Page 39 - 2010 (3) Agricultural Census 2010, Punjab Report

Percent Sampling Error, i.e.

CV:Jw xloo

tf the RV for the aggregate of characteristic (W) for district (g) i. denoted by
RVYII/rand Y,u, is the estimate for the aggregate (f)of characteristic (W)tor
district (g), tn"n ti'te RV for the province with g: I to u districts in the province.

ilnvvw,)ri,
RVYW: c

[?t")'
and

U / \ ^a
ZlnrxwE)x("s
o'
o
RVXW = / i2-
lu ^
'tsv I I
/,"\yg
| I
to I
\o ,/
ft Z estlmates for country levels have also been worked out on the same pattern as
at the province level.

NON.SAMPLING ERROR:

ln Agricultural Census 2010, the non-sampling errors were minimized through careful
planning, elaborate training, intensive care, close supervision and best techniques /
efforts to help the -respondents to provide the best data to be recorded on the
questionnaires. The utmost efforts were made at different levels to minimize the non-
response during fleld enumeration. Comprehensive editing of census questionnaires
in the light of already developed checks / relationships was made before entering the
data. During data processing, the edit checks were applied logically to "clean" the
data and to make it error free before producing tabulations. After producing
tabulations, quality control checks developed for the purpose were applied to verify
the consistency among data tabulations.

CONSISTENCY OF DATA:

Consistency and plausibility of data were adjudged through a number of
interrelationships between virious census items. The users of census data would
certainly find the required:fineness in the quality of census estimates reflecting a
reasonable degree of consistency.

COMPARISON WITH CORRESPONDING SERIES:
The census estimates were also checked / compared with the corresponding series
as far ds possible at the district, province and country levels. The discrepancies
identified were sorted out diligently by digging into the source documents, raising
factors, computational process etc., and corrected after careful considerations-