statzel — Compared with other statistics software

What statzel covers next to widely used statistics software, from getting started to each family of methods. The other products’ entries come from their vendors’ own documentation, checked in September 2026; each entry links to its source.

Compare with
FeaturestatzelSASSPSSStataJASPjamovigretlEViewsPrismOriginMinitabMedCalcAnalyse-it
Getting started
PriceFreePaid annual licence; free for learnersPaid, subscription or perpetualPaid, perpetual or annualFreeFree; paid cloud plansFreePaid, perpetual; academic and student editionsPaid subscriptionPaid, annual or perpetualPaid subscriptionPaid, perpetual or subscriptionPaid, subscription or perpetual
Open sourceNoNoNoNoYes (AGPL-3)Yes (AGPL-3)Yes (GPL)NoNoNoNoNoNo
Runs in a browser, nothing to installYesYes, via SAS OnDemand or ViyaNoNoNo official web versionYes, via jamovi CloudNoNoNo — Prism Cloud views and shares onlyNoYes, via the Minitab web appNoNo
Data stays on your deviceYes, even in the browserOn a local install; the browser versions run on SAS serversYesYesYesDesktop yes; Cloud runs on jamovi serversYesYesYesYesDesktop yes; web app runs on Minitab serversYesYes
PlatformsAny modern browserWindows, Linux, webWindows, macOSWindows, macOS, LinuxWindows, macOS, Linux, ChromeOSWindows, macOS, Linux, webWindows, macOS, LinuxWindows; student editions also macOSWindows, macOSWindowsWindows, webWindowsWindows, inside desktop Excel
Point and click, no code neededYesPartly — SAS Studio tasksYesYesYesYesYesYesYesYesYesYesYes
Opens SPSS, Stata, SAS, R and Parquet filesAll fiveSAS; SPSS and Stata with SAS/ACCESS; Parquet in ViyaSPSS, Stata (to v13), SASSPSS, Stata, SAS, ParquetSPSS, Stata, SAS, RSPSS, Stata, SAS, RSPSS, Stata, SAS transport (.xpt)SPSS, Stata, SAS transport (.xpt), R (with R installed)None of the fiveSPSS, SASNone of the fiveSPSSNone of the five (reads Excel worksheets)
Methods
Classical testst-tests, ANOVA, nonparametric tests, chi-square✓✓✓✓✓✓✓✓✓Higher edition*✓✓✓
Regressionlinear, logistic and count (Poisson / negative binomial)✓✓Higher edition*✓Partial*Add-on*✓✓Partial*Partial*Partial*Partial*Partial*
Mixed-effects modelslinear mixed models or GEE✓✓Higher edition*✓✓Add-on*——Partial*Higher edition*✓——
Survival analysisKaplan-Meier and Cox regression✓✓Higher edition*✓✓Add-on*Partial*—✓Higher edition*✓✓Higher edition*
ROC analysisROC curves with AUC, and comparing two AUCs✓✓✓✓Partial*Add-on*Add-on*—Partial*Partial*Partial*✓Higher edition*
Method comparisonBland-Altman, and Passing-Bablok or Deming regression✓—Partial*—Partial*Add-on*——✓Higher edition*Add-on*✓Higher edition*
Lab method validationCLSI EP05 precision, EP17 detection limits, EP06 linearity✓——————————Partial*Higher edition*
Reference intervalsreference-interval estimation, as in CLSI EP28✓——————————✓Higher edition*
Meta-analysisfixed- and random-effects, with forest plots✓Partial*✓✓✓Add-on*—————✓—
Agreement and reliabilityCohen's and Fleiss' kappa, ICC, Cronbach's alpha✓Partial*✓✓✓Add-on*———Partial*Partial*Partial*Partial*
Factor analysis and PCAexploratory factor analysis and principal components✓✓✓✓✓✓Add-on*✓Partial*Higher edition*✓—✓
Structural equation modelsSEM and confirmatory factor analysis✓✓Higher edition*✓✓Add-on*———Higher edition*———
Bayesian testsBayes-factor t-test, ANOVA, correlation, regression✓Partial*Higher edition*Partial*✓Add-on*———————
Curve fittingnonlinear regression, including 4PL dose-response✓✓Higher edition*✓—Partial*✓✓✓✓✓✓—
Time seriesARIMA modelling and forecasting✓Add-on*Higher edition*✓✓Add-on*✓✓—Higher edition*✓——
Econometric time seriesVAR, Johansen cointegration, GARCH✓Add-on*Partial*✓——✓✓—————
Panel data and IVfixed/random-effects panel models, 2SLS✓Add-on*Partial*✓—Partial*✓✓—————
Control charts and capabilitySPC control charts, Cp/Cpk✓Add-on*✓Partial*✓————Higher edition*✓Partial*Higher edition*
Gauge R&R and DOEmeasurement-system studies, factorial designs✓Partial*——✓————Higher edition*✓——
Machine learningrandom forest, SVM, boosting, neural networks✓Add-on*Partial*Partial*✓Add-on*Partial*——Partial*Partial*——
Power and sample sizepower and sample-size calculation✓✓✓✓✓Add-on*——✓Higher edition*✓✓—
Clusteringk-means and hierarchical clustering✓✓✓✓✓Add-on*Partial*Partial*Higher edition*Higher edition*✓——

* what is missing is noted below the table · ✓ built in, including modules that ship with the product · Higher edition in the product, but only in a higher edition, tier or plan · Add-on needs something installed or licensed separately — a module, app, add-in or package · Partial some of the row, not all · — not offered. A mark counts only if the product does everything the row’s definition lists; statzel’s marks link to the method pages.

Notes

Each entry was checked against the vendor’s public documentation in September 2026, and each links to the page it was checked against. Products change; the source is the place to confirm an entry that has gone out of date. Product names are trademarks of their respective owners, used here only to identify the products.

SAS 9.4M9 and SAS Viya

  • Price: SAS 9.4 is licensed on a periodic (annual) basis; OnDemand for Academics is free for professors, students and independent learners source
  • Runs in a browser, nothing to install: OnDemand is browser-based SAS Studio in SAS's cloud source
  • Data stays on your device: Local SAS 9.4 install on Windows/Linux keeps data on your machine (often deployed on an organisation's server instead); OnDemand for Academics and Viya process data on SAS/cloud servers (uploads, 5 GB quota on ODA) source
  • Platforms: macOS is not a supported platform for SAS 9.4M9; Mac users use the browser (SAS Studio) source
  • Point and click, no code needed: SAS Studio tasks and SAS Enterprise Guide cover common analyses without code; many procedures/options are code-only source
  • Opens SPSS, Stata, SAS, R and Parquet files: Without the PC Files licence only delimited and JMP files import; Parquet LIBNAME engine is Viya-only; no R .rds reader source
  • Classical tests: TTEST, ANOVA/GLM, NPAR1WAY, FREQ (chi-square) source
  • Regression: REG, LOGISTIC, GENMOD (Poisson, negative binomial) source
  • Mixed-effects models: MIXED, GLIMMIX, GEE source
  • Survival analysis: LIFETEST (Kaplan-Meier), PHREG (Cox) source
  • ROC analysis: PROC LOGISTIC ROC/ROCCONTRAST compares AUCs with the DeLong-DeLong-Clarke-Pearson method source
  • Method comparison: No built-in procedure; SAS's own blog shows Passing-Bablok/Deming as user-written SAS/IML programs source
  • Lab method validation: No CLSI protocol procedures in SAS/STAT or SAS/QC source
  • Reference intervals: No dedicated reference-interval procedure (percentiles via UNIVARIATE only) source
  • Meta-analysis: No dedicated meta-analysis procedure or forest-plot routine; random-effects meta-analysis can be coded with MIXED/GLIMMIX/MCMC source
  • Agreement and reliability: Cohen's/weighted kappa (FREQ AGREE) and Cronbach's alpha (CORR ALPHA) built in; Fleiss' kappa and ICC only via SAS-supplied macros %MAGREE and %INTRACC source
  • Factor analysis and PCA: FACTOR, PRINCOMP source
  • Structural equation models: CALIS source
  • Bayesian tests: General Bayesian modelling (MCMC, BGLIMM, BAYES statements) but no Bayes-factor t-test/ANOVA/correlation procedures source
  • Curve fitting: NLIN / NLMIXED; 4PL is entered as a custom formula. source
  • Time series: SAS/ETS: ARIMA, ESM, UCM, FORECAST source
  • Econometric time series: SAS/ETS: VARMAX (VAR, cointegration), AUTOREG (GARCH) source
  • Panel data and IV: SAS/ETS: PANEL (fixed/random effects), SYSLIN (2SLS) source
  • Control charts and capability: SAS/QC: SHEWHART, CUSUM, CAPABILITY source
  • Gauge R&R and DOE: DOE via SAS/QC FACTEX/OPTEX (and PLAN in SAS/STAT); no dedicated gauge R&R procedure source
  • Machine learning: SAS/STAT has HPSPLIT trees only; random forest, gradient boosting, SVM and neural nets (FOREST, GRADBOOST, SVMACHINE, NNET) need SAS Visual Data Mining and Machine Learning on Viya (or SAS Enterprise Miner) source
  • Power and sample size: POWER, GLMPOWER source
  • Clustering: FASTCLUS (k-means), CLUSTER (hierarchical) source

IBM SPSS Statistics 32

  • Price: Subscription (monthly/quarterly/annual) or perpetual licence; campus and student licences; free trial source
  • Runs in a browser, nothing to install: Only desktop client installers (Windows, macOS) and server installers are offered source
  • Data stays on your device: Desktop application; data stays local unless you use distributed mode with SPSS Statistics Server. The optional AI Output Assistant sends selected output to watsonx.ai source
  • Platforms: IBM: 'The IBM SPSS Statistics desktop software is not supported on Linux operating systems' (v32); Linux only for SPSS Statistics Server source
  • Point and click, no code needed: All procedures available from menus/dialogs; syntax optional source
  • Opens SPSS, Stata, SAS, R and Parquet files: Documented Stata support lists versions 4-13 (.dta); no R .rds or Parquet reader documented source
  • Classical tests: t-tests, one-way ANOVA/GLM univariate, nonparametric tests, chi-square in Base edition source
  • Regression: Linear regression in Base; binary logistic needs the Regression option; Poisson/negative binomial need Generalized Linear Models (Advanced Statistics). Both in Standard edition or higher, or 'Custom Tables & Advanced Statistics' subscription add-on source
  • Mixed-effects models: Linear Mixed Models, GLMM and GEE are in Advanced Statistics: Standard edition or higher, or 'Custom Tables & Advanced Statistics' subscription add-on source
  • Survival analysis: Kaplan-Meier and Cox Regression are in Advanced Statistics (Standard edition or higher) source
  • ROC analysis: ROC Analysis in Base; paired-sample design compares two ROC curves (AUC difference test) source
  • Method comparison: Bland-Altman analysis in Base (Analyze > Descriptive Statistics); no Passing-Bablok or Deming regression documented source
  • Lab method validation: No CLSI EP05/EP17/EP06 procedures documented source
  • Reference intervals: No reference-interval (EP28) procedure documented source
  • Meta-analysis: Meta-Analysis (continuous, binary, meta-regression) with forest plots in Base source
  • Agreement and reliability: Cohen's (and weighted) kappa via Crosstabs/Weighted Kappa, Fleiss' multiple-rater kappa, ICC and Cronbach's alpha via Reliability Analysis, all Base source
  • Factor analysis and PCA: Factor Analysis (incl. principal components extraction) in Base source
  • Structural equation models: Requires IBM SPSS Amos (separate application); included in Premium edition, add-on for Base/Standard/Professional source
  • Bayesian tests: Bayesian one-sample/paired/independent t-tests, one-way ANOVA, Pearson correlation, linear regression with Bayes factors; Advanced Statistics (Standard edition or higher) source
  • Curve fitting: Nonlinear Regression (user-specified model, so 4PL can be entered) is in the Regression option (Standard edition or higher); no built-in 4PL template in its list of common models source
  • Time series: ARIMA / Expert Modeler forecasting require the Forecasting option (Professional edition or higher, or 'Forecasting & Decision Trees' subscription add-on) source
  • Econometric time series: VAR models added in v32 (Forecasting option); stationarity/cointegration tests only via the STATS TSTESTS extension; no GARCH documented source
  • Panel data and IV: 2SLS in the Regression option (Standard edition); no dedicated panel fixed/random-effects estimator (random-intercept models via Linear Mixed Models) source
  • Control charts and capability: SPCHART (core system): X-bar, R, s, individuals, p, np, c, u charts and capability indices including Cp and Cpk. The source is IBM's SPSS 24 page; the v32 syntax reference could not be retrieved, and nothing indicates the command was removed. source
  • Gauge R&R and DOE: No gauge R&R or design-of-experiments generator documented (factorial ANOVA can analyse experiments; orthogonal designs only for conjoint in the Conjoint option) source
  • Machine learning: Neural networks need the Premium edition or the Forecasting & Decision Trees add-on; decision trees need Professional; no random forest or SVM. source
  • Power and sample size: Power Analysis (means, proportions, correlations, regression) in Base source
  • Clustering: K-Means, Hierarchical and TwoStep cluster in Base source

Stata 19 / StataNow

  • Price: Annual licences include StataNow; perpetual licences include one year of maintenance source
  • Runs in a browser, nothing to install: Desktop software; StataNow is a continuous-release desktop version, not a web app source
  • Data stays on your device: Desktop application source
  • Point and click, no code needed: 'You can access all of Stata's data management, statistical, and analysis features from the menus and associated dialogs.' source
  • Opens SPSS, Stata, SAS, R and Parquet files: import excel/delimited/spss/sas/parquet; no R .rds reader source
  • Classical tests: ttest, anova/oneway, ranksum/kwallis etc., tabulate chi2 source
  • Regression: regress, logit/logistic, poisson, nbreg source
  • Mixed-effects models: mixed/me commands; xtgee source
  • Survival analysis: sts (Kaplan-Meier), stcox source
  • ROC analysis: roctab/roccomp; roccomp compares AUCs with the DeLong, DeLong and Clarke-Pearson method source
  • Method comparison: No official Bland-Altman, Passing-Bablok or Deming command; only community-contributed packages source
  • Meta-analysis: meta suite with fixed/random-effects models and meta forestplot source
  • Agreement and reliability: kap/kappa (two or multiple nonunique raters, i.e. Fleiss), icc, alpha source
  • Factor analysis and PCA: factor, pca source
  • Structural equation models: sem/gsem incl. CFA; SEM Builder GUI source
  • Bayesian tests: bayes: prefix for 60+ estimators and Bayes factors via bayesstats ic / bayestest model, but no dedicated default-prior Bayes-factor t-test/correlation commands source
  • Curve fitting: nl with built-in log4 (4-parameter logistic) function source
  • Time series: arima with forecasting source
  • Econometric time series: var, vec with Johansen rank tests (vecrank), arch/garch source
  • Panel data and IV: xtreg fe/re, ivregress 2sls source
  • Control charts and capability: Control charts (cchart, pchart, rchart, xchart, shewhart); no process-capability (Cp/Cpk) command. source
  • Gauge R&R and DOE: anova analyses factorial experiments, but no design generation or gauge R&R command source
  • Machine learning: h2oml: random forest and gradient boosting via H2O (Stata 19); lasso/elastic net built in; no built-in SVM or neural networks source
  • Power and sample size: power, ciwidth source
  • Clustering: cluster kmeans and hierarchical linkage methods source

JASP 0.98.1

  • Runs in a browser, nothing to install: JASP's own Chromebook guide points to a third-party hosted version at rollApp; no official web version from the JASP team. source
  • Data stays on your device: Desktop app; data stays on the machine. The optional AI feature added in 0.98 uses a user-supplied API key, and the release post doesn't say what is sent. The third-party rollApp route runs remotely. source
  • Platforms: No official web version. source
  • Point and click, no code needed: Menu- and dialog-driven GUI. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Also .ods, .xpt, .por, .rdata and Minitab. Parquet is not listed. source
  • Classical tests: T-Tests/ANOVA/Frequencies are core modules; Mann-Whitney, Wilcoxon, Kruskal-Wallis and contingency-table chi-square are built in. source
  • Regression: Linear, logistic and GLM (includes Poisson) are core. No negative-binomial regression is listed. source
  • Mixed-effects models: Linear and generalized linear mixed models are in the core Mixed Models module. No GEE. source
  • Survival analysis: In the bundled Survival module (switch it on from the Modules menu): Kaplan-Meier (non-parametric), Cox (semi-parametric) and parametric models. source
  • ROC analysis: Logistic regression reports AUC; the Machine Learning module plots ROC curves. No test comparing two AUCs (e.g. DeLong) was found. source
  • Method comparison: Bland-Altman plots are in the official Reliability module. No Passing-Bablok or Deming regression. source
  • Lab method validation: No CLSI EP protocols listed. source
  • Reference intervals: Not listed. source
  • Meta-analysis: In the bundled Meta-Analysis module (metafor-based): classical/Bayesian, fixed/random effects, forest plots. source
  • Agreement and reliability: In the bundled Reliability module, switched on from the Modules menu: Cohen's/Fleiss' kappa (Rater Agreement), ICC, Cronbach's alpha. source
  • Factor analysis and PCA: PCA and EFA are in the core Factor module. source
  • Structural equation models: CFA is in the core Factor module; full SEM (lavaan-based), with PLS-SEM, mediation and MIMIC, is in the bundled SEM module, switched on from the Modules menu. source
  • Bayesian tests: Bayesian t-tests, ANOVA, correlation, regression and contingency tables are core. source
  • Curve fitting: No nonlinear regression analysis. The bundled JAGS module can fit a hand-written Bayesian model, which is not the same thing. source
  • Time series: In the bundled Time Series module, switched on from the Modules menu: ARIMA with forecasting. source
  • Econometric time series: No VAR, Johansen cointegration or GARCH listed. source
  • Panel data and IV: No panel FE/RE or 2SLS analysis listed. source
  • Control charts and capability: In the bundled Quality Control module, switched on from the Modules menu: variables/attributes/time-weighted control charts and process capability. source
  • Gauge R&R and DOE: In the bundled Quality Control module, switched on from the Modules menu: Type 1-3 gauge R&R, factorial and response-surface designs and their analysis. source
  • Machine learning: In the bundled Machine Learning module, switched on from the Modules menu: random forest, SVM, boosting, neural networks, KNN, decision trees, etc. source
  • Power and sample size: In the bundled Power module (based on jpower): sample size, power and minimum effect size. source
  • Clustering: In the bundled Machine Learning module, switched on from the Modules menu: k-means (listed as Neighborhood-Based), hierarchical, density-, model-based and others. source

jamovi 28.3

  • Data stays on your device: Desktop: data never leaves the device. Cloud: data is processed on jamovi's servers during the session and removed afterwards. source
  • Platforms: Also ChromeOS via Flathub. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Parquet is not listed. source
  • Classical tests: Core t-tests (with Mann-Whitney/Wilcoxon), ANOVA, non-parametric one-way and RM ANOVA (Kruskal-Wallis/Friedman), contingency-table chi-square. source
  • Regression: Core has linear and logistic (binomial, multinomial, ordinal). Poisson/negative-binomial regression needs the GAMLj3 library module. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Mixed-effects models: Linear and generalized mixed models via the GAMLj3 library module. No GEE. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Survival analysis: Kaplan-Meier and Cox via library modules (jsurvival, jYS, deathwatch). Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • ROC analysis: Core binomial logistic regression gives a ROC plot and AUC. DeLong comparison of AUCs needs the meddecide library module (DiagROC also offers ROC). Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Method comparison: Bland-Altman via blandr or SimplyAgree; Deming regression via SimplyAgree. No Passing-Bablok found. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Lab method validation: No CLSI EP module in the library listing. source
  • Reference intervals: No reference-interval module in the library listing. source
  • Meta-analysis: MAJOR (metafor), MetaJam (meta) and esci library modules, with forest plots. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Agreement and reliability: Cronbach's alpha is core (Reliability Analysis). Kappa and ICC need library modules (meddecide, seolmatrix, SimplyAgree). Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Factor analysis and PCA: Core Factor menu: PCA and EFA. source
  • Structural equation models: CFA is core; full SEM via the semlj or pathfy library modules. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Bayesian tests: jsq library module (a port of JASP's Bayesian analyses): Bayesian t-tests, ANOVA, linear models, contingency tables. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Curve fitting: The PGM library module fits Richards growth curves. No 4PL dose-response module found. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Time series: The snowCluster library module has ARIMA (auto-ARIMA with forecasts) and Prophet. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Econometric time series: No VAR, Johansen cointegration or GARCH module found. source
  • Panel data and IV: 2SLS instrumental variables via the endoj library module. No panel fixed/random-effects estimator found. source
  • Control charts and capability: No control-chart or process-capability module found. source
  • Gauge R&R and DOE: No gauge R&R or factorial DoE analysis found. The Randomizer module only randomizes simple designs. source
  • Machine learning: MLwrapj (neural nets, random forest, XGBoost, SVM), snowCluster and jYS library modules. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Power and sample size: jpower (Morey & Selker), pamlj and other library modules. Library modules are community-contributed; the jamovi library does not mark any module as official. source
  • Clustering: snowCluster library module: k-means and hierarchical clustering. Library modules are community-contributed; the jamovi library does not mark any module as official. source

gretl 2026b

  • Runs in a browser, nothing to install: Desktop only. source
  • Data stays on your device: Desktop application. source
  • Point and click, no code needed: Full GUI with menus; the hansl scripting language is optional. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Also Excel, ODS, Gnumeric, EViews .wf1 and JMulTi. It cannot open SAS .sas7bdat, R data files or Parquet; for R it only writes a text format R's read.table can load. source
  • Classical tests: t-tests (meantest), ANOVA, sign/Wilcoxon rank-sum/signed-rank/runs tests, chi-square (xtab). Kruskal-Wallis and Friedman are contributed packages. source
  • Regression: OLS, logit/probit, Poisson and negative binomial are built in. source
  • Mixed-effects models: No linear mixed-effects models or GEE (panel random effects only). source
  • Survival analysis: Kaplan-Meier estimator (kmeier function) and parametric duration models (Weibull, exponential, log-logistic, log-normal). No Cox regression. source
  • ROC analysis: The contributed roc package (gretl's package server) plots ROC curves with AUC and DeLong standard errors; its roccomp function tests whether AUCs are equal. source
  • Method comparison: Not in the command reference or package list. source
  • Meta-analysis: Not in the command reference or package list. source
  • Agreement and reliability: No kappa, ICC or Cronbach's alpha found. source
  • Factor analysis and PCA: PCA is built in (pca); exploratory factor analysis (ML or principal factors, varimax) comes from the contributed staticfactor package. source
  • Bayesian tests: No Bayes-factor tests. Contributed Bayesian regression/BMA/BVAR packages exist. source
  • Curve fitting: General nonlinear least squares (nls). A 4PL model has to be written out by the user as a custom formula; there is no dose-response template. source
  • Time series: ARIMA/SARIMA estimation with forecasting (arima, fcast). source
  • Econometric time series: VAR, VECM, Johansen cointegration and GARCH are built in. source
  • Panel data and IV: Panel fixed/random effects (panel) and 2SLS (tsls) are built in. source
  • Machine learning: SVM built in (LIBSVM) plus LASSO/ridge/elastic net. Random forest (via R), boosting and KNN are contributed packages. No neural nets. source
  • Power and sample size: No power or sample-size routine found. source
  • Clustering: k-means via the contributed kmeans package. No hierarchical clustering found. Contributed function packages are installed from gretl's package server. source

EViews 14

  • Price: Paid, perpetual single-user licence (no annual renewal fee); volume licences with annual renewal. Academic pricing; University Edition sold as a 6-month subscription; free, limited Student Version Lite. EViews 14 merged the former Standard and Enterprise licences into a single product. The Student Version Lite limits data (1,500 obs per series), cannot save workfiles and expires after 1 year. source
  • Runs in a browser, nothing to install: Desktop application that must be installed; no web version. source
  • Data stays on your device: Runs locally on the desktop. University Edition and Student Version Lite need an internet connection for licence checks, and data can optionally be fetched from online databases. source
  • Platforms: Windows (64-bit). University Edition and Student Version Lite also run on macOS. No Linux or web version. Per the FAQ, installing the full single-user licence on a Mac requires contacting sales. source
  • Point and click, no code needed: Menus and dialogs cover estimation; there is also a command/programming language. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Excel (.xls/.xlsx), CSV/text, SPSS .sav and portable, Stata .dta, SAS transport (.xpt) and SAS program files, R data files. No SAS .sas7bdat or Parquet listed. R data support is listed on the EViews 14 feature list (EViews forum staff say it needs a local R installation). Also reads Access, Gauss, ODBC and many time-series databases. source
  • Classical tests: t-tests, ANOVA, Wilcoxon, Mann-Whitney, Kruskal-Wallis and others; cross-tabulation with Pearson chi-square. source
  • Regression: Linear OLS, binary logit/probit, count models with Poisson and negative binomial, plus GLM. source
  • Mixed-effects models: Panel fixed/random-effects estimators only; no general linear mixed models or GEE. source
  • Survival analysis: No Kaplan-Meier or Cox regression. The only related item is a Weibull hazard example for user-specified maximum likelihood. source
  • Agreement and reliability: No kappa, ICC or Cronbach's alpha listed. source
  • Factor analysis and PCA: PCA with scree plots and biplots; factor analysis with over 30 rotation methods. source
  • Bayesian tests: Bayesian VARs only; no Bayes-factor tests. source
  • Curve fitting: General nonlinear least squares; a 4PL curve is fitted by typing its formula (no dedicated dose-response procedure). source
  • Time series: ARMA/ARIMA/ARFIMA, automatic ARIMA selection, exponential smoothing and forecasting. source
  • Econometric time series: VAR/VEC, Johansen cointegration, GARCH family and multivariate GARCH. source
  • Panel data and IV: Panel fixed/random effects, 2SLS/IV, GMM and Arellano-Bond. source
  • Machine learning: EViews 14 has only elastic net, ridge and LASSO. The EViews 15 beta adds random forests, decision trees and LSTM, but it is not a release. source
  • Clustering: k-means only, via the community-written KMeans add-in from the official EViews add-ins library, which is installed separately. No hierarchical clustering. source

GraphPad Prism 11.1

  • Price: Perpetual licence gets no upgrades; several newer analyses need the Pro/Enterprise tier source
  • Runs in a browser, nothing to install: Prism Cloud only views/shares/discusses projects published from the desktop app; you can't enter data or run analyses in the browser. The one exception is the power and sample-size calculator, which runs in Prism Cloud source
  • Data stays on your device: Desktop app; projects only go to Prism Cloud if the user publishes them for sharing source
  • Platforms: Web (Prism Cloud) is for viewing/sharing only source
  • Opens SPSS, Stata, SAS, R and Parquet files: Excel import needs Excel installed on the same Windows PC (Prism Mac cannot import Excel files; paste or CSV instead). No SPSS/Stata/SAS/RDS/Parquet import documented source
  • Classical tests: t tests, ordinary/RM ANOVA, Mann-Whitney, Wilcoxon, Kruskal-Wallis, Friedman, chi-square/Fisher contingency tables source
  • Regression: Multiple linear, simple/multiple logistic and Poisson regression are included; no negative binomial regression in the guides source
  • Mixed-effects models: A mixed-effects model is offered only as 'repeated measures ANOVA that allows missing values' (subjects random); no general LMM or GEE source
  • Survival analysis: Kaplan-Meier with log-rank; Cox proportional hazards regression source
  • ROC analysis: ROC with AUC is included, but the guide says 'Prism does not compare ROC curves' (manual calculation only) source
  • Method comparison: Bland-Altman and Deming regression; no Passing-Bablok source
  • Lab method validation: No CLSI EP05/EP17/EP06 protocols in the Prism 11 guides source
  • Reference intervals: No reference-interval procedure in the Prism 11 guides source
  • Meta-analysis: No meta-analysis in the Prism 11 guides source
  • Agreement and reliability: No kappa, ICC or Cronbach's alpha in Prism itself; GraphPad's separate free QuickCalcs web calculator does kappa from a table source
  • Factor analysis and PCA: PCA (with parallel analysis and PCR) is included; no exploratory factor analysis source
  • Curve fitting: Nonlinear regression with 105 built-in equations, including variable-slope (4PL) dose-response source
  • Gauge R&R and DOE: Multifactor ANOVA (Pro) can analyse factorial data, but there's no gauge R&R or design generation source
  • Machine learning: GraphPad markets its clustering as 'machine learning', but there are no supervised methods (RF/SVM/boosting/NN) source
  • Power and sample size: Runs in Prism Cloud (browser), launched from Prism; the plans FAQ lists it for Standard and above, with some regions excluded source
  • Clustering: K-means and hierarchical clustering need the Prism Pro or Enterprise plan (not Standard) source

Origin / OriginPro 2026b

  • Price: Two editions, Origin and OriginPro. Node-locked, floating or dongle licenses; first year of maintenance included. Academic pricing available; Student Version (1-year term, student-owned PC, 30-window limit) and a free 6-month Learning Edition (12-window limit, university email). source
  • Open source: Proprietary commercial software. source
  • Runs in a browser, nothing to install: Windows desktop application only. The free Origin Viewer (Windows, and an older native Mac build 9.6.5) runs without installation but only views/copies project contents; it is not browser-based and cannot run analyses. source
  • Data stays on your device: Installed desktop application; data and projects are stored locally. source
  • Platforms: 64-bit Windows 10/11 only. No native macOS or Linux version; Mac users run the Windows version via Boot Camp or virtual machine software. source
  • Point and click, no code needed: Analyses run from menus and dialog boxes; LabTalk, Origin C and Python scripting are optional. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Excel and CSV pre-installed; SPSS .sav is an optional component chosen at install; SAS .sas7bdat import ships but must be added to the menu; SAS XPT via a free downloadable connector App. No Stata .dta, R .rds/.RData or Parquet import listed. source
  • Classical tests: t-tests and one/two-way ANOVA built in; nonparametric tests and chi-square (Cross Tabulation) are OriginPro only. source
  • Regression: Linear/multiple regression built in. Logistic regression via a free OriginLab App (the Origin vs OriginPro page lists it for Origin, but the App page says it requires OriginPro); Poisson regression App is OriginPro only; no negative binomial regression found. source
  • Mixed-effects models: OriginPro only; Linear Mixed Effects Model App (GLMM App also available, requires R). No GEE found. source
  • Survival analysis: OriginPro only (Kaplan-Meier and Cox built into OriginPro). source
  • ROC analysis: OriginPro only; ROC curve with AUC, SE and CI, but no test comparing two AUCs (e.g. DeLong) documented. source
  • Method comparison: Bland-Altman plot built in; Deming regression (Fit Linear with X Error) is OriginPro only; no Passing-Bablok. source
  • Lab method validation: No CLSI EP05/EP17/EP06 protocols found. source
  • Reference intervals: No reference-interval (CLSI EP28) tool found; a Tolerance Interval App exists (OriginPro) but is not reference-interval estimation. source
  • Meta-analysis: No meta-analysis; a free Forest Plot App only draws a forest plot from user-supplied estimates and CIs. source
  • Agreement and reliability: OriginPro only: Cohen's kappa in Cross Tabulation and Chi-square; Fleiss' kappa via Attribute Agreement Analysis App. No ICC or Cronbach's alpha found. source
  • Factor analysis and PCA: OriginPro only; PCA built into OriginPro, exploratory factor analysis via Factor Analysis App. source
  • Structural equation models: OriginPro only; Structural Equation Modeling App (Python semopy). source
  • Bayesian tests: No Bayes-factor tests found (only a Bayesian Ridge Regression machine-learning App). source
  • Curve fitting: Nonlinear fitting with built-in DoseResp (4PL) function; user-defined functions supported. source
  • Time series: OriginPro only; ARIMA modelling and forecasting via Time Series Analysis App. source
  • Econometric time series: No VAR, Johansen cointegration or GARCH found; Advanced Time Series Analysis App covers only ADF/KPSS stationarity, Granger causality and prewhitening. source
  • Panel data and IV: No panel-data or instrumental-variable (2SLS) regression found. source
  • Control charts and capability: Basic QC (X-bar/R) chart built in; full control charts and capability analysis (Cp/Cpk) via Statistical Process Control App, OriginPro only. source
  • Gauge R&R and DOE: OriginPro only; Gage R&R Study and Design of Experiments Apps. source
  • Machine learning: OriginPro only; random forest, SVM and neural network via Apps; no boosting found. source
  • Power and sample size: OriginPro only. source
  • Clustering: OriginPro only. source

Minitab Statistical Software 22.5

  • Price: Sold as 'Minitab Solution Center' tiers (Core/Analytics/Copilot); desktop and online access included. No perpetual licence on the pricing page source
  • Runs in a browser, nothing to install: Minitab Web runs in Chrome, Edge or Safari and needs an internet connection source
  • Data stays on your device: The Windows desktop app is local. The web app processes data on Minitab's servers (Minitab says it 'processes but does not store' data), with files kept in OneDrive/Google Drive or local files source
  • Platforms: 22.5 desktop is 'Windows only' (Windows 11, x86); macOS users use the web app. No Linux source
  • Point and click, no code needed: Menus/dialogs plus the Assistant; scripting, Python and R integration are optional source
  • Opens SPSS, Stata, SAS, R and Parquet files: Opens MPX/MPJ/MWX/MTW, XLS/XLSX/XML, CSV/TXT/DAT (web app can't open XLS/XML). No SPSS/Stata/SAS/RDS/Parquet documented source
  • Classical tests: t tests, ANOVA/GLM, sign/Wilcoxon/Mann-Whitney/Kruskal-Wallis/Mood's median/Friedman, chi-square and Fisher's exact source
  • Regression: Linear, binary/ordinal/nominal logistic and Poisson regression are included; no negative binomial regression in the feature list source
  • Mixed-effects models: 'Mixed models' (Fit Mixed Effects Model) under Analysis of Variance; no GEE source
  • Survival analysis: Kaplan-Meier via Nonparametric Distribution Analysis (Reliability/Survival menu) with log-rank; Cox regression listed under Regression source
  • ROC analysis: ROC curve and AUC are output from binary logistic regression (and predictive models); no standalone ROC test or two-AUC (DeLong) comparison documented source
  • Method comparison: Orthogonal (Deming) regression is built in. Bland-Altman is only an official Minitab macro, run from the command line (%BLANDALT). No Passing-Bablok source
  • Lab method validation: Has gage studies (MSA) and stability studies, but no CLSI EP05/EP17/EP06 protocols source
  • Reference intervals: Tolerance intervals only; no reference-interval (EP28) procedure source
  • Agreement and reliability: Fleiss' and Cohen's kappa (Attribute Agreement Analysis) and Cronbach's alpha (Item Analysis) are included; no ICC in the feature list source
  • Factor analysis and PCA: Principal components analysis and factor analysis source
  • Curve fitting: General nonlinear regression with a function catalogue; the docs show no ready-made 4PL, so a 4PL is entered as a custom formula. source
  • Time series: ARIMA, 'Forecast with Best ARIMA Model', exponential smoothing, Winters source
  • Econometric time series: Only ACF/PACF/CCF and the ADF test; no VAR, Johansen or GARCH source
  • Control charts and capability: Variables/attributes/time-weighted/multivariate control charts; normal and non-normal capability, Capability Sixpack source
  • Gauge R&R and DOE: Gage R&R crossed/nested/expanded; two-level, general factorial, RSM, mixture, D-optimal, Taguchi designs source
  • Machine learning: CART, random forests, TreeNet boosting and MARS come in the separately priced Predictive Analytics Module (Analytics and Copilot tiers, not Core); no SVM or neural networks even with it. source
  • Power and sample size: t, proportions, Poisson rates, variances, equivalence, one-way ANOVA, factorial designs, tolerance intervals source
  • Clustering: Cluster K-Means, Cluster Observations (hierarchical), Cluster Variables source

MedCalc 23.7

  • Runs in a browser, nothing to install: The desktop app needs installing. MedCalc's free online calculators work on summary numbers only, not datasets source
  • Data stays on your device: Windows desktop application source
  • Platforms: Native Windows only (x86/x64 and ARM64). For macOS and Linux, MedCalc documents running it under Wine/CrossOver source
  • Point and click, no code needed: Menus/dialogs; scripts are optional source
  • Opens SPSS, Stata, SAS, R and Parquet files: Also Excel (xls, xlsx, xlsm), ODS, DBF, WK1, SYLK, DIF and TXT/PRN. No Stata, SAS, R or Parquet. source
  • Classical tests: t tests, one-/two-way/RM ANOVA, ANCOVA, Mann-Whitney, Wilcoxon, Kruskal-Wallis, Friedman, chi-squared, Fisher, McNemar source
  • Regression: Linear/multiple, logistic, probit and nonlinear regression are included; no Poisson or negative binomial regression source
  • Mixed-effects models: Repeated-measures ANOVA only; no LMM/GEE source
  • Survival analysis: Kaplan-Meier and Cox proportional-hazards regression source
  • ROC analysis: AUC by the DeLong or Hanley & McNeil methods; compares up to 6 ROC curves, plus partial AUC source
  • Method comparison: Bland-Altman (including multiple measurements per subject), Deming (including weighted) and Passing-Bablok regression, mountain plot source
  • Lab method validation: Probit regression reports the limit of detection with a CLSI EP17-A2 reference, and there's CV from duplicates. No EP05 precision, EP06 linearity or LoB procedures source
  • Reference intervals: Parametric, nonparametric percentile and robust methods with 90% CIs per CLSI C28-A3; age-related reference intervals source
  • Meta-analysis: Fixed and random effects, forest and funnel plots, heterogeneity, Egger/Begg tests source
  • Agreement and reliability: Cohen's (weighted) kappa for two raters, ICC, concordance correlation and Cronbach's alpha; no Fleiss' multi-rater kappa found. source
  • Curve fitting: Nonlinear regression from a user-entered equation; the manual's worked example is a 4-parameter logistic, entered as a custom formula. source
  • Control charts and capability: Levey-Jennings-style QC control chart with Westgard rules; no Xbar-R etc. and no Cp/Cpk capability source
  • Power and sample size: Sample size for means, proportions, correlation, logrank, Bland-Altman, AUC and ROC comparison source

Analyse-it 6.23

  • Price: Paid, annual subscription or perpetual licence (perpetual includes 1 year of maintenance). Student pricing for the Standard edition (annual only). Five editions: Standard, Quality Control & Improvement, Medical, Method Validation, Ultimate. Free 15-day trial. source
  • Runs in a browser, nothing to install: COM add-in for desktop Excel for Windows. Does not work in Excel Online or Excel for the web. source
  • Data stays on your device: Runs entirely inside Excel on your PC with no cloud processing. Floating licences need internet for licence management only. source
  • Platforms: Windows only: Excel 365, 2024, 2021, 2019 or 2016 for Windows (32- and 64-bit) on Windows 10/11 or Windows Server 2016+. Not Excel for Mac or Excel for the web. On a Mac, the vendor suggests running Windows in a virtual machine (e.g. Parallels). source
  • Point and click, no code needed: Ribbon tab and task panes inside Excel; no code needed. source
  • Opens SPSS, Stata, SAS, R and Parquet files: Excel workbooks (data must be in an Excel worksheet, so anything Excel opens, e.g. CSV). No SPSS, Stata, SAS, R or Parquet import is documented. source
  • Classical tests: t-tests, ANOVA, Mann-Whitney, Wilcoxon, Kruskal-Wallis, Friedman, chi-square, Fisher exact, McNemar. source
  • Regression: Linear (simple, multiple, polynomial), logistic and probit. No Poisson or negative binomial count regression. source
  • Mixed-effects models: Only within-subjects (repeated-measures) ANOVA; no linear mixed models or GEE. source
  • Survival analysis: Medical or Ultimate edition. The vendor's help article lists survival under Standard, but the edition comparison table ticks only Medical and Ultimate. source
  • ROC analysis: Medical, Method Validation or Ultimate edition. AUC with DeLong CI; compares up to 10 tests with DeLong. source
  • Method comparison: Method Validation or Ultimate edition. The Medical edition has Bland-Altman but not Passing-Bablok or Deming. source
  • Lab method validation: Method Validation or Ultimate edition. Covers EP05-A3, EP17-A2 and EP06-Ed2. source
  • Reference intervals: Medical, Method Validation or Ultimate edition (CLSI EP28-A3c). Verifying and transferring intervals needs Method Validation or Ultimate. source
  • Agreement and reliability: Cronbach's alpha is in Standard. Cohen's and weighted kappa need the Medical, Method Validation or Ultimate edition. No Fleiss' kappa or ICC. source
  • Factor analysis and PCA: PCA with biplots; common factor analysis with 12 rotations. source
  • Curve fitting: Only polynomial, logarithmic, exponential, power and probit fits. No general nonlinear regression or 4PL. source
  • Control charts and capability: Quality Control & Improvement or Ultimate edition. Includes Shewhart, CUSUM and EWMA charts and Cp/Cpk/Pp/Ppk. source
  • Gauge R&R and DOE: No gauge R&R or DOE design generation. The measurement-systems tools are CLSI precision studies in the Method Validation edition. source