We are grateful to Professor Rushton for providing the following
crucial document which was scanned but not completely corrected
for scanning errors. To be sure to avoid any scanning
errors, please get a copy of the original document. Please
contact the webmaster at [email protected] for any
questions or to suggest corrections. This html page is
provided for reference only.

horizontal rule

As an example of the kind of bias and prejudice directed against honorable men like Professor Rushton, and how cowardly his critics are, please note the following description of what "free speech" now means in the "free world" from his closing thoughts in "Race, Evolution, and Behavior":

Q: Why did the Charles Darwin Research Institute publish this Y2K version of the abridged edition? What happened to the original publisher?

A: Transaction Publishers printed 100,000 copies under their copyright. They sent 35,000 to scholars around the world members of the American Anthropological Association, the American Psychological Association, the American Sociological Association, and the American Society for Criminology. Then the Progressive Sociologists, a self-proclaimed radical group within the American Sociological Association, along with some other "anti-racist" groups, threatened Transaction with loss of a booth at its annual meetings, advertising space in journals, and access to mailing lists if they continued to send out the abridged edition. Transaction caved in to this pressure, withdrew from publishing the abridged edition, and even apologized. They claimed that the Transaction copyright should never have appeared on the book and that it had "all been a mistake."

These events sadly confirm what I wrote in the first abridged edition that some vocal groups in academia and the media forbid an open discussion of race. They fear any open discussion of race research, all of which has appeared in peer-reviewed scientific journals. Truth, however, always wins out in the long run.

Closing Thoughts

The information in this book shows that the races differ in important ways. They differ, on average, in brain size, intelligence, sexual behavior, fertility, personality, maturation, life span, crime and in family stability. Orientals fall at one end of the three-way pattern of differences. Blacks fall at the other end, and Whites usually fall in between. Only a theory that looks at both genes and environment in terms of Darwin's theory of evolution can explain why the races differ so consistently throughout the world and over the course of time.

Both science and justice call for us to seek and tell the truth, not to tell lies and spread error. While the research in this book first appeared in peer-reviewed academic journals, many in the media, the government, and unfortunately even in the universities and colleges, skillfully avoid all such evidence. Hopefully this abridged edition will help set the record straight and make the latest scientific findings on race, evolution, and behavior open to all.

If we want to understand human behavior, the social sciences must get back together with the biological sciences. This book is a step in that direction. When we look at both genes and environment we may be able to understand human problems. With that knowledge, society can then go about trying to solve them. The first step is for all of us to be as honest as we can be about race, evolution, and behavior.

horizontal rule

INTELLIGENCE 28(4): 251-265

Copyright � 2000 by Elsevier Science Inc.
All rights of reproduction in any form reserved.

Direct all correspondence to: J.P. Rushton, Department of Psychology, University of Western Ontario, London,
Ontario, Canada NAA 5C2. E-mail: [email protected]

Performance on Raven's Matrices
by African and White University
Students in South Africa


University of Western Ontario, London, Ontario, Canada


University of the Witwatersrand. Johannesburg, South Africa

Untimed Raven's Standard Progressive Matrices (SPM) were administered to 309 17- to 23-year-old
students at the University of the Witwatersrand and the Rand Afrikaans University in Johannesburg,
South Africa (173 Africans, 136 Whites; 205 women, 104 men). African students solved an average
of 44 of ttie 60 problems whereas White students solved an average of 54 of the problems
(p< 0.001). By the standards of the i993 US normative sample, the African university students
scored at the 14th perccntilc and the White university students scored at the 61st percentile (IQ
equivalents of S4 and 104, respectively). The African White differences were found to be greater
on those items of the SPM with the highest item total correlations, indicating a difference in g, or
the general factor of intelligence. A small sex difference favoring males was found in both the
African and the White samples, but unrelated to g.

For many psychologists, South Africa's transition to majority rule in 1994 and the
concomitant dismantling of the "apartheid" system of "separate development" raised
questions about whether Euro-American test norms were equally suitable for the nation's
recognized population groups, namely "Africans," "Coloreds" (mixed-race), "Asians"
(Indians), and "Whites." (In South Africa today, the term "Black" designates alt those
other than Whites, including Indians, Coloreds, and Africans; the term "African" is used
to describe the indigenous habitants who comprise over 80% of the population). The
discussion of test validity centered the on pragmatic problems of assessing cognitive
impairment in African adults and children following motor vehicle accidents, selecting
disadvantaged students for university admission, establishing baselines against which to
evaluate interventions, and examining why the various groups differed in test performance
(Hartshome, 1992; Owen, 1998; Skuy, Zolezzi, Mentis, Fridjbon, & Cockcroft, 1996).
Lower mean test scores are routinely obtained in African samples relative to Euro-
American test norms. For example, Skuy, Schutte, Fridjohn, and O'Carroll (1999) found
scores of 1-2 standard deviations (SD) below American norms in 154 African secondary
school students from Johannesburg on a variety of tests including the Wechsler
Intelligence Scale for Children-Revised (WISC-R), the Rey Auditory Verbal Learning
Test, the Stroop Color Word Test, the Wisconsin Card Sorting Test, the Bender Gestalt
Visual Motor Integration Test, the Rey Osterreith Complex Figure Test, the Trail Making
Test, the Spatial Memory Task, and various Drawing Tasks. Thus, on the WISC-R, the
African students were -1.81 SD below American norms (-1.58 SD with the vocabulary
sub-test excluded).

The question of African test performance came to attention in the US when The Bell
Curve (Hermstein & Murray, 1994, pp. 288-289) examined an often stated hypothesis:

"The test scores of American blacks have been depressed by the experience of slavery and
African blacks will be found to do better (p. 565)." However, black Africans turned out to
be, on average, substantially below black Americans in intelligence test scores. The Bell
Curve cited Richard Lynn's (1991) review of 11 studies from East, West, and Southern
Africa reporting an average IQ of 70 (median = 75), 15 points (1 SD) lower than the mean
of 85 typically found for black Americans and 30 points (2 SD) lower than the mean of
100 typically found for Whites. The tests used included the Standard Progressive Matrices
(SPM), the Colored Progressive Matrices (CPM, a simpler version of the SPM), the Army
Beta, the Junior Aptitude, and the Culture Fair.

Lynn (1978) had earlier summarized seven other African studies, mainly on pupils
using Raven's Matrices, and found average IQ equivalents ranging from 75 to 88 with a
mean of 82. Lynn noted the difficulties of obtaining representative samples as well as
accurate information on ages, both necessities for valid group comparisons. Despite
inadequacies in many samples, Lynn reported the results are consistent. For example,
scores do not vary when samples of Africans are selected in ways that will tend to bias the
results upward-by limiting the sample to people who have completed primary school
(many of the least academically able having dropped out), people who are employed, or
people who live in urban areas.

Subsequent studies (some with quite large TVs) have corroborated the low mean test
scores of Africans Lynn reported. In South Africa, Owen (1992) gave the SPMs without
time limits to 1,093 African, 778 Colored, 1,063 Indian, and 1,056 White 14-year-olds.
Except for the Indians, subjects were tested by school psychologists of their own ethnic
group. Owen (1992) presents the full psychometric profile for the test results (distribu-
tional characteristics, reliability, item difficulty, item discrimination, congruence coeffi-
cients, and discriminant analysis), showing that the test did measure the same construct in
each of the various ethnic groups. He reported the differences in test means, expressed in
SD units, as follows: White-Indian: -.52; White-Colored: -1.35; White-African:

-2.78. Converting Owen's SD differences into IQ equivalents gives Africans an average
IQ of 58 in relation to a mean for Whites of 100, and of 80 for Coloreds. A higher mean
African IQ of 72 results if one uses the percentile equivalents from the SPM standardiza-
tion data of British Whites instead of making the calculations using the noticeably small
SD of the White South African sample in Owen's study-

Zindi (1994), a Zimbabwean, matched 204 black Zimbabwean and 202 white English
pupils from London inner-city schools for age (12-14 years old), sex, and educational
level, both samples being characterized as "working class." Despite the fact that the white
sample was below average for the Whites, with a mean IQ measured by the WISC-R of 95,
the African-White difference was 1.07 SD on the SPM and 2.36 SD on the WISC-R.
Zindi (1994) expressed the SPM results as TQ scores. The means for the Zimbabwean
sample were 72 for the SPM and 67 for the WISC-R. The WISC-R score was depressed by
language considerations, but not by much since the (nonverbal) performance IQ score of
the Zimbabwean sample was 70.

Lynn (1997) reviewed five additional studies of African IQ scores published between
1985 and 1996. Mean IQs were in the range of 60-74. One study reported the results from
a random sample of 1,639 adolescents in Ghana drawn from the entire country (Glewwe &
Jacoby, 1992). Their mean age was 15.2 and their mean score on the CPM (the simpler
version of the SPM) was 12.5, equivalent in British samples to an IQ of 60.

African-White differences are also found on simple reaction time (RT) measures. In
one of these (the "odd-man-out" test), 9- to 12-year-old children are asked to decide
which of several lights stands out from the others, and then press the button that
corresponds to that light. The test is so easy that all children can perform it in less than
1 sec. But even on this very simple test, children with higher IQ scores perform faster than
do children with lower IQ scores, Lynn (1991) found that Black children from South
Africa average slower RTs than do White children from Britain and Ireland. Earlier,
Poortinga (1971) had also shown African-White differences in South Africa on four- and
eight-choice RT tasks for both auditory and visual stimuli. The magnitude of the mean
African-White differences on these RT measures ranged from 1.26 to 1.53 SD (see
Jensen, 1998,, p. 392, for discussion of this study).

An exception to the pattern of low African test scores is a study by Crawford-Nutt
(1976) who found that 228 African high school students from Johannesburg had a mean
score equal to that for Whites. Crawford-Nutt (1976) used a special demonstration
apparatus to administer the SPM to ensure complete understanding of the test
requirements. The mean score for the African pupils was the same as the mean for
the Raven's normative group. The author concluded that "[T]he frequently encountered
poor performance of Blacks on tests of ability could be simply an artifact of the method
of administering the test" (p. 205). Unfortunately, the author did not use a control
group, which did not receive the special instructions. Whether the performance of
Craw ford-Null's testees is attributable to the special test instructions or to the fact that
this particular high school attracts only the top students from feeder schools is not clear.

Only one previous study has reported mental test scores for African post-secondary
students. Poortinga (1971) used the Advanced Progressive Matrices (APM, the more
difficult version of the SPM) at a select college in South Africa. He found an IQ equivalent
of 92. Although this score is 1.5 SD above the mean for Africans as a whole, it is also 2.2
SD below the mean for the white South African college students also tested in that study.
Poortinga (1971) stated that the APM was too difficult for the African students. (On two
other psychometric tests, the groups differed by 2.3 and 1.5 SD).

One South African study found African-While differences were most pronounced
on g, the general factor of intelligence. Lynn and Owen (1994) examined scores of
several thousand South African secondary school students on the 10 sub-tests of the
Junior Aptitude Test (four verbal, six non-verbal). They found an overall African-
White difference of 2 SD, with variation among the subtests correlating 0.62 (p<0.05)
with the g factor as extracted from the African sample (although only 0.23 with g
extracted from the White sample). This finer-grained observation is similar to those
made in the US, where Black-While differences are typically greater on tests with
higher ^-loadings (Jensen, 1998).

The male-female difference in mean IQ is also a topic of current debate. Lynn (1999),
in particular, has marshaled much evidence that supports a 2- to 5-point mean IQ
difference favoring males, including on the SPM (although an earlier review by Court,
1983, failed to find a consistent sex difference on the Raven's). Lynn's results, however,
have been challenged by Jensen (1998, pp. 536-542). The present study, therefore,
examines the sex difference among both African and White university samples.


The primary purpose was to examine performance on the Raven's SPM in a sample likely
to score at least 1 SD above the general South African population mean. First-year
psychology students from the University of the Witwatersrand (Wits) and the Rand
Afrikaans University (RAU) in Johannesburg were invited to take part in an "international
study" of the widely used SPM for which participants would be paid 50 rand (about
US$10). At Wits, students of all races were invited to participate; at RAU only Africans
were invited. Both universities require students to have graduated from high school
(although lower pass scores may be accepted for disadvantaged students). To help obtain
the data, me first author visited the University of the Witwatersrand in October of 1998-

Wits is a pre-eminent English-speaking research institution with a reputation for
political liberalism, while RAU is an Afrikaans-speaking institution (Afrikaans being the
Dutch-derived language of South Africa). Wits' policy since its inception (in 1922) has
been not to discriminate on racial (or any other) grounds and Africans have always been
at least a small part of the student body. African student enrollment at Wits was 8
percent in 1986, 13 percent in 1990, 23 percent in 1994, and 35 percent in 1997. The
figures are higher for "Blacks" (i.e., all non-Whites grouped together): e.g., 36 percent
in 1994 and 48 percent in 1997. When Africans enrol at RAU, their language of
instruction is English.


Historically in South Africa, Blacks and Whites have differed markedly in culture,
language, politics, and history as well as average educational and socio-economic status.
Only recently has there been an end to a century plus of discriminatory policies, including
apartheid. Nonetheless, Black South Africans have achieved higher literacy rates than is
typical elsewhere on the African continent and many South African Black university
students are from middle-class homes. By 1986, the literacy rate reached about 60 percent
with up to 80 percent of all Black children of school age (about 6 million) being in school.
By 1993, Black children spent an average of 11 years enroled in school although the
average attainment of all those leaving school was only 9th grade (Hartshome, 1992;

Macro-Economic Work Group, 1993). Currently, about 85 percent of South African
children attend primary school, m South African schools, standardized examinations are
administered and a requirement for graduating from secondary school (mainly essays, with
some multiple choice questions).

At the university level, until the late 1970s, African, Colored and Indian students
could enrol only at those universities established for them, except for very small numbers
at English language universities such as Wits. Most African students were enroled at one
of five predominantly African universities such as Fort Hare and the University of
Zululand, as well as the University of South Africa (UNISA), a teletuition university
originating in Pretoria with students all over the world. Since South Africa's transition to
majority rule in 1994, Blacks have been entering traditionally White universities in
record numbers.


To facilitate between-group comparisons, an initial pool of 392 subjects was reduced to
309 by eliminating ambiguous categories. There were 173 Africans (49 men, 124 women)
and 136 Whites (55 men, 81 women) aged 17 to 23. Excluded were those who self-
identified as "Indians" (n - 30), "Coloreds" (n - 10), "Other" (n - 3), those who listed
their age as over 23 (n = 33), who failed to give biographical data, who placed their
answers in an inappropriate place, or who left parts of the answer sheet blank (n = 7).

Test Instruments

Raven's SPM is probably the most well-known, most researched, and most widely used of
all culture-reduced tests. Its popularity is evident from the fact that it has been used in well
over 1,000 studies (Raven, Court, & Raven, 1996). As an untimed "capacity" test, and
even as a 20-min "speed" or "efficiency" test, the results have been found to demonstrate
reliability and validity across a wide range of populations. Retest reliabilities ofO.S3-0.93
are found with an interval of approximately 1 year between administrations. Internal
consistency coefficients of 0.80 are found across many cultural groups, including South
African Blacks (Owen, 1992).

The SPM is usually regarded as a good measure of the non-verbal component of
general intelligence not bound by culturally specific information. It was designed to
measure Spearman's (1927) g, the general factor of intelligence, or at least the non-verbal
component thereof. It is also described as a measure of "the ability to identify relation-
ships," "analogical thinking," and the ability to "think clearly" (Raven et al., 1996, SPM
1). It consists of 60 diagrammatic puzzles, each with a missing part that the test taker
attempts to identify from eight options. The 60 puzzles are divided into five sets (A, B, C,
D, and E) of 12 items each. In each set, the first problem is as nearly self-evident as
possible. The problems which follow build on the same reasoning as those thai have gone
before and provide opportunities to grasp the method of thought required to solve the
problems, which become progressively more difficult. To ensure sustained interest and
freedom from fatigue, each problem is boldly presented, accurately drawn, and, as far as
possible, pleasing to look at. No time limit is set and all testees are allowed to complete
the test.


Testing was conducted by both of the authors and five MA research assistants, two of
whom were African, in large examination halls with desks spaced well apart to prevent
copying from others. To ensure the diligence with which participants approached their
tasks, the instructions requested students to wait quietly at their desks if they finished
before 30 min. After 30 min, however, they could come to the front of the room, hand in
their answer sheets and test booklets, and receive payment. A handful of students (all
African) took the full time available. The SPM was administered without any time limits
(up to 1.5 h), but was typically completed within 30 min.

Means, SD, and Internal Consistencies

All calculations are based on raw scores of the SPM, with each of the 60 items scored as 0
(incorrect) and 1 (correct). Internal consistencies based on Cronbach's alpha were .83 for
White males, .73 for White females, .89 for African males, and .92 for African females.
Table 1 shows the means (M) and SD of the various groups for each of the five sub-tests
and for the full test. Fig. 1 shows the percentage of Africans and of Whites who attained
various raw scores. The longer tail of low scores in the African distribution and the ceiling
effect for both groups are clearly visible. The secondary peak of high scores for the
Africans suggests a possible bimodal distribution.

Analysis of variance (ANOVA) with Race and Sex as factors showed significant main
effects and a marginally significant interaction, ^(1,305) = 131.85, p< 0.001; .F(l,305) =
8.89, /?<0.01; and F(l,305) = 3.67, p<0.10. Whites averaged higher scores than
Africans (unweighted means -= 54, 44; SD ~ 5, 9; Ranges - 42-60, 10-59, respectively);

men averaged higher scores than women (unweighted means = 50, 47; SD = 7, 9; Ranges
= 29-60, 10-60, respectively), and the sex difference was marginally greater among
Africans than among Whites (Means 46, 42 vs. 54, 53). Expressed in terms of the White
SD, the African-White differences for sets A to E were: 0.83, 2.01, 2.21, 1.36, and 2.07,
respectively. For the total score the African-White difference was 1.2 SD (based on Black
SD), and 2.6 SD (based on the White SD).

The 1993 US norms for 18- to 22-year-olds show that the White men, with 54 out of
60 correct responses, average at the 61st percentile; that the White women, with 53 correct
responses, average at the 55th percentile; that the African men, with 46 correct responses,
average at the 19th percentile; and the African women with 42 correct responses average at
the llth percentile (Raven et al., 1996, p. 65, Table SPM 12). These SPM grades and
percentile points convert to TQ equivalents of 105 for White men, 102 for White women,
87 for African men, and 82 for African women (Raven et al., 1990, p. 98).

Item Difficulty Values (p)

Table 2 shows the proportion of each of the samples, which selected the correct answer on
each of the 60 items. For all groups, set E was the most difficult, followed by sets C and D
while sets A and B were the easiest. Across the 60 items, the difficulties were virtually
identical for Africans and for Whites (Pearson's r = 0.88; p<0.00\; Spearman's rho =

Most of the SPM items were too easy for these university students. From Table 2, it
is apparent that relatively few of the 60 items display /^-values (proportion passing)
within me optimal range of 0.30-0.70, which provides maximum discriminatory power;

horizontal rule

Table 1. Mean and SD of Raw Scores of the Standard Progressive Matrices (and Subsets of Items) by Race and Sex
A'-weighted N-weighted
Male Female Sex Combined Male Female Sex Combined
	(N=	49)	<N^	124)	(N-	173)	(N=-	55)	</V-	8!)	(N-	136)
A (Items 1-12)	11.37	1.11	10.94	1.49	11.06	1.40	11.62	1.16	11.82	0.45	11.74	0.82
B (Items 13-24)	10.86	1.57	9.RO	2.48	10,10	2.30	11.60	0,66	11.52	0.76	11.55	0.72
C (Items 25-36)	9.43	1.85	8.04	2.39	8.43	2,33	11.16	0.92	10.72	1.21	10,90	1.12
D (Items 37-48)	9.67	1.70	8.98	2.38	9.17	2.23	10.62	1.16	10.56	0.94	10.58	1.03
E (Items 49-60)	4.98	3.14	4.40	2.64	4.56	2.79	9.44	2.57	8.93	1.92	9.13	2.21
Total	46.31	7.19	42.15	9.11	43.32	8.79	54.44	4.57	53.33	3.76	53,90	4.11

horizontal rule

horizontal rule

Figure 1. Percentage of African and White first-year psychology students in South Africa attaining
various scores on the SPM.

horizontal rule

there are only six such items for the White and 12 for the African testees. Using a
proportion of 70 percent of respondents passing as the criterion for judging an item as
"too easy," 54 of the 60 items (90%) proved as too easy for Whites and 41 of the 60
items (68%) too easy for Africans. None of the items was found to be "extremely
difficult" (p <, .10) by the Whites but items C12, E10, Ell, and El 2 were found to be
so for the Africans. Overall, Africans found the items more difficult than did the Whites,
as did women compared to men.

Item Discrimination Values (r;t)

Another index for comparing items across groups is the item-total correlation (/•„). This is
the correlation of each item with the total score on the test. It indicates the extent to which
a particular item measures the construct that is measured by the test as a whole, as well as

horizontal rule

Table 2. Proportion of Sample Selecting the Correct Answer on Items of the Standard Progressive Matrices by Race
	Scf/)			Set B			SetC			Set D	SetE	U/0 C3:
Ilem	African	White	item	African	White	Item	African	White	Item	African	White	Item	African	White	
1	1.00	1-00	13	0.99	0.99	25	0,94	1.00	37	0.96	1.00	49	0.77	0.89	>-n
2	1.00	1.00	14	0.98	1.00	26	0.91	1.00	38	0.94	0.99	50	0.68	0,98	m
3	0.99	1.00	15	0.96	1.00	27	0.92	1.00	39	0.93	0.98	51	0.64	0.93	0>
4	0.98	1.00	16	0.90	1.00	28	0.86	0.93	40	0.89	0.97	52	0.47	0.91	Z
5	0.97	0.98	17	0.90	0.99	29	0.88	i.OO	41	0.92	1.00	53	0.50	0.96	-1m
6	0.98	0.99	18	0.86	0.99	30	0.73	0.94	42	0.87	0.99	54	0.39	0.87	w
7	0.94	0.99	19	0.74	0.90	31	0.86	0.99	43	0.79	0-94	55	0.38	0.76	v>
8	0.90	0.98	20	0,77	0.91	32	0.46	0.87	44	0.80	0.95	56	0.23	0.79	a
9	0.96	0,99	21	0.83	0.97	33	0.72	0.87	45	0.79	0.92	57	0.25	0.79	03
10	0.92	0-99	22	0.85	1.00	34	0.55	0.86	46	0.73	0.95	58	0.51	0.52	m
n	O.S2	0.95	23	0.76	0.93	35	0.46	0.86	47	0.33	0.48	59	0.06	0.35	U)
12	0.60	0.87	24	0.56	0,86	36	0.13	0.57	48	0.2]	0.43	60	0.08	0.38	

horizontal rule

how well the item discriminates among the testees within each group. These correlations
are given in Table 3.

Jensen (1980, p. 445) pointed out that an ideal (i.e., unbiased) test has item-total
correlations that are the same for each item when the two groups are compared. However,
this rarely occurs in practice because of low reliability at the item level and the effect of
marked group differences in item difficulties (as seen in Table 2). Ceiling effects, the result
of extremely easy items (seen in Table 2), also cause low correlations because of restriction
of variance among items.

Nonetheless, the item-total correlations in Table 3 do allow testing a hypothesis of
considerable interest: If the test measures the same ability in the African and the White
groups, then items that best measure ability within each group (i.e., those items with the
largest item-total correlations) should also discriminate most between the groups.
Difference in item difficulties between Africans and Whites were, therefore, correlated
with the items' discrimination values for the total sample. The results support the
hypothesis using either Pearson's (r = 0.70, p < 0.01) or, to ensure against scale artifacts,
Spearman's rank-order correlation (rho = 0.72, p < 0,01). Those items that best measure
individual differences within each ethnic group are the same items that most discriminate
between ethnic groups.

Differences in g

The total score on the Raven's is a very good measure of g, the general factor of
intelligence (Jensen, 1998, p. 38), Thus, the item-total correlation is an estimate of each
item's g loading. This provides an opportunity to test whether African-White differences
are more pronounced on the more g loaded items. The respective Pearson and Spearman
correlations between African-White differences in percentage passing each item (Table 2)
and the item-total correlations (Table 3) were: /- = 0.39 (p < 0.01, N= 58) and rho = 0.43
(p < 0.01, N= 58) using the African item-total correlations; r = 0.34 (p < 0.01, N- 46)
and rho = 0.41 (p<0-01, N - 46) using the White item-total correlations. Sex
differences, however, did not show up in g.

In the analyses reported above, the TV varied because items with 100 percent pass rates
were necessarily eliminated because item-total correlations could not be computed. Also,
the African and White pass rates were first normalized to standard scores before being
subtracted from each other. Alternative correction procedures, such as the odds-ratio
correction for percentiles, where item X - log[%/l—%], or alternative item selection
procedures such as eliminating those items with higher than a 95 percent pass rate, do not
alter the basic finding. (Note that it would be incorrect to use the item-total correlations
from the combined samples because these would reflect the between-groups variance in
addition to the within-groups variance and so inflate the effect.)


The difference in mean test scores of African and White, and of male and female
university students in South Africa on the untimed SPMs corroborate the previously found
group differences reviewed in the Introduction. With Africans and Whites averaging
Raven's scores at the 14th and 61st percentiles, respectively, Whites average 1 to 2 SD
higher than Africans. (The IQ equivalents are 84 and 105, respectively.) Men also
averaged slightly higher than did women.

horizontal rule

Table 3. Item-Total Correlations for Items of the Standard Progressive Matrices by Race
	Set A			SetK	SetC	SetD			SetE	
Item	African	While	Item	African	White	ftem	African	White	Item	African	White	Item	African	White
1	-	-	13	0.37	-0.02	25	0.48	-	37	0.31	-	49	0.38	0.31
2			14	0.46	-	26	0.31	-	38	0.51	-0.02	50	0.45	0.22
3	0.1]	-	15	0.59		27	0.57	-	39	0.42	0.09	51	0.50	0.32
4	0.26	-	16	0.50	-	28	0.36	0.18	40	0.40	0.12	52	0-51	0.47
5	0.32	0.19	17	0.45	-0-02	29	0.51	-	41	0.60	-	53	0.54	0.37
6	0.27	0.17	18	0.40	0-01	30	0.50	0.25	42	0.42	0.24	54	0.41	0.58
7	0.44	0.18	19	0.44	0.29	31	0.52	0.12	43	0.45	0.19	55	0.33	0.43
8	0.27	0.11	20	0-51	0.18	32	0.48	0.43	44	0.54	0.21	56	0-49	0.50
9	0.44	0.17	21	0.46	0.12	33	0.50	0.26	45	0.49	0.08	57	0.35	0,51
10	0,44	0.17	22	0,56	-	34	0.53	0.33	46	0,57	0.32	58	0.15	0.57
11	0.51	0.25	23	0.56	0.21	35	0.40	0.45	47	0.33	0.48	59	0.21	0.44
12	0.40	0.47	24	0.5!	0.35	36	0.31	0,43	48	0.36	0-34	60	0.48	0.55
Note: Hyphen indicates that correlation could not be computed because or lack of variance on item (sec Table 2).

horizontal rule

Black South African university students are a highly selected population. They have
passed standardized school matriculation exams, entered university, and been chosen for a
first-year course in Psychology on the basis of academic performance. Assuming that these
students are I SO above the population mean, the results are in accord with earlier work
finding that Africans, in general, average a tested IQ of 70.

A very wide range of interpretations, however, is possible for these results.
Interpreting them in terms of external IQ norms will carry little conviction for most
readers. It is difficult to believe that "true" scores at the 14th percentile on the Raven's
(with an IQ equivalent of 84) will allow entry to a select university. Can the African
students' motivation to succeed academically override their apparently low scores? Or is
their Raven performance,, unlike other groups, for some reason unconnected with their
academic achievement?

A very important point in testing is that test takers should be sufficiently similar in
cultural, educational, and social background to those on whom the test has been
standardized and the test norms based. If the testee or group differs markedly from the
standardization sample, the use of the norms may be inappropriate. Although the 1993 US
standardization sample (Raven et al., 1996) included a representative sample of black
Americans (i.e., African-Americans), the social context of Blacks in Africa is obviously
very different from that in the US. Although there have been centuries of discriminatory
practices in both countries, the apartheid system imposed on the majority population in
South Africa, along with confinement to "tribal" areas, has been different in kind from the
legacy of slavery of the minority population in the US.

Another problem in applying the test norms pertains to the White students. There
is a lack of fine discriminative power at the upper end of the SPM distribution where
the difference in raw score between percentiles is small (a "ceiling effect"). For 18- to
22-year-olds, the difference between the 61st and me 100th percentile is only six
correct answers, or 6 percentile points per 1 SPM score point. Thus, the mean IQ of
105 for White university students in this study is likely an underestimate. Arthur
Jensen (personal communication, July 29, 1999) reports that University of California,
Berkeley undergraduates, an elite US sample, averages 55 out of 60 items correct on
the SPM (SD of 2). This places them at the 68th percentile with an IQ equivalent of
107 (Raven et al., 1990, p. 98), very similar to the White South African students. But
they, and other White students at good universities in North America, typically have
average IQs of about 115 on other IQ tests (e.g., Rushton, 1992).

Rough-hewn though these results may be, they show population differences in accord
with earlier findings (reviews in Jensen, 1998; Lynn, 1997; Rushton, 1995). Further, items
found easy or difficult by Africans were the same ones found easy or difficult by Whites,
showing that the test measured the same construct in both groups. Since nearly 70 percent
of the items were "easy" for the African students (who averaged 44 correct out of the 60
problems), they can perform the required operations. The item difficulties and inter-item
correlations show that the items that best measure individual differences within each ethnic
group are the same items that best differentiate between the ethnic groups. Whatever the
cause of the substantial mean African-White differences, the systematic relationship
found between the difficulty of an item for Africans relative to its difficulty for Whites,
and that item's discrimination value within these groups, does not support the view that
differences between the groups are caused by biased test instruments.

The African-White TQ differences were greatest on those items correlating most
highly with the total test score, a good measure of g. This is the second demonstration
of this effect in an African population, following Lymi and Owen's (1994) study of
South African secondary school students using the Junior Aptitude Battery. Thus, the
effect seems robust. One implication may be that the causes of African-White
differences are similar to those for the Black-White difference in the US (Jensen,
1998; Rushton, 1995).

However, Olson (1986) argued that much of the supposed "culture fairness" of the
non-verbal SPM is illusory and that it requires the same analytical rules, rules for
analysis, coding, and transforming relationships as are required for analysis of verbal
content. Associated with this interpretation is the view that African languages and black
cultures are more "wholistic," and so not encouraging of the kind of thinking prevalent
in Western cultures. Seen from this perspective. Raven's scores do reflect culture
specifics. More generally, Sowell (1994) has observed a preference for spontaneity
and improvization over abstract thinking in Black cultures throughout the world, from
indigenous West African ancestors to their descendants in modem Brazil, the Caribbean,
and the US. He argues that black cultural patterns are deeply rooted and transmitted in
subtle ways. One can also posit the deleterious effects of the apartheid system's century
plus of institutional discrimination.

All the above interpretations view the low African scores as a valid measure of the
current level of abstract ability of the population, perhaps resulting from the adverse
effects of discrimination. Even now, South African Blacks have greater unemployment,
poorer schools, libraries, and study facilities than do Whites. They live in overcrowded
homes, often with no running water or electricity, and have poorer nutrition. Thus,
Africans may have had less exposure or stimulation to the constructs measured on IQ tests,
and therefore, their poor performance is the result of specific cognitive deficiencies.

Such cognitive deficits have been remedied by intervention techniques. For example,
some evidence from South Africa suggests that "dynamic mediation" with the Learning
Propensity Assessment Device (LPAD; Feuerstein, 1980) has improved the performance
of Black secondary school students on Raven's SPM. A study by Skuy and Shmukler
(1987) found that following such mediation, at least under certain conditions, the skills
learned were generalized to improved performance on the Raven's test. Skuy, Hoffenberg,
Visser, and Fridjhon (1990) found generalized improvements on the Raven's for those
individuals with what was termed a "facilitative temperament."

Rather man abandoning standardized testing in South Africa as "racist," research
should be conducted even more intensively. A useful first step is developing educational
programs to identity, nurture, and recognize more of the talents of more of the pupils. This
requires obtaining normalized distributions for the African population on existing tests and
developing new tests, including those of social intelligence. Experimental treatments such
as teaching problem-solving techniques and assessing the effectiveness of mediation (e.g.,
as defined by Feuerstein, 1980), as well as providing vitamin and mineral supplements to
enhance cognitive functioning, should be examined. (There is evidence that vitamin-
mineral supplements can add necessary trace elements to the brain in those who may have
been deprived of them and so improve test scores; Eysenck & Schoenthaler, 1997.)
Examining these questions will tell us a lot, not just about group differences, but about the
nature and nurture of intelligence as well.


Court, J. H. (1983). Sex differences in performance on Raven's progressive matrices: A review. Alberta Journal
of Educational Research, 29, 54-74.

Crawford-Nutt, D. H. (1976). Are hiack scores on Raven's standard progressive matrices test an artifact of the
method of test presentation? Psychologica Africana. 16, 201-206.
Eysenck. H- J-, & Schoenthaler, S. J. (1997). Raising IQ level by vitamin and mineral supplementation. In

R. J. Steniberg & E, Grigorenko (Eds.), Intelligence, heredity and environment (pp. 363-392). NewYork: Cambridge Univ. Press.

Feuerstein, R. (1980). The dynamic assessment of retarded performers. Glenview. IL: Scott, Foresman.
Glewwe, P., & Jacoby, H. (1992). Estimating the determinants of cognitive achievement in low income countries.

Washington, DC: World Bank.

Hartshome, K. (1992). Crisis and challenge: Black education 19ll)-i991). Cape Town: Oxford Univ. Press.
Hermstein, R. J., & Murray, C. (1994). The Bell curve. New York: Free Press.
Jensen. A. R. (1980). Bias in mental testing. New York: Free Press.
Jensen, A. R. (1998). The g factor. Westport, CT; Praeger.
Lynn, R. (1978). Ethnic and racial differences in intelligence: International comparisons. In R. T. Osbomc,

C. E. Noble, & N. Weyl (Eds.). Human variation: The biopsychology of age, race, and sex (pp.261-286). New York: Academic Press.

Lynn, R. (1991). Race differences in intelligence: A global perspective. Mankind Quarterly, 31, 255-296,
Lynn, R. (1997), Geographical variation in intelligence. InH. Nyborg (Ed.), The scientific study of human nature:

Essays in honor of H. J. Eysenck. New York: Pergamon-
Lynn. R. (1999). Editorial: Sex differences in intelligence and brain size: A developmental theory. Intelligence.

27. 1 -12.
Lynn, R., & Owen, K. (1994). Spearman's hypothesis and test score differences between Whites, Indians, and

Blacks in South Africa. Journal of General Psychology, 121, 27-36-
Macro-Economic Work Group. (1993). Making democracy work: A framework for macroeconomic policy in

South Africa. South Africa: Center for Development Studies.
Olson, D. R. (1986). Intelligence and literacy: The relationship between intelligence and the technologies of

representation and communication. In R. J. Stemberg & R. K. Wagner (Eds.), Practical intelligence:

Nature and origins of competence in the everyday world (pp. 338-360). Cambridge: Cambridge

Univ. Press.
Owen, K. (1992). The suitability of Raven's slandard progressive matrices for various groups in South Africa.

Personality and Individual Differences, 13, 149-159.
Owen, K.. (1998). The role of psychological tests in education in South Africa: Issues, controversies, and benefits.

Pretoria, South Africa: Human Sciences Research Council.
Poortinga, Y. H. (1971). Cross-cultura! comparison of maximum performance tests: Some methodological aspects

and some experiments with simple auditory and visual stimuli. Psychologica Africana (Monograph

Supplement No. 6).
Raven, J., et. al. (1990). Manual for Raven s progressive matrices and vocabulary scales. Research supplement

no. 3: American and international norms, 2nd edn. Oxford. England: Oxford Psychologists Press.
Raven, J. C., Court, J. H., & Raven, J. (1996). Manual for Ravens standard progressive matrices. 1996 edn.

Oxford, England: Oxford Psychologists Press.

Rushton, J- P. (1992). Life-history comparisons between Orientals and Whites at a Canadian university. Person-
ality and Individual Differences, 13, 439-442-

Rushton,J. P. (1995). Race; evolution, and behavior: A life history perspective. New Brunswick, NJ: Transaction.
Skuy, M., Hoffenberg, S-. Visser. L., & Fridjhon, P. (1990). Temperament and cognitive modifiability of acade-
mically superior Black adolescents in South Africa. International Journal of Disability, Development, and Education. 37, 29-44.

Skuy, M., Schutte, E., Fridjhon, P., O'Carroll, S. (1999). Suitability of published neuropsychological test norms for urban African secondary school students in South Africa. Unpublished manuscript. Division of Specialized Education, University of the Witwatersrand, Johannesburg 2050, South Africa.

Skuy, M., & Shmuklcr, D. (1987). Effectiveness of the learning potential assessment device with Indian and Colored adolescents in South Africa. International Journal of Special Education, 12, 131-149.
Skuy, M-, Zolezzi, S., Mentis, M., Fridjhon, P., & Cockcroft, K. (1996). Selection of advantaged and disadvantaged South African students for university admission. South African Journal of Higher Education.
11), 110-118.

Sowell, T. (1994). Race and culture. New York: Basic Books.

Spearman, C. (1927). The abilities of man: Their nature and measurement. New York: Macmilfan.

Zindi, F. (1994). Differences in performance. The Psychologist, 7, 549-552.