To support the use of the Bayesian multivariate analysis methods for multiple binary outcome variables featured in this blog post, I created three open-source tools that work together nicely:
Interactive exploration:
Decision-making with multiple binary outcomes
An interactive online tool that provides a practical introduction to Bayesian analysis of multiple binary outcomes, guiding researchers through the steps required to compare data from two groups. The app makes advanced statistical methods accessible without requiring coding, allowing users to analyze clinical trials and studies where multiple yes/no outcomes need to be evaluated simultaneously.
Multivariate analysis:
bmco
An R-package that provides Bayesian methods for comparing groups on multiple binary outcomes, including basic tests using multivariate Bernoulli distributions, subgroup analysis via generalized linear models, and multilevel models for clustered data. The package is particularly useful for researchers conducting clinical trials or behavioral studies where several correlated binary outcomes must be analyzed together, with proper accounting for their relationships rather than treating them independently.
Power analysis:
bmco-pwr
An interactive Shiny app that performs sample size computations CRAN for Bayesian analysis of multiple binary outcomes. Researchers can use this tool to plan their studies by calculating how many participants are needed to achieve adequate statistical power when comparing groups on multiple correlated binary outcomes — helping to design studies that are neither underpowered nor wastefully overpowered.