Indirect effects w/ cat. predictors a... PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Igor Holas posted on Thursday, February 18, 2010 - 4:04 pm
I have a path analysis model with the following features:

- single nominal (3 category) predictor (run as rotated contrast) (U1)
- several dichotomous covariates, and some continuous ones
- continuous mediators (Y1 and Y2) and outcomes (Y3 and Y4)
- I am using MLR because my residuals failed the assumption tests
- I am using CLUSTER option to correct my stanrard errors, because I have children nested in schools.

my model in short is:
U1 -> Y1, Y2 -> Y3, Y4

I ran the full model as follows

MODEL:
Y1 Y2 on U1#1 U1#2;
Y3 Y4 on U1#1 U1#2 Y1 Y2;

Did I estimate all paths? I usually run MODEL INDIRECT: but I receive an error ( MODEL INDIRECT is not available for analysis with ALGORITHM=INTEGRATION.)

I read through the manual and the discussion. And saw mentions of Thresholds and MODEL CONSTRAIN. However, I did not find enough information on either to understand how to use them.

Thank you for your help!

I
 Linda K. Muthen posted on Friday, February 19, 2010 - 6:49 am
I think you are putting u on the NOMINAL list. This is for dependent variables. You should instead create a set of dummy variables for u and use those as covariates.
 Igor Holas posted on Saturday, February 20, 2010 - 1:38 pm
Dr. Muthen,

that is what I have - I just failed to properly describe it. All my categorical predictors and covariates are indicator coded.

Thank you

I
 Linda K. Muthen posted on Saturday, February 20, 2010 - 2:27 pm
You would need to send the full output and your license to support@statmodel.com so I can see what you are doing.
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: