[HillRat]

As a potted experiment, it really doesn’t generalize well. Weighting identical candidates except for race/gender/etc. may help uncover obvious forms of discrimination, but doesn’t address the path dependency aspects. For example, in the life sciences, women get less funding and fewer staff when founding research groups, while “elite” male researchers train 10-40% fewer women than other labs; as a result, women tend to have “worse” CVs later in their career due to earlier discrimination.