Analysis of driver fatigue causes using log-linear models

Previous studies suggest that at least 25 % of vehicle accidents are caused by fatigue. To better understand this phenomenon, a set of fatigue causes were examined in this study to assess the associations amongst them. The study was conducted through a questionnaire filled by 741 private car drivers in Amman, Jordan, containing questions for men and women drivers as to whether they were exposed to fatigue. In case they were, drivers were asked to identify the time and period of drowsiness. By obtaining independence models in two-way and three-dimensional tables, the data was analyzed using log-linear models. The results showed a strong association between fatigue and gender with an interaction parameter less than (0.05). The risk of fatigue exposure in men was found to be twice the fatigue exposure in women. The partial independence model of sleep amount was independent of fatigue, and the driving period was the best model to describe their relationship. At night, the risk of getting tired was 2.21 times during daytime. The driving period factor was a more powerful effect on fatigue of the driver than gender, with the interaction parameter of the driving period and fatigue being less than (0.0001). The partial independence model of gender was independent of fatigue, and driving time was the best model to describe their association.