What Are Grades For?

At almost all educational institutions, students get grades when they finish a course. Grades are primarily supposed to evaluate how well a student has learned the course material. They should tell other people (future instructors, potential employers, graduate and professional school admissions committees, etc.) roughly how well the student learned what they were supposed to learn in a given course. (This is one of the multiple reasons why grading on a curve is an abomination.)

As hard as it may be to assign a single letter or number to learning, most people would agree that a student who learns the material better than another student should not get a worse grade. Grades should be monotonic with respect to knowledge.

Modified from http://ars.els-cdn.com/content/image/1-s2.0-S0951832013000525-gr1.jpg

So far, the situation is simple. Grades evaluate performance. High grades are desirable because they improve your options later in life, so students typically try to get the highest ones possible, which means learning the material -- or at least the material that will be assessed. As long as instructors create reasonable assessments and students do not resort to cheating, their incentives are aligned and things more or less work out.

However, the desirability of high grades introduces a complication. As instructors, we want students to do things that will benefit their learning. We want them to come to class, do their homework, participate in discussions, and possibly more. Ideally, the desire to learn and to get good grades should be enough to get students to do these things, but sometimes they're not. Sometimes the end of the quarter seems too far away or the student is sure that they know the material and don't need to do all of those chemistry problems. So why not directly include those things as part of the course grade? Students want good grades, so if 10% of the grade is based on completing their homework, they will probably complete their homework. Incentives do work, at least in the short term.

Many, perhaps most, instructors engage in this practice to some degree. Even I do, as I was the freshman chemistry student who found it hard to do ungraded homework. (I do have a second grading scheme that excludes such things.) But there are several problems with it.

First, using grades as incentives to do specific work doesn't distinguish between the student who spends too much time partying and the one who works 20 or more hours a week and supports family members. The latter student is already at a serious disadvantage; making them do a certain amount of work, often at very specific times, no matter how much they actually know is just kicking them while they're down. Furthermore, if one class grades in-class clicker questions or reading quizzes and another one taken by the same students doesn't, students will tend to spend less time on the second one, creating an incentive for its instructor to do the same thing and leading to grading arms races.

The second, less frequently discussed, disadvantage of grading things other than assessments is compromising the monotonicity discussed earlier. The larger the proportion of students' grades based on work, the greater the chances of a student with a better work ethic or organizational skills (or more time) outscoring a student who knows the material better.


This problem could be solved by the kind of multi-category grading my middle school and high school used. For each class, you got both an academic grade and grades in "cooperation" (basically behavior) and work habits. Only the academic grade counted toward your GPA, but the others were still on your report cards and transcript.

The third, deeper problem with using grades to manage the details of student activity is that it very likely undermines the maturation process supposed to occur in college. At some point during development, a person should go from being driven primarily by short-term rewards and punishments to being able to pursue a long-term goal even in the absence of short-term incentives. For many people, this happens in college, where they grow into adulthood by managing the abundance of unstructured time higher education traditionally provides. Often, we learn this lesson the hard way, but we do learn it. However, as colleges start managing the details of what students do and when they do it as much as any high school, this unstructured time is being taken away. If young people don't learn to create their own structure by the time they finish college, when will they learn it? (The heavy reliance of many higher education reformers on external motivators is a subject I will take up in a later post.)

Teaching frequently requires us to balance competing values against each other.  While the mission creep of grades both undermines their original purpose and creates other problems, it can be effective in promoting learning. Every teacher must decide for themselves what to do about this tradeoff, but we can only do so with a clear awareness of what is gained and lost in each decision.