Integrating Trust and Feedback Intervention Theories to Predict Behavioral Change in Response to Algorithmic Feedback

整合信任和反馈干预理论来预测响应算法反馈的行为变化

• 批准号: 2020863
• 负责人: Richard Landers
• 金额: $ 41.87万
• 依托单位: University of Minnesota-Twin Cities
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-08-15 至 2024-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2020863&HistoricalAwards=false
• 关键词: Integrating; Trust; Feedback; Intervention; Theories

As organizations increasingly integrate algorithms into their decision-making, (such as by providing decision aids and algorithmically generated advice to employees), it has become evident that we lack a scientific understanding of when and why people heed such advice. In this project, an integrative theory linking the characteristics of algorithmic advice to advice-taking behavior is proposed and tested. To test this theory, we develop a web-based application that allows internet users to complete automated virtual job interviews and receive algorithmic feedback or human feedback on their performance. In doing so, we improve scientific understanding of how people respond to algorithmic feedback while simultaneously providing people with authentic feedback on their interview performance, a societal benefit. In this project, organizational trust theory, which specifies a theoretical structure for trust and likely consequences, has been integrated into feedback intervention theory, which describes the process by which people act upon received feedback, to better predict behavioral change in response to algorithmic feedback. This study thus fills theoretical gaps about the influence of feedback source on interview performance while also informing broader questions regarding algorithms, trust, and behavioral change. The study’s core propositions will be tested with a between-subjects experimental design and authentic job seekers to maximize generalizability to the present-day workforce.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

随着组织越来越多地将算法集成到决策中（例如通过向员工提供决策辅助和算法生成的建议），很明显，我们对人们何时以及为什么会听取这些建议缺乏科学的理解。在这个项目中，一个整合的理论，算法的建议，建议采取行为的特点，提出和测试。为了测试这一理论，我们开发了一个基于网络的应用程序，允许互联网用户完成自动化的虚拟工作面试，并接收算法反馈或人工反馈。通过这样做，我们提高了对人们如何回应算法反馈的科学理解，同时为人们提供关于他们面试表现的真实反馈，这是一种社会效益。 在这个项目中，组织信任理论，它指定了一个理论结构的信任和可能的后果，已被整合到反馈干预理论，它描述了人们对收到的反馈采取行动的过程，以更好地预测响应算法反馈的行为变化。 因此，这项研究填补了关于反馈源对面试表现影响的理论空白，同时也为算法，信任和行为变化等更广泛的问题提供了信息。 该研究的核心命题将通过受试者之间的实验设计和真实的求职者进行测试，以最大限度地提高对当今劳动力的普遍性。该奖项反映了NSF的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。