In education, grading on a bell curve is an approach of assigning grades designed to yield a desired distribution of grades among the students in a class.
Strictly speaking, grading “on a bell curve” refers to the assigning of grades according to the frequency distribution known as the Normal distribution (also called the Gaussian distribution), whose graphical representation is referred to as the Normal curve or the bell curve. Because bell curve grading assigns grades to students based on their relative performance in comparison to classmates’ performance, the term “bell curve grading” came, by extension, to be more loosely applied to any approach of assigning grades that makes use of comparison between students’ performances, though this type of grading does not necessarily actually make use of any frequency distribution such as the bell-shaped Normal distribution.
In true use of bell curve grading, students’ scores are scaled according to the frequency distribution represented by the Normal curve. The instructor can decide what grade occupies the center of the distribution. This is the grade an average score will earn, and will be the most common. Traditionally, in the ABCDF system this is the ‘C’ grade. The instructor can also resolve what portion of the frequency distribution each grade occupies and whether or not high and low grades are symmetrically assigned area under the curve (i.e. if the top 15% of students earn an ‘A,’ do the bottom 15% fail or might only the bottom 5% fail?). In a system of pure curve grading, the number of students who will receive each grade is already decided at the beginning of a course.
Other forms of “curved” grading differ, but one of the most common is to add to all students’ absolute scores the difference between the top student’s score and the maximum possible score. For example, if the top score on an exam is 55 out of 60, all students’ absolute scores (meaning they have not been adjusted relative to other students’ scores in any way) will be increased by 5 before being compared to a pre-determined set of grading benchmarks (for example the common A>90%>B>80% etc. system). This method keep from unusually hard assignments (usually exams) from unfairly reducing students’ grades but relies on the assumption that the top student’s performance is a good measure of an assignment’s difficulty.
In the U.S., strict bell-curve grading is infrequent at the primary and secondary school levels (elementary to high school) but is common at the university level.