Belief updating is the process of changing your probability estimates in response to new information. Bayesian reasoning provides the mathematical framework: prior belief × likelihood of new evidence = updated posterior belief.
In practice, most people update their beliefs too slowly (conservatism bias) or in the wrong direction (backfire effect). Political scientists have documented systematic failure to update beliefs when confronted with factual corrections.
Healthy belief updating is a teachable skill. Research from the Good Judgment Project shows that forecasters trained in Bayesian updating outperform untrained peers by 14% on long-range geopolitical predictions. The key habits: seeking disconfirming evidence, tracking your prediction history, and being explicit about what evidence would change your mind.