Putting Grades in Their Place

Eben is correct: grades are stupid. Perhaps not everyone has yet realized this, but everyone has probably at least felt this to some degree, perhaps very recently. Evaluations, however, are not stupid. As Eben points out, thorough, frequent, sympathetic, and effective evaluations of a student's work are essential in helping the student both learn more and derive more satisfaction from learning. Eben thinks that grades not only fail to further these ends, but grades actually impede students in their attempt to achieve these ends. I agree: grades in fact often have this effect. But, this does not have to be the case. Grades can be a form of evaluation that serve a practically useful end for people other than the student, however limited that use is. As Eben points out, one cannot honestly think that grades are representative of very much about a student’s ability or that grades help students learn and grow. Yet, it does not follow that grades are inherently stupid or unjust. I contend that we allow them to do what they should not do, thereby making them coincidentally unjust. We take them to be accurate quantitative measurements of one’s intellectual abilities when, in reality, they are crude proxies for what is, at heart, an oversimplified qualitative judgment: how well a student did at a certain task. This oversimplified qualitative judgment can only serve very a limited purpose, as any oversimplification can when serving as a proxy or answer for an involved question. Grades can be a useful beginning in an admissions committee’s quest for a qualified applicant only in practically necessary contexts. Though, this use is quite limited and should never replace utilizing more informed qualitative judgments of a student’s abilities.

Eben argues that grades are inherently unjust because they “purport to reduce the evaluation of learning to a single number on a short, obscure scale.” He elaborates: “Reduction of multi-variate qualitative measurements to grades implies an algorithm implementing a model for collapsing space by throwing away information. Each turns complex surfaces captured in the evaluations…into a single numerical score, by turning a series of qualitative judgments into quantities and then operating on a vector of quantities to produce the score.” While this nicely captures the mechanics of a particular grading system, it is misleading to think that all grading systems have to operate this way in virtue of the nature of grading itself. That is, the quantification of qualitative judgments is not a necessary component of grading. What is necessary is the use of one, or both, of two kinds of qualitative judgments: this work is better than that work, and/or this work is tremendous, good, satisfactory, or unsatisfactory. These are not quantitative judgments. It is true that grading requires labeling those qualitative judgments with a symbol, but this is merely shorthand for an operational definition (such as, B = satisfactory work). This is an oversimplification of a multi-faceted evaluation, but it is not a quantitative judgment. Thinking it is a quantitative judgment entails thinking that any value judgment is quantitative, which does not seem correct.

Eben does not think that the problem with grades is their reliance on qualitative judgments because qualitative judgments are required in order for any evaluation to take place. Eben, though, thinks that the problem with grades lies in their transformation of qualitative judgments into quantitative ones. While I think this is mistaken, Eben does put his finger on what I take to be the crux of the problem. The problem is that in order to assign a label/grade the evaluator needs to jettison much of the fine-grained evaluative judgments. This potentially leads to ignoring relevant information or incorporating irrelevant information. Grades, by their nature, do force an evaluator to make difficult comparisons of work that may excel and lag in very different ways. Thus, it is inherent in the nature of grading to force one to simplify complex evaluative judgments. This is precisely why we should not take grades for being informative of very much.

Can these oversimplified evaluations serve some use? I think so, but the use is not very helpful, at least not for students. I take it that the common view of grades is that they are intended to serve as a representation of the quality of one's knowledge with respect to some set of information or skill set. We then use this representation in two ways: to serve as a guide for the students themselves in the further cultivation of these skills and knowledge, and as a guide for others to predict how the student will perform in a new situation. It is clear that grades fail miserably at the former goal because such an oversimplification cannot function as the effective and thorough feedback required for real learning to take place. For the latter goal of determining how well one will perform in a future situation, the appropriateness of grades to make this determination is dependent upon the context. The only time grades appear useful is if one is trying to sort through a multitude of students very quickly. And even if we don't do this via grades directly, we may do it by the school that one attends (which is, in turn, partly based on grades). But, grades should never completely replace substantive and complex evaluative judgments. At a maximum, grades should only serve as temporary proxies in lieu of these judgments if practical rationality demands it. While this is a rational approach in some contexts, it is not clear that admissions/hiring committees need to put themselves into this situation. Predicting how well-suited one will be at performing future tasks is not easy, and we should not let only grades guide our predictions.