Risk perception, self-efficacy, and disaster preparedness of immigrants in Japan: Applying the risk perception attitude framework

Japan frequently experiences natural disasters, and immigrants are particularly vulnerable to disasters. This cross-sectional study aimed to (1) identify demographic factors associated with risk perception, self-efficacy, and two types of disaster preparedness: cognitive (awareness of information and connections) and behavioral (actual preparedness actions), and (2) examine the relationships between risk perception, self-efficacy, and disaster preparedness using the Risk Perception Attitude Framework, focusing on the moderating role of self-efficacy in the relationship between risk perception and disaster preparedness among immigrants in Japan. Immigrants aged 18 or older were recruited through a self-administered online survey. The final data set included 196 responses (44.9% non-students). We used structural equation modeling to examine the relationships among variables. Unemployed immigrants reported higher risk perception, while those from high-income economies had lower risk perception. Disaster experience in Japan was linked to higher risk perception, self-efficacy, and behavioral preparedness. Unemployed individuals and those with very high Japanese proficiency showed greater cognitive preparedness, while renters had lower behavioral preparedness. Risk perception was significantly related to cognitive preparedness but not to behavioral preparedness. Self-efficacy predicted both cognitive and behavioral preparedness, and higher cognitive preparedness contributed to behavioral preparedness. Self-efficacy had a significant moderating effect: among those with low self-efficacy, the relationship between risk perception and behavioral preparedness was stronger. Thus, targeted interventions are required for less-prepared groups, including students, employed or self-employed individuals, those with language barriers, renters, and people without disaster experience. Limitations include the cross-sectional design and non-random sampling, which may affect the generalizability of the findings.