by JOURNAL OF CONSUMER RESEARCH Inc - PDF document

andThaler1988),whichareoftennotattained(e.g.,Bau-meister2002).Similarly,individualsmakespeciccon-sumptiondecisionsonthebasisoftheirhealthgoalsandarenotalwayssuccessfulinachievingthesegoals(e.g.,dietingtoloseweight;BagozziandEdwards2000;CochranandTesser1996).Thisarticleaddressesthequestion,Howdoesthefailuretoachieveagoalaffectconsumers’sub-sequentbehavior?Werestrictourfocustobehavioralgoalsthatarecon-cernedwiththecompletionofacertaintaskoraspecic JOURNALOFCONSUMERRESEARCHoutcome.Consideranindividualwhowantstosavemoney.Onegoalthispersonmightsetistosave$500eachmonth.Thisunambiguousspecicationofareferencepointsetsatargetthatmustbemetfortheefforttobeconsideredasuccess,andanythingshortofthistargetwouldbecodedasafailure.Werefertosuchgoalsas“all-or-nothing”goals.Incontrast,theindividualmayspecifythattheirgoalistoimprovetheirmonthlysavingsrateovertime.Inthiscase,thereisnoprespeciedtarget,andthenumberofdollarsputasideeachmonthcanmeasuresuccess.Inthiscase,successisnotabinaryattain/violatedistinction,butismoregraded.Werefertosuchgoalsas“graded”goals.Thisperspectiveisconsistentwithpriorresearch(cf.CochranandTesser1996;Heathetal.1999)thatdistin-guishesbetweengoalsfortheeliminationofanundesirablebehavior(e.g.,quittingsmoking)thataretypicallyofanall-or-nothingnature,andgoalsforattainingdesirablebehaviors(e.g.,learning)thataretypicallygraded.Wefocusonall-or-nothinggoalsandinvestigatewhyfailuretoattainthesegoalsresultsinpoorersubsequentperformance.First,weconsiderresearchinsocialpsychology,whichshowsthatactivitiescodedasfailures(ratherthanpartialsuccesses)arelikelytoresultinlowerperceivedself-ef-cacy,whichinturnhasbeenshowntoresultindemotivation,lowergoalcommitmentandconsequentlylowerperfor-mance(BanduraandSimon1977).Theeasieritistocodeanoutcomeasafailure(givenall-or-nothinggoals),themorelikelyitisthatanindividualwillbedemotivatedtocontinuestriving,andhencetheirsubsequentperformancemaysuffer.Asecondstreamofresearchofferingasimilarconclusionisthatintheareaofself-monitoring.Researchershaveshownthatinassessingtheeffectivenessofgoals,individ-ualscompareestablishedstandardsofperformance(i.e.,thegoal)withfeedbackaboutactualperformancetypicallyob-tainedthroughself-monitoringproceduresinwhichtheyobserveandevaluatetheirbehavior(LockeandLatham1990).Inthecaseofall-or-nothinggoals,thisself-moni-toringwillresultinthetrackingofbehaviorsthatarein-consistentwiththegoal,whichhasbeenshowntobedet-rimentaltoperformance(Bandura1986).Forinstance,researchhasshownthattrackingofgoal-inconsistentbe-havioractuallyincreasedtheincidenceofsmokinginin-dividualswhowereattemptingtodecreaseit(seeCochranandTesser1996).However,inthecaseofgradedgoals,theeffectofself-monitoringwillbepositiveandhenceisnotlikelytoresultinadeteriorationofperformance.Third,theviolationofanall-or-nothinggoalresultsinthegenerationofstrongnegativeemotions(Heathetal.1999).Priorresearchsuggeststhatindividualswhoexpe-riencenegativeemotionsengageinemotionalrepairthroughdistractionandbydivertingattentiontoactivitiesthatcangeneratepositiveemotion(cf.Connolly,Ordonez,andCoughlan1997;GarbarinoandEdell1997).Individualswhoviolateagoalinoneparticulardomainmaythereforebemotivatedtoshiftresourcestoanalternatetaskorgoal,All-or-nothinggoalsmaybeconsideredmorespecicthangradedgoals,holdingpeopletoastricterstandard. hurtingperformanceontheoriginaltask.Intherealworldwhereconsumershavemultiplegoals,violationofonegoalmaycausethemtoshiftattentiontoother,moreachievableAfourthstreamofresearchfocusesontheso-calledre-boundeffect.Ithasbeenshowninvariousdomains(e.g.,thoughtsinadults,obedienceinchildren)thatafteraperiodofarticialsuppressionofathoughtorbehavior(asmightbeexpectedingoalstoeliminateundesirablebehaviors),thesuppressedthoughtoractivityisengagedinmorefrequentlythaniftherewerenoattempttosuppressit(Wegneretal.1987).Duringtheperiodofsuppressionindividualsresorttodistractersandimposepsychiccostsonthemselvestokeeptheunwantedactivityatbay.However,asthesepsychiccostsrequireeffort,individualsmightplantoimposethemonlyforalimitedperiodoftime.Oncethisperiodisover,thesepsychicconstraintsaresuddenlyreleasedandthisfacilitatestherecurrenceoftheforbiddenactivity(Wegneretal.1987).Theviolationofanall-or-nothinggoalmightsignalthecom-pletionofaperiodofsuppression,andtheensuingreboundwouldresultinpoorsubsequentperformance.Thus,fourindependentstreamsofresearchleadustohypothesizethatTheviolationofabehavioralgoalthatconstrainsanundesirableactivitywillresultinanultimatedeteriorationofperformance.Oncethegoalisvi-olated,thepreviouslyconstrainedactivitywillbeengagedinwithagreaterlikelihoodthanifthegoalwasyettobeviolated.Consistentwiththisprediction,CochranandTesser(1996)offertheexampleofastudentwhowasstrivingtoreduceweightusingadailycaloricgoal.Oneday,thestu-dentrealizedthatthedailyquotaofcalorieshadaccidentallybeenexceededandproceededtoconsumeapplepie.Thereasoningwas“What-the-hell.SinceI’malreadyovermygoal,itdoesn’tmatter.”CochranandTesser(1996)refertothisresponseasa“whatthehell”effect.RecentresearchbyDharandSimonson(1999)isalsoconsistentwithhypothesis1andthewhatthehelleffect.Theseauthorsshowthatwhenconsumerstradeoffagoal(likethepresenceorabsenceofpleasure)witharesource(likemoney),theypreferahighlightingstrategyinwhichallgoal-consistentconsumptionoccursinoneepisode.Forexample,inselectinganappetizerandentre´e,aconsumerwillselectatastyappetizerandatastyentre´eononedayandatastelessyethealthyappetizerandentre´eonanotherday.Ourframeworkwillmakeasimilarhighlightingpre-dictionforaconsumerwhosegoalistoavoideatingtastyandunhealthyfood.Ifthisconsumerhasalreadyviolatedagoalbyconsumingarichappetizer,ourframeworkpredictsthattheymightgoaheadandconsumethetastyentreanyway.Basedontheearlierdiscussion,weexpecthypothesis1toholdunderthefollowingconditions:1)Thegoalisanall-or-nothingtypeofgoal:Weexpectthedeteriorationofperformancetooccurforall-or- JOURNALOFCONSUMERRESEARCHTABLE1EFFECTOFTARGETDISCRETIONARYSPENDINGANDPREVIOUSDISCRETIONARYSPENDINGONWILLINGNESSTOSPEND:EXPERIMENT1 Targetdis-Previousdiscretionaryspending HK$6,500HK$7,500HK$8,500 HK$8,000Condition1(Group1)4.31(2.28)Condition2(Group2)2.69(1.70)Condition3(Group3)6.23(1.74)HK$7,000Condition4(Group2)2.77(1.92)Condition5(Group3)6.31(2.14)Condition6(Group3)5.92(1.75) Group1representstheconditioninwhichthegoalwillnotbeviolatedbytheadditionalexpense(JustSufÞcientSurplus).Group2representsconditionswherethegoalhasnotbeenviolatedsofarbutwillbeviolatediftheadditionalexpenseisincurred(InsufÞcientGroup3representsconditionsinwhichthesavingsgoalhasalreadybeenviolated(Overspent).TheSDoftheparticipantsÕreportedlikelihoodofspending(SPEND)isreportedinparentheses.Therewere13participantsineachexperimental(e.g.,HeathandSoll1996)predictsthatthelikelihoodofincurringanexpensedependsontheextenttowhichthebudgethasalreadybeendepleted.Inthecontextofthepre-sentdesign,therefore,thepredictionfrompriorresearchwouldbethatparticipantswhosepreviousdiscretionaryspendingwasHK$6,500wouldbethemostlikelytospend,followedbyparticipantswhohadspentHK$7,500,and-nallybythosewhohadspentHK$8,500.However,participantsinsomeoftheconditionshadal-readyviolatedtheirspendingtargetswhileothershadnot.Further,thescenarioimplicitlyledparticipantstotheun-derstandingthatthepersonalspendinggoalwouldbede-sirable.Hence,inadditiontostudyingthemainandinter-actioneffectsofthetwodesignfactors,wewereparticularlyinterestedincontrastingtheresponsesofparticipantswhohadalreadyviolatedtheirgoal,withtheresponsesofthosewhohadnotyetviolatedtheirgoal.ThemeanSPEND(likelihoodofspending)foreachoftheexperimentalconditionsisshownintable1andwasanalyzedina2(TargetDiscretionarySpending)3(PreviousDis-cretionarySpending)ANOVA.Resultsindicatedasignicanttwo-wayinteractionofTargetDiscretionarySpendingwithPreviousDiscretionarySpending(,(2,72)),aswellasasignicantmaineffectofPreviousDis-cretionarySpending(,).Contrary(2,72)11.39topreviousresearch,themaineffectofPreviousDiscre-tionarySpendingsuggestedthatthewillingnesstospendwasthehighestwhenpreviousspendingwasHK$8,500,lowerwhenpreviousspendingwasHK$7,500,andthelow-estwhenitwasHK$6,500.SpendingmoremoneyinthepastactuallyresultedinalargerwillingnesstospendtheadditionalHK$1,500.WhydidthiscounterintuitiveresultWecategorizethesixexperimentalconditionsintothreegroupsasfollows(seetable1):Group1:(JustSufcientSurplus)Thisgroupincludesparticipantswhohadnotviolatedtheirsavingsgoal.IftheyspendtheextraHK$1,500,theywillhavespentuptotheirtargetbutnotexceededit.Participantsincondition1(PreviousSpendingHK$6,500,TargetHK$8,000)areinthisgroup.Group2:(InsufcientSurplus)TheseparticipantshavenotviolatedtheirsavingsgoalbutwoulddosoonspendingtheHK$1,500.Participantsincondition2(PreviousHK$7,500,TargetHK$8,000)andcon-dition4(PreviousSpendingHK$6,500,TargetHK$7,000)areinthisgroup.Group3:(Overspent)Thisgroupincludesparticipantswhohavealreadyviolatedtheirsavingsgoal.TheextraHK$1,500expensewillnotviolatetheHK$10,000avail-ableincomeconstraint.Participantsincondition3(Pre-viousSpendingHK$8,500,Targetcondition5(PreviousSpendingHK$6,500,TargetHK$7,000)andcondition6(PreviousSpendingHK$8,500,TargetHK$7,000)areallpartofthisContrastsrevealedthatthewillingnesstospend(SPEND)forgroup3(Overspent)()wassignicantlyhigherthanSPENDforgroup1(JustSufcientSurplus)(;).SPENDforgroup1(JustSufcientSurplus)()wasinturnsignicantlygreaterthanSPENDforgroup2(InsufcientSurplus)(;).Thewillingnesstospendwashighestwhenthepersonalgoalhadalreadybeenviolated,secondwhentheextraspendingwouldmovecumulativespendinguptothepersonalgoal(butnotviolateit)andthelowestwhentheextraexpensewoulddirectlyresultinaviolationofthepersonalgoal.Resultsfromtherstexperimentsupportedhypothesis1.Specically,wesawparticipantsdisplaythegreatestwill-ingnesstospendwhenthesavingsgoalhadalreadybeenviolated.Conversely,inasituationinwhichtheywereap-proachingthegoalandcouldattainitbyavoidingtheex-pense,participantsshowedthelowestwillingnesstospend.Notethatparticipantswhohadjustsufcientsurplusintheiraccountwerelesslikelytospendascomparedtothosewhohadalreadyoverspent.Itispossiblethattheformerweresomewhatconstrainedbythebudgetceiling;thelatterhad JOURNALOFCONSUMERRESEARCHTABLE2NUMBEROFPARTICIPANTSWHOFINISHTHEASSIGNEDTASK:EXPERIMENT2 ProximalgoalDistantgoalProximal+DistantNogoal A.Participantscompletingbypersonalgoal10(29%)23(64%)33(46%)...B.Participantscompletingafterpersonalgoal12(34%)7(19%)19(27%)...C.Participantsnotcom-pletingin30days(Vio-latorsÑdidnotÞnish)13(37%)6(17%)19(27%)14(38%)Totalparticipantscomplet-ingin30days(A+B)22(63%)30(83%)52(73%)23(62%)Totalnumberofpartici-pantsincondition(A+B+C)35(100%)36(100%)71(100%)37(100%) Participantscompletingbypersonalgoal:Thenumberofparticipantsineachconditionwhocompletedthetaskwithintheself-reportedduration.Participantscompletingafterpersonalgoal:Thenumberofparticipantsineachconditionwhocompletedthetaskaftertheself-reportedduration,butinthe30daystimeperiod.sweredanopen-endedquestion.Participantswerethenthankedfortheirparticipation,paid,anddismissed.ResultsandAnalysisFirst,asamanipulationcheck,welookedattheactualdeadlinessetbyparticipantsinthetwogoalconditions.Intheproximalgoalcondition,theaveragedeadlinewassetat8.6daysfromthetimeofpickingupthebooklet,whileinthedistantgoalconditiontheaveragedeadlinewas17.17daysaway(,).Theinstructiontosetaproximalgoalthereforedidindeedresultinshorterself-generatedpersonaldeadlinesascomparedtotheinstructiontosetadistantgoal.Wenextconsiderperformanceacrossthethreeexperi-mentalconditions.Weusethreemeasuresofperformance.First,welookedatthepercentageofparticipantswhosuc-cessfullynishedthetask(i.e.,gotpaid)withintheentire30-dayperiod(wehereafterrefertothispercentageasFIN-ISH).Foreachofthetwogoalconditions,wealsocomputedFINISH-GOAL,thenumberofparticipantswhonishedbeforetheirself-imposedgoalasafractionofallparticipantsinthatconditionwhonished.Second,welookedatthetotalnumberofdaysittookfortheparticipanttonishthetask(DAYS).Third,wecountedthenumberofmistakesthathadbeencorrectlyidentiedbytheparticipant.Asthetotalnumberofmistakeswasheldconstantat120forallparticipants,thisnumbergaveusanindexoftheaccuracyofeachparticipant(ACCURACY).Thenumberofparticipantswhonishedthetaskatvar-iousstagesofthe30-dayperiodineachofthethreeex-perimentalconditionsisshownintable2.Notsurprisingly,thepercentageofparticipantswhonishedthetaskwithin30daysandclaimedtheirpaymentwashigherinthedistantgoalcondition()thanintheFINISH30/36dist.nogoalcondition(;FINISH23/37,).However,theparticipantsintheproximalgoalconditionfarednobetterthantheparticipantswhohadnogoal(;,FINISH22/35prox.).Settingagoalincreasedthenishingrate,butonlywhenthegoalwasdistantandhenceeasyandnotwhenitwasproximalandchallenging.Theperformanceforprox-imal(i.e.,challenging)goalswasactuallypoorer,providingsupportforprevioussuggestionsthatverydifcultgoalsmayactuallynotbebenecial.Next,withinthetwoconditionsinwhichparticipantssetapersonalgoal,weexaminedthepercentageofparticipantswhonishedbeforetheirpersonalgoal(measuredasaper-centofallthosewhonishedinthe30-dayperiod).Thispercentwassignicantlygreaterforthedistantgoalcon-dition()thantheFINISHGOAL23/30dist.proximalgoalcondition(FINISHGOAL10/22prox.;,).Thisndingisnotsur-prising—participantsinthedistantgoalconditiononanav-eragehadsetforthemselvesadeadlinethatgavethemanadditional8.57()daystonishthetaskascomparedtoparticipantsintheproximalgoalcondition.Forthepurposeoffurtheranalysis,wecategorizedpar-ticipantsinthegoalconditionsintooneoftwosub-groups—thosewhocompletedthetaskbeforetheirpersonaldeadlines(werefertotheseasachievers)andthosewhodidnot(violators).Theviolatorsgroupwasfurthercom-prisedoftwoparts—thosewhoeventuallywentontonishwithinthe30-dayperiod(violators—nished)andthosewhonevernished(violators—didnotnish).Wecomparedthefollowingvegroupsofparticipants:Group1:Nogoalparticipants(,ofwhich23Group2:Proximalgoalachievers()Group3:Proximalgoalviolators(,ofwhich12Group4:Distantgoalachievers()Group5:Distantgoalviolators(,ofwhichseven WHENGOALSARECOUNTERPRODUCTIVETABLE3DISTRIBUTIONOFNUMBEROFDAYSTAKENTOFINISHACROSSEXPERIMENTALCONDITIONSANDNUMBEROFPARTICIPANTSWHOFINISHTHEASSIGNEDTASK:EXPERIMENT2 completingtaskNogoal Proximalgoal Distantgoal %of %of %of %of %of %of %Cum.%%Cum.%%Cum.%%Cum.%%Cum.%%Cum.% 1Ð5days144333141499277666Ð10days52226141662741172693037253111Ð15days4174311271545329274363616Ð20days417611138295563451760145021Ð25days29705433146894393090257526Ð30days73010019627321002063310100883Total:1Ð30days231006222100633010083 %ofcompletersisthefollowingpercentage:NumberofÞnishersinthespeciÞedtime/NumberofÞnishersin30days.%ofparticipantsisthefollowingpercentage:NumberofÞnishersinthespeciÞedtime/Numberofparticipantsineachcondition.Thequestionweseektoansweristhefollowing:Intermsofperformance,isitbettertohavenogoalatall(group1),ortohaveagoalandfailtoachieveit(groups3and5)?Werstcomparethepercentofparticipantswhonishedthetaskbeforethe30-dayperiod.Forgroup1,thispercentwas62.16%().Forgroups3and5,thepercentwas23/37only50%().Intermsofeventuallyn-[12+7]/[25+13]ishingthetask,itwasapparentlybettertohavenogoalatallthantohavehadagoalandhavefailedtoattainit,althoughthisdifferencewasstatisticallynotsignicant(,).However,achieving(orfailingtoachieve)anall-or-nothinggoalmayalsoaffectsubsequentperformanceinagradedmanner.Thiswasanalyzednext,byfocusingonthenumberofdaystakentocompletethetask(DAYS)andthenumberoferrorscorrectlyidentiedWeconsideredallparticipantswhonishedthetaskandcomparedthenumberofdaysittookthemtocompletethetask(asmeasuredfromthedaytheypickedupthebooklet).Notsurprisingly,participantswhosetandachievedtheprox-imalgoaltooktheshortestperiodoftimetocompletethetask()andparticipantswhosetandDAYSgr.2achievedthedistantgoaltooksignicantlylonger(;,).ParticipantsDAYS14.04gr.4whohadsetnopersonalgoalstookevenlongertonishthetask(;,).OfDAYS(43)gr.1fin.specicrelevancetoourpoint,however,participantswhosetgoalsbutfailedtoachievethemtookmuchlongertonishthetaskthanparticipantswhohadsetnogoals.Inparticular,participantswhosetandviolatedaproximalgoaltooksignicantlylongerthanthosewhohadnogoals(,;DAYS25.50DAYSgr.3fin.gr.1fin.,).Andparticipantswhosetandviolatedadistantgoalalsotooksignicantlylongerthanthosewhohadnogoals(,DAYS25.14DAYSgr.5fin.gr.1fin.;,).Therewasnodifferencein3.04thenumberofdaystakenbyproximal(DAYSgr.3fin.)versusdistantgoalviolators(25.50DAYSgr.5fin.;).Overall,participantswhosetagoalandviolatedit(group3andgroup5)tooklonger(DAYS)thanthosewhohadnogoals(25.37DAYSgr.1fin.;,).Table3showsthefre-quencydistributionofthenumberofdaystakentonishthetaskineachoftheexperimentalconditions.Wefurthercomparedtheaccuracylevelsofparticipantswhosuccessfullycompletedthetaskineachofthesevegroupsbystudyingthenumberofmistakescorrectlyiden-tied.Theresultsfollowapatternsimilartothenumberofdaystaken.Participantswhosetandachievedagoal—eitherdifcultoreasy—showedthehighestaccuracy(,;ACCURACY106.90ACCURACYgr.2gr.4).Participantswhohadsetnopersonalgoalswerelessaccurate(;)thanACCURACYgr.1fin.theprevioustwogroups.Ofspecicrelevancetoourpoint,however,participantswhohadsetgoalsbutfailedtoachievethemweretheleastaccurate.Inparticular,participantswhohadsetandviolatedaproximalgoalwerelessratethanthosewhohadnogoals(ACCURACYgr.3fin.,;,91.42ACCURACYgr.1fin.).Andparticipantswhohadsetandviolatedadistantgoalwerealsolessaccuratethanthosewhohadnogoals(,;ACCURACY90.29ACCURACYgr.5fin.gr.1fin.,).Therewasnodifferenceintheac-curacyofproximal()versusACCURACYgr.3fin.distantgoalviolators(;ACCURACYgr.5fin.).Overall,participantswhosetagoalandviolatedit(group3andgroup5)werelessaccurate(ACCURACY)thanthosewhohadnogoals(91.00ACCURACYgr.1fin.;,).Astable2andtheprecedinganalysesshow,goals—especiallydifcultones—canbecounterproductive.Failuretoachievethesegoalscanresultinsubsequentperformancethatisworsethanthatofparticipantswhohadsetnogoalsatall.Whydidtheviolationofgoalsresultinpoorerperfor- WHENGOALSARECOUNTERPRODUCTIVEactioninoneperiodonly.Unlikemanyrealworldgoalsthataresettoaccomplishacomponentofalargerobjective,theparticipantsdidnothavetheopportunitytobroadenthescopeofthegoaltoconvertafailureintoapotentialsuccessbecausetheymadedecisionsforthepresentperiodonly.Giventheopportunitytomakemultipledecisions,weexpectthatparticipantswillbeabletocountersomeofthenegativeeffectsoffailurebybroadeningthegoal,andhenceper-formancemaynotdeteriorategreatly.Inordertotestforthemoderatingeffectoftheabovevariableswewouldneedtoconductanexperimentinwhichthecurrentexperiments’designisembeddedineachofthefourconditionscreatedbyfullycrossingtwolevelsoftheNatureoftheGoal(all-or-nothingvs.graded)withtwolev-elsofOpportunitytoBroadentheScopeoftheGoal(noopportunityvs.opportunity).Asareplicationofthepresentresearch,wewouldexpecttoseethecounterproductiveef-fectsofgoalsforall-or-nothinggoalswhentherewasnoopportunitytobroadenitsscope.Wewouldfurtherexpecttoseeaweakeningofthiseffectwhenthegoalwaseitherconvertedtobeofagradednature,oritsscopecouldbebroadened.Finally,wemightexpectafurtherweaken-ing—orperhapsnotseethenegativeeffectsofgoalsatall—whengoalsaregradedandtheycouldbebroadened.WeplantoinvestigatethesemoderatingeffectsinfutureInthisarticlewedemonstratedthepossiblecounterpro-ductivenatureofgoals—specically,failuretoachieveagoalleadstopoorersubsequentperformance.Onthebasisofpriorresearchongoalachievementwealsospeculatedaboutmechanismsthatcouldexplainthiseffect.Asdis-cussedpreviously,poorperformanceonagoalmayleadtoone(ormore)offourpossibleprocesses.Consumerswhoareunabletoachieveagoalmay:reducetheperceptionofhowlikelytheyaretoachieveagoal(i.e.,decreaseself-efcacy);focustoomuchonthethingstheydidwrong(i.e.,negativeself-monitoring);beunhappyasaresultofthefailure(i.e.,negativeemotion);orengageinmoreoftheforbidden(andlessofthedesirable)activityaftertryingtofocusonthedesirableactivityforsometime(i.e.,reboundeffect).Thesefourpossibleprocessesresultindecreasedmotivationtotrytoachieveagoal,andsubsequentpoorerperformance.Wenowaddressfutureworkthatcouldteaseapartthesepossibleexplanations.First,failingtoachievethepreassignedgoalmayresultindecreasedself-efcacy(BanduraandSimon1977).Therearetwopossiblewaystotestfortheexistenceofthisprocess.One,consumers’perceivedself-efcacycouldbemeasuredbeforetheybeginthersttask,andthenaftertheyachieve(failtoachieve)theirgoal.Measuresofself-efcacycouldincludeasetofstatementssuchas“Ingeneral,IachievethegoalsIsetformyself”or“OnceIsetmymindtoit,IusuallyaccomplishwhatIsetouttodo.”Ifgoalperformanceaffectsself-efcacy,consumerswhosucceedinachieving(failtoachieve)thegoalwouldreporthigher(lower)scores.Moreover,thismeasurewouldmediatetheeffectofgoalachievementonsubsequentperformance(BaronandKenny1986;Kenny,Kashy,andBolger1998).Two,self-efcacycouldbemanipulatedbygivingexperimentparticipantssce-nariosincludingdescriptionsportrayingthetargetindivid-ualsasmore(less)efcacious.Thisprocessexplanationwouldpredictthatindividualsinthelow(high)self-efcacyconditionwouldperformpoorly(better).Second,negativeself-monitoringmaycauseconsumerstocontinuedoingpoorlyfromtheirrstperformancetotheirsecondone(CochranandTesser1996).Ifthisisthemech-anismthroughwhichantecedentfailureaffectssubsequentperformance,amanipulationthatcausesparticipantstofo-cusmoreoninstanceswheretheyactuallydidwell(poorly)willimprove(worsen)subsequentperformance.Suchanelaborationmanipulationmayalsobeusedtoseparatethisexplanationfromthefourthone,asexplainedbelow.Third,negativeemotionfromfailure(Heathetal.1999)mayaffectmotivationtofollowsubsequentsimilargoals(Connollyetal.1997).Inthiscontext,goal-seekingiscon-sideredtobearesourcethatisdepletedbynegativeemotion.Onemanipulationthatwouldtestthispredictionwouldin-volveintroducingawindfallgainafterthersttaskiscom-pleted.SimilartothemanipulationbyArkesetal.(1994),halftheparticipantscouldbetoldthattheyjustwon$100atarandomdrawingattheirlocalradiostation.Thisma-nipulationwouldmoderatetheeffectofgoalsuccess/failureonsubsequentperformance(Kennyetal.1998),byprovid-ingparticipantssomepositiveemotiontooffsetthenegativeemotionexperiencedasaresultoffailingtoattainagoal.Specically,forparticipantswhoreceivethewindfall,therewillbenodifferenceinsubsequentperformanceirrespectiveofwhetherornottheyattainedtheirgoal.However,partic-ipantswhodonotreceiveawindfallwillperformbetter(worse)afterachieving(failingtoachieve)agoal.Fourth,thereboundeffectexplanationthatconsumerswill“letthemselvesgo”aftertryingtoachieveagoalforawhile,couldbetestedbytheelaborationmanipulationmentionedearlierforthenegativeself-monitoringproposition.Ifthereboundeffectactuallyoccurs,however,participantswhobelievetheyhavebeendoingwell(positiveelaborationcon-dition)willbemorelikelytoperformpoorlyinasubsequenttaskascomparedtoparticipantsinthenegativeelaborationcondition,whobelievethattheyhaveperformedpoorlyinthersttask.Notethattheeffectoftheelaborationmanip-ulationgivenareboundeffectwillbeoppositetothatpre-dictedfornegativeself-monitoring.Asmentionedearlier,theseeffectsarenotallmutuallyexclusiveandmaypossiblycoexist.Forinstance,webelievethatfailuretoachieveagoalwillresultinnegativeemotionsanddecreasedself-efcacy.Astudythatmeasuresself-ef-cacywhilesimultaneouslymanipulatingemotion(e.g.,throughawindfallgain)andelaboration(positive/negative)willallowustofurtherstudyexactlywhichofthesefactorsarecrucialantecedentsofpoorperformanceandthuscon-tributetothecounterproductiveeffectofgoals.Insummary,thecurrentprogramofresearchstudiestheeffectofgoalachievementperformanceonsubsequenttasks,thusextendingpreviousresearchongoalachievement.We