Dear list member, as usual I'm trying to perform a very peculiar analysis
and I need help in the syntax process.
I explain:
I have crossed factors with fixed effects and nested factors with random
effect:
First variable-->X1 (random)
Second var. X2 (random nested in X1)
X3 (random nested in X2 and X1 consequently)
X4 (fixed and crossed with X1, X2, X3)
X5 (fixed and crossed with X1 and X2)
this is an UMBALANCED design and the response variable is not normally
distributed (count data of species with a very similar poisson distribution)
What I tried was to use the lme function in the package nlme with this
syntax:
model1<--lme(y~X4+X5, random=~X1+X2+X3 | X3/X2/X1, data=xxx)
I specified the formula following the help page of the lmer formula!
Obviously, R gave me the following result: Not enough memory to write 8000
mAGABYTE.....
In your opinion, a hierarchical design like the one I proposed here, is
analysable in such a way.
Is mine a problem of syntax?
Tank you very much!!
Giovanni
[[alternative HTML version deleted]]