Table of Contents
Sampling Methods in Statistics - Probability and Non-Probability Approaches............................2
Probability Sampling............................................................................................................................................2
a) Simple Random Sampling........................................................................................................................2
b) Systematic Sampling..................................................................................................................................3
c) Stratified Sampling......................................................................................................................................3
d) Cluster Sampling..........................................................................................................................................4
Non-Probability Sampling..................................................................................................................................4
a) Convenience Sampling.............................................................................................................................. 4
b) Purposive Sampling....................................................................................................................................4
c) Snowball Sampling......................................................................................................................................5
d) Quota Sampling............................................................................................................................................5
Conclusion.................................................................................................................................................................5
References................................................................................................................................................................. 6
Sampling Methods in Statistics – Probability and Non-Probability
Approaches
In statistics, sampling methods are systematic procedures used to select a subset of
individuals or items—the sample—from a much larger population. Studying a sample
rather than the entire population saves time, resources, and, in clinical contexts, avoids
unnecessary procedures. For example, a nationwide evaluation of duplex ultrasound for
detecting carotid artery stenosis would be impractical if it required scanning every patient.
Instead, a carefully chosen sample can yield valid, generalisable conclusions (Elfil & Negida,
2017).
Sampling methods are broadly divided into probability and non-probability approaches. In
probability sampling, each population member has a known chance of selection, allowing
for precise error estimation and statistical inference (Lohr, 2022). In non-probability
sampling, selection is not random, making recruitment faster and more flexible but limiting
generalisability (Taherdoost, 2016). This essay defines both types, explains their subtypes
in detail, and considers their application in ultrasound-related research.
Probability Sampling
Probability sampling minimises selection bias by ensuring that every member of the
population has a defined probability of inclusion. This promotes representativeness and
supports rigorous quantitative analysis such as hypothesis testing and the calculation of
confidence intervals (Lohr, 2022).
The general steps for probability sampling are:
1. Define the target population – e.g., all patients receiving abdominal ultrasound in a given
year.
2. Construct a sampling frame – a complete, non-duplicated list of eligible participants.
3. Select participants using a random mechanism – ensuring fairness and impartiality
(Thompson, 2012).
a) Simple Random Sampling
Definition: Each population member has an equal chance of being selected. This is achieved
by numbering all members and using a random process to select the required sample.
Steps:
1. Compile a complete list of all eligible participants.
2. Assign each participant a unique number.
3. Use a random number generator, lottery method, or table of random numbers to select
participants.
Example: In a study assessing the accuracy of ultrasound in detecting gallstones, 2,000
eligible patients are listed, each given a number. Researchers randomly select 300 numbers,
and the corresponding patients are included.
Advantages: High representativeness, low selection bias, and suitability for most statistical
analyses (Taherdoost, 2016).
Limitations: Requires a complete, up-to-date sampling frame and can be logistically difficult
for large or dispersed populations (Lohr, 2022).
b) Systematic Sampling
Definition: Participants are selected at fixed intervals from an ordered population list.
Steps:
1. Arrange the population in a complete, ordered list.
2. Determine the sample size required.
3. Calculate the sampling interval k = Population Size ÷ Desired Sample Size.
4. Choose a random starting point between 1 and k.
5. Select every kth participant until the sample is complete.
Example: In a vascular ultrasound screening program with 2,400 patients scheduled over
six months, the researcher needs 240 participants. The interval is k = 2400 ÷ 240 = 10. If the
starting point is 7, the sample includes patients 7, 17, 27, and so on.
Advantages: Simple, fast, and evenly covers the list; requires fewer random selections than
simple random sampling.
Limitations: Vulnerable to bias if the list has hidden periodic patterns (Elfil & Negida, 2017).
c) Stratified Sampling
Definition: The population is divided into subgroups (strata) that share specific
characteristics, and random samples are taken from each stratum in proportion to its size.
Steps:
1. Identify relevant strata (e.g., patient age, pregnancy trimester, type of scan).
2. Divide the population into these strata.
3. Determine how many participants to take from each stratum.
4. Randomly select participants from each stratum.
Example: In a fetal anomaly ultrasound study, patients are divided into trimesters. A
proportional number of participants are randomly selected from each trimester to ensure
balanced representation.
Advantages: Ensures representation of all key subgroups, improves precision, and allows
subgroup comparisons (Thompson, 2012).
Limitations: Requires accurate classification into strata and is more complex to plan and
execute (Taherdoost, 2016).
d) Cluster Sampling
Definition: The population is divided into clusters, and entire clusters or subsets within
clusters are sampled.
Steps:
1. Divide the population into clusters (often geographical or institutional units).
2. Randomly select the required number of clusters.
3. Either include all members from each selected cluster (one-stage) or randomly select
participants within clusters (two-stage).
Example: In a national ultrasound study, 15 hospitals are randomly selected. In a one-stage
design, all eligible patients in those hospitals are included. In a two-stage design, a random
sample from each hospital’s patients is taken.
Advantages: Cost-effective for large, dispersed populations and reduces travel and logistical
challenges (Lohr, 2022).
Limitations: Higher sampling error due to similarity within clusters; two-stage sampling
improves precision but increases complexity.
Non-Probability Sampling
Non-probability sampling does not involve random selection. It relies on researcher
judgment, convenience, or participant referral, which makes recruitment faster but limits
generalisability (Taherdoost, 2016). It is useful for pilot studies, rare conditions, or
exploratory research.
a) Convenience Sampling
Definition: Selecting participants who are easiest to access.
Steps:
1. Identify the most accessible population segment.
2. Recruit participants directly from that group.
Example: Interviewing patients already waiting for an ultrasound examination.
Advantages: Fast, inexpensive, and easy to implement; useful for early-stage research (Lohr,
2022).
Limitations: High risk of bias and poor external validity.
b) Purposive Sampling
Definition: Selecting participants based on specific characteristics or expertise.
Steps:
1. Identify the traits needed for the study.
2. Use researcher judgment to select participants meeting those criteria.
Example: Choosing only experienced sonographers to evaluate a new Doppler classification
system.
Advantages: Focuses on the most relevant participants and enables in-depth exploration
(Thompson, 2012).
Limitations: Subject to researcher bias and limited generalisability.
c) Snowball Sampling
Definition: Existing participants recruit others from their network who meet the study
criteria.
Steps:
1. Identify initial participants.
2. Ask them to refer others with similar characteristics.
3. Continue until the target sample size is reached.
Example: In a rare congenital liver anomaly study, one patient refers another, who then
refers others.
Advantages: Useful for hard-to-reach populations and efficient for rare conditions (Elfil &
Negida, 2017).
Limitations: May overrepresent certain networks and offers less control over sample
composition.
d) Quota Sampling
Definition: Predetermined numbers are set for each subgroup, but selection within
subgroups is nonrandom.
Steps:
1. Decide quotas for key subgroups (e.g. gender, age group).
2. Recruit participants until each quota is filled.
Example: Recruiting equal numbers of male and female patients for a musculoskeletal
ultrasound study without random selection.
Advantages: Ensures subgroup representation and is more structured than convenience
sampling.
Limitations: Lacks the statistical strength of stratified sampling and is still prone to
selection bias.
Conclusion
The choice between probability and non-probability sampling depends on study objectives,
resources, and the required accuracy. Probability methods simple random, systematic,
stratified, and cluster are ideal for generalizable, guideline informing research. Non-
probability methods convenience, purposive, snowball, and quota offer speed and access to
specific populations, but require careful framing to avoid overgeneralisation. The best
studies use these methods strategically, ensuring that results are both methodologically
sound and practical.
References
Elfil, M., & Negida, A. (2017). Sampling methods in clinical research; an educational review.
Emergency, 5(1), e52. [Link]
Lohr, S. (2022). Sampling: Design and analysis (3rd ed.). CRC Press.
Taherdoost, H. (2016). Sampling methods in research methodology; how to choose a
sampling technique for research. International Journal of Academic Research in
Management, 5(2), 18–27.
Thompson, S. K. (2012). Sampling (3rd ed.). John Wiley & Sons.