CSH researcher Tuan Pham will present his work (done together with Andrew Alexander, Jan Kobel, Rudolf Hanel, and Stefan Thurner) at the Networks 2021, Section “Agent-based models of social networks.”
In the past decades an increasing level of social fragmentation is observed in many societies that leads to the emergence of so-called echo chambers or filter bubbles. We try to understand this phenomenon with an agent-based model where we use spins to represent vectors of (binary) opinions of individuals and use a positive (negative) link weight to represent friendship (enmity), respectively.
We take into account the joint effects of (1) homophily—the tendency of people with similar opinions to establish positive relations, and (2) social balance—the tendency to establish balanced triadic relations. These two mechanisms are incorporated in a localized Hamiltonian that minimizes social stress through the co-evolution of opinions of individuals and their social networks.
Within the presented framework, we show how the likelihood of social fragmentation increases as individuals care more about social balance. We identify the critical size of the social neighbourhood, above which society must fragment into sub-communities that are internally cohesive and hostile towards other groups. This critical value necessarily exists in the presence of social balance for any number of individual opinions and for almost any sparse realistic social network.