[Richard Williams]

If I understand correctly, you solved the problem using factor variable notation, as suggested by Richard. That's just fine. But my previous post was about finding out why xi: was giving the error you mentioned in your opening post. You didn't give enough information to solve that particular problem.

I am often content just to have something work and not care why something else didn't work. In this case though, if it was me, I might have a lingering fear that my installation of Stata was corrupted. I don't see any problem with the original code. But, there may be problems with other code we aren't seeing. Luisa may just want to make sure there isn't some bigger problem lurking out there.

[Carlos Avellaneda]

Perhaps Luisa can try this:

webuse lbw
xi: logit low age lwt i.race*i.smoke ptl ht ui

which runs fine by me. If this doesn't work, that would mean your data set is not corrupt and you have probably some installation issues; then you should probably contact Stata Tech Support.

[Luisa Márquez]

I tried it but I didn't work. Anyway, as when I don't use "xi:" I can get results as well, for me the problem is solved. Thank you once again!

I would like to ask you another question. I am interested in Structural Equation Model, that is to say, command SEM. However, my dependent variable is a binary variable. I have just realized that I cannot use SEM command if the dependent variable is binary. I need to use command GSEM. Unfortunately, as far as I know, this command has been introduced in Stata version 13. Could you let me know if there is another way to develop a Structural Equation Model with a binary variable without using GSEM command?

Thank you so much in advance.

[Richard Williams]

Luisa, I would start a new thread, as this is a major change of topic. Also, if you tried Carlos' example and it didn't work, I would be worried about a corrupted installation. Even if you seem to be ok now, you could encounter problems later. Or you may even be getting problematic results without realizing it. It is up to you, but I'd be worried if a simple problem like Carlos presented failed to run.

[Luisa Márquez]

Yes, you are right Richard. May be it´s a corrupted installation. I´ll review it.