Understanding and predicting the yield in medical school admissions

Graduate and professional schools in general, and medical schools in particular, have traditionally not paid a great deal of attention to applicant "yield"--the proportion of accepted applicants who eventually confirm their intention to matriculate. This study examined data for all students who were accepted for the years 1985-1987 at an urban, private medical school in the Northeast. The author attempted both to estimate the probability of an accepted applicant's matriculating at the institution and to demonstrate the relative effects of selected variables on the applicant's final decision. Through the use of the maximum-likelihood estimation technique, the resulting model indicated that probability ranges for matriculation may be derived using data available from computerized student records. The model was able accurately to classify more than 70% of the students into either the "matriculant" or the "non-matriculant" category.