[R] help: advice on the structuring of ReML models foranalysing growth curves

[Andrew Robinson]

I agree with Bert.  The lme() helper functions are much more developed
than the lmer() helper functions.  This is probably relevant for
Simon's data because temporal autocorrelation is likely for the
measurements within chicks, and is easily handled in lme().  I'm not
sure if it can be done yet in lmer().

Cheers

Andrew

On Tue, Sep 05, 2006 at 09:14:20AM -0700, Berton Gunter wrote:
> 
> > -----Original Message-----
> > From: r-help-bounces at stat.math.ethz.ch 
> > [mailto:r-help-bounces at stat.math.ethz.ch] On Behalf Of Andrew Robinson
> > Sent: Tuesday, September 05, 2006 7:25 AM
> > To: Simon Pickett
> > Cc: r-help at stat.math.ethz.ch
> > Subject: Re: [R] help: advice on the structuring of ReML 
> > models foranalysing growth curves
> > 
> > Hi Simon,
> > 
> > overall I think that lmer is a good tool for this problem.  It's
> > impossible to reply definitively without the full details on the
> > experimental design.
> > 
> > Caveat in place, I have questions and some suggestions.  Are
> > treatment1 and treatment2 distinct factors, or two levels of a
> > treatment, the dietary compound?  Also, what is broodsize?
> > 
> > If you want to nest chick id within brood, I think that you should
> > include the interaction as a random factor.  If you'd like the age
> > effects to differ between chicks then age should be on the left of id.
> > 
> > Thus, start with something like ...
> > 
> > model1 <- lmer(weight ~ treatment +  broodsize + sex + age
> >        + (1|brood) + (age|id:brood), data=H) 
> 
> 
> FWIW, this model can also be easily fit with the lme() function (in the nlme
> package) as the random effects are strictly nested. The only advantage in
> doing so is that the lme tools for examining the model are somewhat more
> developed and extensive (or am I just more familiar with them?)
> 
> Cheers,
> Bert
> 
> - 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

-- 
Andrew Robinson  
Department of Mathematics and Statistics            Tel: +61-3-8344-9763
University of Melbourne, VIC 3010 Australia         Fax: +61-3-8344-4599
Email: a.robinson at ms.unimelb.edu.au         http://www.ms.unimelb.edu.au