SINCE PERSONS SERVING DIFFERENT PERIODS OF TIME MAY NOT BE COMPARABLE PAROLE RISKS, OFFENDERS WERE FIRST CLASSIFIED ACCORDING TO RISK. THEN THE RELATION OF TIME SERVED TO PAROLE OUTCOMES WAS ASSESSED SEPARATELY FOR EACH PAROLE SUBGROUP. AMONG MALES, NINE RISK CATEGORIES WERE DEFINED WITH RATES OF SUCCESS (DEFINED AS AN ABSENCE OF PRISON RETURN DURING THE FIRST YEAR) RANGING FROM 65 TO 99 PERCENT. GENERALLY FOR MEN, SUCCESS RATES DECREASED WITH INCREASING TIME SERVED UP TO 50 MONTHS BUT INCREASED SOMEWHAT THEREAFTER. WHEN THE RISK GROUPS WERE STUDIED SEPARATELY, HOWEVER, EITHER NO RELATION WAS FOUND (FOUR GROUPS) OR NO SINGLE PATTERN OBTAINED (FIVE GROUPS). THE STUDY DEMONSTRATED THAT, AMONG MEN PAROLED BETWEEN 1965 AND 1972, FOR SOME SPECIFIC CLASSIFICATIONS OF OFFENDERS, TIME SERVED IS NOT RELATED TO PAROLE OUTCOME; FOR SOME GROUPS OF OFFENDERS THERE ARE COMPLEX PATTERNS OF ASSOCIATION BETWEEN TIME SERVED AND PAROLE OUTCOMES; AND, THERE CLEARLY IS NO CONSISTENT PATTERN OF INCREASING PAROLE SUCCESS WITH INCREASED TIME SERVED; RATHER, IN GENERAL, SUCCESS RATES DECREASE OR REMAIN FAIRLY CONSTANT WITH INCREASED TIME SERVED IN PRISON. (AUTHOR ABSTRACT)..., BDS