The solution to this? More data. Stipulate that all packets include statistics on each interviewer’s past scores—including number of interviews, range of scores, and mean—so committee members can factor into their decision-making which interviewers grade higher and which clump their scores in the middle of the bell curve.
