II RNCE IAS abril 2015 copy - Educação para o Século...

II Reunião REDE NACIONAL DE CIÊNCIA PARA EDUCAÇÃO 27 de abril de 2015 Instituto Ayrton

Senna, SP

Competências Cognitivas no Século XXI - Medidas de Processos de Raciocínio, Conhecimento e

Criatividade em sistemas de avaliação em larga escala

Dr. Ricardo Primi, Universidade São Francisco, Itatiba, Brasil Instituto Ayrton Senna, São Paulo, Brasil

EduLab21 Centro de Conhecimento

do IAS

Março 2015

Objetivos

• Apresentar uma breve revisão sobre os modelos sobre inteligência (“estado da arte”)

• Relacionar as competências do século XXI com os modelos mais recentes de inteligência (CHC)

• Refletir e advogar a importância desse modelo para mapear o que avaliamos atualmente nos sistemas em larga escala e propor futuros aprimoramentos

Competências do Século XXI

Aprendizagem e inovação

Criatividade

Pensamento crítico e solução de problemas

Comunicação e colaboração

Habilidades para carreira e para vida

Flexibilidade

Adaptação à mudança

Gerenciar objetivos e tempo

Trabalhar independentemente

Aprendizagem com auto-direção

Interagir efetivamente com os outros

Trabalhar efetivamente com a diversidade

Liderança

Responsabilidade para com os outros

www.p21.org

Copyright © National Academy of Sciences. All rights reserved.

Education for Life and Work: Developing Transferable Knowledge and Skills in the 21st Century

Committee on Defining Deeper Learning and 21st Century Skills

James W. Pellegrino and Margaret L. Hilton, Editors

Board on Testing and Assessment and

Board on Science Education

Division of Behavioral and Social Sciences and Education

EDUCATION FOR LIFE AND WORK

Developing Transferable Knowledge and Skil ls in the 21st Century

Competências do Século XXI

Copyright ©

National Academ

y of Sciences. All rights reserved.

Education for Life and Work: D

eveloping Transferable Knowledge and Skills in the 21st C

entury

32 TABLE 2-2 Clusters of 21st Century CompetenciesTABLE 2–2 Clusters of 21st Century Competencies

Cluster Terms Used for 21st Century Skills

O*NET Skills

Main Ability/ Personality Factor

Cognitive Processes and Strategies

Critical thinking, problem solving, analysis, reasoning/argumentation, interpretation, decision making, adaptive learning, executive function

System skills, process skills, complex problem-solving skills

Main ability factor: fluid intelligence (Gf)

COGNITIVE COMPETENCIES

Knowledge

Information literacy (research using evidence and recognizing bias in sources); information and communications technology literacy; oral and written communication; active listening

Content skills Main ability factor: crystallized intelligence (Gc)

Creativity

Creativity, innovation

Complex problem-solving skills (idea generation)

Main ability factor: general retrieval ability (Gr)

Copyright ©

National Academ




entury

33

Cluster

Terms Used for 21st Century Skills

O*NET Skills


Intellectual Openness

Flexibility, adaptability, artistic and cultural appreciation, personal and social responsibility (including cultural awareness and competence), appreciation for diversity, adaptability, continuous learning, intellectual interest and curiosity

[none] Main personality factor: openness

INTRA- PERSONAL COMPETENCIES

Work Ethic/ Conscientious- ness

Initiative, self-direction, responsibility, perseverance, productivity, grit, Type 1 self-regulation (metacognitive skills, including forethought, performance, and self-reflection), professionalism/ ethics, integrity, citizenship, career orientation

[none] Main personality factor: conscientiousness

Positive Core Self- Evaluation

Type 2 self-regulation (self-monitoring, self-evaluation, self-reinforcement), physical and psychological health

[none] Main personality factor: emotional stability (opposite end of the continuum from neuroticism)

continued

Copyright ©

National Academ




entury

34 TABLE 2-2 Continued

SOURCE: Created by committee.

� Cluster

Terms Used for 21st Century Skills�

O*NET Skills


� � � � ��

�7HDPZRUN�DQG�&ROODERUDWLRQ��

&RPPXQLFDWLRQ��FROODERUDWLRQ��WHDPZRUN��FRRSHUDWLRQ��FRRUGLQDWLRQ��LQWHUSHUVRQDO�VNLOOV��HPSDWK\�SHUVSHFWLYH�WDNLQJ��WUXVW��VHUYLFH�RULHQWDWLRQ��FRQIOLFW�UHVROXWLRQ��QHJRWLDWLRQ�

6RFLDO�VNLOOV� 0DLQ�SHUVRQDOLW\�IDFWRU��DJUHHDEOHQHVV��

INTER- PERSONAL COMPETENCIES

� � � �

� ��/HDGHUVKLS��

/HDGHUVKLS��UHVSRQVLELOLW\��DVVHUWLYH�FRPPXQLFDWLRQ��VHOI�SUHVHQWDWLRQ��VRFLDO�LQIOXHQFH�ZLWK�RWKHUV��

6RFLDO�VNLOOV��SHUVXDVLRQ��

0DLQ�SHUVRQDOLW\�IDFWRU��H[WURYHUVLRQ��

� � � � �

� � � � �

� � � � �

�

6285&(��&UHDWHG�E\�WKH�FRPPLWWHH��

Competências do Século XXI• "do século XXI" ou "mais valorizadas em nosso tempo”?

• capacidades já conhecidas

• Qual o “estado da arte" do campo de estudo da inteligência? Qual o conjunto de evidências que baseiam os sistemas de avaliação em larga escala ?

• Quais medidas compõem os sistemas de avaliação em larga escala ? Todos os fatores cognitivos relevantes são cobertos ?

• Abordagens da Inteligência

• Educação foco no conhecimento

• Psicologia foco nos processos mentais

Breve histórico da evolução das teorias sobre a inteligência

Fator g de Charles Spearman

• 1904 General intelligence objectively determined and measured

• "positive main fold”. Invenção da análise fatorial. Modelo bifatorial. “energia mental”. neogenese.

• (a) edução de relações capacidade maior ou menor de estabelecer relações entre duas ou mais idéias;

• (b) edução de correlatos capacidade maior ou menor que as pessoas demonstram de criar novas idéias a partir de uma idéia e uma relação.

• (c) apreensão das experiências capacidade à rapidez e à acuidade com que as pessoas percebem os estímulos, e bem como aos processos de;

Quase 100 anos depois:

Duncan, J., Seitz, R. J., Kolodny, J., Bor, D., Herzog, H., Ahmed, A., ... & Emslie, H. (2000). A neural basis for general intelligence. Science, 289(5478), 457-460.

Teoria do Investimento de R. Cattell

• Inteligência Fluida Gf: capacidade de raciocínio “uso de operações mentais controladas deliberadamente para resolver problemas novos... que não podem ser resolvidos automaticamente” .. “ operações mentais como: fazer inferências, formar conceitos, classificar, gerar e testar hipóteses, identificar relações, compreender implicações, resolver problemas, extrapolar e transformar informações” (MacGrew & Evans, 2004; Kane & Gray, 2005)

• Inteligência Cristalizada Gc Riqueza (profundidade e amplitude) do estoque dos conhecimentos adquiridos

Variáveis Cognitivas e Não Cognitivas naEscola: Modelo de J. B. Carroll (1963)

=

Quantidade de tempo engajado na tarefa

Quantidade de tempo necessário

Quantidade de tempo necessário

Aprendizagem efetiva

=Capacidade

prévia

1

+ Qualidade do ensino

Quantidade de tempo engajado

na tarefa= Oportunidade + Motivação/Per

sistência

Uma loja vende produtos importados e nacionais entre vestidos camisas e casacos. Alguns vestidos e todos os casacos fazem parte dos produtos importados. Não há produto importado disponível em tamanho grande. Assinale qual dentre os fatos enunciados não poderia ser verdadeiro:

A. Carla experimenta uma camisa nacional.B. Luciana está comprando um casaco pequenoC. Alberto pegou um casaco grandeD. Adriana experimenta um vestido pequeno.

2014

MT - 2º dia | Caderno 5 - AMARELO - Página 19

MATEMÁTICA E SUAS TECNOLOGIAS

Questões de 136 a 180QUESTÃO 136

A Figura 1 representa uma gravura retangular com 8 m de comprimento e 6 m de altura.

8 metros

6 m

etro

s

Figura 1Deseja-se reproduzi-la numa folha de papel retangular

com 42 cm de comprimento e 30 cm de altura, deixando livres 3 cm em cada margem, conforme a Figura 2.

Folha de papel3 cm

3 cm

3 cm

3 cm

3 cm

3 cm3 cm

3 cm

42 cm

30 cm

Região disponível para reproduzir a gravura

Região proibida para reproduzir a gravura

Figura 2A reprodução da gravura deve ocupar o máximo

possível da região disponível, mantendo-se as proporções da Figura 1.

PRADO, A. C. Superinteressante, ed. 301, fev. 2012 (adaptado).

A escala da gravura reproduzida na folha de papel é

A 1 : 3.B 1 : 4.C 1 : 20.D 1 : 25.E 1 : 32.

QUESTÃO 137

Uma empresa que organiza eventos de formatura confecciona canudos de diplomas a partir de folhas de papel

cada folha é enrolada em torno de um cilindro de madeira de diâmetro d em centímetros, sem folga, dando-se 5 voltas

cordão no meio do diploma, bem ajustado, para que não

Em seguida, retira-se o cilindro de madeira do meio

Considere que a espessura da folha de papel original seja desprezível.

Qual é a medida, em centímetros, do lado da folha de papel usado na confecção do diploma?

A dB 2 dC 4 dD 5 dE 10 d

QUESTÃO 138

Uma ponte precisa ser dimensionada de forma que possa ter três pontos de sustentação. Sabe-se que a carga máxima suportada pela ponte será de 12 t. O ponto de sustentação central receberá 60% da carga da ponte, e o restante da carga será distribuído igualmente entre os outros dois pontos de sustentação.

No caso de carga máxima, as cargas recebidas pelos três pontos de sustentação serão, respectivamente,

A 1,8 t; 8,4 t; 1,8 t.B 3,0 t; 6,0 t; 3,0 t.C 2,4 t; 7,2 t; 2,4 t.D 3,6 t; 4,8 t; 3,6 t.E 4,2 t; 3,6 t; 4,2 t.

*AMAR25DOM19*

2014

LC - 2º dia | Caderno 8 - ROSA - Página 8

QUESTÃO 100

Óia eu aqui de novo xaxandoÓia eu aqui de novo para xaxar

Vou mostrar pr’esses cabrasQue eu ainda dou no couroIsso é um desaforoQue eu não posso levarQue eu aqui de novo cantandoQue eu aqui de novo xaxandoÓia eu aqui de novo mostrandoComo se deve xaxar

Vem cá morena lindaVestida de chitaVocê é a mais bonitaDesse meu lugarVai, chama Maria, chama LuziaVai, chama Zabé, chama RaqueDiz que eu tou aqui com alegria

BARROS, A. Óia eu aqui de novo. Disponível em: www.luizluagonzaga.mus.br.Acesso em: 5 maio 2013 (fragmento).

A letra da canção de Antônio de Barros manifesta aspectos do repertório linguístico e cultural do Brasil. O verso que singulariza uma forma característica do falar popular regional é:

A “Isso é um desaforo”.B “Diz que eu tou aqui com alegria”.C “Vou mostrar pr’esses cabras”.D “Vai, chama Maria, chama Luzia”.E “Vem cá morena linda, vestida de chita”.

QUESTÃO 101

Em uma escala de 0 a 10, o Brasil está entre 3 e 4 no quesito segurança da informação. “Estamos começando a acordar para o problema. Nessa história de espionagem corporativa, temos muita lição a fazer. Falta consciência institucional e um longo aprendizado. A sociedade caiu em si e viu que é uma coisa que nos afeta”, diz S.P., pós-doutor em segurança da informação. Para ele, devem ser estabelecidos canais de denúncia para esse tipo de situação. De acordo com o conselheiro do Comitê Gestor da Internet (CGI), o Brasil tem condições de desenvolver tecnologia própria para garantir a segurança dos dados do país, tanto do governo quanto da população. “Há uma massa de conhecimento dentro das universidades e em empresas inovadoras que podem contribuir propondo medidas para que possamos mudar isso [falta de segurança] no longo prazo”. Ele acredita que o governo tem de usar o seu poder de compra de softwarese hardwares para a área da segurança cibernética, de forma a fomentar essas empresas, a produção de conhecimento na área e a construção de uma cadeia de produção nacional.

SARRES, C. Disponível em: www.ebc.com.br. Acesso em: 22 nov. 2013 (adaptado).

Considerando-se o surgimento da espionagem corporativa em decorrência do amplo uso da internet, o texto aponta uma necessidade advinda desse impacto, que se resume em

A alertar a sociedade sobre os riscos de ser espionada.B promover a indústria de segurança da informação.C discutir a espionagem em fóruns internacionais.D incentivar o aparecimento de delatores.E treinar o país em segurança digital.

QUESTÃO 102

WILL. Disponível em: www.willtirando.com.br. Acesso em: 7 nov. 2013.

Opportunity é o nome de um veículo explorador que aterrissou em Marte com a missão de enviar informações à Terra. A charge apresenta uma crítica ao(à)

A gasto exagerado com o envio de robôs a outros planetas.B exploração indiscriminada de outros planetas.C circulação digital excessiva de autorretratos.D vulgarização das descobertas espaciais.E mecanização das atividades humanas.

*ROSA25DOM8*

different analytic approaches. In model 1, intelligence and prioracademic achievement were considered together as a direct predictorof final academic achievement. In model 2, both variables were testedas autonomous predictors of final academic achievement, and directand indirect effects were also tested.

In general, results show that 1) when intelligence and prioracademic achievement are considered together, prior academicachievement is the sole significant predictor of final academicachievement; and 2) when intelligence and prior academic achieve-ment are considered as two autonomous predictors of final academicachievement, the mediation model shows that final academicachievement is indirectly predicted by intelligence, yet only throughprior academic achievement. In other words, the inclusion of prioracademic achievement in the pathway between intelligence andfinal academic achievement produces an increase in the predictivevalidity of intelligence.

Several key-points can be derived from these findings. First, resultsshow that intelligence and academic achievement are two variableshighly associated. Not only the intelligence indicators (AR, VR and NR)and prior academic achievement (P7, E7, SC7 andMat7) are highly asso-ciated (ranging from .29 to .55), but also are the latent factor of intelli-gence and prior academic achievement (r = .73, p b .001). Severalstudies corroborate these results, identifying intelligence as a relevantpredictor of academic outcomes (e.g. Deary et al., 2007; Karbach et al.,2013; Lemos et al., 2013; Lynn & Vanhanen, 2012; Primi et al., 2010;Weber et al., 2013).

Second, regardless of the strong association (and an eventualcollinearity) between prior academic achievement and intelligence,results of ESEM analysis suggest that these variables are indepen-dent from each other, working as pure measures of intelligence andacademic achievement. However, even though this methodologyfrees cross-loadings, allowing indicators to load in both latent vari-ables, only Mat7 loads significantly in prior academic achievementand in intelligence. Contrarily, the latent factor of prior academicachievement is only significantly represented by students' academicclassifications. These results show that both latent variables (INTELand AA7) are autonomous variables and pure measures of Gf andacademic achievement. For that reason, intelligence and prioracademic achievement should be considered as two autonomouspredictors of final academic achievement.

Third, when intelligence and prior academic achievement are bothconsidered as predictors of final academic achievement (model 1),

only prior level of academic achievement is a direct predictor of subse-quent learning, explaining about 70% of its variance. However, when theindirect effect of intelligence on final academic achievement throughprior level of academic achievement (model 2) is tested, intelligenceonly has an indirect effect on final academic grades. When includingprior academic achievement in the pathway between intelligence andfinal academic achievement, a considerable increase in the achievedpredictive validity of intelligence is produced. Thus, throughout middleschool, intelligence remains a relevant predictor of academic achieve-ment, only when prior level of knowledge is also considered in theprediction model.

The mediation role played by prior academic achievement in thepathway between intelligence and final academic achievement issupported by the investment theory (Cattell, 1971). Accordingly, inearly stages of schooling (in the present study, 7th grade), intelligence,Gf, plays a more relevant role for learning, helping students to acquirenew skills and abilities, which are required to learn new academicsubjects and topics (Cattell, 1971; Primi et al., 2010). In turn, throughpractice and with school experiences, this potential is transformedinto knowledge, that is, crystallized intelligence (Schneider, 2013). Forthat reason, Gc includes not only developed abilities and knowledge,but is also a reflection of Gf (Kvist & Gustafsson, 2008). The investmenttheory and the fact that Gc is close to academic achievement and to Gfsupport our results. Indeed, final academic achievement is betterexplained when considering the mediated effect of intelligence byprior level of academic achievement. In sum, the relationship betweenintelligence and academic achievement throughout middle schoolemphasizes the role of students' prior academic achievement in thepathway between intelligence and final academic achievement.

5. Conclusion

This study brings a number of contributions to the educationalresearch. Firstly, this research provides a more comprehensive analysisof the relationship between Gf–Gc throughout time, validating assump-tions of the investment theory, and applying such assumptions to theeducational context.

Secondly, two theoretical models were tested, each one following aspecific analytic approach and innovative methodologies. We haveperformed testing models using only latent variables, which take intoaccount measurement error, hence, offering more reliable results.

Furthermore, this study makes a relevant contribution to psycho-logical practice. These results stress the relevance of a multidimen-sional assessment of students' cognitive abilities and learningdifficulties (Lee & Shute, 2010). The assessment should not onlyfocus on the information provided by standard intelligence tests,but also consider “the breadth and depth of knowledge of the lan-guage, information and concepts of a culture”, that is, Gc (Kvist &Gustafsson, 2008, p. 2). Actually, standard intelligence tests are

Fig. 2. Model of indirect effect from intelligence to final academic achievement, through prior academic achievement. INTEL = intelligence; AA7 = academic achievement in 7th grade(prior grade); AA9 = academic achievement in 9th grade (final grade); AR = abstract reasoning; VR = verbal reasoning; NR = numerical reasoning; P7 = Portuguese score in 7thgrade; E7 = English score in 7th grade; Mat7 = mathematics score in 7th grade; SC7 = sciences score in 7th grade; P9 = Portuguese score in 9th grade; E9 = English score in 9thgrade; Mat9 = mathematics score in 9th grade; SC9 = sciences score in 9th grade.

Table 3Comparison of model fit indexes and r2 of each model tested.

χ2 (DF) CMIN/DF RMSEA CFI SRMR r2

Model 1 101.35 (36) 2.81 .07 .97 .03 70%Model 2 74.40 (38) 1.96 .05 .98 .03 71%

5D.L. Soares et al. / Learning and Individual Differences xxx (2015) xxx–xxx

Please cite this article as: Soares, D.L., et al., The relationship between intelligence and academic achievement throughoutmiddle school: The roleof students' prior academic..., Learning and Individual Differences (2015), http://dx.doi.org/10.1016/j.lindif.2015.02.005

The relationship between intelligence and academic achievementthroughout middle school: The role of students' prioracademic performance

Diana Lopes Soares a,⁎, Gina C. Lemos a, Ricardo Primi b, Leandro S. Almeida a

a Centro de Investigação em Educação, Instituto de Educação, Universidade do Minho, Portugalb Departamento de Psicologia, Universidade de S. Francisco, Brazil

a b s t r a c ta r t i c l e i n f o

Article history:Received 4 April 2014Received in revised form 27 January 2015Accepted 18 February 2015Available online xxxx

Keywords:IntelligenceAcademic achievementMediation analysisInvestment TheoryGf–Gc

The association between intelligence and academic achievement is well established. However, how thisrelationship changes throughout schooling remains undefined. In this 3-year longitudinal study, 284Portuguese middle school students completed three reasoning subtests (abstract, numerical, and verbal)by the end of 7th grade (intelligence), and their academic grades were collected at the same time (prioracademic achievement, AA7) and by the end of 9th grade (final academic achievement, AA9). The mainfindings show that i) when intelligence and AA7 are analyzed as two independent predictors of AA9, AA9is best predicted by intelligence when considering the mediation effect of AA7, and ii) the inclusion ofAA7 in the pathway between intelligence and AA9 produces a considerable increase in the predictivevalidity of intelligence. Implications for cognitive assessment and psychological practice are emphasizedbased on this Gf–Gc relationship.

© 2015 Elsevier Inc. All rights reserved.

1. Introduction

An extensive body of research has been developed in order tounderstand the relationship between intelligence and achievementin different life domains, such as job performance (Gottfredson,2002; Schmidt & Hunter, 2004), health outcomes (Der, Batty, &Deary, 2009), or evenwealth (Zagorsky, 2007). Particularly in educa-tional settings, intelligence plays a crucial role in learning andacademic performance. Several studies show high correlationindices between them, ranging from .50 to .70 (Lynn & Vanhanen,2012). For instance, Mackintosh (1998) revealed a .50 correlationbetween 11-year-olds' intelligence scores and later educationalachievement at the age of 16. In addition, in a 5-year longitudinalstudy with 70,000 children, Deary, Strand, Smith, and Fernandes(2007) found a .81 correlation between intelligence at the age of 11and educational achievement at the age of 16 in 25 academicsubjects. Other studies also identify intelligence as a relevant predic-tor of academic achievement (Colom & Flores-Mendoza, 2007;Karbach, Gottschling, Spengler, Hegewald, & Spinath, 2013; Laidra,Pullmann, & Allik, 2007; Lemos, Abad, Almeida, & Colom, 2013;

Primi, Ferrão, & Almeida, 2010; Rohde & Thompson, 2007; Watkins,Lei, & Canivez, 2007; Weber, Lu, Shi, & Spinath, 2013).

General intelligence, also named fluid intelligence – Gf (Cattell,1971), is usually measured by administering tests of inductive anddeductive reasoning, which are assumed to reflect the ability to think,solve problems, make inferences, identify relations, and transforminformation in a significant way (Lemos, Almeida, & Colom, 2011;Nickerson, 2011). Longitudinal growth modeling attests that Gfpredicts, not only the initial level of math achievement, but also, therate of change in learning and achievement (Primi et al., 2010).

In turn, academic achievement is usuallymeasured by administeringtests to assess knowledge that is formally taught in schools. As a broadconcept, achievement could be associated with crystallized intelli-gence – Gc, which is defined as the “depth and breadth of knowledgethat are valued by one's culture” (Schneider & McGrew, 2012, p. 122).

Accordingly, Schneider (2013) claims that intelligence isrelated to potential, and achievement to the execution of potential.Although considered as separate abilities, both are viewed as twoaspects of the g factor according to Cattell's investment theory(1971). In light of Cattell's model, fluid intelligence is one of themain causes of achievement, since more capacity to learn predictsmore efficient and rapid learning. This potential is invested inexperiences, and is transformed into knowledge, that is, crystal-lized intelligence. In the process of transformation of potential(Gf) into fulfilled potential (Gc), other factors play a role, such as

Learning and Individual Differences xxx (2015) xxx–xxx

⁎ Corresponding author at: Centro de Investigação em Educação, Universidade doMinho, Campus de Gualtar, 4710-057 Braga, Portugal. Tel.: +351 927433844.

E-mail address: [email protected] (D.L. Soares).

LEAIND-01051; No of Pages 6

http://dx.doi.org/10.1016/j.lindif.2015.02.0051041-6080/© 2015 Elsevier Inc. All rights reserved.

Contents lists available at ScienceDirect

Learning and Individual Differences

j ourna l homepage: www.e lsev ie r .com/ locate / l ind i f

Please cite this article as: Soares, D.L., et al., The relationship between intelligence and academic achievement throughoutmiddle school: The roleof students' prior academic..., Learning and Individual Differences (2015), http://dx.doi.org/10.1016/j.lindif.2015.02.005

R I

R ML in g . V e rb .

G c

G f

R M G

S ó c . G e o g r.

N a t. M a t.

0 , 2 2

0 , 5 2

0 , 4 9

0 , 4 5

0 , 8 8

0 , 6 9

0 , 5 0

0 , 6 5

0 , 4 3

0 , 4 8

0 , 7 5

0 , 5 7

0 , 0 9 (t= 2 , 5 5 , p < 0 , 0 5 )

0 , 41 (t= 9 , 3 9 , p < 0 , 0 0 1 )

-‐ 1 , 0 4

0 , 8 2

0 , 7 7 C L

R L D D P

47Habilidades b·sicasEstudos de Psicologia 2002, 7(1), 47-55

Com as transformaÁıes pelas quais a sociedade tem

passado, cada vez mais tem sido apontado que ser·

necess·rio ao profissional das prÛximas geraÁıes,

n„o apenas o domÌnio de conhecimentos especÌficos, mas a

capacidade de se adaptar rapidamente e assimilar novas in-

formaÁıes de um mundo em constante transformaÁ„o. Isso

vem se tornando um prÈ-requisito b·sico para o perfil do

profissional do novo milÍnio. Portanto, a universidade deve-

r·, cada vez mais, produzir estratÈgias que privilegiem, n„o

sÛ a aprendizagem de conte˙dos, mas tambÈm a aprendiza-

gem de estratÈgias de adaptaÁ„o a situaÁıes novas.

Essa capacidade, definida pela psicologia como ìinteli-

gÍncia geralî, foi alvo de v·rias pesquisas desde o inÌcio

deste sÈculo (Almeida, 1988; Sternberg, 1981). Os estudos

psicomÈtricos, que investigaram como as habilidades huma-

nas se estruturam, chamaram-na de fator g, enquanto em

estudos mais recentes desenvolvidos pela psicologia

cognitiva, v·rios dos processos que poderiam ser classifica-

Habilidades b·sicas e desempenho acadÍmico em universit·rios

ingressantes

Ricardo PrimiAc·cia A. Angeli dos Santos.

Claudette Medeiros VendraminiUniversidade S„o Francisco

Resumo

Recentes estudos sobre o desenvolvimento cognitivo adulto referem-se ‡ distinÁ„o entre inteligÍncia

fluida como a capacidade geral de relacionar idÈias complexas, formar conceitos abstratos e derivar

implicaÁıes lÛgicas a partir de regras gerais e inteligÍncia cristalizada como a capacidade de derivar

conhecimento a partir de esquemas organizados de informaÁıes sobre disciplinas especÌficas. Para

verificar a possÌvel relaÁ„o entre a habilidade cognitiva requerida e a ·rea de conhecimento, este

estudo foi proposto com o objetivo de investigar as correlaÁıes entre medidas de inteligÍncia fluida e

cristalizada com desempenho acadÍmico em 960 alunos ingressantes dos cursos de Medicina, Odon-

tologia, Engenharia Civil, Matem·tica, Psicologia, Pedagogia, Letras e AdministraÁ„o. As correla-

Áıes encontradas indicam que o desempenho acadÍmico est· associado a diferentes perfis de habili-

dades cognitivas.

Palavras-chave: Habilidades b·sicas, InteligÍncia fluida e cristalizada, AvaliaÁ„o educacional, Teorias de inteligÍncia.

Abstract

Basic abilities and scholastic achievement in freshmen studentsRecent studies of adult cognitive development have distinguished two core abilities: fluid intelligence

as a general capacity to infer complex relationships, abstract concepts, and to deduce logical implications

from general rules; and crystallized intelligence as a capacity to solve problems by using organized

knowledge schemes from specific disciplines. In order to investigate a possible relationship between

cognitive abilities and area of study, this research studied the correlation between fluid and crystallized

intelligence measures and academic achievement among 960 freshmen students of eight areas of

study: Medicine, Dentistry, Psychology, Business, Engineering, Mathematics, Education, and

Literature. The results have indicated that the academic achievement is correlated with different

abilities profile.

Key words: Basic abilities, Fluid and crystallized intelligence, Educational assessment, Intelligence theories.

Meta análise de John B. Carroll

• Carroll (1993): marco histórico nas teorias psicométricas

• Revisou os principais estudos psicométricos dos últimos 60 anos (1500 referências)

• Refez a análise fatorial de 461 matrizes de correlação destes estudos

• Apresentou um modelo integrado de Três Camadas

• Modelo foi integrado na Teoria Cattell-Horn-Carroll (CHC) organizando a inteligência em Camadas 1. Fator g, II 10 Fatores Amplos e III 70 fatores específicos

Contemporary psychometric research has converged on the Cattell-Horn-Carroll (CHC) theory of cognitive abilities as the

consensus working taxonomy of human intelligence

McGrew, K. (2009). Editorial: CHC theory and the human cognitive abilities project: Standing on the shoulders of the giants of psychometric intelligence research, Intelligence, 37, 1-10.

Modelo Cattell-Horn-Carroll

g

Gf GqGc SARGsm Gv Ga TSR

Glm Gs CDS Grw

Gf Gc Gy Gv Gu Gr Gs Gt

A. Carroll Three-Stratum Model

B. Cattell-Horn Extended Gf-Gc Model

D. Tentatively identified Stratum II (broad)

Carroll and Cattell-Horn Broad Ability Correspondence (vertically-aligned ovals represent similar broad domains)

C. Cattell-Horn-Carroll (CHC) Integrated Model

Stratum III (general)

Stratum II (broad)

80+ Stratum I (narrow) abilities have been identified under the Stratum II broad abilities. They

are not listed here due to space limitations(see Table 1)

Gkn Gh Gk Go Gp Gps

D. Tentatively identified Stratum II (broad) domains

Gf GqGc Gsm Gv Ga Glr Gs Gt Grw

g

Gf Fluid reasoning Gkn General (domain-specific) knowledgeGc Comprehension-knowledge Gh Tactile abilitiesGsm Short-term memory Gk Kinesthetic abilitiesGv Visual processing Go Olfactory abilitiesGa Auditory processing Gp Psychomotor abilitiesGlr Long-term storage and retrieval Gps Psychomotor speedGs Cognitive processing speedGt Decision and reaction speed (see Table 1 for definitions)Grw Reading and writingGq Quantitative knowledge

CHC Broad (Stratum II) Ability Domains

(Missing g-to-broad ability arrows acknowledges that Carroll and Cattell-Horn disagreed on the validity of the general factor)

McGrew (2009), McGrew & Schneider (2012, 2013)

Glr Gsm Gf Gs Gt

General Speed Memory

Domain-Independent General Capacities

Acquired Knowledge

Gkn

Gq Grw

Gc Gp

Gh

Sensory Sensory-Motor Domain-Specific Abilities

Motor

Parameters of Cognitive Efficiency

Ga

Go

Gv

Gk

Gps

Conceptual Grouping

Functional Grouping

(From Schneider & McGrew, 2012)

Cattell-Horn-Carroll Theory of Cognitive Abilities

Knowledge

Speed

Perception

GkGoGh

GaGv

Motor

GpPsychomotor

Abilities

Controlled

Attention

Gf

Gsm

Fluid

Reasoning

Short-Term

Memory

GcGq

Grw

Gkn

GpsPsychomotor

Speed

GtSpeed

of Perception

GsAttentional

Fluency

GlrLearning Efficiency

& Retrieval Fluency

Nesse teste gostaríamos que você inventasse metáforas que complete as frases apresentadas. Veja o exemplo abaixo:

O camelo é o/a _________________ do deserto

Metáfora Explicação

1) barco No mar o barco é um meio de transporte que anda

balançando como o camelo no deserto

2) moto Porque a moto é um transporte para uma ou duas pessoas e

anda com pouco combustível como o camelo no deserto que precisa de pouca água

3) lesma Porque anda devagar, marcando o chão e rebolando como camelo

4) Barrichello

Porque quando não está parado está andando devagar

Glr: Fluência de ProduçãoDivergent Productions of Metaphors: Combining Many-Facet Rasch

Measurement and Cognitive Psychology in the Assessment of Creativity

Ricardo PrimiUniversity of São Francisco

This article presents a new method for the assessment of creativity in tasks such as “The camel is________ of the desert.” More specifically, the study uses Tourangeau and Sternberg’s (1981) domaininteraction model to produce an objective system for scoring metaphors produced by raters and themany-facet Rasch measurement to model the rating scale structure of the scoring points, item difficulty,and rater severity analysis, thus making it possible to have equated latent scores for subjects, regardlessof rater severity. This study also investigates 4 aspects of the method: reliability, correlation betweenquality and quantity, criterion validity, and correlation with fluid intelligence. The database analyzed inthis study consists of 12,418 responses to 9 items that were given by 975 persons. Two to 10 raters scoredthe quality and flexibility of each metaphor on a 4-point scale. Raters were counterbalanced in ajudge-linking network to permit the equating of different “test forms” implied in combinations of raters.The reliability of subjects’ latent quality scores was .88, and the correlation between quality and quantitywas low (r ! ".14), thus showing the desired separation between the 2 parameters established for thetask scores. The latent score on the test was significantly associated with the profession that requires ideaproduction (r ! .19), and the latent scores for the correlation between creativity and fluid intelligencewere high, # ! .51, even after controlling for crystalized intelligence (r ! .47). Mechanisms of fluidintelligence, executive function, and creativity are discussed.

Keywords: metaphor production, intelligence, creativity, item response theory, Rasch measurement

The assessment of intelligence has evolved immensely over the lastcentury (Schneider & McGrew, 2012), and there is now consensustaxonomy regarding the classification of cognitive factors and meth-ods of assessment that intelligence tests employ. The assessment ofcreativity, conversely, has not reached such a level of development.Recent works discuss methods, propose new ways to assess, andmake a call for the development of different methods to evaluatecreativity (Silvia et al., 2008). This article is a response to this call andpresents a new method for assessing the quality of metaphors pro-

duced for prompts such as “The camel is ________ of the desert”(example: boat). This study applies principles of cognitive psychologyto produce ratings that are potentially more objective and tied to theunderlying associative process that is believed to be the foundation ofidea production. It also illustrates the application of the psychometricmethod based on the Rasch model, which is useful when severaljudges rate different ideas generated by different people and therebyprovide parameters for simultaneously measuring creativity, itemdifficulty, and rater severity.

Creativity Assessment

One mode of creativity assessment is based on divergent pro-duction tasks that ask people to produce as many ideas as they canfrom a prompt. Examples of prompts include the following: “Listwords that start with ‘s’” (word fluency), “List opposites of theword ‘true’” (associational fluency), and “Create titles for a storyplot” (ideational fluency). These tasks are traced back to Guil-ford’s (1950, 1956, 1957) structure of intellect model, wherein hedefined four factors—fluency (number of ideas), flexibility (shiftin response category or approach to a problem), originality (un-conventionality, remote associations, and response cleverness),and elaboration (amount of details in a response)—that are used intests of creativity. These factors are measured using the widelyknown Torrance Tests of Creative Thinking (TTCT; Kim, 2006a,2011; Plucker, 1999).Some psychometric problems with divergent thinking tests have

been noted (Houtz & Krug, 1995; Plucker & Runco, 1998; Silvia

Ricardo Primi, Graduate Program of Psychology at Univesity of SãoFrancisco.This article was financed by the by the Brazilian National Research

Council (CNPq). The author acknowledges the contributions of DéboraPereira de Barros, Fabiano Koich Miguel, Fernando Pessotto, GleiberCouto, Lucas de Francisco Carvalho Maria Beatriz Zanarella Cruz, Mar-jorie Cristina Rocha da Silva, Marta Petrini, Monalisa Muniz, PriscillaRodrigues Santana, Sanyo Drummond Pires, Tatiana de Cássia Nakano,and Tatiana Freitas da Cunha, who were graduate students working underthe supervision of the author and who contributed in data collection and asraters of the metaphors.Correspondence concerning this article should be addressed to Ricardo

Primi, Universidade São Francisco, Laboratório de Avaliação Psicológicae Educacional [Laboratory of Psychological and Educational Assessment],Rua Alexandre Rodrigues Barbosa, 45, CEP 13251-900, Itatiba, São Paulo,Brazil. E-mail: [email protected]

ThisdocumentiscopyrightedbytheAmericanPsychologicalAssociationoroneofitsalliedpublishers.

Thisarticleisintendedsolelyforthepersonaluseoftheindividualuserandisnottobedisseminatedbroadly.

Psychology of Aesthetics, Creativity, and the Arts © 2014 American Psychological Association2014, Vol. 8, No. 4, 000 1931-3896/14/$12.00 http://dx.doi.org/10.1037/a0038055

1


Knowledge

Speed

Perception

GkGoGh

GaGv

Motor

GpPsychomotor

Abilities

Controlled

Attention

Gf

Gsm

Fluid

Reasoning

Short-Term

Memory

GcGq

Grw

Gkn

GpsPsychomotor

Speed

GtSpeed

of Perception

GsAttentional

Fluency


& Retrieval Fluency

Gf: Raciocínio, Memória de Trabalho e Funções executivas

Procedure

Data were collected for the present study as part of a basicabilities-assessment program for freshmen students that wasfinanced by the university. The goal of the program was toinvestigate the reasoning and basic knowledge abilities of first-year students entering college. The program included the as-sessment of logical reasoning, reading comprehension, generalliteracy, moral judgment, and fluid intelligence. The partici-pants were volunteers, and the testing was part of their regularacademic activities. The participants completed the tests in twosections that lasted an average of 90 minutes each. After anal-ysis of the data, the participants received a personal report thatinformed them of their scores and explained what the scoresmeant.

Test Development

A prototype task used for the measurement of fluid intelli-gence involves geometric inductive matrix items (Marshalek etal., 1983), such as Raven’s Progressive Matrices, which isexemplified in the upper half of Figure 1. In earlier studies,Primi (1995, 1998, 2002) and Primi, Cruz, Muniz, and Petrini(2006) experimentally manipulated matrix item variables toinvestigate the impact of these features on item complexity.These earlier studies provided detailed examples of each itemfeature and an explanation of the link with fluid intelligencecapacities. In summary, three main features were identified thatwere considered the sources of item complexity in geometricinductive matrix problems: (a) the amount of information due tothe increase in the number of elements or in the number of rules

Figure 1. Examples of fluid intelligence items used in the present study and a summary of sources ofcomplexity for geometric matrix items and their link with fluid intelligence capacities.



4 PRIMI

Developing a Fluid Intelligence Scale Through a Combination of RaschModeling and Cognitive Psychology

Ricardo PrimiUniversity São Francisco

Ability testing has been criticized because understanding of the construct being assessed is incompleteand because the testing has not yet been satisfactorily improved in accordance with new knowledge fromcognitive psychology. This article contributes to the solution of this problem through the application ofitem response theory and Susan Embretson’s cognitive design system for test development in thedevelopment of a fluid intelligence scale. This study is based on findings from cognitive psychology;instead of focusing on the development of a test, it focuses on the definition of a variable for the creationof a criterion-referenced measure for fluid intelligence. A geometric matrix item bank with 26 items wasanalyzed with data from 2,797 undergraduate students. The main result was a criterion-referenced scalethat was based on information from item features that were linked to cognitive components, such asstorage capacity, goal management, and abstraction; this information was used to create the descriptionsof selected levels of a fluid intelligence scale. The scale proposed that the levels of fluid intelligencerange from the ability to solve problems containing a limited number of bits of information with obviousrelationships through the ability to solve problems that involve abstract relationships under conditionsthat are confounded with an information overload and distraction by mixed noise. This scale can beemployed in future research to provide interpretations for the measurements of the cognitive processesmastered and the types of difficulty experienced by examinees.

Keywords: inductive reasoning, fluid intelligence, Rasch measurement, matrix reasoning

There has been significant development in cognitive psychologyand the psychometrics of intelligence testing over the last fewdecades (Cornoldi, 2006; Deary, 2001; Whitely, 1980; Whitely &Schneider, 1981). Cognitive task analysis of items that are com-monly used in intelligence tests provides a better understanding ofhow people represent and process information, which in turnimproves a test’s score for construct validity (Sternberg, 1981).Psychometric model-based methods, including item response the-ory (IRT), provide ways to construct scales that show links be-tween test scores and the underlying construct that provide sub-stantial additional interpretations (Embretson, 2006; Wilson,2005). These methods have recently evolved into cognitive diag-nostic assessment models such as that of Tatsuoka (2009).

Despite these developments, test construction procedure has notyet satisfactorily incorporated these new methods (Embretson,1994). Thus, the expectation for the future generation of tests isthat cognitive psychology should play an important role in con-struct representation. By providing a rich theoretical basis for thecreation of purified tasks, it should lead to the development ofinstruments with better theoretical grounding. This article presentsan illustration of these methods as they apply to the developmentof a fluid reasoning test. First, a review of the psychometric andcognitive neuroscience definitions of fluid intelligence is pre-sented. A brief review of how fluid intelligence is measuredfollows, with emphasis on the justification for new tests based onmodern methodology. Finally, an empirical study of test develop-ment is presented.

Nature of Fluid Intelligence

According to Schneider and McGrew (2012), fluid reasoning(Gf) refers to

the deliberate but flexible control of attention to solve novel “on thespot” problems that cannot be performed by relying exclusively onpreviously learned habits, schemas and scripts. Fluid reasoning is amulti-dimensional construct, but its parts are unified in their purpose:solving unfamiliar problems. Fluid reasoning is most evident in ab-stract reasoning that depends less on prior learning. However, it is alsopresent in day-to-day problem solving. Fluid reasoning is typicallyemployed in concert with background knowledge and automatizedresponses. (p. 111)

Fluid intelligence is central to understanding the construct ofintelligence. As it is the broad factor most related to the general

This article was developed as part of a larger project, called Develop-ment of a Computerized Componential Fluid Intelligence Test, which wasfinanced by the Foundation for the Support of Research in the State of SãoPaulo (FAPESP, Process 2000/05913-4), by the Brazilian National Re-search Council (CNPq), and by the University of São Francisco. The authoracknowledges the contributions of Acácia A. Angeli dos Santos, ClaudetteMedeiros Vendramini, Fernanda de Oliveira Soares Taxa, Maria de FátimaLukjanenko, Franz A. Müller, Isabel Sampaio, Fatima Keiko Kuse, andCíntia Heloína Bueno for their valuable comments on drafts of this articleand assistance in collection of the data.Correspondence concerning this article should be addressed to Ricardo

Primi, Universidade São Francisco, Laboratório de Avaliação Psicológicae Educacional, Rua Alexandre Rodrigues Barbosa, 45, CEP 13251-900,Itatiba, São Paulo, Brazil. E-mail: [email protected]



Psychological Assessment © 2014 American Psychological Association2014, Vol. 26, No. 2, 000 1040-3590/14/$12.00 http://dx.doi.org/10.1037/a0036712

1

Working Memory Capacity and Fluid Intelligence Are Strongly RelatedConstructs: Comment on Ackerman, Beier, and Boyle (2005)

Michael J. KaneUniversity of North Carolina at Greensboro

David Z. HambrickMichigan State University

Andrew R. A. ConwayUniversity of Illinois at Chicago

The authors agree with P. L. Ackerman, M. E. Beier, and M. O. Boyle (2005) that working memorycapacity (WMC) is not isomorphic with general fluid intelligence (Gf) or reasoning ability. However, theWMC and Gf/reasoning constructs are more strongly associated than Ackerman et al. (2005) indicate,particularly when considering the outcomes of latent-variable studies. The authors’ reanalysis of 14 suchdata sets from 10 published studies, representing more than 3,100 young-adult subjects, suggests a strongcorrelation between WMC and Gf/reasoning factors (median r ! .72), indicating that the WMC and Gfconstructs share approximately 50% of their variance. This comment also clarifies the authors’ “exec-utive attention” view of WMC, it demonstrates that WMC has greater discriminant validity thanAckerman et al. (2005) implied, and it suggests some future directions and challenges for the scientificstudy of the convergence of WMC, attention control, and intelligence.

Theorists have recently speculated that individual differences inworking memory capacity (WMC) may explain reasoning ability,general intelligence (Spearman’s g), or both (e.g., Engle, 2002;Kyllonen & Christal, 1990; Suß, Oberauer, Wittmann, Wilhelm, &Schulze, 2002). In response to these claims, Ackerman, Beier, andBoyle (2005) conducted a critical review and meta-analysis of theliterature and rejected the notion that WMC is isomorphic to g,general fluid intelligence (Gf), or any other cognitive ability.Indeed, they remained generally skeptical that WMC researchwould significantly advance either the understanding or measure-ment of human intelligence. Perhaps it is surprising, then, that wefind several points of agreement with Ackerman et al. (2005),despite our position that WMC and Gf are closely related and thatexploring this relationship will yield important theoretical insightsinto the nature of normal cognitive variation (e.g., Conway, Kane,& Engle, 2003; Engle & Kane, 2004; Engle, Kane, & Tuholski,1999; Engle, Tuholski, Laughlin, & Conway, 1999; Kane & Engle,2002).

We concur with Ackerman et al. (2005) on three main conclu-sions: (a) WMC is not equivalent to g, Gf, or reasoning ability; (b)

WMC is primarily a domain-general construct; (c) WMC is moreclosely related to Gf and reasoning than is short-term memory(STM). The latter two arguments are now well supported bylarge-sample, latent-variable studies. WMC tasks correlatestrongly with each other and with a wide range of cognitiveabilities regardless of their surface characteristics, and WMC andGf share substantial variance that is independent of STM (forreviews, see Daneman & Merikle, 1996; Kane et al., 2004). Re-garding the central issue of a WMC–intelligence isomorphism,Ackerman et al.’s (2005) meta-analysis suggests that individualtests of WMC and Gf (or reasoning) share only about 20% of theirvariance, on average, and so these constructs are not synonymous(see also Ackerman, Beier, & Boyle, 2002). We have similarlyargued in recent reviews of latent-variable studies that the corre-lations between WMC and Gf constructs do not approach 1.0, evenwhen both are measured broadly, and so speculations about iso-morphism are falsified (Conway et al., 2003; Kane et al., 2004).

Nonetheless, we also find major points of disagreement withAckerman et al. (2005) and discuss them below. Namely, when weconsider the data from latent-variable studies, we find that theWMC–Gf association is much stronger than they suggested. WMCalso shows greater discriminant validity than Ackerman et al.(2005) admitted, correlating more strongly with some constructsthan with others, and so WMC represents much more than a “crudfactor” (p. 52) in its association with other cognitive abilities.Finally, the link between WMC and attention control is potentiallyimportant to a complete understanding of intelligence, but it alsopresents considerable challenges for future theoretical and empir-ical work that we discuss—particularly regarding measurement.To put our critique in proper context, however, we first present anoverview of our theoretical perspective. We also hope that this

Michael J. Kane, Department of Psychology, University of North Caro-lina at Greensboro; David Z. Hambrick, Department of Psychology, Mich-igan State University; Andrew R. A. Conway, Department of Psychology,University of Illinois at Chicago.

Andrew R. A. Conway is now at the Department of Psychology,Princeton University.

We are grateful to Klaus Oberauer for making the complete correlationmatrix available from Suß et al. (2002).

Correspondence concerning this article should be addressed to MichaelJ. Kane, Department of Psychology, University of North Carolina atGreensboro, P. O. Box 26170, Greensboro, NC 27402-6170. E-mail:[email protected]

Psychological Bulletin Copyright 2005 by the American Psychological Association2005, Vol. 131, No. 1, 66–71 0033-2909/05/$12.00 DOI: 10.1037/0033-2909.131.1.66

66

the loading was .40, for Quantitative Reasoning the loading was.51, for Lexical Knowledge (Verbal) the loading was .37, and forMemory Span, the loading on g was .36. In the current SEManalysis, all of these loadings of the first-order factors on g were

larger (by more than .3 to .4) than what the extant literaturesupports, suggesting that the true associations between these latentvariables and g may be substantially lower than that shown in thecurrent SEM analysis. There are two highly salient reasons for thediscrepancy between the current SEM results and those of Car-roll’s (1993) analysis. First, Carroll (1993) did not adjust the rawcorrelations in his data sets for attenuation. Although such correc-tions are common in meta-analytic research, there is some contro-versy about whether these corrections are appropriate, becausethey “confound issues of validity and reliability” and because“equating true scores with construct scores is logically inconsistentwith the classical test theory model itself” (Boorsboom & Mellen-bergh, 2002, p. 505). Second, because Carroll (1993) did not useSEM procedures, but instead used classic principal axis factoring,followed by Schmid-Leiman rotations, the correlations he reportedare not confounded by the degree of model fit inherent in SEMresults reported here.

Discussion

As we noted earlier, SEM with meta-analytically derived cor-relations (i.e., where different correlations are based on differentmeasures, different samples, and different numbers of participants)is not optimal from a psychometric perspective. Taking the SEMresults with some degree of skepticism is thus in order. Nonethe-less, neither SEM solution points to an indication that WM and gare isomorphic to one another. We were able to reject a model inwhich WM measures and ability measures were captured by asingle factor in favor of a model with separate g and WM factors—though the estimated correlation between WM and g latent factorsin this model was .50. The initial model (see Figure 1) seems toprovide a representation that is more concordant with the existingabilities literature, given the presence of both a higher order

Figure 1. A hierarchical factor analytic representation of g and first-orderfactors, based on a structural equation model of meta-analytically derivedcorrelations among ability and working memory (WM) measures.

Figure 2. A nonhierarchical representation of g and working memory (WM) factors, based on a structuralequation model of meta-analytically derived correlations among ability and WM measures.

43WORKING MEMORY AND INTELLIGENCE

Working Memory and Intelligence: The Same or Different Constructs?

Phillip L. Ackerman, Margaret E. Beier, and Mary O. BoyleGeorgia Institute of Technology

Several investigators have claimed over the past decade that working memory (WM) and generalintelligence (g) are identical, or nearly identical, constructs, from an individual-differences perspective.Although memory measures are commonly included in intelligence tests, and memory abilities areincluded in theories of intelligence, the identity between WM and intelligence has not been evaluatedcomprehensively. The authors conducted a meta-analysis of 86 samples that relate WM to intelligence.The average correlation between true-score estimates of WM and g is substantially less than unity (! !.479). The authors also focus on the distinction between short-term memory and WM with respect tointelligence with a supplemental meta-analysis. The authors discuss how consideration of psychometricand theoretical perspectives better informs the discussion of WM–intelligence relations.

Since the 1980s, with the major theoretical and empirical de-velopments of the construct of working memory (WM; see, e.g.,Baddeley, 1986; Richardson, 1996, for reviews) as distinct fromrote or span memory (which is usually referred to as short-termmemory [STM]), several investigators have asserted that WM andintellectual abilities are highly related or identical constructs.These assertions started with demonstrations that significant cor-relations were found between some measures of WM and mea-sures of comprehension (Daneman & Carpenter, 1980), and laterbetween WM and reasoning ability (Kyllonen & Christal, 1990),and other measures, such as the SAT (e.g., Turner & Engle, 1989).Recently, several investigators have claimed that WM and generalintelligence (g; or general fluid intelligence, Gf) are essentially thesame constructs. For example:

So central is the role of WM capacity in individual differences ininformation processing that some cognitive theorists equate WMcapacity with g itself. (Jensen, 1998, p. 221)

Stauffer et al. (1996) found a correlation of " 0.995 between a factorrepresenting general intelligence (g) and a factor representing WM.(Colom, Flores-Mendoza, & Rebollo, 2003, p. 34)1

My colleagues and I used a structural equation modeling analysis totest this and the idea that the construct measured by WM-capacitytasks is closely associated with general fluid intelligence. . . . WM-capacity tasks measure a construct fundamentally important to higher-

level cognition. The construct is distinguishable from STM and is atleast related to, maybe isomorphic to, general fluid intelligence andexecutive attention. (Engle, 2002, pp. 21–22)

No other cognitive factor—knowledge, speed, or learning ability—correlated with g after the working memory factor was partialed out.Thus, we have our answer to the question of what g is. It is workingmemory capacity. (Kyllonen, 2002, p. 433; see also Kyllonen, 1996)

However, the relationship between memory and intelligenceappears to be much more complex than has been asserted by theseinvestigators. Note that this position is not without its critics. Forexample, differential psychologists such as Deary (2000) andKline (2000) have expressed substantial skepticism that WM andgeneral intelligence are even closely linked. In this article, weevaluate these claims in the context of a meta-analysis of correla-tions between WM measures and intellectual ability measures. Inaddition, it is important to note that intelligence theory and theassessment of intelligence have both involved memory abilitiesover the past 110 years. Although models of WM represent rela-tively recent developments in the history of the science, an under-standing of the construct space for individual differences in intel-ligence and WM benefits from a brief review of intelligence andmemory ability research. Thus, we begin with a consideration ofmemory abilities from an intelligence assessment perspective,followed by a review of research on memory abilities and intelli-gence theory. Next, we briefly review memory theory and theunderlying framework for asserting the overlap between WM andintelligence. A meta-analysis of 86 samples that report correlationsbetween measures of WM and measures of intellectual abilities isthen presented. A parallel set of analyses is also provided for STMand intelligence for comparison to the WM–intelligence relations.We then discuss the implications of the meta-analytic results in thecontext of both enduring psychometric measurement and theoryissues.

1 It is interesting to note that in fact the original Stauffer, Ree, andCaretta (1996) article does not state this conclusion. The higher order factorthat Colom et al. (2003) identified as “WM” is actually identified with fourlower order factors: Processing Speed, Working Memory, DeclarativeKnowledge, and Procedural Knowledge, not just with WM.

Phillip L. Ackerman, Margaret E. Beier, and Mary O. Boyle, School ofPsychology, Georgia Institute of Technology.

Margaret E. Beier is now at the School of Psychology, Rice University.This research was partially supported by Air Force Office of Scientific

Research Grant F49620-01-0170 (Phillip L. Ackerman, principal investi-gator). We thank the following individuals for providing correlations fromtheir empirical research: Scott Chaiken, Sandra Hale, Tim Salthouse, LeeSwanson, Bill Tirre, and Werner Wittmann. We also thank Joni Lakin andSarah Molouki for their assistance in obtaining correlations and data entryand processing.

Correspondence concerning this article should be addressed to Phillip L.Ackerman, School of Psychology, MC 0170, Georgia Institute of Tech-nology, 654 Cherry Street, Atlanta, GA 30332-0170. E-mail:[email protected]

Psychological Bulletin Copyright 2005 by the American Psychological Association2005, Vol. 131, No. 1, 30–60 0033-2909/05/$12.00 DOI: 10.1037/0033-2909.131.1.30

30

Kane, M.J., Conway, A.R.A., Hambrick, D.Z., & Engle, R.W. (2007). Variation in working memory capacity as variation in executive attention and control. In A.R.A. Conway, C.Jarrold, M.J. Kane, A. Miyake, and J.N. Towse (Eds.), Variation in Working Memory (pp. 21 - 48). NY: Oxford University Press.


Knowledge

Speed

Perception

GkGoGh

GaGv

Motor

GpPsychomotor

Abilities

Controlled

Attention

Gf

Gsm

Fluid

Reasoning

Short-Term

Memory

GcGq

Grw

Gkn

GpsPsychomotor

Speed

GtSpeed

of Perception

GsAttentional

Fluency


& Retrieval Fluency

Spatial ability and STEM: A sleeping giant for talent identification and development

David Lubinski *

Department of Psychology and Human Development, Peabody 0552, 230 Appleton Place, Vanderbilt University, Nashville, TN 37203, United States.

a r t i c l e i n f o

Article history:Received 27 January 2010Received in revised form 10 March 2010Accepted 18 March 2010Available online 14 April 2010

Keywords:Spatial abilitySTEMGiftedTalent developmentTalent searches

a b s t r a c t

Spatial ability is a powerful systematic source of individual differences that has been neglected in com-plex learning and work settings; it has also been neglected in modeling the development of expertise andcreative accomplishments. Nevertheless, over 50 years of longitudinal research documents the importantrole that spatial ability plays in educational and occupational settings wherein sophisticated reasoningwith figures, patterns, and shapes is essential. Given the contemporary push for developing STEM (sci-ence, technology, engineering, and mathematics) talent in the information age, an opportunity is avail-able to highlight the psychological significance of spatial ability. Doing so is likely to inform researchon aptitude-by-treatment interactions and Underwood’s (1975) idea to utilize individual differences asa crucible for theory construction. Incorporating spatial ability in talent identification procedures foradvanced learning opportunities uncovers an under-utilized pool of talent for meeting the complex needsof an ever-growing technological world; furthermore, selecting students for advanced learning opportu-nities in STEM without considering spatial ability might be iatrogenic.

! 2010 Elsevier Ltd. All rights reserved.

In celebrating the distinguished career of Thomas J. Bouchard,Jr., it seems fitting to recall a vivid memory of what it was like sit-ting in on Bouchard’s course in individual differences (or ‘‘IDs”),first as a student and, subsequently, as his graduate teaching assis-tant for the same course. Twice a week, students would listen toBouchard analyze empirical studies and break them apart bone-by-bone. Many would appear numb as their deeply held supposi-tions about human behavior were found to have little empiricalbasis. Like Donald G. Paterson’s course in IDs, which prior genera-tions of Minnesota graduate students experienced, students leftBouchard’s course with new perspectives. Students having theprivilege of experiencing one of these two intellectual giants – overthe 80-year period they successively taught IDs at Minnesota –could never again look at human behavior through the same lens.Their newly acquired knowledge base forever changed how theyviewed the human condition and the complex tensions surround-ing social order, liberty, and individual differences (Wells, 1937).

One of the more memorable things about Bouchard’s coursewas that, occasionally, after presenting a compelling empiricaldemonstration to the class, on the power that psychological vari-ables can hold for predicting important socially-valued outcomes(educational achievements, occupational accomplishments, or lifein general), Bouchard would turn to the class and say: ‘‘See, see,see what happens when psychologists choose to study real vari-ables.” This point of view was not unrelated to that of Bouchard’scolleague Paul E. Meehl; Meehl would occasionally wonder out

loud whether it would be scientifically prophylactic to requiregraduate students in psychology to take a minor in a natural sci-ence like biology or genetics. By going so, Meehl speculated, theymight be able to recognize a meaningful scientific contribution, ifthey should ever happen to encounter one in psychology! Thesetwo anecdotes set the stage for this contribution, regarding oneof Bouchard’s favorite psychological variables, spatial ability. Fordecades, spatial ability has surfaced as a salient characteristic ofyoung adolescents who go onto develop expertise in science, tech-nology, engineering, and mathematics (STEM), yet applied psy-chologists and talent development researchers have failed tofully recognize this important dimension of human individuality.1

Contemporary discourse on the importance of identifying and nur-turing STEM talent affords an opportunity to correct for this ne-glected ‘‘real variable.”

1. Spatial ability and STEM: decades of longitudinal research

For example, during the same year Sputnik was launched, a lit-tle known report, Scientific Careers, was published on the psycho-logical characteristics of individuals harboring STEM talent. Thisreport was based on a NSF committee chaired by Donald Super,who assembled a distinguished team of psychologists (Harold

0191-8869/$ - see front matter ! 2010 Elsevier Ltd. All rights reserved.doi:10.1016/j.paid.2010.03.022

* Tel.: +1 615 343 1195.E-mail address: [email protected]

1 The referent generality of spatial ability extends well beyond STEM domains andencompasses among other things the creative arts in particular (Humphreys,Lubinski, & Yao, 1993). This article, however, will be restricted to STEM domains.For further reading on the educational and psychological significance of spatialability, see Lohman (1988, 1994, 1996).

Personality and Individual Differences 49 (2010) 344–351

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier .com/locate /paid

Papinsky, Anne Roe, Leona Tyler, and others) to conduct a literaturereview and issue recommendations for future research and theoryconstruction on identifying and developing exceptional careers inengineering and the physical sciences. Their report is impressivefor many reasons. First, many of their conclusions and recommen-dations (Super & Bachrach, 1957), based on the empirical evidenceavailable at the time, were supported by subsequent longitudinalfindings based on normative samples (Austin & Hanisch, 1990;Gohm, Humphreys, & Yao, 1998; Humphreys & Yao, 2002; Humphreyset al., 1993) and on samples of intellectually precocious youth(Shea, Lubinski, & Benbow, 2001; Wai, Lubinski, & Benbow, 2009;Webb, Lubinski, & Benbow, 2007). Like Smith (1964) comprehen-sive review of spatial ability, Super and Bachrach (1957) documentthat exceptional general intellectual potential is characteristic ofengineers and physical scientists at early ages; however, they wentonto stress that specific abilities – especially mathematical reason-ing and spatial ability – are also salient features of their individu-ality. Moreover, they also noted the importance of scientificinterests, and called for additional longitudinal research, for in-stance, following young adolescents over 10- to 15-years for ascer-taining how these and other personal attributes, and contrastingopportunities and supports, factor into differential development.

During the years between Super and Bachrach (1957) and Smith(1964), John C. Flanagan et al. (1962) launched Project Talent,which was expressly the kind of longitudinal study that Super’sNSF team envisioned. Because of its comprehensiveness and size,longitudinal findings from Project TALENT are among the mostcompelling for illustrating the role that spatial ability plays indeveloping expertise in STEM. Project TALENT’s initial data collec-

tion occurred in 1960, and consisted of a stratified random sampleof the USA’s high school population. Students in the 9th through12th grades were assessed on a wide range of tests and question-naires over a one-week period, and the entire sample includedroughly 50,000 males and 50,000 females per grade level, for a to-tal N of approximately 400,000. Included in the tests were a num-ber of measures designed to assess cognitive abilities (e.g., generalintelligence and specific abilities: mathematical, verbal, and spatialreasoning). Project TALENT also included longitudinal data taken 1,5 and 11 years after graduation from high school (Wise, McLaughlin,& Steel, 1979). A number of longitudinal studies based on ProjectTALENT’s 11-year follow-up underscore the importance of spatialability for accomplishments in STEM disciplines (Austin & Hanisch,1990; Gohm et al., 1998; Humphreys & Lubinski, 1996; Humphreys& Yao, 2002; Humphreys et al., 1993). A recent study comparingthese data to modern longitudinal findings from the Study ofMathematically Precocious Youth (SMPY; Lubinski & Benbow,2006), is especially relevant to understanding the development ofSTEM talent (Wai et al., 2009).

Wai et al. (2009) focused on participants’ highest degree re-ceived (bachelor’s, master’s, or doctorate), the disciplines withinwhich their degrees were earned, and their occupations as a func-tion of general (‘‘g”) and specific (mathematical, spatial, and ver-bal) abilities. Fig. 1 graphs the general and specific ability profilesof Project Talent participants earning terminal degrees in variousdisciplines. Because highly congruent findings were observed forall four cohorts, grades 9–12, the cohorts were combined. Highgeneral intelligence and an intellectual orientation dominated byhigh mathematical and spatial abilities, relative to verbal ability,

Fig. 1. Average z-scores of participants on spatial, math, and verbal ability for bachelor’s degrees, master’s degrees, and PhDs are plotted by field. The groups are plotted inrank order of their normative standing on g (S + M + V) along the x-axis and the line with the arrows from each field pointing to it indicates on the continuous scale wherethey are in general mental ability. This figure is standardized in relation to all participants with complete ability data at the time of initial testing. Respective N’s for eachgroup (Males + Females) were (for bachelor’s, master’s, and doctorates respectively): engineering (1143, 339, and 71), physical science (633, 182, and 202), math/computerscience (877, 266, and 57), biological science (740, 182, and 79), humanities (3226, 695, and 82), social science (2609, 484, and 158), arts (615, M = 171), business (2386,M + D = 191), and education (3403, M + D = 1505). *For education and business, masters and doctorates were combined because the doctorate samples for these groups weretoo small to obtain stability (N < 30). From Wai et al. (2009).

D. Lubinski / Personality and Individual Differences 49 (2010) 344–351 345

Síntese das abordagens educacional (conhecimento) e psicológica (processos mentais)

Conclusões …

Conclusões• Confusões atuais ao se tratar das competências cognitivas:

• Tratadas com um único rótulo na literatura "QI"

• Habilidades cognitivas incluídas nos sistemas em larga escala: Gc, Grw e Gq

• Comp do sec XXI enfatizam outros fatores amplos (no âmbito cognitivo):

• Pensamento crítico e solução de problemas: Gf, Gv

• Criatividade inovação: Glr Gf

• Variáveis socioemocionais: O, C, A, N-

• O modelo CHC é importante para mapear quais competências cognitivas estamos avaliando nos sistemas em larga escala

• Será que deveríamos monitorar outras capacidades além daquelas já presentes nos sistemas em larga escala ? Qual seria a utilidade de tal sistema ?

• Temas para Parte 2: Maleabilidade de g e WM, integração entre modelo CHC e a neuropsicologia

Obrigado! [email protected]

mailto:[email protected]

II RNCE IAS abril 2015 copy - Educação para o Século...

Documents

Transcript of II RNCE IAS abril 2015 copy - Educação para o Século...