|
|
 |
| Model comparison (CFA vs. full SEM) u... |
 
|
| Message/Author |
|
|
|
|
I'm interested in comparing my measurement model to my structural model but because the estimator is WLMSV (latent variable indicators are categorical) I don't get given the AIC/BIC/SSA-BIC. I've tried requesting tech10 so that I can compare loglikelihood values, but I'm getting the message that tech10 is only possible when using type=mixture, and I'm not doing an LCA.
Has anyone got any suggestions about how I might compare the two models to see whether the structural model is adding anything to the measuremement model? |
|
|
| Sophie Dan posted on Wednesday, April 26, 2017 - 12:52 am
|
|
|
Hello, Dr. Muthen!
I read some paper of factor analysis, but I'm not sure how to choose the way to do that especially when ESEM is a good way.
For example, I split a data into two parts, so I have different ways to confirm a factor structure. I can 1. EFA - CFA(SEM), 2.ESEM - ESEM, 3. ESEM - CFA(SEM), 4. CFA(SEM) - CFA(SEM), Am I right?
But on thing is not clear to me, according to the model set, ESEM and EFA with mplus seem to have the same result in terms of factor loading and factor correlation, when there is no covariace included, so the way 2.ESEM - ESEM is in fact EFA-EFA, even though there is a two level factor structure. Do I understand right?
Thank you very much!! |
|
|
|
|
| For this type of general analysis question, please use SEMNET. Mplus Discussion is for specific questions about how to use Mplus. Mplus Support is for licensed users who have specific Mplus questions that require us to look at their files. Also, because today you have 6 posts with factor analysis questions I need to say that asking a series of questions on Mplus Discussion is not the way to learn a topic; we can't take the time to teach that way. Instead, take courses, study our videos, read papers, and ask on general discussion lists such as SEMNET. |
|
|
| Sophie Dan posted on Wednesday, April 26, 2017 - 2:23 pm
|
|
|
Dear Dr. Muthen,
OK, thank you!!!! |
|
|
| Back to top |
|
|
