All As is a Sign of Failure

One of the big ideas that haunts discussions of teaching is the strong version of grading on the curve -- the idea that the grade distribution within a class should, or should be forced to, fit a normal distribution. That is, only a few students should be able to get As, and a few students must fail. The alternative to the strong curve philosophy is standards-based grading. In a standards-based system -- of which I am a proponent -- students are graded relative to objectives set out by the teacher. This makes it possible for all students to get As, if they all meet the standard. Indeed, proponents of standards-based grading cite the possibility of a straight-A class with gusto, declaring that such an outcome is proof that the teacher has succeeded. I agree that an all-As class shows that teacher has succeeded relative to the standards. But I also think it shows that the teacher has underperformed when it comes to setting the standards.

There is no hard ceiling for learning about a topic (at least a topic specified broadly enough to be a worthwhile subject for a semester-long class). There is no 100% mastery level to which anyone, let alone all students, can aspire. 100% mastery exists only relative to the teacher-imposed limitations on the scope of the class. Because student aptitudes vary, even the best teacher, who puts in astounding amounts of effort to reach out to students who don't get it or who need different approaches to the material, will have some students who lag, and some who will do exceptionally well. These overachievers, however, may bump into the artificial ceiling created by the limits on the course content.

Thus, if every student is getting an A, it is unlikely to be because all of the students have an equal high level of aptitude. Rather, it is because either low standards or good teaching have allowed even poorer students to meet the standards, and there's simply no room left on the grading scale or in the course material for the exceptional students to shine. This means that there is potential learning for some students that is not taking place. So if every student is getting an A, we should neither accept that*, nor hand out Fs and Ds to those with low A-minuses. Rather, we should raise the standards against which student grades are measured.

And being able to raise standards like this is a wonderful thing! Every teacher worth their salt has encountered the vexing problem of having to limit our course material because we know students can only learn so much in one semester. An upwardly-skewed grade distribution is not an indicator that we've reached success, but rather an opportunity to expand the course and challenge students more fully -- a process which will tend to make a more "normal" grade distribution reappear. The key, however, is that the grade distribution is normalized by a combination of good teaching and rising standards, not by forcing students' work to fit the curve. Rather than saying "I want all my students to get As," we should say "I want all my students to gain at least such-and-such a level of learning, and for those who are able to go as far beyond that as they can."

*A Rawlsian might disagree, arguing that the only thing that matters is the performance of the poorer students. It would be unfair, by this reasoning, to expend effort boosting good students when there are poor students also in need of help. This argument carries weight only if you conceptualize education as being, on the model of income in Rawls's books, primarily of benefit to the student. However, continuing education of prodigies benefits society as a whole, thus placing it into the category of permissible inequalities in Rawls's scheme.