Factor mixture modeling PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
Message/Author
 xiao guo posted on Wednesday, September 12, 2018 - 10:10 am
Hi!
I want to use the factor mixture modeling. but the item of one factor are CATEGORICAL,another are continuous.Can I use the factor mixture modeling analyse my data? thank you!
######## the code :
VARIABLE:
NAMES ARE u1-u30 t1-t30;
CLASS=C(2);
ANALYSIS:
TYPE=MIXTURE;
ESTIMATOR=MLR;
STARTS = 100 20;
PROCESSOR=2;
MODEL:
%OVERALL%
ab BY u1-u30*;
sp BY t1-t30@-1;
ab@1;
sp@1;
ab with sp;
u1-u30 PWITH t1-t30;
%C#1% !
[ab*];
[sp*];
 xiao guo posted on Wednesday, September 12, 2018 - 10:30 am
Add a question on the content above:
how can I set the Correlated residuals Between continuous and CATEGORICAL variables?
eg:u1-u30 PWITH t1-t30;
thank you for your help!
 Bengt O. Muthen posted on Wednesday, September 12, 2018 - 4:50 pm
With ML, it is very hard to do these correlations - you have to create a factor for each of them and you have to use numerical integration which can't be done with so many factors.

Bayes is a possibility and there you simply use PWITH.
 xiao guo posted on Wednesday, September 12, 2018 - 6:39 pm
Hi!
I want to use the factor mixture modeling. but the item of one factor are CATEGORICAL,another are continuous.Can I use the factor mixture modeling analyse my data? thank you!
 Bengt O. Muthen posted on Friday, September 14, 2018 - 1:53 pm
Yes, having different measurement scale types presents no extra consideration.
 xiao guo posted on Saturday, September 15, 2018 - 9:20 am
thank you for your help!
when I use bayes to estimate the model, I do not know how to save the latent variable and the class latent variable from your mplus guide?
thank you!
 Bengt O. Muthen posted on Sunday, September 16, 2018 - 11:59 am
With Bayes you get a distribution for each subject (so like imputation), that is, the nominal variable gets proportions for each category - plus you get the most likely class. See the FAQ on our website:

Cprobabilities with Bayes
 xiao guo posted on Sunday, September 16, 2018 - 11:56 pm
thank you!
when I set the Correlated residuals Between continuous and CATEGORICAL variables,I found that" THE MODEL CONTAINS A NON-ZERO CORRELATION BETWEEN DEPENDENT VARIABLES.
SUCH CORRELATIONS ARE IGNORED IN THE COMPUTATION OF THE FACTOR SCORES.". If I want such correlations affect the factor scores ,how can I do it?
Thank you for your kindness!
 Bengt O. Muthen posted on Monday, September 17, 2018 - 2:05 pm
This should not be a problem using Bayes (unless your Categorical variable is declared as Nominal which Bayes doesn't cover).
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: