Nonparametric estimation of petroleum accident risk to improve environmental protection

The researchers collected and examined 10 years of petroleum-related accidents in the state capital of New York (NY) to develop a preliminary model (N = 1,005). The goal of the research was to propose an evidence-driven methodology to inform urban environmental policy making and emergency preparedness planning. Albany, NY, USA, was a preferentially selected sample site since it was a large city in an environmentally sensitive region with controversial oil–gas fracking policies being debated within government. The objective of the study was to develop a predictive model from petroleum accident data using nonparametric inferential statistical techniques to avoid the constraints inherent of normal distribution assumptions. A statistically significant model was formulated and tested, which indicated that the probability of petroleum accidents in the gas–oil industry was almost six times higher than their occurrence by people in other groups, such as electricity generation, transportation, hospitals, universities, warehouses, government, businesses, and residences.