AI-Supported Instruction and Student Engagement in Higher Education: Evidence from Survey and Regression Modelling
TL;DR
The findings suggest that AI-supported instruction provides additional explanatory value beyond engagement alone and that its effectiveness depends on coherent pedagogical integration.
Abstract
Artificial intelligence (AI) is increasingly being integrated into higher education, yet its educational value depends not only on access to technological tools but also on how these tools are pedagogically embedded within teaching and learning. Although prior studies have reported positive associations between educational technologies and learning outcomes, fewer studies have examined whether AI tool usage explains perceived learning effectiveness beyond established engagement dimensions and instructional integration quality. This study investigates the relationships among AI tool usage, behavioural, cognitive and emotional engagement, instructional integration quality, and perceived learning effectiveness in higher education. Using a quantitative cross-sectional survey design, data were collected from 214 participants, including faculty members and undergraduate students, through a structured Likert-scale questionnaire. Descriptive, correlational, regression and group-comparison analyses were conducted to examine associative and predictive patterns. The findings indicate that AI tool usage was positively associated with all engagement dimensions and with perceived learning effectiveness. In the regression model, AI tool usage remained a significant independent predictor of perceived learning effectiveness even after controlling for engagement dimensions and instructional integration quality, while instructional integration quality also showed a significant positive effect. In addition, AI users reported significantly higher levels of engagement and learning effectiveness than non-users, whereas no significant differences were observed between faculty and students in engagement dimensions or learning outcomes. These findings suggest that AI-supported instruction provides additional explanatory value beyond engagement alone and that its effectiveness depends on coherent pedagogical integration. The study contributes to the AI-in-education literature by offering empirical evidence on the role of AI-supported instruction within an engagement-based framework and by highlighting implications for instructional design, faculty development, and institutional AI integration in higher education.