Let residuals correlate PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 freek bucx posted on Tuesday, February 27, 2007 - 7:26 am
Hi,

I have a question.
I am testing a model in which I have five dummy outcome variables. Beacuse I want them to have the same reference category, I want to let the residuals between these dummy variables be correlated. Which commando can I use best in Mplus?

Thanks in advance,

Freek.
 Linda K. Muthen posted on Tuesday, February 27, 2007 - 2:54 pm
You should use either the NOMINAL option if you have a six-category unordered categorical variable or the CATEGORICAL option if you have an ordered categorical variable.
 freek bucx posted on Thursday, March 01, 2007 - 7:25 am
Hi,

Thanks for your quick answer.

Ideally, I would like to stick to my - separate - dummy variables because, then, I can investigate directly for which categories effects are different (for one six-category variable, when you have found an effect you can only state that there is an effect on this variable, but then you still do not know for which categories effects are different).

If I decide to stick to my five separate dummy variables, I want them to have the same reference category. In that case, which commando can I use best in order to have the residuals between these dummy variables be correlated?
 Linda K. Muthen posted on Thursday, March 01, 2007 - 12:47 pm
Correlations/covariances are specified using the WITH option of the MODEL command.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: