EXP: Data-Driven Support for Novice Programmers

该项目的研究人员将研究完全数据驱动的系统，为学习者提供可扩展和个性化的支持。开放式的，媒体丰富的可视化编程环境，如Scratch和Snap，代表了吸引和激励学生学习编程的下一代流派。然而，解决开放式编程问题对于数据驱动支持来说是一个特别具有挑战性的领域，因为潜在的解决方案空间非常大，而且非常有创意的装饰使它们具有吸引力。该项目将从数据中获得的智能学生支持集成到Snap编程环境中的多媒体开放式问题解决中。更具体地说，研究人员将构建和评估数据驱动的方法，以提供提示，工作示例和积极主动的积极反馈，而学生则致力于多媒体开放式程序，可以吸引和激励不同的学生。研究人员将在一个迭代优化的周期中进行6项研究，以将数据驱动的提示，示例和开放式问题的反馈集成到Snap编程环境中。这使得研究媒体丰富的环境与个性化支持的综合影响成为可能，为适应性支持而创建的结构将为新手如何学习编程提供深入了解。Snap是新手编程环境的代表，与流行的编程语言Scratch几乎相同。更重要的是，所使用的技术是基于计算机程序的通用属性，可以应用于许多编程环境和问题，解决的研究问题具有跨问题解决领域的普遍适用性。

The researchers in this project will study fully data-driven systems to provide both scalable and individualized support for learners. Open-ended, media-rich visual programming environments such as Scratch and Snap represent the next-generation genre for engaging and inspiring students to learn programming. However, solving open-ended programming problems is a particularly challenging area for data-driven support due to the extremely large potential solution spaces and the very creative embellishments that make them attractive. This project integrates intelligent student support that is derived from data into the media-rich, open-ended problem solving in the Snap programming environment.More specifically, the researchers will build and evaluate data-driven methods to provide hints, worked examples, and proactive positive feedback while students work on media-rich, open-ended programs that can interest and motivate diverse students. The researchers will conduct 6 studies in a cycle of iterative refinement to integrate data-driven hints, examples, and feedback for open-ended problems into the Snap programming environment. This enables the study of the combined impact of a media-rich environment with individualized support, and the structures that are created for adaptive support will provide insight into how novices learn to program. Snap is representative of novice programming environments, and is almost identical to Scratch, a popular programming language. More importantly, the techniques used are based on generic properties of computer programs and could be applied to many programming environments and problems, and the research questions addressed have general applicability across problem solving domains.