Role of noise and agents' convictions on opinion spreading in a three-state voter-like model

TL;DR

This study investigates how noise and individual convictions influence opinion spreading in a three-state voter model, revealing phase transitions and the impact of external mass media effects on opinion dynamics.

Contribution

It introduces a three-state voter-like model with heterogeneous convictions and noise, analyzing their effects on consensus formation and opinion distribution.

Findings

Consensus only occurs at zero noise.

Relaxation times grow with system size as a power law.

A threshold noise level causes a majority of undecided agents.

Abstract

In this work we study the opinion formation in a voter-like model defined on a square lattice of linear size $L$. The agents may be in three different states, representing any public debate with three choices (yes, no, undecided). We consider heterogeneous agents that have different convictions about their opinions. These convictions limit the capacity of persuasion of the individuals during the interactions. Moreover, there is a noise $p$ that represents the probability of an individual spontaneously change his opinion to the undecided state. Our simulations suggest that the system reaches stationary states for all values of $p$, with consensus states occurring only for the noiseless case $p=0$. In this case, the relaxation times are distributed according to a log-normal function, with the average value $\tau$ growing with the lattice size as $\tau\sim L^{\alpha}$, where $\alpha\approx$ 0.9. We found a threshold value $p^{*}\approx$ 0.9 above which the stationary fraction of undecided agents is greater than the fraction of decided ones. We also study the consequences of the presence of external effects in the system, which models the influence of mass media on opinion formation.