How To Calculate A Curved Grade

Curving a grade means rescaling marks rather than changing them. The order of the class never changes under any of the standard curves, because every student's mark goes through the same transformation. What changes is the number that appears on the transcript and how it lines up against the grade boundaries.

The reason curves exist is that exams are not calibrated. A test written slightly too hard produces a class average of 62 when the material was understood well, and reporting that as a C misrepresents what happened. A curve restores the intended meaning by shifting the scale.

The proportional curve is the most common and the easiest to justify. Take the highest score in the class as the new top of the scale and scale everything against it. A raw 78 against a class best of 90 becomes 78 divided by 90, times 100, which is 86.7 per cent. If the best score was a perfect 100 the curve does nothing at all.

Its appeal is that it preserves ratios exactly. Someone who scored half of the top score still scores half of the top score afterwards, so the curve compresses or stretches the whole distribution without distorting the gaps between students.

The square root curve takes a different approach. Multiply the square root of the raw score by ten, so 78 becomes ten times 8.8318, or 88.3. Because the square root grows quickly at low values and slowly at high ones, this curve lifts weak marks a great deal and strong marks only slightly.

That asymmetry is the point. A student on 36 jumps to 60, a gain of 24 points, while a student on 81 moves to 90, a gain of 9. The square root curve is popular when a test was hard across the board but the instructor wants the bottom of the class pulled up to a passing level.

The flat curve is the simplest of all: add the same number of points to every score. If the highest mark in the class was 90, adding 10 brings it to 100 and everyone else rises by 10 as well. It is transparent and easy to explain, though it can push marks above 100 if the top score was already high.

A fourth method, curving to a target average, works backwards from a desired class mean. If the average was 68 and the instructor wants 75, every score rises by 7. It is really the flat curve with the adjustment derived from the mean rather than the maximum.

Grade boundaries interact with all of this in ways that catch people out. Curving before applying a letter grade scale and curving after are not the same operation, and a curve that lifts everyone by ten points may push a large block of students over a boundary at once. Instructors usually decide the boundary question first.

There is a fairness argument on both sides. Curving to the top score means your grade depends on someone else's performance: a strong classmate can lower your mark, which many students find hard to accept. A flat curve avoids that but ignores how hard the test actually was.

Curves also interact with the shape of the distribution. In a class where scores cluster tightly, a proportional curve barely changes anything because the top score is close to the rest. In a class with one outstanding student and a wide spread, the same curve lifts everyone considerably.

The practical advice is to know which method is in play before you calculate anything. Instructors usually state it in the syllabus, and the three standard methods give noticeably different answers from the same raw score: 78 becomes 86.7 under proportional, 88.3 under square root, and 88 under a flat curve to a top score of 90.