RETA: Development of Longitudinal Models Linking Student Mathematics Achievement to Instruction Content

该项目旨在开发和评估纵向模型，以考察学生在学校环境中学习什么以及如何学习。它以数学为重点，利用现有的促进严格数学/科学教育成果(PROM/SE)项目收集的独特数据集。这项全面的NSF/数学与科学伙伴关系(MSP)工作涉及密歇根州和俄亥俄州50多个学区的30多万名3年级至12年级的学生和约4,000名教师，他们收集了过去4年(2004-2008)的数据。它包括不同时间点的多年级课程衡量：(A)学生在多个内容主题中的知识和(B)学生学习这些主题的机会。学生样本在人口统计和社会经济因素上都反映了这个国家的情况。纵向模型将涉及三个重要问题：(Ii)学生在特定课程敏感内容领域的数学知识随时间的变化；以及(Iii)这两者之间的关系如何变化。这使研究人员能够为三种不同的纵向设计开发统计模型并对其进行评估：i)一个用于合成队列，涉及单个时间点的多个等级测量；ii)第二个用于在多个时间点重复测量的真实等级队列；以及iii)单个纵向模型，涉及多个时间点的重复测量。纵向模型包含了三个重要的教育问题：i)学生学习机会(OTL)课程内容，即学生随着时间的推移所经历的暴露；ii)学生在特定内容领域的成绩随时间的变化；以及iii)OTL与成绩之间的关系。需要探索与为每个国家开发的统计模型的匹配性，并就导致政策影响的结果和结论进行比较和对比。这一结果有望成为全国提高学生成绩的典范。

The project is to develop and evaluate longitudinal models that examine what and how students learn in the school setting. With a focus on mathematics, it draws on a unique data set gathered by the existing Promoting Rigorous Outcomes in Mathematics/Science Education (PROM/SE) project. This Comprehensive NSF/Math and Science Partnership (MSP) effort involving over 300,000 students in grades 3 through 12 and approximately 4,000 teachers in over 50 districts from the states of Michigan and Ohio collected data over the past 4 years, 2004-2008. It consists of multi-grade curriculum measurements at different time points of (a) student knowledge in multiple content topics and of (b) student's opportunity to learn those topics. The sample of students mirrors that of the nation both in demographics and socio-economic factors. The longitudinal models will involve three important issues: the exposure to important mathematics content students experience over time, (ii) the change in student mathematics knowledge in specific curricular sensitive content areas over time, and iii) how the relationship between these two changes. This enables researchers to develop statistical models and evaluate them for three different longitudinal designs: i) one for synthetic cohorts involving multi-grade measurements at a single time point; ii) a second for true grade cohorts with repeated measurements at multiple time points; and iii) an individual longitudinal model involving repeated measurements at multiple time points. The longitudinal models incorporate three important educational issues : i) students opportunity to learn (OTL) curriculum content, i.e., the exposure students experience over time, ii) the change in student achievement in specific content areas over time, and iii) the relationship between OTL and achievement. The fit with the statistical models developed for each is to be explored, and compared and contrasted for results and conclusions that lead to policy implications. The results promise to inform a national model for the improvement of student achievement.