No adaptive strategy selection without outcome feedback

Abstract

A common view in the judgment and decision making literature is that humans posses a repertoire of decision strategies, from which we choose adaptively when dealing with decision situations. Modeling this adaptive strategy selection process has proved to be difficult, but advances were recently made by modeling it as a reinforcement learning process. Dieckmann and Rieskamp (2007) investigated the influence of information redundancy in a multiple-cue learning setting and found surprising evidence for adaptive strategy selection in situations without outcome feedback, which is difficult to explain by a pure reinforcement learning model. We challenge these findings by pointing out problems in their experimental design. We replicate their experiment, add conditions with stricter experimental controls, and investigate possible underlying mechanisms. In conditions with stricter controls we find no evidence that participants manage to incorporate information redundancy into strategy selection and conclude there is no adaptive strategy selection without outcome feedback.


Back to Table of Contents