Virtual and immersive virtual reality, VR and iVR, provide flexible and engaging learning opportunities, such as virtual field trips (VFTs). Despite its growing popularity for education, understanding how iVR compared to non-immersive media influences learning is still challenged by mixed empirical results and a lack of longitudinal research. This study addresses these issues through an experiment in which undergraduate geoscience students attended two temporally separated VFT sessions through desktop virtual reality (dVR) or iVR, with their learning experience and outcomes measured after each session. Our results show higher levels of enjoyment and satisfaction as well as a stronger sense of spatial presence in iVR students in both VFTs compared to dVR students, but no improvement in learning outcomes in iVR compared to dVR. More importantly, we found that there exists a critical interaction between VR condition and repeated participation in VFTs indicating that longitudinal exposure to VFTs improves knowledge performance more when learning in iVR than through dVR. These results suggest that repeated use of iVR may be beneficial in sustaining students' emotional engagement and compensating the initial deficiency in their objective learning outcomes compared to other less immersive technologies.