Grade curve calculator

A grade curve is a published adjustment that takes every raw exam score in a section and lifts or compresses the set so the class mean matches a target the instructor chose. Teachers reach for a curve when an assessment was harder than intended, when two sections must share a comparable average, or when a department asks for a stated mean without rewriting items. This grade curve calculator accepts a list of non-negative raw scores, a target mean, and a method of linear or square-root scaling. Linear scaling multiplies every raw mark by the ratio of the target mean to the current mean, so relative order stays the same while the new average lands exactly on the target. Square-root scaling first takes the square root of each raw mark—which raises low scores more than high ones—then rescales the transformed values so the curved set still hits the same target average.

Linear scaling versus square-root scaling to a target mean

A classroom grade curve is not a letter on a bell and is not a one-score rescaling to a new maximum. It is a map from an entire roster of raw marks onto a new roster whose mean equals a number you publish in the syllabus or the department memo. Linear scaling is the transparent multiplier: compute the current raw mean, divide the target mean by that raw mean, and multiply every student score by that single scale factor.

Rank order is unchanged, ratios between students are unchanged, and the new class mean equals the target (up to ordinary floating-point rounding). Example: the four marks fifty, sixty, seventy, and eighty have mean sixty-five; a target mean of eighty applies the factor eighty over sixty-five, so each score rises in the same proportion and the curved mean is eighty. Square-root scaling is the concave alternative used when the exam was punishing at the bottom of the range.

First replace every raw mark x with the square root of x, which lifts a forty more, relatively, than it lifts a ninety. Example: the three marks thirty-six, sixty-four, and one hundred have square roots six, eight, and ten, whose mean is eight; a target of eighty uses a factor of ten and returns sixty, eighty, and one hundred. Both methods fail closed when a raw score is negative, because a signed penalty is not a grade that these two published maps are defined on.

Both methods also fail when every score is zero and the target is positive: there is no finite scale factor that turns a zero roster into a positive mean. When the roster is all zeros and the target is also zero, the calculator returns zeros with a scale factor of one. Choose linear when you want a fair proportional lift that a student can check with a calculator; choose square-root when you want extra relief at the low end while still promising a stated class average.

Where a grade-curve calculator earns its keep

Instructors and department chairs need one documented map from a posted roster plus a published target mean onto a list of adjusted scores so the gradebook, the parent portal, and the office-hours explanation never contradict each other. A linear curve is the conversation you can show on a whiteboard: every student is multiplied by the same factor, and anyone with a calculator can verify a neighbor’s new mark.

A square-root curve is the conversation you have after a brutal midterm: lower scores move more, the top of the range still cannot jump over a well-chosen target, and the class average is a number you stated in advance rather than a surprise after ad-hoc bonuses. LMS plugins and registrar scripts can call the same contract instead of each instructor keeping a private spreadsheet with a different formula.

This tool is not a substitute for assigning letters from a normal z-score (that is a different search and a different calculator) and it is not a converter that takes one score from an old maximum to a new maximum. Failed validation does not consume balance.