Discussion on ‘The g-Factor of International Cognitive Ability Comparisons: The Homogeneity of Results in PISA, TIMSS, PIRLS and IQ-Tests Across Nations’ by Heiner Rindermann


Abstract

Geographical Distribution of Mental Abilities and Its Moral Consequences

JÜRI ALLIK

Department of Psychology, University of Tartu, Estonia

Juri.Allik@ut.ee

Rindermann's study provides the most comprehensive evidence so far that national scores of school assessment have systematic differences and the geographical distribution of these differences almost perfectly repeat the distribution of the mean national scores of intelligence. It is argued that without comparison with the random effects of statistical aggregation it is impossible to decide whether additional factors are needed to explain the strong association between national scores of school assessment and intelligence tests. The ignorance about real differences in mental abilities may become a source of social injustice because this does not allow natural inequalities to be arranged such that they are to the greatest benefit of the least advantageous. Copyright © 2007 John Wiley & Sons, Ltd.

The Big G-Factor of National Cognitive-Ability Comparisons: Not Trivial and Not Immutable

JENS B. ASENDORPF

Department of Psychology, Humboldt University Berlin, Germany

jens.asendorpf@rz.hu-berlin.de

Rindermann's analysis identifies a large first factor of cross-national differences, with high loadings of both IQ tests and student achievement tests. This finding is not trivial because correlations at different levels of analysis such as individuals and nations can be very different. The finding should be seriously discussed rather than downplayed because of the enormous political implications if the finding is misinterpreted as evidence for immutable national or regional differences. Copyright © 2007 John Wiley & Sons, Ltd.

The Different Levels of the g-Factor

ROEL J. BOSKER

Institute for Educational Research, University of Groningen, The Netherlands

r.j.bosker@rug.nl

In search for a country-level g-factor, aggregated data were used by Rindermann. This, however, can cause some problems, well-known in the literature on multi-level modelling. Conceptual clarity of the country-level g-factor is lacking, the ‘blown up’ coefficients are well-known in studies on using aggregated data, and are not that remarkable, and the quest for causes and effects of the country-level g-factor is only meaningful if we understand what this factor stands for, and if we are able to analyse data in which the multi-level structure is preserved. Copyright © 2007 John Wiley & Sons, Ltd.

Not Every g is g

MARTIN BRUNNER and ROMAIN MARTIN

Research Unit for Educational Measurement and Applied Cognitive Science, University of Luxembourg, Luxembourg

martin.brunner@emacs.lu; romain.martin@emacs.lu

The target paper identifies a common factor underlying measures of intelligence and student achievement on the cross-national level. Given the level of analysis applied, however, this factor cannot be interpreted as general cognitive ability (g). Rather, it is an indicator of a nation's prosperity. g operates at the individual level and not at the cross-national level. Copyright © 2007 John Wiley & Sons, Ltd.

Little g: Prospects and Constraints

STEPHEN J. CECI and WENDY M. WILLIAMS

Department of Human Development, Cornell University, USA

sjc9@cornell.edu; wmw5@cornell.edu

Rindermann's analyses reveal substantial correlations for intelligence tests in multiple nations, across content domains and over time. Although impressive and supportive of g-theory, high correlations do not necessarily reflect immutability of g over time or high heritability. Multiple studies demonstrate strong training effects for g-loaded tasks, and tendencies for general intelligence to vary by country may reflect resource and experiential factor differences more than heredity. Copyright © 2007 John Wiley & Sons, Ltd.

A g-Factor of International Cognitive-Ability Comparisons: What Else?

FILIP DE FRUYT

Department of Developmental, Personality and Social Psychology, Ghent University, Belgium

Filip.DeFruyt@ugent.be

Rindermann showed that student assessment means across countries are strongly correlated with intelligence means. Potential reasons for this strong relationship are discussed, and alternatives for student assessments are considered and evaluated. Copyright © 2007 John Wiley & Sons, Ltd.

The g-Factor Is Alive and Well But What Shall We Do About It?

ANDREAS DEMETRIOU

University of Cyprus and Cyprus University of Technology, Cyprus

ademetriou@ucy.ac.cy

This commentary first demonstrates, in agreement with the target paper, that g exists, and it specifies the cognitive dimensions involved in it. It then argues that g is malleable and plastic and specifies how it can be increased. It is also maintained that rank ordering groups along a dimension of g is possible but difficult to achieve and that the present international studies depart from this ideal. Finally, it is argued that we need policies and programmes for the cultivation of g in the best interests of everybody. Copyright © 2007 John Wiley & Sons, Ltd.

What Lies Behind g(I) and g(ID)

JAMES R. FLYNN

Emeritus Professor, Political Studies, University of Otago, New Zealand

jim.flynn@stonebow.otago.ac.nz

Rindermann's results suggest that different factors lie behind the emergence of g in international comparisons and the emergence of g when we compare the differential performance of individuals. This renders g(I) and g(ID) so unlike that they have little significance in common. Copyright © 2007 John Wiley & Sons, Ltd.

Parsimony or Reductionism?—Against the g-Factor of Nations

HEDE HELFRICH

Department of Psychology, University of Hildesheim, Germany

helfrich@uni-hildesheim.de

Rindermann presents the thesis that international student assessment studies primarily measure the same cognitive ability as intelligence tests, namely Spearman's g-factor. My comment focuses on two objections to his analysis. First, the uniform correlations taking countries as units of analysis may mask different shapes of correlations within each country. Second, the psychometric approach to assess national differences in general intelligence cannot claim universal validity. Copyright © 2007 John Wiley & Sons, Ltd.

Contributions Made and Contributions Yet to be Made by Rindermann's Analysis of National IQs

EARL HUNT

Department of Psychology, The University of Washington, USA

ehunt@u.washington.edu

Rindermann presented criteria for judging international assessment studies of cognitive competence. He showed that broad trends in the present data are sufficiently strong so that the weaknesses of some data sets can be disregarded. We still need to understand how the nature of the testing situation itself restricts the measurement of important aspects of cognition, at the national level. Copyright © 2007 John Wiley & Sons, Ltd.

IQ and Inequality in Human Conditions: Are Correlates Dependent on the Level of Analysis?

PAUL IRWING

Manchester Business School, University of Manchester, UK

paul.irwing@mbs.ac.uk

Rindermann greatly increases the credibility of Lynn and Vanhanen's (2002) findings, by reproducing those using international studies of student attainment. Additional candidates for the prediction of economic outcomes include: personality, knowledge, motivation, psychopathology, inter-group processes and efficient employment of human resources, the importance of each being dependent on whether analysis is at the individual or cross-national level. Copyright © 2007 John Wiley & Sons, Ltd.

It Does not Help to Ignore It

WENDY JOHNSON1,2

1Department of Psychology, University of Edinburgh, UK

2Department of Psychology, University of Minnesota, Twin Cities, USA

wendy.johnson@ed.ac.uk

Rindermann's (this issue) analysis highlights the large factor of cross-national differences that closely links student achievement and intelligence test results. For some, existence of this factor justifies their dominance by proving their innate superiority. For others, it is an awkward observation to be buried amid hope that it will disappear by itself. Neither is constructive; neither is likely to be accurate. Copyright © 2007 John Wiley & Sons, Ltd.

The Evolutionary Biology of National Differences in Intelligence

RICHARD LYNN

University of Ulster, Northern Ireland

Lynnr540@aol.com

Rindermann's work raises the question of the causes of national differences in intelligence. It is proposed that these are likely adaptations that evolved in the European and East Asian peoples to the cognitive demands of survival during the winter and spring in the temperate and cold climates. Copyright © 2007 John Wiley & Sons, Ltd.

If All Tests Measure the Same Thing, How Can We Evaluate the Quality of Schooling?

GERHARD MEISENBERG

Department of Biochemistry, Ross University, School of Medicine, Dominica

GMeisenberg@rossmed.edu.dm

The near-equivalence of school achievement tests and intelligence tests raises questions about the causal relationships between intelligence and school achievement, and about the evaluation of educational systems on the background of country differences in general intelligence. Causal relationships remain unresolved, but we can conclude that tests of school achievement can only be evaluated conjointly with tests of fluid intelligence. Copyright © 2007 John Wiley & Sons, Ltd.

Does the Globalisation of Assessment Lead to the Globalisation of Education?

FROSSO MOTTI-STEFANIDI

Department of Psychology, University of Athens, Greece

frmotti@psych.uoa.gr

International educational surveys and intelligence tests assess the higher-order cognitive abilities that are important for individuals to adapt to a globalised world, and for countries' economic advance. The globalisation of assessment is believed to have revealed the need to re-examine current educational practices in many parts of the world since they do not face up to the challenges posed by globalisation. Copyright © 2007 John Wiley & Sons, Ltd.

Do Recent Large-Scale Cross-National Student Assessment Studies Neglect General Intelligence g for Political Reasons?

HELMUTH NYBORG

KF Andersen Leadership Academy, Lausanne, Switzerland

helmuthnyborg@msn.com

Rindermann's analysis of international student assessment studies re-confirms previous findings that all intelligence- knowledge- and achievement-scales basically measure psychometric g. Several recent large-scale international assessment studies nevertheless chose literacy over psychometric g as their dependent variable, in order to better promote the notion of considerable student educational malleability. This politically correct choice compromises the studies and educational policy. Copyright © 2007 John Wiley & Sons, Ltd.

The Reciprocal Causation of Intelligence and Culture: A Commentary Based on a Piagetian Perspective

GEORG W. OESTERDIEKHOFF

Institute for Sociology, University of Karlsruhe, Germany

Oesterdiek@aol.com

Social and cultural factors such as special child rearing practices and a lack of formalised education account for the stop of ontogenetic development below the level of formal operations in pre-modern or underprivileged social milieus. Only the socialisation practices in modern societies have been efficient enough to cause the IQ gains (Flynn effect) respective the growth of formal operations. Copyright © 2007 John Wiley & Sons, Ltd.

How Smart are Nations? About Corrections and Correlations of International Data

MANFRED PRENZEL

Leibniz Institute for Science Education, University of Kiel, Germany

prenzel@ipn.uni-kiel.de

The homogeneity of the results of international cognitive assessments which Rindermann claims in his paper will be questioned on two levels: First, Rindermann corrects national data sets by a certain amount if specific test participation rates are available, without testing whether the samples are representative. Second, we will demonstrate with an example that the correlations reported can by no means be considered as evidence for the homogeneity of different cognitive tests and other indicators. Copyright © 2007 John Wiley & Sons, Ltd.

Generalisability, Groups, and Genetics

J. PHILIPPE RUSHTON

Department of Psychology, University of Western Ontario, Canada

Rushton@uwo.ca

Rindermann shows that g is highly generalisable. We can add: (a) predictive validities generalise across cultures; (b) g-loaded items found relatively difficult by the Roma (Gypsies) in Serbia are found relatively difficult by East Asians, Whites, South Asians, Coloreds and Blacks in South Africa and (c) group differences are more pronounced on more heritable items, indicating they are partly genetic. Copyright © 2007 John Wiley & Sons, Ltd.

Profiting From Controversy

MANFRED SCHMITT

Department of Psychology, University of Koblenz-Landau, Germany

schmittm@uni-landau.de

My comment will address the scientific value of Rindermann's contribution, wrong conclusions that might be drawn from it, and his quest for interdisciplinary co-operation. Copyright © 2007 John Wiley & Sons, Ltd.

The Softer the Truth, the Harsher the Fight

BIRGIT SPINATH

Department of Psychology, University of Heidelberg, Germany

birgit.spinath@psychologie.uni-heidelberg.de

Rindermann's target paper is bound to provoke retorts. The present commentary analyses the explosive potential of Rindermann's thesis along five questions: Are international student assessment tests flawed? Should we give up the distinction between intellectual potential and scholastic performance? Is intelligence heritable but school achievement is not? Is school achievement primarily dependent on the educational environment but intelligence is not? Can we advance something that is heritable? Copyright © 2007 John Wiley & Sons, Ltd.

International g: Mixed Feelings and Mixed Messages

FRANK M. SPINATH

Department of Psychology, Saarland University, Germany

f.spinath@mx.uni-saarland.de

This commentary reflects on the misleading amalgamation of behaviour genetics findings on the aetiology of individual differences in intelligence and speculations about the causes of international mean differences in intelligence at the group level. It questions the scientific value of mapping or ranking national mean IQs given that the vast majority of the variation lies on the individual level. Copyright © 2007 John Wiley & Sons, Ltd.

Abilities as Achievements, or is it Achievements as Abilities, or is it Both, or Neither?

ROBERT J. STERNBERG

Tufts University, USA

robert.sternberg@tufts.edu

Rindermann's paper shows that well-known ability and achievement tests all roughly measure the same thing, general ability. Three potential implications are that the distinction between ability and achievement is not clear, that we should use broader psychological theories on which to build tests, and that we should consider teaching for leadership rather than merely for academic facts and skills. Copyright © 2007 John Wiley & Sons, Ltd.

Methodological and Conceptual Notes on National Cognitive Ability

THOMAS VOLKEN

Sociological Institute, University of Zürich, Switzerland

volken@soziologie.unizh.ch

While at first sight, the concept of national cognitive ability is appealing, a closer inspection reveals major conceptual and methodological problems which need to be addressed. In particular, the meaning as well as the scope and reach of the concept remain vague, the aggregation process does not consider estimates of different precision, and (too) many values are estimated. Copyright © 2007 John Wiley & Sons, Ltd.

Something Old, Something New, Something Borrowed and Something Blue: National IQ and the Integration of Cognitive-Ability Research

MARTIN VORACEK

Department of Basic Psychological Research, University of Vienna, Austria

martin.voracek@univie.ac.at

This commentary focuses on methodological issues encountered in estimating national IQ; specifically, on criticism regarding one source of national IQ figures (Buj, 1981), which criticism appears unjustified; and affirms Rindermann's call for a broad integration of different paradigms and lines of cognitive-ability research, illustrated with an example from suicide research. Copyright © 2007 John Wiley & Sons, Ltd.

Methodological Aspects Concerning Rindermann's g-factor of International Cognitive-Ability Comparisons

OLIVER WALTER

Leibniz Institute for Science Education, University of Kiel, Germany

walter@ipn.uni-kiel.de

It is argued that several methodological aspects concerning the broad definition of literacy and intelligence, the heterogeneous samples, the scaling methodology of international student assessments, the highly aggregated data and the requirements of higher-order factor analysis provide sound alternative explanations to Heiner Rindermann's hypothesis that literacy could be subsumed under the intelligence construct. Copyright © 2007 John Wiley & Sons, Ltd.

Percentages of Children Living in Poverty Determine IQ Averages of Nations

VOLKMAR WEISS

Leipzig, Germany

volkmar-weiss@t-online.de

By comparing three bodies of independently collected data sets, Lynn–Vanhanen-IQ, PISA-IQ and children poverty percentages, we have evidence of a downward and hence dysgenic trend in a number of nations, reaching up to 6 points within one generation and even higher losses in Latin American countries. Copyright © 2007 John Wiley & Sons, Ltd.

What is the National g-Factor?

JELTE M. WICHERTS1 and OLIVER WILHELM2

1Department of Psychology, University of Amsterdam, The Netherlands

2Institute for Educational Progress, Humboldt University Berlin, Germany

J.M.Wicherts@uva.nl, oliver.wilhelm@rz.hu-berlin.de

Rindermann correlated the national averages of several student assessment studies and ‘national IQ’ estimates and proposes that these variables are all indicators of a common cognitive ability at the macro-social level, which he denotes the national g-factor. We argue that Rindermann oversimplifies issues of individual differences and applies inappropriate statistical analyses. Therefore, we refute his conclusions. Copyright © 2007 John Wiley & Sons, Ltd.

Ancillary