Data Reconciliation and Estimation in an Agricultural Survey

• Activity Number: 269 - Combining Information from Multiple Sources for Various Purposes
• Type: Topic Contributed
• Date/Time: Tuesday, August 4, 2020 : 1:00 PM to 2:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Habtamu Benecha* and Linda J Young and Denise A Abreu and Rachael Jennings
• Companies: USDA/NASS and NASS and NASS and USDA/NASS
• Keywords: Area-frame ; Census Mailing List; Capture-recapture; Data Reconciliation; Undercoverage; Misclassification

Abstract

USDA’s National Agricultural Statistics Service (NASS) conducts the Puerto Rico Census of Agriculture every 5 years in years ending in 2 and 7. The census is the leading source of statistics about Puerto Rico's agricultural production. To measure undercoverage in the census, NASS collects relevant data using the Agricultural Coverage Evaluation Survey (ACES). The ACES is based on an area-frame that covers all land in Puerto Rico. Census estimates are produced by using a capture-recapture approach that accounts for bias due to undercoverage, non-response and misclassification. Because of the impacts of Hurricane Maria, preparation for the 2017 Census was disrupted and NASS decided to conduct the Puerto Rico Census of Agriculture in 2018, thereby delaying the census by one year. Although the 2017 ACES was conducted as planned, NASS repeated the survey in 2018 using the 2017 sample. In this paper, we discuss the estimation of the number of farms from the census and fully analyze the 2017 and 2018 ACES to determine the number of farms that began and went out of business between the 2017 and 2018 ACES.