Understanding bystander responses to virtual sexual assault: Predictors of intervention and self-protection in virtual environments

Abstract

This study examines predictors of bystander responses to virtual sexual assault in immersive virtual environments. Drawing on a sample of 829 adult users of virtual platforms, it investigates how situational and individual factors shape three responses: intervention, self-protection, and inaction. Multinomial logistic regression showed that incident location, normative beliefs about aggression, preparedness for community-based action, and indirect reporting or monitoring significantly influenced bystander behavior. Incidents occurring in private virtual spaces were associated with higher odds of both helping the victim and engaging in self-protective behavior. Normative beliefs about aggression predicted self-protection, but not intervention. Digital guardianship capacities, especially indirect reporting and monitoring, were associated with a greater likelihood of active responses. By contrast, factors commonly emphasized in offline bystander research, such as the number of offenders and the presence of other witnesses, were not significant. These findings suggest that bystander behavior in immersive environments differs from offline patterns and has implications for platform design, safety policy, and digital literacy.