Advances and Challenges in the Measurement of Latent Traits

Session Submission Type: Created Panel

Session Description

The panel presents projects leveraging novel computational techniques to explore the multifaceted nature of political discourse, ideology, and framing within public policy debates. The papers rely on classical measurement approaches as well as large-language models (LLMs) to obtain measures of latent traits (such as issue frames and ideology) from sources such as parliamentary speeches and open-ended survey responses. The research collectively advances our understanding of how political debates and ideologies evolve and intersect with public opinion and media framing, highlighting the dynamic interplay between traditional political analysis and cutting-edge computational approaches.

Sub Unit

• Division 8: Political Methodology

Individual Presentations

• Binding a Bottomless Barrel: Boundless Dimensions in Ideological Attitude Scales - Philip Warncke, UNC Chapel Hill; Flavio Azevedo, Friedrich Schiller University

• Constructing Machine Learning Index: Cases of Populism and Nationalism - Peitong Jing, University of Notre Dame

• DeepComplete: Generative Adversarial Autoencoders for Multiple Imputation - Umberto Mignozzetti, UC San Diego

• Multiple Imputation for Large Multidimensional Data with Linear Constraints - Jian Cao, Trinity College, Dublin

• The Polarization of the Immigration Debate: Evidence from 9 National Parliaments - Ahra Wu, Princeton University; Rafaela Dancygier, Princeton University; Vincent Heddesheimer, Princeton University; Leah Boustan

Chair

• Matthew Tyler Rice University

Discussants

• Bang Quan Zheng University of Texas at Austin

• Matthew Tyler Rice University