[R] robust standard error

[Frank E Harrell Jr]

On Thu,  9 Oct 2003 13:26:57 +0200
"daniele\.frison" <daniele.frison at email.it> wrote:

> How can I do a 
> Multiple Linear Regression Models
>  with robust standard error?
> Thank in advance.
> D.F.
> 
> 

One way:

install.packages('Hmisc'); install.packages('Design')
library(Design)
f <- ols(y ~ x1 + ...., x=T, y=T)
g <- robcov(f)  # or bootcov to use bootstrap s.e.'s

---
Frank E Harrell Jr    Professor and Chair            School of Medicine
                      Department of Biostatistics    Vanderbilt University