A model for eva‎luating the effect of road factor on national road transportation

10/11/2023 12:00:00 AM

In order to improve the traffic safety of an area, the first and most important step is to have the necessary view of the area’s current safety situation. The existence of this point of view is needful for purposed necessary actions in order to achieve the desired goal. On the other hand, an accident is a multi-causal event. In other words, the occurrence of accidents on the roads is due to the existence of a chain of causes including the human, the vehicle, the road, and the environment. Accordingly, in this study, the aim is to provide a method to find the contribution of the road factor in the severity of traffic accidents in each province for comparing and eva‎luating them. In order to achieve the above goal, first, by reviewing previous studies, the variables influencing the severity of accidents at the macro level (level of each province) and micro level (level of each accident) were identified. After collecting the values of available variables from databases and statistical yearbooks, the values that were not available were estimated using pseudo-inductive exposure methods and statistical models (logistic regression). In the next step, according to the existence of macro and micro variables (in both levels of variables three factors are categorized as the human, vehicle, and road and environment), the framework of the present problem is defined, and a multilevel logit regression model with random effects was made. The results of the model showed that at the micro level, the variables of accident site geometry, accident time, gender, age, type of collision, and type of vehicle, and at the macro level the variables of the number of speed violations per thousand kilometers traveled and injustice in income distribution (Gini coefficient) are the most effect on the severity of accidents. In the next step, the sensitivity analysis of the constructed model was done, based on which the contribution of human factors, vehicle, road, and environment in the severity of the accidents was calculated for each province. The results of the study showed that the highest road share is for Golestan province with 48.5% and the lowest for Tehran province with 16.5%.

شدت تصادفات ترافيكي , رگرسيون لوجستيك , مدل‌هاي چند سطحي , ارزيابي ايمني منطقه‌اي

Severity of traffic accidents , Logistic regression , Multilevel models , Regional safety assessment

Moslem Azizi Bondarabadi

Ali Tavakoli Kashani