3. SUBMITTED BY
DR. MD HANIF TAHSIN
Email: dr.haniftahsin@gmail.com
Department of Public Health
Hamdard University Bangladesh
4. Introduction
• After selecting the research problem, developing the research question and or
hypothesis, and deciding on the research approach (quantitative / qualitative) the
researcher has to select, using appropriate methods, the actual persons, objects, and
events from which information is to be collected.
• Occasionally, the researcher may identify and study the whole population of
interest.
• However, most often the researcher studies a ‘sample’ of the population because it
is impractical, too large, difficult, or time and resource consuming to study the
entire population.
• From the statistical viewpoint, it may even be unnecessary to study the entire
population.
• A sample may provide more accurate findings than a study of the entire population
through collection of more accurate data.
• The researcher should aim to optimize the use of resources and make the study
cost-efficient.
5. Sampling approaches
There are two basic approaches to sampling:
A) Probability sampling or random sampling;
B) Non-probability sampling
• Probability sampling involves random selection, allowing you to
make strong statistical inferences about the whole group.
• Non-probability sampling involves non-random selection based on
convenience or other criteria, allowing you to easily collect data.
6. Population vs sample
• First, my need to understand the difference between a population and a
sample, and identify the target population of my research.
• The population is the entire group that you want to draw conclusions
about.
• The sample is the specific group of individuals that you will collect
data from.
7. Sampling frame
• The sampling frame is the actual list of individuals that the sample will be
drawn from. Ideally, it should include the entire target population (and
nobody who is not part of that population).
• Example: You are doing research on working conditions at Company X.
Your population is all 1000 employees of the company. Your sampling frame
is the company’s HR database which lists the names and contact details of
every employee.
Sample size
• The number of individuals you should include in your sample depends on
various factors, including the size and variability of the population and your
research design. There are different sample size calculators and formulas
depending on what you want to achieve with statistical analysis.
8. A) Probability sampling (or random sampling)
• It implies that all elements in the population have an equal chance of being
included in the sample.
• A probability sample (or random sample) is much more likely to be
representative of the population and to reflect its variations.
• If probability sampling is possible, then one of the common probability
sampling techniques may be used:
1. Simple random sampling,
2. Stratified random sampling,
3. Systematic random sampling and
4. Cluster sampling.
10. 1. Simple random sampling:
• A simple random sample is selected using the basic probability sampling technique.
• Each element to be included in the sample is selected randomly from the sampling frame.
• Each elements in the frame is listed separately so that it has an equal and independent chance of
being included in the sample.
• Simple random sampling is a one-stage selection process.
2. Systematic sampling (or interval sampling):
• Systematic sampling involves selecting elements at equal intervals, such as every third, every
sixth, or every twentieth element. This technique is based on the assumption that elements are not
added to the list in a systematic pattern that coincides with the sampling system.
• If a list of all elements is available, systematic sampling is easy and convenient.
• When careful attention is paid to obtaining an unbiased listing of the population elements, and the
first element to be included in the sample is randomly selected, systematic sampling is classified as
probability sampling. If either of these criteria is not met, non-probability sampling occurs.
example, if the population is 800 and the sample size is 160, the sampling interval is 5 (800 divided
by 160). A number between 1 and 800 is randomly selected as the starting number. Suppose that the
randomly selected number is 12, the next four subjects will be 17, 22, 27 and 32, the process is
continued until 160 individuals are selected.)
11. 3. Stratified Random Sampling:
• In stratified random sampling, the population is divided into subgroups or
‘strata’ according to one or more variables of importance to the study, so that
each element in the population belongs to only one stratum.
• Then, within each stratum, random sampling is performed, using either the
simple random sampling or systematic (interval) sampling technique.
• To use this sampling method, you divide the population into subgroups
(called strata) based on the relevant characteristic (e.g. gender, age range,
income bracket, job role).
Example
• The company has 800 female employees and 200 male employees. You want
to ensure that the sample reflects the gender balance of the company, so you
sort the population into two strata based on gender. Then you use random
sampling on each group, selecting 80 women and 20 men, which gives you a
representative sample of 100 people.
12. 4. Cluster random sampling:
• Cluster random sampling is used in large-scale studies, where the population is
geographically widespread, sampling procedures are difficult and time
consuming, or it is difficult or impossible to obtain a complete listing of the
population.
• Cluster sampling takes place in stages (‘multi-stage’).
• The researcher begins with the largest most inclusive sampling unit, and
progresses to the next most inclusive sampling unit until he/she reaches the final
stage, which is the selection of the elements or participants in the study.
• For example, a researcher who wishes to study cancer patients across a
country may use districts as the largest unit and randomly select a sample from a
list of the districts. Next, he/she identifies the hospitals which admit and treat
cancer patients in each of the districts making up the sample. He/she then selects
a sample of the hospitals, probably by stratified sampling. The final selection is
a sample of cancer patients from the selected sample of hospitals.
13. B) Non-probability sampling
• Non-probability sampling may or may not accurately represent the
population.
• It is usually more convenient and economical, and permits the study of
populations that cannot be sampled using probability sampling techniques.
• Non-probability sampling is often used, for example, when the researcher is
unable to locate the entire population or where access to the subjects or
elements is limited. In this situation the representativeness of the sample
also cannot be determined, because it is impossible for the researcher to
specify whether each element has an equal chance of being included in the
sample
14. The types of commonly used non-probability samples include:
1. Purposive sample or theoretical sample or judgmental sample
2. Convenience sample
3. Quota sample and
4. Snowball sample or network sample
15. The types of commonly used non-probability samples include
16. 1. Purposive / theoretical / judgmental sampling:
Purposive sampling is also known as 'judgmental sampling' or 'theoretical
sampling’
This technique is based on the judgement of the researcher regarding subjects or
objects that are representative of the study phenomenon, or who are especially
knowledgeable about the research question.
Alternatively, the researcher may wish to interview individuals who reflect
different ends of the range of a particular characteristic.
As an example, a researcher who wants to investigate attitudes towards death in
HIV-positive individuals may select subjects who have no symptoms and those
who have active disease and are considered terminal.
This type of sampling is commonly used in qualitative research.
17. 2. Convenience sampling:
• ‘Convenience sampling’ is also known as 'accidental sampling' or 'availability sampling',
and it involves the choice of readily available subjects or objects for the study.
• Although used frequently it is considered a poor type of sampling because it provides little
opportunity to control bias.
• Elements are included in the sample because they happen to be in the right place at the
right time.
• The researcher may choose, for example, the first 20 patients arriving at an antenatal clinic
for an interview, or the patients available in a specific ward on a certain day; or a lecturer
may use students in his/her class. Obviously, this can introduce certain biases, as some
elements may be over-represented or under-represented.
• Generalization based on such samples is extremely uncertain, although the samples so
chosen are convenient for researchers in terms of time and costs. While this type of
sampling is used in studies where probability sampling is not possible, it should be used
only when samples are .unobtainable by other means, especially in quantitative studies.
18. 3. Quota sampling:
• This sampling technique may be considered as the non-probability equivalent
of stratified sampling.
• The aim of quota sampling is to replicate the proportions of subgroups or strata
present in the population.
• However, instead of relying on random selection, this sampling procedure
relies on convenience as the basis of selection.
• The researcher first determines which strata are to be studied. Common strata
are age groups, gender, race, geographic locations, and socio-economic levels.
The researcher then determines a quota, or the number of participants, needed
for each stratum.
• The quota may be determined proportionately or disproportionately. For a
proportionate quota sample, the researcher must obtain information on the
composition of the population. For example, If the population consists of 60%
women, the sample should also consist of 60% women.
19. 4. Snowball / network sampling:
• Snowball sampling involves the using the assistance of already selected study
subjects in obtaining other potential subjects, especially where it is difficult for
the researcher to gain access to the population.
• This type of sampling consists of different stages. First, the researcher
identifies a few people who have the required characteristics. They then help
him/her to identify more people, who also possess the desired characteristics and
who are included in the next stage. The process continues until the researcher is
satisfied that the sample is sufficiently large.
• For example, the researcher wants to determine how to help people to stop
smoking. He/she may know or hear of someone who has been successful in
refraining from smoking for several years. This person is contacted and asked if
he/she knows others who have also been successful.
• This type of networking is particularly helpful in finding people who are
reluctant to make their identity known