You have twenty interview transcripts, a deadline and a supervisor who expects “themes supported by the data”. Where does NVivo fit, and what should you actually click? NVivo coding is the process of linking passages in your data to codes that represent ideas, so you can retrieve every passage about one idea, compare groups of participants and check whether a theme holds across the dataset. The software organizes and retrieves; the interpretation remains yours.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Statistical Analysis service. Sources last checked: November 2026.
This guide covers project setup, coding, the main queries and how to describe the process in your methodology chapter. Menu names follow the NVivo 15 help pages; older versions call codes “nodes”.
What is NVivo coding? In NVivo, a code is a container for references: each time you select a passage of text, an image region or an audio segment and assign it to a code, NVivo stores that reference. Codes can be arranged in a hierarchy (parent and child codes), and the same passage can be coded to several codes. Queries then search the text and the coding to show patterns.
Step 1: How should you set up the project before NVivo coding?
Decisions made before the first round of NVivo coding save hours later.
- Import files consistently. Use one transcript per participant, with the same speaker labels and file naming (for example P01_teacher_female.docx).
- Create cases. A case represents a unit of analysis, usually a participant or an organization. Linking each file to a case lets you compare people rather than documents.
- Add attributes. Attributes such as gender, years of experience or school type are stored in a classification and attached to cases. They are what makes later comparisons possible.
- Write a project memo. Record your research questions and coding decisions in a memo inside the project, dated as you go.
Step 2: Deductive, inductive or both?
Decide how codes will be generated before you start NVivo coding.
| Approach | Where codes come from | Suits |
|---|---|---|
| Deductive (a priori) | Your theoretical framework or interview guide | Studies testing or applying an existing model |
| Inductive (data-driven) | Ideas that emerge while reading | Exploratory studies, grounded theory |
| Hybrid | A starting codebook, extended by new codes | Most master’s theses |
Braun and Clarke’s (2006) thematic analysis, the method most theses using NVivo cite, starts with familiarization and initial coding before themes are searched for and reviewed. Clarke and Braun later stressed that themes are actively constructed by the researcher rather than “emerging” by themselves, which is worth reflecting in the way you write.
📱 Ask a statistician about your data — Send a short description of your instrument and sample size on WhatsApp and get a suitable analysis plan. Every analysis comes with the raw SPSS file and a results report.
Step 3: How do you code in NVivo, pass by pass?
Coding in one heroic pass rarely works. A more defensible routine:
- Familiarize. Read every transcript once without coding and write a short memo on each.
- First pass, open coding. Select a passage, then drag it onto a code or use the right-click menu to code to a new or existing code. Keep codes descriptive and short, such as “workload after reform”.
- Organize. Group related codes under parent codes. Merge duplicates. Write a description for every code in its properties so the meaning stays stable.
- Second pass. Recode all files against the refined codebook, especially the early ones coded before the codebook settled.
- Check with coding stripes. Coding stripes show, in the margin of a file, which codes are applied where; they quickly reveal uncoded sections and inconsistent coding density.
A codebook export (code names, descriptions and number of files and references) makes a useful appendix and shows the examiner that NVivo coding was systematic.
Step 4: Which NVivo queries should you use?
Queries are where NVivo coding turns into findings. The NVivo 15 help describes several query types; these are the ones most theses need:
| Query | What it answers | Typical use in a thesis |
|---|---|---|
| Word frequency | Which words occur most often (stop words excluded) | Early familiarization, possible themes, a word cloud for a presentation |
| Text search | Where a word or phrase appears, optionally with stemmed forms | Checking that a concept was coded everywhere it appears |
| Coding query | Which content is coded to a combination of codes and attributes | Retrieving everything coded “workload” from experienced teachers only |
| Matrix coding | Intersections between two lists of items (codes, files or cases) | Comparing themes across participant groups |
| Crosstab | How coding is spread across codes and cases or case attributes | Counts and percentages of a theme by gender or school type |
| Coding comparison | Agreement between two coders (percentage agreement and Cohen’s kappa) | Reporting intercoder agreement when two people code |
A matrix coding query, for example, can place your main themes in rows and attribute conditions (novice and experienced teachers) in columns. Each cell holds the content coded to both, and double-clicking a cell opens those passages, so you can read what lies behind the numbers.
Step 5: How do you interpret query results without over-counting?
Counts produced by NVivo coding in crosstab or matrix queries describe how often a code was applied, not how important an idea is. A theme mentioned briefly by fifteen participants and one discussed at length by four may both matter. Use the numbers to locate patterns, then read the passages and quote them.
For coding comparison, NVivo’s help notes that kappa can be low even when percentage agreement is high, for instance when two coders each code small, different sections and leave most content uncoded. Report both figures and explain how disagreements were resolved.
Step 6: How do you write up NVivo coding in the methodology chapter?
Examiners want the process, not a software tour. A typical scenario of a methodology paragraph (not from a real study):
Transcripts were imported into NVivo 15 and linked to participant cases with attributes for gender and years of experience. Following Braun and Clarke (2006), coding proceeded in two passes: open coding generated 46 codes, which were refined into a codebook of 28 codes under 6 parent codes. Matrix coding queries compared themes between novice and experienced teachers. A second coder coded 20% of the transcripts; disagreements were discussed and the codebook revised.
Name the software version, the analytic method, the number of coding passes and how themes were reviewed.
What are the common mistakes in NVivo coding?
- Treating word frequency results as themes.
- Coding single words instead of meaningful passages, which strips context.
- Creating hundreds of codes with no descriptions or hierarchy.
- Skipping cases and attributes, then having no way to compare groups.
- Reporting kappa without saying which files and codes it covers.
If you want a second reader for your qualitative analysis plan or codebook, the Mastermind PhD team tells you the specialist’s field and degree before starting and schedules a meeting to agree the scope; see the statistical analysis service page for the scope of analysis support.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
Does NVivo do the analysis automatically?
No. NVivo coding is done by the researcher. The software stores, retrieves and cross-references your coding; features such as word frequency and autocoding only support the reading you still have to do.
Should I code in NVivo or in Word first?
Either works, but coding directly in NVivo keeps every reference linked to its source, which makes queries and the audit trail possible. Coding in Word and re-entering it later doubles the work.
How many codes are normal for a master’s thesis?
There is no standard number. What matters is that every code has a description, related codes are grouped, and the final themes are traceable to the coded passages.
Is intercoder agreement required?
Not always. Some qualitative approaches, including reflexive thematic analysis, do not require it. Follow your method and your program’s expectations, and justify the choice.
What is the difference between a node and a code?
They are the same thing. Older NVivo versions used “node”; newer versions use “code”.
Sources
- Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. https://doi.org/10.1191/1478088706qp063oa
- Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589-597. https://doi.org/10.1080/2159676X.2019.1628806
- Lumivero. Matrix coding query, NVivo 15 Windows help. https://help-nv.qsrinternational.com/15/win/Content/queries/matrix-coding-query.htm (accessed November 2026)
- Lumivero. Crosstab query, NVivo 15 Windows help. https://help-nv.qsrinternational.com/15/win/Content/queries/crosstab-query.htm (accessed November 2026)
- Lumivero. Coding comparison query, NVivo 15 Windows help. https://help-nv.qsrinternational.com/15/win/Content/queries/coding-comparison-query.htm (accessed November 2026)
- Lumivero. Word frequency query, NVivo 15 Windows help. https://help-nv.qsrinternational.com/15/win/Content/queries/word-frequency-query.htm (accessed November 2026)
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.