In the US debate over science-based educational research, experimentation is again coming to the fore as ideal for establishing an evidence base of what works in educational interventions thus contributing to the climate of accountability in educational research and practice. Mixed methods research is being positioned in this discussion as a reasonable alternative, a way of making it possible for both experimentalists and non-experimentalists to join forces in the quest for accountability. This paper argues that mixed methods research is hardly something new; moreover, it is an epistemological response to what is basically a moral-political problem.