Page 45 - 2000 (5) Agricultur Census 2000 Balochitan Province
P. 45
LIMITATIONS OF CENSUS DATA
1. Like faces and watches, the all possible non-sampling errors,
estimates of sample censuses so that these may not loom large
are destined to differ with each in the census estimates.
other. For example, if one
conducts two sample censuses 3. While, comparing census data with
having the same design, same previous sample censuses, one
size and at the same time with has to take into account not
equal efficiency on a given only the estimates but
population, it is not necessary also the associated sampling
that the two w!ll give identical errors and changes I
estimates. The difference will be improvements incorporated in the
due to "sampling and non- census methodology. Further, the
sampling errors" which are bound successive censuses taken after
to occur. Therefore, the estimates a period of ten years, are bound to
presented in this report are reflect the differences not only due
inevitably subjected to sampling to passage of time and legacy of
as well as non-sampling errors as the Government policies but also
already discussed in detail. on account of the socio-political
conditions prevailing at the time
2. The complete elimination of non- of conducting a census. The
sampling errors is not possible in constitution of new tehsils, distr:icts
an operation so extensive and and divisions also renders the
wide spread as that of a census, comparison of historical data
despite the adequate training at these levels meaningless
of enumerators and elaborate without proper adjustment for
arrangernents for supervision. administrative bou ndaries.
Starting with the inaccuracies and
faulty comprehension on the part 4.. An agricultural census primarily
of enumerators, misapprehensions portrays utilization of agricultural
of the respondents, errors resources and provides base
introduced during processing and for structural information for
then during publication of results, formulating long term agricultural
all add up towards the non- policies. lt also provides base
sampling errors. The existence of to study relationships between
such errors might not be sut different factors of production by
rightly ruled out in the census farm size and by tenure.
estiinates. However, every effort The aggregates of eertain
was made to keep strict check on characteristies arrived at during
XLi
