Definition
The Bayesian-frequentist debate is the foundational dispute in probability and statistics over what probability means and how statistical inference should be conducted. Frequentists hold that probabilities are objective long-run frequencies — facts about the world to be measured. Bayesians hold that probabilities are degrees of belief — coherent measures of an agent's uncertainty, updated by Bayes' theorem.
The debate is more than a philosophical curiosity. It shapes how clinical trials are designed, how scientific claims are tested, how legal evidence is weighed, and how machine-learning systems are built.
Why it matters
How it works
A frequentist computes probabilities of data under hypothesised models — p-values measure how surprising the data are if the null hypothesis is true. A Bayesian computes probabilities of hypotheses given data — posterior distributions express how plausible each hypothesis is after the evidence.
In large-data, well-specified problems the two camps usually agree on numerical answers. They diverge most when data is scarce, prior information is strong, or the model is mis-specified — and especially when communicating with non-experts, because frequentist outputs are notoriously easy to misinterpret as Bayesian ones.