Interviews aren't a perfect selector. They have a huge random element. So if minorities consistently get more interview opportunities with the per-interview var

Interviews aren't a perfect selector. They have a huge random element. So if minorities consistently get more interview opportunities with the per-interview variance and per-employer bar variance they are more likely to have an interview where they do better than the bar. This means that over-interviewed minorities will be hired even at a lower "quality". Basically for both groups there is the top quality which will always be hired and the bottom quality which will never be hired. However the difference appears in the middle group that will eventually be hired given enough interviews. For minorities with more interviews a higher portion of this group will be hired. The more over-interviewing and over-hiring you do this group will get more and more drained. But there are still members on the lower end that will eventually be hired given enough opportunity.