Hello,
One of the assumptions of SEM is linearity. I am using Mplus with WLSMV estimator for SEM for ordinal variables. In this case, is linearity assumption still required? If so, is there a way to check linearity assumption in Mplus?
Christian Geiser
You could (and probably should!) still look at bivariate scatterplots of your ordinal variables. But it may be tougher to identify potential non-linear relationships that way (compared to continuous variables), especially when your ordinal variables have only few categories.
3 votes
3 thanks
- Badges
- Science topic
- Similar topics
- Mathematics
More Andrew Tate's questions See All
Similar questions and discussions