Back from his holidays, Chris Bertram has comments about the male brain/female brain multi-choice quiz that I complained about before. Chris stakes out a reasonable position. He’s a bit more precise than I was when I wrote my original post—my excuse is that at the time I was, to quote Raymond Chandler, full of no coffee.

However, I’d stand by what I said. The main point as I see it is that the questionnaire instrument is no good. Both the ‘Type S’ and the ‘Type E’ quizzes suffer from the following serious defects:

  1. They’re full of leading questions.
  2. They assess one’s self-perception rather than actual abilities. The former is much more tightly linked to gender role expectations than the latter.
  3. The examples chosen are correlated with gender-typed behavior and so pick up on typically male (or female) activities when no inference about typically male (or female) brains is warranted. Alternative questions could easily flip the responses, as I showed in the original post.

So, frankly, I wouldn’t believe any strong claims about S-brains and E-brains based on an instrument like this. Nevertheless it gets picked up by the media and put through a stock frame for writing stories about the fundamental differences between men and women.

Chris points out that Cohen’s claim is that there are S-brains (male-type) and E-brains (female-type) but “Not all men have the male brain, and not all women have the female brain. … on average, more males than females have a brain of type S, and more females than males have a brain of type E.” This might well be true, and it’s a more precise formulation than I originally allowed. I just don’t think that the survey instrument can show anything of the sort. The instrument assumes the existence of S and E type brains—even having two separate quizzes with those labels is telling the subject that S and E types exist. Each survey is then filled with questions based on gender-typed activity, so it’s practically guaranteed that more men than women are going to score higher on the S test, and vice versa on the E test. He might as well have printed one on blue paper and one on pink.

Chris goes on to give some examples of sex-linked characteristics (e.g., color perception, prevalence of autism) that can’t be attributed to culture or social structure. He then suggests that skill differences in men and women, if such there be, don’t have strong implications for our moral views. Neither of these points are controversial. Note, though, that we’re now two steps away from the very weak evidence for the very strong original claim. Well-defined sex-linked characteristics like tetrachromat vision are a long way away from ‘Male-type brains.’ Granting differences for the sake of argument is quite different from claiming to empirically establish their biological basis. Again, the real issue is methodological. (Incidentally, tetrachromat vision is 100% sex-linked and so, as Chris himself points out, a very different claim from the partial overlap of S-type/E-type brains on the male and female populations.)

My main gripe with studies of this sort is that they are often badly designed and routinely oversold in popular books and media coverage. This can happen because people are happy to believe in the kind of Mars vs Venus differences that these studies inevitably discover. More often than not, closer inspection reveals that only much more modest claims can really be sustained. But by then attention is usually focused elsewhere.

Again, none of this prejudges the empirical question about the origins of behavioral differences between men and women. It’s just irritating to see the the gun being jumped over and over again in the rush to confirm our common-sense theory.