Variance Modeling in the U.S. Small Area Income and Poverty Estimates Program (SAIPE)

• Activity Number: 86
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Sam Hawala*+ and Parthasarathi Lahiri
• Companies: U.S. Census Bureau and University of Maryland
• Keywords: Small Area Estimation ; Generalized Variance Function ; American Community Survey

Abstract

In the SAIPE program at the U.S. Census Bureau, one of the challenges is the estimation of sampling variances of the direct survey weighted estimators for the counties. Design-based methods can be highly unreliable primarily because of small sample sizes in the area. GVF methods have been previously used in SAIPE to obtain smoothed variance estimates. In this paper, we consider a synthetic approach in which we select a working model by comparing several competing models at the state level where standard direct design-based variance estimates are considered to be reliable. The working variance model relates replication based variance estimates to the corresponding estimates and to survey design variables. It motivates an integrated multilevel model that can be used in the future to produce improved estimates of different parameters of interest simultaneously.