Message/Author |
|
|
I am running a multilevel path analysis model with MLR as the estimator. I read in some other threads that the default setting of handling missing values in Mplus are FIML. I Looked it up the <mplus>, but didn't find relevant information. So my questions are: (1) Is FIML used in MLR? (2)In the outcome, I got the following error message: *** WARNING Data set contains cases with missing on x-variables. These cases were not included in the analysis. Number of cases with missing on x-variables: 55 Does it mean that the 55 cases with missing on x-variables were list wise deleted, while the rest (with missing values in other variables) were calculated using FIML? |
|
|
(1) Yes (2) Yes. |
|
|
Thanks, Dr. Muthen! However, after I use [x] to tell Mplus to model distributional parameters, the error message came out: " THE MODEL ESTIMATION DID NOT TERMINATE NORMALLY DUE TO A NON-ZERO DERIVATIVE OF THE OBSERVED-DATA LOGLIKELIHOOD. THE MCONVERGENCE CRITERION OF THE EM ALGORITHM IS NOT FULFILLED. CHECK YOUR STARTING VALUES OR INCREASE THE NUMBER OF MITERATIONS. ESTIMATES CANNOT BE TRUSTED. THE LOGLIKELIHOOD DERIVATIVE FOR THE FOLLOWING PARAMETER IS 0.78852880D+00: Parameter 29, %BETWEEN%: ICHQ WITH S1" What's wrong with my model? It has no problem to converge with out modeling distributional parameters. |
|
|
A following up question: instead of modeling distributional parameters, can I first use multiple imputation to impute missing variables in predictors, and then use FIML? Thanks! |
|
|
You can, but there are fewer bells and whistles for analyses down the line working with multiple imputation data. Such as no chi-square difference testing. See also our Multiple Imputation teaching in the Short Course videos and handouts for Topic 9 (2 versions). |
|
|
For your error message question we need to see the full output - send to Support along with your license number. |
|
Back to top |