Grades are important. Instructors structure their courses around deciding them; students expend enormous effort pursuing them; they form the only record of student performance every University feels obligated to retain. They’re also consequential. Grades provide students with their primary performance feedback, threshold course passage and the award of diplomas, and are combined to rank students for the receipt of awards and honors.
Unfortunately, grades are worn out, a tattered remnant of the last century’s bureaucratic needs. This exhaustion was anticipated from the start – grades serve too many masters.
“The experiment started by the faculty five years ago must be pronounced a complete failure. And both students and faculty have before now felt it to be a failure. There is no uniformity of grading, but the greatest divergence. It has come to be admitted openly that a student who is anxious to win honors must be careful to elect his work under certain teachers and avoid others as much as possible.”
That’s from a paper Max Meyer wrote in Science in 1908. He was describing just how dreadful it is to pretend to quantify something on an A-F scale when there’s no agreement at all on what the scale means. The image below comes from a 1930 internal report on grading here at the University of Michigan.
Right there at the start, Meyer could see that instructors were using grades in two wildly different ways. At one extreme, their sole purpose was to record relative performance: to rank. Grades awarded on a curve tell us nothing at all about absolute learning. At the other extreme, their sole purpose was to record absolute performance. Grades like this tell us how well the student met the (typically unrecorded) goals of the class. They don’t care about ranking.
Today, most classes award grades using systems torn between these two goals. A few retain a rigid curve, insisting that specific fractions of students will receive each grade, focusing on ranking. Most propose an ‘absolute’ scale, against which, in principal, every student could receive an A. In practice, few instructors are comfortable with that outcome, so they adjust a combination of their course expectations and grading standards to ensure an average grade between B+ and A-.
When a student gets a B in a world like this, we have very little idea what it means. Oh, we try to use context to guess what it means. What kind of context do we have? Usually just the name and number of the course. If it’s a big introductory course in a field like economics, we might assume that this is a pretty good grade. If it’s a grade for participation in marching band, we might guess that it’s not a good sign. But seriously, using a grade to assess what a student learned in a class is worse than a guess: it’s a part of writing a narrative which tells us whatever we want to hear.
There are many ways we might set aside grades as they exist and move to something better. Hopefully, a new system will do what grades do, only much better. Ideally, it will accomplish new goals as well. But whatever we switch to will have to begin by taking on the functions grades perform today. What are they?
We need to provide students with meaningful performance feedback, threshold course passage and progress toward degrees, and help both the institution and the world beyond compare the performance of students. What’s worse, we need to do all of this across a wild array of experiences ranging from huge introductory courses to senior independent study, and from courses in creative writing to abstract algebra.
Sounds like a daunting challenge. There is good news though. Today, we do all this with letter grades, a broken down system which was declared a failure from the start. So whatever we come up with is almost certain to be better. I’m planning to weigh some of the many possibilities in future posts.