Recent studies utilized virtual reality (VR) as an "empathy machine" to mitigate bias towards various social groups. However, studies addressing bias against physical disabilities remain scarce, with inconsistent results based on VR experience design. Moreover, most studies assumed the universal effects of VR simulation on bias reduction, ignoring the potential moderating effects of individual characteristics. This study investigated how experience design components and individual characteristics moderate VR simulation's effect on changes in bias towards physical disabilities. We designed a VR wheelchair experience, manipulating the situational context (negative, neutral) and whole-body avatar visualization (visible, invisible). Participants' implicit and explicit bias levels were assessed to examine the changes according to VR design components and individual characteristics (gender, preexisting bias level). Results indicated that following the VR intervention, implicit bias was reduced in the group with higher preexisting bias but rather increased in the group with lower preexisting bias. In addition, gender interacted with avatar visualization such that male participants' implicit bias was reduced with invisible avatars but increased with visible avatars. Explicit bias, in contrast, was reduced regardless of conditions, suggesting the potential response bias in self-report measures. These findings underscore the importance of considering the complex interplay between experience design and individual characteristics in understanding VR's efficacy as an empathy-inducing tool. This study provides insights and guidelines for developing more effective VR interventions to alleviate implicit bias towards physical disabilities.