Intelligence 34 (2006) 253 – 260

Developmental gender differences on the Naglieri Nonverbal Ability

Test in a nationally normed sample of 5–17 year olds
Johannes Rojahn *, Jack A. Naglieri
George Mason University, United States
Received 22 June 2005; received in revised form 18 September 2005; accepted 26 September 2005
Available online 14 November 2005

Abstract

Lynn [Lynn, R. (2002). Sex differences on the progressive matrices among 15–16 year olds: some data from South Africa.
Personality and Individual Differences 33, 669–673.] proposed that biologically based developmental sex differences produce
different IQ trajectories across childhood and adolescence. To test this theory we analyzed the Naglieri Nonverbal Ability Test
(NNA; [Naglieri, J. A. (1997). Naglieri Nonverbal Ability Test-Multilevel Form. San Antonio: Harcourt Assessment Company.])
standardization sample of 79,780 children and adolescents in grades K-12, which was representative of the US census on several
critical demographic variables. NNAT data were consistent with Lynn’s developmental theory of gender differences insofar as (a)
there were no gender differences between 6 and 9 years; (b) females scored slightly higher between 10 and 13 years; and (c) males
were ahead of females between the ages of 15 and 16. However, the discrepancies between the genders were smaller than predicted
by Lynn. In fact they were so small that they have little or no practical importance. In other words, the NNAT did not reveal
meaningful gender differences at any stage between the ages of 6 and 17 years.
D 2005 Elsevier Inc. All rights reserved.

Gender differences in cognitive ability as well as ders, however, were not uniform across tasks. For in-
academic achievement has been a topic of considerable stance, the effect size for vocabulary was minimal
interest for some time, resulting in a substantial body of (d = .02) but more substantial for speech production
literature on the topic (e.g., Deaux, 1984; Fennema & (d = .33). Geary (1996) found gender differences in
Sherman, 1977; Geary, 1989, 1994, 1996; Halpern, quantitative skills. He reported that bthe male advantage
1986, 1989, 1997; Linn & Peterson, 1985; Lynn & in certain areas of mathematics (e.g., problem solving) is
Irwing, 2004; Maccoby & Jacklin, 1974; Voyer, related to a male advantage in spatial abilitiesQ (p. 236).
Voyer, & Bryden, 1995). Hyde and Linn (1988) con- Females, on the other hand, have been found to have an
ducted a meta-analysis of 165 studies of gender differ- advantage over males on basic arithmetic tests, at least
ences and found a small mean effect size (favoring through junior high school (Hyde, Fennema, & Lamon,
females) of .11 in verbal skills for studies of students 1990). Halpern (1997) summarized the research and
aged 5 through 18 years. The differences between gen- concluded that females outperform males on tests of
verbal fluency, foreign language, fine motor skills,
speech articulation, reading and writing, and math cal-
* Corresponding author. Center for Cognitive Development, George
Mason University, 10340 Democracy Lane, Fairfax VA 22030, Unit- culation. In contrast, males do better on tasks that in-
ed States. Tel.: +1 703 993 4241. volve mental rotation, mechanical reasoning, math and
E-mail address: [email protected] (J. Rojahn). science knowledge, and verbal analogies.
0160-2896/$ - see front matter D 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.intell.2005.09.004
254 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260

Table 1
Demographic information by NNAT level groups
NNAT levels
A B C D
Male Female Male Female Male Female Male Female
N % N % N % N % N % N % N % N %
Ethnic
White 2027 69.6 2060 73.5 2342 68.6 2385 70.5 2595 64.2 2565 63.1 5198 64.8 5281 66.1
African-American 328 11.3 230 8.2 408 12.0 366 10.8 704 17.4 711 17.5 984 12.3 922 11.5
Hispanic 322 11.1 326 11.6 355 10.4 319 9.4 461 11.4 452 11.1 895 11.2 852 10.7
Asian 83 2.9 46 1.6 73 2.1 109 3.2 128 3.2 160 3.9 200 2.5 193 2.4
Native American 19 0.7 26 0.9 55 1.6 47 1.4 45 1.1 54 1.3 79 1.0 79 1.0
Other 30 1.0 28 1.0 57 1.7 39 1.2 35 0.9 41 1.0 104 1.3 89 1.1
Special schooling
Special Ed 38 1.3 11 0.4 104 3.0 22 0.7 234 5.8 145 3.6 433 5.4 195 2.4
GT 9 0.3 12 0.4 61 1.8 101 3.0 232 5.7 144 3.5 555 6.9 580 7.3
Region
Northeast 511 17.5 413 14.7 779 22.8 702 20.7 679 16.8 686 16.9 1729 21.6 1729 21.7
Midwest 818 28.1 790 28.2 896 26.3 902 26.7 875 21.6 828 20.4 1805 22.5 1710 21.4
Southeast 774 26.6 700 25.0 878 25.7 805 23.8 880 21.8 835 20.5 1573 19.6 1600 20.0
West 809 27.8 900 32.1 859 25.2 975 28.8 1610 39.8 1719 42.3 2909 36.3 2945 36.9
Urbanicity
Urban 349 12.0 365 13.0 413 12.1 457 13.5 1055 26.1 1123 27.6 1891 23.6 1887 23.6
Suburban 1218 41.8 1259 44.9 1588 46.5 1566 46.3 1714 42.4 1640 40.3 3464 43.2 3499 43.8
Rural 1026 35.2 929 33.1 1026 30.1 1023 30.2 882 21.8 885 21.8 1768 22.1 1721 21.6
Non-public 319 11.0 250 8.9 385 11.3 338 10.0 393 9.7 420 10.3 893 11.1 877 11.0
SES
1 534 18.3 479 17.1 522 15.3 551 16.3 746 18.4 738 18.1 1356 16.9 1302 16.3
2 489 16.8 557 19.9 411 12.0 455 13.4 963 23.8 1108 27.2 1460 18.2 1559 19.5
3 562 19.3 585 20.9 758 22.2 782 23.1 761 18.8 676 16.6 1415 17.7 1465 18.3
4 499 17.1 447 15.9 701 20.5 587 17.3 606 15.0 528 13.0 1285 16.0 1220 15.3
5 509 17.5 485 17.3 635 18.6 671 19.8 575 14.2 598 14.7 1607 20.0 1561 19.6

NNAT levels
E F G
Male Female Male Female Male Female
N % N % N % N % N % N %
Ethnic
White 4343 56.3 4180 55.3 4935 55.6 5104 55.0 2474 53.1 2397 47.3
African-American 930 12.1 1007 13.3 904 10.2 1121 12.1 501 10.8 566 11.2
Hispanic 983 12.7 960 12.7 1061 12.0 1039 11.2 248 5.3 320 6.3
Asian 251 3.3 238 3.1 250 2.8 169 1.8 70 1.5 107 2.1
Native American 85 1.1 63 0.8 85 1.0 56 0.6 63 1.4 37 0.7
Other 95 1.2 82 1.1 95 1.1 102 1.1 53 1.1 76 1.5
Special schooling
Special Ed 548 7.1 297 3.9 631 7.1 252 2.7 250 5.4 127 2.5
GT 499 6.5 609 8.1 438 4.9 669 7.2 159 3.4 319 6.3
Region
Northeast 1462 18.9 1406 18.6 1928 21.7 2120 22.8 725 15.6 807 15.9
Midwest 1851 24.0 1781 23.6 2422 27.3 2450 26.4 1575 33.8 1385 27.3
Southeast 1458 18.9 1319 17.5 2130 24.0 2251 24.2 1225 26.3 1327 26.2
West 2945 38.2 3050 40.4 2398 27.0 2465 26.5 1131 24.3 1546 30.5
Urbanicity
Urban 1560 20.2 1645 21.8 1356 15.3 1373 14.8 232 5.0 534 10.5
Suburban 3365 43.6 3272 43.3 3974 44.8 4038 43.5 2006 43.1 2167 42.8
Rural 2106 27.3 1903 25.2 2540 28.6 2817 30.3 1865 40.1 1843 36.4
Non-public 685 8.9 736 9.7 1008 11.4 1058 11.4 553 11.9 521 10.3
 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260 255

Table 1 (continued)
NNAT levels
E F G
Male Female Male Female Male Female
N % N % N % N % N % N %
SES
1 1310 17.0 1423 18.8 1457 16.4 1428 15.4 634 13.6 995 19.6
2 1494 19.4 1377 18.2 1615 18.2 1740 18.7 925 19.9 924 18.2
3 1493 19.3 1569 20.8 1521 17.1 1566 16.9 870 18.7 918 18.1
4 1246 16.1 1159 15.3 1588 17.9 1698 18.3 736 15.8 737 14.6
5 1488 19.3 1292 17.1 1689 19.0 1796 19.3 938 20.1 970 19.2

Some researchers have argued that gender differences Africa using the Raven’s Progressive Matrices. Al-
cannot be adequately understood unless males and though his sample was large (N = 3979), it did not pro-
females are compared according to a theoretical model vide a range of ages that could adequately test his
of cognitive functioning (e.g., McHough, Koeske, & expectations. Our goal was to do just that using a sample
Frieze, 1986; Naglieri & Rojahn, 2001). Geary (1989) of children from the United States who ranged in age
further emphasized that conceptual models of cognitive from 5 to 17 years.
differences between the genders should provide an inte-
gration of the neurological and socio-cultural compo- 1. Method
nents that influence the development of cognitive
processes. More recently, Lynn (2002) and Lynn and 1.1. Participants
Irwing (2004) argued that sex differences must be
viewed developmentally and with consideration of the The subjects of this study were 79,780 children and
role played by biology. adolescents from kindergarten to grade 12 who partic-
Based upon his research using Raven’s Progressive ipated in the Naglieri Nonverbal Ability Test (NNAT;
Matrices, Lynn (2002) argued that one would expect that Naglieri, 1997) standardization during the 1995–1996
(a) no sex differences exist between the genders during school years. Of these, approximately 67,000 were
the ages of six to nine; (b) females move one IQ point tested in spring 1996 and the remaining was tested in
ahead of males between the ages of 10 and 13; (c) after fall of 1995. The methods used to collect the sample
age 13 females’ growth begins to decelerate relative to which was representative of the U.S. population in
males and they begin to lose their advantage; and (d) terms of geographic region, socioeconomic status, urba-
from 15 to 16 onward males catch up and overtake nicity, ethnicity, and school setting (public or private)
females ending up with an advantage that reaches 2.4 and the procedures used to create the NNAT norms are
IQ points among adults. Lynn’s (2002) study was based fully described by (Naglieri, 1997). The groups of
on samples of 15 to 16 year old adolescents from South children and adolescents used in this study are further

Table 2
Chronological ages and NAI scores for males and females by NNAT levels
Levels Males Females d-ratio NAI diff t
Age NAI n Age NAI n
M S.D. M S.D. M S.D. M S.D.
A 6.1 0.4 100.0 15.5 2912 6.0 0.4 98.9 16.1 2803 0.07 1.1 2.5*
B 7.1 0.5 99.6 16.0 3412 7.0 0.5 100.9 15.8 3384 0.08 1.3 3.3**
C 8.1 0.5 98.9 15.4 4044 8.0 0.5 98.6 15.5 4068 0.02 0.3 1.0
D 9.6 0.8 100.8 16.7 8016 9.5 0.7 100.5 15.5 7984 0.02 0.3 1.2
E 11.8 0.8 99.0 16.5 7716 11.7 0.7 99.9 15.4 7556 0.06 0.9 3.5***
F 14.2 1.0 99.6 17.1 8878 14.0 1.0 100.3 15.9 9286 0.04 0.7 2.1*
G 17.1 1.0 100.3 17.0 4656 16.9 0.9 99.6 14.7 5065 0.04 0.7 2.1*
* = p b .05; ** = p b .01; *** = p b .001.
NAI diff = gender differences in NNAT NAI scores.
256 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260

102

Mean NNAT NAI Score

100

98

Gender
Male
Female

96

A B C D E F G
NNAT Level

Fig. 1. Mean NNAT NAI scores across NNAT levels (6 to 17 year olds).

described in Table 1 which shows demographic char- 1.2. Instrument

acteristics for males and females by NNAT level and
Table 2 which provides the means and standard devia- The NNAT (Naglieri, 1997) was designed to mea-
tion of the ages of participants who were administered sure general ability of children and adolescents using a
the different NNAT levels. series of progressive matrix items that involve shapes

3.0

2.0
Male vs. Female NAI Score Differences

Advantage of males

1.0

0.0

-1.0

-2.0
Hypothetical Data
Advantage of females
NAI Data

-3.0

Age in Years (6 through 17)

Fig. 2. Mean gender differences in IQ scores as predicted by Lynn (2002) and generated by the empirical NNAT NAI scores.
 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260 257

Table 3
NNAT NAI scores for males and females by Lynne’s age groups
Lynn’s age groups Male Female NAI diff d-ratio t
Mean S.D. N Mean S.D. N
6–9 100.2 16 14,468 100.2 15.6 14,668 0.05 0.000 0.3
10–13 100.0 16.5 14,273 100.2 15.6 14,443 0.25 0.012 1.3
15–17 99.1 17 5681 99.1 15.4 5940 0.03 0.000 0.9
* = p b .05; ** = p b .01; *** = p b .001.
NAI diff = gender differences in NNAT NAI scores.

and geometric designs interrelated through spatial or Grades 7–9; Level G, Grades 10–12. Each level con-
logical organization. Each item within the NNAT is tains eight items common from both the adjacent higher
similar in that the child must realize the relationship and lower levels as well as unique items. The shared
between the parts of the matrix to successfully solve the items were used to develop a continuous scaled score
problem. The NNAT was designed so that it does not across the entire standardization sample.
require the child to read, write, or speak and that the A Nonverbal Ability Index (NAI) standard score
directions require minimal verbal comprehension. The (mean of 100 and S.D. of 15) is converted from the
test is organized into 38 dichotomously scored items in child’s NNAT raw score through an intermediate Rasch
each of seven levels. Each level of NNAT includes value called a Scaled Score. The Scaled Scores for all
items selected to be appropriate for children of different ages are centered on Level D (Grades 3–4). The appro-
grades and ages to maximize the range of ability that priate equating constant was then added to the spring
could be assessed and to achieve good reliability. The standardization Rasch item difficulties of each level to
KR-20 internal reliability coefficients for the NNAT by produce a continuous Rasch ability scale across all
grade found in Naglieri (1997) range from .83 to .93 levels of the test. Thus, each child’s raw score is con-
(median internal reliability across all levels is .87).The verted to a Scaled Score (Rasch value) based on the
seven levels, or forms of the NNAT, and corresponding NNAT level administered and converted to a standard
grades for which they are intended are as follows: Level score with a mean of 100 (S.D. = 15) based on the age
A, kindergarten; Level B, Grade 1; Level C, Grade 2; of the child. The two scores provide different perspec-
Level D, Grades 3–4; Level E, Grades 5–6; Level F, tives from which to understand children’s performance

102
Mean NNAT NAI Score

100

98

Gender
Male
Female
96

6 7 8 9 10 11 12 13 14 15 16 17
Age Groups in Years

Fig. 3. Mean NNAT NAI scores across age groups (6 to 17 year olds) (the boxes identify the critical time periods according to Lynn’s predictions).
258 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260

Table 4
NNAT scaled scores for males and females by NNAT levels
NNAT level Males Females d-ratio SS diff t
Mean S.D. N Mean S.D. N
A 541.7 38.5 2912 537.7 40.3 2803 0.0014 4.0 3.9***
B 568.5 37.9 3412 570.4 37.4 3384 0.0006 1.9 2.1*
C 585.7 34.2 4044 583.0 35.6 4068 0.0007 2.7 3.5***
D 617.3 40.3 8016 614.8 37.3 7984 0.0003 2.4 4.0***
E 630.7 39.4 7716 632.2 37.2 7556 0.0002 1.5 2.4*
F 648.3 41.0 8878 648.7 38.4 9286 0.0001 0.5 0.8
G 661.4 44.0 4656 658.7 38.0 5065 0.0006 2.7 3.3**
* = p b .05; ** = p b .01; *** = p b .001.
SS diff = gender differences in NNAT scaled scores.

on the NNAT. The Scaled Score is useful for exam- based on the average standard deviations. Effect sizes
ining developmental changes across time (average were evaluated according to Cohen’s criteria (1988)
scores are lower for younger children and higher for for small, medium, and large effects (d = .20, .50, and
older children) while the NAI score is useful for .80), respectively.
comparing children based on their score relative to a Two factorial ANOVAs were computed for the NAI
specific age cohort (the mean score is 100 and S.D. 15 scores. The first one examined the NAI scores as a
for all age groups). For more information, see Naglieri function of gender and NNAT levels. Table 2 contains
(l997). the means and standard deviations of the ages and the
NAI scores for males and females across the NNAT
2. Results levels. Fig. 1 illustrates the means across NNAT levels.
A significant interaction effect was found ( F [6,
Examination of the differences between genders 79,766] = 7.0, p b .001), which can be attributed to the
was conducted using two methods, factorial univariate alternating differences in the NAI scores between males
analyses of variance (ANOVA) and d-ratios (Cohen, and females across NNAT levels. The NAI score gender
1988). Dependent variables were either the NAI scores differences ranged from a 1.1 point advantage for males
or the NNAT scaled scores. The d-ratio is an expres- at NNAT level A (mean age = 6.1, S.D. = 0.4) to a 1.3
sion of the difference between the means in S.D. units point advantage for females at NNAT level B (mean

680

660

640
Mean NNAT Scaled Scores

620

Male
600
Female

580

560

540

520
A B C D E F G

NNAT Levels

Fig. 4. Mean NNAT scaled scores across NNAT levels.

 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260 259

age = 7.1, S.D. = 0.5). No significant NAI differences the genders for NNAT levels A (advantage of males,
were found for NNAT levels C (mean age = 8.0, corresponding to the age of 6) and B (advantage of
S.D. = 0.5) and D (mean age = 9.6, S.D. = 0.8). Despite females, corresponding to the age of 7). Despite the
the statistical significance of these differences the results statistical significance, the NAI differences were about
of the d-ratios indicated that there were very small dis- one point and the d-ratios were quite small (b .09). We
crepancies between the mean NAI scores earned by concluded that Lynn’s prediction of no differences at
males and females. The largest d-ratio was .08, consid- the 6 to 9 age span was supported.
erably smaller than the .2 needed for designation as small The second hypothesis that females should be one
by Cohen (1988). Fig. 2 suggests that at the youngest IQ point ahead of males between the ages of 10 and 13
ages (6 and 7 years) NNAT data were not consistent with was also somewhat consistent with our findings. Al-
Lynn’s predictions; at the middle and upper ages there though the differences between the genders at Level D
were similarities between Lynn’s predicted and empiri- were not significant, Levels E and F showed differ-
cally established differences in trend, however not in ences in favor of females that were about 3/4 of a
effect size or in magnitude of differences. point. But again, despite the statistical significance of
The second factorial ANOVA compared NAI scores these findings, the d-ratios were very small (b.07). We
as a function of gender and age groups as defined by concluded that Lynn’s prediction was supported but the
Lynn (2002). Table 3 shows NAI means and standard size of the differences was minor and the effect size
deviations for males and females by the 6–9, 10–13, tiny.
and 15–17 year age groups described by Lynn (2002). The third hypothesis that males should be 2.4 points
Only the main effect for age groups was found to be ahead of females between the ages of 15 and 16 was,
statistically significant ( F [2, 69,467] = 33.4, p b .05). again, somewhat consistent with our findings. Although
Fig. 3 depicts the mean NAI scores across age groups the differences between the genders at Level G were
(including the 14 year olds, which were not explicitly significant, the difference was about 3/4 of a point and
included in Lynn’s, 2002 model). No significant differ- the d-ratio was minuscule (d = .04). We concluded that
ences between the genders were found. Lynn’s prediction of a difference was supported but at
The final factorial ANOVA examined the NNAT much less than 2.4 IQ points.
scaled scores as a function of the factors gender and Finally, we examined the change in scores across the
NNAT levels. Table 4 shows the means and standard age groups using the NNAT scaled scores. Lynn’s
deviations for males and females by NNAT levels. A prediction that mental growth decelerates for females
significant interaction effect ( F [6, 79,766] = 9.1, relative to males was not supported by the trajectory of
p b .001) and a main effect for NNAT levels was the data. Additionally, whereas Lynn predicted that
found ( F [6, 79,766] = 1182.1, p b .001) for the NNAT females lose their advantage from 15–16 onward and
scaled scores, but no significant effect for gender. males begin to catch up and overtake females was not
NNAT scaled score means and S.D.s were remarkably confirmed (Fig. 4).
similar across the seven levels as shown in Fig. 4 In summary, although the NNAT data were partially
indicating very similar rates of growth across the ages. consistent with Lynn’s developmental theory of gender
differences, the discrepancies between the genders were
3. Discussion so small as to render them inconsequential. Statistical
differences were found due to the large sample size and
The overall aim of this study was to test Lynn’s consequent statistical power. Importantly, the d-ratios
(2002) developmental theory of gender differences on indicated that those statistical differences that were
progressive matrices using a large representative sam- found were minute and may have little or no practical
ple of children in the US. In general, we found mixed importance. The data provided in this study suggest that
support for Lynn’s developmental theory. The first of when using the NNAT (Naglieri, 1997) meaningful
Lynn’s predictions was that there would be no sex differences were not found. The differences that were
differences between the ages of 6 and 9. When the detected were small and vacillated between the genders
analysis was conducted for all children ages 6 to 9 and, as suggested by Mackintosh (1998), Lynn’s (2002)
combined we found support for Lynn’s hypothesis. suggestion of a bmale advantage of 5.5 points seems a
However, when the analysis was conducted according serious over-estimateQ (p. 538). We conclude as Mack-
to NNTA levels within the 6 to 9 year olds our findings intosh (1998) did that Lynn’s (2002) assertion of sex
were inconsistent with this expectation. We found very differences in general intelligence as measured by pro-
small but statistically significant differences between gressive matrices was not supported.
260 J. Rojahn, J.A. Naglieri / Intelligence 34 (2006) 253–260

This study like most has limitations that should be Cohen, J. (1988). Statistical power analysis for the behavioral
considered. First, Lynn’s (2002) study involved Raven’s sciences (2nd ed.). San Diego, CA7 Academic Press.
Deaux, K. (1984). From individual differences to social categories:
Progressive matrices which are similar to but not iden- analysis of a decade’s research on gender. American Psychologist,
tical to those developed by Naglieri (1997). Raven’s 39, 105 – 116.
and Naglieri’s progressive matrices are different in Fennema, E., & Sherman, J. (1977). Sex-related differences in math-
appearance, reliability, and standardization sample ematics achievement, spatial visualization, and affective factors.
American Educational Research Journal, 14, 51 – 71.
characteristics (Bracken & Naglieri, 2003) but have
Geary, D. C. (1989). A model for representing gender differences in
been shown to yield results that are highly correlated the pattern of cognitive abilities. American Psychologist, 44,
(see Naglieri, 1985, 2003). It is possible that differ- 1155 – 1156.
ences between the research utilized by Lynn (2002) Geary, D. C. (1994). Children’s mathematical development: research
and ours resulted from the use of different tests but this and practical applications. Washington, DC7 American Psycho-
seems unlikely. Another limitation is that, as Mackin- logical Association.
Geary, D. C. (1996). Sexual selection and sex differences in mathe-
tosh (1998) noted, if the definition of intelligence matical abilities. Behavioral and Brain Sciences, 19, 229 – 247.
changes different results may be found. Researchers Halpern, D. F. (1986). A different answer to the question, bDo sex-
have found sex differences when different methods of related differences in spatial abilities exist?Q American Psycholo-
measuring ability were employed (e.g., Halpern, 1997) gist, 41, 1014 – 1015.
and when different ways of defining intelligence were Halpern, D. F. (1989). The disappearance of cognitive gender differ-
ences: what you see depends on where you look. American
used. Psychologist, 44, 1156 – 1158.
Naglieri and Rojahn (2001) studied sex differences Halpern, D. F. (1997). Sex differences in intelligence. American
in a large sample of children in the US aged 5–17 years Psychologist, 52, 1091 – 1102.
and found some important differences when using the Hyde, J. S., Fennema, E., & Lamon, S. J. (1990). Gender differences
PASS theory to define intelligence as operationalized in mathematics performance: a meta-analysis. Psychological Bul-
letin, 107, 139 – 155.
by the Cognitive Assessment System (CAS; Naglieri & Hyde, J. s., & Linn, M. C. (1988). Gender differences in verbal
Das, 1997). Girls outperformed boys between the ages ability: a meta-analysis. Psychological Bulletin, 104, 53 – 69.
of 5 and 17 years on measures of cognitive processing Linn, M. C., & Peterson, A. C. (1985). Emergence and characteriza-
included in the Planning Scale. The difference of about tion of sex differences in spatial ability: a meta-analysis. Child
5 points (a .33 d-ratio) was consistent with previous Development, 56, 1479 – 1498.
Lynn, R. (2002). Sex differences on the progressive matrices among
research (Bardos, Naglieri, & Prewett, 1992; Warrick & 15–16 year olds: some data from South Africa. Personality and
Naglieri, 1993). Additionally, girls were better than Individual Differences, 33, 669 – 673.
boys on measures of Attention included in the CAS Lynn, R., & Irwing, P. (2004). Sex difference on progressive matrices:
by about the same amount (d-ratio = .35). In contrast a meta-analysis. Intelligence, 32, 481 – 498.
Maccoby, E. E., & Jacklin, C. (1974). The psychology of sex differ-
boys and girls differed minimally on measures of Si-
ences. Stanford, CA7 Stanford University Press.
multaneous (d-ratio = .01) (which includes a measure of Mackintosh, N. J. (1998). Reply to Lynn. Journal of Biosocial
progressive matrices) and Successive (d-ratio = .08) Science, 30, 533 – 539.
cognitive processing. Like McHough et al. (1986) and McHough, M. C., Koeske, R. D., & Frieze, I. H. (1986). Issues to
Geary (1989) we further suggest that greater insights consider in conducting nonsexist psychological research: a guide
into sex differences could be obtained using a theory for researchers. American Psychologist, 41, 879 – 890.
Naglieri, J. A. (1985). Matrix analogies test expanded form. San
that is based on cognitive and neuropsychological the- Antonio7 Harcourt Assessment Company.
ories like PASS rather than those based on the familiar Naglieri, J. A. (1997). Naglieri Nonverbal Ability Test-Multilevel
verbal, quantitative, nonverbal content based conceptu- Form. San Antonio7 Harcourt Assessment Company.
alization. We therefore encourage researchers to exam- Naglieri, J. A. (2003). Naglieri Nonverbal Ability Test individual
ine sex differences using methods that extend beyond administration. San Antonio7 Harcourt Assessment Company.
Naglieri, J. A., & Das, J. P. (1997). Cognitive assessment system.
the general intelligence approach. Itasca7 Riverside Publishing Company.
Naglieri, J. A., & Rojahn, J. (2001). Gender differences in planning,
References attention, simultaneous, and successive (PASS) cognitive process-
es and achievement. Journal of Educational Psychology, 93,
Bardos, A. N., Naglieri, J. A., & Prewett, P. N. (1992). Sex differ- 430 – 437.
ences in planning, attention, simultaneous, and successive cogni- Voyer, D., Voyer, S., & Bryden, M. P. (1995). Magnitude of sex
tive processes. Journal of School Psychology, 30, 293 – 305. differences in spatial Abilities: a meta-analysis and consideration
Bracken, B. A., & Naglieri, J. A. (2003). Assessing diverse populations of critical variables. Psychological Bulletin, 117, 250 – 270.
with nonverbal tests of general intelligence. In C. R. Reynolds, & Warrick, P. D., & Naglieri, J. A. (1993). Gender differences in
R. W. Kamphaus (Eds.), Handbook of psychological and educa- planning, attention, simultaneous, and successive cognitive pro-
tional assessment of children (2nd ed.). New York7 Guilford. cesses. Journal of Educational Psychology, 85, 693 – 701.