Prediction of commitment to wearing seatbelt and response to changes in seatbelt laws with parametric and decision tree models

This study aimed at investigating factors affecting seatbelt use for drivers in Bahrain, and their response to changes in seatbelt laws. An online questionnaire survey was conducted for primary data collection. Multivariate analysis, logistic regression and Classification and Regression Tree (CART) models were used for assessing the effects of traffic laws, crashes, violations and other driver characteristics on use of seatbelt. Compared to traditional logistic regression models, CART models were found to be less sensitive to datasets used for development and provided better accuracies, compared to logistic regression models, for the prediction problems in this study. Observations made from the CART models are as follows. Young female drivers were more likely to be affected by changes in traffic laws. Moreover, violation, age, gender, and driving experience were other factors which affect the drivers’ commitment to seatbelt, with violation having the most prominent effect. However, the findings of this study suggest that drivers’ perception, related to road safety and traffic, plays a bigger role in increased commitment to seatbelt use, as compared to prior violation or crash. In this context, young drivers showed higher tendency to ignore seatbelt use, and being involved in violations and crashes. Hence, it is suggested to introduce measures in the licensing procedure to increase awareness of drivers related to road safety.