Recently, a debate has developed in both the theoretical and applied statistical literature regarding the appropriateness of various techniques in analytical models with limited (particularly, 0-1) dependent variables. In financial research, limited dependent variable models usually arise in one of two ways:1. classification (or discrimination)--assigning observations to discrete, a priori determined groups, or2. regression--relating a qualitative dependent variable to one or more independent variables (which may or may not be qualitative).