A Multinomial and a Uniform Unit-Level Model for Small Area Estimation

• Activity Number: 613 - Practical Applications of Small Area Estimation
• Type: Topic Contributed
• Date/Time: Thursday, August 3, 2017 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Donald Malec*
• Companies: National Center for Health Statistics
• Keywords: Nonparametric Bayes

Abstract

The development and evaluation of a hierarchical multinomial and a hierarchical uniform model for making small area estimates is presented. By limiting these types of models to include constraints at a higher geographic level, many of the design-based characteristics retained by a multinomial or a uniform model may still be retained while borrowing strength from a higher geographic-level model. An evaluation using data from the 2008-2012 NHIS to make estimates of access to care for the counties in States that border Mexico is presented. Here, the model uses sample segments as the unit-level and accounts for the differential selection of segments from within-PSU strata. Covariates available from the U.S. Census Bureau's "American Factfinder" are used in the county-level model.