Collaborative Research: Interactive Online Support for Open-Ended Problem Solving Spanning Science Practices and Domain Topics

在线作业广泛应用于大学水平的科学、技术、工程和数学（STEM）入门课程，目的是促进学生的学习和成功。在线作业系统鼓励学生在整个学期中定期学习，并在对教师要求最低的情况下支持学生解决问题的练习。然而，目前的在线作业系统通常只对学生尝试的多步问题的最终答案提供反馈。这个改进本科STEM教育（IUSE:EHR）开发与实施项目将开发一个改进的在线作业系统，该系统将在学生参与自由形式、多步骤的问题解决过程中跟踪学生，并在整个过程中提供必要的提示和反馈。此外，该界面将促进学生实践构建科学论证，包括用证据支持主张。这些在线工具将在多个机构的普通化学课程中实施、测试和修订，包括卡内基-梅隆大学、爱荷华州立大学和得梅因地区社区学院。因此，高质量的在线资源有望帮助学生培养在化学和其他STEM领域取得成功所需的解决问题的专业知识。该项目将围绕以下假设进行设计研究，即通过在线作业可以促进稳健的学习：(i)支持自由形式、多步骤的问题解决；（ii）让学生参与一系列科学实践；（iii）提供了整合跨领域概念的实践。这项工作将通过两个主要扩展建立在认知导师创作工具的成功基础上。第一个是拖放式表达式构建器，它将允许学生构建解决问题的步骤和科学论证，具有一定程度的灵活性，接近纸笔解决问题，同时允许计算系统通过提示和反馈来监控和支撑工作。第二个是域浏览器，学生可以从中拖放他们的表达式项。这将允许计算机系统监控和指导学生制定解决给定问题的策略。因此，预计该项目将扩大问题解决活动的范围，这些活动可以以研究表明将改善学生学习的方式在线支持。更多开放式、多步骤的问题将为学生提供发展解决问题策略的练习。跨越更大范围的科学实践将帮助学生学习领域内容，同时参与更好地反映化学家的实践。系统设计的迭代将通过：将学生在系统内解决的问题与纸笔解决的问题进行比较；检查学生使用的详细日志文件，以确定系统提供的提示和反馈是否适当；以及评估系统能够支持跨越科学实践的问题的程度，并包括来自整个领域的主题。对从系统中收集的日志文件的分析将允许改进活动的设计迭代，并提供对学生如何解决问题的基本见解。项目的最后一年将包括对不同学生群体的总结性评估。在线工具和评价结果都将广泛传播。

Online homework is widely used in introductory science, technology, engineering and mathematics (STEM) courses at the college level with the goal of contributing to student learning and success. Online homework systems encourage students to study regularly throughout the semester and support student problem-solving practice with minimal demands on instructors. However, current online homework systems typically provide feedback only on the final answers to multistep problems attempted by students. This Improving Undergraduate STEM Education (IUSE:EHR) Development & Implementation project will develop an improved system for online homework that will follow students as they engage in free-form, multistep problem solving, and provide hints and feedback as needed throughout the process. In addition, the interface will facilitate student practice with constructing scientific arguments that involve supporting claims with evidence. The online tools will be implemented, tested, and revised in general chemistry courses at multiple institutions, including Carnegie-Mellon University, Iowa State University, and Des Moines Area Community College. Thus, the high-quality online resources produced are expected to help students develop the problem-solving expertise that is needed for success in chemistry and other STEM fields.The project will be structured as design research around the conjecture that robust learning can be promoted through online homework that: (i) supports free-form, multistep problem solving; (ii) engages students in a range of science practices; and (iii) provides practice in integrating concepts from across the domain. The work will build on the success of the Cognitive Tutor Authoring Tools through two main extensions. The first is a drag-and-drop expression builder that will allow students to construct problem solving steps and scientific arguments with a degree of flexibility that approaches that of paper-and-pencil problem solving, while allowing the computing system to monitor and scaffold the work through hints and feedback. The second is a domain browser from which students will drag and drop items for their expressions. This will allow the computing system to monitor and scaffold students as they develop strategies to solve a given problem. Thus, it is expected that this project will extend the range of problem solving activities that can be supported online in ways that research suggests will improve student learning. Scaffolding more open-ended, multistep problems will provide students practice with developing problem-solving strategies. Spanning a larger range of science practices will help students learn domain content while engaging in practices that better reflect those of chemists. Iterations in the design of the system will be informed by: comparison of student problem solving within the system to that of paper-and-pencil; examination of detailed log files on student use for appropriateness of the hints and feedback provided by the system; and evaluation of the degree to which the system can support problems that span science practices and include topics from across the domain. Analysis of log files gathered from the system will allow for design iterations that improve the activities, as well as provide fundamental insights into how students solve problems. The final year of the project will include a summative evaluation with diverse student populations. Both the online tools and the results of the evaluations will be broadly disseminated.