Novel Methods for Incorporating Sample Designs in Bayesian Inference

• Activity Number: 215 - Evolving Survey Inference in the Big Data Era: Challenges and Opportunities
• Type: Invited
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Michael Elliott* and Yuqi Zhai and Trivellore Raghunathan
• Companies: University of Michigan and University of Michigan and University of Michigan
• Keywords: Survey sampling; importance sampling; small-area estimation; Bayesian bootstrap

Abstract

Blending the Bayesian paradigm, with its emphasis on complex modeling, with the survey sampling paradigm, with its emphasis on non-parametrics and robustness, has been difficult. Particularly problematic has been incorporating weights into analysis, which, depending on the setting, can be either unnecessary, helpful to avoid magnifying model misspecification, or required if sampling is informative. This talk outlines methods to use recently developed methodology for incorporating complex sample designs in a weighted finite population Bayesian bootstrap procedure (Dong et al. 2014; Zhou et al. 2016) to incorporate design effects into Bayesian analyses via importance weighting. We consider this approach in a few simulation settings, and discuss applications to accounting for complex sample design in the setting of small area estimation.