Sampling methods

Dr Aya Mostafa Abdellatif

Lecturer at Community, Environmental and Occupational medicine Department
Faculty of Medicine
Ain Shams University
Intended Learning Objectives (ILOs):
✓ Identify the need of sampling and some definitions
about sample and population.

✓Recognize the different types random and non-random

samples.

✓Identifies the advantages and disadvantages of each

method of sampling
 When we collect data there are
Two options:
Either to collect data from all members of the
“population”
Or
Select a subset from the population and collect the
data from that “sample”
Why investigators use samples instead of the
whole population

Cost and Time

Feasibility Accessibility
effectiveness

Accuracy Bias
Definitions

Target Population – the population we are interested to generalize our

finding upon
Study population – the population selected to be sampled, it is a
subset of the target population.
Sample – a subset of the study population or a collection of sampling
units from the eligible study population
Definitions
Sampling frame – a list of all of the sampling units in the study population

Sampling unit – the basic unit around which a sampling procedure is planned
– Person – a patient, a doctor, an administrator.
– Group – department, ICU, hospital, Primary Health Unit.
– Component – eye, physiological response

Sampling fraction = size of the sample/size of the population. If we have 1000

medical students in the fourth grade and wish to get a sample of 100 students,
the sampling fraction would be 100/1000=10% or 1 of 10.
 Challenges in sampling process
To be representative of the To give accurate
sampled population use estimates use a
appropriate random sample of appropriate
sampling method size

How How many

Representative Accuacy
 Types of samples

A. Random Sample (Probability B. Non-Random Sample (Non-

sample) probability sample)
It is a sampling method in which It is a sampling method in which
every unit in the sampled the probability of being included in
population has an equal probability the sample is not known and not
or chance of being selected. equal for the sampling units.
 Types of sampling:
A. Random (Probability) sampling:
– Simple random sample
– Systematic random sample
– Stratified random sample
– Cluster random sample
– Multistage random sample
B. Non-random (non-probability) sampling
– Convenience sample
– Purposive sample
o Expert sample
o Quota sample
o Snowball sample
 A. Random
(Probability) sampling
A.1. Simple random sample
– Determine the sampling proportion, e.g., 10% of the
population
– Give every case a unique serial number
– Sample by random number tables or a computer
program
Limitations in simple random sample
• Simple but needs sample frame
and may not be representative?!

• Possibility of loss of small subsets

of population

• May be time consuming

A.2. Systematic random sample
– Determine the sampling proportion, e.g., 10% of the
population
– From the first 10% select a random case by random
number table = y
– Select cases y, y+10, y+20 and so on
Limitations in systematic random
sample
• Simple but needs sample frame
and may not be representative?!

• Possibility of systematic
phenomena in the population
A.3. Stratified random sample
• Used if certain strata / group of the
population need to be included and
feared from being lost.
• Stratify the population into different
strata.
• Select from each stratum its
contribution to the whole sample
using a simple or a systematic random
sampling method.
Stratified random sample may be:
➢ Proportionate Stratified random Sampling
➢ Disproportionate Stratified random Sampling
The sampling fraction is the primary differentiating factor
between proportionate and disproportionate stratified
random sampling.
Proportionate Stratified random Sampling:

▪ The size of the sample from each stratum or group is proportionate to the size of
that stratum in the sampled population.
▪ We use the same sampling fraction across all strata

Example:
Disproportionate Stratified random Sampling:

▪ We use different sampling fraction across all strata

Example:
A.4. Cluster sample
▪ This is done when the sampling unit is not the individual, but a
group of individuals called clusters.
▪ So, the original population is divided into clusters each contain
approximately the same number of individuals.
▪ Sample frame needed is a frame of the sampling unit rather
than study unit which is much easier in many situations
▪ Needs special methods in data analysis
A.5. Multi-stage sampling
When sampling is done on many stages and each stage may employ
one or more of the above-mentioned sampling techniques.
Example:
Class 1
To get a sample of secondary First year
Class 2
school students in Cairo School 1 Second year
District A
Secondary schools in Cairo

School 2 Third year

Governmental
District B
schools
Private national
District C
schools
International
schools
 B. Non- Random
(Non-Probability) sampling
Non-probability (Non-random) Sample
It is a sampling method in which the probability of being
included in the sample is not known and not equal for the
sampling units.

Do not involve random selection

Do not depend on probability theory

May represent the population but their

success to do so cannot be guaranteed.
B.1. Convenience Sample:

Includes easily available members of the population.

Typically done with persons in the street to get a quick public opinion.
Often, respondents are selected because they happen to be in the right place at
the right time.
Examples
• Diabetic patients attending a clinic ❖There is no evidence that
• “people on the street” interviews
they are representative of
the populations we are
• persons sitting in the first row of a movie theatre. interested in generalizing
our results to.
B.2. Purposive Sample:

In purposive sampling, we sample with a purpose in mind.

purposive
sampling

Expert Quota Snowball

Sampling Sampling Sampling

It is better than convenient sampling

A proper sample in the view of the expert:
– A sample of all deliveries that occurred in the night
shift to examine the time from decision to incision in
CS candidates
– Patient falls in Geriatric Wards
– Medication errors in Psychiatric Wards
▪ We do not have a sample frame and the process of selecting individuals
according to specified criteria continues till a predetermined sample size is
attained.

▪ A 10% sample of discharges for patient satisfaction, e.g., 120 patient per month:
60 medical patients, 30 surgical patients and 30 pediatric patients
▪ You begin by identifying someone who meets the criteria for inclusion in your
study.
▪ You then ask them to recommend others who they may know and also meet the
criteria.
▪ Used in sensitive issues as drug addicts or
▪ Where it is impossible to have an accurate sample frame as refugees.
Selecting a sampling method

❖Population to be studied

▪Size/geographical distribution

▪Heterogeneity with respect to variable

❖Availability of list of sampling units (sample frame)

❖Level of precision required

❖Resources available
 Classification of Sampling
Methods
Sampling
Methods

Probability Non- Expert

Samples probability

Systematic Stratified Convenience Purposive

Snowball

Multi-stage
Simple
Cluster Quota
Random