AN INTITIAL GROUP OF 650 SUBJECTS WAS SELECTED FOR STUDY OVER THE PERIOD FROM NOVEMBER 1969 TO FEBRUARY 1975, TO SEE WHETHER THEIR ACTUAL POSTRELEASE BEHAVIOR MATCHED THAT WHICH HAD BEEN EXPECTED. THREE HUNDRED OF THE SUBJECTS WERE THOSE WHO HAD APPLIED FOR PAROLE AND BEEN GRANTED IT. FIFTY WERE A CONTROL GROUP OF THOSE WHO HAD APPLIED FOR PAROLE BUT HAD NOT BEEN GIVEN IT AND HAD TO SERVE THEIR FULL TERMS. ANOTHER CONTROL GROUP OF THREE HUNDRED WERE SUBJECTS WHO DID NOT APPLY FOR PAROLE AND SERVED THEIR TERMS. THESE TWO GROUPS WERE GENERALLY TREATED TOGETHER AS FULL-TERM PRISONERS. AMONG THE MAIN CRITERIA FOR POST-RELEASE BEHAVIOR WERE THE NUMBER OF REARRESTS AND RECONFINEMENTS IN THE GROUP DURING THE 5-YEAR STUDY PERIOD. OF THOSE WHO HAD BEEN FULL-TERM PRISONERS, 67 WERE NOT REARRESTED AND 120 NOT RECONFINED, WHILE AMONG THE GROUP OF PAROLEES, 103 WERE NOT REARRESTED AND 153 WERE NOT RECONFINED. IN BOTH THE FULL-TERM GROUP AND THE PAROLEE GROUP, THERE WERE CASES OF AS MANY AS 6 REARRESTS AND 5 RECONFINEMENTS FOR A PARTICULAR SUBJECT DURING THE 5-YEAR PERIOD, BUT IN EVERY CASE THE PAROLEE FIGURES WERE CONSISTENTLY AND SIGNIFICANTLY LOWER. A DETAILED ACCOUNT OF THE 20 POSITIVE OR NEGATIVE FACTORS GIVEN WEIGHT IN EVALUATION FOR PAROLE IS INCLUDED. AMONG THE 20 FACTORS ARE MEMBERSHIP IN GANGS, PREVIOUS RECORD, MARITAL STATUS, EDUCATIONAL BACKGROUND, EMPLOYMENT RECORD, AND SETTLED RESIDENCE. DETAILED STATISTICAL CHARTS AND GRAPHS ARE INCLUDED, AS WELL AS NOTES GIVING BIBLIOGRAPHICAL REFERENCES. --IN JAPANESE.