On May 19, the Faculty of Arts and Sciences of Harvard University voted 458 to 201 to lower A grades in undergraduate courses, limiting the total number of A grades to 20 percent of course enrollment, and four students. Other institutions that look to high-profile peers for insurance will be tempted to follow suit. You shouldn’t. Nearby institutional analogues have produced documented harm. Even Harvard’s 2023 ranking report warned against such policies.
Harvard’s case for the cap may sound intuitive: If everyone gets an A, the A doesn’t matter. But that view confuses meaning with scarcity. If we are going to use general grades, the grade should report what the student did against a defined standard—not where he finished on the class level. If Harvard wants grades to mean more, the answer is academic: clear grades, not random A’s. Rarity is not durability.
The cap moves in a different direction: away from the standards and towards the level of the peers. It includes a fixed quota for the top end of the highest grade in undergraduate courses, with a predetermined maximum number of seats at the top. A Harvard subcommittee predicts the cap will reduce anxiety and encourage intellectual exploration. Research on similar programs suggests otherwise. A pre-registered test found that common-referenced grading—where students are measured against their classmates instead of a fixed standard—produced lower professional orientation, weaker self-esteem and less help-seeking and giving among students. When Cornell University began publishing average course grades online, a follow-up study found an increase in enrollment in less structured courses—a pattern consistent with strategic, grade-preserving course choices. There is no research to support the assumption that students will be less strategic or more analytical. Both describe the incentives that draw students to placement rather than learning.
A clear caveat, however, comes from equity research. Wellesley College has already implemented the closest grade-capping test: a ceiling of B-plus on average grades in 100- and 200-level courses with at least 10 students, since fall 2004. Wellesley economists analyzed what the cap did to grades, enrollment, majors and aptitude tests. The overall drop in grades in departments affected by the policy was approximately 0.17 grade. For black students, it was 0.36—more than double. Undergraduate enrollment in those departments dropped 18 percent; majors are down about 30 percent.
The percentage of magna cum laude graduates in medical departments dropped from 20 percent to 16 percent. Academic evaluations of affected intelligence decreased. The costs were focused on Black students, students with low incoming test scores and faculty in affected departments. Harvard’s policy is not the same as that in place at Wellesley, which the college withdrew in 2019. But the basic idea—a ceiling on the numbers at the top grades—is the same. Harvard could use it more.
The tests that came to the faculty will not come equally, either. A study using a random assignment of students to instructors found that male students rated female instructors about 21 percent of a standard deviation lower than male instructors, although instructor gender did not affect students’ grades or the number of hours they spent studying.
Harvard’s Office of Undergraduate Education has marked the way forward—that expected grades and student teaching evaluations go hand in hand. Its 2023 grading report said, “There is a direct correlation between expected grades and Q [teaching evaluation] ratings that may have an impact on higher rates.” The same report concluded that “institution-wide policies such as capping the percentage of A grades or publishing average course grades on transcripts often lead to unintended negative consequences.”
Three years later, Harvard voted to introduce such a policy. If the cap forces instructors to give lower grades than they otherwise would, the coercion will be paid for with intelligence test scores that already reveal the bias written into that test. A hat can reproduce this imbalance unintentionally.
Princeton University tried the same dominant approach—targeting numbers—and abandoned it. Princeton set targets for A-range grades in 2004 and removed them in 2014, after a review committee said the targets were “often misinterpreted as ratings” and added “a huge element of stress to students’ lives, making them feel like they’re competing for a limited supply of A grades.”
Princeton faculty recognized this. So are Harvard students themselves. In a self-selected survey by the Harvard Undergraduate Association of nearly 800 respondents, 84.9 percent “definitely” opposed average A grades. More than 72 percent definitely or probably opposed a different percentile ranking system that Harvard would use to determine internal honors.
There are other methods that do not measure A. Criterion-referenced grading correlates grades to demonstrated standards rather than peer standards. Transparent rubrics, standards-based assessment and arts-based course design all maintain rigor without treating the class like a competition. Each ends up focusing where teaching should be: on what students have learned, not where they finished on the line.
A Harvard subcommittee said the cap would mean “the A requirement will no longer be a source of anxiety, encouraging students to explore new courses and take intellectual risks.” It is difficult to read the Wellesley data, the Princeton conversion, the Cornell median-grade test and the Harvard OUE report of 2023 and believe that the quota system makes students less anxious or less strategic. Nearby evidence has given warning. Harvard’s own report issued a similar warning three years ago. We know how this ends.
Do not follow.
