Ngatmini

ABSTRACT This study examined differences in sixth-grade students’ conceptual understanding and critical thinking between Problem-Based Learning (PBL) supported by an Augmented Reality (AR)-Based Book and conventional instruction in Solar System learning. A quantitative quasi-experimental study employing a nonequivalent pretest–posttest control group design involved 35 students from two elementary schools in Blora, Indonesia, comprising 17 experimental and 18 control participants. The intervention was implemented across three instructional sessions. Data were collected using validated essay-based assessments and analyzed through analysis of covariance (ANCOVA) and multivariate analysis of covariance (MANCOVA), controlling for pretest performance. The experimental group achieved higher posttest means in conceptual understanding (86.71 vs. 72.67) and critical thinking (85.82 vs. 70.17). ANCOVA revealed significant adjusted between-group differences in conceptual understanding (mean difference = 13.87, 95% CI [11.45, 16.29], p < .001) and critical thinking (mean difference = 15.38, 95% CI [12.19, 18.57], p < .001). MANCOVA confirmed a significant multivariate difference (Pillai’s Trace = .862, F(2, 30) = 95.34, p < .001). These findings suggest that integrating AR-supported visualization within structured problem-based inquiry may support conceptual learning and evidence-based reasoning. Nevertheless, the small sample, nonrandom allocation, and school- and teacher-level confounding preclude definitive causal conclusions. Larger multisite and factorial studies are recommended.