Distribution of Opinions: Insights from Agent‐Based Modeling

Understanding the distribution of opinions is of interest because it shows how a population sees a given subject and how they may behave in certain situations which are important for scientific and policy reasons. In this paper we apply an agent-based opinion dynamics model to investigate the distribution of opinions and opinion cluster sizes. We determine the sensitivity of the distribution of the opinions and opinion cluster sizes to changes in individuals’ tolerance and uncertainty levels through parameter sweeps. Our results support the presence of dynamic distribution for opinion cluster sizes distribution. Moreover, statistical analysis of the simulated opinion data shows that the opinion distribution has fat tail in most configurations of the model and does not comply with Gaussian distribution. In case of the frequency of the opinion cluster sizes, however, in addition to be fat tailed, the simulation results reveal a dynamic transition between lognormal and exponential distribution as parameters change.