A Predictive Learning Engineering Framework for Modeling Active Learning

Active Learning has been consistently linked to improved learning outcomes, persistence, and equity, yet it remains difficult to operationalize in digital and large-scale learning environments. While theory emphasizes internal cognitive processes such as sense-making, reflection, and strategic adaptation, learning-at-scale systems primarily capture observable behaviors. This disconnect has led to analytic approaches that equate engagement with behavioral presence, limiting the ability of learning analytics to represent meaningful learning activity. We present a predictive learning engineering framework that models Active Learning as an emergent, theory-aligned construct inferred from large-scale learner interaction data. Grounded in experiential, constructivist, sociocultural, and metacognitive theory, the framework conceptualizes Active Learning as patterns of generative, evaluative, and adaptive engagement. Using a human-centered, heuristics-guided construct specification process, predictive modeling, and iterative validation, the framework supports actionable and scalable active learning analytics in authentic learning-at-scale environments.