| Message/Author |
|
|
|
|
In Mplus can you use ML to estimate a model with two dichotomous dependent variables and specify that the variables are dependent conditional on the independent variables. A simple example would be
D1 on X1 X2;
D2 on X1 X2;
The independent variables are dummy coded.
Thank you.
Jamie |
|
|
|
|
| Yes. You can't use WITH but you can introduce a factor that is measured by D1 and D2, fixing its variance at 1 and letting the D2 loading pick up the residual covariance. It takes 1 dimension of integration which is fine. |
|
|
|
|
Hi:
Just to check, add the following to the model
F by DV1 DV2;
F@1;
Thank you,
Jamie |
|
|
|
|
| Right. Make sure that F doesn't have a mean (if it's mgroup or mixture) and that F is not correlated with any other variables. |
|
|
| Back to top |
