Why Grade Fairly?

When I was a graduate student teaching assistant at Princeton, we would occasionally have grading sessions where the TAs for a class would meet and mutually grade papers, so that we could make sure we were grading our students by the same standards. There was inevitably a fair amount of divergence – Some people would think that a paper merited a B+ where others thought it deserved a B- – but by and large, by the end, we would be able to grade papers with a surprising amount of consensus.

I teach several sections of Introduction to Ethics here at UNO. There are other professors and lecturers who a few sections as well. We don't follow the same syllabi and we don't make sure that we assign the same amount of work or grade papers the same way. I would be surprised if we all happened to apply the same standards or give out the same grade distributions. I know that the professors in the department view grades differently. UNO has an official interpretation for what different grades mean, but those interpretations are themselves vague and open to interpretation.

I don't think that this is especially problematic. While we could spend effort to make sure that we grade to the same standards (either in different sections of the same class or different classes of the same level). We could have departmental grade calibration sessions or come up with a precise rubric for grading. I don't think that we need to. If we make little to no effort to apply the same standards in different classes, we haven't done anything morally wrong. (I tend to think, also, that students benefit from being subject to different standards in different classes.)

But suppose that I decided to use different standards for my sections. Or suppose that I decided to apply different standards to students within the same section. Would that be wrong?

Let's consider a concrete case to illustrate it. Suppose that I randomly divided up the students into two groups, one group which I graded leniently and another group which I graded stringently. There are multiple possible justifications for doing this. For isntance, I might want to assess how grading standards influence learning, and what better way than a randomized trial?

If it is not inherently wrong for students to be graded by different standards, why should it be wrong for me to grade my students by different standards? Intuitively, to me, it would definitely be wrong. But why?

Here are some possibilities. (Remember, I mean to consider only those possibilities that justify thinking there is a difference between multiple professors doing it, and one professor doing it.)

1. It is wrong because it involves unfairness when fairness would be effortless. It would take effort for different professors to calibrate their grades collectively. It takes no effort for one professor to calibrate his or her grades.

The problems with actually calibrating grades is significant. It is comparatively easy to calibrate grades across sections of one class at one university. But it is somewhat strange to demand this without demanding that the same class is calibrated across universities, or that different classes (at the same level) are calibrated within the same university. Otherwise, you're not exactly fixing the issue of fairness, you're just producing a very small island of fairness.

It is not clear why the same can't be said of a single professor's grading standards. Suppose three professors each teach two sections and each have different grading standards. Professor A grades both her sections easily. Professor B grades one of his sections stringently and one leniently. Professor C grades all her sections stringently. Would it be more or less unfair for B to start grading both sections easily, stringently, or somewhere in between? I don't think it would make a difference, overall, in total fairness. So I am dubious that it is wrong to me to be unfair in my grading because of the unfairness produced.

Second, the effort involved in changing grades isn't necessarily nothing. Calibrating the assignments of grades across 50 papers can be quite taxing. Especially if one wants to do it properly. It is not at all obvious that basic steps towards increasing fairness across sections wouldn't be easier to accomplish than actually calibrating one's own grades.

2. It is wrong because different grading schemes are merited by different teaching philosophy. One professor can't have two teaching philosophy, so one professor can't justifiably assign different grades.

Different grading schemes are warranted by different teaching philosophy, but that doesn't mean that any given professor must be fully committed to a single teaching philosophy. It seems fair for a professor to try out different teaching visions across different semesters. Why not within the same semester?

3. It is wrong because it would be immoral for me to grade students in any way but the way that I think is warranted by the work they produced. It may be that multiple professors think different students merit different grades, but one professor cannot think that the same paper warrants different grades when coming from different students.

The problem with this response is that it requires, I think, too great of a commitment to grade objectivity. I don't think that a 'B' means anything in particular about a paper. I think this even though I assign B's. I have a grading standard that I use, but I use it only because it is practical. I am accustomed to using it, and I don't see any good reason to be either more lenient or more stringent. I think that it encourages students to put in the effort that I want them to, but I am not confident that some alternative might be superior. So I don't think that it would be totally irrational or inconsistent for me to adopt different grading standards when grading different papers. It wouldn't require me to grade in any way that I reflectively regard as inappropriate. There is no absolute fact about the grade a paper deserves.

4. It is wrong because it violates a tacit agreement I make with my students. Part of the student-teacher relationship involves applying a single standard to all students within the same class. If I favor some students over others, I violate part of that relationship.

This forms a plausible explanation in most of the common cases of grading. My students expect me to grade them fairly across sections, and so I should. Suppose, however, that I gave students the following option at the start of the first day of class: if you all agree, I will randomly distribute you into two groups – I will grade the first group leniently and the second group stringently. Suppose that all of my students agreed. Would it then be right? Intuitvely, I think not. There would still be something wrong about it.

I am not sure what justifies, or if it is possible to justify, all of the intuitions I have about grading and the demands of fairness. Possibly, the demands of fairness are less than I think, and it would be ok to grade students differently, so long as it was done in a suitably random fashion. It is possible that when one professor grades unfairly, the unfairness is just easier to recognize, and so it is more psychologically salient. Or possibly, we should do more to calibrate grades across classes on a large scale. Or possibly, I just haven't hit upon the true justification yet.