Although hundreds of published studies have addressed the effects of religion on morbidity and mortality, many investigators may be unaware of this literature. This paper begins with an analysis of an important subset of these studies--those 27 which operationalize 'religiosity' as religious attendance--and which, taken as a whole, point to a consistent salutary effect for frequent attendance. Upon identifying several pervasive epistemological, methodological, and analytical problems with these studies, however, this paper shows that there is insufficient evidence to conclude that religious attendance is positively and significantly related to health. Nevertheless, the authors present a theoretical basis for expecting such associations. This framework is included in a brief primer on religion for epidemiologists and other sociomedical scientists interested in exploring the health-related effects of religious factors. Finally, a possible scenario for the development of an epidemiology of religion is discussed.