Our project aims at developing a method for the assessment of the social impact of scientific research. We review current literature including experimental studies and conduct case studies in four different fields (nanotechnology, ICT, health, social sciences) with various grades of social impact, in both national and supranational settings. Our goal is both to enhance insight in social impact assessment and to develop assessment methods. We focus on social impact of research. However, since most studies of how research has an impact on society show the crucial role of productive interactions between science and society, our prime object of investigation is the identification of these interactions. Thus, social impact is not seen as the ‘logical’ consequence of a unilinear process, but as the outcome of an iterative practice in which researchers and stakeholders each play a role. By productive interactions we mean: Exchanges between researchers and societal actors in collaborative settings (networks) in which knowledge is produced and valued that is at the same time scientifically and socially robust and relevant. Therefore, we engage in our project not only researchers and policy makers, but also other relevant stakeholders in the various research areas. Analytically, we distinguish four main tracks through which such interactions may occur: (1) through direct personal contacts (ranging from mere meetings to complex arrangements for research collaboration), (2) mediated by specific outputs like expert reports, clinical guidelines, scientific advice, or through (3) the transfer of goods (products, social practices, therapies, policy tools), and (4) through funding or other support mechanisms, in short: people, texts, artefacts and support. Our objective is in the case studies to show how these interaction mechanisms form a necessary condition for research to have a social impact.