The first time I saw jamovi was in a tutoring session a year ago. A PhD student showed me her screen with a mix of pride and apology: "My supervisor says this isn't serious, but I've been doing all my analyses here for two months and nothing bad has happened". I asked her how it felt. She said: "For the first time, I understand what I'm doing". That made me pay attention.
In 2026, jamovi has become the silent conversation in many PhD offices in psychology. Students switching in secret because their supervisor is "an old-school SPSS person". Junior researchers defending it at conferences. Master's programs in clinical psychology already teaching it as the first software. And, at the same time, a very real generational resistance among senior faculty who see it as a toy. This article is what I tell every PhD student who asks me whether they should use it for their doctoral thesis.
What is jamovi and why does it matter in 2026
Jamovi is a free, open-source statistical software developed by a team that originally came from JASP. Under the hood it uses R, which means two important things: first, the analyses it runs are the same ones you would do in R with packages like lavaan, psych, or lmerTest; second, any analysis you run in the GUI generates reproducible R code that you can copy and paste.
This is exactly what was missing from the ecosystem. SPSS gives you a friendly interface but locks you into a proprietary system that is expensive and has serious limitations for modern psychometrics. R gives you all the power but demands a learning curve that many people simply cannot afford in the middle of their thesis. Jamovi sits in the middle: free, easy, with results in APA format by default, and with the real possibility of learning R alongside if you want to go further.
In the last year I've seen more manuscripts submitted to Q1 journals stating "analyses were performed in jamovi 2.7" without any reviewer batting an eye. The era when you had to justify using anything other than SPSS to avoid being looked at strangely is ending.
What jamovi does very well
Let me start with the good, because there is plenty. Then I'll be equally honest about its weak points.
APA-formatted results ready to copy. When you run a t-test, ANOVA, or correlation, the tables jamovi produces follow APA guidelines almost without retouching. This doesn't seem important until you've spent three hours reformatting an SPSS table so your thesis committee doesn't make comments. It's the detail that saves you the most time fastest.
Factor analysis with modern metrics. The exploratory and confirmatory factor analysis module (based on R's psych package) calculates McDonald's omega alongside Cronbach's alpha by default. In 2026, any questionnaire validation reported with alpha alone is going to get reviewer comments. Having omega calculated out of the box is a real advantage over SPSS, where it requires an additional macro or a switch to R.
Accessible linear mixed models. The GAMLj module lets you fit mixed models through a GUI using lmerTest under the hood. For longitudinal designs or repeated measures with unbalanced data, this is huge. SPSS has MIXED but the syntax is tedious and the results unfriendly; jamovi gives it to you almost ready.
Integrated Bayesian analyses. If you want to add a Bayes factor to your mean comparisons or include Bayesian analyses as a complement, jamovi does it in a couple of clicks thanks to the jsq module (based on R's BayesFactor). This is something that more and more journals are requesting as a complement to the p-value.
Real reproducibility. The .omv file stores data, analyses, and results in a single container. If your supervisor says "redo that ANOVA but excluding participants under 30", you open the file and the results update in real time. And if you want to share your analysis with a coauthor, you send them the .omv and they will see exactly what you saw. This is something SPSS doesn't do well and that R requires discipline with scripts and seeds.
Where jamovi falls short
Here comes the part that many passionate defenders don't tell you. Jamovi is not the universal solution. There are analyses where it falls short and you need to know that before committing to it for your thesis.
Complex SEM is still limited. The SEMlj module covers basic structural models but if you need models with robust bootstrap, categorical indicators with WLSMV, or multilevel structural models, you'll hit walls. For a classical CFA-based questionnaire validation, it's enough. For more sophisticated models, you still need lavaan in R or Mplus.
Macros and very specific analyses. If your thesis depends on Hayes' PROCESS for moderated mediation with many moderators, or on very specific SPSS macros, jamovi has alternatives but they're not identical. The results will be equivalent but the transition can spark discussions with your supervisor if they're used to a specific output.
Massive datasets. For small or medium-sized datasets (typical in psychology), jamovi handles it easily. If you work with datasets of hundreds of thousands of rows, you'll notice slowness and eventually prefer to load the data directly into R.
Official documentation still scattered. The community is active but smaller than SPSS or R. When you get stuck on something specific, finding the answer can take more time. Stack Overflow has relatively few questions tagged jamovi compared to R.
The realistic workflow I recommend to PhD students
After accompanying several PhD students through the transition to jamovi (some partial, some complete), this is the workflow that works best.
Phase 1: descriptive analyses and exploration. Always start here in jamovi. The interface makes you look at the data, identify outliers, understand the distributions. It's educational and fast. In SPSS, many PhD students skip this phase because the laziness of configuring so many menus pushes them straight into the main analysis.
Phase 2: main analyses (t-tests, ANOVA, correlations, multiple regression). Jamovi is clearly superior to SPSS in terms of output clarity. The tables come out almost ready for your results section. Always activate the effect size and confidence interval options, which are one click away.
Phase 3: basic psychometrics. CFA, reliability with omega, exploratory factor analysis. Jamovi excels. It reports omega and alpha in the same table, gives factor loadings in clean format, and the fit indices with their standard cutoffs.
Phase 4: advanced analyses. Here you decide. If your model is a classical CFA, mediation with three variables, or a mixed model with a couple of random effects, jamovi handles it. If you need something more complex (SEM with categorical variables, multilevel models with many levels, network analysis), consider going out to R or complementing with another software.
This staggered workflow has a hidden advantage: it lets you learn the software gradually. You start with the easy stuff and, as you need more sophisticated analyses, you decide whether to level up in jamovi or jump to R. Many theses are completed 90% with jamovi alone, and the remaining 10% is delegated to a colleague or consultant for the most demanding analyses.
How to report it in your thesis without being looked at strangely
The most frequent question PhD students ask me is: "do I write in my thesis that I did it with jamovi or do I make up that it was SPSS?". The answer is clear: write that you did it with jamovi, exactly as you did. Transparency is always the best policy and, moreover, modern Q1 journals value open-source tools.
The standard format for reporting it in methods is as follows:
"Statistical analyses were conducted in jamovi 2.7 (The jamovi project, 2024), which uses R 4.4 for computation. Confirmatory factor analysis was performed using the SEMlj module based on the lavaan R package (Rosseel, 2012). Reliability was estimated with the Reliability Analysis module using McDonald's omega and Cronbach's alpha."
Cite the exact version. Cite the R packages that jamovi uses underneath when you know them (lavaan, psych, lmerTest, GAMLj). This is exactly what's expected in a transparent analysis report and shows that you know what you're using.
When I do NOT recommend it
Finally, be honest with yourself. Don't use jamovi if:
1. Your supervisor is going to make life hell for using anything other than SPSS. If the political battle will wear you down more than learning SPSS, use SPSS. Theses are not won in software arguments.
2. Your university only teaches SPSS and you'll need to ask for help constantly. In this case use SPSS to avoid fighting the ecosystem, but consider jamovi for future derived publications.
3. Your main analysis is something very specialized that jamovi doesn't cover well. If you need Bayesian SEM or advanced latent class analysis, jamovi will frustrate you. Go directly to R or Mplus.
4. You're going to use lots of nonparametric tests with specific corrections. SPSS and R have better coverage here.
What's coming in the next few years
The jamovi team remains active and the modules available have grown substantially since 2024. What interests me most for 2027 and 2028 is the integration with other open science tools: preregistration in OSF directly from jamovi, export to markdown for reproducible reports, plugins for psychopathology network analysis that are already in beta. This is going to make the "jamovi or not" decision less of a decision and more of a norm.
If you're in the middle of your PhD and still deciding which software to use, I recommend trying it out: spend an afternoon recreating your master's thesis analyses in jamovi. If nothing breaks in that afternoon, jamovi can probably be your main tool. If you hit something it can't handle, you already have useful information about which fragment of your thesis will need another tool.
And if you want us to review together whether jamovi is viable for your specific design, that's exactly the kind of question I work on in my statistical advising sessions for PhD students. Better to decide before starting the analysis phase than in the middle of it.