Objective: The goals of this research are to: (i) investigate the contributing factors of major road traffic accidents using a human factors classification that follows the ideas of the human factors analysis and classification system (HFACS); (ii) quantitatively examine the relationships between the human factors of the cross levels in an entire system. Methods: This study examined 234 major road traffic accidents recorded in 27 Chinese provinces from 1997 to 2014. Odds ratio (OR) was used to quantitatively analyze the relationships among the contributing factors. Results: The frequencies of unsafe acts, violations, and inadequate regulation are the highest in five categories, 15 subcategories, and 63 indicators, respectively. This study has demonstrated a number of associations between the upper and adjacent lower levels. At the outside factors level, “failure to provide supervision for regulatory” can be viewed as a strong predictor to “formal accountability for actions,” “norms and rules,” and “values and beliefs.” At the organizational influences level, “formal accountability for actions,” “norms and rules,” and “values and beliefs” were strong predictors. At the unsafe supervision level, “failure to provide oversight,” “failure to initiate corrective action,” and “failure to enforce rules and regulations” had strong prediction on “fatigue driving.” At the preconditions for unsafe acts level, “visual limitation”, “fatigue driving,” and “vehicle faults” were strong predictors. Conclusions: The generic HFACS failure types were interpreted and applied successfully to the road safety context, and such examination of major accidents has provided significant findings concerning the main contributing factors of those accidents. Using the OR technique, this study has demonstrated a number of associations between the upper level and adjacent lower levels in the entire system and has found the routes to failure, which is particularly important for developing countermeasures and remediation strategies, as it ensures that these countermeasures are targeted to a wider range of systems. Furthermore, these findings demonstrate the efficiency and applicability of the HFACS as a retrospective tool for the analysis of major road traffic accidents.

Studies that measure individual differences leading to risky driving behaviors in pre-crash phase will make contributions to accidents prevention. The primary concern here is what induces drivers to engage in risky behaviors. In this research, 441 valid questionnaires were distributed to examine the impact of risk preference, risk perception, and their interaction on risky driving behaviors, and the moderating effects of gender, age, and driving experience were measured accordingly. Results from the ordered logit regression model analysis demonstrate that the data fit well with our theoretical model. Risk preference and risk perception both predict risky driving behaviors with risk perception having greater predictability, and their interaction significantly affects risky driving behaviors when gender and age variables were added to the model separately. Gender and driving experience moderate the influence of risk perception on risky driving behaviors. The predictive effect of risk perception on risky driving behaviors was more significant for females than males. The effect of risk perception on risky driving behaviors was more pronounced for drivers with 1-3 years of driving experience compared with others. These interesting findings suggest that interventions need to be directed to all parts of the causal chain.

This study aimed to provide a greater understanding of the systemic factors involved in coal mine accidents and to examine the relationships between the contributing factors across all levels of the system. Ninety-four extraordinarily major coal mine accidents that occurred in China from 1997 to 2011 were analyzed using the human factors analysis and classification system (HFACS). The empirical results showed that the frequencies of unsafe behaviors, inadequate regulation and failure to correct hidden dangers were the highest among five levels, 14 categories and 48 indicators, respectively. The odds ratio technique was applied to quantitatively examine the relationships between contributing factors. Various statistically significant associations were discovered and should receive greater attention in future attempts to develop accident measures. In addition, several strategies concerning the main contributing factors and routes to failure are proposed to prevent accidents from reoccurring in an organization.