Well I'm no math genius, but the story that there was no math gender gap seemed strange to me, because it doesn't square with the evidence I've observed. I thought it was perhaps that boys had been horribly dumbed down and worn down by the feminized outcome-based education (math "journals", etc.) which rules today. But now we have confirmation the story is wrong. Canada's National Post, "Larry Summers' Revenge":

Unfortunately, journalists of both sexes tend to not be math geniuses. Few of them anywhere on the continent noticed that Ms. Hyde's data actually come a lot closer to supporting Mr. Summers' hypothesis than they do to refuting it.

The study certainly does confirm that there is probably no difference between males and females in math ability, on average. This means that, if one were to plot out the observed mathematical proficiency of a large number of male and female individuals, the resulting graphical pattern would produce two sex-segregated bell curves centred on roughly the same average point.

But that doesn't mean the two profiles would be identical. Decades worth of data show that male populations exhibit greater variance in their observed mathematical ability (and their intelligence more generally, for that matter). This means that men exhibit "fatter tails" on their bell curve, with more statistical outliers in the far-flung domains of genius on one side, and total dullard on the other. In the case of women, on the other hand, typical psychometric findings show their abilities to be clustered more tightly around the mean.