Skip to main content

Does Research Evidence Inform Policy Choice in Low-income Countries? Exponential Random Graph Models to Determine the Role of Evidence Exchange and Use in Policy Networks

Jessica Shearer

Background: Research evidence is a valuable input into health policy decision-making processes; a consensus exists that research evidence can improve the efficiency and effectiveness of policy-making. Yet, research evidence is not commonly used for reasons relating to the evidence itself, perceptions of the user, and that it competes with other inputs in policy-making. Previous studies have identified the positive effect of interpersonal relationships between research producers and users on the use of evidence in policy-making. This study thus adopts a social network lens in order to describe the determinants of evidence exchange by actors in policy networks, whether those exchanges, or other factors, predict evidence use by individual policy actors, and whether dyadic exchange and nodal use translate to the use of research evidence by the network to inform the resulting policy decision.

Methods: In-depth interviews and social network surveys provided data for three health policy networks and their recent policy decisions in Burkina Faso. Social network methods were used to map and describe these multiplex networks, and exponential random graph models were built to explain determinants of evidence exchange, controlling for structural and individual-level covariates.

Findings: The frequency of evidence exchange varied across policy issues, and actors were more likely to provide, rather than request, evidence. Exchange was better predicted by structural variables in ERGMs, including entrainment and transitivity, than by individual attributes. An actor's use of evidence was positively associated with their degree centrality; however, dyadic and nodal exchange and/or use did not necessarily translate to instrumental use by the aggregate network, hinting at the other social processes at play in policy choice. These findings will be explored in detail in order to suggest a research agenda for network interventions to support evidence use in policy-making.