Narrative analysis: methods, models and an example

Narrative analysis is a family of qualitative methods for interpreting the stories people tell: what a story is about, how it is put together, and to whom and why it is told. Instead of breaking interviews into codes across participants, it usually keeps each story whole and treats the order of events as part of the meaning. It suits research questions about experience over time, identity and how people make sense of turning points.

What is narrative analysis?

The standard textbook is Catherine Kohler Riessman’s Narrative Methods for the Human Sciences (2008). She describes narrative analysis as a family of methods for interpreting texts that share a storied form. A narrative in this sense is more than talk about a topic: the teller selects events, puts them in order and connects them so that they mean something to an audience. Sequence and consequence matter: what happened, in what order, and what followed from it.

Definitions vary in breadth. For William Labov, a minimal narrative needs only two clauses in a fixed order: swap them and the listener understands a different sequence of events. Other researchers treat a whole life story as one narrative, or study the short stories that come up in passing in conversation. Decide which unit you are analysing before you start, because it changes what you transcribe and what you select.

Narrative analysis and narrative inquiry

Narrative inquiry is the wider methodology; narrative analysis is what you do with the stories. D. Jean Clandinin and F. Michael Connelly’s Narrative Inquiry: Experience and Story in Qualitative Research (2000), influential in education research, treats studying experience as studying stories, lived and told. Building on John Dewey, they place every story in a three-dimensional narrative inquiry space: the personal and social (interaction); past, present and future (continuity); and place (situation). Narrative inquirers often work with participants over long periods, moving from field texts such as transcripts and notes to research texts.

Donald Polkinghorne (1995) draws a distinction worth knowing. In an analysis of narratives, you collect stories and look for themes or types across them. In narrative analysis in his narrower sense, you collect events and accounts and compose them into a story, such as a case history. Most student projects do the first, so say which one you mean.

Riessman’s four types of narrative analysis

Riessman (2008) sets out four approaches. They can be combined, but each asks a different main question of the same story.

Thematic narrative analysis resembles thematic analysis but differs in one decisive way: Riessman keeps each story intact and theorises from the case, rather than from codes pulled out of many cases. Structural analysis is the most linguistic, and Labov’s model is its classic tool. Dialogic or performance analysis brings the interviewer into the analysis, since a story told to a fellow student may sound different from the same story told to a lecturer. Visual analysis suits photo elicitation and arts-based projects.

The Labov model: six parts of a narrative

William Labov and Joshua Waletzky (1967) built their model on oral narratives of personal experience told in sociolinguistic interviews, where Labov’s best-known prompt asked whether the speaker had ever been in serious danger of being killed. Their model had five parts. Labov (1972) added a sixth, the abstract, and the six-part version is the one most textbooks teach. The parts tend to come in this order, except evaluation, which can appear anywhere.

1. 1

   Abstract

   A clause or two that sum up what the story is about and signal that a story is coming. Optional.
2. 2

   Orientation

   Who, when, where and what was going on: the background the listener needs.
3. 3

   Complicating action

   The sequence of events, each answering the question ‘and then what happened?’ It is the one part a narrative cannot do without.
4. 4

   Evaluation

   Why the story is worth telling: what the events meant to the teller. It often comes just before the resolution but can run through the whole story, in comments, emphasis, repetition or mentions of what did not or might have happened.
5. 5

   Resolution

   What finally happened: how the complicating action ended.
6. 6

   Coda

   A closing clause that brings the listener back to the present, such as ‘and that’s why I never take the night bus’. Optional.

The model fits short, event-centred stories of personal experience, because that is what it was built on. It fits less well with habitual narratives (‘we would always…’), long life stories and stories built jointly with the interviewer. Use it as a tool for structural analysis, not as a test of whether something counts as a proper story.

Eliciting stories in interviews

Interviews do not automatically produce stories. Elliot Mishler’s Research Interviewing: Context and Narrative (1986) treats the interview as a speech event produced jointly by both speakers, and shows how standard question-and-answer interviewing tends to cut stories short. To hear stories, ask for them and leave room for them.

• Ask about particular events. ‘Can you tell me about a time when…’ or ‘What happened on your first day?’ invites a story; ‘What do you think about…’ invites an opinion.
• Let the story run. Hold back your next question until the participant signals the end, and show you are listening with short responses.
• Use narrative probes. ‘What happened then?’, ‘How did that come about?’ and ‘Can you give me an example?’ keep the story going without steering it.
• Return to stories in the participant’s words. When you follow up, use their terms and the order in which they told things.

Some designs build the whole interview around this. Tom Wengraf’s biographic narrative interpretive method (BNIM), set out in Qualitative Research Interviewing (2001), opens with a single question aimed at inducing narrative (SQUIN), such as an invitation to tell one’s life story, and lets the participant speak without interruption. Follow-up questions come afterwards and use the participant’s own words, in the order they were told. BNIM builds on Fritz Schütze’s narrative interview and Gabriele Rosenthal’s biographical case reconstruction from German sociology. It is demanding, so discuss it with your supervisor before you choose it.

From interview to analysis

Narrative analysis starts from a full transcript, because notes cannot show where a story begins and ends or how it was told. A workable sequence:

• transcribe the whole interview, including your own questions, which are part of how each story was produced;
• mark where each story starts and ends, and note why you drew the line there, since the boundaries are already an interpretation;
• re-transcribe the stories you analyse closely at the level your approach needs: numbered clauses for structural analysis, pauses and overlaps for performance analysis;
• analyse each case in depth before you compare across cases.

Check your university’s rules before you upload recordings anywhere: a voice identifies the speaker, so a recording is personal data even without names. Spirelight offers no data processing agreement; your university, not you, would sign one. Transcription runs on EU servers, files are stored with Google Firebase in the US, and AI features run on OpenAI in the US, which your ethics application and participant information sheet must state.

Where your university allows it, you can transcribe your interviews and get a draft with speaker labels and a timestamp on each speaker turn, which helps you find where each story starts in the recording. Correct the draft while you listen. The tool does not find stories or label their parts; that is your analysis.

Invented example: one story in Labov’s six parts

The study, participant and story below are invented. The research question is: How do international master’s students tell the story of their first weeks at a Norwegian university? The story comes from one interview and is divided using Labov’s model.

In a full study you would compare how other participants evaluate similar stories. A performance analysis would also ask how P4 told the story to this particular interviewer, and why.

Narrative analysis vs thematic analysis

Both can start from the same transcripts, but they cut them differently. Thematic analysis codes segments across the whole dataset and builds themes from the codes, so each participant’s account is split up and recombined with others. Narrative analysis keeps each story whole and treats sequence, structure and telling as data.

Choose narrative analysis when the way events are ordered, connected and told matters to your question, for example how people narrate a turning point or present themselves through a story. Choose thematic analysis when you want patterns of meaning across many participants; the guide to reflexive thematic analysis explains that method. If your question concerns how talk constructs identities and versions of events more generally, compare discourse analysis.

The ethics of using people’s stories

Stories are personal, recognisable and open to more than one reading, which raises particular ethical problems. Ruthellen Josselson (2007) describes narrative research as a relational endeavour in which it is not self-evident how to put the familiar principles of consent, confidentiality and protection from harm into practice.

• Stories identify people. A pseudonym does not hide someone whose story involves a rare job, an unusual event or a small community. Change identifying details where you can without changing the meaning, and tell participants that people who know them might still recognise them.
• Your reading may differ from the teller’s. Say in your participant information sheet how stories will be used and analysed, and decide in advance whether you will share transcripts or interpretations with participants and what you will do if someone disagrees.
• Painful stories. Asking for stories can bring back distressing experiences. Plan how you will pause or end an interview, and where you can point participants for support.
• Consent is a process. In long or repeated interviews, check consent again as the project goes on, and follow your university’s rules on withdrawal.

Your university’s ethics procedure decides what you need to do, so use its templates and talk to your supervisor before your first interview.