TOPIC 5: SAMPLE AND

SAMPLING TECHNIQUES
 Learning objectives
• Sampling concepts
• Learn the reasons for sampling
• Develop an understanding about different
sampling methods
• Distinguish between probability & non
probability sampling
• Discuss the relative advantages &
disadvantages of each sampling methods
 5.1 Sampling Concepts
• Population
• Sample
• Sampling
• Sampling frame
• Sample size
• Sampling unit
 What is a Population?
• The total number of subjects to be studied in a
certain area
• An entire group of individuals, events or objects
having common observable characteristics
• The group of subjects from which you intend to obtain
information to find answers to your research
questions, e.g.
– All patients suffering from HIV/AIDS in the
community
– All students in a school or college
– All indigenous trees in a particular forest
• Study population is denoted by N
 What is a Sample?
• A sample is “a smaller (but hopefully
representative) collection of units from a
population used to determine truths about
that population”
• A small group of the study population
from which you collect the information
• It is a representative group of the study
population
• Members in the sample are referred to as a
subjects or respondents
 What is a Sampling?
• The process of selecting few individuals or
subjects to take part in the study to represent
the large group or study population
• Selection of few members or a fraction of the
study population to become the basis for
estimating or predicting the prevalence of
unknown information regarding the bigger
group
• Sampling is done because collecting data from
every subject of the study population is
sometimes impossible and expensive
Sampling frame and sampling unit
• Sampling frame: is the list from which the
potential respondents are drawn
– e.g., Number of households in a village
– Total number of employees in an
organizations
– Registrar’s office
– Total number students in a college, school etc.
• Sampling unit: a unit of reference, reporting
or analysis e.g. individual, household, hamlet,
school, plot, group, ward or district etc.
 Sample size
• Is the total number of selected subjects to
take part in the study.
• The sample size is denoted by n
• Sample size is the most potent method of
achieving estimates that are sufficiently
precise and reliable for scientific inquiry
and policy decisions.
 Sample size determination
The decision on sample size ought to be
done at an early stage of the survey
research process because:
i. too large n is too costly and leads to
waste of resources without necessarily
increasing the reliability of findings,
ii. too small a sample diminishes the
reliability of the findings
 Strategies for determining
sample size
• In spite of the fact that many factors affect
the sample size required for a particular
study, “what sample size do I need for
my study” is the most frequently asked
question concerning sampling.
• Accordingly, every researcher has to
determine the appropriate sample size
for his/her study.
 Approaches used to determine
sample size
• There are four approaches
i. using a census for small populations,
ii. Using a sample size of a similar study,
iii. Using published tables and
iv. Using formulae.
 Using formulae
• Like in other sample size determination
methods, one has to specify the precision,
variability and level of confidence in using
formulae to determine sample size.
• Although there are many formulae that can be
used in determining sample size the one
provided by Yamane1(1967:886) is
simplified and practical enough for ease
of use by even the statistically
unsophisticated researcher.
 Using formulae….
• The formulae assume simple rondom
sampling technique, a 95% confidence
level. The formulae is as follows;
𝑵
n=
𝟏:𝑵𝒆𝟐
Where:
n = Sample size estimate
N = Popn. size or sampling frame
e = Error of prediction e.g 0.05 or 0.1
 Sample size …
• For example: Given the sampling frame (study
popn.) of 2,500 households, error of estimate/
prediction being 0.05; sample size estimate will
be:
𝑵
n=
𝟏:𝑵𝒆𝟐
2,500
n=
1:2,500/0.052
2,500
n=
1:6.25
= 2,500/ [1+ 6.25]
= 2,500/7.25
= 344
 Exercise
Ms. Mariam Mvula wants to conduct a
survey study in two Villages. Village A and
B with 4,246 households in total, 65%
residing in Village B. Given 0.05 as an error
of the sample size estimation, determine the
representative sample size for Ms. Mariam
Mvula to come up with a representative
sample for meaningful data analysis and
generalizations.
Village A
• Sampling frame in village A; 35% of
4,246= 1,486 households
• Sample size using Yamane (1967)
Formula; n=N/[1+N(e)2]
=1,486/1+1,486(0.05)2

=1,486/4.715 =315
• Sample size for village A is 315
households
Village B
• Sampling frame in village B; 65% of
4,246=2,760
n=N/1+N (e) 2
=2,760/1+2,760 (0.05)2
=2,760/7.9 =349
• Sample size for Village B is 349
households
• Therefore, the representative sample size
for village A and B is 315+349= 664
Sample size (unknown pop)
• Sometimes, a sample size can be
estimated and drawn from unknown
population (Cochran, 1977; Kothari, 2004).
The formulae for unknown population is as
follows;
𝑍 2 ∗𝑝 ∗𝑞 𝑍2 𝑝 𝑞
n= =
𝑒2 𝑒2
can be written
𝑍 2 ∗𝑝 ∗(1;𝑝) 𝑍 2 𝑝 (1;𝑝)
n= =
𝑒2 𝑒2
Sample size (unknown pop)...
• Where;
n=Sample size
Z=level of confidence (90%) =1.64
Z = Confidence level at 95% (Standard value
of 1.96)
P=maximum population proportional assumed to
be 50%=0.5 or
P= Estimate of Standard deviation in the
population which is 0.5
e= Standard Error (0.01, 0.05 0r 0.1)
Sample size (unknown pop)...
• Example1: Find the sample size when Z =
95% and e=5%
𝑍 2 ∗𝑝 ∗𝑞 𝑍2 𝑝 𝑞
n= 𝑒2
= 𝑒2
𝑍 2 ∗𝑝 ∗(1;𝑝) 𝑍 2 𝑝 (1;𝑝)
n= 𝑒2
= 𝑒2
1.962 ∗0.5 ∗0.5 3.92 ∗0.25 0.98
= n= 0.052
= 0.0025
= 0.0025

n = 392
Sample size (unknown pop)...
• Example2: Find the sample size when Z =
90% and e=10%
𝑍 2 ∗𝑝 ∗𝑞 𝑍2 𝑝 𝑞
n= =
𝑒2 𝑒2
𝑍 2 ∗𝑝 ∗(1;𝑝) 𝑍 2 𝑝 (1;𝑝)
n= =
𝑒2 𝑒2
1.642 ∗0.5 ∗0.5 2.6896∗0.25 0.6724
= n= = =
0.12 0.01 0.01

n = 67.24 ≈ 67; 𝑛 = 67
 5.2 Rationale of sampling
Why sampling? What are the
objectives?
i. It is impossible and expensive to collect
data from every body/subject in the
population, but by sampling
– That is, it is often impossible,
impractical and extremely expensive to
collect data from all potential units of
analysis covered by the research
problem.
5.2 Rationale of sampling …
ii. For practical and economy of scientific studies.
– The major reason for sampling is economy,
surveying every individual in a popn using
enumeration is ordinarily much too expensive
in terms of resource time, fiscal resources
and personnel.
– Therefore, sampling is done to save resources
which are always scarce.
iii. Provide accurate estimates of unknown value of
parameters from sample statistics that can be
easily calculated.
5.3 Types of sampling methods
• Sampling methods/procedures are
grouped into two categories.
a. Probability sampling methods/ procedure
b. Non – Probability sampling methods
 5.3.1 Probability sampling methods
• Every subject in the study population has
an equal and independent chance or
probability of being selected or included
in the sample (e.g., persons, households)
and the mathematical probability that any
one of them will be selected can be
calculated.
• Selection of subjects is done randomly, i.e.
by probability.
Types of Probability sampling
• Simple random sampling
• Systematic sampling
• Stratified sampling
• Cluster sampling
• Multistage sampling
Simple Random sampling (SRS)
• Simplest of all the probability sampling methods where
lottery or table of random numbers are used in
selecting respondents
• Procedures
1st Define the population and obtain sampling frame e.g.
list of villages, schools, districts, individuals in a village,
etc.
2nd Give each case a unique number starting with zero
(0) or one (1) i.e. from 0/1 to Nth.
3rd Decide and describe the sample size (n)
4th Select the sample (n) using fishbowl, table of
random numbers or a computer program
Simple Random sampling (SRS) …
Advantages of SRS
• Subjects are selected objectively
• No opportunity for human bias
• Selection doesn’t depend on subjects availability
i.e. one does not have to walk in the interviewees’
proximity to be included in the sample
• Selection is done without subjects knowledge
Disadvantages of SRS:
• The method is tedious i.e, interesting
• It requires good sampling frame.
Systematic sampling (SS)
• Subjects are selected systematically at a fixed interval
directly from the sampling frame, after every nth term.
e.g. 5th person, 10th person etc.
• The starting point is selected randomly
• Procedure for SS
1st Create a sampling frame
2nd Obtain sampling interval, which is given by:
Sampling Interval (SI) = Total Population
Sample size
Systematic sampling (SS) …
• e.g.If the sampling frame is 2,000 persons,
and sample size is 200.

SI = 2000
200
= 10
3rd Select a starting point randomly between 1-10
and make sure that the starting point is not more
than the interval.
Systematic sampling (SS) …
Systematic sampling (SS) …
4th Choose the value/subject after every
nth number (i.e. 10). e.g. say starting point
is 2. next point /subject will be 12,
followed by 22, 32, etc Until you get 200
subjects.
NOTE: Depending on the sampling unit, take
care of sampling men or women only, husbands
or wives only, elders or youth only, etc.
Systematic sampling (SS) …
Advantage of SS
• Subjects are selected objectively when the
technique is accurately applied
Disadvantages of SS
• It is tedious to use it
• There is risk of bias especially when
sampling method follow a particular
system
Stratified sampling
• Used when the study population is heterogeneous, but can
be grouped into certain homogeneous categories or strata
• It is an improvement of systematic sampling
involving grouping of subjects into homogeneous
categories or strata ensure each category is
represented in the sample
• Example
– by location (rural or urban), campus A and
campus B
– by sex (male or female)
– by educ. level (primary, secondary, tertiary)
– by income (high or low; rich or poor etc)
Stratified sampling …
• Stratification is done to ensure representation of subjects by a
variety of categories.
• After stratifying the population, simple random sampling or
systematic sampling is used to select the sample
Stratified sampling …
Advantage
• The technique ensures that resulting sample is
distributed in the same way as the population
by stratification
Disadvantages
• It is applied only when you can identify and
allocate units of strata.
• Using more than one criterion for stratification
makes the sampling exercise tedious.
• It is feasible only when relevant information is
available
Cluster sampling
• It is used where sampling frame does not exist
• It is used where popn. is very much scattered/
wide spread.
• Use the listed blocks e.g. villages, divisions,
schools, instead of a list of subjects or
households.
• The blocked parts (villages, divisions are what
we call clusters.
• Clustering or grouping of the surveyed/
studied popn may be done on the basis of
physical or geographical location.
Cluster sampling …
• After blocking, go to each block e.g. village and
interview every subject in that village/block of
ten cell
Advantage
• Simple to apply as it is easy to block the
population by physical or geographical
location.
Disadvantage
• The technique involves tedious interviewing
exercise, since every subject in the cluster/ block
has to be interviewed
Multistage sampling
• Involves use of several sampling methods at
different stages at different levels
• For example, you want to study an issue of your
choice in an entire country like Tanzania
1st stage: decide and choose 5 regions out of 30 randomly for
a study using simple random or systematic sampling method.
Decide the sample size, say 3000 households
2nd stage: Decide and pick randomly say one
district in every selected region
3rd stage: You may decide to take one ward in
every district
Multistage sampling …
• If there are 5 wards then take 120
subjects from every ward (600/5 =120)
4th stage: from the wards get the villages
• If there are 4 villages in every ward,
take 30 subjects from every village
(120/4 = 30)
• If there are 6 hamlets take 5 subjects from
each hamlet (30/6 = 5); where now you
can get the list of them (sampling frame)
5.3.2 Non- Probability
sampling methods
• With non-probability sampling, in contrast,
population elements are selected on the basis of
their availability (e.g., because they volunteered)
or because of the researcher's personal judgment
that they are representative.
• The consequence is that an unknown portion of
the population is excluded (e.g., those who did
not volunteer).
• Used when the number of elements in the study
population is either not known or can not be
individually identified
5.3.2 Non- Probability
sampling methods …
• Five main non-probability sampling methods
– Quota sampling
– Convenience or accidental sampling
– Judgemental or purposive sampling
– Chunk sampling
– Snowball sampling
Quota sampling
• It aims at producing a sample that reflects a
population in terms of relative proportions of
people in different categories
• For example gender, ethnicity, age groups,
social economic groups, place of residence etc.
• One makes a representative sample either
selected randomly or non-randomly. As he/she
makes/ develops quotas using available
information
• It is claimed by some practitioners to be as good
as probability sampling
• Example: Say IRDP with 500 students
residing in different hostels
A-Centre 50%
B- Centre 40%
C- Centre 10%
• If the sample size is 100 students out of
500: The quota sample representation
will be as follows:
A-Centre 50
B- Centre 40
C- Centre 10
Convenience or
accidental sampling
• In this method subjects are selected for
convenience reasons. That is, subjects are
included in a sample by virtual of their
accessibility
• E.g. all those who happen to be available are
included in the study
• This technique is commonly used when testing
tools (pilot study) like questionnaires
• Is also used where there are possibilities of
having less subjects, for example all men in clinic
during clinic day
Convenience or
accidental sampling …
Advantages
• Easy and cheap to apply
• Quick
• Very convenient one
Limitations
• The sample can be unrepresentative
• One can not know the extent of biasness of
sample
• It is impossible to generalize the findings
because it is not known what population the
sample represents
Judgemental /
Purposive Sampling
• Involves making planned selection of
respondents / areas included in the study
• The subjects / areas are believed to be typical
representative and suitable in the study
For example; Persistence of cholera outbreak in
Mwanza
• Which district is a typical representative?
• Which division or village or street is the typical
representative?
• When you identify such areas or subjects make
your sample
Chunk Sampling
• Chunk sampling usually is employed by
journalists/media people.
• They go in the streets or an area with an
interested issue and ask people about the
interested problem/issue.
• Such as waste management, particular football
match, expected winner for presidency post, or
any other issue /event.
• Then they record the information prior to
processing and disseminating.
Snowball Sampling
• Is a form of convenience sampling but each of the
selected subjects is asked to give address or
contact other people with similar condition or
facing similar problem
• That is, the researcher makes initial contacts with a
small group of people relevant to a study and uses
them to establish contacts with others
• Example: HIV&AIDS Victims, homosexuals,
victims of rape, robberies, drug abusers and so
forth
• To sample them you just need to know one and
ask him/her to show others
Snowball Sampling …
• Note: With such studies snowball
sampling is the only feasible sampling
approach/technique, since it is difficult
to create a sampling frame of say
marijuana and other drug users.
• Major weaknesses for this method are that,
the sample is not representative, and it is
difficult to generalize the results.
• Thus, is more suitable for qualitative
studies only.
Conclusion
• These are just some of the most useful
sampling methods.
• Each has strengths and
weaknesses, it is important that
you identify them.
• Researchers are required to state which
one(s) will be used and why?
 BASIC READING MATERIAL
• Kumar, R. (2005). Research Methodology.
A Step-by-Step Guide For Beginners
(Chapter 12)
Any question?