2020 Theses Bachelor's

Abstract Rule Learning in Gain and Loss Frames

Gandhi, Jay

The ability to abstract rules from prior experiences and apply them in new contexts is critical to adaptive behavior. Prior research has established that the way in which options are framed (i.e. in terms of gains or losses), can shape decisions, and possibly learning; however, the extent to which such framing influences the ability to learn and apply rules remains unknown. In this study, participants learned the correct response to each of several images that were linked by a hidden rule. Though not necessary, this rule could be exploited to maximize performance. As feedback, participants received positive and negative reinforcement (monetary gains and losses) in alternating sessions. We found a significant impact of gain session history on performance, and but a marginal impact of the current frame. To investigate whether these framing sensitivities were related to those classically shown in the decision-making literature, participants also completed a canonical financial decision-making task. We replicated classical findings – participants were more risk-averse in the gain frame – but the relationship with framing effects on inference was equivocal in the present sample. Our results suggest that rule-acquisition and application might be differentially sensitive to framing, such that gain framing appears to improve acquisition more than application. Further investigation is required to determine the extent to which framing sensitivities in the contexts of inference and risky-choice relate.