Assessment of Admission Criteria for Predicting Students' Academic Performance in Graduate Business Programs

Abstract

This study is concerned with quantitative measures currently used or having the potential to be used in the student admission processes for graduate schools of business to produce the best academic performance by admitted students. Methods used to assess the importance of various admission criteria were classified into traditional (linear and nonlinear regression methods) and nontraditional (neural network) multivariate data analysis techniques. In the results for comparing methods, one familiar test, measuring the residual error on a common data set, suggests that neural networks may provide a better predictive model of admitted students' academic performance than traditional quantitative methods of data analysis.