Measuring the policy impact of population health research

Current assessments of research impact have been criticised for capturing what can be easily counted, not what counts. This thesis examines research impact and use in multiple ways, using childhood obesity prevention research and policy conducted and developed in New South Wales (NSW), Australia, between 2000 and 2015, as a case example. A broad systems lens is applied, with sub-studies examining specific aspects of this research-policy system in detail. A mix of quantitative and qualitative methods are used, including: systematic review; bibliometric analysis; citation analysis; document analysis; semi-structured interviews; and case studies. Firstly, the literature on research impact assessment is compared to that measuring policy utilisation of research. Secondly, the knowledge production context, including how and why research was conducted is investigated. Finally, research impact and use are assessed tracing forwards from research and backwards from policy, using a mix of narrow (research projects; policy documents) and broad (policy cases) units of analysis. The comparative approach used, empirically demonstrates how some methods of measuring impact restrict what can be defined and assessed as impact and therefore downplay the diverse ways in which research can influence policy processes. Different methods provide different perspectives of the relationship between research and policy, in terms of causation versus contribution. In addition, they give a different view of what is transferred, knowledge or research. Ultimately assessment needs to go beyond an understanding of research use as instrumental. This view determines the type of research that will be identified as impactful at the expense of other types of research. Looking beyond the efforts of individual researchers and research groups, to examine how systems support policy relevant knowledge production and use, may offer a way forward