The task of assessing posterior distributions from noisy empirical data imposes difficult requirements of modelling, computing and assessing sensitivity to model choice. Seasonal analysis of economic time series is used to illustrate ways of approaching such difficulties.
Discussion on the papers by Zellner, Arnold and Siow, Aloysius, Posterior odds ratios for selected regression hypotheses and by Bernardo, José M., A Bayesian analysis of classical hypotheses testing, both of them part of a round table on Hypothesis testing held in the First International Congress on Bayesian Methods (Valencia, Spain, 28 May - 2 June 1979).
Page 1