Bayesian Spatial Methods and Public Opinion of Geographical Areas

Abstract

Abstract Studying subnational public opinion is often impeded by the limited sample size of survey data. This chapter provides an overview of the issues and methodology involved in estimating public opinion of geographical areas and an introduction of the Bayesian spatial methods. The work focuses in particular on the Bayesian kriging model, which provides a cohesive framework for incorporating different information and for measuring area-level public opinion. The chapter introduces its technical details and related advantages. It then illustrates the method through an example of constituency ideology and concludes with views on the importance of applying various methods to appropriate situations.