Bayesian analysis of liner models with two random components with special referance to the balanced incomplete block design

Abstract

SUMMARY

In this paper we consider liner models which contain two random components and enable representation of general designs with block effects random. We show how, within a Bayesian framework, to combine the intra-and inter-block information about the location parameters. The balanced incomplete block design, an important special case, is considered in detail and illustrated by an example. A simple approximation to the marginal distribution of an individual contrast is also presented and illustrated.