I have a question regarding the behavior of the model-based H0 imputation in Mplus. I noticed that cases are being discarded when only values for covariates in the model have been observed (only dependent variables have missing values). Furthermore, these vases are not featured in the imputed data sets, that is, they are not being imputed. I have observed this for simple regression models and for two-level mixed-effects models.
The corresponding warning message is as follows: "Data set contains cases with missing on all variables except x-variables. These cases were not included in the analysis."
For estimation puposes, I understand that cases for which only the covariate is observed do not contribute to the likelihood and can, thus, be ignored. However, for multiple imputation, I would still like to obtain imputations for the missing dependent variable.
Is there an option that would allow me to keep such cases and obtain imputations for them using H0 imputation?
The only method I am aware of is to feature the covariates in the MODEL specification (e.g., a line that say "x;"). However, this appears to impose distributional assumptions on the covariates, which is not necessary in my opinion.