I want to investigate Y ~ X1 * X2 + (1|ID on this dataset (there's a plot of these data in that post too, it's the same dataframe)
Y is a continuos outcome variable. X1 and X2 are both two 2-level categorical predictors. X1 ("CATEGORIES") accounts for "Year of testing" and X2 "test1 or test2" (MY_GROUP). All data come from the same group of participants (n = 21), each participant has four scores (n = 84). I wanna see the possible effects of X1 and X2 on Y as well as if they're interacting
Problems:
mod1 <- lmer(CONT_Y ~ CATEGORIES * MY_GROUP + (1|PARTICIPANTS), data = myData3)
- My question is: would a robust lmer regression account for these problems? I mean, probably it wouldn't be "THE best" solution, but how well would a robust lmer fit these data?
library("robustlmm")
mod2 <- rlmer(CONT_Y ~ CATEGORIES * MY_GROUP + (1|PARTICIPANTS), data = myData3)
note: I didn't want to exclude any participants since I only have 21...
note2: I've never performed a robust regression/bootstrapping/sandwich before, any tips would be much appreaciated :)
note 3: I'm kinda running out of time to hand in this analysis, so I'm aware that maybe, a robust fit isn't "THE" best, as I've said, but would it be enough for a preliminary analysis? Mainly: what problems would it correct for, which would remain?
A repeated measures Anova would be just fine in my case, bu my date doesn't seem to be normally distributed and I didnt find a non parameteic alternative..
Any thoughts would be much appreciated!!
