Designing and Evaluating Digital Planning Tools to Support Preparation for Peer Tutoring

Abstract

This research explores the design and evaluation of two digital planning tools that support peer tutoring preparation. Peer tutoring is widely recognised for its potential to enhance learning, yet its modest effects indicate room for improvement. For student tutors having neither deep domain nor pedagogical knowledge, better preparation is the only strategy for addressing these challenges. Despite evidence that preparing to teach can promote deep learning, the preparation phase remains empirically under-researched. The study was thus motivated by this gap, aiming to enhance individual tutor learning and collective knowledge construction through supporting the design of tutoring plans. This project employed a design-based research approach and conducted two iterative studies in science education. Guided by a conjecture map framework, the first study developed tuPlan, a dialogue-based planning tool that supported student tutors with structured prompts and visual graphs. An experimental comparison of 50 tuPlans with 50 slide-based plans found that tuPlans contained more tutoring elements, such as question-answer turns and varied guidance levels. Qualitative results revealed that tuPlans more frequently incorporated evaluative thinking and metacognitive knowledge than the slide decks. Nonetheless, there were also considerable differences in quality within the tuPlan group. Building on the findings and lessons learned, the second study developed tuMap, which integrated additional prompts such as goal settings and reflections on guidance and difficulties. An iterative comparison of 50 tuMaps with 50 tuPlans indicated that tuMap led to a higher frequency of metacognitive regulation. Large language model-assisted qualitative coding, grounded in Bloom’s taxonomy, also indicated a greater presence of higher-order cognitive processes and conceptual knowledge in tuMap plans. Moreover, tuMap plans appeared more coherent quality within the group. This study contributes to the learning sciences by illustrating how structured dialogue planning supports a transition in peer tutors from knowledge telling towards deeper processes of knowledge construction. These findings suggest implications for peer-assisted learning in higher education and the training of peer tutors. Given the limitation of focusing on a single content domain and exclusively on the planning phase, future research will expand the evaluation to cross-disciplinary contexts and incorporate additional knowledge technologies to advance the tool design for supporting one-to-one tutoring sessions.