Tde Issue 1-2009 00 Complete

July 2009
Talent Development
&
Excellence
Official Journal of the
Editors-in-Chief: Albert Ziegler

Jiannong Shi
This journal
Talent Development and Excellence is the official scholarly peer reviewed journal of
the International Research Association for Talent Development and Excellence (IRATDE).
The articles contain original research or theory on talent development, expertise,
innovation, or excellence. The Journal is currently published twice annually. All published
articles are assessed by a blind refereeing process and reviewed by at least two
independent referees. Editors-in-Chief are Prof. Albert Ziegler, Ulm University, Germany,
and Prof. Jiannong Shi of the Chinese Academy of Sciences, Bejing. Manuscripts can be
submitted electronically to either of them or to Editor@iratde.org.
Articles will be submitted for abstracting and indexing in Academic Search; Australian
Education Index (AEI); British Education Index; Contents Pages in Education; EBSCO
Online; EBSCO CD Rom Database; Education Journal; Educational Research Abstracts
online (ERA); ERIC; e-psyche; ERIH (European Reference Index for the Humanities,
Pedagogical and Educational Research); Gifted and Talented Abstracts; IBR (International
Bibliography of Book Reviews of Scholarly Literature on the Humanities and Social
Sciences); IBZ (International Bibliography of Periodical Literature on the Humanities and
Social Sciences); ISI Social and Behavioural Sciences; National Database for Research into
International Education (NDRI); psycINFO; PsychLit; Psychological Abstracts; Research
into Higher Education Abstracts and Social Science Citation Index.
Editors-in-Chief:
Albert Ziegler, University of Ulm, Germany
Jiannong Shi, Academy of Sciences, Beijng, China
Editorial Assistant:
Bettina Harder, University of Ulm, Germany
International Advisory Board:

Ai-Girl Tan, Nanyang Technological University, Singapore
Barbara Schober, University of Vienna, Austria
Carmen M. Cretu, University of IASI, Romania
Elena Grigorenko, Yale University, United States
Hans Gruber, University of Regensburg, Germany
Mantak Yuen, The University of Hong Kong, P.R. China
Marion Porath, University of British Columbia, Canada
Osamah Ma'ajeeni, King Abdul Aziz University, Saudi-Arabia
Petri Nokelainen, University of Tampere, Finland
Robert Sternberg, Tufts University, United States
Wilma Vialle, University of Wollongong, Australia
Wolfgang Schneider, University of Wuerzburg, Germany
Impressum:
V.i.S.d.P.: Albert Ziegler, St.Veit-Str. 25, 81673 München, Germany
Welcome note
This issue represents the official inauguration of our scientific journal

published under the umbrella of the International Research
Association for Talent Development & Excellence (IRATDE). This issue
also symbolizes launching of an academic venue with the highest
possible intellectual quality where scientists, educators, and other
professionals interested in the fields of gifted education, talent
development, creativity, innovation and excellence can present their
scientific research. Through IRATDE, we seek to encourage and
support joint research initiatives among scholars of different
nationalities. Our goal is to build bridges for inquisitive
intercommunication that can deepen our understanding and enrich
our research. Through our systematic partnership we aim at fortifying
civilized dialogue among human societies concerning how to
promote the global development of creative minds that care about
constructing an affectionate, merciful, and wise humane society.
My colleagues and I look forward to opportunities to work together
with researchers and practitioners in the field of talent development
and excellence to develop our association to serve our field better
and to promote interdisciplinary and intercultural projects around the
world.
Abdullah M. Aljughaiman
President
The International Research Association for Talent Development and
Excellence
Talent Development & Excellence
Volume 1 Number 1 2009
Contents
Editorial 1
A. Ziegler and J. Shi
The impact of smart fractions, cognitive ability of politicians and average 3

competence of peoples on social development
H. Rindermann, M. Sailer and J. Thompson
Modeling mathematical Actiotopes: The potential role of CLARION 27

S. N. Phillipson and R. Sun
Relative age, talent identification and youth skill development: Do relatively 45

younger athletes have superior technical skills?
J. Schorer, J. Baker, D. Büsch, A. Wilhelm and J. Pabst
Gifted Learners’ Epistemological Beliefs 57

M. Porath, J. Lupart, J. Katz, C. Ngara and P. Richardson
The Leonardo Laboratory: Developing Targeted Programs for Academic 67

Underachievers with Visual-Spatial Gifts
T. M. Newman, W. Brown, L. Hart, D. Macomber, N. Doyle, S. A. Kornilov, L. Jarvin, R. J.
Sternberg and E. L. Grigorenko
Talent Development & Excellence Editorial 1
Vol. 1, No. 1, 2009, 1-2
Editorial
Launching a new scientific journal provokes the question of its raison d’être because of the
now insurmountable flood of publications. As we know, less than 20% of researchers
interested in a given field read more than even 1% of the newly appearing literature.
Nevertheless we believe for several reasons that ‘Talent Development and Excellence’ (TDE)
will play an important role, and in four ways.
As its title says, TDE shall first be a place of engagement for two important research currents
that have been up to now largely separated: Giftedness research and Expertise research. The
former researchers traditionally focus more on talented children, following the development
of their achievement prospectively, while the latter traditionally start above all with persons
who have already excelled, but of course with an interest in how their achievements
developed. We think it is important and desirable for future research to unite these two
branches.
Our second wish in connection with TDE is for interdisciplinary access and a simultaneous
broadening of traditional problems. Typically, talent development and excellence are
investigated by psychologists and pedagogues. By contrast we would like to invite
researchers in further disciplines – (educational) economists, jurists, sociologists, political
scientists, AI researchers, physiologists, neurologists, etc. – to shed light from each of their
specialist perspectives on the process of creativity and achievement development and their
antecedents up to the level of expertise and innovation together. Moreover we explicitly
exclude any limitation to individuals: Excellence can also be an attribute of a group of
persons or a social community, and, in the age of Artificial Intelligence, even the attribute of a
machine or machine network.
While on the one hand publications in modern knowledge societies are more and more
accessible, on the other the necessary expenses of libraries and universities have risen
hugely in the past years. Our third wish is therefore to make TDE freely available, and so the
journal can be read online at www.iratde.org. The print version is generously sponsored by
the King Faisal University in Al-Hassa.
Finally and fourthly, TDE is the official scholarly journal of the International Research
Association for Talent Development and Excellence (IRATDE). We hope that TDE can provide
IRATDE-associated researchers with a bit of intellectual community of their own.
Heiner Rindermann, Michael Sailer and James Thompson present a pioneering study on the
‘smart fraction’ theory. This says that the small share of high-achievers make an
overproportional contribution to societal growth and prosperity. Such a finding would of
course be the best legitimation for society’s support of talent and excellence. The authors
investigated whether there existed relations between the scores of 90 countries in
international educational studies (TIMSS 1995-2007, PISA 2000-2006 and PIRLS 2001-2006)
and a number of societal desiderata such as wealth (GDP), numbers of patents, scientists and
Nobel Prizes, government effectiveness, democracy, but also AIDS and homicides. In
particular they wanted to find out whether the scores of weaker achievers, average achievers,
or of the ‘smart fraction’ could better predict the factors mentioned. And in fact the authors
find clear indications of the prognostic strength of the smart fraction theory: “Our results
emphasize the importance of nurturing the highly gifted.” Could a new scientific journal start
off with any more convincing proof than this of the societal significance of its field of research?
Research areas are only as viable as their methods various. Shane N. Phillipson and Ron Sun
propose a new method for giftedness research, which also expertise research has until now
applied only sporadically: modeling. Their interest is particularly to examine potential
benefits of and problems in modeling the Actiotope Model of giftedness using the cognitive
architecture CLARION. We hope this innovative paper will help make giftedness researchers
ISSN 1869-0459 (print)/ ISSN 1869-2885 (online)
 2009 International Research Association for Talent Development and Excellence
http://www.iratde.org
2 A. Ziegler & J. Shi
aware of the advantages of modeling, as well as establish further bridges to expertise

research. Especially promising is that modeling with CLARION as proposed by the authors
offers a meta-theoretical unifying basis for vigorous scientific exchange.
As circa two decades ago chance became accepted as an independent variable in various
models of talent development, it met with the intense criticism of many traditionally oriented
scholars. But quite soon several factors were found that, while randomly distributed in the
population, do produce a systematic bias. The best studied of these is the birth date:
Relatively older students within an age interval appear to be advantaged. The most significant
explanations for this are held to be, first, maturational differences, and secondly, selection
processes. J. Schorer, J. Baker, D. Büsch, A. Wilhelm, and J. Pabst broaden the state of our
knowledge by pursuing three questions in two very interesting studies: a) Can the ‘relative
age effect’ be demonstrated in two samples of handball players being considered for national
team selection? b) What role do variables play related to body-size and technical skill
development? c) Can these findings be demonstrated also for the differences between those
ultimately selected for the national team and those not selected?
Marion Porath, Judy Lupart, Jennifer Katz, Constantine Ngara, and Pamela Richardson take up
in their contribution what can well be considered a current hot topic in gifted education
research: epistemic beliefs. These are concerned with making meaning of school,
understanding one’s own intentions with regard to learning, and interpreting one’s learning in
the context of knowledge traditions. Their significance is so great because most pedagogic
measures do not act directly, but first must be filtered through the epistemic beliefs of their
addressees. The authors are able to show in this exciting empirical study with pupils in
grades 1-12 that their understanding of learning develops from a focus on the activities of
school (e.g. reading, math), to interpretation of the meaningfulness of different learning
activities, and finally to the meaning of knowledge itself.
The study by Tina M. Newman, William Brown, Lesley Hart, Donna Macomber, Niamh Doyle,
Sergey A. Kornilov, Linda Jarvin, Robert J. Sternberg, and Elena L. Grigorenko is concerned
with a seldom investigated sample: ‘twice exceptional’ children. The authors present with the
Leonardo Laboratory a short-term program for children with coexisting learning disabilities
and spatial gifts. In fact the phenomenon of underachievement, i.e. unexpected low
achievement, is today among the greatest challenges to the promotion of talent. Precisely
twice exceptional children are today considered by schools to be under-identified and
under-served, and therefore to deserve particular attention. Although the program did not
improve academic skills, children demonstrated gains in self-efficacy and improvements in
organizational skills. Given the recognized difficulties in helping twice exceptional children,
these results are definitely encouraging.
Albert Ziegler
Jiannong Shi
Talent Development & Excellence Smart fraction 3
Vol. 1, No. 1, 2009, 3-25
The impact of smart fractions, cognitive ability of

politicians and average competence of peoples on
social development
Heiner Rindermann1,∗, Michael Sailer1 and James Thompson2
Abstract: Smart fraction theory supposes that gifted and talented persons are especially
relevant for societal development. Using results for the 95th percentile from TIMSS 1995-
2007, PISA 2000-2006 and PIRLS 2001-2006 we calculated an ability sum value (N=90
countries) for the upper level group (equivalent to a within country IQ-threshold of 125 or
a student assessment score of 667) and compared its influence with the mean ability and
the 5th percentile ability on wealth (GDP), patent rates, Nobel Prizes, numbers of scientists,
political variables (government effectiveness, democracy, rule of law, political liberty), HIV,
AIDS and homicide. Additionally, using information on school and professional education,
we estimated the cognitive competence of political leaders in N=90 countries. Results of
correlations, regression and path analyses generally show a larger impact of the smart
fractions’ ability on positively valued outcomes than of the mean result or the 5th percentile
fraction. The influence of the 5th percentile fraction on HIV, AIDS and homicide, however,
was stronger. The intelligence of politicians was less important, a longitudinal cross-
lagged analysis could show a positive influence on the cognitive development of nations.
Keywords:
high ability fraction, gifted portion, intelligence, patent rate, growth
A large amount of studies published in the last two decades has shown that cognitive ability
levels of societies are relevant for the development of positively valued aspects of peoples
and countries. Following an economic research tradition “human capital” is relevant for
economic growth and wealth (Hanushek & Kimko, 2000; Lynn & Vanhanen, 2002, 2006; Jones &
Schneider, 2006; Weede, 2006; Rindermann, 2008a). In addition, cognitive ability of nations has
a positive impact on political development, in that it helps building up democracy, the rule of
law and political liberty (Simpson, 1997; Rindermann, 2008b). Intelligence, knowledge and
the intelligent use of knowledge also have beneficial effects on health, for instance they act as
a brake on the spread of HIV (Oesterdiekhoff & Rindermann, 2007; Lakhanpal & Ram, 2008;
Rindermann & Meisenberg, 2009). Finally, cognitive competence is relevant for the
development of modernity as a societal and especially as a cultural phenomenon consisting of
education, autonomy, liberty, morality and rationality (Habermas, 1985/1981; Meisenberg,
2004; Oesterdiekhoff, 2008; Lynn, Harvey & Nyborg, 2009). Societies at a higher ability level
develop more complex, more evidence-based, more ethical and more rational world views.
For some scholars like Georg Oesterdiekhoff (2000) or Michael Hart (2007) intelligence is the
driving force of history.
These broad effects at the cross-national data level are backed in different societies by results
at the individual level for job performance and wealth (Bacharach & Baumeister, 1998; Schmidt
& Hunter, 2004; Irwing & Lynn, 2006; Rindermann & Thompson, 2009), for tolerance, civic
political attitudes and participation in elections (Herrnstein & Murray, 1994; Denny & Doyle,
2008; Deary, Batty & Gale, 2008), for health behavior and health (Goldman & Smith, 2002;
Gottfredson, 2004), moral judgment (Piaget, 1997/1932; Kohlberg, 1987) and more rational
world views (Oesterdiekhoff, 2000; Nyborg, 2009).
1
Institute of Psychology, Karl-Franzens-University Graz, Universitaetsplatz 2, A-8010 Graz,
Austria
∗
Corresponding author. Email: heiner.rindermann@uni-graz.at
2
Department of Psychology, University College London, UK

4 H. Rindermann et al.
The mentioned positive and global effects of cognitive competences are also the reason why
international economic or educational organizations like OECD and UNESCO are conducting
studies on education and competence. But the terms used vary, ranging from “human capital”
by economists (also including personality factors relevant for productive behavior like self-
discipline; e.g. Becker, 1993/1964), “literacy” by educationists (OECD, 2003) and
psychometricians’ and cognitive developmentalists’ focus on the concept of “intelligence”
(Cattell, 1987/1971; Piaget, 2001/1947). The measurement of cognitive abilities in these
diverse research traditions is based on mental tasks that can be solved with thinking and
(within the different approaches) with varying amounts of specific knowledge. Students’
scores from all these different measurement approaches highly correlate at the individual
level (Jensen, 1998; Ceci, 1991) and very highly at the national level (Rindermann, 2007a).
Thus, the manifold of positive correlations among ostensibly dissimilar tests (e.g. verbal,
math, science, figural and Piagetian tasks, psychometric and student assessments) is taken as
an evidence that the same underlying latent factor is involved in all complex cognitive
performance (Ceci, 1991). Formal education (years of schooling or degrees) often is used as a
proxy or causal factor for cognitive ability. We use the term cognitive ability and cognitive
competence interchangeably. They stand for intelligence (thinking ability), the disposal of true
and important knowledge and the intelligent use of that knowledge.
In spite of various terms, research methods, disciplines and paradigms different researchers
came to the same result: Cognitive ability of individuals and societies is important for positive
outcomes. This causal assumption is especially backed by the use of longitudinal designs
controlling for the influence of other factors and of backward effects e.g. of wealth or
democracy (Rindermann, 2008a, 2008b). Of course, intelligence and knowledge are not the
only and single determinants of the positive attributes described. There are additional factors
behind (like education, culture and genes) and beneath (like neighbor effects, personality or
chance factors) cognitive ability and between cognitive ability and the positive outcomes
(like the quality of institutions or meritoric principles). There are positively valued attributes
which may not depend (positively) on cognitive competence (like happiness and mating
success). Therefore, high competence is no guarantee for positive outcomes. Finally,
intelligence not guided by ethics and rationality leads to biased, questionable and
destructive results.
A cognitive theory of action can explain the effect mechanism: An increased ability to
understand information, causal relationships, and consequences of somebody’s own and
others’ behavior in everyday life improves one’s own behavior and the behavior of important
others like parents, friends, classmates, teachers, officials, managers, scientists and political
leaders. Additionally, cognitive ability strengthens attitudes oriented towards a more ethical
lifestyle including perspective taking, considerateness and general pensiveness. But cognitive
ability is also connected – meaning leading to, depending on and relying on common
background factors – to better environments (nurturing, training and gate keeper effects) and
a greater appreciation of civic virtues and bourgeois values like appreciation of future life’s
quality, of security, of education, work and individualized love. Thinking ability forms an
integrated part of a global pattern oriented towards a rational, active, self-controlled and
farsighted lifestyle. They all create positive feedback loops: own intelligence and the
intelligence of others enhance environmental (the physical, social and cultural) quality and by
this again cognitive development.
Smart fraction theory

Thus far we have solely mentioned mean cognitive ability effects. But different authors of the
past and the present have stressed the importance of smart fractions. They place less
emphasis on mean intelligence, but champion the importance of the cognitive ability level of
a smart fraction, say a society’s upper 10, 5, 1 or 0.1%, which should be mainly responsible for
the progress of a society in technological, economic, political and cultural development
(Thomson, 1937; La Griffe du Lion, 2004; Gelade, 2008; Murray, 2008; Weiss, 2009). Because of
Smart fraction 5
the assumed relevance of gifted and talented persons the Scottish pioneer of intelligence
research, Godfrey H. Thomson (1881-1955), recommended the furthering of intelligent
children (similarly Gelade, 2008; Murray, 2008).
The smart fraction is regarded as responsible for progress in a utilitarian sense (wealth, health
and power), but also for achievements in a non-utilitarian sense (music, literature, art, religion,
ethics, philosophy and world-views).
For many this kind of thesis has an elitist taste. But it is first of all an empirical question, which
could be solved with a rigorous orientation towards scientific standards, based on
philosophically justified epistemic principles: Is it true, that smart fractions are more relevant
for important indicators of societal development like wealth and democracy or scientific and
technological progress? Up to now there is no research on this subject using empirical data.
Gelade (2008) has solely used an assumed ability level, he and the anonymous scholar La
Griffe du Lion (2004) have estimated smart fractions from the mean value. But the student
assessment studies (SAS) TIMSS, PISA and PIRLS give results for the 95th percentile
(comparable to an IQ of 125 or higher, mean around IQ 130) enabling us to directly test a
smart fraction theory. Conversely, SAS makes it possible to prove a “non-smart” fraction
theory (5th percentile, comparable to an IQ of 75 or lower, mean around IQ 70): If the cognitive
level of the lower groups is more relevant for outcomes standing for lower cognitive ability
(and at least partly explainable by this) like HIV and AIDS (Oesterdiekhoff & Rindermann,
2007; Lakhanpal & Ram, 2008; Rindermann & Meisenberg, 2009) or criminality (Thomson,
1937; Herrnstein & Murray, 1994; Ellis & Walsh, 2003; Cunha, Heckman, Lochner & Masterov,
2006).
A group in society especially important are statesmen (or rarely stateswomen). Simonton
(1984, 1985, 2006) has done several studies on the measurement of political leader’s cognitive
ability and its impact on political success. Apart from politically (and for the media)
interesting information (e.g. comparatively low cognitive ability of Bush junior, Simonton,
2006, or comparatively high cognitive ability of Obama, Murray, Cardoso & Mendes, 2008)
Simonton and others (Deary et al., 2007) were able to show that cognitive traits of political
leaders are important for their success. Political leadership is a cognitively demanding task
(Suedfeld, Guttieri & Tetlock, 2003, p. 255), so the influence of politicians’ intelligence and
knowledge on the development of societies (apart from their private success in being
reelected or continuing in power) would be expected.
The opponent hypothesis to the political leader thesis would be that it is not the attributes of
politicians that are important but the ones of peoples themselves (e.g. Chomsky, 2009).
Maybe there is no either-or: Both, the ability of peoples and their leaders are important.
Empirical data will show us.
Method
To reduce problems of poor data quality and to ensure that countries are represented in the
sample at all levels of cultural, social and economic development, an average score was
formed from different measures and measurement points to create one construct for each
country.
Cognitive abilities
Three cognitive competence measures were used: mean results in TIMS-, PISA and PIRL-
studies, mean at 5th percentile in TIMS-, PISA and PIRL-studies and mean at 95th percentile in
TIMS-, PISA and PIRL-studies. Sources were TIMSS 1995, 4th and 8th grade, math and science,
TIMSS 1999, 8th grade, math and science, TIMSS 2003, 4th and 8th grade, math and science,
TIMSS 2007, 4th and 8th grade, math and science; PISA (always around 15 year old students)
2000, 2003 and 2006, verbal, math and science literacy, 2003 also problem solving, PIRLS
verbal literacy in 4th grade 2001 and 2006. All results were originally presented in student
assessment scales (SAS M=500, SD=100).
Resources had been for TIMSS 1995 (in the order of the grades and scales: Beaton, Mullis,
Martin, Gonzalez, Kelly & Smith, 1996; Beaton, Martin, Mullis, Gonzalez, Smith & Kelly, 1996;
Mullis, Martin, Beaton, Gonzalez, Kelly & Smith, 1997; Martin, Mullis, Beaton, Gonzalez, Smith &
Kelly, 1997), TIMSS 1999 (Mullis, Martin, Gonzalez, Gregory, Garden, O’Connor, Chrostowski &
Smith, 2000; Martin, Mullis, Gonzalez, Gregory, Smith, Chrostowski, Garden & O’Connor, 2000),
TIMSS 2003 (Mullis, Martin, Gonzalez & Chrostowski, 2004; Martin, Mullis, Gonzalez &
Chrostowski, 2004), TIMSS 2007 (Mullis, Martin & Foy, 2008; Martin, Mullis & Foy, 2008), PISA
2000 (OECD, 2003), PISA 2003 (OECD, 2004a, b), PISA 2006 (OECD, 2007a, 2007b), PIRLS 2001
(Mullis, Martin, Gonzales & Kennedy, 2003) and PIRLS 2006 (Mullis, Martin, Kennedy & Foy,
2007).
A sum value of different scales, grades/age groups, studies and study approaches (grade vs.
age level studies; studies trying to measure abilities defined by curriculum like TIMSS vs.
studies trying to measure abilities defined by cognitive demands in modernity like PISA) is
more convincing (say more representative, reliable and valid). High correlations between
scales within and across studies allow to sum up scales to a general sum value (all factor
loadings on an international G-factor were λ >.90; Rindermann, 2007a, 2007b).
To form a common score the results were at first averaged within one grade, year and study
between different scales (e.g. within TIMSS 1995, 4th grade, across math and science),
secondly within one year and study between different grades (e.g. within TIMSS 1995, across
4th and 8th grade), thirdly within one study between different years (e.g. within TIMSS, across
1995, 1999, 2003 and 2007), fourthly within different grade vs. age study approaches across
TIMSS and PIRLS (TIMSS and PIRLS are studies done in grades, PISA is a study done in a single
age group), fifthly and finally between different study approaches (across grade and age
approach studies: TIMSS-PIRLS-mean and PISA-mean). All averaging was done using z-
transformations calculating means and standard deviations in countries which participated in
all samples used for averaging (so z-formula are based on the same countries and over- or
underestimation is avoided). Subsequently the z-results were re-normed using means and
standard deviations obtained by simple arithmetical averaging of all three study results (SAS-
scale with M=500 and SD=100, mean, 5th percentile, 95th percentile) as an orientation. At the
end the values were transformed to the more usual IQ-scale, using Great Britain as reference
country, SAS-SD were simply transformed to an IQ-scale (“Greenwich-IQ”, M=100, SD=15).
Results are provided for N=90 countries. Means in SAS-scale are 453, 304 and 596, in UK-IQ-
scale 90, 68 and 111. The results are not identical with the formally published cognitive ability
values of Rindermann (2007a), because a) psychometric intelligence test results were not
used here (Lynn & Vanhanen, 2006, provided no information on the tails), b) older student
assessment studies like IEA-Reading and IAEP were not used (for IEA-Reading we could not
find information on the tails, IAEP 1991 only published percentages of solved tasks,
additionally samples were seriously biased), c) newer studies were included (PISA 2006,
PIRLS 2006, TIMSS 2007), and d) the results were not corrected for age and grade or sample
quality. Nevertheless the correlations are very high (with former corrected cognitive ability
sum r=.92, with uncorrected r=.95, N=88).
“Normed” values of all variables at international data level are somewhat arbitrary, e.g. the
student assessment scale with M=500 and SD=100. The norms are estimated by the authors of
the student assessment studies in orientation to results in OECD-countries (and sometimes in
accordance with older results). OECD-membership, however, is no scientific criterion. IQ-
norms depend on the secular rise of intelligence and intelligence test results (“Flynn-effect”).
Student assessment results are biased because only youth in school participated, in several
countries pupils had been too old (especially in older studies and in developing countries),
not all regions participated (especially in older studies and in developing countries) etc. (see
Rindermann, 2007a; Wuttke, 2007). We have not applied corrections here because results at
the higher (“smart”, “gifted”, “talented”) or lower (“non-smart”, “imprudent” or “dull”) levels
are differently affected by the distortions, presumably the high level results are nearly not
affected, but the low level results (and less the means). Corrections should be tried in further
Smart fraction 7
studies. But also in the other variables the norms are arbitrary, e.g. in GDP (inflation, Dollar or
Euro), in democracy (which scale?).
The competence levels of smart fractions, average and non-smart fractions are obtained
through student assessment studies. But students do not work, do not win Nobel Prizes nor do
they vote. We assume that the results of students could be generalized to adults, an
assumption that is backed by high correlations with IQ measures (r=.87, N=86, Lynn &
Vanhanen, 2006; often gained in adult samples), with an adult literacy study (r=.68, N=20;
OECD, 2000) and the educational level of societies (r=.67, N=84, r=.75, N=85; measures see
below). OECD is planning an adult literacy study for a larger country sample, by using their
data on tails it would be possible to prove if our results are stable.
Because in high ability societies there is a larger percentage of a cognitive elite at a higher
level (and inverted for the lower level) the three values are correlated (see also Table 2):
mean with high (r=.97), mean with low (r=.97), high with low (r=.90). As a consequence of
these high correlations suppressor effects are expectable in regression analyses.
For repeated measurements with cognitive abilities (Figure 5) old student assessment studies
collected by Lee and Barro (1997) were used (here N=17; for further information see
Rindermann, 2008b).
Cognitive ability of leading politicians

To estimate the cognitive ability of statesman (stateswomen) based on their formal education
we selected the leading politicians of all the countries in the student assessment studies
between 1960 and 2009. Leading politicians include presidents (33%), heads of government
(64%) and the rest kings, emirs and sultans (3%). It was difficult for many countries to assess
who has the real decisive power or who used or uses it (e.g. for Czech Republic, Iran, Poland,
Russia and especially Switzerland). For these countries the two leading positions were used. A
second problem was the modifications of countries (e.g. Yugoslavia, Soviet Union, Germany).
In these cases the largest successor country represents the older one and vice versa, smaller
successor countries start existing with their formal legalization. We also include (like SAS)
some smaller territories which are not formal states (like Palestine or Taiwan). We always use
the most usual names. A third problem often lies in not exactly knowing the leader’s level of
education. A fourth problem is the assessment of several educational vitae like those of
clerical leaders in Iran (is it a university degree? – we estimated it as a university degree in
this case). We have not assessed the content or the quality of a university degree (e.g. in
Science, Technology, Engineering and Mathematics/STEM or law, highly or lowly ranked
institutions). The most serious problem lies in the low comparability of education and
educational degrees across countries. There is considerable heterogeneity in educational
standards across the world. According to the student assessment results, a secondary degree
in OECD-countries is hardly comparable to a secondary degree in developing countries,
because the former have much higher scores. Furthermore, corruption and forging at
universities and especially in conferring and claiming to have degrees is not impossible. We
tried to exclude all honorary doctorates.
By using databanks (Munzinger-Biography: www.munzinger.de, MSN-Carta, Who is Who and
only rarely Wikipedia) we could find information for N=896 leading politicians. Homepages
of politicians were not used; from experiences in a study on Austrian and German politicians
we found that they try to overestimate their educational record (e.g. university dropout given
as a degree).
We assessed education for school education (as highest level: no school; primary school;
secondary school; high-school diploma; university degree; doctorate; doctorate plus further
scientific degrees, “Habilitation, Venia Legendi”, or scientific achievement like publications),
and education for professional training (as highest level: no vocational or professional training;
vocational training; qualified training like technician; university degree; doctorate plus further
scientific degrees, “Habilitation, Venia Legendi”, or scientific achievement like publications).
For the last category (“doctorate plus further scientific degrees or achievement”) Fernando
Henrique Cardoso from Brazil is an example (published several scientific books, professor of
political science, member of or taught at Institute of Advanced Study, Princeton, Collège de
France, Stanford, Brown, etc.). Both overlapping indicators correlate with r=.90 (Cronbach-
α=.95).
Finally we tried following Gottfredson (2004, 2005) to estimate IQ-levels: For no school IQ 80,
for primary school IQ 90, for secondary school IQ 100, for high-school diploma IQ 113, for
university degree 119, for doctorate 129 and for doctorate plus further scientific degree or
achievement 138. For levels in between and different vocational or qualified trainings we gave
values between 80 and 119. The mean of the two educational indicators and our IQ-estimate
correlate with r=.96, IQ with school education r=.99, with professional training r=.88 (N=896).
The mean IQ of politicians across countries is 118 (SD=7, N=90). The mean IQ is increasing
from 1960 to 2000: in the 60s IQ 114, in the 90s IQ 118, in 00s IQ 119 and the variances are
decreasing (from 12 to 7 and 7). Because we mainly include only countries participating also
in student assessment studies the worldwide average could be lower. The sample for analyses
was N=90 countries (for the map, Figure 2, N=95).
Wealth and development indicators of societies

Wealth: Gross domestic product 2003 (GDP per capita, purchasing power parity/ppp; Human
Development Report/HDR, 2005, here for N=84 countries). GDP (ppp) 1998 per capita from
Lynn and Vanhanen (2002), here for N=85 countries. GDP considers only goods and services
produced within a country, not income received from abroad. GDP is an indicator for
produced wealth. Their sources are UN data sets.
Developmental level of society: Human development index (HDI) from HDR (2005). The HDI is
a composite indicating three basic dimensions of human development: a long and healthy life
(life expectancy at birth), knowledge (adult literacy rate and the combined gross enrolment
ratio for primary, secondary and tertiary schools) and a decent standard of living, as
measured by GDP per capita in ppp US dollars (here for N=85 countries).
Educational level of societies and attributes of educational systems

The general educational level of society was estimated in two ways: 1. Standardized sum of
literate adults’ rate 1991, rate of persons between 12 and 19 years old from 1960 to 1985
having graduated from secondary school, and years of school attendance of persons 25 years
or older 1990-2000 (Rindermann, 2007a). This indicator is used for analyses with data from a
longer time interval (here for N=84 countries). 2. Education index 2003 (sum of adult literacy
rate and combined gross enrollment ratio for primary, secondary and tertiary schools; HDR,
2005). This indicator is used for analyses with data around the turn of the century (here for
N=82 countries).
Young tracking age: Age of initial tracking between schools (reversed: young age), for some
countries (e.g. Hong Kong and Switzerland) the information given by OECD/PISA was
corrected. Class size: Large class sizes, many pupils per class. For the two educational policy
variables see Rindermann and Ceci (2009).
Attributes standing for high cognitive achievement of a society (excellence)

Patent rate: Number of patents of a nation (sum of residents and nonresidents) related to
population size, average annual patents per 1 million people 1960-2007 and 1991-2007 (the
two decades, in which the majority of student assessment studies were carried out, here used
for N=81 and 76 countries). Source is the World Intellectual Property Organization (WIPO,
2009), an agency of the United Nations.
Nobel Prizes: Nobel Prizes in peace, literature and science 1901-2004 related to population
size (sources: http://nobelprize.org/nobel_prizes/lists/all, http://en.wikipedia.org/wiki/
List_of_Nobel_laureates_by_country). Science sums up Nobel Prizes in physics, chemistry,
Smart fraction 9
medicine and economics. Mean correlations between those are around r=.90 (α=.97, here for
N=87 countries).
Scientist rate: Scientists and engineers in research and development per million people, 1985-
1995 (source: Kurian, 2001, p. 388, here for N=51 countries).
High-technology exports: High-technology exports as percentage of manufacturing exports,
1997 (source: Kurian, 2001, p. 389-390, here for N=61 countries).
Political indicators
Government effectiveness: Government effectiveness 1996-2005 as “the quality of public
services, the quality of the civil service and the degree of its independence from political
pressures, the quality of policy formulation and implementation, and the credibility of the
government’s commitment to such policies” (Kaufmann, Kraay & Mastruzzi, 2006, p. 2). The
estimates are given here for N=88 countries.
Democracy: Democracy 1950-2004 and 1996-2000 from Vanhanen (2005) and Marshall and
Jaggers (2000), α=.95. The estimates are used for N=84 countries.
Rule of law: Rule of law 1970-2000 and 2000 with emphasis on ownership law from Gwartney
and Lawson (2003) and Knack and Keefer (1995), α=.90, here N=76 and 73.
Political freedom: Political liberty or freedom was taken from Freedom House (2004) for the
year 1999 (here N=86).
For further description of political indicators see Rindermann (2008b).
HIV, AIDS and criminality indicators

HIV: HIV-infection rate 2001 and 2003 (mean r=.95, α=.99) for adults and adults 15-49 years
old are from UNAIDS/WHO (2003), here N=82.
AIDS: Percentage (per 100.000) of people suffering on AIDS 1995-2000, source Statistisches
Bundesamt (2002, p. 290ff.), here N=83.
Homicide: Homicide rate (per 100.000 inhabitants) 1995 and 2002 (α=.41, here N=80) from
Interpol (2004).
Problems in international data sets

In all international data, problems and flaws are detectable. E.g. students are older than
expected. How we define grade level if in some countries children start attending “school” by
the age of four, in others by the age of seven? How can we compare wealth between First and
Third World countries where much of the economy is based on subsistence farming and
informal trading? Luxembourg, Ireland and the States seem to be too rich or at least richer
than expected, Saudi Arabia is astonishingly constitutional (rule of law). How could literacy of
students in Yemen be measured, where several regions are more or less out of state control?
Kazakhstan seems to be too good in cognitive competence levels.
There are some anomalies in data like decreasing rates of patents in UK and Switzerland or
years without patents in Norway (changed into missing values). Nobel Prizes also seem to be
given with political considerations, e.g. there is a serious decline of Nobel Prizes for Germany
starting with 1933 (Nobel Prizes are given for life’s work achievement and not for one single
achievement in a current year); related to the population size persons from Sweden seem to
receive a Nobel Prize more easily, US-Americans less easily. And not all countries (e.g. Israel)
existed during the whole 20th century. Finally, there is no Nobel Prize in social sciences and
humanities (e.g. in psychology, sociology and philosophy) and mathematics (though there is
the Fields Medal). For future analyses additional awards should be considered.
The scientist rate is slightly misleading because many of the scientists and engineers in
Western countries are of foreign origin (brain gain like in USA and Singapore). The quality of
HIV data is sometimes doubted, especially for countries from Eastern Europe, for developing
countries, Muslim countries, and countries from sub-Saharan Africa.
We tried to solve or at least reduce these problems by aggregating data from different years
and by using different indicators for the very construct, e.g. patent rates, Nobel Prizes,
scientist rates and high-technology exports for cognitive excellence of a society (Figure 3).
Statistical methods
At first we calculated correlations between mean, non-smart and smart fraction competence
measures on the one hand and different societal attributes on the other. Higher correlations
for the non-smart or smart fraction indicate their larger relevance for these attributes.
Secondly we present regression analyses showing relative impact of smart fraction ability or
non-smart fraction ability compared to mean values. Finally we present results of cross-
sectional and longitudinal path-analyses. Because of suppressor-effects the smart and non-
smart fraction ability levels are not put together with mean ability in one regression analysis
(only either smart or non-smart and average).
Regression and path analyses are used to calculate direct, indirect and net effects of
variables. In these analyses the standardized path coefficients (β) between different variables
are to be interpreted. Correlations are always added in parentheses. Correlations help to
quickly estimate the influence of other variables in the model (difference between correlation
and path coefficient), they allow the model to be checked (Σrβ=R²=1-error) and the
proportion of the explained variance in each factor to be calculated (R²=Σrβ). “Good” values
for fit indices are SRMR≤.08 (Hu & Bentler, 1998, 1999) or SRMR≤.05 (Schermelleh-Engel,
Moosbrugger & Müller, 2003) and CFI≥.95 (Hu & Bentler, 1998, 1999) or CFI≥.97
(Schermelleh-Engel et al., 2003), and “acceptable” fit is reached with SRMR≤.10 and CFI≥.95
(Schermelleh-Engel et al., 2003). For the analyses, SPSS 16.0 and LISREL 8.80 were used, while
SAS 9.1 was used to produce the maps.
Significance tests were not used for interpretation (see Rindermann, 2008a; for an in-depth
justification e.g. Cohen, 1994; Falk & Greenbaum, 1995; Gigerenzer, 2004; Hunter, 1997).
Especially in comparisons between countries, they are not appropriate for scientific
reasoning. The results of significance tests depend on the number of observations. The
observations here are for a limited number of countries (around 90), but each country-level
observation is based on thousands of individual observations within each country. Possible
causal relationships are not more or less true if they are significant or not (see Rindermann &
Meisenberg, 2009; Rindermann & Ceci, 2009).
Results
Distributions
The results for the most important variables are shown in Table 1. Figures (Maps) 1 and 2
show the distributions of the smart fractions’ and political leaders’ ability levels across the
world. Because average, upper and lower levels are correlated there are at first sight no large
differences: The highest values for the smart fractions are found in East Asia (1. Singapore IQ
127, 2. South Korea IQ 125, 3. Japan IQ 124, 5. Taiwan IQ 123, 9. Hong Kong IQ 122). A similar
result was found in psychometric (average) intelligence or in student assessment studies (see
Rindermann, 2007a). Different from the SAS, Scandinavia reaches in the cognitive elite not
such a good rank (11. Finland IQ 121, 12. Estonia IQ 121 [the Baltics are added here], 16.
Sweden IQ 120, 25. Denmark IQ 118, 34. Latvia IQ 117, 38. Lithuania IQ 116, 39. Iceland IQ 116,
41. Norway IQ 116). Maybe a homogenizing educational policy furthering weaker but
disadvantaging high ability pupils leads to a smaller standard deviation and lower values for
a gifted subgroup. Better are the traditional Commonwealth countries (5. New Zealand IQ 123,
7. Australia IQ 122 and 8. United Kingdom with IQ 122). They are followed by Western and
Eastern European and North American countries, by South European countries, Arab or
Smart fraction 11
Muslim and Latin American countries and finally by sub-Saharan countries. The countries
with the lowest results are 84. Botswana (IQ 96), 85. Saudi-Arabia (IQ 95), 86. Morocco (IQ 95),
87. Kyrgyzstan (IQ 94), 88. Belize (IQ 90), 89. Ghana (IQ 89) and 90. Yemen (IQ 84). Presumably
many not participating countries would have lower values.
Figure 1. World map of cognitive ability level at 95th percentile (mean of TIMSS 1995-07, IGLU 2001-06,
PISA 2000-06, N=90 nations, darker means higher competence, hachured: no data)
Figure 2. World map of leading politicians‘ cognitive ability level 1960-2009, N=95 nations, darker
means higher competence)
12
Table 1. Means of used variables (most important ones)
Country CA- CA- CA- CA- CA- GDP HDI Patent Patent Nobel Nobel Nobel Scien- High Gov. Demo- Rule Politi- HIV
H. Rindermann et al.
mean 5% 95% politi- politi- 2003 2003 rate rate Peace Literat. Science tists techno- eff. cracy of law cal rate
cians cians 1960- 1991- 1901- 1901- 1901- 1985- logy 1996- 1950- 1970- liberty 2001-
60-09 90-09 2007 2007 2004 2004 2004 95 1997 2005 2004 2000 1999 2003
IQ-scale (M=100) ppp $ 0-1 per mill per 10 mill per mill “IQ”-scale (M=100) %
Albania 81.10 55.84 103.56 120.20 122.00 4584 .780 – – 0 0 0 – 1 92.6 90.0 96.3 92.8 0.05
Algeria 80.56 63.23 97.94 106.63 107.75 6107 .722 5.52 5.80 0 0 0 – 22 90.7 87.2 82.7 85.2 0.06
Argentina 81.50 54.72 105.79 117.83 120.67 12106 .863 51.35 21.25 0.79 0 0.30 671 15 100.4 103.8 95.2 108.0 0.61
Armenia 93.06 69.34 116.94 125.33 125.33 3671 .759 20.12 20.76 0 0 0 – – 93.9 107.3 – 96.6 0.12
Australia 101.12 79.06 121.94 115.55 116.75 29632 .955 252.81 320.48 0 0.77 0.77 3166 39 129.7 128.3 125.8 119.4 0.09
Austria 99.65 78.16 119.34 128.20 123.80 30094 .936 305.42 84.45 2.70 1.35 2.70 1631 24 127.4 129.9 127.7 119.4 0.21
Azerbaijan 84.62 73.01 98.84 119.00 119.00 3617 .729 2.81 2.81 0 0 0 – – 86.4 89.5 – 89.0 0.04
Bahrain 84.24 61.99 105.80 107.00 107.00 17479 .846 42.61 43.73 0 0 0 – – 108.7 83.8 103.6 77.6 0.17
Belgium 99.13 75.02 116.53 121.31 121.50 28335 .945 432.19 40.55 3.12 1.04 1.04 1814 23 125.9 132.9 128.5 115.6 0.18
Belize 63.55 40.93 89.95 119.00 119.00 6950 .753 16.00 28.00 0 0 0 – – 97.9 105.5 103.2 119.4 1.93
Bosnia 90.60 69.88 110.07 121.14 121.14 5967 .786 9.46 10.48 – – – – – 89.8 – – – –
Botswana 73.93 50.79 96.15 116.00 115.00 8714 .565 15.23 – 0 0 0 – – 111.3 104.6 110.1 111.8 33.99
Brazil 81.59 58.43 104.65 116.50 116.50 7790 .792 9.38 8.13 0 0 0 168 18 99.0 101.1 103.5 100.4 0.59
Bulgaria 93.46 67.92 117.22 115.92 124.00 7731 .808 45.19 22.78 0 0 0 – – 96.7 93.6 107.8 108.0 0.04
Canada 101.75 79.59 120.32 120.11 119.00 30677 .949 306.43 211.49 0.46 0 0.92 2656 25 130.9 123.1 127.0 119.4 0.27
Chile 83.62 60.95 105.97 122.75 124.00 10274 .854 18.70 12.33 0 2.02 0 – 19 119.6 101.4 93.6 111.8 0.27
Colombia 80.61 58.15 101.38 119.67 119.00 6702 .785 5.88 4.91 0 0.44 0 – 20 98.3 104.3 87.7 96.6 0.51
Croatia 95.96 77.23 115.06 119.67 119.67 11080 .841 20.67 20.82 0 0 0 1978 19 103.1 104.8 107.5 96.6 0.04
Cyprus 91.59 68.65 112.63 122.33 124.00 18776 .891 54.36 59.70 0 0 0 – – 118.0 119.6 100.8 119.4 0.03
Czech Republic 99.96 78.92 119.96 118.25 120.78 16357 .874 67.35 70.75 0 0.97 0.24 1159 13 111.4 102.8 104.7 115.6 0.05
Denmark 98.46 76.86 118.17 115.00 119.00 31465 .941 181.40 67.90 2.03 6.09 3.55 2647 27 132.1 133.0 128.0 119.4 0.17
Egypt 81.14 53.73 107.28 90.00 90.00 3950 .659 3.26 3.23 0.28 0.28 0.07 458 7 97.4 85.6 91.1 85.2 0.04
El Salvador 77.53 59.36 96.19 118.42 120.00 4781 .722 5.45 3.94 0 0 0 19 16 95.6 96.4 84.1 108.0 0.59
Estonia 102.26 84.40 120.75 120.11 120.11 13539 .853 86.90 107.30 0 0 0 2018 24 113.0 116.7 107.4 115.6 0.84
Finland 102.91 84.96 120.92 120.24 119.00 27619 .941 174.86 197.50 0 2.14 0.54 2812 26 131.7 127.3 119.7 119.4 0.08
France 98.17 77.01 117.77 119.67 119.00 27677 .938 219.04 102.56 1.56 2.74 1.17 2584 31 124.7 122.5 114.2 115.6 0.33
Georgia 87.62 66.16 107.50 123.00 123.00 2588 .732 28.04 26.38 0 0 0 – – 90.5 107.3 – 100.4 0.05
Germany 99.08 75.71 119.72 123.25 125.67 27756 .930 118.02 101.35 0.83 1.17 2.50 2843 26 126.4 128.8 131.0 115.6 0.09
Ghana 61.25 32.86 89.38 118.60 122.33 2238 .520 1.23 0.30 0 0 0 – – 98.2 89.7 84.9 104.2 2.73
Greece 94.37 71.45 115.46 123.07 126.50 19954 .912 87.48 54.56 0 2.21 0 774 12 112.4 117.1 102.6 111.8 0.17
Hong Kong 103.66 83.32 121.54 119.00 119.00 27179 .916 155.19 231.93 0 0 0 98 29 123.3 – 117.7 96.6 0.09
Hungary 99.37 78.07 119.77 119.77 125.00 14584 .862 84.62 64.53 0 .99 0.74 1033 39 111.2 95.9 105.9 115.6 0.05
Iceland 96.45 75.34 116.00 120.08 119.00 31243 .956 73.66 115.52 0 47.66 0 – – 131.3 132.2 122.9 119.4 0.19
Indonesia 81.75 62.00 100.93 119.33 117.40 3361 .697 0.85 0.85 0 0 0 – 20 93.9 89.1 91.0 96.6 0.09
Iran 82.83 60.64 104.46 115.38 119.00 6995 .736 4.66 2.41 0.32 0 0 521 – 93.4 86.4 86.7 81.4 0.09
Table 1. (continued)
60-09 90-09 2007 2007 2004 2004 2004 95 1997 2005 2004 2000 1999 2003
Ireland 99.92 78.55 119.95 117.75 117.80 37738 .946 133.03 101.43 3.19 9.58 0.80 1871 62 126.1 124.7 128.4 119.4 0.09
Israel 92.57 64.65 117.52 117.00 118.25 20033 .915 147.58 170.02 6.24 3.12 0.78 – 33 117.1 126.8 117.5 115.6 0.08
Italy 96.57 74.09 117.45 125.22 123.00 27119 .934 201.68 53.02 0.18 1.11 0.32 1325 15 113.1 133.4 106.9 115.6 0.41
Japan 104.55 82.85 124.30 117.25 119.00 27967 .943 284.74 501.64 0.09 0.19 0.19 6309 38 118.0 122.4 127.1 115.6 0.04
Jordan 86.08 61.37 108.47 113.35 113.91 4320 .753 6.08 3.93 4.03 0 0 106 26 104.3 85.9 94.9 96.6 0.04
Kazakhstan 101.93 79.52 122.11 113.00 113.00 6671 .761 43.78 43.95 0 0 0 – – 88.1 89.8 102.2 85.2 0.13
Korea (South) 106.37 86.11 125.25 118.11 119.80 17971 .901 0.27 0.34 0.30 0 0 2636 39 112.4 103.1 104.5 111.8 0.04
Kuwait 75.72 53.10 97.77 105.00 105.00 18047 .844 – – 0 0 0 – 4 103.4 84.0 91.5 92.8 0.10
Kyrgyzstan 69.93 45.78 93.36 119.00 119.00 1751 .702 10.25 9.50 0 0 0 703 24 89.7 91.9 – 89.0 0.06
Latvia 97.47 77.07 116.96 121.73 121.73 10270 .836 66.28 48.82 0 0 0 1189 15 106.7 121.1 104.9 115.6 0.46
Lebanon 83.61 61.50 105.99 119.22 119.00 5074 .759 30.77 – 0 0 0 – – 96.2 100.6 101.2 85.2 0.09
Liechtenstein 100.93 78.79 121.21 123.00 123.00 – – – – 0 0 0 – – 124.3 – – 119.4 –
Lithuania 96.96 76.70 116.41 123.00 123.00 11702 .852 26.44 20.50 0 0 0 – 21 107.8 124.9 104.0 115.6 0.09
Luxembourg 98.31 76.19 118.84 127.00 124.00 62298 .949 1489.94 87.96 0 0 0 – – 131.9 130.2 126.3 119.4 0.18
Macau 101.11 84.43 117.94 119.00 119.00 – – 13.82 19.30 0 0 0 – – 117.5 – – – –
Macedonia 84.58 60.08 107.12 122.33 122.33 6794 .797 21.82 21.19 0 0 0 – – 95.0 110.0 – 104.2 0.04
Malaysia 95.54 74.74 115.92 119.67 122.33 9512 .796 24.94 46.22 0 0 0 87 67 113.2 104.8 104.3 89.0 0.35
Malta 92.41 63.40 116.70 126.14 129.00 17633 .867 70.59 249.93 0 0 0 – – 111.7 127.1 100.0 119.4 0.13
Mexico 85.37 64.97 105.47 121.22 124.00 9168 .814 20.83 32.43 0.17 0.17 0.04 213 33 101.7 92.4 92.5 100.4 0.26
Moldova 92.29 70.06 112.71 120.43 120.43 1510 .671 26.39 27.86 0 0 0 1539 9 88.6 108.2 – 104.2 0.18
Montenegro 84.22 63.95 104.36 119.00 119.00 – – – – – – – – – – – – – –
Morocco 71.02 47.48 95.36 118.65 123.86 4004 .631 6.88 9.10 0 0 0 – 27 99.5 85.1 96.1 92.8 0.08
Netherlands 101.89 82.74 119.96 124.82 122.33 29371 .943 94.30 74.07 0.76 0 2.28 2656 44 133.5 132.7 129.6 119.4 0.18
New Zealand 100.11 75.94 122.65 114.40 116.00 22582 .933 432.51 556.67 0 0 .87 1778 11 131.0 126.9 127.2 119.4 0.09
Norway 95.80 73.73 115.83 112.47 120.67 37670 .963 264.38 260.45 5.12 7.68 2.56 3678 24 130.7 129.6 126.6 119.4 0.09
Oman 80.64 55.43 104.12 114.00 113.00 13584 .781 – – 0 0 0 – – 111.3 83.7 108.4 81.4 0.09
Palestine 79.96 52.23 106.04 122.33 122.33 – .729 – – – – – – – – – – – –
Peru 74.03 49.77 97.00 118.27 121.50 5260 .762 7.17 7.53 0 0 0 625 10 95.3 99.3 80.1 92.8 0.40
Philippines 73.55 46.61 101.02 121.57 118.50 4321 .758 5.98 7.08 0 0 0 157 56 100.2 102.6 95.3 108.0 0.07
Poland 96.95 74.99 117.89 121.31 121.53 11379 .858 45.51 32.25 0.52 1.04 0 1299 12 109.7 93.1 97.8 115.6 0.08
Smart fraction
Portugal 92.12 70.89 112.14 121.53 121.00 18126 .904 62.28 20.47 0 1.07 0.27 1185 11 117.6 105.4 105.1 119.4 0.38
Qatar 72.11 49.37 96.20 97.67 101.50 19844 .849 – – 0 0 0 – – 109.9 83.7 – 81.4 –
Romania 89.00 65.77 110.77 117.00 123.77 7277 .792 39.63 24.90 0 0 0 1382 7 94.1 92.7 95.2 111.8 0.05
Russia 97.27 75.66 118.09 116.18 124.00 9230 .795 137.30 70.60 0.10 0.21 0.13 3520 19 92.1 95.2 87.5 92.8 0.82
Saudi Arabia 74.40 53.11 95.40 90.00 90.00 13226 .772 3.24 4.21 0 0 0 – 29 97.0 83.7 107.5 73.8 0.05
13
14
Table 1. (continued)
60-09 90-09 2007 2007 2004 2004 2004 95 1997 2005 2004 2000 1999 2003
Singapore 104.56 78.86 127.22 119.00 119.00 24481 .907 433.42 910.35 0 0 0 2728 71 135.7 95.4 125.8 89.0 0.18
Slovakia 97.59 75.61 117.83 125.67 125.67 13494 .849 327.11 58.58 0 0 0 1821 15 107.2 124.7 108.7 111.8 0.04
Slovenia 98.57 78.13 118.27 121.00 121.00 19150 .904 87.20 78.02 0 0 0 2544 16 112.9 127.4 113.0 115.6 0.05
South Africa 63.26 35.69 100.06 114.17 110.40 10346 .658 115.57 – 1.24 0.83 0.10 938 – 108.8 103.7 116.7 115.6 18.64
Spain 95.65 75.36 115.19 122.43 119.00 22391 .928 98.04 25.06 0 1.44 0.07 1210 17 124.9 106.4 114.0 115.6 0.53
Sweden 100.14 79.21 119.98 117.73 116.60 26750 .949 327.75 125.43 4.99 7.48 4.99 3714 34 130.1 129.9 121.1 119.4 0.09
Switzerland 99.83 77.25 120.07 122.33 119.00 30552 .947 559.78 72.71 4.93 3.28 6.16 – 28 135.3 130.0 134.9 119.4 0.37
Syria 80.57 64.91 104.84 120.71 121.00 3576 .721 4.56 – 0 0 0 – 1 85.8 86.0 83.7 73.8 0.04
Taiwan. RoCh 102.93 80.92 122.57 121.33 125.67 – – – – 0 0 0 – – 118.9 91.6 116.0 111.8 –
Thailand 90.11 71.12 109.99 117.12 118.11 7595 .778 3.37 5.45 0 0 0 119 43 104.9 93.7 104.5 108.0 1.07
Trinidad/Tobago 84.55 57.61 110.05 121.50 122.33 10766 .801 46.10 26.89 0 10.03 0 – – 107.3 113.9 102.4 115.6 2.17
Tunisia 80.81 60.33 100.63 109.50 100.00 7161 .753 10.57 7.85 0 0 0 388 11 109.9 84.7 97.8 85.2 0.04
Turkey 87.06 65.69 110.17 120.30 124.00 6772 .750 5.25 8.27 0 0 0 261 9 100.1 112.1 96.1 92.8 0.05
Ukraine 92.99 70.91 113.33 124.00 124.00 5491 .766 51.74 63.53 0 0 0 3173 – 89.3 118.6 89.9 100.4 0.68
United A. Emirates 91.91 67.76 115.05 92.50 91.67 22420 .849 – – 0 0 0 – – 110.3 84.3 91.5 85.2 –
United Kingdom 100.00 76.14 121.92 115.10 116.00 27147 .939 239.84 72.43 1.63 1.45 3.08 2417 41 131.6 126.7 126.6 115.6 0.13
United States 98.41 74.90 120.30 119.36 121.50 37562 .944 168.81 272.74 0.90 0.48 2.03 3732 44 127.3 124.4 134.1 119.4 0.40
Uruguay 87.99 61.08 112.19 118.87 124.00 8280 .840 27.96 13.48 0 0 0 688 8 108.7 115.1 102.4 115.6 0.20
Yemen 63.52 43.27 84.10 100.00 100.00 889 .489 0.28 – 0 0 0 – 0 88.4 85.4 – 85.2 0.09
Yugoslavia/Serbia 90.20 67.81 111.03 117.58 122.27 – – 30.04 10.74 0 0.48 0 – – 91.1 88.9 – 89.0 0.11
89.83 67.53 111.21
M 117.72 118.51 15331 .824 113.11 83.24 0.53 1.37 0.45 1667 24.08 109.4 106.9 107.0 105.0 0.91
(453) (304) (596)
10.95 12.00 9.54
SD 7.14 7.50 11399 .104 197.59 141.55 1.31 5.40 1.10 1301 15.46 14.4 16.6 14.5 13.8 4.24
(73) (80) (64)
Note: As country names the normally used names; all standardized and unstandardized scales are somewhat arbitrary (e.g. GDP and inflation, IQ and reference
groups); CA-mean: cognitive ability mean normed according to UK (“Greenwich-norm”) M=100, as SD used students assessment results (SD=100) transformed into
IQ-scale (SD=15), means calculated across countries, for student assessment results are also reported original mean in SAS-scale (M=500, SD=100); CA-5%:
cognitive ability mean at 5th percentile; CA-95%: cognitive ability mean at 95th percentile; CA-politicians 60-09: estimated cognitive ability of leading politicians
(1960-2009) of a country according to their formal education; CA-politicians 90-09: estimated cognitive ability of leading politicians (1990-2009); GDP 2003: GDP per
capita in purchasing power parity (ppp) US Dollars; HDI 2003: Human Development Index 2003, scale 0 (low) to 1 (high); Patent rate 1960-2007: average annual
patents per 1 million people 1960-2007; Patent rate 1991-2007: average annual patents per 1 million people 1991-2007; Nobel Peace, Literature and Science 1901-
2004: sum of Nobel Prizes in Peace, Literature and Science per 10 million people 1901-2004; Scientists 1985-95: scientists and engineers in research and
development per million people, 1985-1995; High technology 1997: high-technology exports as percentage of manufacturing exports, 1997; Gov. eff. 1996-2005,
Democracy 1950-2004, Rule of law 1970-2000 and Political liberty 1999: government effectiveness 1996-2005, democracy value across 1950-2004, rule of law value
across 1970-2000 and political liberty in 1999 in IQ-scale with M=100 and SD=15 (in world samples, here positively selected by having data in student assessment
results); HIV rate 2001-2003: HIV rate 2001-2003 of adults in percent.
Smart fraction 15
Some astonishing results are observable like the high level of Kazakhstan (6., IQ 122) and the
comparatively low for Israel (31., IQ 118, mean 93). For Kazakhstan we have only results from
TIMSS 2007 (4th grade); Mullis et al. (2008, p. 34) describe sample anomalies, a correction
would be necessary. Israel has participated in several studies, compared to older studies and
Jews in the Western World the results are deteriorating (e.g. Lynn & Longley, 2006). Most
probably multiple reasons are responsible and not only the 20% fraction of Arabs (a thorough
analysis would be necessary).
There are also characteristic differences between mean, upper and lower levels. For instance
between Canada and USA there is no difference in the upper level (IQ 120 and 120), but in the
lower level (IQ 80 and 75). The past history of slavery and a different immigration policy (or
different success of migration policies and geographical distance to societies with lower
mean abilities) may be reflected into this difference. A similar pattern could be found for
Finland and Germany: The difference in the upper level is only 1.20 IQ-points (IQ 121 and
120), but at the lower level 9.60 IQ-points (IQ 85 and 76). Most likely different immigration
histories are reflected here, furthermore differences in educational policy (age of tracking, in
Germany between age 10 and 12, in Finland at age 16). Early tracking increases ability
variance. Using regression analysis (as predictors mean and lower level) the largest residual
(standing for difference between upper level and the rest) is found in South Africa (with its
heterogeneous population of European, Asian and African descent), inverted the largest
residual (standing for difference between lower level and the rest) is found in Belgium
(probably a result of immigration and educational policy).
The highest cognitive ability values (measured by education and educational degrees) of
leading politicians are found in 1. Austria (IQ 128), 2. Luxembourg (IQ 127) and 3. Malta (IQ
126), the lowest in Arab countries: 82. Tunisia (IQ 110), 83. Bahrain (IQ 107), 84. Algeria (IQ
107), 85. Kuwait (IQ 105), 86. Yemen (IQ 100), 87. Qatar (IQ 98), 88. Emirates (IQ 93), 89. Egypt
(IQ 90) and 90. Saudi-Arabia (IQ 90). In all these countries the leaders are not elected by their
population. Austria has the highest statesman cognitive ability level because up to the early
70s it was common (and almost universal) to complete studies with a doctorate (as it is still the
case with medicine in Germany and Austria).
Cross-sectional analyses
The upper and lower levels are highly correlated with the means (r=.97), among each other
with r=.90 (see Table 2). The correlation of the 95th percentile with cognitive ability of political
leaders is slightly higher (r=.364 and .365) than of the average competence values (r=.361
and .354) or of the 5th percentile (r=.345 and .324), indicating that the politicians’ ability
depend more on the smart fraction value. Political leaders most likely stem from this fraction
(and considering the complexity of political tasks this is a reasonable result; s. Suedfeld et al.,
2003). A regression analysis (criterion: IQ of politicians) increases this pattern (for politicians
1960-2009, average ability βaA→PIQ=.13, β95%→PIQ=.24; for politicians 1990-2009, same time
ability measures βaA→PIQ=.00, β95%→PIQ=.37, N=90).
Table 2. Correlations within cognitive ability indicators

CA- CA-
CA-mean CA-5% CA-95% politicians politicians
60-09 90-09
CA-mean (average) -
CA-5% .97 -
CA-95% .97 .90 -
CA-politicians 60-09 .36 .35 .36 -
CA-politicians 90-09 .35 .32 .37 .93 -
N 90 90 90 90 90
Note: See Table 1.
16
Table 3: Correlations between cognitive ability indicators and indicators of societal development
GDP HDI Patent Patent Nobel Nobel Nobel Scien- High Gov. Demo Demo- Rule Rule Politi- HIV AIDS
2003 2003 rate rate Peace Literat. Science tists techno- eff. cracy cracy of law of law cal rate 1995-
1960- 1991- 1901- 1901- 1901- 1985- logy 1996- 1950- 1996- 1970- 2000 liberty 2001- 2000
2007 2007 2004 2004 2004 95 1997 2005 2004 2000 2000 1999 2003
CA-mean (average) .61 .79 .40 .45 .21 .13 .34 .61 .38 .61 .60 .62 .62 .59 .49 -.30 -.21
CA-5% .56 .74 .35 .37 .15 .13 .31 .57 .35 .55 .55 .55 .56 .53 .42 -.31 -.24
CA-95% .61 .78 .42 .50 .23 .13 .34 .64 .40 .63 .62 .65 .66 .63 .53 -.24 -.20
CA-politicians 60-09 .13 .27 .22 .03 .02 .06 .10 .10 .15 .19 .52 .58 .22 .12 .55 -.05 .08
CA-politicians 90-09 .08 .25 .13 .04 .01 .04 .07 .16 .12 .12 .45 .57 .13 .02 .52 -.10 .05
N 84 85 81 76 87 87 87 51 61 88 84 84 76 73 86 82 83
Note: See Table 1; AIDS 1995-2000: percentage of people suffering on AIDS 1995-2000.
Table 4: Regression analyses, predictors: cognitive ability indicators, criteria: indicators of the development of societies
GDP HDI Patent Patent Nobel Nobel Nobel Scien- High Gov. Demo Demo- Rule Rule Politi- HIV AIDS
2003 2003 rate rate Peace Literat. Science tists techno- eff. cracy cracy of law of law cal rate 1995-
1960- 1991- 1901- 1901- 1901- 1985- logy 1996- 1950- 1996- 1970- 2000 liberty 2001- 2000
2007 2007 2004 2004 2004 95 1997 2005 2004 2000 2000 1999 2003
CA-mean (average) .22 .47 -.13 (-.75) -.35 .09 .13 -.12 -.29 -.12 .02 -.13 -.30 -.35 -.41 .03 .15
CA-5% -.34 -.36
CA-95% .40 .33 .55 (1.23) .57 .04 .22 .76 .68 .75 .59 .77 .95 .97 .93
N 84 85 81 76 87 87 87 51 61 88 84 84 76 73 86 82 83
Educational level .54 .56 .61 .58 .47

CA-politicians .27 .32 -.05 -.22 .32
N 83 83 73 71 84 82 83
Note: See Table 1 and 3. For educational level is used for long term criteria (democracy 1950-2004, rule of law 1970-2000) a general long-term educational index,
for more recent criteria (democracy 1996-2004, rule of law 2000 and political liberty1999) a current education index from HDR, the same was done with cognitive
ability estimations of politicians.
Smart fraction 17
The smart fraction (or gifted or high ability fraction) is not only more relevant than the average
cognitive ability level or the non-smart fraction’s level for the cognitive competence levels of
politicians, but also for wealth (GDP; see Tables 3 and 4). For the general human development
level of a society, the 95th is less important than the average, but more than the 5th level.
Even stronger are the effects on all indicators of high intellectual achievement from patent
rates, over Nobel Prizes to the rates of scientists and high technology exports. Only for Nobel
Prizes in Literature the impact is lower, presumably because of the low reliability and validity
of this judgement process (if it should rate literature quality) and because of political
considerations in giving this award (sometimes it seems to be an award given to nations not
receiving science Nobel Prizes, aiming at equal distribution across different cultures and
continents).
Also impressive is the high impact of the smart fraction (exactly: variance in indicators
between different countries statistically explained by differences in cognitive levels of the
95th percentile groups) on all political variables: on government effectiveness, democracy, rule
of law and political liberty. The political development of nations seems to depend heavily on
the cognitive ability of their smart fraction.
On the other hand the development in indicators standing for undesired outcomes (HIV and
AIDS) depends on the cognitive level of the lower fraction. This is a strong support for all
studies demonstrating that health and especially HIV depends on cognitive ability, leading to
more or less risky behavior resulting in the long run in health or death (e.g. Goldman & Smith,
2002; Gottfredson, 2004; Rindermann & Meisenberg, 2009). Additionally, the opposite effects
of the high ability fraction on intellectual outcomes and of the low ability fraction and their
abilities on anti-intellectual outcomes strongly support the theoretical framework: Cognitive
ability specifically matters through a cognitive effect mechanism. Finally, similar results for
crime (homicide) substantiate the pattern: Homicide is higher (negatively) correlated with
the ability level of the 5th percentile than with the average or even with the ability of the 95th
percentile (5th r=-.34, average r=-.33, 95th r=-.23; N=80). The regression analysis strengthens
the result: β5%→Hi=-.35, βaA→Hi=-.01, N=80). Because of this opposite effect pattern an objection,
that the results are a consequence of different reliability (high for 95th percentile ability, mean
for the average value, low for 5th percentile), could be ruled out.
The results for political leaders and their ability levels across the countries are less impressive:
Correlations are always low except for democracy and political freedom. For both variables
political leaders’ competence could explain the variance beneath the general educational
level of society (democracy: βEdu→Dem=.54, βPIQ→Dem=.27, N=83; political freedom: βEdu→PoF=.47,
βPIQ→PoF=.32, N=84).
.38 .60 .45 .86
Patent rate Nobel Prize in Scientist rate High tech

1960-2007 Science rate 1985-1995 export rate
1901-2004 1997
.78 (.78) .64 (.64) .74 (.74) .37 (.37)
GDP 1998 .01

Cognitive 1 (1)
ability of 95% .43 (.75)
Intellectual .94 (.94)
excellence Wealth
(STEM)
Cognitive .32 (.74)
ability mean .98 (.98)
GDP 2003 .04
.44 .12
Figure 3. Standardized path coefficients (and correlations in parentheses) between cognitive ability of
nations, high intellectual achievement and wealth (error term as unexplained variance), N=48 nations
For a path analysis four indicators of intellectual excellence in STEM subjects were used
(patent rate, Nobel Prizes in science rate, scientist rate and high technology export rate),
wealth was measured by GDP 1998 and 2003 (latent variables in circles, measured/manifest
in boxes; see Figure 3). Only countries with data in all measures were used (N=48), the
pattern of results is similar to the larger samples (see Tables 3 and 4). The fit indices are good
(SRMR=.04, CFI=.99). For high STEM achievement the cognitive competence of the smart
fraction rate is more important (β95%→IE=.43, βaA→IE=.32). The impact of cognitive ability on
wealth goes completely through intellectual excellence (βIE→GDP=.94), a further direct effect
would be small and even negative (βaA→GDP=-.05 and β95%→GDP=-.16), indicating, that cognitive
ability not going through STEM-like achievement has no positive impact on wealth (similar:
Schofer, Ramirez & Meyer, 2000).
Longitudinal analyses
Student assessment studies with information on the tails do not allow longitudinal analyses,
the time interval between the 90s and the first decade of 21st century is too short. But we can
do these analyses with political leaders’ cognitive ability from the 60s to the 00s. In the Tables
3 and 4 high correlations and a cross-sectional impact of politicians’ ability on democracy
were observable. Maybe this relationship does not include an effect of ability on democracy,
but of democracy on politicians’ ability. This could be tested by a cross-lagged panel analysis.
Cognitive ability Cognitive ability

of political of political
leaders 1960-64 leaders 2000-09
Democracy Democracy
1960-64 1996-2000
Figure 4. Cross-lagged coefficients (and correlations in parentheses) between cognitive ability of

political leaders and democracy (error term as unexplained variance, correlated error), N=64 nations

of political of political
leaders 1960-69 leaders 2000-09

mean mean
1964-1972 1995-2007
Figure 5. Cross-lagged coefficients (and correlations in parentheses) between cognitive ability of

political leaders and society (error term as unexplained variance), N=17 nations
Smart fraction 19
The relation between politicians’ ability and democracy could be analyzed between the early
60s and the 00s (see Figure 4; the fit of the model is perfect). The years for political leaders
are somewhat misleading because the majority of leaders in 1960-64 had ruled the country
also for several years before, the same is true for 2000-09. The cognitive ability of politicians
has only a small impact on the development of democracy (βPIQ1→Dem2=.04), but democracy
leads to more educated and competent politicians (βDem1→PIQ2=.18). The low competence
levels of Arab leaders described above could at least be partly explained by their
nondemocratic (self-)appointment (and only under rare circumstances non-democratic
leaders like Mohammed VI in Morocco and bin Said in Oman could push their societies
forward).
But politicians’ ability has a positive impact on cognitive development of a society
(βPIQ1→CA2=.21; see Figure 5; SRMR=.01, CFI=1.00). The reverse effect is smaller
(βCA1→PIQ2=.03). Unfortunately the sample with 17 countries is small, but after wealth and rule
of law it is the third factor that longitudinally seems to influence cognitive development at the
country data level. In the long run smart politicians increase their nations’ abilities.
Cognitive ability and educational policy

Large class sizes affect low ability fractions stronger than high ability fractions, the opposite is
true for early tracking, which shows larger benefits for high ability fractions and smaller for
low ability fractions (see Table 5). Both results are in accordance with assumptions and studies
at the national and within-national data level (Rindermann & Ceci, 2009): Less able students
need smaller classes (Krueger, 1999). Early tracking is more beneficial for highly capable
students because they will be together with comparable high able students leading to a more
stimulating instruction, interaction and school climate, the ability of others nurtures one’s own
ability (Rindermann & Heller, 2005), but less able students lose their possible models. Both
patterns of results strongly support former analyses at the country data level.
Rather astonishing is the high correlation between early tracking and cognitive ability of
political leaders (r=.41-.44, N=66; compare them with correlations in Table 3). Maybe in
educational systems with early tracking the political elite stems from the – cognitive
development stimulating – higher tracks and higher education is seen as something
prestigious and valuable for a political career within parties, society and voters (furthering
and gate keeper effect of early tracking). The other way around it may be that more intelligent
and competent politicians are in favor of tracking systems (which seem to benefit cognitive
development). Future longitudinal analyses should prove this.
Table 5: Correlations between cognitive ability indicators and attributes of educational policy
Young tracking age Class size
CA-mean .31 -.45
CA-5% .25 -.47
CA-95% .33 -.38
CA-politicians 60-09 .41 -.19
CA-politicians 90-09 .44 -.22
N 66 74
Note: See Table 1; Young tracking age: tracking age inverted; Class size: large class size and pupil-
teacher-ratio.
Discussion
As other studies before have shown, there are large cognitive ability differences between
societies. They are reflected in the upper and lower levels, but East Asian and Commonwealth
countries seem to have comparatively stronger smart fractions than Scandinavian or other
First-World-countries. Maybe these countries not only benefit from their educational policies
and, especially in East-Asia, from harsh educational efforts with very extensive studying times
per day, week, year and youth, but also from a more successful migration policy. Beneath
cultural background factors relevant for education in family and school genetic factors could
be important, but genes for cognitive ability are not known up to now.
The ability value (intelligence, knowledge and the intelligent use of knowledge) of the smart
fraction (95th percentile, comparable to an IQ 125 or higher in within-country norms) is more
important for country differences in wealth, nations’ intellectual excellence (in STEM fields:
patents, Nobel Prices in science, scientists, high technology exports) and political attributes
of societies (government effectiveness, democracy, rule of law and political liberty) than the
average ability or the ability level of a non-smart fraction (5th percentile, comparable to an IQ
75 or lower in within-country norms). But the cognitive ability level of the non-smart fraction is
more important for country differences in HIV, AIDS and homicide. Wealth differences
between countries could be completely explained through differences in high intellectual
achievement in STEM fields, which itself largely depends on differences in smart fractions
ability. The smart fraction is essentially relevant for beneficial societal development.
The cognitive ability of political leaders is far less important. We could only find higher
correlations to democracy and political liberty, in a longitudinal analysis democracy has a
positive impact on cognitive ability of political leaders. People, if they have the chance to
elect their leaders, prefer more educated ones. Political leaders have, in the long run, a
positive influence on countries’ cognitive ability, presumedly by creating better educational
and social environments increasing cognitive ability.
This study could show how in former studies discovered mean cognitive ability effects on
growth could work: In societies with a higher cognitive average the smart fraction reaches a
higher cognitive level (resp. is also larger from a “real” threshold of high ability on – see
below). This smart fraction pushes growth through excellence in areas relevant for economic
affluence, like in technology and science. We did not expect such a high impact of the smart
fraction on the destiny of societies. The current data do not allow us to present a historical
analysis of modernization processes from the ancient past up to 1960. We can only use our
findings on present day comparisons between countries in terms of cognitive ability as an
analogy to surmise that the same cognitive effects occurred during historical periods. It
seems very likely that the achievements of cognitively eminent persons coming from the
smart fraction, and stimulated by their peers, was decisive for the betterment of their
societies.
But the results are somewhat contradictory: On one hand, international differences in
important attributes of societies (and, we would argue, in the historical development of
nations) depend on the cognitive ability of an elite. On the other hand these differences do
not depend on the cognitive ability levels of politicians. Peoples and nations themselves seem
to be important for the destiny of their countries and societies, more so than politicians.
Leaders stem by a majority from smart fractions of their societies and they seem to reflect
social attributes and especially their cognitive ability levels, rather than influencing them.
One possibility is that in developed societies with a large smart fraction, political leadership
is drawn from somewhat less intelligent elite members (see Cox, 1926, p. 84), willing to earn
less intellectual merits or money in return for the possibility of fame. Another possibility is
that the cognitive ability of politicians is less important because they have competent
consultants and experts who give advice (the first author discussed this in 2006 at an ISIR-
meeting in San Francisco with Dean Keith Simonton, his reply was, that politicians also need to
understand them). Maybe formal education is not the best estimate, better would be real life
criteria like used by Simonton. But we should not use verbal ability here. Rhetorical brilliance
Smart fraction 21
is not solving problems. Verbal fluency could be a dangerous competence, a seduction to talk
people into believing or doing something that they would not do by the use of thinking and
rationality. In the Greek tradition, Plato (Gorgias) similarly assessed rhetorical competence in
a very critical manner. Finally, in German analyses political and “weltanschauliche”
orientations are showing more explanative strength than education of politicians.
One remark on Singapore: Its long-term Prime Minister Lee Kuan Yew has in our data set only
as highest degree “university degree”, no doctorate or an additional scientific degree. But he
has studied at London School of Economics and in Cambridge and finished his studies with
exceptional “Double Starred First Class Honours”. Our assessment procedure seems to
underestimate his cognitive ability level. Furthermore, Singapore has reached the highest
rank in the smart fraction ability (IQ 127) in our list and the second highest rank in average
ability (IQ 105), but “only” the 14th rank in the lower non-smart fraction ability ranking (IQ 79).
In patents (1991-2007) Singapore has the first place. And, that is especially remarkable,
Singapore has reached the first place in government effectiveness. Singapore seems to have
the best government in the world. Lee Kuan Yew’s ability – indicated by his success – seems to
be underestimated again by the solely use of formal education. Of course, Singapore was and
still is no standard-bearer of liberty and democracy and Lee Kuan Yew has attracted criticism
because of this. But, he stands apart from other leaders in terms of his exceptional success for
Singapore in growth, modernization, technology and since several years also in science (up to
now only STEM including biotechnology). Lee is also apparently the only politician who has
read and used the results of intelligence research in his politics. In speeches he has cited
Thomas Bouchard and Richard Lynn (Chan & Chee, 1984), and he is the only statesman, who
has seen that intelligence enhancement not only needs an improvement in the environment
(like in educational policy) but also in demographic policies, because parents transfer
cognitive ability to their children by creating a stimulating environment (especially by
education and modeling) and by transmission of their genes.
Further research should not only use as an indicator for smart fractions a value like the 95th
percentile and the ability at this level, but also a defined threshold of cognitive ability like IQ
130 or SAS 700 and the percentage of population above this threshold. The present editing of
TIMSS, PISA and PIRLS results complicates this. Research using different ability levels seems
to be a fruitful approach. Also further indicators of high achievement like Fields Medal or of
low achievement like traffic accidents should be used as indicators to stand for high or low
intelligence (Dagona, 1994; O‘Toole & Stankov, 1992). Migrations and their impact on
countries’ cognitive ability and development should be analyzed (so Singapore benefits from
a smart fraction import, similarly to Switzerland, Australia and Canada, but Eastern Europe
and less so Germany suffer from a brain flight, others from a low brain immigration; e.g.
Levels, Dronkers & Kraaykamp, 2008; te Nijenhuis, de Jong, Evers & van der Flier, 2004). The
possibilities of education for cognitive enhancement are still not sufficiently explored (e.g.
Heckman, 2000; Nisbett, 2009; Rindermann & Ceci, 2009). Our results emphasize the
importance of nurturing the highly gifted. Their support will be beneficial in the long run not
only for themselves but even more through general effects on societal development including
wealth, health, politics, science, ethics and culture for the less smart and non-smart fractions.
Finally, at the level of societies, there is an integrative theoretical framework on causes and
mechanisms of cognitive competence still missing. By showing the influence of ability on
wealth through high intellectual STEM achievement we hope to provide some further building
block in development of such a theory.
References
Bacharach, V. R., & Baumeister, A. A. (1998). Direct and of general intelligence (pp. 343-365). Oxford:
indirect effects of maternal intelligence, Pergamon.
maternal age, income, and home environment Falk, R., & Greenbaum, Ch. W. (1995). Significance
on intelligence of preterm, low-birth-weight tests die hard. The amazing persistence of a
children. Journal of Applied Developmental probabilistic misconception. Theory &
Psychology, 19, 349-363. Psychology, 5, 75-98.
Beaton, A. E., Mullis, I. V. S., Martin, M. O., Gonzalez, E. Freedom House (2004). Freedom in the world.
J., Kelly, D. L., & Smith, T. (1996). Mathematics Washington: www.freedomhouse.org.
achievement in the middle school years.
Chestnut Hill: TIMSS Study Center. Gelade, G. A. (2008). IQ, cultural values, and the
technological achievement of nations.
Beaton, A. E., Martin, M. O., Mullis, I. V. S., Gonzalez, E. Intelligence, 36, 711-718.
J., Smith, T., & Kelly, D. L. (1996). Science
achievement in the middle school years. Gigerenzer, G. (2004). Mindless statistics. Journal of
Chestnut Hill: TIMSS Study Center. Socio-Economics, 33, 587-606.
Becker, G. S. (1993/1964). Human capital: A Goldman, D. P., & Smith, J. P. (2002). Can patient self-
theoretical and empirical analysis with special management help explain the SES health
reference to education. Chicago: University of gradient? Proceeding of the National Academy of
Chicago Press. Sciences, 99, 10929-10934.
Cattell, R. B. (1987/1971). Intelligence: Its structure, Gottfredson, L. S. (2004). Life, death, and intelligence.
growth and action. Amsterdam: Elsevier. Journal of Cognitive Education and Psychology, 4,
23-46.
Ceci, S. J. (1991). How much does schooling influence
general intelligence and its cognitive Gottfredson, L. S. (2005). Implications of cognitive
components? Developmental Psychology, 27, differences for schooling within diverse
703-722. societies. In C. L. Frisby & C. R. Reynolds (Eds.),
Comprehensive Handbook of Multicultural
Chan, Ch. K. & Chee, H. L. (1984). Singapore 1984: School Psychology (pp. 517-554). New York:
Breeding for Big Brother. In Ch. K. Chan & H. L. Wiley.
Chee (Eds.), Designer genes: IQ, ideology and
biology (pp. 4-13). Kuala Lumpur: Gwartney, J., & Lawson, R. (2003). Economic freedom
Selangor/Institute for Social Analysis (INSAN). of the world. 2003 annual report. Vancouver:
Fraser Institute.
Chomsky, N. (2009). Politics, society, and Obama:
Change we can believe in? The Owl, 15, 16-21. Habermas, J. (1985/1981). The theory of communi-
cative action. Boston: Beacon.
Cohen, J. (1994). The earth is round (p<.05). American
Psychologist, 49, 997-1003. Hanushek, E. A., & Kimko, D. D. (2000). Schooling,
labor-force quality, and the growth of nations.
Cox, C. M. (1926). The early mental traits of three American Economic Review, 90, 1184-1208.
hundred geniuses. Stanford: Stanford University
Press. Hart, M. H. (2007). Understanding human history.
Augusta: Washington Summit.
Cunha, F., Heckman, J. J., Lochner, L., & Masterov, D. V.
(2006). Interpreting the evidence on life cycle Heckman, J. J. (2000). Policies to foster human capital.
skill formation. In E. A. Hanushek, & F. Welch Research in Economics, 54, 3-56.
(Eds.), Handbook of the economics of education Herrnstein, R. J., & Murray, Ch. (1994). The bell curve.
(I, pp. 697-812). Amsterdam: North-Holland. Intelligence and class structure in American life.
Dagona, Z. K. (1994). Substance use and road traffic New York: Free Press.
accidents among Nigerian commercial motor Hu, L., & Bentler, P. M. (1998). Fit indices in covariance
cyclists. Ife PsychologIA, 2, 81-93. structure modeling. Psychological Methods, 3,
Deary, I. J., Batty, G. D., & Gale, C. R. (2008). Childhood 424-453.
intelligence predicts voter turnout, voting Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit
preferences, and political involvement in indexes in covariance structure analysis.
adulthood. Intelligence, 36, 548-555. Structural Equation Modeling, 6, 1-55.
Deary, I. J., Ferguson, K. J., Bastin, M. E., Barrow, G. W. Human Development Report of the United Nations
S., Reid, L. M., Seckl, J. R., Wardlaw, J. M., & (HDR) (2005). Human Development Report 2005.
Maclullich, A. M. J. (2007). Skull size and New York. Retrieved May 25, 2007 from
intelligence, and King Robert Bruce‘s IQ. http://hdr.undp.org/reports/global/2005/pdf/h
Intelligence, 35, 519-525. dr05_HDI.pdf.
Denny, K., & Doyle, O. (2008). Political interest, Hunter, J. E. (1997). Needed: A ban on the
cognitive ability and personality: Determinants significance test. Psychological Science, 8, 3-7.
of voter turnout in Britain. British Journal of
Interpol (2004). International Crime Statistics. Lyon.
Political Science, 38, 291-310.
Retrieved October 22, 2004 from
Ellis, L., & Walsh, A. (2003). Crime, delinquency and www.interpol.int/Public/Statistics/ICS/
intelligence: A review of the worldwide downloadList.asp.
literature. In H. Nyborg (Ed.), The scientific study
Smart fraction 23
Irwing, P., & Lynn, R. (2006). The relation between TIMSS & PIRLS Study Center.
childhood IQ and income in middle age. Journal Martin, M. O., Mullis, I. V. S., Gonzalez, E. J., &
of Social, Political and Economic Studies, 31, 191- Chrostowski, S. J. (2004). TIMSS 2003
196. international science report. Chestnut Hill: TIMSS
Jensen, A. R. (1998). The g factor. The science of & PIRLS International Study Center.
mental ability. Westport: Praeger. Martin, M. O., Mullis, I. V. S., Gonzalez, E. J., Gregory, K.
Jones, G., & Schneider, W. J. (2006). Intelligence, D., Smith, T. A., Chrostowski, S. J., Garden, R. A., &
human capital, and economic growth. Journal of O’Connor, K. M. (2000). TIMSS 1999 international
Economic Growth, 11, 71-93. science report. Chestnut Hill: TIMSS Study
Kaufmann, D., Kraay, A., & Mastruzzi, M. (2006). Center.
Governance matters V: Aggregate and individual Meisenberg, G. (2004). Talent, character, and the
governance indicators for 1996-2005. Worldbank. dimensions of national culture. Mankind
Retrieved November 28, 2006 from Quarterly, 45, 123-168.
www.worldbank.org/wbi/governance. Mullis, I. V. S., Martin, M. O., Beaton, A. E., Gonzalez, E.
Knack, S., & Keefer, Ph. (1995). Institutions and J., Kelly, D. L., & Smith, T. (1997). Mathematics
economic performance. Economics and Politics, achievement in the primary school years.
7, 207-227. Chestnut Hill: TIMSS Study Center.
Kohlberg, L. (1987). Child psychology and childhood Mullis, I. V. S., Martin, M. O., & Foy, P. (2008). TIMSS
education. A cognitive-developmental view. 2007 international mathematics report. Chestnut
New York: Longman. Hill: TIMSS & PIRLS Study Center.
Krueger, A. B. (1999). Experimental estimates of Mullis, I. V. S., Martin, M. O., Gonzalez, E. J., Gregory, K.
education production functions. Quarterly D., Garden, R. A., O’Connor, K. M., Chrostowski,
Journal of Economics, 114, 497-532. S. J., & Smith, T. A. (2000). TIMSS 1999
Kurian, G. Th. (2001). The illustrated book of world international mathematics report. Chestnut Hill:
rankings. Armonk: Sharpe. TIMSS Study Center.
La Griffe du Lion (2004). Smart fraction theory II. 6(2). Mullis, I. V. S., Martin, M. O., Gonzalez, E. J., &
Retrieved July 13, 2009 from Chrostowski, S. J. (2004). TIMSS 2003
www.lagriffedulion.f2s.com/sft2.htm. international mathematics report. Chestnut Hill:
TIMSS & PIRLS International Study Center.
Lakhanpal, M., & Ram, R. (2008). Educational
attainment and HIV/AIDS prevalence: A cross- Mullis, I., Martin, M. O., Gonzales, E. J., & Kennedy, A.
country study. Economics of Education Review, M. (2003). PIRLS 2001 international report.
27, 14-21. Chestnut Hill: IEA.
Lee, J.-W., & Barro, R. J. (1997). Schooling quality in a Mullis, I., Martin, M. O., Kennedy, A. M., & Foy, P.
cross-section of countries. New York. Retrieved (2007). PIRLS 2006 international report. Chestnut
November 18, 2004 from Hill: IEA/TIMSS & PIRLS International Study
www.nber.org/papers/W6198. Center.
Levels, M., Dronkers, J., & Kraaykamp, G. (2008). Murray, Ch. (2008). Real education: Four simple truths
Immigrant children’s educational achievement for bringing America’s schools back to reality.
in Western countries. American Sociological New York: Crown Forum.
Review, 73, 835-853. Murray, Ch., Cardoso, R., & Mendes, D. (2008).
Lynn, R., Harvey, J., & Nyborg, H. (2009). Average Miscigenação diminui o QI dos brasileiros.
intelligence predicts atheism rates across 137 Istoé, 2032(15-10-08), 9-11.
nations. Intelligence, 37, 11-15. Nisbett, R. E. (2009). Intelligence and how to get it.
Lynn, R., & Longley, D. (2006). On the high Why schools and cultures count. New York:
intelligence and cognitive achievements of Jews Norton.
in Britain. Intelligence, 34, 541-547. Nyborg, H. (2009). The intelligence-religiosity nexus.
Lynn, R., & Vanhanen, T. (2002). IQ and the wealth of Intelligence, 37, 81-93.
nations. Westport: Praeger. OECD (2000). Literacy in the information age. Final
Lynn, R., & Vanhanen, T. (2006). IQ and global report of the International Adult Literacy Survey.
inequality. Athens: Washington Summit. Paris: OECD.
Marshall, M. G., & Jaggers, K. (2000). Polity IV Project. OECD (2003). Literacy skills for the world of
Political regime characteristics and transitions, tomorrow. Further results from PISA 2000. Paris:
1800-1999. University of Maryland: OECD.
www.bsos.umd.edu/cidcm/inscr/polity. OECD (2004a). Learning for tomorrow‘s world. First
Martin, M. O., Mullis, I. V. S., Beaton, A. E., Gonzalez, E. results from PISA 2003. Paris: OECD.
J., Smith, T., & Kelly, D. L. (1997). Science OECD (2004b). Problem solving for tomorrow‘s
achievement in the primary school years. world. Paris: OECD.
Chestnut Hill: TIMSS Study Center. OECD (2007a). PISA 2006. Science competencies for
Martin, M. O., Mullis, I. V. S., & Foy, P. (2008). TIMSS tomorrow‘s world. Volume 1: Analysis. Paris:
2007 international science report. Chestnut Hill: OECD.
OECD (2007b). PISA 2006. Volume 2: Data/Données. effects of science on national economic
Paris: OECD. development, 1970-1990. American Sociological
Oesterdiekhoff, G. W. (2000). Zivilisation und Review, 65, 866-887.
Strukturgenese. Frankfurt: Suhrkamp. Simonton, D. K. (1984). Genius, creativity and
Oesterdiekhoff, G. W. (2008). Magic and animism in leadership. Cambridge: Harvard University
old religions: The relevance of sun cults in the Press.
world-view of traditional societies. Croatian Simonton, D. K. (1985). Intelligence and personal
Journal of Ethnology and Folklore Research, 45, influence in groups: Four nonlinear models.
43-66. Psychological Review, 92, 532-547.
Oesterdiekhoff, G. W., & Rindermann, H. (2007). The Simonton, D. K. (2006). Presidential IQ, openness,
spread of AIDS in developing countries: A intellectual brilliance, and leadership: Estimates
psycho-cultural approach. Journal of Social, and correlations for 42 US chief executives.
Political and Economic Studies, 32, 201-222. Political Psychology, 27, 511-526.
O‘Toole, B. I., & Stankov, L. (1992). Ultimate validity of Simpson, M. (1997). Informational inequality and
psychological tests. Personality and Individual democracy in the new world order. In M.
Differences, 13, 699-716. Midlarsky (Ed.), Inequality, democracy, and
Piaget, J. (1997/1932). Moral judgement of the child. economic development (pp. 156-176).
New York: Free Press. Cambridge: Cambridge University Press.
Piaget, J. (2001/1947). The psychology of intelligence. Statistisches Bundesamt (2002). Statistisches Jahrbuch
London: Routledge. 2002 für das Ausland. Suttgart: Metzler-Poeschel.
Rindermann, H. (2007a). The g-factor of international Suedfeld, P., Guttieri, K., & Tetlock, Ph. E. (2003).
cognitive ability comparisons: The homogeneity Assessing integrative complexity at a distance.
of results in PISA, TIMSS, PIRLS and IQ-tests In J. M. Post (Ed.), The Psychological assessment
across nations. European Journal of Personality, of political leaders (pp. 246-270). Ann Arbor: The
21, 667-706. University of Michigan Press.
Rindermann, H. (2007b). The big G-factor of national te Nijenhuis, J., de Jong, M.-J., Evers, A., & van der
cognitive ability. European Journal of Personality, Flier, H. (2004). Are cognitive differences
21, 767-787. between immigrant and majority groups
diminishing? European Journal of Personality, 18,
Rindermann, H. (2008a). Relevance of education and 405-434.
intelligence at the national level for the
economic welfare of people. Intelligence, 36, Thomson, G. H. (1937/1936). Intelligence and
127-142. civilisation: A Ludwig Mond lecture delivered at
the University of Manchester on October 23rd,
Rindermann, H. (2008b). Relevance of education and 1936. Journal of the University of Manchester, 1,
intelligence for the political development of 18-38. Reprinted in: Deary, I. J., Lawn, M., Brett,
nations: Democracy, rule of law and political C. E., & Bartholomew, D. (2009). „Intelligence
liberty. Intelligence, 36, 306-322. and Civilisation“. Intelligence, 37, 48-61.
Rindermann, H., & Ceci, S. J. (in press). Educational UNAIDS/WHO (2003). AIDS epidemic update.
policy and country outcomes in international Geneva: WHO/www.unaids.org.
cognitive competence studies. Perspectives on
Psychological Science. Vanhanen, T. (1997). Prospects of democracy. London:
Routledge.
Rindermann, H., & Heller, K. A. (2005). The benefit of
gifted classes and talent schools for developing Vanhanen, T. (2005). Measures of democracy 1810-
students’ competences and enhancing 2004. Finnish Social Science Data Archive (FSD).
academic self-concept. Zeitschrift für Retrieved January 17, 2006 from www.fsd.uta.fi/
Pädagogische Psychologie, 19, 133-136. english/data/catalogue/FSD1289.
Rindermann, H., & Meisenberg, G. (2009). Relevance Weede, E. (2006). Economic freedom and
of education and intelligence at the national development. Cato Journal, 26, 511-524.
level for health: The case of HIV and AIDS. Weiss, V. (2009). National IQ means transformed from
Intelligence, 37, 383-395. Programme for International Student
Rindermann, H., & Thompson, J. (2009). Parents’ Assessment (PISA) scores, and their underlying
education, and not their money, nurtures the gene frequencies. Journal of Social, Political and
intelligence of their children. Manuscript Economic Studies, 34, 71-94.
submitted for review. World Intellectual Property Organization (WIPO)
Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2009). Patent applications by patent office (1883-
(2003). Evaluating the fit of structural equation 2007), by resident and non-resident. Geneva.
models. Methods of Psychological Research, 8, Retrieved June 22, 2009 from
23-74. www.wipo.int/ipstats/en/statistics/patents.
Schmidt, F. L., & Hunter, J. E. (2004). General mental Wuttke, J. (2007). Uncertainties and bias in PISA. In S.
ability in the world of work. Journal of Personality Th. Hopmann, G. Brinek & M. Retzl (Eds.), PISA
and Social Psychology, 86, 162-173. zufolge PISA - PISA according to PISA (pp. 241-
263). Wien: LIT.
Schofer, E., Ramirez, F. O., & Meyer, J. W. (2000). The
Smart fraction 25
The authors
Heiner Rindermann (born 1966) is Professor for Developmental Psychology at

University of Graz (Austria). He studied psychology and earned his Ph.D. in
Heidelberg (1986-1995). 1994-1999 he worked together with Kurt A. Heller
(LMU Munich) in two research projects on evaluation of programs for the gifted.
1999-2008 he taught as assistant and associate professor psychological
methods (Magdeburg) and educational science (Saarbrucken, Paderborn). His
research topics are development, furthering and consequences of intelligence
at the individual, school and society level, emotional competence, teaching
quality and its evaluation, cross-cultural psychology and educational policy.
Michael Sailer (born 1984) graduated from Secondary School in Leoben and
studied psychology in Graz 2003-2009. He is currently working on his thesis
about the “Impact of politicians and the cognitive ability of smart fractions on
social development.” Since 2007 he is a member for the KFU-Graz students’
association.
Dr James Thompson is Honorary Senior Lecturer in Psychology in the

Department of Psychology, University College London. Currently he teaches
medical students psychology applied to medicine. His clinical work is in
the field of post-traumatic stress disorder. His original thesis was on
the intelligence and scholastic abilities of people who had sustained
localized cortical injuries in childhood. His current interests are on the
ways in which general intelligence can be used as an explanatory variable
in a range of social behaviors, particularly health behaviors.
Talent Development & Excellence Modelling mathematical Actiotopes with CLARION 27
Vol. 1, No. 1, 2009, 27-43
Modeling mathematical Actiotopes: The potential role of

CLARION
Shane N. Phillipson1,* and Ron Sun2
Abstract: The challenges facing research into giftedness are considerable, including the
lack of a clear meta-theoretical framework and agreement on terminology. If giftedness, in
broad terms, refers to the interactions between complex patterns of actions, the
psychological aspects of self and the social environment toward the attainment of
achievement excellence, then it is not surprising that the study of giftedness presents a
number of conceptual and practical difficulties. Using mathematical achievement as an
example, this article explores the potential benefits and problems of modeling the
Actiotope Model of giftedness using the cognitive architecture CLARION. In using
simulation studies to explore giftedness, the meta-theoretical framework, including
terminologies, is clearly defined, helping to direct future field studies using the more
traditional methods of research.
Keywords:
achievement excellence, Actiotope Model of Giftedness, CLARION, cognitive architecture,
mathematics, modeling
Introduction
The current state of research into giftedness has been described by a number of researchers
as fragmented and contradictory. The aim of this article is to argue that the Actiotope Model of
Giftedness (AMG) (Ziegler, 2005) provides a much needed sense of coherence and unity to
this research. The AMG focuses on the development of action repertoires necessary for the
attainment of excellence. In broad terms, the development of action repertoires is dependent
on physiological, psychological and environmental (or social) variables, and excellence is
viewed within social contexts.
Although the AMG is difficult to test experimentally using traditional research methods
because of its reliance on these different conceptual levels, we also argue that it is possible to
simulate the AMG using techniques developed by researchers currently working in the area
of cognitive modeling. Using excellence in the domain of mathematics as an example, we
describe the CLARION cognitive architecture as one possible way to model the development
of higher forms of mathematical thinking from its antecedents of innate number sense. In
describing the general principles of both the AMG and CLARION, this preliminary discussion
then shows how the two can be mapped onto each other and the immediate challenges in
using this approach and thereby providing some suggestions for future research.
Conceptions of giftedness
In reviewing the status of research into giftedness, Eyre (2009) pointed out that, despite the
many decades of work, conceptions of giftedness amongst psychologists are fragmented and
often contradictory. Recent volumes that specifically address different conceptions of
giftedness (Phillipson & McCann, 2007; Shavinina, in press; Sternberg, 2004b; Sternberg &
Davidson, 2005), the related topics of intelligence (Sternberg, 2000), creativity (Kaufman &
Baer, 2006; Kaufman & Sternberg, 2006) and terms such as “high ability”, “talent” and
1
Shane N. Phillipson, Department of Educational Psychology, Counselling and Learning
Needs, The Hong Kong Institute of Education, 10 Lo Ping Road, Tai Po, NT Hong Kong
*
Corresponding author. Email: shane@ied.edu.hk
2
Ron Sun, Cognitive Science Department, Rensselaer Polytechnic Institute, Troy, NY, USA.

28 S. N. Phillipson & R. Sun
“precocity”, for example, would tend to confirm this impression. Given this status, it is not
surprising that the broad field appears confusing to researchers working either within or
outside the field. It is this confusion that provides detractors with ample opportunity to
criticize and, in some cases, deride the research into giftedness altogether (Marsh, 1998;
Seaton et al., 2008).
If conceptions of giftedness amongst psychologists are numerous and fragmented, it is not
surprising that the practice of gifted education is fragmented. The reasons for the
development of the three broad “educational paradigms” (Eyre, 2009) in gifted education lie
partly in the different agendas of the various educational systems around the world, and
partly in the different conceptions of giftedness that are adopted by the policy makers
working within these systems. In Hong Kong, for example, policy documents supporting
gifted education, either explicitly or implicitly, simultaneously refer to multiple intelligences,
creativity, the Marland report, IQ and Gagné’s Differentiated Model of Giftedness and Talent
as conceptions of giftedness. The documents also mention disabilities and underachievement
as characteristics of some gifted individuals (Phillipson et al., in press).
Using the Hong Kong context again, the policy of including as many conceptions of giftedness
is politically attractive. Being “inclusive” means that the cohort of gifted students can be as
wide as possible, reducing the likelihood of alienating certain members of the community
(Eyre, 2009). As Eyre pointed out, the cohort approach depends on the prior identification of
students who are gifted in some way and depend, therefore, on psychological and/or
personality commonalities amongst these students. Perhaps the best known examples of
research involving the cohort paradigm are those that centre on mathematical precocity
(Lubinski & Benbow, 2006). Significantly, this research has identified some of the important
environmental and personal components that work together to enable the attainment of
mathematical excellence.
Eyre (2009) has identified “definitions” as one of the key questions for future research in
gifted education. It is timely that psychologists researching in the field of giftedness begin to
think about how to bring coherence to the diversity of conceptions of giftedness. This
coherence needs to take into account cultural diversity as well as personality variables such
as differences in measured intelligence, creativity and motivation, for example, and
differences in conceptions of achievement. As Eyre concluded, the field must reorganize itself
if it is to move beyond the “educational margins” and take its rightful place in identifying the
“conditions that enable and encourage exceptional ability to flourish” (pp. 17, 18).
Eyre’s (2009) concerns are not new. Ziegler (2005) has argued that most of the theoretical
conceptions of giftedness do not have a sufficient empirical basis, that the relationship
between gifts and achievement is often unclear, and that research methods and definitions are
often characterized by tautologies. In discussing the relationship between intelligence and
giftedness, for example, Callahan (2000) observed that the commonly cited conceptions of
giftedness such as Renzulli’s three-ring definition, Tannenbaum’s psychosocial definition,
Sternberg’s triarchic and WICS models, and Gardner’s multiple intelligences model rely on
different explanatory levels, such as cognitive skills, expressions of appropriate sets of
behaviors, production of knowledge, bio-physical potential, personality, luck and culture.
Furthermore, combinations of these different explanatory levels sometimes exist in the one
model. Of course, models that rely on different explanatory levels are not easily falsifiable.
Last, Ziegler and Heller (2000) pointed out that much of the reported research to date lacks
the rigor normally expected of scientific research, such as appropriate control groups. Heller
and Schofield (2000) also concluded from their content analysis of international trends on
research on giftedness that the field is overwhelmingly dominated by applied rather than
basic research, and that there was an urgent need for improvements in the “quality of
research designs and measurement techniques …[and] … most importantly, basic research
… [requiring] the intensification of cooperation with researchers” (pp. 136-137). Although
there is an urgent need for a systematic analysis of research since 1999, Eyre (2009)
concluded that the situation has probably not changed very much since that time.
Modelling mathematical Actiotopes with CLARION 29
Generally, research on giftedness centers on exceptional performance of some kind, and on

the antecedents and processes that lead to exceptional performance. Despite the problems in
research approaches and methods, there is broad consensus that giftedness involves
interactions between aspects of self (such as intelligence, cognitive processes, self-efficacy
and motivation, for example) and the environment (parents, education, and culture, for
example) that help propel the individual or the group to exceptionality (Sternberg, 2004a).
Last, exceptional performance is contextual, meaning that it refers to different things at
different times and for different sociocultural groups (Phillipson & McCann, 2007).
The critical state view of giftedness

Each of the psychologists working within the field of giftedness brings to their research, of
course, their individual prejudices and preferences. This should be encouraged since it is the
source of debate and the engine of further research. In order to bring a much needed unity to
this debate, Ziegler and Heller (2000) described a meta-theoretical framework for research
into giftedness. Termed the critical state view of giftedness (CSG), the framework describes a
probabilistic rather than deterministic relationship between the critical state (CS) and
achievement excellence (AE). Boundary conditions operate directly on the CS, to either
promote or decrease the probability of an individual reaching AE (Ziegler & Heller). The term
talented refers to someone who has the necessary psychological components, but not yet
fulfilled the conditions of the CS. Gifted refers to someone who has fulfilled the conditions of
the CS but, because of unfavorable boundary conditions, not yet reached AE. Expert refers to
persons who have reached AE.
The two stages of the CSG refer to the temporal relationship between the critical state (CS)
and achievement excellence (AE) (Figure 1). “Gifts” are the psychological components within
the CS but there are other aspects of personality such as motivational states with all being
necessary and sufficient for AE. Individuals achieve AE after fulfilling the conditions within
the CS. Within any domain there can be many types of AE and there may be many ways to
achieve the any one AE.
Achievement excellence (AE) is at first glance the most contentious area of the CSG. Ziegler
and Heller (2000) proposed that AE needs to be understood using a sequential series of
problems beginning with a clarification of domain, focus, frame of reference and significance
respectively. Ultimately, however, solutions to these problems are more likely to be found by
persons working within rather than outside the domain (Phillipson & Callingham, in press).
Development,
learning processes
CS a
Critical State (CS) a

CS b
Personality characteristics Achievement
inducive to gifts (unnecessary, eminence in
sufficient) CS c
Domain X
Gift (insufficient, non-
redundant)
Environmental conditions CS n
Figure 1. Achievement excellence in Domain X originates from one of Critical State (CS) a , CS b … CS
n under optimal boundary conditions for each critical state. Each CS operates independently of the
others and the gift, a psychological process, fulfills the inus condition of Postulate 2. (Adapted from
Ziegler & Heller, 2000, p. 15).
Nature of gifts
Within each CS, the components of interest are the psychological processes. Ziegler and
Heller (2000) referred to these processes as gifts, hence the study of giftedness should focus
on the processes that lead to AE. The gift is a necessary but in itself insufficient for AE.
According to Ziegler and Heller (2000), gifts must satisfy four postulates:
Postulate 1 is that the gift temporally precedes the AE and there must be a causal and non-
trivial relationship between the gift and the achievement. Postulate 2 is the fulfillment of the
inus condition. In other words, the psychological process is in itself an “…insufficient but non-
redundant part of an unnecessary but sufficient condition” (p. 13). Gifts are therefore
necessary but in themselves do not guarantee the AE. Ziegler and Heller pointed out that the
gifts are one component within the CS, and all components must be present if there is a
chance of achieving AE. Postulate 3 is that gifts are personal (i.e. psychological processes) and
not environmental variables. To use the example provided in Ziegler and Heller, intelligence
does not qualify as a gift because a direct and causal relationship between it and AE is not
easily determined. At best, the relationship between intelligence and AE can only be
described as correlational. Postulate 4 is that the gift must hold some significance in the
explanation of the AE, and that a probabilistic rather than deterministic relationship between
the gift and AE is made explicit.
To conclude, the critical state view of giftedness (CSG) clearly defines the nature of giftedness
and how it might be researched. Although Ziegler and Heller (2000) focused on the
psychological processes within the CS, it is clear that any understanding of the development
of AE must also be directed toward the interactions between the various components. The
CSG suggests that the research focus should centre on the psychological processes, termed
gifts, which contribute directly to the achievement of excellence. When applying the CSG, a
number of guiding questions can be formulated (Ziegler & Heller), including:
• What is the domain of excellence and in how many ways can it be defined?
• What are the components of the critical state that lead to AE?
• What are the psychological processes which contribute directly and non-trivially to the
AE?
• What interactions between the psychological, personality and environmental conditions
are likely to contribute to AE?
• Under which conditions is the psychological relationship between CS and AE valid?
Of these questions, the first question is perhaps the easiest to answer. Excellence within a
domain is defined by persons working within that domain and, hence, varies across cultures
and time. The psychological processes that contribute to AE will also be different from
domain to domain, as will the interactions between the processes, the personality and the
environment. These interactions are not easy to describe.
The Actiotope Model of Giftedness

The Actiotope Model of Giftedness (AMG) is a systems approach to the development of
achievement excellence (AE) (Ziegler, 2005; Ziegler & Stöger, 2004). It describes the complex
interactions between a person’s action repertoire, their subjective action space and the
environment. In the AMG, the focus is on the development of behaviors that lead to AE.
The development of action repertoires during each stage depends primarily on the
individual setting and realising new goals. Individuals are free to choose goals and to
determine the best possible pathway towards these goals. In determining the pathways,
individuals make subjective assessments of the relevancy, value and usefulness of their
current action repertoires. These subjective assessments, termed subjective action space, are
psychological structures that represent action opportunities and they may either over or
under-estimate the true nature of the action repertoires. These behaviors are both affected by
and, in turn, affect the environment and so undergo a series of progressive adaptations – but
only as long as the person is pursuing excellence.
Not surprisingly, the terminology in the AMG is consistent with that of the CSG, including the
use of the terms talent, gifted and expert. In the CSG, a gift is a psychological process that
satisfies the four postulates. In the AMG, a gift is also a psychological process and its nature is
dependent on the observer and context. Whether or not actions satisfy the four postulates can
only be answered through research.
The components of the Actiotope, including the psychological components, actions, goals and
personality variables (such as motivation and self-efficacy) are described as a dynamic
system in accordance with the requirements of the critical state (CS) within the CSG. Since
these components, by definition, directly interact with the environment in a purposeful way
toward the AE, the Actiotope has a clear developmental trajectory.
The development of an Actiotope leading to mathematical excellence, for example, proceeds

through a number of phases:
• Phase 1 – Pre-natal and early childhood development. An individual may show

exceptional learning or precocious achievement and their actions are termed talented.
• Phase 2 – Attainment of a critical state. An individual has attained all of the necessary
knowledge and behaviours (actions) necessary for excellence. The actions of an
individual at this stage are termed gifted.
• Phase 3 – Attainment of excellence. An individual demonstrates outstanding performance
in a domain. Their outstanding actions are termed excellence.
Of these three, Phases 1 and 2 are more likely to correspond with the individual’s
development through their formal education, although it is clear that the domain(s) in which
future AE occurs is likely to be unknown and that the domain may or may not correspond with
the immediate aims of education. Phases 1 and 2 are, however, critical periods in the
development of action repertoires necessary for AE.
The probabilistic relationship between an individual’s Actiotope and AE is not, however,
clearly articulated in Ziegler (2005). In this regard, modeling the development of the
Actiotope from talent through to expert stage using Structural Equation Modeling (SEM) may
at first glance offer one way to more precisely describe this relationship (McCoach, 2003).
There are two related problems, however, that limit the usefulness of SEM to model the
development of AE. First, SEM requires a sample size of at least 100 if the modeling is to
produce reliable results. This means that we need to identify at least 100 individuals with a
trajectory toward the same AE. If these individuals are at the expert stage then this
identification is made much easier. If the individuals are at the talent or gifted stage, then it is
almost impossible to identify these individuals since they have not, by definition, reached AE.
One possible solution would be to identify the action repertoires of a number of experts
within a domain, and then use these as markers for actions within individuals at the talent and
gifted stage. There is the complication that some of these individuals will exercise their free
will and choose different domains to be an expert. The situation is in fact much more complex
than this because the AMG proposes that different Actiotopes may develop the same AE.
The second problem is that the AMG focuses primarily on the individual rather than a
population (Ziegler, 2005; Ziegler & Stöger, 2004). Hence, the AMG is a model of individual
differences, albeit at a systems level. The development of an action repertoire, together with
the subjective action space, goals and environment do not guarantee AE. An individual’s
Actiotope, therefore, is an unnecessary but sufficient condition, with the action itself being
both insufficient and non-redundant. In other words, the action within a successful Actiotope
does not in itself guarantee AE since the other components are also necessary, even though
AE without the action is impossible. It is this variability that renders the SEM a conceptually
inappropriate tool.
Gardner’s theory of multiple intelligences (Gardner, 1993) is often criticized for its lack of
heuristic value (Anderson, 1992; Klein, 1997; Morgan, 1996). Anderson’s criticism is based on
the observation that multiple intelligences theory is derived using evidence from a number of
different conceptual levels, including culture, environment, personality, cognition and
neurobiology. Similarly, the AMG is dependent on a very broad conceptual framework,
including environment (for both defining actions and domains), meta-cognitive regulation,
self-evaluation, psychological structures, skills and IQ, rendering it susceptible to the same
criticism.
Despite the difficulties in the heuristic potential of the AMG, it is a unique approach to
understanding the nature of giftedness. The theory is pleasing because of its focus on the
development of an individual’s action repertoire rather than unitary (or pluralistic) views of
intelligence. The AMG takes a holistic perspective of giftedness, focusing on the development
of these action repertoires within a changing and interactive environment. Although it is
possible to identify the action repertoires and subjective action space of an individual at any
given time (Ziegler & Stöger, 2004), modeling the interactions of the many different
components of the Actiotope present greater research challenges.
The Actiotope Model of Giftedness and achievement excellence in

mathematics
In discussing mathematical giftedness for the International Handbook on Giftedness
(Shavinina, in press), Phillipson and Callingham (in press) used the Actiotope Model of
Giftedness (AMG) as their conceptual framework. Drawing on research that focused on the
development of mathematical giftedness, Phillipson and Callingham drew on wide ranging
research that reflected aspects of self (number sense1, intelligence, motivational states, beliefs
and attitudinal states, gender, and self-efficacy and self-esteem, for example), the
development of action repertoires (memory, computation and logical reasoning, for example)
and the environment. Phillipson and Callingham also discussed the domain problem,
concluding that there are at least three broad ways to describe mathematical excellence,
including societal perspectives, educator’s perspectives, and the perspectives of professional
mathematicians.
Phillipson and Callingham (in press) noted that the AMG provided a useful conceptual
framework that helped bring unity to their discussion of mathematical giftedness. At a
fundamental level, however, the AMG emphasized the developmental nature of AE in
mathematics, showing how it may be possible to describe the development of complex
mathematical thinking from the small number of innate numerical abilities such as numerosity
and additive expectations2. Furthermore, Phillipson and Callingham proposed that future
research could meaningfully focus on computer modeling of the complex interactions
between an individual and their environment leading to AE using cognitive architectures such
as SOAR and CLARION (Sun, 2006a, 2006b).
In agreement with the generalizations from almost all conceptions of giftedness to date, the
attainment of mathematical AE in the AMG is dependent on both the social processes
interacting with the individual and the characteristics of the individual. In other words, AE is a
social phenomenon since its attainment depends on the recognition by significant others.
Reaching AE also depends on the characteristics of the individual. In the parlance of cognitive
modeling, it is the interaction of a single agent within a broader multi-agent system.
The next section describes how CLARION might be used as a tool to model the development
of mathematical AE within the framework of the AMG. It begins by describing the essential
features of CLARION, including the interactions between the various components of the
architecture. Next, the various levels of the AMG are mapped onto the components of
CLARION. Finally, some of the challenges in using CLARION to model the development of
mathematical excellence are discussed.
The CLARION cognitive architecture and the Actiotope Model of

Giftedness
Modern computational models of cognition recognise the important role of the social
environment. In exploring the contributions that cognitive science can make to the social
sciences, Sun (2006a) drew a link between the ability of modern cognitive architectures such
as ACT-R, SOAR and CLARION to model complex cognitive processes, and the need of social
scientists to explain complex sociocultural phenomenon. Describing “computer simulations
of social phenomenon” (p. 6) as the third way of doing science3, Sun believed that these
simulations can provide data to support both inductive and deductive approaches to
understanding the processes involved in the phenomenon. According to Sun, however, the
key to modeling complex sociocultural phenomenon is to
incorporate realistic constraints, capabilities, and tendencies of individual agents in terms of their cognitive
processes (and also in terms of their physical embodiment) in their interaction with the environments (both
physical and social) (p. 7).
To date, the simulation of complex cognitive processes using cognitive architectures include
the dynamics of youth subcultures (Holme & Grönlund, 2005) and new religious movements
(Upal, 2005). The general issues surrounding computer simulations of complex social
phenomenon have also been discussed (Sawyer, 2003; Sun, 2006a, 2001; Sun & Naveh, 2007;
Wilensky & Rand, 2007)
The AMG describes a number of broad groupings related to the development of achievement
excellence, broadly grouped into aspects of self (physiological and psychological state) and
the social environment. Possible sources of individual differences into aspects of self and the
social environment are easy to find - we merely need to consult an introductory textbook on
educational psychology. For example, individual differences in the physiological state include
pre-natal environmental damage (i.e. alcohol, drug, nicotine, other), nutrition, gender, genetic
dysfunctions and/or damage, sensory integrity and sensory-motor deficiencies. Individual
differences in the psychological state include existing schemas. In the case of mathematical
thinking, the earliest schemas are likely to be those related to number sense.
Other types of individual differences in the psychological state include goal orientations,
motivations, meta-cognitive skills, self-esteem, self-efficacy, cognitive skills and behaviours,
speed of information processing, learning style, and working memory. In the social
environment, examples include significant other persons such as peers, teachers and parents,
and domain specific mentors.
This list, of course, is not exhaustive. When modelling these individual differences using
cognitive architectures such as CLARION, it is important to recognise the dimensions of
variability and onset. Furthermore, the means to “translate” the physiological and
psychological dimensions into syntax that is recognisable by the architecture is required. For
CLARION, much of this syntax is already available.
In devising a computational model for the AMG, the cognitive architecture must be able to
incorporate the interactions between components within and across different conceptual
levels. Furthermore, the attainment of excellence by an individual in a domain reflects both
the individual’s demonstration of an appropriate suite of actions (action repertoires) and the
production of something of value, as well as recognition by a significant group of other
persons.
The CLARION4 cognitive architecture (Sun, 2006b) was devised in the late 1990s to model
the cognitive processes of individual cognitive agents within a social context. According to
Sun, CLARION is a modular system “intended to capture the essential cognitive processes
within an individual agent” required for learning (p. 82). To date, CLARION has been used to
simulate a variety of cognitive tasks, including the collective processes of academic
publication (Naveh & Sun, 2006), food distribution and enforcement of law in a tribal society
(Sun & Naveh, 2007), and organizational decision making (Sun & Naveh, 2004).
The following description of the overall structure of architecture is derived from Sun (Sun,
2001, 2003, 2006b, 2007; Sun & Naveh, 2007; Sun & Zhang, 2004, 2006; Sun, Zhang, & Mathews,
2006). The essential features of CLARION in relation to the present work include:
1. The relationship between cognition and the environment;
2. The social orientations of both needs and motivations;
3. Two separate dichotomies, including the interactions of declarative processes and
procedural learning processes, and the interactions of explicit (rule based) and implicit
(trial and error) processes;
4. Its focus on both top-down approaches to skill development, where, for example,
procedualization of explicit knowledge leads to skilled performance (implicit
knowledge), and bottom-up approaches, where explicit knowledge can arise from
implicit skills or knowledge;
5. The development of implicit knowledge is gradual and incremental, and as a result, the
development of explicit knowledge may often be graduate and incremental as well;
6. The development of domain specific (abstract) knowledge is possible;
7. Explicit knowledge of needs/desires/motivations can be acquired through bottom-up
learning;
8. The inclusion of innate biases and behavioural propensities (individual differences);
9. The continual interactions of the subsystems;
10. The opportunities of social interactions, for example, through the reflective possibilities
of meta-cognition; and
11. Agents can cooperate with others often because agents can understand the
motivational structures of other agents.
Action-centred subsystem (ACS) Non-action-centred subsystem (NACS)
Top level
Action centred explicit Non-action centred

representation explicit representation
Output
Action centred implicit Non-action centred

representation implicit representation
Input
Reinforcement
Goals
Goal setting
Drives
Filtering goal setting
regulation
Bottom
level
Motivational subsystem (MS) Meta-cognitive subsystem (MCS)
Figure 2. The CLARION architecture. Adapted from Sun (2006, p. 80) and reproduced with permission.
The four subsystems of CLARION are shown in Figure 2 with each having two levels of
knowledge representation, namely explicit (top-level) knowledge and implicit (bottom-level)
knowledge, where:
1. Implicit representational processes are meant to represent the inaccessible nature of
implicit knowledge, including implicit procedural knowledge; and
2. Explicit representational processes represent accessible knowledge.
In CLARION, learning occurs through both top-down (implicit knowledge from explicit
knowledge) and bottom-up processes (explicit knowledge from implicit knowledge).
Implicit knowledge is represented by sub-symbolic, distributed representations, and
captured by backpropagation neural networks. A multi-layered neural network is used to
implement a mapping function, where the parameters of the mapping function are adjusted to
change the input/output mappings through trial-and-error interactions.
Explicit knowledge can be learned in many ways. One-shot learning, based on hypothesis
testing, is preferred when interacting with the world5. Explicit knowledge can be derived
from implicit knowledge (bottom-up), and it can be assimilated into implicit knowledge (top-
down).
Action-centred subsystems (ACS) represent knowledge that is important in action decision
making, where actions change the world in some way. The role of the ACS is to control actions
(physical movements and internal mental operations). The ACS receives information from the
NACS. When comparing the changed state with the original state, the agent learns.
In the ACS, the bottom level is termed the Implicit Decision Networks (IDNs). The top level is
referred to as the Action Rule Store (ARS). The overall algorithm for deciding on an action is:
1. Observe current state x.
2. In the bottom level (the IDNs), compute the Q-values6 of x associated with all possible
actions ai’s: Q(x, a1), Q(x, a2), ... Q(x, an). (The set of all possible actions along with the
process of deciding on an action in the bottom level constitutes the IDNs.)
3. At the top level, find all possible actions (b1, b2, … bm ) (the ARS) based on the rules in
place.
4. Consider values of ai and bj, and choose an action b.
5. Perform b, and observe next state y, and (possibly) reinforcement r.
6. Update bottom level according to Q-learning-backpropagation.
7. Update top level using Rule-Extraction-Revision (RER).
8. Go to Step 1.
(Sun, 2006b, pp. 84-85)
One example is Q-learning, where an evaluation of the “quality” of an action a is calculated
for a given state x as described by a sensory input. One way to choose an action is to use
maximum Q-values, i.e. choose a if Q(x, a) = maxi Q(x, i). To ensure adequate exploration of
potential actions, a stochastic decision process7 based on the Boltzmann distribution8 can be
used.
Bottom level learning allows for both input and output of information. Input to the bottom
level consists of three sets of information:
1. Sensory input. The sensory input has a number of dimensions, each with a number of
possible values.
2. Working memory items. The working memory items have a number of dimensions, each
with a number of possible values.
3. The selected item of the goal structure. The goals have a number of dimensions, each
with a number of possible values.
Thus, the input state x is represented by a set of dimension-value pairs: (d1, v1), (d2, v2) … (dn,
vn).
Output of the bottom level consists of action choices, working memory actions (for temporary
storage of information or removal of information), goal actions, and external actions.
Note that the combination of Q-learning and backpropagation enables the development of
implicit knowledge based solely on the agent exploring the world and does not require an
external teacher or a priori domain-specific knowledge.
Top level learning allows for the development of explicit knowledge through rules and
chunks. The condition of rules consist of three sets of information, namely
1. Sensory input;
2. Working memory; and
3. Current goal.
The output of a rule is an action choice (working memory actions, goal actions, and external
actions). The condition of a rule and the conclusion of a rule constitute a distinct entity known
as a chunk.
Bottom-up learning is captured using Rule-Extraction-Refinement (RER) algorithm. Here, the
rules at the top level are learned by using information from the bottom level.
• If an action of the bottom level is successful, then an explicit rule is extracted and later
refined.
• If the outcome is not successful, then the condition of the rule is made more specific.
Top-down learning may happen when the agent relies on explicit rules at the top level for
making decisions. As more knowledge is acquired by the bottom level through observing
actions directed by explicit rules, the agent becomes less reliant on the top level and more
reliant on the bottom level.
In addition, a set of response time equations specifies the response times of different
components of the ACS, as well as overall response times. Response time equations are based
on “base-level activations” (Anderson & Lebiere, 1998).
Non-action centred subsystem (NACS) represents knowledge that is static, declarative and
generic: That is, it stores general knowledge about the world. The role of the NACS is to
maintain implicit and explicit general knowledge. Information is stored as associative rules
and associative memory. The NACS is under the control of the ACS.
At the top level knowledge is stored as associative rules. At this level, general knowledge
stores (GKS) encode explicit non-action-centred knowledge. At the bottom level, knowledge
is stored as associative memory. Associative memory networks (AMNs) encoded non-action-
centred knowledge by mapping input with output, using, for example, backpropagation
learning algorithms.
As well as associative memory and associative rules, similarity based reasoning may be
employed, whereby a known chunk can be compared with another chunk. If there is sufficient
similarity then the latter chunk can be inferred.
In order to allow for complex patterns of reasoning to emerge, thereby capturing the essential
features of human thinking, similarity and rule based reasoning can be mixed. Furthermore, a
set of response time equations specifies the response times of different components of the
NACS, as well as overall response times. Response time equations are based on “base-level
activations”.
A key point in the AMG is the development of action repertoires and subjective assessments
of their effectiveness by the individual. According to Ziegler (2005), an action within this
repertoire includes schemas that reflect knowledge and rules, both implicit and explicit.
Schemas related to mathematical thinking change and adapt in response to learning. Clearly,
there is a match between the requirements of the AMG and the capabilities of CLARION.
Motivational subsystems (MS) provide the underlying motivations for

perception/action/cognition, including the impetus (reasons) for actions and feedback (are
the outcomes un/satisfactory?). The MS “supervises” (Sun, 2006b, p. 90) processes in the ACS
and NACS (Figure 3). The MS allows for:
• Proportional activation, where activation of drives should be proportional to

corresponding offsets, or deficits, in related aspects such as food or water.
• Opportunism, where an agent’s preferences are taken into account (for example, the
agent may prefer to satisfy x instead of y).
• Contiguity of actions, where current actions sequences are preferred over new actions.
• Persistence, where actions persist beyond a minimal level of satisfaction.
• Interruption when necessary, where current actions are interrupted when a new and more
urgent need arises.
• Combination of preferences, where different drives can be combined to form a higher
overall preference.
Goal
goal to
actions goal stimulus
ACS and
from ACS
MCS
and MCS
Low-level drives
Sensory High-level drives Drive

input strengths
Secondary drives
to MCS
Figure 3. The structure of the motivational subsystems (MS). Adapted from Sun (2006, p. 91) and
reproduced with permission.
A dual system of motivational representations is used, where the explicit goals (“find water”)
of an agent are based on internal drives (“being thirsty”). CLARION refers to primary drives
as those that are essentially “hard-wired” to begin with. They include, for example:
• Get-water Strength is proportional to .95*(water-deficit, water-deficit*water-stimulus). The
actual strength depends on the water-deficit felt by the agent and water-stimulus
perceived by it.
• Get-food Strength is proportional to .95*(food-deficit, food-deficit*food-stimulus). Again,
the actual strength depends on the food-deficit felt by the agent and food-stimulus
perceived by it.
• Avoid-danger Strength is proportional to .98*(danger-stimulus*danger-certainty). It is
proportional to danger signal (its distance, severity) and certainty. Danger signal is
captured by danger-stimulus, and danger certainty captured by danger-certainty.
(Sun, 2006b, p. 92)
These drives implemented as hard-wired instincts using backpropagation neural networks. In
addition, higher level drives include belongingness, esteem, self-actualization, and more can
also be hard-wired. Secondary drives can also be implemented. They are derived mostly
through the satisfaction of primary drives as well as being gradually acquired or externally
set (Sun, 2006b, p. 92).
Meta-cognitive subsystems (MCS) refer to an individual’s knowledge about their cognitive

processes and outcomes. It also refers to the active monitoring and regulation of these
processes, usually toward the fulfilment of a goal. The MCS in CLARION operationalizes these
conceptions (Figure 4). The MCS is linked with the MS to monitor, control and regulate
cognitive processes for the purpose of enhancing their performance. Control and regulation
can be achieved through:
• The setting of goals in the ACS;
• Interrupting and changing ongoing processes in the ACS;
• Setting essential parameters in the ACS and NACS; as well as others.
evaluation reinforcement
state
goal setting
goal
goal action
drives
level selection
reasoning selection
filtering,
learning selection selection
and
Monitoring input selection
regulation
buffer
output selection
goal change
parameter setting
Figure 4. The structure of the meta-cognitive subsystem (MCS). Adapted from Sun (2006, p. 94) and
reproduced with permission.
There are many types of metacognitive processes in the MCS, including:

• Behavioural aiming (setting of reinforcement functions, setting of goals),
• Information filtering (focussing on input dimensions in the ACS and NACS),
• Information acquisition (selection of learning methods in the ACS and NACS),
• Information utilisation (selection of reasoning methods in the ACS and NACS),
• Outcome selection (selection of output dimensions in the ACS and NACS),
• Cognitive mode selection (selection of explicit processing, implicit processing, or a
combination thereof in the ACS), and
• Setting parameters of the ACS and NACS (including the IDNs, ARS, AMNs and GKS).
(Sun, 2006b, p. 93)
The MCS consists of a number of modules in two levels: The bottom level consists of goal
setting network, reinforcement function network, input selection network, output selection
network, parameter setting network and so on, whereas the top level consists of explicit rules
(if they exist) and is similarly divided.
The AMG describes the importance of motivational and metacognitive states, and goal
orientations in the development of action repertoires, including those in the domain of
mathematical thinking. This motivational and metacognitive states change in response to the
needs of the individual and the pressures of the environment. Success, for example, is highly
motivating to the individual. For some social groups, however, motivation can be externally
imposed and just as successful. Goal orientations are also “programmed” within CLARION.
Once again, there is a clear match between the requirements of the AMG and the capabilities
of CLARION.
CLARION parameters describe the constraints of the learning processes as being of two
sets. The fundamental properties of Set 1 include:
• Learning rate of neural network;
• Reliance on top vs. bottom level learning (expressed as probability of choosing each
level);
• Temperature (degree of randomness); as well as others.
In Set 2, the parameters concerning rule-extraction learning include:
• RER positivity threshold (which must be exceeded to count a step as positive),
• RER generalisation threshold (which must be exceeded for a rule to be generalised),
• RER specialisation threshold, as well as other performance metrics.
The AMG describes aspects of the psychological state as being important in the
development of Actiotopes. Manipulating these parameters offers the researcher
opportunities to change learning parameters related to speed and ways of learning, matching
individual differences in the aspects of the psychological self in the AMG with CLARION.
Antecedents of action repertoires

Motivational states, metacognitive processes and starting parameters are relatively simple to
set in CLARION. A more important and perhaps more challenging consideration in using
CLARION to model the development of mathematical AE is to show that CLARION can learn
complex mathematical rules from the small number of innate mathematical abilities,
collectively known as number sense. Several types of number sense are evident in human
species, including numerosity, subitizing, additivity and subtractivity and magnitude
representation. Hence, the first step is to represent each type of schema at the top-level
(explicit level) in both the ACS and NACS as fixed rules (FRs). In CLARION, FRs take the
following syntax:
current-state-condition à action
Chunk ID: (d1, v1)(d2,v2)… à action
(Sun, 2003, 2006b)
To take numerosity as an example, the current state condition for a one-to-one

correspondence of objects would involve identifying them as “same” with a two-to-one
correspondence being identified as “different”. For magnitude representation, a mental
number line is thought to underpin many of these arithmetic skills, where magnitude is
represented on a number line. Two types of number line are postulated – a number line
representing approximate solutions to arithmetic problems, and a number line representing
exact solutions to arithmetic problems. Approximate representations of numbers are thought
to be represented as positions on a line with a logarithmic scale. The fixed rule for position of
the number on the line is represented as a normal distribution with mean logx and a fixed
variance (SD2). Hence the numbers “1” and “2” are represented as:
• approx-1: (total, 1)(1, ND), with fixed SD2; and
• approx-2: (total, 2)(2, ND), with fixed SD2 respectively,
where “total” refers to total number of items and ND is a normal distribution.
On the other hand, exact representations of numbers are represented as positions on a line
with a logarithmic scale. Hence the fixed rules for “1” and “2” are represented as
• exact-1: (total, 1)(1, log1).
• exact-2: (total, 2)(2, log2), respectively.
where “total” refers to total number of items.
Representations of the various types of number sense may present the most challenging
aspects of modeling the development of higher forms of mathematical thinking using
CLARION. However, its success may provide important clues as to the possible pathways and
important milestones in Phase 1 of this development. More importantly, the model may
provide clues regarding the influence of other psychological states such as speed of
information processing and working memory, and other environmental factors such as
nutritional state and schooling on this development.
The way forward

The value of the AMG as a model of giftedness is that it provides a unifying basis for
explaining how aspects of the psychological self, the physical self and the social environment
work together in the attainment of excellence. In the AMG, giftedness is a process by which an
individual acquires a sufficient number of actions that enables them to attain excellence
within a particular domain. There are three phases in this process, defined in terms of stage in
the development of these action repertoires. In Phase 1 an individual may have developed an
action repertoire that is relatively more sophisticated in comparison with their peers. These
individuals are termed talented because of the nature of this repertoire. In Phase 2, an
individual may have attained all the knowledge and behaviors that are necessary for
excellence. The actions of these individuals are gifted. Once these individuals have passed
through Phase 2 and demonstrated outstanding performance in their domain, their actions
are termed excellent.
The process of achieving mathematical excellence begins with the development of talented
mathematical action repertoires from their antecedents in number sense. Modeling this
process within Phase 1 using CLARION should be the first step in testing the AMG. Once the
challenges of creating FRs that reflect the various types of number sense are overcome,
CLARION could be used to find answers to the following questions, including:
• Which of the innate basic arithmetic abilities within number sense are important in the
development of higher levels of mathematical thinking?
• How do higher levels of mathematical thinking develop from the interactions of the
various types of number sense?
• What are the significant learning milestones in the development of higher levels of
mathematical thinking?
• What are the significant aspects of the physiological and psychological self that promote
or inhibit the development of higher levels of mathematical thinking?
• What are the significant variables within the social environment that promote or inhibit
the development of higher levels of mathematical thinking?
• Which combination of self and the social environment lead to talent?
The answers to these questions may be tested in the field using more traditional approaches
to psychological and social science research.
If research into giftedness is to be reinvigorated, new avenues of research need to be
proposed and implemented. If we accept that giftedness is as much a social phenomenon as it
is a reflection of the physiological and psychological state of an individual, then tools such as
CLARION can be employed to simulate the interactions between individuals and their social
environment. Simulations can provide the testable hypotheses that are required for the
advancement of research, not only because they generate data, but because the simulations
themselves depend on a clearly defined meta-theoretical framework.
Acknowledgments
The authors acknowledge a number of helpful discussions with Prof. Dr. Heidrun Stöger at the
University of Regensburg in the development of these ideas. The comments made by the
anonymous reviewers of this article are gratefully appreciated
Notes
1
Number sense refers to a number of innate arithmetic abilities (or schemas) and include
numerosity (Wood & Spelke, 2005) (capacity to make representations of numbers of
objects), use of number words (Lipton & Spelke, 2006), additive and subtractive
expectations (Bisanz, Sherman, Rasmussen, & Ho, 2005) ordinal numerical knowledge
(Brannon, 2005), magnitude (Dehaene, 2003), and arithmetic expectations (Kobayashi,
Hiraki, Mugitani, & Hasegawa, 2004).
2
Recently, Halberda, Mazzocco and Feigenson (2008) established a relationship between
numerosity and mathematical achievement amongst junior high school students.
3
The first (or deductive) approach to conducting social science is to construct mathematical
models of social phenomenon in order to make deductions about changes in variables
within the models. The second (or inductive) approach is to make generalizations from a
large number of observations. According to Sun (2006), the insights from such observations
are mostly qualitative in nature (p. 6).
4
CLARION stands for Connectionist Learning with Adaptive Rule Induction.
5
In this model, an agent explores the world, acquires, representations of the world, and
modifies the representations as needed.
6
Q-learning algorithm used to calculate Q-values. In the algorithm, Q(x, a) estimates the
maximum (discounted) cumulative reinforcement the agent receives from the current state x
on. The updated Q(x, a) is based on:
(x, a) = (r + e(y) – Q(x, a)),
where is a discount factor, e(y) is maxaQ(y, a) and y is the new state resulting from action a.
7
Stochastic decision process describes a framework for modeling optimization problems
involving uncertainty,
1
Q ( x ,a )
α
p (a x ) =
e
1
,
∑e
Q ( x ,a )
α
where controls the degree of randomness (or temperature) of the decision-making

process.
8
The Boltzmann distribution predicts the distribution function for the fractional number of
particles occupying a set of states, each state possessing energy.
References
Anderson, M. (1992). Intelligence and development: A discontinuity designs: Evaluations of gifted and
cognitive theory. Oxford, UK: Blackwell talented programs Journal of Experimental
Publishers. Education, 66(2), 163.
Anderson, J. & Lebiere, C. (1998). The atomic McCoach, D. B. (2003). SEM isn’t just the Schoolwide
components of thought. Mahwah, NJ: Lawrence Enrichment Model anymore: Structural equation
Erlbaum. modeling (SEM) in gifted education. Journal for
Bisanz, J., Sherman, J. L., Rasmussen, C., & Ho, E. the Education of the Gifted, 27, 36-61.
(2005). Development of arithmetic skills and Morgan, H. (1996). An analysis of Gardner's Theory of
knowledge in preschool children. In J. I. D. Multiple Intelligence. Roeper Review, 18, 263-
Campbell (Ed.), Handbook of mathematical 269.
cognition (pp. 143-162). New York, NY: Naveh, I., & Sun, R. (2006). Simulating a simple case of
Psychology Press. organizational decision making. In R. Sun (Ed.),
Brannon, E. M. (2005). What animals know about Cognition and multi-agent interaction:From
numbers. In J. I. D. Campbell (Ed.), Handbook of cognitive modeling to social simulation (pp. 124-
mathematical cognition (pp. 85-107). New York, 150). Cambridge: Cambridge University Press.
NY: Psychology Press. Phillipson, S. N., & Callingham, R. (in press).
Callahan, C. M. (2000). Intelligence and giftedness. In Understanding mathematical giftedness:
R. J. Sternberg (Ed.), Handbook of intelligence Integrating self, action repertoires and the
(pp. 159-175). Cambridge, UK: Cambridge environment. In L. V. Shavinina (Ed.), The
University Press. International Handbook on Giftedness.
Dehaene, S. (2003). The neural basis of the Weber- Amsterdam: Springer Science and Business
Fechner law: A logarithmic mental number line. Media.
TRENDS in Cognitive Science, 7(4), 145-147. Phillipson, S. N., & McCann, M. (Eds.). (2007).
Eyre, D. (2009). Introduction. In D. Eyre (Ed.), Gifted Conceptions of giftedness: Sociocultural
and talented education: Major themes in perspectives. Mahwah, NJ: Lawrence Erlbaum
education (Vol. 1, pp. 1-22). London: Routledge. Associates.
Gardner, H. (1993). Multiple intelligences. New York: Phillipson, S. N., Shi, J., Zhang, G., Tsai, D.-M., Quek, C.
Basic Books. G., Matsumura, N., et al. (in press). Giftedness in
Halberda, J., Mazzocco, M. M. M., & Feigenson, L. East Asia. In L. V. Shavinina (Ed.), The
(2008). Individual differences in non-verbal International Handbook on Giftedness.
number acuity correlate with maths Amsterdam: Springer Science and Business
achievement. Nature, 455(7213), 665-668. Media.
Heller, K. A., & Schofield, N. J. (2000). International Sawyer, K. R. (2003). Artificial societies: Multiagent
trends and topics of research on giftedness and systems and the micro-macro link in
talent. In K. A. Heller, F. J. Mönks, R. J. Sternberg & sociological theory. Sociological Methods
R. F. Subotnik (Eds.), International handbook of Research, 31, 325-363.
giftedness and talent (2 ed., pp. 123-137). Seaton, M., Marsh, H. W., Dumas, F., Huguet, P.,
Oxford, UK: Elsevier Science. Monteil, J.-M., Regner, I., et al. (2008). In search
Holme, P., & Grönlund, A. (2005). Modeling the of the big fish: Investigating the coexistence of
Dynamics of Youth Subcultures [Electronic the big-fish-little-pond effect with the positive
Version]. Journal of Artificial Societies and Social effects of upward comparisons. British Journal of
Simulation, 8 from http://jasss.soc.surrey.ac.uk Social Psychology, 47(1), 73-103.
/8/3/3.html. Shavinina, L. V. (Ed.). (in press). The International
Kaufman, J. C. & Baer, J. (Eds.). (2006). Creativity and Handbook on Giftedness. Amsterdam: Springer
Reason in Cognitive Development. New York: Science and Business Media.
Cambridge University Press. Sternberg, R. J. (Ed.). (2000). Handbook of intelligence.
Kaufman, J. C. & Sternberg, R. J. (Eds.). (2006). The New York: Cambridge University Press.
International Handbook of Creativity. New York: Sternberg, R. J. (2004a). Introduction to definitions
Cambridge University Press. and conceptions of giftedness. In R. J. Sternberg
Klein, P. D. (1997). Multiplying the problems of (Ed.), Definitions and conceptions of giftedness
intelligence by eight: A critique of Gardner's (pp. xxiii-xxvi). Thousand Oaks, CA: Corwin
theory. Canadian Journal of Education, 22, 377- Press.
394. Sternberg, R. J. (Ed.). (2004b). Definitions and
Kobayashi, T., Hiraki, K., Mugitani, R., & Hasegawa, T. conceptions of giftedness. Thousand Oaks, CA:
(2004). Baby arithmetic: One object plus one Corwin Press.
tone. Cognition, 91(2), B23-B34. Sternberg, R. J., & Davidson, J. E. (Eds.). (2005).
Lipton, J. S., & Spelke, E. S. (2006). Preschool children Conceptions of giftedness (2 ed.). New York:
master the logic of number word meanings. Cambridge University Press.
Cognition, 98(3), B57-B66. Sun, R. ( 2001). Cognitive science meets multi-agent
Lubinski, D., & Benbow, C. P. (2006). Study of systems: A prolegomenon. Philosophical
mathematically precocious youth after 35 years. Psychology, 14, 5-28.
Perspectives on Psychological Science, 1(4), 316- Sun, R. (2003). A Tutorial on CLARION 5.0. from
345. http://www.cogsci.rpi.edu/~rsun/sun.tutorial.pdf
Marsh, H. W. (1998). Simulation study of Sun, R. (2006a). Prolegomena to integrating cognitive
nonequivalent group-matching and regression - modelling and social simulation. In R. Sun (Ed.),
Cognition and multi-agent interaction: from Upal, M. A. (2005). Simulating the emergence of new
cognitive modeling to social simulation (pp. 3- religious movements [Electronic Version].
26). Cambridge: Cambridge University Press. Journal of Artificial Societies and Social
Sun, R. (2006b). The CLARION cognitive architecture: Simulation, 8. Retrieved June 12, 2009 from
Extending cognitive modeling to social http://ideas.repec.org/a/jas/jasssj/2004-35-
simulation. In R. Sun (Ed.), Cognition and multi- 2.html.
agent interaction: from cognitive modeling to Wilensky, U., & Rand, W. (2007). Making models
social simulation (pp. 79-99). Cambridge: match: Replicating an agent-based model
Cambridge University Press. [Electronic Version]. Journal of Artificial Societies
Sun, R. (2007). The motivational and metacognitive and Social Simulation, 10 from
control in CLARION. In W. Gray (Ed.), Modeling http://ideas.repec.org/a/jas/jasssj/2007-7-
integrated cognitive systems (pp. 63-75). Oxford 2.html.
University Press: New York. Wood, J. N., & Spelke, E. S. (2005). Infants'
Sun, R., & Naveh, I. (2004). Simulating organizational enumeration of actions: Numerical
decision-making using a cognitively realistic discrimination and its signature limits.
agent model. [Electronic Version]. Journal of Developmental Science, 8(2), 173-181.
Artificial Societies and Social Simulation, 7, from Ziegler, A. (2005). The Actiotope Model of Giftedness.
http://jasss.soc.surrey.ac.uk/7/3/5.html In R. J. Sternberg & J. Davidson (Eds.),
Sun, R., & Naveh, I. (2007). Social institution, cognition, Conceptions of Giftedness (pp. 411-436). New
and survival: a cognitive-social simulation Mind York: Cambridge University Press.
& Society, 6(2), 115-142. Ziegler, A., & Heller, K. A. (2000). Conceptions of
Sun, R., & Zhang, X. (2004). Top-down versus bottom- giftedness from a meta-theoretical perspective.
up learning in cognitive skill acquisition. In K. A. Heller, F. J. Mönks, R. J. Sternberg & R. F.
Cognitive Systems Research, 5, 63-89. Subotnik (Eds.), International handbook of
Sun, R., & Zhang, X. (2006). Accounting for a variety of giftedness and talent (2 ed., pp. 3-21). Oxford,
reasoning data within a cognitive architecture UK: Elsevier Science.
Journal of Experimental & Theoretical Artificial Ziegler, A., & Stöger, H. (2004). Identification based
Intelligence, 18, 169-191. on ENTER within the conceptual frame of the
Sun, R., Zhang, X., & Mathews, R. (2006). Modeling actiotope model of giftedness Psychology
meta-cognition in a cognitive architecture. Science 46, 324-341.
Cognitive Systems Research, 7, 327-338.
The authors
Dr Shane N. Phillipson is an Associate Professor in the Department of
Educational Psychology Counselling and Learning Needs, in the Hong Kong
Institute of Education. His research interests include cultural conceptions of
giftedness and models of achievement.
He has been awarded a number of research grants, resulting in research
publications in many international peer reviewed journals, including High
Ability Studies and Educational Psychology. His books include Phillipson, S. N.
(2007). Learning diversity in the Chinese classroom: Contexts and practice
for students with special needs. Hong Kong: The Hong Kong University Press,
and Phillipson, S. N., & McCann, M. (2007). Conceptions of giftedness: Socio-
cultural perspectives. Marwah, NJ: Lawrence Erlbaum Associates.
Ron Sun is Professor of Cognitive Science at Rensselaer Polytechnic Institute,

and formerly the James C. Dowell Professor of Engineering and Professor of
Computer Science at University of Missouri-Columbia.
His research interest centers around the study of cognition, especially in the
areas of cognitive architectures, human reasoning and learning, cognitive
social simulation, and hybrid connectionist-symbolic models.
He is the founding co-editor-in-chief of the journal Cognitive Systems
Research, and also serves on the editorial boards of many other journals. He
is the general chair and the program chair of CogSci 2006, and the program
chair of IJCNN 2007. He is a member of the Governing Boards of Cognitive
Science Society and International Neural Networks Society.
Talent Development & Excellence Relative age effects and technical skills 45
Vol. 1, No. 1, 2009, 45-56
Relative age, talent identification and youth skill

development: Do relatively younger athletes have
superior technical skills?
Jörg Schorer1,2,*, Joe Baker3, Dirk Büsch2,4, Andreas Wilhelm2,5 and Jan Pabst2,4
Abstract: Relative age effects (RAEs) refer to differences among individuals in age-based
cohorts typically used in sport. These effects usually favour relatively older members of
the cohort and are thought to result from differences in maturation and experience among
athletes of different chronological age. Recently, researchers suggested that relatively
younger participants may not be as disadvantaged as previously thought. In two studies,
we examined whether relatively younger athletes who were able to survive in a system
that advantages their relatively older counterparts would develop superior technical skills.
In study one, participants aged 13-15 years (n=140) drawn from a regional handball talent
selection camp in Germany demonstrated a general relative age effect but no differences
between relatively older and relatively younger athletes in physical body size (i.e.,
height/weight) or technical skills. In study two, similar tests were considered with a larger
sample (n=478) and revealed similar results. Furthermore, there were no differences
between those selected for the national youth team and those not selected. Differences in
RAEs do not seem to be due to technical skills or body size variables. Moreover, the
homogeneity of these results suggests causes of the relative age effect occur early in
development.
Keywords:
birth-date, maturation, expertise, skill acquisition
There are a range of primary and secondary factors influencingg the development of
expertise in sport (Baker & Horton, 2004). One secondary factor believed to affect access to
high quality training and coaching is relative age. Relative age effects (RAEs) refer to
chronological age differences between individuals within annually age-grouped cohorts
(Barnsley, Thompson, & Barnsley, 1985). In most youth sports, annual or biannual age-
groupings, which are thought to create homogenous groups within the competition system of
each sport, unfortunately create a sport structure that perpetuates RAEs. For example,
consider a youth competing in a talent-development program for athletes 10 years of age. If
the youth is born very early in the competition year (i.e., is relatively older), he would be ten
percent older than relatively younger opponents born at the end of the competition year (cf.
Helsen, Starkes, & van Winckel, 2000).
First to report RAEs in sport were Grondin, Deschaies, and Nault (1984), who presented data
for Canadian ice-hockey and volleyball demonstrating an unequal distribution across birth
quartiles. At recreational, competitive, and senior professional levels more players were born
shortly after the cut-off date for their respective sports. Similar findings have been noted in
Canadian ice-hockey within developmental leagues (Barnsley & Thompson, 1988; Barnsley et
al., 1985). In many sports, RAEs are not new phenomena. Wattie, Baker, Cobley, and
Montelpare (2007) found RAEs in Canadian ice-hockey hall of fame athletes going back over
1
Institute for Sport Science, Westfälische Wilhelms-University Münster, Horstmarer Landweg
62b,48149 Münster, Germany
2
Research Group Handball
*
Corresponding author. Email: jschorer@wwu.de
3
School of Kinesiology and Health Science, York University, Toronto, Canada
4
Institute for Applied Training Science, Leipzig, Germany
5
Institute for Sport Science, Christian-Albrechts-University Kiel, Germany

46 J. Schorer et al.
four decades. For professional soccer in Germany, Cobley, Schorer, and Baker (2008)
demonstrated RAEs for players and coaches (but not for referees) since the beginning of the
Bundesliga (Highest Federal league) in the early 1960s. In more recent years, RAEs have been
confirmed internationally for male athletes on many different levels of competition and in a
variety of sports, including Australian Rules football (Abernethy & Farrow, 2005), baseball
(Thompson, Barnsley, & Stebelsky, 1991), basketball (Delorme & Raspaud, 2009), soccer
(Helsen, Starkes, & Van Winckel, 1998, 2000; Helsen, Van Winckel, & Williams, 2005; Jimenez &
Pain, 2008), handball (Schorer, Cobley, Büsch, Bräutigam, & Baker, in press; Schorer, Baker,
Lotz, & Büsch, in press), ice-hockey (Wattie et al., 2007), rugby (Till et al., in press), and
volleyball (Barnsley et al., 1985). Findings for female athletes are less consistent. Vincent and
Glamser (2006) found marginal RAEs at the National level for female soccer players but not at
the state level. Schorer, Cobley, et al. (in press) observed a significant over-representation of
relatively older athletes (i.e., from the first quartile) in young female handball players, but
these effects were smaller than for male athletes. In general, RAEs are confirmed with few
exceptions for a variety of sports for both genders over a range of competition and
development levels (Cobley, Baker, Wattie, & McKenna, 2009).
Our understanding of the mechanisms behind RAEs is much less developed. To date, two
hypotheses have been forwarded to explain relative age differences. Most often cited are
maturational differences (e.g., Barnsley & Thompson, 1988), which assumes that greater
height and mass provide an advantage for relatively older athletes in sports where size is
important (Malina, 1994; Malina, Bouchard, & Bar-Or, 2004; Norikazu, Atsushi, & Toru, 2007).
The relevance of height and weight in youth sport is supported by a study of Brewer, Balsom,
Davis, and Ekblom (1992), who found that elite junior soccer players were above the 95th
percentile of normative data for their population.
A second, related, hypothesis considers the role of selection processes. The above mentioned
maturational differences, which are stronger in the early years of athlete development, lead to
an increased probability of selection by coaches for higher tiers of competition (Schorer,
Cobley, et al., in press; Sherar, Baxter-Jones, Faulkner, & Russell, 2007), which perpetuates a
vicious cycle for the relatively younger. Because they are less likely to make it into the
selection, relatively younger athletes have less access to training with highly qualified
coaches and better medical resources (Helsen et al., 1998). Additionally, they are less likely to
gain match experience at higher levels of competition (considered essential to developing
expertise, Baker & Horton, 2004), thereby increasing the differences between relatively older
and relatively younger athletes. Recently, researchers have discovered that relatively younger
participants may not be as disadvantaged as previous research suggested. One study by
Ashworth and Heyndels (2007) found that relatively younger soccer players had higher
salaries than their relatively older counterparts. Further, Baker and Logan (2007) found that
relatively younger ice-hockey players were more highly sought after during the National
Hockey League draft. In the study by Schorer, Cobley, et al. (in press) relatively younger
athletes had a higher probability of becoming members of the adult national team. More
specifically, early selections (i.e., the first regional and national selections) showed clear RAEs
with significant over representations of relatively older athletes; however, the adult national
team had approximately the same percentage of players born in the first and fourth quartiles
indicating an over representation of relatively younger players relative to the population from
which players were drawn. Although the specific mechanisms of these effects are not known,
Schorer, Cobley, et al. (in press) proposed that systems that perpetuate relative age effects
may turn out to be beneficial for some relatively younger players who develop superior
technical or tactical skills in order to compete successfully against their older, generally
bigger opponents. Ultimately, they hypothesize, relative age effects result in those relatively
younger players who survive the system having a larger repertoire of skills and therefore
demonstrating superior performance. We explore this idea in the present study.
Below we summarize two studies examining these relationships in German handball. Overall,
our objectives were to a) establish relative age effects in two samples of handball players
being considered for national team selection, b) consider differences between relatively
Relative age effects and technical skills 47
older and relatively younger players on variables related to body size and technical skill
development, and c) consider differences between those selected for the National team
versus those not-selected.
Study 1
In this first study, we consider the above relationships with a sample of participants being
considered for national handball team selection in Germany. During this preparation camp
for national selection, five regional D-squads tested their teams (Wohlrab, Landgraf, &
Feldmann, 1998). By the end of the week, an All-Star-team was selected by the coaches based
on their impressions over the week. Based on the preliminary research done to date (Baker &
Logan, 2007; Schorer, Cobley, et al., in press), we hypothesized that the sample would show an
overall relative age effect and that the relatively youngest players would have higher scores in
technical skills while relatively older ones would have larger body sizes, although these
hypotheses were largely exploratory.
Methods
Athlete data were acquired during the “Südcamp 2008” of the Southern German Handball
Federation in Tailfingen. Participants were between 13-15 years of age and most of them
passed two selection levels (district and region/state). For this study, 69 female and 71 male
athletes provided their birth-date, height (in cm), and weight (in kg). Their height ranged
from 1.54-1.81m for the female and from 1.66-1.96m for the male players. Their body mass
was between 46-73kg for females and 52-88kg for males. To test for RAEs, birth-months of the
players were re-coded to reflect the athlete’s birth quartile (Q). Because the handball annual
age-grouping starts with the 1st of January, quartiles were calculated accordingly: Q1: January-
March, Q2: April-June, Q3: July-September; and Q4: October-December. As with most of the
previous research, comparisons were drawn on the assumptions of an equal distribution (see
Cobley et al., 2008). For Germany, a recent study by Schorer, Baker, et al. (in press)
demonstrated equal distributions between quartiles. Chi-square analyses were conducted to
test for differences in quartile distributions among the overall sample as well as between
players selected for the All-Star-team and those not selected.
In addition to these tests we conducted a technical talent assessment where respondents
were required to throw ten balls as precisely and as quickly as possible alternating between
the upper two corners of the goal. This task was chosen by the national coaches because it
represents one of the main skills athletes at this age need to have acquired since it is the basis
for most throwing skills in handball. The players had to pick up the ball from a box positioned
twelve meters away from the goal in the middle of the field and then throw the balls after
three approaching steps from approximately nine metres. The duration of these ten throws
was measured by an electronic watch. Due to time limitations during the camp only 43 male
and 48 female players were able to perform the test. During these ten throws two
independent raters evaluated the over arm throw (German: Schlagwurf) in four different
categories (run up, shooting position, throwing movement, and final position). Scores
demonstrated a high degree of inter-rater-reliability (ρ(overall)=.97) and the ratings for both
judges were amalgamated into a general technical rating (Pabst, Büsch, Wilhelm, & Schorer,
2009). Additionally, the speed of throws was measured by Speed-Trac (Sport Thieme,
Germany) and the percentage of correct hits to the goal corners provided a measure of throw
accuracy. This task resulted in four dependent variables (1) test duration, (2) expert ratings,
(3) throw speed, and (4) throw accuracy.
Instead of using a descriptive statistical approach, a more exploratory data analysis was
conducted for physiognomic data and technical skill data, because cell sizes were small and
varied among the quartiles (cf. Sedlmeier, 1996). Figures are presented as boxplots, based on
a rank scale, which are less influenced by outliers. The line in the middle of the box is the
median. The box shows 25% and 75% quartiles and the lines above and below show the
highest and lowest values that are not outliers. Circles (1.5 to 3.0 standard deviations) and
asterisks (more than 3.0 standard deviations) above or below these lines indicate outliers. The
width of the boxes allows a relative comparison of cell sizes per group (cf. Benjamini, 1988;
Sedlmeier, 1996).
For the inferential statistical analysis, a conservative Exact Test by Kruskal-Wallis for
independent samples was calculated, which works well with small samples and rank scales.
For the Monte Carlo sampling the default values corresponding to a sample size of 10,000 and
confidence levels of 99% were used. Due to the small sample size, the exact or asymptotic p-
value seemed inappropriate; therefore, Monte Carlo p-values are reported. When applicable,
post-hoc comparisons of independent samples were calculated using Mann-Whitney-U-Tests
(alpha criterion set on .05 and for technical skills alpha was Bonferroni adjusted). For
calculating effect sizes and power, G*Power 3.0.10 was used (Faul, Erdfelder, Lang, & Buchner,
2007).
Results
The results are presented in three sections. First, we provide the analyses of overall relative
age effect. Second, we test birth quartile differences in physiognomic variables (i.e., weight
and height). In the final section, we investigate differences between quartiles in the technical
skill test. For all three sections, we differentiate those selected for the all-star game and those
not selected, to test whether the selection affected the strength of the RAE (as hypothesized
by Musch & Grondin, 2001)
The analysis of the overall sample revealed a significant relative age effect, ²(3, n=140) =
24.29, p<.01, w=.42. Differentiating between selected and not-selected players, significant
differences from the normal distribution were found for selected players, ²(3, n=43) = 6.76,
p=.04, w=.40, and for not-selected players, ²(3, n=97) = 18.59, p<.01, w=.44. Comparing
distributions of not-selected players to selected players, no significant differences were
found, ²(3, n=97) = 3.51, p=.33, w=.19, 1 – =.45 (see Figure 1).
Quartile 1 Quartile 2 Quartile 3 Quartile 4
45
40
Player per quartile (in %)
35
30
25
20
15
10
0
Overall (n = 140) Selected (n = 43) Not selected (n = 97)
Figure 1. Distribution of players’ birth quartiles overall (n=140) as well as differentiated between
selected (n=43) and non-selected (n=97) for study 1.
200 90 70
58
190 80
180 70
Weight (in kg)

Height in cm
170 60
160 50
150 40
Quartile 1 Quartile 2 Quartile 3 Quartile 4 Quartile 1 Quartile 2 Quartile 3 Quartile 4
Birth-quartiles Birth-quartiles
58
35 30
72
62
57
28
30
Ratings by experts (in points)
26
Duration of test (in ms)
24
25
22
20
20 26
18 101
15
100 8
90
6
Accuracy of throws (in n of hits)
Speed of throws (in km/h)
80
70
2
60
50 0
Figure 2. Boxplots for height (in cm/upper left), weight (in kg/upper right), test duration (in
secs/middle left), expert ratings (in points/middle right), throw speed (in km/h/lower left), and throw
accuracy (in n of hits/lower right) differentiated by birth quartile.
As can be seen in Figure 2 (upper row), there were no differences among the birth quartiles
for either height, H(3)=1.20, p=.76, or weight, H(3)=0.55, p=.91. Further, differentiating
between selected and not-selected players revealed no differences between the groups;
height, selected: H(3)=6.31, p=.09, not-selected: H(3)=2.87, p=.41, and weight, selected:
H(3)=2.14, p=.56, not-selected: H(3)=0.59, p=.90. Finally, we considered differences between
players from different quartiles concerning their technical skills. As can be seen in Figure 2
(middle and lower rows), there were no differences among the groups for any of the four
technical variables. Inferential statistics confirmed no significant differences for (1) expert
ratings, H(3)=3.10, p=.38, (2) throw accuracy, H(3)=1.77, p=.64, (3) throw speed, H(3)=1.46,
p=.70, and (4) test duration, H(3)=0.90, p=.83. Similarly, no differences were found between
selected and not-selected players, (1) expert ratings, selected: H(3)=0.95, p=.83, not-selected:
H(3)=4.39, p=.22, (2) throw accuracy, selected: H(3)=0.22, p=.98, not-selected: H(3)=3.35,
p=.35, (3) throw speed, selected: H(3)=1.48, p=.71, not-selected: H(3)=1.51, p=.68, and (4) test
duration, selected: H(3)=2.94, p=.42, not-selected: H(3)=4.46, p=.22.
Discussion
Analyses indicated very little difference among the birth quartiles for the study outcomes
(height, weight, or technical skills). Furthermore, these variables did not differentiate who
was selected versus who was not-selected. While these results are intriguing from a
theoretical point of view, one main methodological limitation in this study needs to be
considered that warrants caution in our discussion of the results. The sample size of this study
was rather small, which resulted in cell sizes of five or less especially for the relatively
younger and selected athletes. This small number is reached even without differentiating
between genders, a known moderator of RAEs (Cobley et al., 2009; Schorer, Cobley, et al., in
press). For the comparisons of birth quartile distributions to normal distributions these
limitations are not as important as for testing the difference hypothesis on physiognomic
factors (Malina, 1994; Malina et al., 2004; Norikazu et al., 2007) and technical skills (Schorer,
Cobley, et al., in press).
Based on the limitations noted above, a replication of this study was needed with a larger
number of participants. While replications are often not seen as an attractive alternative for
researchers, their necessity for the consolidation and improvement of a scientific disciplines
is unquestionable (Amir & Sharon, 1990). We undertake this task in Study Two.
Study 2
The national handball try-outs in Germany provide the opportunity to collect data from up to
480 young athletes. Out of this sample, up to 80 players are chosen for a second selection
camp out of which the German national youth team is selected. This larger sample should
allow for hypothesis testing inferential statistics, at least for the combined data set.
The aims of Study Two were the same as in Study One. We investigated whether a sample of
female and male players considered for the youth national team would show an overall RAE
and whether body seize and technical skill measures would differ between relatively older
and younger athletes or predict who was selected for the National team. In addition to these
objectives from the first study, we examined differences between male and female sub-
groups. Based on the preliminary research done to date (i.e., Baker & Logan, 2007; Schorer,
Cobley, et al., in press) and similar to Study 1, we hypothesized that the sample would show an
overall relative age effect and that the relatively youngest players would have higher scores in
technical skills and the relatively older ones would have larger body seizes. We also
hypothesized that these effects would be smaller in female than in male players.
Methods
Athlete data were acquired during the 2009 try-outs for the female and male German national
youth teams. Participants in these camps were between 13-15 years of age and most of them
passed through two previous selection levels (district and region/state). Their height ranged
from 1.53-1.86m for the female players and from 1.62-1.98m for the males. Their body mass
ranged from 43-85kg and from 55-106kg for the females and males respectively. By the end of
these try-outs, lists of female and male candidates for the national youth teams were
aggregated. For this study, 238 female and 240 male athletes provided their birth-date, height
(in cm), and weight (in kg). To test for RAEs, birth-months of the players were re-coded to
reflect the athlete’s birth quartile (Q). Similar to Study One, quartiles were calculated as Q1:
January-March, Q2: April-June, Q3: July-September; and Q4: October-December and
comparisons were drawn on the assumption of an equal distribution (Cobley et al., 2008;
Schorer, Baker, et al., in press). Chi-square analyses were conducted to test for differences in
quartile distributions among the overall sample as well as between players selected for the
second national selection level and those not selected. Additionally, the technical skill test
described in Study One was performed by 469 out of the 478 try-outs participants. Nine
participants were not able to fulfil this test due to injuries. Due to the amount of time
necessary to evaluate the technique of the throws it was impossible to conduct this rating
during the try-outs. Therefore, only accuracy and mean speed of throwing were utilized as
dependent variables for the technical test.
Results
As in Study One, results are presented in three main sections: overall RAEs, differences
between the birth quartiles for height and weight, and differences for technical skills (i.e.,
throwing speed, throwing accuracy, and test duration). In each section we also consider males
and females separately.
As expected, significant overall RAEs were revealed, ²(3, n=478) = 41.72, p<.01, w=.29. As
can be seen in Figure 3, RAEs were also found in the female, ²(3, n=238) = 20.49, p<.01,
w=.29, and male sub-groups, ²(3, n=240) = 24.23, p<.01, w=.32.
Quartile 1 Quartile 2 Quartile 3 Quartile 4

60
50
Player per quartile (in %)
40
30
20
10
0
Overall (n = 240) Selected (n = 50) Not selected (n = 190) Overall (n = 238) Selected (n = 39) Not selected (n = 198)
Male Female
Figure 3. Distribution of male players’ birth quartiles (left side) overall (n=240) as well as differentiated
between selected (n=50) and non-selected (n=190) and female players birth quartiles (right side)
overall (n=238) as well as differentiated between selected (n=39) and non-selected (n=198) for study
2.
Differentiating between selected and not-selected female players, significant differences

from the equal distribution were found for both selected, ²(3, n=39) = 13.82, p<.01, w=.59,
and not-selected players, ²(3, n=199) = 11.99, p<.01, w=.24. Comparing the distributions of
the not-selected players against the selected players, significant differences were also found,
²(3, n=199) = 65.81, p<.01, w=.41. Considering male athletes, significant differences from the
normal distribution were revealed for selected players, ²(3, n=50) = 18.00, p<.01, w=.60, and
not-selected players, ²(3, n=190) = 12.53, p<.01, w=.26; further, significant differences were
also found between these distributions, ²(3, n=190) = 36.36, p<.01, w=.39.
Because there were few players in the third and fourth quartile for the selected players (as
small as only 3 players), only differences between quartiles in the combined group were
considered for height, weight and technical skills. For the physiognomic variables Kruskal-
Wallis-tests revealed no significant differences between birth quartiles for height, H(3)=1.00,
p=.80, or for weight, H(3)=0.72, p=.87, for males. As can be seen in Table 1, similar results
were found for the females. There were no significant differences between birth quartiles for
height, H(3)=1.32, p=.72, or for weight, H(3)=0.57, p=.90.
For males, no significant differences between groups could be revealed for mean speed of
throws, H(3)=3.72, p=.29, and accuracy of throwing, H(3)=1.55, p=.67. For females, no
significant differences between birth quartiles were revealed for mean speed of throws,
H(3)=2.08, p=.56, or for accuracy of throwing, H(3)=2.79, p=.43.
Table 1. Comparison of means (standard deviations) per birth quartiles concerning height (in m),
weight (in kg), accuracy of throws (n out of ten), and mean speed of throw (in km/h).
Gender Dependent variable Quartile 1 Quartile 2 Quartile 3 Quartile 4
Height 1.82 (0.07) 1.81 (0.07) 1.82 (0.07) 1.81 (0.06)
Weight 75.3 (8.6) 74.5 (7.8) 74.7 (9.9) 74.3 (7.7)

Male
Accuracy of throwing 3.81 (1.96) 3.66 (1.71) 3.47 (1.81) 3.38 (1.54)
Mean speed of throws 63.7 (6.8) 64.8 (7.3) 62.3 (6.4) 64.8 (8.4)
Height 1.68 (0.06) 1.67 (0.07) 1.67 (0.06) 1.67 (0.07)
Weight 61.3 (7.4) 60.7 (7.1) 61.1 (7.7) 60.8 (8.1)

Female
Accuracy of throwing 3.97 (1.78) 3.66 (1.82) 3.38 (1.67) 3.71 (1.66)
Mean speed of throws 54.3 (5.5) 54.3 (5.7) 53.8 (5.5) 52.8 (5.1)
Discussion
As expected (Schorer, Cobley, et al., in press), young handball talents showed a general
relative age effect; however, contrary to our hypotheses, there were no relative age
differences in any of the player-specific outcomes examined in this study. Players across the
birth quartiles were similar in height, weight, and technical skills (i.e., test duration, expert
rating, throw speed and accuracy). Moreover, there were no differences between players
selected for the national youth team and those not selected.
While we were able to overcome one of Study One’s limitations by differentiating between
genders, the small number of selected athletes limited the sophistication of our analyses. As
in Study One, the cell sizes for the third and fourth quarter within the selected male and
female groups became too small to conducted differentiated parametric analyses. An attempt
to use data from the previous selection year was not possible, because the technical test was
altered for the selection of 2009. This analysis may be possible in the future, provided the test
remains constant for the next two to three years. By choosing random comparison samples,
null-hypothesis testing – as suggested by our results – should be possible.
General Discussion
Taken together, results from these studies support previous work (e.g., Cobley et al., 2008;
Schorer, Cobley, et al., in press) suggesting that the primary mechanism driving relative age
effects acts early in the athlete development process and that latter levels of selection, such as
being selected for the national youth team, do not add to the effect. If this result bears out in
further research, it provides important information for reducing or eliminating relative age
effects. Specifically, that the earliest levels of talent identification and selection need to be
targeted and modified to provide a more balanced selection process.
On the other hand, these results do little to explain the preliminary work by Baker and Logan
(2007) as well as by Ashworth and Heyndels (2007) indicating that relatively younger athletes
are more highly valued than their relatively older counterparts. The clear similarities among
the birth quartiles in this study indicates either that relatively younger athletes are superior
on some other outcome not measured in this study (e.g., tactical or decision-making skill) or
that a similar advantage for relatively younger athletes is not present in handball (which
would be contrary to Schorer, Cobley et al., in press). Both explanations might be influenced
by missing moderator variables. Playing position, for instance, has been seen as important for
understanding the RAE in handball (Schorer, Cobley, et al., in press), soccer (Ashworth and
Heyndels, 2007) and ice hockey (Edwards, 1994; Grondin & Koren, 2000). As a result, the
expected difference between relatively younger and relatively older would need to be
considered relative to playing position, which, unfortunately, was not possible with the sample
investigated here.
It is also possible that a very homogenous group of players is being considered as ‘talented’
right from the first regional level of selection (which in handball occurs around 10 years of
age). As the present results indicate, they have similar stature and weight (similar results have
been noted in ice-hockey by Baker, Cobley, Montelpare, Wattie, & Faught, in press; Sherar et
al., 2007). It is possible that compensation for the relatively younger occurs during earlier
selection levels since the developmental differences are typically much larger pre-puberty
than post-puberty. Additionally, it would be helpful to expand the measures of growth and
maturation beyond simple measures of height and weight to more objective measures of
biological maturity. These measures would rule out whether relatively younger athletes are
simply maturing more quickly than their relatively older peers.
Another reason for the homogeneity of the groups in technical skills might be the consistency
in types and amounts of training after the first selection. As Helsen et al. (1998) hypothesized
more qualified training on technical skills might happen in their selection team training
resulting in similar movement patterns for the players, although the ‘homogeneity of training’
hypothesis does not explain the results by Baker and Logan (2007) or Ashworth and Heyndels
(2007). Furthermore, observation of individual players at most levels of play clearly suggests
a degree of individual difference even among players with the same coach (reflecting a given
coach’s ‘signature’). It is possible that the technical skill examined in the present study
(throwing) was too simple and as a result a large degree of homogeneity across the birth
quartiles is not surprising. Future work may consider more sophisticated measures of
technical skills to consider whether these differences are affected by RAEs.
Overall, future studies considering differences between relatively younger and older athletes
should investigate the very first selection level, to investigate why relatively younger athletes
are selected for further talent developmental programmes as well as later levels of selection
to determine why specific selection decisions are being made. Especially helpful would be
longitudinal data showing the development of technical as well as perceptual and tactical
skills in players across the selection process. From a broader perspective, recent studies
(including the data presented here) reinforce the need for a more elaborate theory of RAEs,
which considers the complex interaction among moderator variables and main effect
predictors. A comprehensive understanding of this phenomenon would facilitate behavioural
and policy changes to eliminate this inequality.
Acknowledgements
We would like to thank the German Handball Federation for their cooperation. Additionally,
we would like to express our gratitude to Sebastian Bagats, Christina Janning, Rebecca
Rienhoff, Lennart Fischer and all the students from the seminar “talent detection in handball”
for their helping hand with the data collection. This study was in part funded by the Federal
Institute of Sport Science, Germany, IIA1-070704/09.
References
Abernethy, B., & Farrow, D. (2005). Contextual factors Delorme, N., & Raspaud, M. (2009). The relative age
influencing the development of expertise in effect in young French basketball players: a
Australian athletes. Proceedings of the 11th World study on the whole population. Scandinavian
Congress of Sport Psychology, Sydney, Australia. Journal of Medicine and Science in Sport, 19, 235-
Amir, Y., & Sharon, I. (1990). Replication Research: A 242.
"must" for the scientific advancement of Edwards, S. (1994). Born too late to win? Nature, 370,
psychology. In J. W. Neuliep (Ed.), Handbook of 186.
replication research in the behavioral and social Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A.
sciences (pp. 51-69). Corte Medera, CA: Select. (2007). G*Power 3: A flexible statistical power
Ashworth, J., & Heyndels, B. (2007). Selection bias and analysis program for the social, behavioral, and
peer effects in team sports: The Effect of Age biomedical sciences. Behavior Reseach Methods,
Grouping on Earnings of German Soccer 39, 175-191.
Players. Journal of Sports Economics, 8, 355-377. Grondin, S., Deschaies, P., & Nault, L. P. (1984).
Baker, J., Cobley, S., Montelpare, W. J., Wattie, N., & Trimestres de naissance et rendement scolaire.
Faught, B. (in press). Exploring proposed Apprentissage et Socialisation, 16, 169-174.
mechanisms of the relative age effect in Grondin, S., & Koren, S. (2000). The relative age effect
Canadian minor hockey. International Journal of in professional baseball: A look at the history of
Sport Psychology. major league baseball and at current status in
Baker, J., & Horton, S. (2004). A review of primary and Japan. Avante, 6, 64-74.
secondary influences on sport expertise. High Helsen, W. F., Starkes, J. L., & Van Winckel, J. (1998).
Ability Studies, 15, 211-228. The influence of relative age on success and
Baker, J., & Logan, A. J. (2007). Developmental dropout in male soccer players. American
contexts and sporting success: birth date and Journal of Human Biology, 10, 791-798.
birthplace effects in national hockey league Helsen, W. F., Starkes, J. L., & Van Winckel, J. (2000).
draftees 2000-2005. British Journal of Sports Effect of a change in selection year on success
Medicine, 41, 515-517. in male soccer players. American Journal of
Barnsley, R. H., & Thompson, A. H. (1988). Birthdate Human Biology, 12, 729-735.
and success in minor hockey: the key to the Helsen, W. F., Van Winckel, J., & Williams, A. M. (2005).
NHL. Canadian Journal of Behavioural Science, 20, The relative age effect in youth soccer across
167-176. Europe. Journal of Sports Sciences, 23, 629-636.
Barnsley, R. H., Thompson, A. H., & Barnsley, P. E. Jimenez, I. P., & Pain, M. T. G. (2008). Relative age
(1985). Hockey success and birthdate: The RAE. effect in Spanish association football: Its extent
Canadian Association for Health, Physical and implications for wasted potential. Journal of
Education, and Recreation, 51, 23-28. Sports Sciences, 26, 995-1003.
Benjamini, Y. (1988). Opening the box of a boxplot. Malina, R. M. (1994). Physical growth and biological
The American Statistician, 42, 257-262. maturation of young athletes. Exercise and Sport
Brewer, J., Balsom, P. D., Davis, J. K., & Ekblom, B. Sciences Reviews, 22, 389-434.
(1992). The influence of birth date and physical Malina, R. M., Bouchard, & Bar-Or, O. (2004). Growth,
development on the selection of a male junior maturation, and physical activity Champaign
international soccer squad. Journal of Sport Illinois: Human Kinetics.
Science, 10, 561-562.
Musch, J., & Grondin, S. (2001). Unequal competition
Cobley, S., Baker, J., Wattie, N., & McKenna, J. M. as an impediment to personal development: A
(2009). Annual age-grouping and athlete review of the relative age effect in sport.
development: A meta-analytical review of Developmental Review, 21, 147-167.
relative age effects in sport. Sports Medicine, 39,
Norikazu, H., Atsushi, H., & Toru, F. (2007).
235-256.
Relationship between date of birth, skeletal age,
Cobley, S., Schorer, J., & Baker, J. (2008). Relative age and anthropometric characteristics in
effects in elite German soccer: a historical adolescent elite soccer players. Paper
analysis. Journal of Sports Sciences, 26, 1531- presented at the 12th Annual Congress of the
1538. ECSS 2007.
Pabst, J., Büsch, D., Wilhelm, A., & Schorer, J. (2009). 879-886.
Haben es gute Handballer (selbst) in der Hand? Thompson, A. H., Barnsley, R. H., & Stebelsky, G.
[Can good handball players be identified by (1991). Born to Play Ball - the Relative Age Effect
hand?], Manuscript in preparation. and Major-League Baseball. Sociology of Sport
Schorer, J., Baker, J., Lotz, S., & Büsch, D. (in press). Journal, 8, 146-151.
Influence of early environmental constraints on Till, K., Cobley, S., Wattie, N., O'Hara, J., Cooke, C., &
achievement motivation in talented young Chapman, C. (in press). The Prevalence,
handball players. International Journal of Sport Influential Factors and Mechanisms of Relative
Psychology. Age Effects in UK Rugby League. Scandinavian
Schorer, J., Cobley, S., Büsch, D., Bräutigam, H., & Journal of Medicine and Science in Sport.
Baker, J. (in press). Influences of competition Vincent, J., & Glamser, F. D. (2006). Gender
level, gender, player nationality, career stage differences in the relative age effect among US
and playing position on relative age effects. Olympic Development Program youth soccer
Scandinavian Journal of Medicine and Science in players. Journal of Sports Sciences, 24, 405-413.
Sports.
Wattie, N., Baker, J., Cobley, S., & Montelpare, W. J.
Sedlmeier, P. (1996). Jenseits des Signifikanztest- (2007). Tracking relative age effects over time in
Rituals: Ergänzungen und Alternativen. Methods Canadian NHL players. International Journal of
of Psychological Research Online, 1, 41-63. Sport Psychology.
Sherar, L. B., Baxter-Jones, A. D. G., Faulkner, R. A., & Wohlrab, S., Landgraf, L., & Feldmann, K. (1998).
Russell, K. W. (2007). Do physical maturity and Talentcamp - Erlebnis-Event und Ausbildungs-
birth date predict talent in male youth ice Check. Handballtraining, 20, 27-31.
hockey players? Journal of Sports Sciences, 25,
The authors
Dr. Jörg Schorer is a post-doctoral fellow at the University of Münster in the

Department of Sport Psychology and is currently finishing his habilitation. His
research considers the development and maintenance of expertise across the
life-span from multiple perspectives.
Dr. Joe Baker is an associate professor in the School of Kinesiology and Health
Science at York University, Canada and a visiting researcher at Leeds
Metropolitan University in the UK. His research examines the factors affecting
athlete development across the lifespan. Dr. Baker is the current president of
the Canadian Society for Psychomotor Learning and Sport Psychology.
PD Dr. Dirk Büsch is head of the Department of Game and Combat Sports and
has responsibility for elite youth sports at the Institute for Applied Training
Science University of Leipzig, Germany.
Prof. Dr. Andreas Wilhelm is currently working at the Institute for Sport and
Sport Science at the Christian-Albrechts-University Kiel, Germany. His
research interests include topics like groups, leadership, coaching, and sport
games.
Dr. Jan Pabst is currently employed as the Head of the Research Group
Handball at the Institute for Applied Training Science in Leipzig, Germany.
Talent Development & Excellence Gifted Learners’ Epistemological Beliefs 57
Vol. 1, No. 1, 2009, 57-66
Gifted Learners’ Epistemological Beliefs

Marion Porath1,*, Judy Lupart2, Jennifer Katz1, Constantine Ngara1 and Pamela Richardson1
Abstract: This paper focuses on gifted children's and adolescents' narrative

interpretations of learning using a neo-Piagetian theory of conceptual development
(Case, 1992; Case & Okamoto, 1996) as a framework. Children's narratives develop
from action based reasoning to intentional reasoning that incorporates understanding
of their own and others' mental states. In adolescence, thinking becomes more
interpretive, with a focus on the psychology of individuals involved. Similarly,
understanding of learning develops from a focus on the activities of school (e.g.,
reading, math) to interpretation of the meaningfulness of different learning activities
and the meaning of knowledge itself. Students’ narrative accounts of learning are used
to articulate the development of epistemological beliefs across childhood and
adolescence.
Keywords:
Learning, conceptual development, giftedness, epistemological beliefs
There are significant gaps in our knowledge of how children understand the educational
enterprise, and of how this understanding may impact their adjustment to school and
realization of their talents. In particular, it is imperative that we understand how children
with significantly developmentally advanced ability profiles understand their own
learning. Familiar norms do not apply to these children. Significant adaptations need to be
made to curricula and educational programming to meet their needs (Robinson, Zigler, &
Gallagher, 2000). Listening to children and providing environments and educational
strategies that engage them in understanding and directing their own learning are critical
to the notion of “gift-creation” advocated by Hymer (2009) and Huxtable (2009) and
consonant with the vision of contemporary classrooms where growth, enquiry, and
personally relevant education are the focus (Hymer, Whitehead, & Huxtable, 2009).
Harter (1996) pointed out how important it is to know how children perceive different
aspects of school life, in order that we can take these perceptions into account in
supporting them through the school years. However, education typically disregards
learners' perspectives on knowledge and their understanding of their capacity for
learning (Bruner, 1996). Children are thinkers (Bruner) with competent and legitimate
ways of making sense of the world. Children's understanding of learning is a necessary
starting point for pedagogy. Children's understandings are rarely engaged in this way,
but when they are, the result for children is deeper and more meaningful learning
(Bransford, Brown, & Cocking, 2000; Fosnot, 1996; Griffin & Case, 1996; McKeough, 1992).
Epistemic beliefs, or beliefs about knowledge and knowing, are related to self-regulated
learning and achievement (Muis, 2007) and intellectual performance (Kuhn, Cheney, &
Weinstock (2000).
What do children understand about learning and knowledge? This paper investigates the
epistemological beliefs of gifted learners, concentrating on how they understand the
relationship between themselves as knowers/learners and knowledge (Burr & Hofer,
2002). Following Burr and Hofer, ‘personal epistemology’ is the focus rather than
‘epistemology’ in the philosophical sense.
1
Faculty of Education, The University of British Columbia, 2125 Main Mall, Vancouver, BC,
Canada V6T 1Z4
*
Corresponding author. Email: marion.porath@ubc.ca
2
University of Alberta

58 M. Porath et al.
Research on epistemological beliefs has been conducted primarily with adults and some
hypothesize that epistemological reasoning is not possible before the acquisition of
formal thinking (Hofer & Pintrich, 1997). There may, however, be developmental
precursors to epistemological thought, with abstract conceptions building on first
thoughts about what it means to learn, as suggested by Burr and Hofer (2002), Kitchener
(2002), and Kuhn et al. (2000). Very little research has been done on children’s
epistemological beliefs; the “origins of epistemological awareness” (Burr & Hofer, p. 200)
are not well understood. Studies on children include Burr and Hofer’s study of 3- to 5-year-
olds focused on the relationship of theory of mind and epistemological beliefs and Kuhn
et al.’s studies of 7- and 8-year-olds and 10-year-olds. The former suggested a relationship
between ability to predict and justify what others will think and do and theory of mind
ability characterized by development from inability to do either set of tasks, to the ability
to predict and justify in a relevant way, to the ability to coordinate both abilities
successfully. Burr and Hofer saw their study as only beginning to address a critical
developmental connection and call for more fine-grained, developmental studies that
investigate process. Similarly, they note the need to connect such studies to education,
highlighting that another gap in our understanding is that learners’ beliefs and the school
context have not been linked in any practical way (Ziegler, Stoeger, & Mundi, 2004). Kuhn
et al.’s work, while tangentially related to education in its articulation of development of
conceptions of knowledge from childhood to adulthood, also does not address the
educational implications of the influence of different views of knowledge on the process
of learning.
Mansfield and Clinchy’s (2002) study focused on children aged 10, 13, and 16, finding an
increasing ability to coordinate events with the “inner worlds of individual knowers” (p.
225). While they included a vignette featuring a teacher, the focus was on justifications for
why the child thought the teacher was nice or mean. Their identification of the age-related
increase in complexity and nature of epistemological position is important in
understanding the general nature of epistemological development but needs practical
translation to be meaningful in schools and the education of gifted learners.
A Neo-Piagetian Perspective on the Understanding of Learning
Neo-Piagetian theory is the framework for the work presented in this paper. Neo-
Piagetian theories allow for the mapping of "precise changes in the development of
academic skills and understandings" (Mascolo, Li, Fink, & Fischer, 2002, p. 120). Because
of its articulation of the structure and content of children’s representations of the academic
and social dimensions of school, and the developmental course of these representations,
Case’s (1992; Case & Okamoto, 1996) neo-Piagetian theory provides an appropriate
framework for this study. Case (1992) described children’s central conceptual structures in
a number of domains (mathematical, spatial, social narrative; see Case & Okamoto, 1996).
These structures are “blueprints” of children’s understanding, the internal mental entities
that consist of the relations among a number of concepts. Because the relations are
semantic, articulating the meanings children assign to concepts, structures are conceptual.
Structures are considered central to a domain because they form the basis for
understanding a wide range of related tasks within that domain.
Studies of gifted learners framed within Case’s (1992) neo-Piagetian theory have found
development to be relevant to the level of conceptual understanding they demonstrate
(see, for example, McKeough, Genereux, & Jeary, 2006; Okamoto, Curtis, Jabagchourian, &
Weckbacher, 2006; Porath, 2006). That is, maturation is a factor in the level of conceptual
complexity demonstrated, even among highly able learners, when responses are
considered from the perspective of central conceptual structures. For this reason, work
with children who represented a typical range of abilities was used to inform this study of
gifted learners’ conceptions of learning.
Children’s conceptions of learning. Bickerton (1994) articulated a developmental

progression in how children understand learning, using Case’s (1992) theoretical model.
This research provides a basis for determining the sort of understanding that may be
central to children's conceptions of academic tasks. Bickerton asked elementary school
children what learning means, and what happens when they learn. She found that
children’s understanding of the meaning of learning develops in the following manner.
• Four-year-olds think of learning as behavioural events (e.g., playing, going to school,

building a tower). They also think of learning as involving the presence of a “learning
agent,” such as Mom or Dad.
• Six-year-olds define learning as a relationship between a behavioural event and an
internal state. Internal states include thoughts, feelings, and judgments. For example,
“If I do good work, I get happy.” “Sometimes learning gets me frustrated because I
make so many mistakes” (Bickerton, pp. 6-7).
• Eight-year-olds are able to relate a behavioral event to two internal states when
defining learning. “You get better at it if you try your hardest” (Bickerton, p. 7).
• Ten-year-olds add a personal element to their definitions of learning. “To me, learning
means knowing how to do something without having any problems doing it
(Bickerton, p. 7).
• Twelve-year-olds’ responses are more psychological in nature, recognizing “states of
mind” as important in learning. “Learning is developing a smarter mind.” “Learning
is knowing and understanding things you didn’t know before” (Bickerton, p. 8).
Bickerton (1994) conceptualized children’s understanding of learning as related to
narrative thought, that is, as one of the understandings related to a central social narrative
structure (Case & Okamoto, 1996; McKeough, 1992). Children’s responses clearly reflected
human action and motivations related to the social context and traditions of schooling and
education more broadly defined; they were making meaning of their learning
experiences.
As children enter the world of formal schooling, they face exceptional demands to make
meaning of their experiences, including understanding their own learning and why
others behave the way they do. Bruner (1986) described meaning making as the
simultaneous construction of two landscapes.
One is the landscape of action, where the constituents are the arguments of action: agent, intention, or goal,
situation, instrument…. The other landscape is the landscape of consciousness: what those involved in the
action know, think, and feel, or do not know, think or feel. (p. 12).
In a similar vein, Perner (1991) stated, "Mental states need to be systematically linked to
externally observable events" (p. 101).
Method
This study focused on gifted children's perspectives on learning. The task and analyses
reported are from a larger study focused on gifted children’s understanding of learning
and teaching, their learning identities, and social aspects of education (Porath & Lupart,
2007). Eighty-one students from Grades 1 through 12 (43 boys; 38 girls) identified as
gifted participated in the larger study. Seventeen of the participants were enrolled in a 2-
year program that facilitated early entrance to university (11 in Year 1; 6 in Year 2). All
grades except 8, 9, and 11 were represented in the sample. (Canada’s school systems
comprise twelve grades, preceded by kindergarten.) The study included public and
parochial schools in two large western Canadian cities. The students were identified in
different ways, as is characteristic across school districts in western Canada. Identification
strategies included combinations of teacher nomination, superior academic achievement,
and/or superior intelligence or cognitive ability test scores. The participants attended a
variety of programs for gifted learners, including segregated classes and pullout
programs.
60 M. Porath et al.
Table 1. Development of Thinking about Learning
Level Nature of thought about learning Examples
1 • Action-based You do a lot of new things; “doing stuff”; brain

is remembering all that stuff
2 • Gaining knowledge through a variety of To do … no, not to do but to think about what
modes and learning in different ways something means.
• Understanding/meaning is key Learning can mean different things like my one
• Rudimentary notion of knowledge example is probably you can learn about
• Future-oriented yourself, like what…yourself means to you,
• General idea of brain’s role in learning what is really inside you…or you can learn
knowledge based like… math, and reading,
and drawing, PE, all sorts of learning …to do
with knowledge well actually PE isn’t to do with
knowledge but whatever …and…I guess
learning can also mean learning about…
relationships like sometimes you might have
problems with your friends like you might get
into a fight but you always learn from your
mistakes…so… I think that’s what learning
means.
3 • Acquisition of competence – increasing I’m finding out new things, researching things,
knowledge; learning new things and … basically passing on what others have
• Awareness of interests, weaknesses, learned. It doesn’t always happen to a kid.
boredom, curiosity Learning doesn’t stop when you become an
• Multiple modes of learning adult, so being a scientist … it’s always
• Learning as continuous learning new stuff.
4 • Links you to knowledge traditions Learning to me means to find new things, new
• Transformative understandings of things.
• Awareness of the breadth of learning You absorb information from others and you
(variety of sources, vicarious learning) try to use it in your own life.
5 • Views of learning and knowledge related Changing behaviour through experience and
to personal preferences and goals applying the knowledge.
• Recognition of ideas and different It’s a function of mind.
perspectives on knowledge There is a spark when you are happy about
• Characterization of learning as abstract what you are learning, asking questions, and
• Recognition of neuropsychological factors making your own connections.
in learning (perception, associative It’s interesting to learn new things and open
learning) your point of view.
Putting information together to see if it can be
combined to come up with an idea.
Conceptions of Learning
This task was the first of a series of research tasks completed by the participants. Each
child participated in an individual semi-structured interview, using the questions below as
a framework for discussing their perspectives on learning. Their responses were recorded
and transcribed for analysis.
• What does learning mean?
• What is happening when you are learning?
• Where does learning come from? (Bickerton, 1994)
Coding guidelines for responses to the first two questions were devised based on
inductive analyses of children’s responses and Case’s (1992) theoretical framework. The
scheme also drew on McKeough’s (1992; McKeough & Genereux, 2003) work on
development in the narrative domain, viewing learning, generally and specifically, as an
intentional act focused primarily on knowledge acquisition during middle childhood and
an interpretive epistemological act in adolescence (Table 1). Coding guidelines were the
result of an iterative process in which data were examined in light of Case’s (1992) neo-
Piagetian theory of development; prototypical characteristics of stages in child and
adolescent development defined; and children’s best responses assigned to a level of
development. Coding was done first by four individual coders who were research
assistants on the project; codes were then compared and discussed in pairs; the final
stage involved comparing and discussing codes in the group of four to reach consensus.
The process used was a “constructivist revision of grounded theory” (Henwood &
Pidgeon, 2003, p. 134); the data guided interpretation within the conceptual framework of
developmental theory using the steps of open coding, constant comparison, and
theoretical integration (Henwood & Pidgeon). Analyses were primarily qualitative in
nature. However, following other neo-Piagetian studies, a numerical score was assigned to
levels of development and used in a statistical analysis of developmental trend.
Responses to the question regarding the source of learning were content analyzed,
focusing on whether the source was external, internal, or a combination of the two and the
degree of sophistication evident in the response. This analysis was also informed by
Case’s (1992) theory of development; however, no numerical scores were assigned.
Results
Development of Conceptions of Learning
Participants were grouped by grades; these groupings approximated the age ranges
specified by Case (1992) for each of his hypothesized developmental stages (Grades 1
and 2, n=6; Grades 3 and 4, n=12; Grades 5 and 6, n=24; Grade 7, n=12; Early Entrance 1
and 2 with Grade 10, n=23; and Grade 12, n=4). Mean (SD) learning scores by group are
reported in Table 2. Learning scores were analyzed for age effect using a oneway ANOVA.
A significant between groups effect was found, F(5, 76)=6.73, p=.000. The linear trend was
significant, F(1,80)=26.02, p=.000, with no significant deviation from a linear trend, F(4,
77)=1.90, p=.119. Post hoc comparisons were not done due to the unequal group sizes.
Table 2. Mean (SD) learning level by age group
Group Mean age (SD) in months Learning level
Grades 1 and 2 93.67 (7.66) 1.50 (.45)
Grades 3 and 4 109.42 (8.10) 1.83 (.62)
Grades 5 and 6 124.75 (6.16) 2.40 (.64)
Grade 7 145.17 (3.07) 2.54 (.72)
Early entrance 1 and 2; Grade 10 176.35 (8.49) 3.15 (1.07)
Grade 12 204a 2.38 (.85)
a
Average age is an estimate based on the usual age of Grade 12 students. Age for these students was
not noted during data collection.
62 M. Porath et al.
Children’s responses form a developmental progression generally related to age, as

illustrated in Table 1. The youngest children understand learning as “doing stuff” and
learning new things. Children in the middle grades of elementary school understand
learning as acquisition of knowledge that is important to them in the future and as more
than just academic. Learning makes one conscious of one’s own interests. Children of this
age also have a rudimentary understanding that meaning is important in learning. Without
meaning, learning is boring. In pre-adolescence, there is recognition of the competence
that comes with learning, the link to established knowledge, and the human drive to learn.
Early adolescent participants talked about learning as a transformative experience (“You
think in ways you didn’t before) and their emergent place in a knowledge tradition. Some
of the oldest participants, in particular students from the program designed to support
early entrance to university, demonstrated sophisticated epistemological beliefs –
epistemic inclination (one must be interested and willing), epistemic acquisition (the
availability of knowledge and its perceivability), and attributes of the perception (its
certainty and the uniqueness of the perception) (Ziegler et al., 2004). They began to play
with notions of the relativity of knowledge and unique ideas that emerge when individual
perceptions interact with established knowledge domains, although these ideas were not
yet well formed.
Sources of Learning
Children’s attributions for the sources of learning developed from external sources, often
with little relationship to learning, in young children to a blend of internal and external
sources later in middle childhood. Through adolescence a trend was observed that saw
movement from primarily internal to primarily external to a blend of the two, all in more
sophisticated statements that reflected growing recognition of the complexity of learning
(Table 3).
Table 3. Developmental Progression in Attributions for the Source of Learning

Source
Developmental Stage
Examples
Middle childhood - Early primary External

school (Grades 1 and 2) Scientists, teachers and other smart people; the
government
Middle childhood - Elementary school Internal and external
(Grades 3 to 6) Some learning comes from your brain and some learning
comes from experiences.
Early adolescence (Grade 7) Internal
I think learning comes from our instincts. We just have to
keep exploring new things. That’s what drives evolution.
Adolescence I (Early entrance and External (more sophisticated notions)
Grades 10 and 12) Life itself provides a huge learning opportunity.
I think it can come anywhere.
Adolescence II (Early entrance and Internal (more sophisticated notions) and external;
Grades 10 and 12) interaction between brain/mind and environment
Your brain senses because you need to understand.
Biologically, there are associations of stuff – what we see,
perceive, think.
Discussion
While the research was limited by unequal sample sizes, with few participants at the
youngest and oldest points in the age range, and the cross-sectional design, it provides a
beginning to articulating how gifted learners think about learning and knowledge. The
youngest children’s responses suggest that there are developmental precursors to
epistemological beliefs and that these earliest understandings lay the foundations for
more sophisticated understandings of learning and knowledge. As the research was
conducted in school settings, a bridge was made to what learning and knowledge mean
to children in schools, contributing to our knowledge of how personal epistemologies are
relevant to school learning. The children’s responses also offer insight into how some of
our brightest students conceive of learning.
Understanding of Learning: A Social Narrative or Epistemological Structure?

Children’s understanding of learning can be conceptualized as part of a central social
narrative structure. It is concerned with making meaning of actions, understanding one’s
own intentions with regard to learning, and in adolescence, interpreting one’s learning in
the context of knowledge traditions. However, understanding of learning may be broader
than the domains conceived by Case as having unique conceptual and operational
structures (Case, 1992; Case, Demetriou, Platsidou, & Kazi, 2001). In Case’s system, central
conceptual structures have been proposed in the domains of number, spatial
representation, and social narrative.
These structures (1) represent the core content in a domain of knowledge, (2) help children to think about
the problems that the domain presents, and (3) serve as a tool for the acquisition of higher order insights
into the domain in question. (Case et al., p. 328)
Children’s understandings of learning, while they can be explained by a central social
narrative structure, appear to be relevant to other central conceptual structures as well.
A central epistemological structure might be proposed. However, given the domain-
specific nature of central conceptual structures, a central epistemological structure, to be
theoretically consistent, would need to explain understandings in a delimited domain of
understanding. Understanding of learning appears to entail a more general conceptual
understanding. It may be that beliefs about learning are processes that underlie or
interconnect (Case et al., 2001) central conceptual structures. Case’s (1992) executive
processes or Demetriou, Efklides, and Platsidou’s (1993) hypercognitive system, for
example, entail self-understanding and self-regulation. These “enable the thinker to select
and organize the particular computations and concepts required by a task addressed to
any of the specific domains or to any combination of them” (Case et al., p. 322). These
constructs resemble the reflective abstraction that was, for Piaget, a central
developmental process (Campbell, 1993). Analysis of the relationships of these beliefs to
children’s conceptions of themselves as learners in different academic disciplines
(reading, writing, and mathematics) suggests strong parallels with the development of
understanding of learning (Porath & Lupart, in press), providing some support for
understanding of learning and oneself as a learner as a central developmental process, an
underlying mechanism for children’s understandings of learning in general and academic
subjects in particular. Epistemological beliefs are concerned with making meaning of
school, understanding one’s own intentions with regard to learning, and interpreting one’s
learning in the context of knowledge traditions. These beliefs develop from action-based
understanding of the educational endeavour to the awareness of personal preferences
and needs and, ultimately, to sophisticated reflections on the nature of knowledge and
one’s role in knowledge development.
Educational Implications
Understanding the epistemological underpinnings of students’ approaches to learning
and academic subjects can help educators to support students to consolidate their
64 M. Porath et al.
current constructions of learning and move them forward to more sophisticated

constructions (McKeough, Okamoto, & Porath, 2002). Appreciating and supporting the
foundations of learning in this way may contribute to students achieving richer and
deeper knowledge of academic disciplines. Mansfield and Clinchy (2002) support this
approach, arguing that teaching should start from students’ perspectives on what is to be
studied, allowing for personal meaning to be considered in education. Through helping
students to consolidate their understanding and then supporting them to build bridges to
more sophisticated understanding, teachers also provide students the support necessary
to achieve optimal levels of understanding (Kitchener, 2002). Without this support, their
understandings, even those of the brightest students, can remain at a functional level
(Kitchener).
Directions for Research

This study provides a starting point for continued exploration of gifted learners’
epistemological beliefs and the role those beliefs play in achieving excellence. Further
research could take a longitudinal approach, tracking students’ growing sophistication in
beliefs about learning and knowledge. Fine-grained analyses of process, such as can be
accomplished with microgenetic analysis that examines changes while they occur during
instruction, could provide knowledge about what instructional strategies and supports
facilitate development. The fine-grained and intensive examination of processing that
microgenetic analysis allows can provide suggestions about cognitive change
mechanisms relevant to personal epistemology and indicate directions for extension of
instructional models (McKeough & Sanderson, 1996; Pellegrino, Chudowsky, & Glaser,
2001). It facilitates the tracking of developmental processes from “specifiable beginnings”
(as described by central conceptual understanding) through increasingly competent
approximations to a more sophisticated understanding (Catan, 1986). Importantly, this
tracking of process would take place in the context of learning more about what children
have to tell us about their perspectives on themselves as learners and their views of
knowledge.
Acknowledgements
A version of this paper was presented in the symposium, Developmental Trajectories in
Narrative Thought: Contributing Factors and Underlying Mechanisms, XIIth European
Conference on Developmental Psychology, University of La Laguna, Tenerife.
This research was made possible by a grant from the Social Sciences and Humanities
Research Council of Canada.
References
Bickerton, G. (1994, June). Narrative knowledge as Campbell, R. L. (1993). Commentary.
revealed in children’s perceptions of learning. Epistemological problems for neo-Piagetians.
Paper presented at the Canadian Society for In A. Demetriou, A. Efklides, & M. Platsidou,
the Study of Education, The University of The architecture and dynamics of developing
Calgary. mind (pp. 168-191). Monographs of the Society
Bransford, J. D, Brown, A. L. & Cocking, R. C. for Research in Child Development, 58(5-6,
(Eds). (2000). How people learn: Brain, mind, Serial No. 234).
experience, and school. Washington, DC: Case, R. (1992). The mind’s staircase: Investigating
National Academy Press. the conceptual underpinnings of children’s
Bruner, J. (1986). Actual minds, possible worlds. thought and knowledge. Hillsdale, NJ:
Cambridge, MA: Harvard University Press. Erlbaum.
Bruner, J. (1996). The culture of education. Case, R., Demetriou, A., Platsidou, M., & Kazi, S.
Cambridge, MA: Harvard University Press. (2001). Integrating concepts and tests of
Burr, J. E., & Hofer, B. K. (2002). Personal intelligence from the differential and
epistemology and theory of mind: developmental traditions. Intelligence, 29,
Deciphering young children’s beliefs about 307-326.
knowledge and knowing. New Ideas in Case, R., & Okamoto, Y. (1996). The role of central
Psychology, 20, 199-224. conceptual structures in the development of
children’s thought. Monographs of the Society of folk epistemology. New Ideas in

for Research in Child Development, 61(1-2, Psychology, 20, 309-328.
Serial No. 246). Kuhn, D., Cheney, R., & Weinstock, M. (2000). The
Catan, L. (1986). The dynamic display of process: development of epistemological understand-
Historical development and contemporary ing. Cognitive Development, 15, 309-328.
uses of the microgenetic method. Human Mansfield, A. F., & Clinchy, B. McV. (2002).
Development, 29, 252-263. Toward the integration of objectivity and
Demetriou, A., Efklides, A., & Platsidou, M. (1993). subjectivity: Epistemological development
The architecture and dynamics of developing from 10 to 16. New Ideas in Psychology, 20,
mind. Monographs of the Society for Research 225-262.
in Child Development, 58(5-6, Serial No. 234). Mascolo, M. F., Li, J., Fink, R., & Fischer, K. W.
Fosnot, C. T. (Ed.) (1996).Constructivism. Theory, (2002). Pathways to excellence: Value
perspectives, and practice. New York: presuppositions and the development of
Teachers College Press. academic and affective skills in educational
Griffin, S., & Case, R. (1996). Evaluating the contexts. In M. Ferrari (Ed.), The pursuit of
breadth and depth of training effects when excellence through education (pp. 113-146).
central conceptual structures are taught. In R. Mahwah, NJ: Erlbaum.
Case & Y. Okamoto (Eds.), The role of central McKeough, A. (1992). A neo-structural analysis of
conceptual structures in the development of children's narrative and its development. In R.
children’s thought (pp. 83-102). Monographs Case, The mind's staircase: Exploring the
of the Society for Research in Child conceptual underpinnings of children's thought
Development,61(1-2, Serial No. 246). and knowledge (pp. 171-188). Hillsdale, NJ:
Harter, S. (1996). Teacher and classmate Lawrence Erlbaum.
influences on scholastic motivation, self- McKeough, A., & Genereux, R. (2003).
esteem, and level of voice in adolescents. In J. Transformation in narrative thought during
Juvonen & K. R. Wentzel (Eds.), Social adolescence: The structure and content of
motivation: Understanding children's school story compositions. Journal of Educational
adjustment (pp. 11-42). New York: Cambridge Psychology, 95, 537-552.
University Press. McKeough, A., Genereux, R., & Jeary, J. (2006).
Henwood, K., & Pidgeon, N. (2003). Grounded Structure, content, and language usage: How
theory in psychological research. In P. M. do exceptional and average storywriters
Camic, J. E. Rhodes, & L. Yardley (Eds.), differ? High Ability Studies, 17, 203-223.
Qualitative research in psychology: Expanding McKeough, A., Okamoto, Y., & Porath, M. (2002,
perspectives in methodology and design (pp. April). A design for development: The legacy
131-155). Washington, DC: American of Robbie Case. Paper presented at the
Psychological Society. American Educational Research Association,
Hofer, B. K., & Pintrich, P. R. (1997). The New Orleans.
development of epistemological theories: McKeough, A., & Sanderson, A. (1996). Teaching
Beliefs about knowledge and knowing and storytelling: A microgenetic analysis of
their relation to learning. Review of developing narrative competency. Journal of
Educational Research, 67, 88-140. Narrative and Life History, 6, 157-192.
Huxtable, M. (2009). Creating inclusive and Muis, K. R. (2007). The role of epistemic beliefs in
inclusional understandings of gifts and talents self-regulated learning. Educational Psycholo-
through living educational theory research. In gist, 42, 173-190.
T. Balchin, B. Hymer, & D. J. Matthews (Eds.), Okamoto, Y., Curtis, R., Jabagchourian, J. J., &
The Routledge international companion to Weckbacher, L. M. (2006). Mathematical
gifted education (pp. 292-298). London: precocity in young children: A neo-Piagetian
Routledge. perspective. High Ability Studies, 17, 183-202.
Hymer, B. J. (2009). Beyond compare? Thoughts Pellegrino, J. W., Chudowsky, N., & Glaser, R.
towards an inclusional, fluid and non- (Eds.) (2001). Knowing what students know.
normative understanding of giftedness. In T. The science and design of educational
Balchin, B. Hymer, & D. J. Matthews (Eds.), assessment. Washington, DC: National
The Routledge international companion to Academy Press.
gifted education (pp. 299-307). London: Perner, J. (1991). Understanding the representa-
Routledge. tional mind. Cambridge, MA: MIT Press.
Hymer, B., Whitehead, J., & Huxtable, M. (2009). Porath, M. (2006). The conceptual underpinnings
Gifts, talents and education: A living theory of giftedness: Developmental and educational
approach. Chichester, UK: Wiley-Blackwell. implications. High Ability Studies, 17, 145-157.
Jackson, N. E. (2000). Strategies for modeling the Porath, M., & Lupart, J. (2007, May). The child as
development of giftedness in children. In R. psychologist: Gifted students making meaning
C. Friedman & B. M. Shore (Eds.), Talents of learning and knowledge. Paper presented
unfolding: Cognition and development (pp. 27- at the Canadian Society for the Study of
54). Washington, DC: American Psychological Education, University of Saskatchewan.
Association. Porath, M., & Lupart, J. (in press). Gifted children’s
Kitchener, K. S. (2002). Skills, tasks, and representations of learner identities.
definitions: Discrepancies in the Exceptionality Education International.
understanding and data on the development
66 M. Porath et al.
Robinson, N. M., Zigler, E., & Gallagher, J. J. Ziegler, A., Stoeger, H., & Mundi, M. (2004). The
(2000). Two tails of the normal curve: epistemic learner model. In Tourón, J. (Ed.),
Similarities and differences in the study of Educational technology for gifted education:
mental retardation and giftedness. American From information age to knowledge era (pp.
Psychologist, 55, 1413-1424. 13-20). European Council for High Ability:
Pamplona, Spain.
The authors
Marion Porath is a Professor in the Department of Educational and Counselling

Psychology, and Special Education in the Faculty of Education at the University of British
Columbia, Vancouver, Canada. Her research interests include different forms of
giftedness, young children’s social development, instructional applications of
developmental theory, and problem-based learning. Her current research focuses on
gifted and gifted/learning disabled children’s and adolescents' views of themselves as
learners and their understanding of teaching and learning and the roles of intrapersonal
and interpersonal competencies in pedagogy.
Judy L. Lupart is a Professor and Canada Research Chair in Special Education, beginning
January 2003, in the Department of Educational Psychology at the University of Alberta.
She has served as the Founding Director of the University of Calgary Centre for Gifted
Education, and the Founding Editor of the journal Exceptionality Education Canada. Her
research, publication and teaching interests include inclusive education and school
transformation; cognition, transfer and educational applications; learning disabilities,
giftedness and at-risk learners; and girls, women and achievement.
Dr. Jennifer Katz has recently joined the Department of Educational Administration,
Foundations and Psychology in the Faculty of Education at the University of Manitoba,
Canada. Dr. Katz received her PhD in 2008 from the University of British Columbia and is
currently writing a book on “Teaching to Diversity: Universal Design for Learning and
Inclusive Learning Communities” for Pearson Education Press. Dr. Katz has been a
successful special education teacher, sessional lecturer, educational consultant,
classroom teacher, editor, and guidance counselor. Following her early years as an
elementary teacher, her work as an advocate of inclusive education is evident by her
written works and conference presentations. Dr. Katz brings to the faculty broad research
interests in the areas of: Inclusive special education, teaching to diversity, assessment,
and social and emotional learning development.
Constantine Ngara (PhD, M. Ed, PGDE, B. Ed.) is an Assistant Professor at University of

Bahrain Teachers College. In his doctoral thesis (at University of British Columbia),
Constantine studied Shona stone sculptors of Zimbabwe’s talent attributions and
proposed a dynamic interactive process model (DIPM) that explains talent development
from an African perspective espoused in Shona culture. His research agenda is to
understand giftedness and creativity from a cultural perspective with a view to inform
gifted education from a sociocultural perspective. His collaborative work and
publications focus on conceptions of giftedness and creativity.
Pamela Richardson is a Doctoral Candidate in Special Education (High Ability) at the

University of British Columbia in Vancouver, British Columbia, Canada, and an Instructor
in the Faculty of Education at Vancouver Island University in Nanaimo, British Columbia,
Canada. In her research, she employs arts-based, narrative and poetic inquiry methods
to explore experiences and representations of giftedness within actual and imaginary
contexts.
Talent Development & Excellence The Leonardo Lab 67
Vol. 1, No. 1, 2009, 67-78
The Leonardo Laboratory:

Developing Targeted Programs for Academic
Underachievers with Visual-Spatial Gifts
Tina M. Newman1, William Brown2, Lesley Hart1, Donna Macomber1, Niamh Doyle3,#,
Sergey A. Kornilov4, Linda Jarvin5,#, Robert J. Sternberg5,# and Elena L. Grigorenko1,4,6*
Abstract: This article presents a short-term program, The Leonardo Laboratory, for
children with coexisting learning disabilities and spatial gifts. The program was
developed in collaboration with a museum and administered as an after-school
program to a group of elementary-school children (20 boys and 5 girls) so that its
effectiveness could be evaluated preliminarily. Although the students in the program
did not demonstrate statistically significant gains in academic skills, they
demonstrated gains in self-efficacy and improvements in organizational skills. Thus,
the program we developed has promise for certain kinds of skills and attitudinal
development. Our work shows the importance of addressing the needs of children
with disabilities and gifts (twice exceptional children).
Kewords:
Twice exceptional children, learning disabilities, spatial gifts, after-school program
Students with coexisting learning disabilities and special talents (twice exceptional
children) are an often under-identified and under-served segment of the school-age
population (Baum, 2004; Dix & Schafer, 2005; Kalbfleisch & Iguchi, 2008). Although some of
these students are provided services for either their gifts or their learning difficulties,
very few are identified for services that both develop their areas of weakness and allow
them to explore their areas of strength (Bianco, 2005; Brody & Mills, 1997). This oversight
may have significant consequences, both direct and indirect, on students’ opportunities to
succeed in careers that utilize their areas of strength (Johnsen & Kendrick, 2005). Directly,
the students have little or no opportunity to develop their abilities and learn to
compensate for their disabilities. Indirectly, this lack of services may create a lessened
sense of self-efficacy. Ultimately, successfully intelligent individuals both capitalize on
strengths and compensate for or correct weaknesses, so instruction should address both
strengths and weaknesses (Sternberg & Grigorenko, 2007).
Given that these students require services both for their gifts and for their learning
disabilities, there have been a number of documented interventions aimed at developing
services specifically for this group of students. Predominantly, these programs are
described in literature that covers case studies and anecdotal reports (e.g., Hua, 2002;
Thrailkill, 2005; Turk & Campbell, 2005; Volker, Lopata, & Cook-Cottone, 2006); yet, there
are a few illustrations of small-scale group-administered programs (see below).
In terms of their objectives, these programs can be classified into three groups. First,
some are designed to address specific weaknesses these students possess. For example,
1
Yale University, Child Study Center, 230 South Frontage Road, New Haven, CT 06519-
1124,USA
2
The Eli Whitney Museum, USA
3
Syracuse University, USA
#
At Yale University during the completion of the Leonardo Lab project
4
Moscow State University, Russia
5
Tufts University, USA
6
Columbia University, USA
*
Corresponding author. Email: elena.grigorenko@yale.edu
68 T. M. Newman et al.
one program utilizes a process-based remediation strategy to improve the reading skills
of three students who have both gifts and learning disabilities (Crawford & Snart, 1994).
Second, other programs address areas of weakness but are primarily designed to develop
areas of strength through enrichment activities (e.g., Baum, 1988; Cooper, Baum, & Neu,
2004). One such intervention, Project High Hopes, sought to identify and develop
scientific talents in students with special needs (Cooper et al. 2004). Finally, a number of
programs are designed to develop areas of strength while remediating weaknesses (e.g.,
Bees, 1998; Weinfeld, Barnes-Robinson, Jeweler, & Shevitz, 2002). For example, the Wings
program (Weinfeld et al., 2002) was developed to ensure that students with both gifts and
learning disabilities have access to an enriched and accelerated curriculum, while also
receiving instruction, adaptations, and accommodations in their areas of weakness
through a cascade of services from integrated to self-contained classes, depending on
each student’s needs.
In addition to focusing on academic and cognitive skills, many programs acknowledge an
additional goal of increasing student self-efficacy and self-regulation. Both the literature
on students with gifts and the literature on students with learning disabilities describe
children who are at risk for socio-emotional difficulties, such as poor self-concept and
high levels of frustration and anxiety (Baldwin, 1999; Dole, 2000). Thus, a significant
benefit can be provided by programs designed to increase students’ self-efficacy.
While limited in number and scale, the literature on interventions for twice exceptional
students provides some generalizations regarding advisable approaches to educating
such students. One focal finding has been the importance of providing these students with
a curriculum relevant to their gifts and talents (Baum, Cooper, & Neu, 2001; Neilsen &
Mortoff-Albert, 1989; Weinfeld et al., 2002). The benefits of gifted-and-talented
programming have been seen in both academic achievement and self-concept. For
example, in a study of 76 grade 3 to 5 twice exceptional students, researchers (Neilsen &
Mortoff-Albert, 1989) found that those receiving a combination of gifted-and-learning-
disability services or only gifted services reported higher self-concept than those
students receiving interventions exclusively focused on remediating weaknesses.
Another finding in the literature has highlighted the importance of having an
individualized education plan that sets both strength-building and weakness-correcting
goals (Baum et al., 2001; Cooper et al., 2004; Shevitz, Weinfeld, Jeweler, & Barnes-
Robinson, 2003; Weinfeld et al., 2002). For example, the Wings program ensures that
students have individualized and appropriate instruction and accommodations in their
area of disability, while securing access to gifted programs or mentorships depending
upon need (Weinfeld et al., 2002).
Pedagogical programs for students with both gifts and learning disabilities have not been
rigorously evaluated. Yet, the preliminary evidence suggests that approaches addressing
both strengths and weaknesses appear to have a positive impact on students’ attitudes
toward school and their commitment to academic work.
Prompted by the success of programs for students with learning disabilities whose gifts
are in the analytical domain, and recognizing that strengths in other areas can promote
success in life, we developed and implemented the Leonardo Laboratory program. This
program was designed to address the creative and practical gifts of children who
experience difficulties in academic subjects, but who have strong spatial abilities. Gifts
that allow people to be successful in life can be found in a variety of domains, including
strengths in visual-spatial skills or creativity and innovation (Aaron, Joshi, & Ocker, 2004;
Mann, 2006; VonKarolyi & Winner, 2004). One goal of the Leonardo Laboratory program
was to develop an enrichment curriculum that allows students to develop their identified
strengths while at the same time identifying and remediating areas of weakness
(Sternberg & Grigorenko, 2007). A second goal was to increase self-efficacy and enhance
the self-regulation of these students, whose combination of gifts and disability has often
not led to the development of adequate related skills.
The Leonardo Lab 69
The present project was designed to (1) identify students who have academic difficulties
in reading, spelling, or math, while also displaying exceptional abilities in the creative arts
(e.g., drawing, painting, or design) or in practical/spatial skills (e.g., model building); (2)
develop an after-school program designed to address both the strengths and weaknesses
of these students; and to (3) preliminarily evaluate the effectiveness of this program. Thus,
we wanted to appraise findings from prior case-study and small-group research
suggesting that when children with both gifts and learning disabilities are provided with a
program that addresses their strengths, they demonstrate growth in areas of weakness
that are targeted within the program (e.g., task approach and follow-through), even
though they might not demonstrate improvement in areas of weakness that are not
targeted (e.g., reading-related and math skills). We also wanted to validate a claim from
prior research that students with both gifts and learning difficulties demonstrate growth in
self-efficacy and self-regulation when provided with an intervention program that
highlights their strengths and gives them opportunities to meet other students with
similar learning profiles.
To summarize, this project was conceived to develop an intervention for children with
coexisting spatial gifts and difficulties in reading/writing and/or math. In this article, we
describe the intervention, comment on the feasibility of its implementation in after-school
settings, and present some pilot data reflecting the promise of the intervention. Of note is
that, although there are no formal estimates of the prevalence of double exceptionalities,
the literature suggests that the number of twice exceptional children is low. Given this fact,
recruiting a second group of twice exceptional children to serve as controls was not
feasible, while a group of more typical students would not serve as adequate controls,
being drawn from too different a population. Hence our study was done as a preliminary
study only to assess whether there were any effects apparently associated with the
program, with plans for a future controlled study if these minimal conditions showed any
effects.
Method
Participants
The participating students were recruited through a brochure sent out to local elementary
schools and families who had previously participated in programs at the Eli Whitney
Museum (http://www.eliwhitney.org/) asking teachers and parents to nominate students
in grades 4 through 6 who demonstrated, simultaneously, both learning difficulties and
talents in drawing or building activities. A total of 33 students were nominated by their
teachers and parents; of those, 30 students attended an evaluation session. Of these
students, 11 were in grade 4, 14 in grade 5, 3 in grade 6, and 2 were being homeschooled.
There were 24 male students and 6 female students, and the average age of the
participants was 10 years 2 months (SD=.81). The program inclusion criteria were (1)
reading, writing, or mathematics difficulties and (2) talent in at least one other area of
endeavor (building, practical problem-solving, drawing). Of the 30 students assessed, one
was reported as not having any learning difficulties at school, and one was not interested
in the initial building project, and therefore deemed not well matched to the program. A
total of 28 students began the program; 25 students completed the program. Table 1
presents the group by age, gender, and ethnicity.
Table 1. Means and Standard Deviations of Age, Grade, Gender, and Ethnicity of Program
Participants
Age Grade Gender Ethnicity

M SD M SD Male Female Asian African-Amer Caucasian
10.12 0.67 4.60 0.71 20 5 4% 12% 84%

Project Site
The intervention program was designed in collaboration with and implemented by the Eli
Whitney Museum (http://www.eliwhitney.org/). The Museum is founded on the ideas of
industrial revolutionist Eli Whitney and seeks to encourage the resourcefulness,
innovation, and entrepreneurship exhibited by its namesake through a unique tradition of
learning via self-guided, hands-on problem-solving and experiences beyond the
conventional classroom.
The Museum offers projects that begin in play and grow into art, architecture, engineering
and science, and they include Japanese Toys and Technology, Rubber-Band Powered Cars,
Aeromodeling, and Boat Building. The intervention program at the Eli Whitney Museum
featured here was based on the works of Leonardo DaVinci and was led by the director of
the museum, William Brown.
Procedure
To collect pilot data on the outcomes of the intervention, participants were assessed
before beginning the Leonardo Laboratory program (Time 1, Pretest), then re-evaluated
again at the completion of the program (Time 2, Posttest). The tests used were group-
administered measures of nonverbal reasoning, math calculation, phonological and
orthographical knowledge, and perception of task approach and follow-through skills. For
performance assessment, students were asked to build a rubber-band powered car and
their projects were assessed on two dimensions: (1) practical problem-solving and (2)
creativity. In addition, parents completed a rating scale assessing their child’s task
approach and follow-through skills.
Targeted Skills
In designing the intervention program, the primary goal was to focus on students’
strengths while developing areas of weakness. In a recent surge of research in developing
self-regulated learning skills for children with learning disabilities (Harris, Reid, &
Graham, 2004), self-regulation and the related concept of metacognition are described as
encapsulating a number of skills, including (1) setting goals; (2) developing strategies to
meet those goals (planning) (Butler, 1998); (3) consciously monitoring progress (Butler &
Winne, 1995); (4) adjusting approaches as required (Butler & Winne, 1995); and (5) using
motivation to cope with setbacks or obstacles (Corno, 1993). There is evidence that
students with learning disabilities in postsecondary education continue to have
difficulties with these task-approach and follow-through strategies, and that those college
students with learning disabilities who employ self-regulation strategies are more
successful (Butler, 1998; Trainin & Swanson, 2005).
The purpose of this intervention was to help students develop these self-regulation
strategies – specifically task approach and follow through – while they completed projects
that capitalized on their strengths and helped them practice compensating for
weaknesses, including elements of reading/writing and doing math. The program also
intended to increase children’s self-efficacy.
Materials
Pretest measures included (1) a measure of nonverbal cognitive skills, (2) baseline
measures of academic skills that were areas of difficulty for students as specified in the
recruitment inclusion/exclusion criteria, and (3) a performance-based measure of
practical problem-solving and creativity. For the academic skills, specifically, we assessed:
(1) reading and reading-related skills (phonological processing, orthographical
processing, and print exposure); (2) math calculation skills; and (3) achievement as
indicated by a parental rating. We also collected measures on target areas that are often
challenging for students with learning difficulties, including parent and child ratings of
task approach and follow-through skills. We also administered a hands-on performance
The Leonardo Lab 71
measure to assess students’ potential areas of strengths, including visual/spatial building

skills, practical problem solving, and creative skills.
At posttest, we collected academic skill measures, including the reading and reading-
related skills, math calculation, and student and parent ratings of task-approach and
follow-through skills. A qualitative feedback measure was also obtained.
Nonverbal ability. The Cattell Test of “g”: Culture Fair Scale 2, Form B (Cattell & Cattell,
2002) is a measure of nonverbal reasoning in which the student analyzes visual patterns,
for example, determining missing parts or finding similar patterns. This test was
administered at Time 1 to provide a benchmark level of cognitive functioning and to
ensure that all students were of average to above average ability in nonverbal reasoning.
Scores on this measure indicated that all students had average to superior ability: the
average score for the group was 114.84 (SD=14.30).
Academic skills measures. The Title Recognition Task (Cunningham & Stanovich, 1990)
gauges students' exposure to print. This test, consisting of a list of real and false book
titles, was originally designed to predict reading experience in third- and fourth-grade
children, but has been correlated with other reading-related skills in students in grades 5
through 9 (McBride-Chang, Manis, Seidenberg, Custodio, & Doi, 1993). Participants were
instructed to put a mark next to the titles that they knew were real books. In addition, they
were told that the list contained both real and false titles and authors, and therefore they
were not to guess. This test has two versions and administration was randomly assigned at
pretest, with the other version administered at posttest. Both versions have 25 real books;
version 1 has 15 false titles, version 2 has 16 false titles. Scoring of this measure consisted
of the percentage of real book titles selected minus the percentage of false titles selected.
The same measure was administered at both pretest and posttest. The Phonological
Choice Task (Gayan & Olson, 2001) is a measure of phonological-processing skill
consisting of forced-choice decisions between pairs of nonwords that sound like common
words if read aloud and two nonwords that do not sound like known words. Students were
asked to circle the word in each row that sounded like a real word if read aloud. There are
five practice items and 60 test items that were randomly divided into 2 sets of 30 items to
create A and B versions of the test. Students randomly received either the A or B version at
pretest and were given the alternate version at posttest. Scoring of this measure consisted
of the percentage of correct responses. The same measure was administered at both
pretest and posttest. The Orthographic Choice Task (Olson, Forsberg, Wise, & Rack, 1994)
is a measure of orthographic-processing skill consisting of a forced-choice decision
between a word and a phonetically identical pseudohomophonic nonword (e.g., take –
taik). The task requires the participant to recognize the correct orthographic pattern for
the word independent of its phonology. There are five practice items and 78 test items that
were randomly divided into 2 sets of 39 items to create an A and B version of the test.
Students randomly received either the A or B version at pretest and were given the
alternate version at posttest. Scoring of this measure consists of the percentage of correct
responses. The same measure was administered at both pretest and posttest. The
Woodcock-Johnson (WJ-III) Math Calculation Subtest (Woodcock, McGrew, & Mather, 2001)
provides a measure of a student’s mathematical-computation ability. Questions progress
through single-, double-, and triple-digit addition, subtraction, multiplication, and
division. Harder items require students to perform calculations with fractions, decimals,
and negative numbers. Students in the study were asked to complete as many questions
as they could. Standard scores were obtained. The same measure was administered at
both pretest and posttest.
Self-regulated learning measures. The Organizational Rating Scale (Parent-Report) is a
measure developed for this project that asks parents to rate their child’s behavior on a 5-
point scale. The measure consists of 33 items assessing task-approach and follow-through
skills such as planning, preparing, overcoming obstacles, and independence in
completing home chores, school tasks, and extra-curricular tasks. This measure was
administered at both pretest and posttest. The Organizational Rating Scale (Child-Report) is
a measure developed for this project that asks students to rate their own behavior on a 5-
point scale. The scale consists of 36 items measuring self-perception of task approach and
follow-through skills such as planning, preparing, and completing tasks at home and
school. This measure was administered at both pretest and posttest.
Performance measures. The Rubber-band-powered car task, developed by the Eli Whitney
Museum, gives students the parts and instructions for building a basic car powered by
rubber-bands. The task was assessed on two factors (1) function (practical) and (2) form
(creative). For the functional assessment, the student had to compete to see how far their
car could travel compared with the cars of other students. The students were given trial
and revision opportunities for testing their car. These trial and revision opportunities were
limited in time but not in number, so students could try their car and revise as many times
as they chose during a time period that was set for everyone. For the form assessment,
students were provided with resources to make their car unique.
Qualitative measures. Qualitative feedback was solicited from both parents and students
following the completion of the program. Parents were given a feedback form asking
three questions: (1) please describe any benefits of the program for your child; (2) please
describe any negative outcomes of the program for your child; and (3) please provide any
additional information that you feel may be important for us in evaluating this program.
Students were also given a feedback form with three questions: (1) please describe three
things that made this program good for you; (2) please describe anything about the
program that was not so good for you; and (3) please describe anything else about the
program that you think was great or not so great.
Intervention
General overview. The program included a series of ten workshops that invited students to
complete experimental building projects derived from Leonardo DaVinci’s notebooks.
Using activities initially sketched by DaVinci and further developed for the purposes of
this intervention, the program integrated exercises developing important learning skills
such as: (1) following the demands of the instructor while making projects on their own;
(2) coping with environmental constraints, such as availability of materials and tools; (3)
progressing through steps of the projects, i.e., planning, preparation, time scheduling,
exploring, and follow-through; and (4) specific skills in following directions, overcoming
obstacles, and recognizing their own successes.
Specific application of the goals. Each assignment allowed students to take responsibility
and realize the choices available in directing or ‘steering the project’. It also
acknowledged the restrictions that can be placed on an assignment by facilitators (i.e.,
museum staff members), reminding the student that in most learning situations, there are
boundaries that must be considered. Each project was also subject to environmental
constraints, such as resources and time constraints that can push an endeavor forward,
backward or in a different direction. At the start of each project, students were provided
with the basic materials and instructions for completing the project. A demonstration of
the steps for completing the basic model and examples of different ways of
individualizing the project were provided. Additional resources were provided (e.g., bins
of markers, scraps of wood, pieces of cloth, pipe cleaners, pieces of wire), and students
were reminded that they could use any other resources found at the museum. During
project building, facilitators circulated among the students, asking them about their plans,
how they were going to reach their goal, and what they were going to need. The
facilitators also reminded the students to consider the time available. No specific ideas
were provided to the students, although they were free to talk amongst themselves and
were encouraged to make the project their own. Each obstacle was presented as an
opportunity for choice, a time for decision-making, or a lesson to be learned. In deciding
how to overcome obstacles, students were supported by choices such as back-tracking
and finding alternate ideas, methods, and materials. At the end of each session, students
were taken aside individually and encouraged to consider the obstacles they had
The Leonardo Lab 73
encountered and overcome, the opportunities that had arisen, the opportunities captured,
the set-backs faced, the discoveries or successes, the disappointments, and the lessons
learned in their projects.
Results
Descriptive Statistics
To determine any significant effects of age and gender on the results, correlations were
calculated between these three variables and the rating scales and the academic-skills
measures. No significant correlations were found for age or grade with the Child-Report
measure, the Parent-Report measure, or any of the academic-skills measures
(phonological, orthographical, title recognition, or calculation) at either pretest or
posttest.
Intervention Outcomes
The Parent-Report Organizational Rating Scale and The Child-Report Organizational Rating
Scale were two new measures developed for this study. As these rating scales were new,
we wanted to ensure their reliability across time. The child and parent scales showed a
high degree of consistency with each other, indicating that children and parents tended to
agree on the abilities of the children. With respect to the internal consistency of the items,
the Child Rating Scale had a Cronbach’s alpha of 0.88 and the Parent Rating Scale had a
Cronbach’s alpha of 0.91, demonstrating a high degree of internal consistency.
Given the pre/post design of this pilot study and the small sample size, paired-sample t-
tests were conducted to determine any significant differences between the pre- to post-
test academic skills and the pre/post child and parent-ratings on task approach and
follow-through skills.
Academic skills assessed by group-administered tests of reading-related skills and math
calculation skills showed no significant differences from pre to post test, neither on the
omnibus, nor on individual tests. As these were not areas of intervention, this finding was
expected (see Table 2 for descriptive and inferential statistics).
Table 2. Means, Standard Deviations, and T-test Data of Pretest and Posttest Measures of Reading-
and Math-Related Skills
N Pretest Posttest Difference
M SD M SD t-score p-value
Title recognition task 21 0.21 0.22 0.27 0.18 -1.00 0.33
Phonological choice task 18 0.64 0.25 0.65 0.27 -0.19 0.85
Orthographic choice task 22 0.84 0.12 0.85 0.13 -0.17 0.86
Math calculation task 22 93.06 14.69 95.39 15.64 -0.85 0.40
The parent-rating scale and child-rating scale assessed skills that were targeted in the
intervention, and therefore, differences between pre and post ratings of parents’ and
children’s perceptions of task approach and follow-through skills were expected. The
parent rating scale showed a significant difference between pre and post ratings, with
scores rising following the intervention, indicating that, overall, parents saw an
improvement in their children’s task-approach and follow-through skills. In particular, the
following items showed a significant positive change or approached a significant change
in rating from pretest to posttest: (1) overcoming obstacles in completing a homework
task; (2) leaving the house on time for school with minimal prompting; (3) following
multiple steps to complete a household chore; (4) overcoming obstacles in completing a
household chore; and (5) following instructions to complete a one-step household chore.
The child-rating scale demonstrated no comparable rise (see Table 3), however, given
that parents are the ones who provide all the support and/or reminders, they are likely the
ones who notice change first. Of note also are the dynamics of the correlations between
parent and child ratings at pre- and posttests (see Table 4).
Table 3. Means, Standard Deviations, and T-test Data of Pretest and Posttest Parent and Child Rating
Scales of Task Approach and Follow-Through Skills
N Pretest Posttest Difference
M SD M SD t-score p-value
Parent Rating Scale 17 56.42 12.78 59.84 2.37 -2.41 0.03
Child Rating Scale 13 122.62 19.09 126.15 16.20 -0.62 .55
Table 4. Correlations of the Pre and Post Administrations of the Parent and Child Rating Skills
Pre Parent Post Parent Pre Child
Pre Parent -
Post Parent 0.89*** -
Pre Child 0.72** 0.65* -
Post Child 0.30 0.32 0.67*
*** p < .0001, ** p < .01, * p < .05
Qualitative feedback from parents converged on three themes: (1) the development of
areas of weakness, for example, “he is more willing to sit down and do homework, less
resistant, more persevering”; (2) the development of areas of strength, for example, “he
continued to work on each project when he got home”; and (3) the development of self-
efficacy, for example, “he doesn’t shrug off his accomplishments.” In general, parent
qualitative feedback confirmed the findings of the parent rating scale that children
showed improvement in areas of weakness. In addition, parents also indicated that
children experienced benefits to their self-efficacy and developed further their areas of
strength. Qualitative feedback from the students reflected primarily themes of self-
efficacy, for example, “showed me I could do good things that people love and want.”
The Leonardo Lab 75
Discussion
The results of the current study are consistent with the literature that suggests that
students who have both gifts and learning disabilities benefit from programs targeted to
developing their areas of strength. Although very preliminary and based only on
informants’ reports, this study provides some quantification of positive effects on the
improvement of targeted skills by the intervention. The students who participated in this
study diverged from those in previous studies in the fact that the focus was not on students
who demonstrated cognitive gifts, but on those whose gifts were in the creative and
visual-spatial domains. The intervention offered also differed in that students were
delivered a program out of the typical school curriculum, focused on developing not only
their strengths but also on areas of weakness that impact school performance. In
particular, the intervention targeted skills in task initiation and follow-through, areas often
found to be weak in students with learning difficulties. To quantify these skills, new
measures were devised to obtain ratings from parents and students on the application of
these skills from the intervention setting to outside settings, specifically, the home
environment and especially homework tasks.
The new measures offered a way to capture parents’ perceptions of students’ strengths
and weaknesses in task approach and follow-through. Parents reported significant
improvement on a number of items. Finally, parent and student qualitative statements
corroborated these findings.
One of the most important findings reported by both parents and students was the growth
in self-efficacy and self-recognition of an area of strength. At pretest, many of these
students were reported as experiencing social and emotional difficulties, including
anxiety and sadness. Parents reported concerns that their children were exhibiting
indications of learned helplessness in their school environments and were demonstrating
less effort in their schoolwork. At the time of the posttest, both parents and students
reported a re-engagement in goal-directed tasks and a sense that the student could
accomplish something worthwhile. One indicator of this program’s positive benefit to
these students is the number of students who have continued to engage in programs at the
Eli Whitney museum as both students and, later, facilitators, after first being introduced to
the Leonardo Laboratory.
Yet, although this work resulted in the development of a new and engaging program, the
data we present here are only preliminary and, though encouraging, should be
interpreted with caution. To validate and strengthen our claims about the potential of the
Leonardo Laboratory, a number of steps need to be taken. First, a formal effectiveness
study with a large sample of participants and a control group needs to be carried out.
Second, a battery of performance-based assessments capturing the skills targeted by the
program needs to be administered using the pre-/posttest design. Here we used
convergent information, provided by parents and the students themselves, but, as
indicated above, this is only the first step in carrying out an evaluation of the program.
Clearly, other adults, such as students’ teachers and museum personnel can be included
as additional informants. The positive feedback from parents and students might be
indicative of wishful thinking for change, rather than change itself. Finally, the
relationships that a number of the students subsequently built with the museum after their
participation in the Leonardo Laboratory, although encouraging, should also be
interpreted with caution: their comfort and satisfaction with the atmosphere of the
museum might not be indicative of learning and change.
In summary, this report is only an initial presentation of a promising intervention program
for children who have coexisting gifts and disabilities. With the program developed, the
next step is to formally evaluate it. Hopefully, there will be future opportunities for that
step to be made.
Acknowledgements
This research was supported by Grant R206R950001 under the Javits Act program as
administered by the Office of Educational Research and Improvement, U.S. Department of
Education. Grantees undertaking such projects are encouraged to express freely their
professional judgment. This article, therefore, does not necessarily represent the position
or policies of the U.S. Department of Education, and no official endorsement should be
inferred. We express our gratitude to Ms. Mei Tan for her editorial assistance.
References
Aaron, P. G., Joshi, R. M., & Ocker, E. S. (2004). Education, 15, 162-169.
Dyslexia and visual-spatial talents: Are they Corno, L. (1993). The best laid plans: Modern
connected? In T. M. Newman & R. J. conception of volition and educational
Sternberg (Eds.), Students with both gifts research. Educational Researcher, 22, 14-22.
and learning disabilities: Identification, Crawford, S., & Snart, F. (1994). Process-based
assessment, and outcomes (pp. 199-234). remediation of decoding in gifted LD
New York, NY: Kluwer Academic/ Plenum students: Three case studies. Roeper Review,
Publishers. 16, 247-252.
Baldwin, A. Y. (1999). Learning disability: The Cunningham, A. E., & Stanovich, K. E. (1990).
mysterious mask - The USA perspective. In Assessing print exposure and orthographic
A. Y. Baldwin & W. Vialle (Eds.), The many processing skill in children: A quick
faces of giftedness (pp. 103-134). Belmont, measure of reading experience. Journal of
CA: Wadsworth Publishing. Educational Psychology, 82, 733-740.
Baum, S. M. (1988). An enrichment program for Dix, J., & Schafer, S. (2005). From paradox to
gifted learning disabled students. Gifted performance: Practical strategies for
Child Quarterly, 32, 226-230. identifying and teaching Gifted/LD
Baum, S. M. (Ed.). (2004). Twice-exceptional and students. In S. K. Johnsen & J. Kendrick
special populations of gifted students. (Eds.), Teaching gifted students with
Thousand Oaks, CA: Corwin Press. disabilities (pp. 153-159). Waco, TX: Prufrock
Baum, S. M., Cooper, C. R., & Neu, T. W. (2001). Dual Press.
differentiation: An approach for meeting Dole, S. (2000). The implications of the risk and
the curricular needs of gifted students with resilience literature for gifted students with
learning disabilities. Psychology in the learning disabilities. Roeper Review, 23, 91-
Schools, 38, 477-490. 96.
Bees, C. (1998). The GOLD program: A program for Gayan, J., & Olson, R. K. (2001). Genetic and
gifted learning disabled adolescents. environmental influences on orthographic
Roeper Review, 21, 155-161. and phonological skills in children with
Bianco, M. (2005). The effects of disability labels reading disabilities. Developmental
on special education and general education Neuropsychology, 20, 483-507.
teachers’ referrals for gifted programs. Harris, K. R., Reid, R. R., & Graham, S. (2004). Self-
Learning Disability Quarterly, 28, 285-293. regulation among students with LD and
Brody, L. E., & Mills, C. J. (1997). Gifted children ADHD. In B. Wong (Ed.), Learning about
with learning disabilities: A review of the learning disabilities (3rd ed., pp. 167-195).
issues. Journal of Learning Disabilities, 30, San Diego, CA: Elsevier Academic Press.
282-296. Hua, C. B. (2002). Career self-efficacy of the
Butler, D. L. (1998). The strategic content learning student who is gifted/learning disabled: A
approach to promoting self-regulated case study. Journal for the Education of the
learning. In B. J. Zimmerman & D. Schunk Gifted, 25, 375-404.
(Eds.), Developing self-regulated learning: Johnsen, S. K., & Kendrick, J. (Eds.). (2005).
From teaching to self-reflective practice (pp. Teaching gifted students with disabilities.
160-183). New York, NY: Guilford Press. Waco, TX: Prufrock Press.
Butler, D. L., & Winne, P. H. (1995). Feedback and Kalbfleisch, M. L., & Iguchi, C. M. (2008). Twice-
self-regulated learning: A theoretical exceptional learners. In J. A. Plucker & C. M.
synthesis. Review of Educational Callahan (Eds.), Critical issues and practices
Research(65), 245-281. in gifted education: What the research says
Cattell, R. B., & Cattell, A. K. S. (2002). Culture Fair (Vol. 707-719). Waco, TX: Prufrock Press.
Intelligence Tests. Champaign, IL: Institute Mann, R. L. (2006). Effective teaching strategies for
for Personality and Ability Testing, Inc. gifted/learning-disabled students with
Cooper, C. R., Baum, S. M., & Neu, T. W. (2004). spatial strengths. The Journal of Secondary
Developing scientific talent in students with Gifted Education, 17, 112-121.
students with special needs: An alternative McBride-Chang, C., Manis, F. R., Seidenberg, M. S.,
model for identification, curriculum, and Custodio, R. G., & Doi, L. M. (1993). Print
assessment. The Journal of Secondary Gifted
The Leonardo Lab 77
exposure as a predictor of word reading students with learning disabilities. Learning

and reading comprehension in disabled Disability Quarterly, 28, 261-262.
and nondisabled readers. Journal of Turk, T. N., & Campbell, D. A. (2005). What's wrong
Educational Psychology(85), 230-238. with Doug? The academic struggles and
Neilsen, M. E., & Mortoff-Albert, S. (1989). The triumphs of a gifted student with ADHD
effects of special education service on the from preschool to college In S. K. Johnsen &
self-concept and school attitude of learning J. Kendrick (Eds.), Teaching gifted students
disabled/gifted students. Roeper Review, 12, with disabilities (pp. 105-141). Waco, TX:
29-36. Prufrock Press
Olson, R. K., Forsberg, H., Wise, B., & Rack, J. (1994). Volker, M. A., Lopata, C., & Cook-Cottone, C.
Measurement of word recognition, (2006). Assessment of children with
orthographic, and phonological skills. In G. intellectual giftedness and reading
R. Lyon (Ed.), Frames of reference for the disabilities. Psychology in the Schools, 43,
assessment of learning disabilities: New 855-869.
views on measurement issues (pp. 243-277). VonKarolyi, C., & Winner, E. (2004). Dyslexia and
Baltimore, MD: Paul H. Brookes. visual-spatial talents: Are they connected?
Shevitz, B., Weinfeld, R., Jeweler, S., & Barnes- In T. M. Newman & R. J. Sternberg (Eds.),
Robinson, L. (2003). Mentoring empowers Students with both gifts and learning
gifted/learning disabled students to soar! disabilities: Identification, assessment, and
Roeper Review, 26, 37-40. outcomes (pp. 95-118). New York, NY:
Sternberg, R. J., & Grigorenko, E. L. (2007). Kluwer Academic/ Plenum Publishers.
Teaching for successful intelligence (2nd Weinfeld, R., Barnes-Robinson, L., Jeweler, S., &
ed.). Thousand Oaks, CA: Corwin Press. Shevitz, B. (2002). Academic programs for
Thrailkill, C. (2005). Patrick's story: A gifted and talented/learning disabled
gifted/learning-disabled child. In S. K. students. Roeper Review, 24, 226-233.
Johnsen & J. Kendrick (Eds.), Teaching Woodcock, R. W., McGrew, K. S., & Mather, N.
gifted students with disabilities (pp. 99-103). (2001). Woodcock-Johnson III Tests of
Waco, TX: Prufrock Press. Achievement. Itasca, IL: Riverside
Trainin, G., & Swanson, H. L. (2005). Cognition, Publishing.
metacognition, and achievement of college
The authors
Tina Newman, Ph.D. received her doctorate in School/Applied Child Psychology

from McGill University in Canada. Dr. Newman is currently an Associate Research
Scientist at the Yale University Child Study Center. Her clinical and research
interests are in assessment and intervention with children who have individual
learning needs, including learning disabilities, ADHD, and giftedness. Dr.
Newman is particularly interested in children with double exceptionalities (e.g.,
giftedness and learning disabilities) and circumscribed interests in autism
spectrum disorders.
William Brown, MSW, is the Director of the Eli Whitney Museum and Workshop in
Hamden, Connecticut. He is a designer, educator, and social worker who trained
in child development at Columbia University in New York. In addition to directing
all aspects of the Eli Whitney Museum’s programs, Mr. Brown runs an
apprenticeship program and designs and teaches core summer workshops.
Dr. Hart received her Ph.D. in cognitive psychology from the University of
Pittsburgh, with a concentration in neuroscience through the Center for the
Neural Basis of Cognition. Currently, Dr. Hart is an Associate Research Scientist at
the Yale University Child Study Center. Her research interests focus on reading
development, skilled reading performance, and impairments in language and
reading processes.
78 Tina M. Newman et al.
Donna Macomber has a Master’s Degree in Psychology and is a Research

Assistant at the Child Study Center of Yale University. Her primary interests
include childhood development, at-risk youth, educational psychology, and
learning/cognition. She is currently working with at-risk preadolescent and
adolescent children toward improving social problem solving skills. She also
works closely on other projects within her department including a project
committed to improving the delivery of educational services to court involved
youth while in detention.
Niamh Doyle is originally from Dublin, Ireland where she received a B. Ed in Elementary
Education from St. Patrick's College (Ireland) and a B.Sc in Psychology from Open
University (U.K.). Niamh received her MA in Psychology and is currently in her 5th year of a
doctoral program in School Psychology at Syracuse University (USA). Prior to pursuing
graduate studies, she was an elementary school teacher in Ireland and Australia and a
research assistant at the Yale University Child Study Center and PACE center. Research
interests include school and family integration, parent-teacher relationships, early
detection of developmental disabilities, and transitions from pre-k to kindergarten for
children with disabilities. Niamh’s master's thesis examined parents’ perspectives on
education and teacher quality and levels of involvement in school life. Niamh hopes to
extend that line of research examining the validity of parents’ perspectives on teacher
quality and its relationship to classroom quality.
Sergey Kornilov received his MSc/MA in Educational Psychology from Moscow

State University and is currently a graduate student at the University of
Connecticut. His research interests include giftedness and gifted education,
creativity, spoken language comprehension and learning disabilities. His master’s
thesis examined longitudinal changes in academic achievement of Russian
college students as predicted by ability measures developed within R.
Sternberg’s theory of successful intelligence and was awarded the Russian
National Contest for Best Student Scientific Paper Award in 2009.
Linda Jarvin, Ph.D., is an Associate Research Professor in the Department of

Education at Tufts University, and director of its Center for Enhancing Learning
and Teaching (CELT). She received her PhD in Cognitive Psychology from the
University of Paris V (France) and her postdoctoral training at Yale University. She
has extensive experience with curriculum planning and development, designing
and implementing professional development opportunities for k-12 and college
teachers focusing on teaching and assessment, and facilitating programmatic
evaluation plans.
Robert J. Sternberg is Dean of the School of Arts and Sciences, Professor of

Psychology, and Adjunct Professor of Education at Tufts University (formerly-IBM
Professor of Psychology and Education, Professor of Management, and Director of
the Center for the Psychology of Abilities, Competencies, and Expertise at Yale).
Sternberg was the 2003 President of the American Psychological Association and
is President of the International Association for Cognitive Education and
Psychology. He is the author of about 1200 journal articles, book chapters, and
books, and has received over $20 million in government and other grants and
contracts for his research.
Dr. Elena L. Grigorenko is Associate Professor of Child Studies and Psychology at
Yale and Adjunct Professor of Psychology at Columbia University (USA) and
Moscow State University (Russia). She has published more than 250 peer-
reviewed articles, book chapters, and books. She has received awards for her
work from five different divisions of the American Psychological Association. In
2004, Dr. Grigorenko won the APA Distinguished Award for an Early Career
Contribution to Developmental Psychology. Dr. Grigorenko’s research has been
funded by NIH, NSF, DOE, Cure Autism Now, the Foundation for Child
Development, the American Psychological Foundation, and other federal and
private sponsoring organizations.
Instructions to authors
The Journal
• Talent Development and Excellence is an international scholarly journal.
• Manuscripts should be written in accordance with the publication manual of the
American Psychological Association. See http://apastyle.apa.org/
• Manuscripts must be written in English. It is recommended that non-native English
speakers have their papers checked in regard to language accuracy prior to
submission.
Submission and Reviewing

• Manuscripts may be submitted electronically to one of the Editors-in-Chief.
• All personal data will be removed by the Editor prior to the refereeing process to
allow for a masked review.
Manuscript
• Length: A paper submitted should not exceed 6000 words including abstract,
references, illustrations and appendixes.
• Abstract: An abstract should consist of a maximum 125 words. The abstract must, if it
is the result of an empirical research, briefly outline theoretical basis, research
question/s, methodology and instrumentation, sample/s and important characteristics
(e.g. number, gender and age) as well as the main findings of the study. Also, it should
include main conclusions. An abstract for a theoretical article or a review should
describe the topic, the objective or thesis and the scope of the article. It should also
outline the main conclusions.
• Footnotes should be numbered consecutively with superscript Arabic numerals.
• If reporting statistics, sufficient information must be included according to the APA
Manual.
Figures and Tables

• Figures and Tables must be placed on separate pages; not included in the text. They
should be submitted as separate file/s. Figures and Tables must have an Arabic
number, an explaining text and a title. Their approximate place in the text should be
clearly indicated.
• Authors should follow APA format in designing tables and figures and consider the
fact that illustrations supplements - not duplicates - the text. In the text, refer to every
table and figure and tell the reader what to look for. Discuss only the table's
highlights.
Presentation of the Authors

For the presentation of the Author(s) please hand in a digital photograph (min. 300 dpi,
4x5 cm/ 1.6x2 inch) and a description of yourself (max. 100 words).
If you have any questions about references or formatting your article, please contact the
Editors-in-Chief.
Copyright
It is a condition of publication that authors assign copyright or licence the publication
rights in their articles, including abstracts, to the International Research Association for
Talent Development and Excellence. Authors are themselves responsible for obtaining
permission to reproduce copyright material from other sources.
ISSN 1869-0459 (print)/
ISSN 1869-2885 (online)

Tde Issue 1-2009 00 Complete

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tde Issue 1-2009 00 Complete

Uploaded by

Copyright:

Available Formats

July 2009

Official Journal of the

Editors-in-Chief: Albert Ziegler

International Advisory Board:

This issue represents the official inauguration of our scientific journal

The impact of smart fractions, cognitive ability of politicians and average 3

Modeling mathematical Actiotopes: The potential role of CLARION 27

Relative age, talent identification and youth skill development: Do relatively 45

Gifted Learners’ Epistemological Beliefs 57

The Leonardo Laboratory: Developing Targeted Programs for Academic 67

Vol. 1, No. 1, 2009, 1-2

aware of the advantages of modeling, as well as establish further bridges to expertise

The impact of smart fractions, cognitive ability of

ISSN 1869-0459 (print)/ ISSN 1869-2885 (online)

Smart fraction theory

Cognitive ability of leading politicians

Wealth and development indicators of societies

Educational level of societies and attributes of educational systems

Attributes standing for high cognitive achievement of a society (excellence)

HIV, AIDS and criminality indicators

Problems in international data sets

Table 2. Correlations within cognitive ability indicators

Educational level .54 .56 .61 .58 .47

.38 .60 .45 .86

Patent rate Nobel Prize in Scientist rate High tech

.78 (.78) .64 (.64) .74 (.74) .37 (.37)

GDP 1998 .01

Cognitive ability Cognitive ability

Figure 4. Cross-lagged coefficients (and correlations in parentheses) between cognitive ability of

Cognitive ability Cognitive ability

Cognitive ability Cognitive ability

Figure 5. Cross-lagged coefficients (and correlations in parentheses) between cognitive ability of

Cognitive ability and educational policy

Young tracking age Class size

CA-mean .31 -.45

CA-5% .25 -.47

CA-95% .33 -.38

CA-politicians 60-09 .41 -.19

CA-politicians 90-09 .44 -.22

Heiner Rindermann (born 1966) is Professor for Developmental Psychology at

Dr James Thompson is Honorary Senior Lecturer in Psychology in the

Vol. 1, No. 1, 2009, 27-43

Modeling mathematical Actiotopes: The potential role of

ISSN 1869-0459 (print)/ ISSN 1869-2885 (online)

Generally, research on giftedness centers on exceptional performance of some kind, and on

The critical state view of giftedness

Critical State (CS) a

The Actiotope Model of Giftedness

The development of an Actiotope leading to mathematical excellence, for example, proceeds

• Phase 1 – Pre-natal and early childhood development. An individual may show

The Actiotope Model of Giftedness and achievement excellence in

The CLARION cognitive architecture and the Actiotope Model of

Action-centred subsystem (ACS) Non-action-centred subsystem (NACS)

Action centred explicit Non-action centred

Action centred implicit Non-action centred

Motivational subsystem (MS) Meta-cognitive subsystem (MCS)

Motivational subsystems (MS) provide the underlying motivations for

• Proportional activation, where activation of drives should be proportional to

Sensory High-level drives Drive

Meta-cognitive subsystems (MCS) refer to an individual’s knowledge about their cognitive

There are many types of metacognitive processes in the MCS, including:

Antecedents of action repertoires

To take numerosity as an example, the current state condition for a one-to-one

* p < .0001, p < .01, * p < .05