认知诊断理论在数学教育评价中的应用

摘要

质性地比较论证了基于认知诊断理论的数学教育评价工具开发的可行性。通过分析实际测试数据，比较几种不同认知诊断模型中的参数估计方法实际应用于分析数学评价测验的可能性。研究发现，认知属性概念可以帮助研究者和实践者分析影响学生解答数学题目背后的认知结构。人工神经网络模型能够充分利用理论设计的认知模型，克服测验题目有限、题型多样、认知属性差异大等不易分析的困难，较好评价学生的数学学业成就，为后续教学提供诊断性信息，达到了诊断性测验的目的。%Qualitatively illustrating the feasibility of developing mathematics education assessment tool based on Cognitive Diagnose Model. Several different parameter estimation model based on survey data were compared. The results indicated that the concept of cognitive attributes could help both the researchers and the practitioners to analyze the cognitive structure behind students’ solutions of mathematical problems. Artificial neural networks model could fully make use of the cognitive model to solve the problems of limited amount of rubrics as well as diversity of different cognitive attributes, and assess students’ mathematical achievement with high validity, which provide diagnosestic information for future instruction.