Are there any parametric assumptions for Linear Mixed Models - lmer (multivariate model)? I saw a few videos or journals that have reported results without resulting in any assumptions? any answer with reference will be appreciated
The fact that they didn't mention assumptions doesn't mean there aren't any. Many strong assumptions are made, e.g.,
- The right hand side of the model is structurally correct with regard to nonlinearities and interactions
- The overall error term is normal with constant variance
- The random effects are normal with constant variance
- The within-subject correlation is constant. If analyzing longitudinal data this is unlikely to be true, as typically correlations are stronger for measurements taken closer together in time.