Former Google engineer James Damore was hardly the first person to argue that biological differences between men and women determine career outcomes. Many people — even smart, science-minded ones — have asserted that biological differences can explain the gender gap in math, engineering, and science. A 2005 Gallup poll found that 21% of Americans believed men were better than women in terms of their math and science abilities (though 68% believed men and women were about the same). The fact that this argument keeps coming up means that we need to engage with it and clarify which claims are supported by evidence and which are not.
To address these claims, we need to examine three interrelated questions: Are there gender differences in outcomes achieved by men and women? If so, is there evidence that they are due to biological differences? Is there stronger evidence that they are due to bias?
To answer the first question: Yes, there are gender differences in the participation of men and women in some STEM fields among college students, and these differences do contribute to the underrepresentation of women in STEM professions. Women are also significantly underrepresented in top leadership positions.
But are these outcome differences due to biological differences? While there are (of course) biological differences between the sexes, social science has shown that men and women are more similar than different on a wide range of characteristics, from personality to ability to attitude — and that these factors have a larger effect on career outcomes than biology does.
My former colleague Janet Hyde, a developmental psychologist and an authority on gender differences, reviewed 46 meta-analyses that had been conducted on psychological gender differences from 1984 to 2004. (A meta-analysis examines the results from a large number of individual studies and averages their effects to get the closest approximation of the true effect size.) Hyde’s review spanned studies looking at differences between men and women in cognitive abilities, communication, personality traits, measures of well-being, motor skills, and moral reasoning.
She found that 78% of the studies in her sample revealed little to no difference in these measures between men and women; this supports her gender similarities hypothesis, which states that men and women are far more similar than they are different. The only large differences she found related to girls being better than boys in spelling and language, and testing higher than boys on the personality variable of agreeableness/tendermindedness; boys tested higher than girls on motor performance, certain measures of sexuality (masturbation, casual attitudes about sex), and aggression. So there are some gender differences, but most are small to nonexistent.
But can these differences truly be classified as biological? Or are they due to differences in socialization? It’s the old nature/nurture debate — a debate that can be a false one because most human behavior involves complex interactions between genetic, environmental, and epigenetic influences. For example, one study that Damore cited did find gender differences in personality across cultures, but the researchers described the differences as relatively small to moderate and concluded that “human development—long and healthy life, access to education, and economic wealth—is a primary correlate of the gap between men and women in their personality traits.”
And a review of studies on levels of prenatal exposure to testosterone found resultant differences in empathy, aggression, and toy preference between males and females, but found no significant differences in dominance/assertiveness or ability. Unless all of the differences in men’s representation in STEM and leadership are the result of their lack of empathy, high levels of aggression, or toy preferences, there is little evidence that biological differences affect work-related outcomes. In fact, based on the research on leadership, we would expect to see that a lack of empathy and high levels of aggression would hurt a person’s chances of becoming a successful leader, not help them.
On the other hand, there is a great deal of evidence to support the impact that environment has on gender differences in society. For example, a review of research on gender differences in math test scores shows that the already small effects have declined over time and tend to be greater in countries with less gender equality. In terms of behavior, a study by economists showed that in cultures where women are dominant, they tend to be more competitive than men. Meta-analytic evidence on gender differences in leadership aspirations showed that differences are decreasing over time — women are closing the gap in terms of wanting to be leaders — suggesting that the gap is more due to society than to biology.
Other data also contradicts the idea that women are biologically predisposed to lower levels of leadership. One meta-analysis of 95 studies found that female leaders tend to be rated by others as significantly more effective than male leaders, and this effect is stronger after 1996. (On the flip side, men rated themselves as significantly better leaders than women, particularly before 1982.) But this data does tell us something about the impact of gender roles (as women tend to rate themselves as less effective leaders) and societal changes (since the effects are diminishing over time).
If the evidence on biological differences is too thin to explain the large gender gaps in leadership roles and STEM careers, is the evidence on gender bias any stronger?
Several studies have shown that employers do discriminate against women and minorities. One robust vein of research uses résumés to test how people respond to different candidates with identical qualifications. For example, in one study, professors rated the identical applications of fictional male or female students. When a male name was used, faculty members rated them as significantly more competent and hirable than the female applicant, and they offered the male applicant a higher starting salary and more career mentoring. The reason for this was that women were perceived as less competent by the faculty members; faculty who had greater bias against women rated female students worse. The effect sizes here were moderate to large, unlike those shown in sex-differences studies. And numerous other studies have had similar results, not just in hiring but in promotion rates, performance evaluations, getting credit for good work, and project assignments.
This body of research also shows why advocating for a “pure meritocracy” — rather than explicitly pursuing diversity — doesn’t help companies overcome bias. In fact, companies that highlight “meritocracy” may actually cause greater bias against women: Experimental studies show that when an organization is referred to as a meritocracy, individuals in managerial positions favor male employees over equally qualified female employees and give them larger rewards. The author theorizes that calling the organization a meritocracy may create moral credentialing (when one’s track record as egalitarian makes them feel justified in making nonequalitarian decisions) or greater self-perceived objectivity, giving them license to discriminate against women.
Calling for a meritocracy and denying that workplace inequality still exists captures what scientists refer to as modern sexism. Modern sexism is characterized by “beliefs that discrimination against women is a thing of the past, antagonism towards women who are making political and economic demands, and resentment about special favors for women. Notably, individuals espousing such views do not regard these notions as sexist or unfair and…conclude that, given the even playing field upon which the two sexes now compete, the continuing under-representation of women in certain roles (e.g., management positions…) must be a result of women’s own choices or inferiority as opposed to discrimination.”
In his memo, Damore wrote, “We need to stop assuming that gender gaps imply sexism,” and that we should assume “people have good intentions.” But the gender gap in the workforce can be explained by sexism, just as the race gap can be explained by racism. When workplace practices aim to support underrepresented groups, that does not mean they are unfairly biased against overrepresented groups. It just means that we need more than good intentions to change biased behavior.
We all want systems that are fair. But we need to consider how to make them fair for everyone.