<![CDATA[Science learning in elementary schools still faces obstacles, particularly because students struggle to understand the state of matter and its changes. This struggle arises since the concept is abstract and requires strong visual-spatial abilities. To address this, the present study aims to determine the effectiveness of augmented reality-based learning media in improving students' visual-spatial abilities, especially in science. The research method employed was a quasi-experimental design with a pretest-posttest control group. Seventy fourth-grade students participated, divided into two control groups and one experimental group. To measure outcomes, the research instrument consisted of a visual-spatial ability test developed based on five main indicators. For analysis, data were examined using a t-test to assess differences in learning outcomes between groups, and an N-Gain test was used to measure the improvement. Results revealed a significant difference between the two groups, as indicated by the t-test with a significance value of 0.002 is smaller than 0.05. Additionally, the N-Gain test demonstrated that the experimental group achieved an average N-Gain of 64.63%, categorizing it as fairly effective; in contrast, the control group obtained only 46.10%, placing it in the less effective category. Therefore, using augmented reality significantly enhances students' visual-spatial abilities in the context of states of matter and their transformations. These findings imply that augmented reality can serve as an alternative, relevant, and innovative learning medium to support science learning in elementary schools.]]>
