"They also found that participating firms attached less importance to interrupted work histories or other negative signals for minority candidates. This suggests that anonymizing resumes may have made it more difficult for employers that were originally more favorable to minority candidates, to identify these candidates, and to assess their application in light of the adverse employment conditions many of them face."

That makes sense. It's sort of like how I wouldn't expect someone from Kentucky to have the same internship opportunities as someone from the bay area, but if I cant see their location then all I have to go off of is the listed experience.