THE GROUP COMPRISED 69 SUCCESSFUL AND 33 UNSUCCESSFUL WORK RELEASE INMATES WHO WERE IN THE PROGRAM ON OR BEFORE FEBRUARY 1, 1969. THE VARIABLES WERE SELECTED FROM THOSE WHICH HAD PROVED MEANINGFUL PREDICTORS OF PAROLE PERFORMANCE FOR MALE FELONS, AS INDICATED BY A REVIEW OF THE LITERATURE, OR WERE IMPORTANT VARIABLES FOR THE SCREENING OF WORK RELEASE CANDIDATES AT LA TUNA. THESE INCLUDED TEST SCORES, PERSONAL HISTORIES, PREVIOUS CRIMINAL CONVICTIONS, TYPE OF PREVIOUS CRIMINAL ACTIVITY, EDUCATIONAL CLASSES/CLUBS PARTICIPATED IN AT THE PRISON, AND WAGES ON THE PROGRAM. THESE WERE GROUPED INTO 'INTELLIGENCE,' 'PERSONALITY,' 'PAST HISTORY,' 'PRESENT HISTORY,' AND 'AGE' VARIABLES. STATISTICAL ANALYSIS SHOWED THAT THE MOST SIGNIFICANT VARIABLES IN EACH GROUP WERE: (1) INTELLIGENCE, GATB-V APTITUDE AND BETA IQ (INTELLIGENCE QUOTIENT), SCORE 3 (2) PERSONALITY, MMPI-PD SCALE; (3) PAST HISTORY, LONGEST TIME ON ONE JOB, HIGHEST SCHOOL GRADE COMPLETED, PRIOR ARRESTS, AND LONGEST PRIOR SENTENCE; (4) PRESENT HISTORY, WAGES ON WORK RELEASE; AND (5) AGE, AGE AT WHICH FIRST ARRESTED/RECEIVED. THE VARIABLES WERE STUDIED BY MULTIPLE LINEAR REGRESSION TO DEVELOP A SERIES OF PREDICTION TABLES. IT WAS FOUND THAT PRESENT AGE AND BETA IQ COULD BE ELIMINATED WITHOUT SIGNIFICANT CHANGE. A PREDICTIVE TABLE IS DEVELOPED. THE STATISTICAL ASPECTS OF THE STUDY ARE PRESENTED IN DETAIL AND THE RESULTS OCCUPY A SERIES OF TABLES. THE FINAL PREDICTIVE TABLE IS INCLUDED, TOGETHER WITH WEIGHTED SCORES FOR THE VARIABLES. IT IS CONCLUDED THAT A USEFUL PREDICTIVE TABLE BASED ON MULTIPLE LINEAR REGRESSION COEFFICIENTS COULD BE PRODUCED TO AID CLASSIFICATION COMMITTEES IN SCREENING INMATES FOR WORK RELEASE AT THE FEDERAL CORRECTIONAL INSTITUTION IN LA TUNA. AN EXTENSIVE BIBLIOGRAPHY IS APPENDED. (GLR)