| dc.description.abstract | Resilience is an important concept in the context of natural disasters because it helps reconcile why similar events can have such disparate impacts on various communities. However, little progress has been made on using data-driven approaches to improve assessments of community resilience. Here, we use factor analysis to develop a schema for comparing a popular class of frameworks aimed at measuring community resilience, called resilience indices. Our findings suggest that 50 of the 130 variables that comprise six of the most established indices in the field effectively load onto five main dimensions of community resilience: wealth, poverty, agencies per capita, elderly populations, and non-English speaking populations. The resulting model serves as an intuitive set of features from which researchers can predict disaster outcomes and establish construct validity. Additionally, we develop empirically-based decision models that can be used to predict and generate responses in common-pool resource dilemmas pertaining to public infrastructure investments that bolster a community’s resilience to disasters. We find that individuals tend to over-contribute relative to Pareto efficient strategies in this setting and that the personality trait Openness has a significant and nonlinear effect on responses, suggesting that it may be warranted to reexamine results of previous studies that assume linear relationships between participants' dispositions and decisions in cooperative dilemmas. We implement these decision models in an agent-based model that predicts state investments toward the development of a publicly operated port along the Mississippi River that has the ability to mitigate regional production losses due to floods. We find that state investments often lead to inefficient monetary outcomes that are not conducive to sustained cooperation. However, when decisions are viewed as subjective valuations of the port, decision-makers can better understand the motivation behind one another's tendencies and biases, and paths toward resolving the dilemma are revealed. | |
