Mary Stewart Van Leeuwen
Neurohormonal Wars, Part 2
Jordan-Young applies these quasi-experimental research standards to her hefty sample of studies designed to test BOT. She concludes that, far from pointing to likely causal relationships between "brain sex" and sex-typed behavior, both the theory and its research methods are in need of drastic overhaul. Consider, for example, the many studies that have tried to demonstrate more "masculine" personality traits, skills, sexual interests, and the like in females with the intersex condition known as congenital adrenal hyperplasia (CAH) than in the normal population of females. When these are taken together (and they must be: even the findings of a double-blind trial require at least one replication, and less-controlled quasi-experiments are held to an even higher standard), the evidence is thin. For example, there is no consistent evidence that CAH females have superior spatial skills, of the sort that help you orient yourself geographically or do jigsaw puzzles. In normal populations, there is a modest but statistically significant difference in some spatial skills that favors males. But when assessing CAH females or males most studies show no advantage, or (more embarrassingly) poorer spatial skills in CAH persons than in normal controls of their own genotype. This clearly violates the dose-response rule: just as you would expect increasing rates of smoking to correlate with increasing risks of lung cancer, so (according to BOT) there should be a positive correlation between people's prenatal and/or postnatal "doses" of testosterone and their scores on sex-related cognitive tasks. At the very least, the correlation shouldn't be reversed. Nor are CAH girls reliably more aggressive, assertive, competitive, or dominant than their non-CAH peers. Nor are they more likely than girls in general to engage in rough-and-tumble play as children, or to prefer male playmates in childhood, or seek out female sexual partners in adulthood. And all this is despite the fact that, having been born with masculinized genitals, there may be the expectation on the part of parents and others who know of their condition that they will be stereotypically more like boys. Remember, you can't do a double blind experiment when you're raising children.
There are two exceptions to these disconfirmations of brain organization theory. One is that CAH girls, compared to matched controls, are more likely to say they prefer so-called boys' toys, such as building blocks or vehicles, to toys like dolls or cooking sets. But here we run into two problems. The first is that what people say and what they do are often discrepant, and very few studies have looked at what toys CAH girls actually do play with compared to normal controls. A 2003 study that did so (in an individualized playroom setting) found that CAH girls were indeed more likely than others to play with a toy garage, cars, and Lincoln logs than with baby dolls. This was taken as support for a version of BOT that sees people with testosterone-organized brains as more likely to be "systematizers" (interested in how things are put together, both concretely and abstractly) than "empathizers" (interested in understanding other people). However, a closer look at the data shows that even the normal control girls spent three times as much time playing with the garage and cars as with the baby doll, and six times as much with the Lincoln logs. And at the end of the play period, when offered the choice of a doll, a car, or a ball to take home, they chose the doll least often, by a wide margin.
If you're wondering why, in 2003 (well after girls started playing soccer in droves), BOT researchers were coding "masculine" and "feminine" toys the same way they did in the 1950s, that's the second problem. It reflects the essentialist presumption that gendered behaviors, along with normal genitals, are parts of a fixed package, and so their measures can remain as unchanging as measures of genital anatomy. It presumes that normal girls with low testosterone can be only minimally influenced by cultural shifts to start preferring Lincoln logs to baby dolls (once an empathizer, always an empathizer), even when data gathered by BOT researchers themselves indicate otherwise.