|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Modeling latent variables interaction |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
Maddy Ong posted on Saturday, September 03, 2016 - 1:25 pm
|
|
|
I have an interaction in my model between two latent variables. xy | x XWITH y; I am only interested in the effect of my interaction term (xy) on 3 dependent variables. Theoretically, I don't care and do not hypothesize about the effects of the two latent variables (x and y) on the 3 dependent variables. Do I still need to model the two original latent variables, or do I just model the interaction term? z1 on x y xy; z2 on x y xy; z3 on x y xy; OR z1 on xy; z2 on xy; z3 on xy; |
|
|
You typically want to include the "main" effects as well. |
|
Back to top |
|
|