Investigating Spatial-Behavioral Patterns in Hazards: A Virtual Reality Study as A Data Gathering Method

Analyzing multi-hazards requires a comprehensive approach, involving complexities in studying multiple hazards and challenges in visualizing numerous risks due to the abundance of information (Kappes, et.al.2012). Risk perception research, on the other hand, has emerged to aid decision-makers in understanding how people characterize and evaluate different hazards, anticipating behavioral responses, and guiding risk communication. Although the risk perception concept has been integrated into various behavioral theories applied to examine preparedness for numerous hazard types, there remains a gap in understanding which theories are suitable for examining multiple hazard types simultaneously (Gill & Malamud, 2017). Therefore, anthropogenic factors indirectly influencing multi-hazard risk assessment need addressing. Studying human behavior in multi-hazard scenarios presents inherent challenges, primarily due to the retrospective nature of analyses conducted after the event. The lack of direct observation during occurrences hampers the formulation of questions and modeling beforehand, limiting the ability to address perception and recall biases in real time. Despite these challenges, a thorough examination of catastrophes necessitates understanding not only how people behave but also delving into the underlying reasons for their behavior, a longstanding challenge in economics and social sciences (Wilson, 2017).

Virtual Reality (VR) environments emerge as valuable tools for overcoming these challenges. VR facilitates a more natural interaction among participants, providing an ideal setting to explore complex behavioral dynamics in disaster scenarios, previously nearly impossible in controlled settings. Combining the internal validity of laboratory experiments with the external validity of field or natural experiments (Fiore et al., 2009), VR enables repeated experiments with large subject pools, a challenge in real disaster situations. This allows researchers to achieve realistic yet replicable results that traditional methods struggle to attain. In contrast to real-world disasters, VR experiments avoid participant attrition, a common issue in natural research studies introducing biases. Conducting numerous identical experiments with a significant number of participants allows researchers to subtly manipulate factors and interactions, exploring specific questions comprehensively. Participants in VR experiments can engage in multiple scenarios, facilitating the exploration of learning behavior beyond one-time event analyses. Fiore et al. (2009) emphasize that VR participants can experience long-term scenarios in a short time, generating multiple counterfactual scenarios.

While traditional laboratories played a central role in advancing behavioral economics, VR is poised to be vital for inclusive multidisciplinary behavioral research in more realistic environments. It not only addresses methodological challenges associated with disaster research but also opens avenues for nuanced exploration of the reasons behind human behavior in disasters. This study examines the use of VR as an innovative tool for new risk assessment in complex contexts, considering behavioral differences and mobility preferences of participants with and without familiarity with the spatial environment.