The recent report Re-Centering Academics at Harvard College revealed that in 2025, As accounted for 60% of grades at Harvard’s undergraduate college. Because Harvard has outsized influence on conversations about higher ed, this report inspired a flurry of op-eds about whether the university has “gone soft.” These concerns have, unsurprisingly, focused on alleged grade inflation at Harvard and more broadly, but as I have been arguing, people who worry about grade inflation are actually concerned about grade compression, or the narrowing of distributions around high grades.
Grade compression is most clearly a problem for those seeking to identify the highest achieving students—a group I call Team Sort. But I believe those who champion learning for all students—a group I call Team Support—are also implicated by compression’s downstream effects.
My core point is this: Even if grades fail to sort students, the sorting function itself is too big to fail. And this reality is almost as big a problem for Team Support as Team Sort. I find the Harvard Report especially useful for illustrating why this is the case, so I want to spend some time digging into this document.
According to the report, authored by Dean of Undergraduate Education Amanda Claybaugh, grades have three key purposes: motivation, information, and distinction, but grades currently fulfill none of these functions at Harvard. Now, I believe traditional grading systems can only meaningfully achieve the third purpose, but let’s examine the reasons cited for these failures in their turn.
Insofar as motivation is concerned, Claybaugh explains: “We expect grading to encourage students to engage with their courses. But faculty perceive—and students confirm—that in our current system grades are as much a limit on effort as a spur.” Instead, “Students will do what they need to do to earn an A, but they are seldom moved to do more than that.”
This observation should not have been surprising; decades of research illustrate the flaws in motivating people with extrinsic rewards like high grades (and the inseparable threat of low grades). Within systems structured around extrinsic motivation, it is rational to do the least necessary to obtain the desired reward—what Susan Blum calls the minimax principle. Students, of course, are more motivated to learn when they find the material intrinsically interesting, a point Claybaugh concedes: “What does encourage students to engage is their sense that a course is meaningful.” With all due respect … duh.
Regarding the information function, Claybaugh clarifies: “We expect grades to give students reliable information about whether they have mastered skills or material.” But Harvard faculty believe that “grades currently distinguish between work that meets expectations or fails to meet expectations, but beyond that grades don’t distinguish much at all.” (Note that even this explanation bleeds into the distinction function, implying that whatever information grades provide about learning cannot be dissociated from sorting students.) I have previously argued against the premise that all courses should focus on acquiring mastery. But setting aside this objection about mastery, there are inherent limits to how much information grades can provide about learning in general.
In his book The Score, C Thi Nguyen explains that grades are a form of quantitative reasoning (essentially a scoring system) built for portability across contexts. By design, the single letter grade washes out the nuances and specificities of what individual students did to receive their letters. In practice, an A can mean all kinds of things depending on the grade breakdown—what counted and what didn’t, whether students were evaluated against one another, whether students were allowed to revise work or retake tests—meaning grades do not (and cannot) provide reliably useful information about whether students have “mastered” course material. Feedback, a form of qualitative reasoning that allows for nuance and context-specificity, can provide considerably more information, but at the cost of efficiency. In a system that centralizes portability over specificity, it is again unsurprising that students often receive feedback that is either minimally useful or that, for one reason or another (including their tendency to minimax), they are ill-equipped to process.
That leaves distinction, of which Claybaugh states: “We rely on grades to distinguish the strongest student work for the purposes of honors, prizes, and applications to professional and graduate schools.” In other words, we use grades to rank and sort … their one honest and true purpose. But when 60% of grades are As—i.e., when grades are compressed—their capacity to distinguish is rendered largely futile. Claybaugh explains: “What is true of prizes is true of graduate and professional school admissions as well. Increasingly, admissions committees tell us that it can be difficult to tell Harvard students apart.”
Throughout this series on the inherent tensions between sorting and supporting, I have maintained that despite high-minded, support-centric rhetoric about education as society’s “great equalizer,” formal education in the US strongly prioritizes sorting. In practice, this system has failed to meaningfully capture student learning, and it has encouraged students to minimax, for a long time—as proponents of alternative grading routinely remind us. But so long as grades (more or less) contributed significantly to the sorting function, higher education could (more or less) carry on with a nod and a wink to the illusion that grades—imbued with Tinkerbell-like magic powers—simultaneously motivate, inform, and distinguish students.
If, however, grades’ role in sorting students were to weaken, this Tinkerbell effect would inevitably weaken as well. And while Claybaugh never explicitly states that distinction is the one function to rule them all, her messaging is quite clear.
In what I find the most telling paragraph in the report, Claybaugh explains that twenty years earlier, Harvard called on faculty to prioritize support-centric teaching practices. After briefly praising faculty for answering this call, Claybaugh turns to the “unintended consequences” of these efforts:
“For instance, many of us shifted from high-stakes exams to more frequent lower-stakes assignments, believing that this would help students retain the material. A number have found, however, that lower-stakes assignments are more effective at rewarding effort than at evaluating performance, giving students the false sense that they’d mastered material that still eludes them. Similarly, faculty shifted from exams and papers to alternate modes of assessment, such as creative assignments and group projects, in the hopes of increasing student engagement with their courses. A number struggled, however, to assess these assignments in a sufficiently differentiated way.”
Remarkably, the “creative assignments” Claybaugh cites as cultivating a “false” sense of mastery—e.g., collaborative and project-based learning—are considered high-impact practices, whereas allowing students to retake quizzes and tests and to revise work are key to Robert Talbert and Dave Clark’s four pillars of alternative grading. In other words, these teaching and assessment methods are designed to help students learn more than in lecture-based and traditionally graded courses. But Claybaugh scarcely entertains the possibility that these methods might have worked as intended, enhancing learning for students who struggle in more conventional course environments. Instead, she suggests—without any evidence, mind you—that low-stakes learning opportunities allow lower-performing students to earn high grades without mastering the material—whatever that pesky, undefined word “mastery” is supposed to mean.
Let’s explore the insufficiently “differentiated” assessment of collaborative work. That’s kind of the point, isn’t it? If students are genuinely collaborating, they are helping one another learn, so everyone benefits. And plenty of research indicates that community-building is one of the most effective ways to promote student engagement, learning, and persistence. It’s hard to imagine these reciprocal benefits accruing if we tell students to work in teams while making sure to “differentiate” themselves in the process. What are students supposed to do if they are told, essentially, to collaborate and compete at the same time?
Claybaugh concludes the paragraph by directly impugning alternative graders:
“Finally, some faculty have eschewed conventional grading, turning instead to ‘ungrading’ or ‘contract-based learning’ or other systems in which students earn As for completing all assigned work. There is a pedagogical case to be made for these alternate approaches, but they’re fundamentally at odds with our current grading system, which relies on grades to differentiate.”
That last line, where Claybaugh acknowledges—or rather, confesses—that these practices are pedagogically defensible but institutionally indefensible, gives up the game entirely. While Harvard faculty possess a range of views about how to address grade compression—and some surely remain committed to the practices cited by Claybaugh—the institution has taken the position that Harvard can no longer even pretend to prioritize support for all students. The Tinkerbell grading fairy, dear readers, has died at Harvard College.
So why should this matter to Team Support? Because the consequence of widespread grade compression at Harvard, far from slaying the sorting beast, has been to displace sorting to other domains of university life. Claybaugh observes that once students have minimaxed their A, they “direct their energies elsewhere, either adding more courses to their schedules or taking on more demanding extracurriculars.” To put this another way: The sorting function abhors a vacuum of mechanisms for sorting students.
Members of Team Support should find such displacement deeply troubling. After all, Team Support—especially alternative graders—wants students spending less time obsessing over and being distracted by grades. But if students—still grappling with precarity, still feeling caught in the logic of high-stakes sorting, still keeping up with the collegiate Joneses—believe they must obtain high grades and do all those extracurricular things too, they will have even less time, energy, and headspace (let alone motivation) to engage the learning process in the ways Team Support wants. As I turn to in the next post, this development will cause more stress and misery while further exacerbating inequity … and I don’t think this is just a Harvard problem.
Teaching Inquiry 