【作 者】：
吴 斓， 董 艳， 郑娅峰， 余胜泉
【关 键 词 】：
基于计算机的协作问题解决； 能力评估； 评估框架； 评估工具； 评估方法
【栏 目】：
学习环境与资源
【中文摘要】：
协作问题解决能力是人工智能时代学习者必备的高阶能力之一，开展相关评估工作有助于培养学生的协作问题解决能力。文章对基于计算机的协作问题解决能力评估的典型研究与项目进行回顾，从评估框架、工具和方法三个方面对影响评估工作科学性和有效性的关键问题进行剖析，并针对性地提出发展建议，以期为后续评估工作改进提供参考：（1）遵循“理论突围—应用牵引—数据循证”的评估思路；（2）基于“聚焦个体—精进小组”的视角丰富评估信息；（3）坚持“适度协作支架—平衡同伴影响”的原则设计评估工具；（4）面向“智能采集—精准分类”实现丰富协作情境下多模态数据的自动化编码；（5）纳入“语境、时序、语义、社会文化等因素”自动分析会话数据质性特征；（6）结合“多维—交互”特征探索新的测量模型。
【英文摘要】：
Collaborative problem-solving ability is one of the necessary higher-order abilities for learners in the era of artificial intelligence, and conducting relevant assessments can help develop students' collaborative problem-solving abilities. This paper reviews the typical research and projects of computer-based collaborative problem-solving ability assessment, analyzes the key issues affecting the scientificity and effectiveness of the assessment from three aspects of assessment framework, tools and methods, and makes targeted recommendations for the improvement of subsequent assessments: (1) the assessment idea of "theory breakthrough—application traction—evidence-based data" needs to be followed; (2) Assessment information is to be enriched based on the perspective of "focusing on individuals-improving groups"; (3) The principle of "moderate collaborative scaffolding—balanced peer influence" is to be adhered to in designing assessment tools; (4) Automatic coding of multimodal data in rich collaborative contexts is to be realized through "intelligent collection—accurate classification"; (5) The qualitative characteristics of session data are to be automatically analyzed by incorporating "contextual, temporal, semantic and sociocultural factors"; (6) A new measurement model is explored with the feature of "multi-dimensional—interaction".