Exploring Group Opinion Polarization Based on an Opinion Dynamics Model Considering Positive and Negative Social Influence

Regarding the evolution of public opinion from plurality of opinions to polarization, as well as leading to further attacks on each other by individuals holding different opinions and generating public opinion conflicts, it attracts the focused attention of research scholars. Therefore, constructing models that capture the complexity of social influence in networks should play an important role in the design of communication systems. In this study, based on the social judgment theory in social psychology, we add the repulsion threshold mechanism to the bounded confidence opinion dynamics model, combine assimilative and repulsive social influence, and analyze the characteristics of group opinion evolution. With the agent-based model, we find that the combination of assimilative and repulsive social influence produces complex outcomes that arise in social networks. Sometimes, more assimilative social influence actually leads to more, rather than less, opinion polarization. Similarly, the propensity of users to communicate with like-minded individuals sometimes diminishes opinion polarization. While only occurring in specific parts of the parameter space, these counterintuitive dynamics are robust, as demonstrated by simulation experiments. In addition, we explore the effect of WS small-world network structure on the evolution of opinions.