03 · What You Need to Know
What Determines the Number of Qualitative Interviews You Need?
Qualitative Interview Sample Size Serves a Different Purpose from Survey Sample Size
In many quantitative surveys, sample size is planned to achieve a desired level of statistical precision or power under specified assumptions. Qualitative interview studies often pursue different objectives.
They may seek to understand how participants interpret experiences, how practices operate within particular settings, or how processes differ across relevant circumstances.
The adequacy of an interview sample therefore depends on whether the material can support the intended analysis, rather than whether the number of participants satisfies a conventional statistical threshold.
This does not mean that qualitative sample size is unimportant or that any number is acceptable. Too few interviews may leave important experiences unexplored, while collecting substantially more interviews than the analysis requires can create unnecessary participant burden and analytical workload.
The planning task is to establish a reasoned relationship between the research question, the sample, and the proposed analytical claims.
Start with the Scope and Specificity of the Research Question
A narrowly focused question involving a relatively specific group may require fewer interviews than a broad question encompassing substantially different experiences and contexts.
For example, a study investigating how newly appointed instructional designers at one university experience a particular onboarding program has a more bounded focus than a study examining how educators across multiple sectors adapt to artificial intelligence.
The second investigation may involve different professional responsibilities, organizational environments, and technology practices. Adequately examining those differences could require a broader sample.
However, breadth alone does not determine the number. Researchers must also consider whether the study intends to analyze these differences systematically or merely include them as contextual variation.
A useful starting question is: What range of experiences must the sample contain for the proposed interpretation to be credible?
Participant Specificity Influences Information Power
Malterud, Siersma, and Guassora (2016) proposed the concept of information power as a way to reason about qualitative sample adequacy.
The central proposition is that the more relevant information a sample provides for the study's purpose, the fewer participants may be required.
Their framework considers five dimensions:
- The breadth or narrowness of the study aim.
- The specificity of the sample in relation to that aim.
- The contribution of established theory.
- The quality of interview dialogue.
- The analytical strategy.
For example, participants selected because they have direct, substantial experience with a narrowly defined phenomenon may provide more relevant information than a loosely recruited group with only incidental exposure.
Information power is not a formula that converts these dimensions into a fixed interview count. It is a framework for evaluating whether the planned and achieved sample can support the study's purpose.
Sample Diversity Matters When Comparing Experiences
Researchers often use purposive sampling to recruit participants whose characteristics or experiences are relevant to the research question.
Sometimes the objective is to recruit a relatively homogeneous group. In other studies, meaningful variation is essential.
Consider research examining how university teachers experience generative AI governance. The researcher may want to include teachers from different disciplines, employment arrangements, or levels of AI experience.
If these differences are central to the analysis, the sample must contain enough relevant cases to examine them meaningfully.
However, qualitative diversity should not be confused with statistical representativeness. Including one participant from each academic discipline does not necessarily provide an adequate basis for comparing disciplinary experiences.
The researcher should distinguish characteristics included to broaden contextual understanding from groups that will be treated as substantive analytical categories.
The Analytical Approach Changes What Counts as Adequate
Qualitative analysis is not a single procedure, and different analytical traditions have different expectations about sampling.
A focused qualitative descriptive study may seek recurring patterns across a relatively bounded set of experiences.
An interpretative phenomenological analysis may prioritize detailed idiographic examination of individual cases, often using relatively small and carefully selected samples.
Grounded theory research may require iterative recruitment informed by developing theoretical categories and relationships.
Narrative inquiry may concentrate on a small number of extended accounts, while a multi-site comparative study may require enough cases to examine contextual variation.
Consequently, sample size should be justified in relation to the actual analytical approach rather than the broad label "qualitative research."
Interview Depth and Dialogue Quality Affect Sample Adequacy
Not all interviews generate the same amount or kind of relevant information.
A detailed conversation with a participant who has substantial experience of the phenomenon may contribute more to a narrowly focused analysis than a brief interview containing mostly general statements.
Interview quality depends on participant engagement, question design, interviewer skill, context, and the opportunity for meaningful follow-up.
The choice among structured, semi-structured, and unstructured interviews may therefore influence the depth and comparability of the material.
However, researchers should not assume that longer interviews are automatically richer or that a small number of articulate participants can adequately represent every important perspective.
Dialogue quality should be evaluated in relation to the evidence needed for analysis.
What Does Existing Research Say About Interview Numbers?
Methodological studies provide useful empirical observations, but they do not establish universal sample-size requirements.
Guest, Bunce, and Johnson (2006) examined thematic saturation in an analysis of interviews with a relatively homogeneous group. Their findings suggested that many commonly identified themes emerged within the first 12 interviews in that particular study.
Hennink, Kaiser, and Marconi (2017) distinguished code saturation from meaning saturation. Their analysis showed that identifying a range of thematic issues and developing a more nuanced understanding of those issues can require different amounts of data.
Hennink and Kaiser (2022), in a systematic review of empirical studies, reported that saturation was often observed within approximately 9 to 17 interviews in studies involving relatively homogeneous populations and narrowly defined objectives, while emphasizing important contextual variation.
These findings should not be converted into a rule that every qualitative study requires 12, 15, or 17 participants.
The studies examined particular forms of analysis, populations, and operational definitions of saturation. Their results may inform planning where the methodological conditions are sufficiently comparable, but they cannot determine adequacy for every qualitative tradition.
What Is the Difference Between Code Saturation and Meaning Saturation?
Code saturation concerns the point at which additional interviews yield few or no new thematic codes under a specified analytical procedure.
Meaning saturation concerns whether the researcher has developed sufficient depth, variation, and understanding of the issues represented by those codes.
For example, several teachers may mention "lack of institutional support" early in a study. The researcher may quickly identify that category.
However, understanding how support differs across departments, why some teachers perceive it as inadequate, and how it shapes instructional decisions may require additional interviews.
Identifying a topic is not equivalent to understanding its complexity.
Researchers considering whether sampling should continue until saturation should specify what kind of analytical adequacy they are seeking and whether saturation is appropriate to their methodological orientation.
Should You State a Fixed Sample Size in Your Proposal?
A research proposal generally needs a credible estimate of recruitment requirements, resources, and expected data volume.
One defensible approach is to specify a provisional target or range and explain the assumptions supporting it.
For example, a researcher might plan approximately 18 to 24 interviews because the study includes several analytically important participant categories and requires detailed comparisons across experiences.
That range would be a planning judgment, not a universal methodological requirement.
Where iterative recruitment is appropriate, the protocol can explain how analysis will inform decisions about further interviews.
Any changes must remain consistent with ethics approval, consent arrangements, recruitment feasibility, and institutional requirements.
A provisional target should be neither an arbitrary promise nor an unlimited commitment to continue interviewing indefinitely.
When Might a Small Interview Sample Be Defensible?
A relatively small sample may be appropriate when the research question is narrowly focused, participants have highly relevant experiences, and the analytical approach emphasizes detailed interpretation of individual cases.
For example, an idiographic study of a small, clearly defined group of specialist educators may prioritize extensive examination of each participant's account.
However, a small sample does not become adequate simply because the study is qualitative.
The researcher must demonstrate how the selected cases support the intended interpretation and acknowledge the limits of the evidence.
Claims about broad variation, subgroup differences, or common experiences across diverse settings generally require additional justification.
When Might More Interviews Be Necessary?
Additional interviews may be justified when the study encompasses heterogeneous participants, multiple institutions, contrasting contexts, or developing theoretical categories.
They may also be necessary when initial interviews reveal important experiences that were not represented in the original sample.
For example, a study of AI adoption across universities may discover that teachers working under restrictive institutional policies describe substantially different practices from those in permissive environments.
If that contrast becomes analytically important, the researcher may need further participants from each context.
Nevertheless, increasing the interview count does not automatically improve the analysis. Researchers must still examine the relevance, depth, and interpretive contribution of additional material.
Should You Allocate a Minimum Number to Every Subgroup?
Not necessarily, but planned subgroup comparisons require particular attention.
If the study seeks to compare novice and experienced teachers, it should recruit enough participants within each category to support meaningful examination of similarities and differences.
One or two participants may provide valuable illustrative cases but may not adequately represent the range of experiences within a heterogeneous subgroup.
There is no universal minimum per subgroup that guarantees qualitative adequacy.
The required coverage depends on within-group variation, the specificity of the research question, and the depth of comparison intended.
Researchers should avoid creating numerous demographic categories that the planned sample cannot support analytically.
How Do Recruitment and Nonparticipation Affect the Plan?
The number of participants approached will usually differ from the number ultimately interviewed.
Some eligible individuals may decline, be unavailable, or withdraw before the interview.
Researchers should therefore distinguish the intended number of completed interviews from the number of people who must be contacted.
Recruitment planning may include alternative eligible participants, appropriate follow-up procedures, and realistic scheduling allowances.
However, recruiting additional participants should not override voluntary participation or place inappropriate pressure on people who decline.
In sensitive or highly specialized populations, recruitment feasibility may be a substantial constraint that needs to be acknowledged in the study design.
Interviewing and Analysis Should Inform Each Other When the Methodology Permits
In iterative qualitative research, analysis can begin while interviews are still being conducted.
Early examination of transcripts, field notes, and emerging interpretations may reveal gaps in participant coverage or issues requiring further exploration.
Researchers may then refine recruitment priorities or interview prompts within the boundaries of the approved study.
This does not mean every qualitative project must use iterative sampling. Some studies have a fixed recruitment plan for legitimate methodological or operational reasons.
The important point is to explain how the chosen sampling procedure supports the analytical approach.
What Should You Document to Justify the Final Interview Count?
The methods section should explain the initial sampling rationale and how the achieved sample supported the analysis.
Relevant information may include participant eligibility, sampling strategy, recruitment process, intended variation, interview depth, analytical approach, and the reasons recruitment ended.
If saturation or information power informed the decision, describe how these concepts were applied rather than merely stating that sufficient data were obtained.
Reporting frameworks such as COREQ and the Standards for Reporting Qualitative Research encourage transparent descriptions of sampling and data collection.
Researchers should also distinguish the number of completed interviews from the number of participants when some individuals are interviewed repeatedly.
The justification should demonstrate methodological coherence, not simply satisfy an expected numerical convention.
