Learning and decision making under uncertainty

How do (and should) people learn and make decisions when they lack knowledge about important elements of the environment? Various forms of this question have been investigated through people’s (hypothetical and real) behavior on monetary gambles. In the present thesis, we asked (broadly) what we can learn about how people learn and make decisions under uncertainty from their often seemingly perplexing behavior in such gambles. We investigated this and related issues in two experimental chapters and a theoretical chapter focused on an experience-based variation of monetary gambles. The main empirical findings can be categorized into three groups. First, we replicated and extended findings of people’s attempts to identify and exploit the temporal structure of such tasks. Most (but not all) participants engaged in such behavior even when outcomes were randomly generated. Second, we investigated the effects of various manipulations of the way in which outcome feedback was provided to the participants. Incentivizing people to attend to more aspects of the feedback had a small but reliably positive effect on people’s performance, while making feedback relevant and diagnostic with respect to participants’ hypotheses improved their performance considerably. Lastly, we provided an in-depth analysis of the often-ignored individual variability on such tasks, which indicated that people often differ in their task-representations quite substantially. Additionally, we outlined a general theoretical framework for how people learn and decide in such contexts based on the metaphor that considers people intuitive scientists (and contrasted it with the currently more popular metaphor that considers people intuitive statisticians). In this framework, we explained the three main results as at least partly the result of people’s background knowledge and the diagnosticity and relevance of feedback for their hypotheses about the experimental environment. We also highlighted the many benefits of this type of explanation for more general questions about how people learn and make decisions under uncertainty