Association for Public Policy Analysis & Management

Discourse Network Analysis

1. Purpose and Idea

Studies of public opinion and public discourse enjoy increased attention, not only since the somewhat surprising outcome of the recent U.S. election but also because of the ongoing efforts of companies, parties, individuals, states and other political actors to shape and influence the public opinion. A particularly interesting area of public opinion research is the study of discourse-coalitions and discourse-developments in the public arena. Questions like

-          Who is setting the agenda of the public discourse?

-          Who determines the framing of relevant topics?

-          What actors share opinions about relevant topics?

-          What actors are particularly influential in the public discourse?

-          What are the frequently cited “advocacy coalitions”?

are most interesting for scholars of public opinion as well as decision makers in government and the private sector.

2. The Tool: Discourse Network Analyzer

The “Discourse Network Analyzer” (DNA), developed by Dr. Philip Leifeld (University Konstanz, University of Glasgow) in 2011, is a most useful tool for answering questions like the ones stated above. It is a java-based software that can be downloaded free of charge and that is very user friendly. The DNA can be used for generating data that is very easy to visualize and to communicate to both academic and non-academic audiences. In general, a discourse analysis is used to analyze either discourse contents (“content-based discourse analysis”) or to analyze the relations between participants of the discourse (“actor-based discourse analysis”). This analysis is possible through the systematical capturing of the believes and opinions of the actors and groups of interest.

It is possible to classify and analyze different kinds of public opinion data with the Discourse Network Analyzer, ranging from traditional media like newspapers to new media such as Facebook-comments. It combines attributes of qualitative content analysis with attributes of quantitative methods, such as traditional network analysis. The graphical representation of the discourse with the identification of “central” actors is both relevant from an analytical point of view as well as for the communication of scientific results to decision makers.

3. Examples

In my own research, I used the Discourse Network Analyzer for several different studies in the areas of international relations and public policy. Currently, I supervise a project that is dedicated to identify and to analyze the relevant political actors in the discourse regarding energy policy in Germany. The project team consists out of 6 graduate students and 8 undergraduate students assisting with the collection of data. The project is part of the initiative “ThinkLab Energy – Society – Change”. So far, we analyzed over 1000 newspaper articles and identified hundreds of organizations and their respective policy preferences regarding Germany’s long term energy policy (“Energiewende”). We were able to identify changes in the public discourse, e.g. due to external events such as the Crimea Crisis as well as strategic coalitions between different sectors in the economy. In August 2017, we will present a concluding report about the development of policy preferences during the last 5 years.