Measuring Polarization in Courts and Legislatures via Bipartite Networks

• In Event: Political Networks, Influence, and Communication in American Politics

Abstract

While conventional wisdom---both within the academy and in the larger public discourse---holds that polarization has intensified throughout America's political institutions, its measurement has thus far failed to keep pace with its growing conceptual importance. I propose a new measure of polarization rooted in network science, one which links political actors to the documents they endorse via a series of bipartite networks. The primary advantages of the measure are that it (1) applies to both judicial and legislative settings; (2) shifts the spotlight from voting behavior to position-taking behavior; and (3) permits the disentangling of partisan polarization from polarization along other axes. I define polarization, measure it using bipartite modularity, and demonstrate that it responds in intuitive ways to changing ideal point distributions using simulated data. I then measure polarization in the opinion-joining behavior of U.S. Supreme Court justices (1946--2017) and in the cosponsorship behavior of U.S. senators and representatives (1973--2016).

Author

• Jason Douglas Todd, Duke Kunshan University