The aim of this article is to assess whether the track attended in upper secondary education affects student competencies in Italy, by disentangling the genuine effects of track choices from selection biases related to the characteristics of students enrolled in different tracks. We contribute to the literature by relying on a more detailed measure of tracking, by focusing on between-school tracking and exploring whether track effects vary systematically by student social background, a largely overlooked issue in previous research. We adopt a counterfactual approach and rely on population panel data on a recent cohort of students assessed in 5^th, 8^th and 10^th grade. We rely on a difference-in-difference strategy integrated with marginal mean weighting with stratification and inverse probability weighting, which are used respectively to better control for selection into tracks and account for missing data. First, we document strong social selection into tracks, along various students’ characteristics. Second, we find that track effects are smaller once accounting for selection processes, but are still substantial on both reading and mathematics competencies, albeit slightly larger in the latter subject. Beyond the anticipated advantage of the academic track over vocational education, we also find differential effects of attending different curricula within these tracks. Third, the benefits of attending the academic tracks appear to be rather homogeneous across students from different social backgrounds.

本文旨在评估参加高中教育的课程是否会影响意大利的学生能力，方法是将课程选择的真正影响与与不同课程注册学生的特征相关的选择偏差分开。我们依靠更详细的跟踪测量，通过关注校际跟踪并探索跟踪效应是否因学生社会背景而系统地变化，这是以前研究中很大程度上被忽视的问题，从而为文献做出贡献。我们采用反办法，依靠人口面板数据在5评估学生的近队列^日，8^日和10^日年级。我们依赖于差异中的差异策略，该策略与具有分层的边际均值加权和逆概率加权相结合，分别用于更好地控制选择到轨道并考虑丢失的数据。首先，我们根据不同学生的特征将强烈的社会选择记录到轨道中。其次，我们发现，一旦考虑到选择过程，轨道效应就较小，但在阅读和数学能力方面仍然很重要，尽管在后一科目中略大。除了学术轨道相对于职业教育的预期优势之外，我们还发现在这些轨道内参加不同课程的不同影响。第三，来自不同社会背景的学生参加学术课程的好处似乎相当相似。