Wilson,J.Q.&Herrnstein,R.J.(1985)CrimeandHumanNature:TheDe - PDF document

Wilson,J.Q.&Herrnstein,R.J.(1985)CrimeandHumanNature:TheDe“nitiveStudyoftheCausesofCrime.Touchstone,NewYork.Wirls,D.(1986)Reinterpretingthegendergap.PublicOpinionQuarterly,316…330.Wol”e,L.M.(1980)Theenduringeffectsofeducationonverbalskills.SociologyofEducation,104…114.Intelligenceandhomosexuality Lee,R.B.(1979)The!KungSan:Men,Women,andWorkinaForagingSociety.CambridgeUniversityPress,Cambridge.Leutenegger,W.&Kelly,J.T.(1977)Relationshipofsexualdimorphismincaninesizeandbodysizetosocial,behavioral,andecologicalcorrelatesinanthropoidprimates.Primates,117…136.LeVay,S.(2010)Gay,Straight,andtheReasonWhy:TheScienceofSexualOrientation.OxfordUniversityPress,Oxford.Levinson,D.(EditorinChief)(1991…1995)EncyclopediaofWorldCultures(10Volumes).G.K.Hall,Boston.Miller,A.S.&Hoffmann,J.P.(1995)Riskandreligion:anexplanationofgenderdifferencesinreligiosity.JournalfortheScienti“cStudyofReligion,63…75.Miller,A.S.&Stark,R.(2002)Genderandreligiousness:cansocializationexplanationsbesaved?AmericanJournalofSociology107,1399…1423.Miner,J.B.(1957)IntelligenceintheUnitedStates:ASurvey…WithConclusionsforManpowerUtilizationinEducationandEmployment.Springer,NewYork.Murray,S.O.(2000)Homosexualities.UniversityofChicagoPress,Chicago.Muscarella,F.(2000)Theevolutionofhomoeroticbehaviorinhumans.JournalofHomosexuality,51…77.Mustanski,B.S.,Chivers,M.L.&Bailey,J.M.(2002)Acriticalreviewofrecentbiologicalresearchonhumansexualorientation.AnnualReviewofSexResearch,89…140.Nash,G.(2001)Thesubversivemale:homosexualandbestialimagesonEuropeanMesolithicrockart.InBevan,L.(ed.)IndecentExposure:Sexuality,SocietyandtheArchaeologicalRecord,CruithnePresspp.43…63.Pickford,M.(1986)Ontheoriginsofbodysizedimorphisminprimates.InPickford,M.&Chiarelli,B.(eds)SexualDimorphisminLivingandFossilPrimates.IlSedicesimo,Florence,pp.77…91.Rahman,Q.,Abrahams,S.&Wilson,G.D.(2003)Sexual-orientation-relateddifferencesinverbal”uency.Neuropsychology,240…246.Rahman,Q.,Bhanot,S.,Emrith-Small,H.,Ghafoor,S.&Roberts,S.Gendernonconformity,intelligence,andsexualorientation.ArchivesofSexualBehavior(inpress).Shapiro,R.Y.&Mahajan,H.(1986)Genderdifferencesinpolicypreferences:asummaryoftrendsfromthe1960stothe1980s.PublicOpinionQuarterly,42…61.Sundquist,J.L.(1983)DynamicsofthePartySystem(revisededition).BrookingsInstitution,WashingtonDC.Symons,D.(1990)Adaptivenessandadaptation.EthologyandSociobiology,427…444.Tooby,J.&Cosmides,L.(1990)Thepastexplainsthepresent:emotionaladaptationsandthestructureofancestralenvironments.EthologyandSociobiology,375…424.Trivers,R.(2000)Theelementsofascienti“ctheoryofself-deception.AnnalsofNewYorkAcademyofSciencesoftheUSA,114…131.Turner,G.(1996)FindinggenesontheXchromosomebywhichHomomayhavebecomesapiens.AmericanJournalofHumanGenetics,1109…1110.Tuttle,G.E.&Pillard,R.C.(1991)Sexualorientationandcognitiveabilities.ArchivesofSexualBehavior,307…318.Weinrich,J.(1978)Nonreproduction,homosexuality,transsexualism,andintelligence.JournalofHomosexuality,275…290.Whitten,N.E.Jr(1976)SachaRuna:EthnicityandAdaptationofEcuadorianJungleQuichua.UniversityofIllinoisPress,Urbana.Wilson,G.&Rahman,Q.(2005)BornGay:ThePsychobiologyofSexOrientation.PeterOwen,London.S.Kanazawa Hamer,D.H.,Hu,S.,Magnuson,V.L.,Hu,N.&Pattatucci,A.M.I.(1993)AlinkagebetweenDNAmarkersontheXchromosomeandmalesexualorientation.Science,321…327.Harvey,P.H.&Bennett,P.M.(1985)Sexualdimorphismandreproductivestrategies.InGhesquiere,J.,Martin,R.D.&Newcombe,F.(eds)HumanSexualDimorphism.TaylorandFrancis,London,pp.43…59.Haselton,M.G.&Nettle,D.(2006)Theparanoidoptimist:anintegrativeevolutionarymodelofcognitivebiases.PersonalityandSocialPsychologyReview,47…66.Herrnstein,R.J.&Murray,C.(1994)TheBellCurve:IntelligenceandClassStructureinAmericanFreePress,NewYork.Hewlett,B.S.&Hewlett,B.L.(2010)SexandsearchingforchildrenamongAkaforagersandNgandufarmersofcentralAfrica.AfricanStudyMonographs,107…125.Hilbe,J.M.(2007)NegativeBinomialRegression.CambridgeUniversityPress,Cambridge.Hill,K.&Hurtado,A.M.(1996)AcheLifeHistory:TheEcologyandDemographyofaForagingPeople.Aldine,NewYork.Huang,M-H.&Hauser,R.M.(1998)Trendsinblack…whitetest-scoredifferentials:II.TheWORDSUMvocabularytest.InNeisser,U.(ed.)TheRisingCurve:Long-termGainsinIQandRelatedMeasure.AmericanPsychologicalAssociation,Washington,DC,pp.303…332.Kanazawa,S.(2001)Degustibusestdisputandum.SocialForces,1131…1163.Kanazawa,S.(2002)Bowlingwithourimaginaryfriends.EvolutionandHumanBehavior167…171.Kanazawa,S.(2004a)TheSavannaPrinciple.ManagerialandDecisionEconomics,41…54.Kanazawa,S.(2004b)Generalintelligenceasadomain-speci“cadaptation.PsychologicalReview111,512…523.Kanazawa,S.(2008)Temperatureandevolutionarynoveltyasforcesbehindtheevolutionofgeneralintelligence.Intelligence,99…108.Kanazawa,S.(2009)Evolutionarypsychologyandcrime.InWalsh,A.&Beaver,K.M.(eds)BiosocialCriminology:NewDirectionsinTheoryandResearch.Routledge,NewYork,pp.90…110.Kanazawa,S.(2010a)Whyliberalsandatheistsaremoreintelligent.SocialPsychologyQuarterly,33…57.Kanazawa,S.(2010b)Evolutionarypsychologyandintelligenceresearch.AmericanPsychologist,279…289.Kanazawa,S.&Hellberg,J.E.E.U.(2010)Intelligenceandsubstanceuse.ReviewofGeneralPsychology,382…396.Kanazawa,S.&Novak,D.L.(2005)Humansexualdimorphisminsizemaybetriggeredbyenvironmentalcues.JournalofBiosocialScience,657…665.Kanazawa,S.&Perina,K.(2009)Whynightowlsaremoreintelligent.PersonalityandIndividualDifferences,685…690.Kauth,M.R.(2000)TrueNature:ATheoryofSexualAttraction.Kluwer,NewYork.Kirk,K.M.,Bailey,J.M.,Dunne,M.P.&Martin,N.G.(2000)Measurementmodelsforsexualorientationinacommunityoftwinsample.BehaviorGenetics,345…356.Kirkpatrick,L.A.(2005)Attachment,Evolution,andthePsychologyofReligion.Guilford,NewYork.Kirkpatrick,R.C.(2000)Theevolutionofhumanhomosexualbehavior.CurrentAnthropology,385…413.Kluegel,J.R.&Smith,E.R.(1986)BeliefsaboutInequality:AmericansViewofWhatisandWhatOughtToBe.Aldine,NewYork.Lake,C.C.&Breglio,V.J.(1992)Differentvoices,differentviews:thepoliticsofgender.InRies,P.&Stone,A.J.(eds)TheAmericanWoman,1992…93:AStatusReport.Norton,NewYork,pp.178…201.Intelligenceandhomosexuality Blanchard,R.&Bogaert,A.F.(1996)Biodemographiccomparisonsofhomosexualandhetero-sexualmenintheKinseyinterviewdata.ArchivesofSexualBehavior,551…579.Bogaert,A.F.&Blanchard,R.(1996)Physicaldevelopmentandsexualorientationinmen:height,weightandageofpubertydifferences.PersonalityandIndividualDifferences,77…84.Boyer,P.(2001)ReligionExplained:TheEvolutionaryOriginsofReligiousThought.Basic,NewYork.Chagnon,N.(1992)Yanomamo(4thedition).HarcourtBraceJovanovich,FortWorth,TX,USA.Chivers,M.L.,Seto,M.C.&Blanchard,R.(2007)Genderandsexualorientationdifferencesinsexualresponsetosexualactivitiesversusgenderofactorsinsexual“lms.JournalofPersonalityandSocialPsychology,1108…1121.Cochran,G.&Harpending,H.(2009)The10,000YearExplosion:HowCivilizationAcceleratedHumanEvolution.Basic,NewYork.Crapo,R.H.(1995)Factorsinthecross-culturalpatterningofmalehomosexuality:areappraisaloftheliterature.Cross-CulturalResearch,178…202.Crawford,C.B.(1993)Thefutureofsociobiology:countingbabiesorproximatemechanisms?TrendsinEcologyandEvolution,183…186.Cronk,L.(2004)FromMukogodotoMaasai:EthnicityandCulturalChangeinKenya.Westview,Boulder.deWaal,F.B.M.(1989)Foodsharingandreciprocalobligationsamongchimpanzees.JournalofHumanEvolution,433…459.deWaal,F.B.M.(1992)Appeasement,celebration,andfoodsharinginthetwoPanspecies.InNishida,T.,McGrew,W.C.&Marler,P.(eds)TopicsinPrimatology:HumanOrigins.Uni-versityofTokyoPress,Tokyo,pp.37…50.deWaal,F.B.M.,Luttrell,L.M.&Can“eld,M.E.(1993)Preliminarydataonvoluntaryfoodsharinginbrowncapuchinmonkeys.AmericanJournalofPrimatology,73…78.Deary,I.J.,Batty,G.D.&Gale,C.R.(2008)Brightchildrenbecomeenlightenedadults.PsychologicalScience,1…6.Deary,I.J.,Whiteman,M.C.,Starr,J.M.,Whalley,L.J.&Fox,H.C.(2004)Theimpactofchildhoodintelligenceonlaterlife:followinguptheScottishMentalSurveysof1932and1947.JournalofPersonalityandSocialPsychology,130…147.Diamond,L.M.(2008)SexualFluidity:UnderstandingWomensLoveandDesire.HarvardUni-versityPress,Cambridge.Ellis,L.(1998)NeoDarwiniantheoriesofviolentcriminalityandantisocialbehavior:photo-graphicevidencefromnonhumananimalsandareviewoftheliterature.AggressionandViolentBehavior,61…110.Ellis,L.&Ames,M.A.(1987)Neurohormonalfunctioningandsexualorientation:atheoryofhomosexuality…heterosexuality.PsychologicalBulletin101,233…258.Evans,P.D.,Gilbert,S.L.,Mekel-Bobrov,N.,Vallender,E.J.,Anderson,J.R.,Vaez-Azizi,L.etal(2005)Microcephalin,ageneregulatingbrainsize,continuestoevolveadaptivelyinhumans.Science309,1717…1720.Gallup,G.G.Jr.&Suarez,S.D.(1983)Homosexualityasaby-productofselectionforoptimalheterosexualstrategies.PerspectivesinBiologyandMedicine,315…322.Gallup,G.G.Jr(1995)Haveattitudestowardhomosexualsbeenshapedbynaturalselection?EthologyandSociobiology,53…70.Grov,C.,Bimbi,D.S.,Nanšn,J.E.&Parsons,J.T.(2006)Race,ethnicity,gender,andgenera-tionalfactorsassociatedwiththecoming-outprocessesamonggay,lesbian,andbisexualindi-viduals.JournalofSexResearch,115…121.Guthrie,S.E.(1993)FacesintheClouds:ANewTheoryofReligion.OxfordUniversityPress,NewYork.S.Kanazawa werecompletingtheireducation,anditisstillfreeinScotlandtoday,soitisunlikelythatthenegativeassociationbetweeneducationandhomosexualitystemsfromparentsreluctancetopayforcollegefortheirgaychildren.Onepossibilityisthatthestressandstigmaassociatedwithbeinggayandcomingoutmakeitmoredif“cultforgaychildrenandadolescentstopursuehighereducation.ThepresentstudyrepresentsonlythemostrecentattempttousetheSavanna-IQInteractionHypothesistoexploreandexplaintheoriginofindividualvalues,preferencesandtastes.EventhoughIprovidesupportiveempiricalresults,futurestudiesareneces-sarytoestablishtheassociationbetweengeneralintelligenceandhomosexualbehaviourmore“rmly.Inaddition,moreempiricalworkisclearlynecessary,bothtotesttheHypothesisrigorouslyinothervaluedomainsandtoinvestigatetheoriginofindividualvalues,includingsexualbehaviour.AcknowledgmentsThisresearchusesdatafromAddHealth,aprogrammeprojectdesignedbyJ.RichardUdry,PeterS.BearmanandKathleenMullanHarris,andfundedbyagrantP01-HD31921fromtheEuniceKennedyShriverNationalInstituteofChildHealthandHumanDevelopment,withcooperativefundingfromseventeenotheragencies.SpecialacknowledgmentisduetoRonaldR.RindfussandBarbaraEntwisleforassistanceintheoriginaldesign.Personsinterestedinobtainingdata“lesfromAddHealthshouldcontactAddHealth,CarolinaPopulationCenter,123WestFranklinStreet,ChapelHill,NC27516-2524,USA(addhealth@unc.edu).NodirectsupportwasreceivedfromgrantP01-HD31921forthisanalysis.IthankDaviddeMeza,PatrickM.Markey,QaziRahman,DianeJ.Reyniersandanonymousreviewersfortheircommentsonearlierdrafts.ReferencesAdams,H.E.,Wright,L.W.&Lohr,B.A.(1996)Ishomophobiaassociatedwithhomosexualarousal?JournalofAbnormalPsychology105,440…445.Alexander,R.D.,Hoogland,J.L.,Howard,R.D.,Noonan,K.M.&Sherman,P.W.(1979)Sexualdimorphismsandbreedingsystemsinpinnipeds,ungulates,primatesandhumans.InChagnon,N.A.&Irons,W.(eds)EvolutionaryBiologyandHumanSocialBehavior:AnAnthropologicalPerspective.DuxburyPress,NorthScituate,pp.402…435.Arabsheibani,G.R.,Marin,A.&Wadsworth,J.(2005)GaypayintheUK.Economica,333…347.Atran,S.(2002)InGodsweTrust:TheEvolutionaryLandscapeofReligion.OxfordUniversityPress,Oxford.Bagemihl,B.(2000)BiologicalExuberance:AnimalHomosexualityandNaturalDiversity.MartinsPress,NewYork.Bailey,J.M.&Pillard,R.C.(1991)Ageneticstudyofmalesexualorientation.ArchivesofGeneralPsychiatry,1089…1096.Bell,A.P.&Weinberg,M.S.(1978)Homosexualities:AStudyofDiversityamongMenandWomen.SimonandSchuster,NewYork.Bell,A.P.,Weinberg,M.S.&Hammersmith,S.K.(1981)SexualPreference:ItsDevelopmentinMenandWomen.IndianaUniversityPress,Bloomington.Intelligenceandhomosexuality InteractionbetweengeneralintelligenceandsexForthemostpart,theHypothesisdoesnotmakesex-speci“cpredictions.Itpredictsthatbothmoreintelligentmenandmoreintelligentwomenaremorelikelytoacquireandespouseevolutionarilynovelpreferencesandvaluesthantheirlessintelligentcounter-parts.Asaresult,theHypothesisisunabletoaccountforthestatisticallysigni“cantinter-actionsbetweensexandgeneralintelligencefoundinTables2,3and4.Itisobviousthatmanyfactorsin”uenceanindividualspropensitytoengageinhomosexualbehaviour,andgeneralintelligenceisonlyoneofthem.Itislikelythatsomeoftheotherfactors…suchassocialandculturalin”uencesonhomosexualbehaviouranditsacceptabilityformenorwomeninsociety…mayin”uencewhymoreintelligentmenormoreintelligentwomenaremorelikelytoengageinhomosexualbehaviour.Thisappearsespeciallytobethecase,becausethedirectionofthesigni“cantinteractiondependsontheparticularmeasureofhomosexualbehaviour.Moreintelligentwomenaremorelikelytoexpressadulthomosexualattractionthanmoreintelligentmen,butmoreintelligentmenaremorelikelytohavemorehomosexualpartnersandhomosexualcohabitationpartnersthanmoreintelligentwomen.Itisthereforelikelythatnosinglefactorcanexplainallofthesigni“cantsexinteractioneffectsthatI“nd,iftheyareindeedrobust.AlternativeexplanationsAnearlierstudy(Rahmanetal.,2003)showsthatgaymenhavesuperiorverbal”uencythanstraightmen.GiventhatmymeasuresofintelligenceinStudies1and2arelargelyverbal,thiscanpotentiallyexplainwhymoreintelligentboysmaygrowuptobemorehomosexualasadults.However,sincelesbianshavetheoppositepro“letogaymen,havinglowerverbal”uencythanstraightwomen,Rahmanetal.s(2003)“ndingscannotexplainwhymoreintelligentgirlsalsogrowuptobemorehomosexualasadults.Further,IuseagenuinemeasureofgeneralintelligenceinStudy3.Anotherpossibilityisthatmoreintelligentindividuals,ratherthanbeingtrulymorehomosexualintheirsexualidentity,expressedattraction,andsexualbehaviour,aremorelikelyopenlytoadmitthattheyarehomosexualthanlessintelligentindividuals.Itmayalsobethatmoreintelligentindividualsaremoreself-awareandlessself-deceptivethanlessintelligentindividuals.Ifso,thenmoreintelligentindividualsaremorelikelytobeawareoftheirhomosexualdesiresandfantasiesthanlessintelligentindividuals.Inote,however,thatself-awarenessandself-deceptionmusthavedeepevolutionaryorigin(Trivers,2000),sotheHypothesiswouldpredictthatthecapacityforself-awarenessisunrelatedtogeneralintelligence.Thereisnowayformetodiscoverwhetherrespondentsmaybemisrepresentingthemselvesinmydata;justlikeanyotheruserofthesesurveydata,Iamatthemercyoftheirrecordedresponses.However,ifmoreintelligentindividualsareindeedmorelikelyopenlytoadmitthattheyarehomosexual,thenonewouldthinkthatmoreedu-catedindividualsareequallymorelikelytobesocandid.Thisalternativehypothesisthereforecannotexplainwhyeducationhasasigni“cantlynegativeassociationwithhomosexualityinStudies1and3andnoassociationatallinStudy2.Butthenwhyareeducationandhomosexualitynegativelyassociated?Higheredu-cationwasentirelyfreeinEnglandandWalesuntil1998,whentheNCDSrespondentsS.Kanazawa heterosexualpartners(contrarytotheHypothesis).Further,inStudy1,childhoodintel-ligenceisnotassociatedwithadultexpressedhomosexualattractionamongmen;inStudy3,childhoodgeneralintelligenceisnotassociatedwiththelifetimenumberofhomosexualcohabitationpartnersamongwomen.Second,becauseallofthesamplesarenationallyrepresentativeandbecausehomo-sexualityisrelativelyrareinthegeneralpopulation,thedependentmeasureshaveverylowfrequencies.Forexample,inStudy1,90%ofAddHealthrespondentsidentifythemselvesas100%straight.InStudy3,onlyaverysmallproportion(0.5%)oftheNCDSrespondentshaveeverhadanyhomosexualcohabitationpartners.Thesmallincidenceofhomosexualityinthemeasuresmayin”uencethestabilityoftheestimates.Third,evenwhentheresultsunequivocallysupportthepredictionoftheHypothesis,themagnitudeoftheassociationbetweenintelligenceandhomosexualityisoftenverysmall.Forexample,inStudy1,verybrightindividualsandverydullindividualsareseparatedbylessthanone-sixthofapointona“ve-pointscaleofadultsexualidentityorlessthan7%intheprobabilityofeverexperiencingadulthomosexualattraction,althoughthelatterrepresentsnearlydoublingoftheprobability.InStudy2,verybrightindividualshavehadmerely2.11morehomosexualpartnersintheirlifetimesthanverydullindividuals,althoughitdoesrepresentaneightfoldincrease.PracticalimportanceIemphasizethatmyscienti“cinterestinhomosexualbehaviourinthispaperisstrictlytheoretical.Homosexualbehaviourrepresentsyetanotherdomaininlifewheresomeindividualsmaychoosetoacquireandespouseevolutionarilynovelpreferencesandvalues.TheHypothesishaspreviouslybeentestedinmanyotherdomains,andineverycasemoreintelligentindividualshavebeenshowntobemorelikelytoacquireandespouseevolutionarilynovelpreferencesandvaluesineachdomain.Homosexualbehavioursimplyrepresentsyetanotherdomainoflifewheresomeindividualsmaychooseevolutionarilynovelbehaviour(homosexuality)whileothersmaychooseevolu-tionarilyfamiliarbehaviour(heterosexuality),andgivesmeanopportunitytoexamineitsassociationwithchildhoodgeneralintelligence.SmallstatisticalassociationsbetweenintelligenceandhomosexualitythatI“ndinmystudiessupportthepredictionoftheHypothesis.However,Iemphasizethatmy“ndingshaveabsolutelynopracticalimportance.Itisnotlikewecannowusesome-onesintelligencetoassesstheirhomosexualityaccurately.Myapproachtoscienceisdecidedlybasic,notapplied(clinicalormedical).Iamentirelydrivenbythedesiretodiscoverknowledge,notbyitspotentialapplicationsorimplications.Mypaperrepresentsoneofthemostcomprehensiveempiricalattemptstoestablishtheassociationbetweenintelligenceandhomosexuality,usinglarge,prospectivelylongi-tudinalandnationallyrepresentativesamplesfromtwodifferentnations.Whileotherstudieshavenotedapotentialempiricalassociationbetweenintelligenceandhomo-sexuality,tothebestofmyknowledge,theHypothesisprovidestheonlytheoreticalexplanationforwemayexpectsuchanassociationtoexistatall.Itexplainswhymoreintelligentchildrenmaygrowuptoidentifythemselvesashomosexual,expresshomosexualattraction,havemorehomosexualsexualandcohabitationpartnersthanlessintelligentchildren.Intelligenceandhomosexuality cohabitationpartnersintheirlifetimesthanlessintelligentindividuals.Sinceheterosexualunionsareeminentlyevolutionarilyfamiliar,thisisconsistentwiththepredictionoftheHypothesis.DiscussionandConclusionsTheSavanna-IQInteractionHypothesis,derivedfromalogicalconjunctionoftheSavannaPrincipleandatheoryoftheevolutionofgeneralintelligence,suggeststhatmoreintelligentindividualsmaybemorelikelytoacquireandespouseevolutionarilynovelpreferencesandvaluesthanlessintelligentindividuals,whilegeneralintelligencemayhavenoeffectontheacquisitionandespousalofevolutionarilyfamiliarpreferencesandvalues.SeveralearlierstudieshaveempiricallysupportedtheHypothesisinvariousdomainsoflife(politicalideology,religiosity,sexualexclusivity,circadianrhythmandconsumptionofpsychoactivesubstances).Inthispaper,IhaveextendedtheHypothesistosexualidentity,expresseddesiresandbehaviour.Whileindividualstruesexualorienta-tionsareprobablydeterminedpriortobirth,theirsexualidentity,desiresandbehaviourmaydeviatefromtheirtruesexualorientation,andtheHypothesissuggestsaroleofgeneralintelligenceinthem.Asurveyofethnographiesoftraditionalsocietiesshowsthatexclusivelyorpredomi-nantlyhomosexualidentityandbehaviourwereprobablyrareintheancestralenviron-mentandmaythusbeevolutionarilynovel.TheSavanna-IQInteractionHypothesisthereforepredictsthatmoreintelligentindividualsaremorelikelytoidentifythemselvestobehomosexualandengageinhomosexualbehaviourthanlessintelligentindividuals.Threelarge,nationallyrepresentativesamples(twoofwhichareprospectivelylongitu-dinal)fromtwodifferentnationsconvergetosupportmyprediction.ThisstudysanalysisofAddHealthdata(Study1)suggeststhat,netofage,sex,race,maritalstatus,parentalstatus,education,earnings,religion,religiosityandpoliticalattitude,childhoodintelli-genceissigni“cantlypositivelyassociatedwithadulthomosexualidentityandexpressedadulthomosexualattraction.Moreintelligentchildrenaremorelikelytogrowuptobeadultswhoidentifythemselvestobehomosexualandexpresshomosexualattractionthanlessintelligentchildren.TheanalysisoftheGeneralSocialSurveys(Study2)sug-geststhatmoreintelligentindividualshavehadsigni“cantlymorehomosexualpartnersintheirlifetimesthanlessintelligentindividuals.Whileintelligenceisalsoassociatedwiththelifetimenumberofheterosexualpartners,theassociationistwiceasstrongwithhomosexualpartnersaswithheterosexualpartners.TheanalysisoftheNationalChildDevelopmentStudy(Study3)showsthatmoreintelligentchildren(beforetheageof16)havesigni“cantlymoresame-sexcohabitationpartnersthanlessintelligentchildren30yearslater.Incontrast,moreintelligentindividualsdonothavemoreopposite-sexcohabitationpartnersintheirlifetimes.LimitationsofthecurrentstudiesThereareseveralimportantlimitationstomycurrentempiricalstudies.First,whilemydataanalysesinthreestudiesareconsistentwiththepredictionsfromtheHypothesis,theempiricalsupportisnotentirelyunequivocal.Inparticular,theanalysisoftheGSSdatashowsthatintelligenceisassociatedwithnotonlythelifetimenumberofhomo-sexualpartners(consistentwiththeHypothesis)butalsowiththelifetimenumberofS.Kanazawa withlifetimenumberofhomosexualcohabitationpartnersamongmen(0.01)butnotamongwomen(0.002,SE0.014,ns).Table4,column(3),showsthat,incomparisonwiththelifetimenumberofhomo-sexualcohabitationpartners,childhoodgeneralintelligenceisnotsigni“cantlyasso-ciatedwiththelifetimenumberofheterosexualcohabitationpartners.Netofthesamecontrolvariablesasbefore,moreintelligentindividualsdonothavemoreheterosexual Table4.Theassociationbetweenchildhoodintelligenceandlifetimenumberofhomosexualcohabitationpartners,NationalChildDevelopmentStudyLifetimenumberofhomosexualcohabitationpartnersLifetimenumberofheterosexualpartners(1)(2)(3)Childhoodintelligence0.023(0.010)0.003(0.012)0.001(0.001)Sex(1ifmale)0.033(0.243)4.515*(10.855)0.018(0.019)Childhoodintelligencesex0.042*(0.017)Currentlymarried(1ifyes)0.234(0.441)0.216(0.441)0.141***(0.026)Evermarried(1ifyes)1.974***(0.438)2.001***(0.437)0.583***(0.049)Parentalstatus(1ifparent)0.724*(0.296)0.774**(0.297)0.179***(0.029)Education0.101*(0.052)0.108*(0.053)0.002(0.002)Earnings6.558(8.3057.267(8.4282.271(3.275ReligionRomanCatholic0.795(0.463)0.853(0.467)0.049(0.039)Anglican0.925**(0.291)0.950**(0.293)0.034(0.028)OtherChristian1.333***(0.370)1.338***(0.370)0.028(0.030)Other0.795(1.062)0.869(1.066)0.135(0.091)Frequencyofchurchattendance0.018(0.170)0.003(0.171)0.025*(0.011)Constant1.741(1.109)0.443(1.386)0.269(0.088)Likelihoodratio(df12/13)168.372174.553558.650414141414141Mainentriesareunstandardizedregressioncoef“cients.Entriesinparenthesesarestandarderrors.0.01;***0.001(two-tailed).Intelligenceandhomosexuality tests(ReadingComprehensionTestandMathematicsComprehensionTest).I“rstperformafactoranalysisateachagetocomputetheirgeneralintelligencescoreforeachage.Allcognitivetestscoresateachageloadonlyononelatentfactor,withreasonablyhighfactorloadings(age7:CopyingDesignsTest0.671,Draw-a-ManTest0.696,SouthgateGroupReadingTest0.780andProblemArithmeticTest0.762;age11:VerbalGeneralAbilityTest0.920,NonverbalGeneralAbilityTest0.885,ReadingComprehensionTest0.864,MathematicalTest0.903,andCopy-ingDesignsTest0.486;age16:ReadingComprehensionTest0.909,andMathe-maticsComprehensionTest0.909).ThelatentgeneralintelligencefactorsateachageareconvertedintothestandardIQmetric,withameanof100andastandarddeviationof15.ThenIperformasecond-orderfactoranalysiswiththeIQscoresatthreedifferentagestocomputetheoverallchildhoodgeneralintelligencescore.ThethreeIQscoresloadonlyononelatentfactorwithveryhighfactorloadings(age70.867;age110.947;age0.919).IusethechildhoodgeneralintelligencescoreinthestandardIQmetricasthemainindependentvariable.Controlvariables.Inadditiontochildhoodgeneralintelligence,Icontrolforthefollowingvariablesintheregressionequations:sex(0female,1male;measuredatbirth),whethercurrentlymarried(1yes;measuredat47),whetherevermarriedyes;measuredat47),whethereveraparent(1yes;measuredat47),education(ageatwhichtherespondentleftformalschooling,measuredat42),earnings(inGBP,measuredat47),religion(infourdummies…RomanCatholic,Anglican,otherChristianandotherreligion…withnoneasthereferencecategory;measuredat42),frequencyofchurchattendance(1neverorveryrarely,2lessthanonceamonth,3morethanonceamonth,4onceaweekormore,measuredat42).ResultsTable4,column(1),showsthat,netofsex,whethercurrentlymarried,whetherevermarried,whethereveraparent,education,earnings,religionandfrequencyofchurchattendance,moreintelligentindividualsintheUnitedKingdomhavehadsigni“cantly(0.05)moresame-sexcohabitationpartnersintheirlifetimesthanlessintelligentindividuals.Themoreintelligenttheyarebeforetheageof16,themoresame-sexpartnerstheyhavebeforetheageof47.Quitepredictably,thosewhohaveeverbeenmarriedandthosewhohaveeverhadchildrenhavesigni“cantly0.001and0.05,respectively)fewersame-sexcohabitationpartners,asdoAnglicansandthosewhosubscribetonon-Christianreligions(0.01and0.001,respectively)comparedwithatheists.ReplicatingtheanalysesaboveofAddHealthdata(presentedinTables1and2),whilemoreintelligentindividualsaremorehomosexualjudgedbythelifetimenumberofhomosexualcohabitationpartners,moreeducatedindividualsarelesshomosexualbythesamemeasure.Table4,column(2),showsthattheinteractiontermbetweenchildhoodgeneralintelligenceandsexisstatisticallysigni“cant(0.05).Itsuggeststhattheeffectofchildhoodintelligenceonadulthomosexualityissigni“cantlystrongeramongmenthanamongwomen.Childhoodgeneralintelligenceissigni“cantlypositivelyassociatedS.Kanazawa TheNationalChildDevelopmentStudy(NCDS)isalarge-scaleprospectivelylongi-tudinalstudywhichhasfollowedapopulationofBritishrespondentssincebirthformorethanhalfacentury.Thestudyincludesbabies(17,419)borninGreatBritain(England,WalesandScotland)duringoneweek(3rd…9thMarch1958).Therespondentsweresubsequentlyre-interviewedin1965(Sweep1atage7;15,496),in1969(Sweep2atage11;18,285),in1974(Sweep3atage16;14,469),in1981(Sweep4atage23;12,537),in1991(Sweep5atage33;11,469),in1999…2000(Sweep6atage41…42;11,419)andin2004…2005(Sweep7atage46…47;9534).TherearemorerespondentsinSweep2thanintheoriginalsample(Sweep0)becausetheSweep2sampleincludeseligiblechildrenwhowereinthecountryin1969butnotin1958whenSweep0interviewswereconducted.Ineachsweep,personalinterviewsandques-tionnaireswereadministeredtotherespondents,totheirmothers,teachersanddoctorsduringchildhood,andtotheirpartnersandchildreninadulthood.Nearlyall(97.8%)oftheNCDSrespondentsareCaucasian.Therearesofewre-spondentsinotherracialcategoriesthat,ifIcontrolforracewithaseriesofdummiesingeneralizedlinearmodels,itoftenresultsincompleteseparationofdata,andthemaximumlikelihoodestimationbecomesimpossible.IthereforedonotcontrolforrespondentsraceintheanalysisoftheNCDSdata.Includingaracedummy(whitevs.others)veryslightlystrengthenstheassociationbetweenchildhoodintelligenceandhomosexualityreportedbelow,butotherwisedoesnotaltermysubstantiveconclusions.Dependentvariable:lifetimenumberofhomosexualandheterosexualcohabitationpartners.Forameasureofhomosexuality,thenumberofsame-sexcohabitationpartnersthattherespondentshaveeverhadintheirlifeisused,de“nedassomeoneofthesamesexwithwhomtherespondentshavelivedasmarriedandsharedanaccommodationfor6monthsorlonger.Oftherespondents,0.5%(94)reportatleastonelifetimehomosexualcohabitationpartner.Becausethelifetimenumberofsame-sexcohabitationpartnersisacountmeasurewithoverdispersion(0.067),Iusenegativebinomialregressiontoanalyseit(Hilbe,2007).Giventhathomosexualmenoftenhaveaverylargenumberofsexpartners,evenwhiletheyareincommittedrelationships(Bell&Weinberg,1978),thenumberofcohabitationpartnersisnottheidealmeasureofhomosexuality.IusetheNCDSdataforthestrengthofthemeasureoftheindependentvariable(generalintelligence),notforthestrengthofthemeasureofthedependentvariable(homosexuality).Asacomparison,Ialsousethenumberofopposite-sexcohabitationpartnersthattherespondentshaveeverhadintheirlife.Becausethelifetimenumberofopposite-sexcohabitationpartnersisacountmeasurewithoutoverdispersion(2.907,1.090),IusePoissonloglinearregressiontoanalyseit(Hilbe,2007).Bothdependentvariablesaremeasuredatage47.Independentvariable:childhoodgeneralintelligence.TheNCDSrespondentstakemultipleintelligencetestsatages7,11and16.Atage7,therespondentstakefourcognitivetests(CopyingDesignsTest,Draw-a-ManTest,SouthgateGroupReadingTestandProblemArithmeticTest).AtAge11,theytake“vecognitivetests(VerbalGeneralAbilityTest,NonverbalGeneralAbilityTest,ReadingComprehensionTest,MathematicalTest,andCopyingDesignsTest).Atage16,theytaketwocognitiveIntelligenceandhomosexuality Fig.2.Bivariateassociationsbetweenchildhoodintelligenceandadultcohabitation.(a)Meanlifetimenumberofhomosexualsexpartnersbycognitiveclass.(b)Meanlife-timenumberofheterosexualsexpartnersbycognitiveclass.Errorbarsrepresentthestandarderrorforthemean.S.Kanazawa Onceagain,becausetheassociationbetweenintelligenceandhomosexualbehav-iourmaydifferbysex,anintelligencesexinteractiontermisincludedinthenegativebinomialregressionequation,presentedinTable3,column(2).Itssigni“cantly0.001)positiveassociationsuggeststhattheassociationbetweenintelligenceandthenumberofhomosexualpartnersissigni“cantlystrongeramongmenthanamongwomen.Infact,netofthesamecontrolvariables,intelligenceissigni“cantlynegativelyassociatedwiththenumberofhomosexualpartnersamongwomen(0.008,0.01),whileitisverystronglyandsigni“cantlypositivelyassociatedwiththenumberofhomosexualpartnersamongmen(0.001).Thecomparisonoftheunstandardizedcoef“cientsrevealsthatthepositiveassociationamongmenisnearly“vetimesasstrongasthenegativeassociationamongwomen.Table3,column(3),presentstheresultsofanegativebinomialregressionanalysisofthelifetimenumberofheterosexualpartners.Theyshowthat,contrarytothepredic-tionoftheHypothesis,intelligenceissigni“cantlypositivelyassociatedwiththenumberofheterosexualpartners.Becauseheterosexualbehaviouriseminentlyevolutionarilyfamiliar,theHypothesiscannotexplainthepositiveassociationbetweenintelligenceandthenumberofheterosexualpartnersintheGSS.However,theassociationbetweenintelligenceandthenumberofhomosexualpartnersistwiceasstrongasitsassociationwiththenumberofheterosexualpartners(vs.0.011).AsFig.2shows,theassociationbetweenintelligenceandthelifetimenumberofhomosexualpartnersismonotonicandverystrong,whereasitsassociationwiththelifetimenumberofheterosexualpartnersisnotmonotonicandismuchweaker.Verybrightindividuals(withIQ125)havehadeighttimesasmanyhomosexualpartnersasverydullindividuals(withIQ75)(2.42vs.0.31).Insharpcontrast,verybrightindividualshavehadlessthan40%moreheterosexualpartnersthanverydullindivid-uals(9.79vs.7.10).Infact,brightindividuals(with110125)havehadmoreheterosexualpartnersthanverybrightindividuals.Theinteractioneffectbetweensexandintelligenceonthenumberofheterosexualpartnersisstatisticallysigni“cantly(0.05)negative,whichsuggeststhatthepositiveassociationbetweenintelligenceandthenumberofheterosexualpartnersissigni“cantlystrongeramongwomenthanamongmen.Infact,netofthesamecontrolvariables,intelligenceistwiceasstronglyassociatedwiththenumberofheterosexualpartnersamongwomen(0.014)asamongmen(0.007)(bothStudy3Data:NationalChildDevelopmentStudy(NCDS).AddHealthandGSShaveoneimportantshortcomingforthepurposesofthisstudy:theyonlyhavemeasuresofverbalintelligence,notgeneralintelligence.Further,theirsamplesarelimitedtocon-temporaryAmericans.Inordertomakesurethattheassociationbetweenchildhoodintelligenceandadulthomosexualityislimitedneithertoverbalintelligencenorcon-temporaryAmericansnorthespeci“cmeasuresofsexualbehaviouremployed,Inowuseadifferentmeasureofhomosexualitywithdatafromanothernation,whichhaveaverystrongmeasureofgeneralintelligence.Intelligenceandhomosexuality Table3.Theassociationbetweenintelligenceandthelifetimenumberofhomosexualandheterosexualpartners,GeneralSocialSurveysLifetimenumberofhomosexualpartnersLifetimenumberofheterosexualpartners(1)(2)(3)(4)Intelligence0.022***(0.002)0.009***(0.003)0.011***(0.001)0.013(0.001)Sex(1ifmale)1.584***(0.045)2.882***(0.322)1.147***(0.024)1.564***(0.167)Intelligencesex0.043***(0.003)0.004*(0.002)Age0.016***(0.002)0.017***(0.002)0.002**(0.001)0.002**(0.001)Race(1ifblack)0.060(0.067)0.109(0.066)0.471***(0.037)0.476***(0.037)Socialclass0.208***(0.033)0.201***(0.033)0.068**(0.018)0.067***(0.018)Education0.011(0.009)0.009(0.009)0.014**(0.005)0.014**(0.005)Earnings0.006*(0.003)0.008**(0.003)0.021***(0.002)0.021***(0.002)Currentlymarried(1ifyes)0.394***(0.055)0.392***(0.055)0.439***(0.028)0.437***(0.028)Evermarried(1ifyes)1.146***(0.061)1.130***(0.062)0.405***(0.036)0.404***(0.036)Numberofchildren0.002(0.017)0.016(0.017)0.030***(0.008)0.029***(0.008)ReligionCatholic10.605***(0.106)10.510***(0.107)0.168**(0.059)0.167**(0.059)Protestant0.936***(0.102)0.844***(0.103)0.230***(0.058)0.230***(0.058)Jewish10.354(0.158)10.270***(0.160)10.069***(0.091)10.067***(0.091)Other0.577***(0.134)0.398**(0.135)0.595***(0.085)0.593***(0.085)Religiosity0.161***(0.037)0.123***(0.037)0.236***(0.020)0.237***(0.020)Politicalattitude0.317***(0.015)0.308***(0.015)0.056***(0.008)0.055***(0.008)Year0.061***(0.003)0.065***(0.003)0.020***(0.002)0.020***(0.002)Constant125.303(6.723)130.030(6.752)38.495(3.918)38.525(3.918)Likelihoodratio(df17/18)8482.9298673.1304347.0044353.4099362936294449444Mainentriesareunstandardizedregressioncoef“cients.Entriesinparenthesesarestandarderrors.0.01;***0.001(two-tailed).S.Kanazawa totalscore(thenumberofcorrectresponses)variesfrom0to10.TherawscoreisthentransformedintothestandardIQmetric,withameanof100andastandarddeviationof15.Asmentionedabove,verbalintelligenceisknowntobehighlycorrelatedwithgeneralintelligence(Miner,1957;Wol”e,1980;Huang&Hauser,1998).Itisimpor-tanttonotethat,unliketheAddHealthdataaboveandtheNCDSdatabelow,theGSSrespondentsintelligenceismeasuredatthesametimeastheirnumberofsexualpartners.Itisthereforeimpossibletoestablishthecausalorderunambiguously.How-ever,itwouldbeverydif“culttoimaginehowthenumberofhomosexualandhetero-sexualpartnerscanaffectonesintelligence,whenthelatterislargelyheritableandstableacrosslifeaftertheageof10or11(Dearyetal.,2004).Controlvariables.Inthenegativebinomialregressionequation,Icontrolfortherespondentssex(1ifmale),age(inyears),race(1ifblack),socialclass(1lowerclass,2workingclass,3middleclass,4upperclass),education(yearsofformalschooling),earnings(measuredby12…25equidistantordinalcategoriesheretreatedasinterval),whethercurrentlymarried(1ifyes),whetherevermarried(1ifyes),numberofchildren,religion(measuredbyfourdummiesforCatholic,Protestant,Jewishandother,withnoneasthereferencecategory),religiosity(1noreligion,2somewhatstrong,3notverystrong,4strong),politicalattitude(1extremelyconservative,conservative,3slightlyconservative,4moderate,5slightlyliberal,6liberal,7extremelyliberal)andsurveyyear.ResultsTable3,column(1),presentstheresultsofanegativebinomialregressionanalysisofthelifetimenumberofhomosexualpartners.Theyshowthat,netofsex,age,race,socialclass,education,earnings,whethercurrentlymarried,whetherevermarried,numberofchildren,religion,religiosity,politicalattitudeandsurveyyear,moreintel-ligentindividualshavemorelifetimehomosexualpartnersthanlessintelligentindi-viduals(0.022,0.001).Itisinterestingtonotethat,whilewomenarefarmorehomosexualthanmenintheirsexualself-identityandexpressedattraction,mennone-thelesshavesigni“cantlymorehomosexualpartnersintheirlifetimesthanwomendo.Thismaybeduetothefactthatwomenonaveragecomeoutlaterinlifethanmendo(Grovetal.,2006)orthatmenarerelativelymoreunrestrictedintheirsociosexualorientationthanwomenandasaresulthomosexualmenarefarmoresexuallypromis-cuousthanhomosexualwomen(Gallup&Suarez,1983,pp.317…318).Asexpected,individualswhoarecurrentlymarriedandwhohaveeverbeenmarriedhavesigni“cantlyfewerlifetimehomosexualpartners.Olderindividuals,individualsinlowersocialclasses,andliberalshavehadmorehomosexualpartnersthanyoungerindividuals,individualsinhighersocialclassesandconservatives.Relativetoatheistsandagnostics,allreligiousindividuals(exceptforJews)havehadfewerhomosexualpartners.Veryinterestingly,controllingforreligiousaf“liation,religiosityissigni“cantlypositivelyassociatedwiththenumberofhomosexualpartners.Netofallthecontrolvariables,morestronglyreligiousindividualshavehadmorehomosexualpartnersthanlessstronglyreligiousindividuals.Thesigni“cantlypositiveassociationofsurveyyearmeansthat,duringthe35-yearhistoryoftheGSS,respondentsinmorerecentyearshavehad(oradmitto)morehomosexualpartners.Intelligenceandhomosexuality 125);andverybright(IQ125).Figure1adepictsthebivariateasso-ciationbetweencognitiveclassandadultsexualidentity,treatingtheordinalmeasureofthelatterasinterval.Itshowsthatthereisamonotonicpositiveassociationbetweenchildhoodintelligenceandmeanadultsexualorientation;themoreintelligentAddHealthrespondentsareaschildren,themorehomosexualtheybecomeasadults.How-ever,onlythedifferencesbetweennormalandbrightandbetweenbrightandverybrightarestatisticallysigni“cant.Figure1bdepictsthebivariateassociationbetweencognitiveclassandexpressedadulthomosexualattraction.Itshowsthatthereisamonotonicpositiveassociationbetweenchildhoodintelligenceandthemeanproportionwithadulthomosexualattrac-tion.Nearlytwiceasmanyverybrightchildrenasverydullchildrengrowuptoexperiencehomosexualattraction(0.15vs.0.08).However,onlythedifferencesbetweennormalandbrightandbetweenbrightandverybrightarestatisticallysigni“cant.Study2Data:GeneralSocialSurveys(GSS).TheAddHealthdataareverysuitableformypurposesherebecausetheyareprospectivelylongitudinalandmeasureintelligenceinchildhoodandtheoutcomemeasuresinearlyadulthood.Nevertheless,theyhaveoneshortcoming:theyonlymeasureadultsexualidentity(straight…bisexual…gay)andadultexpressedhomosexualattraction;theydonotmeasureactualhomosexualbehaviour,onlyheterosexualbehaviour.IthereforeusetheGeneralSocialSurvey(GSS)datatosupplementtheanalysisoftheAddHealthdata,toexaminewhetherintelligenceincreaseshomosexualbehaviouraswellasidentityandattraction.TheNationalOpinionResearchCenterattheUniversityofChicagohasadministeredtheGSSeitherannuallyorbienniallysince1972.Personalinterviewsareconductedwithanationallyrepresentativesampleofnon-institutionalizedadultsintheUS.Thesamplesizeisabout1500foreachannualsurvey,andabout3000foreachbiennialone.Theexactquestionsaskedinthesurveyvarybytheyear.Dependentvariable:lifetimenumbersofhomosexualandheterosexualpartners.manysurveyyears,theGSSasksitsrespondentshowmanymenandwomentheyhaveeverhadsexwithsincetheir18thbirthday.Thesevariablesareusedtoconstructthelifetimenumbersofhomosexualandheterosexualpartners.OftheGSSrespondents,5.5%reportatleastonelifetimehomosexualpartner,and3.7%reportmorethanone.Thesemeasurescorrespondtothesecondde“nitionofhomosexuality…actualsexualbehaviour(withwhomindividualshavesex)…byWilson&Rahman(2005).Becausethelifetimenumberofbothhomosexualpartners(186.70)andhetero-sexualpartners(810.71)arecountmeasuresthatsufferfromextremeoverdispersion,Iusethenegativebinomialregressiontoanalyseit(Hilbe,2007).Independentvariable:verbalintelligence.TheGSSmeasurestheverbalintelligenceofitsrespondentsbyaskingthemtoselectasynonymforawordoutof“vecandidates.HalfoftherespondentsineachGSSsampleanswer10ofthesequestions,andtheirS.Kanazawa Fig.1.Bivariateassociationsbetweenchildhoodintelligenceandhomosexuality.(a)Meanadultsexualidentitybycognitiveclass.(b)Meanproportionwithadultexpressedhomosexualattractionbycognitiveclass.Errorbarsrepresentthestandarderrorforthemean.Intelligenceandhomosexuality Table2.Theassociationbetweenchildhoodintelligenceandadultexpressedhomosexualattraction,AddHealthAdultexpressedhomosexualattraction(1)(2)Childhoodintelligence0.010(0.003)0.016(0.003)Sex(1ifmale)1.042***(0.071)0.990*(0.480)Childhoodintelligencesex0.020***(0.005)Age0.037(0.020)0.034(0.020)RaceAsian0.397**(0.135)0.397**(0.135)Black0.038(0.085)0.040(0.085)NativeAmerican0.039(0.136)0.022(0.136)Maritalstatus(1ifcurrentlymarried)0.334***(0.100)0.338***(0.100)Parentalstatus(1ifparent)0.254**(0.092)0.230*(0.092)Education0.113***(0.019)0.114***(0.019)Earnings0.001(0.002)0.001(0.002)ReligionCatholic0.483***(0.099)0.488***(0.099)Protestant0.792***(0.126)0.795***(0.126)Jewish0.020(0.287)0.020(0.290)Other0.304***(0.092)0.305***(0.092)Religiosity0.084*(0.041)0.085*(0.042)Politicalattitude0.506***(0.044)0.504***(0.044)Constant3.100(0.501)3.677(0.521)Likelihoodratio(df16/17)576.609594.48212,40912,409Mainentriesareunstandardizedregressioncoef“cients.Entriesinparenthesesarestandarderrors.0.05;**0.01;***0.001(two-tailed).S.Kanazawa forJews)aresigni“cantlylesslikelytoidentifythemselvestobehomosexual.Religiousindividualsaresigni“cantlylesshomosexualintheiradultsexualidentity,andliberalsaresigni“cantlymorehomosexualthanconservatives.Itisinstructivetonotethat,whileeducationandchildhoodintelligencearesigni“-cantlypositivelycorrelated(0.3320,0.0001,14,429),educationandchild-hoodintelligencehaveoppositeassociationswithadultsexualidentity;moreintelligentindividualsaremorelikelytoidentifythemselvestobehomosexual,whilemoreeducatedindividualsarelesslikelytodoso.Eventhoughanearlierstudyshowsthatchildhoodgeneralintelligenceissigni“cantlypositivelyassociatedwithadultliberalismandsigni-“cantlynegativelyassociatedwithadultreligiosity(Kanazawa,2010a),childhoodintelligenceisstillsigni“cantlypositivelyassociatedwithadultsexualidentityevennetofliberalismandreligiosity.Becausetheassociationbetweenchildhoodintelligenceandadultsexualidentitymaydifferbysex,achildhoodintelligencexsexinteractiontermisenteredintotheequation.Table1,column(2),showsthattheinteractiontermisnotsigni“cant.Itsuggeststhatchildhoodintelligencehasasimilarassociationwithbothmensandwomensadultsexualidentity.Table2presentstheresultsofbinarylogisticregressionanalysisofadultexpressedhomosexualattraction.Theyshowthat,netofthesamecontrolvariablesasbefore,childhoodintelligenceissigni“cantlypositivelyassociatedwithadultexpressedhomo-sexualattraction.Theassociationsofcontrolvariableswithadultexpressedhomo-sexualattractionarenaturallysimilartotheirassociationswithadultsexualidentity,exceptforrace;itisAsians,notblacks,whoaresigni“cantlylesshomosexualwhenmeasuredbyadultexpressedhomosexualattraction.Onestandarddeviationincreaseinchildhoodintelligence(15IQpoints)isassociatedwithgreateroddsofexpressedadulthomosexualattractionby16%((0.0101.1618).Consistentwithearlier“ndingsoftheirgreatersexual”uidity(Diamond,2008),womenaresigni“cantlyandstronglymorelikelytoexpresshavingexperiencedadulthomosexualattraction;womenhavenearlythreetimestheoddsofexpressingitasdomen(1/1.0422.8349).Bothcurrentlymarriedindividualsandparentsaresigni“-cantlylesslikelytoexpresshavingexperiencedadulthomosexualattraction,asaremoreeducatedindividualsandallreligiousindividuals(exceptforJews).AswithadultsexualidentityinTable1,adultexpressedhomosexualattractionisnegativelyasso-ciatedwithreligiosityandpositivelyassociatedwithliberalpoliticalattitude,butchildhoodintelligenceisstillsigni“cantlypositivelyassociatedwithadultexpressedhomosexualattractionnetoftheseconfounds.Table2,column(2),includesachildhoodintelligencebysexinteractionterm.Itssigni“cantly(0.001)negativeassociationsuggeststhattheassociationbetweenchildhoodintelligenceandadulthomosexualattractionissigni“cantlystrongeramongwomenthanamongmen.Netofthesamecontrolvariables,childhoodintelligenceissigni“cantlypositivelyassociatedwithadultexpressedhomosexualattractionamongwomen(0.017,SE0.003,0.001),butnotamongmen(0.006,SE0.004,ns).ForthepurposeofgraphicpresentationinFig.1only,theAddHealthsampleisdividedinto“vecognitiveclasses(Herrnstein&Murray1994)bychildhoodintelli-gence:verydull(IQ75);dull(7590);normal(90110);brightIntelligenceandhomosexuality Table1.Theassociationbetweenchildhoodintelligenceandadultsexualidentity,AddHealthAdultsexualidentity(1)(2)Childhoodintelligence0.013***(0.003)0.015***(0.003)Sex(1ifmale)1.095***(0.070)0.406(0.498)Childhoodintelligencesex0.007(0.005)Age0.035(0.020)0.034(0.020)RaceAsian0.147(0.118)0.146(0.118)Black0.291**(0.089)0.291**(0.089)NativeAmerican0.060(0.134)0.053(0.134)Maritalstatus(1ifcurrentlymarried)0.404***(0.101)0.405***(0.101)Parentalstatus(1ifparent)0.292**(0.093)0.284**(0.093)Education0.084***(0.018)0.084***(0.018)Earnings0.004(0.003)0.004(0.003)ReligionCatholic0.442***(0.094)0.444***(0.094)Protestant0.778***(0.123)0.779***(0.123)Jewish0.450(0.292)0.451(0.293)Other0.297***(0.089)0.298***(0.089)Religiosity0.166***(0.041)0.166***(0.041)Politicalattitude0.613***(0.044)0.612***(0.044)Threshold13.925(0.493)4.106(0.510)25.231(0.495)5.413(0.513)36.012(0.498)6.194(0.516)46.556(0.502)6.738(0.520)Likelihoodratio(df16/17)811.366813.29812,30012,300Mainentriesareunstandardizedregressioncoef“cients.Entriesinparenthesesarestandarderrors.0.01;***0.001(two-tailed).S.Kanazawa analysedinStudies2and3below,asAddHealthunfortunatelydoesnotaskitsrespondentsabouthomosexualbehaviour,onlyheterosexualbehaviour.Independentvariable:childhoodintelligence.AddHealthmeasuresrespondentsintelligencewiththePeabodyPictureVocabularyTest(PPVT).Therawscores(0…87)areage-standardizedandconvertedtotheIQmetric,withameanof100andastandarddeviationof15.ThePPVTisproperlyameasureofverbalintelligence,notgeneralintelligence.However,verbalintelligenceisknowntobehighlycorrelatedwith(andthusheavilyloadson)generalintelligence.Miners(1957)extensivereviewof36studiesshowsthatthemediancorrelationbetweenvocabularyandgeneralintelligenceis0.83.Wol”e(1980)reportsthatthecorrelationbetweenafull-scaleIQtest(ArmyGeneralClassi“cationTest)andtheGSSsynonymsmeasure(whichisusedlaterinStudy2)is0.71.Asaresult,theGSSsynonymsmeasurehasbeenusedwidelybyintelligenceresearcherstoassesstrendsingeneralintelligence(Huang&Hauser,1998).Inordertoestablishthedirectionofcausalitymoreclearly,IusethemeasureofintelligencetakeninWaveI(in1994…1995whentherespondentsareinjuniorhighandhighschool)topredicttheiradulthomosexualityinWaveIII(in2001…2002whentherespondentsareintheirearlyadulthood).Despitethefactthatthecorrelationbe-tweenmeasuresofintelligenceatWavesIandIII(takensevenyearsapart)isnotextremelyhigh(0.5844,13,943),allthesubstantiveconclusionsremainexactlythesameifIuseWaveIIImeasureofintelligence,oralatentfactorforchildhoodgeneralintelligenceextractedfromtheWaveIandWaveIIImeasures.Controlvariables.Inthemultipleregressionanalysis,Icontrolforthefollowingvariables:age(eventhoughthereisverylittlevarianceinitgiventhatthesearecohortdata);sex(1ifmale);race(withthreedummiesforAsian,blackandNativeAmerican,withwhiteasthereferencecategory);maritalstatus(1ifcurrentlymarried);parentalstatus(1ifparent);education(yearsofformalschooling);earnings(inUS$1K);reli-gion(withfourdummiesforCatholic,Protestant,Jewishandother,withnoneasthereferencecategory);religiosity(Howimportantisreligiontoyou?:0noreligion,notimportantatall,2fairlyunimportant,3fairlyimportant,and4important);andpoliticalattitudes(Intermsofpolitics,doyouconsideryourselfconservative,liberal,ormiddle-of-the-road?:1veryconservative,2conservative,middleoftheroad,4liberal,and5veryliberal).ResultsTable1,column(1),presentstheresultsofordinalregressionanalysisofAddHealthrespondentsadultsexualidentity.Theyshowthat,netofsex,age,race,maritalstatus,parentalstatus,education,earnings,religion,religiosityandpoliticalattitude,childhoodintelligenceissigni“cantly(0.001)positivelyassociatedwithadultsexualidentity;themoreintelligentAddHealthrespondentsareintheirchildhood,themorehomosexualtheyareintheiradultsexualidentity.Menaresigni“cantlylesshomo-sexualintheiradultsexualidentity,asareblacks.Asexpected,currentlymarriedindi-vidualsandparentsaresigni“cantlylesslikelytobehomosexual,asaremoreeducatedindividuals.Relativetoatheistsandagnostics,thosewithreligiousaf“liations(exceptIntelligenceandhomosexuality thattherearenostatisticallysigni“cantdifferencesinestimatedfull-scaleIQamongstraightmen,straightwomenandgaymen.However,amongheterosexuals,they“ndthatrecalledchildhoodgendernon-conformityispositivelyassociatedwithadultintelli-gence.Moremasculinegirlsandmorefeminineboysgrowuptohavehigherintelligenceasadultsthantheirmoregender-conformingcounterparts.Inhisrecentsurveyofthecausesandcorrelatesofsexualorientation,LeVay(2010,pp.113…114)discussesonlyonestudy(Weinrich,1978)ontheassociationbetweenintelligenceandsexualorientation.Inaddition,LeVaynotes,Itseemsverypossible,however,thatthoseearlystudiessufferedfromvolunteerbias,suchthatonlyrela-tivelyintelligentgaypeoplewereavailableforstudy.Inordertoavoidsuchsampleselectionbias,Iusenationallyrepresentativesamplesinallofmystudiesbelow.Study1Data:NationalLongitudinalStudyofAdolescentHealth(AddHealth).TheNationalLongitudinalStudyofAdolescentHealth(AddHealth)isaprospectivelylongitudinalstudyofalarge,nationallyrepresentativesampleofAmericanyouths.Asampleof80highschoolsand52middleschoolsfromtheUSwasselectedwithunequalprobabilityofselection.Incorporatingsystematicsamplingmethodsandimplicitstrati“cationintotheAddHealthstudydesignensuresthissampleisrepresentativeofUSschoolswithrespecttoregionofcountry,schoolsize,schooltypeandethnicity.Asampleof20,745adolescentswerepersonallyinterviewedintheirhomesin1994…1995(WaveI)andagainin1996(WaveII;14,738).In2001…2002,15,197oftheoriginalWaveIrespondents,nowaged18…28,wereinterviewedintheirhomes.MysampleconsistsofWaveIIIrespondents.Dependentvariable.Iusetwodifferentmeasuresofhomosexuality.BothmeasuresareusedinWaveIIIwhentherespondentsareintheirearlyadulthood.First,AddHealthasksitsrespondents:Pleasechoosethedescriptionthatbest“tshowyouthinkaboutyourself:1100%heterosexual(straight)(13,466);2mostlyheterosexual(straight),butsomewhatattractedtopeopleofyourownsex(1017);3bisexual…thatis,attractedtomenandwomenequally(245);4mostlyhomosexual(gay),butsomewhatattractedtopeopleoftheoppositesex(96);5100%homosexual(gay)(131).Icallthisvariableameasureofadultsexualidentity,andanalyseitwithanordinalregressionmodel.Second,AddHealthasksitsrespondentstwoquestions:Haveyoueverhadaromanticattractiontoafemale?andHaveyoueverhadaromanticattractiontoamale?Fromthesequestions,Iconstructabinaryvariableadultexpressedhomosexualattraction,whichis1iftherespondentexpresseseverhavinghadaromanticattractiontoamemberofthesamesex,and0ifotherwise.OftheAddHealthrespondents,9.3%1416)reporthavingeverhadaromanticattractiontomembersofthesamesex.Ianalysethisdependentvariablewithabinarylogisticregressionmodel.ThesetwomeasuresofhomosexualityinAddHealthcorrespondtoself-identi“edlabelsandself-reportedsexualfeelingsinWilson&Rahman(2005).ActualsexualbehaviourwillbeS.Kanazawa bestiality,particularlyonrockart,arerare,(Nash,2001,p.44).Imaythereforesurmisethathomosexualbehaviourintheancestralenvironmentwascorrespondinglyrare.Itisveryimportanttopointout,however,thatevenveryextensiveethnographies,basedonlong-term“eldworkbyveryexperiencedanthropologistsfamiliarwiththelocalculture,maynotalwaysdetectinstancesofhomosexuality,especiallyifitiscondemnedandnegativelysanctionedinthelocalculture.Sotheabsenceofreferencestohomosexualityintheseethnographiesisnotbyitselfconclusiveevidenceofitsabsenceintraditionalsocieties.However,thesameethnographersandanthropologistshavenonethelessbeenadeptatuncoveringevidenceofothernegativelysanctionedandconcealedbehaviourlikemurder,theft,infanticideandextramaritalaffairs.Sotheneartotalabsenceofanydocumentationofhomosexualbehaviourasanindividualchoicemaysuggestthatitmayberelativelyrareinsuchsocieties.Itmayalsosuggestthatwidespreadpracticeofhomosexualbehaviourmayhavebeenrareintheancestralenvironment,anditmaythereforebeevolutionarilynovel.Whilesomeformofhomosexualityisobservedinmanyspecies(Bagemihl,2000),thebasicbiologicaldesignofallspeciesisheterosexualreproduction,andexclusiveandpredominanthomosexualityisrareinnature.Kirkpatricks(2000)surveyofhomo-sexualityintraditionalsocietiesthroughouttheworldalsosuggeststhatvirtuallyallinstancesofhomosexualitywereconcurrentwithheterosexualbehaviour.Mostimpor-tantly,wearenotdescendedfromancestorswhowereexclusivelyhomosexual,soitisunlikelythathomosexualorientationhasbeenpartofhumannaturethroughoutevolu-tionaryhistory.Somedisagree,however,andsuggestthathomosexualbehaviourmayhavebeenadaptiveintheancestralenvironment.Kauth(2000),Kirkpatrick(2000)andMuscarella(2000)allvariouslyarguethathomosexualattractionandbehaviourmighthavefacili-tatedsame-sexcoalitionsandaf“liations,whichmayhavebeencrucialtoourancestorssurvivalandreproductivesuccess.Mycontentionthathomosexualbehaviourisevolu-tionarilynovelisalsoinconsistentwithevidencethatsuggeststhathomophobia…negativeattitudestowardhomosexuals…maybeanevolvedpsychologicalmechanism(Gallup,1995).Ifhomosexualidentityandbehaviourareevolutionarilynovel,thentheSavanna-IQInteractionHypothesiswouldpredictthat,regardlessoftheirtruesexualorientation,moreintelligentindividualsmaybemorelikelytoidentifythemselvesashomosexual,reporthomosexualfeelingsanddesires,andengageinhomosexualbehaviourthanlessintelligentindividuals.Itestthispredictionbelowwiththreelarge,nationallyrepresen-tativesamples(twoofwhichareprospectivelylongitudinal)fromtheUnitedStatesandtheUnitedKingdom.Weinrichs(1978)meta-analysisoftherelationshipbetweensexualorientationandintelligenceshowsthathomosexualsaregenerallymoreintelligentthanheterosexuals,exceptinsomesamplesofprisoners.Tuttle&Pillard(1991)studiedasmallsampleofhomosexualandheterosexualmenandwomenandconcludethatthereisnodifferenceinintelligencebetweenhomosexualsandheterosexuals.Arabsheibanietal.(2005)“ndthathomosexualsintheUnitedKingdomearnmorethanheterosexuals,althoughgaymenmakelessthanstraightmenoncetheirhumancapitaliscontrolled.Theirstudydoesnotmeasureindividualsintelligence,however.Rahmanetal.(inpress)showIntelligenceandhomosexuality (Blanchard&Bogaert,1996;Bogaert&Blanchard,1996),thesamongtheAchemighthavebeengeneticallyandhormonallypredisposedtohomosexuality.ButHill&Hurtadomakeitclearthattheynonethelessdonotengageinhomosexualbehaviour.Inarecentin-depthethnographicstudyofAkaforagersandNgandufarmersincentralAfrica,Hewlett&Hewlett(2010)reportthathomosexualityisunknownorrareinbothpopulations.TheAkahavedif“cultyunderstandingtheconceptofhomo-sexualityanddonothaveawordforit.TheNganduarefamiliarwiththeconceptbuthavenowordforitandmaintainthatitdoesnotexistintheirpopulation.Evencomprehensivesurveysofhomosexualitymakelittlereferencetoitintraditionalsocieties.Crapo(1995)distinguishesfourdifferenttypesofhomosexuality:1)pederastyormentorship,inwhichthereisasigni“cantagedifferencebetweenthepartners,(Crapo,1995,p.184);2)pathecism,inwhichthenon-dominantpartnerundergoesarolechange(including,e.g.transvestismorotherformsofgendermixing),(p.184);3)homophilia,inwhichadultpartnersofequivalentagebothmaintainthegenderrolesthatareusuallyassignedtothoseoftheirbiologicalsex,(pp.184…185);and4)youthfulexperimentation,inwhichadolescentsareinvolvedinhomosexualrelationshipspriortotheirentryintoadultstatuswhentheyareexpectedtobeginaheterosexualmarriagecareer,(p.185).WhatCrapocallshomophiliaisclosesttowhatImeanbyhomo-sexualityinthispaper,yetCraposextensivesurveyofhomosexualityintraditionalsocietiesintheStandardCross-CulturalSampleisentirelyaboutthe“rsttwotypesofhomosexualityanddoesnotatalldiscussthelattertwo.Ina500-pagecompendiumandencyclopedicreviewofallinstancesofhomosexualityinrecordedhumanhistorythroughouttheworld,Murray(2000)devotesonlyfourpagestoegalitarianfemalehomosexualities(pp.359…360)andegalitarianmalehomosexual-ities(pp.363…365)intribalsocietiesinsub-SaharanAfrica,enumeratingonlyfourinstancesofhomosexualityreportedbyfourethnographers.GiventhatMurraysbookisotherwisetrulycomprehensiveandincludeshundredsofexamples,theneartotalabsenceofanymentionofhomosexualbehaviourinsub-SaharanAfricaisremarkable.Nash(2001)claimsthathomosexualactsaredepictedinsomeMesolithiccaveart.However,hisevidencecomessolelyfromtwocavepaintings,onefromNorwayandtheotherfromSpain.Thesepaintingsappeartodepictactsofsexualintercourseintherear-entrypositionandfellatio,respectively.Theyarethereforenotunambiguousdepictionsofhomosexualbehaviour.Infact,itappearsthatNashistheonlypersonwhointerpretsthesepaintingsasdepictinghomosexualbehaviour:Twoofthehuman“guresarelockedinrear-entrysexualintercourseandhavepreviouslybeeninterpretedasbeingmaleandfemalebyHallstrom.Thesmaller“guremaypossiblyhavebreasts,but,equally,thetwolinesherecouldwellbeapairofarms....Thepositionofthepenissuggeststhatpenetrationisviatheanus.GustaveHallstrom,however,regardstheareaofpenetrationasthevulva.(Nash,2001,pp.47…48)Nashdeclares,withoutanylogicorevidence,thatthe“gureperformingthefellatiointheSpanishpaintingismale.However,Idonotsee(andNashdoesnotprovide)anyreasonwhythis“guremustnecessarilybemale.Theartcouldverywelldepictaheterosexualactoffellatio.Infact,Nashhimselfnotes:Imagesofsexualitywithinprehistoryarenotuncommon,while,incontrast,scenesdepictinghomosexualityandS.Kanazawa ancestorsintheancestralenvironment,theyarethebestanaloguesthatwehaveavail-ableforcloseexaminationandarethusoftenusedforthepurposeofmakinginferencesaboutourancestrallife.Theten-volumecompendiumTheEncyclopediaofWorldCultures(Levinson,1991…1995),whichextensivelydescribesallhumanculturesknowntoanthropology(morethan1500),mentionsmalehomosexualityinsevendifferentcultures(Foi,Gebusi,Kaluli,Keraki,Kiwai,Marind-animandSambia).However,thesearephylogeneticallycloselyrelatedtribesallinPapuaNewGuinea,andallpracticesofhomosexualityinthesePapuaNewGuineanculturesoccurlargelyaspartofinitiationritesforboys.So,forexample,Gebusibelieveboysmustbeorallyinseminatedtoobtainmalelifeforceandattainadulthood.Inseminationcontinuesduringadolescenceandculminatesinthemaleinitiation(wakawala,orchildbecomingbig)betweenages17and23,(Levinson,1991…1995,Vol.2,p.79).AndamongtheSambia,Malematurationrequireshomoeroticinseminationtoattainbiologicalcompetence.Initiationritualsthusinvolvecomplexhomosexualcontactfromlatechildhooduntilmarriage,whenitstops,(Levinson,1991…1995,Vol.2,p.285).SuchhomosexualpracticesinPapuaNewGuineaappearhighlyritualizedandculturallymandated.Thereappearsverylittleindividualchoiceinvolvedand,assuch,homosexualitydoesnotappeartobeanindividual-differencevariable(wheresomepeoplepractiseitwhileothersdont).Itthereforeappearsquitedifferentfromwhatwenormallymeanbysexualrelations,whichinvolvechoice,emo-tionsandattachment.Atanyrate,itisverydif“culttosuggestthathomosexualitywasroutinepartofourancestorslifeifitspresent-daypracticeamongtraditionalsocietiesislimitedonlytooneislandintheSouthPaci“cfaroutsideoftheancestralenvironmentofsub-SaharanAfrica.Inaddition,Ihaveconsultedthefollowingextensive(monograph-length)ethnogra-phiesoftraditional(hunter-gatherer,pastoralandhorticultural)societiesaroundtheworld:Yanomamo(Chagnon,1992);FromMukogodotoMaasai:EthnicityandCulturalChangeinKenya(Cronk,2004);AcheLifeHistory:TheEcologyandDemographyofaForagingPeople(Hill&Hurtado,1996);The!KungSan:Men,Women,andWorkinaForagingSociety(Lee,1979);andSachaRuna:EthnicityandAdaptationofEcuadorianJungleQuichua(WhittenJr,1976).Inalloftheseethnographies,thereisnomentionofexplicithomosexualrelationshipsamongthemembersofthesocietiesunderstudy.TheonlypotentialexceptionisthepanegiamongtheAche(Hill&Hurtado,1996,pp.276…277;emphasisadded).Somemeninoursampleneverhadanychildrenandothersneveracquiredawife.OnecategoryofmeninAchesocietyoptsoutofthemalematingpoolaltogether.Thesemen,calledpanegi,takeonafemalesocioeconomicrole(thewordpanemeansunsuccessfulorunluckyathunting).Menwhoarepanegigenerallydonothunt,butinsteadcollectplantresourcesandinsectlarvae.Theyweavebaskets,matsandfans,andmaketoothnecklaces,bowstringsandotherfemalehandicrafts.Theyspendlonghourscooking,collecting“re-woodorwater,andcaringforchildren.Mostinformantsstatedthatpanegisdidnoteverengageinhomosexualbehavior(oraloranal)priorto“rstcontact.Afewinformantssaidtheywerenotsure,buthadneverheardofsuchbehavior.Nowgiventhatpanegisareapparentlysmallinstature(Hill&Hurtado,1996,p.277)and,atleastinNorthAmerica,homosexualmenareshorterthanheterosexualmenIntelligenceandhomosexuality De“nitionsandmeasuresofhomosexualityMustanskietal.(2002,pp.122…127)andWilson&Rahman(2005,pp.13…16)enumeratefourdifferentmeasuresofsexualorientation:1.Self-identi“edlabels(homosexual,bisexual,heterosexual)2.Actualsexualbehaviour(withwhomindividualshavesex)3.Self-reportedsexualfeelings(fantasiesanddesires)4.Genitalorbrainresponses(physiologicallymeasuredarousaltomaleorfemaleimages)Wilson&Rahman(2005,pp.13…16)notethatself-identi“edlabelscanbein”uencedbypoliticsandculturalclimate(manyhomosexualsthroughouthistoryhavebeenforcedtoremainintheclosetduetosocialpressureandthreatoflegalpunishment),andthatactualsexualbehaviourcanbein”uencedbyopportunitiesandcircumstances(hetero-sexualmenoftenhavesexwithothermenwhileinprisonduetothecompleteabsenceofpotentialfemalesexualpartners).Incontrast,sexualfeelingsandphysiologicalmea-suresaremorestableandclosertoindividualstruesexualorientation;forexample,self-identi“edheterosexualmenwhoareopenlyhomophobicmaynonethelessshowgenitalresponseofarousaltosexualimagesofothermen(Adamsetal.,1996).Wilson&Rahman(2005,p.15)alsonotethathomosexualfantasiesarequitecommoninheterosexualmenandwomenasaformofmentalexplorations,andthatmeasuringhomosexualitywithreportedsexualfantasiesanddesiresassumesthatsurveyrespondentsarecompletelyhonest.Allinall,Wilson&Rahmanconcludethatphysiologicallymea-suredarousal(genitalorbrainresponsestosexualimagesofmenorwomen)isprobablythemostaccuratemeasureofsexualorientation,andtheotherthreemeasuresmaycorrelatepoorlywithitandmaydeviatefromtheirtruesexualorientation,especiallyamongwomen(Chiversetal.,2007),althoughmosthomosexualmentendtobeexclu-sivelyhomosexual(Belletal.,1981).Giventhatanindividualstruesexualorientation,atleastformen,maybeprenatallydetermined,eitherbygeneticorprenatalhormonalfactors(Ellis&Ames,1987;Bailey&Pillard,1991;Kirketal.,2000),itisnotlikelythatmoreintelligentindividualsaremorelikelytobehomosexual.Thereisapossibility,however,thatthe(asyetundiscovered)genesforintelligencearesomehowlinkedtothe(asyetundiscovered)genesforhomosexuality,asgenesforbothintelligenceandhomosexualityappeartobelocatedonthechromosomeXq28(Hameretal.,1993;Turner,1996).Giventhatthe“rstthreemeasuresofsexualorientationaremoremalleableandsubjecttocon-sciouschoiceandself-presentation,itmayalsobepossiblethatmoreintelligentindivid-ualsaremorelikelytoappearhomosexualbythesemeasures,thatis,homosexualidentityandbehaviourareevolutionarilynovel.Regardlessoftheirsexualorien-tation,moreintelligentindividualsmayidentifythemselvesashomosexual,engageinhomosexualbehaviourorreporthomosexualfantasiesanddesires.EvolutionarynoveltyofhomosexualidentityandbehaviourInordertoascertaintheextenttowhichourancestorsmighthaveengagedinhomo-sexualbehaviour,Ihaveconsultedethnographicrecordsoftraditionalsocietiesthrough-outtheworld.Whilecontemporaryhunter-gatherersarenotexactlythesameasourS.Kanazawa Hellberg,2010).Thisispossiblybecausethehumanconsumptionofsuchpsychoactivesubstancesisevolutionarilynovel,alloriginatinglessthan10,000yearsago.Finally,criminalsonaveragehavelowerintelligencethanthegeneralpopulation(Wilson&Herrnstein,1985;Herrnstein&Murray,1994).ThisisconsistentwiththeHypothesisbecause,whilemuchofwhatwecallinterpersonalcrimetodayisevolu-tionarilyfamiliar,theinstitutionsthatcontrol,detectandpunishsuchbehaviourareevolutionarilynovel(Kanazawa,2009).Murder,assault,robberyandtheftwereprob-ablyroutinemeansofintrasexualmalecompetitionforresourcesandmatesintheancestralenvironment.Wemayinferthisfromthefactthatbehavioursthatwouldbeclassi“edascriminalifengagedinbyhumansarequitecommonamongotherspecies(Ellis,1998),includingotherprimates(deWaal,1989,1992;deWaaletal.,1993).However,therewasverylittleformalthird-partyenforcementofnormsintheancestralenvironment,onlysecond-partyenforcement(byvictimsandtheirkinandallies)orinformalthird-partyenforcement(ostracism).ItthereforemakessensefromtheperspectiveoftheHypothesisthatmenwithlowintelligencemaybemorelikelytoresorttoevolutionarilyfamiliarmeansofcompeti-tionforresources(theftratherthanfull-timeemployment)andmatingopportunities(raperatherthancomputerdating)andnottocomprehendfullytheconsequencesofcriminalbehaviourimposedbyevolutionarilynovelentitiesoflawenforcement.Italsoexplainstheexceptionthatprovestherule,whymoreintelligentindividualsaremorelikelytoconsumeillegaldrugs(Kanazawa&Hellberg,2010).Unlikemostinter-personalandpropertycrimes,theconsumptionofsuchsubstancesisevolutionarilynovel.Itsnotlegalitypersethatmatters,butevolutionarynoveltyofthebehaviour.Mostevolutionarypsychologistsandbiologistsconcurthathumanshavenotunder-gonesubstantialevolutionarychangesinthelast10,000years,sincetheendofthePleistoceneEpoch,astheenvironmentduringthisperiodhasnotprovidedastablebackgroundagainstwhichnaturalandsexualselectioncanoperateovermanygenera-tions.ThisistheassumptionbehindtheSavanna-IQInteractionHypothesis.Morere-cently,however,somescientistshavevoicedopinionsthathumanevolutionhascon-tinuedandevenacceleratedduringtheHoloceneEpoch(Evansetal.,2005;Cochran&Harpending,2009).WhilethesestudiesconclusivelydemonstratethatnewalleleshaveindeedemergedinthehumangenomesincetheendofthePleistocene,theimplica-tionsandimportanceofsuchnewallelesforevolutionarypsychologyarenotimmediatelyobvious.Inparticular,withthesoleexceptionoflactosetolerance,itisnotclearwhetherthesenewalleleshaveledtotheemergenceofnewphysicalorpsychologicaladaptationsinthelast10,000years.Inthispaper,IapplytheSavanna-IQInteractionHypothesistoonedomainoflife…sexualbehaviour…andexplainwhy,regardlessoftheirgeneticpredisposition,moreintelligentindividualsmaybemorelikelytoengageinhomosexualbehaviourthanlessintelligentindividuals.Itestmypredictionwithdatafromthreelarge,nationallyrepre-sentativesamples(twoofwhichareprospectivelylongitudinal)fromtheUnitedStatesandtheUnitedKingdom.ConsistentwiththeHypothesis,myanalysesshowthatmoreintelligentindividualsaremorelikelytoidentifythemselvesashomosexual,experiencehomosexualattraction,engageinhomosexualbehaviourandhavemorehomosexualcohabitationpartnersthanlessintelligentindividuals.Intelligenceandhomosexuality havelessdif“cultywithsuchstimulithanlessintelligentindividuals.Incontrast,generalintelligencemaynotaffectindividualsabilitytocomprehendanddealwithevolutionarilyfamiliarentitiesandsituations.Evolutionarilynovelentitiesthatmoreintelligentindividualsarebetterabletocomprehendanddealwithmayincludeideasandlifestylesthatformthebasisoftheirpreferencesandvalues;itwouldbedif“cultforindividualstopreferorvaluesomethingthattheycannottrulycomprehend.Hence,appliedtothedomainofpreferencesandvalues,theHypothesissuggeststhatmoreintelligentindividualsaremorelikelytoacquireandespouseevolutionarilynovelpreferencesandvaluesthatdidnotexistintheancestralenvironmentthanlessintelligentindividuals,butthatgeneralintelligencehasnoeffectontheacquisitionandespousalofevolutionarilyfamiliarpreferencesandvaluesthatexistedintheancestralenvironment.TherehasbeenemergingevidencefortheHypothesisasanexplanationforindividualpreferencesandvalues.First,moreintelligentchildrenaremorelikelytogrowuptoespouseleft-wingliberalism(Dearyetal.,2008;Kanazawa,2010a),possiblybecausegenuineconcernswithgeneticallyunrelatedothersandwillingnesstocontributeprivateresourcesforthewelfareofsuchothers…liberalism…maybeevolutionarilynovel.Eventhoughpaststudiesshowthatwomenaremoreliberalthanmen(Shapiro&Mahajan,1986;Wirls,1986;Lake&Breglio,1992),andblacksaremoreliberalthanwhites(Sundquist,1983;Kluegel&Smith,1989),theeffectofchildhoodintelligenceonadultliberalismistwiceaslargeastheeffectofsexorrace.Second,moreintelligentchildrenaremorelikelytogrowuptobeatheists(Kanazawa,2010a),possiblybecausebeliefinhigherpowers,asaconsequenceofover-inferenceofagencybehindotherwisenaturalphenomena,maybepartofevolvedhumannature(Guthrie,1993;Boyer,2001;Atran,2002;Kirkpatrick,2005;Haselton&Nettle,2006),andatheismmaythereforebeevolutionarilynovel.Eventhoughpaststudiesshowthatwomenaremuchmorereligiousthanmen(Miller&Hoffmann,1995;Miller&Stark,2002),theeffectofchildhoodintelligenceonadultreligiosityistwiceaslargeasthatofsex.Third,moreintelligentboys(butnotmoreintelligentgirls)aremorelikelytogrowuptovaluesexualexclusivity(Kanazawa,2010a),possiblybecausehumanswerenaturallypolygynousthroughoutevolutionaryhistory(Leutenegger&Kelly,1977;Alexanderetal.,1979;Harvey&Bennett,1985;Pickford,1986;Kanazawa&Novak,2005).Eitherundermonogamyorpolygyny,womenareexpectedtobesexuallyexclu-sivetoonemate;insharpcontrast,meninpolygynousmarriagesarenotexpectedtobesexuallyexclusivetoonemate,whereasmeninmonogamousmarriageare.Sotheexpectationofsexualexclusivitymaybeevolutionarilynovelformen,butnotforwomen.Fourth,moreintelligentchildrenaremorelikelytogrowuptobenocturnal,goingtobedandwakinguplater(Kanazawa&Perina,2009),possiblybecausenocturnallifewasrareintheancestralenvironmentwhereourancestorsdidnothavearti“cialsourcesofilluminationuntilthedomesticationof“re.Ethnographiesofcontemporaryhunter-gathererssuggestthatourancestorsmayhavewokenupshortlybeforedawnandgonetosleepshortlyafterdusk.Nightlifemaythereforebeevolutionarilynovel.Fifth,moreintelligentchildrengrowuptoconsumemorealcoholmorefrequently,smokemoretobacco(butonlyintheUS)andusemoreillegaldrugs(Kanazawa&S.Kanazawa J.Biosoc.Sci.,(2012),595…623,CambridgeUniversityPress,2011doi:10.1017/S0021932011000769Firstpublishedonline31Jan2012INTELLIGENCEANDHOMOSEXUALITYSATOSHIKANAZAWA -ournaO of BLosocLaO ScLenceKttp//MournaOs.cambrLdJe.orJ/-BSAddLtLonaO servLces for -ournaO of BLosocLaO ScLenceEmaLO aOerts COLcN KereSubscrLptLons COLcN KereCommercLaO reprLnts COLcN KereTerms of use  COLcN Kere ,NTE//,*ENCE AND +OMOSE;UA/,T