The core of what he is asking, and failed to drilled into sufficiently in my view, is exactly what the OP says: what % of the gender gap in various fields can b

The core of what he is asking, and failed to drilled into sufficiently in my view, is exactly what the OP says: what % of the gender gap in various fields can be explained (within that field specifically) due to demographically correlated population preferences vs other factors. I think there are two unforced errors: - Trying to tie things back to causation due to biological factors gave cherry pickers sufficient ammo to snipe him as some kind of supremacist, sexist, bigot, whatever. Explaining causation was unnecessary for his argument. All that was needed was highlighting statistical correlations measured on demographics which could influence career choice. The question then reduces down to "do these correlations actually influence career choice or not" not anything about causation. From there, it's easy to ask "if so, how much? Zero is an acceptable answer." - He should have made a good faith argument of what the world looks like if you assume these demographically correlated attributes have no effect. It was important that he really highlight other potential explanations, since to people who disagree with his premise fundamentally, "extraordinary claims require extraordinary evidence." Once illustrated, leave it to readers to decide on the absurdity or validity of that counter example. Some would consider it a sufficient "proof" by counter example to make his point, others would not be convinced but at least could digest his argument more likely as one of good faith not bigotry. In general, a lack of empathy for the lens through which his biggest, most vocal critic will read his argument was his undoing. The best path to having this discussion (regardless of merit) would be to continually buttress the argument on all sides with qualifiers and signs of good faith to those who will read it in the worst possible light. He did some of that, but not nearly enough.