Sampling:

Process of Selection of a sample from a population to generate precise and valid estimates of
population parameter(s) like mean or proportion is called Sampling while process of
collecting relevant information/data in respect of each and every member/unit of the
population is called Census.
Advantages of sampling: Sampling has the following advantages:
(a) Less expensive: If data are collected for the entire population, the cost may be quite high.
A sample is a small proportion of a population. Obviously the cost will be lower if the data
are collected for a part of population.
(b) Saves Time: Use of sampling takes less time than census technique.
(c) Scope of sampling: To study the whole population would be impractical in a few cases.
Some populations are so large that before the census is completed, the population would have
changed. For e.g. to test blood, the entire blood can’t be taken from the body.
(d) Accuracy of data is high: A sample permits a high degree of accuracy since a careful
execution of field work is possible.
(e) Suitable in case of limited resources: The resources available within an organization are
usually limited so studying the entire universe is not a viable proposition.
(f) Other advantages: If we use sampling; it is possible to determine the extent of error.
Further, the Non Sampling errors are likely to be less; even the Census Results are verified by
sampling.
Disadvantages of sampling
(a) Chances of bias: The serious limitation of the sampling method is that it may have biased
selection and thereby leads to wrong conclusions.
(b) Difficulties in selecting a truly representative sample: A sample produces reliable and
accurate results only when it is representative of the whole group.
(c) Lack of adequate subject knowledge: Use of sampling method requires adequate subject
specific knowledge.
(d) Changeability of units: Some of the units of sample may not cooperate with the researcher
and some others may be inaccessible.
(e) Impossibility of sampling: Deriving a representative sample is difficult, when the universe
is too small or too heterogeneous.

1. Simple Random Sampling (SRS):

It is the simplest form of random/statistical sampling; it consists of selecting the
sample units ensuring equal probability of selection to every unit of the population. It
is of two types:
(a) Simple Random Sampling with Replacement (SRSWR): In SRSWR, a unit is selected
from the sampling frame (list of units in the population); the unit is replaced back and the
next unit is selected; the process is repeated till a sample of the desired size is selected. As a
result it is possible for a unit to be included in the sample more than once. In practice
SRSWR is not attractive; same units can be selected more than once which may not add any
value/additional information to audit. But in mathematical terms, it is simpler to relate the
sample to population by SRSWR.
(b) Simple Random Sampling without Replacement (SRSWOR): In SRSWOR, a unit is
selected for inclusion in the sample, it is removed from the sampling frame and the next unit
is selected. Thus, in this type of sampling a unit cannot be selected again. If a sample is
selected with the help of random number table; a random number selected more than once is
ignored in SRSWOR. SRSWOR has two advantages:
i) Elements are not repeated so resources are not wasted.
ii) Variance estimation is smaller (efficiency is higher) than SRSWR with same sample size.
We select a slip from 10 slips and replace the slip before selecting the next slip, then it is
sampling with replacement; if we do not replace the slip before selecting the second slip, it is
sampling without replacement.
Advantages
(i) One of the greatest advantages of simple random sampling method is that it needs
only a minimum knowledge of the study group of population in advance.
(ii) It is free from errors in classification/stratification.
(iii) It is totally free from bias and prejudice.
(iv) The method is simple to use.
(v) It is very easy to assess the sampling error in this method.
Disadvantages
(i) As compared to stratified sampling, it gives less efficient results for the same
sample size.
(ii) The study of sample becomes time consuming if the units or items are widely
dispersed.
(iii) It cannot be employed if the units of the population are heterogeneous in nature.
(iv) This method does not use available knowledge about the population.
(v) It needs complete list of units in the population (sampling frame)
(vi) It does not always produce a representative sample.

2. Systematic sampling:
In Systematic Sampling the sample is chosen by selecting a random starting point and
then picking every kth (k is sampling interval) unit in succession from the sampling
frame. The sampling interval ‘k’ is the ratio of population size (N) to sample size (n),
rounded to the nearest integer. Systematic sampling is of 2 types (a) Linear and (b)
Circular Systematic Sampling.
Example of Systematic Sampling: (i) Say, Target Population N= 54000 vouchers
and sample size n = 6000
 (ii) Sample Fraction (k) = Target Population / Sample size = 54000/6000 = 9
(iii) Number all vouchers of the population from 1 to 54000
(iv) Select a number between 1 to 9 (here k = 9) randomly
(v) Say, number 5 is selected then 5th voucher is selected
(vi) Next 5+9 = 14th, 14 + 9 = 23rd , 32nd, 41st vouchers are selected and so on
…

Advantages
1. Systematic sampling is less costly and easier to implement than SRS. It is
because in this method random selection is done only for the first unit.
2. It ensures representativeness across the list (population) and is easy to
implement.
3. It can help eliminate cluster selection i.e. selection of nearby units of the
population

Disadvantages
1. It works well only if the complete and up-to-date frame is available and if the
units are randomly arranged in the frame; for this reason the units are arranged in
some order say alphabetical or in increasing/decreasing order of value before
selecting a sample.
2. If the population has a periodicity of the trait, this sampling technique may not
give a representativeness sample. So it should not be used if the population
already has some pattern.

3. Stratified sampling:
It is a two-step process in which the entire population is partitioned into
subpopulations, or strata. The strata should be mutually exclusive and collectively
exhaustive; it means, every population unit should be assigned to one and only
one stratum (singular for strata) and no population unit should be omitted. From
each stratum, units are selected randomly, usually by SRS. The population units in
each stratum should be as homogeneous as possible. A major objective of
stratified sampling is to increase reliability without increasing cost. Stratification
may be done on the basis of income, age, rural-urban, states, Revenue-Capital,
Treasuries, major heads, etc.

Examples of stratified sampling in Audit:

(i) To select BPL households for a social audit; divide the population of BPL
into three categories (strata) say top 25%, Middle 50% and Bottom 25%
and select separate samples from 3 categories/strata; the sample from the
three strata may or may not be equal.
(ii) Dividing contracts into value ranges and then selecting separate samples
from each value range; say 100% contracts from highest value range, 50%
contracts from next highest and so on.
(iii) Dividing the entire state into 3-4 geographical regions and then selecting
required sample of districts from each of these regions.
 Advantages
1. Stratified sampling is more precise if variable of interest is associated
with strata.
2. In this kind of sampling all the subgroups are represented, allowing
separate conclusion about each of them; say separate conclusion for each
state/District/treasury.
3. The stratified random sample also improves the representation of the
various groups within the population, as well as ensures that strata are not
over/under-represented. It helps the researcher/auditor to compare strata,
as well as make more valid inferences about the population from the
sample.

Disadvantages
1. In this method sampling error is difficult to measure
2. There is a loss of precision if a small number of units are sampled in
individual heterogeneous strata.
3. This method is useful when there is a sufficient knowledge about the
spread/variability of the population so that the strata are homogeneous.

4. Cluster sampling:
In this type of sampling, the target population is first divided into mutually
exclusive and collectively exhaustive sub-populations or clusters. It means, each
unit should belong to one and only one subgroups and that none of the units of the
population is left out. Then a random sample of clusters is selected, based on a
probability sampling technique such as SRS or Systematic Sampling. For each
selected cluster, either all the units or a sample of units is drawn and audited.
Heterogeneity within the cluster should be the same as that in population, ideally
each cluster should be a small-scale representation of the population.

Advantages

1. It is simple as complete list of units (sampling frame) is required only

for the clusters selected in the sample which reduces efforts of finding
out the complete sampling frame for all the clusters. 2. For this kind of
sampling less travel/resources are required.

Disadvantages
1. It is imprecise if clusters are homogeneous (Large sample as compared to
SRS is required for the same precision)
2. In this kind of sampling; Sampling Error is difficult to measure.