ABSTRACT
This chapter seeks to provide the basic tools for the study of the asymptotic properties of a Bayesian experiment. It reports the subsequent development of the asymptotic properties of conditional models and of finite parameter spaces is found in Florens, Mouchart and Rolin and Florens and Mouchart. The chapter describes those probabilistic tools which are most useful for the analysis of the properties of the limit of a sequence of conditional independence properties. It presents some asymptotic properties of sequential experiments before considering the asymptotic admissibility of marginalization or of joint reductions in an unreduced experiment. Te chapter examines the estimability from both a Bayesian and an asymptotic point of view. More specifically, in a Bayesian framework, consistency is exact estimability in an asymptotic experiment.
