Page 65 - SREENARAYANAGURU OPEN UNIVERSITY

The other method is to take an equal number from each stratum
irrespective of strata’s shares (like 10 boys and 10 girls). This
method is called disproportionate stratified sampling.
C. Systematic sampling or fixed interval method

This method considers every k item from the population as a
th
sample after a random start. For example, if an employer wants
to take a survey of 30 employees from a group of 350, then he ◆ Every kth item after
can divide 350/30 and the quotient he gets would be 11.6 (ignore a random start
the fraction and consider the integer or whole number). So he can
take either 11 or 12 as the k item. Then the first sample must be
th
drawn randomly from the sample frame. Say if he gets 14 then the
14 employee in the roll will be taken as the first sample. Then
th
25 (14+11), 36 (25+11), 47 (36+11)... and so on till the required
number 30 is reached. The interval for selecting samples is fixed
(that is, the quotient calculated) and hence this method is also
known as fixed interval method.
This method is simple, less time consuming and cheaper. It is
also easy to understand and execute. As the sample is drawn from
a chronologically arranged frame it gives a highly representative
sample.

D. Cluster sampling
When we have a scattered population, cluster sampling method
can be used. This involves dividing the population into different ◆ Homogeneous
clusters which are homogeneous and then choosing samples clusters with
systematically. Elements within each cluster are heterogeneous. heterogeneous
Suppose we need to select banks all over Kerala. We can divide elements
the geographical region of Kerala into 3 as northern, central and
southern clusters and then choose the required number of samples
from each cluster. This provides a chance for all units to be selected
in particular regions.
This differs from stratified sampling as the clusters are the
sampling elements and strata the population element. Clusters are
made on geographical areas but strata are based on demographic
variables. Strata yield higher representative samples than clusters
and sampling error is comparatively less in stratified sampling. ◆ Cluster-
Cluster sampling is well suited when there is a large population geographical area
or larger geographical area to be covered. This method involves ◆ Strata- demographic
variables
lesser cost. Easy replacement of units is possible from the same
random section. The varying size of clusters can be a matter of
concern.
Area sampling
This is a form of cluster sampling which is used in large field
surveys. Geographical areas like city, blocks, village, taluks,
districts etc. are selected as samples. Hence it is called area

SGOU - SLM - MCom Research Methodology 59