Diversity Debate, Science / Tech, Top

Sex and STEM: Stubborn Facts and Stubborn Ideologies

Many academics in the modern world seem obsessed with the sex difference in engagement with science, technology, mathematics, and engineering (STEM) fields. Or rather they are obsessed with the fact that there are more men than women in some of these fields. There is particular concern about the lack of women in prestigious STEM fields, such as Ph.D.-level faculty positions, but surprisingly there is no concern about the under-representation of women in lower-level technical jobs, such as car mechanics or plumbing.

The concerned academics have been especially effective in convincing others, or at least intimidating them, into accepting their preferred interpretations regarding the source of these sex differences (as illustrated in the Google memo debate). These interpretations are not surprising and they include sexism, stereotype threat, and more recently implicit bias and microaggression. Each of these ideas has gained traction in the mainstream media and in many academic circles but their scientific foundations are shaky. In this essay, we’ll provide some background on the STEM controversy and consider multiple factors that might contribute to these sex differences.

The U.S. National Science Foundation reports that women are awarded 57 percent of undergraduate STEM degrees, but with substantial differences across fields. Women earn the majority of degrees in the life and social sciences but less than 20 percent of the degrees in computer science and engineering, sex differences that have held steady for several decades. The STEM debate is primarily about sex differences in educational and later occupational choices in inorganic fields, those focused on understanding non-living things. These differences are socially important because these tend to be prestigious occupations, and practically important because the different numbers of men and women in these fields contribute, in part, to the sex difference in earnings.1

Melba Roy at NASA, 1964

At the core of the obsession is the zeitgeist that there should be gender equity – equal outcomes – for anything of monetary or social value. The combination of an extreme agenda among some feminists and a stubborn sex difference has created a cottage industry focused on rectifying this ‘injustice.’ The federal governments in the U.S., U.K., and other Western nations have devoted hundreds of millions of dollars to interventions to close the gap. Some of the activities funded by these initiatives make sense and are possibly helpful in some ways, such as programs to increase interest in mathematics or programming among girls. Other programs, such as developing mentoring programs exclusively for women who are junior faculty in science and engineering in university settings (e.g., U.K. Athena Swan’s programs) are ethically problematic because they assume men do not need that level of support. From an evidence-based perspective, the most questionable and perhaps the most favored of these interventions are focused on stereotype threat, implicit bias, and microaggression.

Stereotype threat occurs when one is confronted with tasks or situations that trigger negative stereotypes (e.g., that “women are not as proficient at math as men”), that in turn result in a preoccupation about performing in a way that confirms the stereotype.2 Critically, the preoccupation is said to undermine actual performance, even when there is no factual basis to the stereotype. Implicit bias is a related concept and involves an unconscious association between group membership (e.g., sex) and stereotypical positive or negative attributes that can also, in theory, result in prejudicial behavior towards individuals within that group.3 Microaggressions are subtle behaviors (e.g., facial expressions) or statements that are not explicitly hostile but are nevertheless interpreted by the receiver as conveying contempt, stereotypical attitudes, or other negative beliefs.

Proponents of these theories and their activist followers believe that some significant proportion of the sex differences in STEM fields – but curiously only those in which men outnumber women – are thought to be caused by pervasive negative stereotypes about women’s abilities in these fields that in turn undermine their performance. Their argument is that in school and in the workplace, women in these fields are subjected to microaggressions by teachers and colleagues that seep from their unconscious belief in these same stereotypes. The result is the creation of unsupportive and even subtly hostile classrooms and work environments. These types of explanations fit hand-in-glove with the narrative of some feminist scholars; that the sex difference is largely due to oppressive social and cultural factors that undermine women’s pursuit of degrees and occupations in STEM fields.4

These concepts have been embraced by the mass media and beyond, and include accusations made in the New York Times that the wording of several SAT items trigger stereotype threat and undermine girls’ performance on the mathematics section of the test, and the publication of self-help books to purge one’s own unconscious biases.5 On the face of it, there is nothing wrong with an academic and mass media focus on these topics. The real issues concern the magnitude of these effects on women’s STEM participation and the foregone opportunities of not focusing on other factors that might have an even stronger impact on their participation.

Let’s start with the magnitude of stereotype threat on girls’ and women’s mathematics achievement.6 Given the prominence of the topic and the resources devoted to it, we carried out the first meta-analysis (i.e., statistical aggregation of experimental results across many studies) of the effect of stereotype threat on sex differences in mathematics performance.7 We reasoned that if stereotype threat had a substantive effect on girls’ and women’s mathematics performance then the most basic experimental manipulation of the effect should replicate across studies. It replicated in about half of the studies that used the same and most basic experimental design. And of the half that replicated, half of these used a questionable statistical approach. The summary of the other half did not show a stereotype threat effect. Thus, if you accept the questionable statistical approach, you may still argue that a small stereotype threat exists.

In a related analysis, Flore and Wichert found a similar overall effect, but when they corrected for publication bias – the tendency for positive but not negative results to be published – the effect essentially disappeared.8 Because studies that do not find an effect tend not to get published, this means that even when there is evidence for a small stereotype threat effect in some reports, the real-world impact could be close to zero. Currently, a large replication effort is being carried out, and we are optimistic that this will be a significant step towards finally determining whether or not stereotype threat can undermine girl’s and women’s performance in mathematics, and if so to what extent.

It should be noted, though, that the largest study to date included nearly 1000 children (9-14 years old) and found no effects.9 This latter study is of particular interest, because it included adolescents, whereas most other stereotype threat studies were carried out with university students. If stereotype threat discourages girls from pursuing math-intensive STEM coursework and careers, its effect should be evident in high-school. The fact that a large and well-designed study could not find any effect, in our opinion, suggests either the effect does not exist or it is unmeasurably small.

In any case, the existing evidence indicates that stereotype threat has received outsized attention from educational policy makers and opinion makers. Thus, the considerable efforts at addressing this ‘problem’ will almost certainly have little if any effect on girls’ and women’s participation in inorganic STEM fields.

We suspect the same is true for implicit bias. For a variety of cultural and legal reasons, the level of explicit sexism has dropped considerably over the years in most school and work environments. But, as noted, girls’ interest and women’s participation in inorganic STEM fields has remained stubbornly low.4 So, there is now a fork in the road. Down one path is the conclusion that explicit sexism is no longer keeping girls and women away from these fields and so something other than sexism or bias must going on, as we’ll discuss below. The other path maintains the conceptual grasp on sexism but switches focus to an ‘unconscious’ and subtle form of sexism that results from implicit bias and its behavioral companion, microaggression.10

It seems that most academics and the general public have wandered down this second path. Indeed, implicit bias has achieved a cult-like status in some academic circles and in the wider culture. There are now on-line tests to assess one’s implicit bias in a number of areas, including sex differences in work and family. We are not doubting that people have all sorts of implicit beliefs that may or may not be accurate. At issue here is whether or not we can rigorously and accurately assess these biases, and whether or not the strength of any such biases is sufficient to explain the sex differences in STEM fields. The assessment of implicit bias is often done using the implicit associations test, whereby the strength of people’s associations between sex and certain attributes, such as work or science, is assessed by a series of categorization tasks. The difference between the speed of categorizing certain attributes (e.g., scientist, engineer) to one sex or the other is taken as an index of implicit bias. Nosek and colleagues found that people are generally quicker to associate men with science and women with literature, which is taken as an implicit bias against women in science, although they do concede that their results may reflect people’s knowledge of actual occupational sex differences.11

Although it doesn’t typically reach the ears of the general public, there is vigorous debate within the scientific community regarding what exactly is being measured by implicit bias tests, and whether they actually influence behavior.1213 Even if the tests are measuring bias, the influence on actual behavior is small at best, although proponents argue that these small effects add up over time.14 The ways in which implicit attitudes are thought to influence real-world outcomes include promoting stereotype threat and microaggressions.15 As we noted above for stereotype threat, there are serious concerns about the ability to accurately measure microaggressions, whether or not they are related to implicit bias at all, and – if it is indeed a valid concept – whether or not ‘victims’ suffer long-term consequences.16 These issues have not stopped the development of yet another rent-seeking industry to put a stop to this ‘aggression’ on college campuses, in the workplace, and in daily life.

We suspect that concepts like stereotype threat, implicit bias, and microaggression have gained traction because they fit the inequalities-equals-oppression narrative.17 In many cases, explicit oppression is hard to find in classrooms and the workplace and, thus, the resort to explanations from unconscious bias and fleeting behaviors (microaggressions) that continually ‘assault’ and undermine the ‘victim.’ In this case, the victims are girls’ and women’s aspirations towards and performance in certain STEM fields, especially engineering, computer science, and the physical sciences. The logical response to this narrative is the development of interventions to reduce stereotype threat, implicit bias, and microaggressions. But, what if these factors have much smaller effects on girls and women than proponents argue?  The associated time and resources devoted to addressing these problems will have little or no long-term effect on girls’ interest in or women’s participation in inorganic STEM fields.

If not implicit and subtle oppression, then what’s really going on?  

We’ve recently found that countries renowned for gender equality show some of the largest sex differences in interest in and pursuit of STEM degrees, which is not only inconsistent with an oppression narrative, it is positive evidence against it.18 Consider that Finland excels in gender equality, its adolescent girls outperform boys in science, and it ranks near the top in European educational performance.19 With these high levels of educational performance and overall gender equality, Finland is poised to close the sex differences gap in STEM. Yet, Finland has one of the world’s largest sex differences in college degrees in STEM fields. Norway and Sweden, also leading in gender equality rankings, are not far behind. This is only the tip of the iceberg, as this general pattern of increasing sex differences with national increases in gender equality is found throughout the world.20

The recent uptick in interest in concepts such as stereotype threat, implicit bias, and microaggression may be a reaction to the low female STEM participation in highly developed nations. At one time, there were substantive social and educational impediments to women’s participation in these (and other) fields, but as explicit sexism and restricted educational opportunities faded into history, the sex differences (e.g., fewer women than men physicists) attributed to them should have faded as well. Some of them have even reversed, such that more women than men attend and graduate from college and women may now have structural advantages (e.g., hiring practices) in STEM fields.21 Even with these changes, many other sex differences remain or have become larger over time. The latter are serious problems for anyone with strong beliefs about purely or largely social influences on sex differences; if the obvious social causes have been addressed, then there must be other, more subtle oppressive factors afoot. This is where stereotype threat, implicit bias, microaggression and related concepts enter the oppression narrative.

We believe that with economic development and advances in human rights, including gender equality, people are better able to pursue their individual interests and in doing so more basic sex differences are more fully expressed.22 The differences in STEM are related in part to student’s personal and occupational interests and relative academic strengths. Sex differences in occupational interests are large, well-documented, and reflect a more basic sex difference in interest in things versus people.23 Men prefer occupations that involve working with things (e.g., engineering, mechanics) and abstract ideas (e.g., scientific theory) and women prefer working with and directly contributing to the wellbeing of others (e.g., physician, teacher). The sex difference in interest in people extends to a more general interest in living things, which would explain why women who are interested in science are much more likely to pursue a career in biology or veterinary medicine than computer science.24

Programs designed to steer women into inorganic STEM fields would in effect steer these same women away from the life sciences. Such programs would, in our opinion, only be justifiable if women are not provided a fair opportunity to pursue inorganic STEM fields (for which there is no good evidence). The main argument from gender activists is that inorganic STEM fields are a better choice for women either because these jobs lead to higher incomes or that there is a labor market demand for them. Both arguments are fundamentally capitalist and dehumanizing in the sense that considerations of personal interest are overridden by considerations of societal demand. This is ironic, given that the agenda arguing for more women in STEM seems most popular among left-leaning people.

In any event, on top of differences in career preferences, there are important and largely overlooked sex differences in relative strengths in reading, mathematics, and science.25 Students who are relatively better in reading-related areas (e.g., literature) than they are in science or mathematics, independent of their absolute level of performance are more likely to pursue college degrees in the humanities and enter non-science occupations. The reverse is true for students who are relatively better in science and mathematics than literature.26 This is where the results from Finland and elsewhere make sense. Although Finnish girls perform as well or better than Finnish boys in science, the gap is even larger in reading. The result is that more Finnish girls have relative advantages in reading than science. Most adolescent boys in contrast are relatively better at science or mathematics than reading, independent of their absolute level of performance. Individuals with this academic profile are likely to enter STEM areas, either as research scientists or technicians, and there are more boys than girls with this pattern throughout the world.25

We found that the sex difference in academic strengths explains part of the gap between the proportion of adolescent girls who have the absolute level of science and mathematical competencies needed to pursue a STEM degree in college and the proportion that actually obtain such a degree. The other contributing factor is their stated interest in and enjoyment of science. At the same time, there was still a gap between the number of women capable of obtaining a STEM degree and those that actually completed one. There was no sex difference in the math and science abilities needed to pursue a STEM degree. This leaves some room for stereotype threat, implicit bias, or related factors, but their relative contribution (assuming they exist) would be small.

A better route to increasing women’s participation in STEM might be to focus on the substantive numbers of girls with relatively higher science or mathematics than reading achievement; 24 percent of Finnish girls, for instance. These girls have the academic profile that is common in boys that pursue STEM-related careers but fewer of these girls than boys actually pursue them. It seems to us that interventions focused on this group of girls (e.g., individual mentoring) holds much more promise for increasing the number of women in inorganic STEM professions than do currently vogue interventions that focus on purging the wider society of stereotypes, implicit bias, and microaggressions.

Adapted from a chapter for D. Allen and B. Howell, editors, Groupthink in Science: Greed, Pathological Altruism, Ideology, Competition, and Culture. New York: Springer.


David C. Geary studies sex differences and is a Curators’ Distinguished Professor in the Department of Psychological Sciences and Interdisciplinary Neuroscience at the University of Missouri.

Gijsbert Stoet is Professor of Psychology in the School of Social Sciences at Leeds Beckett University (UK).



1 Del Río, C., & Alonso-Villar, O. (2015). The evolution of occupational segregation in the United States, 1940–2010: Gains and losses of gender–race/ethnicity groups. Demography52(3), 967-988.
2 Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women’s math performance. Journal of Experimental Social Psychology, 35, 4–28.
3 Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology74, 1464;  Greenwald, A. G., Poehlman, T. A., Uhlmann, E. L., & Banaji, M. R. (2009). Understanding and using the Implicit Association Test: III. Meta-analysis of predictive validity. Journal of Personality and Social Psychology, 97, 17-41.
4 Hill, C., Corbett, C., & St Rose, A. (2010). Why so few? Women in science, technology, engineering, and mathematics. Washington, DC: American Association of University Women.
5 Thiederman, S. (2015). 3 keys to defeating unconscious bias. San Diego, CA: Cross-Cultural Communications.
6 Walton, G. M., Logel, C., Peach, J. M., Spencer, S. J., & Zanna, M. P. (2015). Two brief interventions to mitigate a “chilly climate” transform women’s experience, relationships, and achievement in engineering. Journal of Educational Psychology107, 468-485.
7 Stoet, G., & Geary, D. C. (2012). Can stereotype threat explain the sex gap in mathematics performance and achievement?  Review of General Psychology, 16, 93-102.
8 Flore, P. C., & Wicherts, J. M. (2015). Does stereotype threat influence performance of girls in stereotyped domains? A meta-analysis. Journal of School Psychology53, 25-44.
9 Ganley, C. M., Mingle, L. A., Ryan, A. M., Ryan, K., Vasilyeva, M., & Perry, M. (2013). An examination of stereotype threat effects on girls’ mathematics performance. Developmental psychology49, 1886-1897.
10 Basford, T. E., Offermann, L. R., & Behrend, T. S. (2014). Do you see what I see? Perceptions of gender microaggressions in the workplace. Psychology of Women Quarterly38, 340-349.
11 Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). Harvesting implicit group attitudes and beliefs from a demonstration web site. Group Dynamics: Theory, Research, and Practice6, 101-115.
12 Greenwald, A. G., Banaji, M. R., & Nosek, B. A. (2015). Statistically small effects of the Implicit Association Test can have societally large effects. Journal of Personality and Social psychology108, 553-561;  Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J., & Tetlock, P. E. (2013). Predicting ethnic and racial discrimination: A meta-analysis of IAT criterion studies. Journal of Personality and Social Psychology105, 171-192;  Rothermund, K., & Wentura, D. (2004). Underlying processes in the implicit association test: Dissociating salience from associations. Journal of Experimental Psychology: General133, 139-165.
13 Blanton, H., Jaccard, J., Klick, J., Mellers, B., Mitchell, G., & Tetlock, P. E. (2009). Strong claims and weak evidence: Reassessing the predictive validity of the IAT. Journal of Applied Psychology94, 567.
14 Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J., & Tetlock, P. E. (2013). Predicting ethnic and racial discrimination: A meta-analysis of IAT criterion studies. Journal of Personality and Social Psychology105, 171-192;  Greenwald, A. G., Banaji, M. R., & Nosek, B. A. (2015). Statistically small effects of the Implicit Association Test can have societally large effects. Journal of Personality and Social psychology108, 553-561.
15 Miller, D. I., Eagly, A. H., & Linn, M. C. (2015). Women’s representation in science predicts national gender-science stereotypes: Evidence from 66 nations. Journal of Educational Psychology107, 631-644;  Sue, D. W. (2010). Microaggressions in everyday life: Race, gender, and sexual orientation. Hoboken, NJ: John Wiley & Sons.
16 Lilienfeld, S. O. (2017). Microaggressions: Strong claims, inadequate evidence Perspectives on Psychological Science12, 138-169.
17 Hicks, S. R. (2004). Explaining postmodernism: Skepticism and socialism from Rousseau to Foucault. Scholarly Publishing, Inc.
18 Stoet, G., & Geary, D. C. (2018). The gender equality paradox in STEM education. Psychological Science. Advanced online http://journals.sagepub.com/doi/full/10.1177/0956797617741719 See also Pinker, S. (2008). The sexual paradox: Men, women and the real gender gap. New York, NY: Simon and Schuster.
19 World Economic Forum (2015). The Global Gender Gap Report 2015. Geneva, Switzerland: World Economic Forum; Programme for International Student Assessment, 2016; https://nces.ed.gov/surveys/pisa/
20 This is not restricted to STEM fields. Many sex differences are larger in gender-equal countries. Lippa, R.A., Collaer, M.L, & Peters, M. (2010). Sex Differences in Mental Rotation and Line Angle Judgments Are Positively Associated with Gender Equality and Economic Development Across 53 Nations. Archives of Sexual Behavior, 39, 990-997.

21 Ceci, S. J., & Williams, W. M. (2015). Women have substantial advantage in STEM faculty hiring, except when competing against more-accomplished men. Frontiers in Psychology6. e1532; Williams, W. M., & Ceci, S. J. (2015). National hiring experiments reveal 2: 1 faculty preference for women on STEM tenure track. Proceedings of the National Academy of Sciences USA112, 5360-5365.
22 Geary, D. C. (2010). Male, female: The evolution of human sex differences (second ed)Washington, DC: American Psychological Association.
23 Su, R., Rounds. J., & Armstrong, P. I. (2009). Men and things, women and people. Psychological Bulletin, 135, 859-884.
24 Lofstedt, J. (2003). Gender and veterinary medicine. The Canadian Veterinary Journal44, 533-535.
25 Stoet, G., & Geary, D. C. (2015). Sex differences in academic achievement are not related to political, economic, or social equality. Intelligence, 48, 137-151.
26 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 Psychology78, 250-261.


Listen to this article
Voiced by Amazon Polly


  1. Pingback: New top story on Hacker News: Sex and STEM: Stubborn Facts and Stubborn Ideologies – Tech + Hckr News

  2. Sarka says

    While I am not a believer in the “blank slate” – I’m quite prepared to believe that some differences in educational preferences have something to do with innate sex differences – I am becoming a little irritated by this constantly wheeled out assertion that countries with the greatest sexual equality show the highest degree of “freely chosen” (i.e. natural not cultural) difference in subject preference, outcomes etc.

    My major problem with it (as a social historian by training) is with base lines of comparison. The assertion cannot possibly be meant to apply in some universal way, because it would follow that the countries with the least degree of gender equality should show the least differentiation between male and female educational choices. Which countries are they? Saudi Arabia? Iran? (Or if we wanted to look in a comparative historical way – the UK in the 1900s, or 1950s?) Countries (or periods) with very high or even impossible barriers to female participation in some subjects and careers can hardly provide the sort of baseline the assertions about modern Scandinavian countries seem to imply. To put it at its starkest (most absurd) I do not believe that a higher proportion of women chose scientific careers in Britain in the 19th century than in the early 21st century!

    So I presume that the comparisons involved are quite narrow – i.e. between advanced western countries with relatively small differences between them within a relatively short recent time-frame. Here I certainly think the findings are interesting, but it is very dodgy indeed to assume that we are dealing with contrasting situations reflecting different degrees of proximity to some pure “natural” choice. But at all events, proponents of the theory should invest a little time in trying to work out what – in a situation of more gender inequality – would be pushing girls to take up e.g. STEM subjects…It’s all rather murky.

    • I think the key is an egalitarian society, but in combination with widespread prosperity and safety, in which Scandinavian countries lead the pack.

      For example, in China a woman studying STEM has an opportunity to go to a major domestic city, Europe or the US to earn many times more than she would at home, to provide not only wealth but safety/security for her and her family. But the average engineer in Oslo doesn’t earn that much more than a qualified public sector worker with a few years experience and families are already protected by social safety nets that don’t exist in poorer countries. I think this equation leads to interest trumping the desire for slightly more money.

      In less egalitarian societies there are of course additional variables, but where there is rough equality of opportunity it seems that poverty and lack of social safety nets are key drivers of women choosing jobs that pay more over jobs that they would otherwise choose under better circumstances.

    • David C Geary says

      We’re more than happy to send the original article to you and you can judge for yourself. Dave Geary

      • Great article! One of my favorites published here, which is saying a lot because I enjoy many articles here.

      • Manuel Cruz says

        This article, like all other articles about gender equality, suffer from severe ideological blindness.
        Replace “gender equality” with “discrimination against men”, and everything becomes evident. That’s it, it’s not that there is less women in STEM, is that men are disappearing from all non-STEM fields. The reason is that they are being actively discriminated, in two ways.

        1. Public Universities nepotism has turned them into slurry pits of Cultural Marxists who openly discriminate against men and conservatives, because instead of scientists they are numerologist morons that only care about the “Feng Shui of gender ratios”. These people openly brag about grading the men poorly, and created abominations such as Common Core because they found in tests that it impaired the performance of boys more than it impaired that of the girls (yes! they sabotage both sexes for the sake of equality!). As a defense, men are leaving those fields that have subjective grading to those in which you can prove that an answer is objectively right.
        2. Feminism, which is the excuse the elite made up to divert public money to their spouses and concubines, has created an environment of useless women with plenty of free time and starved to find ANYTHING that justifies their existance. As a result, any kind of interaction with women is twisted to turn it into rape or discrimination, resulting in ruined lives for men. As a defense, men are leaving in droves any field that has any kind of social interaction.

        In fact, we can see in this study

        that in countries where women are openly discriminated, they go into STEM at higher rates than men. We are seeing the same effect as men in feminist hellholes like Sweden.

    • You make a good point. All we can really conclude from the comparison of highly gender-egalitarian countries and countries with more misogynist cultures is that gender gaps in STEM are not amenable to simplistic cultural explanations. Freely-made decisions may or may not be at work here, but at least we can say that the simplistic narratives about societal-level sexism are inadequate for explaining what’s going on.

      It may be that women in Scandinavia and similar countries choose against STEM because they have more freedom to do so. Or it may be that men in STEM are similarly sexist in the Middle East and Scandinavia, but in Scandinavia they are much more sexist than other professions, while in the Middle East they are unremarkable. Or it may be that women in more traditionalist cultures are more affected by parental pressures to choose fields that seem to be more practical. Or maybe any number of other explanations apply. I find some of them to be more plausible than others, but the international comparisons proffered here do not suffice to choose among them.

    • The comparisons in the study included many non-Western countries. The gender gap in STEM degree attainment is smallest in countries like Algeria, Tunisia and Albania, largest in Scandinavia, Chile, Belgium.

    • If you dont have the freedom to make a truly free decision, you are statistically more likely to choose a field of study that you dont enjoy but fits your skillsets. As this article states, women who are strong in math and science tend to also be stronger than men in verbal/reading skills. If free to choose based on their innate preferences then they would likely choose humanities over science, however the more innate preference is disregarded, the more likely they are to pick a career that fits their skillsets rather than their desires. If innate preference is removed from the equation, you would expect a more statistically even distribution of women in certain fields since we know competence is about the same between genders.

      So which do you believe to be the case? That gender roles in free societies force women out of the sciences, while harsher gender roles in less free societies do not? Or that innate preference is more easily expressed in free societies, while in less free societies innate preference is less likely to influence career choice?

      Also keep in mind in your example of 19th century Britain, women likely werent allowed in the workforce altogether. Women being heavily involved in the workplace is only a recent part of human history. We are still trying to work out the details.

  3. Pingback: New top story on Hacker News: Sex and STEM: Stubborn Facts and Stubborn Ideologies – ÇlusterAssets Inc.,

  4. stephen buhner says

    I find your reasoning elegant and clear, thanks for the work. I do have a couple of observations. The first is that when i trained as a mathematician I found that my professors had little interest in the gender of their students but a great deal of interest in intellectual companionship, that is, finding anyone who was interested in being able to converse in the language they loved. Unlike most people i know, i trained and worked extensively in the trades, as a remodeling carpenter and furniture maker of one of a kind pieces with high aesthetic value. When training in that field (I had to pay for college some how) I experienced what would now be termed microagressions though using the prefix micro- would be lessening its nature. I experienced that even more when i decided, at age 57, to become a musician, a player really (players are street musicians rather than schooled musicians). Rather than microagressions the proper term is hazing. I found the carpenters to be far more kind in it than the musicians . . . or the venues in which I began to play. In looking back on it all I keep hearing something an old blues player once told me, “You got to pay your dues.” It seems to me that “paying dues” is simply a phrase that captures a common human experience that anyone who wishes to become part of a particular tribe experiences from the people already in that field. My experience in working in more academic technical fields is pretty similar. Most of the discussions about this topic never seem to address this common dynamic among human beings. And one last comment: “lower-level technical jobs, such as car mechanics or plumbing” is actually class-oriented and while not necessarily intended is somewhat disparaging of the trades. I don’t actually find them to be lower-level in any meaningful sense. But it does remind me of that old saying: “Cultures that value philosophy more than plumbing will end up with both bad philosophy and bad plumbing.” I found William Deresiewicz’s article “The Disadvantages of an Elite Education” to be rather succinct on this point.

  5. Charles Earl says

    May freedom, equality, and access always be in vogue.

  6. This is a very well argued and sourced piece. Kudos to Quillette for obtaining this excerpt for publication. And it’s impressive that a single courageous, heterodox piece (by James Damore) should have been sufficient to inject this important topic into the popular discourse to such a great extent.

    That said, one must continue to hold out hope that society will find ways to encourage more smart women at the university level and beyond to shift from fields that are peripheral to general progress or oversubscribed (e.g. women’s studies, law) to the fields that are predominantly driving progress and are the most costly for society to staff because of tight talent markets.

    I assume that much of that is to ensure labour market price signals ring as loudly and clearly for women as they do for men by reducing various public, private and even household distorting subsidies that bring socially perverse outcomes. And, yes, to continue at least for a few more decades to roll out the welcome mat so that traditionally male work environments are more inviting in order to see whether such rudimentary efforts can move the needle.

    We should concede that there is enough cause for skepticism that we should not measure progress against possibly illusory goal of parity in these fields or the related goal of achieving gender equality of earnings. However, from the standpoint of larger social benefit and general welfare, the key metric — easy to conceptualize, but perhaps challenging to operationalize — remains how much talent is underutilized or wasted because of counterproductive public policies and outdated social norms and practices. The question remains how best to identify and tackle those in detail.

  7. steve says

    Firstly I would check the stats – the article says:
    ” For a variety of cultural and legal reasons, the level of explicit sexism has dropped considerably over the years in most school and work environments. But, as noted, girls’ interest and women’s participation in inorganic STEM fields has remained stubbornly low.4 ”
    However Misa in 2011 wrote:
    “The most recent NSF figures suggest that women may account for just one in seven undergraduate computing students, or around 15%: a catastrophic drop from the peak of 37%.” (the peak was in the mid eighties)
    In CS (not sure about the rest of STEM) there has genuinely been a reduction in female participation – not a “stubbornly low” constant.

    The explanation given by Margolis among others – is that it was in the late 80s when the computer first became an affordable consumer device, with particularly marketing aimed at boys – particularly regarding computer games. Consequently, male 1st year degree students would already be far ahead of female 1st year degree students because of greater familiarity with the computer. As a result the cs cohort became a much less cohesive group and may have presented a less welcoming environment for female undergrads.

    • Taupe Pope says

      Why would the CS cohort have become less cohesive when it came to consist of more experienced, more predominantly male students?

      Isn’t it an indictment of women that the mere presence of men would prevent them from pursuing a lucrative field of study in which they were interested? Or do you mean that they could not out-compete their more advanced, male counterparts?

  8. Maybe there is nothing to the idea, maybe there is. I have noticed one thing – how come most or all of these “critical of” articles are written by male researchers in the “soft sciences”, which, according to the info here is 57% female graduates?

    Can we get some women researchers to analyze some data and maybe provide some of their own anecdotes in an article or two? Or would I be guilty of “forcing equal outcomes” by asking this?

    Perhaps an AI could provide a more neutral narrative?

    • This article apears to be written by two angry, women hating dudes who have never experienced the shit women have a single day of their red pill lives. Its reductionist rubbish.

  9. Beyond sex oriented talent, which I personal doubt, and potentially social constructed gaps, I believe that there are some aspects that are hardly addressed when people talk about the gap in academia. I am a physicist and i can say that to success in academia and research one has to work hard and a lot,and furthermore, leave behind most of what people would call a personal life. This is not new. If one looks into the past one can notice the little number of famous scientists who had normal lives, if any. Indeed, many, like Newton, did not have families at all. Or think of Einstein who left his family behind to construct GR in Berlin and later married a woman who relieved him of most of the common man worries.

    In my opinion, as I have witnessed it, the gap is a not gender gap, but actually a parenthood gap. When a selection committee has to choose among a heap of young early 30’s candidates, it would select the one who has the larger number of publication and citations. Those who have children always underachieve in comparison with those who have not, independently of their sex.

    Now, concerning the sex gap, it seems to me, but this is merely an on-field observation, that women in their early 30’s that are less willing to abandon to construct a family in post of academic career than men. In more than an occasion we had discussed with my colleagues about an incredible talented woman, holding a postdoctoral position in a top 10 university, who left academia because she decided to start a “normal life”. Only once I have see a man takes that decision.

    • David C Geary says

      This is not uncommon and has been documented by David Lubinski and Camilla Benbow in their longitudinal study (age 13 to 50s) of mathematically gifted people.

      • Could it be that what is happening IS a good thing, maybe women (and many smart women) choosing parenting above a prestigious academic job are making the collectively wise choice, even though from the narrow perspective of “increasing women in STEM” it seems disappointing. What if, from a wider perspective, what is happening is an adaptive, wiser response and not a regression? Given the hours necessitated by current academic practice, who would be there for the children in the developmentally appropriate amount? The cost-benefit analysis of rigorous academic achievement versus well-adapted, nurtured progeny (the future generation) in a bigger scheme of social life could well be more meaningful and appropriate.

    • So why is this parenthood gap less pronounced on the “organic” side of STEM?

    • Absolutely spot on observation. Mums are not seen as technical, nor do they have the capacity to be as prolific.

  10. gonzalo says

    Excellent article! Only one comment. You wrote at the very end “It seems to us that interventions focused on this group of girls (e.g., individual mentoring) holds much more promise for increasing the number of women in inorganic STEM professions…” What about no interventions at all? Because “this group of girls” (a grouping/categorization already not accurate enough) may not be interested in inorganic STEM and many other girls outside “this group may”, so it will be unfair for the latter not having access to mentoring programs. What about promoting all STEM equally among girls and boys, and let them decide if they want any STEM at all.

    • Because if their choice doesn’t result in this Utopian equality of distribution then the reason is…sexism?

      • gonzalo says

        You are right! But to repeat myself here, the last paragraph of this article is not helping at all. It was so good until the authors envisioned a “solution” to the “problem”. As much as I can see, there is no solution to this “problem”, but further promote inorganic STEM among kids in general, because “interventions” are always in detriment of another “group” or all the rest, due to the limited amount of resources available to promote anything. In a perfect world, you would have access to all areas of knowledge, but you will always be restricted by time. So even if there is unlimited resources for mentoring, kids will pick what is interesting for them due to time.

  11. Gabriele Mari says

    Great read, thanks. I find though that the argument “countries renowned for gender equality show some of the largest sex differences” in interests, personality traits and so on has a false premise. Yes, Scandinavian countries are renowned for gender equality and yes, if you compare male and female employment rates, you’ll find one example supporting this. At the same time, Finalnd has one of the largest gender wage gap in OECD countries (https://stats.oecd.org/); studies for Denmark (https://www.henrikkleven.com/uploads/3/7/3/1/37310663/kleven-landais-sogaard_gender_feb2017.pdf) and Sweden (http://www.journals.uchicago.edu/doi/abs/10.1086/684851) show that women suffer sizeable wage losses due to childbirth. I think we would all benefit from abandoning this idea that Scandinavian countries are hallmarks of gender equality; in many respects (particularly women’s career attainment), they are not.

    We also shall not overlook the fact that STEM jobs have features that make them, today at least, hard to combine with family life. Taking time out (if on family leave), reducing working hours, missing training opportunities, all this can harm women’s wages. Why discounting that women may anticipate this and choose different lines of work (http://www.ucl.ac.uk/~uctpb21/Cpapers/fertility_resubmitJPE.pdf)?

    I think that in the US the debate has been hijacked by the theories the authors convincingly review and criticise. Here in Europe we are concerned more on designing better policies for working mothers, for instance. And probably the US should consider this too.

    • David C. Geary says

      By gender equality we meant per typical measures such as access to school, legal rights, women’s representation in parliament and so on. These are social rules and can’t change basic biology. You’re correct about the career costs of investing in children. Lubinski and Benbow’s study shows this for women on the STEM track but also no drop in life satisfaction. It’s one of the tradeoffs in life that could be reduced but it’s hard to see how to completely eliminate it.

    • Taupe Pope says

      The “maternity penalty” argument is special pleading. To wit, a man would suffer the same ‘penalties’ if he behaved like the typical female worker. The solutions to this ‘inequality’ are as discriminatory as girls-only STEM tutoring/mentorship programmes.

  12. David C Geary says

    I agree that promoting math and science more generally and for everyone is the way to go. The problem is there is a very large cottage industry promoting STEM for girls and women. If they are going to intervene anyway, it seems to make sense to do so in a way that might be more productive than what has been already tried.

  13. Interesting that Finnish girls outperform boys in science. Does that include life science? How much do you think the earlier development and maturing of girls might affect that? Are there cultural explanations – white working class boys in the UK for instance generally underperform for cultural reasons. Do educational methods affect differential outcomes for boys and girls?

    • Overall, girls generally do better in school. Girls are more conscientious when it comes to homework, paying attention in class, and listening to teachers, all of which contribute to better grades and sometimes better performance on achievement tests. The PISA test was a general science literacy test, so assessing your knowledge of what you would have been exposed to in school. It wouldn’t get at more complex areas (e.g., advanced physics) where boys usually do better.

  14. 1. Mowgli doesn’t exist. Leave a baby Einstein with monkeys and he will become a monkey. The CULTURE makes humans who they are.
    2. The discussion “can females do this or that” is wrong in its core. It shows already biased culture in the favor of “can’t”.
    3. The initial statement has to be “everyone can do everything”, no gender, or race, or ethnic bias. Limitations may have physiological nature, but they are not so broad and strong to exclude a certain type of people from any professional area.
    4. The people who impose culture on children are parents and teachers. When teachers have a gender bias (or any sort of bias) they organize their practice according to this bias. This happens on a subconscious (!) level.
    5. When children do not feel a support from teachers who do not believe students could succeed, students do not believe they could succeed and avoid any activities which could make them feel uncomfortable.
    5. Any assessments of what male or female students do better (reading v. math) essentially demonstrate the teacher’s bias.
    6. The problem of professional selection students make is the problem of psychological bias teachers (and parents) have.
    7. When teachers will be teaching everyone equally, there will be no more problem. But this moves us in a different field – teacher professional development.

    • Debbie says

      It would be easy to test your assertions by designing an on-line learning experiment. No male teachers. No female teachers. No implicit bias because the curricula could be designed by bias-free people like you. Tee it up — let’s see how it plays.

    • Dave J says

      @Valentin Voroshilov

      Assertion after unsupported assertion.

      There seems to be little or no point to your comment.

  15. From my perspective as someone working in tech, it seems like overt sexism is not rare – not just minor microaggressions, but obviously bad behavior going unpunished. While I appreciate this piece in pointing out the possibility of personal interest affecting gender ratios, I think we need to put much more emphasis in rooting out the jackasses first. There are male software engineers who genuinely assume their female colleagues are less capable because they are female and there are enough of these people to make women engineers’ life miserable. This is why unconscious bias and microaggression training exists.

    • David C Geary says

      I believe that this happens; my son works in the area and says the same. That said, I have two points.

      First, the academic profiles that predict entry into STEM fields are evident years before students enter college, and long before they have any experiences in a tech work environment. Wang et al. (cited above) found that academic strengths, attitudes, and interest in people vs. things in high school predicted occupation at age 33. Once these were controlled, there were no sex differences in occupational choices.

      Second, I agree that there is no room for sexist behaviors in the workplace. But, the current approaches to reducing them are not based on solid evidence. Intervening in ways that are not effective ensures that the issues will continue.

  16. So if 57% of STEM workers are female, all these girls into STEM programs are taking away from equality?

  17. I think much commentary here is myoptic, I.e. focused on recent and short termed studies, social change, and trendy developments, while ignoring longer term history (such as the last 2500 years.) Also ignoring the enormous gender differentials in hard core technical work, both historically and more recently. What does that long term history suggest? Doesn’t the fact that the whole corpus of science and mathematical discovery over the last 2500 years is for all practical purposes a male creation, deserve some weight in such a discussion? Might it suggest males and females have intrinsically different propensities and motivations ?

    Has there ever been a female Archimedes, Copernicus, Kepler, Galileo, Newton, Maxwell, Euler, Gauss, Poincare, Einstein, Heisenberg, Schrodinger, Dirac, Von Neumann, and hundreds of others ? In the field of invention has there been a female DaVinci, Harrison (chronometers), James Watt, Eli Whitney, Morse, Bell, Marconi, Edison, Wright Brothers, to name a few out of hundreds. Do girls and women, on their own, hang out in garages and rebuild cars and engines? Do they build home built airplanes? Do they build computers and ham radios from scratch? Even among present day female hi tech CEOs, how many were actually founders of the underlying technology, rather than merely appointed by boards of directors of already mature corporations? Do girls and women do these things in whatever numbers they do, because of intrinsic interests, or from artificial push for social change ?

    It is obvious that the propensities of men and women are profoundly different. It is misleading to cite that there are more women in STEM today, and then assume there will be corresponding achievements proportional to those greater numbers. Mere numbers nominally entering a field say little about that. Or that the number entering STEM truly reflects intrinsic interest as opposed to influence by social change push, which has become very systematic and pervasive.

    • David C Geary says

      Right, this is a related but somewhat different question. Our focus was on who could, based on ability, obtain an undergraduate degree in a STEM field. This broader question concerns the extremes of accomplishment. For the latter, I highly recommend Charles Murray’s Human Accomplishment.

    • Actually, I would give the historical experience very little weight. In overwhelmingly subsistence societies, the demands of raising the next generation dominated sex roles. When one looks at the patterns of women writers before the mid C19th, they were either childless (Sappho, Jane Austen), had adult children (Anna Commena), had others in their family to do the childcare (Christine de Pizan) or their children were raised by others. Before the Pill gave women easy control over their own fertility, the capacity of women to do extended education, though expanding, was extremely limited, which particularly limited their STEM possibilities.

      This study is informative on the last point.

      • Absolutely, Degler’s “In Search of Human Nature” provides a good historical narrative of some of the arguments for why women should not attend college; among other issues.

        In any case, there was a surge in women’s participation in inorganic STEM in the 1980s (US), but as one of the other commenters noted participation then dropped and has been pretty stable since then. Of course, college access is not an issue now, as more women attend than men.

      • Contrarywise, women in the upper classes had much leisure and were free to do most any intellectual activity. In fact they were routinely schooled in music, as a social grace. Yet there are no Bach’s, Beethoven’s, Chopins, etc. Mathematics requires nothing more than a few books, paper and pencil, yet there are no first rank female mathematicians. The women you mention-Jane Austen, etc., although free to do these things, chose to write about human relations, romance and family. That choice is a palpable and pervasive preference among women, ultimately rooted in biology. Rare exceptions do not alter the generality that women seek to occupy themselves with interpersonal relations, personalities, family, nurture and child rearing, as opposed to the impersonal and abstract sciences. Thus women seek to be nurses and elementary school teachers, and men seek to become mechanics and engineers. I said rooted in biology because modern socio-biology makes this entirely plausible in terms of evolutionary selection pressures as well as being what we actually see.

  18. Debbie says

    I think it would be interesting to see the results of an implicit bias test based on how quickly test-takers could differentiate between the categories animal/mineral and snowboard/mountainbike. I suspect the primary differences are the result of changing the test-taker’s mental programming when the categories are re-ordered.

  19. Alexandre says

    As a historian with a PhD in Sociology I feel comfortable to say: the Humanities and some of the Social Sciences have completely ruined their possibilities of understanding social behaviour, social action, agency, you name it, in a proper way due to their commitment to ideology and politics. When I hear someone from History, Sociology, etc. talking about bias, I just laugh.
    I’m glad that I’ve found those studies on sex differences. I don’t read contemporary sociology anymore.

    • It is unfortunate that these internal power struggles damaged many of these disciplines and seeped into the wider culture. Including yourself, there are many reasonable scholars within these fields, they are just out numbered, at least for the time being.

      The postmodern, critical theory agenda has, in my opinion, set back research on sex differences by 4 or 5 decades, including studies of sex-specific disease risks and treatments.

      The Jan/Feb 2017 issue of Journal of Neuroscience Research was exclusively devoted to this problem. I’ve published a few recent articles on how an evolutionary perspective can facilitative our identification and understanding of sex-specific vulnerabilities, e.g.:

      Geary, D. C. (2017). Evolution of human sex-specific cognitive vulnerabilities. The Quarterly Review of Biology, 92, 361-410.

      Geary, D. C. (2016). Evolution of sex differences in trait- and age-specific vulnerabilities.
      Perspectives on Psychological Science, 11, 855-876.

      I’ll send copies to whomever might be interested.

  20. Chris Corrigan says

    Curiously not mentioned by the author is the fact that reason that men, on average, score better in tests of mathematical ability is that they dominate the higher end of the distribution. Something like 90% of those above the second standard deviation of performance on math tests are men. This result is independent of the test or the country where the tests are given. This is almost certainly the reason why PhD’s in math and physics are mostly men.

    • There are important differences at the extremes of the distribution of mathematical competence. Jonathan Wai’s work documents these; https://jonathanwai.wordpress.com/research/

      The data set we analyzed and the question addressed most recently – why cross-national variation in the numbers of women getting degrees in inorganic STEM fields – weren’t well suited to addressing this issue.

      There is an older monograph on the broader issue that is still relevant and I believe freely available:

      Halpern, D., Benbow, C., Geary, D. C., Gur, R., Hyde, J., & Gernsbacher, M. A. (2007). The science
      of sex differences in science and mathematics. Psychological Science in the Public Interest, 8 (No 1, 1-52).

  21. Pingback: BrokenPla.net - Reading Between the Headlines

  22. Lorne Carmichael says

    A brief note on the intellectual history of one of the important ideas in this debate.

    Over the years, much has been made of the fact that women do as well on mathematical tests as do men, as if it were a direct implication that in an unbiased world there would be as many women in mathematical fields as there are men. See, for example, the well known debate between Steven Pinker and Catherine Spelke (https://www.edge.org/event/the-science-of-gender-and-science-pinker-vs-spelke-a-debate).

    More recently people have recognised that personal choice depends on a comparison of alternatives. My decision to enter a STEM field does not depend on my abilities in math relative to others (although this will affect, say, whether I become an electrical engineer or an electrician). My decision to enter a STEM field depends on how good I am at math relative to how good I am at other things. It’s a personal choice. What is the best option for me?

    This idea first appeared about 200 years ago in the works of David Ricardo in his book “The Principles of Political Economy and Taxation (1817)”. He was concerned with the reasons why countries (rather than people) specialize in one commodity or another but the principle is the same. If you ask people why Canada exports softwood lumber to the United States, most people might say “Because trees grow really well in Canada.”. But softwood trees grow terribly in Canada – they grow much faster in North Carolina. The reason Canada specializes in growing trees is that there is precious little else that can be done with the land.

    In Economics the concept is called “comparative advantage” and (at least for Economics) the implications run quite deep. It’s not just that it helps to explain occupational choice. It also tells you that there is a cost to encouraging more women into STEM fields. If a woman with her technical and social skills has the potential to be an effective doctor and we encourage her to enter Computer Science, our society may be losing more with the loss of a physician that we are gaining with the addition of another software engineer. As well, if we are nonetheless committed to a more even distribution of genders across occupations, we might accomplish this as easily by encouraging boys to read. This might be better than teaching girls to code because it will also help to address the general under-representation of boys in higher education.

    I address these issues in a bit more detail here:


    So far as I know, the first paper to look at the impact of relative grades on the choice of university programme is here:


  23. Becky says

    Does the data on STEM graduates by country refer to the degrees awarded in that country or to graduates of that nationality? Because when I studied engineering in the UK 10 years ago there were only 2 other British women on the course, but we had a lot of students from China and Malaysia who had maybe a 50/50 gender ratio that probably brought us up to 30% overall. Without them the imbalance would have been enormous.

    Also a big thank you to David Geary for engaging in the comments here.

    • David C Geary says

      Great question. The data are degrees awarded in the country, and so you’re correct that this could skew the results such that the actual sex ratio in these degrees is larger, once students from developing countries are excluded. We didn’t have that information.

Comments are closed.