Predicting medical specialty choice: a model based on students' records

A discriminant analysis of objective and subjective measures from the records of 628 students who graduated from the University of Medicine and Dentistry of New Jersey-New Jersey Medical School over a six-year period was used to generate a model for the prediction of medical specialty choice. The authors found that National Board of Medical Examiners Part II examination scores, sex, race, grades given by preceptors, and a score derived from narrative comments by preceptors during clinical experiences in psychiatry contained information for predicting such a choice. With this model, the correct prediction rate for all specialties was 41 percent. The correct prediction rate for individual specialties ranged from a low of 28 percent for family practice to a high of 68 percent for psychiatry.