[R] Functions in R like lincom and nlcom of Stata

[Thomas Lumley]

On Sun, 14 Dec 2008, John Fox wrote:

> Dear Stas and David,
>
> A couple of people have mentioned linear.hypothesis() in the car package,
> which has methods for several kinds of models and a default method that will
> work for any model that responds to coef() and vcov(). The delta.method()
> function in the alr3 package is similarly general.
>

Despite its name, svycontrast() in the survey package also works for anything with coef() and vcov() methods, for both linear and nonlinear combinations.

       -thomas

Thomas Lumley			Assoc. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle