One of the first questions every qualitative researcher gets asked — by a supervisor, a reviewer, or an ethics board — is some version of “how many participants?” Unlike a survey, where sample size comes from a power calculation, qualitative sample size is usually justified after the fact, by reaching saturation: the point where new data stops producing new insight. This guide covers how researchers actually select participants and how saturation is defined, argued for, and — increasingly — measured.
Purposive sampling: selecting for insight, not representativeness
Qualitative research generally uses purposive sampling — deliberately selecting participants because they have the knowledge, experience, or characteristics relevant to the research question, rather than aiming for a statistically representative cross-section of a population (Ahmad & Wilkins, 2024; DeJonckheere & Vaughn, 2019). A related, more iterative version is theoretical sampling, where each new participant or data source is chosen based on what’s been learned from the data collected so far, rather than being fixed in advance (SAGE Research Methods Foundations, 2020) — this is the sampling logic grounded theory is built on.
The common thread across both is that “who to include” is an analytical decision, not a statistical one. A study of first-time fathers’ experience of postnatal depression isn’t weakened by only including first-time fathers who experienced it — that specificity is the point.
What saturation actually means
Saturation is usually described as the point where continuing to collect data stops surfacing anything new relevant to the research question. But there are at least two different things researchers mean when they say this, and conflating them causes real confusion:
- Code saturation — no new codes are emerging from new data.
- Meaning saturation — the full depth and nuance of existing themes is understood, even if no new codes appear.
This distinction matters because meaning saturation typically takes noticeably more data than code saturation to reach (Bouncken, Czakon, & Schmitt, 2025). A study can hit “no new codes” fairly early while still not fully understanding the themes it already has.
What the evidence actually shows
A widely cited empirical study of saturation in interview-based research found that the basic set of “metathemes” was present by six interviews, and saturation of the full codebook was reached by twelve (Guest, Bunce, & Johnson, 2006) — a finding that’s often cited as informal justification for smaller qualitative samples. More recent systematic reviews of saturation across many published studies converge on a broader range: many homogeneous-population interview studies saturate somewhere around 9 to 17 interviews, while studies aiming for meaning saturation — or working with more heterogeneous populations — often need more, sometimes in the range of 16 to 24 interviews (Hennink & Kaiser, 2021; Bouncken et al., 2025).
Sample size in practice depends on several things at once: how homogeneous the population is, how tightly scoped the research aim is, how skilled and consistent the interviewing is, and which analytic approach is being used (Hennink & Kaiser, 2021; DeJonckheere & Vaughn, 2019). There is no single correct number — the honest answer to “how many interviews” is “enough to reach the kind of saturation your study needs, which you won’t know for certain until you’re coding the data” (Ahmad & Wilkins, 2024).
What this means for planning a study
In practice, most researchers propose a sample size range rather than a fixed number, justify it against comparable published studies, and stay explicit in their methods section about which kind of saturation they’re claiming — code or meaning — and how they checked for it. Reporting “interviews continued until no new codes emerged across the final three interviews” is a defensible, specific saturation claim; “saturation was reached” on its own is not.
Tracking saturation in Byleron QDA: because codes stay linked to the documents they came from, you can watch your codebook’s growth rate directly — seeing exactly how many new codes each additional document adds makes a saturation claim something you can actually show, not just assert.
References
- (2020). Theoretical Sampling. SAGE Research Methods Foundations. https://doi.org/10.4135/9781526421036785176
- Ahmad, M., & Wilkins, S. (2024). Purposive sampling in qualitative research: a framework for the entire journey. Quality & Quantity, 59, 1461–1479. https://doi.org/10.1007/s11135-024-02022-5
- Bouncken, R., Czakon, W., & Schmitt, F. (2025). Purposeful sampling and saturation in qualitative research methodologies: recommendations and review. Review of Managerial Science, 20, 579–615. https://doi.org/10.1007/s11846-025-00881-2
- DeJonckheere, M., & Vaughn, L. (2019). Semistructured interviewing in primary care research: a balance of relationship and rigour. Family Medicine and Community Health, 7. https://doi.org/10.1136/fmch-2018-000057
- Guest, G., Bunce, A., & Johnson, L. (2006). How Many Interviews Are Enough? Field Methods, 18, 59–82. https://doi.org/10.1177/1525822×05279903
- Hennink, M., & Kaiser, B. (2021). Sample sizes for saturation in qualitative research: A systematic review of empirical tests. Social Science & Medicine, 114523. https://doi.org/10.1016/j.socscimed.2021.114523