溢油事故危害程度评估行为特征识别

Identification of behavior characteristics for hazard assessment of oil spill accidents

为识别应急决策者(emergency decision maker,EDM)对溢油事故危害程度评估的行为特征,引入有序概率模型和D-efficient意向选择实验,对其评估行为进行研究。基于意向选择实验数据和极大似然法对模型参数进行标定,根据参数显著性和模型拟合度指标选择最优模型进行边际效应分析,以揭示EDM的评估行为特征。结果表明:有序Logit模型优于有序Probit模型;溢油量和溢油影响水域对EDM的评估行为影响最大,专业背景和从业年限也会影响评估行为。研究结果能够客观描述决策者的评估行为特征,为合理使用评估结果提供理论基础。

To identify the behavior characteristics of emergency decision makers(EDMs)for hazard assessment of oil spill accidents,the assessment behaviors of EDMs are studied by introducing ordered probability models and D-efficient stated choice experiments.Both the stated choice experimental data and the maximum likelihood method are used to calibrate OPM parameters.According to indexes of parameter significance and model fitting,the best OPM is selected for marginal effect analysis so as to reveal EDMs’assessment behavior characteristics.The results show that,the ordered Logit model outperforms the ordered Probit model;both the oil spill size and its influencing area have the greatest effect on the assessment behaviors of EDMs,and education background and working years also affect assessment behaviors.The findings can objectively describe the characteristics of EDMs’assessment behaviors,and provide a theoretical basis for rational use of assessment results.