Skip to content
Prev 987 / 7420 Next

R-sig-ecology Digest, Vol 22, Issue 3

Hi,
To add to all the good advice you have gotten, I would second Scott's
advice about using a GAM. 
Two issues that come to mind are that the interpretation might be tricky
as population, I assume
geographical unit, is your "Experimental unit" and individual nests are
within the populations
so fluctuations in population size are going to be correlated with
nesting success. The second is that 
contrasts in GAM's are tricky. Here the model would be success ~ s(time,
by=population), so you have
a curve for each population, but your interest is in the difference of
nesting success. Simon 
in his mgcv package has a tricky way of doing those kinds of
comparisons, you need to get the lp
matrix from the predict function, and then simulate to get the
confidence intervals. There is an example in
in the documentation for predict. This may be a can of worms. If your
time series are not too wiggly
you may be able to get away with using Harrel's rms package, and use
restricted cubic splines in
a logistic regression model. The contrasts may be easier to get.

Hope this helps. Even simple models can get complicated.

Nicholas