I think the challenge is how to correct for unconscious bias (e.g. the classic NBER finding that "black" names have to send in 150% of the resumes of "white" na

I think the challenge is how to correct for unconscious bias (e.g. the classic NBER finding that "black" names have to send in 150% of the resumes of "white" names to get the same interviews[0]) without over-correcting. When people think they're making objective, merit-based judgements, they are actually being racist, sexist, classist, etc. So how do you fix this? One way is what you suggested: "companies ... should be completely blind to their workforce's background. This is doable in the interview process; for example a test which results in an offer if passed. (Though normal interview processes are not like this.) But recruiting is more difficult, because right it involves humans reaching out to other humans who they judge to be a good fit. So being truly merit-based there requires some correction of one's own intuition. I mean, if recruiters were really instructed to "purge the hiring pool" of non-diverse candidates then that's obviously pretty blatant. But I just wanted to note that being 100% merit-based is not easy. [0] http://www.nber.org/digest/sep03/w9873.html http://www.nber.org/digest/sep03/w9873.html