Opinion dynamics involving contrarian and independence behaviors based on the Sznajd model with two-two and three-one agent interactions

TL;DR

This paper studies opinion dynamics using an extended Sznajd model with contrarian and independent behaviors, revealing phase transitions, critical points, and consensus probabilities in social agent interactions.

Contribution

It introduces a novel Sznajd model with two-two and three-one agent interactions incorporating contrarian and independence behaviors, analyzing phase transitions and universality class.

Findings

Critical points decrease exponentially with contrarian and flexibility factors.

Scenario three-one reaches a smaller critical point than two-two.

Contrarian-involving scenarios have higher consensus probability than independence.

Abstract

We investigate the opinion evolution of outflow dynamics based on the Sznajd model on a complete graph involving contrarian and independence behaviors. We consider a group of four spins representing the social agents with the following scenarios: (1) scenario two-two with contrarian agents or independence agents and (2) scenario three-one with contrarian or independence agents. All of them undergo a second-order phase transition according to our simulation. The critical point decreases exponentially as $\lambda$ and $f$ increases, where $\lambda$ and $f$ are contrarian and flexibility factors, respectively. Furthermore, we find that the critical point of scenario three-one is smaller than that of scenario two-two. For the same level of $\lambda$ and $f$, the critical point of the scenario involving independence is smaller than the scenario with contrarian agents. From a sociophysics point of view, we observe that scenario three-one can likely reach a stalemate situation rather than scenario two-two. Surprisingly, the scenarios involving contrarians have a higher probability of achieving a consensus than a scenario involving independence. Our estimates of the critical exponents indicate that the model is still in the same universality class as the mean-field Ising model.