Institute for Mathematical Behavioral Sciences

Loss aversion refers to the finding that many people would prefer a small amount of cash rather than accept a mixed gamble with equal or higher expected value. For example, many people would rather accept a sure gain of $1, rather than accept a 50-50 gamble to either win $100 or lose $80, even though the gamble has a higher expected value ($10).There are two major theories to account for this phenomenon. The first is Cumulative Prospect Theory (CPT), which holds that the loss of $80 has stronger negative utility than the gain of $100 has positive utility. A configural weight model known as the Transfer of Attention Exchange (TAX) model assumes that losing $80 has greater configural weight than winning $100 does in this gamble. In this model more attention or weight is placed on the worst negative consequence than the best positive consequence of a gamble. The property of Gain-Loss Separability, investigated by Wu and Markle (2004), can provide a test between these two ideas. A new experiment will be reported which compares these two models, based on their parameters fit to previous data. Gain-Loss separability is implied by CPT under any parameters, but the observed violations were predicted from the TAX model fit to previous data.