BASED ON THE FINDINGS OF RESEARCH PSYCHOLOGISTS CONCERNING THE INTERACTION BETWEEN HUMAN DECISIONMAKERS AND THE MODE OF PROVISION OF INFORMATION TO THEM AND PRINCIPLES OF MODERN DECISION THEORY, THE FOLLOWING SEQUENCE OF STAGES IS PROPOSED FOR CORRECTIONAL DECISIONMAKING: (1) THE EVIDENCE AVAILABLE TO DECISIONMAKERS SHOULD BE PRESENTED TO THEM SEQUENTIALLY; (2) DECISIONMAKERS SHOULD BE ASKED TO RECORD THEIR ESTIMATE OF THE SIGNIFICANCE OF EACH PIECE OF EVIDENCE RELATIVE TO THE DECISION IN QUESTION; (3) THESE WEIGHTING ESTIMATES WILL BE MADE FOR ALL THE AVAILABLE DATA ITEMS CONSIDERED RELEVANT TO THE DECISION; (4) THE WEIGHTING ESTIMATES, WHICH ARE IN THE NUMERICAL FORM OF A LONG FRACTION, WILL THEN BE MULTIPLIED OUT; AND (5) MAKE AS LIBERAL A DECISION AS POSSIBLE FOR THE INMATE, WITHIN THE BOUNDARY CONDITIONS OF THE VALUES OF SOCIETY. THE OPERATIONAL HYPOTHESIS UNDERLYING THIS DECISIONMAKING STRUCTURE MAINTAINS THAT A SIMPLE QUANTIFICATION PROCEDURE OF PROBABILITIES OF SUCCESS AND SOCIAL VALUE UTILITIES, COMBINED IN A BAYESIAN MANNER, PRODUCES A HIGHER DEGREE OF CONSISTENCY OF DECISION AMONG CORRECTIONAL DECISIONMAKERS THAN THE NORMAL UNSTRUCTURED DECISION PROCESS; FOR THEORETICAL REASONS, THIS ACHIEVED LEVEL OF DECISION AGREEMENT WILL FAVOR THE OPTIMAL DECISION. IN A TESTING OF THE PROPOSED DECISIONMAKING MODEL ON THREE IMPORTANT CORRECTIONAL DECISIONS IN CALIFORNIA, THE CHANGE TO A FORMALIZED STRUCTURED METHOD PRODUCED NO DETERIORATION IN CONSISTENCY ACROSS DECISIONMAKERS, AND IT WAS THE OPINION OF MOST OF THE SUBJECTS, THAT, GIVEN TIME TO BECOME FAMILIAR WITH THE NEW SYSTEM, IT WOULD BOTH IMPROVE THEIR OWN UNDERSTANDING OF THEIR DECISION PROCESSES AND CONTRIBUTE TO CLARITY AND CONSISTENCY IN THE DECISIONS AMONG DIFFERENT PERSONNEL. REFERENCES ARE PROVIDED. (RCB)