Sampling is the decision about who or what can provide evidence for the dissertation and how those cases will be selected. A sample is not justified simply because its size resembles another study. The selection strategy needs to fit the research question, design, access, and type of claim the dissertation intends to make.
Define the population before choosing a sampling method
Be clear about the larger group, set of cases, documents, records, organisations, or observations from which the study is selecting. If that boundary is vague, the sampling strategy will also be difficult to justify.
Match the sampling logic to the research design
Different designs need different forms of selection. A quantitative study seeking population estimates may require a different strategy from a qualitative study seeking information-rich cases. Start from the design rather than choosing a familiar sampling label first. See research design help when that foundation is uncertain.
Know when probability sampling is relevant
Probability approaches depend on a defined population and a selection process in which units have a known chance of inclusion. The exact method—such as simple random, systematic, stratified, or cluster sampling—should be selected for a reason rather than listed without explanation.
Probability sampling becomes most useful when the study has a reasonably defined sampling frame and wants to make population-level estimates or comparisons. Simple random sampling can reduce systematic selection, stratified sampling can protect the representation of important subgroups, and cluster approaches may make geographically dispersed populations more manageable. The dissertation still needs to explain why the chosen form fits the population and research objective.
A probability label does not automatically make a sample representative. Non-response, an incomplete sampling frame, inaccessible groups and recruitment failures can still introduce bias. These limitations belong in the sampling justification rather than being hidden behind the name of the technique.
Use non-probability sampling deliberately
Purposive, convenience, snowball, quota, and other non-probability approaches solve different access and evidence problems. Their limitations also differ. A dissertation should explain why the chosen approach fits the research purpose and what it prevents the study from claiming.
Non-probability sampling is often appropriate when the research needs participants with particular experiences, when the population cannot be enumerated reliably, or when the dissertation prioritises depth rather than population estimation. Purposive sampling can identify information-rich cases, snowball sampling can help reach difficult-to-identify participants, and convenience sampling may be feasible where access is restricted.
The important question is what the selection process allows the researcher to claim. A convenience sample should not be described as though every member of the wider population had an equal chance of inclusion, while a purposive qualitative sample should be judged according to the relevance and richness of the selected cases rather than statistical representativeness.
Explain inclusion and exclusion criteria
Criteria make the sample boundary transparent. They should follow from the research question and evidence need rather than being added after recruitment has already begun.
Justify sample size from the study rather than a universal number
There is no single correct sample size for every dissertation. Quantitative studies may need a defensible statistical basis. Qualitative studies may justify depth, information power, or saturation depending on the design and discipline. Institutional guidance and supervisor expectations should be followed where they specify requirements.
For quantitative work, sample-size reasoning may depend on the intended statistical analysis, expected effect, variability, number of predictors or groups, and the amount of usable data expected after non-response. Where formal power analysis is appropriate, its assumptions should be stated rather than reporting a number without explanation.
Qualitative justification follows a different logic. The researcher may consider the specificity of the participants, the complexity of the question, the quality of the interviews or observations, and whether additional cases continue to add meaningful information. The justification should therefore be tied to the actual design and evidence need.
Account for access and recruitment
A sampling plan must be feasible. Consider whether participants, cases, documents, or datasets are actually accessible and whether permissions or ethics approval are required. For ethics-specific questions, use dissertation ethics help.
Connect sampling limitations to the claims you make
The sampling strategy affects the scope of interpretation. Limitations should therefore be discussed in relation to representativeness, transferability, selection bias, coverage, or other design-relevant concerns rather than as a generic weakness list.
For the data-collection decision that follows sampling, continue to data collection help. For broader methodology integration, return to Methodology Chapter Help.