Academic

Two years ago we discussed the new pass-fail grading system at Harvard Law School. That didn't last long:

There are three main changes to the law school's grading scheme. The recommended grade distribution for large classes, which was published last year for the first time in the law school's history, will no longer be publicly available. It will only be disclosed by the administration to faculty. Second, the Handbook previously allowed up to two Dean's Scholar Prizes to be awarded per class. It no longer contains any recommendation for how many should be awarded. Finally, but perhaps most importantly for students heading to interviews, the Law School has adopted a new grade point average formula for calculating Latin Honors that more closely resembles a traditional 4.0 GPA scale. It awards a student five points for each Dean's Scholar Prize credit, four for each Honors credit, three for per Pass credit, two for a Low Pass credit, and zero for a Failing grade.

The folks at ATL are claiming a conspiracy to hide the change from the students -- and that has its own interesting features -- but as to the merits of the new system, we see again the eternal tension between the desire to rank student performances and the desire to avoid doing that ranking with too much precision.

Debates about grading systems always end up at a single question: how many levels of differentiation do we want?

Yale ends up with two levels, for all intents and purposes, and that is too few for most schools. A system with more than a handful of levels seems like too many to me, so I like the impulse at Harvard to fix the levels at four or five. But this relatively limited number of options will tug at those who want more nuance for those students at the margin -- "this student isn't quite as good as the Honors students, but is better than the Pass students" -- which is why we end up with A- and B+ and so forth.

To me the most interesting feature of grading systems is that they always seem to be in flux, either as to the macro elements (e.g., letters v. numbers) or the micro elements (e.g., fixing the median or the recommended curve). I have just come to terms with the fact that students or faculty want to revisit grading discussions on a regular basis, but I am happy to let others sweat the details.