This methods lecture will introduce discourse network analysis, a methodological toolbox for analysing policy debates and their development over time. At its core, the open-source software Discourse Network Analyzer (DNA) and its associated R package rDNA allow researchers to code actors’ policy beliefs in text data, for instance on how to solve a complex policy problem like the pension gap in the face of demographic change or on the adoption of a specific health or environmental policy. Using the annotated text data, the researcher can export various kinds of network data, such as an actor congruence network, in which actors (e.g., legislators, interest groups, charities) are connected to each other by shared policy beliefs, a concept congruence network, in which policy beliefs are connected through co-referral by actors, and affiliation networks, in which actors are connected to policy beliefs, possibly longitudinally. The resulting networks serve to uncover informal coalitions in the policy process, their development and polarisation over time, defection by actors, opinion leadership and brokerage, policy beliefs of particular structural importance for the debate, phase transitions in the structure of a debate, and the relative positions of actors in their coalitions. Temporal statistical network models can be employed to learn the micro-level mechanisms underlying contributions to a debate, and the models can predict the next actor-belief statements at any given time of the debate, with uncertainty quantification. The presentation introduces the method and provides an overview of different ways in which the resulting network data are analysed in practice.

——————

This seminar is part of the Department of Social Policy and Intervention Modern Methods in Social Policy and Intervention Research Seminar Series.

Booking required for people outside of the Department of Social Policy and Intervention (DSPI). DSPI Members do not need to register.