基于自我调节理论的在线学习行为分析及促进机制研究

Self-regulated learning (SRL) refers to the self-directive processes that occurred during learning, or the self-beliefs of one’s regulation. It is a consensus in educational psychology that self-regulation related constructs are critical predictors of academic achievement. It is expected that SRL will continue to play an important role in Internet-based learning. However, current studies of SRL in online learning have either heavily relied on questionnaires and follow the traditional paradigms of correlational study, or turned to data-driven methods but mainly described the qualitative patterns of online learning activities. Studies employing both approaches would benefit the field theoretically and practically..The proposed project intends to combine self-report measure of SRL and log data analysis of online learning activities. It is hypothesized that these methods could reflect the macro-level and micro-level of monitoring and control of individual’s learning process. These measures also can augment each other in explaining SRL processes that underlie online learning. A large educational data set from the online learning platform ALEKS will be used to build a quantitative model for SRL. This model could map learning activities on to the underlying SRL processes. Educational data-mining techniques will be utilized to build the models. The project will then incorporate small-scale experimental studies, as well as large-scale online surveys, to investigate the relationship between autonomy motivation, self-regulation, and learning outcomes in the context of online learning. This comprehensive model will be examined using structural equation modeling technique. A state-of-art optimization algorithm—Adaptive Design Optimization—will be implemented to help find personalized interventions to regulate online learning activities. The interventions can be tailored to the learner characteristics and ongoing learning tactics. Experimental studies will be conducted to examine whether the adaptive regulating algorithm will increase performance in online learning. The project not only will advance SRL theories in the context of Internet education, but also can showcase methodological enhancement.

在强调自主学习的在线教育环境中，学习者的自我调节学习行为和自我决定学习动机，对于学习效果具有直接的影响。现有研究或沿用理论驱动的心理测量范式，或偏重于数据驱动的学习行为模式挖掘，如果综合这两条路径，可以推动网络学习的相关理论研究，考察提高学生网络学习效果的可行途径，并结合技术手段，进行个性化干预。为此，本项目将首先运用教育数据挖掘技术，基于在线教育工具自动记录的学习行为数据，分析学生的自我调节学习行为，构建数学模型进行定量描述，并通过实验室实验对模型进行验证。其次，拟通过大样本实证研究，探索在线教育环境、学习动机、学习行为、学习效果之间的影响，建立整合的在线学习促进机制模型，进而尝试开发自适应学习算法，以学习者自主动机和自我调节学习方法为切入点进行干预，提高学习效果。本项目能通过教育心理学与信息科学的交叉，丰富和发展在线学习中的理论研究，改进网络学习心理的研究方法。

随着教育信息化的发展和在线教育需求的扩大，分析在线学习的认知过程，探讨学习行为和学习动机对在线学习效果的影响，既能从理论层面拓展学习科学研究，又具备优化在线教学设计的应用前景。本项目基于自我调节学习理论，以在线学习的多个要素作为切入点展开研究。首先，收集智能导学系统的日志数据，采用教育数据挖掘的研究范式，使用两层隐马尔可夫模型对序列学习行为进行建模；结合自我调节学习过程，对抓取的学习行为模式进行比较，以描述有效的学习行为模式；尝试从学习行为模型进行学业预测，获得了较高的准确率。其次，检验自我调节学习支架对促进学习行为、提高学习效果的作用和认知机制。选择在线学习中最为普及的视频学习作为研究情境，通过系列实验，重点比较了自我解释提示、总结提示、教学解释等常见自我调节学习支架的作用，并探索了多种学习者因素的调节效应。对于持续时间较短的视频学习，选择合适的支架、引发主动或建构学习活动，可以提高学习者的认知投入水平，增强对学习材料的记忆。引导学习者进行学习判断也能帮助学习者更好地监控学习过程。实验结果也为检验ICAP学习理论在视频学习的适用性提供了实证数据。再次，整合自我调节学习理论和动机的自我决定理论，通过检验学习动机、策略元认知知识、数学学习成绩之间的相互关系，拓展了自我决定动机的双路径模型，强调了自主动机对增强学习策略、提高学习效果的积极效应。进而采用启动实验的范式，通过操纵课程目标介绍和课程评价信息，观察动机相关因素对学习投入和学习效果的作用。本项目通过教育心理学、教育信息技术等学科领域的交叉，依托教育大数据分析、实验室学习实验、问卷法等多项研究范式，围绕在线教育的重点场景开展多方面研究，丰富了在线学习的学习理论研究，为配置合理的学习活动、完善在线教学设计、优化自主学习行为建立了研究基础。