Extending the Dual-System Estimation for the Census of Agriculture

• Activity Number: 430 - Record Linkage and Auxiliary Data Sources
• Type: Contributed
• Date/Time: Wednesday, August 10, 2022 : 10:30 AM to 12:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Habtamu Benecha* and Luca Sartore and Grace Yoon and Bruce A. Craig and Denise A. Abreu and Linda J Young
• Companies: USDA National Agricultural Statistics Service and USDA NASS and National Agricultural Statistics Service and Purdue University and USDA National Agricultural Statistics Service and USDA National Agricultural Statistics Agency
• Keywords: Capture-Recapture; Coverage; Non-Response; Misclassification; List Frame; Area Frame

Abstract

USDA’s National Agricultural Statistics Service (NASS) conducts the Census of Agriculture every five years. The Census is the leading source of information on U.S. agriculture, providing characteristics of farms and the people who operate them. However, the Census Mail List (CML) is incomplete. That is, the CML does not contain all operations that are farms, and some records on the CML are not farms. To quantify the incompleteness in the CML, NASS uses the June Area Survey (JAS), which is based on an area frame. Census weights are adjusted by applying a capture-recapture method that accounts for undercoverage, non-response, and misclassification. Historically, only JAS data and CML records that are linked to the JAS are used for producing these adjustment weights. This paper proposes an alternative capture-recapture approach that utilizes all the available JAS & Census information for the estimation of Census weights. Results from simulation studies and an application of the method to data from the 2017 Census of Agriculture are presented.