From years of internets reading when I wasn't necessarily "looking" for such information, I've found the following food for thought:
1) So one study says there's a correlation between how much dad and mom push stereotypes on their girl.
A longitudinal study out of the University of Michigan said that if a father held on to "traditional" gender stereotypes, that females were less likely to go into a math-related profession. From the website:
They found that parents provided more math-supportive environments for their sons than for their daughters, including buying more math and science toys for the boys. They also spent more time on math and science activities with their sons than with their daughters.
Davis-Kean and colleagues, including the late Janis Jacobs of Pennsylvania State University, Martha Bleeker of Mathematica Policy Research, Inc., and U-M psychologists Jacquelynne Eccles and Oksana Malanchuk, also found that parents' attitudes, particularly stereotypes they hold about whether math and science are more important for boys than for girls, have a significant effect on their children's later math achievement, and even on their eventual career choices.
They found that girls' interest in math decreases as their fathers' gender stereotypes increase, whereas boys' interest in math increases as their fathers' gender stereotypes increase.
"Fathers' gender stereotypes are very important in supporting—or in undermining—daughters' choices to pursue training in math and science," Davis-Kean said.
2) One psychology study from PNAS (irony in too many ways), said that elementary female teachers' math anxiety transfers over to their students.
Description from 80beats, Discover Magazine Blog:
The findings are the product of a year-long study on 17 first-and second-grade teachers and 52 boys and 65 girls who were their students [Science Daily]. Researchers recruited the female teachers from a Midwestern school district and assessed their level of math anxiety. They also gave math tests to 117 of these teachers’ students and jotted down their beliefs about math and gender at the beginning and end of the year. By the end of the year, the more anxious teachers were about their own math skills, the more likely their female students – but not the boys – were to agree that “boys are good at math and girls are good at reading” [AP].
3) Another psychological PNAS study found that countries with better gender equality, girls performed as well as boys on math tests.
Another description from 80beats:
They found that countries with poor gender equality, like India, had a larger gender gap in math, while in countries with excellent gender equality, like the Netherlands, girls performed as well as boys. If males really did have an innate advantage in math, the researchers note, that advantage should be obvious throughout all these cultures. Instead, the study suggests that cultural issues are the basis of the math gender gap.
4) Not a study, but teacher's experience in the culture of Mathematics:
The crucial difference of mathematics from many other walks of life is that its power games are deeply personal in the purest possible sense. To recognize someone as a fellow mathematician means to accept that she is intellectually equal (or even superior) to you and that she has the right to wear, like knight’s armor, her aura of intellectual confidence and independence. Too many men will still feel uncomfortable with that.
In that entry, she also cited and criticized this idea that women think more "collaboratively" as opposed to "independently." That made me think of why minority groups in general, whether racial, ethnic, religious, sexual, seem to always be described as more "collaborative."
5) Finally, the blog Sociological Images, run by two Sociology professors found the following and summarized some more findings:
And, since girls often outperform boys in a practical setting, math aptitude (even measured at the levels of outstanding instead of average performance) doesn’t explain sex disparities in science careers (most of which, incidentally, only require you to be pretty good at math, as opposed to wildly genius at it). In any case, scoring high in math is only loosely related to who opts for a scientific career, especially for girls. Many high scoring girls don’t go into science, and many poor scoring boys do.
In Japan, though girls perform less well than the boys, they generally outperform U.S. boys considerably. So finding that boys outperform girls within a country does not mean that boys outperform girls across all countries.