# Glm or glmm, repeated measures?

I would really apreciate some help for my statistical analysis. My dependent variable is the number of insects selecting one habitat from another(count). There where 4 stands, two from each kind. In each stand there where two kind of habitat, from which the insects had to choose. 2 of the stands had habitat A and habitat Z. The other 2 stand had Habitat B and habitat Z. Insects where obteined in the forest and released in the middle between 2 of these two habitat.

The stands where closed, so we knew how many insects where inside the stands and we where able to free them in the middle between the two habitat kinds and then recapture them whenever they had chose one habitat or the other, having our result. In each replication I released 20 individuals, replicated 7 times per stand, so 14 times per stand type, always with different insects. Afterwards, I changed Habitat Z in both stands types, to Habitat x. The stands were the same but changed the conditions of Habitat Z for assesing the difference in the insect habitat selection.

I analyse the data with a glm, binomial distribution (as there were two possible answers, habitat A/B or Habitat Z, and afterwards, Habitat A/B or Habitat X so:

Model<- glm(cbind(N°insects recaptured in Hab A/B, total insects- habitat recaptured in Hab A/B) ~ A/B + Z/X + stand + replicates, family= poisson)

But i have received some comments that make me doubt this. For example, they say this are repeated measures (as the insects are always meassured in the same plots) but i believe that for being repeated meassures the dependent measure shoud be the one repeated, and the insects were always different.

They also say that as i used two stands for each type of treatment, i should make a random factor called stand and do a glmixed model.

I would like your opinion. Eather way my factor "stand" and "replicates" where really non significant so I simplified the model just for the first two factors, as i have been told that it is better this way for seing better the results. Here a fragment of the results (just for explanation, not how I had them in r project) ﻿﻿ ﻿ ﻿