THE DIFFERENCE BETWEEN PROGRAM EFFECTIVENESS AND PROGRAM EFFICIENCY IS DISCUSSED AND THE USE OF EVALUATION TO IMPROVE BOTH IS SUGGESTED. VARIOUS TYPES OF EVALUATION ARE BRIEFLY REVIEWED, INCLUDING CLASSICAL 'BEFORE AND AFTER' EXPERIMENTAL DESIGN AND ONGOING ANALYSIS OF A PROGRAM IN OPERATION. IT IS CONCLUDED THAT EVALUATION CAN BE OF GREAT USE TO AN ONGOING PROGRAM, AND THIS THESIS IS THEN ILLUSTRATED BY EXAMINING SEVERAL EVALUATIONS OF CALIFORNIA PROGRAMS. AN EVALUATION OF EARLY DISCHARGE FROM PAROLE FOUND THAT THE PROGRAM WAS EFFECTIVE BECAUSE RECIDIVISM DID NOT INCREASE AND PAROLE PERSONNEL TIME WAS CONSERVED. THE SECOND STUDY INVOLVED A SHORT-TERM RETURN PROJECT, IN WHICH PAROLE VIOLATORS WHO ONLY NEEDED A SHORT TIME BACK IN PRISON TO 'DRY OUT' FROM ALCOHOL OR GET A DRUG PROBLEM UNDER CONTROL WERE RETURNED TO THE STREETS AS SOON AS POSSIBLE. THIS FOUND THAT 80 TO 90 PERCENT OF PAROLEES RETURNED AS TECHNICAL VIOLATORS, IN CONTRAST TO THOSE WITH NEW FELONY COMMITMENTS, COULD BE SAFELY RETURNED TO THE COMMUNITY IN 3 TO 6 MONTHS. ONGOING EVALUTION ALERTED OFFICIALS TO THE POTENTIAL FAILURE OF A REVISED SYSTEM FOR HANDLING PAROLE CASES WHICH INVOLVED HIGHER CASE LOADS FOR EACH PAROLE OFFICER. THE PROGRAM WAS DROPPED. THE FINAL CASE STUDIED INVOLVED PRESENTENCE EVALUATION, SEVERAL ADULT AUTHORITY ACTIONS, AND A NUMBER OF JUVENILE DIVERSION PROGRAMS. IN EACH CASE, SHIFTS IN POLICY WERE MADE AS THE RESULT OF EVALUATION. TABLES PRESENT DATA FOR EACH STUDY. REFERENCES ARE APPENDED. (GLR)