Who knows what? Bayesian competence inference guides knowledge attribution and information search

One of the main challenges of social cognition is inferring the competence of others, which often occurs in contexts of limited information. Recent research suggests that people can successfully infer the competence of others through Bayesian inference, but prior studies have relied largely on artificial stimuli and competence benchmarks, leaving it unclear whether these rational principles generalize to naturalistic settings with rich prior knowledge. Using trivia questionnaires, we test whether people can infer others' competence and search for informative evidence in a near-optimal way, consistent with rational Bayesian principles. In Studies 1 and 2, participants were presented with an individual's performance on a trivia question and predicted the individual's ability to answer other trivia questions from the same theme. Replicating and extending past results, we observe that participants accurately predict performance from limited information. Computational modeling shows that participants' inferences are better described by Bayesian processes than by plausible heuristics, suggesting that participants rationally integrate new information with their prior expectations about others' competence. Study 3 shows that participants can select which information would be most diagnostic for inferring an individual's competence, again in a manner consistent with Bayesian rationality. Overall, our results suggest that people approximate a rational Bayesian model both when searching for and when integrating information about others' competence.