Reviving the Meaning of Grades

As Harvard College debates the magnitude of grade inflation and considers competing solutions to the problem, it must not compromise valuable educational goals looking for a quick fix. Top-quality work deserves to receive high grades, but the talent of Harvard’s student body does not dictate that the University withhold grades below B for only the most unusual circumstances.

Although a successful approach to grade inflation will certainly result in a reduction in the average student GPA, that reduction must not itself be the goal. Instead, Harvard should aim to reform its grading policies so that grades better serve their original intent; a shift in the overall grade distribution will come as a corollary of that process.

Harvard being an educational institution, the crucial role grades play in teaching must come before any role they play for outsiders. Because grades differentiate student work based on its quality, and because students produce different quality work, grades will inevitably differentiate between students. To that extent, outsiders can rightfully use grades to compare Harvard graduates with each other and with other graduates nationwide. But this effect should be an incidental by-product, not a central part, of any reform efforts.

In their educational function, grades are, most basically, a means of feedback. Whether we speak of grades received on individual assignments or in courses on the whole, they provide a measure of performance that, together with detailed comments and criticisms, point students toward future improvement. High grades reinforce strengths, and low grades emphasize the need to improve. The motivational value of grades is encompassed in this feedback purpose, because good work is rewarded and mediocre work is honestly appraised.

Unfortunately, current grading practices fail miserably at providing meaningful feedback on student work, and as a result, grades as pedagogical tools are ineffective at best and useless at worst. Because 86 percent of grades were B or better last year, professors and TFs were effectively limited to four grades—B, B-plus, A-minus and A—when evaluating all but the weakest student work. Not only are there too few grades to be precise, but worse yet, the meaning of those grades is utterly unclear.

Even if the difference between grades were clearly understood, grades provide little feedback to students if they separate quality into only four categories—especially if those categories are at the very top of the grade scale. A student might earn a B-plus on three different papers for the same course, and although each paper may be significantly better than the previous, the grading scale is still too coarse to indicate that difference in quality. With respect to motivation, grades are of little help if a rushed paper—composed hastily and with little preparation—will earn a B, while hours of additional effort may only bring it up to a B-plus. Instructors must feel comfortable awarding grades in the B and C ranges if grades are to have any precision in their feedback. Harvard students are extremely talented and adept in many subjects, but not every student can excel in every area. There is significant variation in their quality of work, more than instructors can accurately reflect with only four grades.

No matter how many grades instructors have at their disposal, those grades are useless to students if no one agrees on their meaning. Individual TFs, professors and departments are forced to guess at the proper meaning of inflated grades. One TF’s A-minus may be another TF’s A or B-plus. In lieu of clear standards across the Faculty and even across TFs in the same course, interpreting a grade can be as challenging as earning it in the first place.

Although creating some standard meaning for grades is a challenging process, it is essential that we do so using quality-based grading instead of relative grading based on a curve. Curves only describe student performance relative to the rest of their class, and so students must guess about the performance of their classmates in order to determine the overall quality of their work thereby destroying the pedagogical purpose of grades.

Moreover, by introducing such overt competition, curves discourage students from taking intellectual risks. If all students do extraordinary, innovative and original work, they all deserve As. Math 55 is often cited by Dean of the College Harry R. Lewis ’68 as an example of why a curve makes no sense. Most students in that class have competed in math competitions at the international level and have extensive college-level math experience. By any standards, nearly all produce “exceptional” work, and many justly receive As.

Professors and well-trained TFs are better judges of quality than statistics. Even if curves in some large courses happen to match the grades that would have been given on an individual basis, a little extra effort on the part of professors and TFs is a small price to pay for the confidence that grades correspond to quality, not arbitrary quotas.

Of course, an A in an introductory course would not mean the same thing as an A in an advanced course. Although both grades would demonstrate mastery of the material and critical thinking about the topics, the introductory course would work with more basic material and analysis. And not all departments would be expected to end up with the same grade distributions. Some, like music and social studies, self-select because they have higher requirements for their concentrators.

Creating and enforcing such broad standards of quality is a difficult process about which we will present our suggestions in more detail on Friday. Letter grades must have a standard translation in terms of quality of work and thought so instructors are able to give meaningful grades that students understand. Such standards would revive grading as a useful educational tool.