LCA w/ covariates PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Corey Savage posted on Thursday, August 06, 2015 - 2:27 pm
1)The items I am using for model identification and selection in my LCA have no missing values. Unfortunately, the variables I want to use for covariates have about 20% to 30% missing. What would you recommend I do in this case? Obviously when then attempting to add covariates, the gamma and rho estimates change from the initial model identification. I have a nice sample of about 1100 future teachers. Iím afraid if I go back and drop observations with missing values across the covariates, Iíll lose a huge portion of my data.

2) If I am able to get that resolved, is there a way I could ask ďConditional on [some covariates], does class membership predict [some outcome]?

I believe the outcomes I am thinking about also have missing issues at about the same percentage as above
 Bengt O. Muthen posted on Friday, August 07, 2015 - 5:42 pm
Try Auxiliary R3STEP. It doesn't resolve the missing on covariates issue but at least the covariates - and the lowered sample size their missingness causes - don't affect the class formation.
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: