Abstract We account for a methodological problematique in doing research in mathematics classrooms. We discuss how the definition of unit of analysis varies depending on