Binary outcome two groups PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 Geetanjali Kamath posted on Friday, January 06, 2017 - 9:54 am
Dear Dr. Muthen,

I modeled and did invariance testing for a CFA with 15 measured variables and 6 latent factors for two groups using the GROUPING option. All worked fine. Now I want to regress these 6 factors on a binary (observed) dependent variable. I used Estimator=MLR since data were non-normal (not missing though). This above SEM ran correctly when I did not specify GROUPING, and gave me odds ratios. But when I try to do SEM for two groups, it gives me this error:

ALGORITHM=INTEGRATION is not available for multiple group analysis. Try using the KNOWNCLASS option for TYPE=MIXTURE.

However, I get even more issues if I try to do the above. Is there a solution for SEM with two groups and a binary observed DV? Something obvious that I'm overlooking? Note that if I don't declare my DV as CATEGORICAL=, it gives me estimates with no problems. However, I'm not sure if this is the way I'm supposed to do it. Any help is much appreciated!

Thanks,
Geetu
 Bengt O. Muthen posted on Friday, January 06, 2017 - 3:23 pm
Knownclass is the way to go. Send the output for which you had "even more issues" to Support along with your license number.
 Geetanjali Kamath posted on Sunday, January 15, 2017 - 4:38 pm
Dear Prof Muthen,

I have tried the following code using KNOWNCLASS: (putting only a part of my code since all won't fit).
I'm now getting this error message:
Invalid value in KNOWNCLASS option: MALE

Please can you help me figure out why this is happening? My outcome is adhered and subgroups are defined by gender.

Many thanks!
Geetanjali

CATEGORICAL = adhered;
KNOWNCLASS = GENDER (1=Male 2=Female);
CLASSES = GENDER (2);

MODEL:
%OVERALL%
f1 BY c56 c57 c58;
f2 BY c44 c45 c47 c49;
f3 BY c48 c50;
f4 BY c51 c55;
f5 BY c43 c46 c52;
f6 BY c61;
adhered ON f1-f6;

ANALYSIS:
ESTIMATOR = MLR;
TYPE = MIXTURE;
 Bengt O. Muthen posted on Monday, January 16, 2017 - 9:59 am
Please send your full output to Support along with your license number.
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: