Constant Comparison

The core analytic procedure of Grounded Theory. Each new datum is compared with previous data, with codes already assigned, and with the developing categories — moving from incident-to-incident comparison early on toward comparison of incidents with concepts and concepts with concepts as analysis matures.

Procedure

Glaser and Strauss (1967) set out four overlapping stages. (1) Comparing incidents applicable to each category: as a code is applied to an incident, the analyst compares it with prior incidents in the same category to surface its properties. (2) Integrating categories and their properties: comparison shifts from incident-to-incident to incident-to-category, moving the category from a descriptive label toward a conceptual entity with stable properties and dimensions. (3) Delimiting the theory: the framework tightens as further comparisons stop generating new properties and the analyst becomes more selective about what counts. (4) Writing the theory: memos accumulated across the earlier stages are integrated into the analytic account.

Function

Constant comparison disciplines the analyst against premature closure. It forces the question "how is this case similar to and different from the cases already seen?" at every step. The procedure produces categories that are conceptually bounded — each has properties along which cases vary, rather than being a topic bucket — and drives Theoretical Saturation, the point at which further comparisons no longer alter the framework.

Beyond grounded theory

Constant comparison has migrated outside grounded theory into general qualitative practice. Many studies that do not aim at theory generation still use the procedure as a coding discipline, comparing each new extract against the existing scheme. In this borrowed form it functions as a quality control on coding consistency rather than as part of a theory-building programme.

Use in applied linguistics

Constant comparison appears in interview studies of teacher cognition, classroom discourse studies, and longitudinal learner research where the analyst tracks how categories evolve across data points. Reporting conventions ask the analyst to be explicit about what was compared with what, and at what level of abstraction.

References

• Glaser, B. G., & Strauss, A. L. (1967). The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine.
• Charmaz, K. (2006). Constructing Grounded Theory: A Practical Guide Through Qualitative Analysis. SAGE.