<![CDATA[Students' low conceptual understanding remains a fundamental problem in science education at the junior high school level, mainly because teacher-centered learning causes students to merely memorize without understanding the concepts. This condition requires the application of a learning model that can foster deep conceptual understanding and reflective awareness through discovery and direct experience. This study used a quantitative approach with a One Group Pretest–Posttest design to measure the effectiveness of applying the Discovery Learning model with a Deep Learning approach. The research subjects consisted of 32 seventh-grade students at SMP Negeri 10 Kupang in the odd semester of the 2025/2026 academic year. Data were collected through learning outcome tests (pretest and posttest), observations of learning implementation, and student reflection sheets. Data analysis was performed using descriptive statistics, including calculations of mean, median, mode, standard deviation, and Normalized Gain (N-Gain) to assess the increase in conceptual understanding. The students' pretest average score of 54.22 increased to 78.09 on the posttest, with an increase of 23.87 points and an average N-Gain score of 0.51 (moderate category). This shows that Discovery Learning-based learning is effective in improving students' conceptual understanding of the relationship between particles and changes in the form of matter. The results of student reflections showed 83.1% positive responses, with the highest aspect being “Impressions and Messages about Learning” (90.6%), indicating that students felt that learning became more enjoyable, interactive, and meaningful. The Discovery Learning model with a Deep Learning approach has proven effective in improving students' conceptual understanding and reflective awareness of Matter States and Particle Models. This approach is recommended for implementation in science learning based on the Merdeka Curriculum to foster critical, reflective learners with scientific character.]]>
Copyrights © 2026
