Narrative Inquiry for Course Evaluation: Reading Student Stories as Wholes, Not Fragments

In brief

Narrative inquiry treats a student's account of a course as a story — an experience configured over time with a beginning, a turning point, and a resolution — rather than as a bag of themes to be counted. Where thematic, framework, and grounded-theory coding deliberately fragment feedback into categories, narrative analysis keeps each account whole and asks what the plot reveals about how the course was actually lived. It is powerful for understanding why a course worked or failed and for surfacing turning points, but it is labour-intensive, small-scale, interpretive, and unsuited to summative comparison across many courses.

What the research says

Narrative inquiry entered educational research through F. Michael Connelly and D. Jean Clandinin (1990), whose Educational Researcher paper "Stories of experience and narrative inquiry" defined narrative as both the phenomenon and the method: people lead storied lives and tell stories of them, and the researcher describes, collects, and re-tells those stories. Their framework locates every account in a three-dimensional space — temporality (past, present, future), sociality (personal and social conditions), and place — insisting that experience is understood only when situated in time and context, not extracted from it. A student saying "the course finally clicked in week eight" is making a claim about sequence and change that a decontextualised code cannot hold.

Donald Polkinghorne (1995) drew the sharpest methodological distinction in the field, between two operations analysts routinely conflate. Analysis of narratives (which he calls paradigmatic) moves from stories to common elements: it collects accounts and sorts them into categories and themes — this is what thematic and framework coding do. Narrative analysis moves the opposite way, from elements to stories: it configures the events and happenings a person reports into a coherent plot by means of emplotment, producing an explanatory story rather than a category list. The two answer different questions, and treating open-text feedback only paradigmatically discards the temporal, causal structure students actually convey.

Jerome Bruner (1991), in "The narrative construction of reality," gave the epistemological grounding: humans think in two irreducible modes, the paradigmatic (logical, categorical) and the narrative (story-based, concerned with human intention and consequence over time). Course evaluation almost exclusively uses the first and largely ignores the second, even though students reason about their learning narratively. Catherine Kohler Riessman (2008), in Narrative Methods for the Human Sciences, provided the working typology practitioners use — thematic, structural, and dialogic/performative narrative analysis — and stressed that narrative work is co-constructed between teller and researcher and demands reflexivity about that co-construction.

Why it matters for course evaluation in practice

Standard evaluation is relentlessly paradigmatic: a mean, a top-box percentage, a list of the five most frequent themes. That machinery is blind to structure that only appears when an account is read as a whole.

Turning points. The single most actionable thing a student can tell you is when and why their experience shifted — the assignment that suddenly made the subject make sense, the week the workload became unmanageable. These pivot points are the levers a redesign can pull, and they are precisely what theme-frequency counts erase by scattering the pieces of one story across separate codes.

Causal texture. Students narrate consequence: "because the first lab was so rushed, I never caught up." A thematic code tags "pacing" and "labs" separately and loses the causal link between them. Narrative analysis preserves the because, which is what a programme director needs to understand a failure rather than merely tally it.

Divergent journeys. Two students can end at the same 3/5 having travelled opposite paths — one who started engaged and lost faith, one who started lost and recovered. The mean hides this completely; the stories make it visible, and they imply different interventions.

Closing the loop credibly. A well-told student story, respectfully anonymised, communicates to a teaching team what a bar chart never will, and is often what actually motivates change. Narrative evidence complements the numbers rather than replacing them.

Limitations and honest caveats

Narrative inquiry is easy to romanticise, and a rigorous reader will hold it to account.

First, it does not scale and is not meant to. Configuring and interpreting whole accounts is labour-intensive; a narrative study handles tens of cases, not thousands. It is a depth instrument for understanding, not a monitoring instrument for a whole institution, and using it to compare hundreds of courses is a category error.

Second, generalisability is limited by design. Narrative findings are particular and contextual. Their value is transferability — insight a reader can judge as relevant to their own setting — not statistical representativeness. Presenting a compelling story as typical without warrant is a real risk.

Third, the researcher co-constructs the story. Because meaning is negotiated between teller and analyst (Riessman), a different analyst may emplot the same events differently. This demands explicit reflexivity, an audit trail, and ideally member-checking with the student, or the "findings" reduce to the analyst's preferred reading.

Fourth, selection and voice bias. The students who tell vivid stories are not a random sample; articulate or aggrieved voices dominate, and the quiet middle goes unheard. Narrative work must be honest that it privileges the tellable.

Fifth, it is unsuited to summative, high-stakes comparison. Narrative evidence should inform formative improvement and understanding, not rank instructors for personnel decisions, where its subjectivity and non-representativeness would be indefensible.

How Koji incorporates this

Koji is unusually well suited to generating the material narrative inquiry needs, because its core is a conversation rather than a form — but it applies the method with the caveats above, never overclaiming.

• Eliciting the arc, not just the rating. Koji's AI-moderated conversational interview can follow a student through the story of their term — asking what changed, when, and why — producing accounts with the temporal and causal structure that Connelly and Clandinin, and Polkinghorne, treat as the raw material of narrative analysis, rather than the flattened snapshot a Likert grid yields.
• Structured probes that surface turning points. Using open_ended questions with adaptive follow-ups, Koji can gently probe the pivot points ("what made it click?", "when did it start to slip?") that theme-counting misses, capturing the because that carries the actionable insight.
• Both modes of analysis, kept distinct. Koji's automatic thematic analysis serves Polkinghorne's paradigmatic mode (themes across accounts), while the full transcript is preserved so a human analyst can do genuine narrative analysis (configuring one student's events into a plot). Keeping the transcript intact is what makes the second mode possible at all.
• Reflexive, ethical reporting. Koji supports careful anonymisation and keeps the student's own words available for member-checking-style review, and it pairs any illustrative story with the quantitative and representativeness context so a single vivid narrative is never mistaken for the typical case.

Koji's core research platform at koji.so applies the same AI-moderated interview engine to product and customer research, where narrative and journey-based analysis of user experience is a mainstay — the education product simply points that engine at the student's experience of a course.

Frequently asked questions

How is narrative inquiry different from thematic analysis?

Thematic analysis fragments accounts into codes and themes counted across respondents; narrative inquiry keeps each account whole and analyses its plot — the sequence, turning points, and causal links a student conveys over time. Polkinghorne calls these two directions paradigmatic analysis (stories to elements) and narrative analysis (elements to a story). They answer different questions and are complementary.

Does narrative inquiry replace numeric course evaluations?

No. It complements them. Numbers give representative, comparable summaries; narrative gives depth, mechanism, and turning points that numbers cannot. A mature quality process uses narrative for formative understanding and improvement, and quantitative ratings for monitoring and fair comparison, keeping each to what it does well.

Can I use narrative inquiry to compare instructors or courses?

Not for high-stakes summative comparison. Narrative work is small-scale, interpretive, co-constructed, and non-representative, so ranking instructors on it would be indefensible. Use it to understand why a course succeeds or struggles and to inform redesign, not to score people.

How many student stories do I need?

Far fewer than a survey — narrative studies typically work with tens of cases, chosen for richness rather than representativeness. The goal is deep understanding and transferable insight, not statistical power, so adequacy is judged by whether the accounts illuminate the question, not by a sample-size formula.

What is the main methodological risk?

That the analyst's interpretation substitutes for the student's meaning, since narratives are co-constructed. Guard against it with reflexivity, an audit trail of how stories were configured, member-checking where possible, and honesty that vivid tellers are not a random sample of the class.

Where does narrative inquiry fit in a course-evaluation strategy?

As the depth layer. Run it selectively — on a redesigned course, a struggling module, or a purposive sample — alongside routine quantitative evaluation, to explain the numbers and locate the levers for change. It is a formative, understanding-oriented method, not a replacement for institution-wide monitoring.

Related resources

• Grounded Theory for Open-Text Course Feedback
• The Qualitatively Different Ways Students Experience Your Course: Phenomenography
• The Critical Incident Technique for Course Feedback
• The Most Significant Change Technique
• The Framework Method for Open-Text Course Feedback
• What Can Open-Text Student Comments Tell You That Likert Scores Cannot?

References

• Connelly, F. M., & Clandinin, D. J. (1990). Stories of experience and narrative inquiry. Educational Researcher, 19(5), 2-14. https://doi.org/10.3102/0013189X019005002
• Polkinghorne, D. E. (1995). Narrative configuration in qualitative analysis. International Journal of Qualitative Studies in Education, 8(1), 5-23. https://doi.org/10.1080/0951839950080103
• Bruner, J. (1991). The narrative construction of reality. Critical Inquiry, 18(1), 1-21. https://doi.org/10.1086/448619
• Clandinin, D. J., & Connelly, F. M. (2000). Narrative Inquiry: Experience and Story in Qualitative Research. Jossey-Bass.
• Riessman, C. K. (2008). Narrative Methods for the Human Sciences. Sage.