1

I am trying to analyse the effect of a treatment on the abundance of three response variables (either as absolute values, in which case they are counts with some zeros, or as relative proportions). Plots are stratified across 5 sites, each with 2 or 3 replicates of each of the 5 treatment levels. Each plot was sampled 3 times.

Because of the structure of my response variable I think I should use a generalised linear model, and was planning to analyse each of my three response variables separately. To take into account the repeated measures on the same individual, I suspect a generalised linear mixed effects model is likely to be the only way to proceed. The treatment would be my fixed effect, and site/block/season the random effects.

Can anyone suggest whether a more simple analysis might be possible? More specifically, is it possible to take into account repeated measures (at three distinct time points) within a generalised linear model?

Thanks in advance for any feedback, and apologies if this is too vague.

flummoxed
  • 11
  • 1
  • Welcome to the site! Generally speaking, GLM assumes independent observations and thus cannot handle repeated measures. GLMM can provide subject (or site/block/season in your example) specific estimates, but the marginal model (e.g. generalized estimating equations) is another solution. The clarification between subject specific and population estimates can be found [here](http://stats.stackexchange.com/questions/59137/subject-specific-vs-population-average-predictions/68531#68531). – Randel Sep 23 '13 at 00:01

0 Answers0