Teachers grade a test on a curve by making the highest score from the class represent a 100 percent score, regardless of the actual top score, and then assigning different grades to percentages of the class based on the top score. Typically, this means that students who earn scores in the top 20 percent of the class receive an A or B, students in the bottom 20 percent get a D, and all other students get a C.
Grading on a curve uses the bell curve, also known as the normal distribution or Gaussian curve. This is a familiar, bell-shaped graph of a mathematical function. The area between the curve and the x-axis of the graph represents the percent of the population being studied, in this case the scores that students earn on a test.
Based on the bell curve, about 60 percent of a population lies near the average, which is the highest point on the graph. Only about 20 percent of the population lies in each of the trailing portions of the graph, known as the tails.
Some teachers choose to curve a test because they believe it reflects the students' actual performance on the test more accurately than the original scores. Curving is more common on college tests, where averages are sometimes as low as 60 percent, meaning that most students would receive a D if the exam were not curved.