Strengthening Economic Learning Based on Deep Learning to Improve Financial Literacy of SMAN Tondano Students

Abstract

Digital financial services expand adolescents' transactional convenience while exposing them to credit, fraud, privacy, and impulsive-consumption risks. This exploratory qualitative study examined how students in three public senior high schools in Tondano, Indonesia, organize economic concepts when reasoning about everyday financial decisions and what these patterns imply for deep learning-oriented economics instruction. Thirty students completed open-ended problem-based tasks used to map variation in conceptual structure; 15 were then purposively selected for semi-structured interviews and think-aloud activities. Data were organized in NVivo and analyzed thematically, with the Structure of Observed Learning Outcomes (SOLO) taxonomy used to interpret structural complexity and thematic-network mapping used to represent relationships among analytic domains. The findings indicate that students could identify saving, consumption, investment, and risk concepts but often treated them as disconnected elements rather than integrated decision rules. Operational familiarity with QRIS, e-wallets, and paylater did not consistently include recognition of fees, debt, fraud, or digital-security risks. Peer and promotional contexts shaped spending considerations, while experiential reflection and self-regulatory practices appeared to support more connected reasoning. The study does not test a Deep Learning intervention; instead, it provides a diagnostic basis for designing economics tasks that require students to relate concepts, justify trade-offs, evaluate digital risks, and transfer principles to new situations.