Measuring Learner Performance to Support Educational Management Using Waiting-Time Distributions

This paper reviews and synthesizes existing frameworks for developing adaptive assessments based on waiting-time distributions. By integrating the time required to achieve a predefined sequence of correct responses into the model, the reviewed approaches can be used to infer in real time each examinee’s latent ability, cognitive load, and engagement level, dynamically adjusting item difficulty and order. The reviewed approaches offer concrete benefits for educational administration: the resulting test data deliver timely, trustworthy insights that support the design of targeted learning interventions, more efficient resource allocation, and evidence-based decision-making in both academic institutions and corporate training programs. Simulation studies reported in the literature indicate that the adaptive design substantially reduces expected test length without compromising statistical validity, providing an effective tool for modern adaptive testing systems and the administrative processes that rely on them.