Designing Multisite Mediation Studies to Track Teacher Development Processes in Mathematics

人们普遍认识到，关于可塑性教师属性的研究是改善和推进STEM教育所必需的基础知识的关键组成部分。了解教师如何影响学生的一种研究方法是中介分析。中介分析通过探究实现这些关系的教师的中间发展过程和行为，解释了属性如何以及为什么与学生的进步有关。多个学校或多个地点的调解研究可以建立关于特定属性是否、如何以及为什么有助于改善学生结果的多层次证据。他们可以描述干预措施(例如，通过专业发展计划)如何改善教师和学生的结果。他们可以同时规划教师发展过程，创造机会发现、拆开和了解这些过程中的变化，并适应学校内部自然发生的变化。遗憾的是，在教师发展和有效性的背景下，缺乏有意识地设计有效和高效的多学校调解研究的方法和战略方面的文献和工具。在这个项目中，将通过为多学校调解研究开发一个理论和实证框架来加强这一探索线。拟议的工作旨在获得有用的工具，即新的统计估计器、原则和战略，用于广泛适用于STEM地区的多站点设计。它还打算开发基于经验的策略和软件，用于数学教师发展的研究。这些工具整合了四个关键部分：(A)开发统计估计器，以指导观察和实验多地点调解研究的设计；(B)制定经验基础，在丰富的教师发展进程中实施这些战略；(C)开发自由和可访问的软件，以实施这些结果；以及(D)开发培训材料，以提高实地能力。将开发一套工具--统计、经验、软件和培训--以前瞻性地设计和实施多地点调解教师发展研究，以便它们能够提供有效、高效和全面的证据。这项工作分为三个部分。第一部分旨在发展多地点调解研究有效和高效设计的统计理论，并有四个目标：(A)在观察性和试验性设计中得出多地点调解的统计功率估计值，(B)扩展这些公式以适应调解过程的调节者(在不同的教师发展过程下)，(C)扩展这些公式以包括协变量，以及(D)推导公式以确定给定约束的最优(最有效)抽样方案。第二部分将把统计理论应用于数学教师发展的研究。该项目旨在通过对数学教育中一系列丰富的教师属性和中介因素进行元分析，对设计教师发展研究所需的参数进行经验估计。具体地说，该项目计划制定(A)对众多教师和学生变量的必要参数的经验估计，以及(B)以经验为基础的指导方针和战略，以提高多地点教师发展研究的效果。最后一个组成部分是开发软件，利用免费和用户友好的软件进行上述公式和计算，并举办讲习班，培训范围广泛的研究人员。最终，该项目寻求改善教师发展调解研究产生的证据的范围和质量。核心平台旨在转移到其他STEM教育领域和STEM学科。此外，该项目开发的软件和讲习班旨在以可获得的方式将成果分发给STEM教育领域和STEM学科的广泛和多样化的研究和从业人员群体。因此，该项目有可能极大地扩大通过数学教师发展研究产生的证据的范围和质量，更广泛地说，通过整个STEM地区的多地点调解研究来产生证据。随着这项工作提高研究人员设计有效、高效和全面研究的能力，这种潜力就存在了。该项目由本科生教育部门改善本科生STEM教育计划作为开发和实施项目提供支持。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

There is widespread recognition that research on malleable teacher attributes represents a critical component of the foundational knowledge necessary to improve and advance STEM education. A method of research to understand how teachers affect students is mediation analyses. Mediation analyses decode how and why attributes are connected to student improvement by probing the intermediate developmental processes and behaviors of teachers that operationalize these relationships. Multi-school or multisite mediation studies can build a multilayered body of evidence regarding if, how, and why specific attributes help improve student outcomes. They can delineate how interventions (e.g., through professional development programs) might improve both teacher and student outcomes. They can concurrently map out teacher development processes, create opportunities to discover, unpack and learn about variation in these processes, and accommodate naturally occurring variation within schools. Unfortunately, there are scant literature and tools available on methods and strategies to intentionally design effective and efficient multi-school mediation studies in the context of teacher development and effectiveness. In this project, this line of inquiry will be enhanced by developing a theoretical and empirical framework for multi-school mediation studies. The proposed work aims to derive useful tools, that is new statistical estimators, principles, and strategies for multisite designs that are widely applicable across STEM areas. It also intends to develop empirically-based strategies and software for studies of teacher development in mathematics. These tools integrate four key components: (a) the development of statistical estimators to direct the design of observational and experimental multisite studies of mediation, (b) the development of an empirical basis from which to operationalize these strategies for a rich array of teacher development processes, (c) the development of free and accessible software to implement these results, and (d) the development of training materials to improve the capacity of the field.This project aims to develop a framework to direct the effective and efficient design of multisite mediation studies of teacher development in mathematics. A suite of tools--statistical, empirical, software, and training--will be developed to prospectively design and implement multisite mediation teacher development studies so that they are positioned to produce effective, efficient, and comprehensive evidence. The work partitions into three components. The first component aims to develop the statistical theory underlying the effective and efficient design of multisite mediation studies and has four objectives: (a) derive statistical power estimators for multisite mediation in observational and experimental designs, (b) extend these formulas to accommodate moderators of the mediation processes (under heterogeneous teacher development processes), (c) extend these formulas to include covariates, and (d) derive formulas to identify the optimal (most powerful) sampling plans given constraints. The second component will apply the resulting statistical theory to the study of teacher development in mathematics. The project intends to develop empirical estimates of the parameters needed to design teacher development studies by conducting a meta-analysis on a rich array of teacher attributes and mediators in mathematics education. Specifically, the project plans to develop (a) empirical estimates of the necessary parameters for numerous teacher and student variables and (b) empirically-based guidelines and strategies for improving the efficacy of multisite teacher development studies. The final component is the development of software to carry out the above formulas and calculations using free and user-friendly software and workshops to train a broad range of researchers. Ultimately, this project seeks to improve the scope and quality of evidence produced by mediation studies of teacher development. The core platform is intended to be transferable to other STEM education areas and STEM disciplines. Moreover, the software and workshops developed in this project are intended to be used to distribute the results in an accessible manner to broad and diverse sets of research and practitioner groups across STEM education areas and STEM disciplines. Thus, this project has the potential to substantially expand the scope and quality of evidence generated through mathematics teacher development studies and, more generally, multisite mediation studies throughout STEM areas. This potential exists as this work increases researchers' capacity to design effective, efficient, and comprehensive studies. This project is supported by the Division of Undergraduate Education's Improving Undergraduate STEM Education program as a Development and Implementation project.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.