Alternative class ranks using z-scores

Abstract

Grades at US colleges and universities have increased precipitously over the last 50 years, suggesting that their signalling power has become attenuated. Moreover, average grades have risen disproportionately in some departments, implying that weak students in departments with high grades may obtain better class ranks than strong students in departments with low grades. Using a unique data set, we show that Colby College has experienced both trends and that variation in student ability does not fully explain variation in grades awarded in different departments. We then propose a simple alternative measure of calculating class ranks based on z-scores, and demonstrate that this relative measure of rank mitigates the bonus (penalty) associated with taking courses in departments with high (low) grading distributions.

Keywords:

• grades
• class rank
• normalisation

Notes

1. Throughout this paper, 0.0 reflects an F, 1.0 reflects a D, 2.0 reflects a C, 3.0 reflects a B, and 4.0 reflects an A. ‘Pluses’ add 0.3 points to the grade and ‘minuses’ reduce the grade by 0.3 points, except for F grades. Thus, grades range from 0 (F) to 4.3 (A+).

2. As such, the phenomenon of rising grades is often dubbed ‘grade inflation’. In this paper, we deliberately avoid using this potentially incendiary terminology.

3. Heterogeneity in distaste for defending low grades and meeting with ‘nuisance’ students may contribute to variation in grading across instructors of the same courses (Grant 2007).

4. Grade sensitivity has also been shown to influence the gender composition of majors such as economics (Rask and Tiefenthaler 2008).

5. At the request of the Office of Institutional Research, departments are identified by number rather than name. On a positive note, such anonymity may facilitate discussions of class rank without alienating high- or low-grading departments.