About SEM

[Bob Wen]

I want to run sem on the famous data set mroz.dta from Wooldridge's introductory econometrics.

sem (hours<-lwage educ age kidslt6 nwifeinc) (lwage<-hours educ exper expersq), cov(e.hours*e.lwage)

but it doesn't work.

[Joseph Coveney]

Maybe avoiding (hours <- lwage) (lwage <- hours) might help.

[Joseph Coveney]

Well here, to be clearer:

(hours <- lwage) (lwage <- hours)

covvariance(e.hours*e.lwage)

Pick one.

[Bob Wen]

Why when i add cov(e.hours*e.lwage) to sem, it doesn't converge, without the cov, it give the result?