JournalofAbnormalChildPsychology,Vol.33,No.1,February2005,pp.55 - PDF document

JournalofAbnormalChildPsychology,Vol.33,No.1,February2005,pp.55Ð68( C  2005) DOI:10.1007/s10802-005-0934-z UnderstandingTeasing:LessonsFromChildrenWithAutism ErinA.Heerey, 1  2  4 LisaM.Capps, 3 DacherKeltner, 3 andAnnM.Kring 3 Teasingrequirestheabilitytounderstandintention,nonliteralcommunication,pretense,andsocial context.ChildrenwithautismexperiencedifÞcultywithsuchskills,andconsequently,areexpected tohavedifÞcultywithteasing.Tobetterunderstandteasingconceptsandbehaviors,childrenwith autism,theirparents,andageandVerbal-IQ-matchedcomparisonchildrenandparentsdescribed conceptsandexperiencesofteasingandengagedinaparentÐchildteasinginteraction.Theteasingof childrenwithautismwaslessplayfulandprovocativeandfocusedlessonsocialnormsthanthatof onatheoryofmindtaskaccountedforseveraloftheobserveddifferences.Discussionfocusedonthe importanceofunderstandingsocialcontextandplayfulbehaviorduringteasing. KEYWORDS: autism;socialinteraction;theoryofmind;teasing. ÒTherearesomethingsIdonÕtknowsomuchabout.  Teasingisoneofthem.Ó ÐAparticipantwithautism,11yearsold. Teasingisacomplexyetvitalsocialinteraction throughwhichpeoplesocializeeachother,enterintoand maintainrelationships,andnegotiategroupmembership inganditisespeciallycommonamongfamilymembers andpeers.Anecdotally,teasingappearstobeespecially problematicforchildrenwithautism:theyteaseineffec- tivelyandseemtohavedifÞcultyunderstandingwhythey arebeingteased(Grandin,1995).Aconceptualanalysis ofteasingsuggestswhy.Thecomprehensionofteasing requiresformsofsocialunderstandingthatchildrenwith autismÞnddifÞcult,includingtheabilitytounderstand intention,nonliteralcommunication,pretense,andsocial Onthebasisofthisreasoning,wetestedhypothesesrelat- ingautismto(1)teasingbehavior,(2)recountedexperi- encesofteasing,and(3)positiveandnegativeconceptsof 1 MarylandPsychiatricResearchCenter,UniversityofMaryland- BaltimoreMedicalSchool,Baltimore,Maryland. 2 MentallllnessResearch,Education,andClinicalCenter,Departmentof VeteransÕAffairsMarylandHealthCareSystem,Baltimore,Maryland. 3 PsychologyDepartment,UniversityofCalifornia,Berkeley,California. 4 (Suite6A),10NorthGreeneStreet,Baltimore,Maryland21201; e-mail:eheerey@mprc.umaryland.edu. teasing.High-functioningchildrenwithautismandtheir parentsandcomparisonchildrenandtheirparentspartic- ipatedinaninterviewaboutteasingconceptsandexperi- encesaswellasaparentÐchildteasinginteractionwherein theyinventednicknamesforoneanother. DeÞnitionsofTeasing ThetermÒteasingÓhasbeenappliedtomyriadsocial behaviors,rangingfromhostilebullying(Smith&Brain, 2000),totheaffectionate,playfulidiomsofromanticpart- geneouscategoryofbehaviors,wehavedeÞnedteasingas aprovocationthatcommentsonsomethingofrelevance tothetarget(Keltneretal.,2001).Theprovocationcan beverbal(e.g.,aninsultorcommentondeviantbehavior) ornonverbal(e.g.,apokeintheribs).Toreducethehos- tilityofthisprovocation,theteasermayemployplayful orÒoff-recordÓmarkers,whichconveythattheprovoca- tionistobetakenpartlyinthespiritofplay.Off-record markersincludeverbalcomments(e.g.,ÒjustkiddingÓ), repetition,exaggeration),andprosodiccues(e.g.,singÐ songvoice)thatsignalthattheprovocationisnotentirely serious.Teasingthen,isaformofprovocation(criticism orhostility)mitigatedbyoff-recordmarkers(playfulges- tures),eachofwhichmaybepresenttovaryingdegrees. Thistypeofteasingisconsideredprosocial,inthatits 55 0091-0627/05/0200-0055/0 C  2005SpringerScience  BusinessMedia,Inc. 56 Heerey,Capps,Keltner,andKring generalaimistocommentuponorcorrectanaspectof thetarget Õ ssocialbehavior. SocialUnderstandingandTeasing Althoughitmaybeparsedinotherways,ourde Þ - nitionsuggeststhatthegenerationandcomprehensionof teasinghingesonseveralabilities.Elementaltotheunder- standingofteasingarefairlysophisticatedabilitiesinthe comprehensionofintention.Inprovokinginplayfulfash- ion,theteaseconveysbothintenttocriticize,andplayful, affectionateintent.Theeffectiveteaserthen,mustbeable toconveythesecon ß ictingintentionsandtherecipientto decipherthem,oftenduringrelativelybrief,emotionally chargedexchanges. Teasingalsorequiresadroitnessinnonliteralcom- munication.Muchoftheplayfulcontentofateaseisnon- literal,seeninthesmiles,prosodicvariations(e.g.,sing Ð songvoice),andgrammaticaldevices(e.g.,exaggeration) thatindirectlyrendertheprovocationlesshostile.Thus,in understandingteasing,onemustinfertheimpliedmean- ingsbasedonthejuxtapositionoftheliteralprovocation andthenonliteralmeaninginasetofsubtleparalinguistic acts. Teasingfurtherinvolveselementsofpretensethat,to beunderstoodproperly,requiretheabilitytoassumepre- tendrolesandstances(forelaborationonthisnotion,see Clark,1996;Clark&Gerrlg,1984;Keltneretal.,2001). Thatis,teasingconveyscoexistingliteralandnonliteral representationsofthetargetofatease,whichnecessitate thatindividualsdifferentiaterealityfrompretense(Leslie, 1987).Asinotherformsofostensiblecommunication,the teaserandtargetassumehypotheticalidentities,andthe teaserdeliversaprovocationtothetarget.Thehypothet- icalteaserintendstheprovocationtobeserious,andthe hypotheticaltargetreceivesitassuch.Theactualtarget however,isexpectedtodiscoverthepretenseandtounder- standtheactualteaser Õ sattitudetowardthehypothetical teaser,thehypotheticaltarget,andtheprovocation(Clark &Gerrig,1984).Theabilitytoproduceandinterpretacts ofpretenseisthereforecriticaltoteasing. Finally,teasingtypicallycommentsuponsomede- viationfromthesocialnorm(e.g.,Keltner,Young,Heerey, Oemig,&Monarch,1998;Shapiro,Baumeister,&Kessler, 1991).Socialnormsconstitutethesetofexpectationsthat governbehaviorandinteractionsbothamonggroupmem- bers(e.g.,fraternitybrothers,romanticpartners,family members)andbetweenthegroupandtheoutsideworld. Thenormrelatedfocusofmuchteasingpoints,somewhat ironically,totheprosocialendsofteasing:inteasing,the teaserimpliesaninterestinthetarget Õ sengagementin sociallyappropriatebehavior(Bell,Buerkel-Rothfuss,& Gore,1987;Eder,1993;Eisenberg,1986).Thus,under- standingsocialnormsandtheactionsthatviolatethemis acentralelementintheprovocationofatease. Childrenunderstandtheprovocativenatureofteas- ingearlyon,evidentintheiruniversallynegativeviews thereof.Byages10 Ð 11however,theyhavebeguntocon- ceptualizeteasingasmorepositiveandprosocial,andto teasebothmoreplayfullyandprovocatively,whilecon- tinuingtoreportnegativefeelingsaboutteasing(Warm, 1997).Thisdevelopmentalshiftislikelyrelatedtothe engagementoftheseinterrelatedabilities,uponwhich wehavearguedprosocialteasinghinges.Childrenwith autismhavedif Þ cultieswithsuchskills,suggestingthat thepositivesideofteasingwillproveelusive. AutismandtheUnderstandingofIntention Ahallmarkofautisticdisorderisdif Þ cultywiththe- oryofmind,theabilitytounderstandmentalstates,such astheintentionsofothers(Flavell,1999).Theoryofmind hasbeenlinkedtoformsofspeechandsocialreasoning thatpresupposeanunderstandingofothers Õ mentalstates (Frith,Happe,&Siddons,1994;Happe,1993).Thesetof abilitiesthatweproposetounderlieteasingareallaspects oftheoryofmind. Higherfunctioningchildrenwithautismhavedif Þ - cultyinusingacharacter Õ sprobablementalstateorinten- tiontoexplainthecharacter Õ sambiguousspeechorbehav- ior(Abell,Happe,&Frith,2000;Tager-Flusberg,1999). Forexample,inHappe Õ s Ò StrangeStories Ó task,children wereaskedtointerpretshortsocialvignettesinwhichan ambiguousstatementwasmade.Inonevignette,achild coughsthroughoutlunch.Herfathersays, Ò PoorEmma. Youmusthaveafroginyourthroat. Ó Relativetocom- parisonchildren,childrenwithautismweremorelikely toconcludethatsuchstatementswereliterallytrueand usedfewerexplanationsinvolvingmentalstates(Happe, 1994).Thiskindofevidencesuggeststhatchildrenwith autismwillhavedif Þ cultydiscerningtheplayfulintentof ateasethatistypicallyconveyedinanambiguous,nonlit- eralfashion.Thisdiminishedcapacitytorecognizesocial intentionsmayservetomakethetruemeaningofatease remainopaque. NonliteralCommunicationandChildrenWithAutism Autismisassociatedwithahostofdif Þ cultiesinthe realmofcommunication(Tager-Flusberg,1999),includ- ingdif Þ cultiesdecipheringnonliterallanguage(Perner, UnderstandingTeasing 57 Frith,Leslie,&Leekam,1989).Childrenwithautism tendtoexplainsocialcommunicationsusingsystemsof interpretationthatrelyontheliteralmeaningofanut- terance(Hobson,1989).Forexample,individualswith autismfrequentlyhavedif Þ cultyinterpretingspeechacts thatrelyonthelistener Õ simplicitunderstandingofgen- eralsocialconventionsorpriorknowledge(e.g., Ò Doyou knowwhattimeitis? Ó ).Despitethefactthatthespeaker hasnotspeci Þ callyinquiredaboutthetime,thelistener isexpectedtounderstandthatthetimeofdayisbeing requestedandtorespondaccordingly.Individualswith autismmayrespondliterallytotheutterance(e.g., Ò Yes. Iknowwhattimeitis. Ó ),ratherthantothenonliteral,in- directmeaningthespeakerhasimplied(Bara,Bosco,& Bucciarelli,1999;Frith,1989). Theuseofidiomaticlanguage,anothervarietyof nonliteralcommunicationthatiscommontoteasing,par- ticularlyintheuseofnicknames(Bell&Healey,1992), hasbeenshowntobedif Þ cultforindividualswithautism. Inonestudy,agroupofhigh-functioningadultswith autismshowedperformancede Þ cits,relativetoacompar- isongroup,onataskinwhichtheyidenti Þ edidioms(e.g., Ò potluck, ÓÒ fatchance, ÓÒ pointblank, Ó etc.,Strandburg, Marsh,Brown,&Asarnow,1993).Sarcasmisalsodif Þ - cultforchildrenwithautism.Insteadofattendingtothe voicetoneandprosodyusedindelivery,childrenwith autismarelikelytoerroneouslyrespondonlytothever- balcontentofaremark,missingitsintendedmeaning (Frith,1989).Becausetheintendedmeaningofatease iscommonlyconveyedinanonliteralorindirectfashion (Keltneretal.,2001),teaseinterpretationislikelytobe problematicforchildrenwithautism. Pretense,Play,andChildrenWithAutism Spontaneouspretendplay,commonamongtoddlers andyoungchildren,appearstopresentdif Þ cultyforindi- vidualswithautism.Althoughchildrenwithautismhave beenshowntoengageinpretendplay,inparticularwhen initiatedbyothers(Jarrold,Boucher,&Smith,1993),they arelesslikelytospontaneouslyuseobjectsinimaginative waysandtendtopreferliteralplaytothatinvolvingpre- tense(Charman&Baron-Cohen,1997).Moreover,they havebeenshowntoengageinlessspontaneousfunctional playandimitativeplaythantheirpeers(Jarroldetal., 1993).Insocialinteraction,de Þ citsinspontaneousplay maymanifestasadearthinnonverbalgesturesandplayful verbalandnonverbalactions(Mundy,Sigman,&Kasari, 1990).Becauseteasingcanbeviewedasaformofsocial pretense,childrenwithautismmay Þ nditsplayfulaspects bothdif Þ culttoinitiateandtounderstand. UnderstandingofSocialContext AmongChildrenWithAutism Asteasingbecomesmoresophisticateditincreas- inglyrevolvesaroundsocialization(Keltneretat.,2001). Childrenwithautismhavebeenshowntobelesssen- sitivetothesocialenvironmentthancomparisonchil- dren(Capps&Sigman,1996).Forinstance,individuals withautismseemtobemoreinclinedtoresistfollow- ingnormsgoverningsocialcomportment(Frith,1989; Szatmari,Offord,Siegel,&Finlayson,1990),whichmay stemfromdif Þ cultyinidentifyinginappropriatebehav- iors.Instudiesofjudgmentsofbehaviors,childrenwith autismhadmoredif Þ cultyidentifyingandexplainingin- appropriatesocialbehaviors(Loveland,Pearson,Tunali- Kotoski,Ortegon,&Gibbs,2001),andfauxpasthan didcomparisonparticipants(Baron-Cohen,O Õ Riordan, Stone,Jones,&Plaisted,1999).Thus,someofthesocial de Þ citsmanifestedinchildrenwithautismmaystemfrom theirapparentlackofknowledgeaboutimportantsocial norms,implyingdif Þ cultyinunderstandingteasing. CURRENTRESEARCH Thedevelopmentalliteraturesuggeststhatwithin- creasedsocialunderstanding,teasingbecomesmoresym- bolicandplayful(Warm,1997),morefocusedonsocial norms(Keltneretal.,2001),andmorepositive(Shapiro etat.,1991).Inconcreteterms,thisliteraturesuggests thatwhereasthehostile,provocativecomponentofteas- ingvariesonlyslightlyacrosslevelofsocialunderstand- ing,themoreplayful,prosocialcomponentsofteasing increaseinvolumeandsophisticationwithsocialdevel- opment,promptingchangeinindividuals Õ interpretations ofteasing(Lightner,Bollmer,Harris,Milich,&Scambler, 2000).Onthebasisofourde Þ nitionofteasingandwhat isknownaboutautism-relateddif Þ cultiesinsocialunder- standing,weexpectboththeunderstandingandgeneration ofteasingtobeproblematicforchildrenwithautism. Asa Þ nalnoteaboutteasing,parentsplayanim- portantandearlyroleinchildren Õ steasing(Keltneretal., 2001).Parentsofchildrenwithautismhaveagreatdealof concernabouttheteasingtheirchildrenrelateand,anec- dotally,attempttosoftenthestingofthisoftenpainful interaction(Barron&Barron,1992).However,parentsof childrenwithautismreportoccasionalboutsoffriendly, ritualizedteasingorjokingwiththeirchildren,oftencen- teredondailyroutines(McDonnell,1993).Itistherefore likelythatparentsofchildrenwithautism,likethoseof, typicallydevelopingchildren,adapttheirteasingtomeet thecapabilitiesoftheirchildren. 58 Heerey,Capps,Keltner,andKring METHODS Aspartofalargerstudyofsocialunderstanding, childrenwithhigh-functioningautism,typicallydevelop- ingchildren,andtheirparentswereinterviewedabouttheir conceptsaswellastheirexperienceswithteasing.Inaddi- tion,parentsandtheirchildrenengagedinadyadicteasing interactioninwhichtheyinventedandsharednicknames foroneanother. Participants Participantsincluded43children:23nonretarded childrenwithdiagnosesofautism( n  10)orAsperger Õ s syndrome( n  13)comprisedthehigh-functioning autism/Asperger Õ ssyndrome(HFA)group,and20typi- callydevelopingchildrenmadeupthecomparisongroup. HFAparticipantswererecruitedthroughclinicians,none ofwhomwereaf Þ liatedwiththeproject.Tocon Þ rmdi- agnosesofeitherautismorAsperger Õ ssyndrome,one orbothparentsofHFAchildrenwereadministeredthe AutismDiagnosticInterview-Revised(ADI-R;Lord, Rutter,&LeCouteur,1994),whichgeneratesdiagnoses basedon DSM-IV criteria.TheADI-Rachievesreliable diagnosesofautismandAsperger Õ ssyndromeacrossgen- dersandagegroups(forspeci Þ c Þ ndings,seePilowsky, Yirmiya,Shulman,&Dover,1998).Nosigni Þ cantdif- ferencesemergedbetweenparticipantswithautismand thosewithAsperger Õ ssyndromeonanymeasuredvari- ables.Thus,analyses(below)didnotdistinguishpartici- pantsbydiagnosis. OnlychildrenwhosecurrentVerbal,Performance, andFullScaleIQscores,asmeasuredusingtheWechsler IntelligenceScaleforChildren,EditionIII(WISC-III; Wechsler,1991),were80orhigherwereincludedinthe study.Acutoffof80ensuredthatstudyparticipantswere withinthenormalrangeontheWISC-IIIandthatnochild hadmentalretardation.AllHFAchildrenwereworkingat ageappropriategrade-levelsand21ofthe23children werebeingeducatedinmainstreamedschoolenviron- ments.Twoattendedaprivateschoolforchildrenwith learningdisabilities.AutismandAsperger Õ ssyndromeaf- fectboysathigherratesthangirls(Capps&Sigman, 1996).Thisbiaswasre ß ectedinourHFAsample,18of whomwereboys. Comparisonchildrenwererecruitedfromlocal schoolsandrecreationprogramsinalargemetropolitan area.Asindicatedduringatelephonescreening,noneof thechildrenhadpriorpsychiatrichistoriesnorwerethey beingtreatedforpsychologicaldif Þ cultiesofanykind. Inaddition,thepresenceofaPervasiveDevelopmental Disorderwasruledoutusingaparent-reportquestion- naire,thePervasiveDevelopmentalDisorderScreening Test(Siegel,1986).Groupswerestatisticallymatchedon verbalIQ,chronologicalage,andgender(foradiscus- sionofageandverbalIQmatchinginstudiesofhigh- functioningautismseeOzonoff,Rogers,&Pennington, 1991).TherewerenogroupdifferencesonVerbal IQ( M HFA  103  82(14  86); M Control  106  00(7  97); t (38)   86  ns ),fromtheWISC-III.Allchildrenwere aged8to15years,andthegroupsdidnotdifferonage, ( M HFA  10  92(2  62); M Control  10  57(1  21); t (38)  1  18  ns )orgender(HFA:5girls;18boys;Control:4girls; 16boys;  2 (1)  0  75  ns ).Allchildrenreceived$10/hr inappreciationoftheirparticipation. Procedure Testingwascompletedintwo2-hrsessions,typi- callyoccurring5to7daysapart,duringwhichchildren completedtheWISC-IIIandabatteryoftasksassessing socialunderstanding,emotion,theoryofmind,andnar- rativecompetence.SessionswerevideotapedusingPana- sonicvideocamerasthatwerevisibletoparticipants.All participantswereseatedatatableacrossfromanexperi- menterandcameraswerepositionedsuchthatthepartic- ipant Õ sface,upperbody,andaportionofthetablewere inviewatalltimes.Oneofthetaskspresentedduringthe Þ rststudysession,aninterviewaboutteasing,generated datarelevanttothepresentinvestigation.Aparent Ð child teasinginteraction(adaptedfromKeltneretal.,1998)was conductedduringthesecondstudysessionandthesedata werealsoincludedinthepresentstudy.Finally,atheory ofmindtask,StrangeStories(Happe,1994),wasadmin- isteredtochildrenforthepurposesofrelatingtheoryof mindtochildren Õ steasingbehaviorsandconcepts. Parent Ð ChildTeasingInteraction Childrenandtheirparentswereseatedside-by-side, acrossatablefromtwoexperimenters.Theexperimenters, bothofwhomhadworkedcloselywithchildrenandtheir parentsduringthestudysessions,introducedandmod- eledthetask.Participantsweretoldthattheywouldbe playinga Ò nicknamegame Ó inwhichtheywouldeachbe askedtoinventanicknamefortheotherandexplainthat nickname.Toallowparticipantsto Ò warm-up Ó totheinter- actionsetting,theywereaskedtodescribeanynicknames theyhadforoneanotherathomeandwhetherothermem- bersoftheirfamilieshadnicknames.Experimentersthen explainedthattheywouldinventnicknamesforonean- other,whichcouldberealorfanciful,andprovideabrief UnderstandingTeasing 59 rationaleforeachnickname.Noconstraintswereplaced onthegenerationofnicknames.Experimenter1demon- stratedthegamebyteasingExperimenter2(i.e., Ò Iwould callyou,T.F.forTomatoFacebecauseofthewayyou blushwhenyougetcallednicknames. Ó ).Experimenter2 likewisedemonstratedtheprocedurebyinventinganick- nameforthechild(i.e., Ò IwouldcallyouS.P.,forSmarty Pantsbecauseyouweresosmartinourstudytoday. Ó ). Allparticipantsheardthesametwointroductorynick- names.Participants Õ questionswereansweredandchil- drenwerepromptedtodelivertheirnicknames(e.g., Ò Can youthinkofanicknameforyourmom? Ó ).Childrende- liveredtheirnicknamesand,iftheydidnotspontaneously generatearationaleforthenickname,werepromptedto doso(e.g., Ò Whywouldyoucallherthat? Ó ).Parentswere thenasked,usingsimilarprompts,toinventandexplaina nicknameforthechild. Ifachildwasunabletogeneratethe Þ rstnickname, aswasthecasewith13childrenacrossthegroups,the parentwaspromptedtoproduceandexplainanickname forthechild.Followingtheparent Õ sdelivery,thechild wasagainpromptedtoinventanicknamefortheparent. Ifthechildwasstillunabletogenerateanickname(two comparisonchildrenandfourHFAchildren),thegame wasconcludedwithoutthechild Õ stease.Noparentwas unabletogenerateanickname. ChildandParentTeasingInterviews TeasingConcepts Teasinginterviewsforbothchildrenandparentsbe- gansimilarly.Participantswereasked, Ò Howdoyoude Þ ne teasing? Ó Iftheparticipantwasunsureaboutthemean- ingofthequestionorunabletoanswer,theexperimenter providedasecondprompt: Ò Whenyouthinkofteasing, whatcomestomind? Ó Allparticipantsunderstoodandan- sweredthesecondprompt,providingthreeto Þ vedifferent ideasaboutteasing. TeasingExperiences Followingthepromptsforconceptsofteasing,par- ticipantswereaskedtorecountteasingexperiencesbyde- scribingtimeswhentheyteasedandwereteasedbyothers. Speci Þ cally,childrenwereaskedtodescribeoneincident ofteasingwhentheyteased,andoneincidentofteasing whentheywereteasedbypeersatschool,parentsathome, and,ifapplicable,siblingsathome.Thus,childrenwith siblingswerepromptedtorecountatotalofsixteasing episodes:(1)asthetargetofpeerteasing,(2)whenteas- ingapeer,(3)asthetargetofaparent Õ steasing,(4)when teasingaparent,(5)asthetargetofsiblingteasing,and (6)whenteasingasibling.Childrenwithoutsiblingswere askedtorecountfourteasingepisodes,excludingprompts aboutsiblingteasing.Parentswereaskedtorecallatime inwhich(1)theyteasedthechildinthestudyand(2)the childinthestudyteasedthem.Parentsofmorethanone childwereadditionallypromptedtodescribeteasingin- teractionswitheachoftheirotherchildren,astheyhad with,thechildinthestudy.Participantswereaskedto provideaccountsoftheirteasinginteractionsbutwerenot speci Þ callypromptedfordetails(e.g., Ò Tellmeabouta timewhenyouwereteasedbyyourbrother.Tellmewhat happened. Ó or Ò Tellmeaboutatimewhenyouteasedyour momordad.Tellmewhathappened. Ó ). StrangeStoriesTask Atheoryofmindmeasure,consistingof12short vignettesinwhichstorycharactersproduceambiguous speechoractions(e.g.,achild,playingwithafriend,picks upabananafromafruitbowl,holdsittoherearandsays, Ò Look!Thisbananaisatelephone! Ó ),wasadministered (foradescriptionofthe Ò strangestories Ó theoryofmind taskaswellasscoringinformationandrelated Þ ndings seeHappe,1994).Theexperimenterreadeachvignette aloudtothechildandthenaskedthechildtoanswertwo questionsaboutthestory: Ò Wasittrue,what[astorychar- acter]said? Ó and Ò Whydid[thestorycharacter]saythat? Ó Positivecommentsweremadeduringtestingbutchildren weregivennofeedbackaboutwhethertheiranswerswere correct. CodingofChild Ð ParentTeasingInteractions Teasinginteractionswerecodedfromvideotapefor thefollowingitems: TeaseClassi Þ cation Eachteasewasclassi Þ edaccordingtotypeofteasing: socialnormviolation (ateaserelatingtoasocialnorm, e.g., Ò pant-a-balloon Ó forachildwhowears Ò pantsthat look5sizestoobig Ó ), characterteasing (ateaserelating toaconsistentaspectofanindividual Õ sbeing,e.g., Ò little go-go Ó forachildwho Ò isalwaysonthegoandnever seemstogettired Ó ),or endearingnickname (oftenacur- rentnicknameoftheparentoranicknameofthechild duringinfancy,e.g., Ò shin-shin Ó meaning Ò littlestar Ó in Chinese).Itwasseldomapparentfromthenicknamealone 60 Heerey,Capps,Keltner,andKring howateaseshouldbeclassi Þ ed.Therefore,inallcases bothnicknameandrationalewereusedinclassi Þ cation. Ninety-twopercentofteaseswereuniquelyclassi Þ edinto thethreecategories.Theremainingnineteases,allgen- eratedbychildren(twocomparisonandsevenHFA),in- cludednicknamesthatwereeitherentirelynonverbal,or wereacombinationofnonverbalgestures(e.g.,sticking outtongue)andverbalizedsoundsthatdidnotinclude words(e.g.,snif ß ing,sighing).Sixofthenineteasesin- cludedverbalrationales.Theseteases,includingthenick- nameandrationale,wereclassi Þ edandanalyzedforboth verbalcontentandnonverbalbehavior(seebelow).The remaining3teases,allgeneratedbyHFAchildren,con- tainednoverbalcontentandwereanalyzedfornonverbal behavioronly. VerbalContent Theverbalcontentofeachteasewasratedontwo, unipolar,7-pointLikertscalesaccordingtohow af Þ liative andhow critical itwasoftherecipient(1  notatall af Þ liative/critical ;7  extremelyaf Þ liative/critical ).Af- Þ liativeteaseswerethoseinwhichtheverbalcontentof eitherthenicknameorrationaleincludedpraise,endear- ment,referencestopositivebehaviors,etc.Forexample, onechildwasnicknamed Ò Super-Student Ó byhismother duetoastraight-Areportcard.Thisteasewasconsidered af Þ liativebecausetheteasecenteredonapositivetopic, beingagoodstudent,andinvolvedpraise.Criticalteases involvedtopicssuchasnegativeaspectsoftherecipient Õ s character,accidents,mistakes,orteasesinwhichthere- cipientwasaccusedofuntowardbehavior.Forexample, onechildcalledherfather Ò CreativeFarts Ó duetopublic ß atulence. PlayfulandCriticalBehaviors Onthebasisofasummaryoftheliteratureregard- ingnonverbalmarkersofteasing(Keltneretal.,2001), wecodedplayfulbehaviors,includingfriendlylaughter, smiles,playfulprosody(sing Ð songvoice),playfulges- tures(waving),reassuringphysicalcontact(hand-holding, pattingtheinteractionpartner),andplayfulmimicry(play- fulparodyofanactionorotherpersondesignedtopro- motelaughter).Criticalbehaviorsincludedfrowns,sneers, grimaces,politesmiles,criticalprosody(e.g.,sarcastic tones),criticalgestures(e.g.,stickingouttongue),aggres- sivephysicalcontactoractions(e.g.,slapping,poking, pretendingtohitanother),andhostilemimicry(e.g.,par- odyofanactionorotherdesignedtomock).Foreach tease,wesummedthenumberofobservationsofplayful, off-recordmarkersandthenumberofcriticalgesturesto generateatotalnumberofplayfulandcriticalbehaviors perparticipant. CodingTeasingInterviews Interviewswerecodedfromvideotapesforverbal content.Unlessotherwisenoted,alldatageneratedwere frequencycounts. TeasingConcepts Wecodedparticipants Õ conceptsofteasingaseither (1) positive :iftheresponseindicatedthatteasingcould beviewedasenjoyable(e.g., Ò teasingisplayful, ÓÒ joking around, ÓÒ fun Ó )or(2) negative :iftheresponseindicated thatteasingwasunpleasant(e.g., Ò teasingisbeingmean, Ó Ò bullying, ÓÒ hurtingsomeone Õ sfeelings Ó ).Eachpartici- pantspontaneouslygeneratedatleasttwodistinctconcepts ofteasingthatwerecodedaseitherpositiveornegative. Thetotalnumberofnegativeconceptsre ß ectedthenum- berofuniquenegativestatementsparticipantsmadeabout teasing.Totalpositiveconceptsweresimilarlycalculated. TeasingExperiences Wecodedparticipants Õ recountedteasingexperiences asfollows: SocialContext Recountedteasingexperienceswerecodedforthe totalnumberofreferencesmadeineachofthefollow- ingdomains:(1) antecedents :orwhytheteasehappened, werecodedfromdescriptionsoftheactionsleadingupto theteaseitself(e.g., Ò Ikeptmissingmyshotatbasket- ball  Ó );and(2) consequences :theresultsofthetease, werecodedfromdescriptionsoftheoutcomeofatease (e.g., Ò  soIconcentratedhardandmadealltherestof myshotsexceptone,then[theteaser]couldn Õ tsayany- thingaboutthatanymore. Ó ). Valence Theverbalcontentofateasecanbeclassi Þ edasei- therpositiveornegativeinvalenceaccordingtowhether theteaseisaboutpositiveornegativebehaviorsortraits (seeKeltneretal.,1998).Forexample,achildwhore- portsbeingteasedaboutoutstandingathleticoracademic abilityisbeingteasedaboutapositiveability,eventhough UnderstandingTeasing 61 thechildmaynotreportfeelingpositiveduringthetease. Teasingwithnegativecontentinvolvesnegativetraitsor behaviors,suchasirritability,asinthecaseofachild teasedaboutpicking Þ ghtswithasibling.Recountedteas- ingexperienceswereclassi Þ edaseither(1) positive or (2) negative incontentvalence.Valencewascodedinde- pendentlyofthechild Õ sinterpretationofthetease. TeasingCategories Teasingexperienceswereclassi Þ edintofourcate- gories(seeWarm,1997).Teasesaboutviolationsofso- cialnorms,e.g.,thingssomeone does suchasdroppinga lunchtrayintheschoolcafeteria,wereclassi Þ edas social normviolations .Teasingexperiencesthatfocusedonan individual Õ sphysicalorpsychologicalcharacter,suchas apersonalitytrait(beingverynice),physicaltrait(hav- ingalargenose),ormentalcharacteristic(beingforget- ful),wereclassi Þ edas characterteasing .Descriptions ofmimicry,namecalling,mocking,andphysicalteasing (poking,hitting,touching)werecategorizedas taunting . Practicaljokes,pranks,tricks,whitelies,anddescriptions ofsituationsinwhichapersonisledtobelievesomething falsewereclassi Þ edas trickery .Allbuttwooftheteasing descriptions(about99%)wereclassi Þ edintooneofthe fourcategories. Forthethreemeasuresofrecountedexperiences Ñ socialcontext,valence,andteasecategory Ñ wesummed thescoresacrossallrecountedexperiencesforeachmea- sureandthendividedbythenumberofexperiencesthatthe participantrecounted.Inourdataanalysis,wecollapsed acrossteaserandtargetexperiences,becauseanumberof HFAchildren(7)didnotrecountatimewhentheyteased someone. CodingofStrangeStories Children Õ sresponsestothestrangestoriestaskwere codedasfollows.Childrenwhogeneratedcorrectanswers tothe Þ rstquestion( Ò Wasittrue,what[storycharac- ter]said? Ó ),andexplainedtheiranswerswithreference tothesocialprocessofthestory,received2points.Those whogeneratedcorrectanswerstothe Þ rstquestionbut explainedtheiranswersusingincorrectmentalorsocial processesortheperceptualfeaturesofthestoryreceived ascoreof1point.Childrenwhoansweredincorrectly andexplainedtheiranswersreferringtonomentalorso- cialprocessesorsaid, Ò Idon Õ tknow, Ó receivedscores of0.Scoresweresummed,yieldingamaximumscore of24points(forcompleteresults,seeSobel,Capps& Gopnik,1999). CodingandReliability Allcoderswereblindtobothparticipants Õ groupsta- tusandtoexperimentalhypotheses.Agroupoffourun- dergraduateresearchassistantsweretrainedtoclassify teasesandcodeplayfulandhostilecontentandbehavior. Eachratercodedtwointeractionsunderthesupervision ofthe Þ rstauthorandworkedindependentlythereafter. Excludingthesixchildrenwhowereunabletogenerate nicknames,atotalof80teasesweregeneratedandcoded duringtheinteraction. Toassessreliability,eachteasewasindependently codedbytworaters.Thereliabilityofraters Õ teaseclassi- Þ cationswasassessedusingakappacoef Þ cient(    84). Intraclasscorrelationcoef Þ cientswerecalculatedforrat- ingsofverbalcontentaswellasparticipants Õ nonverbal behavior.Theserangedfrom.64to.91.Analyseswere conductedusingadatasetthatre ß ected,foreachtease, theaverageofthetworaters Õ codes. Twoundergraduateresearchassistants,notassoci- atedwiththecodingoftheteasinginteractions,worked tocodeteasinginterviews.Eachcodedseveralinterviews underthesupervisionofthe Þ rstauthor,andthereafter workedindividuallyfromvideotapestogeneratedatarel- evanttotheinvestigationofteasinginterviews.Coders overlappedtheircodingsuchthatabouthalfoftheinter- views(46)werecodedindependentlybytwocoders. Forthechild-teasinginterviews,theintraclasscorre- lationsbetweencoders Õ ratingsofitemsrelatingtocon- ceptsandexperiencesofteasing,includingde Þ nitions, antecedents,consequences,socialcontext,functions,and nonverbalbehaviorsrangedfrom.69to.89.Inaddition, eachteasingepisodewasclassi Þ edaccordingtoitscon- tentcategory:socialnormviolation,characterteasing, taunting,ortrickery(    76).Forparent-teasinginter- views,theintraclasscorrelationsbetweencoders Õ ratings ofitemsrelatingtoconceptsandexperiencesofteasing rangedfrom.71to.94.Classi Þ cationofparent-reported teasingepisodesaccordingtocontentcategoriesyielded akappacoef Þ cientof.73. Tworaterscodedthetheoryofminddatafromtran- scripts.Thesecoderswereindependentofthosewhohad codedtheteasinginteractionsandinterviews.Theyat- tainedahighlevelofagreement(    87). RESULTS Becauseoftheinteractivenatureofthetask,teas- inginteractionswereanalyzedbydyad.Weconducted amixedmodelANOVA,treatinggroup(HFA,compari- son)asthebetween-dyadvariableandparticipant(parent, 62 Heerey,Capps,Keltner,andKring child)asthewithin-dyadvariable.Posthoctestscompar- ingchildrenfromthetwogroupsandparentsfromthetwo groupswereconductedonsigni Þ cantomnibus F susing theNewman-KeulscorrectionfortypeIerrorrates(Glass &Hopkins,1996).Forteasinginterviews,parents Õ and children Õ sdatawereanalyzedseparately,againusingthe Newman-Keulsmethodtocontrolerrorrates. TeasingBehavior Wehypothesizedthatchildrenwithautismandtheir parentswouldhaveparticulardif Þ cultygeneratingthe playfulcomponentofteasing,whichhingesontheuse ofnonliteral,off-recordmarkersandpretense.Consistent withthishypothesis,amixedmodelANOVAfoundthat comparisonchildrenandtheirparentsusedmoreplayful behaviors,suchassmiles,unusualintonation,andexag- geratedgesturesintheirteasingthandidchildrenwith autismandtheirparents, F (1  40)  8  06; p  01(see TableIformeans).Morespeci Þ canalysesrevealedthat comparisonchildrenusedmoreplayfulbehaviorsthandid HFAchildren, t (38)  2  42; p  05,asdidcomparison groupparentswhencomparedwithHFAgroupparents, t (38)  2  56; p  05.Parentsandchildrendidnotdiffer intheextenttowhichtheyincorporatedplayfulbehaviors intotheirteasing, F (1  40)   90; ns . Givenanecdotalevidencesuggestingthatprovoca- tiveteasingisproblematicforchildrenwithautism,we expectedHFAgroupparticipantstoteaseinmoreaf Þ l- iative,lesscriticalwaysthancomparisongrouppartici- pants.Mixed-modelANOVAsdemonstratedthatcompar- isonparticipantswere(1)lessaf Þ liative, F (1  40)  6  64; p  01,and(2)morecritical, F (1  40)  8  93; p  01, intheirteasingthanwereHFAparticipants.Thesesame analysesrevealedthatregardlessofgroup,parentswere TableI. ContentandBehaviorinTestingInteractions ChildrenParents HFA( N  23)Comparison( N  20)HFA( N  23)Comparison( N  20) Nonverbalbehaviors Playful1.61(1.23)2.27(1.08)1.39(1.15)2.07(1.19) Critical0.36(0.88)0.17(0.43)0.25(0.53)0.13(0.38) Teasingcontent Af Þ liative3.65(1.51)3.38(1.40)5.20(1.23)4.20(1.66) Critical2.39(1.75)3.27(1.53)1.36(1.07)2.54(1.48) Teasingtypes Socialnorms0.41(0.68)0.68(0.65)0.48(0.67)0.93(10.79) Character0.24(0.54)0.11(0.38)0.32(0.58)0.42(0.75) Endearingnames0.39(0.49)0.09(0.29)0.40(0.50)0.36(0.49) Note .Allvariablesarelistedasmeanspertease.Standarddeviationsareinparentheses. lesscriticaloftheirchildrenthanchildrenwereoftheir parents, F (1  40)  8  67; p  01,andalsomoreaf Þ lia- tivewhileteasingthanwerechildren, F (1  40)  17  50; p  001.Morefocusedcomparisonsrevealedthatchil- drenintheHFAgroupdidnotteaseinmoreaf Þ liativefash- ionthancomparisonchildren, t (38)  0  95; ns ,although parentsofHFAchildrenwereindeedmoreaf Þ liativein theirteasingthancomparisonparents, t (38)  2  70; p   01.Comparisonchildrendid,however,teaseinmorecrit- icalwaysthandidHFAchildren, t (38)  2  01; p  05. Likewise,comparisongroupparentsweremorecriticalof theirchildrenthanwereHFAgroupparents, t (38)  3  40; p  01. Intheanalysisofcriticalbehaviors,includingcritical gestures,frowns,andsarcasticintonation,mixedmodel ANOVAyieldedneitherdifferencesbetweentheHFA andcomparisongroups, F (1  40)  0  89  ns ,norbetween parentsandchildren, F (1  40)  0  17; ns  Wenowturntothenormativecontentoftheteas- ing.WehadpredictedthatHFAchildrenandtheirpar- entswouldbelesslikelytoteaseaboutsocialnormsthan comparisonparents.Recallthattheteaseswereclassi Þ ed accordingtowhethertheteasereferredtoasocialnorm violation,theindividual Õ scharacter,oranendearingnick- name.Therewerenogroupdifferencesinthenumberof timesparticipantsteasedaboutaspectsofcharacter(means appearinTableI; F (1  40)  1  82  ns ).Thenumberof teasesclassi Þ edasendearingnicknamesdidnotdiffer amongparentsandchildren, F (1  40)  1  18  ns ,nordid itdifferacrossHFAandcomparisongroups, F (1  40)  1  75  ns ,althoughchildrenintheHFAgrouptendedtouse moreendearingnicknamesthandidchildreninthecom- parisongroup, t (38)  1  80; p  10.Consistentwithour expectations,comparisongroupparticipantsteasedabout socialnormviolationsmorefrequentlythandidHFAgroup participants, F (1  40)  8  04; p  01,and,asonemight UnderstandingTeasing 63 intuit,parentsteasedaboutsocialnormviolationsmore frequentlythandidtheirchildren, F (1  40)  5  60; p   05.Morefocusedanalysesshowedthatcomparisonchil- drenteasedtheirparentsaboutsocialnormviolations morefrequentlythandidHFAchildren, t (38)  2  15; p  05.Likewise,comparisongroupparentsteasedmore frequentlythanparentsofchildrenintheHFAgroupabout socialnormviolations,thuscon Þ rmingpredictions, t (38)  2  29; p  05. Analysesoftheactualteasingbehaviorofparents andchildrenyieldedresultslargelysupportiveofourhy- potheses.Whencomparedwithappropriatecomparison individuals,childrenintheHFAgroupandtheirparents werelessplayfulandlesscriticalintheirteasing,andtheir teasingwaslesslikelytofocusonsocialnormviolations. Wenowturntoanalysesofparticipants Õ recountedexperi- encesofteasing,whichmorespeci Þ callyaddresspossible differencesintheawarenessofsocialcontext Ñ acritical partofunderstandingteasing. TeasingExperiences Weexpectedcomparisongroupparticipantstoshow agreaterawarenessofsocialcontextintheirrecounted teasingexperiences.UsingamixedmodelANOVAwith group(HFA,comparison)asthebetween-participantsfac- torandsocialcontext(antecedents,consequences)asthe within-participantsfactor,wefoundthatcomparisonchil- dren,ingeneral,referredtothesocialcontextintheir self-reportedteasingexperiencesmorefrequentlythan didchildrenwithautism, F (1  40)  12  07, p  01.As showninTableII,comparisonchildrenreferredtoan- tecedents, t (38)  2  90; p  05,andconsequencesof teasingmorefrequentlythandidHFAchildren, t (38)  TableII. TeasingConceptsandRecountedTeasingExperiences ChildrenParents HFA( N  23)Comparison( N  20)HFA( N  23)Comparison( N  20) Teasingconcepts Positive0.20(0.41)1.14(1.01)1.23(1.03)1.00(0.86) Negative2.48(1.43)2.82(0.29)3.45(2.57)1.36(0.73) SocialContestvariables Antecedent1.55(1.20)3.14(1.59)2.20(1.58)4.07(0.96) Consequences1.65(1.72)2.85(1.63)3.60(1.50)4.78(1.78) Teasingtypes Socialnorms0.10(0.30)1.64(1.79)0.48(0.55)1.54(0.88) Character0.78(1.00)0.93(0.98)1.14(0.97)1.27(1.04) Taunting0.63(0.81)0.71(0.71)0.81(0.72)0.88(0.73) Trickery0.35(0.58)0.07(0.26)0.03(0.09)0.07(0.15) Note .Meansofsocialcontextvariablesrepresenttheaveragenumberofinstancesofeachitemperrecountedteasing experience.Standarddeviationsarenotedinparentheses. 2  15; p  05.Asimilaranalysiswasconductedtoexam- ineparents Õ data.Aswithchildren,parentsinthecom- parisongroupmademorespeci Þ creferencestosocial contextinteasingexperiencesthandidparentsinthe HFAgroup, F (1  40)  21  30, p  01.Comparisonpar- entsexplicitlyreferredtobothantecedents, t (38)  3  82; p  01,andconsequences, t (38)  2  09; p  05,more frequentlythandidHFAparents. Intermsofthecategoriesofteasing,childreninboth groupsrecalled,withequivalentfrequency,teasingexpe- riencesrelatedtocharacterteasing, t (38)   60  ns ,and taunting, t (38)   59  ns ,(seeTableII).Aspredicted, childreninthecomparisongroupreportedmoreteasing experiencesrelatedtosocialnormviolationsthanHFA groupchildren, t (38)  3  01, p  01.HFAchildrenre- portedexperiencesofteasingclassi Þ edastrickerymore frequentlythancomparisongroupchildren, t (38)  2  00, p  05.Parentsofcomparisonchildrenlikewise recountedmoreinstancesofteasingaboutsocialnormvi- olationsthandidparentsofHFAchildren, t (38)  3  51, p  01.Therewerenogroupdifferencesamongparents foranyothercategoryofteasing. ConceptsofTeasing Children Our Þ nalinterestwasinthegeneralcontentofpartic- ipants Õ conceptsofteasing,elicitedwhenparticipantspro- videdde Þ nitionsofteasing.Regardlessofgroup,among children,negativeconceptsoccurredwithgreater frequency(89%)thanpositiveconcepts(11%) Ñ a Þ nding thatonewouldexpectfromthedevelopmentalliterature, whichshowsthatchildrenupthroughage11de Þ neteasing 64 Heerey,Capps,Keltner,andKring largelyinnegativeterms(Lightneretal.,2000;Warm, 1997).Childreninthecomparisongroup,however,of- feredpositiveconceptsofteasingatasigni Þ cantlygreater frequencythandidHFAgroupchildren, t (38)  3  15, p  01,consistentwithourexpectation.Thetwogroups didnotdifferintermsofthefrequencywithwhichthey de Þ nedteasinginnegativefashion, t (38)  1  12  ns (see TableII). Toascertainthatchildrenwithautismdidnotfailto expresspositiveconceptsofteasingsimplybecausethey areteasedinmorenegativewaysthancomparisonchil- dren,weexaminedthevalence(positiveornegative)of theirrecountedteasingexperiences.Recallthatrecounted teasingexperienceswerecodedasbeingeitherpositive ornegativeincontentvalence,independentofthechild Õ s emotionalevaluationoftheexperience.Thatis,itwaspos- sibleforateasetobecodedaspositiveincontentvalence (e.g.,breakingthecurveonatest),eventhoughthechild experiencedtheteaseasnegative,andviceversa.Forex- ample,onechildreportedateasethatwasnegativeincon- tentvalence(clumsilyspillingmilkatlunch)butfound hisfriends Õ teasingtobefunnyandreportedapositive experience.Codesreferringtocontentvalencearethus independentofchildren Õ sinterpretationsoftheteasing. Wefoundthatallchildrenrecalledmorenegativeteas- ingexperiences(84%)thanpositiveteasingexperiences (16%).Therewerenogroupdifferencesinrecountedteas- ingexperiencesclassi Þ edaseitherpositiveincontentva- lenceornegativeincontentvalence( positive : M control  1  10(0  62), M HFA   88(0  74), t (38)  1  36  ns ; nega- tive : M control  3  59(1  89),M HFA  4  21(1  66), t (38)   82  ns ).Thus,HFAchildrendonotappeartoofferfewer positiveconceptsofteasingsimplybecausetheyexperi- encenegativeteasingmorefrequently. Parents Liketheirchildren,parentsofHFAgrouppartici- pantsofferedmorenegativeconceptsofteasingthandid parentsofcomparisonchildren, t (38)  3  14, p  01. Interestingly,analysesrevealednogroupdifferencesin parents Õ positiveconceptsofteasing, t (38)   44  ns . LinkingParents Õ andChildren Õ sTeasing Takentogether,the Þ ndingsfromtheteasinginter- actionandinterviewsuggesttwothemes:(1)thattyp- icallydevelopingchildrenandtheirparentsappreciate thepositiveandplayfulaspectsofteasingmoresothan dohigh-functioningchildrenwithautismandAsperger Õ s syndromeandtheirparents,and(2)thattheyengagein teasingthatismorecloselylinkedtothesocialcontext thandoHFAgroupparticipants.Howmightparents Õ and children Õ steasingbehaviorsrelate?Weusedasetofcor- relationalanalysestoexplorethisquestion. Individuals Õ teasingstylesandbehaviorsduringin- teractionsarethoughttorelatetooneanother(Keltner etal.,2001).Wefoundthistobethecase,atleastamong comparisonparticipants.Criticalcontentamongchildren Õ s andparents Õ nicknameswaspositivelyassociated, r   69; p  01.Moreover,comparisonchildren Õ saf Þ liative contentwasnegativelycorrelatedwithparents Õ critical content, r   46; p  05,althoughparentswereaf- Þ liativeregardlessoftheirchildren Õ scritical, r   57; p  05,oraf Þ liativecontent, r   53; p  05.Parents Õ criticalbehaviorstendedtodecreaseaschildren Õ scriti- calteasecontentincreased, r   38; p  10.,Finally, children Õ splayfulbehaviorswerepositivelyrelatedtopar- ents Õ criticalbehaviors, r   54; p  05.Thus,among comparisondyads,participantsteasingstylesappeared related. ThestorywasmuchdifferentfortheHFAchildren andtheirparents.Correlationsbetweenparents Õ andchil- dren Õ steasingbehaviorssuggestedlittlereciprocityorin- terdependenceintheirteasing.Nosigni Þ cantcorrelation betweenHFAchildren Õ steasingbehaviorsandthoseof theirparentswerefound(seeTableIII). TheoryofMindandTeasingBehavior Might Þ ndingsrelatetochildren Õ stheoryofmind? Hierarchicalmultipleregressionanalysesaddressed whetherthegroupdifferencesinchildren Õ sreferencesto socialcontextandplayfulteasingbehaviorcouldbeac- countedforbydifferencesintheoryofmind.Weused groupasthedependentvariableineachregression.Each analysisincludedonlychildren Õ sdata,aswehadnotas- sessedparents Õ theoryofmindskills. Groupdifferencesinchildren Õ sreferencestosocial contextastheyrecountedteasingexperienceswereex- aminedaftercontrollingfortheoryofmind.Atstep1, thevariablesageandverbalIQwereentered.Asprevi- ouslyreported,nogroupdifferencesemergedasthisstep (  R 2   07; F  1  03; ns ).AtStep2,theoryofmind scoreswereenteredandweresigni Þ cantlyassociatedwith groupstatus(  R 2   18; F  6  03; p  05).Atstep3, thefrequencywithwhichparticipantsreferencedsocial contextduringteasinginteractionswasentered.Afterac- countingfortheoryofmind,teasingaboutsocialnormsno longerdifferentiatedparticipantsbygroup(  R 2   01; F   34; ns ). Theuseofplayfulnonverbalbehaviorsduringsocial interactionhasbeenrelatedtotheoryofmind UnderstandingTeasing 65 TableIII. TeasingBehaviorsAmongParentsandChildren Children Õ sbehaviors CriticalcontentAf Þ liativecontentCriticalbehaviorPlayfulbehavior HFAControlHFAControlHFAControlHFAControl Parent Õ sbehaviors Criticalcontent   09  69    04   46  .03   11.05.18 Af Þ liativecontent.14  57  .04.53    04.00   15   07 Criticalbehaviors   12   38  .23.27.09   14.05.54  Playfulbehaviors   18   07   11   27   09   25   15   20 Note .DatarepresentPearsoncorrelationsbetweenparentsandtheirchildren.  p  10.  p  05.  p  01  (Charman,1997).Asabove,inathree-stepregression,age andverbalIQdidnotdifferacrossgroups(  R 2   07; F  1  03; ns ),thoughtheoryofmindabilitydid(  R 2   18; F  6  03; p  05).Aspredicted,thegroupdiffer- enceinchildren Õ suseofplayfulbehaviorduringteasing interactionsdisappearedwhentheoryofmindwascon- trolled(  R 2   02; F   80; ns ).Differencesintheory ofmindthen,appeartoaccountfortheuseofplayfulac- tionswhileteasingandthetopicsaboutwhichchildren tease. DISCUSSION Whenhigh-functioningchildrenwithautismand theirparentsrecalledpriorepisodesofteasinganden- gagedinateasinginteraction,theytendedtoneglecttwo crucialcomponentsofteasing:(1)theplayfulbehaviors thatmitigatetheseriousnessofateaseand(2)theideathat teasingisasocialcommentaryaboutthebehaviorofan- otherindividual.Thisstudyhighlightsaninterestingpara- doxinteasing.Althoughcomparisongroupparticipants teasedinmorecriticalandmoralisticways,engagingin teasingaboutviolationsofsocialnormsmorefrequently thandidparticipantsintheHFAgroup,theirconceptsof teasingweremorepositiveandlessnegativethanthoseof HFAparticipants.Incontrast,themoreaf Þ liativeteasers inthestudy,thoseintheHFAgroup,experiencedmore dif Þ cultyseeingthepositivesideofteasing. Additionally,wefoundthatduringteasing,compar- isonparticipantsweremuchmorelikelytoshowrelat- ednessintheirteasingstyles,associatingtheirlevelsof provocationandplayfulness.These Þ ndingssuggestthat comparisongroupparticipantshadmorevariedandadapt- ableconceptsandstylesofteasing.Finally,theoryofmind accountedfordifferencesbetweencomparisonchildren Õ s andHFAchildren Õ suseofplayfulbehaviorwhenteas- ing,aswellasdifferencesinteasingaboutsocialnorms. Takentogether,these Þ ndingscorroboratepriorresearch, suggestingthattheoryofmindabilityaccountsfordiffer- encesinspontaneoussocialbehavior(e.g.,Tager-Flusberg, 1999),andtheunderstandingofsocialinteraction(e.g., Baron-Cohen&Hammer,1997). AutismandtheAbsenceofPlayinTeasing Aspredicted,HFAgroupparticipantsusedlessplay- fulbehaviorduringteasingthandidcomparisonpartici- pantsandtheoryofmindaccountedforthesedifferences. Thereareseveralpossibleaccountsofthis Þ nding.Re- latedtotheoryofminddif Þ culties,childrenwithautism andAsperger Õ ssyndromeengageinlessplay(Mundy etal.,1990).Inaddition,parentsofchildrenwithautism andchildrenwithautismthemselvesusefewernonver- balgesturesincommunication,havemoredif Þ cultypro- ducingcommunicativenonverbalgestures(Piven,Palmer, Jacobi,&Childress,1997),anddemonstrateapreference forhighlyliteralcommunicativedisplays(Baron-Cohen &Hammer,1997).Thesedifferencesintheuseofplay andnonverbalcommunicativebehaviormaygeneralize toteasing,therebyaccountingforourobservedde Þ citin playfuloff-recordmarkers. Alternately,theHFAgroupparticipantsteasedin lesscriticalfashion.Thecriticalcomponentofteasing, somewhatparadoxically,positivelycorrelateswiththein- creaseduseofoff-recordmarkers.Itmaybethecasethat HFAchildrenandtheirparentsdidnotteaseinplayful fashionbecausetheyteasedinamoreaf Þ liationfashion andthereforehadlesscriticalcontenttorenderplayful. Althoughour Þ ndingontherelationshipbetweentheory ofmindandplayfulbehaviorlendsplausibilitytothefor- merexplanation,themethodsutilizedinthepresentstudy donotallowustodistinguishconclusivelybetweenthem. AutismandtheSocial Ð MoralContentofTeasing Referencestogroupandsocialnorms,andviolations thereof:oftencomprisethemajorgrounduponwhicha 66 Heerey,Capps,Keltner,andKring teaseisconstructed(Eder,1993).TheteasingofHFA childrenandtheirparentslackedthissocialcontextual richness.Relativetocomparisonparticipants,HFAgroup participantsreferredlessfrequentlytoantecedentsand consequencesintheirrecountedexperiencesofteasing. Additionally,theirteasingcenteredlessfrequentlyonso- cialnormviolations.Thispatternofresultsisquitecon- sistentwithpreviousstudiesthathavedocumentedtheas- sociationbetweenautismandde Þ citsinsocialcontextual understandingincludingdiminishedsocial-referencing behaviors,de Þ citsinjointattention,anddecreaseduseof communicativenonverbalbehaviors(Capps&Sigman, 1996;Frith,1989).Moreover,theyhavedif Þ cultylink- ingemotionstothesocialsituationsinwhichtheyoccur (Capps,Yirmiya,&Sigman,1992),andshowde Þ cient understandingofself-consciousemotions,inwhichan- other Õ ssocialevaluationoftheselfplaysarole(Heerey, Keltner&Capps,2003). Thetendencyforchildrenwithautismtoignorethe antecedentsandconsequencesofteasing,anditsnorm- relatedfocusmoregenerally,mayhelpexplaintheanec- dotalreportsofthedif Þ cultychildrenwithautismfacein learningsocialnormsandabidingbysocialrules(Klinger &Dawson,2001).Thepresentcontext,afriendlyparent Ð childinteraction,islikelyquitedissimilartothepeerand siblingteasingmostchildrenexperience.Nonetheless,this studysuggeststhatteasingdoesnotofferchildrenwith autismanarenaforlearningsocialnorms,asitdoesfor typicallydevelopingchildren. Inthisvein,manyofthehigh-functioningchildren withautismandAsperger Õ ssyndromeinoursamplespon- taneouslyreportedthattheydidnotknowwhypeople teasedthemandoftendidnotevenunderstandthatthey werebeingteaseduntiltheybegantofeel Ò bad. Ó Whereas anumberofcomparisongroupchildrenreportedtheexpe- rienceofembarrassmentasaconsequenceofteasing,only twoHFAgroupchildrendidso.Findingsamongparents paralleledthoseoftheirchildren,suggestingthatparents ofHFAchildrenmayalsohavedif Þ cultyspontaneously linkingteasingwiththesocialcontextFinally,onthebasis ofteasinginterviews,itappearsthatHFAchildreninitiate teasinglessfrequentlyandmayhavefewersocialrelation- shipswithinwhichteasingisappropriate.Takentogether, these Þ ndingssuggestthatcomparisongroupmembers haveamorethoroughlydevelopedunderstandingofthe functionsofteasinginsocialcontextandofthewaysin whichteasingmayrelatetosocialrelationships. TeasingConcepts Conceptsofteasingarerelatedtoteasingexperi- encesandbothconceptsandexperiencesprogresswith socialandlinguisticdevelopmentAssuchdevelopment occurs,positiveconceptsofteasingbegintocoexistwith morenegativeones,allowingindividualstorecognizethe bene Þ tsofteasing.Thechildreninourstudywerejust attheagewhenthisdevelopmentalshifttypicallytakes place(Warm,1997).Accordingly,comparisonchildren evidencedbothpositiveandnegativeteasingconcepts whereastheHFAgroupchildrenhaddif Þ cultyappreciat- ingthemorepositiveaspects.Parents Õ conceptsmirrored thoseoftheirchildren.BothcomparisonandHFAgroup parentsunderstoodthatteasingcouldbepositive,how- evertheteasingconceptsofparentsofHFAchildrenwere muchmorenegative,suggestingthatalthoughtheycon- ceptuallyunderstandthatteasinghaspositiveside,they appeartointerpretitmorenegatively. CONCLUSION Inteasingoneanother,peoplelearnaboutsocial norms,roles,andexpectations.Theyexpressaffectionand explorepossiblerelationships.Theydiscovertheprefer- ences,attitudes,andbeliefsofothers.Theoryofmindand theabilitytounderstandsocialintentionsunderpinteasing andmuchofthesocialdiscourseinwhichhumansengage, anideaillustratedinthecoemergenceofearlyteasingand theoryofmindinyoungchildren(e.g.,Dunn&Munn, 1985).Our Þ ndingshintatthedif Þ cultyfacedbychil- drenwithautismandAsperger Õ ssyndromeastheytryto navigatesocialrelationshipsand Þ tintotheirsocialenvi- ronments.Unfortunately,thesedif Þ cultiesarenotunique toautismspectrumdisorders.Onthebasisoftheabili- tiesnecessarytounderstandteasingwewouldpredictthat teasingdif Þ cultieswouldextendtoanyindividualswho experiencedif Þ cultywiththeoryofmind-relatedskills, suchaschildrenwithattentionde Þ cithyperactivitydis- order(Hinshaw,1987).Becausethesedif Þ cultiesareso characteristicofchildrenwithautism,thefundamentalso- cialinteractionofteasingmayeludeeventhebestefforts ofthesechildren,despitetheirlaborstounderstandit. ACKNOWLEDGMENT WewishtothankDrDavidSobelforhiscontribution ofthetheoryofminddatasetaswellashisastutecom- mentarythroughoutthisproject.Thisworkisdedicatedto thememoryofLisaM.Capps. REFERENCES Abell,F.,Happe,F.,&Frith,U.(2000).Dotrianglesplaytricks?Attri- butionofmentalstatestoanimatedshapesinnormalandabnormal development. CognitiveDevelopment , 15, 1 Ð 16. UnderstandingTeasing 67 Bara,B.G.,Bosco,F.M.,&Bucciarelli,M.(1999).Developmental pragmaticsinnormalandabnormalchildren. BrainandLanguage , 68, 507 Ð 528. Baron-Cohen,S.,&Hammer,J.(1997).Isautismanextremeform ofthe Ò malebrain Ó ? AdvancesinInfancyResearch , 11, 193 Ð 217. Baron-Cohen,S.,O Õ Riordan,M.,Stone,V.,Jones,R.,&Plaisted,K. (1999).Recognitionoffauxpasbynormallydevelopingchildren withaspergersyndromeorhigh-functioningautism. Journalof AutismandDevelopmentalDisorders , 29, 407 Ð 418. Barron,J.,&Barron,S.(1992). There Õ saboyinhere .NewYork:Simon &Schuster. Bell,R.A.,Buerkel-Rothfuss,N.L.,&Gore,K.E.(1987). Ò Didyou bringtheyarmulkeforthecabbagepatchkid? Ó Theidiomaticcom- municationofyounglovers. HumanCommunicationResearch , 14, 47 Ð 67. Bell,R.A.,&Healey,J.G.(1992).Idiomaticcommunicationandinter- personalsolidarityinfriends Õ relationalcultures. HumanCommu- nicationResearch , 18, 307 Ð 335. Capps,L.,&Sigman,M.(1996).Autisticaloneness.InR.D.Kavanaugh, B.Zimmerberg,etal.(Eds.), Emotion:Interdisciplinaryperspec- tives (pp.273 Ð 296).Hillsdale,NJ:Erlbaum. Capps,L.,Yirmiya,N.,&Sigman,M.(1992).Understandingofsimple andcomplexemotionsinnon-retardedchildrenwithautism. Jour- nalofChildPsychologyandPsychiatryandAlliedDisciplines , 33, 1169 Ð 1182. Charman,T.(1997).Therelationshipbetweenjointattentionandpretend playinautism. DevelopmentandPsychopathology , 9, 1 Ð 16. Charman,T.,&Baron-Cohen,S.(1997).Briefreport:Promptedpretend playinautism. JournalofAutismandDevelopmentalDisorders , 27, 325 Ð 332. Clark,H.H.(1996).Communities,rommonalities,andcommunica- tion.InJ.J.Gumperz&S.C.Levinson(Eds.), Rethinkinglin- guisticrelativity (pp.324 Ð 355).NewYork:CambridgeUniversity Press. Clark,H.H.,&Gerrig,R.J.(1984).Onthepretensetheoryofirony. JournalofExperimentalPsychology:General , 113, 121 Ð 126. Dunn,J.F.,&Munn,P.(1985).Becomingafamilymember:Family con ß ictandthedevelopmentofsocialunderstandinginthesecond year. ChildDevelopment , 56, 480 Ð 492. Eder,D.(1993). Ò GogetyaaFrench! Ó :Romanticandsexualteasing amongadolescentgirls.InD.Tannen(Ed.), Genderandconver- sationalinteraction (pp.17 Ð 31).NewYork:OxfordUniversity Press. Eisenberg,A.R.(1986).Teasing:VerbalplayintwoMexicanohomes.In B.B.Schieffelin&E.Ochs&etal.(Eds.), Languagesocialization acrosscultures (pp.182 Ð 198).NewYork:CambridgeUniversity Press. Flavell,J.H.(1999).Cognitivedevelopment:Children Õ sknowledge aboutthemind. AnnualReviewofPsychology , 50, 21 Ð 45. Frith,U.(1989).Autismand Ò theoryofmind. Ó InC.Gillberg(Ed.), Diagnosisandtreatmentofautism (pp.33 Ð 52).NewYork:Plenum. Frith,U.,Happe,F.,&Siddons,F.(1994).Autismandtheoryofmind ineverydaylife. SocialDevelopment , 3, 108 Ð 124. Glass,G.V.,&Hopkins,K.D.(1996). Statisticalmethodsineduca- tionandpsychology (3rded.).NeedhamHeights,MA:Allyn& Bacon. Grandin,T.(1995).Howpeoplewithautismthink.InE.Schopler&G.B. Mesibov(Eds.), Learningandcognitioninautism .(pp.137 Ð 156). NewYork:Plenum. Happe,F.G.E.(1993).Communicativecompetenceandtheoryofmind inautism:Atestofrelevancetheory. Cognition , 48, 101 Ð 119. Happe,F.G.E.(1994).Anadvancedtestoftheoryofmind:Under- standingofstorycharacters Õ thoughtsandfeelingsbyableautistic, mentallyhandicapped,andnormalchildrenandadults. Journalof AutismandDevelopmentalDisorders , 24, 129 Ð 154. Heerey,E.A.,Keltner,D.,andCapps,L.M.(2003).Makingsenseof self-consciousemotion:Linkingtheoryofmindandemotionin childrenwithautism, Emotion , 3, 394 Ð 400. Hinshaw,S.P.(1987).Onthedistinctionbetweenattentional de Þ cits/hyperactivityandconductproblems/aggressionin childpsychopathology. PsychologicalBulletin , 101, 443 Ð 463. Hobson,J.A.(Ed.).(1989). Abnormalstatesofbrainandmind . Cambridge,MA:Birkhaeuser. Jarrold,C.,Boucher,J.,&Smith,P.(1993).Symbolicplayinautism:A review. JournalofAutismandDevelopmentalDisorders , 23, 281 Ð 307. Keltner,D.,Capps,L.,Kring,A.M.,Young,R.C.,&Heerey,E.A. (2001).Justteasing:Aconceptualanalysisandempiricalreview. PsychologicalBulletin , 127, 229 Ð 248. Keltner,D.,Young,R.C.,Heerey,E.A.,Oemig,C.,&Monarch,N.D. (1998).Teasinginhierarchicalandintimaterelations. Journalof PersonalitvandSocialPsychology , 75, 1231 Ð 1247. Klinger;L.G.,&Dawson,G.(2001).Prototypefonnationinautism. DevelopmentandPsychopathology , 13, 111 Ð 124. Leslie,A.M.(1987).Pretenseandrepresentation:Theoriginsof Ò theory ofmind. Ó PsychologicalReview , 94, 412 Ð 426. Lightner,R.M.,Bollmer,J.M.,Harris,M.J.,Milich,R.,&Scambler, D.J.(2000).Whatdoyousaytoteasers?Parentandchildevalu- ationsofresponsestoteasing. JournalofAppliedDevelopmental Psychology , 21, 403 Ð 427. Lord,C.,Rutter,M.,&LeCouteur,A.(1994).AutismDiagnostic Interview-Revised:Arevisedversionofadiagnosticinterviewfor caregiversofindividualswithpossiblepervasiyedevelopmentaldis- orders. JournalofAutismandDevelopmentalDisorders , 24, 659 Ð 685. Loveland,K.A.,Pearson,D.A.,Tunali-Kotoski,B.,Ortegon,J.,& Gibbs,M.C.(2001).Judgementsofsocialappropriatenessbychil- drenandadolescentswithautism. JournalofAutism&Develop- mentalDisorders , 31, 367 Ð 376. McDonnell,J.T.(1993). Newsfromtheborder:Amother Õ smemoirof herautisticson .NewYork:TicknorandFields. Mundy,P.,Sigman,M.,&Kasari,C.(1990).Alongitudinalstudyofjoint attentionandlanguagedevelopmentinautisticchildren. Journalof AutismandDevelopmentalDisorders , 20, 115 Ð 128. Ozonoff,S.,Rogers,S.J.,&Pennington,B.F.(1991).Asperger Õ ssyn- drome:Evidenceofanempiricaldistinctionfromhigh-functioning autism. JournalofChildPsychologyandPsychiatryandAllied Disciplines , 32, 1107 Ð 1122. Perner,J.,Frith,U.,Leslie,A.M.,&Leekam,S.R.(1989).Exploration oftheautisticchild Õ stheoryofmind:Knowledge,belief,andcom- munication. ChildDevelopment , 60, 689 Ð 700 Pilowsky,T.,Yirmiya,N.,Shulman,C.,&Dover,R.(1998).TheAutism DiagnosticInterview-RevisedandtheChildhoodAutismRating Scale:Differencesbetweendiagnosticsystemandcomparisonbe- tweengenderes. JournalofAutismandDevelopmentalDisorders , 28, 143 Ð 151. Piven,J.,Palmer,P.,Jacobi,D.,&Childress,D.(1997).Broaderautism phenotype:Evidencefromafamilyhistorystudyofmultiple- incidenceautismfamilies. AmericanJournalofPsychiatry , 154, 185 Ð 190. Shapiro,J.P.,Baumeister,R.F.,&Kessler,J.W.(1991).Athree- componentmodelofchildren Õ steasing:Aggression,humor,and ambiguity. JournalofSocialandClinicalPsychology , 10, 459 Ð 472. Siegel,B.(1986).Empiricallyderivedsubclassi Þ cationoftheautistic syndrome. JournalofAutismandDevelopmentalDisorders , 16, 275 Ð 293. Smith,P.K.,&Brain,P.(2000).Bullyinginschools:Lessonsfromtwo decadesofresearch. AggressiveBehavior , 26, 1 Ð 9. Sobel,D.M.,Capps,L.M.,&Gopnik,A.(1999,May).Explor- ingtherelationshipbetweenautisticchildren Õ sperceptionofam- biguous Þ guresandtheoryofmind.Paperpresentedatthean- nualmeetingoftheSocietyforResearchinChildDevelopment, Denver,CO. Strandburg,R.J.,Marsh,J.T.,Brown,W.S.,&Asamow,R.F. (1993).Event-relatedpotentialsinhigh-functioningadultautistics: 68 Heerey,Capps,Keltner,andKring Linguisticandnonlinguisticvisualinformationprocessingtasks. Neuropsychologia , 31, 413 Ð 434. Szatmari,P.,Offord,D.R.,Siegel,L.S.,&Finlayson,M.A.(1990).The clinicalsigni Þ canceofneurocognitiveimpairmentsamongchildren withpsychiatricdisorders:Diagnosisandsituationalspeci Þ city. JournalofChildPsychologyandPsychiatryandAlliedDisciplines , 31, 287 Ð 299. Tager-Flusberg,H.(1999).Apsychologicalapproachtounderstanding thesocialandlanguageimpairmentsinautism. InternationalRe- viewofPsychiatry , 11, 325 Ð 334. Warm,T.R.(1997).Theroleofteasingindevelopmentandviceversa. JournalofDevelopmentalandBehavioralPediatrics , 18, 97 Ð 101. Wechsler,D.(1991). WechslerIntelligenceScaleforChildren (3rded.). SanAntonio,TX:PsychologicalCorporation.