So what's the data? Why even bother discussing (this specific case) without that? Assume the ideal of simmilar pay for simmilar positions: Here google might e

So what's the data? Why even bother discussing (this specific case) without that? Assume the ideal of simmilar pay for simmilar positions: Here google might employ 3 top tier female programers and 7 top tier male programmers with equally high pay, while the rest of the workforce is split even, also with simmilar pay. So they might have sourced from the top tier of progammers where there might be a gender imbalance and it would cause the average pay of men to rise. On the other hand discrimination might have a far bigger impact on which roles equaly qualified men or women get in the first place and how many promotions they get. I also wonder how broad job "roles" are defined. You'd need that bo be quite detailed to compare equal roles and avoid the scenario above.