EVALUATIVE RESEARCH STUDIES ATTEMPT TO APPLY METHODS OF SCIENCE TO SERVICE AND ACTION PROGRAMS IN ORDER TO OBTAIN MEASURES OF PROGRAM ACCOMPLISHMENT. THREE ELEMENTS ARE MEASURED TO EVALUATE PROGRAM EFFECTIVENESS: (1) DESIRABILITY OR VALUE OF THE PROGRAM GOAL; (2) THE ABILITY OF THE PLANNED PROGRAM TO INTERVENE AND ACHIEVE THE GOAL; AND (3) THE POSSIBILITY FOR RESEARCHERS TO DETERMINE THE DEGREE TO WHICH THE DESIRED GOAL IS ATTAINED. THE NEED FOR EVALUATION AND ITS FUNCTIONS ARE DISCUSSED. THE EVALUATION CAN TAKE PLACE AT THE RESEARCH, PLANNING, DEMONSTRATION, OR OPERATION LEVEL OF A PROGRAM. HOWEVER, THE MAJOR EMPHASIS OF EVALUATION STUDIES IS ON THE DEMONSTRATION AND OPERATION STAGES, AND THESE TWO STAGES ARE CONSIDERED IN RELATION TO PREFERRED EVALUATION SCHEMES. THE TYPE OF RESEARCH DESIGN, USUALLY AN ADAPTATION OF THE EXPERIMENTAL OR SYSTEMS MODEL, SHOULD RELATE TO THE TYPE OF PROGRAM OR SERVICE BEING EVALUATED. THE CASE STUDY, SURVEY DESIGN, AND PANEL OR PROSPECTIVE STUDY DESIGN ARE DESCRIBED. THE DEFINITION OF THE RESEARCH GOAL OR OBJECTIVE IS THE MOST CRUCIAL STEP IN EVALUATIVE RESEARCH. OBJECTIVES ARE COMMONLY CLASSIFIED ACCORDING TO LEVELS OF GENERALITY RANGING FROM IMMEDIATE TO INTERMEDIATE TO ULTIMATE. DESCRIPTIONS OF INSTITUTIONAL AND ORGANIZATIONAL OBJECTIVES ARE INCLUDED. EVALUATIVE CRITERIA ARE DISCUSSED, AND THE DECISIONMAKING PROCESS IS CONSIDERED ACCORDING TO THE INFORMATION, INTERPRETATION, AND ACTION COMPONENTS. THE EVALUATOR SHOULD LOOK FOR VARIABLES WHICH PERMIT INTERFERENCE WITH THE CAUSAL PROCESS; CRITERIA MAY MONITOR A STEADY STATE ALLOWING RECOGNITION OF PROBLEMS, IDENTIFY ALTERNATIVES IN A PROBLEM SITUATION, WEIGH ALTERNATIVE COURSES OF DECISIONMAKING, OR DETERMINE CORRECTIVE ACTION AND THE ERROR AND RISKS INVOLVED. ADMINISTRATIVE PROBLEMS WILL VARY ACCORDING TO EVALUATION TYPE, AND PROBLEMS COMMON TO SYSTEM-ORIENTED AND PROGRAM-ORIENTED EVALUATIONS ARE DISCUSSED. REFERENCES ARE PROVIDED. (DAG)