Inference under Informative Sample Designs
Conference
Category: International Association of Survey Statisticians (IASS)
Abstract
Model-based analyses of survey data are crucial for analytic inference and small area estimation. When the selection probability and the response variable are dependent, after conditioning on model covariates, the design is informative for the model. Ignoring an informative design can render invalid inferences. This session offers innovative methods for small area estimation and analytic inference in the presence of an informative design.