Latent growth curve modeling in STATA 13 sem framework

[SRL]

Hello,

I have been following Acock's Discovering SEM using STATA like a bible for my research.
However, there are some unresolved issues that are not addressed in the book and I am unsure whether I can ignore and proceed or not.

1. I have a panel data with survey/sampling weights. I am not sure if I ever saw or know how to apply survey weights when running sem. Any hints or can I proceed with the raw data?
2. My dependent variable is a binary outcome (satisfaction coded as 0: not satisfied 1: satisfied). It seems like other software use maximum likelihood estimation method to go about binary response variables (correct me if I am wrong....this is based on quick google search). I have not found satisfactory answer in STATA . Acock's chapter on latent growth curve does not address restrictions on dependent variable too much. Does this mean that I can run the latent growth model on binary outcomes? Or should I be more cautious?

Thanks so much in advance!

[Rich Goldstein]

please read the FAQ

re: your first question, it is clear from the help and from the manual that pweights are allowed - what is confusing about this?

your second question is not at all clear to me - if you're outcome is binary, how does "growth" come into it? there is only one possible change (either 0 -> 1 or 1 -> 0) - please clarify after reading the FAQ

[Nick Cox]

The request by Clyde Schechter made in January http://www.statalist.org/forums/foru...ysis-using-sem to re-register with a full real name still stands.

Rich Goldstein just suggested that you read the FAQ, which is at http://www.statalist.org/forums/help Our practice of using full real names is explained in Section 6.