EACH APPROACH IS EXAMINED IN TERMS OF ITS CENTRAL QUESTIONS, INTELLECTUAL ORIGINS, MINIMUM KNOWLEDGE REQUIREMENTS, LOGIC FOR DRAWING EVALUATIVE CONCLUSIONS, METHODS FOR ANALYZING DATA, AND THE POLITICAL ROLE OF THE EVALUATOR. THE GOAL-ATTAINMENT APPROACH TRIES TO SEE IF THE PROGRAM GOALS, DEVISING OPERATIONAL MEASURES, AND COLLECTING DATA ON THE EXTENT OF GOAL ACHIEVEMENT. LABORATORY EXPERIMENT METHODOLOGY IS APPLIED, ASSIGNING SERVICES AND GOALS TO THE ROLES OF INDEPENDENT AND DEPENDENT VARIABLES. THE SYSTEMS PARADIGM SEES THE PROGRAM AS AN OPEN SYSTEM WITH MULTIPLE GOALS OF COMPETING FOR RESOURCES. THE EVALUATION TRIES TO DETERMINE IF THE RESOURCES AVAILABLE TO THE PROGRAM ARE BEING OPTIMALLY ALLOCATED. A MODEL LINKS RESOURCE ALLOCATION WITH PERFORMANCE AND IS USED TO PREDICT THE POSSIBLE EFFECTS OF DIFFERENT ALLOCATION PATTERNS. THE DECISION-THEORETIC PARADIGM IS BASED ON PSYCHOLOGICAL THEORIES OF OPTIMAL DECISIONMAKING. A MATRIX INCLUDING ALL POSSIBLE COURSES OF ACTION OPEN ESTIMATES THE PROBABILITIES OF MAXIMIZING VALUES FOR SPECIFIED ACTIONS. THE EVALUATION OF DRUG ABUSE PROGRAMS MAY BE BASED ON THE ASSUMPTIONS THAT SUCH PROGRAMS ARE USUALLY IN A STATE OF CHANGE, THAT NEW TYPES OF TREATMENT ARE CONTINUALLY BEING DEVELOPED, THAT DEBATE WILL CONTINUE TO EXIST OVER APPROPRIATE GOALS AND THEIR ACHIEVEMENT, AND THAT IT IS DIFFICULT TO ISOLATE THE EFFECTS OF A SPECIFIC PROGRAM GIVEN THE REPETITIVE NATURE OF DRUG ABUSERS. OPTIMAL USE OF THE GOAL-ATTAINMENT PARADIGM WOULD BE FOR SMALL SCALE EVALUATION OF NEW TREATMENT TECHNIQUES. THE DECISION-THEORETIC PARADIGM WOULD BE MOST USEFUL WHEN A PROGRAM IS TO UNDERGO A LARGE PLANNED CHANGE EFFORT. THE SYSTEMS PARADIGM IS MOST USEFUL IN THE ONGOING EVALUATION OF A WELLESTABLISHED, IN-PLACE PROGRAM. A TOTAL OF 19 REFERENCES ARE INCLUDED.