pomontiglio-uqam

    set.seed(0)
    nobs = 100
    nind = 100
    nlevels <- 4
    levels.cat <- rnorm(n = nlevels, mean = 0, sd = 2)
    indbt <- rnorm(n = nind, mean = 0, sd = 2)
    id <- rep(1:nind, each = nobs)
    categorique <- sample(x = 1:4, size = nobs*nind, replace = TRUE)
    table.cat <- data.frame(categorique = 1:4, val = levels.cat)