CorporateReputationintheVolkswagen EmissionsScandal TheRoleofInvolvement,Emotions,ResponsibilityandPerson-CompanyFit MasterThesis MasterMediaStudies–MediaandBusiness ErasmusSchoolofHistory,CultureandCommunication ErasmusUniversityRotterdam StudentName: StudentNumber: LouisaWanjek 437075 Supervisor: Dr.YijingWang June20,2016 CorporateReputationintheVolkswagenEmissionsScandal TheRoleofInvolvement,Emotions,ResponsibilityandPerson-CompanyFit Abstract Whilesomescholarsarguethatastrongreputationcanprotectanorganizationfrom reputationallossduringacrisis,othersbelievethatahighreputationleadstohigher expectationsamongthepublic,whichareviolatedduringacrisisandmaketheorganization suffer.Astheroleofcorporatereputationinthecrisiscontextisstillinconclusiveinliterature,it isthusofinteresttoinvestigatetowhatextentitaffectstheoutcomesofaglobalcrisisfora corporation.Inthisstudy,theemissionsscandalofthehighlyreputedcorporationthe VolkswagenGroup(VW)wasinvestigated.Thefocusisonwhetherthecrisisresultedin reputationallossofandanincreasednegativeword-of-mouthintentiontowardstheVWGroup amongtheGermanpublic.TheSituationalCrisisCommunicationTheory(SCCT)isemployedto conceptualizethecrisiscontext.Inadditiontocrisisresponsibilityandanger,whichhavebeen identifiedaskeypredictorsofcrisisoutcomesintheSCCT,thisstudyalsoexaminestwo potentialpredictors:Thenewlyintroducedconceptcrisisinvolvement,andthepositivethusless regardedemotion-sympathy.Moreover,theimpactoftheperson-companyfitisinvestigatedin thiscrisiscontext. Usingaquantitativeonlinesurvey,thisstudyinvestigatesthemechanismthroughwhich theemissionsscandalinfluencedtheVolkswagenGroup’sreputationandthenegativeword-ofmouthintentionamongtheGermanpublic.Theseoutcomesarecomparedbetweentheaffected andnon-affectedGermanpublic.Thedataconsistsof1475Germanrespondentsintotalandthe dataanalysiswasconductedbyemployingthestructuralequationmodelingmethod.Theresults suggestthattheGermanrespondentsevaluatethepost-crisisreputationoftheVWGrouponly onamodestlevel,though,theirintentiontoexpressnegativeword-of-mouthisratherlow.Both emotions-angerandsympathy-amongtheGermanpublicmediatetheimpactofperceived crisisresponsibilityandcrisisinvolvementonpost-crisisreputationandnegativeword-ofmouthintention.However,theimpactofcrisisresponsibilityonangerandthatofangeronboth crisisoutcomesisstrongeramongtheaffectedGermanpublicthanthenon-affectedGerman public,whilenodifferenceisobservedwithrespecttosympathy.Nexttothis,person-company fitisidentifiedasmoderatorintherelationships.Itweakenstheimpactsofcrisisresponsibility andcrisisinvolvementonanger,aswellstrengthenstheimpactofcrisisresponsibilityon sympathy. I TheresultsofthisstudyimplyanextensionoftheSCCTframeworkthroughidentifying theroleofcrisisinvolvementandperson-companyfitinthecrisiscontext.Managerial implicationsareprovidedwithregardtocorporatecrisiscommunication.Corporationsshould considerthatnotonlyangermightinfluencethepost-crisisreputationandnegativeword-ofmouthintentionbutalsosympathy.Thisimpliesthatcorporationsshouldontheonehand mitigateangerandontheotherhandreinforcesympathyinordertosavethemselvesfrom negativecrisisoutcomes.Toachievethis,oneoptionsistocarryoutlowperceivedcrisis responsibilityandcrisisinvolvement.Further,asperson-companyfitmayvarythenegative crisisoutcomesforbothaffectedandnon-affectedgeneralpublic,itisimportantfor corporationstobuildastrongrelationshipwithcurrentandpotentialcustomers. Keywords:CrisisCommunication,CorporateReputation,NegativeWord-of-Mouth,Crisis Responsibility,CrisisInvolvement,Emotion,Anger,Sympathy,Person-CompanyFit,SCCT II Abbreviations AT - AttributionTheory DV - DependentVariable Engl. - English EV - ExpectancyViolation IV - IndependentVariable KMO - Kaiser-Meyer-Olkin MI - ModificationIndex/ModificationIndices NWOM - NegativeWordofMouth SCCT - SituationalCrisisCommunicationTheory SE - StandardError SEM - StructuralEquationModeling SIT - SocialIdentityTheory VW - Volkswagen VWGroup - VolkswagenGroup WOM - WordofMouth III TableofContents ABSTRACT.........................................................................................................................................................I ABBREVIATIONS...........................................................................................................................................III 1. INTRODUCTION......................................................................................................................................1 2. THEORETICALFRAMEWORKANDLITERATUREREVIEW.......................................................6 2.1. CORPORATEREPUTATIONINCRISISCOMMUNICATION.................................................................................6 2.1.1. TheSituationalCrisisCommunicationTheory..................................................................................9 2.2. POST-CRISISREPUTATION..................................................................................................................................11 2.3. NEGATIVEWORD-OF-MOUTH...........................................................................................................................13 2.4. IMPACTOFPERSONALPERSPECTIVESINCRISISCOMMUNICATION...........................................................15 2.4.1. Crisis-Emotions.............................................................................................................................................15 2.4.2. CrisisResponsibility....................................................................................................................................18 2.4.3. CrisisInvolvement.......................................................................................................................................20 2.4.4. Person-CompanyFit...................................................................................................................................23 2.5. CONCEPTUALMODEL..........................................................................................................................................26 3. METHODOLOGY...................................................................................................................................27 3.1. CHOICEOFMETHOD............................................................................................................................................27 3.2. SAMPLEANDSAMPLINGMETHOD....................................................................................................................28 3.3. DATAANALYSIS....................................................................................................................................................29 3.3.1. Pre-Test,DataCleaningandPreparationfortheDataAnalysis............................................30 3.4. RESPONDENTS......................................................................................................................................................32 3.5. OPERATIONALIZATIONANDMEASUREMENTS................................................................................................32 3.5.1. Post-crisisreputation.................................................................................................................................33 3.5.2. NegativeWordofMouthIntention(NWOM)..................................................................................35 3.5.3. CrisisResponsibility....................................................................................................................................36 3.5.4. CrisisInvolvement.......................................................................................................................................37 3.5.5. Mediators:AngerandSympathy..........................................................................................................38 3.5.6. Anger.................................................................................................................................................................38 3.5.7. Sympathy.........................................................................................................................................................39 3.5.8. Moderator:Person-CompanyFit..........................................................................................................40 4. RESULTS.................................................................................................................................................43 4.1. TESTINGOFMEDIATIONEFFECTS....................................................................................................................48 4.2. HYPOTHESES9THROUGH12–MODERATIONEFFECTS..............................................................................49 4.3. COMPARISONOFAFFECTEDANDNON-AFFECTEDPUBLICS........................................................................50 IV 4.4. FURTHERFINDINGS.............................................................................................................................................52 5. DISCUSSIONANDCONCLUSION......................................................................................................53 5.1. INTERPRETATIONOFRESULTS..........................................................................................................................53 5.2. MANAGERIALIMPLICATIONS.............................................................................................................................59 5.3. CONCLUSION.........................................................................................................................................................60 5.4. STRENGTHSANDLIMITATIONS.........................................................................................................................61 5.5. FUTURERESEARCH..............................................................................................................................................63 REFERENCES.................................................................................................................................................65 APPENDICES..................................................................................................................................................76 APPENDIXA-QUESTIONNAIRES...................................................................................................................................76 AppendixA1-QuestionnaireGerman.................................................................................................................76 AppendixA2-QuestionnaireEnglish..................................................................................................................81 APPENDIXB–OVERVIEWOFITEMS.............................................................................................................................86 APPENDIXC–FURTHERTABLES..................................................................................................................................88 V ListofFiguresandTables FIGURE1:CRISISSITUATIONMODELOFSCCT(BASEDONCOOMBS,2007A)...........................9 FIGURE2:CONCEPTUALMODEL............................................................................................................26 FIGURE3:MEASUREMENTMODELFORPOST-CRISISREPUTATION.........................................34 FIGURE4:MEASUREMENTMODELFORNWOMINTENTION........................................................35 FIGURE5:MEASUREMENTMODELFORCRISISRESPONSIBILITY..............................................36 FIGURE6:MEASUREMENTMODELFORCRISISINVOLVEMENT..................................................38 FIGURE7:MEASUREMENTMODELFORANGER................................................................................39 FIGURE8:MEASUREMENTMODELFORSYMPATHY.......................................................................40 FIGURE9:MEASUREMENTMODELFORPERSON-COMPANYFIT................................................41 FIGURE10:STRUCTURALMODEL…...............…...…..............................................................................44 TABLE1:SUMMARYOFRESULTS..........................................................................................................45 TABLE2:BOOTSTRAPPINGRESULTSFORMEDIATIONEFFECTS...............................................49 TABLE3:MULTIGROUPANALYSIS-AFFECTEDPUBLICVS.NON-AFFECTEDPUBLIC..........51 TABLEB1:CORRELATIONMATRIXPOST-CRISISREPUTATION..................................................88 TABLEB2:CORRELATIONMATRIXNWOM........................................................................................88 TABLEB3:CORRELATIONMATRIXANGER........................................................................................88 TABLEB4:CORRELATIONMATRIXSYMPATHY................................................................................88 TABLEB5:CORRELATIONMATRIXINVOLVEMENT........................................................................89 TABLEB6:CORRELATIONMATRIXCRISISRESPONSIBILITY.......................................................89 TABLEB7:CORRELATIONMATRIXPCFIT.........................................................................................89 TABLEB8:EXPLORATORYFACTORANALYSISPOST-CRISISREPUTATION............................90 TABLEB9:EXPLORATORYFACTORNWOM.......................................................................................90 TABLEB10:EXPLORATORYFACTORANALYSISCRISISRESPONSIBILITY...............................91 TABLEB11:EXPLORATORYFACTORANALYSISCRISISINVOLVEMENT...................................92 TABLEB12:EXPLORATORYFACTORANALYSISANGER................................................................92 TABLEB13:EXPLORATORYFACTORANALYSISSYMPATHY........................................................93 TABLEB14:EXPLORATORYFACTORANALYSISPERSON-COMPANYFIT.................................93 VI 1. Introduction Itisacasethatdominatedthenewsformonths–notonlyinGermanybutalsoalloverthe world:theVolkswagenemissionsscandal(VolkswagenfurtherstatedasVW).Germany,beinga countryinwhich“oneinsevenpeopleearntheirliving,directlyorindirectly,fromautomaking“ (Bender,2015,para.3)andwhereVWcarsarethe“mostfamousexport“(Bender,2015,para4), wasshocked.Whatfollowedthescandal’sdisclosurewasalargeproductrecallthatis comparabletothatoftheToyotarecallin2010-andfromthis,theToyotabrandhasnot completelyrevived,yet(Murphy,2015;Vizard,2015). TheVolkswagenGroup(furtherstatedasVWGroup)isaGermanautomobile manufacturer.Entailingtwelvebrands,suchasVolkswagen,Audi,Seat,Skoda,Porscheand Lamborghini,itis“thelargestcarmakerinEurope”(“TheGroup”,2014).OnSeptember18, 2015,thecorporationwasaccusedofintentionallymanipulatingVWandAudicarswith sophisticatedsoftwaretobypassCleanAirActstandards–thisincidentbecameknownasthe VWemissionsscandal.Thesoftwarewasinstalledinmillionsofdieselcarsandenabledthecars toproduceupto40timesmoreemissionthanpermitted(Geier,2015;Kollewe,2015; Woodyard,2015).Thisisespeciallysignificant,astheVWGroupsolditscarswiththepromise ofhavinglowemission(Vizard,2015).SincethentheVWemissionsscandalhasspreadtoother countriesincludingtheUnitedKingdom,GermanyandAustralia(Kollewe,2015),having affectedmorethan11millioncarsofthebrandsVW,Audi,Seat,SkodaandPorscheworldwide (Kollewe,2015),ofwhich2.4millionaloneinGermany(heiseonline,2016).Sincetheemissions scandalhasledtoseveralissuesforthecorporation,suchasfallingshares(Geier,2015)and decreasingsales(“VWglobalsalesfell“,2016),itcanberegardedasacrisisforthecarmaker. Productrecallsconstituteasevereandoften-facedproblemintheautomotiveindustry(Birchall &Milne,2009)andalthoughcommunicationinproductrecallcrisesissignificant,itisnotwell researcheduntilnow(Laufer&Jung,2010).Moreover,as“Volkswagenexecutivessetoutto deliberatelyandcriminallybreakthelaw”(Vizard,2015,para.7),theVWemissionsscandalis anespeciallyseverecrisis.Lastly,researchoutsideNorthAmericaisscarce(Lee,2004)anda greaterunderstandingofinternationalcrisesisneeded(Coombs,2014).Thus,analyzingtheVW emissionsscandalwouldaddvaluetothefieldofcrisiscommunicationresearchandwould providecorporationswithvaluableknowledgeabouthowcrisiscommunicationcanbeapplied inordertosaveanorganization’sreputationfromacrisis(Coombs,2007a). TheVWGrouphadbeenknownforitssolidityandreliability(Griffin,2015)andhadhad astrongreputationforyears(Fombrun,2015).Notonlywasthecorporationrankedamongthe 1 first15companiesintheGlobalRepTrak1001inthepastthreeyears(ReputationInstitute, 2013,2014,2015),italsohad,accordingtothereputationmonitoroftheeconomicresearch instituteDr.Doebler,thehighestreputationofallDAX30companiesin2015withintheGerman population(Reidel,2015).Areputationiswidelyacceptedasanintangibleassetforan organization(Coombs,2007a;Coombs&Holladay,2006)andafavorablereputationcan provideseveraladvantages,suchascredibilityamongcustomers,commitmentofemployees andabetterfinancialperformance(Fombrun&vanRiel,2004).Acrisis,suchastheVW emissionsscandal,canhoweverleadtonegativeoutcomesforanorganization(Coombs,2007a). Itcan,forexample,harmacorporatereputation(Coombs,2007a;Coombs&Holladay,2008) andcausepeopletousenegativeword-of-mouth(NWOM),meaningtomakenegative statementsabouttheorganization(Schultz,Utz&Göritz,2011).Itisthusofinterestofthisstudy whethertheVWemissionsscandalaffectedthecorporation’sreputationandwhetherithas causedpeopletoexpressNWOMabouttheVWGroup. Despitethelargeextentoftheemissionsscandal,themajorityoftheGermanpopulation remainedfaithfultotheVWGroupafterthescandalhadbecomepublic.Accordingtoasurvey thatwasconductedtwoweeksafterthebreakoutofthescandalbythemanagementconsultancy Prophet,twothirdsoftherespondentsstatedthattheystilltrustedVW.75percentsaidthat theywouldcontinuetobuyVWcars“iftheylikedthevehicleandtheprice“(Prophet,2015,as citedinLöhr,2015)and63percenthadtheopinionthatthescandalwouldbeforgottenwithina year(Prophet,2015,ascitedinLöhr,2015).TakingintothefactthattheVWGrouphadhada favorablereputationforyears,thisstudyaimstoexaminehowtheGermanpublicassesspostcrisisreputationofthecorporationandwhattheirevaluationsreplyon.Severalscholarshave addressedtheroleofafavorablepre-crisisreputationinpreviousstudies.Ontheonehand,they claimthatastrongreputationcanprotectanorganizationfromreputationallossduringacrisis (e.g.Coombs&Holladay,2006).Ontheotherhand,scholarsarguethatahighreputationleads tohigherexpectationsamongthepublic,whichareviolatedduringacrisisandresultinthe sufferingoftheorganization(e.g.Rhee&Haunschild,2006).Thus,theVWGroup’spost-crisis reputationisexploredinthisstudy. AccordingtoLee(2004),researchoncrisiscommunicationhasbeenconductedontwo stages.Firstly,responsestrategiesinspecificcriseshavebeenassessedandsecondly,the characteristicsofcrisesthatforecastthechoiceofsuitableresponsestrategieswereidentified. However,mostexistingresearchwasnotaudience-oriented(Lee,2004)andlittleresearchhas 1TheRepTrakPulseisameasurementforthepublicopinionofcompaniesdetectedbyanannually conductedglobalsurveybytheReputationInstitute.Therating„describeshowmuchconsumerstrust, likeandadmireacompany“(Fombrun,2015)andcanthereforebeevaluatedasausefulindicatorfora company’sreputation. 2 includedconsumervariables,suchasemotionsorinvolvement(Choi&Lin,2009a).Thisis crucial,though,inordertoassesshowindividualsbothunderstandandreacttoacrisis(Lee, 2004)andthuswhicheffectsthecrisishasfortheorganization.Hence,anincreasingnumberof authors(e.g.Coombs&Holladay,2002;Dawar&Pillutla,2000)havecalledforresearchonthe publicopinionofsuchevent(Choi&Lin,2009b).Moreover,severalscholarshavesuggested includingindividualvariablesincrisiscommunicationresearch(Coombs&Holladay,2014;Lee, 2004).Thepresentstudyrespondstothisperceivedbiasinresearch(Coombs,2014)and focusesontheroleofindividualvariablesthatarepossiblycausingdifferentcrisisreactions.By includingsuchindividualperspectives,knowledgeaboutindividuals’perceptionsofand reactionstothecrisiscanbegained.Basedonthisknowledge,managerialimplicationscanbe madeonhowcorporationscanreactproperlytoacrisis. AsCoombsandHolladay(2014)pointout,crisisexperthaveaninterestinthe knowledgeaboutcrisisreactionsofimportantpublics.Thereby,importantpublicsare customersbutnotonlycustomers(Coombs,2007a)andthusalsonon-stakeholdersandnonaffectedpeople.Oftheexistingstudiesincrisiscommunicationresearch,manyfocusedonthe reactionsofstakeholderstocrises,though(e.g.Coombs&Holladay,2014;Choi&Lin,2009b; Kiambi&Shafer,2015).BynotonlyincludingaffectedpeopleorstakeholdersoftheVWGroup, thisstudyprovidesinsightsofhowalsolessinvolvedpeopleperceiveandreacttoacrisis.This enablesthecomparisonofperceptionsandcrisisoutcomesbetweenaffectedandnon-affected publics,whichmakesthisstudyevenmoreworthwhile. Crisisresponsibilityhasbeenfoundtoplayakeyroleintheperceptionofacrisis (Coombs,2007a;Coombs&Holladay,2005;Kiambi&Shafer,2015).Responsibilityrelatesto whetherpeoplebelievethatthecompanycanbeblamedforthecrisis(Coombs,1995). Dependingontheperceivedresponsibility,peoplethenexperiencedifferentemotions.Anger andsympathyhavebeenidentifiedasmainemotionsinthecontextofacrisis(Coombs& Holladay,2005)andareassumedtocausebehavioralresponses(Coombs,2007a).Beinga product-recallcrisis,theVWemissionsscandalcanbecategorizedasapreventablecrisis (Coombs,2007a;Choi&Chung,2013)andisthusprobabletobeattributedwithahighlevelof responsibility.Ahighlevelofresponsibilityhasbeenfoundtoleadtomoreanger(Choi&Lin, 2009b)andlesssympathy(Coombs&Holladay,2005;Jin,2014).Emotionsinturncanhavean impactonpeoples’evaluationofanorganization’spost-crisisreputationandcaninfluence behavioralintentions,suchastheintentiontoexpressNWOM(Coombs,2007a). Nexttothis,theconceptofinvolvementhasbecomeofinterestincrisiscommunication. Theconceptreferstoaperson’ssubjectiverelevanceaboutatopic(McDonald&Härtel,2010) andahigherlevelofinvolvementinacrisiswasfoundtocausemoreangerandlesssympathy 3 (McDonald,Sparks&Glendon,2010).However,crisisinvolvementisnotmuchresearched,yet (Choi&Lin,2009a).Furthermore,companiestodayaimtohaveastrongrelationshipwithits customersbecauseitprovidesseveraladvantages,suchasloyalty(Bhattacharya&Sen,2003). Thecustomer-companyrelationshiphasbeenarguedbyBhattacharyaandSen(2003)byusing socialidentitytheory.Thetheoryreferstothefeelingofgroupaffiliationandisrelatedtovalues andemotions.Consumerswhoidentifystronglywithacompanyactinasupportivemanner (Chu&Li,2012).ThepresentstudytransfersthisconcepttotheGermanpublicasthepersoncompanyfit.ItisarguedthatGermanshaveanoverallhighidentificationwiththeVWGroupdue tothecorporation’shighreputationandtherelevanceofthecarmakerforGermans.Itis exploredwhethertheperson-companyfithasanimpactontherelationshipofinvolvementand responsibilitywithemotionsintheVWcrisis. Thepresentstudyexaminestheabove-mentionedtheoreticalconceptsbyusingthe SituationalCrisisCommunicationTheory(SCCT)(Coombs2007a)asgroundwork.SCCTisa usefultheoryforunderstandingcomponentsofacrisisinordertoapplypropercrisis communicationandthustoprotectanorganizationfromreputationaldamage(Coombs,2007a). However,notalloftherelevantconceptsandposedlinksbetweenconceptsareincludedin Coomb’sSCCT.Thus,theexistingframeworkisnotonlytestedonarealcrisisbutalsoextended byaddingtheconceptsofinvolvementandperson-companyfitaswellasthelinkbetween emotionsandreputationtotheframework. Concluding,thisthesisaimstoresearchtheoutcomesoftheVWemissionsscandaland therolethattheVWGroup’sreputationplayedinthecrisis.Itfurtherexaminestheimpactof individualperspectivesonsuchcrisisoutcomes.Thefollowingresearchquestionsarehence introduced: RQ1:HowdoGermansperceivetheVWGroup’sreputationandtowhatextentdotheyhave theintentiontoexpressnegativeword-of-mouthabouttheVWGroupaftertheoccurrence oftheemissionsscandal? RQ2:Towhatextentdocrisisinvolvement,crisisresponsibilityandemotions(i.e.anger andsympathy)influencetheVWGroup’spost-crisisreputationandpeople’snegativewordof-mouthintention?AndwhetherandhowdotheimpactsvarybetweenaffectedandnonaffectedGermanpublic? RQ3:Whetherandhowdoesperson-companyfitaffectthecrisisoutcomesinthiscontext? 4 Afterhavingintroducedthetopicofthisthesisandtheresearchquestions,thesecond chapterpresentsthetheoreticalframeworkunderlyingthisstudyaswellasresultsofprevious research.Accordingtothetheoreticalconstructsthatarerelevantforthisstudy,thechapterwill bestructuredintosub-sectionsthatwillconcludewithoneormorehypotheses.Followingthis, thethirdchapterwillgiveanoverviewoftheresearchdesignaswellastheoperationalizationof thetheoreticalconstructsandthemeasurementmodelsoftheperformedstructuralequation modeling.Inthefourthchapter,theresultsofthedataanalysisarepresented.Thesefindings willbefurtherdiscussedinchapterfiveofthisthesisbeforemanagerialimplicationswillbe given,limitationsprovidedandrecommendationsforfutureresearchgiven. 5 2. TheoreticalFrameworkandLiteratureReview Inthischapter,therelevanttheoreticalconceptswillbereviewedaswellaspriorresearch findingspresented.Attheendofeachsub-chapter,thetheoreticalknowledgeisappliedtothe VWemissionsscandal,whichwillthenleadtothehypothesesofthisstudy.Inorderto summarizeandvisualizethetheoreticalframework,aconceptualmodelwillbeprovidedatthe endofthischapter. 2.1. CorporateReputationinCrisisCommunication Businessscandalsofthepastyearshaveindicatedhowimportanttheestablishing,maintaining andprotectionofreputationisfororganizationsofallkinds(Doorley&Garcia,2007).Inorder tounderstandthecompletesignificanceofreputation,though,itiscrucialtofirstdefinethe concept.Untilnow,severaldifferentviewsonreputationcanberecognized(Love&Kraatz, 2009)andmanydifferentdefinitionsofreputationexist(vanRiel&Fombrun,2007).Forthis study,thedefinitionofCoombsandHolladay(2006)isused.Accordingtotheresearchers“[a] reputationisanevaluationstakeholdersmakeaboutanorganization”(Coombs&Holladay, 2006,p.123).Thereby,stakeholdersareindividuals,groupsororganizationsthathave”interest orconcerninanorganization”andcaninfluenceorbeinfluencedbyit(“Stakeholder”,n.d.). Someexamplesforstakeholdersarecustomers,employeesandshareholders(“Stakeholder”, n.d.). Areputationisdependentonanorganization’spastactions(Kiambi&Shafer,2015;van Riel&Fombrun,2007)andisgeneratedfromcognitiveassociations,whicharederivedfrom informationthatstakeholdersreceiveaboutanorganizationovertime(Fombrun&vanRiel, 2004;Rhee&Haunschild,2006;vanRiel&Fombrun,2007;Turketal.,2012).Thisinformation canbegainedthroughpersonalexperiencewiththecompany,secondhandinformationofother personssuchasfriendsorcolleaguesandthemassmedia(Bromley,2000;Coombs,2007a; Fombrun&vanRiel,2004;vanRiel&Fombrun,2007;Turketal.,2012).Thereby,direct personalexperiencehasthegreatestimpactonreputation,whereasmostoftheinformation stemsfromthemassmedia(Coombs,2007a;vanRiel&Fombrun,2007).Stakeholderscompare theirinformationaboutanorganizationinordertoevaluatewhetheritmeetstheirexpectations oftheorganization’sactions(Fombrun&vanRiel,2004).Incaseofanexpectationgap,meaning thattherespondents’expectationsarenotmetbytheorganization,issuesfortheorganization canemerge(Coombs,2007a;Reichart,2003). Asevaluations,reputationscanbefavorableandunfavorable(Coombs,2007a)orin otherwordspositiveornegative(Walker,2010).Favorablereputationsareacceptedas 6 intangibleassetsthathavebeenrelatedtopositiveoutcomesforanorganization(Coombs, 2007a;Coombs&Holladay,2006;Rhee&Haunschild,2006),orasFombrunandvanRiel(2004) putit:“Agoodreputationislikeamagnet:Itattractsustothosewhohaveit”(p.3).Such benefitscanbebothtangibleandintangible(Doorley&Garcia,2011).Intangiblepositive outcomesincludeforinstancetheperceptionofacompany’sproductsasmoreattractive,a highercredibilityamongcustomersandahighercommitmentofemployees(Fombrun&van Riel,2004).Thiscanpotentiallybetranslatedintotangiblepositiveoutcomes,suchasthe advantageofpayinglessforsuppliers(Davies,Chun,daSilva&Roper,2003;Doorley&Garcia, 2011)oranimprovedfinancialperformanceofanorganization(Fombrun&vanRiel,2004; Turketal.2012).Thisisforinstancebecauseagoodreputationcanaffectthereceivingof positivefeedbackfromfinancialanalystsandgainingmoreandalsomorefavorablemedia coverage(Daviesetal.,2003;Fombrun&vanRiel,2004).Companieswithabadreputationon theotherhandcanhavemoredifficultiesindrawingattentionofinvestorsandreceivingfunding (Aula&Mantere,2008,ascitedinKiambi&Shafer,2015)aswellascanreceivemorenegative mediacoverage(Daviesetal.,2003).CorporateReputationhasadditionallybeendemonstrated toinfluencestakeholder’sresponsestoproduct-harmcrises(Laufer&Coombs,2006),suchas emotionsandbehavioralintentions(Coombs,2014).Concluding,apositivereputationisof importanceforanorganizationandcanevenbeseenas„thesinglemostvaluedorganizational asset“(Gibsonetal.,2006,p.15). Basesontheseadvantages,FombrunandvanRiel(2004)arguethat"reputations[…] mustbenurturedandprotected"(p.7).Thisisespeciallythecaseintimesofcrisis.Accordingto Coombs(2007a),acrisiscanbedefinedas“asuddenandunexpectedeventthatthreatensto disruptanorganization’soperationsandposesbothafinancialandareputationalthreat.Crises canharmstakeholdersphysically,emotionallyand/orfinancially“(p.164).Acrisiscanoccur whenstakeholdersperceiveviolationsoftheirexpectationsofanorganization(Coombs,2014). Manydifferenttypesofstakeholders,includingemployees,suppliers,customersand stockholders,canbenegativelyaffectedbyacrisis(Coombs,2007a).Thus,crisescancreate victims,meaningpeoplewhoareactuallyharmedbyit,andpotentialvictims,referringtothose whocouldbeaffectedbytheincident.Inadditiontothis,acrisiscangeneratewitnesses, meaningpeoplewhogaininformationaboutandrespondtoacrisis(Bies,1987,ascitedin Coombs&Holladay,2007,p.300).Thereby,mostofthestakeholdersarenotdirectlyaffected (Coombs&Holladay,2005)butallofthesethreetypesofpersonsmightfeelsomesortof ”emotionalinvolvementinthecrisis“(Bies,1987,ascitedinCoombs&Holladay,2007,p.300) andareconnectedtotheorganizationduetotheincident(Coombs&Holladay,2007).Thus,the presentstudyisinterestedinthereactionsofnotonlyvictimsbutalsonon-victimsand thereforeconsidersthereactionsofthegeneralGermanpublic. 7 Asmentionedabove,crisesusuallycausenegativepublicity(Daviesetal.,2003;Dean, 2004).Asmoststakeholdersgaintheirinformationaboutacrisisfromnewsmediaandthe Internet,badpublicitycanresultinpeoplethinkingbadlyofanorganization,whichmaydamage theorganization’sreputation.This,inturn,maychangethewaystakeholdersperceiveand interactwiththeorganizationandmaythenleadtoalossoftheabove-mentionedbenefitsthata favorablereputationprovides(Coombs,2007a;Coombs,2014).Inaddition,acrisiscanresultin stakeholdersquestioninganorganization’scredibility(Arpan,2002),endingtheirrelationswith anorganizationand/orsayingnegativethingsaboutit(Coombs,2007a). Ascrisescanhavesuchnegativeeffectsforacompany,thesignificanceofcorrectly managingtheeventshouldnotbeunderestimated(Laufer&Coombs,2006)andtimely decisionsaboutthecrisisresponseshouldbemade(Doory&Garcia,2007;McDonaldetal., 2010).Whenacrisishasoccurred,organizationshavetocommunicatewithstakeholdersfor severalreasons.Firstofall,theyprovideinformationonhowtoshieldthemselvesfromthe crisis(instructinginformation)andinformationthatsupportsthemwithpsychologically managingtheincident(adjustinginformation).Onlythenextstepistoaddressthereputational threatbyusingseveralreputation-buildingstrategies(Coombs,2007a;Coombs&Holladay, 2005,2009).Thereby,anadequatecrisisresponsecanfunctionasabenefitandmighteven increasetheorganization’sreputation.Aninadequatecrisisresponseontheotherhandcan damageanorganization’sactions,itsreputationandthreatenitsexistence(Doorley&Garcia, 2007). Crisiscommunicationgenerallyrefersto„thecollection,processing,anddissemination ofinformationrequiredtoaddressacrisissituation“(Coombs,2010,p.20).Asespecially avoidablecrisescanendangeranorganization’sreputation,themaingoalofcrisis communicationistorebuildboththeorganization’sreputationandthestakeholder’strust(Utz, Schultz&Glocka,2013).Thus,crisismanagement,includingcrisiscommunication,and reputationmanagementarehighlyintertwined(Carroll,2009).Crisiscommunicationis furthermoreaprocessthatcanbedividedintothethreephasespre-crisiscommunication,crisis communicationandpost-crisiscommunication(Coombs,2010).Pre-crisiscommunication addressesthepreparationforpossiblethreats,crisiscommunicationregardsthedecisions duringthecrisisaswellastheprovidingofinformationtothepeopleandpost-crisis communicationincludestheanalysisofpreviouscommunicationandthepossibleprovisionof “follow-upcrisismessages“(Coombs,2010,p.21).Especiallyinthecontextofproductrecalls, effectivecommunicationisessentialbecauseitpotentiallyreducesdamage(Desai,2014).Asthe VWemissionsscandalhasalreadygonepublicoversixmonthsago,theVWGroupiscurrently situatedinthepost-crisiscommunicationphase.Ithastobekeptinmind,though,thatnew detailsaboutthecrisisarestillbeingrevealed. 8 2.1.1. TheSituationalCrisisCommunicationTheory Inordertobuildandtesttheory,onelineofresearchinthefieldofcrisiscommunicationhas begantodeveloptheSituationalCrisisCommunicationTheory(SCCT)(Coombs,2004,2007a; Coombs&Holladay,2002;Kim&Cameron,2011).TheSCCTmodel(seefigure1)hasbeen developedsince2002(Coombs&Holladay,2002)andprovidesastructureforthe comprehensionof“howcrisiscommunicationcanbeusedtoprotectreputationalassetsduring acrisis“(Coombs,2007a,p.163).Itcanbeappliedtodifferenttypesoforganizations(Coombs, 2007a)andtakesanaudience-centeredapproach,whichdeterminestheimpactofimportant aspectsofthecrisissituationonthestakeholders’perceivedreputation(Coombs,1998;Coombs, 2007a).Knowledgeabouthowstakeholdersreacttoacrisiscanthenagaininfluencea company’spost-crisiscommunication(Coombs,2007a).EmpiricalresearchbasedonSCCT offersdirectionsforcrisismanagersonhowresponsestrategiescanbeusedinordertosavea reputationfromacrisis(Coombs,2007a).Thismakestheframeworkespeciallyvaluable. Previousresearch(e.g.Choi&Chung,2013;Coombs&Holladay,1996;Dean,2004)has confirmedsuchbenefitoftheusageoftherightcrisisresponsestrategy(Coombs&Holladay, 2008). Figure1:CrisisSituationmodelofSCCT(basedonCoombs,2007a) 9 SCCToriginatedfromAttributionTheory(AT)(Weiner,1985)(Coombs,1995;2007a), whichisa“theoryofmotivationandemotion”(Weiner,1985,p.548)thatsuggeststhat individualsseekforunderlyingcausesofeventsthathappenaroundthem(Weiner,1985). Peopleconstantlyask‘why’becausetheyhavetheneedtocomprehendandcontroltheir environment(Weiner,1985).AccordingtoFolkes(1988),studiesaboutATinthecontextof consumerbehaviorfoundthatconsumersmakeattributionsforinstanceforwhytheyhave boughtaproductortofindthereasonforwhyaservicefailed.ThetwokeyaspectsofATare unexpectancyandnegativitybecausetheyinduceaperson’sneedtolookforthereasonofan incident(Weiner,1985,1986;Coombs,2007b).Sincecrisesareperceivedasunpredictable, negativeevents,stakeholdersassociateblametotheinvolvedactorsinacrisis(Coombs,2007a, 2015;Dean,2004;Laufer&Coombs,2006). Inordertoreactproperlytoacrisis,thepotentialcrisisthreatforanorganization’s reputationneedstobeassessed.AsCoombs(2007a)explains,“threatistheamountofdamagea crisiscouldinflictontheorganization’sreputationifnoactionistaken”(p.137).Thereby,the reputationalthreatisdeterminedbythethreedeterminantsinitialcrisisresponsibility,crisis historyandpriorreputation(Coombs,2007a).Usingatwo-stepprocess,crisismanagers estimatethethreattoareputation.First,theyevaluatetheinitialcrisisresponsibility,whichis basedonthetypeofthecrisis.AccordingtotheSCCT,crisescanbedividedintothreetypes,also knownasframes,whicheachpresentaspecificaspectoftheparticularcrisis:victimcrisis, accidentalcrisisandlastlythepreventableorintentionalcrisis(e.g.human-errorproductharm ororganizationalmisdeed).Eachcrisistypedefineshowmuchresponsibilitythestakeholders attributetotheorganization.Ofthethreetypes,theintentionalcrisishasthestrongest attributionofcrisisresponsibilityandposesaseverereputationalthreat(Coombs&Holladay, 2002;Coombs,2007a).AccordingtoClaeys,CaubergheandVyncke(2010),inthecaseof preventablecrises,evenallresponsibilityisassignedtotheorganization.Basedonthecrisistype,theSCCTthenproposesdifferentstrategiestorespondtothecrisis(Coombs,2007a).The dieselenginemanipulations,whichresultedintheVWemissionsscandal,wereperformedby employeesoftheVWGroup.Moreover,severalmanagersofthecorporationknewaboutthese illegalactions(Neate,2016,para.1,10).Aspreventablecrisesinvolve,forexample, managementmisbehavior,whichintentionallyendangersstakeholdersand/orbreakslaws (Coombs,2004;2007a),theemissionsscandalandrecallofVWcarsclearlyfallsintothe preventablecrisiscluster.Asthisisthemostseverecrisistype,itmakestheexaminationofits consequencesespeciallysignificant. Inthesecondstageoftheassessmentofthereputationalthreat,theintensifyingfactors crisishistory,referringtowhethertheorganizationhashadalikelyeventbefore,andpre-crisis reputation,meaningthereputationbeforethecrisis,areexamined(Coombs,2007a).The 10 existenceofacrisishistoryoranegativepre-crisisreputationwillreinforcethereputational threatofanorganization(Coombs,2007a).However,crisishistorywasfoundnottobeas importantasthepre-crisisreputationofacompany(Coombs&Holladay,2001).Thus,prior reputationisconsideredtobemoreimportantandwillthereforebeofinterestthisstudy. 2.2. Post-crisisreputation Asexplainedinthepreviouschapter,onenegativeoutcomefororganizationsinacrisisisthe lossofreputation(Coombs,2007a;Coombs&Holladay,2008;Dutta&Pullig,2011;Kiambi& Shafer,2015).Inthiscontext,itisassumedthatthepublic’sevaluationofacompany’s reputationbeforethecrisisinfluencesthecompany’srecovery(Turketal.,2012).Thereby, differentpointsofviewexist. Ontheonehand,itcansimplybesupposedthatafavorablepriorreputationoperatesas abankaccountthatconsistsofreputationalcapital(Alsop,2004,ascitedinCoombs&Holladay, 2006).Reputationalcapitalreferstoanorganization’s“stockofperceptualassetsandsocial assets”(Fombrun&vanRiel,2004,p.32).Ahigherreputationalcapitalcouldresultinmore supportiveactionsbystakeholders(Fombrun&vanRiel,2004).Asacrisiswillcausesomeharm toanorganization’sreputation,italsoleadstoalossofreputationalcapital(Coombs,2007a).If anorganizationhasafavorablepre-crisisreputation,thusmeaningthatithasplentyof reputationalcapital,itcanallowitselftospendorlosesomecapitalinacrisis.Inthatcase,it sustainsastrong,favorablereputationafterthecrisis(Alsop,2004,ascitedinCoombs& Holladay,2006;Coombs,2007a;Dowling,2001).Thisindicatesthatorganizationswitha favorablereputationcansufferasmuchasthosewithanunfavorablereputationbutwillstill maintainabetterreputationaftersuchanevent(Claeys&Cauberghe,2015).Thus,agoodprecrisisreputationindicatesthatanorganizationisharmedlessandrecoversmorerapidly (Coombs,2007a). Anothermorecomplexexplanationisthatafavorablepriorreputationcanfunctionasa shield(Claeys&Cauberghe,2015)orhalo(Coombs&Holladay,2006)thatprotectsan organizationfromthelossofreputation(Coombs&Holladay,2006;Ulmer,2001).Thetheories underlyingthisassumptionareexpectancyconfirmationtheoryandcognitivedissonance (Claeys&Cauberghe,2015;Coombs&Holladay,2006).Itisassumedthatindividualshave expectationsaboutsocialissuesandthattheytrytoavoidexperiencingcognitivedissonance. Thismeansthatindividualstrytoobtaininformationthatcorrespondswiththeirprioropinions ofanissue.Becausepeopletrytodiminishcognitivedissonance,conflictinginformationis interpretedinawaythatiscoherentwiththeindividuals’previousbeliefs(Claeys&Cauberghe, 2015;Coombs&Holladay,2006;Edwards&Smith,1996;Perloff,2010).Likewise,ifpeople 11 receivecrisisinformationaboutafavoredorganization,theymightfacecognitivedissonance (Perloff,2010,ascitedinClaeys&Cauberghe,2015,p.65).Consumerswithapositiveattitude maytendtopaymoreattentiontopositiveinformationandavoidnegativeaspectsaboutthe organization(Coombs&Holladay,2006;Claeys&Cauberghe,2015).Thisformofinformation processingcouldthenresultintheavoidanceofreputationalloss.Additionally,stakeholders couldtoacertainextentrejectthefactthatthecrisisoccurredandthusholdontotheir favorableopinionaboutit.Hence,thecrisiswillnothavealargeeffectontherelationship betweenstakeholdersandorganizations (Coombs&Holladay,2001,2006).Consequently,an organizationwithafavorablepre-crisisreputationwouldsufferlessreputationallossthanan organizationwithanunfavorablepre-crisisreputation.GrunwaldandHempelmann(2011)even assumethatthecognitivedissonancemechanismonlytakesplaceforwell-knownandwellreputedorganizations.AstheVWGroupdidhaveafavorablepre-crisisreputation,itisassumed thatthisprocessisapplicable. Severalscholarshaveinvestigatedtheroleofagoodpre-crisisreputationandare positiveaboutitsbenefitsforanorganizationduringacrisisbyhelpingtoprotectacompany’s reputationalassetsaswellasbeinganaidtoitsrepair(Coombs&Holladay,2006).Studies foundforinstancethatagoodpriorreputationinfluencedthepublic’sopinionandbehavioral intentionstowardstheorganization(e.g.Carroll,2009;Lyon&Cameron,2004),resultedina betterevaluationof(e.g.Kiambi&Shafer,2015)andledtoamorepositiveattitudetowardsthe organization(e.g.Turketal.,2012).KiambiandShafer(2015)thusevaluateagoodreputation ascriticalforanorganizationandemphasizethenecessitytoestablishafavorablepre-crisis reputation.However,sincemostexistingresearch“isopinionbasedratherthanwell researched”(Dowling,2001,p.252),thereisashortageofempiricalevidenceforthis assumption(Coombs&Holladay,2006).Thus,moreempiricalproofisneededinorderto provideadviceontheprotectionofanorganization’sreputationfromacrisis(Coombs,2007a; Rousseau,2006;Kiambi&Shafer,2015).Moreover,theshieldingfunctionofagoodreputation hasnotoftenbeenconfirmedinpreviousstudies(Claeys&Cauberghe,2015).Coombsand Holladay(2006)onlyfoundevidencefortheoccurrenceofahaloeffectcouldfororganizations withanextremelyfavorablepre-crisisreputation(Coombs&Holladay,2006).However,Claeys andCauberghe(2015)wereabletovalidatetheshieldingeffectandfoundevidenceforthe consumers’endeavortoavertcognitivedissonanceregardingtheirpre-crisisattitudetowards anorganization.Theconsumerswerenotonlytoassociatelowercrisis-responsibilitytothe organizationbutalsodisregardednegativepublicity.Thus,consumerswereaversetoaltertheir originalattitudetowardsanorganization(Claeys&Cauberghe,2015). Despitetheabove-describedrelevanceofagoodpre-crisisreputationfororganizations, someresearchers(e.g.Dean,2004;Grunwald&Hempelmann,2011;Rhee&Haunschild,2006) 12 arguethatafavorablereputationcanalsoleadtonegativeconsequencesforanorganizationina crisis,beingconcerned“aboutthepossibilitythatagoodreputationwillbackfireandinflicteven moreseveredamagetofirms”(Sohn&Lariscy,2015,p.238).Comparedtoaloworaneutral reputation,afavorablereputationcouldleadtohigherexpectationsofanorganizationamong stakeholders(Rhee&Haunshild,2006;Sohn&Lariscy,2015).Iftheseareviolatedinacrisis, well-reputedorganizationswillbepunishedmoresternly(Sohn&Lariscy,2015),forinstance, bycausingtheorganizationstopayhigherrestitutionsinordertoresolvetheincident (Grunwald&Hempelmann,2011).SohnandLariscy(2015)callthisthe‘boomerangeffect’ofa favorablereputationandjustifythisphenomenonwiththeexpectancyviolations(EV)theory. Contradictingtocognitivedissonance-basedviews(Sohn&Lariscy,2015),thistheorysuggests thatinsteadoffullyignoringordenyingcontradictinginformation,peopletendtocomparetheir pre-interactionexpectationswiththetargetobject’sbehaviors(Burgoon&LePoire,1993).The violationofaperson’sexpectancyratheroperatesasatriggerforcognitiveprocessing,hence affectingthetarget’spost-interactionjudgmentinasensethat“positiveandnegativeviolations (disconfirmation)leadtomorepositiveandnegativeinteractionoutcomesrespectivelythan doesconformitytoexpectations”(Burgoon&LePoire,1993,p.69).Initially,theEVtheorywas developedfortheclarificationofinternalcommunication(Sohn&Lariscy,2015).Aspeopleare likelytohumanizeorganizations(Daviesetal.,2003;Love&Kraatz,2009)andstakeholders considerthemasexchangepartnerswithcharacteristicssuchasreliability,whichtheyare evaluatedon(Fombrun,1996;Love&Kraatz,2009;Sohn&Lariscy,2015),thistheoryisalso suitabletotheinteractionofstakeholdersandorganizations(Sohn&Laricsy,2015). Althoughmostresearchershaveaddressedtheadvantagesofagoodreputationduringa crisis,otherscholarsargueforamorepessimisticview.Asabove-described,opponentempirical findingsexist,whereastheoptimisticviewisleading(Sohn&Lariscy,2015).Givensuchan ongoingdebate,thestudyathandexploresthepost-crisisreputationoftheVWGroupafterthe emissionsscandalhasoccurred. 2.3. NegativeWord-of-Mouth Besidesthepotentiallynegativeeffectonreputation,crisescanalsoinfluenceanindividual’s behaviorintentions,suchasNWOMintention(Coombs,2010;Coombs&Holladay,2008).Wordof-mouth(WOM)cangenerallybedefinedasinformal,non-commercialperson-to-person communicationamongcommunicatorsaboutbrands,products,servicesororganizations (Anderson,1998;Harrison-Walker,2001;Richins,1984;Goyette,Ricard,Bergeron& Marticottte,2010).Inthecontextofthisthesis,WOMrelatestostatementsthatstakeholders makeaboutacorporation(Schultzetal.,2011),namelytheVWGroup.WOMcantakeplaceface 13 toface,byphone,emailoranyothercommunicationchannels(Silverman,2001).WOMcan furtherbeofpersonalorigin,forinstancefromfriendsorfamily,andofimpersonalorigin,such asjournalists(Goyetteetal.,2010).WOMhaslongbeenacceptedasadominantpowerin buildingconsumers’opinionsandbehaviors(Brown&Reingen,1987;Herr,Kardes&Kim, 1991)andisevenreferredtoas„themostpowerfulforceinthemarketplace“(Silvermann, 2001,p.47).ThisismainlybecauseWOMisnormallygeneratedbycrediblesourcesandisthus believedtohaveastrongerimpactconsumers’judgmentsthaninformationobtainedthrough commercialprintsources,likeadvertising(Silvermann,2001;Herretal.1991),becausetheyare more“accessible”and“diagnostic”(Herretal.,1991,p.459).Inaddition,WOMcaneasilyreach alargeamountofpeopleduetoitsabilitytospreadquickly(Silvermann,2001),especially throughnewchannels,suchasonlineforums(Hennig-Thurau,Gwinner,Walsh&Gremler, 2004). WOMcanbepositive,neutralornegative(Anderson,1998).NWOM„denigratesthe objectofthecommunication“andrefersto„aconsumerresponsetodissatisfaction“(Richins, 1984,p.697).Asitisdamagingtoacompany’ssuccess(Richins,1984),NWOMcanbeseenasa threattoorganizations(Coombs,Fediuk&Holladay,2007).PositiveWOMontheotherhandis anadvantagefororganizations(Coombs,Fediuk&Holladay,2007).Moreover,itwasfoundthat NWOMhasamoreintensiveinfluenceoncustomerevaluationsthanpositiveWOM(Herretal., 1991;Laczniak,DeCarlo&Ramaswami,2001;Mizerski,1982).Moreprecisely,itsignificantly influencestheevaluationofbrands(Laczniaketal.,2001),products(Rea,Wang&Stoner,2014) andorganizations(Kiambi&Shafer,2015).Additionally,itisconsideredtobeanantecedentfor consumerbehavior,forinstance,itmaychangeaperson’spresentandfuturepurchasedecisions (Chu&Li,2012;Coombs&Holladay,2007;Schultzetal.,2011). NWOMisconsideredtobeparticularlypowerfulandproblematicbecauseitcanreach manyreceivers,includingpeoplewhodidnotknowabouttheincidentbefore.Itcanalsopersist (online)evenafteracrisisandevenafterpeoplehavealreadyforgottenabouttheincident. Therefore,itisathreattoorganizations(Coombs,2007a,2010,2014;Coombs&Holladay, 2007)andasmanagersaimtopreventthepossiblenegativeoutcomesofNWOM(Coombs, 2007b,2007a)theythustrytoavoidNWOM. ScholarshavearguedthatstakeholdershaveatendencytouseNWOMinacrisis(e.g. Coombs&Holladay,2007;Coombsetal.,2007;Kiambi&Shafer,2015).Thisrelationshiphas beeninvestigatedinseveralstudies.Utzetal.(2013),forinstance,foundintheirresearchabout theFukushimaDaiichinucleardisasterthatrespondentsusedmoresecondarycrisis communication,suchasNWOM,inthecaseofanintentionalcrisis.Furthermore,Kiambiand Shafer’s(2015)studyrevealedthatanorganizationwithabadpre-crisisreputationtendsto 14 undergomoreintentionsofNWOMthanthosewithagoodreputation.However,moreempirical researchonNWOMincrisiscommunicationisneeded(Kiambi&Shafer,2015). DuetopossiblenegativebehaviorsfollowingNWOM,theconceptisofhighimportance forcorporations.AstheintentiontouseNWOMcanbeevokedinacrisis,thepresentstudy examinestheoccurrenceofNWOMintheVWemissionsscandal.SincetheVWemissions scandalcanbecategorizedasanintentionalcrisis,theNWOMintentionoftheGermanpublicis expectedtoberatherhighafterthecrisis.Onthecontrary,theVWGrouphadhadavery favorablereputationpriortotheemissionsscandal,whichiswhytheNWOMintentioncouldbe potentiallylow,aswell.ThepresentstudythusfurtherinvestigatestheGermanpublic’s intentiontoexpressNWOMabouttheVWGroup. 2.4. ImpactofPersonalPerspectivesinCrisisCommunication Studiesinthefieldofcrisiscommunicationhavemainlyappliedtwotheoriesinordertoexplore publics’responsestoorganizations’usageofcrisiscommunicationstrategies(Coombs& Holladay,2014),namelycontingencytheory(e.g.Jin&Cameron,2007)andSCCT(e.g.Coombs, 2007;CoombsandHolladay,2007).Thereby,researchoftentimesfocusedonfictitious companiesandcrisissituations(e.g.Dean,2004;Claeys&Cauberghe,2015).Thismadeit difficulttoincludeindividuals’attitudestowardsacompanyintheanalysis(Dean,2004). Severalresearchers,however,suggesttheincorporationofindividualvariables,suchas involvementoremotionswhenanalyzingtheimpactofcrisesoncorporatereputation(e.g.Choi &Lin,2009a;Choi&Lin,2009b;Claeys&Cauberghe,2015;Dean,2004).AccordingtoChoiand Lin(2009b),notmuchisknown“abouthowapotentiallyaffectedpublicwillrespondtoacrisis andhowtheirresponsesshouldbeincorporatedintoSCCTwhentestingthatmodel“(p.199). However,thereissignificanceinthecomprehensionofstakeholderreactionstocrises(Härtel, McColl-Kennedy&McDonald,1998;Kim&Cameron,2011)becauseitcanbeusefulknowledge fortheorganization’spost-crisiscommunication(Coombs,2007a).Thus,theaimofthepresent studyistoexaminetheimpactofpersonalperspectivesontheabove-explainedpossible outcomesofacrisis.Throughthis,thestudywilldrawaconnectionbetweentheindividualand corporateleveloftheVWemissionsscandal. 2.4.1. Crisis-Emotions Asnotmuchresearchhasaddressedemotionsincrisissofar(McDonaldetal.,2010),scholars havebeguntoinvestigatetheroleofemotionalresponsesinthecontextofsucheventsinrecent years(e.g.Choi&Lin,2009a,2009b,2009c;Coombsetal.,2007;Jin,2009,2010;Jin,Pang& Cameron,2012,2014;Kim&Cameron,2011).Theemotionsfeltbystakeholderstowardsan 15 organizationcanhaveanimpactontheorganization’sdevelopmentandsurvival,makingthe comprehensionofaffectsespeciallyessential(Coombs&Holladay,2005).AsCoombsand Holladay(2005)furtherposit,crisiscommunicationcanmoreeffectivelyprotectan organization’sreputationwhenitconsidersthestakeholder’saffectivereactions(Coombs& Holladay,2005)becausecrisismanagerscanrespondmoreproperlytotheincident(Laufer& Coombs,2006). AsCoombs’SCCTframeworkusesanaudience-directedapproach,itcanbeappliedfor thecomprehensionofstakeholders’reactionsincrises(Jin,2010).CoombsandHolladay(2005) arguethatcriseswillnotonlytriggerattributionsbutalsocreateemotionalresponsesamong individuals(Coombs&Holladay,2005).InaccordancewithAT,angerandsympathyhavebeen statedtobethemainemotionsinthecontextofpost-crisiscommunication(Coombs&Holladay, 2005,2008).Althoughthesetwoemotionshavedifferentvalences,sympathyhasapositiveand angerhasanegativevalence,theyarebothprobabletobefeltbynon-victimpublics(Jin,2014). Transferredtothestudyathand,thiswouldsignifythatthegeneralGermanpubliccouldfeel emotionsabouttheVWGroupandtheemissionsscandal. Angertowardsaneventismainlyfeltwhenpeopleattributeresponsibilitytocertain agents,suchasorganizations,foraviolationorsorrow(Iyer&Oldmeadow,2006).Lindner (2006)arguesthatindividualsfeelangerwhentheyfeelhurtandwhenevertheyaretreated withdisrespect.Transferredtoanorganizationincrisis,itwouldmeanthatpersonswouldfeel angertowardstheorganizationwhentheybelievethattheorganizationistobeblamedforthe event(ascitedinJin,2010,p.527).Sympathyontheotherhandisevokedthroughthe “awarenessofothers’suffering[…]especiallywhenthesufferingisseenasundeserved“ (Salovey&Rosenhan,1989;ascitedbyIyer&Oldmeadow,2006,p.637).Thiscanbeexplained bythefactthatsympathyinvolvesasenseofcompassionthatiscausedbythedistressofthe otherperson(Gruen&Mendelsohn,1986).IyerandOldmeadow(2006)indicate,though,that notallpeoplewhoobserveotherssufferalsofeelsympathy.Theysuggestthatsympathyisnot onlyevokedbythefeelingforsomeoneelsebutthatitalsorequiressomedistancefromwhatis happening.Basedonthis,itisarguedthatmostprobablypeoplewhoperceiveacrisisbutare notdirectlyaffectedbyitwouldtendtofeelsympathy(Jin,2014).Incaseofanorganization experiencingacrisis,itcanbeassumedthatapersonwouldfeelsympathytowardsthe organizationwhenevaluatingtheorganizationasvictimthatissufferingduetothehappening. Inlinewiththis,severalstudiesfoundthatpersonsexperiencedifferentemotionsbasedonthe crisistype(e.g.Jin,2009;Jinetal.,2012). Beingonlyconfrontedwithstakeholderswhofeelangerorsympathytowardsan organizationwouldnotbeverysignificantfororganizationsincrisis.However,Coombsand 16 Holladay(2005)arguethatemotionswillhaveanimpactonastakeholder’sfuture organizationalinteractions,suchassupportingtheorganization(Coombs&Holladay,2005), andarethereforecrucialtoconsider.Furthermore,Liu,AustinandJin(2011)explainthat certainemotionshavebeenrelatedtoundesirablecrisisoutcomes.InthemostcurrentSCCT model,emotionisincorporatedasapredictorforbehavioralintentionsbutnotforreputation (seefigure1,arrowE)(Coombs,2007a;Choi&Lin,2009b).Jin,PangandCameron(2007) argue,though,thatemotionsinacrisiscanhaveanimpactonpeople’sopinionaboutan organization.ChoiandLin(2009b)thusproposedandtestedarevisedmodelthatcontained suchadirectpathfromemotionstoreputation.Theyfoundthatangersignificantlypredictedthe company’sperceivedreputation:Ahigherlevelofangerledtoaworsereputation.This highlightsthesignificanceofemotionsinSCCTandtheneedtotakeemotionalreactionsinto accountwhenaimingtoprotectanorganization’sreputation(Choi&Li,2009b).Basedontheir findings,ChoiandLi(2009b)suggestarevisedSCCTmodelthatcontainsadirectlinkfrom angertoreputation. Onthecontrary,onlyfewstudiesincrisisresearchhavecenteredontheimpactof positiveemotions,althoughthesignificanceofpositiveaffectsincommunicationhasbeen apparent(Jin,2014).Forinstance,FolkmanandMoskowitz(2000)claimthatpositiveemotions canemergetogetherwithdistressinastressfulsituation.Asfortherelationshipofsympathy andpost-crisisreputation,itcanbeassumedthatpersonswhofeelsympatheticforan organizationhaveapositiveattitudeandthusratetheorganization’sreputationbetterthan thosewhodonotfeelsympathyfortheorganization.Basedonthisbackground,thefollowing hypothesescanbeformulated: H1:AngerhasanegativeimpactontheVWGroup’spost-crisisreputation. H2:SympathyhasapositiveimpactontheVWGroup’spostcrisisreputation. AlthoughbeingconsideredasapredictorforbehavioralintentionsinSCCT(Choi&Lin, 2009b;Coombs,2007a),potentialeffectsofcrisisemotionsonbehaviorhavenotmuchbeen investigated,yet(Coombs&Holladay,2005).Suchresearchwouldhelp,though,“todevelop effectivecrisismanagementstrategies”(McDonaldetal.,2011,p.333).Onepossiblenegative behavioralintentionthatistriggeredbyemotionsisthatofNWOM(Coombs&Holladay,2007; Coombs,2007a,2014). AsLernerandTiedens(2006)explain,severalstudieshaveconfirmedthatangrypeople believethattheyhavetheabilitytomakeanimpactonormanageacertainsituationandtrigger abehavior.Inlinewiththis,angerhasbeenfoundtoleadtoNWOMintentionbecausepeople wanttoexpresstheirfeelingsorwanttoavenge(Wetzer,Zeelenberg&Pieters,2007).Coombs 17 etal.(2007)positthatunhappycustomershaveahigherproclivitytotellclosepersonsabout productsandservicesthanthosewhoarehappy.Theauthorsfurtherarguethatstakeholders whoareangrybecauseofacrisisarelesslikelytousepositiveWOMbutmorelikelytouse NWOMinstead.Theauthorscallthisthe“negativecommunicationdynamic”(Coombsetal., 2007).However,intheirstudyabouttherelationshipbetweencrisisresponsibility,angerand WOM,bothresponsibilityandangerwerefoundtoincreasepositiveWOMinsteadofNWOM. Coombsetal.(2007)concludethatthemoderatelevelofangertowardsthecrisisintheirstudy maynothavebeensufficienttodeveloptheirproposednegativecommunicationdynamic.They suggestexaminingtheimpactofangerinfuturestudies.Otherstudies,though,havefound evidencefortherelationshipbetweenangerandNWOM.Forinstance,Utzetal.’s(2013)study revealedthatangerhadanimpactonsecondarycrisiscommunication,suchasNWOM. Moreover,thehigherthelevelofnegativeemotions,theleastthepersonshadbehavioral intentionsthatweresupportiveforanorganizationandthemoretheytendedtouseNWOM (McDonaldetal.,2010).McDonaldetal.(2010)evenfoundthatangerwasoneofthestrongest predictorsfortheintentiontouseNWOM. Sympathy,ontheotherhand,couldnotplaysuchanimportantroleinacrisisbecauseits positiveaffectmightnotinfluencestakeholderstoalargeextent(Coombs&Holladay,2005). ResultsofastudybyStockmyer(1996)revealforinstancethatsympathytowardsanaffected companydidnotinfluencepeopletopurchasefromitafterthecrisis.CoobsandHolladay(2005) suppose,“customersaremorelikelytocomplainaboutabadexperiencewithaproductor servicethanreportapositiveexperience“(p.275).Moreover,thescholarsarguethatsympathy mightresultinsupportingactionsbystakeholders(Coombs&Holladay,2005).Furtherresearch isnecessary,though,inordertoassesstheactualbenefitsofsympathyforcrisiscommunication (Coombs&Holladay,2005).Thus,itcanbearguedthatsympathytowardsthecorporation wouldnotleadtoNWOMandahigherlevelofsympathywoulddecreasetheNWOMintention. Hence,thefollowingcanbehypothesized: H3:Angerincreasestheintentionfornegativeword-of-mouthabouttheVWGroup. H4:Sympathydecreasestheintentionfornegativeword-of-mouthabouttheVWGroup. 2.4.2. CrisisResponsibility Inordertoshieldanorganizationfromacrisis,understandingabouthowtheincidentcauses damagetotheorganizationisneeded(Coombs,2015).Previousresearchhasrecognizedfour situationalfactorsthatsupportthecomprehensionoftheharmfulimpactofacrisis,oneofthem crisisresponsibility(Coombs,2015).Sinceitisanessentialconceptforunderstanding “stakeholders’reactionstocrisisresponses“(Browm&Ki,2013,p.2)andtakesonapivotalpart 18 inSCCT(Coombs,2004;2007a;2015),itisespeciallysignificantforthepresentstudy. Crisisresponsibilityreferstotheamountofresponsibilityforacrisisthatstakeholders attributetoanorganization(Coombs,1998,2004),alsocalledtheblameoftheorganization (Coombs,1998).Asexplainedinchapter2.1.1,theconstructisderivatedfromAT(Coombs, 2015),inwhichcausalattributionsplayapivotalrole(Weiner,1985).Responsibilitycaneither beattributedtothepersonororganizationembroiledinthecrisisevent(internal)or circumstantial(external)factors(Coombs,2010).Theattributionofinternalorexternal responsibilityisessentialininducingaffectivereactionsorbehaviorstotheactorsthatarepart ofthecrisis(Weiner,1986).Inthecaseofahighdegreeofinternalresponsibility,behavioral reactionsarenegative.Ontheotherhand,iftheexternalresponsibilityisperceivedtobehigh, behavioralreactionsarepositive(Weiner,2006).Theprocessofblamingisbasedonthe knowledgethatapersonpossessesaboutwhetheranorganizationisresponsibleforacrisisand whetheritcouldhavecontrolledoravoidedthecrisis(Jin,2010). Asexplainedinchapter2.1.1,basedonthecrisistype,managersareabletodetect whetherstakeholderstendtoconsidertheorganizationasresponsibleforthecrisisornot (Coombs,2015).TheVWemissionsscandalfallsintothecategoryofapreventablecrisis,thus,a highattributionofcrisisresponsibilitytheVWGroupisexpected(Coombs,2007a).The emphasizingofcertainaspectsofacrisis,thereforeframingitinacertainway,caninfluencea person’sopinion(Coombs,2007a;Druckman,2001)andonhowstakeholdersevaluatethecrisis (Coombs&Holladay,2002).Forinstance,existingresearchinthefieldofcrisiscommunication hasfoundadirectlinkbetweencrisisresponsibilityandreputation(e.g.Coombs,2004,2007a, 2014,2015;Coombs&Holladay,1996;2002;Laczniaketal.,2001;Turketal.,2012)and behavioralreactions(Coombs,2007a),suchaspurchaseintention(Laufer&Coombs,2006)and NWOMintention(Coombs,2015). Furthermore,aspeopleascriberesponsibilityforanevent,theywillexperiencevarious commonemotions(Weiner,1985;Coombs,2007a).MainemotionsinATareangerand sympathy(Coombs,2007a).InlinewithAT,CoombsandHolladay(2005)positthat“[c]risis responsibilityshouldberelatedtotheaffectcreatedbyacrisis”(p.269).TheSCCTcontainsthe so-called‘CrisisResponsibility–AffectProposition’(seefigure1,arrowC),whichindicatesthata strongerassociationofcrisisresponsibilityhasapositiveimpactonthefeelingsofanger whereasalowerassociationofcrisisresponsibilityithasapositiveimpactonsympathy (Coombs,2007a;Coombs&Holladay,2005).Themoreresponsibilityastakeholderattributesto anorganization,thegreatertheriskofthecrisis(Coombs,2014). Previousstudieshavedemonstratedtheoccurrenceofcertainemotionsdueto stakeholders’attributionsofcrisisresponsibility(Utzetal.,2013).Severalstudiesfoundthat 19 personsexperiencedifferentemotionsbasedonthecrisistype(e.g.Jin,2009;Jinetal.,2012). Thereby,intentionalcrisescreatethestrongestangerduetothehighlevelofcrisis responsibility(e.g.Coombs&Holladay,2005).Inaccordancewiththis,ChoiandLin(2009b, 2009c)foundthatangerwasthemostandsympathytheleastoccurringemotionafteraproduct recallcrisis.ResultsofChoiandLin(2009b)furtherrevealedthatcrisisresponsibility significantlypredictedemotionssuchasanger,surpriseorfear,whereasangerhadthestrongest relationtoresponsibility.However,theycouldnotproveanassociationofresponsibilitywith sympathy.Thisresultcouldbeduetothesmallsamplesizeofsympathy(n=2)inthecontent analysis,though,whichmayhaveledtoinsignificantresults(Choi&Lin,2009b).Inlinewith this,CoombsandHolladay(2005)foundthatcrisisresponsibilitywasstronglypositively correlatedwithanger.Theauthorsalsofoundthough,thatcrisisresponsibilityisstrongly negativelycorrelatedwithsympathy.Basedonthis,McDonaldetal.(2010)proposethata strongercrisisresponsibilityprognosticatesstrongeremotions.Thus,whenapersonor organizationisjudgedresponsible,angerisexperiencedandbehavioralactionsarenegative.On theotherhand,whenapersonororganizationisnotjudgedasresponsible,sympathyisevoked andactionsarepositive(Kiambi&Shafer,2015;Weiner,1985;Weiner,2006,ascitedin Coombs,2007a).Basedontheseassumptionsandempiricalresults,thefollowingcanbe hypothethized: H5:AhigherlevelofperceivedcrisisresponsibilityleadstomoreangertowardstheVW Group. H6:AhigherlevelofperceivedcrisisresponsibilityleadstolesssympathytowardstheVW Group. 2.4.3. CrisisInvolvement Despitethecallforincludingtheconceptofstakeholderinvolvementintocrisiscommunication research(e.g.Dean,2004;Coombs&Holladay,2005;McDonald&Härtel,2000)onlyfewstudies haveinvestigatedthemeaningoftheconceptinthiscontext(Choi&Lin,2009a).Asaresultof thenoveltyofinvolvementincrisiscommunication,studiesthatdidincorporatetheconceptdid notconsideritinthesamemanner.Thismakesthecomparisonandgeneralizationofresults difficult.Forinstance,whilesomescholarsincludedproductinvolvement(e.g.Choi&Lin, 2009a;Choi&Chung,2013),otherappliedcrisisinvolvement(e.g.McDonaldetal.,2010).Since theemissionsscandalandnottheproductsoftheVWGroupareofinterestofthisstudy,the involvementwiththecrisisanditsconsequencesareexamined. Asmanyresearchersconsiderpersonalimportance(Petty&Cacioppo,1981)or relevanceasthecrucialaspectofinvolvement(e.g.Celsi&Olson,1988;Zaichkowsky,1985),the 20 conceptcangenerallybedefinedas“aperson’sperceivedrelevanceoftheobjectbasedon inherentneeds,valuesandinterests(Zaichkowsky,1985).Thus,thelevelofperceivedpersonal relevancedefineshowmuchaconsumerisinvolved“withanobject,situationoraction”(Celsi& Olson,1988,p.211).CelsiandOlson(1988)implythatsomethingisofpersonalrelevancewhen consumersregarditasself-relatedorsomehowcrucialforaccomplishingindividualaimsand values.Generally,theconceptofinvolvementisusedinthecontextofproductsorbrands(Peter &Olson,1990,ascitedinMcDonald&Härtel,2000,p.801;Zaichkowsky,1985),however, consumerscouldalsobeinvolvedwithvariousissues,suchasevents(Peter&Olson,1990,as citedinMcDonald&Härtel,2000,p.801;Petty&Cacioppo,1986). Inthecaseofanindividual’ssubjectivesenseofpersonalrelevance,itcanbereferredto “feltinvolvement“(Celsi&Olson,1988).Itisaperceptionthatistiedtoanobjectoreventand solelyexistsatparticulartimesandsituations,whichemphasizesthesituationalroleofthe concept(Celsi&Olson,1988).AccordingtoCelsiandOlson(1988),feltinvolvementhas motivationalcharacteristics,whichhaveanimpactoncognitiveprocesses,includingattention andcomprehension,andonbehavior,suchasconsumptionbehavior.AsPettyandCacioppo (1981,1986)argue,involvementhasaninfluenceonpeople’smotivationformessage processingandthusontheirattitudechange.Whilehighlyinvolvedconsumerprocess informationonthecentralroute,meaningtheypaymoreattentiontothequalityofthe arguments,low-involvedconsumersprocessitontheperipheralrouteandwillratherpay attentiontoaspectssuchasthesourcecredibilityofthemessage.Thus,thehigheraperson’s involvement,themoredifficultitistochangetheirattitude. Researchinthefieldofpublicrelations(Heath&Douglas,1990,1991,ascitedinChoi& Chung,2013)hashighlightedthecrucialroleofinvolvementregardingaudience’sreceptivityto informationandissues.Theconceptofinvolvementwasfirstappliedtothefieldofcrisis communicationandtoorganizationalcrisesbyMcDonaldandHärtel(2000)(Choi&Lin,2009a) becausetheyassumedthatpersonalrelevanceisimportantforthedeterminationofcrisis outcomes(McDonald&Härtel,2000).Accordingtotheauthors(2000),mostoftheexisting crisiscommunicationresearchappliesATinordertoproveconnectionsbetweenattributions andangeraswellaspurchaseintention.AlthoughATviewstheevaluationofindividual relevanceofaneventascritical,itdoesnotintegrateitintothemodel.Thus,thescholars suggestconsideringAffectiveEventsTheory(AET)(Weiss&Cropanzano,1996)inorderto examinetheroleofbothpersonalrelevanceaswellasemotionsincrisis.Moreprecisely,AET suggeststhatthelevelofpersonalrelevancedefinestheintensityoffeltemotions(McDonald& Härtel,2000). 21 Applyinginvolvementtocrisisevents,McDonaldandHärtel(2000)arguethatthe conceptdefineswhetherapersonisawareofamessageandhowmuchattentionheorshepays toit.Inthecaseofacompanycrisis,aperson’sintrinsicsourcesofpersonalrelevance,suchas itsgoals,valuesorneeds,butalsochangeablesituationalsourcesofpersonalrelevanceinthe consumer’senvironment,includingthemediacoverageofsuchevent,determinehisorher motivationtodevoteherselforhimselftoamessage.Theleveloffeltinvolvementwouldthen influencetheprocessesofattributionandemotion,whichwouldinturnhaveanimpacton behaviorintentionsafteracrisis(McDonald&Härtel,2000;Choi&Lin,2009a).Accordingto Weiner(1995),eventsthataremorepersonallyrelevantleadtostrongeremotions.Thus, McDonaldandHärtel(2000)arguethatinvolvementisanimportantfactorindefiningthe effectsofacompanycrisisbecauseitcaninfluencethelevelofpeople’sattentiontoamessage andthustheefficiencyofcrisiscommunication.Moreprecisely,thescholarsproposethata consumer’sinvolvementwithacompanycrisispredictsthelevelofemotions.Asfelt involvementinfluencesconsumer’sanger,McDonaldandHärtel(2000)proposethatthelevelof feltinvolvementdefinestheintensityofangertheyfeelinacompanycrisis.Furthermore,since crisisvariablesareconsiderstobe„dynamicandchangeable“(McDonald&Härtel,2000,p.801), feltinvolvementistemporary.Thus,asangerisassumedtoresultfromfeltinvolvement,itis alsomomentary.Thisindicatesthatangerwilldisappearastimepasses,exceptforin„high impactcrisesthatgenerateextremelyhighlevelsofanger“(Coombs&Holladay,2007,p.302). Basedonthis,CoombsandHolladay(2005)supposethatconsumerinvolvementcouldincrease theemotionsand/orcrisisresponsibilitythatwerecreatedinacrisis,whichcouldleadtoan extensionofSCCT. FollowingMcDonaldandHärtel’s(2000)introductionofinvolvementtothefieldofcrisis communication,severalresearchers(e.g.Choi&Chung,2013;Choi&Lin,2009a;McDonaldet al.,2010)haveappliedtheconceptofproductinvolvement,meaningaperson’sperceptionof relevanceregardingaproduct(Choi&Chung,2013),tocrisiscommunication.ChoiandChung (2013),forinstance,foundthatinvolvementhadasignificanteffectonreputation.Astudyon crisisinvolvement,conductedbyMcDonaldetal.(2010),revealedthatinvolvementand responsibilitycausedbothpositiveandnegativecrisisemotions.Thestrongesteffectswere foundforanger,fearandsympathy.Incomparisonwithresponsibility,though,crisis involvementwasonlyaweakpredictorforemotions(McDonaldetal.,2010).Inlinewiththis, ChoiandLin(2009a)foundintheirstudythathighandlowinvolvedconsumersperceivea crisisdifferentlyandfoundalinkbetweenproductinvolvementandangerfortheMattel productcrisis.TheauthorsconcludethattheinclusionofconsumerinvolvementintotheSCCTis a„logicalnextstepforfutureresearchincrisiscommunication“(p.21). 22 BasedonChoiandChung’s(2013)argumentationfortheToyotarecallcase,fortheVW emissionsscandal,itcanbeassumedthattherearespecificgroupsofpeople(e.g.current ownersofacaroftheVWGrouporownersofanaffectedcar)whoarelikelytoregardtheVW emissionsscandalaspersonallyrelevant.Thus,itcanbeexpectedthattheyhaveahigherlevel offeltinvolvementwiththeVWemissionsscandalthanthosewhoarenotownersofsuchcars. Basedontheaboveexplainedtheoreticalassumptionsandempiricalfindings,itcanfurtherbe assumedthatpeoplewithahigherlevelofcrisisinvolvementwillfeelmoreangerandsympathy towardstheVWGroupafterthecrisis.Consequently,thefollowinghypothesiscanbeproposed: H7:AhigherlevelofinvolvementwiththeVWemissionsscandalleadstoahigherlevelof angerafterthecrisis. H8:AhigherlevelofinvolvementwiththeVWemissionsscandalleadstoalowerlevelof sympathyafterthecrisis. 2.4.4. Person-CompanyFit AsBhattacharyaandSen(2003)state,anincreasingnumberofcompaniesseektoachieve significantandenduringrelationshipswiththeircustomersbecausetheymaybringseveral corporatebenefits.Suchbenefitsincludecustomerloyalty(Bhattacharya&Sen,2003; Lichtenstein,Drumwright&Braig,2004),emotionallyattachedcustomers(Lichtensteinetal., 2004),theusageofpositiveWOM(Bhattacharya&Sen,2003),betterevaluationsofand attitudestowardsthecompany(Einwilleretal.,2006;Sen&Bhattacharya,2001)aswellas commitmentwiththecompany(Kim,Lee,Lee&Kim,2010).Severalofthesestudieshave appliedtheconceptofconsumer-companyidentificationtotherelationshipbetweenconsumers andacompany(e.g.Lichtensteinetal.,2004;Bhattacharya&Sen,2003)becauseitisbeneficial forexplainingpeople’scausesandmotivesforassociatingwithcompanies(Pérez,2009;Du, Bhattacharya&Sen,2007;Marín&Ruiz,2007).Otherstudiesexaminedtherelationship betweenemployeesandtheircompany,(e.g.Berger,Cunningham&Drumwright,2006;Kimet al.,2010),customersandbrands(e.g.Underwood,Bond,&Baer,2001)oralumniandtheir formercollege(e.g.Mael&Ashforth,1992).ManyofthesestudiesappliedSocialIdentityTheory (SIT)toexplainsuchidentificationprocesses(e.g.Ashforth&Mael,1989;Bhattacharya&Sen, 2003;Dutton,Dukerich&Harquail,1994;Kimetal.,2010;Mael&Ashforth,1992;Pérez,2009) aswellastheconceptoforganizationalidentification(Ashforth&Mael,1989;Bhattacharya& Sen,2003). SITreferstoasocial-psychologicaltheorythatwasmainlyestablishedbyTajfeland Turner(e.g.Tajfel,1974;Tajfel&Tuner,1985)(Ashforth&Mael,1989).Thetheorypostulates thatpeoplearelikelytocategorizethemselvesintosocialgroups,suchasgenderorreligious 23 groups(Ashforth&Mael.1989;Tajfel&Turner,1985)becauseitenablesthem„tosituate themselvesintheirsocialenvironment“(Pérez,2009.p.179).Socialidentificationthereby refersto„theperceptionofonenesswithorbelongingnesstoagroup,involvingdirector vicariousexperienceofitssuccessesandfailures“(Ashforth&Mae,1989,p.34). AshforthandMael(1989)transferredtheconceptofsocialidentificationtoan organizationalcontextandarguethatorganizationalidentificationisaparticulartypeofsocial identification.Theorganizationtherebyfunctionsasasocialcategorythatmight“fulfill[…] motivesfortheindividual”(Ashforth&Mael,1989,p.22)andthattheindividualusestobuild upself-confidence(Ashforth&Mael,1989).Duetal.(2007)arguethattheconsumer-company identificationisapsychologicalattachmentwiththecompanythatdrivesbehaviors,whichare favorabletothecompany(Duetal.,2007;Pérez,2009).SITpostulatesthatindividualswho identifythemselveswithacompanyaremorelikelytojudgethecompanypositivelyinorderto increaseandenhancetheirself-concept(Tajfel&Turner,1979;ascitedinPéres,2009).A consumerthatidentifieshim-orherselfwithacompanyhasamentalconnectionwithit (Duttonetal.,1994;Bhattacharya&Sen,2003)andwillthenadjusthisactionstothecompany’s aimsandinterests(Mael&Ashforth,1992).Thus,AshforthandMael(1989)assumethata higherlevelofidentificationincreases“supportforandcommitmentto“thecompanyaswellas „loyaltyto,andpridein,the[company]anditsactivities“(p.26).Peoplewhohaveastrong identificationwithanorganizationbehaveinawaythatiscoherentwiththeorganization’s „values,beliefsandculture“(Xiao&Hwan(Mark)Lee,2014,p.1242).Moreover,theidentity withacompanyisrelatedtovaluesoremotionsthattheindividualexperiencesforbeingpartof agroup.Whentheindividualevaluateshisidentitytobeequaltothatofthegroup,hecaneasily connecttoit,indicatingthatemotionalidentificationtendstobecreatedorimproved.If consumersstronglyidentifythemselveswithacompany,theywillbecomesupportivetowards theorganization.Theywillnotonlyshowloyaltytothecompanybutwillalsoshowenthusiasm aboutcompanyactivities(Chu&Li,2012). AccordingtoDuttonetal.(1994),organizationalidentificationtakesplacebasedonthe individual’sperceptionoforganizationalattributesorperceivedidentity.Thereby,the individual’sidentificationisdependentonhowattractiveheorsheevaluatestheorganization (Duttonetal.,1994).Itisarguedthatcorporatereputationpositivelyinfluencestheemergence ofcustomeridentificationbecauseacompanymaybeperceivedasmoreattractive (Bhattacharya&Sen,2003;Duttonetal.,1994;Keh&Xie,2009).KehandXie(2009)givetwo reasonsforthis:Firstly,well-reputedcompanieshaveatendencyfor“superiorfinancial profitability,productsorservices,andfrequentmediacoverage,whichsubsequentlyenhance theirrelativeadvantageanddistinctiveidentityinthemarketplace,whichinturncontributeto theiridentityattractiveness“(Keh&Xie,2009).Secondly,agoodreputationstandsforhigh 24 prestige,meaningthattheorganizationisregardedinapositiveway(Bergami&Bagozzi,2000). Intheirstudy,KehandXie(2009)foundthatcorporatereputationpositivelyinfluenced customeridentification.Basedonpreviousdefinitions,organizationalidentificationisinthis contextreferredtoasthedegreetowhichapersonfeelsconnectedtotheorganizationand defineshim-/herselfwiththeorganization(Bhattacharya&Sen,2003;Ashforth&Mael,1989; Xiao&Hwan(Mark)Lee,2014). Socialidentification„withacollectivity“canevenemergeinthecasethatno interpersonalconnectionorinteractionexistsandcanstillhaveastrongeffectonemotionand behavior(Ashforth&Mael,1989,p.26).Basedonthis,itcanbearguedthatcustomersofcars oftheVWGroupbutalsonon-customerscanfeelcertainidentificationwiththecorporation. ThiswouldsignifythatGermansasawholecoulddevelopsomesortofidentificationwiththe corporation.AccordingtoAshforthandMael(1989),theidentificationwithagroupcaneven endurewhen“groupfailureislikely”(p.35).Transferredtoanorganization,afailurecouldbe forinstanceacrisis.Therefore,itcanbeexpectedthatevenaftertheemissionsscandal,the identificationwiththecorporationwouldremainhigh.However,Bergeretal.(2006)arguethat changedbehaviorsofmembersofanorganizationcanresult in changed beliefs and identification. FortheGermanpopulation,whichevaluatedthereputationoftheVWGroupespecially high,itcanbeassumedthattheidentificationwiththecorporationishigh.Asmentionedabove, theidentificationwithabrandoranorganizationhasresultedinpositiveandmoresupportive outcomes,includingpositiveemotionalresponses.Basedonthesefindings,itcanbeexpected thatahigheridentificationwiththeVWGroupresultsinsuchpositiveoutcomes,meaningthat theperson-companyfitinfluencesthewaypeoplefeelabouttheVWGroup.Duetothis supportivemanner,itcanbeexpectedthatperson-companyfitweakenstherelationshipsof involvementandresponsibilityandangeraswellasenhancestherelationshipsofinvolvement andresponsibilitywithsympathy.Thiswouldforinstanceindicatethatpersonswhoperceivea highresponsibilityforthecrisismaytendtoexpresslessangerwhentheirperson-companyfit ishigh,comparedtothosewhoidentifywiththecompanyless.Thus,thefollowinghypothesis canbestated: H9:Ahigherperson-companyfitweakenstherelationshipbetweenperceivedcrisis responsibilityandanger. H10:Ahigherperson-companyfitintensifiestherelationshipbetweenperceivedcrisis responsibilityandsympathy. H11:Ahigherperson-companyfitweakenstherelationshipbetweencrisisinvolvementand anger. 25 H12:Ahigherperson-companyfitintensifiestherelationshipbetweencrisisinvolvement andsympathy. 2.5. ConceptualModel Insummary,inthischapter,twelvehypotheseswereformulatedbasedontheoreticaland empiricalknowledge.Figuretwoillustratestheconceptualmodelofthepresentstudy,which providesavisualoverviewofallhypotheses.Thereby,thetwoemotionsangerandsympathy takeonamediatingroleintherelationshipbetweencrisisresponsibilityandcrisisinvolvement withpost-crisisreputationandNWOM.Moreover,themoderatingfunctionofperson-company fitintherelationshipsbetweencrisisresponsibilityandcrisisinvolvementwiththetwo emotionsisportrayed. Figure2:ConceptualModel 26 3. Methodology Thethirdchapterstatesandexplainsthemethodologythatwasappliedinordertotestthe above-explainedconceptualmodel.Itthusprovidestherationaleforthedecisionsmade regardingtheresearchdesign,datacollection,samplingmethodaswellasthe operationalizationofthetheoreticalconceptsthatarerelevantforthisstudy. 3.1. ChoiceofMethod Inordertoanswertheresearchquestionunderlyingthisstudy,thedecisionforaquantitative methodologywastaken.AsBabbie(2011)pointsout,quantitativeresearchmethodsoftenusea deductiveapproach,implyingthattheresearcherformulatesseveralexpectationsthatarebased ontheory.Thesearethentestedinordertofindoutwhethertheyactuallyoccur(Babbie,2011; Zhou&Sloan,2009).Besidestestingsuchhypotheses,quantitativeresearchmethodsenablethe investigationofrelationsbetweenvariables(Punch,2014).Additionally,suchresearchmethods permitnotonlygeneralizationsbutalsothepredictionofcertaineffects(Zhou&Sloan,2009). Astheaimofthepresentstudyistotestrelationshipsbetweentheoreticalconcepts,a quantitativemethodologywasthebestoptionforthisstudy. Moreprecisely,aquantitativeonlinesurveywasconducted.Oneaimofquantitative surveysistomakestatementsaboutspecificpersonsthatarerelevanttoaresearch(Brosius, Haas&Koschel,2012).Furthermore,asurveyisoftenusedtoaskforthebehavior,attitudes, expectationsandcharacteristicsofpeopleanditallowstheself-classificationoftheparticipants (Neuman,2014).Asexplainedabove,thisstudyisinterestedinpersonalperspectives,whichare alllatentvariables.AccordingtoTheo,TingTsai&Yang(2013),latentvariablesare unobservableand“cannotbemeasureddirectly”(p.4).Thus,theymustbedefinedby indicators,whichareforinstancemeasuredbyself-reportedresponsesonanattitudescale (Byrne,2013).Hence,asurveyisanappropriateresearchmethodforthisstudy. Anonlinesurveyisaspecialtypeofasurvey,whichhasseveraladvantagesthatare relevantforthisstudy.Firstofall,themethodisinexpensive,fastandlocation-independent. Furthermore,thedropoutrateandtheresponsetimearecaptured.Besides,thismethod providesanonymityandissuitableforsensibletopics,suchaspersonality(Möhring&Schlütz, 2010).Anonlinesurveyalsoenablestherandomizationofitems,whichcanpreventthe occurrenceofsequenceeffects(Scholl,2009).Lastly,77.6percentoftheGermanpopulationare internetusers(Statista,2016),makingthismethodsuitableforreachingalargepartofGermans. However,besidestheseadvantages,themethodalsohasseveraldisadvantagesthatneedtobe considered.Themostimportantdisadvantageistheself-selectionoftherespondents,which 27 oftentimesleadstoalowresponserate(Möhring&Schlütz,2010).Therespondentsdecideon theirownwhethertheywanttoparticipateinthesurveyornot.Thismakesthedrawingofa randomandrepresentativesamplebarelypossible(Punch,2014;Scholl,2009).Nevertheless, duetothemajorityofadvantages,theapplicationofthismethodissuitable. Lastly,thechoiceforthisspecificmethodologycanbesupportedbythedemandfor quantitativesurveymethodsincrisiscommunicationresearch.Overthepastyears,researchin thefieldofcrisiscommunicationhasincreasedrapidly(Kim&Cameron,2011)butexisting researchhasmostlyusedcasestudymethods(Coombs&Holladay,2008;Kiambi&Shafer, 2015).Althoughcasestudiesprovidevaluabledescriptivedata(Coombs&Holladay,2008),they oftentimesofferlittletheoreticalunderstandingofcrisiscommunication(Dean,2004)andare hardtogeneralize(Carroll,2009).AccordingtoRousseau(2006),achangetoevidence-based managementincrisiscommunicationisnecessary,though,inordertobasemanagerialdecisions onscientificevidence.Overthelastdecade,researchinthefieldhasstartedtoapply experimentaldesignmethods(Kiambi&Shafer,2015),meetingthecallforquantitative research(Dawar&Pillutla,2000;Dean,2004).However,mostexperimentsfocusedonfictitious organizationsand/orcrises(e.g.Claeys&Cauberghe,2015;Dean,2004;Kiambi&Shafer,2015). Thisleadsmayhaveleadtotheproblemofartificialitybycreatingafavorablereputation throughonlyoneexposure(Lyon&Cameron,2004)andmayhavehadanimpactontheresults ofsuchstudies(Kiambi&Shafer,2015).Intheirreviewofcrisiscommunicationresearch, JohnsonAvery,WeaverLariscy,KimandHocke(2010)implytousesurveymethodsinorderto improvethemethodologicaldiversityinthefield.Thepresentstudythuscontributestoexisting researchbyapplyingaquantitativesurveymethodthatexaminesarealcrisisofanon-fictive corporation.Byusingaquantitativesurveymethodandincludingindividualperspectives,it furtherdrawsaconnectionbetweentheindividualandthecorporatelevelinacrisis. 3.2. SampleandSamplingMethod Asaforementioned,thisresearch’stargetpopulationcomprisestheGermanpublic.Asthis populationisverylarge,itisexpensiveanddifficult“tocollectinformationfromeveryoneinthe group”(deVaus,1996,p.60).Therefore,asampleofthepopulationwasobtained. Asoneofthesamplingmethod,thepurposivesamplingwaschosen.Purposivesampling isaformofnon-probabilitysamplinginwhichtheresearcherdecides,basedonselection criteria,whichpersonissuitableforthesample(Babbie,2011;Walliman,2006).Althoughthis samplingmethoddoesnotensurerepresentativeness,usefulinformationcanstillbesupplied (deVaus,1996).Additionally,sinceaspecifictargetaudiencehasbeenidentifiedforthisstudy (i.e.theGermanpublic),thissamplingmethodseemsplausible.Besides,thesnowballsampling 28 methodwasapplied.Snowballsamplingreferstoasamplingtechniquewherefurther participantsarereachedthroughotherrespondents(Babbie,2011)byforwardingittotheir relativesandfriends(Zhou&Sloan,2011).Althoughthissamplingmethodhasseveral disadvantages,suchasnotenablingtogainarepresentativesampleandthusmaking generalizationsimpossible(Brosiusetal.,2012),itissuitable.Thisismainlyforthereasonthat itenablestoreachalargenumberofparticipantswithinashortamountoftime(Möhring& Schlütz,2010).Lastly,theconveniencesamplingmethodwasapplied.Thissamplingmethod referstousingthosepersonsasparticipantswhoareinstantlyaccessible(Walliman,2006). Thus,itenablesthereachingofalargeamountofpeopleandisthusappropriateforthisstudy. Duetotheonlinenatureofthesurvey,theparticipantswererecruitedonline.Firstofall, emailsweresenttosecretariesofGermanuniversitiesaskingtoforwardthesurveytotheir studentsthroughinternalmailinglists.Additionally,thelinkwaspostedinseveraldifferent GermanFacebookgroups,suchas“DuitseenNederland–DeutscheinHolland”(engl.“Germans intheNetherlands”),“VWAbgasSkandalMotorEA189”(engl.:“VWemissionsscandalengine EA189”)or“IchhaltezuVolkswagen,egalwaspassiert”(engl.“IwillstandbyVolkswagen,no matterwhathappens”).Lastly,thelinktothesurveywassharedontheresearcher’sown FacebookwallandsenttoherownpersonalnetworkviaE-Mailandmessages.Allpotential participantsreceivedthelinktogetherwithashorttextaskingfortheparticipationinthestudy andarequesttoforwardthelinktothesurveytotheirpeergroup.Thisapproachallowedto reachapopulationthatisbeyondtheresearcher’sownpersonalnetwork. 3.3. DataAnalysis InordertotestthehypothesesinChapter4,structuralequationmodeling(SEM)isapplied.For thispurpose,thetwosoftwareIBMSPSSStatistics22andtheSPSSrelatedsoftwareAmosare used.SEMreferstoacollectionofstatisticalanalysismethodologiesthatareusedtotest hypothesesaboutthedirectandindirectrelationsbetweenvariables(Byrne,2013;Hoyle,1995, ascitedinTheoetal.,2013;In’nami&Koizumi,2013;Raykov&Marcoulides,2000).The hypothesizedmodelisbasedontheoryandistestedinonesimultaneousanalysiscontainingall proposedvariables.Dependingonthefitofthemodel,theaprioripostulatedrelationships amongthevariablescanbesupportedorrejected(Byrne,2013;Raykov&Marcoulides,2000). ThevariablesintheSEMcanbebothobservedvariables,alsocalledindicators,aswellaslatent variables(Theoetal.,2013).Incomparisontoothermultivariatetechniques,SEMhasseveral advantages.First,duetoitsconfirmatoryapproach,itenablesthetestingofhypotheses.Second, SEMtakesintoaccountspecificestimatesoferrorvarianceparametersandisthusmore accuratethanothermethods.Third,SEMconsidersbothobservedandlatentvariables.It 29 investigatestherelationshipsbetweenthetwokindsofvariablesbutalsotherelationsamong differentlatentvariables(Byrne,2013;In’nami&Koizumi,2013;Nachtigall,Kroehne,Funke& Steyer,2003).Infact,SEMwasevendevelopedtoanalyzetherelationshipsbetweenlatent variables(Nachtigalletal.,2003).Lastly,noeasilyapplicablealternativemethodsareexistent thatposemultivariaterelationsorestimateindirecteffectsthroughamediatingvariable(Byrne, 2013;In’nami&Koizumi,2013;Nachtigalletal.,2003). Basedonthesecharacteristics,itcanbearguedthatSEMisasuitablesetof methodologiesforthisthesis.Firstofall,thisstudyaimstoanalyzetherelationshipsbetween latentvariables,suchasinvolvement,angerorreputation,whichweremeasuredbyseveral indicators(seechapter3.5).Inaddition,theconceptualmodelandthehypothesesthatare underlyingthisresearchandwillbetested(seechapter2)arederivedfromtheory.Finally,SEM enablestheestimationoftheindirecteffectsofthemediatingvariablesangerandsympathyin theproposedconceptualmodel. 3.3.1. Pre-Test,DataCleaningandPreparationfortheDataAnalysis TheonlinesurveyquestionnaireforthisresearchwasconductedusingQualtrics.Beforethe actualsurveywasinitiated,thequalityandfunctionalityoftheresearchinstrumentwastested inapretest(Möhring&Schlütz,2010;Zhou&Sloan,2009).Forthepurposeofthistest,three personsreceivedthelinktothequestionnaireandthreepersonsreceiveditasadigitaltext documentoraprintedquestionnaire.Throughthis,valuablefeedbackonthecomprehensionof thequestionsandsuggestionsforimprovementweregained.Mostimportantly,questionswere evenfurtheradjustedtotheVWGroupandtheemissionsscandalandsomeitemswere eliminatedfromthequestionnaireoftheactualsurvey. Thedatacollectiontookplacebetween3and19April2016.Inthistimeperiod,atotalof 2072peopleparticipatedinthesurvey.Oftheseparticipants,1510finishedthequestionnaire, makingtheresponserate72.91percentandthusquitehigh.However,thissampleincluded30 respondentswhowerenotofGermannationality(e.g.Turkish,RussianorItalian)andthushad tobeexcluded,astheydidnotfitintothetargetpopulation.Additionally,respondentsthatwere identifiedtohavegivenwronganswers,suchasanageof“00”or“0”wereexcludedfromthe sample.Lastly,onepersonwhoneededlessthantwominutestocompletethewholesurveywas excludedbecauseitcanbeassumedthathe/shedidnotanswerthequestionsinanelaborate way.Thisleadstoafinalsampleof1475cases,whichwereincludedintheanalysis. Furthermore,aspartofthedatacleaningprocess,someitemsthatwerereverse-coded hadtoberecoded.Throughthis,aneasierinterpretationofthedataanditsusageforthe analysiswasenabled.Anexampleitemfromthereputationscale,whereahigheragreementof 30 anitemreferstoahigherreputation,is“TheVWGroupisbasicallydishonestconcerningthe emissionsscandal”.Finally,openanswersforeducationandnationalitywereexaminedandif necessaryaddedtotherightvariables.Thiswasforinstancethecaseforpersonswhoopenly answeredthattheyhadobtainedahighschooldiploma. InordertoapplySEMthedataneedstocomplywithseveralassumptions(Kline,1998). Firstofall,thedatasetshouldnotcontainanymissingdata(Kline,1998).Asonlyrespondents whofinishedthewholequestionnaireweremaintainedinthedataset,thiswasnotthecaseand thefirstpremisemet.Second,nomulticollinearity,meaningnostrongintercorrelationsbetween variables,shouldoccur(Kline,1998).Forthispurpose,correlationsbetweenall37variablesof thisstudywerecalculated.Mostofthecorrelationsaresignificantbutweakormoderate. However,alsosomesignificantstrongcorrelationscouldbedetected,especiallyamong variablesoftheconceptsperson-companyfitandinvolvement.Allofthesevaluesarebelow.8, though,andthecorrelationsareonlywithinconceptsandnotbetweendifferentconcepts.Thus, allvariablesweremaintainedinthedatasetbutthestrongcorrelationswerekeptinmindfor thedataanalysis(forallcorrelationmatrices,seeAppendixC).Accordingtothethirdpremise, thedatashouldnotcontainanymultivariateoutliers.Thisassumptionisalsofulfilled,asoutliers havealreadybeendeletedinthedatacleaningprocess.Next,thedatasetwasexaminedinterms oflinearityandhomoscedasticitybutnodeviationsweredetectedandthispremiseisalso accepted.Last,thedatawasinvestigatedregardingitsnormality.Despitethefactthat “maximumlikelihoodmethodsarerobustagainstnon-normality,itisstillimportanttoassess whetherthedatasatisfytheassumptionofnormality”(In’Nami&Koizumi,2013,p.34).Forthis purpose,theskwenessandkurtosisofthedatacanbetested(In’Nami&Koizumi,2013).Ifthese valuesarezero,“datanormalityisensured”(In’Nami&Koizumi,2013,p.34).Inliterature,there isnoconsensusaboutwhichlevelofnon-normalityisacceptable(In’Nami&Koizumi,2013), however,West,FinchandCurran(1995)suggestthatthevaluesforskwenessandkurtosis shouldnotexceed2and7.Forthepresentstudy,theselimitvaluesareused.Thetestfor skewnessandkurtosisrevealedthatnovariablesshowedanydeviationsfromthedesired valuesexceptfortwovariablesoftheconceptperson-companyfit.Thereby,thetwovariables “WhensomeonecriticizestheVWGroup,itfeelslikeapersonalinsult”and“Thesuccessesofthe VWGrouparemysuccesses”showedabnormalitiesregardingtheirskewnessvalues (skewnessinsult=2.138;skewnesssuccesses=2.016).Thus,thesetwovariablesareexcludedfromthe datasetinordertoassumethepremiseofnormalityforallconcepts. 31 3.4. Respondents Ofthe1475respondentsthatwereincludedinthedataanalysis,44.9%arefemale(N=662)and 55.1%male(N=813).Further,therespondentsarebetween17and70yearsold.However,the meanageis25.48(SD=7.09)andthemedianisonly24.Regardingtheeducationallevelofthe participants,itcanbenoticedthattheyareratherhighlyeducated.Only.3%(N=5)donothave anykindofhighschooldiploma,.5%(N=7)statedtohavegainedthelowesthighschool diploma(“Hauptschulabschluss”)andfurther4.1%(N=60)haveobtainedthemiddleleveled highschooldiploma(“Realschulabschluss”).Mostoftherespondents,however,claimedtohave thehighesthighschooldiploma(“Abitur”/”(Fach-)Hochschulreife”)(60.3%,N=889)and further34.1%havealreadyobtainedauniversitydegree(N=503).Inadditiontothis,threeof therespondentshavealreadyfinishedtheirPhD.Lastly,sevenrespondents(.5%)madeanopen answerfortheireducation,whichwasastatementabouttheircurrentoccupationandmadeit impossibletoclearlyidentifytheirhighesteducationlevel. Besidesthisdemographicinformation,thequestionnaireaskedwhethertherespondents ownedatleastonecarofthebrandsVW,Audi,Seat,SkodaorPorsche,whichwereaffectedby theemissionsscandal.50.8%(N=749)oftheparticipantsindicatedtocurrentlyownacarof suchbrands.Ofthese749respondents,themajoritydoesnotownanaffectedcar(73.7%, N=552)andonly21.8%statedtoownanaffectedcar(N=163).Further4.5%(N=34)claimedto notknowwhethertheircarwasaffectedbytheemissionscandalornot. 3.5. OperationalizationandMeasurements Inthissub-chapter,theoperationalizationofthetheoreticalconceptswillbepresented. Additionally,descriptivestatisticsofthetheoreticalconceptswillbeprovidedaswellas measurementmodelsdeployedandtested.Inthecontextofasurvey,operationalizationrefers tothetransformationoftheoverallresearchquestionintospecifictest-questionsthatcanbe askedintheactualquestionnaire.Forthis,themeaningandthecontentoftherelevant constructshavetobeexplored(Möhring&Schlütz,2010).Forthepurposeofthisstudy,already existingandpreviouslytestedmeasurementswereusedandadjustedtothetopicathand, meaningthattheywereadaptedtotheVWGroupandtheemissionsscandal.Sincethesurvey wasdirectedattheGermanpublic,thequestionnairewasgeneratedinGerman.Thus,measures thatwereoriginallyconstructedinEnglishlanguageweretranslatedintoGerman.Forthis process,existingGermantranslationswereusedaswellastheback-translationmethodapplied. Inaddition,whenformulatingthequestionnaire,theusageoftheterm“crisis”wasavoided,asit isnegativelyconnotated(Doorley&Garcia,2007).Instead,itwasreferredtothe“VWemissions scandal”.ForthefullquestionnairesinEnglishandGerman,seeAppendixAandforthe 32 overviewofallitems,seeAppendixB. 3.5.1. Post-crisisreputation Thedependentvariablepost-crisisreputationwasmeasuredusingthefive-itemversionofthe ten-itemOrganizationalReputationScale(Coombs&Holladay,1996,2002;Coombs,2004).This scalewasoriginallyadaptedfromMcCroskey’s(1966)Charactersubscaleformeasuringethos (Coombs&Holladay,2002).Theshorterfive-itemscalehadahighCronbach’salphavalueof α=.87(Coombs&Holladay,2002)andofα=.81(Coombs,2004)inpreviousstudiesandwasthus appropriatetoapply.Inthepresentstudy,thescalewasmeasuredusingaseven-pointLikert scalewithverbalizedendpoints(1=stronglydisagree,7=stronglyagree). Beforethemeasurementmodelofthepost-crisisreputationwasdeployedandtested,an exploratoryfactoranalysis(EFA)wasconductedinordertoexplore,whethertheconceptonly consistsoftheproposedonedimension.AstheKaiser-Meyer-Olkin(KMO)measurehadavalue of.706andtheBartlett’stestforSphericitywassignificant,itwasappropriatetoconductthe factoranalysis.ToallEFA’sdeployedinthisstudy,themaximumlikelihoodfactoringwitha Varimaxrotationwasapplied.ThisisfirstlybecausetheAmossoftwareusestheMaximum Likelihoodestimationmethod(Theoetal.,2013)andsecondlybecausethisparticularmethodis especiallysuitableforlargesamples(Weiber&Mühlhaus,2014),asitisthecaseinthisstudy. AccordingtotheKaisercriterionoftheEFAofpost-crisisreputation,theextractionofonlyone factorisproposed,confirmingthetheoreticalassumptionofonesinglereputationdimension. Furthermore,allfiveitemshaveasufficientfactorloading(seeAppendixC). Next,themeasurementmodelofthepost-crisisreputationistestedusingthe ConfirmatoryFactorAnalysis(CFA)inAmos.Having5degreesoffreedom,themodelisoveridentifiedandcanthusbetested.However,thesignificantChi-Squaretestindicatestherejection ofthemodel(χ2(5)=681.385;p=0.000).Ithastobenoted,though,thattheChi-Squaretest becomeseasilysignificantforlargesamples(Theo,TingTsai&Yang,2013).Therefore,theChiSquarevalueshouldbeseeninrelationtothedegreesoffreedom,whichresultsinameasure thatshouldnotexceedavalueof2,5(Weiber&Mühlhaus,2014).Inthepresentcase,thevalue isexceededbyfar(χ2/df=136,277)andalsoothermodelfitindicesdonotofferanacceptable modelfit2(GFI=.837;NFI=.738;CFI=.739;RMSEA=.303).Thus,amodificationofthe measurementmodelwasimplied.Inordertogainabettermodelfit,co-variancesbetweenthe errortermsoftheitemswithinthisconstructwereallowedandthemodelwiththebestfit 2FortheRMSEA,valuesof≤.08signifyanacceptableandvaluesof≤.05signifyagoodmodelfit(Browne &Cudek,1993).FortheGFIandCFI,valuesof≥.90signifyanacceptableandvaluesof≥.95agoodmodel fit(Weiber&Mühlhaus,2014).ThesameisthecaseforNFI(Bentler,1992). 33 indiceswasthenselected.Forthismodelmodification,theproposedmodificationindices(MI) wereregardedandtheco-varianceswiththehighestmodificationindicesconsidered.Basedon this,aco-variancebetweentheerrortermsoftheitemsreputation1andreputation5 (MI=422.22)wasaddedtothemodel.Astheoverallmodelfitwasstillnotsatisfactoryafterthis modification,anotherco-variancewasaddedbetweentheerrortermsoftheitemsreputation1 andreputation4(MI=68.82),resultinginanalmostsufficientmodelfit(seefigure3).Although themodelfitcouldnotbeconsideredasgood,thismeasurementmodelwasmaintaineddueto tworeasons:First,theseco-variancesweretheoreticallyplausible:reputation1andreputation4 bothrefertotherespondent’sbeliefoftheVWGroup’sgoodintentionsregardingtheemissions scandalandreputation1andreputation5bothrelatetothesamestatementthatisoncephrased positively(reputation1)andoncephrasednegatively(reputation5).Moreover,addingothercovariancesdidnotleadtoabettermodelfit.3 Samplesize=1475;standardizedestimates χ2(3)=13.386;p=0.004;χ2/df=4.462;GFI=.996;NFI=.995;CFI=.996;RMSEA=.048 Figure3:MeasurementModelforPost-CrisisReputation Forthedescriptionofthepost-crisisreputation,anindexofthefivereputationitems wasformed(N=1475),representingthemeanpost-crisisreputationoftheVWGroup.According tothat,theaveragereputationoftheVWGroupaftertheemissionsscandalisM=3.7(SD=1.25; Min=3.22;Max=4.4)onaseven-pointLikertscale.Thus,thepost-crisisreputationliesslightly abovethescale’smiddleof3.5,indicatingamoderatepost-crisisreputation.Forthepresent study,theinternalconsistencyfortheorganizationalreputationscalewasα=.795,whichimplies analmostgoodreliability. 3Theprocessofallowingtheerrortermstocorrelatewithinamodelandselectingthemodelwiththe bestfitwasalsoappliedtotheothermeasurementmodels(i.e.foreachconstruct)thathadfourormore items.However,itcouldnotbeappliedfortheconstructofonlythreeitems. 34 3.5.2. NegativeWordofMouthIntention(NWOM) InordertomeasuretheintentionforNWOM,threeitemsfrompreviousstudieswereapplied (e.g.Coombs&Holladay,2008,2009;Kiambi&Shafer,2015).Inpriorstudies,thescalehada reliabilitycoefficientofCronbach’sα=.76(Coombs&Holliday,2008)andα=.71(Coombs& Holliday,2008;Kiambi&Shafer,2015).Asampleitemofthismeasurementis“Becauseofthe emissionsscandal,IwouldsaynegativethingsabouttheVWGroupanditscarstootherpeople”. Aspost-crisisreputation,thisconceptwasalsomeasuredusinga7-pointLikertscale (1=stronglydisagree,7=stronglyagree). Asforthepreviousconcept,anEFAusingmaximumlikelihoodfactoringwithaVarimax rotationwasalsoappliedtoNWOM.HavingaKMOvalueof0.667andasignificantBartlett’stest forSphericity,theperformanceofanEFAwasappropriate.Astheoreticallyassumed,the extractionofonlyonefactorfortheconceptisproposedandallthreeitemshaveasufficient factorloading(seeAppendixC).However,aCFOcouldnotbeperformedbecausethemodelhad nodegreeoffreedomandwasthusunder-identified. Inordertosolvethisissue,aconstraintwasimposed.Imposingaconstraintreferstothe procedureoffreelyestimatingparametersthatwere“fixed-to-zero”before(Kline,1998,p.132). AsthefirstandtheseconditemwithwhichtheNWOMintentionwasmeasuredarerelatively similar,itcanbearguedthatthepathestimatesofbothconstructscanbesetto1.Asa consequence,anadditionaldegreeoffreedomwasgainedandthusthemeasurementmodel couldjustbeidentified.Althoughthefitofthismeasurementmodelwasnotideal,astheChiSquareandRMSEAvaluesweretoohigh,thismodelwasmaintainedbecausenobettermodelfit couldbeachieved(seefigure4). Samplesize=1475;standardizedestimates χ2(1)=19.026;p=0.000;χ2/df=19.026;GFI=.992;NFI=.986;CFI=.987;RMSEA=.111 Figure4:MeasurementModelforNWOMintention ForthedescriptionoftheNWOMintention,anindexofthethreemeasureditemswas formed(N=1475).Accordingtothat,theaverageNWOMintentionoftherespondentsisquite 35 low(M=2.27;SD=1.43;Min=2.26;Max=3.29).TheCronbach’salphacoefficientofthisstudywas α=.780,indicatinganacceptable,almostgood,reliability. 3.5.3. CrisisResponsibility ToassesswhethertheVWGroupwasheldresponsibleforthecrisis,thenewlyinventedscaleby BrownandKi(2013)wasusedandmeasuredona7-pointLikertscale(1=stronglydisagree, 7=stronglyagree).The12-itemsscaleconsistsofthethreedimensionsintentionality,locality andaccountabilityandhadareliabilityofα=.95(Brown&Ki,2013).Asaresultofthepretestof thepresentstudy,thescaleforcrisisresponsibilitywasreducedtoaneight-itemscale,makingit shorterandlessrepetitive(seeAppendixA). Inordertoexplorewhethertheconceptconsistsofoneoremoredimensions,anEFA wasperformed.AstheKMOmeasurehadavalueof.704andtheBartlett’stestforSphericity wassignificant,theEFAwasappropriatetoconduct.AccordingtotheKaisercriterion,the extractionofthreefactorsissuggested,confirmingBrownandKi’s(2013)theoretical assumptionofthethreedimensions“intentionality”,“accountability”and“locality”.Asthe seconddimension(firstfactor)accountsformostoftheexplainedvariance(25.13%)and consistsoffouroftheeightvariableswithacceptablefactorloadings,onlythisdimensionwas usedforthefurtheranalysis(seeAppendixC). Asanextstep,themeasurementmodelofthefourremainingitemsofthecrisis responsibilitymeasurewastestedinaCFAinAmos.With2degreesoffreedom,themodelwas justover-identifiedandcouldthusbetested.AsthelowChi-Squarevalueindicated,themodel hadagoodfit(χ2(2)=2.685;p=0.261),whichwassupportedbyotherfitindices(GFI=.999; NFI=.998;CFI=.999;RMSEA=.015).Hence,nomodelmodificationwasappliedandtheoriginal modeloffouritemsmaintained(seefigure5). Samplesize=1475;standardizedestimates χ2(2)=2.685;p=0.261;χ2/df=1.342;GFI=.999;NFI=.998;CFI=.999;RMSEA=.015 Figure5:Measurementmodelforcrisisresponsibility 36 Theindexofthefourcrisisresponsibilityitems(N=1475),formedinordertodescribe theconceptcrisisresponsibility,showsthattheparticipantsaveragelyassignedaquitehigh crisisresponsibilitytotheVWGroup(M=5.91;SD=1.19;Min=5.63;Max=6.18).Theinternal consistencyofthemeasurementwasα=.686,thusratherlowbutjustacceptable. 3.5.4. CrisisInvolvement Thecrisisinvolvementoftheparticipantswasmeasuredbyusingthesix-item,7-pointbipolar scalebyWigleyandPfau(2010).ThescaleisbasedontheinvolvementscalebyZaichkowski (1985)andhadareliabilityofα=.95inapreviousstudy(Wigley&Pfau,2010).However,after thepretest,thisscalewasreducedbytheitempairsignificant/insignificantbecausethe distinctionfromotheritempairswasclearenough.Furthermore,therespondentswereasked whethertheyhavebeenowningacarofthebrandsVW,Audi,Seat,SkodaorPorscheand,ifso, whetherithadbeenaffectedbytheproductrecall,assumingthatthoseparticipantswhowere affectedwouldperceivetheemissionsscandalasmorerelevant. TheKMOmeasurehadavalueof.810andtheBartlett’stestforSphericitywas significant,indicatingthatanEFAwasappropriatetoperform.TheEFAoftheinvolvement conceptproposedaonefactorsolution,whichwasinlinewiththetheoreticalassumptionsof onesingleinvolvementdimension.Allfiveitemshadasatisfactoryfactorloading(seeAppendix C).Furthermore,themodelwasover-identifiedandcouldthusbetested(df=5).Aswiththe conceptofpost-crisisreputation,thismeasurementmodelalsohadaveryhigh,significantChiSquarevalue(χ2(5)=490.858;p=0.000),implyingtherejectionofthemodel.Inlinewiththis, othermodelfitindicesdidnotsuggestanacceptablemodelfit(GFI=.886;NFI=.882;CFI=.883; RMSEA=.257),either.Sincethethirditemofthescale(meansnothing/meansalot)washighly correlatedwiththefourthandfifthitems,itwasexcludedfromthemodel.Moreover,a constraintwasimposedforthefirstandfifthitembecausetheirpathestimateswerethesame (β=.88)andtheyweretheoreticallyverysimilar(unimportant/important,irrelevant/relevant). Theimpliedmodificationofthemeasurementmodelthusledtoanadjustedmodeloffouritems thathadagoodfit(seefigure6). Aswiththepreviousvariables,ameanindexfromthefourremainingmeasureditems wascreated(N=1475)inordertopresentthedescriptiveanalysisoftheconcept.Accordingto theindex,theaveragecrisisinvolvementoftheparticipantswasM=4.27(SD=1.57;Min=3.41; Max=4.57),whichindicatesthattherespondentswereinvolvedwiththeVWemissionsscandal abovethescale’saverage.Theinternalconsistencyofthefour-itemcrisisinvolvementscalewas α=.852,indicatingalmostanexcellentreliabilityofthescale. 37 Samplesize=1475;standardizedestimates χ2(3)=8.383;p=0.039;χ2/df=2.794;GFI=.997;NFI=.997;CFI=.998;RMSEA=.035 Figure6:MeasurementModelforCrisisInvolvement 3.5.5. Mediators:AngerandSympathy Theemotionsangerandsympathyfunctionasmediatorvariablesinthepresentstudy.A mediatorvariableisathirdvariablethat“reflectsanintermediatelinkbetweenanindependent anddependentvariable”(Pawar,2009,p.110).Appliedtothisstudy,thismeansthatthetwo independentvariablescrisisinvolvementandcrisisresponsibilityaffectangerandsympathy whichinturninfluencethedependentvariablespost-crisisreputationandtheintentionfor NWOM.AsrecommendedbyZhao,Lynch&Chen(2010),thebootstrappingmethodwasapplied inordertotestmediationeffects.Bootstrappingisamethodthatisusedfor“estimating propertiesofestimatorsbasedonsamplesdrawnfromtheoriginalobservations”(Bollen,1990, p.117).Itisausefulapproachforthisstudybecauseitenablestheestimationofdirect,indirect andtotaleffectsandprovidesthe95%confidenceintervalsofeacheffect(Bollen,1990). Themediatorvariablesangerandsympathywereassessedusingtwofour-itemscales fromMcDonaldetal.(2011).Theauthorscriticized“theabsenceofscalesusingwordsthat incorporateconsumers’owncrisisemotionlexiconandwhicharepsychometricallyrobust“ (McDonaldetal.,2011,p.337).Thus,theydevelopedandtestedscalesspecificallyforcrisis emotions(McDonaldet.al.,2011).Theangerscalecontainedtheitemsangry,disgusted, annoyed,outragedandhadareliabilityofα=.91(McDonaldetal.,2011).Thesympathyscale positedbyMcDonaldetal.(2011)consistedoftheitemssympathetic,sorry,compassion, empathyandhadaninternalconsistencyofα=.83.Inthisstudy,bothconceptsweremeasured usinga7-pointLikertscalewhereas1means“notatall”and7means“verymuch”. 3.5.6. Anger TheKMOmeasurehadavalueof0.780andtheBartlett’stestforSphericitywassignificant, indicatingthattheperformanceofanEFAwassuitable.Inlinewiththetheoreticalassumptions, 38 theKaiser’scriterionsuggestedtheextractionofonlyonefactorandallfactorloadingswere sufficient(seeAppendixC).Havingtwodegreesoffreedom,thismeasurementmodelwasjust over-identified,whichiswhyaCFAcouldbeperformed.However,thehighChi-Squarevalue (χ2(2)=87.598;p=0.000)andthehighRMSEAvalue(RMSEA=0.170)indicatedtherejectionof themodel.Onthecontrary,othermodelindicessuggestedanacceptablemodelfit(GFI=.974; NFI=.968;CFI=.969).However,byapplyingtheabove-expainedadjustmentmethods,abetter measurementmodelcouldbefound.Addingaconstrainttothepathestimatesofitem1(angry) anditem2(annoyed)aswellasaddingaco-variancebetweentheerrortermsofitem2 (annoyed)anditem3(disgusted)(MI=45.47)werenotonlytheoreticallyreasonablebutalso resultedinamodelwithanacceptablefit(seefigure7).Sinceotherco-varianceswithlower modificationindicesdidnotincreasethemodelfit,onlythisco-variancewasadded. Samplesize=1475;standardizedestimates χ2(2)=20.272;p=0.000;χ2/df=10.136;GFI=.993;NFI=.993;CFI=.993;RMSEA=0.079 Figure7:Measurementmodelforanger Asforthepreviousconcepts,ameanindexthatisrepresentingtheaveragelevelofpostcrisisangertowardstheVWGroupwascreated.Theindexofthefouritemsshowsthatthe participants’angerafterthecrisiswasbelowaverage(M=3.32;SD=1.59;Min=2.56;Max=3.97). Thereliabilityscoreforthisstudywasα=.855,implyingagoodinternalconsistencyofthe measurement. 3.5.7. Sympathy Inordertofindoutwhetherthefouritemsthatmeasuredsympathyrepresentonedimension, anEFAshouldbeperformed.AccordingtotheKMOmeasureof0.720andthesignificant Bartlett’stestforSphericity,suchananalysiswasappropriate.Confirmingthetheoretical assumptionsofsuchuni-dimensionality,theEFAofferedonlyonefactorforsympathywith sufficientfactorloadingsforeachitem(seeAppendixC).TherelationbetweentheChi-Square valueandthedegreesoffreedomintheCFAwasnotacceptable(χ2(2)=179.480;p=0.000)and 39 alsotheRMSEAvalueindicatesapoorfitofthemeasurementmodel(RMSEA=.245).Other modelfitindices,ontheotherhand,impliedanacceptablemodelfit(GFI=.942;NFI=.918; CFI=.919).Nevertheless,asthemodificationindicessuggested,anadjustmentofthe measurementmodelshouldbemade.Theproposedco-variancebetweentheerrortermsof item1anditem4hadthehighestmodificationindex(MI=159,73),wastheoreticallyplausible andresultedinamodelwithagoodfit.Thus,thismeasurementmodelwasselected(seefigure 8). Samplesize=1475;standardizedestimates χ2(1)=1.481;p=0.224;χ2/df=1.481;GFI=.999;NFI=.999;CFI=1.0;RMSEA=0.018 Figure8:Measurementmodelforsympathy Forthedescriptiveanalysisofsympathy,anindexwascreatedthatconstitutesthe averagelevelofsympathytowardstheVWGroupafterthecrisis.Astheindexindicates, participantsdidonaveragenotfeelverysympatheticabouttheVWGroupaftertheemissions scandal(M=2.14;SD=1.22;Min=1.84;Max=2.59).Thereliabilityscorefortheindexisα=.799, indicatingagoodinternalconsistencyofthefour-itemmeasurement. 3.5.8. Moderator:Person-CompanyFit Thefinalconceptthatisofinterestforthisstudy,person-companyfit,functionsasamoderator variable.Amoderatorvariablecanbedefinedasavariablethatinfluences“thedirectionand/or strengthoftherelationbetween”theindependentanddependentvariable(Baron&Kenny, 1986,p.1174).Inthecaseofthisstudy,thismeansthatthedegreeofperson-companyfitis assumedtoinfluencetherelationshipsoftheindependentvariablescrisisinvolvementand crisisresponsibilitywiththemediatorvariablesangerandsympathy.Itisexpectedthatthe person-companyfitstrengthenstherelationshipofcrisisresponsibilityandinvolvementwith sympathybutweakenstherelationshipofthetwoindependentvariableswithanger. Themoderatorvariableperson-companyfitwasmeasuredusingeightitems.These 40 itemswerepreviouslydevelopedbyLin,Chen,ChiuandLee(2011)basedonscalesbyKehand Xie(2009)andMaelandAshforth(1992)aswellasitemsofMaelandAshforth(1992) themselves.Thescalewasmeasuredusinga7-pointLikertscale(1=stronglydisagree, 7=stronglyagree). Astwoitemswereexcludedduetotheirhighskewnessvalues(seeChapter3.3.1),an EFAwasperformedinordertoexaminewhetherthesixremainingitemsrepresentonesingle dimension.AccordingtotheKMOmeasure(KMO=.831)andthesignificantBartlett’stestfor Sphericity,anEFAwasapplicable.AstheperformedEFAsuggestsaone-factorsolution,the theoreticalassumptionofonedimensionofperson-companyfitcanbeconfirmed.Supporting this,allfactorloadingaresufficient(seeAppendixC).Havingninedegreesoffreedom,this measurementmodelwasclearlyover-identifiedandaCFAcanbeperformed.However,thehigh Chi-Squarevalue(χ2(9)=295.618;p=0.000)andthehighRMSEAvalue(RMSEA=0.147)indicated therejectionofthemodel.Ontheotherhand,othermodelindicesalreadysuggestedan acceptablemodelfit(GFI=.935;NFI=.916;CFI=.918).Nevertheless,amodelwithamuchbetter fitcouldbefoundbyeliminatingitemtwoanditemeightoftheconcept(seefigure9). Samplesize=1475;standardizedestimates χ2(2)=5.149;p=0.076;χ2/df=2.574;RMSEA=0.033;GFI=.998;NFI=.998;CFI=.998 Figure9:Measurementmodelforperson-companyfit Aswiththepreviousconcepts,ameanindexofthefourremainingitemswascreated (N=1475)inordertopresentthedescriptivesoftheperson-companyfit.Accordingtotheindex, therespondentshaveanaveragelylowperson-companyfit(M=2.73;SD=1.32;Min=1.98; Max=2.07),whichindicatesthattherespondentsdonothighlyidentifythemselveswiththe company.Theinternalconsistencyoftheperson-companyscaleisα=.802,indicatingagood reliabilityofthescale. Inordertotestthefourhypothesesofthemoderationeffects,twointeractionvariables arecreated.Forthis,thescoresoftwolatentvariablesaremultipliedforeachmoderator 41 variable(Schumacker&Lomax,2004).Moreprecisely,forthefirstmoderator,thestandardized meanindexofcrisisresponsibilityismultipliedbythestandardizedmeanindexofpersoncompanyfit.Inlinewiththis,forthesecondmoderator,thestandardizedmeanindexofcrisis involvementismultipliedbythestandardizedmeanindexofperson-companyfit.Next,the skwenessandkurtosisofthetwointeractionvariablesareexamined.Forthefirstmoderator, bothvaluesexceedtherequiredvaluesof2forskewnessand7forkurtosis(skewnessmoderatorI=3.66;kurtosismoderatorI=21.11).However,asthismoderatorvariableiscentraltotestingthe hypothesesofthisstudy,itismaintained.Thedeviationshavetoberememberedforthe interpretationofthedata,though.Forthesecondmoderatorvariable,noproblematicvaluescan bedetected.Thus,bothmoderatorvariablesareaddedtothemodelinordertotestthe moderatingeffects.Nexttothis,correlationsbetweenthetwomoderatorvariables,thepersoncompanyfitindexandtheindependentvariablescrisisresponsibilityandcrisisinvolvementare addedtothemodel.Sincethemoderatorvariablesaregeneratedfromtheseothertwovariables, correlationsbetweenthemareexpectedtooccur. 42 4. Results Afterthetheoreticalconceptshavebeentested,thischaptercontainsthetestingofthetwelve hypotheses,includingthemediationandmoderationeffects.Theinterpretationanddiscussion oftheresultswillfollowinchapterfive. Inordertotestthehypotheses,allabove-describedmeasurementmodelsareincludedin aFullLatentVariableModel,whichisbasedonthetheoreticallyderivedconceptualmodel(see chapter2.5).ThisisbecausetheFullLatentVariableModelenables“thespecificationof regressionstructureamongthelatentvariables”(Byrne,2001,p.6).Themodelconsistsofboth theoverallmeasurementmodelandthestructuralmodel(Byrne,2001).Thereby,theformer describestherelationshipsbetweeneachlatentconceptwiththeirobservedindicators(see Chapter3)andthelatterdescribestherelationshipsbetweenthelatentvariables(Byrne,2001). Thus,theimpliedcausalrelationshipsbetweenthelatentconcepts,asposedinthehypotheses ofthisstudy,canbetestedbyusingthestructuralmodel(Caruana&Erwing,2010).Before beingabletotestthehypotheses,themodelfitoftheFullLatentVariableModelhastobe examined.Despitetheadjustmentsthatwereundertakenforeachmeasurementmodel(see Chapter3),theoverallstructuralmodeldoesnothaveagoodmodelfit(χ2(12)=183.733; p=0.000;RMSEA=0.099;GFI=.973;NFI=.964;CFI=.966)becausetheChi-Squarevalueisbyfar toohighandtheRMSEAvalueisnotacceptableeither.AsthehighChi-Squarevaluecanbe explainedbythelargesamplesizeandthevaluesoftheotherthreemodelfitindicesaregood, themodeliskeptforthisanalysis,though.Figuretenshowsthestructuralmodelwithwhichthe relationshipsbetweenthelatentvariablesaretestedandtableonegivesanoverviewofallpath estimates4. Hypothesis1 Thefirsthypothesisassumesthatangernegativelyinfluencesthepost-crisisreputationofthe VWGroup.Itwasfoundthatangerhasahighlysignificantnegativeimpactonthepostcrisis reputation.ThismeansthattheangriertherespondentsaretowardstheVWGroupinthe contextoftheemissionsscandal,theworsetheyevaluatethepost-crisisreputationofthe corporation.Moreprecisely,whenangergoesupby1,thepost-crisisreputationgoesdownby .192.Thus,thefirsthypothesisissupported. 4Throughouttheresultsanddiscussionsections,thestandardizedbetaestimatesarepresented. 43 Figure10:StructuralModel Note:***p≤.001,**p≤.01,*p≤.05 Samplesize=1475;Standardizedestimatesareportrayed χ2(12)=183.733;p=0.000;χ2/df=15.311;RMSEA=0.099;GFI=.973;NFI=.964;CFI=.966 44 Table1:Summaryofresults Paths ModelSpecifications Beta SE coefficient Percentile confidence intervals Lower Upper Directeffectsofmediators AngeràReputation -.436*** .018 -.470 -.397 onDV(bpaths) SympathyàReputation .505*** .018 .469 .540 AngeràNWOM .644*** .016 .613 .676 SympathyàNWOM -.255*** .017 -.288 -.220 IVtomediators(apaths) ResponsibilityàAnger .213*** .021 .170 .252 InvolvementàAnger .587*** .018 .551 .621 ModeratorIàAnger -.078*** .026 -.131 -.030 ModeratorIIàAnger -.089*** .021 -.127 -.047 Person-CompanyFitàAnger -.065** .022 -.106 -.021 ResponsibilityàSympathy -.220*** .024 -.267 -.173 InvolvementàSympathy -.064*** .019 -.102 -.028 ModeratorIàSympathy .050** .030 -.002 .114 ModeratorIIàSympathy -.027 .025 -.078 .022 Person-CompanyFità .635*** .021 .590 .675 DirecteffectsofIVonDV(c ResponsibilityàReputation -.344*** .025 -.391 -.294 paths) InvolvementàReputation -.272*** .025 -.320 -.223 ResponsibilityàNWOM .209*** .022 .164 .249 InvolvementàNWOM .381*** .022 .335 .422 DirecteffectsofIVonDV ResponsibilityàReputation -.060** .023 -.105 -.014 whenmediatorsare InvolvementàReputation .032 .024 -.018 .078 included(c’paths) ResponsibilityàNWOM -.040* .020 -.081 -.001 InvolvementàNWOM -.0.042 .023 -.086 .004 Sympathy Rsquared Reputation .548 NWOM .556 Anger .458 Sympathy .550 Note:***p≤.001,**p≤.01,*p≤.05 Samplesize=1475;Standardizedestimatesareportrayed Levelofconfidenceforconfidenceintervals=95% Numberofbootstrapresamples=2,000 45 Hypothesis2 ThesecondhypothesispostulatesthatsympathypositivelyinfluencestheVWGroup’spost-crisis reputation.Theanalysishasshownthatsympathyhasahighlysignificantpositiveeffectonthe VWGroup’spost-crisisreputation.Aspeoplefeltmoresympathetictowardsthecorporationin thecontextoftheemissionsscandal,theyratedtheVWGroup’spost-crisisreputationhigher.In fact,whensympathyincreasesby1,theevaluationofthepost-crisisreputationincreasesby.272. Comparedtotheimpactofangeronthepost-crisisreputation,theeffectofsympathyiseven stronger.Hence,thesecondhypothesisissupported,aswell.Furthermore,bothpredictorsof post-crisisreputationexplain54.8percentofthevarianceofpost-crisisreputation(R2=.548).This meansthatthetwoemotionsdeterminethevarianceofpost-crisisreputationtoalargeextentbut notcompletely. Hypothesis3 Accordingtothethirdhypothesis,angerincreasestheintentionforNWOMabouttheVWGroup. Astheanalysishasshown,angerhadahighlysignificantnegativeimpactontheparticipants’ NWOMintention.ThemoreangertherespondentsfelttowardstheVWGroupregardingthe emissionsscandal,thehigherwastheirintentiontosaynegativethingsaboutthecorporation. Whenangerincreasesby1,theNWOMintentionrisesby.500.Thisimpactwasevenstronger thanthatofangeronpost-crisisreputationandthatofsympathyonpost-crisisreputation. Concluding,thishypothesiscanalsobesupported. Hypothesis4 HypothesisfourpostulatesthatsympathyhasanegativeimpactontheintentionforNWOMabout theVWGroup.ItwasfoundthatsympathysignificantlydecreasedtheNWOMintentionofthe respondents.Assympathygoesupby1,therespondents’NWOMintentiongoesdownby.242.In comparisontoanger,sympathyhadaweakereffectontheNWOMintentionoftheparticipants, though.Additionally,theimpactofsympathyontheNWOMintentionisweakerthanthatof sympathyandangeronthepost-crisisreputation.Nevertheless,thefourthhypothesiscanbe supported.Nexttothis,bothpredictorsofNWOMexplain55.6percentofthevarianceofthe dependentvariable(R2=.556).Thus,thetwoemotionsdeterminemorethanhalfofthevarianceof NWOM. 46 Hypothesis5 Thefifthhypothesisstatesthatthehigherthelevelofperceivedcrisisresponsibility,themore angertherespondentsfeeltowardstheVWGroup.Theresultsofthedataanalysissupportthis hypothesis.Whentheperceivedcrisisresponsibilityincreasesby1,thefeelingofangerraisesby .293.Hence,respondentswhobelievedthattheVWGroupwasresponsiblefortheemissions scandalexpressedmoreangertowardstheGroup. Hypothesis6 Incontrarytohypothesisfive,hypothesissixpostulatesthatahigherlevelofperceivedcrisis responsibilityleadstolesssympathytowardstheVWGroup.Asindicatedbythedataanalysis,the morecrisisresponsibilitytherespondentsascribedtotheVWGroup,thelowerwastheirfeeling ofsympathytowardsthecorporation.Whentheperceivedcrisisresponsibilitygoesupby1,the feltsympathytowardstheVWGroupgoesdownby.247.Incomparisontotheimpactonanger, crisisresponsibilityhasaminimalstrongereffectonsympathy.Concluding,thesixthhypothesis canbesupported. Hypothesis7 Hypothesissevenassumesthatinvolvementhasapositiveimpactonanger.Theresultsofthe analysissupportthisassumption:ThemoreinvolvedtherespondentswerewiththeVW emissionsscandal,theangriertheyweretowardstheVWGroup.Infact,whentheinvolvement withtheVWemissionsscandalincreasesby1,theangerrisesby.576.Basedonthehighly significantpositiveimpactofcrisisinvolvementonanger,thishypothesiscanbesupported. Hypothesis8 Onthecontrarytohypothesisseven,hypothesiseightpostulatesthatahigherlevelofcrisis involvementleadstoalowerlevelofsympathytowardstheVWGroup.Thedataanalysisrevealed thatcrisisinvolvementhadasmallsignificantnegativeimpactonsympathy.Whenthe involvementwiththeemissionsscandalincreasesby1,thesympathyfelttowardstheVWGroup decreasesby.051.Comparingthebetavaluesoftherelationshipsbetweeninvolvementandthe twoemotionsangerandsympathy,itbecomesclearthatinvolvementhasamuchstrongerimpact onangerthanonsympathy.Nevertheless,hypothesiseightcanbesupported.Inaddition,45.8 percentofthevarianceofanger(R2=.458;)and55percentofsympathy(R2=.550)arepredictedby crisisresponsibilityandcrisisinvolvement.Thismeansthatbothemotionscanbeexplainedby thetwoindependentvariablestoalargeextentbutnotcompletely. 47 4.1. TestingofMediationEffects Next,itwastestedwhetherthetwoemotionsangerandsympathyfunctionasmediatorsthat formanintermediatelinkbetweenthetwoindependentvariablescrisisinvolvementandcrisis responsibilityandthetwodependentvariablespost-crisisreputationandNWOM.Accordingto BaronandKennedy(1986),threerequirementsmustbefulfilledinordertoarguethatavariable isamediator.First,theindependentvariablesignificantlypredictsthedependentvariable(c path).Second,themediatorvariablesignificantlypredictsthedependentvariable(bpath)and last,whenthemediatorisaddedtothemodel,therelationshipbetweentheindependentand dependentvariables(cpath)isreduced(c’path)(Baron&Kennedy,1986;Little,Card,Bovaird, Preacher&Crandall,2007).Asitisdisplayedintableone,thesethreeassumptionsaremetforthe relationshipsbetweenthevariables-thusmediationeffectsoccur.Sincethedirecteffects betweencrisisresponsibilityandbothcrisisoutcomes(i.e.post-crisisreputationandNWOM) remainsignificantafterbothmediatorsareadded(c’paths),partialmediationsoccurinthese relationships(Littleetal.,2007).Asfortherelationshipsbetweencrisisinvolvementandboth crisisoutcomes,fullmediationsoccurbecausethedirecteffectisnotsignificantanymoreonce bothmediatorsareaddedtothemodel(c’path)(Littleetal.,2007). Tabletwoprovidesanoverviewofallmediationeffectsofthismodel,includingthe bootstrappingresults.Itisshownthatallmediationeffectsarehighlysignificant,meaningthat angerandsympathyfunctionasmediatorsforalleffectsbetweentheindependentanddependent variables.Whilenegativemediationeffectsoccurintherelationshipsbetweencrisisresponsibility andcrisisinvolvementwithpost-crisisreputation,positivemediationeffectsexistinthe relationshipsofthetwoindependentvariableswithNWOMintention.Thereby,themediation effectintherelationshipbetweencrisisinvolvementandNWOMintentionisthestrongest. However,notonlythetotalmediationeffectsofbothemotionsineachrelationshipareofinterest butalsotheirseparateeffects.Astabletwoshows,angerhasastrongermediatingeffectthan sympathyinallrelationshipsexceptforthatofcrisisresponsibilityandcrisisreputation.Thereby thestrongestmediationeffectofangeroccursintherelationshipbetweencrisisinvolvementand NWOM.Whilebothemotionstakeonamediatingroleinthisstudy,theseresultsindicatethat angerisamoreimportantmediatorintheVWemissionsscandalthansympathy. 48 Table2:BootstrappingResultsforMediationEffects Indirect effects of IV on DV through proposed mediators (ab paths) DV=Post-CrisisReputation IV=CrisisResponsibility Total Anger Sympathy IV:CrisisInvolvement Total Anger Sympathy DV=NWOM IV=CrisisResponsibility Total Anger Sympathy IV:CrisisInvolvement Total Anger Sympathy Beta -.204 -.093 -.112 -.288 -.258 -.032 .193 .138 .056 .394 .380 .016 SE .015 .010 .013 .016 .013 .010 .015 .014 .007 .016 .016 .005 Percentileconfidence intervals Lower -.233 -.113 -.139 -.318 -.285 -.051 .163 .110 .044 .364 .350 .007 Upper -.174 -.073 -.087 -.256 -.231 -.014 .222 .165 .072 .427 .413 .026 Note:***p≤.001,**p≤.01,*p≤.05 Samplesize=1475 Levelofconfidenceforconfidenceintervals=95% Numberofbootstrapresamples=2,000 4.2. Hypotheses9through12–ModerationEffects Thefollowingfourhypothesesaddressthemoderatingroleofperson-companyfitonthe relationshipsofcrisisinvolvementandcrisisresponsibilitywiththetwoemotions.Hypothesis nineassumesthatahigherperson-companyfitweakenstherelationshipbetweenperceivedcrisis responsibilityandanger.Thissignifiesthatifpeoplehaveahighpersoncompanyfit,theywould feellessangereveniftheyattributehighcrisisresponsibilitytotheVWGroup.Asthetestingof themoderationeffectofthisrelationshipshows,theinteractioneffectbetweencrisis responsibilityandperson-companyfithasasignificantweaknegativeeffectonanger.This indicatesthatwithahigherperson-companyfit,theinfluenceofcrisisresponsibilityonanger decreases,leadingtolessanger.Thus,thishypothesiscanbesupported:Person-companyfit functionsasamoderatorinthisrelationship. Hypothesistenpostulatesthatahigherperson-companyfitintensifiestherelationship betweenperceivedcrisisresponsibilityandsympathy.Thismeansthatifpeoplehaveahigh 49 person-companyfit,theytendtofeelmoresympathytowardstheVWGroupdespiteaperceived crisisresponsibility.Thiswouldindicatethatcrisisresponsibilitymatterslessifpeoplehavea highperson-companyfit.Resultsofthedataanalysisrevealthattheinteractioneffectbetween crisisresponsibilityandperson-companyfithasaweaksignificantpositiveimpactonsympathy. Hence,thishypothesiscanbeaccepted,aswell:Person-companyfitmoderatestherelationship betweencrisisresponsibilityandsympathy. Hypothesiselevenassumesthatahigherperson-companyfitwillweakentherelationship betweenfeltcrisisinvolvementandanger.Thissignifiesthatifpeoplehaveahighpersoncompanyfit,theywouldbelessangrywiththeVWGroupregardingtheemissionsscandaleven thoughtheyarehighlyinvolvedwiththeemissionsscandalitself.Resultsindicatethataweak significantnegativeeffectoccurredandthusthehypothesisforthisinteractioneffectcanbe supported:Person-companyfitfunctionsasamoderatorinthisrelationship. Lastly,hypothesistwelvepostulatesaninteractioneffectofperson-companyfitandcrisis involvementonsympathy,indicatingthatahigherperson-companyfitwillintensifythe relationshipbetweenfeltcrisisinvolvementandsympathy.Inthecaseofhighinvolvement,ahigh person-companyfitwouldleadtomoresympathytowardstheVWGroupaswithlowpersoncompanyfit.Testingthisassumptionshowsaminimalnegativebutnotsignificanteffect.This leadstotherejectionofthelasthypothesis:Person-companyfithasnomoderatingeffectonthe relationshipbetweeninvolvementandsympathy.Summarized,elevenofthetwelvehypothesesof thisstudycanbeaccepted. 4.3. ComparisonofAffectedandNon-AffectedPublics Inordertoexaminewhethertheabove-foundeffectsdifferforpeoplewhohavebeenaffectedby theVWemissionsscandalcomparedtothosewhowerenotdirectlyaffectedbyit,amultigroup analysiswasperformed.Asexplainedinthequestionnaire,affectedpeoplearethosewhoowna carofthebrandsVW,Audi,Seat,SkodaorPorschebecausethesehavebeenaffectedbythe emissionsscandal.Consequently,non-affectedpeoplearetheoneswhodonotownsuchcars.Itis assumedthattheeffectsofresponsibilityandinvolvementonangeraswellasthatofangeron post-crisisreputationandNWOMwouldbestrongeramongaffectedpeoplecomparedtothose whowerenotaffected.Onthecontrary,theeffectsthatinvolvedsympathywouldbeweakerfor theaffectedgroup.TheChi-Squaredifferencetestofthemultigroupanalysisrevealsthatthetwo testedmodels(affectedvs.non-affected)aresignificantlydifferent.Whilethemodelofthe affectedparticipantshasaChi-Squarevalueofχ2(12)=65.915(p=0.000),themodelofthenonaffectedparticipantshasaChi-Squarevalueofχ2(12)=148.345(p=0.000).Inordertofindout whichspecificrelationshipsdifferthetwomodels,thesinglepathestimateswereregarded.For 50 this,singlepathswerecomparedbyfreelyestimatingthetwomodelsexceptforsinglepathsthat wereconstrained(Byrne,2013). Table3:MultigroupAnalysis-affectedpublicvs.non-affectedpublic Paths ModelSpecifications Directeffectsof AngeràReputation mediatorsonDV Sympathyà Reputation AngeràNWOM Affected Non-affected Beta SE Beta SE coefficient coefficient -.438*** .008 -.435*** .008 .550*** .010 .494*** .009 .672*** .014 .640*** .014 SympathyàNWOM -.286*** .017 -.247*** .013 IVand moderatorsto mediators Responsibilityà Anger InvolvementàAnger .226*** .029 .204*** .026 .584*** .020 .583*** .020 ModeratorIàAnger -.104*** .031 -.074*** .029 ModeratorIIàAnger -.112*** .030 -.085*** .027 PersonCompanyFità Anger Responsibilityà Sympathy Involvementà Sympathy ModeratorIà Sympathy ModeratorIIà Sympathy PersonCompanyFità Sympathy Rsquared Reputation NWOM Anger Sympathy -.076*** .022 -.069*** .022 -.218*** .022 -.217*** .023 -.056*** .014 -.062*** .014 .054* .023 .042* .023 -.034 .022 -.028 .020 .630*** .016 .634*** .015 .608 .623 .426 .521 .534 .554 .459 .551 Difference Beta coefficient -.003 (p=.006) .056 (p=.353) .032 (p=.001) -.039 (p=.756) .022 (p=.026) .001 (p=.150) -.030 (p=.698) -.027 (p=.352) .007 (p=.845) .001 (p=.211) -.006 (p=.716) .012 (p=.099) -.006 (p=.556) -.004 (p=.826) Note:***p≤.001,**p≤.01,*p≤.05 Samplesizeaffected=163,Samplesizenon-affected=1312; Chi-Squareaffected=χ2(12)=148.345(p=0.000);Chi-Squarenon-affected=χ2(12)=148.345(p=0.000) Standardizedregressionweightsofstructuralweightsareportrayed 51 Tablethreeprovidesanoverviewofthepathestimatesofbothmodelsaswellasthe differencesforeachpath.Itbecomesclearthatthedifferencesbetweenthepathestimatesare verysmallandmostofthemarenotsignificant.Theonlysignificantdifferencebetweenthetwo modelscouldbefoundintheeffectthatcrisisresponsibilityhasonangerandontheimpactthat angerhasonbothcrisisoutcomes.Accordingtotheanalysis,responsibilityhadaslightlystronger impactonangeramongtheaffectedpublic.Moreover,angerhadaminimalstrongernegative impactonpost-crisisreputationandaminimalstrongerpositiveimpactonpeople’sNWOM intentionamongthosewhowereaffected.Thissignifiesthatamongparticipantswhowere affected,crisisresponsibilityresultedinmoreanger,whichinturnledtoaworseevaluationof theVWGroup’spost-crisisreputationandastrongerintentiontoexpressNWOMcomparedtothe notaffectedparticipants.Althoughthesedifferencesareminimal,thecrucialroleofangerinthis studyisunderlined. 4.4. FurtherFindings Finally,thedataanalysisrevealedsomefurtherrelevantfindings.Forthetestingofthe moderationeffects,theconceptperson-companyfitwasaddedtothemodel.Resultsofthe analysisshowedthatperson-companyfitnotonlyhasamoderatingroleinthismodelbutalso directlyaffectsonthetwoemotionsangerandsympathy.Moreprecisely,person-companyfithad aweaksignificantnegativeimpactonanger(β=-.065;p=.002)andastronghighlysignificant positiveimpactonsympathy(β=.635;p=.001).Whentheperson-companyfitincreasedbyone, angerdecreasedby.067andsympathyincreasedby.542.Thismeansthatparticipantswho identifiedthemselvesmorewiththeVWGroupfeltslightlylessangerandbyfarmoresympathy towardsthecorporationcomparedtothosewhoidentifiedthemselveslesswiththeVWGroup. Thereby,theeffectsizeperson-companyfithasonsympathyisstrongerthanalleffectsofthetwo independentvariablesonthetwoemotions(seetable1). 52 5. DiscussionandConclusion Inthischapter,theprecedingresultsofthedataanalysiswillbeinterpretedanddiscussedagainst thebackgroundoftheabove-explainedtheory.Thereby,theresearchquestionswillbeanswered aswellasmanagerialimplicationsgiven.Inaddition,conclusionsfromtheresultswillbedrawn, strengthsandlimitationsofthepresentstudypresentedaswellasimplicationsforfuture researchproposed. 5.1. InterpretationofResults OneofthemaininterestsofthisstudywastheexplorationoftheVWGroup’spost-crisis reputationandtheNWOMintentionoftheGermanpublic.Inthiscontext,itwasarguedthatthe corporation’sfavorablepre-crisisreputationcouldhavefunctionedasahalothatprotectedthe VWGroupfromreputationlossandnegativebehaviorintentionsafterthecrisis.Ontheother hand,thefavorablepriorreputationcouldhaveincreasedexpectationsofthecorporation,which wereviolatedduetotheemissionsscandalandthusresultedinnegativecrisisoutcomes.Results indicatethattherespondents’evaluationoftheGroup’spost-crisisreputationwasmodest (M=3.7;SD=1.25)andtheirNWOMintentionwasverylow(M=2.27;SD=1.43).Withrespecttothe post-crisisreputation,itwasassumedthattheVWGroup’sfavorablepriorreputationwouldsave thecorporationreputationalloss.Duetothemoderatereputationafterthecrisis,theexistenceof thehalo-effectcanbechallenged,though.However,notonlythepre-crisisreputationbutalso otherfactorscanaffectanorganization’spost-crisisreputation(seechapter2.1.1)andthusneed tobeconsideredwheninterpretingthisoutcome.Thehighlevelofcrisisresponsibilitythatis attributedinanintentionalcrisisortheVWGroup’spriorcrisishistoryandcrisisresponsesneed tobeconsidered.AsrespondentsattributedahighlevelofcrisisresponsibilitytotheVWGroup, thiscouldhavehadastrongeffectonthecorporation’spost-crisisreputation.Infact,thepostcrisisreputationcouldhaveevenbeenlowerwithoutsuchafavorablepriorreputation.This wouldthenbeanargumentfortheoccurrenceofthehaloeffect. ThelowNWOMintentionoftheparticipantswasverysurprising,consideringthefactthat theemissionsscandalwasanintentionalcrisisthatleadtotheascriptionofahighcrisis responsibility.However,asKiambiandShafer(2015)found,peoplehavealowerintentionto expressNWOMforahighreputedorganizationcomparedtoalowreputedorganization.Thus,the favorablepre-crisisreputationcouldnotonlyhaveprotectedthecorporationfrommore reputationallossduetotheemissionsscandalbutalsofromnegativebehaviorsandbehavioral intentionsdirectedtotheVWGroup.Thiswouldmeanthatpeopledidnotwanttoengagein NWOMaboutVWGroupduetotheirpositiveattitudeaboutthecorporationpriortothecrisis. 53 Thiswouldsuggesttheshieldingfunctionofafavorablepre-crisisreputationregardingnegative behaviorintentionsandindicatethattheVWGroup’spriorreputationdidplayanimportantrole inprotectingitfromNWOM. Onefurtherexplanationfortheoccurrenceofamodestpost-crisisreputationandalow NWOMintentionamongtheGermanpubliccouldbethecountry-of-origineffect.Previousstudies haveimpliedthatconsumersapplyaproduct’scharacteristics,suchasthecountryoforigin,asan evidenceforproductqualityinordertocompareaproduct’squalitytothatofothers.Thecountry oforigintherebysignalsstereotypesofproducts(Lee,Yun&Lee,2005),forinstancethe“made in”tag(Yun,Lee,andSego2002).Whileapositiveimageofanationresultsinapositive assessmentofitsproducts,anegativeoneleadstoanegativeassessment(Zhukov,Bhuiyan& Ullah,2015).Transferringthecountry-of-origineffecttoacorporatelevel,EtayankaraandBapuji (2009)concludefromtheirliteraturereviewofproductrecallsthatthemagnitudeofcompany lossesdependsnotonlyontheseverityofthecrisisorthecompany’sreputationbutalsoonthe imageofthecountry.AstheVWGroupisaGermancorporationandthelabel“madeinGermany” hasbeenassociatedwithahighproductquality(Haucap,Wey&Barmbold,1997),itcanbethus assumedthatitpreventedthecorporationfrommorenegativecrisisoutcomes. Moreoverthepresentstudyinvestigatedtherespondents’emotionalresponsestotheVW emissionsscandal.Basedonthecrisistype,aratherhighlevelofangerandalowlevelof sympathytowardstheVWGroupwereexpected.Despitetheseassumptions,therespondentshad onlyamoderatedegreeofangertowardstheVWGroupandtheemissionsscandal(M=3.32; SD=1.59),thusnotconfirmingthepriorassumption.However,asthestandarddeviationwasquite highforthisemotion,itcanbeconcludedthattheparticipantsdidhardlyagreeregardingtheir levelofanger.TakingtheVWGroup’sfavorablepre-crisisreputationintoaccount,itcouldbe assumedthatitprotectedtheGroupfrombeingfacedwithahighlevelofanger.Kiambiand Shafer(2015)forinstancefoundintheirstudythatrespondentswerelessangrytowardsan organizationwithafavorablepre-crisisreputationthantowardsanorganizationwithaprior unfavorablereputation.Anotherexplanationforthismodestlevelofangercouldbethetime betweentheoutbreakoftheemissionsscandalandthesurvey.AccordingtoCoombsandHolladay (2007),emotionalresponsescandecreaseovertimebecausethestakeholdersforgetabouta crisis.Asthisstudywasconductedaboutsixmonthsafterthefirstinformationontheemissions scandalwaspublished,itcouldbethecasethattherespondentsweresimplynotangryanymore. Onthecontrary,theassumedlowlevelofsympathyfeltbytherespondentswasconfirmed bythisstudy(M=2.14;SD=1.22).Thisprovesthatalsopositiveemotionscanemergeinacrisis,as FolkmanandMoskowitz(2000)hadproposed.CoombsandHolladay(2005)foundthatthemost sympathywasfeltinthecaseofavictimcrisisinwhichlowresponsibilitywasattributedtothe 54 organization.Anintentionalcrisisontheotherhandresultedinalowlevelofsympathy.TheVW emissionsscandalcannotonlytheoreticallybecategorizedasanintentionalcrisiswithahigh crisisresponsibility,butalsotherespondentsofthisstudyattributedahighresponsibilitytothe corporation.Therefore,itisinlinewithpreviousexpectationsthattheyfeltonlylittlesympathy towardstheVWGroupaftertheemissionsscandalhadoccurred. Regardingtheeffectofbothemotionsonthetwoinvestigatedcrisisoutcomes,itwas foundthatangerhadanegativeimpactontheVWGroup’spost-crisisreputationandsympathya positiveeffectonthiscrisisoutcome.Thus,thesefindingsarenotonlyinlinewiththatofChoiand Lin(2009b)butalsoexpandit,asnotonlyaneffectofangerbutalsoofsympathyonreputation wasfound.Moreover,theseresultsconfirmoneextensionoftheSCCTmodelwithadirectpath fromemotionstoreputation,asChoiandLin(2009b)proposed.Further,itisespecially interestingthatsympathyhadanevenstrongereffectonreputationthananger.Thisimpliesthe importanceofconsideringnotonlythenegativebutalsoespeciallythepositiveemotionina crisis. Moreover,bothemotionshadanimpactontheparticipants’intentionforNWOM,with angerresultinginahigherintentiontoexpressnegativestatementsabouttheVWGroupand sympathyleadingtoalowerNWOMintention.ThesefindingsarethusinlinewithMcDonaldet al.’sstudy(2010),whichfoundthatahigherlevelofangerresultedinahigherNWOMintention. Furthermore,theresultsconfirmedpartofthenegativecommunicationdynamicthatwasposed byCoombsetal.(2007).However,whileCoombsetal.(2007)didnotfindevidencefortheir hypothesisandconcludedthatthemoderatelevelofangerwasnotenoughtoresultinNWOM,the studyathandprovedthatalreadyalowlevelofangerstronglyincreasedpeople’sintentionto expressNWOM.Duetoitseffectsize,thesignificanceofangerisevenmoreemphasized.However, assympathywasfoundtodecreasepeople’sNWOMintention,thispositiveemotionshouldnotbe leftouteither.Thus,whileangerhadahigherimpactonreputationthanontheNWOMintention, sympathyhadastrongereffectonpeople’sintentiontouseNWOMthantheirevaluationofthe VWGroup’spost-crisisreputation.Nevertheless,allfoureffectswerequitehigh,underliningthe decisiveroleofbothemotionsinthecontextoftheVWemissionsscandal. Besideshavingfoundevidencefortheimpactsofbothemotionsonthetwocrisis outcomes,thepresentstudyalsoconfirmedthemediatingroleofangerandsympathyinthe relationshipsofcrisisresponsibilityandcrisisinvolvementwiththepost-crisisreputationand NWOMintention.Thereby,angerformedanespeciallystronglinkbetweenthetwoindependent variablesandNWOM.Overall,sympathydidnotmediatetherelationshipsbetweenthe independentanddependentvariablesasstronglyasanger,withoneexceptionbeingthelink betweencrisisresponsibilityandthepost-crisisreputation,wheresympathyhadaslightly 55 strongerimpactthananger.Thisemphasizesthesignificanceofavoidingangerinacrisiseven further. SincetheVWemissionsscandalcanbecategorizedintheintentionalcrisiscluster,itwas assumedthattherespondentswouldattributeahighlevelofcrisisresponsibilitytotheVW Group.ResultsshowthattherespondentstendedtobelievethattheVWGroupwasresponsible fortheoccurrenceoftheemissionsscandal(M=5.91;SD=1.19).Hence,thisstudy’sfindings confirmthisproposedassumption.Wheninterpretingthesefindings,itshouldbeconsidered, though,thattheinitialmeasurementofcrisisresponsibilitywasreducedinthedataanalysis process.Astheremainingitemsbelongtotheaccountabilitydimensionoftheoriginal measurementbyBrownandKi(2013),theresponsibilitysolelystandsfor“thedegreetowhich theorganizationcouldhaveavoidedthecrisis“(p.14).Thus,wheninterpretingtheVWGroup’s crisisresponsibilityinthisstudy,itdoesnotrefertointentionality,meaning“thedegreetowhich thecrisiswascreatedpurposefullybyamemberormembersoftheorganization,”norlocality, referringto“thedegreetowhichthecrisisisaninternalmatter”(Brown&Ki,2013,p.14).Hence, theparticipantsbelievedtoahighextentthattheVWGroupisaccountablefortheemissions scandalandthatitcouldhaveavoidedthecrisisfromoccurring. Thisstudyrevealedthatcrisisresponsibilityresultedinbothangerandsympathy,thus confirmingCoombs’(2007a)andCoombsandHolladay’s(2005)assumptionsthatcrisis responsibilitytriggersthetwomainemotionsofAT.AshypothesizedandinlinewithSCCT (Coombs,2007a),ahigherperceivedcrisisresponsibility,thusholdingtheVWGroupaccountable fortheoccurrenceoftheemissionsscandal,resultedinmoreangerbutlesssympathytowardsthe corporationcomparedtothosewhoassignedlesscrisisresponsibilitytotheVWGroup.Thereby, theeffectsofcrisisresponsibilityonbothemotionshadapproximatelythesamestrength.Thus, contradictingtoChoiandLin’s(2009b)study,thepresentstudyalsofoundevidenceforthe relationofresponsibilityandsympathy,thereforeconfirmingitsroleasapredictorforpositive andnegativeemotionsinacrisis. Meetingthecallforexaminingtheconceptofinvolvementincrisiscommunication research(McDonald&Härtel,2000),thisstudyincorporatedcrisisinvolvementasapredictorfor emotions.BecauseoftheVWGroup’simportancefortheGermanpopulationanditsfavorable priorreputation,itwasarguedthattheemissionsscandalwouldbeperceivedashighlyrelevant bytheGermanpublic.Furthermore,theemissionsscandalwasquiteunexpectedanddeveloped toasevere,internationalcrisis.Thisisanotherreasonwhyitwasassumedthatitdidmatterto theGermanrespondents.Thisstudy’sresultsindicate,though,thattherespondentswereonly involvedonamodestlevel(M=4.27;SD=1.57),thusnotconfirmingthisassumption.Onepossible explanationforthisisthechangeabilityanddynamicoftheconceptoffeltinvolvement 56 (McDonald&Härtel,2000).Itcouldbethecase,thatthefeltcrisisinvolvementwasespecially highwhentheemissionsscandalfirstwentpublic.However,asfeltinvolvementisapersonal statethatcanchangeovertime(Celsi&Olson,1988),itmightbethecasethatthelevelofcrisis involvementhasdecreasedsincethen.Inaddition,theemissionsscandalcouldhavebeen perceivedaslessrelevantbecausemoreandmoreinformationonalsootherautomotive manufacturerswhohavemanipulatedcarswerepublished(e.g.Weingartner,2015). ResultsprovideevidenceforthesignificanceoftheconceptintheVWemissionsscandal. Respondentswhoweremoreinvolvedwiththeemissionsscandal,meaningthattheyperceived thecrisisaspersonallyrelevant,feltsignificantlyangriertowardsthecorporation.Thisfinding supportsMcDonaldandHärtel’s(2000)andCoombsandHolladay’s(2005)assumptionsthatthe levelofinvolvementdeterminesaperson’sintensityofemotionsinacrisis.Consideringthe strengthofthiseffect,crisisinvolvementwasveryimportantforpredictingangerintheVW emissionsscandal.Theimpactofcrisisinvolvementwasevenstrongerthanthatofcrisis responsibilitywhendetermininganger.ThisiscontrarytothefindingsofMcDonaldetal.(2010), whofoundthatcrisisresponsibilitywasamoreimportantpredictorforemotionalreactionsina crisisthaninvolvement.GiventhefactthatMcDonaldetal.(2010)usedthesameinitial measurementforcrisisinvolvementbyMcQuarrieandMunson(1992),thisfindingisespecially interesting.Theseopposingfindingscanbeexplainedbythedifferentresearchapproachesthat wereusedbyMcDonaldetal.(2010)andinthestudyathand.WhileMcDonaldetal.(2010)used anexperimentalapproachwithanartificialairlinecompanythathasexperiencedanartificial crash,thepresentstudyusedasurveymethodandarealcrisisscenario.Regardingtheinfluence ofcrisisinvolvementonsympathy,thisstudyfoundthatahigherleveloffeltcrisisinvolvement resultedinlesssympathytowardstheVWGroup,thusconfirmingtheresultsofMcDonaldetal.’s (2010)study.Theimpactofcrisisinvolvementonsympathywasratherlow,though,comparedto thatonanger.Furthermore,incomparisonwiththeeffectthatcrisisresponsibilityhadon sympathy,theeffectofcrisisinvolvementwasonlyabouthalfasstrong,whichisinlinewiththe studyofMcDonaldetal.(2010)aswell.Concluding,crisisinvolvementwasastrongerpredictor foranger,whilecrisisresponsibilitywasastrongerpredictorforsympathy.Thisindicatesthe crucialroleofinvolvementinacrisisandimpliestheincorporationoftheconceptintheSCCT framework. Inordertodrawaconnectionbetweentheindividualandcorporateleveloftheemissions scandal,theidentificationofGermanswiththeVWGroupwasincludedinthisstudyasthe person-companyfit.DuetotheVWGroup’sfavorablepre-crisisreputation,itwasassumedthat therespondentswouldhaveahighidentificationwiththecorporation.Despitethisexpectation, therespondentshadanaveragelylowperson-companyfit(M=2.73;SD=1.32)andthusalow identificationwiththeVWGroup.Onereasonforthelowlevelofidentificationcouldbethe 57 impactoftheVWemissionsscandalontheparticipants’identificationwiththecorporation. Althoughthequestionsabouttheperson-companyfitwereaskedbeforeevenmentioningthe emissionsscandalinthequestionnaireinordertokeeptheimpactofthecrisisaslowaspossible, itcouldstillbethecasethattheemissionsscandalaffectedtheparticipants’identificationwith thecorporation.ThisiscontradictingtoAshforthandMael’s(1989)assumptionthatthe identificationwithagroupenduresevenafterthefailureofagroup.Ontheotherhand,itisinline withBergeretal.(2006)whosuggestthatchangedmembersofanorganizationcanaffectbeliefs andidentification.AnotherpossiblereasoncouldbethesizeandcompositionoftheVWGroup.As thecorporationconsistsofseveralsubsidiaries,productsandbrands,itcouldbeverydifficultfor peopletoidentifythemselveswiththecorporationasawhole.Instead,itcouldbethecasethat theyaremorelikelytoidentifythemselveswithcertainbrands,suchasVWorPorsche,or products,suchastheVolkswagenCamperT2. Itwasarguedthattheleveloftherespondent’sidentificationwiththeVWGroupwould affecttherelationshipsofcrisisresponsibilityandcrisisinvolvementwiththeemotional responsesangerandsympathy.Resultsofthisstudyshowthatperson-companyfitfunctionedas amoderatorintherelationshipsbetweencrisisresponsibilitywithbothemotionsandbetween involvementandanger.InthecasethatpeopleidentifiedthemselvesmorewiththeVWGroup, theinfluenceofcrisisresponsibilityandcrisisinvolvementonangerwasweakened,resultingin lessangertowardstheVWGroup.Moreover,ahigheridentificationwiththeVWGroup strengthenedtheimpactofcrisisresponsibilityonsympathy,leadingtomoresympathy.Further findingsalsorevealed,thatperson-companyfitnotonlymoderatedsuchrelationshipsbutalso directlyaffectedbothemotions.Thereby,ithadaparticularlystrongimpactonsympathy.Allof thesefindingsimplythatbuildingastrongerperson-companyrelationshipcanreducethe negativeemotionalreactionandincreasethepositiveaffect.Thisunderlinesthesignificanceofa highperson-companyfitintheVWemissionsscandal.TheyalsoconfirmAshforthandMael’s (1989)argumentthatmoreidentificationresultsinmoresupportforthecorporation–inthis caseemotionalsupport.However,thisstudydidnotfindprooffortheassumptionthatahigh corporatereputationresultsinahighidentificationwiththecompany(Bhattacharya&Sen,2003; Duttonetal.,1994;Keh&Xie,2009).Nevertheless,theseresultssuggesttheconsiderationof person-companyfitincrisiscommunicationandSCCT. Finally,followingKiambiandShafer’s(2015)suggestionofcomparingvictimsandnonvictimsofacrisis,thisstudyperformedacomparisonbetweenparticipantswhowereaffectedby theemissionsscandalandthosewhowerenotaffected.Itwasfoundthatcrisisresponsibility increasedthelevelofangermoreamongaffectedpublicscomparedtonon-affectedpublics. Furthermore,angerhadastrongernegativeeffectontheperceivedpost-crisisreputationofthe VWGroupandastrongerpositiveeffectontherespondents’intentiontoexpressNWOMamong 58 theaffectedparticipants.Thus,angerleadtomorenegativecrisisoutcomesamongtheaffected personsthanamongthegeneralpublic.Assumingthataffectedpeopleregardedtheemissions scandalasmoreself-related(Celsi&Olson,1988)andthusevaluateditasmorepersonally relevant,itseemsplausiblethattheirangeralsohadmoreseverenegativeoutcomesthanfor peoplewhowerenotaffected.Thisresultparticularlyemphasizesthatdifferentgroupsof stakeholderscanreactdifferentlytoacrisisandthatespeciallyaffectedpeoplecouldreactmore negativelytosuchanevent. 5.2. ManagerialImplications Basedontheresultsofthisstudy,severalpracticalsuggestionscanbemade.Asthisstudy confirmed,angerplaysapowerfulroleinacrisis.Notonlydiditnegativelyinfluencethe participants’evaluationontheVWGroup’spost-crisisreputation,italso,andmorestrongly, increasedtherespondents’intentiontoexpressNWOMaboutthecorporation.Thestudyfurther foundthatalreadyalowlevelofangercouldresultinthesenegativecrisisoutcomes.Moreover, thisstudyrevealedthatsympathyfunctionedasapredictorforpost-crisisreputationandNWOM intention,aswell.Althoughcorporationsaimtoavoidnegativeconsequencesandenhance positiveoutcomesofacrisis,theycannotcontrolpeople’semotions.Whattheycaninfluence, though,arethedrivingforcesofangerandsympathy. Thisimpliesfirstofall,thatcorporationsshouldunderstandandtakeseriously(Choi& Lin,2009b)bothemotionalresponsesandtheirconsequencesinordertoreactproperlytoa crisis(Laufer&Coombs,2006).Inordertomitigateangerandreinforcesympathy,theyshould carefullycommunicatewiththeirpublicsbasedontheiremotionalstateatacertainmoment.For this,corporationsshouldconstantlymonitornewscoverageandthepublic’sreactionstoacrisis, suchasonsocialmediaplatformslikeTwitterandFacebook.Basedonsuchknowledge, corporationscouldthenidentifyandapplythebestresponsetotheincident(Jin,2014).Byusinga propercrisisresponse,whichforinstancefocusesonthewellbeingoforcaringforvictims,also theattributionofcrisisresponsibilitycouldbeweakened.Thisisespeciallysignificantascrisis responsibilitywasfoundtoaffectthelevelofangerandsympathy.Thus,byreactingproperlytoa crisis,theusageoftherightcrisisresponsestrategycouldpreventtheoccurrenceoforreducethe levelofangerandenhancethelevelofsympathy. Evenmoreimportantly,thisstudyhasrevealedthatcrisisinvolvementisanevenstronger predictorofangerthancrisisresponsibilityandalsohasanimpactonsympathy.Thus,when respondingtoanincident,corporationsshouldconsiderpeople’sdegreeofinvolvement(Choi& Chung,2013)andthereforeadjusttheirresponsestoeachgroupofpeople.Thereby,theyshould payspecialattentiontotheirdifferentgroupsofstakeholders,suchascustomersorshareholders, 59 astheyareassumedtoevaluateacrisisasparticularlyrelevant.Asthisstudyhasshown,alsothe generalpublicshouldnotbeforgotten,though,whenrespondingtoacrisis.Thus,public statements,forinstanceinthemedia,shouldbeprovidedbytheaffectedcorporation,aswell. Thisstudyhasfurtherfoundevidencefortheimportanceofastrongperson-companyfit forreducingangerandincreasingsympathyinacrisis.Basedonthisstudy’sfindings,itisimplied thatcorporationsshouldbuildandmanageastrongrelationshipnotonlywiththeirstakeholders butalsothegeneralpublic.Forthis,corporationsshouldgettoknowtheneedsandwantsof differentpublics,forinstancebyconductingregularsurveys.Thisway,corporationscannotonly performproperrelationshipmanagementwithpublicsbutcanalsoofferandcommunicate productsandservicesthatthepublicscanidentifywith.Thesecouldbeforinstancebethosethat dogoodfortheenvironmentorsociety,thusmeetingtheincreasingdemandforenvironmentallyfriendlyproducts(iwd,2013). Lastly,resultsofthisstudyimplythatcorporationsshouldbeawareabouttheoccurrence andeffectofemotionalresponsesonpost-crisisreputationandNWOMintentionevenseveral monthsafteracrisishasoccurred.Thus,suchcorporationsshouldnotonlyimplementsuch actionsrightafteracrisishashappenedbutalsointhelongterm. 5.3. Conclusion TheaimofthisthesiswastoexplorecrisisoutcomesoftheVWemissionsscandal,analyzethe impactofpersonalperspectivesontheseoutcomesaswellascomparetheserelationships betweenaffectedandnon-affectedpublics.Thereby,thisstudyfocusedontheGermanpublic’s evaluationoftheVWGroup’spost-crisisreputationandtheirintentiontouseNWOM.Applying theSCCTasgroundworkandextendingtheframeworkwiththeconceptsofcrisisinvolvement andperson-companyfit,aswellasthelinkbetweenemotions(i.e.angerandsympathy)andpostcrisisreputation,thisthesisexaminedtheroleofcrisisresponsibility,crisisinvolvement, emotionsandperson-companyfitinthecrisiscontext.AstheVWGrouphadagoodpre-crisis reputationandahighrelevanceasacarmanufacturerespeciallyinGermany,theVWemissions scandalwasevaluatedasparticularlyrelevantinthiscountry.Thus,asurveyamongtheGerman publicwasconductedforthepurposeofthisstudy.Both,thehighnumberofparticipants (N=2072)andthehighresponserate(72.91%)ofthissurveyconfirmedtherelevanceofthe emissionsscandalforGermans. Thisstudyhasshownthatalloftheconceptsplayedanimportantroleinthetested relationships.FindingsrevealedthattherespondentsevaluatedtheVWGroup’spost-crisis reputationonlyonamodestlevel,leadingtotheassumptionthatthecorporationdidnotmaintain itsfavorablereputationafterthecrisis.Thestudyfurtherfoundthattherespondents’NWOM 60 intentionwasratherlowafterthecrisis.Thus,itwassupposedthattheVWGroup’sfavorable priorreputationcouldhavesavedthecorporationfrombeingconfrontedwithahigherlevelof NWOM.Besides,thisstudyconfirmedtheeffectsofemotionsintheemissionsscandaloncrisis outcomes.Whileangerledtoamorenegativeperceptionofpost-crisisreputationandtoahigher intentiontouseNWOM,sympathyresultedinamorefavorablepost-crisisreputationandalower intentiontouseNWOMabouttheVWGroup.Hence,evidencewasfoundfortheinSCCTexistent linkbetweenemotionsandNWOMintentionbutalsofortheproposedlinkbetweenemotionsand post-crisisreputation.Moreover,thesignificanceofthetwoemotionalresponseswasevenmore emphasizedbyprovingitsmediatingrolesintherelationshipsbetweencrisisresponsibilityand crisisinvolvementwithpost-crisisreputationandNWOMintention.Thereby,angerwasfoundto beastrongermediator,highlightingtherelevanceofthisnegativeaffectinacrisis.However, sympathywasalsoshowntobeimportantinpredictingcrisisoutcomes,thusconfirmingits proposedrelevanceinthecrisis. Furthermore,thisstudynotonlyconfirmedthecrucialroleofcrisisresponsibilitybutalso thatofcrisisinvolvementasapredictorforemotions:Bothconceptsincreasedthelevelofanger anddecreasedthelevelofsympathyamongtherespondents.Hence,itwasproventhatcrisis involvement,whichisarathernewconceptinthecontextofcorporatecrises,isarelevant predictorforcrisisemotions,aswell.BesidestestingtheextensionoftheSCCTframeworkwith crisisinvolvement,thisstudyalsoincludedtheconceptofperson-companyfit.Accordingtothe findings,person-companyfitmoderatedthelinksbetweencrisisresponsibilityandbothemotions aswellasbetweencrisisinvolvementandanger.Thereby,astrongeridentificationwiththeVW Groupresultedinmoresympathyandlessanger.Finally,thecomparisonofalleffectsamong affectedandnon-affectedpublicsshowedthatcrisisresponsibilityresultedinmoreangeramong theaffectedpubliccomparedtothenon-affectedpublic.Inaddition,theimpactofangerwas strongeramongtheaffectedthanamongthenon-affectedGermanpublic,whilenodifferencewas observedregardingsympathy. Hence,havingconfirmedelevenofthetwelveposedhypotheses,thisstudynotonlyfound evidenceforexistingpathsoftheSCCTframeworkbutalsofortheproposedextensions(i.e.the linkbetweenemotionsandpost-crisisreputation,aswellastheincorporationofinvolvementand person-companyfit).Thus,greatvaluewasaddedtocrisiscommunicationresearchbyvalidating theframeworkforthisrealcrisisscenario. 5.4. StrengthsandLimitations Thisstudyaddedvaluetoexistingcrisiscommunicationresearchforseveralreasons.Firstly,it notonlyfoundevidenceforexistingpathsoftheSCCTbutalsoforproposedextensionswhen 61 testingtheframeworkinarealcrisisscenario.Thereby,thisstudynotonlymetthecallfor researchbyexaminingthepersonalperspectivesemotionsandinvolvementinacrisis.Italso confirmedtheimpactofbothemotions(i.e.angerandsympathy)onpost-crisisreputationandthe significantroleofcrisisinvolvementasapredictorforemotionalresponses.Thus,notonly productinvolvement,asproposedbyChoiandLin(2013),butalsocrisisinvolvement,wasfound tobeimportantwhenstudyingcrisisemotionsandshouldbeincorporatedintheSCCT.Nextto this,thisstudyappliedperson-companyfittotheorganizationalcrisiscontextandfoundevidence foritsimpactonemotions. Inaddition,theVWemissionsscandalconstitutedasignificantcasetostudy.Itenabledthe investigationofasevereandinternationalproductrecallcrisisoutsideoftheUnitedStates.By analyzingthiscrisis,theartificialityofafictitiousorganizationand/orcrisiswasavoided. Moreover,andmoreimportantlythecomparisonofeffectsbetweenaffectedandnon-affected groupswasenabledanditwasrevealedthatangerhadastrongerimpactoncrisisoutcomes amongaffectedpersons.Inadditiontothis,thepresentstudycontributedtoexistingresearchby applyingaquantitativesurveymethod,thusgoingbeyondcasestudyresearch,whichhadbeen dominatingthefieldofcrisiscommunication.Inthiscontext,thehighnumberofparticipants shouldbementionedthatenabledtheapplicationofSEMforthedataanalysis. Despitethesestrengthsandtheencouragingresultsofthisstudy,certainlimitations shouldbetakenintoconsiderationwheninterpretingthesefindings.Firstofall,thestudyathand onlymeasuredtheVWGroup’sreputationafterthecrisisbutnotthatbeforetheoccurrenceofthe emissionsscandal.AlthoughseveralsourcesagreeontheVWGroup’sfavorablereputationbefore thecrisis,noreputationlosscouldbeexplicitlybedetectedbasedonthisstudy’sresults. Moreover,thepost-crisisreputationwasonlymeasuredonceandthusthedynamicoftheconcept ofreputationwasnotconsidered.However,anorganization’sreputationdevelopsovertimeand canchangequicklyduetotheappearanceofnewevidence(Choi&Chung,2013).Inthiscontext,it hastobeconsideredthatthisstudywasconductedabouthalfayearafterthefirstinformationon theemissionsscandalwasdisclosed.Thiscouldhavehadaneffectontheevaluationofthe perceptionofthepost-crisisreputationaswellasontheotherrelevantconceptsofthisstudy, meaningforinstancethatthereputationhadalreadyrecoveredorthelevelofangerhadalready decreasedduetothetimepassed. Furthermore,thisstudyappliedSCCTbutdidnotincludeallelementsoftheframework. Forinstance,itdidnotconsidercrisisresponsestrategiesthatwereusedbytheVWGroupto reacttothecrisisortheVWGroup’scrisishistory.Theseareimportantfactors,though,when evaluatingcrisisoutcomes(seechapter2.6.). 62 Anotherlimitationofthisistheway,inwhichperson-companyfitwasincludedinthe study.Firstly,asmentionedinchapter5.1.1.,itcouldhavebeendifficultfortherespondentsto evaluatetheiridentificationwiththeVWGroupbecauseitisalargecorporationwithseveral brandsandproducts.Peoplemightratheridentifythemselveswithabrandorproductthough, insteadwithsuchalargeentity.Secondly,thescalethatwasusedtomeasureperson-companyfit includedsomeitemsthatweredifficulttoanswerbyrespondentsbecausetheyinitiallystemfrom theconceptofemployer-employeeidentification.Asseveralrespondentsleftacommentabout thisdifficultyofansweringinthefeedbackfieldintheendofthesurvey,itissupposedthatthis problemhadoccurred. Nexttothis,themodelfitofthestructuralmodelwasonlyacceptableandnotashighas desired.Thishastobeconsideredwheninterpretingtheresultsofthisstudy.Finally,some limitationsarearesultofthechoiceofmethodandsamplingmethodthatwasusedforthedata collection.Althoughsurveymethodshavethestrengthtomeasurepeople’sopinionsand behaviors,itisdebatablewhethertheyarethebestmethodtomeasureemotionsaswell. Moreover,itshouldbeconsideredthatduetotheself-reportinginasurvey,somepeoplecould havethetendencytoanswerinasociallydesirablemanner.Asaresultoftheusageofpurposive samplingandtheself-selectionofrespondents,norepresentativesamplecouldbeachieved. Instead,thesampleconsistsofmostlyyoungandhighlyeducatedrespondentswhileolderand lowereducatedpartsoftheGermanpopulationareunder-represented.Thisbiasneedstobe takenintoconsiderationwheninterpretingtheresultsofthisstudy. 5.5. FutureResearch AstheVWemissionsscandaloriginatedintheUnitedStatesandevolvedintoaninternational crisis,itwouldbeworthwhiletoreplicatethepresentstudyintheUnitedStates.Conductingthe samestudyinanothercountrythatwasaffectedbytheemissionsscandalwouldalsoshedmore lightintothemeaningofboththeVWGroupandtheemissionsscandalforGermanyandother countries.OnlywhencomparingresultsfromaGermansamplewiththatofothernationalities, thefullsignificanceofthisstudy’sresultswouldbecomeclear.Inaddition,theVWemissions scandalconstitutesasuitablecaseforalong-termstudy.Byrepeatingthesamesurveyin Germanyaftersometime,valuableknowledgeaboutthelong-termcrisisoutcomescouldbe gained.Basedonthis,suggestionscorporationsonhowtohandleacrisisinthelongruncouldbe given. Moreover,futurestudiesontheVWemissionsscandalcouldincludeotherindependentor dependentvariablesthathavenotbeenconsideredinthepresentstudy.Forinstance,other behavioralintentionsthatareofinterestfororganizations,suchastheintentiontore-purchasea 63 productortheboycottofacorporation,couldbeincorporated.Inadditiontothis,thecrisis historyandcrisisresponsestrategiesthatwereusedbytheVWGroupcouldbeinvestigated,for instance,regardingtheirimpactonthecorporation’spost-crisisreputation.Inthiscontext,the roleoftheVWGroup’sformerCEOMartinWinterkornwouldbeanaspectworthconsidering.As previousresearchhasshown,defensivecrisisresponseandCEOvisibilityinimmediatecrisis responsewasthemostefficientforgeneratingthemostpositiveattitudeandmostpositive purchaseintentioninacrisis(Turketal.,2012). Thepresentstudyhasfoundevidenceforthesignificantroleofinvolvementinacrisis. SincetheconceptisstillnewinbothcrisiscommunicationresearchandtheSCCTframework, though,prospectiveresearchshouldcontinuetoinvestigatecrisisinvolvementinthiscontext. Additionally,inpreviousresearch,involvementhadbeenoperationalizedindifferentways,for instanceasproductinvolvementorascrisisinvolvement.Inordertoconfirmtheimportanceof thisconceptandtheresultsofthisstudy,moreconsistentresearchisnecessary.Thesameapplies forperson-companyfit.Thestudyathandintroducedtheconceptasaninfluencingfactoron emotionsinacrisis.Although,thisrelationshipcouldbeconfirmedtoalargeextentinthisstudy, futureresearchshouldcontinuetoexploreperson-companyfitanditseffectsonemotionsand crisisoutcomes,suchaspurchaseintention,inorganizationalcrises. Finally,inordertogainarepresentativesampleandthusgenerizableresults,thisstudy shouldbereplicatedusinganon-purposivesamplingmethod,suchasquotasampling.Thiswould enabletheinclusionofindividualswithcertainrelevantcharacteristicsinthesampleandhencea lessbiasedsample(Möhring&Schlütz,2010). 64 References Anderson,E.W.(1998).Customersatisfactionandwordofmouth.JournalofServiceResearch, 1(1),5-17. Arpan,L.M.(2002).WheninRome?Theeffectsofspokespersonethnicityonaudienceevaluation ofcrisiscommunication.JournalofBusinessCommunication,39(3),314-339. Ashforth,B.E.,&Mael,F.(1989).Socialidentitytheoryandtheorganization.TheAcademyof managementreview,14(1),20-39. Babbie,E.R.(2011).Introductiontosocialresearch.Stamford,CT:WadsworthCengagelearning. Baron,R.M.,&Kenny,D.A.(1986).Themoderator–mediatorvariabledistinctioninsocial behaviours.JournalofPersonalityandSocialPsychology,51(6),443-464. Bender,R.(2015,Sept.25).VolkswagenScandalTestsAuto-LovingGermany.TheWallStreet Journal.Retrievedfromhttp://www.wsj.com/articles/vw-scandal-tests-auto-lovinggermany-1443217183 Bentler,P.M.(1992).OnthetofmodelstocovariancesandmethodologytotheBulletin. PsychologicalBulletin,112,400–404. Bergami,M.,&Bagozzi,R.P.(2000).Self-categorization,affectivecommitmentandgroupselfesteemasdistinctaspectsofsocialidentityintheorganization.BritishJournalofSocial Psychology,39(4),555−577. Berger,I.E.,Cunningham,P.H.,&Drumwright,M.E.(2006).Identity,identification,and relationshipthroughsocialalliances.JournaloftheAcademyofMarketingScience,34(2), 128-137.doi:10.1177/0092070305284973 Bhattacharya,C.B.,&Sen,S.(2003).Consumer-companyidentification:Aframeworkfor understandingconsumers'relationshipswithcompanies.JournalofMarketing,67(2),76-88. doi:10.1509/jmkg.67.2.76.18609 Birchall,J.&Milne,R.(2009,November13).Recallsechoroundglobeinsocialnetworkera.The FinancialTimes.Retrievedfromhttp://www.ft.com/intl/cms/s/0/1af4dfe6-d09c-11deaf9c-00144feabdc0.html#axzz3yqzpb9GL Bollen,K.A.,&Stine,R.(1990).Directandindirecteffects:Classicalandbootstrapestimatesof variability.Sociologicalmethodology,20(1),15-140. Bromley,D.B.(2000).Psychologicalaspectsofcorporateidentity,imageandreputation. CorporateReputationReview,3(3),240-252.doi:10.1057/palgrave.crr.1540117 Brosius,H.B.,Haas,A.&Koschel,F.(2012).MethodenderempirischenKommunikati-onsforschung. EineEinführung(6thedition).Wiesbaden:SpringerVS. Brown,J.J.,&Reingen,P.H.(1987).Socialtiesandword-of-mouthreferralbehavior.Journalof ConsumerResearch,14(3),350-362. Brown,K.A.,&Ki,E.J.(2013).Developingavalidandreliablemeasureoforganizationalcrisis responsibility.Journalism&MassCommunicationQuarterly,90(2),363-384.doi: 10.1177/1077699013482911 Browne,M.W.,&Cudeck,R.(1993).Alternativewaysofassessingmodelt.InK.A.Bollen&J.S. Long(Eds.),Testingstructuralequationmodels(pp.136–162).NewburyPark,CA:Sage. 65 Burgoon,J.K.,&LePoire,B.A.(1993).Effectsofcommunicationexpectancies,actual communication,andexpectancydisconfirmationonevaluationsofcommunicatorsandtheir communicationbehavior.Humancommunicationresearch,20(1),67-96. Byrne,B.M.(2013).StructuralequationmodelingwithAMOS:Basicconcepts,applications,and programming.NewYork,NY:Routledge. Carroll,C.(2009).DefyingaReputationalCrisis–Cadbury’sSalmonellaScare:WhyareCustomers WillingtoForgiveandForget?CorporateReputationReview,12(1),64-82. Caruana,A.,&Ewing,M.T.(2010).Howcorporatereputation,quality,andvalueinfluenceonline loyalty.JournalofBusinessResearch,63(9),1103-1110.doi:10.1016/j.jbusres.2009.04.030 Celsi,R.L.,&Olson,J.C.(1988).Theroleofinvolvementinattentionandcomprehension processes.Journalofconsumerresearch,15(2),210-224. Choi,J.,&Chung,W.(2013).AnalysisoftheInteractiveRelationshipBetweenApologyand ProductInvolvementinCrisisCommunication:StudyontheToyotaRecallCrisis.Journalof BusinessandTechnicalCommunication,27(1),3-31.doi:10.1177/1050651912458923 Choi,Y.,&Lin,Y.H.(2009a).Consumerresponsetocrisis:Exploringtheconceptofinvolvementin Mattelproductrecalls.PublicRelationsReview,35,18–22.doi: 10.1016/j.pubrev.2008.09.009 Choi,Y.&Lin,Y.H.(2009b).ConsumerResponsestoMattelProductRecallsPostedonOnline BulletinBoards:ExploringTwoTypesofEmotion.JournalofPublicRelationsResearch, 21(2),198-207.doi:http://dx.doi.org/10.1080/10627260802557506 Choi,Y.&Lin,Y.H.(2009c).Individualdifferenceincrisisresponseperception:Howdolegal expertsandlaypeopleperceiveapologyandcompassionresponses?PublicRelations Review,35,452-454.doi:10.1016/j.pubrev.2009.07.002 Chu,K.K.,&Li,C.H.(2012).Thestudyoftheeffectsofidentity-relatedjudgment,affective identificationandcontinuancecommitmentonWOMbehavior.Quality&Quantity,46(1), 221-236.doi:10.1007/s11135-010-9355-3 Claeys,A.S.,&Cauberghe,V.(2015).Theroleofafavorablepre-crisisreputationinprotecting organizationsduringcrises.PublicRelationsReview,41(1),64-71.doi: http://dx.doi.org/10.1016/j.pubrev.2014.10.013 Claeys,A.,Cauberghe,V.,&Vyncke,P.(2010).Restoringreputationsintimesofcrisis:An experimentalstudyofthesituationalcrisiscommunicationtheoryandthemoderating effectsoflocusofcontrol.PublicRelationsReview,36,256–262. doi:10.1016/j.pubrev.2010.05.004 Coombs,W.T.(2004).Impactofpastcrisesoncurrentcrisiscommunicationinsightsfrom SituationalCrisisCommunicationTheory.JournalofbusinessCommunication,41(3),265289.doi:10.1177/0021943604265607 Coombs,W.T.(2007a).Protectingorganizationreputationsduringacrisis:Thedevelopmentand applicationofsituationalcrisiscommunicationtheory.CorporateReputationReview,10(3), 163–176. Coombs,W.T.(2007b).AttributionTheoryasaguideforpost-crisiscommunicationresearch. PublicRelationsReview,33,135-139.doi:10.1016/j.pubrev.2006.11.016 Coombs,W.T.(2010).Parametersforcrisiscommunication.InW.T.Coombs&S.J.Holladay (Eds.),Thehandbookofcrisiscommunication(pp.17-53).Malden,MA:Blackwell. 66 Coombs,W.T.(2014).CrisisCommunication:ADevelopingField.InW.T.Coombs(SeriesEd.), CrisisCommunication:Vol.I(pp.3-18).London:Sage. Coombs,W.T.,&Holladay,S.J.(1996).Communicationandattributionsinacrisis:An experimentalstudyincrisiscommunication.JournalofPublicRelationsResearch,8(4),279295. Coombs,W.T.,&Holladay,S.J.(2001).Anextendedexaminationofthecrisissituations:Afusion oftherelationalmanagementandsymbolicapproaches.JournalofPublicRelationsResearch, 13(4),321-340. Coombs,W.T.,&Holladay,S.J.(2002).Helpingcrisismanagersprotectreputationalassetsinitial testsoftheSituationalCrisisCommunicationTheory.ManagementCommunication Quarterly,16(2),165-186.doi:10.1177/089331802237233 Coombs,W.T.&Holladay,S.J.(2005).AnExploratoryStudyofStakeholderEmotions:Affectand Crises.TheEffectofAffectinOrganizationalSettings,1,263-280.doi: http://dx.doi.org/10.1016/S1746-9791(05)01111-9 Coombs,W.T.&Holladay,S.J.(2006).Unpackingthehaloeffect:reputationandcrisis management.JournalofCommunicationManagement,10(2),123-137.doi: http://dx.doi.org/10.1108/13632540610664698 Coombs,W.T.,&Holladay,S.J.(2007).Thenegativecommunicationdynamic.Exploringthe impactofstakeholderaffectonbehavioralintentions.JournalofCommunication Management,11(4),300–312.doi:http://dx.doi.org/10.1108/13632540710843913 Coombs,W.T.&Holladay,S.J.(2008).Comparingapologytoequivalentcrisisresponsestrategies: Clarifyingapology’sroleandvalueincrisiscommunication.PublicRelationsReview,34,252257.doi:10.1016/j.pubrev.2008.04.001 Coombs,W.T.,Fediuk,T.A.&Holladay,S.J.(2007).FurtherExplorationsofPost-Crisis CommunicationandStakeholderAnger:TheNegativeCommunicationDynamicModel. PaperpresentedattheInternationalPublicRelationsResearchConference. Davies,G.,Chun,R.,DaSilva,R.V.,&Roper,S.(2003).CorporateReputationandCompetitiveness. NewYork,NY:Routledge. Dawar,N.&Pillutla,M.(2000).Impactofproduct-harmcrisesonbrandequity:Themoderating roleofconsumerexpectations.JournalofMarketingResearch,37,215-226. deVaus,D.D.(1996).SurveysinSocialResearch(4thedition).London:UCLPress. Dean,D.H.(2004).Consumerreactiontonegativepublicityeffectsofcorporatereputation, response,andresponsibilityforacrisisevent.JournalofBusinessCommunication,41(2), 192-211.doi:10.1177/0021943603261748 Desai,P.(2014).TheRoleOfPrintAdvertisingDuringProductRecallCrisis.InnovateJournalof BusinessManagement,2(1). Doorley,J.,&Garcia,H.(2010).Reputationmanagement:Thekeytosuccessfulpublicrelationsand corporatecommunication.NewYork,NY:Routledge. Dowling,G.(2001).Creatingcorporatereputations:Identity,image,andperformance.Oxford: OxfordUniversityPress. Druckman,J.N.(2001).Theimplicationsofframingeffectsforcitizencompetence.Political Behavior,23(3),225-256. 67 Du,S.,Bhattacharya,C.B.,&Sen,S.(2007).Reapingrelationalrewardsfromcorporatesocial responsibility:Theroleofcompetitivepositioning.Internationaljournalofresearchin marketing,24(3),224-241.doi:10.1016/j.ijresmar.2007.01.001 Dutta,S.&Pullig,C.(2011).Effectivenessofcorporateresponsestobrandcrises:Theroleofcrisis typeandresponsestrategies.JournalofBusinessResearch,64,1281–1287.doi: 10.1016/j.jbusres.2011.01.013 Dutton,J.E.,Dukerich,J.M.,&Harquail,C.V.(1994).Organizationalimagesandmember identification.AdministrativeScienceQuarterly,39(2),239-263. Edwards,K.,&Smith,E.E.(1996).Adisconfirmationbiasintheevaluationofarguments.Journal ofPersonalityandSocialPsychology,71(1),5-24.doi:10.1037//0022-3514.71.1.5 Einwiller,S.A.,Fedorikhin,A.,Johnson,A.R.,&Kamins,M.A.(2006).Enoughisenough!When identificationnolongerpreventsnegativecorporateassociations.JournaloftheAcademyof MarketingScience,34(2),185-194.doi:10.1177/0092070305284983 Etayankara,M.,&Bapuji,H.(2009,June).Productrecalls:areviewofliterature.Paperpresentedat AdministrativeSciencesAssociationofCanada(30),NiagaraFalls,Canada. Folkes,V.S.(1988).Recentattributionresearchinconsumerbehavior:Areviewandnew directions.JournalofConsumerResearch,14(4),548-565. Folkman,S.,&Moskowitz,J.T.(2000).Positiveaffectandtheothersideofcoping.American psychologist,55(6),647-654.doi:10.1037//0003-066X.55.6.647 Fombrun,C.(2015,October7).AboutVolkswagen,Reputation,andSocialResponsibility[Blog post].Retrieved26January2016,from http://blog.reputationinstitute.com/2015/10/07/about-volkswagen-reputation-andsocial-responsibility/ Fombrun,C.J.&vanRiel,C.B.M.(2004).Fame&Fortune:HowSuccessfulCompaniesBuildWinning Reputation.UpperSaddleRiver,NJ:PearsonEducation. Geier,B.(2015,September22).EverythingtoknowaboutVolkswagen'semissionscrisis.Fortune. Retrievedfromhttp://fortune.com/2015/09/22/volkswagen-vw-emissions-golf/ Gibson,D.,Gonzales,J.L.andCastanon,J.(2006).Theimportanceofreputationandtheroleof publicrelations,PublicRelationsQuarterly,51(3),15–18. Goyette,I.,Ricard,L.,Bergeron,J.&Marticottte,F.(2010).e-WOMScale:Word-of-Mouth MeasurementScalefore-ServicesContext*.CanadianJournalofAdministrativeSciences,27, 5-23.doi:10.1002/CJAS.129 Griffin,A.(2015,September26).Volkswagenscandal:Thereisawayoutofthecrisisanditstarts now.TheTelegraph.Retrievedfrom http://www.telegraph.co.uk/finance/comment/11893883/Volkswagen-scandal-There-isa-way-out-of-the-crisis-and-it-starts-now.html Gruen,R.J.,&Mendelsohn,G.(1986).Emotionalresponsestoaffectivedisplaysinothers:The distinctionbetweenempathyandsympathy.JournalofPersonalityandSocialPsychology, 51(3),609-614. Grunwald,G.&Hempelmann,B.(2011).ImpactsofReputationforQualityonPerceptionsof CompanyResponsibilityandProduct-relatedDangersintimesofProduct-recallandPublic ComplaintsCrises:ResultsfromanEmpiricalInvestigation.CorporateReputationReview, 13(4),264–283.doi:10.1057/crr.2010.23 68 Harrison-Walker,L.J.(2001).TheMeasurementofWord-of-MouthCommunicationandan InvestigationofServiceQualityandCustomerCommitmentasPotentialAntecedents. JournalofServiceResearch,4(1),60-75. Härtel,C.E.,McColl-Kennedy,J.R.,&McDonald,L.(1998).Incorporatingattributionaltheoryand thetheoryofreasonedactionwithinanaffectiveeventstheoryframeworktoproducea contingencypredictivemodelofconsumerreactionstoorganizationalmishaps.Advancesin ConsumerResearch,25,428-432. Haucap,J.,Wey,C.,&Barmbold,J.F.(1997).LocationChoiceasaSignalforProductQuality:The Economicsof"MadeinGermany".JournalofInstitutionalandTheoreticalEconomics (JITE)/ZeitschriftfürdiegesamteStaatswissenschaft,153(3),510-531. heiseonline.(2016,January27).Abgas-Skandal:VWbeginntmitRückruf–zunächstfürdas ModellAmarok.Retrievedfromhttp://www.heise.de/newsticker/meldung/Abgas-SkandalVW-beginnt-mit-Rueckruf-zunaechst-fuer-das-Modell-Amarok-3085837.html Hennig-Thurau,T.,Gwinner,K.P.,Walsh,G.&Gremler,D.D.(2004).ELECTRONICWORD-OFMOUTHVIACONSUMER-OPINIONPLATFORMS:WHATMOTIVATESCONSUMERSTO ARTICULATETHEMSELVESONTHEINTERNET?JournalofInteractiveMarketing,18(1),3852. Herr,P.M.,Kardes,F.R.,&Kim,J.(1991).Effectsofword-of-mouthandproduct-attribute informationonpersuasion:Anaccessibility-diagnosticityperspective.Journalofconsumer research,17(4),454-462. In’nami,Y.&Koizumi,R.(2013).StructuralEquationModelinginEducationalResearch:APrimer. InM.S.Khine(Ed.),Applicationofstructuralequationmodelingineducationalresearchand practice(pp.23-54).Rotterdam:SensePublishers. iwd.(2013,August8).GrüneProduktesindgefragt.Retrieved10Juni2016from https://www.iwd.de/artikel/gruene-produkte-sind-gefragt-120512/ Iyer,A.,&Oldmeadow,J.(2006).Picturethis:Emotionalandpoliticalresponsestophotographsof theKennethBigleykidnapping.EuropeanJournalofSocialPsychology,36(5),635-647.doi: 10.1002/ejsp.316 Jin,Y.(2009).Theeffectsofpublic’scognitiveappraisalofemotionsincrisesoncrisiscopingand strategyassessment.PublicRelationsReview,35,310–313.doi: 10.1016/j.pubrev.2009.02.003 Jin,Y.(2010).Makingsensesensiblyincrisiscommunication:Howpublics’crisisappraisals influencetheirnegativeemotions,copingstrategypreferences,andcrisisresponse acceptance.CommunicationResearch,37(4),522-552.doi:10.1177/0093650210368256 Jin,Y.(2014).Examiningpublics'crisisresponsesaccordingtodifferentshadesofangerand sympathy.JournalofPublicRelationsResearch,26(1),79-101.doi: 10.1080/1062726X.2013.848143 Jin,Y.,&Cameron,G.T.(2007).Theeffectsofthreattypeanddurationonpublicrelations practitioner’scognitive,affective,andconativeresponsesincrisissituations.Journalof PublicRelationsResearch,19,255–281. Jin,Y.,Pang,A.,&Cameron,G.T.(2007).Integratedcrisismapping:Towardsapublics-based, emotion-drivenconceptualizationincrisiscommunication.SpheraPublica,7(1),81-96. 69 Jin,Y.,Pang,A.&Cameron,G.T.(2012).TowardaPublics-Driven,Emotion-Based ConceptualizationinCrisisCommunication:UnearthingDominantEmotionsinMulti-Staged TestingoftheIntegratedCrisisMapping(ICM)Model.JournalofPublicRelationsResearch, 24,266–298.doi:10.1080/1062726X.2012.676747 JohnsonAvery,E.,WeaverLariscy,R.W.,Kim,S.&Hocke,T.(2010).Aquantitativereviewofcrisis communicationresearchinpublicrelationsfrom1991to2009.PublicRelationsReview,36, 190–192.doi:10.1016/j.pubrev.2010.01.001 Jorgensen,B.K.(1996).ComponentsofConsumerReactiontoCompany-RelatedMishaps:A StructuralEquationModelApproach.AdvancesinConsumerResearch,23(1),346-351. Keh,H.T.,&Xie,Y.(2009).Corporatereputationandcustomerbehavioralintentions:Therolesof trust,identificationandcommitment.IndustrialMarketingManagement,38(7),732-742. doi:10.1016/j.indmarman.2008.02.005 Kiambi,D.M.&Shafer,A.(2015).CorporateCrisisCommunication:ExaminingtheInterplayof ReputationandCrisisResponseStrategies,MassCommunicationandSociety,1-22.doi: 10.1080/15205436.2015.1066013 Kim,H.J.,&Cameron,G.T.(2011).Emotionsmatterincrisis:Theroleofangerandsadnessinthe publics'responsetocrisisnewsframingandcorporatecrisisresponse.Communication Research,38(6)826–855.doi:10.1177/0093650210385813 Kim,H.R.,Lee,M.,Lee,H.T.,&Kim,N.M.(2010).Corporatesocialresponsibilityandemployee– companyidentification.JournalofBusinessEthics,95(4),557-569.doi:0.1007A10551-0100440-2 Kline,R.B.(1998).Principlesandpracticeofstructuralequationmodeling.NewYork,NY:Guilford publications. Kollewe,J.(2015,December10).Volkswagenemissionsscandal–timeline.TheGuardian. Retrievedfromhttp://www.theguardian.com/business/2015/dec/10/volkswagenemissions-scandal-timeline-events Laczniak,R.,DeCarlo,T.,&Ramaswami,S.(2001).Consumers'responsestonegativeword-ofmouthcommunication:Anattributionstheoryperspective.JournalofConsumerPsychology, 11(1),57-73. Laufer,D.(2015).Emergingissuesincrisismanagement.BusinessHorizons,2(58),137-139.doi: 0.1007A10551-010-0440-2 Laufer,D.&Jung,J.M.(2010).Incorporatingregulatoryfocustheoryinproductrecall communicationstoincreasecompliancewithaproductrecall.PublicRelationsReview,36, 147–151.doi:10.1016/j.pubrev.2010.03.004 Laufer,D.,&Coombs,W.T.(2006).Howshouldacompanyrespondtoaproductharmcrisis?The roleofcorporatereputationandconsumer-basedcues.BusinessHorizons,49(5),379-385. doi:10.1016/j.bushor.2006.01.002 Lee,B.K.(2004).Audience-OrientedApproachtoCrisisCommunication:AStudyofHongKong Consumers’EvaluationofanOrganizationalCrisis.CommunicationResearch,31(5),600618.doi:10.1177/0093650204267936 Lee,W.N.,Yun,T.,&Lee,B.K.(2005).Theroleofinvolvementincountry-of-origineffectson productevaluation:Situationalandenduringinvolvement.JournalofInternational ConsumerMarketing,17(2-3),51-72.doi:10.1300/J046v17n02_04 70 Lerner,J.S.,&Tiedens,L.Z.(2006).Portraitoftheangrydecisionmaker:Howappraisal tendenciesshapeanger'sinfluenceoncognition.JournalofBehavioralDecisionMaking, 19(2),115-137.doi:10.1002/bdm.515 Lichtenstein,D.R.,Drumwright,M.E.,&Braig,B.M.(2004).Theeffectofcorporatesocial responsibilityoncustomerdonationstocorporate-supportednonprofits.Journalof marketing,68(4),16-32. Lin,C.P.,Chen,S.C.,Chiu,C.K.,&Lee,W.Y.(2011).Understandingpurchaseintentionduring product-harmcrises:Moderatingeffectsofperceivedcorporateabilityandcorporatesocial responsibility.JournalofBusinessEthics,102(3),455-471.doi:10.1007/sl0551-011-0824- Little,T.D.,Card,N.A.,Bovaird,J.A.,Preacher,K.J.&Crandall,C.S.(2007).Structuralequation modelingofmediationandmoderationwithcontextualfactors.InT.D.Little,J.A.Bovaird,& N.A.Card(Eds.),Modelingcontextualeffectsinlongitudinalstudies(pp.207-230).Mahwah, NJ:LawrenceErlbaumAssociates. Liu,B.F.,Austin,L.,&Jin,Y.(2011).Howpublicsrespondtocrisiscommunicationstrategies:The interplayofinformationformandsource.PublicRelationsReview,37(4),345-353.doi: 10.1016/j.pubrev.2011.08.004 Löhr,J.(2015,October20).Two-thirdsofGermansstilltrustVolkswagenafteremissionsscandal. TheGuardian.Retrievedfromhttp://www.theguardian.com/business/2015/oct/20/twothirds-of-germans-still-trust-volkswagen-after-emissions-scandal Love,E.G.,&Kraatz,M.(2009).Character,conformity,orthebottomline?Howandwhy downsizingaffectedcorporatereputation.AcademyofManagementJournal,52,314-335. Lyon,L.,&Cameron,G.T.(2004).Arelationalapproachexaminingtheinterplayofprior reputationandimmediateresponsetoacrisis.JournalofPublicRelationsResearch,16(3), 213-241.doi:10.1207/s1532754xjprr1603_1 Mael,F.,&Ashforth,B.E.(1992).Alumniandtheiralmamater:apartialtestofthereformulated modeloforganizationalidentification.JournalofOrganizationalBehavior,13(2),103−123. Marin,L.,&Ruiz,S.(2007).“Ineedyoutoo!”Corporateidentityattractivenessforconsumersand theroleofsocialresponsibility.JournalofBusinessEthics,71(3),245-260.doi: 10.1007/s10551-006-9137-y McDonald,L.,&Härtel,C.E.(2000).Applyingtheinvolvementconstructtoorganisationalcrises. ProceedingsoftheAustralianandNewZealandMarketingAcademyConference,GoldCoast, Australia. McDonald,L.,Glendon,A.I.,&Sparks,B.(2011).MeasuringConsumers’EmotionalReactionsto CompanyCrises:ScaleDevelopmentandImplications.AdvancesinConsumerResearch,39, 333-340. McDonald,L.M.,Sparks,B.&Glendon,A.I.(2010).Stakeholderreactionstocompanycrisis communicationandcauses.PublicRelationsReview,36,263–271.doi: 10.1016/j.pubrev.2010.04.004 McQuarrie,E.&Munson,J.(1992).Arevisedproductinvolvementinventory-improvedusability andvalidity.AdvancesinConsumerResearch,19,108-115. Mizerski,R.W.(1982).Anattributionexplanationofthedisproportionateinfluenceof unfavorableinformation.JournalofConsumerResearch,9(3),301-310. 71 Möhring,W.&Schlütz,D.(2010).DieBefragunginderMedien-undKommunikations-wissenschaft: EinepraxisorientierteEinführung(2ndedition).Wiesbaden:SpringerVS. Murphy,S.(2015,Sept.29).VWscandal:HowitcomparestoToyota'srecallcrisis.YouGov. Retrievedfromhttps://yougov.co.uk/news/2015/09/29/volkswagen-crisis-lessonshistory/ Nachtigall,C.,Kroehne,U.,Funke,F.,&Steyer,R.(2003).(Why)ShouldWeUseSEM?Prosand ConsofStructuralEquationModeling.MethodsofPsychologicalResearchOnline,8(2),1-22. Neate,R.(2016,March2).VWCEOwastoldaboutemissionscrisisayearbeforeadmittingto cheatscandal.TheGuardian.Retrievedfrom https://www.theguardian.com/business/2016/mar/02/vw-ceo-martin-winterkorn-toldabout-emissions-scandalhttps://www.theguardian.com/business/2016/mar/02/vw-ceomartin-winterkorn-told-about-emissions-scandal Neuman,W.L.(2014).SurveyResearch.InW.L.Neuman(Ed.),SocialResearchMethods: QualitativeandQuantitativeApproaches(7thedition)(pp.201-244).NewYork,NY:Pearson EducationLimited. Pawar,B.S.(2009).Theorybuildingforhypothesisspecificationinorganizationalstudies.New Delhi:SagePublicationsIndia. Pérez,R.C.(2009).Effectsofperceivedidentitybasedoncorporatesocialresponsibility:therole ofconsumeridentificationwiththecompany.CorporateReputationReview,12(2),177-191. Perloff,R.M.(2010).Thedynamicsofpersuasion:communicationandattitudesinthetwenty-first century.NewYork,NY:Routledge. Petty,R.E.&Cacioppo,J.T.(1981).AttitudesandPersuasion:ClassicandContemporaryApproaches. Dubuque,Iowa:Wm.C.BrownCompanyPublishers. Petty,R.E.&Cacioppo,J.T.(1986).CommunicationandPersuasion.CentralandPeripheralRoutes toAttitudeChange.NewYork,NY:SpringerVerlag. Podolny,J.M.,&Phillips,D.J.(1996).Thedynamicsoforganizationalstatus.Industrialand CorporateChange,5(2),453-471. Punch,K.F.(2014).Introductiontosocialresearch:Quantitativeandqualitativeapproaches(3rd edition).London:Sage. Raykov,T.,&Marcoulides,G.A.(2000).Afirstcourseinstructuralequationmodeling.NewYork, NY:Routledge. Reichart,J.(2003).Atheoreticalexplorationofexpectationalgapsinthecorporateissue construct.CorporateReputationReview,6(1),58-69.doi:10.1057/palgrave.crr.1540190 Reidel,M.(2015,August4).VWundBMWhabenbeidenDeutschendasbesteAnsehen.Horizont. Retrievedfromhttp://www.horizont.net/marketing/nachrichten/Reputation-VW-undBMW-haben-bei-den-Deutschen-das-beste-Ansehen-135685 ReputationInstitute.(2013).TheGlobalRepTrak®100:TheWorld’sMostReputableCompanies (2013).RIReportonConsumerPerceptionsofCompaniesin15Countries[Report]. Retrievedfromhttp://www.reputationinstitute.com/Resources/Registered/PDFResources/The-2013-Global-RepTrak®-100-Results-and-Report.aspx 72 ReputationInstitute.(2014).TheGlobalRepTrak®100:TheWorld’sMostReputableCompanies (2014).RIReportonConsumerPerceptionsofCompaniesin15Countries[Report]. Retrievedfromhttp://www.reputationinstitute.com/Resources/Registered/PDFResources/2014-Global-RepTrak-100.aspx ReputationInstitute.(2015).TheGlobalRepTrak®100:TheWorld’sMostReputableCompanies (2015).RIReportonConsumerPerceptionsofCompaniesin15Countries[Report]. Retrievedfromhttps://www.reputationinstitute.com/Resources/Registered/PDFResources/Global-RepTrak-100-2015.aspx Rhee,M.,&Haunschild,P.R.(2006).Theliabilityofgoodreputation:Astudyofproductrecallsin theUSautomobileindustry.OrganizationScience,17(1),101-117.doi:10.1287/orsc. 1050.0175 Richins,M.L.(1984).WORDOFMOUTHCOMMUNICATIONANEGATIVEINFORMATION. Advancesinconsumerresearch,11(1),697-702. Rousseau,D.M.(2006).Istheresuchathingas“evidence-basedmanagement”?.Academyof managementreview,31(2),256-269. Scholl,A.(2009).DieBefragung(2ndedition).Konstanz:UVK-Verl.-Ges. Schultz,F.,Utz,S.&Göritz,A.(2011).Isthemediumthemessage?Perceptionsofandreactionsto crisiscommunicationviatwitter,blogsandtraditionalmedia.PublicRelationsReview,37, 20-27.doi:10.1016/j.pubrev.2010.12.001 Schumacker,R.E.,&Lomax,R.G.(2004).Abeginner'sguidetostructuralequationmodeling. Mahwah,NJ:LawrenceErlbaumAssociates. Sen,S.&Bhattacharya,C.B.(2001).DoesDoingGoodAlwaysLeadtoDoingBetter?Consumer ReactionstoCorporateSocialResponsibility.JournalofMarketingResearch38(2),225-243. Silverman,G.(2001).Thepowerofwordofmouth.DirectMarketing,64(5),47-52. Sohn,Y.J.,&Lariscy,R.W.(2015).A“Buffer”or“Boomerang?"—TheRoleofCorporateReputation inBadTimes.CommunicationResearch,42(2),237-259.doi:10.1177/0093650212466891 Stakeholder.(n.d.).InBusinessDictionary.Retrievedfrom http://www.businessdictionary.com/definition/stakeholder.html Statista.(2016).AnteilderInternetnutzerinDeutschlandindenJahren2001bis2015.Retrieved fromhttp://de.statista.com/statistik/daten/studie/13070/umfrage/entwicklung-derinternetnutzung-in-deutschland-seit-2001/ Stockmyer,J.(1996).Brandsincrisis:Consumerhelpfordeservingvictims.AdvancesinConsumer Research,23(23),429-435. Tajfel,H.(1974).Socialidentityandintergroupbehaviour.SocialScienceInformation,13(2),6593. Tajfel,H.,&Turner,J.C.(1985).Thesocialidentitytheoryofintergroupbehavior.InS.Worchel& W.G.Austin(Eds.),Psychologyofintergrouprelations(2nded.,pp.7-24).Chicago:Burnham IncPub. Terry,D.,Hogg,M.andWhite,K.(2000)‘Attitude-behaviourrelations:Socialidentityandgroup membership’,inD.TerryandM.Hogg(eds.),Attitudes,BehaviourandSocialContext:The RoleOfNormsandGroupMembership,LawrenceErlbaumAssociates,Mahwah,NJ,USA,pp. 67–95. 73 TheGroup.(2014).RetrievedfromVolkswagenAG http://www.volkswagenag.com/content/vwcorp/content/en/the_group.html Theo,T.,TingTsai,L.&Yang,C-C.(2013).ApplyingStructuralEquationModeling(SEM)in EducationalResearch:AnIntroduction.InM.S.Khine(Ed.),Applicationofstructural equationmodelingineducationalresearchandpractice(pp.3-22).Rotterdam:Sense Publishers. Tucker,L.,&Melewar,T.C.(2005).Corporatereputationandcrisismanagement:Thethreatand manageabilityofanti-corporatism.CorporateReputationReview,7(4),377-387. Turk,J.V.,Jin,Y.,Stewart,S.,Kim,J.,&Hipple,J.R.(2012).Examiningtheinterplayofan organization'spriorreputation,CEO'svisibility,andimmediateresponsetoacrisis.Public RelationsReview,38(4),574-583.doi:10.1016/j.pubrev.2012.06.012 Ulmer,R.R.(2001).Effectivecrisismanagementthroughestablishedstakeholderrelationships MaldenMillsasacasestudy.ManagementCommunicationQuarterly,14(4),590-615. Underwood,R.,Bond,E.,&Baer,R.(2001).Buildingservicebrandsviasocialidentity:Lessons fromthesportsmarketplace.JournalofMarketingTheoryandPractice,9(1),1-13. Utz,S.,Schultz,F.,&Glocka,S.(2013).Crisiscommunicationonline:Howmedium,crisistypeand emotionsaffectedpublicreactionsintheFukushimaDaiichinucleardisaster.Public RelationsReview,39(1),40-46.doi:10.1016/j.pubrev.2012.09.010 vanRiel,C.B.M.&Fombrun,C.J.(2007).EssentialsofCorporateCommunication.NewYork,NY: Routledge. Vizard,S.(2015).WhyVolkswagencannotsurvivetheemissionsscandalunscathed.Marketing Week.Retrievedfromhttps://www.marketingweek.com/2015/09/24/why-volkswagencannot-survive-the-emissions-scandal-unscathed/ VWglobalsalesfell2%inyearemissionsscandalhit.(2016,January8).TheGuardian.Retrieved fromhttp://www.theguardian.com/business/2016/jan/08/vw-global-sales-fell-yearemissions-scandal-2015 Walker,K.(2010).Asystematicreviewofthecorporatereputationliterature:Definition, measurement,andtheory.CorporateReputationReview,12(4),357-387. Walliman,N.(2006).SocialResearchMethods.London:Sage. Weiber,R.&Mühlhaus,D.(2014).Strukturgleichungsmodellierung.Berlin,Heidelberg:Springer Gabler. Weiner,B,(1985).AnAttributionalTheoryofAchievementMotivationandEmotion.Psychological Review,92(4),548-573. Weiner,B.(1986).AnAttributionalTheoryofMotivationandEmotion.NewYork,NY:Springer. Weiner,B.(2006)SocialMotivation,Justice,andtheMoralEmotions:AnAttributionalApproach. Mahwah,NJ:LawrenceErlbaumAssociates. Weingartner,M.(2015,November11).AuffälligeAbgaswertebeimehrerenHerstellern. FrankfurterAllgemeineZeitung.Retrievedfromhttp://www.faz.net/aktuell/wirtschaft/vwabgasskandal/kraftfahrtbundesamt-stellt-auffaellige-abgaswerte-bei-mehreren-herstellernfest-13907055.html 74 Weiss,H.M.,&Cropanzano,R.(1996).Affectiveeventstheory:Atheoreticaldiscussionofthe structure,causesandconsequencesofaffectiveexperiencesatwork.Researchin OrganizationalBehaviour,18,1-74. West,S.G.,&Finch,J.F.&Curran.P.J.(1995).Structuralequationmodelswithnonnormal variables:problemsandremedies.Structuralequationmodeling:Concepts,issues,and applications,56-75. Wetzer,I.M.,Zeelenberg,M.,&Pieters,R.(2007).“Nevereatinthatrestaurant,Idid!”:Exploring whypeopleengageinnegativeword-of-mouthcommunication.Psychology&Marketing, 24(8),661-680.doi:10.1002/mar.20178 Wigley,S.&Pfau,M.(2010).CommunicatingBeforeaCrisis:AnExplorationofBolstering,CSR, andInoculationPractices.InW.T.Coombs&S.J.Holladay(Eds.),Thehandbookofcrisis communication(pp.607-634).Malden,MA:Blackwell. Winton,N.(2015,October19).VolkswagenSalesStartToFeelTheImpactOfDieselScandal. Forbes.Retrievedfrom http://www.forbes.com/sites/neilwinton/2015/10/19/volkswagen-sales-start-to-feel-theimpact-of-diesel-scandal/#4856a9575e91 Woodyard,C.(2015,November20).Chronology:HowVW'semissionsscandalhasmushroomed. USAToday.Retrievedfrom http://www.usatoday.com/story/money/cars/2015/11/20/vw-volkswagen-chronologyemissions/76122812/ Xiao,N.,&Hwan(Mark)Lee,S.(2014).Brandidentityfitinco-branding:Themoderatingroleof CBidentificationandconsumercoping.EuropeanJournalofMarketing,48(7/8),1239-1254. doi:10.1108/EJM-02-2012-0075 Yun,T.W.,Lee,W.-N.,&Sego,T.(2002).DirectandIndirectUseofCountryofOriginCuesfor HybridandNon-hybridProducts.AdvancesinInternationalMarketing,12,195-214. Zaichkowsky,J.L.(1985).MeasuringtheInvolvementConstruct.JournalofConsumerResearch,12, 341-352. Zhao,X.,Lynch,J.G.,&Chen,Q.(2010).ReconsideringBaronandKenny:Mythsandtruthsabout mediationanalysis.Journalofconsumerresearch,37(2),197-206.doi:10.1086/651257 Zhou,S.&Sloan,WM.D.(2011).ResearchMethodsinCommunication(2nd.edition).Northport: VisionVPress. Zhukov,D.V.,Bhuiyan,M.A.,&Ullah,A.(2015).Utilizationofthecountryoforigineffectin product-harmcrisismanagement:anoverviewofliteratureandaconceptualmodel proposition.InternationalJournalofManagementScienceandBusinessAdministration,1(2), 54-70. 75 Appendices AppendixA-Questionnaires AppendixA1-QuestionnaireGerman SehrgeehrteTeilnehmerin,sehrgeehrterTeilnehmer, vielenDank,dassSiesichdazubereiterklärthaben,anmeinerUmfrageüberdenVolkswagen Konzern(imFolgenden:VWKonzern)teilzunehmen.DerVWKonzernumfasstunteranderem dieAutomarkenVW,Audi,Seat,Skoda,LamborghiniundPorsche. DieUmfrageistTeilmeinerMasterarbeitanderErasmusUniversitätRotterdam.Demnachhatdie UmfragekeinerleikommerziellesInteresseunddieErgebnissedienenausschließlich wissenschaftlichenZwecken.DerFragebogendauertnuretwa10Minuten.BeidenFragengibt eskeinerichtigenoderfalschenAntworten–esgehtalleinumIhrepersönlicheMeinungund Einstellungengegenüber,sowieErfahrungenmitdemVWKonzern.AlleDatenwerdennatürlich strengvertraulichbehandeltundanonymisiertausgewertet.SiewürdenmirmitIhrer Teilnahmesehrhelfen. SolltenSieFragenhabenoderandenErgebnissenderStudieinteressiertsein,könnenSiemich [email protected]. VielenDankimVorausfürIhreTeilnahme. LouisaWanjek ErasmusUniversitätRotterdam [email protected] ___ 1. ZuallererstwürdeichgernevonIhnenwissen,obSiejemalsvomVWKonzerngehört haben? □ja □nein 2. MenschenkönnenganzunterschiedlicheMeinungengegenüberdemVWKonzern haben.WieistesbeiIhnen,wiesehrstimmenSiedenfolgendenAussagenzu? BitteordnenSiesichfürjedeAussageaufderSkalazwischen1bis7ein,wobei1„stimme überhauptnichtzu“und7„stimmevollundganzzu“bedeuten.MitdenZifferndazwischen könnenSieIhreMeinungabstufen. IchidentifizieremichstarkmitdemVWKonzern,wennichmit 1□□□□□□□□□□7 anderendarüberspreche. IchbevorzugeAutosvomVWKonzern,wennichsiemitdenenvon 1□□□□□□□□□□7 anderenAutomobilherstellernvergleiche. IchstehedemUnternehmensimagevomVWKonzernpositiv 1□□□□□□□□□□7 gegenüber. 76 WennjemanddenVWKonzernkritisiert,fühltessichfürmichwie 1□□□□□□□□□□7 einepersönlicheBeleidigungan. Ichbinsehrdaraninteressiert,wasandereüberdenVWKonzern 1□□□□□□□□□□7 denken. ErfolgedesVWKonzernsfühlensichanwiemeineeigenen 1□□□□□□□□□□7 Erfolge. WennjemanddenVWKonzernlobt,empfindeichesals 1□□□□□□□□□□7 persönlichesKompliment. WennderVWKonzernindenMedienkritisiertwird,istesmir 1□□□□□□□□□□7 peinlich. WieSievielleichtmitbekommenhaben,hatderVWKonzernzugegeben,absichtlichMotorenvon Diesel-Fahrzeugenmanipuliertzuhaben,umdieerlaubtenHöchstwertefürEmissionenin Prüfungssituationeneinzuhalten.SeitSeptember2015sindweltweitetwa11MillionenAutosder MarkenVW,Audi,Seat,SkodaundPorschebetroffen.DieserVorfallwirdoftalsder„VWAbgasskandal“bezeichnet. DiefolgendenFragenbeziehensichaufIhreMeinungüberdenVWKonzernnachdemdie InformationenüberdenAbgasskandalöffentlichwurden. 3. HabenSiejemalsvomVW-Abgasskandalgehört? □ja □nein 4. BesitzenSiederzeiteinAutoderMarkenVW,Audi,Seat,SkodaoderPorsche?(Filter) □ja □nein 5. IstIhrAuto/mindestenseinsIhrerAutosvomAbgasskandalbetroffen?Mitbetroffen istgemeint,dasseszudenAutomodellenderMarkenVW,Audi,Seat,Skodaoder Porschegehört,dievomVWKonzernindieWerkstattzurückgerufenwurden. □ja □nein □Weißnicht 6. DiefolgendenAussagenbetreffenIhrenEindruckvomVWKonzernunddem Abgasskandal.WiesehrstimmenSiediesenAussagenzu? BitteordnenSiesichfürjedeAussageaufderSkalazwischen1bis7ein,wobei1„stimme überhauptnichtzu“und7„stimmevollundganzzu“bedeuten.MitdenZifferndazwischen könnenSieIhreMeinungabstufen. BezüglichdesAbgasskandalsistderVWKonzernbesorgtumdas 1□□□□□□□□□□7 WohlseinerAnspruchsgruppen(z.B.Kunden,Mitarbeiter, Investoren). DerVWKonzernistinBezugaufdenAbgasskandalimGrunde 1□□□□□□□□□□7 unehrlich. 77 IchtrauedemVWKonzernnichtzu,dieWahrheitüberden 1□□□□□□□□□□7 Abgasskandalzuerzählen. IchwürdeunterdenmeistenUmständenwahrscheinlichglauben, 1□□□□□□□□□□7 wasderVWKonzernüberdenAbgasskandalsagt. BezüglichdesAbgasskandalsistderVWKonzernnichtbesorgtum 1□□□□□□□□□□7 dasWohlseinerAnspruchsgruppen(z.B.Kunden,Mitarbeiter, Investoren). 7. InwiefernstimmenSiedenfolgendenAussagenzu?BittedenkenSieauchhieranden VWKonzernimZusammenhangmitdemAbgasskandal. BitteordnenSiesichfürjedeAussageaufderSkalazwischen1bis7ein,wobei1„stimme überhauptnichtzu“und7„stimmevollundganzzu“bedeuten.MitdenZifferndazwischen könnenSieIhreMeinungabstufen. IchwürdeFreundeoderVerwandteermutigen,aufgrunddes 1□□□□□□□□□□7 AbgasskandalskeineAutosvomVWKonzernzukaufen. WegendesAbgasskandalswürdeichzuanderenLeutennegative 1□□□□□□□□□□7 DingeüberdenVWKonzernunddessenAutossagen. Ichwürdejemandem,dernachmeinemRatfragt,auchnachdem 1□□□□□□□□□□7 AbgasskandalAutosvomVWKonzernempfehlen. 8. WennSieandenAbgasskandaldenken,wasempfindenSiegegenüberdemVW Konzern? BitteordnenSieIhrEmpfindengegenüberdemVWKonzernmitdenfolgendenAdjektivenauf derSkalazwischen1und7ein.Dabeibedeutet1„überhauptnicht“und7„sehr“.MitdenZiffern dazwischenkönnenSieIhrEmpfindenabstufen. WennichandenVWKonzernunddenAbgasskandaldenke,binich... wütend überhauptnicht 1□□□□□□□□□□7 sehr verärgert überhauptnicht 1□□□□□□□□□□7 sehr angewidert überhauptnicht 1□□□□□□□□□□7 sehr empört überhauptnicht 1□□□□□□□□□□7 sehr WennichandenVWKonzernunddenAbgasskandaldenke,empfindeich... Verständnis überhauptnicht 1□□□□□□□□□□7 sehr Mitleid überhauptnicht 1□□□□□□□□□□7 sehr Mitgefühl überhauptnicht 1□□□□□□□□□□7 sehr Sympathie überhauptnicht 1□□□□□□□□□□7 sehr 9. WasdenkenSieüberdenAbgasskandalselbst? BittebewertenSieIhreEinstellunggegenüberdemAbgasskandalmitdenfolgendenAussagen. 78 DerAbgasskandal(ist)… unwichtig □□□□□□□□□□ wichtig nichtbesorgniserregend □□□□□□□□□□ besorgniserregend bedeutetmirnichts □□□□□□□□□□ bedeutetmirviel spieltkeineRollefürmich □□□□□□□□□□ spielteineRollefürmich irrelevant □□□□□□□□□□ relevant 10. WiesehrstimmenSiedenfolgendenAussagenzu? BitteordnenSiesichfürjedeAussageaufderSkalazwischen1bis7ein,wobei1„stimme überhauptnichtzu“und7„stimmevollundganzzu“bedeuten.MitdenZifferndazwischen könnenSieIhreMeinungabstufen. DerAuslöserfürdenAbgasskandalwareinevorsätzliche 1□□□□□□□□□□7 HandlungvonjemandemimKonzern. JemandimKonzernhatdieUrsachefürdenAbgasskandal 1□□□□□□□□□□7 wissentlichherbeigeführt. DerKonzernhattedieFähigkeit,dasAuftretendesAbgasskandals 1□□□□□□□□□□7 zustoppen. DerAbgasskandalwarvomKonzernvermeidbar. 1□□□□□□□□□□7 DerKonzernhättedenAbgasskandalvermeidenkönnen. 1□□□□□□□□□□7 DerKonzernsolltefürdenAbgasskandalzurVerantwortung 1□□□□□□□□□□7 gezogenwerden. DerAbgasskandalwurdedurcheineSchwächeinderOrganisation 1□□□□□□□□□□7 verursacht. InterneorganisatorischeProblemehabenzumAbgasskandal 1□□□□□□□□□□7 beigetragen. 11. ZumAbschlussmöchteichSiebittennocheinpaarAngabenzuIhrerPersonzumachen. 11.1. BittegebenSieihrGeschlechtan: □weiblich □männlich 11.2. WiealtsindSie?BittegebenSieIhrAlterinJahrenan: ___Jahre 11.3. WasistIhrhöchsterBildungsabschluss? □(noch)keinAbschluss □Hauptschulabschluss(Volksschulabschluss) □Realschulabschluss(MittlereReife) □Abitur/(Fach-)Hochschulreife □(Fach-)Hochschulabschluss □Andere,undzwar:______________________ 79 11.4. WasistIhreNationalität? □Deutsch □Andere,undzwar:______________________ DamitsindSienunamEndederBefragungangekommen. FallsSienochAnmerkungenoderKritikhaben,könnenSiegernenochfolgendesFeldausfüllen. Ende VielenDanknocheinmalfürIhreTeilnahmeundIhrerUnterstützungbeimeinerAbschlussarbeit! Ichwürdemichfreuen,wennSiedenuntenstehendenLinkzumeinerUmfragenochanIhre Familie,Freunde,BekannteoderKollegenweiterleitenwürden.JemehrPersonenanmeiner Umfrageteilnehmen,destoaussagekräftigersinddieErgebnissemeinerStudie. https://erasmushcc.qualtrics.com/SE/?SID=SV_a9MJdOahXuuae4B SolltenSieFragenzumeinerStudiehabenoderandenErgebnisseninteressiertsein,könnenSie michgerneunter437075lw@student.eur.nlkontaktieren.EineZuordnungIhrerE-Mail-Adresse mitdenAngabenimFragebogenistnichtmöglich. BesteGrüße LouisaWanjek 80 AppendixA2-QuestionnaireEnglish Dearparticipant, thankyouverymuchfortakingpartinthissurveyabouttheVolkswagenGroup(following:VW Group). The VW Group comprises among others the brands VW, Audi, Seat, Skoda, Lamborghini andPorsche. ThesurveyispartofmyMastersThesisattheSchoolofHistory,CultureandCommunicationof the Erasmus University Rotterdam. Thus, this survey does not have any commercial interest and the results are only used for scientific purposes. The questionnaire will take about 10 minutes. There are no right and wrong answers - I am simply interested in your attitudes and opinionstowardstheVWGroup.Allofyouranswerswillbecompletelyanonymousandtreated confidentially.Yourparticipationwouldhelpmeverymuch. Ifyouhaveanyquestionsorifyouareinterestedintheresultsofmystudy,pleasedonothesitate tocontactme([email protected]). Thankyouinadvanceforyourparticipation. LouisaWanjek ErasmusUniversityRotterdam [email protected] --- 1. Firstofall,IwouldliketoknowifyouhaveeverheardoftheVWGroup? □yes □no 2. PersonscanhaveverydifferentopinionsabouttheVWGroup.Howaboutyou,how muchdoyouagreewiththefollowingstatements? Pleaseratehowmuchyouagreewiththestatementsonascalefrom1to7,whereas1means thatyou“stronglydisagree”and7meansthatyou“stronglyagree”.Withthenumbersin between,youcangraduateyouropinion. I have strong identification with the VW Group when talking to 1□□□□□□□□□□7 othersaboutit. IprefercarsoftheVWGroupwhencomparingitwiththatofother 1□□□□□□□□□□7 automobilemanufacturers. IampositiveaboutthecompanyimageoftheVWGroup. 1□□□□□□□□□□7 WhensomeonecriticizestheVWGroup,itfeelslikeapersonal 1□□□□□□□□□□7 insult. IamveryinterestedinwhatothersthinkaboutVWGroup. 1□□□□□□□□□□7 ThesuccessesoftheVWGrouparemysuccesses. 1□□□□□□□□□□7 WhensomeonepraisestheVWGroup,itfeelslikeapersonal 1□□□□□□□□□□7 compliment. 81 IfastoryinthemediacriticizestheVWGroup,Ifeelembarrassed. 1□□□□□□□□□□7 Asyoumayhaveheard,theVWGrouphasadmittedtohaveintentionallymanipulatedenginesof dieselcarstoincreasetheirperformanceonemission,whenbeingtested.SinceSeptember2015, about11millioncarsofthebrandsVW,Audi,Seat,SkodaandPorschehavebeenaffected.This incidentisoftenreferredtoasthe“emissionsscandal”. ThefollowingquestionswillaskyouropinionabouttheVWGroupaftertheinformationabout this“emissionsscandal”hasbeenrevealed. 3. HaveyoueverheardoftheVWemissionsscandal? □yes□no 4. DoyoucurrentlyownacarofthebrandsVW,Audi,Seat,SkodaorPorsche? □yes□no 5. Hasyourcar/atleastoneofyourcarsbeenaffectedbythe“emissionsscandal”? WithaffecteditismeantthatyourcarbelongstothoseofthebrandsVW,Audi,Seat, SkodaorPorschethathavebeenrecalledbyVW. □yes□no □don’tknow 6. TheitemsbelowconcernyourimpressionoftheVWGroupandthe“emissions scandal”.Howmuchdoyouagreeordisagreewiththesestatements? Pleaseratehowmuchyouagreewiththestatementsonascalefrom1to7,whereas1means thatyou“stronglydisagree”and7meansthatyou“stronglyagree”.Withthenumbersin between,youcangraduateyouropinion. Regardingtheemissionsscandal,theVWGroupisconcernedwith 1□□□□□□□□□□7 thewell-beingofitspublics(e.g.customers,employees,investors). TheVWGroupisbasicallydishonestconcerningtheemissions 1□□□□□□□□□□7 scandal. I do not trust the VW Group to tell the truth about the emissions 1□□□□□□□□□□7 scandal. Undermostcircumstances,Iwouldbelikelytobelievewhatthe 1□□□□□□□□□□7 VWGroupsaysabouttheemissionsscandal. Regardingtheemissionsscandal,theVWGroupisnotconcerned 1□□□□□□□□□□7 withthewell-beingofitspublics(e.g.customers,employees, investors). 82 7. Towhatdegreedoyouagreewiththefollowingitems?PleasethinkagainoftheVW Groupinthecontextoftheemissionsscandal. Pleaserateyouragreementwiththestatementsonascalefrom1to7,whereas1meansthat you“stronglydisagree”and7meansthatyou“stronglyagree”.Withthenumbersinbetween, youcangraduateyouropinion. IwouldencouragefriendsorrelativesnottobuycarsfromtheVW 1□□□□□□□□□□7 Groupbecauseoftheemissionsscandal. Becauseoftheemissionsscandal,Iwouldsaynegativethings 1□□□□□□□□□□7 abouttheVWGroupanditscarstootherpeople. Even after the emissions scandal, I would recommend cars of the 1□□□□□□□□□□7 VWGrouptosomeonewhoaskedmyadvice. 8. HowdoyoufeelabouttheVWGroupduetothe“emissionsscandal”? Foreachadjectivebelow,pleaserateyourfeelingstowardstheVWGrouponascalefrom1 to7,whereas1means“notatall”and7means“verymuch”.Withthenumbersinbetween, youcangraduateyourfeelings. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... angry notatall 1□□□□□□□□□□7 verymuch annoyed notatall 1□□□□□□□□□□7 verymuch disgusted notatall 1□□□□□□□□□□7 verymuch outraged notatall 1□□□□□□□□□□7 verymuch WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... sympathetic notatall 1□□□□□□□□□□7 verymuch sorry notatall 1□□□□□□□□□□7 verymuch compassion notatall 1□□□□□□□□□□7 verymuch empathy notatall 1□□□□□□□□□□7 verymuch 9. Whatdoyouthinkoftheemissionsscandalitself? Pleaserateyourattitudetowardstheemissionsscandalwiththefollowingitems. Theemissionsscandalis... unimportant □□□□□□□□□□ important ofnoconcern □□□□□□□□□□ ofconcern meansnothing □□□□□□□□□□ meansalot doesnotmatter □□□□□□□□□□ matterstome irrelevant □□□□□□□□□□ relevant 10. Howmuchdoyouagreeordisagreewiththefollowingitems? Pleaseratehowmuchyouagreewiththestatementsonascalefrom1to7,whereas1means thatyou“stronglydisagree”and7meansthatyou“stronglyagree”.Withthenumbersin between,youcangraduateyouropinion. 83 Thecauseoftheemissionsscandalwasanintentionalactby 1□□□□□□□□□□7 someoneintheorganization. Someoneintheorganizationknowinglycreatedthecauseofthe 1□□□□□□□□□□7 emissionsscandal. Theorganizationhadthecapabilitytostoptheemissionsscandal 1□□□□□□□□□□7 fromoccurring. Theemissionsscandalwaspreventablebytheorganization. 1□□□□□□□□□□7 Theorganizationcouldhaveavoidedtheemissionsscandal. 1□□□□□□□□□□7 The organization should be held accountable for the emissions 1□□□□□□□□□□7 scandal. The emissions scandal was caused by a weakness in the 1□□□□□□□□□□7 organization. Internal organizational issues contributed to the emissions 1□□□□□□□□□□7 scandal. 11. Finally,Iwouldliketoaskyoutoprovidesomegeneralinformationaboutyourself. 11.1. Pleaseindicateyourgender: □Female □Male 11.2. Howoldareyou?Please,indicateyourageinyears: ___years 1.1. Whatisyourhighestlevelofeducationachieved? □IhavenotyetcompletedHighSchool □HighSchoolDiploma(lowest) □HighSchoolDiploma(middle) □HighSchoolDiploma(highest) □UniversityDegree □Other:______________________ 1.2. WhatisyourNationality? □German □Other:______________________ Feedback Youhavereachedtheendofthesurvey. Ifyouhaveanyfurthercommentsorsuggestionsonthequestionnaire,pleaseletmeknowby fillinginthefollowingfield. 84 TheEnd ThankyouagainforyourparticipationandsupportingmyMaster’sthesis!Iwouldbegladifyou sentthefollowinglinkofthesurveytoyourfamily,friendsorcolleagues.Themorepeople participateinmysurvey,themoreinformativewillbetheresultsofmystudy. https://erasmushcc.qualtrics.com/SE/?SID=SV_a9MJdOahXuuae4B Ifyouhaveanyquestionsaboutmystudyorareinterestedintheresults,feelfreetocontactme via437075lw@student.eur.nl.Anassociationofyoure-mailaddresstoyourstatementsinthe questionnaireisnotpossible. Kindregards, LouisaWanjek Contact:[email protected] 85 AppendixB–OverviewofItems ItemName Operationalization Reputation1 Regardingtheemissionsscandal,theVWGroupisconcerned withthewell-beingofitspublics(e.g.customers,employees, investors). Reputation2 TheVWGroupisbasicallydishonestconcerningtheemissions scandal. Reputation3 IdonottrusttheVWGrouptotellthetruthabouttheemissions scandal. Reputation4 Undermostcircumstances,Iwouldbelikelytobelievewhatthe VWGroupsaysabouttheemissionsscandal. Reputation5 Regardingtheemissionsscandal,theVWGroupisnot concernedwiththewell-beingofitspublics(e.g.customers, employees,investors). NWOM1 Iwouldencouragefriendsorrelativesnottobuycarsfromthe VWGroupbecauseoftheemissionsscandal. NWOM2 Becauseoftheemissionsscandal,Iwouldsaynegativethings abouttheVWGroupanditscarstootherpeople. NWOM3 Evenaftertheemissionsscandal,Iwouldrecommendcarsof theVWGrouptosomeonewhoaskedmyadvice. Anger1 Involvement1 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... angry. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... annoyed. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... disgusted. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... outraged. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... sympathetic. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... sorry. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... compassion. WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeel... empathy. Theemissionsscandalis...unimportant/important. Involvement2 Theemissionsscandalis...ofnoconcern/ofconcern. Involvement3 Theemissionsscandalis...meansnothing/meansalot. Involvement4 Theemissionsscandalis...doesnotmatter/matterstome. Involvement5 Theemissionsscandalis...irrelevant/relevant. Anger2 Anger3 Anger4 Sympathy1 Sympathy2 Sympathy3 Sympathy4 86 Responsibility1 Thecauseoftheemissionsscandalwasanintentionalactby someoneintheorganization. Responsibility2 Someoneintheorganizationknowinglycreatedthecauseofthe emissionsscandal. Responsibility3 Theorganizationhadthecapabilitytostoptheemissions scandalfromoccurring. Responsibility4 Theemissionsscandalwaspreventablebytheorganization. Responsibility5 Theorganizationcouldhaveavoidedtheemissionsscandal. Responsibility6 Theorganizationshouldbeheldaccountablefortheemissions scandal. Responsibility7 Theemissionsscandalwascausedbyaweaknessinthe organization. Responsibility8 Internalorganizationalissuescontributedtotheemissions scandal. PCFit1 IhavestrongidentificationwiththeVWGroupwhentalkingto othersaboutit. PCFit2 IprefercarsoftheVWGroupwhencomparingitwiththatof otherautomobilemanufacturers. PCFit3 IampositiveaboutthecompanyimageoftheVWGroup. PCFit4 WhensomeonecriticizestheVWGroup,itfeelslikeapersonal insult. PCFit5 IamveryinterestedinwhatothersthinkaboutVWGroup. PCFit6 ThesuccessesoftheVWGrouparemysuccesses. PCFit7 WhensomeonepraisestheVWGroup,itfeelslikeapersonal compliment. PCFit8 IfastoryinthemediacriticizestheVWGroup,Ifeel embarrassed. 87 AppendixC–FurtherTables TableB1:CorrelationMatrixPost-CrisisReputation Reputation1 Reputation2 Reputation3 Reputation4 Reputation5 Reputation1 1 .29** .31** .40** .70** Reputation2 1 .58** .43** .39** Reputation3 1 .55** .49** Reputation4 1 .34** Reputation5 1 Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation TableB2:CorrelationMatrixNWOM NWOM1 NWOM2 NWOM3 NWOM1 1 .64** -.57** NWOM2 1 .44** NWOM3 1 Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation TableB3:CorrelationMatrixAnger Anger1 Anger2 Anger3 Anger4 Anger1 1 .73** .59** .61** Anger2 1 .48** .66** Anger3 1 .52** Anger4 1 Sympathy3 1 .57** Sympathy4 1 Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation TableB4:CorrelationMatrixSympathy Sympathy1 Sympathy2 Sympathy3 Sympathy4 Sympathy1 1 .32** .46** .54** Sympathy2 1 .71** .45** Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation 88 TableB5:CorrelationMatrixInvolvement Involvement 1 Involvement 2 Involvement 3 Involvement 4 Involvement 5 Involvement 1 1 Involvement 2 Involvement 3 Involvement 4 Involvement 5 .63** 1 .59** .52** 1 .52** .47** .72** 1 .78** .63** .55** .52** 1 Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation TableB6:CorrelationMatrixCrisisResponsibility Resp1 Resp2 Resp3 Resp4 Resp5 Resp6 Resp7 Resp8 Resp1 1 .50** .20** .26** .27** .21** .17** .20** Resp2 1 .17** .27** .28** .22** .14** .18** Resp3 1 .35** .36** .21** .16** .15** Resp4 1 .65** .36** .09** .13** Resp5 1 .33** .11** .13** Resp6 1 .09** .10** Resp7 1 .60** Resp8 1 PC_Fit6 1 .79** .47** PC_Fit7 1 .50** PC_Fit8 1 Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation TableB7:CorrelationMatrixPCFit PC_Fit1 PC_Fit2 PC_Fit3 PC_Fit4 PC_Fit5 PC_Fit6 PC_Fit7 PC_Fit8 PC_Fit1 1 .59** .48** .67** .58** .73** .70** .46** PC_Fit2 1 .57** .44** .40** .45** .43** .30** PC_Fit3 1 .38** .34** .40** .39** .24** PC_Fit4 1 .52** .71** .71** .54** PC_Fit5 1 .56** .52** .46** Note:*p≤.05,**p≤.01;Samplesize=1475;PearsonCorrelation 89 TableB8:ExploratoryFactorAnalysisPost-CrisisReputation Items FactorLoadings Regardingtheemissionsscandal,theVWGroupisconcernedwiththewellbeingofitspublics(e.g.customers,employees,investors).(Reputation1) .651 TheVWGroupisbasicallydishonestconcerningtheemissionsscandal. (Reputation2) .636 IdonottrusttheVWGrouptotellthetruthabouttheemissionsscandal. (Reputation3) .691 Undermostcircumstances,IwouldbelikelytobelievewhattheVWGroup saysabouttheemissionsscandal.(Reputation4) .640 Regardingtheemissionsscandal,theVWGroupisnotconcernedwiththe well-beingofitspublics(e.g.customers,employees,investors).(Reputation5) .694 Cronbach’sAlpha .795 Eigenvalue 2.196 %ofVariance 43.93 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.706;factorloadingsbelow.4arenot includedinthetable TableB9:ExploratoryFactorNWOM Items FactorLoadings IwouldencouragefriendsorrelativesnottobuycarsfromtheVWGroup becauseoftheemissionsscandal.(NWOM1) .904 Becauseoftheemissionsscandal,IwouldsaynegativethingsabouttheVW Groupanditscarstootherpeople.(NWOM2) .707 Evenaftertheemissionsscandal,IwouldrecommendcarsoftheVWGroupto someonewhoaskedmyadvice.(NWOM3) .627 Cronbach’sAlpha .780 Eigenvalue 1.708 %ofVariance 56.94 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.667;factorloadingsbelow.4arenot includedinthetable 90 TableB10:ExploratoryFactorAnalysisCrisisResponsibility Items FactorLoadings accountability locality intentionality Theemissionsscandalwaspreventablebythe organization.(Responsibility4) Theorganizationcouldhaveavoidedthe emissionsscandal.(Responsibility5) Theorganizationhadthecapabilitytostopthe emissionsscandalfromoccurring. (Responsibility3) Theorganizationshouldbeheldaccountablefor theemissionsscandal.(Responsibility6) .803 .780 .423 .397 Theemissionsscandalwascausedbyaweakness intheorganization.(Responsibility7) .905 Internalorganizationalissuescontributedtothe emissionsscandal.(Responsibility8) .646 Someoneintheorganizationknowinglycreated thecauseoftheemissionsscandal. (Responsibility2) .691 Thecauseoftheemissionsscandalwasan intentionalactbysomeoneintheorganization. (Responsibility1) .643 Cronbach’sAlpha .686 .751 .664 Eigenvalue 1.692 1.285 1.012 %ofVariance 21.16 16.06 12.66 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.704;factorloadingsbelow.4arenot includedinthetable 91 TableB11:ExploratoryFactorAnalysisCrisisInvolvement Items FactorLoadings Theemissionsscandalisunimportant/important(Involvement1) .869 Theemissionsscandalisirrelevant/relevant(Involvement5) .856 Theemissionsscandalisofnoconcern/ofconcern(Involvement2) .730 Theemissionsscandalmeansnothing/meansalot(Involvement3) .709 Theemissionsscandaldoesnotmatter/matterstome(Involvement4) .660 Cronbach’sAlpha .852 Eigenvalue 2.961 %ofVariance 59.21 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.810;factorloadingsbelow.4arenot includedinthetable TableB12:ExploratoryFactorAnalysisAnger Items FactorLoadings WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelangry. (Anger1) .855 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelannoyed. (Anger2) .842 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeeloutraged. (Anger4) .752 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeeldisgusted. (Anger3) .643 Cronbach’sAlpha .855 Eigenvalue 2.419 %ofVariance 60.48 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.780;factorloadingsbelow.4arenot includedinthetable 92 TableB13:ExploratoryFactorAnalysisSympathy Items FactorLoadings WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelcompassion. (Sympathy3) .908 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelsorry. (Sympathy2) .760 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelempathy. (Sympathy4) .640 WhenIthinkoftheVWGroupandtheemissionsscandal,Ifeelsympathetic. (Sympathy1) .529 Cronbach’sAlpha .799 Eigenvalue 2.093 %ofVariance 52.33 Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.720;factorloadingsbelow.4arenot includedinthetable TableB14:ExploratoryFactorAnalysisPerson-CompanyFit Items FactorLoadings IhavestrongidentificationwiththeVWGroupwhentalkingtoothersaboutit. (PCFit1) .890 WhensomeonepraisestheVWGroup,itfeelslikeapersonalcompliment. (PCFit7) .772 IamveryinterestedinwhatothersthinkaboutVWGroup.(PCFit5) .662 IprefercarsoftheVWGroupwhencomparingitwiththatofotherautomobile manufacturers.(PCFit2) .652 IampositiveaboutthecompanyimageoftheVWGroup.(PCFit3) .558 IfastoryinthemediacriticizestheVWGroup,Ifeelembarrassed.(PCFit8) .549 Cronbach’sAlpha .838 Eigenvalue 2.863 Variance 47.72% Note:MaximumLikelihoodwithVarimaxrotationwasapplied;KMO=.831;factorloadingsbelow.4arenot includedinthetable 93
© Copyright 2026 Paperzz