Multilevel SEM (diary study) PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Aislin Graham posted on Tuesday, February 01, 2011 - 5:08 pm
Hi there,

I've had success running a ML model via path analysis similar to example 9.2 in the mplus manual.My syntax is at the end of this message.

What I was hoping to do, was run the same model with latent variables for each construct. I tried adding the latent variable syntax under both the within and between conditions (for the respective latent variables) but it didn't work. Example 9.10 uses SEM in the context of a ML model. However, in this example the predictor (fw) is the latent variable.My DVs need to be a latent variables. Since this is a diary study, the within predictor variable is "day".


Perhaps the problem comes from the cross over from the within to the between analyses when the variables are treated as intercepts in the between model?

Any help would be greatly appreciated. Thank you.


USEVARIABLES ARE

ID DAY MDI PROM POMSD PM_EDDS N_MDI SPP;

MISSING = ALL(999);

CLUSTER = ID;

WITHIN = DAY;

BETWEEN = SPP;

analysis:

type = twolevel random;

estimator=ml;

model:

%WITHIN%

SLOPE | MDI PROM POMSD PM_EDDS N_MDI ON DAY;

PROM ON MDI;

POMSD ON PROM;

PM_EDDS ON POMSD;

N_MDI ON PM_EDDS;

%BETWEEN%

SLOPE MDI PROM POMSD PM_EDDS N_MDI ON SPP;

MDI PROM POMSD PM_EDDS N_MDI WITH SLOPE;

output: standardized;
 Linda K. Muthen posted on Tuesday, February 01, 2011 - 6:55 pm
Please send the full output and your license number to support@statmodel.com so I can see what problem you are running into.
 Aislin Graham posted on Saturday, February 05, 2011 - 9:37 pm
Hi again,

I have some variables that are measured on the within level. I have other variables measured on the between level. I would like to model, using SEM, various cyclical relationship between the within level variables, while also predicting each within level variable by the between level variables.

My question is, where should I be specifying all the variables?
- Between level variables on the BETWEEN
- WITHIN level variables ?
- If I model the WITHIN level variables on the WITHIN level then I can't have them predicted by the BETWEEN level variables right?

Another aspect that might complicate my analyses is that one within variable is "day" while the rest are measured constructs (via questionnaires).

I was able to run the model I want with path analysis with the following syntax, but when I tried to add in latent variable specification I have encountered a lot of problems.

%WITHIN%
SLOPE | MDI PROM POMSD PM_EDDS N_MDI ON DAY;
PROM ON MDI;
POMSD ON PROM;
PM_EDDS ON POMSD;
N_MDI ON PM_EDDS;

%BETWEEN%
SLOPE MDI PROM POMSD PM_EDDS N_MDI ON SPP;
MDI PROM POMSD PM_EDDS N_MDI WITH SLOPE;

I would appreciate any guidance on how to proceed. Thanks!
 Linda K. Muthen posted on Sunday, February 06, 2011 - 4:48 pm
I think you will find the answers to your questions in Examples 9.1 and 9.2. If you have problems, please send them and your license number to support@statmodel.com.
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: