> It seems like some highly quantitative fields which have traditionally been male-dominated (CS, physics) may very well be practicing some sort of entry discrimination

Maybe I misread, but I interpreted this follow up statement as "50% is correct, here's why (CS, physics) may not be at this level".

About 20% of people in CS are women. Ordinary male variability, difference in median aptitude, and employer selectivity could explain some of the disparity. But 20% is extraordinary. And people rationalized the disparity in biology the same ways people rationalize the disparity in CS.