Dear readers,
I have a basic question about how to use lme for my design. I haven't been able
to find an example in r-help that made it clear to me how to tackle this
problem and unfortunately I also cannot get hold of Pinheiro & Bates 2000. I
hope someone can help.
Data for the response variable "foan" were collected in:
· 60 plots
· plots were re-sampled yearly for 15 years (certain years are missing
for certain plots)
· plots are divided over 4 reserves (but not equally, so the design is
unbalanced)
· rain is a continuous covariate, which has the same value for all plots
within one year.
Since plots are random within reserves, but reserves can be seen as fixed
factors, it seems I should use a mixed-model and so I have come up with the
following models:
model.1 <- lme(foan ~ year * reserve, data=data, random= ~1 | plot)
model.2 <- lme(foan ~ year * reserve, data=data, random= ~year | plot)
Model.2 does not provide a better fit than model.1, so apparently the slope of
the relationship between "foan" and "year" does not vary for plots.
I want to know if (and how) the response variable changes over the years, while
correcting for variance explained by rain. My question is how to incorporate
"rain" in this model. Can I simply use:
model.3 <- lme(foan ~ year * reserve * rain, data=data, random= ~ 1 | plot)
This somehow seems wrong (forgive my ignorance) since rain has the same value
within each year. On the other hand it has different values within each plot. I
hope someone can point me in the right direction.
Thanks in advance,
Robert Buitenwerf
South African Environmental Observation Network
Phalaborwa
South Africa
_________________________________________________________________
[[elided Hotmail spam]]
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
