学习行为分析支持的差异化教学模式研究——以概率论与数理统计课程为例

Research on the Model of Differentiated Instruction Supported by Learning Behavior Analysis——Taking the Course of Probability and Statistics as an Example

差异化教学尊重学习者个体内在认知逻辑,通过挖掘并诊断学习者的个体潜质和优势,提供适合的学习环境和学习干预。然而,现有差异化教学存在时效性差、维度单一等问题,停留在浅层次干预阶段。数据驱动的学习行为分析为设计和实施高阶差异化干预策略提供了新的途径。本研究依据概率论与数理统计课程创新班30名注册大学生在超星学习通平台上的学习日志数据,应用相关分析、聚类分析和滞后序列分析对在线学习行为特征和个性化干预进行实证研究。根据观察结果发现,定义的10种学习行为所形成的62个行为序列中,7个行为序列与期末成绩显著相关。以62个行为序列的频数作为分类变量,应用k-means算法将30名大学生样本根据其学习行为模式划分为3种类别。应用滞后序列分析法发现和识别3个类别大学生的学习风格:混合/无序型,孤立/分解型,整体/规划型,提出差异化干预框架和具体的干预策略。本研究期望构建技术增强的差异化教学模式,为建设和优化以大学生为中心的个性化课堂提供新范式。

Differentiated instruction respects learners'internal cognitive logic,and provides suitable learning environments and learning interventions by tapping and diagnosing learners'individual potentials and strengths.However,the existing differentiated instruction has problems such as poor timeliness and single dimension,which remains at the shallow level of intervention.Data driven learning behavior analysis provides a new way to design and implement high-level differentiated intervention strategies.Based on the learning log data of 30 college students registered in the innovation class of Probability and Statistics on the Chaoxing learning platform,an empirical study on online learning behavior characteristics and personalized intervention was conducted using correlation analysis,clustering analysis,and lag sequential analysis.According to the observation results,out of the 62 behavioral sequences formed by the 10 defined learning behaviors,7behavioral sequences were significantly correlated to the final test score.Using the frequencies of 62 behavioral sequences as classification variables,30 students were divided into three categories based on their learning behavior patterns using kmeans algorithm.Three learning styles were discovered and identified using the lag sequence analysis method,including mixed/disordered pattern,isolated/decomposed pattern,and holistic/planned pattern,and a differentiated intervention framework as well as specific intervention strategies were proposed.This study is helpful for constructing a technologyenhanced differentiated instruction model and providing a new paradigm for building and optimizing personalized studentcentered classrooms.