The Uncomfortable Science Of Sex Differences - Steve Stewart-Williams

Steve argues that because of the long, sorted history of sexist science (e.g., 19th-century claims that intelligent women are as rare as two-headed gorillas), it's understandable that people are nervous about evolved sex differences. However, he believes the science disproves old sexist ideas and that we can treat people well even if men and women differ on average. He advocates a careful approach: emphasize that most psychological sex differences are small (Cohen’s d often 0.2–0.5), that there is always massive overlap, and that no 'is' implies an 'ought.' This protocol is his way to bridge the gap between legitimate concerns and scientific honesty. He does not want to minimize the differences, but to contextualize them to avoid the twin dangers of moralizing and denial. He also adds his personal philosophy: 'let people be themselves,' meaning we should neither push people into nor out of traditional roles.

Steve warns that the soft bigotry of male expectations—treating male-typical traits as the standard—leads to viewing any underrepresentation of women in certain fields as evidence of discrimination. He argues we must consider that the people vs. things interest difference (d ≈ 1.0) is large, consistent across cultures, and resistant to socialization efforts. If we ignore this, we will implement costly interventions that fail, then progressively coerce outcomes, potentially discriminating against men and creating backlash. He advocates telling people all the facts and then respecting their choices, removing bias but not engineering equal outcomes.

Steve repeats this phrase as his core philosophy. He contrasts it with two extremes: one that moralizes differences to justify traditional gender roles, and one that denies differences and tries to enforce a unisex ideal. Both, he says, are straightjackets. By letting people be themselves, we honor variation and avoid the harms of both exaggeration and denial. He notes that because differences are moderate and there is huge overlap, there will always be individuals who don’t fit the average, and that’s fine. This principle also underlies his advice on communicating about differences—it’s about providing information and then stepping back.

Steve points out that this difference is gigantic—essentially all men are attracted to women and all women to men, with very few exceptions. He notes that even in academic papers on evolved sex differences, this one is often neglected. The reason might be its obviousness: people don’t think to list it because it seems too obvious. Yet the effect size is enormous, and the evolutionary explanation (it takes one male and one female to produce offspring) is straightforward. He argues that if you ask people to guess the biggest psychological sex difference, almost no one names this, and when revealed, they realize its magnitude. He contrasts this with other sex differences, which are often modest, to highlight how this one is off-scale. The oversight shows how even researchers can miss what’s right in front of them because they’re looking for more controversial or politically charged differences.

Steve explains that the gender equality paradox is the deeply counterintuitive finding that in nations with greater gender equality and less strict gender roles, many sex differences (e.g., in personality, career interests, even some cognitive abilities and physical traits like height) are actually larger. He originally assumed, even from an evolutionary perspective, that patriarchal environments would pry the sexes further apart. The discovery that the reverse happens was a genuine surprise to him. He notes a rebuttal from social-role advocates: in patriarchal societies, men and women are more segregated, so they compare themselves only to their own sex on self-report measures, making differences appear smaller. However, this doesn’t explain why the pattern also appears for non-self-report traits like spatial ability and height. Steve remains open-minded but emphasizes that the paradox undermines simplistic socialization accounts and suggests that when people are freer to choose their paths, inherent differences in interests and preferences can manifest more strongly.

Steve notes that among mammals, only about 5–10% of species form pair bonds with biparental care, while around 90% of bird species do. Humans fall squarely in the bird-like camp: we tend to form relatively durable pair bonds, fall in love, and both parents (not just females) invest heavily in offspring. This has major implications for sex differences. Because males invest, the ceiling on male reproductive output is lowered, which dampens the reproductive variance that drives many sex differences. That is why human sex differences in aggression, sociosexuality, and parenting are present but more moderate than in, say, elephant seals. He finds it amusing that in our primary reproductive arrangement, we are more like the average bird than the average mammal, and jokingly concludes that this makes us more like dinosaurs. This insight helps explain why many stereotypes about male sexual behavior, while having a kernel of truth, are not as extreme in humans as in many other species.

Steve acknowledges hearing the 40/80 figure but believes the reproductive skew in humans is quite a bit lower than in most mammals. His reasoning: males in our species typically invest a lot in offspring through provisioning and protection, not just the sex act. High biparental care reduces the maximum offspring number for males relative to females, narrowing the reproductive variance. Therefore, while some reproductive skew existed, it was not as extreme as often portrayed. He declines to give a precise alternative number but is confident the classic figure is inflated. This matters because if the skew was smaller, then the selection pressure for hyper-competitive male traits would be correspondingly weaker, aligning with the observed moderate sex differences in humans compared to highly dimorphic species.

Steve uses the example of aggression: men commit over 90% of homicides. One way to close that gap would be to reduce male violence; another would be to increase female violence. The latter would be disastrous, because violence is the problem, not the difference in violence. Therefore, equality is not an unqualified good. This principle extends to other domains: we should aim to reduce harmful behaviors and remove discriminatory barriers, but not force equal outcomes if they stem from freely chosen preferences. He contrasts this with the soft bigotry of male expectations, where male-typical traits (like risk-taking or career ambition) are taken as the standard, and deviations by women are framed as deficits needing correction. Real fairness, he argues, is letting people be themselves and judging traits by their harm or benefit, not by whether they eliminate a statistical gap.