Deep Learning in Blended Learning Environments: A Systematic Literature Review

Abstract

This study aimed to review research on students’ deep learning in blended learning environments. The study examined 29 peer-reviewed scholarly publications from the Web of Science using the PRISMA approach. Results show that effective deep learning practices in blended environments depend on several interconnected strategies including student-centered approaches like flipped classrooms, multimodal resources, experiential learning, strong teaching presence, and social interaction. Assessment strategies have evolved from recall tests to authentic assessments, reflection prompts, formative feedback, and AI-driven analytics. For implementation, educators should align online materials with face-to-face activities, provide appropriate scaffolding, embed metacognitive reflection, integrate technology thoughtfully, and offer personalized feedback. Approaches should be tailored to different learning contexts and disciplines. This systematic literature review identifies gaps in previous research and suggests directions for future studies including longitudinal investigations of deep learning trajectories and more robust combinations of AI-driven analytics with qualitative data.