Sampling
What is sampling?
• Sampling is the process involving the selection of a finite number of
elements from a given population of interest, for purposes of inquiry.
12/05/2023 1
What is a sample?
• A sample is a representative part of a population.
• A decision is often made, therefore, to study only a small fraction of
the population, or a “sample” of it, from which conclusions can be
drawn about the whole population.
12/05/2023 2
What are the characteristics a sample should possess?
• A sample should possess all the characteristics of the population
from which it is drawn, if possible, so that is fully representative of the
population.
• The method of sample selection usually determines its
representative nature.
12/05/2023 3
The sampling process
Reasons for sampling:
i) Sampling for an exploratory purpose, in other words
to get a general impression of the total population of
interest (quiet subjective).
ii) Sampling for purposes of obtaining estimates on
certain characteristics of the population (objective).
12/05/2023 4
The sampling process…
• Several reasons make sampling more useful than complete
enumeration.
• These include considerations regarding:
• Time
• Costs and available resources, and
• Practicability.
12/05/2023 5
What are the common methods of sample selection?
• There are two basic/general types of sampling techniques:
Probability (or random) and non probability sampling.
• The nature of your research study will determine which type of
sampling you should use.
• Large-scale descriptive studies almost always use probability-sampling
techniques.
• Intervention studies sometimes use probability sampling but also
frequently use non-probability sampling.
• Qualitative studies almost always use non-probability samples.
12/05/2023 6
Probability Sampling
• This is a sampling technique which employs random procedures
whereby the selection of the sampling unit is done on the basis of
chance.
• This chance selection ensures that every member of the population
has an equal chance of being included in the sample.
12/05/2023 7
Probability Sampling…
• Probability sampling is a technique you can use to
maximize external validity or generalizability of the
results of the study.
• Descriptive studies and large operational research
intervention studies are frequently designed so that
their results are as generalizable as possible.
12/05/2023 8
Non probability Sampling
• Non-probability sampling refers to the selection of a sample that is
not based on known probabilities.
• Non-probability sampling procedures are not valid for obtaining a
sample that is truly representative of a larger population.
12/05/2023 9
Sample Size Determination
• Many handbooks contain formulae for estimating sample size because
the size of the sample is one of the most important determinates of
the accuracy of survey estimates.
• We will focus on the basic questions the researcher should ask when
determining sample size and choosing a sample
12/05/2023 10
Basic questions that should be asked when choosing a sample.
1. How large a sample can you collect?
The best advice that can usually be given to a new researcher is
that as large a sample as possible should be used.
The larger the sample the smaller the chance that the sample
will be markedly different from the population it should
represent (Minimize sampling error).
2. What level of budget do you have for the study?
The safest procedure is to select as large as your fund can
support, within reason.
12/05/2023 11
Basic questions…
3. What is the prevalence of the condition you are studying?
If you are studying a condition that appears quite often in a
population, you can take a smaller sample than if the condition is
quite rare.
4. What staff are available to gather the sample?
Limited human resources may be a constraint on sample size.
12/05/2023 12
Basic questions…
5. How much time do you have for the research?
• You can only study a limited number of people in a
certain time.
6. Into how many cells or categories are you going to
divide your data for analytical purposes?
• The more categories planned for analysis, the larger the
sample must be.
7. What is the general rule concerning minimum sample
12/05/2023
size? 13
Sample size…
It is generally recommended that a sample size of at
least 10, or preferably 20 to 30, be selected for each
sub-part of data or cell of the design that will be
analyzed.
The absolute size of the sample is more important to
accuracy of estimates than the proportion of the
population sampled.
A smaller sample can be used if the population is
relatively homogenous
12/05/2023 14
Sample Size to Estimate a Single
Population Proportion
• If the main objective of the study is to estimate single population proportion, then
the sample size can be determined using the formula:
• Where
• n is the minimum sample size required for very large population
(≥10,000)
• Z is the critical value for a given confidence interval
• P is expected proportion of the event to be studied (to be estimated
based on findings of previous studies)
12/05/2023 • d is margin of error 15