2. Sampling Design
Population
• The whole group of people, items, objects, events etc. having some
common characteristic and which is being researched on is called a
population.
• A single person, item etc. is called an element of the population.
Sample
• A sample is a part of a population.
• The process of selection of one or more elements from the
population is called sampling.
• The elements in a sample are selected in such a way that the
sample represents the population in every possible characteristic
and feature.
5. Why Sampling?
Benefits of sampling
• Cost effective
• Saves time
• Can be more accurate
• In a situation where the research results in destruction,
deformation, mutilation or contamination of the elements
sampled sampling is essential
Total study
• in this case sampling is not required
• this is done in a case where the population is too small
• this is done when the research so intends that all the
elements in the population must be included in the study
6. Sampling design
• Sampling design is created keeping in view the purpose and focus
of the research.
• Sampling design consists of five sequential and interrelated steps.
• Each step has relevance to all aspects of the research.
• A sample selected using a biased or technically wrong method will
lead to irrelevant information, which in turn will lead to incorrect or
distorted conclusions of the research.
• Steps in a sampling design
• 1. Define the target population
• 2. Specify the sampling frame
• 3. Decide on a sampling procedure
• 4. Determine a method of determining the sample size
• 5. Determine the optimum sample size
7. Define the target population
• The target population is that entire group of items or individuals or
cases from where the information is to be collected.
• This may also be called the ‘subject’ of research.
• In an empirical study, the target population consists of physical
objects like people or items or events.
• In a case study it contains just one object or event.
• In fundamental research it can be infinite, where it is required to
know something that is true for every object or event of the given
type in the universe.
• A total study gives a complete and accurate description of the
population, but it is possible only if the population is not too large
and if all the elements in the population are available for study.
8. Specify the sampling frame
• A sampling frame is a list of all the
elements in the target population
• Telephone directory, mailing list, register
maintained at office are examples of
sampling frames
10. Probabilistic Methods
Simple Random Sampling
Every element in the population has an equal
probability of being selected in the sample
Pros Cons
• 1. Comparatively easy method • If the population has
• 2. Softwares for generating subgroups that may be of
random numbers are available research interest, they may not
and simple to use. be adequately and
proportionately represented by
the sample.
• 2. If the population size is too
large, allocating numbers to its
elements and
11. Stratified random sampling
The population is divided into strata. A stratum is a subset
of the population that share at least one common
characteristic. Every element in a stratum has an equal
probability of being selected in the sample.
• This is a more representative • Calculation more complicated
form of the population. that simple random sampling.
• Gives good results when • Population and sample size for
studies involve subgroups as each strata should be known.
gender, age, income group,
education level, socio-
economic category, religion,
geographical location, etc.
12. Cluster sampling
divides the population into groups or clusters. A
number of clusters are selected randomly to represent the
population, and then all units within selected clusters are
included in the sample. No units from non-selected clusters
are included in the sample. This differs from stratified
sampling, where some units are selected from each group.
• 1. Clustering helps in • Usually the general
reducing data collection assumption is that the
time and cost. clusters are alike – if this
• 2. In case it is impossible is violated the sample will
and impractical to get the be biased.
list of the entire • 2. It is better to increase
population, this method is the number of clusters
useful. and thereby reduce the
cluster size.
13. Systematic sampling
Every ith numbered element is selected. That is there is
uniform gap between selected elements .
• Very convenient to • If there is an existing
use. recurring pattern in
• Only the 1st element the population, this
needs to be selected may produce a bias in
randomly. the sample.
14. Multi-stage sampling
similar to cluster sampling, but involves selecting a
sample within each chosen cluster.
• 1. This does not require a • Same as in cluster
complete list of members sampling
in the target population,
which greatly reduces
sample preparation cost.
• 2. The list of members is
required only for those
clusters used in the final
stage.
15. Non Probabilistic Methods
Convenient sampling :Elements in the population who / that
are readily available is included in the sample.
• Cost-effective, time- • Since the sample is
saving practical so chosen, it is
method unlikely to be
representative of the
population.
16. Judgement sampling
(Purposive sampling)
The researcher selects the sample based on judgment.
• 1. Useful in • 1. The researcher
exploratory research. should be fully aware
• 2. Makes certain that of the purpose and
the widest variety of objective of the
elements is chosen in research.
the sample. • 2. The bias of the
researcher may affect
the
representativeness of
the sample.
17. Quota sampling
The population is divided into groups. Then convenience or
judgment sampling is used to select the required number of
subjects from each group.
• Useful when prior • Records relating to
knowledge of groups proportions of groups
exist. must be complete,
correct and up-to-da
18. Snowball sampling
This method is used when the desired sample
characteristic is difficult to find or cost prohibitive
Snowball sampling relies on references from initial
subjects to generate additional subjects .
• 1. This unique technique • Data collected from
can reduce research ‘snowballed’ sample may
costs. not be a measure of what
• 2. This is a good method is to be actually collected.
for such populations that
are not well defined or
properly listed
19. Determine a method of
determining the sample size
Sample Size is the number of elements in the sample. The concern is to decide
on the size of a sample, so that the sample
• will not lose its usability,
• will give us data reliable enough about the population,
• will be able to represent the population in its truest form.
Five interrelated factors that play a role in decision about the sample size:
• heterogeneity of the population
• required precision level
• sampling procedure
• resources available
• time constraints
• number of major sub-divisions in the research, each needing separate
sampling, and sample sizes to be determined for each of these samplings.
20. Determine the optimum sample
size
The size of the sample may be determined in two ways – subjective
and objective.
Subjective
• the researcher decides subjectively on the size according to his
understanding of the research,
• past experience with similar type of research
• time and cost constraints
• availability of elements to be included in the sample
• this method has no consideration for statistical theory
Objective
• statistically decided sample size,
• predetermined limits of sampling error and confidence level are
required,
• the researcher’s subjectivity has no role to play in this method.
21. The following table gives optimum sample
sizes for ± 5% sampling error with 95%
confidence level.
Population Size Sample Size Percentage of the population
size
10 10 100
20 19 95
50 44 88
100 80 80
250 152 61
500 217 43
1,000 278 28
2,500 333 13
5,000 350 7
10,000 370 4
22. Exercise
A medical inspector desires to estimate the
overall average monthly occupancy rates
of the cancer wards in 80 different
hospitals that are evenly located in the
different suburbs of Delhi NCR
23. Exercise
In an article in the wall street journal titled “Kellogg to study work of
salaried staff, setting stage for possible job cutbacks”, it was
started that the kellogg’s earnings remained under heavy
competitive pressure and its cereal market continued to slip. It was
also stated that kellogg was seeking to regain its lost momentum
through the first three strategies listed below, to which last two are
added.
1. Increasing production efficiencies
2. Developing new products
3. Increasing product promotion through advertising effectiveness
4. Tapping creative ideas from organizational members at different
levels
5. Assessing perception of organizational health and vitality
Discuss sampling design for each of the five strategies. Give the
reasons for ur choice.
24. Exercise
Care for elderly relatives is a concern for
many working parents. If you were to do a
scientific study of this , what kind of
sampling design you would use? Discuss
your response with reasons for the choice
of population and sample