Modelling the Emergence of Shared Attitudes from Group Dynamics Using an Agent-Based Model of Social Comparison Theory

We propose a novel agent-based implementation of Festinger's social comparison theory (SCT). The social comparison model (SCM) consists of connectionist networks that simulate agent-level social comparison processes. Agent networks are combined into an adaptive network structure that is shaped by social comparisons between individual agents. Simulations show how the SCM produces behavior consistent with the empirical literature on group dynamics. In addition, experimental results are reported that show how the SCM can simulate how critical and conformist norms affect interpersonal processes and emergent attitudes. We conclude that the coupling of simulations and experiments, and the use of psychologically plausible agent models within adaptive network structures, can provide new impetus to the development of models of individual and social cognition. An integrated framework such as the SCM allows investigating key theoretical predictions around the origin and maintenance of socially shared information through social comparisons in fundamentally novel ways.