Hello All,

I am trying to do a ROC Curve Comparison looking at three different models in the same dataset. Each model has different pieces of missing data (retrospective data), so I'd like to be able to make three ROC Curves on the same graph maximizing the n for EACH model. When I simply use roccomp it drops any case that has any missing variable in any of the three models.

Is there a way to do this in STATA?

Much appreciated!!!


- Sasha