Andy Cohen posted on Monday, July 09, 2007 - 12:56 pm
I just recently purchased version 4.21. I don't see mention of negative binomials in the user guide but wanted to check here to see whether Mplus has the functionality to model a negative binomial distribution for my ultimate outcome variable. What can you tell me?
I am conducting a simple negative binomial regression that only has missing data on the predictors, which are both categorical and continuous. Is Mplus able to use all available data to estimate parameters from this model?
For any model, missing data theory does not apply to the observed exogenous variables. The model is estimated conditioned on these variables and any observation with missing on one or more of these variables is eliminated from the analysis. You can bring these variables into the model by mentioning their variances in the MODEL command. When you do this, they are treated as endogenous variables and distributional assumptions are made about them.