I'm looking for the appropriate test to perform to determine if there is a three-way interaction effect on my (non-parametric) likert-scale data.
My data currently has a dependent variable with values from 1-7 on a likert scale. The study includes three dependent variables, all of which are nominal. I need to check whether there is a relationship between these three that would explain the pattern of responses in the dependent variable. The study design is 2x4x4.
Something like Scheirer Ray Hare but including a three-way interaction? Or should I just go for a non-parametric model? If so, what model would you suggest?
TIA!
