Checking the Assumptions of a Mixed Effects Model in R

Hi all was hoping of a little bit of advice here.
As far as I am aware, 4 of the assumptions of a mixed effects mode are:
-Normally distributed residuals
-Heteroskedasticity of residuals
-Normally distributed residuals of the random effects potion of the model (i.e. Blups)
-Heteroskedasicity of the random effects potion of the model (i.e. Blups)

Is this correct? If so, then how would you test for it in R?
I am struggling the most with the last one, as I can get qqnorm plots for looking at normality. I have been using the package 'nlme' -

'''model1<-lme(Outcome~Height*Weight,data=data,random=~1|Individual,method="REML")'''
'''residual_1<-fitted(model1,type="pearson")'''
'''hist(residual_1)'''
'''qqnorm(residual_1)""

I think I also figured how to plot the fitted residuals against the he fitted residuals against the residuals to check heteroskedasicit

'''fits<-fitted(model1)'''
'''plot(residual_1~fits)'''

or
"plot(model1)"

Is this correct? What would this look like if the assumption of homoskedasicity was met?
And how do I get R to run this in the random effect residuals instead?

Many thanks!

Hi, and welcome!

Preliminarily, see FAQ: What's a reproducible example (`reprex`) and how do I do one? for how to attract more answers by reducing the setup time to recreate the problem and better understand the question.

Take a look at [this vignette from the DHARMa package(https://cran.r-project.org/web/packages/DHARMa/vignettes/DHARMa.html) for a discussion of the issues.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.