Modeling educational effectiveness and equity at the school level

学校效能研究涉及学校投入变量(例如学生组成、学校资源和学校结构)、学校内部过程(例如学校氛围、校长的管理风格和教学质量)和学校产出指标之间的关系。输出主要关注学生的成绩，但也关注动机或情感变量。在实证研究中，学校效能的可操作性有不同的方法。在提议的项目中，我们区分质量和公平。质量是由学校根据学生背景特征调整的具体成绩收益(“增值”)来实施的。公平是通过估计学生背景变量(如性别、社会地位、移民背景)与教育结果之间的特定关系来评估的。该项目利用LAU和KESS研究的数据，旨在回答三组研究问题：1.学校有效性的不同操作有多一致？2.哪些输入变量与学校有效性有关？以及3.学校有效性随时间的变化有多稳定？什么是学校有效性变化的预测因素？第一组和第二组的问题可以分别使用LAU和KESS的数据进行分析，结果可以与各自另一项研究的数据进行交叉验证。为了回答来自第三组的问题，利用里程数据的特定结构，将两项研究都视为学校层面上的两次“重复测量”。对于这两者，质量和公平都是使用学生层面的纵向数据进行操作的。这使得可以检查学校效能的稳定性和随时间的变化，并预测第二次学校效能(KESS)，并在较早的时间点(LAU)评估学校水平预测指标。除了与学校效能有关的研究问题外，拟议的项目还将探讨与所需分析技术有关的方法学问题。具体地说，将审查决定基于模型的学校效能变量的可靠性的因素，比较潜在变量模型和观测变量模型，并探索学校效能变量的可信数值的产生。

School effectiveness research deals with associations between school input variables (e.g. student composition, school resources, and school structure), processes within school (e.g. school climate, management style of the school principal, and instructional quality), and school output measures. Output predominantly focuses on student achievement, but also on motivational or affective variables. There are different ways to operationalize school effectiveness in empirical research. Within the proposed project, we distinguish between quality and equity. Quality is operationalized by the school specific gain in achievement adjusted for student background characteristics ("value added"). Equity is assessed by estimating school specific relations between student background variables (e.g. gender, SES, migration background) and educational outcomes.Using data from the LAU and KESS studies, the proposed project aims at answering three groups of research questions: 1. How consistent are different operationalizations of school effectiveness?, 2. What input variables are associated with school effectiveness?, and 3. how stable is school effectiveness across time and what are predictors for change in school effectiveness? Analyses for questions from groups 1 and 2 can conducted with data from LAU and KESS separately, and results can be cross-validated with data from the respective other study. To answer questions from group 3, the specific structure of the MILES data is utilized by treating both studies as two "repeated measurements" on school level. For both, quality and equity are operationalized using longitudinal data on student level. This allows to examine stability and change in school effectiveness across time and to predict school effectiveness at the second occasion (KESS) with school level predictors assessed at the earlier time point (LAU). In addition to the research questions related to school effectiveness, the proposed project will also pursue methodological questions related to the required analysis techniques. Specifically, factors determining the reliability of model based school effectiveness variables will be examined, latent and observed variable models will be compared, and the generation of plausible values for school effectiveness variables will be explored.