Exploring the relationship between cyclists' perceived unsafety, crash risk, and exposure in Dutch cities

Road safety of cyclists can be investigated in terms of objective measures based on crashes and conflicts and in subjective measures based on perceptions of cyclists. The number of studies investigating these two safety measures simultaneously is limited, in particular studies that also include exposure metrics. Therefore, the present study aims to find out to what extent perceived unsafety and crash risk of cyclists correlate and spatially align and how this relates to exposure to cyclists and motorised vehicles in the four largest Dutch cities. For this purpose, data and models estimated in earlier work on objective road safety and perceived safety of cyclists are combined. Perceived unsafety is expressed as the probability that a cyclist indicates a road section as unsafe while crash risk is expressed as the probability of a bicycle crash occurring. Results show a significant positive correlation between perceived unsafety and crash risk. It is also shown that perceived unsafety increases stronger than crash risk, which is particularly related to an increase in exposure to cyclists, followed by exposure to motorised vehicles. Conversely, crash risk remains relatively low with an increase in exposure to cyclists, which might indicate a safety-in-numbers effect. However, from a certain point in the exposure to cyclists, crash risk increases more strongly. Presumably, this hints at a situation beyond the safety-in-numbers effect where increasing cycling volumes affect cycling safety more negatively. It can be concluded that the probability of perceiving a road section as unsafe significantly follows the same direction as the probability of a bicycle crash occurring, but with a increasing exposure to cyclists perceived unsafety increases stronger than crash risk. With higher exposure to motorised vehicles, on the other hand, the increase in perceived unsafety and crash risk is more gradual.