Abstract:
|
Statistical agencies face demands for official estimates at increasingly finer levels of disaggregation. The United States Department of Agriculture's National Agricultural Statistics Service (NASS) is striving to meet such demands. NASS first implemented nationwide probability sampling for its crops county estimates in 2011. The underlying probability sample for crops county estimates has been an integral part of the production of its official statistics. The existing publication standard is based on either having sufficient number of reports (30 or more responses for that crop) or sufficient coverage (the sum of unweighted reports accounts at least 25% of the estimated harvested area total). NASS is transitioning to the use of model-based county estimates. For publication, these estimates need to protect confidentiality and be precise enough to be fit for use. At the same time, the publication standard must be easily described to NASS's clientele, who include farmers, researchers, other USDA agencies, and Congress. In this paper, we discuss the challenges and intricacies of developing such a publication standard.
|