Association for University Business and Economic Research
Member Login | Contact | Site Search

NASS to Issue Improved 2002 Census of Agriculture Data

USDA’s National Agricultural Statistics Service (NASS) announced that 2002 Census of Agriculture data tables were released on June 3. All farms in the United States are represented. Representation for all U.S. farms is an improvement over earlier census publications that included data only for farms on the census mail list. It has long been recognized that many small farm operations have been overlooked in the census process. NASS Administrator R. Ronald Bosecker commented “a new approach will present the truest picture of U.S. farms. In the past, the number of farms identified in the census could be quite different from other official estimates. This was because no allowance was made in the census for farm operations that did not appear on the census list. Different farm numbers raised questions and concerns among data users.”

To determine the detailed characteristics of farms and farm operators missed by the census mail list, NASS conducts an independent comprehensive survey of sample geographic areas. Using the characteristics of farms and farm operators identified in these area-based surveys, the key methodology improvement reweights the completed census forms for selected farms to compensate for very similar farms missing from the mail list. Appendix C of the June 3 publication provides a detailed explanation of the new procedures. Appendix C also summarizes the effects of reweighting on many key variables and size groups.

Most farms missed on the mail list are relatively small operations. These small farms are less likely to appear on lists used to build the census mail list. Minority operators will be better represented in the 2002 Census of Agriculture because higher percentages of these operations are in the small farm category. The new approach will also make comparability between censuses less dependent upon comparability in census mail lists.

Administrator Bosecker stated that “since the area samples were drawn at the state level, the adjusted numbers will have their best quality at the state level. Reweighting may result in some over- and underexpansions for specific items at the county level, but representation of items on all sizes and types of farms should be much improved.”

The 2002 Census of Agriculture contains the first census data set on a coverage-adjusted basis representing all farms to the county level. Most censuses since 1969 have used detailed interviews in sample geographic areas to measure census coverage for selected items. Only in 1978 was there a prior attempt to provide coverage totals for all items, but data were presented only at the state and U.S. level.

To aid data users, the 1997 Census of Agriculture files were reprocessed using the new procedures to provide comparable data for 1997 and 2002 in the June 3 tables. This will provide the bridge from the earlier census data series to the new levels. There are no plans to attempt to use the same methodology for any census data sets before 1997.

The 2002 Census of Agriculture is the only source of uniform agricultural data across the United States. To learn more about the 2002 census, visit NASS’s website at: www.nass.usda.gov/census/.

Back to the homepage