Predictors of Individual Resilience a Case Study of the Deepwater Horizon Oil Spill

Background and Purpose: The Deepwater Horizon (DWH) Oil Spill is regarded as the worst hydrocarbon disaster in the history of the United States. The impact from hydrocarbon disaster results in uncertainty regarding the immediate and long-term future, a situation exacerbated by lack of equity and vulnerability. The 2010 DWH disaster provides a unique opportunity to gain a scientific understanding of the understudied area of individual resilience within the context of hydrocarbon disaster events. The purpose of this paper is to present findings from a recently completed study on predictors of individual resilience in three natural resource dependent Gulf Coast communities (Port Sulphur & Galliano, LA and Bayou La Batre AL) impacted by the DWH Oil Spill.

Methods: Data was collected from a total of 326 individuals living in three communities in the Gulf Coast region after the Deep-Water Horizon Oil Spill via face-to-face interviews. The analysis in this study included two phases: (1) latent profile mixture models; and a step-wise logistic regression to identify factors that predicted latent class membership. Ten items drawn from the abbreviated version of the Connor Davidson Scale for Resilience were used as indicators for the latent classes. Once identified, a new variable “class” was used for the second phase of analysis. Predictors in this phase included, education level, employment status, age, race, gender, marital status, data collection site, disaster exposure severity, preparation for future disasters, and worry regarding the likelihood of future disasters occurring.

Results: A series of latent profile mixture models were estimated. Goodness of fit statistics, including AIC, BIC, Entropy value, Lo-Mendall Rubin and Bootstrap tests, in conjunction with the interpretability of the classes, were used to compare models. The two-class solution offered the best fit for the data and was selected as the final model, the two classes were termed “Below Average Resilience” (n=84) and “Above Average Resilience” (n=242). Subjects were allocated to their most likely latent class using posterior probability. Being employed, (OR=2.09, p<0.05) having a college degree (OR=2.36, p<0.05), and total levels of preparation for future disasters (OR=1.73, p<0.05) were indicators that increased the odds of belonging to the “Above Average Resilience” latent class. Alternatively, in comparison to white counterparts, being Vietnamese (OR=0.23 p<0.05) decreased the odds of membership in the “Above Average Resilience” latent class.

Conclusion and Implication: Despite high levels of exposure to disasters, the majority of the sample (74%) were allocated to “Above Average Resilience” latent class, suggesting in the wake of trauma and disaster the majority of individuals bounce back. These findings highlight the importance of social capital in promoting resilience and particularly the vulnerable nature of ethnic minority groups in the context of disasters. These findings have implications for the ways in which we allocate resources pre and post disaster. More specifically, increasing access to education and employment may help disaster affected individuals bounce back. Similarly, increasing efforts to prepare for disasters before they occur may also help survivors bounce back. Lastly, additional resources may be needed for resource-constrained communities, particularly among ethnic minority groups.