Developing prescriptive environmental protection models from descriptive human accident behavior

“

" The researchers demonstrated that an empirically grounded environmental planning risk management model could be developed and validated from oil– gas accident data. The operations research and decision-making literature was reviewed for empirical models related to environmental disasters.Rigorous nonparametric statistical techniques were applied to build and validate the model. A contingency table was developed to calculate the marginal, joint and conditional probabilities. Two well-known nonparametric techniques were applied to estimate the reliabilities of the results, namely, the Cochran and Mantel–Haenszel tests of conditional independence. The Mantel–Haenszel odds ratio was 5.762 (3.979, 8.346), which suggests it was almost six times more likely for an oil– gas service station to have an oil– gas accident during the period 2003-2013. This study makes an original contribution to the practice by comparing our approach with those of others (in the literature) and by applying nonparametric statistical techniques on retrospective data to test the predictive model. The value of this research is through its inclusion in the literature which increases the choice of alternatives practitioners now have in their intellectual toolkit to prevent the frequency and severity of future human-caused petroleum-related accidents. Or to put it another way, the exclusion of this study from choices that practitioners can make would not serve the better interests of the profession." (p. 449)

”