Rapid and comprehensive power analysis in R, with a focus on planned error control for multiple outcomes

Friday, November 8, 2024

2:30pm to 4:00pm ET

Presenters:

Kristen Hunter, Harvard University

Luke Miratrix, Harvard University

 Researchers planning on assessing impacts on multiple outcomes in a randomized trials need to account for multiple testing adjustments for controlling error rates when thinking through whether their experiment is adequately powered, or for calculating necessary sample sizes to achieve a given level of power. With multiple testing, however, even the definition of power is more complex: are you powering to detect all effects?  To detect some fraction of effects? In this webinar we will learn about defining power under multiplicity, and also learn how to use PUMP, an R package designed to calculate power when planning to use false discovery rate control or family-wise error control methods such as Bonferroni or Westfall-Young.  With PUMP researchers can easily calculate how small of an effect could reliably be detected, or how large of a sample would need to be obtained, or how powerful a given experiment would be, for several definitions of power and several options for multiple-testing adjustment, given a specific experimental design and planned analytic model (along with the usual design parameters).

 This webinar will be most helpful for those with basic R familiarity and at least some comfort with basic (single-outcome) power analyses.  It is particularly targeted for those involved in research demanding precise power analysis in complex experimental setups.

 As part of the webinar materials, we will provide vignettes that walk through power analysis for a variety of illustrative scenarios. We will also point attendees to a Shiny app that wraps the core package, allowing for a web interface for those who prefer that modality.

 ____________________________________________________________________________________________________________________________________________