On the multivariate probit model for exchangeable binary data with covariates

Area of Science:

• Biostatistics
• Statistical Modeling
• Epidemiology

Background:

• Analyzing correlated binary data is crucial in various scientific fields.
• Existing models may lack flexibility in handling intracluster associations and covariates.
• Familial disease aggregation and developmental toxicity present complex binary outcomes.

Purpose of the Study:

• To present a multivariate binomial probit model for correlated binary data.
• To demonstrate its capability in accommodating cluster and individual-level covariates.
• To illustrate Bayesian estimation methods for parameter inference.

Main Methods:

• Utilized a multivariate binomial probit model.
• Employed Bayesian estimation techniques.
• Applied Gibbs sampling for posterior density derivation.

Main Results:

• The proposed model effectively analyzes correlated exchangeable binary data.
• It allows for flexible intracluster association structures.
• Demonstrated applicability in familial disease aggregation and developmental toxicity studies.

Conclusions:

• The multivariate binomial probit model offers a robust framework for correlated binary data analysis.
• Bayesian estimation with Gibbs sampling provides a viable inference approach.
• The model is well-suited for complex datasets in public health and toxicology.