Proceedings of the Annual Meeting of the Cognitive Science Society

Older adults (OA) need to make many important and difficultdecisions. Often, there are too many options available to ex-plore exhaustively, creating the ubiquitous tradeoff betweenexploration and exploitation. How do OA make these com-plex tradeoffs? We investigated age-related shifts in solvingexploration-exploitation tradeoffs depending on the complex-ity of the choice environment. Participants played four andeight option bandit problems with numbers of gambles and av-erage rewards available on the screen. OA reliably performedworse in a more complex choice environment and were alsomore deviant from an optimality model (Thompson sampling),which keeps track of uncertainty beyond just the mean or lastreward. OA seem to process important information in morecomplex choice environments sub-optimally, suggesting lim-ited representations of future rewards. This interpretation fitsto multiple contexts in the complex cognitive aging literature,in particular to the context of challenges in the maintenance ofgoal-directed learning.