Exploring the effect of parental involvement on student engagement and academic performance using process data from learning management system

Parental involvement is essential to children’s learning engagement activities and academic performance. Much research revolves around the impact of parental involvement on students’ academic performance or the relationship between student engagement and grades. However, few studies have used process data to examine the relationship between parental involvement, student learning engagement, and grades. This study classified students into three groups using hierarchical clustering (Inactive Engagers, Selective Engagers, and Proactive Engagers) based on four online engagement behaviors. Each behavioral feature, parental school satisfaction, parental answering of surveys, and student grade levels were different across clusters. Different influencing factors were found for each cluster, the student absent days feature was the most critical factor influencing student grades. Artificial neural networks (ANN) outperformed other classifiers in the inactive engagers group. For the selective engagers group, the performance of decision trees (DT) and ANN were similar, and both were better than naïve Bayes (NB). For the proactive engagers group, the DT model achieved the best precision and ANN achieved the best F1-Measure. These results provide insight into the effects of parental involvement on both student online engagement behaviors and academic performance and will enable parents and teachers to develop strategies to support students’ academic success.