Hi,
I am currently running a GSEM model using the following syntax:
gsem( ///
(b6.means_of_travel <- sex i.seg i.household_composition ///
journey_minutes sameresschoolED sameresschool SA ///
corine entropy ///
roads active bus rail house ///
deprivation ///
M1[powscSA], family(multinomial) link(logit)) ///
(journey_minutes sameresschoolED sameresschool SA <- corine entropy ///
roads active bus rail house ///
deprivation) ///
, nocapslatent latent(M1) startvalues(fixedonly))
This GSEM specification works fine. But typically after such models, I run the following postestimation:
margins, dydx(*) atmeans
However, this margins syntax, and most of the other variations [i.e., margins, dydx(*) predict(outcome(*) fixed)] are not working. Specifically, the following error emerges:
r(498): prediction is a function of possibly stochastic quantities other than e(b).
Any indication on where I am going wrong, and advice on how to correctly run my margins syntax, would be greatly appreciated. Thanks in advance.
Conor
I am currently running a GSEM model using the following syntax:
gsem( ///
(b6.means_of_travel <- sex i.seg i.household_composition ///
journey_minutes sameresschoolED sameresschool SA ///
corine entropy ///
roads active bus rail house ///
deprivation ///
M1[powscSA], family(multinomial) link(logit)) ///
(journey_minutes sameresschoolED sameresschool SA <- corine entropy ///
roads active bus rail house ///
deprivation) ///
, nocapslatent latent(M1) startvalues(fixedonly))
This GSEM specification works fine. But typically after such models, I run the following postestimation:
margins, dydx(*) atmeans
However, this margins syntax, and most of the other variations [i.e., margins, dydx(*) predict(outcome(*) fixed)] are not working. Specifically, the following error emerges:
r(498): prediction is a function of possibly stochastic quantities other than e(b).
Any indication on where I am going wrong, and advice on how to correctly run my margins syntax, would be greatly appreciated. Thanks in advance.
Conor
