This study examines the influence of spatial effect and demographic characteristics on postdisaster households' relocations at the block group level. Data for 479 block groups were collected in the city of New Orleans following Hurricane Katrina. The explanatory variables were transformed into five uncorrelated principal components using principal component analysis and spatial autocorrelation of the relocation decision of households tested using Moran's I index. The results indicate households' decisions exhibit significant spatial clustering. A better statistical fit was found for spatial lag regression than standard linear regression and other spatial models. Four statistically significant components influencing the block group level percentage of relocated households were identified. Specifically, places characterized by a greater proportion of the population with higher income population and out-of-county employment are more likely to experience high rates of relocation. In addition, family ties and place attachment had significant negative effects on the rate of block group-level relocation.
|Journal||Journal of Infrastructure Systems|
|State||Published - Mar 1 2021|
- Hurricane Katrina
- Principal component
- Relocation decision
- Spatial effect