Adapting Courses and Assessments to Enable Fair Grade Differentiation

Why fair grade differentiation matters

The 2025 FAS Grading Report speaks of grades as having three primary functions: motivating students to engage seriously with course material, clarifying for them what they have and have not mastered, and distinguishing the strongest work for the purpose of honors, prizes, and graduate and professional admissions. Grades cannot do any of these well when they cluster at the top. As one faculty member put it, "There's not much resolving power at the top." Per the FAS Faculty Handbook, the grade of A is reserved not only for work "whose excellent quality indicates a full mastery of the subject" but also work that is "of extraordinary distinction." A course produces fair grades when its assessments acknowledge the quality of the work submitted. The aim is not to grade harder. It is to grade so that the grade carries real information.

To produce a fair, differentiated distribution:

Place greater weight on assignments that produce meaningfully differentiated grades.
Anchor a large proportion of the grade in work you can confirm is the student's own.
Build genuine difficulty into required work.
Reserve A’s for work that is of extraordinary distinction rather than for work that checks every box.

Getting Started

A few things you can do right now:

Review your grade distribution. As per the new grading policy, your distribution should support identifying no more than a small proportion of students who have demonstrated extraordinary distinction in the course, and it should separate degrees of mastery below that.
Audit your grade weights. Consider how much of the final grade rests on effort and completion and how much rests on demonstrated mastery and excellence, then rebalance toward work that differentiates as needed.
Incorporate assignments that provide open-ended challenges. Make sure at least one required assignment gives the strongest students room to show distinction.
Check where you draw your differentiation. Ensure that a substantial portion of the grade is reserved for work you can confirm is the student's own. Add an in-person component to a major out-of-class assignment if needed.
Set your course policies. Setting Course Policies that Center Academics covers in-person engagement and the appropriate use of generative AI.

Detailed Guidance

Raise the bar on required work

Ensure students can demonstrate distinction on a required assignment by including a part that’s open-ended, a problem that asks them to apply their learning in a new context, or a question with no limit at the top end.

Use assignments that span the range of mastery. An assignment with a broad range has an entry point that works for all students and no limit on how far the strongest students can take it.
Assign open-ended challenges in required work, not optional extras. Optional challenge problems and extra credit tend to compress grades, because they shift the grades toward a measure of effort rather than mastery and excellence. Optional challenge work can be valuable, but it is unlikely to differentiate grades fairly.
Ask students to apply ideas to something new. A familiar, well-rehearsed task is one most prepared students can complete, so it cannot reveal the strongest work. A novel problem draws on the conceptual depth and transfer that differentiate student work.
Include genuinely hard items in every major assessment.
Audit your assignments for low difficulty. As a rule of thumb, if over 15 percent of students earn the maximum score/grade on a component of an assignment or on the assignment overall, that component/assignment is hiding the strongest work rather than identifying it.

Anchor grades in verifiable work

Grades are fair only when the work on which they are based is reliable evidence of what the individual student knows. Students have many sources of outside help, including other people, published solutions, and generative AI. A take-home assignment can be designed to differentiate students yet still produce submissions that do not reflect a student's own work. When that happens, grades do not track who learned the most.

Pair major out-of-class work with an in-person component. Follow a take-home project with a short oral defense, scaffold a paper with in-class writing or presentation activities, a problem set with an in-class oral or written understanding-check, a reading assignment with an in-class response.
Consider a seated exam if you do not use one. Seated exams reward engagement with all the material, tend to differentiate well, and give reliable evidence of individual learning in the age of generative AI.
Make in-class activities contribute in a measurable way to meaningful participation grades.
Make end-of-semester assessments cumulative, so they build on the whole course rather than the last unit alone.

Use rubrics to differentiate, not to bind

A rubric should sharpen judgment about quality, not compel you and your teaching team to give a high grade for work that checks every box but does not excel.

Leave the top end open. Reserve the top grade for work that is excellent as a whole, and tell students that is what it is for. An "exceeds expectations" band lets you recognize excellence you did not pre-specify or anticipate, and keeps a complete but ordinary submission from earning the top grade automatically.
Do not make the top grade conjunctive. A conjunctive top grade is one a student earns by satisfying each element of a rubric. Under that design, work that is solid on each criterion but distinguished on none earns the top grade, while genuinely excellent work with one small flaw is kept out.
Distinguish excellent from very good, not just passing from failing. A criterion that separates adequate work from inadequate work will not, on its own, separate the best work from the merely strong.

Calibrate grading and tell students what grades mean

Invest in calibration, not in a more elaborate rubric. A rubric keeps grading consistent across graders, but one detailed enough to script every decision rewards box-checking and flattens the top. Keep the rubric broad enough for a holistic judgment of quality, then calibrate: work from exemplars of genuine A-level work and have graders reconcile on shared samples.
Calibrate across sections before final grades are set. Have your teaching team grade collectively, each taking certain questions, or norm on shared samples before grading.
Review graded work for accuracy before grades go out, so you can stand behind a TF's decision on appeal.
Tell students what each grade represents in your course. Students often understand grades in terms of effort. Be explicit that grades measure mastery of content and skill, not just the amount of work put in, and describe what excellent work looks like.