Sampling:

• The process of selecting a sample from a population is called

sampling.

• In sampling, a representative sample or portion of elements of a

population or process is selected and then analyzed. Based on
sample results, called sample statistics, statistical inferences are
made about the population characteristic.

• For instance, a political analyst selects specific or random set of

people for interviews to estimate the proportion of the votes that
each candidate may get from the population of voters; a doctor
examines a few drops of blood to draw conclusions about the
nature of disease or blood constitution of the whole body.
Reasons of Sample Survey:
A census is a count of all the elements in a population.
Few examples of census are: population of eligible voters;
census of consumer preference to a particular product, buying
habits of adult Indians. Some of the reasons to prefer sample
survey instead of census are given below:

1. Movement of Population Element The population of

fish, birds, snakes, mosquitoes, etc. are large and are constantly
moving, being born and dying. So instead of attempting to
count all elements of such populations, it is desirable to make
estimates using techinques such as counting birds at a place
picked at random, setting nets at predetermined places, etc.
2. Cost and/or Time Required to Contact the Whole
Population
A census involves a complete count of every individual
member of the population of interest, such as persons in a
state, households in a town, shops in a city, students in a
college, and so on. Apart from the cost and the large
amount of resources (such as enumerators, clerical
assistance, etc.) that are required, the main problem is the
time required to process the data. Hence the results are
known after a big gap of time.
3. Destructive Nature of Certain Tests
The census becomes extremely difficult, if not impossible,
when the population of interest is either infinite in terms
of size (number); constantly changing; in a state of
movement; or observation results required destruction. For
example, sometimes it is required to test the strength of
some manufactured item by applying a stress until the unit
breaks. The amount of stress that results in breakage is the
value of the observation that is recorded. If this
procedure is applied to an entire population, there
would be nothing left. This type of testing is called
destructive testing and requires that a sample be used in
such cases.
Population Parameters and Sample Statistics:
Parameters An exact, but generally unknown measure (or
value) which describes the entire population or process
characteristics is called a parameter. For example,
quantities such as mean 𝜇, variance 𝜎 2 , standard deviation
𝜎, median, mode, and proportion 𝑝 computed from a data
set (also called population) are called parameters. A
parameter is usually denoted with letters of the lower case
Greek alphabet, such as mean 𝜇 and standard deviation
𝜎.
Sample Statistics A measure (or value) found from
analyzing sample data in called a sample statistic or simply
a statistic. Inferential statistical methods attempt to
estimate population parameters using sample statistics.
Sample statistics are usually denoted by Roman letters
such as mean 𝑥, standard deviation 𝑠, variance 𝑠 2 , and
proportion 𝑝.
Sampling Methods:
Table 7.1 Types of Sampling Methods
1. Probability Sampling Methods:
Simple Random (Unrestricted) Sampling
• In this method, every member (or element) of the
population has an equal and independent chance of
being selected again and again when a
sample is drawn from the population.

• To draw a random sample, we need a complete list

of all elements in the population of interest so that
each element can be identified by a distinct number.
Such a list is called frame for experiment.
• For instance, in drawing the random sample of 50 students
from a population of 3500 students in a college we make a
list of all 3500 students and assign each student an
identification number. This gives us a list of 3500
numbers, called frame for experiment. Then we generate
by computer or by other means a set of 50 random numbers
in the range of values from 1 and 3500. The procedure gives
every set of 50 students in the population an equal chance of
being included in the sample. Selecting a random sample
is analogous to using a gambling device to generate
numbers from this list.

• This method is suitable for sampling, as many statistical

tests assume independence of sample elements. One
disadvantage with this method is that all elements of the
population have to be available for selection, which many a
times is not possible.
Stratified Sampling
• This method is useful when the population consists
of a number of heterogeneous subpopulations and
the elements within a given subpopulation are
relatively homogeneous compared to the population
as a whole.

• Thus, population is divided into mutually

exclusive groups called strata that are relevant,
appropriate and meaningful in the context of the
study. A simple random sample, called a sub-sample,
is then drawn from each strata or group, in-
proportion or a non-proportion to its size.
• As the name implies, a proportional sampling
procedure requires that the number of elements in
each stratum be in the same proportion as in the
population. In non-proportional procedure, the
number of elements in each stratum are
disproportionate to the respective numbers in the
population.

• The basis for forming the strata such as location,

age, industry type, gross sales, or number of
employees, is at the discretion of the investigator.
Individual stratum samples are combined into
one to obtain an overall sample for analysis.
• This sampling procedure is more efficient than the
simple random sampling procedure because, for the
same sample size, we get more representativeness
from each important segment of the population and
obtain more valuable and differentiated information
with respect to each strata.

• Disproportionate sampling decisions are made either

when strata are either too small, too large, or when
there is more variability suspected within a
particular stratum. For example, the educational
levels in a particular strata might be expected to
influence perceptions, so more people will be
sampled at this level. Disproportionate sampling is
done when it is easier, and less expensive to collect
data from one or more strata than from others.
• For this method of sampling to be more effective
in terms of reliability, efficiency, and precision,
any stratification should be done which ensures
(i) maximum uniformity among members of
each strata,
(ii) largest degree of variability among various
strata.
Cluster Sampling
• This method, sometimes known as area sampling
method, has been prepared to meet the problem of
costs or inadequate sampling frames (a complete
listing of all elements in the population so that each
member can be identified by a distinct number).

• The entire population to be analyzed is divided into

smaller groups or chunks of elements and a sample
of the desired number of areas selected by a
simple random sampling method. Such groups
are termed as clusters. The elements of a cluster are
called elementary units. These clusters do not have
much heterogeneity among the elements. A
household where individuals live together is an
example of a cluster.
• If several groups with intragroup heterogeneity
and intergroup homogeneity are found, then a
random sampling of the clusters or groups can be
done with information gathered from each of the
elements in the randomly chosen clusters. Cluster
samples offer more heterogeneity within groups
and more homogeneity among groups—the
reverse of what we find in stratified random
sampling, where there is homogeneity within
each group and heterogeneity across groups.
• For instance, committees formed from various
departments in an organization to offer inputs to
make decisions on product development, budget
allocations, marketing strategies, etc. are
examples of different clusters. Each of these
clusters or groups contains a heterogeneous
collection of members with different interests,
orientations, values, philosophy, and vested
interests. Based on individual and combined
perceptions, it is possible to make final decision
on strategic moves for the organization.
Multistage Sampling
• This method of sampling is useful when the
population is very widely spread and random
sampling is not possible. The researcher might
stratify the population in different regions of the
country, then stratify by urban and rural and then
choose a random sample of communities
within these strata. These communities are then
divided into city areas as clusters and randomly
consider some of these for study. Each element in
the selected cluster may be contacted for desired
information.
• For example, for the purpose of a national pre-
election opinion poll, the first stage would be to
choose as a sample a specific state (region). The
size of the sample, that is the number of
interviews, from each region would be
determined by the relative populations in each
region. In the second stage, a limited number of
towns/cities in each of the regions would be
selected, and then in the third stage, within the
selected towns/cities, a sample of respondents
could be drawn from the electoral roll of the
town/city selected at the second stage.
• The essence of this type of sampling is that a
subsample is taken from successive groups or
strata. The selection of the sampling units at each
stage may be achieved with or without
stratification. For example, at the second stage
when the sample of towns/cities is being drawn,
it is customary to classify all the urban areas in
the region in such a way that the elements
(towns/cities) of the population in those areas are
given equal chances of inclusion.
Systematic Sampling
• This procedure is useful when elements of the population
are already physically arranged in some order, such as an
alphabetized list of people with driving licenses, list of
bank customers by account numbers. In these cases one
element is chosen at random from first k element and
then every kth element (member) is included in the
sample. The value k is called the sampling interval.

• For example, suppose a sample size of 50 is desired from

a population consisting of 1000 accounts receivable.
The sampling interval is k = N/n = 1000/50 = 20.
Thus a sample of 50 accounts is identified by moving
systematically through the population and identifying
every 20th account after the first randomly selected
account number.
2. Non-Random Sampling Methods:
Convenience Sampling
• In this procedure, units to be included in the sample
are selected at the convenience of the investigator
rather than by any pre-specified or known
probabilities of being selected.

• For example, a student for his project on ‘food

habits among adults’ may use his own friends in the
college to constitute a sample simply because they
are readily available and will participate for little or
no cost. Other examples are, public opinion surveys
conducted by any TV channel near the railway
station; bus stop, or in a market.
• Convenience samples are easy for collecting data on
a particular issue. However, it is not possible to
evaluate its representativeness of the population and
hence precautions should be taken in interpreting
the results of convenient samples that are used to
make inferences about a population.

Purposive Sampling
• Instead of obtaining information from those who are
most conveniently available, it sometimes becomes
necessary to obtain information from specific
targets–respondents who will be able to provide the
desired information either because they are the only
ones who can give the desired information or
because they satisfy to some criteria set by
researcher.
Judgment Sampling
• Judgment sampling involves the selection of respondents
who are in the best position to provide the desired
information. The judgment sampling is used when
a limited number of respondents have the
information that is needed.
• This sampling method may reduce the generalizability of
the findings due to the fact that we are using a sample of
respondents who are conveniently available to us. It is the
only viable sampling method for obtaining the type of
information that is required from very specific
section of respondents who possess the knowledge and
can give the desired information.
• However, the validity of the sample results depend on the
proper judgment of the investigator in choosing the
sample. Great precaution is needed in drawing conclusions
based on judgment samples to make inferences about a
population.
Quota Sampling
• Quota Sampling is a form of proportionate
stratified sampling in which a predetermined
proportion of elements are sampled from
different groups in the population, but on
convenience basis.

• In other words, in quota sampling the selection of

respondents lies with the investigator, although in
making such selection he/she must ensure that
each respondent satisfies certain criteria which is
essential for the study.
• For example, the investigator may choose to
interview ten men and ten women in such a way that two
of them have annual income of more than two lakh rupees
five of them have annual income between one and
two lakh rupees and thirteen whose annual income is
below one lakh rupees. Furthermore, some of them
should be between 25 and 35 years of age, others between
36 and 45 years of age, and the balance over 45 years.
This means that the investigator’s choice of respondent is
partly dictated by these ‘controls’.

• Quota sampling has been criticized because it does not

satisfy the fundamental requirement of a sample, that is,
it should be random. Consequently, it is not possible
to achieve precision of results on any valid basis.