Geary, D.C. (1996). Sexual selection and sex differences in
mathematical abilities. Behavioral and Brain Sciences 19 (2): 229-284.
The final published draft of the target article, commentaries and Author's
Response are currently available only in paper.
For information about
subscribing or purchasing offprints of the published version, with commentaries
and author's response, write to: journals_subscriptions@cup.org
(North America) or journals_marketing@cup.cam.ac.uk
(All other countries).
SEXUAL SELECTION AND SEX DIFFERENCES IN MATHEMATICAL ABILITIES
David C. Geary
Department of Psychology
210 McAlester Hall
University of Missouri at Columbia
Columbia, MO 65211
psycorie@mizzou1.missouri.edu
Keywords
sexual selection, sex differences, mathematical ability,
spatial ability, sex-role stereotypes, social style
Abstract
The principles of sexual selection were used as an
organizing framework for interpreting cross-national patterns of sex
differences in mathematical abilities. Cross-national studies suggest that
there are no sex differences in biologically primary mathematical abilities,
that is, for those mathematical abilities that are found pan-culturally, in
nonhuman primates, and show moderate heritability estimates. Sex differences in
several biologically secondary mathematical domains (i.e., those that emerge
primarily in school) are found throughout the industrialized world. In
particular, males consistently outperform females in the solving of
mathematical word problems and in geometry. Sexual selection and any associated
proximate mechanisms (e.g., sex hormones) appear to influence these sex
differences in mathematical performance indirectly. First, sexual selection
appears to have resulted in the greater elaboration of the neurocognitive
systems that support navigation in 3-dimensional space in males than in
females. Knowledge implicit in these systems appears to reflect an
understanding of basic Euclidean geometry, and thus appears to be one source of
the male advantage in geometry. Males also co-opt these spatial systems in
problem-solving situations more readily than females, which provides males with
an advantage in word problems and geometry. Moreover, sex differences in social
styles and interests, which also appear to be related, in part, to sexual
selection, result in sex differences in engagement in mathematics-related
activities, which further increases the male advantage in certain mathematical
domains. A model that integrates these biological influences with sociocultural
influences on the sex differences in mathematical performance is
presented.
Introduction
The principles of sexual selection provide a useful
theoretical framework for understanding human sex differences in certain social
and sexual behaviors (Buss 1995), as well as sex differences in certain
cognitive domains (Gaulin 1992; Kimura & Hampson 1994). In this article,
sexual selection provides the primary theoretical context for examining the
cross-national pattern of sex differences in mathematical abilities. The
argument is not that complex mathematical abilities, or any associated sex
differences, have been directly shaped by evolutionary pressures, but rather
that sexual selection appears to have directly shaped the social and cognitive
styles of males and females, which, in turn, influence mathematical development
and performance and contribute to sex differences in certain mathematical
domains.
In order to fully consider the potential influence of evolutionary pressures
on mathematical abilities, a general framework for making inferences about the
relative degree of biological and cultural influences on cognition is needed
(Geary 1995). This is because children's mathematical development occurs
primarily in school settings, and, as a result, the assessment of mathematical
performance necessarily reflects some cultural influences. In fact, the
administration of all psychometric tests, such as the Scholastic Achievement
Test (SAT, formerly the Scholastic Aptitude Test) necessarily reflects
culturally-taught skills, such as reading. Although performance on these
measures must to a large degree reflect schooling and other sociocultural
influences, this does not preclude more primary biological influences on
psychometric test performance. In the first section below, a framework for
making inferences about forms of cognition that are largely influenced by
biological factors and forms of cognition that are more culturally-specific is
presented; these respective forms of cognition are called biologically-primary
and biologically-secondary (Geary 1995; Rozin 1976).
On the basis of this framework, a systematic assessment of sex differences
in mathematical abilities requires a consideration of whether the differences
in question are evident for biologically- primary or biologically-secondary
mathematical domains, which are defined in Section 2. If sexual selection were
directly related to sex differences in mathematical abilities, then any such
sex difference should be most evident for biologically-primary mathematical
domains, those least affected by sociocultural influences (e.g., schooling). In
contrast, if there are no sex differences in biologically-primary mathematical
domains but consistent sex differences in secondary mathematical domains, then
there are two general potential sources of these sex differences. The first and
most obvious is a difference in the schooling of boys and girls, given that
school is the primary cultural context within which secondary mathematical
abilities appear to emerge (Geary 1995). The second and less obvious source of
sex differences in secondary mathematical domains is the secondary effects of
sexual selection on the cognitive and social styles of boys and girls. Section
3 presents a framework for considering any such secondary influences of sexual
selection on sex differences in mathematical abilities.
The present article is not the first to argue that cognitive sex differences
in general and sex differences in mathematics in particular have biological
origins. Indeed, there are many theoretical reviews in the literature on the
potential biological influences on cognitive sex differences (e.g., Benbow,
1988; McGee 1979; McGuinness & Pribram 1979; Ounsted & Taylor 1972;
Sherman 1967), and many well reasoned arguments that any such differences are
largely the result of the differential socialization of boys and girls (e.g.,
Fennema & Sherman 1977; Sherman 1980). Benbow, for instance, presented
evidence suggesting that the greater number of males than females at the upper
end of the distribution of SAT scores reflects, at least in part, a sex
difference in the functional organization of the left- and right-hemisphere.
McGee argued that a sex difference in certain forms of spatial cognition has
biological origins and contributes to sex differences in certain mathematical
areas. Sherman (1980, 1981), in contrast, presented evidence suggesting that
the sex difference in mathematics was primarily related to the greater
mathematical confidence of boys than girls. On this view, the greater
mathematical confidence of boys results in a sex difference in mathematical
activities (e.g., course taking), favoring boys, and ultimately a male
advantage in mathematical performance. In fact, the emergence of any complex
cognitive skill almost certainly reflects an interaction between biologically
based differences in the types of activities that boys and girls prefer to
engage in and the environments made available to them by parents and by peers
(McGuinness & Pribram 1979; Scarr & McCarthy 1983).
The present article builds on and extends previous theoretical treatments of
the source and nature of sex differences in mathematics in several ways. First,
as noted above, potential biological influences on mathematical sex differences
are considered explicitly within the context of sexual selection. In
particular, sexual selection is used as a theoretical context for examining sex
differences in those cognitive and social domains that appear to influence
mathematical development. As an example, it is argued in section 3.1 that
sexual selection has resulted in a sex difference in the social preferences of
males and females, with males being more object oriented and females being more
person oriented, which, in turn, influences math-intensive career choices and
mathematics course taking. Although the position that the sex difference in
person/object orientations influences the sex difference in mathematical
development is not novel (Chipman & Thomas 1985; Thorndike 1911; McGuinness
& Pribram 1979), the present article contributes to this literature by
attempting to place the potential relation between person/object orientations
and sex differences in mathematical abilities in the wider context of human
evolution.
Second, the article provides a theoretical framework for examining how
evolved cognitive abilities might be manifested in evolutionarily novel
contexts (such as schools). In the section below, it is argued that there are
at least two ways in which biologically primary forms of cognition, such as
certain spatial abilities, can be expressed in biologically secondary domains,
such as complex mathematics. By more fully articulating the potential relation
between primary spatial abilities and secondary mathematical abilities, this
framework expands on McGee's (1979) position that the sex difference in certain
mathematical areas is secondary to more primary sex differences in certain
spatial abilities. Finally, the review of sex differences in mathematical
abilities is more comprehensive in many ways than other treatments of this
issue (e.g., Benbow 1988; Hyde, Fennema & Lamon 1990). Specifically, the
review of mathematical sex differences provides a greater emphasis on
cross-national patterns than have most previous reviews and includes a review of
potential sex differences in very early numerical competencies, those that
appear to be biologically primary.
1. Evolution, culture and cognition
In some respects, all forms of cognition are supported by neurocognitive
systems that have evolved to serve some function or functions related to
reproduction or survival. These basic neurocognitive systems appear to be found
in human beings throughout the world, and appear to support the emergence of
species-typical cognitive domains, such as language (Pinker & Bloom 1990;
Witelson 1987). However, some forms of cognition, such as reading, emerge in
some cultures and not others. This pattern suggests that while the emergence of
some domains of cognition are driven largely by biological influences, other domains
emerge only with specialized cultural practices and institutions, such as
schools, that are designed to facilitate the acquisition of these cognitive
skills in children (Geary 1995). Thus, when assessing the source of group or
individual differences in cognitive abilities, it seems necessary to consider
whether the ability in question is part of a species-typical
biologically-primary cognitive domain, or whether the ability in question is
culturally-specific, and therefore biologically-secondary.
Although biologically-secondary abilities appear to emerge only in specific
cultural contexts, at least for large segments of any given population, they
must perforce be supported by neurocognitive systems that have evolved to
support primary abilities. Indeed, these culture-specific abilities might
involve the co-optation of biologically-primary neurocognitive systems or
access to knowledge implicit in these systems for purposes other than the
original evolution-based function (S. J. Gould & Vrba 1982; Rozin 1976).
The basic premise is that in terms of children's cognitive development, the
interface between culture and biology involves the co-optation of highly
specialized neurocognitive systems to meet culturally-relevant goals. In this
section, a basic framework for distinguishing between primary and secondary
abilities is presented (a more complete presentation can be found in Geary
1995).
1.1 Biologically-primary and biologically-secondary abilities. One way to
organize our understanding of both biologically-primary and
biologically-secondary forms of cognition is to hierarchically organize them
into domains, abilities, and neurocognitive systems. Domains, such as language
or arithmetic, represent constellations of more specialized abilities, such as
language comprehension or counting. Individual abilities, in turn, are
supported by neurocognitive systems, and might consist of three types of
cognitive competencies; goal structures, procedural skills, and conceptual
knowledge (Gelman 1993; Siegler & Crowley 1994). The goal of counting, for
instance, is to determine the number of items in a set of objects. Counting is
achieved by means of procedures, such as the act of pointing to each object as
it is counted. Pointing helps the child to keep track of which items have been
counted and which items still need to be counted (Gelman & Gallistel 1978).
Counting behavior, in turn, is constrained by conceptual knowledge (or skeletal
principles for the initial emergence of primary domains), so that, for
instance, each object is pointed at or counted only once.
Although there are similarities in the cognitive competencies associated
with primary and secondary abilities (Siegler & Crowley 1994), there also
appears to be a number of important differences, two of which are relevant to
the current discussion. First, inherent in the neurocognitive systems that
support primary abilities is a system of skeletal principles (Gelman 1990).
Skeletal principles provide the scaffolding upon which goal structures and
procedural and conceptual competencies emerge. For instance, one skeletal
principle that appears to be associated with biologically-primary counting
abilities is "one-one correspondence" (Gelman & Gallistel 1978).
Here, as noted above, the counting behavior of human children and even the
common chimpanzee (Pan troglodytes) is constrained by an implicit understanding
that each item in an enumerated set must be tagged once and only once (Boysen
& Berntson 1989; Starkey 1992). Second, it appears, at least initially,
that the knowledge that is associated with primary domains is implicit. That
is, the behavior of children appears to be constrained by skeletal principles,
but children cannot articulate these principles (Geary 1995; Gelman 1993).
While the initial structures for the cognitive competencies that might be
associated with primary abilities appear to be inherent, the goal structures as
well as procedural and conceptual competencies for secondary abilities are
likely to be induced or learned from other people (e.g., teachers) and might
emerge, at least in part, from primary abilities. For the latter, there appear
to be two possibilities. As noted above, the first involves the co- optation of
the neurocognitive systems that support primary abilities. Second, knowledge
that is implicit in the skeletal principles of primary abilities can be made
explicit and used in ways unrelated to the evolution of these principles (Rozin
1976).
For a germane example, consider the possibility that the development of
geometry as a formal discipline involved, at least in part, the co-optation of
the neurocognitive systems that have evolved to support navigation in the
three-dimensional physical universe and access to the associated implicit
knowledge. Arguably all terrestrial species, even invertebrates (e.g.,
insects), have cognitive systems that enable navigation in three-dimensional
space (Gallistel 1990; J. L. Gould 1986; Landau et al. 1981; Shepard 1994).
Cheng and Gallistel (1984), for instance, showed that laboratory rats appear to
develop a "Euclidean representation of space for navigational
purposes" (p. 420), and, as a result, are sensitive to changes in basic
representation, such as shape, and metric, such as angle. Implicit in the
functioning of the associated neurocognitive systems is a basic understanding
of geometric relationships amongst objects in the physical universe. So, for
example, even the behavior of the common honey bee (Apis mellifera) reflects an
implicit understanding that the fastest way to move from one location to another
is to fly in a straight line (J. L. Gould 1986).
Even though an implicit understanding of geometric relationships appears to
be a feature of the neurocognitive systems that support habitat representation
and navigation, this does not mean that individuals have an explicit
understanding of the formal principles of Euclidean geometry. Rather, the
development of geometry as a formal discipline might have been initially based
on early geometer's access to the knowledge that is implicit in the systems
that support habitat navigation. In keeping with this position, in the
development of formal geometry, Euclid apparently "started with what he
thought were self-evident truths and then proceeded to prove all the rest by
logic" (West et al. 1982; p. 220). The implicit understanding, or
"self-evident truth," that the fastest way to get from one place to
another is to go "as the crow flies," was made explicit in the formal
Euclidean postulate, "a line can be drawn from any point to any point"
(In Euclidean geometry, a line is a straight line)" (West et al 1982; p.
221). The former appears to represent implicit, biologically-primary knowledge
(i.e., a skeletal principle) associated with the neurocognitive systems that
support habitat navigation, whereas the latter represents the explicit
formalization of this knowledge as part of the formal discipline of geometry.
Moreover, although the neurocognitive systems that support habitat
navigation appear to have evolved in order to enable movement in the physical
universe (Shepard, 1994), they can also be co-opted or used for many other
purposes. For an example of what I would call cognitive co-optation, consider
the use of spatial representations to aid in the solving of arithmetic word
problems. Lewis and Mayer (1987) showed that word problems that involve the
relative comparison of two quantities are especially difficult to solve. For
instance, consider the following compare problem from Geary (1994): "Amy
has two candies. She has one candy less than Mary. How many candies does Mary
have?" The solution of this problem requires only simple addition, that
is, 2+1. However, many adults and children often subtract rather than add to
solve this type of problem; the keyword "less" appears to prompt
subtraction rather than addition. Moreover, the structure of the second
sentence (i.e., Mary is the object rather than the subject of the sentence)
also leads individuals to conclude that Amy has more candy than Mary. Lewis
(1989) showed that one way to reduce the frequency of errors that are common
with these types of relational statements is to diagram (i.e., spatially
represent) the relative quantities in the statements (in this example the
number of Amy's and Mary's candies).
It is very unlikely that the evolution of spatial abilities was in any way
related to the solving of mathematical word problems. Nevertheless, spatial
representations of mathematical relationships are used, that is co-opted, by
some people to aid in the solving of such problems (Johnson 1984). More
important, there appear to be differences in the ease with which these systems
can be used for their apparent evolution-based functions and co-opted tasks.
The use of spatial systems for moving about in one's surroundings or developing
cognitive maps of one's surroundings appears to occur more or less
automatically (Landau et al. 1981). However, most people need to be taught,
typically in school, how to use spatial representations to solve, for instance,
mathematical word problems (Lewis 1989). In short, the formal step-by-step
procedures that can be used to spatially represent mathematical relationships
are secondary with respect to the evolution of spatial cognition--these systems
did not evolve for this purpose but nevertheless can be used for this purpose.
In other words, the practices that occur within some cultural institutions,
such as schools, can, in a sense, create cognitive skills that otherwise would
not emerge. Any such culture-based skill must perforce be built upon more
primary forms of cognition (Geary 1995).
In general, I am arguing that the neurocognitive systems that have more
likely evolved to support movement in the three- dimensional physical universe
(Gaulin 1992; Shepard 1994) can be adapted by human beings for purposes other than
the original evolution-based function (Rozin 1976). With regard to mathematics,
there are two resulting predictions. The first is that the relationship between
spatial and mathematical abilities should be rather selective. Specifically,
the prediction is that measures of 3-dimensional spatial abilities should be
more strongly related to abilities in the domain of Euclidean geometry (e.g.,
an intuitive understanding of basic geometric relationships and graphing
relationships in Euclidean space) than to other forms of mathematics (e.g.,
solving algebraic expressions). The prediction is for tests of 3-dimensional
spatial abilities because the neurocognitive systems that appear to support
habitat navigation evolved in the 3-dimensional physical universe (Shepard
1994). Second, in addition to Euclidean geometry, a relationship between
spatial abilities and mathematical information that can be represented
spatially is predicted. Example of such areas would include the solving of word
problems, tables, graphs, etc.
From this perspective, a thorough assessment of sex differences in
mathematical abilities should be based on a consideration of whether the
abilities in question are likely to be biologically primary or biologically
secondary. For secondary domains, we should consider whether any differences
that might emerge could possibly involve the co-optation of biologically
primary cognitive systems or access to knowledge implicit in these systems.
Thus, as noted earlier, the sections below provide a brief overview of
potential primary and secondary mathematical domains. Moreover, given the
potential for co-optation, potential biologically-primary sex differences in
cognitive and social styles that might influence mathematical development are
considered in Section 3.
2. Potential primary and secondary mathematical abilities
The following subsections present a brief overview of potential
biologically-primary and biologically-secondary mathematical domains.
2.1 Biologically-primary mathematical abilities. There is some evidence for
the pan-cultural existence of a biologically-primary numerical domain which
consists of at least four numerical abilities; numerosity, ordinality,
counting, and simple arithmetic (see Geary 1995). A brief description of these
abilities is provided in Table 1.
Numerosity represents the ability to quickly determine the quantity of a set
of 3 to 4 items without the use of counting or estimating. The ability to
quickly and accurately make numerosity judgments is evident in human infants in
the first week of life, as well as in the laboratory rat, an African grey
parrot (Psittacus erithacus), and the common chimpanzee (Pan troglodytes)
(Antell & Keating 1983; Boysen & Berntson 1989; Davis & Memmott
1982; Pepperberg 1987). Moreover, numerosity judgments appear to be based on an
abstract representation of quantity rather than on modality-specific processes,
as these judgments can be made by human infants for auditory and visual
information (Starkey et al. 1983; 1990). In support of this view is the finding
that certain cells in the parietal-occipital cortex of the cat are selectively
responsive to small quantities, whether the quantities are presented in the
visual, auditory, or tactile modalities (Thompson et al. 1970).
A sensitivity to ordinal relationships, for example, that 3 is more than 2
and 2 is more than 1, is evident in 18-month-old infants (Cooper 1984; Strauss
& Curtis 1984). Moreover, well- controlled studies have shown that nonhuman
primates are able to make very precise ordinal judgments (Boysen 1993; Washburn
& Rumbaugh 1991). For instance, after learning the quantity associated with
specific Arabic numbers, a rhesus monkey (Macaca mulatta) named Abel could
correctly choose the larger of two Arabic numbers more than 88% of the time
(Washburn & Rumbaugh 1991). More important, Abel could choose the larger of
two previously unpaired numbers more than 70% of the time.
Counting appears to be a pan-cultural human activity that is, at least
initially, supported by a set of skeletal principles, or implicit knowledge,
before children learn to use number words (Crump 1990; Geary 1994; Gelman &
Gallistel 1978; Saxe 1982; Starkey 1992; Zaslavsky 1973). As noted earlier, one
basic principle that appears to constrain counting behavior is one-one correspondence
(Gelman & Gallistel 1978). Implicit knowledge of this skeletal principle is
reflected in the act of counting when each item is tagged (e.g., with a number
word) or pointed to once and only once (Gelman & Gallistel 1978). Some
human infants as young as 18-months-of-age are able to use some form of tag in
order to determine the numerosity of sets of up to three items (Starkey 1992),
as can the common chimpanzee (Boysen 1993; Rumbaugh & Washburn 1993).
Other research suggests that 5-month-old infants are aware of the effects
that the addition and subtraction of one item has on the quantity of a small
set of items (Simon, Hespos & Rochat in press; Wynn 1992). Similar results
have been reported for 18- month-olds (Starkey 1992) and for the common chimpanzee
(Boysen & Berntson 1989). Moreover, these competencies in simple arithmetic
appear to be qualitatively similar in the chimpanzee and human infants and
young children (Gallistel & Gelman 1992). Preschool children appear to be
able to add quantities up to and including three items using some form of
preverbal counting, while at least some chimpanzees appear to be able to add
items up to and including four items also by means of preverbal counting
(Boysen & Berntson 1989; Starkey 1992).
Finally, psychometric and behavioral genetic studies support the argument
that some numerical and arithmetical skills are biologically primary and
cluster together. In the psychometric literature, human cognitive abilities are
hierarchically organized. At the top are abilities that span many or all more
specific lower- order ability domains (Cattell 1963; Gustafsson 1984; Horn
& Cattell 1966; Spearman 1927; Thurstone & Thurstone 1941; Vernon
1965). These general abilities are often subsumed under the terms general intelligence
(g) or fluid and crystallized intelligence and appear to represent cognitive
skills, such as speed of processing or working memory, that support performance
in many cognitive domains (e.g., Kyllonen & Christal 1990; Vernon 1983).
From an evolutionary perspective, g might index, at least in part, individual
differences in the ability to co-opt primary abilities, that is, use these
abilities for purposes unrelated to their evolutionary functions (e.g.,
adapting primary abilities for school learning; Geary 1995). Either way, of
particular importance to this discussion is the emergence of lower-order
numerical and mathematical factors, or ability domains.
For five-year-olds, tests that assess number knowledge, memory for numbers,
as well as basic counting and arithmetic skills cluster together and define a
Numerical Facility factor (Osborne & Lindsey 1967). In fact, the Numerical
Facility factor is one of the most stable lower-order factors ever identified
through decades of psychometric research (e.g., Coombs 1941; Thurstone 1938;
Thurstone & Thurstone 1941), and has been found throughout the lifespan, as
well as with studies of American, Chinese, and Filipino students (Geary 1994;
Guthrie 1963; Vandenberg 1959). Even Spearman, an ardent supporter of the view
that individual differences in human abilities are best explained by g stated
that basic arithmetic skills "have much in common over and above...g"
(Spearman 1927; p. 251). Behavioral genetic studies of tests that define the
Numerical Facility factor have yielded heritability estimates of about .5,
suggesting that roughly 1/2 of the variability in at least some components of
arithmetical ability are due to genetic differences across people (Vandenberg,
1962; 1966).
2.2. Biologically-secondary mathematical abilities. The argument that
certain features of counting, number, and arithmetic are biologically primary
should not be taken to mean that all numerical and arithmetical abilities are
biologically primary. In fact, there are many features of counting and
arithmetic that are probably biologically secondary. These features include
skills and knowledge taught by parents (e.g., the names of number words),
concepts that are induced by children during the act of counting (e.g., that
counted objects are usually tagged from left to right), and skills that are
formally taught in school (e.g., the base-10 system, trading, fractions,
multiplication, exponents, etc.) (Briars & Siegler 1984; Fuson 1988; Geary
1994; Ginsburg et al. 1981). Moreover, it is likely that most features of
complex mathematical domains, such as algebra, geometry, and calculus are
biologically- secondary, given that the associated abilities only emerge with
formal education. Of particular relevance to the goal of this article is the
development of geometric and mathematical problem-solving abilities, because
cross-national studies consistently find sex differences in these areas (see
Section 4).
Even though I have argued that basic geometric knowledge might emerge from
the neurocognitive systems that support habitat navigation, many formal
geometric skills, as taught in high school in the United States, are very
likely to represent secondary cognitive abilities. This is because the
knowledge that appears to be implicit in the neurocognitive systems that
support navigation is limited and imprecise (Gallistel 1990). Thus, while
access to this knowledge might facilitate the learning of basic geometric
principles, many other features of geometry probably need to be explicitly
taught, such as geometric proofs, that the sum of the angles in a triangle is
180 degrees, etc.
Mathematical problem-solving abilities are typically assessed by the solving
of arithmetical and algebraic word problems. The solving of mathematical word
problems requires the ability to spatially represent mathematical relations, as
described earlier, as well as the ability to translate word problems into
appropriate equations, and an understanding of how and when to use mathematical
equations (Mayer 1985; Schoenfeld 1985). Psychometric studies suggest that
mathematical problem solving, or mathematical reasoning (as psychometric
researchers call it), is a biologically-secondary cognitive domain. Although a
lower-order Mathematical Reasoning factor has been identified in many psychometric
studies (Dye & Very 1968; Thurstone 1938), this factor does not emerge in
all samples, not even in some samples of college students (e.g., Guthrie 1963).
In fact, a distinct Mathematical Reasoning factor is consistently found only
with groups of older adolescents (i.e., end of high school or early college)
who have taken a lot of mathematics courses (e.g., Very 1967). Thus, unlike the
Numerical Facility factor which has emerged in nearly all studies that have
included arithmetic tests, a Mathematical Reasoning factor emerges only in
samples with prolonged mathematical instruction, suggesting that many
individuals do not easily acquire the competencies that are associated with the
solving of arithmetical and algebraic word problems.
The fact that a Mathematical Reasoning factor does not emerge until
adolescence does not of course preclude biological influences on the
acquisition of the associated skills. Indeed, factor analytic studies suggest
that mathematical reasoning abilities initially emerge from
biologically-primary mathematical abilities and general reasoning abilities,
both of which are partly heritable (Vandenberg 1966). However, by the end of
high school, the abilities subsumed by the Numerical Facility and Mathematical
Reasoning factors are largely independent (Geary 1994). The point is that the
abilities that are subsumed by the Mathematical Reasoning factor only appear to
emerge with sustained mathematical instruction, and are therefore more likely
to represent secondary abilities rather than primary abilities. In other words,
it appears that for most individuals direct instruction (e.g., teaching the use
of diagrams to solve word problems; Lewis 1989) is necessary for the
co-optation of primary abilities and the eventual emergence of a coordinated
system of secondary mathematical abilities.
3. Evolution of sex differences
One potential ultimate source of sex differences in social behavior and
cognition is sexual selection (Darwin 1859, 1871; Buss 1995; Gaulin 1992). For
most mammalian species, sexual selection is thought to operate through
intramale competition and female choice of mating partners. For the present
argument, sexual selection is broadly defined to include characteristics that
directly influence the outcome intramale competition, such as physical
strength, as well as less direct influences on the outcome of any such
competition. On this view, the factors that might have influenced the evolution
of sex differences in certain forms of spatial cognition, for instance, are
included under the broader definition of sexual selection. Similarly, some
influences on the sexual division of labor might also be considered under the
broader definition of sexual selection. Those social and cognitive
characteristics that might contribute to any sex difference in the level of
investment in offspring would be examples of germane features of the sexual
division of labor.
Indeed, in a seminal paper, Trivers (1972) argued that sex differences in
the level of parental investment in offspring "governs the operation of
sexual selection" (p. 141), and, as such, is the ultimate cause of any
associated sex differences. From this perspective, the sex that invests the
least in offspring will show less discriminant mating, and, relative to the
higher investing sex, show more intrasexual competition over access members of
the opposite sex, be physically larger, and have higher mortality rates, among
other things. The higher-investing sex, in contrast, is expected to be much
more discriminating in terms of choosing sexual partners, with discrimination
focusing on a potential partner's physical characteristics or behaviors that
might benefit future offspring. In most mammalian species, particularly
polygynous species, males invest less than females in offspring, and are
physically larger, show more intrasexual competition, and have higher mortality
rates than females (Daly & Wilson 1983). For these species, the sex
difference in investment in offspring is likely to reflect the internal
gestation of offspring, which perforce makes the initial female investment
larger than the initial male investment (see Trivers 1972).
There are a number of findings that indicate that human males and females
differ on the above dimensions (e.g., sexual behavior, physical size), suggesting
that sexual selection has quite likely influenced the physical, social, and
cognitive evolution of males and females. For instance, although most human
mating systems are functionally monogamous, most preliterate societies are
polygynous, that is, a few males have more than one wife, and some males have
no wife (Symons 1979). In keeping with the view that human males are
potentially polygynous (if only in fantasy for most), are consistent sex
differences in sexual behavior. In a large-scale meta-analysis, Oliver and Hyde
(1993) reported that across cohorts (i.e., historical periods) males reported
greater sexual activity and more permissive attitudes towards casual sex than
females. Moreover, human males are, on average, physically larger than human
females, although the magnitude of this difference appears to have decreased
over the course of human evolution (Frayer & Wolpoff 1985), more frequently
engage in extreme forms of intra- and inter-sexual aggression (e.g., homicide;
Daly & Wilson 1988; Wilson & Daly 1985), have higher mortality rates,
and invest less in their offspring, on average, than human females (Furstenberg
& Nord 1985; Stillion 1985). The overall pattern for humans is consistent
with Trivers' (1972) model, as well as with the pattern of sex differences
found in most (i.e., polygynous) mammalian species (Daly & Wilson 1983).
Any related sex differences are likely to be directly associated with
differences in the reproductive strategies of males and females and, as noted
above, to any associated division of labor (Buss & Schmitt 1993;
Eibl-Eibesfeldt 1989; Frayer & Wolpoff 1985; Ghiglieri 1987; Ruff 1987). We
cannot of course recreate human evolution in order to directly test the effects
of sexual selection on the psychological development of males and females.
However, indirect evidence can be accrued. First, sex hormones are likely to be
an important proximate mechanism for the development of any sex differences
associated with sexual selection. Thus, social behaviors and cognitive abilities
that are sensitive to fluctuations in hormonal levels are good candidates for
primary attributes that have been shaped by sexual selection. Second, sex
differences in social behaviors and cognitive abilities that are relatively
insensitive to historical and cultural changes also need to be considered as
potentially related to sexual selection.
Third, it is possible that sex differences in early play patterns influence
later sex differences in cognitive abilities (Serbin & Connor 1979). If sex
differences in these cognitive abilities are related to sexual selection, then
sex differences should be evident in the play patterns that facilitate the
acquisition of these abilities (Geary 1992; McGuinness & Pribram 1979).
Moreover, these play patterns should also be influenced by sex hormones; this
does not preclude social influences as well (MacDonald 1988). Finally, the
associated social and cognitive skills should serve some plausible function
related to reproductive success. With regard to potential sex differences in
mathematical abilities, sex differences in social and cognitive styles need to
be considered.
3.1. Social sex differences. Sex differences for a variety of social and
sexual behaviors might be understood based on the principles of sexual selection
(e.g., mate preferences; Buss 1989). However, this section only provides a
brief consideration of two forms of social behavior, the relative degree of
competitive versus cooperative social styles and the relative degree of object
versus people preferences. A consideration of these two forms of social
behavior is necessary because there is some evidence of sex differences on
these dimensions, and some research that suggests that any such differences
might influence the relative achievement of boys and girls in competitive and
cooperative classroom environments (Peterson & Fennema 1985), as well as
one's relative interest in science and mathematics-related careers (Chipman et
al. 1992). Of course, any such influences would likely be applicable to academic
domains other than mathematics. The goal of this article is to consider these
potential influences within the context of children's mathematical development,
and not to argue that they are exclusively related to mathematics.
As noted in the preceding section, males tend to be more aggressive and
dominance oriented than females in most mammalian species, including humans
(e.g., Daly & Wilson 1983; Pratto, Sidanius & Stallworth 1993). There
are a number of factors that suggest that the male propensity toward
aggression, competitiveness, and dominance seeking has its roots in intramale
competition (i.e., sexual selection). Human males are verbally and physically
more aggressive than females across cultures (Eibl- Eibesfeldt 1989; Rohner
1976). Sex differences in the concern for status and the use of
rough-and-tumble play to establish dominance emerges early in the preschool
years and at about the same age in all cultures that have been studied (Maccoby
1988). A sex difference in the relative degree of competitive versus
cooperative social styles is found in most nonhuman primates as well, although
these differences are also related to kin relationships within social groups
(Ghiglieri 1987; Manson & Wrangham 1991). Moreover, Susman et al. (1987)
found that during adolescence, adrenal and gonadal hormones were related to
aggression and delinquency in human males but not females. As with males in
many other mammalian species, aggression in human males is most extreme (i.e.,
homicide) during the courtship period, that is, late adolescence and young
adulthood (Daly & Wilson 1983). During this time, homicide tends to be one
male killing another male, either for resources (robbery), status, or as a
result of sexual jealousy (Wilson & Daly 1985).
None of this should be taken to mean that the sex difference in competitive
and aggressive behavior is not influenced by child- rearing or other
sociocultural practices. In fact, the expression of many, if not all,
biologically based sex differences are likely to be facultatively influenced by
developmental experiences and contextual factors (Hall 1992). Indeed, the
magnitude of the sex difference in the level of aggression varies across
cohorts, age, context, and culture (Eagly & Steffen 1986; Eibl-Eibesfeldt
1989; Hyde 1984). Sex differences in the level of aggression are rather small
and unimportant in many contexts, such as laboratory settings (Eagly &
Steffen 1986). However, sex differences in competitive and aggressive behavior
are often very large under conditions that represent a threat to the male's
status, or might directly influence reproduction, e.g., sexual jealousy (Daly
& Wilson 1988; Wilson & Daly 1985). The point is, in terms of mean
differences and the ratio of extremely aggressive males to females, sex
differences are expected to be largest in evolutionarily significant contexts,
such as a threat to status, and much smaller in other areas. Nevertheless,
these differences might influence how boys and girls respond to competitive, as
contrasted with cooperative, classroom environments.
The nature of the social relationships of males and females also tends to
differ, across cultures, in other ways that might create a sex difference in
responsiveness to competitive and cooperative classroom environments
(Eibl-Eibesfeldt 1989). During the preschool years and into adulthood, girls'
social styles reflect, relative to boys' styles, a greater concern for
egalitarian relationships, more cooperation, and a greater concern for the
feelings of other group members (usually other girls; Buhrmester & Furman
1987; Maccoby 1988; 1990). In a longitudinal study of personality development,
Block (1993) found rather interesting sex differences in the relation between
the self-esteem and social relationships: "young women with high self-esteem
scores seemed happy, warmly extraverted, and deeply concerned about
interpersonal relationships. Young men with high self esteems seemed
self-focused and defensively critical, uneasy, and unready for a connection
with others (Block 1993; p. 28).
Indeed, there is reason to suspect that a more self-serving personality
provided males with a reproductive advantage over less self-serving males at
certain points in our evolutionary history; that is, this style might be
facultatively expressed in competitive contexts. Betzig (1993), for instance,
described the social structure of humanity's first six civilizations as being
despotic. Across all six civilizations, "powerful men mate with hundreds
of women, pass their power on to a son by one legitimate wife, and take the
lives of men who get in their way" (Betzig, 1993, p. 37). No matter how
socially abhorrent, the reproductive advantage that these high status males
achieved over other males was likely facilitated by striving for and achieving
dominance, and by an insensitivity to the effects that their exploitative
behavior had on the well-being of other human beings, male and female.
Perhaps related to the above are sex differences in the relative degree to
which males and females are object versus people oriented; males appear to be
relatively more object oriented, and females more people oriented (McGuinness
1993; Thorndike 1911). Again, this sex difference is also found in many
different cultures and early in development (Eibl-Eibesfeldt 1989). For instance,
in one study it was found that people were the focus of 80% of the stories of
2-year-old girls, relative to 10% of the boy's stories; boys primarily talked
about objects, such as cars and trains. By 4- years-of-age, 100% of the girls
and about 60% of the boys talked about people, as opposed to objects
(Goodenough 1957). If the sex difference in object versus people interests was
simply due to socialization, then the interests of boys and girls should have
diverged not converged with development (McGuinness 1985; McGuinness &
Pribram 1979). This is not to say that this pattern necessarily supports a
proximate mechanism associated with sexual selection, but rather suggests that
parental socialization is not likely to be the sole cause of the sex difference
in object versus people preferences.
One might argue, for instance, that the pattern found in this study
(Goodenough 1957) simply reflects a sex difference in the rate of the
psychological development of boys and girls, favoring girls. However, the sex difference
in object versus people orientations is found in throughout the lifespan
(McGuinness & Pribram 1979; McGuinness 1993); the sex difference is not
likely to be due to a "developmental delay" in boys. McGuinness and
Pribram, for instance, noted that female infants are preferentially attentive
to communicative signals (e.g., facial expressions), as compared to male
infants' preference for "blinking lights, geometric patterns, colored
photographs of objects and three- dimensional objects" (p. 19).
While levels of competitiveness, degree of social cooperation, and object
orientation will almost certainly be influenced by cultural and child-rearing
practices (MacDonald 1988; Rohner 1976), the overall pattern suggests that we
should entertain the possibility that at least a portion of the sex difference
on these social dimensions is biologically primary, and perhaps related to
sexual selection. Indeed, there have been a number of models that have
attempted to explain these differences based on the principles of sexual
selection. For instance, as described above, it has been argued that males
compete with one another more intensely than females (females compete as well,
but differently; see Buss 1989) in order to gain social status and resources,
which influence mate value and reproductive success more strongly in males than
in females (Betzig 1993; Buss 1989; Symons 1979). The object oriented bias of
males might reflect an inherent interest in resource acquisition or perhaps,
under some circumstances, allow for a depersonalized, and therefore potentially
exploitative, view of other human beings. Indeed, even for preschool boys it
has been noted that "people appear to be assessed ... more on the basis of
their function or utility (can they play games or build a tree house?) than on
the basis of their personal characteristics" (McGuinness & Pribram
1979, p. 21). Nevertheless, it is more likely that the object preferences of
boys reflect a bias toward learning about the physical, as opposed to the social,
environment.
Eibl-Eibesfeldt (1989) suggested that throughout human evolution, the social
style of females provided the nucleus for maintaining the long-term stability
of social groups. In particular, the cooperative and sociable nature of female groups,
relative to male groups, might have provided an important context for the
socialization of children and for the sharing of child care responsibilities.
Whatever the reason, the consistency of these sex differences across historical
periods and cultures, combined with the finding that some of these differences
fluctuate with sex and adrenal hormones and are most pronounced during the
courtship years, suggest that they might reflect the operation of sexual
selection during our evolutionary history. More to the point of the present
discussion, these sex differences might contribute to the later described sex
differences in an interest in pursuing math- intensive careers (e.g.,
engineering) and in responsiveness to competitive and cooperative classroom environments.
3.2. Cognitive sex differences. It has been proposed that
biologically-primary sex differences exist for some cognitive abilities, in
particular language fluency and certain forms of spatial cognition (e.g.
Buffery & Gray 1972; McGuinness & Pribram 1979). The focus of this
section is on potential sex differences in spatial skills, because it has been
frequently argued that the male advantage in certain areas of mathematics
(e.g., problem solving) is related to a male advantage in spatial abilities
(e.g., Benbow 1988; McGee 1979).
Across various types of spatial ability measures, the magnitude and
significance of the male advantage varies considerably (e.g., Caplan et al.
1985). In fact, the lack of consistency across spatial tasks, among other
things, led Caplan et al. to conclude "sex differences in spatial
abilities do not exist or at least it is by no means clear as yet" (p.
797). Despite this claim, consistent and substantive sex differences, favoring
males, have been found for spatial tests that require the mental manipulation
of images in 3-dimensional space, and for more dynamic measures of complex
spatial abilities (Linn & Petersen 1985; Masters & Sanders 1993; Gilger
& Ho 1989; Law et al. 1993; Voyer, Voyer & Bryden 1995). For instance,
in a meta-analysis of sex differences on the Mental Rotation Test (MRT;
Vandenberg & Kuse 1978), a measure of the ability to mentally rotate 3-
dimensional geometric figures, Masters and Sanders found a substantial male
advantage (3/4 to 1 1/4 standard deviations) in 14 of the 14 studies assessed.
The magnitude of the sex difference was not related to the year the study was
published, suggesting that the overall male advantage on the MRT does not
fluctuate with short-term cultural changes; although the magnitude of the sex
difference on some other spatial tests does appear to vary from year to year
(Voyer et al. 1995).
The sex difference on the MRT is especially interesting because of the
nature of the cognitive processes used in the rotational task from which the
MRT was developed. Based on the results of a detailed series of studies of the
strategic and cognitive processes underlying spatial test performance, Just and
Carpenter (1985) stated that the rotation of 3-dimensional geometric figures
(the figures used to develop the MRT) "is not open to alternative
strategies...(and) the cognitive coordinate system within which the figures are
represented is the standard environmentally defined one" (p. 165). Stated
differently, there appears to be little across subject and across item
variability in the processes used to rotate 3- dimensional figures and the
rotation of these figures appears to involve the engagement of the same
neurocognitive systems that support navigation in three dimensional space
(Shepard 1994). In addition to the sex difference in skill at rotating
3-dimensional geometric figures, sex differences are also evident on more
dynamic measure of spatial cognition. Law et al. (1993), for instance, found
that male college students were better at judging the relative distance and the
relative velocity of moving objects than were female college students.
Performance on these spatial tasks improves with practice and experience for
both males and females. However, the magnitude of the male advantage on these
tasks does not diminish with practice (Baenninger & Newcombe 1989; Law et
al. 1993).
Furthermore, "there is now substantial evidence that cognitive patterns
may vary with phases of the menstrual cycle in normally cycling women and with
seasonal variations in androgens in men" (Kimura & Hampson 1994; p.
57). For normally cycling females, performance on spatial tests, including the
MRT, is at its lowest when estrogen and progesterone levels are at their
highest (Hampson & Kimura 1988; Kimura & Hampson 1994; Silverman &
Phillips 1993). For young males, performance on spatial tests varies with
testosterone levels (though not linearly) and peaks in the Spring (Kimura &
Hampson 1994). Prenatal exposure to sex hormones (e.g., androgens; however, the
process is complex and is not always related to androgens in the
straightforward manner) increases the spatial abilities of human females and
females of other species (Diamond et al. 1979; Resnick et al. 1986). In the
laboratory rat, for instance, hormonally treated females and normal males often
outperform castrated males and normal females on spatial tasks (e.g., Williams
et al. 1990).
In one study it was found that "rats exposed neonatally to gonadal
steroids normal males and (hormonally) treated females| selectively attended to
geometric cues when they (were) presented in compound with other types of cues;
landmarks (were) overshadowed by a coordinate system obtained by the geometry
of the room" (Williams et al. 1990; p. 95). Castrated males and normal females
did not selectively attend to geometric cues, although they did attend to these
cues at times. Gaulin (1992) found that the male advantage on laboratory-based
spatial tasks was evident in polygynous but not monogamous species of voles
(Microtus). In natural settings, males in polygamous species have much larger
home ranges than females, but only during the mating season. For monogamous
vole species, the male and female home ranges are equivalent.
In preliterate societies, and for children in Western societies, human males
also tend to have larger "ranges" (or play areas) than human females
(Eaton & Enns 1986; Gaulin 1992; Maccoby 1988). It has been suggested that
for humans, intramale competition favored males with well developed spatial and
navigational abilities (Symons 1979). The selection for strong spatial
abilities in human males might have been directly related to male-male
aggression, which often occurs in the context of small scale warfare between
kin-based groups (Alexander 1979; Geary in press). During these periods of
conflict, groups of related males often travel relatively long distances to
ambush males from other groups or to capture females, which makes foraging far
from the home base dangerous for females in some societies (e.g., Chagnon 1977).
Moreover, skill in the use of primitive weapons, such as a bow and arrow, might
be facilitated by more dynamic spatial skills, such as tracking moving
trajectories (Kolakowski & Malina 1974).
Alternatively, the sex difference in spatial and navigation abilities might
have been indirectly related to male-male competition through their relation to
hunting success (e.g., Hill 1982). In many preliterate societies, "a man's
hunting prowess is directly related to the number of wives he can obtain" (Symons
1979; p. 159), and, as a result, skilled hunters often have more offspring than
their less skilled peers. Hunting skill is presumably dependent, in part, on
the neurocognitive systems that support habitat navigation, as well as on more
dynamic forms of spatial cognition (e.g., tracking moving trajectories). This
is not to say that hunting is the only way that males can achieve high status
in preliterate societies. The point is---if males with superior spatial
abilities (which facilitate hunting and, for example, group migration or
warfare) had even a slight reproductive advantage over their low-ability peers,
then a sex difference, favoring males, in certain spatial abilities (e.g.,
those involving the processing of 3- dimensional information) would have emerged
over the course of human evolution.
Whatever the ultimate reason, for humans, it is likely that hormones have
direct and indirect effects on the development of spatial abilities (Geary
1989). Direct effects would include the influence of sex hormones on the
development of the neural substrate that supports spatial cognition (Diamond et
al. 1979). Indirect influences would involve engagement in behaviors that are
likely to provide spatial-related experiences to the individual. Engagement in
behaviors that are likely to influence spatial ability development (e.g.,
Serbin & Connor 1979) appear to be more strongly influenced by early
exposure to sex hormones (Berenbaum & Hines 1992) than to parental
reinforcement of sex-typed behavior (Lytton & Romney 1991), although the
definitive study of these relationships has yet to be conducted (Baenninger
& Newcombe in press). In addition, sex differences that are likely to
facilitate some component skills associated with hunting/fighting are evident
in children's play. In many preliterate societies, boys practice throwing
sticks and rocks (often at small animals) about three times as often as girls
(Eibl-Eibesfeldt 1989). Two of the largest (often 1.5 to 3.5 standard
deviations) sex differences that have been documented are the velocity and
distance with which objects can be thrown. A male advantage in throwing
velocity and distance is even evident in the preschool years, before large sex
differences in sports-related activities (e.g., baseball; Thomas & French
1985).
In summary, human males have a substantial and robust advantage over human
females on dynamic spatial measures and for measures that require the
representation of information in 3- dimensional space, though much smaller or
no sex differences are found for some less complex spatial measures (e.g.,
rotating images in 2-dimensional space). For complex spatial measures, the male
advantage is found across historical periods and cultures, is influenced by
prenatal exposure to sex hormones, and fluctuates with circulating levels of
sex hormones. At the same time, early exposure to gonadal steroids appears to
predispose boys to engage in activities that will facilitate the development of
spatial skills. At least for rats, sex hormones also appear to make males especially
sensitive to geometric relationships in their habitat, and might explain the
earlier described sex difference in human infants' preferences for social
versus environment (e.g., geometric shapes) stimuli (McGuinness & Pribram
1979). Finally, consistent with the principles of sexual selection (Trivers
1972), the work of Gaulin (1992) suggests that the male advantage in spatial
abilities is restricted to polygynous species, that is, those species, which
include humans, where males are typically larger than females, have higher
mortality rates, etc. The overall patten of results suggests that sexual
selection should be seriously considered as one potential source of human sex
differences in complex spatial abilities, especially those supporting
navigation in the 3- dimensional physical universe.
This position should not be taken to mean that psychological factors, such
as sex-roles, cannot influence performance on spatial tests (e.g., Antill &
Cunningham 1982). Baenninger and Newcombe (in press) have recently argued that
the just described sex differences in spatial abilities, and the later
described sex differences in certain mathematical domains, reflects, at least
in part, the differential experiences of boys and girls. They argue that the
sex difference in early experiences is driven, in part, by sociocultural
influences such as sex role stereotypes, patterns of parental expectations and
encouragement, etc. In fact, the emergence of the just described sex
differences likely represent an interaction between biological biases and
cultural influences. The position here is that at least a portion of the sex
difference in spatial abilities appears to reflect biological influences that
have been shaped by sexual selection. The proximate mechanisms governing the emergence
of these sex differences include sex hormones and a biological bias in the
spatial-related activities of boys and girls, a bias that is also likely to be
influenced by sociocultural factors.
3.3 Intrasexual variability. Although not a primary focus of this article,
the position that sexual selection might be related to the tendency of males to
show more variability in some cognitive domains than females is relevant to the
issue of sex differences in mathematics (Benbow 1988; Feingold 1992; Thorndike
1911). Although it was noted earlier that human males tend toward polygyny, at
least under some circumstances, there is also evidence to suggest that more
monogamous pair-bonding, as well as polygyny, has been selected for at some
points in our evolutionary history (Smuts & Gubernick 1992; Lovejoy 1981).
It is very possible that human males are simply more variable than human
females in terms of reproductive strategies, some males favoring short-term
mating opportunities and others showing high levels of investment in the
martial relationship and the resulting offspring (Buss & Schmitt 1993).
Greater variability in reproductive strategies would presumably result in
greater variability in any associated social or cognitive domains. In fact, it
appears that for many cognitive domains there is greater variability in the
scores of males than in the scores of females (Lubinski & Benbow 1992;
Lubinski & Dawis 1992). In other words, there are more males than females
at the high and low ends of many ability distributions, even for domains where
there are no mean sex differences.
The greater male variability across many cognitive domains has often been
interpreted to reflect something fundamental about males (Ounsted & Taylor
1972). Fausto-Sterling (1985), in contrast, has argued that any such
differences in variability are rather small and when they are found can be
attributed to social bias (e.g., differences in the numbers of boys and girls
referred to remedial or gifted programs) or more males than females at the
lower end of the ability distribution. A this point, the source, or sources, of
the sex difference in cognitive variability is not clear.
If sexual selection were somehow related to this sex difference, then larger
heritability estimates should be found for males than females on any associated
attributes (Wilcockson, Crean & Day 1995). In theory, there might be two
ways in which sexual selection might be related to this sex difference
intrasexual variability. First, the proximate mechanisms, presumably sex
hormones, that induce sex-dimorphic behaviors and cognitive styles might also
contribute to the greater variability in males. For instance, not only are
males more variable in many cognitive abilities, they are also at much greater
risk than females for an array of neurodevelopmental and other physiological
disorders (e.g., Gualtieri & Hicks 1985; Stillion 1985). The greater
morbidity and mortality of males is not restricted to humans, but is seen in
most polygynous species (Daly & Wilson 1983; Ferguson 1985; Mitchell 1981).
Sex hormones have been implicated as one source of these physiological sex
differences, although other models have been presented as well (e.g., Gualtieri
& Hicks 1985). In other words, "androgens not only induce males to violent
and risky behavior but probably also hasten degeneration and senescence
(Ferguson 1985; p. 448). Such deleterious effects of male hormones might
explain the greater number of males than females at the low end of many ability
distributions, but does not explain the greater number of males at the high end
of these distributions. Alternatively, it has been suggested that males are
more sensitive than females to environmental effects, perhaps due to a slower
developmental rate which, in turn, might be influenced by sex hormones (e.g.,
Ounsted & Taylor 1972; Juraska 1986). If so, then poor environments would
negatively impact larger numbers of males than females, whereas larger numbers
of males than females would benefit from enriched environments. The overall result
would be greater male variability in the distribution of many social and
cognitive attributes.
Second, as noted above, the relatively greater intrasexual competition among
males in comparison to females might have resulted in greater intermale
variability in reproductive strategies. In most mammalian species and in most
polygynous human societies, there is much greater variability in the number of
offspring produced by different males, in comparison to females (Daly &
Wilson 1983; Symons 1979; Wade 1979). The greater variability in the
reproductive success of males, relative to females, might have created
pressures for males who were not successful in modal forms of intramale
competition to develop alternative reproductive strategies (Le Boeuf 1974). For
instance, although the reproductive success of male elephant seals (Mirounga
angustirostris) is directly related to social status, which is determined by
means of intermale aggression, some low-status males also father offspring (Le
Boeuf 1974; Le Boeuf & Peterson 1968). These males "sneak into the
harem and occasionally succeed in copulating with females...by apparently
passing for females" (Le Boeuf 1974; p. 173). In addition to the above
described differences across males in the level of paternal investment in
marriage and children, there is further evidence for variability in the
reproductive strategies of human males. For instance, Symons (1979) stated that
in many preliterate societies, some high ranking males were good hunters,
others were skilled at leading group migrations, and still others had well
developed oratory skills. It seems that there are many different routes to high
status for males. Presumably, different routes to high status would have led to
different patterns of cognitive abilities being selected for in males who used
different reproductive strategies. Any such differences in reproductive
strategy would presumably have resulted in greater intra- and inter-individual
variability in the cognitive profiles of males, relative to females. However,
any such sex difference in variability would presumably be restricted to those
domains, such as 3-dimensional spatial abilities, associated with intramale
competition, rather than across all ability domains. Males are, in fact, more
variable than females on spatial-related ability measures but are not more
variable on many verbal tests (e.g. Feingold 1992).
Whatever the reason for the sex difference in variability, there are two
important points. First, understanding sex differences in the variability of
cognitive abilities is probably as important as understanding mean sex
differences, because this variability has implications for understanding any
difference in the numbers of males and females at the high and low ends of
ability distributions (Feingold 1995; Humphreys 1988). Here, sexual selection
and its consequences might be one theoretical perspective from which the issue
of the sex difference in intrasexual variability can be addressed. Second, sex
differences in the distributions of cognitive abilities has important
implications for interpreting mean sex differences in mathematical abilities,
because mathematics is a domain where a consistent sex difference in
variability is found; males are more variable than females (e.g., Feingold
1992). For measures on which there is a mean male advantage, such as
mathematics, there could be no sex difference or a female advantage for low
ability samples, or for samples with poorly developed skills. Moreover,
relative to average-ability samples, there could be larger mean differences,
favoring males, for higher ability samples, and greater numbers of males than
females in these samples (Benbow 1988).
4. Sex differences in mathematical abilities
This section focuses on potential sex differences, or a lack thereof, in
biologically-primary and biologically-secondary mathematical abilities. In
preview, research in these domains suggests no sex differences in
biologically-primary mathematical abilities, but consistent differences,
favoring males, in some biologically secondary areas, in particular
mathematical problem solving and geometry.
4.1. Biologically-primary abilities. Although there have been many reviews
of sex differences in mathematical abilities, these reviews have been largely
confined to kindergarten and older children (Hyde et al. 1990; Kimball 1989).
As a result, there have been no systematic assessments of sex differences in
biologically- primary mathematical domains, most of which can be assessed in
infancy and during the preschool years.
Of particular interest with the infancy studies is whether boys and girls
differ in their sensitivity to numerosity, and in their basic understanding of
the effects of addition and subtraction on quantity (Starkey et al. 1983; Wynn
1992); sex effects were not assessed in the ordinality studies. Many of the
associated studies did not examine sex differences in infant's sensitivity to
numerosity (Starkey & Cooper 1980). Those studies that did examine sex
differences are very consistent in their findings: Boy and girl infants do not
differ in their ability to discriminate small numerosities (Antell &
Keating 1983; Starkey et al. 1990; Strauss & Curtis 1981). Wynn (1992), in
her assessment of 5-month-old infants' understanding of the effects of addition
and subtraction on quantity, did not examine sex differences. In a personal
communication, however, she stated that there was no sex difference at all on
this task (Wynn, personal communication, August 20, 1993). In a set of related
studies, with 18- to 42- month-olds, Starkey (1992) also found no sex
difference in the basic understanding of the effects of addition and
subtraction on quantity.
As described in Section 2.1, an array of important and potentially
biologically-primary numerical abilities emerge during the preschool years.
Ginsburg and Russell (1981), in an extensive study of social class and racial
differences in the basic numerical skills of 4- to 5-year-olds, found only a
single sex difference on measures that appear to assess biologically-primary
abilities; girls were slightly better than boys on a basic addition and
subtraction task. There were no sex differences, across race or social class,
in the ability to count and enumerate, or on any task that assessed basic
counting and number knowledge. In a similar study, Song and Ginsburg (1987)
examined the basic numerical skills of 4- to 8- year-old children from Korea
and the United States. Again, no sex differences for either Korean or American
4- to 5-year-old children were found for measures that appeared to assess
biologically- primary numerical abilities. Finally, Lummis and Stevenson (1990)
examined the pattern of sex differences in reading and mathematical skills from
three cross-national studies of kindergarten, and first- and fifth-grade
children from the United States, Taiwan, and Japan. The results showed no sex
differences in the counting skills, conceptual knowledge, or the simple
arithmetic skills of kindergarten children in any of the three cultures.
The pattern of results suggests that there are no sex differences in
biologically-primary mathematical abilities. This conclusion seems to be
especially sound for preschool and kindergarten children, because the results
are robust across studies and across cultures. For the infancy research,
however, this conclusion must be considered tentative, because the measures
used in these studies, combined with the small sample sizes, might not be
sensitive enough to detect any potentially more subtle differences.
Nevertheless, given the results for preschool children, there appears to be
little reason to suspect that more subtle sex differences exist in these basic
skills. The overall pattern of findings for the infancy and preschool studies
suggest that boys are not biologically primed to outperform girls in basic
mathematics. In other words, the later sex differences in mathematical problem
solving and geometry do not appear to have their antecedents in fundamental
numerical abilities.
4.2. Biologically-secondary abilities. This section focuses on sex differences
in mathematical problem solving, geometry, and other spatially-related areas.
But first, a brief overview of sex differences in complex arithmetic skills
(e.g., solving 35 + 97), but not arithmetic word problems, is presented,
because differences are sometimes found in this area (Hyde et al. 1990).
4.2.1. Complex arithmetic. A sex difference, favoring girls, is often found
on arithmetic achievement tests, especially in the elementary-school and
junior-high-school years (Hyde et al. 1990; Marshall & Smith 1987). Hyde et
al., in a meta-analysis of sex differences in mathematical skills, found that
this advantage is modest (about 1/5 of a standard deviation) but robust, at
least for studies conducted in the United States. Cross-national studies that
have examined sex differences in complex arithmetic suggest that the advantage
of girls in this area is largely an American (i.e., U.S.) phenomenon (Husn
1967). It appears that the sex difference for complex arithmetic skills is
found primarily in countries with low-achieving children, which includes the
United States, and some other nations, such as Sweden (Husn 1967). Carefully
conducted cross-national studies, in nations with higher-achieving children,
typically do not find a sex difference in this area, or sometimes find a male
advantage (Husn 1967; Lummis & Stevenson 1990). Finally, it should be noted
that Hyde et al. (1990) found no sex differences in the understanding of
arithmetical concepts in elementary school through high school. Even though
most of the studies used in the Hyde et al. meta-analysis were conducted in the
United States, cross-national studies support the same conclusion (Fuson &
Kwon 1992; Ginsburg et al. 1981a; 1981b; Lummis & Stevenson 1990; Song
& Ginsburg 1987).
4.2.2. Mathematical problem solving and geometry. The issue of sex
differences in mathematical problem solving, or mathematical reasoning, and
geometry is complex, because there are many different types of component skills
associated with these domains, some of which show a sex difference, favoring
males, and some of which do not. Moreover, in keeping with patterns of
intrasexual variability described in Section 3.3, the magnitude and the
practical importance of any such sex difference appears to vary with the level
of ability of the sample (Hyde et al. 1990). For these reasons, the issue of
sex differences in mathematics is addressed in two sections. The first focuses
on sex differences in general samples, and separately examines performance on
word problems and geometry (and other spatially-based domains, such as
measurement). The second section focuses on the issue of sex differences in
gifted samples (Benbow 1988; Benbow & Stanley 1983). In both sections, the
assessment of sex differences is based on performance on ability and
achievement measures, because performance on these measures is predictive of
later academic and job-related performance (Benbow 1992; Humphreys et al. 1993;
McGee 1979; Rivera-Batiz 1992).
4.2.3. Sex differences in general samples. The results described in this
section are from large-scale studies of what appear to be representative
samples of boys and girls from many different nations. These studies represent
differences across a broad range of abilities, including the gifted. Hyde et
al. (1990), and others before them (e.g., Dye & Very 1968), argued that the
sex difference in mathematical problem solving, for instance, is not evident
until adolescence. As described in Section 2.2, factor- analytic studies
suggest that the Mathematical Reasoning factor does not emerge until high
school, and it is the emergence of this factor that is typically associated
with the emergence of sex differences in mathematical skills, at least in the
United States (Very 1967).
In contrast, a number of cross-national studies of elementary-school
children have found that boys have a performance advantage over girls in
several areas of mathematical problem solving, such as the solving of
arithmetic word problems. Lummis and Stevenson (1990) found a reliable sex
difference, favoring boys, for the solving of arithmetic word problems for
first- and fifth-grade children from the United States, Taiwan, and Japan.
Stevenson and his colleagues also found a reliable sex difference, again
favoring boys, for performance on arithmetic word problems for fifth-grade
children from mainland China and the United States (Stevenson et al. 1990). In
all of the comparisons, the Asian children outperformed their American peers,
but the magnitude of the sex difference (about 1/5 of a standard deviation) was
about the same across nations. Similarly, Marshall and Smith (1987) found that
sixth-grade boys, in the United States, committed fewer errors than sixth-grade
girls when solving arithmetic word problems. Error patterns suggested that
girls found translating relational information into appropriate equations more
difficult than boys (Marshall & Smith 1987).
Harnisch et al. (1986), in a reanalysis of data from a large- scale
international study of the mathematics of achievement of adolescents (Husn
1967), found a male advantage on a mathematics achievement measure in each of
the ten nations included in this reassessment, although the magnitude of the
difference varied greatly across nations. The smallest overall sex difference
was found in the United States and Sweden, and the largest in Great Britain and
Belgium. For 13-year-olds, sex differences were the largest for solving word
problems and for geometry in all ten nations (Steinkamp et al. 1985); for
17-year- olds, this same pattern was found in 8 of the 10 countries. Johnson
(1984) found that male college students consistently (i.e., across 9
experiments) outperformed their female peers for solving algebraic word
problems. The problem set used in this study was largely the same as that used
in a series of studies conducted in the 1950s. Across studies, the overall
magnitude of the sex difference (about 1/2 of a standard deviation) was
unchanged comparing the results from the 1950s to those in the 1980s, although
the male advantage in solving word problems did not generalize to other non-
mathematical problem-solving tasks (Johnson 1984; Maccoby & Jacklin 1974).
In all, these studies indicate that beginning in the elementary-school years
and continuing into adulthood, boys often show an advantage over girls in the
solving of arithmetical and algebraic word problems.
Except for the solving of algebraic word problems, there appear to be no
other consistent sex differences in algebraic skills (Hyde et al. 1990), but
modest sex differences are found in geometry and calculus. As noted earlier,
Harnisch et al. (1986) found a consistent male advantage (of about 1/2 of a
standard deviation) in geometry for both 13- and 17-year-olds across ten
nations. Similarly, Stevenson and his colleagues have consistently found sex
differences on many mathematical tasks that, like geometry, can be solved by
co-opting spatial skills (Lummis & Stevenson 1990; Stevenson et al. 1990).
The associated tasks involved measurement, estimation, and the visualization of
geometric figures. Differences on these tasks, which always favored boys,
emerged as early as the first grade and were found in mainland China, Taiwan,
Japan, the United States, as well as in Great Britain (Wood 1976). In all,
these studies show a consistent advantage of boys over girls in geometry and
other spatially-related mathematical areas, such as measurement, beginning as
early as the first grade.
4.2.4. Sex differences in gifted samples. Perhaps the most striking sex difference
in mathematical problem solving (or mathematical reasoning) emerges in studies
of gifted adolescents (Benbow 1988; Benbow & Stanley 1980; 1983). In the
early 1970s, Stanley began a project designed to identify mathematically
precocious adolescents (Study of Mathematically Precocious Youth; SMPY). To
achieve this end, children who scored in the top 2 to 5% on standard
mathematics achievement tests in the seventh grade were invited to take the
SAT. The mathematics section of the SAT, the SAT-M, assesses the individual's
knowledge of some arithmetic concepts, such as fractions, as well as basic
algebraic and geometric skills (Stanley et al. 1986). Twelve- and
thirteen-year-olds who scored above the mean for high school girls on the SAT-M
were considered to be mathematically gifted, that is, in the top 1% of
mathematical ability (for an overview see Benbow 1988).
The consideration of the sex difference in SAT-M performance for SMPY
adolescents has been on two levels, mean differences and the ratio of boys to
girls at different levels of performance (Benbow 1988; Benbow & Stanley
1983). Across cohorts, American boys, on average, have been found to
consistently outperform American girls on the SAT-M by about 30 points (about
1/2 of a standard deviation). This sex difference has also been found in the
former West Germany and in mainland China (Benbow 1988; Stanley et al. 1986),
although it is of interest to note that the mean performance of gifted Chinese
girls (M = 619) was between 50 and nearly 200 points higher, depending on the
cohort, than the mean of the American boys identified through SMPY (Stanley et
al. 1986). Stanley et al. argued that the advantage of gifted Chinese children
over gifted American children on the SAT-M was probably due to more homework in
China and the fact that some of the material covered on the SAT-M is introduced
in the seventh grade in China, but not until high school in the United States.
The finding that Chinese individuals do not have better developed spatial
abilities than Americans indicates that this national difference in SAT-M
performance is not related to a national difference in spatial abilities
(Stevenson et al. 1985).
In the United States, the sex difference in the ratio of boys to girls at
different levels of mathematical reasoning ability, as measured by the SAT-M,
increases dramatically as the level of mathematical performance increases. The
ratio of boys to girls at the lower end of SAT-M scores is a rather modest
1.5:1, but increases to 13:1 for those scoring > 700 (Benbow & Stanley
1983). The over-representation of boys at the high end of SAT-M performance is
not limited to SMPY samples. Dorans and Livingston (1987), for instance,
reported that across two administrations of the SAT to high-school seniors, 96%
of the perfect scores (i.e., 800) on the SAT-M were obtained by males.
A more recent study indicated that the sex difference in the mathematical
problem- solving abilities of mentally gifted children is evident even during
the elementary school years (Mills et al. 1993). In this study, relatively
large samples (ns greater than 400) of mentally gifted boys and girls in grades
2, 3, 4, 5, and 6 were administered a quantitative test that assessed the
ability to problem solve in mathematics (no computations were needed). The
content of the test included fractions, proportions, and geometry. For all five
grade levels, the boys outperformed girls on the overall measure by roughly 1/2
of a standard deviation. A component analysis indicated that boys and girls were
equally skilled at determining whether or not enough information was available
to answer the question, but boys outperformed girls in the remaining areas.
4.2.5. Summary and conclusion. Consistent sex differences in mathematical
performance are found in some domains, such as geometry and word problems, but
not other domains, such as algebra (Hyde et al. 1990). It has generally been
argued that when a sex difference in mathematical skills is found, it is
typically not found until adolescence (Benbow 1988; Hyde et al. 1990). This
conclusion has been based, for the most part, on comparisons of American
children. Multinational studies, in contrast, show that a male advantage in the
solving of arithmetic word problems and on tasks that are solvable through the use
of spatial skills, such as visualizing geometric shapes, is often evident in
elementary school (Lummis & Stevenson 1990). Differences across studies
primarily conducted in the United States and those conducted in other countries
might be related to the overall level of achievement of the associated samples.
Harnisch et al. (1986) found that sex differences tended to be smallest in
nations with the lowest achieving children, which includes the United States.
Thus, the relatively poor mathematical skills of American children appear to
mask differences that might otherwise be found (see Section 5.2.1).
Even in areas where a sex difference in performance exists across cultures,
the difference tends to be selective. For instance, the results of Marshall and
Smith (1987) suggest that the advantage of elementary-school boys over girls
for solving arithmetical word problems might be more pronounced for problems
that require the translation of important relationships, rather than being a
general male advantage in solving word problems. Similarly, Senk and Usiskin
(1983), in a large-scale national (U.S.) study, found no sex difference in
high-school students' ability to write geometric proofs, after taking a
standard high-school geometry course, even though adolescent males typically
perform better than their female peers on geometric ability tests (Hyde et al.
1990). Thus, the male advantage in geometry also appears to be selective, that
is, associated with certain features of geometry rather than the entire domain.
The apparent selectivity of the male advantage in secondary mathematical
domains is important, because it suggests that the finding of no sex
differences in primary domains and select sex differences in secondary domains
is not simply due to differences in the complexity of primary and secondary
mathematical tasks. Finally, where sex differences emerge, the magnitude of the
difference appears to increase as the level of skill of the associated sample
increases, at least for adolescents and adults.
5. Sex differences in mathematical abilities: Potential causes
In this section, proximate cognitive and psychosocial factors that appear to
contribute to the just described sex differences in mathematical performance
are considered. The relationship between these factors and the
biologically-primary sex differences discussed in Section 3 are explored in
Section 6.
5.1. Cognitive style. The relationship between spatial skills and
mathematical performance is complex and inconsistent (Pattison & Grieve
1984; McGee 1979; Sherman 1967). Spatial skills might be useful for some types
of mathematical tasks, such as visualizing geometric shapes, but are probably
relatively unimportant for seemingly similar tasks, such as providing geometric
proofs. Also, many mathematical problems that could be solved with the use of
spatial representations can also be solved with the use of non- spatial
strategies (Fennema & Tartre 1985; McGuinness 1993), while performance in
other areas of mathematics, such as solving algebraic equations, might not be
strongly facilitated by spatial skills at all (Halpern 1992). In other words,
the relationship between spatial and mathematical skills is probably selective,
even within mathematical areas that appear to have important spatial
components. To illustrate, Ferrini-Mundy (1987) found that spatial skills were
important for solving solids of revolution problems, but less important for
solving other types of calculus problems. Thus, it is not surprising that
correlational studies sometimes find a relationship between spatial skills and
mathematical performance (see McGee 1979), and sometimes do not (Armstrong
1981). It is likely that the content of the mathematical and spatial tests, and
the skill level of the sample strongly influence the degree of correlation
between tests of mathematical and spatial abilities.
Nevertheless, there appears to be a moderate relationship between certain
spatial skills and certain mathematical skills (e.g., Burnett et al. 1979;
Fennema & Sherman 1977; Friedman 1995; Sherman 1980). Johnson (1984), for
instance, found spatial skills and word problem-solving skills to be correlated
.52 and .63, respectively, for male and female college students. Moreover,
providing diagrams (i.e., spatial representations) of the mathematical
relationships presented in the word problems improved the problem-solving
performance of females but not males. Burnett et al. statistically eliminated
an advantage of male college students over their female peers on the SAT-M, by
partialing a sex difference on a spatial visualization test, which included 2-
and 3-dimensional spatial features. Cross-national studies also show that sex
differences in mathematical problem solving primarily reside on tasks that
appear to be solvable through the use of spatial representations (Harnisch et
al. 1986; Lummis & Stevenson 1990). Moreover, although not intuitively
obvious, the male advantage in solving mathematical word problems might also be
related to the sex difference in spatial abilities. Recall that elementary-school
girls appear to have particular difficulty in solving word problems that
involve relational comparisons (Marshall & Smith 1987). As noted in Section
2.2, the errors that are associated with solving these types of problems can
often be avoided, if important relationships described within the problem are
diagramed as part of the problem-solving process (Lewis 1989).
This should not be taken to mean that all cognitive psychologists agree that
the male advantage is spatial abilities contributes to the sex difference in
mathematical abilities. Lubinski and Humphreys (1990a), for instance, argued
that the relationship between sex differences in spatial and mathematical
abilities is a statistical artifact. Friedman (1995) showed that mathematical
achievement scores are just as highly, and sometimes more highly correlated,
with verbal ability than with spatial ability. Clearly, spatial abilities are
not the only source of individual differences in performance on mathematical
problem solving and geometry tests. Nevertheless, the finding that experimental
manipulations, that is, teaching subjects to spatially represent mathematical
relationships, reduces the frequency of errors associated with the solving of
mathematical word problems (Lewis 1989), especially for females (Johnson 1984),
suggests that there is a direct relation between spatial abilities and some
mathematical skills.
The teaching of diagramming skills facilitates the performance of females
more than males, presumably because more males than females spontaneously use
spatial representations to aid in their mathematical problem solving. Indeed,
McGuinness (1993) has recently reported that males, from the age of 4-years,
are much more likely to resort to spatial-related strategies in problem-solving
situations than are females; Johnson and Meade (1987) also found an early
(i.e., elementary school) sex difference, favoring males, in spatial abilities.
In particular, males spontaneously use dynamic 3- dimensional representations
of problem situations much more frequently than do females. This sex difference
in strategic approaches to problem solving easily accommodates the cross-
national male advantage in certain areas of geometry, as well as performance on
estimation and measurement tasks (Harnisch et al. 1986; Lummis & Stevenson
1990), and for solving word problems.
In fact, the final answer to this issue will likely require detailed
trial-by-trial assessments of problem solving strategies across mathematical
domains, as has been done with children's arithmetic (Siegler 1986). Any such
studies will likely show that both males and females use a variety of problem
solving strategies to solve mathematics problems. The prediction is that such
studies will also show that males and females differ in the frequency with
which they use spatial-related strategies for solving certain classes of
problem, such as word problems that involve relational comparisons. Indeed,
many of the contradictory results in this area are quite likely the result of
using measures, such as complex multi-item achievement tests, that reflect an
averaging of performance across items for which different strategies have been
used for problem solving (Siegler 1987).
5.2. Psychosocial factors. In this section, psychosocial influences on the
magnitude of the sex differences in mathematical performance are considered.
Any such influence should be considered as complimentary to the influence that
the sex difference in spatial skills appears to have on mathematical
performance, rather than an alternative explanation. Psychosocial factors
appear to influence the level of participation in mathematics and
mathematics-related activities. A sex difference in the level of participation
in mathematical activities might increase the male advantage in mathematical
and spatial performance, but is not likely to create a sex difference in the
spontaneous use of spatial strategies in problem-solving situations, especially
in 4-year-olds (McGuinness 1993). In all, three general psychosocial influences
on the sex differences in mathematical performance are considered; historical
trends, perceived competence and perceived usefulness of mathematics, and
classroom experiences.
5.2.1. Historical trends. Several recent meta-analyses have suggested that
the magnitude of the sex differences in mathematical performance has declined
over the last several decades (Feingold 1988; Hyde et al. 1990). These trends
support the conclusion that the magnitude of the sex differences in certain
mathematical skills is responsive to social changes, such as increased
participation of girls in mathematics courses (Travers & Westbury 1989).
Hyde et al. (1990) reported that the overall, averaged across all areas, male
advantage in mathematics had been reduced by about 1/2, comparing studies
published before 1973 with studies published in 1974 or later. The authors did
not provide analyses for separate mathematical areas, such as arithmetic or
geometry, so it is unclear whether this represents a general or selective
phenomenon. Also, two data points are certainly not enough to argue for any
type of trend. Feingold, on the other hand, did report separate historical
trends, across four time periods, for the Preliminary SAT (PSAT), taken by
high-school juniors, and the SAT. For the PSAT, there was a 65% decline in the
relative advantage of boys over girls from 1960 to 1983. For the SAT-M, in
contrast, the male advantage remained relatively constant from 1960 to 1983.
Benbow and her colleagues have found no evidence for a declining male advantage
on the SAT-M for SMPY adolescents assessed in different years (Benbow 1988).
Similarly, as noted earlier, Johnson (1984) found that the overall magnitude,
across nine studies, of the male advantage for solving algebraic word problems
remained unchanged from the 1950s to 1980s.
The pattern of results in this area is obviously conflicting. It seems
likely that, in the United States, the magnitude of the sex differences in
mathematical performance is declining in some areas, especially for the general
population (Hyde et al., 1990). This general trend probably reflects a variety
of influences, including the greater participation of girls in
mathematics-related activities and courses (Linn & Hyde 1989; Travers &
Westbury 1989), changes in test items, and potential "floor effects"
(Harnisch et al. 1986; Steinkamp et al. 1985). Recall that Harnisch et al.
found that the advantage of 17-year-old adolescent boys over same-age girls was
smaller in the United States and in Sweden than in eight other nations. The
mean difference in the mathematical performance of American boys and the
performance of boys in the top eight countries was 1.7 standard deviations,
whereas the same comparison for girls produced a difference of 1.2 standard
deviations. In other words, the lack of emphasis on mathematics education in
American culture (Stevenson & Stigler 1992) appears to have a larger impact
on the mathematical achievement of boys than girls.
Moore and Smith (1987) found a similar pattern of sex differences even
within the United States, that is, poorly educated adult females showed better
mathematical skills than their male peers, but for better educated adults,
males had higher mathematics achievement scores than their female peers. Thus,
the historical trend for the magnitude of the sex differences in mathematical
performance, which has been generated based primarily on data collected within
the United States, needs to be interpreted cautiously. In addition to the
increase in the percentage of girls in higher-level mathematics courses (see
Section 5.2.2), it is very possible that the "disappearance" of some
mathematical sex differences might reflect a more general cultural trend in
education that is affecting the achievement of boys more than girls. From this
perspective, and based on patterns of intrasexual variability on cognitive
tests, the sex differences in mathematical problem solving and geometry should
increase as the level of mathematical performance increases. This is exactly
the pattern that has been found with gifted and general samples within the
United States (Benbow 1988; Dorans & Livingston 1987; Moore & Smith
1987), as well as in cross-national comparisons (Steinkamp et al. 1985).
Nevertheless, these same cross-national comparisons show that 17-year-old
girls from eight nations had higher mean mathematics achievement scores than
boys in the United States (Harnisch et al., 1986). The cross-national pattern
of results suggests that with appropriate educational experiences, American
girls, on average, can achieve a level of mathematical competence that far
exceeds the current skill level of their male peers. This appears to be the
case for average as well as gifted children (Stanley et al. 1986). These data
also suggest that with any such improved educational experiences, the magnitude
of the male advantage in certain areas of mathematics will likely
"reappear" or increase.
5.2.2. Perceived competence and usefulness of mathematics. In addition to
spatial skills, attitudes about mathematics appear to be related to the
magnitude of the sex differences in mathematical performance. Of particular
importance is the influence that attitudes have on participation in mathematics
courses. In this section, the sex difference in mathematics course taking is
first reviewed, and is followed by a consideration of psychosocial factors that
influence this sex difference.
Participation in mathematics courses is important, because the number of
mathematics courses taken in high school and beyond will influence the
development of mathematical skills and will influence career options in
adulthood (Sells 1980; Wise 1985). Many studies, across many different nations,
have shown that girls take fewer advanced mathematics courses in high school
than do boys (e.g., Armstrong 1981; Husn 1967; Nevin 1973; Sherman 1981;
Travers & Westbury 1989). This trend lessened somewhat from the 1960s to
the 1980s, although in the 1980s male high school students still outnumbered
female high school students in advanced courses by about 2 to 1 in the United
States and in many other nations (e.g., Chipman & Thomas 1985; Travers
& Westbury 1989). Overall, it appears that the greater participation of
males in mathematics and mathematics-related courses (e.g., physics)
contributes to the sex differences in mathematical performance, at least in
high school (Armstrong 1981; Fennema & Sherman 1977; Husn 1967). Thus, it
is important to understand why women participate in mathematics courses less
frequently than men.
Two important influences on the sex difference in mathematics course taking
in high school include sex differences in perceived mathematical competence and
in the perceived usefulness of mathematics. Eccles et al. (1993) recently
reported that as early as the first grade, children differentiate between their
perceived competence in various academic areas, including math, and the value
or usefulness of the area. Eccles et al. (1993) found no sex difference in the
perceived value or usefulness of mathematics for elementary-school children.
However, during the high-school years female students begin to value English
courses more highly than mathematics courses, whereas male students show the
opposite pattern, valuing mathematics more than English (Eccles et al. 1984;
Lubinski et al. 1993). Eccles et al. (1984) also found that the sex difference
in the relative valuation of English and mathematics mediated a sex difference,
favoring males, in the number of upper- level mathematics courses taken in
twelfth grade. Husn (1967) found the same pattern for male and female
adolescents across twelve nations. Thus, during the high-school years, male
students consistently perceive mathematics as being a more useful skill than
female students.
The perceived usefulness of mathematics appears to be largely related to
long-term career goals (Chipman & Thomas 1985; Wise 1985). Not
surprisingly, those students who aspire to professions that are math intensive,
such as engineering or the physical sciences, take more mathematics courses in
high school and college than those students who have aspirations for less math-
intensive occupations. Chipman and Thomas showed that women were much less
likely to enter math-intensive professions than equal ability males. For
example, in the United States, by the 1970s, women received just over 3% of the
undergraduate engineering degrees, and fewer than 20% of the degrees in
computer science and the physical sciences (i.e., chemistry and physics). The
proportion of women earning graduate degrees in these areas is even lower
(Chipman & Thomas 1985). Longitudinal studies of mathematically gifted
individuals show a similar trend. "Less than 1% of females in the top 1%
of mathematical ability are pursuing doctorates in mathematics, engineering, or
physical science. Eight times as many similarly gifted males are doing so"
(Lubinski & Benbow 1992; p. 64).
The relatively small number of women entering these math- intensive areas is
not simply due to the fact that they are male- dominated professions, although
this likely makes some women hesitant to enter, or remain in, these areas. It
also appears that many girls and women do not believe that work in these areas
will be especially interesting, even women with very high SAT-M scores (Chipman
& Thomas 1985; Lubinski et al. 1993). Sherman (1982), for instance,
assessed the attitudes of ninth-grade girls toward mathematics and
science-related careers. One interview question asked the girls to imagine
working as a scientist for a day. "A clear majority of girls (53%)
disliked that day somewhat or very much" (Sherman 1982; p. 435). Even
those girls who found working as a scientist for a day acceptable did not
consider it to be a preferred activity. This same question was not posed to
ninth- grade boys, so it is not known how many boys of this age would have also
imagined that working as a scientist would be unrewarding. Either way, by
mid-adolescence there is a clear sex difference, favoring boys, in intentions
to pursue scientific or other math-intensive careers (Wise 1985).
The relative lack of interest in math-intensive careers on the part of
girls, in turn, appears to be related to the sex difference in the relative
orientation toward people and human relationships (McGuinness 1993; Maccoby
1988; 1990). As noted in Section 3.1, girls show a greater preference for and
interest in social relationships than boys, whereas boys show a greater
tendency than girls to be object oriented (Eibl-Eibesfeldt 1989; Thorndike
1911). Object-oriented preferences are related to interests in math- intensive
careers, such as engineering or the physical sciences (Chipman & Thomas
1985; Roe 1953). Chipman et al. (1992) found that among college women, an
orientation toward objects, rather than people, was related to an interest in a
career in the physical sciences. This is not the whole story, however, since
female college students are less likely to major in physical science areas than
their male peers, even after controlling for sex differences in object versus
people preferences and the number of mathematics courses taken in high school
(Chipman & Thomas 1985). The fact that most math-intensive areas are male
dominated likely dissuades some women from pursuing careers in these areas
(Halpern 1992). Nevertheless, it also appears that many girls perceive
math-intensive careers (e.g., engineering) to be incompatible with their
interests in human relationships, which, in turn, appears to contribute to the
sex difference in the relative valuation of mathematical skills, and the
resulting sex difference in the number of mathematics courses taken in high
school (Lubinski et al. 1993).
The second important psychosocial factor that appears to influence
participation in mathematics is perceived competence (Meece et al. 1982). At a
general level, perceived competence in academic areas appears to be related to
at least two factors (Marsh et al. 1985). The first is one's performance
relative to peers; the better the relative performance, the higher the
perceived competence. The second factor is intraindividuality. Children who are
relatively better at mathematics than reading tend to have a higher perceived
competence in mathematics than in reading, independent of their skills relative
to other children (Marsh et al. 1985). In the United States, adolescent males
are typically more confident of their mathematical abilities than are their
female peers (Eccles et al. 1984). Harnisch et al. (1986) found that male 17-
year-olds had generally more positive attitudes toward mathematics than female
17-year-olds in 8 of the 10 countries included in their assessment. Eccles et
al. (1993) and Marsh et al. (1985) have found that elementary-school boys feel
better about their mathematical competence than elementary-school girls,
despite the finding that the girls sometimes had higher achievement scores in
mathematics.
Lummis and Stevenson (1990) also found that elementary- school boys in the
United States, Taiwan, and Japan tended to believe that they were better at
mathematics than reading, whereas girls showed the opposite pattern.
Nevertheless, in Taiwan and the United States, most of the children thought
that boys and girls were equally skilled in mathematics. In Japan, roughly 1/3
of the boys and girls thought that there was no sex difference in mathematical
ability, whereas 1/3 of the children thought that boys were better, and the
remaining 1/3 thought that girls were better. Thus, the finding that
elementary-school boys have better perceptions of their mathematical skills
than girls might be related to intraindividual comparisons, that is, because
girls tend to prefer reading over math, whereas boys tend to prefer math over
reading (Lummis & Stevenson 1990). It is likely that for adolescents,
performance differences in high school reinforce these perceptions, as do sex-
role stereotypes that portray mathematics as a male domain (Becker 1981; Linn
& Hyde 1989).
Regardless of why a sex difference in perceived mathematical competence
emerges, it has been argued that perceived competence and the associated
expectancies for success might influence task persistence after experiencing
failure, and influence the likelihood that the individual will aspire to a
math- intensive career (Chipman et al. 1992). Eccles et al. (1984), however,
found no sex difference in the tendency to persist on a mathematical
problem-solving task following failure, although Kloosterman (1990) found that
for algebraic problem solving, high- school boys tended to increase their
efforts following failure, whereas high-school girls tended to show less effort
following failure. Meece et al. (1990) found that basic abilities, as indexed
by earlier grades, and expectancies both contributed to mathematical
performance. In all, it seems likely that the sex difference in perceived
mathematical competence likely contributes to the sex difference in aspirations
to math-intensive careers and the associated sex difference in mathematics
course taking in high school (Chipman et al. 1992).
5.2.3. Classroom experiences. It has been suggested that the classroom
experiences of boys and girls might impact the development of reading and
mathematical skills (Becker 1981; Kimball 1989). In high-school mathematics
courses, for instance, a number of studies have found that boys tend to receive
more personal contact with teachers than girls. Becker (1981), for instance, collected
quantitative and qualitative data on student- teacher interactions across 100
geometry lessons and found some important differences in the way in which
teachers interacted with males and females. For example, in comparison to
females, males received many more encouraging comments and were asked more
process oriented questions. The asking of process oriented questions is
important because these questions tend to facilitate the student's conceptual
understanding of the material (Geary 1994). Thus, it is likely that the
mathematics classroom experiences differ for some males and females, especially
in high school, and might contribute to the sex difference in perceived
mathematical competence. Though problematic, the nature of these student-
teacher interactions is not likely to be the only source of the sex difference
in perceived mathematical competence. This is so because females typically
receive higher grades than males in these classes which presumably would
facilitate their mathematical competence (Kimball 1989).
Indeed, an intriguing study conducted by Peterson and Fennema (1985)
suggests that more general teaching styles might have stronger influences on
the mathematical, and probably general academic, development of boys and girls
than individual interactions between students and teachers. In this study, the
activities of teachers and students in 36 fourth-grade classrooms were recorded
during mathematics instruction for at least 15 days. Student-teacher
interactions were observed, such as the amount of time the teacher spent
helping individual students, as well as the amount of time each student spent
in mathematical and nonmathematical (e.g., socializing) activities. There was
no sex difference in the amount of time boys and girls spent learning
mathematics, and no overall sex difference in the change in mathematics
achievement scores from the beginning to the end of the school year. However,
there were important across-class differences in the relative gains of boys and
girls in mathematics achievement. These gains appeared to be related to
frequency with which the children engaged in competitive or cooperative
classroom activities. Engagement in competitive activities had a considerable
negative impact on the mathematics achievement of girls, but slightly improved
the performance of boys. Engagement in cooperative mathematical activities
(e.g., problem solving in small groups), in contrast, was associated with
improvements, across the academic year, in the mathematical performance of
girls, but poorer performance for boys.
6. Culture, sexual selection, and sex differences in mathematical abilities:
An integrative model
In this section, an integrative consideration of cultural and biological influences
on mathematical development and the earlier described sex differences in
mathematical performance is presented.
Figure 1: Schematic representation
of the relations among biological, cognitive, and psychosocial influences in
mathematical performance.
The general pattern of relations is show in Figure
1. Here, I have divided biologically-secondary mathematical domains into
two general areas, because different factors might influence sex differences in
these areas. The first includes the solving of mathematical word problems
(i.e., mathematical problem solving), geometry, and other spatially-related
areas, such as measurement. The second includes other areas of mathematics that
should not be as strongly influenced by spatial abilities, such as algebra (except
word problems), probability, etc.
Before examining the pattern of relations shown in Figure
1, I want to suggest that the general level of mathematical achievement
within any given culture influences the degree to which sex differences in
biologically-secondary mathematical domains (e.g., mathematical problem
solving) are expressed, but is not the primary cause of these sex differences.
In fact, given that sex differences only appear to emerge for secondary, and
not primary, mathematical domains, these differences will only emerge in
societies with prolonged formal mathematical instruction. Although the mode of
instruction (i.e., competitive versus cooperative classrooms) appears to
differentially influence the mathematical development of boys and girls,
instructional differences are not likely to be the only source of the sex
differences in mathematical abilities. Rather, math instruction in school
provides an important context within which more primary sex differences (e.g.,
spatial abilities) can be expressed.
Moreover, the finding that males tend to be more variable in mathematical
and other cognitive abilities and more sensitive to environmental influences
(see Section 3.3) in comparison to females, suggests that the degree to which
sex differences in mathematical performance are expressed should vary directly
with the overall degree of mathematical development. The magnitude of the sex
differences in mathematical performance should, and appears to be, the largest
for samples (gifted and nongifted) with well developed mathematical skills
(Benbow 1988; Harnisch et al. 1986). At the other end of the continuum, it
appears that these sex differences disappear or are reversed in some cases
(Moore & Smith 1987). Thus, the overall level of skill development
influences the expression of sex differences in mathematical abilities, but is
not likely to be the primary causes of these differences. Stated differently,
sex differences in mathematical performance appear to be the largest,
especially by adolescence, in those cultures that greatly facilitate the
mathematical development of children. As the emphasis on mathematics
achievement declines within a culture, such as the United States, the
performance of boys appears to fall more quickly than that of girls. As a
result, any sex differences in component abilities will likely be masked.
Nevertheless, there do appear to be some more specific cultural influences
that contribute to the magnitude of the sex differences in mathematical
performance, that is, sex-role stereotypes and cultural influences on perceived
mathematical competence. These factors are shown in the bottom left-hand corner
of Figure
1 and are important because they appear to influence participation in
mathematics course taking and related activities (Becker 1981; Chipman et al.
1992; Linn & Hyde 1989). However, in comparison to the sex difference in
social preferences (described below), stereotypes and perceived competencies
probably have a relatively minor, but nonetheless important, influence on
participation in mathematics course taking and related activities. Fennema et
al. (1981), for instance, found that focusing on the utility of mathematics had
a stronger effect on increasing girl's mathematics course taking than did
changing sex-role stereotypes. Moreover, stereotypes are the smallest in those
groups (i.e., high-ability groups) where the sex differences in mathematics are
the largest (Lubinski & Humphreys 1990b; Raymond & Benbow 1986). Stated
differently, high-ability girls tend to be less sex-typed than other girls, but
still participate much less frequently in mathematics-related activities than
their male peers (Lubinski & Benbow 1992). Nevertheless, these findings do
not preclude the possibility that some high-ability girls (and other girls) who
do not have strong sex-role stereotypes might be discouraged from entering
mathematical areas by individuals with sex-role stereotypes (Becker 1981).
Sex differences in social preferences and general interests appear to be
important influences on the sex difference in the perceived utility of
mathematics and the resulting sex difference in mathematics course taking
(Lubinski et al. 1993). Lubinski et al. argued that "the personal
attributes of females compared to males will lead them to choose scientific
(and mathematical) careers less frequently, as a group, and to distribute their
educational development across artistic, social and investigative areas more
evenly" (Lubinski et al. 1993; p. 704). Moreover, "preference and
ability profiles are in place long before high school" (Lubinski et al.
1993; p. 704), that is, before the sex difference in mathematics course taking
emerges. It is not unreasonable to argue that some of these sex differences in
social preferences are influenced by stereotypes and perceived competencies; in
fact, it is likely that preferences and stereotypes have bidirectional
influences on one another, as shown in Figure
1. However, the finding that interest in math-intensive careers is related
to object rather than people orientations (e.g. Chipman et al. 1992), and that
the sex difference in object versus people orientations is evident in cultures
where there are no scientists, mathematicians, or high-school mathematics
curriculums, suggests that sex differences in social preferences and interests
largely transcend specific cultural influences. Culture, will of course,
influence how these preferences can be expressed, but is probably not the
primary cause of these differences.
In fact, as depicted in Figure
1, I am arguing, as others have (e.g., Thorndike 1911), that the sex
difference in social preferences, that is, object versus people orientations,
and social styles might arise, at least in part, from biological sex
differences. I argued in Section 3.1 that sexual selection might have operated
to make the nature of social relationships different for males and females. For
instance, developing intimate social relationships seems to be an end in itself
for young females but not young males (Block 1993). Young males, of course,
have companions and develop social alliances, but these might be more of a
means to an end, rather than an end in itself. Regardless, the point is that
males and females differ in terms of the importance of social relationships and
interests in objects, which appear to influence career aspirations. Career
aspirations appear to influence the perceived utility of mathematics, which, in
turn, influences participation in mathematics-related activities (and other
types of activities, as well).
It was also suggested that sexual selection operated to make males more
competitive than females, and, as such, might influence how boys and girls
perform, in mathematics and other academic areas, in competitive and
cooperative classroom environments. In keeping with this view, is the finding
that the mathematical achievement of girls is the highest in cooperative
settings (e.g., problem solving small groups); the performance of boys,
however, drops in these setting (Peterson & Fennema 1985). Similarly, the
mathematical achievement of girls drops in competitive classrooms, while the
achievement of boys improves slightly. Sociocultural models for these social
sex differences could be developed, I'm sure. The overall pattern, however,
seems to be consistent with the earlier described sex differences in social
styles (Section 3.1). At the very least, we should consider the possibility
that fundamental differences between males and females, perhaps shaped by
sexual selection, influence social styles and preferences and, as such, are
indirect sources of the sex difference in participation in mathematics courses
and in related activities.
I also argued that one consequence of sexual selection was different
cognitive ability patterns in males and females. Most relevant to the current
discussion is the argument that sexual selection might be the primary source of
the male advantage in complex spatial abilities. Specifically, it was argued
that sexual selection, and any associated proximate mechanisms (e.g., sex
hormones) resulted in the more elaborate development of the neurocognitive
systems that evolved to support navigation in the three-dimensional physical
universe in males than in females. >From this perspective, sex differences
in spatial abilities are expected, and appear to be the largest, for tasks that
require the mental manipulation or tracking of information in 3-dimensional
space (e.g., Masters & Sanders 1993). Sex differences in the systems that
support habitat navigation are expected to show a selective relationship to
mathematical performance. In particular, knowledge implicit in the systems that
support habitat navigation should result in a sex difference, favoring males,
in an intuitive understanding of the basic principles of Euclidean geometry.
This is because Euclid's basic postulates seem to reflect, to some extent,
knowledge that appears to be implicit (i.e., skeletal principles) to the
neurocognitive systems that support habitat navigation.
Moreover, it is expected that the systems that have evolved to support
habitat navigation can be co-opted by both males and females; the use of
spatial representations of mathematical relationships to solve mathematical
word problems is one example of co-optation (Lewis 1989). Sex differences are
anticipated, however, in the complexity of the spatial representations that can
be spontaneously developed by males and females, and perhaps in the frequency
with which these representations are spontaneously used by males and females
during problem solving (McGuinness 1993). The relatively better developed
spatial abilities of males, on average, in comparison to females might also
contribute to the sex difference, favoring males, in participation in
mathematics-related courses (e.g., drafting) and activities, as shown in Figure
1. A sex difference in such activities is also likely to further increase
the male advantage in complex spatial abilities.
Either way, the cross-national pattern of sex differences in mathematical
abilities is generally consistent with the argument that these differences
emerge primarily for geometry and other areas where the use of spatial
representations might facilitate performance. However, none of the studies
described in Section 4 assessed the sex differences in mathematical performance
with the fidelity that I am proposing; for example, the sex differences in
spatial abilities should be the largest for 3-dimensional tests, and that
performance on 3-dimensional tests of spatial ability should be more strongly
related to performance in specific areas of geometry (e.g., solving problems
represented in 3-dimensional space) than, for instance, algebra or even other
areas of geometry (e.g., writing proofs). Thus, although the overall pattern of
sex differences in mathematical abilities seems to be consistent with the
argument that these differences are, in part, due to sex differences in the
degree to which the neurocognitive systems that support habitat navigation are
elaborated, the validity of the specific predictions presented here must await
further research.
The model presented in Figure
1 also predicts some sex differences in mathematical areas that are not
facilitated by spatial abilities. For these areas, sex differences would be
primarily driven by sex differences in mathematics course taking and related
experiences. As a result, sex differences in these areas, if they arise, are
expected to be smaller than for geometry and mathematical word problems, and
emerge only after the emergence of sex differences in mathematics course taking
(i.e., the latter part of high school).
Finally, in closing, even though I have argued that biological sex
differences are an important source, though not the only source, of the sex
differences that emerge for some biologically-secondary mathematical domains,
this should not be taken to mean that the mathematical development of girls
cannot be improved. In fact, there are a number of steps that can be taken to
improve the mathematical development of girls. First, stressing the utility of
mathematics for future career options increases girl's mathematics course
taking (Fennema et al. 1981). Second, even though girls do not appear to
spontaneously use spatial representations in problem-solving situations as
frequently as boys do, they can be taught to do so (Lewis 1989; Johnson 1984).
Teaching girls to use diagrams during mathematical problem solving
significantly improves their performance, but the male advantage does not
disappear (Johnson 1984). Third, highly competitive classroom environments
should be avoided; highly cooperative environments should be avoided as well,
since the achievement of boys appears to drop in these environments. Classroom
teaching styles should either be gender neutral, or boys and girls should be
educated in separate classrooms. Finally, relative to international standards,
the mathematical development of American children, boys and girls, is very
poor. As noted earlier, with improved mathematical instruction, American girls
should be able to develop a level of mathematical skill that far exceeds the
current level of their male peers. At the same time, a greater emphasis on
mathematical instruction in the United States will not likely result in
disappearing sex differences. In fact, based on cross-national studies, sex
differences in mathematics are likely to increase, as the level of mathematical
achievement increases.
References
Alexander, R. (1979) Darwinism and human affairs. University
of Washington Press.
Antell, S. E. & Keating, D. P. (1983) Perception of numerical invariance
in neonates. Child Development 54: 695-701.
Antill, J. K. & Cunningham, J. D. (1982) Sex differences in performance
on ability tests as a function of masculinity, femininity, and androgyny.
Journal of Personality and Social Psychology 42: 718-728.
Armstrong, J. M. (1981) Achievement and participation of women in
mathematics: Results from two national surveys. Journal for Research in
Mathematics Education 12: 356-72.
Baenninger, M. & Newcombe, N. (1989) The role of experience in spatial
test performance: A meta-analysis. Sex Roles 20: 327- 44.
Baenninger, M. & Newcombe, N. (in press) Environmental input to the
development of sex-related differences in spatial and mathematical ability.
Learning and Individual Differences.
Becker, J. R. (1981) Differential treatment of females and males in
mathematics classes. Journal for Research in Mathematics Education 12: 40-53.
Benbow, C. P. (1988) Sex differences in mathematical reasoning ability in
intellectually talented preadolescents: Their nature, effects, and possible
causes. Behavioral and Brain Sciences 11: 169-232.
Benbow, C. P. (1992) Academic achievement in mathematics and science of
students between ages 13 and 23: Are there differences among students in the
top one percent of mathematical ability? Journal of Educational Psychology 84:
51-61.
Benbow, C. P. & Stanley, J. C. (1980) Sex differences in mathematical
ability: Fact or artifact? Science 210: 1262-64.
Benbow, C. P. & Stanley, J. C. (1983) Sex differences in mathematical
reasoning ability: More facts. Science 222: 1029- 31.
Berenbaum, S. A. & Hines, M. (1992) Early androgens are related to
childhood sex-typed toy preferences. Psychological Science 3: 203-6.
Betzig, L. (1993) Sex, succession, and stratification in the first six
civilizations: How powerful men reproduced, passed power on to their sons, and
used power to defend their wealth, women, and children. In: Social
stratification and socioeconomic inequality Volume 1: A comparative biosocial
analysis, ed. L. Ellis. Praeger.
Block, J. (1993) Studying personality the long way. In: Studying lives
through time: Personality and development, ed. D. C. Funder, R. D. Parke, C.
Tomlinson-Keasey, & K. Widaman. American Psychological Association.
Boysen, S. T. (1993) Counting in chimpanzees: Nonhuman principles and
emergent properties of number. In: The development of numerical competence:
Animal and human models, ed. S. T. Boysen & E. J. Capaldi. Erlbaum.
Boysen, S. T. & Berntson, G. G. (1989) Numerical competence in a
chimpanzee (Pan troglodytes). Journal of Comparative Psychology 103: 23-31.
Briars, D. & Siegler, R. S. (1984) A featural analysis of preschoolers'
counting knowledge. Developmental Psychology 20: 607-18.
Buffery, A. W. H. & Gray, J. A. (1972) Sex differences in the
development of spatial and linguistic skills. In: Gender differences: Their ontogeny
and significance, ed. C. Ounsted & D. Taylor. Churchill Livingstone.
Buhrmester, D. & Furman, W. (1987) The development of companionship and
intimacy. Child Development 58: 1101-13. Burnett, S. A., Lane, D. M. &
Dratt, L. M. (1979) Spatial visualization and sex differences in quantitative
ability. Intelligence 3: 345-54.
Buss, D. M. (1989) Sex differences in human mate preferences: Evolutionary
hypotheses tested in 37 cultures. Behavioral and Brain Sciences 12: 1-49.
Buss, D. M. (1995) Psychological sex differences: Origins through sexual
selection. American Psychologist 50: 164-68.
Buss, D. M. & Schmitt, D. P. (1993) Sexual strategies theory: An
evolutionary perspective on human mating. Psychological Review 100: 204-12.
Caplan, P. J., MacPherson, G. M. & Tobin, P. (1985) Do sex- related
differences in spatial abilities exist? A multilevel critique with new data.
American Psychologist 40: 786-99.
Cattell, R. B. (1963) Theory of fluid and crystallized intelligence: A
critical experiment. Journal of Educational Psychology 54: 1-22.
Chagnon, N. A. (1977) Yanomamo, the fierce people. Holt, Rinehart and
Winston.
Cheng, K. & Gallistel, C. R. (1984) Testing the geometric power of an
animal's spatial representation. In: Animal cognition, ed. H. L. Roitblat, T.
G. Bever, & H. S. Terrace. Erlbaum.
Chipman, S. F. & Thomas, V. G. (1985) Women's participation in
mathematics: Outlining the problem. In: Women and mathematics: Balancing the
equation, ed. S. F. Chipman, L. R. Brush & D. M. Wilson. Erlbaum.
Chipman, S. F., Krantz, D. H. & Silver, R. (1992) Mathematics anxiety
and science careers among able college women. Psychological Science 3: 292-95.
Coombs, C. H. (1941) A factorial study of number ability. Psychometrika 6:
161-89.
Cooper, R. G. Jr. (1984) Early number development: Discovering number space
with addition and subtraction. In: Origins of cognitive skills: The eighteenth
Carnegie symposium on cognition, ed. C. Sophian. Erlbaum.
Crump, T. (1990) The anthropology of numbers.
Daly, M. & Wilson, M. (1983) Sex, evolution, and behavior (second
edition). PWS Publishers.
Daly, M. & Wilson, M. (1988) Evolutionary social psychology and family
homicide. Science 242: 519-24.
Darwin, C. (1859) The origin of species by means of natural selection. John
Murray.
Darwin, C. (1871) The descent of man and selection in relation to sex.
Modern Library.
Davis, H. & Memmott, J. (1982) Counting behavior in animals: A critical
examination. Psychological Bulletin 92: 547-71.
Diamond, M. C., Johnson, R. E. & Ehlert, J. (1979) A comparison of
cortical thickness in male and female rats-normal and gonadectomized, young and
adult. Behavioral and Neural Biology 26: 485-91.
Dorans, N. J. & Livingston, S. A. (1987) Male-female differences in
SAT-Verbal ability among students of high SAT-Mathematical ability. Journal of
Educational Measurement 24: 65-71.
Dye, N. W. & Very, P. S. (1968) Growth changes in factorial structure by
age and sex. Genetic Psychology Monographs 78: 55- 88.
Eagly, A. H. & Steffen, V. J. (1986) Gender and aggressive behavior: A
meta-analytic review of the social psychological literature. Psychological
Bulletin 100: 309-30.
Eaton, W. O. & Enns, L. R. (1986) Sex differences in human motor
activity level. Psychological Bulletin 100: 19-28.
Eccles (Parsons), J., Adler, T. & Meece, J. L. (1984) Sex differences in
achievement: A test of alternative theories. Journal of Personality and Social
Psychology 46: 26-43.
Eccles, J., Wigfield, A., Harold, R. D. & Blumenfeld, P. (1993) Age and
gender differences in children's self- and task perceptions during elementary
school. Child Development 64: 830-47.
Eibl-Eibesfeldt, I. (1989) Human ethology. Aldine de Gruyter.
Fausto-Sterling, A. (1985) Myths of gender. Basic Books.
Feingold, A. (1988) Cognitive gender differences are disappearing. American
Psychologist 43: 95-103.
Feingold, A. (1992) Sex differences in variability in intellectual
abilities: A new look at an old controversy. Review of Educational Research 62:
61-84.
Feingold, A. (1995) The additive effects of differences in central tendency
and variability are important in comparisons between groups. American
Psychologist 50: 5-13.
Fennema, E. & Sherman, J. (1977) Sex-related differences in mathematics
achievement, spatial visualization and affective factors. American Educational
Research Journal 14: 51-71.
Fennema, E. & Tartre, L. A. (1985) The use of spatial visualization in
mathematics by girls and boys. Journal of Research in Mathematics Education 16:
184-206.
Fennema, E., Wolleat, P. L., Pedro, J. D. & Becker, A. D. (1981)
Increasing women's participation in mathematics: An intervention study. Journal
for Research in Mathematics Education 12: 3-14. Ferguson, M. W. J. (1985) Short
and sweet: The classic male life? Behavioral and Brain Sciences 8: 448-449.
Ferrini-Mundy, J. (1987) Spatial training for calculus students: Sex
differences in achievement and in visualization ability. Journal for Research
in Mathematics Education 18: 126-40.
Frayer, D. W. & Wolpoff, M. H. (1985) Sexual dimorphism. Annual Review
of Anthropology 14: 429-73.
Friedman, L. (1995) The space factor in mathematics: Gender differences.
Review of Educational Research 65: 22-50.
Furstenberg, F. F. Jr. & Nord, C. W. (1985) Parenting apart: Patterns of
childrearing after marital disruption. Journal of Marriage and the Family 47:
893-904.
Fuson, K. C. (1988) Children's counting and concepts of number.
Springer-Verlag.
Fuson, K. C. & Kwon, Y. (1992) Korean children's understanding of
multidigit addition and subtraction. Child Development 63: 491- 506.
Gallistel, C. R. (1990) Organization of learning. MIT Press/Bradford Books.
Gallistel, C. R. & Gelman, R. (1992) Preverbal and verbal counting and
computation. Cognition 44: 43-74.
Gaulin, S. J. C. (1992) Evolution of sex differences in spatial ability.
Yearbook of Physical Anthropology 35: 125-51.
Geary, D. C. (1989) A model
for representing gender differences in the pattern of cognitive abilities.
American Psychologist 44: 1155-56.
Geary, D. C. (1992) Evolution of human cognition: Potential relationship to
the ontogenetic development of behavior and cognition. Evolution and Cognition
1: 93-100.
Geary, D. C. (1994) Children's mathematical development: Research and
practical applications. American Psychological Association.
Geary, D. C. (1995) Reflections of evolution and culture in children's
cognition: Implications for mathematical development and instruction. American
Psychologist 50: 24-37. Geary, D. C. (in press). Sexual selection and sex differences
in spatial cognition. Learning and Individual Differences.
Gelman, R. (1990) First principles organize attention to and learning about
relevant data: Number and the animate-inanimate distinction as examples.
Cognitive Science 14: 79-106.
Gelman, R. (1993) A rational-constructivist account of early learning about
numbers and objects. In: The psychology of learning and motivation: Advances in
research and theory, ed. D. L. Medin. Academic Press.
Gelman, R. & Gallistel, C. R. (1978) The child's understanding of
number. Harvard University Press.
Ghiglieri, M. P. (1987) Sociobiology of the great apes and the hominid
ancestor. Journal of Human Evolution 16: 319-57.
Gilger, J. W. & Ho, H. Z. (1989) Gender differences in adult spatial
information processing: Their relationship to pubertal timing, adolescent
activities, and sex-typing of personality. Cognitive Development 4: 197-214.
Ginsburg, H. P., Posner, J. K. & Russell, R. L. (1981) The development
of mental addition as a function of schooling and culture. Journal of
Cross-Cultural Psychology 12: 163-78.
Ginsburg, H. P. & Russell, R. L. (1981) Social class and racial
influences on early mathematical thinking. Monographs of the Society for
Research in Child Development 46 (No. 6, Serial No. 193).
Goodenough, E. W. (1957) Interest in persons as an aspect of sex differences
in the early years. Genetic Psychology Monographs 55: 287-323.
Gould, J. L. (1986) The locale map of honey bees: Do insects have cognitive
maps? Science 232: 861-63.
Gould, S. J. & Vrba, E. S. (1982) Exaptation--a missing term in the
science of form. Paleobiology 8: 4-15.
Gualtieri, T. & Hicks, R. E. (1985) An immunoreactive theory of
selective male affliction. Behavioral and Brain Sciences 8: 427- 441.
Gustafsson, J. E. (1984) A unifying model for the structure of intellectual
abilities. Intelligence 8: 179-203. Guthrie, G. M. (1963) Structure of
abilities in a non-Western culture. Journal of Educational Psychology 54:
94-103.
Hall, B. K. (1992) Evolutionary developmental biology. Chapman & Hall.
Halpern, D. F. (1992) Sex differences in cognitive abilities (second
edition). Erlbaum.
Hampson, E. & Kimura, D. (1988) Reciprocal effects of hormonal
fluctuations on human motor and perceptual-spatial skills. Behavioral Neuroscience
102: 456-59.
Harnisch, D. L., Steinkamp, M. W., Tsai, S. L. & Walberg, H. J. (1986)
Cross-national differences in mathematics attitude and achievement among
seventeen-year-olds. International Journal of Educational Development 6:
233-44.
Hill, K. (1982) Hunting and human evolution. Journal of Human Evolution 11:
521-44.
Horn, J. L. & Cattell, R. B. (1966) Refinement and test of the theory of
fluid and crystallized general intelligence. Journal of Educational Psychology
57: 253-70.
Humphreys, L. G. (1988) Sex differences in variability may be more important
than sex differences in means. Behavioral and Brain Sciences 11: 195-96.
Humphreys, L. G., Lubinski, D. & Yao, G. (1993) Utility of predicting
group membership and the role of spatial visualization in becoming an engineer,
physical scientist, or artist. Journal of Applied Psychology 78: 250-61.
Husn, T. (1967) International study of achievement in mathematics: A
comparison of twelve countries, vol. 2. Wiley.
Hyde, J. S. (1984) How large are gender differences in aggression? A
developmental meta-analysis. Developmental Psychology 20: 722-36.
Hyde, J. S., Fennema, E. & Lamon, S. J. (1990) Gender differences in
mathematics performance: A meta-analysis. Psychological Bulletin 107: 139-55.
Johnson, E. S. (1984) Sex differences in problem solving. Journal of
Educational Psychology 76: 1359-71.
Johnson, E. S. & Meade, A. C. (1987) Developmental patterns of spatial
ability: An early sex difference. Child Development 58: 725-40.
Juraska, J. M. (1986) Sex differences in developmental plasticity of
behavior and the brain. In: Developmental Neuropsychobiology, ed. W. T.
Greenough & J. M. Juraska. Academic Press.
Just, M. A., & Carpenter, P. A. (1985) Cognitive coordinate systems:
Accounts of mental rotation and individual differences in spatial ability.
Psychological Review 92: 137- 72.
Kimball, M. M. (1989) A new perspective on women's math achievement.
Psychological Bulletin 105: 198-214.
Kimura, D. & Hampson, E. (1994) Cognitive pattern in men and women is
influenced by fluctuations in sex hormones. Current Directions in Psychological
Science 3: 57-61.
Kloosterman, P. (1990) Attributions, performance following failure, and
motivation in mathematics. In Mathematics and gender, ed. E. Fennema & G.
C. Leder. Teachers College Press.
Kolakowski, D. & Malina, R. M. (1974) Spatial ability, throwing accuracy
and man's hunting heritage. Nature 251: 410-12.
Kyllonen, P. C. & Christal, R. E. (1990) Reasoning ability is (little
more than) working-memory capacity?! Intelligence 14: 389- 433.
Landau, B., Gleitman, H., & Spelke, E. (1981) Spatial knowledge and
geometric representation in a child blind from birth. Science 213: 1275-1278.
Law, D. J., Pellegrino, J. W. & Hunt, E. B. (1993) Comparing the tortoise
and the hare: Gender differences and experience in dynamic spatial reasoning
tasks. Psychological Science 4: 35-40.
Le Boeuf, B. J. (1974) Male-male competition and reproductive success in
elephant seals. American Zoologist 14: 163-176.
Le Boeuf, B. J. & Peterson, R. S. (1969) Social status and mating
activity in elephant seals. Science 163: 91-93.
Lewis, A. B. (1989) Training students to represent arithmetic word problems.
Journal of Educational Psychology 81: 521-31.
Lewis, A. B. & Mayer, R. E. (1987) Students' misconceptions of
relational statements in arithmetic word problems. Journal of Educational
Psychology 79: 363-71.
Linn, M. C. & Hyde, J. S. (1989) Gender, mathematics, and science.
Educational Researcher 18: 17-19, 22-27.
Linn, M. C. & Petersen, A. C. (1985) Emergence and characterization of
sex differences in spatial ability: A meta- analysis. Child Development 56:
1479-98.
Lovejoy, C. O. (1981) The origin of man. Science 211: 341-50.
Lubinski, D. & Benbow, C. P. (1992) Gender differences in abilities and
preferences among the gifted: Implications for the math-science pipeline.
Current Directions in Psychological Science 1: 61-66.
Lubinski, D., Benbow, C. P. & Sanders, C. (1993) Reconceptualizing
gender differences in achievement among the gifted. In: International handbook
of research and development of giftedness and talent, ed. K. A. Heller, F. J.
Monks & A. H. Passow. Pergamon Press.
Lubinski, D. & Dawis, R. V. (1992) Aptitudes, skills, and proficiencies.
In: The handbook of industrial/organizational psychology (second edition), ed.
M. D. Dunnette & L. M. Hough. Consulting Psychologists Press.
Lubinski, D. & Humphreys, L. G. (1990a) Assessing spurious
"moderator effects": Illustrated substantively with the hypothesized
("synergistic") relation between spatial and mathematical ability.
Psychological Bulletin 107: 385-393.
Lubinski, D. & Humphreys, L. G. (1990b) A broadly based analysis of
mathematical giftedness. Intelligence 14: 327-55.
Lummis, M. & Stevenson, H. W. (1990) Gender differences in beliefs and
achievement: A cross-cultural study. Developmental Psychology 26: 254-63.
Lytton, H. & Romney, D. M. (1991) Parents' differential socialization of
boys and girls: A meta-analysis. Psychological Bulletin 109: 267-96.
Maccoby, E. E. (1988) Gender as a social category. Developmental Psychology
24: 755-65.
Maccoby, E. E. (1990) Gender and relationships: A developmental account.
American Psychologist 45: 513-20.
Maccoby, E. E. & Jacklin, C. N. (1974) The psychology of sex differences.
Stanford University Press.
MacDonald, K. B. (1988) Social and personality development: An evolutionary
synthesis. Plenum Press.
Manson, J. H. & Wrangham, R. W. (1991) Intergroup aggression in
chimpanzees and humans. Current Anthropology 32: 369-90.
Marsh, H. W., Smith, I. D. & Barnes, J. (1985) Multidimensional
self-concepts: Relations with sex and academic achievement. Journal of
Educational Psychology 77: 581-96.
Marshall, S. P. & Smith, J. D. (1987) Sex differences in learning
mathematics: A longitudinal study with item and error analyses. Journal of
Educational Psychology 79: 372-83.
Masters, M. S. & Sanders, B. (1993) Is the gender difference in mental
rotation disappearing? Behavior Genetics 23: 337-41.
Mayer, R. E. (1985) Mathematical ability. In: Human abilities: An
information processing approach, ed. R. J. Sternberg. Freeman.
McGee, M. G. (1979) Human spatial abilities: Psychometric studies and
environmental, genetic, hormonal, and neurological influences. Psychological
Bulletin 86: 889-918.
McGuinness, D. (1985) When children don't learn. Basic Books.
McGuinness, D. (1993) Gender differences in cognitive style: Implications
for mathematics performance and achievement. In: The challenge of mathematics
and science education: Psychology's response, ed. L. A. Penner, G. M. Batsche,
H. M. Knoff, & D. L. Nelson. American Psychological Association.
McGuinness, D. & Pribram, K. H. (1979) The origins of sensory bias in
the development of gender differences in perception and cognition. In: Cognitive
growth and development: Essays in memory of Herbert G. Birch, ed. M. Bortner.
Brunner/Mazel.
Meece, J. L., Parsons, J. E., Kaczala, C. M., Goff, S. B. & Futterman,
R. (1982) Sex differences in math achievement: Toward a model of academic
choice. Psychological Bulletin 91: 324-48.
Meece, J. L., Wigfield, A. & Eccles, J. S. (1990) Predictors of math
anxiety and its influence on young adolescents' course enrollment intentions
and performance in mathematics. Journal of Educational Psychology 82: 60-70.
Mills, C. J., Ablard, K. E. & Stumpf, H. (1993) Gender differences in
academically talented young students' mathematical reasoning: Patterns across
age and subskills. Journal of Educational Psychology 85: 340-346.
Mitchell, G. (1981) Human sex differences: A primatologist's perspective.
Van Nostrand Reinhold Company.
Moore, E. G. & Smith, A. W. (1987) Sex and ethnic group differences in
mathematics achievement: Results from the national longitudinal study. Journal
for Research in Mathematics Education 18: 25-36.
Nevin, M. (1973) Sex differences in participation rates in mathematics and
science at Irish schools and universities. International Review of Education
19: 88-91.
Oliver, M. B. & Hyde, J. S. (1993) Gender differences in sexuality: A meta-analysis.
Psychological Bulletin 114: 29-51.
Osborne, R. T. & Lindsey, J. M. (1967) A longitudinal investigation of
change in the factorial composition of intelligence with age in young school
children. Journal of Genetic Psychology 110: 49-58.
Ounsted, C. & Taylor, D. C. (ed.) (1972) Gender differences: Their
ontogeny and significance. Churchill Livingstone.
Pattison, P. & Grieve, N. (1984) Do spatial skills contribute to sex
differences in different types of mathematical problems? Journal of Educational
Psychology 76: 678-89.
Pepperberg, I. M. (1987) Evidence for conceptual quantitative abilities in
the African grey parrot: Labeling of cardinal sets. Ethology 75: 37-61.
Peterson, P. L. & Fennema, E. (1985) Effective teaching, student
engagement in classroom activities, and sex-related differences in learning
mathematics. American Educational Research Journal 22: 309-35.
Pinker, S. & Bloom, P. (1990) Natural language and natural selection.
Behavioral and Brain Sciences 13: 707-84.
Pratto, F., Sidanius, J. & Stallworth, L. M. (1993) Sexual selection and
the sexual and ethnic basis of social hierarchy. In: Social stratification and
socioeconomic inequality Volume 1: A comparative biosocial analysis, ed. L.
Ellis. Praeger.
Raymond, C. L. & Benbow, C. P. (1986) Gender differences in mathematics:
A function of parental support and student sex typing? Developmental Psychology
22: 808-19.
Resnick, S. M., Berenbaum, S. A., Gottesman, I. I. & Bouchard, T. J.,
Jr. (1986) Early hormonal influences on cognitive functioning in congenital
adrenal hyperplasia. Developmental Psychology 22: 191-98.
Rivera-Batiz, F. L. (1992) Quantitative literacy and the likelihood of
employment among young adults in the United States. Journal of Human Resources
27: 313-28.
Roe, A. (1953) A psychological study of eminent psychologists and
anthropologists, and a comparison with biological and physical scientists.
Psychological Monographs: General and Applied 67 (No. 352).
Rohner, R. P. (1976) Sex differences in aggression: Phylogenetic and
enculturation perspectives. Ethos 4: 57-72.
Rozin, P. (1976) The evolution of intelligence and access to the cognitive
unconscious. In: Progress in psychobiology and physiological psychology, ed. J.
M. Sprague & A. N. Epstein. Academic Press.
Ruff, C. (1987) Sexual dimorphism in human lower limb bone structure:
Relationship to subsistence strategy and sexual division of labor. Journal of
Human Evolution 16: 391-416.
Rumbaugh, D. M. & Washburn, D. A. (1993) Counting by chimpanzees and
ordinality judgments by macaques in video- formatted tasks. In: The development
of numerical competence: Animal and human models, ed. S. T. Boysen & E. J.
Capaldi. Erlbaum.
Saxe, G. B. (1982) Culture and the development of numerical cognition:
Studies among the Oksapmin of Papua New Guinea. In: Children's logical and
mathematical cognition: Progress in cognitive development research, ed. C. J.
Brainerd. Springer- Verlag.
Scarr, S., & McCarthy, K. (1983) How people make their own environments:
A theory of genotype-->environment effects. Child Development 54: 424-35.
Schoenfeld, A. H. (1985) Mathematical problem solving. Academic Press.
Sells, L. W. (1980) The mathematics filter and the education of women and
minorities. In: Women and the mathematical mystique, ed. L. H. Fox, L. Brody
& D. Tobin. The Johns Hopkins University Press.
Senk, S. & Usiskin, Z. (1983) Geometry proof writing: A new view of sex
differences in mathematics ability. American Journal of Education 91: 187-201.
Serbin, L. A. & Connor, J. M. (1979) Sex typing of children's play
preferences and patterns of cognitive performance. Journal of Genetic
Psychology 134: 315-16.
Shepard, R. N. (1994) Perceptual-cognitive universals as reflections of the
world. Psychonomic Bulletin & Review 1: 2-28.
Sherman, J. A. (1967) The problem of sex differences in space perception and
aspects of intellectual functioning. Psychological Review 74: 290-99.
Sherman, J. (1980) Mathematics, spatial visualization, and related factors:
Changes in girls and boys, grades 8-11. Journal of Educational Psychology 72:
476-82.
Sherman, J. (1981) Girls' and boys' enrollments in theoretical math courses:
A longitudinal study. Psychology of Woman Quarterly 5: 681-89.
Sherman, J. A. (1982) Mathematics the critical filter: A look at some
residues. Psychology of Women Quarterly 6: 428-44.
Siegler, R. S. (1986) Unities across domains in children's strategy choices.
In Perspectives for intellectual development: Minnesota symposia on child
psychology, ed. M. Perlmutter. Erlbaum.
Siegler, R. S. (1987) The perils of averaging data over strategies: An
example from children's addition. Journal of Experimental Psychology: General
116: 250-64.
Siegler, R. S. & Crowley, K. (1994) Constraints on learning in nonprivileged
domains. Cognitive Psychology 27: 194-226.
Silverman, I. & Phillips, K. (1993) Effects of estrogen changes during
the menstrual cycle on spatial performance. Ethology and Sociobiology 14:
257-70.
Simon, T. J., Hespos, S. J. & Rochat, P. (in press) Do infants
understand simple arithmetic? A replication of Wynn (1992). Cognitive
Development.
Smuts, B. B. & Gubernick, D. J. (1992). Male-infant relationships in
nonhuman primates: Paternal investment or mating effort? In: Father-child
relations: Cultural and biosocial contexts, ed. B. S. Hewlett. Aldine de
Gruyter.
Song, M. J. & Ginsburg, H. P. (1987) The development of informal and
formal mathematical thinking in Korean and U.S. children. Child Development 58:
1286-96.
Spearman, C. (1927) The abilities of man. MacMillan.
Stanley, J. C., Huang, J. & Zu, X. (1986) SAT-M scores of highly
selected students in Shanghai tested when less than 13 years old. College Board
Review 140: 10-13, 28- 29.
Starkey, P. (1992) The early development of numerical reasoning. Cognition
43: 93-126.
Starkey, P. & Cooper, R. G., Jr. (1980) Perception of numbers by human
infants. Science 210: 1033-35.
Starkey, P., Spelke, E. S. & Gelman, R. (1983) Detection of intermodal
numerical correspondences by human infants. Science 222: 179-81.
Starkey, P., Spelke, E. S. & Gelman, R. (1990) Numerical abstraction by
human infants. Cognition 36: 97-127.
Steinkamp, M. W., Harnisch, D. L., Walberg, H. J. & Tsai, S. L. (1985)
Cross-national gender differences in mathematics attitude and achievement among
13-year-olds. The Journal of Mathematical Behavior 4: 259-77.
Stevenson, H. W., Lee, S. Y., Chen, C., Lummis, M., Stigler, J., Fan, L.
& Ge, F. (1990). Mathematics achievement of children in China and the
United States. Child Development 61: 1053-66.
Stevenson, H. W. & Stigler, J. W. (1992) The learning gap: Why our
schools are failing and what we can learn from Japanese and Chinese education.
Stevenson, H. W., Stigler, J. W., Lee, S. Y., Lucker, G. W., Kitamura, S. &
Hsu, C. C. (1985) Cognitive performance and academic achievement of Japanese,
Chinese, and American children. Child Development 56: 718-734.
Stillion, J. M. (1985) Death and the sexes: An examination of differential
longevity, attitudes, behaviors, and coping skills. Hemisphere Publishing
Corporation.
Strauss, M. S. & Curtis, L. E. (1981) Infant perception of numerosity.
Child Development 52: 1146-52.
Strauss, M. S. & Curtis, L. E. (1984) Development of numerical concepts
in infancy. In: Origins of cognitive skills: The eighteenth Carnegie symposium
on cognition, ed. C. Sophian. Erlbaum.
Susman, E. J., Inoff-Germain, G., Nottelmann, E. D., Loriaux, D. L., Cutler,
G. B. Jr. & Chrousos, G. P. (1987) Hormones, emotional dispositions, and
aggressive attributes in young adolescents. Child Development 58: 114-34.
Symons, D. (1979) The evolution of human sexuality. Oxford University Press.
Thomas, J. R. & French, K. E. (1985) Gender differences across age in
motor performance: A meta-analysis. Psychological Bulletin 98: 260-82.
Thompson, R. F., Mayers, K. S., Robertson, R. T. & Patterson, C. J.
(1970) Number coding in association cortex of the cat. Science 168: 271-73.
Thorndike, E. L. (1911) Individuality. The Riverside Press.
Thurstone, L. L. (1938) Primary mental abilities. Psychometric Monographs
(No. 1).
Thurstone, L. L. & Thurstone, T. G. (1941) Factorial studies of
intelligence. Psychometric Monographs (No. 2).
Travers, K. J. & Westbury, I. (1989) The IEA study of mathematics I:
Analysis of mathematics curricula. Pergamon Press.
Trivers, R. L. (1972) Parental investment and sexual selection. In: Sexual
selection and the descent of man: 1871-1971, ed. B. Campbell. Aldine.
Vandenberg, S. G. (1959) The primary mental abilities of Chinese students: A
comparative study of the stability of a factor structure. Annals of the New
York Academy of Sciences 79: 257-304.
Vandenberg, S. G. (1962) The hereditary abilities study: Hereditary
components in a psychological test battery. American Journal of Human Genetics
14: 220-37.
Vandenberg, S. G. (1966) Contributions of twin research to psychology.
Psychological Bulletin 66: 327-52.
Vandenberg, S. G. & Kuse, A. R. (1978) Mental rotations, a group test of
three-dimensional spatial visualization. Perceptual and Motor Skills 47:
599-604.
Vernon, P. A. (1983) Speed of information processing and general
intelligence. Intelligence 7: 53-70.
Vernon, P. E. (1965) Ability factors and environmental influences. American
Psychologist 20: 723-33.
Very, P. S. (1967) Differential factor structures in mathematical ability.
Genetic Psychology Monographs 75: 169-207.
Voyer, D., Voyer, S. & Bryden, M. P. (1995) Magnitude of sex differences
in spatial abilities: A meta-analysis and consideration of critical variables.
Psychological Bulletin 117: 250-70.
Wade, M. J. (1979) Sexual selection and variance in reproductive success.
American Naturalist 114: 742-47.
Washburn, D. A. & Rumbaugh, D. M. (1991) Ordinal judgments of numerical
symbols by macaques (Macaca mulatta). Psychological Science 2: 190-93.
West, B. H., Griesbach, E. N., Taylor, J. D. & Taylor, L. T. (1982) The
Prentice-Hall encyclopedia of mathematics. Prentice- Hall.
Wilcockson, R. W., Crean, C. S. & Day, T. H. (1995) Heritability of a
sexually selected character expressed in both sexes. Nature 374: 158-59.
Williams, C. L., Barnett, A. M. & Meck, W. H. (1990) Organizational
effects of early gonadal secretions on sexual differentiation in spatial
memory. Behavioral Neuroscience 104: 84-97.
Wilson, M. & Daly, M. (1985) Competitiveness, risk taking, and violence:
The young male syndrome. Ethology and Sociobiology 6: 59-73.
Wise, L. L. (1985) Project TALENT: Mathematics course participation in the
1960s and its career consequences. In: Women and mathematics: Balancing the
equation, ed. S. F. Chipman, L. R. Brush & D. M. Wilson. Erlbaum.
Witelson, S. E. (1987) Neurobiological aspects of language in children.
Child Development 58: 653-88.
Wood, R. (1976) Sex differences in mathematics attainment at GCE ordinary
level. Educational Studies 2: 141-60.
Wynn, K. (1992) Addition and subtraction by human infants. Nature 358:
749-50.
Zaslavsky, C. (1973) Africa counts: Number and pattern in African culture.
Prindle, Weber, & Schmidt.