> I’d be in favor of rigorous research on this topic. There's a lot of pretty high quality research. Unfortunately, it can only tell you about things that are

> I’d be in favor of rigorous research on this topic. There's a lot of pretty high quality research. Unfortunately, it can only tell you about things that are correlated, but it can't establish causation. You can only run observational trials of various kinds-- case control studies, looking at time series data, etc. > My own experience in the tech industry at many companies leads me to believe that if there is discrimination, it must be either historical or occurring earlier in the pipeline than tech company hiring. OK-- and if there's some 25 year old who has faced discrimination "earlier in the pipeline"-- what is the proper redress? If candidate A measures 90% as good as candidate B, but candidate B had a much easier road to get to this point, which candidate has more potential? I'm a teacher and I see it "earlier in the pipeline". By middle school, a whole bunch of kids have figured out that typical engineers don't look like them, and they should maybe consider doing something else-- irrespective of class. Some of my highest performers have been in these groups. > with the overall effect of selecting for class above all else. I don't think this is very true of most interventions. But it's true that it's true to some extent for some.