I posted this a while ago but it got lost when the server crashed. Since I find myself wanting to refer to it in conversation, I'm rewriting my conception of the social hierarchy. This ordered list represents who's higher than whom in the social pecking order, in descending order of power, based on common understandings of success and desirability.
1. Top 1% of men.
2. Top 10% of women.
3. Next 9% of men.
4. Next 40% of women.
5. Bottom 90% of men.
6. Bottom 50% of women.
If you consider this representation I think you'll find it interesting, even if you want to quibble about the numbers. Note that the least desirable half of women are below all men, but the most desirable half of women are above most men. Other divisions are possible, with finer gradations, but this conveys my intentions well enough.
A distribution similar to this shows up in my math course grades. In a class of 40, my top 10 scorers are usually about 50-50 males and females, and my bottom 10 are 90-10 males and females. This holds true even when the class as a whole is 75-25 males and females (in that case, the top 10 are usually 60-40 instead of 50-50). Women in my classes have an much higher pass rate than men. And if there is a female re-entry student in the class (usually an older single mom), odds are good she's the top student.
I thought maybe this was due to my being a woman and my teaching style somehow resonates better with women, but now I'm not so sure.
WH: That's a really interesting application I hadn't considered. If you want to post any hard data (actual grade breakdowns by gender, with a significant sample size) I'd be very curious.