People are complex, as are the social structures they build (including but not limited to educational institutions, workplaces, and professional communities).
Accordingly, the appropriate causal stories to account for the behaviors and choices of humans, individually and collectively, are bound to be complex. It will hardly ever be the case that there is a single cause doing all the work.
However, there are times when people seem to lose the thread when they spin their causal stories. For example:
The point of focusing on innate psychological differences is not to draw attention away from anti-female discrimination. The research clearly shows that such discrimination exists—among other things, women seem to be paid less for equal work. Nor does it imply that the sexes have nothing in common. Quite frankly, the opposite is true. Nor does it imply that women—or men—are blameworthy for their attributes.
Rather, the point is that anti-female discrimination isn’t the only cause of the gender gap. As we learn more about sex differences, we’ve built better theories to explain the non-identical distribution of the sexes among the sciences. Science is always tentative, but the latest research suggests that discrimination has a weaker impact than people might think, and that innate sex differences explain quite a lot.
What I'm seeing here is a claim that amounts to "there would still be a gender gap in the sciences even if we eliminated anti-female discrimination" -- in other words, that the causal powers of innate sex differences would be enough to create a gender gap.
To this claim, I would like to suggest:
1. that there is absolutely no reason not to work to eliminate anti-female discrimination; whether or not there are other causes that are harder to change, such discrimination seems like something we can change, and it has negative effects on those subject to it;
2. that is is an empirical question whether, in the absence of anti-female discrimination, there would still be a gender gap in the sciences; given the complexity of humans and their social structures, controlled studies in psychology are models of real life that abstract away lots of details*, and when the rubber hits the road in the real phenomena we are modeling, things may play out differently.
Let's settle the question of how much anti-female discrimination matters by getting rid of it.
* This is not a special problem for psychology. All controlled experiments are abstracting away details. That's what controlling variables is all about.