Compare six curve methods
GradeCompute includes flat bonus, proportional scaling, square-root curve, target-mean shift, min-max linear rescaling, and z-score rescaling. Each method answers a different statistical question. The calculator shows the formula rather than declaring one method best.
Flat, proportional, and target-mean curves
A flat bonus adds the same number of points to every score. Proportional scaling multiplies every score so a chosen current top becomes a target top. Target-mean shift adds the same amount needed to move the class mean to a chosen value.
Linear and z-score rescaling
Linear rescaling maps the raw low and high scores to target endpoints. Z-score rescaling preserves relative standardized distance while changing the target mean and population standard deviation. It does not force a quota of letter grades.
If every raw score is identical, the standard deviation is zero and z-score rescaling is undefined. Use a target-mean shift instead. Read the neutral curve-method guide for policy questions.
Grade curve FAQs
What does grading on a curve mean?
It means transforming raw scores using a stated mathematical method or policy.
Which curve methods are included?
Flat bonus, proportional, square-root, target-mean, linear, and z-score rescaling.
Can a curve lower a grade?
Yes. Some target settings can lower scores, especially when moving a mean or endpoints downward.
Should results be capped at 100%?
The default cap is on, but the unclamped formula remains visible. Follow the applicable policy.
Is a z-score curve the same as forced letter-grade quotas?
No. This rescaling changes the numerical distribution and does not assign a fixed quota of letters.
How is the square-root curve calculated?
Multiply 100 by the square root of the raw percentage divided by 100.