'That's not quite the way we see it': the epistemological challenge of visual data

In research textbooks, and much of the research practice, they describe, qualitative processes and interpretivist epistemologies tend to dominate visual methodology. This article challenges the assumptions behind this dominance. Using exemplification from three existing visual data sets produced through one large education research project, this article considers the affordances and constraints of the research process focusing particularly on analysis. It examines how and when the visual can be incorporated, gives some critical reflections on the role and use of visual methods to fulfil different research intents, and, in particular, considers combining large, open-ended data sets with acceptable and rigorous analysis techniques. We then explore arguments about the nature of visual data, what is considered epistemologically appropriate and the decision-making which accompanies any appraisal of process in education research. The intention is to challenge ourselves, and fellow visual methods researchers, to develop a more complete understanding of the theory and practice of visual research.