Not Everyone Has a Voice, or an Ear, on Social Media: The Poster-Observer Model of Opinion Dynamics

Abstract

Classic modeling of opinion dynamics via the Friedkin-Johnsen (FJ) model relies on the unrealistic assumption of synchronous communication among network users. Social media, however, exhibit extreme participation asymmetries: a vast majority of users silently observe, while a tiny minority creates and broadcasts most of the content. To capture this reality, we introduce the Poster-Observer FJ (PO-FJ) model, which decouples user communication into distinct posting and observing behaviors. This formulation fundamentally alters network dynamics by preventing convergence to a static equilibrium, driving opinions into perpetual oscillations. We establish closed-form expressions for these new dynamics and generalize classic metrics of public opinion, polarization, and disagreement under the PO-FJ model. By leveraging the novel Poster-Observer dynamics, we introduce a realistic and scalable intervention algorithm for conflict mitigation. Through experiments on synthetic and real-world networks, we demonstrate the efficacy and efficiency of our approach in reducing social-network conflict.