Analytic induction (AI) is a label Florian Znaniecki invented in the 1930s for a particular conception of the process of scientific inquiry. This conception was later developed and applied by Alfred Lindesmith, Donald Cressey, Howard S. Becker, and others, and has been subject to considerable variation in interpretation since that time. It was originally set up in opposition to “statistical method,” which Znaniecki labeled “enumerative induction,” the use of which became more prevalent, and then dominant, in sociology during the first half of the twentieth century. AI focused on the development and testing of explanatory theories through the investigation of individual cases, rather than statistical analysis of data at the aggregate level. It drew on philosophical ideas about scientific method, such as those of Aristotle and Mill. It was the focus for considerable debate among some US sociologists in the 1950s but since that time has attracted rather less attention, despite occasional articles dealing with it or seeking to apply it. To a large extent, it came to be overshadowed not just by developments in quantitative analysis but also by the widespread influence of grounded theory, with which it shares some features while also differing in significant ways. It is an approach which engages with difficult issues about the nature of both social causality and sociological theory.