Analyzing Interaction Patterns and Content Dynamics in an Online Mathematics Discussion Board

Research by: Jinnie Shin, Renu Balyan, Michelle P. Banawan, Tracy Arner, Walter L. Leite, & Danielle S. McNamara

EXECUTIVE SUMMARY

The paper “Analyzing Interaction Patterns and Content Dynamics in an Online Mathematics Discussion Board” employs Natural Language Processing (NLP) and Social Network Analysis (SNA) to comprehensively analyze interaction patterns within online algebra education. Drawing from a dataset of 170,000 posts across 14,000 threads involving 4,649 high school students in Florida, the study delves into both content-specific discussions (cognitive presence) and social interactions (social presence), offering insights into how these dynamics shape the learning experience.

Central to the methodology is the use of topic modeling and semi-supervised machine learning to uncover latent themes in the discussion posts by decomposing high-dimensional text data into a set of interpretable topics. This technique allowed the authors to distinguish between algebra-related content and non-content social interactions, providing a detailed analysis of cognitive engagement.

Beyond the linguistic analysis, the study applies Social Network Analysis (SNA) to quantify the relationships and engagement patterns within the discussion boards. Key SNA metrics such as degree centrality, which measures how actively participants engage, and closeness centrality, which evaluates how efficiently participants can spread information within the network, were used to explore the interaction dynamics. Flow hierarchy was also analyzed to assess the structure of information flow and the influence of facilitators on interaction networks. These techniques allowed the authors to map out complex student-to-student and student-to-facilitator interactions, highlighting the critical role of peer support in fostering academic success.

The results reveal a nearly equal split between content-driven discourse (52%) and social interactions (47%), demonstrating the dual purpose of online platforms in supporting both academic engagement and social interaction. High-performing students emerged as key facilitators in peer-driven discussions, playing a significant role in fostering content-specific support for their peers. This study showcases the potential of integrating AI Data Science techniques like NLP and SNA to analyze educational environments, providing actionable insights into designing more effective online learning platforms that promote both cognitive and social learning experiences.

To cite this article:  Shin, J., Balyan, R., Banawan, M. P., Arner, T., Leite, W. L., & McNamara, D. S. (2024). Analyzing interaction patterns and content dynamics in an online mathematics discussion board. Interactive Learning Environments, 1–24. https://doi.org/10.1080/10494820.2024.2392619

To access this article: https://doi.org/10.1080/10494820.2024.2392619

About the Journal

Interactive Learning Environments publishes articles on all aspects of the design and use of interactive learning environments in the broadest sense, encompassing environments that support individual learners through to environments that support collaboration amongst groups of learners or co-workers.

Relevant domains of application include education and training at all levels, life-long learning and knowledge sharing. Relevant topics for articles include: adaptive systems, learning theory, pedagogy and learning design, the electronically-enhanced classroom, computer mediated communications of all kinds, computer aided assessment, the design and use of virtual learning environments and learning management systems, facilitating organisational change, applying standards for courseware reuse, tracking, record keeping and system interoperability, the use of learning content management systems, including workflow design and publication to a range of media, and issues associated with scaling up delivery to large cohorts of students and trainees within the corporate, educational and other public sectors.

Journal ranking