Using Dynamic Item Response Theory to Model Ability Development in Online Learning

• In Event: AER Roundtable Session Friday 1:45 pm - Gold 4
  In Roundtable Session: Innovations in Item Response Theory for Modeling Student Performance and Growth

Abstract

Online learning increases the importance of understanding how engagement patterns affect learning outcomes, given students’ flexibility in interacting with course materials. This study uses Dynamic Item Response (DIR) models and latent class analysis to examine the relationship between engagement, latent traits, and learning outcomes in an online statistics course. Six student clusters emerged, with high performers showing consistent, spaced engagement, and low performers demonstrating cramming behaviors. Clusters also differed in math anxiety, memorization beliefs, prior performance, and coding experience. Findings underscore the need for adaptive strategies—such as spaced learning and anxiety management—to support diverse learners in digital environments. This research contributes to a nuanced understanding of engagement and ability development in digital learning contexts.

Authors

• Alice Xu, University of California - Los Angeles

• Zhihua Ma, Shenzhen University

• Guanyu Hu, Michigan State University

• Icy Yunyi Zhang, University of Wisconsin - Madison