AN ATTEMPT WAS MADE TO DETERMINE THE PREDICTIVE ACCURACY OF THREE MEDICAL SCHOOL SELECTION DEVICES--PREMEDICAL GRADES, THE MEDICAL COLLEGE ADMISSION TEST (MCAT), AND THE BIOGRAPHICAL INVENTORY FOR MEDICINE (BIM). THE BIM REPRESENTS A NEWLY DEVISED SELECTION INSTRUMENT FOR THE MEDICAL SCHOOL. IN THE STUDY, EACH INSTRUMENT WAS VALIDATED AGAINST IMMEDIATE CRITERIA OF MEDICAL SCHOOL PERFORMANCES AND THE MORE ULTIMATE CRITERIA OF ON-THE-JOB PERFORMANCES. THE MORE ULTIMATE CRITERIA MEASURES EMPLOYED WERE THE RESULTS OF AN EXTENSIVE YEAR-AND-A-HALF STUDY OF ABOUT 800 PRACTICING PHYSICIANS REPRESENTING THREE MAJOR AREAS OF MEDICAL PRACTICE--GENERAL PRACTICE, SPECIALITY PRACTICE, AND MEDICAL SCHOOL FACULTY "PRACTICE." A TOTAL OF 30 FACTOR SCORES (CRITERIA MEASURES) HAD BEEN CALCULATED FROM THE PREVIOUS STUDY. THE STUDY DEMONSTRATED THAT NO SINGLE PREDICTIVE DEVICE NOW IN USE CAN MEET THE STRINGENT REQUIREMENTS OF PREDICTING THE MULTIFACETED DIMENSIONS WHICH COMPRISE TOTAL PHYSICIAN PERFORMANCE. THE INVESTIGATORS RECOMMEND REFINEMENT AND EXTENSION OF BOTH THE CRITERIA AND PREDICTORS. (JC)