Although much is claimed by practitioners for the value that ethnographic film and other visual research methodologies offer to anthropological research, much remains at the level of assertion rather than evidence. As visual methodologies are rarely if ever used in isolation from other methodological strategies (such as interviewing) it is sometimes difficult to disentangle what findings were produced by which method. This paper considers the problem of verifying that the data produced by visual research methodologies are robust and that they reveal sociological or anthropological insights that could not have been generated by other means. The first part of the paper takes ‘research findings’ in a largely positivistic way and considers a number of examples that claim to have uncovered sociological or anthropological data simply not accessible by any other means of enquiry. The paper then goes on to consider claims of visually-derived evidence from a more phenomenological or post-structuralist perspective, within which issues of ‘evidence’, ‘findings’ and indeed ‘data’ are critiqued or rejected. The paper concludes by seeking to square the circle: to propose some means by which visual methodologies employed within the now-dominant post-structuralist paradigm can nonetheless satisfy the criteria of empirical social science.