Interpretative Phenomenological Analysis (IPA): From Transcript to Group Experiential Themes

Interpretative Phenomenological Analysis (IPA) is a specific, procedurally structured approach to qualitative analysis, developed by Jonathan Smith and colleagues in psychology from the mid-1990s onward and now widely used across health psychology, clinical, and applied dissertation research. This guide covers the part most treatments skip: the actual analytic sequence, case by case, from a raw interview transcript to a cross-case set of group experiential themes (GETs). For IPA’s philosophical position relative to other phenomenological schools, see Phenomenological Research: Method, Approaches, and When to Use It. For the mechanics of a different, non-idiographic coding approach that IPA is often confused with, see Thematic Analysis: A Step-by-Step Guide to Braun and Clarke’s Six Phases.

Two commitments that separate IPA from generic qualitative coding

A study that interviews a handful of people and codes the transcripts for recurring ideas is not automatically doing IPA, even if it borrows the label. IPA adds two specific commitments that a generic coding pass does not require:

• Idiographic commitment. IPA starts with, and stays anchored to, the single case. Each transcript is analyzed in full, on its own terms, before any comparison across participants is attempted. This is the opposite analytic direction from approaches that code across the whole dataset from the outset — IPA deliberately delays cross-case work until each individual account has been understood in its own right.
• The double hermeneutic. The participant is already making sense of their own experience by describing it; the researcher is then making sense of the participant making sense of their experience. IPA’s analysis is explicitly interpretive on both levels, not just a descriptive summary of what participants said.

Both commitments follow from IPA’s small-sample logic: because each case gets this much individual attention, IPA studies are almost always small and purposively homogeneous — commonly 3 to 15 participants who share the specific experience under study, sometimes a single detailed case. See Purposive Sampling: Choosing Cases on Purpose for how to build and justify a sample like this in a methods section.

Data collection: usually one semi-structured interview per case

IPA is almost always paired with semi-structured, one-to-one interviews built around open questions that let participants talk at length in their own terms, with follow-up probes rather than a fixed script. Diaries and, less commonly, focus groups appear in the literature, but the semi-structured interview is the dominant instrument because it produces the kind of rich, first-person, reflective account the idiographic analysis that follows depends on. See Designing Semi-Structured Interviews: A Practical Guide for question design.

The IPA analytic sequence: transcript to group experiential themes

The sequence below follows the process set out by Smith, Flowers, and Larkin (2009) and updated by Smith and Nizza (2022) — see the terminology note further down. It is worked one case at a time; only the final step moves across cases.

1. Reading and re-reading

Before any coding, the researcher reads the full transcript multiple times, ideally alongside the original audio, to become immersed in that one participant’s account before imposing any analytic structure on it.

2. Exploratory noting

Close, line-by-line annotation of the transcript, conventionally worked in three registers: descriptive comments (what the participant said, in their own terms), linguistic comments (word choice, tone, pauses, metaphor, repetition — how something was said), and conceptual comments (more interpretive, often question-framed notes about what might underlie what the participant is expressing). This is the most time-intensive step and the one most often shortcut in practice; thin exploratory noting is the single most common reason a published “IPA” analysis reads as generic thematic coding wearing IPA’s name.

3. Constructing experiential statements and personal experiential themes (PETs)

The exploratory notes are distilled into concise experiential statements — still tied to the specific place in the transcript they came from — and then clustered into themes for that one case. In current IPA terminology these within-case themes are called Personal Experiential Themes (PETs).

4. Searching for connections across PETs

The case’s PETs are examined for relationships to each other — abstraction (grouping similar PETs under a superordinate label), subsumption (one PET absorbing related others), polarization (themes that oppose each other), and contextualization (situating a PET in time, place, or relationship) are the patterns most often used. The output of this step is a structured table of PETs for that single case, each with supporting verbatim extracts and transcript line references.

5. Moving to the next case

Steps 1 through 4 are repeated in full for the next transcript, as far as possible on its own terms rather than through the lens of the first case’s themes. Whether this “bracketing” of prior cases is ever fully achievable is a genuine, unresolved methodological debate in the phenomenological literature generally, not one IPA claims to have settled; most IPA guidance treats it as an aspiration to be reflexively managed rather than a clean, achievable reset, and expects the researcher to document that reflexive process. A documented reflexivity/positionality statement is standard practice here — see Reflexivity and Positionality Statements: A Template With Worked Examples.

6. Looking for patterns across cases: group experiential themes (GETs)

Only once every case has been analyzed individually does the cross-case work begin: the researcher lays the PET tables from every case alongside each other and looks for patterns of convergence and divergence — which experiential themes recur across most or all participants, which are present in only some cases, and where cases genuinely diverge. The themes that emerge from this comparison are the study’s Group Experiential Themes (GETs) — the level most readers think of as “the findings,” even though it is the product of the idiographic work underneath it, not a substitute for it. A GET is not simply a theme that happened to appear in every transcript; it should be traceable back to how it showed up, and varied, in each individual case, which is why the write-up (see below) keeps returning to individual participants by name or pseudonym rather than presenting GETs as anonymous aggregate categories.

PETs and GETs: a terminology update worth knowing before you read older sources

Smith, Flowers, and Larkin’s original 2009 text (Interpretative Phenomenological Analysis: Theory, Method and Research) described the within-case and cross-case outputs as “superordinate/subordinate themes” and “master themes” respectively. Smith and Nizza’s 2022 text (Essentials of Interpretative Phenomenological Analysis) revised this to the “experiential themes” language used above — Personal Experiential Themes (PETs) within a case, Group Experiential Themes (GETs) across cases — partly to keep the terminology explicitly tied to IPA’s experiential, phenomenological aims rather than borrowing the more generic “master theme” language also used outside IPA. Both vocabularies describe the same underlying process; a great deal of online guidance and older published studies still use the 2009 terms, so expect to see both in the literature, and be explicit in a methods section about which terminology a given study is using.

Sample size: small is not a limitation to justify, it’s the design

Because each case receives this much individual analytic attention, IPA does not aim for the kind of numeric saturation logic used elsewhere in qualitative research. A single, richly analyzed case is a complete and legitimate IPA study; most published IPA work sits in the 3-to-15-participant range, with smaller samples for professional doctorates or single-site clinical studies and slightly larger samples where the research question calls for more variation within a still-homogeneous group. What is not appropriate is treating IPA’s small-N idiographic design as a fallback for “we couldn’t recruit enough people for thematic analysis” — the small sample is a direct consequence of the analytic commitments above, and should be justified on those terms in a methods section, not apologized for.

Writing up IPA findings

An IPA results section is organized by GET, but each GET is substantiated with verbatim extracts from multiple participants, explicitly linked back to how it appeared differently in different individual cases — not just an aggregated theme name with a representative quote or two. The double hermeneutic should be visible in the prose itself: the write-up moves between what the participant said and the researcher’s interpretive commentary on it, rather than presenting only one or the other. Reviewers evaluating IPA work generally look for this idiographic traceability specifically — a GET that cannot be traced back to specific, cited moments in specific transcripts reads as thematic analysis with IPA’s name attached to it.

IPA vs. reflexive thematic analysis: how to tell which one you’re actually doing

IPA and Braun and Clarke’s reflexive thematic analysis are both common qualitative approaches and are sometimes used interchangeably in practice, which is a real problem for a methods section’s credibility. The distinguishing questions:

• Does the analysis stay anchored to one case at a time before comparing across cases? IPA does, by design (steps 1–5 above, worked per case before step 6). Thematic analysis typically codes across the whole dataset from an early stage.
• Is the sample small, homogeneous, and purposively selected for a shared experience? This is close to a requirement for IPA. Thematic analysis places no such constraint on sample composition or size.
• Is the double hermeneutic an explicit analytic aim, not just a general acknowledgment of researcher influence? IPA treats interpreting the participant’s sense-making as the point of the analysis. Thematic analysis (in its reflexive form) acknowledges researcher subjectivity throughout but is not built around this two-layer interpretive structure specifically.

See Thematic Analysis: A Step-by-Step Guide to Braun and Clarke’s Six Phases for the six-phase process on the other side of this comparison, and Qualitative Research Methods: The Major Methodologies and How to Choose for where both sit among the broader set of qualitative methodologies, including grounded theory.

Common pitfalls

• Borrowing IPA’s name without its design logic. Coding a mid-sized, heterogeneous sample across the whole dataset and calling it IPA is the same “methodological slurring” problem that affects grounded theory — the label without the underlying commitments.
• Rushing to GETs. Moving to cross-case comparison before each transcript has had its own full exploratory-noting and PET-construction pass undermines the idiographic foundation the cross-case themes are supposed to rest on.
• Presenting GETs as anonymous aggregate categories. A results section that reports GETs without tracing them back to specific participants and transcript moments loses the idiographic traceability reviewers look for.
• Treating small-N as a limitation rather than the design. A methods section that apologizes for sample size, rather than justifying it on IPA’s own idiographic terms, signals the design wasn’t fully understood.
• Using a heterogeneous sample. IPA’s cross-case comparison assumes participants share the specific phenomenon under study; a sample that mixes genuinely different experiences undermines what a GET is supposed to represent.

Frequently asked questions

What does IPA stand for in qualitative research?

Interpretative Phenomenological Analysis — not to be confused with other unrelated uses of the same acronym (e.g., “in-principle acceptance” in registered-reports publishing, or intergovernmental personnel exchanges). In a qualitative-methods context it always refers to the analytic approach described on this page.

How many participants does an IPA study need?

There is no fixed number. Most published IPA studies use somewhere between 3 and 15 participants, chosen for homogeneity around the specific experience under study rather than for statistical representativeness; a single, richly analyzed case is also a legitimate IPA study.

What is the difference between a personal experiential theme (PET) and a group experiential theme (GET)?

A PET is a theme identified within one participant’s transcript, during the case-by-case analysis stage. A GET is a theme identified by comparing PETs across every case in the study — the cross-case, study-level finding.

Can IPA use focus groups instead of individual interviews?

Occasionally, but it works against IPA’s idiographic logic, which depends on each participant’s account being analyzed as their own bounded case. One-to-one semi-structured interviews remain the standard data-collection method for IPA for this reason.

Is bracketing possible in IPA?

Whether a researcher can fully set aside prior assumptions and prior cases’ themes when approaching a new transcript is a genuinely contested question in the phenomenological literature, and IPA does not claim to resolve it. Most IPA guidance treats bracketing as a reflexive practice to document and manage, not a clean state to achieve.