Opening the book…
One assumption that has run through most of history was a supposed connection between skull size and intelligence. Researchers treated skull size or capacity as a measure of brain size and assumed brain size was a measure of intelligence. Both of those inferential steps are wrong. Skull capacity correlates with brain volume only imperfectly, and brain volume is correlated more strongly with body size than with intelligence (Eliot et al., 2021). None of the 19th-century measurements we are about to discuss were measuring what their authors claimed to measure. They were measuring skulls.
In 1879, Gustave Le Bon published a long article in the Revue d'Anthropologie on the variations of skull capacity and their relation to intelligence. The article draws sweeping conclusions about women's inherent inferiority. It also, quite accidentally, records evidence that contradicts those conclusions. The contradiction is worth examining because it illustrates how data was interpreted through a biased framework that justified the status quo, and thus the (male) researchers’ place in it.
Le Bon observed that the gap in average skull capacity between men and women was larger among Parisians than among the groups he considered less advanced, where there was overlap between men and women. He attributed this to Parisian women's educational environment. The schooling a Parisian woman received, he wrote, "in no way exercises her intelligence, and tends rather to restrict it than to develop it" (Le Bon, 1879, p. 58). Women in “less civilized societies” where there was overlap between male and female skull sizes, shared men's labor and were “made industrious by necessity” he noted. Despite clearly observing this environmental difference for women, Le Bon interpreted the wider gap between Parisian men and women as evidence of all women's inferiority. A more justifiable reading of his own data could be that the “cognitive differences” he was measuring were produced by what women were allowed to learn and to do. And remember that he wasn’t even measuring cognitive differences, just skull size.
Le Bon also observed that male and female skull-capacity distributions overlap. A “certain number” of women, he acknowledged, had skulls larger than a “certain number” of men of the same population. His response was to call the overlap small, dismiss the overlapping women as “merely mediocre”, and compare all women’s brains to those of gorillas (Le Bon, 1879, p. 60). He did this despite the fact that, by his own measurements, the overlapping “merely mediocre” women’s skulls were larger than the average male skull.
Figure 7.1. Schematic illustration (not plotted from data) of overlapping male and female score distributions, after Hyde (2005). Figure by the author, CC BY-NC 4.0.
The overlap Le Bon’s bias treated as noise was the more meaningful pattern for the psychology of women. Nearly a century later, Janet Hyde (2005), in a meta-analysis of meta-analyses (how meta!), showed that across gender differences research, gender similarities were far more striking than differences. Hyde’s gender similarities hypothesis does not claim that men and women are identical. It claims that within-group variation is so much larger than between-group variation that population averages tell us almost nothing about what any individual man or woman can do. Figure 7.1 illustrates how the similarity between male and female score distributions is much larger than the difference between their averages. This means that if we looked at the score of a random participant from a gender difference study included in Hyde’s meta-analysis, we would not be able to say with confidence whether that person was male or female. Le Bon had the data to reach the same conclusion; he even noticed the environmental difference. He chose to ignore environment to support the illusion of female inferiority.
The rest of the gender differences research history follows similar patterns. Phrenology, in the early 1800s, claimed to identify character and ability from skull shape and used the framework to justify excluding women from higher education. Craniometry, which Le Bon contributed to, replaced phrenology by mid-century and produced the same conclusions with more precise instruments. Paul Broca's (1861) measurements were used to rank women and non-European peoples; anatomist Robert Bennett Bean reported in 1906 that Black brains had smaller frontal lobes, a finding his own mentor Franklin Mall debunked three years later (Mall, 1909). Mall's correction, and Woolley's broader condemnation of the entire field from the second opening quote, received far less attention than the original claims. Sexist assumptions continued to shape popular and scientific understanding for decades.
This pattern of letting bias drive interpretation, now called neurosexism (introduced in Chapter 3), continued into the twentieth century. In the mid-20th century, functional arguments replaced structural ones. Men were said to have more hemispherically lateralized brains, meaning that left and right hemispheres were more specialized for distinct functions. This was offered as the basis for male spatial advantages and was critique by Alper (1985) and others. Lateralization of function has not held up in large samples (Eliot et al., 2021; Hirnstein et al., 2022). The arrival of brain imaging in the 1990s produced a new wave of claimed sex differences, most of which have not replicated. Eliot and colleagues (2021) synthesized three decades of MRI (structural and functional Magnetic Resonance Imaging) and postmortem data and found that once body size is accounted for, sex and gender explain roughly one percent of variance in brain structure. And because brains change as a result of experiences throughout life, even that one percent could reflect the different experiences of people socialized as men and women rather than any innate biological difference. Eliot (2024) has since argued that the default assumption in this research should be similarity rather than difference.
|
Research Highlight: Dump the "Dimorphism" and Remember the Null A sexually dimorphic trait is one that shows two distinct forms in males and females, with little to no overlap between them. Animal features like peacocks' tails, antlers, and body size differences are common examples. For decades, neuroscientists assumed the human brain would fit this pattern. Lise Eliot and colleagues (2021) reviewed three decades of neuroimaging research on male-female brain differences in order to directly address the neurosexism introduced in chapter 3. They found that almost every widely cited "difference" dissolves once brain size is accounted for. Males' brains are about 11 percent larger on average, and this size difference drives most of the reliable findings (white/gray matter ratios, regional volumes, connectivity patterns). When size is controlled, sex/gender explains only about one percent of the variance in brain structure. The authors note that task-based fMRI has been particularly unreliable, with most claimed activation differences failing to replicate across studies. Their article title sums up the argument: it's time to stop calling the human brain sexually dimorphic. In a 2024 follow-up, Eliot urged neuroscientists to take the null hypothesis (the one that assumes no differences) seriously when searching for brain sex differences. |
The pattern across time is consistent. In every generation, the technology of the day has produced observations that could have included a developmental and environmental account of cognitive differences between men and women. In every generation, those observations were instead forced into a framework that treated the differences as natural, fixed, and favoring men in power, including those doing the research. Such history should make us cautious about any confident claim of brain-based cognitive sex difference, including claims made today.
This page titled 6.2: A History of Looking for Differences, and Finding Them was last modified on Sun, 13 Sep 2026 21:12:15 GMT and is shared under a CC BY-NC-SA 4.0 license and was authored, remixed, and/or curated by .
The book could not be loaded just now.