Plotting latent variables after SEM

[Tim Goetz]

Hi at all,

I am currently running a SEM Analysis but I really dont know how to calculate margins of my latent IVs.
Just assume following code:

sem ///
(OUTCOME1 -> y1 y2 y3 ) ///
(LATENT1 -> x1 x2 x3 x3 ) ///
(LATENT2 -> z1 z2 z3 z4) ///
(INT12 -> int1 in2 in3) ///
(OUTCOME <- LATENT1 LATENT2 INT12 covariates)
, latent(OUTCOME LATENT1 LATENT2 INT12)

I want to calculate the margins of my latent IVs LATENT1 and LATENT2 on OUTCOME, but I really dont know how. Afterwards I want to plot the graph with marginsplot.
When trying to use margins the same way as I do after a Regression command I get the failure that the postulated variables (e.g. LATENT1) cant be found within my specified covariates, e.g. the following code:

margins, at(LATENT1=(1(1)5) LATENT2=(1(1)5))
   marginsplot

Does anyone know hot to calculate margins with latent variables and how to specifiy the commands correctly?
If it is not possible plotting the graphs this way, is there another one? Any hints may be helpful..

+ If anyone knows whether there are differences in how to plot latent variable interactions (I think the same way, as I specify the interaction the same way in Sem as normal latent variables [via double mean-centering Approach (INT12)], let me know.

Thanks for your help!! I Need to find a solution quickly, I hope you can help me..

Best Regards,
Tim

[Francesco Visconti]

Dear Tim,

I have a similar issue (I want to plot the marginal effects of a latent variable). Did you come up with a solution?

Thanks for your attention.

Best regards,

FV