| Rank | Tool | Owner (Aug 2026) | Runs | Best for | The catch |
|---|---|---|---|---|---|
| 1 | Whatever your university licenses | — | — | Support, training, a supervisor who can open your file | Access usually ends when your student status does |
| 2 | MAXQDA | VERBI (independent) | Windows, Mac | Mixed-methods projects; universities can obtain a free course licence | Its AI assistant is described as GDPR-compliant, which is a European frame, not a Canadian one |
| 3 | Taguette | Open-source community project | macOS, Windows, Linux, or self-hosted | Any project where the data must not leave your machine — and any budget of zero | Tagging and export only; no AI, no visual networks, no statistics |
| 4 | ATLAS.ti | Lumivero | Desktop and Web | Dense theory-building work and network visualisation | AI analysis is processed in the US or Europe; pairs with OpenAI |
| 5 | NVivo | Lumivero | Desktop | Large text corpora; the default in many Canadian faculties | The heaviest tool here, and the one students most often stop using after graduation |
The recommendation: use the package your university licenses, and install Taguette as well. MAXQDA is the pick if you are choosing freely and paying. That ranking will make more sense after the next section, because the market changed and most methods courses have not caught up.
The fact that reframes this comparison
Every methods seminar poses the choice as “NVivo or ATLAS.ti?” As of 2026, that is a question about one company.
ATLAS.ti’s own homepage carries the announcement: “Lumivero Acquires ATLAS.ti. We are thrilled to share with you that ATLAS.ti has been acquired by Lumivero, a global leader in data-intelligence software.” Lumivero is the company that publishes NVivo, and its own product navigation lists ATLAS.ti, Citavi, NVivo and XLSTAT together under research and QDA software.
Three consequences for someone choosing this week:
- “Switch to the competitor” is no longer a lever. If a licensing term, a price change or a cloud-migration decision pushes you off one product, the obvious alternative has the same owner.
- Comparative advice written before the acquisition describes a market that no longer exists. That includes most textbook chapters and most library guides you will find.
- Independence is now a real differentiator. MAXQDA (published by VERBI) and Taguette (open-source) are the remaining options outside the consolidated pair, and for some projects that matters more than any feature.
The three criteria that actually decide it
Feature lists do not decide this, because all four tools will let you code a thesis-sized set of transcripts. Three other things do.
1. What your university licenses
This outranks everything. An institutional licence brings campus workshops, a help desk, a supervisor who can open your project file, and peers who can unstick you at 11 p.m. Ask your library or research-computing service before you evaluate a single feature, and ask two follow-ups: whether the licence covers your own laptop or only lab machines, and on what date your access ends relative to your convocation.
MAXQDA advertises a free course licence route for universities, so if your department is teaching a qualitative methods course, that licence may already exist and nobody has told you.
2. Where your participant data goes
This criterion barely existed three years ago, and in Canada it is now the one most likely to produce an awkward question at your ethics review.
Every one of these tools has added AI features, and AI features mean your interview transcripts are processed somewhere. What the vendors publish:
- ATLAS.ti states that you can “choose where your AI analysis is processed – the US or Europe,” and that “ATLAS.ti pairs with OpenAI for AI-driven analysis.” Read that offer carefully: it is a genuine and unusually transparent choice, and neither option is Canada.
- MAXQDA describes its AI Assist as “powerful, transparent and GDPR-compliant” — a European data-protection frame, which is a real standard but not the one your Québec or federal privacy obligations are written against.
- Taguette is the outlier: “Taguette works both on your local computer (macOS, Windows, Linux) and on a server. When running Taguette locally, your data is as secure as your computer.”
The practical rule: your consent form is the binding document, not the vendor’s page. If you told participants their recordings would be stored on encrypted university systems and seen only by the research team, uploading those transcripts to a third-party AI service is a change to what you promised, and it is your Research Ethics Board’s call rather than yours — our guide to REB approval and TCPS 2 in Canada sets out who makes that determination. Any AI-assisted coding also falls under your graduate school’s disclosure rules, covered in what Canadian universities actually permit for AI in a thesis.
3. What you still have the day after you graduate
Your corrections, your first article and your job talk all happen after your student status ends. A project file you cannot open is a project you cannot revisit.
This is the same trap that catches people with reference managers, and for the same reason — our comparison of Zotero, Mendeley and EndNote for a Canadian thesis works through the ownership question in that category. In QDA the exposure is worse, because a coded project is months of irreplaceable interpretive work rather than a list of references you could rebuild.
Taguette’s export list is the benchmark to judge the others against: your whole project, the codebook, all highlighted quotes or those for a specific tag, and the highlighted documents. Before you commit to any tool, run a test export on week one and confirm you can read the output without the software.
The four tools, honestly
NVivo (Lumivero)
The default in a great many Canadian faculties, which is its strongest argument: your supervisor probably knows it and your library probably licenses it. It handles large text corpora comfortably and Lumivero now markets it around AI-assisted analysis. It is also the heaviest of the four to learn, and the one students most often abandon at graduation, which makes the licence-end-date question especially worth asking here.
ATLAS.ti (Lumivero)
Strong for theory-building and for visual networks linking codes, quotations and memos, and it runs both as a desktop application and on the web. Its published position on AI processing location is the most specific of the four, which is to its credit even though neither location is Canadian. Choose it if network visualisation is genuinely part of your analysis rather than a feature you admired in a demo.
MAXQDA (VERBI)
The pick if you are choosing freely. It is the strongest of the four for mixed-methods work — it carries quantitative text analysis and statistical tools alongside the coding environment — it offers transcription with or without a MAXQDA licence, and it has a free course licence route for universities. It also remains independent of the consolidated pair, which is now a genuine consideration.
Taguette (open-source)
Free, open, and deliberately small: import materials, highlight, tag, export. No AI, no networks, no statistics. For a master’s thesis with twenty interviews and a reflexive thematic analysis, that is not a limitation — it is the whole job, and the tool does not pretend to do the interpreting for you. It is developed in the open on GitLab and funded through donations, so treat it as a community project rather than a supported product: keep your own backups and run the export early. Its language options include French, which matters for a bilingual Canadian project.
On prices, and what we could not verify
We tried to read current prices from the vendors’ own pages in August 2026 and could not, so we are not going to reprint figures from a third party.
MAXQDA’s product page loaded but served no price as static text. Two pricing paths we tried on the Lumivero and ATLAS.ti sites returned 404s rather than price lists — one of them a 200-kilobyte “page not found” that looks like a real page until you check the status code. Lumivero was, at the time of writing, running a promotional pricing event on NVivo starting 18 August, which is itself a reason not to treat any figure you read as stable.
What we can tell you is more useful than a list price anyway: most Canadian graduate students never pay it. The commercial tools reach students through institutional licences, and the free option is free permanently. Ask the library first, and treat any quoted price as the fallback rather than the starting point.
The thing no tool does
None of these applications will analyse your data. They store codes, retrieve extracts and count things. Every decision that makes an analysis — what counts as a code, which codes cluster, what the central concept of a theme is, which quotation earns its place — is made by you, and it is made the same way whether you are working in a licensed commercial package or with a stack of highlighters. The procedure itself is set out in our step-by-step guide to doing a thematic analysis for your thesis, and it is worth reading before you choose a tool, because it will tell you which features you actually need.
If your project also has a quantitative arm, the equivalent decision on that side runs on different criteria again — see our comparison of SPSS, R, jamovi and Stata for a Canadian thesis.
Once the coding is done, the findings chapter still has to be written and kept consistent through every rewrite. Draft your findings chapter in Tesify — the structure and references stay stable while every interpretation remains 100% written by you.
Frequently asked questions
Which QDA software is best for a master’s thesis?
The one your university licenses, because support and a supervisor who can open your file matter more than features at thesis scale. Choosing freely, MAXQDA. On no budget, Taguette.
Are NVivo and ATLAS.ti made by the same company?
Yes as of 2026. ATLAS.ti’s own homepage announces its acquisition by Lumivero, which publishes NVivo, and Lumivero lists both products together in its research software range.
Is there genuinely free qualitative analysis software?
Yes. Taguette is free and open-source, runs on macOS, Windows and Linux or on your own server, and lets you import materials, highlight, tag and export.
Will AI coding features get me in trouble with my ethics board?
Possibly, if your consent form did not contemplate them. Sending participant transcripts to a third-party service is a change in how the data are handled, and that is a determination for your REB rather than for you.
Where does ATLAS.ti process my data?
Its site states that you can choose whether AI analysis is processed in the US or Europe, and that it pairs with OpenAI. No Canadian processing location is published.
Is MAXQDA GDPR compliance enough for a Canadian project?
It is a meaningful standard but it is not the standard your obligations are written against. Canadian institutional and provincial privacy requirements are separate, so confirm with your REB and your research-data-management service.
Can I get a free MAXQDA licence?
MAXQDA publishes a free course licence route for universities alongside its student offerings, so ask whether your department has one before buying anything.
What happens to my project when my student licence expires?
Typically you lose the ability to open it. Export your project, codebook and highlighted extracts in a readable format well before that date, and test the export rather than assuming it works.
Can I switch software mid-project?
It is painful but possible if you exported cleanly. Codes and highlights rarely transfer perfectly between packages, so the realistic plan is to finish in the tool you started in and export at the end.
Do I need software at all for twenty interviews?
No. Printouts, highlighters and a spreadsheet complete a thesis-scale analysis. Software helps with retrieval and organisation, not with interpretation.
Does using QDA software make my analysis more rigorous?
No. Rigour comes from the transparency of your procedure and your engagement with the data. Naming the software in your methods chapter without describing the procedure is a common and visible weakness.
Should I name the software in my methods chapter?
Yes, with its version, alongside a description of the analytic procedure. The version matters because features and defaults change between releases.