Mining Spatial Reasoning in Mathematics: Problem-Solving Patterns in a Geometry Digital Game Environment

• In Event: AERA Poster Session Friday 9:45am
  In Poster Session: STEM Education and Computational Learning Innovations

Abstract

Understanding how learners engage in complex digital mathematics tasks can help advance personalized learning. This study classified users’ problem-solving types in Shadowspect, a 3D spatial-reasoning game for geometry and math problem-solving, by their total task attempts, task successes, and time spent per successful attempt. We also examined how task difficulty and user performance related to sequential behavior patterns and contrasted strategies of effective versus less effective users on equally difficult tasks. Using gameplay data from 17 users, k-means clustering yielded three profiles: Efficient Achievers, Effortful Achievers, and Disengaged Users. Sequential pattern mining analysis revealed distinct strategic approaches across user groups and task difficulty levels. This work informs data-driven tools for targeted feedback, adaptive scaffolding, and enhanced spatial-reasoning outcomes.

Authors

• Shuchen Liu, University of Illinois at Urbana-Champaign

• Qingqing Zhou, University of Illinois at Urbana-Champaign

• Oktaria Kolnel, University of Illinois at Urbana-Champaign

• Jessica R. Gladstone, University of Illinois at Urbana-Champaign

• Jina Kang, University of Illinois at Urbana-Champaign