Statistical Analysis of Accident Patterns and The Causes at A State Road in Perlis

Abstract

Data from the World Health Organization have stated that each year, approximately 1.3 million road users are at high risk of a road traffic accident. Therefore, this study aims to evaluate the accident patterns and causes at a state road of Kangar to Alor Setar. The methods used to evaluate the accident patterns are the chi-squared test (CST) and level of service (LOS) survey. The Pearson CST shows that the significant factors for accidents are the age of the drivers, in which people below 30 years old are likely to be involved in accidents (? = 0.037), and type of vehicles used, in which motorcycles are more prone to be involved in accidents (? = 0.000). However, gender shows an equal risk for accident involvement (? = 0.911). The LOS E is observed on evenings of Monday and Friday, indicating the unstable flows of traffic during those periods. Partial least squares structural equation modelling is used to evaluate the causes of accidents from the road user’s perspective via a questionnaire. Before distribution, the questionnaire underwent a pilot study to ensure its validity and reliability. Four hypotheses were developed with the following independent variables: human factors, vehicle factors, environment factors and road condition factors. Results show that the majority of the road users (33.3%) travelled 1 to 3 times per week and 23.1% of them were involved in accidents while travelling on this road. The validity test via the structural model shows that the most significant factors affecting the relationships to the accidents are the road factors (t = 6.166, p = 0.000), the vehicle factors (t = 4.3399, p = 0.000) and the human factors (t = 2.893, p = 0.005). Thus, the factors are significant to the causes of accidents on this road. Countermeasures that emphasise these factors are very important to be considered by authorities to reduce accidents on this road.