How Many Interviews Do You Need for a Qualitative Study? There Is No Magic Number
There is no universal minimum number of qualitative interviews. The sample needed depends on the study aim, participant specificity and heterogeneity, methodological approach, quality and depth of interviews, analytic strategy and the type of understanding sought. Studies showing saturation around 9, 12 or 20 interviews are empirical examples from particular datasets, not rules that can be copied into every thesis.
“How many interviews should I do?”
Twelve.
Fifteen.
Twenty.
You will hear all three numbers with remarkable confidence.
The correct answer is less satisfying:
It depends on the information your study needs and how your methodology defines adequate sampling.
That is not an excuse to avoid planning.
It means the plan needs reasoning.
Why does everyone say 12 interviews?
A frequently cited 2006 study by Guest, Bunce and Johnson examined 60 interviews with women in two West African countries and documented the development of thematic saturation.
That study became influential.
But it was an empirical study of a particular:
- topic;
- relatively homogeneous participant group;
- interview approach;
- analytic process.
It did not establish:
“All qualitative studies require 12 interviews.”
A finding from one dataset is evidence.
It is not legislation.
A published sample size can inform your reasoning. It cannot substitute for it.
What did later research show?
Hennink, Kaiser and Marconi examined 25 in-depth interviews and distinguished:
Code saturation
The point at which the range of issues or codes had largely been identified.
In their dataset, this occurred around nine interviews.
Meaning saturation
The point at which the analysis had developed richer understanding, nuance and dimensions of those issues.
That required approximately 16-24 interviews in their study.
The lesson is important:
Hearing the main topics is not the same as understanding them deeply.
So a study seeking a broad list of issues may need a different information depth from a study trying to understand mechanisms, variation and meaning.
Information power: a more useful planning idea
Malterud, Siersma and Guassora proposed the concept of information power.
The basic idea is:
The more information relevant to the study that each participant contributes, the fewer participants may be needed.
They identify factors including:
- study aim;
- sample specificity;
- use of established theory;
- quality of interview dialogue;
- analysis strategy.
This is not a calculator.
It is a framework for reasoning.
Factor 1: how broad is your study aim?
Broad aim:
Explore barriers to healthcare among adults in Uganda.
That covers:
- many populations;
- many services;
- many barriers;
- many contexts.
A small interview sample is unlikely to capture enough variation.
Narrow aim:
Explore how women who discontinued injectable contraception because of side effects describe the decision to discontinue.
Much more focused.
Participants share a specific relevant experience.
The sample may have greater information power for that question.
Factor 2: how heterogeneous is your sample?
Suppose you plan 15 interviews across:
- adolescents;
- adult women;
- men;
- health workers;
- community leaders;
- urban and rural settings.
That is not really “15 interviews.”
It is a handful of people representing several very different perspectives.
If subgroup comparison matters, sample adequacy should be considered within those analytical groups.
A total number can hide thin coverage.
You wrote:
“Fifteen interviews will be conducted.”
But you have five participant categories.
Three per category.
And the analysis plans to compare all five.
The issue is not whether 15 is “allowed.”
The issue is whether the sample can support the comparisons you intend to make.
Factor 3: how specific is the sample?
A highly specific purposive sample may provide concentrated information.
For example:
mothers who discontinued a contraceptive implant within six months because of side effects.
That group has direct experience of the phenomenon.
Compare:
all women aged 15-49.
The second population contains much more variation relative to the focused question.
Specificity can increase information power.
But highly specific sampling can also narrow transferability.
That may be appropriate.
Be explicit.
Factor 4: how rich will the interviews be?
A 15-minute interview with:
- five closed questions;
- one optional free-text response;
does not provide the same analytic depth as a 60-minute interview probing:
- experience;
- meaning;
- sequence;
- contradictions;
- social context.
Sample adequacy depends partly on the quality of dialogue and data generated.
You cannot fix shallow interviews by recruiting 100 people.
Factor 5: what analysis are you doing?
A cross-case thematic analysis may require different sampling logic from:
- narrative analysis;
- phenomenology;
- grounded theory;
- case study;
- framework analysis;
- discourse analysis.
Methodological traditions also use “saturation” differently.
Do not take sample-size language from grounded theory and attach it to every qualitative study.
planning a thesis study
Question:
How do women who discontinued injectable contraception because of side effects describe the decision to stop?
Suppose:
- aim is focused;
- sample is purposively selected for direct experience;
- one participant group;
- interviews are expected to be in-depth;
- analysis seeks common patterns and variation;
- the researcher expects some variation by age and parity.
A reasonable proposal might set an initial recruitment target such as:
15-20 participants
But the justification should not be:
“Qualitative studies require 15-20 participants.”
Instead explain the reasoning.
For example:
The study has a focused aim and will purposively recruit women with direct experience of discontinuing injectable contraception following side effects. We will initially recruit approximately 15-20 participants, while assessing sample adequacy iteratively in relation to the depth and variation of information generated and the development of the analysis.
The number is now a planning estimate.
Not a magic minimum.
What if my institution requires a fixed number before ethics approval?
Very common.
You still need a planned number or range.
The solution is not to write:
“Until saturation.”
and leave the sample size blank.
Instead:
- specify an initial recruitment target or range;
- justify it using the study aim and expected information power;
- explain whether and how adequacy will be reviewed during data collection and analysis;
- remain within approved recruitment limits unless amendments are obtained where necessary.
Ethics and budgeting need a plan.
Qualitative methodology still allows iterative judgement.
What do I actually write in my proposal?
Weak:
Fifteen participants will be interviewed because qualitative research requires 10-20 participants.
Better:
The study has a focused aim and will purposively recruit participants with direct experience of [phenomenon]. An initial sample of approximately 15-20 interviews is planned to capture variation across [relevant dimensions]. Sample adequacy will be assessed iteratively during analysis, considering the depth and diversity of information generated and whether further interviews contribute substantively new understanding.
Important:
This wording does not fit every methodology.
If your approach is grounded theory, phenomenology, case study or another tradition with its own sampling logic, adapt it.
Can I calculate qualitative sample size with a formula?
Usually not in the same way as probability-based quantitative estimation.
Qualitative sampling is not designed to estimate prevalence with a margin of error.
A Cochran formula does not become more rigorous simply because you apply it to interview recruitment.
Different question.
Different inference.
Different sample-size logic.
Should I stop the moment no new code appears?
Not necessarily.
Our guide on saturation goes deeper into this.
Hennink's work shows why identifying the range of codes can occur before the researcher has developed rich understanding of those codes.
A final interview may not create a brand-new theme.
It may still:
- add nuance;
- identify conditions;
- reveal exceptions;
- deepen interpretation;
- show subgroup differences.
“No new code” is not the only possible definition of adequate data.
Write your proposed qualitative sample size.
Under it, answer:
- How focused is my aim?
- How similar or heterogeneous are participants?
- How many analytical subgroups matter?
- How deep will interviews be?
- What methodology am I using?
- What type of understanding do I need?
- How will I assess whether additional interviews still add value?
If your only answer is:
“Guest et al. said 12”
you have cited a paper.
You have not yet justified your sample.
Frequently asked questions
- Are 12 interviews enough for qualitative research?
Sometimes they may be. Sometimes they may be far too few. Guest et al.'s findings should not be treated as a universal threshold.
- Is 20 interviews enough for a Master's thesis?
The degree level does not determine adequacy. The research question, sample, methodology and analytic depth do.
- Should I specify a range?
A justified initial range can be useful, particularly when iterative assessment of adequacy is planned.
- Does purposive sampling need a sample-size calculation?
It needs a sample-size justification, but usually not a probability-estimation formula.
- What is information power?
A framework proposing that sample adequacy depends on how much relevant information the sample provides for the particular study.
Use a Methods Bench Qualitative Sample Planning Worksheet:
- Aim breadth
- Sample specificity
- Heterogeneity
- Key subgroups
- Expected interview depth
- Analytic approach
- Initial target
- Evidence that would justify continuing/stopping recruitment
References and further reading
- 1.Guest G, Bunce A, Johnson L. How Many Interviews Are Enough? An Experiment with Data Saturation and Variability. Field Methods. 2006;18(1):59-82. doi:10.1177/1525822X05279903.
- 2.Hennink MM, Kaiser BN, Marconi VC. Code Saturation Versus Meaning Saturation: How Many Interviews Are Enough? Qualitative Health Research. 2017;27(4):591-608. doi:10.1177/1049732316665344.
- 3.Malterud K, Siersma VD, Guassora AD. Sample Size in Qualitative Interview Studies: Guided by Information Power. Qualitative Health Research. 2016;26(13):1753-1760. doi:10.1177/1049732315617444.
- 4.COREQ reporting guideline, EQUATOR Network.
Take it further
Better research, once a week.
Practical methods guidance, research tools and funding opportunities from Methods Bench.
Get practical research notes and new opportunities
Short, useful emails for researchers in Uganda and East Africa. No spam.