WendyWoodDepartmentofPsychologyTexasAMUniversityCollegeStationTexas77843emailwlwpsyctamueduKeyWordsin ID: 107951
Download Pdf The PPT/PDF document "Annu.Rev.Psychol.2000.51:539" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Annu.Rev.Psychol.2000.51:539±5702000byAnnualReviews.AllrightsreservedPersuasionandSocial WendyWoodDepartmentofPsychology,TexasA&MUniversity,CollegeStation,Texas77843;e-mail:wlw@psyc.tamu.eduKeyWordsin¯uence,motives,fearappeals,socialidentity Thischapterreviewsempiricalandtheoreticaldevelopmentsinresearchonsocialin¯uenceandmessage-basedpersuasion.Thereviewemphasizesresearchpublishedduringtheperiodfrom1996±1998.Acrosstheseliteratures,threecentralmotiveshavebeenidenti®edthatgenerateattitudechangeandresistance.Theseinvolveconcernswiththeself,withothersandtherewards/punishmentstheycanprovide,andwithavalidunderstandingofreality.Themotiveshaveimplicationsforinformationprocessingandforattitudechangeinpublicandprivatecontexts.Motivesinpersuasionalsohavebeeninvestigatedinresearchonattitudefunctionsandcognitivedissonancetheory.Inaddition,thechapterreviewstherelativelyuniqueaspectsofeachliterature:Inpersuasion,itconsidersthecognitiveandaffectivemech-anismsunderlyingattitudechange,especiallydual-modeprocessingmodels,recipi-ents'affectivereactions,andbiasedprocessing.Insocialin¯uence,thechapterconsidershowattitudesareembeddedinsocialrelations,includingsocialidentitytheoryandmajority/minoritygroupin¯uence.MotivesforAgreeingwithOthersPublicVersusPrivateIn¯uence.............................................................542MotivesinPersuasionResearchFunctionalTheories...........................................................................544CognitiveDissonanceTheory................................................................546MultipleAttitudesSourcesofMultipleAttitudes................................................................548In¯uenceandMultipleAttitudes.............................................................549Dual-ModeProcessingModelsofPersuasionCognitiveResponseMediationofAttitudeChange.......................................552Dual-ModeProcessingModelsandSocialIn¯uence....................................553MotivatedProcessingandBiasCorrection................................................554AffectandIn¯uenceEffectsofMood................................................................................555FearAppeals...................................................................................556GroupandSelf-Identity WOOD SocialConsensusandValidityofInformation............................................558MultipleMotivesInstigatedbyGroups.....................................................559OpinionMinorityandMajorityGroupsThischapterreviewstheresearchonattitudechangefromwhattraditionallyhavebeentwoseparateareasofinquiry,thestudyofmessage-basedpersuasionandthestudyofsocialin¯uence.Inthepersuasionparadigm,in¯uenceappealstyp-icallyincludedetailedargumentationthatispresentedtoindividualrecipientsinacontextwithonlyminimalsocialinteraction.Socialin¯uenceappeals,incon-trast,usuallyconsistsolelyofinformationaboutthesource'sposition,butthesearedeliveredinmorecomplexsocialsettingsthatmayincludeinteractionamongparticipants.Becauseofthemarkedcontinuitiesinthetheoreticalanalysesandintheempirical®ndingsthathaveemergedacrosstheseresearchareasinthepastfewyears,thisreviewdrawsfrombothliteratures.Itemphasizesinparticularresearchpublishedduring1996to1998,sincethepriorreviewofPettyetalGivingsocialin¯uenceresearchasigni®cantroleinthecurrentreviewrequiresthatlimitedattentionbegiventosomeotherresearchareasthathavebeenfeaturedprominentlyinthepast.Workonattitudestructureandonattitude-behaviorrela-tionshascontinuedto¯ourish,andEagly&Chaiken(1998)provideanexcellentreviewelsewhere.Alsonoteworthy,despitesomeoverlapwiththecurrentreview,arePetty&Wegener(1998a),Cialdini&Trost(1998),andChaikenetal(1996b).Anotherresearchareabeyondthescopeofthischapteristheextensiveworkonintergroupattitudesandstereotypes(Brewer&Brown1998,Fiske1998).MOTIVESFORAGREEINGWITHOTHERSAhallmarkofsocialin¯uenceresearchisthedelineationofthemultiplemotivesthatspuragreementordisagreementwithothers.Forover40years,thecentralorganizingperspectiveinthisareahasbeenadual-motiveschemethatdifferen-tiatesbetweeninformationalin¯uence,whichinvolvesacceptinginformationobtainedfromothersasevidenceaboutreality,andnormativein¯uence,whichinvolvesconformitywiththepositiveexpectationsof``another,''whocouldbe``anotherperson,agroup,orone'sself''(Deutsch&Gerard1955:629).Yetcontemporarytheoriesofmotivesforattitudechangeandresistanceappeartobeconverginginsteadonatripartitedistinction(e.g.Chaikenetal1996a,Cialdini&Trost1998,Johnson&Eagly1989,Wood1999;foranearlypresen-tationofthiskindofframework,seeKelman1958).Althoughthesetypologieseachpossessuniquefeatures,acommonthreadistherecognitionthatattitude ATTITUDECHANGE changecanbemotivatedbynormativeconcernsfor()ensuringthecoherenceandfavorableevaluationoftheself,and()ensuringsatisfactoryrelationswithothersgiventherewards/punishmentstheycanprovide,alongwithaninforma-tionalconcernfor()understandingtheentityorissuefeaturedinin¯uenceappeals.Thus,forexample,Cialdini&Trost(1998)identifythebehavioralgoalsofsocialin¯uencerecipientsasmanagingtheself-concept,buildingandmain-tainingrelationships,andactingeffectively.Similarly,Chaikenetal(1996a)dis-tinguishedbetweenpeople'sego-defensivemotivestoachieveavalued,coherentself-identity,impression-relatedmotivestoconveyaparticularimpressiontooth-ers,andvalidity-seekingmotivestoaccuratelyassessexternalreality.Socialin¯uenceresearcherstraditionallyassumedthatinformationalandnor-mativemotivesareeachassociatedwithuniquemechanismsthatgenerateattitudechangeandwithuniqueformsofchange.Thedesireforaninformed,correctpositionsupposedlyorientsmessagerecipientstoprocessthecontentoftheappealandresultsinenduringprivatechangeinjudgments.Thedesiretomeetnormativeexpectationssupposedlyyieldslessinformationalanalysisandpublic,context-dependent,transitoryjudgmentchange.Thisviewhasbeenchallengedbydual-modeprocessingmodelsofpersuasion(Eagly&Chaiken1993,Petty&Wegener1998a),especiallybythedemonstrationthatinformational,accuracy-seekingmotivescanleadeithertoextensiveprocessingandenduringattitudechangeortomoresuper®cialprocessingandtemporarychange.Inthedual-modeframe-work,motivesforchangearenotpreferentiallyrelatedtochangemechanismsorTworecentstudiessupportthepersuasionanalysisbyprovidingevidencethatnormativeandinformationalmotivesaffectin¯uencethroughacommonsetofinformation-processingmechanisms(Chenetal1996,Lundgren&Prislin1998).Lundgren&Prislin(1998)foundthat,whenparticipantsweremotivatedtobeaccurate,theyselectedargumentstoreadonbothsides(i.e.proandcon)ofthetargetissue,generatedthoughtsthatwererelativelybalancedinevaluationofbothsides,andindicatedrelativelyneutralattitudes.Incontrast,whenparticipantsweremotivatedtoconveyafavorableimpressiontoaninteractionpartner,theyselectedargumentstoreadthatwerecongruentwiththeviewostensiblyheldbythepartnerandgeneratedthoughtsandattitudesthatwerecongenialwiththeirpartner'sposition.Finally,whenparticipantsweremotivatedtodefendtheirownposition,theyselectedargumentstoreadthatsupportedtheirview,generatedthoughtssupportiveoftheirposition,andindicatedrelativelypolarizedattitudes.Furthermore,analysestotestmediationrevealedthatthefavorabilityofpartici-pants'thoughts(atleastpartially)mediatedtheeffectsofmotivesonattitudechange.Thus,itappearsthattheattitudeeffectsemergedinpartbecauseaccuracymotivesgeneratedarelativelyopen-mindedprocessingorientation,impressionmotivesgeneratedanagreeableorientation,anddefensemotivesgeneratedapro-tectiveorientationthatmaintainedexistingjudgments.Thesetwostudiesalsochallengetheassumptionthatrecipients'motivesareassociatedwithuniqueformsofattitudechange.Regardlessoftheinitialmotive WOOD directingattitudejudgments,theattitudesparticipantsexpressedtotheirosten-siblediscussionpartnerspersistedwhenparticipantslaterindicatedtheirjudg-mentsprivately(Chenetal1996,Lundgren&Prislin1998).Especiallyimpressiveisthepersistenceofattitudesdesignedtoconveyafavorableimpres-sion.Contrarytoclassictheoriesofsocialin¯uence,attitudesdirectedbyimpression-relatednormativemotiveswerenomore``elastic''thanwereattitudesdirectedbyaccuracy-seeking,informationalmotives.Instead,itseemsthatimpressionanddefensemotives,muchliketheaccuracymotivesstudiedexten-sivelyinmessage-basedpersuasionresearch,canyieldcareful,systematicpro-cessingofrelevantinformationthatresultsinstablejudgments.This®ndingaugmentstheresultsofearlierresearchinwhichimpressionmotiveswerelinkedtosuper®cialprocessingandtemporaryjudgmentshifts(e.g.Cialdinietal1976).Thefactorsthatdeterminewhethermotivesinstigateextensiveormoresuper®cialprocessingarediscussedbelow.PublicVersusPrivateIn¯uenceInsocialin¯uenceparadigms,researchersoftenhavediagnosedthemotiveforattitudechangefromthecontinuityofrecipients'judgmentsacrosspublicandprivatesettings.Inpublicsettings,recipientsbelievethatthesourceoftheappealormembersoftheirexperimentalgrouphavesurveillanceovertheirresponses,whereasinprivatesettings,recipientsbelievethattheseothersareunawareoftheirjudgments.Supposedly,attitudesthatmaintainacrosspublicandprivatemeasuresareinternalizedresponsesthatresultfromthethoughtfulprocessingassociatedwithaccuracymotives,whereasattitudesthatareexpressedinpublicbutnotprivatere¯ectnormativepressuressuchasacceptancefromthesourceorRecentempirical®ndingssuggestinsteadthatlackofcontinuitybetweenpub-licandprivatejudgmentsisnotreliablydiagnosticofrecipients'motives.Asdescribedabove,enduringattitudechangeisnottheuniqueprovinceofinfor-mationalmotives;italsocanemergefromsocialmotivessuchasthedesiretoaccommodateothers(Chenetal1996,Lundgren&Prislin1998).Evidenceofjudgmentstabilityacrosspublicandprivatesettingshasemergedalsoinin¯uenceintheAsch-typejudgmentparadigm.Inthisresearch,participantsareexposedtoothers'obviouslyincorrectjudgmentsoflinelengthandparticipants'agreementwithotherstypicallyisinterpretedaspublic,super®cialacquiescencetosocialpressure.Yetthemeta-analyticsynthesisbyBond&Smith(1996)of97studiesusingtheAsch-typesocialin¯uenceparadigmrevealednogreateragreementinpublicthaninprivatecontextsofattitudeexpression.Itseems,then,thatsocialmotivesforagreementaffectedattitudesinpublicaswellasprivatesettings.Thelackofsystematicdifferencesbetweenpublicandprivateexpressionsofjudgmentcanbeattributedtoanumberoffactors.Oneisthatrecipients'motivesforagreementcanhaverelativelyextendedeffectsthatgeneralizetonewcontextsinwhichtheoriginalmotivesarenolongersalientorrelevant(e.g.Hardin& ATTITUDECHANGE Higgins1996,Higgins1981,Ruscher&Duval1998).Extendedeffectscanoccurwhentheinitialmotivatedjudgmentisretrievedinnewsettingsorwhentheinformationonwhichthejudgmentwasbasedisretrieved,giventhatthemoti-vatedprocessingyieldedabiasedrepresentationoftheoriginalinformation.Thus,becausemotivesaffectthejudgmentsandthejudgment-relevantinformationavailableinmemory,initialmotivationsforprocessingmayhaveeffectsthattranscendcontext,andpositionsstatedinpubliccontextsmaybemaintainedinprivate.Kassin&Kiechel(1996)provideacompellingexampleoftheextendedconsequencesofmotivatedprocessing.Theysimulatedproceduressometimesusedintheinterrogationofcrimesuspectsby(falsely)accusingresearchpartic-ipantsofanactofnegligencewhiletheyweretypingdataintoacomputer.Whenparticipantsweresubjectivelyuncertainabouttheirinnocence(becausetheyweretypingatafastspeed),theyacceptedawitness'sreportoftheiractionsand(incor-rectly)confessedtotheallegation.Forthemajorityofparticipants,theconfessionwasnotmerecompliance.Overhalfreportedinasubsequentdiscussionthattheyhadperformedthenegligentact,andoverathirdactuallyconfabulateddetailsinsupportofthefalseallegation.Furthermore,thedistinctionbetweenpublicandprivatesettingssuggestsanoverlysimpli®edviewofsocialimpact,onethatequatessocialpresencewithsurveillance.Allport's(1985)famousde®nitionofsocialpsychologyprovidedaconsiderablymoredifferentiatedviewofsocialimpact,inwhichtheeffectsofothersemergewhethertheirpresenceis``actual,imagined,orimplied.''Becauseimportantfeaturesofsocialimpactmayholdacrosspublicandprivatecontexts,attitudesthatareaffectedbythesefeaturesmayalsoholdacrosssettings.Forexample,themanipulationbyBaldwin&Holmes(1987)ofsocialimpactinvolvedsimplyinstructingfemaleparticipantstothinkabouttwooftheirolderrelatives.Thewomenwerelatergivensexuallyexplicitmaterialtoreadinasupposedlyunrelatedcontext,andtheyreportednotlikingitmuch.Presumably,others'conservativemoralstandardswereactivatedintheinitialmanipulationandcontinuedtoexertimpactonsubsequentexperiences.Theoreticalperspectivesneedtoprogressbeyondthesimpledistinctionbetweenpublicandprivateattitudeexpressionandconsiderwhetherthefeaturesofsocialpressurethatarerelevanttoattitudechangearestableacrosssettings.Forexample,inameta-analyticsynthesisoftheminority-in¯uenceliterature(Woodetal1994),thein¯uenceofopinion-minority,low-consensussourcesprovedcomparableinpublicandprivatesettings.Thus,itseemedthatattitudechangewasnotcontrolledbysurveillanceandthefearthataligningwithadeviantminoritysourceinpublicwouldleadtosocialembarrassmentandrejectionbyothers.Agreementdidvary,however,withanotherfeatureofthein¯uencecon-text;howdirectlyattitudesweremeasured.``Direct''measuresassessattitudesontheissueintheappeal,andrecipientsareawarethattheir(publicorprivate)judgmentcanalignthemwiththesource'sposition.``Indirect''measuresmight,forexample,assessattitudesonissuestangentiallyrelatedtotheappeal,andrecipientsarelessawarethattheirjudgmentscanalignthemwiththein¯uence WOOD source.Minorityimpactwassmallerondirectthanonindirectmeasures.Woodetal(1994)concludedthatrecipients'resistanceondirectmeasuresisduetotheirownpersonalknowledgethattheirjudgmentscouldalignthemwithadevi-antminoritysource.Itseems,then,thatminorityin¯uencewasinhibitedbyrecip-ients'concernforthefavorabilityandintegrityoftheirself-conceptandtheirplaceintheirreferencegroup,andthatthesemotivesheldinbothpublicandprivatecontexts(seebelow).Thecurrentanalysisofattitudeexpressionsinpublicandprivatecontextsalsocallsintoquestionthecommonassumptionthatwhenpublicandprivatejudg-mentsdifferinaccuracy,privatelyexpressedonesaregenerallymoretrustworthybecausepublicexpressionsmaybebiasedtoachievesocialmotives.Althoughsomefeaturesofpubliccontexts(e.g.politenessnorms)maycompromisetheaccuracyofattitudeexpressions,otherfeaturesappeartoenhancethoughtfulanal-ysesandsometimestoincreaseaccuracy.Cowan&Hodge(1996)demonstratedthattotheextentpubliccontextsenhanceperceivedaccountabilityforjudgments,peoplegiveespeciallythoughtful,reasonedresponsesinpublic.Similarly,Lam-bertetal(1996)arguedthattheexpectationofpublicdiscussionfocusespeopleontheirownattitudesandencouragesthemtobolstertheirbeliefs;thus,attitudeswerefoundtoplayagreaterroleinguidingthoughtandactioninpublicthaninprivatesettings.Finally,giventhatinsocial-in¯uenceparadigmsrespondentsoftengivetheirjudgments®rstpubliclyandthenagainprivately,continuityacrossjudgmentcon-textscanemergefromtheeffectsofinitialjudgmentsonsubsequentones.Researchontheimpactofbehavioronattitudeshasdemonstratedthatpeople'sinterpretationsoftheirpublicstatementsandotherattitude-relevantbehaviorscaninstigateshiftsinprivatelyheldattitudestocorrespondtopublicacts(seeChaikenetal1996b).Thisresearchalsohasdemonstratedthatpublic-attitudestatementsthatareofquestionableveracity(e.g.whenapublicstatementisgivenwithlowchoiceorhighreward)donotaffecttheattitudesexpressedinprivatesettings(seebelow).However,Maio&Olson(1998)provideintriguingevidencethatevenunderlow-choiceconditions,theactofprovidinganattitudejudgmentcanenhancetheaccessibilityofone'sownattitudeinmemory;accessibleattitudesthenmayaffectsubsequentattitude-relevantjudgmentsinseeminglyunrelatedMOTIVESINPERSUASIONRESEARCHFunctionalTheoriesThemotivesunderlyingattitudechangeinmessage-basedpersuasionparadigmshavebeeninvestigatedprimarilyinresearchonattitudefunctions(seeEagly&Chaiken1998).Inadditiontothebasicadaptivefunctionofenablingpeopletoevaluateandappraisestimuliintheirenvironment,attitudesalsoarethoughttoservemorespeci®cfunctions.Functionsidenti®edinearlyworkincludesecuring ATTITUDECHANGE utilitarianoutcomes,egodefense,valueexpression,andsocialadjustment(Katz1960,Smithetal1956).Thesefunctionsarere¯ectedinthetripartitemotiveschemesuggestedabove:Accuracymotivescorrespondgenerallytoautilitarianconcernwithmaximizingrewardsandminimizingpunishments,self-conceptmotivescorrespondtoconcernsfordefendingtheegoagainstpotentialthreatsandforexpressingone'svalues,andsocialrelationmotivescorrespondtocon-cernsforsocialadjustmentandforobtainingsocialrewardsandavoidingsocialInoneaccountoftheroleofattitudefunctionsinin¯uence,persuasiveattemptsarelikelytobeeffectivetotheextentthatthefunctionof,orreasonforholding,thepositionoutlinedintheappealmatchesthefunctionunderlyingrecipients'attitudes(Lavine&Snyder1996,Murrayetal1996:Study2).Forexample,Lavine&Snyder(1996)reportedthatforpeoplewhoaregenerallysensitivetothesocialconsequencesoftheirbehavior(i.e.highself-monitors),appealsthatemphasizedthesocialadjustivefunctionsofvoting(e.g.enhancingone'sattrac-tivenesstoothers)elicitedmorefavorableevaluationsandgreaterattitudechangethanappealsthatemphasizeditsvalue-expressivefunctions(e.g.awaytoexpressvalues).Forpeoplewhorelyoninnerdispositions(i.e.lowself-monitors),appealswithvalue-expressiveargumentsyieldedmorefavorableevaluationsandweremorepersuasive.Furthermore,certainissuesmaybeassociatedwithcertainatti-tudefunctionsformostpeople.Althoughnotspeci®callycouchedwithinafunc-tionalframework,theanalysisbyRothman&Salovey(1997)ofhealth-relatedmessagessuggestedthatin¯uenceisgreatestwhentheorientationofanappealmatchestheorientationintrinsictothehealthissueitself.Recommendationstoperformillness-detectingbehaviors(e.g.breastself-exams)potentiallyincurriskandthusloss-framedappealsarelikelytobeeffective,whereasrecommendationsforpreventativebehaviors(e.g.exercise)potentiallyincurpositiveoutcomesandthusgain-framedappealsarelikelytobeeffective.Inanotheraccount,matchingbetweenattitudefunctionandmessageorienta-tiondoesnotalwaysenhancepersuasionbutinsteadenhancescarefulthoughtaboutanappeal.Petty&Wegener(1998b)demonstratedthatmatchedmessagesincreasedscrutinyofmessagecontentbutenhancedpersuasiononlywhenthemessagecontainedstrong,cogentargumentsandnotwhenitcontainedweakarguments.Yetbecausefunctionallymatchedmessagespotentiallyaddressimpor-tantaspectsofrecipients'self-concepts,thiscarefulprocessingwillnotalwaysbeobjectiveandunbiased.Suchappealsmayinstigateathoughtfulbutdefensiveorientation,asrecipientstrytomaintainvaluedaspectsoftheself.Forexample,Tykocinskietal(1994)reasonedthatmessagesframedtomatchpeople'scurrentexperiencesandconcernscanelicitdistressbyidentifyingseeminglyrelevantgoalsthathavenotbeenadopted.Thus,suchmessagesareespeciallylikelytoyieldcounterarguingandresistance.Similarly,Marshetal(1997)reportedthatpersuasivemessagesthataddressanimportantattitudefunction(i.e.forcollegestudents,thevalue-relevantissueofsororities/fraternitiesoncollegecampuses)areprocessedcarefullyyetdefensivelyandasaresultareminimallyin¯uential. WOOD CognitiveDissonanceTheoryThisclassicmotivationaltheoryofhowattitudeschangetomaintaincognitiveconsistencycontinuestosparkinterest.TheoriginalnotionofFestinger(1957)thatdissonancearisesfrompsychologicalinconsistencybetweenlinkedcogni-tionshasbeenmodi®edextensivelyinsubsequentresearch.InCooper&Fazio's(1984)``newlook''approach,dissonancearisesnotfromsimpleinconsistencybutratherfromfreelychosenbehaviorthatbringsaboutsomeforeseeableneg-ativeconsequence.Acentralquestionfordissonanceresearchershasbeenthemotivationalbasesfordissonanceandthecauseoftheaversivestateofdissonancearousal.InAron-son's(1992)self-conceptanalysis,dissonancearisesfrominconsistentcognitionsthatthreatentheconsistency,stability,predictability,competence,ormoralgood-nessoftheself-concept.InSteele's(1988)self-af®rmationtheory,dissonancearisesfromtheviolationofgeneralself-integrity.Fromtheseself-relatedper-spectives,negativeconsequencesarepowerfulinducersofdissonancebecauseitisinconsistentwithmostpeople'sself-viewstoactinawaythatresultsinfore-seeableaversiveconsequences.Analternateperspectiveondissonancearousal,whichhasyettobeintegratedintomainstreamtheorizing,istheargumentbyJoule&Beauvois(1998)thatdissonancereductionisorientedtowardrational-izingbehaviorratherthanattainingpsychologicalconsistency.Inaninterestingintegrationthatrecognizesthatbothself-conceptthreatandaversiveconsequencescaninstigatedissonance,Stone&Cooper(seePetty&Wegener1998a)proposedthatdissonanceariseswhenpeoplefailtobehaveinamannerconsistentwithsomevaluedself-standard.Thespeci®cmotivationbehinddissonancesupposedlydependsonthetypeofself-standardinvolved.Dissonancecanemergefrombehaviorthatisinconsistentwithpersonalself-standardsanddoesnotre¯ectthewaypeoplewanttobe(idealself)orthinktheyshouldbe(oughtself),ordissonancecanemergefrombehaviorthatgeneratesaversiveconsequencesanddoesnotre¯ecthowotherswantthemtobe(normativeself-Severalstudiessupporttheconclusionthatdissonancemotivationcanemergeincontextsdevoidofnegativeconsequences.ParticipantsinastudybyHarmon-Jonesetal(1996)freelyengagedinthenonconsequentialbehaviorofprivatelytakingacounterattitudinalposition,yettheyexperiencedincreasedarousalandattitudeshiftstowardtheirexpressedposition.Similarly,Prislin&Pool(1996)foundlittleevidencethatdissonancearisesonlywhenbehaviorhasidenti®ablynegativeconsequencesandinsteadconcludedthatdissonanceemergeswhenbehavioranditsconsequenceschallengeexistingideasabouttheself.Thehypocriticaladvocacyparadigmwasdevelopedtostudydissonancemoti-vationintheabsenceofnegativeconsequences.Inthisresearch,participantsadvocateaproattitudinalposition,aremadeawareoftheirpastfailurestoactinaccordancewiththisposition,and(inordertoreducedissonance)thenengageinactscongruentwiththeposition(e.g.Fried1998,Stoneetal1997).Althoughthe ATTITUDECHANGE lackofimmediatenegativeconsequencestothisproattitudinaladvocacymightappeartoindicatethatsuchconsequencesarenotcriticalforproducingdisso-nance,proponentsofthenegativeconsequencesviewcanrightlynotethathigh-lightingpastfailurestoupholdone'sstatedbeliefsidenti®esaversiveconsequencesofpastacts.However,Fried&Aronson(1995)arguedfortheimportanceoftheselfininstigatingdissonancebynotingthatdissonanceinthisparadigmdoesnotemergewhenpeopleareonlyremindedofpasttransgressions(i.e.negativeconsequences)anddonotengageinproattitudinaladvocacyandexperiencetheself-relatedimplicationsofthisbehavior.Thesecondmajorissueaddressedindissonanceresearchisthemultipleroutesormodesthroughwhichdissonancecanbereduced.Althoughself-af®rmingbehaviorthatreestablishespersonalintegrityhasbeenshowntoreducedissonance(Steele1988),self-af®rmationsarenotalwaysthemodeofchoice.Whenmultipleroutesareavailable,peopleapparentlyprefertoreducedissonancedirectlybychangingattitudesandbehaviors(i.e.modifyingtheinconsistentcognitions)ratherthantoalleviateitindirectlythroughself-af®rmations(Stoneetal1997).Evenpeoplewithhighself-esteem,whoshouldpossesstheresourcestoreducedissonancebyfocusingonpositiveaspectsoftheself-concept,havebeenfoundinsteadtomodifycognitions(Gibbonsetal1997).Otherresearchhasidenti®edboundaryconditionsfortheusefulnessofself-af®rmations.AccordingtoBlantonetal(1997),af®rmationsdonotreducedissonanceiftheyremindpeopleoftheviolatedself-standard(e.g.reassuranceofone'scompassionwhenonehasactedinanoncompassionateway),presumablybecausesuchaf®rmationsmakeitmoredif®culttojustifythedissonance-inducingactandleadonetodwellonthedis-sonantbehavior.Otherresearchonmodesofdissonancereductionhasrevealedindividualdif-ferencesintherouteofchoice.Peoplewhoarehighinattributionalcomplexityandcharacteristicallysearchforabstract,complexexplanationsappeartoreducethedissonancecausedbycounterattitudinaladvocacybyconsideringpossibleexternaljusti®cationsfortheattitude-discrepantactratherthanbychangingatti-tudes(Stadler&Baron1998).Inaddition,suggestingthatresearchhasonlybeguntotapthevarietyofmodesavailable,Burrisetal(1997)documentedthedissonance-reducingeffectsoftranscendence(reconcilinginconsistentbeliefsunderabroaderprinciple)andreaf®rmationoftheattackedbelief.Arecognitionofthefullrangeofroutesavailablefordissonancereductioncanhelptoaccountforsomeoftheseeminginconsistenciesinthedissonanceliterature.Forexample,whenFried(1998)modi®edthehypocrisyparadigmtomakepublicparticipants'pastfailurestoliveuptotheirattitudes,participantsdidnotperformattitude-congruentcompensatingbehaviorstoreducedissonance.Instead,theydecreasedthedissonancecausedbypublictransgressionsbychangingtheirattitudestobecongruentwiththebehavioraltransgression.Anewperspectiveworthwatchingisthedevelopmentofformalmathematicalmodelsofdissonance-relatedprocesses.Parallelconstraintsatisfactionsystemsrepresentdissonanceasadynamic,holisticprocessandprovideameanstoeval- WOOD uateconsistencyandotheraspectsoftherelationbetweenbeliefs(Readetal1997).Inoneapplicationofthisapproach,researchersspecifyanetworkofatti-tudesandothercognitions,simulatechangesinthenetwork(accordingtopre-speci®edrulesofhowtheattitudes/cognitionsarerelatedtoeachother)untilitreachesastateofoverallconsistency,andcomparetheresultsofthesimulationtodatafromresearchparticipants.Shultz&Lepper(1996)wereabletousethistechniquetosuccessfullyaccountforjudgmentchangeinseveralclassicdisso-nanceparadigms.MULTIPLEATTITUDESSourcesofMultipleAttitudesAcentralassumptionofmuchattitudetheorizingisthatpeople'sevaluationsofagivenobjectarestableacrosstime,context,andformofassessment.Empiricalevidenceofthiscoherencehasemergedwithincreasingclarityinrecentyears,inpartduetoimprovedunderstandingofthedeterminantsofcoherenceinatti-tudinalresponses[e.g.strongattitudes(Petty&Krosnick1995)]andinparttoimprovedmethodstodocumentcoherence[e.g.compatibilityinmeasurement(Ajzen1996)].However,empiricalevidencethatpeoplecanholdmultipleatti-tudestowardagivenobjectisemergingaswell,andthispro®leofdissociationhasbeenexplainedinanumberofways.Theevidenceformultipleattitudeshassometimesbeendismissedasre¯ectingepiphenomena(e.g.contexteffects,``nonattitudes''orweaklyheldattitudes).Yetmultipleattitudesalsocanstemfrommoreenduringeffects.Forexample,atti-tudesthatvarywithcontextcanrepresenttemporaryconstructions(e.g.differ-entialuseofratingscales,anchoring,andadjustmenteffects)ormoreinternalizedtendenciestorespondthatarestablylinkedtocertaincontextualfeatures(McConnelletal1997).Inaddition,multipleevaluationsofanattitudeobjectcanemergefromattitudestructure.Theycanre¯ectthesuper®cialresponsesassociatedwithattitudesthathaveminimalcognitiveandaffectivebasesorthemorestablereactionsassociatedwithattitudesthathaveinconsistentcomponents,aswhenattitudesarestructurallyinconsistent(e.g.Chaikenetal1995,Prislinetal1998),heldwithambivalence(Priester&Petty1996,Thompsonetal1995),orassociatedwithvariedaffectiveresponses[e.g.immediateversusanticipatedfutureaffect(Richardetal1996a,b,VanderPligtetal1998)].Inrecentyears,multipleattitudeshavebeguntogenerateinterestintheirownright,asincreasedtheoreticalunderstandingprovidesabasisforpredictingboththecoherenceandthedissociationthatoccursinattitudejudgments(e.g.Mackie&Smith1998,Wilson&Hodges1992).Thisisastill-developingresearchareathatencompassesasomewhatdiversesetofeffects,includingcontext-dependentattitudes,multiplycategorizableattitudeobjects,explicitversusimplicitattitudes,subjectiveconstruals,andissueframing.Thecommonthemelinkingthesevarious ATTITUDECHANGE researchareasisthatpeople'sresponsestoaparticularattitudeobjectcanre¯ectdiverseevaluations,cognitiverepresentationsandinterpretations,andaffectiveOnesourceofdissociationsisthevarietyofmotivationsthatcanunderlieattitudes.Dissociationisimpliedintheclaimofsocialin¯uencetheoriststhatpeoplearemotivatedtoadoptattitudesofrelevantreferencegroupstotheextentthatthegroupidentityissalientanddesirable(Kelman1958,Turner1991).Thatis,peoplemaypossessmultiplecognitiverepresentationsofattitudesonanissuethatisimportanttomorethanoneofthegroupswithwhichtheyidentify.Prelim-inarysupportforthisideawasprovidedin®ndingsbyWood&Matz(unpublisheddata)thatcollegestudents'attitudestowardwelfareprogramsweremorefavor-ablewhentheirsocialidentitiesasreligiouspeopleweresalient(relativetoabase-lineattitudemeasureobtainedwithnosalientgroupidentity),whereasstu-dents'attitudeswerelessfavorablewhentheiridentitiesaspoliticalconservativesweresalient.Furthermore,theseattitudeshiftsemergedmoststronglyamongparticipantswhoconsideredreligiousorconservativesocialgroupstobeself-Dissociationsalsocanarisefromcognitiveprocesses.Totheextentthatanattitudeobject(e.g.yogurt)isrelevanttoadiversesetofissuesandvalues,peoplecanselectfrommultiplecategories(e.g.dairyproducts,healthfoods)whencon-struingtheobject.Anumberoffactorsdeterminethecategoryselected,includingtheaccessibilityofcategoryattitudes(ERSmithetal1996).Furthermore,asingleattitudeobjectmaygeneratemultiplerepresentationsinmemorywhenattitude-relevantinformationisstoredseparatelyfromoverallevaluationsoftheobject(Hastie&Park1986).Indeed,McConnelletal(1997)demonstratedthatdifferentattitudescanbegeneratedindifferentcontextsforasingleattitudeobjectwhenperceiversdonotattempt,on-line,toformanintegratedattitude.Multiplecog-nitiverepresentationsofanissuealsocanemergefrominformationprocessinginsocialin¯uencesettings.Afterbeingexposedtothejudgmentsofothersinanin¯uenceappeal,recipientsmaylaterretrievethisinformationwithoutrecogniz-ingitssourceand,undersomecircumstances,unwittinglyadoptitastheirownresponse(Betzetal1996).Dissociationsalsocanariseasby-productsofthevarietyofprocessesthroughwhichattitude-relevantjudgmentsaregenerated.Explicit,consciousjudgmentsdifferinanumberofwaysfromjudgmentsthatareimplicitlyheld,includingthekindsofinformationconsidered(Ajzen1996,Greenwald&Banaji1995).Forexample,Wittenbrinketal(1997)suggestthatexplicitandimplicitmeasuresofracialprejudiceareonlymoderatelycorrelatedbecausepeoplearemorelikelytobaseexplicitjudgmentsonanegalitarianideology;thusexplicitattitudestowardracialminoritiesaremorefavorablethanimplicitones.In¯uenceandMultipleAttitudesIn¯uencestrategieshavecapitalizedonthemultiplecognitiverepresentations,affectivereactions,andevaluationsthatpeoplecanholdconcerningagivenobject.Asch(1940)arguedearlyonthattheprimaryprocessinin¯uenceisnot WOOD changeinattitudestowardanobjectbutratherchangeinthede®nitionandmean-ingoftheobject.Whenmeaningchanges,attitudeschangeaccordingly.Thelinkbetweenmeaningandevaluationissuggestedinthe®ndingbyBosveldetal(1997)thatpeoplearemorefavorabletoward``af®rmativeaction''whenothersclaimthatitreferstoequalopportunityratherthan,forexample,reversediscrim-ination.Evensubtleaspectsofthewayanissueisframedorrepresentedinanappeal,suchastheapparentlocationandtimeatwhichaproposalwilltakeplace,canaffectrecipients'attitudes(Liberman&Chaiken1996).Thislatter®ndingraisestroublingquestionsforpersuasionresearchthathasvariedpersonalrele-vancethroughthesupposedtimeorlocationofamessageproposal(e.g.insti-tutingseniorcomprehensiveexamsatone'sownuniversityoranotheruniversity).Typically,ithasbeenassumedthatrecipientsarerespondingtocomparableatti-tudeobjectsregardlessofwhetherthetimeandlocationofthemessageproposalrendersitpersonallyrelevant.Issueframingisanin¯uencestrategythatcapitalizesonmultipleattitudes.Inthisapproach,theimportanceandrelevanceofcertainconsequencesorattributesofanissueareemphasizedoverotherpotentialconsequences(Ball-Rokeach&Loges1996,Nelson&Kinder1996,Nelsonetal1997,Price&Tewksbury1997).Forexample,Eagly&Kulesa(1997)illustratehowpersuasiveappealsfrombothsidesofenvironmentaldebateshaveframeddesiredpositionsasachievingvaluesfavoredbyrecipients(e.g.ahealthyeconomybasedontourismorbasedonloggingandindustry)andframedthenondesiredpositionasachievingvaluesmostrecipientsreject.Readersmaywonderhowappealsthatuseframingdifferfromonesthatusethebeliefchangestrategiestypicalofmessage-basedpersuasionparadigms.Stan-dardpersuasionappealstypicallyaddressintraattitudinalrelationsanddescribethespeci®cattributespossessedbyanattitudeobject.Incontrast,asimpliedinthelabel,framingappealstypicallyaddressinterattitudinalrelationsandplaceanissueorobjectinthecontextofotherattitudeissues,values,andgoals.Further-more,framingappearstohighlighttherelevanceorimportanceofexistingknowl-edgestructuresandvalues.Nelsonetal(1997)reportedthatframingeffectsaretypicallystrongerforparticipantsfamiliarwithanissue(i.e.possessingtherele-vantknowledgestructures)thanforthoseunfamiliarwithit.Despitetheseuniquefeaturesofframing,establishedmodelsofattitudechangeseemappropriateforunderstandingitseffects.Forexample,expectancy-valueformulationscanaccountforframingeffectsthroughvariationsinthesalienceandlikelihoodofrelevantgoalsandvalues,aswellasthroughvariationsintheevaluationgiventothem(Ajzen1996).Themechanismsunderlyingframingeffectsprobablydepend,asinstandardpersuasionparadigms,ontheextenttowhichrecipientsaremotivatedandabletoprocessmessagecontentandotherrelevantinformation(seediscussionofthedual-modeinformationprocessingmodelsinnextsection).Forexample,theimpetustoadoptanewinterpretationorframeforanissuecanarisefrommoti-vationalgoalssuchasrejectingastronglydislikedgroupidentity.Inademon- ATTITUDECHANGE strationofthisprocess,collegestudentsinformedthattheirattitudescorrespondedtothoseofahatedgroup,theKuKluxKlan,shiftedtheirowninterpretationsoftheissueawayfromaseeminglyracistconstrual;theythenwereabletoshifttheirattitudesawayfromthoseoftheKlan(Woodetal1996,Pooletal1998).DUAL-MODEPROCESSINGMODELSOFPERSUASIONPersuasionresearchhascontinuedinthehighlysuccessfultraditionofthedual-modeprocessingtheories,theheuristic/systematicmodel(Chaikenetal1996a)andtheelaborationlikelihoodmodel(Petty&Wegener1998a).Thecentraltenetofthesetheoriesisthatthedeterminantsandprocessesofattitudechangedependonpeople'smotivationandabilitytoprocessissue-relevantinformation.Whenpeoplearenothighlymotivated(e.g.theissueisnotpersonallyinvolving)ortheyhavelowability(e.g.theyaredistracted),attitudejudgmentsarebasedoneasilyavailableattributesofasource,message,orsituationthatareevaluatedviaef®cientprocessingstrategies.Forexample,theymightusetheheuristicrule,consensusimpliescorrectness,andthusagreewithamajorityposition.Whenpeoplearebothmotivatedandabletoprocessinformationcarefully,thenattitudesarebasedonamorethoughtful,systematicassessmentofrelevantinformation.Accordingtotheelaborationlikelihoodmodel,suchhigh-elaborationprocessesincludelearningmessagecontent,generationofcognitiveresponses,anddissonance-inducedreasoning(Petty&Wegener1998a).Forevaluationsofthesimilaritiesanddifferencesbetweentheheuristic/systematicmodelandtheelab-orationlikelihoodmodel,seeChaikenetal(1996a),Eagly&Chaiken(1993),andPetty&Wegener(1998a).Researchhascontinuedapaceidentifyingthefactorsthatenhancesystematic,thoughtfulprocessing.Accordingtotheheuristic/systematicmodel,peoplearemotivatedtoengageinsystematicthoughtinordertoachieveasuf®cient``desiredlevelofcon®dence''intheirjudgments.Factorsthathavebeenfoundtoincreasesystematicprocessing(presumablybydecreasingactualcon®denceorincreasingdesiredcon®dence)includethefollowing:framingofpersuasivemessagesinanunexpectedmanner(Smith&Petty1996);self-relevanceofmessages,eitherbecauserecipientsself-referenceoraremadeself-aware(Turco1996);anduseinmessagesoftokenphrasesthatambiguouslysignalbroadervalues,suchas,forDemocrats,``familyvalues''rhetoric(Garst&Bodenhausen1996).System-aticprocessingalsohasbeenfoundwhenrecipientsholdstrong,accessibleatti-tudesonthemessagetopic(Fabrigaretal1998),holdambivalentattitudes(Maioetal1996),orenjoyeffortfulcognitiveactivity(i.e.arehighinneedforcognition;forreview,seeCacioppoetal1996).Inaddition,recipientsengageinsystematicprocessingincircumstancesinwhichcarefulthoughtislikelytogeneratejudg-mentcon®dence,suchaswhenrecipientsbelieveintheirownef®cacyandabilitytoevaluate(Bohneretal1998b)andwhenthemessageispresentedinaccessible,notoverlycomplexlanguage(Haferetal1996). WOOD Systematicprocessingalsomaybeimplicatedinresistancetoin¯uence.Pfau's(1997)insightfulreviewofresistanceresearchfocusedinparticularoninoculationprocedures,inwhichrecipientsreceiveinformationthatstrengthenstheirattitudesbeforeexposuretopersuasiveattack.Hearguesthatinoculationeffectsemergewhenanticipatedthreats(e.g.warningofapotentialattacktoone'sattitudes)motivatethoughtfulprocessingtosupportone'sownpositionortocounteroppos-ingones.Furthermore,resistancetoattitudechangeappearstoincreaseatmidlife(Visser&Krosnick1998).Middle-agedpeopleappeartohaveespeciallystrongattitudesthatenablethemtocounterargueopposingpositions.Heuristicprocessingisusedwhenalow-effortstrategyyieldsattitudeswithasuf®cientlevelofcon®dence(Chaikenetal1996a).Avarietyofheuristiccueshavebeenidenti®edinrecentresearch.Forexample,familiarsayingscanprovidecuestoagreement(Howard1997),andthesubjectiveexperienceofthinkingaboutanissuecanbeacue,sothat,forexample,peopleadoptpositionswhentheycaneasilygeneratesupportiveargumentsandrejectpositionswhentheycannot(Roth-man&Schwarz1998).Inaddition,relevanttounderstandingtheheuristiccuesusedineverydaycontextsareDickerson's(1997)observationsofpoliticiansbol-steringtheirfavoredpositionswithreferencestounbiased,expertsourcesandsourceswithanapparentallegiancetotheopposition.CognitiveResponseMediationofAttitudeChangeResearchhascontinuedtoaddressthemechanicsunderlyingsystematic,high-elaborationprocessing.Onequestioniswhetherthevalenceofrecipients'cog-nitiveresponsesmediatespersuasionorwhethervalencerepresentseitheranalternatemeasureofattitudechangeoranafter-the-factjusti®cationofchange.Totestcausalprecedence,Romeroetal(1996)andKilleya&Johnson(1998)directlymanipulatedthoughtsandexaminedtheeffectsonattitudes.Consistentwiththeperspectiveinwhichthoughtsmediatechange,acceptanceofthepositionintheappealcorrespondedtotheextentandvalenceofthegeneratedthoughts.Furthermore,correlationalanalysesinthesestudies,aswellasinseveralstudiesthatassessedrecipients'spontaneouslygeneratedthoughts(e.g.Friedrichetal1996,Haferetal1996),provedconsistentwiththemediationalroleofthoughts.Conclusionsconcerningcausalorderarecomplicated,however,bythefewstudiesthatreorderedthetypicalcausalsequencetestedincorrelationaldesignsandfoundthatattitudescansuccessfullymediatetheeffectsofindependentvari-ablesonthoughts(Friedrichetal1996,Maio&Olson1998).Yettheevidenceofmutualmediationthroughattitudesaswellasthoughtsdoesnotnecessarilyrepresentachallengetothecognitiveresponsemodel.Mutualmediationcouldre¯ectthesimultaneoususeofdualprocessingmodes.Peoplemayengagesimul-taneouslyineffortfulprocessing,inwhichanappealinstigatesthoughtsthatthenaffectattitudes,alongwithlesseffortfulprocessing,inwhichattitudesaredirectlyaffectedbyanappealthrough,forexample,heuristicanalyses(Chaikenetal ATTITUDECHANGE 1996a).Whenappealshaveadirecteffectonattitudes,thenthoughtsmayre¯ectanafter-the-factjusti®cationforattitudejudgments.Anotherquestioniswhethercognitiveresponsesprovideasuf®cientmodelofmediation.Munro&Ditto(1997)andZuwerink&Devine(1996)reportthatattitudechangeonprejudice-relatedissuesislinkedtomessagerecipients'sub-jectiveexperienceofaffectinadditiontothefavorabilityoftheirthoughts.Althoughitispossiblethattheinsuf®ciencyofthoughtsasamediatorstemsfrommeasurementlimitations(e.g.poorreliabilityorvalidityofcodingofthoughtprotocols),italsoispossiblethatforprejudiceandotherattitudeswithastrongaffectivebasis,changesinaffectivereactionsimpactattitudesindependentlyofcognitiveresponses.Dual-ModeProcessingModelsandSocialIn¯uenceDual-modemodelsalsocanprovideaframeworkforunderstandingattitudechangeinsocialin¯uencesettings.Forexample,carefulscrutinyofothergroupmembers'answerstoajudgmenttaskappearstodependonparticipants'moti-vationtoperformwellandtheirabilitytoconducttheirownevaluationsofthetask.Baronetal(1996)reportedthathighlymotivatedparticipantsreliedontheirownevaluationsexceptwhenjudgmentstimuliwerepresentedtoobrie¯ytoidentifythecorrectanswer;thentheyappearedtoadopttheheuristicstrategyofrelyingonothers'judgments.Incontrast,participantswhowereonlymoderatelymotivatedusedtheheuristic-likestrategyofrelyingonothers'estimatesregard-lessofwhethertheycoulddeterminethecorrectanswerthemselves.Theheuristiccuesimportantinsocialin¯uencesettingsincludenotonlyothers'judgments,butalsoaspectsofsocialinteractionandothers'self-presentation.Forexample,groupmemberswhoaremorecon®dentintheirjudgmentshavebeenfoundtobemorein¯uentialindiscussions,regardlessoftheiractualtaskaccuracy(Zar-noth&Sniezek1997).ThegroupdiscussionresearchbyKellyetal(1997)alsocanbeinterpretedfromadual-modeperspective.Whenmembersweremotivatedandabletoachieveaccuratesolutions(i.e.thetaskhadaseeminglycorrectsolution,timewassuf®cient),groupdiscussionsentailedconsiderablereasoningandargumen-tation.Furthermore,systematicreasoningwasapparentlysuccessfulinyieldingvalidsolutions;moresystematicthoughtduringdiscussionwasassociatedwithgreatersolutionaccuracy.Alsointerpretablefromthisperspectiveisthe®ndingbyShestowskyetal(1998)thatindyadicdiscussions,participantswhoweremotivatedtoengageincognitiveactivities(i.e.werehighinneedforcognition)hadthegreatestimpactongroupdecisions.Althoughnodirectevidencewasobtainedthatmotivatedparticipantsengagedincarefulanalysisduringthedis-cussionorthatthisiswhatmadethemin¯uential,participants'self-ratingssug-gestedthatthosewhovaluedthinkingactivitiespresentedmanyarguments,presentedvalidarguments,andtriedhardtobepersuasive. WOOD MotivatedProcessingandBiasCorrectionResearchhascontinuedtodocumenthowrecipients'motives(i.e.todefendself,maintaindesiredrelationswithothers,haveaccuratejudgements)instigateanddirectsystematicprocessingandyieldmorefavorableevaluationsofgoal-promotingthanhinderinginformation(Munro&Ditto1997,Zuwerink&Devine1996).Motivesalsocanyieldselectiveuseofheuristiccues,suchasrelyingonsocialconsensuswhenitprovidesadequatesupportforone'sdesiredposition(Giner-Sorolla&Chaiken1997).Avarietyoffactorsappeartomotivatebiasedprocessing,includingrecipients'broadervalues(Eagly&Kulesa1997,Maio&Olson1998,Seligmanetal1996)andself-interests(Giner-Sorolla&Chaiken1997),andattitudeissuesthatarehighlyimportantandinvolving(Zuwerink&Devine1996).However,theroleofabilityinbiasedprocessingremainsunclear.Althoughsomehavespeculatedthatbiasemergeswhenpeoplerelyontheir(presumablypredominantlyattitude-supportive)personalbeliefsandknowledgetoevaluateanissue,Bieketal(1996)foundknowledgealonetobeinsuf®cient;thoughtfulbiasedprocessingemergedonlywhenknowledgeablepeoplewerealsohighlymotivated(e.g.bystrongaffect)toholdaparticularposition.Asymmetriesalsohavebeennotedinmotivatedprocessing.Peoplesometimesrespondmoreintenselytothreateninginformationthatdiscon®rmstheirdesiredviewthantocongenialinformationthatcon®rmsit(seealsoCacioppoetal1997).Dittoetal(1998)speculatethatbecausethreateningstimuliarelikelytorequireanimmediatebehavioralresponse,itisadaptiveforpeopletoconductanobjec-tive,criticalanalysisofpreference-inconsistentinformationwhilerespondingmorepassivelytocongenialinformation(seealsoEdwards&Smith1996).Itseemslikely,however,thatpeopleuseavarietyofprocessingstrategiestomeetdefensivegoals.Forexample,astrategyofdefensiveinattentiontochallenginginformationcanexplainthe®ndingbySlater&Rouner(1996)thatpeoplemorecarefullyprocesscongenialthanthreateninghealth-relatedinformation(seeAninterestingquestionthathasemergedinthepastfewyearsiswhetherpeopleareawareofandcancounteractbiasesandshortcutsintheirinformationprocessingstrategies.Wegener&Petty(1997)suggestthatpeopleengageinbiascorrectionprocessestotheextentthattheybelievefactorsunrelatedtothetruequalitiesoftheattitudeissuehavein¯uencedtheirjudgmentsandtotheextentthattheyaremotivatedandabletocounteractthebias.Althoughanimplicitassumptionseemstobethatvalidity-seekingmotivesinstigatebiascorrection,othermotivesalsoareplausible(e.g.theself-relatedmotiveofbeinganobjective,impartialjudge,theother-relatedmotiveofconveyingthisimpressiontoothers).Inatestofbiascorrection,Pettyetal(1998)simulatedacontextinwhichitwouldbeillegitimatetorelyonheuristicrules.Participantswereinstructedtoavoidlettingtheirbiasesaboutaseeminglylikableorunlikablesourcein¯uence ATTITUDECHANGE theirjudgmentsofthesource'sproposal.Theresultwasanapparentover-correctioninwhichthedislikablesourcewasmorepersuasivethanthelikableone.Furthermore,becausethecorrectioninstructionsappearedtoaffectprocess-ingindependentlyofparticipants'motivationtoscrutinizethemessage,Pettyetal(1998)concludedthatcarefulmessagescrutinydoesnotspontaneouslyincludeattemptstocounteractpotentialbiases.Biascorrectionalsoemergesinjurytrialswhentheevidencepresentedissubsequentlyruledinadmissiblebythecourt.Inexperimentaltrialsimulations,whethersuchevidenceisdiscountedappearstodependinpartonthereasonforitbeinginadmissible.Evidenceexcludedbecauseitwasunreliableorbecauseitwaspresentedforsomepersonalmotivehadlittleeffectonjudgments,whereasevidenceexcludedforproceduralreasonscontinuedtoexertimpact(Feinetal1997,Kassin&Sommers1997).Inaddition,consistentwithWegener&Petty's(1997)model,peopleattemptingtocompensateforinadmissibleevidencehavebeenfoundtobothover-andundercorrect,dependingonwhethertheybelievetheevidencewaslikelytohaveastrongoraweakimpactontheirjudgments(Schul&Goren1997).AFFECTANDINFLUENCEEffectsofMoodSeveralmodelshavebeendevelopedtoexplaintheeffectsofmoodoninfor-mationprocessingandattitudechange.AccordingtoWegener&Petty(1996),moodeffectsvarywithelaborationlikelihood.Directeffectsofmoodonagree-mentemergethroughlow-elaborationprocesses,includingassociationofaper-suasiveappealwithpositiveornegativefeelings(e.g.classicalconditioning)anduseofheuristicrulesbasedonthosefeelings(e.g.``IfeelbadsoImustdislikeit'').Whenpeoplearemoreextensivelyprocessing,howanattitudeobjectmakesthemfeelcanserveasapersuasiveargument.Italsocanbiastheinformationconsidered,suchaswhenpeopleattendmoretomessagesthatmatchtheirmoodorwhentheyrecallsuchinformationmoreaccurately(Rusting1998).Whenamountofelaborationisatsomemiddlelevel,peoplerespondstrategicallyto``manage''moods.Happypeopleselectivelyprocessinordertomaintaintheirpositivemood(e.g.attendingtoinformationthatmakesthemfeelgood),whereassadpeoplearelessselectivebecausethereisgreaterpotentialforanyactivitytobemoodenhancing.Alternatively,inthefeelings-as-informationaccount,moodssignalappropriateprocessingstrategies(Blessetal1996,Schwarz1997,Schwarz&Clore1996).Similartothediscussioninthepriorsectionconcerningasymmetricaleffectsofmotivesonprocessing,positivemoodssuggestabenignenvironmentappropriateforheuristicstrategies,whereasnegativemoodsindicateapotentialproblemthatmayrequiresystematicevaluation.Yetallnegativeemotionsdonotappearto WOOD havethesameprocessingimplications.Ottatietal(1997)foundthatanger(mar-ginally)reducedsystematicprocessingrelativetoneutralmoods,perhapsbecauseangerimpliesagonisticcontextsrequiringquickresponse.Controversyhasemergedconcerningwhethernegativeaffectenhancespro-cessingthroughmoodmanagement,assuggestedbytheelaborationmodelofWegener&Petty(1996),orthroughsignalingaproblematicsituation,assug-gestedbythefeelings-as-informationaccountofSchwarz&Clore(1996).Atpresent,theavailableempiricaldatacanbeinterpretedassupportingeitherper-spectives.Inaddition,research®ndingshavebeentakentosupportathird,affect-infusionmodel(Forgas1995),inwhichaffectinfusesthoughtsandbehaviorsprimarilywhenpeopleengageinsystematic,substantiveprocessing.AlthoughForgas(1998)concludedthatempiricaltestssupportaffectinfusion,the®ndingthatmoodsbiasanddirectsystematicprocessingofin¯uenceappealsalsoisconsistentwiththeothertwomodelsconsideredhere.FearAppealsTheeffectsoffear-inducingappealshavebeenofparticularinterestinthehealthdomain.Severaltheorieshaveoutlinedtheproximalbeliefsthroughwhichfear-inducingappealsaffectin¯uence.Accordingtoprotectionmotivationtheory(Rogers1983),appealsthatarethreateningandthatofferaneffectivemeansofcopingwiththethreatinstigatedangercontrolprocesses,whichincludeacceptingtherecommendedcopingstrategyandchangingthemaladaptivebehavior(Prentice-Dunnetal1997,Sturges&Rogers1996).Rogers&Prentice-Dunn(1997)concludedthatabouthalfofthestudiestodatethathaveperformedappro-priatetestshavesupportedthetheory'spredictionofmaximumacceptancewhenperceivedthreatandcopingarebothhigh.Theextendedparallelprocessmodel(Witte1992,1998)haslinkedin¯uencetofear-controlaswellasdanger-controlresponses.Whenthreatisgreaterthancoping,fearreactionscaninstigatemessagerejectionthroughdefensiveresponses(McMahanetal1998,Witteetal1998).Consistentwiththisview,Aspinwall&Brunhart(1996)reportedthatpeoplewhoarenotveryoptimisticabouttheirownhealthcopingstrategiesarelesslikelytoattendtothreateninghealthinformationthanthosewhoaremoreoptimistic.Becausefear-controlresponsesinhibittheadoptionofself-protectiveactssuggestedintheappeal,theycanaccountforthesometimesobtained``boomerang''shiftsinattitudesawayfrommessageswhenpeopledonotbelievetheycancopeeffectively(Rogers&Prentice-Dunn1997).Researchalsohascontinuedtoinvestigatetheeffectsoffearonmessagepro-cessingandtointegratethestudyoffearappealsintothebroadermodelsofmoodandin¯uencecoveredabove.Fear,likeotheraffectiveresponses,appearstoimpactextentofprocessing.Atlow-to-moderatelevels,fearfunctionslikeper-sonalinvolvementandincreasesprocessing(Rogers&Prentice-Dunn1997).Increasedprocessing(atleastofstrongmessages)mayexplaintheoften-obtained®ndingthatfearfacilitatesin¯uenceandacceptanceofnewcopingstrategies(e.g. ATTITUDECHANGE Dillardetal1996,Millar&Millar1996).Butfearalsocanbiasprocessinginawaythatjusti®esexistingcoping(ornoncoping)behaviors(Bieketal1996).Athighlevels,fearappearstoreducesystematicprocessing(e.g.Jepson&Chaiken1990).Ingeneral,thenonlineareffectsoffearonprocessingarecompatiblewiththeideathatfearismultidimensionalandinvolvesbotharousalandnegativeaffect.Measurementscalesthathavebeendesignedtotaptheactivationcom-ponenthavefoundthatgreaterreportedfear(i.e.arousal)enhancesin¯uence,whereasscalesthathavebeendesignedtotapthenegativetensioncomponenthavefoundthatgreaterfear(i.e.negativeaffect)inhibitsin¯uenceorhasminimaleffect(Celuchetal1998,LaTour&Rotfeld1997).GROUPANDSELF-IDENTITYSocialidentitytheory(Tajfel1981,1982)hassparkedconsiderableinterestingroupin¯uenceaswellasinotheraspectsofgroupbehavior(seeBrewer&Brown1998).Inthesocialidentityview,whenpeoplecategorizethemselvesasaningroupmember,theingroupservesasareferenceforsocialcomparison,andpeopleadopttheprototypicingroupattitudesandbeliefsastheirown.Buildingonthisanalysis,Turner(1982,1991)proposedthatgroupsexertin¯uencethroughaspeci®cprocess,whichhecalledreferentinformationalin¯uence.Inthisview,agreementfromotherscategorizedassimilartoselfenhancesone'ssubjectivecertaintyandsuggeststhatthesharedattitudesre¯ectexternalrealityandtheobjectivetruthoftheissue.Disagreementfromsimilarothersyieldssubjectiveuncertaintyandmotivatespeopletoaddressthediscrepancythrough,forexample,mutualsocialin¯uenceorattributionalreasoningtoexplainthedisagreement.Empiricaldatafromavarietyofresearchparadigmsarecongruentwiththesocialidentityapproach.ContentanalysesbyReicher&Hopkins(1996a,b)ofpoliticalandsocialspeechesillustratethecentralroleofingroupandoutgroupde®nitionsineverydaypersuasion.Forexample,theiranalysisofanantiabortionspeechdeliveredtomedicaldoctorsrevealedthatitde®neddoctorsandanti-abortionactivistsasacommoningroupintheirconcernforothers'welfare,de®nedantiabortionactivityasconsonantwiththemedicalidentity,andde®nedabortion-rightsproponentsasaderogatedoutgroup.Otherworkhassupportedthesocialidentityclaimthatin¯uencestemsfromprototypicgroupattitudes.InthesmallgroupdiscussionsexaminedbyKamedaetal(1997),®naldecisionswerein¯uencedmoststronglyby``cognitivelycentral''groupmembers,whoseinitialbeliefsaboutthediscussiontopicoverlappedthemostwithothermembers.Thein¯uenceadvantageofprototypicmemberswasindependentofwhethertheirinitialjudgmentpreferencesplacedtheminthegroupmajorityorminority.Kamedaetal(1997)arguedthatthesharedbeliefsandknowledgeprovidedsocialvalidationforothermembers'viewsandabasisonwhichotherscouldrecognizeprototypicmembers'expertise. WOOD Theself-categorizationanalysisdiffersfromstandardpersuasionmodelsinlocatingthedeterminantsofattitudechangeinpeople'sconstructionofgroupidentityratherthanintheirunderstandingofattitudeissues.Onekindofevidencepresentedinsupportofthisanalysisisthatin¯uencevarieswithgroupmember-ship(e.g.Haslametal1995).Otherevidenceisthatthein¯uenceofagroupdoesnotdependonrecipientslearningthecontentofthein¯uenceappeal(McGartyetal1994;seealsoHaslametal1996).Althoughsomeresearchhassuggestedthatingroupin¯uenceisaccompaniedbyacceptanceandlearningofthemessage(e.g.Mackieetal1992),Haslametal(1996)arguethatsuchlearningoccursafteradoptionoftheingroupposition,aspeopletrytounderstandthegroupviewinordertobeaneffectivegroupmember.Theself-categorizationtheoryclaimthatgroupidentitycanhaveadirecteffectonin¯uenceiscompatiblewithaheuristic-likereasoningprocessinwhichpeoplearepersuadedbecauseoftheseemingvalidityofingrouppositions(e.g.mykindofpeoplebelieveX).Yetevidencealsoexistsforotherkindsofprocessing.Empiricalresearchhasidenti®edanumberofconditionsunderwhichgroupiden-titymotivatesrecipientstoconductasystematic,carefulevaluationandinterpre-tationoftheappeal,andattitudechangeorresistancedependsonthisevaluation.Systematicprocessinghasbeenfoundwhentheidentityofthesourcegroupisrelevanttorecipients'ownself-de®nitions(Woodetal1996),whentheissueisrelevanttorecipients'membershipgroup(Crano&Chen1998,Mackieetal1990),whenthemessagepositionisrepresentativeofingroupconsensusandthusisinformativeaboutthetruegroupnorms(vanKnippenberg&Wilke1992),andwhenthein¯uenceappealevaluatesaningroupmemberandthushasimplicationsforrecipients'self-evaluations(Budesheimetal1996).Systematicprocessingalsoemergeswhenrecipientsareunabletoprocessheuristically,suchaswhenthegrouppositionisprovidedafterthemessage(Mackieetal1992),andwhenthecontextenablescarefulanalysis,suchaswhenpeoplehavesuf®cienttimetoprocessamessagefromasalient,potentiallyimportantingroup(Hogg1996).Thus,theempiricaldatasuggestthatgroupin¯uencedoesnotoperatethroughasingleprocess.People'smotivestoalignwithordifferentiatefromsocialgroupscanyieldheuristicanalysesandotherformsofrelativelysuper®cialinformationprocessing,ortheycanyieldcareful,systematicprocessingofrelevantSocialConsensusandValidityofInformationSocialidentityandself-categorizationtheoriesofferanewperspectiveonthequestionofwhatmakesinformationin¯uential(seealsoMoscovici1976).Inmessage-basedpersuasionparadigms,strong,cogentargumentstypicallyhavebeende®nedasonesthatlinktheattitudeissuetohighlyvaluedoutcomesandthatgeneratefavorablereactionsfromrecipients,whereasweakargumentsareonesthatlinktheissuetoless-valuedoutcomesandthatgeneratemorenegativereactions(Eagly&Chaiken1993).Recentworkwithinthistradition,forexample, ATTITUDECHANGE hasexaminedhowinsertingweakargumentsintoanotherwisestrongmessagecanimpairpersuasion(Friedrichetal1996,Friedrich&Smith1998).Socialidentityperspectivesinsteadhavefocusedonthesocialdeterminantsofinformationvalidity.ThisviewwasdevelopedinoppositiontoFestinger's(1954)claimthatpeoplepreferobjectiverealitytestingandusesocialrealitytesting(e.g.comparisonwithothers'views)primarilywhenobjectivestandardsarenotavailable.Inself-categorizationtheory,realitytestingisasingleprocessinwhichpeopleusethenormativestandardsofrelevantreferencegroupstoachieveavalid,consensuallysharedunderstandingofrealitycongruentwiththeirsocialidentity(Turner1991).Inastrongstatementoftheimplicationsofthisposition,Turner&Oakesarguedthatconsensualjudgments``arerationallymorelikelytore¯ectadeepertruthabouttheworld,notbecauseagreementalwaysindicatesaccuracy,butbecausetheyhaveemergedfrom,andsurvivedprocessesofdiscussion,argument,andcollectivetesting''(1997:369).Empiricaldata,however,suggestthattherelationbetweensocialconsensusandtheapparentaccuracyofinformationiscontingentonavarietyoffactors.Subjectivevalidityemergesfromconsensusthatisestablishedthroughthecon-vergenceofindependentratherthandependentviewsandthroughvalidationbyanindividual'sown,privatecognitiveprocessing(Levine1996,Mackie&Skelly1994).Furthermore,socialconsensusappearstobemoreimpactfulforsomekindsofissuesthanforothers,presumablybecauseingroupconsensusimpliessubjec-tivevaliditymorestronglyforsomeissuesthanforothers.Forexample,majorityconsensushasanespeciallystrongimpactonjudgmentsofpersonalpreference,andlessimpactonjudgmentsofobjectivestimuli,presumablybecauseconsensusindicatespreferenceslikelytobesharedwithsimilarothers(Crano&Hannula-Bral1994,Woodetal1994).MultipleMotivesInstigatedbyGroupsInsocialidentityandself-categorizationtheories,themotiveforin¯uencederiveslargelyfromthedesiretoestablishandmaintainapositiveevaluationoftheself.Insupport,Pooletal(1998)demonstratedthatpeoplemaintainafavorableself-viewbyshiftingtheirattitudestoalignwithpositivelyvaluedgroupsandtodeviatefromnegativelyvaluedones.Attitudechangemotivatedbysocialidentityalsocanbedrivenbyotherself-relatedconcerns,suchasstrivingtobetruetooneselfandtoachieveacoherent,certainself-view(Abrams&Hogg1988).Thetripartiteanalysisofmotivespresentedinthischaptersuggestsadditionalreasonswhypeopleadoptorrejectgrouppositions(seealsoWood1999).Oneinvolvesthepositiveandnegativeoutcomesprovidedbyothers.Forexample,goingalongwithothersinordertogetalongwiththem(i.e.receivesocialrewardsandavoidpunishments)islikelytobeimportantincloserelationshipsandotheringroupsettingsinwhichsocialharmonyisvalued.Anotherpotentialmotiveforadoptingorrejectinggrouppositionsisinordertoachieveanaccurate,valid WOOD understandingofreality.Thispossibilitychallengesperspectivesthathaveasso-ciatedgroupin¯uencewithbias.Forexample,inananalysisofdemocraticpar-ticipation,Pratkanis&Turner(1996)proposedthatactive,deliberativediscussionandanalysisofpoliticalissuesoccurswhencitizensviewissuesaspersonallyrelevant,whereasuncriticalacceptanceofpoliticalsolutionsandpropagandapro-videdbyarulingeliteoccurswhencitizenswishtoassumeacertainsocialidentity(e.g.partyallegiance).However,totheextentthatsomesocialandgroupiden-tities(e.g.informedvoter,responsiblecitizen)motivatepeopletoadoptvalid,accuratepositions,thengroup-relatedmotivesarenotnecessarilyassociatedwithbiasedprocessing.OPINIONMINORITYANDMAJORITYGROUPSTheseemingparadoxthat``fewcanin¯uencemany''hascontinuedtosparkinterestinminorityin¯uence.Thisworktypicallyde®nesminoritiesassourcesadvocatinginfrequent,lowconsensuspositions.Recentresearchhasprogressedbeyondtheoriginalnotionthatminoritysourcesarein¯uentialbecausetheyelicitinformationalcon¯ictandchallengerecipients'understandingofissues(Moscovici1985).Instead,themostconsistent®ndingappearstobethatthesocialidentityofminoritiesinhibitsin¯uence.Minoritiesaremostin¯uentialwhentheirimpactisassessedon``indirect''judg-mentmeasuresÐonesonwhichrecipientsarerelativelyunawarethattheirjudg-mentscouldalignthemwiththedeviantminoritysource(Woodetal1994).Yetevenonindirectmeasures,evidenceofminorityimpactoftenfailstoemerge(Martin1998).Minoritiesseemtofacetwoimpedimentstoexertingin¯uence.Theyare,byde®nition,notimportantreferencegroups.Thussourcegroupiden-tityisunlikelytomotivatepeopletoattendtoandevaluateanappeal(DeVriesetal1996).Furthermore,evenwhenpeopleprocesstheappeal,thesource'slowconsensus,deviantpositionislikelytoyieldanegativelybiasedprocessingori-entation.Asaresult,recipientsresistin¯uence(Mugny1980,Erbetal1998)ordemonstrate``boomerang''shiftsawayfromtheminorityview(Pooletal1998,Woodetal1996).Inaseriesofelegantexperiments,Cranoandhiscolleagues(Alvaro&Crano1996,1997;Crano&Alvaro1998;Crano&Chen1998)demonstratedthattheseimpedimentstominorityin¯uencecanbesurmountedbyingroupminoritysources(seealsoDavid&Turner1996).Itappearsthatthedistinctivenessofingroupminoritypositionsencouragescarefulmessageprocessing,yetwiththelenient,open-mindedorientationtypicallyaccordedtoingroupmembers.Thedevianceoftheminorityviewattenuatesacceptanceondirectattitudemeasuresbuttheimbalanceinattitudestructurethatresultsfrommessageprocessingyieldsattitudechangetowardtheminorityviewonmeasuresindirectlyrelatedtothe ATTITUDECHANGE Severallinesofresearchhavedocumentedthekindsofthoughtrecipientsgeneratetominorityappeals.Especiallyinproblem-solvingcontextsinwhichnovelsolutionsarevalued(Nemeth1986),minoritiesencouragerecipientstothinkaboutissuesinadivergentmannerandtoconsidernovelideasandsolutionstrategies(Buteraetal1996,Erbetal1998,Gruenfeldetal1998,Nemeth&Rogers1996,Peterson&Nemeth1996).Attributionalreasoningisalsolikelytomediatein¯uence,giventhatminoritypositionsareoftenunexpectedandrequireexplanation.Research®ndingsgenerallysupporttheanalysisbyEaglyetal(1981)thatin¯uenceisimpairedwhentheadvocatedpositioncanbeattributedtoapotentiallybiasingattributeofthesourceorsituation,suchasthesource'spersonalself-interest(Moskowitz1996).Inlikemanner,in¯uenceisenhancedwhentheadvocatedpositioncanbeattributedtoexternalrealityorthetruthabouttheissue(Bohneretal1996,1998a).Alsosuggestingattributionalreasoningisthe®ndingthatminoritiesarein¯uentialwhentheyexpresstheirviewsincon-texts,suchasface-to-faceinteraction,inwhichdeviancycanincurcosts(McLeodetal1997).Advocacythensuggestsasourcewhoissuf®cientlycommittedandcertaintorisksocialrejection.Ausefulgoalforfutureresearchwillbetoidentifythespeci®caspect(s)ofminorityidentitythatmotivateagreementorresistance.Followingthethree-motiveschemeofthischapter,itmaybethattheminoritysourcehasrelevanceforrecipients'personalidentitiesandrepresents,forexample,adeviantotherfromwhomtheywishtodifferentiateoravaluedinnovatortheywishtoemulate(Woodetal1996).Anothermotiveconcernsrecipients'relationswithothers,suchasthedesiretobelenientwithingroupminoritysources(Crano&Chen1998).Finally,recipientscouldachieveanaccurateunderstandingofanissuebyagreeingwithminoritieswhosepositionsappeartore¯ectexternalreality(Bohneretal1996).Thischapterhasdrawnfromtheliteraturesonmessage-basedpersuasionandsocialin¯uencetoidentifycommonthemesinattitudechangeresearch.Inthesetraditionallyseparateareasofinvestigation,theoreticalandempiricalworkhasbeguntodelineatethemotivesunderlyingrecipients'responsestoin¯uenceappealsandthevarietyofcognitiveandaffectiveprocessesinvolvedinattitudechangeandresistance.AnnualReviewchapteronattitudestraditionallyhasfocusedonmessage-basedpersuasionresearchthatexaminesattitudesattheindividuallevel.Fromthisstandpoint,thecurrentinclusionofsocialin¯uence®ndingshighlightsthesometimesneglectedpointthatattitudesaresocialphenomena,thattheyemergefromandareembeddedinsocialinteraction.Yet,enthusiasmforaggregatingknowledgeacrossthesetwoareasofinvestigationshouldnotoverwhelmthe WOOD equallyimportantpointthateachareaisassociatedwithuniquepredictorsandprocesses.Modelsofin¯uenceviacomplexlyarguedpersuasivemessageswillneedtoaddressuniquefactorsthatarenotasimportantinthestudyofsimplemessages,includingrecipients'abilitytoengageinextensivecognitiveprocessingandtheirknowledgeaboutthemessagetopic.Inlikemanner,modelsofin¯uenceingroupsandothercomplexsocialcontextsaddressuniquefactorssuchasthelikelihoodofinformationexchangeandtheinteractionstructureinthesocialAchallengeforfutureinvestigationwillbetocontinuetodevelopmodelsofsocialandcognitiveprocessesthataresuf®cientlyinclusivetocaptureattitudechangeinthevarietyofsocialandinformationalsettingsinwhichitoccurs.Animportantaspectofthischallengeistoplacepersuasionandsocialin¯uencewithinaframeworkthatrecognizescross-culturalandethniceffects.Itisappro-priatetoendthechapterwithanoteworthyexampleofsuchanapproach,themeta-analyticsynthesisbyBond&Smith(1996)ofsocialin¯uenceexperimentsfrom17countries.Theusefulnessofintegratingcultural-levelphenomenaintoattitudetheoriesisevidentinthegreaterlevelsofconformityinexperimentsconductedinnationscharacterizedbycollectivethanbyindividualisticvalues.ACKNOWLEDGMENTSPreparationofthischapterwassupportedbyNationalScienceFoundationgrant#SBR95±14537.TheauthorthanksWilliamDCrano,AliceHEagly,DavidMatz,RichardEPetty,andRadmilaPrislinfortheirhelpfulcommentsonanearlierdraftofthechapter.VisittheAnnualReviewshomepageatwww.AnnualReviews.org.LITERATURECITEDAbramsD,HoggMA.1988.Commentsonthemotivationalstatusofself-esteeminsocialidentityandintergroupdiscrimination.Eur.J.Soc.Psychol.AjzenI.1996.Thedirectivein¯uenceofatti-tudesonbehavior.SeeGollwitzer&Bargh1996,pp.385±403AllportGW.1985.Thehistoricalbackgroundofsocialpsychology.SeeLindzey&Aron-son1985,1:1±46AlvaroEM,CranoWD.1996.Cognitiveresponsestominority-ormajority-basedcommunications:factorsthatunderlieminorityin¯uence.Br.J.Soc.Psychol.AlvaroEM,CranoWD.1997.Indirectminor-ityin¯uence:evidenceforleniencyinsourceevaluationandcounterargumenta-J.Pers.Soc.Psychol.AronsonE.1992.Thereturnoftherepressed:Dissonancetheorymakesacomeback.chol.Inq.AschSE.1940.Studiesintheprinciplesofjudgmentsandattitudes.II.Determinationofjudgmentsbygroupandbyegostan-J.Soc.Psychol.AspinwallLG,BrunhartSM.1996.Distin-guishingoptimismfromdenial:optimisticbeliefspredictattentiontohealththreats.Pers.Soc.Psychol.Bull. ATTITUDECHANGE BaldwinMW,HolmesJG.1987.Salientpri-vateaudiencesandawarenessoftheself.Pers.Soc.Psychol.Ball-RokeachSJ,LogesWE.1996.Makingchoices:mediarolesintheconstructionofvalue-choices.SeeSeligmanetal1996,pp.BaronRS,VandelloJA,BrunsmanB.1996.Theforgottenvariableinconformityresearch:impactoftaskimportanceonsocialin¯uence.J.Pers.Soc.Psychol.BetzAL,SkowronskiJ,OstromTO.1996.Sharedrealities:socialin¯uenceandstim-ulusmemory.Soc.Cogn.BiekM,WoodW,ChaikenS.1996.Workingknowledge,cognitiveprocessing,andatti-tudes:onthedeterminantsofbias.Pers.Soc.Psychol.Bull.BlantonH,CooperJ,SkurnikI,AronsonJ.1997.Whenbadthingshappentogoodfeedback:exacerbatingtheneedforself-justi®cationwithself-af®rmations.Pers.Soc.Psychol.Bull.BlessH,CloreGO,SchwarzN,GolisanoV,RabeC,WolkM.1996.Moodandtheuseofscripts:Doesahappymoodreallyleadtomindlessness?J.Pers.Soc.Psychol.BohnerG,ErbHP,ReinhardMA,FrankE.1996.Distinctivenessacrosstopicsinminorityandmajorityin¯uence:anattri-butionalanalysisandpreliminarydata.Br.J.Soc.Psychol.BohnerG,FrankE,ErbHP.1998a.Heuristicprocessingofdistinctivenessinformationinminorityandmajorityin¯uence.Eur.J.Soc.Psychol.BohnerG,RankS,ReinhardMA,EinwillerS,ErbHP.1998b.Motivationaldeterminantsofsystematicprocessing:Expectancymod-erateseffectsofdesiredcon®denceonpro-cessingeffort.Eur.J.Soc.Psychol.BondR,SmithPB.1996.Cultureandcon-formity:ameta-analysisofstudiesusingAsch's(1952,1956)linejudgmenttask.Psychol.Bull.BosveldW,KoomenW,VogelaarR.1997.Construingasocialissue:effectsonatti-tudesandthefalseconsensuseffect.Br.J.Soc.Psychol.BrewerMB,BrownRJ.1998.Intergrouprela-tions.SeeGilbertetal1998,2:554±94BudesheimTL,HoustonDA,DePaolaSJ.1996.Persuasivenessofin-groupandout-grouppoliticalmessages:thecaseofneg-ativepoliticalcampaigning.J.Pers.Soc.Psychol.BurrisCT,Harmon-JonesE,TarpleyWR.1997.``Byfaithalone'':religiousagitationandcognitivedissonance.BasicAppl.Soc.Psychol.ButeraF,MugnyG,LegrenziP,PerezJA.1996.Majorityandminorityin¯uence,taskrepresentationandinductivereasoning.Br.J.Soc.Psychol.CacioppoJT,GardnerWL,BerntsonGG.1997.Beyondbipolarconceptualizationsandmeasures:thecaseofattitudesandevaluativespace.Pers.Soc.Psychol.Rev.CacioppoJT,PettyRE,FeinsteinJ,JarvisB.1996.Dispositionaldifferencesincognitivemotivation:thelifeandtimesofindividualsvaryinginneedforcognition.Psychol.CeluchK,LustJ,ShowersL.1998.Atestofamodelofconsumers'responsestoproductmanualsafetyinformation.J.Appl.Soc.Psychol.ChaikenS,Giner-SorollaR,ChenS.1996a.Beyondaccuracy:defenseandimpressionmotivesinheuristicandsystematicinfor-mationprocessing.SeeGollwitzer&Bargh1996,pp.553±78ChaikenS,PomerantzEM,Giner-SorollaR.1995.Structuralconsistencyandattitudestrength.SeePetty&Krosnick1995,pp.ChaikenS,WoodW,EaglyAH.1996b.Prin-ciplesofpersuasion.SeeHiggins&Krug-lanski1996,pp.702±42ChenS,SchechterD,ChaikenS.1996.Gettingatthetruthorgettingalong:accuracy-ver-susimpression-motivatedheuristicandsys- WOOD tematicprocessing.J.Pers.Soc.Psychol.CialdiniRB,LevyA,HermanCP,KozlowskiLT,PettyRE.1976.Elasticshiftsofopin-ion:determinantsofdirectionanddurabil-ity.J.Pers.Soc.Psychol.CialdiniRB,TrostMR.1998.Socialin¯uence:socialnorms,conformity,andcompliance.SeeGilbertetal1998,2:151±92CooperJ,FazioRH.1984.Anewlookatdis-sonancetheory.Adv.Exp.Soc.Psychol.CowanG,HodgeC.1996.Judgmentsofhatespeech:theeffectsoftargetgroup,public-ness,andbehavioralresponsesofthetarget.J.Appl.Soc.Psychol.CranoWD,AlvaroEM.1998.Thecontext/comparisonmodelofsocialin¯uence:mechanisms,structure,andlinkagesthatunderlieindirectattitudechange.Eur.Rev.Soc.Psychol.CranoWD,ChenX.1998.Theleniencycon-tractandpersistenceofmajorityandminor-ityin¯uence.J.Pers.Soc.Psychol.CranoWD,Hannula-BralKA.1994.Context/categorizationmodelofsocialin¯uence:minorityandmajorityin¯uenceinthefor-mationofanovelresponsenorm.J.Exp.Soc.Psychol.DavidB,TurnerJC.1996.Studiesinself-categorizationandminorityconversion:Isbeingamemberoftheout-groupanadvan-Br.J.Soc.Psychol.DeutschM,GerardHB.1955.Astudyofnor-mativeandinformationalsocialin¯uencesuponindividualjudgment.J.Abnorm.Soc.Psychol.DeVriesNK,DeDreuCKW,GordijnEH,SchuurmanM.1996.Majorityandminorityin¯uence:adualroleinterpretation.Eur.Rev.Soc.Psychol.DickersonP.1997.`It'snotjustmewho'ssay-ingthis....'Thedeploymentofcitedoth-ersintelevisedpoliticaldiscourse.Br.J.Soc.Psychol.DillardJP,PlotnickCA,GodboldLC,Frei-muthVS,EdgarT.1996.ThemultipleaffectiveoutcomesofAIDSPSAs:Fearappealsdomorethanscarepeople.mun.Res.DittoPH,ScepanskyJA,MunroGD,Apano-vitchAM,LockhartLK.1998.Motivatedsensitivitytopreference-inconsistentinfor-J.Pers.Soc.Psychol.EaglyAH,ChaikenS.1993.ThePsychologyofAttitudes.Orlando,FL:HarcourtBraceEaglyAH,ChaikenS.1998.Attitudestructureandfunction.SeeGilbertetal1998,1:269±EaglyAH,ChaikenS,WoodW.1981.Anattri-butionanalysisofpersuasion.InNewDirectionsinAttributionResearch,ed.JHHarvey,WJIckes,RFKidd,3:37±62.Mah-wah,NJ:ErlbaumEaglyAH,KulesaP.1997.Attitudes,attitudestructure,andresistancetochange:impli-cationsforpersuasiononenvironmentalissues.InEnvironment,Ethics,andBehav-ior:ThePsychologyofEnvironmentalValu-ationandDegradation,ed.MHBazerman,DMMessick,AETenbrunsel,KAWade-Benzoni,pp.122±53.SanFrancisco:NewEdwardsK,SmithEE.1996.Adiscon®rma-tionbiasintheevaluationofarguments.Pers.Soc.Psychol.ErbH,BohnerG,SchmalzleK,RankS.1998.Beyondcon¯ictanddiscrepancy:cognitivebiasinminorityandmajorityin¯uence.Pers.Soc.Psychol.Bull.FabrigarLR,PriesterJR,PettyRE,WegenerDT.1998.Theimpactofattitudeaccessi-bilityonelaborationofpersuasivemessa-Pers.Soc.Psychol.Bull.FeinS,McCloskeyAL,TomlinsonTM.1997.Canthejurydisregardthatinformation?Theuseofsuspiciontoreducethepreju-dicialeffectsofpretrialpublicityandinad-missibletestimony.Pers.Soc.Psychol.FestingerL.1954.Atheoryofsocialcompar-isonprocesses.Hum.Relat.FestingerL.1957.ATheoryofCognitiveDis-sonance.Evanston,IL:RowPeterson ATTITUDECHANGE FiskeST.1998.Stereotyping,prejudice,anddiscrimination.SeeGilbertetal1998,ForgasJP.1995.Moodandjudgment:theaffectinfusionmodel(AIM).Psychol.Bull.ForgasJP.1998.Askingnicely?TheeffectsofmoodonrespondingtomoreorlesspolitePers.Soc.Psychol.Bull.FriedCB.1998.Hypocrisyandidenti®cationwithtransgressions:acaseofundetectedBasicAppl.Soc.Psychol.FriedCB,AronsonE.1995.Hypocrisy,mis-attribution,anddissonancereduction.Pers.Soc.Psychol.Bull.FriedrichJ,FetherstonhaughD,CaseyS,Gal-lagherD.1996.Argumentintegrationandattitudechange:suppressioneffectsintheintegrationofone-sidedargumentsthatvaryinpersuasiveness.Pers.Soc.Psychol.FriedrichJ,SmithP.1998.Suppressivein¯u-enceofweakargumentsinmixed-qualitymessages:anexplorationofmechanismsviaargumentrating,pretesting,andordereffects.BasicAppl.Soc.Psychol.GarstJ,BodenhausenGV.1996.``Familyval-ues''andpoliticalpersuasion:impactofkin-relatedrhetoriconreactionstopoliticalJ.Appl.Soc.Psychol.GibbonsFX,EgglestonTJ,BenthinAC.1997.Cognitivereactionstosmokingrelapse:thereciprocalrelationbetweendissonanceandJ.Pers.Soc.Psychol.GilbertDT,FiskeST,LindzeyG,eds.1998.TheHandbookofSocialPsychology,Vols.1,2Boston:McGraw-Hill.4thed.Giner-SorollaR,ChaikenS.1997.Selectiveuseofheuristicandsystematicprocessingunderdefensemotivation.Pers.Soc.Psy-chol.Bull.GollwitzerPM,BarghJA,eds.1996.ThePsy-chologyofAction:LinkingCognitionandMotivationtoBehavior.NewYork:GreenwaldAG,BanajiMR.1995.Implicitsocialcognition:attitudes,self-esteem,andPsychol.Rev.GruenfeldDH,Thomas-HuntMC,KimPH.1998.Cognitive¯exibility,communicationstrategy,andintegrativecomplexityingroups:publicversusprivatereactionstomajorityandminoritystatus.J.Exp.Soc.Psychol.HaferCL,ReynoldsK,ObertynskiMA.1996.Messagecomprehensibilityandpersuasion:effectsofcomplexlanguageincounterat-titudinalappealstolaypeople.Soc.Cogn.HardinCD,HigginsET.1996.Sharedreality:Howsocialveri®cationmakesthesubjec-tiveobjective.InHandbookofMotivationandCognition,ed.RMSorrentino,ETHig-gins,3:28±84.NewYork:GuilfordHarmon-JonesE,BrehmJW,GreenbergJ,SimonL,NelsonDE.1996.Evidencethattheproductionofaversiveconsequencesisnotnecessarytocreatecognitivedisso-J.Pers.Soc.Psychol.HaslamSA,McGartyC,TurnerJC.1996.Salientgroupmembershipsandpersuasion:theroleofsocialidentityinthevalidationofbeliefs.InWhat'sSocialAboutSocialCognition?ResearchonSociallySharedCognitioninSmallGroups,ed.JLNye,AMBrower,pp.29±56.ThousandOaks,CA:SageHaslamSA,OakesPJ,McGartyC,TurnerJC,OnoratoRS.1995.Contextualchangesintheprototypicalityofextremeandmoderateoutgroupmembers.Eur.J.Soc.Psychol.HastieR,ParkB.1986.Therelationshipbetweenmemoryandjudgmentdependsonwhetherthejudgmenttaskismemory-basedoron-line.Psychol.Rev.HigginsET.1981.``Thecommunicationgame'':implicationsforsocialcognitionandpersuasion.InSocialCognition:TheOntarioSymposium,ed.ETHiggins,CP WOOD Herman,MPZanna,1:343±92.Mahwah,NJ:ErlbaumHigginsET,KruglanskiAW,eds.1996.Psychology:HandbookofBasicPrinciples.NewYork:GuilfordHoggMA.1996.Intragroupprocesses,groupstructureandsocialidentity.InGroupsandIdentities:DevelopingtheLegacyofHenriTajfel,ed.WPRobinson,pp.65±93.Oxford,UK:Butterworth-HowardDJ.1997.Familiarphrasesasperiph-eralpersuasioncues.J.Exp.Soc.Psychol.JepsonC,ChaikenS.1990.Chronicissue-speci®cfearinhibitssystematicprocessingofpersuasivecommunications.J.Soc.Behav.Pers.JohnsonBT,EaglyAH.1989.Effectsofinvolvementonpersuasion:ameta-Psychol.Bull.JouleR,BeauvoisJ.1998.Cognitivedisso-nancetheory:aradicalview.Eur.Rev.Soc.Psychol.KamedaT,OhtsuboY,TakezawaM.1997.Centralityinsociocognitivenetworksandsocialin¯uence:anillustrationinagroupdecision-makingcontext.J.Pers.Soc.Psy-chol.KassinSM,KiechelKL.1996.Thesocialpsy-chologyoffalseconfessions:compliance,internalization,andconfabulation.Psychol.KassinSM,SommersSR.1997.Inadmissibletestimony,instructionstodisregard,andthejury:substantiveversusproceduralconsid-Pers.Soc.Psychol.Bull.KatzD.1960.Thefunctionalapproachtothestudyofattitudes.PublicOpin.Q.KellyJR,JacksonJW,Hutson-ComeauxSL.1997.Theeffectsoftimepressureandtaskdifferencesonin¯uencemodesandaccu-racyindecision-makinggroups.Pers.Soc.Psychol.Bull.KelmanHC.1958.Compliance,identi®cation,andinternalization:threeprocessesofatti-tudechange.J.Con¯ictResolut.KilleyaLA,JohnsonBT.1998.Experimentalinductionofbiasedsystematicprocessing:thedirectedthoughttechnique.Pers.Soc.Psychol.Bull.LambertAJ,CronenS,ChasteenAL,LickelB.1996.Privatevspublicexpressionsofracialprejudice.J.Exp.Soc.Psychol.LaTourMS,RotfeldHJ.1997.Therearethreatsand(maybe)fear-causedarousal:theoryandconfusionsofappealstofearandfeararousalitself.J.Advert.LavineH,SnyderM.1996.Cognitiveprocess-ingandthefunctionalmatchingeffectinpersuasion:themediatingroleofsubjectiveperceptionsofmessagequality.J.Exp.Soc.Psychol.LevineJ.1996.SolomonAsch'slegacyforgroupresearch.PresentedatAnnu.Meet.Soc.Pers.Soc.Psychol.,Sturbridge,MALibermanA,ChaikenS.1996.Thedirecteffectofpersonalrelevanceonattitudes.Pers.Soc.Psychol.Bull.LindzeyG,AronsonE,eds.1985.ofSocialPsychology,Vols.1,2NewYork:RandomHouseLundgrenSR,PrislinR.1998.Motivatedcog-nitiveprocessingandattitudechange.Pers.Soc.Psychol.Bull.MackieDM,Gastardo-ConacoMC,SkellyJJ.1992.Knowledgeoftheadvocatedpositionandtheprocessingofin-groupandout-grouppersuasivemessages.Pers.Soc.Psy-chol.Bull.MackieDM,SkellyJJ.1994.Thesocialcog-nitionanalysisofsocialin¯uence:contri-butionstotheunderstandingofpersuasionandconformity.InSocialCognition:ImpactonSocialPsychology,ed.PDevine,DHamilton,TOstrom,pp.259±89.NewYork:AcademicMackieDM,SmithER.1998.Intergrouprela-tions:insightsfromatheoreticallyintegra-tiveapproach.Psychol.Rev. ATTITUDECHANGE MackieDM,WorthLT,AsuncionAG.1990.Processingofpersuasiveingroupmessages.J.Pers.Soc.Psychol.MaioGR,BellDW,EssesVM.1996.Ambiv-alenceinpersuasion:theprocessingofmes-sagesaboutimmigrantgroups.J.Exp.Soc.Psychol.MaioGR,OlsonJM.1998.Attitudedissimu-lationandpersuasion.J.Exp.Soc.Psychol.MarshKL,Hart-O'RourkeDM,JulkaDL.1997.ThepersuasiveeffectsofverbalandnonverbalinformationinacontextofvaluePers.Soc.Psychol.Bull.MartinR.1998.Majorityandminorityin¯u-enceusingtheafterimageparadigm:aseriesofattemptedreplications.J.Exp.Soc.Psychol.McConnellAR,LeiboldJM,ShermanSJ.1997.Within-targetillusorycorrelationsandtheformationofcontext-dependentJ.Pers.Soc.Psychol.McGartyC,HaslamSA,HutchinsonKJ,TurnerJC.1994.Theeffectsofsalientgroupmembershipsonpersuasion.GroupRes.McLeodPL,BaronRS,MartiMW,YoonK.1997.Theeyeshaveit:minorityin¯uenceinface-to-faceandcomputer-mediatedgroupdiscussion.J.Appl.Psychol.McMahanS,WitteK,MeyerJ.1998.Theper-ceptionofriskmessagesregardingelectro-magnetic®elds:extendingtheextendedparallelprocessmodeltoanunknownrisk.HealthCommun.MillarMG,MillarKU.1996.Effectsofmes-sageanxietyondiseasedetectionandhealthpromotionbehaviors.BasicAppl.Soc.Psy-chol.MoscoviciS.1976.SocialIn¯uenceandSocialChange.London:AcademicMoscoviciS.1985.Socialin¯uenceandcon-formity.SeeLindzey&Aronson1985,MoskowitzGB.1996.Themediationaleffectsofattributionsandinformationprocessinginminoritysocialin¯uence.Br.J.Soc.Psy-chol.MugnyG.1980.ThePowerofMinorities.don:AcademicMunroGD,DittoPH.1997.Biasedassimila-tion,attitudepolarization,andaffectinreactionstostereotype-relevantscienti®cPers.Soc.Psychol.Bull.MurraySL,HaddockG,ZannaMP.1996.Oncreatingvalue-expressiveattitudes:anexperimentalapproach.SeeSeligmanetal1996,pp.107±33NelsonTE,KinderDR.1996.Issueframesandgroup-centrisminAmericanpublicopin-J.Polit.NelsonTE,OxleyZM,ClawsonRA.1997.Towardapsychologyofframingeffects.Polit.Behav.NemethCJ.1986.Differentialcontributionsofmajorityandminorityin¯uence.Psychol.Rev.NemethCJ,RogersJ.1996.Dissentandthesearchforinformation.Br.J.Soc.Psychol.OttatiV,TerkildsenN,HubbardC.1997.Happyfaceselicitheuristicprocessinginatelevisedimpressionformationtask:acog-nitivetuningaccount.Pers.Soc.Psychol.PetersonRS,NemethC.1996.Focusversus¯exibility:Majorityandminorityin¯uencecanbothimproveperformance.Pers.Soc.Psychol.Bull.PettyRE,KrosnickJA.1995.Strength:AntecedentsandConsequences.Mahwah,NJ:ErlbaumPettyRE,WegenerDT.1998a.Attitudechange:multiplerolesforpersuasionvari-ables.SeeGilbertetal1998,1:323±90PettyRE,WegenerDT.1998b.Matchingver-susmismatchingattitudefunctions:impli-cationsforscrutinyofpersuasivemessages.Pers.Soc.Psychol.Bull.PettyRE,WegenerDT,FabrigarLR.1997.Attitudesandattitudechange.Annu.Rev.Psychol. WOOD PettyRE,WegenerDT,WhitePH.1998.Flex-iblecorrectionprocessesinsocialjudg-ment:implicationsforpersuasion.Cogn.PfauM.1997.Theinoculationmodelofresis-tancetoin¯uence.Prog.Commun.Sci.PoolGJ,WoodW,LeckK.1998.Theself-esteemmotiveinsocialin¯uence:agree-mentwithvaluedmajoritiesanddisagreementwithderogatedminorities.Pers.Soc.Psychol.PratkanisAR,TurnerME.1996.Persuasionanddemocracy:strategiesforincreasingdeliberativeparticipationandenactingsocialchange.J.Soc.Issues.Prentice-DunnS,JonesJL,FloydDL.1997.Persuasiveappealsandthereductionofskincancerrisk:therolesofappearanceconcern,perceivedbene®tsofatan,andef®cacyinformation.J.Appl.Soc.Psychol.PriceV,TewksburyD.1997.Newsvaluesandpublicopinion:atheoreticalaccountofmediaprimingandframing.Prog.Com-mun.Sci.PriesterJR,PettyRE.1996.Thegradualthresholdmodelofambivalence:relatingthepositiveandnegativebasesofattitudestosubjectiveambivalence.J.Pers.Soc.Psychol.PrislinR,PoolGJ.1996.Behavior,conse-quences,andtheself:IsallwellthatendsPers.Soc.Psychol.Bull.PrislinR,WoodW,PoolGJ.1998.Structuralconsistencyandthedeductionofnovelfromexistingattitudes.J.Exp.Soc.Psy-chol.ReadSJ,VanmanEJ,MillerLC.1997.Con-nectionism,parallelconstraintsatisfactionprocesses,andgestaltprinciples:(re)introducingcognitivedynamicstosocialpsychology.Pers.Soc.Psychol.Rev.ReicherS,HopkinsN.1996a.Seekingin¯u-encethroughcategorizingself-categories:ananalysisofanti-abortionistrhetoric.Br.J.Soc.Psychol.ReicherS,HopkinsN.1996b.Self-categoryconstructionsinpoliticalrhetoric;ananal-ysisofThatcher'sandKinnock'sspeechesconcerningtheBritishminers'strikeEur.J.Soc.Psychol.RichardR,VanderPligtJ,DeVriesNK.1996a.AnticipatedaffectandbehavioralBasicAppl.Soc.Psychol.RichardR,VanderPligtJ,DeVriesNK.1996b.Anticipatedregretandtimeper-spective:changingsexualrisk-takingbehavior.J.Behav.Dec.Mak.RogersRW.1983.Cognitiveandphysiologicalprocessesinfearappealsandattitudechange:arevisedtheoryofprotectionmoti-vation.InSocialPsychophysiology:ASourcebook,ed.JTCacioppo,REPetty,pp.153±77.NewYork:GuilfordRogersRW,Prentice-DunnS.1997.Protectionmotivationtheory.InHandbookofHealthBehaviorResearch,ed.DGochman,1:113±32.NewYork:PlenumRomeroAA,AgnewCR,InskoCA.1996.Thecognitivemediationhypothesisrevisited:anempiricalresponsetomethodologicalandtheoreticalcriticism.Pers.Soc.Psy-chol.Bull.RothmanAJ,SaloveyP.1997.Shapingper-ceptionstomotivatehealthybehavior:theroleofmessageframing.Psychol.Bull.RothmanAJ,SchwarzN.1998.Constructingperceptionsofvulnerability:personalrele-vanceandtheuseofexperientialinforma-tioninhealthjudgments.Pers.Soc.Psychol.Bull.RuscherJB,DuvalLL.1998.Multiplecom-municatorswithuniquetargetinformationtransmitlessstereotypicalimpressions.Pers.Soc.Psychol.RustingCL.1998.Personality,mood,andthecognitiveprocessingofemotionalinfor-mation:threeconceptualframeworks.chol.Bull.SchulY,GorenH.1997.Whenstrongevidencehaslessimpactthanweakevidence:bias, ATTITUDECHANGE adjustment,andinstructionstoignore.Cogn.SchwarzN.1997.Moodsandattitudejudg-ments:acommentonFishbeinandMid-J.Consum.Psychol.SchwarzN,CloreGL.1996.Feelingsandphe-nomenalexperiences.SeeHiggins&Krug-lanski1996,pp.433±65SeligmanC,OlsonJM,ZannaMP,eds.1996.ThePsychologyofValues:TheOntarioVol.8.Mahwah,NJ:ErlbaumShestowskyD,WegenerDT,FabrigarLR.1998.Needforcognitionandinterpersonalin¯uence:individualdifferencesinimpactondyadicdecision.J.Pers.Soc.Psychol.ShultzTR,LepperMR.1996.Cognitivedis-sonancereductionasconstraintsatisfaction.Psychol.Rev.SlaterMD,RounerD.1996.Value-af®rmativeandvalue-protectiveprocessingofalcoholeducationmessagesthatincludestatisticalevidenceoranecdotes.Commun.Res.SmithER,FazioRH,CejkaMA.1996.Acces-sibleattitudesin¯uencecategorizationofmultiplycategorizableobjects.J.Pers.Soc.Psychol.SmithMB,BrunerJS,WhiteRW.1956.ionsandPersonality.NewYork:WileySmithSM,PettyRE.1996.Messageframingandpersuasion:amessageprocessinganal-Pers.Soc.Psychol.Bull.StadlerDR,BaronRS.1998.Attributionalcomplexityasamoderatorofdissonance-producedattitudechange.J.Pers.Soc.Psy-chol.SteeleCM.1988.Thepsychologyofself-af®rmation.Adv.Exp.Soc.Psychol.StoneJ,WiegandAW,CooperJ,AronsonE.1997.Whenexempli®cationfails:hypoc-risyandthemotiveforself-integrity.Pers.Soc.Psychol.SturgesJW,RogersRW.1996.Preventivehealthpsychologyfromadevelopmentalperspective:anextensionofprotectionmotivationtheory.HealthPsychol.TajfelH.1981.HumanGroupsandSocialCategories:StudiesinSocialPsychology.Cambridge,UK:CambridgeUniv.PressTajfelH.1982.Socialpsychologyofinter-grouprelations.Annu.Rev.Psychol.ThompsonMM,ZannaMP,Grif®nDW.1995.Let'snotbeindifferentabout(attitudinal)ambivalence.SeePetty&Krosnick1995,pp.361±86TurcoRM.1996.Self-referencing,qualityofargument,andpersuasion.Curr.Psychol.Dev.Learn.Pers.Soc.TurnerJC.1982.Towardsacognitiverede®-nitionofthesocialgroup.InSocialIdentityandIntergroupRelations,ed.HTajfel,pp.15±40.Cambridge,UK:CambridgeUniv.TurnerJC.1991.SocialIn¯uence.Grove,CA:Brooks/ColeTurnerJC,OakesPJ.1997.Thesociallystruc-turedmind.InTheMessageofSocialPsy-chology:PerspectivesonMindinSociety,ed.CMcGarty,SAHaslam,pp.355±73.Cambridge,MA:BlackwellTykocinskiO,HigginsET,ChaikenS.1994.Messageframing,self-discrepancies,andyieldingtopersuasivemessages:themoti-vationalsigni®canceofpsychologicalsitu-Pers.Soc.Psychol.Bull.VanderPligtJ,ZeelenbergM,VanDijkWW,DeVriesNK,RichardR.1998.Affect,atti-tudes,anddecisions:Let'sbemorespeci®c.Eur.Rev.Soc.Psychol.vanKnippenbergD,WilkeH.1992.Prototyp-icalityofargumentsandconformitytoingroupnorms.Eur.J.Soc.Psychol.VisserPS,KrosnickJA.1998.Developmentofattitudestrengthoverthelifecycle:surgeanddecline.J.Pers.Soc.Psychol.WegenerDT,PettyRE.1996.Effectsofmoodonpersuasionprocesses:enhancing,reduc-ing,andbiasingscrutinyofattitude-relevantinformation.InStrivingand WOOD Feeling:InteractionsBetweenGoals,Affect,andSelf-Regulation,ed.LLMartin,ATesser,pp.329±62.Mahwah,NJ:WegenerDT,PettyRE.1997.The¯exiblecor-rectionmodel:theroleofnaivetheoriesinbiascorrection.Adv.Exp.Soc.Psychol.WilsonTD,HodgesSD.1992.Attitudesastemporaryconstructions.InTheConstruc-tionofSocialJudgment,ed.LLMartin,ATesser,pp.37±65.Mahwah,NJ:ErlbaumWitteK.1992.Puttingthefearbackintofearappeals:theextendedparallelprocessCommun.Monogr.WitteK.1998.Fearasmotivator,fearasinhib-itor:usingtheextendedparallelprocessmodeltoexplainfearappealsuccessesandfailures.InHandbookofCommunicationandEmotion:Research,Theory,Applica-tions,andContexts,ed.PAAndersen,LKGuerro,etal,pp.423±50.SanDiego:WitteK,BerkowitzJM,CameronKA,Mc-KeonJK.1998.Preventingthespreadofgenitalwarts:usingfearappealstopromoteself-protectivebehaviors.HealthEduc.Behav.WittenbrinkB,JuddCM,ParkB.1997.Evi-denceforracialprejudiceattheimplicitlevelanditsrelationshipwithquestionnaireJ.Pers.Soc.Psychol.WoodW.1999.Motivesandmodesofpro-cessinginthesocialin¯uenceofgroups.InDual-ProcessTheoriesinSocialPsychol-ogy,ed.SChaiken,YTrope,pp.547±70.NewYork:GuilfordWoodW,LundgrenS,OuelletteJ,BuscemeS,BlackstoneT.1994.Minorityin¯uence:ameta-analyticreviewofsocialin¯uencePsychol.Bull.WoodW,Matz.1999.Attitudesvarywithsalientgroupidentity.Unpublisheddata,TexasA&MUniv.WoodW,PoolGJ,LeckK,PurvisD.1996.Self-de®nition,defensiveprocessing,andin¯uence:thenormativeimpactofmajorityandminoritygroups.J.Pers.Soc.Psychol.ZarnothP,SniezekJA.1997.Thesocialin¯u-enceofcon®denceingroupdecisionmak-J.Exp.Soc.Psychol.ZuwerinkJR,DevinePG.1996.Attitudeimportanceandresistancetopersuasion:It'snotjustthethoughtthatcounts.J.Pers.Soc.Psychol.