Optimizing Class Timing: AI Driven Group Engagement’s Role in Academic Success
, M M Munshi, Sanjay V Hanji, Chakradhar PabbaThis study investigates the relationship between time of class and the academic performance of Master of Business Administration (MBA) students with ‘group engagement’ serving as the moderator. Notably, ‘group engagement’ is measured using a novel computer vision-based deep learning approach.
Generally, the first year of MBA programs is a critical phase for students, marked by academic and personal growth challenges. The timing of MBA classes, particularly morning sessions, can disrupt students’ circadian rhythms, leading to decreased engagement and academic performance. Existing literature highlights the potential benefits of active learning methods, such as blended learning and collaborative learning, in improving individual student engagement. However, there is a gap in understanding the impact of these methods on group engagement and how this, in turn, influences academic performance.
We collated video-recorded data from 54 first-semester MBA students when they were attending their morning and afternoon classes. Notably, we adopted blended and collaborative learning methods in the morning classes to check their impact. The study variables included time of class, group engagement, and academic performance. While we measured academic performance through proctored exams, group engagement was estimated using a vision-based system, whereby we analyzed facial expressions from the recorded videos, employing a Convolutional Neural Network (CNN) model. We trained the CNN model to measure group engagement by categorizing specific emotions (sleepy, bored, yawning, frustrated, confused, and focused). We used these emotions to assess group engagement levels (i.e., low, medium, and high).
This study establishes a definite link between morning classes with blended and collaborative learning methods, resulting in improved academic performance. Besides, through group engagement moderation, we get crucial insights into how high group engagement effectively enhances academic performance. Based on the findings, we propose strategies for educators to optimize their teaching methods and foster a conducive learning environment by leveraging insights from students’ affective states.
Broadly, the results indicate that students performed better in morning classes in which both blended and collaborative learning methods were used. The average marks in morning classes increased to 18.74 compared to 15.39 in afternoon classes. There was a high level of group engagement in the mornings, significantly impacting the relationship between class time and academic performance. However, switching from morning to afternoon classes decreased the effect (academic performance) from 0.03 to -5.42. This shows that the use of blended and collaborative learning methods and the presence of high group engagement in morning classes are essential for better educational outcomes.
We recommend integrating active learning methods, such as blended and collaborative learning, into morning MBA classes to optimize academic performance. Our results show that implementing technology-driven group engagement measurement tools enhances real-time insight into emotional states. Educators can use this information to tailor teaching approaches and foster a positive learning environment.
Additional research should explore the longitudinal impact of AI-based engagement measurement. Studies could also investigate the scalability and applicability of group-focused strategies across diverse educational settings.
This study highlights the need to improve teaching methods and group engagement in MBA education. It offers tools for active learning and AI-based engagement tracking. The findings would help both educators and policymakers create better learning experiences.
Future research should include more diverse samples to enhance external validity. Studies can explore various learning methods across different cultural contexts. Researchers can develop hybrid systems integrating physiological sensors with computer vision to provide more comprehensive results.
Back