AIC/BIC with EFA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Jeremy Miles posted on Friday, October 10, 2008 - 2:31 pm
Hi, Is there any reason that EFA doesn't give AIC/BIC? Would it be a bad idea to calculate those using the chi-square and df from the model?

Thanks,

jeremy
 Linda K. Muthen posted on Friday, October 10, 2008 - 4:32 pm
EFA using the maximum likelihood estimator does give AIC and BIC. These statistics are not relevant for weighted least squares.
 Haigen Huang posted on Wednesday, November 06, 2013 - 12:21 pm
It seems when I use the following code, it doesn't give me BIC or AIC. I have categorical variables. Do you know how to obtain BIC or AIC in this case? I will appreciate!

ANALYSIS:TYPE = TWOLEVEL EFA 1 3 UW 1 2 UB;
 Linda K. Muthen posted on Wednesday, November 06, 2013 - 1:31 pm
The default for categorical variables is the weighted least squares estimator for which BIC and AIC are not available. You would need maximum likelihood estimation to obtain these values. With categorical variables and maximum likelihood estimation, each factor requires one dimension of integration. This could become computationaly demanding.
 Haigen Huang posted on Thursday, November 07, 2013 - 11:28 am
Thanks so much for the prompt response! I specified with ESTIMATOR=ML; But the BIC and AIC still don't show up...Am I doing anything wrong?

The following are the code I used:

ANALYSIS: TYPE = TWOLEVEL EFA 1 3 UW 1 2 UB; ESTIMATOR=ML;
 Haigen Huang posted on Thursday, November 07, 2013 - 12:12 pm
By they way, the warning says:

*** WARNING in ANALYSIS command
Estimator ML is not available for TYPE=EFA TWOLEVEL with categorical variables.
Default estimator will be used.

I will appreciate if you give me some suggestions on how to obtain AIC and BIC in this case.

Thanks, Thanks, and Thanks!
 Linda K. Muthen posted on Thursday, November 07, 2013 - 1:13 pm
If ML is not available, you cannot obtain AIC and BIC.
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: