Accuracy of confidence: Relationship with decision-making and decision outcomes in a complex social-ecological system

当在有风险和不确定的环境中做决定时，人们对自己对情况的理解的信心程度往往比理解本身的程度更能影响决策。这种效应存在于医学、金融和政治等各个领域，也存在于群体决策中，人们也会利用他人的信心来指导自己的决策。例如，对自己诊断的准确性过于自信的医生不太可能寻求额外的诊断信息，高估自己专业水平的官员更有可能支持冒险政策，人们更愿意听从更自信的建言者的建议。尽管这是风险决策、不确定环境和自然涉及群体决策的主要例子，但关于人类如何集体管理公共池资源使用的研究在很大程度上忽视了信心对理解的潜在影响。这里将解决三个基本问题：当决定使用公共资源时，即使在控制理解的情况下，自己对理解资源动态（资源随时间的发展）的信心是否能预测决策？其他团队成员的自信是否能预测自己的决定，即使他们的自信是毫无根据的？潜在的自信驱动的决策如何与团队在可持续管理资源方面的成功相关？将采取多实验、多方法的方法，通过以下方式提供可靠的答案：(i)分析现有的实验数据，在这些实验数据中，人们在不同的外部风险水平下集体管理共同资源；（ii）进行一项专门设计的后续实验，以调查作为群体互动功能的理解和信心的最新测量方法的变化；（iii）计算建模方法，其中可能的基于信心的决策策略被形式化并与行为数据进行比较。总而言之，该项目将有助于我们理解是什么驱动了个人和群体对公共资源使用的决策，并将提供对解释这些风险决策的理解与信心的相对权重的可靠估计。

When making decisions in risky and uncertain environments, the level of confidence humans have in their understanding of the situation influences decision-making, often more so than the level of understanding itself. This effect exists in areas as diverse as medicine, finance, and politics, as well as in group decision-making, where people also use other’s confidence to guide own decisions. For example, physicians who are overconfident about the accuracy of their diagnosis are less likely to seek out additional diagnostic information, bureaucrats who overestimate their level of expertise are more likely to endorse risk-taking policies, and people follow more the advices of more confident advice-givers. Despite being a prime example of risky decisions, uncertain environments, and naturally involving group decision-making, research on how humans collectively manage the use of common-pool resources, however, has largely neglected the potential influence of confidence in understanding. Here three fundamental questions will be addressed: When deciding about the use of a common resource, does own confidence in understanding the resource dynamics (development of the resource over time) predict decisions, even when controlling for understanding? Does other group members’ confidence predict own decisions, potentially even when their confidence is unwarranted? How do potentially confidence-driven decisions relate to groups’ success in sustainably managing the resource? A multi-experiment, multi-method approach will be taken to deliver robust answers by (i) analyzing existing experimental data where groups of people collectively managed a common resource under externally varied levels of risk; (ii) conducting a specifically designed follow-up experiment that allows for investigating changes in state-of the art measures of understanding and confidence as a function of group interaction; and (iii) a computational modeling approach where possible confidence-based decision strategies are formalized and compared against the behavioral data. In sum, the project will contribute to our understanding of what drives individual and group decisions on common-pool resource use, and will deliver solid estimates of the relative weights of understanding versus confidence in explaining these risky decisions.