MG-CFA PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 HY, LEE posted on Friday, January 17, 2014 - 2:56 am
Hi, Linda

I have a question.
Can I run a multiple group CFA with Mplus when groups have different number of indicators?

Thank you
 Linda K. Muthen posted on Friday, January 17, 2014 - 8:54 am
See the FAQ on the website:

Different number of variables in different groups
 HY, LEE posted on Friday, January 17, 2014 - 5:13 pm
Thank you, Linda
 HY, LEE posted on Wednesday, January 22, 2014 - 12:20 am
Hi, Linda

I have another question about missing data generation.

I am trying to generate data set of MG-CFA(two group) for a Monte Carlo study.

I wonder how generate missing data only for Group2.

Thank you.
 Linda K. Muthen posted on Wednesday, January 22, 2014 - 10:06 am
MODEL MISSING can be used for only one group. You do this using the group label.
 HY, LEE posted on Wednesday, January 22, 2014 - 10:30 pm
I am including model syntax below.

MONTECARLO:
NAMES=Y1-Y5;
NOBSERVATIONS= 400 400;
NGROUPS=2;
NREPS=500;
REPSAVE=ALL;
SAVE=rep*.DAT;
RESULTS=results.sav;
missing=Y4-Y5;


MODEL POPULATION:
F1@1;
F1 BY Y1@1.0 Y2@0.9 Y3@0.8 Y4@0.7 Y5@.6;
Y1@.33 Y2@.54 Y3@0.16 Y4@0.21 Y5@.19;
[Y1@1.0 Y2@1.0 Y3@1.0 Y4@1.0 Y5@1.0];
[F1@0];

MODEL POPULATION-g2:
F1 BY Y4-Y5@0;
[Y4@1.0 Y5@1.0];

MODEL MISSING-G2:
[Y4@1.0 Y5@1.0];

But the generated data sets include missing data both group1 and group2.
Would you please give me an advise and information?
 Linda K. Muthen posted on Thursday, January 23, 2014 - 10:18 am
Please send the output and your license number to support@statmodel.com.
 Anonymous  posted on Saturday, October 17, 2015 - 10:57 am
Hello,

I am attempting to run a multigroup CFA and I received an error message stating my standard errors could not be estimated and my model may not be identified. I utilized the Tech1 output to identify the problematic parameter. The parameter was in the PSI matrix. Next, I utilized the tech 4 output to view the covariance and it did not appear to be problematic. After some tinkering, I found that fixing my latent variable variances @1 solved the issue. I was wondering if you might be able to discuss some reasons for this, or at least point to some literature to help understand why this worked?


Thank you for your time.
 Linda K. Muthen posted on Saturday, October 17, 2015 - 3:17 pm
If you did not free the first factor loadings for each factor, then fixing the factor variances to one is not proper. The Mplus default is to fix the first factor loading to one to set the metric of the factor. You do not want to also set the metric by fixing the factor variance to one. Please send the output before you did this and your license number to support@statmodel.com.
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: