MODINDCES in EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 fredrik saboonchi posted on Thursday, November 14, 2013 - 9:07 pm
Hello!
After conducting an EFA and asking for MODINDICES, Mplus produces a matrix of Theta for the indicators, which is somehow different from the MODINDICES in ESEM.

I'd like to incorporate the information in the EFA MOD into the model that approximates ok fit (CFI =0.887), to improve it, so I add a WITH in a Model statement
between the two indicators with largest MOD in theta , but nothing changes in the model fit, am I reading the MODINDICES wrongly?

this i the section that contains the largest MOD (other items not shown)


THETA
COPE_18 COPE_19 COPE_20
________ ________ ________ ________

COPE_28 47.499 2.640 0.263



and I add :


Analysis: type = efa 5 6;
Model: COPE_28 with COPE_18



But the output looks the same as before. Is it possible to "improve" an EFA, that is what I gathered from the 2009 paper, but maybe I have to rely on ESEM?

Best

Fredrik
 Linda K. Muthen posted on Friday, November 15, 2013 - 11:39 am
The MODEL command is not used with TYPE=EFA. You can do your EFA via ESEM and then add the residual covariance. See Example 5.24. If you remove the covariates and direct effects, this is an EFA.
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: