Examining risky driving behaviours: A comparative analysis of SUVs and other car types

Abstract

The aim of the study was to analyze whether the driving behaviour changes with different types of cars (Hatchback, Sedan and Sports Utility Vehicle (SUV)), particularly on expressways. To achieve the study's objective, an exclusive DBQ was designed. Self-reported frequency of instances related to violations, mistakes and lapses were gathered which can collectively be used to study driver's behaviour in different circumstances. A total of 546 samples were collected through an online survey from drivers across India who possessed a valid driving license. Results of exploratory factor analysis confirmed two factor solution (errors and violations) and the same was verified by confirmatory factor analysis. With the help of structural equation modelling, individual risky driving score of each driver was calculated. Based on the calculated scores, drivers were clustered into different risk levels using K-Means clustering technique. Further, ANOVA test revealed that SUV drivers performed riskier behaviour more compared to other types of cars. To compare the relationships among different risk levels with drivers' demographics and driving characteristics, two ordered logit models were developed for SUV and Standard car drivers. The results indicated that drivers' age, crash history, and speeding behaviour were statistically significant predictors of risky driving behaviour. Comparison of risk levels between SUV and Standard cars with respect to different predictor variables presented various interesting findings. These findings would be helpful in understanding the differences among risky driving behaviour performed by drivers of different types of cars and to identify the potential road safety countermeasures.