Today I noticed this question, and I thought it would be helpful if we had a thread that listed resources that people could conveniently access for power analysis / sample size calculations, perhaps analogous to this thread: Resources for learning R.
Asked
Active
Viewed 2,544 times
1 Answers
18
Power analysis refers to analytical procedures that attempt to determine the power of a statistical test (i.e., the probability of rejecting a false null hypothesis) or the sample size (i.e., $N$) required to achieve a given power. You can search Cross Validated for more information about power analysis by clicking here: power-analysis, here: power, or possibly here sample-size.
Free software for power analyses:
- G*Power
- PS: Power and sample size calculator
- Optimal Design software - statistical power analysis for multi-level and longitudinal research
Websites with free sample size calculators:
- Russ Lenth has a java-based web app
- Michael Friendly has a power analysis / sample size calculator for ANOVA designs
- Rollin Brant's webpage for comparing two proportions
- OpenEpi has sample size/power calculators for basic epidemiological study designs.
Open source programming languages with power analysis libraries / modules:
R packages:
- pwr is the oldest power-analysis library; some introductory info can be found on Quick-R
- PoweR: Computation of power and level tables for hypothesis tests
- Power2Stage: Power and Sample size distribution of 2-stage BE studies via simulations
- powerAnalysis: Power analysis in experimental design
- powerGWASinteraction: Power Calculations for Interactions for GWAS
- powerMediation: Power/Sample size calculation for mediation analysis, simple linear regression, logistic regression, or longitudinal study
- powerpkg: Power analyses for the affected sib pair and the TDT design
- powerSurvEpi: Power and sample size calculation for survival analysis of epidemiological studies
- PowerTOST: Power and Sample size based on two one-sided t-tests (TOST) for (bio)equivalence studies
- longpower: Power and sample size for linear model of longitudinal data
There is a Python module, statsmodels, that can run some power analyses. There is some information here.

Steve
- 611
- 6
- 26

gung - Reinstate Monica
- 132,789
- 81
- 357
- 650
-
7Anyone who is aware of a freely available resource that is not listed already is encouraged to edit this to include that information. – gung - Reinstate Monica Jul 05 '13 at 04:12