Grounded Theory Methodology Explained
Building theory from data rather than testing it, coding stages, constant comparison, theoretical sampling and saturation in grounded theory.
Grounded theory (GT) is a qualitative methodology for generating theory from data rather than testing pre-existing theory against it. Developed by Glaser and Strauss, it is ideal where little existing theory explains a phenomenon. Its defining features, constant comparison, theoretical sampling and coding toward saturation, set it apart from generic qualitative analysis.
Theory Emerges from Data
Unlike deductive research, GT does not start with a hypothesis. You approach the field with an open question and let concepts emerge from systematic analysis of the data. This inverts the usual logic: data first, theory as the product.
The Coding Stages
- Open coding: break data into discrete incidents and label them, staying close to the data.
- Axial coding: reassemble codes into categories, identifying relationships between them.
- Selective coding: integrate categories around a core category that captures the central phenomenon, forming the theory.
Constant Comparison
Throughout, you continually compare new data with existing codes and categories, incident to incident, category to category. This constant comparative method is the engine of GT, forcing categories to earn their place and evolve as evidence accumulates.
Theoretical Sampling
GT sampling is not fixed in advance. As your emerging theory raises questions, you deliberately sample new data to develop and test the categories, following the theory where it leads. Who you interview next depends on what the analysis so far has revealed.
Theoretical Saturation
Data collection continues until new data no longer yield new properties of your categories, the point of saturation. This, not a preset number, determines sample size in a true GT study.
Memo-Writing
Throughout, you write analytic memos capturing your developing ideas about codes and their relationships. Memos are the bridge between coding and the final written theory, and skipping them is a common reason GT studies stall at description.
A Caution
GT is frequently claimed but rarely done in full. Using a few coding steps does not make a study grounded theory. If you invoke GT, commit to constant comparison, theoretical sampling and saturation, or describe your work more modestly as thematic analysis.
Checklist
- Did concepts emerge from data rather than a prior hypothesis?
- Did I use constant comparison across all coding stages?
- Was sampling theoretical, driven by the emerging theory?
- Did I collect to saturation and write analytic memos?
- Have I named which version of grounded theory I am following?
- Does my write-up present a theory rather than a set of themes?
Three Versions, and Why the Choice Matters
Grounded theory is not one method. It split into distinct traditions, and they disagree about things that affect what you do daily, including whether you should read the literature first.
| Version | Associated with | Distinctive stance |
|---|---|---|
| Classic | Glaser | Theory emerges from data. Delay the literature review to avoid forcing concepts |
| Straussian | Strauss and Corbin | Structured coding procedures, including a coding paradigm for conditions and consequences |
| Constructivist | Charmaz | Theory is constructed by researcher and participants together. Reflexivity is central |
Constructivist grounded theory is the most common choice in current student work, partly because its position on researcher influence is easier to defend than a claim of pure emergence. Whichever you choose, name it in the methodology and let it govern your decisions consistently. Mixing Glaser's stance on literature with Strauss's coding procedures is a contradiction an examiner may well probe.
The Coding Sequence
| Stage | What you do | Output |
|---|---|---|
| Open or initial coding | Code line by line or incident by incident, staying close to the data | Many provisional codes, often gerunds such as "managing uncertainty" |
| Focused or axial coding | Test the most frequent and significant codes across the data, relate categories to each other | A smaller set of working categories |
| Theoretical coding | Specify how categories relate, and identify the core category | An integrated theory with a central explanatory concept |
Coding in gerunds is a small technique with a large effect. "Managing uncertainty" preserves action and process, where "uncertainty" flattens into a topic. Since grounded theory is trying to explain how something happens, keeping the verbs alive keeps the analysis moving toward process rather than description.
Constant Comparison, Concretely
Constant comparison is the engine of the method and the part most often described but not done. In practice it means four kinds of comparison, running throughout.
- Incident with incident, to see whether two pieces of data really belong under the same code.
- Incident with category, to test whether new data fits the category you built or forces you to change it.
- Category with category, to establish how they relate and whether two are actually one.
- Data with the emerging theory, to check the theory still accounts for what you are seeing.
The consequence is that analysis and collection interleave. You cannot do grounded theory by collecting everything and then analysing it, because the analysis is supposed to determine what you collect next.
Theoretical Sampling and Saturation
Theoretical sampling means your emerging categories decide who or what you approach next. If a category about disclosure appears strongest among newer staff, you deliberately seek longer-serving staff to test its limits. This is not convenience or representativeness. It is sampling driven by conceptual need.
Saturation follows from that. It is reached when further data stops generating new properties of your categories, not when you feel you have enough interviews. State how you judged it, and be specific about which categories reached it. Blanket claims of saturation after a fixed number of interviews are a familiar and unpersuasive move.
Memos Are Not Optional
Memo writing is where grounded theory analysis actually happens. Codes are labels. Memos are where you argue with yourself about what a code means, how it relates to another, and what it would take to disprove your reading.
Write them from the first coding session, date them, and keep them. They become the audit trail that demonstrates dependability, they supply much of the eventual prose of your findings chapter, and they are the difference between a theory you can defend in a viva and a set of categories you can only describe.
Common Mistakes and Fixes
| Mistake | Fix |
|---|---|
| Collecting all data first, then analysing | Interleave. Without that, theoretical sampling is impossible |
| Calling thematic analysis grounded theory | If you did not sample theoretically or build a theory, describe it accurately as thematic analysis |
| Presenting categories with no relationships | Specify how categories connect, and identify the core category |
| No memos | Start now. They are evidence as well as thinking |
| Mixing versions without noticing | Name one tradition and apply it consistently |
| Claiming saturation by interview count | Explain what stopped changing, and in which categories |
Is Grounded Theory Realistic for Your Project?
Be honest about the demands. Interleaved collection and analysis needs ethics approval flexible enough to allow evolving sampling, and it needs time you may not have in a one-year masters. Many students would produce better work by running a well-executed reflexive thematic analysis and saying so.
There is no shame in that choice, and examiners respond far better to an accurately named method than to an aspirational label the work does not support.
Where Ethical Support Fits
Discussing emerging categories with a supervisor is standard, and in constructivist grounded theory the researcher's engagement is treated as part of the method rather than a contaminant. Ethics committees will also need to understand why your sampling cannot be fully specified in advance, and a supervisor can help you write that section.
What stays yours is the coding, the memos and the theory. The theory is the deliverable, and it has to be built by the person who read the data.
Frequently Asked Questions
Should I read the literature before starting?
Classic grounded theory says delay it. Constructivist grounded theory treats a working familiarity as unavoidable and asks you to be reflexive about it instead. In practice most degrees require a literature review up front, so follow your version's logic and explain how you handled the tension.
How many participants does grounded theory need?
It is determined by saturation, not set in advance. Published student studies commonly land between 15 and 30 interviews, but the number is an outcome rather than a plan.
What is a core category?
The central concept that accounts for most variation in the data and that the other categories relate to. Identifying it is what turns a set of categories into a theory.
Can I use software?
Yes. NVivo, MAXQDA and ATLAS.ti all support memoing and code comparison. They organise the work, they do not do it.
Is grounded theory suitable for document analysis?
It can be, since the method is not restricted to interviews. Theoretical sampling of documents is harder, so explain how you selected further material as categories developed.
How do I present the theory in the write-up?
Lead with the core category and the model, then show how each category supports it with evidence. A diagram of the relationships is usually worth including, because a theory expressed only in prose is hard for an examiner to hold.
Your Next Step Today
Open your codes and check whether they are nouns or gerunds. Rewriting a handful as processes, from "uncertainty" to "managing uncertainty", is a small mechanical change that reliably pushes the analysis from description toward theory.
Trusted Sources
- APA Dictionary of Psychology, grounded theory. Accessed 11 August 2026.
- Tenny S, Brannan JM and Brannan GD, Qualitative Study, StatPearls, NCBI Bookshelf. Accessed 11 August 2026.
- University of Southern California, Qualitative Methods. Accessed 11 August 2026.
- Charmaz K, Constructing Grounded Theory (2nd edn, Sage 2014), the standard constructivist account.
- Glaser BG and Strauss AL, The Discovery of Grounded Theory (Aldine 1967), the founding text.
Grounded theory traditions differ substantially. Where this guide and your supervisor's preferred version differ, follow your supervisor.
Related services
Related guides
- Thematic Analysis: Braun and Clarke's Six Phases
- Writing the Methodology Chapter: Design, Sampling and Justification
- Sampling Methods in Research Explained
- Research Ethics for Students: Consent, Privacy, and Integrity
All academic writing guides or browse our academic writing services.