Skip to content

Dispersion in summary.glm() with binomial & poisson link

2 messages · John Maindonald, Brian Ripley

#
That is why I did not submit a bug report.  The problem is that in
many application areas phi is much greater than one.
I agree with the sentiments.  So would it be feasible to have
quasi-poisson and quasi-binomial errors?  Would an immediate recourse
be to create functions summary.quasi and predict.quasi?  Or perhaps
summary.phi, etc.



John Maindonald               email : john.maindonald at anu.edu.au        
Statistical Consulting Unit,  phone : (6249)3998        
c/o CMA, SMS,                 fax   : (6249)5549  
John Dedman Mathematical Sciences Building
Australian National University
Canberra ACT 0200
Australia
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
#
On Tue, 9 May 2000, John Maindonald wrote:

            
All you need is a function to estimate dispersion as you want
(and there are other ways for gammas, e.g. in the MASS library).
So you call, e.g.

summary(object, dispersion=chisquare.dispersion(object))

but a simpler interface is desirable, and yes, quasibinomial and
quasipoisson look an excellent idea.

Because I was aware of the (in)consistency issues I am looking into this.
Expect something along these lines for 1.1.x.

(BTW, now is a good time to raise such design issues. As you will see from
http://developer.r-project.org 1.1 is about a month off, and our plans are
for 1.1.1 to be the version for teaching in academic year 2000/1 in Europe
and N. America.)

Brian