Std coeff. with continuos and binary ... PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Tobias Ludwig posted on Monday, December 12, 2016 - 5:45 am
Hey all,

I have a model with 4 exogenous and 4 endogenous LVs in a SE model.

en1 ON ex1 ex 2 ex3 ex 4;
en4 ON ex1 ex ex3 ex4;

Now I want to expand this model so that a observed binary outcome gets regressed on the 4 endogenous LVs (logit regression on the endogenous LVs)

obsBinary ON en1 en2 en3 en4;

I think this is equivalent to a mediation model with 4 IVs, 1 DV and 4 mediators. The thing is that all the coefficients for the regressions between the 4 IVs and the 4 mediators then change to a different scale (standardized and unstandardized output).

Is this because the end DV is binary?
Why does it happen?
How can I interpret these coefficients then?

(It would be just nice to have r-like coefficients for the regressions between the IVs and the mediators and odds for the regressions between the mediators and the outcome. I do not necessarily need to test for mediation, it would be nice to control for some influence of the IVs before regressing the binary outcome onto the 4 mediators / endogenous LVs.)

Furthermore, when I add the binary observed outcome to my model I can't report fit indices anymore. What´s best practice for that?

I would be very happy if somebody could help me out.
 Bengt O. Muthen posted on Monday, December 12, 2016 - 4:24 pm
We need to see the 2 outputs - send to Support along with your license number.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message