Eliciting the influence of roadway and traffic conditions on hurricane evacuation decisions using regression-content analysis approach

Abstract

Hurricane evacuation decisions can be complex depending on the available information and transportation options. Although significant research has been performed to understand the associated factors for hurricane evacuation decisions, the influence of roadway transportation and associated traffic conditions has received less attention. This study applied regression analysis and text mining approaches on publicly available survey data conducted after Hurricane Irma to understand the influence of roadway transportation and associated traffic conditions on evacuation decisions. Only respondents who received evacuation orders were used in this study. The regression analysis and text mining results revealed that roadway transportation-related factors, including traffic and roadway conditions and gasoline availability, were significantly associated with a lower likelihood of evacuation. Further, respondents were less likely to evacuate if they received voluntary, shelter-in-place, and any other combination of orders other than mandatory orders. Higher-income people were also less likely to evacuate. Considering building type, people in apartments and single homes were less likely to evacuate than mobile homes. Additionally, people who experienced hurricanes and/or tsunamis were less likely to evacuate. On the other hand, high education level, male respondents, people who live in flood risk areas or zone as determined by the Federal Emergency Management Agency, people with experience with terrorism, and those with prior evacuation experience were more likely to evacuate. The study findings can be used to determine the critical characteristics of vulnerable people who are less likely to evacuate and direct more resources to make sure that they evacuate to potentially save their lives.