A comparison between p-values for goodness-of-fit checking

M.J. Bayarri and M.E. Castellanos

The problem of checking the compatibility of a proposed parametric model with the observed data is an old one. If no alternative models are proposed, Bayes factors are precluded and only measures of `surprise' can be given. The p-value is, by far, the most ubiquitous of such measures. Also, model checking is often done in a `casual' manner, using statistics that are not distribition-free, nor even parameter-free under the assumed model. In these situations, we compare several p-values that can be used for model checking. We investigate both their distribution under the `null' model, and their power under families of alternatives.

Postscript File (792 Kb)