|M.Sc Student||Nizri Tal Ori|
|Subject||Search Behaviors: The Evolvement of Stopping Criteria and|
Confidence Judgments with Accumulation of
|Department||Department of Industrial Engineering and Management||Supervisors||ASSOCIATE PROF. Rakefet Ackerman|
|DR. Kinneret Tedorescu|
Search is an important part of our lives; we search for food to eat and what job to accept. How and when do we decide when to terminate our searches? This question was studied in sequential search tasks in behavioral economics literature. The robust finding in that field is that people terminate their search too early rather than at the optimal point. In the meta-cognition literature, confidence judgments are shown to guide the termination of mental searches. The current work is aimed to combine both behavioral economics and meta-cognition research fields to investigate the relationship between external search, in a repeated search task, and internal confidence judgments. We used a search task which requires people to choose from a sequence of values presented sequentially, under the constraint that it is not possible to return to an earlier alternative. After each choice, the participants assessed the value's relative place among the alternatives and rate their confidence in that assessment. By manipulating the environments' variance and the experience of repeated environment, we aimed to alter search behavior and by that, to shed light on its relationship with confidence judgments. The results of two experiments demonstrate that higher environment variance leads to changes in search, but not in confidence. Since confidence is equivalent in both low and high variance environments, this may imply that it serves as a stopping rule for search termination. We also found that the repeated experience of the same environment was associated with higher levels of confidence, better accuracy, and improved performance, but it did not affect search. Thus, people increased their target level of confidence with the accumulation of experience, became more accurate, and performed better. The similar amount of search over repetitions could be attributed to more explorative search in the first environments and more exploitative search in later environments. In addition, an overall positive relationship between search and performance was found, in line with previous under-search findings. Further theoretical and methodological implications are discussed.