<![CDATA[Earth and space concepts are difficult for elementary students because the phenomena are abstract, spatial, and rarely observable. This study examined whether Assemblr Edu-based augmented reality (AR) embedded in problem-based learning (PBL) improves sixth graders’ conceptual understanding more than identical PBL using two-dimensional (2D) media. A quasi-experimental pretest–posttest nonequivalent control group design involved two intact classes at a public elementary school in Ternate, Indonesia (n = 25 each). A 10-item essay test measured six cognitive processes of understanding. Because some distributions were non-normal, normalized gain (N-Gain) was compared with the Mann–Whitney U test, supported by Wilcoxon tests, ANCOVA, and Quade’s rank ANCOVA. The experimental class improved from 62.00 to 91.20 (N-Gain = 0.78, high) and the control class from 54.80 to 74.80 (N-Gain = 0.43, moderate). N-Gain was significantly higher with AR (U = 4.5, p < .001, r = .85, Cliff’s δ = .99), and the pretest-adjusted posttest advantage persisted (F(1, 47) = 131.14, p < .001, ηp² = .74). AR was superior on five of six indicators, notably inferring and comparing, but not interpreting. AR thus adds value over 2D media within PBL when 3D models serve as evidence for spatial reasoning; the unusually large effect requires confirmation.]]>
Copyrights © 2026
