STEM Berbasis Proyek dan Kesenjangan Infrastruktur Digital: Studi Komparatif Implementasi di Sekolah Perkotaan dan Sekolah Pinggiran

Abstract

This article investigates how digital infrastructure mediates the effectiveness of project-based STEM learning in secondary science classrooms. Using a comparative case study, it contrasts two implementation contexts frequently found in Indonesia: an urban school with stable connectivity and rich device access, and a peri-urban school with limited connectivity and intermittent access. The study draws on public school quality reports, policy documents, and recent empirical studies on STEM project learning. Analysis focuses on how infrastructure conditions shape task design, collaboration, modelling, and feedback. Findings show that project-based STEM can improve problem solving and conceptual integration in both settings, but the mechanism differs. In better connected contexts, digital tools support iterative modelling and distributed collaboration. In constrained contexts, teacher-designed low-tech scaffolds and community-based resources become more decisive than platform sophistication. The article argues that the effectiveness of digital STEM learning depends less on technology abundance than on pedagogical redesign that aligns tools, tasks, and local constraints. Its novelty lies in reframing infrastructure not as a binary deficit variable but as a condition that interacts with teacher design and assessment logic. The study contributes to science education research by specifying how equity and contextual adaptation shape STEM outcomes beyond generic claims about educational technology.