Artificial Intelligence-Driven Adaptive Testing: A Psychometric Approach to Personalized Learning in Computer Science Education

Abstract

The integration of Artificial Intelligence (AI) into educational assessment has revolutionized how learners are evaluated and supported. This study, Artificial Intelligence-Driven Adaptive Testing: A Psychometric Approach to Personalized Learning in Computer Science Education, investigates the fusion of psychometric modeling and AI-based adaptive testing systems to enhance individualized learning pathways for students in computer science. The research adopts a hybrid framework combining Item Response Theory (IRT) and reinforcement learning algorithms to dynamically adjust question difficulty based on learner performance in real time. A dataset of undergraduate computer science learners was used to develop and validate the adaptive system through parameters such as accuracy, response time, and knowledge progression. The psychometric evaluation demonstrated high reliability and discriminant validity, while the AI model optimized test adaptivity and reduced assessment bias. Findings indicate that AI-driven adaptive testing significantly improves learning engagement, reduces cognitive overload, and enhances conceptual retention compared to static assessment methods. The study contributes to the growing discourse on intelligent educational systems, presenting a scalable, data-driven psychometric model that fosters personalized, equitable, and efficient learning environments in computer science education.