Description
Bayesian Analysis for Multivariate Categorical Outcomes.
Description
Provides Bayesian methods for comparing groups on multiple binary outcomes. Includes basic tests using multivariate Bernoulli distributions, subgroup analysis via generalized linear models, and multilevel models for clustered data. For statistical underpinnings, see Kavelaars, Mulder, and Kaptein (2020) <doi:10.1177/0962280220922256>, Kavelaars, Mulder, and Kaptein (2024) <doi:10.1080/00273171.2024.2337340>, and Kavelaars, Mulder, and Kaptein (2023) <doi:10.1186/s12874-023-02034-z>. An interactive shiny app to perform sample size computations is available.
README.md
bmco
Bayesian methods for comparing groups on multiple binary outcomes.
Installation
# From CRAN (when accepted)
install.packages("bmco")
# Development version
# devtools::install_github("XynthiaKavelaars/bmco")
Quick Example
library(bmco)
# Generate data
set.seed(123)
data <- data.frame(
treatment = rep(c("control", "drug"), each = 50),
outcome1 = rbinom(100, 1, 0.5),
outcome2 = rbinom(100, 1, 0.5)
)
# Analyze
result <- bmvb(
data = data,
grp = "treatment",
grp_a = "control",
grp_b = "drug",
y_vars = c("outcome1", "outcome2"),
n_it = 10000
)
print(result)
Functions
bmvb(): Basic group comparisonbglm(): Subgroup analysisbglmm(): Multilevel data
Getting Help
See vignette("introduction") for detailed examples.
Acknowledgements
The statistical underpinnings of this package were developed with financial support of a NWO (Dutch Research Council) research talent grant (no. 406.18.505) and the theoretical insights of Maurits Kaptein (Eindhoven University of Technology, The Netherlands) and Joris Mulder (Tilburg University, The Netherlands).