Rewards and punishments in iterated decision making:An explanation for thefrequency of thecontingent event effect

Iterated decision making can be studied in laboratory using situations, like the Iowa Gambling Task (IGT), in which participants face repeatedly the same decision problem getting feedback after each choice. In the paper we focus on a recurring finding in experiments carried out with the IGT, the frequency of the contingent event effect\u2014i.e., the fact that people consistently prefer options associated with rare losses, independently of their attractiveness, expected value and loss magnitude\u2014 that has not yet received a satisfactory explanation. An experiment reveals that the effect relies on simply experiencing rewards and punishments, not being influenced by the net outcome (loss or win) to which they are associated, and a computational model, implemented in the ACT-R cognitive architecture, corroborates the idea that punishments and losses on one hand, and rewards and wins on the other, play the same functional role in determining the participants\u2019 behavior in IGT