Virtual reality (VR) has become widely accessible through the development of more commercially available head-mounted displays (HMDs). This accessibility has particularly increased the use of VR in children. However, much of the previous research on understanding perception and action in virtual reality has only been conducted on adults, leaving many open questions about how children perceive and interact with virtual environments. In this paper, we examine whether there are age-related differences in judging the ability to step over a gap using immersive VR. Affordances are a useful measure for understanding objective perceptions of the actions that can be performed in an immersive virtual environment. Such measures are particularly well suited for children given they can easily respond yes or no as to whether they perceive that they can step over a gap. Further, manipulations of the size of virtual body parts could allow us to ascertain how much children rely on the perceived size of their bodies to make decisions about actions. In Experiment 1, adults and children saw motion-tracked virtual feet that were either larger or smaller than their actual foot size. They had to respond as to whether they could step over gaps that varied in width. They also gave perceptual estimates of the width of the gap in feet or meters. The results showed that adults who experience the smaller virtual feet underestimated their stepping ability more than adults with the larger feet. However, children had the opposite effect, such that seeing smaller virtual feet led to an overestimation of their stepping ability. To test whether this age-related difference in body scaling was due to misperception of foot size, adults and children matched virtual shoes to their actual feet size in Experiment 2. This matching task showed no perceptual differences between the age groups. Across our two experiments, we showed that children scale gap affordance judgments differently than adults and this difference cannot be explained by difference in perceptions of the size of virtual feet. The results suggest that children can effectively make perception and action judgments in virtual reality, but do not always do so in the same manner as adults. Such a finding has implications for the design of virtual reality games, educational tools, and training systems that are becoming increasingly common for children.