Harvard is overlooking two solutions to its problem of grade inflation. The first one is prohibitively expensive, and for many other reasons impossible. The second one just might work.

Personalized evaluations would allow Harvard to scrap its system of ranking students with grades. Each professor could write a couple paragraphs about each student, succinctly identifying the student’s academic strengths and weaknesses, in addition to explaining the material the student learned in the course. Comparison of the student to the rest of the students in the class would be forbidden. The upside to this system is that the evaluations would describe the student’s performance more precisely than grades, and all of this could be done in a non-competitive manner. The downside is that it would take too much effort, and far too much time.

So Harvard uses grades instead. Inherent in letter grading is the comparison of students to one another:

According to the Information for Faculty Offering Instruction in Arts and Sciences, “B+, B and B- represent work of good quality which, however, does not merit special recognition.” When there is comparison among students, then there will be competition. It is not necessarily a bad thing, but it is a fact. Many people insist that grade inflation is a fair way to get around the problem of comparing so many strong students. The quality of the work is so high across the board, they say, that everyone should receive A’s and B’s. This is the how grade inflation is justified.

The trouble with grade inflation is that it makes it impossible to standardize evaluations within and across departments. Some teachers see a B as a default grade for those who don’t receive A’s. A different teacher may distribute B’s according to a Bell curve. Not only does this penalize students who happen to take courses with tougher-grading professors, it also creates incentives for students to select courses with easier-grading instructors. And this in turn puts pressure on tougher-grading professors to inflate their grades, or else face dwindling enrollment in their courses. Grades are supposed to be an evaluation of how a student performs in class, but without standardization, grades start to affect what classes students take.

That’s where the “achievement index” comes into play. Developed by Valen E. Johnson, a professor of statistics at Duke University, the achievement index is a system of ranking students that takes into account how each student performs relative to the people in his course, and how the people in his course performed in their other courses. Johnson compares it to the rankings in college sports, which take into account the difficulty of the schedule in addition to the team’s record. The Harvard football team is a great team to watch, but no sane person would rank them above the twentieth best team in the country, Florida State University, even though Harvard’s 8-0 record looks better than FSU’s record of 6-3.

With the achievement index, a B in a course where many C’s are given out, would be worth more than the same B in a class where everyone gets an A or a B. And that B would be worth even more to students’ ranks if their classmates did very well in other classes. This way there would be no incentive to find classes with easy graders. Rather, students would be encouraged to take courses that challenge them, and then be compensated by this system for the risk involved.

At Duke this system was not instituted, due to students who feared a shake-up in grades, and professors who didn’t want to relinquish their power over the grading process. Of course this system isn’t perfect, but it is worth considering. Once we’ve accepted a system of letter grading, we might as well make it fair.