New study on gender gap in mathematics/reading

[durt]

Dammit I want to punch you in the face, yenlee.

[MysticTerminator]

dude, chill

why would you even type something like that

it could not possibly serve any purpose

[JBL]

durt is a troll -- "not serving any purpose" is its purpose. Ignore it and hope it goes away.

[orl]

[luthien]

I recommend the article "Cross-Cultural Analysis of Students with Exceptional Talent in Mathematical Problem Solving," by Titu Andreescu, Joseph A. Gallian, Jonathan M. Kane, and Janet E. Mertz, which appeared in Notices of the American Mathematical Society, Vol. 55, No.10, pages 1248-1260, November 2008.

The participation by girls in the IMO is addressed in some detail in this article, with interesting comparisons across different countries. The authors particularly contrast Slovakia vs. the Czech Republic: from 1998-2008, Slovakia had 10 girls out of 66 on the IMO team, while the Czech Republic had 2 out of 66. (These two countries were formerly united.) In the same time frame, South Korea had 7 girls out of 66 IMO team members, while Japan had 0 out of 66. Up to 1990 (pre-reunification), East Germany had 15 girls on the IMO teams out of 202, while West Germany had 0 out of 84. The authors comment: "Chi-square analysis indicated that it is virtually impossible that this observed country-to-country fluctuation was caused by random chance."
The p value is less than 0.0001.

The article makes evident the role of cultural influences on the participation by girls. Personally, I would not actually anticipate equality of interest or success (I'm female), but I think we do not yet know what the "natural" ratio of girls to boys would be. As long as the % of female IMO participants is rising in the U.S., I'd suspect we have not yet reached the "natural" ratio.

[Kalle]

I don't think anyone was doubting that (except durst). Since qualification for IMO is by an (I believe) unbiased competition, I think it illustrates the problem - "recruitment" of women at the lower levels. yenlee seems to have argued a few months ago that discrimination in recruitment at the top levels is part of the issue, but I want to believe it is not the case for academia.

[Altheman]

There may be small differences in learning ability of men and women, but I don't think that our culture has developed by chance. That is to say, I think that men take dominant roles in society more naturally than women because men are naturally physically stronger.

So one of the advantages that men have is that they create a society that is not advantageous for women.

Another hypothesis that I have is that women are more inclined to be side-tracked by child birth than men are. Of course this is true in our society, but I think it is something that is inherent in our biology.


So where does this leave us? I think that people can take the liberal point of view but it will always remain that people want to be able to claim superiority based on a group that they belong to. Many men will always be biased in the sense that they'd rather think of women as inferior because by comparison, it would make men superior. Likewise, women will want to exclaim their opposition to the male dominance.

I think that people are more likely to use group identification as a means of gaining self-esteem rather than bolstering oneself by claiming to fight the evils of society.

But I think things will generally settle into the equilibrium where men take more dominant roles because of the clear biological differences in physical strength.

So are there significant gender differences in absolute intelligence? probably not. But will a human society believe in such a difference? probably.

[jmerry]

Here's another new study (PDF link), along with a story summarizing it. This study is focused on the contest data, mostly the AMC.

[JBL]

Altheman, it sounds to me like you're trying to extrapolate from your personal experience to human universals in a way that's unlikely to be valid. The United States is only middling when it comes to political participation by women, for example: even Pakistan, of all places, does better. (link) I think, for example, that one lesson to be learned from Sweden and other Nordic countries is that you can (largely through modest changes to public policy) create a society in which men do their fair share of child-care even while birth rates remain relatively high. It's not at all clear why, in light of examples like this, we should assume that the gender distribution in various different kinds of employment (e.g., politics or mathematics) is at some biologically-determined equilibrium.

[Kent Merryfield]

In reaction to the link jmerry just posted (or at least the summary article, which is all that I read):

This squares with an anecdotal observation that I made some years ago. The top female math competitors have tended to come from the places with the largest, strongest and best organized school-based competition groups. Many male competitors also came from those places - they were, after all the strongest and best organized schools - but some other top male competitors could pop up almost anywhere. For a Southern California example of this, Lauren Williams and her younger sisters all came through the Palos Verdes schools, and at that time the Palos Verdes schools were by far the strongest and best organized for contests of the region.

Every rule has its exceptions, of course. Wasn't Sherry Gong already well-established as a top competitor before she transfered to Exeter? I don't think there are any top-handful schools in Puerto Rico.

But what I wonder now is whether this whole observation is outdated. Have we moved into a time in which the schools are less important than before? That AoPS and other on-line sources, and competition-oriented weekend schools and tutoring programs, and summer camps, and all of that non-school-based contest support structure is now so important that even going to a top school isn't enough any more? And if schools and their programs are relatively less important, who is going to produce the next generation of female math competitors? Can the non-school programs do it?

[jmerry]

Virtually all of the data used in that article is from 2007 or 2008. In fact, one of the main justifications for that assertion is the CGMO team composition- see my first-page quote:

[jmerry]

As far as I can tell, that table I objected to in the last post is comparing students in both 2007 and 2008 to 2008 statistics excluding those students, being somewhat inconsistent in how that exclusion is handled, and comparing to a rather unhelpful sample of all AMC schools.
Here's my version of the table, strictly looking at USAMO qualification numbers:
Table (very big)

Interpretation: A line such as "43-61" in the percentile column means that 43% of USAMO qualifiers (excluding homeschoolers and non-USA residents) went to schools with strictly fewer USAMO qualifiers than that student did, and 61% went to schools with that many or fewer.

If I average those percentile numbers, I can create a "cluster index", in which a random sample of USAMO qualifiers would score about 50. Cluster index by year:

It's a dramatic result; the boys on the IMO teams are distributed almost exactly like USAMO qualifiers overall, while the girls on the CGMO teams are heavily concentrated in schools with lots of qualifiers.

[Kalle]

[JBL]

Modestly different regulatory regimes have substantially different outcomes, ergo human nature is not a stronger constraint on division of labor than some rather small details of public policy. In other words, it demonstrates that the American division of child-tending labor is not the only one consistent with human nature and therefore disproves the notion that the current division of labor is the "natural" one.