The (Neural) Dynamics of Stochastic Choice
从心理学和神经科学的累积模型推导出随机效用模型，揭示决策时间如何影响随机效用分布，并可能导致偏好参数估计偏差，可通过加入反应时间等观测数据来缓解。
The standard framework for modeling stochastic choice, the random utility model, is agnostic about the temporal dynamics of the decision process. In contrast, a general class of bounded accumulation models from psychology and neuroscience explicitly relate decision times to stochastic choice behavior. This article demonstrates that a random utility model can be derived from the general class of bounded accumulation models, and characterizes how the resulting distribution of random utility depends on response time. This relationship can bias the estimation of structural preference parameters. The bias can be alleviated via the inclusion of standard observables directly in the econometric specification, or through incorporating novel observables such as response time or neurobiological data. Examples of estimating risk and brand preferences are pursued. This paper was accepted by Matthew Shum, marketing.