Students ask which statistics package is best. For a thesis that is the wrong question, because all four of these will run the analyses a typical graduate project needs. The right question is which one will still be open in front of you during corrections, eighteen months from now, when an examiner asks you to re-run a model with one variable removed.
Judged that way, the decision has four criteria and none of them is a feature list.
A note on prices
We checked each vendor’s own pages in August 2026. Neither IBM nor StataCorp publishes a plain numeric price on the pages we could read — SPSS leads with a free trial, and Stata presents pricing behind a licence-term selector rather than as static text. jamovi and R state clearly that they are free.
Rather than reprint figures from a third party, we have built this guide on the criteria that decide the outcome anyway. For most Canadian graduate students the commercial packages arrive through an institutional licence, which makes list price close to irrelevant — and makes the expiry date of that licence very relevant indeed.
The four criteria that decide it
| Criterion | Why it outranks features |
|---|---|
| What your supervisor can open | Someone has to check your analysis. If they cannot run your file, they cannot help you fix it |
| What your university licenses | Determines cost, support, and campus training you can actually attend |
| What your field publishes in | Reviewers read output conventions; matching them removes friction from your first paper |
| What survives graduation | Corrections, publications and job talks all happen after your student status ends |
The four packages, honestly
SPSS
Commercial, owned by IBM, and menu-driven. It remains the package most Canadian students first meet, because it is what a great many undergraduate methods courses in psychology, nursing, education and health sciences teach. IBM currently markets it with AI-assisted features and offers a free trial.
Choose it when: your department teaches it, your supervisor uses it, and your analyses are standard — descriptives, t-tests, ANOVA, regression, common non-parametrics. The point-and-click interface genuinely lowers the barrier when statistics is not your strength.
The catch: it is the package most likely to disappear when your licence does, and menu-driven work is the easiest to do irreproducibly. See the syntax section below, which matters most for exactly this tool.
Stata
Commercial, and strong in economics, epidemiology, political science, public health and quantitative sociology. Its licensing is unusually explicit about term: StataCorp offers 6-month, annual and perpetual licences, and states that all 6-month and annual licences give access to the latest features through StataNow.
That structure is worth understanding rather than skimming. A perpetual licence keeps working after your studies but freezes at the version you bought; a term licence tracks new features but stops when you stop paying. For a thesis student the term option usually wins during the degree; if you expect to publish from the thesis afterwards, price a perpetual licence into the decision before you graduate.
Choose it when: your field runs on it, or your analysis is panel data, survey-weighted estimation, or survival analysis, where Stata’s conventions are well established and its documentation is excellent.
R
Free software, and the environment most of contemporary statistical methodology is written for. New methods appear as R packages years before they appear anywhere else, which is why it dominates in statistics, ecology, bioinformatics and any data-heavy discipline.
Choose it when: your method is not standard, your dataset is large or awkward, you need reproducibility to be effortless, or you want a skill with value outside the university.
The catch: the learning curve is real and it arrives at the worst possible moment — the analysis stage, when your deadline is closest. Starting R in your final six months is a decision students regret. Starting it in your first term is one they rarely do.
jamovi
jamovi describes itself as “a free and open statistical spreadsheet, designed to be easy to use,” founded to develop a free and open statistical platform, built by researchers for researchers, with a cloud option available free.
Its significance for thesis students is specific: it presents an SPSS-like spreadsheet and menu interface while running on R underneath. That makes it the natural bridge for someone trained on SPSS who has lost access to it, or who wants their analysis to be reproducible without learning to program first.
Choose it when: you were taught SPSS, you no longer have a licence, and your analyses are conventional. It costs nothing and it will not expire.
Rough guide by discipline
Treat this as a starting point to check against your own department, not as a rule. Local convention beats any general mapping.
| Field | Commonly encountered | Free route if you lose access |
|---|---|---|
| Psychology, education | SPSS | jamovi |
| Nursing, health sciences | SPSS, Stata | jamovi, R |
| Epidemiology, public health | Stata, R | R |
| Economics, political science | Stata, R | R |
| Sociology | SPSS, Stata, R | R, jamovi |
| Biology, ecology | R | R |
| Engineering, computer science | R, Python | R |
The habit that matters more than the choice
Whatever you use: save the syntax.
Every one of these tools can produce a result from a sequence of clicks and every one of them can produce the same result from a saved script. Only the second is a thesis asset. SPSS has a syntax window and a paste button that converts your menu choices into commands; Stata has do-files and logs; R and jamovi are script-backed by construction.
The concrete argument, which lands better than any appeal to good practice:
- Corrections. An examiner asks you to exclude a subgroup and re-run. With a script that is four minutes. Without one it is an afternoon of trying to remember which twelve dialog boxes you used, and no guarantee you reproduce your own published table.
- Error recovery. When a number in your results chapter does not match your appendix — and at some point one will not — the script tells you which is wrong. Memory does not.
- Defensibility. “How did you handle missing data?” is a standard defence question. The honest answer is in your script, with a date on it.
- Continuity. A script runs after your licence lapses, on a different machine, in a different package if need be. Clicks do not port.
Store the script alongside your data and your codebook, and record which version of the software produced each table. Results can shift subtly across versions, and “SPSS” without a version number is not a reproducible statement.
Before you install anything
- Ask your supervisor what they use, and whether they will be checking your analysis files. This one answer settles most cases.
- Ask the library or IT what the university licenses, and — critically — when that access ends relative to your graduation date.
- Check what your target journals show. Open three recent papers in your subfield and look at the analysis section.
- Install the free option too. R or jamovi cost nothing and give you a fallback the day a licence server refuses you during your final week.
One last point specific to Canadian work. If your analysis runs on national statistical files, the access tier may constrain the software before your preference does — analysis inside a secure Research Data Centre happens in the environment that facility provides, not on your laptop. Our guide to Statistics Canada microdata sets out which tiers those are, and it is worth reading before you commit to a workflow.
Once the numbers exist, they still have to become a results chapter that an examiner can follow. Draft your results chapter in Tesify and keep tables, references and structure consistent while every interpretation stays 100% written by you.
Frequently asked questions
Is SPSS still worth learning?
If your department teaches and supports it, yes — the fastest route to a finished analysis is usually the tool the people around you can help you with.
Is jamovi as capable as SPSS?
For the analyses a typical graduate thesis needs, it covers the standard ground and runs on R underneath. Highly specialised procedures may require R directly.
Should I learn R for my thesis?
If you are early in your programme, yes. If your analysis starts in eight weeks, learn it after submission rather than during.
What licence term should I buy for Stata?
Term licences (6-month or annual) include access to the latest features through StataNow; a perpetual licence keeps working afterwards but at the version purchased. Match the choice to whether you will publish from the thesis after graduating.
Will my university give me the software free?
Many Canadian universities license at least one commercial package. Ask the library or IT service desk, and ask when access ends.
Can I switch packages mid-thesis?
Yes if you have saved syntax and clean data, because you are re-running a documented procedure. Without syntax you are reconstructing from memory, which is where errors enter.
Does the software choice affect my results?
It should not for standard procedures, but defaults differ — handling of missing data, contrast coding and estimation options vary between packages and versions. State your software and version in your methods.
Do I need ethics approval to analyse a dataset?
It depends on whether the data are anonymous and how they were obtained. See REB approval and TCPS 2 for the exemptions.
What should I cite for the software?
Cite the software with its version, following your programme’s required style. Your reference manager can hold these entries alongside your literature — see our comparison of reference managers.
Does a course-based master’s need any of this?
Only if your major research paper is quantitative. Whether your programme requires empirical work at all depends on its structure — see thesis, dissertation or major research paper.
What is the single most common mistake?
Running the analysis by clicking, publishing the table, and discarding the trail. Save the syntax from the first analysis you ever run.