<p>Organisations whose mission is to fund health research are increasingly concerned with ensuring that the research they fund is used productively. The resulting interest in the concept of “knowledge transfer” has involved introducing policies to prompt researchers to think about their role, not just as knowledge producers, but as translators of research findings. In New Zealand, researchers can be asked, in their application for funds, to provide an account of what will happen to their research results. They are then judged on the quality of that account. However, little is known about how effectively this type of policy influences researchers to do more to make connections with those who use their findings. Using the explanatory power of the realist evaluative approach, this thesis examines the implementation of new instructions by the Health Research Council of New Zealand (HRC) for providing knowledge transfer pathways in research applications. A focus of the research is on how these instructions change (or do not change) the mind-set of researchers. Key informant interviews were held, and the scholarly and grey literature examined, to develop an initial theory on how researchers would be influenced by such instructions. Individual interviews were then held with researchers, seeking their reflections on what they had originally written in a specific knowledge transfer pathway and how this then matched up with what actually happened; these interviews were then used to refine the initial theory. Finally, an on-line survey was conducted with those who sat on the HRC’s research assessing committees in the 2014/15 funding round in order to refine the theory further. The final theory identified six mechanisms, which under different contexts, explain how the HRC’s knowledge transfer policy works (or does not work) to prompt researchers to reason differently. A continuum of reasoning in the form of a dimmer switch was used to explain circumstances where researchers may become more mindful of what is involved in knowledge transfer, but were not likely to markedly change their behaviours. Based on the assumption that the HRC wants to be more active in encouraging researchers to undertake activities other than producing research results, two recommendations are made: (1) knowledge transfer policies should support self-reflexivity by different groups of researchers rather than creating more hoops within the research application process, and (2) the processes by which knowledge transfer sections are judged needs to be strengthened if researchers are going to be confident that this is a “serious” part of the application process.</p>