Who Framed Qualitative Synthesis?: Thematic versus Framework approaches and how to choose.

Qualitative evidence synthesis hinges on a practical choice: whether to build explanations by clustering data into themes (thematic synthesis) or by mapping data onto an existing conceptual structure and then revising it where it fails (framework synthesis). The core finding is that method selection should follow the state of theory in a topic, the richness of the data, and the purpose of the synthesis—whether the goal is to describe, test, generate, or explain mechanisms. Getting that match right matters because it determines how transparently conclusions can be justified and how well the synthesis can handle complexity without forcing data into the wrong mold.

The session frames the decision around three common qualitative synthesis approaches recommended in major guidance: thematic synthesis, framework synthesis, and meta-ethnography. Thematic synthesis is presented as the most intuitive and widely taught option, especially when data are “thin” and may not support strong theory-building. It typically proceeds by extracting data, coding it, translating codes into themes, and then examining relationships among themes to produce higher-order interpretations. A key strength is accessibility and the ability to integrate with quantitative synthesis; a key risk is oversimplification—turning nuanced qualitative meaning into tidy labels that may not translate well into guidance or policy.

Framework synthesis is positioned as a method for topics with existing theory or frameworks. When multiple models or frameworks already exist, mapping data to a pre-existing structure becomes a prime candidate. The method can handle complexity because frameworks often encode relationships, and it can support both theory testing and theory generation. A central concept is “best fit” framework synthesis: start with a framework that explains a substantial portion of the data (pragmatically, more than 50%), map what fits, and then use an inductive process to capture what the framework misses—producing a revised model with a clear “switch point” from deductive mapping to inductive development. Transparency is emphasized through explicit documentation of what data map to the original framework and what does not.

Thematic versus framework differences are illustrated with a holiday analogy: thematic synthesis groups excerpts into themes and then links themes (e.g., climate leads to enjoyment), while framework synthesis maps excerpts onto an existing model (e.g., an “optimal holiday” model) and updates it when new data add elements such as humidity. The analogy underscores how framework synthesis can refine theory rather than merely rename patterns.

Meta-ethnography is treated as the most interpretive and academically prominent approach, often likened to grounded theory. It is most appropriate when data are plentiful and “thick,” and when the aim is to generate higher-order conceptual explanations. Because it abstracts from original study contexts, it can be harder for readers to see how conclusions were reached, and it may require a more experienced team.

The practical decision pathway boils down to four questions: Is the topic well theorized (favoring framework synthesis)? Are there large numbers of studies (favoring sampling and potentially thematic synthesis when no framework exists)? Is the data rich enough to sustain theory generation (favoring meta-ethnography when thick)? And is the goal to explain mechanisms (pointing toward realist approaches) or to generate theory more broadly (pointing toward meta-ethnography). The session closes by stressing that framework choice should be tested early for receptivity and credibility, since frameworks can become “a gallows” if stakeholders reject the lens even when findings are useful.

Concrete examples show framework synthesis in action: adapting frameworks for attitudes to mobile apps across chronic kidney disease, diabetes, and dementia; using a framework from a review of reviews of qualitative synthesis to code enabler factors for childhood obesity in ethnic minorities more broadly; and applying a pathway model to organize data by temporal stages in weight management programs. Overall, the guidance makes method choice less about tradition and more about matching theory availability, data thickness, and the intended explanatory payoff.