GPower is Heinrich-Heine-Universität Düsseldorf’s free statistical power analysis tool, and it has been the graduate-student standard since 2007. It is free, it is well-documented in peer-reviewed papers, and it works. It is also Windows-first, Intel-only on Mac (a native Apple Silicon version is under development), and it has no way to import a dataset — you enter effect sizes and sample sizes by hand. Here are seven GPower alternatives for desktop that either do more or do it better.
Quick comparison
| App | Best for | Free plan | Starting price | Standout feature |
|---|---|---|---|---|
| JASP | Modern power + full stats | Yes | Free | SPSS-like UI with Bayesian too |
| jamovi | R-backed stats without R syntax | Yes | Free | Runs R behind a friendly interface |
| R (pwr package) | Reproducible, scriptable power | Yes | Free | Every power test in one script |
| PASS | Comprehensive commercial suite | Trial | $1,295/yr | Widest test coverage |
| nQuery | Clinical trial sample sizing | Trial | Quote | Adaptive design and gatekeeping |
| SPSS SamplePower | IBM SPSS shops | Trial | ~$99/mo | Familiar SPSS interface |
| WebPower | Free web-based analysis | Yes | Free | Browser-based, no install |
Why researchers move off G*Power
The dataset gap. GPower takes hand-entered parameters and returns power or sample size. It does not read a CSV, does not compute effect sizes from real data, and does not export tables for a manuscript. If you want to run a post-hoc power analysis on an existing dataset, GPower expects you to compute the effect size elsewhere and type the result in.
The Mac situation. GPower 3.1.9.6 is the current Mac build, and it is compiled for Intel processors. On Apple Silicon Macs it runs under Rosetta 2, which works but adds latency. HHU is working on GPower 4 as a native Apple Silicon build, but it has been in development for several years.
The Linux gap. There is no official Linux build. Researchers on Linux run G*Power under Wine, which is functional but not officially supported.
The scriptability gap. Every GPower run is a manual GUI operation. Reproducible research workflows (Rmarkdown, Jupyter, Quarto) cannot embed a GPower call the way they can embed an R or Python one. This makes G*Power an outlier in modern computational science.
1. JASP — Best modern power analysis with a full stats suite
JASP is the open-source statistical analysis suite from the University of Amsterdam. Since version 0.16 it has included a full power analysis module that covers most G*Power tests plus Bayesian equivalents. UI is SPSS-familiar; runs on Windows, macOS, and Linux.
Where it falls short: The power module is newer than the rest of JASP and does not cover every G*Power test yet. Highly specialized designs (multi-arm gatekeeping, adaptive) still need a dedicated tool.
Pricing:
- Free (AGPL)
- vs G*Power: same free tier, native Apple Silicon, dataset-aware
Migrating from G*Power: No file format shared. Rebuild your analysis in JASP’s Power module using the same parameters.
Download: jasp-stats.org
Bottom line: Pick JASP as a modern G*Power replacement plus a full stats package. Skip it for niche designs.
2. jamovi — Best for R-backed analysis without R syntax
jamovi wraps R in a friendly, SPSS-like interface. The community-maintained jpower module adds power analysis for common designs (t-tests, ANOVA, correlations, regressions). Runs on Windows, macOS, and Linux.
Where it falls short: jpower is community-maintained and not as comprehensive as G*Power or JASP’s built-in module. Reproducibility depends on the jpower version.
Pricing:
- Free (AGPL)
- vs G*Power: same free tier, extensible via R
Migrating from G*Power: No file format shared. Set parameters in jamovi’s jpower module.
Download: jamovi.org
Bottom line: Pick jamovi if your team already prefers R-flavoured stats but wants a GUI. Skip it for one-off power calculations.
3. R (pwr package) — Best for reproducible workflows
R with Stéphane Champely’s pwr package is the scriptable, reproducible standard. A single line of R computes power, sample size, or effect size for any of the common tests. Fits inside Rmarkdown, Quarto, and Jupyter notebooks natively.
Where it falls short: Command-line first. Non-R users face a learning curve. Some specialized designs (multi-arm trials, adaptive) need other packages like Superpower or simr.
Pricing:
- Free (GPL)
- vs G*Power: same free tier, radically more capable when you know R
Migrating from G*Power: Every G*Power test has a pwr equivalent. Manuals cross-reference the two.
Download: r-project.org plus install.packages("pwr") in R console
Bottom line: Pick R plus pwr if your research already runs in R. Skip it if you have never opened a terminal.
4. PASS — Best comprehensive commercial power software
PASS by NCSS is the commercial gold standard for power analysis in clinical and pharmaceutical research. Coverage exceeds G*Power by an order of magnitude: over 1,000 procedures, all validated against published methods. Windows-only.
Where it falls short: Windows-only. Pricey. Overkill for casual academic use.
Pricing:
- Trial: 7-day
- Individual: $1,295/year
- Academic: $995/year
- vs G*Power: much pricier, dramatically wider coverage
Migrating from G*Power: No shared format. Manuals cross-reference procedures.
Download: ncss.com/software/pass
Bottom line: Pick PASS for regulated clinical research budgets can support. Skip it for classroom or exploratory work.
5. nQuery — Best for clinical trial sample sizing
nQuery is a specialized tool for adaptive clinical trial design, group sequential methods, and multi-arm sample sizing. Used in pharma submissions to the FDA and EMA. Windows and macOS.
Where it falls short: Narrow scope. If you are not designing a clinical trial, nQuery is more than you need.
Pricing:
- Quote-based, typically several thousand dollars per year
- vs G*Power: significantly pricier, focused on regulatory-grade designs
Migrating from G*Power: No shared format. Parameters and effect sizes re-entered.
Download: statsols.com
Bottom line: Pick nQuery for clinical trial biostatisticians. Skip it for anything else.
6. SPSS SamplePower — Best for SPSS shops
SPSS SamplePower is IBM’s add-on for SPSS Statistics that brings power analysis into the SPSS interface. Familiar to anyone already using SPSS; workflow integrates with existing SPSS syntax.
Where it falls short: Requires an SPSS Statistics license first. Windows-first, macOS available. Not as comprehensive as PASS.
Pricing:
- Requires SPSS Statistics subscription: ~$99/month for Standard tier
- vs G*Power: paid, only worthwhile if SPSS is already in use
Migrating from G*Power: No file share.
Download: ibm.com/products/spss-statistics
Bottom line: Pick SPSS SamplePower if the lab already licenses SPSS. Skip it as a standalone purchase.
7. WebPower — Best free web-based analysis
WebPower from Zhiyong Zhang’s team at Notre Dame is a free web-based power analysis tool with a matching R package. Runs in any desktop browser on Windows, macOS, and Linux. Covers common tests plus longitudinal and multilevel designs G*Power skimps on.
Where it falls short: Web UI is spartan. Requires internet.
Pricing:
- Free
- vs G*Power: same free tier, browser-based, stronger for multilevel and longitudinal designs
Migrating from G*Power: Re-enter parameters in WebPower’s form.
Download: webpower.psychstat.org
Bottom line: Pick WebPower for a quick, no-install power calculation. Skip it if offline reliability matters.
How to choose
Pick JASP as the modern default. It replaces G*Power for most academic tests and adds full statistical analysis.
Pick jamovi if your workflow prefers R behind a GUI.
Pick R plus pwr if you already write R for the rest of your analysis. It is the reproducibility win.
Pick PASS for regulated clinical research budgets that need every test procedure.
Pick nQuery specifically for adaptive clinical trial designs.
Pick SPSS SamplePower in an SPSS-committed lab.
Use WebPower for a fast one-off calculation, especially for multilevel or longitudinal designs.
Stay on G*Power if you are teaching a research methods class where every textbook cites GPower, or if your reviewer expects GPower screenshots. The tool is not going away; it just is no longer the default best choice.
FAQ
Is G*Power still maintained? Yes. The Heinrich Heine University group continues to fix bugs and issue new builds. A native Apple Silicon Mac version is under active development as G*Power 4.
What is the best free G*Power alternative? JASP and jamovi are the two strongest free options for most academic researchers. Both run on Windows, macOS (native Apple Silicon), and Linux and include modern statistical procedures alongside power analysis. R with the pwr package is the strongest scriptable option.
Is JASP better than G*Power? For most academic use, yes. JASP includes power analysis, Bayesian statistics, and full frequentist analysis under one interface. G*Power is more established but narrower in scope.
Can I run G*Power on Linux? Not officially. G*Power under Wine works for most researchers. Native Linux users typically move to JASP, jamovi, or R with pwr.
Which alternative works on Apple Silicon Macs? JASP, jamovi, R, WebPower (browser), and SPSS all have native Apple Silicon builds. GPower currently runs under Rosetta 2 pending the GPower 4 release.
