromtheUSAand13%
fromIndia.Theremaining27%ofrespondentswerefromothercountries.85%
hadboughtorownedSonyproductsinthepast(M=1.15,SD=0.359)and71%of
ourrespondentsclaimedtouseFacebookonadailybasis(M=4.35,SD=1.154).
33%ofrespondentsalsoreportedthatFacebookisanimportantsourceoftheir
news‘mostofthetime’(M=2.89,SD=1.3).Themedianageoftherespondents
was25yearsold(M=28.84,SD=10.032).
Table 3.7.1. Demographic Descriptives
AGE GENDER FB USE FB NEWS OWNED SONY
Mean 28.84 1.49 4.35 2.89 1.15
Median 25.00 1.00 5.00 3.00 1.00
Std. Deviation 10.032 .528 1.154 1.293 .359
Table 3.7.2. Correlation Matrix
1 2 3 4 5 6 7 8 9 10 11
1 DENIAL_OR_CORRECTIVE 1
2 BOLSTERING_WITH_OR_WITHOUT 0.013 1
3 PRE_CRISIS_REPUTATION 0.053 0.016 1
4 RELATION_SONY(N=124) 0.021 -0.01 .756** 1
5 ATTRIBUTION_CRISIS 0.097 -0.085 0.011 0.028 1
6 POST_CRISIS_REPUTATIOn -0.095 0.023 .327** .251** -.198* 1
7 SECONDARY_CRISIS_COMMUNICATIONS -0.01 -0.124 .166* .214* -0.08 .191* 1
8 SECONDARY_CRISIS_REACTIONS 0.203* 0.113 .281** .338** -0.112 .491** -0.138 1
9 Age 0.009 0.082 0.037 -0.013 0.123 .168* 0.106 0.141 1
10 Gender 0.065 -0.118 -0.1 -0.028 0.16 -0.054 -0.034 0.024 0.076 1
11 Facebook_Use 0.07 -0.16 0.088 -0.004 0.126 0.004 -0.059 0.12 -0.099 0.111 1
12 FB_as_News_Source -0.076 -0.089 0.055 0.02 -0.051 0.083 .290** -0.067 -0.106 .211* .308***Correlationissignificantatthe0.05level(2-tailed).
**Correlationissignificantatthe0.01level(2-tailed).
Ourcorrelationmatrixshowsthat‘RelationSony’iscorrelatedwith‘Pre-
CrisisReputation’atthep<.01levelandwithastrongstrengthofr=0.756.‘Post-
CrisisReputation’iscorrelatedwith‘Pre-CrisisReputation’and‘Relation’atthe
p<.01levelswithamediumrelation(r=0.327;r=0.252).PostCrisisReputationis
negativelycorrelatedwithAttributionatthep<.05levelswithasmallnegative
relation(r=-0.198).SecondaryCrisisCommunicationsiscorrelatedwithPre-Crisis
Rep,RelationandPostCrisisReputationatthep<.05levels.Aweakrelationwas
Page 40
MAMediaStudies:Media&BusinessMasterThesis
40
observed(r=0.166;r=0.214;r=0.191).SecondaryCrisisReactionsiscorrelated
withResponseStrategyatthep<.05levelwithaweakrelation(r=0.203).SCRis
alsocorrelatedwithPreCrisisReputation,RelationandPostCrisisReputationat
thep<.01levelandwithaweak-to-mediumrelation(r=0.281;0.338;0.491).Age
wasfoundtobecorrelatedwithPost-CrisisReputationatthep=.05levelwitha
weakrelation(r=0.168).FacebookasNewsSourcewasrelatedtoSecondaryCrisis
CommunicationsandFacebookUseatthep<.01levelwithamediumrelation(r=
0.290;r=0.308).FacebookasNewsSourcewasalsorelatedtoGenderatthep<.05
levelwithaweakrelation(r=0.211).
Toadjustoursixcomputedvariablesforvariancesinthescales,weused
SPSStostandardizeourvariables.Thesestandardized(z-score)variableswillbe
usedinthefollowinganalysischapterfortheregressions.
Page 41
N.M.Leeflang--359167ErasmusUniversityRotterdam
41
4.Results
Byaskingparticipantstocorrectlyidentifyeachoftheconditions,weare
abletoassessthesuccessofthemanipulations.Thiswasdonewithtwo
questionstowardstheendoftheexperiment.Onequestionaskedthe
respondenttoidentifythecorrectresponsestrategyoutoffouroptionsofwhich
twowerenotusedinanycondition.Thesecondquestionaskedrespondentsif
theynoticedanyBolsteringintheresponsebySony.Achi-squaretestconfirmed
thatthemanipulationofResponseStrategywassuccessful,χ2(4)=34.513,p<
.001.SuccessoftheBolsteringStrategywastestedinthesameway,χ2(2)=
37.111,p<.001.
4.1.Hypotheses1:TheeffectofBolsteringonPost-Crisis
ReputationToanalyzeiftheuseornotuseofBolstering(FixedFactor1)in
combinationwitheitheroneofourtworesponsestrategies(FixedFactor2)has
aneffectonPost-CrisisReputation(DV)weperformedtwo-wayanalysisof
variance(ANOVA).Levene’stestindicatedequalvariances(F=0.031,p=.993).
TheinteractionbetweenbolsteringandresponsestrategyonPost-Crisis
Reputationwasnotfoundtobesignificantatthep<.05level:F(1,142)=0.661,
p=.417.OurresultsalsoindicatedthatBolstering/NoBolstering(p=.767)and
DenialStrategy/CorrectiveActionStrategy(p=.269)allhavenosignificanteffect
onPost-CrisisReputation.
Table 4.1: Results of two-way ANOVA Post-Crisis Reputation (N= 146)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 1.240 1 1.240 1.232 .269 0.009
BOLSTERING_WITH_OR_WITHOUT .089 1 .089 .088 .767 0.001
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
.666 1 .666 .661 .417 0.005
Error 142.942 142 1.007
Total 145.000 146
*p<.1,**p<.05,***p<.01
Page 42
MAMediaStudies:Media&BusinessMasterThesis
42
ThiswouldmeanthatthedifferencebetweenusingDenialwith(M=
3.213,SD=0.922)orwithout(M=3.046,SD=0.948)bolsteringorCorrective
Actionwith(M=3.258,SD=0.852)orwithout(M=3.335,SD=0.907)bolsteringis
statisticallyinsignificantonPost-CrisisReputation.Becauseofthis,weaccepted
ournullHypothesis1thatstatesthereisnodifferenceoftheeffectbetweenthe
responsestrategiesandtheuseornouseofbolsteringonPost-CrisisReputation.
WetherebyalsorejectedbothourH1aandH1bastheresultsfromthisanalysis
indicatedthatusingcorrectiveaction+bolsteringordenial+bolsteringto
minimizethenegativeeffectsofapreventablecrisisonorganizationspost-crisis
reputationdonotyieldsignificantresultscomparedtothesamestrategies
withoutbolstering.
4.2.Hypotheses2:TheeffectofBolsteringonSecondaryCrisis
Communication Weperformedanothertwo-wayanalysisofvariancetoassessifthereis
aneffectoftheuseofBolsteringorNoBolstering(FixedFactor1)andourtwo
ResponseStrategies(FixedFactor2)onSecondaryCrisisCommunications(DV).
Levene’stestindicatesequalvariances(F=0.150,p=.929).Ourresultsforthis
analysisalsoindicatedthatthereisnointeractionbetweentheuseofBolstering
/NoBolsteringandDenial/CorrectiveActionStrategiesonSecondaryCrisis
Communicationatthep<.05significancelevel:F(1,142)=0.022,p=.882.The
resultsalsoindicatedtheseparateBolstering/NoBolsteringconditionwas
insignificant(p=.140),aswellastheDenial/CorrectiveActionstrategies(p=
.916).
Page 43
N.M.Leeflang--359167ErasmusUniversityRotterdam
43
Table 4.2.1: Results of two-way Secondary Crisis Communications (N= 146)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 0.11 1 0.11 0.11 0.916 0.00
BOLSTERING_WITH_OR_WITHOUT 2.213 1 2.213 2.201 0.140 0.015
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
0.022 1 0.022 0.022 0.882 0.00
Error 142.756 142 1.005
Total 145.000 146
*p<.1,**P<.05,***P<.01
ThiswouldmeanthatthedifferencebetweenusingDenialwith(M=
2.344,SD=1.557)orwithout(M=2.772,SD=1.705)bolsteringorCorrective
Actionwith(M=2.411,SD=1.475)orwithout(M=2.761SD=1.578)bolsteringis
statisticallyinsignificantonSecondaryCrisisCommunications.Ourresults
indicatedtheeffectofBolsteringandResponseStrategyonSecondaryCrisis
Communicationsisinsignificant.WethereforeacceptedthenullHypothesis2and
assumedtherenodifferenceoftheeffectbetweentheresponsestrategiesand
theuseornouseofbolsteringonSecondaryCrisisCommunications.Thismeans
wealsorejectedourH2aandH2bastheresultsindicatedthatusingcorrective
action+bolsteringordenial+bolsteringtominimizethenegativeeffectsofa
preventablecrisisintheformofSecondaryCrisisCommunicationsdonotyield
significantresultscomparedtothesamestrategieswithoutbolstering.
However,wealsonotedthattheincorporationofthebolsteringstrategy
isclosetosignificantatthep<.1levels(p=.14).Whenomittingthetypeof
responsestrategy,wecanseethatusingtheBolsteringstrategy(M=2.377,SD=
1.506)isslightlylessefficientinreducingthelikelihoodofSecondaryCrisis
CommunicationthannotusingBolstering(M=2.766,SD=1.63).Accordingtothis
result,bolsteringactuallyincreasesthelikelihoodofstakeholdersengagingin
SecondaryCrisisCommunicationscomparedtowhennotusingbolstering.We
reportedtheplotbelowtohighlightthisandshowthedifferenceinheight
betweenthetwoconditions.
Page 44
MAMediaStudies:Media&BusinessMasterThesis
44
Table4.2.2:BolsteringPlot
4.3.Hypotheses3:TheeffectofBolsteringonSecondaryCrisis
ReactionsAfinaltwo-wayanalysisofvariancewasperformedwiththeuseorno
useofbolstering(FixedFactor1)andresponsestrategies(FixedFactor2)on
SecondaryCrisisReactions(DV).Levene’stestindicatedequalvariances(F=
0.371,p=.774).Theresultstoldusthereisnosignificantinteractionoftheuseof
Bolstering/NoBolsteringandDenial/CorrectiveActiononSecondaryCrisis
Reactionsatthep<.05significancelevel:F(1,142)=0.23,p=.632.Theseparate
Bolstering/NoBolsteringconditionwasalsofoundtobeinsignificant(p=.178)
but,surprisingly,theDenial/CorrectiveActionconditionswerefoundtobe
significantforthisanalysis(p=.016).
Page 45
N.M.Leeflang--359167ErasmusUniversityRotterdam
45
Table 4.3.1: Results of two-way Secondary Crisis Reactions (N= 146)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 5.769 1 5.769 5.769 .016** 0.040
BOLSTERING_WITH_OR_WITHOUT 1.765 1 1.765 1.829 .178 0.013
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
0.222 1 0.222 .230 .632 0.002
Error 137.039 142 0.965
Total 145.000 146
*p<.1,**P<.05,***P<.01
ThiswouldmeanthatthedifferencebetweenusingDenialwith(M=
3.776,SD=1.063)orwithout(M=3.621,SD=1.09)bolsteringorCorrective
Actionwith(M=3.428,SD=0.992)orwithout(M=3.103SD=1.119)bolsteringis
statisticallyinsignificantonSecondaryCrisisReactions.Theseresultsindicated
theeffectofBolsteringandResponseStrategyonSecondaryCrisisreactionsis
insignificant.ThenullHypothesis3isthereforeacceptedandwecanassume
thereisnodifferenceoftheeffectbetweentheresponsestrategiesandtheuseor
nouseofbolsteringonSecondaryCrisisReactions.Thismeanswealsorejected
ourH3aandH3bastheresultsindicatedthatusingcorrectiveaction+bolstering
ordenial+bolsteringtominimizethenegativeeffectsofapreventablecrisisin
theformofSecondaryCrisisReactionsdoesnotyieldsignificantresults
comparedtothesamestrategieswithoutbolstering.Whatisinterestingtonote
fromthisanalysisisthat,althoughbolsteringandbolsteringincombinationwith
ourresponsestrategiesisinsignificant,itappearsthattheeffectoftheresponse
strategyonSecondaryCrisisReactionsissignificant.Evenmoreinterestingisthe
findingthatitappearsasiftheDenialStrategy(M=3.696,SD=1.072)ismore
effectivethantheCorrectiveActionstrategy(M=3.257,SD=1.066)inminimizing
negativeSecondaryCrisisReactions.
However,wealsonotedthattheincorporationofthebolsteringstrategy
isclosetosignificantatthep<.1levels(p=.178).Whenomittingthetypeof
responsestrategy,wecanseethatusingtheBolsteringstrategy(M=3.602,SD=
1.036)isslightlymoreefficientinreducingthelikelihoodofSecondaryCrisis
Page 46
MAMediaStudies:Media&BusinessMasterThesis
46
ReactionsthannotusingBolstering(M=3.355,SD=1.128).Accordingtothis
result,bolsteringdecreasesthelikelihoodofstakeholdersengaginginSecondary
CrisisCommunicationscomparedtowhennotusingbolstering.Wereportthe
plotbelowtohighlightthisandshowthedifferenceinheightbetweenthetwo
conditions.
Table4.3.2:BolsteringPlot
4.4.AdditionalResultsonPost-CrisisReputation ToinvestigatetheeffectonPost-CrisisReputation(DV)byotherfactors,
weperformedaLinearRegressionanalysis.WewantedtodiscoverhowPre-
CrisisReputation,LevelofAttribution,Bolstering,ResponseStrategy,Age,
Gender,FBUseandFBNews(IV’s)influencethePost-CrisisReputationandin
whatway.ThePre-CrisisReputation,Attribution,FBUseandFBnewsIV’sfor
thisregressionanalysisandtheonesfollowingarestandardizedvariables.The
samegoesforthePost-CrisisReputationDVinthisanalysisandtheSCCandSCR
variablesinthenextanalyses.Asignificantregressionequationwasfound(F(8,
145)=4.117,p=.00)withanR2of0.194.
4.4.1.Pre-CrisisReputation
Pre-CrisisReputationwasasignificantpredictorofPost-CrisisReputation(β=
0.321,t=4.126,p=.00).PostCrisisReputationincreasedwith0.321(ona7-point
Page 47
N.M.Leeflang--359167ErasmusUniversityRotterdam
47
scalefrom1=negativeto7=positive)foreverypointincreaseinPre-Crisis
Reputation.ThismeansastrongerPre-CrisisReputationwillminimizethenegative
effectsofacrisisonPost-CrisisReputation.
4.4.2.LevelofAttribution
Levelofattribution(howguiltytheorganizationisperceived)wasa
significantpredictorofPost-CrisisReputation(β= -0.213, t=-2.666, p=.009).ThismeansthatforeveryincreaseinperceivedAttributionofthecrisis,
Post-CrisisReputationdecreaseswith-0.213.Thismeansthatthemore
attribution(guilt)canbeascribedtoanorganization,thelowerthe
organization’sreputationwillbePost-Crisis.
4.4.3.Age
AgeoftherespondentappearstobeanothersignificantpredictorofPost-
CrisisReputation(β= 0.19, t= 2.466, p=.016).Post-CrisisReputationincreased0.019foreveryextrayear(Age)oftherespondent.Thismeansthat
olderpeople,thosewithmoreyears,appeartobemorelenientwhenitcomesto
judgingacompanyduringacrisis,aswecanseefromtherelationshipwithPost-
CrisisReputation
4.4.4.Gender,Bolstering,ResponseStrategy&Facebook Gender(p=.875),theuseornouseofBolstering(p=.9),thetypeof
ResponseStrategy(p=.264)andtheamountofFacebookUsed(p=.934)orthe
importanceofFBasaNewsSource(p=.42)wereallnotfoundtobesignificant
predictorsofPost-CrisisReputation.Gendertherebyseemstonotmatteratall
withbothmalesandfemalesbehavingsimilarly.InterestingisthatBolstering
andResponseStrategyseemtohavenoeffect,whichwealreadyknewfromour
previousanalysis.
Page 48
MAMediaStudies:Media&BusinessMasterThesis
48
Table4.4:Post-CrisisReputationInfluencers
Unstandardized
B(Effect)
Standard
Error
Standardized
B
t-value p(sig.)
Pre-CrisisRep 0.321 0.078 0.321 4.126 0.00***
Attribution -0.213 0.080 -0.213 -2.666 0.009***
Age 0.019 0.008 0.192 2.446 0.016**
Gender -0.024 0.154 -0.013 -0.158 0.875
ResponseStrategy -0.174 0.155 -0.034 -1.121 0.264
(No)Bolstering -0.020 0.157 -0.087 -0.125 0.9
FBUse 0.007 0.083 0.007 0.082 0.934
FBNews 0.068 0.084 0.068 0.809 0.420
R-Square 0.194
F-Test 4.117 0.000
*p<.1,**P<.05,***P<.01
4.5.AdditionalResultsonSecondaryCrisisCommunicationsToinvestigatetheeffectonSecondaryCrisisCommunication(DV)by
otherfactors,weperformedaLinearRegressionanalysis.Wewantedtodiscover
howPre-CrisisReputation,LevelofAttribution,Bolstering,ResponseStrategy,
Age,Gender,FBUseandFBNews(IV’s)influenceSecondaryCrisis
Communicationsandinwhatway.Asignificantregressionequationwasfound
(F(8,137)=3.783,p=.000)withanR2of0.181.
4.5.1.Pre-CrisisReputation
Pre-CrisisReputationwasasignificantpredictorofSecondaryCrisis
Communications(β=0.147,t=1.877,p=.063),butonlywhenweadjustthe
significanceleveltop<.1.SecondaryCrisisCommunicationsincreasedwith0.147
(ona7-pointscalefrom1=Unlikelyto7=Likely)foreverypointincreaseinPre-
CrisisReputation.ThismeansastrongerPre-CrisisReputationwillactually
increasethelikelihoodofsharing,commentingorinformingfollowingacrisis.
4.5.2.BolsteringStrategy
TheuseoftheBolsteringstrategywasasignificantpredictorofSecondaryCrisis
Reactions(β=-0.302,t=-1.912,p=.58),butonlywhenweadjustthesignificance
Page 49
N.M.Leeflang--359167ErasmusUniversityRotterdam
49
leveltop<.1.SecondaryCrisisCommunicationsdecreasedwith0.302(ona7-
pointscalefrom1=Unlikelyto7=Likely)whenusingtheBolsteringstrategy.
ThismeansusingtheBolsteringstrategywilldecreasethelikelihoodofsharing,
commentingorinformingfollowingacrisis.Thisiscontrarytoourprevious
analysis,whichprovedthebolsteringstrategyhasnoeffectonSecondaryCrisis
Communications.
4.5.2.Attribution,Age,Gender&ResponseStrategy
OurregressionanalysisfoundnosignificanteffectforAttribution
(p=.453),Age(p=.127),Gender(p=.241)andResponseStrategy(p=.661).It
appearsasifthechosenelementsdonotpredictSecondaryCrisis
Communications.
4.5.3.Facebook&SecondaryCrisisCommunications
TheamounttherespondentusedFacebook(β=-0.176,t=-2.098,p=.038)
andtheimportanceofFacebookastheirnewssource(β=0.358,t=4.247,p=.00)
werebothsignificantpredictorsofSecondaryCrisisCommunications.FBUsage
andNewsSourcewerecodedasafrequencyrangingfromdailytonever.
LikelihoodofSecondaryCrisisCommunicationsdecreased-0.176themore
frequenttherespondentwasonFB(FBUse)andincreasedwith0.358themore
importantFBwasasanewssource(FBNews)totherespondent.Thismeans
thatthemorefrequentarespondentwasonFB,thelesslikelytoshare,comment
orinform.Vice-versaistrueforFBasNewsSourceastheresultsindicatethe
moreimportantFBwasanewssourcetoourrespondents,themorelikelythey
weretoshare,commentorinform.
Page 50
MAMediaStudies:Media&BusinessMasterThesis
50
Table4.5.1:SecondaryCrisisCommunicationsInfluencers
Unstandardized
B(Effect)
Standard
Error
Standardized
B
t-value p(sig.)
Pre-CrisisRep 0.147 0.079 0.147 1.877 0.063*
Attribution -0.61 0.080 -0.061 -0.753 0.453
Age 0.015 0.008 0.148 1.869 0.064*
Gender -0.183 0.155 -0.096 -1.178 0.241
ResponseStrategy -0.069 0.157 -0.034 0.439 0.661
(No)Bolstering -0.302 0.158 -0.151 -1.912 0.058*
FBUse -0.176 0.084 -0.176 -2.098 0.038**
FBNews 0.358 0.084 0.358 4.247 0.00***
R-Square 0.181
F-Test 3.783 0.000
*P<.1,**P<.05,***P<.01
4.6.AdditionalResultsonSecondaryCrisisReactionsToinvestigatetheeffectonSecondaryCrisisReactions(DV)byother
factors,weperformedaLinearRegressionanalysis.Wewantedtodiscoverhow
Pre-CrisisReputation,LevelofAttribution,Bolstering,ResponseStrategy,Age,
Gender,FBUseandFBNews(IV’s)influenceSecondaryCrisisReactionsandin
whatway.Asignificantregressionequationwasfound(F(8,137)=4.08,p=.000)
withanR2of0.192.
4.6.1.Pre-CrisisReputation
Pre-CrisisReputationwasasignificantpredictorofSecondaryCrisis
Reactions(β=0.266,t=3.416,p=.001).SecondaryCrisisReactionsincreasedwith
0.266(ona7-pointscalefrom1=negativeto7=positive)foreverypoint
increaseinPre-CrisisReputation.ThismeansastrongerPre-CrisisReputation
willminimizethenegativeeffectsofacrisisonSecondaryCrisisReactions.
4.6.1.Attribution
AttributionofguiltwasasignificantpredictorofSecondaryCrisis
Reactions(β=-0.179,t=-2.249,p=.026).SecondaryCrisisReactionsincreased
with0.179(ona7-pointscalefrom1=negativeto7=positive)foreverypoint
Page 51
N.M.Leeflang--359167ErasmusUniversityRotterdam
51
increaseinAttributionofguilt.Thismeansmoreattributionofguiltwillincrease
thenegativeeffectsofacrisisonSecondaryCrisisReactions.
4.6.3.ResponseStrategy
TypeofResponseStrategy(DenialorCorrectiveAction)wasasignificant
predictorofSecondaryCrisisReactions(β=0.355,t=2.281,p=.024).Secondary
CrisisReactionsincreasedwith0.355(ona7-pointscalefrom1=negativeto7=
positive)whenusingtheDenialStrategy(ResponseStrategy).Thismeansthe
DenialstrategywillminimizethenegativeeffectsofacrisisonSecondaryCrisis
Reactions.Thisalsobecameclearfromourpreviousanalysis
4.6.4.FacebookUseThefrequencyofFacebookusewasasignificantpredictorofSecondary
CrisisReactions(β=0.166,t=1.997,p=.048).SecondaryCrisisReactions
increasedwith0.166(ona7-pointscalefrom1=negativeto7=positive)when
frequenciesofFBuseincreases.ThismeanspeoplewhofrequentlyuseFacebook
arelesslikelytospreadthenegativeeffectsofacrisisonSecondaryCrisis
Reactions.
4.6.5.Age,Gender,Bolstering&FBNews
Age(β= 0.014, t= 1.787, p=.076)couldbeconsideredsignificantwhenadaptingoursignificanceleveltop<.1.Agedisplaysasmallprediction,
meaninglikelihoodofSCRactionsincreasesby0.014foreveryyearolderthe
respondentwas.Gender(p=.348),FBasNewsSource(p=.152)andtheuseorno
usesofBolstering(p=.185)wereallnotfoundtobesignificantpredictorsof
SecondaryCrisisReactions.
InterestingisthatAttribution(guiltoforganization)wasfoundsignificant
andnegativelyrelatedtoSCR.Thismeansmoreguiltresultsinahigher
likelihoodofstakeholdersspreadingnegativemessagesabouttheorganization.
Genderseemstonotmatteratallwithbothmalesandfemalesbehaving
similarly.InterestingisthatBolsteringseemstohavenoeffectwhiletypeof
responsestrategydoesseemtomatterforSecondaryCrisisReactions.
Page 52
MAMediaStudies:Media&BusinessMasterThesis
52
Table4.6:SecondaryCrisisReactionsInfluencers
Unstandardized
B(Effect)
Standard
Error
Standardized
B
t-value p(sig.)
Pre-CrisisRep 0.266 0.078 0.266 3.416 0.001***
Attribution -0.179 0.080 -0.179 -2.249 0.026**
Age 0.014 0.008 0.141 1.787 0.076*
Gender 0.145 0.154 0.077 0.942 0.348
ResponseStrategy 0.355 0.155 0.178 2.281 0.024**
(No)Bolstering 0.209 0.157 0.105 1.333 0.185
FBUse 0.166 0.083 0.166 1.997 0.048**
FBNews -0.121 0.084 -0.121 -0.121 0.152
R-Square 0.164
F-Test 4.549 0.00
*P<.1,**P<.05,***P<.01
4.7.RobustnessCheck Theresultsintheprevioussectionsindicateourassumptions
(hypotheses)appeartoberejected.Tocontrolourresults,wewillperforma
robustnesscheckinwhichweonlyincluderesultsinwhichtherespondentgot
ourattentioncheckright.Thismeansthatweonlyincludedataforwhicha
respondentwho,forexample,hadDenial+Bolsteringasaconditionalso
answeredtheattentioncheckaboutthetypeofresponsestrategyandthe
questionaboutthepresenceofbolsteringright.Thesetypesofrobustness
checksareacommonpracticeinempiricalresearchandaimtodistinguish
evidenceofstructuralvaliditybyomittingcertaininfluencersthatmighthave
compromisedthedataset(Xu&White,2014).Inthiscase,weomitresponsesfor
whichwearenotsureiftherespondentpaidenoughattentionbyfilteringfor
correctattentionchecks.Thisresultedinadatasetof53(N=53)respondents.
4.7.1.H1:PostCrisisReputation
Weperformedthesametypeofanalysis(two-wayANOVA)onthis
correcteddatasetasinsection4.1.Theinteractionbetweenresponsestrategies,
useofbolsteringandpostcrisisreputationwasonceagainfoundinsignificantat
Page 53
N.M.Leeflang--359167ErasmusUniversityRotterdam
53
thep>0.05level:F(3,49)=2.396,p=.079.Levene’stestindicatedequal
variances(F=0.272,p=.845).Noothersignificantresultswerefound.
Table4.7.1:Results of two-way ANOVA Post-Crisis Reputation (Robust N= 53)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 3.837 1 3.837 3.721 .060* 0.071
BOLSTERING_WITH_OR_WITHOUT 1.098 1 1.098 1.065 .307 0.021
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
1.506 1 1.506 1.461 .233 0.029
Error 50.522 49 1.031
Total 58.093 53
*P<.1,**P<.05,***P<.01
However,wecanalsonotethatdifferentresponsestrategiesarecloseto
significantatthep<.1levels(p=.06).WhenomittingtheBolsteringstrategy
conditions,wecanseethatusingtheDenialstrategy(M=2.766,SD=1.007)isless
efficientinreducingnegativePost-CrisisReputationfalloutthanusingthe
CorrectiveActionStrategy(M=3.381,SD=0.869).Accordingtothisresult,
CorrectiveActionasaresponsestrategyresultsinahigherPost-Crisis
ReputationlevelthanwhenanorganizationwouldusetheDenialstrategy.We
reporttheplotbelowtohighlightthisandshowthedifferenceinheightbetween
thetwoconditions.
Page 54
MAMediaStudies:Media&BusinessMasterThesis
54
Table4.7.2:ResponseStrategyPlot(RobustN=53)
4.7.2H2:SecondaryCrisisCommunications
Weperformedthesametypeofanalysis(two-wayANOVA)onthis
correcteddatasetasinsection4.2.Theinteractionbetweenresponsestrategies,
useofbolsteringandSecondaryCrisisCommunicationswasonceagainfound
insignificantatthep>.05level:F(3,49)=1.460,p=.237.Levene’stestindicated
equalvariances(F=0.698,p=.558).Noothersignificantresultswerefound,
althoughwecouldnoteasimilarresultfortheBolsteringstrategyasinsection
4.3inwhichthesignificanceofthisfindingisneartothep<.1levelsanddisplays
thesamekindofinteraction:Bolsteringresponsesresultsinalowerlikelihoodof
engaginginSecondaryCrisisCommunications.
Table4.7.2:Results of two-way ANOVA Secondary Crisis Com. (Robust N= 53)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 0.990 1 0.990 1.571 .216 0.031
BOLSTERING_WITH_OR_WITHOUT 1.511 1 1.511 2.398 .128 0.047
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
0.018 1 0.018 0.028 .867 0.001
Error 30.863 49 0.630
Total 35.923 53
*P<.1,**P<.05,***P<.01
Page 55
N.M.Leeflang--359167ErasmusUniversityRotterdam
55
4.7.3H3:SecondaryCrisisReactions
Weperformedthesametypeofanalysis(two-wayANOVA)onthis
correcteddatasetasinsection4.3.Theinteractionbetweenresponsestrategies,
useofbolsteringandSecondaryCrisisReactionswasonceagainfound
insignificantatthep>.05level:F(3,49)=0.156,p=.926.Levene’stestindicated
equalvariances(F=2.762,p=.052).Noothersignificantresultswerefound.
Table4.7.3:Results of two-way ANOVA Secondary Crisis Reactions (Robust N= 53)
Sum of
Squares
df Mean
Square
F p η2
DENIAL_OR_CORRECTIVE 0.006 1 0.006 0.006 .940 0
BOLSTERING_WITH_OR_WITHOUT 0.01 1 0.01 0.01 .920 0
DENIAL_OR_CORRECTIVE *
BOLSTERING_WITH_OR_WITHOUT
0.468 1 0.468 0.462 0.5 0.009
Error 49.647 49 1.013
Total 50.864 53
*p<.1,**p<.05,***p<.01
4.7.3.AdditionalResultsfromtheRobustData
Asimilarlinearregressionasinsection4.4wasperformed.Asignificant
regressionwasfound:F(8,52)=3.632,p=.03withanR2of0.398.
PostCrisisReputationwasonceagainasignificantpredictor(β=0.492,
t=2.75,p=.009).PostCrisisreputationincreasedwith0.492(ona7-pointscale
from1=negativeto7=positive)foreverypointincreaseinPre-CrisisReputation.
Thisisinlinewithresultsfromourpreviousregressionalthoughtheeffectis
largerinourcontrolleddataset(0.492vs0.321).
Attribution(ofguilt)alsoseemstobeasignificantpredictorwhen
assumingp<.1(β=-0.295,t=-2,p=.052).Thismeansthatforeveryincreasein
perceivedattributionofthecrisis,Post-CrisisReputationdecreaseswith-0.295.
Thismeansthatthemoreattribution(guilt)canbeascribedtoanorganization,
thelowertheorganization’sreputationwillbePost-Crisis.Thiseffectisonce
againlargerinourcontrolleddata(-0.295vs-0.213).
Theothervariableswereallnotsignificant.
Table4.7.3.1:Post-CrisisReputationInfluencers(RobustN=53)
Page 56
MAMediaStudies:Media&BusinessMasterThesis
56
Unstandardized
B(Effect)
Standard
error
Standardized
B
t-
values
p(sig.)
Pre-CrisisRep 0.492 0.179 0.355 2.75 0.009***
Attribution -0.295 0.148 -0.266 -2 0.052*
Age 0.005 0.019 0.039 0.283 0.799
Gender -0.312 0.274 -0.149 -1.136 0.252
ResponseStrategy -0.285 0.268 -0.134 -1.062 0.285
(No)Bolstering -0.398 0.282 -0.19 -1.410 0.146
FBUse -0.004 0.152 -0.004 -0.27 0.979
FBNews 0.024 0.158 0.023 0.151 0.88
R-Square 0.397
F-Test 5.055 0.00**
*p<.1,**p<.05,***p<.01
Asimilarregressionasinsection4.5wasalsoperformed(secondary
crisiscommunications).Nosignificantregressionwasfound:F(8,52)=1.85,
p=.093withandR2of0.252.Theregressioncouldhoweverbeconsidered
significantatthep<.1level.
FacebookUseprovedtobetheonlysignificantvariablefromthis
regression(p=.033).ThismeanslikelihoodofSCCdecreasesby0.285themore
therespondentusedFacebook.Thisisinlinewithourpreviousresultthatthe
likelihoodofSecondaryCrisisCommunicationsdecreasesthemorefrequenta
personisonFacebook.Thiseffectwasalsolargerinourcontrolleddata(-0.285
vs-0.17).
Table4.7.3.2:SecondaryCrisisCommunicationInfluencers(RobustN=53)
Page 57
N.M.Leeflang--359167ErasmusUniversityRotterdam
57
UnstandardizedB
(Effect)
Standarderror StandardizedB t-values p(sig.)
Pre-CrisisRep 0.046 0.152 0.044 0.304 0.762
Attribution -0.137 0.125 -0.162 -1.094 0.280
Age 0.013 0.016 0.129 0.849 0.4
Gender -0.187 0.233 -0.117 -0.803 0.426
ResponseStrategy -0.202 0.228 -0.125 -0.888 0.379
(No)Bolstering -0.287 0.240 -0.18 -1.2 0.236
FBUse -0.285 0.129 -0.354 -2.203 0.033**
FBNews 0.141 0.134 0.175 1.048 0.3
R-Square 0.252
F-Test 1.85 0.093*
*P<.1,**P<.05,***P<.01
Finally,asimilarregressionasinsection4.6wasperformed(secondary
crisisreactions).Nosignificantregressionwasfound:F(8,52)=1.456,p=.102
withanR2of0.209.Nosignificantresultswerefoundinthisregressionandthey
arethereforenotreportedhere.
Page 58
MAMediaStudies:Media&BusinessMasterThesis
58
5.Discussion
ThisresearchwasgroundedintheSituationalCrisisCommunication
Theory(Coombs,2007a),whichisatheorythataimstoprovidepractitionersin
thereputationindustrywithguidelinesonhowtorespondtoorganizational
crisesbasedonacademicresearch.Thistheoryimpliesthatthebolstering
strategy(asecondarycrisisresponse)isasupplementalstrategytobeusedwith
oneoftheprimarycrisisresponsesthatoffersa“minimalopportunitytodevelop
reputationalassets”.(Coombs,2007a)Thisstatementwascontraryto
indicationswereceivedfromotherresearch.Mostnotablewasanextensive
studyofcasestudiesincrisiscommunicationthatindicatedthat,inpractice,the
mostusedresponsestrategywasinfactbolstering.Bolsteringwasusedmost
oftenincombinationwiththedenialorcorrectiveactionstrategies(Kimetal,
2009).Otherresearchintothematteralsoindicatedthatbolsteringmighthave
beenunderplayedintheSCCTbutdidnotstudythestrategyoritseffectiveness
inminimizingnegativereputationfalloutfollowingacrisisspecifically(Sheldon
&Sallot,2008;Wan,2004;Weberetal,2011).Thepremiseofthisstudywas
thereforetostudythebolsteringstrategyastheoryandpracticeseemedto
displayadifferenceinascribedeffectiveness.
Ourresearchsetouttoinvestigatetheinfluenceoftheuseofbolsteringas
asupplementarycrisisresponsetechniqueonthreedifferentfactorsthatmight
beinfluencedbyacrisis.ThesefactorswerePost-CrisisReputation,Secondary
CrisisCommunicationsandSecondaryCrisisReactions.Theeffectofthe
bolsteringstrategywasmeasuredusinganexperimentalresearchdesignwith
fourconditionsinwhichtwoinitialcrisisresponses(denialandcorrective
action)incombinationwithbolstering(usingbolsteringornotusingbolstering)
werepresentedtorespondentsrandomly(N=146).Ourresultsindicatethat
thereisnosignificanteffectoftypeofresponsestrategyandtheincorporationof
bolsteringonanyofthethreefactors.Wedidfindconfirmationofexisting
researchthatindicates(apositive)Pre-CrisisReputationisthemostimportant
factorinminimizingnegativereputationfalloutdirectlyfollowingacrisis.Our
researchquestionwas:
Page 59
N.M.Leeflang--359167ErasmusUniversityRotterdam
59
Towhatextent,ifany,doesthebolsteringstrategyaffectpostcrisisreputation,
secondarycrisiscommunicationsandreactionsduringtimesoforganizational
crisis?
Basedonthefindingsfromthisresearchwecananswerourresearch
questionandsaythatthereisnone,ornegligible,effectofthebolsteringstrategy
onpost-crisisreputation,secondarycrisiscommunicationsandsecondarycrisis
reactions.
TheSCCTincorporatesanumberoffactorsrelatedtoreputationfor
whichitaimstominimizethenegative:Post-CrisisReputation,SecondaryCrisis
CommunicationsandSecondaryCrisisReactions.Hypotheseswereconstructed
aroundthesefactorsandthesewillbediscussedbelow,aswellasotherfindings,
basedonourresultsandanalysis.
5.1.H1:Bolstering&Post-CrisisReputation TherecommendationsmadeintheSCCTareaimedatminimizingthe
negativeeffectsofacrisisonpostcrisisreputation(Coombs,2007a).Thismeans
thatusingtherightstrategywillresultinmoreforgivingstakeholderswith
betteropinionsoftheorganizationthanwhenusingthewrongstrategy.The
designofourresearchsomewhatchallengedtheserecommendationsaswe
basedourconditionsonwhatpractitionersinthefieldofcrisiscommunication
usemostoften(Kimetal,2009).Wechoosedenial(notrecommendedtobeused
withbolsteringintheSCCT)andcorrectiveaction(recommendedintheSCCT)
asourtwoprimaryresponsestrategiesandtestedtheseincombinationor
absenceofbolsteringasasecondarycrisisresponsestrategy(supplement).
Basedontheprevalenceofbolsteringamongstpractitioners,mostused
duringpreventablecrises,wehypothesizedthatincorporatingbolsteringwith
eitherofourprimaryresponsestrategieswouldresultinlessreputational
damageafterthecrisis:ahigherpost-crisisreputationthanwhennot
incorporatingthebolsteringstrategy.Basedonourfindingswerejectedourfirst
hypothesesandcanthereforestatethattheincorporationofbolsteringhasno
significanteffectonPost-CrisisReputation.Itwasfoundthatthetypeofprimary
responsestrategyandtheincorporationofbolsteringdonothelpinlimiting
negativefalloutontothepost-crisisreputationfactor.
Page 60
MAMediaStudies:Media&BusinessMasterThesis
60
Wedidfindaresultsignificantatthep<.1levelsinourrobustdatasetthat
onlyincludedcorrectanswersfromourmanipulationchecks.Thisresult
indicatedthatusingthecorrectiveactionresponsestrategycauseshigherPost-
CrisisReputationthanwhenusingthedenialstrategy.Thiscouldbeduetothe
factthattheserespondentspaidbetterattention,hencetheygotthe
manipulationchecksright,andasaresultweremoreawareoftheinternalguilt
oftheorganizationwhichasaconsequencemadethedenialstrategyless
effectivethanthestrategythataimstoadmitandcorrectafaultbyan
organization.ThisisalsoinlinewithrecommendationsmadeintheSCCT
(Coombs,2007a).
5.2.H2:Bolstering&SecondaryCrisisCommunications ThesecondthingtheSCCTaimstominimizeistheintenttocommunicate
aboutthecrisis(Coombs,2007a).However,theSCCTislimitedinthis
incorporation,asitdoesnotspecificallydealwithmediatype.Socialmediaisan
increasinglyimportantsourceofnewsandcommunicationtool,andoffersthe
opportunitytoeasilyreacttoorspread,amongstothers,organizationalcrises.
Tosupplementthis,weusedtheSocialMediatedCrisisCommunicationModel
thatdealswithhowstakeholderscommunicateonsocialmediafollowingacrisis
(Austinetal,2012).
WehypothesizedthattheincorporationoftheBolsteringsecondarycrisis
responsestrategywouldminimizethelikelihoodofourrespondentstoshare,
commentorinformpeersaboutthecrisis,effectivelyhelpingtocontainthe
crisis.Ourresultshoweverindicatethatthetypeof(primary)responsestrategy
andtheincorporationofabsenceofbolsteringhavenoeffectonsecondarycrisis
communications.Wethereforerejectedoursecondhypothesesand,basedon
thisstudy,canstatethatresponsestrategyandbolsteringdonotinfluencethe
likelihoodofourrespondentstoengageinsecondarycrisiscommunications.
Wedidhoweverfindaresultthatcouldnotbeconsideredsignificantat
thep<.1levels,butcomesclosetoit.Wealsofoundasimilarsignificancelevelin
therobustdatasetanalysisforSCC.ThisresultindicatedthatusingBolsteringin
theinitialcrisisresponseresultsinalowerlikelihoodofstakeholdersengaging
inSecondaryCrisisCommunicationsthanstrategiesthatdonotuseBolsteringin
Page 61
N.M.Leeflang--359167ErasmusUniversityRotterdam
61
thecrisisresponse.Thiscouldindicatethebolsteringstrategyiseffective,but
remainsunprovenandcouldthereforewarrantfutureresearch.
5.3.H3:Bolstering&SecondaryCrisisReactions ThefinalfactortheSCCTaimstominimizethenegativefallouton
followingacrisisisthesecondarycrisisreactions(Coombs,2007a).Secondary
CrisisReactionsrefertothebehavioralintentionsastakeholdermighthave
followingacrisisandincludethewaytheymighttalkaboutanorganization
(negativeorpositiveword-of-mouth)oradapttheirpurchaseintentions
followingacrisis(Schultzetal,2011;Stockmyer,1996).TheSCCTaimsto
minimizethespreadofnegativeopinionsaboutanorganizationandalsoto
ensurepeoplekeeponbuyingtheorganizationsproducts.
OurhypothesisassumedtheincorporationoftheBolsteringstrategy
wouldresultinlessnegativesecondarycrisisreactionsamongstour
respondentsbutwasfoundtobeincorrect.Wethereforerejectedourthird
hypothesesandcanstatethat,basedonourresearch;incorporatingbolstering
doesnothelpreducethenegativeeffectsofacrisisonsecondarycrisisreactions.
Wedidhoweverfindasignificantresultfortheresponsestrategyonsecondary
crisisreactions.Thedenialstrategyprovedtobe(slightly)moreeffectivethan
thecorrectiveactionstrategyinreducingsecondarycrisisreactions.Asurprising
resultasthiswouldmeandenyingattributionduringacrisishelpstoreducethe
intenttospreadnegativeword-of-mouthoradaptpurchaseintentions.This
resultwasalsofoundduringfurtheranalysis.Thisisacounter-intuitiveresult
andcouldwarrantfutureresearchintothematter.
Inthisanalysiswealsofoundaresultthatcouldnotbeconsidered
significantatthep<.1levels,butcomesclosetoit.Thisresultalsoindicatedthat
usingBolsteringintheinitialcrisisresponseresultsinalowerlikelihoodof
stakeholdersengaginginSecondaryCrisisReactionsthanstrategiesthatdonot
useBolsteringinthecrisisresponse.Thisresultwasnotreplicatedinourrobust
analysis,however.Aswehavenowfoundtworesultsthatareclosetosignificant
thatindicateBolsteringmightbeeffectiveinatleastminimizingthelikelihoodof
stakeholdersengaginginSCCorSCC,wehighlyrecommendfutureresearchto
furtherexplorethisinteraction.
Page 62
MAMediaStudies:Media&BusinessMasterThesis
62
5.4.Pre-CrisisReputationShieldsOrganizationsfromNegative
ReputationFallout ResearchintotheimportanceofPre-CrisisReputation,thereputation
heldbyanorganizationpriortoacrisis,indicatedthathavingapositivepre-
crisisreputationcanhelpminimizethenegativeeffectsofacrisisonthe
reputationheldbyandorganizationafteracrisis(Claeys&Cauberghe,2015;
Coombs,2007a;Coombs&Holladay,2001;2002;2006;Sheldon&Sallot,2009).
Thissocalled‘Halo-Effect’(Coombs&Holladay,2006)statesthatapositivepre-
crisisreputation(‘beinggood’)canactasashieldinreflectingnegativefallout
ontothepost-crisisreputationbecausestakeholdersarereluctanttochange
theirowncognitionorexpectationbias.
Ourresearchconfirmsthisonmultiplelevels.Pre-CrisisReputationwas
foundtobeasignificantpredictor(p<.05)ofPost-CrisisReputationand
SecondaryCrisisReactions.Thismeansthatahigher(morepositive)pre-crisis
reputationwillminimizethenegativeeffectsofacrisisonpost-crisisreputation
(perceptionofthecompanypost-crisis)andsecondarycrisisreactions(toneof
communication).Asignificanteffectwasalsofoundatalowersignificancelevel
(p<.1)forthepredictivepowerofPre-CrisisReputationonSecondaryCrisis
Communications.ThismeansthatahigherPre-Crisisreputationalsoreducesthe
likelihoodofourrespondentstoshare,commentorinformpeersonthecrisis
andeffectivelyaidsincontainingthecrisis.
Thesefindingsareinlinewithpreviousresearchandwecanconfirmwith
almostcertaintythatthemostimportantfactorinreducingtheinitialnegative
effectsofacorporatecrisisishavingapositivereputationpriortosaidcrisis.
Pre-CrisisReputationdoesindeedactasashieldagainstcorporatecrisesand,as
ourresultsindicate,willresultinahigherreputationafterthecrisis,lesslikely
stakeholderswhomightsharenegativeinformationandalowerlikelihoodof
communicatingaboutsaidcrisistotheirpeers.
Theseimplicationsarenotnovelforbothacademicsorprofessionalsbut
doconfirmthatitisimportant,ifnotvital,fororganizationstomaintaina
positivereputationasitwillprotecttheorganizationfromnegativityfollowinga
crisisandalsoresultinbenefitsforthecompanywhennotincrisis(vanRiel&
Fombrun,2007,ch2).
Page 63
N.M.Leeflang--359167ErasmusUniversityRotterdam
63
5.5.OtherFindings Wefoundmultipleothereffectsbesidespre-crisisreputationofour
measuresonPost-CrisisReputation.Attribution,thelevelofguiltascribedby
ourrespondenttotheorganization,wasapredictorofpost-crisisreputation.The
mainfindingherewasthattheguiltiertherespondentthoughttheorganization
was,thelesspositivetheywouldbeabouttheorganizationafterthecrisis.This
isinlinewiththeSCCT,whichstatesthatahigherlevelofattributionresultsina
greaterreputationalthreatfortheorganization’spost-crisisreputation(Coombs,
2007a).
Anotherfactorfoundtobeasignificantpredictorofpost-crisisreputation
wastheageoftherespondent.Itwasfoundthattheolderrespondentsappeared
tobemorelenienttowardsthecompanyandascribedthemwithahigherpost-
crisisreputationthantheiryoungercounterparts.Thisissomethingthathasnot
beenstudiedyetinthefieldofreputationmanagement,andthereforewecannot
giveavalidatedreasonforthisfinding.Wecan,however,imagineanolder
respondentmighthavebecomemore‘jaded’thantheyoungeronesdueto
possiblybeingexposedtomorecrisesintheirlife.
Inrelationtosecondarycrisiscommunicationsandsecondarycrisis
reactionswefoundthatourrespondentswerelesslikelytospreadinformation
oropinionduringacrisiswhentheywereonFacebookmorethanthe
respondentswhoreportedbeingonFacebookless.Thisisagainacounter-
intuitivefindingaswemighthaveexpectedpeoplewhoareontheplatform
moretoalsobemorelikelytoexpressthemselvesinonewayoranotheronit.
WealsofoundthatthemoreimportantFacebookisasasourceofnews,the
morelikelyourrespondentsweretoshare,commentorinformpeersabouta
crisis.ThislastresultwasonlyfoundforSCCandnotforSCR.
5.6.Conclusion Basedonthisresearchwecananswerourresearchquestionandstate
thattheeffectofthebolsteringstrategyonpost-crisisreputation,secondary
crisiscommunicationsandsecondarycrisisreactionsisnegligible,ifnotnon-
existent.WecantherebyconfirmtheSCCTisrightwhenitstatesbolsteringisa
minimalopportunityforreputationalgainsandwouldrecommendpractitioners
Page 64
MAMediaStudies:Media&BusinessMasterThesis
64
incrisiscommunicationtoconsiderthiswhenrespondingtocrises.This
researchimpliesthatbolsteringhaslittleeffectbesidesallowingpractitionersto
putsomepositiveinformationnexttothenegativeone,butinrealityhasvery
littleeffectontherespondents’opinionsandactionstowardstheorganization
followingthecrisisresponse.Bolsteringisthereforemore‘puffery’andnotso
mucha‘underratedstrategy’.
Whatwedidfind,andthisisimportantforacademicsbutespeciallyfor
practitioners,isthatthemosteffectivewayofreducingnegativefalloutfollowing
acrisisisthereputationanorganizationholdsbeforesaidcrisis.Thisfinding
reconfirmsexistingliteraturethatstatesapositivereputationwillactasa
‘shield’duringorganizationalcrisesandisinlinewithwhatisincorporatedin
theSCCTandotherreputationorcrisiscommunicationresearch.Whatis
interestingtonoteisthattheuseofdifferentresponsestrategiesdidnotseemto
mattergreatlyeitherinourinitialcrisis-responseresearch,althoughitisstated
intheSCCTthatitshould.
5.7.Limitations&FutureResearch Thisresearchwassubjecttoafewlimitationsthat,whilenotdetrimental,
dolimittheimpactofthisstudyslightly.Oneofthemainlimitationswasthe
scopeandsizeofthisstudy.Oursampleof146respondentsisenoughforan
experimentalstudyofthisnature,butahighersamplewouldincreasethe
validityandpossiblygiveusdifferentresults.Ourrespondentsarealsofroma
varietyofnationalitiesandcultures.Futureresearchmightwanttofocuson
specificregionstoaccountforculturalandpersonalfactorswhenstudyingcrisis
communication.Anotherlimitationwecouldnoteistheabsenceofanynovel
findingsinthisstudy.Ourresultsre-confirmthetheorythatPre-Crisis
Reputationshieldsorganizationsduringcrises.
WedidfindthattheincorporationoftheBolsteringstrategyhadan
almostsignificanteffectonbothSecondaryCrisisCommunicationsand
SecondaryCrisisReactions.Becauseoftheproximitytoasignificancelevel,we
recommendfutureresearchfurtherexploretheeffectofbolsteringincrisis
containment,possiblyinadifferentresearchdesign.
Page 65
N.M.Leeflang--359167ErasmusUniversityRotterdam
65
Whilewedonotknowifadifferentresearchdesignmightgiveus
differentresults,itcouldbethatourdesignandconditionswerenotrefined
enoughforthisresearch.Ourfocusononlytwocrisisresponsestrategies,one
singlecompanywithagoodreputationandacrisisinthepreventableclusterin
combinationwiththebolsteringstrategyisalimitedscopeandfutureresearch
wouldbenefitfromamoreextensivestudythatincorporatesallpossibleaspects
andresponsesintheSCCT.Ouruseofonlyonecompanywithan(near)
impeccablepre-crisisreputationisalimitationofthisstudyasitmightbethat
thishighlevelofreputation‘shielded’ourcompany.Futureresearchmight
replicatethisstudywithacompanyoflowpriorreputationorincorporatea
comparativedesignbetweenhighandlowpre-crisisreputations.
Ideally,futureresearchwouldalsolookattheuseofthebolstering
strategyduringcrisiscommunicationovertime.Ourresearchwasbasedona
‘snapshot’ofacrisisandtheinitialcrisisresponsestrategy.Itmightbethat
bolsteringworksbetterovertimebycontinuouslyrepeatingpositivefactsto
contrastthenegativeones,therebyreducingtheoffensivenessofacrisisinlight
ofpastgooddeeds.Asthisishardtoreplicateinanexperimentalsetting,itcould
beinterestingtostudyareallifecrisisasitunfoldsbypresentingasurveyat
differentintervalstoaselectionofrespondents.Sucharesearchmightreveal
howacrisisunfolds,comestothepublic’sattention,whatthecompanydoesto
reducenegativereputationalfallout,howthepublicreactstoitandhowpublic
opinioncanchangeovertime.
Otherrecommendationsforfutureresearchwecouldmakearean
investigationintotherelationshipbetweenageandorganizationalcrisis,why
ourresultsindicatethedenialstrategypositivelyinfluencessecondarycrisis
reactionsandtherelationshipbetweenFacebookuseandsecondarycrisis
communicationsandreactions.Finally,consideringtheuseofbolsteringby
practitioners,futureresearchmightbenefitfromaqualitativedesignaimedat
understandingsaidpractitioners’motivationsandexperiencewiththe
bolsteringstrategy.
Page 66
MAMediaStudies:Media&BusinessMasterThesis
66
ReferencesAustin,L.,FisherLiu,B.,&Jin,Y.(2012).Howaudiencesseekoutcrisisinformation:Exploringthe
social-mediatedcrisiscommunicationmodel.JournalofAppliedCommunicationResearch,40(2),188-207.http://dx.doi.org/10.1080/00909882.2012.654498
Avery,E.J.,Lariscy,R.W.,Kim,S.,&Hocke,T.(2010).Aquantitativereviewofcrisiscommunicationresearchinpublicrelationsfrom1991to2009.PublicRelationsReview,36(2),190-192.http://dx.doi.org/10.1016/j.pubrev.2010.01.001
Benoit,W.L.(1997).Imagerepairdiscourseandcrisiscommunication.PublicRelationsReview,23(2),
177-186.http://dx.doi.org/10.1016/S0363-8111(97)90023-0Bi,G.,Zheng,B.,&Liu,H.(2014).Secondarycrisiscommunicationonsocialmedia:Theroleof
corporateresponseandsocialinfluenceinproduct-harmcrisis.InPACIS(p.93).http://aisel.aisnet.org/pacis2014/93
Bland,J.M.,&Altman,D.G.(1997).Statisticsnotes:Cronbach'salpha.Bmj,314(7080),572.doi:
https://doi.org/10.1136/bmj.314.7080.572Box,J.F.(1980).RAFisherandthedesignofexperiments,1922–1926.TheAmerican
Statistician,34(1),1-7.http://dx.doi.org/10.1080/00031305.1980.10482701Bundy,J.,&Pfarrer,M.D.(2015).Aburdenofresponsibility:Theroleofsocialapprovalattheonsetof
acrisis.AcademyofManagementReview,40(3),345-369.doi:10.5465/amr.2013.0027Bundy,J.,Pfarrer,M.D.,Short,C.E.,&Coombs,W.T.(2016).CrisesandcrisismanagementIntegration,
interpretation,andresearchdevelopment.JournalofManagement,doi.org/10.1177/0149206316680030
Buhrmester,M.,Kwang,T.,&Gosling,S.D.(2011).Amazon'sMechanicalTurkanewsourceof
inexpensive,yethigh-quality,data?.PerspectivesonPsychologicalScience,6(1),3-5.Brinson,S.L.,&Benoit,W.L.(1999).ThetarnishedstarrestoringTexaco’sdamagedpublic
image.ManagementCommunicationQuarterly,12(4),483-510.https://doi.org/10.1177/0893318999124001
Chiciudean.I.&David,G.(2013).Considerationsonusingthesituationalcrisiscommunicationtheory
inthecrisiscommunicationplanningactivitiesofRomanianarmedforces'informationandpublicrelationsstructures.JournalofDefenseResourcesManagement,4(1),159-167.
Choi,J.,&Chung,W.(2012).Analysisoftheinteractiverelationshipbetweenapologyandproduct
involvementincrisiscommunication:Studyonthetoyotarecallcrisis.JournalofBusinessandTechnicalCommunication.https://doi.org/10.1177/1050651912458923
Claeys,A.S.,&Cauberghe,V.(2015).Theroleofafavorablepre-crisisreputationinprotecting
organizationsduringcrises.PublicRelationsReview,41(1),64-71.https://doi-org.eur.idm.oclc.org/10.1016/j.pubrev.2014.10.013
Claeys,A.S.,Cauberghe,V.,&Vyncke,P.(2010).Restoringreputationsintimesofcrisis:An
experimentalstudyoftheSituationalCrisisCommunicationTheoryandthemoderatingeffectsoflocusofcontrol.PublicRelationsReview,36(3),256-262.https://doi.org/10.1016/j.pubrev.2010.05.004
Claeys,A.S.,&Opgenhaffen,M.(2016).Whypractitionersdo(not)applycrisiscommunicationtheory
inpractice.JournalofPublicRelationsResearch,1-16.http://dx.doi.org.eur.idm.oclc.org/10.1080/1062726X.2016.1261703
Page 67
N.M.Leeflang--359167ErasmusUniversityRotterdam
67
Coombs,W.T.(1999).Crisismanagement:Advantagesofarelationalperspective.InJ.A.Ledingham,&S.D.Bruning(Eds.),Relationshipmanagement:Arelationalapproachtothestudyandpracticeofpublicrelations(pp.75–93).Mahwah,NJ:LawrenceErlbaumAssociates.
Coombs,W.T.(2004).ImpactofpastcrisesoncurrentcrisiscommunicationinsightsfromSituational
CrisisCommunicationTheory.JournalofBusinessCommunication,41(3),265-289.https://doi.org/10.1177/0021943604265607
Coombs,W.T.(2006).Theprotectivepowersofcrisisresponsestrategies:Managingreputational
assetsduringacrisis.JournalofPromotionManagement,12(3-4),241-260.http://dx.doi.org/10.1300/J057v12n03_13
Coombs,W.T.(2007a).Protectingorganizationreputationsduringacrisis:Thedevelopmentand
applicationofsituationalcrisiscommunicationtheory.CorporateReputationReview,10(3),163-176.DOI:10.1057/palgrave.crr.1550049
Coombs,W.T.(2007b).Attributiontheoryasaguideforpost-crisiscommunicationresearch.Public
RelationsReview,33(2),135-139.http://dx.doi.org/10.1016/j.pubrev.2006.11.016Coombs,W.T.(2008).Crisiscommunicationandsocialmedia.EssentialKnowledgeProject,Institute
forPublicRelations.RetrievedonFebruary19,2017,fromhttp://www.instituteforpr.org/essential_knowledge/detail/crisis_communication_and_social_media/
Coombs,W.T.(2015).Thevalueofcommunicationduringacrisis:Insightsfromstrategic
communicationresearch.BusinessHorizons,58(2),141-148.http://dx.doi.org.eur.idm.oclc.org/10.1016/j.bushor.2014.10.003
Coombs,W.T.,&Holladay,S.J.(2001).Anextendedexaminationofthecrisissituations:Afusionof
therelationalmanagementandsymbolicapproaches.JournalofPublicRelationsResearch,13(4),321-340.Retrievedfrom:https://goo.gl/BgVlFF
Coombs,W.T.,&Holladay,S.J.(2002).Helpingcrisismanagersprotectreputationalassetsinitialtests
oftheSituationalCrisisCommunicationTheory.ManagementCommunicationQuarterly,16(2),165-186.https://doi.org/10.1177/089331802237233
Coombs,T.W.,&Holladay,S.J.(2005).Anexploratorystudyofstakeholderemotions:Affectandcrises.
InTheeffectofaffectinorganizationalsettings(pp.263-280).Bradford:EmeraldGroupPublishingLimited.
Coombs,W.T.,&Holladay,S.J.(2006).Unpackingthehaloeffect:Reputationandcrisismanagement.
JournalofCommunicationManagement,10(2),123-137.http://dx.doi.org/10.1108/13632540610664698
Coombs,W.T.,&Holladay,S.J.(2007).Thenegativecommunicationdynamic:Exploringtheimpactof
stakeholderaffectonbehavioralintentions.JournalofCommunicationManagement,11(4),300-312.http://dx.doi.org/10.1108/13632540710843913
Coombs,W.T.,&Holladay,S.J.(2008).Comparingapologytoequivalentcrisisresponsestrategies:
Clarifyingapology'sroleandvalueincrisiscommunication.PublicRelationsReview,34(3),252-257.http://dx.doi.org/10.1016/j.pubrev.2008.04.001
Coombs,W.T.,&Holladay,S.J.(Eds.).(2011).Thehandbookofcrisiscommunication(Vol.22).
Chichester:JohnWiley&Sons.Coombs,W.T.,&Holladay,S.J.(2014).Howpublicsreacttocrisiscommunicationefforts::Comparing
crisisresponsereactionsacrosssub-arenas.JournalofCommunicationManagement,18(1),40-57.http://dx.doi.org/10.1108/JCOM-03-2013-0015
Page 68
MAMediaStudies:Media&BusinessMasterThesis
68
Coombs,W.T.,Holladay,S.J.&Claeys,A.S.(2016).Debunkingthemythofdenial’seffectivenessincrisiscommunication:contextmatters.JournalofCommunicationManagement,20(4),381-395.http://dx.doi.org/10.1108/JCOM-06-2016-0042
Decker,W.H.(2012).Afirm'simagefollowingallegedwrongdoing:Effectsofthefirm'spriorreputationandresponsetotheallegation.CorporateReputationReview15(1),20-34.DOI:10.1057/crr.2011.27
Davey,N.(2010).Implosionofabrand:WhatcanwelearnfromToyota'scrisismismanagement?March,10,2014.Retrievedfromhttp://www.mycustomer.com/topic/customer-experience/implosion-brand-what-canwe-learn-toyotascrisis-mismanagement/103652
Dutta-Bergman,M.J.(2006).CommunityparticipationandInternetuseafterSeptember11:
Complementarityinchannelconsumption.JournalofComputer-MediatedCommunication,11(2),469-484.DOI:10.1111/j.1083-6101.2006.00022.x
Fombrun,C.,&Foss,C.(2004).Businessethics:Corporateresponsestoscandal.CorporateReputation
Review,7(3),284-288.DOI:10.1057/palgrave.crr.1540226Fombrun,C.J.,Gardberg,N.A.,&Sever,J.M.(2000).TheReputationQuotientSM:Amulti-stakeholder
measureofcorporatereputation.JournalofBrandManagement,7(4),241-255.DOI:10.1057/bm.2000.10
Fombrun,C.J.andvanRiel,C.B.M.(2004)Fame&Fortune:HowSuccessfulCompaniesBuildWinning
Reputations,NewYork:Prentice-HallFinancialTimes,Fombrun,C.J.,Ponzi,L.J.,&Newburry,W.(2015).Stakeholdertrackingandanalysis:TheRepTrak®
systemformeasuringcorporatereputation.CorporateReputationReview,18(1),3-24.doi:10.1057/crr.2014.21
Fricker,R.D.,Jr.(2012).Samplingmethodsforwebande-mailsurveys,In:J.Hughes,ed.,SAGE
InternetResearchMethods,London:SAGEPublications.ReprintedfromTheSAGEHandbookofOnlineResearchMethods,N.Fielding,R.M.LeeandG.Blank,eds.,chapter11,London:SAGEPublications,195-216.http://hdl.handle.net/10945/38713
Gregory,J.R.(1998).Doescorporatereputationprovideacushiontocompaniesfacingmarket
volatility?Somesupportiveevidence.CorporateReputationReview,1(3),288-290.doi:10.1057/palgrave.crr.1540050
Heckathorn,D.D.(2011).Comment:snowballversusrespondent-drivensampling.SociologicalMethodology,41(1),355-366.DOI:10.1111/j.1467-9531.2011.01244.x
Huang,Y.H.(2008).Trustandrelationalcommitmentincorporatecrises:Theeffectsofcrisis
communicativestrategyandformofcrisisresponse.JournalofPublicRelationsResearch,20(3),297-327.http://dx.doi.org/10.1080/10627260801962830
Jeong,S.H.(2009).Public'sResponsestoanoilspillaccident:Atestoftheattributiontheoryand
situationalcrisiscommunicationtheory.PublicRelationsReview,35(3),307-309.http://dx.doi.org/10.1016/j.pubrev.2009.03.010
Jin,Y.,&Liu,B.F.(2010).Theblog-mediatedcrisiscommunicationmodel:Recommendationsfor
respondingtoinfluentialexternalblogs.JournalofPublicRelationsResearch,22(4),429-455.DOI:10.1080/10627261003801420
Jin,Y.,Liu,B.F.,&Austin,L.L.(2011).Examiningtheroleofsocialmediaineffectivecrisis
management:Theeffectsofcrisisorigin,informationform,andsourceonpublics’crisisresponses.CommunicationResearch.41(1).74-94.https://doi.org/10.1177/0093650211423918
Page 69
N.M.Leeflang--359167ErasmusUniversityRotterdam
69
Kahn,W.A.,Barton,M.A.,&Fellows,S.(2013).Organizationalcrisesandthedisturbanceofrelationalsystems.AcademyofManagementReview,38(3),377-396.doi:10.5465/amr.2011.0363
Kaplan,T.(2005).TheTylenolcrisis:HoweffectivepublicrelationssavedJohnson&Johnson.URL:
http://www.grif.com.au/Tylenol-Poisonings.79.0.htmlKazoleas,D.(2008).Testingcrisisresponsestrategies:Acomparativequantitativeinvestigationofthe
impactofreputationalbolsteringonperceptionsofwrongdoingandtrustduringorganizationalcrises.PedagogyPosters108,pp49-58.https://goo.gl/T7Hfqc
Kelley,H.H.(1967).Attributiontheoryinsocialpsychology.InNebraskasymposiumonmotivation.Vol
15,192-238.Nebraska:UniversityofNebraskaPress.https://goo.gl/R7ES1nKelley,H.H.,&Michela,J.L.(1980).Attributiontheoryandresearch.Annualreviewof
psychology,31(1),457-501.DOI:10.1146/annurev.ps.31.020180.002325
Kidd,R.F.(1976).Manipulationchecks:advantageordisadvantage?.RepresentativeResearchinSocialPsychology.7,160-165.https://goo.gl/z9BaMi
Kim,S.,Avery,E.J.,&Lariscy,R.W.(2009).Arecrisiscommunicatorspracticingwhatwepreach?:An
evaluationofcrisisresponsestrategyanalyzedinpublicrelationsresearchfrom1991to2009.PublicRelationsReview,35(4),446-448.http://dx.doi.org/10.1016/j.pubrev.2009.08.002
Kim,S.,&Sung,K.H.(2014).Revisitingtheeffectivenessofbasecrisisresponsestrategiesin
comparisonofreputationmanagementcrisisresponses.JournalofPublicRelationsResearch,26(1),62-78.http://dx.doi.org.eur.idm.oclc.org/10.1080/1062726X.2013.795867
Knight,R.F.,&Pretty,D.J.(1999).Corporatecatastrophes,stockreturns,andtradingvolume.
CorporateReputationReview,2(4),363-378.doi:10.1057/palgrave.crr.1540092Lance,C.E.,Butts,M.M.,&Michels,L.C.(2006).Thesourcesoffourcommonlyreportedcutoffcriteria
whatdidtheyreallysay?.OrganizationalResearchMethods,9(2),202-220.Lee,B.K.(2004).Audience-orientedapproachtocrisiscommunication:AstudyofHongKong
consumers’evaluationofanorganizationalcrisis.Communicationresearch,31(5),600-618.https://doi.org/10.1177/0093650204267936
Luoma-aho,V.,Moreno,A.,&Verhoeven,P.(2017).CrisisresponsestrategiesinFinlandandSpain.
JournalofContingenciesandCrisisManagement.DOI:10.1111/1468-5973.12163Loewenthal,K.M.(2001).Anintroductiontopsychologicaltestsandscales.London:PsychologyPress.Lowry,P.B.,D’Arcy,J.,Hammer,B.,&Moody,G.D.(2016).“CargoCult”scienceintraditional
organizationandinformationsystemssurveyresearch:Acaseforusingnontraditionalmethodsofdatacollection,includingMechanicalTurkandonlinepanels.TheJournalofStrategicInformationSystems,25(3),232-240.
McDonald,L.M.,Sparks,B.,&Glendon,A.I.(2010).Stakeholderreactionstocompanycrisis
communicationandcauses.PublicRelationsReview,36(3),263-271.10.1016/j.pubrev.2010.04.004
Ma,L.,&Zhan,M.(2016).Effectsofattributedresponsibilityandresponsestrategieson
organizationalreputation:Ameta-analysisofsituationalcrisiscommunicationtheoryresearch.JournalofPublicRelationsResearch,1-18.http://dx.doi.org/10.1080/1062726X.2016.1166367
Macias,W.,Hilyard,K.,&Freimuth,V.(2009).Blogfunctionsasriskandcrisiscommunicationduring
HurricaneKatrina.JournalofComputer-MediatedCommunication,15(1),1-31.DOI:10.1111/j.1083-6101.2009.01490.x
Page 70
MAMediaStudies:Media&BusinessMasterThesis
70
Mahon,J.F.,&Wartick,S.L.(2003).Dealingwithstakeholders:Howreputation,credibilityand
framinginfluencethegame.CorporateReputationReview,6(1),19-35.DOI:10.1057/palgrave.crr.1540187
Mangold,W.G.,&Faulds,D.J.(2009).Socialmedia:Thenewhybridelementofthepromotion
mix.BusinessHorizons,52(4),357-365.http://dx.doi.org/10.1016/j.bushor.2009.03.002O'Brien,C.(2011).Theemergenceofthesocialmediaempoweredconsumer.IrishMarketing
Review,21(1/2),32.Retrievedfrom:https://goo.gl/Nt0RbAPark,H.,&Reber,B.H.(2010).Usingpublicrelationstopromotehealth:Aframinganalysisofpublic
relationsstrategiesamonghealthassociations.JournalofHealthCommunication,15(1),39-54.http://dx.doi.org/10.1080/10810730903460534
Ponzi,L.J.,Fombrun,C.J.,&Gardberg,N.A.(2011).RepTrak™pulse:Conceptualizingandvalidatinga
short-formmeasureofcorporatereputation.CorporateReputationReview,14(1),15-35.DOI:10.1057/crr.2011.5
ReputationInstitute.(2016a).2016GlobalRepTrakReport.RetrievedFebruary4,2017,from
ReputationInstitutewebsite:https://goo.gl/bvonPNReputationInstitute(2016b).Howvwlostthepublic’strust.RetrievedMay24,2017,from
https://www.reputationinstitute.com/volkswagen.aspxSarstedt,M.(2009).Reputationmanagementintimesofcrisis.JournalofBrandManagment,16,499-
503.DOI:10.1057/bm.2009.7Schultz,F.,Utz,S.,&Göritz,A.(2011).Isthemediumthemessage?Perceptionsofandreactionsto
crisiscommunicationviatwitter,blogsandtraditionalmedia.PublicRelationsReview,37(1),20-27.http://dx.doi.org/10.1016/j.pubrev.2010.12.001
Sheldon,C.A.,&Sallot,L.M.(2008).Imagerepairinpolitics:Testingeffectsofcommunication
strategyandperformancehistoryinafauxpas.JournalofPublicRelationsResearch,21,25–50.doi:10.1080/10627260802520496
Seeger,M.W.,Sellnow,T.L.,&Ulmer,R.R.(1998).Communication,organization,andcrisis.Annalsof
theInternationalCommunicationAssociation,21(1),231-276.http://dx.doi.org/10.1080/23808985.1998.11678952
Seeger,M.W.,Sellnow,T.L.,&Ulmer,R.R.(2003).CommunicationandOrganizationalCrisis.
Westport:GreenwoodPublishingGroup.Sisco,H.F.,Collins,E.L.,&Zoch,L.M.(2010).Throughthelookingglass:AdecadeofRedCrosscrisis
responseandsituationalcrisiscommunicationtheory.PublicRelationsReview,36(1),21-27.http://dx.doi.org/10.1016/j.pubrev.2009.08.018
Stockmyer,J.(1996).Brandsincrisis:Consumerhelpfordeservingvictims.NA-AdvancesinConsumer
ResearchVolume23.Retrievedfrom:https://goo.gl/9CbWqdTucker,L.,&Melewar,T.C.(2005).Corporatereputationandcrisismanagement:Thethreatand
manageabilityofanti-corporatism.CorporateReputationReview,7(4),377-387.DOI:10.1057/palgrave.crr.1540233
Turk,J.V.,Jin,Y.,Stewart,S.,Kim,J.,&Hipple,J.R.(2012).Examiningtheinterplayofanorganization's
priorreputation,CEO'svisibility,andimmediateresponsetoacrisis.PublicRelationsReview,38(4),574-583.http://dx.doi.org/10.1016/j.pubrev.2012.06.012
Page 71
N.M.Leeflang--359167ErasmusUniversityRotterdam
71
VanRiel,C.B.,&Fombrun,C.J.(2007).Essentialsofcorporatecommunication:Implementingpracticesforeffectivereputationmanagement.London:Routledge.
Veil,S.R.,Buehner,T.,&Palenchar,M.J.(2011).Awork-in-processliteraturereview:Incorporating
socialmediainriskandcrisiscommunication.JournalofContingenciesandCrisisManagement,19(2),110-122.DOI:10.1111/j.1468-5973.2011.00639.x
Verhoeven,P.,Tench,R.,Zerfass,A.,Moreno,A.,&Verčič,D.(2014).Crisis?Whatcrisis?:How
Europeanprofessionalshandlecrisesandcrisiscommunication.PublicRelationsReview,40(1),107-109.https://doi.org/10.1016/j.pubrev.2013.10.010
Verschoor,L.(2014)Crisiscommunication:Howtorespondtothepublic.(Masterthesis),Tilburg
University,Tilburg.Retrievedfrom:https://goo.gl/dNz5hv
Wan,H.H.(2004).Inoculation,Bolstering,andCombinedApproachesinCrisisCommunication.JournalofPublicRelationsResearch,6(3),301-328.Retrievedfrom:https://goo.gl/dYRhul
Weber,M.,Erickson,S.L.,&Stone,M.(2011).Corporatereputationmanagement:Citibank'suseof
imagerestorationstrategiesduringtheUSbankingcrisis.JournalofOrganizationalCulture,CommunicationandConflict,15(2),35.Retrievedfrom:https://goo.gl/eR2g8f
Weiner,B.(1985).Anattributionaltheoryofachievementmotivationandemotion.Psychological
Review,92(4),548.http://dx.doi.org/10.1037/0033-295X.92.4.548
Page 72
MAMediaStudies:Media&BusinessMasterThesis
72
AppendixA-experimentconditionsA1-FictionalNewsPost
Page 73
N.M.Leeflang--359167ErasmusUniversityRotterdam
73
A2-FictionalFacebookPosts
A2.1-Denial:WeregretthestoriesthatarecirculatingtheInternetandcanguaranteeourcustomersthatweareinvestigatingthematter.Ourmanagementfullheartedlydeniestheaccusationsthattheywereinformedofthemalfunctionpriortorelease.
A2.2-Denial+BolsteringWeregretthestoriesthatarecirculatingtheInternetandcanguaranteeourcustomersthatweareinvestigatingthematter.Ourmanagementfullheartedlydeniestheaccusationsthattheywereinformedofthemalfunctionpriortorelease.Ourlaptopsareofaward-winningqualityandwehavecontinuouslyworkedtoincreaseourcorporateresponsibilityefforts.WeatSonyhavealwaysaimedatprovidingourcustomerswithhighendproductsandservices,andwillcontinuetodosointhefuture.
A2.3-CorrectiveActionWeregretthereportsthatarecirculatingtheInternetandcanguaranteeourcustomersthatweareinvestigatingthematter.AnyonewhohasboughtaVAIO7Laptopcanexchangehisorherbatteryforanewonethroughoursite.
A2.4-CorrectiveAction+BolsteringWeregretthereportsthatarecirculatingtheInternetandcanguaranteeourcustomersthatweareinvestigatingthematter.AnyonewhohasboughtaVAIO7Laptopcanexchangehisorherbatteryforanewonethroughoursite.Ourlaptopsareofaward-winningqualityandwehavecontinuouslyworkedtoincreaseourcorporateresponsibilityefforts.WeatSonyhavealwaysaimedatprovidingourcustomerswithhighendproductsandservices,andwillcontinuetodosointhefuture.
Page 74
MAMediaStudies:Media&BusinessMasterThesis
74
AppendixB-ExperimentSurveyQ1.1Dearparticipant,Thissurveywillaskyouafewquestions.Pleasereviewalltextsbeforeansweringthequestions.Theestimatedtimetofinishthissurveywillbenolongerthan12minutesandallanswerswillbekeptconfidentialandanonymous.Pleasetakeyourtimetoanswerthesequestions.Therearenorightorwronganswers.Ifyouhaveanyquestionsregardingthesurvey,pleasecontactmeviae-mail:359167nl@eur.nlThankyouforyourparticipation.Q2AreyoufamiliarwiththeconsumerelectronicscompanySony?m Yesm Maybem NoQ5Pleaseindicatetowhatdegreeyouagreewiththefollowingstatements:Q3IhaveagoodfeelingaboutSonyasacompany.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ4ItrustSony.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ6IadmireandrespectSonyasacompany.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem Stronglyagree
Page 75
N.M.Leeflang--359167ErasmusUniversityRotterdam
75
Q7IthinkSonyhasagoodoverallreputation.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ8HaveyouboughtanySonyproductsinthepast?m Yesm NoDisplayThisQuestion:
IfHaveyouboughtanySonyproductsinthepast?YesIsSelectedQ9Sonyhasalwaysofferedgoodservicetome.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeDisplayThisQuestion:
IfHaveyouboughtanySonyproductsinthepast?YesIsSelectedQ10IamsatisfiedwiththeSonyproductsIown(ed).m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem Stronglyagree
Page 76
MAMediaStudies:Media&BusinessMasterThesis
76
DisplayThisQuestion:IfHaveyouboughtanySonyproductsinthepast?YesIsSelected
Q50Sonyoffersgoodvalueformymoney.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem Stronglyagree
Page 77
N.M.Leeflang--359167ErasmusUniversityRotterdam
77
Q51CorporateCrisesaresuddenandsignificanteventsthatcanhaveanegativeimpactonanorganizationsreputation.Q11AreyouawareofanycorporatecrisesSonyhashadinthepast?m Yesm NoDisplayThisQuestion:
IfAreyouawareofanycorporatecrisesSonyhashadinthepast?YesIsSelectedQ12Canyouverybrieflydescribethepastcrisisyouareawareof(max140characters).
Page 78
MAMediaStudies:Media&BusinessMasterThesis
78
Q13YouwillnowbepresentedwithanewsreportaboutacrisisexperiencedbySony.Pleasereadthisshortarticlecarefullyandproceedtothenextquestions.Theimagemaytakeafewsecondstoload.
Q14
Page 79
N.M.Leeflang--359167ErasmusUniversityRotterdam
79
Q15TowhatdegreedoyouthinkSonyistoblameforthecrisis?m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ16Sonycouldhaveavoidedthecrisis.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ56Thecrisiswascausedbyaprobleminsidetheorganization.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ19DoyouhaveaFacebookaccount?m Yesm No
Page 80
MAMediaStudies:Media&BusinessMasterThesis
80
Q21NextyouwillbepresentedwiththeresponsefromSonyonthesocialmediasiteFacebook.PleasetreatthisasifyouhadjustreadthenewsarticleandcameacrossthispostfromSonyonyourFacebooktimeline.Q20
Q17
Q22
Page 81
N.M.Leeflang--359167ErasmusUniversityRotterdam
81
Q23
Q24PleaseindicatetowhatdegreeyouagreewiththefollowingstatementsafterhavingreadthenewsarticleandresponsepostfromSony.Q27Sonyisconcernedwiththewell-beingofitsconsumers.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ28Sonyisbasicallybeingdishonest.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ29IdonottrustSonytotellthetruthaboutthesituation.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem Stronglyagree
Page 82
MAMediaStudies:Media&BusinessMasterThesis
82
Q30UndermostcircumstancesIwouldbelikelytobelievewhatSonysays.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ31Sonyisnotconcernedwiththewell-beingofitsconsumers.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ32PleaseindicatethelikelihoodofyouundertakinganyofthefollowingactionsafterhavingreadSony'sresponse.Q33IwouldsharethepostonFacebook.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym ExtremelylikelyQ34IwouldinformmyfriendsandpeersonFacebook.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym Extremelylikely
Page 83
N.M.Leeflang--359167ErasmusUniversityRotterdam
83
Q35IwouldcommentunderSony'spostonFacebook.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym ExtremelylikelyQ38Pleaseindicatehowlikelyyouwouldbetoundertakethefollowingactions.Q39IwouldencourageothersnottobuySonyproducts.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym ExtremelylikelyQ40IwouldsaynegativethingstoothersaboutSony.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym ExtremelylikelyQ41IwouldrecommendSonytosomeonewhoaskedmyadvice.m Extremelyunlikelym Moderatelyunlikelym Slightlyunlikelym Neitherlikelynorunlikelym Slightlylikelym Moderatelylikelym Extremelylikely
Page 84
MAMediaStudies:Media&BusinessMasterThesis
84
Q43ThelikelihoodofmepurchasingSonyproductsinthefutureishigh.m Stronglydisagreem Disagreem Somewhatdisagreem Neitheragreenordisagreem Somewhatagreem Agreem StronglyagreeQ52InthepostbySony,whichofthefollowingtypesofresponsesdidyounotice?m Denialm CorrectiveActionm Justificationm AttacktheAccuserQ45Pleasereadthefollowingdefinitionofbolstering:Bolstering:Usedtomitigatethenegativeeffectsofacrisisbystrengtheningtheaudience'spositiveideaoftheaccused.Theymayremindtheaudienceofpreviousgooddeedsorgoodreputation.Q46DidyounoticeanybolsteringintheFacebookpostbySony?m Yesm Maybem NoQ45Howoldareyou?Q46Whatisyourgender?m Malem Femalem RathernotsayQ55Whatisyournationality?Q47HowoftenareyouonFacebook?m Neverm Onceaweekm 2-3timesaweekm 4-6timesaweekm Daily
Page 85
N.M.Leeflang--359167ErasmusUniversityRotterdam
85
Q48IsFacebookanimportantsourceofyournews?m Alwaysm Mostofthetimem Abouthalfthetimem Sometimesm NeverQ53Howdidyoucomeacrossthissurvey?m Afriendm SocialMediam Websitem Emailm Other____________________Q47Thankyouforyourhelp,itismuchappreciated.Allanswersrecordedinthissurveyaretreatedwithconfidentiality.Thecrisisandresponsesbroughtforwardinthissurveyareentirelyfictional.NosuchsituationhaseveroccuredwithSonythatwecouldfind.