Learning how to use NVivo basics for qualitative analysis takes about an afternoon. Create a project, bring in your transcripts, build a small code list, tag the passages that matter, and export your evidence. That core workflow covers most dissertation and thesis work.
NVivo structures and organises your material. It does not interpret it for you, and it will not hand you a finding. Researchers on PhD and social science forums describe it as intimidating on day one and straightforward once a handful of buttons become familiar. The trick is learning five of them, not all of them.
The workflow below follows three phases: prepare and import your data, code your data, then explore, query and visualise. Every menu path below refers to the current ribbon layout in the Windows and Mac releases. Plenty of tutorials online still show the older NVivo 10 and 11 menu bar, where you found things under an Analyze menu. That was the Analyse and Code Selection at Existing Node era, and it is no longer how the software is arranged.
Table of Contents
- 1What You Need
- 2Step-by-Step
- 3How to Use NVivo Basics: Create a Project and Prepare Data
- 4How to Use NVivo Basics: Add and Organize Source Material
- 5How to Use NVivo Basics: Start Initial and Focused Coding
- 6How to Use NVivo Basics: Build and Classify a Code Hierarchy
- 7How to Use NVivo Basics: Code a Complete Interview or Document
- 8How to Use NVivo Basics: Use Memos and Attribute Notes
- 9How to Use NVivo Basics: Review Code Frequencies and Extract Examples
- 10How to Use NVivo Basics: Create a Basic Theme or Query View
- 11How to Use NVivo Basics: Check Reliability and Save the Project
- 12Common Mistakes
- 13Frequently Asked Questions
- 14Is NVivo suitable for beginners doing qualitative analysis?
- 15What is the difference between a code, a category, and a theme in NVivo?
- 16Should every interview participant be created as a case?
- 17How many codes are enough for a qualitative research project?
- 18Can I use NVivo for thematic analysis?
- 19How do I export quotations and coded text from NVivo?
- 20Conclusion
What You Need
You need four things ready before you open the software, and none of them is a new computer.
- Access to NVivo. Most universities hold an institutional licence, and it often covers every student and staff member on campus through a single sign-in. Ask your library or research support office before buying anything, because a free trial also exists if your institution has nothing.
- Your source material. Clean transcripts, field notes, open-ended survey responses, PDFs, audio files, video clips or photographs. One DOCX or RTF transcript per participant works well for a first project.
- A written research question. You cannot tell whether a passage is worth coding without knowing what you are coding for.
- A starting coding framework. Even a rough one. A dozen themes drawn from your literature review or interview guide saves you from coding everything that moves.
Do the prep outside NVivo. Convert scanned documents to text or audio you can search, remove names you do not need from transcripts, and give every file a consistent name such as P01_interview. NVivo will happily hold a messy project, but the mess shows up later in your queries and your exports.
Step-by-Step
How to Use NVivo Basics: Create a Project and Prepare Data
Open NVivo, choose File then New, and give the project a short descriptive name with no spaces if you plan to script anything later. The project file is local to your machine, so put it in a folder you back up rather than on a desktop you will wipe in April.
Before importing, decide whether this project will hold one strand of data or several. Beginners often try to load interviews, survey exports and a literature corpus into one project, then lose a day untangling it. Keep the scope to your main data sources and add the rest later if a reviewer asks for it.
How you know it worked: the project title appears in the title bar and an empty project opens in the main window with no error message.
How to Use NVivo Basics: Add and Organize Source Material
Go to Data then Import then Files, and select a folder or a selection of documents. NVivo converts what it can into searchable text. Word documents and text files import cleanly, PDFs import with their text layer intact, and audio and video import as media you can code by time span or by frame.
Watch what happens with PDFs and scans. A PDF made from a scan may look like a document but contain no readable text, which means searches and queries will return nothing. Test one file first by opening it and pressing Ctrl+F on Windows or Cmd+F on Mac. If the search cannot find a word you can see on screen, that file needs OCR before it enters the project.
Rename files inside NVivo if you need friendlier labels, and group them into folders if you have more than about twenty. Interviewers get separated into a Files folder, observation notes into another. A researcher on a sociology forum described colour-coding folders and nodes as the single trick that kept a 40-interview project navigable at month four.
How you know it worked: every file appears in the Files pane, opens on double click, and a text search finds a word you can see in the document.
How to Use NVivo Basics: Start Initial and Focused Coding
A code is a short label attached to a passage of text that relates to your research question. Initial codes come straight from the data and stay close to the participant’s own words. Focused codes group those initial codes into broader categories you can defend in a write-up.
To create one, highlight a sentence or paragraph, then use Coding then Codes then Add to New Code, or drag the selection onto a code you already have. NVivo creates the code, files the selection under it, and adds a coloured stripe in the margin so you can see at a glance what got tagged.
Code in vivo is the other route. Highlight a phrase the participant used that carries weight, such as working around the system, and turn that phrase itself into a code. Interviewers love it because the label stays in the participant’s voice, and later you can search for that phrase across every other transcript.
Work selectively. Code what speaks to your research question and leave the rest. Coding everything is the most common beginner habit and the fastest way to end up with two hundred shallow labels and no argument.
How to Use NVivo Basics: Build and Classify a Code Hierarchy
Once you have fifteen or so initial codes, stop coding and sort them. Create a parent code such as Institutional response, then drag the related children under it: Delayed response, Informal workaround, Silent compliance. In the Codes pane, codes and files live in two separate trees, and beginners often confuse them. Codes are your labels. Files are your material.
Drag and drop in that tree is the whole mechanic. Drag a code onto another code to make it a child, and drag it out again to promote it back to the top level. Doing this a week after coding is normal, because the shape of your argument usually appears only once you have seen the material.
Use classifications when a grouping is not a hierarchy. Cases get case classifications, files get file classifications, and each member carries attributes you can filter on later. A case classification with Gender, Role and Site fields lets you ask which themes appear in which group without reading a single transcript again.
How to Use NVivo Basics: Code a Complete Interview or Document
Try a small end-to-end exercise on one transcript. Read it once without coding and jot down anything that surprises you. Read it a second time, coding the passages that relate to your question, and expect twenty to forty codes from a one-hour interview.
Then run a coding summary on that single file. The summary shows every code applied to the document, how many references each one has, and how much of the text it covers. It is the fastest way to spot a code you applied to a single stray word and decide whether it deserves to stay.
How you know it worked: the transcript shows continuous coloured stripes down the left margin where coded passages sit, and clicking a code shows you every excerpt you assigned to it.
How to Use NVivo Basics: Use Memos and Attribute Notes
Memos are where your thinking lives, and they are the feature beginners skip and then miss most. A memo attaches to a code, a case, a file or a selection, and it holds your reasoning: why you made a coding decision, what an early theme might mean, what a quiet participant implied rather than said.
Attach memos while coding, not afterwards. The reasoning is freshest at the moment you tag the passage, and reconstructing why you coded something three weeks later is guesswork. The Journal, reached from the Home tab, works differently: it is a running diary for the project as a whole, useful for recording sampling decisions, changes to the interview guide or a supervisor’s advice.
Two small habits help. Give every memo a title that states the idea rather than the activity, so Cost overruns are framed as a virtue beats Memo 14. And attach memo links to the codes you wrote about, so the chain from evidence to interpretation stays traceable when an examiner asks how you got there.
How to Use NVivo Basics: Review Code Frequencies and Extract Examples
Run a coding summary and read it sceptically. A code applied forty times may mean a theme is central, or it may mean you coded a single word that appears throughout. Frequency is a pointer, never a finding. Ask what the passages actually say before you put a number in your write-up.
Pull the coded text into a view. Any code you select can be opened as a list of all its excerpts, grouped by file or by reference, which is how you gather quotations for a results section. Keep quotations short and load-bearing. Three good lines beat a page of transcript that no reader will finish.
Look for the negative cases as well. Search your corpus for a theme you expected and check where it is absent or contradicted. Researchers repeatedly say the counter-examples shape the argument more than the confirming ones do, and finding them takes minutes inside NVivo rather than hours of rereading.
How to Use NVivo Basics: Create a Basic Theme or Query View
A text search query finds a word or phrase across everything you have imported and lets you turn matches into codes automatically. Use it when you suspect a term matters but have not decided what to call the code. A word frequency query shows the words that cluster in your corpus, which is a fast sanity check on whether your documents are actually about what you think they are.
A matrix coding query is the one worth learning properly. You pick rows and columns, such as themes against cases or themes against classifications, and NVivo fills a grid showing where each theme appears. That grid is where you spot a pattern held by four of six participants and one outlier, and it is the basis of most thematic analysis write-ups.
For visualisation, the hierarchy chart draws your code tree so you can see structure at a glance, and a word cloud gives you a rough sense of vocabulary, though size is frequency only and says nothing about meaning. Treat both as conversation starters for your own analysis, not as evidence.
How to Use NVivo Basics: Check Reliability and Save the Project
Before you call the analysis finished, go back through the coded excerpts and recode the ones that bother you. Move passages that landed in the wrong code, merge two codes that turned out to be the same idea, and delete the codes with a single reference that never made it into your framework.
Document the codebook. For every code, record its definition, why it exists, an example excerpt and a note on when not to use it. Export the code list with references so you have a version your appendix can quote.
If someone else coded part of your data, compare codes rather than trusting a percentage. Coding the same three transcripts independently and discussing the disagreements is a practical reliability check, and it tends to surface definitional problems that no statistic would.
Finally, save, close, reopen and export. Run the retrieval of your main themes so you can see your evidence as a reader will, and keep at least two copies of the project file in separate places. Projects get corrupted, laptops get stolen, and NVivo cannot recover a file it never wrote.
Common Mistakes
Coding everything. The fix: write your research question at the top of a blank page and code only against it. If a passage could appear in any study, it belongs to your background notes, not your code list.
Building one giant flat list of codes. The fix: stop at fifteen initial codes and sort them into parents before coding the next document. A hierarchy you build in week one is worth more than a perfect list you never restructure.
Treating a quotation as an interpretation. The fix: keep quotes and claims in different places. Store the quotation in NVivo, write the interpretation in a memo linked to the code, and the evidence chain stays inspectable.
Trusting code frequency as importance. The fix: every number gets opened and read. A frequent code often just reflects a common word, and the rare quotation that contradicts your theme is usually the one that earns you marks.
Ignoring negative cases. The fix: run a matrix coding query, then deliberately search for where a theme is missing or framed differently. Note the disagreement in a memo rather than quietly dropping the participant.
Saving in one place and one format. The fix: keep a dated copy after each analysis phase and export your tables and coded text as a second record. It takes seconds and it has rescued more than one dissertation.
Copying tutorial menu paths blindly. The fix: check the version label in your own title bar first. Paths from NVivo 10 and 11 tutorials land you nowhere in the current ribbon, and search the interface instead of assuming the function is gone.
Frequently Asked Questions
Is NVivo suitable for beginners doing qualitative analysis?
Yes. Most beginners cover their whole project with four operations: create a project, import files, build a small code tree, and code selections of text. Queries, visualisations and advanced automation can wait months. Researchers with no prior coding experience report becoming productive within a single afternoon, as long as they resist the urge to learn every button in the ribbon before starting.
What is the difference between a code, a category, and a theme in NVivo?
In NVivo all three are stored as nodes in the same tree, but they play different roles. A code is a label attached to a passage of text. A category is a parent code that gathers related codes beneath it. A theme is an analytical concept you build from categories and codes, usually documented with a memo. The software treats a theme as a node like any other; the difference lives in your writing, not in the tool.
Should every interview participant be created as a case?
Usually yes, if your study compares people. A case gives you somewhere to record attributes such as role, site or demographic group, and it lets matrix coding queries compare themes across participants. If you are analysing documents rather than people, such as a corpus of newspaper articles, make each article the case instead. Cases organise who or what; codes organise the ideas.
How many codes are enough for a qualitative research project?
There is no correct number, and anyone giving you one is guessing. A workable first pass on a one-hour interview produces roughly twenty to forty initial codes, which you then group into a smaller set of categories and themes. The test is whether each code has a clear definition and appears in more than one place in the data. A code used once is usually noise.
Can I use NVivo for thematic analysis?
NVivo supports thematic analysis well, because the method needs organised codes, retrievable evidence and a clear trail from data to write-up. Map it onto the software by coding excerpts to themes, grouping codes into categories, running matrix coding queries to check theme coverage, and attaching memos that explain each theme. NVivo does not perform the analysis, so the interpretation and the quality of your writing still carry it.
How do I export quotations and coded text from NVivo?
Open the code, case or query you want, choose Retrieve, then Export from the results window. You can save coded text as a Word document, spreadsheet or text file, and the coding summary exports as a table listing each code with its reference count. Export before you reorganise your codes, since retitling a code changes the label on every quotation underneath it.
Conclusion
Start with one small document set, three transcripts at most. Build a code list of about twelve, code one interview end to end, and read the coding summary before you touch a second one. Sorting those codes into parents early is what stops the project turning into a list nobody can navigate.
Attach a memo every time you make a judgement call, keep your project file in two places, and export your evidence as you go. If your institutional licence runs out, ask about the free trial. 2026 moves fast in qualitative work, and the researchers who get clean results are the ones who documented their decisions while they still remembered making them.
