Deflategate: How to Fix The Problem of Grades (with Information)

This article is the first installment of The Clerk’s feature and opinion series on grading at Haverford and beyond. In upcoming articles, we will explore the perspectives of Haverford students and faculty. We will also examine alternative grading methods used at peer institutions.

If the grade 3.0 is average, then there are three kinds of above average: 3.3, 3.7, and 4.0. If 3.3 is average, then there are only two kinds of above average. As grades inflate, the range of reasonable grades shrinks, lowering the accuracy of the grades which can be assigned. For this reason, I suspect that there is no one in favor of grade inflation.

Grade inflation, or at least its threat, has affected particular American colleges and universities since the 1960’s. An obvious solution to the threat of grade inflation is to deflate grades. However, because of Haverford’s small size, a unilateral move to deflate our grades would make our average graduate seem to be a grade worse than he or she is.

The real problem is that grade deflation is not a panacea: moving the average grade from 3.3 to 3.0 does not change much. The system still affords little accuracy and is likely only marginally more useful for the people who are interested in evaluating students by their grades.

So, a solution to the problem of grade inflation must be compatible with the existing system-otherwise, our graduates will be impacted. However, to be a an actual solution, our new system must provide more accurate and useful information about our students.

Part of the problem with the present system of grading is that we expect a single number to convey two pieces of information: how well a student learned the material taught in a course and how a student is doing compared to his or her peers. This second piece of information can be provided more accurately through the class percentile of the student—or percentile in the major.

However, conveying what a student learned from a course is more difficult, and the present system of grading fails to address this at all. Everyone at Haverford knows which courses are easy and teach nothing, and which courses are hard or teach a lot. However, on a transcript, an identical grade in two such dissimilar courses can look quite similar.

The solution to this problem is to expand what is contained in a grade. Certainly a percentile mark is useful at a glance, but there is no reason that a grade should stop there. In the present state of technology, it is possible to include far more information about courses than just a primitive number.

I envision “the transcript of the future” as a digital portfolio of the college student’s work: when someone evaluates a transcript, he or she will initially see a list of courses and marks similar to what we have now. But, if curious about a particular course, he or she can click it to learn more about the course. This individual should be able to see the syllabus for the course, as well as problem sets and other forms of evaluation from the course. This should allow people to determine how much a course actually taught as well as how proficient the student was with the material.

Importantly, a system like this would address one of the key issues that graduates of liberal arts colleges face: people often do not know what we are and do not know what to expect from us. To fix this, we should just show them what we can do.

In some sense, what I have described above is not a solution to the problem of grades. Instead, I am proposing that instead of worrying too much about what the grades themselves mean, we should make available enough information that people can find out for themselves on a need-to-know basis. We cannot fix the 4.0 grading system on our own. However, we can enrich our 4.0-style grades with information that will make them both meaningful and useful for others.