Page 40 - 2000 (5) Agricultur Census 2000 Balochitan Province

PUBLECATION SF GENSUS *AT'A affected by a number of lapses taking
The census data as mentioned
above have been tabulated at district,
divislon, province and national levels.
The tabulations thus PrePared
published in the fallowing reports :-
Vol.1: All Pakistan Report contains
tabulations at country and
province levels in hectares for
the local as well as foreign
users.
Vol.2: Provinciai Reports consist of
four parts, i.e. , one for
each province. Each contains
tabulations at Province,
division and district levels in
acres for the local users.
Vol.3: Specia! Reports consist of
two reports, i.e. one each for
Northern Areas and Azad
Jammu & Kashmir. Each
contains tabulations at the
district and overall levels in
acres for the local users.
QUALffY OF CENSUS DATA
eensus taking is a gigantic
project of nailcnal imPortance
involvinE a great deal cf sophistieation
in lts planning, executlon and
proeessinE. In spite of b*st efforts to
*nsljr* quaiity at eaeh stag*, some
margin cf err*r !s always pcssible.
!n ieehnEcal tei'ms, the qualitY of
the census ciaia gathei"*ci through
semple enumeraii*n might had been
place in the course of all the staEes
of: conducting the census as well
as processing the census data.
Therefore, all data estimates arrived at
are the basis of sample enumeration are
destined to have errorsi as usual. All
such errors are classified into (1)
sampling errors; and (2) non-sampling
errors. So, the important consideration
would not be that the estimates are
error free but that of controlling the
extent of error, both, in sampling and
non-sampling errors.
ln general terms thg qualitY of
the census data may be adjudged
on the criterion of (1) internal
consistencies and reasonableness;
and (2) by comparison with other
corresponding series. The estimates
presented in this report were checked
for quality for technical terrns as well
as for general terms.
t1) Sampling Error
Sampling error is a measure to
determine the Precision I
consistency of the samPle
results, i.e. the closeness of
sample results to those
obtained through comPlete
count conducted with exactlY
sirnilar arrangements of training,
enumeration, suPeryision, Pro-
cessing, etc. ln other words,
sarnpling error is not 'error', as
the word error portraYs, but is
the probable difference beiuYeen
the populatlon mean and the
XXX\ii