A foundational principle of ethnomethodological and conversation-analytic (EM/CA) research is that if people's interactions with one another are to be studied, they must be recorded as faithfully and as fully as possible. In practice, this principle throws up ethical challenges at every stage of a typical EM/CA study: negotiating access to the scene; recruiting stakeholders and participants; soliciting their informed consent and attending to their expectations; recording interactions; anonymizing data; publishing findings; and curating, storing, and sharing data. The contributors to this special collection of Research on Language and Social Interaction articles identify and analyze the problems at each stage, and give an account of how these challenges may be met.