I'm on a journey trying to get my mind around the normality of residuals assumption for linear mixed-effects models, this is a FOLLOW-UP POST
While studying to better understand what I could do to solve the non-normality problem, I came across different types of residuals such as standartized or studentized or conditional or marginal residuals, for example. (also, ref. Field, 2012, chapter 7)
mod1 <- lmer(SCORE ~ X1_c * X2 + (1|PARTICIPANT), data = data)
With THIS dataset. (this is a Git link)
### model residuals:
myResiduals <- tibble(residuals = resid(mod1)) ### this is the function I've been using so far
std and studentized residuals:
myResiduals$standardized.residuals <- rstandard(mod1) ### doesn't work: error message
myResiduals$studentized.residuals <-rstudent(mod1) ### works
- Simply put, my question is: am I testing the normality assumption on the right type of residuals for this model? Any thoughts would be much appreciated! Thanks in advance.
