In this essay, Gary Thomas argues that education research repeatedly makes a mistake first noted by Dewey: it misunderstands our science. This misunderstanding has led to attempts to import various putatively scientific precepts into education inquiry. But in reality, he argues, those "scientific" precepts do not characterize scientific endeavor, which is fluid and plural: science flexes to any angle to answer the questions that are posed in any field. Questions in education concern worlds of practice and social relations where change and corrigibility draw the parameters for inquiry. Education research becomes valuable only when it takes account of the reality of the educational endeavor. Thomas urges us to strive to forge a new science of education based on singular and shared understandings of such practice. (Contains 10 notes, 1 table, and 2 figures.)