A Meta-analysis of Developmental Change in Explore-Exploit Behavior

• In Event: Poster Session 07
  In Poster Session: PS 07 Topic - Cognitive Processes

Abstract

Developmental changes in explore-exploit decision-making processes, a key mechanism through which individuals learn novel information, have garnered a great deal of interest in recent years. Some studies provide evidence consistent with increased exploration in young children relative to adults (Gopnik, 2020), while others find evidence for lower levels of exploration in youth (Somerville et al., 2017). Results may vary across studies due to inconsistencies in which explore-exploit task is chosen and which age groups are compared. The aim of this meta-analysis is to examine age-related differences in explore-exploit behavior across different types of tasks, including bandit, foraging, and sequential search tasks. Specific age group (early childhood, middle childhood, adolescence), and task type will be tested as potential moderators of the overall effect of age on explore-exploit decision making. Three primary objectives of this project are to:
1. Determine the extent to which children show differences in explore-exploit behavior relative to adults.
2. Determine whether specific age groups show heterogenous differences in explore-exploit behavior relative to adults (e.g., do young children explore more, while adolescents exploit more?)
3. Determine the extent to which age group differences in explore-exploit behavior depend on the type of task used.
Published data and unpublished pre-prints and conference proceedings will be eligible for inclusion in the meta-analysis if they include at least one age comparison between a child/adolescent group with a mean age of 16 years or younger, and an adult group with a mean age of 20 years or older; focus on neuro-typical development; and include a behavioral measure of explore-exploit behavior based on choices participants made during a task (either gross measures or patterns computed through computational modeling).
Studies eligible for inclusion are being identified through database search, conference proceedings, and solicitation of unpublished data through society list-servs. Studies for inclusion will be collected by Dec 1, and data will be analyzed before the Biennial Meeting. Results are expected to clarify the role of context x age group interactions in explore-exploit decision-making, and more broadly, will have implications for how learning strategies change across age. This project is pre-registered at osf.io/2mn9d.

Authors

• Madeline B Harms, University of Minnesota - Duluth
  Presenting Author

• Cooper Tunac, Macalester College
  Non-Presenting Author