中文摘要：

使用Amos建立了31个省市道路交通事故的结构方程，SEM道路交通结构模型选用了潜变量交通事故、人口素质、经济因子和车辆道路，每个潜变量包含三至五个指标观察变量，所有数据均通过信度检验。 SEM道路交通结构模型采用bootstrap Stine p估计加以修正卡方值，通过了适配性检验和显著性检验，并拟合好全部路径系数、潜变量因子得分权重等，以此可计算得31个省市道路交通事故得分和排名，实证结果与31个省市道路交通事故实际情况相符，最后根据分析得出了影响道路交通事故的因素，并有针对性地提出了相关政策建议。

英文摘要：

A structural equation model （ SEM） was established by applying Amos based on road traffic accidents occurred in 31 provinces and cities.In this SEM, multiple latent variables were selected , including traffic acci-dents, population quality, economic factors and vehicles on the road , and each factor consists of 3 to 5 observed variables.All data passed confidence verification test .In the SEM, bootstrap stine p was used to correct chi square value.Through the adaptation and significance test and a good matching of all path coefficients , etc., SEM was used to score and get rank of the road traffic accidents of 31 provinces and cities , and the ranking was consistent with the actual situation .According to the analysis on influencing factors of road traffic accidents , some correspond-ing policy recommendations were proposed .