Hello, nlme-experts among the list members,
I have encountered a problem with the function pairs.lme() from the
nlme-library, version 3.3.1 (using S-PLUS, Ver. 6.0 Rel. 1 for Sun SPARC,
SunOS 5.6), about which I haven't found anything in the archive:
If there are more than 2 random coefficients in the model the attempt to
create a conditional pairs plot of the random coefficients (using the |
operator to define separate trellis panels) produces a (for me
untractable) error message.
For example, take a look at the groupedData object Orthodont in Dr.
Pinheiro's and Dr. Bates' book "Mixed-Effects Models in S and S-PLUS":
> Orthodont[ c( 1:3, 108),]
Grouped Data: distance ~ age | Subject
distance age Subject Sex
1 26 8 M01 Male
....
108 28 14 F11 Female
Fitting the model
> O1.lme <- lme( fixed= distance ~ Sex * age, random= ~ age | Subject,
+ data= Orthodont)
and checking the assumptions on the random effects by
> pairs( O1.lme, ~ ranef( .) | Sex)
works perfectly fine yielding a trellis plot with two panels (one for
Male, one for Female), each of which containing the age coefficients vs.
the intercepts.
However, if I extend the model (for what reason ever) to one that is
quadratic in age
> O2.lme <- update( O1.lme, fixed= . ~ . + age^2, random= ~ . + age^2)
the attempt to produce a pairs plot
> pairs( O2.lme, ~ ranef( .) | Sex)
results in the following error mesage:
Problem in eval(expr[[1]], data): Object ".x" not found
Is this a bug or a normal limitation of pairs.lme(), or have I ovelooked
anything? And if it is the first, does anybody have a work-around for the
mentioned problem?
Thank you for any help.
Best regards -- Gerrit
-------------------------------------------------------------------------
Dr. Gerrit Eichner Mathematical Institute of the
gerrit.eichner@math.uni-giessen.de Justus-Liebig-Univ. Giessen
Tel: +49-(0)641-99-32104 Arndtstr. 2, D-35392 Giessen
Fax: +49-(0)641-99-32029 http://www.uni-giessen.de/AG_Stochastik
-------------------------------------------------------------------------
|