<![CDATA[This study aims to determine the effectiveness of the STEM-based project-based learning (PjBL) model in improving students' basic science learning outcomes. The study was conducted at MIS Darul Huda Campang Jaya, Bandar Lampung. This study employed a quantitative method with a quasi-experimental design, specifically the pretest and posttest control group design. The sample consisted of 78 students, divided equally into an experimental class (39 students) and a control class (39 students), selected via simple random sampling. The research instrument was a 20-item multiple-choice science achievement test, which was validated by two science education experts and one science teacher. The validity of each test item was analyzed using Pearson Product-Moment Correlation, while reliability testing using Cronbach's Alpha obtained a coefficient of 0.86, indicating high reliability. Data analysis techniques included normality testing with the Kolmogorov–Smirnov test, homogeneity testing with Levene's Test, and hypothesis testing with the independent-samples t-test. The results showed that the significance value of 0.000 < 0.05 indicated a significant difference in students' science learning outcomes between the experimental and control classes. The findings confirmed that implementing the STEM-based PjBL model effectively improved conceptual understanding, critical thinking, and collaboration skills through contextual, hands-on learning experiences. Therefore, this model can be recommended as an innovative approach to enhancing 21st-century competencies in elementary school science learning.]]>
