Three-level factor analysis and SEM PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
 lamjas posted on Tuesday, July 10, 2007 - 6:38 pm
I currently have Mplus version 3.12 Base and Combination Add-on. My unbalanced data from schools are quite complex and I am thinking to use three-level factor analysis and SEM (students are nested within classes between schools). I am wondering whether I can run such kinds of analyses using my current copy of Mplus.
 Linda K. Muthen posted on Wednesday, July 11, 2007 - 6:22 am
Neither Version 3.12 not the current version can handle three levels unless one of the is repeated measures over time. However, in the current version, you can use TYPE=COMPLEX TWOLEVEL; where the two levels would be student and classroom and COMPLEX would refer to variation across schools. This reflects the experience that most of the variation is on the student level, followed by the variation on the classroom level, with relatively little variation across schools. The variation on the first two level is modeled whereas the variation on the school level is taken into account by correcting the standard errors.
 lamjas posted on Wednesday, July 11, 2007 - 7:26 pm
Hi Linda,

According to your explanation, once I upgrade my copy to version 4.2 I can analyze the 3-level school data (CFA and SEM) appropriately even if the data are unbalanced, right? Also, do you have any examples for such kind of data? Thanks again.
 Linda K. Muthen posted on Thursday, July 12, 2007 - 6:05 am
You can model two levels and adjust standard errors for the third level. If you go to the user's guide on the website, you will find examples in the Version 4.2 addendum.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message