THE STUDY IS COMPREHENSIVE IN THAT SEVERAL DIFFERENT OPERATIONAL DEFINITIONS ARE EMPLOYED, AN ARRAY OF DATA RELATED TO THE CRITERIA OF RECIDIVISM IS EXAMINED, DATA WERE COLLECTED FROM FIVE DISTINCT TIME PERIODS, AND THE ENTIRE POPULATION WAS STUDIED. GOALS OF ANALYSES INCLUDED COMPARING 13 DIFFERENT OPERATIONAL DEFINITIONS OF RECIDIVISM AND DETERMINING THE FACTORS ASSOCIATED WITH RECIDIVISM AT VARIOUS PHASES TO IDENTIFY THE OPTIMUM POINT IN TIME AT WHICH DATA COLLECTION MAY BE PREDICTIVE OF RECIDIVISM. DATA COLLECTION INVOLVED PSYCHOLOGICAL TESTING OF ENTERING INMATES, INMATE QUESTIONNAIRE EVALUATION, PHYSICAL EXAMINATION, AND STRUCTURED INTERVIEWS CONDUCTED BY A PSYCHOLOGIST. SYSTEMATIC RECORDS WERE MAINTAINED OF EACH INMATE'S ADJUSTMENT THROUGHOUT INCARCERATION. IMMEDIATELY PRIOR TO RELEASE INMATES WERE RETESTED AND AGAIN GIVEN THE VALUES/PRISONIZATION QUESTIONNAIRES. FUNDS WERE NOT AVAILABLE TO PERMIT FOLLOWUP STUDY OF THE RELEASED INMATES IN THE COMMUNITY. THE 13 OPERATIONAL DEFINITIONS OR MEASURES OF RECIDIVISM INCLUDED NUMBER OF ARRESTS, CONVICTIONS, INCARCERATIONS, OFFENSES, SEVERITY OF OFFENSE, AND RECIDIVISM RATE. VARIABLES COLLECTED UPON THE INMATES' ENTRANCE INTO THE SYSTEM, WHETHER DEMOGRAPHIC, TEST, OR OBSERVATIONAL VARIABLES, YIELDED POOR RESULTS REGARDING RELATIONSHIPS INDICATED WITH THE FOUR CRITERIA OF RECIDIVISM, BOTH IN TERMS OF FIRST ORDER CORRELATIONS AND MULTIPLE CORRELATIONS. SIMILARLY, PRERELEASE DATA WERE ALSO UNIFORMLY LOW AND INSIGNIFICANT IN TERMS OF CORRELATION. THE REGRESSION ANALYSIS DID PROVE TO YIELD HIGHER MULTIPLE CORRELATIONS THAN WERE FOUND IN THE DERIVATION SAMPLES. PROBLEMS IDENTIFIED IN THIS STUDY WHICH SHOULD BE CONSIDERED IN FUTURE RESEARCH INCLUDE OBTAINING AN ADEQUATE DATA BASE CONCERNING SUBSEQUENT CRIMINAL BEHAVIOR AND DEVISING DEFINITIONS WHICH PROVIDE GRADATIONS OF SUCCESS RATHER THAN JUST GRADATIONS OF FAILURE. TABLES AND REFERENCES ARE PROVIDED IN THE STUDY. (LWM)