Political polarization is a central concern in democratic societies, as extreme ideological divides hinder democratic functioning. While cross-cutting exposure is often seen as depolarizing, recent evidence suggests it operates on a continuum—from agreement to fundamental disagreement—with only moderate disagreement effectively reducing polarization. This nuance is particularly critical in multi-party systems characterized by greater ideological diversity. We propose a theoretical framework that integrates varying agreement levels with interpersonal, mass media, and social media contexts of opinion exposure, each with distinct frequency, intensity, and impact profiles. Using an agent-based model (ABM), we simulate how everyday opinion exposure across contexts shapes polarization over time, systematically varying media environments and initial opinion distributions. Our simulations reveal a general trend toward polarization, amplified under high-choice mass media conditions. Notably, social media platforms that increase exposure to moderate disagreement could mitigate overall polarization, underscoring the significance of algorithm design in democratic discourse.

Schindler, J. & Haim, M. (6/2026). The Differential Impact of Disagreement on Polarization: Simulating Opinion Dynamics Across Interpersonal, Mass Media, and Social Media Contexts. Presented at the 76th Annual Conference of the International Communication Association, Cape Town. (content_copy)