The call for ‘evidence-based policy making’ has provoked responses amongst academics, which range from challenging the meaning of ‘evidence’, to trying to fit their research into ‘evidence’s’ presumed demands, or simply running for cover. All such responses reinforce the rise to ascendancy of ‘evidence’. They all bolster the idea that policy makers can legitimately demand ‘evidence’ as an essential raw material for their craft. This paper questions that assumption. Using approaches from social anthropology, it examines recent policies to reform universities in OECD countries and specifically in Denmark. An evidence-based process would suggest that decision making followed a rational, linear path. It also evokes an old image of policy making as a hierarchical process, with politicians and officials, armed with evidence, devising policy which ‘trickles down’ through new, more stringent forms of governance, to change the leadership of institutions and thence the behaviour of practitioners and clients (or lecturers and students) on the ground floor. In contrast, this study of Danish university reform draws on recent anthropological approaches to the study of large scale processes of transformation. The study takes a more democratic approach to policy and reveals a continuous and contested process involving a multitude of actors – from OECD scenario writers, to national policy makers, university leaders, academics and students – all involved in imagining and enacting the future university.