Search

Learning analytics (LA) provides tools to analyze historical data with the goal of better understanding how curricular structures and features have impacted student learning. Forward-looking curriculum design, however, frequently involves a degree of uncertainty. Historical data may be unavailable, a contemplated modification to curriculum may be unprecedented, or we may lack data regarding particular learner populations. To address this need, we propose using curriculum modelling and learner simulation (CMLS), which relies on well-established modelling theory and software to represent an existing or contemplated curriculum. The resulting model incorporates relevant research-based principles of learning to individually simulate learners and estimate their learning achievement as they move through the modelled curriculum. Results reflect both features of the curriculum (e.g., time allocated to different learning outcomes), learner profiles, and the natural variability of learners. We describe simulations with two versions of a college-level curriculum, explaining how results from simulations informed curriculum redesign work. We conclude with commentary on generalizing these methods, noting both theoretical and practical benefits of CMLS for curriculum (re)design.