Latent Variable Mixture Ignoring Missing PreviousNext
Mplus Discussion > Missing Data Modeling >
 Evgenia  posted on Wednesday, June 13, 2012 - 5:38 am
I want to do Latent Variable Mixture Modeling
My data have missing (MNAR) BUT I want to do analysis ignoring missing values to see differences at estimates.
What changes I have to make at the following input?

DATA: FILE = DatasetMisSc1.dat;

VARIABLE: NAMES = u1-u12;!Var u7-u12 are indicators of missingness
MISSING ARE all (9); !Missing
CLASSES = c (2); 2 classes
STARTS =30 10;
f by u1-u12@0;!

f by u7-u12*1;

f by u7-u12*1;

Thanks in advance
 Linda K. Muthen posted on Wednesday, June 13, 2012 - 5:41 am
Add LISTWISE=ON; to the DATA command.
 Evgenia  posted on Thursday, June 14, 2012 - 2:46 am
Folowing your suggestion
I add LISTWISE=ON; to the DATA command.
But then
I take error message *** ERROR
Categorical variable U7 contains less than 2 categories

which is true since U7-U12 are missing indicators and deleting listewise missing at vars U1-U6 make U7-U12 being constant equal to 1.

Thanks again
 Linda K. Muthen posted on Thursday, June 14, 2012 - 11:23 am
You should not use listwise deletion with missing value indicators.
 Evgenia  posted on Wednesday, June 20, 2012 - 10:30 pm
I want one more clarrification.
Having both data and missing value indicators if I use only MISSING ARE all (9); and no listwise deletion (previous post and reply), what is the analysis MPLUS do? It ignores missing in each case? Could you explain to me more?

Thanks alot

 Linda K. Muthen posted on Thursday, June 21, 2012 - 3:06 pm
Mplus uses all available data to estimate the model according to Little and Rubin. See the user's guide for the reference.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message