In September 2017, Hurricane Irma prompted one of the largest evacuations in U.S. history of over six million people. This mass movement of people, particularly in Florida, required considerable amounts of public resources and infrastructure to ensure the safety of all evacuees in both transportation and sheltering. Given the extent of the disaster and the evacuation, Hurricane Irma is an opportunity to add to the growing knowledge of evacuee behavior and the factors that influence a number of complex choices that individuals make before, during, and after a disaster. At the same time, emergency management agencies in Florida stand to gain considerable insight into their response strategies through a consolidation of effective practices and lessons learned. To explore these opportunities, we distributed an online survey (n = 645) across Florida with the help of local agencies through social media platforms, websites, and alert services. Areas impacted by Hurricane Irma were targeted for survey distribution. The survey also makes notable contributions by including questions related to reentry, a highly under-studied aspect of evacuations. To determine both evacuee and non-evacuee behavior, we analyze the survey data using descriptive statistics and discrete choice models. We conduct this analysis across a variety of critical evacuation choices including decisions related to evacuating or staying, departure timing, destination, evacuation shelter, transportation mode, route, and reentry timing.