Launch the power and precision app — run_power_precision

Opens an interactive Shiny app for exploring how sample size, explanatory variables, interactions, and repeated measurements affect statistical power and confidence intervals for a known, fictional caffeine effect.

Usage

run_power_precision_app(...)

Arguments

...: Optional arguments passed to shiny::runApp(). By default, statsapps opens apps in the system default web browser using launch.browser = TRUE. Use launch.browser = FALSE to disable this behavior.

Value

run_power_precision_app() opens the Shiny app locally.

Details

All data are simulated. Sample size ranges from 12 to 180 people in steps of 6, with 42 people by default. Equal numbers are assigned to each combination of drink and age group. The true caffeine-minus-decaf effect is fixed at -25 milliseconds.

Optional repeated measurements provide 1 to 5 observations per person. When person identity is accounted for, a random-intercept model is fitted by restricted maximum likelihood (REML). Predictors are constant within each person. Mixed-model confidence intervals and power are approximate. This is not a general-purpose mixed-model fitter.

Power is calculated from the known simulation parameters, not from estimates fitted to the displayed outcomes. Confidence intervals describe the caffeine-minus-decaf effect estimated from the displayed sample. When the drink-by-sleep interaction is included, this contrast is evaluated at 7 hours of sleep.

Examples

if (FALSE) { # interactive()
run_power_precision_app()
}