Citation Context

... an arbitrary extent with appropriate choice of priors and posteriors. This is not a new observation, and even proponents of prior odds analysis admit to “some arbitrary aspects” (see e.g., Robert A. =-=Connolly, 1991-=- p. 57). Donald W. K. Andrews (1994) provides a correspondence between classical hypothesis tests and Bayesian posterior odds ratios, showing that for certain choices of priors the Bayesian posterior ...