Intelligence and childlessness Satoshi Kanazawa Managerial Economics and Strategy Group - PDF document

Intelligenceandchildlessness SatoshiKanazawa ManagerialEconomicsandStrategyGroup,DepartmentofManagement,LondonSchoolofEconomicsandPoliticalScience,HoughtonStreet,LondonWC2A2AE,UnitedKingdom articleinfoArticlehistory:Received1March2013Revised30May2014Accepted15June2014Availableonline24June2014 Keywords: Demographersdebatewhypeoplehavechildreninadvancedindustrialsocietieswherechildrenareneteconomiccosts.Fromanevolutionaryperspective,however,theimportantquestioniswhysomeindividualschoosenottohavechildren.Recenttheoreticaldevelop-mentsinevolutionarypsychologysuggestthatmoreintelligentindividualsmaybemorelikelytoprefertoremainchildlessthanlessintelligentindividuals.AnalysesoftheNationalChildDevelopmentStudyshowthatmoreintelligentmenandwomenexpresspreferencetoremainchildlessearlyintheirreproductivecareers,butonlymoreintelligentwomen(notmoreintelligentmen)aremorelikelytoremainchildlessbytheendoftheir 1.Introduction Demographershavelongdebatedwhypeoplehavechildren,especiallyinadvancedindustrialsocietieswherechildrenareneteconomiccosts(Friedmanetal.,1994;Schoenetal.,1997).Amongothers,thequestioniskeytounderstanding Explanationsforparenthoodandfertilitydeclinerangefromnormative(Blake,1968;Lesthaeghe,1983;Preston,1986;Ryder,1979;Westoff,1986),toculturalorreligious(HayfordandMorgan,2008;Heaton,1986;Mosheretal.,1992;Pearce,2002;WestoffandBumpass,1973),andtorationalchoice(Becker,1960,1981;ButzandWard,1979;Easterlinetal.,1980;Friedmanetal.,1994;Schoenetal.,1997).Empirically,weknowthatindividualsdesireforandintentionsaboutchildrenstronglyin”uencetheiractualfertilitybehavior(Barber,2001;Williamsetal.,1999);quiteunsurprisingly,individ- Fromanevolutionaryperspective,however,whypeoplehavechildrenisthewrongquestiontoask.Humans,likeallotherspeciesinnature,areevolutionarilydesignedtoreproduce.Itisthereforenotatalltheoreticallyproblematicthatmostindividualschoosetohavechildrenintheirlifetimes,evenwhentheyrepresentneteconomiccosts.Humanshavechildrenbecausetheyareevolutionarilydesignedtodoso.Reproductionistheultimate(albeitlargelyunconscious)goalofall2014ElsevierInc.Allrightsreserved. SocialScienceResearch48(2014)157…170 ContentslistsavailableatScienceDirectSocialScienceResearchjournalhomepage: biologicalexistence,includinghumans,andeverythingelse,likemoneyandwealth,isameanstoit.Thetheoreticalpuzzlefromthisperspectiveiswhysomeindividualschoosenottohavechildrendespitetheirbiologicaldesign. Aleadingevolutionaryexplanationforfertilitydeclineinindustrialsocietiesisthetrade-offtheory(Beauchamp,1994;Kaplanetal.,2002).Thetheoryproposesthatparentsmaximizetheirreproductivesuccessbymaximizingthejointproductofthequantityandqualityofoffspring.Foranygivenlevelofparentalresources,thefewerchildrentheyhave,themoretheycaninvestineach.Sinceindustrialsocietiesrequirethattheparentsinvestheavilyineachchild,theyareforcedtohavefewerchildren,hencethefertilitydeclineafterindustrialization. ArecentempiricaltestintheNetherlands(Kaptijnetal.,2010),however,doesnotsupportthetrade-offtheory.Whilethenumberofchildreninthefamilyisnegativelyassociatedwiththeirquality,measuredbysex-speci“cmatevalue(thefewerchildrenparentshave,thehigherthematevalueofeachchild),thereisalsoanegativeassociationbetweenthechildrensmatevalueandtheirfertility.Inotherwords,takenaltogether,parentswhohavefewerchildrenendupwithfewergrand-childrenthanthosewhohavemorechildren.Thusanytendencytolimitthenumberofchildreninordertoincreasetheirqualitycannotevolve.Perhapsthisresultcouldhavebeenpredictedfromtheearlierbehaviorgenetic“ndingthatfertilityispartlyheritable(Kohleretal.,1999;Rodgersetal.,2001).Thefewerchildrenyouhave,thefewerchildrenyourchildrenaregeneticallypredisposedtohave. Inthispaper,Iproposeanewevolutionarypsychologicalexplanationforwhysomeindividualsdesiretohavechildrenmorethanothers.Thetheorypredictsthatmoreintelligentindividualsarelesslikelytovalueparenthoodandmorelikelytochoosetoremainchildless.Analysesofaprospectivelylongitudinal,large,nationallyrepresentativesamplefromtheUnitedKingdomshowthatmoreintelligentmenandwomenaremorelikely,earlyintheirreproductivecareers,toexpressadesiretoremainchildlessforlife.However,onlymoreintelligentwomen(notmoreintelligentmen)areactuallymorelikelytoremainchildlessforlife. 2.Evolutionaryoriginsofindividualpreferencesandvalues Wheredoindividualspreferencesandvalues,suchasvaluesfororagainstchildren,comefrom?Whydopeoplelikeorwantwhattheydo?Theoriginofindividualpreferencesandvaluesisoneoftheremainingtheoreticalpuzzlesinsocialandbehavioralsciences(Kanazawa,2001 Recenttheoreticaldevelopmentsinevolutionarypsychologymaysuggestonepossibleexplanation().Ontheonehand,evolutionarypsychology(Crawford,1993;Symons,1990;ToobyandCosmides,1990)positsthatthehumanbrain,justlikeanyotherorganofanyotherspecies,isdesignedforandadaptedtotheconditionsoftheancestralenvironment(roughlytheAfricansavannaduringthePleistoceneEpoch),notnecessarilytothoseofthecurrentenvironment.Itmaythereforehavedif“cultycomprehendinganddealingwithentitiesandsituationsthatdidnotexistintheancestralenvironment(Kanazawa,2002,2004a).Ontheotherhand,anevolutionarypsychologicalthe-oryoftheevolutionofgeneralintelligenceproposesthatgeneralintelligencemayhaveevolvedasadomain-speci“cadaptationtosolveevolutionarilynovelproblems,forwhichtherearenopredesignedpsychologicaladaptationsKanazawa,2004b,2008 Thelogicalconjunctionofthesetwotheories,theSavanna-IQInteractionHypothesisKanazawa,2010a),impliesthatthehumanbrainsdif“cultywithevolutionarilynovelstimulimayinteractwithgeneralintelligence,suchthatmoreintelligentindividualshavelessdif“cultywithsuchstimulithanlessintelligentindividuals.Incontrast,generalintelligencemaynotaffectindividualsabilitytocomprehendanddealwithevolutionarilyfamiliarentitiesandsituations. Evolutionarilynovelentitiesthatmoreintelligentindividualsarebetterabletocomprehendanddealwithmayincludeideasandlifestyles,whichformthebasisoftheirpreferencesandvalues;itwouldbedif“cultforindividualstopreferorvaluesomethingthattheycannottrulycomprehend.Hence,appliedtothedomainofpreferencesandvalues,theHypothesissuggeststhatmoreintelligentindividualsaremorelikelytoacquireandespouseevolutionarilynovelpreferencesandvaluesthatdidnotexistintheancestralenvironmentthanlessintelligentindividuals,butgeneralintelligencehasnoeffectontheacquisitionandespousalofevolutionarilyfamiliarpreferencesandvaluesthatexistedintheancestralenvironmentKanazawa,2010b TherehasbeenemergingevidencefortheHypothesisasanexplanationforindividualpreferencesandvalues.First,moreintelligentchildrenaremorelikelytogrowuptoespouseleft-wingliberalism(Dearyetal.,2008;Kanazawa,2010apossiblybecausegenuineconcernswithgeneticallyunrelatedothersandwillingnesstocontributeprivateresourcesforthewelfareofsuchothers…liberalism…maybeevolutionarilynovel.Eventhoughpaststudiesshowthatwomenaremoreliberalthanmen(LakeandBreglio,1992;ShapiroandMahajan,1986;Wirls,1986),andblacksaremoreliberalthanwhitesKluegelandSmith,1986;Sundquist,1983),theeffectofchildhoodintelligenceonadultliberalismistwiceaslargeastheeffectofsexorrace(Kanazawa,2010a Second,moreintelligentchildrenaremorelikelytogrowuptobeatheists(Kanazawa,2010a),possiblybecausebeliefinhigherpowers,asaconsequenceofoverinferenceofagencybehindotherwisenaturalphenomena,maybepartofevolvedhumannature(Atran,2002;Boyer,2001;Guthrie,1993;HaseltonandNettle,2006;Kirkpatrick,2005),andatheismmaythereforebeevolutionarilynovel.Eventhoughpaststudiesshowthatwomenaremuchmorereligiousthanmen(andHoffmann,1995;MillerandStark,2002),theeffectofchildhoodintelligenceonadultreligiosityistwiceaslargeasthatofsex(Kanazawa,2010a S.Kanazawa/SocialScienceResearch48(2014)157…170 Third,moreintelligentboys(butnotmoreintelligentgirls)aremorelikelytogrowuptovaluesexualexclusivityKanazawa,2010a),possiblybecausehumanswerenaturallypolygynousthroughoutevolutionaryhistory(etal.,1979;HarveyandBennett,1985;KanazawaandNovak,2005;LeuteneggerandKelly,1977;Pickford,1986).Eitherundermonogamyorpolygyny,womenareexpectedtobesexuallyexclusivetoonemate;insharpcontrast,meninpolygynousmarriagearenotexpectedtobesexuallyexclusivetoonematewhereasmeninmonogamousmarriageare.Sotheexpectationofsexualexclusivitymaybeevolutionarilynovelformen,butnotforwomen. Fourth,moreintelligentchildrenaremorelikelytogrowuptobenocturnal,goingtobedandwakinguplater(andPerina,2009),possiblybecausenocturnallifewasrareintheancestralenvironmentwhereourancestorsdidnothavearti“cialsourcesofilluminationuntilthedomesticationof“re.Ethnographiesofcontemporaryhunter-gathererssuggestthatourancestorsmayhavewokenupshortlybeforedawnandgonetosleepshortlyafterdusk.Nightlifemaythereforebeevolutionarilynovel. Fifth,thehumanconsumptionofpsychoactivesubstances,suchasalcohol,tobacco,anddrugs,isevolutionarilynovel,alloriginatinglessthan10,000yearsago.ThustheHypothesiswouldpredictthatmoreintelligentindividualsaremorelikelytoconsumealcohol,tobacco,anddrugs.TheanalysesoftwoprospectivelylongitudinaldatasetswithnationallyrepresentativesamplesintheUKandtheUSsupporttheprediction.Moreintelligentindividualsconsumemorealcoholmorefrequently,smokemoretobacco(butonlyintheUS),andusemoreillegaldrugs(KanazawaandHellberg,2010).Moreintelligentindi-vidualsaremorelikelytoengageinbingedrinkingandgetdrunk(Kanazawa,2012,pp.163…167 Finally,criminalsonaveragehavelowerintelligencethanthegeneralpopulation(WilsonandHerrnstein,1985;HerrnsteinandMurray,1994).ThisisconsistentwiththeHypothesisbecause,whilemuchofwhatwecallinterpersonalcrimetodayisevolutionarilyfamiliar,theinstitutionsthatcontrol,detect,andpunishsuchbehaviorareevolutionarilynovelKanazawa,2009).Murder,assault,robberyandtheftwereprobablyroutinemeansofintrasexualmalecompetitionforresourcesandmatesintheancestralenvironment.Wemayinferthisfromthefactthatbehaviorthatwouldbeclassi“edascriminalifengagedinbyhumansisquitecommonamongotherspecies(Ellis,1998),includingotherprimates(Waal,1989,1992;deWaaletal.,1993).Italsoexplainstheexceptionthatprovestherule,whymoreintelligentindivid-ualsaremorelikelytoconsumeillegaldrugs(KanazawaandHellberg,2010).Unlikemostinterpersonalandpropertycrimes,theconsumptionofsuchsubstancesisevolutionarilynovel.Itisnotlegalitypersethatmatters,butevolutionarynoveltyofthebehavior. Therewasverylittleformalthird-partyenforcementofnormsintheancestralenvironment,onlysecond-partyenforce-ment(retaliationbyvictimsandtheirkinandallies)orinformalthird-partyenforcement(ostracism).ItthereforemakessensefromtheperspectiveoftheHypothesisthatmenwithlowintelligencemaybemorelikelytoresorttoevolutionarilyfamiliarmeansofcompetitionforresources(theftratherthanfull-timeemployment)andmatingopportunities(raperatherthancomputerdating)andnottocomprehendfullytheconsequencesofcriminalbehaviorimposedbyevolutionarilynovelentitiesoflawenforcement. ItisimportanttoemphasizethattheHypothesispredictsthatmoreintelligentindividualsaremorelikelytoacquireandevolutionarilynovelpreferencesandvalues,notnecessarilymoreadaptive(ormoredesirable,orbetterorsmarter)ones.Infact,manyoftheevolutionarilynovelpreferencesandvaluesespousedbymoreintelligentindividualsareoftenmaladaptiveinthecontextofthecurrentenvironment(Kanazawa,2012).Itisveryimportanttokeepthistheoreticalpointinmind,asvoluntarychildlessnessasanevolutionarilynovelvalueisbybiologicalde“nitionhighlymaladaptive. 3.Voluntarychildlessnessasanevolutionarilynovelvalue Ifanyvalueisdeeplyevolutionarilyfamiliar,itisreproductivesuccess.Alllivingorganisms,includinghumans,areevolutionarilydesignedtoreproduce;reproductivesuccessistheultimateendofallbiologicalexistence.Whilehavingchildrenisnottheonlymeanstoincreaseinclusive“tness(representationofonesgenesinthenextgeneration),asitcouldbeachievedbyinvestmentinclosegeneticrelatives,itisnonethelesstheprimarymeansofmaximizingreproductivesuccess.Noneofusaredescendedfromancestorswhoremainedchildless,andwearedisproportionatelydescendedfromindividualswhoachieveddisproportionatereproductivesuccess. Havingchildren,andhavingasmanychildrenasonecanpotentiallyraisetosexualmaturitysothatthechildrenthemselvescanreproduce,isanevolutionarilyfamiliargoal.Incontrast,voluntarychildlessnessisevolutionarilynovel.TheHypothesiswouldthereforepredictthatmoreintelligentindividualsaremorelikelytoexpressanevolutionarilynovelvalueforvoluntarychildlessnessthanlessintelligentindividuals.Giventhatindividualsinadvancedindustrialnationshavereasonablecontrolovertheirfertility,andthustheabilitytoimplementtheirvaluesfororagainstchildrensuccessfully,theHypothesiswouldalsopredictthatmoreintelligentindividualsaremorelikelytoremainchildlessnessthanlessintelligent Therehavebeennumerousstudiesthatestablishthenegativeassociationbetweenintelligenceandfertilityorparent-Maxwell(1969)showsthatchildhoodintelligenceisnegativelyassociatedwiththenumberofchildrenforbothmenandwomeninthe1932and1947ScottishMentalSurveys.LynnandVanCourt(2004)establishanegativeassociationbetweenverbalintelligenceandthenumberofchildreninrepresentativesamplesofAmericansintheGeneralSocialSurveys.However,educationisnotcontrolledineitherofthesestudies,soitisnotclearwhetheritisintelligenceitself S.Kanazawa/SocialScienceResearch48(2014)157…170 159 oritscorrelate,education,thatdecreasesfertility.ItiswellknownthateducationisnegativelyassociatedwithfertilityMusicketal.,2009;Rindfussetal.,1996 IntheiranalysisoftheWisconsinLongitudinalStudy,RetherfordandSewell(1988,1989)showthatchildhoodintelli-gence,measuredin11thgrade,decreasesfertility,measuredat35.Theiranalysisfurthershowsthattheeffectofintelligenceonfertilityislargerforwomenthanformen,andthatitismediatedbyeducation.Netofeducation,childhoodintel-ligencedoesaffectcompletedfertilityintheiranalysis.Similarly,Meisenberg(2010;MeisenbergandKaul2010)that,netofeducation,childhoodintelligencehasnoindependenteffectonthenumberofchildren(exceptforwhitemales)intheNationalLongitudinalSurveyofYouth1979.Rodgersetal.(2008)showthatintelligencenetofeducationhasnoeffectonageat“rstbirthamongfemaleDanishtwinsintheMiddle-AgedDanishTwinsurvey,andNeissetal.(2002)reachasimilarconclusionintheirbehaviorgeneticanalysisoftheNationalLongitudinalSurveyofYouth1979.Reeveetal.showthatbothintelligenceandeducationhaveindependenteffectsonparenthoodintheProjectTALENTdata,buteducationhasamuchstrongereffectthanintelligenceforbothmenandwomen.Finally,Chenetal.(2013)showthatintelligencedoesnotatallmediatetheeffectofeducationonfertilityamongTaiwanesewomen,whilethereisnoassociationbetweenintelligenceoreducationandfertilityamongmen. InthispaperIwillattempttoreplicatethese“ndingsontheassociationbetweenintelligence,educationandfertilitywithalarge,nationallyrepresentative,andprospectivelylongitudinalsamplefromtheUnitedKingdom.Iwillexaminetheeffectofchildhoodintelligenceonindividualspreferenceforparenthoodearlyintheirreproductivecareersandtheircompletedfertility(lifetimeparenthoodorchildlessness)attheendoftheirreproductivecareersintheirlate40s.ItisimportanttonotethattheHypothesisisabouttheeffectofgeneralintelligenceonevolutionarilynovelpreferencesandvalues,suchasvolun-tarychildlessness.Therefore,contraryto“ndingsbyRetherfordandSewell(1988,1989),Meisenberg(2010;MeisenbergandKaul2010),Reeveetal.(2013)Chenetal.(2013),Idoexpecteducationtomediatetheeffectofgeneralintelligenceonlifetimeparenthoodatall. 4.Empiricalanalyses 4.1.Data TheNationalChildDevelopmentStudy(NCDS)isalarge-scaleprospectivelylongitudinalstudywhichhasfollowedaofBritishrespondentssincebirthformorethanhalfacentury.Thestudyincludesbabies(=17,419)borninGreatBritain(England,Wales,andScotland)duringoneweek(03…09March1958).Therespondentsaresubsequentlyreinterviewedin1965(Sweep1at7;=15,496),in1969(Sweep2at11;=18,285),in1974(Sweep3at16;=14,469),in1981(Sweep4at23;=12,537),in1991(Sweep5at33;=11,469),in1999…2000(Sweep6at41…42;=11,419),andin2004…2005(Sweep7at46…47;=9534).TherearemorerespondentsinSweep2thanintheoriginalsample(Sweep0)becausetheSweep2sampleincludeseligiblechildrenwhowereinthecountryin1969butnotwhenSweep0interviewswereconductedin1958.IneachSweep,personalinterviewsandquestionnairesareadministeredtotherespondents,totheirmothers,teachers,anddoctorsduringchildhood,andtotheirspousesandchildreninadulthood. Virtuallyall(97.8%)oftheNCDSrespondentsareCaucasian.Therearesofewrespondentsinotherracialcategoriesthat,ifIcontrolforracewithaseriesofdummiesingeneralizedlinearmodels,itoftenresultsincompleteseparationofdata,andthemaximumlikelihoodestimationbecomesimpossible.IthereforedonotcontrolforrespondentsraceinmyanalysisoftheNCDSdata.Theappendixtablepresentsthemeans,standarddeviations,andafullcorrelationmatrixforallthevariables(describedbelow)thatareusedinmylogisticregressionanalyses,separatelybysex. 4.2.Dependentvariable:preferenceforparenthood At23,NCDSrespondentsindicateiftheyintendtohaveanychildren.Theirresponseiscoded0iftheyintendtoremainchildlessand1iftheyintendtohavechildren.Iftherespondentsalreadyhavechildrenbefore23,theyautomaticallyreceivethevalueof1forpreferenceforparenthood.However,excludingthosewhoalreadyhavechildrenbefore23andlimitingtheanalysestothosewhoarechildlessat23doesnotatallalteranyofthesubstantiveconclusionspresentedbelow.TheNCDSdoesnotaskrespondentsfertilityintentionsafter23.Ianalyzethedichotomousvariablewithbinarylogisticregression. 4.3.Dependentvariable:lifetimeparenthood At47,NCDSrespondentsindicatewhethertheyhavehadanybiologicalchildrenintheirlives.FromthisIcreateabinarydependentvariableforlifetimeparenthood,whichis0iftherespondenthasremainedchildlessforlifeand1iftherespon-denthashadanychildrenintheirlives. FielderandHuber(2007)showthat99.7%ofwomenand96.5%ofmeninarepresentativeSwedishsamplehavecompletedtheirlifetimereproductionby45.ImaythereforereasonablyassumethattheNCDSrespondentshavelargely(ifnotentirely)completedtheirlifetimereproductionby47. S.Kanazawa/SocialScienceResearch48(2014)157…170 4.4.Independentvariable:Generalintelligence TheNCDSrespondentstakemultipleintelligencetestsat7,11,and16.At7,therespondentstakefourcognitivetests(CopyingDesignsTest,Draw-a-ManTest,SouthgateGroupReadingTest,andProblemArithmeticTest).At11,theytake“vecognitivetests(VerbalGeneralAbilityTest,NonverbalGeneralAbilityTest,ReadingComprehensionTest,MathematicalTest,andCopyingDesignsTest).At16,theytaketwocognitivetests(ReadingComprehensionTest,andMathematicsComprehen-sionTest).I“rstperformafactoranalysisateachagetocomputetheirgeneralintelligencescoreforeachage.Allcognitivetestscoresateachageloadonlyononelatentfactor,withreasonablyhighfactorloadings(Age7:CopyingDesignsTest=.671,Draw-a-ManTest=.696,SouthgateGroupReadingTest=.780,andProblemArithmeticTest=.762;Age11:VerbalGeneralAbilityTest=.920,NonverbalGeneralAbilityTest=.885,ReadingComprehensionTest=.864,MathematicalTest=.903,andCopyingDesignsTest=.486;Age16:ReadingComprehensionTest=.909,andMathematicsComprehensionTest=.909).ThelatentgeneralintelligencescoresateachageareconvertedintothestandardIQmetric,withameanof100andastandarddeviationof15.Then,Iperformasecond-orderfactoranalysiswiththeIQscoresatthreedifferentagestocomputetheoverallchildhoodgeneralintelligencescore.ThethreeIQscoresloadonlyononelatestfactorwithveryhighfactorloadings(Age7=.867;Age11=.947;Age16=.919).IusethechildhoodgeneralintelligencescoreinthestandardIQmetricasmymainindependentvariable. 4.5.Controlvariables4.5.1.Educationat23 Inordertoseparatetheeffectofchildhoodintelligencefromthatofeducation,Icontrolfortherespondentseducationatthebeginningofreproductivecareerat23.Duetoahighlycomplexsystemofexaminations,quali“cations,andcerti“cationsintheBritishschoolsystem,educationinNCDSismeasuredqualitativelybya“ve-pointordinalscale:0=noquali“cation;1=CSE2…5/NVQ1;2=Olevels/NVQ2;3=Alevels/NVQ3;4=higherquali“cation/NVQ4;5=degree/NVQ5…6.4.5.2.Earningsat23 Icontrolfortherespondentsearningsat23.EarningsaremeasuredasnetannualpayofcurrentjobinGBP1K.Respondentswhoarenotinthelaborforceareincludedinthesampleandgiventheearningsof0 Only9.1%oftheNCDSrespondentswenttouniversities(Education=5),andeventheywouldhavecompletedtheirformaleducationbyAge21,asuniversitiesareonlyforthreeyearsinEnglandandWales.Veryfewwouldstillbeinfull-timeeducationatAge23. Ichoosetomeasureearningsat23,nearthebeginningoftherespondentsreproductivecareers,inordertoavoidtheproblemofendogeneity.Earningslaterinlife,especiallyforwomen,maybeendogenousto,andpartiallyin”uencedby,therespondentsreproductivebehaviorearlierintheirlives.Asitturnsout,however,itdoesnotmuchmatterwhentheearningsaremeasured,becauseearningsat23issigni“cantlypositivelycorrelatedwithearningsat33(=.052,.001),at42(=.093,.001),andat47(.095,.001).Thesecorrelationsaresurprisinglysmall,butitisbecausetheAge23measureistakenveryearlyintherespondentscareersorbecauseofthelongtimeinterval.ThecorrelationbetweenAge42and47measuresisequallysmall(=.071,.001)andthatbetweenAge33and42measuresisnotevenstatisticallysigni“cant(=.019, 4.5.3.Religiosityat23 Icontrolfortherespondentsreligiosityat23bythefrequencyofchurchattendance:0=noreligion;1=rarelyornever;2=lessthanmonthly;3=monthlyormore;4=weeklyormore.4.5.4.Lifetimenumberofmarriages Becausemarriageisoften(thoughnotalways)apreconditionforparenthood,Icontrolforthelifetimenumberoftimesthattherespondenthasbeenlegallymarriedbefore47.4.5.5.Socialclassoffamilyoforientation Icontrolfortherespondentssocialclassinthefamilyoforientationwiththreeseparatemeasures:Fatherssocialclassatbirth(0=unemployed,dead,retired,nofatherpresent,1=unskilled,2=semiskilled,3=skilled,4=white-collar,5=professional);motherseducation;andfatherseducation(bothmeasuredastheageatwhichtheparentleftfull-timeeducation).Motherseducationandfatherseducationaremeasuredat16. Excludingrespondentsnotinthelaborforceat23andlimitingtheanalysestoonlythosewithsomeearningsat23doesnotaltertheresultsforeitherdesireforparenthoodat23orlifetimeparenthood,excepttoreducethesigni“canceoftheeffectofchildhoodgeneralintelligenceonlifetimeparenthoodforwomenslightly,from=.038to=.059.IfIsubstitutemeanlifetimeearnings(anarithmeticmeanofearningsmeasuredat23,33,42,and47)forearningsat23,noneoftheresultsfordesirefchildren(presentedinTable1below)changes.Forlifetimeparenthood,amongwomen,thenegativeeffectofchildhoodgeneralintelligenceisnolongerstatisticallysigni“cant,whilethemeanlifetimeearningshasasigni“cantlynegativeeffect.However,thisisnodoubtlargelyduetothefactthatwomencandevotemoreoftheirtimeandenergythroughouttheirlivestotheircareersthanwomenwithchildrencan.Amongmen,thepositiveeffectofchildhoodintelligencebecomesstatisticallysigni“cant,whilemeanlifetimeearningsdoesnothaveasigni“canteffect. S.Kanazawa/SocialScienceResearch48(2014)157…170 161 4.5.6.Numberofsiblings Sincethereissomeevidencethatthenumberofchildrenonehasispartlyheritableandgeneticallyin”uenced(etal.,1999Rodgersetal.,2001),Icontrolforthenumberofsiblingstherespondenthas,asanindicatoroftherespondentsparentsfertility.Numberofsiblingsismeasuredat16.AllofthesubstantiveconclusionsfromthelogisticregressionanalysesbelowremainthesameifImeasurethenumberofsiblingsdichotomouslyaswhethertherespondentisanonlychildorhassiblings. 5.Results Table1presentstheresultsofthebinarylogisticregressionanalysisofpreferenceforparenthoodat23.Columns(1)and(4)showthat,consistentwiththepredictionoftheHypothesis,childhoodgeneralintelligenceissigni“cantlynegativelyassociatedwithpreferenceforparenthoodbothamongwomenandmen.Fig.1showsthatwomenwhointendtobecomeparentshaveameanchildhoodIQof99.94whereaswomenwhointendtoremainchildlessforlifehaveameanchildhoodIQof105.50(=7.173,=3544,.001).Similarly,menwhointendtobecomeparentshaveameanchildhoodIQof100.02whereasmenwhointendtoremainchildlessforlifehaveameanchildhoodIQof104.35(=5.310,=3439,.001).Thestandardizedcoef“cientsonoddsinTable1,Columns(1)and(4),suggestthatonestandarddeviationincreaseinchildhoodgeneralintelligence(15IQpoints)decreasestheoddsofpreferenceforparenthoodby35%amongwomenandby27%amongmen.Moreintelligentboysandgirlsareindeedmorelikelytoacquireandespousetheevolutionarilynovelpreferenceofvoluntarychildlessnessinearlyadulthood. Table1,Columns(2)and(5),showthatcontrollingforeducationdoesnotatallattenuatethenegativeassociationbetweenchildhoodgeneralintelligenceanddesireforparenthoodat23,whileeducationitself,netofchildhoodgeneralintelligence,isnotsigni“cantlyassociatedwithdesireforparenthoodat23.However,onceIfurthercontrolforearnings,religiosity,socialclassatbirth,numberofsiblings,motherseducationandfatherseducation,childhoodgeneralintelligence Table1Theeffectofgeneralintelligenceondesireforchildrenatage23,NationalChildDevelopmentStudy(1958…2005).WomenMen(1)(2)(3)(4)(5)(6)ChildhoodIQ(.004)(.006)(.009)(.004)(.006)(.009).655.715.780.735.758.898Educationat23(.057)(.088)(.062)(.093).885.975.916.825Earningsat23(.061)(.045).5741.411Religiosityat23.249(.087)(.106)1.3341.266FamilyoforientationSocialclassatbirth.033(.108)(.114)1.035.869Numberofsiblingsat16.039.077(.061)(.063)1.0741.153Motherseducation.074(.081)(.085)1.109.911Fatherseducation.055.020(.066)(.075).9141.033Intercept5.0994.7464.1484.1534.1702.952(.433)(.577)(.848)(.401)(.559)(.820)(df=1/2/8)52.757Numberofcases354627771458344125051243Mainentriesareunstandardizedregressioncoef“cients.Numbersinparenthesesarestandarderrors.Italicizednumbersarestandardizedeffectsonodds..05..01..001.S.Kanazawa/SocialScienceResearch48(2014)157…170 isonlymarginallysigni“cantlyassociatedwithpreferenceforparenthoodamongwomen(=.057)andnotatallamongmen(=.007,).Itisinterestingtonotethatearningsat23issigni“cantlyassociatedwithpreferenceforparenthoodamongwomenandsigni“cantlyassociatedwithitamongmen.Womenwhoearnmoremoneyat23aremorelikelytowanttoremainchildlesswhereasmenwhoearnmoremoneyat23arelesslikelytowanttodoso.Religiosityhasacomparableeffectonthepreferenceforparenthoodamongbothmenandwomen. Table2presentstheresultsofthebinarylogisticregressionanalysisofthelifetimeparenthoodoftheNCDSrespondents.Column(1)showsthat,consistentwiththepredictionoftheHypothesis,childhoodgeneralintelligenceissigni“cantlynegativelyassociatedwithlifetimeparenthoodamongwomen.Thestandardizedcoef“cientonodds(.747)suggeststhatone-standard-deviationincreaseinchildhoodgeneralintelligencedecreasestheoddsofbecomingaparentsometimeinlifeby25%. Insharpcontrast,andcontrarytothepredictionoftheHypothesis,Column(4)showsthatchildhoodgeneralintelligenceisnotatallassociatedwithlifetimeparenthoodamongmen.Thedifferenceinchildhoodgeneralintelligencebetweenlifetimeparentsandnonparentsisstatisticallysigni“cantamongwomen(105.3vs.101.7,=5.135,.001),butnotamongmen(102.2vs.103.0,).Columns(2)and(5)showthattheseresultsdonotchangeatallforeithermenorwomenwhenIcontrolforeducation.Childhoodgeneralintelligence,netofeducation,isstillstatisticallysigni“cantlynegativelyassociatedamongwomen,butnotamongmen. Column(3)showsthatfurthercontrollingforearnings,religiosityat23,lifetimenumberofmarriages,childhoodsocialclass,numberofsiblings,motherseducation,andfatherseducationdoesnotmuchattenuatetheeffectofchildhoodgeneralintelligenceonlifetimeparenthoodamongwomen.Withallthecontrols,one-standard-deviationincreaseinchildhoodgeneralintelligenceisnowassociatedwith21%decreaseintheoddsoflifetimeparenthood. Acomparisonofstandardizedeffectsonoddsshowsthatthechildhoodgeneralintelligencehasagreatereffectonlifetimeparenthoodthananyothervariablesintheequationexceptforearningsat23.Womenwhoearnmoreat23are Fig.1.MeanchildhoodIQbypreferenceforparenthoodatage23. S.Kanazawa/SocialScienceResearch48(2014)157…170 signi“cantlylesslikelytobecomeparentsintheirlifetimes;onestandard-deviationincreaseinearningsat23decreasestheoddsoflifetimeparenthoodby26%.Moreimportantly,contrarytotheearlier“ndingsbyRetherfordandSewell(1988,1989),Meisenberg(2010;MeisenbergandKaul,2010),Reeveetal.(2013),andChenetal.(2013),educationnetofchildhoodgeneralintelligenceisnotsigni“cantlyassociatedwithlifetimeparenthoodeitherforwomenormen. Additionalanalysesshowthatchildhoodgeneralintelligencealoneentirelymediates(orexplainsaway,aschildhoodgeneralintelligenceiscausallypriortoeducation)theassociationbetweeneducationandparenthood.Whileeducationissigni“cantlynegativelyassociatedwithparenthoodamongwomen(.001)butnotmen(netonlyofchildhoodgeneralintelligence,partialcorrelationbetweeneducationandparenthoodisnolongerstatisticallysigni“cantforeitherwomen()ormen(),asColumns(2)and(5)inTable2show.TheSobeltestshowsasigni“cantstatisticalmediationoftheassociationbetweeneducationandparenthoodbychildhoodgeneralintelligencewomen(.05)butnotamongmen(=1.00, Nowcriticsmightclaimthatitisnotfairtocomparetherelativeeffectsofchildhoodgeneralintelligenceandeducation,whentheformerismeasuredonacontinuousscaleandthelatterismeasuredonasix-pointordinalscale,because,acontinuousvariableisexpectedtoaccountforgreatervarianceinadependentvariablethananordinalvariable.InNCDS,educationismeasuredcontinuouslyonlyonce,inSweep6,astheageatwhichtherespondentleftfull-timecontin-uouseducation.Despitethefactthatthecontinuousmeasureofeducationisnotextremelyhighlycorrelatedwiththeordi-nalmeasure(=.491),allofmysubstantiveconclusionslargelyremainthesameifIusethecontinuousmeasure.Educationisstillsigni“cantlynegativelyassociatedwithparenthoodamongwomen(.001)butnotmen(Netofchildhoodgeneralintelligence,educationisnolongersigni“cantlyassociatedwithparenthoodamongwomen)butitisamongmen(.05).TheSobeltestshowsasigni“cantmediationoftheassociationbetweeneducationandparenthoodbychildhoodgeneralintelligenceamongwomen(.01)butnotamongmen(=1.00, Table2TheeffectofgeneralintelligenceonlifetimeparenthoodNationalChildDevelopmentStudy(1958…2005).WomenMen(1)(2)(3)(4)(5)(6)ChildhoodIQ.004.004.014(.004)(.005)(.008)(.004)(.005)(.008).747.758.7921.0631.0631.240Educationat23.071.008.001.029(.049)(.080)(.050)(.083).9031.012.9991.042Earningsat23(.052)(.040).7391.338Religiosityat23.005.048(.068)(.080)1.0061.053Lifetimenumberofmarriages(.859)(1.244).921.891FamilyoforientationSocialclassatbirth(.100)(.102).985.827Numberofsiblingsat16.028.149(.055)(.056)1.0521.317Motherseducation.005.013(.072)(.079).9931.018Fatherseducation.029(.063)(.067)1.048.951Intercept3.7163.7843.571.910.964(.423)(.492)(.780)(.372)(.426)(.729)(df=1/2/8)26.7391.4591.00825.588Numberofcases26542486117723192159946Mainentriesareunstandardizedregressioncoef“cients.Numbersinparenthesesarestandarderrors.Italicizednumbersarestandardizedeffectsonodds..05..01..001.S.Kanazawa/SocialScienceResearch48(2014)157…170 6.Discussion 6.1.Discrepancywiththepreviousstudies Myresultsabove,whichshowthatchildhoodgeneralintelligencehasadirecteffectonwomenslifetimechildlessnessevennetofeducationandthat,netofchildhoodgeneralintelligence,educationhasnoindependenteffectonwomensparenthood,areinconsistentwithsomepreviousstudiesontheassociationbetweenintelligence,education,andfertility.RetherfordandSewell(1988,1989),Meisenberg(2010;MeisenbergandKaul,2010),Reeveetal.(2013),andChenetal.allshoweitherthatbothintelligenceandeducationhaveindependenteffectsonfertilitynetofeachotherorthateducationentirelymediatestheeffectofintelligenceonfertility.Howcanwereconcilethesediscrepancies? Thereareseveralmethodologicalreasons.First,previousstudiesmeasureintelligencewithonlyoneorafewIQtests,whiletheNCDSmeasuresitwith11IQtestsadministeredatthreedifferentages.Forexample,theWisconsinLongitudinalStudy(RetherfordandSewell,1988,1989)measuresintelligencewithonlyoneIQtest(Henmon…NelsonTest),theNationalLongitudinalSurveyofYouth1979(Meisenberg,2010;MeisenbergandKaul,2010;Neissetal.,2002)usesfoursubtestsoftheArmedForcesQuali“cationTest(arithmetic,wordknowledge,paragraphcomprehension,andmathematicsknowledge),andtheMiddle-AgedDanishTwinsurvey(Rodgersetal.,2008)uses“ve(”uency,digit-spanforward,digit-spanbackward,digitsymbolsubstitution,anddelayedrecall).WhiletheProjectTALENTdatausedbyReeveetal.(2013)containscoreson11IQtests(thesamenumberasinNCDS),intheirfactoranalysisofthescores,Reeveetal.(2013,p.360)forsomereasonrestrictthestandarddeviationoftheIQmetricto10,ratherthantheusual15or16commonlyusedinintelligenceresearch.Thearti“ciallyrestrictedvarianceoftheIQmeasuremayhavecontributedtotheirconclusionthatgeneralintelligencedoesnothaveanindependenteffectonfertilitynetofeducation.WhiletheTaiwaneseversionofWAIS-IIIthatChenetal.(2013)usealsohas11subtests,theirsampleisextremelysmall(=38women,27men),soitisuncleartowhatextenttheir“ndingthateducationandintelligenceareindependentlyassociatedwithfertilitynetofeachotherisgeneralizable.Furtherresearchisclearlynecessarytoinvestigatetheindependenteffectsofintelligenceandeducationonparenthoodandfertility.Thecurrentstudyonlyopens,notcloses,thedebate. 6.2.Sampleattrition OnepotentialproblemwiththeNCDSdataisthesampleattritionrate.Whiletheinitial(Sweep0)sampleincludesaofbabiesbornduringoneweekinMarch1958intheentireGreatBritain,only54.7%oftheinitialrespondentshaveparticipatedinallsevensweepsofNCDS.(Giventhenearlyhalf-centuryspanofNCDS,however,theretentionratemaybeconsideredveryhigh.)Whilesampleattrition,ifsystematic,oftenposesathreattointernalvalidity,itmaynotthrea-tenthesubstantiveconclusionofthispaperbecausemoreintelligentindividualsandparentsaremorelikelytostayinthe NCDSrespondentswhohaveparticipatedinallsevensweepshaveasigni“cantlyhigherchildhoodgeneralintelligencethanthosewhohavedroppedout(102.1vs.96.4,=17.932,.001).Atthesametime,netofchildhoodgeneralintelli-gence,thosewhoareparentsat33or42aresigni“cantlymorelikelytoparticipateinallsevensweepsofNCDSthantheirnonparentalcounterparts(Age33:=.077,SE=.039,.05;Age42:=.173,SE=.048,.001).Netofchildhoodgeneralintelligence,parentalstatusat23isnotsigni“cantlyassociatedwiththelikelihoodofparticipatinginallsevensweeps.018,SE=.037,).Forexample,82.7%ofNCDSrespondentswhoareparentsat42participateinallsevensweeps,comparedto79.4%ofthosewhoarechildless(=3.423,.001). Giventhatthemainpredictiontestedinthecurrentpaperisthatmoreintelligentindividualsaremorelikelytoremainchildless,thefactthatmoreintelligentindividualsandparentsaresimultaneouslymorelikelytostayinthesamplegoesagainstthepredictionandmakesmystatisticaltestmoreconservative.IthereforebelievethatarelativelyhighrateofattritionintheNCDSdatadoesnotseriouslyunderminetheinternalvalidityofthestudy. 6.3.Sexdifferencesintheeffectofintelligenceonlifetimeparenthood TheNCDSdataprovidesupportfortheHypothesisinmyanalysisofexpressedpreferenceforchildrenearlyinthereproductivecareers;moreintelligentmenandwomenaremorelikelytoprefertoremainchildlessforlife.However,inmyanalysisoflifetimeparenthood,theNCDSdataprovidesupportfortheHypothesisonlyforwomen.Moreintelligentwomenaremorelikelytoremainchildlessthanlessintelligentwomenbytheendoftheirreproductivecareers.Incontrast,moreintelligentmenarenotmorelikelytoremainchildlessthanlessintelligentmen.Whataccountsforthesexdifference? My“ndingreplicatesRetherfordandSewells(1988,1989)earlierdiscoverythattheeffectofgeneralintelligenceoncompletedfertilityismuchlargerforwomenthanformen.However,itisnotclearwhygeneralintelligencedoesnothaveanegativeeffectonmensparenthoodpredictedbytheHypothesis,asitdoesforwomen.Amongotherthings,thesexdifferencesintheeffectofintelligenceonfertilityimplythatBritishcouplesarenotveryendogamousonintelligence. theywerehighlyendogamousonintelligenceandifmoreintelligentwomenaremorelikelytoremainchildless,thenmore IowethisinsighttoLenaEdlund. S.Kanazawa/SocialScienceResearch48(2014)157…170 165 intelligentmenmustalsonecessarilybemorelikelytoremainchildless.Myresultsinsteadsuggestthatmoreintelligentwomen,whoaremorelikelytoremainchildless,aremarriedtomoreintelligentmen,andviceversa. Onepossibilityisthatwomen“ndintelligentmenattractiveasmates.Miller(2000a,2000b)hasconsistentlyarguedthatwomenpreferentiallyselectmenwithhigherlevelsofintelligencetomatewith.Giventhatmatingformammalianspeciesislargelyafemalechoice(Trivers,1972),womenspreferenceforintelligentmenasmatescanpotentiallyexplainwhymoreintelligentmenmaybeaslikelytobecomeparentsaslessintelligencemendespitetheirexpressedpreferencetoremainchildlessat23. TheredoesnotappeartobemuchevidenceforthispossibilityintheNCDSdata,however.Netofthesamecontrolvariablesasabove,childhoodgeneralintelligencedoesnotsigni“cantlyincreasethenumberoftimesthatmenhavebeenmarriedortheprobabilitythattheyhaveeverbeenmarried.However,moreintelligentmenaresigni“cantly(.05)morelikelytobemarriedat47.Atanyrate,thelifetimenumberofmarriagesiscontrolledinTable2,Column(6). Anotherpossibilityisthatmoreintelligentwomenmustmakeadif“culttradeoffbetweencareerandfamily,andthusenduphavingfewerchildrenornochildrenatall,whereasmoreintelligentmendonothavetomakesuchachoice.Thiscanpotentiallyexplainwhymoreintelligentwomenaremorelikelytoremainchildless,whereasintelligencedoesnotaffectmensparenthood. Howeverrealsuchtradeoffsthatintelligentwomenfaceintheirlivesare,however,itisnotlikelytoexplainthesexdifferencesintheeffectofchildhoodgeneralintelligenceonparenthoodthatI“ndabove.Inthemultiplebinarylogisticregressionequationabove,botheducationandearningsarecontrolledfor.Ifmoreintelligentwomenaremorelikelytooptforcareersratherthanmotherhood,onewouldexpectthemtohavemoreeducationandhigherearningsthanlessintelligentwomen,andsuchwomenwithgreatereducationandearningsshouldbemorelikelytoremainchildless. TheresultsinTable2showthat,whileearningsat23doessigni“cantlyaffectwomenslifetimeparenthood,educationdoesnot.Netofgeneralintelligenceandalltheothercontrolvariablesincludedintheequation,moreeducatedwomenarelesslikelytobecomeparentsbytheendoftheirreproductivecareers,althoughwomenwithhigherearningsare.Mostimportantly,theeffectofchildhoodgeneralintelligenceonlifetimeparenthoodremainsevenaftercontrollingforeducationandearnings.Moreintelligentwomenaremorelikelytoremainchildlessforlifeevennetofeducationandearnings. Alargenumberofpaststudieshaveshownthatmoreeducatedwomenaremorelikelytoremainchildlessforlife(andTrussell,1984;JacobsonandHeaton,1991;Kiernan,1989;MosherandBachrach,1982;PostonandKramer,1986;seeBloomandPebley,1982forreview).However,ofthesestudiesmeasureandcontrolforthewomensgeneralintelli-gence.TheresultspresentedinTable2,Column(3),suggeststhateducationinthesestudiesmayhavebeenconfoundedwithintelligence,andthatitmaybewomensgeneralintelligence,notnecessarilytheireducation,whichin”uencestheirlikelihoodofremainingpermanentlychildless. 6.4.Potentialmacrolevelconsequences Generalintelligenceisknowntobehighlyheritable(Jensen,1998,butseeTaylor,2010),andthegenesthatin”uencegeneralintelligencearethoughttobelocatedontheXchromosomes(Lehrke,1972,1997;Turner,1996a,1996b).Itmeansthatboysinherittheirgeneralintelligencefromtheirmothersonly,whilegirlsinherittheirgeneralintelligencefromtheirmothersandfromtheirfathers.Sowomenin”uencethegeneralintelligenceofthefuturegenerationsmorethanmendo. moreintelligentwomenaremorelikelytoremainchildless,asmyanalysesoftheNCDSdataaboveseemtosuggest,thenonepotentialconsequenceisthattheaveragelevelofgeneralintelligencemaydeclineovertime. Throughoutthe20thcentury,theaveragelevelofintelligenceinmostwesternindustrialnationssteadilyincreased,inaphenomenoninitiallyknownastheFlynnEffect(Flynn,1984,1987)butmorerecentlyknownastheLynn-FlynnEffectBeaujeanandGuiling,2006;Lynn,1982;Rushton,1997;Voracek,2006).AlthoughthereisnoconsensusonwhatcausedtheLynn-FlynnEffect(Neisser,1998),onelikelycandidateistheincreasinglevelsofinfantandchildhoodnutritionandhealth(Lynn,1990,1998).Thesefactorslikelymorethancompensatedforthedysgenicfertility…wherelessintelligentpar-entsaremorelikelytohavechildrenasdocumentedinthispaper…throughoutthe20thcentury,andtheaveragelevelofintelligencehasincreasedinmostadvancedindustrialnations. ImprovedhealthandnutritionasapotentialcauseoftheLynn-FlynnEffect,however,wouldpredictthatthesecularriseingeneralintelligencewouldhaltinadvancedindustrialnations:Theoptimallevelofnutritionhaslongpassedandnowobesityanddiabeteshavebecomeseriousproblemsinsuchsocieties.Ifimprovedhealthandnutritionarechie”yresponsi-bleforthesecularincreaseinintelligence,andifthesefactorsnolongercontributetotheincrease,thenthenegativeeffectofthedysgenicfertilityshouldleadtoalevelofaverageintelligenceinadvancedindustrialnations(Lynnand ThankPaulaEnglandandDianeJ.Reyniersforindependentlysuggestingthispossibility.IowethisinsighttoChristopherBadcock.Ifmothershavegreaterin”uenceonchildrensintelligencethanfathersdo,itimpliesthatthemother…chcorrelationinIQisgreaterthanthefather-childcorrelation.Thereissomesuggestionthatthismightindeedbethecase.Jesteretal.s(2009)equationmodelshowsthattheeffectofmothersIQonchildsIQ(=.54)isslightlystrongerthantheeffectoffathersIQ(=.52).NeissandRowe(2000)showthatthecorrelationbetweenmotherseducation(asaproxyforIQ)andchildsIQis=.41,whereasthecorrelationbetweenfatherseducationandchildsIQis=.36.AtAge33,theNCDSselectsasubsetofrespondentsandadministers2…5IQtests(dependingonage)toalloftheirchildren.IcomputethecorrelationbetweentherespondentschildhoodIQandtheeldestchildsIQ.Themother…childcorrelationisstatisticallysigni“cantlypositive=.190,.01),whilethefather-childcorrelationisonlymarginallyso(=.127,=.088).Ithankoneanonymousreviewerforpointingoutthisimplicationofthegreaterin”uenceofmothersthanfathersintelligenceonchildrensintelligence. S.Kanazawa/SocialScienceResearch48(2014)157…170 Harvey,2008).ThereisinfactstrongevidencethattheLynn-FlynnEffecthasendedattheendofthe20thcentury,andtheaveragelevelofintelligencehasbeguntodeclineatthebeginningofthe21stcenturyinsuchadvancedindustrialnationsasAustralia(Cottonetal.,2005),Denmark(TeasdaleandOwen2005),Norway(Sundetetal.,2004),andtheUnitedKingdomShayeretal.,2007;ShayerandGinsburg2009 PrestonandCampbell(1993)employmathematicalmodelsandsimulationstoshowthat,evenwhenlow-IQindividualshavehigherfertilitythanhigh-IQindividuals,themeanIQofthepopulationdoesnotcontinuetodeclineforeverandinsteadreachesasteady-stateequilibriuminmanycases.Thisisbecausethereissuf“cientintergenerationalmobilitybetweenIQcategories;childrendonothaveexactlythesameIQastheirparents.Intheirmodels,PrestonandCampbellusedatafromReedandReed(1965)toestimatetheassociationbetweenparentsandchildrensintelligence.ReedandReeds(1965)containchildrensIQandtheirmothersandfathersIQs.Heritabilityofintelligenceisknowntobemuchlowerinchildhood(around.4)thaninadulthood(around.8)(Boomsmaetal.,2008 EvenwiththeReedandReeddata,withtheirlowheritabilities,however,thereisverylittleintergenerationalmobilitybetweenIQclasses.Tables2and3inPrestonandCampbell(1993,p.1002)showthatchildrenofparentsbothofwhomhaveIQsabove126haveameanIQof123.5withavarianceof13.0(SD=3.61).Giventhesedata,andassumingthatthedistri-butionoftheirIQisnormal,theprobabilitythatchildrenoftheseparentsendupintheIQcategories74orlower,75…84,85…95,and96…105iseachzerotothesixthdecimalpoint.EvenintheReedandReeddata,whichunderestimateheritabilityofadultIQ,itisimpossibleforchildrenbothofwhoseparentshaveIQsabove126tohaveIQsbelow105.Similarly,intheReedandReeddata,itisimpossibleforchildrenbothofwhoseparentshaveIQsbelow74tohaveIQsabove Nevertheless,PrestonandCampbellsresultslargelyshowthatthemeanIQofthepopulationisheavilyin”uencedbydifferentialfertilityandmovesclosertotheIQofthehigh-fertilitygroup.Forexample,underrandommating(whereindi-vidualsmaterandomlywithrespecttotheIQoftheirmates),whenheritabilityishigh(or,intheirlanguage,whentheleak-agebetweenIQcategoriesislow)andfertilityandIQarepositivelyassociated(suchthathigher-IQindividualshavehigherfertilitythanlower-IQindividuals),then,atequilibrium,41.7%ofthepopulationwillhaveIQsabove126,another40.9%willhaveIQsbetween116and125,andwillhaveIQsbelow94(PrestonandCampbell,1993,p.1012,Table4).Underendogamousmating(whenindividualsmateandreproduceonlywithmatesintheirownIQcategory),ifIQandfertilityarepositivelyassociated,themeanIQofthepopulationatequilibriumwillbeabout113,whereasifIQandfertilityareneg-ativelyassociated,themeanIQofthepopulationatequilibriumwillbeabout95(PrestonandCampbell,1993,p.1014Figure3). Thedysgenicfertilitydocumentedinthispaper,coupledwiththeendoftheLynn-FlynnEffect,maypotentiallyleadtoagradualdeclineoftheaverageintelligenceofthepopulationofadvancedindustrialnationsinthe21stcentury. 7.Conclusion TheSavanna-IQInteractionHypothesis,derivedfromthelogicalconjunctionoftheSavannaPrincipleandthetheoryoftheevolutionofgeneralintelligence,predictsthatmoreintelligentindividualsaremorelikelythanlessintelligentindivid-ualstoacquireandespouseevolutionarilynovelpreferencesandvaluesthatourancestorsdidnotpossessthroughoutevolutionaryhistory,whilegeneralintelligencedoesnotaffecttheacquisitionandespousalofevolutionarilyfamiliarpreferencesandvalues.Ifanysinglevaluegoesagainsttheevolutionarydesignofbiologicalorganisms…includinghumans…morethananyother,itisvoluntarychildlessness.TheHypothesiswouldthereforepredictthatmoreintelligentindividualsaremorelikelytoprefertoremainchildlessandinfactdoremainchildlessthanlessintelligentindividuals. TheanalysesoftheNationalChildDevelopmentStudyintheUnitedKingdom,whichhasfollowedallbabiesborninGreatBritaininoneweekinMarch1958formorethanhalfacentury,showthatmoreintelligentboysandgirlsaremorelikelytoacquireandespousetheevolutionarilynovelvalueoflifelongchildlessnessinearlyadulthood.However,onlymoreintelli-gentwomenaremorelikelytoremainchildlessbytheendoftheirreproductivecareers;childhoodgeneralintelligencedoesnotappeartoaffectmenslifetimeparenthood.Sincewomenmayhavegreaterimpactonthegeneralintelligenceofthefuturegenerations,thedysgenicfertilityamongwomen…wherelessintelligentwomenaremorelikelytohavechildrenthanmoreintelligentwomen…mayleadtotheprediction,supportedbyrecentstudies,thattheaveragelevelofintelligencemaysteadilydeclinethroughoutthe21stcentury. Acknowledgments IthankPaulaEnglandandDianeJ.Reyniersfortheircommentsonearlierdrafts.IdedicatethispapertothememoryofDebraFriedman.MyintellectualinterestinparenthoodbeganwithmycollaborationwithDebrawhileingraduateschool.Sadly,shedidnotlivelongenoughtoseethislateststatementfromherrebelliousintellectualchildbeforeherlifewastragicallycutshort. AppendixA Descriptivestatistics S.Kanazawa/SocialScienceResearch48(2014)157…170 167 (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(1)Desire(2)Parenthood.164(3)IQ.034.322(4)Education.003.638001.307(5)Earnings.065006.142(6)Religiosity.066.001.064.003.084(7)Numberof.010.001009.004(8)Socialclass.010.300(9)Siblings.050(10)Mothers.015.292.007.056.010.281(11)Fathers.012.315.001.087.004.363Mean.897.832100.0182.1481.8421.129.0112.8032.4423.9523.926SD.304.37414.5911.4341.7171.156.1101.0321.8191.3981.627Mean.887.79099.9832.3083.366.798.0132.8132.4773.8833.883SD.316.40815.3861.4162.3111.068.1121.0221.8501.3541.617Descriptivestatisticsforwomenareabovethediagonalandinitalics.DescriptivestatisticsformenarebelowthediagonalandinRoman..05..01..001. References Alexander,RichardD.,Hoogland,JohnL.,Howard,RichardD.,Noonan,KatharineM.,Sherman,PaulW.,1979.Sexualdimorphismsandbreedingsystemspinnipeds,ungulates,primatesandhumans.In:Chagnon,ApoleonA.,Irons,William(Eds.),EvolutionaryBiologyandHumanSocialBehaviorAnAnthropologicalPerspective.DuxburyPress,NorthScituate,pp.402…435 Atran,Scott,2002.GodsWeTrust:TheEvolutionaryLandscapeofReligion.OxfordUniversityPress,Oxford Barber,JenniferS.,2001.Ideationalin”uencesonthetransitiontoparenthood:attitudestowardchildbearingandcompetingalternatives.Soc.PQuart.64,101…127 Beauchamp,Guy.,1994.Thefunctionalanalysisofhumanfertilitydecisions.Ethol.Sociobiol.15,31…53 Beaujean,A.Alexander,Guiling,Shawn.F.,2006.TheLynn-Flynneffectandschoolpsychology:acallforresearch.Sch.Psychol.60,17…20 Becker,GaryS.,1960.AnEconomicAnalysisofFertility.In:DemographicandEconomicChangeinDevelopedCountries,editedbytheUniversities…NationalBureauCommitteeforEconomicResearch.Princeton:PrincetonUniversityPress.pp.209…31. Becker,GaryS.,1981.ATreatiseontheFamily.HarvardUniversityPress,Cambridge Blake,Judith,1968.AreBabiesConsumerDurables?Popul.Stud.22,5…25 Bloom,DavidE.,Pebley,AnneR.,1982.Voluntarychildlessness:areviewoftheevidenceandimplications.Popul.Res.PolicyRev.1,203…224 Bloom,DavidE.,Trussell,James,1984.WhatarethedeterminantsofdelayedchildbearingandpermanentchildlessnessintheUnitedStates?Demogra21,591…611 Boomsma,DorretI.,vanBeijsterveld,ToosC.E.M.,Beem,A.Leo,Hoekstra,RosaA.,Polderman,TincaJ.C.,Bartels,Meike,2008.Intelligenceandbirthorderinboysandgirls.Intelligence36,630…634 Boyer,Pascal,2001.ReligionExplained:TheEvolutionaryOriginsofReligiousThought.Basic,NewYork Butz,WilliamP.,Ward,MichaelP.,1979.TheemergenceofcountercyclicalUSfertility.Am.Econ.Rev.69,318…328 Chen,Hsin-Yi,Chen,Yung-Hua,Liao,Yung-Kun,Chen,Hsin-Ping,2013.Relationshipoffertilitywithintelligenceandeducationintaiwan:abriefreport.J.Biosoc.Sci.45,567…571 Cotton,SueM.,Kiely,PatriciaM.,Crewther,DavidP.,Thomson,Brenda,Laycock,Robin,Crewther,SheilaG.,2005.AnormativeandreliabilitystudyfortheravenscolouredprogressivematricesforprimaryschoolagedchildrenfromVictoria,Australia.PersonalityIndivid.Differ.39,647…659 Crawford,CharlesB.,1993.Thefutureofsociobiology:countingbabiesorproximatemechanisms?TrendsEcol.Evol.8,183…186 Davis,Kingsley,Bernstam,MikhailS.,Ricardo-Campbell,Rita(Eds.),1987.Below-ReplacementFertilityinIndustrialSocieties:Causes,Consequences,Policies.CambridgeUniversityPress,Cambridge. deWaal,FransB.M.,1989.Foodsharingandreciprocalobligationsamongchimpanzees.J.Hum.Evol.18,433…459 deWaal,FransB.M.,1992.Appeasement,celebration,andfoodsharinginthetwopanspecies.In:William,C.(Ed.),TopicsinPrimatology:HumanOrigUniversityofTokyoPress,Tokyo,pp.37…50 deWaal,FransB.M.,Luttrell,LesleighM.,EloiseCan“eld,M.,1993.Preliminarydataonvoluntaryfoodsharinginbrowncapuchinmonkeys.Am.J.Pri29,73…78 Deary,IanJ.,DavidBatty,G.,Gale,CatharineR.,2008.Brightchildrenbecomeenlightenedadults.Psychol.Sci.19,1…6 Easterlin,RichardA.,Pollak,RobertA.,Pollak,MichaelL.,1980.Towardamoregeneraleconomicmodeloffertilitydetermination:endogenouspreandnaturalfertility.In:Easterlin,RichardA.(Ed.),PopulationandEconomicChangeinDevelopingCountries.UniversityofChicagoPress,Chicago,pp. Ellis,Lee,1998.Neodarwiniantheoriesofviolentcriminalityandantisocialbehavior:photographicevidencefromnonhumananimalsandareviewoftheliterature.Aggress.ViolentBehav.3,61…110 Fielder,Martin,Huber,Susanne,2007.TheeffectsofsexandchildlessnessontheassociationbetweenstatusandreproductiveoutputinmodernsociEvolut.Hum.Behav.28,392…398 Flynn,JamesR.,1984.ThemeanIQofAmericans:massivegains1932to1978.Psychol.Bull.95,29…51 Flynn,JamesR.,1987.MassiveIQgainsin14nations:whatIQtestsreallymeasure.Psychol.Bull.101,171…191 Friedman,Debra,Hechter,Michael,Kanazawa,Satoshi,1994.Atheoryofthevalueofchildren.Demography31,375…401 Guthrie,StewartElliott,1993.FacesintheClouds:ANewTheoryofReligion.OxfordUniversityPress,NewYork S.Kanazawa/SocialScienceResearch48(2014)157…170 Harvey,P.H.,Bennett,P.M.,1985.Sexualdimorphismandreproductivestrategies.In:Ghesquiere,J.,Martin,R.D.,Newcombe,F.(Eds.),HumanSexDimorphism.TaylorandFrancis,London,pp.43…59 Haselton,MartieG.,Nettle,Daniel,2006.Theparanoidoptimist:anintegrativeevolutionarymodelofcognitivebiases.Person.Soc.Psychol.Rev.10,47…66 Hayford,SarahR.,2009.Theevolutionoffertilityexpectationsoverthelifecourse.Demography46,765…783 Hayford,SarahR.,Morgan,S.Philip,2008.ReligiosityandfertilityintheUnitedStates:theroleoffertilityintentions.Soc.Forces86,1163…11 Heaton,TimB.,1986.Howdoesreligionin”uencefertility?Thecaseofmormons.J.Sci.StudyRelig.25,248…258 Herrnstein,RichardJ.,Murray,Charles,1994.TheBellCurve:IntelligenceandClassStructureinAmericanLife.FreePress,NewYork Jacobson,CardellK.,Heaton,TimB.,1991.VoluntarychildlessnessamongAmericanmenandwomeninthelate1980s.Soc.Biol.38,79…93 Jensen,ArthurR.,1998.ThegFactor:TheScienceofMentalAbility.Praeger,Westport Jester,JenniferM.,Nigg,JoelT.,Puttler,LeonI.,Long,JeffreyC.,Fitzgerald,HiramE.,Zucker,RobertA.,2009.Intergenerationaltransmissionofneuropsychologicalexecutivefunctioning.BrainCogn.70,145…153 Kanazawa,Satoshi,2001.Degustibusdisputandum.Soc.Forces79,1131…1163 Kanazawa,Satoshi,2002.Bowlingwithourimaginaryfriends.Evolut.Hum.Behav.23,167…171 Kanazawa,Satoshi,2004a.TheSavannaPrinciple.Manage.Decis.Econ.25,41…54 Kanazawa,Satoshi,2004b.Generalintelligenceasadomain-speci“cadaptation.Psychol.Rev.111,512…523 Kanazawa,Satoshi,2008.Temperatureandevolutionarynoveltyasforcesbehindtheevolutionofgeneralintelligence.Intelligence36,99…108 Kanazawa,Satoshi,2009.Evolutionarypsychologyandcrime.In:Anthony,Walsh,Beaver,KevinM.(Eds.),BiosocialCriminology:NewDirectionsinandResearch.Routledge,NewYork,pp.90…110 Kanazawa,Satoshi.,2010a.Whyliberalsandatheistsaremoreintelligent.Soc.Psychol.Quart.73,33…57 Kanazawa,Satoshi,2010b.Evolutionarypsychologyandgeneralintelligence.Am.Psychol.65,279…289 Kanazawa,Satoshi.,2012.TheIntelligenceParadox:WhytheIntelligentChoiceIsntAlwaystheSmartOne.Wiley,NewYork Kanazawa,Satoshi,Hellberg,JosephineE.E.U.,2010.Intelligenceandsubstanceuse.Rev.Gen.Psychol.14,382…396 Kanazawa,Satoshi,Novak,DeannaL.,2005.Humansexualdimorphisminsizemaybetriggeredbyenvironmentalcues.J.Biosoc.Sci.37,657…665 Kanazawa,Satoshi,Perina,Kaja,2009.Whynightowlsaremoreintelligent.PersonalityIndivid.Differ.47,685…690 Kaplan,Hillard,Lancaster,JaneB.,TroyTucker,W.,Anderson,K.G.,2002.Evolutionaryapproachtobelowreplacementfertility.Am.J.Hum.Biol.14,233… Kaptijn,Ralf,Thomese,Fleur,vanTilburg,TheoG.,Liebroer,AartC.,Deeg,DorlyJ.H.,2010.Lowfertilityincontemporaryhumansandthematevalueoftheirchildren:sex-speci“ceffectsonsocialstatusindicators.Evolut.Hum.Behav.31,59…68 Kiernan,KathleenE.,1989.Whoremainschildless?J.Biosoc.Sci.21,387…398 Kirkpatrick,LeeA.,2005.Attachment,Evolution,andthePsychologyofReligion.Guilford,NewYork Kluegel,JamesR.,Smith,EliotR.,1986.BeliefsAboutInequality:AmericansViewofWhatIsandWhatOughttoBe.Aldine,NewYork Kohler,Hans-Peter,Rodgers,JosephL.,Christensen,Kaare,1999.Isfertilitybehaviorinourgenes?FindingsfromaDanishtwinstudy.Popul.Develop.Rev.25,253…288 Lake,CelindaC.,VincentJ.Breglio.,1992.Differentvoices,differentviews:thepoliticsofgender.In:Paula,Ries.,AnneJ.Stone,(Eds.),TheAmericanWoman,1992…93:AStatusReport,NewYork:Norton,pp.178…201. Lehrke,Robert,1972.AtheoryofX-linkageofmajorintellectualtraits.Am.J.Ment.De“c.76,611…619 Lehrke,Robert,1997.SexLinkageofIntelligence.TheX-Factor.Praeger,Westport Lesthaeghe,Ron,1983.AcenturyofdemographicandculturalchangeinWesternEurope:Anexplorationofunderlyingdimensions.Popul.Develop.Rev.9, Leutenegger,Walter,Kelly,JamesT.,1977.Relationshipofsexualdimorphismincaninesizeandbodysizetosocial,behavioral,andecologicalcorrelatesinanthropoidprimates.Primates18,117…136 Lynn,Richard,1982.IQinJapanandtheUnitedStatesshowsagrowingdisparity.Nature297,222…223 Lynn,Richard,1990.Theroleofnutritioninthesecularincreasesofintelligence.PersonalityIndivid.Differ.11,273…286 Lynn,Richard,1998.Insupportofthenutritiontheory.In:Neisser,Ulric(Ed.),TheRisingCurve:Long-TermGainsinIQandRelatedMeasures.AmeriPsychologicalAssociation,Washington,DC,pp.207…215 Lynn,Richard,Harvey,John,2008.ThedeclineoftheworldsIQ.Intelligence36,112…120 Lynn,Richard,VanCourt,Marian,2004.NewevidenceofdysgenicfertilityforintelligenceintheUnitedStates.Intelligence32,193…201 Maxwell,James,1969.Intelligence,educationandfertility:acomparisonbetweenthe1932and1947Scottishsurveys.J.Biosoc.Sci.1,247…271 Meisenberg,Gerhard,2010.Thereproductionofintelligence.Intelligence38,220…230 Meisenberg,Gerhard,Kaul,Anubhav,2010.Effectsofsex,race,ethnicityandmaritalstatusontherelationshipbetweenintelligenceandfertility.MankindQuart.50,151…187 Miller,GeoffreyF.,2000a.TheMatingMind:HowSexualChoiceShapedtheEvolutionofHumanNature.Doubleday,NewYork Miller,Geoffrey,2000b.SexualSelectionforIndicatorsofIntelligence.In:Bock,GregoryR.,Goode,JamieA.,Webb,Kate(Eds.),TheNatureofIntWiley,NewYork,pp.260…275 Miller,AlanS.,Hoffmann,JohnP.,1995.Riskandreligion:anexplanationofgenderdifferencesinreligiosity.J.Sci.StudyRelig.34,63…75 Miller,AlanS.,Stark,Rodney,2002.Genderandreligiousness:cansocializationexplanationsbesaved?Am.J.Sociol.107,1399…1423 Mosher,WilliamD.,Bachrach,ChristineA.,1982.ChildlessnessintheUnitedStates:estimatesfromthenationalsurveyoffamilygrowth.J.Fam.Issues3, Mosher,WilliamD.,Williams,LindaB.,Johnson,DavidP.,1992.ReligionandfertilityintheUnitedStates:newpatterns.Demography29,199…214 Musick,Kelly,England,Paula,Edgington,Sarah,Kangas,Nicole,2009.Educationdifferencesinintendedandunintendedfertility.Soc.Forces88,543…572 Neiss,Michelle,Rowe,DavidC.,2000.ParentaleducationandchildsverbalIQinadoptiveandbiologicalfamiliesinthenationallongitudinalstudyofadolescenthealth.Behav.Genet.30,487…495 Neiss,Michelle,Rowe,DavidC.,Rodgers,JosephL.,2002.DoeseducationmediatetherelationshipbetweenIQandageof“rstbirth?Abehaviouralgenanalysis.J.Biosoc.Sci.34,259…275 Neisser,Ulric.(Ed.),1998.TheRisingCurve:Long-TermGainsinIQandRelatedMeasures.Washington,DC:AmericanPsychologicalAssociation. Pearce,LisaD.,2002.Thein”uenceofearlylifecoursereligiousexposureonyoungadultsdispositionstowardchildbearing.J.Sci.StudyRelig.41,325…340 Pickford,Martin.1986.Ontheoriginsofbodysizedimorphisminprimates.In:Martin,Pickford.,Brunetto,Chiarelli.(Eds.),SexualDimorphisminLivingandFossilPrimates.Florence.IlSedicesimo,pp.77…91. PostonJr.,DudleyL.,Kramer,KathrynB.,1986.Patternsofchildlessnessamongcatholicandnon-catholicwomenintheU.S.:alog-linearanalysis.Inq.56,507…522 Preston,SamuelH.,1986.Changingvaluesandfallingbirthrates.Populat.Develop.Rev.12,s176…s195 Preston,SamuelH.,Campbell,Cameron,1993.Differentialfertilityandthedistributionoftraits:thecaseofIQ.Am.J.Sociol.98,997…1019 Reed,ElizabethW.,Reed,SheldonC.,1965.MentalRetardation:AFamilyStudy.W.B.Saunders,Philadelphia Reeve,CharlieL.,Lyerly,JordanE.,Peach,Hannah,2013.Adolescentintelligenceandsocio-economicwealthindependentlypredictadultmaritalreproductivebehavior.Intelligence41,358…365 Retherford,RobertD.,Sewell,WilliamH.,1988.Intelligenceandfamilysizereconsidered.Soc.Biol.35,1…40 Retherford,RobertD.,Sewell,WilliamH.,1989.Howintelligenceaffectsfertility.Intelligence13,169…185 Rindfuss,RonaldR.,PhilipMorgan,S.,Offutt,Kate,1996.EducationandthechangingagepatternofAmericanfertility:1963…1989.Demography33, S.Kanazawa/SocialScienceResearch48(2014)157…170 Rodgers,JosephLee,Hughes,Kimberly,Kohler,Hans-Peter,Christensen,Kaare,Doughty,Debby,Rowe,DavidC.,Miller,WarrenB.,2001.Geneticinhelpsexplainvariationinhumanfertility:evidencefromrecentbehavioralandmoleculargeneticstudies.Curr.Dir.Psychol.Sci.10,184…188 Rodgers,JosephLee,Kohler,Hans-Peter,McGue,Matt,Behrman,JereR.,Petersen,Inge,Bingley,Paul,Christensen,Kaare,2008.Educationandcogabilityasdirect,mediating,orspuriousin”uencesonfemaleageat“rstbirth:behaviorgeneticmodels“ttoDanishtwindata.Am.J.Sociol.114,S20 Rushton,J.Philippe,1997.Reviewessay.Society34,78…82 Ryder,NormanB.,1979.ThefutureofAmericanfertility.Soc.Probl.26,359…370 Schoen,Robert,Kim,YoungJ.,Nathanson,ConstanceA.,Fields,Jason.,Astone,Nan.Marie.,1997.WhydoAmericanswantchildren?Populat.Develop.Rev.23,333…358 Shapiro,RobertY.,Mahajan,Harpreet,1986.Genderdifferencesinpolicypreferences:asummaryoftrendsfromthe1960stothe1980s.PublicOpin.Q50,42…61 Shayer,Michael,Ginsburg,Denise,2009.Thirtyyearson…alargeanti-Flynneffect?(II):13-and14-year-olds.Piagetiantestsofformaloperationsnorms1976…2006/7.Br.J.Educ.Psychol.79,409…418 Shayer,Michael,Ginsburg,Denise,Coe,Robert,2007.Thirtyyearson…alargeanti-”ynneffect?ThePiagetiantestvolume&heavinessnorms1975…20Br.J.Educ.Psychol.77,25…41 Sundet,JonMartin,Barlaug,DagG.,Torjussen,ToreM.,2004.TheendoftheFlynneffect?Astudyofseculartrendsinmeanintelligencetestscoresofnorwegianconscriptsduringhalfacentury.Intelligence32,349…362 Sundquist,JamesL.,1983.DynamicsofthePartySystem,Reviseded.BrookingsInstitution,Washington,DC Symons,Donald,1990.Adaptivenessandadaptation.Ethol.Sociobiol.11,427…444 Taylor,Peter,2010.Threepuzzlesandeightgaps:whatheritabilitystudiesandcriticalcommentarieshavenotpaidenoughattentionto.Biol.Philos.25,1… Teasdale,ThomasW.,Owen,DavidR.,2005.Along-termriseandrecentdeclineinintelligencetestperformance.theFlynneffectinreverse.PersonaIndivid.Differ.39,837…843 Tooby,John,Cosmides,Leda,1990.Thepastexplainsthepresent:emotionaladaptationsandthestructureofancestralenvironments.Ethol.Sociobiol.11, Trivers,RobertL.1972.Parentalinvestmentandsexualselection.In:Bernard,Campbell,(Ed.),SexualSelectionandtheDescentofMan1871…1971.Chicago:Aldine,pp.136…179. Turner,Gillian,1996a.IntelligenceandtheXchromosome.Lancet347,1814…1815 Turner,Gillian,1996b.FindinggenesontheXchromosomebywhichmayhavebecome.Am.J.Hum.Genet.58,1109…1110 Voracek,M.,2006.Phlogiston,”uidintelligence,andtheLynn-Flynneffect.Behav.BrainSci.29,142…145 Westoff,CharlesF.,1986.Perspectiveonnuptialityandfertility.Populat.Develop.Rev.12,s155…s170 Westoff,CharlesF.,Bumpass,LarryL.,1973.TherevolutioninbirthcontrolpracticesofU.S.RomanCatholics.Science179,41…44 Williams,Lindy,Abma,Joyce,Piccinino,LindaJ.,1999.Thecorrespondencebetweenintentiontoavoidchildbearingandsubsequentfertility:apranalysis.Fam.Plann.Perspect.31,220…227 Wilson,JamesQ.,Herrnstein,RichardJ.,1985.CrimeandHumanNature:TheDe“nitiveStudyoftheCausesofCrime.Touchstone,NewYork Wirls,Daniel,1986.Reinterpretingthegendergap.PublicOpin.Quart.50,316…330 S.Kanazawa/SocialScienceResearch48(2014)157…170