Effects of studying abroad during high school - A reanalysis of the LAU study

本项目旨在调查高中期间出国留学对学生成绩的影响。尽管在过去的几年里，教育流动性显著增加（OECD, 2014），但只有少数研究关注个人和情境预测因素以及高中期间出国留学的影响。关于留学效果的研究数量有限的原因之一是为了得出有意义的推论，研究设计和数据必须达到很高的要求。利用法学研究的纵向数据（Behörde f<s:1> r Schule und Berufsbildung, 2012; Lehmann, Hunger, Ivanov, Gänsfuß & Hoffmann, 2003; Lehmann, Peek, Gänsfuß & Husfeldt, 2002; Trautwein, Köller, Lehmann & l<s:1> dtke, 2007）将使我们能够应对这些挑战。我们建议解决三个研究问题：(1)学生和学校的哪些特征预测学生是否会在高中期间出国留学？我们将特别关注学生的家庭背景特征，以及他们通过标准化成绩测试所衡量的成绩特征。法律研究包括在出国之前评估的一套丰富的措施，这些措施非常适合这一目的。(2)在控制了已有差异后，高中时期在英语国家留学对英语技能发展（以标准化成就测试衡量）和高中英语成绩的影响是什么？law -13的数据将使我们能够分析英语表现在各个方面（阅读/写作/听力）的差异发展。我们还将测试课程选择行为（例如，将英语作为一门高级课程）是否会调节出国留学对英语成绩的影响，以及出国留学的持续时间是否会系统性地影响出国留学的有效性。(3)高中期间出国留学是否也会对其他科目的表现和成绩产生影响，尤其是数学？由于学生们在国外学习期间往往有很大的灵活性选择课程，对其他学科的影响似乎是可能的，但之前还没有系统的研究。为了充分利用LAU数据集的优势，有必要首先准备复杂的数据，并实施适当的策略来考虑缺失值。为了确定高中期间出国留学的因果关系，必须确定和考虑指导选择过程的协变量。为此，我们将使用倾向评分方法，允许灵活的选择模型规范。我们将使用敏感性分析检查估计效应对未观察到的混杂因素的稳健性。

The proposed project aims to investigate the effects of studying abroad during high school on student achievement. Although there has been a significant increase in educational mobility during the past years (OECD, 2014), there only have been a few studies that focus on individual and contextual predictors and the effects of studying abroad during high school. One of the reasons for the limited amount of research on the effects of studying abroad is that the research designs and data must meet high requirements in order for meaningful inferences to be drawn. Using longitudinal data from the LAU-study (Behörde für Schule und Berufsbildung, 2012; Lehmann, Hunger, Ivanov, Gänsfuß & Hoffmann, 2003; Lehmann, Peek, Gänsfuß & Husfeldt, 2002; Trautwein, Köller, Lehmann & Lüdtke, 2007) will allow us to address these challenges.We propose to address three research questions: (1) What student and school characteristics predict whether students will study abroad during high school? We will focus in particular on students' family background characteristics as well as on their achievement characteristics as measured by standardized achievement tests. The LAU-study includes a rich set of measures assessed prior to going abroad that is ideally suited for this purpose. (2) What are the effects of studying abroad in English-speaking countries during high school on the development of English skills (as measured by standardized achievement tests) and on high school grades in English after controlling for pre-existing differences? The LAU-13 data will allow us to analyze differential developments in various dimensions of English performance (reading/writing/listening). We will also test whether course-choice behavior (e.g., English as an advanced course) mediates the effects of stays abroad on English achievement and whether the duration of a stay abroad systematically influences the effectiveness of such a stay. (3) Does studying abroad during high school also have an effect on performance and grades in other subjects, particularly in mathematics? As students tend to have plenty of flexibility in choosing courses during their time abroad, an impact on other subjects seems likely, but has not been systematically investigated before.In order to fully exploit the strengths of the LAU dataset, it is necessary to first prepare the complex data and implement an adequate strategy for taking missing values into account. In order to identify causal effects of studying abroad during high school, the covariates that guide the selection process have to be identified and accounted for. For this purpose, we will use propensity score methods that allow for flexible specifications of the selection model. We will check the robustness of the estimated effects against unobserved confounders using sensitivity analyses.