Qualitative data analysis - the role of theoretical framework in data analysis

Qualitative research can use an established theoretical framework effectively—but only if the framework is tied to how data were collected. The key decision is how strongly the framework shaped the interview guide (or other data collection methods). If the framework was used to design questions that directly target specific elements, analysis can legitimately be more deductive, with coding aimed at those predetermined concepts. If the framework was only part of the background reading—without driving the interview questions—then analysis should stay more inductive, letting themes emerge from participants’ accounts before the framework is brought back in.

A practical way to think about it starts at the data collection stage. When researchers build an interview guide entirely around a framework, the interview tends to probe only the framework’s components. In that scenario, thematic analysis and the resulting thematic framework (and later codes and themes) will closely mirror the original model, leaving limited room for unexpected findings. This approach isn’t “wrong,” especially when the study’s purpose is to test whether a previously established framework applies in a new context or to replicate earlier work. Still, it offers less opportunity for the study to contribute anything beyond confirming that the framework’s elements show up in the new setting.

For studies that aim to do more than replication—especially when researchers are already investing time in literature review, method design, and analysis—there’s a stronger case for building in space for emergent findings. Even when the goal is to assess whether a framework fits, researchers can start interviews with open-ended questions that invite participants to discuss what influences their experiences or identity in their own terms. For example, in a migrant identity study, researchers might first ask what factors shape someone’s identity or lived experience, and only then ask about specific factors drawn from the framework (such as language competence). If participants independently raise the same factors the framework predicts, that alignment strengthens the framework’s credibility; if they raise additional factors, the study can extend or refine the framework.

Data analysis follows from that design choice. When interviews were tightly structured around framework elements, coding can be done deductively—creating codes in advance that represent the framework’s factors and then applying them to the transcripts. When interviews were more exploratory, the recommended move is to temporarily “forget” the framework during initial coding: analyze from the ground up using an inductive, bottom-up approach so findings are grounded in the data rather than forced into preexisting categories. This reduces the risk of overlooking important themes that don’t neatly match the framework.

The framework comes back at the end. After developing themes and a thematic framework from the data, researchers compare their wording and concepts with the established model to ensure clarity for readers. The goal isn’t to rewrite participants’ meaning to fit the theory, but to align terminology where frameworks overlap—so concepts aren’t confusingly labeled differently across models. Finally, if a theoretical framework wasn’t used to build the interview guide at all, researchers can still discuss similarities and differences later, without treating the framework as a driver of data collection or analysis. The overall payoff is stronger validity: analysis isn’t prematurely shaped by the theory, yet the final interpretation can still connect to established scholarship.