Against the widespread insistence that educational research should be "scientific", I attempt to change the terms of the debate. Instead of asking whether research is robust and rigorous, and whether it "works"--all terms derived from the "scientific" view--I argue that we should ask whether we can have "confidence" in research. This way of putting it enables us to place the idea of interpretation back at the centre of our understanding of education, and to see that when we have confidence in interpretation, it is because good interpretation has quite different qualities from scientific knowledge. At the centre of interpretation lies the idea of text and of responsiveness to text. Good interpretation requires good listening or "attentiveness". It does not pretend to represent the world accurately and conclusively, as "scientific" research does: it opens up space for conversation, and brings a world into being.