Safeguarding Democracy: Polarization, Populism, and Post-Truth ISS Sessions

Safeguarding Democracy: Polarization, Populism, and Post-Truth Symposia Sessions

Why do citizens support leaders who undermine democratic institutions? Recent research suggests that authoritarian leaders mobilize followers through identity-based appeals, deploy gendered imagery, and exploit online platforms to spread their message, where algorithms and user preferences amplify hostile and moralized content. Interventions such as fact-checks and polarization-reduction strategies—including personalized approaches powered by large language models—demonstrate measurable benefits, yet their effects often fade in divisive media environments.  

This symposium examines these dynamics and considers how democracies can build lasting resilience in the face of evolving authoritarian tactics. 

  • Jay Van Bavel, New York University, USA
  • Renée DiResta, Georgetown University, USA
  • Christina Pagel, University College London, United Kingdom
  • Alexander Haslam, The University of Queensland, Australia

16:15 – 17:15 (4:15 PM – 5:15 PM)

Presenting Author: Aldo M Barrita, Michigan State University, USA

Abstract: We surveyed 202 Latinx immigrant college students to examine how immigration-status microaggressions relate to distress and school motivation. Microaggressions predicted higher distress and lower motivation, partially via psychological distress. Moderated mediation analyses showed that resistance and discrimination-related education buffered harmful effects, highlighting protective coping processes within diverse learning environments.

Presenting Author: Yi Zhang, University of Southern California, USA

Abstract: In a preregistered live-conversation study (N=182), participants discussed moral disagreements under mindful, suppression, or natural reaction instructions. Mindful acceptance increased positive affect, openness in expressing disagreement, and moral subjectivism. Suppression, by contrast, reduced expression. Results show that emotion regulation causally shapes interpersonal disagreement dynamics and offers a pathway for depolarization.

Presenting Author: Ahoo Hekmati, Paris Cité University, CNRS, France

Abstract: We examined relationships between maternal cultural values, mental-state talk, and children’s ToM in Turkiye (N=112, 53 boys, M=51.68, SD=5.25). While SES was linked to maternal cognitive talk, maternal individualism was negatively associated with children’s false belief understanding. Findings suggest high SES supports linguistic scaffolding, yet achievement-oriented values may hinder ToM.

Presenting Author: Will Blakey, Stanford University, USA

Abstract: We created a brief (20-minute), social-psychological intervention to change how people think about and communicate in political disagreements. In a preregistered longitudinal RCT with a U.S. census–matched sample of 1,027 university students, treated participants wrote more respectfully to outpartisans at each survey-wave, including one month later in a disconnected survey.

Presenting Author: Tehila Kogut, Ben-Gurion University, Israel

Abstract: Across a large European survey and an experiment, we show that gendered perceptions of discrimination shape democratic support in opposite ways: women strengthen democratic commitment when threatened, whereas men reduce theirs under status insecurity. These divergent reactions reveal how subjective disadvantage fuels democratic erosion and opens pathways for populist mobilization.


17:30 – 19:00 (5:30 PM – 7:00 PM)

Presenters:

  • Steve J. Rathje, Carnegie Mellon University, USA
  • Jon Roozenbeek, University of Cambridge, United Kingdom
  • Ili Ma, Leiden Institute for Brain and Cognition, Leiden University, Leiden, The Netherlands

Abstract: We plan to provide a hands-on tutorial on recent advances in AI that can be used to improve methods for studying democracy-related questions (for instance, by classifying large-scale text data and creating interactive experiments). We will also discuss the strengths, challenges, and ethical issues involved in using large language models in research.