03 · What You Need to Know
Understanding and Assessing Saturation in Qualitative Sampling
What Does Data Saturation Actually Mean?
Data saturation is commonly used to describe a stage in qualitative inquiry at which collecting additional material no longer contributes sufficiently new information to the analysis.
However, this description leaves an important question unanswered: new information about what?
Researchers may be concerned with identifying additional topics, developing richer interpretations, elaborating conceptual categories, or understanding variation across participants.
These are different analytical objectives. A study may identify most recurring topics relatively early while still requiring additional interviews to understand how participants interpret those topics or why their experiences differ.
Consequently, saturation should be defined in relation to the specific analytical purpose rather than treated as a self-explanatory stopping point.
What Is the Difference Between Code Saturation and Meaning Saturation?
Hennink, Kaiser, and Marconi (2017) distinguished code saturation from meaning saturation in qualitative interview research.
Code saturation concerns identifying the range of issues or codes relevant to the analysis. Meaning saturation concerns developing a sufficiently rich understanding of those issues, including their dimensions, nuances, and contextual variation.
| Consideration |
Code saturation |
Meaning saturation |
| Primary concern |
Identifying relevant issues or codes. |
Understanding the depth and variation of those issues. |
| Possible indication |
Additional interviews contribute few or no new relevant codes. |
Additional interviews contribute little further nuance or contextual understanding. |
| Analytical emphasis |
Breadth of identified content. |
Richness and complexity of interpretation. |
| Important limitation |
Repeated codes may conceal insufficient understanding. |
Adequacy requires substantive judgment and cannot be reduced to code counts alone. |
A researcher might identify recurring concerns about assessment integrity after several interviews but still lack sufficient understanding of how teachers negotiate those concerns when designing different assessment types.
In that situation, code saturation may appear plausible while meaning saturation remains uncertain.
What Is Theoretical Saturation?
Theoretical saturation is associated particularly with grounded theory traditions.
It concerns the development of conceptual categories and their relationships through iterative analysis and theoretical sampling.
Rather than asking only whether new topics are appearing, the researcher considers whether further data are needed to elaborate the properties, variation, and connections of emerging categories.
For example, a developing explanation of how teachers negotiate institutional AI policies might require additional interviews with teachers who have experienced conflicting expectations.
Theoretical saturation should not be equated automatically with code saturation. It involves a different analytical purpose and depends on the grounded theory approach being used.
Does Saturation Mean That Nothing New Can Ever Be Found?
No. Qualitative inquiry cannot ordinarily establish that no additional participant could ever contribute a new perspective.
Experiences are shaped by context, history, participant characteristics, and the questions asked. Further interviews may always reveal unusual cases or alternative interpretations.
A defensible saturation claim is therefore bounded by the study's research aim, sample, analytical framework, and operational definition of new information.
It is better understood as a reasoned judgment about diminishing analytical contribution than as proof that the phenomenon has been exhaustively documented.
How Does Saturation Influence Initial Sample-Size Planning?
Because saturation is ordinarily evaluated during data collection and analysis, it cannot be known with certainty before the study begins.
Researchers may nevertheless establish a provisional recruitment target or range based on the research aim, participant specificity, methodological literature, and anticipated analytical requirements.
For example, a researcher might initially plan fifteen to twenty interviews while specifying how the developing analysis will be reviewed for additional relevant information.
That range is an illustrative planning decision, not a universal recommendation or a guarantee that saturation will occur within it.
The broader task of determining an appropriate qualitative sample size requires attention to more than saturation alone.
What Does Empirical Research Say About the Number of Interviews Needed?
Guest, Bunce, and Johnson (2006) investigated thematic saturation using a particular interview dataset and found that major recurring themes appeared relatively early in their analysis.
Hennink, Kaiser, and Marconi (2017) demonstrated that identifying codes and developing their meanings may require different amounts of data.
Hennink and Kaiser (2022) subsequently synthesized empirical studies examining saturation and found that sample requirements varied with research conditions and the form of saturation assessed.
These findings provide useful evidence for planning, but they do not establish that six, twelve, or any other fixed number of interviews is sufficient across qualitative studies.
Differences in participant diversity, research scope, interview depth, and analytical purpose limit direct transfer of reported saturation counts.
Can You Assess Saturation Using a Structured Procedure?
Yes, when the study's analytical approach supports a sufficiently explicit definition of new information.
Guest, Namey, and Chen (2020) proposed a systematic approach for assessing thematic saturation using a base set of interviews, a subsequent run of interviews, and a threshold for the amount of new information.
The procedure can make decisions more transparent in studies using compatible coding approaches.
However, its numerical thresholds are methodological choices, not universal standards. The procedure should not be transferred automatically to interpretive approaches that reject the assumption of a finite inventory of themes.
How Can a Saturation Tracking Matrix Help?
Researchers may maintain a structured record of how additional interviews contribute to the analysis.
The matrix can document newly identified codes, elaborations of existing concepts, contrasting accounts, and unresolved questions.
| Interview batch |
New analytical contribution |
Remaining uncertainty |
| Interviews 1–5 |
Initial codes concerning assessment design, workload, and academic integrity. |
Limited understanding of disciplinary differences. |
| Interviews 6–10 |
Further examples and contrasting experiences across assessment types. |
Unclear how teachers negotiate conflicting institutional expectations. |
| Interviews 11–15 |
Additional explanation of policy interpretation and contextual variation. |
Some questions remain about teachers with limited institutional support. |
| Interviews 16–18 |
Primarily elaboration of established categories, with limited additional variation. |
Assess whether the material is sufficient for the stated analytical objective. |
This table is hypothetical and illustrates documentation, not a recommended saturation sequence.
The researcher must interpret the significance of additional information rather than simply count new codes. A single contrasting case may be analytically important even when it introduces no entirely new code label.
Why Is Concurrent Data Collection and Analysis Important?
When saturation guides recruitment, researchers need to examine data while collection is ongoing.
If all interviews are completed before any analysis begins, saturation cannot meaningfully inform decisions about whether additional participants are required during that collection period.
Concurrent analysis may involve reviewing transcripts, developing provisional codes, writing analytical memos, and identifying gaps that subsequent interviews could address.
This does not require a fully completed analysis after every interview. It does require enough analytical engagement to make informed recruitment decisions.
Should Saturation Be Assessed Across the Entire Sample or Within Subgroups?
The answer depends on the research question.
If the study aims to understand one relatively specific participant group, an overall adequacy assessment may be appropriate.
If the research aims to compare distinct groups, such as teachers from different disciplines or institutions, overall repetition may conceal insufficient information within particular groups.
For example, repeated accounts from engineering teachers do not necessarily establish adequate understanding of experiences among humanities teachers.
Researchers should therefore consider whether separate analytical adequacy assessments are needed for important subgroups.
Such assessments should follow the intended comparisons rather than mechanically require identical interview counts in every category.
How Do Negative Cases Affect Saturation?
Negative or discrepant cases challenge emerging patterns and may reveal important limitations in an interpretation.
Suppose most teachers describe generative AI as increasing assessment-design efficiency, but one participant reports that it substantially increased verification workload.
That account should not be dismissed merely because it differs from an otherwise recurring pattern.
Researchers may need to examine the circumstances producing the difference, including institutional expectations, assessment complexity, or professional experience.
A saturation decision should not be based solely on the frequency of dominant accounts while ignoring meaningful contradictions.
Does Repetition of Interview Responses Prove Saturation?
No. Repetition may indicate that certain topics are common, but it does not necessarily establish sufficient analytical depth.
Participants may repeat broad statements because interview questions are too general or leading.
For example, asking every teacher whether AI creates ethical concerns may repeatedly produce affirmative answers without revealing how those concerns influence actual instructional decisions.
Improving probing and asking for concrete examples may generate richer information without necessarily increasing participant numbers substantially.
The relationship between interview quality and sample size is therefore relevant to saturation assessment.
Can Saturation Be Claimed After Very Few Interviews?
Possibly, under narrowly defined conditions, but such a claim requires careful justification.
A focused study involving highly specific participants and a limited analytical objective may identify recurring information quickly.
However, an early absence of new codes could reflect restricted recruitment, repetitive questioning, or superficial analysis.
The researcher should examine whether the material is sufficiently rich, whether relevant variation has been explored, and whether additional participants might address unresolved questions.
The methodological concerns surrounding saturation claims after very few interviews deserve particular attention when conclusions extend beyond a narrowly defined sample.
Is Saturation Appropriate for Reflexive Thematic Analysis?
Its use is contested and may be inconsistent with the assumptions of reflexive thematic analysis as developed by Braun and Clarke.
In that approach, themes are actively constructed through the researcher's interpretive engagement with data rather than treated as a fixed collection of entities waiting to be discovered.
The claim that no new themes can emerge may therefore conflict with the analytical orientation.
Researchers using reflexive thematic analysis should consider justifications based on research purpose, sample relevance, analytical depth, and methodological coherence instead of automatically invoking saturation.
The broader question of whether saturation is appropriate for a particular qualitative methodology should be addressed before using it as a stopping criterion.
How Does Information Power Differ From Saturation?
Information power, proposed by Malterud, Siersma, and Guassora (2016), considers how much relevant information a sample provides in relation to the study's aim.
It emphasizes the interaction among study aim, sample specificity, established theory, dialogue quality, and analysis strategy.
Unlike a simple code-saturation approach, information power does not depend on demonstrating that no new codes appear.
It may be useful when sample adequacy is better understood through the relevance and depth of information than through the accumulation of identifiable categories.
However, information power is not an exact sample-size formula, and its suitability still depends on methodological fit.
When Should Recruitment Stop?
When saturation is an appropriate criterion, recruitment may stop once the specified form of saturation has been supported by a sufficiently transparent analytical assessment.
The researcher should consider whether additional interviews are likely to contribute information consequential to the research question, including relevant variation and unresolved conceptual relationships.
Practical constraints may also influence recruitment, but they should be distinguished from analytical adequacy.
If recruitment ends because funding, time, or access has been exhausted, the report should not automatically describe that decision as saturation.
The decision about when qualitative participant recruitment should stop should be documented honestly and consistently with the chosen methodology.
How Should Saturation Be Reported in a Research Article?
Reports should explain the form of saturation assessed, the analytical procedure used, and the evidence supporting the decision.
For example, a study might describe how interviews were analyzed in successive batches and how the contribution of later interviews was evaluated.
If a structured threshold was used, its definition and rationale should be reported.
Researchers should avoid unsupported statements such as "data saturation was achieved after fifteen interviews" without describing how that conclusion was reached.
Transparency allows readers to assess the adequacy of the sampling process rather than simply accept the author's declaration.
