ripley@stats.ox.ac.uk ripley at stats.ox.ac.uk
Mon Mar 10 17:07:16 CET 2003

R does not contain anything equivalent to step.glm: we do not copy 
incorrect functions!

There is step() in base R and stepAIC() in package MASS, both of which do 
what step.glm in S claims to do (but does not).

[step.glm has an incorrect definition of AIC, and uses Wald tests to 
approximate likelihood ratio tests, despite the Hauck-Donner effect. It 
can give seriously misleading results: this was discovered in a real 
application about 8 years ago, when I wrote stepAIC.]

On Mon, 10 Mar 2003, laurans wrote:

> I would to know the equivalent of the function step.glm for R.
> Thank you

Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

More information about the R-help mailing list