Most qualitative studies use between 6 and 30 participants, but the right number depends on your methodology, your research question, and how similar your participants are to one another. There is no single correct answer, and that ambiguity frustrates many researchers. The real goal isn’t hitting a magic number. It’s collecting enough data that new interviews stop producing new insights.
The Concept That Replaces a Fixed Number
In quantitative research, you calculate a sample size with a formula. In qualitative research, the guiding principle is saturation: the point at which additional interviews or observations become redundant. As one early description put it, “when the researcher begins to hear the same comments again and again, data saturation is being reached.” Saturation has become the gold standard for determining sample size in qualitative health and social science research, and most review committees expect you to reference it when justifying your participant count.
There are slightly different versions of this idea depending on your approach. In grounded theory, the term is “theoretical saturation,” meaning the categories in your emerging theory are fully developed and no new properties appear. In interview-based studies using thematic analysis, the focus is on “data saturation,” meaning no new themes or codes are surfacing. The practical effect is the same: you keep collecting data until the returns diminish to near zero.
Ranges by Methodology
Different qualitative approaches call for different levels of depth per participant, which directly affects how many you need. Here are the commonly cited ranges:
- Phenomenology and IPA: These methods explore lived experience in fine detail. Sample sizes are intentionally small, often 3 to 10 participants, because each interview generates dense, layered data that takes significant time to analyze.
- Grounded theory: Because the goal is building a theoretical framework, you need enough participants to develop and test your emerging categories. Recommendations typically fall between 20 and 35, though some sources cite ranges as low as 5 when the population is narrow and well-defined.
- Ethnography: Academic ethnographic studies typically involve 6 to 10 participants for formal interviews, though the researcher also draws on extensive field observation. The participant count for interviews is smaller because the method relies heavily on watching daily interactions and cultural activities firsthand.
- Case study research: A single-case design might involve 4 to 30 participants within that case, depending on the complexity. Multiple-case designs multiply accordingly, but each case is treated as its own unit of analysis.
- Focus groups: The ideal group size is 8 to 10 people per session. Research on focus group saturation found that two to three groups are enough to capture about 80% of themes in a homogeneous population, and three to six groups capture around 90%. By the fourth focus group session, studies have documented over 90% of thematic codes already identified.
Five Factors That Push the Number Up or Down
A widely cited model proposes that your required sample size depends on the “information power” your participants hold. The more relevant information each person can provide, the fewer people you need. Five specific factors shape this:
How broad your research aim is. A narrow, focused question (how do nurses in one ICU experience moral distress during end-of-life care) requires fewer participants than a broad exploratory question (how do healthcare workers experience workplace stress). Broad aims need more voices to cover the terrain.
How specific your sample is. If your participants share key characteristics relevant to the research question, you’ll reach saturation faster. A group of first-generation college students at one university will produce overlapping insights more quickly than a mixed sample spanning different institutions, ages, and backgrounds.
Whether you’re using established theory. If your study builds on or tests an existing theoretical framework, you have a scaffold that guides your analysis. This means you can work with fewer participants because you’re looking for specific things rather than discovering everything from scratch.
The quality of your interviews. Skilled interviewers who build strong rapport and ask layered follow-up questions produce richer transcripts per session. A 90-minute conversation with thoughtful probing generates more usable data than a 30-minute interview that stays surface-level. Better dialogue means fewer interviews needed.
Your analysis strategy. Methods that require deep, case-by-case analysis (like interpretive phenomenology) work best with smaller samples. Methods that look for patterns across a larger dataset (like some forms of thematic analysis) can accommodate and often require more participants.
What the Empirical Evidence Actually Shows
One of the most frequently cited studies on this question analyzed 60 interviews and tracked when new codes appeared. The researchers found that the vast majority of codes emerged within the first 12 interviews, with basic themes identifiable even earlier. This study is often used to support a minimum of 12 interviews for studies aiming at thematic saturation in a relatively homogeneous group.
For focus groups specifically, a large-scale analysis of 40 group discussions found that averaging across different ordering scenarios, 90% of themes surfaced by the fourth to sixth group session. The most prevalent, high-frequency themes appeared even sooner, within two to three sessions. This suggests that for straightforward research questions with similar participants, you don’t need dozens of focus groups to reach solid thematic coverage.
Some researchers have proposed more structured approaches to sample size planning. One method developed for thematic analysis uses a calculation that factors in how common you expect a theme to be in your population, how many instances of that theme you want in your data, and the level of confidence you want in capturing it. This kind of tool can help when a review board asks you to justify your number before data collection begins, which is exactly the situation most researchers face.
How to Justify Your Sample Size
Review boards and dissertation committees typically ask two things: is your sample large enough to answer the research question, and is your justification clearly described? Saying “I’ll interview people until I reach saturation” is necessary but rarely sufficient on its own. You need to pair it with a specific estimate and explain why that estimate fits your study.
A strong justification usually includes three elements. First, reference the accepted range for your methodology (for example, 20 to 30 for grounded theory). Second, explain the information power of your sample, noting factors like population homogeneity or the focused nature of your research question that might push toward the lower or higher end. Third, describe your plan for recognizing saturation during data collection, such as coding each transcript before conducting the next interview and tracking whether new codes are still emerging.
If your proposed number falls below what’s typical for your method, the information power model gives you language to defend it. A highly specific sample, a narrow research question, a strong theoretical lens, and a deep analytic approach can all justify a smaller count. Conversely, if your topic is broad, your population is diverse, or your analysis is more descriptive, plan for more participants than the minimum.
Common Mistakes With Sample Size
The most frequent error is treating saturation as a checkbox rather than a process. Some researchers predetermine a number, collect exactly that many interviews, and then claim saturation without evidence. Review boards and experienced readers can spot this. If you say you reached saturation at 15 interviews, you should be able to show that your last few interviews produced no new codes or themes.
Another common problem is confusing repetition with saturation. Hearing the same stories doesn’t necessarily mean your categories are fully developed. In grounded theory specifically, theoretical saturation means the properties and dimensions of each category are complete, not just that participants are saying similar things. Asking the same questions of every participant will produce repetitive answers quickly, but that reflects the limits of your interview guide, not genuine saturation of the topic.
Finally, some researchers overrecruit out of anxiety, conducting 40 or 50 interviews when 20 would have been sufficient. This isn’t just inefficient. It creates an ethical issue: you’ve asked more people to give their time and share personal experiences than the study required. The goal is a sample large enough to provide sufficient data without involving more participants than necessary.