I agree, but after some time I started wondering: What happens if the employers measurement of quality (height) has an uncertainty? If the employer has two candidates very close together, using Bayesian statistics, the employer must conclude that the male (for height) is more likely to be better, because there are more men that are this tall than women. I think this is "fair" for the employer to do, as their goal is to hire the best person. But obviously it's unfair to the tall woman.