Universities are increasingly expected to fulfill a third mission in addition to those of research and education. Universities must demonstrate engagement with society through the application and exploitation of knowledge. As societal impact of research is uncertain, long term and always dependent on other factors, we argue here that evaluation should focus on the conditions under which societal impact is generated rather than on the impact itself. Here we focus on a specific set of those conditions: the interactions between academic researchers and societal actors. Instead of speculating about potential impacts of research, we argue that current productive interactions of researchers with societal stakeholders improve the probability that future societal impact will occur. This article supports this idea by examining in detail several, mainly retrospective examples. As productive interactions are field specific, we restrict ourselves to 'professional adhocracy fields', especially to information and communication technologies (ICT) research. We address the patterns of productive interactions that result in societal impact within this field and we discuss whether differences are observed in contrast to other fields, such as social sciences and humanities (fragmented adhocracies). We end by discussing the implications that these patterns have for societal impact assessment. Shifting the focus to interactions allows assessment of short-term knowledge transfer and other collaborative efforts with stakeholders that contribute to long-term societal impact.

Many countries have amended legislation and introduced policies to stimulate universities to transfer their knowledge to society. The effects of these policies on scientists are relatively unexplored. We employ principal-agent theory to increase our understanding of the relationship between impact policies and scientific practice. Our methodology includes the analysis of policy documents and of data gathered in focus groups. We conclude that there is a gap between policy on the one hand and how scientists perceive it on the other. Policy documents put forward a broad notion of impact, but scientists perceive them as focusing too narrowly on commercial impacts. Scientists are further puzzled by how societal impact is evaluated and organised, and their perceptions frame their behaviour. Our policy recommendations focus on improving the interaction between intermediaries, such as universities and research councils, and scientists so as to include the latter's perspective in policy-making.

No abstract

Science is increasingly heterogeneous, posing new questions for research evaluation. How can we evaluate the between scientific and societal quality of research, taking into account differences between research fields and between research groups? In this paper we present the findings of two case studies in fields where societal and scholarly output of research are highly intertwined (architecture and law). We analyze the nature of the two fields in terms of research areas and specific aspects of knowledge dynamics. This results in an approach and indicators for contextual research evaluation. Websites of faculty and research groups