Atrios notes that one of his commenters has pointed out that the real driver behind the college gender gap is the larger discrepancy among nonwhite students. Unfortunately, that doesn't make very much difference. If you look at the absolute numbers - 698,000 female graduates versus 529,000 male grads - the assumption that the number of white female grads is about the same as the number of white male grads leads to the conclusion that white females are outnumbered by a 2-1 margin by nonwhite females in universities. White males would thus be outnumbered by about a 9:7 ratio in this scenario. Obviously, that ain't so.

A more reasonable assumption is that the white to nonwhite ratio at universities is about 3:1 or 4:1. Under those conditions, the gender gap for whites is about 5:4, or 55.5% to 45.5%. The greater the ratio between white and nonwhite, the closer the white gender gap will come to the overall 57:43 ratio.

(I'm skipping the math because it basically boils down to the kind of related-rate word problems that most normal people hated in high school algebra. And it's late and I did a bunch of back-of-the-envelope calculating in my head. If you insist on making me show my work, let me know in the comments.)

The bottom line is that the larger nonwhite gender gaps have a small overall effect. A better line of inquiry is the one that Fritz raised, which is the effect of the larger incarceration rate among men. If, for example, the ratio of college age black women who are not in jail to college age black men who are not in jail is 60:40, then the gender gap for blacks is merely reflective of the general population. I don't have the time or the gumption to look into this right now, but maybe I will later.