[R] R 3.0.0

Ben Bolker bbolker at gmail.com
Fri May 24 20:23:34 CEST 2013


David Winsemius <dwinsemius <at> comcast.net> writes:

> 
> 
> On May 24, 2013, at 8:29 AM, Simmons, Susan J. wrote:
> 
> > In the newest release of R (R 3.0.0), the "glm" package
>  no longer supports logistic nor probit regression. 
> Was this intentional?  What package is best to use in the 
> newest version of R to perform logistic or probit regression?
> > 
> 
>  I would be very skeptical about that claim. 
> Where is your evidence? Have you looked at:
> 
> ?family  # ?
> ?quasi   # ?

  I'm skeptical too. 'glm()' is the 'stats' package which comes
with R and is automatically loaded.

  glm(...,family=binomial)  runs a logistic regression

(the default link is logit, so this is equivalent to 
family=binomial(link="logit"), if you prefer explicitness
over brevity)

  glm(...,family=binomial(link="probit"))

runs a probit regression.

   I'd actually be quite curious to hear how you reached
this conclusion, in case there is confusing or mis-information
going around that we can correct ...

  Ben Bolker



More information about the R-help mailing list