Social learning with confirmation bias
This paper investigates a model of how confirmation bias, i.e. the tendency to ignore contrary evidence and interpret it as consistent with one's own belief, influences the way we learn from others on a network. Confirmation bias leads to slower learning and information loss in any network structure. We identify a subset of agents that become more/less influential with confirmation bias, and show that homophily is increasing in the strength of the bias. The networks that minimize the impact of the bias are symmetric, unweighted, regular and minimize average path length. In an application, we show that confirmation bias leads to more volatile elections in a standard voting model.