Abstract
We investigated whether virtual reality (VR) training transfers to real-world performance. Participants were trained to complete a real-world maze task, which involved quickly moving an object through a tabletop maze without hitting the maze walls. Participants either trained in VR, trained by physically simulating the task (PS), or received no training (NT). After five timed training trials (or no training in the control condition), participants completed five timed test trials using the real-world maze. The VR training condition completed the maze significantly faster than the other conditions across each test trial. Other than the initial test trial where the NT condition was the slowest, the PS and NT conditions were comparable on the remaining test trials. These results suggest that VR provided trainees with an opportunity for meaningful encoding of the task, beyond the act of physically simulating the motions of the task.