[R] Centered variables and mixed-model

[Berton Gunter]

(Posted because private reply to sender was kicked back).

Note that you said "centered **response** variables in your original
original post, not centered **explanatory** variables, which is what (your
misquote of) Bates and Pinheiro said. ** Big Difference!** If you don't know
what the difference is, I strongly suggest you consult a statistician.

Centering of explanatory variables can sometimes help numerical accuracy,
but otherwise produces an exactly equivalent linear model with a different
intercept. Obviously, it changes NONlinear models.

-- Bert Gunter
Genentech Non-Clinical Statistics
South San Francisco, CA
 
"The business of the statistician is to catalyze the scientific learning
process."  - George E. P. Box
 
 

> -----Original Message-----
> From: r-help-bounces at stat.math.ethz.ch 
> [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Martin Julien
> Sent: Friday, January 14, 2005 1:44 PM
> To: r-help at stat.math.ethz.ch
> Subject: [R] Centered variables and mixed-model
> 
> I work in biology and I use mixed-model for my data analysis 
> In a scientific paper, the author wrote: 
> "All continuous exploratory variables were centred on their 
> median value
> prior to inclusion in the analysis (Pinheiro & Bates, 2000)."
> They refer to the book "Mixed-effects model in S and S-Plus" 
> by Pinheiro et
> Bates in 2000.
> I feel a bit strange with that paper because I can't find in 
> the book why
> they centred the variables on their median.
> So I have two question:
> First, is it correct to centred the variables on their median in a
> mixed-model?
> Second, why they do that?
> Thank
> Julien 
> 
> -----Message d'origine-----
> De : Liaw, Andy [mailto:andy_liaw at merck.com] 
> Envoyé : 13 janvier 2005 19:26
> À : 'Martin Julien'; r-help at stat.math.ethz.ch
> Objet : RE: [R] Lme and centered variables
> 
> I'm by no means expert in such matter, but where did you get 
> the idea that
> such manipulation is needed?  If you center the response by 
> its mean, that's
> the same as taking out the intercept.  There's no reason to do that
> beforehand.  As to centering by median, that's a first for 
> me, and quite
> frankly I don't see what it buys you.
> 
> Andy
> 
> > From: Martin Julien
> > 
> > I want to fit a linear mixed-model with my data but I want to 
> > know if I have
> > to center all my responses variables on their median and why 
> > I have to do
> > it?
> > 
> > Thank
> > 
> > Julien Martin
> > 
> >  
> > 
> > 
> > 	[[alternative HTML version deleted]]
> > 
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide! 
> > http://www.R-project.org/posting-guide.html
> > 
> > 
> 
> 
> --------------------------------------------------------------
> --------------
> --
> Notice:  This e-mail message, together with any 
> attachments,...{{dropped}}
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
>