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

## Description

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

## Usage

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

## Arguments

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

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23``` ```##---- 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) } ```

WRS documentation built on May 31, 2017, 4:03 a.m.