Negative social comparisons and psychosis proneness in a healthy adolescent population

(1)

Original

article

Negative

social

comparisons

and

psychosis

proneness

in

a

healthy

adolescent

population

F.A.

Cotier

a

,

T. Toulopoulou

a,b,c,

*

a_Department_of_Psychology,_the_University_of_Hong_Kong,_Pokfulam_Road,_Hong_Kong_SAR,_China b_Department_of_Psychology,_Bilkent_University,_Main_Campus,_Bilkent,_Ankara,_Turkey c

DepartmentofBasicandClinicalNeuroscience,InstituteofPsychiatry,PsychologyandNeuroscience,King’sCollegeLondon,16DeCrespignyPark,London SE58AF,UK

1. Introduction

Severalinﬂuentialtheoriesofmentaldisorder,andinparticular psychosis, have highlighted the potentially important role of negative social comparisons (NSC). Gilbert and Allan [1], for example,suggestedthatindividualswhoperceivethemselvesas more subordinate or lower ranking compared to others are at greater risk of developing mental disorder. Similarly, in their introductiontotheSocialDefeatHypothesis,Seltenetal.propose perceptions of holding ‘‘a subordinate position’’ or ‘‘outsider status’’ to be signiﬁcant risk factors for the development of psychosis[2].Morerecently,inrelationtopsychosis,researchers

have proposed a paranoia hierarchyin which social evaluative concerns,orfeelingsofinadequacycomparedtoothers,formthe basisoftheparanoiaexperience[3,4].

There is now growing evidence to support these proposals implicatingNSCinthedevelopmentofpsychosis.Negativesocial comparativebeliefs,forexample,havebeenfoundtobeassociated withhallucinations,suchaspatients’perceptionsofthepowerof theirvoices[5–7].Socialrank(oneformofNSC)hasalsobeenreported topredictparanoiainthegeneralpopulation[8]andplayakeyrolein thetransitionfromsubclinicaltoclinicallevelsofparanoia[3,4].Most recently, NSC have been shown to predict positive and negative psychotic experiences, and play a mediatory role between social adversity and negative psychotic experiences [9]. Using advanced modelling techniques andlongitudinal approaches,several studies havealsoexaminedthedirectionalityoftheseeffects,andfoundNSC to exist before the emergence of these subclinical psychotic ARTICLE INFO

Articlehistory: Received2June2017

Receivedinrevisedform21August2017 Accepted21August2017

Availableonline1September2017 Keywords: Socialrank Group-ﬁt Belonging Psychosis Twins ABSTRACT

Thereis growingevidenceof anassociation between negativesocial comparisons(NSC)and both

psychosis,andpsychosisproneness.Themajorityoftheworkthusfar,however,hasfocusedlargelyon

onetypeofNSC,namely,socialrank.Whilstsocialrankisclearlyanimportantfactor,anindividual’s

perceptionofbelongingislikelyalsoofimportance;particularly,whenconsideringindividualsfrom

collectivisticculturessuchasChina,wheregreateremphasisisplacedonﬁttingintothegroup.Thereis

alsolimitedresearchinvestigatingwhatfactorsmaycontributetowardstherelationshipbetweenNSC

andpsychosisproneness,andtowhatextentthisrelationshipmaybeduetocommonfamilialfactors.To

addresstheseissues,weexaminedwhether(1)Socialrankandperceivedbelongingpredictnegative,

positiveanddepressivepsychoticexperiencesinaChinese,adolescent,twinandsiblingpopulation,(2)

copingstylesmoderatetheimpactoftheserelationshipsand(3),thereisafamilialassociationbetween

NSCandpsychosisproneness. Bothsocialrankandperceivedbelongingwerefoundto predictthe

negativeanddepressivedimensionsofpsychosis.Theserelationshipsweremoderatedby

problem-focusedcopingstyles.Interestingly, theassociation betweenperceptionofbelonging,andnegative

psychoticexperiences wasfamilial—andstrongerin Monozygotictwins—indicatingperhapsshared

aetiologyduetocommongenes.Our ﬁndingshighlightNSCaspotentialvulnerability markersfor

negativeanddepressivepsychoticexperiences,andsuggestpotentiallydifferentaetiologicalpathways

amongstdifferentNSCanddifferentpsychoticexperiences.Onaclinicallevel,ourﬁndingsemphasize

theneedtoconsidercopingstyleswhentreatingat-riskindividuals.

C 2017ElsevierMassonSAS.Allrightsreserved.

* Correspondingauthor.Department ofPsychology,BilkentUniversity,Main Campus,Bilkent,Ankara,Turkey.Fax:+903122664960.

E-mailaddress:timothea.toulopoulou@kcl.ac.uk(T.Toulopoulou).

ContentslistsavailableatScienceDirect

European

Psychiatry

j our na l ho me p a ge : ht t p: / / w ww . e ur opsy -j ou rna l . c om

http://dx.doi.org/10.1016/j.eurpsy.2017.08.006

(2)

experiences[e.g.10].ThissuggeststhatNSCmayrepresentapotential vulnerabilitymarkerforthedevelopmentofpsychosis.

Althoughthesestudiesprovideimportantpreliminaryevidence intotheassociationbetweenNSCandpsychosis,themajorityhave tendedtoeitheruseonlythesocial‘‘rank’’scale([e.g.4])oratotal scorecomprisedofthetwosubscales(socialrankandbelonging). Whilstsocialrank,oranindividual’sbeliefsregardingtheirsocial standingorpositionin comparisontoothers [11], isclearly an importantfactor,anindividual’sperceptionsofbelongingorﬁtting into‘‘thegroup’’–anotherformofNSC–seemsequallyimportant toconsider.Thismaybeparticularlysigniﬁcantwhenconsidering individualsfrom collectivisticcultures, such as China,who are thoughtto have more interdependentself-construals and thus placegreateremphasisonthegroupandinterpersonalrelations

[12], and consequently may be more impacted by lower perceptionsofbelonging.Basedonthesepredictions,wesought to investigate how different types of NSC beliefs (rank and belonging) relate to the different dimensions of psychosis proneness (positive, negative, depressive) within a Chinese population.

In addition to further clarifying the link between NSC and psychosisproneness,itisimportanttodeterminewhatfactorsmay act as moderators of this relationship, as these may represent importanttargetsforintervention.Thisisparticularlyimportant giventhefindings ofa recentstudy,which suggestedthat NSC beliefs alone may not be sufficient to induce paranoia. In an interestingstudy,Asconeetal.[13]inducedfeelingsoflowerrank inparticipantsthrough theuseof anexperimental task, which mimickedsocial networkinteractions, and foundsocial rank to affect sadness, but not paranoia. Furthermore, a significant moderationeffectoffearofrejectionbetweentheexperimental conditions(i.e.pre-andpost-experimentalmanipulationofsocial rank) was found. These findings highlight the importance of consideringwhichfactorsmayexacerbateorbuffertheimpactof NSConthedevelopmentofpsychosis.

Onefactorthatmayplayamoderatingroleisthewayinwhich anindividualcopes,ordeals,withtheirNSCbeliefs.Copingstyle seemstoplayacrucialroleinthedevelopmentandmaintenanceof psychoticsymptoms in psychosisprone and at-risk individuals

[14–18],andhenceitisimportanttoidentifytheirpreciserolein therelationshipbetweenNSCandpsychoticexperiences. Howev-er,toourknowledge,nostudytodatehasspeciﬁcallyexamined theroleofcopingstrategiesinthisrelationship.

Coping styles arecommonly classiﬁed intoemotion-focused andproblem-focusedstrategies[19].Emotion-focusedstrategies aimat soothing thenegative associated emotions and moving away from the problem (e.g. thinking reassuring thoughts). Problem-focusedstrategiesaiminsteadatactivelytryingtosolve theproblemathandandfocusonthesourceoftheproblem(e.g. engaginginsocialactivitiestocopewithfeelingsofisolationor seekingsocialsupporttofeelincluded)[20,21].Problem-focused strategieshavebeenfoundtobeassociatedwithbettermental healthoutcomesforadolescents[20]andmoreeffectivedealing withdaily stressors [22]. Yet, individuals withearly psychosis generallyreportmoreemotion-focusedcopingstrategies[22].We therefore sought to examine the role of coping styles in the relationshipbetweenNSCandpsychosisproneness. Understand-ingwhetherdifferentcopingstylesmayexacerbateorbufferthe harmfuleffectsofNSConpsychosisproneness,willallowformore effectiveandtargetedinterventions.

Finally,whenexaminingvulnerabilitymarkersforpsychosisit is also important to consider possible familial inﬂuences. Speciﬁcally,whetherfamilial(geneticandsharedenvironmental) ornon-shared environmental factorsmayexplain this relation-ship,therebyopeningupnewavenuesfortranslationalresearch. Whilstlittleisknownabouttheextenttowhichgeneticandshared

environmentalinfluences(i.e.factorssharedwithinfamilies)may contributetonegativesocialcomparisons,researchsuggestsahigh percentageofvarianceinself-esteem—acloselyrelatedconcept— tobeexplainedbysuchfactors[23].Furthermore,thereissome evidencethatotherfactors,whichmaybelinkedwithnegative socialcomparisons,suchasstressreactivity,maycongregatewith psychosisproneness[24].Thatis,stressreactivityandpsychosis pronenessseemtobeexplainablebycommongeneticinfluences.It therefore seems possible that negative social comparisons and psychosispronenessmayalsosharecommonetiologicalinfluences includingsharedenvironmental(andpotentiallygenetic)factors. Wethereforealsoexaminedtheextenttowhichtherelationship betweensocialcomparisonsandpsychosispronenessisinfluenced byfamilialandnon-familialfactors.Iffamilial,thiscouldindicate that the two share, to some extent, their genetic or shared environmentalaetiologicalfactors.

Aimsofthestudy:

todeterminehowdifferenttypesofsocialcomparativebeliefs (i.e. rank and belonging) are associated with the three dimensionsofpsychosisproneness(i.e.positive,negativeand depressive)inaChinesepopulation;

toexaminewhethercopingstylesmoderatetheimpactofNSC onpsychosisproneness;

toexplorewhetherthereisafamilialassociationbetweenNSC andpsychosisproneness.

We hypothesized that NSC will predict all three forms of psychotic experiences (negative, positive and affective). We further hypothesized that problem-focused coping styles (e.g. seekingsocial support)willbuffertheharmfuleffectofNSCon psychosis proneness, and that emotion-focused strategies (e.g. thinking reassuring thoughts)will exacerbate this relationship. Finally,wehypothesizedthatNSC(bothbelongingandrank)will share familial aetiology with all three forms of psychotic experiences.

2. Materialandmethods 2.1. Procedure

EthicalapprovalwasobtainedfromtheHumanResearchEthics Committeeof theUniversity of HongKong. Participantssigned writtenconsentbeforebeginningtheprotocol,andinthecaseof minors,alegalguardianalsoprovidedwrittenconsent.

2.2. Participants

The sample included healthy, Chinese adolescents from the generalpopulation.ParticipantswererecruitedfromHongKong, Beijing, Guangzhou, Nanjing and Qingdao. Participants from BeijingwererecruitedthroughtheChineseNationaltwinregister. ParticipantsfromHongKong,andallotherlocationsinMainland Chinawererecruitedthroughsecondaryschoolsanduniversities. Schools and universitiesfrom many regionsof HongKong and Mainland China were included to ensure a sample truly representativeofChineseadolescents.Onlynon-clinical(i.e.those withno history ofmental disorderor current mental disorder) participantswereincluded.OnlyindividualsofChineseethnicity whowerenativeChinesespeakerswereincluded.

The initial sample consisted of 201 individuals including 76twinpairsand49siblings.Thirteenindividualswereremoved from thesample due to missing data, thus resulting in a ﬁnal sample of188 participantsincluding44 siblings,70 twinpairs (MZ=37,DZ=25,Unknown=8),3twinswithouttheirco-twins

(3)

(1DZ,2MZ) and1 lonesibling.For thequestionoffamiliarity (sample2),onlythetwinpairswithcompletezygosityinformation wereincluded(n=62).Anadditional69twinpairswererecruited forthepurposeofthisquestion,resultinginatotalof131(MZ=83, DZ=46)twinpairsforthefamiliarityanalyses(Table1). 2.3. Measures

Allinstrumentsweretranslated,back-translatedandadapted forusewithChineseparticipantsaccordingtoexistingguidelines

[25].

2.3.1. Socialcomparisons

SocialcomparisonswereassessedwiththeSocialComparison Scale(SCS)[11].TheSCSemploysasemanticdifferentialapproach toassessanindividual’sperceivedsocialrankandbelongingness. Thescaleconsistsof11bipolaritems(e.g.Inferior–Superior).Items areratedona10-pointlikertscalewithlowerscoresindicating lowerperceivedbelongingnessandrank.Thetotalscoreforthe socialcomparisonmeasure,whichwasthesumofallitems,was calculated.Separatevariablesforbelongingandsocialrankwere alsocomputed.TheCronbachalphaforthetotalscalewas0.92and 0.87(rank)and0.87(belonging)forthetwosubscales.

2.3.2. Psychosisproneness

Self-reportedsubclinicalpsychoticexperienceswereassessed usingtheChineseversionoftheCommunityAssessmentofPsychic Experiences (CAPE-C15) [26–28]. The translated version of the CAPE includes three subscales (Positive, Negative, Depressive) consistingof4,5and6itemsrespectively.Anexampleitemis:‘‘Do

youeverfeelthereisaconspiracyagainstyou?/

FurtherdescriptionoftheCAPE-C15andthetranslationprocess canbefoundinourlabspreviousreport[26].Inthecurrentsample theCAPE-C15reachedaninternalconsistencyofCronbachAlpha 0.86, representing good/excellent reliability, and the positive, negativeanddepressivesubscaleshadaCronbachalphacoefﬁcient of0.79,0.71,and0.79,respectively.Frequencyscoreswererated ona4-pointLikertscalerangingfromnevertonearlyalways.Mean frequencyscoreswerecalculatedforeachofthethreedimensions withhigherscoresindicatinggreaterpronenesstothatdimension.

In order to normalize the distribution of the mean frequency scores,logtransformationswereperformed.

2.3.3. Copingstyles

The Utrecht Coping list (UCL) (Schreurs, Van de Willige, Tellegen,&Brosschot,1993)wasadoptedheretomeasurecoping styles.This44-iteminstrumentmeasuressevenscalesincluding: activetackling,seekingsocialsupport,palliativereacting,avoiding, passivereacting,reassuringthoughtsandexpressionofemotions (seeTable3). Ofthesetheﬁrsttwoareconsideredas problem-focusedcopingstrategies(i.e.activelytryingtosolvetheproblem) andthelatter5areconsideredtobeemotion-focused(i.e.soothing negative associated emotions through moving away from the problem). Participants wereasked to indicatehow they would typicallyrespondtoaproblem.Forexample,‘‘Youseeproblemsas

a challenge/ ’’

(Prob-lem-focused) and ‘‘You avoid difﬁcult situations as much as

possible/ ’’

(Emotion-focused). Theinternal consistency ofthis questionnairefor the current sample was good to excellent with a Cronbach Alpha coefﬁcientof0.90forthetotalscale.Thereliabilitycoefﬁcientsfor theindividualscaleswere0.84(activetackling),0.69(palliative reacting), 0.64 (avoiding), 0.81 (seeking social support) 0.76 (depressive reactivepattern),0.61(expressionof emotions)and 0.62 (reassuring thoughts), which werein line withpreviously establishedinternalconsistencyestimatesforwesternpopulations (Schreursetal.,1993).

2.4. Statisticalanalyses

All analyses were conducted in SPSS 20 for Mac. Given the hierarchical structure of the data, multilevel regression analysis was performed with individuals (level 1) nested within-twin pairs (level 2) for all cross-trait, within-twin analyses. All hierarchical models were conducted using the MIXEDcommand.

2.4.1. Associationsbetweennegativesocialcomparisonsand psychoticexperiencesandcopingstylesasmoderators(cross-trait, within-twinanalyses)

2.4.1.1. Main effects. In order to examine the main effects of perceived social comparison on psychosis proneness within individuals, separate multilevel models were conducted with the three measures of psychosis proneness as the dependent variables(DV’s:positive,negativeanddepressive)andthethree measuresofsocialcomparisonasindependentvariables(IV:total SCS,belongingness,rank).

2.4.1.2. Moderationanalyses. Inordertoinvestigatethe moderat-ing roleof copingstrategies on therelationshipbetweensocial comparisons and psychoticexperiences, thesevencoping style measureswereindividuallyaddedtotheTotalSocialComparison within person models as moderators. The primary focus was therefore theinteractionterm. Moderatorswere only addedto models that were foundto be signiﬁcantin theﬁrst round of analyses.

2.4.2. Familialassociationsbetweennegativesocialcomparisonsand psychoticexperiences(cross-trait,cross-twinassociations)

2.4.2.1. Main effects. Cross-trait cross-twin associations were examined through the use of simple linear regression models. Speciﬁcally, we assessed whether the three social comparison measuresintwin1wouldpredictthethreedimensionofpsychosis proneness intwin2.The IV’swerethethreesocial comparison variables in twin1 andthe DV’swerethethreedimensions of Table1 Descriptivestatistics. Sample1(n=188) Sample2(n=262) Site HongKonga ₁₃₁₍₃₀₎a ₆₈₍₂₁₎a Beijinga 57(70)a 54(26)a Nanjing n/a 8(24) Guangzhou n/a 64(3) Qingdao n/a 68(26) Age,mean(SD) 17.67(1.93) 17.66(1.89) Sex,n(%) Female 124(66) 167(64) Male 64(34) 95(36)

CAPEnegativedimension,mean(SD) 2.07(0.59) 1.90(0.55) CAPEpositivedimension 1.32(0.52) 1.25(0.40) CAPEdepressivedimension 1.95(0.54) 1.80(0.47) SCStotal 60.10(14.2) 61.53(13.31) SCSbelonging 30.32(7.51) 31.34(7.31) SCSrank 29.91(7.51) 30.19(7.31) SD: standard deviation; SCS: Social Comparison Scale; CAPE: Community AssessmentofPsychicExperiences;n/a:notapplicable.

a

Indicateswheretherewasoverlapinthetwosamples.

(4)

psychosisintwin2.TwinswererandomlyassignedanIDofeither ‘‘twin1’’or‘‘twin2’’.Modelswereadjustedforthecorresponding psychosisdimensionintwin1[24].

2.4.2.2. Moderationanalyses. Inordertoexaminewhethergenetic factorsmayexplainanyfamilialassociationsfound,weexamined thetwo-wayinteractionbetweenthedifferentformsofnegative social comparisons in twin 1 (those which were found to be signiﬁcantpredictors)Zygosity(withtwolevels;MZ&DZ).In thecasethata strongerfamilialassociationwasfoundbetween twin1 negative social comparisons and twin 2 dimensions of psychosisinMZtwinsitcanbeassumedthatageneticcorrelation exists. Zygosity was only examined as a moderator in those relationshipsthatwerefoundtobefamilialinnature.

3. Results

3.1. Associationsbetweennegativesocialcomparisonsandpsychotic experiencesandcopingstylesasmoderators

3.1.1. Maineffects

Thecross-traitwithin-twinanalysesrevealedthatallmeasures of social comparison signiﬁcantly predicted the negative and depressivedimensionsofpsychosisproneness (Table2). Whilst there was a non-statistical trend for the association between belongingnessandthepositivedimension,neitherofthemeasures of social comparison were found to signiﬁcantly predict the positive psychosis dimension (Table2). These resultstherefore suggestthatNSCareassociatedwiththenegativeanddepressive dimensionsofpsychosisproneness.

3.1.2. Moderationanalyses

The moderation analysis revealed active coping to be a signiﬁcantmoderatoroftherelationshipbetweenperceivedsocial comparisons,andboththenegativeanddepressivedimensionsof psychosisproneness.Socialsupportwasfoundtobeamoderator fortheassociationbetweenNSCand thedepressivedimension.

Morespeciﬁcally,bothsocialsupportandactivecopingwerefound tobuffertheharmfuleffectsofNSConpsychosisproneness.None of the emotion-focused coping strategies were foundto act as signiﬁcantmoderators(seeTable3fordetailedresults).

3.2. Familialassociationsbetweennegativesocialcomparisonsand psychoticexperiences

3.2.1. Maineffects

The cross-trait,cross-twin analysesrevealedall three mea-suresofperceivedsocialcomparisonsintwin1tosignificantly predictnegativepsychoticexperiencesintheco-twin.Senseof belonging in twin 1 also significantly predicted depressive psychotic experiences in the co-twin. Sense of belonging and total social comparisons in twin 1 remained statistically significantpredictorsofnegative psychoticexperiencesintwin 2 after adjusting for negative psychotic experiences in twin 1. Sense of belonging in twin 1 did not, however, remain a statistically significantpredictor ofdepressive psychotic expe-riences in twin 2 after controlling for depressive psychotic experiencesintwin1(seeTable4).

3.2.2. Moderationanalyses

Zygosity was found to be a significant moderator of the relationship between total social comparisons in twin 1 and negative psychotic experiences in twin 2 with a statistically significant increase in total variation explained of 6.7%, F(1, 126)=9.659,P=0.002.Specifically,theassociationwasstrongerin MZtwins(B= 0.10;P<0.01,95%CI[ 0.015, 0.005])compared toDZtwins(B=0.001;P=0.763,95%CI[ 0.004,0.006]).Similarly, zygosity also moderated the relationship between sense of belongingintwin1 andnegativepsychoticexperiencesintwin 2 as evidenced by a statistically significant increase in total variation explained of 5.4%, F(1, 125)=7.984, P=0.005 with a greaterassociationfoundinMZtwins(B= 0.020;P<0.01,95%CI [ 0.028, 0.011])compared toDZtwins(B= 0.001; P=0.809, 95%CI[ 0.011,0.009]).

Table2

Associationsbetweennegativesocialcomparisonsandpsychoticexperiences. CAPEpositive Dimension CAPEnegative Dimension CAPEdepressive Dimension B 95%CI P B 95%CI P B 95%CI P Rank 0.004 0.010to0.002 0.180 0.014 0.019to 0.009 0.000 0.012 0.018to 0.008 0.000 Belonging 0.005 0.01to0.0006 0.080 0.015 0.02to 0.01 0.000 0.012 0.017to 0.007 0.000 TotalSCS 0.002 0.01to0.18 0.125 0.008 0.01to 0.005 0.000 0.007 0.009to 0.004 0.000 SCS:SocialComparisonScale;CAPE:CommunityAssessmentofPsychicExperiences.TheboldtextindicatesthatthePvalueis<0.05,andthussigniﬁcant.

Table3

Resultsofmoderationanalysis.

Mean SD Min–Max CAPEnegative Dimension CAPEdepressive Dimension B 95%CI P B 95%CI P Problem-focused Activetackling 18.37 3.79 8–28 0.0008 0.00009to0.002 0.027 0.0008 0.0001to0.001 0.017 Seekingsocialsupport 15.12 3.54 7–24 0.0007 4.5233to0.0015 0.065 0.001 0.0003to0.002 0.003 Emotion-focused Palliativereacting 19.63 3.88 6–30 0.0002 0.0004to0.0009 0.524 0.0005 0.0001to0.0011 0.116 Avoiding 19.50 3.25 11–30 0.0004 0.0002to0.0012 0.191 0.0004 0.0002to0.0011 0.198 Passivereacting 12.45 3.56 0–23 0.0004 0.0003to0.0011 0.213 0.0001 0.0005to0.0007 0.709 Reassuringthoughts 12.69 2.61 6–20 0.0003 0.0006 0.558 0.0003 0.0005to0.0011 0.462 Expressionofemotions 6.47 1.84 3–11 0.0010 0.0004to0.0025 0.160 0.0002 0.0011to0.0015 0.821 SD:standarddeviation;SCS:SocialComparisonScale;CAPE:CommunityAssessmentofPsychicExperiences.TheboldtextindicatesthatthePvalueis<0.05,andthus signiﬁcant.

(5)

4. Discussion

Thisstudyexaminedtherelationshipbetweenperceivedsocial comparisons and psychosis proneness in ethnically Chinese individuals.Toourknowledgethiswasthefirststudytoexplore associations between differenttypes of social comparisons (i.e. rankandbelongingness)anddimensionsofpsychosisproneness. In addition it was the first study toexamine coping styles as potentialmoderatorfactors,andfamilialassociationsbetweenNSC and psychosis proneness. Three major findings were revealed: firstly,bothbelongingnessandrankwerefoundtopredictnegative anddepressivepsychoticexperiences,butnotpositivepsychotic experiences.Secondly,problem-focusedcopingstyleswerefound tobuffertheimpactofnegativesocialcomparisonsonpsychotic experiences. Finally, a familial association was found between belongingandnegativepsychoticexperiences,andthis relation-ship was stronger in monozygotic twins, suggesting perhaps sharedaetiologyduetocommongenes.Thesefindingsshedlight ontheexactnatureoftherelationshipbetweenNSCandpsychosis proneness,andhaveimplicationsforearlyinterventionandfuture translationalstudiesinthefieldofpsychosis.

Aswashypothesized, negativesocial comparisons,including bothbelongingnessandrank,significantlypredictedthenegative anddepressivedimensionsofpsychosisproneness.Thesefindings areconsistentwithastudy,whichfoundexperimentallyinduced socialranktoleadtosignificantincreasesinsadness[13].These findings are also in line with studies, which have found associationsbetweennegativeinterpersonalconceptsand nega-tivesymptomsinpsychosis[29],aswellastheresearchsuggesting acentralroleforinterpersonalinsecuritiesinthedevelopmentof depressiveandanxioussymptoms[1,30].Ifanindividualfeelsleft out or views himself/herself morenegatively in comparison to others,itmakessensethattheindividualmaybelessmotivatedor experiencelesspleasureinengaginginsuchactivities[29].

Ourhypothesisregardingthepositivedimensionofpsychosis proneness was not supported. Even though this finding is consistentwithatleastoneotherstudy[13],severalotherstudies havefoundassociationsbetweenNSCandparanoiaatasubclinical level[3,4,8]aswellasbetweenNSCandhallucinations[5–7].There are several potential explanations for these findings. One possibilityis that NSCare simplyanindirectmeasureof social adversity and that, in terms of positive symptoms, the actual relationshipiswithsocialadversitynotNSC[9].Thiswouldalso explainwhypreviouswork([e.g.8])whichdidnotconsidersocial adversityfoundasignificantrelationship,andwhyJayaetal.[9]

failed to ﬁnd a relationship between social rank and positive symptomswhentheyconsidereditasamediatorbetweensocial adversityandpositivesymptoms.Anotherpotentialexplanationis

thattheseﬁndingsreﬂectanaffectivepathwayaspostulatedby keycognitivemodelsofpsychosis([e.g.31]).Thatis,NSCmaylead toincreasednegativeemotione.g.sadness,whichinturnleadsto positivesymptoms[13].

Consistentwithourhypotheses,morefrequentuseof problem-focusedcopingstylesincludingactivecopingandsocialsupport seekingwerefoundtobuffertheharmfuleffectofNSConpsychotic experiences.Theseﬁndingsareconsistentwithpriorresearchthat highlightstheprotectiveroleofmoreproactiveproblem-focused coping styles against the development of psychopathology

[20].Thesefindingsalsosupportcognitivemodelsthatpropose a protective role for problem-focused coping strategies in the genesisofpsychoticdisorder[32].Emotion-focusedstrategiesdid not seem to exacerbate the relationship between NSC and psychotic experiences. Prior research suggested that more maladaptivecopingstrategiesmaymoderate(specifically, exacer-bate) the relationship between stressful experiences and the developmentofpsychosis,asinthestudybyPhillipsetal.[16], which found the coping strategies employed rather than the stressfulexperiencespersetopredictpsychosis.Ourfindingson theotherhand suggestthat emotion-focused strategiesneither exacerbatenorbuffertherelationshipbetweenNSCandpsychosis proneness.Onepossibilityisthatemotion-focusedstrategiesmay onlyplayaroleoncethepsychoticsymptomsreachacertainlevel ofseverity.ThiswouldexplainwhythestudybyPhilipsetal.[16], forexample,foundmaladaptivecopingstylestoplayamoderately role.Anotherpossibleexplanationisthat,astheliteraturereferred tothusfarhasfocusedonwesternpopulations,thisfindingmay reflectculturaldifferences.Thatis,usingemotion-focused strate-gies(i.e.thosethatmoveawayfromtheproblemwiththeaimof soothingnegativeassociatedemotions)maybelessdetrimental,in termsofpsychosis,inChinesepopulations.Futureresearchmay wishtoexploreculturaldifferencesinhowcopingstylesimpact thedevelopmentofpsychosisproneness.

Belongingand totalsocial comparisonin thefirst twinwere associatedwithnegativepsychoticexperiencesinthesecondtwin, therebysuggestingafamilialassociationbetweenanindividual’s senseofbelongingandnegativepsychoticexperiences. Further-more,thisrelationshipwasfoundtobestrongeramongstMZtwins suggesting it maybe due toshared genetic influences. Yet,no familial association was found between depressive psychotic experiences in twin one and NSC in twin two. These findings suggestpossibleetiologicaldifferencesintheroutetothenegative and depressive dimensions of psychosis. More specifically, belonging and negative psychotic experiences seem to share genetic and environmental influences that are common to membersofthesamefamily,whilstNSCanddepressivesymptoms appear toshare unique or idiosyncratic environmental factors. Researchershavehighlightedthepossibilityofdifferentroutesto differentformsofpsychosis.Myin-GermeysandvanOs[31],for example,suggestedonepathwaytothepositivesyndrome, non-deficit, good outcome form of schizophrenia I, characterized primarilybyalteredstresssensitivity,andanothertothenegative syndrome, deficit poor outcome Schizophrenia II, characterized primarily by cognitive impairment [31]. These differences in familialassociationsthuslikelyreflectsuchdifferences,perhaps evensuggestingathirdpathwaytothemoredepressiveformof schizophrenia:Schizoaffectivedisorder.Alternatively,depressive symptomsmaydevelopastheresultofdistressassociatedwiththe negativeandpositivesymptoms[27].

Theﬁndingsofthecurrentstudyshouldbeinterpretedinlight ofitslimitations.Firstly,thecross-sectionalnatureofthestudy doesnotallowforthedirectionofcausalitytobeestablished.Itis possible that individuals with increased psychosis proneness, particularlysubclinicalparanoia,mayleadtomoreNSC.Onthe other hand, NSC can be a vulnerability marker for the later Table4

Familial associations between negative social comparisons and psychotic experiences.

Twin1 Twin2

CAPE:negativepsychosis dimension

CAPE:depressivepsychosis dimension B t P B t P Unadjusted Rank 0.007 2.06 0.041 0.004 1.24 0.218 Belonging 0.012 3.60 0.000 0.006 2.12 0.036 TotalSCS 0.005 2.83 0.005 0.003 1.67 0.098 Adjusted Rank 0.006 1.57 0.120 0.002 0.609 0.544 Belonging 0.011 3.12 0.002 0.004 1.19 0.234 TotalSCS 0.005 2.32 0.022 0.001 0.858 0.392 SCS:SocialComparisonScale;CAPE:CommunityAssessmentofPsychic Expe-riences.TheboldtextindicatesthatthePvalueis<0.05,andthussigniﬁcant.

(6)

developmentofpsychosis.Futurestudiesmaywishtoemploya longitudinaldesigntoconfirmthedirectionofcausality.Asecond limitation is that measures in this study involved subjective reports and thus may be influenced by self-reportbiases. Our questionnaires nonetheless have been found to have high reliabilityand validity,particularlyforadolescent samplessuch astheoneexaminedhere[11,33,34],sothepresentfindingsare unlikelytobeprimarilyduetoself-reportbiases.Thirdly,wewere notabletoemploytwinmodellingduetothelimitedsamplesize. The results do, however, justify the implementation of this approachina largersample inordertodeterminewhether the established familial associations can be explained by shared environmentalorgeneticfactorsandquantifytheextent.Finally, whilstsomeofthepresentfindingswereinterpretedintermsof cross-culturaldifferencesinself-construals,nosuchmeasurehad beentaken.Ontheotherhand,thereissubstantialevidencefor moreinterdependentself-construalsinindividualsincollectivist societies such as China and Hong Kong than in the more individualistic Western societies [12]. Future research may consider direct comparisons of individuals from Chinese and Westerncultures,includingmeasuresofself-construals,tofurther evaluatethesehypotheses.

Despitetheselimitations,thepresentresearchhasimportant implicationsforbothearlyinterventionandfurthertranslational research.Firstly,thesefindingsspeakdirectlytotheimportanceof usingcognitivebehaviourtherapyapproacheswhichplacestrong emphasisonthedevelopmentofmoreeffectivecopingstrategies andmorepositiveinterpersonalschemaswithat-riskindividuals. Ourfindingssuggestthatguidingat-riskindividualstoengagein problem-focusedcopingstrategiesmaybeparticularlyimportant astheymayhelpbufferthepotentiallynegativeeffectsofNSCand, as a result, hinder the development of full-blown psychotic symptoms. Secondly, these findings tentatively suggest that measuresofNSCmayinformassessmentsfortheearly identifica-tionofat-riskindividualsinChinesepopulations.This,however, wouldneedtobeconfirmedinfuturestudies.Thirdly,ourfindings suggest a need for further work confirming the relationship between negativesocial comparisons and psychosisproneness/ Psychosis and it’s precise nature using genetic modelling and longitudinalapproaches.Finally,thisworkhighlightstheneedto considercross-culturaldifferences(particularlyindividualisticvs. collectivisticcultures)inpathwaystopsychosisandencourages furtherresearchinthisregard.

Disclosureofinterest

Theauthorsdeclarethattheyhavenocompetinginterest. References

[1]GilbertP,AllanS.Theroleofdefeatandentrapment (arrestedﬂight)in depression: an exploration of an evolutionary view. Psychol Med 1998;28:585–98.

[2]SeltenJP,Cantor-GraaeE.Socialdefeat:riskfactorforschizophrenia?BrJ Psychiatry2005;187:101–2.

[3]BebbingtonPE,McBrideO,SteelC,KuipersE,RadavanovicM,BrughaT,etal. The structure of paranoia in the general population. Br J Psychiatry 2013;202:419–27.

[4]FreemanD,GaretyPA,BebbingtonPE,SmithB,RollinsonR,FowlerD,etal. Psychologicalinvestigationofthestructure ofparanoia ina non-clinical population.BrJPsychiatry2005;186:427–35.

[5]BirchwoodM,GilbertP,GilbertJ,TrowerP,MeadenA,HayJ,etal. Interper-sonalandrole-relatedschemainﬂuencetherelationshipwiththedominant ‘voice’ in schizophrenia: a comparison of three models. Psychol Med 2004;34:1571–80.

[6]BirchwoodM,MeadenA,TrowerP,GilbertP,PlaistowJ. Thepowerand omnipotenceofvoices:subordinationandentrapmentbyvoicesand signiﬁ-cantothers.PsycholMed2000;30:337–44.

[7]GilbertP, Birchwood M,GilbertJ, TrowerP, HayJ, Murray B, etal.An explorationofevolvedmentalmechanismsfordominantandsubordinate behaviourinrelationtoauditoryhallucinationsinschizophreniaandcritical thoughtsindepression.PsycholMed2001;31:1117–27.

[8]FreemanD,NicoleE,ListerR,AntleyA,DunnG,SlaterM.Height,social comparison,andparanoia:animmersivevirtualrealityexperimentalstudy. PsychiatryRes2014;218:348–52.

[9]JayaES,AsconeL,LincolnTM.Socialadversityandpsychosis:themediating roleofcognitivevulnerability.SchizophrBull2016;43:557–65.

[10]FowlerD,HodgekinsJ,GaretyP,FreemanD,KuipersE,DunnG,etal.Negative cognition,depressedmood,andparanoia:alongitudinalpathwayanalysis usingstructuralequationmodeling.SchizophrBull2012;38:1063–73.

[11]AllanS,GilbertP.ASocialComparisonScale:psychometricpropertiesand relationshiptopsychopathology.PersIndividualDifferences1995;19:293–9.

[12]PfundmairM,AydinN,DuH,YeungS,FreyD,GraupmannV.Excludemeifyou can:culturaleffectsontheoutcomesofsocialexclusion.JCrossCultPsychol 2014;46:579–96.

[13]AsconeL,JayaES,LincolnTM.Theeffectofunfavourableandfavourablesocial comparisonsonparanoidideation:anexperimentalstudy.JBehavTherExp Psychiatry2016;56:97–105.

[14]JalbrzikowskiM,SugarCA,ZinbergJ,BachmanP,CannonTD,BeardenCE. Copingstylesofindividualsatclinicalhighriskfordevelopingpsychosis.Early IntervPsychiatry2012;8:68–76.

[15]MeloSS,BentallRP.Copinginsubclinicalparanoia:atwonationsstudy. PsycholPsychother2010;83:407–20.

[16]PhillipsLJ,EdwardsJ,McMurrayN,FranceyS.Comparisonofexperiencesof stressandcopingbetweenyoungpeopleatriskofpsychosisandanon-clinical cohort.BehavCognPsychother2012;40:69–88.

[17]SchuldbergD,KarwackiSB,BurnsGL.Stress,coping,andsocialsupportin hypotheticallypsychosispronesubjects.PsycholRep1996;78:1267–83.

[18]ZongJG,ChanRC,StoneWS,HsiX,CaoXY,ZhaoQ,etal.Copingﬂexibilityin youngadults:comparisonbetweensubjectswithandwithoutschizotypal personalityfeatures.SchizophrRes2010;122:185–92.

[19]FolkmanS,LazarusRS.Ifitchangesitmustbeaprocess:studyofemotionand copingduring threestages of acollege examination.J Pers SocPsychol 1985;48:150–70.

[20]CompasBE,Connor-SmithJK,Saltzman H,ThomsenAH, WadsworthME. Copingwithstressduringchildhoodandadolescence:problems,progress, andpotentialintheoryandresearch.PsycholBull2001;127:87–127.

[21]MurisP,MayerB,ReindersE,WesenhagenC.Person-relatedprotectiveand vulnerabilityfactorsofpsychopathologysymptomsinnon-clinical adoles-cents.CommunityMentHealthJ2009;47:47–60.

[22]MacdonaldEM,ElspethM,PicaS,McdonaldS,HayesRL,BaglioniJrAJ.Stress andcopinginearlypsychosis:roleofsymptoms,self-efﬁcacy,andsocial supportincopingwithstress.BrJPsychiatry1998;172:122–7.

[23]RaevuoriA,DickD,Keski-RahkonenA,PulkkinenL,RoseRJ,RissanenA,etal. Geneticandenvironmentalfactorsaffectingself-esteemfromage14to17:a longitudinalstudyofFinnishtwins.PsycholMed2007;37:1625–33.

[24]LatasterT,WichersM,JacobsN,MengelersR,DeromC,ThieryE,etal.Does reactivitytostresscosegregatewithsubclinicalpsychosis?Ageneral popula-tiontwinstudy. ActaPsychiatrScand2008;119:45–53.

[25]SperberAD.Translationandvalidationofstudyinstrumentsforcross-cultural research.Gastroenterology2004;126:S124–8.

[26]MarkW,ToulopoulouT.ValidationoftheChineseversionofCommunity AssessmentofPsychicExperiences(CAPE)inanadolescentpopulation.AsianJ Psychiatry2017;26:58–65.

[27]StefanisNC,HanssenM,SmirinsNK,etal.Evidencethatthreedimensionsof psychosis have a distribution in the general population. Psychol Med 2002;32:347–58.

[28]MarkW,ToulopoulouT.Psychometricpropertiesof‘‘communityassessment of psychic experiences’’: review and meta-analyses. Schizophr Bull 2016;42:34–44.

[29]Lincoln TM,MehlS, KestingML,RiefW.Negative symptomsand social cognition:identifyingtargetsforpsychologicalinterventions.SchizophrBull 2011;37:S23–32.

[30]GilbertP,AllanS,BroughS,MelleyS,MilesJNV.Relationshipofanhedoniaand anxietytosocialrank,defeatandentrapment.JAffectDisord2002;71:141–51.

[31]Myin-GermeysI,vanOsJ. Stressreactivity inpsychosis:evidenceforan affectivepathwaytopsychosis.ClinPsycholRev2007;27:409–24.

[32]NuechterleinKH,DawsonME,VenturaJ,GitlinM,SubotnikKL,SnyderKS,etal. Thevulnerability/stressmodelofschizophrenicrelapse:alongitudinalstudy. ActaPsychiatrScand1994;89:58–64.

[33]SchreursPJG,VandeWilligeG,TellegenB,BrosschotJF.Herzienehandleiding UtrechtseCopingLijst(UCL).Lisse:Swets&Zeitlinger,B.V;1993.

[34]TurnerH,Bryant-WaughR,PevelerR,BucksRS.Apsychometricevaluationof an English version of the Utrecht Coping List. Eur Eating Disord Rev 2012;20:339–42.