Linear contrasts for the one way analysis of variance: A Bayesian approach

• Linear contrasts between means are dealt with for the first time as objective Bayesian model selection problems.
• A specific solution for the homoscedastic case is proposed.
• The pp-value and the posterior probability of the null hypothesis are compared through calibration curves.

Linear contrasts between means for the one way analysis of variance are studied for the first time as objective model selection problems. For it, Bayes factors for intrinsic priors are used and classical and Bayesian measures of evidence are compared.