Immersive virtual reality (VR) is predicted to have a significant impact on education; but most studies investigating learning with immersive VR have reported mixed results when compared to low-immersion media. In this study, a sample of 118 participants was used to test whether a lesson presented in either immersive VR or as a video could benefit from the pre-training principle, as a means of reducing cognitive load. Participants were randomly assigned to one of two method conditions (with/without pre-training), and one of two media conditions (immersive VR/video). The results showed an interaction between media and method, indicating that pre-training had a positive effect on knowledge (d = 0.81), transfer (d = 0.62), and self-efficacy (d = 0.64) directly following the intervention; and on self-efficacy (d = 0.84) in a one-week delayed post-test in the immersive VR condition. No effect was found for any of these variables within the video condition.