The next question is how to close the loop. If people say they are unhappy with part of a model specification, perhaps because they do not want the model to be so flattering or sycophantic, how does the lab return and show what changed because of that input?