Analysis of safety climate and individual factors affecting bus drivers' crash involvement using a two-level logit model. (May 2021)

• Record Type: Journal Article
• Authors: Wang, Xuesong; Jiao, Yujun; Huo, Junyu; Li, Ruirui; Zhou, Chu; Pan, Hanzhong; Chai, Chen
• Is Part Of: Accident analysis and prevention. Volume 154(2021)
• Publication Date: 2021-05
• Subjects: Bus driver -- Crash involvement -- Safety climate -- Fleet-level -- Individual characteristics -- Random-effects two-level logit model; Accidents -- Prevention -- Periodicals; Accident Prevention -- Periodicals; Accidents -- Prévention -- Périodiques; 363.106
• DOI: 10.1016/j.aap.2021.106087
• Languages: English
• ISSNs: 0001-4575

Abstract

Highlights: A questionnaire was conducted on 725 bus drivers of an intercity bus company in Shanghai. A random-effects two-level logit model is conducted considering fleet and individual factors. Safety climate and drivers' characteristics are found to be significant variables predicting self-reported crashes. Psychological distress associates with the high frequency of risky driving behavior and high severity of driving anger. Abstract: Although traffic crashes involving buses are less frequent than those involving other vehicle types, the consequences of bus crashes are high due to the potential for multiple injuries and casualties. As driver error is a primary factor affecting bus crashes, driver safety education is one of the main countermeasures used to mitigate crash risk. In China, however, safety education is not as focused as it should be, largely due to the limited research identifying the specific driver behaviors, and potential influences on those behaviors, that are correlated with crashes. The aim of this study is, therefore, to explore the fleet- and driver-level risk factors underlying bus drivers' self-reported crash involvement, including analyzing the effect of psychological distress on the most influential driver-level factors. A survey was conducted of 725 drivers from a large Shanghai bus company, and a random-effects two-level logit model was developed to integrate fleet and individual variables. Results showed that: 1) the fleet-level safety climate explained about 8.5% of the model's variance, indicating it was a valid predictor of self-reported crash involvement; 2) the driver-level factors of drivers' age, seniority, marital status, positive behavior, and driving anger influenced drivers' self-reported crash involvement, but ordinary violations, lapses, aggressive violations, and insomnia were the most influential variables; 3) psychological distress appeared to associate with the high frequency of risky driving behavior and the high severity of driving anger. This study's findings will help bus companies to give more attention to their safety climate and implement more targeted improvements to their driver safety education programs.