Abstract
Humans are traditionally depicted as suboptimal decision-makers, since they often fail to maximize expected utility. However, recent studies claim people choose optimally in tasks called motor lotteries, where participants reach to different targets within a maximum time window in order to get a reward. A difference is that, while in classical tasks probability information is explicitly given, in motor tasks it is implicit in each participant’s motor variability. Once this variability is known, a target can be designed to match a specific probability of being hit. This manipulation is normally implemented through size, but little has been done to explore other ways to represent probability in a motor task: for instance, through distance. Our experiment studied differences in expected utility maximization between these two ways of representing probabilities. In each of our two different conditions, trials consisted in the presentation of two targets, one with a lower probability to be hit but higher gain (risky) and another with higher probability and lower gain (safe). Participants decided to reach for one or another. In the size condition, both targets appeared at the same distance, but the risky was smaller. In the distance condition, both targets had equal size, but the risky was further away. Probabilities were manipulated to sample various expected gain differences between both targets. Results showed clear differences. Risk aversion was more present in the distance condition: participants tended to reach for the safe target even if the optimal choice was the risky target. In the size condition, participants were more sensitive to expected value differences: the more this difference favored the risky target, the more it was chosen. These differences may be interpreted as participants considering additional cost functions (e.g. biomechanical) in the distance condition not captured by the mere probability of hitting.
Meeting abstract presented at VSS 2015