Page 32 - 1990(3) Cesus Of Agriculture, provience report-Punjab
P. 32
questionnaires. During this pha~;e of various items. Editing was do_ne
it was also ensured that no section strictly in accordance with the
of any questionnaire was left written instructions and . all impor-
blank. These operations were · tant decisions were taken at ap-
performed , in Coding Section which propriate level in writing.
was organized as part of the
Control Section. . Data Entry and Verification
Manual Editing The manually edited question-
naires were passed on to the Com-
In order to ensure that data puter Centres of Federal Bureau of
on various characteristics · of one Statistics for . data entry and
farm are apparently coherent their verification.
intermatching was made in the light
of. their already known relation-, Data entry was carried out in
ships. Similarly, various totals free format (code and data) to ex-
and sub-totals were checked. In pli?di te data entry operations . and
view of the length ot question~aire avoid unnecessary key depres-
and interrelationship of various sions for blank fields. A few data
census items, a lbng list of edit- validity checks were also applied
ing checks was drawn. In the pre- at this stage to ensure accuracy of
vious censuses and their pretests, entered data.
it was established that in view of
co-i•nIDlexj_ t.y of relationships of cen- A sample verification of en-
sus ' data only a few Jtems could be tered data was carried out to main-
programmed for self correction on tain standards. of accuracy of data
computer,· and the rest had to be entry operators.
:manuetlly correctecl by examining the
original Gi@cumehts after the errors Computer Editing
were detected ~and pointed out by
computer. In view of the enormity A comprehensive computer edit-
oi manual correction work, that ar- ing programme was prepared and all
rangement proved extrem~ly time census returns were chec'ked on the
consuming and expensive and often c@mputer. Besides normal editing
fraught with danger of mis-editing· checks, · range tests were includ@d
under pressure . of high volume of to· identify Glata entry errors, if
work. Manual editing preceding .com- any. Errors found in data were ex-
puter editing· proved to be helpful hibited in the form 0f computer
~ in reducing the correction . work to print-outs. The error re.ports were
be . performed manually. Further, passed on to a special section in
this arrangement alsq fiiicilitated the offices of the Agricultural
the data entry operati0ns as il- Census . Organization for correction
legible . ~:~tri,~s were made legible by examining the o_riginal docu-
during edVting~ , :- ments. The corrections were
' 1 '' . .
inserted in . the· computer recorqs
AceorditrglY, the censi:is and the same were re-edited until
return:s were subjected to manual these were declared f~ee of errors.
editing during which a nu1nber of
checks particularly relatirig to Estimation of census results
cons•istency of data', were . applied
and illegible entries were clearly In. order to raise the sample
written to facilitate their key observations to population level,
punching. This work :. was entrusted an . overall raising factor repre- ·
to the Census Master Trainers 'i'lho s'enting the . inverse of the proba-
had already worked in the field and bility of selection of respec~ive
were well acq-nainted with the cen- . household was f°ed to th:e . computer
sus concept's <and interrelationship along with dth~r data. in case of
(XXX)
I
