Proceedings of the Annual Meeting of the Cognitive Science Society

It is widely assumed that when belief polarization occurs – people updating their beliefs in opposing directions in response to the same evidence – people's reasoning must be in some way irrational. Yet, several authors have shown that Bayesian reasoning can, theoretically, cause polarization. I provide an accessible explanation for why Bayesians polarize, unifying some previous approaches. I elucidate the conditions under which Bayesian polarization can occur in two scenarios, providing non-technical explanations, with examples, and using Bayesian Networks to set up mathematical proofs. I show how these two scenarios can provide explanations for some prior documented cases of belief polarization in the literature.