Hi, in the framework of my internship i have made competition bioassays. To highlight competition, I must present my result as a population rate: Species1/(Species1+Species2) for instance, and that is where it gets tricky: If i only had counts I would be using a poisson or quasipoisson law, however, as I am working with proportions I don't realy know which law would fit the best my glm.
Here's my model: (AMF, my explanatory variable correspond to mycorrhizal treatments)
m0<-glm(evansi~AMF, Comp., family=poisson,
control=glmerControl(optimizer="bobyqa"))
So, if you could help replace poisson by the best family and explain why, it would be so helpful !
Thanks you !