Home

mvreprobit is an R package for estimating multivariate random-effect probit models using efficient Gibbs sampling procedures.

The mvreprobit R package contains code associated with the article

Steele, F., Zhang, S., Grundy, E., and Burchardt, T. (2022). Longitudinal analysis of exchanges of support between parents and children in the UK.

Why mvreprobit?

• Efficient. Tailored code to handle random effects sampling and sampling of correlation matrix parameters.
• A general algorithm which can handle a range of random effect probit models and can be extended easily.
• Concise code with pure R.

Install the mvreprobit package from GitHub

## install required packages
required_packages <- c("remotes", "truncnorm", "mvtnorm", "MCMCpack", "matrixcalc")
new_packages <- required_packages[!(required_packages %in% installed.packages()[,"Package"])]
if(length(new_packages)) install.packages(new_packages)

remotes::install_github("slzhang-fd/mvreprobit")

Usage

Check two examples in

• Demo 1: Four-process 2-level (1 random effect) probit model
• Demo 2: Four-process 3-level (3 random effects) probit model