Small-Area Estimation for Government Surveys

• Activity Number: 422
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Bac Tran*+ and Carma Ray Hogue
• Companies: U.S. Census Bureau and U.S. Census Bureau
• Keywords: Cut-off sampling ; Decision-based ; Direct Estimator ; Synthetic Estimator ; Composite Estimator ; Step-wise Ratio

Abstract

Small sample sizes and no sampled units in subdomains are common problems in surveys. Recently we applied Small Area Estimation methodology to two major surveys in our Governments Division surveys: the Annual Survey of Public Employment and Payroll (ASPEP), and the Annual Finance Survey (AFS). In ASPEP we used the composite estimator to estimate the total gross payroll and employees for each state and function. In AFS we used the step-wise ratio to estimate the Revenues, Expenditures, Debt, and Assets for each state. The main idea is to borrow strength from larger areas as well as from previous years. The efficiency and the quality of the data in the sample were improved by applying Decision-based Estimation methods to a modified cutoff Probability Proportional to Size sample. The Decision-based Estimation improved model fit from which we benchmarked the final estimates.