ggpower is an R package and Shiny application for
statistical power analysis. It provides a scriptable API through
power_compute() and a wide-screen GUI with six modules:
Choose a test family and statistical test.
Choose a power-analysis type.
Enter input parameters.
Calculate output parameters.
Inspect distribution plots, X-Y power plots, and protocol
history.
Supported Analysis Modes
A priori: compute required sample size.
Compromise: compute alpha and beta from beta / alpha.
Criterion: compute alpha for a target power.
Post hoc: compute achieved power.
Sensitivity: compute the effect size required for target power.
Supported Test Families
t tests: one-sample, paired, two independent means, point-biserial,
generic t.
F tests: one-way ANOVA, special ANOVA effects, multiple-regression
omnibus, multiple-regression R2 increase, two variances.
Clinical: superiority, non-inferiority, equivalence, Simon
two-stage, cluster RCT.
Some advanced procedures use documented approximations. The result
object includes method notes so these cases are visible in the GUI and
protocol. See the approximation
catalog.