Professional Learning  ›  Assessment  ›  Workshops  ›  Grading on a Curve
a ready-to-run workshopGrading on a Curve
Assessment · Workshop

Grading on a Curve (Workshop)

Everything you need to lead this with your staff or PLC: a minute-by-minute run-of-show, what to say, what to hand out, and how to make it stick. Pick the length that fits your slot.

Key takeaway

Grading on a curve forces grades into a preset distribution based on how students rank against each other rather than what they learned, which means a student's grade depends on their classmates and some must fail by design, a practice that conflicts directly with grading for learning, so the honest move is usually to fix the real problem a curve is covering rather than to curve.

Format
PLC / team meeting
Length
45 minutes
Group size
Any (works 4-40)
Who can run it
Any teacher-leader
You will need
Slides (5)Printed handoutA time you were tempted to curveA timer
Aligned to
Learning Forward: Rigorous ContentInTASC 6: AssessmentResearch on norm-referenced versus criterion-referenced grading
Share / assign

Why it works

Grading on a curve means adjusting grades so they fit a predetermined distribution, often a bell curve, based on how students rank relative to one another rather than on how their work measures against a fixed standard. In its strict form, a fixed percentage of students receive each grade regardless of their absolute performance: the top slice get As, the bottom slice fail, and the rest are distributed between, so that grades are a ranking of students against each other. It is worth examining critically because, although it persists and can be tempting, it conflicts directly with grading for learning and carries consequences many teachers who reach for it have not fully considered. The core problem is that curving makes a grade norm-referenced, a measure of a student's rank among classmates, rather than criterion-referenced, a measure of what the student actually learned against the standard, and those are very different things with very different implications. Under a curve, a student's grade depends not only on their own learning but on who else is in the class: the same performance earns a high grade in a weak class and a low grade in a strong one, which is arbitrary from the student's point of view and unfair. A curve also builds in that some students must fail (or get low grades) by design, regardless of whether they met the standard, because the distribution requires a bottom, which means the grade can tell a student they failed even if they learned the material adequately, simply because others did better. It pits students against one another, since one student's better performance can lower another's grade, which poisons collaboration and the class climate and is the opposite of a learning community. And it disconnects the grade from the standard, so the grade no longer communicates what a student knows and can do, which is what a grade is supposed to mean. Teachers are usually tempted to curve for understandable reasons, and the better response is almost always to address those reasons directly rather than to curve. If an assessment was too hard or flawed, so that scores are uniformly low, the honest fix is to examine and adjust the assessment or reteach and reassess, not to curve scores that reflect a bad test. If the grades seem too low across the board, the question is whether the teaching, the assessment, or the standard needs attention, not whether to redistribute the existing scores. If the concern is that criterion-referenced grading produced too many low grades, that is a signal to investigate, not to paper over with a curve. The constructive alternatives, then, are criterion-referenced and standards-based grading, where every student is measured against the standard and any number can succeed or fail on their own merits; fixing flawed assessments rather than curving their results; and reteaching and reassessment when learning genuinely fell short. A caution on terminology: teachers sometimes say curve loosely for adjusting an unfair assessment (adding points to a flawed question, dropping a bad item), which is really correcting the assessment and is legitimate; the problematic practice is the true norm-referenced curve that forces a distribution and makes grades about rank. The workshop helps teachers understand what grading on a curve really does and why it conflicts with grading for learning, and address the real problems that tempt them to curve rather than curving.

The research: Research on norm-referenced versus criterion-referenced grading and standards-based grading.

The run-of-show

Choose your slot. The agenda, timings, and length update to match.

    The core activity: fix the real problem instead of curving

    The heart of the session is understanding what grading on a curve really does and addressing the problems that tempt teachers to curve rather than curving.

    Fix the problem the curve is coveringA curve makes a grade a ranking against classmates, so some must fail by design and the same work earns different grades in different classes. The honest move is almost always to fix the real problem, a flawed test, the teaching, the standard, rather than redistribute scores.

    Facilitator notes

    Correcting a bad test is not curvingTeachers sometimes say curve for adjusting a genuinely flawed assessment, dropping a broken item, adding points for an unfair question, which is legitimately correcting the assessment. The problematic practice is the true norm-referenced curve that forces a distribution and makes grades about rank.

    Adapt it

    Participant handout

    One page for every teacher. Print it, or save it as a PDF.

    K12 Academics · Professional Learning

    Grading on a Curve: fix the cause, not the distribution

    Understand what curving really does and choose better: keep grades about learning against the standard, fix flawed assessments, and reteach, rather than forcing a distribution based on rank.

    1. A time I was tempted to curve:
    2. What curving would have made the grade mean (rank, not learning):
    3. The real problem behind the temptation (test, teaching, standard):
    4. The alternative I will use (fix the test, reteach, criterion-referenced):
    5. How I keep grades measuring learning, not rank:

    Make it stick

    Fixing the cause beats curving the scores. Close by setting the follow-up:

    Common questions

    If a test comes out with everyone scoring low, isn't curving the fair thing to do?
    Uniformly low scores are a real signal that something needs fixing, but curving is usually the wrong fix, because it papers over the problem instead of addressing it, and the honest response depends on diagnosing why the scores are low. If the scores are low because the assessment was flawed, too hard, poorly worded, testing things not taught, with broken items, then the right fix is to correct the assessment: drop or rescore the bad items, adjust for a genuinely unfair question, or in a serious case readminister, which addresses the actual cause and is legitimately different from curving. If the scores are low because students genuinely did not learn the material, then the fix is to reteach and reassess, so that grades come to reflect actual learning, rather than to redistribute scores that honestly show students did not learn it. What curving does instead is force the low scores into a nicer-looking distribution based on rank, which makes the top of a low-scoring class look fine and the bottom fail regardless of what anyone actually learned, hiding the real problem rather than solving it. So uniformly low scores should prompt the question why, with the answer pointing to fixing the test or reteaching, both of which serve learning, rather than to a curve that just cosmetically redistributes the result.
    Doesn't curving protect students from an unusually hard test or a tough grader?
    It can make grades look better in those situations, but it does so by introducing new unfairness and disconnecting grades from learning, so the better protections are the ones that fix the actual cause. If the test was unusually hard because it was flawed, the genuine protection is to correct the flawed assessment, which is legitimate and directly addresses the problem; students are protected because the bad items no longer count, and the grade still reflects learning on the fair parts. If the concern is a tough grader whose standards are miscalibrated, the fix is to calibrate the grading against a shared, appropriate standard, so grades are fair and consistent, rather than to curve, which just redistributes miscalibrated scores. A true norm-referenced curve, by contrast, protects students only in the narrow sense of raising some grades, while creating the deeper unfairness that a grade now depends on classmates rather than the student's own learning: the same work earns different grades depending on who else is in the room, and some students fail by design regardless of what they learned. That is not real protection; it is a cosmetic adjustment that trades one problem for several. So when the worry is an unfair test or inconsistent grading, address those directly, fix the assessment, calibrate the standard, which protects students in a way that keeps grades meaningful, rather than curving.
    Isn't some version of curving necessary to rank students for things like class rank or competitive admissions?
    Ranking students and grading their learning are different purposes, and conflating them is part of the problem; where a genuine ranking is needed, it can be produced without making everyday grades norm-referenced. Class rank, competitive selection, and similar uses do require comparing students to one another, and there are legitimate ways to generate such comparisons when they are genuinely needed. But that is a separate function from what a grade is primarily supposed to do, which is communicate what a student has learned against a standard. Building norm-referencing into ordinary grades by curving every assessment corrupts that primary function across the board for the sake of a ranking purpose that applies only in specific contexts, which is a poor trade. The better approach keeps everyday grades criterion-referenced, so they accurately communicate each student's learning, and derives any needed ranking separately from those grades or from purpose-built measures, rather than forcing all grading into a competitive distribution. This also avoids the serious costs of pervasive curving, students pitted against each other, collaboration discouraged, grades dependent on classmates, which damage the learning environment in pursuit of a ranking most of the grading does not need. So the existence of some legitimate uses for student ranking does not justify curving ordinary grades; it argues for keeping grades about learning and handling ranking as the separate, limited function it is.

    Go deeper

    Build it into a bigger day

    Workshop
    Standards-Based Grading
    Workshop
    Grading for Equity
    Workshop
    Zeros & Grade Averaging