Re: st: Adding a time-invariant covariate in SEM builder

[kmacdonald@stata.com (Kristin MacDonald, StataCorp LP)]

Robert Long <W.R.Long@leeds.ac.uk> is fitting a latent growth curve model with
time-invariant covariates and asked how to include an intercept in an equation
for a latent variable when using the SEM Builder.

> After fitting the growth model (that is, the model with just the latent
> growth factors and the measured variables), I add a time-invariant (in
> this case "male") covariate in the do file like this:
>  
> 1. add the path: (Intercept Slope <- male _cons)
> 2. remove the means() option
> 3. and add e. to the factor covariances (if any)
> 
> No problem with that.
>  
> My question is how to achieve the the same thing with SEM builder ?
> Specifically, how do I get _cons into the paths from the time-invariant
> covariate to the latent growth factors ?

Robert can include a box in the diagram representing the constant.  To do this

  1.  Use the Add Observed Variable tool to drop a box on the path diagram;
  2.  In the contextual toolbar, type _cons in the box next to "Variable:"
      and press Enter;
  3.  Draw paths from the _cons box to the latent variables.

Robert can now fit the same model from the Builder that he fits using command
syntax.

-- Kristin
kmacdonald@stata.com




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/