Hi, I'm a biology grad student in Canada and I need some help.
I have abundance data for almost 400 populations and I am trying to estimate trends in population sizes over time. I am using the glimmix procedure to estimate the effect size of time using a log link and a negative binomial error structure. My random effect is each different population and my problem is that I don't know what code to use. I thought I had a G-side random effect, but when I try to run it that way ("random popid") the program considers all my data to be from one subject, when really there are 384 different populations. Do I really have an R-side random effect, and if so do I consider my random effect and an intercept or residual?
Any help, direction, or advice anyone could offer me would be greatly appreciated.