Soooo it’s happened again. There’s another pop science article arguing that there are “real differences” between “female and male brains” and citing hormones as the root of gender differences in cognitive ability. There are so many of these that I usually just scowl at them as they go by, but every so often there’s one so blatantly terrible that I can’t help but swing at it. This recent Slate article was one such doozy. Get out your baseball bats.
According to the initial report of the study that the Slate article is writing about, these researchers took pictures of the brains of 18 trans individuals undergoing hormone therapy as part of their transition to living as male. Over the course of 4 weeks, these researchers found statistically significant decreases in the gray matter of two brain regions involved in language processing. They also found significant increases in the white matter connecting with these areas, which they say they found “surprising” — translation: “does not support our conclusion.” Their conclusion, of course, is that hormones, specifically testosterone, may contribute to gender differences in verbal ability. The pop science article extends this to imply that hormones could totally explain all such functional gender differences.
We’ve got great stuff to work with here.
Part 1: The Obvious
This study is looking at a group of 18 people. That’s a pretty small number, and your guess is as good as mine as to whether that number is sufficient to state with any level of confidence that these results are typical of transgender individuals, let alone all 7+ billion members of humanity.
However, arguably more damning: this study does not have a control group. What would have been a very interesting and informative study is if these researchers had taken 36 individuals about to undergo hormone therapy with testosterone and randomly assigned 18 of them to a wait-list condition. Repeating the original report on the study, Slate implies that this kind of research is simply unbearably tricky to do ethically and therefore has not even been attempted before due to ethics concerns. Although not explicitly stated, I suspect that “ethical concerns” is the researcher’s reasoning for why they don’t have a control group.
However, if this kind of research truly hasn’t been done before because it wasn’t considered ethical, then why are people doing it now? As it turns out, randomized, controlled research on hormone supplement therapy has actually been around since 1950 among women being given hormone therapy for the treatment of menopause. Consequently, I doubt ethics had much to do with the delay in conducting this kind of research with transgender individuals. The 65 years that have passed since that 1950 study seem to suggest that medical and psychology researchers historically have just not given a shit about including transgender individuals as participants in their work until they saw an opportunity for trans bodies to be useful to them.
Accordingly, I don’t buy the potential defense that a randomized, controlled design would have been impossible to do ethically via a wait-list condition. Medical and psychology researchers do this all the time when testing out a new medication or therapy. With proper informed consent, I feel pretty confident that the researchers could have found participants willing to chance waiting 4 weeks to begin hormone therapy while taking a placebo for the sake of science and/or appropriate compensation.
The point being, the design of this study is pretty unfortunate and has opened up its results to any number of possible explanations. Without a control group, there is simply no way to know if the hormone treatment is what is causing the effect that we see here. For instance, perhaps there is research out there indicating that JUST THINKING OF YOURSELF AS MALE VAPORIZES GENDER DIFFERENCES IN COGNITIVE ABILITY, ostensibly by reducing stereotype threat. Beginning hormone therapy and starting a physical transition process might have prompted the participants to begin thinking of themselves more as male, resulting in more stereotypically male cognitive patterns that altered the shape of their brain structures, since thoughts and experiences can influence our brain structures (I probably just lost all my credibility by citing Wikipedia, but I promise, neuroplasticity is a thing). The explanation for the brain changes observed in this study is pretty much up for grabs, because with no control group, the true cause of those structural changes is anybody’s guess.
Part 2: The Less Obvious.
A related reason why I hate pop science articles (I realize this blog post is dangerously close to becoming one, but I feel that I am still safely in the territory of a self-righteous rant) is because reporting on a single study while ignoring the rest of literature on the subject tends to neglect important information. New studies should be viewed in context of other work on the subject, and all results should be replicated before we view them as reliable, which happens way less often than it should.
Some of the information that we miss when ignoring the broader context of research can be extremely important. For example, if hormones are primarily responsible for altering brain structures thought to cause gender differences in verbal ability, we might expect to see a comparable effect in other species. As it turns out, there is some fascinating work on this subject which did not make it into the Slate article because why spend time reading background literature or consulting with experts when you can just — not.
Zebra finches have song patterns differentiated by sex. They also have sex hormones that include both estrogen and testosterone and, in the 1980’s, were thought to result in the different brain structures that correspond with sex differences in song patterns. However, about 10 years ago, scientists discovered a zebra finch with with physical structures divided almost perfectly in half between male and female sex characteristics — including its brain. The right half of the zebra finch’s brain had structural characteristics for male song production, whereas the left half of the finch’s brain had structural design for female song production. Given that hormones circulate throughout the entire body, the researchers concluded that hormones could not be the cause for the dual sexes observed in the zebra finch’s brain structure and plumage, nor could hormones explain the finch’s male-typical courtship behavior.
True, this is a bird, not a human. But if hormones are the primary cause of sex differences in brain structure and, consequently, verbal ability, we would expect that to be the case in other species that have both sex hormones and sex differences in vocal abilities (in this case, song patterns). That does not seem to be the case for this glorious bird.
Part 3: The Totally Obscure
I originally wrote a section here critiquing the researcher’s analysis methods. This was before I saw the abstract. As it turns out, their analysis actually looks pretty solid, with large effect sizes and correcting across the whole brain for family-wise error – that is, correcting for the fact that the more times you conduct the same type of statistical test, the greater your chance is of getting a false positive, a result that appears significant but actually isn’t. This is important: researchers have found that not correcting for family-wise error can result in findings of significant brain activity that indicate active contemplation of emotional stimuli in the brain of a dead salmon. Read it and weep.
Anyway, these researchers’ analysis actually looks pretty sound, even though I’m still not thrilled with their design. Which leads us to the main takeaway.
The Main Takeaway
The major problem with this study is not as much with the study itself as it is with the way that Slate chose to interpret and present the findings. Unlike the straw-person feminists presented in the Slate article’s opening salvo, no one is saying that hormones have nothing to do with sex, gender, and gender identity. In fact, Luz has an amazing series on the role of biology in sex and gender (“Take the Red Pill“) that if you haven’t read yet, you totally should. However, the difference is that, as Luz points out so eloquently, just because hormones and genetics play a role in their development does not mean that sex, gender, and gendered behavior are not also socially constructed.
Overall, as a general rule of thumb, it should take a lot more than a single study with 18 people, no control group, and findings presented out of context from the relevant body of literature to convince us that all gender differences in verbal ability can be chalked up to hormones. But nice try, Slate.