This article focuses on the role of scientific inference in the study of bureaucracy. Its focus, in particular, is on the microand neoinstitutional foundations of choice. The context for scintific inference is that offered by King, Keohane, and Verba (KKV) in their book, Designing Social Inquiry. The article notes that KKV essentially stipulates an idea of normal scientific inquiry to allow for the evaluation of competing empirical claims. Two fundamental research programs, those of bounded rationality and rational choice, are assessed for their theoretical and scientific contributions to the study of bureaucracy. The latest turns in the bounded rationality program for organizational behavior have moved to some degree from that program`s roots in cognitive psychology to a form of cultural anthropology, while rational choice models have become more sensitized to information costs. Thus, some work originating in the rational choice mode has come to borrow significantly from ideas within the bounded rationality paradigm. Rational choice models also tend to emphasize principal-agent relationships, the most interesting aspects of which, however, are mainly rooted in the peculiarities of Americanl political institutions. Although neoinstitutionalist rational choice theories of bureaucracy have develope a clear and compelling normal science program allowing disputes to be taken out of the realm of theology and into the realms of specification, measurement, and other issues associated with empirical claims, evidence also persists to suggest limit to economic theories of organization and a role for leadership and norms