PROJECT START SERVED 276 NONVIOLENT PROPERTY FELONS ON PROBATION IN BOTH DETROIT RECORDERS COURT AND WAYNE COUNTY CIRCUIT COURT, MICHIGAN. THE PROJECT WAS DESIGNED WITH EVALUATION IN MIND. THUS TWO FELONS WERE ASSIGNED TO THE START TREATMENT GROUP FOR EVERY ONE ASSIGNED TO THE CONTROL GROUP. AS A RESULT, SUFFICIENT RECORDS WERE AVAILABLE FOR COMPARISON WITH A CONTROL SAMPLE OF 240. THE PROJECT'S LEAA-FUNDED DEMONSTRATION PHASE BEGAN NOVEMBER 1, 1975 AND ENDED FEBRUARY 28, 1978. EVALUATION WILL CONTINUE FOR AN ADDITIONAL YEAR TO ASSESS BOTH POSTPROBATION EFFECTS OF THE ONE-TO-ONE COUNSELING OFFERED BY THE VOLUNTEERS AND THE EFFECTS OF VARIOUS CLIENT CHARACTERISTICS ON RECIDIVISM. BOTH THE TREATMENT AND CONTROL GROUPS WERE RANDOMLY SELECTED AND WERE COMPARABLE ON KEY DEMOGRAPHIC VARIABLES. THIS DATA IS SUMMARIZED IN TABULAR FORM. PART I FOCUSES ON SERVICES DELIVERED TO CLIENTS BY THE PROBATION DEPARTMENT, EMPLOYMENT OF OFFENDERS, AND PROBATION-OFFICER CONTACT. IN ADDITION, EVALUATION OF THE SHORT-TERM EFFECTS OF THE VOLUNTEER INTERVENTION UPON CLIENT RECIDIVISM IS PRESENTED. PART 2 EXAMINES THE CHARACTERISTICS OF THE COMMUNITY VOLUNTEERS, THE VOLUNTEERS' PERCEPTIONS OF CRIME, THEIR PERCEPTION OF THEIR ROLE IN THE PROJECT, AND THE CHARACTERISTICS OF THE VOLUNTEERS WHO COMPLETED THE PROGRAM. THE INSTRUMENT DEVELOPED TO MEASURE VOLUNTEERS' ATTITUDES IS INCLUDED. THE INITIAL EVIDENCE SUGGESTS THAT PROJECT START SIGNIFICANTLY LOWERED THE RECIDIVISM RATES OF ITS CLIENTS, AS MEASURED BY WARRANTS, ARRESTS, AND CONVICTIONS. IN THE START GROUP 4 PERCENT HAD WARRANTS SWORN AGAINST THEM, 5 PERCENT WERE ARRESTED, AND 22 PERCENT WERE CONVICTED COMPARED TO 12 PERCENT, 7 PERCENT, AND 28 PERCENT RESPECTIVELY IN THE CONTROL GROUP. PROJECT START APPEARED TO SHOW MORE EFFECTS IN RECORDERS COURT THAN CIRCUIT COURT. A COMPARISON OF THE TWO CONTROL GROUPS SHOWED NO DIFFERENCE IN RECIDIVISM BETWEEN THE TWO COURTS AMONG REGULAR PROBATIONERS, SUGGESTING BOTH COURTS SERVE THEIR CLIENTS EQUALLY WELL. HOWEVER, THERE ARE NOTABLE DIFFERENCES IN DEMOGRAPHIC CHARACTERISTICS IN PROBATIONERS BETWEEN THE TWO COURTS, LEADING TO THE HYPOTHESIS THAT THESE DIFFERENCES INFLUENCE THE VOLUNTEERS' SUCCESS. FURTHER INVESTIGATION OF THIS IS SUGGESTED. A PRELIMINARY ANALYSIS INDICATES START COSTS $71 LESS PER CLIENT THAN REGULAR PROBATION. STATISTICAL DATA, DEMOGRAPHIC DATA FOR BOTH PROBATIONERS AND VOLUNTEERS, AND LISTS OF SERVICES PROVIDED BY BOTH THE REGULAR PROBATION DEPARTMENTS AND THE VOLUNTEERS ARE INCLUDED. (GLR)