Rethinking Rationality: Expert Performance in Bayesian Reasoning Challenges Dominant Theories

We tested Bayesian reasoning performance across experts in math and statistics (n=31), novice academics (n=33), and novice students (n=55). Experts greatly outperformed both novice groups (accuracy: 90.3% vs. 30.3% and 12.3%). This novel finding challenges current theoretical accounts of Bayesian reasoning in favour of an expertise account.