ESEM with correlation matrix? PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Hyun Joo, Jung posted on Wednesday, November 13, 2013 - 11:59 am
I'm wondering if It's really impossible to use correlation matrix type as input data when I want to use ESEM.

Please let me know how to use ESEM with summary data such as correlation matrix.

Here is my syntax:
DATA: FILE IS "L.txt";
TYPE IS CORRELATION;
NOBSERVATIONS ARE 300;

VARIABLE: NAMES ARE V1-V10;
USEVARIABLES ARE V1-V10;

ANALYSIS: ROWSTANDARDIZATION IS KAISER;
ESTIMATOR IS MLR;

MODEL: F1-F2 BY V1-V10(*1);

ERROR message:
EFA factors are not allowed with summary data.
EFA factors are declared with (*label).
 Linda K. Muthen posted on Wednesday, November 13, 2013 - 1:13 pm
We require individual-level data for ESEM.
 Hyun Joo, Jung posted on Wednesday, November 13, 2013 - 2:01 pm
Linda, Thanks a alot!!!
 Rimantas Vosylis posted on Wednesday, February 17, 2016 - 5:11 am
Dear Muthens,

I have used ESEM approach to EFA, by testing loading invariance in two groups. My results provide support for metric invariance, but loadings are equal only for unstandardized solution. I know that standardized coefficients are different in two groups because raw coefficients are standardized using group-specific model estimated standard deviations.

However, for reviewers these results seem to be somewhat confusing and they asked for some additional explanation.

My question: is there a way to run MG-EFA analysis that would also give me equal standardized coefficients?
 Linda K. Muthen posted on Wednesday, February 17, 2016 - 10:08 am
No. Standardization is done using the standard errors for each group. The only way they would be equal is if each group has the same standard errors.
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: