Anne Appelbaum pooh-poohs the Summers critics who think sexism is a problem:

Outside of a handful of institutions, the evidence of unthinking discrimination is slim too. It is true, of course, that men continue to earn more than women -- approximately $1 for every 75 cents that women make. But economists such as June O'Neill or Harvard's Claudia Goldin, who have accounted for different job choices, hours worked and time taken off for raising children, have concluded that it is these factors, not discrimination, that account for most of the difference.

Maybe someone who knows more about economics can help me out here, but this doesn't seem to be Goldin's position at all! She's argued that gender 'as such' is declining in significance when it comes to the workplace, true. She's also noted—in her "pollution theory of discrimination"—that discrimination against women isn't driven so much by a desire not to interact with women (as is the case with racial discrimination), but rather a fear that women could reduce the prestige of a male-dominated occupation. To wit: "Men are averse to having women enter their occupation when women's productivity is not observable and verifiable to all."

In sum, discrimination does exist, and the best way to erase its lingering and historically-based effects is to issue credentials for various occupations (so that women can "prove" that their productivity is up to snuff). At any rate, I can't find anything Goldin's written on the university—I'm not sure her work can "prove" that the evidence of unthinking discrimination is slim.

Moreover, plenty of other researchers have found that motherhood and hours worked can't account for the entire gender wage gap. So "slim" evidence? It's not like this is something people just make up for the fun of it.