Online Program

Agile Processing: Modifying the Editing and Imputation Procedures to Quickly Handle Questionnaire Changes to the 2017 Census of Agriculture (307948)

Keywords: editing, imputation, agriculture

The census of agriculture (COA) is the only source of uniform, comprehensive agricultural data for every state and county in the United States. The United States Department of Agriculture’s (USDA’s) National Agricultural Statistics Service (NASS) conducts the COA every five years, in years ending in 2 and 7. In 2015, a panel of experts recommended that the COA questionnaire should be updated to include information collected about women and new or beginning farmers. Subsequently, NASS redesigned the demographics section of the 2017 COA. Some of the updates included allowing multiple principal operators and adding detailed farm operation decision-making questions to the 2017 COA. This major redesign to the questionnaire required changes/updates to downstream processes such as editing and imputation. One of the updates to the editing and imputation process included developing an algorithm to designate a primary producer when a primary producer was not specified. We describe challenges that arise in some of the downstream processing to accommodate major changes to the questionnaire in a short time frame and how multiple divisions worked together to meet those challenges. We will also discuss the performance of the decision algorithm by comparing the frequency of respondent-designated primary producers to algorithm-designated primary producers.