Justifying Alpha Levels for Hypothesis Tests

Functions to justify alpha levels for statistical hypothesis tests by avoiding Lindley's paradox, or by minimizing or balancing error rates. For more information about the package please read the following: Maier & Lakens (2021) ).

JustifyAlpha

Daniel Lakens and Maximilian Maier 2021-06-08

R Package accompanying “Justify Your Alpha: A Primer on Two Practical Approaches”

The goal of JustifyAlpha is to provide ways for researchers to justify their alpha level when designing studies. Two approaches are currently implemented. The first function optimal_alpha allows users to computed balanced or minimized Type 1 and Type 2 error rates. The second approach uses the function ttestEvidence or ftestEvidence to lower the alpha level as a function of the sample size to prevent Lindley’s paradox.

Installation

You can install the released version of JustifyAlpha from GitHub with:

devtools::install_github("Lakens/JustifyAlpha")

Preprint

A preprint explaining how to use this package and the Shiny app is available from here:

Vignette

A vignette explaining how to use this package and the Shiny app is available from here: https://lakens.github.io/JustifyAlpha/articles/intro_to_justifieR.html

Shiny App

You can run the shiny app locally, but an online version is available from https://shiny.ieis.tue.nl/JustifyAlpha/ and https://maxma1er.shinyapps.io/JustifyAlpha/.