Grade inflation

What gets called grade inflation is actually a change in the scale we use for evaluation (Brandon has made this general point many times).

If you look at the traditional descriptions of the letter grades, it’s clear that they are describing a bell curve with a two standard deviations and a C as its apex. The 70-79% range and the decades on either side of it were doing exactly the same work as the IQ test does with its 100 average and standard deviations of 15. On this model of grading, the goal was to maximize students in the 70 range, and to have about as many A’s in the class as students who flunk out. The B and D grades would inform you about some acceptable deviation from what was most common.

But no one takes the grades as meaning this now. Teachers expect the student to acquire a certain amount of information, and “A” means to acquire it perfectly while “F” means a complete failure to acquire any of it. The B and C/D grades serve to indicate which of these poles you’re closest too, with the C/D together serving as a broad warning zone that one is approaching failure.

Clearly, the letters have stayed the same while taking on completely incomparable meanings. We haven’t inflated the grades, we’ve utterly changed what they mean. No teacher I know ever tries to get Gaussian results, though they are just as easy to get as any other result (it’s no harder to correct or construct a test with a bell-curve paradigm than the one I have now.) Part of the problem is that we’ve lost the sense that bell-curves should always be descriptive of achievement; but a more salient reason for the change is that we’ve started to see education as more the imparting of information than as a contest to identify those who can best internalize and manipulate that information. If you want a bell-curve, you have to do more than just ask the students to get the information you are telling them, and we tend to see this as unnecessary and even unfair. We view an “A” as something that anyone should be able to achieve, which requires that the path to it should be laid out in advance. The bell-curve, on the other hand, sees “A’s” as an exercise of genius and uncommon insight, that is, as found down a path that relatively few can find for themselves.

The new system is more egalitarian, but also more moralistic: since an A is defined as open to everybody and requiring no genius, we view each letter grade that falls away from it as indicating more and more laziness and failure to act. The new system thus hits the student twice: he is not only more and more ignorant as he falls away from the A but also more and more lazy, shiftless, willful etc. The older system lacks this overbearing moral character, but it is certainly far more elitist. There is no a priori answer to whether a system should target egalitarian moralism or elite deviations from a bell-curve.

Seen in this way, much of the scandal of grade inflation collapses into tautology: if “A” means complete achievement of rules set down in advance, then of course the average grade at an ivy-league school should be an A. The Ivy league selects for just the sort of person who habitually does this. We should expect A’s in general to be more common now since ability to follow a pre-set path is far more common than genius. Again, we can lament that “B is the new C” or even that A- is the new C, but all this only tells us about comparable relative numbers that are made by completely different standards. It’s not an inflated currency (where a dime is now a dollar) but a changed standard of value (where a dime changes to happiness).

All this doesn’t mean that we’re at a standoff – there are arguments to be made both ways about the merits of both systems. Just what standard you use will depend – indeed is already depending – on what you think education is. If it is largely or entirely about conveying information and measuring how well persons do at following a trail marked out in advance, then our present grading system is a better metric. If you think education is about identifying those who have most penetrated into information so as to creatively manipulate it and understand its ramifications beyond any path that can be laid out in advance, then the bell-curve is a better metric.


  1. Cj aka Elderofzyklons Blog said,

    January 20, 2015 at 1:22 pm

    Reblogged this on ElderofZyklon's Blog!.

  2. January 21, 2015 at 9:01 am

    This is really good. Two thoughts:

    1) It strikes me that, despite the equivocation, “grade inflation” still has some merit (no pun intended). The standard has changed qualitatively, as you note, but the fact remains that what was once a C is now an A. And—far more importantly to me—what was once an A is still an A, except that we now have no way of identifying and differentiating it.

    2) A serious flaw in the old system is precisely the reliance on the Gaussian distribution. As Nassim Taleb points out, the bell curve fits the randomness of casino games and certain physical features like height. But educational achievement is more like socioeconomic distributions: fractal, with no real limit to the possible level of performance. So, if he’s right, both grading systems are inadequate measures of the actual reality of students’ capabilities and performance. There may be some extent to which this is unavoidable, but I wonder if we can’t yet build a better mousetrap….

  3. Aron Wall said,

    January 28, 2015 at 9:55 am

    A student who is brilliant in 5 subjects but lazy or lousy in 1 ought not to be regarded as a worse student than one who is diligent and minimally competent in all 6. Yet the later is what the modern system values. The new system produces a servile mentality, not a free one (in the old sense of “liberal arts”).

%d bloggers like this: