Data-Driven Modeling of Opinion–Behavior Coevolution in Social Networks: Evidence from Environmental Cooperation and Digital Participation
Main article
Abstract
Understanding how environmental opinions and cooperative behaviors co-evolve within social networks is essential for designing effective governance mechanisms in the context of growing ecological crises and expanding digital connectivity. This study develops a data-driven coevolution model that integrates opinion dynamics theory, bounded-confidence updating rules, and agent-based simulation to analyze how individual environmental attitudes and collective cooperation behaviors mutually shape each other over time. Using data from the Chinese General Social Survey (CGSS 2021, N = 3,842), we empirically examine the structural and relational determinants of environmental cooperation and digital participation, and then validate a multi-layered simulation framework that reproduces the observed empirical patterns. Our regression analysis reveals that network density (β = 0.312, p < 0.001), opinion homophily (β = 0.241, p < 0.001), environmental awareness (β = 0.334, p < 0.001), and digital participation (β = 0.187, p < 0.001) are robust predictors of environmental cooperation. Simulation experiments under seven distinct network scenarios demonstrate that cooperation equilibria are highly sensitive to the confidence bound parameter (ε), network topology, and social trust levels. Specifically, scale-free networks with high digital participation rates generate cooperation rates approximately 22.9 percentage points higher than sparse networks with low social trust. These findings offer both theoretical contributions to the growing literature on socio-ecological coevolution and practical guidance for policymakers seeking to leverage digital platforms and social network structures to promote environmental collective action.
