Page 35 - 1990(3) Cesus Of Agriculture, provience report-Punjab

work on computer programme w_r i ting that all the national and provin-
initiated at the Lahore Centre. cial totals are subjected to a sub-
Computer programmes were tested on stantial rounding effect as in a
specially prepared Test Deck to en- large majority of the cases; · such
sure that the same were error free. effect cancel ·out each other at
As soon as the finally edited data various stages, diminishing its
for a sub-di vision/ district . ·was ~ize at the natiorial or provincial
ready, the tabulation for that sub- level to a minimum . Similar . dif-
division/district was carried out. ferences may be noticed in ' Number
In all, 90 tables repeated a~ dis- of Farms' reporting a certain
trict, division, province a:hd na-· characteristic as available in acre
tional levels presenting the census and hec;::tare version of a table.
data and their various cross tabu- Al though the acre and hectare
lations were prepared. At country t'abulations are based on the same
and province levels, the census data and multiplied with a similar
tabulations were presented both in raising factor, yet the use of
acres as well as in hectares. slightly different size classes and
separ?te processing for the two
Quality control sets· of data subjected different
· total to a di f ferent extent of
In order to safeguard against · rounding effect.
the publication of any glaring er-
rors, the census tabulations were Publication of Ce nsus Data
checked for quality. Data were ex-
i:lmined to investigate errors th,at The census data have been
could not be detected or corrected tabulated at sub-division, dis-
during editing or any errors w~ich trict, division , province and
might have crept in during coding, national levels. The tabulations
punching, verification, processing are published in ·the following
and tabulation. The staff had been reports:
specially trained to check the in-.
consistencies in various tables and Preliminary Conta ining tabulations
to detect inter-table differ_ences Report at national, province,
and probe backward to detect the division and district
discrepancies; which were th~n cor- levels based on un-
rected after careful consideration. edited data for
important census i terns.
In some cases, the figures in a Country and province
row or ·a · column may not add up tables are both in
exactly t6 their respective row or acres and hectares and
column total ~ These differenc~s are other in acres only.
due to rounding which was un-
avoidable in view of the fact that Vol.I Containing tabulation
ea9h area figure •was collected in . All Pakistan at country and province
terms of acres and kanals and con- Report Levels both in
verted into acres and hectares. ( Excluding. acres and hectares.
Further, each data item was multi- Northern
plied with its raising factor Areas)
worked out upto three places of
decimal. The si&e of rounding ef- Vol.2 ·consisting of four
fect in a total normally depends Provincial parts i.e. one for
upon the number of independently Reports ea6h province contain~
worked out components included ing tabulations at ·
therein. Consequently, it will be ·province, division and
larger in i n~tional or provincial district levels in
total than in a district total. It acres only.
should not, however, be presumed

Cxxxiii)

I