Thematic analysis helps researchers find meaningful patterns in qualitative data and is adaptable across many disciplines. This article explains how thematic analysis compares to other methods, breaks down its key phases, and shows how tools like NVivo and the Lumivero AI Assistant can streamline the process.
Thematic analysis is a widely used method in qualitative research for identifying and interpreting patterns within data. It offers a structured way to handle large amounts of unstructured information, making it easier to recognize repeated ideas and meanings across a dataset. Qualitative researchers use thematic analysis to examine how participants talk about a particular topic, staying close to the data while also interpreting broader significance. The method is flexible and applicable across a range of disciplines, from psychology to education to public health.
This article outlines what thematic analysis is and how it compares to other approaches such as discourse analysis, content analysis, grounded theory, and framework analysis. It also explains how each phase of thematic analysis unfolds, particularly when using software like NVivo and ATLAS.ti. After that, we'll look at some key concepts relevant to the method, including data saturation and data triangulation.
Finally, we'll look at how NVivo and its analysis tools can help you with conducting thematic analysis easily and effectively. We will highlight the capabilities of analysis tools like NVivo's AI Assistant which can support you through the thematic analysis process and other qualitative research methodologies by improving your efficiency and enabling you to dive deeper into your data.
Supercharge your qualitative research with NVivo!
Thematic analysis is a method for identifying, analyzing, and interpreting meaningful patterns or themes within qualitative data. It focuses on what participants say and how they say it, using their own words as the basis for analysis. Themes reflect recurring patterns of shared meaning relevant to the research question. Quality thematic analysis goes beyond labeling and examines how meaning is constructed. This approach helps researchers organize data in detail and, depending on the aim, interpret it within a broader context.
Most commonly associated with Braun and Clarke, thematic analysis follows a six-phase process, from familiarizing oneself with the data to producing a report of the findings. It can be applied to various data types including interview transcripts, focus group discussions, open-ended survey responses, and written documents.
The method is flexible. It works inductively, where themes emerge from the data, or deductively, where analysis is guided by existing concepts or theories. This makes it suitable for both exploratory and theory-driven research. Thematic analysis also does not require specialized training or commitment to a specific theory, though researchers must be transparent about their assumptions and choices. Reflexivity, or thinking about how one's background shapes interpretation, is especially important.
Researchers should document their process: how codes were grouped, how themes were formed, and how findings relate to the data. This helps ensure transparency and trustworthiness. While thematic analysis can be done manually, software like ATLAS.ti or NVivo helps organize data and manage coding more efficiently.
Thematic analysis shares some features with other qualitative research methods but differs in its goals, procedures, and theoretical commitments. Understanding these differences helps clarify when thematic analysis is an appropriate choice and how it compares with other approaches that also involve coding and theme development.
Discourse analysis focuses on how language is used in social contexts. Rather than identifying recurring themes across a dataset, it examines the structure, function, and meaning of talk or text in specific settings. The focus is often on the way language shapes social realities and reflects power relationships, identities, or ideologies.
In contrast, thematic analysis prioritizes what is said rather than how it is said. It can involve some attention to discourse, but its primary aim is to identify shared meanings across the dataset. Discourse analysis tends to be more interpretive and theory-driven, while thematic analysis can be either data-driven or theory-informed depending on the research design.
Content analysis and thematic analysis are often confused because both involve coding and categorizing textual data. However, content analysis usually involves quantifying the presence of specific words, phrases, or concepts. It often relies on frequency counts and can be used to make inferences about the presence or absence of certain ideas in a dataset.
Thematic analysis is more interpretive. It goes beyond counting to understand the significance of themes and how they relate to each other. While qualitative content analysis might tell you how often a concept appears, thematic analysis focuses on what that concept means in the context of the data. The two qualitative methods can complement each other, but they serve different purposes and answer different types of research questions.
Grounded theory is a method for developing new theories grounded in data. It involves a series of tightly structured steps, including open coding, axial coding, and selective coding. The process is iterative and often includes theoretical sampling and constant comparison to refine emerging concepts and their relationships.
While both grounded theory and thematic analysis involve coding and theme development, grounded theory is more systematic and aims to build explanatory models. Thematic analysis does not aim to produce theory unless explicitly stated as part of a theoretical framework. It is more flexible and does not require adherence to a specific set of procedures. Researchers who want to describe and interpret themes without generating a full theory often prefer thematic analysis.
Get the guide: Download “The Essential Guide to Grounded Theory” to start building insightful, actionable theory from your qualitative data.
Framework analysis is commonly used in applied policy research. It involves a matrix-based approach to organizing data, with rows representing cases (e.g., interview participants) and columns representing codes or themes. The method is highly structured and designed for research with specific questions, pre-defined objectives, and clear time constraints.
Thematic analysis is less rigid and more adaptable. It allows researchers to revise codes and themes more freely as the analysis progresses. Framework analysis is useful for comparing data across multiple cases in a structured format, while thematic analysis supports more flexible, interpretive engagement with the data. Both can be used with qualitative software tools, but framework analysis tends to require a more predefined structure from the start.
Qualitative data analysis software (QDA software or QDAS) helps manage large volumes of text, organize coding work, and track decisions throughout a qualitative research project. While software does not analyze data for the researcher, it can support a more efficient and systematic application of thematic analysis. Tools like NVivo allow users to analyze transcripts, apply codes, retrieve coded segments, and visualize connections among themes. The following sections outline how these tools can support each phase of thematic analysis.
The first phase involves reading and re-reading the data to become deeply familiar with its content. This includes not just reading the transcripts but also listening to audio files or watching videos if available. The goal is to grasp the overall tone, flow, and range of meanings present in the dataset.
In QDAS, researchers can upload their data and add memo fields for each document. These memos can be used to note initial impressions, highlight surprising responses, or flag parts of the data that may be relevant later. Some software also supports annotations directly within the text. At this stage, researchers are not yet applying formal codes, but they are beginning to engage with the data analytically. Familiarization memos can be revisited in later phases to check whether early observations remain relevant.
The second phase involves creating codes to label important features of the data. Codes can describe content explicitly stated by participants (semantic codes) or imply underlying meanings (latent codes). This process helps organize the data into manageable units for a deeper qualitative data analysis.
Analysis tools streamline the coding process by allowing users to highlight text and apply one or more codes to each segment. Researchers can create a code list that grows organically as they move through the data. Most QDAS programs let users merge or rename codes, add code definitions, and keep a log of when codes were applied. Coding stripes or highlights can be turned on to visualize overlapping or related codes, which is especially helpful when returning to the same transcript for multiple rounds of coding.
Once the data has been coded, the third phase focuses on identifying patterns among the codes. Related codes are grouped together to form initial themes that reflect shared meanings across the dataset. This stage is interpretive and involves organizing codes into broader categories without finalizing their definitions.
In QDAS, researchers can use code groups or code networks to cluster similar codes. Some platforms provide visual tools like code co-occurrence tables or network diagrams to support this process. These features help users identify which codes tend to appear together or within similar contexts. Researchers can also return to the coded segments to verify whether the groupings reflect a coherent idea or if the codes need to be split or regrouped.
This phase involves reviewing the initial themes to ensure they are consistent, distinct, and grounded in the data. Researchers check that the themes work across the dataset and reflect the meanings captured in the original transcripts. Some themes may be broken apart, refined, or discarded during this stage.
QDAS tools support theme review by allowing researchers to retrieve all data coded under a specific code or theme. Viewing all segments side by side makes it easier to assess whether the theme holds together or if it is too broad or vague. Researchers can also use memoing features to document decisions about refining or adjusting themes. Linking analytic memos to specific themes provides a clear audit trail of how interpretations evolved.
After reviewing themes, the fifth phase involves refining and naming each one. This includes defining what each theme captures, determining its scope and boundaries, and giving it a concise and meaningful name. Researchers aim to distinguish each theme clearly from the others.
In software environments, researchers can revise the structure of code groups or rename codes to reflect the refined themes. Some platforms allow the creation of summary reports that include theme names, definitions, and representative quotes. These features support transparency by providing a snapshot of how each theme is grounded in the data. Writing theme descriptions as memos can also help prepare for the final report.
The final phase involves presenting the analysis in a report, article, or other format. The goal is to describe the themes clearly and support them with selected excerpts from the data. Researchers interpret the significance of each theme and link their findings back to the research question.
QDAS can assist in report writing by exporting coded segments, theme summaries, and analytic memos. Many tools allow researchers to export data directly into word processing or reference management software. This helps streamline the integration of raw data and interpretation. Researchers can retrieve quotes efficiently, check where specific interpretations came from, and track how each finding links to earlier stages of analysis.
NVivo helps researchers manage, code, and interpret large amounts of qualitative data. It is commonly used for thematic analysis because it allows for structured coding, the organization of themes, and the integration of memos and notes throughout the process. NVivo also includes an AI Assistant that can support early stages of analysis by summarizing transcripts and suggesting potential themes. The following subsections describe how NVivo supports key steps in thematic analysis.
Coding qualitative data in NVivo involves highlighting segments of text and applying labels – called "codes" – that represent meaningful units of information. These codes can represent topics, concepts, emotions, actions, or any other feature relevant to the research question.
To begin qualitative coding, researchers import data sources such as interview transcripts or focus group responses into the NVivo project. Text can be selected and coded to new or existing codes. Each time a segment is coded, NVivo stores a reference to the original data, allowing for quick retrieval and comparison. Researchers can also use code hierarchies to organize codes into parent and child categories, which is useful when analyzing complex or layered themes.
NVivo supports descriptive, interpretive, and pattern coding, allowing researchers to use the software in both inductive and deductive approaches. Memos can be attached to specific codes or sources to explain coding decisions or record observations for later use.
NVivo also allows users to visualize code coverage, code density, and co-occurrence with other codes. These tools help track the distribution of ideas across the dataset and provide insights into which themes are most developed.
Once initial codes have been created, NVivo helps researchers move into theme development by examining patterns across coded data. This involves reviewing coded segments to identify shared meanings, organizing codes into categories, and refining them into meaningful themes.
Researchers can group related codes into folders or use NVivo’s “sets” to gather data segments across different sources or cases. The software also includes tools for comparing coding across cases, which can support theme refinement based on participant type, demographic group, or other attributes.
NVivo’s query tools allow for searches based on codes, text strings, or coding combinations. Matrix coding queries and coding comparison queries are especially useful when reviewing how different codes interact or overlap. These outputs can support the interpretation of relationships between codes and guide decisions about theme boundaries.
Visualizations like coding stripes, word clouds, and code co-occurrence maps offer alternative views of the data. These tools do not generate themes automatically, but they help researchers test ideas and make coding decisions more transparent.
The NVivo AI Assistant is a feature designed to help researchers work more efficiently and generate early insights from their data.
When used during the data familiarization or early coding stages, the AI Assistant can help identify points of interest that might otherwise take longer to detect. Suggested topics are based on semantic similarity and can serve as provisional codes or prompts for deeper analysis.
The AI Assistant also supports the creation of a thematic map by clustering related ideas and offering candidate labels. These suggestions are intended to help researchers think critically about the data rather than replace their judgment. All themes and summaries can be reviewed, edited, or rejected by the researcher.
Importantly, AI Assistant is designed to support the researcher by assisting with summarizing the data and suggesting child codes to apply to the data. By accelerating early-phase tasks, it allows researchers to focus more on interpretation and insight-building later in the process.
NVivo users should treat the AI Assistant as one part of a larger analytic workflow. The value of the tool lies in its ability to support – not substitute – the researcher’s role in interpreting meaning and making analytical decisions.
Thematic analysis is a method used to identify, analyze, and interpret patterns – or themes – within qualitative data. It helps researchers understand how participants express ideas and experiences related to a specific topic.
The six steps of thematic analysis are:
NVivo and ATLAS.ti are the best software for thematic analysis, reviewed as the most trusted and best-rated QDA software. Both allow researchers to code data, organize codes, and generate themes in a structured and efficient way. As part of Lumivero’s portfolio, NVivo also offers AI-assisted features for early-stage analysis.
AI tools in software like NVivo can summarize transcripts, suggest key topics, and propose initial themes. These features help researchers save time and focus more on interpretation, while still allowing full control over the final analysis.
Yes, thematic analysis is well-suited to focus groups. It allows researchers to identify shared views, disagreements, and patterns across participants.
No, it is a flexible method that can be applied inductively or deductively and adapted to a range of epistemological positions.
If you're ready to take your qualitative research to the next level, NVivo is your go-to tool for powerful, efficient thematic analysis. It helps you organize, code, and interpret large volumes of unstructured data – so you can move from raw transcripts to meaningful patterns with clarity and confidence.
