Sei sulla pagina 1di 7

PsychologicalReports, 2005, 97, 891-897.

O Psychological Reports 2005

IMPLICATIONS OF T H E CATTELL-HORN-CARROLL
THEORY O N ETHNIC DIFFERENCES I N IQ '

OLIVER W. EDWARDS AND DAVID P. FULLER

University of Central Florida

Summar?l.-The Cattell-Horn-Carroll theory of intelligence and the Spearman-


Jensen hypothesis were applied to help interpret mean I Q differences among ethnic
groups. The Cattell-Horn-Carroll theory classifies cognitive abilities on three strata,
which differ by magnitude of generality including specific, broad, and general abilities.
The Spearman-Jensen hypothesis advances the belief that mean I Q differences be-
tween African Americans and Euro-Americans occur as a function of the test's g load-
ings. Although I Q differences between African Americans and Euro-Americans occur,
the differences should be larger on tests with higher g loadings and smaller on tests
with lower g loadings. The analysis described in this paper may serve as the catalyst to
encourage all intelligence test publishers to supply test users with data regarding mean
I Q differences among various ethnic groups.

Intelligence tests are used widely in education and are usually adminis-
tered to students. The test scores are important in educational decision-mak-
ing (Sattler, 2001). These tests are designed deliberately to produce score
variance (Wesson, 2000). The generation of a broad range of individual
scores allows scholars to make judgments about the rank distinctions be-
tween and within different groups. Knowledge of between and within group
differences allows scholars to interpret the distribution of test scores that
lead to various decisions including eligibility for placement in special educa-
tion programs.
Previous research data have shown that on average and when unad-
justed for differences in socioeconomic status, Asian Americans' mean score
is approximately three points higher than Euro-Americans, African Ameri-
cans' score is approximately 15 points lower than Euro-Americans, and His-
panic Americans score is somewhere in between the latter two groups (Jen-
sen, 1980, 1998; Herrnstein & Murray, 1994; Bartholomew, 2004). However,
many of the studies on ethnic differences in IQ have been criticized for their
failure to use socioeconomic status across more than one generation when
comparing intelligence test scores of different ethnic groups. Therefore, some
researchers contend that the comparisons are not made using similar groups
(Jencks & Phillips, 1998; Onwuegbuzie & Daley, 2001).
Judgments regarding test selection and administration when differences

'Address correspondence to Oliver W. Edwards, University of Central Florida, De artment of


Child, Family, and Community Sciences, Orlando, FL 12816-1250 or e-mail (oweBward@mail.
ucf.edu).
892 0.W. EDWARDS & D. P. FULLER

in mean IQs are obtained from two statistically sound tests such as the Kauf-
man Assessment Battery for Children-Second Edition (Kaufman & Kaufman,
2004) and the Wechsler Intelligence Scale for Children-Third Edition
(Wechsler, 2004) will influence educational decision-making (Kanaya, Scul-
lin, & Ceci, 2003). These differences may occur as a result of the recency of
the tests' norms, or more importantly given the puproses of this paper, their
theoretical foundations. Thus, researchers should investigate the effect of new
or revised theories of intelligence upon test development, test scores, and
mean 1Q differences between ethnic groups. The purpose of this review is to
analyze the potential effect of the Cattell-Horn-Carroll theory of intelligence,
one of psychology's most recent and comprehensive theories of intelligence,
upon the measurement of mean I Q differences between ethnic groups.
THEORIES OF INTELLIGENCE
The Cattell-Horn-Carroll theory provides the theoretical model for this
empirical review. The theory's historical antecedents can be found in Spear-
man's two-factor theory of intelligence (1927), Cattell and Horn's fluid and
crystallized theory of intelligence (Horn & Cattell, 1996; Horn & Noll,
1997))and Carroll's three-stratum theory of cognitive abilities (1993).
Spearman's g
Spearman used factor analytic techniques to demonstrate the viability of
g as the first factor to emerge when analyzing factor scores on intelligence
tests. Tests that contain substantial loadings of g demand conscious and com-
plex mental effort, often evident in analytical, abstract, and hypothesis-test-
ing tasks (Sattler, 2001). Conversely, tests which require less conscious and
complex mental effort load low on g. Intelligence tests with lower g loadings
emphasize specific s factors such as recognition, recall, speed, visuomotor
abilities, and motor abilities (Sattler, 2001). The s is not a common factor,
but a unique factor specific to each and every measurement variable (Jen-
sen, 1998). Several theorists consider g to be the most parsimonious method
to describe human intelligence and should be used when examining mean
IQ differences between African Americans and Euro-Americans (e.g., Carrol,
1993; Jensen, 1998; Rushton & Jensen, 2005).
Other theorists have used factor analysis to suggest that IQ depends
upon a number of independent factors, not upon a large general factor
(Thurstone, 1938; Gardner, 1983). Although researchers may disagree about
the structure of intelligence, they agree that I Q is a function of a general fac-
tor and reflects multidimensional aspects of intellectual functioning (Sattler,
2001).
Cattell and Horn: Fluid and Crystallized Intelligence
Cattell and Horn (Cattell, 1963; Horn & Cattell, 1966) developed a the-
CATTELL-HORN-CARROLLTHEORY AND ETHNIC DIFFERENCES IN IQ 893

ory of intelligence based on two factors, fluid and crystallized abilities. Fluid
intelligence is measured by tasks requiring inductive, deductive, conjunctive,
and disjunctive reasoning to understand, analyze, and interpret relationships
among stimuli. Crystallized intelligence requires familiarity with the salient
culture through such qualities as vocabulary and general information (Sat-
tler, 2001).
Carroll's Three-stratum Theory of Cognitive Abilities
Carrol's development of a three-stratum theory of intelligence (1993)
was crucial to understanding of the structure of human intellect (Sternberg,
1994; Eysenck, 1998). Carroll's three-stratum theory proposes a hierarchical
model of cognitive abilities which includes the following three levels or stra-
ta: specific, broad, and general ability categories.
The Cattel-Horn-Carroll Theory of Intelligence
McGrew (1997)) a coauthor of the frequently administered Woodcock-
Johnson I11 Tests of Cognitive Abilities (McGrew & Woodcock, 2001) first
proposed the integrated Cattell-Horn-Carroll theory. The theory classifies
cognitive abilities on three strata which differ in generality. Carroll identified
more than 69 specifc abilities, described as Stratum I abilities. Stratum I
abilities correlate positively and apparently do not reflect independent traits
(Carroll, 1993; Flanagan & Ortiz, 2001). Therefore, Stratum I abilities were
regrouped to form approximately nine broad Stratum I1 abilities. The Stra-
tum I1 abilities correlate and thus help form a third-order general factor of
intelligence. This general factor, or g, is at the apex of the hierarchical mod-
el and is termed Stratum I11 (Carroll, 1993).
Extensive factor analytic, neurological, developmental, and heritability
evidence support the Cattell-Horn-Carroll theory (Flanagan & Ortiz, 2001).
In addition, research suggests the theory provides equal explanatory power
across sex and ethnicity (Carroll, 1993; Gustafsson & Balke, 1993; Keith,
1999). Increasingly, new and recently revised tests of intelligence are based
extensively on the Cattell-Horn-Carroll theory (e.g., Kaufman Assessment
Battery for Children-Second Edition, the Wechsler Intelligence Scale for
Children-Third Edition (Wechsler, 2003; Woodcock-Johnson I11 Tests of
Cognitive Abilities; Woodcock, McGrew, & Mather, 200 1). The frequent
use of the Cattell-Horn-Carroll theory in development of intelligence tests
raises important questions regarding the theory's effect on the measurement
of I Q and the interpretation of mean IQ differences between ethnic groups.
The Spearman-Jensen Hypothesis
Spearman's hypothesis proposes that the subgroup differences observed
in mean intelligence test scores occurs as a function of intelligence tests' g
loadings (Jensen, 1998). Spearman suggested that persons of sub-Saharan
894 0. W. EDWARDS & D. P. FULLER

African origin tend to score lower than Euro-Americans on cg-loaded tests


than on tests of more narrow abilities (Jensen, 1998). Jensen's research has
supported and supplemented Spearman's hypothesis. The term Spearman-
Jensen hypothesis is used in this paper to reflect the theory that ethnic dif-
ferences in mean intelligence test scores occur as a function of a test's g load-
ing. Jensen (1998) found that I Q differences between African Americans and
Euro-Americans on mental tests were closely related to the g component in
score variance and do not result from the tests' factor structure, cultural
loading, or test bias. Intelligence tests generally have the same reliability and
validity scores for native, English-speaking ethnic groups (Jensen, 1998).Jen-
sen cites research which suggests the magnitude of a test's g loading predicts
the magnitude of the standardized mean subgroup difference. As a result,
differences between African Americans and Euro-Americans are expected to
be larger on Stratum 111, which is a measure of g, than on Stratum 11, which
measures more specific abilities. If the Spearman-Jensen hypothesis is correct,
I Q differences between African Americans and Euro-Americans should be
less than 15 points on intelligence tests which extensively assess broad Stra-
tum I1 abilities than on those which rely more considerably on general Stra-
tum 111 ability. However, if g explains similar or more of the variance on re-
vised compared to earlier versions of intelligence tests, and mean I Q differ-
ences between ethnic groups are smaller on the revised versions, the Spear-
man-Jensen hypothesis should be re-evaluated.
Interpretive ImpZications of tbe Cattell-Horn-Carroll Tbeoqv
O n most intelligence tests, approximately 50% of the variance of each
subtest is specific to that subtest. As such, its source of variance is partly
comprised by g and is partly separate of g. The technical term for a subtest's
specific variance is specificity. Specificity is defined as the proportion of a
test's true-score variance unaccounted for by a common factor such as g
(Jensen, 1998). The WISC-IV and Woodcock-Johnson I11 Tests of Cogni-
tive Abilities are excellent tests to investigate the effect of the Cattell-Horn-
Carroll theory on interpretations of mean IQ differences between ethnic
groups. These tests are said to use the Cattell-Horn-Carroll theory for their
theoretical underpinnings. Results of studies indicated the factor structure of
the measuring devices was consistent across ethnic groups (McGrew &
Woodcock, 2001; Wechsler, 2003). This finding allows one to test the Spear-
man-Jensen hypothesis and the effect of the Cattell-Horn-Carroll theory on
interpretations of mean IQ differences among ethnic groups with greater
confidence. However, analyses of IQ differences between ethnic groups as a
function of g loadings and subtest specificity were not reported or not per-
formed by the publishers of either of the two tests.
A fundamental issue for tests developed using the Cattell-Horn-Carroll
CATTELL-HORN-CARROLL THEORY AND ETHNIC DIFFERENCES IN I Q 895

theory is the potential loss in the predictive validity of I Q as a proxy for g.


Jensen and others have demonstrated repeatedly that a variety of real world
outcomes can be predicted by the magnitude of a test's loadings on g. When
tests are constructed utilizing theories that affect the mechanics of formulat-
ing I Q so that they reflect more specific factors and less g, the result may be
a loss in predictive power. Therefore, the less a test loads on g, the less the
I Q obtained is predictive of later overall academic achievement in reading,
arithmetic, and writing as well as attainment in any number of areas (Jen-
sen, 1998). The aforementioned notwithstanding, when properly designed,
tests developed using the Cattell-Horn-Carroll theory should explain a sig-
nificant portion of variance in different skills (McCrew, 1997). Since each
subtest that comprises the test may have substantial specificity, the test
should function effectively in detecting strengths and deficits of specific abil-
ity. For example, tests based on the Cattell-Horn-Carroll theory may im-
prove intelligence tests' capacity to identify various skills related to reading.
As such, tests based on the Cattell-Horn-Carroll theory should attenuate the
need for cross-battery assessment (see Flanagan & Ortiz, 2001 for a descrip-
tion of cross-battery assessment).
A number of popular intelligence tests, e.g., Stanford-Binet Fifth Edi-
tion (Roid, 2003). WISC-IV, and Woodcock-Johnson I11 Tests of Cognitive
Abilities) currently do not provide information about g loadings, subtest
specificity, and mean I Q differences by ethnic groups in their manuals. These
data cannot be considered when interpreting test scores if they are not pro-
vided. Test developers are encouraged to make standardization data avail-
able to allow such types of analyses. Researchers could then investigate po-
tential reductions in mean IQ differences between ethnic groups whenever
new or renormed intelligence tests are published. Results of these investiga-
tions would assist psychologists in ascertaining the fairness and social conse-
quences of intelligence test scores for use with various ethnic groups.
CONCLUSION
Mean I Q differences between groups may influence the proportion of
students enrolled in special education programs merely as a function of the
intelligence test administered (Kanaya, et al., 2003). For example, an Afri-
can-American child who obtains a score of 69 on one intelligence test may
achieve a score of 74 on another as a function of the test's g loadings. A
score of 69 has greater potential to lead to placement in a program for stu-
dents with mental retardation than a score of 74. Practices such as this con-
tribute greatly to the increasingly disproportionate numbers of ethnically di-
verse students in special education (Kanaya, et al., 2003). Using data to un-
derstand better the nature of this dilemma is the first step in identifying so-
lutions to improving educational services for ethnically diverse students.
896 0. W. EDWARDS & D. P. FULLER

Although disability labels may benefit students as the first step in ob-
taining services, there are potential negative consequences as well. Many be-
lieve that labels suggest something is "wrong" with the student so teachers
have lower expectations which may significantly impede development or ac-
quisition of new skills. Of course, children who experience significant diffi-
culty learning without special education support should receive such support
(Donovan & Cross, 2002). However, a cost benefit analysis of such assistance
should be undertaken.
The consequences of differences in I Q among ethnic groups are of sub-
stantial importance. In light of the repeated use of intelligence tests with stu-
dents, this paper may serve as the catalyst to encourage all publishers of in-
telligence tests to supply test users with data regarding tests' g loadings, sub-
test specificity, and mean IQs for various ethnic groups so differences may
be evaluated. Such data allow psychologists to make informed decisions
about which tests will provide the most meaningful information to help ex-
aminees. Data-based decision-making is essential for effective psychology.
Political correctness should not subjugate scholarly precision.
REFERENCES
BARTHOLOMEW, D. J. (2004) Measuring bztelligence: facts and fallacies. New York: Cambridge
Univer. Press.
CARROLL, J. B. (1993) Human cognitive abilities: a survey of factor-analytic studies. New York:
Cambridge Univer. Press.
CATTELL,R. B. (1963) Theory of fluid and crystallized intelligence: a critical experiment. Jour-
nal of Educational Psj~cholo~y, 54, 1-22.
DONOVAN, M. S., & CROSS,C. T. (2002) Minority students in special and gifted education:
committee on minority representation in special education. Washington, DC: National
Academy of Education.
EYSENCK,H. J. (1998) A new look at intelligence. New Brunswick, NJ: Transaction Books.
FLANAGAN, D. I?, &ORTIZ,S. (2001) Essentials of cross-batteql assessment. New York: Wiley.
GARDNER, H. (1983) Frames of mind: the theoq of multiple intelligences. New York: Basic
Books.
GUSTAFSSON, J. E., & BALKE,G. (1993) General and specific abilities as predictors of school
achievement. Multivariate Behavioral Research, 28, 407-434.
HERRNSTEIN, R. J., &MURRAY, C. (1994) The bell curve: intelligence and class structure in Amer-
ican lge. New York: Free Press.
HORN,J. L., & CATTELL, R. B. (1966) Refinement and test of the theory of fluid and crystal-
lized general intelligences. Journal ofEducattbna1 Psychology, 57, 253-270.
HORN,J. L., &NOLL,J. (1997) Human cognitive capabilities: Gf-Gc theory. In D. P. Flanagan,
J. L. Genshaft, & P. L. Harrison (Eds.), Contemporary intellectual assessment: theories,
tests, and issues. New York: Guilford. Pp. 53-91.
JENCKS,C., & PHILLIPS, M. (Eds.) (1998) The black-white test score gap. Washington, DC:
Brookings Institution.
JENSEN,A. R. (1980) Bias in mental testing. New York: Free Press.
JENSEN,A. R. (1998) The g factor: the science of mental ability. Westport, CT: Praeger.
KANAYA,T., SCULLIN, M. H., & CECI,S. J. (2003) The rise and fall of IQ in special ed: historical
trends and their implications. Journal of School Psychologj~,41, 453-465.
KAUFMAN,A. S., & KAUFMAN, N. L. (2004) Kaufman Assessment Batteq for Children. (2nd ed.)
Circle Pines, MN: AGS.
CATTELL-HORN-CARROLL THEORY AND ETHNIC DIFFERENCES IN IQ 897
KEITH,T. Z. (1999) Effects of general and specific abilities on student achievement: similarities
and differences across ethnic groups. School Psychology Quarterly, 14, 239-262.
MCGREW,K. S. (1997) Analysis of the major intelligence batteries according to a proposed
comprehensive Gf-Gc framework. In D. P. Flanagan, J. L. Genshaft, & P. L. Harrison
(Eds.), Contemporary intellectual assessment: theories, tests, and issues. New York: Guil-
ford. Pp. 151-180.
MCGREW,K.S., &WOODCOCK, R. W. (2001) Technical manual: Woodcock-Johnson 111. Itasca,
IL: averside.
ONWUEGBUZIE, A. J., & DALEY,C. E. (2001) Racial differences in IQ revisited: a synthesis of
nearly a century of research. Journal of Black Psychology, 27, 209-220.
ROID,G. H. (2003) Stanford-Binet Intelligence Scales. (5th ed.) Itasca, IL: Riverside.
RUSHTON, J. P., & JENSEN, A. R. (2005) Thirty years of research on race differences in cognitive
ability. Psychology, Public Policy, and Law, 11, 235-296.
SATTLER,J. M. (2001) Assessment of children: cognitive applications. (4th ed.) San Diego, CA:
Author.
SPEARMAN, C. E. (1927) The abilities of man. New York: Macmillan.
STERNBERG, R. J. (1994) A triarchic model for teaching and assessing students in general psy-
chology. General Psychologist, 30, 42-48.
THURSTONE, L. L. (1938) Primary mental abilities. Chicago, IL: Univer. of Chicago.
WECHSLER, D. (2003) The Wcchslcr Intelligence Scale for Children-Third Edition. New York:
The Psychological Corp.
WESSON,K.A. (2000) The Volvo effect-questioning standardized tests. Education Week, 20,
34-36.
WOODCOCK, R. W., MCGREW,K. S., &MATHER, N. (2001) Woodcock-Johnson 1I1 Tests of Cogni-
tive Abilities. Itasca, IL: Riverside.

Accepted December 1, 2005

Potrebbero piacerti anche