Hello everybody!

Currently, I am fitting a structural equation model using gsem in Stata 16.


My code looks as follows:

Code:
gsem (L1 -> X1, family(ordinal) link(logit)) (L1 -> X2, family(ordinal) link(logit)) (L1 -> X3, family(ordinal) link(logit)) (L1 -> L2, ) (L1 -> L3, ) (L1 -> L4 ) (L2 -> Y1, family(ordinal) link(logit)) (L2 -> Y2, family(ordinal) link(logit)) (L3 -> L2, ) (L3 -> Y3, family
(ordinal) link(logit)) (L3 -> Y4, family(ordinal) link(logit)) (L3 -> Y5, family(ordinal) link(logit)) (L3 -> L4, ) (L4 -> L2, ) (L4 -> Y6, family(ordinal) link(logit)) (L4 -> Y7, family(ordinal) link(logit)) (L4 -> Y8, family(ordinal) link(logit)) (L4 -> Y9, family(ordinal) link(logit)), latent(L1 L2 L3 L4 )
Where L1, L2, L3 & L4 are the latent variables (L1 is the only exogenous latent variable). All of the observed items (Y1...Y9 & X1...X3) are ordinal scaled.


Does anybody know how to compute the average marginal effect of e.g. L1 on L2? Unfortunately, I found no proper solution.