Class membership in EFMA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
Message/Author
 Masako Okamoto posted on Saturday, January 03, 2015 - 3:41 pm
Dear Mplus support team,

I have questions about exploratory factor mixture analysis.

(1) Is there any way of saving class assignments? When I tried gSAVEDATAh option, I received warning messages as follows:
*** WARNING in SAVEDATA command
Class probabilities cannot be saved for TYPE=EFA when the range of factors
requested is greater than 1.
***************************************

(2) Again about obtaining information related to class membership. When I run exploratory factor mixture analysis on categorical data, I found gESTIMATED PROBABILITIESh and gESTIMATED STANDARDIZED THRESHOLDSh in the output file. However, when I run the same program on count data (only changed variable type), it seems I do not get these information. Do I need to specify something in the input file?

Thank you for your help,
 Linda K. Muthen posted on Saturday, January 03, 2015 - 4:23 pm
This cannot be done for EFA.
 M. O. posted on Saturday, January 03, 2015 - 10:53 pm
Dear Dr. Muthen,

Thank you for your reply. Maybe I did not explain my question well, so I just would like to confirm:

I have conducted EFMA by specifying 2 classes and 2 factors. In the output file, I found that about half of the respondents being assigned to class 1. Is there any way to know who is assigned to class 1? Input file that I used was this.
----
TITLE:EFMA
DATA: FILE IS 60itemN888EFMA.dat;
VARIABLE: NAMES ARE u1-u60;
usevariables u1-u60;
CATEGORICAL ARE u1-u60;
CLASSES = c(2);
ANALYSIS: TYPE = MIXTURE EFA 2 2;
INTEGRATION=MONTECARLO;
--------

Sorry for taking your time again.
 Linda K. Muthen posted on Sunday, January 04, 2015 - 10:32 am
Please send the full output and your license number to support@statmodel.com.
 M. O. posted on Monday, January 05, 2015 - 7:58 pm
Dear Dr. Muthen,

Thank you for your reply. In mixture analysis, I have been referring
to LO-MENDELL-RUBIN ADJUSTED LRT TEST to determine appropriate number
of classes. Given that TECHNICAL 11 is not available in EFMA, do you
suggest any alternative when deciding number of classes?@
Here is the analysis I am trying:
----
TITLE:EFMA
DATA: FILE IS test.dat;
VARIABLE: NAMES ARE u1-u10;
usevariables u1-u10;
CATEGORICAL ARE u1-u10;
CLASSES = c(2);
ANALYSIS:
TYPE = MIXTURE EFA 2 2;
OUTPUT: TECH1 TECH7 TECH11;
-----
Thank you again.
 Linda K. Muthen posted on Tuesday, January 06, 2015 - 8:25 am
I would use 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: