[R] lme and varFunc()

[Andrew Robinson]

Christoph,

----- Original Message -----
From: Christoph Scherber <Christoph.Scherber at uni-jena.de>
Date: Wednesday, January 26, 2005 7:57 pm
Subject: Re: [R] lme and varFunc()

> Dear all,
> 
> I am expecting a Poisson error distribution in my lme with 
> weights=varFunc().
> 
> The "weigths= varPower (form= fitted (.))" doesn´t work due to 
> missing 
> values in the response:
> 
> Problem in lme.formula(fixed = sqrt(nrmainaxes + 0...: Maximum 
> number of iterations reached without convergence. 
> Use traceback() to see the call stack

Are you certain that this is caused by missing values in the response?  It looks like something quite different to me.  How did the response come to have missing values?  

> That´s why I´ve used one of my most important explanatory 
> variables as a variance covariate:
> 
> weigths= varPower (form=~explanatory)
> 
> With that, it worked out properly so far.
> 
> What would your suggestion in such a case be?

Does it satisfy your needs with regards to the model diagnostics?  If so, then ask yourself: am I fitting a Poisson because it is an important and intrinsic part of the model that I need to construct, or am I fitting it in order to satisfy the model assumptions?  If the latter then it seems that the Poisson might be unnecessary.

I hope that this helps,

Andrew