作者：Hou, C., Zhu, G., Yang, Y.* (杨玉芹), & Tang, S.

出版刊物：Computers & Education

出版时间：2025年

内容摘要：

Knowledge Building (KB) is a pedagogical approach emphasizing students' collective responsibility to continuously improve their community knowledge. During KB activities, various emotions may arise due to students’ diverse ideas, theory-building, as well as cognitive equilibrium and disequilibrium. Depending on the different development levels of the inquiry threads, the emotions may differ. An inquiry thread is a series of notes that address the same problem or topic. Despite increasing recognition of the importance of emotions in KB, or more generally, in computer-supported collaborative learning (CSCL), there is a lack of empirical studies systematically examining the dynamics of these emotions, particularly how they vary across different types of collaborative inquiry threads. Addressing this gap, this study analyzed 148 threads containing 6,240 notes from a learning science course at a public university over three years. Through clustering analysis, we identified productive and improvable KB inquiry threads recorded in the Knowledge Forum (KF) with productive threads, which are characterized by deeper cognitive efforts and more sustained discussion than improvable threads. Integrating three levels of analysis, namely frequency, lag-sequential analysis, and sequential pattern mining, we aim to comprehensively capture students’ emotional dynamics across different kinds of inquiry threads. Our study identified distinct emotional dynamics and constructed emotion evolution models for productive and improvable inquiry threads. Productive threads frequently exhibited transitions between negative emotions such as confusion, anxiety, and frustration, indicating their deeper cognitive engagement and suggesting that while students experienced challenges in reaching a consensus, they remained cognitively engaged in the inquiry. Conversely, improvable threads were characterized by sequences involving positive emotions like joy and curiosity, often appearing at the beginning of discussions or in threads that lacked depth, indicating that while initial interest was present, these discussions failed to evolve into meaningful inquiries. Emotional transitions from activating emotions (e.g., joy) to deactivating emotions (e.g., boredom) in improvable threads further suggest a disengagement from the discussion. These findings highlight the intricate emotional dynamics during learning activities and provide valuable insights for future research focused on enhancing productive discussions in CSCL environments through effective emotional regulation strategies.

知识建构作为一种强调学生集体责任以持续优化共同体知识的教学范式，在其活动过程中常因观点碰撞、理论构建及认知平衡状态的打破而引发多样情绪。受探究主题发展水平差异的影响，不同讨论线程中的情绪表现往往呈现系统性差异。尽管情绪在知识建构乃至计算机支持的协作学习中的重要性日益受到关注，但当前仍缺乏对不同类型协作探究线程中情绪动态的系统性实证研究。为填补这一空白，本研究对某公立大学一门持续三年的学习科学课程数据展开分析，从知识论坛中提取包含6,240条笔记的148条讨论线程，通过聚类分析识别出生产性与待改进型两类知识建构探究线程，其中生产性线程表现出更深入的认知投入与更持续的讨论特征。研究融合频次分析、滞后序列分析与序列模式挖掘三个分析层次，系统揭示不同类型探究线程中的情绪动态特征。研究发现，生产性线程频繁出现困惑、焦虑、沮丧等消极情绪间的相互转化，体现出学生在共识形成过程中保持持续认知投入的特征；待改进型线程则多呈现以喜悦、好奇等积极情绪开端的序列模式，且常伴随由激活情绪向倦怠等抑制情绪的转变，反映此类讨论虽具初始兴趣却未能发展为深度探究。这些发现揭示了学习活动中复杂的情绪演化机制，为未来在CSCL环境中通过情绪调节策略促进高质量讨论提供了重要启示。