Given the scientific and social importance attached to productive interactions between science and policy practices, there is a striking lack of insight into current knowledge practices and the dilemmas they lead to. Our special issue can’t solve this deficiency but it can provide an impetus for opening up current knowledge practices, reflect on the role of science in them and instigate a more systematic exchange of methods. A warning is given for the reification of boundary work and Gabrielle Bammers’ Implementation and Integration Sciences is introduced as framework for the analyses.

When scientific values like objectivity, validity and reliability are inadequate for designing research that enables society’s capacity for dealing with unstructured problems, which values or criteria should we use for designing adequate knowledge practices? Based on the articles in this special issue we answer this question by analysing the methods researchers have used for selecting stakeholders, knowledges, synthesis, context and outcome in new knowledge practices. Although a common language for comparison and documentation is lacking, the analysis provides recommendations for better designing interaction between scientific and other practices. The most important message however is that we need a designated platform for exchanging and evaluating experiences and discussing methods and the outcomes they yield.