Try talking about whether single sex education is better for boys, or why there aren't more female science professors at Harvard, or whether male financiers are innately more aggressive, and sooner or later someone will evoke that handy, biological explanation of sex difference: fetal testosterone. The usual argument is that early hormonal exposures mold "male" and "female" minds. That is, a prenatal marinade helps shape men and women's later sexual desires, intellectual talents, personality traits and career interests—in ways that typically differ by gender. A loud chorus of researchers and popular writers, including psychologists Simon Baron-Cohen, Susan Pinker and Steven Pinker, psychiatrist Louann Brizendine, and therapist Michael Gurian, frequently invoke prenatal hormones to explain male and female behavior.
At first glance, the science seems to offer strong backup: Hundreds of articles report a relationship between prenatal hormone levels and, say, which toys girls and boys like as kids or how aggressive they are or how easily they can mentally rotate objects or (when they're a bit older) how much they masturbate. The usual assumption is that since males have more prenatal T, and T is the male hormone, of course it must explain the hallmarks of maleness in the mind.
This view sucks up a lot of oxygen, and it takes painstaking effort to refute it. I tried to in a Slate series on psychological sex differences that sent me deep into the weeds for months. (Every writer I know who's taken on this literature shudders a little at the memory.) So it was with appreciation verging on glee that I read Barnard professor Rebecca Jordan-Young's devastatingly smart and definitive critique: Brain Storm: The Flaws in the Science of Sex Differences. Jordan-Young argues that the science of prenatal hormones, gender, and the mind "better resembles a hodgepodge pile than a solid structure." And she knows of what she speaks.
An expert on measures and study designs, Jordan-Young has spent the last 13 years combing the literature on brain organization, unpacking assumptions, questioning methods and statistical practices, holding one paper up against another. She stresses that fetal hormones must matter to the brain—somehow. But after picking apart more than 400 studies that try to understand the genesis of particular psychological sex differences (real or supposed), she concludes that fetal T looks like an awfully anemic explanation.
Prenatal hormones are, indisputably, important to genital development. But the more controversial question has always been whether the mind is also shaped in categorically "male" or "female" ways, by testosterone or estrogen. For decades, a series of researchers have doggedly tried to prove that it is, treating the brain as an "accessory reproductive organ," as Jordan-Young writes.
This presumption has led to some strange contortions and flip-flopping. In the early 1970s, for instance, one study reported that boys whose mothers took a synthetic estrogen, which was thought to prevent complications in pregnancy, tended later on to show less masculine behavior like lower assertiveness or aggression. Soon, however, animal research found that testosterone converted to estrogen in the brain actually seems to do the opposite—to make boys act more typically masculine. And soon after that, the reversal showed up in human research, too: Another study suggested that boys exposed prenatally to a synthetic estrogen might show more masculine behavior, according to Jordan-Young. The data on synthetic estrogen and progesterone exposures are complicated. But what's striking, in Jordan-Young's telling, is that the researchers' findings tended to dovetail suspiciously well with their assumptions.
The red flags only multiply as Jordan-Young compares more-recent results. Of course, scientists cannot go in and vary prenatal exposure to hormones and then see what happens in a controlled experiment. Nor can they measure anything directly in the fetal brain. Instead, they must rely on "quasi-experiments" and proxy variables for hormone levels.This puts particular pressure on how well different findings fit together, Jordan-Young argues.To take one example: Psychologist Melissa Hines has measured testosterone in pregnant women's blood and linked higher levels of the hormone to more masculine behavior in 3½-year-olds, measured as "involvement with sex-typical toys, games and activities." Meanwhile, psychologist Simon Baron-Cohen has measured testosterone in amniotic fluid and linked higher levels of hormone to tendencies he deems typical of males (less eye contact at age 1, poorer social relationships and more restricted interests at age 4). The only problem is that Hines found her link for girls but not for boys. On some measures, Baron-Cohen has found a relationship just for boys and on other measures for boys and girls taken together. For play behavior, he reports no link with T levels for either gender. So it's hard to put these projects together and come up with a consistent story. Instead, it's a mishmash.
Research on people with intersex disorders comes in for similar needed scrutiny. A large number of studies look at women and girls with congenital adrenal hyperplasia, a genetic disorder that involves the overproduction of androgens like testosterone. Because these females' hormonal exposures are closer to males', the theory is their behaviors and interests ought to tend toward the "masculine" as well. And in fact, some data suggest that CAH females may be more likely than their sisters to play with vehicles and construction-related toys as children. They may be "less likely to prioritize marriage and motherhood over career," writes Jordan-Young. Also, they may be more likely to express an interest in male-dominated careers like engineer and airline pilot. This evidence has played a starring role in debates over whether fewer women hold tenured positions in these fields because they are innately less interested in the subject matter. Why push for parity in any male-dominated field if you will just run up against nature, as inscribed by fetal T?
But it's never been clear how relevant the research on CAH girls and women is to other females. Many CAH patients are born with masculinized genitalia. That might make them feel freer to express preferences that are less common or acceptable for girls, like a desire to fly planes for a living. CAH girls also go through extensive medical monitoring and treatment, and that, too, may make a difference. Jordan-Young cites an intriguing experiment from the mid-1980s in which researcher Froukje Slijper looked at girls with diabetes, as well as girls with CAH and girls without either condition. What she found was that "both groups of girls with chronic illness scored in the more masculine range" than the normally healthy girls. Whatever the reason, the finding should give us pause: It's tricky to link up girls with CAH and those who don't have a serious, T-related disease.
So where does this leave us? In light of Jordan-Young's meticulous synthesis, it's hard to name any specific feature of male-typical or female-typical behavior that consistently matches up with prenatal T levels across several models of research. No measure holds up: not aggressiveness, or masturbation habits, or even the ability to rotate objects in the mind, long viewed as the gold standard of sex-difference research because it is a skill on which average men and average women reliably differ. There are studies to cite for all of these claims. But what's missing is corroboration across approaches—from studies that look at different kinds of people (say, those with a disorder like CAH or those without), or that try to gauge prenatal hormone exposures in different ways (say, through amniotic fluid or maternal blood or finger-digit ratio). After decades of determined research, if robust links between prenatal hormones and "male" or "female" minds really exist, shouldn't we see those links across lots of different kinds of studies?
This matters because the obsession with prenatal T can easily become a distraction. It can make us forget how much gender norms have changed—think of all those female accountants, lawyers, and doctors who weren't around 50 or even 30 years ago—and how remarkably similar men's and women's brains and minds actually are. All this unwarranted hammering away at difference (and its putative explanations) causes real trouble, too. As a growing body of research shows, cues that foreground gender and bring stereotypes to mind can dampen men's performance on tests of social sensitivity, women's scores on math tests, and women's stated interest in quantitative pursuits Jordan-Young has done an enormous amount of work to untangle the gender claims. We ought to read her, cite her, thank her. And then, let's move on.