This paper re‐examines Kurt Lewin's classic leadership studies, using them as a concrete example to explore his wider legacy to social psychology. Lewin distinguished between advanced “Galileian” science, which was based on analysing particular examples, and backward “Aristotelian” science, which used statistical analyses. Close examination of the way Lewin wrote about the leadership studies reveals that he used the sort of binary, value‐laden concepts that he criticised as “Aristotelian”. Such concepts, especially those of “democracy” and “autocracy”, affected the way that he analysed the results and the ways that later social scientists have understood, and misunderstood, the studies. It is argued that Lewin's famous motto—“there is nothing as practical as a good theory”—is too simple to fit the tensions between the leadership studies and his own views of what counts as good theory.