Modeling risk perception and risk taking in dynamic situations

本研究旨在探讨风险认知与风险承担行为的基本机制与过程。虽然在这方面已经有很多研究，但主要是静态任务。然而，这些任务并没有反映这样一个事实，即现实生活中的风险决策通常会导致改变进一步决策风险的条件。因此，一个有趣的问题是，决策者在多大程度上能够发现这些变化并加以考虑。为了回答这个问题和相关的问题，本项目集中在基本的动态方面的风险行为。最近，已经开发了几个非常适合实现这一目标的动态任务。一个是哥伦比亚卡片任务（CCT），在这个任务中，必须做出几个连续的风险决策，每一个决策都会改变后续决策的风险。CCT允许对准备-暴露-应对（PEC）周期进行详细调查，这是研究单元的主要课题。该项目的重点是顺序决策的动态性。相应的性能不仅将进行实验研究，而且还与静态和动态模型建模。更令人感兴趣的是决策中的动态。也就是说，它将审查在决策过程中如何以及何时考虑到不同类型的信息，并将其用于评价不同的备选方案。这些问题也将通过扩散模型的决策性能建模进行分析。总体而言，8个实验计划，其中一些是关于调查的基本属性的动态任务和相应的风险决策。在其他实验中，将测试动态任务中的表现与静态任务中的表现有何不同。将与其他子项目合作，测试所开发的概念和模型在临床等方面的适用性。

The goal of this project is to investigate basic mechanisms and processes of risk perception and risk-taking behavior. Although there is already much research in this area, mainly static tasks have been applied. These tasks, however, do not reflect the fact that risky real-life decisions usually lead to conditions that change the risk for further decisions. An interesting question therefore is, to what extent decision makers are able to detect these changes and to take them into account. To answer this and related questions, the present project focusses on basic dynamic aspects of risky behavior. Recently, several dynamic tasks have been developed that are well suited for achieving this goal. One is the Columbia Card Task (CCT), where several sequential risky decisions have to be made, each of which modifies the risk for the subsequent decision. The CCT allows a detailed investigation of the Preparedness-Exposure-Coping (PEC) cycle, which is the main topic of the research unit. The focus of this project is the dynamics of sequential decisions. The corresponding performance will not only be investigated experimentally, but also modeled with static as well as dynamic models. Of further interest is the dynamics within decisions. That is, it will be examined how and when different types of information are taken into account during the decision process and are used for the evaluation of the different options. These issues will also be analyzed by modeling decision performance with diffusion models. Overall, 8 experiments are planned, of which some are concerned with the investigation of basic properties of dynamic tasks and corresponding risky decisions. In other experiments it will be tested how performance in dynamic tasks differ to that in static tasks. The applicability of the developed concepts and models, e.g. in clinical contexts, will be tested in cooperation with other sub-projects.