Identification of key determinants of personal safety perception at bus stops using proportional odds logistic regression

Complex vehicle-pedestrian interaction near the bus stop affects the safety perception of users, but the subjective nature of perceived safety and the lack of specific measurement techniques makes its estimation further laborious. This paper identifies factors influencing safety perceptions of bus users by employing Ordered Logistic Regression analysis. The study focussed on five dimensions; bus stops' design and surrounding facilities, traffic characteristics, and travellers' individual, travel and accident characteristics. The results demonstrated that increased vehicular flow, absence of footpath, crossing facility and safety barrier would scale down the perceived safety. Furthermore, the respondent's age, education, frequency of travelling by bus, familiarity with the bus stop, and previous accident experience are significant predictors of perceived safety. Results concluded that females perceived less safe, and no effects can be attributed to the household vehicles and income. Regarding road shoulder width, 1-2 meters is the most preferred width from the travellers' safety perspective.