Opinion Dynamics of Skeptical Agents

How does skepticism affect opinion formation in networks? In many settings, agents exhibit skepticism in the presence of people whose beliefs radically different from their own, and they are reluctant to be persuaded by such individuals. We present a model of opinion dynamics where agents are receptive toward other agents that have similar opinions, but remain skeptical of agents holding disparate opinions. We analyze how agents with extreme opinions affect the general population, using simulations on Barabási-Albert random graphs, and modified Erdös-Rényi random graphs that incorporate homophily. Finally, we show that even skeptical agents are able to come to an early consensus and take coordinated action to reach a final opinion in most settings; but, agents in homophilic networks may fail to converge to a single opinion. Paradoxically, this happens when agents are least skeptical, and are able to stabilize themselves by balancing influence from extremists from opposing camps.