A parametric study of opinion progression in a divided society

Abstract

In this paper, a probabilistic finite state automaton framework is used to model the temporal evolution of opinions of individuals in an ideologically divided society in the presence of social interactions and influencers. In such a society, even quantifiable and verifiable facts are not unqualified absolute but are only viewed through the prism of the individual’s biases which are almost always strongly aligned with one of the few prominent actors’ viewpoint. The gradual progression of divisiveness and clustering of opinion or formation of consensus in a scale free network is studied within the framework of bounded-confidence interaction between nodes. Monte Carlo simulations were conducted to study the effect of different model parameters, such as the initial distribution of opinion, confidence bound, etc. in the behavior of the society. We have shown that in absence of influencers, government policies are the important factors in the final distribution of the society unless a specific group has higher number of members initially. Also, even very small groups of influencers proved to be highly effective in changing the dynamics.