From Variables to States to Trajectories (VaSSTra): A Method for Modelling the Longitudinal Dynamics of Learning and Behaviour

Abstract

Research in learning analytics needs longitudinal studies that explore the learner’s behaviour, disposition, and learning practices across time, a gap this article aims to bridge. We present VaSSTra: an innovative method for the longitudinal analysis of educational data that can be applied at different time scales (e.g., days, weeks, or courses), and allows the study of different aspects of learning as well as the factors that explain how such aspects evolve over time. Our method combines life-events methods with sequence analysis and consists of three steps: (1) converting variables to states (where variables are grouped into homogenous states); (2) from states to sequences (where the states are used to construct sequences across time), and (3) from sequences to trajectories (where similar sequences are grouped in trajectories). VaSSTra enables us to map the longitudinal unfolding of events while taking advantage of the wealth of life-events methods to visualize, model and describe the temporal dynamics of longitudinal activities. We demonstrate the method with a practical case study example. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.