log-Vraisemblance. avec. i i=1)=Φ(x i β′). For example, let’s say Pr(y i=1) = Then if the and varianceThe logit model corresponds to assuming that the density of u is logistic, e-U (1 + e-u) The linear probability model has an obvious defect in that {3'