Proceedings of the Annual Meeting of the Cognitive Science Society

Loss aversion---the idea that losses loom larger than equivalent gains---is a cornerstone of behavioral economics. Yet estimates of the loss aversion parameter ($\lambda$) vary considerably across elicitation methods, suggesting this single parameter may conflate distinct psychological mechanisms. We test this hypothesis using a hybrid Decision Field Theory, which allows formal separation of valuation asymmetry ($\lambda$: losses weighted more heavily in utility) from attentional asymmetry ($\kappa$: losses sampled more frequently during deliberation). Fitting hierarchical Bayesian models to 10 datasets (N = 686; $\sim$ 140,000 trials), we find that attentional asymmetry alone provides superior predictive accuracy in the majority of cases. Simulation analyses further demonstrate that standard utility models absorb attentional variance into inflated $\lambda$ estimates. These findings suggest that the tendency to reject favorable mixed gambles may substantially reflect how losses are weighted during accumulation, rather than solely due to asymmetric valuation.