Framing and policy change after shooting rampages: a comparative analysis of discourse networks

Abstract

In this contribution, we comparatively analyse the discourse networks generated by three shooting rampages. We formulate hypotheses on the extent to which (a) the framing cohesion of the status quo coalition and (b) the perceived causal complexity of the event are associated with varying degrees of subsequent gun policy change. Drawing on news reports in major newspapers, we collect information on actors and frames and systematically analyse those data with the tool Discourse Network Analyzer. The networks show that major gun policy change is possible if the status quo coalition is internally divided and the event's causal complexity is low. Incremental adjustments are also likely if the status quo coalition lacks cohesion, but if additionally causal complexity is high, i.e., the problem signified by the event is disputed. Finally, if the status quo coalition manages to retain its framing cohesion, deadlock is likely to occur regardless of the event's perceived causal complexity.

Keywords:

• Discourse network analysis
• framing
• gun policy
• morality policy
• policy change

ACKNOWLEDGEMENTS

We thank our three anonymous referees for their constructive comments. This contribution is based on the project MORAPOL (ERC Advanced Grant). We gratefully acknowledge the generous funding by the European Research Council.

Notes

Note that this argument is also closely linked to the Advocacy Coalition Framework (Sabatier 1998). However, Sabitier's framework explicitly requires the researcher to look at a decade or more, while our focus lies on rapid policy change after an external shock.

These networks can be visualized by other software packages such as visone or UCINET.

Germany: Frankfurter Allgemeine Zeitung and Süddeutsche Zeitung; United Kingdom: The Guardian and The Independent; Switzerland: Neue Zürcher Zeitung and Der Tagesanzeiger.

Dunblane: 171 articles, 106 statements; Erfurt: 408 articles, 311 statements; Zug: 76 articles, 84 statements. A list of the analysed articles can be made available upon request.

We decided not to differentiate between diffuse and focused statements in our networks, because this would have made them more confusing and harder to interpret. In Erfurt and Dunblane, we observed that diffuse statements generally spread across different types of actors, but not so much in Zug.