THE MAIN THESIS OF THE PAPER IS THAT THE FORMAL THEORY OF DECISIONMAKING CAN BE APPLIED TO PROBLEMS OF CORRECTIONAL DECISIONS AND, IN PARTICULAR, TO PROBATION DECISIONS. AT EACH DECISION POINT, IT IS ASSUMED THAT DECISIONMAKERS (PROBATION OFFICERS) USE ALL AVAILABLE RELEVANT INFORMATION TO REACH A DECISION ABOUT JUVENILES. THE DECISION MODEL USES TWO SETS OF CRITERION CATEGORIES: THE NUMBER OF OFFENSES AFTER PROBATION DISPOSITION (RECIDIVISM) AND A MEASURE THAT RATES YOUTH ON A BEHAVIOR IMPROVEMENT-DETERIORATION SCALE. AN ONLINE EXPERIMENTAL COMPUTER SYSTEM DESIGNED TO IMPROVE THE PROBATION PROCESS HAS BEEN DEVELOPED. IT IS KNOWN AS SIMBAD (SIMULATION AS A BASIS FOR SOCIAL AGENT DECISIONS) AND ALLOWS INFORMATION ABOUT EACH YOUTH TO BE COMBINED AND PRESENTED TO PROBATION OFFICERS AT THEIR REQUEST. THE BASIS FOR THE DEVELOPMENT OF THE SYSTEM INVOLVED DATA OBTAINED ON 2,290 JUVENILES REFERRED TO PROBATION DEPARTMENTS IN SEVEN CALIFORNIA COUNTIES. ONE COMPONENT OF THE SYSTEM IS A BAYESIAN DECISION MODEL INTENDED TO REDUCE THE UNCERTAINTY OF PROBATION DECISIONMAKERS BY ALLOWING THEN TO SELECT WHATEVER DISPOSITION-TREATMENT ALTERNATIVE WILL HAVE THE HIGHEST PROBABILITY OF SUCCESS. SUPPORTING DATA ON THE CALIFORNIA STUDY AND EQUATIONS EMPLOYED IN THE COMPUTER-AIDED PROBATION DECISIONMAKING MODEL ARE PROVIDED. (DEP)