The influence of augmented reality on learner-generated explanations during chemical hands-on experiments

Augmented Reality (AR) enables the presentation of models at the submicroscopic level during chemical hands-on experiments and may thus support learners in connecting macroscopic observations, submicroscopic processes, and symbolic representations. Such connections are essential for understanding chemical phenomena, yet many students struggle to establish them, leading to fragmented or purely descriptive explanations. However, little is known about how AR influences the structure and quality of learner-generated explanations during experimentation. In this quasi-experimental pre-post study, three conditions were compared in grade 11 chemistry classes: AR, animation, and filmstrip. Students (N = 104) worked in pairs on two acid-base experiments and created explanatory screencasts, which were analyzed in two complementary ways: network analysis was used to examine transitions between representational levels, while a separate holistic coding scheme was applied to assess overall explanatory quality. Across both experiments, students in the AR group integrated submicroscopic entities into their explanations more frequently than students in the other groups. However, this increased integration was not consistently associated with higher explanatory quality. The findings suggest that AR can foster representational integration, but that explicit scaffolding might be necessary to translate this into high-quality causal explanations.