Mary Stewart Van Leeuwen
Neurohormonal Wars, Part 2
Intersex persons are burdened enough by disputes about how discrepancies between their genotype and genitals should be handled. But their psychological makeup is also part of the battlefield on which the nature-nurture dispute is fought. To explain why, Rebecca Jordan-Young, in Brain Storm, cites experiments with animals that traced the effects of deliberate pre- or post-natal hormonal interference on their subsequent social and sexual behavior. Brain Organization Theory (BOT) researchers have used these findings to suggest that every mammalian brain (human included) "is a sort of accessory reproductive organ." In other words, along with differing gonads and other structures needed for heterosexual mating, "[m]ales and females also need different brains so they are predisposed to complementary sexual desires and behaviors that lead to reproduction."
As Jordan-Young summarizes:
This theory suggests that regardless of chromosomal sex, having a male-typical hormonal milieu in utero leads to male-looking genitals and "masculine" psychological traits, including erotic orientation toward women, as well as broadly masculine cognitive patterns and interests. Likewise, a female-typical hormonal milieu leads to feminine-appearing genitals and "feminine" psychology, including erotic orientation to men …. Moreover, sexual differentiation is not restricted to those behaviors that are directly involved in reproduction, or even in courting. Instead, brain organization theory is used to explain a very wide range of differences related to gender and sexuality—in humans, including everything from spatial [ability], verbal ability, or math aptitude, to a tendency to display nurturing behavior, to sexual orientation …. [T]he core assumption is that masculinity and femininity are package deals with reproductive sexuality at the core.
How well is the theory substantiated? To answer this, Jordan-Young reviewed over 300 studies, covering the various research designs used to test the hormone-psychology connections summarized above. She also interviewed almost two dozen prominent brain organization scientists. The result is a very detailed assessment, and one that could initially intimidate readers with little training in research design. But Jordan-Young is a clear and careful writer, so her book will repay persistent readers at all levels of methodological literacy. She is no Cartesian dualist: she clearly respects the constraints of embodiment in animals and humans alike. She also defends the freedom of scientists (of whom she is one) to use deterministic models to explain psychological and social processes, not just more obviously physical ones. But she thinks that BOT research is plagued by bad theory and shoddy methodology, often in the service of defending the theory in the face of ambiguous or contrary data. Ironically, despite the much-touted scientific trappings of their work, many BOT researchers routinely violate the central Popperian criterion for being scientific—namely, willingness to revise a theory when it is confronted by consistent empirical failures. Jordan-Young shows that some BOT adherents are also too quick to accuse their critics of blindly following the forces of postmodern political correctness, when in fact many are motivated by legitimate methodological concerns.
What are some of these concerns? Jordan-Young notes that when you can't do with humans the kind of controlled experiments that can be done with animals, you have to settle for a series of quasi-experiments, in which you look for consistent patterns of correlation across time and place, and across differing samples and research designs. There are rules for doing good quasi-experiments. For instance, we obviously couldn't establish a causal link between cigarette smoking and lung cancer by randomly assigning some people to smoke real cigarettes and others to smoke fake ones for several decades, then comparing their differential cancer rates. So epidemiologists did the next best thing: longitudinal studies tracking the health of existing smokers and non-smokers. (This is known as a cohort study design.) They also used the opposite approach: finding people who already did or didn't have lung cancer, then retrospectively assessing their rates of cigarette consumption. (This is known as a case-control design.)
The reason the smoking-cancer link is so well accepted is that, over the years, these quasi-experimental studies shared several important features. First, the "input" and "output" variables, i.e., "smoking" and "lung cancer"—were consistent, both in terms of what they represented and how they were measured over time. Second, the number of people studied was very large. Third, the smoking-cancer correlations showed up in different kinds of research designs (e.g., both cohort and case-control studies). Fourth, most of the correlations between smoking and lung cancer were large and statistically significant. Finally, and critically, there was a consistent "dose-response" relationship across all types of studies: more smoking, more lung cancer risk; less smoking, less lung cancer risk. (There would hardly have been a legal case against the cigarette companies if a significant percentage of studies showed that smoking a pack a day for years made no difference, or worse, even reduced the chances of lung cancer.)