A senior scholar who has taught at a large number of top law schools writes:
USC's grading system was the best (when I was there, is it still in use?). Lots of grades is a lot better than just a few. Don't know why people don't see this. If the grades you have are A, B, and C, the students at the B/C cut-off will always be arguing post-mortem, and, more important, if you make a mistake it is a mistake that matters. Whereas, the mistakes with smaller grades tend to wash out. (If I mistakenly give someone at USC a 78 instead of a 79, that is not going to hurt so much.) It doesn't matter much that there is no absolute objective difference between a 78 and a 79, because what one does is rank all the papers in a pile and then parcel out the grades, which is what one does with fewer grades too, but then one must worry a lot more about the margins.
The situation is worst if you have only grades that amount to A and Not-A. The great bunch of Not-A's does not have much incentive to become more accurate and well-prepared (certainly saw this at Berkeley, though I loved their students). The top Not-A doesn't get any more recognition than the bottom Not-A. Employers and judges have to take to calling up professors and asking who is "really" good. That incentivizes students to be kiss-ups. Those who don't play this game will suffer. My impression has been that those who don't play this game have been disproportionately women and minorities, at least in the past. I thought this was happening for a long time at Yale; those who were touted by Owen [Fiss] and Bruce [Ackerman] could get better clerkships and launching into teaching jobs than those who weren't; meant that Owen and Bruce had a lot of would-be student friends.




To be worth using, a detector needs not only (A) not get very many false positives, but also (B) get…