If the test's questions are well designed, the students' scores should be roughly normally distributed. Say the mean score on the test is 100, with a standard deviation of 10 points.