NVivo Qualitative Analysis Tutorial: Coding, Themes & Reporting

Last Updated on 6 days ago by Bernard Mugo

What this tutorial covers: This complete NVivo tutorial walks you through the full qualitative analysis process for interview data — from importing raw transcripts through coding, developing themes using Braun and Clarke’s framework, writing up your findings, and visualizing your results using nine different NVivo tools.

If NVivo feels overwhelming and you’re not sure how to analyze your interviews properly, you’re in exactly the right place.

In this complete step-by-step tutorial, I’ll take you through the full qualitative analysis process in NVivo — from raw transcripts through qualitative coding, developing themes, reporting your findings, and visualizing your results clearly and professionally. By the end, you’ll know exactly how to use NVivo for your thesis or dissertation.

My name is Bernard Mugo. Over the past four years, I’ve helped more than 600 PhD students analyze qualitative data and finally complete their dissertations. This tutorial brings together the most important NVivo skills into one place. For software documentation throughout this tutorial, the NVivo official resource hub is a reliable reference.

This tutorial covers four parts:

1. How to Code Interview Transcripts in NVivo
2. How to Develop Themes from Your NVivo Codes
3. How to Write Up and Report Your Findings from NVivo
4. 9 Ways to Visualize Your Qualitative Data in NVivo

One of the most important skills in qualitative data analysis is generating qualitative codes — what we simply call “coding.” Coding is the process of tagging important quotes from your original data and interpreting them in a way that lets you combine those codes to form themes at a later stage.

A code is a short label or interpretive statement you attach to a segment of your data. You’re not summarizing — you’re interpreting. You’re tagging information with a concise label that captures its meaning in relation to your research questions.

 Here’s a real example. I’m reading through a transcript about the experiences of women with immune treatments for fertility. One participant says:

“…myself and my husband started trying to have a baby in March 17, tried for a year before we went to our GP… had one round of NHS IVF, which wasn’t successful, so then went to another clinic in London which offers immune treatment…”

From this excerpt, I generate two codes:

• Tried to conceive for a year — attached to the first sentence
• Switched clinics for fertility treatment — attached to the second sentence

Notice I’m not summarizing what the participant said. I’m interpreting the meaning of their words in relation to the research objective. That is the core discipline of qualitative coding.

I’m using NVivo 15 — the current version as of 2025. The core coding process works the same way in NVivo 12, 14, and 15. For this tutorial, I’m working with a study on the opinions of physiotherapists on assessing the psychological well-being of stroke patients. If you want to go deeper into the inductive coding approach before continuing, my guide on inductive thematic analysis using NVivo (Saldana’s method) covers that in full.

Setting up your NVivo project:

1. Open NVivo and click New Project. Name it anything — I’ll call mine “Qual coding projects”
2. In the Files section, drag and drop your interview transcripts to import them. You can also go to Import → Files if you prefer
3. In the Codes section, create a folder called “Initial codes” to keep your work organized. It’s not mandatory, but it helps you track your work as you move toward theme development later
4. Open the first transcript by double-clicking it

How to color-code by interview question:

Before coding, I like to color-code the interview questions in the transcript so I can later see which codes came from which question. In NVivo, highlight an interview question and mark it red for question one, green for question two, blue for question three, and so on. This becomes very useful in Step 2 when you start grouping codes into themes.

Generating and applying codes:

1. Read through a full paragraph before generating any codes — understand the context before you label it
2. Highlight the key passage → right-click → select Code Selection → navigate to Initial Codes → Top Level Codes → type your code name
3. Make your codes interpretive, not just descriptive. “Patients reluctant to disclose mental health challenges” is a code. “Mental health” alone is just a topic
4. Assign a color to each code matching the question it came from — this is how you’ll group them into themes later

Some examples of codes generated from the physiotherapy study:

• Individuals with mental health issues might find it difficult to seek help
• Difficulties communicating
• Considers assessment of psychological well-being as part of scope of practice
• Psychological well-being assessment not currently provided to stroke patients

Coding across multiple transcripts:

When you move to the second transcript, look for passages that express the same meaning as codes you’ve already created. Instead of making a new code, highlight that passage, then drag and drop it inside the existing code in the Codes panel. NVivo automatically updates the file and reference count.

For example — when the second participant also mentions difficulties communicating, I drag that quote straight into the “Difficulties communicating” code. The counter updates from 1 file and 1 reference to 2 files and 2 references. If you double-click the code, you see both participant quotes side by side. That is your evidence base.

Key rule: Only code sections that are relevant to your research questions. NVivo tracks everything you code, so be selective and interpretive. We do not code everything — we code what matters.

Once your coding is complete, the next step is developing themes. If you’re using Braun and Clarke’s six-step thematic analysis framework, theme development is Step 3 — it comes after familiarizing yourself with the data and generating initial codes.

A theme is not a topic. A theme is a pattern of shared meaning across multiple codes. You’re looking for codes that — when grouped together — tell a coherent story in relation to your research questions.

I’ll demonstrate using a study on teachers’ challenges with alternatives to corporal punishment in secondary schools. I analyzed two interview transcripts and generated initial codes, color-coded by interview question.

Step-by-Step: Building Themes in NVivo

1. In the Codes section, create a folder called “Preliminary themes” and copy all your initial codes into it
2. Look at your color-coded codes — codes sharing the same color came from the same interview question and are likely related to the same theme
3. Right-click in blank space → Create New Code → name it as a theme. Note: this is now a theme, not a code, even though NVivo stores them together
4. Drag all related codes into the theme. These codes become the evidence base for that theme
5. Write a description for each theme — 1 to 2 sentences explaining what the theme represents. This is the analytical memo that will later become your findings write-up
6. Repeat until you have 3 to 7 well-defined, distinct themes

In this study, the five themes that emerged were:

• Stakeholders involved in disciplining learners

• Teachers’ views on the effectiveness of corporal punishment

• Common cases of indiscipline in school

• Causes of indiscipline among learners

• Strategies teachers use to discipline learners in schools

Colour Coded Codes

Important: When grouping codes into themes, some codes will not fit neatly anywhere. Do not force them. Leave them in a “miscellaneous” folder or discard them. Your themes should have clear, defensible shared meaning — not be a catch-all.

To see how theme development works using Saldana’s inductive approach instead of Braun and Clarke, I cover that in detail in my guide on inductive thematic analysis using NVivo.

Once you have your themes, the next step is writing the findings chapter. If you’re new to qualitative research and unsure how to plan your interviews before analysis, my guide on how to conduct a qualitative research interview is a good starting point.

A qualitative findings report has two main sections:

1. The data analysis process — a description of the steps you used to go from raw transcripts to themes. If you used Braun and Clarke’s inductive approach, you describe all six steps: familiarization, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report.
2. The findings discussion — a theme-by-theme write-up supported by participant quotes as evidence.

Exporting Your Data from NVivo

NVivo gives you several ways to export your analysis for use in your report:

• Export theme list: Select all themes → Ctrl+A → right-click → Export → Export List. This gives you an Excel file with all themes, codes, file counts, and reference counts.
• Export codebook: Go to Share → Export → Export Codebook. This exports your full coding structure with theme descriptions, subthemes, and codes — ready for your dissertation appendix.
• Participant quotes: Double-click any theme or code to access the exact participant quotes it contains. Copy these into your report as direct evidence, formatted in italics.

For each theme in your findings chapter:

1. Introduce the theme and explain what it represents (this is where your theme description from NVivo becomes your opening paragraph)
2. Present the supporting codes and explain what they show about the theme
3. Provide participant quotes in italics as direct evidence — these are the voices of your research participants
4. Interpret what the theme means in relation to your research question

Include NVivo visualizations where they add clarity to your write-up. Hierarchy charts and mind maps work particularly well as figures in findings chapters — you’ll see how to create both in Part 4 below. For a comparison of how NVivo and other tools handle the analysis and reporting process, see my guide on qualitative analysis of interviews with ATLAS.ti.

NVivo offers powerful visualization tools that can strengthen both your analysis and your findings report. Here are all nine methods available in NVivo 14 and NVivo 15.

A word cloud shows the most frequently used words by a participant or across your whole dataset. To generate one for a specific participant:

1. Right-click the participant’s transcript → Auto Code → select Speaker Name
2. Insert the speaker initials (e.g. I for interviewer, M for Michelle) → Next → create a new case classification (e.g. “Michelle input”) → Finish
3. Go to Case Classifications in the left menu → right-click Michelle → Visualize → Word Cloud
4. Export with Ctrl+Shift+E

The resulting word cloud shows what that participant talked about most. In my example, Michelle’s word cloud prominently features words like fashionable, temper, feelings, and emotional. This kind of visualization works well as a figure in your methodology chapter.

Use regular charts to visualize demographic data you collected from participants — age, gender, education level, and so on. Right-click any participant → Visualize → Chart Cases by Attribute Value.

If your chart shows “unassigned” or “not applicable” values, right-click the chart → Select Data → on the X-axis attribute, choose “All attributes except unassigned and not applicable” → click OK. This gives you a clean, professional chart.

Rename the axis labels (right-click → Edit Labels) to plain English: “Age of participants” instead of “name, age” and “Number of participants” instead of “number of matching cases.” Export with Ctrl+Shift+E.

Hierarchy charts show the relative weight of each code within a theme — how much data was attributed to each code. Right-click a subtheme → Visualize → Hierarchy Chart of Codes.

You get two views: the tree map and the sunburst. The sunburst is my preference. In my reflexive thematic analysis example, the sunburst clearly shows that the code “temperamental” has far more references than the others under self-perception. That’s not a value judgment — it means more participant data was coded to that concept. Export with Ctrl+Shift+E.

To create a table from your NVivo data, select all themes (Ctrl+A) → right-click → Export → Export List. This gives you an Excel file with your full coding structure, including file counts and reference counts.

The numbers in the Excel export mean: 1 = major themes, 2 = subthemes and codes. The Files column shows how many participants were coded to each theme. The References column shows how many quotes support each theme.

Copy this table into Word for your appendix. To create a bar chart from it, go to Insert → Charts → Bar → OK. An Excel pop-up appears. Paste your theme names into the category column, rename “files” to “number of participants” and “references” to “number of references,” paste your data, and close the Excel window. NVivo’s exported data now drives a professional bar chart in your findings report

Project maps show your codes and themes as a connected hierarchical diagram. Go to Explore → Map → Project Map. Name it after the theme you want to visualize (e.g. “Project Map Ethnicity”). Drag the theme into the map → click the plus sign → highlight all codes → drag them in. Click the diagram → Layout → Hierarchical for a clean, structured output. Export with Ctrl+Shift+E.

Mind maps show the full hierarchical structure of a theme — from major theme down to subthemes and codes. Go to Explore → Maps → Mind Map. Copy your theme and paste it in as the central node.

Use Child Idea to add nodes below another idea (e.g. subthemes under the main theme). Use Sibling Idea for nodes at the same level (e.g. multiple subthemes). Apply the Top Down layout from the left formatting panel for a clear reading. Export with Ctrl+Shift+E.

Concept maps show the relationships between different themes — how they connect or influence each other. Go to Explore → Maps → Concept Map. Create a shape for each major theme, then use the Connector tool to draw relationship lines between them. You can label the connectors to explain the nature of each relationship. This type of map works well in a theoretical framework or discussion chapter.

Comparison diagrams compare what two participants (or files) talked about. Go to Explore → Diagrams → Compare Files → select two participants → click OK. NVivo draws a Venn-style diagram showing codes that both participants mentioned in the centre, and codes unique to each participant on their respective sides.

In my example comparing Michelle and Sapan, the diagram shows they both discussed ethnicity, negative perceptions, and positive perceptions — but each also raised unique codes. This is useful for showing convergence and divergence across your participant group.

Matrix coding query tables show how much each participant contributed to each theme. Go to Explore → Matrix Coding Query → click the plus sign. Add your participants as rows, and your themes or subthemes as columns. Click Run Query.

Click the Column Percentage button to see the proportion of references each participant contributed. In my example: Michelle 34.9%, Sapan 20.23%, Karen 15.98%, Kashif 10.7%, David 18.8%. This does not mean some participants are more important than others — in qualitative research we don’t interpret percentages quantitatively. But this table is useful in your appendix to show the distribution of data across participants.

That covers all nine visualization methods in NVivo 14 and 15. For the parallel process using a different tool, see my guide on qualitative analysis with MAXQDA, which covers coding, theming, and reporting in the same way.

Frequently Asked Questions

Can I use NVivo for thematic analysis?

Yes — NVivo is one of the most widely used tools for thematic analysis. It supports both the Braun and Clarke six-step framework and Saldana’s inductive method. The coding, grouping, and theme development process maps directly onto what both frameworks require.

What version of NVivo should I use?

NVivo 15 is the current version as of 2025. The core coding and theming process is the same across NVivo 12, 14, and 15. The visualization features in Part 4 apply to NVivo 14 and 15. Check the Lumivero NVivo page for the latest version details.

How many codes should I generate before moving to themes?

There is no fixed number. For most PhD-level interview studies with 6 to 15 participants, you will typically generate 30 to 80 initial codes before grouping them. Focus on coverage — make sure every significant topic raised by participants has a code — rather than aiming for a specific count.

Should I code all transcripts before developing themes?

Yes. Code every transcript before you start developing themes. Theme development requires seeing the full picture across all your participants. Theming too early leads to themes that do not hold up once all the data is coded.

Can I use NVivo for focus groups as well as interviews?

Yes. The same coding and theming process applies to focus group transcripts. The main difference is that you code by group rather than by individual participant. For a comparison of how ATLAS.ti handles the same process, see my guide on qualitative analysis of interviews with ATLAS.ti. For more on qualitative research methodology, Scribbr’s guide to thematic analysis is a reliable external reference.

Key Takeaways

Code all transcripts first, then develop themes — never the other way around

Coding = tagging and interpreting important quotes from your transcripts. You are not summarizing — you are labeling meaning in relation to your research questions

Use color codes in NVivo to organize codes by interview question — this makes theme development visual and fast

A theme is a pattern of shared meaning across multiple codes. Every theme needs a written description explaining what it represents

Your findings report has two parts: the data analysis process (how you analyzed) and the findings discussion (what you found), supported by participant quotes

NVivo can export theme lists, codebooks, and participant quotes directly — these become the raw material for your report and appendix

Use at least two or three visualizations in your findings chapter. Hierarchy charts and mind maps are the most academically appropriate for a thesis or dissertation

Struggling with your ATLAS.ti thematic analysis? If you’re working through your qualitative data in ATLAS.ti and feeling stuck — on the coding, the categories, the themes, or writing it all up — I can take it off your plate completely. I offer a done-for-you qualitative analysis service for PhD students. You share your transcripts and research questions, and I handle the full ATLAS.ti analysis: coding, categorizing, theme development, analytic memos, and a written summary of your findings. I’ve helped more than 300 PhD students get their qualitative analysis done — on time, to examiner standard, and without the stress. Get in touch here to discuss your ATLAS.ti project.