CAREER: From Assignment to Evaluation: The Design of School Choice Systems

在过去的十年里，市场设计领域有着大量的活动和兴奋，这是一个规模虽小但不断发展的经济学领域，研究如何设计个人互动的机制和制度，并试图改善现有系统，推荐新机制，并报告其绩效。 特别是，在学生分配制度的研究方面取得了进展，例如在美国和世界各地的许多公立学区的选择计划中发现的那些。 关于这些系统的最初文献建立在早期的机制设计和匹配文献的基础上，而最近的工作已经成熟到经济学家能够就少数学区的分配程序提供建议的程度。 迄今为止，数十万学生已通过受这项工作启发的新分配机制分配到学校。此外，基于这些新机制的持续性以及绩效的总结衡量标准，这些机制的初步经验是有希望的。该提案概述了关于设计、实施和研究的长期研究和教育议程学生分配制度和学校选择。 该建议包括三个部分。 首先，我将研究在实践中应用机制的挑战所激发的重要的未解决的理论问题。 第二，我在波士顿、芝加哥和纽约的经验和与学区的合作关系为我提供了研究参与者如何对特定入学规则的激励做出反应的机会。 反过来，这将使我能够严格衡量基于选择的改革的分配后果。例如，使用新机制的政策变化所产生的变化，我打算研究参与者之间博弈的性质和后果。 参与者是否会对系统做出反应，这些系统会对参与者产生不正当的激励，以揭示他们的真实偏好？ 一些老练的申请人是否从特定的招生规则中获益，而损害了其他不老练的申请人？ 我将研究的另一个问题是关于机制的学习和信息获取。申请人如何了解好学校，以及如何最好地利用选择？ 对该领域机制的实证问题的回答将最终导致更好的系统的设计，并更深入地理解机制之间的相互作用及其实际执行。第三，我将利用我对学生作业程序的了解，利用准实验变异来研究教育生产和学校绩效的长期问题。例如，我打算研究选择性教育的因果关系，同龄人和学生分组的影响，以及各种学校模式对学生成绩的影响。 哪些学校适合哪些孩子？ 孩子们是否从与能力相似的同龄人一起上学中受益？ 这种分层会伤害表现较差的学生吗？ 使用本提案中开发的丰富新数据集的研究将加强公共政策辩论和教育经济学。这项研究自然支持我的教育和外展活动。 我将开发一门新的市场设计研究生课程，为学区提供分配算法方面的帮助，并通过关于学校分配的年度会议将从业者和研究人员聚集在一起。学生分配知识的扩展代表了经济学学科中市场设计的重要发展和机遇。 实地执行机制的经验导致新的理论问题、政策问题和经验分析。 从设计练习中获得的结果和教训将促进创建面向实践的设计文献，以补充现有的机制设计理论文献。 与学区的联系提供了一个独特的机会，可以用新的数据集和经验方法以创新的方式研究长期存在的教育经验问题。该项目具有直接的公共利益：它将帮助学区设计更好的学生分配机制。它还将提高决策者对影响学生成绩的因素的理解。 由于从业人员的投入的重要性，该提案的一个核心组成部分是在NBER主办的一次学术界和从业人员参加的会议，旨在促进和传播与学生分配最佳做法有关的奖学金。 我还将在本科生和研究生阶段开展一些指导活动。

In the last decade, there has been a great deal of activity and excitement in market design, a small but growing field of economics which studies how to design mechanisms and institutions through which individuals interact and also attempts to improve existing systems, recommends new mechanisms, and reports on their performance. In particular, there has been progress in the study of student assignment systems, such as those found in choice plans in many public school districts in the United States and around the world. The initial literature on these systems built on an earlier mechanism design and matching literature, while recent work has matured to a point where economists have been able to advise a handful of school districts on their allocation procedures. To date, hundreds of thousands of students have been assigned to schools via new assignment mechanisms inspired by this work. Moreover, based on the persistence of these new mechanisms as well as summary measures of performance, the initial experience with these mechanisms has been promising.This proposal outlines a long-term agenda of research and education on the design, implementation, and study of student assignment systems and school choice. The proposal consists of three parts. First, I will study important unresolved theoretical questions motivated by challenges in applying mechanisms in practice. These include issues related to dynamics, efficiency in stochastic mechanisms, and the overall desirability of school choice.Second, experiences and partnerships with school districts in Boston, Chicago, and New York provide me the opportunity to study how participants respond to the incentives of particular admissions rules. This, in turn, will allow me to rigorously measure the distributional consequences of choice-based reforms. For example, using variation generated by policy changes from new mechanisms, I intend to study the nature and consequences of gaming among participants. Do participants react to systems which create perverse incentives on participants to reveal their true preferences? Do some sophisticated applicants gain from knowledge of particular admissions rules to the detriment of other unsophisticated applicants? Another issue I will study is learning and information acquisition about mechanisms. How do applicants learn about good schools and how to best utilize choice options? Answers to empirical questions about mechanisms in the field will ultimately lead to the design of better systems and to a deeper understanding of the interaction between mechanisms and their practical implementation.Third, I will use my knowledge of student assignment procedures to exploit quasi-experimental variation to study long-standing questions on education production and school performance. For instance, I intend to study the causal effects of selective education, the impacts of peers and student grouping, and the influence of various school models on student achievement. Which schools are best for which kids? Do children benefit from being in schools with schools with peers of similar ability? Does this stratification hurt lower performing students? Studies using the rich new datasets developed in this proposal will enhance public policy debates and the economics of education.This research naturally supports my education and outreach activities. I will develop a new graduate course in market design, provide school districts assistance with their assignment algorithms, and bring together practitioners and researchers through an annual meeting on school assignment.The expansion of knowledge on student assignment represents an important development and opportunity for market design within the discipline of economics. Experiences implementing mechanisms in the field lead to new theoretical issues, policy questions, and empirical analyses. The results and lessons derived from design exercises will foster the creation of a practically-oriented design literature to complement the existing theoretical literature on mechanism design. Connections to school districts offer a unique opportunity to study long-standing empirical questions on education in innovative ways with new data sets and empirical approaches.This project has a direct public benefit: it will help school districts design better student assignment mechanisms. It will also enhance policy makers' understanding of factors that impact student achievement. Because of the importance of input from practitioners, a central component of the proposal is a conference hosted at the NBER involving academics and practitioners, intended to catalyze and disseminate scholarship related to best practices in student assignment. I will also pursue a number of mentoring activities at the undergraduate and graduate level.