Relative opinion similarity leads to the emergence of large clusters in opinion formation models

TL;DR

This paper introduces a variant of opinion formation models where the likelihood of opinion change depends on the relative similarity of opinions among neighbors, leading to larger opinion clusters and reduced polarization.

Contribution

It proposes a new opinion formation model incorporating relative opinion similarity, showing its impact on cluster size and polarization in social networks.

Findings

Larger weights on relative similarity increase the largest opinion cluster size.

The threshold parameter exhibits an inverse-U relationship with cluster size.

Considering relative similarity prevents polarization into small clusters.

Abstract

This study considers a variant of the bounded confidence opinion formation model wherein the probability of opinion assimilation is dependent on the relative similarity of opinions. Agents are located on a social network and decide whether or not they adopt the opinion of one of the neighbors (called a role agent). Opinion assimilation is more (less) likely to occur when the distance from the opinion of the role agent is smaller (larger) than the average opinion distance from other neighbors. Thus, assimilation probability is reliant not only on opinion proximity with the role agent considered in conventional models but also on relative similarity that considers other neighbors. The simulation results demonstrate that large weights on relative similarity in determining assimilation probability increase the size of the largest opinion cluster. The size of the threshold parameter of the bounded confidence model displays inverse-U relationships with the largest cluster size. The findings imply that consideration of relative opinion similarity, as observed in recent empirical studies, prevents polarization into small opinion clusters.