Sunday, December 26, 2010

My Christmas Reading: The Science of Sex Differences

I finished Lise Eliot's Pink Brain, Blue Brain. I then read Cordelia Fine's Delusions of Gender, and then Rebecca M. Jordan-Young's Brainstorm.The Flaws in The Science of Sex Differences. I have read the other side of the debate in the past.

Short conclusions (much more will appear in later posts:

1. Do not read all these in two days or so. You get tremendously angry and then you kick holes in the garage door.

2. The history of sex differences in science is a nasty one, and there's not much reason to expect we are doing much better right now.

The general tendency seems to be to grab onto any particular innate explanation as the right one. In the nineteenth century it was women's smaller brains or something odd about the spinal cord connecting the pelvis to the brain or wandering wombs.

In the twentieth century it was any observable difference in some brain part, and any differences found were then interpreted as meaning that they show why men are better, pretty much. In some cases the differences turned out to be reversed, but suddenly THAT was why men did better in some cognitive field.

And all this was always interpreted as "hard-wiring", not possible to change and eternal, despite the fact that the actual behavior of men and women was changing at the very same time, and despite more recent findings which show the brain changing with how it is actually used. In short, we (or some group of people) have a bias to welcome findings of innate differences and a bias to resist a more complicated treatment of the topic. I have seen exactly the same with the recent chimpanzees and sticks study.

3. Anyone willing to study innate sex differences between men and women is already biased in one particular direction. This is ignored almost totally, even though anyone finding sex similarities is labeled as biased.

4. Really bad research is regarded seriously when it reinforces basic stereotypes. Really good research is ignored when it fails to reinforce basic stereotypes.

To give an example of the former, studies which measure the correlation between unusual prenatal testosterone exposure in girls and "masculine" behavior almost always use a large number of measures for "masculine" behavior. Almost all of the published studies in this field find one or two significant differences. But they may have studied fifteen such measures or even twenty measures! Yet the finding of those one or two significant differences is publicized as an important one. Statistically speaking, such findings are quite likely to be flukes.

5. The "file drawer" aspect (i.e., that certain types of studies fail to get published) is tremendously strong in this field.

It takes two forms: Studies which don't find differences are often not published at all OR are published in a form which focuses on something else than the sex difference, and later studies which cannot reproduce the differences found in an early (and now very famous) study are not disseminated outside the narrow research community.

The consequences of these are that people in general believe science has found a humongous number of purely innate differences and people go around believing research results which have been falsified in later (invisible) research.

6. I'm becoming very suspicious of a few "researchers" or popularizers as actually knowing that they are lying or utterly not caring one way or the other, just to sell books or ideas. I even wonder about the actual research of a few of them. This is because nobody has been able to replicate the oh-so-famous findings, even in studies which were much better done in the methods sense.

7. If a difference is found which appears to handicap men or boys (say, verbal skills and reading skills), the next step is to argue for compensatory actions in the society. If a difference is found which appears to handicap women or girls (say, three-dimensional mental rotation abilities), the next step is to argue that nothing can be done at all, and that any compensatory treatment should cease this minute.