There is no formula. Published empirical tests converge on a narrow range: Hennink and Kaiser’s 2022 systematic review found studies reached saturation within 9 to 17 interviews, or 4 to 8 focus group discussions, particularly where the population was homogenous and the objectives narrow. What a Philippine panel actually assesses is your justification, not your number.
Why can’t I just use Slovin’s formula?
Because it answers a question your study is not asking.
Slovin’s formula estimates how many respondents you need in order to generalise a measurement to a whole population within a stated margin of error. That is a quantitative goal, and it depends on having a population size, a numeric variable and a sampling frame.
A qualitative study is not trying to generalise a measurement. It is trying to understand how something is experienced, why it happens, or what it means to the people it happens to. Depth and range matter; representativeness in the statistical sense does not. Applying a formula built for the other goal produces a number that looks rigorous and justifies nothing — and a panellist who knows the difference will ask you what the margin of error means for a lived experience.
If you are still deciding which design your study needs, that decision comes before this one and is set out in the comparison of quantitative and qualitative research.
What is data saturation?
Saturation is the point at which collecting more data stops producing anything new — no new codes, no new themes, no new information relevant to your research questions.
It is a stopping rule, not a target. That distinction matters more than it sounds, because it has a practical consequence most undergraduate groups miss: you cannot observe saturation if you collect all your interviews first and analyse them afterwards. Saturation is only visible if you transcribe and code as you go, so you can see the new codes tapering off.
A group that runs fifteen interviews in one week and codes them the following month has not reached saturation. It has run fifteen interviews and hopes that was enough.
What does the published evidence actually say?
Two studies are cited more than any others on this question, and both are worth citing by name in your Chapter 3 rather than repeating a number you saw in a batchmate’s thesis.
Guest, Bunce and Johnson (2006) analysed 60 in-depth interviews with women in two West African countries and documented, interview by interview, when new themes stopped appearing. They reported that saturation occurred within the first twelve interviews, although the basic elements for metathemes were present as early as six.
Hennink and Kaiser (2022) conducted a systematic review of studies that had empirically tested saturation, identifying 23 articles — 17 using empirical data and 6 using statistical modelling. Their finding: studies reached saturation within a narrow range of 9 to 17 interviews, or 4 to 8 focus group discussions, particularly those with relatively homogenous study populations and narrowly defined objectives. They also note the exceptions — multi-country research and meta-theme analysis needed larger samples.
| Source | Method | Finding |
|---|---|---|
| Guest, Bunce & Johnson (2006) | 60 in-depth interviews, two West African countries | Saturation within the first 12 interviews; metatheme elements by 6 |
| Hennink & Kaiser (2022) | Systematic review of 23 empirical tests | 9–17 interviews or 4–8 focus group discussions |
Read those two rows together and a defensible planning figure for a typical Philippine undergraduate qualitative study — one site, one fairly homogenous group, two or three research questions — is ten to fifteen participants, with a stated intention to continue until saturation. That is a planning figure, not a promise, and the next section explains why the difference matters.
Is saturation a settled idea?
No, and saying so in your manuscript is a strength rather than a weakness.
Braun and Clarke, whose approach to thematic analysis many Philippine theses cite, published a direct challenge to the concept — “To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales” — arguing that saturation sits awkwardly with an interpretive approach in which meaning is generated by the analyst rather than discovered in a fixed quantity in the data.
You do not have to resolve that debate. What you should do is be consistent: if your Chapter 3 says you used reflexive thematic analysis, do not also claim you stopped because the data were exhausted. Say instead that you continued until additional interviews were no longer generating material relevant to your research questions, and that your sample size reflects the depth of engagement your questions required.
How do I write the justification into Chapter 3?
Four sentences, in this order. This is the section a panel reads, and a filled-in version is more useful than a rule.
Fifteen participants were recruited through purposive sampling, using the criteria stated above. Sample size was determined by data saturation rather than by a statistical formula, following the empirical guidance that studies with homogenous populations and narrowly defined objectives commonly reach saturation between nine and seventeen interviews (Hennink & Kaiser, 2022). Transcription and coding were conducted concurrently with data collection so that the emergence of new codes could be monitored. No new codes were identified after the thirteenth interview; two further interviews were conducted to confirm this, and recruitment then closed.
That paragraph does four things a bare number cannot: it names the sampling method, it states the basis for the size, it explains how saturation was observable, and it gives the point at which it occurred plus the confirmation interviews. Where this paragraph sits within the chapter, and what the surrounding sections must contain, is set out in the guide to writing Chapter 3 section by section.
How do I actually track saturation?
Keep a table. It costs you a few minutes per interview and it is the single most convincing exhibit you can bring to a defense.
- Transcribe each interview before the next one, or at minimum within a day.
- Code it and list every code. For the first interview, all codes are new.
- Log two numbers per interview: total codes so far, and new codes from this interview.
- Watch the second column. It falls steeply, then flattens, then reaches zero.
- Run two more interviews after your first zero before you stop. One zero can be an unusually similar participant.
Put that table in your appendix. When a panellist asks how you knew you had enough, you turn to it. This is the qualitative equivalent of showing your Slovin computation, and it carries the same weight.
What if my adviser insists on a specific number?
Use it, and justify it anyway.
Many Philippine research manuals and many advisers set an expected range for undergraduate qualitative work — ten, twelve, fifteen. Your manual governs you, and arguing methodology with an adviser over a number is a poor use of the weeks you have left.
What you can always do is write the justification paragraph regardless. A thesis that says “fifteen participants were selected, in line with the department’s guidance, and coding was conducted concurrently; no new codes emerged after the thirteenth interview” satisfies both the manual and the panel. The number came from the department; the defensibility came from you.
Does the number differ by qualitative design?
Yes, and this is worth one sentence in your Chapter 3 because it shows you know your own design.
- Phenomenology works with small, deeply engaged samples — the point is the depth of each account, not the count.
- Case study is bounded by the case, so the unit is the case rather than the participant; you may interview several people about one case.
- Focus group designs count groups, not people. The empirical range above is 4 to 8 discussions, and a group of eight participants is one discussion, not eight interviews.
- Grounded theory uses theoretical sampling — who you recruit next is decided by what the analysis so far suggests you still need, which makes the final number genuinely unknowable in advance.
Whichever you use, name it, and make your sampling logic match it. A manuscript claiming grounded theory with a fixed sample recruited in one week has a mismatch a panel will find.
What will the panel ask about my sample?
Four questions, reliably. Rehearse the answers before you build your deck — the deck itself is covered in the guide to preparing your defense presentation slide by slide.
- “Why fifteen?” Answer with your stopping rule and your saturation table, not with a citation alone.
- “How did you know you reached saturation?” This is the question the tracking table exists for.
- “How were participants selected?” Name the sampling technique and the inclusion criteria. Never call purposive sampling random.
- “Can your findings be generalised?” No, and say so plainly. Qualitative findings are transferable to similar contexts, not generalisable to a population, and claiming otherwise is the fastest way to lose the room.
Write the justification, not just the number
Tesify structures your thesis chapter by chapter in the format your research manual expects and keeps every citation attached to a source you actually opened, so a methodology section that has to justify a sample size reads like a decision rather than an assertion. Over 9,000 students have used it to write more than 15,000 chapters, and every word is still written by you.
Frequently asked questions
How many respondents do you need for a qualitative thesis?
There is no formula. Published empirical tests converge on 9 to 17 interviews or 4 to 8 focus group discussions for studies with homogenous populations and narrow objectives. Ten to fifteen participants is a defensible planning figure for a typical Philippine undergraduate study, provided you state how you determined it.
Can I use Slovin’s formula for a qualitative study?
No. Slovin’s estimates the sample needed to generalise a measurement within a margin of error, which is a quantitative goal. A qualitative study seeks depth and range of meaning, so the formula answers a question your research is not asking.
What is data saturation in simple terms?
The point where new interviews stop giving you anything new — no new codes, no new themes. It is a rule for stopping, not a number to hit, and it can only be seen if you analyse while you collect.
How many interviews is too few?
Fewer than about eight becomes hard to defend for an interview study with more than one research question, because you have not given saturation a chance to appear. If circumstances force a smaller sample, say so as a stated limitation rather than leaving the panel to notice.
How many focus groups do I need?
Four to eight discussions, on the systematic review evidence. Count discussions, not participants — one group of eight people is one focus group discussion.
Do I need to reach saturation to pass?
Not necessarily. Many undergraduate timelines make true saturation impractical, and a manuscript that states honestly that recruitment closed at the timetable rather than at saturation, and lists that as a limitation, is stronger than one that claims saturation it cannot evidence.
Should I state my sample size in the abstract?
Yes, and it must match Chapter 3, Chapter 4 and any table that counts participants. A number that drifts between sections is one of the most frequently caught defects in Philippine manuscripts.
Can I mix interviews and focus groups?
You can, and it is common, but treat them as two data sources with two stopping rules and say how they relate. Do not add nine interviews to five focus groups and report fourteen of anything.
Does my qualitative study need statistical treatment?
Not in the inferential sense. You describe the analytic approach — thematic analysis, content analysis — rather than naming statistical tests. What belongs in that section for either design is covered in the guide to the statistical treatment of data.
Where in the five chapters does this go?
The sample size and its justification belong in Chapter 3, and the number reappears in Chapter 4 whenever you describe your participants. The overall arrangement is set out in the overview of Chapters 1 to 5.
Should I cite these studies even though they are not Filipino?
Yes. Methodological guidance is cited for its method, not its locale, and both of these are standard references in the international literature. Pair them with any local study whose design resembles yours, using the synthesis method in the guide to writing a review of related literature.