SOME COMMENTS ON MISSPECIFICATION OF PRIORS IN BAYESIAN MODELING OF MEASUREMENT ERROR PROBLEMS

In this paper we discuss some aspects of misspecification of prior dis tributions in the context of Bayesian modelling of measurement error p roblems. A Bayesian approach to the treatment of common measurement er ror situations encountered in epidemiology has been recently proposed. Its implementation involves, first, the structural specification, thr ough conditional independence relationships, of three submodels a meas urement model, an exposure model and a disease model - and secondly, t he choice of functional forms for the distributions involved in the su bmodels. We present some results indicating how the estimation of the regression parameters of interest, which is carried out using Gibbs sa mpling, can be influenced by a misspecification of the parametric shap e of the prior distribution of exposure.