Much Ado about Inflation

Grade inflation that is.

A couple days ago, Brian Solomon '11 reprised his data-driven critique of grade inflation at Dartmouth (See Parts I, II, III, IV). I find it refreshing that he backs up his case with data (particularly data visualizations). What the figures lack, however, makes all the difference: the figures cannot show that grade inflation exists or that it should be addressed. Let me explain my reasoning.

Mr. Solomon writes that:

With the administration and faculty not paying attention, poor work has become mediocre, and mediocre has become good. That’s the academic world we live in now, and it shows us that a big first step in the grading battle would simply be to curb further inflation. We can worry about making grades more meaningful again after we’ve stopped the bleeding.

Such a claim requires a few conditions to be met that, as of yet, have not been. To argue against grade inflation, one must make the case that:

1. Grade inflation exists. This sounds obvious, but such a claim presents more challenges than one sees on the surface. To show grade inflation, one must show (a) that grades are rising, and (b) that this rise is a function of falling grade value.

Mr. Solomon shows the first clearly: median grades are rising, and quickly. But is this really inflation? To borrow from economics, grade inflation really means grade devaluation; in other words, grade inflation means that students need exchange less quality work today for an "A" than they did three decades previous. Put another way, an "A" grade purchases less quality work than it would have in the past.

The alternative to grade devaluation is grade compression. According to this hypothesis, an "A" buys the same quality -- perhaps better quality -- work as it did in the past. The rise in grades is merely a function of a rising median quality of work. Since grades are retroactively immutable and naturally bounded (i.e. grade ϵ (0,4)), what some term grade "inflation" is actually the consequence of the grade distribution being compressed near the top of the grade space. Put another way: Dartmouth's students are abler, smarter, more driven or otherwise just do better quality work, on average, than they did in the past.

2. Grade inflation is a problem. Even if one can show that the College values grades less than before, one must then make the case that this marks a bad trend. Some argue (pdf) that it does: that harsher grading metrics push students to learn more, achieve more, and that it provides more opportunity to separate good students from middling ones. Others retort (pdf) that students suffer when the grading system fosters competitive, uncooperative or unpleasant working environments.

Mr. Solomon tries letting the data speak for itself, but rising grades do not a normative case make. He does show -- quite convincingly, I believe -- that grade discrepancies between departments pose a serious problem. Given that honors, Phi Beta Kappa and many other recognitions tend to be awarded on raw GPA, not performance relative to others in your department, one need seriously question the fairness of the current grading system.

***

Unfortunately, debates over grade inflation invoke ideology over the interest of students. Grade inflation has been a cause célèbre amongst conservative revolutionaries within the academy. Conservative commentators on life at the College (see Powerline and Dartblog for examples) tend to find rising grades most troubling. And as Harvey Mansfeld showed, grade inflation often gets implicated in a constellation of other issues, like diversity, political correctness and affirmative action.

Mr. Solomon has done a great job so far, but to discuss grade inflation at Dartmouth requires more work. The purpose of this exploratory post is to invite thoughts and comments on how we can extricate grade inflation from its ideological prison and discuss what grade inflation might mean for the future of the Dartmouth experience.