Cultural and opinion dynamics in small-world “social” networks

This study shows the importance of considering two social network features in agent-based models of cultural and opinion dynamics – network segregation by groups (e.g. race, class, or ideology) and interaction frequency by structural embeddedness. I formalize the two features as modeling conditions and apply them to an existing model that shows cultural or opinion polarization can emerge in a small-world network by the bridging of long-range ties. I find that when a small-world network is segregated, the inactivity of long-range ties (i.e. infrequent interactions) – instead of the bridging – becomes the key feature that contributes conversely to consensus or more extreme polarization. This implies that a systemic understanding of dyadic-level tie characteristics and suitable approaches to considering social networks in agent-based models are necessary.