Evaluation of university-based research already has a reasonably long tradition in the UK, but proposals to revise the framework for national evaluation aroused controversy in the academic community because they envisage assessing more explicitly than before the economic, social and cultural 'impact' of research as well as its scientific quality. Using data from the 2009 public consultation on the proposals for a Research Excellence Framework, this paper identifies three main lines of controversy: the threats to academic autonomy implied in the definition of expert review and the delimitation of reviewers, the scope for boundary-work in the construction of impact narratives and case studies, and the framing of knowledge translation by the stipulation that impact 'builds on' research. Given the behaviour-shaping effects of research evaluation, the paper demonstrates how the proposed changes could help embed impact considerations among the routine reflexive tools of university researchers and enhance rather than restrict academic autonomy at the level of research units. It also argues that the REF could constitute an important dialogical space for negotiating science-society relations in an era of increasing heteronomy between academia, state and industry. But the paper raises doubts about whether the proposed operationalisation of impact is adequate to evaluate the ways that research and knowledge translation are actually carried out. (C) 2011 Elsevier B.V. All rights reserved.
