associated with the sample estimate: 
SE (CR ) 
CV = ^100 % 
CR 
I 
where SE(CRi) is the standard error of the capture- 
recapture estimate for data item i. This relative 
measure allows the reliability of a range of estimates 
to be eompared. For example, the standard error is 
often larger for large population estimates than for 
small population estimates, but the large population 
estimates may have a smaller CV, indieating a more 
reliable estimate. For eounty-level estimates, a 
generalized coeffieient of variation (GCVs) was 
determined for each estimate within a State. A 
generalized varianee funetion relates a funetion of 
the varianee of an estimator to a funetion of the 
estimator. Within a State, the standard error of an 
estimate for a data item was often found to be 
linearly related to the estimate of that item with an 
intereept of zero. Based on this modeled relationship, 
the GCV is the slope of the line relating the standard 
error to the estimate, multiplied times 100 to 
represent the GCV as a pereentage. 
The standard error is the produet of the CV (or GCV 
for eounty estimates) and the estimate divided by 
100. As an example, if the GCV for a State is 25 
percent and a county’s estimate is 4, then the 
standard error is 25 (4)/ 100 = 1. The standard error of 
an estimated data item from the eensus provides a 
measure of the error variation in the value of that 
estimated data item based on the possible outeomes 
of the census eolleetion, ineluding variants as to who 
was on the CML, who returned a census form, who 
was miselassified either as a farm or as a nonfarm, 
and the uneertainty assoeiated with ealibration and 
integerization. With 95 pereent eonfidenee, an 
estimate is within two standard errors of the true 
value being estimated. For this example, with 95 
pereent eonfidenee, the estimate of 4 is within 2(1) = 
2 of the true eounty value. 
Table B presents the fully adjusted estimates with 
the eoefficient of variation for seleeted items. 
NONMEASURED ERRORS IN THE CENSUS 
PROCESS 
As noted in the previous seetion, sampling errors ean 
2012 Census of Agriculture 
USDA, National Agricultural Statistics Service 
be introdueed from the coverage, nonresponse and 
miselassifieation adjustment proeedures. This error 
is measureable. However, nonsampling errors are 
imbedded in the eensus proeess that eannot be 
directly measured as part of the design of the census 
but must be contained to ensure an aecurate count. 
Extensive efforts were made to eompile a eomplete 
and aecurate mail list for the census, to elicit 
response to the census, to design an understandable 
report form with elear instructions, to minimize 
proeessing errors through the use of quality eontrol 
measures, to reduee matehing error associated with 
the eapture-reeapture estimation process, and to 
minimize error assoeiated with identifieation of a 
respondent as a farm operation (referred to as 
elassifieation error). The weight adjustment and 
tabulation proeesses reeognize the presenee of 
nonsampling errors; however, it is assumed that 
these errors are small and that, in total, the net effect 
is zero. In other words, the positive errors eaneel the 
negative errors. 
Respondent and Enumerator Error 
Incorreet or ineomplete responses to the eensus 
report form or to the questions posed by an 
enumerator can introduce error into the eensus data. 
Steps were taken in the design and exeeution of the 
eensus of agrieulture to reduce errors from 
respondent reporting. Poor instruetions and 
ambiguous definitions lead to misreporting. 
Respondents may not remember aeeurately, may 
give rounded numbers, or may reeord an item in the 
wrong eell. To reduee reporting and recording errors, 
the report form was tested prior to the eensus using 
industry accepted eognitive testing procedures. 
Detailed instruetions for eompleting the report form 
were provided to eaeh respondent. Questions were 
phrased as elearly as possible based on previous tests 
of the report form. Computer-assisted telephone 
interviewing software ineluded immediate integrity 
ehecks of recorded responses so suspeet data could 
be verified or correeted. In addition, each 
respondent’s answers were checked for completeness 
and eonsisteney by the complex edit and imputation 
system. 
Processing Error 
Proeessing of eaeh eensus report form was another 
potential source of nonsampling error. All mail 
APPENDIX A A- 17 
