OK, I'll bite. Sides of this issue you might not be considering

I'd also like to suggest a good search term for discussions about this: "leaky pipe". The analogy is that the process of getting women into open source is like a pipe with leaky points all along the way. All of the leaks need to be stopped up, and stopping any particular leak (whether it is early childhood influences or sexism at tech conferences) will help more deliver more women to the end of the pipe.

Nobody's saying that there aren't other leaks. There absolutely are. But the ones that the open source community can best address are the ones that are specific to the open source community.

If you are interested in eg. encouraging girls in STEM (science/tech/eng/math) education at early ages, there are many other organisatinos working on that. Many of them take donations, or would welcome your volunteer time.

> 1) Studies which show a lack of difference tend to not get published.

Just to expand for the benefit of those without a lot of experience in social science: this has two unfortunate consequences.

First, social scientists constantly have to ask themselves whether some pattern they observe in their data is a result of some underlying principle, or just a coincidence. (If you flip a coin 10 times, and got 8 heads, is that because the coin is unfair? If you had 10 men and 10 women perform some task, and 8 of the women did better than the average man, is that because women in general are better at your task?) That's what statistical testing is for. Unfortunately, statistical tests are never perfect -- they won't tell you that getting 10 heads in a row means your coin is unfair, just that if not then that's one *heck* of a coincidence.

But if you keep trying long enough, then eventually you'll get that coincidence. And "science says women are <...>!" gets press, so lots of the time, when someone's running some random study, they'll do a quick check for gender effects, just in case. If 20 people do this, then 19 of them will get nothing, shrug, and forget about it; 1 of them will flip 10 heads in a row and publish a really excited paper! They don't know they're the 20th person to try, after all. (And that's leaving out the effects of confirmation bias, etc.)

Second, once a claim like that is out there in the literature, it's hard to disprove; if you just repeat the study and don't see a difference, then maybe you just did it wrong or something -- it's hard to get that published. (And even if you do, it's not as exciting, so it won't get press coverage, so a heap of people will go on believing that they Know Something About Men and Women that's just wrong.)

The end result is that the literature on gender differences has heaps of confusing nonsense in it. There are real gender differences too, but they're hard to pin down, and after all that nonsense it's hard to imagine that people would have *missed* anything so dramatic as to cause 98.5%/1.5% differences in participation a specific field invented in the last 30 years. Seriously, that'd be Nobel-worthy.

This isn't my area of specialty, but AFAICT, whether you're right or left handed has more of an effect on your general cognition than what you keep in your pants (and your culture matters a lot more than either).