Sampling bias and logistic models - Discussion