On Bayesian Inference with Complex Survey Data.

Submit Manuscript | http://medcraveonline.com on updating assumed prior distributions with observed data likelihood. Much of this disconnect may be a function of differing goals; Bayesian approaches are focused on reliable statistical models [8] rather than on assessing the degree to which their estimates are nationally representative or not. However, Bayesian approaches, which have been successfully applied to multilevel data [8], missing data, and measurement errors [9], may represent a natural partner in complex survey data analysis. Measurement errors, missing data, and multilevel variables in complex survey data sets can be all treated as unobserved random variables in the Bayesian framework and they can be assessed by updating assumed prior distributions of related parameters with observed data sets [8,9].