THE PROJECT CONSIDERED INTENDED TO MAKE A VARIETY OF INHOME SERVICES AVAILABLE TO SERIOUSLY DISABLED, INDIGENT PERSONS. BEYOND THIS FUNDAMENTAL COMMITMENT, SPECIFIC CHOICES REGARDING THE DEVELOPMENT OF THE PROJECT WERE LEFT TO RESEARCH METHOLOLOGISTS IN THE INTERESTS OF DEVISING A PROGRAM WHOSE OUTCOMES COULD BE PRECISELY MEASURED. WHILE PROJECT DEVELOPMENT WAS GUIDED BY EVALUATION RESEARCH REQUIREMENTS, HOWEVER, CERTAIN SOCIAL POLICY IMPLICATIONS WERE NEGLECTED TO THE DETRIMENT OF EFFECTIVE PROGRAM DEVELOPMENT. THE PARTICULARS OF THIS PROBLEM ARE DISCUSSED. IT IS INDICATED THAT EVALUATION RESEARCHERS PLAY A USEFUL ROLE IN PROGRAM DEVELOPMENT ONLY IF THEIR WORK IS ADDRESSED TO THE MOST EFFECTIVE IMPLEMENTATION OF THE PROJECT AS A SOCIAL SERVICE. IF INVESTMENT IN EVALUATION RESEARCH IS TO BE PRODUCTIVE, POLICYMAKERS SHOULD SPECIFY THE INFORMATION THEY HOPE TO DERIVE FROM INNOVATIVE INTERVENTIONS AND THEN PARTICIPATE IN NEGOTIATIONS WITH EVALUATION RESEARCHERS REGARDING BASIC EVALUATION STRATEGY. SO THAT ADHERENCE TO RIGID EVALUATION DESIGN DOES NOT CRIPPLE OR DISTORT THE DEVELOPMENT OF A PROJECT AT ITS OUTSET, ALTERNATE EVALUATION STRATEGIES ARE SUGGESTED. A DELAYED OR PHASED APPROACH TO THE IMPLEMENTATION OF EVALUATION PROCEDURES IS SUGGESTED AS A WAY OF AVOIDING THE SOMETIMES OVERBURDENING TASK OF DEVELOPING PROGRAM ELEMENTS AND SOPHISTICATED EVALUATION DESIGN SIMULTANEOUSLY. THE HAZARDS TO EVALUATION OF SUCH AN APPROACH ARE ACKNOWLEDGED, AND WAYS OF COUNTERING THE MAJOR DANGERS ARE SUGGESTED. (RCB)