Many studies have tried to associate water supply and sanitation facilities with health status. The published results are confusing and contradictory because of methodological problems. This review attempts to establish a cause-effect relationship between water and sanitation and diarrheal morbidity, nutritional anthropometry, and mortality by systematically assessing reports according to research design and health outcome. It finds that excreta disposal facilities are more effective than improved water supplies for lowering diarrheal morbidity and mortality rates and for improving child growth. It also finds that water supply improvements by themselves may have limited effectiveness in improving health. A conceptual framework is developed to reduce methodological problems when evaluating present and future water/sanitation programs. Appended are 7 tables and 4 figures and a 69-item bibliography (1952-84). (Author abstract, modified).