Census of Agriculture 2022

Direct data capture was ensured by the CAWI method. The complete questionnaires received by mail were scanned and an Optical Mark Recognition (OMR) was used to capture categorical responses and to identify entries in numeric and alpha-character answer zones. Data entry operators keyed data from the scanned images. Answer zones with entries, identified in the earlier OMR analysis were presented to the data entry operators. The keyer evaluated the contents and captured pertinent responses. Ten percent of the identified answer zones were keyed a second time for independent quality control. If differences existed between the first keyed value and the second, an adjudicator handled resolution. Captured data were processed through a computer formatting program that verified that records were valid. Rejected records were referred to analysts for correction.

Accepted records were sent to a complex computer batch edit process. The computer edit determined whether a reporting operation met the qualifying criteria to be counted as a farm (in-scope). The edit systematically checked reported data section-by-section with the overall objective of achieving an internally consistent and complete report. The edit determined the best value to impute for reported responses that were deemed unreasonable and for required responses that were absent. Sources used for imputation included previous census and administrative data from the Farm Service Agency. When deterministic editing logic and previously reported data sources were unable to provide a current value, data from a reporting farm of similar type, size, and location were considered through the nearest neighbour imputation method. In cases where automated imputation was unable to provide a consistent report, the record was referred to an analyst for resolution.

Prior to publication, tabulated totals were reviewed by statisticians to identify inconsistencies and potential coverage problems. Weighting adjustments were used to address under-coverage, non-response and misclassification errors. A description of those adjustment can be found in Annex A of the final report. After tabulation and review of the aggregates, a comprehensive disclosure review was conducted. Farm counts were not considered sensitive and were not subject to disclosure controls. Cell suppression was used to protect the cells that were determined to be sensitive to disclosure of information. NASS used SAS software for editing and summary.