ALittleThanksGoesaLongWay:ExplainingWhyGratitudeExpressionsMotivatePro - PDF document

ALittleThanksGoesaLongWay:ExplainingWhyG
ALittleThanksGoesaLongWay:ExplainingWhyGratitudeExpressionsMotivateProsocialBehaviorAdamM.GrantUniversityofPennsylvaniaFrancescaGinoUniversityofNorthCarolinaatChapelHillAlthoughresearchhasestablishedthatreceivingexpressionsofgratitudeincreasesprosocialbehavior,littleisknownaboutthepsychologicalmechanismsthatmediatethiseffect.Weproposethatgratitudeexpressionscanenhanceprosocialbehaviorthroughbothagenticandcommunalmechanisms,suchthatwhenhelpersarethankedfortheirefforts,theyexperiencestrongerfeelingsofself-efficacyandsocialworth,whichmotivatethemtoengageinprosocialbehavior.InExperiments1and2,receivingabriefwrittenexpressionofgratitudemotivatedhelperstoassistboththebeneficiarywhoexpressedgratitudeandadifferentbeneficiary.Theseeffectsofgratitudeexpressionsweremediatedbyperceptionsofsocialworthandnotbyself-efficacyoraffect.InExperiment3,weconstructivelyreplicatedtheseeffectsinafieldexperiment:Amanager’sgratitudeexpressionincreasedthenumberofcallsmadebyuniversityfundraisers,whichwasmediatedbysocialworthbutnotself-efficacy.InExperiment4,adifferentmeasureofsocialworthmediatedtheeffectsofaninterpersonalgratitudeexpression.Ourresultssupportthecommunalperspectiveratherthantheagenticperspective:Gratitudeexpressionsincreaseprosocialbehaviorbyenablingindividualstofeelsociallyvalued.gratitude,prosocialbehavior,helping,agencyandcommunion,socialworthWearebetterpleasedtoseethoseonwhomweconferbenefitsthanthosefromwhomwereceivethem.—LaRochefoucauld,Gratitudeisomnipresentinsociallife.Peoplefeelgratefulwhentheybenefitfromgifts,assistance,kindness,help,favors,andsupportfromothers(Tesser,Gatewood,&Driver,1968).Gratefulfeelingshaveseveralbeneficialeffects:Theyenableindividualstosavorpositiveexperiences,copewithstressfulcircumstances,andstrengthensocialrelationships(Lyubomirsky,Sheldon,&Schkade,Psychologicalresearchhighlightsthebenefitsofgratitudeasatrait,demonstratingthatdispositionalgratitudeisassociatedwithhigherlevelsofsubjectivewell-being(McCullough,Tsang,&Emmons,2004),andasastate,demonstratingthattheactofcountingone’sblessingscanincreasepositiveemotions,subjectivewell-being,andhealth(Emmons&McCullough,2003;Seligman,Steen,Park,&Peterson,2005).Behaviorally,gratitudeisaproso-cialtraitandstate:Itmotivatesindividualstoengageinprosocialbehaviorstoreciprocatetheassistancetheyreceivefromothers(Bartlett&DeSteno,2006;Tsang,2006).Althoughresearchprovidesvaluableinsightsintobeneficiaries’experiencesofgratitude,itofferslessinformationabouthowbeneficiaries’expressionsofgratitudeaffecthelpers.Becausegrat-itudeis,bydefinition,asocialemotionproducedinsocialex-changes(McCullough,Kilpatrick,Emmons,&Larson,2001),itiscriticaltoexaminehowgratitudeaffectsbothpartnersinsocialexchanges.Towardthisend,anumberofstudieshaveprovidedinitialevidencethatgratitudeexpressionsmotivateprosocialbe-havior(forareview,seeMcCulloughetal.,2001).However,littleresearchhasbeendonetoexaminewhygratitudeexpressionsmotivateprosocialbehavior.Throughwhatpsychologicalpro-cessesdoesbeingthankedleadtohigherlevelsofhelping?Weaddressthisquestionbydrawingontheclassicdistinctionbetweenagencyandcommunion.Psychologistshavelongarguedthatindividualshavebasicmotivestofeelboth,orper-sonallycompetentandcapable,and,orconnectedtoandvaluedbyothers(Bakan,1966;McAdams&deSt.Aubin,1992;Wiggins,1979).Wecomparetheagenticandcommunalmechanismsthatmaymediatetheeffectsofgratitudeexpressionsonprosocialbehavior.Fromanagenticperspective,expressionsofgratitudemayenhancehelpers’feelingsofself-efficacy,whichwillmotivatethemtoengageinprosocialbehaviorbyreducingtheirfeelingsofuncertaintyaboutwhethertheycanhelpeffec-tively.Fromacommunalperspective,expressionsofgratitudemayenhancehelpers’feelingsofsocialworth,whichwillmotivatethemtoengageinprosocialbehaviorbyreducingtheirfeelingsofuncertaintyaboutwhethertheirhelpwillbevaluedbybeneficia-ries.Acrossfourexperiments,wecomparetheseagenticandcommunalmechanismstoexplainwhygratitudeexpressionsin-creaseprosocialbehavior.GratitudeExpressionsandProsocialBehaviorGratitudeisafeelingofthankfulnessdirectedtowardothersthatemergesthroughsocialexchangesbetweenhelpersandbeneficia-AdamM.Grant,TheWhartonSchool,UniversityofPennsylvania;FrancescaGino,Kenan-FlaglerBusinessSchool,UniversityofNorthCarolinaatChapelHill.Forassistancewithdatacollectionandentry,wethankRebeccaBram-lett,BethBraxton,StanCampbell,JennyDeveau,HowardHeevner,YuxiLiu,andAaronMaas.CorrespondenceconcerningthisarticleshouldbeaddressedtoAdamM.Grant,TheWhartonSchool,UniversityofPennsylvania,3620LocustWalk,Suite2000SH/DH,Philadelphia,PA19104-6370.E-mail:JournalofPersonalityandSocialPsychology©2010AmericanPsychologicalAssociation2010,Vol.98,No.6,946–9550022-3514/10/$12.00DOI:10.1037/a0017935ries(Blau,1964).Beneficiariesexperiencegratitudewhentheyattributetheirfavorablecircumstancestotheeffortsofahelper(Weiner,1985).Beneficiariesoftenexpressgratitudebythankinghelpersfortheircontributions.Indeed,psychologistshavemanip-ulatedgratitudethrough“gratitudevisits”inwhichbenefic

iariesexpressthankstohelpers(Seligmaneta
iariesexpressthankstohelpers(Seligmanetal.,2005).Researchhasshownthatthesegratitudevisitspositivelyaffectnotonlythebeneficiariesbutalsothehelpersthemselves.Grati-tudeexpressionsappeartoserveasmoralreinforcersinenhancinghelpers’prosocialbehavior(McCulloughetal.,2001).Anumberofexperimentshaveshownthatwhenhelpersarethankedbythebeneficiariesoftheirhelp,helpersaremorewillingtohelpthesebeneficiariesagain(Carey,Clicque,Leighton,&Milton,1976;H.B.Clark,Northrop,&Barkshire,1988;McGovern,Ditzian,&Taylor,1975;Rind&Bordia,1995)andtohelpothers(R.D.Clark,1975;Goldman,Seever,&Seever,1982;Moss&Page,1972).However,weknowlittleaboutthemediatingpsychologicalprocessesunderlyingtheseeffects:Whydobeneficiaries’gratitudeexpressionsmotivatehelpers’prosocialbehavior?Becausegratitudeexpressionsaredeliveredbybeneficiariestohelpersaspartofasocialexchangeprocess,gratitudeexpressionsarelikelytoinfluencehowhelpersviewthemselvesinthesocialworld.Arichhistoryoftheoryandresearchinpsychologysug-geststhatindividuals’self-viewsinthesocialworldvaryalongtwodimensions:agencyandcommunion(Bakan,1966;Fiske,Cuddy,&Glick,2007;Wiggins,1979).Agencyreferstofeelingsofpersonalcompetenceorself-efficacy,andcommunionreferstofeelingsofinterpersonalwarmthorconnectednesstoothers(forareview,seeJudd,James-Hawkins,Yzerbyt,&Kashima,2005).Weproposethatgratitudeexpressionscanincreasehelpers’proso-cialbehaviorsbyincreasingtheiragenticfeelingsofself-efficacyandtheircommunalfeelingsofsocialworth.AgenticMechanism:Self-EfficacyFromanagenticperspective,gratitudeexpressionsmayincreaseprosocialbehaviorbyenablinghelperstoexperiencegreaterself-efficacy.Self-efficacyisthefeelingofbeingcapableandcompe-tenttoacteffectivelytoorchestrateanoutcome(Bandura,1977).Psychologistsagreethatthisdesiretofeelcapableandcompetentisabasichumanmotivation(White,1959;Ryan&Deci,2000).Extensiveresearchhasshownthatwhenindividualsfeeleffica-ciousinanactivity,theyaremorewillingtoinvesttimeandenergyinitbecausetheybelievethattheireffortscanleadtosuccess(Bandura,1977;Ryan&Deci,2000).Individualsoftenwithholdhelpbecausetheyareuncertainaboutwhethertheywillbeabletohelpcompetentlyandeffectively(Rosen,Mickler,&Collins,1987).Becausehelpingatthewrongtimeorinthewrongwaycanharmorembarrassbeneficiaries,helpersmaybereluctanttogiveassistance.Anexpressionofgratitudefromabeneficiarycanreducethehelper’sexperienceofuncertaintyaboutbeingcapableofhelpingeffectively.Gratitudesignifiesthatabeneficiaryisconfidentinahelper’sabilitytoofferassistancesuccessfully(Penner,Dovidio,Piliavin,&Schroeder,2005).Gratitudeprovidespositivefeedbacktohelpersthattheyhavesucceeded—andcansucceed—inbenefitingrecipients,therebysatisfyinghelpers’basicmotivestofeelcapableandef-fective(McAdams&deSt.Aubin,1992;Ryan&Deci,2000;White,1959).Whenhelpersfeelefficacious,theybecomemorewillingtoprovidehelpbecausetheyfeelthattheireffortswillincreasetheiroddsofgenuinelyhelpingothers(Bandura,1977).Thus,weproposethatwhenbeneficiariesexpressgratitude,help-erswillfeelgreaterself-efficacy,whichwillmotivatethemtoengageinprosocialbehavior.CommunalMechanism:SocialWorthFromacommunalperspective,gratitudeexpressionsmayalsoincreaseprosocialbehaviorbyenablinghelperstofeelvalued.Psychologistshavearguedthatthepursuitofsocialworth—asenseofbeingvaluedbyothers—isafundamentalhumanmoti-vation(McAdams&deSt.Aubin,1992;Ryan&Deci,2000).Whenindividualsexperiencesocialworth,theyfeelthattheiractionsmatterinotherpeople’slives(Elliott,Colangelo,&Gelles,2005;Rosenberg&McCullough,1981),whichconfersasenseofbelongingness(Baumeister&Leary,1995;Keyes,1998).Socialworthfulfillsthe“desiretobeneededbyothers...oneexpressionofcommunion”(McAdams&deSt.Aubin,1992,p.1005).Whenindividualsexperiencesocialworth,theyfeelneeded,caredabout,andvaluedbyothers,whichsignifiesaninterpersonalbondorpositiverelationship(Bakan,1966;Kaplan&Kaplan,2003;Wr-zesniewski,Dutton,&Debebe,2003).Individualsoftenwithholdhelpbecausetheyareuncertainaboutwhetherbeneficiarieswillvaluetheirhelp.Becausegivinghelpcanleadbeneficiariestofeelincompetent,helpless,andpowerless,beneficiariesoftenrejecttheoffersofhelpers(Fisher,Nadler,&Whitcher-Alagna,1982),leavinghelpersfeelingspurned,angry,andreticenttoofferhelpagain(Rosenetal.,1987).Anexpressionofgratitudecanreducethehelper’sexperienceofuncertaintyaboutwhetherthehelpwillbeappreciated.Expressionsofgrati-tudesignifythatabeneficiaryvalues,needs,appreciates,andacceptsone’sassistanceratherthanrejectingordevaluingit.Gratitudeexpressionsprovideconcreteevidencethathelpers’ac-tionsmatterinthelivesofbeneficiaries,thussatisfyinghelpers’basicmotivestofeelvalued(Baumeister&Leary,1995;Elliottetal.,2005;Ryan&Deci,2000).Whenhelpersfeelvalued,theybecomemoremotivatedtohelpbecausetheyfeeltheiractionswillimprovethewell-beingofbeneficiaries(Batson,1998).Feelingvaluedencouragesprosocialbehaviorbyreducingthehelper’suncertaintyaboutwhetherbeneficiarieswillwelcomeassistance.Thus,weproposethatwhenbeneficiariesexpressgratitude,help-erswillfeelmoresociallyvalued,whichwill

motivatehelperstoengageinprosocialbehavi
motivatehelperstoengageinprosocialbehavior.OverviewofthePresentResearchWecomparetheseagenticandcommunalmechanismsasme-diatorsoftheeffectsofgratitudeexpressionsonhelpers’prosocialbehaviorsacrossfourexperiments.InExperiment1,weexaminewhetherself-efficacyandsocialworthmediatetheeffectsofreceivingabriefexpressionofgratitudeontheprosocialbehaviorofvoluntarilyhelpingastudentimproveajobapplicationcoverletter.InExperiment2,weinvestigateself-efficacyandsocialworthasmediatorsofaspillovereffectofgratitudeexpressionsfromonebeneficiaryonprosocialbehaviortowardadifferentbeneficiary.InExperiment3,weassesstheexternalvalidityofthemediatingmechanismsinafieldexperimentwithuniversityfund-raisers.InExperiment4,weconstructivelyreplicateoureffectsGRATITUDEEXPRESSIONSANDPROSOCIALBEHAVIORwithnewmeasuresoftheproposedmediatorsandanewmanip-ulationofgratitudeexpressions.Experiment1Weexaminedtheeffectsofgratitudeexpressionsontheproso-cialbehaviorofhelpingastudentimprovehisorherjobapplica-tioncoverletters(Grantetal.,2007).Participantseditedastu-dent’scoverletterandthenreceivedeitheraneutralorgratefulmessagefromthestudent,whosubsequentlyaskedforhelponanothercoverletter.Wetrackedtheeffectofthegratitudeexpres-siononwhetherparticipantsengagedinprosocialbehaviorbyhelpingwiththesecondletterandassessedwhetherthiseffectwasmediatedbyperceptionsofself-efficacyandsocialworth.Wealsomeasuredaffectasanalternativeexplanation.Itmaybethecasethatbycommunicatingthebenefitsofhelping,gratitudeexpres-sionsincreasehelpers’feelingsofpositiveaffectorreducetheirfeelingsofnegativeaffect.Inturn,thesechangesinhelpers’ownfeelingsmayenhancetheirprosocialbehaviorbyleadinghelperstoperceivebeneficiariesinamorefavorablelightandencouraginghelperstomaintaintheirpositivemoods(e.g.,Carlson,Charlin,&Miller,1988;Fredrickson,2001).Participants,design,andprocedures.Sixty-nineundergrad-uateandgraduatestudents(25male,44female,3.55)atauniversityintheSoutheastUnitedStatesparticipatedinthisstudy.Theirmajorswerepredominantlysocialscience(67.8%),naturalscience(27.5%),humanities(11.6%),andmathematics(2.9%).Werecruitedtheparticipantsthroughanadvertisementonauniversity-widewebsitethatoffered$10inexchangeforparticipationinanonlinestudyaboutwritingskillsandfeedback.Whentheysignedup,participantsreceivedanelectronicmessagefromtheexperimenterexplainingthattheywouldbeprovidingfeedbackonastudent’sjobapplicationcoverletteraspartofastudyrunbytheuniversity’scareercenter.Participantsreceivedacoverletterfromtheexperimenterandwereaskedtosendtheircommentsbye-maildirectlytothestudent,EricSorenson(ericsor2006@gmail.com),within24hr.Thementeralsoaskedparticipantstosendherane-mailoncetheyhadsenttheirfeedbackdirectlytoEric,justtolettheexperimenterknowtheyhadcompletedthetask.Whentheysenttheirfeedback,onthenextday,theexperimentersentthemareplyfromthestudent’se-mailwhichcontainedourmanipulation.Werandomlydividedparticipantsbetweentwoconditions:gratitude(35)andcon-trol(34).Inbothconditions,themessagefromEricSorensonaskedforhelpwithasecondcoverletter;themessagevariedonlyintheamountofgratitudeexpressedforthehelpthatparticipantshadprovidedonthefirstcoverletter.Inthecontrolcondition,participantsreceivedthefollowingmessagefromEricSorenson’se-mailaccount:“Dear[name],IjustwantedtoletyouknowthatIreceivedyourfeedbackonmycoverletter.IwaswonderingifyoucouldhelpwithasecondcoverletterIpreparedandgivemefeedbackonit.Thecoverletterisattached.Canyousendmesomecommentsinthenext3days?”Inthegratitudecondition,themessageread:“Dear[name],IjustwantedtoletyouknowthatIreceivedyourfeedbackonmycoverletter.Thankyousomuch!Iamreallygrateful.IwaswonderingifyoucouldhelpwithasecondcoverletterIpreparedandgivemefeedbackonit.Thecoverletterisattached.Canyousendmesomecommentsinthenext3days?”Thus,thetwomessageswereidenticalexceptfortheadditionofagratitudeexpressioninthelatter.Theinitialinstructionsaskedparticipantstoe-mailtheexperimenteraftersendingtheirfeedbacktoEricSorenson.Afterreceivingthise-mailandsendingpartici-pantsoneofthetwomessagesfromEricwithourgratitudemanipulation,theexperimentersentparticipantsalinktoanonlinequestionnairethatcontainedourmeasuresofself-efficacy,socialworth,positiveandnegativeaffect,andamanipulationcheck.Afterparticipantscompletedthefinalquestionnaire,theexperi-mentersentinstructionsforobtainingthe$10thatparticipantshadearned.Wemeasuredobjectiveprosocialbehaviorbytrackingwhetherparticipantsprovidedhelponthesecondcoverletterinthefollowing3days.Unlessotherwiseindicated,allitemshada7-pointLikert-typeresponsescaleanchoredat1disagreestronglyagreestronglyProsocialbehavior.Weassessedprosocialbehaviorwithadichotomousmeasureofwhetherparticipantsvoluntarilyprovidedhelponthesecondcoverletter.Weassessedself-efficacywithathree-itemscaleadaptedfromBandura(1990),whichaskedparticipantstoindicatetheextenttowhichtheyfeltcapable,competent,andabletohelpinthisspecifictask(Socialworth.Weassessedtheextenttowhichparticipantsfeltvaluedwithathree-itemscaleadaptedfrommeasuresbyKeyes(1998)andGra

nt(2008),whichaskedparticipantstoindicat
nt(2008),whichaskedparticipantstoindicatetheextenttowhichtheyfeltvaluedasapersonbythestudent,feltappreciatedasanindividualbythestudent,andfeltthattheyhadmadeapositivedifferenceinthestudent’slife(Positiveandnegativeaffect.Participantscompletedthe20-itemstateversionofthePositiveandNegativeAffectSchedule(Watson,Clark,&Tellegen,1988),whichhas10itemseachforpositiveaffect(e.g.,enthusiastic,inspired;.95)andnegativeaffect(e.g.,upset,distressed;Manipulationcheck.Toensurethatourgratitudemanipula-tionwaseffective,weaskedparticipantstoindicatetheextenttowhichthestudent’se-mailmessageexpressedgratitudeandthanks.95).Frombothconceptualandempiricalperspectives,itisimportanttoaddresshowthismanipulationcheckisdistinctfromsocialworth.Conceptually,thegratitudemanipulationcheckas-sessesthehelper’sperceptionthatthebeneficiary’sspecificcom-municationexpressedthanks.Themeasureofsocialworth,ontheotherhand,assessesthehelper’smoregeneralfeelingofbeingvaluedasapersonbythebeneficiary.Empirically,thetwovari-ablessharedonly15.8%oftheirvariance(.01).Weconductedbothexploratoryandconfirmatoryfactoranalysestoassesswhethertheyloadedondistinctfactors.Fortheexploratoryfactoranalysis,weusedprincipalaxisfactoringandmaximumlikelihoodestimationprocedureswithobliquerotation.Theanal-ysisreturnedtheexpectedtwo-factorsolution(eigenvaluesand1.28):Thethreesocialworthitemsloadedstronglyonthefirstfactor(.90,.89,.58),withverylowcross-loadingsonthesecondfactor(.06,.10,.07),andthetwomanipulationcheckitemsloadedstronglyonthesecondfactor(.91,.99),withverylowcross-loadingsonthefirstfactor(.01,.02).Intheconfirmatoryfactoranalysis,weusedEQSsoftwareVersion6.1withmaximumGRANTANDGINOlikelihoodestimationprocedures(e.g.,Bentler&Dudgeon,1996;Kline,1998).Thetwo-factormodeldisplayedexcellentfitaccord-ingtoHuandBentler’s(1999)criteria,comparativefitindex(CFI).98,standardizedroot-mean-squareresidual(SRMR).045,whereastheone-factormodeldisplayedverypoorfit,106.43,CFI.58,SRMRTakentogether,theseresultssuggestthatthegratitudemanipula-tioncheckwasindeedconceptuallyandempiricallydistinctfromthesocialworthmeasure.ResultsandDiscussionMeansandstandarddeviationsbyconditionappearinTable1.Insupportofthevalidityofourmanipulation,anindependent-testshowedthatparticipantsinthegratitudeconditionperceivedEric’se-mailmessageasexpressingmoregratitude0.66)thandidparticipantsinthecontrolcondition(1.67.Insupportofourhypothesisthatgratitudeexpressionswouldincreaseprosocialbehavior,thepercentageofparticipantsvolun-tarilyprovidinghelpbyeditingthesecondletterwaslargerinthegratitudecondition(23/35,66%)thanintheneutralcondition(11/34,32%),.01.Independent-testsshowedthatparticipantsinthegratitudeconditionfeltsignificantlygreaterself-efficacy(0.51)thandidparticipantsinthecontrolcondition(0.49.Inaddition,participantsinthegratitudeconditionfeltsignificantlymoresociallyvalued(0.55)thandidparticipantsinthecontrolcondition0.77.However,thegratitudemanipulationdidnotinfluencepositiveornegativeaffect.Participantsinthegratitudeconditiondidnotdiffersignif-icantlyinpositiveaffect(1.34)fromthoseinthecontrolcondition(Participantsinthegratitudeconditionalsodidnotdiffersignifi-cantlyinnegativeaffect(0.80)fromthoseinthecontrolcondition(Toexaminewhetherself-efficacyorsocialworthmediatedtheeffectofgratitudeonprosocialbehavior,wefollowedthestepsrecommendedbyBaronandKenny(1986).Thefirstandsecondcriteriaspecifythattheindependentvariableshouldsignificantlyaffectthedependentvariableandthemediators.Theprioranalysesshowedthatthesetwocriteriaweremet,asthegratitudemanipu-lationhadasignificanteffectonthedependentvariableofproso-cialbehaviorandthemediatorsofself-efficacyandsocialworth.Toassessthethirdandfourthcriteria,weconductedahierarchicalordinaryleast-squares(OLS)regressionanalysispredictingproso-cialbehaviorfromtheindependentvariableofgratitudecondition(Step1)andthemediatorsofself-efficacyandsocialworth(Step2).Thethirdcriterionspecifiesthatthemediatorshouldsignifi-cantlypredictthedependentvariablewhilecontrollingfortheindependentvariable.Theresultsmetthiscriterionforsocialworth:Withthegratitudemanipulationcontrolledfor,socialworthsignificantlypredictedhigherprosocialbehavior(.05.Includingsocialworthincreasedvarianceexplainedsignificantlyby9%from.11to.01.However,theresultsdidnotmeetthiscriterionforself-efficacy,whichdidnotpredicthigherprosocialbehavior(.97.Thus,self-efficacydidnotmediatetheeffectofgratitudeonprosocialbehavior.Tocompletethetestofmediationforsocialworth,thefourthcriterionholdsthattheeffectoftheindependentvariableonthedependentvariableshoulddecreaseaftercontrollingforthemedi-ators.Aftercontrollingforsocialworth,theeffectofthegratitudemanipulationonprosocialbehaviordecreasedfrom.01to.07.Totestwhetherthesizeoftheindirecteffectofthegratitudemanipulationonprosocialbehaviorthroughsocialworthdifferedsignificantlyfromzero,weusedabootstrapproceduretoconstructbias-correctedconfidenceintervalsbasedon1,000randomsampleswithreplacementfromthefullsample,asrecommendedbymeth-

odologistsandstatisticians(MacKinnon,Fai
odologistsandstatisticians(MacKinnon,Fairchild,&Fritz,2007;Shrout&Bolger,2002).Thesizeoftheindirecteffectfromthefullsamplewas.12,andthe95%confidenceintervalexcludedzero,95%CI[0.04,0.26].Thus,socialworthmediatedtheeffectofgratitudeonprosocialbehavior.Theseresultsprovideevidencethatgratitudeexpressionsin-creaseprosocialbehaviorthroughthecommunalmechanismofenablinghelperstofeelmoresociallyvalued,ratherthanthroughtheagenticmechanismofenablinghelperstofeelmoreefficaciousorthroughpositiveornegativeaffect.However,inthisstudy,wefocusedonprosocialbehaviordirectedtowardthesamebenefi-ciarywhoexpressedgratitude.Amorestringenttestofourhy-pothesisrequiresexaminingwhetherself-efficacyandsocialworthmediatetheeffectofgratitudeexpressionsonprosocialbehaviordirectedtowardathirdparty.Doesanexpressionofgratitudefromonebeneficiarycausehelperstofeelmoreefficaciousorsociallyvaluedandthusmotivatethemtoprovideadditionalhelpbeyondthisdyadicrelationshiptoadifferentbeneficiary?Experiment2Inoursecondstudy,weexaminewhethersocialworthmediatesthespillovereffectsofonebeneficiary’sgratitudeexpressiononhelpers’prosocialbehaviortowardanotherbeneficiary.Tostrengthencausalinferencesabouttheprimacyofself-efficacyand/orsocialworthdrivingprosocialbehavior,wemeasureself-efficacyandsocialworthbeforeprovidingparticipantswiththeopportunitytoengageinprosocialbehavior.Wealsomeasuredfeelingsofpositiveandnegativeaffect.Furthermore,tocaptureamorespecificaffectivestate,wealsoexploredthepossibilitythatgratitudeexpressionsincreaseprosocialbehaviorbyenhancinghelpers’feelingsofempathytowardbeneficiaries(Batson,1998).Participants,design,andprocedures.Fifty-sevenunder-graduateandgraduatestudents(28male,29female,3.47)atlocaluniversitiesintheNortheastUnitedStatesparticipatedinthisstudy.Theirmajorswerepredominantlyinmathematics,engineering,information,andcomputerscience(43.9%);socialscience(31.6%),naturalscience(14%);andtheInExperiments1and2,becauseourdependentvariablewasbinary,wereranthemediationanalyseswithMacKinnonandDwyer’s(1993)logisticregressionmethodandfoundthesamepatternofresults.Wereportthemoretraditionalapproachintheinterestofparsimony.GRATITUDEEXPRESSIONSANDPROSOCIALBEHAVIORhumanities(8.8%).Weusedthesamerecruitingmethodasinthepreviousstudy,advertisingtheexperimentasanonlinestudyofwritingandfeedbackthatwouldpay$10.Uponsigningup,par-ticipantsreceivedanelectronicmessagefromtheexperimenteraskingthemtoreadastudent’sjobapplicationcoverletterandsendthecommentsbye-maildirectlytothestudent,EricSoren-son,within24hr.Whenparticipantssubmittedtheirfeedback,wesentthemareplyfromthestudent’se-mailaccountcontainingourmanipulation.Asinthepreviousexperiment,werandomlydividedparticipantsbetweentwoconditions:gratitude(29)andcon-trol(Inthecontrolcondition,participantsreceivedthefollowingmessagefromEricSorenson’se-mailaccount:“Dear[name],IjustwantedtoletyouknowthatIreceivedyourfeedbackonmycoverletter.”Inthegratitudecondition,themessageread:“Dear[name],IjustwantedtoletyouknowthatIreceivedyourfeedbackonmycoverletter.Thankyousomuch!Iamreallygrateful.”Thenextday,theexperimentersentallparticipantsalinktoanonlinequestionnaire,whichincludedameasureofself-efficacy,socialworth,andamanipulationcheck.Afterparticipantscompletedthefinalquestionnaire,theexperimentersentparticipantsinstructionsforobtainingthe$10theyhadearned.Onedaylater,theexperimentersentallparticipantsane-mailmessagefromtheaccountofadifferentstudent,StevenRogoff.Themessageread:“Hi[name],IunderstandthatyouparticipatedinaCareerCenterstudytohelpstudentsimprovetheirjobappli-cationcoverletters.Iwaswonderingifyoucouldgivemefeed-backonacoverletterIprepared.Thecoverletterisattached.Wouldyoubewillingtohelpmebysendingmesomecommentsinthenexttwodays?”Wetrackedparticipants’objectivelevelsofprosocialbehaviorbyassessingwhethertheyhelpedStevenRo-goffbysendinghimfeedback.Unlessotherwiseindicated,allitemshada7-pointscaleanchoredat1disagreestronglyand7agreestronglyProsocialbehavior.Weassessedprosocialbehaviorwithabinarymeasure,codingwhetherparticipantsvoluntarilyprovidedhelptoStevenRogoffonhiscoverletter.Mediatingmechanisms:Self-efficacyandsocialworth.measuredself-efficacywiththesix-itemscalefromRyan,Koest-ner,andDeci(1991),whichincludesitemssuchas,“IthinkIwasprettygoodatthistask”and“Iwasprettyskilledatthistask”(.89).Wemeasuredsocialworthwiththesamescaleasinthepreviousexperiment(Alternativeexplanations:Affectivestates.Asalternativeex-planations,wemeasuredaffectatthesametimeastheself-efficacyandsocialworthscales.WemeasuredpositiveandnegativeaffectusingthePANAS(Watsonetal.1988),andboththepositiveaffect.96)andnegativeaffect(.93)scalesshowedhighinternalconsistency.Wemeasuredparticipants’feelingsofempathytowardthestudentwithBatson’s(1987)six-itemadjectivescale(Manipulationcheck.ParticipantsratedthedegreetowhichEricSorenson’smessageexpressedgratitudeandthanks(.96).Themanipulationcheckandthesocialworthmeasureshared36.6%oftheirvariance(.001).Inaconfirmatoryfactoranalysis,thetwo-factormodelachie

vedgoodfit,.98,SRMR.086,whereastheone-fa
vedgoodfit,.98,SRMR.086,whereastheone-factormodeldidnot,101.20,CFI.56,SRMRResultsandDiscussionMeansandstandarddeviationsbyconditionaredisplayedinTable2.Ourgratitudemanipulationwaseffective:Participantsinthegrat-itudeconditionratedEric’smessageasexpressingmoregratitude1.06)thandidparticipantsinthecontrolcondition2.86.Consistentwithourpredictionthatgratitudeexpressionswouldincreaseproso-cialbehaviortowardathirdparty,thepercentageofparticipantswhovoluntarilyprovidedhelptothenewstudent,Steven,wassignificantlyTable1Experiment1MeansbyConditionSelf-efficacySocialworthPositiveaffectNegativeaffectMSDMSDMSDMSDMSDGratitude666.030.516.050.554.061.341.450.806.270.66Control325.651.005.440.993.591.151.550.764.181.6835;controlTable2Experiment2MeansbyConditionSelf-efficacySocialworthPositiveaffectaffectEmpathyMSDMSDMSDMSDMSDMSDGratitude555.650.765.740.473.401.421.220.483.521.406.221.06Control255.050.834.630.603.571.401.390.343.551.213.021.2229;controlGRANTANDGINOhigherinthegratitudecondition(16/29,55%)thanintheneutralcondition(7/28,25%),.05.Participantsinthegratitudeconditionfeltsignificantlymoreefficacious(0.76)thandidparticipantsinthecontrolcondition(0.77.Participantsinthegratitudeconditionalsofeltsignificantlymoresociallyvalued(0.47)thandidparticipantsinthecontrolcondition(2.07.Onceagain,thegratitudemanipulationdidnotinfluencepositiveornegativeaffect.Participantsinthegratitudeconditiondidnotdiffersignificantlyinpositiveaffect(1.42)fromparticipantsinthecontrolcondition(.Participantsinthegratitudeconditionalsodidnotdiffersignificantlyinnegativeaffect(0.48)fromparticipantsinthecontrolcondition(.Finally,thegratitudemanipulationalsodidnotinfluencefeelingsofempathy;participantsinthegratitudecondition(1.40)andparticipantsinthecontrolcondition(1.21)reportednearlyidenticallevelsofempathy,Totestwhetherself-efficacyandsocialworthmediatedtheeffectofgratitudeonprosocialbehavior,weusedthesamepro-ceduresasinthepreviousexperiment.AhierarchicalOLSregres-sionanalysisshowedthatwhilecontrollingforthegratitudema-nipulation,self-efficacydidnotpredicthigherprosocialbehavior.37,butsocialworthdid(.01.Includingsocialworthincreasedvarianceexplainedsignificantlyby10%from.10to(1,54).01.Aftercontrollingforsocialworth,theeffectofthegratitudemanipulationonprosocialbehaviorde-creasedfrom.05to.80.Abootstrapanalysisshowedthatthe95%bias-correctedconfidenceintervalsforthesizeoftheindirecteffect(.33)excludedzero,95%CI[0.08,0.58].Thus,socialworth—butnotself-efficacy,positiveaffect,neg-ativeaffect,orempathy—mediatedtheeffectofanexpressionofgratitudefromonebeneficiaryonprosocialbehaviordirectedtowardadifferentbeneficiary.Thefactthattheopportunityforprosocialbehavioroccurred24hrafterthemeasureofsocialworthstrengthenscausalinferencesabouttheroleofsocialworthindrivingprosocialbehavior.Havingestablishedtheinternalvalidityofourresults,followingtheguidelinesoffull-cyclepsychologicalresearch(Cialdini,1980),weconductedathirdexperimentinthefieldtoexaminetheexternalvalidityofourfindings.Experiment3Thisstudyfocusesonfundraisersresponsibleforsolicitingalumnidonationstoauniversity.Becausefundraisingisoftenathanklessjobthattriggersrudefeedbackandregularrejections(Seligman&Schulman,1986),gratitudeexpressionsinthisrealmarelikelytobememorableandsalient.Weexaminetheeffectsofagratitudeexpressionfromanannualgivingdirectoronfundrais-ers’prosocialbehaviorsinraisingmoneytobenefittheuniversity,asmediatedbytheirperceptionsofself-efficacyandsocialworth.Participants,design,andprocedures.Forty-onefundraisers(10male,31female,4.5months,5.17)atapublicU.S.universityparticipatedinthisstudy.Werandomlydividedfundraisersbetweentwoconditionsindifferentshiftstopreventtreatmentdiffusionordiscussionsabouttheintervention.Inthecontrolcondition(21),fundraisersreceivednotreatment.Inthegratitudecondition(20),adirectorofannualgivingvisitedtheorganizationtothankthefundraisersfortheirwork.Sheexplainedtothefundraisers,“Iamverygratefulforyourhardwork.Wesincerelyappreciateyourcontributionstotheuniver-sity.”Bothgroupsoffundraisersreceiveddailyfeedbackontheireffectiveness;theonlydifferencebetweenthetwogroupswaswhethertheyreceivedanexpressionofgratitudefromthedirector.Prosocialbehavior.Weassessedprosocialbehaviorwithanobjective,automaticallyrecordedmeasureofthenumberofvol-untarycallsthateachfundraisermadeduringtheweekbeforeandtheweekaftertheintervention.Thisisanappropriateindicatorofprosocialbehaviorbecausethefundraisersreceivedafixedsalaryandwerenotrewardedforeffort;anyvoluntarycallsthatfund-raisersmadewerepurelytohelptheuniversity.Mediatingmechanisms:Self-efficacyandsocialworth.ingtheweekaftertheintervention,wemeasuredfundraisers’perceptionsofself-efficacyandsocialworth.Wemeasuredself-efficacywithSpreitzer’s(1995)scale,askingthefundraiserstoreporttheextenttowhichtheyfeltcapable,confident,andabletosucceedinthetaskofraisingmoney(.83).Wemeasuredsocialworthwiththesamescaleasintheprevioustwostudies,askingthefundraiserstoreporttheextenttowhichtheyfelt,asemploy

ees,valuedandappreciatedbymanagers(Resul
ees,valuedandappreciatedbymanagers(ResultsandDiscussionMeansandstandarddeviationsbyconditionappearinTable3.Arepeated-measuresanalysisofvarianceindicatedasignificanttime–conditioninteractiononprosocialbehavior,(1,39).10.Paired-samplestestswithineachconditionovertimeindicatedthatthegratitudeconditionincreasedinproso-cialbehavior,0.75,whereasthecontrolTable3Experiment3MeansbyConditionPretestprosocialbehaviorPosttestprosocialbehaviorPosttestself-efficacyPosttestsocialworthMSDMSDMSDMSDGratitude41.4023.2762.6032.456.180.525.680.83Control39.7625.2141.3822.055.800.635.041.0220;controlGRATITUDEEXPRESSIONSANDPROSOCIALBEHAVIORconditiondidnot,0.44.Inaddition,comparedwithfund-raisersinthecontrolcondition,fundraisersinthegratitudecondi-tionreportedstrongerperceptionsofself-efficacy,0.66,andsocialworth,Totestwhetherself-efficacyandsocialworthmediatedtheeffectofgratitudeonchangesinprosocialbehavior,wefollowedJudd,Kenny,andMcClelland’s(2001)guidelinesfortestingme-diationwithOLSregressioninwithin-subjectdesigns.Whilecontrollingforthegratitudemanipulation,self-efficacydidnotsignificantlypredictincreasesinprosocialbehavior(.95,butsocialworthdid(.05.Aftercontrollingforsocialworth,theeffectofthegratitudemanipulationdecreasedfrom.05to.20.Includingsocialworthincreasedvarianceexplainedsignificantlyby10%from(1,36).05.Abootstrapanalysisindicatedthatthe95%bias-correctedconfidenceintervalsforthesizeoftheindirecteffect(6.15)excludedzero,95%CI[0.29,20.79].Accordingly,thegratitudeexpressionincreasedtheproso-cialbehaviorofmakingmorecallstohelptheuniversitybystrengtheningfundraisers’feelingsofsocialworth,notbyenhanc-ingtheirfeelingsofself-efficacy.Experiment4ConsistentwiththefindingsinExperiments1and2,ourthirdstudyprovidesfurtherevidencethatgratitudeexpressionsincreaseprosocialbehaviorbyenablingindividualstofeelsociallyvalued.OnestrengthofExperiment3isthatwemanipulatedgratitudeinaface-to-faceinteraction,ratherthaninanelectroniccommuni-cation,asinExperiments1and2.Atthesametime,thismanip-ulationpresentsapotentialconfound;itmaybetheinterpersonalinteractionitself,ratherthanthegratefulcontentoftheinteraction,whichcausestheobservedeffects.Toaddressthisissue,weconductedafourthexperimentinwhichwevarythegratefulcontentofaninterpersonalinteraction.Inaddition,tofurtherstrengthensupportforthemediatingroleofthecommunalmech-anism,weuseadifferentmeasuretoassesssocialworth.Participants,design,andprocedures.Seventy-nineunder-graduatesataMidwestuniversity(25male,54female)partici-patedinthisstudyforcoursecreditintheirintroductorypsychol-ogyclasses.Participantsarrivedindividuallyatthelaboratory,andweinformedthemthatwewereworkingwiththeuniversity’scareercentertounderstandpeerfeedbackprocesses.Weaskedthemtohelpastudent,EricSorenson,byeditinghisjobapplica-tioncoverletters.TheyeditedaninitialcoverletterforEric,andthenaconfederateactingasEricarrived,purportedlytodeliverformstotheexperimenter.TheconfederateintroducedhimselfasEricSorensontoeachparticipantandstruckupaconversationabouttheweather.Inthegratitudecondition,Ericsaid“Thankyouforyourfeedback,”andinthecontrolcondition,hedidnot.TheexperimenterthendismissedEricandgaveeachparticipantasecondcoverlettertoedit,informingthemthattheycouldstopwhenevertheywerefinished.Afterparticipantsannouncedtaskcompletion,theyfilledoutabriefsurvey.Unlessotherwiseindicated,allitemshada7-pointscaleanchoredat1disagreestronglyand7agreestronglyProsocialbehavior.Theexperimenterrecordedthenumberofminutesandsecondsthatparticipantsvoluntarilyspenteditingthesecondcoverlettertohelpthestudent.Mediatingmechanisms:self-efficacyandsocialworth.measuredself-efficacywiththesix-itemscalefromRyanetal..73).WemeasuredsocialworthwithfouritemsadaptedfromKeyes’s(1998)socialintegrationscale,whichwemodifiedtoclearlytapintothecommunalelementofsocialworth.Weaskedparticipantstoreporttheextenttowhichtheyfeltclosetothestudent,feltthatthestudentvaluesthemasaperson,feltstrongtrustfromthestudent,andfeltimportanttothestudent(Manipulationcheck.Participantsratedthedegreetowhichthestudent’smessageexpressedgratitudeandthanks(Themanipulationcheckandthesocialworthmeasureshared12.96%oftheirvariance(.01),andfactoranalysesonceagainsupportedthedistinctivenessofthetwomeasures.ResultsandDiscussionMeansandstandarddeviationsbyconditionaredisplayedinTable4.Participantsinthegratitudeconditionratedtheinteractionwiththestudentasexpressingmoregratitude(1.09)thandidparticipantsinthecontrolcondition(0.45.Thegratitudemanipulationincreasedprosocialbehavior:Participantsinthegrat-itudeconditionspentsignificantlymoretimehelpingthestudentonthesecondcoverletter(6.02)thandidparticipantsinthecontrolcondition(0.49.Therewerenosignificantdifferencesbetweenconditionsinself-efficacy,However,participantsinthegratitudeconditionfeltsignificantlymoresociallyvalued(0.81)thandidparticipantsinthecontrolcondition(Table4Experiment4MeansbyConditionProsocialbehaviorSelf-efficacySocialworthManipulationcheckMSDMSDMSDMSDGratitude22.836.024.790.823.840.814.891

.09Control19.836.344.590.763.330.894.401
.09Control19.836.344.590.763.330.894.401.1040;controlGRANTANDGINOTotestwhethersocialworthmediatedtheeffectofgratitudeonprosocialbehavior,weusedthesameproceduresasinthepreviousexperiments.AhierarchicalOLSregressionanalysisshowedthatwhilethegratitudemanipulationwascontrolledfor,socialworthpredictedhigherprosocialbehavior(Includingsocialworthincreasedvarianceexplainedsignificantlyby6%from.06to(1,76).02,andcontrollingforsocialworthreducedtheeffectofthegratitudemanipulationfrom.03to.16.Inabootstrapanalysis,the95%bias-correctedconfidenceintervalsfortheindirecteffect(1.00)ex-cludedzero,95%CI[.16,2.27].GeneralDiscussionAcrossfourexperiments,perceptionsofsocialworthmediatedtheeffectsofbeneficiaries’gratitudeexpressionsonhelpers’prosocialbehavior.Theexperimentsshowedtheseeffectsforprosocialbehaviordirectedtowardthesamebeneficiary(Experi-ments1and4),adifferentbeneficiary(Experiment2),andauniversity(Experiment3).Inaddition,gratitudeexpressionsin-creasedboththeinitiationandmaintenanceofprosocialbehavior,motivatingparticipantstoprovidehelpasecondtimewhenasked(Experiments1and2)andtopersistlongerintheircurrenthelpingactivitieswithoutbeingasked(Experiments3and4).Thesecon-vergenteffectsonobjectiveprosocialbehaviorsareimportant,giventhatthemajorityofgratituderesearchhasfocusedonpsy-chologicaleffects,leadingpsychologiststocallformoreresearchonitsbehavioraleffects(Tsang,2006).Thepresentresearchanswersthesecallsbydemonstratingthecausalimpactofgratitudeonhelpers’concrete,observableprosocialbehaviors.Theconsis-tentlystrongeffectsofrelativelysmallgratitudemanipulationsarenoteworthy(Prentice&Miller,1992).Inourfirsttwoexperiments,amereexpressionofthanksmorethandoubledthelikelihoodthathelperswouldprovideassistanceagain(from25%to55%andfrom32%to66%).Inourthirdexperiment,gratitudeproducedmorethan50%increasesinthenumberofcallsthattheaveragefundraisermadeinasingleweek.Inourfourthexperiment,asinglegratitudeexpressionyieldedanincreaseof15%intheaverageamountoftimespenthelping.Ourresultsrevealthatgratitudeexpressionsincreaseprosocialbehaviorthroughcommunalmechanismsratherthanagenticmechanisms.Inourexperiments,althoughgratitudeexpressionsincreasedbothfeelingsofself-efficacyandsocialworth,onlysocialworthexplainedtheeffectsofgratitudeexpressionsonprosocialbehavior.Thesefindingssuggestthatwhenhelpersarethankedfortheirefforts,theresultingsenseofbeingsociallyvalued,morethanthefeelingsofcompetencetheyexperience,arecriticalinencouragingthemtoprovidemorehelpinthefuture.Together,ourexperimentssupportacommunalperspectiveonwhygratitudeexpressionsincreaseprosocialbehavior.Further-more,wedidnotfindanyevidencethattheseeffectsoperatethroughinfluencinghelpers’positiveornegativeaffect.Additionally,oursecondandthirdexperimentsshowedthatgratitudeexpressionshadspillovereffectsonprosocialbehaviorstowardotherbeneficiaries.Thisfindingsuggeststhatbeneficia-ries’gratitudeexpressionsmayprompthelperstoconsiderassist-ingabroadergroupofbeneficiaries(Batson,1998).Bysignalingtohelpersthattheireffortsarevalued,gratitudeexpressionsmaybesufficientlypotenttoinfluencehelpers’effortsonbehalfofthelargergroupstowhichbeneficiariesbelong.Thisevidencedove-tailswithresearchdemonstratingspillovereffectsofgratefulfeel-ings,whichencourageindividualstoengageinupstreamreciproc-ityby“payingforward”helptootherbeneficiaries(Bartlett&DeSteno,2006).Ourfindingscomplementthislineofresearchbydemonstratingthatexpressionsofgratitude,notonlyexperiencesofgratitude,havespillovereffectsonprosocialbehavior.Assuch,ourresearchsuggeststhatgratitudeexpressionsmayhaveimpor-tanttheoreticalandpracticalimplicationsforencouragingproso-cialbehaviorsthatpromotecooperation(seealsoNowak&Roch,Thesecontributionsmustbequalifiedinlightofseveralimpor-tantlimitationsofourresearch.First,therewereinconsistenciesacrossstudiesintheeffectsofourgratitudemanipulationsonself-efficacy.InExperiment3,themanipulationhadequallystrongeffectsonself-efficacyandsocialworth.However,inExperiments1and2,themanipulationhadstrongereffectsonsocialworththanonself-efficacy.AndinExperiment4,themanipulationaffectedsocialworthbutdidnotinfluenceself-efficacy.Ononehand,therelativelyweakeffectsonself-efficacyinthreeofourfourexper-imentsmaybeanartifactofthegratitudemanipulationsthatpreventedusfromdetectingamediatingroleofself-efficacy.Ontheotherhand,itispossiblethatgratitudeexpressionsaresimplylesslikelytoinfluenceself-efficacythansocialworth.Futureresearchisnecessarytoresolvethesecompetinginterpretations.Furthermore,beyondself-efficacy,socialworth,andaffect,theremaybeadditionalmechanismsthroughwhichgratitudeexpres-sionsmotivateprosocialbehavior.Forexample,itwillbeworth-whiletoexaminewhethergratitudeexpressionspromoteprosocialbehaviorbyincreasinghelpers’feelingsofself-esteemandtheirperceptionsofreciprocity,strengtheningtheirconfidencethattheireffortswillberepaid(McCulloughetal.,2001).Infact,inlightoftheperspectivefromsociometertheorythatself-esteemcanserveasanindicatorofone’ssocialworth(Leary&Baumeist

er,2000),itmaybethecasethatgratitudeexpr
er,2000),itmaybethecasethatgratitudeexpressionsstrengthenfeelingsofsocialworthbybuildingself-esteem.Inthefuture,researcherscouldinvestigatethispossibilitythatself-esteemisamicromedia-toroftheeffectsofgratitudeexpressionsonsocialworth.Second,wedidnotinvestigatemoderatorsoftheeffectsofgratitudeexpressionsonprosocialbehavior.Forinstance,researchhasshownthatgratitudeexpressionsaremorelikelytomotivatehelpingamongindividualswithahighneedforapprovalthanamongthosewithalowneedforapproval(Deutsch&Lamberti,1986).Third,ourmanipulationscreatedbroaddifferencesinthelevelofgratitudeexpressionswithoutattendingtoimportantvari-ationsintheiremotionalandlinguisticcontent.Werecommendthatresearchersstudyhowemotionscommunicatedinfacialandothernonverbalcuesinfluencehelpers’reactionstogratitudeex-pressions,aswellassubtlelinguisticvariationsinthewordingofgratitudeexpressions,suchasinwhetherthecontentofthemes-sagedirectlyinvolvesgratefulemotionsorwhetheritsimplyconveysthatonehasbeenhelped.Finally,ourresearchfocusedonthebenefitsofgratitudewithlittleattentiontoitscosts.Researcherscouldexplorewhetherandwhengratitudeexpressionsviolatehumilitynorms,causinghelp-erstofeeluncomfortable,orburdenedandpressured(Graziano,Habashi,Sheese,&Tobin,2007),toassisttobeneficiariesbeyondalevelthattheyfindreasonableoruseful.Fornow,ourresearchGRATITUDEEXPRESSIONSANDPROSOCIALBEHAVIORtakesameaningfulsteptowardexplainingwhyalittlethankscangoalongway:Smallexpressionsofgratitudecanmotivateproso-cialbehaviorsbyleadinghelperstofeelsociallyvalued.Bakan,D.(1966).Thedualityofhumanexistence:Isolationandcommun-ioninWesternman.Boston,MA:BeaconPress.Bandura,A.(1977).Self-efficacy:TowardaunifyingtheoryofbehavioralPsychologicalReview,84,Bandura,A.(1990).Multidimensionalscalesofperceivedself-efficacy.Stanford,CA:StanfordUniversity.Baron,R.M.,&Kenny,D.A.(1986).Themoderator–mediatorvariabledistinctioninsocialpsychologicalresearch:Conceptual,strategic,andstatisticalconsiderations.JournalofPersonalityandSocialPsychology,Bartlett,M.Y.,&DeSteno,D.(2006).Gratitudeandprosocialbehavior:Helpingwhenitcostsyou.PsychologicalScience,17,319–325.Batson,C.D.(1998).Altruismandprosocialbehavior.InD.T.Gilbert,S.T.Fiske,&G.Lindzey(Eds.),Thehandbookofsocialpsychology(Vol.2,4thed.,pp.282–316).NewYork,NY:McGraw-Hill.Baumeister,R.F.,&Leary,M.R.(1995).Theneedtobelong:Desireforinterpersonalattachmentsasafundamentalhumanmotivation.logicalBulletin,117,Bentler,P.M.,&Dudgeon,P.(1996).Covariancestructureanalysis:Statisticalpractice,theory,anddirections.AnnualReviewofPsychology,Blau,P.(1964).Exchangeandpowerinsociallife.NewYork,NY:Wiley.Carey,J.R.,Clicque,S.H.,Leighton,B.A.,&Milton,F.(1976).Atestofpositivereinforcementofcustomers.JournalofMarketing,40,98–Carlson,M.,Charlin,V.,&Miller,N.(1988).Positivemoodandhelpingbehavior:Atestofsixhypotheses.JournalofPersonalityandSocialPsychology,55,Cialdini,R.B.(1980).Full-cyclesocialpsychology.AppliedSocialPsy-chologyAnnual,1,21–47.Clark,H.B.,Northrop,J.T.,&Barkshire,C.T.(1988).Theeffectsofcontingentthank-younotesoncasemanagers’visitingresidentialcli-EducationandTreatmentofChildren,11,Clark,R.D.(1975).Theeffectsofreinforcement,punishment,anddepen-dencyonhelpingbehavior.BulletinofPersonalityandSocialPsychol-ogy,1,596–599.Deutsch,F.M.,&Lamberti,D.M.(1986).DoessocialapprovalincreasePersonalityandSocialPsychologyBulletin,12,149–157.Elliott,G.C.,Colangelo,M.F.,&Gelles,R.J.(2005).Matteringandsuicideideation:Establishingandelaboratingarelationship.PsychologyQuarterly,68,Emmons,R.A.,&McCullough,M.E.(2003).Countingblessingsversusburdens:Anexperimentalinvestigationofgratitudeandsubjectivewell-beingindailylife.JournalofPersonalityandSocialPsychology,84,Fisher,J.D.,Nadler,A.,&Whitcher-Alagna,S.(1982).Recipientreac-tionstoaid.PsychologicalBulletin,91,Fiske,S.T.,Cuddy,A.J.C.,&Glick,P.(2007).Universaldimensionsofsocialcognition:Warmthandcompetence.TrendsinCognitiveSci-ences,11,77–83.Fredrickson,B.L.(2001).Theroleofpositiveemotionsinpositivepsychology:Thebroaden-and-buildtheoryofpositiveemotions.icanPsychologist,56,218–226.Goldman,M.,Seever,M.,&Seever,M.(1982).Sociallabelingandthefoot-in-the-dooreffect.JournalofSocialPsychology,117,19–23.Grant,A.M.(2008).Thesignificanceoftasksignificance:Jobperfor-manceeffects,relationalmechanisms,andboundaryconditions.ofAppliedPsychology,93,108–124.Grant,A.M.,Campbell,E.M.,Chen,G.,Cottone,K.,Lapedis,D.,&Lee,K.(2007).Impactandtheartofmotivationmaintenance:Theeffectsofcontactwithbeneficiariesonpersistencebehavior.OrganizationalBe-haviorandHumanDecisionProcesses,103,53–67.Graziano,W.G.,Habashi,M.M.,Sheese,B.E.,&Tobin,R.M.(2007).Agreeableness,empathy,andhelping:Apersonsituationperspective.JournalofPersonalityandSocialPsychology,93,Hu,L.,&Bentler,P.M.(1999).Cutoffcriteriaforfitindexesincovariancestructureanalysis:Conventionalcriteriaversusnewalternatives.turalEquationModeling,6,Judd,C.M.,James-Hawkins,L.,Yzerbyt,V.,&Kashima,Y.(2005).Fundamentaldimensionsofsocialjudgment:Understandingtherela-tionsbet

weenjudgmentsofcompetenceandwarmth.Journ
weenjudgmentsofcompetenceandwarmth.JournalofPer-sonalityandSocialPsychology,89,899–913.Judd,C.M.,Kenny,D.A.,&McClelland,G.H.(2001).Estimatingandtestingmediationandmoderationinwithin-subjectdesigns.icalMethods,6,Kaplan,S.,&Kaplan,R.(2003).Health,supportiveenvironments,andthereasonablepersonmodel.AmericanJournalofPublicHealth,93,1484–Keyes,C.L.M.(1998).Socialwellbeing.SocialPsychologyQuarterly,Kline,R.B.(1998).Principlesandpracticeofstructuralequationmod-NewYork,NY:GuilfordPress.Leary,M.R.,&Baumeister,R.F.(2000).Thenatureandfunctionofself-esteem:Sociometertheory.InM.P.Zanna(Ed.),Advancesinexperimentalsocialpsychology(Vol.32,pp.1–62).SanDiego,CA:AcademicPress.Lyubomirsky,S.,Sheldon,K.M.,&Schkade,D.(2005).Pursuinghappi-ness:Thearchitectureofsustainablechange.ReviewofGeneralPsy-chology,9,MacKinnon,D.P.,&Dwyer,J.H.(1993).Estimatingmediatedeffectsinpreventionstudies.EvaluationReview,17,144–158.MacKinnon,D.P.,Fairchild,A.J.,&Fritz,M.S.(2007).MediationAnnualReviewofPsychology,58,593–614.McAdams,D.P.,&deSt.Aubin,E.(1992).Atheoryofgenerativityanditsassessmentthroughself-report,behavioralacts,andnarrativethemesinautobiography.JournalofPersonalityandSocialPsychology,62,McCullough,M.E.,Kilpatrick,S.D.,Emmons,R.A.,&Larson,D.B.(2001).Isgratitudeamoralaffect?PsychologicalBulletin,127,249–McCullough,M.E.,Tsang,J.-A.,&Emmons,R.A.(2004).Gratitudeinintermediateaffectiveterrain:Linksofgratefulmoodstoindividualdifferencesanddailyemotionalexperience.JournalofPersonalityandSocialPsychology,86,McGovern,L.P.,Ditzian,J.L.,&Taylor,S.P.(1975).Theeffectofpositivereinforcementonhelpingwithcost.PsychonomicSocietyBul-letin,5,421–423.Moss,M.K.,&Page,R.A.(1972).Reinforcementandhelpingbehavior.JournalofAppliedSocialPsychology,2,360–371.Nowak,M.A.,&Roch,S.(2007).Upstreamreciprocityandtheevolutionofgratitude.ProceedingsoftheRoyalSocietyB:BiologicalSciences,605–609.Penner,L.A.,Dovidio,J.F.,Piliavin,J.A.,&Schroeder,D.A.(2005).Prosocialbehavior:Multilevelperspectives.AnnualReviewofPsychol-ogy,56,Prentice,D.A.,&Miller,D.T.(1992).Whensmalleffectsareimpressive.PsychologicalBulletin,112,160–164.Rind,B.,&Bordia,P.(1995).Effectofserver’s“thankyou”andperson-alizationonrestauranttipping.JournalofAppliedSocialPsychology,Rosen,S.,Mickler,S.E.,&Collins,J.E.(1987).Reactionsofwould-beGRANTANDGINOhelperswhoseofferofhelpisspurned.JournalofPersonalityandSocialPsychology,53,288–297.Rosenberg,M.,&McCullough,B.C.(1981).Mattering:Inferredsignifi-canceandmentalhealthamongadolescents.ResearchinCommunityandMentalHealth,2,Ryan,R.M.,&Deci,E.L.(2000).Self-determinationtheoryandthefacilitationofintrinsicmotivation,socialdevelopment,andwell-being.AmericanPsychologist,55,68–78.Ryan,R.M.,Koestner,R.,&Deci,E.L.(1991).Variedformsofpersis-tence:Whenfree-choicebehaviorisnotintrinsicallymotivated.vationandEmotion,15,Seligman,M.E.P.,&Schulman,P.(1986).Explanatorystyleasapredictorofproductivityandquittingamonglifeinsurancesalesagents.JournalofPersonalityandSocialPsychology,50,832–838.Seligman,M.E.P.,Steen,T.A.,Park,N.,&Peterson,C.(2005).Positivepsychologyprogress:Empiricalvalidationofinterventions.Psychologist,60,410–421.Shrout,P.E.,&Bolger,N.(2002).Mediationinexperimentalandnon-experimentalstudies:Newproceduresandrecommendations.logicalMethods,7,422–445.Spreitzer,G.M.(1995).Psychologicalempowermentintheworkplace:Dimensions,measurement,andvalidation.AcademyofManagementJournal,38,Tesser,A.,Gatewood,R.,&Driver,M.(1968).SomedeterminantsofJournalofPersonalityandSocialPsychology,9,Tsang,J.-A.(2006).Gratitudeandprosocialbehaviour:Anexperimentaltestofgratitude.Cognition&Emotion,20,138–148.Watson,D.,Clark,L.A.,&Tellegen,A.(1988).Developmentandvali-dationofbriefmeasuresofpositiveandnegativeaffect:ThePANASJournalofPersonalityandSocialPsychology,54,Weiner,B.(1985).AnattributionaltheoryofachievementmotivationandPsychologicalReview,92,548–573.White,R.W.(1959).Motivationreconsidered:Theconceptofcompe-PsychologicalReview,66,Wiggins,J.S.(1979).Apsychologicaltaxonomyoftrait-descriptiveterms:Theinterpersonaldomain.JournalofPersonalityandSocialPsychol-ogy,37,395–412.Wrzesniewski,A.,Dutton,J.E.,&Debebe,G.(2003).Interpersonalsensemakingandthemeaningofwork.InB.M.Staw&R.KramerResearchinOrganizationalBehavior(Vol.25,pp.93–135).NewYork,NY:ElsevierScience.ReceivedFebruary22,2009RevisionreceivedSeptember21,2009AcceptedSeptember28,2009CorrectiontoFordandCollins(2010)Thearticle“Self-EsteemModeratesNeuroendocrineandPsychologicalResponsestoInterpersonalRejection,”byMa´ireB.FordandNancyL.Collins(JournalofPersonalityandSocialPsychology2010,Vol.98,No.3,pp.405–419),containedamisspellinginthelastnameofthefirstauthorinthebelowreference.Thecompletecorrectreferenceisbelow.Theonlineversionofthearticlehasbeencorrected.Stinson,D.A.,Logel,C.,Zanna,M.P.,Holmes,J.G.,Cameron,J.J.,Wood,J.V.,&Spencer,S.J.(2008).Thecostoflowerself-esteem:Testingaself-andsocial-bondsmodelofhealth.ofPersonalityandSocialPsychology,94,412–428.DOI:10.1037/a0019744GRATITUDEEXPRESSIONSANDPROSOCIALBE