Are male brains different from female brains? If so, how? And does it matter?
This week, five researchers debated these questions at the annual meeting of the Society for Neuroscience. Their panel session, “The Promise and Peril of Research on Sex Differences,” didn’t settle the controversy, because it isn’t binary, and evidence is complex. But the exchange did clarify common mistakes to watch out for. Here’s a guide.
1. Ideology. All the panelists recognized that sex-difference research could be abused to justify sexism. But Larry Cahill, a behavioral neurobiologist at the University of California-Irvine, raised the opposite concern: His colleagues are so afraid of being called “neurosexists” that they’ve refused to study or acknowledge differences. This anxiety about lending credence to sexism was manifest on the panel, as three of the presenters repeatedly emphasized similarities and downplayed differences. Afterward, they were challenged by two female scientists in the audience who called the aversion to studying innate differences anti-scientific and an impediment to understanding mental illness in women. The exchange, in which one panelist repeatedly portrayed sex-difference research as a waste of time, confirmed the problem: Fear of sexism has produced a bias against conceding sex differences, which gets in the way of frank discussion and investigation.
2. Monocausality. Melissa Hines, a psychologist at the University of Cambridge, proposed a good rule for screening studies and news reports: Beware any explanation that relies on a single factor. Hormones matter, but so does socialization. A study in animals might illuminate the role of genes, but it won’t capture the effects of culture on humans. Conversely, anyone who dismisses boy-girl differences as cultural artifacts (the panelists criticized Cordelia Fine’s Delusions of Gender in particular) isn’t accounting for similar patterns in animals, such as research showing that male monkeys prefer to play more with cars and less with dolls than female monkeys do. Hines also mentioned a male-female gap in “maze performance” among rodents. You can’t blame that on society.
3. Casual extrapolation. The problem with genetic or cultural theories of sex difference isn’t that they’re false. It’s that they’re limited. They work better in some contexts than in others. Hines recalled an incident in which, after she had described data on toy preference among girls, a male physicist said she had just explained why it was hard to recruit women to teach physics. The leap from dolls to doctorates was effortless, though groundless. Another psychologist on the panel, Janet Hyde of the University of Wisconsin, noted that sex differences in math performance had largely evaporated over the past 20 years. But not all differences: A stubborn gap remains in mental rotation, which requires the imaginary realignment of three-dimensional shapes. Cahill offered a sensible way to think about these uneven findings: The effects of sex difference on behavior and performance vary, and as a researcher of emotional stress, he acknowledged, “I may be working in a domain of neuroscience where these effects are maximal.” We need more of that humility, and less glib generalization.
4. Self-fulfillment. Maryjane Wraga, a psychologist at Smith College, presented research on stereotype threat, showing that women perform worse at mental rotation (compared with other women) when they’re told that men are better at it. So if scientists go around saying girls are bad with numbers, tests might appear to validate that prediction, but the prediction itself will be the culprit. The panelists were particularly alarmed by the single-sex education movement and the brain theorists behind it, authors Michael Gurian and Leonard Sax. These authors grossly exaggerate boy-girl differences, the panelists argued. But the greater danger is that single-sex education, by preaching and practicing segregated socialization, may exacerbate these differences.
5. Stereotypes. Girls differ from boys, but girls also differ from other girls. This in-group variation gets obscured by composite models such as The Female Brain (a book by Louann Brizendine, panned by the presenters) and binary metaphors like Men Are From Mars, Women Are From Venus. Sex differences don’t show up as separate clusters. They show up, in Cahill’s words, as “overlapping distributions.” Hyde explained that to compute the “effect size” of sex, you have to factor in the variability of scores among males and among females. Otherwise, you have no perspective on how meaningful the gap is between the male and female averages, relative to being Jane rather than Sally, or being Mike rather than Bill. You certainly can’t infer from a person’s sex how well he or she will do on a test.
6. Either/or. Hyde took a blowtorch to former Harvard president Larry Summers, presenting data that showed no sex difference in K-12 math scores. Then, to be fair, she acknowledged Summers’ more complicated argument: not that boys are better than girls at math on average, but that boys are more spread out, with lots of boys scoring very high or very low compared with girls, who tend to cluster more in the middle. Hyde argued that the K-12 data didn’t support Summers because the ratio of male variation in scores to female variation in scores wasn’t high enough to explain a shortage of women in some Harvard faculty departments. But on her presentation slide, every grade level showed a ratio between 1.11 and 1.21. In other words, the data did show greater male variation. This might contribute to professional gender gaps, though it doesn’t fully explain them.
7. Overinterpretation. With today’s technology, it’s easy to scan and measure brains and compute sex differences in size or activity. The hard part is figuring out what these differences mean. Yes, the brains of male fetuses and boys get bathed in testosterone. But does this really affect their math skills or their ability to communicate and process emotions, as some theorists assert? Yes, men have more gray matter, and women have more white matter. But does that justify CBS anchor Harry Smith’s conclusion—captured in a video clip by panelist Lise Eliot, a Slate contributor and neuroscientist at Rosalind Franklin University—that this is “why women are such good multitaskers”? The fishy part of neuroscience isn’t the data. It’s the spin we put on the data in the guise of explanation.
8. Inferred immutability. Before you attribute sex differences in behavior or success to evolution, check the record. Today’s differences may not have existed yesterday and may not exist tomorrow. The percentage of math Ph.D.’s awarded to women in the 1950s, according to Hyde, was half what it was in the 1890s and one-sixth of what it is today. Several panelists targeted the word hardwired as a misleading metaphor for explaining the brain. Brains, unlike computers, are constantly altered by experience. So while scans may show differences between men’s and women’s brains, that doesn’t prove the differences are innate. Wraga’s scans, for instance, showed different patterns of activation, but the patterns corresponded to social inputs. Even in animals, Eliot noted that male rats are licked and groomed more than female rats are, which could affect sex differences in stress response. And Cahill described new research indicating that birth control pills alter patterns of emotional memory. So, yes, hormones influence how we think. But we, in turn, can influence our hormones.
9. Data pooling. Beware broad generalizations based on the blending of data about various traits or activities. The panel made much ado about Hyde’s finding that 78 percent of effect sizes in studies of psychological sex difference were small or near zero. But that aggregate figure obscures the fine print: Half the effect sizes were between .11 and .35. In aggression, they averaged around .50, and in mental rotation, they were even higher. And if you read Hyde’s paper carefully, you’ll find that she breaks down these differences further. So while she’s right that males and females are largely similar, the details are intriguing. Likewise, Eliot’s observation that “most of our behavioral sex differences are quite a bit smaller than [sex differences in] height” obscures the curious fact, mentioned on one of Hines’ slides, that boys and girls differ more in toy preference than in height.
10. Comparison games. In science, as in politics, you can make a difference look big or small by choosing the basis of comparison. So while Hines’ slide compared the height gap to the toy-preference gap, Eliot’s slide compared the height gap to the much smaller “empathy” gap. While Eliot and Hyde characterized the effect sizes in sex-difference studies as small or near zero, Cahill argued that these effect sizes were no smaller than those typically found in other neuroscience research. (Indeed, I heard no complaints from the panel about small effect size when Wraga cited a 6-percent effect on math scores as evidence of stereotype threat.) Hyde acknowledged that boys scored “an itty-bitty bit better” than girls in math in the United States, Taiwan, and Japan, but she pointed out that the bigger difference is between the three countries, with Taiwanese and Japanese girls outscoring American boys.
These 10 warnings don’t add up to an answer on the overall question of sex differences. There is no answer. There’s only a complex, preliminary array of evidence on various questions, and an evolving menu of research to explore those questions further. Let’s not be afraid to pursue the research. And let’s not jump to conclusions.
William Saletan’s latest short takes on the news, via Twitter: