Interpretative Phenomenological Analysis

What IPA is

Interpretative Phenomenological Analysis (IPA) is a qualitative method developed by Jonathan Smith in the mid-1990s for the close study of how individuals make sense of major life experiences. Its theoretical commitments are three: phenomenology (a concern with lived experience), hermeneutics (interpretation as the central analytical act), and idiography (a focus on the particular before the general). The combination gives IPA its distinctive small-sample, deep-reading character (Smith, 1996; Smith, Flowers & Larkin, 2022).

IPA is most at home in health psychology, counselling research, illness-narrative work, and other contexts where the research question is some version of “what is it like to live through X, and how do people make sense of it?” It is not the right method for questions about prevalence, about social processes, or about theory construction; chapters 2 and 3 of this guide cover better-fitting alternatives for those.

The double hermeneutic

The phrase Smith introduced — the double hermeneutic — is the most-cited and least understood part of IPA. It names the layered interpretation involved in IPA work: the participant is interpreting their own experience, and the researcher is interpreting the participant’s interpretation. The analyst is one step removed from the experience itself, working on the participant’s account of it.

The practical consequence is not philosophical decoration; it is methodological. Three things follow:

• Verbatim quotation is load-bearing. IPA papers quote heavily from transcripts because the participant’s words are the raw material of the first hermeneutic. Paraphrase loses evidence.
• The researcher’s interpretation is explicit, not hidden. A good IPA analysis announces its interpretive moves — “what Maya is doing here is locating the disruption in the body rather than in the role” — and lets the reader judge them against the transcript.
• Positionality matters more than in some other methods. The researcher’s relationship to the phenomenon shapes the second hermeneutic. IPA methods sections routinely include a positionality statement, and the statement is not pro forma.

The four-step procedure

Smith, Flowers and Larkin (2022, chapter 5) describe IPA analysis as a six-step procedure; the working summary most analysts actually use is the four-step version. The phases apply to one transcript at a time before cross-case work begins.

1. Reading and re-reading

Read the transcript multiple times before analysing. Listen to the audio if you have it. The aim is to let the participant’s account become familiar without yet imposing analytical structure. This phase mirrors the “familiarisation” step in thematic analysis but is more pronounced; IPA treats this reading as part of the first hermeneutic.

2. Exploratory comments

Annotate the transcript line by line with three layers of comment: descriptive (what is the participant saying), linguistic (how are they saying it — metaphor, pronoun use, repetition, pauses), and conceptual (what does the analyst notice at a more interpretive register). The three layers can be done in passes or interleaved; the output is a transcript with dense margin notes.

3. Emergent themes

Group exploratory comments into emergent themes for this case. Themes at this stage are still case-specific; the goal is not yet cross-case synthesis but a coherent map of how this participant makes sense of the phenomenon. Smith, Flowers and Larkin suggest 8–15 emergent themes per case as a working range.

4. Connecting themes across cases

After 4–10 cases have been analysed individually, group-level analysis identifies patterns across participants — convergence (where participants share an interpretive move), divergence (where they diverge in instructive ways), and superordinate themes that organise the case-level themes into a higher-order structure. The final write-up moves back and forth between superordinate themes and individual exemplars.

Why small samples

IPA studies typically use 4–10 participants. This is a design choice grounded in the idiographic commitment, not a pragmatic compromise. The argument: deep interpretive engagement with each case is incompatible with large samples; doubling N halves the analytic attention available per participant, and that attention is what distinguishes IPA from a thinner thematic analysis.

Smith, Flowers and Larkin (2022) suggest specific numbers: undergraduate or master’s projects, 3 participants; doctoral theses or papers, 4–10 in homogeneous samples; up to 15 only when the comparison structure of the design warrants it. Larger samples push IPA toward thematic analysis.

When IPA fits (and when it doesn’t)

IPA is the right method when the research question is about meaning-making in relation to a significant experience, when the sample can be reasonably homogeneous around that experience, and when the timeline allows for deep per-case analysis.

Three signals that IPA is not the right method:

• The question is about social process or context. “How do hospital teams negotiate handover practices?” is a grounded-theory question, not an IPA question. IPA’s unit of analysis is the individual’s sense-making, not the collective practice.
• The sample is heterogeneous in ways that matter. IPA assumes participants share enough of the phenomenon for cross-case synthesis to be meaningful. A study of “the experience of chronic illness” across 10 different conditions is too heterogeneous; a study of “the experience of receiving a Type 1 diabetes diagnosis in late adolescence” is the kind of focused sample IPA expects.
• The data are not narrative. IPA works on first-person accounts — semi-structured interviews, focus groups (with caveats), diaries, written reflections. Survey free-text responses are usually too thin for the line-by-line work IPA requires.

Common mistakes in published IPA

Larkin, Watts and Clifton (2006) and Pietkiewicz and Smith (2014) catalogue the most-frequent errors in published IPA. The three worth flagging here:

Thin interpretation. Many studies labelled IPA produce what is effectively a descriptive thematic analysis with verbatim quotes — the participant’s first hermeneutic is reported, but the analyst’s second hermeneutic is missing. The give-away in writing is themes that read as topic labels (“experience of disclosure”) rather than interpretive claims (“disclosure as managed risk”).

Over-claiming the sample. A four-participant IPA study cannot warrant claims about “people with X.” The findings are about how these four participants make sense of X. That is a real finding; treating it as a population-level claim is a methodological error reviewers will catch.

Skipping the idiographic stage. Studies that go straight to cross-case themes without per-case analysis are not doing IPA. The case-level analysis is what licenses the cross-case work; doing one without the other produces a paper that looks like IPA but reads like thematic analysis.

Further reading

• Smith, J. A., Flowers, P., & Larkin, M. (2022). Interpretative Phenomenological Analysis: Theory, Method and Research (2nd ed.). SAGE. The contemporary reference.
• Smith, J. A. (1996). Beyond the divide between cognition and discourse: Using interpretative phenomenological analysis in health psychology. Psychology & Health, 11(2), 261–271.
• Larkin, M., Watts, S., & Clifton, E. (2006). Giving voice and making sense in interpretative phenomenological analysis. Qualitative Research in Psychology, 3(2), 102–120.
• Pietkiewicz, I., & Smith, J. A. (2014). A practical guide to using interpretative phenomenological analysis in qualitative research psychology. Psychological Journal, 20(1), 7–14.
• Larkin, M., & Thompson, A. R. (2012). Interpretative phenomenological analysis in mental health and psychotherapy research. In Qualitative Research Methods in Mental Health and Psychotherapy (pp. 99–116). Wiley-Blackwell.

This chapter is in draft. It has not yet been peer-reviewed by an external methodologist. Reviewer contact: [email protected].