Exploring Expert and Novice Graphical Communication Through Digital Sketching

通过绘图等图形手段进行交流的能力是工程教育和实践的核心，但很少有人知道学生如何，为什么以及何时使用工程草图或如何学习使用它们。该项目研究新手和专家如何生成和响应他们自己的草图，并建立在现有的数字草图识别技术。学生的表达流畅性将通过探索专家和新手在解决问题时使用和制作工程草图的异同来检验，重点关注两个领域的草图：数字逻辑和桁架。这些领域使用互补的草图风格，因为数字逻辑草图在视觉上与现实世界的实现不同，而桁架依赖于现实世界对象的复制品。互补领域的使用将提供更丰富的观察结果，可以为理论和教学工具的未来发展提供信息。对学生素描的教育数据挖掘将提供对素描行为（例如，笔速，压力）的新观察，这些行为可能对人类研究人员来说是模糊的或不可见的。当学生和研究生助教解决数字逻辑和桁架的分析和设计问题时，将进行有声思考访谈。这些访谈的数据将用定性分析技术进行分析，对新手和专家的行为进行分类。此外，将对数字笔输入进行初步数据挖掘，以确定可以通过算法检测到的显着变量。这个项目的智力价值是建立在对专家-新手差异研究的丰富传统的基础上的。这个项目是由先前关于新手和专家如何在他们的推理中理解和使用静态图像的研究特别通知的。将教育数据挖掘与素描的定性分析结合起来，将为混合方法研究的新形式的发展提供信息，推动教育研究方法的发展。这个项目的更广泛的影响将通过三个不同的方面来看：i)为工程教育提供更丰富的资格认证的新途径，ii)通过数据挖掘增强和最大化教学方法及其工具的有效性，以及iii)产生大量的草图语料库，同时建立一个可以使用该数据库的开发人员/用户社区。

The ability to communicate through graphical means such as sketching is central to engineering education and practice, yet little is known about how, why, and when students use engineering sketches or how they learn to use them. This project studies how novices and experts produce and respond to their own sketches, and builds on existing techniques for digital sketch recognition. Students' representational fluency will be examined by exploring the similarities and distinctions between experts and novices in their use and production of engineering sketches during problem solving by focusing on sketching in two domains: digital logic and trusses. These domains use complementary styles of sketches as digital logic sketches are visually distinct from real-world implementations, and trusses rely on facsimiles of real-world objects. The use of complementary domains will provide richer observations that can inform the future development of theory and pedagogical tools. Educational data mining of students' sketches will provide new observations on sketching behaviors (e.g., pen speed, pressure) that might be obscured or invisible to human researchers. Think-aloud interviews will be conducted as students and graduate teaching assistants solve analysis and design problems for digital logic and trusses. Data from these interviews will be analyzed with qualitative analysis techniques to categorize behaviors of novices and experts. Additionally, preliminary data mining will be performed on digital pen inputs to identify salient variables that can be detected algorithmically. The intellectual merit of this project is established by its basis on a rich tradition in the study of expert-novice differences. This project is particularly informed by prior research on how novices and experts comprehend and use static images in their reasoning. Aligning educational data mining with qualitative analysis of sketching will inform the development of a new form of mixed methods research, advancing the state of the art in educational research methods. The broader impacts of this project will be seen through three different facets: i) providing new avenues for richer credentialing in engineering education, ii) enhancing and maximizing the effectiveness of pedagogies and their tools through data mining, and iii) producing a large corpus of sketching, simultaneously building a developer/user community that can use that database.