<![CDATA[This study aims to examine differences in IPAS learning outcomes between students taught through Problem-Based Learning (PBL) assisted by Assemblr Edu and those taught through conventional instruction, as well as to analyze the interaction between the learning model and students’ self-confidence. The study employed a quasi-experimental design using a pretest–posttest non-equivalent control group approach. The participants were 115 fifth-grade students from Cluster II Tegallalang Elementary Schools, comprising 57 students in the experimental group and 58 in the control group. The research instruments included a multiple-choice IPAS learning outcomes test and a self-confidence questionnaire, both of which were validated and tested for reliability. Content validity was assessed by two expert lecturers using Gregory’s formula. The learning outcomes test demonstrated very high reliability (KR-20 = 0.958), and the self-confidence questionnaire also showed very high reliability (Cronbach’s α = 0.929). Data were analyzed using ANCOVA. The results indicate that the learning model significantly affected IPAS learning outcomes (F = 56.791, p < 0.001, Partial η² = 0.340), while self-confidence also had a significant effect (F = 43.628, p < 0.001, Partial η² = 0.284). Furthermore, a significant interaction was observed between the learning model and self-confidence (F = 6.486, p = 0.012, Partial η² = 0.056). These findings suggest that PBL assisted by Assemblr Edu is more effective than conventional instruction, indicating that the integration of problem-based learning with AR-supported media can reduce learning barriers and enhance conceptual understanding in a meaningful way.]]>
