Investigating jaywalker crossing risks from the sequential-conflict perspective: A grouped random parameters generalized ordered probit model.

Jaywalking at the mid-blocks without crossing facilities has been widely acknowledged as a dangerous crossing behavior. Due to the "unexpected" nature of jaywalking activities, jaywalkers may encounter multiple conflicts in sequence, called sequential jaywalker-vehicle (J-V) conflicts in this study. Sequential conflicts are usually complex and hazardous, but receive little attention in previous studies. To fill such research gaps, this study investigates the jaywalking risks from the sequential-conflict perspective, providing a more comprehensive depiction of the jaywalking hazards. Moreover, the "lagged effects" of the previous conflict-related factors on the subsequent conflict are also explored. Such considerations could help with predicting J-V conflicts in advance, further assisting the drivers to adopt precautionary actions. The results indicate that, as the conflict sequence increases, a higher proportion of jaywalkers would behave aggressively, and the proportion of slight conflicts increases accordingly. Furthermore, the grouped random parameters generalized ordered probit model is established to identify the influencing factors of sequential conflicts, which could account for the panel effects and unobserved heterogeneity simultaneously. The model results indicate that jaywalkers' previous evasive actions, such as "decelerating while swerving behind the vehicle", could lead to severer subsequent conflicts. In addition, jaywalkers' previous risky actions (i.e. rushing before the vehicle, using phones) are prone to convey to the subsequent conflicts, further increasing the subsequent conflict severity level. Moreover, non-motor vehicles are found to be a threat to the jaywalkers at mid-blocks without crossing facilities. In light of the above-mentioned results, practical suggestions are provided to reduce the jaywalking activities and prevent the jaywalker-involved accidents.