Phillips Perry <pcp35 at ...> writes:
Hi Thomas, I'm anxious to see what more experienced folks say, but in the mean time, as I understand it, your first model is the right one: glmer(species richness ~ Ca + (year|plot), family=poisson, data=data) In this model, "year" is a fixed variable and "plot" is the random variable and you're testing for an interaction between "year" and "plot".
You might want to try r-sig-mixed-models at r-project.org with this one. (Do mention that you tried r-sig-ecology first) Ben Bolker