03 · What You Need to Know
What Makes Purposive Sampling Methodologically Defensible?
What Exactly Is Purposive Sampling?
Purposive sampling, also called purposeful sampling, is a non-probability approach in which researchers deliberately select cases according to characteristics or experiences relevant to the study.
Rather than giving every population member a known probability of selection, the researcher uses a defined rationale to identify cases capable of contributing evidence to the research question.
For example, a study examining institutional responses to generative AI misconduct might recruit academic integrity officers, faculty members involved in disciplinary proceedings, and students who have experienced relevant institutional processes.
The participants are not selected because they are assumed to represent the average university member. They are selected because their positions or experiences make them relevant to the phenomenon being investigated.
Purposive sampling is therefore not synonymous with arbitrary selection. A defensible design requires a clear connection between the research question, the characteristics used for selection, and the intended analytical claims.
When Is Purposive Sampling Particularly Appropriate?
Purposive sampling is useful when the research question requires evidence from people or cases possessing specific characteristics that cannot be assumed to occur frequently in an unrestricted sample.
Common situations include studies of specialized professional practice, unusual institutional experiences, implementation processes, and participants who have direct knowledge of a particular event.
For instance, researchers studying how universities establish AI assessment policies may need participants who have participated in policy development. Recruiting faculty members who have never encountered the process may add relatively little evidence to that particular question.
Purposive sampling can also be appropriate when the study seeks contrasting experiences. A researcher may deliberately select institutions with different implementation conditions to examine how those conditions shape practice.
Its value depends on the information required, not simply on the difficulty of obtaining a random sample.
How Is Purposive Sampling Different from Convenience Sampling?
The distinction concerns the principal reason participants are selected.
Convenience sampling
Participants are recruited primarily because they are accessible, available, or easy to contact.
Purposive sampling
Participants are selected primarily because they satisfy criteria relevant to the research question.
Suppose a researcher interviews ten teachers from the same department because they are colleagues. That recruitment is principally convenience-based.
If the researcher identifies teachers who have implemented distinct AI assessment practices and selects them to examine those practices, the selection has a purposive rationale.
These approaches can overlap. Researchers may identify relevant participants purposively but recruit only those accessible within one institution. The methodology should acknowledge both the selection rationale and the practical restrictions.
When accessibility is the dominant selection principle, the justification for convenience sampling may be more appropriate than presenting the sample as fully purposive.
Which Purposive Sampling Strategies Can Researchers Use?
Purposive sampling encompasses several strategies. The appropriate choice depends on the kind of evidence the study needs.
| Strategy |
Selection principle |
When it may be useful |
| Criterion sampling |
Select cases meeting predefined conditions |
Studying participants with a particular experience or eligibility characteristic |
| Maximum variation sampling |
Select cases differing across important dimensions |
Examining variation and shared patterns across contrasting contexts |
| Homogeneous sampling |
Select cases sharing a narrowly defined characteristic |
Investigating a specific group's experiences in depth |
| Expert sampling |
Select participants with relevant specialist knowledge |
Investigating technical decisions or professional judgments |
| Typical-case sampling |
Select cases considered typical according to justified criteria |
Understanding ordinary processes within a defined context |
| Critical-case sampling |
Select cases with particular analytical significance |
Examining a proposition under strategically informative conditions |
These strategies are not interchangeable. Selecting experts to explain institutional decision-making differs from selecting diverse students to examine variation in learning experiences.
For qualitative studies, the decision may involve maximum variation or homogeneous sampling, depending on whether breadth of perspectives or depth within a particular group is more important.
How Should Selection Criteria Be Established?
Selection criteria should be derived from the research question and the evidence required to answer it.
For example, a study examining faculty decisions about AI-assisted grading might require participants who have directly used, evaluated, or formally rejected such systems within a specified period.
The researcher should distinguish essential eligibility conditions from characteristics used to obtain useful variation.
Essential criteria determine whether a participant belongs to the relevant study population. Additional criteria may guide the selection of contrasting cases, institutional settings, or levels of experience.
Criteria should be sufficiently precise to apply consistently, but not so restrictive that they exclude relevant perspectives without justification.
Researchers should also explain how they will verify that potential participants satisfy the criteria. Self-reported experience may be appropriate in some studies, while others may require administrative confirmation or documentary evidence.
What Makes a Participant Information-Rich?
An information-rich participant is someone whose experiences or knowledge can contribute substantively to understanding the phenomenon being studied.
Information richness is relational. It depends on the research question rather than a participant's general expertise or professional status.
A university administrator may be highly knowledgeable about institutional policy but unable to describe how students actually experience AI-generated feedback during coursework.
Conversely, a student with no formal expertise in educational technology may provide valuable first-hand evidence about how feedback is interpreted and acted upon.
The researcher should therefore identify information-rich participants according to the particular knowledge or experience needed, rather than assuming that seniority, accessibility, or professional reputation guarantees relevance.
How Many Participants Are Needed?
Purposive sampling does not have a universal minimum or maximum sample size.
The appropriate number depends on the research question, analytical approach, specificity of the participants, quality of the data, and intended depth of analysis.
Malterud and colleagues describe information power as one way of thinking about qualitative sample adequacy. Under this perspective, the more relevant information a sample provides for the study's aim, the fewer participants may be needed, although adequacy remains dependent on the study design and analysis.
Some qualitative approaches also consider saturation, but the meaning and usefulness of saturation vary across methodological traditions.
Researchers should avoid asserting that a fixed number of interviews is sufficient for every purposive sample.
A defensible sample size rationale should explain what analytical work the selected cases must support and how adequacy will be evaluated.
Can Purposive Sampling Change During Data Collection?
Yes, particularly in iterative qualitative designs.
Early interviews may reveal perspectives or circumstances that were not sufficiently represented in the initial sample. Researchers may then refine recruitment to examine those emerging issues.
For example, an investigation of AI assessment practices might initially focus on experienced faculty members but subsequently identify the need to include instructors who discontinued AI use after implementation difficulties.
Such adaptation can be defensible when it follows the study's analytical logic and is documented transparently.
However, changes should not be used merely to recruit participants likely to confirm the researcher's preferred interpretation.
In methodologies such as grounded theory, theoretical sampling has a more specific purpose: selecting additional cases to develop and refine emerging theoretical categories.
Can Purposive Sampling Support Generalization?
Purposive sampling does not ordinarily support conventional design-based statistical generalization because the inclusion probabilities of population units are not known.
Nevertheless, qualitative findings may contribute to theoretical understanding, analytical generalization, or transferability when the research design and evidence support those interpretations.
For example, a study of how experienced teachers respond to AI-generated assessment recommendations may identify processes that are relevant beyond the participating institutions.
Whether those processes apply elsewhere depends on contextual similarities, theoretical arguments, and additional evidence.
Researchers should distinguish identifying a plausible process from estimating how frequently it occurs in a population.
A purposive sample can provide substantial evidence about how or why a phenomenon occurs without establishing its population prevalence.
What Are the Main Risks of Purposive Sampling?
The principal risks arise when selection criteria are poorly justified, important perspectives are excluded, or the researcher treats selected cases as though they represent the wider population statistically.
Researcher judgment can introduce confirmation bias if participants are chosen because they are expected to support a preferred explanation.
Gatekeepers may also influence recruitment by recommending cooperative or institutionally favored participants while discouraging access to critical voices.
To reduce these risks, researchers should document selection criteria, recruitment pathways, relevant exclusions, and changes made during data collection.
Where appropriate, deliberate attention to contrasting experiences or negative and deviant cases may strengthen the analysis by challenging emerging interpretations.
These practices improve methodological transparency and analytical credibility, but they do not convert purposive sampling into probability sampling.
How Should Purposive Sampling Be Reported?
A strong methodology explains who was eligible, why the selected characteristics mattered, how potential participants were identified, and how recruitment decisions were made.
For qualitative interview studies, reporting frameworks such as COREQ and the Standards for Reporting Qualitative Research provide useful guidance on describing participant selection, research context, and methodological decisions.
Researchers should also report relevant limitations, including restricted institutional access, gatekeeper influence, and perspectives that may not have been captured.
The justification should demonstrate that the sample was designed to generate evidence relevant to the research question rather than merely assembled from willing participants.
04 · A Practical Example
Selecting Faculty Members to Study AI Assessment Decisions
Hypothetical Example
A Qualitative Study of Generative AI in Assessment
A researcher wants to understand how university faculty members decide whether to permit, restrict, or integrate generative AI in student assessments.
The research question concerns decision-making processes rather than the percentage of faculty members adopting a particular policy.
The researcher therefore proposes purposive sampling of faculty members with direct experience making or implementing AI-related assessment decisions.
Define eligibility.
Recruit faculty members who have made substantive decisions about generative AI use in assessment during the preceding two academic years.
Identify meaningful variation.
Seek participants from different disciplines, institutional sectors, and approaches to AI assessment.
Verify relevance.
Use a short screening procedure to confirm that potential participants have directly encountered the decisions under investigation.
Review emerging evidence.
Examine whether the interviews reveal relevant differences or perspectives that require further recruitment.
Document the final sample.
Report the selection criteria, recruitment channels, participant characteristics, and reasons for any changes.
Suppose the researcher interviews 18 faculty members, including those who encourage AI use, those who restrict it, and those who have revised their policies after implementation difficulties.
The interviews reveal contrasting approaches to academic integrity, assessment authenticity, and student disclosure.
The researcher may use these accounts to develop an interpretation of how faculty members negotiate competing assessment priorities.
However, the researcher should not conclude that a particular percentage of all university faculty members follows one approach merely because that percentage appeared among the 18 interviewees.
The purposive sample is defensible because the selected participants have direct experience with the decisions being studied and the analysis remains focused on understanding those decisions.
