On the point of offering a friend a glass of water, we do not ask, “How long has it been since you last drank any water?” or “If I offer you a glass of water, what are the chances you will accept it?” We ask, “Are you thirsty?” The answer tells us all we need to know. In an experimental analysis, however, we need a better account of the conditions that affect hydration and a better measure of the probability that a subject will drink. A report of how thirsty the subject feels will not suffice. (p. 11)
