Pedagogical Considerations for Designing Automated Grouping Systems: the Case of the Parabola

Abstract

This article queries how learning analytics systems can support content-specific group formation to develop students’ thinking about a specific mathematical concept. Automated group formation requires identifying personal characteristics, designing tasks to probe students’ perceptions, and grouping them to increase individual learning chances. Designers of automated group formation recommendation modules (GFRMs) rarely consider content-specific objectives. We draw on theories on conceptual learning in mathematics and dialogic thinking to emphasize the role of a dialogic gap between students’ voices to enhance individual learning. In an experiment, fifty 8th and 9th grade students solved three mathematical tasks in a pre-intervention-post-set-up: individually, then in dyads, and then individually again. We used a learning analytics system to collect fine-grained content-specific data on students’ responses based on four pre-defined aspects of the parabola concept. We compared students’ answers with those of their peers in order to identify interpersonal relations. The experiment results indicate that students’ thinking about the parabola concept was the most successfully developed when every group member had a different perception of this concept. We illustrate the learning trajectories of four students and elaborate on the learning sequence of one of these students in particular. This article suggests that the centrality of a dialogic gap in developing personal learning is probably content independent. We thus call for software engineers to think about GFRMs that can support content-specific learning and instruction.