THE MULTIPLE REGRESSION ANALYSIS MODEL WAS APPLIED FOR PREDICTING RECIDIVISM, AND FUNDAMENTAL VALUES FOR STATISTICAL ANALYSIS WERE FOUND IN ORDER TO DEVELOP THE SELECTION OF SUBJECT GROUPS AND ITEMS. DISTANCE BETWEEN GIVEN DATA AND THE CENTER OF EACH GROUP WAS CALCULATED, AND SINGULAR DATA WERE VERIFIED WITH THE PROBABILITY THEORY. SUBJECTS USED IN THE STUDY WERE YOUTHS RELEASED FROM THE JUVENILE CLASSIFICATION HOMES; 254 WERE JUVENILES UNDER PROBATIONARY SUPERVISION, AND 181 WERE JUVENILES SENT TO THE JUVENILE TRAINING SCHOOL. DATA FOR THE STUDY WAS GATHERED FROM THE STAFF AT THE PROBATION OFFICE AND JUVENILE TRAINING SCHOOL, FROM CARDS FOR STATISTICAL USE OF CLASSIFICATION, AND FROM THE SURVEY MADE DURING 1972 RELATING TO RECIDIVISM OF SUBJECTS. FINDINGS SHOW THAT JUVENILES UNDER PROBATIONARY SUPERVISION HAD A HIGH RECIDIVISM RATE, PARTICULARLY IN THE LATER STAGES OF PROBATION. PROBLEMS OF PERSONALITY, INCLUDING ATTITUDES TOWARD OTHERS, RECORDS OF MISBEHAVIOR, AND ENVIRONMENTAL PROBLEMS SHOULD BE CAREFULLY TAKEN INTO CONSIDERATION WHEN TREATING JUVENILES. THE GROUP OF SUBJECTS FROM THE TRAINING SCHOOL SHOWED A HIGHER RATE OF RECIDIVISM AMONG INMATES WHO HAD BEEN ADMITTED TO THE HOME MORE OFTEN AND HAD HAD PREVIOUS TROUBLE WITH AUTHORITIES IN SCHOOL, IN VOCATIONAL OR ECONOMIC LIFE, AND WITH LEGAL AUTHORITIES. THE LENGTH OF THE PERIOD BETWEEN THE FIRST DISPOSITION GIVEN BY THE AUTHORITY AND THE PRESENT CASE SHOWS A CLOSE RELATION WITH EFFECTS OF TREATMENT AT THE JUVENILE TRAINING SCHOOL. THE MAIN EMPHASIS IN DEALING WITH YOUTH HAVING GRAVE PERSONALITY PROBLEMS SHOULD BE REFORMATION OF PERSONALITY THROUGH THERAPY, BUT THE MORE SERIOUS THE DISTURBANCE, THE MORE DIFFICULT TREATMENT WILL BE. IT WAS FOUND THAT GOOD ENVIRONMENTAL SITUATIONS AFTER THE RELEASE FROM TRAINING SCHOOL ARE FACTORS IN THE PREVENTION OF RECIDIVISM. (DAG)