Agriculture Reference
In-Depth Information
Table 6.1
Comparison of sample survey method and census method of data collection
Sl no.
Sample survey method
Census method
1
Only a representative part of the population (sample) comes
under investigation
Every element of the population comes
under investigation
2
Comparatively less accurate, if not done properly
Accurate
3
Economical
Costly
4
Less time- and labor-consuming
Time- and labor-consuming
5
Possible in case of infinite population
Not possible for infinite population
6
Possible for large population
Rarely possible and feasible for large
population
7
With both sampling and non-sampling errors
With absent sampling errors
8
Nonresponse errors can be solved
Nonresponse problem cannot be worked
out
9
Parameters are to be estimated and tested
Parameters are directly worked out
10
Used frequently
Rarely used (e.g., population census)
Parameter
is a real valued function of the
population values. For example,
of sample is more reliable and acceptable than
from a biased sample.
The
Sample size
n
) of a sample is the number
of elements/units with which the sample is
constituted of. Sample size is determined by a
number of factors, namely, (1) objective and
scope of the study, (2) nature of population and
sampling unit, (3) sampling technique and estima-
tion procedure to be used, (4) structure of
variability in the population, (5) structure of time
and cost component, and (6) size of the popula-
tion.
(
N
i¼
1
Y
i
;
1
N
¼ Y ¼
Population mean
N
i¼
1
Y
i
Y
1
N
2
2
Population variance
¼ σ
Y
¼
ð
Þ
;
¼ C
Y
¼
σ
Y
Population coefficient of variation
Y
:
: A sample is a representative part of a
population. If the sample fails to represent the
population adequately, then there is every chance
of drawing wrong inference about the population
based on such sample because of the fact that it
will overestimate or underestimate some popula-
tion characteristics. Let us suppose that we want to
know the average economic background of the
students of an agriculture college. If the college
is a coeducation college and consists of students
from different parts of the country and one draws a
sample (1) of either boys or girls only, or (2) from
a particular class, or (3) from the students of a
particular state only, then the average economic
condition obtained from the sample may fail to
infer about the true average economic condition of
the students of the college (the population). This
type of sample is called a
Sample
An efficient and optimum sample size either
minimizes the mean squared error of the estimator
for a fixed cost or minimizes the cost for a fixed
value of mean squared error
.Fixingofoptimum
sample size becomes complicated when more
than one parameter is to be estimated or more
than one variable is under study. In fact, it is
very difficult to have a fixed rule for getting the
sample size. However, based on past information
or information gathered through the pilot study
conducted before the main study and giving
due consideration to the above decisive factors,
sample sizes are fixed for specific studies.
Generally a sample is regarded as
large sample
ifthesamplesize
.
Krejcie and Morgan in their 1970 article
“Determining
n
30; otherwise,
small sample
Sample
Size
for Research
biased sample
.Onthe
Activities”
(
Educational and Psychological
other hand, an
is statistically
almost similar to its parent population, and thus,
inference about the population based on this type
unbiased sample
Measurement
, #30, pp. 607-610) have used the
following formula to determine sampling size
provided the population is definite.
