A1516
Title: Bayesian alternatives to model the variances of direct estimates
Authors: Sirapat Watakajaturaphon - University of California, Davis (United States) [presenting]
Abstract: The estimates of variances in small area estimation play an important role. Common problems of assuming a frequentist framework for modeling the variances of small area estimates are discussed and a new Bayesian framework is proposed to deal with these problems in practice. Suitable Markov chain Monte Carlo algorithms are proposed, and the theoretical properties of the proposed model is studied. Finally, the model in a real data set is implemented.
