An interval containing a specified probability mass of the posterior distribution (e.g. the central 95% credible interval). Unlike a frequentist confidence interval, it admits the direct probabilistic interpretation that the parameter lies within the interval with the stated probability, conditional on the model and data.