Free point-and-click statistics on your computer: jamovi, JASP, NuriStat and GNU PSPP

jamovi keeps preparation, analysis and the file you hand over in one place. JASP suits data that is already clean when you want classical and Bayesian results side by side. NuriStat offers SPSS-style menus with more setup; GNU PSPP suits SPSS-style syntax on GNU/Linux.

PRACTICAL PICKSPreferredPickOur Preferred PickPrepare and analyse without code$0

jamovi

A spreadsheet beside live results, with computed columns and filters, common analyses built in and one file that holds the data, settings and results for whoever comes next.

Use an updated Solid release: 2.7.41 includes the October certificate update. The Current line changes almost weekly. Usage statistics are sent when it connects, and cannot be switched off in the app.

Why jamovi is our Pick
PRACTICAL PICKSRisingStarOur Rising StarSPSS-style menus and variable metadata$0

NuriStat

A free Python desktop program with computed and recoded columns, common tests, survival and agreement methods, and editable project files, according to its documentation.

Expect Python/source installation on Mac, conflicting Windows installer instructions and less complete security documentation.

Why consider NuriStat

Who is this article for? People with a nonsensitive table, such as survey responses, course results or a business export, who want to clean it and run common analyses on their own computer without writing code, then pass the work to someone else. Not advice on choosing a statistical method.

About 27 min read

What Preferred Pick, Rising Star and Ready to Rock mean

The quick list

Choose jamovi for preparing and analysing a table, JASP for clean data and Bayesian results, NuriStat for SPSS-style variable metadata if you can manage the installation, or PSPP for SPSS syntax on GNU/Linux. All four desktop programs are free. Compare what each lets you do to the data before analysing it, what you can hand over and what connects to the internet.

jamovi, JASP and GNU PSPP are free to download and use, checked October 19, 2026; NuriStat offers free MIT source, checked October 28, 2026. Optional costs sit outside them: jamovi Cloud’s Priority plan is priced by region and Teams by quote; JASP’s AI assistant uses your own provider key, which may charge per use; JASP Services BV sells support plans. We could not confirm comparable current prices for BlueSky Statistics, SPSS, Stata or JMP. Licenses: AGPL3 with some GPL2+ parts (jamovi), AGPLv3 interface and GPLv2 engine (JASP), GPLv3 or later (GNU PSPP).

The comparison at a glance

Compare what each program lets you do to the data, what you keep and hand over, and the platform, privacy and security conditions that come with it.

jamovi leads for preparing and analysing a table without code; JASP follows for clean data and Bayesian results, NuriStat for SPSS-style menus with more setup, then PSPP for SPSS syntax on GNU/Linux. All tools remain selectable.

Free desktop statistics programs for point-and-click analysis
FeaturejamoviJASPNuriStatGNU PSPP
Best fitPreferred Pick: prepare and analyse a table without code, then hand over one fileAlso consider: data already clean, classical and Bayesian results side by sideRising Star: SPSS-style variable metadata, menu analyses and editable projects; more installation workSPSS-style syntax and data files, mainly on GNU/Linux
Desktop costFree for any use; optional paid Cloud plansFree; optional AI uses your own provider key, which may chargeFree MIT source; Python and scientific-library dependencies; no paid recipient licence shownFree; no license fee, expiry or paid add-ons
Preparing data in the programComputed and transformed columns, row filters; reshaping and merging through modulesRecode, filter and computed columns by menu or R; an independent reviewer rates its data management minimalCompute, recode, bin, rank, select and weight cases; spreadsheet filtering and variable metadataTransformation commands; several datasets can be merged, joined or concatenated; exact dialogs not verified
Built-in analysesDescriptives, t-tests, ANOVA, regression, frequencies, factor analysis; 70+ library modulesClassical and Bayesian t-tests, ANOVA, regression and frequencies; online module library since March 202627 documented modules: common tests, ANOVA, regression, survival, ROC, ICC and agreement methodsDescriptives, t-tests, ANOVA, linear and logistic regression, factor, cluster, non-parametric; no modules
Saved workOne .omv file: data, analyses, options and resultsOne .jasp file: data, options and results; optional password encryptionComplete session in .swb: ZIP, Parquet data and JSON metadata, per maker; cross-version reopening unknownData file plus syntax; command journal on by default; no one-file project verified
Desktop platformsWindows, macOS, Linux; Chromebooks via Flathub; current minimum versions not establishedWindows 10 22H2 or current Windows 11; macOS 15 or later; Linux Flatpak; Chromebooks from 2019Windows 10/11, macOS and Linux per release notes; Python 3.10+ source installation; installer instructions conflictLinux distribution packages; volunteer Windows and Mac builds, which the project ranks low priority
Internet connectionsFiles stay local; welcome screen, updates and modules connect and log usage statistics you cannot switch off in the appRuns without internet; collects no user data, per JASP; optional AI sends requests to the provider you configureLocal project files and public source; exact connections and installed dependency versions not documentedNo desktop privacy statement found

Desktop statistics for nonsensitive tabular data; we do not recommend statistical methods. Public-source research checked October 19, 2026, with NuriStat, release and security updates checked October 28: no installs, imports, analyses or AI use by us. All four desktop programs are free. Keep the original data and a note of exclusions and recoding beside the project file, and rerun important results after upgrading. Next substantive review: December 18, 2026, or sooner after a major release or security change.

All features and sources 16 criteria

Open the groups that matter to you. They compare the same selected tools. “Planned” means future work, not an available feature. Sources include the published conditions.

Preparing and analysing data

Show 5 criteria
Preparing and analysing data
FeaturejamoviJASPNuriStatGNU PSPP
Best fitPreferred Pick: prepare and analyse a table without code, then hand over one fileAlso consider: data already clean, classical and Bayesian results side by sideRising Star: SPSS-style variable metadata, menu analyses and editable projects; more installation workSPSS-style syntax and data files, mainly on GNU/Linux
Opens directlyCSV, Excel, SPSS, SAS, Stata and R data filesCSV and text, Excel, OpenDocument, SPSS, SAS, Stata, R and Minitab filesCSV, TSV, Excel .xlsx, SPSS .sav and clipboard, per makerSPSS data and syntax; Gnumeric and LibreOffice spreadsheets; Excel via CSV or OpenDocument
Preparing data in the programComputed and transformed columns, row filters; reshaping and merging through modulesRecode, filter and computed columns by menu or R; an independent reviewer rates its data management minimalCompute, recode, bin, rank, select and weight cases; spreadsheet filtering and variable metadataTransformation commands; several datasets can be merged, joined or concatenated; exact dialogs not verified
Built-in analysesDescriptives, t-tests, ANOVA, regression, frequencies, factor analysis; 70+ library modulesClassical and Bayesian t-tests, ANOVA, regression and frequencies; online module library since March 202627 documented modules: common tests, ANOVA, regression, survival, ROC, ICC and agreement methodsDescriptives, t-tests, ANOVA, linear and logistic regression, factor, cluster, non-parametric; no modules
Distinctive capabilitySpreadsheet and live results side by side; Rj editor runs R insidePublic verification against SPSS, SAS, Minitab and R (run on JASP 0.13.1); optional AI assistantSPSS-style Variable View with labels, missing rules and scales; Python/PySide6 applicationSPSS-compatible syntax; terminal or graphical interface
Sources for preparing and analysing data

NuriStat package documentation and releases, checked October 28, 2026, supply maker claims for its column. Checked October 19, 2026: jamovi features, FAQ, spreadsheet and homepage; JASP features, FAQ, getting started, release notes and verification project; GNU PSPP home page and FAQ (Internet Archive copies; gnu.org did not respond); Muenchen, August 4, 2026. Lists are documented features, not evidence of matching defaults or correct results.

Keeping, checking and sharing the work

Show 3 criteria
Keeping, checking and sharing the work
FeaturejamoviJASPNuriStatGNU PSPP
Saved workOne .omv file: data, analyses, options and resultsOne .jasp file: data, options and results; optional password encryptionComplete session in .swb: ZIP, Parquet data and JSON metadata, per maker; cross-version reopening unknownData file plus syntax; command journal on by default; no one-file project verified
Seeing how it was doneSyntax mode shows R for each analysis (jmv package); import step omitted; no one-step workflow exportR syntax can be shown in results; no one-step workflow export; full standalone R path not establishedAuto-generated syntax log for analyses; whole-workflow reproduction not establishedThe syntax is the record; which menus generate syntax not verified
Readable outputCopy into Word or email; .docx and .odt export on the Current line since September 30, 2026Annotated PDF or HTMLStructured tables, HTML output and PNG/SVG charts, per makerText, PostScript, PDF, OpenDocument or HTML
Sources for keeping, checking and sharing the work

NuriStat project format and output, checked October 28, 2026. jamovi features, results, jamovi and R and release notes; JASP teaching guide, section 5 and release notes; PSPP home page and release notes; Muenchen on workflow export. We did not open, export or regenerate any file.

Cost, platforms and risk

Show 8 criteria
Cost, platforms and risk
FeaturejamoviJASPNuriStatGNU PSPP
Desktop costFree for any use; optional paid Cloud plansFree; optional AI uses your own provider key, which may chargeFree MIT source; Python and scientific-library dependencies; no paid recipient licence shownFree; no license fee, expiry or paid add-ons
Desktop platformsWindows, macOS, Linux; Chromebooks via Flathub; current minimum versions not establishedWindows 10 22H2 or current Windows 11; macOS 15 or later; Linux Flatpak; Chromebooks from 2019Windows 10/11, macOS and Linux per release notes; Python 3.10+ source installation; installer instructions conflictLinux distribution packages; volunteer Windows and Mac builds, which the project ranks low priority
Release and pace, checked October 28, 2026Current 28.14 and Solid 2.7.41 (October 21, 2026); both update the CA certificate bundle0.99.0 (October 21, 2026), plot fixes; 0.97.1 fixed a misleading rounding error in June3.6.1 (June 11, 2026), English UI; July human source fixes not confirmed in that release2.1.2 (September 19, 2026), bug and translation fixes; Debian stable ships 2.0.1
Internet connectionsFiles stay local; welcome screen, updates and modules connect and log usage statistics you cannot switch off in the appRuns without internet; collects no user data, per JASP; optional AI sends requests to the provider you configureLocal project files and public source; exact connections and installed dependency versions not documentedNo desktop privacy statement found
Security recordUpdate to 2.7.41 or 28.14 or later for compromised TLS certificates; older affected releases need updating before module installs; no GitHub advisories; old CVE affects up to 1.6.18Signed releases; Windows R sandbox since 0.95.0; no published GitHub advisories; HECVAT publishedNo published security policy or advisories; signing/sandbox not documented; no comprehensive security clearanceEight 2025 CVEs for crafted input; Debian lists seven unfixed in 2.1.2, three rated unimportant
Help and supportForum, documentation, GitHub; no formal helpdesk or SLAForum, GitHub, teaching materials; paid support plans from JASP Services BVPublic GitHub issues and contribution guide; no formal support commitment establishedManual and mailing lists; private list for security reports
Who maintains itThe jamovi Project Pty Ltd (Australia), a small team funded by community, Patreon and Cloud subscriptionsJASP Team, University of Amsterdam; long-term grants and tenured core developers, per JASPGitHub maintainer baramgay; human commits in July 2026; funding not establishedGNU volunteers; releases announced by Ben Pfaff
LicenseAGPL3, some parts GPL2 or laterInterface AGPLv3, engine GPLv2MIT per package READMEGPLv3 or later
Sources for cost, platforms and risk

NuriStat requirements, licence and history, releases and security page, checked October 28, 2026; no published security policy or advisory is not proof of no vulnerabilities. jamovi privacy policy, September 24, 2026, Cloud plans, security policy, advisories, license and CVE-2021-28079; JASP download and requirements, security, FAQ, AI announcement and JASP Services BV; PSPP obtaining PSPP, 2.1.2 announcement and Debian security tracker. jamovi’s download page redirected our requests, so its minimum OS versions are unconfirmed. Release notes for jamovi and JASP, jamovi’s certificate advisory and the PSPP tracker were checked October 28, 2026; other evidence remains dated October 19. Cloud, AI provider charges and support-plan prices are outside the desktop price.

What matters when choosing

jamovi is our Practical Pick for someone who wants to prepare and analyse a nonsensitive table on their own computer without writing code, then hand the work to someone else. The desktop app is free for any purpose, including commercial work, and opens CSV, Excel, SPSS, SAS, Stata and R data files directly. Its spreadsheet sits beside the results: you can add computed and transformed columns and row filters, and the results update as you change them. One .omv file keeps the data, every analysis with its settings and the results together, so a colleague can open it and change an option instead of rebuilding your work, and syntax mode shows the R code behind each analysis. The most important compromise is the pace of change. jamovi’s Current line had twelve releases between July 28 and October 15, 2026, one of them a fix to how computed variables handle missing values. The project describes its Solid line as more thoroughly tested; use it unless you need a new feature. FAQ, features, release notes, security policy

jamovi’s desktop privacy policy, updated September 24, 2026, says your data files and analyses never leave your computer. The app does go online to show its welcome screen, check for updates and install modules, and on those connections it records your operating system, language, country, downloaded modules and a shortened IP address. You cannot switch this off inside jamovi; the policy says to block two jamovi web addresses instead. Desktop privacy policy

Consider JASP instead if your data is already clean and you want classical and Bayesian versions of the same analysis side by side. JASP lists both versions for its t-tests, ANOVA, regression and frequency analyses, and publishes a project comparing its results with SPSS, SAS, Minitab, R and hand calculations, although that comparison used an older JASP version (0.13.1). Its weakness for this reader is preparation: an independent reviewer of R-based statistics programs wrote in August 2026 that almost none of the usual data preparation can be done in JASP. The current version needs macOS 15 or later on a Mac. Its new AI assistant is off until you add a key for an AI provider; once it is on, your requests go to the provider you choose. JASP features, verification project, Muenchen’s comparison, download and requirements, AI announcement

NuriStat is our Rising Star for people who want a free, SPSS-style desktop workspace with variable metadata, menu-driven analyses and editable projects. Its maker documents common tests as well as survival analysis, ROC and agreement methods. The main limit is installation: on a Mac, expect to install Python and the application from source; current Windows installer instructions conflict. jamovi remains the easier choice for preparing and handing over a table. NuriStat documentation, releases, checked October 28, 2026.

Choose GNU PSPP only if you need SPSS-style syntax and data files and work mainly on GNU/Linux. Its project says Windows and Mac are “low on our list of priorities”, builds for them come from volunteers, and several 2025 vulnerability reports about crafted files are still listed as unfixed by Debian’s security tracker for the current 2.1.2 package. Use it with data files you created or trust. PSPP FAQ, Debian security tracker

These are researched judgments from documentation, release notes, policies and public reports checked October 19, 2026, with the release, security and NuriStat updates dated October 28 below. We did not install any program, import a file or run an analysis. Nothing here tells you which statistical method suits your question, or that a listed analysis will match another program’s defaults. Three conditions would change our recommendation: clean data and a need for Bayesian results (choose JASP), an institution that needs signed installers, a formal security questionnaire or paid support (JASP documents all three), or SPSS syntax you must keep running (consider PSPP). Whichever you choose, keep the original data file and a short note of the rows you excluded and the variables you changed beside the project file. A project shows what was done; it does not record why.

Practical Picks is researched and written with Claude. JASP advertises support for Anthropic among its optional AI providers; AI features do not affect our recommendation for this task. Release notes, checked October 28, 2026

jamovi

Preferred Pick · prepare, analyse and hand over without code

$0AGPL3 · optional paid Cloud plans

Choose jamovi when you want to clean a table and run common analyses in one program, then give someone a single file they can reopen and change. Use the Solid release line, and keep the original data beside the project.

Reasons to choose it

  • Computed and transformed columns and row filters beside live results
  • One .omv file holds the data, analyses, settings and results
  • Shows the R code behind each analysis

What to watch

  • The Current line changes almost weekly; one October fix concerned missing values
  • Usage statistics are sent when it goes online and cannot be switched off in the app
  • Generated R code leaves out the data import step

jamovi is a free statistics program built on R, published by The jamovi Project Pty Ltd, an Australian company. Its window pairs a spreadsheet with a results panel. You can enter or paste data, open CSV, Excel, SPSS, SAS, Stata and R files, correct the types jamovi guesses for each column, add computed variables from formulas, create transformed (recoded) copies that leave the source column untouched, and filter rows out of an analysis. When you open an SPSS or Stata file, the FAQ is explicit that your data comes across but your earlier analyses do not, so plan to rerun them. FAQ, first steps, spreadsheet, desktop privacy policy

The reason to hesitate is change. jamovi ships on two lines. Current had twelve releases from July 28 to October 15, 2026 (28.1 to 28.13), including a fix to missing values in computed variables on October 1 and a fix to decimal missing-value comparison on September 18. Solid, 2.7.41 as of October 21, 2026, is described in the project’s security policy as more thoroughly tested and comparable in spirit to a long-term-support release, and it receives security fixes too. The documentation also warns that automatic type guessing can be wrong, so check that numeric codes for categories are not being treated as measurements. Release notes, security policy, spreadsheet

Read the full jamovi analysis and sources

jamovi saves the data, analyses, their options and results in one .omv file, and you can annotate the results with headings and notes before sharing. Results copy into Word or an email; the Current line added .docx and .odt export on September 30, 2026. Syntax mode shows equivalent R code for each analysis, using jamovi’s own jmv package, but the code omits the step that imports the data, and Robert Muenchen’s August 2026 comparison notes there is no single export of the whole workflow: you copy each step. The optional Rj editor runs R inside jamovi with bundled R or your own installation; with your own R, anyone rerunning the work needs the same packages. A file containing Rj code shows a warning on opening and does not run the code unless you allow it. Features, results, jamovi and R, FAQ, Muenchen’s comparison

Built-in analyses cover descriptives, t-tests, ANOVA including repeated measures and ANCOVA, correlation, linear and logistic regression, frequencies and factor analysis. The project says its library holds more than 70 community modules, adding methods such as mixed models, meta-analysis and Bayesian analyses; modules are written by third parties, and since version 2.7.32 (June 15, 2026) jamovi checks module signatures. Muenchen’s comparison gives jamovi the most graphics features among the R-based programs he reviewed and says reshaping and merging datasets are available through modules. Features, homepage counts, release notes, Muenchen’s comparison

The privacy policy keeps the usage records for 60 days on servers in the Netherlands and says they are not sold or shared for advertising. Blocking www.jamovi.org and library.jamovi.org stops them, at the cost of the welcome screen and new module installs. jamovi Cloud is a separate service that processes data on jamovi’s servers; its Guest plan is free, Priority is priced by region and Teams by quote. Cloud is outside this comparison. Desktop privacy policy, Cloud plans

The project describes itself as a small open-source team supported by the community, Patreon and Cloud subscriptions. Its security policy says there is no formal helpdesk or service-level agreement and that confirmed vulnerabilities are typically fixed within 24 hours; we cannot verify that timing. GitHub shows no published security advisories. The project reports compromised TLS certificates and describes the impact as low. Solid 2.7.41 and Current 28.14, released October 21, 2026, stop trusting those certificates. Update to one of those versions or later; if you use 2.7.31 or earlier, or 28.1–28.4, update before installing or updating modules. Until then, avoid installing modules over networks you do not trust. The older CVE-2021-28079 concerns versions up to 1.6.18. Certificate advisory, checked October 28, 2026 If jamovi will not start on Windows, the FAQ’s suggestion is another computer or jamovi Cloud. The source code is AGPL3, with some parts GPL2 or later. Security policy, advisories, CVE record, FAQ, license

Choose JASP if your data is already prepared and you want classical and Bayesian results paired by default, or signed installers and a published security questionnaire for an institution. Choose PSPP if your work must stay in SPSS syntax.

Used jamovi? Share your experience to help other readers choose.

JASP

Also consider · clean data, classical and Bayesian results

$0AGPLv3 / GPLv2 · AI uses your own key

Consider JASP when your data is already prepared and you want classical and Bayesian versions of the same analysis, or an institution needs signed installers, a security questionnaire and paid support.

Reasons to choose it

  • Classical and Bayesian versions of core analyses side by side
  • Signed releases, a Windows R sandbox and a published HECVAT
  • Public comparison of its results with SPSS, SAS, Minitab and R

What to watch

  • Little data preparation, by an independent reviewer’s count
  • Current version needs macOS 15 or later on a Mac
  • The optional AI assistant sends requests to the provider you choose

JASP is a free statistics program from the JASP Team at the University of Amsterdam’s Psychological Methods group. Its feature list pairs classical and Bayesian versions of t-tests, ANOVA, regression and frequency analyses, with modules for areas such as mixed models, structural equation modelling, machine learning, meta-analysis and quality control; since March 2026 modules can be added or updated from an online library. It opens CSV and text, Excel, OpenDocument, SPSS, SAS, Stata, R and, since 0.96.0, Minitab files. The release notes list 0.99.0, dated October 21, 2026, with plot fixes and Persian translation support, checked October 28, 2026. JASP also publishes a verification project that compares its output with SPSS, SAS, Minitab, R and hand calculations; the comparison was run on JASP 0.13.1, not the current release. Features, FAQ, release notes, download, validation, verification project

The strongest reason not to choose it for this reader is preparation. JASP documents recoding, filters and computed columns through menus or R, but Robert Muenchen’s August 2026 comparison of R-based statistics programs says JASP offers “just a handful” of data management functions and that almost none of the usual data preparation can be done in it. Its own guide warns that changing a column to Scale turns non-numeric entries into missing values, and that some analyses force a variable’s type, marked with an asterisk. Do your cleaning in a spreadsheet first if you choose JASP. On a Mac, the current version requires macOS 15 or later; JASP says Windows devices with ARM processors may run it in slower emulation but are not yet supported. Getting started, Muenchen’s comparison, download and requirements

Read the full JASP analysis and sources

A .jasp file holds the data, analysis options and results, and annotated output exports to PDF or HTML. Since 0.97.0 (May 13, 2026) you can protect a file with a password, or encrypt it so only the JASP team can read it when reporting a problem. JASP can show R syntax in the results and keeps that setting in the file; we did not establish that the code reproduces every module’s analysis outside JASP. Teaching guide, section 5, release notes

Upgrades can change results, so rerun important analyses after updating. Version 0.97.1 (June 5, 2026) fixed numbers over four significant figures that were sometimes rounded in a misleading way, including a wrong sum in descriptives, when you had not fixed the number of decimals; the team apologised for it. The same release fixed older files whose labels or categorical filters broke in 0.97. Version 0.98.8 fixed a crash on saving on Windows systems using the Encrypting File System, and a crash on broken AI input. The known-issues page listed no important issues for 0.98.8 when checked October 19; that does not establish the status of 0.99.0. Release notes, known issues

The AI assistant added in 0.98.0 (July 2026) can create and run analyses and write annotations from a chat. It is off until you turn it on in Preferences and provide an API key, and it works with OpenAI-compatible services, including local models. A July 2026 post on JASP’s blog points readers to free tiers and notes that “for free” means “you pay with your data”. JASP’s security page says it never uploads data and does not use the cloud; in our judgment, that statement cannot cover requests you choose to send through the assistant, so check your provider’s terms before using it with real data. Without the assistant, the FAQ says the software collects no user-identifiable information; the 2018 privacy policy covers the website and forum, not the program. AI announcement, release notes, free API key post, security, FAQ, privacy policy

JASP’s institutional paperwork is the most complete of the four. Releases are digitally signed on every platform; on Windows, since 0.95.0, the R engine runs in a sandbox that by default cannot reach the internet or files outside its workspace. The project publishes a HECVAT security questionnaire and an accessibility statement, says it is funded by long-term grants with tenured core developers, and has no published GitHub security advisories. JASP Services BV sells support plans for organisations; prices are not published. The interface is licensed AGPLv3 and the engine GPLv2. It needs at least 4 GB of memory (8 GB recommended) and about 4 GB of disk, and no internet connection to run. Security, FAQ, JASP Services BV, download and requirements

Choose jamovi if you need to prepare the data in the same program or use an older Mac. Choose PSPP if SPSS syntax compatibility matters more than modern analysis menus.

Used JASP? Share your experience to help other readers choose.

NuriStat

Rising Star · SPSS-style variable metadata and menu analyses

$0MIT source · Python and scientific dependencies

Consider NuriStat if you want SPSS-style variable definitions, common tests and editable projects, and can manage a source installation. jamovi remains the easier route for preparing and handing over a table.

Reasons to choose it

  • Variable metadata, computed columns, recoding and filters
  • Common tests plus survival, ROC and agreement methods, per maker
  • Complete-session SWB projects and HTML output

What to watch

  • Expect Python/source installation on Mac; Windows installer instructions conflict
  • No published security policy; exact dependencies and connections undocumented
  • July source correctness fixes are not confirmed in the June release

NuriStat is a free, MIT-licensed desktop statistics program built with Python and PySide6. Its maker describes a spreadsheet with cell editing, undo and redo, sorting and filtering, plus a Variable View for labels, missing-value rules and measurement scales. Menus cover computing and recoding variables, binning, ranking, selecting and weighting cases. It opens CSV, TSV, Excel and SPSS .sav files. Choose it if that SPSS-style organization suits your work and you can manage the installation; choose jamovi for a simpler route from preparation to results. Documentation, checked October 28, 2026.

The documented analyses include descriptives, t-tests, ANOVA, regression, factor and cluster analysis, survival analysis, ROC curves, ICC and Bland–Altman plots. These are maker claims about available methods, not evidence that their results match another program. A .swb project saves the complete analysis session; the maker describes its contents as ZIP, Parquet data and JSON metadata. A syntax log records analyses, and output exports to HTML. Compatibility between project versions has not been established. Keep your original data and check important results after upgrading. Documentation, release notes.

Read the full NuriStat analysis and sources

The June 2026 releases and repository history place this entrant within the past two years, including its earlier StatWorkbench name. Version 3.6.1, released June 11, adds an English interface; the release notes explicitly list Windows 10/11, macOS and Linux support. The installation route needs care: the README requires Python 3.10 or later, lists PySide6 6.6 or later, and says the PyPI package is coming soon. Earlier releases list Windows installers, while the June 11 release says an installer is coming soon. Use the source-install instructions as the dependable documented route, and note that the current package sits in the repository’s nuristat folder. This is more setup than downloading jamovi or JASP. Releases, package and requirements, checked October 28, 2026.

The public repository shows human maintenance in July 2026. It has no published security policy or advisories; that does not establish that the program has no vulnerabilities. Exact installed dependency versions and internet connections are not documented in the pages we could read. July source changes include a fix for a Wilcoxon dialog that previously showed placeholder output; we could not establish that the June release includes that fix. For nonsensitive tables, the public source and portable project format make NuriStat worth considering, but its installation and security documentation are less complete than the leading alternatives. Choose JASP instead for paired classical and Bayesian analyses and institutional security documentation. Package history, security page, checked October 28, 2026.

Used NuriStat? Share your experience to help other readers choose.

GNU PSPP

SPSS-style syntax on GNU/Linux

$0GPLv3+ · volunteer builds

Consider GNU PSPP only when you must keep SPSS-style syntax and data files and work mainly on GNU/Linux. Open only data files you created or trust.

Reasons to choose it

  • Syntax and data files compatible with SPSS, per the project
  • No license fee, expiry, case limits or paid add-ons
  • Terminal or graphical interface

What to watch

  • Debian lists seven 2025 vulnerability records as unfixed in 2.1.2
  • Windows and Mac are low priority, with volunteer builds
  • No one-file project; which menus exist was not verified

GNU PSPP describes itself as a free replacement for SPSS, with syntax and data files compatible with SPSS, a graphical or terminal interface, and no expiry, case limits or paid add-ons. It lists descriptive statistics, t-tests, ANOVA, linear and logistic regression, measures of association, cluster, reliability and factor analysis and non-parametric tests, and can open, merge, join or concatenate several datasets at once. Version 2.1.2, announced September 19, 2026 by maintainer Ben Pfaff, contains translation updates and bug fixes; 2.0.0 added the CTABLES command and made command journaling the default when PSPP runs interactively. It is licensed GPLv3 or later. PSPP home page, archived October 8, 2026, 2.1.2 announcement, release notes

The reasons to hesitate are platform and security. The project releases source code only; packages come from Linux distributions and volunteers, and its FAQ says Windows and Mac are low priorities. Debian’s current stable release ships 2.0.1. Eight vulnerability records published in 2025 concern crafted input to PSPP 2.0.1 or the development code of that time. On October 19, 2026, Debian’s security tracker listed seven of the eight as still unfixed in its 2.1.2 package, three of them rated unimportant. One is CVE-2025-47814, a heap overflow when reading SPSS output (.spv) files, whose upstream bug report remains open. In September 2025 the maintainer said such reports “may be current” and matter if you download data files from sites you do not trust, while doubting that vulnerability sites check their claims. Obtaining PSPP, archived September 26, 2026, FAQ, archived September 17, 2026, Debian security tracker, CVE-2025-47814, upstream bug 67074, maintainer response

Read the full GNU PSPP analysis and sources

PSPP imports Gnumeric, LibreOffice and OpenOffice spreadsheets directly through File, Import Data. For Excel, the FAQ recommends saving as CSV or OpenDocument first. A bundled tool, pspp-convert, can decrypt encrypted SPSS data and syntax files. “Compatible” syntax and data files do not mean an SPSS project, its output and every procedure transfer intact; check the commands you rely on. FAQ, archived September 17, 2026, PSPP home page, archived

Your saved work is a data file plus syntax, not a single project with results. Output can be written as text, PostScript, PDF, OpenDocument or HTML. We could not read enough of the current graphical-interface manual to confirm which procedures have dialogs, so if avoiding syntax is a firm requirement, check the dialog you need before relying on it. PSPP home page, archived

PSPP is a GNU volunteer project. Help comes from the manual and the pspp-users mailing list, and security reports go to a private pspp-security list. We found no privacy statement specific to the desktop program. An experimental rewrite in the Rust language is described by the project as “very incomplete”. PSPP home page, archived, Obtaining PSPP, archived

Choose jamovi for point-and-click preparation and a one-file handover, or JASP for paired classical and Bayesian analyses with signed installers.

Used GNU PSPP? Share your experience to help other readers choose.

Which product are we missing?

Know a product that belongs in this article? Tell us what it does differently and why it deserves a closer look. Newcomers welcome.

Readers and product teams can suggest candidates. Suggestions inform our research; they do not guarantee inclusion or a recommendation.

Suggest a product

Opens an email draft with this article attached as a link. Nothing is sent until you send it.

No email app? Send the product name, website, and this article’s title to admin@practicalpicks.guide.

What people are saying

Robert Muenchen, who has long published reviews of point-and-click programs built on R, updated his comparison on August 4, 2026. Counting features, he ranks BlueSky Statistics Pro (a commercial product) first overall with jamovi close behind; jamovi leads on graphics; JASP is close to the top on ease of use but has minimal data management. He also notes that neither jamovi nor JASP exports a whole workflow as R code in one step. Muenchen is also the author of BlueSky Statistics’ user guide, published by BlueSky Statistics LLC, and his scores count features rather than measure quality. R GUI comparison, user guide announcement

An August 2025 discussion of SPSS alternatives includes recommendations for jamovi and JASP alongside a first-person warning about the learning curve when moving to them; broad claims there that PSPP is identical to SPSS are not adopted here. An older PSPP discussion reports installation difficulties and bugs without version details. Both were last read September 7, 2026; they are small, self-selected accounts that suggest questions to ask, not satisfaction or reliability scores.

Our October 19 search for vulnerabilities, lawsuits, regulatory action, ownership changes and recurring complaints involving the three projects found nothing beyond the items in this article that changed our conclusions. A search that finds nothing is not proof that nothing exists.

How we researched

This is public-source research. We did not install any program, open or import a data file, run an analysis or use any AI feature. No project supplied access, payment or input, and we use no affiliate links. We first compared jamovi, JASP and GNU PSPP on September 7, 2026, and rechecked every material claim on October 19, 2026: releases, platforms, licenses, imports, saved-work formats, privacy policies, security records and support routes. We read the projects’ websites, documentation, release notes, privacy and security pages and source-code repositories, Debian’s package and security records, the GNU announcement archive, Robert Muenchen’s comparison and public discussions. The GNU PSPP website did not respond to our requests on October 19, so we read its home, download and FAQ pages in Internet Archive copies from September 17 to October 8, 2026. jamovi’s desktop download page redirected our automated requests to its home page, so we could not confirm the minimum operating-system versions for its current installers.

We wrote this for someone with a nonsensitive table, such as survey responses, course results or a business export, who wants to clean it and run common analyses on their own computer without writing code, and pass the work to someone else. We fixed five criteria and weights before comparing: preparing data in the program (20%), coverage of common analyses (20%), keeping and sharing inspectable work (25%), desktop fit and release stability (15%), and privacy, security and continuity (20%). jamovi scored highest. Shifting most of the weight to analysis coverage for data that is already clean puts JASP narrowly ahead, which is why we name it as the alternative for that case. R, Python and RStudio require code for this job; browser services do not meet the desktop requirement. We also considered BlueSky Statistics, IBM SPSS Statistics, Stata, JMP, RKWard and R AnalyticFlow; their current documentation did not establish a complete comparison of editable handover, platforms and security. We could not confirm comparable current prices for these alternatives.

Our October 28, 2026 examination of launch and project pages identified NuriStat as the strongest eligible newer entrant for this desktop task. Its June releases document Windows and Mac support, menu-driven preparation and analyses, and editable projects; the installation burden is its main compromise. BioStat Prime’s first release, Mac support and responsible operator could not be established. MetricGate and SavQuick serve browser or server workflows. JASP’s AI assistant and jamovi’s release line are changes to established programs. NuriStat releases, documentation, BioStat Prime, MetricGate, SavQuick, checked October 28, 2026.

ToolPublic sourcesWhat remains uncertain
jamoviFAQ, features, release notes, desktop privacy policy (September 24, 2026), Cloud plans, user documentation, GitHub license, security policy and advisories, Muenchen’s comparison.Minimum OS versions for current installers, how often Current-line fixes change results, the exact scope of the new .docx export, module quality.
JASPDownload and requirements, release notes, FAQ, features, getting-started guide, teaching guide, security, validation and known-issues pages, verification project, AI announcement and blog post, privacy policy, JASP Services BV, Muenchen’s comparison.What the AI assistant sends to a provider, whether exported R syntax reproduces every module outside JASP, verification on current versions, support-plan prices.
NuriStatPublic package README, June releases, July human commit history and security page, checked October 28, 2026.Conflicting installer instructions, dependency versions and connections, cross-version project compatibility, whether July correctness fixes are included in the June release.
GNU PSPPArchived home, download and FAQ pages, 2.1.2 announcement, release notes as packaged by Debian, Debian security tracker and bug 1105106, upstream bug 67074, maintainer response.Which graphical dialogs exist, whether the open 2025 vulnerabilities affect 2.1.2, quality of volunteer Windows and Mac builds, desktop privacy practice.

None of this establishes statistical accuracy, speed, import fidelity or security on your computer. Next substantive review is due December 18, 2026, or sooner after a major release or security change. Found a changed release or policy? Contact the editors.

Sources and evidence

Sources were checked October 19, 2026 unless another date is given. Release notes and security pages for jamovi and JASP, the Debian PSPP security tracker, and the candidate pages cited above were checked October 28, 2026. Project material establishes documented features, policies and releases, not behaviour on your computer. Independent reviews and community posts are attributed opinions, not representative samples.

This is a researched comparison, not a product review. Method: evidence standard, editorial picks.