Multimodal learning analytics infrastructure and adaptive digital assessment

Abstract This study empirically validates a theoretically grounded framework linking Multimodal Learning Analytics (MMLA) infrastructure components to adaptive digital assessment outcomes. Drawing on Universal Design for Learning (UDL) principles and learning analytics theory, we developed and tested a structural model examining how MMLA infrastructure is associated with educational practice. A cross-sectional survey was conducted with 476 educators and instructional designers across diverse educational contexts in India. The initially proposed 11-construct framework integrating UDL principles with MMLA infrastructure underwent systematic empirical validation using Partial Least Squares Structural Equation Modelling (PLS-SEM). A rigorous three-iteration refinement process reduced the initial 44-item instrument to a validated 24-item scale measuring seven constructs. The final validated model comprised MMLA infrastructure components (data quality, technology platforms, real-time analytics) and educational outcomes (adaptive intervention, formative feedback, inclusive assessment, SRL support). Real-time analytics emerged as the dominant predictor across all outcomes (β = 0.546, f 2 = 0.421, large effect on adaptive intervention), while MMLA technology demonstrated consistent moderate effects, and MMLA data showed limited direct association. Together, MMLA infrastructure explained 19–46% of the variance in adaptive assessment practices. Notably, UDL constructs failed to achieve acceptable psychometric properties despite theoretical grounding, revealing important insights about measurement challenges in practitioner self-reports. Findings suggest institutions should prioritise real-time analytics capabilities over raw data collection or basic technology infrastructure. The validated 24-item instrument enables practitioners to benchmark MMLA implementation and identify infrastructure gaps. This study provides the first empirically validated framework linking MMLA infrastructure with adaptive assessment outcomes, distinguishing data quality, technology platforms, and real-time analytics as separate constructs with differential effects. The UDL measurement challenges contribute to implementation science by revealing gaps between theoretical articulation and practical measurement.