International Society for Intelligence Research (ISIR)...International Society for Intelligence Research (ISIR) 20th Annual Conference July 11–13, 2019 Minneapolis, MN (United States)

InternationalSocietyforIntelligenceResearch(ISIR)

20thAnnualConference

July11–13,2019Minneapolis,MN(UnitedStates)RenaissanceMinneapolisHotel,TheDepot

www.isironline.org

ISIR2019Schedule

2

TABLEOFCONTENTS

PAGE3 ConferenceInformation4 Dailyschedule7 InvitedTalksandSymposia8 Abstracts39 Posters58 Acknowledgements59 Usefulinformation64 Blankpagesforyourownuse

ISIR2019Schedule

3

ISIRPRESIDENTANDBOARD2019President:WilliamRevelle

PresidentElect:RexJung

PastPresident:TimothyBates

Secretary/Treasurer:ThomasCoyle

Board:RosalindArden,JamesLee,JacobPietschnig,HeinerRindermann

LOCALHOST

ProfessorJamesJ.Lee DepartmentofPsychology [email protected]

CONFERENCEVENUE

RenaissanceMinneapolisHotel,TheDepot225,3rdAvenueSouth,MinneapolisMN55401UnitedStates

ISIR2019Schedule

4

Thursday11thJuly2019

8:00-9:00 Registration(alsoopennightbefore)

9:00–9:10 President’sWelcomeAddress: Prof.WilliamRevelle,NorthwesternUniversity9:10-10:10 LifetimeAchievementAward: Prof.DavidLubinski,&DeanCamillaBenbow,VanderbiltUniversity

10:10-10:30 Cognitiveenhancementandnetworkeffects DrJonathanAnomaly,OxfordUniversityandUniversityofSanDiego

10:30–11:00 MORNINGCOFFEE

11:00–13:00 President’sSymposiumII:ScienceandEthicsofGeneticEngineering. GregoryCochran,UniversityofUtah StephenHsu,MichiganStateUniversity RazibKhan,Insitome BruceLahn,UniversityofChicago NevenSesardić

13:00–14:15 LUNCH

14:15–14:30 Workingmemory,measuredwitha3Dvideogame,correlateswithstandardworkingmemory tasks,fluidreasoningandspatialability,butnotwithverbalability MsAliciaDiaz,EdinburghUniversity

14:30–14:45 AnegativeFlynneffectinrecentcognitiveabilityscores DavidSchroeder,JohnsonO'ConnorResearchFoundation

14:45–15:00 LookingforaFlynneffect:ExaminingshiftsincognitiveabilitywithintheSAPAproject Mrs.ElizabethDworak,NorthwesternUniversity15:00-15:15 SmallerFlynneffectsforcrystallizedintelligencemayberootedinitemobsolescence:results fromarchivaldataandadirecttestofgenerationalIQchanges Dr.JakobPietschnig,UniversityofVienna

15:15–15:45 AFTERNOONTEA15:45–16:05 Academicaccelerationingiftedyouthandthefruitlessconcernaboutpsychologicalwell-being:a

35-yearlongitudinalstudy Mr.BrianBernstein,VanderbiltUniversity16:05–16:25 HowintelligentandeducatedaretheAmericanelite?Astudyof26,000U.S.leadersacross 30sectors Dr.JonathanWai,UniversityofArkansas16:25-16:45 FindingthemissingEinsteins:searchingforexceptionalabilityintalentsearches,K-12,and

highereducation Dr.JoniLakin,AuburnUniversity16:45-17:05 ThepathtoSTEMleadership:consistentprioritiesbetweenage25andage50 Dr.KiraMcCabe,VanderbiltUniversity

17:05 PosterSession

ISIR2019Schedule

5

Friday12thJuly9:30–10:00 TheinfluenceoffamilialfactorsontheassociationbetweenIQandeducationalandoccupational

achievement:asiblingapproach Ms.EmilieHegelund,UniversityofCopenhagen10:00-10:15 Theassociationbetweenpolygenicscoresforeducationalattainmentandintelligenceis

mediatedbyfibrenetworkefficiency Dr.ErhanGenc,RuhrUniversityBochum10:15-10:30 Gene-environmentinterplayinSESassociationswithIQandeducationalattainment Prof.WendyJohnson,UniversityofEdinburgh10:30-10:45 Istest-takingmotivationaproblemforintelligenceresearch?Noandyes Dr.GillesGignac,UniversityofWesternAustralia

10:45-11:00 Whyisnarcissismthestrongestpredictorofsubjectiveintelligence? Dr.MarcinZajenkowski,FacultyofPsychology,UniversityofWarsaw11:00–11:30 MORNINGCOFFEE11:30-11:50 Familynetworksversusgeneticsinsocialoutcomes,England1750-2019 Prof.GregoryClark,UniversityofCalifornia,Davis11:50-12:10 Economicperformance,strategicbehaviourandintelligence Prof.AldoRustichini,UniversityofMinnesota12:10-13:00 HoldenMemorialLecture RazibKhan,Insitome13:00–14:15 LUNCH14:15-16:15 President’sSymposiumI:Standardizedtestingincollegeadmissions StephenHsu,MichiganStateUniversity NathanKuncel,UniversityofMinnesota CharlesMurray,AmericanEnterpriseInstitute RichardH.Sander,UniversityofCalifornia,LosAngeles AmyWax,UniversityofPennsylvania 16:15–16:45 AFTERNOONTEA16:45-17:00 Usingreactiontimetodifferentiategroupfactorsofopennesstoexperience Ms.YuriKim,UniversityofMinnesota17:00-17:15 Measurementofattentioncontrol Prof.RandallEngle,GeorgiaInstituteofTechnology17:15-17:30 ExaminingconvergentvalidityoftheICARandWAIS-IV Mr.JamisonCarrigan,TheUniversityofTexasatAustin17:30-17:45 Machinelearningpsychometrics:improvedcognitiveabilityvalidityfromsupervisedtraining onitemleveldata Mr.EmilO.W.Kirkegaard,UlsterInstituteforSocialResearch

6

Saturday13thJuly

9:30–9:50 DetectingtheScarr-RoweeffectusingpolygenicscoresintwolargeUSsamples Mr.MatthewSarraf,UniversityofRochester9:50–10:10 Sexdifferencesinsocialandspatialperspectivetaking:areplicationandextensionof Tarampietal.(2016) Ms.ElyssaGeer,FloridaStateUniversity10:10–10:30 Genderdifferencesinmathematicsandreadingastudyonover10,000,000 observations Dr.DavidGiofre,UniversityofGenova10:30–10:50 Intelligenceandhappinessstability:evidencefromRussia Ms.NargizaIbragimova,ERGOResearchandConsultancy10:50–11:20 MORNINGCOFFEE11:20–11:40 Therelevanceoflesionstudiesforunderstandingthereasonforthepositivemanifold DrJohnProtzko,UniversityofCalifornia,SantaBarbara11:40–12:00 GettingatunderlyingmechanismsoffartransfertoGf-themediatingroleofnear transfer Dr.AnjaPahor,UniversityofCalifornia,Riverside12:00–12:20 Intelligenceisassociatedwithintrinsicstabilityoffunctionalbrainmodules Dr.KirstenHilger,GoetheUniversityFrankfurt12:20–12:40 Higherintelligenceisassociatedwithamoreeffectiveadaptationofbrainactivityto cognitivedemands Dr.UlrikeBasten,GoetheUniversityFrankfurt12:40–14:00 LUNCHANDBUSINESSMEETING14:00–14:20 Thefirstlongitudinalinvestigationofcognitiveabilityinadultadoptivefamilies MsEmilyWilloughby,UniversityofMinnesotaTwinCities14:20–14:40 Techtiltpredictsjobs,collegemajors,andspecificabilities:supportforinvestment theories Dr.ThomasCoyle,UniversityofTexasatSanAntonio14:40–15:00 Intelligence'splaceinthepsychologycurriculum DrRussellWarne,UtahValleyUniversity15:00–15:30 AFTERNOONTEAStudentAwardCommitteeshouldmeettodiscussAwards15:30-16:30 Prof.DavidLubinskiinterviewsProf.RandallEngle19.00 BANQUETandpresentationofStudentAwards

7

INVITEDTALKSANDSYMPOSIA

WELCOMEADDRESS

WilliamRevelle1

1NorthwesternUniversity,[email protected]

LIFETIMEACHIEVEMENTAWARDADDRESSProfoundlygiftedadolescentsandtopSTEMgraduatestudentsatage50:

Creativity,productivityandlifestyleatmidlife

DavidLubinskiandCamillaP.Benbow1

1VanderbiltUniversity,[email protected],[email protected]

PRESIDENT’SINVITEDSYMPOSIUMII

Scienceandethicsofgeneticengineering

GregoryCochran1,StephenHsu2,RazibKhan3,BruceLahn4andNevenSesardić

1UniversityofUtah,[email protected],[email protected],[email protected]

4UniversityofChicago,[email protected] [email protected]

HOLDENMEMORIALADDRESSFORDISTINGUISHEDSCIENTIFICJOURNALISM

Beingintelligentabouttheendofthe20thcentury

RazibKhan,Insitome,[email protected]

PRESIDENT’SINVITEDSYMPOSIUMI

Standardizedtestingincollegeadmissions

StephenHsu1,NathanKuncel2,CharlesMurray3,RichardH.Sander4andAmyWax5

1MichiganStateUniversity,[email protected],[email protected],[email protected],LosAngeles,[email protected]

5UniversityofPennsylvania,[email protected]

ISIR2019Presentations

8

Abstracts

Abstractsarepresentedinalphabeticalorderbyfirstauthor’slastname.


9

COGNITIVEENHANCEMENTANDNETWORKEFFECTS

Dr.JonathanAnomaly1andDr.GarettJones2

1OxfordUniversityandUniversityofSanDiego,[email protected] 2GeorgeMasonUniversity

Acentraldebateintheethicsofbiomedicalenhancementconcernstheconditionsunderwhichpeopleshouldbefreetoinfluencethegeneticcharacteristicsoftheirchildren.Wearguethatthecaseforusingmateselection,embryoselection,orgeneticengineeringtocreatechildrenwhoarelikelytoscorehigheronstandardmeasuresofintelligenceandpatienceisstrengthenedbythefactthatthesetraitsarelikelytohavestrongnetworkeffects.Thesenetworkeffectsincludeincreasedcooperationincollectiveactionproblems,whichcontributes to social trustandeconomicprosperity.We thenargue that if individualwelfare is largely a functionof group traits,we should strive to createpolitical institutions and fostersocialnormsthatare likely to increase intelligence.Weendbyofferingsomeconjecturesaboutwhatthese institutions would look like. A guiding feature of these institutions is that they can minimizecoerciveinterferencebyfosteringtheemergenceofsociallybeneficialreproductivenorms.


10

HIGHERINTELLIGENCEISASSOCIATEDWITHAMOREEFFECTIVEADAPTATIONOFBRAINACTIVITYTOCOGNITIVEDEMANDS

Dr.UlrikeBasten,Ms.RebeccaA.Mayer,Ms.RebekkaWeygandtandProf.ChristianFiebach1

1GoetheUniversityFrankfurt,[email protected]

A prominent theory on the neural basis of human intelligence states that brains of more intelligentpeople are more efficient. Empirical evidence for this so-called neural efficiency hypothesis ofintelligence is inconclusive as previous studies have associated intelligence with both weaker andstrongeractivationtocognitivechallenges.Thistalkpresentsanewstudyontheroleoftaskdifficultyasapotentialmoderatoroftheassociationbetweenintelligenceandbrainactivation.For63healthyadultparticipants,weusedfMRItomeasurebrainactivityduringadecisionmakingtaskwith5differentlevelsof difficulty. Intelligence was assessed with a matrix-reasoning test (BOMAT-Advanced). In two brainregions,we observed an interaction of intelligence and task difficulty on brain activation that can bedescribed as a moderating effect of difficulty on the association between intelligence and brainactivation.InrightDLPFC,participantswithhigherintelligencetestscoresshowedastrongerincreaseofactivity with difficulty than participants with lower scores. Conversely, in DMPFC, higher intelligencescoreswereassociatedwitha strongerdecreaseofactivitywithdifficulty.Consequently, in these twobrainregions,intelligence-relateddifferencesinbrainactivationweremorelikelyobservedinthemoredifficulttrialsofourtaskthanintheeasiertrials. Insum,thecurrentstudyconfirmsthatthebrainsofmore intelligent people are not generally more or less efficient. Instead, observed associations maydependontaskdifficulty.Thecurrentfindingsfurthermoresuggestthatinmoreintelligentpeoplebrainactivationismoreeffectivelyadaptedtotaskdemands.


11

ACADEMICACCELERATIONINGIFTEDYOUTHANDTHEFRUITLESSCONCERNABOUT

PSYCHOLOGICALWELL-BEING:A35-YEARLONGITUDINALSTUDY

Mr.BrianBernstein,Dr.DavidLubinskiandDr.CamillaBenbow1

1VanderbiltUniversity,[email protected]

Studies have consistently supported the educational efficacy of acceleration for intellectually precociousyouth. This educational practice has been widely documented for facilitating growth among precociouslearners;and,short-termpositivereportsarerepeatedlyobtainedfromthestudentsthemselves(Assoulineetal.,2015;Colangeloetal.,2004;NationalMathematicsAdvisoryPanel,2008).Yet,teachers,parents,andcounselorscontinuetoexpressconcernaboutpossiblelong-termregretsandpsychologicalmaladaptationengendered by such practices. This study is designed to definitively answer the question of whetheracademicaccelerationresultsinanylong-termnegativeeffectsonthepsychologicaladjustmentandwell-beingofgiftedchildren.

Weutilize longitudinaldatafromthefirstthreecohortsoftheStudyofMathematicallyPrecociousYouth(SMPY),whereallparticipantswereinthetop1%inintellectualabilitywhenidentified,atage13,between1972-1975(Cohort1,N=1133),1976-1978(Cohort2,N=431),or1980-1983(Cohort3,N=237).Theireducational histories prior to high school graduation were subsequently assessed at age 18. They werefollowed-up again at age 50when they completed a number of standardizedmeasures of psychologicaladjustmentandwell-being.Theseincludedfivewell-knownscalesofwell-being:psychologicalflourishing,positiveaffect,negativeaffect(reversed),coreself-evaluations,andlifesatisfaction.Participants’statusonthesemeasures were evaluated in twoways: First, as a function of age of high school graduation and,second, as a function of how much educational acceleration they experienced prior to high schoolgraduation.Psychologicalwell-beingwasnot foundtocovarywitheitherhowearly theygraduated fromhighschoolorhowmucheducationalaccelerationtheyexperienced.

Whenstructuralequationmodelingwasemployedtoaccountforpossibleconfounds(e.g.,socioeconomicstatus),theaforementionedfindingsmaintainedtheirrobustnessandconclusionswerenotaltered.Inthiscontext,abriefdiscussionisgivenontheassumptionsofsuchexpostfactodesigns(Meehl,1970),whichfrequentlyremoveconstructvalidvarianceofthephenomenaunderanalysisand,assuch,aresuboptimal.

Finally,followingLykken(1968,1991),aconstructivereplicationofthesefindingswasconductedusinganindependentsampleofhigh-potentialstudentsidentifiedbyothercriteria.In1992,SMPYsurveyed714topSTEMgraduatestudents(48%females),duringtheirfirsttwoyearsofgraduateschoolatoneofthetop15STEMgraduatetrainingprogramsintheUnitedStates.Thesestudentswerethenassessedatage50withthesamemeasuresnotedabove.Thelackofarelationshipbetweeneducationalaccelerationonthelong-termpsychologicaladjustmentandwell-beingofhigh-potentialstudentswasagainconfirmed.

Counselors, educators, and parents should take comfort in knowing that allowing children learn at theirdesiredrateanddepthdoesnotresultinlong-termpsychologicalmaladjustment.


12

EXAMININGCONVERGENTVALIDITYOFTHEICARANDWAIS-IV

Mr.JamisonCarrigan,Ms.StephanieYoungandDr.TimothyKeith1

1TheUniversityofTexasatAustin,[email protected]

To address the lack of freely accessible cognitive abilitymeasures for primary researchers, Condon andRevelle(2014)developedtheInternationalCognitiveAbilityResource(ICAR),apublic-domainassessmenttoolwithfourinitialitemtypesandseveralothertypesunderdevelopment.ThegoaloftheICARprojectisto encourage broader assessment of cognitive abilities in the social sciences and healthcare fields byprovidingflexible,unrestrictedtestitemstoresearchers(Revelleetal.,2014).OneofthemainbarrierstotheutilityoftheICARisthatnoresearchtodatehasbeenconductedonitsconstructvalidityusingawellestablished,theoretically-guidedcognitiveassessmentbattery.

The aim of the current study is to examine the relations between the ICAR items and the CHC broadabilities as measured by the Wechsler Adult Intelligence Scale—Fourth Edition (WAIS-IV). Correlationsbetween scoreson the respectiveassessments froma convenience samplewillbeexamined foraddressthe following research questions: 1) Is there evidence of the convergent validity of the ICAR whencomparedwithawell-established,long-formmeasureofintelligence?And,2)WhichCHCconstructsarerelatedtotherespectiveICARsubtests?

Participantsincludeasampleof~100universitystudents.Thissamplecombinesstudentsobtainedthroughconvenience sampling (N ~ 70)with students in a clinical sample (N ~ 35). These students completed ademographicquestionnaire,anonlineversionoftheICAR,andselectedsubtestsontheWAIS-IV.

Correlationsbetween thegeneral factors from the ICARandWAIS-IVwillbeexaminedandcompared tothose between other brief assessments of cognitive ability and traditional intelligence batteries. Thegeneral factor for the ICARwill include the four item types, and the general factor on theWAIS-IVwillinclude the four composite scores (VCI, PRI,WMI, and PSI). Range and reliability corrected correlationsbetweenthetotalscoreontheICAR16andtheFSIQandGAIobservedscoreswillalsobeexamined.

Correlational methods of examining convergent and discriminant validity will also be employed.Correlationsofvariableshypothesizedtomeasurethesameconstructwillbecomparedwiththosethoughttomeasuredisparateconstruct (Campbell&Fiske,1959).Correlationsbetweencomposite scoreson therespectiveICARitemtypesandtheWAIS-IVsubtestswillbeevaluated.

ItisexpectedthattotalscoresontheICARwillcorrelatewithgeneralintelligencescoresontheWAIS-IVatamagnitude similar to thoseobserved in correlationsbetween theWAIS-IVandotherbriefmeasuresofintelligence(r>.7)(Salthouse,2014).ItisalsoexpectedthattheICARitemtypeswillbelargelyrelatedtomeasures of fluid reasoning on theWAIS-IV, with the exception of the three-dimensional rotation task,which isexpectedtobemostrelatedtomeasuresofvisual-spatialabilities.Becausethe ICAR isuntimed,lowcorrelationsareexpectedwithmeasuresofprocessingspeedacrossallICARitemtypecomposites.


13

FAMILYNETWORKSVERSUSGENETICSINSOCIALOUTCOMES,ENGLAND1750-2019

Prof.GregoryClark1andProf.NeilCummins21UniversityofCalifornia,Davis,[email protected] 2LondonSchoolofEconomics

In any society individuals are embedded in family networks: parents, grandparents, uncles, aunts, andcousins.Most social science disciplines - anthropology, economics, and sociology - assume these familynetworks play an important role in the social outcomes for children. It is certainly the case that evencontrolling for parent characteristics, the social characteristics of other relatives typically significantlypredictchildoutcomes.

Inthispaperweshow,however,usinganextensivelineageofEnglishfamilieswithraresurnames,330,000personsborn1750-2012,thatthereisgoodevidencethatfamilyconnectionsplayednocausalroleinchildsocialoutcomes.Inthislineageweknowwhichrelativeswerealiveordeadwhenchildrenwereborn.Wealsoknowwhichrelativesweregeographicallyproximateatbirth,andwhichdistant.Themeasuredsocialoutcomesforchildrenwerebeingatschoolaged11-20,attainmentofhighereducation,adultoccupationalstatus, and wealth at death. We find the characteristics of dead and of distant relatives were just asstrongly predictive of social outcomes as living and proximate ones. The characteristics of relativesseeminglyprovideonlyinformationabouttheunderlyingcharacteristicsofparentsthatarebeinginherited.Parentsalonedeterminedchildsocialoutcomes.

Anotherfeaturethatwouldsupporttheimportanceofsocialnetworksinstatusdeterminationwouldbeanasymmetry in effects between relatives on the paternal line compared to those on the maternalline.Throughmuchoftheperiod1750-2019,paternalfamilialconnectionsarebelievedtohavebeenmoreimportant thanmaternalones, in influencingsocial statusoutcomes.We findnosuchasymmetry in thepredictiveeffectsofrelatives’status.

We further show that the size of the relative effects, at least in the case of grandparents, is of themagnitudeasimplefirstorderMarkovmodelofstatustransmissionwouldpredict,wheretheoutcomesofparents derive from their underlying transmittable status onlywith noise.Genetic transmission of socialoutcomes is one such first order Markov model. The English genealogy is thus consistent in itscharacteristicswithgeneticsplayingasurprisinglystrongroleindetermininggeneralsocialstatus.


14

TECHTILTPREDICTSJOBS,COLLEGEMAJORS,ANDSPECIFICABILITIES:SUPPORTFORINVESTMENTTHEORIES

Dr.ThomasCoyle,Ms.KarrieElpers,Mr.MiguelGonzalezandMr.TylerMinnigh1

1UniversityofTexasatSanAntonio,[email protected]

Background: Specific cognitive abilities include ability tilt, based on the difference in math and verbalscores on standardized tests (SAT, ACT). Ability tilt yields math tilt (math>verbal), which predicts STEMcriteria(science,technology,engineering,math),andverbaltilt (verbal>math),whichpredictshumanitiescriteria.Thecurrentstudyexaminedanewtypeof tilt: tech tilt,basedondifferences in technical scoresandacademicscores(mathorverbal).(Technicalscorestappedvocationalskillsforelectronics,mechanics,cars, and tools.) The difference yielded two types of tilt: tech tilt (tech>academic) and academic tilt(academic>tech).Toexamine its validity, tech tiltwascorrelatedwithmathandverbal scoresoncollegeaptitudetests(SAT,ACT)andjobsandcollegemajorsinSTEMandhumanities.

Predictions were based on investment theories. Such theories assume that investments in one domain(math/STEM) boost abilities in similar domains but retard abilities in competing domains(verbal/humanities). Tech tilt reflects a non-academic (vocational) ability, which should compete withacademic abilities (verbal and math). Therefore, tech tilt was expected to correlate negatively withacademic abilities (based on other tests). In contrast, tech tiltwas expected to correlate positivelywithmajorsandjobsinSTEM,whichincorporatestechnicalknowledge(e.g.,electronics,mechanics).

Method:Datawere fromtheNationalLongitudinalSurveyofYouth (N=1950).Tech tiltandacademic tiltwerebasedthe12testsoftheArmedServiceVocationalAptitudeBattery(ASVAB).Academicabilitieswerebasedonmathandverbalscoresonthreecollegeaptitudetests(SAT,ACT,PSAT).Techtiltwasthewithin-subject difference in tech scores (electronics,mechanics, cars) and academic scores (math, verbal) (techminus academic; positive scores indicate tech tilt), yielding tech tilt (tech>academic) and academic tilt(academic>tech).Collegemajorsand jobswereclassifiedasSTEM(chemistry,engineering)orhumanities(journalism, literature). Effects are standardized and are reported as significant at p<.05. The averageofanalogouseffectsisreportedinparentheses(Mr).

Results: Tech tilt correlated negatively and significantly with academic abilities based on themath andverbal scores of the three college tests (Mr=-.25, range = -.41 to -.11), indicating that tech tilt wasassociated with poor performance on the college tests. In addition, tech tilt correlated positively and(generally)significantlywithjobsandmajors(Mr=.20,range=.01to.35),indicatingthattechtiltpredictedSTEMcriteria.Theeffectsreplicatedusingdifferentanalyticalapproaches(e.g.,regressionsandstructuralequationmodeling)andaftercontrollingforgeneralintelligencebasedontheASVAB.

Discussion: The negative effects of tech tilt with academic abilities support investment theories, whichpredictthat investments inonedomain(non-academic,technical)comeattheexpenseof investments incompetingdomains (academic). In addition, theeffectsdemonstrate the validityof vocational aptitudes,extendingpriorresearchonability tilt,whichhas focusedonacademicaptitudes.Futureresearchshouldconsider factors thatmoderate the effects of tech tilt (e.g., education and personality) aswell as othertypesoftilt(e.g.,spatialtilt).


15

LOOKINGFORAFLYNNEFFECT:EXAMININGSHIFTSINCOGNITIVEABILITYWITHINTHESAPAPROJECT

Mrs.ElizabethDworak1,Dr.WilliamRevelle1andDr.DavidCondon21NorthwesternUniversity,[email protected]

Using items from the InternationalCognitiveAbilityResource (ICAR), this studywill examine if there is aFlynnEffectacross12yearsofcross-sectionallycollecteddatafromtheUnitedStates(N=710,139).ICARisan open-source multiple choice cognitive ability assessment used by the web-based Synthetic AperturePersonalityAssessment(SAPA)projecttoexaminethecognitiveperformanceofits’participants.Whileweexpect thatsomedifferencesaredueto temporal trends inself-selection,wehopeto furtherexamine ifcovariates, such as gender, education level, and collegemajor are influencing natural changes in abilityscores. Preliminary analyses comparing 18 and 19 year old participants from 2010 to 2012 and 2014 to2016showedsimilarmatrixreasoningscores(M=0.23;M=0.23),verbalreasoningscores(M=0.36;M=0.36),and letter and number series scores (M=0.25;M=0.25) from 35 ICAR items. Other preliminary analysescomparing34and35yearoldparticipantsbetween thesameyears showedamarginaldecline inmatrixreasoning scores (M=0.30;M=0.27), verbal reasoning scores (M=0.41;M=0.39), and letter and numberseriesscores(M=0.31;M=0.29)from35ICARitems.


16

MEASUREMENTOFATTENTIONCONTROL

Prof.RandallEngle1

1GeorgiaInstituteofTechnology,[email protected]

Many of the most interesting findings and concepts in psychology rely on reaction time (RT) and adifference score between twoRT’s. Reliability of RT and evenRT difference scores are pretty reliable inexperimentalstudies.TaskssuchastheStroop,AttentionNetworkTask,EricksonFlankerTaskandothersfrompersonalityandsocialareassuchastheImplicitAssociationTestareeasilyreplicatedandoftenusedingroupsettingssuchasclassrooms.However,measuringindividualanddevelopmentaldifferenceswithRT and RT difference scores represents a major problem. First, there are large individual anddevelopmentaldifferencesinspeed/accuracytradeoffandverysmallchangesinaccuracycancorrespondto quite large differences in RT. Second, when subtracting two conditions such as congruent andincongruent in theStrooptask, thehigher the reliabilityofeachcondition, the lessvariability remains todistinguish among individual and developmental groups. I report results of an attempt to developmeasuresofattentioncontrolbasedonthresholdaccuracymeasuresandshowtheimpactofthosemorereliablemeasureson several latent variablesandonpredictive validityof theArmedServicesVocationalAptitudeBatteryforaproxyofcomplexwork.


17

SEXDIFFERENCESINSOCIALANDSPATIALPERSPECTIVETAKING:AREPLICATIONANDEXTENSIONOFTARAMPIETAL.(2016)

Ms.ElyssaGeerandDr.ColleenGanley1

1FloridaStateUniversity,[email protected]

Tarampiandcolleagues(2016)conductedastudytobetterunderstandsexdifferencesinspatialandsocialperspective-taking skills. Their findings suggested that adding social information enhanced femaleperformance on perspective-taking tasks.In this preregistered replication, we tested an alternativeexplanation:thatthetaskswithaddedsocialinformationalsoprovidedadditionalspatialinformationwhichcouldexplainimprovementsinperformance.Itispossiblethattheirredesigned‘social’taskalsoprovidedadditional spatial information. We asked: Does social information lead to improved performance on aspatial task? Does additional spatial information lead to improved performance on a spatial task?3) Doeitheroftheseeffectsdifferbysex?

WereplicatedTarampietal.’s(2016)study inasampleof283undergraduatestudents(134males)usingthetwotasksfromtheirstudy:theSOT(HegartyandWaller,2004)andtheroad-maptest(Moneyetal.,1965; modified by Zacks et al., 2002). We used the conditions from their study (one with no social oradditional spatial information [original tasks]andonewithboth social andadditional spatial information[theirsocialtasks]),andalsoaddedtwonewconditionsthathelpisolatethepotentialeffectsofsocialandadditional spatial information.Wealso includedamath fluency taskand several self-reportmeasuresofsocialskillsandanxieties.

ToreplicateTarampietal.(2016),werantwoindependentsamplest-testcomparingperformanceoneachof the two perspective-taking tasks between females in the no spatial, no social condition (their spatialcondition) and females in the spatial, social condition (their social condition). There was no significantdifference in performance on the road map task (t(74)=1.07, p=.289) or the SOT (t(74)=-1.06, p=.293)suggestingthat,contrarytothefindingsfromTarampietal.(2016),femalesintheirsocialconditiondidnotout-performfemalesintheirspatialcondition.

Totestouralternativehypothesis,werana2(additionalspatialinformation:yes,no)x2(socialinformation:yes,no)x2(sex:female,male)foreachoftheperspective-takingtasks.FortheRoadMaptasktherewasamain effect of sex (F(1,275)=32.82,p<.001) such thatmales outperformed females. In partial support ofTarampi et al.’s (2016) hypothesis, therewas amain effect of social information (F(1,275)=3.94,p=.048)suchthatperformancewasbetterontheversionsofthetaskthathadadditionalsocialinformation.Therewas a main effect of additional spatial information (F(1,275)=11.28, p=.001) such that adding spatialinformationledtoimprovedperformanceonthetask,aswehypothesized.FortheSOTtherewasamaineffectofsex(F(1,275)=25.42,p<.001)suchthatmalesoutperformedfemales,buttherewerenosignificantmain effects for either social (F(1,275)=2.79, p=.096) or additional spatial (F(1,275)=3.726, p=.055)information.

The results for the road map task show some support that additional social information does improveperformanceonaspatialtask,butthiseffectdoesnotdifferbysexasTarampietal.(2016)posited.Theseresultsalsosuggestthatadditionalspatialinformationimprovesperformanceonaspatialtask,regardlessofsex,whichlendssomesupporttoourhypotheses.


18

THEASSOCIATIONBETWEENPOLYGENICSCORESFOREDUCATIONALATTAINMENTANDINTELLIGENCEIS

MEDIATEDBYFIBERNETWORKEFFICIENCYDr.ErhanGenc,Mr.ChristophFraenz,Ms.CarolineSchlüter,Prof.ManuelVoelkle,Prof.OnurGüntürkün,

Dr.SebastianOcklenburgandProf.RobertKumsta11RuhrUniversityBochum,[email protected]

Individualdifferencesincognitiveabilitiesaresubstantiallyinfluencedbygeneticvariability.Intelligenceisahighlypolygenictrait,withthousandsofallelescontributingwithsmalleffects.Aninnovativewaytoassessthecumulativeeffectofgeneticvariationontraitsof interest is tousepolygenicscores (PGS),whicharederived from large genome-wide association studies (GWAS). These PGS are increasingly used inintelligence research. Whereas PGS for IQ or educational attainment reliably predict interindividualdifferences in cognitive performance, the intermediate steps in the pathway from genes to cognitiveabilitiesare largelyunknown.Giventheassociationbetweenthebrain'smacrostructureanditsstructuralnetwork properties with cognitive performance, we tested whether the association between PGS andintelligenceismediatedviabrainvolumeandfibernetworkconnectivity.

Structural network efficiency and intelligence was assessed in a sample of 300 healthy individuals.Intelligence was measured with a broad cognitive test battery. Polygenic scores - a summary geneticmeasureusedtoassess thecombinedeffectof singlenucleotidepolymorphisms (SNPs) - foreducationalattainmentwerederivedusingdatafromthelatestGWASonadulteducationalattainmentwithasamplesizeof1.1millionindividuals.Furthermore,weusedstandardMRIalongwithDTIandfibertractographytoexaminedifferentestimatesofbrainvolumeaswellasbrainfibernetworkconnectivitybymeansofgraphtheory.

Consistent with previous findings, PGS significantly predicted cognitive test performance (R2=.062).Furthermore,participantswithlargerbrainsandmoreefficientfiberconnectivityachievedhigherIQscores.Mediationanalysesshowedthattheassociationbetweenpolygenicscoresandintelligencewasmediatedbyfibernetworkefficiency,bothmeasuredforthewholebrainandforsinglebrainregionspredominantlylocatedinfronto-parietalareas.Brainvolumewasnotasignificantmediator.Theresultsofourstudyshowthat the pathway from genes to cognitive performance involves the shaping of fiber network efficiency.Overall, our findings represent a crucial step in the endeavor of identifying the missing links betweengeneticvariabilityandcognitiveperformance.


19

ISTEST-TAKINGMOTIVATIONAPROBLEMFORINTELLIGENCERESEARCH?NOANDYES

Dr.GillesGignac11UniversityofWesternAustralia,[email protected]

It has been long assumed that examinees need to apply their best effort when completing intelligencetests. However, individual differences in test-taking motivation exist, and these differences have beenargued to compromise interpretations of IQ scores as valid representations of intelligence. To date, theassociation between test-taking motivation and intelligence test performance has been examined onlyrelatively rarely.In study 1 (N=173), test-takingmotivationwasmeasured objectively with the AnagramPersistence Task. In study 2 (N=218), test-taking motivation was measured subjectively with aquestionnaire.Acrossbothstudies,intelligencewasmeasuredwithabatterydesignedtomeasuregeneralintelligence.Study 2 also included a within-subjects design to evaluate the degree to which test-takingmotivation might reduce across a 50 minute testing session. In both studies 1 and 2, the positiveassociationbetweenobjectivelyandsubjectivelymeasuredtest-takingmotivationandgeneralintelligencewas found to be curvilinear: the majority of the positive effect was isolated between relatively low tomoderatelevelsofthetest-takingmotivation.However,acrossthe50minutetestingsessioninstudy2,thepercentageofinsufficientlymotivatedexamineesmorethandoubledfrom6%to13%.Theresultssuggestthatonlyamoderateleveloftest-takingmotivationmayberequiredforvalidintelligencetesting.However,testing that exceeds 30minutes should probably be discouraged, given the increase in the insufficientlymotivated.


20

GENDERDIFFERENCESINMATHEMATICSANDREADINGASTUDYONOVER10,000,000OBSERVATIONS

Dr.DavidGiofre1,Dr.EnricoToffaliniandProf.CesareCornoldi2

1UniversityofGenova,[email protected] 2UniversityofPadova

The question of gender differences in mathematics is very important and captured the attention ofscientists,practitioners,andteachersworldwide.Despitethisimportance,mostoftheresearchongenderdifferencesinmathematicshasbeenbasedonEnglishspeakingcountries,e.g.,UK,USA,orCanada,whileonlylimitedevidenceisavailableongenderdifferencesinothercontextsandrealities(Else-Quest,Hyde,&Linn,2010).

Efforts to measure the mathematics achievement cross-nationally have produced two large-scaleassessments, the Trends in InternationalMathematics and Science Study (TIMSS) and PISA. TIMSS is aninternational assessment ofmathematics and science learning in eighth graders, conducted on a 4-yearcycle.While, PISA is an international assessment ofmathematics, reading, science, and problem-solvingliteracyin15-year-olds,conductedona3-yearcycle.Thesemeasuresareinfactveryrobustandareveryhighlycorrelatedwithformalintelligencetests.

Results from TIMSS and PIRLS have produce an invaluable contribution to the evaluation of genderdifferences inmathematics.However, there are several limitswith these two international assessments.Forastart,theresultstendtobelimitedwithcertaingradesandagegroups,mainlyeightgradersand15-yearsoldchildrenonly.Moreover,thesestudiesarenotperformedeveryyear, thusmaking itdifficulttoassesschangesoverthecourseoftime.Finally,themathematics`curriculaindifferentcountriesmightalsopresentseveraldifferences,thusmakingthecomparisonbetweencountriessomewhatdifficult.However,TIMSSandPIRLSarenottheonlyevaluationsandseveralalternativesmathematicsassessmentshavebeenpromotedinsomecountries.

TheINVALSI(IstitutoNazionaleperlaValutazionedelSistemaEducativodiIstruzioneediFormazione)isanationalexaminationofmathematicsandreadingcompetencesofchildrenintheir2nd,5th,6th,8th,and10th grades. This assessment typically involves, two subjects, i.e., reading and mathematics, and isperiodicallyperformedeverysingleyear.DatafromtheINVALSIareveryvaluablebecause,contrarytotheaforementioned evaluations (PISA and TIMSS) they are performed yearly, and they allow to comparechildrenacross several grades. Finally, results from INVALSI canbedifferentiatedamongdifferent Italianregions, and are hierarchically structured, thus allowing to control for the effect of the school and classsettings. For all these reasons, this database seem to be particularly appropriate in order to test thedevelopmentofgenderdifferencesacrossseveralyearsandseveralgrades.

Inthispaperover10,000,000observationsonmathematicsandreadingwereanalysedoverthecourseof8years. This sample is representative of the Italian population. Rasch scores were calculated for eachindividual subject. Findings were then collapsed using ameta-analytic approach. Results showed a veryimportant geographical gradient behind the gender gap inmathematics. Surprisingly, gender differenceswere more marked in some reached areas of the country. The present findings have very importantimplicationsforourunderstandingofgenderdifferencesonhighercognitiveabilities.


21

THEINFLUENCEOFFAMILIALFACTORSONTHEASSOCIATIONBETWEENIQANDEDUCATIONALANDOCCUPATIONALACHIEVEMENT:ASIBLINGAPPROACH

Ms.EmilieHegelund1,Ms.TrineFlensborg-Madsen1,Dr.JesperDammeyer1,Mr.LaustHvasMortensen2andProf.ErikLykkeMortensen1

1UniversityofCopenhagen,[email protected] 2StatisticsDenmark

Intelligence test scores have been shown to correlate positively with educational and occupationalachievement innumerousobservational studies.However,becauseof theobservational studydesigns, itcannot be ruled out that part of the observed associations is explained by unmeasured familial factorsoperatinginchildhood.Studiesoftwinsandnon-twinsiblingpairshavefoundthatbothgeneticfactorsandsharedenvironmental factorsare in fact responsible forpartof theobservedassociations.However, thegeneralizabilityofthestudyfindingstothegeneralpopulationisuncertainbecausethestudypopulationsare restricted to samples of twins or samples of biological and adoptive sibling pairs. It is therefore aninteresting question how much of the associations can be attributed to familial factors operating inchildhoodinstudypopulationsthataremorerepresentativeofthegeneralpopulation.

Usinga largecohortconsistingofnearly365,000siblingswithinmore than170,000sibships, thepresentregister-based study tried to come one step closer to an answer to this question by investigating theinfluence of familial factors shared by siblings on the association between IQ and educational andoccupationalachievementamongyoungpeopleinDenmark.ThestudypopulationconsistedofallDanishmenwithat leastone fullbrotherwhereboth the individualandhisbrotherswerebornsince1950andappearedbefore a draft board in 1968-1984 and1987-2015 (N=364,193 individuals&171,037 sibships).IntelligencewasmeasuredbyBørgePriensPrøveatage18.

Educationalandoccupationalachievementwasmeasuredbygradepointaverage(GPA)inlowersecondaryschool, by time to receiving social benefits at ages 18-30, and by gross income at age 30. Covariatesincludedthestudypopulation’syearofbirth,ethnicity,birthregion,binaryindicatorsofout-of-homecareinchildhood,psychiatricdiagnoses inchildhood,neurologicaldiagnoses inchildhood,perinataldiagnosesandcongenitaldeformities,andparentalsocioeconomicpositionatbirth.

Thestatisticalanalysesconsistedof twodistinct statisticalanalysesofall the investigatedassociations:Aconventionalcohortanalysisandawithin-sibshipanalysisinwhichtheassociationunderinvestigationwasanalysed within siblings while keeping familial factors shared by siblings fixed. The findings of theconventionalcohortanalysisandthewithin-sibshipanalysiswereafterwardscomparedbyvisualinspectionoftheconfidenceintervalstodeterminetheimportanceofconfoundingfromunmeasuredfamilialfactorsoperatinginchildhood.

The results showed thatanappreciablepartof theassociationsof IQwitheducational andoccupationalachievementcouldbeattributedto familial factorssharedbysiblings.However,mostof theassociationswerenotattributabletosuchfamilialfactors.Infact,visualinspectionoftheconfidenceintervalsrevealedthatonlytheassociationbetweenIQandGPAinlowersecondaryschoolwithinsiblingsdifferedstatisticallysignificantly from the association observed in the cohort analysis after covariates had been taken intoaccount. This might either suggests that proximal family factors, for which we have no informationavailable, only influence educational achievement during upbringing or that these factors have a muchsmallerinfluenceonlatereducationalandoccupationalachievementinyoungadulthood.


22

INTELLIGENCEISASSOCIATEDWITHINTRINSICSTABILITYOFFUNCTIONALBRAINMODULES

Dr.KirstenHilger1,Dr.MakotoFukushima2,Prof.ChristianFiebach1andProf.OlafSporns31GoetheUniversityFrankfurt,[email protected] 2GraduateSchoolofFrontierBiosciences,OsakaUniversity3IndianaUniversityBloomington

Intelligencepredicts important lifeoutcomes likeeducationalsuccess,health,or longevity.Previousworkdemonstratedthatvariationsinthemodularorganizationofresting-stateconnectivityrelatesignificantlytoindividualdifferences inhumanintelligence.However,while intelligentbehavior impliesgreaterflexibilityinadaptingtochangingenvironmentaldemands,itissofarunclearwhetherandhowdynamicchangesinthe configuration of brain networks, for example in their modular organization, relate to generalintelligence.Toclarifythisrelationship,weusedmulti-bandfMRIresting-statedatafromN=281healthyadult subjects and estimated subject-specific time-dependent functional connectivity from fMRI timeseries. A modularity optimization algorithm was applied to determine individual time-variant modulepartitionsandnetworkmetricsthatcapturedfluctuationsinmodularityacrosstime.Weshowthathigherintelligence,indexedbyanestablishedcompositemeasure,theWechslerAbbreviatedScaleofIntelligence(WASI), is associated with higher stability of brain network modularity. Post hoc analyses reveal thatsubjects with higher intelligence scores engage in fewer periods of very high modularity – which arecharacterizedbya greaterdisconnectionof task-positive from task-negativenetworks. Further,we showthatattention-relatedbrainregionsofthedorsalattentionnetworkcontributemosttotheobservedeffect.In sum, our results suggest that higher general cognitive abilities are associatedwithmore stable brainnetwork dynamics, particularly in the dorsal attention systems of the human cortex. Our studydemonstrates that the investigation of temporal dynamics of brain networks can contribute to ourunderstandingoftheneuralbasesofindividualdifferencesingeneralcognitiveabilities.


23

GENE-ENVIRONMENTINTERPLAYINSESASSOCIATIONSWITHIQANDEDUCATIONALATTAINMENT

Prof.WendyJohnson1,Prof.MattMcGueandProf.WilliamG.Iacono2

1UniversityofEdinburgh,[email protected] 2UniversityofMinnesota

SeveralstudiesintheUnitedStateshaveindicatedthatgeneticvarianceinIQtendstobegreaterinyouthsfrom higher-SES backgrounds, and theMinnesota Twin Family Study (MTFS; ~2500 twin pairs) is amongthem.Thisisgenerallyinterpretedasanexampleof'vantagesensitivity'--peoplevarygeneticallymoreinabilitytobenefitfromadvantageousenvironmentsbutnotsomuchinhowdisadvantageousenvironmentsundermine development. Contrasting this is 'stress-diathesis -- people vary in genetic vulnerability todisadvantageous environments but not so much in how advantageous environments support them. InMTFS, genetic variance in educational attainment is greater among young adults from lower-SESbackgrounds--anexampleofstress-diathesis.SchoolengagementisgenerallyassociatedwithbothIQandeducational attainment; MTFS is no exception. There, its variance follows the stress-diathesis pattern.WhenitisconsideredanenvironmentaffectinglaterIQandeducationalattainment,bothshowthestress-diathesis pattern for genetic variance, though shared environmental variance in educational attainmentshowthevantage-sensitivitypattern.Iwilldiscusshowtheseobservationsmighthelpusunderstandhowgenesareinvolvedinmaintainingintergenerationalsocio-economicpatterns.


24

USINGREACTIONTIMETODIFFERENTIATEGROUPFACTORSOFOPENNESSTOEXPERIENCE

Ms.YuriKim,Ms.EmilyWilloughbyandDr.JamesLee11UniversityofMinnesota,[email protected]

Intelligence and the nature of its relationship with personality has been fraught with disagreement.Intelligencehasbeenarguedtobecompletelyseparatefrompersonality;related,butdistinct;oranaspectof it. However, intelligence has consistently shown meaningful relationships to the Big Five factors ofpersonality,suggestingthatintelligencecouldbeanattributeofpersonality(Stanek,2014;DeYoung,2011).IntheBigFiveAspectsScale(BFAS)model,eachfactorispartitionedintolower-levelAspects(DeYoungetal.,2007).OpennesstoExperienceiscomprisedofOpennessandIntellect.IntelligenceisnotequivalenttoIntellect (intellect is defined as an individual’s perceived intelligenceor intellectual engagement); rather,intelligenceisoneofmanyfacetsencompassedbyIntellect.

IpresentlyseektofurtherexploretherelationshipbetweenBigFive—specificallyOpennessandIntellect—andintelligenceviacognitiveability.Inthisproject,mycollaboratorsandIassessbothpersonalitythroughthe BFAS, and cognitive ability through reaction time on two speeded cognitive tasks. In the first task,participantsarepresentedwithanumber(0to4and6to9)andmustdecideifthatnumberisgreaterthanorlessthan5.Inthesecondtask,participantsarepresentedwithtwotonesanddecideifthesecondtonewas higher or lower in pitch. If intelligence is a facet of Intellect,we expect reaction time to negativelycorrelatewithIntellect.IndividualswithhigherIQstendtocometothecorrectdecisionmoreswiftly.WedonotexpecttheOpennesstohaveasignificantcorrelationwithreactiontime

The design of the study also allows us to test implications of the Cybernetic Big Five Theory, or CB5T(DeYoung,2015).TheCB5Tprovidesamechanisticexplanationforhowpersonalityproducesgoal-directedbehaviors.IndividualswhoscorehighlyonIntellectmayhavequickerreactiontimesinpartbecause,astheCB5T proposes, intellect relates to the detection of logical patterns. For example, in the number task,decidingthat1islessthan5islessdifficultthandecidingthat4islessthan5.Theoretically,individualswithhighIntellectshouldbelesshinderedbythedifficultyofdecision(e.g.thestimulusnumber’sdistancefrom5). Furthermore, according to the CB5T, Openness relates to the detection of correlational patterns insensory or perceptual information. In the experiment, we also test the difficulty of perception. In thenumbertask,thecolorofthestimulusnumbervariesinhowstronglyorweaklyitstandsoutfromthecolorofthebackground.TheCB5TpredictsthatindividualswhoscorehighlyonOpennesswillbelesshinderedbyvariationinthestimulusnumber’ssaturation

Wearethusseekingevidenceforseparateinformationprocessingaswellasevidenceforpossibleneuralbasesofthetwopersonalityaspects.

Wearecurrentlycollectingparticipantsandexpecttohaveatleast300bytheconference.


25

MACHINELEARNINGPSYCHOMETRICS:IMPROVEDCOGNITIVEABILITYVALIDITYFROMSUPERVISEDTRAININGONITEMLEVELDATA

Mr.EmilO.W.Kirkegaard1,Dr.CurtisDunkel2,Dr.ShaneMcLoughlin3andMr.AndrewCutler41UlsterInstituteforSocialResearch,emil@emilkirkegaard.dk2WesternIllinoisUniversity3UniversityofChester4BostonUniversity

Has psychometrics overlooked machine learning methods? We investigated whether machine learningmethodscouldimprovethescoringofcognitiveitemdata.Todothis,wecollected7largedatasetsofitemdata(totaln=37k).Mostdatasetsprovidedresponse-leveldatai.e.whichresponsesubjectgave,notjustbinary(correct/incorrect).Datasetscollectivelyhadmanyoutcomesofinterest,butwefocusedonasmallnumberthatmostlyoverlappedbetweendatasets:age,sex,educationalattainmentandincome(all[quasi-]continuousasidefromsex).Wethenappliedstandardpsychometricscoringmethods,sumscoresanditemresponse theory (IRT) scores, to the data as well as a variety of supervised machine learning methodsincluding random forest, lasso/ridge regression, deep neural networks, as well as unpenalized ordinaryleastsquares(OLS)forcomparison.Parametersweretunedusingefficientleaveoneoutcross-validationonatrainingset.Performancewasmeasuredon20%holdoutdata.

Our results indicate thatmachine learningmethods regularly outperform standard psychometric scoringmethods.Acrossdatasets,ameangaininvalidityof47%,17%,and13%wereseenforage,education,andincome, respectively. This gain was fairly consistent across datasets and test item types, i.e. machinelearning methods were able to use all tested item types to extract extra validity, including vocabulary,memory,verbalfluency,matrices/Raven’s,generalknowledge,andmath.

Of themachine learningmethods,manywere roughly equivalent. Ridge regressionwas overall the bestmethod.This indicates that:1)Sparsityofeffects isnotagoodassumption for thesedata, i.e. thateachresponseoptionwasuniqueinutility.Inotherwords:thedifferent‘distractor’(wrongoptions)inmultiplechoicequestionsaredifferentially informative,notequivalentasassumedbythebinaryscoringmethods.2)thatinteractionsbetweenitemsdonotseemtoplayimportantrolesforpredictivepurposes,inlinewithtraditionalpsychometricresults.

Weconclude thatstandardscoringapproachesofcognitivedataaremissingextravaliditypresent in thedata, sometimes a lot of it, depending on the outcome. This finding has implications for tests used forpractical purposes (such as selection, dementia screening), where their validity has likely beenunderestimated.


26

FINDINGTHEMISSINGEINSTEINS:SEARCHINGFOREXCEPTIONALABILITYINTALENTSEARCHES,K-12,ANDHIGHEREDUCATION

Dr.JoniLakin1andDr.JonathanWai2

1AuburnUniversity,[email protected] 2UniversityofArkansas

Standardized tests have played and continue to play an important role in gifted talent searches, K-12education, and higher education selection. These testsmeasureg to a large degree (Frey&Detterman,2004),butalso tap intosecondaryquantitativeandverbal reasoning factors.Theseassessmentsserveasgatekeepers to specialized services including summer programs and elite universities. One criticism ofcommonly used screening tests is that they disproportionately select students from more advantagedgroups,whichhas ledadmissionsdecisions tomovetowardsholistic reviewto improvethediversityandbackgrounds of admitted students. This, however, means that selection becomes more subjective,especiallywhencoupledwith recentevidence that“universally screening”or testingall studentsactuallyimprovestherepresentationofstudentsfromdisadvantagedbackgrounds(Card&Giuliano,2016).Insteadof jettisoning effective screening tools, which are more sensitive to educational opportunity andenrichment, for subjective selection, we propose broadening the scope of screening tests to includemeasuresofspecificabilitiesthatarelessdirectlyinfluencedbyformaleducation.Likeexistingmeasures,theseadditionalscreeningtoolscouldbeaddedonthebasisoftheincrementalpredictivevalidityofsuchtestsonnumerousreal-worldoutcomes(Kunceletal.,2004).

Richard Snow (1999) argued that the breadth and depth of admissions testing needed expanding, withspatialabilitiesbeingonesetofmeasuresthatcouldbepromising,givenwhatweknowaboutthestructureof cognitive abilities. Research has indicated a link between spatial abilities and long-term STEM andcreative outcomes (Wai et al., 2009; Kell et al., 2013). However, the question remains as to how tosimultaneouslyexpandthesetsofmeasuresused inadmissionswhileadequatelyaddressingtheneedtoensuretalentedbutdisadvantagedstudentsarerepresentedtotheirfullestcapacity.

In this studywe aim to examinewhich of awide array of cognitivemeasuresmight bemost fruitful inobjectively identifying disadvantaged talent (e.g., low SES, from rural areas, underrepresentedminority)across threenationally representeddatasets: Project Talent (1960),High School andBeyond (1980), andtheNLSY (1997).We leverage these older datasets and allow allmeasures to compete to determine towhat extent 1. Broader screening can identify more disadvantaged talent proportional to their grouprepresentation and 2.Which of these assessments simultaneously have predictive validity for importanteducationaloutcomes.

AnalysisoftheProjectTalentsampleiscomplete.Findingssofarindicatethatspatialreasoningmeasuresbroadly—inparticular2DSpatialReasoning—improvetheproportionalrepresentationoflowincome,rural,and underrepresented minority talent, and that such measures have reasonable predictive validity onshort- and long-term educational outcomes of interest. Next we intend to explore the extent to whichthese findings replicate in the HSB and NLSY samples. These findings build upon the fundamentalunderstandingthatallstandardizedtestsmeasurecognitiveabilities,andwillhelpdeterminetheempiricalstartingpointfromwhichexpandingthebreadthofadmissionstestingtohelpidentifydisadvantagedtalentingiftedtalentsearches,K-12,andhighereducation,mightbemostfruitful,


27

THEPATHTOSTEMLEADERSHIP:CONSISTENTPRIORITIESBETWEENAT

AGE25ANDAGE50

Dr.KiraMcCabe,Dr.DavidLubinskiandDr.CamillaBenbow11VanderbiltUniversity,[email protected]

In1992,714first-andsecond-yeargraduatestudents(48.5%female)attendingU.S.universitiesrankedinthetop15byscience,technology,engineering,andmathematics(STEM)fieldweresurveyed.ParticipantsprovidedtheirGraduateRecordExamscores,andtheycompletedstandardized interest,personality,andvaluesquestionnaires;theyalsoweresurveyedontheirworkandlifestylepreferences.Twenty-fiveyearslater,theirpublicallyavailableliferecordswereevaluatedtoidentifywhobecameaSTEMleaderandwhodid not become a STEM leader (labeled “nonleaders”). Our approach classified all 714 participants.Following the Berkeley Studies of Creativity, Institute of Personality Assessment Research, we reasonedthatifsomeonewastrulyeminentintheirfield,theyshouldbepubliclyidentifiablebyleadersinthefield.The results were as follows: Male STEM leaders (n=97), Female STEM leaders (n=55), Male nonleaders(n=271),andFemalenonleaders(n=291).

We conducted a series of discriminant function analyses to determine whether Time 1 (age 25) datadistinguished these four groups. Each analysis used a different and unique set ofmeasures: a) abilities,interests,andvalues,b)specific interests,c) lifestylepreferences,andd)workpreferences.Eachanalysisshowed twocleardiscriminant functions,one thathighlighted leadershipdifferences, and theother thathighlightedgenderdifferences.Theseresultstellanimportantstoryaboutthemultiplicityofdeterminantsthat combine to produce STEMeminence. STEM leaders have distinctive abilities, STEM interests, fewercompeting interests, and dedicate an inordinate amount of time to their career. These results highlightamongthenonleadersmorecompetinginterestsandlifestyleprioritiesrelativetoSTEMleaders.Thereisnorightorwrong,asdifferentpeoplerequiredifferentthingstocreateameaningfulandsatisfyinglife.

Further analyses that examined leadership differences and gender differences among the participantsincluded the extent to which age 50 data reflects similar priorities and preferences reported at age 25during graduate school. We also examined the potential impact of having children. We found similardifferencesbetweenleadersandnonleadersintheirlifestyleandworkpreferencesatage25andatage50.Leadersreportedworkingmorehoursthannonleaders,andtheyplantoworkmorehoursthannonleadersover the next 15 years.On average,male leaders also reportedwilling to retire 5 years later thanmalenonleaders. Takenas awhole, these results showa consistentpatternofpriorities amongSTEM leadersoverthecourseoftheirlives.


28

GETTINGATUNDERLYINGMECHANISMSOFFARTRANSFERTOGF–THEMEDIATINGROLEOFNEARTRANSFER

Dr.AnjaPahor1,Prof.AaronSeitz1andProf.SusanneJaeggi21UniversityofCalifornia,Riverside,[email protected] 2UniversityofCalifornia,Irvine

The extent to which working memory training leads to improved performance on tasks that are of adifferent nature than those used for training remains unclear. Moreover, little is known about themechanisms underlying these so-called far transfer effects tomeasures of Gf such asmatrix reasoning.Hereweusemediationmodels to test thehypothesis that far transfer ismediatedbynear transfer.Thesampleconsistedofhealthyadults(Ncohort1=94;Ncohort2=177)thatwererandomlyassignedto10daysofN-backtrainingortoapassivecontrolgroupandcompletedabatteryoftasksatpre-testandpost-test.Atpost-test,performanceonuntrainedN-backtasks(neartransfer)improvedinbothactivegroupsrelativetopre-test (Cohort 1: d = .80; Cohort 2: d = .96), whereas little-to-no improvement was observed in thepassive control groups (Cohort 1:d = .06; Cohort 2:d = .14). Similarly, post-test performance onmatrixreasoningtasks(fartransfer)improvedinbothactivegroupsrelativetopre-test(Cohort1:d=.49;Cohort2:d=.43)butshowedverylittleimprovementinthepassivegroups(Cohort1:d=-.02;Cohort2:d=.19).Multiplelinearregressionshowedthattraininggroupwasasignificantpredictorofpost-testperformanceon matrix reasoning, even after accounting for pre-test performance and cohort effects. Subsequentanalyses demonstrated that this relationship was mediated by performance on untrained N-back tasks,whichwasreplicated ineachofthecohorts(Cohort1: IE=0.03,95%CI=0.01–0.05;Cohort2: IE=0.04,95%CI=0.02–0.06).Incontrast,training-relatedimprovementonaSimon-liketaskdidnotmediatepost-testmatrix reasoning performance. These findings suggest thatN-back training does notmerely lead totask-specificlearning,but,atleastforsomeindividuals,leadstochangesintheunderlyingcomponentsofworkingmemorythatleadtofartransfereffects.


29

SMALLERFLYNNEFFECTSFORCRYSTALLIZEDINTELLIGENCEMAYBEROOTEDINITEMOBSOLESCENCE:RESULTSFROMARCHIVALDATAANDADIRECT

TESTOFGENERATIONALIQCHANGES

Dr.JakobPietschnig,Prof.GeorgGittler,Ms.FranziskaHöltl,Dr.UlrichTranandProf.MartinVoracek11UniversityofVienna,[email protected]

Background:GenerationalIQtestscorechanges(i.e.,theFlynneffect)havebeenpositiveovermostofthe20thcenturywithrecentstudiesindicatingstagnationandareversalinsomecountriesaroundtheturnofthecentury.ForthepositiveFlynneffect,someofthebestreplicatedfindingsrelatetostrongerfluid(i.e.,reasoning) than crystallized (i.e., knowledge) gains. Whilst some researchers attributed these domaindifferencestoabilitydifferentiationordifferentialsocialmultipliereffectfunctioning,conclusiveevidencefor a plausible cause is still missing. One viable hypothesis pertains to domain-specific effects of itemobsolescence. This hypothesis suggests that weaker crystallized intelligence gainsmay be due to verbalitemsbecomingcontent-wiseoutdatedovertime(orsolutionsbecomingoutrightfalse),whilstsolutionstonon-verbal (fluid) tasks remain qualitatively unaffected. Consequently, test revisions of fluid intelligence(sub-)testsareoftenconfinedtomereupdatingoftestnormswhilst incrystallizedintelligence(sub-)tests,itemsareoftenrevisedorreplacedaltogether.Thesedomain-specificdifferencesinobsolescenttestitemtreatmentmayconceivablymaskcrystallizedIQgains.

Methods:First,wemeta-analyticallyinvestigatedinfluencesofageontestscoregainsonperformance(k=98;N=10,769)andverbalIQ(k=97;N=10,749)forsamplesthathadtakenaWAISorWISCtestandoneofitsrestandardizationsatthesametime-point.Second,weexaminedinteractioneffectsofageanddatacollectionyearonanunrevisedmeasureofcrystallizedintelligenceinalargenumberofadultpatientsfrom1978to1994(N=5,441).Finally,weprovideadirect,targeted,andpreregisteredtestofitemobsolescenceeffectsoncrystallizedIQchangesinasampleofGerman-speakers.Wecross-sectionallyadministeredthreeoriginal and revised subtests (number series, sentence completion, similarities) of the IST, a well-established IQtest inGerman language,toaconveniencesampleof158healthyadults (78f;meanage=23.7).

Results: Consistent with our predictions, meta-analytical data showed positive (albeit non-significant)associationsofagewithcrystallized,butnotfluidIQgains.Moreover,weobservedasignificantinteractionbetweenparticipantageanddatacollectionyearinourlongitudinalsample,indicatingstrongerIQgainsforolderthanforyoungerparticipants.Inourcross-sectionalsample,weobservedapositiveFlynneffectfornumber series, virtually none for similarities, but a negative one for sentence completion. Although IQsdecreased over time in sentence completion (i.e., our most crystallized subtest; d = -1.076), scores onreplacedandreviseditemsweresignificantlylargerthanscoresontheoriginalitems(d=1.161and0.285),indicatingnon-replacedand-reviseditemsasthedriversofdecreasingIQs.

Discussion: We show convergent evidence from three studies for influences of item obsolescence oncrystallized Flynn effect estimations. In the light of our findings, it seems likely that past evidence forcrystallized IQgainsmayhaveunderestimatedtheFlynneffectandtruedomaindifferencesareprobablysmaller thantypicallyobserved. Inall,our resultssuggest that true fluidandcrystallized IQchangesovertimemayhaveprogressedatamoresimilarratethanpreviouslyreported.


30

THERELEVANCEOFLESIONSTUDIESFORUNDERSTANDINGTHEREASONFORTHEPOSITIVEMANIFOLD

Dr.JohnProtzko1andProf.RobertoColom21UniversityofCalifornia,SantaBarbara,[email protected] 2UniversidadAutónomadeMadrid

Thepositivemanifoldisamongthemostreliablepsychologicalfindingsworldwide.Theexplanationforwhyall cognitive tests covary, however, are many. Explanations tend to fall into one of three categories:commoncause;sampling/network;orcausalinteractionistmodels.Fitstatistics,weargue,areinadequateas a means to demarcate between models. Each model, however, is at its heart a causal model withcausallytestablepredictions.Onewaytotestthemodelsagainsteachotheristolookatthemanipulabilityofpartsofeachmodelandtestthecausallinksthemodelsimply.Brainlesionsrepresentonesuchpseudo-causalapproachtounderstandingmeasurementmodels.Thebasicpatternoftheeffectsofbrainlesionsis:focal cortical lesions lead to local, not global, deficits. Here we review the evidence of not only whatabilitiesareaffectedbywhichlesions,butcriticallywhichabilitiesareunaffectedbywhichlesions.Wefindthathierarchicalmodels,bifactormodels,andsomebutnotallversionsofsampling/networkmodelsareabletoaccommodatethepatternoflesionevidenceonthepositivemanifold.Causalinteractionistmodels,some versions of sampling/network models, and correlated factor models, however, are unable toaccommodate such evidence. The causal connections implied by ourmeasurementmodels represent anadditional way of demarcating between equally well-fitting models for understanding the structure ofcognitiveabilitiesandthecauseofthepositivemanifold.


31

WORKINGMEMORY,MEASUREDWITHA3DVIDEOGAME,CORRELATESWITH

STANDARDWORKINGMEMORYTASKS,FLUIDREASONINGANDSPATIALABILITY,BUTNOTWITHVERBALABILITY

Prof.MariangelesQuiroga1,Mr.JavierDelaFuente2,Dr.JesusPrivado1,Dr.JuanRamos-Cejudo1,

Dr.F.J.Román2andProf.RobertoColom2

1UniversidadComplutensedeMadrid,[email protected]ónomadeMadrid,

Introduction:Thisstudypresentsthefirst3Dvideogamedesignedfromtheoutsettotapspatialworkingmemory:ForgottenDepths.Thevideogame(a)followstheprinciplesofgamedesign(Adams,2013)and(b)can be used for research and applied purposes. It consists of 12 level-maps of increased cognitivecomplexity (short-term storage capacity) that the participant has to rememberwhile trying to exit fromeachlevel(processingrequirement).ThegameincludesaModfeaturesothatresearcherscancreatetheirownmaps.Itspsychometricpropertieshavebeentestedintwoindependentstudies.Measuresofstandardworking memory were considered in the first study, whereas in the second study fluid reasoning (Gf),spatialability (Gv),andverbalability (Gc)werealsomeasuredfortesting itsconvergentanddiscriminantvalidity. We expect substantial correlations with standard working memory measures, Gf, and Gv(convergentvalidity),butnotwithGc(discriminantvalidity).

First study: The video gamewasplayed in a computer room, alongwith two standardworkingmemorytasks(DotMatrixandCorsiBlock).Onehundredandelevenundergraduatesparticipatedinthestudy(87women).Mean agewas 21.01 (SD = 4.65, range from 18 to 49). No sex differences in agewere found(t(109)= -.53;p= .595).Reliability indices for thescoresobtained in thevideo-gamewereexcellent (α=0.92,=0.93),andconvergentvalidityevidencewithDotMatrixandCorsiBlockscoresweresubstantial(r=0.44** and r = 0.41** respectively). Latent relationship between video game and standard workingmemory measures were computed trough structural equation modeling, obtaining a remarkablecorrelation(r=0.70,p<.01).

Secondstudy:Thevideogamewasplayedinacomputerroom,alongwithstandardworkingmemory(DotMatrix and Corsi Block), and intelligence tests (Advanced Progressive Matrices, Verbal Test –PrimaryMental Abilities- and Spatial Reasoning –Differential Aptitudes Test). Ninety-three undergraduatesparticipated in the study (81 women). Mean age was 20.35 (SD = 3.18, range from 18 to 44). No sexdifferences inagewerefound(t (109)=-1.58;p= .123).Reliability indices forthescoresobtained inthevideo-gamescorewereexcellent(α=0.87).Latentrelationshipbetweenvideogameandstandardworkingmemory measures were computed trough structural equation modeling, and their correlation wassubstantial (r = 0.60, p < .01). Furthermore, video game performance (a) still correlates with workingmemorywhen fluid reasoning (Gf) and spatial ability (Gv) is controlled for (r = 0.30**) and (b) doesn’tcorrelatewithverbalability(Gc)(r=0.10).

Discussion:ForgottenDepthscanbeusedformeasuringindividualdifferencesinspatialworking-memory.Performancedifferences in thevideogamecorrelateswithworkingmemory,GfandGvbutnotwithGc,whichisconsistentwiththepredictionregardingitsconvergentanddiscriminantvalidity.

References:Adams,E.(2013).Fundamentalsofgamedesign.NewRidersPublishing.


32

ECONOMICPERFORMANCE,STRATEGICBEHAVIORANDINTELLIGENCE

Prof.AldoRustichini11UniversityofMinnesota,[email protected]

Thetalkwill survey recent results linking theanalysisofeconomicperformance (for instance, income)ofindividualsandthestrategicbehavior(thatis,thebehaviorininteractivesituations),andtheirIntelligenceanditsheritability.

Inour studyofeconomicperformance,wesetup theanalysis in a standardeconomicmodelofoptimalparentalinvestmentandintergenerationalmobility,extendedtoincludeafullyspecifiedgeneticanalysisofskilltransmission,andshowthatthemodel'spredictionsonmobilitydiffersubstantiallyfromthoseofthestandardmodel.Forinstance,thecoefficientofintergenerationalincomeelasticitymaybelarger,andmaydifferacrosscountriesbecausethedistributionof thegenotype isdifferent,completely independentlyofany difference in institution, technology or preferences. We then study how much of the educationalachievement is explained by the Polygenic Score (PGS) for education, thus estimating howmuch of thevariance of education can be explained by genetic factors alone.We find a substantial effect of PGS onperformanceinschool,yearsofeducationandcollege.FinallywestudythechannelsbetweenPGSandtheeducationalachievement,distinguishinghowmuch isdue tocognitiveskillsand topersonality traits.Weshow that theeffectof PGS is substantially strongeron Intelligence thanonother traits, likeConstraint,which seem natural explanatory factors of educational success. For educational achievement, bothcognitive and non-cognitive skills are important, although the larger fraction of success is channeled byIntelligence.

The analysis of the effect of Intelligence on strategic behavior is experimental: we study the level ofcooperationinrepeatedgames(withfourpossiblestagegames)whereplayersareseparatedintwogroupsofhigherandlowerIntelligence.Thefractionofcooperationissimilarintheearlystagesofthesession,butdiverges significantly between the two groups, with the high Intelligence reaching full cooperation, andlower Intelligence defecting. The difference in behavior is not due to an unconditional propensity tocooperateofhigherIntelligenceindividual:withlowerdiscountrates,wherecooperationisnotprofitable,cooperation fails also in the high Intelligence group. The analysis of the data provides a completeexplanationofthedifference.


33

INTELLIGENCEANDHAPPINESSSTABILITY:EVIDENCEFROMRUSSIA

Dr.RaufxonSalaxodjaevandMs.NargizaIbragimova11ERGOResearchandConsultancy,[email protected]

Over the past decade there has been growing interest in the association between intelligence andhappiness.Withinthisliterature,anewstrandhasemergedthatinvestigatesthelinkbetweenintelligenceandhappinessstabilityorvariability.Thisstudyexploresthe linkbetween intelligence(IQ)andhappinessstability,measuredby the standarddeviationof life satisfaction.Data for this studywas taken from theRLMS,aseriesofnationallyrepresentativesurveysdesignedtomonitortheeffectsofRussianreformsonthehealthandeconomicwelfareofhouseholdsandindividualsintheRussianFederation.ApanelnatureofRLMS individualdata setwasusedcovering20waves fromtheyearof1994 to2015.Thesamplecoversmore than71,000observationsofwhicharound27,000aremenand45,000arewomen.UsingOrdinaryLeastSquaredRegressionanalysis,wefindthatmoreintelligentcohortsaremorestableintheirhappiness.Thisassociationishighlystatisticallysignificant(β=-0.004,ρ<0.01)evenwhenwecontrolforalargesetofcontrolvariables.Theresultsremainrobustunderaseriesofrobustnesstests.


34

DETECTINGTHESCARR-ROWEEFFECTUSINGPOLYGENICSCORESINTWOLARGEUSSAMPLES

Mr.MatthewSarraf1,Dr.MichaelWoodleyofMenie2,Dr.JonatanPallesen3andDr.CurtisDunkel41UniversityofRochester,[email protected] 2VrijeUniversiteitBrussel3Independentresearcher4WesternIllinoisUniversity

Polygenic scores by virtue of their phenotypic correlations imperfectly but very directly index traitheritability,with the best polygenic scores derived for cognitivemeasures accounting for approximately10%ofthevarianceinthesemeasures.Byexamininginteractionsbetweenpolygenicscoresandcognitivemeasuresasa functionofchildhoodSES, it ispossible touse these todetectScarr-RoweEffects, i.e. thereductionofIQ-heritabilitywithlowerchildhoodSES.

Two large genotyped samples of the US population were tested for Scarr-Rowe Effects, the WisconsinLongitudinalStudy(WLS)andtheHealthandRetirementStudy(HRS),theformersamplingthepopulationof Wisconsin born in the 1930s and 40s, and the latter sampling the US population at retirement,specifically individuals born from the 1910s to 60s. A measure of childhood SES, along with individualpolygenic scores (EA3) and cognitive ability measures were collected from 6256 individuals inWLS and5275 individuals inHRS. Interactionswereestimatedusing theContinuousParameter EstimationModel,whichallowsfordirectestimationtheeffectofonevariable (childhoodSES)onthechange incovarianceamongtwoothers (cognitiveabilityandEA3).Positivemodelβvalues indicatethepresenceof theScarr-RoweEffect, as increasing childhoodSESpredicts increasing covarianceamongEA3and cognitiveability.ThiscanbethoughtofintermsofincreasinggeneticexpressivityofthepolygenicscorerelatedtocognitiveabilitywithhigherchildhoodSES.

Theeffect,whilesmallinmagnitude,waspresentintheWLS(β=.08,p=4.71x10−10),andwasrobusttonormalizationoftheSESparameter,couldberecoveredusinga‘conventional’two-wayinteractionmodel,anddidnot yield sexdifferences (theeffectswerepresentand significant forboth sexes,however). Theeffect was replicated in HRS (β= .05,p= .002,N=5275), was present in both sexes, and was robust toremoving‘old’participants(thoseaged>55).ThereplicabilityoftheEffectacrossthesetwodatabasesalsoindicates substantial robustness, as they could be recovered using different operationalizations of bothchildhoodSESandcognitiveability.

The use of polygenic scores in robustly detecting the Scarr-Rowe Effect yields a significant potentialincrementofpower for studies investigating thiseffect,whichare frequentlyhandicappedby theuseofsmaller samples of twins. These results also remove ambiguities concerning the direction of causation.Here, lower quality childhood environments are directly suppressing the expressivity of genetic variantspredictiveofcognitiveability–preventingsuch individuals fromattaining their fullgeneticpotentialwellintolatelife.


35

HOWINTELLIGENTANDEDUCATEDARETHEAMERICANELITE?ASTUDYOF26,000U.S.LEADERSACROSS30SECTORS

Dr.JonathanWai1,Mr.StephenAnderson2,Ms.KajaPerina3,Dr.FrankWorrell4and

Dr.ChristopherChabris51UniversityofArkansas,[email protected] 2PennsylvaniaStateUniversity3PsychologyToday4UniversityofCalifornia-Berkeley5GeisingerHealthSystem

Thisstudyinvestigatesthecognitiveabilitiesandhighereducationbackgroundsof26,198membersoftheAmerican elite across 30 sectors. Broad groups include business leaders (e.g., Fortune 500 CEOs, 4-stargenerals,startupfoundersandCEOs),scientificachievers(e.g.,NobelPrizewinners,NationalAcademyofScience members), influential journalists and literary prize winners, (e.g., New York Times editors andwriters, Pulitzer Prize winners), politicians (e.g., House members and Senators, Presidents and VicePresidents),andelite faculty (e.g.,HarvardUniversityprofessors).We investigatedschoolattendance foreitherundergraduateorgraduateeducation:HarvardUniversity,oneofthe8IvyLeagueschools,andasetof34most selectiveundergraduateandgraduate schoolsbasedonaverageadmissions test scores (e.g.,SAT,ACT,GMAT,LSAT).Attendingoneofthese34selective institutions indicated,basedonstandardizedtestscores,thattheindividualswerelikelyinthetop1%ofcognitiveabilitygiventestssuchastheSATorACTareproxymeasuresofgeneralintelligence(g).

Top1%incognitiveabilityrepresentation.Acrossall30groups,roughlyhalf(54.2%)ofindividualsattendedoneofthe34highlyselectiveschoolswhosegraduatesaremostlyinthetop1%ofcognitiveability,rangingfrom11.2%—25.9% (Four StarGenerals, Four StarAdmirals,Housemembers)up through78.9%—80.9%(Forbesmostpowerfulmen,HarvardFaculty,AmericanPhilosophicalSociety).ExamplesofgroupsthatfellintothemiddlerangesincludedFortune500CEOs(41.9%),WallStreetJournaleditorsandwriters(50.8%),National Academy ofMedicinemembers (60.5%), and National Academy of Sciencesmembers (70.5%).Giventhattop1%inabilitypeoplewouldbeexpected,onbaserate,tomakeup1%ofthesepopulations,thismeanstop1%inabilitypeoplewereoverrepresentedintheAmericaneliteoverallbyafactorofabout54,rangingfrom11to70timesfordifferentgroups,suggestingwidecognitivevariationevenacrossthesehighlyselectoccupationalandleadershipgroups.

Higher education representation. Of all American adults in the general population, about 32.5% havereceived a bachelor’s degree. Within this 32.5%, about 1.9% have a degree from one of the 34 highlyselectiveschoolsinourstudy,about0.6%haveadegreefromoneofthe8IvyLeagueschools,andabout0.2%haveadegree fromHarvardUniversity.This suggests that thepercentagesof individuals ineachofthesegroupsofleadersofU.S.societyarequitehighrelativetopopulationbaserates.Forexample,if54%ofthe26,198individualsinoursampleattendedoneofthe34highlyselectiveschoolsandthebaserateisabout 1.9%, this suggests a factor of overrepresentation of roughly 28 times base rate expectations(calculatedas54/1.9).Acrossallgroups,roughly36%attendedanIvyLeagueschool,suggestingafactorofoverrepresentation of roughly 60 times (36/0.6). Across all groups, roughly 16% attended HarvardUniversity,suggestingafactorofoverrepresentationofroughly80times(16/0.2).


36

INTELLIGENCE'SPLACEINTHEPSYCHOLOGYCURRICULUM

Dr.RussellWarneandMr.JaredBurton11UtahValleyUniversity,[email protected]

Prior research has shown that psychology textbooks' treatment of intelligence is often erroneous andshallow(Pesta,McDaniel,Pozananski,&DeGroot,2015;Warne,Astle,&Hill,2018).However,itisunclearhowmuchthebroadcollege-levelpsychologycurriculumcontributestothewayintelligenceistaught.Todiscernhowintelligencefits intothebroaderpostsecondarycurriculum,wecollectedcoursedescriptionsandcataloginformationfrom303Americancollegesanduniversities.Theseinstitutionsweredrawnfromalist of (a) leadinguniversities, (b) leading liberal arts colleges, and (c) leading regional universities in theUnitedStates.

Wefoundthatcollegecoursesdedicatedtomainstreamintelligencesciencearerare,withonly12coursesincluding g-based theories of intelligence in their descriptions.We also found 195 courses dedicated toemotional intelligence (often offered in business or management departments)and 47 dedicated tomultiple intelligences (frequently found in education departments). On the other hand, 200 courses onpsychological testing included intelligence in their description, and for this reasonwebelieve that somepsychology students get exposed to at least information about intelligence. However, the near-totalabsenceofclassesdedicatedtointelligence(only4at303universities)showsthatmanyimportantaspectsofintelligenceresearcharemissingfromthepostsecondarycurriculum.Weseethenear-totalabdicationofteaching about intelligence as a possible contribution tothe proliferation of classes on emotionalintelligenceandmultipleintelligencesinnon-psychologydepartmentsandthewidespreadmisconceptionsaboutintelligenceamongthegeneralpublic.


37

THEFIRSTLONGITUDINALINVESTIGATIONOFCOGNITIVEABILITYINADULTADOPTIVEFAMILIES

Ms.EmilyWilloughby1,Prof.MattMcGueandDr.JamesLee21UniversityofMinnesotaTwinCities,[email protected] 2UniversityofMinnesota

Does theeffectof rearingenvironmenton childhood IQpersistwell intoadulthood?Despitedecadesofgenetically-informed studies of IQ, this question has yet to be answered to satisfaction. Early adoptionstudiesgenerallyfoundthattheIQsofyoungchildrencorrelatedwiththoseoftheiradoptedparents,butthat inadolescentsthisrelationshipwas lessstableandoftendisappearedaltogether (Scarr&Weinberg,1983). Newer research, by contrast, has suggested a substantial role of the family environment thatamounts to about 4.4 IQ points (SE = 0.75) in 18–20-year-old Swedish adoptees (Kendler, Turkheimer,Ohlsson, Sundquist, & Sundquist, 2015). It is unclear, however, whether this relationship would remainrobustaftertheoffspringleavethefamilyhomeandcarveouttheirownlives,orwhetheritwouldshowa“fadeout effect” of the parental environment, consistent with observations of many phenotypes(Poldermanetal.,2015).

AttheMinnesotaCenterforTwinandFamilyResearch(MCTFR),theSiblingInteractionandBehaviorStudy(SIBS) has become the longest-running longitudinal study of adopted siblings in the world. This uniquesampleof409adoptiveand208non-adoptivefamiliesincludesatotalof1,164parentsand1,232offspring,withanaverageageofplacementinadoptivehomesof4.7months(SD=3.4)McGueetal.,2007).Thesefamilieswerefirstassessedonabilityandachievementmeasuresbeginningin1998,whentheaverageageoftheoffspringwas14.9(SD=1.9).Asoffall2018,thesenowadultindividualsandtheirparentsareintheprocessofbeingassessedforthefourthtime.Atotalof574adoptedandbiologicaloffspringbetweentheagesof26and40havebeeninterviewedsofaronmeasuresofachievementandability,withameanageof32.2(SD=2.6).

Attimeofintake,IQscoresofbiologicalchildren(N=396;meanage=14.8)correlatedwiththeirparents’scoresatr=.365(p<.001),andIQscoresofadoptedchildren(N=519;meanage=14.9)correlatedwithparental scores at r = .09 (p= .04). At this time,we have assessed cognitive ability in about half of oursampleofparentsandoffspringusingtheICAR16sampletest,withcomprehensiveverbalIQassessmentontheway.ICARscoresofbiologicaloffspring(N=165;meanage=31.75)correlatewiththeirparents’scoresatr=.205(p=.008),andadoptedoffspring(N=171,meanage=32.23)withtheirparents’scoresatr=–.05 (p= .497). This association holds similarly for related cognitive phenotypes, such as GPA andeducationalattainment.Takentogether,theseincipientcontrastingpatternsofassociationareconsistentwith the diminishing effect of parental environment on measure of offspring cognitive ability well intoadulthood.


38

WHYISNARCISSISMTHESTRONGESTPREDICTOROFSUBJECTIVEINTELLIGENCE?

Dr.MarcinZajenkowski1

1FacultyofPsychology,UniversityofWarsaw,[email protected]

Intelligence is regarded as highly agentic and socially desirable characteristic. Grandiose narcissists aretypicallypreoccupiedwith the topicof intelligence,however theassociationbetween theseconstructs isstillpoorlyinvestigated.Inthecurrentresearchtherelationshipbetweengrandiosenarcissism,objectivelyassessedintelligenceandbeliefsaboutintelligencewereexamined.InStudies1(n=233)and2(n=255)itwas found that narcissism was the strongest predictor of subjectively assessed intelligence (SAI) andexplainedmorevariance(12%-18%)inSAIthanobjectiveintelligence(measuredwithstandardtestssuchas Raven’s and Cattell’s) and basic personality traits (Big Five). Additionally, narcissism did not correlatewithobjectiveintelligence.

Subsequently,wetestedhownarcissistsperceiveintelligence.InStudy3(n=362)weaskedparticipantstojudgehowadvantageousintelligenceisforseveral lifedomains,e.g. lifesuccess, jobperformance,schoolachievements, income, social status, relationships, popularity andphysical attractiveness.We found thatnarcissistsbelievedthatintelligenceisbeneficialmainlyforsocialinteractionsandphysicalattractiveness.Finally,inStudy4(n=222)weexaminedhownarcissistsexperiencethesituationoftakinganIQtest.Theresultsindicatedthatnarcissistsdidnotenjoysolvingthecognitiveabilitytestandreportedlowmotivationincognitiveperformance.

Thecurrentstudies indicatethatabelief in their intellectualsuperiority isan importantbuildingblockofgrandiose narcissists’ self-concept. Furthermore, narcissists feel that high intelligence is a resource thatbuys people benefits in social interactions domains and they feel that they possess that resource.Moreover,theydonotshowgenuineengagementincognitiveactivities,butconsider intelligencetobeasourceofbenefitsinmultipledomains,includinginterpersonalrelationships.


39

POSTERSINALPHABETICALORDERBYAUTHORLASTNAME

UTILITYOFTHECATTELL-HORN-CARROLLTHEORYOFCOGNITIVEABILITIESTOINTERPRETPERFORMANCEONSPEECH-LANGUAGETESTS

Dr.VincentAlfonso1

1GonzagaUniversity,[email protected]

The Cattell-Horn-Carroll (CHC) Theory of Cognitive Abilities (CHC Theory) has been well researched andvalidatedasacomprehensivemodelofabilitiesandprocesses(Flanagan,Ortiz,&Alfonso,2013;McGrew,2009;Schneider&McGrew,2018). Itprovidesstructuretoclarifyandcategorizethevariousabilitiesandprocessesthatimpactfunctioningwithinandoutsidetheclassroom.WithintheCHCframework,thereare17 second stratum or broad ability constructs including fluid reasoning, comprehension-knowledge,workingmemory,learningefficiency,andauditoryprocessingtonameafew.Inaddition,therearedozensofsub-abilitiesornarrowabilitiesthatarespecificmeasuresofthebroadabilities.Someofthesenarrowabilities are induction, lexical knowledge, auditory short-term storage, associativememory and phoneticcoding.

CHC Theory has been applied to cognitive tests for decades. For example, the Woodcock-Johnson IV(Schrank,McGrew,&Mather,2014)weredesignedwithCHCasitsblueprintortheoreticalframework.Thevenerable Wechsler Scales, although a-theoretical, have been revised to include more and more CHCabilities and processes (Flanagan & Alfonso, 2017). Flanagan, Ortiz, and Alfonso (2017) have used CHCTheorytoclassifymorethan1,000subtestsfoundonvariouscognitive,achievement,memory,andspeech-languagebatteries toaid in testperformance interpretation.Although several speech-languagebatterieshavebeen revisedornewoneshavebeen released, their subtestshavenotbeenclassifiedaccording toCHCTheory.

Theauthorsexaminedthesubtestsofseveralspeech-languagebatteriesandusingtheirexpertknowledgeandanextantdatabase,classifiedeachsubtestaccordingtoitsCHCbroadandnarrowabilities.Amajorityof thesubtestsappear tomeasureabilitiesandprocesses thatmakesensegiventhat theyare language-based subtests. For example, many subtests measure lexical knowledge, a comprehension-knowledgenarrowabilityandphoneticcoding,anauditoryprocessingnarrowability.However,othersubtestsseemtomeasureCHCbroadandnarrowabilitiesthatarenotintuitivesuchasworkingmemorycapacity,auditoryshort-term storage, retrieval fluency, and ideational fluency. Knowledge of the underlying abilitiesmeasured by speech-language subtests can aid in test performance interpretation and assist in makingsound instructionaland interventionrecommendations.TheutilityofCHCTheory inthe interpretationofperformanceonspeech-languagetestsisexplainedandfutureresearchissuggested.

ISIR2019PosterPresentations

40

ALONGITUDINALANALYSISOFTHEEFFECTSOFCOGNITIVESKILLSANDNON-COGNITIVESKILLSONCOUNTERPRODUCTIVEANDCITIZENSHIPWORKPLACEBEHAVIORS:

ATESTOFROBUSTNESSINTWOSAMPLES

Ms.EliseAnderson,Prof.MattMcGueandDr.PaulSackett1

1UniversityofMinnesota,[email protected]

Recently,therehasbeenanincreasinginterestinhowadolescentandchildhoodfactorsimpactlateradultbehaviorsandoutcomes.Heckman(2012)notedtheimportanceofbothcognitiveandnon-cognitiveskillson later educational and occupational success, while Vergunst et al. (2019) described a negativerelationship between childhood disruptive behavior and later income level. However, there has beenlimited research on the impact of childhood influences on adult non-task job performance. Jobperformance,an indicatorofoccupationalsuccess,hastypicallybeenbrokendownintothreecategories:taskperformance,citizenshipbehaviors(positivelyaddingtothesocialorpsychologicalenvironment)andcounterproductivebehaviors(actionsthatharmthecompanyanditsmembers)(Rotundo&Sackett,2002).Generalcognitiveabilityhasbeenshowntobearobustpredictoroftaskperformance(Schmidt&Hunter,2004),althoughlessisknownabouttheextenttowhichcognitiveandnon-cognitiveskillscontributetotheothertwocomponentsofjobperformance,particularlyfromalongitudinalstandpoint.

Buildingonthepreliminaryfindingspresentedat ISIR lastyear(Andersonetal.,2018,oralpresentation),this study employs two discrete longitudinal datasets spanning adolescence through middle age, todetermine whether cognitive and non-cognitive factors contribute differentially to the behavioralcomponentsofjobperformance,andwhetherthefindingsarerobustacrossthetwosamples.

Longitudinal research conducted by Roberts and colleagues (2007) suggests that both personality andintelligenceatage17correlatewithcounterproductiveworkbehaviorslaterinlife,thoughtheirfindingsinregardstointelligence(apositivecorrelation)iscountertothatofotherstudies(e.g.,Dilchertetal.,2007).The current research will add to the existing body of literature on intelligence and counterproductiveworkplacebehaviors,andextendthisresearchareabyexaminingcitizenshipbehaviorsand“softskills”(asummarymeasure of the personality, school related behavior, and reversed externalizing behavior; seeMcGueet al., 2017) aswell. Sample1will includeover 400 individuals from theMinnesota Twin FamilyStudy (MTFS) who were initially assessed at the target age 17 with the most recent follow-up whenparticipantswere in their early 40s. A second sample over 500 adopted and non-adopted offspringwillallow us to examine the robustness of the effects found in sample one. These individuals were initiallyassessedinadolescence,withthemostrecentfollow-upwhiletheparticipantswereinyoungadulthood.

Preliminary analysis from the first 307 individuals identified soft skills as significantly and moderatelyrelatedtobothcounterproductiveworkbehaviorsandorganizationalcitizenshipbehaviors.Similartothefindings presented by Roberts et al. (2007), general cognitive ability was significantly and positivelyassociatedwith counterproductiveworkplacebehavior.No relationshipwas identifiedbetween generalcognitiveabilityandcitizenshipbehavior.


41

THEGINGRAMMAR:GRAMMARUSAGEISANINDICATOROFGENERALCOGNITIVEABILITY

Prof.AprilBleske-RechekandMs.KatiePaulich1

1UniversityofWisconsin-EauClaire,[email protected]

Grammarusagereferstohowwords,phrases,andpunctuationarecombinedtocreateandcommunicatemeaning. Here,we describe three findings that, taken together, imply grammar usage is an indicator ofgeneralcognitiveability.First,usinganexperimentaldesigninwhichwemanipulatedtherateofgrammarusageerrorsinajobapplicationcoverletter,wefoundthatapplicants(whowouldotherwisebeviewedasstrong candidates) were judged as less capable and less skilled if their letter contained grammar usageerrors. Second, using a longitudinal pre-post design with ten weeks of distributed practice on typicalgrammar usage errors, we found that students’ ability to identify and rectify common grammar usageerrors is highly consistent over time andrelatively unresponsive to intervention. Third, in that samelongitudinaldesignwefoundthatyoungadults’abilityto identifyandcorrectgrammarusageerrorsonafive-minuteassessmentishighlypredictiveoftheirperformanceinacoursedesignednotforwritingbutforanalyticalandstatisticalreasoning.Giventhestrongverbalabilitycomponent inmanymeasuresofg,wearenotnecessarilysurprisedthatgrammarusageshowssimilarcharacteristicstothoseseenwithgeneralcognitiveability.However,wemayhaveserendipitouslyuncoveredafive-minutemeasureofg.


42

COGNITIVEABILITIES,DIVERGENTTHINKING,ANDTHEASPECTSOFOPENNESS

Dr.AshleyBrown,Dr.LindaHouser-MarkoandDr.DavidSchroeder1

1JohnsonO'ConnorResearchFoundation,[email protected]

Among the Big Five personality factors, Openness to Experience is notable for its significant, positiverelationship to cognitive ability (Costa & McCrae, 1992). But the devil is ever in the details: DeYoung,Peterson,andHiggins(2005)notedthatsomefacetsofNEO-PI-ROpenness(Fantasy,Aesthetics,Feelings)were correlated with crystallized ability whereas others (Ideas) were correlated with fluid ability andworking memory. Later, DeYoung, Quilty, and Peterson (2007) used these results to buttress theirargument in favorof thedistinctnessof their two“aspects”ofOpennesstoExperience;namely, IntellectandOpenness(hereafter,theterm“Openness”willbeusedtoreferonlytotheaspect,notthebroadtrait).We sought to replicate and extend these findings using an age-diverse sample of 331 individuals withscoresontheJohnsonO’ConnorAbilityBattery(JOAB)andtheSyntheticAperturePersonalityAssessment(SAPA;Condon,2017).

BecausetheJOABconsistsofmanydifferenttests,wewonderedwhichabilities,ifany,wouldberelatedtoOpennessratherthanor inadditiontoIntellect.WehypothesizedthatallcognitiveabilitytestswouldbepositivelyrelatedtoIntellect;wealsopredictedthatameasureofdivergentthinking(DT;operationalizedusing scores on the JOAB Foresight test)would be positively related toOpenness.We set about testingthese predictions using scores on DT as well as on a subsample of nine JOAB tests (N = 43,540) whichrecreated the four-factormodelof abilitydiscussed inHaieret al. (2009): SpeedofReasoning,hereafter“Speed;” Spatial; Numerical; andMemory; plus a single-test “factor” of Verbal ability (the JOAB EnglishVocabularytest).Amongthe331 individualswithscoresonbothSAPAandJOAB,all fivecognitiveabilityfactorsaswellasDTweresignificantlypositivelycorrelatedwithIntellect,whereasonlyDTwassignificantlypositivelycorrelatedwithOpenness.

Althoughthesefindingswerelargely in linewithexpectation,wewereintriguedtonotenotonlythatDTwasalsorelatedtoIntellect,butalsothatOpennesswasmarginallypositivelyrelatedtoseveralcognitiveabilities(namely,SpeedandSpatial).SimultaneousregressionanalysesshowedthatamodelinwhichthefivecognitiveabilitiesandDTpredictedIntellectwassignificant(adjustedR2=.296),althoughonlythreeofthepredictorshadsignificantcoefficients(Speed,ß=.164;Verbal,ß=.317;DT,ß=.142).AmodelinwhichthefiveabilitiesandDTpredictedOpennesswasalsosignificant(adjustedR2=.083)and,ashypothesized,onlytheregressioncoefficientforDTachievedsignificance(ß=.297).Notablemarginallysignificantresults,however,wereobtainedforSpatial(ß=.124,p=.052)and,unexpectedly,Numerical(ß=-.113,p=.086)abilities.Thelatterismostlikelyasuppressoreffect;thatis,wemayinferthatthepartofNumericalabilitythatisindependentofSpeed,Spatial,Verbal,Memory,andDivergentThinkingabilityisnegativelyrelatedtotheOpennessaspect.


43

CROSS-BATTERYRELATIONSOFCHILDREN’SCOGNITIVE

ABILITIESANDREADINGANDWRITING

Dr.JacquelineCaemmerer1,Dr.TimothyKeith2andMs.EuniceBlemahdoo1

1HowardUniversity,[email protected]

Research that incorporatesmultiple intelligenceandachievement tests,knownascross-batteryanalyses,can better address questions about the broader influences of children’s cognitive abilities on theirachievement.Suchcross-batteryresearchwillextendpsychologists’understandingofhowintelligenceandachievement relate not just at the test-level, but at the broader construct level. Previous cross-batterystudieswereoftenlimitedtotwotestsandweremostlybasedontheWoodcock-Johnsontests(McGrew&Wendling,2010).Thus,findingsarenotnecessarilygeneralizableacrossothertestsandindeeddifferencesbasedon the testsusedhavebeen found.Whetherornot intelligence testsmeasure thesameabilities,despite differences across tasks, and whether or not cognitive-achievement relations are reproducibleacrossseveralbatteries,arequestionswithboththeoreticalandclinicalimplications.

Six intelligence tests (KABC-II,WJ III,WISC-III,WISC-IV,WISC-V, andDAS-II) and three achievement tests(KTEA-II,WIAT-II,WIAT-III) were included in a cross-battery cognitive-achievement analysis. Participantswere3,930youthaged6to16drawnfrom7nationallyrepresentativestandardizationandlinkingsamplesprovidedbyPearson.Thesamplesizesrangedfrom181to2,520persample.Sampleswerelinkedtoeachother by tests they shared in common. In order to simultaneously analyze several tests a plannedmissingness methodological approach was used, which addresses concerns regarding bias and power(Enders, 2010;Grahamet al., 2006). Also,modern structural equationmodeling techniques for handlingmissing data, maximum likelihood estimation, allow researchers to analyze the data without discardingincompletecases.

Ahigher-orderCattell-Horn-Carrollmodelwith6broadabilities (Gc,Gf,Gv,Gsm,Gs,andGlr)andgwastested;each latentbroadability variablewas indicatedby7–12measuredvariables, the subtests. Thiscomprehensivecross-batterycognitivemodelwasusedtotest2models,abroadwritingandbasicreadingmodel.Abroadwritinglatentvariablewasindicatedby6subtestswhichmeasuredstudents’spellingandwritten expression. A basic reading latent variable was indicated by 4 subtests which measured wordrecognitionandpseudoworddecoding.The fitof the2modelswasadequate togood.Gc,Glr,andGsmwerestatisticallysignificantlyrelatedtobasicreading(β= .40, .25,and.22respectively).AndGc,Glr,Gs,andGsmwerestatisticallysignificantlyrelatedtobroadwriting(β= .29, .29, .23, .22respectively).Whentestedseparately,ghadlargeeffectsonbothbasicreadingandbroadwriting(β=.75and.82respectively).


44

PUBLICPERCEPTIONANDPRESSCOVERAGEOFINTELLIGENCE

RESEARCHINTHEFRENCHREPUBLIC

Mr.JulienDelhez1andMr.RobertChapman2

1Georg-August-UniversitaetGoettingen,[email protected],UniversityofLondon

The French seem tobe largelyunderrepresentedamong intelligence researchers (Becker,Rindermann&Coyle, 2018: fig. 5). We hypothesize that this might be due to poor knowledge among the generalpopulation,ortoinsufficientorunrepresentativepresscoverageoftoday’sintelligenceresearch.Toverifythese hypotheses, we use the data from the French participants (N = 54) of the International GeneticsLiteracyandAttitudesSurvey(iGLAS),aninternationalsurveyevaluatingpublicknowledgeofgeneticsandtheheritabilityof intelligenceandotherbehaviouraltraits.Wealsoassessthethemescovered inarticlesonhumanintelligencepublishedbetween1999and2019inFrenchnation-widegeneralistnewspapers(N=103), local generalist newspapers (N = 15), and scientific magazines (N = 76), and we discuss in whatmeasurethesethemesreflectthecurrentevolutionofintelligenceresearch.FortheFrenchparticipantsofiGLAS,therateofcorrectanswersonGeneticKnowledgeitemsis67%,apercentagecomparabletothatofotherWesternEuropeancountries.ThetwomainthemescoveredintheFrenchpressappeartobe,ontheonehand,environmentalfactorspresentedasinfluencingIQ(33%innation-widenewspapers,0%inlocalnewspapers,19.74%inscientificmagazines),whetherpositive(e.g.breastfeedingorfishconsumption)ornegative(e.g.pollution);ontheotherhand,high-IQchildrenandtheirsocialdifficulties(18.4%innation-widenewspapers,26.7%inlocalnewspapers,5.3%inscientificmagazines).Presscoveragewasscarcerforareasofintelligenceresearchthatarenonethelessveryactivenowadays,e.g.theneurobiologicalbasisofintelligence (1.9% in nation-wide newspapers, 0% in local newspapers, 2.6% in scientificmagazines) andgenome-wide association studies related to intelligence (1% in nation-wide newspapers, 0% in localnewspapers, 0% in scientific magazines). The number of iGLAS participants and collected newspaperarticlesisstillincreasingatthetimeofsubmissionofthisabstractandisexpectedtobesubstantiallyhigherinJuly.


45

RELATIONSAMONGVERBALABILITY,SELF-COMPETENCE,ANDTASKMOTIVATIONINTHE

TRANSITIONFROMPRIMARYTOSECONDARYSCHOOL.

Ms.AliceDiazandProf.WendyJohnson1

1UniversityofEdinburgh,[email protected]

A meta-analysis of TIMSS and PISA test found negligible sex differences (d=.01-.11) in mathematicsachievementacross87countries (Else-Quest,Hyde,&Linn,2010).However, inameta-analysisof schoolgrades, girls outperformed boys across all subjects, size effects are small but consistent (d=.154-.374)(Voyer&Voyer,2014).

UnderEccles’(1983)model,motivationincludesstudents’taskvalueandself-competence.Someresearchshowsthatgirlstendtoseethemselvesmorecompetentinlanguagesthanphysicsormathematics,whilstthe opposite is true for boys, even when there are no achievement differences. In addition, self-competence and task value tend to decline with agein a sex and subject-specific fashion during thetransition from primary to secondary school, though the extents vary across countries.Moreover, self-competenceseemstoberelatedtohighability forboysbutnotsomuchforgirls.Weexploredrelationsamong these observations inWaves 5 and 6 (11 and 14 years old, n=13,000+) of theMillennium StudyCohort,anongoing longitudinalstudycarriedout intheUKsince2000,usingregression.Ourhypotheseswere:

• Sexdifferencesfavoringgirlsinoverallacademicachievementandineachspecificsubject.

• Sexdifferencesinself-competenceandtask-value(relativelyhigherinlanguagethanmathingirlsandtheoppositeinboys)inbothwaves.

• Verbal ability ismost strongly associatedwith academic achievement, followedby self-competenceandtaskvalueatbothwaves.

• Self-competenceandtaskvaluedeclinebetweenages11and14 issex-andsubject-specific: slowerforlanguagethanmathematicsingirls,andtheoppositeforboys.

• Verbal ability moderates relations between sex and self-competence and task value so that high-abilitychildrenexperience relatively lowerdeclines; thiseffect isnotas strong inhigher-abilitygirlsthaninhigher-abilityboys.

Wewilldiscusshowresultsmayshedlightonsourcesofthecurrenttendencyforgirlstomaintainhigheracademicachievementthanboys.


46

RELATIONSBETWEENGENERALINTELLIGENCE,EMPATHIZING,SYSTEMIZING,ANDTHEORYOFMIND

Ms.KarrieElpersandDr.ThomasCoyle1

1TheUniversityofTexasatSanAntonio,[email protected]

Background:Previousresearchsuggeststhatmeasuresoftheoryofmind,empathizing,andsystemizingareunrelated to general intelligence. However, more recent research suggests that theory of mind ismoderatelyrelatedtogeneral intelligence.Thepresentstudyaimedtoreplicatethepreviouslysuggestedrelationship between theory ofmind and general intelligence. Furthermore, it extends the literature byexaminingtherelationshipbetweengeneral intelligenceandempathizingandsystemizing. Inaddition,asallofthesemeasureshavebeenrelatedtocollegemajorchoice(i.e.,STEMversushumanities),thepresentstudy also examined the relationship between general intelligence, theory of mind, empathizing,systemizing,andfeelingstowardsSTEMandhumanitiesmajors.

Method: A general intelligence factor was calculated using the total scores of the 4 sections of theInternational Cognitive Ability Resource (ICAR). Supplemental analyses were conducted using a generalintelligence factor calculated from the subsections of the ACT college admissions test and (to increasesample size) composite ACT scores. Empathizing and systemizing were based on the Empathy QuotientShortForm(EQ)andtheSystemizingQuotientShortForm(SQ).TheoryofmindwasbasedontheReadingthe Mind in the Eyes Revised Version (RMET). Feelings towards STEM and humanities were based onwithin-subjectsdifferencescoresfromLikertscalesonwhichparticipantsrankedtheextenttowhichtheylikeanddislikeSTEMandhumanities.SamplesizewasN=431.Pearson’scorrelationswereconductedforall variables. In addition, an exploratory mediation was conducted to determine if SQ mediates therelationshipbetweengeneralintelligenceandfeelingstowardsSTEMmajorsusingthePROCESSplug-inforSPSS.

Results:Asexpected, theoryofmind,empathizing, and systemizingwereall at leastmodestly related togeneral intelligence (r = .27, -.15, .27 respectively). However, the relationship between empathizing andgeneral intelligence was unexpectedly negative. Also unexpectedly, theory of mind (RMET) was notsignificantly related to empathizing (EQ) (r = -.00). General intelligence and systemizing (SQ) were notsignificantlyrelatedtofeelingstowardshumanities(r=.05,.02respectively),andtheoryofmindwasnotsignificantlyrelatedtofeelingstowardsSTEM(r=-.04).SupplementalanalysesusingACTbasedmeasuresofgeneralintelligencesupporttheresultsofprimaryanalyses.AnexploratorymediationsuggeststhatSQsignificantlymediatestherelationshipbetweengeneralintelligenceandfeelingstowardsSTEMmajors(β=.11, 95% BC CI [.07, .15]). However, this mediation only explains half of the variance shared betweengeneralintelligenceandSTEMdifferencescores.

Discussion:Theresults supporta relationshipbetweengeneral intelligence, theoryofmind,empathizing,and systemizing. However, the negative relationship between empathizing and general intelligencemaysuggestthatempathizingisbeingusedtocompensateforlowgeneralintelligence.Theoryofmindwasnotrelatedtoempathizing,possiblysuggestingthatpeopledonotaccuratelyself-reportempathy.Unexpectedpatterns in the relationships between variables and feelings towards STEMand humanitiesmay suggestthat certain skills that are necessary to be successful in either STEM or humanities are not universallynecessaryinallfields.


47

GENERALINTELLIGENCE(G)ANDTHEORYOFMINDDIFFERENTIALLYPREDICTPERFORMANCEINAMULTIPLAYERGAME:AMULTILEVEL

ANALYSISOFINDIVIDUALANDGROUPLEVELEFFECTSMr.MiguelGonzalez1,Dr.ThomasCoyle1,Ms.KarrieElpers1,Mr.TylerMinnigh1,Dr.JacobFreeman2,

Dr.JacopoBaggio3andDr.DavidPillow11UniversityofTexasatSanAntonio,[email protected],2UtahStateUniversity,3UniversityofCentralFloridaBackground:Thecurrentstudyusedamultiplayergametoexaminetheeffectsofgeneralintelligence(g)andtheoryofmind(ToM)insmall(4people)andlargegroups(8people).Thegamerewardedgroupsthatcoordinatedeffectivelyandadoptedapatient(ratherthanselfish)strategy.MultilevelanalysesexaminedtheeffectsofgandToMatthegroup(meanofgroup)andindividuallevels(afterremovinggroupeffects).AkeyquestionwaswhethertheeffectsofgandToMvariedatthetwolevelsandwithgroupsize.

Predictions were based on the functional intelligences proposition (FIP; Freeman et al., 2016). The FIParguesthatdiversesetsofabilitiespredictperformanceincomplexgamesbetterthanhomogeneoussetsofabilities,providedtheabilitiesare:(a)relativelyindependentand(b)functionallyrelevant(well-suitedtothetask).TheseassumptionsapplytoToMandg,whichsharelimitedvariance.Inaddition,whereasToMfacilities social problem solving (via social inferencing and mind reading), which in turn facilitatescoordinationinmultiplayergames,gtapstheabilitytoadaptquicklytocomplex,dynamicgames(likethegameusedhere).

Method:gwasbasedontheACT,acollegeadmissionstestthatcorrelatesstronglywithIQandgobtainedfromdiversetestssuchastheASVAB(r=.77,Koenigetal.,2008).ToMwasbasedontheReadingtheMindintheEyesTest(RMET),whereparticipantsinferemotionsfromeyes,andtheShortStoryTest(SST),whereparticipants infer characters’ mental states. College students (N=336) played six rounds of the foraginggame,amultiplayerstrategygame.Subjectscollectedtokensingroupsof4or8inthefirstthreerounds,with group size switched in the last three rounds. Tokens (dependent variable) regenerated under acomplexproximityrule,whichrewardedgroupsthatadoptedapatientstrategyandrefrainedfromquicklyoverharvesting.MultilevelanalysesofgandToMexaminedwithineffects(obtainedafterremovinggroupeffects),andbetweeneffects(basedongroupmeans).Effectsareunstandardizedcoefficients.

Results: Initialanalysesoftokenscollapsedacrossgroupsandrounds.Betweeneffectswerepositiveandsignificant for all predictors (3.24, 4.35, 5.58) indicating that higher g/ ToM predicted more tokens. Incontrast,withineffects(ACT,RMET,SST)werenegativeandsignificant(p<.05)forallpredictorsbutRMET(-1.32, -.46, -1.49), indicating that higher g/ToM generally predicted fewer tokens. The sign of effectsreplicated when accounting for group size (4 vs. 8) and round (1-3 vs. 4-6), with negligible differencesbetweengroupsizesandrounds.

Discussion: The results supported the FIP. Between effects were positive, indicating that higher g/ToMpredicted better group performance. In contrast, within effects were negative, indicating that higherg/ToM predicted lower individual performance. The latter finding suggests that individuals with higherg/ToMadoptapatientstrategytooptimizetokencollection,whereasindividualswithlowerg/ToMadoptaselfish strategy that leads to overharvesting. Future research might consider personality traits(agreeableness)thatcontributetostrategicbehaviorinmultiplayergames.


48

INTELLECTUALPERFORMANCEINCHRONOTYPES:EFFECTSOFSLEEPQUALITY

Dr.KonradJankowski1

1UniversityofWarsaw,[email protected]

Existing studies provide conflicting results regarding links between intelligence and the dimension ofmorningness-eveningness.Thisstudyaimedtodeepenunderstandingofassociationsbetweenintellectualperformance and chronotype – an individual characteristic describing functioning during 24 hours day.Evening chronotypes often suffer lowered sleep quality like inadequate sleep length or other sleepproblemsduetosocial jet lagtheyexperience(sleepinginbiologicallynon-optimalhours).Loweredsleepquality, on the other hand, can impact on cognitive performance includingworkingmemory,which hasdetrimentaleffectonperformanceintasksmeasuringfluidintelligence.Thiseffectcanbeimmediateandpresumably best observable during the morning following night with lowered sleep quality. Thus, weexpectedthatsynchronyeffectbetweenchronotypeandtimeofdayonperformance in tasksmeasuringfluidintelligencediminisheswhensleepqualityiscontrolledfor.Totestthishypothesisuniversitystudentscompletedmeasuresofchronotypeandperformedtests for fluid intelligence in themorning,while theirsleep quality a night before testing was assessed by actigraphy and self-reports. The results will bediscussedintermsoftheirrelevanceforsynchronyeffect.


49

INVARIANCEOFAWJIVCOGNITIVEABILITIESMODELFORGIFTED,GIFTED/LEARNINGDISABLED,ANDNON-GIFTEDINDIVIDUALS

Dr.DanikaMaddocksandDr.TimothyKeith1

1TheUniversityofTexasatAustin,[email protected]

Intelligence test results inform diagnosis and support of individuals with exceptional characteristics,including those who are gifted or have a learning disability. Such exceptional populations have uniquecognitive characteristics, however,and few studies have examined whether intelligence testsarecomparable for these groups and the population as a whole.This study compared the structure andmeasurementofintelligenceacrossgifted,gifted/learningdisabled(GLD),andnon-giftedindividuals.

DatacamefromthestandardizationsamplefortheWoodcockJohnson(WJIV)TestsofCognitiveAbilities,OralLanguage,andAchievement.Thesamplecontains6,817individualsages5to93andisrepresentativeof the U.S. population in terms of ten census variables.Individuals in the sample were identified aspotentially gifted or GLD based on theirWJ IV cognitive and academic scores.Amultigroupmeans andcovariance structures analysis was used to test configural, metric, and scalar invariance of a cognitiveabilitiesmodelacrossgifted,GLD,andnon-giftedgroups.

AninitiallatentabilitiescognitivemodelbasedonaWJIVvaliditystudybyNiilekselaandcolleagues(2016)demonstratedadequate fitwith theentire sample,χ2(190)=9597.29,CFI= .872,RMSEA= .086,SRMR=.046, AIC = 9767.29 but demonstrated poor fit with the gifted group and would not run with the GLDgroup.Multiple revisions were attempted to improve the fit; some revisions were informed bynon-significant andnegative correlations among cognitive abilities in the gifted andGLD groups.The revisedcognitive abilities model was hierarchical with a latent general intelligence (g) factor and measuredvariablesforvariousCattell-Horn-Caroll(CHC)broadabilities,withacombinedGf-Gccompositescoreduetoastrongnegativecorrelation(-.63)betweenGfandGcfortheGLDgroup.

Thecognitivemodeldemonstratedconfiguralandmetric invarianceacrossallthreegroups.Partialscalarinvariance for theGLD/non-giftedmodelwasachievedwhenthe intercept forGf-Gcwasallowedtovaryacrossgroups;theGf-GcinterceptfortheGLDgroup(115.12)wasmorethanonestandarddeviationhigherthan the intercept for the non-gifted group (97.97) given the same level ofg.The gifted/GLD scalarinvariancemodelwouldnotrunbecausethevarianceofgwasnearzerofortheGLDgroupwhenmeasuredintercepts were constrained to be equal across groups.Despite adequate fit of the overall model, theloadingofGsongwassmallandnotsignificantforthegifted(.02)andGLDgroups(.03).

From a theoretical standpoint, the cognitive abilities of gifted and GLD individuals violated the positivemanifoldobservedbySpearman.ResultsalsochallengetheutilityofthetypicalCHCmodelofintelligence(and related scoring on intelligence batteries) for gifted and GLD populations.Specifically, full-scale IQscores may underestimate the verbal abilities and fluid reasoning of GLD individuals and the role ofprocessing speed in gifted intelligence is unclear.Although these results are due in part to selectioneffects, the same effects also apply to the real-life selection of gifted individuals and are therefore stillrelevanttopractice.


50

MIGRATORYEVOLUTIONOFIQANDEMINENCE

Prof.HelmuthNyborg1

1UniversityofAarhus(1968-2007),[email protected]

Eons ago a few sons of East-African haplotype Amales andmaternal group I1 females beganmigratingnorthward.About275.000yearslatersomeoftheirdescendantssettleddowninanarrowpolygon-shapedcoreof today’sEuropeaspredominantlyhaplotype I-L22males.There they,andsomeof their sonswhoeventuallymigrated to offshoot countries like the US, brought about 97% of all generally agreedmajorhumanaccomplishmentstofruition-mainlybetween1400-1950-asdocumentedbyMurray(2003)usingstandardhistoriometricmethods.

ThisstudyexaminestheideathatcurrentpopulationdifferencesinIQandtheircovarianttraits,includingeminence,arelargelypunctuatedmigratoryspin-offsofunforgivingprehistoriclatitudinalgeo-bio-climaticnaturalselection.

Testofthisideaincludesmodelingofcognitivedata(5th.degreepolynomial,binomial,anddual-linelinearregression)forgeographicalareaswithin±25olongitude.Thisindicates1)thatthelatitudinalprogressioninmigratoryIQis11.5timessteeperwithinAfricathanoutsideit,2)thattheIQcurvereachesanasymptotejust north of the Mediterranean Sea ( ≈ 390latitude), and 3) that IQ does not evolve under east-westmigration(notevenalongagloballongitudinalgradient).

A 5-part econiche/ecotype classification is then construed to check whether grading of average currentecologic ecotype IQs and covariate trait patterns correspondswith grading of physical average selectivegeo-bio-climaticeconiche conditions. Thehypothesis tobe tested is that the5-point seriesofpunctually“frozen-in-time-and-space” ecotype IQs and their covariate trait patterns for those who settled downpermanently in various geo-bio-climatic zones along the latitudinal gradient, mirror the correspondingseriesofprehistoricselectiveforcesthatdominatedtheirpermanentsettling-downineconichescallingforsite-specific fine-tuning of their covariant Gaussian ecotype trait pattern optimal for survival andreproductionrightthere.Theotherhypothesisisthatextraordinaryscientificandartisticsuccessisvirtuallythe evolutionary prerogative of high-IQ male European polygon settlers, the forefathers of which hadsurvived the full lateral progression of harsh geo-bio-climatic selection through optimized polygenestructuresforIQnexustraits.

With both hypotheses supported, it is concluded that themodern latitudinal co-distribution of geo-bio-climatic econiche characteristics, haplotypes, brain size, IQ, and behavioral traits including eminence isconsistent with a migratory theory of serial progressive polygene and behavioral optimizations duringprehistoricnorthboundmigration.Nodoubt the recentexplosion in large-scalegenomewideassociationstudies will throw further light on the evolutionary basis for modern individual and group differencespredictableintermsofprogressivenorthwardlyextendedstructuralpolygeneselectionforIQandcovarianttraitpatterns,anditrequirespervasivegeneticanddemographicevidencetocausallyexplaintheevolutionofstillmoresophisticatedsocial infrastructures,agriculture, industrialization,modernwelfarestates,andthesupremescientificandartisticaccomplishmentsincoreEurope.

Murray,C.(2003).HumanAccomplishment:ThepursuitofexcellenceintheArtsandSciences,800B.Bto1950.NewYork,USA:HarperCollinsPubl.


51

DIFFERENTIALRELATIONSOFEXTERNALIZINGTOINTELLIGENCEANDCREATIVITY

Ms.AnnaPeddle,Mr.ScottBlainandDr.ColinDeYoung11UniversityofMinnesota,[email protected]

Externalizing behaviors such as aggression and substance abuse have been widely found in thepsychological literature to be negatively correlated with specific aspects of cognitive ability such asemotional intelligence, and more limited research links externalizing to lower general intelligence.Conversely, externalizing and other psychopathological characteristicsmay actually confer advantages incertaincognitivedomains,suchascreativity.Therelationshipbetweencreativityandgeneral intelligencehas been widely researched, and though it seems intelligence is generally positively correlated withcreativity,otherfactorsplayarolewhenitcomestoindividualdifferencesincreativity.Thepresentstudysought to investigate the relationshipbetween IQ,externalizingbehavior, and creativeachievement.Wehypothesized that, somewhat paradoxically, despite their negative correlation with one another, bothexternalizingandintelligencewouldpositivelypredictcreativity.Tothisend,weadministeredmeasuresofcreativity,externalizing,andintelligencetoparticipantsfromtwoindependentsamples.

ParticipantsinSample1(N=304)completedtheExternalizingSpectrumInventory,andthoseinSample2(N=234)completedacollectionofsixself-reportexternalizingmeasures.Participantsinbothsamplesalsocompleted theCreativeAchievementQuestionnaireand four subtestsof theWeschlerAdult IntelligenceScale. Structural equationmodelingwasused toexamine theeffectsof externalizingand intelligenceoncreativeachievement.

LatentvariablescorrespondingtoexternalizingandintelligencewerenegativelycorrelatedinbothSamples1(r=-.15,p=.033)and2(r=-.24,p=.018).InSample1,creativeachievementintheartswaspositivelypredictedbybothexternalizing(β= .33,p< .001)and intelligence(β= .27,p .004). InSample2,creativeachievementintheartswaspositivelypredictedbyexternalizing(β=.25,p=.035),butnotintelligence(β=-.06,p=.551).

Takenintandemwithpastresearch,findingssuggestthatbothexternalizingandintelligencemaypositivelycontributetocreativeachievement,despitetheirnegativecorrelationwithoneanother.Thispresentsanexample of statistical suppression, which should be considered by future researchers individuallyinvestigatingthefactorsofexternalizingand intelligence inrelationtocreativity. Implicationsforbroaderresearchonpsychopathology,creativity,andcognitiveabilitywillbediscussed.


52

COGNITIVEABILITIES,DIVERGENTTHINKING,ANDTHEASPECTSOFOPENNESS

Dr.DavidSchroeder,Dr.AshleyBrownandDr.LindaHouser-Marko1

1JohnsonO'ConnorResearchFoundation,[email protected]

Among the Big Five personality factors, Openness to Experience is notable for its significant, positiverelationship to cognitive ability (Costa & McCrae, 1992). But the devil is ever in the details: DeYoung,Peterson,andHiggins(2005)notedthatsomefacetsofNEO-PI-ROpenness(Fantasy,Aesthetics,Feelings)were correlated with crystallized ability whereas others (Ideas) were correlated with fluid ability andworking memory. Later, DeYoung, Quilty, and Peterson (2007) used these results to buttress theirargument in favorof thedistinctnessof their two“aspects”ofOpennesstoExperience;namely, IntellectandOpenness(hereafter,theterm“Openness”willbeusedtoreferonlytotheaspect,notthebroadtrait).We sought to replicate and extend these findings using an age-diverse sample of 331 individuals withscoresontheJohnsonO’ConnorAbilityBattery(JOAB)andtheSyntheticAperturePersonalityAssessment(SAPA;Condon,2017).BecausetheJOABconsistsofmanydifferenttests,wewonderedwhichabilities, ifany, would be related to Openness rather than or in addition to Intellect. We hypothesized that allcognitiveabilitytestswouldbepositivelyrelatedtoIntellect;wealsopredictedthatameasureofdivergentthinking (DT; operationalized using scores on the JOAB Foresight test) would be positively related toOpenness.We set about testing thesepredictionsusing scoresonDTaswell asona subsampleofnineJOAB tests (N =43,540)which recreated the four-factormodelofabilitydiscussed inHaieretal. (2009):Speed of Reasoning, hereafter “Speed;” Spatial; Numerical; and Memory; plus a single-test “factor” ofVerbalability(theJOABEnglishVocabularytest).Amongthe331individualswithscoresonbothSAPAandJOAB, all five cognitive ability factors aswell asDTwere significantly positively correlatedwith Intellect,whereasonlyDTwassignificantlypositivelycorrelatedwithOpenness.Althoughthesefindingswerelargelyinlinewithexpectation,wewereintriguedtonotenotonlythatDTwasalsorelatedtoIntellect,butalsothatOpennesswasmarginallypositivelyrelatedtoseveralcognitiveabilities (namely,SpeedandSpatial).Simultaneous regression analyses showed that a model in which the five cognitive abilities and DTpredictedIntellectwassignificant(adjustedR2=.296),althoughonlythreeofthepredictorshadsignificantcoefficients (Speed, ß= .164; Verbal, ß= .317; DT, ß= .142). Amodel inwhich the five abilities and DTpredicted Openness was also significant (adjusted R2 = .083) and, as hypothesized, only the regressioncoefficient for DT achieved significance (ß= .297). Notablemarginally significant results, however, wereobtained for Spatial (ß= .124,p = .052) and, unexpectedly,Numerical (ß= -.113,p = .086) abilities. Thelatter is most likely a suppressor effect; that is, wemay infer that the part of Numerical ability that isindependentofSpeed,Spatial,Verbal,Memory,andDivergentThinkingabilityisnegativelyrelatedtotheOpennessaspect.


53

COGNITIVEOPERATIONSUNDERLYINGSOLVINGOFDOMAIN-SPECIFICCASESBYNOVICESANDEXPERTSINMEDICINE

Dr.OlgaShcherbakova 1,Ms.DariaMakarova 1,Ms.VictoriaMorgunova 2,Mr.JosefKlushkin 3andMs.KseniaTeslenko 4

1SaintPetersburgStateUniversity,o.scherbakova@gmail.com2Independentresearcher3NationalMedicalResearchCenterofOncologyn.a.N.N.Petrov4VsevolozhskInterdistrictHospitalDevelopment of expert performance in medicine is well described in literature (Patel et al., 1990;Boshiuzen, Schmidt, 1992; Mylopulous, Regehr, 2009). However, specific cognitive mechanisms andoperationsunderlying intellectualproductivityof ‘novices’and ‘experts’ inmedicineremainobscure.Ourstudyaimedat:1)definingarolethat IQplays insolvingofdomain-specific tasksbynovices inmedicineandmedicalexperts;2)describingcognitiveoperationsthatunderliesolvingofsuchtasksinbothgroups.We predicted that: 1) IQ is of minor importance for successful solving of domain-specific tasks in both‘novices’ and ‘experts’ in medicine; 2) ‘novices’ and ‘experts’ in medicine involve different cognitiveoperationswhilesolvingdomain-specifictasks.

Participantswere30doctorsofvariousmedicalspecializations(10males,aged21–48,mean–27y.o.):15‘novices’(<3yearsofpractice),15‘experts’(>3yearsofpractice).AfterbeingtestedwithJ.Raven’sSPMT,they solved 5 professional cases during indepth interviews. Caseswere based on real practice; requireddiagnosis and suggestion of treatment options for a somatic disease; did not require any narrowly-specializedknowledge;couldbesolvednotthroughapplyingstereotypedsuperficialsolutions,butthroughsystematic analysis of the issue. For each case, a group of experts (N = 9) suggested an etalon answer.Interviews (t = 45 hours) were audio-recorded and then transcribed verbatim; written protocols wereassessedbytwoexpertswhoscoredtheanswerforeachcase0,1,2or3dependingonhowclose/far itwasfromtheetalon.Basedonthesumofthescores,participantsweredividedbymedianin‘successful’(0–6scores,N=15)and‘unsuccessful’(>6scores,N=15)groupswhichdidnotoverlapwith‘novices’and‘experts’.Mann-Whitney test showed that ‘experts’ weremore successful in solving cases compared to‘novices’ (p = 0,015). No differences in IQ between ‘novices’ and ‘experts’ (p = 0,058), nor between‘successful’ and ‘unsuccessful’ participants (p = 0,298) were found, which supports our prediction andcorresponds to the results of our previous study (Shcherbakova, Makarova, 2016). We interpret thesefindings that mechanisms of experts’ intellectual productivity differ from those underlying high level ofpsychometricintelligence.

Qualitativeanalysisoftheinterview’sprotocolsrevealeddifferencesincognitiveoperationsthat‘novices’and‘experts’involvewhilesolvingcases.‘Novices’tendtochooseonesolutionanddevelopitignoringanylogical incongruenciesandnotaccountingfeedback;donotembraceallaspectsoftheproblemsituation;do not critically assess the results of their thinking; do not doubt hypotheses they suggested; preferstereotyped solutions, simplified representations of problem situations and display ‘naïve’ intellectualbehavior.‘Experts’activelyseekforextrainformation;changetheirhypothesesafterreceivingfeedback;donotoperatewithdiscretesymptomsbutdisplaysystematicanalysisofaproblem;acknowledgetheirownmistakes;displayinsistenceoftheircognitiveattempts.

Ourfutureresearchwillfocusonexaminingpatternsofelectrophysiologicalactivitydisplayedwhilesolvingdomain-specificcasesbymedical‘novices’and‘experts’.

SupportedbytheGovernmentofRussianFederationgrant№14.W03.31.0010


54

THEIRTDIFFERENCESINRUSSIANANDYAKUTSAMPLES

Mr.VladimirShibaev11VSUES(vladivostokstateuniversityofeconomicsandservice),[email protected]

Differences in intelligence of Yakuts and Russians has been founded by the direct research byStandardProgressivesMatricePlusversion(non-verbal,cross-culturaltest).

TherewascalculatedthedifferencesbetwennethnicRussians(n=968)andethnicYakuts(n=702)bytheIRTtheory(DIf,Lord(1980)andalsoMantel-Haenszelmethodwereusedalso).

Herethepictureturnedout likethis.B9,B10,B11,B12,C11aresignificantlyshifted,andC3,C4,C12aremarginallysignificantlyshifted infavoroftheRussians;D3,D6,E1,E2,E3,E4,E6aresignificantlyshifted,andC7,D4,D8,E10aremarginallysignificantlyshiftedinfavoroftheYakuts.

DIFintwoofthethree,atleast,methodsaredetectedbypointsB9,B10,etc.AconnectedgroupconsistsofparagraphsB9-B12.Thesetaskslookquitesimple,butforsomereasonrepresentacertaindifficultyfortheYakuts.

Thefollowingexplanationsarepossible.

1.InEuropeanculturethereisa“scheme”,akindofgestalt:Aa,AB,Bb,ab.Itfacilitatesthefillingofoneoftheelementsincaseofitsabsence;

2.ItismoredifficultforYakutstoswitchfromaperceptualdecisiontoananalyticalone.


55

GLOADINGSOVERTHE20THCENTURY:ANALYSISOFCARROLL’SHUMANCOGNITIVEABILITIESDATASET

Dr.EkaterinaValuev,Dr.SofyaBelovaandProf.DmitryUshakov1

1InstituteofpsychologyRussianAcademyofSciences,[email protected]

TheFlynneffectisaphenomenonoftemporaldynamicsofpsychometricintelligence,namelytheincreaseinintelligencescoresoverthe20thcentury.Theresearchquestionofthecurrentstudywastoestimatetwomoretemporalcharacteristicsofintelligencewhichpresumablywerechangingoverthe20thcentury.Theyareintelligencetests’gloadings,andtheamountofvarianceexplainedbygfactor.

Thedataon21intelligencetests(orsubtests)presentinatleast3correlationmatrixduring1930-1991wasextracted fromHumanCognitiveAbilitiesDataSetArchive.Among themtherewere9 subtestsofWAIS,somesubtestsKitofFactorReferencedCognitiveTests,SPM,CPM,etc. Wehaveanalyzed77correlationmatriceswith 214 tests in total. Principle component factor analysiswas conducted on each correlationmatrix which allowed to calculate factor loadings on the first factor for each intelligence test.We alsoestimated g factor variance (i.e. variance explained by a first factorwithout rotation) in each study. Foreachintelligencetestthecorrelationbetweenyearofthestudyandgloadingwascalculated.Correlationcoefficients took a range of values from -0.74 to 0.25. Themeta-analysis based on the random effectsmodeldemonstratedthatmeanvaluewas-0.165(p=0.04).Besidesforeachintelligencetestthecorrelationbetweenyearofthestudyandvarianceexplainedbygfactorwascalculated.Correlationcoefficientstookarangeof values from -0.90 to 0.62. Themean valuewas -0.2 (p=0.0179). These results indicate that theFlynneffect) is accompaniedbyweakeningof general intelligence factor (i.e. bydecreaseof intelligencetests’loadingsongfactorandbydecreaseoffactorgvariance).

The obtained results can be considered in the framework of structural-dynamic theory of intelligencedevelopedbyD.Ushakov.Thekeynotionofthetheoryistheindividual’scognitivepotential,whichdefineshow quickly an individual forms cognitive systems that can provide the problem solution. The theoryassumesthatintelligencestructureisbeingformedbyinvestmentofcognitivepotentialtodifferentabilitydomains determined by influence of social and cultural environment. The increase of intelligence underfavorable environmentmay follow any direction but its limit is determinedby genetics behind cognitivepotential.Assoonasthis limit is facedthe increasemaystillkeepgoingup insomedirectionsbutattheexpenseofotherdomains.Thustheredistributionofthecognitivepotentialfromonedomaintoanotheristaking place. The Flynn effect registered by means of intelligence tests reflect development of rationalconstituent of cognitive potential over the 20th century caused by transition from industrial topostindustrial society.Alongwith that thediversityofenvironmentalconditions is increasing,andduetogeneralhumanizationandliberalizationofsocietypeopleobtainthepossibilitiestochoosefromawidersetof alternative scenarios of specialization and investment of cognitive potential. This phenomenon maymanifestitselfingfactor’sweakeninganddecreaseofintelligencetests’gloadings.

ThisresearchwassupportedbytheRussianFoundationforBasicResearch,project№17-06-00574


56

ANALYSINGVISUALSEARCHSTRATEGIESINARTISTSANDENGINEERS–ANEYETRACKINGSTUDYONRPM

Ms.SravyaVatsavayiandDr.KavitaVemuri 1

1InternationalInstituteofInformationandTechnology,Hyderabad,[email protected]

Understanding visual search process and correlating to performance in fluid intelligence tests providesinsights to the strategies applied (constructive, elimination techniques) and thinking styles.Weexploredthe role of skill training on strategy for two diverse groups – fine art students (divergent thinking) andengineeringstudents(convergentthinkingstyle)withbasicartisticskills.ThestandardRaven’sprogressivematrices (RPM)wasconsidered.TwosectionsofRPM(D,E levels)wasadministeredto8 (3subjectsdatapoints were discarded due to low% of eye data) students of fine arts ( age = 23-31, 2 females) and 5artistically inclined engineering students (age = 18-23, 3 females). Levels D,E were selected for theirdifficultylevelandnovelpatterns.QuestionsinDand10questionsofEareconstructedbaseduponcertainsetsofrules.D-levelcontainsmultiplepatternsinasinglequestionpresentingperceptualchallenge,whiletheE-levelexhibitspatternswithsimplearithmeticoperationsandidentificationofsimilaritiesthustestingfor analytical ability. The questions were ported as a computer-based application implemented using agameengine(Unity-3D).Theorderofpresentationofquestionswererearrangedtoreducespatialmemoryconfoundsbyensuringthatnotwoquestionsfollowthesameruleofdeductionandtodisruptrepetitivepatternlogic.EyetrackingdatawascollectedwithTobiiT-120andX2-30systems.Fixationsandscan-pathofeachparticipant/questionwasanalyzed.

Theengineeringstudents(scorerange:24to12,M=18.8,SD=4.62)performancewasbetterthanartists(scorerange:12to7,M=10.8,SD=1.93).HigherdifferencewasnoticedintheE-level(Engineering:M=9.2, SD = 3.18; Artists : M = 3.2, SD = 1.32). The analysis of scan-path indicated that both sets ofparticipants majorly apply either of the 3 pattern-identification techniques : horizontal (row-wise) ,vertical (column-wise ), diagonal (pair-wise scan of diagonal elements). It was observed all participantsmostquestionswereansweredusinghorizontalscan-paththoughnosignificantadvantage(performance)wasobservedforanyofthethreescanstrategies.AStructuralSimilarityIndexMethodalgorithmappliedtocompare image structural similarity in the respective scan-pathdirections foundno significantdifferencetoo.Artists tended to fixateon the last rowand themissingmatrixelement,whileengineering studentsrarelydoso.Apossibleexplanationbasedonthethinkingstylesis,artistsconstructthemissingelementbyevaluatingmultiple solutionswhile engineers deduce a single solution. A secondary possible process byartistscanbeanattempt tocreateadesign to fit themissingelementderived fromtheshape/structurecontinuityoftheotherelements.

The options (provided at the bottom of the matrix) were not considered by the engineering students(fixation count <25.5%) indicative constructive technique,while artists (fixation count >41.1%) seem tohave applied the elimination technique. Though preliminary, the results point at possible preference ofartists towards constructingananswerattributed to their trainingor ability in creatingpatterns. Furthertestingwithlargerparticipantsetarerequiredtoarriveatstrongerobservations.


57

COGNITIVEPREDICTORSOFTHEVARIATIONINSTANDARDIZEDTESTPERFORMANCEINRUSSIAANDKYRGYZSTAN

Dr.IvanVoronin 1,Prof.TatianaTikhomirova 1,Ms.IrinaLysenkova 2andProf.SergeyMalykh 1

1PsychologicalInstituteofRussianAcademyofEducation;MoscowStateUniversity,[email protected] 2Kyrgyz-RussianSlavicUniversityTheUnifiedStateExam(USE)hasbeenservingasatoolforassessmentoflearningoutcomesatschoolandforuniversityadmissionsinRussiasince2006.Inourpreviousresearch,thevariationofMathematicsandRussian USE performance remained largely unexplained after taking into account mathematical ability,intelligence,basiccognitivefunctionsanddifferencesinmathematicalcurricula(Voroninetal.,2018).TheaimofcurrentstudyistoprovidemoreevidenceonthevariabilityinstandardizedtestperformanceandtocompareUSEtoNationwideTesting(NT),aKyrgyzreplicaofSAT.AllschoolstudentsinKyrgyzstanhavetopassMainNTforfurthercollegeadmission.MainNTassessescognitiveandacademicskillsandliteracyinmathematicsand innative language (KyrgyzorRussian). The students canalso choose topassadvancedNTsinspecificsubjects(Mathematics,Biology,Physicsetc.).

Thesampleincluded266Russianstudents(meanage17.7years,SD=0.4years)and829Kyrgyzstudents(meanage17.5years,SD=0.4years).Inbothgroupsweassesseduntaughtnumericabilities(NumberLine,NumberSense),mathematicalskills(ProblemVerificationTask,UnderstandingNumber),workingmemorycapacity (Corsi Block), reaction time (four-choice simple reaction time) and general cognitive ability(Raven’s Progressive Matrices). For Russian students we acquired Mathematics USE and Russian USEscores.ForKyrgyzstudentsweobtainedMainNTandMathematicsNTscores.

During thedatapreparationwe identifieda groupofKyrgyz studentswithhigh response speedand lowaccuracyinNumberSensetask(35%oftheKyrgyzsample).Theirdatawereclearlydistinguishablewithinthe responsedistributions inall cognitive testsbutnot inRaven’sProgressiveMatricesorNTscores.Weexcludedcognitivedataofthesestudentsfromfurtheranalysis.

MathematicsUSEscorewasassociatedwithallcognitivecharacteristics,correlationsbetween0.147(CorsiBlock)and0.414(UnderstandingNumber).RussianUSEscorewasassociatedwithacquiredmathematicalknowledge (Problem Verification, 0.302, and Understanding Number, 0.280) and Number Line (-0.153).Raven’s score was associated with Mathematics USE (0.466) but not Russian USE (0.115, ns). Whenintroduced into regression model altogether (using SEM approach), predictors explained 35% ofMathematics USE variation and 18% of Russian USE variation. However, the separate contributions ofcognitivepredictorswerenon-significant,exceptlearnedmathematicalskills(RussianUSE)andintelligence(MathematicsUSE).

For Kyrgyz sample,MainNT scorewas associatedwith all cognitive characteristics included, correlationsrangedbetween0.117(NumberSense)and0.336(UnderstandingNumber).ThecorrelationbetweenMainNT and Raven’s score was 0.293. Neither of cognitive predictors were associated withMathematics NTscore(correlationsunder0.107,ns).Inregressionmodel,allcognitivepredictorsexplained20%ofMainNTand3%ofMathematicsNT.

The results of current study support out previous findings suggesting that cognitive characteristics andintelligencehave limitedpower inexplaining individualdifferences inUnifiedStateExamscores. In turn,althoughNationwideTestinginKyrgyzstanissimilartoSATinmanyaspects,themajorpartofitsvariabilityremainedunexplained.

ThestudywassupportedbytheRussianScienceFoundation,grant17-78-30028.


58

INTELLIGENCE,SENSORYDISCRIMINATION,ANDIMPLICITREWARDLEARNING

Mr.MuchenXi,Ms.DaiqingZhao,Mr.ScottBlainandDr.ColinDeYoung 1

1UniversityofMinnesota,[email protected]

Introduction: An implicit reward-learning task has been frequently used as a behavioral marker forphenotypes such as anhedonia and inattention. The task asks participants to discriminate between twosimilarstimuli, rewardingaportionofcorrectanswers,usingdifferential reinforcementrates for thetwostimuli. Studies using this task have found negative associations between self-reported depressivity andparticipants’tendenciestodeveloparesponsebiastowardthemorefrequentlyrewardedstimulus.Despitethewidespreaduseof this task, toourknowledge,noonehasexaminedtaskperformance inrelationtointelligence.Previousstudiesusingdifferenttaskshaveshownthatintelligenceisassociatedwithabilitytoperformlow-levelsensorydiscriminationandexplicit,butnotimplicit,rewardlearning.Ourcurrentstudyattempted to look at the effect of intelligence on performanceon the implicit reward-learning task.Wehypothesized intelligencewouldbepositivelyassociatedwith theability todiscriminatebetweenstimuli,butnotwithlevelsofimplicitrewardlearning.

Method: In the current study, participants (N = 303) were asked to complete an implicit probabilisticreward-learning task, whichmeasured reward sensitivity, as well as the ability to discriminate betweensimilarvisualstimuli.Weimplementedsignaltocalculateparticipants’responsebiastowardonestimulusortheother,aswellastheaccuracyofstimulusdiscrimination,whencontrollingforbias,foreachofthreetaskblocks.Change in responsebias fromBlocks1 to2wasalso calculated.Participantsalso completedfour subtests from the Weschler Adult Intelligence Scale. Structural equation modeling was used toexamine the effects of latent intelligence on latent variables corresponding to stimulus discrimination,responsebias,andchangeinresponsebias.

Results:Allmanifestvariablessignificantly loadedontotheircorresponding latentvariables,asexpected.Intelligencepositivelypredictedstimulusdiscrimination(β=.37,p<.001).However,intelligencepredictedneitherresponsebias(β=-.10,p=.32)norchangeinresponsebias(β=-.07,p=.207).

Conclusions: Findings dovetail with previous work suggesting that people with higher intelligence arebetter ability to distinguish between stimuli differing in subtle, low-level sensory characteristics.Conversely, intelligence had no effect on implicit reward learning, as measured by response bias andchangeinresponsebias.Thisindicatesthatintelligencemaynotinfluencetheabilitytolearnfromrewardcuesthatarenotexplicitlypartofatask,alsosuggestingthosewithhigher intelligencemaybebetteratattendingtothestatedinstructionsanddemandsofthesetasks.Futureresearchusingthistaskshouldtakenote of these findings and consider controlling for intelligence in individual difference analyses.Researchersshouldalsoconsidertheuseofdiscriminabilityasameaningfulvariableofinterest(ratherthananuisancevariable).

U

59

AcknowledgmentsSincerethankstolocalhostProf.JamesJ.Leeandco-hostProf.MattMcGue,forpreparingasuper20thAnnualmeetingforISIRSincerethanksarealsoextendedtoCescaEatonforpreparingtheprogram

Inadditionwehaveseveralfurthervotesofappreciation:

OurthankstoTheInstituteforMentalChronometryfortheirsupportforISIR2019

WethankElsevierwarmlyfortheirgeneroussupportofthePostersessionandReception

OurthankstoalltheReviewerswhoreadsubmissionsforthisconference

Our thanks to allmemberswho have given of their timemost generously on numerous committees toenabletheconferencetorunsmoothlyOursincerethankstothejudgesoftheStudentAwards

U

60

UsefulInformation

TheDepotHotel225South3rdAvenueMinneapolis,MN55401Phone:612-375-1700

61

UsefulInformation

GettingAroundMinneapolis1MetroTransit(https://www.metrotransit.org/)-MetroTransitprovidespublictransportationtheTwinCitiesviabusandMETROLightrailroutes.TheMETROLightrail’sBlueLinetrainisaparticularlyeasyandefficientwaytotravelbetweendowntownandtheMSPairport.UberLyftTaxiservices(manymoreavailable)-MinneapolisTaxi/onCabs(https://oncabs.com/booking/minneapolis)-MSPAirportTaxiService(http://www.airporttaxiservicemsp.com/)-RedandWhiteTaxiService(http://www.redwhitetaxi.com/)NiceRide(https://www.niceridemn.org/)-Quickbicyclerentalwith200+locationstopickupanddropoff1LocalinformationpreparedbyJeffreyDahlkeandCaseyGiordano

RestaurantsneartheDepotHotel(Sortedbydistance(closesttofarthest);0.5to1.3miles;additionaldetailsinseparatehandout)NameContactInformationAddressSpoonandStablehttps://www.spoonandstable.com/Phone:612-224-9850211FirstStreetNorthMinneapolis,MN55401TheBachelorFarmerhttp://thebachelorfarmer.com/Phone:612-206-392050North2ndAvenueMinneapolis,MN55401112Eateryhttp://www.112eatery.com/Phone:612-343-7696112North3rdStreetMinneapolis,MN55401Astercafehttp://astercafe.com/Phone:612-379-3138125SEMainStMinneapolis,MN55414Vic’sRestauranthttp://vicsminneapolis.com/Phone:612-312-2000201MainSt.SEMinneapolis,MN55414

62

SevenSushiandSteakhousehttps://www.7mpls.com/Phone:612-238-7770700HennepinAveSMinneapolis,MN55403OceanaireSeafoodhttp://www.theoceanaire.com/Phone:612-333-227750SouthSixthStreetMinneapolis,MN55402Tugg’sTavernhttp://tuggsminneapolis.com/Phone:612-379-4404219MainStSEMinneapolis,MN55414RestaurantAlmahttps://www.almampls.com/Phone:612-379-4909528UniversityAve.SEMinneapolis,MN55414BarLaGrasahttp://www.barlagrassa.com/Phone:6123333837800WashingtonAve.NMinneapolisMN55401Cravehttps://www.craveamerica.com/Phone:612-332-1133825HennepinAveMinneapolis,MN55402PizzaLucehttps://pizzaluce.comPhone:612-333-7359119North4thStreetMinneapolis,MN55401Butcher&theBoarhttp://butcherandtheboar.com/Phone:612-238-88881121HennepinAve.Minneapolis,MN55403LotusRestauranthttps://www.lotusmplsmn.com/Phone:612-870-1218113WestGrantStreetMinneapolis,MN55403

63

PlacestoVisitUniversityofMinnesota’sMinneapolisCampus(Sortedbydistance(closesttofarthest);1.9to2.3miles;additionaldetailsinseparatehandout)NameContactInformationAddressWeismanArtMuseumhttp://wam.umn.edu/Phone:612-625-9494333ERiverPkwyMinneapolis,MN55455CoffmanMemorialUnionhttp://sua.umn.edu/locations/coffman/Phone:612-625-5000300WashingtonAveSEMinneapolis,MN55455ElliottHall(PsychologyDept.)https://cla.umn.edu/psychologyPhone:612-625-281875ERiverPkwyMinneapolis,MN55455McNamaraAlumniCenterhttp://mac-events.org/Phone:612-624-9831200SEOakStMinneapolis,MN55455

Museums,AthleticVenues,&HistoricalSites(Sortedbydistance(closesttofarthest);0.4to11.3miles;additionaldetailsinseparatehandout)NameContactInformationAddressMillCityMuseumhttp://www.millcitymuseum.org/Phone:612-341-7555704South2ndSt.Minneapolis,MN55401GuthrieTheaterhttps://www.guthrietheater.org/Phone:612-377-2224818S2ndStMinneapolis,MN55415StoneArchBridgeandSt.AnthonyFallsScenicareaontheMississippiRiver,activitytrails,andthefamousfallsthatpoweredMinneapolis’millingindustryfordecades.100PortlandAve,Minneapolis,MN55401

64

U.S.BankStadiumhttps://www.usbankstadium.com/Phone:612-777-8700401ChicagoAveMinneapolis,MN55415WalkerArtCenterhttps://walkerart.org/Phone:612-375-7600725VinelandPlMinneapolis,MN55403MinneapolisSculptureGardenhttps://walkerart.org/visit/gardenPhone:612-375-7600726VinelandPlMinneapolis,MN55403FoshayMuseumandObservationDeckhttp://www.wminneapolishotel.com/Phone:612-215-3783821MarquetteAvenueMinneapolis,MN,55402ScienceMuseumofMinnesotahttps://www.smm.org/Phone:651-221-9444120WKelloggBlvdStPaul,MN55102MallofAmericahttps://www.mallofamerica.com/60EastBroadwayBloomington,MN55425MinnesotaHistoricalSocietyhttp://www.mnhs.org/Multiplelocations(seewebsite)

LeisureActivitiesintheTwinCitiesRunningOptionsMinneapolis’MississippiRivertrailsareeasilyaccessiblefromdowntown.ThetrailsexistalongboththeeastandwestbanksoftheMississippiandcanbeaccessedattheSt.AnthonyFalls/FatherHennepinParkarea.Amapofpopulartrailsandparkscanbefoundathttps://minneapolisrunning.com/twin-citiestrail-map/

65

NOTES

66

NOTES

67

NOTES

Documents

International Society for Intelligence Research (ISIR)...International Society for Intelligence Research (ISIR) 20th Annual Conference July 11–13, 2019 Minneapolis, MN (United States)