基于协作编程多模态数据的学习投入可视化及关联分析——理解行为、认知、社会、情感的交互关系及对学习的影响

Visualization and Correlation Analysis of Learning Engagement Based on Multimodal Data in Collaborative Programming:Understanding the Interplays Among Behavioral,Cognitive,Social,and Emotional Dimensions and Their Impacts on Learning

基于多模态数据可视化展示和关联来分析学生的协作学习投入水平,可有效对协作学习实施精准评价。为探究协作小组结构对学习投入的影响机理,研究梳理了多模态协作学习投入研究的分析框架,构建了涵盖学生行为、认知、社会、情感四要素的协作投入分析模型和编码方案,并基于这一编码框架,采集H高校66名学生参与小组协作编程学习的交互音视频、编码录屏、代码文本等多模态数据,再使用Nvivo进行编码和分析。研究进一步借助R语言可视化工具直观展示不同结构特征的协作小组在学习投入水平和学习成绩上的差异,并采用皮尔逊相关分析方法探究各协作学习投入维度的复杂内在联系及其对学习成绩的影响。研究表明:开展协作学习评价、实施精准干预以及促进知识建构提供科学依据。对这一问题的探讨可为多元视角探索协作学习投入提供参考。

Visualization and Correlation analyses of students’engagement based on multimodal data can effectively implement accurate evaluation of collaborative learning.To delve into the intricate mechanisms how group structures influence learning engagement,an analytical framework and coding scheme for multimodal collaborative learning engagement are constructed,covering students’behaviors,cognition,social interactions,and emotions.Utilizing this coding framework,multimodal data from 66 participants at H University,including interactive audio-video recordings,coding screen captures and code texts,are collected,coded and analyzed by using Nivivo.The diverse learning engagement and academic performance among collaborative groups with various structural features are visualized by using R tools.The complex intrinsic connections within each collaborative learning engagement dimension and their impacts on learning achievements are exploredby using Pearson correlation analysis.The research shows that the evaluation of collaborative learning,the implementation of precise intervention and the promotion of knowledge construction provide scientific basis.The research on this problem can provide reference for exploring collaborative learning input from multiple perspectives.