Bayesian Reasoning

Bayes' theorem, formulated by the Reverend Thomas Bayes in the 18th century and published posthumously in 1763, provides a mathematical framework for updating the probability of a hypothesis in light of new evidence. The formula relates the prior probability — what we believe before seeing new information — to the likelihood — the probability of observing the data if the hypothesis were true — to yield the posterior probability — our updated belief. This process of constant updating is what rationalists simply call "updating." In its most intuitive form, Bayesian reasoning tells us that the strength of evidence depends not only on how probable it is under the hypothesis we want to confirm but also on how probable it is under alternative hypotheses.

One of the most common errors in probabilistic reasoning is base rate neglect. When a medical test comes back positive for a rare disease, most people — including many physicians — dramatically overestimate the probability of actually having the disease because they ignore the low prevalence of the disease in the general population. Gigerenzer and Hoffrage (1995) demonstrated that presenting information in terms of natural frequencies rather than probabilities substantially improves people's ability to perform correct Bayesian reasoning. This finding has direct applications in clinical communication: telling patients that "out of 1,000 people, 10 will have the disease and of those 10, 9 will test positive; of the 990 who don't have the disease, 99 will have a false positive" is far more comprehensible than talking about 90% sensitivity and 90% specificity.

The rationalist community, especially through LessWrong and Eliezer Yudkowsky's writings, adopted Bayesian reasoning as the centerpiece of its epistemic philosophy. "Updating" one's beliefs in response to evidence became a fundamental value. Calibration exercises — where people make predictions with explicit confidence levels and then verify their accuracy — became a standard practice for improving epistemic rationality. This perspective has influenced the forecasting movement, where Tetlock's superforecasters demonstrate that disciplined Bayesian thinking can consistently outperform expert predictions.

In clinical psychology, Bayesian reasoning has important practical applications. Diagnostic reasoning — evaluating the probability that a patient has a particular disorder in light of symptoms and risk factors — is inherently Bayesian. The concepts of positive predictive value (PPV) and negative predictive value (NPV) are direct applications of Bayes' theorem. In evidence-based practice, the therapist updates their understanding of the case as new information arrives: treatment response, new symptoms, or test results. Training clinicians in Bayesian thinking can improve diagnostic accuracy and reduce errors stemming from base rate neglect and confirmation bias.