Grounded Theory
Generating Theory from Data
Produces inductively derived theory grounded in the systematic analysis of social process
Grounded theory is among the most widely used and most frequently misapplied qualitative research methodologies in the social, health, and educational sciences. It was designed not to test hypotheses but to generate theory — substantive, verifiable, conceptual accounts of social processes derived from rigorous, iterative engagement with empirical data. Understanding it correctly demands close attention to its philosophical commitments and exact procedural requirements.
Definition & Scope of Grounded Theory
Grounded theory is a qualitative research methodology whose purpose is the generation of theory from the systematic analysis of empirical data. It proceeds inductively: the researcher enters the field with minimal preconceptions, collects data, analyzes them concurrently, and allows conceptual categories and their relationships to emerge from the data rather than imposing frameworks derived from existing literature. The result is a theory that is grounded in the data from which it was constructed.
The foundational text is The Discovery of Grounded Theory by Barney Glaser and Anselm Strauss (1967), published as an explicit challenge to the then-dominant assumption in sociology that qualitative work could only verify pre-existing theory, not generate new theory. Glaser and Strauss argued that systematic inductive analysis of qualitative data could produce rigorous, credible, and generative theory on equal footing with deductive, hypothesis-testing research.
The Central Question
Grounded theory is appropriate when the researcher is asking: "What is happening here? What social process is at work? What theory can account for this pattern of behavior?" The question is process-oriented, not descriptive. It seeks not to catalogue what people experience but to explain the social, psychological, or organizational process through which they navigate it.
Charmaz (2014, p. 1) is precise on this point: grounded theory studies must aim at theory construction, not rich description alone. A study that collects qualitative data, codes it thematically, and presents findings without developing conceptual propositions or relational statements between categories is not a grounded theory study, regardless of the labels it applies to its procedures.
When Grounded Theory Is the Appropriate Choice
According to Creswell and Poth (2018, pp. 83–84) and Birks and Mills (2023), grounded theory is methodologically indicated when:
- No adequate existing theory explains the phenomenon or social process under investigation
- Existing theories require substantial modification to account for the observed social process
- The research question concerns a social process, action, or interaction that unfolds over time
- The researcher seeks to generate a conceptual or theoretical account that can be tested, refined, or transferred to related settings
- The study population has not been adequately theorized in existing literature
Grounded Theory Distinguished from Other Qualitative Designs
| Design | Central Aim | Core Question | Primary Output | Temporal Focus |
|---|---|---|---|---|
| Grounded Theory | Generate theory of a social process | What theory explains this process? | Substantive or formal theory | Process, over time |
| Phenomenology | Describe lived experience | What is the essence of this experience? | Essential description | Present, immediate |
| Ethnography | Describe a culture-sharing group | What does this culture look like? | Cultural description | Sustained, immersive |
| Case Study | In-depth case understanding | What can be learned from this case? | Thick case description | Bounded case |
| Narrative Inquiry | Understand individual story | How does this person story experience? | Restoried narrative | Longitudinal, biographical |
Historical Development
The intellectual genealogy of grounded theory is inseparable from mid-twentieth century debates within sociology about the status of qualitative research. Its development traverses three distinct generational phases, each associated with specific methodological commitments and philosophical revisions.
Symbolic Interactionism: The Sociological Root
All three grounded theory traditions share roots in symbolic interactionism, the sociological perspective associated with George Herbert Mead, Herbert Blumer, and the Chicago School. Symbolic interactionism holds that human beings act toward things on the basis of the meanings those things have for them, that meanings arise through social interaction, and that these meanings are modified through an ongoing interpretive process (Blumer, 1969, p. 2). This commitment to meaning as socially constructed and processual — rather than fixed, pre-given, or measurable — is what grounds grounded theory in qualitative rather than quantitative inquiry (Chamberlain-Salaun et al., 2013; Charmaz, 2025).
The 1967 Origin: A Methodological Challenge
The publication of The Discovery of Grounded Theory in 1967 was a direct challenge to the prevailing functionalist sociology of Talcott Parsons, which held that the proper role of empirical research was the verification of grand theoretical frameworks. Glaser and Strauss argued that this arrangement had it backwards: theory should emerge from systematic engagement with data, not be imposed upon it from above. Their text introduced the core methodological strategies — constant comparison, theoretical sampling, and saturated categories — that remain central to all grounded theory variants today (Glaser & Strauss, 1967, pp. 21–43).
The 1992 Split: Glaser versus Strauss and Corbin
When Strauss and Corbin published Basics of Qualitative Research in 1990, Glaser responded with Basics of Grounded Theory Analysis: Emergence vs. Forcing (1992), a sustained critique arguing that Strauss and Corbin's axial coding paradigm and verification procedures imposed conceptual structures on the data rather than allowing theory to emerge from it. This methodological rupture produced two distinct traditions that remain in tension, with researchers required to choose between them on both philosophical and practical grounds.
The Three Traditions
The three principal traditions of grounded theory — Glaserian, Straussian, and Constructivist — share a recognizable family of methodological characteristics while diverging significantly in their philosophical commitments, coding procedures, role of the literature, and relationship between researcher and data. Selecting the appropriate tradition is not a matter of convenience or preference; it requires principled alignment between the researcher's ontological and epistemological position and the tradition's underlying assumptions (Birks & Mills, 2023; Charmaz, 2014, pp. 7–11).
Ontology: Naive realism. An objective social reality exists independently of the observer. The researcher discovers — rather than constructs — theory that already exists in the data waiting to be uncovered (Glaser, 1978, 1992).
Epistemology: Post-positivist. The researcher is a neutral instrument who applies rigorous procedures to allow the data to speak for themselves. The researcher's role is to enable emergence, not to impose structure.
Literature review timing: Delayed until after initial analysis. Prior literature should not contaminate theoretical emergence. The researcher enters the field with as few preconceptions as possible (Glaser, 1978, pp. 31–32). This is the most distinctive — and most contested — feature of classical grounded theory.
Coding sequence: Substantive coding (open/initial) followed by theoretical coding. Theoretical codes conceptualize how substantive codes relate to each other. Glaser identified sixty-four theoretical coding families (e.g., the Six C's: causes, contexts, contingencies, consequences, covariances, conditions) as heuristic guides, not prescriptions (Glaser, 1978, pp. 72–82).
Core category: Central to classical grounded theory. The core category is the concept around which all other categories integrate. It must have the greatest explanatory power and must account for the pattern of behaviour that is relevant and problematic for those involved. The core category becomes the basis of the substantive theory.
Primary reference: Glaser, B. G. (1978). Theoretical sensitivity. Sociology Press.
Ontology: Critical realism. Multiple realities exist and are always interpreted; there is no single pre-existing objective reality. However, patterns and regularities are identifiable through rigorous analysis (Strauss & Corbin, 1998, pp. 10–11).
Epistemology: Post-positivist with constructivist leanings. The researcher's subjectivity is acknowledged but managed through systematic procedures. Strauss and Corbin moved toward acknowledging the researcher's interpretive role while maintaining verification as a quality criterion.
Literature review timing: A preliminary review is recommended to clarify the research problem and identify gaps. The literature is revisited after theory emerges to situate the theory within existing knowledge (Strauss & Corbin, 1998, pp. 48–49; Corbin & Strauss, 2015, pp. 37–38).
Coding sequence: Three-stage sequential coding: (1) open coding — fracturing data into discrete concepts; (2) axial coding — reassembling data around categories using the paradigm model (phenomenon, causal conditions, context, intervening conditions, action/interaction strategies, consequences); (3) selective coding — integrating categories around a central/core category. This is the most prescriptive coding framework in the grounded theory tradition.
Paradigm model: The conditional matrix and paradigm model are distinctive features of the Straussian approach. They provide a visual and conceptual framework for understanding how categories relate. Critics — including Glaser — argue these impose structure rather than allow it to emerge.
Primary references: Strauss, A., & Corbin, J. (1990, 1998). Basics of qualitative research. Sage. Corbin, J., & Strauss, A. (2015). Basics of qualitative research (4th ed.). Sage.
Ontology: Relativist/constructivist. Multiple realities exist and are always socially constructed. There is no single truth waiting to be discovered — only perspectives, meanings, and interpretations that are themselves co-constructed between researcher and participants (Charmaz, 2014, pp. 12–14).
Epistemology: Constructivist/interpretivist. Knowledge is constructed between researcher and participants. The researcher's positionality, theoretical sensitivities, and prior experiences actively shape the analytic process. Rather than managing or bracketing subjectivity, the researcher makes it visible through reflexivity (Charmaz, 2014, pp. 15–17; Charmaz, 2024).
Literature review timing: A preliminary conceptual review is recommended to identify the research problem. The literature is used throughout analysis to develop theoretical sensitivity and is integrated after theory construction (Charmaz, 2014, pp. 304–307).
Coding sequence: Three stages: (1) initial coding — line-by-line, word-by-word, or incident-by-incident coding to remain close to the data; (2) focused coding — selecting the most significant codes and applying them analytically across larger data segments; (3) theoretical coding — integrating categories into coherent theoretical relationships. Gerund-form (action-based) codes are preferred to capture process (Charmaz, 2014, pp. 111–152).
Theoretical output: Charmaz describes the resulting theory as an interpretive rendering, not a discovered truth. Quality criteria foreground reflexivity, resonance, and usefulness rather than verification and generalizability (Charmaz, 2014, pp. 236–243).
Primary reference: Charmaz, K. (2014). Constructing grounded theory (2nd ed.). Sage. Charmaz, K. (2024). Grounded theory. In J. A. Smith (Ed.), Qualitative psychology: A practical guide (pp. 81–110). Sage.
Selecting on philosophical alignment: The first criterion is consistency between the researcher's ontological and epistemological position and the tradition's assumptions. A researcher who holds that objective social patterns exist and can be discovered should work in the Glaserian tradition. A researcher who believes reality is always interpreted and constructed should work in the Straussian or Constructivist tradition (Birks & Mills, 2023; Charmaz, 2014).
Selecting on research purpose:
- Classical GT (Glaser): Best suited when the researcher wants maximum inductive emergence with minimal a priori structure; when the phenomenon is poorly understood; when the researcher has sufficient theoretical sensitivity to work without prescriptive coding guides
- Straussian GT (Strauss & Corbin): Best suited for researchers who want systematic, verifiable procedures; who work in health sciences or social work where methodological transparency is valued; who want a prescribed analytical framework
- Constructivist GT (Charmaz): Best suited for researchers who embrace an interpretivist or constructivist worldview; who work in social sciences or education; who foreground reflexivity and researcher positionality; who want a flexible framework that is nevertheless conceptually rigorous
Key principle: Whichever tradition is selected, it must be applied consistently and completely. Mixing coding procedures from Straussian and Constructivist traditions without justification — a common error — produces methodological incoherence (Birks & Mills, 2023, p. 45; Corbin & Strauss, 2015, p. 6).
| Feature | Glaserian (Classic) | Straussian (Systematic) | Constructivist (Charmaz) |
|---|---|---|---|
| Ontology | Naive realism | Critical realism | Relativism/constructivism |
| Epistemology | Post-positivist | Post-positivist/constructivist | Constructivist/interpretivist |
| Role of researcher | Neutral detector | Systematic analyst | Active co-constructor |
| Literature review | After analysis | Preliminary + after | Preliminary + throughout |
| Coding approach | Substantive + theoretical | Open + axial + selective | Initial + focused + theoretical |
| Core output | Core category + theory | Paradigm model + theory | Conceptual theory (interpretive) |
| Quality criterion | Fit, work, relevance, modifiability | Fit, applicability, concepts, logic | Credibility, originality, resonance, usefulness |
| Primary risk | Undisciplined emergence | Overstructured forcing | Insufficient theoretical abstraction |
Core Concepts in Grounded Theory
Grounded theory employs a set of technical concepts that carry precise methodological meanings distinct from their colloquial usage. Precision in these definitions is essential: using grounded theory terminology loosely without performing the corresponding analytical acts is one of the most frequently cited quality failures in published grounded theory research (Birks & Mills, 2023; Charmaz, 2014; Corbin & Strauss, 2015).
Coding Procedures in Grounded Theory
Coding in grounded theory is not the assignment of labels to text segments. It is an active analytical process through which the researcher moves from raw data toward theoretical abstraction. Each tradition prescribes a distinct coding sequence; these are not interchangeable. The researcher must understand not only what each stage requires but why each stage is necessary to the theory-building process.
Classic GT
Systematic
Constructivist
Initial / Open / Substantive Coding in Depth
The first stage of coding in all traditions involves close, line-by-line engagement with the data. The goal is to fracture the data into conceptually distinct units and to label each unit with a code that captures its analytical significance. At this stage, the researcher must resist the temptation to apply pre-existing conceptual frameworks. The data, not theory, must drive the codes.
Charmaz's (2014, pp. 111–132) instruction to use gerund-form codes — present participles that capture action and process — is one of the most practical and analytically powerful prescriptions in grounded theory methodology. Gerund codes keep the analysis oriented toward process: rather than coding a passage as "uncertainty," the researcher codes it as "managing uncertainty," "tolerating uncertainty," or "denying uncertainty." These process-oriented codes are more analytically productive because they capture how people act, not just what they experience.
Axial Coding and the Paradigm Model (Straussian)
Axial coding is the analytical procedure through which the categories identified during open coding are reassembled around axes of relationships. Strauss and Corbin (1998, pp. 123–142) developed the paradigm model as a heuristic framework for this reassembly. The model organizes the relationship between:
- Causal conditions — events or circumstances that give rise to the central phenomenon
- The central phenomenon — the main idea, event, or incident around which the theory is built
- Context — the specific properties of the situation in which action/interaction occurs
- Intervening conditions — broader conditions that facilitate or constrain action/interaction
- Action/interaction strategies — purposeful responses to the phenomenon under specific conditions
- Consequences — outcomes of action/interaction strategies
Glaser (1992) argued this model constitutes forcing — imposing structure on data that should be allowed to emerge. Researchers using the Straussian framework must justify their use of the paradigm model and demonstrate that it fits the data rather than being applied mechanically.
Selective Coding and the Storyline
Selective coding is the final analytical stage in which the researcher integrates all categories around the core category. Strauss and Corbin (1998, pp. 143–161) recommend writing the storyline — a descriptive narrative statement that identifies the central phenomenon and describes the process through which it operates — as the first step in selective coding. The storyline provides the integrating frame around which the theory is constructed.
Charmaz's (2014, p. 163) guidance: "Stop and memo whenever an interesting idea occurs to you." Memos should be dated, titled by code or category name, and kept separate from field notes. They grow in complexity as analysis proceeds: early memos are exploratory; later memos are integrative, developing theoretical relationships between categories.
Key principle: Memos serve as the bridge between initial coding and theoretical writing. Without sustained memo-writing, grounded theory data collection and coding cannot produce genuine theory — they can only produce description (Birks & Mills, 2023, pp. 97–118; Corbin & Strauss, 2015, pp. 117–131).
- The Six C's: Causes, contexts, contingencies, consequences, covariances, conditions — basic causal-process relationships
- Process: Stages, phases, progressions, passages, transitions — temporal sequence relationships
- Degree: Limit, range, intensity, extent, amount — dimensional variation relationships
- Strategy: Tactics, techniques, maneuvers, mechanisms — action-oriented relationships
- Identity-self: Self-image, self-concept, social worth, identity transformations — identity process relationships
- Cutting point: Boundaries, thresholds, turning points, critical junctures — event-marking relationships
The Grounded Theory Research Process
Grounded theory research does not proceed as a linear sequence of discrete phases. Data collection, coding, constant comparison, memo-writing, and theoretical sampling occur simultaneously and iteratively from the outset of the study. This simultaneity is what distinguishes grounded theory from research designs that complete all data collection before beginning analysis. The following procedure draws on the Straussian framework as the most comprehensively specified, with comparative notes on Constructivist adaptations (Corbin & Strauss, 2015; Charmaz, 2014; Creswell & Poth, 2018).
Data Sources Beyond the Interview
While in-depth, semi-structured interviews remain the dominant data source in grounded theory research, the tradition permits a range of data types. Glaser and Strauss (1967, pp. 65–78) explicitly permitted the use of documentary data, observation field notes, and secondary sources. Charmaz (2014, pp. 25–56) provides practical guidance on using personal documents, organizational records, and internet data. The researcher must be able to justify why each data source contributes to theoretical development and how it is integrated into the constant comparative process.
Rigour, Quality Criteria, and Limitations
The criteria for evaluating the quality of grounded theory differ across traditions and are not reducible to the trustworthiness framework developed for generic qualitative research by Lincoln and Guba (1985), though that framework is frequently applied. Charmaz (2014, pp. 236–243) developed a constructivist-specific set of criteria that have become the most widely cited quality framework in contemporary grounded theory literature.
Glaser (1978, pp. 4–5) proposed four criteria for evaluating the quality of a grounded theory:
- Fit: The theory must fit the data from which it emerged. Categories must be grounded in data, not imposed from without. Participants familiar with the substantive area should recognize the theory as relevant to their experience.
- Work: The theory must explain, predict, and interpret what is happening in the substantive area. It must be able to generate hypotheses about how the process will operate in new situations.
- Relevance: The theory must address the main concern of participants in the substantive area — it must explain what is actually problematic for those who are living the phenomenon.
- Modifiability: A grounded theory is never finished; it should be open to modification when new data emerge that require revision of existing categories or the integration of new ones.
Strauss and Corbin (1990, pp. 250–261) and Corbin and Strauss (2015, pp. 339–354) proposed evaluative criteria organized around the validity of the research process and the resulting theory:
- Concepts: Are the concepts grounded in data? Does the theory consist of conceptual categories rather than mere description?
- Contextualization of concepts: Are properties and dimensions developed for each major category, demonstrating contextual variation?
- Logical coherence: Are the relationships between categories logically consistent? Is the theoretical framework internally coherent?
- Usefulness: Does the theory offer insight into the substantive area? Can it be applied to generate further understanding or guide practice?
- Variation: Does the theory account for variation in the phenomenon — the conditions under which it operates differently?
- Significance: Does the theory contribute meaningfully to knowledge? Does it go beyond what was already known?
Charmaz (2014, pp. 236–243) proposed four criteria for constructivist grounded theory, oriented toward the interpretive and reflexive dimensions of the approach. These criteria are now the most widely cited in social science grounded theory research:
- Credibility: Has the researcher achieved intimate familiarity with the setting and topic? Is the data sufficient and varied? Do the theoretical claims rest on a comprehensive analysis that goes beyond isolated quotations?
- Originality: Does the theory offer new conceptual renderings? Does it challenge, extend, or refine existing ideas? Does it make a significant analytic contribution to the literature?
- Resonance: Do the categories portray the fullness of the studied experience? Does the theory reveal the intersection of individual experience and broader social processes? Would participants recognize themselves in the theory?
- Usefulness: Can the theory be used outside the study itself? Does it offer interpretations that can be applied to related settings? Does it contribute to better practice, policy, or further research?
Charmaz subsequently added a fifth criterion, Sincerity (reflexivity and transparency), to the framework in later iterations of her work (Charmaz, 2024).