Gioia Methodology: Data Structure Guide with Example
Qualitati Research Team · 2026-10-18 · 11 min read
Short answer: The Gioia methodology is an inductive approach to qualitative analysis, set out by Dennis Gioia, Kevin Corley and Aimee Hamilton in Organizational Research Methods (2013). Researchers code informant language into 1st-order concepts, group them into researcher-centric 2nd-order themes, distill those into aggregate dimensions, and then build a grounded model showing how the concepts relate.
If you are doing an inductive interview study in management, information systems, entrepreneurship or organization studies, a reviewer will probably ask to see your "data structure." That request comes from the Gioia methodology, one of the most widely used templates for presenting qualitative rigor in business-school research.
This guide explains what the method is and walks through the steps with a short worked example. It also covers the main critiques you should know before you commit to it and what recent research says about using large language models to build a Gioia data structure. It ends with a checklist you can use before submitting.
Key takeaways
- Three levels, one figure. 1st-order concepts (informant terms) feed 2nd-order themes (researcher concepts), which feed aggregate dimensions. The "data structure" figure shows this progression.
- The data structure is not the finding. Gioia, Corley and Hamilton (2013) treat the static data structure as a step toward a dynamic grounded model, which is where the theoretical contribution lives.
- Informants are treated as knowledgeable agents. The method gives prominence to participants' own interpretations before the researcher theorizes.
- It is contested as a template. Mees-Buss, Welch and Piekkari (2022) argue that applying it as a fixed recipe does not solve interpretive problems, and they propose hermeneutic heuristics instead.
- LLMs handle the bottom level better than the top. A 2025 AMCIS study found a reasoning model produced plausible 1st-order concepts but overgeneralized at the theme and dimension levels.
What is the Gioia methodology?
The Gioia methodology is a systematic approach to inductive concept development and grounded theory articulation. It was designed to make qualitative work look and feel rigorous to readers without killing its capacity to reveal new concepts. Its signature output is a three-level data structure followed by a grounded theoretical model.
The canonical statement is Gioia, Corley and Hamilton, "Seeking Qualitative Rigor in Inductive Research: Notes on the Gioia Methodology", Organizational Research Methods, vol. 16, no. 1, pp. 15–31 (2013). The approach grew out of earlier work, notably Gioia and Chittipeddi's 1991 study of strategic change in a university. In 2023, Magnani and Gioia published an application guide for international business and entrepreneurship researchers in International Business Review (vol. 32, no. 2, article 102097). It describes three stages: building the data structure, building the grounded model, and writing the findings as a persuasive narrative.
Two assumptions underpin the method. First, the organizational world is socially constructed. Second, the people who construct it are "knowledgeable agents" who can explain their own thoughts, intentions and actions. The researcher's job is to represent those voices faithfully and then to theorize at a level the informants would not themselves use.
The three levels of a Gioia data structure
Each level answers a different question, and confusing them is the most common reason reviewers send the figure back.
| Level | Whose language | What it answers | Typical form |
|---|
| 1st-order concepts | Informant-centric | What are participants saying, in close to their own words? | Short descriptive phrases, often many of them early on |
| 2nd-order themes | Researcher-centric | What concepts help describe and explain what is going on here? | Theoretically meaningful labels, some drawn from the literature, some new |
| Aggregate dimensions | Researcher-centric | What overarching ideas do the themes add up to? | A small number of high-level dimensions that anchor the model |
Gioia and colleagues describe moving between the data and the literature once 2nd-order themes start to stabilize. The researcher asks whether an emerging theme already has a name in existing theory or whether it is a new concept. That back-and-forth is why the method is often described as inductive in spirit with an abductive move in practice. Our guide to abductive analysis covers that logic in more detail.
How to do a Gioia analysis: step by step
Here is how the 2013 article and the 2023 guide describe the method, written as a sequence you can follow. Treat it as a working order rather than a strict recipe. The authors themselves warn against using it as a cookbook.
1. Design for informant voice
Start with a well-specified research question that leaves room for concepts you have not yet named. Use semi-structured interviews that let informants describe events in their own terms, and revise the protocol as new concepts emerge. Our guide to semi-structured interviews covers protocol design.
2. Code 1st-order concepts
Code close to the informants' terms and resist early abstraction. Expect a large, messy list at this stage. In vivo codes are a natural fit here (see in vivo coding).
3. Reduce and group
Look for similarities and differences among the 1st-order concepts and consolidate them into a manageable set. Give each group a label or phrase that stays recognizable to informants.
4. Develop 2nd-order themes
Ask what concepts would describe and explain the patterns. Now bring in the literature, and pay particular attention to themes that have no adequate name in existing theory. These are your candidate contributions.
5. Distill aggregate dimensions
Where it makes sense, collapse the 2nd-order themes into a few aggregate dimensions. Draw the data structure figure: concepts on the left, themes in the middle, dimensions on the right.
6. Build the grounded model
Turn the static structure into a dynamic model that shows how the concepts relate over time or causally, with arrows, phases and feedback loops. Without this step, a data structure is description rather than theory.
7. Write the findings with "power quotes" and "proof quotes"
Use a small number of vivid quotes in the narrative and put fuller supporting evidence in a table organized by theme. Then show readers how each theme connects to the model.
A worked example (illustrative)
Imagine an interview study of 25 university researchers who adopted AI coding tools. The example below is invented to show the logic of the levels. It is not drawn from a published study.
| 1st-order concepts (informant terms) | 2nd-order theme | Aggregate dimension |
|---|
| "I re-read every AI code before I trust it"; "I check the quote is really in the transcript" | Verification rituals | Calibrating trust in the machine |
| "It's fine for first pass, not for themes"; "I'd never let it name a theme" | Bounded delegation |
| "My supervisor asked who did the coding"; "Reviewer 2 wanted the prompts" | Anticipated audience scrutiny | Defending methodological legitimacy |
| "I wrote a paragraph on it in the methods chapter"; "I kept a log of what the AI suggested" | Documenting the human role |
A grounded model might then propose that anticipated scrutiny drives verification rituals, and that these rituals widen the scope of bounded delegation over time. The model would make a claim about process that the static table cannot make.
Original framework: the Gioia Data Structure Readiness Checklist
Qualitati's research team built this checklist from the sources cited in this article. Use it before you send a manuscript to a reviewer or supervisor. Each "no" is a likely revision request.
| # | Check | Why it matters |
|---|
| 1 | Every 1st-order concept can be traced to at least one quote in the proof table | Shows the structure is grounded, not asserted |
| 2 | 1st-order labels use informant language, not theory terms | Keeps the informant-centric and researcher-centric levels separate |
| 3 | Each 2nd-order theme is defined in one sentence in the text | Lets readers judge whether concepts were grouped consistently |
| 4 | You state which themes match existing concepts and which are new | Makes the theoretical contribution explicit |
| 5 | Aggregate dimensions are few and clearly distinct | Prevents a model so generic it explains nothing |
| 6 | A dynamic grounded model follows the data structure | The static figure alone is not the contribution |
| 7 | Disconfirming or negative cases are reported | Answers the critique that templates smooth over interpretive tension |
| 8 | If AI assisted coding, the methods section says where, how, and how output was verified | Lets readers assess the human interpretive role |
Critiques and limitations of the Gioia methodology
The Gioia methodology works well as a way to present rigor, but it is not a guarantee of insight. Know the debate before you choose it.
- Template versus interpretation. Mees-Buss, Welch and Piekkari, "From Templates to Heuristics: How and Why to Move Beyond the Gioia Methodology" (Organizational Research Methods, 25(2), 405–429, 2022), argue that the method's inductive route does not address the challenges of interpretation. They recommend hermeneutic heuristics: probing questions rather than prescribed steps.
- Homogenization. Gioia's 2022 curated debate in the Journal of Management Inquiry (record) collected scholars' positions on whether widespread template use narrows what qualitative research in management looks like. Gioia himself argues the method should be treated as a flexible orientation, not a formula.
- Epistemological fit. The method assumes informants are knowledgeable agents. Critical, discursive or psychoanalytic studies that read against participants' accounts may fit poorly. See our guide to discourse analysis for one alternative.
- Not a fit for every field. The method is dominant in management journals. In health or education research, reflexive thematic analysis or template analysis may be more familiar to reviewers.
Human-review note: whether the Gioia methodology suits your study depends on your epistemology, field and target journal. Discuss the choice with your supervisor or co-authors.
Gioia methodology vs grounded theory vs thematic analysis
| Criterion | Gioia methodology | Grounded theory (Glaser/Strauss/Charmaz) | Reflexive thematic analysis |
|---|
| Main output | Three-level data structure plus a grounded model | A substantive theory around a core category | A set of themes telling a story about the data |
| Role of literature | Brought in once themes start to stabilize | Varies by school; classic version delays it | Flexible; theory can inform the analysis throughout |
| Signature artifact | Data structure figure | Theoretical memos and a core category | Thematic map |
| Typical home | Management, IS, entrepreneurship | Sociology, nursing, management | Psychology, health, education |
For the coding vocabulary behind these approaches, see open, axial and selective coding and thematic analysis vs grounded theory.
Can AI build a Gioia data structure?
Partly. AI is most useful at the 1st-order level and least reliable at the aggregate-dimension level.
The clearest evidence so far comes from Laato, Mäntymäki, Kordyaka and Laato, "Automating Qualitative Data Analysis with Chain-of-Thought Reasoning Models: A Study with the Gioia Method" (AMCIS 2025 Proceedings, published 15 August 2025). They used DeepSeek R1 to try to replicate a published Gioia analysis of 17 expert interview transcripts, and checked with Levenshtein distance that the quotes the model returned actually existed in the transcripts. The model produced 1st-order concepts similar to the original study's but lost some of their meaning. It then overgeneralized, producing 2nd-order themes and aggregate dimensions the authors judged inaccurate and of little informational value.
That pattern matches what we see in broader research on AI qualitative coding (see where AI qualitative coding breaks down). A practical division of labor:
- Let AI help: drafting candidate 1st-order codes, retrieving every passage tagged with a concept, checking that quotes exist verbatim, and building the proof-quote table.
- Keep for the researcher: naming 2nd-order themes, deciding which concepts are theoretically novel, distilling aggregate dimensions, and building the dynamic model.
Where Qualitati fits
Qualitati is a European qualitative research platform for universities and research companies, designed with data privacy and GDPR as central priorities. For a Gioia-style study it supports the parts of the workflow that are mechanical, and it leaves the interpretive moves to you:
- Data collection: AI-moderated interviews in text or voice, in 10 languages, with transcripts saved automatically. Or use Active Listener mode to get real-time prompts while you conduct the interview yourself.
- 1st-order coding: QDA Workspace supports AI-assisted inductive and deductive coding and manual coding, along with inter-rater reliability.
- A draft hierarchy to argue with: QDA Workspace's Three-Level Codebook report produces a themes → categories → codes hierarchy. Treat it as a starting scaffold to challenge, not as your data structure. The evidence above suggests the upper levels are exactly where AI overgeneralizes.
- Cross-transcript synthesis: ThemeLens maps codes to research questions across up to 100 transcripts, with participant-anchored quotes that help when you assemble a proof table.
Qualitati does not draw the Gioia data structure figure or the grounded model for you, and it should not. Those are the theoretical contribution.
Who this is for, and when not to use it
Use the Gioia methodology if you are doing inductive, interview-based theory building and targeting management, IS or entrepreneurship journals whose reviewers expect a data structure. It also suits PhD students who need a transparent, teachable analytic trail.
Do not use it if your study is deductive or hypothesis-testing, if your epistemology reads against informants' accounts, if your aim is descriptive rather than theory-building, or if your field's reviewers expect reflexive thematic analysis or classic grounded theory.
FAQ
What is the difference between 1st-order concepts and 2nd-order themes?
1st-order concepts stay close to informants' own words and describe what participants say. 2nd-order themes are researcher-centric concepts that explain what is going on theoretically. Aggregate dimensions then group the 2nd-order themes at a higher level.
Is the Gioia methodology the same as grounded theory?
No. It draws on grounded theory and Gioia and colleagues describe it as a way to articulate grounded theory, but it has its own presentation conventions, notably the three-level data structure and the move to a dynamic grounded model. It also brings in the literature once themes start to stabilize.
How many interviews do I need for a Gioia analysis?
The method does not prescribe a number. Justify sample size with saturation or information power, as in any qualitative interview study. See our guide to qualitative sample size.
What software can I use to build a Gioia data structure?
Any QDA tool that supports hierarchical codes works, including NVivo, ATLAS.ti, MAXQDA and Qualitati's QDA Workspace. The figure itself is usually drawn in a diagramming or slide tool. Software organizes the codes; it does not perform the abstraction.
Can ChatGPT or another LLM do a Gioia analysis?
Not reliably on its own. A 2025 AMCIS study found a reasoning model produced reasonable 1st-order concepts but inaccurate, overgeneralized 2nd-order themes and aggregate dimensions. Use AI for drafting and retrieval, and verify every quote.
What is the main criticism of the Gioia methodology?
Critics such as Mees-Buss, Welch and Piekkari (2022) argue that using it as a fixed template does not resolve interpretive challenges and can homogenize qualitative research. They propose hermeneutic heuristics instead of prescribed steps.
Conclusion
The Gioia methodology gives inductive researchers a shared language for showing their work: informant terms, researcher themes, aggregate dimensions and a grounded model. Used thoughtfully, it makes interpretation visible to reviewers. Used mechanically, as both its critics and Gioia himself warn, it produces tidy figures with thin theory. AI tools can speed up the bottom of the data structure, but the higher levels and the model remain the researcher's work.
To run your next interview study and code it in one place, you can start free with 30 monthly credits, no credit card required. See transparent pricing for larger studies, or compare Qualitati with NVivo, ATLAS.ti and MAXQDA.
Last updated: 2026-10-18.
Written by the Qualitati Research Team. Methodological claims are attributed to the cited sources; please review them against your discipline's standards before use in a thesis or publication.