Hana Taha¹, Thomas W. James¹; ¹Indiana University Bloomington

A relatively new development in decision science involves studying decisions from experience. In this context, risk is particularly interesting to study, because learning can lead to preferences among risk conditions, but risk can also shape learning. To investigate the interplay between risk and learning, we systemically manipulated risk, expected value, and skew of resource distributions in a foraging task called Learning about Expected Value & Risk (LEVR). The task used a four-factor design: EV (-4, +4), risk level (low, medium, high), skew type (negative, no skew, positive), and grouping (One-Skew Multi-Risk, One-Risk Multi-Skew). Across 240 trials, subjects chose between six options that combined these factors. For One-Risk, options for each subject group varied skew type, but were all the same risk level. For One-Skew, options for each group varied risk level, but were all the same skew type. EV4 choices were picked more than EV-4 choices across most subject groups, indicating significant learning. Learning was most consistent for One-Skew, suggesting that choosing among options with different risk levels aided learning. In contrast, for One-Risk, learning was much greater for the low risk level group than the medium or high-risk level groups. This was supported by a significant 3-way interaction between risk, grouping, and learning. Grouping also influenced preferences among risk levels, with high-risk level choices picked most of all, specifically in the One-Skew condition. Finally, there were no significant effects of skew in the results. Overall, the grouping together of multiple risk levels emerged as the most influential factor in shaping choice behavior. We suspect that grouping of risk provided environmental scaffolding for learning. In the One-Skew Multi-Risk group, learning about low-risk level options may have aided learning of high-risk level options, which wasn’t evident in the One-Risk group.