Predicting Academic Performance of MBA Program Applicants*

We report on an approach for aiding and strengthening the MBA admissions process that was used at the Stanford University Graduate School of Business. Multiple regression models were used to make predictions of the academic performance of an applicant, if admitted to the MBA program. Applicants who were predicted to have a “substantial chance” of an inadequate academic performance were considered further only if a detailed reading of the application indicated exceptional circumstances or characteristics not adequately captured by the models. From the remaining applicants, selection was made by the admissions officer(s) based mainly on management potential. The present paper focuses on the prediction of academic performance. We discuss in detail the development of criterion variables, the specification of predictor variables, and the development, estimation and validation of models to predict academic performance as well as issues associated with the implementation of the approach.