GPA, GMAT, and SCALE: A method for quantification of admissions criteria

Abstract

Multiple regression analysis is used to establish a scale, measuring involvement in campus activities, work experience, technical background, references, and goals. This scale is then tested to see whether it improves the prediction of success in school. For graduate business students, previous studies have shown a coefficient of determination .14 when both GMAT score and grade point average are used. Introduction of the scale raises the coefficient to .19 for the 1976–1977 graduate class. Appropriate predictive cross-validation techniques indicate that the regression is validated on data from the 1980–1981 academic year. Using standard statistical computer software, any school can choose variables, experimentally build a scale, and test its predictive value. This type of scale helps to integrate much of the miscellaneous information collected on the standard college and graduate school application.