Integrating Formative Assessment of Computational Thinking with Self-Regulated Learning

这是一个为期两年的探索性项目，提交给了发现研究PreK-12 （DRK-12）计划的评估链，也响应了NSF 17-149亲爱的同事信：发现研究PreK-12：推进STEM+计算。它旨在开发、完善、测试和验证一个原型，以说明计算机科学（CS）学科中教师（n=10）和学生（n=160）的形成性评估（FA）、计算思维（CT）和自我调节学习（SRL）之间的协同作用。该项目将重点关注FA，也称为学习评估，它生成用于反馈的信息，以修改教学活动。通过明确地激发学生在解决问题过程中的SRL学习，评估任务将说明FA和SRL之间的协同作用。在尝试解决之前，学生将被提示分析CT任务，明确设定目标，并反思教师的反馈，以指导他们在随后的尝试中实现这些目标的行为。为了引出SRL行为，CT任务设计将遵循一个三相模型。第一阶段将重点放在目标设定和任务分析上。第二阶段将侧重于任务执行过程中的自我监控，而第三阶段将侧重于反思。每个任务将有两次性能迭代，允许学生和教师将先前的性能反馈用于使用。任务的内容将取自中学课堂上使用的计算机科学课程。他们将在拥有不同学生群体的公立学校系统的真实课堂环境中得到验证。本工作计划取得以下成果：(1)创建4个原型CT评估任务；(2)开发任务的信度和效度论证；(3)过程型教师反馈与SRL外显使用对学生CT成绩的影响研究；(4)教师专业知识的发展。该项目将解决两个关键的研究问题：(1)当教师反馈对SRL过程和绩效结果提供特定目标的反馈时，教师反馈是否会对学生的绩效产生积极影响？(2)与没有明确提示参与SRL的学生相比，参与SRL循环的学生在CT任务中的表现是否有所提高？为了回答这些问题，该项目将进行两项研究。第一项研究将调查教师对SRL过程和任务表现的反馈对学生CT成绩的积极影响程度，而不仅仅是基于结果的反馈。本研究将采用准实验，前后测试设计与对照组。实验组的学生将收到关于SRL过程和CT任务表现的反馈。在控制条件下的学生只会收到成绩的反馈。第二项研究将调查通过外显提示参与SRL的学生与没有SRL提示的学生相比，在多大程度上提高了他们在CT任务中的表现。本研究将采用实验混合模型设计来检验实验组之间的差异以及学生在三个重复测量周期中的表现变化。这项工作的主要成果将是一个原型，说明CT、FA和SRL之间潜在的协同作用，这将在未来用于进一步研究这些领域之间的交叉点。在这个项目下制定的任务将通过多个从业者网络免费提供给公众。咨询委员会和外部评价人员将就工作的进展和质量提供反馈意见和建议。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This is a two-year exploratory project submitted to the assessment strand of the Discovery Research PreK-12 (DRK-12) Program, also responsive to NSF 17-149 Dear Colleague Letter: Discovery Research PreK-12: Advancing STEM+Computing. It is aimed at developing, refining, testing, and validating a prototype to illustrate the synergy among formative assessment (FA), computational thinking (CT), and self-regulated learning (SRL) within the discipline of computer science (CS) with teachers (n=10) and students (n=160) in 9th-12th grades. The project will focus on FA, also known as assessment for learning, which generates information that is used for feedback to modify teaching and learning activities. By explicitly eliciting student SRL learning during problem-solving, assessment tasks will illustrate the synergy between FA and SRL. Students will be prompted to analyze a CT task before attempting to solve it, explicitly setting goals, and reflecting on teacher feedback to direct their behaviors towards achieving these goals in subsequent attempts. To elicit SRL behavior, CT task design will follow a three-phase model. Phase 1 will focus on goal setting and task analysis. Phase 2 will focus on self-monitoring during task performance, while Phase 3 will center on reflection. Each task will have two performance iterations, allowing students and teachers to put feedback about prior performance into use. Tasks' content will be drawn from the CS curricula used in secondary school classrooms. They will be validated in authentic classroom contexts in a public school system with a diverse student population. This effort plans to achieve the following outcomes: (1) creation of four prototype CT assessment tasks; (2) development of reliability and validity arguments for the tasks; (3) research into impacts on student CT performance of process-oriented teacher feedback and explicit use of SRL; and (4) development of teacher expertise in FA of CT. The project will address two key research questions: (1) Does teacher feedback positively affect student performance when it provides goal-specific feedback on SRL processes, as well as performance outcomes?; and (2) Do students who engage in SRL cycles improve performance on CT tasks, as compared to students who complete tasks without explicit prompts to engage in SRL? To answer these questions, the project will conduct two research studies. The first study will investigate the extent to which teacher feedback on both SRL processes and task performance positively affects student CT achievement, more than outcome-based feedback only. This study will use a quasi-experimental, pre- post-test design with comparison group. Students in the treatment condition will receive feedback on SRL processes and CT task performance. Students in the control condition will receive feedback on performance outcomes only. The second study will investigate the extent to which students who engage in SRL through explicit prompting improve performance on CT tasks, as compared to students without SRL prompts. This study will employ an experimental mixed-model design to examine differences between experimental groups and changes in students' performance over three repeated measurement cycles. The main outcome of this effort will be a prototype illustrating the potential synergies among CT, FA, and SRL which will be used in the future to further research on the intersections among these domains. The tasks developed under this project will be made available online at no cost to the public through multiple practitioner networks. An advisory board and external evaluator will provide feedback and recommendations in regard to the progress and quality of the work.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.