多标准录取中优先级序的作用
The Role of Precedence Order in Matching with Multi-Criteria Admissions
浏览论文内容
中文总结 AI 辅助
该研究分析多标准录取中优先级序的影响，推导相关比较静态，指出延迟接受机制下反馈是比较反转的唯一来源，大市场中失败会消失，波士顿机制下比较恒成立。
中文摘要 AI 辅助
录取系统常通过依次应用各标准的排名来填补名额。当排名可能完全不一致，且各标准通过响应偏好评估整个录取集合时，我们推导了关于优先级序的标准比较静态分析。对于固定的申请者集合，反转最后两个块会使每个标准更偏好其块靠后的结果。若两个块后还有另一标准，即使在储备型排名下，该比较也可能反转。在学生提议的延迟接受机制下，因拒绝链会改变焦点高校的申请者集合，两个标准下也可能出现反转，这种反馈是唯一可能的失败来源。在常规大市场中，几乎所有高校的失败都会消失，且在足够密集时，所有高校间的失败也会均匀消失。在波士顿机制下，最终录取可阻止反馈，因此对于固定提交的秩次列表和标准独立的可接受性，该比较在每个有限市场中均成立。
英文摘要
Admission systems often fill seats by applying criterion-specific rankings sequentially. We derive criterion-wise comparative statics with respect to precedence order when rankings may be arbitrarily nonaligned, and each criterion evaluates the entire admitted set through a responsive preference. For a fixed applicant set, reversing the final two blocks makes each criterion weakly prefer the outcome in which its block is later. This comparison can reverse if two blocks are followed by another criterion, even under reserve-type rankings. Under student-proposing deferred acceptance, it can also reverse with two criteria because rejection chains change the focal college's applicant set; such feedback is the only possible source of failure. In regular large markets, failures vanish for almost all colleges and, under sufficient thickness, uniformly across colleges. Under the Boston mechanism, final acceptance prevents feedback; thus the comparison holds in every finite market for fixed submitted rank-order lists and criterion-independent acceptability.
