SEM model with categorical and continuous variables

[Cameron Whitley]

Hello,

I have 2 exogenous continuous variables and 2 endogenous variables (one continuous and one dummy variable).

(DV1<- X1, X2) (DV2<-DV1 X1 X2) (DV3<-DV1 DV2 X1 X2)

DV1-continuous (endogenous)
DV2-categorical (endogenous)
X1-continuous (exogenous)
X2-continuous (exogenous)
DV3-continous (exogenous)

Question: I'm fairly new to SEM and I want to run a model with these vars, but my understanding is that sem works with linear models and gsem works with categorical exogenous variables, but what if you have both? Is it appropriate to run a gsem model?

Question: I am very familiar with seeing the R2 as variance explained by the model, but what syntax do you use to get the proportion of total variance explained by each variable?

Thanks for the assistance. Like I said, I'm somewhat new to these models.

Cam

[Richard Williams]

I think you can just use gsem. The manual says

In gsem, responses are continuous or binary, ordinal, count, or multinomial. Models are linear regression, gamma regression, logit, probit, ordinal logit, ordinal probit, Poisson, negative binomial, multinomial logit, and more.

Having said that, I am not sure you need to get that fancy. Doesn't the model have 0 d.f.? You could just run separate logit/ologit/regress commands, I think

[Cameron Whitley]

Thanks. That's what I was thinking about gsem as well, but I wasn't sure.