December 24, 2008

Math gender gap news

A couple of new papers have appeared recently, which paint two different pictures of the gender gap in science/math ability.

The New Scientist reports on the first one:
Why are there so few female Einsteins? Many people share a belief that while women can do science, there are far fewer women than men at the very top of the science hierarchy because women just aren't as innately good at science as men. Others feel this view is wrong but cannot easily put their finger on why.

They should be able to now. There are few women at the top of science because there are so few women in science. It's simple statistics.

...

To test this idea, Bilalic's team checked the records of the German chess federation – in which males outnumbered females 16 to one. They found that the statistical effect of this difference in numbers accounted for 96% of the observed difference in performance between the sexes. "There is little left for biological differences to explain," says Bilalic.

While statistics may explains the absence of females at the top, they don't explain why there are fewer females in the first place. If this is due to any innate differences in chess ability between the sexes it would have to influence whether children start to play the game at all, because the dropout rates for girls and boys once they do start are similar.

The sentence in bold, of course, undermines the whole argument. The pool of male chess players is much larger than the corresponding pool of female chess players. Hence, it cannot be assumed that male and female chess players represent random samples of the same portion of the ability curve.

To give a different example, there are probably far fewer female than male lumberjacks or miners. Those that exist may be every bit as good at their job as their male counterparts. But, the fact that there are so few of them isn't so much a question of preference, but of the physiological fact that men are on average much better at mining or cutting trees.

Women were, of course, denied access to certain occupations in the past. But, nowadays, and for several generations, women have had access to the professions and higher education, so it is becoming less tenable that their under-representation in science and math achievement is due to their lack of preference for these fields, rather than a small difference in ability, manifesting itself strongly at the high end of the ability curve.

The second paper undermines the thesis of the first, by demonstrating inherent differences between male and female non-verbal infants. From ScienceDaily:

"We've known for approximately 30 years that men and women can see an object from one perspective and then recognize that object after it has been rotated in space into a new position," said David S. Moore, professor of psychology at Pitzer College and Claremont Graduate University, both in Claremont, Calif., and an expert in the development of perception and cognition in infants. "In addition, while we have known that all people can do this, it turns out that men are quite a bit faster at it than women are. Previous studies have shown that this sex difference can be detected in children as young as 4 years of age, but our study is the first to have successfully found a way to assess the situation in young infants.

"Although we did not expect to find any sex differences in babies this young, our results suggest that the 5-month-old boys in our study used mental rotation to complete our task while the 5-month-old girls in our study did not," Moore said.

On the same subject, La Griffe du Lion has posted an analysis on the Math Sex Gap Revisited: a theory for everyone:
At the annual meeting of Women Against the Gap, Prodigy unveils a model of mathematical ability that brings together seemingly isolated facts. He demonstrates that there is a single math ability gap between the sexes, biological in origin, and independent of race, culture and geography. Prodigy introduces the theory of Everyone which accounts for all available data.

No comments: