Workshop 7

Bayesian power analysis for multilevel models in R with epower and brms

Barneche D1, Fisher R1
1Australian Institute Of Marine Science

This half-day workshop aims to provide an introduction to power analysis using Bayesian multilevel models in R. We will work through two example cases: 1) a dataset from a Before-After-Control-Impact (BACI) design which will be fitted through the epower R package (backend is INLA); 2) a dataset which will be used to fit a multilevel model using the R package brms (backend is Stan). Therefore, participants are required to have the latest version of R, Stan, rstan, INLA, epower and brms **installed and functioning** on their machines prior to the start of the workshop. This workshop will be free of charge apart from an eventual overhead cost required by the conference organisers.

We will cover the following content:

  1. Introduction to Bayesian power analysis
  2. The challenge of calculating power in Bayesian Statistics
  3. BACI designs and power analysis in epower (theory, code and exercise)
  4. Power analysis of multilevel models in brms (theory, code and exercise)
  5. Live Q&A

The content will be pre-recorded, so the participants are encouraged to listen through the content with ample time and go through the exercises at their own pace, to then join in the live Q&A.