Saving factor scores in ESEM PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Gareth posted on Wednesday, August 20, 2008 - 2:02 am
Is it possible to save factor scores in exploratory SEM, or do I have to respecify the model as a CFA?
 Linda K. Muthen posted on Wednesday, August 20, 2008 - 9:34 am
You can save factors scores with exploratory factor analysis in SEM. If you have categorical indicators, there is a problem so you should send your input, data, and license number and I can run it for you in our developmental version. It is fine for continuous outcomes.
 Susan Skidmore posted on Wednesday, February 25, 2009 - 6:54 am
Is it possible to use the ESEM syntax from the ESEM paper or is this not available yet on the current Mplus version?
Also is it possible to use type=complex in ESEM?
Thank you for your help.
 Linda K. Muthen posted on Wednesday, February 25, 2009 - 9:15 am
Yes, ESEM became available in Version 5.1. See the Version 5.1 Language and Examples Addendums on the website with the Mplus Users' Guide.

Yes, TYPE=COMPLEX can be used with ESEM.
 statstudent posted on Wednesday, February 25, 2009 - 9:26 am
Thank you very much!
 Xu, Man posted on Monday, March 19, 2012 - 4:19 am
Dear Dr. Muthen,

I just saved factor scores in ESEM using SAVE=FSCORES, and I just wanted to check about these scores.

I wanted to get a summary of some items. The CFA of them have terrible fit so I thought EFA scores would be better. I decided to work with factor scores as the ESEM model would otherwise become too large to handle easily. Since I could not expor scores from EFA, I tried the ESEM set up and it worked!

Now I would like to check if these ESEM factor scores are the same to the factor scores that would have obtained from EFA scores with the same number of factors (although I understand Mplus does not have this option for EFA).

I have MPlus version 6.1. My items are categorical, the latent variable is continuous.

Thanks a lot!

Best,

Kate
 Linda K. Muthen posted on Monday, March 19, 2012 - 6:07 am
ESEM without covariates or indirect effects is EFA for the same number of factors. The factor scores from both approaches would be identical.
 Xu, Man posted on Monday, March 19, 2012 - 9:16 am
Thanks! Yes, I checked that ESEM and EFA with same number of factors gave exactly the same results!
PS: is it possible to fit SEM with ESEM as when CFA is specified, e.g. using ESEM factors as predictors of other CFA or growth curve factors?
 Bengt O. Muthen posted on Monday, March 19, 2012 - 9:33 am
Yes.
 Andreas Hirschi posted on Friday, March 29, 2013 - 7:35 am
Linda, above you said that EFA factor scores and ESEM factor scores are identical. Do you happen to have a reference for this? Thank you for your response in advance.
 Linda K. Muthen posted on Friday, March 29, 2013 - 10:09 am
This follows from the factor the ESEM without indirect effects and covariates is EFA. See Papers on the website. There is a section on ESEM.
 Zachary Isoma posted on Friday, April 12, 2013 - 7:15 am
Dr. Muthen-

I am attempting to run an ESEM with categorical variables to save factor scores, then explore the interrelationship of these factor scores with factors from two other measures. Is it possible to run an ESEM with categorical variables? If so, are there any helpful resources for how to run this analysis? Thank you!

Zack
 Linda K. Muthen posted on Friday, April 12, 2013 - 8:58 am
The ESEM examples start with Example 5.24. You would add the CATEGORICAL option to these examples to specify which factor indicators are categorical.
 Mariah Mantsun Cheng posted on Monday, June 23, 2014 - 9:04 am
Hi Linda/Bengt,
I have been using ESEM to output my EFA factor scores with no problem but until lately that I ran into this problem:

While the EFA model results seem reasonable, but when I used ESEM to output factor scores, I got the warning that "THE RESIDUAL COVARIANCE MATRIX (THETA) IS NOT POSITIVE DEFINITE" and therefore no factor scores can be saved.

My questions are:
1. Why did I not get the same warning with my EFA run? The ESEM run did give me same model estimates as the EFA run.
2. What should I do in order to resolve this problem and to get the factor scores? Is this problem related to the missing data issue?
Thanks for your advice in advance!
Best,
Mariah
 Bengt O. Muthen posted on Monday, June 23, 2014 - 10:11 am
Please send the two outputs to Support along with your license number.
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: