Computational Modeling of Decision Making by Prosecutors and Jurors in Criminal Justice

错误定罪的可能性激发了检察官、法院和其他人对改善刑事司法系统决策的兴趣，部分原因是为了防止错误定罪。为了研究这一问题，本项目将研究检察官和陪审员在决定嫌疑人是否有罪时如何权衡证据和犯罪的其他方面。该项目将采用实验方法，测试检察官和陪审员如何看待刑事调查和审判期间出现的各种问题，这些问题可能涉及判定被告有罪的决定。因此，该项目将研究刑事案件中可能出现的错误，这些错误是如何发生的，以及为什么错误如此难以纠正。该项目将结合实验设计和计算模型，这些模型已经成功地研究了其他情况下的决策，因为这种方法将许多复杂的决策分解为基本的认知步骤，并分析关键变量对每个步骤的影响，认知过程首先，该项目测试犯罪描述如何影响潜在陪审员对指控该罪行的证据的重视程度。第二，这项研究探讨了早期证据，被告是有罪的，可以引发一种倾向，折扣随后的证据有利于被告。最后，该项目测试陪审员和专业检察官是否怀有相似或不同的认知偏见，这可以作为在时间压力，压力或其他分散注意力或消耗认知资源的条件下最有可能发生的错误类型的指南。这种高通量设计与统计建模兼容，统计建模隔离了个体变量的影响及其对被指控犯罪的人是否有罪的决定的潜在相互作用。这项研究包括真实的陪审员和专业检察官作为研究对象，使之有可能比较决策的法律的行为者发挥关键，但不同的，在刑事司法系统中的作用。

The potential for wrongful convictions has spurred interest among prosecutors, courts, and others in improving decision-making in the criminal justice system, in part to prevent wrongful convictions. To study this issue, this project will examine how prosecutors and jurors weigh evidence and other aspects of a crime when deciding whether a suspect is guilty. The project will use an experimental approach to test how prosecutors and jurors view various issues that arise during a criminal investigation and trial, factors which may implicate the decision to find a defendant guilty. Consequently, this project will examine what can go wrong in a criminal case, how those errors occur, and why errors are so difficult to correct.The project will incorporate an experimental design and computational models that have successfully investigated decision making in other contexts, as this approach breaks down many kinds of complex decisions into underlying cognitive steps and for analyzing the impact of key variables on each of those cognitive processes. First, the project tests how the description of a crime affects the weight that potential jurors give to the evidence against someone accused of that crime. Second, the study examines how early evidence that a defendant is guilty can trigger a tendency to discount subsequent evidence favorable to the defendant. Finally, the project tests whether jurors and professional prosecutors harbor similar or different cognitive biases, which can serve as a guide to the kinds of errors that are most likely to occur under time pressure, stress, or other conditions that distract attention or deplete cognitive resources. This high-throughput design is compatible with statistical modeling that isolates the effects of the individual variables and their potential interactions on decisions about the guilt or innocence of someone accused of a crime. The study includes real jurors and professional prosecutors as research subjects, making it possible to compare decision making by legal actors that play critical, but distinct, roles in the criminal justice system.