Role of place and space in a displaced population's post-disaster recovery: A geographically weighted regression analysis

Hurricane Katrina displaced an estimated 645,000 individuals from Louisiana and 66,000 from Mississippi. Many were unable to return to their pre-Katrina because of sharply increased rental costs, uninhabitable neighborhoods, or insufficient financial means to return home. Research has shown that place can predict health outcomes, and social models of disaster recovery stress the interaction of social processes and geographic place. This study tests whether place influences post-disaster recovery; specifically, whether the pre-Katrina neighborhood influences an individual's current self-reported state of recovery. Furthermore, it is posited that the further away an individual lives from their pre-Katrina residence, the poorer their recovery. This theory will be tested using a global regression model as well as a Geographically Weighted Regression (GWR) model to test for spatial processes not accounted for by the global regression model. An advantage of using a GWR is that the study space is continuous, which means the issues of edge-effects in spatial analyses are avoided, as well as making a priori assumptions as to what is and is not a ‘neighborhood.' The address database utilized is composed of the pre-Katrina and current residence for respondents from the Gulf Coast Child and Family Health Study, a longitudinal cohort study of 1,079 displaced or greatly impacted households in Louisiana and Mississippi. The GWR Findings indicate distance does not predict recovery but housing stability and income greatly influence recovery, and to a lesser degree sense of community. Several variables were found to exhibit significant spatial variation not accounted for in the global model. These findings emphasize the importance of place in health and post-disaster recovery, although a more robust model includes compositional (individual) and contextual (household) effects. Appropriate application of the GWR could help elucidate many place effects and begin to uncover the black box of geography and health.