On Barres On Women in Science

Stanford professor Ben Barres has an editorial in Nature this week arguing against what he terms the "Larry Summers Hypothesis:" Women don't make as strong a showing in science because of innate ability.

First of all, I think it's more than a little unfair to term this the Larry Summers Hypothesis. Summers said in his notorious speech, "different socialization and patterns of discrimination in a search" is the third most important factor in why women aren't achieving. He also said the most important factor was the "high-powered job hypothesis:" Men are prepared to commit more hours to their jobs, at least partially because women have to choose between high-commitment jobs and family life. The middle factor was intrinsic aptitude, and he frames the difference in aptitude as a difference in standard deviation, which only affects the extremes to a severe extent. He doesn't deal with average differences. So that's why I think the name of this hypothesis is unfair.

Briefly, dwelling on Summers, Barres indirectly acknowledges that Summers is on target with his MOST important reason why women aren't achieving: "Women faculty, in particular, need help from their institutions in balancing career and family responsibilities." This means there is a real effect of women not being able to commit time. Great. You two boys are in agreement.

Barres spends relatively little his space actually arguing that the data say intrinsic aptitude is not a factor, although some of his numbers are compelling. A third of top Putnam scorers are women, and a study of 20,000 kids showed no difference in math scores between the genders. I'm a little suspicious of that because, again, we're not looking at means, just extremes, and there are about 5 kids in that 20,000 who fit our definition of extreme. Finding no difference there is still not really significant. (Someone tell me if this is flawed statistical reasoning) [Update: I realize in the morning light that this is, indeed, horrible reasoning. With 20,000 subjects you can easily determine a curve and a standard deviation and extrapolate this to the extremes, even if your extreme data doesn't match this perfectly. Still, I would like to see that these curves and not just their averages are identical.]

Most of Barres's essay is showing evidence that there is real, real discrimination. This is nicely underscored by his personal story. Ben Barres was Barbara Barres until only about 10 years ago, and so he is a unique (read: rare) position to compare how people treat men and how people treat women. Admittedly, it's not a scientific comparison...the difference between Barbara and Ben isn't PURELY gender; how a personality and set of mannerisms interact with one's gender is a strong, strong factor in how a person is perceived. Still, his position is special and should be considered. So after he transition, people treated him and his research with an obscene increase in respect. The money quote from the editorial is from a professor who saw him give a talk right after he transitioned, "Ben Barres gave a great seminar today, but then his work is much better than his sister's." Ooh. Zing.

The most persuasive part of the essay, I'd say, is the evidence that women face discrimination, "one study found that women applying for a research grant
needed to be 2.5 times more productive than men in order to be considered equally competent." There is perhaps too little statistical evidence in the editorial, but there is some.

And some of his studies are a little shaky, "A 2002 study did find a gender gap in competitiveness in financial tournaments, but the authors suggested that this was due to differences in self confidence rather than ability." Suggested? I don't yet know why to take his word on it. Even if the problem is self-esteem, I don't want to take anyone's word that this is caused by discouragement and not by nature. Gimme more info.

In sum, he convinces me that there is legitimate discrimination, but I am not totally convinced that both ability at the extremes and willingness to forsake family to commit are not equally important factors. If that study on productivity is totally legit, that would strengthen his claim a lot. I'm just initially wary. It's a decidedly important editorial, and I think this conversation should be ongoing. I'm glad he's kicked it up again.