IN CORRECTIONS RESEARCH THE BASE EXPECTANCY TABLE HAS BEEN DEVELOPED TO ESTIMATE THE VARYING DEGREES TO WHICH INDIVIDUALS IN A GIVEN PRISON POPULATION OR TREATMENT GROUP ARE AT RISK OF CONTINUING THEIR CRIMINAL CAREERS AFTER RELEASE. IT IS A CLASSIFICATION TECHNIQUE FOR ASSIGNING INDIVIDUALS TO RISK GROUPS. IN EVALUATING PROGRAMS, RECIDIVISM RATES FOR PERSONS WITH SIMILAR RISKS BUT WHO DID NOT RECEIVE THE TREATMENT ARE COMPARED WITH RATES FOR THOSE IN THE TREATMENT PROGRAM. THE HISTORY OF THE DEVELOPMENT OF BASE EXPECTANCY TABLES IS GIVEN AND REFINEMENTS MADE OVER THE YEARS EXPLAINED. THE TABLES USED BY THE MASSACHUSETTS DEPARTMENT OF CORRECTIONS ARE EXPLAINED IN DETAIL. RECIDIVISM RATES ARE THEN DEVELOPED FOR THE BOSTON STATE PRE-RELEASE PROGRAM AND THE SHIRLEY PRE-RELEASE PROGRAM. VARIABLES USED IN ANALYSES ARE COMMITMENT VARIABLES, PERSONAL BACKGROUND CHARACTERISTICS CRIMINAL HISTORY, AGE AT RELEASE COMBINED WITH LENGTH OF TIME SERVED ON PRESENT INCARCERATION AND TYPE OF RELEASE, AND A RECIDIVISM VARIABLE. APPENDIXES EXPLAIN THESE VARIABLES IN DETAIL AND GIVE THE POINT VALUES ASSIGNED TO EACH. COMPUTATIONS FOR THE PRERELEASE PROGRAMS ARE GIVEN. THESE FIGURES SHOW THAT THE EXPECTED RECIDIVISM RATE OF THE SHIRELY GROUP WAS 30.9 PERCENT WHILE THE RATE OF THE TREATMENT GROUP WAS 24 PERCENT. FOR THE BOSTON STATE SAMPLE THE EXPECTED RATE WAS 21.5 PERCENT, WHICH WAS LOWER THAN RATE OF THE TREATMENT GROUP (24 PERCENT). THIS STUDY CONCLUDES THAT THE PROGRAM COMPLETION SAMPLES DO NOT DIFFER SIGNIFICANTLY FROM GROUP IN TERMS OF RECIDIVISM RISKS FOR THEIR POPULATION AND IN THE FUTURE RANDOM SELECTION CAN BE USED WHEN ASSESSING PROGRAM EFFECTS. (GLR)