st: SEM: Using individually-varying time scores in latent growth model

[Jacqueline Homel <jhomel@uvic.ca>]

Dear List,

I am using SEM (Stata 12) to estimate a latent growth model of 
depression measured at four time points between adolescence and early 
adulthood. Previously I have estimated this model in MPlus, and now I am 
trying to learn how to produce the same results using SEM. The below 
model works fine-

sem (dep_3 <- Idep@1 Sdep@0 _cons@0)
(dep_5 <- Idep@1 Sdep@1 _cons@0)
(dep_7 <- Idep@1 Sdep@2 _cons@0)
(dep_9 <- Idep@1 Sdep@3 _cons@0),
latent(Idep Sdep)
 var(e.dep_3@var e.dep_5@var
 e.dep_7@var e.dep_9@var)
 noconstant means(Idep Sdep) method(mlmv)

However, although observation points were two years apart, not all 
participants were interviewed at the same time, leading to 
individually-varying times of observation. Earlier analyses using MPlus 
used 'time in study' to represent time, rather than 0, 1, 2, 3. Is there 
a way to incorporate these varying time scores into a growth model using 
SEM?

Kind regards,

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