Introduction to Bayesian Statistics

Alan Agresti, Maria Kateri, Ranjini Grove et al. · Chapman & Hall/CRC eBooks · 2026

The frequentist approach to statistical inference applies probabilities and probability distributions to random variables that yield the observed data. For that approach, parameter values are fixed values and do not themselves have probability distributions. By contrast, Bayesian statistical inference applies probabilities and probability distributions to parameters as well as to data. Using Bayes&s; Theorem, it finds probability distributions for parameters, given the data, using probability distributions for the data, given the parameters. It assumes a prior distribution for the parameters, chosen before observing the data. Using Bayes&s; Theorem, the prior distribution combines with the information that the data provide through the likelihood function to generate a posterior distribution for the parameters. Bayesian statistical inferences for a parameter that are analogs of frequentist point estimators, confidence intervals, and significance tests are based on the posterior distribution of that parameter.

Cited by 0 works. doi:10.1201/9781003715924-1

Authors