Networked Models of Social Influence: Explaining Left-Right Political Landscapes in Europe Through Opinion Dynamics and Network Structure

Traditional models of social influence typically use assimilative or repulsive influence to study how consensus or polarization emerge. Given simple network structures, such as fully connected graphs, traditional models often fail to account for the multi-modal opinion distributions found in empirical data. In this study, we focus on more realistic social network structures in terms of clustering coefficient and average shortest path length and construct a model that allows both assimilative and repulsive influence to drive opinion changes in individuals. We find that non-trivial patterns emerge when the forces of assimilative and repulsive influence are kept at a specific ratio and the network structure is highly clustered. Comparisons with empirical left-right political opinion landscapes show that our model produces realistic results that share the multi-model characteristics as observed in data collected by the European Social Survey program.