ThatDevGuy

Kinda hard since journalists never bother to properly cite their sources. I looked up the journal they mentioned, but the onle I have access to stopped publishing in 2001. Must be a different one with the wame name, perhaps.

Also note that typos have no correlation with intelligence.

Article FTW

Males have greater g: Sex differences in general mental ability from
100,000 17- to 18-year-olds on the Scholastic Assessment Test☆
Douglas N. Jackson, J. Philippe Rushton ⁎
The University of Western Ontario, Canada
Received 29 August 2005; received in revised form 17 February 2006; accepted 11 March 2006
Abstract
In this study we found that 17- to 18-year old males averaged 3.63 IQ points higher than did their female counterparts on the
1991 Scholastic Assessment Test (SAT). We analysed 145 item responses from 46,509 males and 56,007 females (total
N=102,516) using a principal components procedure. We found (1) the g factor underlies both the SAT Verbal (SAT-V) and the
SAT Mathematics (SAT-M) scales with the congruence between these components greater than 0.90; (2) the g components predict
undergraduate grades better than do the traditionally used SAT-V and SAT-M scales; (3) the male and the female g factors are
congruent in excess of .99; (4) male-female differences in g have a point-biserial effect size of 0.12 favoring males (equivalent to
3.63 IQ points); (5) male-female differences in g are present throughout the entire distribution of scores; (6) male-female
differences in g are found at every socioeconomic level; and (7) male-female differences in g are found across several ethnic
groups. We conclude that while the magnitude of the male-female difference in g is not large, it is real and non-trivial. Finally, we
discuss some remaining sex-difference/brain-size/IQ anomalies.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Brain size; Intelligence tests; Gender differences; Evolutionary psychology; Psychometrics
1. Introduction
For approximately a century a consensus has existed
that there are no sex differences in overall general intelligence.
It was British psychologist Cyril Burt who first
advanced this conclusion based on the results from a
series of reasoning tasks he developed and administered
to both boys and girls in various secondary schools in
Liverpool (inspired partly by Galton's work and that of
Binet in France; Burt & Moore, 1912). He thereby overturned
traditional Victorian wisdom (Ellis, 1904). Terman
(1916, pp. 69-70) further advanced the conclusion
with his American standardization sample of the Stanford-
Binet on approximately 1000 4- to 16-year-olds.
Their findings of "no sex difference in intelligence" have
since been replicated many times on other standardization
samples with other test batteries. However,
males are often observed to average higher scores on
some tests of spatial ability, mathematical reasoning, and
targeting, while females are often found to average
higher on some tests of memory, verbal ability, and motor
coordination within personal space (Halpern, 2000;
Kimura, 1999). Also, males have been found to have
Intelligence xx (2006) xxx-xxx
+ MODEL
INTELL-00316; No of Pages 8
☆ D.N.J. carried out the statistical analyses of these data and
presented them at the December 2002 meeting of the International
Society for Intelligence Research (ISIR) in Nashville, TN. Following
D.N.J.'s death in September 2004, J.P.R. completed the write up as
presented in this article.
⁎ Corresponding author. Department of Psychology, University of
Western Ontario, London, Ontario Canada N6A 5C2. Tel.: +1 519 661
3685; fax: +1 519 850 2302.
E-mail address: Rushto­nowuc­a (J.P. Rushton).
0160-2896/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.intell.2006.03.005
ARTICLE IN PRESS
greater test score variance than females, being overrepresented
at both the high and the low extremes
(Deary, Thorpe,Wilson, Starr, &Whalley, 2003; Hedges
& Nowell, 1995). When marked sex differences are
found in particular abilities, such as the 1 standard
deviation male advantage in rotating imaginary objects
(Kimura, 1999; Voyer, Voyer, & Bryden, 1995), they are
typically considered "sex-biased" and excluded from
general test batteries.
Two recent sets of observations raise anew the
question of sex differences in general intelligence in
normal populations. The first is that the general factor of
mental ability-g-permeates all tests to a greater or lesser
extent, as initially proposed by the British psychologist
Charles E. Spearman (1904, 1923; Jensen, 1998). More
than any other factor, the magnitude of the test's g
loading best determines a test's power to predict academic
achievement, creativity, career potential, and job
performance (Kuncel, Hezlett, & Ones, 2004; Lubinski,
2004). Thus, a "spatial" test may be relatively high on g
(mental rotation) or low (perceptual speed), a "verbal"
test may be relatively high (reasoning) or low (fluency),
as may a "memory" test be high (repeating a series in
reverse order) or low (repeating a series in presented
order). As Spearman (1923, p. 198) noted, there is an
"indifference of the indicator," meaning that the specific
content or form of the item is inconsequential so long as
it is apprehended or perceived in the same way by all
persons taking the test. The question of sex differences in
general intelligence can therefore be formulated more
precisely as: "Are there sex differences on the g factor?"
The second set of observations concerns the sex
difference found in brain size and the relation between
brain size and cognitive ability. In 1992, C. Davison
Ankney re-analysed 1000 brain weights at autopsy from
a study published by Ho, Roessmann, Straumfjord and
Monroe (1980) and discovered that males averaged a
larger brain size than females even after adjusting for
body size (140 g before adjustments; 100 g after adjustments).
Ankney's findings were immediately corroborated
by Rushton (1992) using cranial capacity data
calculated from head size measurements gathered from a
stratified random sample of 6325 U.S. military personnel
with individual adjustments made for many different
body size variables using analysis of covariance. At the
time, Ankney's and Rushton's findings were considered
"revolutionary" (Maddox, 1992) because prior studies of
sex differences in brain size argued that they "disappeared"
when adjusted for body size (by using an
inappropriate adjustment based on brain- to body-size
ratios, it turns out; Ankney, 1992). Additional data on sex
differences in brain size were reviewed by Rushton and
Ankney (1996) who confirmed the male advantage; they
also reviewed the relation between brain size and
cognitive ability, which they found to be 0.20 using
external head size measures and 0.40 using magnetic
resonance imaging, or MRI. Subsequently, Pakkenberg
and Gundersen (1997) documented that men have 15%
more neurons than women (22.8 versus 19.3 billion), and
over two-dozen MRI studies have confirmed the brainsize/
IQ correlation of about 0.40 (e.g., Ivanovic et al.,
2004; McDaniel, 2005).
Lynn (1999, p. 1) dubbed the findings on sex
differences in brain size "the Ankney-Rushton anomaly."
He argued that if brain size is linked to IQ, and males
average larger brains than females, then men should have
higher average intelligence than women. Lynn (1994,
1999) reviewed data from a number of published tests
such as the well-standardized highly-g-loaded Wechsler
Adult Intelligence Scale-Revised (WAIS-R) from
countries as varied as Britain, Belgium, Greece, China,
Israel, the Netherlands, Norway, Sweden, Japan, India,
and Indonesia, as well as the United States, and found that
men average 3.8 IQ points higher than women. Also in
support, Lynn and Irwing (2004) carried out a metaanalysis
of 57 studies of general population samples to
examine sex differences on the Standard and Advanced
Progressive Matrices, one of the most highly-g-loaded
tests of non-verbal reasoning, and found that adult men
exceeded adult women by an average of 5.0 IQ points.
Subsequently, Irwing and Lynn (2005) carried out a
meta-analysis of 22 studies of university samples on the
Progressive Matrices and found the male advantage
averaged between 3.3 and 5.0 IQ points, with 4.6 being
their best estimate.
Lynn (2005) has also pointed to other research supportive
of his hypothesis, such as that by Baron-Cohen
(2003; who prefers the terminology that men have
greater "synthesising ability"). He cited studies showing
that people of both sexes consistently rate their fathers as
more intelligent than their mothers (Furnham, 2001) and
of self-ratings showing that males rate themselves higher
than do females even after adjusting for any effects of sex
differences in personality (Furnham & Buchanan, 2005).
He also pointed to the evidence from history that males
made 98% of the world's contributions to knowledge
(Murray, 2003).
Age turns out to be an important factor for determining
sex differences in IQ because the male advantage does not
emerge until the late adolescent growth spurt when brain
size differences peak. Girlsmature faster than boys, which
give them an early advantage in language development
and may mask later cognitive differences. One study of
1400 3-year-olds in Mauritius found girls had a nearly
2 D.N. Jackson, J.P. Rushton / Intelligence xx (2006) xxx-xxx
ARTICLE IN PRESS
2 IQ point advantage over boys in verbal ability and also
in visual-spatial ability on the Boehme Test of Basic
Concepts (Lynn, Raine, Venables, & Mednick, 2005).
However, among children 5 to 15 years of age, Lynn
(1994, 1999) found no consistent sex differences, a general
finding which he suggested led generations of researchers,
who relied on school samples, to miss the later
emerging sex difference.
Lynn's procedure for determining sex differences in g
has been the straightforward one of summing items and
subtests from g loaded batteries to produce a total score.
Jensen (1998) criticized this method as producing outcomes
dependent on the particular mix of questions in the
battery and suggested the use of principal components
analysis to extract g instead, a procedure he had found
useful in examining ethnic and other group differences.
When Jensen (1998) carried out this analysis with five
large data sets he observed considerable inconsistency:
although males averaged higher in three studies (by 5.49,
2.81, and 0.18 IQ points), females averaged higher in two
others (by 7.91 and 0.11 IQ points), which he averaged to
give a negligible male advantage of 0.11 IQ points.
Jackson (2002) endorsed Jensen's use of the
principal components procedure but criticized his choice
of tests, suggesting that the General Aptitude Test
Battery (the one on which females had scored 7.91 IQ
points higher), biased the results because it included a
large number of low g tests of mechanical aptitude on
which males excelled but which resulted in their
achieving a low g score. He suggested limiting the use
of principal components analysis to highly-g-loaded
tests such as theWechsler or his own Multi-Dimensional
Aptitude Battery (Jackson, 1984). When Jackson (2002)
analysed previously published data this way, including
that by Jensen and Reynolds (1982) on the U.S.
standardization of the Wechsler Intelligence Scale for
Children-Revised (WISC-R), and by Lynn and Mulhern
(1991) on the Scottish WISC-R standardization, he
found that males outperformed females on the more g
loaded subtests.
Most recently, Nyborg (2003, 2005) has argued that
the principal components type of analysis used by Jensen
and Jackson is insufficiently sensitive to detect true score
differences and so provides little or no improvement over
Lynn's summing of subtests because the magnitude of the
sex difference is modest and the number of subtests being
analysed is typically 12 or fewer. In this type of analysis
the test of significance is often based on the number of
subtests and so the result can be thrown off by the
inclusion or exclusion of even one subtest. Nyborg noted
that although more powerful multivariate methods (such
as hierarchical factor analysis and structural equation
modeling) are available for identifying and controlling
factorial impurities in the measures employed (e.g., spatial
visualization rather than g) and although these are used
routinely when looking at other group differences, they
are rarely used in studying sex differences.
To date, only two studies have been entirely
satisfactory from a multivariate perspective and both
have other imperfections. Nyborg (2005) used hierarchical
factor analysis with the Schmid-Leiman transformation
on a battery of 20 widely differing ability tests
given to carefully matched groups of Danish 17-yearolds
and found a g difference favoring males equivalent
to 8.55 IQ points. Although this was a longitudinal study
with repeated measures, the main analyses rested on a
very small sample of only 31 males and 31 females. The
other study, by Colom, Garcia, Juan-Espinosa and Abad
(2002), analysed 703 female and 666 male participants in
the Spanish standardization of theWAIS-III and reported
a male advantage of 3.6 IQ points. The clarity of the
result was reduced, however, when the authors combined
it with other findings and concluded there were "null sex
differences in general intelligence." It was only when
Nyborg (2003, 2005) disaggregated the data of Colom et
al. (2002) and applied a statistical test that the male
advantage on g emerged.
2. Method
The present study provides a more definitive test of
the hypothesis that males average higher in g than
females by factor analysing the 145 item scores (correct
or incorrect) of 46,509 male and 56,007 female 17- to
18-year-olds (total N=102,516) from the "validity study
sample" of the 1991 administration of the well-known
Scholastic Assessment Test (SAT; College Entrance
Examination Board, 1992). Before 1990 the test was
called the Scholastic Aptitude Test; after 1994 it was
renamed simply SAT-not an acronym.
The SAT provides a reasonable way of testing the
hypothesis of sex differences in g. It is anexamthat several
generations of high school students have taken for
admission to college in theUnited States thatwas carefully
developed from a psychometric perspective to maximize
prediction of college-level academic achievement while
minimizing extraneous factors such as potential sex,
ethnic, and social class bias. About 50% of the U.S.
population now go to college and SAT test takers are
representative of those who aspire to do so. Over decades
of research it has demonstrated substantial reliability and
validity and is considered the "gold standard" for academic
achievement tests (The lowest scoring SAT group are
likely below the average IQ of 100.).
D.N. Jackson, J.P. Rushton / Intelligence xx (2006) xxx-xxx 3
ARTICLE IN PRESS
Although the SAT is widely believed to measure
mainly academic achievement, Frey and Detterman
(2004) have shown that it is an excellent measure of
general cognitive ability (g). For example, they found
that scores on the SAT correlated highly (0.82) with g
extracted from the Armed Services Vocational Aptitude
Battery (ASVAB), as it also did with g loaded mental
ability tests such as the California Test of Mental
Maturity (0.82), the Otis-Lennon Mental Ability Test
(0.78), the Lorge-Thorndike Intelligence Test (0.79),
the Differential Aptitude Test (0.78), the Henmon-
Nelson Test of Mental Maturity (0.65), the Coop School
and College Ability Test (0.53), and the Raven's Advanced
Progressive Matrices (0.48; 0.72 when corrected
for restricted range).
The item scores to be analysed are the responses to
multiple-choice questions arranged in four sections, each
lasting 30 min: a 45-item verbal section, a 40-item verbal
section, a 25-item mathematics section, and a 35-item
mathematics section. The verbal questions were of four
types: analogies, reading comprehension, antonyms, and
sentence completions; the mathematics questions were
also of four types: arithmetic, algebra, geometry, and
quantitative comparisons. These sum to yield two scores:
SAT Verbal (SAT-V) and SAT Mathematics (SAT-M).
However, we disregard question type, section, and subtest
in most of our analyses and instead focus on extracting the
g component from each item. Such procedures have
proven very useful in establishing the convergent and
divergent validities of mental ability subtests as well as
their g loadings (Jackson, 1984; Jensen, 1998). They have
also proven useful for determining the structure of
intelligence in different ethnic groups (Jensen, 1998;
Rushton & Jensen, 2005; Rushton, Skuy, & Bons, 2004).
3. Results
Both the SAT Verbal (SAT-V) scale and the SAT
Mathematics (SAT-M) scale have a mean of 500 and a
standard deviation of 100. In our 1991 validity sample,
males averaged 499 on the SAT-M and 434 on the SATV,
while females scored 457 on the SAT-M and 425 on
the SAT-V-a magnitude of sex difference reported for
the last 32 years (Halpern, 2000, p. 129, Figs. 3.8 and
3.9). In 1995, both the SAT-M and SAT-V scores were
re-centered upwards by a total of about 100 points or
one standard deviation to adjust for the fact that the
mean scores had been declining.
We first carried out a principal components analysis
of the items for the SAT-Vand SAT-Msubtests to extract
their corresponding g scores. Table 1 provides the
correlations between scores from the SAT-V and the
SAT-M and their corresponding first principal components.
Note that while the correlation between SAT-Vand
SAT-M is 0.67, the corresponding correlation for their g
scores is 0.90. Given their respective reliabilities, this
latter value approaches unity when corrected for attenuation.
These results support Frey and Detterman's
(2004) conclusion that the SAT-Vand SAT-M are primarily
measures of g, with the other factors they measure
being secondary. When sex is entered into the matrix as a
dummy variable, the point-biserial correlations are
positive, indicating that males score higher on average
than females. Extracting the g factor score from all the
SAT-V and SAT-M items together gave a point-biserial
effect size favoring males of 0.12, which is equivalent to
0.24 standard deviation units (see Jensen, 1998, p. 543,
n. 12). Given a standard deviation of 15 typical for tests
such as the Wechsler, the difference between men and
women is equivalent to 3.63 IQ points.
The g factors extracted from the male and female
samples were highly congruent, in fact virtually
identical, in excess of 0.99. This means that the items
"behave" in the exact same way for males as they do for
females and have the same "meaning." For example,
those items that men find difficult, women do too.
We examined how well the SAT-Vand SAT-M scores
predict freshman year university grades before and after
g has been partialed out. Fig. 1 depicts the correlations
between SAT-V and SAT-M scores and grades for
English, Foreign Language, Natural Science, Mathematics,
and Social Science. The first two bars in each set
are those for the traditional SAT-V and SAT-M scales.
The second two bars are the respective correlations with
the g factor removed. Note that the non-g components
make a negligible contribution for all academic subjects
except Mathematics. For Mathematics there is a modest
residual correlation of approximately 0.12, suggesting
that, in addition to g, differences in experience with
mathematics or some residual component such as spatial
ability might contribute to correlations with grades.
Table 1
Correlations between sex⁎, SAT scale scores, and SAT g scores
(N=102,516)
Sex SAT-V SAT-M SAT-V
g score
SAT-M
g score
Sex 1.00 0.05 0.19 0.10 0.12
SAT-V 1.00 0.67 0.91 0.89
SAT-M 1.00 0.86 0.84
SAT-V
g score
1.00 0.90
SAT-M
g score
1.00
Note: ⁎Females = 1, males = 2.
4 D.N. Jackson, J.P. Rushton / Intelligence xx (2006) xxx-xxx
ARTICLE IN PRESS
We examined the male-female differences in g across
a representation of the entire distribution (Fig. 2). Males
are over-represented in each and every category above
the middle category, and females are over-represented in
each and every category below the middle category. This
indicates that male-female differences in g occur across
the entire distribution of g scores. Category 1, representing
the lowest block of scores, contains persons who
scored below chance (perhaps misled by distracter
items). The ratio of females to males who did so was
approximately 5:3. In the past, when males were
observed to obtain a higher mean score, it was often
attributed to males having greater representation at the
extreme high end of the distribution (Hedges & Nowell,
1995). Fig. 2 indicates this is not correct.
Fig. 3 contains the male and female g score
distributions (mean standardized to 50, with SD of 10)
for eleven levels of family income. Note that the male-
0
1
2
3
4
5
6
7
8
9
10
1 2 3 4 5 6 7 8 9 10 11
SAT levels from low to high
Percentage
Males Females
Fig. 2. Distribution of males and females at eleven SAT g factor score
levels. Males are more numerous at upper range of the distribution and
females in the lower range.
0
10
20
30
40
50
60
70
80
Family income levels from low to high
Mean = 50; SD = 10
1 2 3 4 5 6 7 8 9 10 11
Males Females
Fig. 3. Male and female SAT g scores (mean standardized) for eleven
levels of family income.
American Indian
Asian
Black
Mexican American
Puerto Rican
Latin American
White
Total
Ethnic groups
0
10
20
30
40
50
60
70
Mean = 50; SD = 10
Males Females
Fig. 4. Male and female SAT g scores (mean standardized) for seven
different ethnic groups.
-0.1
0
0.1
0.2
0.3
0.4
0.5
English
Foreign language
Natural science
Mathematics
Social science
Academic subject
Correlation
SAT-V SAT-M SAT-V g-score SAT-M g-score
Fig. 1. SAT correlations with grades. The effects of removing g factor
variance from SAT scores reduces correlation to zero.
D.N. Jackson, J.P. Rushton / Intelligence xx (2006) xxx-xxx 5
ARTICLE IN PRESS
female difference holds at each level. The same
conclusion of higher scores for males also holds for
nine levels of father's education and nine levels of
mother's education (not shown as figures). Hence, it is
most unlikely that the sex differences are due to these type
of family background factors.
Fig. 4 shows the overall male-female g score
differences (mean standardized to 50, with SD of 10)
across seven ethnic groups, representing the largest
broad groups who complete the SAT. All seven groups
show sex differences although there is variation among
them on the magnitude of the difference. The two far
right bars show the overall mean is 105 for men and 95
for women.
4. Discussion
This study found a point-biserial effect size of 0.12
favoring males on the SAT, which provides a good
measure of g as manifested through "school-learned
abilities" in high school graduating samples. (This is
equivalent to a d of 0.24 on a test with a conventional SD
of 15, or 3.63 IQ points.) Despite its modest size,
particularly in comparison with the variance of scores
within the sexes, the results reported here support Lynn's
(1994, 1999) hypothesis of significant sex differences in
average cognitive ability. It is in accord with the previous
estimates of the male advantage reviewed in the
Introduction and approximates the 3.8 IQpoints estimated
by Lynn (1999) but shy of the 5.0 IQ points estimated by
Lynn and Irwing (2004) using the Progressive Matrices.
However, Lynn and Irwing's analyses gave a 0.16d for
17-year-olds (IQ=2.4) and a 0.30d for 20- to 29-year-olds
(IQ=4.5) so our result for 17-year-olds corroborates
Lynn's rather closely for this age range.
A point-biserial effect size of 0.12 is considered
"small." Nonetheless, it indicates that if selection occurred
at the mean, then 55% of males would pass compared to
only 45% of females (Rosnow, Rosenthal, & Rubin,
2000). If a more stringent criterion for selection were
applied, say at the 85th percentile, the ratio of males
selected over females would be considerably higher.
Moreover, Jackson (2002) suggested that because test
constructors such as himself and the Educational Testing
Service (which developed the SAT) often eliminate items
showing marked sex differences in order to reduce the
perception of bias, it is possible that the results reported
here might be a lower-bound estimate of the "true" sex
difference. It is also possible that the male advantage is
underestimated because of restriction in range on g in this
higher performance group; if so, the sex difference would
be larger in the general population.
It might be questioned, however, whether findings
about sex differences on the high-end SAT are generalizable
to the general population.Also, because 55%of
our SATsample was female, it might be hypothesized that
we drew from lower in the female IQ pool, and that this is
why a lower mean g was found for females. Nonetheless,
as shown in Figs. 2, 3, and 4, the sex difference was found
in the very highest and the very lowest SAT levels, at the
very highest and the very lowest socioeconomic status
levels, and for each and every ethnic group, including
Whites examined alone. To maintain that selection bias
caused the sex difference in this data set, therefore, would
require the assumption that there are hypothetical
respondentswho, if tested, would provide a compensating
female-male advantage in g that would counterbalance
the findings. They would have to be found at every level
of SAT performance, in every level of family income, for
every level of fathers' and of mothers' education, and for
every ethnic group examined.
It could still be argued, however, that some
ambiguity remains in interpreting the results. Although
there is a clear sex difference in g among students taking
the test, it is perhaps less clear that this sex difference
captures faithfully the sex difference in the general
population. Better sampling of both sexes from the SAT
pool might answer this question more definitively, as
may additional parametric studies of sex differences
from other parts of the g distribution.
The g factors extracted from the male and female
samples were highly congruent indicating they were
measuring the same thing for both sexes. Sex differences
in g probably rest on sex differences in brain functioning.
Both Ankney (1992) and Rushton (1992) found that
males average 100 g more brain weight than females
even after correction for body size. Ankney estimated the
sex difference in brain size as 0.78 standard deviations.
Assuming a 0.35 correlation between brain size and IQ,
therefore, Lynn (1999) and Nyborg (2005) predicted that
the male advantage in average IQ is 0.78×0.35=0.273
standard deviation units, equal to 4.10 IQ points on a test
like the Wechsler standardized with SD=15. This
predicted outcome is very close to the 4.3 mid-point of
the range of observed outcomes for large samples, i.e.,
IQ=3.6 in the present study and IQ=5.0 in Lynn and
Irwing's (2004) review.
However, several sex-difference/brain-size/IQ anomalies
still require resolution. First, there is a major gap in
Lynn's resolution of the Ankney-Rushton anomaly.
Males are found to average a larger brain size from birth
onwards, even after controlling for body size. For
example, in a study of 100 East Asian children followed
from birth to age seven, boys at birth averaged a cranial
6 D.N. Jackson, J.P. Rushton / Intelligence xx (2006) xxx-xxx
ARTICLE IN PRESS
capacity 5 cm3 larger than girls, a difference that increased
to 40 cm3 by 4months, and 50 cm3 by age 1 year and then
remained stable through to age 7 years (Rushton, 1997;
controlling for body size). Other data show the 50 cm3
male advantage in brain size at 1 year remains stable until
adolescence when male brains grow to become 140 to
160 cm3 larger than female brains (Rushton & Ankney,
1996; also controlling for body size).
Brain tissue is metabolically expensive. It would be
interesting to know what these "extra male neurons" are
doing from birth to age 16 before the male advantage in
IQ manifests itself. Ankney (1995) hypothesized that
they are related to the male advantage in dynamic spatial
abilities (not measured by IQ tests) such as in throwing at
targets. A male advantage at targeting shows up among
3- to 5-year-olds even when the tasks are simple
underhanded throws and avoid sex differences in
skeletomuscular structure (Kimura, 1999). Dynamic
spatial ability may also explain the additional anomaly
that Black males average a larger brain size than do
White or East Asian females, even after adjustments for
body size, and despite averaging 11 IQ points or more
lower than White or East Asian females (Rushton &
Ankney, 1996). Additional research using magnetic
resonance imaging and a wider array of cognitive tasks
could surely untangle these further conundrums and
interrelationships.
References
Ankney, C. D. (1992). Sex differences in relative brain size: The
mismeasure of woman, too? Intelligence, 16, 329−336.
Ankney, C. D. (1995). Sex differences in brain size and mental
abilities: Comments on R. Lynn and D. Kimura. Personality and
Individual Differences, 18, 423−424.
Baron-Cohen, S. (2003). The essential difference. London: Penguin/
Perseus.
Burt, C. L., & Moore, R. C. (1912). The mental differences between the
sexes. Journal of Experimental Pedagogy, 1, 273-284, 355-388.
College Entrance Examination Board. (1992). Validity study sample of
the 1991 SAT administration. New York, NY: College Entrance
Examination Board.
Colom, R., Garcia, L. F., Juan-Espinosa, M., & Abad, F. (2002). Null
sex differences in general intelligence: Evidence from the WAISIII.
Spanish Journal of Psychology, 5, 29−35.
Deary, I. J., Thorpe, G., Wilson, V., Starr, J. M., & Whalley, L. J.
(2003). Population sex differences in IQ at age 11: The Scottish
Mental Survey 1932. Intelligence, 31, 533−542.
Ellis, H. (1904). Man and woman (4th edition). New York: Houghton
Mifflin.
Frey, M. C., & Detterman, D. K. (2004). Scholastic assessment or g?
The relationship between the scholastic assessment test and general
cognitive ability. Psychological Science, 15, 373−378.
Furnham, A. (2001). Self-estimates of intelligence. Personality and
Individual Differences, 31, 1381−1405.
Furnham, A., & Buchanan, T. (2005). Personality, gender and selfperceived
intelligence. Personality and Individual Differences, 39,
543−555.
Halpern, D. F. (2000). Sex differences in cognitive abilities (3rd edition).
Mahwah, NJ: Erlbaum.
Hedges, L. V., & Nowell, A. (1995). Sex differences in mental test
scores, variability, and numbers of high-scoring individuals.
Science, 269, 41−45.
Ho, K. C., Roessmann, U., Straumfjord, J. V., & Monroe, G. (1980).
Analysis of brain weight: I and II. Archives of Pathology and
Laboratory Medicine, 104, 635−645.
Irwing, P., & Lynn, R. (2005). Sex differences in means and variability
on the Progressive Matrices in university students: A metaanalysis.
British Journal of Psychology, 96, 505−524.
Ivanovic, D. M., Leiva, B. P., Castro, C. G., Olivares, M. G., Jansana,
J. M. M., Castro,V. G., et al. (2004). Brain development parameters
and intelligence in Chilean high school graduates. Intelligence, 32,
461−479.
Jackson, D. N. (1984). Multidimensional aptitude battery. Port Huron,
MI: Research Psychologists Press.
Jackson, D. N. (2002, December 5-7). Evaluating g in the SAT:
Implications for the sex differences and interpretations of verbal
and quantitative aptitude. Paper presented at the International
Society for Intelligence Research, Nashville, TN.
Jensen, A. R. (1998). The g factor. Westport, CT: Praeger.
Jensen, A. R., & Reynolds, C. R. (1982). Race, social class and ability
patterns on the WISC-R. Personality and Individual Differences, 3,
423−438.
Kimura, D. (1999). Sex and cognition. Cambridge, MA: MIT Press.
Kuncel, N. R., Hezlett, S. A., & Ones, D. S. (2004). Academic
performance, career potential, creativity, and job performance: Can
one construct predict them all? Journal of Personality and Social
Psychology, 86, 148−161.
Lubinski, D. (2004). Introduction to the special section on cognitive
abilities: 100 years after Spearman's (1904) "'General intelligence,'
objectively determined and measured". Journal of
Personality and Social Psychology, 86, 96−111.
Lynn, R. (1994). Sex differences in intelligence and brain size: A
paradox resolved. Personality and Individual Differences, 17,
257−271.
Lynn, R. (1999). Sex differences in intelligence and brain size: A
developmental theory. Intelligence, 27, 1−12.
Lynn, R. (2005). Sex differences in IQ. The Psychologist, 18, 471
(Letter).
Lynn, R., & Irwing,

(blah blah blah)

Crap, didn't realize it would paste so long. Here's a link, click "Download File" near the top.

http://upload.ohshare.com/v/5721687/study.pdf.pdf.html