The opinion dynamics model which is being widely investigated by physicists can be largely divided into two types: discrete opinion models and continuous opinion models. The former uses a finite number of values to express agents’ opinions. The latter, unlike the discrete opinion models, takes a real number to represent agents’ opinions.
Some researchers studied the SJT model, one of the continuous opinion models. They proposed two thresholds: an acceptance threshold (d1) and a repulsion threshold (d2). According to their rules, when the difference in the inner opinion between two randomly selected agents is smaller (or larger) than d1 (or d2), the agents simultaneously update their opinions in the direction where their difference gets smaller (or bigger).
Jang Paek San, a researcher at the Institute of Nano Science and Technology, proposed a modified SJT model with a new opinion updating mechanism. Unlike the previous works, every individual affects the others’ opinions and the extent of the influence depends only on the magnitude in the opinion difference. To incorporate the convergence parameter of the assimilation effect and the divergence parameter of the repulsion effect into one parameter, he introduced a new function and used Monte Carlo simulations to study the opinion evolution process of the new model.
The simulation results showed that as d1 increases, the number of opinion clusters increases, but with increasing d2, the opinion dynamics changes from a bipolarization of opinions to a consensus of opinions.
For more information, you can refer to his paper “Social Judgment Theory model with continuous interaction between assimilation and repulsion mechanisms” in “SN Social Sciences” (SCI).