A violation of the monotonicity axiom: experimental evidence on the conjunction fallacy

The conjunction fallacy is an anomaly in human reasoning for which the conjunction of two events is rated more likely to occur than one of the events alone. In the context of decision under uncertainty, this violates the monotonicity axiom of probability, and consequentially also Bayes’ Rule and the monotonicity axiom of preferences. Our experiments show how dynamic feedback and monetary incentives affect the fallacy rate, and how the complexity (and possibly the presentation) of the decision problem and an averaging heuristic might determine outcomes and reasoning.