Their evidence for lack of bias in their process is pretty weak: "Men and women are invited to interview, advanced to a second interview, and admitted at the same rates."<p>All this means is that the person doing interview #1, interview #2 and admission decision are biased at the same rate. It doesn't mean that rate is zero.<p>The proper test - compare admission rates against a truly objective and unbiased measure like a final exam (particularly if students anonymize their names). If gender information does not leak to the grader, bias is impossible.<p>[edit: Wow, I've clearly made a blatant failure of reasoning - apparently my knowledge of statistics is far less than I thought. I hope someone can explain my error - I'm not afraid of math, so don't skip the details.]