Akteure handeln rational. Sie maximieren ihren Nutzen. Sie versuchen jedenfalls, mit ihren Handlungen ihren Nutzen zu maximieren. Aber: Welche Rationalität ist die richtige, um das Handeln von Akteuren vorherzusagen? Das, was Max Weber die objektive Richtigkeitsrationalität genannt hat oder das, was Herbert Simon “bounded rationality” genannt hat?
Die Antwort auf diese Frage hängt mit dem Interesse dessen, der forscht zusammen. Wir präsentieren heute das Pizza-Experiment als Beispiel dafür, wie man daraus, dass tatsächliches Verhalten von der objektiven Richtigkeitsrationalität abweicht, Erkenntnisse über handlungsleitende Normen gewinnen kann.
“A local pizza parlor offered an all-you-can-eat-lunch for $3. You pay at the door and the waiter brings you as many slices of pizza as you like. One of my colleagues performed this experiment: He had an assistant serve as the waiter for one group of tables. The ‘waiter’ selected half of the tables at random and gave everyone at those tables a $3 refund before taking orders. The remaining half of his tables got no refund. He then kept careful count of the number of slices of pizza each diner ate. What difference, if any, do you predict in the amounts eaten by these two groups?
Dines in each group confront the question, ‘Should I eat another slice of pizza?’ Here the activity x consists of eating one more slice. For both groups, C(x) [Costs of x] is exactly zero: even members of the group that did not get any refund can get as many additional slices as they want at no extra charge. Because the refund group was chosen at random, there is no reason to suppose that its members like pizza any more or less than the others. For everyone, the decision rule says keep eating until there is no longer any extra pleasure in eating another slice. Thus, B(x) [Benefit of x] should be the same for each group, and people of both groups should keep eating until B(x) falls to zero.
By this reasoning, the two groups should eat the same amount of pizza, on the average. The $3 admission fee is a sunk cost and should have no influence on the amount of pizza one eats. In fact, however, the group that did not get the refund consumed substantially more pizza”. (Frank, Microeconomics and Behavior, pp.13-14)