Page 24 - 1990(3) Cesus Of Agriculture, provience report-Punjab
P. 24
number of households according to ting very small probabiljties were
1981 Population Census. merged together and treated as one
mouza for sampling purposes.
FIRST STAGE SELECTION OF
PATWAR CIRCLES: The nµmber of THIRD STAGE SELECTION OF
patwar circles to be selected from CLUSTERS OF HOUSEHOLDS: A sys-
each sub-division _was determined tematic selection of clusters of
keeping in view the number of nearly 30 households each was made
enumerators which could be with equal probability from all the
effecti~ely trained and properly selected mouzas with the help of
supervised besides inherent the list of households. The number
variance of the population in of clusters to be selected, with a
respect of items under study. The minimum of two, was determined with
patwar circles falling · in various regard to the raising factor ot
kanungo circles were arranged by mouza and the pre-fixed overall
their cultivated area within the raising factors at cluster level,
respective kanungo circles. Each for a parti9ular area. The limit of
patwar circle was given a overall raising factor wa~ fixed at
probability proportional to its 18 for rural settled areas of NWFP,
measure of size in the sub-division Sindh and Balochistan and at 24 for
multiplied by the number 6f patwar those of the Punjab. In exceptional
circles t.o be selected therefrom. cases like big cities of Lahore and
Patwar circles getting a Karachi, the limit was allowed to
probability of 0:9 or more were go upto 34 and 40 respectively. ,
s~lected with certainty and sifted
out from the list. The remaining In this manner 1. 14 million
non-certainty patwar circles were households representing roughly 9
re-assigned a probability by percent of the total households
repeating the above mentioned were selected for enumeration. They _
process. Cumulative probability was included: ·
worked out and shown against each r
patwar circle in a sub-di vision. i) The households contained in the
For the . first sub-di vision to be selected clusters.
sampled, a random number over zero .
and upto o. 999 was se·lected. The ii) The households located in the
first patwar circle with cumulative selected urban w~rds~
probability equal to or greater
than the · random number , was thus iii) The households located in the
selected. Simi,larly ,- random number selected mouzas designated as
·plus 1 ( one) selected the second Mouza Certainty Holdings;
patwar circle, random number plus 2
(two) selected the third and so on iv) The households located . any
until the allocated number of pat- where in the country included
war circles was selected. in the list of National Cer-
tainty Holdings.
SECOND STAGE SELECTION OF
MOUZAS: It was decided to select The procedure adopted for
two mouzas per selected patwar selection of sample, involved sub-
circle. Mouzas falling in certainty stantial calculation worl. It was
and selected patwar circles were feared that . while • the design was
arranged by their cultivated area effective in red~cing sampling er-
within respective pat war · circles. rors, its implementation might in-
The method employed in the f trs.t troduce considerable non-sam~ling
stage was repeated at patwar circle errors. The use of electronic com-
level ' for se_gregating the certainty puter for the first and second
mouzas and the selection of non- stage selections and working out
certainty mouzas. The mouzas get- feasible sampling fractions for the
(xxii)
