corb: Compute a .95 confidence interval for a correlation
In WRS: A package of R.R. Wilcox' robust statistics functions

The default correlation is the percentage bend.

The function corfun is any function that returns a correlation coefficient in corfun$cor. The functions pbcor and wincor follow this convention.

When using Pearson's correlation, and when n<250, use lsfitci instead.

The default number of bootstrap samples is nboot=599

1	corb(x, y, corfun = pbcor, nboot = 599, SEED = TRUE, ...)

`x`
`y`
`corfun`
`nboot`
`SEED`
`...`

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (x, y, corfun = pbcor, nboot = 599, SEED = T, ...) 
{
    est <- corfun(x, y, ...)$cor
    if (SEED) 
        set.seed(2)
    data <- matrix(sample(length(y), size = length(y) * nboot, 
        replace = T), nrow = nboot)
    bvec <- apply(data, 1, corbsub, x, y, corfun, ...)
    ihi <- floor(0.975 * nboot + 0.5)
    ilow <- floor(0.025 * nboot + 0.5)
    bsort <- sort(bvec)
    corci <- 1
    corci[1] <- bsort[ilow]
    corci[2] <- bsort[ihi]
    phat <- sum(bvec < 0)/nboot
    sig <- 2 * min(phat, 1 - phat)
    list(cor.ci = corci, p.value = sig, cor.est = est)
  }