Collaborative Research: Development of Concept Inventories for Computer Science

本计画透过为计算机科学入门科目开发三个概念量表，改善学生对计算机科学学习的评估。“力概念量表”是为了评估学生对牛顿物理学的学习情况而开发的，该量表以成功的“力概念量表”为模型，以一种能够可靠地比较不同大学课程的方式，测试学生对关键计算机科学概念的理解。伊利诺伊大学厄巴纳-香槟分校、罗斯-霍尔曼研究所和华盛顿大学之间的多机构合作伙伴关系，通过为不同的学生群体和不同的项目目标提供机会，有助于确保拟议工具的有效性。有了标准化的评估工具，计算机科学教育界可以对不同教学方法的有效性进行有意义的比较，极大地促进了计算机科学教育的研究。概念量表的设计是为了测试学生对困难概念的理解，通过强迫学生在正确答案和由学生常见误解构成的干扰物之间做出选择。作为关于哪些主题是困难的以及最常见的误解是什么的主要信息来源，学生们直接通过自省、讨论、访谈和“大声思考”来指导概念清单问题的发展。然后通过同行评议、定性分析和心理测量分析对这些问题进行改进和验证。一个由经验丰富的概念清单开发人员组成的咨询小组正在为项目进度提供建议和年度评估。

This project is improving the assessment of student learning in computer science by developing three concept inventories for introductory computer science subjects. Modeled after the successful Force Concept Inventory that was developed to assess student learning of Newtonian physics, the concept inventories are testing understanding of key computer science concepts in a manner that enables reliable comparisons between courses at different universities. This multi-institution partnership among the University of Illinois at Urbana-Champaign, Rose-Hulman Institute, and Washington University is helping to ensure the validity of the proposed instruments by providing access to diverse student populations and different program objectives. With a standardized assessment tool, the computer science education community can make meaningful comparisons of the effectiveness of different pedagogical approaches, greatly facilitating computer science education research.Concept inventories are designed to test student comprehension of difficult concepts by forcing a choice between the correct answer and distractors constructed from common student misconceptions. As the primary source of information about which topics are difficult and about what the most common misconceptions are, students are being directly engaged through introspections, discussions, interviews, and "think alouds" to direct the development of questions for the concept inventories. These questions are then being refined and validated through peer review, qualitative, and psychometric analyses. An advisory panel, comprised of experienced concept inventory developers, is advising and annually assessing project progress.