Abstract
Deep learning has become a central theme in contemporary educational reform, representing a critical indicator of learning quality. Peer feedback, as an interactive and learner-centered approach, has been shown to foster students' cognitive and meta-cognitive growth and holds significant potential for facilitating deep learning. This study constructed a peer assessment framework to promote deep learning in blended teaching and designed corresponding activities and implementation procedures. Drawing on CIMO-logic, the study examined how peer assessment triggered mechanisms such as personal engagement, seeking and providing relevant feedback, iterative exploration, and understanding one's own learning process. Data were collected through the SOLO taxonomy, rubrics, and questionnaires, complemented by discourse analysis of peer feedback comments. The linguistic analysis revealed that metalinguistic explanations and elicitation questions were associated with cognitive and ability development, while praise and politeness strategies primarily supported emotional engagement. The findings provide empirical evidence that peer assessment promotes deep learning across cognitive, ability, and emotional dimensions, and demonstrate that linguistic strategies in feedback are integral to how students process and internalize learning. This study provides theoretical insights into the occurrence of deep learning and offers practical implications for designing peer feedback activities to enhance learning quality in blended educational settings.

This publication has 53 references indexed in Scilit: