It's not just Mathematics. I've heard of Science professors who use English scores to sift through PhD applicants. They reason this to be an effective way to

It's not just Mathematics. I've heard of Science professors who use English scores to sift through PhD applicants. They reason this to be an effective way to find the strongest amongst otherwise equal candidates. (There is also a benefit in terms of writing papers, but writing strengths cannot make up for scientific weakness. Thus, writing is not the central point, as mathematical ability is not the central point in this econlib.org essay.) I've not seen "gold standard" (large, randomized, double-blinded, ...) studies of the effectiveness of this selection strategy. And I don't know how such studies could even be done: in research departments, it costs a lot to take on PhD students, and a lot rests on their success.