Skip to content
Prev 301 / 20628 Next

boneheaded (?) question about SS in anova (lmer vs lme)

On 8/27/07, Ben Bolker <bolker at zoo.ufl.edu> wrote:

            
Thanks for the questions, Ben, and for the well-constructed example.

Before going into detail about how the analysis of variance table is
constructed for an lmer model, I thought that I would do a bit of
plotting of the data.  Some of the plots indicate that a linear mixed
model is probably not appropriate.  Consider

library(lattice)
print(dotplot(reorder(PATCH, ALGAE) ~ ALGAE,  urchins, ylab = "Patch"))

and the same plot for the TREAT variable (PDF files attached).

I would be much more inclined to use a generalized linear mixed model
with a Poisson conditional distribution than a normal conditional
distribution for these data.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Urchins1.pdf
Type: application/pdf
Size: 8948 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-mixed-models/attachments/20070831/aef3e375/attachment.pdf>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Urchins2.pdf
Type: application/pdf
Size: 7926 bytes
Desc: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-mixed-models/attachments/20070831/aef3e375/attachment-0001.pdf>