I have a question about this, does GLMM model like above show inflated type I error as the sample size seems limited. If so, what is a proper way to analyze the effect of "var" and how to do post-hoc?