A RECIDIVIST IS DEFINED AS ANY RELEASED SUBJECT WHO WAS RETURNED TO A FEDERAL OR STATE CORRECTIONAL INSTITUTION OR TO A COUNTY HOUSE OF CORRECTION OR JAIL FOR 30 DAYS OR LONGER. THE FOLLOWUP PERIOD WAS 1 YEAR FROM THE DATE OF THE SUBJECTS' RELEASE TO THE COMMUNITY. THE ANALYSIS COVERS FIVE CATEGORIES OF VARIABLES: COMMITMENT; PERSONAL BACKGROUND CHARACTERISTICS; CRIMINAL HISTORY; FURLOUGH; AND RECIDIVISM. MANY TRENDS ISOLATED IN EARLIER STUDIES ARE CONTINUING. IN 1966, THE MEAN RATE OF RECIDIVISM WAS 30 PERCENT; IN 1971, 25 PERCENT; IN 1972, 22 PERCENT; IN 1973, 19 PERCENT; AND IN 1974, 19 PERCENT. THE LAST TWO FIGURES SUGGEST A POSSIBLE LEVELING-OFF OF THE DOWNWARD TREND IN RECIDIVISM. OTHER MAJOR FINDINGS OF THE ANALYSIS ARE AS FOLLOWS: (1) FURLOUGH PROGRAM PARTICIPATION APPEARS TO BE AN IMPORTANT FACTOR IN REDUCING RECIDIVISM; (2) PERSONS RELEASED FROM MAXIMUM AND MEDIUM-SECURITY INSTITUTIONS ARE MORE LIKELY TO BECOME RECIDIVISTS THAN ARE PERSONS RELEASED FROM MINIMUM-SECURITY AND PRERELEASE FACILITIES; AND (3) THERE IS A SHIFT AWAY FROM RETURN FOR TECHNICAL VIOLATION OF PAROLE AND TOWARD RETURN AS A RESULT OF A NEW ARREST OR NEW COMMITMENT FROM THE COURTS. IT IS CONCLUDED THAT MASSACHUSETTS' DECREASING RECIDIVISM RATES ARE ATTRIBUTABLE TO INTRODUCTION OF THE FURLOUGH AND GRADUATED RELEASE PROGRAMS, AND TO IMPLEMENTATION OF A MODIFIED PAROLE REVOCATION PROCESS. SUPPORTING DATA ARE INCLUDED. (AUTHOR ABSTRACT MODIFIED).