基于协作编程多模态数据的学习投入可视化及关联分析OA北大核心CSSCI

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 engage-ment,an analytical framework and coding scheme for multimodal collaborative learning engagement are constructed,covering stu-dents'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 fea-tures 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.