Being great proponents of home schooling around here you'll never guess what the result was: schooling exacerbates differences. (I can't figure out if girls improve objectively/absolutely; abstract of paper doesn't say): alternative summary = if you want your girl(s) to fall behind in math, TAKE THEM TO PUBLIC SCHOOL!

It'll be interesting to confirm that assumption by comparing the graphs for kids that have been homeschooled for one year. Is it the school environment, or is the fact of learning math for one year (be it at home, or be it in a traditional school) that causes such divergence? Probably harder to find "equivalent" groups of kids where the only difference is the one of being homeschooled, but who knows.

Also lovely to see that initial graph, almost a validation of Greater Male Variability hypothesis right there.

EDIT: Never heard of this hypothesis, the Wikipedia article seems to be fairly objective on a topic that could be targeted by social justice warriors.