The role of attentional processes in the emergence and prevention of learning traps

Learning traps are cycles of suboptimal decision-making where a false belief about the structure of the environment leads to avoidance of rewarding options. Two experiments (N= 324) examined the role of selective attention in the emergence and prevention of learning traps. Participants learned to approach or avoid members of two categories that were associated with either gains or losses. A rule involving two visual feature dimensions predicted category membership, while a third dimension was irrelevant. Category feedback was provided only when an item was approached. Selective attention during learning was assessed using an eye tracker. Experiment 1 found that many participants fell into the trap of using a single-dimension rule to guide approach/avoid decisions and consequently missed rewards. Eye tracking confirmed that these participants narrowed their visual attention to a single dimension. In Experiment 2, we manipulated visual feature salience across three groups and found evidence for a causal role of selective attention in determining the emergence and prevalence of learning traps.