Title: The quality of school track assignment decisions by teachers Show Chinese title

Title: 教师在学校田径队分配决策中的质量

Abstract: This paper analyzes the effects of educational tracking and the quality of track assignment decisions. We motivate our analysis using a model of optimal track assignment under uncertainty. This model generates predictions about the average effects of tracking at the margin of the assignment process. In addition, we recognize that the average effects do not measure noise in the assignment process, as they may reflect a mix of both positive and negative tracking effects. To test these ideas, we develop a flexible causal approach that separates, organizes, and partially identifies tracking effects of any sign or form. We apply this approach in the context of a regression discontinuity design in the Netherlands, where teachers issue track recommendations that may be revised based on test score cutoffs, and where in some cases parents can overrule this recommendation. Our results indicate substantial tracking effects: between 40% and 100% of reassigned students are positively or negatively affected by enrolling in a higher track. Most tracking effects are positive, however, with students benefiting from being placed in a higher, more demanding track. While based on the current analysis we cannot reject the hypothesis that teacher assignments are unbiased, this result seems only consistent with a significant degree of noise. We discuss that parental decisions, whether to follow or deviate from teacher recommendations, may help reducing this noise. Show Chinese abstract

Abstract: 本文分析了教育分轨及其分轨决策质量的影响。 我们使用在不确定性下的最优分轨模型来驱动我们的分析。 该模型对分轨在分配过程边缘的平均影响产生了预测。 此外，我们认识到平均影响并不能衡量分配过程中的噪声，因为它们可能反映了正负分轨效应的混合。 为了验证这些想法，我们开发了一种灵活的因果方法，能够分离、组织并部分识别任何符号或形式的分轨效应。 我们在荷兰的回归不连续设计背景下应用这种方法，其中教师会提出分轨建议，这些建议可能根据考试成绩截止值进行修改，并且在某些情况下家长可以推翻这些建议。 我们的结果表明分轨影响显著：重新分配的学生中有40%到100%因进入更高轨道而受到积极或消极影响。 然而，大多数分轨影响是积极的，学生从被分配到更高、更具挑战性的轨道中受益。 虽然基于当前分析我们无法拒绝教师分配无偏的假设，但这一结果似乎仅与相当程度的噪声一致。 我们讨论了家长决定是否遵循或偏离教师建议，可能有助于减少这种噪声。