I am trying to figure out the best method of repeated measures analysis. Here is what I have. 19 years of data collected at 259 grid cells each. Data collected at each grid cell each year (for a total of 4921 observations) includes relative species abundance for 6 species, total acreage shellfish aquaculture and further those aquaculture totals split by acres cultivated of clams, geoduck, oysters and mussels. I am trying to see if there is a relationship between acres shellfish cultivated and bird relative abundances over 19 years, including how might each species respond differently, and how to responses differ by shellfish cultivated.
I have performed a mixed model repeated measures analysis for each species (for a total of 6 analyses) using the following construct model effects: acres (nominal bin, 11 bins total); gridcell[acres] & Random; year; acres*year
I have unsure if I am running the best analysis of my data and would appreciate any advice or feedback.
How is relative species abundance distributed? I'm guessing it is highly skewed and non-normally distributed. You may need to run this model using Proc GLIMMIX in SAS and specify a negative binomial or poisson distribution for the errors. I don't think you can do this in JMP.