Question: Consider a model with a zero-one dependent variable. You have a multiple regression program and a program for the logit and probit models. You have computed the coefficients of the linear probability model and the logit and probit models:
(a) How will you transform the coefficients of the three models so that they are comparable?
(b) How will you compute the R2s for the three models?
(c) By what criteria will you choose the best model?