It's probably true that a person trying to self-correct for bias will tend to over-correct, to no one's benefit. But the sort of things you do to correct for w

It's probably true that a person trying to self-correct for bias will tend to over-correct, to no one's benefit. But the sort of things you do to correct for widespread unconscious bias are different. For instance, the New York Symphony Orchestra found that female applicants were making it through the tape screening round more often when their tapes were blinded, replacing identifying information with a serial number. Women were still being rejected at the audition stage in higher numbers than men, so they started having the applicants play from behind a screen so the judges couldn't see the gender of the audition, and once again, the number of women making it through that round increased[1]. This was all at a time when the NYSO was consciously trying to practice sexless-ness, if not actively increase the number of women in the orchestra. Several years ago, GoGaRuCo brought in blind judging of their talk applicants pool after ensuring the pool was seeded with half women. Suddenly their roster was half women too, on judging the talks purely on the merits of the proposal. We constantly see that when we successfully block stimuli to unconscious bias, the bias is disarmed. [1] Interestingly, they didn't significantly increase until they made auditions walk on stage in their socks, because the sound of their formal-wear shoes were obviously different.