designpower: Design Power Analysis Tools

Onkelinx, Thierry¹²³ Research Institute for Nature and Forest (INBO)⁴⁵

keywords: sample size, power analysis, design optimization

Tools for iterative power analysis and design optimization using simulation and adaptive sampling strategies.

Installation

You can install the development version from GitHub with:

# install.packages("remotes")
remotes::install_git("https://github.com/inbo/designpower")

Example

The package provides tools for iterative power analysis. Here’s a basic example:

library(designpower)

# Define your simulation function
my_sim <- function(design, n_sim, ...) {
  replicate(n_sim, {
    # Your simulation logic here
    # Should return p-values
    runif(1)
  }) |>
    list(p = _)
}

# Find optimal design parameters
result <- find_power(
  design = list(effect_size = 0.3, n_per_group = 30),
  design_digits = c(effect_size = 2, n_per_group = 0),
  opti = "effect_size",
  sim_power = my_sim,
  extra_args = list(),
  power = 0.8,
  alpha = 0.05
)

result

Functions

The main functions in designpower are:

• find_power() - Iteratively search for optimal design parameters using adaptive simulation
• sim_power() - Example power simulation function
• observed_power() - Calculate observed power from simulations (internal)
• get_design_id() - Manage design identifiers in the database (internal)
• sample_new_design() - Determine next design parameters using GAM modelling (internal)