What Can We Learn from Exam Grade Distributions?

Abstract

A course with good learning outcomes is one in which most of the enrolled students achieve the mastery specified in the predefined learning objectives. Since the enrolment is invariably a mix of students with heterogeneous capabilities, the class average grade is a poor indicator of how the class is divided into at least two groups, of high- and low-performers. Clearly, achieving the desired outcomes implies increasing the proportion of high-performing students and their mean grade by providing appropriately designed teaching protocols. In this paper, the actual class grade distribution is approximated by a bimodal probability distribution function, whose parameters enable the proportions and average performance of these two groups to be quantified. This paper describes the methodology to achieve this and demonstrates its usage to diagnose example exam grade distributions, as well as to provide quantification of the impact of pedagogic changes on the degree of achieving teaching objectives.

Keywords

• assessment
• bimodal distributions
• outcomes