Policy Networks in Complex Governance Subsystems: Observing and Comparing Hyperlink, Media, and Partnership Networks

Methods for observing policy networks have not kept up with the development of new network analytic techniques required to understand governance in complex settings. We compare three unobtrusive methods for observing policy networks based on hyperlinks between policy actor web sites, on media reports, and on public policy partnerships. Observations of one complex local water policy arena with all three methods provide a comparison of the general as well as actor‐specific network characteristics in the three observed networks. The core network of actors observed by all methods has similar network level statistics, highly correlated relationships measured by Quadratic Assignment Procedures models, and the same significant network microstructures as measured by Exponential Random Graph Models. The full networks including actors observed by any method also exhibit similar actor‐level characteristics, although the correlations across networks are stronger for bridging capital measures than for bonding capital measures, and each method has different apparent biases. Once biases are accounted for, similarities suggest that these methods may provide useful proxies for each other and for other relationships that are more difficult or impossible to measure, particularly when combined to offset each method's biases. If so, they can extend the range of policy networks observable with limited resources across space and time.