Policy network theory at the science-policy interface

Improving the ability of a diversity of research voices to influence policy in critical domains for Aotearoa New Zealand.

The need and impact

How do funding flows and other measures of academic prestige such as appointments affect the ways in which researchers engage with governments and government-like organisations? How do we quantitatively identify the relationship between money, research, and power at the science-policy interface?

Using policy network theory, this project draws on recent digitisation of policy records and developments in natural language processing to develop an empirical study that will robustly test established political science theory regarding the spread and impact of influence in the public domain.

The significance of this study, using new methods to test established political science theory, has impact internationally, where intergovernmental, governmental, and community organisations are concerned about the deterioration of democracy and increasing public distrust in scientific certainty. These are pressing questions in Aotearoa New Zealand, as a small country with a small research sector which often cannot support more than a small number of experts, even in critical to Aotearoa New Zealand domains.

Our impact for Aotearoa New Zealand, therefore, focuses on improving the ability of a diversity of research voices to influence policy in critical domain areas. Focusing on water quality, climate change, human health, and fisheries policy domains, the project plans for impact on the ways in which Aotearoa New Zealand organises, solicits, and manages expert advice and evidence in forming public policy.

The approach

This empirical study enables one of the first robust quantitative tests of policy network theory at the science-policy interface using digital records of the policy-making process to measure the influence of research and researchers on the outcomes of that process. We have:

  • Information about the influence of researchers within their own research community (bibliometric data).
  • Measures of the influence of researchers in the public domain (e.g., mainstream media reporting).
  • Information about public funding of researchers, a mechanism by which policymakers can exert influence in return.

Linking these sources of information together at the level of individual researchers, we can then use natural language processing to analyse policy content to assess the relationship between researcher characteristics and their influence on policy change. The project utilises natural language processing tools developed by Te Pūnaha Matatini spin-out company Nebula.

We have key international links with Associate Professor Kathryn Oliver (London School of Tropical Medicine) and Professor Paul Cairney (University of Stirling), who will use our method to study similar examples in the United Kingdom.

Research aims

  • Estimate the influence of researchers in policy-relevant domains using bibliometric, mainstream media, and funding data.
  • Develop models for influence based on both individual characteristics such as assumed career stage alongside structural impediments such as sexism and racism.
  • Characterise researchers by strategy (e.g. insider, outsider, or otherwise), identifying changes over time.
  • Measure policy change using natural language processing.
  • Examine the relationship between influence and policy change.


  • Professor Shaun Hendy (Project Lead)
  • Dr Mubashir Qasim
  • Professor Ann Brower
  • Professor Troy Baisden
  • Dr Anna Matheson