CAREER: Engineering opportunity: Manipulating choice architecture to attenuate social bias

社会上许多最重要的社会决策都是由一群人做出的。例如，招聘决策通常涉及对一组求职者的评估；住房决定包括从一组申请人中进行选择；投票决定是由一组竞选公职的候选人做出的。理性选择理论认为，无论其他选择的数量或质量如何，决策者在一组选择中的任意两个选择之间的偏好应该保持不变。然而，研究表明，人们对选择集中每个选项的偏好会以可预测的方式变化，这取决于其他可用的选择。我们在消费者行为的研究中看到了这一点，例如，第三次劣质产品的引入改变了消费者对两种原始产品的偏好（所谓的“诱饵效应”）。当选择涉及到其他人时，社会陈规定型观念和相关情绪往往导致系统性歧视，尤其是针对边缘社会群体的歧视。在这个项目中，哈佛大学的研究员Mina Cikara研究了选择集的构建——选择架构——是如何影响歧视的。过去大多数减少社会决策偏见的努力都集中在改变感知者的刻板印象和偏见上。相比之下，该项目侧重于改变选择的背景，将其作为减少歧视的一种方式。从认知心理学和计算生物学的正式决策模型中，该研究解决了关于选择结构在歧视中的作用的知识空白，并提供了可能减少各种相应社会背景下的歧视行为的见解。该项目采用跨学科的方法，整合了认知和社会心理学、神经经济学和计算生物学的方法和发现，来研究一个被忽视的、但潜在的强大的减少歧视的来源：选择架构。正式的决策模型对社会“诱饵”影响决策的机制和决策过程背后的时间动态作出具体预测。将这些模型的见解整合到社会决策研究中，可以提高预测精度，并激发减少偏见的创新策略。在招聘、住房和投票决策领域的一系列实验和实地研究测试了社会诱饵是否会增加决策者对其他方面处于不利地位的群体（如老年人、少数民族、妇女）的偏好。进一步的实验利用参与者对诱饵效应的敏感性来量化人口统计学和刻板印象相关属性对社会决策的影响。最后一组实验操纵参与者对目标属性信息的暴露，以转向减少偏见的干预，不需要在选择集中制造或选择性地包括社会诱饵。这个研究项目为不同的学生群体提供培训。研究结果将广泛传播，并特别努力使工业管理人员了解。许多接触点与科学探究的其他领域，包括神经科学，社会学，经济学和生物学。该项目具有广泛的社会影响，包括为公共和私营机构提供机会，深入了解如何制定政策，从而在许多不同规模上减少社会偏见。

Many of society's most significant social decisions are made over sets of individuals. For example, hiring decisions typically involve the evaluation of a set of job candidates; housing decisions involve selection from among a set of applicants; voting decisions are made from among a set of candidates running for office. Rational theories of choice suggest that decision makers' preferences between any two options within a set should remain the same regardless of the number or the quality of other options. Yet, research has shown that people's preferences for each option in a choice set shift in predictable ways depending on the other available alternatives. We see this in the study of consumer behavior, for example, where the introduction of a third inferior product changes consumers' preferences for the two original products (the so-called "decoy effect"). When choices involve other people, social stereotypes and associated emotions often lead to systematic discrimination, especially against marginalized social groups. In this project, investigator Mina Cikara of Harvard University examines how the construction of choice sets -- choice architecture -- influences discrimination. Most past efforts to reduce bias in social decisions have focused on changing perceivers' stereotypes and prejudices. In contrast, this project focuses on changing the context in which choices are presented as a way to reduce discrimination. Drawing from formal models of decision making in cognitive psychology and computational biology, the research addresses a major gap in knowledge about the role of choice architecture in discrimination and provides insights that may reduce discriminatory practices in a variety of consequential social contexts.This project adopts an inter-disciplinary approach, integrating methods and findings from cognitive and social psychology, neuroeconomics, and computational biology to examine a neglected, but potentially powerful source of discrimination reduction: choice architecture. Formal models of decision-making make specific predictions about both the mechanisms by which social "decoys" should influence decisions and the temporal dynamics underlying the decision process. Integrating insights from these models into the study of social-decision making allows for greater predictive precision and stimulates innovative strategies for reducing bias. A series of experiments and field studies in the domains of hiring, housing, and voting decisions test whether social decoys can increase decision-makers' preferences for groups that are otherwise disadvantaged (e.g., elderly, racial minorities, women). Further experiments use participants' susceptibility to decoy effects to quantify the influence of demographic and stereotype-related attributes on social decision-making. A final set of experiments manipulate participants' exposure to targets' attribute information to move toward bias-reduction interventions that do not require fabricating or selectively including social decoys in choice sets. This research program provides training to a diverse group of students. The research results will be broadly disseminated, with special efforts to reach managers in industry. Many points of contact are made with other areas of scientific inquiry, including neuroscience, sociology, economics and biology. The project offers broad societal impacts, including the opportunity for public and private institutions to gain insight into shaping policy that can reduce social bias at many different scales.