Proceedings of the Annual Meeting of the Cognitive Science Society

Rational agents should update their beliefs in the light of newevidence. Equally, changes in belief should depend only onthe quality of the evidence, and not on factors such as the or-der in which the evidence is acquired, or whether intermedi-ate judgements are requested during evidence acquisition. Incontrast we show that requests for intermediate judgments caninhibit belief updating for real decision makers, which repre-sents a new type of decision making fallacy. This behaviour isparadoxical from the point of view of classical Bayesian mod-els, but we show that it is consistent with an a priori, parameterfree prediction of a cognitive model based on quantum theory.