Desktop virtual reality with scaffolded interaction and reflective explanation: effects on cognitive, motivational, and learning outcomes in engineering education

Abstract This study examined how a sequenced virtual reality (VR)–supported learning activity influenced learning and motivation in undergraduate engineering education. Drawing on cognitive load theory, generative learning theory, and the cognitive theory of multimedia learning, the study evaluated a desktop VR instructional sequence that included exploration of a simulated mechanical system, scaffolded interaction tasks, and a reflective self-explanation activity. A small sample of 38 undergraduate students was randomly assigned to either a VR-supported instructional condition or a traditional instruction control group. Students in the VR condition interacted with a desktop simulation of an excavator, completed scaffolded measurement and analysis tasks, and later responded to a structured self-explanation prompt before solving a novel statics problem. Compared with the control group, students in the VR-supported condition demonstrated higher procedural knowledge, stronger problem-solving performance, and greater self-efficacy. They also reported lower extraneous and intrinsic cognitive load. Declarative knowledge scores were higher following the introduction of scaffolded guidance, although the between-group difference did not reach statistical significance, and no significant differences were observed for germane load or maintained situational interest. In summary, the results suggest that outcomes varied across phases of the VR-supported activity relative to traditional instruction. While the study does not isolate the independent effects of individual instructional components, its small sample and use of a single VR application limit the generalizability of the findings. Nevertheless, the findings provide evidence that structuring VR learning activities with scaffolded interaction and reflective explanation may support learning efficiency and learner confidence in engineering contexts. These results offer design-relevant insights for structuring instructional supports within simulation-based learning environments.