UTILIZATION, EVALUATION'S WEAKEST LINK, DEPENDS ON MANAGEMENT PERSONNEL'S ADEQUATE CONCEPTUAL PREPARATION. THE DECISIONMAKING PRACTICED BY MANAGERS HAS BEEN TERMED 'SATISFACTING,' OR MANAGEMENT BY CRISIS,' RESEARCH HAS BEEN MORE A HINDRANCE THAN AN AID. TO HELP IMPROVE DECISIONMAKING, THE EVALUATOR SHOULD PROVIDE MANAGEMENT WITH RELIABLE ALTERNATIVE PROJECTIONS. A MULTIATTRIBUTE UTILITY MEASUREMENT IS NEEDED TO AGGREGATE ASSUMPTIONS ABOUT A PROGRAM AREA INTO A MATRIX MODEL. ITS AIM WOULD BE STRUCTURING AN EXTENDED RANGE OF PROJECTIONS TO IDENTIFY IMPLICATIONS RATHER THAN A DEFINITIVE ANSWER. THE METHOD PROVIDE AN ONGOING, INTERACTIVE PROCESS OF PROJECTIONS AND FEEDBACK. THE PROPOSED MODEL HAS THREE SECTIONS: THE POLITICAL CONTEXT, FEASIBILITY EXPECTATIONS, AND OUTCOME EXPECTATIONS. FEASIBILITY FACTORS ARE THOSE THAT HINDER OR SUPPORT PROGRAM IMPLEMENTATION AND ARE USED TO CLARIFY PRIORITIES. DATA COLLECTION FOR EVOLVING THE MATRIX INVOLVES SEVEN STEPS, STARTING WITH QUESTIONNAIRE RESPONSES OF INVOLVED PERSONNEL AND GENERATING ASSUMPTIONS ON A CASE-BY-CASE BASIS. THESE ASSUMPTIONS ARE COMPILED AND EVALUATED BY RESPONDENTS. KEY ASSUMPTIONS ARE CRYSTALLIZED, AND A PLAN FOR COLLECTING TEST DATA IS GIVEN THE RESPONDENTS. VARIOUS POSSIBLE RESULTS FOR THE DATA COLLECTIONS ARE PROJECTED ALONG WITH THE IMPLICATIONS FOR THE MODEL. RESPONDENTS ARE ASKED HOW THEY WOULD TAKE ACTION IN RESPONSE TO EACH PROJECTION, AND A DATA COLLECTION IMPLEMENTED. COLLECTED DATA ARE ANALYZED IN RELATION TO PROJECTIONS, RESPONDENTS ARE SOLICITED FOR CHANGES IN THEIR ASSUMPATIONS AND THE DATA COLLECTION IS CONTINUED TO ASSESS VARIATIONS OVER TIME OR WITH ACTION CHANGES. EVENTUALLY, AS THE INTERRELATIONSHIPS BETWEEN PERFORMANCE AND OUTCOME BEGIN TO BE SUGGESTED, A NEW SET OF BASIC QUESTIONS IS DEVELOPED AND THE PROCESS BEGINS ANEW. THIS MODEL SHOULD DECREASE INAPPROPRIATE DECISIONS BASED ON EASING THE PROGRAM'S PRESSURES. TABLES AND REFERENCES ARE INCLUDED. (MRK)