About trends estimation wtih mplus PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 robin hao posted on Monday, June 18, 2007 - 12:47 pm
we hae a project to analyze the trends of a index.
this index is produce with structrual equation modeling based on 24 variable.
when calculating the index, we first divide those 24 variables into 5 groups, then use factor analysis to calculate the factor scores of every group, next step, use 4 groups factor scores to regress against the 5th groups factor scores and use the pridictive value as the index value.

Now we would like to analyze this index's trends with several years data?

I think every the index values should have autocorrelation error among different year.

Is it possiable to process this project with mplus?

Thank you very much!

Hao yunze
 Linda K. Muthen posted on Monday, June 18, 2007 - 1:47 pm
Yes. See Example 6.17 in the Mplus User's Guide.
 robin hao posted on Monday, June 18, 2007 - 7:14 pm
Hi Linda,

Thank you for your kind reply!

Could I ask a extension question about this project?

Is it possible to find out relation of the index growth and the variables growth with Mplus?


Thank you very much!

Hao yunze
 Linda K. Muthen posted on Tuesday, June 19, 2007 - 4:03 pm
When I first read your initial question, it made me think that what you want is multiple indicator growth. See Example 6.14 and see if this would help you.
 robin hao posted on Wednesday, June 20, 2007 - 1:33 pm
Hi Linda,

Thank you for your kind help!

I am very new to MPLUS, and used SAS in past.

I wonder what is the difference between the MPLUS Example 6.17 and SAS Proc Autoreg?

They all seem to process the autocorrelation question.

What is the difference between MPLUS growth model and SAS time series procedure?

Have a great day!

Hao yunze
 Linda K. Muthen posted on Wednesday, June 20, 2007 - 2:56 pm
I think this example is SAS PROC MIXED with a first-order autoregressive structure on the residuals.

Mplus does not do time series analysis. This is usually for one observation and many repeated measures.
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: