Hi folks! I need to build a binary dependent variable model, but my independent variables have some serious outliers (they are not data errors.) I was thinking of using the robustbase package because I noticed one of the functions accepts a binary depvar. Can I evaluate this model like any other? Could anyone recommend a primer on robust methods? I really, really, hate to adjust my raw data for outliers first and then estimate my model! Thanks. Best, john [[alternative HTML version deleted]]