THE EVALUATION SOUGHT TO ASSESS THE OVERALL IMPACT OF THE PROJECT ON RECIDIVISM (DEFINED AS 'RETURNS TO THE STATE CORRECTIONS DIVISION'). THE STUDY COMPARED 'MATCHED' SAMPLES COMPOSED OF PROJECT AND PREPROJECT GROUPS AND CONCLUDED THAT THE RATE OF RECIDIVISM WAS LOWER AMONG PROJECT CLIENTS AND THAT PROJECT CLIENTS WHO DID RECIDIVATE DID SO WITHIN A SHORTER PERIOD OF TIME FOLLOWING RELEASE. A SECOND PART OF THIS REPORT PRESENTS DATA ON DISCHARGEES, SERVICES, AND RECIDIVISM RATES. AMONG THE MORE NOTABLE FINDINGS, WERE THAT THE MOST FREQUENT NEEDS AT TIME OF RELEASE RESULTING IN ACCEPTED REFERRALS OR SERVICES ACTIONS WERE RELATED TO EMPLOYMENT AND HOUSING SHELTER. RECIDIVISM RATES ARE NOTED TO BE CORRELATIVE TO AGE AT FIRST ARREST, INSTITUTION FROM WHICH RELEASED, AND TYPE OF OFFENSE, AND COMMUNITY-BASED SUBSIDIES. RETURNS WERE SIGNIFICANTLY HIGHER AMONG INMATES WHO RECEIVED COMMUNITY SUBSIDIES THAN AMONG THOSE WHO DID NOT. HOWEVER, AMONG THOSE CLIENTS WHO WERE EMPLOYED, THE RATES OF RECIDIVISM WERE ALMOST ENTIRELY THE SAME FOR THOSE RECEIVING SUBSIDIES AND THOSE NOT RECEIVING SUBSIDIES. AMONG RECIDIVISTS, THERE WAS A DIRECT CORRELATION BETWEEN EMPLOYMENT STATUS AT RELEASE AND THE NUMBER OF 'DAYS OUTSIDE:' THOSE EMPLOYED TENDED TO 'STAY OUT' LONGER.