Re: R-beta: Various R Questions and Comments

Bill Venables (wvenable@attunga.stats.adelaide.edu.au)
Wed, 18 Feb 1998 08:13:38 +1030


Date: Wed, 18 Feb 1998 08:13:38 +1030
Message-Id: <9802172143.AA06075@attunga.stats.adelaide.edu.au>
From: Bill Venables <wvenable@attunga.stats.adelaide.edu.au>
To: Jim Lindsey <jlindsey@luc.ac.be>
Subject: Re: R-beta: Various R Questions and Comments
In-Reply-To: <9802170822.AA21240@alpha.luc.ac.be>
 <9802170822.AA21240@alpha.luc.ac.be>

Jim Lindsey writes:
 > Offsets only are really useful if the link function is not the
 > identity. Hence little use in lm. Jim

They are convenient, though.  It is easy just to correct the
dependent variable by subtracting the offset, but then when you
want to predict from the fitted model object you predict the
offset-corrected response mean, not the response itself.  So, I
know it's easy to add it back on, but that's another little thing
not to forget...

I would favour having offset available for all lm/aov/glm/gam/...
fitting functions.

Bill Venables.

-- 
Bill Venables, Head, Dept of Statistics,    Tel.: +61 8 8303 5418
University of Adelaide,                     Fax.: +61 8 8303 3696
South AUSTRALIA.     5005.   Email: Bill.Venables@adelaide.edu.au
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._