|
|
 |
| Cox proportional hazards model - FIML |
 
|
| Message/Author |
|
|
| Claire Gorey posted on Wednesday, January 08, 2020 - 10:01 am
|
|
|
Hi!
I am running a continuous survival model in Mplus. I am wondering what command, if any, allows FIML to be engaged for this analysis. When using Estimator = ML, it still retains the same number of observations as listwise deletion. Thanks, in advance! |
|
|
|
|
This is what Mplus does when the missing values are for the covariates. You have two options.
1. Introduce a model for the covariates using something like this
x1-x5 with x1-x5;
That can be increasingly complicated if some of the covariates are not normally distributed. It could also be inaccurate if the covariates model is misspecified.
2. Use multiple imputation like in User's guide example 11.5 followed by the analysis model like in User's guide example 13.13 |
|
|
| Back to top |
|
|
