Replication 
Replication is utilized to improve efficiency and 
reduce respondent burden. To adjust for 
nonresponse associated with criteria records in the 
2007 Census of Agriculture, NASS replicated a set 
of respondents determined to be in-scope from the 
last mailing of the Agricultural Identification Survey 
(AIS), conducted in December 2006. The replicated 
records represented operations that were relatively 
small in size and homogeneous in nature. Replicated 
records were assumed to be in-scope, based on their 
AIS reported data. 
For the 2012 Census of Agriculture, a first mailing 
was sent to the criteria records, a subpopulation 
consisting of all of the approximately 74,000 
respondents to the 2011 NACS mailing. This 
included pre-notification using a pre-recorded 
message, the first mailing, and the thank-you 
reminder post card. No further follow-up efforts 
were conducted on this subpopulation. As in 2007, 
the agricultural operations in this subpopulation were 
relatively small in size and homogeneous in nature. 
The responses from the criteria records were used to 
estimate the in-scope rate for the 20,168 
nonrespondents from this subpopulation. 
Records were selected randomly for replication or 
coding as out-of-scope based on the estimated in- 
scope rate. The use of the in-scope rate after one 
mailing is supported by analysis of 2007 census data, 
which indicated the early in-scope rate was a 
reasonable proxy for the in-scope rate for the 
subpopulation of criteria records that did not respond 
to the NACS immediately preceding the census 
mailing. Of the 20,168 NACS records with no 
response, 16,762 records were selected to be in- 
scope. 
Data relationships between the 2012 responses and 
their respective NACS data were applied to the 
NACS data for the nonrespondents selected to be in- 
scope to derive values to seed replication. Then 
replication was conducted through imputation. 
Criteria records with no response to the December 
2011 NACS were excluded in the capture-recapture 
adjustments for coverage, response, or correct 
classification. The in-scope records were each given 
an initial weight of one. However, for calibration, the 
2012 Census of Agriculture 
USDA, National Agricultural Statistics Service 
replicated in-scope records were eligible for a 
coverage adjustment. 
REPORT FORM PROCESSING 
Data Capture 
The Census Bureau’s National Processing Center 
(NPC) in Jeffersonville, IN was contracted to 
process returned mail packets. NASS staff on site at 
the NPC provided technical guidance and monitored 
NPC processing activities. All report forms returned 
to the NPC were immediately checked in, using bar 
codes printed on the mailing label, and removed 
from follow-up report form mailings. All forms with 
any data were scanned and an image was made of 
each page of a report form. Optical Mark 
Recognition (OMR) was used to capture categorical 
responses and to identify the other answer zones in 
which some type of mark was present. 
Data entry operators keyed data from the scanned 
images using OMR results that highlighted the areas 
of the report forms with respondent entries. The 
keyer evaluated the contents and captured pertinent 
responses. Ten percent of the captured data were 
keyed a second time for quality control. If 
differences existed between the first keyed value and 
the second, an adjudicator handled resolution. The 
decision of the adjudicator was used to grade the 
performance of the keyers, who were required to 
maintain a certain accuracy level. 
The images and the captured data were transferred to 
NASS’s centralized network and became available to 
field offices and headquarters on a flow basis. The 
images were available for use in all stages of review. 
Images were computer generated for reports 
obtained from the telephone interviews and the 
Internet. 
Editing Data 
Captured data were processed through a computer 
formatting program, which verified that records were 
valid - that the record identification number was on 
the list of census records, that the reported counties 
of operation and production were valid, and other 
related criteria. Rejected records were referred to 
APPENDIX A A- 7 
