CAREER: Cognitive Diagnosis in E-Learning: A Nonparametric Approach for Computerized Adaptive Testing

这是一个教师早期职业发展计划（Career）项目。CAREER计划是一项国家科学基金会范围内的活动，提供最负盛名的奖项，以支持通过杰出的研究，优秀的教育和教育与研究的整合来体现教师学者角色的初级教师。研究者将开发一种认知诊断计算机自适应测试（CD-CAT）系统，以支持大学水平的入门统计学学习。这项工作代表了教育测试的重大进步，提供个性化的评估，从而根据学生的熟练程度概况生成诊断信息，而不是抽象的总体测试分数。该研究特别处理了一些困难，如对大样本的要求，底层模型识别的不发达，以及可能限制基于模型的CD-CAT对教育从业者的可用性的技术复杂性；尤其是在小型教学环境中的教师和学生。鉴于这些考虑，该项目首先提出了一种创新的一般非参数分类（GNPC）替代方案，这是非参数分类（NPC）方法的一般版本。GNPC方法不仅可用于分析大型数据，也可用于分析小规模数据，例如符合广泛认知诊断模型的基于课程的数据，这一优势使那些最需要认知诊断分析的人能够获得认知诊断分析。这种随时可用的启发式方法的发展随后将该领域进一步推向非参数CD-CAT在教育微环境中的常规使用。所提出的算法将应用于在线统计学课程，在这些课程中，高性能的形成性评估被用来对学生的表现提供有用和及时的反馈。拟议的调查路线通过开发在线CD-CAT系统将研究和教育结合起来，以支持统计学的学习。教育活动既包括研究生水平的教学，也包括教师和研究人员的专业发展活动。

This is a Faculty Early Career Development Program (CAREER) project. The CAREER program is a National Science Foundation-wide activity that offers the most prestigious awards in support of junior faculty who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research. The investigator will develop a Cognitive Diagnostic Computerized Adaptive Testing (CD-CAT) system to support the learning of introductory statistics at the university level. The work represents a significant advance in educational testing to provide personalized assessments which in turn generate diagnostic information based on students' proficiency profiles, as opposed to an abstract overall test score. The research deals specifically with difficulties such as the requirement of large samples, underdevelopment in identification of the underlying model, and possible technical complexity limiting the availability of the model-based CD-CAT to educational practitioners; especially teachers and their students in smaller instructional settings. In light of these considerations, the project first proposes an innovative general nonparametric classification (GNPC) alternative, which is a general version of the nonparametric classification (NPC) method. The GNPC method can be used to analyze not only large- but also small-scale data, such as course-based data, conforming to a broad range of cognitive diagnostic models, an advantage that makes cognitive diagnosis analysis accessible to those who are in its highest need. The development of this ready-to-use heuristic method then moves the field further toward the routine use of the nonparametric CD-CAT within educational micro-environments. The proposed algorithms will be applied to online statistics courses, where high-performance formative assessments are employed to provide useful and timely feedback on how well students perform. The proposed line of inquiry integrates research and education through the development of an online CD-CAT system to support the learning of statistics. The education activities include both teaching at the graduate level and professional development activities for teachers and researchers.