Mining Students’ Engagement Pattern in Summer Vacation Assignment

Abstract

Learning Analytics (LA) is an emergent field which aims at a better understanding of students and providing intelligence to learners, teachers, and administrators using learning log data. Although the use of technology in class is increasing in the K-12 sector as well as territory education, cases of effective implementation of LA in secondary schools were rarely reported, especially in Japan. In this paper, we offer an example where LA is implemented at a junior-high Math class in Japan. We introduce our LA platform, LEAF - LMS and e-book integrated learning analytics dashboard - and its usage during summer vacation period in the target class. We analyzed 121 students’ question answering logs and their exam performance after the vacation by K-means clustering method. As a result, we found that students’ progress patterns were able to be categorized as four types: early engagement, late engagement, high engagement, and low engagement and the early and high engagement group got significantly higher scores than the low engagement group. It implies the importance of the engagement at the beginning of the vacation. Moreover, by comparing the previous studies in MOOCs, we concluded that self-regulation skills are an important factor for student success in a long vacation period, too. Finally, we introduce a monitoring tool which aims to detect and send messages to at-risk students at an early stage in the next summer vacation period. Our case will become the first model case of how to implement LA in secondary school in Japan.