Page 37 - 1990(3) Cesus Of Agriculture, provience report-Punjab

under: u
~
t, rl q. N -N' -N" k y' ,i, g=l
k
k
hijkl
p:; J -------- + RVYW = ------- --- .------- -
yh' 1=~. ~
J k l" I Phijk u ,I\ 2
( ~ y ).
Q ; N" g=l - wg
k y~ijkl"
+1::3 ~ -------] and
k l'' p~ij u
x_
~ "2
g=l wg
r2 q. N -NI -N" RVXW = -- -----------
I • k . k k y~2jkl"'
[-~ J -~ -------- + u
yh' 2=~. l , 2
J k l" I p~2jk ( -~ X )
g=l . wg
Q. N" k yh2jkl"
+~J !: -------] Rel-variaric.e estimates for
k l" P112j .provinces and country level have
also been worked out on the same
then, Rel-variance for the sub- pattern as at district level.
division estimate of the attribute
is obtained by taking square of Different ca~es of Sampie
differences between estimates for Selection Conditions
succe~sive pairs of half strata
divided by the Sub-division as fol- It would have been better for
lows: variance estimation purposes, if
there had been no such occurrenc es
I' ,._ in both ~ertainty and non-certainty
, 2 L 1 . y h ' 1 - y h ' 2 ,2 mouzas within certainty patwar
(C.V.Y) ~ [-----~-----]
h' y circles. But an agricultural census
is . a practical exercise and one
must cope with problems ·as they oc-
Similarly; r~l-variance for
the sub-di vision -estimate Qf uni ts cur; although, in hindsight, some
of tnese "problems" could have been
having the attribute is:
averted. Fortunately over _96 per-
cent of the sample falls into the
L' "" I\
r 2 · xh • 1 - xh • 2 2 standard pattern and the excep-
(C.V.X) ~ [-------- . --] .
/, . tional cases requiring special
h X
treatments in variance calculations
amount to less than 4 percent, The
Percent Sampling Error.
standard piocedure for c~lcul~ting
variances calls for stage by stage
( c. V) = J Rel. Variance x 100
expansion of the household obi~rva-
tions in each sample patwar circle
Rel-Vari~nce Estimates at District, to ~he stratum level and tak ~~q
Province and country le~els
square of the differences bejtween
the so expanded values of the two
If the rel-variance for sub- patwar circles in each stratum. The
di vision g is denoted -by RVYWg and special cases requiring some depar-
Y is the estimate of the ag- ture from the standard procedure
g~igate for sub-di vision g_, t~en for · various calculations are as
the rel-variance for the district follows:
with g=l to · u sub-~ivisions in the
district is i) Two non-certainty mouzas
selected in a Certainty Patwar
Circle in which case the

(lCXXV)