Comparing Discourse and Policy Networks: Empirical Evidence from the Emerging Policy Field of Micropollutants in Surface Waters

Policy networks and discourse networks both have been applied in literature to analyze policy processes. Policy network analysis typically uses survey data to measure interdependencies with regard to collaboration and shared beliefs. Discourse network analysis often uses media data in order to analyze interdependencies between actors based on their beliefs or policy preferences. In this regard, discourse network analysis builds on the assumption that discourse networks represent the discursive layer of policy processes and hence follow similar patterns as policy networks. However, there are reasons to challenge the assumption and expect differences in both networks. In this study, we systematically compare policy networks with discourse networks in the new emerging policy field of micropollutants in surface waters in Germany. In this regard, we pose several research questions: What are the differences and similarities between policy and discourse networks? Do both networks possess similar network structures, e.g. actor coalitions? Are there differences in the type of policy proposals discussed in policy compared to discourse networks? Theoretically, one can expect both similarity and discrepancy between policy and discourse networks. Based on Sabatiers’ early work (Sabatier 1988), it has been argued that advocacy coalitions consist of actors with shared beliefs or policy preferences. Subsequently, discourse networks and policy networks should identify similar network structures when analyzing the same empirical policy process. Even if different structures were identified, this would be due to measurement, e.g. data collection at different points in time. On the other hand, there is literature suggesting that the discursive layer of politics follows different patterns. Based on thoughts by Schattschneider (1960) and Baumgartner et al. (2009), there is reason to expect that different actor types dominate the respective networks. Furthermore, one can argue that discourse networks represent a more deliberative process that is open to marginalized or new actors and new ideas compared to policy networks that are less inclusive and more likely bound to path dependency. As a consequence, we should expect differences in participating actors, policy preferences, and network structures eventually. To sum up, this study compares policy networks and discourse networks on the policy issue of micropollutants in surface waters. In order to identify similarities and differences, we formulate several hypotheses based on theoretical underpinnings, which we then test by applying both network approaches on empirical data.