Abstract: Computer programming is a difficult course for many students. Prior works advocated for group learning pedagogies in pursuit of higher-level reasoning and conceptual understanding. However, the methodological gaps in existing implementations warrant further research. This study conducted a three-armed cluster-randomized controlled trial to comparatively evaluate the social and cognitive effects of group learning pedagogies in computer programming. Following an apprenticeship model, each group has a designated master: drivers in pair programming (PP), peer leaders in peer-led team learning (PLTL), and practitioners in practitioner-assisted group learning (PAGL). In all course deliverables, the PP group received the lowest mean scores. Meanwhile, no significant difference was found between the PLTL and PAGL groups. Except for psychological safety, social factors such as task cohesion, interdependence, and group potency were significantly different between the groups. Both PLTL and PAGL groups reported a significant increase in social factors after 14 weeks of intervention. These findings provide a rationale for educational leaders and teachers to formulate curricular plans that integrate PLTL and PAGL in computer programming education. Overall, this study contributes to the literature on group learning, expands the pedagogies in computer programming, and serves as additional empirical evidence on cognitive apprenticeship and sociocultural perspectives of learning.

Keywords: Computer Programming, Group Learning, Active Learning, Peer-Led Team Learning, Pair Programming, Practitioner-Assisted Group Learning