NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Academy of Sciences (US), National Academy of Engineering (US), and Institute of Medicine (US) Committee on Maximizing the Potential of Women in Academic Science and Engineering. Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering. Washington (DC): National Academies Press (US); 2006.

Cover of Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering

Biological, Social, and Organizational Components of Success for Women in Academic Science and Engineering.

Show details

Panel 1: Cognitive and Biological Contributions

  • Panel Summary
  • Gender Differences and Similarities in Abilities, Janet Hyde, Department of Psychology, University of Wisconsin at Madison
  • Sexual Dimorphism in the Developing Brain, Jay Giedd, National Institute of Mental Health, National Institutes of Health
  • Environment-Genetic Interactions in the Adult Brain: Effects of Stress on Learning, Bruce McEwen, The Rockefeller University
  • Biopsychosocial Contributions to Cognitive Performance, Diane F. Halpern, Berger Institute for Work, Family, and Children, Claremont McKenna College
  • Selections from the Question and Answer Session Moderated by committee member Ana Marie Cauce


The panel considered whether there are differences between males and females in brain development and in average performance on cognitive tasks and whether those differences account for the large discrepancies in male and female representation among academic scientists.

Janet Hyde, of the University of Wisconsin-Madison, proposed the “novel concept of gender similarities” in cognitive abilities, noting that the mathematical, verbal and spatial skills involved in scientific work are all gender-stereotyped. Meta-analyses of 100 studies of math ability involving 3 million persons, including nine state assessments, show that the highly touted and widely reported gender differences in mathematical ability are in fact small or insignificant.

Diane Halpern, of Claremont McKenna College, observed that men and women are in fact both similar and different and “what you see depends on where you look.” The differences or similarities found depend on which tests and measures are used. She also emphasized that nature and nurture form a “false dichotomy,” are not independent variables, and “do not just interact.” The factors are instead “inextricably intertwined” because experience alters the biological underpinnings of behavior, and the resultant biology influences the types of experiences people have. Instead of the old two-part paradigm, she proposed a biopsychosocial conceptualization of the issue and the recognition that even small differences may have large effects over time because small effects accumulate into large ones.

Jay Giedd, of the National Institute of Mental Health, presented data from magnetic resonance imaging (MRI) studies of brain structure and development during adolescence showing both gender differences in the trajectory of brain development and the strong and lifelong influence of experience on the brain. MRI studies show “gray boxes,” not individual neurons, and behavioral interpretations are therefore “speculative.” The sex hormones estrogen and testoster-one are present both in males and females, and play a role in brain development, although hormones are not sole factors driving sex differences in the brain. Male brains show more morphological variance than female brains, but observations are based on group averages and not individuals, and overall, the brains of males and females are more alike than different.

Panelist Bruce McEwen, of Rockefeller University, presented evidence of complex sex differences in nonhuman brain response to stress and of the brain’s high adaptability and plasticity throughout the lifespan. Males and female humans differ in the processes and priorities they use in processing information. Genes, hormones, and experience exert different influences on human males and females, he concluded, but the cognitive differences between men and women appear to involve differing strategies of information processing rather than different “abilities.”


Janet Hyde

Department of Psychology, University of Wisconsin-Madison

Janet Hyde’s presentation emphasized what she called “the novel concept of gender similarities” and focused on mathematical, verbal, and spatial abilities as basic to science ability. Those abilities are “gender stereotyped,” with boys believed to excel on mathematics and spatial tests and girls on verbal measures.1

Hyde described the power of meta-analysis (see Box 1-1) and discussed a particularly large study of gender differences in mathematics performance that pooled the results of 100 studies that tested more than 3 million people and included a wide variety of data sources, such as assessments from nine states. Averaged over all samples of the general population, the d was equal to minus 0.05, “a tiny gender difference.” Another team of investigators obtained very similar results using somewhat different meta-analytic techniques.2

Box Icon

BOX 1-1

Meta-Analysis. Hundreds of studies examine gender differences in performance. Rather than conduct an additional study, one can synthesize the existing studies to find an overall outcome. Meta-analysis refers simply (more...)

Might there be an increasing gender gap in performance with age? Second, do the mathematics tests tap lower level math computation, or a deeper conceptual understanding of mathematics and complex problem solving, which is needed to do science?

Using meta-analytic methods to investigate these questions, Hyde found that girls are better than boys at computation by a small amount in elementary and middle school. For the deeper understanding of mathematical concepts, she found no gender difference at any age level. Finally, at the highest cognitive level, complex problem-solving, she found no gender difference in elementary school or middle school, but a small difference among high school and college students. Although that difference deserves attention, it is not large.

The important point is that within-gender differences are enormous compared to between-gender differences.

—Janet Hyde

One explanation for the gender difference in problem-solving favoring high-school and college-age males is the difference in patterns of course taking. Girls have been less likely to take optional advanced mathematics classes in high school, although this gender gap has closed in the last five years. Girls now take calculus in high school at the same rate as boys. Nonetheless, they are less likely to take science courses in high school than boys, especially in chemistry and physics. This handicaps girls in pursuing science careers, and it also handicaps their performance on standardized mathematics tests, because students experience mathematical problem-solving in physics and chemistry classes.

Concerning gender differences in verbal ability, meta-analysis of 165 studies representing the testing of 1.4 million persons showed superior performance by females but the difference is very small (d = −0.11).3The question of gender differences in spatial ability, a relevant skill in many fields of science and engineering, is complicated because there are many types of spatial ability and many tests to measure them. With regard to gender differences in three dimensional mental rotation, which is crucial in fields such a engineering,4 two meta-analyses have been conducted. One found a large gender difference favoring males, and the other found a medium gender difference favoring males,5 both more substantial than for mathematical and verbal abilities. That does not mean that girls cannot succeed at engineering; research shows that spatial skills can be trained.6

One major factor in determining mathematics performance is student high school course choice. In investigating what factors influence adolescents’ choice of courses and careers, Eccles found that students value what they think they will learn in a course, and that is heavily influenced by intended career.7 Many occupations in the U.S. are highly gender-segregated. That makes it more likely that girls will not imagine themselves in science or engineering careers and therefore they will not value mathematics or physics courses as much as boys do.

Parents play an important role. Research shows that even in elementary school, parents estimate the math ability of sons to be higher than those of daughters, despite the absence of any gender difference in actual grades or test scores at this point. One particularly impressive longitudinal study found that mothers’ estimates of their 6th grader’s likelihood of mathematics success predicted the child’s actual mathematics career choice at age 25.8

Schools play a third important role on the gender difference in advanced mathematics and science performance. Research shows, for example, that hands-on laboratory experiences in the physical sciences improved the science achievement of girls but not of boys, and helped to close the gender gap in achievement.

Cultural influences at the broadest level also play a role. In a cross-national study of 5th graders’ math performance,9 one could focus on the small difference in performance between girls compared with boys. However, the bigger picture shows that girls in Taiwan and Japan dramatically outperform American boys. Many features probably account for this, among them differences in the way mathematics is taught, in cultural values attached to mathematics, and in different attitudes about the importance of ability vs. effort in producing excellent performance.

FIGURE 1-1. Cross-cultural differences in fifth-grade mathematics performance.


Cross-cultural differences in fifth-grade mathematics performance.

SOURCE: M Lummis and HW Stevenson (1990). Gender differences in beliefs and achievement: A cross-cultural study. Developmental Psychology 26:254-263.

Another study looked at the magnitude of the gender difference in mathematics performance in different countries and correlated it with the United Nations standardized measure of gender stratification.10 The correlation between mathematics performance and the percentage of women in the paid workforce was an impressively large −0.55 across nations. Countries with the greatest gender stratification tended to have the largest gender difference favoring males.

All those findings led Hyde to propose the gender similarities hypothesis.11 She subjected 46 relevant meta-analyses to a meta-analysis. The studies spanned a wide range of psychological characteristics, including abilities, communication, aggression, leadership, personality and self-esteem. She found 78% of the gender differences effect sizes were small or close to 0. Psychologically, women and men are more similar than they are different. Large gender differences are found in a few cases, but the big picture is one of gender similarities.

On the basis of these data, Hyde suggested some policy recommendations: (1) a spatial learning curriculum should be instituted in primary and secondary schools, (2) colleges of engineering should have a spatial skills training program for entering students, (3) four years of math and four years of science should be required in high school or at least for university admission, (4) the mathematics curriculum in many states needs far more emphasis on real problem solving, and (5) teachers and high school guidance counselors need to be educated about the findings on gender similarities in mathematics performance, or teachers will believe the stereotypes about girls’ mathematics inferiority that pervade our culture and those expectations will be conveyed to the students.


Jay Giedd

National Institute of Mental Health, National Institutes of Health

Jay Giedd began by noting his focus on the adolescent brain. In child psychiatry nearly everything has different prevalences, ages of onset, and symptomatology between boys and girls and nearly every disorder is more common in boys. His studies use MRI, magnetic resonance imaging, which because it does not require radiation can be used in children to perform longitudinal studies.

To the MRI machine, the brain is boxes of gray or white measuring about 1 cubic millimeter. Within each of these boxes are millions of neurons and trillions of synaptic connections. Using much finer resolution microscopic techniques, one can see synapses and connections, but MRI currently cannot do that. MRI pictures and images can bequite colorful, but interpretations are necessarily speculative.

—Jay Giedd

What we call the gray matter consists mostly of the neuronal cell bodies, where the nucleus and the DNA are housed; the antenna-like dendrites reaching for connections to other brain cells; and the terminal branches of the axons, the location of the synapses, and the connections to other brain cells. The white matter is myelin, the insulation material wrapped around the axon that speeds communication between the brain cells.

Giedd and his colleagues performed longitudinal MRI scans of 2,000 subjects. They found that white matter volume increased at least through the fourth decade in women and through the third in men (Figure 1-2). At no time during development did white matter volume decrease.

FIGURE 1-2. Longitudinal development of white matter.


Longitudinal development of white matter. SOURCE: JN Giedd, J Blumenthal, NO Jeffries, FX Castellanos, H Liu, A Zijdenbos, T Paus, AC Evans, and JL Rapaport (1999). Brain development during childhood and (more...)

The white matter has been of interest in the study of sexual brain-structure differences, or sexual dimorphism, because one of the first reports of a brain difference not related to reproduction concerned the corpus callosum, the white matter tract connecting the two brain hemispheres. In over 100 published papers, the results are inconclusive—the corpus callosum of females is bigger than, smaller than, or not different from that of males. The key to understanding these results is in considering developmental windows. At young ages the corpus callosum is sexually dimorphic; between ages 9 and 14 it is not; and then it becomes so again. These changes happen throughout life.

Brain areas have intersecting developmental trajectories. This is a very important concept in how to interpret the findings. Often, the literature will combine data from people across seven or eight decades, andreport that average as the difference between male and female brains.

—Jay Giedd

The most robust sex difference is total brain size. From autopsy studies, even when correcting for total body mass, male brains have been found to be about 10% larger than female brains, but bigger isn’t better and size is not related to intelligence. A lot of the literature is really murky on how to account for total brain size difference.

The other part that MRI can see—the gray matter—has a distinct developmental trajectory from that of white matter. Instead of a general linear increase in volume, gray matter has an upside down “U” path in development. Changes in cortical thickness are not due to an increase in the number of neuronal cells, but to an increase in arborization, or the number of branches, twigs, and roots of existing individual neurons. Although both progressive and regressive processes occur throughout life, during childhood there is a net increase in the degree of branch ing and during adolescence there is a net decrease. Growth reaches a peak in the frontal part of the brain at 11 in girls and 14 years in boys. Pruning then begins: the cells and connections that are used survive and flourish, and those that are not wither and die.

There is a lot of regional variation in the process. Maturation starts in the parts of the brain needed to keep us alive, such as those controlling heart rate and breathing. The next parts of the brain to mature are those involved in processing the five senses, followed by the parts of the brain that link together the primary senses. Then there is a cascade of hierarchies linking the linkings. The final stop is the frontal lobe, which doesn’t reach adult levels until about age 25.

By adulthood, once you correct for the total brain-size differences, the sex differences are quite subtle. But if you look at the path the brains took to get there, the differences are far more robust. It’s the journey, not the destination.

—Jay Giedd

The most variable parts of the brain seem to be those that mature last, and are the least heritable. The structure that we have examined thus far that is the most different between males and females is the cerebellum. Because it is one of the last brain areas to mature, the cerebellum is under the influence of the environment for a long period. Accounting for overall brain size increase, the cerebellum is larger and it reaches adult volume later in males than in females. Overall, male brains have a greater variation in cortical thickness; this is a very robust phenomenon that occurs throughout the brain.

Giedd summarized with two points: First, male and female adolescent brains are much more alike than different; there is enormous overlap. Second, with regard to developmental trajectories, there are more marked sex differences. Male brain structure appears more variable. Whether the variability is biological or social in origin, the data are robust. Work is underway on the effects of sex chromosomes and hormones. In ending, Giedd emphasized that differences are group average differences, and are not to be implied as constraints for individual boys or girls.


Bruce McEwen

The Rockefeller University

Bruce McEwen presented data on sex-based differences in the effects of stress, which have implications for learning. He and his colleagues study brain regions that are involved in memory, emotions, and executive function or deci sion making. He commented on the translation of animal-model studies to humans, and complemented the discussion on the continuing interaction throughout the life course between genes, hormones, and environment/experience.12

The adult brain is a very adaptable organ, and through our adult life there is a continual functional and structural remodeling.

—Bruce McEwen

McEwen briefly summarized the plasticity literature. When the brain is damaged, there is collateral sprouting and functional reprogramming in many cases. Even without damage, there is continual remodeling of connections with use and disuse, which has been demonstrated for the visual system and also in the motor system in terms of practice effects, such as in playing musical instruments and doing repetitive motor tasks. There are progenitor cells and even some stem cells in the adult brain; and in the dentate gyrus of the hippocampus and the olfactory bulb there is a continuous replacement of nerve cells throughout adult life. There is remodeling of the dendrites—the tree-like structures of neurons— and of synaptic connections in animals undergoing both acute and chronic stress.

Examples from an ongoing study on the prefrontal cortex illustrate the latter point.13 In male rats that have been repeatedly stressed, neuronal dendrites become shorter and less branched and the number of synaptic connections is reduced. The overall reduction is as much as 30%, which has functional implications. However, in the amygdala, an area of the brain that is associated with fear, with aggression and emotional responses, repeated stress of the same kind causes neurons and dendrites to grow and increase their synaptic connectivity.14 That may explain why repeated stress causes animals to become more fearful and more aggressive.

The sex hormones testosterone and estradiol have effects throughout adult life and widespread influences throughout the brain.15 Receptors for both sex hormones are found in most brain areas, meaning that hardly any area of the brain is not influenced by circulating sex hormones. There is also evidence of a direct effect of the X and Y chromosomes on certain aspects of brain development and differentiation.

Testosterone and estradiol and receptors for them are present in both males and females. Their effects in the two sexes are subtly different, depending on developmental programming. For example, estrogen affects motor coordination, vulnerability to seizures, aspects of the premenstrual syndrome or pre-menstrual dysphoric disorder, depression, vulnerability to stroke, and the amount of damage from stroke, pain mechanisms, cognitive function, and vulnerability for dementia. Estrogen influences functions both at the level of the cell nucleus through the traditional mechanism, but also through relatively newly discovered cell-surface signaling mechanisms. Similarly broad effects are seen with testosterone and other androgens in males.

There is virtually no function that is not influenced by reproductive hormones.

—Bruce McEwen

These broad effects should be kept in mind when thinking about how the male and female brain, with and without circulating sex hormones, responds to stressful experiences. We know that acute stress generally enhances the learning of survival-related information. Repeated stress results in adaptive plasticity. The resulting changes in dendritic branching and synaptic connectivity in areas like the amygdala, prefrontal cortex, and hippocampus, an area of the brain involved in memory, are largely reversible: when the stress ends, these effects disappear.

Recent evidence indicates that a single episode of traumatic stress results in a delayed and relatively prolonged increase in anxiety in the animal and actual growth of new synaptic connections in the amygdala and the prefrontal cortex. There is also evidence that repeated stress increases vulnerability to other traumas such as a stroke or a seizure.

In the response to stress, there are sex differences in brain remodeling. Female rats do not show the increased dendritic branching seen in the hippocampus of male rats. In contrast, dendritic branching in the amygdala appears to be enhanced by estrogen. When circulating estrogen in female rats is depleted by removing the ovaries, the stress response becomes similar to that in a male rat. Other studies have shown a greater initial effect of acute stress in the male on food intake and fear. It also appears that it takes longer for the female rat than the male rat to recover to baseline levels from a stressor.

Sex differences are neurobiologically and psychologically more complicated than we had thought. There are opposite effects in males and females of an acute stress on the conditioning of a classical Pavlovian response. Work of Gwendolyn Wood and Tracey Shors16 shows that conditioning in male rats is enhanced by stress. Exactly the same stress regimen in female rats profoundly suppresses conditioning. These results can be reversed by manipulating hormonal sex early in development. More recently, Shors has shown that giving the male and female rat control over the amount of shock makes the sex differences disappear.

How might some of this translate from animals to humans? McEwen suggested the key may lie in behavioral strategy. Research on rats in a water maze, where they have to swim and find a hidden platform to rest on, shows that the male and females tend to use different exploratory strategies. Without spatial cues, male rats reach the platform faster. When spatial cues are provided, females decrease the time it takes to reach the platform and do as well as or better than males. Karyn Frick and colleagues put student volunteers into an outdoor spatial maze tested memory of local contextual cues.17 Men and women did not differ in their performance in the spatial maze but women had a better memory of objects and their location than men did.

Arguments go back and forth, and the data makes it much more complicated to reach some simple generalizations.

—Bruce McEwen

In summary, McEwen explained there are sex differences that are products of genes, of hormones, and of experience throughout the life span. Males and females do respond differently to stressors, although the differences are complex and depend on the kind of stressor and the circumstances. There appears to be modulation by circulating sex hormones, at least in the animal models. What is described in the animal literature, and also perhaps in some of the human literature, is that there are differences in processing—maybe in priorities and strategies—that are far more important than what are commonly called “abilities.”


Diane F. Halpern

Berger Institute for Work, Family, and Children, Claremont McKenna College

Diane F. Halpern began her presentation referring to a paper she had written several years ago, entitled, “What You See Depends on Where You Look.”18 Whether male and female cognitive abilities seem similar or different depends on which data are used. To address the question, whether fewer women than men have the ability to become scientists and engineers, requires an examination of how men and women are similar and different.19

We are not talking about whether men and women are similar or different, which is debatable, because in fact women and men are both similar and different. The real question is in what ways are men and women similar and different, and how to understand the relevance of the similarities and differences.

—Diane F. Halpern

Women are graduating in very high numbers with degrees in science fields, so women obviously have the innate ability to do science. But women are not graduating in equal numbers from all of the sciences. To explain this discrepancy, some people have said that women prefer biological sciences, whereas men prefer physical sciences. Alternatively, psychologists have said that women seem to prefer people-oriented careers and men prefer thing-oriented careers. Career choice and trajectory involves a complex of traits, including abilities, interests, personality variables, opportunities, and the knowledge of available career options.

Society has many sex differences. One is the wage gap, which is not just between men and women. Overwhelmingly women are poorer than men, but the largest wage differences are between women who have children and other people. Women have fewer leadership positions overall, not just in science, not just in academia, but in corporations. College students tell us gender differences are a thing of the past, but men in college spend several more hours a week playing video games than women, among many other differences.

We don’t like to talk about sex differences. Sex differences are simply not popular. It’s much more popular to talk about similarities, there is no doubt about that. But when we talk about differences, then at least we much prefer to acknowledge that they are embedded in environment. But this concept is embedded in the false idea that nature and nurture form a dichotomy. There is not a number out there that we can pin on nature or nurture. We have got to get away from the idea of a nature/nurture dichotomy and interaction, because nature and nurture are not independent variables, and they do not merely interact. We need to replace that whole idea with a model that is biopsychosocial. Nature and nurture are inextricably intertwined; they cannot be separated.

—Diane F. Halpern

Experience alters the biological underpinnings of behavior which in turn influences the experiences to which we are exposed. A graphic model of biopsychosocial interactions is presented in Figure 1.3.

FIGURE 1-3. Biopsychosocial model.


Biopsychosocial model. SOURCE: DF Halpern (2000). Sex Differences in Cognitive Abilities. 3rd Ed. Mahwah, NJ: Erlbaum.

Some cognitive tasks show sex differences. Some of these differences are lost in aggregated data. Halpern disagreed with Janet Hyde regarding assigning values to small and large effect sizes, stating that small differences in fact accumulate to make very large differences.20

Some differences that favor females:

  • Rapid access to and use of phonological, semantic, episodic information and long-term memory.
  • Production and comprehension of complex prose.
  • School grades and tests closely aligned to school curricula.
  • Fine motor tasks and speech articulation.
  • Perceptual threshold tasks.

Some differences that favor males:

  • Visual transformations and visuospatial working memory.
  • Moving objects and aiming at targets.
  • Fluid reasoning tasks
  • Novel tasks unrelated to things that are taught in school.

Males are overrepresented in both extremes of performance—among the retarded and the gifted. That finding has been used to explain why there are fewer females in science and mathematics, but does not explain why there are fewer females in these professions overall. Not just are there fewer gifted women in science and mathematics, there are just fewer women.

International data also show sex differences. The PERLS reading study shows statistically significant effects on reading literacy at age 15, favoring girls. The mathematics test score difference is rather unimpressive and tends to be insignificant. The science test score difference at 8th grade tends to favor males and gets larger in college and graduate school as the student samples become more selective.

A test-grade disparity is part of the puzzle. Girls get higher grades in school in every subject, even when they are getting lower grades on the achievement tests. Women are graduating at a substantially higher rate than males from college, 133 women for every 100 men.

Despite those successes, women score significantly lower on many tests of science and mathematics, particularly on tests that have questions not closely related to materials taught in school. This discrepancy leads some to ask whether teachers in schools are biased against boys or whether achievement tests are biased against girls.

Cognitive processes are involved. As Bruce McEwen discussed, some have suggested that males and females are using different problem-solving strategies. Like Janet Hyde, Halpern called for education in visuospatial skills. But in trying to answer the underlying question, are there too few women with the highest levels of ability to be scientists and engineers, Halpern pointed beyond cognitive processes to a larger framework in academe: the tenure system. Marriage and having children have an adverse effect on the research productivity of women in academia.

That tenure clocks and biological clocks run in the same time zone is the more likely and proximal cause for some of these problems than cognitive differences.

—Diane F. Halpern

The take-home message: females and males are similar and different, depending on what is measured. The types and sizes of cognitive differences vary between men and women. Some of the measures favor females, some favor males. There are consistent differences internationally. Halpern called for a biopsychosocial model to replace the nature/nurture dichotomy and for consideration of the larger academic and societal context.


DR. AGOGINO: Hi, I’m Alice Agogino from the University of California at Berkeley. I have a question about how authentic these assessments or these features are in terms of actual practice and success. Janet, you mentioned the Linn Peterson study, a meta-analysis on spatial reasoning and found the greatest differences were for three-dimensional rotation, as measured by the Shepherd Test. I worked with Marcia Linn when I taught a Mechanical Engineering freshman design class where spatial reasoning skills were important. We looked at expert spatial reasoners in industry and found that they did even worse on some of those tests than the students at the lowest end of the scale. The big difference was timing. If we added 30 seconds onto a test, we got rid of a lot of the differences. We did a two hour workshop and developed strategies that improved the performance of both men and women and got rid of all the gender differences in performance on these tests. My question is, before we start creating courses, do they really matter in terms of success, and their authenticity for success in practice?

DR. HALPERN: People often ask that question. Spatial reasoning is correlated with grades in engineering schools; it’s been used in dental schools as a grade predictor; and the ability to see things from multiple angles is used in imagining what a molecule will look like if you rotate it in space. In some of my own work recently we have found that males were imaging a lot of the material when they were reading it, and some of the females also. While we are teaching people how to read, we’re teaching people how to do math. Cognitively this is another one of those dimensions that we have just not paid attention to in the educational system.

DR. BICKLE: Janet Bickle, formerly of the Association of American Medical Colleges, and now a career development coach. I wonder if anyone else noticed this week, a very small article in the Post that was a study of monkeys, finding that male monkeys were more likely to play with cars, and the female monkeys were more likely to play with dolls, including looking at the dolls’ bottoms. And the males actually playing with the cars the way little boys do. I was wondering what sense the panelists could help us make of this type of finding.

DR. HYDE: I think partly because I’m a meta-analyst, I’m very keenly aware of how many behavioral studies in psychology don’t replicate. And so, I would really want to see that study replicated before I made any interpretations, because studies like that are so quickly picked up by the media. Everybody loves them. And then there are 10 failures to replicate, and they never get attention. I think we really have to ask for the standard of replicability in a lot of these phenomena.

DR. MCEWEN: I might add that while I have no comment about that particular study, it’s well established in animal behavior studies on both rodents and on rhesus monkeys that there is an androgen-dependent rough and tumble play behavior which is very typical of the male of both species, and can be influenced by testosterone, and can be produced by exposure of females at the right time of development to testosterone. So, there is a phenomenon there. How it has to do with playing with any particular toy, I have no idea.

DR. GIEDD: If the studies are done well, it is a great insight into the role of socialization and media exposure and all these other sort of things, and the biology itself. So, I think it’s a very worthwhile direction to pursue, if it’s done well.

DR. WEYUKER: I’m Elaine Weyuker. I’m at AT&T Labs, and I’m a member of the committee. In terms of the swimming rats, one of the things I was struck by was the female rats’ strategy was to swim along the edge, whereas the male strategy was to go down the middle and to look. But one of the other things I noticed was that you stuck the platform in the middle. And so, had you stuck the platform at the edge, it sounds to me like the female rats would have been the stars. Are we using as measures of “success” the things that the women don’t do as well?

DR. MCEWEN: The point you make gets back to this idea of strategy, and obviously, the way you set up the task can give you different results. I can give you more kinds of experiments not involving that swimming task, where again, you can establish that there are not only sex differences, but also giving estrogen to ovariectomized female rats actually improves their choice of a place strategy over a response strategy, perhaps by enhancing the function of the hippocampus over the function of other brain areas.

DR. WEYUKER: And what the measure of success is.

DR. MCEWEN: Yes, that’s a good point. But like the example from Karyn Frick’s studies, when you are looking at the memory of location and identity of objects, on the average the women did better than the men in remembering these things. That may contribute to the success of women in handling certain kinds of spatially-related, contextual tasks where they have to remember locations of things in order to make choices.

DR. GARMIRE: I’m Elsa Garmire from Dartmouth College. The subject of this convocation is women in academe. From my point of view, I would imagine that most of women in academe would be in perhaps let’s say the top 20% of whatever group that you are investigating. And what I want to know particularly in the meta-analyses, which seem to give you the average of all humans, have there been studies that have looked at the top 20%, and compared the top 20% of males to the top 20% of females in any of these studies?

DR. HYDE: There is a series of studies originally begun at Johns Hopkins by Julian Stanley and Camilla Benbow of gifted youth. They recruit mathematically precocious children in the seventh or eighth grade who score 700 or more on the mathematics SAT. Stanley and Benbow do find a disproportionate number of boys in their group compared with girls. I have never been able to pin down exactly how they recruit them though, because for example, if it’s partly by teacher recommendation, then you wonder if teachers don’t tend to see more mathematical talent in boys, even when it’s present in girls as well.

DR. GARMIRE: Yes, but they have started out already selecting. What I’m suggesting is in all of these studies, if one went back and said, okay we are not going to look at the average for everyone, we’re going to fit everyone to a bell curve, and then take the top 20% of that data, I think you could do a meta-analysis without any pre-selection of people and analyze exactly how males and females compare in the upper reaches.

DR. CAUCE: Those are a series of studies that I’m fairly familiar with. Part of what is interesting is that there are many more men in both tails of the performance distribution. But what is interesting is that even though you have more men than women in the tails, if you look at the differences in the career trajectories of the men and women in the upper tail, so we are talking about the upper 1% in terms of mathematics talent and ability, a much higher percentage of those men follow the trajectory into mathematics and science. Women are much more likely to go into particularly medicine and law than in science. I’m not aware of any studies that have tried to particularly truncate at about 20 or 25%.

DR. VOGT: Hi, Christina Vogt, National Academy of Engineering. I think that we need to look a little bit more at social determinants of engineering and science careers than spatio-visualization skills.

DR. HYDE: I agree, and I think some of the panels later today are going to be getting at some of the factors like that, so it’s definitely important.

DR. CAUCE: I couldn’t agree with you more. There is no question but that workplaces and how people react to them are different. But then also there is some work that suggests there might be some biologically based differences in motivations, so, that women would be motivated more towards going into social careers, which are defined, and I would say erroneously, as being non-science careers.

DR. SAENZ: My name is Delia Saenz. I’m a social psychologist at Arizona State University, and I do work on tokenism. Much of that work, at least in the early part of my career, demonstrated that tokens suffer cognitive deficits. I remember when I found the first result, I wanted to hide the research, because I thought, okay, nobody is going to want to hire women or minorities, because they will bring them in, and they will do poorly, not because of their capacity, but rather because of the environmental configuration that is having them concerned with self-presentation. One of my best friends said, you know what? You got the same finding for males and for white males. So, it’s not a matter of who you are, but the context.

I agree with what you all suggested earlier, that if you match the person to the task, and you have a good fit, things will go better. And in fact, my more recent work on tokenism suggests that there are cognitive surfeits if you are a token. So, because you are concerned with self-presentation, you’re better able to take perspectives, and you are good at negotiating, which is a good thing, and it happens for women and minorities, as well as for males and whites. That’s very exciting. So, we will get to the point where we are not just focusing on differences in ability, but differences in outcome, differences in being able to make a living and having your contributions validated.



For more details, figures, and references, see Janet Hyde’s paper in Section 2.


LV Hedges and A Nowell (1995). Sex differences in mental test scores, variability, and numbers of high-scoring individuals. Science 270:364-365.


JS Hyde and MC Linn (1988). Gender differences in verbal ability: A meta-analysis. Psychological Bulletin 104:53-69.


M Hegarty and VK Sims (1994). Individual differences in mental animation during mechanical reasoning. Memory and Cognition 22(4):411-430.


MC Linn and AC Petersen (1985). Emergence and characterization of sex differences in spatial ability: A meta-analysis. Child Development 56:1479-1498; D Voyer, S Voyer, and MP Bryden (1995). Magnitude of sex differences in spatial abilities: A meta-analysis and consideration of critical variables. Psychological Bulletin 117:250-270.


Hyde referred to a study by Sheryl Sorby and her colleagues, who have developed a multi-media software program that improves the spatial performance of students and has improved the retention of women in the engineering major from 47% to 77%. See: N Boersma, A Hamlin, and S Sorby (2005). Work in progress—Impact of a remedial 3-D visualization course on student performance and retention. Presentation at 34th ASEE/IEEE Frontiers in Education Conference, October 20-23, 2004, Savannah, GA.http://fie​.engrng.pitt​.edu/fie2004/papers/1391.pdf.


JS Eccles (1994). Understanding women’s educational and occupational choices: Applying the Eccles et al. model of achievement-related choices. Psychology of Women Quarterly 18:585-610.


JE Jacobs and JS Eccles (1992). The influence of parent stereotypes on parent and child ability beliefs in three domains. Journal of Personality and Social Psychology 63(6):932-44.


M Lummis and HW Stevenson (1990). Gender differences in beliefs and achievement: A cross-cultural study. Developmental Psychology 26:254-263.


DP Baker and DP Jones (1993). Creating gender equality: Cross-national gender stratification and mathematical performance. Sociology of Education 66:91-103.


JS Hyde (2005). The gender similarities hypothesis. American Psychologist 60:581-592.


For an overview, see BS McEwen and EN Lasley (2005). The end of sex as we know it. Cerebrum 7(4):65-79.


JJ Radley, AB Rocher, M Miller, WG Janssen, C Liston, BS McEwen, and JH Morrison (2005). Repeated stress induces dendritic spine loss in the rat medial prefrontal cortex. Cerebral Cortex 16(3):313-320.


A Vyas, S Bernal, and S Chattarji (2003). Effects of chronic stress on dendritic arborization in the central and extended amygdala. Brain Research 965(1):290-294.


BS McEwen (1999). The molecular and neuroanatomical basis for estrogen effects in the central nervous system. Journal of Clinical Endocrinology and Metabolism 84(6):1790-1797.


GE Wood and TJ Shors (1998). Stress facilitates classical conditioning in males but impairs classical conditioning in females through activational effects of ovarian hormones. Proceedings of the National Academies of Sciences 95(7):4066-4071.


LJ Levy, RS Astur, and KM Frick (2005). Men and women differ in object memory, but not performance of a virtual radial maze. Behavioral Neuroscience 119(4):853-862.


DF Halpern (1989). The disappearance of cognitive gender differences: What you see depends on where you look. American Psychologist 44:1156-1158.


For more details, figures, and references, see Diane Halpern’s paper in Section 2.


V Valian (1999). Why So Slow: The Advancement of Women. Cambridge, MA: MIT Press.

Copyright © 2006, National Academy of Sciences.
Bookshelf ID: NBK23768


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.7M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...