Using data from the Department of Health funded study "Making the most of policy evaluations", this paper explores the views of those working within the policy process about the role of evidence. It highlights a tension between formative and summative uses of research, which appeared to be exacerbated by a lack of clarity surrounding the objectives that policy leads had for the research. Additionally, the data reveal an uncertainty about the status of "pathfinder", "demonstration" and "pilot" sites within the policy process.