Harvard’s grade cap won’t restore academic standards

Harvard University’s decision to cap A grades has reignited a debate that has long been smouldering across higher education: how can universities address grade inflation without undermining fairness and genuine learning?

The problem certainly demands a solution. Across many universities, and especially elite institutions, grades have drifted steadily upward for decades, raising understandable concerns about academic standards and the meaning of distinction.

Generative artificial intelligence intensifies the problem further. If students can use AI tools to improve assignments without corresponding gains in understanding, grades become even weaker indicators of actual learning.

But in that environment, artificially redistributing grades addresses appearances more than substance.

A quota may reduce the number of As statistically, but it does not necessarily increase rigour or improve learning. More importantly, it shifts grading towards a relative system in which students are judged partly against one another, rather than against clearly defined standards of mastery.

Students protesting Harvard’s policy are correct to view it this way – and to regard it as unfair, that if an unusually strong class exceeds the cap, some students who genuinely earned top marks must still receive lower grades. A student’s grade should reflect demonstrated learning, not the performance distribution of classmates in a particular semester.

Yale University’s recent committee on grading offered a far better principle: “Grade like we mean it.” If only 15 per cent of students earn As in a demanding course, so be it. If 30 per cent genuinely demonstrate mastery, they should receive the grades they earned.

But this requires a significant resetting of the institutional incentive system that makes grade inflation predictable. It requires rigorous expectations, meaningful assessments and, crucially, institutional support for faculty members willing to uphold standards consistently.

Universities increasingly rely on student evaluations of teaching in promotion, tenure and compensation decisions. And whatever their original purpose, these systems often reward student satisfaction more than intellectual challenge. When rigorous grading risks lower evaluations, it can become professionally costly. As the management theorist W. Edwards Deming observed, “A bad system will beat a good person every time.”

Encouragingly, some institutions are already recognising the limitations of traditional teaching evaluations. The TEval initiative, involving institutions such as the universities of Colorado Boulder, Kansas, and Massachusetts Amherst, has begun developing broader systems for evaluating teaching through peer review, teaching portfolios, classroom observation and evidence of student learning, rather than relying primarily on student satisfaction surveys.

Just as research assessment has begun in recent years to focus much more on real-world outcomes and societal impact, so teaching assessment should focus less on grade distributions and more on whether students are genuinely developing analytical ability, ethical reasoning, critical thinking and problem-solving skills.

Academic standards will not be restored by engineering scarcity. They will be restored when universities align incentives with rigour, integrity and meaningful learning – so that grades reflect genuine mastery, not quotas, curves or administrative optics.

Mohan Rao is professor of decision sciences at Texas A&M University-Corpus Christi.