Pedagogical design of adaptive courses for the development of high school students’ metacompetencies based on learning analytics and artificial intelligence

Abstract

The article is devoted to the justification and empirical testing of a model of pedagogical design of adaptive courses aimed at developing metacompetencies (self-regulation, critical thinking, creativity, learning motivation) of high school students in the context of digital transformation of schooling. The purpose of the study was to develop and validate a digital course architecture that uses learning analytics and artificial intelligence tools to construct personalized learning trajectories based on digital trace data. A longitudinal quasi-experiment was conducted in seven schools involving 482 students aged 15-17, divided into control and experimental groups; instruction in the experimental group was carried out in an LMS with an AI module implementing dynamic adaptation of task difficulty and content. The methods included standardized metacompetency questionnaires, critical thinking tests, motivation scales, machine learning (k-means clustering, recurrent neural networks for trajectory prediction), NLP methods for essay analysis, as well as ANOVA, correlation, and regression analysis. The results demonstrated a statistically significant superior increase in integral indicators of critical thinking, self organization, and creativity in the experimental group compared to the control group, a reduction in the variance of final outcomes, and the leveling of educational achievements among lower-performing students. It was shown that adaptive trajectories ensure a higher knowledge acquisition coefficient with equal or reduced time expenditure, and the cognitive comfort index of participants in adaptive learning increases substantially throughout all stages of the experiment. It is discussed that pedagogically meaningful integration of learning analytics and AI contributes to reducing unproductive cognitive load, transforming errors into a resource for reflection, and fostering sustainable metacompetencies, while the effectiveness of the system is determined by the quality of pedagogical scenarios and the balance between algorithmic regulation and student agency. The prospects for scaling the model and the ethical-methodological limitations of its implementation in school practice are outlined.