Abstract
In this paper we use simulation to examine the small sample behavior of a series of increasingly complicated stopping rules in the context of sequential estimation of a multivariate mean/regression parameter vector.This is done under a set ot hierarchical bayes models. some.but not all.of the rules are known to be he non-deficient [Ghosh and Hoekstra ('89, '95)] Thus.this study extends the work of those papers to finite sample number behavior.