|
|
Algorithms for handling missing data |
|
Message/Author |
|
Dex posted on Wednesday, June 07, 2017 - 9:54 pm
|
|
|
Dr. Muthens, Could you please help me clarify the missing data handling algorithms used in Mplus under the following three situations? 1. Missing data with categorical variable using limited information (wlsmv), I assume Mplus use pairwise deletion, right? 2. Missing data with categorical variable with “ESTIMATOR=ML”. I do not think this is the same as “FIML” in continuous case. Then what is the actual algorithm used? 3. How does Mplus do bootstrapping with missing data, i.e. using MISSING ARE ALL (999)+ BOOTSTRAP Thanks for your help! |
|
|
1. Right 2. It is the same. 3. The procedure doesn't change. |
|
Back to top |
|
|