Advances in Business Statistics, Methods and Data Collection

35Adopting Previously Reported Data into the 2022 Census of Agriculture: Lessons Learned from the 2020 September Agricultural Survey

Linda J. Young¹, Joseph B. Rodhouse¹, Zachary Terner^1,2, and Gavin Corral¹

¹USDA National Agricultural Statistics Service, Washington, DC, USA

²National Institute of Statistical Sciences, Washington, DC, USA

35.1 Introduction

Each year the US Department of Agriculture's National Agricultural Statistics Service (NASS) conducts more than 100 surveys and produces more than 500 reports on all facets of US agriculture. As the agricultural industry has become increasingly concentrated, precise estimates of agricultural production require participation of the large producers. Consequently, the burden on these large producers has been increasing with some being asked to respond to more than 20 surveys each calendar year. This has contributed to increased nonresponse. In addition to surveys, NASS conducts the Census of Agriculture in years ending in two and seven. Although producers are required by law to complete the Census questionnaire, the penalty for failure to do so is small and seldom enforced. The response rate for the Census has been declining, falling from 78.2% in 2007 to 74.6% in 2012 to 71.8% in 2017. To reduce respondent burden, mitigate the rise in nonresponse, and improve data quality, NASS has begun exploring the use of previously reported data (PRD).

NASS has been reluctant to use PRD, or its methodological cousin‐dependent interviewing ...