Heterogeneous Interaction Network Analysis (HINA): A New Learning Analytics Approach for Modelling, Analyzing, and Visualizing Complex Interactions in Learning Processes

A team of researchers from Monash University and the University of Edinburgh has published a new paper introducing Heterogeneous Interaction Network Analysis (HINA), a novel framework designed to model the complex, multi-faceted interactions in modern learning environments. The core innovation is transforming traditional sequential learning process data into Heterogeneous Interaction Networks (HINs), which can map connections between diverse entities like students, their specific behaviors, AI agents, and task designs. This moves beyond older methods that treat interactions as simple sequences or homogeneous relationships, allowing for a more nuanced, distributed view of how learning actually unfolds.

HINA integrates a suite of original analytical methods, including new summative measures and a non-parametric clustering technique, with established statistical testing and interactive visualization tools. The researchers demonstrated its utility in a case study on AI-mediated small-group collaboration. The analysis revealed clear, quantifiable differences in how students interact with AI agents compared to human peers, uncovering distinct engagement patterns and showing, for example, that specific learning behaviors like planning or asking questions are directed differently depending on the recipient. This provides a powerful new toolkit for learning scientists and educational technologists to dissect the complex dynamics of AI-augmented classrooms.

By providing a dedicated, multi-level analytical approach, HINA opens new avenues for understanding educational dynamics. It allows researchers and instructors to move from observing single data streams to visualizing how all elements of a learning ecosystem—students, tools, tasks, and AI—interact and co-influence each other. This paradigm shift is crucial for designing more effective, responsive educational technologies and pedagogical strategies, especially as AI becomes a more integrated participant in the learning process.