THE CHANGE ALTERED THE PRIMARY ROLES OF PROBATION TO INCLUDE MAKING CLIENT NEED ASSESSMENTS, REFERRING CLIENTS TO APPROPRIATE COMMUNITY AGENCIES, AND MONITORING THE PROGRESS OF CLIENTS IN COURT-ORDERED PROGRAMS. SINCE ALL CLIENTS DID NOT REQUIRE THE SAME SERVICES, IT WAS DECIDED THAT THEY ALSO DID NOT REQUIRE THE SAME AMOUNT OF SUPERVISION OR MONITORING. THE TOTAL PROBATION CASELOAD WAS DIVIDED INTO CATEGORIES BASED ON THE NEED FOR HIGH OR LOW SUPERVISION. THIS CLASSIFICATION TOOK PLACE AT THE TIME PRESENTENCE REPORTS WERE PREPARED. BECAUSE OF THE LACK OF OBJECTIVITY IN THE CLASSIFICATION METHOD, IT WAS DETERMINED THAT A MORE STANDARD WAY OF DETERMINING HIGH-RISK AND LOW-RISK CLIENTS SHOULD BE USED BY PROBATION COUNSELORS. A POINT SYSTEM WAS ESTABLISHED BY WHICH EACH DEFENDANT COULD BE RATED AND THEN ASSIGNED TO A RISK CATEGORY. ALL DEFENDANTS REFERRED BY THE COURTS FOR PRESENTENCE REPORTS BETWEEN MARCH AND AUGUST 1974 WERE SCORED TO PREDICT RECIDIVISM. A SMALL RANDOM SAMPLE OF 88 INDIVIDUALS WAS THEN SELECTED, AND ARREST CHECKS WERE MADE TO FIND OUT IF ANY HAD BEEN REARRESTED FOR NEW OFFENSES SUBSEQUENT TO STARTING PROBATION. MULTIPLE REGRESSION ANALYSIS WAS EMPLOYED TO IDENTIFY FACTORS IN THE SCORING SYSTEM THAT PREDICTED RECIDIVISM. TO ASCERTAIN THE VALIDITY OF THE REARREST RISK PREDICTION SCALE, A 9-MONTH FOLLOWUP WAS PERFORMED. OF ALL HIGH-RISK PERSONS, 24.1 PERCENT WERE REARRESTED AT LEAST ONCE WITHIN THE FOLLOWUP PERIOD. ONLY 10.7 PERCENT OF LOW-RISK PERSONS WERE REARRESTED. FOR THE 142 HIGH-RISK RECIDIVISTS, THERE WAS A MEAN AVERAGE OF 1.50 REARRESTS. THE 33 LOW-RISK RECIDIVISTS HAD A MEAN AVERAGE OF 1.27 REARRESTS. THE STUDY FINDINGS DEMONSTRATE THAT MISDEMEANANT PROBATIONERS CAN BE OBJECTIVELY CLASSIFIED ON THE BASIS OF RECIDIVISM RISK. SUPPORTING DATA ARE PROVIDED. APPENDIXES CONTAIN THE INITIAL AND REVISED SCORING FORM AND A LIST OF SPECIFIC CHARGES WITHIN OFFENSE CATEGORIES. REFERENCES ARE CITED. (DEP)