This study aims to analyze the differences in the effectiveness of the Discovery Learning, Video-Based Learning, and models on students' cognitive learning outcomes in food additives material. The research approach used a quantitative approach with a posttest-only control group design involving three eighth-grade classes at SMPN 27 Bandung as research subjects selected by purposive sampling. Data were collected through a validated learning outcome test instrument and then analyzed using the Kruskal–Wallis nonparametric test and the Mann–Whitney follow-up test because the normality assumption was not met. The results showed that the highest median learning outcome score was obtained in the Discovery Learning model (90.00), followed by Video-Based Learning (80.00) and Direct Instruction (70.00). The Kruskal-Wallis test showed a significance value of 0.029 (p < 0.05), indicating a significant difference in learning outcomes between the three learning models. Specifically, the follow-up test showed that the Discovery Learning model had a significantly different impact compared to Direct Instruction, while the Video-Based Learning model had a different impact compared to Direct Instruction. These findings confirm that knowledge construction through independent discovery activities is more effective in improving science concept understanding. Teachers are advised to integrate exploratory learning to minimize the gap in student learning outcomes.
Copyrights © 2025