<%BANNER%>

Influence of Product Design on Information Processing


PAGE 1

INFLUENCEOFPRODUCTDESIGNONINFORMATIONPROCESSINGByJOANDREAHOEGGADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2006

PAGE 2

Copyright2006byJOANDREAHOEGG

PAGE 3

Thisdissertationisdedicatedtomyparents.

PAGE 4

ivACKNOWLEDGMENTSIwouldliketothankmyadvisor,JosephW.Alba,towhomIamextremelygratefulforguidance,encouragement,andpatience.Joewasincrediblygenerousofhistime,andIwouldnothavesucceededwithouthissupport.IwouldalsoliketothankAlanCooke,RichLutz,andMikeLewisfortheirkindandvaluableassistancethroughoutmytimeattheUniversityofFlorida.IthankChrisJaniszewskiforhisexcellentadvice,andmyexternalcommitteemember,JasonColquitt,forhisinsightfulcommentsonmywork.IalsoexpressmygratitudetoMickeyandJoeBelch,whoencouragedmetopursueaPhDattheUniversityofFlorida.IamblessedwithmanysupportivefriendsbothinsideandoutsidethePhDprogram.AtFlorida,IwouldparticularlyliketothankEliseChandon,QiongWang,JuliaBelyavsky,DanRice,AnneMagi,andWouterVanhouche.Outsideoftheuniversity,IthankJackieLevy,AnnLane,andLeahThomsonformanyyearsofwonderfulfriendship.AlsoIamdeeplygratefultomyfamilyforsupportingmydecisiontoreturntoschooltopursuemyPhD.Finally,IwouldliketothankTimothySilkforbelievinginmeallthosetimesthatIdidnotbelieveinmyself,andforteachingmealwaystolookonthebrightside.Icouldnothavedonethiswithouthim.

PAGE 5

vTABLEOFCONTENTSPage ACKNOWLEDGMENTS.................................................................................................ivLISTOFTABLES...........................................................................................................viiiABSTRACT.......................................................................................................................ixCHAPTER1INTRODUCTION........................................................................................................12REVIEWOFRESEARCHONAPPEARANCE.........................................................4AestheticInfluence.......................................................................................................5EvidencefromResearchonAttractiveness...........................................................5EvidencefromOtherDomains..............................................................................7Affectversusevaluation.................................................................................7Motivatedreasoning.....................................................................................11Haloeffects..................................................................................................17Awarenessofaffectiveandmotivationalinfluences....................................17Non-AestheticInfluenceofProductAppearance.......................................................18Summary.....................................................................................................................213OVERVIEWOFEXPERIMENTS............................................................................234PILOTSTUDY...........................................................................................................26Method........................................................................................................................27Pretest..................................................................................................................27DesignandProcedure..........................................................................................28Results.........................................................................................................................30BackpackRatings................................................................................................30RunningShoeRatings.........................................................................................31Discussion...................................................................................................................325EXPERIMENT1........................................................................................................36Method........................................................................................................................36Stimuli.................................................................................................................36DesignandProcedure..........................................................................................36

PAGE 6

viResults.........................................................................................................................37Discussion...................................................................................................................386EXPERIMENT2........................................................................................................41Method........................................................................................................................42Stimuli.................................................................................................................42DesignandProcedure..........................................................................................43Results.........................................................................................................................44Discussion...................................................................................................................457EXPERIMENT3........................................................................................................48Method........................................................................................................................48Stimuli.................................................................................................................48DesignandProcedure..........................................................................................48Results.........................................................................................................................49Discussion...................................................................................................................508EXPERIMENT4........................................................................................................53Method........................................................................................................................54Stimuli.................................................................................................................54DesignandProcedure..........................................................................................55Results.........................................................................................................................56Discussion...................................................................................................................589EXPERIMENT5........................................................................................................61Method........................................................................................................................62Stimuli.................................................................................................................62DesignandProcedure..........................................................................................62Results.........................................................................................................................62Discussion...................................................................................................................6310GENERALDISCUSSION.........................................................................................65Implications................................................................................................................67Limitations..................................................................................................................68FutureResearch..........................................................................................................69APPENDIXAPRODUCTPICTURESUSEDINPILOTSTUDY..................................................71BPRODUCTPICTURESUSEDINEXPERIMENT1................................................72CPRODUCTPICTURESUSEDINEXPERIMENTS2THROUGH5......................73

PAGE 7

viiDPRODUCTDESCRIPTIONSUSEDINEXPERIMENT2,4AND5......................75EPRODUCTDESCRIPTIONSUSEDINEXPERIMENT3......................................77FPRODUCTPICTURESUSEDINEXPERIMENT4................................................79LISTOFREFERENCES...................................................................................................81BIOGRAPHICALSKETCH.............................................................................................88

PAGE 8

viiiLISTOFTABLESTable Page 4-1.PilotStudyStimuli......................................................................................................344-2.PilotStudyBackpackFeatureRatings.......................................................................354-3.PilotStudyChoiceProportionsofSuperiorBackpack..............................................354-4.PilotStudyRunningShoeFeatureRatings................................................................354-5.PilotStudyChoiceProportionsofSuperiorRunningShoe.......................................355-1.RunningShoeFeatureDescriptionsUsedinExperiment1.......................................405-2.Experiment1RunningShoeFeatureRatings.............................................................405-3.Experiment1ChoiceProportionsofSuperiorRunningShoe....................................406-1.PretestResultsforPerformancePictures....................................................................466-2.PretestResultsforFeatureDescriptions.....................................................................466-3.ResultsofExperiment2.............................................................................................466-4.ManipulationCheckforExperiment2.......................................................................477-1.DimensionsUsedinExperiment3.............................................................................517-2.ResultsofExperiment3.............................................................................................517-3.ManipulationCheckforExperiment3.......................................................................527-4.ResultsofExperiment3withHelmetReplicateRemoved........................................528-1.PretestPictureRatings................................................................................................608-2.ResultsofExperiment4.............................................................................................608-3.ManipulationCheckforExperiment4.......................................................................609-1.ResultsofExperiment5.............................................................................................64

PAGE 9

ixAbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulfillmentoftheRequirementsfortheDegreeofDoctorofPhilosophyINFLUENCEOFPRODUCTDESIGNONINFORMATIONPROCESSINGByJoAndreaHoeggAugust2006Chair:JosephW.AlbaMajorDepartment:MarketingDespitethegrowingimportanceofproductdesigninthemarketplace,relativelylittleconsumerresearchhasbeenconductedaddressingtheinfluenceofproductappearanceontheprocessingofotherproductinformation.Thecurrentinvestigationpresentstheresultsoffiveexperimentsthatpitproductappearanceagainstfunctionalfeaturedescriptionstoexaminehowconflictingvisualinformationandverbalinformationarereconciled.Inallstudiesparticipantswereaskedtocomparethequalityoftwobrandsofagivenproductbasedonareadingofallegedlyunbiasedprofessionalreviews.Theprimaryexperimentalmanipulationwasthatforsomeparticipantsthereviewswereaccompaniedbypicturesofthetwobrands.Foreachpairofproducts,onebrandwassuperiorvisuallybutwasactuallyinferiorbasedontheobjectivereview,andonebrandwasinferiorvisuallybutwasobjectivelysuperior.Insomeexperimentsthevisualsuperioritywasaesthetic;thatis,oneproductwasmoreattractivethantheother.Inotherexperimentsthevisualsuperioritywasperformancebased:oneproductlookedlikeitwouldoffersuperiorperformanceonagivendimension.Theprimaryfindingwas

PAGE 10

xthatproductappearanceexertedaninfluenceonpeoplestendencyandabilitytoevaluateproductquality,butonlywhenthevisualdifferencewasperformancebasedandnotaestheticallybased.Diagnosticityappearedtobethedrivingforce.

PAGE 11

1CHAPTER1INTRODUCTIONProductdesignisofcriticalimportanceformanagers.Inaneraofsophisticatedmanufacturingprocesses,whereitisdifficultforcompaniestogaincompetitiveadvantagesinotherareassuchaspriceorreliability(Jordan,ThomasandMcClelland1996;Kalins2003),designcanactasastrategictoolthatfirmscanuseasapointofdifferentiation(PageandHerr2002,p.133).Moreover,withdesignerstylesavailableataffordableprices(e.g.,IsaacMizrahiinTarget;TommyHilfigerinJCPenney),theaverageconsumerisnowabletopurchaseandhencedemandproductsthatarenotonlyfunctionallysound,butalsoaestheticallypleasing(MillerandAdler2003).Marketersandacademicsagreethatproductdesignisimportant,andthereexistsalargebodyofmanagerialandquantitativeresearchondesignasacomponentofthenewproductdevelopmentprocess.Consumerresearchondesign,however,hasbeensomewhatlimited.Consumerbehaviorinvestigationsintoproductappearancehaveincludedexaminationsofbehavioralresponsestoproductform(Silvera,Josephs,andGiesler2002;VeryzerandHutchinson1998),theinteractionofbrandstrengthanddesign(PageandHerr2002),theeffectsofrepeatedexposuretodesigns(CoxandCox2002),andtheroleofindividualdifferencesinresponsetodesignfeatures(Bloch,BrunelandArnold2003;HolbrookandSchindler1994).Bloch(1995)developedamodelofconsumerresponsetoproductformthatoutlinedseveralstagesinthedesignprocessfrommanagerialgoalsthroughthecreationoftheproductsformtotheconsumersresponse.Blochsmodeladdressedthepossibilityofbothcognitiveandaffective

PAGE 12

2responsestoproductform,butdidnotspeakspecificallytohowvariationsinproductdesignmightinfluencethenatureofthecognitiveoraffectivereactions.Inadiscussionofthedifferentrolesofproductdesign,CreusenandSchoormans(2005)similarlymentionedthepsychologicalrolethatproductappearancemightplaybutdidnotofferpredictionsastothenatureanddirectionoftheinfluence.Thecurrentresearchcontributestothegrowingliteratureonproductdesignbyinvestigatinghowaproductsappearancecaninfluencetheprocessingoffunctionalproductinformation.Iexaminetheextenttowhichdesignsignalsquality,andhowsuchasignalmightinterferewiththeevaluationofmoreobjectiveproductinformation.Priorconsumerresearchinproductdesignhasexaminedhowaestheticscouldinfluencequalityevaluationsintheabsenceofotherinformation(PageandHerr2002).Otherworkhasalsolookedatdifferencesbetweenvisualandverbalinformation,primarilyintermsofpictureandworduseinadvertising(e.g.,Hirschman1986;HolbrookandMoore1981).However,noresearchhasspecificallyaddressedhowtheappearanceofaproductmightaltertheevaluationofobjectiveproductinformation.Inaddition,priorresearchhasnotaddressedthepossibilityofdifferenttypesofdesigninfluence.Theresearchreportedherecomparestheinfluenceofaestheticdesign,ordesignforthesakeofbeauty,totheinfluenceofperformance-baseddesign,ordesignasasignalofproductperformance,andinvestigateshowthesetworolesmightexertdifferentialinfluenceonconsumerevaluations.Fiveexperimentswereconductedthatpitteddesignagainstwrittenproductinformationtoexaminehowconflictingvisualandverbalinformationwouldbereconciled.Inallstudiesparticipantswereprovidedwithobjectivereviewsabouttwo

PAGE 13

3brandsandwereaskedtoevaluatethebrandsonparticularfeaturesbasedonthereviews.Accompanyingthetworeviewswerepicturesofthebrandsthatconflictedwiththereview.Thatis,thebrandthatwasbetterintermsoffeaturedescriptionswasinferiorvisually,andthebrandthatwasworsebasedonfeaturedescriptionshadthesuperiordesign.Thevisualsuperiorityhadtwoforms.Insomestudiesthesuperiorityofthedesignwasaesthetic:oneproductwasmoreattractivethantheother.Inotherexperimentsthesuperiorityofthedesignwasperformancebased;oneproductlookedlikeitwouldoffersuperiorperformancealongcertaindimensions.Thepaperisorganizedasfollows.Chapter2providesareviewofliteratureonvisualeffectsinmarketingandpsychology,aswellasotherrelevantresearchtooutlinethescopeoftheinvestigationandinformthekeyquestions.Becausethestudiesdistinguishaestheticdesignfromperformance-baseddesigntheliteraturereviewisdividedintotwosectionsbased.Firstisareviewofaestheticeffects,wheretheinfluenceofanobjectsappearancerelatesspecificallytoitsattractiveness.Secondisareviewofperformance-basedappearanceeffects;thatis,whereanobjectsappearancecommunicatessomelevelofperformanceorquality.Chapter3providesanoverviewofthestudies.Chapters4through9presentthemethodology,dataanalysis,andfindingsofapilotstudyandfiveexperiments.Chapter10includesageneraldiscussionofthekeyresults,alongwithmarketingimplications,limitations,andavenuesforfutureresearch.

PAGE 14

4CHAPTER2REVIEWOFRESEARCHONAPPEARANCEDatingbacktoclassicaltimes,scholarshaveconsideredtheaestheticsofthevisualform(Dickie1971),mostlyattemptingtodeterminewhatisconsideredbeautifulandwhy.Inmorerecentyears,psychologistshaveattemptedtodetermineexperimentallywhatelementsofanobjectmakeitattractiveorunattractive(e.g.,Berlyne1974;Green1995;Langloisetal.2000;Martindale1984;MartindaleandMoore1988;Rhodesetal.2002;Rubenstein,LangloisandRoggman2002;Zebrowitz1997),andwithinmarketing,asmallbutgrowingnumberofresearchershavejoinedthequesttodeterminewhatdesignelementsconsumersfindattractive(e.g.,CoxandCox2002;RaghubirandGreenleaf2006;VeryzerandHutchinson1998).Separatefromtheliteratureonaestheticsperse,severalstreamsofresearchexistthatexamine,directlyorindirectly,waysinwhichappearancemightinfluencetheprocessingofotherinformation.Researchaddressingthisissuecomesprimarilyfromfaceperception,whichhasdemonstratedthatappearancecaninfluenceratingsofpersonality(AmbadyandRosenthal1993;Eaglyetal.1991),jobsuitability(HassinandTrope2000)andevenworkquality(LandyandSigall1974;MuellerandMazur1996).Intheproductrealm,thereisevidencethatevenclearlyirrelevantdesignelementslikeaninvitationcardorawebpagebackgroundcanhaveaninfluenceonproductpreference(MandelandJohnson2002;RaghubirandGreenleaf2006).Althoughsucheffectshavebeenobservedonpreferenceratherthanmoreobjectiveratings,thepresenceofinfluenceinanunequivocallyirrelevantcontextsuggeststhepotentialpowerofdesign.

PAGE 15

5Whetherthereareconditionsunderwhichdesignmightexertinfluenceoverobjectiveproductqualityjudgmentsremainslargelyunresolved.Someresearchinmarketinghasshownthattheattractivenessofaproductdesigncaninfluenceproductqualityratings,althoughtheinfluenceappearstooccuratthemargins(PageandHerr2002;YamamotoandLambert1994).Tohelpilluminatetheissue,thereviewpresentedherewillfocusonresearchinpersonandobjectappearance,butwillalsoborrowfromseveralliteraturestreamsnotcenteredonthevisualform,notablyworkinmotivatedreasoning,haloeffects,theinterplayofaffectandcognition,andaccessibility-diagnosticity.Althoughtheseliteraturesdonotspecificallyaddressappearance,theyareusefulinasmuchastheyspeaktothepotentialimpactastimuluscanhaveoninformationprocessing.Thisreviewdividesthetopicoftheinfluenceofappearanceintotwoparts.Thefirstsectionfocusesspecificallyonaesthetics,thatis,onappearanceasasourceofbeautyorpleasure.Thesecondsectionfocusesonappearanceasasignalofqualityorperformance.AestheticInfluenceEvidencefromResearchonAttractivenessThereisafairlylargebodyofresearchdemonstratingeffectsthatphysicalattractivenesscanhaveonjudgmentsofotherpeoplespersonalitytraitsandcapabilities(Eaglyetal.1991;LandyandSigall1974;Rubensteinetal.2002;Zebrowitz1997).Forexample,attractivepeopleareperceivedashavingsuperiorskillssuchasleadership,socialabilityandjobcompetence(Langloisetal.2000).Thereisevensomeresearchsuggestingthatphysicalattractivenesscaninfluenceevaluationsofworkquality.LandyandSigall(1974)hadparticipantsrateessayssupposedlywrittenbystudentswhose

PAGE 16

6picturesaccompaniedtheessays.Despitetheobviousirrelevanceofappearancetothequalityofanessay,participantsratedboththeessaysandthewritingskillsofmoreattractivepeopleassuperiortothoseoflessattractivepeople.Toafarmorelimitedextent,attractivenesseffectshavebeenfoundintheconsumerrealm.RaghubirandGreenleaf(2006)demonstratedadramaticattractivenesseffectinanexperimentthatinvestigatedaestheticpreferenceforthegoldenratioofrectangles.Participantswerepresentedwithaninvitationtoaconcert.Theonlydifferencebetweenexperimentalconditionswasthattheinvitationwasprintedonrectangularcardsofoneoftworatios:1.38:1or1.62:1(thegoldenratio).ParticipantsreceivingtheinvitationconformingtothegoldenratioratedtheconcertasmoreharmoniousandindicatedahigherlikelihoodofpurchasingaCDoftheperformancethandidparticipantsreceivingtheinvitationwitha1.38:1ratio.PageandHerr(2002)havesuggestedthattheattractivenessofaproductdesignproducesanaffectiveresponseassociatedwithfeeling-based,relativelyautomaticevaluativeprocesses,whilethequalityoftheproductisbasedmoreonhigher-ordercognitiveevaluation.Insupportoftheirresults,PageandHerr(2002)foundthatlikingjudgmentsoccurredfasterthanqualityjudgments.Moreimportantly,productlikingwasdrivenbyreactiontotheproductsdesignandwasnotinfluencedbybrandnameorfunctionality,whereasevaluationofthequalityoftheproductseemedtointegratedesign,functionandbrandinformation.Intermsofproductevaluation,thedesignofaproductmaybeimpossibletoignore,evenwhentheevaluationoftheproductshouldbetotallyindependentofattractiveness.Someevidenceforthiscomesfromastudyonindustrialproducts

PAGE 17

7conductedbyYamamotoandLambert(1994).Theyhadbusinesspeopleinvariousorganizationalpositionsfromsalestoengineeringindicatepreferencesforvariousindustrialproducts.Productswerepresentedoncardsconsistingofproductattributespecificationsandaphotograph.ThroughaconjointanalysistechniqueYamamotoandLambertfoundthattheattractivenessoftheproductdidhaveaninfluence,evenforengineersandbuyerswhowerewellinformedaboutproductcharacteristics.Acrosssevenindustrialproductsthemoreattractiveproductwassignificantlyfavored.Thisinfluencedidnotappeartobenon-normative,however,sinceattractivenesswasoneofthelowestvariablesfoundtoaffectchoiceandtheinfluencewasonlyatthemargins.Nonetheless,aninfluenceofattractivenesswaspresentdespiteoverwhelmingself-reportsthatattractivenessshouldplaynoroleintheselectionofasuperiorproduct.Consideringthattheproductswereindustrialandtheparticipantswereexperts,thefindingthatattractivenessplayedanyroleatallissrtiking,andsuggeststhatsuchaninfluencemayexistforconsumers.EvidencefromOtherDomainsThereareseveralstreamsofresearchoutsidetheaestheticsliteratureshowingthataninitialreactiontoastimuluscaninfluenceitssubsequentevaluation.Threeareasaredirectlyrelevanttothecurrentdiscussion:theinterplayofaffectivereactionsandcognitiveevaluations,motivatedreasoning,andhaloeffects.Eachoftheseisreviewedbriefly.AffectversusevaluationManyauthorshavearguedthataffectandevaluationaredistinctconcepts,withaffectreferringtofeelingstatesthatariseduetointeractionwithastimulusandevaluationreferringmoretoacognitiveprocessofjudgmentbasedonbeliefs(Breckler

PAGE 18

8andWiggins1989;CohenandAreni1991).Bothcanoccurforagivenobject,althoughsomehavearguedthataffectivereactionscanoccurentirelywithoutcognitiveevaluations(Berkowitz1993;Zajonc1980;ZajoncandMarkus1982,1985).Inanattempttodisentangleaffectivereactionfromamorecognitiveappraisal,BrecklerandWiggins(1989)askedparticipantstoevaluateblooddonationsondifferentdimensions.Theyaskedforaseriesofrelativelycognitiveevaluationsofblooddonations,usingsemanticdifferentialscalesoftheextenttowhichdonatingbloodwasuseful,good,wise,etc.Theyalsoaskedaseriesoffeelings-basedquestionsregardinghowpeoplefeltwhentheygaveblood.Thekeyfindingwasamongnon-donors,whohadpositivecognitiveevaluationsofblooddonation,butnegativeaffectivereactionstowardblooddonation,indicatingthattheiraffectiveandcognitiveevaluationsweredistinct.Thereisevidencetosuggestthatanaffectivereactiontoastimulusisresistanttochange(Edwards1990)andmaybeabetterpredictorofthetypeofthoughtspeoplehaveaboutproductsthancognitiveevaluations(Phametal.2001).AssuggestedbyZajonc(1980)andEdwards(1990),astrongaffectiveresponsemaybehighlyresistanttoproductinformationthatcontradictstheindividualsfeelingabouttheproduct.Edwards(1990)examinedtheresistancetochangeofbothcognitivelyandaffectivelyformedattitudes,lookingspecificallyathowattitudesformedthroughaffectivemeansorthroughcognitivemeanscouldbealteredusingaffectiveorcognitivepersuasion.Shedemonstratedthattheeffectivenessofarational(cognitive)oremotional(affective)persuasivemessagedependedonthenatureoftheattitudesorigin.Attitudesoriginallyformedthroughanaffectiveprocesswereresistanttorationalpersuasiveattemptsandcouldonlybealteredthroughaffectivemeans,supportingZajoncs(1980)argumentthat

PAGE 19

9affectivejudgmentsarenoteasytochange.Cognitiveattitudes(thoseformedthroughrationalmeans)ontheotherhand,couldbealteredthroughbothcognitiveandaffectivemeansofpersuasion.Edwards(1990)usedprimacytomanipulatewhetheranattitudewasformedthroughaffectiveorcognitivemeans,assumingthattheattitudeformationoccurredwiththefirstpersuasiveexposure.Ifvalid,thispossibilityhasrelevancewhenconsideringtheaffectivereactiontoproductdesignversusacognitiveappraisal.Inthemarketplaceitiscommonforconsumerstoseeaproduct,throughadvertisingorstoredisplays,priortolearningaboutitsfeatures.Ifapersonweretoencouteraproductsdesignfirst,andthenlaterreadinformationaboutproductfunction,theattitudeformedduringtheobservationofthedesigncouldberesistanttocontraryinformationinthewrittendescription.Ontheotherhand,ifapersonreadtheproductinformationbeforeseeingthedesignitself,anyconflictingaffectivereactiontothedesignmightstillinfluencetheattitudeformedduringthereadingoftheinformation,althoughitislikelytheinfluencewouldberelativelyweaker.Phametal.(2001)examinedtheroleofaffectivereactiontoaproductandcomparedthatreactiontoamorereasoned,cognitiveanalysis.Theypredictedthatfeelingresponses(affectivereactions)wouldbebetterpredictorsofthenumberandvalenceofthoughtspeoplehaveaboutproductsthanreason-basedresponses.Inaccordancewithmodelsofaffectiveresponse(e.g.Berkowitz1993;CohenandAreni1991),Phametal.proposedthatfeeling-basedresponsesaregenerallyfasterthanreason-basedresponses,andthataninitialaffectiveresponsepromptsthoughtgenerationthroughbothautomaticandcontrolledprocesses(p.170).Instudy1(Phametal.2001),participantsevaluated

PAGE 20

10picturesfrommagazinesusingeitherareason-basedanalysisorafeeling-basedanalysis.Theauthorsfoundthatfeelingresponseswerefasterthancognitiveresponses,suggestingthatthepicturestriggeredarelativelyautomaticaffectivereaction.Instudy2Phametal.(2001)examinedthenumberofvalencedthoughtsgeneratedthroughfeeling-basedorreason-basedanalysis.Thevalenceofsubjectsspontaneouslygeneratedthoughtswasmorestronglyrelatedtothevalenceofthefeeling-basedresponsesthanthereason-basedresponses.Phametal.(2001)alsofoundthatfeeling-basedassessmentsshowedmoreinterpersonalagreementthanreason-basedassessments.Theyconcludedthatfeeling-basedevaluationsaremorenaturalandmoreconsistent.ThissuggestionisconsistentwithTakahashi(1995),whoseworkinthepsychologyofaestheticsfoundthatpeoplewereverygoodatguessingtheemotionportrayedthroughadrawing,andalsowithresearchinfacialattractivenessshowingthatpeoplegenerallyagreeintheirratingsofattractiveandunattractivefaces(Rubensteinetal.2002).Theaffect-as-informationmodelproposesthatpeopleusetheirfeelingsduringjudgmentbecausetheybelievetheirfeelingstoberelevanttothejudgmenttask(Schwarz2002;Schwarz&Clore,1983).Suchfeelingscouldstemfromtheproductitself,withtheproductsdesignactingasaformofmoodmanipulation(Norman2004).Avisuallypleasingdesigncouldenhancemood,causingpeopletobemorecreativeproblemsolversandmorewillingtoignoredetailsorsmallproblems.Anunattractivedesigncoulddepressmood,leadingpeopletobemoreanalyticalandheighteningtheirtendencytoexpectandaddressproblems.Accordingtothislogic,visuallypleasingproductscouldactuallyappeartoworkbetter.

PAGE 21

11ShivandFedorikhin(1999)demonstratedhowanaffectivereactiontoastimuluscandominateacognitiveevaluationwhencognitiveresourcesareconstrained.InaccordancewithBerkowitz(1993)theauthorsproposedthatuponexposuretoachoice,bothaffectiveandcognitiveprocessesmaybeengendered;theaffectivereactionoccurringrelativelyautomaticallyandthecognitiveprocessesoccurringinamorecontrolled,resource-taxingmanner.Duetothisautomaticnatureoftheaffectivereaction,whenprocessingresourcesareconstrained,onlycognitiveprocessingisimpaired;thus,affectiveprocessesdominate.Theinterplayofaffectandcognitioncanbeconsideredinthecontextofproductaesthetics.Aproductsdesignisquicklyandeasilyperceivedandmaypromptanimmediateaffectivereactionpriortoacognitiveappraisalofproductfeatures.Theaffectivereactionmaybeconsistentorinconsistentwithalatermoreconsciousorcontrolledevaluation.Theextenttowhichthelaterreasonedanalysiswouldcorrectfortheinitialreactionisuncertain.Thefeeling-basedreactionmaybedifficulttoalter(Edwards1990),or,inaccordancewithrecentdiscussionsofdualprocesses,thecognitiveappraisalmaybeabletoreviewandcorrectaninitialassessment(e.g.,KahnemanandFrederick2002).MotivatedreasoningMotivatedreasoningissaidtooccurwhenapersonaldesireforaparticularoutcomeinterfereswiththetendencyorabilitytoevaluatechoicesobjectively.Thenotionthatmotivations(hopes,fears,desires,etc.)canaffectjudgmentshasbeenstudiedanddebatedfordecades(e.g.DittoandLopez1992;Kruglanski1990;MillerandRoss1975;PyszczynskiandGreenberg1987;TetlockandLevi1982).Althoughsomehavearguedthatwhatiscalledmotivatedreasoningcaneasilybeexplainedentirelyinnon-

PAGE 22

12motivational,orpurelycognitivetermsbasedonpriorbeliefsorexpectations(e.g.,MillerandRoss1975),themotivationalcampcontinuestoproducecompellingevidencethatmotivationdoesplayaroleinjudgment(e.g.Boiney,KennedyandNye1997;DittoandLopez1993;MarkusandKunda1986).Forexample,inastudyonbusinessdecisions,Boineyetal.(1997)reportedthatpeopletendedtobiastheirjudgmentsmoreorlessasneededtosupporttheirdesiredconclusions.Ifpeoplewereactinginapurelycognitivemanner,andsimplycomingtothemostprobableresultgivenpriorbeliefsandexpectations,itisunlikelytheywouldhaveadjustedthedegreeofbiasdependingonwhatwasneededtoarriveatthedesiredconclusion.Inherreviewoftheliterature,Kunda(1990)suggestedthatmotivationandcognitioninteract,withmotivationaffectingreasoningthroughrelianceonabiasedsetofcognitiveprocessesforaccessing,constructingandevaluatingbeliefs.Kunda(1990)separatedmotivatedreasoningintotwocategories:accuracygoals,forwhichthemotiveistoarriveatanaccurateconclusionanddirectionalgoals,forwhichthemotiveistoarriveataparticulardesiredconclusion.Shearguedthatbothkindsofgoalsinfluencethechoiceofbeliefsandstrategiesappliedtoagivenproblem.Accuracygoalsleadtotheuseofbeliefsandstrategiesconsideredmostappropriateforthetask;directionalgoalsleadtotheuseofthosethatareconsideredmostlikelytoresultinthedesiredconclusion.Thus,assuggestedbyKruglanski(1996)thereisinterplayofcognitionandmotivation:Motivationdetermineswhichcognitiveprocesseswillbeusedinagivensituation.Inaproductcontext,whenonechoiceisaestheticallypreferabletoanothertheremaybeamotivation,perhapsbelowawareness,tounder-critiquetheattractiveproductandover-critiquetheunattractiveproduct,resultinginachoiceconsistentwiththe

PAGE 23

13aestheticpreference,butpossiblynon-normative.Thistypeofmotivatedreasoningwouldbedrivenbyadirectionalgoalcreatedbytheproductaesthetics.However,asnotedbyseveralauthors(Boineyetal.1997;Hsee1995;Kunda1990),peoplegenerallywanttomakerational,correctdecisions;thus,theirdesiresareconstrainedbyreasonableness.Theextenttowhichmotivationsinfluencejudgmentscanbelargelydependentonthestrengthandclarityoftheinformationtobejudged.Whenambiguityispresentmotivationmayplayalargerrole.Ambiguityofproductinformation.Hsee(1996)hasreportedthatifdecisioncriteriaaresomewhatambiguous,peoplewillusefactorsthattheyshouldnotinordertoachieveamotivatedoutcome.Inhiselasticityhypothesis(Hsee1995;1996;SchweitzerandHsee2002),Hseedefinestwotypesofjudgmentfactors,justifiablefactors,whicharerelevanttothetaskandwhoseusecanbejustified,andnon-justifiablefactors,whicharenotrelevanttothetaskbutarepersonallyappealing.Whentwooptionsarepresented,whereoneoptionissuperioronajustifiablefactorandtheotherisinferioronthejustifiablefactorbutissuperioronanunjustifiablefactor,peopleschoiceswilldependontheextenttowhichthejustifiablefactorsvaluesaredifficulttoevaluate.Thatis,whenthevaluesofthejustifiablefactorarefixedandclear,peoplewillchoosetheoptionthatissuperioronthatfactor;however,whenthelevelofthejustifiablefactorisnotclearlyindicated(e.g.,ispresentedinarange),peoplewillusetheunjustifiablefactortochoosethemorepersonallyattractiveoption.Thisoccursbecausewithoutambiguityinthejustifiablefactor,peoplehavenoexcuseforconsideringtheunjustifiablefactorintheirdecisions.Whenthejustifiablefactorispresentedinarange,peoplecandistorttheirevaluationofthisfactorinthedirectionoftheunjustifiablefactor.Thismakestheir

PAGE 24

14choiceappeartobebasedonthejustifiablefactorwhenitisactuallybasedontheunjustifiablefactor.Hsee(1995)demonstratedelasticjustificationinanexperimentusingacoverstoryofaproofreadingassignment.Participantsproofreadadsandwerepaidonapererrorbasis.Theadswereoftwoforms:furnitureads,whichwereuninterestingbuthadmoreerrorsperpage,andpersonalads,whichweremuchmoreinteresting,buthadfewererrorsperpage,andparticipantscouldchoosewhichtypeofadstoread.Thedependentmeasurewasthenumberofpagesreadfromeachtype.Whenthenumberoferrorsperpagewasfixed,participantsbehavednormativelyandchosetoreadthefurnitureads.However,whenthenumberoferrorsperpagewaspresentedinarange,proofreadersweremorelikelytochoosethepersonalads.Presumably,presentingthejustifiablefactorinarangemadeiteasiertochoosethemorepersonallyappealingbutobjectivelyinferioroption.Russoandhiscolleagues(Russo,Medvec,andMeloy1996;Russo,Meloy,andMedvec1998)usedthetermpredecisionaldistortiontodescribeaformofbiasedreasoningthatoccurswhenevaluatingproductsonattributespresentedinasequence.Theauthorsdemonstratedthatapreexistingpreferenceforoneoptioncouldleadtothedistortionofambiguousinformationtoconformtothepriorpreference.Inthetypicalpredecisionaldistortionparadigm,aninitialpreferenceforoneoftwoproductswouldbecreatedthroughsomekindofpersonalrelevancemanipulation.Thus,uponengagingintheevaluationtask,subjectswouldalreadybemotivatedtochooseaparticularproduct.Relativelyambiguousproductinformationwouldthenbeevaluatedand(inmostcases)afinalchoicewouldbemade.Resultsindicatedthatfeature

PAGE 25

15informationwasdistortedtoconformtothepriorpreference.Asnotedbytheauthors,predecisionaldistortionresultsareconsistentwithliteratureonconfirmationbias,wherepeopledistortevidencetosupportaninitialhypothesisorpriorbeliefs(e.g.DarleyandGross1983;Deighton1984;HochandHa1986;Nickerson1998).AlthoughRussoetal.(1996,1998)supportacognitiveexplanationforpredecisionaldistortion,theyacknowledgethattheiroriginalmanipulationwasmotivationalandcannotruleoutamotivationalexplanation.Ananalysisoftheirevidence,inparticulartheirsecondexperiment(Russoetal.1998),suggeststhatbothmotivationalandcognitivefactorswereoperating.Thebasicmotivatedreasoningprocessappearedtooperateinitially,withdirectionalmotivationinfluencingbeliefs,butoncetheambiguitywasremovedthemotivationalbiasdisappearedandastandardconfirmationbiasobtained.InsomecasesthedistortionofjudgmentsbasedonpriorbeliefscanbejustifiedfromaBayesianperspective(Koehler1993;Lord,RossandLepper1979;Russoetal.1998).Itmaynotalwaysbeeasytodeterminethediagnosticvalueofproductattributes,particularlywhentheyareambiguous,sogivenuncertainty,itmaybeperfectlylogicaltouseonespriorbeliefinthesuperiorityofonebrand(oronespreferenceforit)aspartofanoverallevaluation.IgnoringthispriorinformationwouldcontradictaBayesianupdatingapproachwherepriorbeliefsareupdated(butnotnecessarilydiscarded)whennewinformationispresented.Individualdifferences.Individualdifferencesmayinfluencetheformandextentofmotivatedreasoning.ThenotionthatsomepeoplearemoreorientedtowardaestheticqualitiesthanotherswasoriginallysuggestedbyLoewy(1951)andcontinuestopermeatecurrentaestheticsresearch(e.g.Blochetal.2003).Blochetal.suggestthat

PAGE 26

16peoplecanbemeasuredfortheirdegreeofcentralityofvisualproductaesthetics(CVPA).CVPAistheoveralllevelofsignificancethatvisualaestheticsholdforaparticularconsumerinhisorherrelationshipswithproducts.ConsumershighinCVPAhavegreaterconcernforvisualaestheticsandmayplacehigherweightonaestheticsasaproductdimension.Theseindividualsassignhighvaluetoproductappearanceinenhancingtheirpersonalwell-being.Theyperceivethemselvestobebetterabletorecognizeorevaluateproductdesigns,andtheyrespondmorestronglythanotherstoaestheticelements.ThelevelofconsumersCVPAmayinfluencetheextenttowhichtheyaredirectionallymotivatedbyproductaesthetics.Becausetheseindividualsexperiencestrongerreactionstoaesthetics,theymaybemorelikelytoengageinbiasedprocessingofotherproductattributes.Anotherindividualdifferencethatmaymoderatetheaestheticeffectisexpertise.Itislikelythatforexpertsinaproductcategory,theaestheticappealoftheproductdoesnotexertamotivationalinfluencebecauseknowledgeoffunctionalproductinformationiswellestablished.Expertsarelessinfluencedbyexternalfactorsandtendtoprocessinformationmoredeeply(AlbaandHutchinson1987).Ontheotherhand,themotivatedreasoningliteraturehasdemonstratedthatpeopleselectivelyretrieveinformationfrommemorythatsupportstheirdesiredgoal(Kunda1990),suggestingthatwhenfacedwithanaestheticallypleasingproductexpertsmayselectivelyrecallproductinformationsupportingtheirchoice.Indeed,recentresearchinwebsiteatmosphericshasdemonstratedthatproductchoicesbybothnovicesandexpertsareinfluencedbywebpagebackgroundprimes(MandelandJohnson2002),sotheeffectofproductdesignonexpertsisuncertain.

PAGE 27

17HaloeffectsThehaloeffect,whereglobalimpressionsofanotherpersoncanunconsciouslyinfluencesubsequentevaluationsofthatindividualspersonalitytraitsorabilitiesisawell-establishedphenomenon(e.g.Goldman,CowlesandFlorez1983;NisbettandWilson1977).Researchonhaloeffectshasmostlyfocusedonpersonperception,butithasbeendemonstratedthathaloeffectscananddoexistforproducts(e.g.BeckwithandLehman1975;HolbrookandHuber1979).HolbrookandHuber(1979)suggestthatbecauseofthecomplexityofsomeproductsattributesandthedifficultyofcalculatinganoverallevaluationusingsomethingliketheMultiattributemethod,peopleoftenworkinreverse,lettingtheiroverallaffectiveimpressionoftheproductguidetheirjudgmentsoftheproductsindividualattributes.Intermsofproductaesthetics,anattractivelyorunattractivelydesignedproductcouldcreateahalothatwouldinfluencesubsequentproductevaluation.AwarenessofaffectiveandmotivationalinfluencesRelativelylittleworkhasdirectlyaddressedtheextenttowhichmotivatedreasoningoccursbeyondconsciousawareness,althoughavarietyoffindingssuggestitdoes.Forexample,Boineyetal.(1997)foundthatthebiasexhibitedinmotivatedreasoninghadnoimpactonjudgmentconfidence;thatis,biaseddecisionmakerswerenomoreorlessconfidentthanunbiaseddecisionmakers.Alackofconfidencecouldhavesuggestedawarenessofthebias.OtherindirectevidencecomesfromKoehler(1993),whofoundthatscientistswerebiasedinevaluatingthequalityofresearchthatagreedordisagreedwiththeirpriorbeliefs.Despitedisplayinganagreementbias,therespondentsbelievedtheycouldevaluatetheevidenceinanunbiasedmannerandmoreover,stronglybelievedthatitwasimportanttodoso.Thisresearchwasfocusedonpriorbeliefsrather

PAGE 28

18thandirectionalmotivations,butitdoeshintatalackofawarenessoftheextenttowhichbiasedevaluationoccurs.Researchersinpredecisionaldistortionhavealsoclaimedthatawarenessofthesourceoftheirbiasisunlikely(Russoetal.1998).Theyarguedthatintheirparadigm,ifsomeofthesesourcesofinitialpreferencewererecognized,consumersmightdismissthemasunacceptablebasesforbrandcomparison(p.449).Haloeffectstudiesalsosuggestthatpeoplearenotawaretheyarebeinginfluenced.NisbettandWilson(1977)hadcollegestudentsevaluateapsychologyinstructorwhowasspeakinginavideotapedinterview.Halfthesubjectssawaninterviewwheretheinstructorwaswarmandapproachable,andhalfsawaninterviewwheretheinstructorseemedcoldandunfeeling.Afterwatchingthevideo,participantswereaskedtoratetheinstructoronvariouscharacteristics.Consistentwithhaloeffects,participantsratedtheindividualtraitsmorepositivelyforthewarmteacher.However,theinterestingfindingisinregardtoawareness.Mostparticipantsbelievedthattheiroveralllikingfortheteacherhadnotinfluencedtheirratings,andactuallybelievedthat,ifanything,thereverseoccurred,withtheinstructorstraitsinfluencingoverallliking(NisbettandWilson1977).Non-AestheticInfluenceofProductAppearanceIndependentofattractiveness,thereisresearchtosuggestthattheextenttowhichapersonlooksappropriateforaroleinfluencesevaluationsofobjectivecriteriasuchastheirresumeorperformance.Inastudyonphysicaltraitsandpersonalcharacteristics,HassinandTrope(2000)foundthatpeopleallowedinferencesmadeonthebasisofapersonsappearancetoinfluenceevaluationsofthepersonssuitabilityforparticularjobs.Thepicturesusedinthestudycontrolledforattractivenessandratherfocusedondifferencesinphysiognomy,ortheextenttowhichaparticularfacelookslikeitwould

PAGE 29

19possessparticularcharactertraits.Ofcourse,suchratingsarenecessarilycorrelatedwithperceptionsofattractiveness,butsinceattractivenesswascontrolledforduringpretests,itcannotexplaintheresults.HassinandTropefoundthatpeopleweregenerallyunabletoavoidusingphysicalcharacteristicsinevaluatingpersonalityorabilitytraits,despitebelievingthattheycould.Inonestudytheauthorsspecificallyinstructedparticipantstoignoreaccompanyingphotographswhenevaluatingtheirresumesaspotentialjobcandidates,andmost(60%)indicatedthattheywereabletodoso.Nonetheless,experimentalevidenceshowedthatparticipantswereinfluencedbythephotographs,suggestingphysicalcharacteristicscanplayaroleinevaluationsbeyondawarenessandperhapsbeyondattemptstocorrectforthem.Otherstudiesonfaceshavedemonstratedbiasingeffectsofappearanceindependentofattractiveness.MuellerandMazur(1996)reportthatmalemilitaryofficerswhoappeareddominant-looking(asdefinedbyahighforeheadandasquarejaw)weremorelikelytobepromotedwithinmilitaryranksthanthosewhoappearedlessdominant-looking.Todorovetal.(2005)foundeffectsofappearanceinthepoliticalreal.Theauthorshadstudyparticipantsratepicturesofcongressionalcandidatesalongsevenpersonalitydimensions,includingcompetence,intelligence,charisma,trustworthiness,honesty,likeability,andleadership.Theyfoundthatthecompetenceratingscouldpredictwith68.8%accuracythewinneroftheelection.Inaproductcontext,productscancommunicateaparticularlevelofperformancethroughtheirdesign.Certainlyadditionalfeaturesvisibletotheeyewouldinformtheconsumerofproductperformance,butevenwhenthenumberoffeaturesisequivalentacrossproducts,thedesignofthosefeaturescancommunicatedifferentinformation.

PAGE 30

20Heavierconstructionmaterialsareoftenequatedwithdurability;sizecanbeasignalofpower;asleekshapecansuggestaerodynamics.Suchdesignelementscouldinfluencetheprocessingoffunctionalproductinformationindifferentways.Performance-basedproductdesigncouldoperatemuchliketheeffectsobservedinphysiognomyexperimentsanddistorttheprocessingofothermorereliableproductinformationinaconfirmation-biastypemanner.Alternatively,theinformationcommunicatedthroughdesigncouldprovidequalitycuesthatreducetheextenttowhichotherproductinformationisweighted.Theextenttowhichdesignwouldbeusedandperhapsoverweightedasadecisioncuemaybeafunctionoftheextenttowhichitwereperceivedtobediagnostictothejudgment(FeldmanandLynch1988;SlovicandLichtenstein1971).Ifthevisualdimensionsoftheproduct,i.e.,itsdesignelements,communicateaparticularlevelofquality,itmightberelieduponattheexpenseofothercues.Unlikeahaloorconfirmatoryprocessingeffect,thissuggestsarelativelydeliberateuseofdesign,andingeneralthiscouldberelativelynormative.Inthemarketthereisoftenacorrelationbetweentheextenttowhichproductslooksliketheywillperformwellandtheextenttowhichtheyactuallydo,sofromacost/benefitperspective,usingdesignasaqualityheuristicmaybeperfectlyreasonable.However,therearecaseswhereappearanceisnotanaccuratesignalofquality,suchaswithcopycatbrands(HutchinsonandAlba1991;Loken,RossandHinkle1989).Insuchsituationsthequestioniswhetherconsumersareabletodisregarddesignandattendtotrulydiagnosticinformation.Whetherthroughdistortionorashiftindimensionweights,thepotentialimplicationsofperformance-baseddesignbiasaredramatic.Anexampleistherecent

PAGE 31

21decisionbytheUSDepartmentofDefensetochooseLockheedMartinoverBoeingasthemanufacturerofitsnewX-plane.Atanestimated$225billion,withanother$175billioninpotentialsalestoforeignallies,thecontractforthenext-generationfighterjetswasthelargestdefensecontractinUShistory.ThiscontractwasawardedafteralongcompetitionbetweenBoeingandLockheedMartin.BothcompaniesdesignedfighterjetstomeettherequirementsoftheUSNavy,suchasaccelerationandverticalliftcapability.Theclearperformancecriteriashouldhavefacilitatedanunbiasedcomparativeevaluation;however,thetwocompaniesachievedtheNavysrequirementsinverydifferentways.Oneareainwhichtheydifferedsignificantlywasaesthetics.TheLockheedMartinplanelookedaerodynamicandveryclosetotheprototypicalfighterjetshape.TheBoeingplanehadanunusualdesignthatwasdescribedbyaPentagonsourceaslookinglikeaflyingfrogwithitsmouthwideopen(Prasso2001).AlthoughthePentagonflatlydeniedthatdesignplayedanyroleinthedecisionandtherewasnoevidencethattheBoeingplanewasinfactfunctionallysuperior,themeresuggestionmadebysomeindustryexperts(citedinNovavideo2003)thatthePentagonwasinfluencedbytheappearanceoftheplanesimpliesthepotentialseriousnessofdesigninfluence.SummaryThisliteraturereviewexaminedresearchhighlightingdifferentrolesproductappearancecouldplayininformationprocessing.Thefirstsectiondiscussedwaysinwhichareactiontotheaestheticappealofaproductmightbiasproductevaluation.Evidencefromliteratureinhaloandattractivenesseffects,affectandcognition,andmotivatedreasoningseemstosupportthenotionthatproductappearancecoulddistorttheprocessingoffunctionalproductinformation,impairingconsumersabilitytomake

PAGE 32

22objectiveproductjudgments.Ontheotherhand,literaturefocusedspecificallyonproductaestheticshasthusfarsuggestedthataestheticinfluenceonqualityjudgmentsoccursprimarilyatthemargins,influencingprocessingonlyintheabsenceofconcreteinformation.Asyetitisunclearunderwhatconditionsproductaestheticsmightbiasqualityjudgmentsinthefaceofobjectiveproductinformation.Suchinfluencemaydependonseveralfactors,includingthestrengthoftheaffectivereactiontothedesign,thestrengthofthefunctionalproductinformation,orindividualdifferences.Thesecondsectionreportedresearchsuggestingthatproductappearancecouldhaveaneffectonjudgmentindependentofanaestheticinfluence.Mostofthisworkhasfocusedonevaluationsofpeoplesfaces,whererecentresearchhasindicatedthatjobperformancebypeoplewhoappeartofitaparticularprofileisjudgeddifferentlythanthatofpeoplewhodonotfittheprofile.Thesamemaybetrueofproducts.Thatis,designcouldappeartobeasourceofdiagnosticinformationaboutthequalityoftheproduct,andassuchcouldbereliedupon,evenincaseswhereitshouldnotbe.

PAGE 33

23CHAPTER3OVERVIEWOFEXPERIMENTSPreviousresearchhasdemonstratedthatproductdesigncanbeacuetoqualityandthatintheabsenceofotherinformationsuchasbrandnameorproductspecifications,productdesigncaninfluencechoice(e.g.CreusenandSchoormans2005;PageandHerr2002).However,littleproductdesignresearchhasbeenconductedonhowdesigninfluencestheprocessingofotherproductinformation,andmostpreviousstudieshavefoundafairlyhighdegreeofrationalityintheuseofdesignindecisionmaking.Thereisevidencefromresearchonfacialappearancethatappearancecanbiasevaluationofpersonalitytraitsandperformance;however,underwhatconditionsthismightintheproductrealmisunclear.Itmaybethatpeoplearedeliberateinhowtheyweight(oroverweight)design,butgiventhewealthofresearchonhowpeopleareoftenunawareofhowtheyuseavarietyofdecisioncues,itisplausiblethatdesigncouldexertanon-normativeeffectondecisionmaking.Thestudiespresentedinthenextchaptersextendthefindingsofthevisualeffectsseenwithfacesintotheproductrealm.Theexperimentsinvestigateseveralquestions:(1)doesdesignbiasinformationprocessing,(2)aretheredifferenttypesofdesignbiasesanddotheyexertdifferentialinfluence,and(3)whereinthedecisionprocessdoesdesignhavethemost/leastinfluence.Thegoaloftheinitialexperimentswastodeterminewhetherdesigninterfereswithproductqualityjudgments.Toinvestigatethisissue,thestudyemployedamodifiedversionofthepredecisionaldistortionparadigm(Russoetal.1996;1998).Inthestudies

PAGE 34

24byRussoandhiscolleagues,participantswereaskedtocomparethequalityoftwobrandsofbackpacks.Priortotheevaluationtaskparticipantsweregivenapersonalreasontohopethatonebrandwassuperiortotheother(i.e.,anexternalmotivation).Theywerethenpresentedwithdescriptionsoftheproductfeaturesforeachbrandandevaluatedthemonebyone.Overtimeabiastowardtheirinitialpreferenceemerged,despitetheirbeingaskedtoevaluateeachfeatureonanindividualbasis,ratherthanthebackpackoverall.Thepilotstudyandexperiment1ofthecurrentinvestigationusedasimilarprocedure.However,inplaceofanexternalmotivation,picturesofthedesignsoftheproductwereusedtocreatetheinitialleaningtowardoneofthetwobrands.Priortoevaluatingthefeaturedescriptions,experimentalparticipantssawpicturesofthetwobrands,andonepicturewassignificantlymoreattractivethantheother.Thesubsequentwritteninformationsuggestedthattheunattractiveproductwasinfactsuperior.Comparingtheratingsofparticipantswhoinitiallysawthepicturestotheratingsofthosewhodidnotseethepicturesmadeitpossibletodeterminewhetherproductappearancebiasedprocessingoffeatureinformation.Afterestablishingthepresenceoftheproductappearancebias,theremainderofthedissertationfocusesonhowthebiasoccursandonsomeoftheconditionsofitsinfluence.Todothis,experiments2through5usedaslightlydifferentparadigm.Thelargestchangewasthatratherthanpresentingthefeatureinformationoneattributeatatime,inexperiments2through5allattributeinformationwaspresentedatonetimeinparagraphformat.Instudies2and3theprimarycomparisonwasstillbetweenratingswithpicturesandratingswithoutpictures.Instudies4and5comparisongroupssawboth

PAGE 35

25thepicturesandthewritteninformation.Intheseexperimentsthekeymanipulationconsistedofalteringtheorderofpresentationofthevisualandverbalinformation.Comparingtheevaluationsofthosewhosawthepicturespriortothewrittenreviewstotheevaluationsofthosewhosawthepicturesafterthereviewsprovidedinsightintotheprocessunderlyingthedesigninfluenceandhelpedtoruleoutsomealternativeexplanations.

PAGE 36

26CHAPTER4PILOTSTUDYThepilotstudyusedamodifiedversionofthepredecisionaldistortionparadigm(Russoetal.1996)toinvestigatewhetheraproductsappearanceleadstobiasedprocessingofproductfeatures.InRussosexperiments,participantsweretoldtheywouldbeevaluatingtwobrandsofbackpacksandweregivenapersonalmotivationtohopethatonebrandwassuperiortotheother.Theywerethenpresentedwithdescriptionsoftheproductsfeatures,onefeatureatatime.Foreachfeature,theysawadescriptionforeachbrand.Bothdescriptionswerepresentedonthescreenatthesametimeandparticipantsindicatedwhichbackpackwasfavoredbasedoneachfeature.Afterevaluatingthetwobackpacksontheparticularfeature,participantswereaskedwhichbackpackappearedtobesuperioroverall,basedontheinformationtheyhadseenuptothatpoint.Thiswasreferredtoasthehorserace,withthebackpacksrepresentingthehorses.Participantsindicatedwhichhorsewasintheleadatthatpoint.ThepilotstudyreplicatedthedesignoftheRussopredecisionaldistortionstudies.Inplaceofapriorpersonalmotivationtochooseoneofthebrands,participantsintheexperimentalconditionwerepresentedwithpicturesofthetwobrandsbeforereadingthefeatureinformation.Oneofthebrandswassignificantlymoreattractivethantheother.Twoproductcategorieswereused:backpacks,tomirrortheRussostudies,andrunning(racing)shoes.Runningshoeswerechosenasasecondcategorybecausetheyrepresentedaproductwherelooksshouldnotmattertoaseriousrunnerwhoisshoppingforaperformanceshoe.

PAGE 37

27MethodPretestPicturesofbackpacksandrunningshoeswerepretestedtoobtainpairsofbrandsinwhichonebrandwassignificantlymoreattractivethantheother.Toprepareforthepretests,24picturesofbackpacksand24picturesofrunningshoesweretakenfrominternetwebsites.UsingPhotoshop,thepicturesweremanipulatedtoremovebrandlabelsandtomakethemuniforminsize.Tworoundsofpretestratingswereconducted.First,20participantsratedthebackpacksindividuallyonascalefrom0(veryunattractive)to100(veryattractive).Thethreemostattractiveandunattractiveproductswereselected.Next,30participantsratedtherelativeattractivenessoftheninepossiblecombinationsofpairsofunattractiveandattractiveproducts,andthepairswiththelargestdifferencewereselected.Picturesofthefinalpairsofbackpacksandrunningshoesappearinappendixa.Intermsoffeaturedescriptions,thegoalwastohaveoneproductbesuperiortotheotheroverall,butnotoverwhelminglyso.Foreachfeature,twodescriptions,oneforeachbrand,werecreatedandpretested.Forbackpacks,twooftheactualfeaturedescriptionsusedbyRussowereselectedtomirrorhisprocedureascloselyaspossible.Pretestsresultedinstimuliliststhatwhenconsideredoverall,favoredonebrandovertheother,butnotalwayswhenthefeatureswereconsideredindividually.Theunattractivepicturewasthenpairedwiththesuperiorproductlist,andtheattractivepicturewiththeinferiorlist.Becausetheattributelistsdidnotfavortheunattractiveproductoneveryfeature,theorderofpresentationoftheattributeswascontrolledsothatinitiallytherewassomeambiguityastowhichproductwassuperior.Forcingasingleattributeordercouldhave

PAGE 38

28resultedingeneralizabilityconcernsbutforthepurposesofthepilotstudyitwasconsideredworthwhile.Thefeaturelistsandorderofpresentationappearintable4-1.DesignandProcedureThestudyhadtwobetweensubjectsconditions,no-pictureandpicture.Thereweretwoproductreplicates,backpacksandrunningshoes(ordercounterbalanced).Participantsintheno-pictureconditionsawonlythewrittenfeatureinformation.Pictureparticipantssawthefeatureinformationalongwithpicturesofthetwobrands,whichremainedonthescreenthroughouttheprocedure.Thestudywasconductedoncomputer.Allinformationwasprovidedinwrittenformonthecomputerscreen.Itwasimportanttoencourageparticipantstofocusonqualityratherthanaesthetics,sobeforepresentationofthefeatureinformation,specificinstructionswereprovided.Forthebackpacks,participantsweretoldthattheywereshoppingforanewbackpacktoreplaceonetheyhadboughtpreviously.Thepreviousonehadbeenpoorqualityandthistimetheyweredeterminedtogetabetterbackpack.Theyhadnarrowedthesearchdowntotwomodels,bothofwhichhadthebasicfeaturestheywerelookingfor,sotheirdecisionhadcomedowntodeterminingwhichpackwasbetterquality.Forrunningshoes,participantsweretoldtoimaginethattheywereserious,experiencedrunnerswhoregularlyparticipatedinroadraces.Theyweretoimaginethattheywerebuyingshoesforanupcomingraceandhadnarrowedthesearchtotwomodels,bothofwhichweremadefortheirtypeoffootandbothofwhichwereequallycomfortable.Theirdecisionhadcomedowntodeterminingwhichshoewasbetterquality.

PAGE 39

29Aftertheintroductoryscreen,participantsreceivedfeatureinformationaboutthetwoavailablebrands,onefeatureatatime.EachfeaturehadadescriptionforbrandAandoneforbrandB.Forexample,forbackpacks,thefirstfeaturedescribedwasthedurabilityofthematerial,andthetwodescriptionswerethefollowing:BrandA:Hasalayerofleathersewninsidethebottomofthepacktoenhanceitsintegrity.Inaddition,alayerofwaterrepellentmaterialhasbeensewnbetweentheinteriorlayerandtheexterior.BrandB:Double-stitchedanddouble-linedfordurability.Inadditionthematerialhasbeenchemicallytreatedtoenhanceitsstrengthandwaterrepellencywhileretainingitssuppleness.Bothdescriptionsappearedonthescreenatthesametime,oneoneachsideofthescreen(counterbalanced).Forpictureparticipants,thecorrespondingpictureofthebackpacksappearedabovethefeaturedescriptions.Participantsreadbothdescriptionsandevaluatedwhichbrandwasfavoredbasedonthetwodescriptionsusinga9-pointscalewhere9=descriptionfavorsBrandA(theobjectivelysuperiorbutlessattractivebrand)and1=descriptionfavorsBrandB(theobjectivelyinferiorbutmoreattractivebrand).ConsistentwithRussoshorseraceparadigm,immediatelyafterratingthefavorabilityofagivenfeature,participantsindicatedwhichbrandtheythoughtwasleadingbasedonallfeatureinformationvieweduptothatmoment.Thiswasabinarychoice.Theythensawthenextfeaturedescriptionandcompletedthesametwotasks.Thiscontinueduntilallfivefeatureswerepresented.Afterseeingandratingallfivefeatures,participantswereasked,Youhavenowseenallthedifferentiatingfeaturesofthetwo[backpacks/runningshoes].Consideringalltheinformationyouhaveseen,which[backpack/runningshoe]doyouthinkisbetterquality?Participantsindicatedtheirresponsetothisbinarychoice.Theywerethenaskedhowconfidenttheywerethattheyhadselectedtheobjectivelysuperiorbrandandindicatedtheirresponseona9-pointscale

PAGE 40

30fromnotatallconfidenttoextremelyconfident.Afterprovidingtheirconfidencerating,theythencompletedtheentiretaskagainforthesecondproductcategory.Finallytheythencompletedaproductknowledgequestionandansweredsomegeneraldemographicquestions.Onehundredthirty-eightundergraduatestudentsparticipatedinthestudyinexchangeforcoursecredit.ResultsBackpackRatingsEighteenparticipantsfailedanattentioncheckandwereremoved,leaving120intheanalysis.Asignificantordereffectofreplicaterequiredthatonlythefirstproductreplicateforeachparticipantbeanalyzed.Table4-2presentsthemeanfeatureratingsforbackpacks.Highernumbersindicatethebeliefthattheinformationsupportstheobjectivelysuperior(lessattractive)item.Thereisasignificanteffectofthepicture,F(1,41)=5.64,p<.05.Thereisalsoanunsurprisingeffectofattribute,F(4,164)=9.69,p<.001,andnointeraction,F<1.Aftereachattributerating,participantsindicatedwhichbrandtheybelievedwasleadingusingabinarychoicemeasure.Table4-3presentstheproportionofparticipantschoosingtheobjectivelysuperiorbackpackastheleaderafterthepresentationoftherespectiveattribute.AChi-squaretestofindependencewasconductedoneachmeasure.Thethirdandfifthmeasuresweresignificant(p-values<.05),andthefourthmeasurewasdirectional,butnotsignificant(p=.10)Onthefinalchoicemeasure,participantswhodidnotseethepictureselectedthesuperiorbackpack78%ofthetime(18outof23),whereasthosewhosawthepicturesselectedthesuperior(lessattractive)backpack40%(8outof20)ofthetime,p<.05.

PAGE 41

31Notethatthisquestionwasessentiallyidenticaltothehorseracequestionafterthefifthfeature,sothechoiceproportionsarethesameasthoseofthefifthhorseracequestion.Aftercompletingthefinalbackpackchoice,participantswereaskedhowconfidenttheywerethattheyhadselectedthehigherqualitybackpack.Ona9-pointscale,no-pictureparticipantsreportedconfidenceof6.35andpictureparticipantsreportedconfidenceof6.20,indicatingnodifferenceinconfidenceasafunctionofthepresenceofthepictures,t<1.RunningShoeRatingsTable4-4presentsthemeanratingsfortherunningshoefeatures.Thereisamarginaleffectofthepicture,F(1,75)=3.98,p<.06,butincontrasttothebackpackresults,theeffectisinthereversedirection.Participantsseeingthepicturesratedthefeaturedescriptionsasmoresupportiveofthesuperiorshoethandidparticipantswhodidnotseethepictures.Aftereachfeaturerating,participantsindicatedwhichbrandtheybelievedwasleadingusingabinarychoicemeasure.Table4-5presentstheproportionofparticipantschoosingtheobjectivelysuperiorshoesastheleaderafterthepresentationoftherespectiveattribute.Similartothefeatureratings,theyareallintheoppositedirection.Thereisasignificanteffectinthereversedirectiononfeature1,2=5.74,p<.05,andadirectionaldifferenceonfeature5,2=2.64,p=.10.Forthefinalchoice,peoplewhodidnotseethepictureselectedthesuperiorrunningshoe59%ofthetime(23outof39),whilepeoplewhosawthepicturesselectedthesuperiorshoe76%(29outof38)ofthetime.Aftercompletingthefinalchoicemeasure,participantswereaskedhowconfidenttheywerethattheyhadselectedthehigherqualityrunningshoe.Ona9-pointscale,no-

PAGE 42

32pictureparticipantsreportedconfidenceof6.54andpictureparticipantsreportedconfidenceof5.95.Thisdifferencewasnotsignificant,t(75)=1.48,p>.14.DiscussionThebidirectionaleffectofthepicturesacrossthetwoproductsisperplexing.Thebackpackresultssuggestanattractivenesseffect,afindingthatisconsistentwithresearchinfacialattractiveness.Thequestioniswhytheshoeresultswenttheotherway.Onepossibilityisthatthepicturesoftheshoesdifferedonsomeadditionaldimensionbesidesattractiveness.Inthefaceperceptionliterature,physicalcharacteristicsotherthanattractivenesshavebeenfoundtoinfluencejudgmentsofcompetence,jobsuitability,andjobperformance(MuellerandMazur1996;Todorovetal.2005).Suchcharacteristicshavebeenshowntobeextremelydifficulttoignore(HassinandTrope2000).Ifphysicalcharacteristicsotherthanattractivenesscansignalindividualperformance,itisplausiblethatdesignelementsotherthanaestheticonescouldsignalacertainlevelofproductperformance.Perhapstheunattractiveshoeappearedsuperioralongoneormoreperformancedimensions,andparticipantsusedthisinformationintheirevaluations.Ifso,itmaybethatthisappearanceofperformancedominatedtheaestheticdifferenceandbiasedtheprocessingofthefeaturedescriptionsintheoppositedirection.Thispossibilityisexploredinexperiment1.Anotherpossibleexplanationforthesurprisingrunningshoeresultsisaproblemwiththestimuli.Forthebackpacks,noneofthefeaturedescriptionscouldbeconfirmedwithavisualcomparisonofthebrands.However,withtheshoes,severaloftheverbalattributesmentionedfeaturessuchasmeshupperorstripedrubbercouldbeconfirmedordisconfirmedbythepictures.Itmaybethatparticipantsdidnotbelievethe

PAGE 43

33writtendescriptionsbecausethepicturescontainedconflictinginformation.Inexperiment1thewrittendescriptionswerealteredtoavoidthisproblem.

PAGE 44

34Table4-1.PilotStudyStimuliFeatureBrandA(superior)BrandB(inferior) Backpacks 1.DurabilityHasalayerofleathersewninsidethebottomofthepacktoenhanceitsintegrity.Inaddition,alayerofwaterrepellentmaterialhasbeen sewnbetweentheinteriorlayerand theexterior.Double-stitchedanddouble-linedfordurability.Inadditionthematerialhasbeenchemicallytreatedtoenhanceitsstrengthandwaterrepellencywhileretainingitssuppleness.2.ExteriormaterialExteriormaterialismadefrommicrofiberandnylon,givingthebagbothflexibilityanddurability.Exteriormaterialisacombinationofduraliteandnylonforstrengthwiththeabilitytostretch.3.CompartmentsThemaincompartmentislargeenoughtofitbooksoralaptopcomputer.Smallerinteriorpocketsareorganizedforkeys,pens,etc.Hasalargemaincompartment,plussmallercompartmentsandseveralaccessorypocketsinsidetostoresmallitems.4.CarryingcomfortContouredloadliftershoulderstrapseasetheweightontheshoulders.Dual-densityfoamintheshoulderstrapsincreasescarryingcomfort.5.ZipperHasaheavy-weightreinforcedzippertreatedwithlubricantforsmootherrunning.InadditionaflapofmaterialtiesdownoverthetopofthebackpackHasaplasticzipper,coatedwithathinlayerofblackplasticforsmootherrunning.RunningShoes 1.UppermaterialVentilatedupperandmidsolewithmicromesh.Meshusedthroughoutformaximumbreathability.2.WeightLightweighttooptimizerunningefficiency.Lightweightwhilemaximizingenergytransferproperties.3.MidsoleLow-to-thegroundmidsoleprofile, astabletwodensitymidsole,anda flexible,lightweightupperandoutsole.Midsoleandoutsolecomponentsareengineeredtoset thefootupinanefficient,balanced positionfromheelstriketotoeoff, offeringstabilityandflexibility.3DUltralyteonepiecemidsole/outsolematerialforsuperiorfit.MidfootTorsionsupportdevicethataddsmedialsupportandstability.4.OuterSoleHigh-densitycarbonrubberincreasestractionandextendsthelifeoftheshoe.Stripedrubberoutsideforgripanddurability.5.InteriorcomfortMeshinnersleevethatstretchestoadjusttoyourfootoverlongrunsMeshupperandalacingsystemthatwrapsaroundthearchforcomfortandsupport.

PAGE 45

35Table4-2.PilotStudyBackpackFeatureRatingsTable4-3.PilotStudyChoiceProportionsofSuperiorBackpack*p=.10**p<.05Table4-4.PilotStudyRunningShoeFeatureRatingsTable4-5.PilotStudyChoiceProportionsofSuperiorRunningShoe*p=.10**p<.05Feature ConditionDurabilityExteriorMaterialCompartmentsCarryingComfortZipper NoPicture4.655.526.485.787.22 Picture3.704.854.854.106.40 ChoiceOpportunities Condition AfterFirst AttributeAfterSecondAttributeAfterThirdAttribute**AfterFourthAttribute*AfterFifthAttribute** NoPicture30%39%70%70%78% Picture30%40%35%45%40% Feature ConditionUpperMaterialWeightMidsoleOuterSoleInteriorComfort NoPicture3.674.385.626.905.41 Picture4.374.476.827.116.08 ChoiceOpportunities ConditionAfterFirstAttribute**AfterSecondAttributeAfterThirdAttributeAfterFourthAttributeAfterFifthAttribute* NoPicture31%38%54%64%59% Picture58%50%68%71%76%

PAGE 46

36CHAPTER5EXPERIMENT1Inthepilotstudythesuperiorshoewaspretestedtobelessattractive,butmayhavelookedlikeitwouldperformbetter,whichconsequentlyinfluencedfeatureratings.Thatis,ratherthananattractivenessbias,therewasaperformance-baseddesignbias.Toaddressthispossibility,experiment1usedpicturesthatwerepretestedtobeequallyattractive,butsignificantlydifferentontheextenttowhichtheylookedlikehighorlowqualityproducts.Inaddition,thewrittendescriptionswerealteredsothatacloseinspectionofthepicturescouldnotconfirmordisconfirmthewrittenfeatureinformation.MethodStimuliTheexperimentwasconductedusingrunningshoesonly.Thewrittendescriptionsoftheattributeswerealteredtocontainonlyinformationthatcouldnotbeassessedvisually.Table5-1presentstherevisedfeaturedescriptions.Newrunningshoepictureswerepretestedtocreateapairofbrandsthatwereequallyattractive(56.63and61.29ona0-100scale,p>.73)butappearedsignificantlydifferentintermsoftheextenttowhichtheylookedlikesuperiorperformingrunningshoes(relativeperformanceratingof-37.22onascalefrom-100to+100,where0meanstheshoesareequivalent,p<.002).Seeappendixbforpicturesoftheshoesusedinthestudy.DesignandProcedureIncontrasttothepilotstudy,experiment1hadthreebetweensubjectsconditions.Theno-pictureconditionfromthepilotstudywasretainedasthecontrolgroup.In

PAGE 47

37additionthereweretwolevelsofthepicturecondition:onefromthepilotstudy,wheretheinferiorpicturewaspairedwiththesuperiorattributes(hereaftercalledtheinferiorpicturecondition),andanewconditioninwhichthesuperiorpicturewaspairedwiththesuperiorattributes(hereaftercalledthesuperiorpicturecondition).Inallotherrespectsthedesignandproceduremirroredthepilotstudy.Ninety-threeundergraduatestudentsparticipatedinthestudyinexchangeforcoursecredit.ResultsTwentyparticipantswereremovedforfailinganattentioncheckmeasure.Table5-2presentsthemeanfeatureratingsfortheremaining73participantsbycondition.Asinthepilotstudy,thereisaneffectofthepicture,F(2,70)=4.04,p<.05butnopicturexattributeinteraction,F<1.Thereisapicturebiasintheinferiorpicturecondition.Peoplearebiasedtowardthesuperiorlookingshoewhenitispairedwithinferiorfeatures.Therewasnobiasintheoppositedirectionwhenthehighperformanceshoewaspairedwithgoodfeatures.Contrastsperformedonthedatarevealedthatthefeatureratingsintheno-picturecontrolconditionweresignificantlyhigherthantheinferiorpictureconditionforfeatures1,2and4(one-tailedt-tests,p<.05),andmarginallyhigheronfeature5(p<.06).Therewerenosignificantdifferencesbetweenthecontrolconditionandthesuperiorpicturecondition,althoughfeature2wasmarginallysignificantinadirectioninconsistentwithexpectations(p<.10).Theproportionsofparticipantswhochosetheobjectivelysuperiorshoeasleadingafterthepresentationofeachfeaturearepresentedintable5-3.Noneofthechoicesharesweresignificantlydifferentfromoneanother,butacrossallfivechoiceopportunities,theinferior-picturechoiceproportionwaslowerthantheno-picturecondition.Moreover,in

PAGE 48

38threeofthefivechoices,thesuperior-picturechoiceproportionwashigherthanthecontrolcondition,suggestingsomeinfluenceofthepicturesinbothdirections.Asinthepilotstudy,therewasnodifferenceinhowconfidentparticipantswerethattheyhadselectedthesuperiorshoe,F<1.Onaonetoninescalewhere1=notatallconfidentand9=completelyconfident,themeanratingswere6.18,5.73,5.52fortheno-picture,superior-picture,andinferior-pictureconditionsrespectively.DiscussionExperiment1demonstratedaperformance-baseddesignbiaswhencontrollingforattractiveness.Whentheobjectivelyinferiorshoewaspairedwithapicturethatsuggestedsuperiorperformance,participantsratingswerebiasedinthedirectionofthatshoe.Theseresultssupportthecontentionfromthepilotstudythatthebiasinthedirectionoftheunattractiveshoewastheresultofitsappearingtobefunctionallysuperior.Whenthesuperior-lookingshoewaspairedwithsuperiorfeatures,nobiaswasobservedinthedirectionofthesuperiorshoeforthefeatureratings.Nonetheless,thechoiceshares,whilenotsignificant,weredirectionallysupportiveofaninclinationtowardthesuperiorshoebeyondthatofthecontrolcondition.Thefeatureratingsareinterestingbecausethereisnojustifiablereasontorelyonthepicturewhenratingeachfeature.However,theuseoftheRussoparadigmmadetheparticipantsmajorgoaltofindthesuperiorshoeoverall.Thus,itispossiblethatthepicturewasdistortingtheratingsinanoverallhalo-likefashion,particularlywhenthefeatureswererelativelyambiguous,ratherthaninaspecificfeature-by-featurefashion.Moreover,thepictureswerepretestedtoensurethatoneshoeappearedsuperioroverall,butitispossiblethatitmighthavelookedsuperioronlyoncertaindimensions.Thislack

PAGE 49

39ofspecificitymayhaveinfluencedresults.Tofocusmorenarrowlyonindividualdimensions,subsequentexperimentsusedproductsthatwerepretestedtodominateonlyononespecificdimensionratherthanoverall.Predecisionaldistortionisaformofconfirmationbias,wheresequentiallypresentedinformationisdistortedinadesireddirection.AsacknowledgedbyRussoandhiscolleagues(Russo,Medvec,andMeloy1998),thistypeofsequentialratingformatcanleadtobiasevenforacontrolcondition.Tobetterisolateadesigneffectandensurethatitisnotanartifactoftheexperimentalformat,experiment2movesawayfromtheRussoparadigmandfocusesonasinglefeaturerating.Whenconsideredinconjunctionwiththerunningshoeresultsfromthepilotstudy,theresultsofthecurrentexperimentcallintoquestionthebackpackresultsobservedinthepreviousexperiment.Ifthedesigneffectforrunningshoeswasduetotheappearanceofperformance,couldithavebeenthatthebackpackresultswerealsoduetoaperformance-baseddesignbiasratherthananattractivenessbias?Thebackpackpretestsdidnotcontrolfortheappearanceofperformance,soitmayhavebeenthatthemoreattractivebackpackwasalsotheonethatappearedsuperioronperformance.Toexaminethispossibility,afollow-uptestwasconductedonthebackpackstimulimeasuringthecorrelationbetweenperceivedperformanceandperceivedattractiveness.Therewasamarginallysignificantpositivecorrelationbetweenattractivenessandperformance,r=.24,p=.10;thus,itisimpossibletoconcludethattheresultsforthebackpacksinthepilotstudyweredueentirelytoattractiveness.Becausethereisnostrongevidenceofbiasedprocessingduetoaesthetics,thenexttwoexperimentsfocusonlyontheappearanceofperformance.However,theissueofaestheticinfluenceremains

PAGE 50

40unresolved,andinexperiment4thepotentialforanattractivenessbiasisreexaminedinadifferentcontext.Table5-1.RunningShoeFeatureDescriptionsUsedinExperiment1FeatureBrandA(superior)BrandB(inferior) 1.WeightLightweighttooptimizerunningefficiency.Lightweightwhilemaximizingenergytransferproperties. 2.MidsolematerialEthyleneVinylAcetateairinfusedfoammidsoleprovidessuperiorcushioningwithoutsacrificingdurability.Midsoleismadeofhighqualitypolyurethanefoam,whichisdenseandstable.3.MidsoleDesignMidsolecomponentsaredesignedtokeepthefootinthemostefficientposition,maintainingbalancebutallowingittoflexnaturally.Midsolecomponentsarecompressionmolded,withsolidconstructionthatensuresthattheshoewillmaintainitsstructureandshapeovertime.4.OuterSoleDurablehigh-densitycarbonrubberoutsoleextendsthelifeoftheshoe. Blownrubbertractionenhancing outsoleincreasesgriponsoftersurfaces.5.InteriorcomfortDoublestrikeinsoleismadewithorthotic-typematerialthatadjuststothefootshapeovertimetomaximizeshockabsorption.Wellcushionedinsoleprovidesgoodarchsupport. Table5-2.Experiment1RunningShoeFeatureRatingsTable5-3.Experiment1ChoiceProportionsofSuperiorRunningShoe*p<.10Feature ConditionUpperMaterialWeightMidsoleOuterSoleInteriorComfort NoPicture5.687.686.185.457.73 SuperiorPicture5.236.88a5.855.087.38InferiorPicture4.40a6.84a5.924.24a6.96aCellmeandiffersfromcontrolgroupmeanatp<.05significance. ChoiceOpportunities Condition AfterFirst AttributeAfterSecondAttribute*AfterThirdAttributeAfterFourthAttributeAfterFifthAttribute NoPicture45%82%73%59%91% SuperiorPic58%92%88%54%85%InferiorPic36%68%82%56%80%

PAGE 51

41CHAPTER6EXPERIMENT2Experiment2marksasmallshiftintheformatofthestudies,withfourkeychanges.First,basedonthefindingsofexperiment1,experiment2focusedsolelyonavisualdifferenceinperformance,ratherthanavisualdifferenceinattractiveness.Second,whereasthepilotstudyandexperiment1madeuseofasequentialpresentationformat,withfeatureinformationpresentedoneatatime,inexperiment2allproductinformationwaspresentedatonceinparagraphformat.Participantssawtwoparagraphsonthescreen,onedescribingeachbrand.Eachparagraphprovidedinformationonseveralproductfeatures.Inaddition,forparticipantsinthepicturecondition,thewritteninformationwaspresentedsimultaneouslywiththepicture.Thesechangeseliminatedthepossibilityofdistortionoccurringovertimeinaconfirmationbiasmanner.Third,tomakethevisualdifferenceinperformanceasclearandasspecificaspossible,thestudyfocusedononlyonedimensionperproduct.Pictureswerechosentoreflectsuperiorityalongonegivendimension,ratherthanoverallquality,andparticipantswereonlyaskedtoratetheproductonthatsingledimension.Thecorrespondingparagraphssuggestedthattheinferior-lookingbrandwasinfactsuperioronthegivendimension.Finally,tobroadengeneralizability,andtoavoidsomeofthedifficultiesofthestimulifromthepilotstudy,experiment2andallsubsequentexperimentsusedfivedifferentproductcategories:cookware,stereospeakers,in-lineskates,bicyclehelmets,

PAGE 52

42andelectricmixers.Thesecategorieswereselectedbecausetheywereproductsforwhichappearancecouldpotentiallyprovideinformationaboutperformance,butcouldnotunequivocallysignalhighorlowquality.MethodStimuliPretestswereconductedtoselectpicturesandfeatureinformationforthenewsetofproducts.Thegoalforeachpairofpictureswastohaveoneproductthatappearedsuperioralongaspecificdimension,yetappearedequallyorlessattractivethantheotherproduct.Severalpicturesforeachofthefiveproductcategorieswereselectedforpretesting.Twenty-fourparticipantsratedpairsofpicturesfromthefivecategoriesonrelativeattractivenessandsuperiorityonthekeyperformancedimension.Thequestionorderwascounterbalancedacrossparticipants.ThemeasuresweresevenpointscalesfromBrandAisfarsuperiortoBrandBisfarsuperior.One-samplet-testswereconductedontheattractivenessanddimensionratings.Productschosenwerethoseforwhichtheattractivenessratingwasnotsignificantlydifferentfromthemid-pointrating(picturesareequallyattractive)buttheperformancedimensionwassignificantlydifferentfromthemid-point.Thepretestledtotheselectionofonepairofpicturesforeachproductcategory.Thefinalpicturepairsselectedappearinappendixc,andtheratingsforattractivenessandperformancealongthefocaldimension,meancenteredforeaseofcomparison,appearintable6-1.Paragraphdescriptionsofproductfeatureswerewrittenintheformatofanexpertreview,supposedlybyConsumerReports.Thewritteninformationwaspretestedsothatonthefocaldimension,onereviewwassignificantlybetterthantheother.Thereviewswereevaluatedinpairsonaseven-pointscalewhere1=BrandAisfarsuperiorand7

PAGE 53

43=BrandBisfarsuperior.Inadditioncarewastakentoensurethatthefocaldimensioncouldnotbeconfirmedbyavisualinspectionofthepictures.Theproductcategories,keydimension,andpretestratings,meancentered,appearintable6-2.Theparagraphdescriptionsusedinthestudyappearinappendixd.DesignandProcedureExperiment2hadtwobetweensubjectsconditions(pictureandno-picture)andfiveproductreplicates.Forpictureparticipants,thewrittendescriptionsthatindicatedtheproductwassuperioralongthekeydimensionwerematchedwiththepicturethatsuggestedtheproductwasinferioronthedimension.Theexperimentwasconductedoncomputer.ParticipantssawanintroductoryscreenthatinformedthemtheywouldbereadingConsumerReportsreviewsforpairsofbrandsinfivedifferentproductcategoriesandwouldbeaskedtoevaluatethequalityoftheproductsbasedonthereviews.Afterthisintroductoryscreen,theywerepresentedwiththefirstproductcategory,inarandomlyselectedorder,andsawtwoparagraphssidebysideonthescreen(counterbalanced),onedescribingeachbrand.Participantsinthepictureconditionalsosawthetwopicturesabovethecorrespondingreviews.Afterthereviewshadbeenonthescreenfor40seconds,participantswereaskedthedegreetowhichoneproductortheotherwassuperioralongthekeydimension.Thiswasindicatedona9-pointscalefrom1(BrandAisfarsuperioronthisdimension)to9(BrandBisfarsuperioronthisdimension).Aratingof9wouldmeanthattheobjectivelysuperiorbrand,basedonthewritteninformation,wassuperior.ToavoidanyevaluativeinferencesrelatedtothescalepointsthenumbersonthescalewereblackedoutsothatparticipantsonlysawnineradiobuttonsrangingfromBrandAisfarsuperiortoBrandBisfarsuperior.Thepictureandwritteninformationdidnotdisappearuntilaftersubjects

PAGE 54

44providedtheirrating.Thenextproductcategorythenappearedonthescreenandtheprocesswasrepeated.Afterprovidingratingsforthefiveproductcategories,participantsweregivenanunrelatedfillertaskthatlastedapproximatelytenminutes.Uponcompletionofthefillertaskamanipulationcheckwasconducted.Participantswereshownthepicturepairsagain.Theywereaskedtoimaginetheyhadnotreadanythingabouttheproductspreviouslyand,basedsolelyonavisualinspectionoftheproducts,toindicatewhichlookedsuperioronattractivenessandonthefocaldimension.Ofcourse,despitethefillertask,aconsistencybiaswasaseriousconcernforthepictureparticipants,sotheusefulnessofthemanipulationcheckwassomewhatlimited.Afterthemanipulationcheck,participantswereaskedquestionsregardingtheirknowledgeoftheproducts,aswellassomedemographicinformation.Theywerethenthankedanddismissed.Eighty-threeundergraduatestudentsparticipatedinthestudyinexchangeforcoursecredit.ResultsThemeanratingsforthefiveproductsappearintable6-3.Higherratingsindicatethattheobjectivelysuperiorbrandaccordingtothereviewswasratedassuperior.Asthetableindicates,bothgroupswererelativelyaccurate;bothwereabletodifferentiateandtellwhichreviewwassuperior.Nonetheless,thesignificanteffectofpicture,F(1,81)=4.28,p<.05,indicatesthatthepicturedidhaveaneffect.Acrossallreplicates,themeanratingforthepicturegroupwaslowerthanthemeanratingfortheno-picturegroup.Themanipulationchecksconfirmedthatthestimulisatisfiedthecriteriaforthestudy.Aratingof0wouldindicatethatthetwopicturesappearedequal.Anegativevaluewouldindicatethatthepicturepretestedtobesuperioralongthefocaldimensionwasperceivedassuperior,andapositivevaluewouldindicatethatthepicturewasperceived

PAGE 55

45asinferior.Thus,asuccessfulmanipulationwouldhaveattractivenessratingsequaltoorgreaterthan0andperformanceratingssignificantlylessthan0.Table6-4presentsthemeanratingsforthemanipulationcheck.Allfivepicturepairshadsignificantdifferencesinthedesireddirectionforperformanceandhadeithernodifferencesordifferencesinthereversedirectionforattractiveness.Becauseparticipantsinthepictureconditionmighthavebeenpronetoconsistencybiasintheirpictureratings,themanipulationcheckwasperformedasecondtimewithonlyno-pictureparticipants.Theresultsmirroredthosethatincludedallparticipants.DiscussionTheresultsofexperiment2suggestthatpeopleareusingthepicturesintheirevaluationoftheproducts.However,thereasonforthisisstillunclear.Itmaybethatparticipantshadaninitialreactiontothepicturesthatdistortedsubsequentevaluationofproductdimensions.Thisisconsistentwiththeresultsofpilotstudyandexperiment1.Inexperiment2,allvisualandverbalinformationwaspresentedsimultaneously,whichcouldminimizetheeffectsofaninitialreactiontothevisualstimuli.However,avisualreactiontodesignisusuallyrapidandrelativelyautomatic(PageandHerr2002),anditishighlyunlikelythatparticipantsreadtheparagraphsbeforelookingatthepictures.Thus,thenotionthattheeffectisafunctionofaninitialreactioninfluencingsubsequentevaluationinahalo-likemannerisstillarealpossibility.Iftheresultwereduetoahaloeffect,thenthereactiontothevisualinformationshouldcarryovertootherdimensionsonwhichthepicturedoesnotdominate.Anotherpossibilityisthatthepicturesarebeingrelieduponbecausetheyappeardiagnosticofquality.Ifthisisthecase,thenifparticipantswereaskedaboutdimensionsonwhichthesuperiorpicturedoesnotdominate,thereshouldbenoeffectofthepicture.

PAGE 56

46Experiment3addressesthesetwopossibilitiesbyreplicatingtheformatofexperiment2andaddingquestionsabouttwonon-focaldimensions.Table6-1.PretestResultsforPerformancePictures*p<.05**p<.01***p<.001Table6-2.PretestResultsforFeatureDescriptions*p<.07**p<.01Table6-3.ResultsofExperiment2CategoryPictureN=43NoPictureN=40 Cookware7.307.58 Stereospeakers7.307.45In-lineskates6.607.83Bicyclehelmets6.867.55Electricmixers7.938.18Average7.207.72 CategoryAttractivenessRatingKeyDimensionPerformanceRating Cookware-0.37Easeofcleaning1.08* Stereospeakers0.54Power1.08**In-lineskates-0.33Speed1.13**Bicyclehelmets0.50Comfort1.08***Electricmixers0.13Power2.00*** CategoryKeyDimensionDimensionRating CookwareEaseofcleaning-2.08** StereospeakersPower-1.71*In-lineskatesSpeed-2.62**BicyclehelmetsComfort-2.87**ElectricmixersPower-2.71**

PAGE 57

47Table6-4.ManipulationCheckforExperiment2CategoryAttractivenessRatingPerformanceRating Cookware0.07-1.04** Stereospeakers1.46*-0.83*In-lineskates0.99*-1.80***Bicyclehelmets-0.60-1.02***Electricmixers0.61-2.49*** *p<.05**p<.01***p<.001

PAGE 58

48CHAPTER7EXPERIMENT3Experiments1and2demonstratedaperformance-baseddesignbias,butthenatureoftheeffectwasunclear.Experiment1suggestedthatpeoplehadageneralreactiontothepicturethatcoloredtheirevaluationoffeatures.Experiment2focusedononedimensiononlyandconfirmedtheexistenceofadesignbias,butdidnothelptoclarifywhetheritwasduetoahaloeffectofthepictureoraresultofthepictureappearingdiagnosticofqualityonthefocaldimension.Inexperiment3participantsagainsawalltheproductinformationonthescreenatonetimeandagainevaluateddimensionalperformance.However,inthiscase,participantsevaluatedtheproductsalongthreedimensions,thefocaldimensiononwhichtheproductsdifferedvisuallyandtwoadditionaldimensionsonwhichtheproductsdidnotdiffervisually.MethodStimuliTheproductpicturesfromexperiment2wereusedinexperiment3.Twonon-focaldimensionswereaddedforeachproduct,andfeaturedescriptionswerealteredslightlytoprovidesomeinformationaboutthenon-focaldimensions.Theadditionaldimensionsappearintable7-1,andtherevisedparagraphsappearinappendixe.DesignandProcedureThedesignwassimilartoexperiment2.Thereweretwobetweensubjectsconditions,pictureandnopicture.Therewerefiveproductreplicates,butinthisstudy

PAGE 59

49eachreplicatewasaccompaniedbythreedimensionquestions,oneaboutthefocaldimensionandtwoaboutnon-focaldimensions.Participantsweregiventhesameinitialinstructionsasinexperiment2.Theywerethenpresentedwitheachproductreplicate,inrandomorder.Foreachreplicate,paragraphdescriptionsofthetwobrandsappearedonthescreen,positioncounterbalanced.Forpictureparticipants,apictureofeachbrandaccompaniedthedescriptions.After40secondsthefirstofthethreedimensionquestionsappeared.TheorderofpresentationofthethreequestionswasdeterminedusingaLatinSquare.Onceparticipantsrespondedtothefirstquestion,itdisappearedandwasreplacedbythesecondandthenthethirdquestion.Asinexperiment2,thepicturesanddescriptionsremainedonthescreenuntilallthreequestionswereanswered,atwhichtimetheydisappearedandwerereplacedbythenextproductreplicate.Afteransweringthedependentmeasuresforallfivereplicates,participantscompletedafillertaskandrespondedtothemanipulationcheckquestions.Finally,theyansweredproductknowledgeanddemographicquestions,andweredismissed.Eighty-nineundergraduateparticipantscompletedthestudy.ResultsEightparticipantsfailedanattentioncheckmeasureandwereremoved,leaving81intheanalysis.Meanratingsforthefocalandnon-focaldimensionsappearintable7-2.Thedimensionxpictureinteractionwasnotsignificantbutwasdirectionallysupportive,F(2,158)=2.19,p<.12.However,amanipulationcheckonthefocaldimensionsrevealedthatforthisgroupofparticipants,bicyclehelmetsdidnotconstituteasuccessfulmanipulationofperceivedperformanceonthekeydimension.Incontrasttotheotherfourreplicates,thetwohelmetsdidnotdiffersignificantlyintheextenttowhichthey

PAGE 60

50appearedcomfortableforthisgroupofparticipants.Themanipulationcheckratings,meancentered,appearintable7-3.Droppinghelmetsfromtheanalysisandrerunningitontheremainingfourreplicatesresultedinasignificantinteractionofdimensionandpicture,F(2,158)=3.16,p<.05.Forcompletiontherewasalsoasignificanteffectofreplicate,F(2,158)=166.39,p<.001andnoeffectofpicture,F(1,79)=2.47,p>.05.Theoverallmeanswithhelmetsremovedappearintable7-4.Followupcontrastsontherevisedmeansrevealedthatthedifferencebetweenthepictureandno-pictureconditionswassignificantforthefocaldimension(p<.05)butnotforeitherofthenon-focaldimensions.DiscussionExperiment3wasdesignedtodeterminewhetherthepictureeffectobservedinpreviousstudieswastheresultofageneralhaloortheresultofthepicturebeingperceivedasdiagnostic.Theresultsofexperiment3indicatethelatter.Thepicturesprovidedinformationaboutthefocaldimensionbutnotabouttheotherdimensions,andthepictureonlyhadaneffectontheratingsofthefocaldimension.Forthefocaldimensions,thepresenceofthepictureloweredtheratingsofthestrengthoftheverbalinformation,butfornon-focaldimensions,thepresenceofthepicturehadnoeffect.Thissuggeststhatparticipantsusedthepicturesasasourceofinformationonlywhentheyperceivedthemtobediagnostic.Thehelmetreplicatedidnotconformtothispattern,butitalsodidnotpassthemanipulationcheckforthisgroupofsubjects.Thusfar,allthestudieshavecomparedinformationprocessinginthepresenceofvisualinformationtoprocessingintheabsenceofvisualinformation.However,withsuchdesignsthepicturesareanadditionalsourceofinformation,soassuminganaveragingmodelitisnotentirelysurprisingthattherewasaneffectofthepicture.A

PAGE 61

51strongertestoftheuseofvisualinformationwouldbetocompareconditionsthathavethesameinformationbutvarytheorderinwhichtheinformationispresented.Experiments4and5involvedsuchtests.Inaddition,althoughitappearedthatthepictureswerenotoperatingashalosandcoloringevaluationofallfeatures,itmaystillhavebeenthattheirinfluencewastheresultofaninitialreactiontothekeydimension.Sincethepictureswerepretestedtodiffersignificantlyalongonedimension,arelativelyautomaticreactiontothisdifferencecouldhaveinfluencedtheratingsofthatdimension.Manipulatingtheorderofpresentationofthevisualandverbalinformationinexperiments4and5enabledatestofthispossibility.Table7-1.DimensionsUsedinExperiment3ProductFocaldimensionNon-focaldimensionNon-focaldimension CookwareEaseofcleaningFitoflidsHeatdistribution StereospeakersPowerSoundclaritySounddistributionIn-lineskatesSpeedComfortEaseofuseBicyclehelmetsComfortProtectionAerodynamicsElectricmixersPowerLoudnessEaseofuse Table7-2.ResultsofExperiment3ProductConditionFocalDimensionDimension2Dimension3 CookwarePicture7.583.794.23 NoPicture8.003.924.47StereospeakersPicture7.055.164.88NoPicture7.534.454.68In-lineskatesPicture6.633.194.40NoPicture7.553.455.58BicyclehelmetsPicture7.026.166.33NoPicture6.265.745.03ElectricmixersPicture6.654.476.07NoPicture7.713.955.58AveragePicture6.994.555.18 NoPicture7.414.305.07

PAGE 62

52Table7-3.ManipulationCheckforExperiment3CategoryAttractivenessRatingPerformanceRating Cookware0.51-0.73* Stereospeakers1.25**-1.10**In-lineskates0.12-0.68*Bicyclehelmets0.38-0.42Electricmixers0.31-2.41** *p<.05**p<.001Table7-4.ResultsofExperiment3withHelmetReplicateRemovedProductConditionFocalDimensionDimension2Dimension3 AveragePicture6.984.154.90 NoPicture7.703.945.08

PAGE 63

53CHAPTER8EXPERIMENT4Inpreviousstudiesfeatureevaluationwithoutapicturewascomparedtofeatureevaluationwithapicture,andasignificanteffectofthepresenceofthepicturewasobserved.Certainlythissuggestsaninfluenceofproductdesignbutassuminganaveragingmodel,theuseofthepictureasadditionalinformationdoesnotprovideinsightintotherolethepictureplayed.Thus,inexperiment4,thekeycomparisonsoccurbetweengroupsthatsawbothpicturesandwritteninformation.Experiment3suggestedthatthepictureshadinfluencebecausetheywereperceivedtobediagnostic.Thepictureindicatedsuperiorityalongthefocaldimension,soeventhoughparticipantswereencouragedtofocusontheConsumerReportsreviews,thedesignmayhaveprovidedwhatwasperceivedtobeusefulinformation.Thissuggestsarelativelydeliberateuseofthepictureasasourceofinformation.Anotherpossibilityisthatpeoplehadanimmediatereactiontothedimensionalinformationinthepicture,whichdistortedevaluationofthecorrespondingfeatureinformation.Toaddressthesepossibilities,participantsinexperiment4receivedbothpicturesandwritteninformation,buttheorderofpresentationofthevisualandverbalinformationwasvaried.Halftheparticipantssawthepicturesfirstandthenthereviews.Theotherhalfsawthereviewsfirst,andthenthepictures.Iftheresultswereduetoaninitialreactiontotheinformationprovidedbythepictureandaninsufficientadjustmentuponseeingtheverbalinformation,thenthepictureshouldhaveaneffectonlywhenitcomes

PAGE 64

54first.Iftheresultsofthepreviousstudieswereduetodiagnosticity,assuggestedbyexperiment3,thereshouldbenodifferenceasafunctionoftheorderofpresentation.Anotherissuethatisaddressedinexperiment4isonethataroseinthepilotstudy.Thepilotstudywasoriginallydesignedtoexamineattractivenesseffectsinaproductcontext,buttheresultsmappedmorecloselyontoaperformance-baseddesignbiasthananattractivenessbias.However,researchinfacialattractivenessdoesreportabiasingeffectofaesthetics.Mostconsumerresearchonproductdesignhasfocusedonaestheticsandhasdemonstratedavarietyofeffectsonlikingratingsorproductevaluationsintheabsenceofreliableinformation.Inthecurrentseriesofstudies,thekeymeasurehasbeenaqualityratingofparticularfeatures,andprioraestheticsresearchinconsumerbehaviorismuteonwhetherattractivenessmightbiassuchevaluation.Thus,toaddressthispossibilityandtoconfirmtheresultsfromthepilotstudy,anadditionalfactorwasincludedinexperiment4.Halftheexperimentalparticipantssawthesamepicturesasthoseusedinexperiment3;thatis,picturesinwhichtheproductslookedequallyattractivebutonelookedsuperioronthekeydimension.Fortherestoftheexperimentalparticipants,pictureswereusedwheretheoppositewastrue;thatis,theproductslookedequalonperformancebutonewassuperioronattractiveness.Suchadesignhasinherentdifficultiesbecausethepicturesarenotuniformacrossconditions.However,theuseoffiveproductreplicateshelpedtoreducethemagnitudeofthisdesignissue.MethodStimuliTheConsumerReportsreviewsusedinthisstudywerethesameasthoseusedinexperiment2.Similarly,theperformancepictureswerealsothesame.Theonlychangewastheadditionofpairsofpicturesforeachofthefivecategoriesinwhichthepairshad

PAGE 65

55equalratingsonperformancealongthekeydimension,butsignificantlydifferentratingsinattractiveness.Thepretestresultsfortheseadditionalpicturesarepresentedintable8-1.Foreaseofcomparison,thetablealsoincludesthepretestresultsfromtheperformancepicturesoriginallypresentedinexperiment2.Thepicturesofthenewstimuliappearinappendixe.DesignandProcedureThestudywasa2(orderofpresentation:picturefirstversusreviewfirst)x2(typeofpictures:performancedifferenceversusaestheticdifference)betweensubjectsdesign.Inadditiontherewasareview-onlyconditionasacontrolgroup.Theexperimentwasrunoncomputer,andallinstructionsweregivenonscreen.Asinexperiments2and3participantsweretoldtheywouldbereadingConsumerReportsreviewsforpairsofbrandsinfiveproductcategoriesandthattheywouldbeaskedtoassessthequalityoftheproductsbasedonthereviews.Aftertheintroductoryscreenthefirstofthefivecategorieswasdisplayedinarandomlydeterminedorder.Thepresentationofeachproductcategoryproceededasfollows.Participantsinthepicture-firstconditionssawpicturesofthetwobrands.Participantslookedatthepicturesfor20secondsbeforeproceeding.Thepicturesthendisappearedfromthescreenandwerereplacedbythewrittenreviews.Participantscouldnotproceedfurtheruntil40secondshadpassed.Participantsinthereview-firstconditionssawthetworeviewsfirst.After40secondsparticipantscouldproceedfromthewrittenreviewstothepictures,whichremainedonthescreenfor20seconds.Participantsinthebaselinecontrolconditionsawonlythereviews.Afterthepresentationofthepicturesanddescriptions,participantsproceededtoanewscreenandwereaskedtoindicatethedegreetowhichoneortheotherbrandwassuperioralongthekeydimension.Thisprocedurewasthenrepeatedfor

PAGE 66

56eachofthefiveproductcategories.Therestoftheprocedure,includingthefillertask,manipulationchecks,productknowledgeanddemographicquestions,mirroredexperiments2and3.Onehundredeighty-sevenundergraduatestudentsparticipatedintheexperimentforclasscredit.ResultsAcomputererrorcaused19participantstobeunabletocompletethestudy.Anadditional23participantsfailedanattentioncheckandwereremoved,leaving145intheanalysis.Table8-2presentstheoverallresultsforthefiveproductreplicates.A2(order)x2(picturetype)x5(replicate)mixedANCOVAusingknowledgeoftheproductsascovariateswasperformedonthefourexperimentalconditions(thebaselineconditionwasnotincludedinthisanalysis).Theanalysisrevealednothreewayinteraction(F<1),amarginalinteractionofpicturetypeandreplicate,F(4,448)=2.39,p<.06,nointeractionoforderandreplicate,F(4,448)=1.12,p>.34,noeffectofpicturetype,F(1,112)=1.34,p>.25,andnoeffectoforder,F<1.Importantly,however,therewasasignificantinteractionoforderandpicturetype,F(1,112)=4.47,p<.05.Asthetableindicatesthepictureshadlittleeffectwhentheydifferedonaesthetics,regardlessofwhetherthepicturecamebeforeorafterthewritteninformation.Whenthepicturesdifferedonperformance,theyhadaneffect,butonlywhentheyfollowedthereviews.Pairwisecomparisonsontheoverallmeansconfirmedthattherewasasignificantdifferencebetweenthepicture-firstandreview-firstconditionsfortheperformancepictures,(p<.05).Thedifferencebetweentheaestheticandperformanceconditionswhenthereviewprecededthepictureswasmarginallysignificant,(p<.07),aswasthe

PAGE 67

57differencebetweenthereview-firstperformancegroupandthebaselinecontrolcondition,(p<.08).Allothercomparisonswerenon-significant.Themanipulationcheckresultsappearintable8-3.Fortheperformancepictures,thegoalwastohavemean-centeredratingsthatweresignificantlylowerthan0onperformance(indicatingsuperiorityforthesuperiorpicture),andequaltoorhigherthan0onattractiveness.Fortheaestheticpicturesthegoalwastheopposite.Asindicatedintable8-3,allproductsmetthisgoalforthisgroupofsubjectsexceptin-lineskates.Theremovalofthisreplicatedidnotmateriallychangetheresults.AnANCOVAontheremainingreplicatesagainrevealedthecriticalsignificantinteractionoforderandpicturetype,F(1,113)=6.45,p<.05.Forcompletiontherewasnothreewayinteraction(F<1),asignificantinteractionofpicturetypeandreplicate,F(3,339)=3.53,p<.05,nointeractionoforderandreplicate,F(3,339)=1.51,p>.20,noeffectofpicturetype,F(1,113)=1.23,p>.20,andnoeffectoforder,F<1.Therevisedoverallmeanswiththein-lineskatereplicateremovedwereslightlyhigherthantheoriginalmeans,butmaintainedthesamedirectionalpatterns.Forthepicture-firstconditions,meanratingswere7.56forperformancepicturesand7.21foraestheticpictures.Forthereview-firstconditions,meanratingswere6.77forperformancepicturesand7.53foraestheticpictures.Thereview-onlycontrolconditionhadameanratingof7.23.Follow-upcontrastsagainrevealedasignificantdifferencebetweenthetwoperformancepictureconditions(p<.05)andamarginaldifferencebetweenthereview-firstconditions,(p<.07).Theonlynotablechangewasthatthedifferencebetweenthereview-firstperformanceconditionandthebaselinereview-onlygroupbecamenon-significantaftertheremovalofthein-lineskatereplicate.

PAGE 68

58DiscussionWhatisimmediatelynotablefromexperiment4isthattherewaslittleeffectofaesthetics.Thepilotstudysuggestedanaestheticbiasforbackpacks.Althoughtherunningshoesdemonstratedaperformance-baseddesignbias,itwaspresumedthatbotheffectscouldhaveexistedbuttheaestheticbiasmayhavebeendominatedbytheperformancebias.However,theresultsofexperiment4suggestthataestheticsmaynotinfluencequalityevaluations.Ofcourse,onecouldarguethatthiscouldhavebeenacalibrationissue;indeed,themeanratingsforspeakersandtoalesserextenthelmets,didsuggestanattractivenesseffectwhenthepicturesprecededthereviews.Acomparisonofthepretestandmanipulationcheckvaluesforattractivenessversusperformancerendersthispossibilitysomewhatlesslikely.Asshownintable8-1,thepretestratingsforthepicturestendedtohavesimilarsizeddifferencesfortheperformanceandaestheticpictures.Themanipulationcheckratingsprovideasimilarpicture.Asindicatedintable8-3,forthisgroupofparticipants,thedifferenceratingsfortheattractivenesspairsweregenerallymuchlargerthanfortheperformancepairs.Ofcoursethescalesmayhavebeenquitedifferent,butitatleastsuggeststhattheresultwasnotsolelyduetocalibration.Priorproductdesignresearchwouldsuggestthataestheticsmighthaveaninfluenceoninformationprocessing,althoughmostsuchresearchhasfocusedonaestheticsbiasingpreferenceorevaluationsintheabsenceofotherinformation(PageandHerr2002).Theexpectationofanaestheticinfluenceisthatpeoplewouldhaveanautomaticaffectivereactionthatwouldcolorfurtherprocessing.Itispossiblethatatleastinsomecasesaestheticsdidhaveaninfluencewhenthepicturecamefirst,butoverallatleast,nostrongaestheticsbiasemergedoverall.Thepossibilityofanaestheticbiascannotberuledouthere,andfurtherstudieswouldbeneededtoaddressthisissue.However,basedonthe

PAGE 69

59resultsofthecurrentstudy,thefocusofthedissertationwillremainontheperformance-baseddesignbias.Anadditionalsurprisingresultwasthatfortheperformancepictures,theeffectofdesignwasobservedwhenthepicturefollowedthereviewbutnotwhenitprecededthereview.Thusthenotionthattheinfluenceofdesignisbecausepeoplehaveanimmediatereactionwhichbiasessubsequentevaluationappearsincorrect.Theresultsfortheperformanceconditionsinexperiment4seemconsistentwitharecencyeffect.Intheexperimentparticipantsknewtheywouldbemakingsomekindofqualityrating,buttheydidnotknowaheadoftimewhatdimensiontheywouldbeevaluating.Thus,uponseeingthequestion,whichappearedaftertheproductinformationhadbeenremovedfromthescreen,theyhadtorelyontheirmemoriestorespond.Themostrecentlyviewedinformationwasprobablytheeasiesttorecall,andsoitwasused,leadingtothedivergentresults.However,recencycannotentirelyaccountforallexperimentalconditions.Theaccuracyofthereview-first/aestheticsconditiondemonstratesthatpeoplewereabletorecalltheearlierwritteninformation.Fortheseparticipants,thepictureswerethemostrecentlyviewedpiecesofinformationandwerecertainlyaccessible,buttheywerenotdiagnostic,soitwasnecessarytodelvefurtherintomemorytoretrievetask-relevantinformation.Forparticipantsintheperformance/review-firstcondition,thepictureswerealsohighlyaccessibleinmemory.Unliketheaestheticpicturegroups,however,fortheseparticipantsthepictureswerediagnostic,sopeoplecouldusethemintheirjudgment.Thus,ratherthanbeingmerelyarecencyeffect,itappearsthepictureresultsaremoreconsistentwithacombinationofaccessibilityanddiagnosticity(seeFeldmanandLynch1988).Ifthisisthecase,thenalertingparticipants

PAGE 70

60tothejudgmenttaskpriortopresentingtheproductinformationmightaltertheinfluenceofthepictureandwritteninformation.Thispossibilityisexploredinexperiment5.Table8-1.PretestPictureRatingsPictureTypePerformanceDifferenceAestheticDifference CategoryAttractivenessRatingPerformanceRatingAttractivenessRatingPerformanceRating Cookware-0.371.08*-1.79***0.37 Stereospeakers0.541.08**-1.42**0.46In-lineskates-0.331.13**-1.04*0.33Bicyclehelmets0.501.08***-1.04*0.50Electricmixers0.132.00***-1.04*0.13*p<.05 **p<.01***p<.001Table8-2.ResultsofExperiment4PerformancePicturesAestheticPicturesControl PictureFirstReviewFirstPictureFirstReviewFirstReviewOnly Cookware7.436.457.527.176.71 Stereospeakers7.867.616.877.877.33In-lineskates6.096.066.526.467.08Bicyclehelmets6.896.266.166.506.54Electricmixers8.066.778.298.588.33Average7.266.637.077.327.20 Table8-3.ManipulationCheckforExperiment4PictureTypePerformanceDifferenceAestheticDifference CategoryAttractivenessRatingPerformanceRatingAttractivenessRatingPerformanceRating Cookware1.01***-1.19***-2.03***1.17*** Stereospeakers1.68***-0.56*-1.21***1.18***In-lineskates0.74**-1.18***0.210.91***Bicyclehelmets-0.19-0.94***-2.39***0.02Electricmixers0.68**-2.12***-2.50***-0.14*p<.05 **p<.01***p<.001

PAGE 71

61CHAPTER9EXPERIMENT5Experiment5focusedsolelyontheperformance-designbias.Theexperimentattemptedtoaddresstheordereffectthatwasobservedintheperformance-pictureconditionsinexperiment4.Inexperiment4participantsinthepicture-firstconditiondidnotshowadesignbias,butthoseinthereview-firstconditiondid.However,participantsinneithergroupknewaheadoftimewhatdimensiontheywouldbeevaluating.Alltheyknewwasthattheywouldbeevaluatingsomeelementofquality.Thusitisplausiblethatparticipantsinthepicture-firstgroupdidnotfocusonthekeydimension,butratheronsomeoverallqualityjudgment.Thereviewsthenprovidedinformationonthekeydimension,whichwassubsequentlyusedtomakethejudgment.Forparticipantsinthereview-firstcondition,thereviewfocusedparticipantsonthekeydimension,whichmayhavethenalertedthemtotheconflictinginformationprovidedbythepicture.Thus,whenitcametomakingajudgment,theywereinfluencedbythepicture.Inexperiment5participantsweretoldwhatdimensiontheywouldbeevaluatingpriortoviewinganyinformationabouttheproduct.Thefinalquestionstilloccurredafterbothpiecesofinformationwereremovedfromthescreen,butbyprovidingparticipantswiththedimensionaheadoftime,itwasmorelikelythatboththepictureandwritteninformationwereequallyaccessibleatthetimeofjudgment.

PAGE 72

62MethodStimuliThesamesetofpicturesusedintheperformanceconditionsofexperiment4wereusedinexperiment5.Thesameproductinformationwasalsoretained.DesignandProcedureExperiment5consistedofonebetween-subjectsfactoroforder,withthreelevels:picturefirst,reviewfirst,andareview-onlycontrolgroup.Againtherewerefiveproductreplicates.Theprocedurewassimilartothatofexperiment4.Theonlydifferencewasthatparticipantsweretoldwhatdimensiontheywouldbeevaluatingpriortoseeingthepicturesorthewrittenreviewsoftheproducts.Ninety-oneundergraduatestudentsparticipatedintheexperimentforpartialclasscredit.ResultsElevenparticipantswereremovedforfailinganattentioncheck,leaving80intheanalysis.Themeanratingsappearintable9-1.Therewasasignificanteffectofreplicate,F(4,308)=12.50,p<.01andnointeractionoforderandreplicate,F(8,308)=1.45,p>.15.Therewasasignificanteffectoforder,F(2,77)=4.83,p<.05,butthedirectionoftheordereffectwasreversedfromthatobservedinexperiment4.Inthecurrentstudy,whenpeoplewereinstructedaheadoftimewhichdimensiontofocuson,thepicturehadastrongereffectwhenitprecededthereview.Followupcomparisonsrevealedasignificantdifferencebetweenthepicture-firstandreview-firstconditions,t(77)=2.55,p<.05andbetweenthepicture-firstandreview-onlyconditions,t(77)=2.78,p<.05,butnodifferencebetweenthereview-firstandreview-onlyconditions,t<1.

PAGE 73

63Themanipulationchecksrevealedthatforthisgroupofparticipants,boththestereospeakersandthebicyclehelmetswereunsuccessfulmanipulationsofperformancedesign.Removingthesetworeplicatesfromtheanalysisresultedinrevisedoverallmeansof7.07,7.55,and7.98forthepicture-first,review-first,andreview-onlyconditionsrespectively.Theeffectoforderwasstillsignificant,F(2,77)=3.48,p<.05,butthepairwisecomparisonsresultedinonlyasignificantdifferencebetweenthepicture-firstandreview-onlyconditions,p<.05.DiscussionByprovidingparticipantswiththefocaldimensionbeforetheinformation,experiment5enabledatestoftheeffectoforderthatavoidedtheproblemofrecency.Inthisstudytheappearanceofthequestionafterthepresentationofthepicturewasnosurprise,soparticipantsdidnothavetosearchtheirmemoriesforrelevantjudgmentinformation.Thekeyresultfromexperiment5isthatthepicturehadthegreatestinfluencewhenitprecededthewrittenreview,areversaloftheordereffectobservedinexperiment4.Inthisstudy,participantsknewwhatfeaturetheywouldbeevaluating.Whenthepictureappeared,itseemeddiagnosticofqualityandthereforewasusedinarrivingatajudgmentofthefeature.Thesubsequentwrittenreviewwasofcoursealsodiagnostic.However,whenitcametimetosupplytheevaluation,thejudgmentmadebasedonthepicturewasprobablystillhighlyaccessible,andsinceitwasalsoperceivedtobediagnosticitinfluencedtherating.Whenthereviewprecededthepicture,participantswereeasilyabletoassessthequalityofthetwoproductsalongthedimensionofinterestbeforeseeingthepicture.Thus,eventhoughthepictureprovidedconflictinginformation,itsinfluencewasweaker,

PAGE 74

64perhapsbecausethejudgmentmadebasedonthewrittenreviewswasveryclear.Thepicturedidstillappeartohavesomeinfluence,however.Anexaminationoftheoverallmeansaftertheremovalofthestereospeakersandbicyclehelmetsdoessuggestaninfluenceofthepicturewhenitfollowedthereview,albeitaweakeronethanwasobservedinthepreviousstudy.Experiments3and4suggestedthatthepictureeffectwasduetodiagnosticity.However,inthecurrentstudy,adiagnosticityresultshouldhaveledtoequalratingsforthepicture-firstandreview-firstconditions.Thereview-firstconditionshowedsomeeffectofthepictures,particularlyforthein-lineskatesandelectricmixers,butoveralltheresultswerenotsupportive.Ifthecurrentresultsarereliable,thedesigneffectmaynotbesolelyduetodiagnosticity,butrathermaybetheresultofanimmediatereactiontothepicturebiasingsubsequentevaluation,butonlywhenattentionisdrawntodimensionswherethepicturedominates.Additionaldataareneededtoconfirmtheresultsofthereview-firstcondition.Table9-1.ResultsofExperiment5PictureFirstReviewFirstReviewOnly Cookware6.397.657.50 Stereospeakers6.547.967.19In-lineskates7.147.197.77Bicyclehelmets7.047.657.42Electricmixers7.687.818.69Average6.957.657.71

PAGE 75

65CHAPTER10GENERALDISCUSSIONThegoalofthispaperwastoexaminesomeofthewaysproductdesignmightinfluencejudgment.Apilotstudyandfiveexperimentspittedproductappearanceagainstconflictingwritteninformationtoaddresswhetherdesigncouldinfluenceapersonstendencyorabilitytoobjectivelyevaluateobjectiveproductinformation.Consistentwithpriorresearchinpsychologyaswellasconsumerresearchonproductdesign,theinitialpilotstudyfocusedonattractivenessasasourceofbias.Anattractivenesseffectobtainedforoneofthetwoproductreplicates,butattractivenesswasconfoundedwithperformanceforthatstimulus.Fortheotherstimulus,noattractivenesseffectwasobserved.Rathertheresultssuggestedabiasinthedirectionoftheunattractivebrand,whichsubsequentanalysisrevealedlookedlikeasuperiorperformer.Thisdiscrepancywasresolvedinexperiments1and2,whichconfirmedthattheappearanceofperformancebiasedtheprocessingofotherinformation.Experiments3through5attemptedtoassessthesourceoftheperformance-designbias.Experiment3investigatedwhetherdesignoperatedasatypeofhalothatinfluencedratingsofotherattributes.Resultsindicatedthatdesignonlyinfluencedratingsforfeaturesonwhichitappearedtodominate,suggestingtheeffectwasnotduetoahaloeffect.Experiments4and5examinedwhetherthebiaswasafunctionofaninitialreactiontothevisualinformationcoloringsubsequentprocessingofwritteninformation.Experiment4didnotsupportsuchanexplanation,andinsteadappearedmoreconsistentwithaaccessibility-diagnosticityexplanation.Inthatexperiment,participantswerenot

PAGE 76

66focusedonthekeydimensionpriortoviewingtheinformation.Whenaskedthedependentmeasure,participantsusedthemostaccessibleinformationthatwasdiagnosticforthetask.Thistendedtobethemostrecentlyviewedinformation.Experiment5waslessconclusive,butstillhintedatadiagnosticityresult.Whenparticipantswerefocusedonthekeydimensionaheadoftime,thepicturehadaninfluencewhenitappearedpriortothewritteninformationbecausethetaskwasknownatthetimeandthepictureswerediagnostic.Whenthepicturefollowedthereview,however,lessinfluenceofthepicturewasobserved.Thisresultisinconsistentwithadiagnosticityexplanationandmoreinlinewithanearlyreactiondistortingsubsequentevaluation.However,stimulicalibrationconcernsmayhavebeenpartiallyresponsible.Overalltheresultsoftheexperimentssuggestthatforproducts,whenitcomestobeingevaluatedforquality,seeingisbelieving.Whenaproductappearstooffersuperiorperformancealongagivendimension,peoplebelievethatitdoes,andreadinginformationthatsuggestsotherwisemaybeinsufficienttocounteractthatbelief.Whetherthereisanaestheticsbiasisunclear.Experiment4addressedtheissueofattractivenessasasourceofbiasinitiallyinvestigatedinthepilotstudy,andfoundnoeffectofaesthetics.However,thestudydidnotaskforchoiceorpreferencejudgments.Suchjudgments,whichareinherentlymorepersonal,maybemoresusceptibletoaestheticeffects.Certainlyifoneconsiderstheresearchinfacialattractivenessandthedramaticeffectsthathavebeenreported(e.g.,LandyandSigall1974),itislikelythatanaestheticbiasinproductswouldoccurundersomeconditions.Theresultsofsomeofthereplicatesinexperiment4hintthattheremaybeaestheticsbias;furtherresearchis

PAGE 77

67neededtoconfirmthisandtounderstandtheconditionsunderwhichanaestheticdesignbiasmightobtain.ImplicationsProductdesignandstylingareoftenamongthemostimportantdeterminantsofnewproductsuccess(CooperandKleinschmidt1987;Vriensetal.1998).Agooddesigncanattractattention,communicateinformationaboutaproduct,andinfluencepreference.Thus,understandinghowconsumersprocessandreacttoproductdesign,particularlywithregardtonewproductdesignsusedinmarkettesting,isofgreatimportance.Formanagers,thekeytakeawayfromthisresearchmaybethatconsumersbelievewhattheysee.Ifthegoalofanewproductistocommunicatesomelevelofperformance,itiscriticalthatthisiscommunicatedvisually.Otherwisemarketingeffortsaimedatconvincingconsumersthattheproductexcelsonagivendimensionmayfallflat.Intermsofdesignteamgoals,theresearchsuggeststhatacompanymaybebetteroffexpendingeffortsonmakingaproductlooklikeitworkswellratherthansimplymakingitlookgood.Fromaconsumerperspective,ofparticularimportanceisabetterunderstandingoftherolethatproductappearanceplaysindecisionmaking.Theexperimentsdemonstratethatconsumersareabletodiscountattractivenessintheevaluationofproductquality,andyetseemunabletoignoreappearancewhendesignsuggestsperformance.Itmaybethatthisisadeliberatechoiceandthatconsumersconsciouslydiscountthewritteninformationwhentheireyestellthemotherwise.However,ifconsumersareconvincedbywhattheyseeratherthanotherinformationtheyencounterandareunabletoignorevisualinformationwhenitshouldbeignored,thereispotentialfornon-normativedecisionmaking.Asanextremeexample,productsthatappeartodominateonsafety,

PAGE 78

68suchaslifejackets,flashlights,orcars,butactuallydonotprovidethelevelofperformancethattheirappearancesuggests,couldresultinseriouslydetrimentalconsequencesfortheconsumer.LimitationsTheresearchreportedherefocusedonafairlysmallnumberofproductsthatwerechosenbecausetheywereproductswheredesigncouldbeimportantbutwasnotcriticaltothefunctionoftheproduct.Becauseofthesenarrowselectioncriteria,itisunclearwhethertheresultsobservedinthedissertationwouldextendtoawiderarrayofproductcategories.Moreover,inseveralexperimentsnotallthepicturemanipulationsweresuccessful.Itwillbeimportanttoimprovethestimulicalibrationandreplicatethestudiestoconfirmthevalidityoftheresults.Theresearchwasconductedusinganundergraduatestudentsubjectpool.ThisisarelativelyhomogeneousgroupandnotrepresentativeofthetypicalAmericanconsumer.Althoughthisisaconcerncommontonearlyallconsumerbehaviorresearch,itmakesitisdifficulttoassessthegeneralizabilityoftheresultstothepopulation.Thestudiespresentedinthisdissertationsuggestedthatthereislittleinfluenceofattractivenessintheprocessingofproductinformation.Itiscertainlypossible,however,thatundersomeconditions,theaestheticsofaproductmayindeedinterferewithaconsumersabilityorinclinationtoevaluateproductquality.Itispossiblethattheproductsusedinthecurrentstudywerenotsufficientlyattractiveorinterestingtothestudentparticipants,andhadmoreattractiveorstudent-relevantproductsbeenused,anattractivenesseffectmighthaveemerged.

PAGE 79

69FutureResearchExperiments3and4suggestedthattheeffectofproductdesignobservedinthisresearchwereduetoadiagnosticityeffect.Theresultsofexperiment5,however,lefttheissuesomewhatunresolved.Experiment5hadsomeevidenceconsistentwithadiagnosticityeffect,butalsosuggestedthatthedesigneffectcouldhavebeentheresultofaninitialreactioncoloringsubsequentevaluation.Thestudyhadtwofailedstimulimanipulations,whichwouldneedtobecorrectedinafutureexperiment.Inaddition,astudywhichmanipulatesorderofinformationpresentation,asinexperiment5,butalsohasparticipantsratefeaturesonwhichthepicturesarenotdiagnostic,asinexperiment3,couldhelptoresolvetheissue.Theresearchreportedheredemonstratedthatconsumersabilitytoprocesswritteninformationisinfluencedbyconflictingvisualinformation.Theevidenceshowedthatwhenitcomestoqualityjudgments,consumersarenotbiasedbyattractiveness,butarebiasedbytheappearanceofperformance.Thissuggeststhattheeffectmayberelativelydeliberativeandconscious,butitmaynotbe.Additionalresearchisnecessarytodeterminethedegreetowhichpeopleareawareoftheeffectofdesignandtheextenttowhichpeoplecancorrectforit.Consistentwithresearchinmentalcontamination(WilsonandBrekke1994)itcouldbethatpeopleareawareofthebiasingeffectofdesignandyetarestillunabletocounteractit.Asstatedabove,thedissertationfoundlittleaestheticinfluence.Priorresearchhasfoundthatattractivenessdoesplayaroleinlikingevaluationsandmayinfluencequalityevaluationsintheabsenceofotherconcretefunctionalinformation(PageandHerr2002).Thecurrentresearchisconsistentwiththesefindings,butdoesnotconcludethataestheticscannotinterferewithinformationprocessing.Researchinfaceprocessinghas

PAGE 80

70demonstratedthatundercertainconditions,anattractivefacecanalterthemannerinwhichfunctionalinformationisassessed(LandyandSigall1974).Theremaybesomecasesinwhichasufficientlyattractiveproductcanhavethesameeffect.Forexample,Norman(2004)suggeststhatproductdesigncanactasamoodmanipulation,withgooddesignelevatingmoodandmakingpeoplelesslikelytonoticesmallfunctionalproblems.Ifaproductisattractiveenoughtocauseanelevationinmood,thenitmayleadtomotivatedreasoningwithregardtofunctionalattributeinformation.Futureresearchisneededtoinvestigatetheconditionsunderwhichproductaestheticswillandwillnotinterferewithprocessingofproductinformation.

PAGE 81

71APPENDIXAPRODUCTPICTURESUSEDINPILOTSTUDYUnattractiveShoeAttractiveShoe UnattractiveBackpackAttractiveBackpack

PAGE 82

72APPENDIXBPRODUCTPICTURESUSEDINEXPERIMENT1HighperformanceshoesLowperformanceshoes

PAGE 83

73APPENDIXCPRODUCTPICTURESUSEDINEXPERIMENTS2THROUGH5HighPerformanceBrandLowPerformanceBrand

PAGE 84

74

PAGE 85

75APPENDIXDPRODUCTDESCRIPTIONSUSEDINEXPERIMENT2,4AND5ProductCategoryLowPerformanceBrandHighPerformanceBrand CookwareThiscookwaresetincludesalargestockpot,threepots,andtwofryingpans.Thepotshavewellfittinglidsthatkeepheatinnicelyandsturdyhandlesthatstaycooltothetouch.Thesetisdurableandcanbeusedatthehighestheat,evenwhencookingwithagasstove.Foodcooksveryevenlybecauseoftheflatbottomsthatsitsquarelyontheelements,maximizingheatdistributionandminimizingburning.Ifpanssitoutforanextendedperiod,remainingfoodcanbedifficulttoremove.Thelidsforthestockpotandmediumpotalsofitthetwofryingpans.Thiscookwarehasgoodheatdistribution;foodgenerallycooksevenly.Thepotsandpanshaveasolidbase,durableconstructionandlidsthatsealinheatverywell.Thestay-coolhandlesareergonomicallydesignedandcomfortabletohold.Verylittlewillsticktothesepans,regardlessofcookingmethod,sotheycanbewashedoutwithdishsoapandacloth;it'srarelynecessarytouseascrubbrush.Thesetcomesincludesthreehandledpotsofvaryingsizeswithlids,alargestockpotwithlid,andtwofryingpansthatfitthelidsofthestockpotandmid-sizedpot. StereospeakersThesespeakerspackseveralpatentedtechnologiesintotheirslendercabinetstoensurestrongresults.Theyhaveastraight-linesignalpathcrossovernetworkwhichdistributesthesignalwithminimalprocessing.Atrelativelylowvolumesthespeakersprovideclearandaccuratesound,butslightdistortionmayoccurathighervolumes.Theycanbeusedwithacomputer,astereooraspartofahometheatresystem.Thesemagneticallyshieldedspeakersareappropriateforanapartmentordorm,butwouldbeinadequateforalargerhome.Thespeakershousetwoangled,full-rangedriversengineeredtogeneratemoreairdistributionfromasmallenclosurethanconventionaldrivers.Theyprovideconsistentsoundatbothlowandhighvolumes.Youcanalmosttelltherelativepositionofthedifferentinstrumentsintherecordingstudio.In-lineskatesThisin-lineskateissolidanddurable.Itoffersaneasy-entrysystemmakingforveryquickchanges.ItusesaTrueFitinterior,wrappinganylonandplasticfilmaroundthefootforextrastabilityandcontrol.Thefitissnugandwellcushioned.Therearetinyslitsinthefrontwhichinterfereslightlywithaerodynamics,butwhichhelpcoolthefeet.Thewheelsareslightlysofterthanonothermodels,makingforamorecomfortableridebyincreasingshockabsorption.Arigid,well-alignedframeimprovesmaneuverability.Thisin-lineskatehasacloserfitintheanklecollarthanoldermodels,givingstabilitywhilestillallowingtheskatertoflexforward.Lightermaterialsthroughoutgivetheskateanimpressiveweightofonlythreepounds,whichincreaseseaseofmovement.Thewheelsarehard,tolastlongerandimprovespeed,butasaresult,donotofferthatmuchshockabsorption.Thisiscompensatedforsomewhatbytheinteriorcomfort.Theskateshaveathicklinerandasoftfoaminteriortoincreasecomfortandkeepfeetcool.

PAGE 86

76BicyclehelmetsThishelmethasaroll-cageinteriorbracingsomewhatreminiscentofrollbarsinjeeps,madetoprotecttherideronimpact.Despitethisadditionalbracing,thehelmetisreasonablylight.Theventilationisadequate.Althoughitonlycomesinonesize,thehelmetissuitableformostheadsizes.Thefoamontheinteriorliningisalittlestiff,butthepadscanbeadjustedtosuityourheadsize.Thebuckleisspeciallydesignedtostaysecurebutallowforaquickreleaseincaseofemergency.Thishelmetisdurablewithareinforcedribcageframe.Ofcourse,thereinforcedstructureaddssomeweighttothehelmet,makingitabitheavierthanmost.Buttheliningissoft,thickandwellpaddedwithfoam,sotheextraweightisnotmuchofaburden.Superiorfitisprovidedbyhavingthreesizes.Also,ithasasleekwindfunnelingtechnologytopullcoolairinandblowhotairout,coolingtheheadduringrides.Strapsaresturdyandwillstayinplaceonimpact.ElectricmixersThismixerisrelativelycompact,fittingeasilyonmostcounters,butthebowlisdeepenoughtohandlelargesizedjobs.Ithasagradualstarter,minimizingsplatter.Itchangesspeedsverysmoothly,anditsplasticgearsmakethemachinesurprisinglyquiet.Ithasanautoshut-offmotorprotectionfeatureifthemotorgetsoverloaded,whichdoeshappenoccasionally.Beaterssnapinandoutveryeasilyandaredishwashersafe.Thisisanefficientmixerwithawide-setbowlthatcanbesnappedinorliftedoutofthecradlequicklyandeasily.Themachinehasallsteelgearsandatwin-motordesignthatrotatesthebowlwhilemixing.Itcanhandleeverythingfromsmall,highspeedtaskstolargeheavydoughs.Occasionallythebatterwillclingtothebeaters,whichrequiresstoppingthemachinetoscrapethecontentsbackintothebowltocompletethemixing.

PAGE 87

77APPENDIXEPRODUCTDESCRIPTIONSUSEDINEXPERIMENT3ProductCategoryLowPerformanceBrandHighPerformanceBrand CookwareThiscookwaresetincludesalargestockpot,threepots,andtwofryingpans.Thepotshavewellfittinglidsthatcompletelysealinheatandsturdyhandlesthatstaycooltothetouch.Thesetisdurableandcanbeusedatthehighestheat,evenwhencookingwithagasstove.Foodcooksevenlybecauseoftheflatbottomsthatsitsquarelyontheelements,maximizingheatdistributionandminimizingburning.Ifpanssitoutforanextendedperiod,remainingfoodcanbedifficulttoremove.Thelidsforthestockpotandmediumpotalsofitthetwofryingpans.Thiscookwarehasgoodheatdistribution;foodcooksveryevenly.Thepotsandpanshaveasolidbase,durableconstructionandlidsthatkeepheatinnicely.Thestay-coolhandlesareergonomicallydesignedandcomfortabletohold.Verylittlewillsticktothesepans,regardlessofcookingmethod,sotheycanbewashedoutwithdishsoapandacloth;it'srarelynecessarytouseascrubbrush.Thesetcomesincludesthreehandledpotsofvaryingsizeswithlids,alargestockpotwithlid,andtwofryingpansthatfitthelidsofthestockpotandmid-sizedpot. StereospeakersThesespeakershaveastraight-linesignalpathcrossovernetworkwhichdividesthesignalwithminimalprocessing.Bassresponseisadequatebutnotstrong,makingthesespeakersperfectforasmallerspacelikeadorm.Thesoundisclearandaccurate,veryeasytolistento.Youcanalmosttelltherelativepositionofthedifferentinstrumentsintherecordingstudio.Thebalancingtechnologyblendsdirectandindirectsoundalmostanywhereintheroom,givingthefeelingofsurroundsoundwithonlytwospeakers.Thesemagneticallyshieldedspeakershousetwoangled,full-rangedriversengineeredtogeneratemoreairmovementfromasmallenclosurethanconventionaldrivers.Theydeliverclean,uncolored,realisticsound,whichismaintainedathighervolumes,providedofcoursethattheaccompanyingstereosystemisadequate.Theywouldbesuitableforahometheatersystembecausethedirect/reflectsystemavoidsthecommonproblemofaspeakersweetspotandensuresthesamequalityofsoundthroughouttheroom.In-lineskatesThisin-lineskateissolidanddurable.Itoffersaonestrapeasy-entrysystemmakingforveryquickchanges.ItusesaTrueFitinterior,wrappinganylonandplasticfilmaroundthefootforextrastabilityandcontrol.Thefitissnugandverywellcushioned.Therearetinyslitsinthefrontwhichinterfereslightlywithaerodynamics,butwhichcoolthefeet.Thewheelsareslightlysofterthanonothermodels,makingforamorecomfortableridebyincreasingshockabsorption.Thisin-lineskatehasaclosefitintheanklecollar,givingstabilitywhilestillallowingtheskatertoflexforward.Lightermaterialsthroughoutgivetheskateaweightofonlythreepounds,increasingeaseofmovement.Thewheelsarehard,tolastlongerandimprovespeed,butasaresult,donotofferthatmuchshockabsorption.Thisiscompensatedforverywellbyathicklinerandasoftfoaminterior.Nolacesandeasy-tugentrygripsmaketheskateverysimpletoputonandtakeoff.

PAGE 88

78BicyclehelmetsThishelmethasaroll-cageinteriorbracingmadetoprotecttherideronimpact.Despitethisadditionalbracing,thehelmetisnottooheavy.Astreamlineddesignandwellpositionedventsoptimizespeed.Althoughitonlycomesinonesize,thehelmetissuitableformostheadsizes.Thefoamontheinteriorliningisalittlestiff,butthepadscanbeadjustedtosuityourheadsize.Thebuckleisspeciallydesignedtostaysecurebutallowforaquickreleaseifnecessary.Thishelmetisstrengthenedwithareinforcedribcageframeforsafety.Ofcourse,thereinforcedstructureaddsabitofweighttothehelmet.Buttheliningissoft,thickandwellpaddedwithfoam,sotheextraweightisnotmuchofaburden.Superiorfitisprovidedbyhavingthreesizes.Strategicallypositionedairventsandadvancedwindfunnelingtechnologyhelptocooltheheadwithoutsacrificingspeed.Strapsaresturdyandwillstayinplaceonimpact.ElectricmixersThismixerisrelativelycompact,fittingeasilyonmostcounters,butthebowlisdeepenoughtohandlelargesizedjobs.Ithasagradualstarter,minimizingsplatter.Itchangesspeedsverysmoothly,anditsplasticgearsmakethemachinesurprisinglyquiet.Ithasanautoshut-offmotorprotectionfeatureifthemotorgetsoverloaded,whichdoeshappenoccasionally.Beaterssnapinandoutveryeasilyandaredishwashersafe.Thisisanefficientmixerwithawide-setbowlandbeatersthatcanbesnappedinorliftedoutofthecradlequicklyandeasilywithonehand.Themachinehasallsteelgearsandatwin-motordesignthatrotatesthebowlwhilemixing.Thegearsandmotorsareinsulatedtominimizenoise.Itcanhandleeverythingfromsmall,highspeedtaskstolargeheavydoughs.

PAGE 89

79APPENDIXFPRODUCTPICTURESUSEDINEXPERIMENT4AttractiveBrandUnattractiveBrand

PAGE 90

80

PAGE 91

81LISTOFREFERENCESAlba,JosephW.,andJ.WesleyHutchinson(1987),DimensionsofConsumerExpertise,JournalofConsumerResearch,13(March),411-454.Ambady,Nalini,andRobertRosenthal(1993),HalfaMinute:PredictingTeacherEvaluationsfromThinSlicesofNonverbalBehaviorandPhysicalAttractiveness,JournalofPersonalityandSocialPsychology,64(March),431-41.Beckwith,NeilE.,andDonaldR.Lehman(1975),TheImportanceofHaloEffectsinMulit-AttributeAttitudeModelsl,JournalofMarketingResearch,12(August),265-75.Berkowitz,Leonard(1993),TowardsaGeneralTheoryofAngerandEmotionalAggression:ImplicationsoftheCognitive-NeoassociationisticPerspectivefortheAnalysisofAngerandOtherEmotions,inAdvancesinSocialCognition,Vol.6,eds.RobertS.WyerandThomasK.Srull,Hilsdale,NJ:Erlbaum,1-46.Berlyne,D.E.(1974).Novelty,ComplexityandInterestingness,StudiesintheNewExperimentalAesthetics,ed.D.E.Berlyne,NewYork:JohnWiley&Sons,Inc.,175-80.Bloch,PeterH.(1995),SeekingtheIdealForm:ProductDesignandConsumerResponse,JournalofMarketing,59(July),16-29.Bloch,PeterH.,FredericF.Brunel,andToddJ.Arnold(2003),IndividualDifferencesintheCentralityofVisualProductAesthetics:ConceptandMeasurement,JournalofConsumerResearch,29(March),551-65.Boiney,LindsleyG.,JaneKennedy,andPeteNye(1997),InstrumentalBiasinMotivatedReasoning:MoreWhenMoreIsNeeded,OrganizationalBehaviorandHumanDecisionProcesses,72(October),1-24.Breckler,StevenJ.,andElizabethC.Wiggins(1989),AffectversusEvaluationintheStructureofAttitudes,JournalofExperimentalSocialPsychology,25(May),253-71.Cohen,JoelB.,andCharlesS.Areni(1991),AffectandConsumerBehavior,HandbookofConsumerBehavior,eds.ThomasS.RobertsonandHaroldH.Kassarjian,EnglewoodCliffs,NJ:Prentice-Hall,188-240.

PAGE 92

82Cooper,RobertG.,andElkoJ.Kleinschmidt(1987),NewProducts:WhatSeparatesWinnersfromLosers?JournalofProductInnovationManagement,4(September),169-84.Cox,Dena,andAnthonyD.Cox(2002),BeyondFirstImpressions:TheEffectsofRepeatedExposureonConsumerLikingofVisuallyComplexandSimpleProductDesigns,JournaloftheAcademyofMarketingScience,30(spring),119-30.Creusen,MarielleE.H.,andJanP.L.Schoormans(2005),TheDifferentRolesofProductAppearanceinConsumerChoice,JournalofProductInnovationManagement,22(January),63-81.Darley,JohnM.,andPagetH.Gross(1983),AHypothesis-ConfirmingBiasinLabelingEffects,JournalofPersonalityandSocialPsychology,44(January),20-33.Deighton,John(1984),TheInteractionofAdvertisingandEvidence,JournalofConsumerResearch,11(December),763-70.Dickie,George(1971),Aesthetics:AnIntroduction,NewYork:TheBobs-MerrillCompany,Inc.Ditto,PeterH.,andDavidF.Lopez(1992),MotivatedSkepticism:UseofDifferentialDecisionCriteriaforPreferredandNonpreferredConclusions,JournalofPersonalityandSocialPsychology,63(October),568-84.Eagly,AliceH.,RichardD.Ashmore,MonaG.Makhijani,andLauraC.Longo(1991),WhatIsBeautifulIsGood,But:AMeta-AnalyticReviewofResearchonthePhysicalAttractivenessStereotype,PsychologicalBulletin,110(January),109-28.Edwards,Kari(1990),TheInterplayofAffectandCognitioninAttitudeFormationandChange,JournalofPersonalityandSocialPsychology,59(August),202-16.Feldman,JackM.,andJohnG.Lynch,Jr.(1988),Self-GeneratedValidityandOtherEffectsofMeasurementonBelief,Attitude,Intention,andBehavior,JournalofAppliedPsychology,73(August),421-35.Goldman,Morton,MarkD.Cowles,andCharlesA.Florez(1983),TheHaloEffectofanInitialImpressionuponSpeakerandAudience,JournalofSocialPsychology,120(August),197-201.Green,ChristopherD.(1995),AllthatGlitters:AReviewofthePsychologicalResearchontheAestheticsoftheGoldenSection,Perception,24(August),937-68.Hassin,Ran,andYaacovTrope(2000),FacingFaces:StudiesontheCognitiveAspectsofPhysiognomy,JournalofPersonalityandSocialPsychology,78(November),837-52.

PAGE 93

83Herr,PaulM.,FrankR.Kardes,andJohnKim(1991),EffectofWord-of-MouthandProduct-AttributeInformationonPersuasion:AnAccessibility-DiagnosticityPerspective,JournalofConsumerResearch,17(March),454-62.Hirschman,ElizabethC.(1986),TheEffectofVerbalandPictorialAdvertisingStimulionAesthetic,UtilitarianandFamiliarityPerceptions,JournalofAdvertising,15(2),27-34.Hoch,StephenJ.,andYoung-WonHa(1986),ConsumerLearning:AdvertisingandtheAmbiguityofProductExperience,JournalofConsumerResearch,13(September),221-233.Holbrook,MorrisB.,andJoelHuber(1979),SeparatingPerceptualDimensionsfromAffectiveOvertones:AnApplicationtoConsumerAesthetics,JournalofConsumerResearch,5(March),272-83.Holbrook,MorrisB.,andWilliamL.Moore(1981),FeatureInteractionsinConsumerJudgmentsofVerbalversusPictorialPresentations,JournalofConsumerResearch,8(June),103-13.Holbrook,MorrisB.,andRobertM.Schindler(1994),Age,Sex,andAttitudeTowardthePastasPredictorsofConsumersAestheticTastesforCulturalProducts,JournalofMarketingResearch,31(August)412-22.Hsee,ChristopherK.(1995),ElasticJustification:HowTemptingbutTask-IrrelevantFactorsInfluenceDecisions,OrganizationalBehaviorandHumanDecisionProcesses,62(June),330-37._________(1996),ElasticJustification:HowUnjustifiableFactorsInfluenceJudgments,OrganizationalBehaviorandHumanDecisionProcesses,66(April),122-29.Hutchinson,J.Wesley,andJosephW.Alba(1991),IgnoringIrrelevantInformation:SituationalDeterminantsofConsumerLearning,JournalofConsumerResearch,18(December),325-345.Jordan,PatrickW.,B.Thomas,andIanL.McClelland(1996),IssuesforUsabilityEvaluationinIndustry:SeminarDiscussions,UsabilityEvaluationinIndustry,ed.PatrickW.Jordan,London:TaylorandFrancis,237-43.KahnemanDaniel,andShaneFrederick(2002),RepresentativenessRevisited,HeuristicsandBiases:ThePsychologyofIntuitiveJudgment,eds.ThomasGilovich,DaleGriffin,andDanielKahneman,NewYork:CambridgeUniversityPress,49-81.Kalins,Dorothy(2003),Design2004:ItsCooltoBeWarm,Newsweek,October27,2003,58-73.

PAGE 94

84Koehler,JonathanJ.(1993),TheInfluenceofPriorBeliefsonScientificJudgmentsofEvidenceQuality,OrganizationalBehaviorandHumanDecisionProcesses,56(October),28-55.Kruglanski,ArieW.(1990),MotivationsforJudgingandKnowing:ImplicationsforCausalAttribution,TheHandbookofMotivationandCognition,eds.E.ToryHigginsandR.M.Sorrentino,NewYork:TheGuilfordPress,333-68._________(1996),MotivatedSocialCognition:PrinciplesoftheInterface,SocialPsychology:HandbookofBasicPrinciples,eds.E.ToryHigginsandArieW.Kruglanski,NewYork:TheGuilfordPress,493-520.Kunda,Ziva(1990),TheCaseforMotivatedReasoning,PsychologicalBulletin,108(November),480-98.Landy,David,andHaroldSigall(1974),BeautyisTalent:TaskEvaluationasaFunctionofthePerformersPhysicalAttractiveness,JournalofPersonalityandSocialPsychology,29(March),299-304.Langlois,JudithH.,LisaE.Kalakanis,AdamJ.Rubenstein,AndreaD.Larson,MonicaJ.Hallam,andMonicaT.Smoot,(2000),MaximsorMythsofBeauty:AMeta-AnalyticandTheoreticalReview,PsychologicalBulletin,126(May),390-423.Loewy,Raymond(1951),NeverLeaveWellEnoughAlone,NewYork:Simon&Schuster.Loken,Barbara,IvanRoss,andRonaldL.Hinkle(1989),ConsumerConfusionofOriginandBrandSimilarityPerceptions,JournalofPublicPolicyinMarketing,5,195-211.Lord,CharlesG.,LeeRoss,andMarkR.Lepper(1979),BiasedAssimilationandAttitudePolarization:TheEffectsofPriorTheoriesonSubsequentlyConsideredEvidence,JournalofPersonalityandSocialPsychology,37(November),2098-19.Mandel,Naomi,andEricJ.Johnson(2002),WhenWebPagesInfluenceChoice:EffectsofVisualPrimesonExpertsandNovices,JournalofConsumerResearch,29(September),235-45.Markus,Hazel,andZivaKunda(1986),StabilityandMalleabilityoftheSelf-Concept,JournalofPersonalityandSocialPsychology,51(May),858-66.Martindale,Colin(1984),ThePleasuresofThought:ATheoryofCognitiveHedonics,TheJournalofMindandBehavior,5(winter),49-80.Martindale,Colin,andKathleenMoore(1988),Priming,Prototypicality,andPreference,JournalofExperimentalPsychology:HumanPerceptionandPerformance,14(November),661-70.

PAGE 95

85Miller,DaleT.,andMichaelRoss(1975),Self-ServingBiasesinAttributionofCausality:FactorFiction?PsychologicalBulletin,82(March),213-25.Miller,Mark,andJerryAdler(2003),Design2004:IsaacHitsHisTarget,Newsweek,October27,2003,74-77.Mueller,Ulrich,andAllanMazur(1996),FacialDominanceofWestPointCadetsasaPredictorofLaterMilitaryRank,SocialForces,74(March),823-50.Nickerson,RaymondS.(1998),ConfirmationBias:AUbiquitousPhenomenoninManyGuises,ReviewofGeneralPsychology,2(June),175-220.Nisbett,RichardE.,andTimothyD.Wilson(1977),TheHaloEffect:EvidenceforUnconsciousAlterationofJudgments,JournalofPersonalityandSocialPsychology,35(April),250-56.Norman,D.A.(2004).Emotionaldesign:WhyWeLove(orHate)EverydayThings.NewYork:BasicBooks.NovaVideo(2003),BattleoftheX-Planes:InsidetheHigh-Tech,High-StakesCompetitiontoCreateAmericasNewestFighterPlane,Boston:WGBHEducationalFoundation.Page,Christine,andPaulM.Herr(2002),AnInvestigationoftheProcessesbyWhichProductDesignandBrandStrengthInteracttoDetermineInitialAffectandQualityJudgments,JournalofConsumerPsychology,12(2),133-47.Pham,MichelTuan,JoelB.Cohen,JohnW.Pracejus,andG.DavidHughes(2001),AffectMonitoringandthePrimacyofFeelingsinJudgment,JournalofConsumerResearch,28(September),167-88.Prasso,Sheridan(2001),BoeingGetsBeat,Businessweek,October29,12.Pyszczynski,Tom,andJeffGreenberg(1987),TowardanIntegrationofCognitiveandMotivationalPerspectivesonSocialInference:ABiasedHypothesis-TestingModel,AdvancesinExperimentalSocialPsychology,Vol.20,ed.L.Berkowitz,SanDiego:AcademicPress,297-340.Raghubir,Priya,andEricGreenleaf(2006),RatiosinProportion:WhatShouldtheShapeofthePackageBe?JournalofMarketing,70,forthcoming.Rhodes,Gillian,KateHarwood,SakikoYoshikawa,MiwaNishitani,andIanMcLean(2002),TheAttractivenessofAverageFaces:Cross-CulturalEvidenceandPossibleBiologicalBasis,FacialAttractiveness,eds.GillianRhodesandLeslieA.Zebrowitz,WestportCT:AblexPublishing,35-58.

PAGE 96

86Rubenstein,AdamJ.,JudithH.Langlois,andLoriA.Roggman(2002),WhatMakesaFaceAttractiveandWhy:TheRoleofAveragenessinDefiningFacialBeauty,FacialAttractiveness,eds.GillianRhodesandLeslieA.Zebrowitz,WestportCT:AblexPublishing,1-34.Russo,J.Edward,VictoriaHustedMedvec,andMargaretG.Meloy(1996),TheDistortionofInformationDuringDecisions,OrganizationalandHumanDecisionProcesses,66(April),102-10.Russo,J.Edward,MargaretG.Meloy,andVictoriaHustedMedvec(1998),PredecisionalDistortionofProductInformation,JournalofMarketingResearch,35(November),438-52.Schwarz,Norbert(2002),FeelingsasInformation:MoodsInfluenceJudgmentsandProcessingStrategies,HeuristicsandBiases:ThePsychologyofIntuitiveJudgment,eds.ThomasGilovich,DaleGriffinandDanielKahneman,NewYork:CambridgeUniversityPress,534-47.Schwarz,Norbert,andGeraldL.Clore(1983),Mood,Misattribution,andJudgmentsofWell-Being:InformativeandDirectiveFunctionsofAffectiveStates,JournalofPersonalityandSocialPsychology,45(September),513-23.Schweitzer,MauriceE.,andChristopherK.Hsee(2002),StretchingtheTruth:ElasticJustificationandMotivatedCommunicationofUncertainInformation,TheJournalofRiskandUncertainty,25(September),185-201.Shiv,Baba,andAlexanderFedorikhin(1999),HeartandMindinConflict:TheInterplayofAffectandCognitioninConsumerDecisionMaking,JournalofConsumerResearch,26(December),278-92.Silvera,DavidH.,RobertA.Josephs,andR.BrianGiesler(2002),BiggerisBetter:TheInfluenceofPhysicalSizeonAestheticPreferenceJudgments,JournalofBehavioralDecisionMaking,15(July),189-202.Slovic,Paul,andSarahLichtenstein(1971),ComparisonofBayesianandRegressionApproachestotheStudyofInformationProcessinginJudgment,OrganizationalBehaviorandHumanPerformance,6(November),649-744.Takahashi,Shigeko(1995),AestheticPropertiesofPictorialPerception,PsychologicalReview,102(October),671-83.Tetlock,PhilipE.,andArielLevi(1982),AttributionBias:OntheInconclusivenessoftheCognition-MotivationDebate,JournalofExperimentalSocialPsychology,18(January),68-88.Todorov,Alexander,AnesuN.Mandisodza,AmirGoren,andCrystalC.Hall(2005),InferencesofCompetencefromFacesPredictElectionOutcomes,Science,10(June),1623-26.

PAGE 97

87Veryzer,RobertW.Jr.,andJ.WesleyHutchinson(1998),TheInfluenceofUnityandPrototypicalityonAestheticResponsetoNewProductDesigns,JournalofConsumerResearch,24(March),374-94.Vriens,Marco,GerardH.Loosschilder,EdwarRosbergen,andDickR.Wittink(1998),VerbalversusRealisticPictorialRepresentationsinConjointAnalysiswithDesignAttributes,JournalofProductInnovationManagement,15(September),455-67.Wilson,TimothyD.,andNancyBrekke(1994),MentalContaminationandMentalCorrection:UnwantedInfluencesonJudgmentsandEvaluations,PsychologicalBulletin,116,(July),117-42.Yamamoto,Mel,andDavidR.Lambert(1994),TheImpactofProductAestheticsontheEvaluationofIndustrialProducts,JournalofProductInnovationManagement,11(September),309-24.Zajonc,RobertB.(1980),FeelingandThinking:PreferencesNeedNoInferences,AmericanPsychologist,35(February),151-75.Zajonc,RobertB.,andHazelMarkus(1982),AffectiveandCognitiveFactorsinPreferences,JournalofConsumerResearch,9(September),123-31._________(1985),MustAllAffectBeMediatedbyCognition?JournalofConsumerResearch,12(December),363-64.Zebrowitz,Leslie(1997).ReadingFaces:WindowstotheSoul?Boulder:WestviewPress.

PAGE 98

88BIOGRAPHICALSKETCHJoAndreaHoeggearnedbachelorsdegreesinEnglishandeducationfromtheUniversityofBritishColumbiainVancouver,Canada.In1998shemovedtoSanDiego,whereshecompletedaMasterofBusinessAdministrationatSanDiegoStateUniversity.In2001sheenteredthePh.D.programinmarketingattheUniversityofFlorida.ShejoinedtheUniversityofBritishColumbiaasanassistantprofessorofmarketinginJuly2006.


Permanent Link: http://ufdc.ufl.edu/UFE0015683/00001

Material Information

Title: Influence of Product Design on Information Processing
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015683:00001

Permanent Link: http://ufdc.ufl.edu/UFE0015683/00001

Material Information

Title: Influence of Product Design on Information Processing
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0015683:00001


This item has the following downloads:


Full Text












INFLUENCE OF PRODUCT DESIGN ON INFORMATION PROCESSING


By

JOANDREA HOEGG













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2006

































Copyright 2006

by

JOANDREA HOEGG


























This dissertation is dedicated to my parents.















ACKNOWLEDGMENTS

I would like to thank my advisor, Joseph W. Alba, to whom I am extremely

grateful for guidance, encouragement, and patience. Joe was incredibly generous of his

time, and I would not have succeeded without his support. I would also like to thank Alan

Cooke, Rich Lutz, and Mike Lewis for their kind and valuable assistance throughout my

time at the University of Florida. I thank Chris Janiszewski for his excellent advice, and

my external committee member, Jason Colquitt, for his insightful comments on my work.

I also express my gratitude to Mickey and Joe Belch, who encouraged me to pursue a

PhD at the University of Florida.

I am blessed with many supportive friends both inside and outside the PhD

program. At Florida, I would particularly like to thank Elise Chandon, Qiong Wang,

Julia Belyavsky, Dan Rice, Anne Magi, and Wouter Vanhouche. Outside of the

university, I thank Jackie Levy, Ann Lane, and Leah Thomson for many years of

wonderful friendship. Also I am deeply grateful to my family for supporting my decision

to return to school to pursue my PhD. Finally, I would like to thank Timothy Silk for

believing in me all those times that I did not believe in myself, and for teaching me

always to look on the bright side. I could not have done this without him.















TABLE OF CONTENTS

Page

A C K N O W L E D G M E N T S ................................................................................................. iv

L IST O F TA B LE S ................ .......................................... ........... .... ..... viii

A B S T R A C T ......... .................................. ................................................. ix

CHAPTER

1 IN TR O D U C T IO N ................... .... .......................... .. ............ ............. ....

2 REVIEW OF RESEARCH ON APPEARANCE........................ ... ................4

A esth etic Infl u en ce .................. .... ........................................................ .............. ... 5
Evidence from Research on Attractiveness................................ ..................5
Evidence from Other D om ains .......................................................... ..... .......... 7
A effect versus evaluation ............................................ ........................... 7
Motivated reasoning ..... ........................................................11
H alo effects ................ ................. ................................... 17
Awareness of affective and motivational influences............... ................17
Non-Aesthetic Influence of Product Appearance ....................................................18
Sum m ary ................................... ................................... ......... 21

3 OVERVIEW OF EXPERIMENTS ........................................ .......................... 23

4 PILO T STU D Y .................. .................. ......................... .... .. ....... ..... 26

M eth o d ............................................................................................................ 2 7
P rete st ............................................................................ 2 7
D esign and Procedure ............................................................. ............... 28
R results .................. 0................................................
Backpack Ratings ..................................... ......... .................. 30
R running Shoe R atings .................................................. ............................. 31
D iscu ssio n ........................................................................................3 2

5 EXPERIMENT 1 .......................................... .......... ............ 36

M eth o d ..................................................................................................... ....... . 3 6
S tim u li ............................................................................................................ 3 6
D esign and Procedure .............................................. ....... ........................ 36









R e su lts .............. ...... ........... ....................................... ......................... ............3 7
D iscu ssio n ......... ...... ............ ................................... ...........................3 8

6 EXPERIM ENT 2 ...................................... .......... ........ ..... 41

M ethod .................................................................. ............... 42
S tim u li ............................................................................................................ 4 2
Design and Procedure .............. ............ ..... ....................... 43
Results ............. ..................... ............. ...............44
D iscu ssio n ......... ...... ............ ................................... ...........................4 5

7 EXPERIM ENT 3 .................................... ..... .......... ....... ..... 48

M ethod .................................................................. ............... 48
S tim u li ............................................................................................................ 4 8
Design and Procedure ............... ........... ..... ....................... 48
Results ............. ..................... ............. ...............49
D iscu ssio n ......... ...... ............ ................................... ...........................5 0

8 EXPERIM ENT 4 ...................................... .......... ........ ..... 53

M ethod .................................................................. ............... 54
S tim u li ............................................................................................................ 5 4
D design and Procedure ........ ...... ................... ...... ........ ................ 55
Results ............. ..................... ............. ...............56
D iscu ssio n ......... ...... ............ ................................... ...........................5 8

9 EXPERIM ENT 5 ..................................... .......... ........ ..... 61

Method ....................... ...................... ...............62
S tim u li ............................................................................................................ 6 2
Design and Procedure ............... ........... ..... ....................... 62
Results ............. ..................... ............. ...............62
D iscu ssio n ......... ...... ............ ................................... ...........................6 3

10 G EN E R A L D ISC U SSIO N ........................................ ..........................................65

Im p licatio n s ................................................................6 7
L im itatio n s ................................................................6 8
F utu re R research ................................................................69

APPENDIX

A PRODUCT PICTURES USED IN PILOT STUDY ...................................... 71

B PRODUCT PICTURES USED IN EXPERIMENT 1 ................... ........... 72

C PRODUCT PICTURES USED IN EXPERIMENTS 2 THROUGH 5 ...................73









D PRODUCT DESCRIPTIONS USED IN EXPERIMENT 2, 4 AND 5....................75

E PRODUCT DESCRIPTIONS USED IN EXPERIMENT 3 ....................................77

F PRODUCT PICTURES USED IN EXPERIMENT 4.........................................79

L IST O F R E FE R E N C E S ......... ..... ............ ................. ................................................8 1

BIOGRAPHICAL SKETCH ............. ...................... ............. 88















LIST OF TABLES

Table Page

4-1. Pilot Study Stim uli ................................... ........................... .. ...... 34

4-2. Pilot Study Backpack Feature Ratings ............................................ ............... 35

4-3. Pilot Study Choice Proportions of Superior Backpack ............................................35

4-4. Pilot Study Running Shoe Feature Ratings ..................................... .................35

4-5. Pilot Study Choice Proportions of Superior Running Shoe ...............................35

5-1. Running Shoe Feature Descriptions Used in Experiment 1 ....................................40

5-2. Experiment 1 Running Shoe Feature Ratings ..................................................40

5-3. Experiment 1 Choice Proportions of Superior Running Shoe.................................40

6-1. Pretest Results for Perform ance Pictures............................... ................. .............. 46

6-2. Pretest Results for Feature Descriptions..... ............... ...................46

6-3. Results of Experim ent 2 ..................................................................... 46

6-4. M manipulation Check for Experim ent 2 ............................................ ............... 47

7-1. D im tensions U sed in Experim ent 3 ....................... ................................................ 51

7-2. R results of Experim ent 3 ........................................ .................. ............... 51

7-3. Manipulation Check for Experiment 3 ..................... ...........................52

7-4. Results of Experiment 3 with Helmet Replicate Removed ..................................52

8-1. P retest P picture R ratings ............................................................... ........................ 60

8-2 R results of E xperim ent 4 ..................................................................... ..................60

8-3. M manipulation Check for Experim ent 4 ............................................ ............... 60

9-1. R results of E xperim ent 5 ..................................................................... .................. 64















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

INFLUENCE OF PRODUCT DESIGN ON INFORMATION PROCESSING

By

JoAndrea Hoegg

August 2006

Chair: Joseph W. Alba
Major Department: Marketing

Despite the growing importance of product design in the marketplace, relatively

little consumer research has been conducted addressing the influence of product

appearance on the processing of other product information. The current investigation

presents the results of five experiments that pit product appearance against functional

feature descriptions to examine how conflicting visual information and verbal

information are reconciled. In all studies participants were asked to compare the quality

of two brands of a given product based on a reading of allegedly unbiased professional

reviews. The primary experimental manipulation was that for some participants the

reviews were accompanied by pictures of the two brands. For each pair of products, one

brand was superior visually but was actually inferior based on the objective review, and

one brand was inferior visually but was objectively superior. In some experiments the

visual superiority was aesthetic; that is, one product was more attractive than the other.

In other experiments the visual superiority was performance based: one product looked

like it would offer superior performance on a given dimension. The primary finding was









that product appearance exerted an influence on people's tendency and ability to evaluate

product quality, but only when the visual difference was performance based and not

aesthetically based. Diagnosticity appeared to be the driving force.














CHAPTER 1
INTRODUCTION

Product design is of critical importance for managers. In an era of sophisticated

manufacturing processes, where it is difficult for companies to gain competitive

advantages in other areas such as price or reliability (Jordan, Thomas and McClelland

1996; Kalins 2003), design can act as a "strategic tool" that firms can use as a point of

differentiation (Page and Herr 2002, p. 133). Moreover, with designer styles available at

affordable prices (e.g., Isaac Mizrahi in Target; Tommy Hilfiger in JC Penney), the

average consumer is now able to purchase and hence demand products that are not only

functionally sound, but also aesthetically pleasing (Miller and Adler 2003).

Marketers and academics agree that product design is important, and there exists a

large body of managerial and quantitative research on design as a component of the new

product development process. Consumer research on design, however, has been

somewhat limited. Consumer behavior investigations into product appearance have

included examinations of behavioral responses to product form (Silvera, Josephs, and

Giesler 2002; Veryzer and Hutchinson 1998), the interaction of brand strength and design

(Page and Herr 2002), the effects of repeated exposure to designs (Cox and Cox 2002),

and the role of individual differences in response to design features (Bloch, Brunel and

Arnold 2003; Holbrook and Schindler 1994). Bloch (1995) developed a model of

consumer response to product form that outlined several stages in the design process

from managerial goals through the creation of the product's form to the consumer's

response. Bloch's model addressed the possibility of both cognitive and affective









responses to product form, but did not speak specifically to how variations in product

design might influence the nature of the cognitive or affective reactions. In a discussion

of the different roles of product design, Creusen and Schoormans (2005) similarly

mentioned the psychological role that product appearance might play but did not offer

predictions as to the nature and direction of the influence.

The current research contributes to the growing literature on product design by

investigating how a product's appearance can influence the processing of functional

product information. I examine the extent to which design signals quality, and how such a

signal might interfere with the evaluation of more objective product information. Prior

consumer research in product design has examined how aesthetics could influence quality

evaluations in the absence of other information (Page and Herr 2002). Other work has

also looked at differences between visual and verbal information, primarily in terms of

picture and word use in advertising (e.g., Hirschman 1986; Holbrook and Moore 1981).

However, no research has specifically addressed how the appearance of a product might

alter the evaluation of objective product information. In addition, prior research has not

addressed the possibility of different types of design influence. The research reported

here compares the influence of aesthetic design, or design for the sake of beauty, to the

influence of performance-based design, or design as a signal of product performance, and

investigates how these two roles might exert differential influence on consumer

evaluations.

Five experiments were conducted that pitted design against written product

information to examine how conflicting visual and verbal information would be

reconciled. In all studies participants were provided with objective reviews about two









brands and were asked to evaluate the brands on particular features based on the reviews.

Accompanying the two reviews were pictures of the brands that conflicted with the

review. That is, the brand that was better in terms of feature descriptions was inferior

visually, and the brand that was worse based on feature descriptions had the superior

design. The visual superiority had two forms. In some studies the superiority of the

design was aesthetic: one product was more attractive than the other. In other

experiments the superiority of the design was performance based; one product looked like

it would offer superior performance along certain dimensions.

The paper is organized as follows. Chapter 2 provides a review of literature on

visual effects in marketing and psychology, as well as other relevant research to outline

the scope of the investigation and inform the key questions. Because the studies

distinguish aesthetic design from performance-based design the literature review is

divided into two sections based. First is a review of aesthetic effects, where the influence

of an object's appearance relates specifically to its attractiveness. Second is a review of

performance-based appearance effects; that is, where an object's appearance

communicates some level of performance or quality. Chapter 3 provides an overview of

the studies. Chapters 4 through 9 present the methodology, data analysis, and findings of

a pilot study and five experiments. Chapter 10 includes a general discussion of the key

results, along with marketing implications, limitations, and avenues for future research.














CHAPTER 2
REVIEW OF RESEARCH ON APPEARANCE

Dating back to classical times, scholars have considered the aesthetics of the visual

form (Dickie 1971), mostly attempting to determine what is considered beautiful and

why. In more recent years, psychologists have attempted to determine experimentally

what elements of an object make it attractive or unattractive (e.g., Berlyne 1974; Green

1995; Langlois et al. 2000; Martindale 1984; Martindale and Moore 1988; Rhodes et al.

2002; Rubenstein, Langlois and Roggman 2002; Zebrowitz 1997), and within marketing,

a small but growing number of researchers have joined the quest to determine what

design elements consumers find attractive (e.g., Cox and Cox 2002; Raghubir and

Greenleaf 2006; Veryzer and Hutchinson 1998).

Separate from the literature on aesthetics per se, several streams of research exist

that examine, directly or indirectly, ways in which appearance might influence the

processing of other information. Research addressing this issue comes primarily from

face perception, which has demonstrated that appearance can influence ratings of

personality (Ambady and Rosenthal 1993; Eagly et al. 1991), job suitability (Hassin and

Trope 2000) and even work quality (Landy and Sigall 1974; Mueller and Mazur 1996). In

the product realm, there is evidence that even clearly irrelevant design elements like an

invitation card or a web page background can have an influence on product preference

(Mandel and Johnson 2002; Raghubir and Greenleaf 2006). Although such effects have

been observed on preference rather than more objective ratings, the presence of influence

in an unequivocally irrelevant context suggests the potential power of design.









Whether there are conditions under which design might exert influence over

objective product quality judgments remains largely unresolved. Some research in

marketing has shown that the attractiveness of a product design can influence product

quality ratings, although the influence appears to occur at the margins (Page and Herr

2002; Yamamoto and Lambert 1994).

To help illuminate the issue, the review presented here will focus on research in

person and object appearance, but will also borrow from several literature streams not

centered on the visual form, notably work in motivated reasoning, halo effects, the

interplay of affect and cognition, and accessibility-diagnosticity. Although these

literatures do not specifically address appearance, they are useful inasmuch as they speak

to the potential impact a stimulus can have on information processing.

This review divides the topic of the influence of appearance into two parts. The

first section focuses specifically on aesthetics, that is, on appearance as a source of beauty

or pleasure. The second section focuses on appearance as a signal of quality or

performance.

Aesthetic Influence

Evidence from Research on Attractiveness

There is a fairly large body of research demonstrating effects that physical

attractiveness can have on judgments of other people's personality traits and capabilities

(Eagly et al. 1991; Landy and Sigall 1974; Rubenstein et al. 2002; Zebrowitz 1997). For

example, attractive people are perceived as having superior skills such as leadership,

social ability and job competence (Langlois et al. 2000). There is even some research

suggesting that physical attractiveness can influence evaluations of work quality. Landy

and Sigall (1974) had participants rate essays supposedly written by students whose









pictures accompanied the essays. Despite the obvious irrelevance of appearance to the

quality of an essay, participants rated both the essays and the writing skills of more

attractive people as superior to those of less attractive people.

To a far more limited extent, attractiveness effects have been found in the consumer

realm. Raghubir and Greenleaf (2006) demonstrated a dramatic attractiveness effect in an

experiment that investigated aesthetic preference for the golden ratio of rectangles.

Participants were presented with an invitation to a concert. The only difference between

experimental conditions was that the invitation was printed on rectangular cards of one of

two ratios: 1.38:1 or 1.62:1 (the golden ratio). Participants receiving the invitation

conforming to the golden ratio rated the concert as more harmonious and indicated a

higher likelihood of purchasing a CD of the performance than did participants receiving

the invitation with a 1.38:1 ratio.

Page and Herr (2002) have suggested that the attractiveness of a product design

produces an affective response associated with feeling-based, relatively automatic

evaluative processes, while the quality of the product is based more on higher-order

cognitive evaluation. In support of their results, Page and Herr (2002) found that liking

judgments occurred faster than quality judgments. More importantly, product liking was

driven by reaction to the product's design and was not influenced by brand name or

functionality, whereas evaluation of the quality of the product seemed to integrate design,

function and brand information.

In terms of product evaluation, the design of a product may be impossible to

ignore, even when the evaluation of the product should be totally independent of

attractiveness. Some evidence for this comes from a study on industrial products









conducted by Yamamoto and Lambert (1994). They had business people in various

organizational positions from sales to engineering indicate preferences for various

industrial products. Products were presented on cards consisting of product attribute

specifications and a photograph. Through a conjoint analysis technique Yamamoto and

Lambert found that the attractiveness of the product did have an influence, even for

engineers and buyers who were well informed about product characteristics. Across

seven industrial products the more attractive product was significantly favored. This

influence did not appear to be non-normative, however, since attractiveness was one of

the lowest variables found to affect choice and the influence was only at the margins.

Nonetheless, an influence of attractiveness was present despite overwhelming self-reports

that attractiveness should play no role in the selection of a superior product. Considering

that the products were industrial and the participants were experts, the finding that

attractiveness played any role at all is srtiking, and suggests that such an influence may

exist for consumers.

Evidence from Other Domains

There are several streams of research outside the aesthetics literature showing that

an initial reaction to a stimulus can influence its subsequent evaluation. Three areas are

directly relevant to the current discussion: the interplay of affective reactions and

cognitive evaluations, motivated reasoning, and halo effects. Each of these is reviewed

briefly.

Affect versus evaluation

Many authors have argued that affect and evaluation are distinct concepts, with

affect referring to feeling states that arise due to interaction with a stimulus and

evaluation referring more to a cognitive process of judgment based on beliefs (Breckler









and Wiggins 1989; Cohen and Areni 1991). Both can occur for a given object, although

some have argued that affective reactions can occur entirely without cognitive

evaluations (Berkowitz 1993; Zajonc 1980; Zajonc and Markus 1982, 1985).

In an attempt to disentangle affective reaction from a more cognitive appraisal,

Breckler and Wiggins (1989) asked participants to evaluate blood donations on different

dimensions. They asked for a series of relatively cognitive evaluations of blood

donations, using semantic differential scales of the extent to which donating blood was

useful, good, wise, etc. They also asked a series of feelings-based questions regarding

how people felt when they gave blood. The key finding was among non-donors, who had

positive cognitive evaluations of blood donation, but negative affective reactions toward

blood donation, indicating that their affective and cognitive evaluations were distinct.

There is evidence to suggest that an affective reaction to a stimulus is resistant to

change (Edwards 1990) and may be a better predictor of the type of thoughts people have

about products than cognitive evaluations (Pham et al. 2001). As suggested by Zajonc

(1980) and Edwards (1990), a strong affective response may be highly resistant to

product information that contradicts the individual's feeling about the product. Edwards

(1990) examined the resistance to change of both cognitively and effectively formed

attitudes, looking specifically at how attitudes formed through affective means or through

cognitive means could be altered using affective or cognitive persuasion. She

demonstrated that the effectiveness of a rational (cognitive) or emotional (affective)

persuasive message depended on the nature of the attitude's origin. Attitudes originally

formed through an affective process were resistant to rational persuasive attempts and

could only be altered through affective means, supporting Zaj once's (1980) argument that









affective judgments are not easy to change. Cognitive attitudes (those formed through

rational means) on the other hand, could be altered through both cognitive and affective

means of persuasion.

Edwards (1990) used primacy to manipulate whether an attitude was formed

through affective or cognitive means, assuming that the attitude formation occurred with

the first persuasive exposure. If valid, this possibility has relevance when considering the

affective reaction to product design versus a cognitive appraisal. In the marketplace it is

common for consumers to see a product, through advertising or store displays, prior to

learning about its features. If a person were to encounter a product's design first, and then

later read information about product function, the attitude formed during the observation

of the design could be resistant to contrary information in the written description. On the

other hand, if a person read the product information before seeing the design itself, any

conflicting affective reaction to the design might still influence the attitude formed during

the reading of the information, although it is likely the influence would be relatively

weaker.

Pham et al. (2001) examined the role of affective reaction to a product and

compared that reaction to a more reasoned, cognitive analysis. They predicted that feeling

responses (affective reactions) would be better predictors of the number and valence of

thoughts people have about products than reason-based responses. In accordance with

models of affective response (e.g. Berkowitz 1993; Cohen and Areni 1991), Pham et al.

proposed that feeling-based responses are generally faster than reason-based responses,

and that an initial affective response prompts thought generation through "both automatic

and controlled processes" (p. 170). In study 1 (Pham et al. 2001), participants evaluated









pictures from magazines using either a reason-based analysis or a feeling-based analysis.

The authors found that feeling responses were faster than cognitive responses, suggesting

that the pictures triggered a relatively automatic affective reaction. In study 2 Pham et al.

(2001) examined the number of valenced thoughts generated through feeling-based or

reason-based analysis. The valence of subjects' spontaneously generated thoughts was

more strongly related to the valence of the feeling-based responses than the reason-based

responses.

Pham et al. (2001) also found that feeling-based assessments showed more

interpersonal agreement than reason-based assessments. They concluded that feeling-

based evaluations are more natural and more consistent. This suggestion is consistent

with Takahashi (1995), whose work in the psychology of aesthetics found that people

were very good at guessing the emotion portrayed through a drawing, and also with

research in facial attractiveness showing that people generally agree in their ratings of

attractive and unattractive faces (Rubenstein et al. 2002).

The affect-as-information model proposes that people use their feelings during

judgment because they believe their feelings to be relevant to the judgment task (Schwarz

2002; Schwarz & Clore, 1983). Such feelings could stem from the product itself, with the

product's design acting as a form of mood manipulation (Norman 2004). A visually

pleasing design could enhance mood, causing people to be more creative problem solvers

and more willing to ignore details or small problems. An unattractive design could

depress mood, leading people to be more analytical and heightening their tendency to

expect and address problems. According to this logic, visually pleasing products could

actually appear to work better.









Shiv and Fedorikhin (1999) demonstrated how an affective reaction to a stimulus

can dominate a cognitive evaluation when cognitive resources are constrained. In

accordance with Berkowitz (1993) the authors proposed that upon exposure to a choice,

both affective and cognitive processes may be engendered; the affective reaction

occurring relatively automatically and the cognitive processes occurring in a more

controlled, resource-taxing manner. Due to this automatic nature of the affective reaction,

when processing resources are constrained, only cognitive processing is impaired; thus,

affective processes dominate.

The interplay of affect and cognition can be considered in the context of product

aesthetics. A product's design is quickly and easily perceived and may prompt an

immediate affective reaction prior to a cognitive appraisal of product features. The

affective reaction may be consistent or inconsistent with a later more conscious or

controlled evaluation. The extent to which the later reasoned analysis would correct for

the initial reaction is uncertain. The feeling-based reaction may be difficult to alter

(Edwards 1990), or, in accordance with recent discussions of dual processes, the

cognitive appraisal may be able to review and correct an initial assessment (e.g.,

Kahneman and Frederick 2002).

Motivated reasoning

Motivated reasoning is said to occur when a personal desire for a particular

outcome interferes with the tendency or ability to evaluate choices objectively. The

notion that motivations (hopes, fears, desires, etc.) can affect judgments has been studied

and debated for decades (e.g. Ditto and Lopez 1992; Kruglanski 1990; Miller and Ross

1975; Pyszczynski and Greenberg 1987; Tetlock and Levi 1982). Although some have

argued that what is called motivated reasoning can easily be explained entirely in non-









motivational, or purely cognitive terms based on prior beliefs or expectations (e.g., Miller

and Ross 1975), the motivational camp continues to produce compelling evidence that

motivation does play a role in judgment (e.g. Boiney, Kennedy and Nye 1997; Ditto and

Lopez 1993; Markus and Kunda 1986). For example, in a study on business decisions,

Boiney et al. (1997) reported that people tended to bias their judgments more or less as

needed to support their desired conclusions. If people were acting in a purely cognitive

manner, and simply coming to the most probable result given prior beliefs and

expectations, it is unlikely they would have adjusted the degree of bias depending on

what was needed to arrive at the desired conclusion.

In her review of the literature, Kunda (1990) suggested that motivation and

cognition interact, with motivation affecting reasoning through reliance on a biased set of

cognitive processes for accessing, constructing and evaluating beliefs. Kunda (1990)

separated motivated reasoning into two categories: accuracy goals, for which the motive

is to arrive at an accurate conclusion and directional goals, for which the motive is to

arrive at a particular desired conclusion. She argued that both kinds of goals influence the

choice of beliefs and strategies applied to a given problem. Accuracy goals lead to the use

of beliefs and strategies considered most appropriate for the task; directional goals lead to

the use of those that are considered most likely to result in the desired conclusion. Thus,

as suggested by Kruglanski (1996) there is interplay of cognition and motivation:

Motivation determines which cognitive processes will be used in a given situation.

In a product context, when one choice is aesthetically preferable to another there

may be a motivation, perhaps below awareness, to under-critique the attractive product

and over-critique the unattractive product, resulting in a choice consistent with the









aesthetic preference, but possibly non-normative. This type of motivated reasoning would

be driven by a directional goal created by the product aesthetics. However, as noted by

several authors (Boiney et al. 1997; Hsee 1995; Kunda 1990), people generally want to

make rational, correct decisions; thus, their desires are constrained by reasonableness.

The extent to which motivations influence judgments can be largely dependent on the

strength and clarity of the information to be judged. When ambiguity is present

motivation may play a larger role.

Ambiguity of product information. Hsee (1996) has reported that if decision

criteria are somewhat ambiguous, people will use factors that they should not in order to

achieve a motivated outcome. In his elasticity hypothesis (Hsee 1995; 1996; Schweitzer

and Hsee 2002), Hsee defines two types of judgment factors, justifiable factors, which are

relevant to the task and whose use can be justified, and non-justifiable factors, which are

not relevant to the task but are personally appealing. When two options are presented,

where one option is superior on a justifiable factor and the other is inferior on the

justifiable factor but is superior on an unjustifiable factor, people's choices will depend

on the extent to which the justifiable factor's values are difficult to evaluate. That is,

when the values of the justifiable factor are fixed and clear, people will choose the option

that is superior on that factor; however, when the level of the justifiable factor is not

clearly indicated (e.g., is presented in a range), people will use the unjustifiable factor to

choose the more personally attractive option. This occurs because without ambiguity in

the justifiable factor, people have no excuse for considering the unjustifiable factor in

their decisions. When the justifiable factor is presented in a range, people can distort their

evaluation of this factor in the direction of the unjustifiable factor. This makes their









choice appear to be based on the justifiable factor when it is actually based on the

unjustifiable factor.

Hsee (1995) demonstrated elastic justification in an experiment using a cover

story of a proofreading assignment. Participants proofread ads and were paid on a per

error basis. The ads were of two forms: furniture ads, which were uninteresting but had

more errors per page, and personal ads, which were much more interesting, but had fewer

errors per page, and participants could choose which type of ads to read. The dependent

measure was the number of pages read from each type. When the number of errors per

page was fixed, participants behaved normatively and chose to read the furniture ads.

However, when the number of errors per page was presented in a range, proofreaders

were more likely to choose the personal ads. Presumably, presenting the justifiable factor

in a range made it easier to choose the more personally appealing but objectively inferior

option.

Russo and his colleagues (Russo, Medvec, and Meloy 1996; Russo, Meloy, and

Medvec 1998) used the term "predecisional distortion" to describe a form of biased

reasoning that occurs when evaluating products on attributes presented in a sequence.

The authors demonstrated that a preexisting preference for one option could lead to the

distortion of ambiguous information to conform to the prior preference.

In the typical predecisional distortion paradigm, an initial preference for one of two

products would be created through some kind of personal relevance manipulation. Thus,

upon engaging in the evaluation task, subjects would already be motivated to choose a

particular product. Relatively ambiguous product information would then be evaluated

and (in most cases) a final choice would be made. Results indicated that feature









information was distorted to conform to the prior preference. As noted by the authors,

predecisional distortion results are consistent with literature on confirmation bias, where

people distort evidence to support an initial hypothesis or prior beliefs (e.g. Darley and

Gross 1983; Deighton 1984; Hoch and Ha 1986; Nickerson 1998). Although Russo et al.

(1996, 1998) support a cognitive explanation for predecisional distortion, they

acknowledge that their original manipulation was motivational and cannot rule out a

motivational explanation. An analysis of their evidence, in particular their second

experiment (Russo et al. 1998), suggests that both motivational and cognitive factors

were operating. The basic motivated reasoning process appeared to operate initially, with

directional motivation influencing beliefs, but once the ambiguity was removed the

motivational bias disappeared and a standard confirmation bias obtained.

In some cases the distortion of judgments based on prior beliefs can be justified

from a Bayesian perspective (Koehler 1993; Lord, Ross and Lepper 1979; Russo et al.

1998). It may not always be easy to determine the diagnostic value of product attributes,

particularly when they are ambiguous, so given uncertainty, it may be perfectly logical to

use one's prior belief in the superiority of one brand (or one's preference for it) as part of

an overall evaluation. Ignoring this prior information would contradict a Bayesian

updating approach where prior beliefs are updated (but not necessarily discarded) when

new information is presented.

Individual differences. Individual differences may influence the form and extent

of motivated reasoning. The notion that some people are more oriented toward aesthetic

qualities than others was originally suggested by Loewy (1951) and continues to

permeate current aesthetics research (e.g. Bloch et al. 2003). Bloch et al. suggest that









people can be measured for their degree of centrality of visual product aesthetics

(CVPA). CVPA is the overall level of significance that visual aesthetics hold for a

particular consumer in his or her relationships with products. Consumers high in CVPA

have greater concern for visual aesthetics and may place higher weight on aesthetics as a

product dimension. These individuals assign high value to product appearance in

enhancing their personal well-being. They perceive themselves to be better able to

recognize or evaluate product designs, and they respond more strongly than others to

aesthetic elements. The level of consumers' CVPA may influence the extent to which

they are directionally motivated by product aesthetics. Because these individuals

experience stronger reactions to aesthetics, they may be more likely to engage in biased

processing of other product attributes.

Another individual difference that may moderate the aesthetic effect is expertise.

It is likely that for experts in a product category, the aesthetic appeal of the product does

not exert a motivational influence because knowledge of functional product information

is well established. Experts are less influenced by external factors and tend to process

information more deeply (Alba and Hutchinson 1987). On the other hand, the motivated

reasoning literature has demonstrated that people selectively retrieve information from

memory that supports their desired goal (Kunda 1990), suggesting that when faced with

an aesthetically pleasing product experts may selectively recall product information

supporting their choice. Indeed, recent research in website atmospherics has

demonstrated that product choices by both novices and experts are influenced by web

page background primes (Mandel and Johnson 2002), so the effect of product design on

experts is uncertain.









Halo effects

The halo effect, where global impressions of another person can unconsciously

influence subsequent evaluations of that individual's personality traits or abilities is a

well-established phenomenon (e.g. Goldman, Cowles and Florez 1983; Nisbett and

Wilson 1977). Research on halo effects has mostly focused on person perception, but it

has been demonstrated that halo effects can and do exist for products (e.g. Beckwith and

Lehman 1975; Holbrook and Huber 1979). Holbrook and Huber (1979) suggest that

because of the complexity of some products' attributes and the difficulty of calculating an

overall evaluation using something like the Multiattribute method, people often work in

reverse, letting their overall affective impression of the product guide their judgments of

the product's individual attributes. In terms of product aesthetics, an attractively or

unattractively designed product could create a halo that would influence subsequent

product evaluation.

Awareness of affective and motivational influences

Relatively little work has directly addressed the extent to which motivated

reasoning occurs beyond conscious awareness, although a variety of findings suggest it

does. For example, Boiney et al. (1997) found that the bias exhibited in motivated

reasoning had no impact on judgment confidence; that is, biased decision makers were no

more or less confident than unbiased decision makers. A lack of confidence could have

suggested awareness of the bias. Other indirect evidence comes from Koehler (1993),

who found that scientists were biased in evaluating the quality of research that agreed or

disagreed with their prior beliefs. Despite displaying an agreement bias, the respondents

believed they could evaluate the evidence in an unbiased manner and moreover, strongly

believed that it was important to do so. This research was focused on prior beliefs rather









than directional motivations, but it does hint at a lack of awareness of the extent to which

biased evaluation occurs.

Researchers in predecisional distortion have also claimed that awareness of the

source of their bias is unlikely (Russo et al. 1998). They argued that in their paradigm, "if

some of these sources of initial preference were recognized, consumers might dismiss

them as unacceptable bases for brand comparison" (p. 449).

Halo effect studies also suggest that people are not aware they are being influenced.

Nisbett and Wilson (1977) had college students evaluate a psychology instructor who was

speaking in a videotaped interview. Half the subjects saw an interview where the

instructor was warm and approachable, and half saw an interview where the instructor

seemed cold and unfeeling. After watching the video, participants were asked to rate the

instructor on various characteristics. Consistent with halo effects, participants rated the

individual traits more positively for the warm teacher. However, the interesting finding is

in regard to awareness. Most participants believed that their overall liking for the teacher

had not influenced their ratings, and actually believed that, if anything, the reverse

occurred, with the instructor's traits influencing overall liking (Nisbett and Wilson 1977).

Non-Aesthetic Influence of Product Appearance

Independent of attractiveness, there is research to suggest that the extent to which a

person looks appropriate for a role influences evaluations of objective criteria such as

their resume or performance. In a study on physical traits and personal characteristics,

Hassin and Trope (2000) found that people allowed inferences made on the basis of a

person's appearance to influence evaluations of the person's suitability for particular

jobs. The pictures used in the study controlled for attractiveness and rather focused on

differences in physiognomy, or the extent to which a particular face looks like it would









possess particular character traits. Of course, such ratings are necessarily correlated with

perceptions of attractiveness, but since attractiveness was controlled for during protests, it

cannot explain the results. Hassin and Trope found that people were generally unable to

avoid using physical characteristics in evaluating personality or ability traits, despite

believing that they could. In one study the authors specifically instructed participants to

ignore accompanying photographs when evaluating their resumes as potential job

candidates, and most (60%) indicated that they were able to do so. Nonetheless,

experimental evidence showed that participants were influenced by the photographs,

suggesting physical characteristics can play a role in evaluations beyond awareness and

perhaps beyond attempts to correct for them.

Other studies on faces have demonstrated biasing effects of appearance

independent of attractiveness. Mueller and Mazur (1996) report that male military

officers who appeared dominant-looking (as defined by a high forehead and a square jaw)

were more likely to be promoted within military ranks than those who appeared less

dominant-looking. Todorov et al. (2005) found effects of appearance in the political real.

The authors had study participants rate pictures of congressional candidates along seven

personality dimensions, including competence, intelligence, charisma, trustworthiness,

honesty, likeability, and leadership. They found that the competence ratings could predict

with 68.8% accuracy the winner of the election.

In a product context, products can communicate a particular level of performance

through their design. Certainly additional features visible to the eye would inform the

consumer of product performance, but even when the number of features is equivalent

across products, the design of those features can communicate different information.









Heavier construction materials are often equated with durability; size can be a signal of

power; a sleek shape can suggest aerodynamics. Such design elements could influence

the processing of functional product information in different ways. Performance-based

product design could operate much like the effects observed in physiognomy experiments

and distort the processing of other more reliable product information in a confirmation-

bias type manner. Alternatively, the information communicated through design could

provide quality cues that reduce the extent to which other product information is

weighted.

The extent to which design would be used and perhaps overweighted as a decision

cue may be a function of the extent to which it were perceived to be diagnostic to the

judgment (Feldman and Lynch 1988; Slovic and Lichtenstein 1971). If the visual

dimensions of the product, i.e., its design elements, communicate a particular level of

quality, it might be relied upon at the expense of other cues. Unlike a halo or

confirmatory processing effect, this suggests a relatively deliberate use of design, and in

general this could be relatively normative. In the market there is often a correlation

between the extent to which products looks like they will perform well and the extent to

which they actually do, so from a cost/benefit perspective, using design as a quality

heuristic may be perfectly reasonable. However, there are cases where appearance is not

an accurate signal of quality, such as with copycat brands (Hutchinson and Alba 1991;

Loken, Ross and Hinkle 1989). In such situations the question is whether consumers are

able to disregard design and attend to truly diagnostic information.

Whether through distortion or a shift in dimension weights, the potential

implications of performance-based design bias are dramatic. An example is the recent









decision by the US Department of Defense to choose Lockheed Martin over Boeing as

the manufacturer of its new X-plane. At an estimated $225 billion, with another $175

billion in potential sales to foreign allies, the contract for the next-generation fighter jets

was the largest defense contract in US history. This contract was awarded after a long

competition between Boeing and Lockheed Martin. Both companies designed fighter jets

to meet the requirements of the US Navy, such as acceleration and vertical lift capability.

The clear performance criteria should have facilitated an unbiased comparative

evaluation; however, the two companies achieved the Navy's requirements in very

different ways. One area in which they differed significantly was aesthetics. The

Lockheed Martin plane looked aerodynamic and very close to the prototypical fighter jet

shape. The Boeing plane had an unusual design that was described by a Pentagon source

as looking like "a flying frog with its mouth wide open" (Prasso 2001). Although the

Pentagon flatly denied that design played any role in the decision and there was no

evidence that the Boeing plane was in fact functionally superior, the mere suggestion

made by some industry experts (cited in Nova video 2003) that the Pentagon was

influenced by the appearance of the planes implies the potential seriousness of design

influence.

Summary

This literature review examined research highlighting different roles product

appearance could play in information processing. The first section discussed ways in

which a reaction to the aesthetic appeal of a product might bias product evaluation.

Evidence from literature in halo and attractiveness effects, affect and cognition, and

motivated reasoning seems to support the notion that product appearance could distort the

processing of functional product information, impairing consumers' ability to make









objective product judgments. On the other hand, literature focused specifically on product

aesthetics has thus far suggested that aesthetic influence on quality judgments occurs

primarily at the margins, influencing processing only in the absence of concrete

information. As yet it is unclear under what conditions product aesthetics might bias

quality judgments in the face of objective product information. Such influence may

depend on several factors, including the strength of the affective reaction to the design,

the strength of the functional product information, or individual differences.

The second section reported research suggesting that product appearance could

have an effect on judgment independent of an aesthetic influence. Most of this work has

focused on evaluations of people's faces, where recent research has indicated that job

performance by people who appear to fit a particular profile is judged differently than

that of people who do not fit the profile. The same may be true of products. That is,

design could appear to be a source of diagnostic information about the quality of the

product, and as such could be relied upon, even in cases where it should not be.














CHAPTER 3
OVERVIEW OF EXPERIMENTS

Previous research has demonstrated that product design can be a cue to quality and

that in the absence of other information such as brand name or product specifications,

product design can influence choice (e.g. Creusen and Schoormans 2005; Page and Herr

2002). However, little product design research has been conducted on how design

influences the processing of other product information, and most previous studies have

found a fairly high degree of rationality in the use of design in decision making. There is

evidence from research on facial appearance that appearance can bias evaluation of

personality traits and performance; however, under what conditions this might in the

product realm is unclear. It may be that people are deliberate in how they weight (or

overweight) design, but given the wealth of research on how people are often unaware of

how they use a variety of decision cues, it is plausible that design could exert a non-

normative effect on decision making

The studies presented in the next chapters extend the findings of the visual effects

seen with faces into the product realm. The experiments investigate several questions:

(1) does design bias information processing, (2) are there different types of design biases

and do they exert differential influence, and (3) where in the decision process does design

have the most/least influence.

The goal of the initial experiments was to determine whether design interferes with

product quality judgments. To investigate this issue, the study employed a modified

version of the predecisional distortion paradigm (Russo et al. 1996; 1998). In the studies









by Russo and his colleagues, participants were asked to compare the quality of two

brands of backpacks. Prior to the evaluation task participants were given a personal

reason to hope that one brand was superior to the other (i.e., an external motivation).

They were then presented with descriptions of the product features for each brand and

evaluated them one by one. Over time a bias toward their initial preference emerged,

despite their being asked to evaluate each feature on an individual basis, rather than the

backpack overall.

The pilot study and experiment 1 of the current investigation used a similar

procedure. However, in place of an external motivation, pictures of the designs of the

product were used to create the initial leaning toward one of the two brands. Prior to

evaluating the feature descriptions, experimental participants saw pictures of the two

brands, and one picture was significantly more attractive than the other. The subsequent

written information suggested that the unattractive product was in fact superior.

Comparing the ratings of participants who initially saw the pictures to the ratings of those

who did not see the pictures made it possible to determine whether product appearance

biased processing of feature information.

After establishing the presence of the product appearance bias, the remainder of the

dissertation focuses on how the bias occurs and on some of the conditions of its

influence. To do this, experiments 2 through 5 used a slightly different paradigm. The

largest change was that rather than presenting the feature information one attribute at a

time, in experiments 2 through 5 all attribute information was presented at one time in

paragraph format. In studies 2 and 3 the primary comparison was still between ratings

with pictures and ratings without pictures. In studies 4 and 5 comparison groups saw both






25


the pictures and the written information. In these experiments the key manipulation

consisted of altering the order of presentation of the visual and verbal information.

Comparing the evaluations of those who saw the pictures prior to the written reviews to

the evaluations of those who saw the pictures after the reviews provided insight into the

process underlying the design influence and helped to rule out some alternative

explanations.














CHAPTER 4
PILOT STUDY

The pilot study used a modified version of the predecisional distortion paradigm

(Russo et al. 1996) to investigate whether a product's appearance leads to biased

processing of product features. In Russo's experiments, participants were told they would

be evaluating two brands of backpacks and were given a personal motivation to hope that

one brand was superior to the other. They were then presented with descriptions of the

products' features, one feature at a time. For each feature, they saw a description for each

brand. Both descriptions were presented on the screen at the same time and participants

indicated which backpack was favored based on each feature. After evaluating the two

backpacks on the particular feature, participants were asked which backpack appeared to

be superior overall, based on the information they had seen up to that point. This was

referred to as the "horse race," with the backpacks representing the horses. Participants

indicated which "horse" was in the lead at that point.

The pilot study replicated the design of the Russo predecisional distortion studies.

In place of a prior personal motivation to choose one of the brands, participants in the

experimental condition were presented with pictures of the two brands before reading the

feature information. One of the brands was significantly more attractive than the other.

Two product categories were used: backpacks, to mirror the Russo studies, and

running (racing) shoes. Running shoes were chosen as a second category because they

represented a product where looks should not matter to a serious runner who is shopping

for a performance shoe.









Method

Pretest

Pictures of backpacks and running shoes were pretested to obtain pairs of brands in

which one brand was significantly more attractive than the other. To prepare for the

protests, 24 pictures of backpacks and 24 pictures of running shoes were taken from

internet websites. Using Photoshop, the pictures were manipulated to remove brand

labels and to make them uniform in size. Two rounds of pretest ratings were conducted.

First, 20 participants rated the backpacks individually on a scale from 0 (very

unattractive) to 100 (very attractive). The three most attractive and unattractive products

were selected. Next, 30 participants rated the relative attractiveness of the nine possible

combinations of pairs of unattractive and attractive products, and the pairs with the

largest difference were selected. Pictures of the final pairs of backpacks and running

shoes appear in appendix a.

In terms of feature descriptions, the goal was to have one product be superior to the

other overall, but not overwhelmingly so. For each feature, two descriptions, one for each

brand, were created and pretested. For backpacks, two of the actual feature descriptions

used by Russo were selected to mirror his procedure as closely as possible. Pretests

resulted in stimuli lists that when considered overall, favored one brand over the other,

but not always when the features were considered individually. The unattractive picture

was then paired with the superior product list, and the attractive picture with the inferior

list.

Because the attribute lists did not favor the unattractive product on every feature,

the order of presentation of the attributes was controlled so that initially there was some

ambiguity as to which product was superior. Forcing a single attribute order could have









resulted in generalizability concerns but for the purposes of the pilot study it was

considered worthwhile. The feature lists and order of presentation appear in table 4-1.

Design and Procedure

The study had two between subjects conditions, no-picture and picture. There were

two product replicates, backpacks and running shoes (order counterbalanced).

Participants in the no-picture condition saw only the written feature information. Picture

participants saw the feature information along with pictures of the two brands, which

remained on the screen throughout the procedure.

The study was conducted on computer. All information was provided in written

form on the computer screen. It was important to encourage participants to focus on

quality rather than aesthetics, so before presentation of the feature information, specific

instructions were provided. For the backpacks, participants were told that they were

shopping for a new backpack to replace one they had bought previously. The previous

one had been poor quality and this time they were determined to get a better backpack.

They had narrowed the search down to two models, both of which had the basic features

they were looking for, so their decision had come down to determining which pack was

better quality.

For running shoes, participants were told to imagine that they were serious,

experienced runners who regularly participated in road races. They were to imagine that

they were buying shoes for an upcoming race and had narrowed the search to two

models, both of which were made for their type of foot and both of which were equally

comfortable. Their decision had come down to determining which shoe was better

quality.









After the introductory screen, participants received feature information about the

two available brands, one feature at a time. Each feature had a description for brand A

and one for brand B. For example, for backpacks, the first feature described was the

durability of the material, and the two descriptions were the following:

Brand A: Has a layer of leather sewn inside the bottom of the pack to enhance its
integrity. In addition, a layer of water repellent material has been sewn between the
interior layer and the exterior.
Brand B: Double-stitched and double-lined for durability. In addition the material has
been chemically treated to enhance its strength and water repellency while retaining its
suppleness.

Both descriptions appeared on the screen at the same time, one on each side of the

screen (counterbalanced). For picture participants, the corresponding picture of the

backpacks appeared above the feature descriptions. Participants read both descriptions

and evaluated which brand was favored based on the two descriptions using a 9-point

scale where 9 = "description favors Brand A" (the objectively superior but less attractive

brand) and 1 = "description favors Brand B" (the objectively inferior but more attractive

brand). Consistent with Russo's horse race paradigm, immediately after rating the

favorability of a given feature, participants indicated which brand they thought was

leading based on all feature information viewed up to that moment. This was a binary

choice. They then saw the next feature description and completed the same two tasks.

This continued until all five features were presented. After seeing and rating all five

features, participants were asked, "You have now seen all the differentiating features of

the two [backpacks/running shoes]. Considering all the information you have seen, which

[backpack/running shoe] do you think is better quality?" Participants indicated their

response to this binary choice. They were then asked how confident they were that they

had selected the objectively superior brand and indicated their response on a 9-point scale









from "not at all confident" to "extremely confident." After providing their confidence

rating, they then completed the entire task again for the second product category. Finally

they then completed a product knowledge question and answered some general

demographic questions. One hundred thirty-eight undergraduate students participated in

the study in exchange for course credit.

Results

Backpack Ratings

Eighteen participants failed an attention check and were removed, leaving 120 in

the analysis. A significant order effect of replicate required that only the first product

replicate for each participant be analyzed. Table 4-2 presents the mean feature ratings for

backpacks. Higher numbers indicate the belief that the information supports the

objectively superior (less attractive) item. There is a significant effect of the picture, F (1,

41) = 5.64, p < .05. There is also an unsurprising effect of attribute, F(4, 164) = 9.69, p <

.001, and no interaction, F< 1.

After each attribute rating, participants indicated which brand they believed was

leading using a binary choice measure. Table 4-3 presents the proportion of participants

choosing the objectively superior backpack as the leader after the presentation of the

respective attribute. A Chi-square test of independence was conducted on each measure.

The third and fifth measures were significant (p-values < .05), and the fourth measure

was directional, but not significant (p = .10)

On the final choice measure, participants who did not see the picture selected the

superior backpack 78% of the time (18 out of 23), whereas those who saw the pictures

selected the superior (less attractive) backpack 40% (8 out of 20) of the time, p < .05.









Note that this question was essentially identical to the horse race question after the fifth

feature, so the choice proportions are the same as those of the fifth horse race question.

After completing the final backpack choice, participants were asked how confident

they were that they had selected the higher quality backpack. On a 9-point scale, no-

picture participants reported confidence of 6.35 and picture participants reported

confidence of 6.20, indicating no difference in confidence as a function of the presence of

the pictures, t < 1.

Running Shoe Ratings

Table 4-4 presents the mean ratings for the running shoe features. There is a

marginal effect of the picture, F(1, 75) = 3.98, p < .06, but in contrast to the backpack

results, the effect is in the reverse direction. Participants seeing the pictures rated the

feature descriptions as more supportive of the superior shoe than did participants who did

not see the pictures.

After each feature rating, participants indicated which brand they believed was

leading using a binary choice measure. Table 4-5 presents the proportion of participants

choosing the objectively superior shoes as the leader after the presentation of the

respective attribute. Similar to the feature ratings, they are all in the opposite direction.

There is a significant effect in the reverse direction on feature 1, 2 = 5.74, p < .05, and a

directional difference on feature 5, 2 = 2.64, p = .10. For the final choice, people who did

not see the picture selected the superior running shoe 59% of the time (23 out of 39),

while people who saw the pictures selected the superior shoe 76% (29 out of 38) of the

time.

After completing the final choice measure, participants were asked how confident

they were that they had selected the higher quality running shoe. On a 9-point scale, no-









picture participants reported confidence of 6.54 and picture participants reported

confidence of 5.95. This difference was not significant, t(75) = 1.48, p > .14.

Discussion

The bidirectional effect of the pictures across the two products is perplexing. The

backpack results suggest an attractiveness effect, a finding that is consistent with research

in facial attractiveness. The question is why the shoe results went the other way. One

possibility is that the pictures of the shoes differed on some additional dimension besides

attractiveness. In the face perception literature, physical characteristics other than

attractiveness have been found to influence judgments of competence, job suitability, and

job performance (Mueller and Mazur 1996; Todorov et al. 2005). Such characteristics

have been shown to be extremely difficult to ignore (Hassin and Trope 2000). If physical

characteristics other than attractiveness can signal individual performance, it is plausible

that design elements other than aesthetic ones could signal a certain level of product

performance. Perhaps the unattractive shoe appeared superior along one or more

performance dimensions, and participants used this information in their evaluations. If so,

it may be that this appearance of performance dominated the aesthetic difference and

biased the processing of the feature descriptions in the opposite direction. This possibility

is explored in experiment 1.

Another possible explanation for the surprising running shoe results is a problem

with the stimuli. For the backpacks, none of the feature descriptions could be confirmed

with a visual comparison of the brands. However, with the shoes, several of the verbal

attributes mentioned features such as "mesh upper" or "striped rubber" could be

confirmed or disconfirmed by the pictures. It may be that participants did not believe the









written descriptions because the pictures contained conflicting information. In experiment

1 the written descriptions were altered to avoid this problem.









Table 4-1. Pilot Study Stimuli
Feature Brand A (superior)
Backpacks


1. Durability


2. Exterior
material


3. Compartments



4. Carrying
comfort

5. Zipper


Running Shoes
1. Upper
material

2. Weight

3. Midsole








4. Outer Sole


5. Interior
comfort


Brand B (inferior)


Has a layer of leather sewn inside Double-stitched and double-
the bottom of the pack to enhance lined for durability. In addition
its integrity. In addition, a layer of the material has been chemically
water repellent material has been treated to enhance its strength
sewn between the interior layer andand water repellency while
the exterior. retaining its suppleness.
Exterior material is made from Exterior material is a
microfiber and nylon, giving the combination of duralite and
bag both flexibility and durability, nylon for strength with the
ability to stretch.
The main compartment is large Has a large main compartment,
enough to fit books or a laptop plus smaller compartments and
computer. Smaller interior pockets several accessory pockets inside
are organized for keys, pens, etc. to store small items.
Contoured load lifter shoulder Dual-density foam in the
straps ease the weight on the shoulder straps increases
shoulders. carrying comfort.
Has a heavy-weight reinforced Has a plastic zipper, coated with
zipper treated with lubricant for a thin layer of black plastic for
smoother running. In addition a smoother running.
flap of material ties down over the
top of the backpack

Ventilated upper and midsole with Mesh used throughout for
micro mesh. maximum breathability.


Lightweight to optimize running
efficiency.


Lightweight while maximizing
energy transfer properties.


Low-to-the-ground midsole profile, 3D Ultralyte one piece
a stable two-density midsole, and a midsole/outsole material for
flexible, lightweight upper and superior fit. Midfoot Torsion
outsole. Midsole and outsole support device that adds medial
components are engineered to set support and stability.
the foot up in an efficient, balanced
position from heel strike to toe-off,
offering stability and flexibility.
High-density carbon rubber Striped rubber outside for grip
increases traction and extends the and durability.
life of the shoe.
Mesh inner sleeve that stretches to Mesh upper and a lacing system
adjust to your foot over long runs that wraps around the arch for
comfort and support.










Table 4-2. Pilot Study Backpack Feature Ratings
Feature
Exterior Carrying
Durability Exterior Compartments Carrying Zipper
Condition Material Comfort
No Picture 4.65 5.52 6.48 5.78 7.22
Picture 3.70 4.85 4.85 4.10 6.40

Table 4-3. Pilot Study Choice Proportions of Superior Backpack
Choice Opportunities
After First After Second After Third After Fourth After Fifth
Condition Attribute Attribute Attribute** Attribute* Attribute**
No Picture 30% 39% 70% 70% 78%
Picture 30% 40% 35% 45% 40%
*p =.10
**p <.05

Table 4-4. Pilot Study Running Shoe Feature Ratings
Feature
Upper Interior
Upe .Weight Midsole Outer Sole I
Condition Material W t M e O r Se Comfort
No Picture 3.67 4.38 5.62 6.90 5.41
Picture 4.37 4.47 6.82 7.11 6.08

Table 4-5. Pilot Study Choice Proportions of Superior Running Shoe
Choice Opportunities
After First After Second After Third After Fourth After Fifth
Condition Attribute** Attribute Attribute Attribute Attribute*
No Picture 31% 38% 54% 64% 59%
Picture 58% 50% 68% 71% 76%
*p =.10
**p <.05














CHAPTER 5
EXPERIMENT 1

In the pilot study the superior shoe was pretested to be less attractive, but may have

looked like it would perform better, which consequently influenced feature ratings. That

is, rather than an attractiveness bias, there was a performance-based design bias. To

address this possibility, experiment 1 used pictures that were pretested to be equally

attractive, but significantly different on the extent to which they looked like high or low

quality products. In addition, the written descriptions were altered so that a close

inspection of the pictures could not confirm or disconfirm the written feature information.

Method

Stimuli

The experiment was conducted using running shoes only. The written descriptions

of the attributes were altered to contain only information that could not be assessed

visually. Table 5-1 presents the revised feature descriptions. New running shoe pictures

were pretested to create a pair of brands that were equally attractive (56.63 and 61.29 on

a 0-100 scale, p > .73) but appeared significantly different in terms of the extent to which

they looked like superior performing running shoes (relative performance rating of -37.22

on a scale from -100 to +100, where 0 means the shoes are equivalent, p < .002). See

appendix b for pictures of the shoes used in the study.

Design and Procedure

In contrast to the pilot study, experiment 1 had three between subjects conditions.

The no-picture condition from the pilot study was retained as the control group. In









addition there were two levels of the picture condition: one from the pilot study, where

the inferior picture was paired with the superior attributes (hereafter called the inferior

picture condition), and a new condition in which the superior picture was paired with the

superior attributes (hereafter called the superior picture condition). In all other respects

the design and procedure mirrored the pilot study. Ninety-three undergraduate students

participated in the study in exchange for course credit.

Results

Twenty participants were removed for failing an attention check measure. Table 5-

2 presents the mean feature ratings for the remaining 73 participants by condition. As in

the pilot study, there is an effect of the picture, F(2, 70) = 4.04, p < .05 but no picture x

attribute interaction, F< 1. There is a picture bias in the inferior picture condition.

People are biased toward the superior looking shoe when it is paired with inferior

features. There was no bias in the opposite direction when the high performance shoe was

paired with good features. Contrasts performed on the data revealed that the feature

ratings in the no-picture control condition were significantly higher than the inferior

picture condition for features 1, 2 and 4 (one-tailed t-tests, p < .05), and marginally

higher on feature 5 (p < .06). There were no significant differences between the control

condition and the superior picture condition, although feature 2 was marginally

significant in a direction inconsistent with expectations (p < .10).

The proportions of participants who chose the objectively superior shoe as leading

after the presentation of each feature are presented in table 5-3. None of the choice shares

were significantly different from one another, but across all five choice opportunities, the

inferior-picture choice proportion was lower than the no-picture condition. Moreover, in









three of the five choices, the superior-picture choice proportion was higher than the

control condition, suggesting some influence of the pictures in both directions.

As in the pilot study, there was no difference in how confident participants were

that they had selected the superior shoe, F < 1. On a one to nine scale where 1 = "not at

all confident" and 9 = "completely confident," the mean ratings were 6.18, 5.73, 5.52 for

the no-picture, superior-picture, and inferior-picture conditions respectively.

Discussion

Experiment 1 demonstrated a performance-based design bias when controlling for

attractiveness. When the objectively inferior shoe was paired with a picture that

suggested superior performance, participants' ratings were biased in the direction of that

shoe. These results support the contention from the pilot study that the bias in the

direction of the unattractive shoe was the result of its appearing to be functionally

superior.

When the superior-looking shoe was paired with superior features, no bias was

observed in the direction of the superior shoe for the feature ratings. Nonetheless, the

choice shares, while not significant, were directionally supportive of an inclination

toward the superior shoe beyond that of the control condition.

The feature ratings are interesting because there is no justifiable reason to rely on

the picture when rating each feature. However, the use of the Russo paradigm made the

participants' major goal to find the superior shoe overall. Thus, it is possible that the

picture was distorting the ratings in an overall halo-like fashion, particularly when the

features were relatively ambiguous, rather than in a specific feature-by-feature fashion.

Moreover, the pictures were pretested to ensure that one shoe appeared superior overall,

but it is possible that it might have looked superior only on certain dimensions. This lack









of specificity may have influenced results. To focus more narrowly on individual

dimensions, subsequent experiments used products that were pretested to dominate only

on one specific dimension rather than overall.

Predecisional distortion is a form of confirmation bias, where sequentially

presented information is distorted in a desired direction. As acknowledged by Russo and

his colleagues (Russo, Medvec, and Meloy 1998), this type of sequential rating format

can lead to bias even for a control condition. To better isolate a design effect and ensure

that it is not an artifact of the experimental format, experiment 2 moves away from the

Russo paradigm and focuses on a single feature rating.

When considered in conjunction with the running shoe results from the pilot study,

the results of the current experiment call into question the backpack results observed in

the previous experiment. If the design effect for running shoes was due to the appearance

of performance, could it have been that the backpack results were also due to a

performance-based design bias rather than an attractiveness bias? The backpack protests

did not control for the appearance of performance, so it may have been that the more

attractive backpack was also the one that appeared superior on performance. To examine

this possibility, a follow-up test was conducted on the backpack stimuli measuring the

correlation between perceived performance and perceived attractiveness. There was a

marginally significant positive correlation between attractiveness and performance, r =

.24, p = .10; thus, it is impossible to conclude that the results for the backpacks in the

pilot study were due entirely to attractiveness. Because there is no strong evidence of

biased processing due to aesthetics, the next two experiments focus only on the

appearance of performance. However, the issue of aesthetic influence remains









unresolved, and in experiment 4 the potential for an attractiveness bias is reexamined in a

different context.


Table 5-1. Running Shoe Feature Descriptions Used in Experiment 1


Feature
1. Weight

2. Midsole
material


3. Midsole
Design



4. Outer Sole


5. Interior
comfort


Brand A (superior)
Lightweight to optimize running
efficiency.
Ethylene Vinyl Acetate air infused
foam midsole provides superior
cushioning without sacrificing
durability.
Midsole components are designed
to keep the foot in the most
efficient position, maintaining
balance but allowing it to flex
naturally.
Durable high-density carbon
rubber outsole extends the life of
the shoe.
Double strike insole is made with
orthotic-type material that adjusts
to the foot shape over time to
maximize shock absorption.


Brand B (inferior)
Lightweight while maximizing
energy transfer properties.
Midsole is made of high quality
polyurethane foam, which is
dense and stable.

Midsole components are
compression molded, with solid
construction that ensures that the
shoe will maintain its structure
and shape over time.
Blown rubber traction enhancing
outsole increases grip on softer
surfaces.
Well cushioned insole provides
good arch support.


Table 5-2. Experiment 1 Running Shoe Feature Ratings
Feature
Upper Weight Midsole Outer Sole Interior
Condition Material Comfort
No Picture 5.68 7.68 6.18 5.45 7.73
Superior Picture 5.23 6.88a 5.85 5.08 7.38
Inferior Picture 4.40a 6.84a 5.92 4.24a 6.96
aCell mean differs from control group mean atp < .05 significance.

Table 5-3. Experiment 1 Choice Proportions of Superior Running Shoe
Choice Opportunities
After First After Second After Third After Fourth After Fifth
Condition Attribute Attribute* Attribute Attribute Attribute
No Picture 45% 82% 73% 59% 91%
Superior Pic 58% 92% 88% 54% 85%
Inferior Pic 36% 68% 82% 56% 80%
*p< .10















CHAPTER 6
EXPERIMENT 2

Experiment 2 marks a small shift in the format of the studies, with four key

changes. First, based on the findings of experiment 1, experiment 2 focused solely on a

visual difference in performance, rather than a visual difference in attractiveness.

Second, whereas the pilot study and experiment 1 made use of a sequential

presentation format, with feature information presented one at a time, in experiment 2 all

product information was presented at once in paragraph format. Participants saw two

paragraphs on the screen, one describing each brand. Each paragraph provided

information on several product features. In addition, for participants in the picture

condition, the written information was presented simultaneously with the picture. These

changes eliminated the possibility of distortion occurring over time in a confirmation bias

manner.

Third, to make the visual difference in performance as clear and as specific as

possible, the study focused on only one dimension per product. Pictures were chosen to

reflect superiority along one given dimension, rather than overall quality, and participants

were only asked to rate the product on that single dimension. The corresponding

paragraphs suggested that the inferior-looking brand was in fact superior on the given

dimension.

Finally, to broaden generalizability, and to avoid some of the difficulties of the

stimuli from the pilot study, experiment 2 and all subsequent experiments used five

different product categories: cookware, stereo speakers, in-line skates, bicycle helmets,









and electric mixers. These categories were selected because they were products for which

appearance could potentially provide information about performance, but could not

unequivocally signal high or low quality.

Method

Stimuli

Pretests were conducted to select pictures and feature information for the new set of

products. The goal for each pair of pictures was to have one product that appeared

superior along a specific dimension, yet appeared equally or less attractive than the other

product. Several pictures for each of the five product categories were selected for

protesting. Twenty-four participants rated pairs of pictures from the five categories on

relative attractiveness and superiority on the key performance dimension. The question

order was counterbalanced across participants. The measures were seven point scales

from "Brand A is far superior" to "Brand B is far superior." One-sample t-tests were

conducted on the attractiveness and dimension ratings. Products chosen were those for

which the attractiveness rating was not significantly different from the mid-point rating

("pictures are equally attractive") but the performance dimension was significantly

different from the mid-point. The pretest led to the selection of one pair of pictures for

each product category. The final picture pairs selected appear in appendix c, and the

ratings for attractiveness and performance along the focal dimension, mean centered for

ease of comparison, appear in table 6-1.

Paragraph descriptions of product features were written in the format of an expert

review, supposedly by Consumer Reports. The written information was pretested so that

on the focal dimension, one review was significantly better than the other. The reviews

were evaluated in pairs on a seven-point scale where 1 = "Brand A is far superior" and 7









= "Brand B is far superior." In addition care was taken to ensure that the focal dimension

could not be confirmed by a visual inspection of the pictures. The product categories, key

dimension, and pretest ratings, mean centered, appear in table 6-2. The paragraph

descriptions used in the study appear in appendix d.

Design and Procedure

Experiment 2 had two between subjects conditions (picture and no-picture) and five

product replicates. For picture participants, the written descriptions that indicated the

product was superior along the key dimension were matched with the picture that

suggested the product was inferior on the dimension.

The experiment was conducted on computer. Participants saw an introductory

screen that informed them they would be reading Consumer Reports reviews for pairs of

brands in five different product categories and would be asked to evaluate the quality of

the products based on the reviews. After this introductory screen, they were presented

with the first product category, in a randomly selected order, and saw two paragraphs side

by side on the screen (counterbalanced), one describing each brand. Participants in the

picture condition also saw the two pictures above the corresponding reviews. After the

reviews had been on the screen for 40 seconds, participants were asked the degree to

which one product or the other was superior along the key dimension. This was indicated

on a 9-point scale from 1 (Brand A is far superior on this dimension) to 9 (Brand B is far

superior on this dimension). A rating of 9 would mean that the objectively superior brand,

based on the written information, was superior. To avoid any evaluative inferences

related to the scale points the numbers on the scale were blacked out so that participants

only saw nine radio buttons ranging from "Brand A is far superior" to "Brand B is far

superior." The picture and written information did not disappear until after subjects









provided their rating. The next product category then appeared on the screen and the

process was repeated. After providing ratings for the five product categories, participants

were given an unrelated filler task that lasted approximately ten minutes. Upon

completion of the filler task a manipulation check was conducted. Participants were

shown the picture pairs again. They were asked to imagine they had not read anything

about the products previously and, based solely on a visual inspection of the products, to

indicate which looked superior on attractiveness and on the focal dimension. Of course,

despite the filler task, a consistency bias was a serious concern for the picture

participants, so the usefulness of the manipulation check was somewhat limited. After the

manipulation check, participants were asked questions regarding their knowledge of the

products, as well as some demographic information. They were then thanked and

dismissed. Eighty-three undergraduate students participated in the study in exchange for

course credit.

Results

The mean ratings for the five products appear in table 6-3. Higher ratings indicate

that the objectively superior brand according to the reviews was rated as superior. As the

table indicates, both groups were relatively accurate; both were able to differentiate and

tell which review was superior. Nonetheless, the significant effect of picture, F(1, 81) =

4.28, p < .05, indicates that the picture did have an effect. Across all replicates, the mean

rating for the picture group was lower than the mean rating for the no-picture group.

The manipulation checks confirmed that the stimuli satisfied the criteria for the

study. A rating of 0 would indicate that the two pictures appeared equal. A negative value

would indicate that the picture pretested to be superior along the focal dimension was

perceived as superior, and a positive value would indicate that the picture was perceived









as inferior. Thus, a successful manipulation would have attractiveness ratings equal to or

greater than 0 and performance ratings significantly less than 0. Table 6-4 presents the

mean ratings for the manipulation check. All five picture pairs had significant differences

in the desired direction for performance and had either no differences or differences in

the reverse direction for attractiveness.

Because participants in the picture condition might have been prone to consistency

bias in their picture ratings, the manipulation check was performed a second time with

only no-picture participants. The results mirrored those that included all participants.

Discussion

The results of experiment 2 suggest that people are using the pictures in their

evaluation of the products. However, the reason for this is still unclear. It may be that

participants had an initial reaction to the pictures that distorted subsequent evaluation of

product dimensions. This is consistent with the results of pilot study and experiment 1. In

experiment 2, all visual and verbal information was presented simultaneously, which

could minimize the effects of an initial reaction to the visual stimuli. However, a visual

reaction to design is usually rapid and relatively automatic (Page and Herr 2002), and it is

highly unlikely that participants read the paragraphs before looking at the pictures. Thus,

the notion that the effect is a function of an initial reaction influencing subsequent

evaluation in a halo-like manner is still a real possibility. If the result were due to a halo

effect, then the reaction to the visual information should carry over to other dimensions

on which the picture does not dominate.

Another possibility is that the pictures are being relied upon because they appear

diagnostic of quality. If this is the case, then if participants were asked about dimensions

on which the superior picture does not dominate, there should be no effect of the picture.









Experiment 3 addresses these two possibilities by replicating the format of experiment 2

and adding questions about two non-focal dimensions.

Table 6-1. Pretest Results for Performance Pictures


Category
Cookware
Stereo speakers
In-line skates
Bicycle helmets
Electric mixers


Attractiveness
Rating
-0.37
0.54
-0.33
0.50
0.13


Key
Dimension
Ease of cleaning
Power
Speed
Comfort
Power


Performance
Rating
1.08*
1.08**
1.13**
1.08***
2.00***


*p<.05
**p <.01
***p <.001

Table 6-2. Pretest Results for Feature Descriptions
Category Key Dimension Dimension Rating
Cookware Ease of cleaning -2.08**
Stereo speakers Power -1.71*
In-line skates Speed -2.62**
Bicycle helmets Comfort -2.87**
Electric mixers Power -2.71**
*p <.07
**p <.01

Table 6-3. Results of Experiment 2
Picture No Picture
Category N=43 N=40
Cookware 7.30 7.58
Stereo speakers 7.30 7.45
In-line skates 6.60 7.83
Bicycle helmets 6.86 7.55
Electric mixers 7.93 8.18
Average 7.20 7.72






47


Table 6-4. Manipulation Check for Experiment 2
Attractiveness Performance
Category Rating Rating
Cookware 0.07 -1.04**
Stereo speakers 1.46* -0.83*
In-line skates 0.99* -1.80***
Bicycle helmets -0.60 -1.02***
Electric mixers 0.61 -2.49***
*p<.05
**p<.01
***p <.001














CHAPTER 7
EXPERIMENT 3

Experiments 1 and 2 demonstrated a performance-based design bias, but the

nature of the effect was unclear. Experiment 1 suggested that people had a general

reaction to the picture that colored their evaluation of features. Experiment 2 focused on

one dimension only and confirmed the existence of a design bias, but did not help to

clarify whether it was due to a halo effect of the picture or a result of the picture

appearing diagnostic of quality on the focal dimension. In experiment 3 participants again

saw all the product information on the screen at one time and again evaluated

dimensional performance. However, in this case, participants evaluated the products

along three dimensions, the focal dimension on which the products differed visually and

two additional dimensions on which the products did not differ visually.

Method

Stimuli

The product pictures from experiment 2 were used in experiment 3. Two non-focal

dimensions were added for each product, and feature descriptions were altered slightly to

provide some information about the non-focal dimensions. The additional dimensions

appear in table 7-1, and the revised paragraphs appear in appendix e.

Design and Procedure

The design was similar to experiment 2. There were two between subjects

conditions, picture and no picture. There were five product replicates, but in this study









each replicate was accompanied by three dimension questions, one about the focal

dimension and two about non-focal dimensions.

Participants were given the same initial instructions as in experiment 2. They were

then presented with each product replicate, in random order. For each replicate,

paragraph descriptions of the two brands appeared on the screen, position

counterbalanced. For picture participants, a picture of each brand accompanied the

descriptions. After 40 seconds the first of the three dimension questions appeared. The

order of presentation of the three questions was determined using a Latin Square. Once

participants responded to the first question, it disappeared and was replaced by the second

and then the third question. As in experiment 2, the pictures and descriptions remained on

the screen until all three questions were answered, at which time they disappeared and

were replaced by the next product replicate. After answering the dependent measures for

all five replicates, participants completed a filler task and responded to the manipulation

check questions. Finally, they answered product knowledge and demographic questions,

and were dismissed. Eighty-nine undergraduate participants completed the study.

Results

Eight participants failed an attention check measure and were removed, leaving 81

in the analysis. Mean ratings for the focal and non-focal dimensions appear in table 7-2.

The dimension x picture interaction was not significant but was directionally supportive,

F(2, 158) = 2.19,p < .12. However, a manipulation check on the focal dimensions

revealed that for this group of participants, bicycle helmets did not constitute a successful

manipulation of perceived performance on the key dimension. In contrast to the other

four replicates, the two helmets did not differ significantly in the extent to which they









appeared comfortable for this group of participants. The manipulation check ratings,

mean centered, appear in table 7-3.

Dropping helmets from the analysis and rerunning it on the remaining four

replicates resulted in a significant interaction of dimension and picture, F(2, 158) = 3.16,

p < .05. For completion there was also a significant effect of replicate, F(2, 158) =

166.39, p < .001 and no effect of picture, F(1, 79) = 2.47,p > .05. The overall means with

helmets removed appear in table 7-4. Follow up contrasts on the revised means revealed

that the difference between the picture and no-picture conditions was significant for the

focal dimension (p < .05) but not for either of the non-focal dimensions.

Discussion

Experiment 3 was designed to determine whether the picture effect observed in

previous studies was the result of a general halo or the result of the picture being

perceived as diagnostic. The results of experiment 3 indicate the latter. The pictures

provided information about the focal dimension but not about the other dimensions, and

the picture only had an effect on the ratings of the focal dimension. For the focal

dimensions, the presence of the picture lowered the ratings of the strength of the verbal

information, but for non-focal dimensions, the presence of the picture had no effect. This

suggests that participants used the pictures as a source of information only when they

perceived them to be diagnostic. The helmet replicate did not conform to this pattern, but

it also did not pass the manipulation check for this group of subjects.

Thus far, all the studies have compared information processing in the presence of

visual information to processing in the absence of visual information. However, with

such designs the pictures are an additional source of information, so assuming an

averaging model it is not entirely surprising that there was an effect of the picture. A









stronger test of the use of visual information would be to compare conditions that have

the same information but vary the order in which the information is presented.

Experiments 4 and 5 involved such tests.

In addition, although it appeared that the pictures were not operating as halos and

coloring evaluation of all features, it may still have been that their influence was the

result of an initial reaction to the key dimension. Since the pictures were pretested to

differ significantly along one dimension, a relatively automatic reaction to this difference

could have influenced the ratings of that dimension. Manipulating the order of

presentation of the visual and verbal information in experiments 4 and 5 enabled a test of

this possibility.

Table 7-1. Dimensions Used in Experiment 3
Focal Non-focal Non-focal
Product dimension dimension dimension
Cookware Ease of cleaning Fit of lids Heat distribution
Stereo speakers Power Sound clarity Sound distribution
In-line skates Speed Comfort Ease of use
Bicycle helmets Comfort Protection Aerodynamics
Electric mixers Power Loudness Ease of use

Table 7-2. Results of Experiment 3
Focal
Product Condition Dimension Dimension 2 Dimension 3
Cookware Picture 7.58 3.79 4.23
No Picture 8.00 3.92 4.47
Stereo speakers Picture 7.05 5.16 4.88
No Picture 7.53 4.45 4.68
In-line skates Picture 6.63 3.19 4.40
No Picture 7.55 3.45 5.58
Bicycle helmets Picture 7.02 6.16 6.33
No Picture 6.26 5.74 5.03
Electric mixers Picture 6.65 4.47 6.07
No Picture 7.71 3.95 5.58
Average Picture 6.99 4.55 5.18
No Picture 7.41 4.30 5.07









Table 7-3. Manipulation


Category
Cookware
Stereo speakers
In-line skates
Bicycle helmets
Electric mixers
*p<.05
**p <.001


Check for Experiment 3
Attractiveness Performance
Rating Rating
0.51 -0.73*
1.25** -1.10**
0.12 -0.68*
0.38 -0.42
0.31 -2.41**


Table 7-4. Results of Experiment 3 with Helmet Replicate Removed
Focal
Product Condition Dimension Dimension 2 Dimension 3
Average Picture 6.98 4.15 4.90
No Picture 7.70 3.94 5.08














CHAPTER 8
EXPERIMENT 4

In previous studies feature evaluation without a picture was compared to feature

evaluation with a picture, and a significant effect of the presence of the picture was

observed. Certainly this suggests an influence of product design but assuming an

averaging model, the use of the picture as additional information does not provide insight

into the role the picture played. Thus, in experiment 4, the key comparisons occur

between groups that saw both pictures and written information.

Experiment 3 suggested that the pictures had influence because they were

perceived to be diagnostic. The picture indicated superiority along the focal dimension,

so even though participants were encouraged to focus on the Consumer Reports reviews,

the design may have provided what was perceived to be useful information. This suggests

a relatively deliberate use of the picture as a source of information. Another possibility is

that people had an immediate reaction to the dimensional information in the picture,

which distorted evaluation of the corresponding feature information.

To address these possibilities, participants in experiment 4 received both pictures

and written information, but the order of presentation of the visual and verbal information

was varied. Half the participants saw the pictures first and then the reviews. The other

half saw the reviews first, and then the pictures. If the results were due to an initial

reaction to the information provided by the picture and an insufficient adjustment upon

seeing the verbal information, then the picture should have an effect only when it comes









first. If the results of the previous studies were due to diagnosticity, as suggested by

experiment 3, there should be no difference as a function of the order of presentation.

Another issue that is addressed in experiment 4 is one that arose in the pilot study.

The pilot study was originally designed to examine attractiveness effects in a product

context, but the results mapped more closely onto a performance-based design bias than

an attractiveness bias. However, research in facial attractiveness does report a biasing

effect of aesthetics. Most consumer research on product design has focused on aesthetics

and has demonstrated a variety of effects on liking ratings or product evaluations in the

absence of reliable information. In the current series of studies, the key measure has been

a quality rating of particular features, and prior aesthetics research in consumer behavior

is mute on whether attractiveness might bias such evaluation. Thus, to address this

possibility and to confirm the results from the pilot study, an additional factor was

included in experiment 4. Half the experimental participants saw the same pictures as

those used in experiment 3; that is, pictures in which the products looked equally

attractive but one looked superior on the key dimension. For the rest of the experimental

participants, pictures were used where the opposite was true; that is, the products looked

equal on performance but one was superior on attractiveness. Such a design has inherent

difficulties because the pictures are not uniform across conditions. However, the use of

five product replicates helped to reduce the magnitude of this design issue.

Method

Stimuli

The Consumer Reports reviews used in this study were the same as those used in

experiment 2. Similarly, the performance pictures were also the same. The only change

was the addition of pairs of pictures for each of the five categories in which the pairs had









equal ratings on performance along the key dimension, but significantly different ratings

in attractiveness. The pretest results for these additional pictures are presented in table 8-

1. For ease of comparison, the table also includes the pretest results from the performance

pictures originally presented in experiment 2. The pictures of the new stimuli appear in

appendix e.

Design and Procedure

The study was a 2 (order of presentation: picture first versus review first) x 2 (type

of pictures: performance difference versus aesthetic difference) between subjects design.

In addition there was a review-only condition as a control group. The experiment was run

on computer, and all instructions were given on screen.

As in experiments 2 and 3 participants were told they would be reading Consumer

Reports reviews for pairs of brands in five product categories and that they would be

asked to assess the quality of the products based on the reviews. After the introductory

screen the first of the five categories was displayed in a randomly determined order. The

presentation of each product category proceeded as follows. Participants in the picture-

first conditions saw pictures of the two brands. Participants looked at the pictures for 20

seconds before proceeding. The pictures then disappeared from the screen and were

replaced by the written reviews. Participants could not proceed further until 40 seconds

had passed. Participants in the review-first conditions saw the two reviews first. After 40

seconds participants could proceed from the written reviews to the pictures, which

remained on the screen for 20 seconds. Participants in the baseline control condition saw

only the reviews. After the presentation of the pictures and descriptions, participants

proceeded to a new screen and were asked to indicate the degree to which one or the

other brand was superior along the key dimension. This procedure was then repeated for









each of the five product categories. The rest of the procedure, including the filler task,

manipulation checks, product knowledge and demographic questions, mirrored

experiments 2 and 3. One hundred eighty-seven undergraduate students participated in

the experiment for class credit.

Results

A computer error caused 19 participants to be unable to complete the study. An

additional 23 participants failed an attention check and were removed, leaving 145 in the

analysis. Table 8-2 presents the overall results for the five product replicates.

A 2 (order) x 2 (picture type) x 5 (replicate) mixed ANCOVA using knowledge of

the products as covariates was performed on the four experimental conditions (the

baseline condition was not included in this analysis). The analysis revealed no three way

interaction (F< 1), a marginal interaction of picture type and replicate, F(4, 448) = 2.39,

p < .06, no interaction of order and replicate, F(4, 448) = 1.12, p > .34, no effect of

picture type, F(1, 112) = 1.34, p > .25, and no effect of order, F < 1. Importantly,

however, there was a significant interaction of order and picture type, F(1, 112) = 4.47,p

<.05. As the table indicates the pictures had little effect when they differed on aesthetics,

regardless of whether the picture came before or after the written information. When the

pictures differed on performance, they had an effect, but only when they followed the

reviews.

Pairwise comparisons on the overall means confirmed that there was a significant

difference between the picture-first and review-first conditions for the performance

pictures, (p < .05). The difference between the aesthetic and performance conditions

when the review preceded the pictures was marginally significant, (p < .07), as was the









difference between the review-first performance group and the baseline control condition,

(p < .08). All other comparisons were non-significant.

The manipulation check results appear in table 8-3. For the performance pictures,

the goal was to have mean-centered ratings that were significantly lower than 0 on

performance (indicating superiority for the superior picture), and equal to or higher than 0

on attractiveness. For the aesthetic pictures the goal was the opposite. As indicated in

table 8-3, all products met this goal for this group of subjects except in-line skates. The

removal of this replicate did not materially change the results. An ANCOVA on the

remaining replicates again revealed the critical significant interaction of order and picture

type, F(1, 113) = 6.45, p < .05. For completion there was no three way interaction (F <

1), a significant interaction of picture type and replicate, F(3, 339) = 3.53,p < .05, no

interaction of order and replicate, F(3, 339) = 1.51, p > .20, no effect of picture type, F(1,

113) = 1.23,p > .20, and no effect of order, F < 1. The revised overall means with the in-

line skate replicate removed were slightly higher than the original means, but maintained

the same directional patterns. For the picture-first conditions, mean ratings were 7.56 for

performance pictures and 7.21 for aesthetic pictures. For the review-first conditions,

mean ratings were 6.77 for performance pictures and 7.53 for aesthetic pictures. The

review-only control condition had a mean rating of 7.23. Follow-up contrasts again

revealed a significant difference between the two performance picture conditions (p <

.05) and a marginal difference between the review-first conditions, (p < .07). The only

notable change was that the difference between the review-first performance condition

and the baseline review-only group became non-significant after the removal of the in-

line skate replicate.









Discussion

What is immediately notable from experiment 4 is that there was little effect of

aesthetics. The pilot study suggested an aesthetic bias for backpacks. Although the

running shoes demonstrated a performance-based design bias, it was presumed that both

effects could have existed but the aesthetic bias may have been dominated by the

performance bias. However, the results of experiment 4 suggest that aesthetics may not

influence quality evaluations. Of course, one could argue that this could have been a

calibration issue; indeed, the mean ratings for speakers and to a lesser extent helmets, did

suggest an attractiveness effect when the pictures preceded the reviews. A comparison of

the pretest and manipulation check values for attractiveness versus performance renders

this possibility somewhat less likely. As shown in table 8-1, the pretest ratings for the

pictures tended to have similar sized differences for the performance and aesthetic

pictures. The manipulation check ratings provide a similar picture. As indicated in table

8-3, for this group of participants, the difference ratings for the attractiveness pairs were

generally much larger than for the performance pairs. Of course the scales may have been

quite different, but it at least suggests that the result was not solely due to calibration.

Prior product design research would suggest that aesthetics might have an influence

on information processing, although most such research has focused on aesthetics biasing

preference or evaluations in the absence of other information (Page and Herr 2002). The

expectation of an aesthetic influence is that people would have an automatic affective

reaction that would color further processing. It is possible that at least in some cases

aesthetics did have an influence when the picture came first, but overall at least, no strong

aesthetics bias emerged overall. The possibility of an aesthetic bias cannot be ruled out

here, and further studies would be needed to address this issue. However, based on the









results of the current study, the focus of the dissertation will remain on the performance-

based design bias.

An additional surprising result was that for the performance pictures, the effect of

design was observed when the picture followed the review but not when it preceded the

review. Thus the notion that the influence of design is because people have an immediate

reaction which biases subsequent evaluation appears incorrect.

The results for the performance conditions in experiment 4 seem consistent with a

recency effect. In the experiment participants knew they would be making some kind of

quality rating, but they did not know ahead of time what dimension they would be

evaluating. Thus, upon seeing the question, which appeared after the product information

had been removed from the screen, they had to rely on their memories to respond. The

most recently viewed information was probably the easiest to recall, and so it was used,

leading to the divergent results. However, recency cannot entirely account for all

experimental conditions. The accuracy of the review-first/aesthetics condition

demonstrates that people were able to recall the earlier written information. For these

participants, the pictures were the most recently viewed pieces of information and were

certainly accessible, but they were not diagnostic, so it was necessary to delve further into

memory to retrieve task-relevant information. For participants in the performance/review-

first condition, the pictures were also highly accessible in memory. Unlike the aesthetic

picture groups, however, for these participants the pictures were diagnostic, so people

could use them in their judgment. Thus, rather than being merely a recency effect, it

appears the picture results are more consistent with a combination of accessibility and

diagnosticity (see Feldman and Lynch 1988). If this is the case, then alerting participants









to the judgment task prior to presenting the product information might alter the influence

of the picture and written information. This possibility is explored in experiment 5.


Table 8-1. Pretest Picture Ratings
Picture Type Performance Difference
Attractiveness Performance
Category Rating Rating
Cookware -0.37 1.08*
Stereo speakers 0.54 1.08**
In-line skates -0.33 1.13**
Bicycle helmets 0.50 1.08***
Electric mixers 0.13 2.00***
*p<.05
**p <.01
***p <.001


Table 8-2. Results of Experiment 4
Performance Pictures
Picture Review
First First
Cookware 7.43 6.45
Stereo speakers 7.86 7.61
In-line skates 6.09 6.06
Bicycle helmets 6.89 6.26
Electric mixers 8.06 6.77
Average 7.26 6.63


Table 8-3. Manipulation Check for Experiment 4
Picture Type Performance Difference
Attractiveness Performance
Category Rating Rating
Cookware 1.01*** -1.19***
Stereo speakers 1.68*** -0.56*
In-line skates 0.74** -1.18***
Bicycle helmets -0.19 -0.94***
Electric mixers 0.68** -2.12***
*p<.05
**p <.01
***p <.001


Aesthetic Difference
Attractiveness Performance
Rating Rating
-1.79*** 0.37
-1.42** 0.46
-1.04* 0.33
-1.04* 0.50
-1.04* 0.13


Aesthetic Pictures
Picture Review
First First
7.52 7.17
6.87 7.87
6.52 6.46
6.16 6.50
8.29 8.58
7.07 7.32


Control
Review
Only
6.71
7.33
7.08
6.54
8.33
7.20


Aesthetic Difference
Attractiveness Performance
Rating Rating
-2.03*** 1.17***
-1.21*** 1.18***
0.21 0.91***
-2.39*** 0.02
-2.50*** -0.14














CHAPTER 9
EXPERIMENT 5

Experiment 5 focused solely on the performance-design bias. The experiment

attempted to address the order effect that was observed in the performance-picture

conditions in experiment 4. In experiment 4 participants in the picture-first condition did

not show a design bias, but those in the review-first condition did. However, participants

in neither group knew ahead of time what dimension they would be evaluating. All they

knew was that they would be evaluating some element of quality. Thus it is plausible that

participants in the picture-first group did not focus on the key dimension, but rather on

some overall quality judgment. The reviews then provided information on the key

dimension, which was subsequently used to make the judgment. For participants in the

review-first condition, the review focused participants on the key dimension, which may

have then alerted them to the conflicting information provided by the picture. Thus, when

it came to making a judgment, they were influenced by the picture.

In experiment 5 participants were told what dimension they would be evaluating

prior to viewing any information about the product. The final question still occurred after

both pieces of information were removed from the screen, but by providing participants

with the dimension ahead of time, it was more likely that both the picture and written

information were equally accessible at the time of judgment.









Method

Stimuli

The same set of pictures used in the performance conditions of experiment 4 were

used in experiment 5. The same product information was also retained.

Design and Procedure

Experiment 5 consisted of one between-subjects factor of order, with three levels:

picture first, review first, and a review-only control group. Again there were five product

replicates.

The procedure was similar to that of experiment 4. The only difference was that

participants were told what dimension they would be evaluating prior to seeing the

pictures or the written reviews of the products. Ninety-one undergraduate students

participated in the experiment for partial class credit.

Results

Eleven participants were removed for failing an attention check, leaving 80 in the

analysis. The mean ratings appear in table 9-1. There was a significant effect of replicate,

F(4, 308) = 12.50,p < .01 and no interaction of order and replicate, F(8, 308) = 1.45, p >

.15. There was a significant effect of order, F(2, 77) = 4.83, p < .05, but the direction of

the order effect was reversed from that observed in experiment 4. In the current study,

when people were instructed ahead of time which dimension to focus on, the picture had

a stronger effect when it preceded the review. Follow up comparisons revealed a

significant difference between the picture-first and review-first conditions, t(77) = 2.55, p

< .05 and between the picture-first and review-only conditions, t(77) = 2.78,p < .05, but

no difference between the review-first and review-only conditions, t < 1.









The manipulation checks revealed that for this group of participants, both the stereo

speakers and the bicycle helmets were unsuccessful manipulations of performance

design. Removing these two replicates from the analysis resulted in revised overall means

of 7.07, 7.55, and 7.98 for the picture-first, review-first, and review-only conditions

respectively. The effect of order was still significant, F(2, 77) = 3.48, p < .05, but the

pairwise comparisons resulted in only a significant difference between the picture-first

and review-only conditions, p < .05.

Discussion

By providing participants with the focal dimension before the information,

experiment 5 enabled a test of the effect of order that avoided the problem of recency. In

this study the appearance of the question after the presentation of the picture was no

surprise, so participants did not have to search their memories for relevant judgment

information.

The key result from experiment 5 is that the picture had the greatest influence when

it preceded the written review, a reversal of the order effect observed in experiment 4. In

this study, participants knew what feature they would be evaluating. When the picture

appeared, it seemed diagnostic of quality and therefore was used in arriving at a judgment

of the feature. The subsequent written review was of course also diagnostic. However,

when it came time to supply the evaluation, the judgment made based on the picture was

probably still highly accessible, and since it was also perceived to be diagnostic it

influenced the rating.

When the review preceded the picture, participants were easily able to assess the

quality of the two products along the dimension of interest before seeing the picture.

Thus, even though the picture provided conflicting information, its influence was weaker,









perhaps because the judgment made based on the written reviews was very clear. The

picture did still appear to have some influence, however. An examination of the overall

means after the removal of the stereo speakers and bicycle helmets does suggest an

influence of the picture when it followed the review, albeit a weaker one than was

observed in the previous study.

Experiments 3 and 4 suggested that the picture effect was due to diagnosticity.

However, in the current study, a diagnosticity result should have led to equal ratings for

the picture-first and review-first conditions. The review-first condition showed some

effect of the pictures, particularly for the in-line skates and electric mixers, but overall the

results were not supportive. If the current results are reliable, the design effect may not be

solely due to diagnosticity, but rather may be the result of an immediate reaction to the

picture biasing subsequent evaluation, but only when attention is drawn to dimensions

where the picture dominates. Additional data are needed to confirm the results of the

review-first condition.

Table 9-1. Results of Experiment 5
Picture Review Review
First First Only
Cookware 6.39 7.65 7.50
Stereo speakers 6.54 7.96 7.19
In-line skates 7.14 7.19 7.77
Bicycle helmets 7.04 7.65 7.42
Electric mixers 7.68 7.81 8.69
Average 6.95 7.65 7.71














CHAPTER 10
GENERAL DISCUSSION

The goal of this paper was to examine some of the ways product design might

influence judgment. A pilot study and five experiments pitted product appearance against

conflicting written information to address whether design could influence a person's

tendency or ability to objectively evaluate objective product information.

Consistent with prior research in psychology as well as consumer research on

product design, the initial pilot study focused on attractiveness as a source of bias. An

attractiveness effect obtained for one of the two product replicates, but attractiveness was

confounded with performance for that stimulus. For the other stimulus, no attractiveness

effect was observed. Rather the results suggested a bias in the direction of the

unattractive brand, which subsequent analysis revealed looked like a superior performer.

This discrepancy was resolved in experiments 1 and 2, which confirmed that the

appearance of performance biased the processing of other information.

Experiments 3 through 5 attempted to assess the source of the performance-design

bias. Experiment 3 investigated whether design operated as a type of halo that influenced

ratings of other attributes. Results indicated that design only influenced ratings for

features on which it appeared to dominate, suggesting the effect was not due to a halo

effect. Experiments 4 and 5 examined whether the bias was a function of an initial

reaction to the visual information coloring subsequent processing of written information.

Experiment 4 did not support such an explanation, and instead appeared more consistent

with a accessibility-diagnosticity explanation. In that experiment, participants were not









focused on the key dimension prior to viewing the information. When asked the

dependent measure, participants used the most accessible information that was diagnostic

for the task. This tended to be the most recently viewed information. Experiment 5 was

less conclusive, but still hinted at a diagnosticity result. When participants were focused

on the key dimension ahead of time, the picture had an influence when it appeared prior

to the written information because the task was known at the time and the pictures were

diagnostic. When the picture followed the review, however, less influence of the picture

was observed. This result is inconsistent with a diagnosticity explanation and more in line

with an early reaction distorting subsequent evaluation. However, stimuli calibration

concerns may have been partially responsible.

Overall the results of the experiments suggest that for products, when it comes to

being evaluated for quality, seeing is believing. When a product appears to offer superior

performance along a given dimension, people believe that it does, and reading

information that suggests otherwise may be insufficient to counteract that belief.

Whether there is an aesthetics bias is unclear. Experiment 4 addressed the issue of

attractiveness as a source of bias initially investigated in the pilot study, and found no

effect of aesthetics. However, the study did not ask for choice or preference judgments.

Such judgments, which are inherently more personal, may be more susceptible to

aesthetic effects. Certainly if one considers the research in facial attractiveness and the

dramatic effects that have been reported (e.g., Landy and Sigall 1974), it is likely that an

aesthetic bias in products would occur under some conditions. The results of some of the

replicates in experiment 4 hint that there may be aesthetics bias; further research is









needed to confirm this and to understand the conditions under which an aesthetic design

bias might obtain.

Implications

Product design and styling are often among the most important determinants of

new product success (Cooper and Kleinschmidt 1987; Vriens et al. 1998). A good design

can attract attention, communicate information about a product, and influence preference.

Thus, understanding how consumers process and react to product design, particularly

with regard to new product designs used in market testing, is of great importance. For

managers, the key takeaway from this research may be that consumers believe what they

see. If the goal of a new product is to communicate some level of performance, it is

critical that this is communicated visually. Otherwise marketing efforts aimed at

convincing consumers that the product excels on a given dimension may fall flat. In

terms of design team goals, the research suggests that a company may be better off

expending efforts on making a product look like it works well rather than simply making

it look good.

From a consumer perspective, of particular importance is a better understanding of

the role that product appearance plays in decision making. The experiments demonstrate

that consumers are able to discount attractiveness in the evaluation of product quality,

and yet seem unable to ignore appearance when design suggests performance. It may be

that this is a deliberate choice and that consumers consciously discount the written

information when their eyes tell them otherwise. However, if consumers are convinced

by what they see rather than other information they encounter and are unable to ignore

visual information when it should be ignored, there is potential for non-normative

decision making. As an extreme example, products that appear to dominate on safety,









such as life jackets, flashlights, or cars, but actually do not provide the level of

performance that their appearance suggests, could result in seriously detrimental

consequences for the consumer.

Limitations

The research reported here focused on a fairly small number of products that were

chosen because they were products where design could be important but was not critical

to the function of the product. Because of these narrow selection criteria, it is unclear

whether the results observed in the dissertation would extend to a wider array of product

categories. Moreover, in several experiments not all the picture manipulations were

successful. It will be important to improve the stimuli calibration and replicate the studies

to confirm the validity of the results.

The research was conducted using an undergraduate student subject pool. This is a

relatively homogeneous group and not representative of the typical American consumer.

Although this is a concern common to nearly all consumer behavior research, it makes it

is difficult to assess the generalizability of the results to the population.

The studies presented in this dissertation suggested that there is little influence of

attractiveness in the processing of product information. It is certainly possible, however,

that under some conditions, the aesthetics of a product may indeed interfere with a

consumer's ability or inclination to evaluate product quality. It is possible that the

products used in the current study were not sufficiently attractive or interesting to the

student participants, and had more attractive or student-relevant products been used, an

attractiveness effect might have emerged.









Future Research

Experiments 3 and 4 suggested that the effect of product design observed in this

research were due to a diagnosticity effect. The results of experiment 5, however, left the

issue somewhat unresolved. Experiment 5 had some evidence consistent with a

diagnosticity effect, but also suggested that the design effect could have been the result of

an initial reaction coloring subsequent evaluation. The study had two failed stimuli

manipulations, which would need to be corrected in a future experiment. In addition, a

study which manipulates order of information presentation, as in experiment 5, but also

has participants rate features on which the pictures are not diagnostic, as in experiment 3,

could help to resolve the issue.

The research reported here demonstrated that consumers' ability to process written

information is influenced by conflicting visual information. The evidence showed that

when it comes to quality judgments, consumers are not biased by attractiveness, but are

biased by the appearance of performance. This suggests that the effect may be relatively

deliberative and conscious, but it may not be. Additional research is necessary to

determine the degree to which people are aware of the effect of design and the extent to

which people can correct for it. Consistent with research in mental contamination

(Wilson and Brekke 1994) it could be that people are aware of the biasing effect of

design and yet are still unable to counteract it.

As stated above, the dissertation found little aesthetic influence. Prior research has

found that attractiveness does play a role in liking evaluations and may influence quality

evaluations in the absence of other concrete functional information (Page and Herr 2002).

The current research is consistent with these findings, but does not conclude that

aesthetics cannot interfere with information processing. Research in face processing has









demonstrated that under certain conditions, an attractive face can alter the manner in

which functional information is assessed (Landy and Sigall 1974). There may be some

cases in which a sufficiently attractive product can have the same effect. For example,

Norman (2004) suggests that product design can act as a mood manipulation, with good

design elevating mood and making people less likely to notice small functional problems.

If a product is attractive enough to cause an elevation in mood, then it may lead to

motivated reasoning with regard to functional attribute information. Future research is

needed to investigate the conditions under which product aesthetics will and will not

interfere with processing of product information.















APPENDIX A
PRODUCT PICTURES USED IN PILOT STUDY

Unattractive Shoe Attractive Shoe


I i. I:


/ 'I ~a 'a -


Unattractive Backpack



,.;. .:


Attractive Backpack
Allk












APPENDIX B
PRODUCT PICTURES USED IN EXPERIMENT 1


High performance shoes


Low performance shoes


,q













APPENDIX C
PRODUCT PICTURES USED IN EXPERIMENTS 2 THROUGH 5


High Performance Brand


Low Performance Brand


6i a-














1.**,*

















APPENDIX D
PRODUCT DESCRIPTIONS USED IN EXPERIMENT 2, 4 AND 5


Product Low Performance Brand High Performance Brand
Category


Cookware


This cookware set includes a large stock
pot, three pots, and two frying pans. The
pots have well fitting lids that keep heat
in nicely and sturdy handles that stay
cool to the touch. The set is durable and
can be used at the highest heat, even
when cooking with a gas stove. Food
cooks very evenly because of the flat
bottoms that sit squarely on the
elements, maximizing heat distribution
and minimizing burning. If pans sit out
for an extended period, remaining food
can be difficult to remove. The lids for
the stock pot and medium pot also fit the
two frying pans.

These speakers pack several patented
technologies into their slender cabinets
to ensure strong results. They have a
straight-line signal path crossover
network which distributes the signal
with minimal processing. At relatively
low volumes the speakers provide clear
and accurate sound, but slight distortion
may occur at higher volumes. They can
be used with a computer, a stereo or as
part of a home theatre system.

This in-line skate is solid and durable. It
offers an easy-entry system making for
very quick changes. It uses a TrueFit
interior, wrapping a nylon and plastic
film around the foot for extra stability
and control. The fit is snug and well
cushioned. There are tiny slits in the
front which interfere slightly with
aerodynamics, but which help cool the
feet. The wheels are slightly softer than
on other models, making for a more
comfortable ride by increasing shock
absorption. A rigid, well-aligned frame
improves maneuverability.


This cookware has good heat distribution;
food generally cooks evenly. The pots
and pans have a solid base, durable
construction and lids that seal in heat very
well. The stay-cool handles are
ergonomically designed and comfortable
to hold. Very little will stick to these
pans, regardless of cooking method, so
they can be washed out with dish soap
and a cloth; it's rarely necessary to use a
scrub brush. The set comes includes three
handled pots of varying sizes with lids, a
large stock pot with lid, and two frying
pans that fit the lids of the stock pot and
mid-sized pot.

These magnetically shielded speakers are
appropriate for an apartment or dorm, but
would be inadequate for a larger home.
The speakers house two angled, full-
range drivers engineered to generate more
air distribution from a small enclosure
than conventional drivers. They provide
consistent sound at both low and high
volumes. You can almost tell the relative
position of the different instruments in the
recording studio.

This in-line skate has a closer fit in the
ankle collar than older models, giving
stability while still allowing the skater to
flex forward. Lighter materials
throughout give the skate an impressive
weight of only three pounds, which
increases ease of movement. The wheels
are hard, to last longer and improve
speed, but as a result, do not offer that
much shock absorption. This is
compensated for somewhat by the interior
comfort. The skates have a thick liner and
a soft foam interior to increase comfort
and keep feet cool.


Stereo
speakers











In-line
skates










Bicycle
helmets













Electric
mixers


This helmet has a roll-cage interior
bracing somewhat reminiscent of roll
bars injeeps, made to protect the rider
on impact. Despite this additional
bracing, the helmet is reasonably light.
The ventilation is adequate. Although it
only comes in one size, the helmet is
suitable for most head sizes. The foam
on the interior lining is a little stiff, but
the pads can be adjusted to suit your
head size. The buckle is specially
designed to stay secure but allow for a
quick release in case of emergency.

This mixer is relatively compact, fitting
easily on most counters, but the bowl is
deep enough to handle large sized jobs.
It has a gradual starter, minimizing
splatter. It changes speeds very
smoothly, and its plastic gears make the
machine surprisingly quiet. It has an
auto shut-off motor protection feature if
the motor gets overloaded, which does
happen occasionally. Beaters snap in
and out very easily and are dishwasher
safe.


This helmet is durable with a reinforced
rib cage frame. Of course, the reinforced
structure adds some weight to the helmet,
making it a bit heavier than most. But the
lining is soft, thick and well padded with
foam, so the extra weight is not much of a
burden. Superior fit is provided by having
three sizes. Also, it has a sleek wind
funneling technology to pull cool air in
and blow hot air out, cooling the head
during rides. Straps are sturdy and will
stay in place on impact.


This is an efficient mixer with a wide-set
bowl that can be snapped in or lifted out
of the cradle quickly and easily. The
machine has all steel gears and a twin-
motor design that rotates the bowl while
mixing. It can handle everything from
small, high speed tasks to large heavy
doughs. Occasionally the batter will cling
to the beaters, which requires stopping the
machine to scrape the contents back into
the bowl to complete the mixing.

















APPENDIX E
PRODUCT DESCRIPTIONS USED IN EXPERIMENT 3


Product Low Performance Brand High Performance Brand
Category


Cookware


This cookware set includes a large stock
pot, three pots, and two frying pans. The
pots have well fitting lids that
completely seal in heat and sturdy
handles that stay cool to the touch. The
set is durable and can be used at the
highest heat, even when cooking with a
gas stove. Food cooks evenly because of
the flat bottoms that sit squarely on the
elements, maximizing heat distribution
and minimizing burning. If pans sit out
for an extended period, remaining food
can be difficult to remove. The lids for
the stock pot and medium pot also fit the
two frying pans.

These speakers have a straight-line
signal path crossover network which
divides the signal with minimal
processing. Bass response is adequate
but not strong, making these speakers
perfect for a smaller space like a dorm.
The sound is clear and accurate, very
easy to listen to. You can almost tell the
relative position of the different
instruments in the recording studio. The
balancing technology blends direct and
indirect sound almost anywhere in the
room, giving the feeling of surround
sound with only two speakers.

This in-line skate is solid and durable. It
offers a one strap easy-entry system
making for very quick changes. It uses a
TrueFit interior, wrapping a nylon and
plastic film around the foot for extra
stability and control. The fit is snug and
very well cushioned. There are tiny slits
in the front which interfere slightly with
aerodynamics, but which cool the feet.
The wheels are slightly softer than on
other models, making for a more
comfortable ride by increasing shock
absorption.


This cookware has good heat distribution;
food cooks very evenly. The pots and
pans have a solid base, durable
construction and lids that keep heat in
nicely. The stay-cool handles are
ergonomically designed and comfortable
to hold. Very little will stick to these
pans, regardless of cooking method, so
they can be washed out with dish soap
and a cloth; it's rarely necessary to use a
scrub brush. The set comes includes three
handled pots of varying sizes with lids, a
large stock pot with lid, and two frying
pans that fit the lids of the stock pot and
mid-sized pot.

These magnetically shielded speakers
house two angled, full-range drivers
engineered to generate more air
movement from a small enclosure than
conventional drivers. They deliver clean,
uncolored, realistic sound, which is
maintained at higher volumes, provided
of course that the accompanying stereo
system is adequate. They would be
suitable for a home theater system
because the direct/reflect system avoids
the common problem of a speaker sweet
spot and ensures the same quality of
sound throughout the room.

This in-line skate has a close fit in the
ankle collar, giving stability while still
allowing the skater to flex forward.
Lighter materials throughout give the
skate a weight of only three pounds,
increasing ease of movement. The wheels
are hard, to last longer and improve
speed, but as a result, do not offer that
much shock absorption. This is
compensated for very well by a thick liner
and a soft foam interior. No laces and
easy-tug entry grips make the skate very
simple to put on and take off.


Stereo
speakers














In-line
skates












Bicycle
helmets













Electric
mixers


This helmet has a roll-cage interior
bracing made to protect the rider on
impact. Despite this additional bracing,
the helmet is not too heavy. A
streamlined design and well positioned
vents optimize speed. Although it only
comes in one size, the helmet is suitable
for most head sizes. The foam on the
interior lining is a little stiff, but the
pads can be adjusted to suit your head
size. The buckle is specially designed to
stay secure but allow for a quick release
if necessary.

This mixer is relatively compact, fitting
easily on most counters, but the bowl is
deep enough to handle large sized jobs.
It has a gradual starter, minimizing
splatter. It changes speeds very
smoothly, and its plastic gears make the
machine surprisingly quiet. It has an
auto shut-off motor protection feature if
the motor gets overloaded, which does
happen occasionally. Beaters snap in
and out very easily and are dishwasher
safe.


This helmet is strengthened with a
reinforced rib cage frame for safety. Of
course, the reinforced structure adds a bit
of weight to the helmet. But the lining is
soft, thick and well padded with foam, so
the extra weight is not much of a burden.
Superior fit is provided by having three
sizes. Strategically positioned air vents
and advanced wind funneling technology
help to cool the head without sacrificing
speed. Straps are sturdy and will stay in
place on impact.


This is an efficient mixer with a wide-set
bowl and beaters that can be snapped in
or lifted out of the cradle quickly and
easily with one hand. The machine has all
steel gears and a twin-motor design that
rotates the bowl while mixing. The gears
and motors are insulated to minimize
noise. It can handle everything from
small, high speed tasks to large heavy
doughs.









APPENDIX F
PRODUCT PICTURES USED IN EXPERIMENT 4

Attractive Brand Unattractive Brand


t-- 2P -- -r
^s TP>'3^^^


I












9
'I

*~1
Q 3 J
C















LIST OF REFERENCES


Alba, Joseph W., and J. Wesley Hutchinson (1987), "Dimensions of Consumer
Expertise," Journal of Consumer Research, 13 (March), 411-454.

Ambady, Nalini, and Robert Rosenthal (1993), "Half a Minute: Predicting Teacher
Evaluations from Thin Slices of Nonverbal Behavior and Physical Attractiveness,"
Journal ofPersonality and Social Psychology, 64 (March), 431-41.

Beckwith, Neil E., and Donald R. Lehman (1975), "The Importance of Halo Effects in
Mulit-Attribute Attitude Modelsl," Journal ofMarketing Research, 12 (August),
265-75.

Berkowitz, Leonard (1993), "Towards a General Theory of Anger and Emotional
Aggression: Implications of the Cognitive-Neoassociationistic Perspective for the
Analysis of Anger and Other Emotions," in Advances in Social Cognition, Vol. 6,
eds. Robert S. Wyer and Thomas K. Srull, Hilsdale, NJ: Erlbaum, 1-46.

Berlyne, D. E. (1974). "Novelty, Complexity and Interestingness," Studies in the New
ExperimentalAesthetics, ed. D. E. Berlyne, New York: John Wiley & Sons, Inc.,
175-80.

Bloch, Peter H. (1995), "Seeking the Ideal Form: Product Design and Consumer
Response," Journal ofMarketing, 59 (July), 16-29.

Bloch, Peter H., Frederic F. Brunel, and Todd J. Arnold (2003), "Individual Differences
in the Centrality of Visual Product Aesthetics: Concept and Measurement," Journal
of Consumer Research, 29 (March), 551-65.

Boiney, Lindsley G., Jane Kennedy, and Pete Nye (1997), "Instrumental Bias in
Motivated Reasoning: More When More Is Needed," Organizational Behavior and
Human Decision Processes, 72 (October), 1-24.

Breckler, Steven J., and Elizabeth C. Wiggins (1989), "Affect versus Evaluation in the
Structure of Attitudes," Journal ofExperimental Social Psychology, 25 (May), 253-
71.

Cohen, Joel B., and Charles S. Areni (1991), "Affect and Consumer Behavior,"
Handbook of Consumer Behavior, eds. Thomas S. Robertson and Harold H.
Kassarjian, Englewood Cliffs, NJ: Prentice-Hall, 188-240.









Cooper, Robert G., and Elko J. Kleinschmidt (1987), "New Products: What Separates
Winners from Losers?" Journal of Product Innovation Management, 4
(September), 169-84.

Cox, Dena, and Anthony D. Cox (2002), "Beyond First Impressions: The Effects of
Repeated Exposure on Consumer Liking of Visually Complex and Simple Product
Designs," Journal of the Academy ofMarketing Science, 30 (spring), 119-30.

Creusen, Marielle E. H., and Jan P. L. Schoormans (2005), "The Different Roles of
Product Appearance in Consumer Choice," Journal of Product Innovation
Management, 22 (January), 63-81.

Darley, John M., and Paget H. Gross (1983), "A Hypothesis-Confirming Bias in Labeling
Effects," Journal of Personality and Social Psychology, 44 (January), 20-33.

Deighton, John (1984), "The Interaction of Advertising and Evidence," Journal of
Consumer Research, 11 (December), 763-70.

Dickie, George (1971), Aesthetics: An Introduction, New York: The Bobs-Merrill
Company, Inc.

Ditto, Peter H., and David F. Lopez (1992), "Motivated Skepticism: Use of Differential
Decision Criteria for Preferred and Nonpreferred Conclusions," Journal of
Personality and Social Psychology, 63 (October), 568-84.

Eagly, Alice H., Richard D. Ashmore, Mona G. Makhijani, and Laura C. Longo (1991),
"What Is Beautiful Is Good, But...: A Meta-Analytic Review of Research on the
Physical Attractiveness Stereotype," Psychological Bulletin, 110 (January), 109-28.

Edwards, Kari (1990), "The Interplay of Affect and Cognition in Attitude Formation and
Change," Journal ofPersonality and Social Psychology, 59 (August), 202-16.

Feldman, Jack M., and John G. Lynch, Jr. (1988), "Self-Generated Validity and Other
Effects of Measurement on Belief, Attitude, Intention, and Behavior," Journal of
Applied Psychology, 73 (August), 421-35.

Goldman, Morton, Mark D. Cowles, and Charles A. Florez (1983), "The Halo Effect of
an Initial Impression upon Speaker and Audience," Journal of Social Psychology,
120 (August), 197-201.

Green, Christopher D. (1995), "All that Glitters: A Review of the Psychological Research
on the Aesthetics of the Golden Section," Perception, 24 (August), 937-68.

Hassin, Ran, and Yaacov Trope (2000), "Facing Faces: Studies on the Cognitive Aspects
of Physiognomy," Journal ofPersonality and Social Psychology, 78 (November),
837-52.









Herr, Paul M., Frank R. Kardes, and John Kim (1991), "Effect of Word-of-Mouth and
Product-Attribute Information on Persuasion: An Accessibility-Diagnosticity
Perspective," Journal of Consumer Research, 17 (March), 454-62.

Hirschman, Elizabeth C. (1986), "The Effect of Verbal and Pictorial Advertising Stimuli
on Aesthetic, Utilitarian and Familiarity Perceptions," Journal ofAdvertising, 15
(2), 27-34.

Hoch, Stephen J., and Young-Won Ha (1986), "Consumer Learning: Advertising and the
Ambiguity of Product Experience," Journal of Consumer Research, 13
(September), 221-233.

Holbrook, Morris B., and Joel Huber (1979), "Separating Perceptual Dimensions from
Affective Overtones: An Application to Consumer Aesthetics," Journal of
Consumer Research, 5 (March), 272-83.

Holbrook, Morris B., and William L. Moore (1981), "Feature Interactions in Consumer
Judgments of Verbal versus Pictorial Presentations," Journal of Consumer
Research, 8 (June), 103-13.

Holbrook, Morris B., and Robert M. Schindler (1994), "Age, Sex, and Attitude Toward
the Past as Predictors of Consumers' Aesthetic Tastes for Cultural Products,"
Journal ofMarketing Research, 31 (August) 412-22.

Hsee, Christopher K. (1995), "Elastic Justification: How Tempting but Task-Irrelevant
Factors Influence Decisions," Organizational Behavior and Human Decision
Processes, 62 (June), 330-37.

(1996), "Elastic Justification: How Unjustifiable Factors Influence
Judgments," Organizational Behavior and Human Decision Processes, 66 (April),
122-29.

Hutchinson, J. Wesley, and Joseph W. Alba (1991), "Ignoring Irrelevant Information:
Situational Determinants of Consumer Learning," Journal of Consumer Research,
18 (December), 325-345.

Jordan, Patrick W., B. Thomas, and Ian L. McClelland (1996), "Issues for Usability
Evaluation in Industry: Seminar Discussions," Usability Evaluation in Industry, ed.
Patrick W. Jordan, London: Taylor and Francis, 237-43.

Kahneman Daniel, and Shane Frederick (2002), "Representativeness Revisited,"
Heuristics and Biases: The Psychology ofIntuitive Judgment, eds. Thomas
Gilovich, Dale Griffin, and Daniel Kahneman, New York: Cambridge University
Press, 49-81.

Kalins, Dorothy (2003), "Design 2004: It's Cool to Be Warm," Newsweek, October 27,
2003, 58-73.









Koehler, Jonathan J. (1993), "The Influence of Prior Beliefs on Scientific Judgments of
Evidence Quality," Organizational Behavior and Human Decision Processes, 56
(October), 28-55.

Kruglanski, Arie W. (1990), "Motivations for Judging and Knowing: Implications for
Causal Attribution," The Handbook of Motivation and Cognition, eds. E. Tory
Higgins and R.M. Sorrentino, New York: The Guilford Press, 333-68.

(1996), "Motivated Social Cognition: Principles of the Interface," Social
Psychology: Handbook ofBasic Principles, eds. E. Tory Higgins and Arie W.
Kruglanski, New York: The Guilford Press, 493-520.

Kunda, Ziva (1990), "The Case for Motivated Reasoning," Psychological Bulletin, 108
(November), 480-98.

Landy, David, and Harold Sigall (1974), "Beauty is Talent: Task Evaluation as a
Function of the Performer's Physical Attractiveness," Journal of Personality and
Social Psychology, 29 (March), 299-304.

Langlois, Judith H., Lisa E. Kalakanis, Adam J. Rubenstein, Andrea D. Larson, Monica J.
Hallam, and Monica T. Smoot, (2000), "Maxims or Myths of Beauty: A Meta-
Analytic and Theoretical Review," Psychological Bulletin, 126 (May), 390-423.

Loewy, Raymond (1951), Never Leave Well Enough Alone, New York: Simon &
Schuster.

Loken, Barbara, Ivan Ross, and Ronald L. Hinkle (1989), "Consumer Confusion of
Origin and Brand Similarity Perceptions," Journal of Public Policy in Marketing,
5, 195-211.

Lord, Charles G., Lee Ross, and Mark R. Lepper (1979), "Biased Assimilation and
Attitude Polarization: The Effects of Prior Theories on Subsequently Considered
Evidence," Journal ofPersonality and Social Psychology, 37 (November), 2098-
19.

Mandel, Naomi, and Eric J. Johnson (2002), "When Web Pages Influence Choice: Effects
of Visual Primes on Experts and Novices," Journal of Consumer Research, 29
(September), 235-45.

Markus, Hazel, and Ziva Kunda (1986), "Stability and Malleability of the Self-Concept,"
Journal ofPersonality and Social Psychology, 51 (May), 858-66.

Martindale, Colin (1984), "The Pleasures of Thought: A Theory of Cognitive Hedonics,"
The Journal of Mind and Behavior, 5 (winter), 49-80.

Martindale, Colin, and Kathleen Moore (1988), "Priming, Prototypicality, and
Preference," Journal ofExperimental Psychology: Human Perception and
Performance, 14 (November), 661-70.









Miller, Dale T., and Michael Ross (1975), "Self-Serving Biases in Attribution of
Causality: Fact or Fiction?" Psychological Bulletin, 82 (March), 213-25.

Miller, Mark, and Jerry Adler (2003), "Design 2004: Isaac Hits His Target," Newsweek,
October 27, 2003, 74-77.

Mueller, Ulrich, and Allan Mazur (1996), "Facial Dominance of West Point Cadets as a
Predictor of Later Military Rank," Social Forces, 74 (March), 823-50.

Nickerson, Raymond S. (1998), "Confirmation Bias: A Ubiquitous Phenomenon in Many
Guises," Review of General Psychology, 2 (June), 175-220.

Nisbett, Richard E., and Timothy D. Wilson (1977), "The Halo Effect: Evidence for
Unconscious Alteration of Judgments," Journal ofPersonality and Social
Psychology, 35 (April), 250-56.

Norman, D. A. (2004). Emotional design: Why We Love (or Hate) Everyday Things. New
York: Basic Books.

Nova Video (2003), "Battle of the X-Planes: Inside the High-Tech, High-Stakes
Competition to Create America's Newest Fighter Plane," Boston: WGBH
Educational Foundation.

Page, Christine, and Paul M. Herr (2002), "An Investigation of the Processes by Which
Product Design and Brand Strength Interact to Determine Initial Affect and Quality
Judgments," Journal of Consumer Psychology, 12 (2), 133-47.

Pham, Michel Tuan, Joel B. Cohen, John W. Pracejus, and G. David Hughes (2001),
"Affect Monitoring and the Primacy of Feelings in Judgment," Journal of
Consumer Research, 28 (September), 167-88.

Prasso, Sheridan (2001), "Boeing Gets Beat," Businessweek, October 29, 12.

Pyszczynski, Tom, and Jeff Greenberg (1987), "Toward an Integration of Cognitive and
Motivational Perspectives on Social Inference: A Biased Hypothesis-Testing
Model," Advances in Experimental Social Psychology, Vol. 20, ed. L. Berkowitz,
San Diego: Academic Press, 297-340.

Raghubir, Priya, and Eric Greenleaf (2006), "Ratios in Proportion: What Should the
Shape of the Package Be?" Journal ofMarketing, 70, forthcoming.

Rhodes, Gillian, Kate Harwood, Sakiko Yoshikawa, Miwa Nishitani, and Ian McLean
(2002), "The Attractiveness of Average Faces: Cross-Cultural Evidence and
Possible Biological Basis," Facial Attractiveness, eds. Gillian Rhodes and Leslie
A. Zebrowitz, Westport CT: Ablex Publishing, 35-58.









Rubenstein, Adam J., Judith H. Langlois, and Lori A. Roggman (2002), "What Makes a
Face Attractive and Why: The Role of Averageness in Defining Facial Beauty,"
Facial Attractiveness, eds. Gillian Rhodes and Leslie A. Zebrowitz, Westport CT:
Ablex Publishing, 1-34.

Russo, J. Edward, Victoria Husted Medvec, and Margaret G. Meloy (1996), "The
Distortion of Information During Decisions," Organizational andHuman Decision
Processes, 66 (April), 102-10.

Russo, J. Edward, Margaret G. Meloy, and Victoria Husted Medvec (1998),
"Predecisional Distortion of Product Information," Journal ofMarketing Research,
35 (November), 438-52.

Schwarz, Norbert (2002), "Feelings as Information: Moods Influence Judgments and
Processing Strategies," Heuristics and Biases: The Psychology ofIntuitive
Judgment, eds. Thomas Gilovich, Dale Griffin and Daniel Kahneman, New York:
Cambridge University Press, 534-47.

Schwarz, Norbert, and Gerald L. Clore (1983), "Mood, Misattribution, and Judgments of
Well-Being: Informative and Directive Functions of Affective States," Journal of
Personality and Social Psychology, 45 (September), 513-23.

Schweitzer, Maurice E., and Christopher K. Hsee (2002), "Stretching the Truth: Elastic
Justification and Motivated Communication of Uncertain Information," The
Journal of Risk and Uncertainty, 25 (September), 185-201.

Shiv, Baba, and Alexander Fedorikhin (1999), "Heart and Mind in Conflict: The
Interplay of Affect and Cognition in Consumer Decision Making," Journal of
Consumer Research, 26 (December), 278-92.

Silvera, David H., Robert A. Josephs, and R. Brian Giesler (2002), "Bigger is Better: The
Influence of Physical Size on Aesthetic Preference Judgments," Journal of
Behavioral Decision Making, 15 (July), 189-202.

Slovic, Paul, and Sarah Lichtenstein (1971), "Comparison of Bayesian and Regression
Approaches to the Study of Information Processing in Judgment," Organizational
Behavior and Human Performance, 6 (November), 649-744.

Takahashi, Shigeko (1995), "Aesthetic Properties of Pictorial Perception," Psychological
Review, 102 (October), 671-83.

Tetlock, Philip E., and Ariel Levi (1982), "Attribution Bias: On the Inconclusiveness of
the Cognition-Motivation Debate," Journal ofExperimental Social Psychology, 18
(January), 68-88.

Todorov, Alexander, Anesu N. Mandisodza, Amir Goren, and Crystal C. Hall (2005),
"Inferences of Competence from Faces Predict Election Outcomes," Science, 10
(June), 1623-26.









Veryzer, Robert W. Jr., and J. Wesley Hutchinson (1998), "The Influence of Unity and
Prototypicality on Aesthetic Response to New Product Designs," Journal of
Consumer Research, 24 (March), 374-94.

Vriens, Marco, Gerard H. Loosschilder, Edwar Rosbergen, and Dick R. Wittink (1998),
"Verbal versus Realistic Pictorial Representations in Conjoint Analysis with
Design Attributes," Journal of Product Innovation Management, 15 (September),
455-67.

Wilson, Timothy D., and Nancy Brekke (1994), "Mental Contamination and Mental
Correction: Unwanted Influences on Judgments and Evaluations," Psychological
Bulletin, 116, (July), 117-42.

Yamamoto, Mel, and David R. Lambert (1994), "The Impact of Product Aesthetics on
the Evaluation of Industrial Products," Journal of Product Innovation Management,
11 (September), 309-24.

Zajonc, Robert B. (1980), "Feeling and Thinking: Preferences Need No Inferences,"
American Psychologist, 35 (February), 151-75.

Zajonc, Robert B., and Hazel Markus (1982), "Affective and Cognitive Factors in
Preferences," Journal of Consumer Research, 9 (September), 123-31.

(1985), "Must All Affect Be Mediated by Cognition?" Journal of Consumer
Research, 12 (December), 363-64.

Zebrowitz, Leslie (1997). Reading Faces: Windows to the Soul? Boulder: Westview Press.















BIOGRAPHICAL SKETCH

JoAndrea Hoegg earned bachelor's degrees in English and education from the

University of British Columbia in Vancouver, Canada. In 1998 she moved to San Diego,

where she completed a Master of Business Administration at San Diego State University.

In 2001 she entered the Ph.D. program in marketing at the University of Florida. She

joined the University of British Columbia as an assistant professor of marketing in July

2006.