Skip to content

how to compure R-squared in glm

5 messages · David Winsemius, Aravindhan, K, Bert Gunter

#
Team,
Can some one help me in computing the R-squared value in glm.

Thanks
Aravindhan
#
Which version of a pseudo-R^2? What?s the model?
This is a plain text mailing list
Please:
David Winsemius, MD
Alameda, CA, USA
#
Hi David,
I am using generalized linear models (glm command with family="poisson"). 

Thanks
Aravindhan

-----Original Message-----
From: David Winsemius [mailto:dwinsemius at comcast.net] 
Sent: Saturday, November 15, 2014 1:16 PM
To: Aravindhan, K
Cc: R-help at r-project.org
Subject: Re: [R] how to compure R-squared in glm
Which version of a pseudo-R^2? What?s the model?
This is a plain text mailing list
Please:
David Winsemius, MD
Alameda, CA, USA
#
You still have not answered what sort of pseudo-R^2 measure you expected.

http://www.econ.ucdavis.edu/faculty/cameron/research/jbes96preprint.pdf

? 
David.
David Winsemius, MD
Alameda, CA, USA
#
... and I would further add to David's comments that R^2 is basically
nonsense for glm's (and mostly for lm's also --  but not a topic for
further discussion here). Likelihood ratio tests are what should be
used to judge model fits for glm's -- also not a topic for further
discussion here. Consult a local statistician, do your glm homework,
or post on stats.stackexchange.com for statistical follow-ups.

Cheers,
Bert


Bert Gunter
Genentech Nonclinical Biostatistics
(650) 467-7374

"Data is not information. Information is not knowledge. And knowledge
is certainly not wisdom."
Clifford Stoll
On Sat, Nov 15, 2014 at 9:00 PM, David Winsemius <dwinsemius at comcast.net> wrote: