Maximum Likelihood Estimator Robust f... PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Dylan Johnson posted on Thursday, July 06, 2017 - 2:10 pm
In my dataset of 6000, I have about 500 cases that are incomplete (at least on missing value on an indicator variable). I am conducting a LPA with three indicator variables:

1. Depression

2. Conduct Problems

3. ADHD

Due to the fact that these are problem behaviours in children, the averages are somewhat skewed (the average child scores low on all 3).

If I want to run the LPA so that it uses maximum likelihood estimates for the incomplete data and is robust for skewness, do I simply need to add the following:

ANALYSIS:
estimator=ml;



Many thanks!
 Bengt O. Muthen posted on Thursday, July 06, 2017 - 7:00 pm
Use MLR.
 Dylan Johnson posted on Thursday, July 06, 2017 - 7:26 pm
Thank you.

Does it make sense to include these incomplete cases as long as they are not missing data for all three indicators?
 Dylan Johnson posted on Thursday, July 06, 2017 - 7:40 pm
Also, is it okay to use the MLR function simultaneously with BLRT?
 Bengt O. Muthen posted on Friday, July 07, 2017 - 4:32 pm
Yes on both of your posts.
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: