A prior probability provides the starting point for Bayesian reasoning. It may come from historical data, established scientific knowledge, a previous posterior, or a deliberately broad assumption when little is known. For several possible values, the prior is represented as a probability distribution.
Priors make assumptions explicit rather than eliminating them. An overly strong or poorly justified prior can dominate limited evidence, while a weak prior may allow noisy data to have too much influence. Sensitivity analysis shows whether reasonable prior choices materially change a conclusion.
Acronyms and aliases
prior belief synonymprior distribution variant
Related terms
Frequently asked questions
Where does a prior probability come from?
It can come from historical evidence, expert knowledge, earlier analysis, a previous posterior distribution, or a deliberately broad starting assumption.
Can two analysts choose different prior probabilities?
Yes. They should document and test reasonable alternatives so readers can see how much the final result depends on that choice.
