Re: [R] nonparametric mixed effect model

Hi, Usually species richness data have very heterogeneous variance, so they are better described by a Poisson distribution than by a Gaussian. you should probably analyse your data with Generalized Linear Models using function lmer (from package lme4) and specify family=poisson. hope this helps

[R] nonparametric mixed effect model

Hi all, I'd like to compare species richness patterns of two groups with paired observations with an extra continuous fixed effect as a covariate. Normally I would use a standard general mixed effect model for such data, with one group variable and a continuous variable as fixed effects, plus a r