The idea that educational research should be "scientific", and ideally based on randomised control trials, is in danger of becoming hegemonic. In the face of this it seems important to ask what other kinds of educational research can be respectable in their own different terms. We might also note that the demand for research to be "scientific" is characteristically modernist, and thus arguably local and temporary. It is then tempting to consider what non-modernist approaches might look like. The purpose of this article is to sketch a case for one particular reaction against modernist thinking: romanticism. How might our understanding (apprehension, sense) of education be changed by readmitting the insights and perspectives of romanticism? And, crucially, what confidence could we have in educational research that was thus inspired and that took the "romantic turn"?