[BlueTemplar]

A significant factor seems to be that the way (not only STEM) academia is currently organized means that women have to pick between a career or having any children (at ages favorable to childbearing):

https://www.scottaaronson.com/blog/?p=4522

And I kind of have the impression that the situation here actually got worse during the last decades?

[hnracer]

I'm also talking about regular career choices (not academia) that only involve a three or four year undergraduate degree. Women self select out of technical pursuits (CS) and self select into more verbal or empathic areas (Law or Medicine, and an even more extreme example being liberal arts) even though such areas might entail even more study than CS. This self selection is magnified when women have more choice and freedom. Women in software is more common the more poor a country is.

I mean it's not only career choices where such gender differences manifest, we also see it in the choices that young children make. The burden of proof is on the camp that proposes that it's entirely sociological since the extant empirical evidence, along multiple independent lines, contradicts it.

[BlueTemplar]

Yes, I'm specifically talking of women loss after undergraduate level.

[hnracer]

Agree that this is an issue

[tasogare]

This theory doesn’t any hold scrutiny: some academic or academic-like fields are dominated by women despite having the same formal organization as the STEM ones.

[BlueTemplar]

I would guess that they were already dominated by women at the undergraduate level, perhaps even more so ?