ABSTRACT
A key problem in sociology and social anthropology is the relationship between theory and research. Some of the discussions concerning this relationship in social science have been oversimplified. Many social science textbooks on methodology have begun by outlining an ideal scientific procedure of: theory-developing hypotheses-data collection-testing of hypotheses-conclusion. Such a procedure, it is claimed, occurs in the natural sciences and can be applied to research in the social sciences. However, such discussions do not question the relevance of this research procedure for the natural sciences. Indeed, Medawar (1964) has suggested that the way in which scientific papers are written up constitutes a fraud, as the actual process of doing research follows a different pattern from that represented in the papers. He indicates that natural science can be a creative and imaginative activity, in which hypotheses appear along uncharted paths in the research process. Among social scientists, similar remarks have been made by Wright Mills (1959), who criticises empirical researchers who surround their data with ‘theory’ after data collection has been completed. Such procedures, he argues, can mislead a reader into thinking that the study was specifically designed and conducted to test broader conceptions.
