THE CULTURAL MALLEABILITY OF
INTELLIGENCE AND ITS IMPACT ON THE
Lisa Suzuki and Joshua Aronson
New York University
This commentary highlights previous literature focusing on cultural and environ-
mental explanations for the racial/ethnic group hierarchy of intelligence. Assump-
tions underlying definitions of intelligence, heritability/genetics, culture, and race
are noted. Historical, contextual, and testing issues are clarified. Specific attention is
given to studies supporting stereotype threat, effects of mediated learning experi-
ences, and relative functionalism. Current test development practices are critiqued
with respect to methods of validation and item development. Implications of the
genetic vs. culture-only arguments are discussed with respect to the malleability
Rushton and Jensen (2005) review decades of literature to support a genetic
basis for the racial/ethnic group hierarchy in intelligence, a position they have
held unwaveringly for over 30 years. Their report gives little mention to findings
that point to the impact of environment and race (i.e., race as a social construction)
on intellectual development or performance—what they term the culture-only
perspective. We are not among the culture-only adherents as characterized by
Rushton and Jensen. While acknowledging the impact of biological factors on
intelligence test performance, we have examined the impact of cultural/environ-
mental factors that affect performance on aptitude and achievement measures. Our
work, and that of others (e.g., Aronson, 2002; Sternberg, 1996), show us that
intellectual performance is much more fragile and malleable than what is often
noted in the current literature. The goals of our commentary are to highlight,
briefly, assumptions underlying definitions (i.e., intelligence, heritability, genet-
ics, culture, race) and clarify historical, contextual, and testing issues that were
only briefly mentioned by Rushton and Jensen. Finally, we comment on the
heuristic value and on policy implications of the research.
Problematic Assumptions Underlying Definitions
Rushton and Jensen’s (2005) argument rests on particular definitions of
intelligence, genetics (i.e., heritability), culture, and race. It is, in part, differences
in definitional assumptions that have allowed researchers to claim support for
distinctly different perspectives (i.e., environment vs. genetics) based on the same
data (Hayman, 1998).
Lisa Suzuki and Joshua Aronson, Department of Applied Psychology, New York University.
We would like to thank John Kugler, Leo Wilton, Jacqueline Mattis, and Muninder Ahluwalia
for their feedback on earlier versions of this commentary.
Correspondence concerning this article should be addressed to Lisa Suzuki, Department of
Applied Psychology, New York University, East Building 239 Greene Street 409, New York, NY
10003. E-mail: email@example.com
Psychology, Public Policy, and Law
2005, Vol. 11, No. 2, 320–327
Copyright 2005 by the American Psychological Association
Numerous theories of intelligence have been framed and reframed over the
years as scholars have ruminated about what constitutes “intelligence.” In this
section we highlight a few of the issues that complicate the linkage between race
and IQ as presented by Rushton and Jensen (2005). As Fagan and Holland (2002)
argued, IQ scores represent a composite of how well one does in comparison with
one’s peers. Test performance is a measure of a person’s intellectual ability that
is dependent on one’s genetic makeup and affected by environment and cultural
experiences (e.g., informal learning and schooling).
Psychometric definition of g.
Although different theories of intelligence
have been noted throughout the literature, intelligence has been defined to a large
extent by the tests designed to measure it. It should be noted that others have
challenged the emphasis on measurement and focused on the processing compo-
nent of intelligence (Fagan, 2000). The psychometric definition of intelligence has
led to debates about what should constitute the focus of IQ tests. Some have
argued that the measurement of intelligence is based primarily on the concept of
g (general intelligence) and related subabilities, whereas other researchers have
posed that intelligence should be measured as numerous intelligences of more or
less equal status (e.g., Gardner, 1999). The focus on g has been predominant in the
literature and has proved to be one of the most controversial issues in psychology
with respect to race (Deary, 2000).
As Rushton and Jensen (2005) concede, researchers have challenged the
derivation of g as being a statistical artifact based on factor analysis. Even
Spearman’s (1927) early work notes the limitations of g as “a hypothetical and
purely quantitative factor” (p. 5). Rushton and Jensen cite Spearman’s hypothesis
indicating that racial group differences would be largest on g-loaded measures.
Though they note that particular tasks are more “g saturated” than others, the
discussion alludes to g as a unitary construct in relation to various measures of
aptitude. This is clearly not the case, given that standardized IQ tests measure
multiple abilities and therefore have differential loadings in relationship to g.
In response to accusations that g is a statistical “artifact” of factor (or
principal-components) analysis, others have noted that “it need not occur. If, in
fact, there were mental abilities that were independent of others they would be
uncorrelated and they would not load on g” (Hayman, 1998, p. 9). While this is
true in theory, in practice, new IQ tests that do not correlate with popular
measures currently in existence are considered to be problematic in terms of
validity. It is clear that among the “best sellers” in the testing domain, the way to
validate a new test is by correlating it with other well-established cognitive
instruments (Valencia & Suzuki, 2001). Based on this practice, it is unlikely that
a measure unrelated to g will emerge as a winner in current practice. Thus, it is
no wonder that the intelligence hierarchy for different racial/ethnic groups re-
mains consistent across different measures. The tests are highly correlated among
each other and are similar in item structure and format. In addition, many
predictive validity studies note correlations among IQ, level of education, income,
and socioeconomic status. As noted by White (2000), “these are anything but
independent variables; they are criteria for one another” (p. 40).
321GENETICS VERSUS CULTURE ONLY
Full scale IQ (FSIQ), g, and racial/ethnic group differences.
the score that is cited as the basis for the racial/ethnic group hierarchy in
intelligence, with a mean of 100 and standard deviation of 15 (e.g., Wechsler,
1997). As the literature indicates, the FSIQ is not a pure indicator of g. Subtest g
loadings for different subtests have been found to vary by racial group. Thus, the
order of magnitude for g loadings for Blacks and Whites can be “considerably
unique” (Kaufman, 1990, p. 254). Some may argue that this is unimportant,
because regardless of whether a test is a pure measure of g, it can still measure
something meaningful. Yet given that tests measure more than just g, the psy-
chometric definition of intelligence may be challenged and performance on
intelligence tests may be more malleable than assumed in past theories.
The FSIQ is
Genetics and Heritability
In support of their genetic arguments, Rushton and Jensen (2005) cite re-
search documenting results of twin and sibling studies, anatomical differences
(e.g., brain size, brain metabolism), processing speed differences, as well as other
factors that differentiate between racial groups. However, their either-or method
of scoring the evidence between the genetic versus culture-only data implies a
misleading dichotomy (Deary, 2000). There are clear interactions among genetic
factors, anatomical structures, culture, and environment. The importance of par-
ticular interactions may vary depending on an individual’s circumstances and not
their racial group membership.
The genetic explanation for the racial/ethnic hierarchy of intelligence is also
based largely on estimates of heritability. Heritability estimates are based on
correlations of traits between biologically related individuals (Lewontin, Rose, &
Kamin, 1984). Most often, correlations are derived from twin and adoption
studies. These are limited given that relatives resemble one another because they
share genetic traits and live in similar environments. In addition, research on
heritability estimates for minority populations is limited, given small sample sizes
and geographic regionalism (Suzuki & Valencia, 1997). Thus, the complexities of
the culture and genetic interactions make teasing apart the individual contribu-
tions of each difficult, if not impossible.
Over the years, culture has been assigned various definitions. The complex-
ities and ambiguities of the definition of culture are extensive and incorporate
multiple levels of meaning across generations (Geertz, 1973).
According to Rushton and Jensen (2005), there are four data sources that are
believed to remove the cultural component in support of the genetic argument.
These include neurological studies (e.g., reaction time), physiological studies
(e.g., anatomical), inheritance studies, and adoption studies. Limitations in these
research bases from a cultural–environmental perspective have also been noted in
the literature (Hayman, 1998) but are not mentioned by Rushton and Jensen. In
particular, the major assumption that differences in culture do not affect these
supposedly culture-free measures is questionable. Physiological measures in this
case are being used to approximate psychological variables (i.e., intelligence).
Evidence supports that culture affects nearly all psychological phenomena; there-
322SUZUKI AND ARONSON
fore, it is entirely possible that biological indicators of intelligence are also
Although Rushton and Jensen (2005) adhere to a biological definition of race,
other theorists such as Loury (2001) have emphasized the important social
underpinnings of this construct. In this view, although race refers to physical
characteristics, the emphasis is placed on the social meanings or interpretations of
these features made in society.
If race is, therefore, as much a social category as a biological one, then it
would follow that race differences in intellectual performance are not simply
mediated by genetics to the exclusion of cultural and environmental factors. The
reality is, “under the skin, there is very little order to real human genetic variation”
(Cohen, 2002, p. 211). Cohen (2002) noted that of the 15,000 to 20,000 gene pairs
that exist, only 6, or 0.03%, are linked to skin color. In addition, it should be noted
that skin color and other phenotypic markers are only grossly related to race
(Cohen, 2002). Therefore, the associations made by Rushton and Jensen (2005)
between race and IQ are questionable.
A related issue with respect to racial group differences in intelligence has
been the consistent finding that the variance within racial groups is much greater
than that found between racial groups (Valencia & Suzuki, 2001). “Average group
differences in g are simply aggregated individual differences in g, so the com-
position of racial group differences and individual differences are of the same
essential nature” (Jensen, 2000, p. 124). This conclusion, however, has been
challenged by Fagan and Holland (2002), whose research suggests that the
“average difference of 15 IQ points between Blacks and Whites is not due to the
same genetic and environmental factors, in the same ratio, that account for
differences among individuals within a racial group in IQ” (p. 382). These results
indicate the need to seek further explanations for intelligence differences and to
look beyond racially aggregated intelligence test data.
Historical, Contextual, and Testing Issues
Ruston and Jensen (2005) acknowledge in a few sentences the contribution of
other theoretical and empirical work supporting an environmental–cultural per-
spective. These include stereotype threat, mediated learning, and the impact of
relative functionalism with respect to particular marginalized groups.
Stereotype threat is defined as anxiety regarding one’s performance in a
particular domain (e.g., intelligence) based on negative stereotypes that exist in
reference to one’s group (e.g., racial/ethnic group; Aronson, 2002; Steele &
Aronson, 1995). This anxiety is not related to the individual’s ability but rather to
the situation in which a negative stereotype (e.g., “Blacks are unintelligent”) may
be confirmed by one’s performance. Evidence for stereotype threat’s effects is
now abundant. Numerous studies show that it can depress the standardized test
performances on a variety of groups for whom stereotypes allege inferior abilities
in some domain (see Aronson, 2002, for a review).
323GENETICS VERSUS CULTURE ONLY
Rushton and Jensen (2005) minimize the stereotype threat evidence, arguing
that it cannot account for cases in which Blacks are in the majority, such as in the
sub-Sahara, where despite outnumbering Whites, Blacks perform less well on IQ
tests. This work demonstrates that Blacks and Whites experience testing situations
differently, often in ways that have a meaningful impact on scores. This effect
does not require numerical minority status. Studies have replicated the stereotype
threat effect even in all-Black colleges (Aronson, 2002), so it is certainly con-
ceivable that sub-Saharan Blacks could be affected. In addition, Rushton and
Jensen ignore the fact that people exist in sociopolitical contexts that have a
profound impact on their experience and worldview. Sub-Saharan Blacks operate
within a context of racism and colonialism that, in turn, creates and shapes
stereotypes. Therefore, when one applies tests constructed by Whites within one
cultural context (i.e., American) and then applies them to Blacks and Whites in
another, the tests do not mysteriously lose their bias. Stereotype threat may
therefore partly explain why any group alleged to be inferior may underperform
groups thought to be superior, regardless of their numerical representation in a
classroom, in a community, or in a country.
Effects of Mediated Learning Experiences
Studies have also indicated that performance on highly g-loaded tasks can be
affected through intervention such as exposure to information and dynamic
assessment procedures. For example, Skuy et al. (2002) indicated that perfor-
mance on a highly g-loaded task (i.e., Raven’s Standard Progressive Matrices
[RSPM]) can be improved significantly through mediated learning experiences.
Skuy et al. concluded that “African students, by virtue of their sociopolitical
history, are especially likely to have been deprived of mediated learning experi-
ence” (Skuy et al., 2002, p. 230). Thus, scores on the RSPM may be “more related
to schooling, literacy, and the cognitive demands imposed by the environment,
and, thus, they may vary more from culture to culture” (Skuy et al., 2002, pp.
230–231). Other studies also indicate that mediated learning interventions were
effective in raising the measured indicators of cognitive ability for Black children
(see Fagan & Holland, 2002; Sternberg et al., 2002).
Rushton and Jensen (2005) also report findings indicating the relatively high
intelligence of Asians in comparison with other racial/ethnic groups. They fail to
mention explanations such as relative functionalism that have been used to
explain the high achievements of Asians in terms of the educational achievement
and the intelligence hierarchy. Relative functionalism suggests that groups will
pursue opportunities for achievement in particular contexts (e.g., academic, social,
vocational) when it is perceived that other avenues to success are closed. Sue and
Okazaki (1990) refuted the notion that Asians are genetically superior to other
racial/ethnic groups. On the contrary, they cited relative functionalism as account-
ing for the high achievement of Asian Americans beyond their measured IQ. This
theory posits that Asian Americans experience opportunities for upward mobility
in educational areas and exclusion from other noneducational pursuits (e.g.,
entertainment, politics) because of social discrimination or limited English lan-
324 SUZUKI AND ARONSON
guage skills. Though relative functionalism has been difficult to test empirically,
anecdotal evidence in terms of the experiences of Asians in the United States
seems to support this explanation. Arguments based on relative functionalism
could also be made with respect to the limited educational achievements of
African Americans due to slavery and historically little access to educational
Test Development Practices
Current test development practices have served to maintain the racial/ethnic
group hierarchy of intelligence test scores. Strategies used to address issues of
cultural bias are limited to expert review panels and various statistical formula-
tions. However, these practices have been criticized on the basis of their concep-
tual limitations (Valencia & Suzuki, 2001).
Sternberg (2000) criticized current methods of establishing test validity. He
noted that intelligence can be represented in terms of a person’s talents and the
abilities that are valued in a particular sociocultural context. To the extent that
one’s behavior is discrepant from that valued by society, these individuals will be
viewed as less successful and intelligent. Sternberg stated, “tests are validated
almost exclusively against the societally approved criteria, giving tests an appear-
ance of validity that they may not have within a given sociocultural group”
(Sternberg, 2000, p. 165). Issues of how one adapts to particular environmental
contexts that may differ from the status quo are not considered.
With respect to specific tests, accusations of “cultural” and “statistical” bias
are still noted for popular tests such as the SAT (Freedle, 2003). Freedle (2003)
contended that a corrective scoring method, the Revised–SAT (R-SAT), be used
to address the “nonrandom ethnic test bias patterns found in the SAT” (p. 1) by
focusing on the “hard” items of the SAT. These hard items are often dependent
on “rare vocabulary” (Freedle, 2003, p. 2). Freedle cited work using differential
item functioning, which reflects a “small” but “highly patterned nature; that is
many easy items show a small but persistent effect of African Americans’
underperformance, while many hard items show their overperformance” (Freedle,
2003, p. 3). Freedle referenced the cultural unfamiliarity hypothesis that “many
easy verbal items tap into a more culturally specific content and therefore are
hypothesized to be perceived differently, depending on one’s particular cultural
and socioeconomic background” (Freedle, 2003, p. 7). Hard items are less
ambiguous given that they are most often used in an academic setting. The R-SAT
has reduced the Black–White test gap by one third. Verbal scores are particularly
affected as Freedle noted that scores on the Verbal R-SAT are increased by as
much as 200 to 300 points for individual minority test takers.
Further challenges to the SAT are noted by Rosner (2003), executive director
of the Princeton Review Foundation. His research on the 1998 version of the SAT
indicates that the percentage of White students answering questions correctly was
higher than the percentage of Black students for all 138 items. Items with higher
percentages of Black students answering correctly in comparison with Whites
were “systematically” rejected during the pretesting phase of the instrument
development (Rosner, 2003).
325 GENETICS VERSUS CULTURE ONLY
It is evident that to reach Rushton and Jensen’s (2005) position on the
meaning of the race differences in test performance, one has to accept a particular
definition of intelligence and believe in the validity of IQ tests to measure it. There
is also growing documentation of the powerful effects of context on intellectual
performance (e.g., stereotype threat) and learning (e.g., mediated learning). Even
admitting the possibility of racially based differences in intelligence, there appears
to be considerable research supporting environmental/cultural justification for
race differences—enough at least to make one question a steadfast belief in a
It appears that the culture versus genetic debate will continue despite the fact
that most would adhere to an interactionist perspective (Reynolds, 2000). As
noted in the beginning of this article, our concerns focus on the implications of the
genetic argument. Where society stands on the malleability of intelligence will
affect the allocation of resources (e.g., affirmation action) and the promotion of
particular methods of intervention (e.g., educational programs like Head Start).
Our commentary has only briefly highlighted the literature with respect to
possible cultural and environmental explanations for the racial/ethnic group
hierarchy on intelligence tests. The theoretical and empirical work appear prom-
ising in this area. In addition, questions may be raised regarding current test
development practices (from item selection to validation). There appears to be
many opportunities to think “outside the box” in our examination of what
constitutes an intelligence measure and how we examine issues of bias (White,
2000). In addition, given growing concerns regarding the usage of intelligence
tests for selection purposes, Jensen (2000) suggested using criteria that go beyond
standardized measures and the inclusion of indicators of past performance (e.g.,
work history). The goal for all of us is to discover “truth” in whatever form it may
take. Reynolds (2000) called for members of the profession to base interpretations
of racial differences on mental tests on empirical data and continually challenge
assumptions about the meaning of these differences.
It appears that many challenges remain in explaining fully the racial/ethnic
group hierarchy of intelligence whether one adheres to the culture-only or genetic
perspective. We believe that the answer resides most likely in the interaction
between the two and that data supporting the malleability of IQ will prevail.
Aronson, J. M. (Ed.). (2002). Improving academic achievement: Impact of psychological
factors in education. San Diego, CA: Academic Press.
Cohen, M. N. (2002). An anthropologist looks at “race” and IQ testing. In J. M. Fish (Ed.),
Race and intelligence: Separating science from myth (pp. 201–224). Mahwah, NJ:
Deary, I. J. (2000). Looking down on human intelligence: From psychometrics to the
brain. New York: Oxford University Press.
Fagan, J. F. (2000). A theory of intelligence as processing: Implications for society.
Psychology, Public Policy, and Law, 6, 168–179.
Fagan, J. F., & Holland, C. R. (2002). Equal opportunity and racial differences in IQ.
Intelligence, 30, 361–387.
326 SUZUKI AND ARONSON
Freedle, R. O. (2003). Correcting the SAT’s ethnic and social-class bias: A method for Download full-text
reestimating SAT scores. Harvard Educational Review, 73, 1–42.
Gardner, H. (1999). Intelligence reframed: Multiple intelligences for the 21st century.
New York: Basic Books.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Hayman, R. L., Jr. (1998). The smart culture: Society, intelligence and law. New York:
New York University Press.
Jensen, A. R. (2000). Testing: The dilemma of group differences. Psychology, Public
Policy, and Law, 6, 121–127.
Kaufman, A. S. (1990). Assessing adolescent and adult intelligence. Boston: Allyn &
Lewontin, R. C., Rose, S., & Kamin, L. (1984). Biology, ideology, and human nature: Not
in our genes. New York: Pantheon Books.
Loury, G. C. (2001). The anatomy of racial inequality. Cambridge, MA: Harvard Uni-
Reynolds, C. R. (2000). Why is psychometric research on bias in mental testing so often
ignored. Psychology, Public Policy, and Law, 6, 144–150.
Rosner, J. (2003, April 14). On White preferences. Nation, 276, 24.
Rushton, J. P., & Jensen, A. R. (2005). Thirty years of research on race differences in
cognitive ability. Psychology, Public Policy, and Law, 11, 235–294.
Skuy, M., Gewer, A., Osrin, Y., Khunou, D., Fridjhon, P., & Rushton, J. P. (2002). Effects
of mediated learning experience on Raven’s matrices scores of African and non-
African university students in South Africa. Intelligence, 30, 221–232.
Spearman, C. (1927). The abilities of man: Their nature and measurement. New York:
Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance
of African-Americans. Journal of Personality and Social Psychology, 69, 797–811.
Sternberg, R. J. (1996). Myths, countermyths, and truths about intelligence. Educational
Researcher, 25, 11–16.
Sternberg, R. J. (2000). Implicit theories of intelligence as exemplar stories of success:
Why intelligence test validity is in the eye of the beholder. Psychology, Public Policy,
and Law, 6, 159–167.
Sternberg, R. J., Grigorenko, E. L., Ngorosho, D., Tantufuye, E., Mbise, A., Nokes, C., et
al. (2002). Assessing intellectual potential in rural Tanzanian school children. Intel-
ligence, 30, 141–162.
Sue, S., & Okazaki, S. (1990). Asian American educational achievements: A phenomenon
in search of an explanation. American Psychologist, 45, 913–920.
Suzuki, L., & Valencia, R. R. (1997). Race-ethnicity and measured intelligence: Educa-
tional implications. American Psychologist 52, 1103–1114.
Valencia, R. R., & Suzuki, L. A. (2001). Intelligence testing and minority students:
Foundations, performance factors and assessment issues. Thousand Oaks, CA: Sage.
Wechsler, D. (1997). Wechsler Adult Intelligence Scale—Third edition. San Antonio, TX:
White, S. H. (2000). Conceptual foundations of IQ testing. Psychology, Public Policy, and
Law, 6, 33–43.
327GENETICS VERSUS CULTURE ONLY