An Investigation into the Sensitivity of Social Opinion Networks to Heterogeneous Goals and Preferences

As research into the dynamics and properties of opinion diffusion on social networks has increased, so too has the attention paid to modeling such systems. Simulations using agent-based modeling (ABM) analyze aggregate network outcomes when individual agents act on typically limited information, and tend to focus on agents that are conforming and homophilic-that is, they prefer to be around similar others, and they update their own personal state over time to be more like their friends. In this work, we illustrate the value of diverse agent modeling in environments that allow for strategic unfriending. We focus on network dynamics generated by three agent models, or archetypes. Our work shows that polarization and consensus dynamics, as well as topological clustering effects, may rely more than previously known on the interplay between individuals' goals for the composition of their neighborhood's opinions.