Why do people reject mixed gambles

Decision makers often reject mixed gambles offering equal probabilities of a larger gain and a smaller loss. This behavior contradicts the predictions of expected utility-based theories of risk taking, and is widely seen as evidence for loss aversion, a psychological mechanism according to which losses are given higher utility weights than gains. In this paper we consider an alternate mechanism capable of generating high rejection rates: A predecisional bias towards rejection without the calculation of utility. We use a drift diffusion model of decision making to simultaneously specify and test for the effects of these two psychological mechanisms in a gambling task. We document the existence of a robust predecisional bias for rejection, and find that this bias provides the largest quantitative contribution to individual and group-level model fits. This bias also predicts novel response time patterns for acceptance and rejection, which we observe in our data. Our results indicate that high rejection rates for mixed gambles can be a result of multiple different psychological mechanisms, and that a predecisional bias applied prior to the computation of utility (rather than loss aversion) is the primary determinant of this important behavioral tendency. These results have important implications for how we model behavior in risky choice tasks, and how we interpret its relationship with various psychological and neurobiological variables.