Prediction of Success in Graduate School at Rutgers University

Abstract

THE PRESENT STUDY is part of a broad and j continuing program of research on graduate student selection at Rutgers University. Although this pa per concerns the Department of Applied and Mathe matical Statistics, several other departments are also cooperating, namely the departments of English and Chemistry and the schools of E ng i neering and Agriculture. These results will be presented in forthcoming articles upon their completion. The Department of Applied and Mechanical Sta tistics offers courses leading to the Master of Sci ence degree for part-time evening students employ ed in local industry. Most of these students have returned to academic study after a lapse of several years. For this reason, itwasfelt undergraduate grades did not have the predictive significance they would have for recent graduates and another predic tor was desired, preferably in the form of a con venient test. A total of 150 students enrolled in courses given by this department were tested and complete data for this study were finally obtained for 56 of these students. This constituted the ex perimental group, all of whom had been accepted for graduate study. Three predictors of success were used, namely, undergraduate grade averages and two tests of The Psychological Corporation, restricted to use in graduate student selection, the Miller Analogies Test (1) and the newer Doppelt MathematicalRea soning Test (2). 2 Three criteria of success were employed. They consisted of graduate grades, an academic competence rating, and a job-effective ness rating. Further description of the predictors and criteria is given in the later sections of the ar ticle. In summary, three types of data were gath e r e d : 1) Doppelt and Miller tests admi nistered during the regular classroom period, 2) undergrad uate and graduate grade averages and 3) rankings by instructors of academic competence and job ef fectiveness. Predictors