Sampling Methods: Probability vs Non-Probability Sampling

Reviewed by Prof. Dr. Alexander Holtermann · Updated 16 October 2026 · 5 min read

Probability sampling gives everyone a known chance of selection and supports generalization; non-probability sampling is quicker but limits it.

Key takeaways

Probability sampling (simple random, systematic, stratified, cluster) requires a sampling frame and allows statistical generalization to the population.
Non-probability sampling (convenience, purposive, quota, snowball) is quicker and cheaper but does not support statistical inference to a population.
Most student surveys use convenience samples. That is acceptable if you describe it honestly and discuss the limits on generalizability.
Qualitative research uses purposive sampling: participants are chosen because they can inform the research question.
A large sample does not fix a biased sampling method. How you sample matters more than how many you sample.

Key terms: population, sampling frame, sample

The population is the whole group you want to make statements about, for example all undergraduate students in England. The sampling frame is the list or source from which you actually draw, such as a university’s enrollment list. The sample is the subset that takes part.

Mismatches between these levels cause coverage error: if your frame is a student society mailing list, students who are not members can never be selected. Sampling error is the random difference between sample and population that occurs even with perfect random sampling; it shrinks as the sample grows. Non-response bias arises when those who respond differ systematically from those who do not.

Probability sampling methods

In probability sampling, every member of the population has a known, non-zero chance of selection. Only then can you calculate margins of error and confidence intervals in the strict sense.

Method	How it works	Example	When useful
Simple random	Every member has the same chance; selection by random number generator	300 students drawn randomly from the enrollment list	Complete frame available
Systematic	Every k-th person from a list after a random start	Every 10th name on the list, starting at 7	Long lists; avoid if the list has a periodic pattern
Stratified	Population divided into strata, random sample from each	Random samples from each faculty, proportional to size	Subgroups must be represented
Cluster	Random selection of groups, then all or some members	Randomly select 10 seminar groups and survey everyone in them	No individual list, but a list of groups

Non-probability sampling methods

Here selection depends on accessibility or on the researcher’s judgment. You cannot calculate a true margin of error, but these methods are often the only practical option.

Method	How it works	Example	Main risk
Convenience	Whoever is easy to reach	Survey link posted in course groups	Strong self-selection bias
Purposive (judgmental)	Deliberate choice of information-rich cases	Interviewing nurses with 5+ years on night shifts	Researcher bias; must be justified
Quota	Convenience sampling until quotas (e.g. gender, age) are filled	50% first-years, 50% final-years	Looks representative but is not random
Snowball	Participants recruit further participants	Students with caring responsibilities refer peers	Networks of similar people

Purposive sampling in qualitative research

Qualitative studies do not aim for statistical representativeness. They select participants who can provide rich information about the phenomenon. Common purposive strategies:

Maximum variation: deliberately different cases to capture a range of perspectives.
Homogeneous: similar cases to explore one experience in depth.
Typical case: cases that represent the average situation.
Extreme or deviant case: unusual cases that reveal something about the norm.
Criterion sampling: everyone who meets defined criteria (e.g. completed a placement abroad).
Theoretical sampling: in grounded theory, the next cases are chosen based on the emerging theory.

How to choose a sampling method

Ask yourself three questions:

What do you want to conclude? Statistical generalization (e.g. “X% of students …”) needs probability sampling. Exploring relationships or experiences can work with non-probability samples.
Do you have a sampling frame? Without a list of the population, simple random sampling is impossible; consider cluster sampling or accept a non-probability approach.
What are your resources? Time, budget, access and data protection rules (many universities do not release student email lists) limit your options.

Describing your sample in the methods chapter

Examiners do not expect a perfect random sample in a dissertation. They expect transparency. State:

the target population and the sampling frame (if any)
the sampling method, named with the correct term
recruitment channels and the period of data collection
inclusion and exclusion criteria
the achieved sample and its key characteristics compared with the population, where known
the consequences for generalizability in the limitations section

Common mistakes

These errors are frequent:

Calling a convenience sample “random” because the link was shared widely.
Calling a quota sample “representative”.
Believing that a very large online sample removes self-selection bias.
Using statistical generalization language for purposive interview samples.
Not reporting how many people were invited, started and completed.

Tools and institutional requirements

Our sample size calculator shows how many responses you need, and the methods tool helps you choose a design. You can distribute your survey via a link or QR code from /umfragen. Whether a convenience sample is acceptable depends on your department and level of study; your university’s own guidelines take precedence. Upload them to Professor Digital and the advice is tailored to them.

Frequently asked questions

What is the difference between probability and non-probability sampling?

In probability sampling every member of the population has a known chance of selection, which allows statistical generalization. In non-probability sampling, selection depends on accessibility or judgment, so results cannot be generalized statistically.

Which sampling method is best for a dissertation?

For surveys, often a convenience or stratified sample, depending on access. For interviews, purposive sampling. The best method is the one you can justify and describe transparently.

Is convenience sampling acceptable?

Yes, in most student projects, provided you name it correctly and discuss the limits on generalizability in your limitations section.

What is the difference between stratified and quota sampling?

Both divide the population into groups. Stratified sampling then selects randomly within each group; quota sampling fills each group with whoever is available, so it is non-probability.

What is snowball sampling used for?

For hard-to-reach populations without a sampling frame, such as people with rare experiences. Participants refer others, which helps access but tends to produce similar networks.

Sources and further reading

Clark, T., Foster, L., Sloan, L., & Bryman, A. (2021). Bryman’s Social Research Methods (6th ed.). Oxford University Press.
Creswell, J. W., & Creswell, J. D. (2023). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (6th ed.). Sage.

Your university’s examination regulations and style guide always take precedence.