COLLEGE MAJOR AND GENDER DIFFERENCES IN THE PREDICTION OF COLLEGE GRADES

Abstract

ABSTRACTRecent studies have found substantial reductions in gender differences in the prediction of academic achievement in college when variations in grading standards among courses were taken into account. The purpose of this project was to examine gender differences in the prediction of freshman grades after controlling for differential course grading based on college majors. This method involved deriving a variable that measured grading leniency using residual scores from the within‐gender regressions of freshman grades on high school grades and scores on the SAT for the non‐Latino white group. The procedure worked quite well and generalized to other groups not involved in the derivation of the grading‐leniency scale. Nevertheless, there were modest, sometimes statistically significant, gender differences in prediction that remained after this control variable was introduced into the regressions. The largest and smallest differences for females between actual grades and grades predicted from the males' regressions tended to be found in the African American and Asian American groups, respectively. The results imply that the use of information on college majors is a reasonable, practical procedure for controlling for grading leniency.