A posteriori

A posteriori (literally: ‘from the latter’) hypotheses are formed after observing data or empirical evidence, as opposed to being pre-defined. These hypotheses are developed through inductive reasoning and are used to explain the underlying cause of an observation or unexpected finding. In statistics, a posteriori tests (post-hoc tests) are not planned before data collection. Compared with a priori tests, this can introduce bias (and, therefore, can be viewed with some scepticism) because the choice of tests and methods can be influenced by the observed results. Multiple comparison methods can be employed to correct for this potential bias.

See posteriors for the definition within a Bayesian context.