Detecting study regularity in trace data: A comparison of different operationalisations

Background: Spreading study sessions over time is considered beneficial for lasting learning. Trace data from digital learning platforms offer new ways of assessing study regularity over time and in naturalistic contexts. In addition, the fine-grained resolution of trace data allows for very different operationalisations of study regularity. As of now, however, it is unclear how study regularity should best be operationalised with trace data.Objectives: We developed and compared different operationalisations of study regularity in a naturalistic online learning setting, using objective trace data.Methods: 331 medical students preparing for a high-stakes exam on a digital learning platform provided fine-grained longitudinal data which allowed us to compare four different operationalisations of study regularity: 1) distributing studying over more vs less days, 2) the growth in questions answered before the exam (via growth curve modelling), 3) the shape of study trajectories (via time-series clustering), and 4) the predictability and repetitiveness in study trajectories (via time-series entropy). The resulting scores were used to predict students’ exam performance, controlling for the total amount studied.Results and Conclusion: All four approaches could capture differences in study regularity but only distributing study activities over more days positively predicted exam scores over and above the total amount of studying done by students. To boost learning outcomes, students should be supported to learn every day, if even for a little while.