Disentangling the effects of COVID-19 on pedestrian injury severity in motorcycle-related crashes: Evidence from mixed logit modelling

Past research has thoroughly investigated pedestrian safety, whereby most studies analysed vehicle-pedestrian interactions and the underlying pedestrian injury determinants. However, motorcyclists exhibit significantly different behaviours than car drivers, which may lead to comparable or even greater injuries to pedestrians, and possibly with different injury severity determinants. Further, with COVID-19 encouraging increased walking and single-person travel (like motorcycle riding), whether and how pedestrian injury severities vary in motorcycle-related crashes need further exploration. This study answers these questions. Using three-year data from the UK STATS19, separate year-wise mixed logit models are estimated to examine pedestrian injury severity outcomes in motorcycle-related crashes during COVID-19. The likelihood ratio tests are performed to investigate temporally unstable patterns in pedestrian injury severity. The results confirm that the presence of temporal variation in pedestrian severity outcomes is linked to pronounced effects of COVID-19. The developed models suggest some pedestrian injury severity determinants are stable, whereby speed limit (40 mph or 64 kph), evening time, and junction indicators show consistent effects across years, highlighting that these factors are not impacted by COVID-19. Contrastingly, some factors reveal varied effects across years, highlighting the effect of COVID-19, for example, the 30 mph (or 48 kph) speed limit, pedestrian crossing, urban areas, pedestrian age, and motorcyclist location change signs. Injury severity determinants are compared/contrasted between vehicle- and motorcyclist-pedestrian crashes, highlighting the need to separately analyse these interactions and frame targeted, tailored policies to minimise pedestrian injury severity. The study findings indicate pedestrian injury determinants when interacting with motorcyclists shift over time, which is attributed to COVID-19, thus requiring tailored safety interventions for pedestrians and motorcyclists. The results will underpin developing effective countermeasures, which will promote walking and contribute to the national agendas on increasing active travel uptake and achieving net-zero goals. • Pedestrian injury determinants are examined in motorcycle-related crashes. • Mixed logit models with heterogeneity in means are developed. • COVID-19 affects pedestrian injury severity and results in temporally unstable injury severity patterns. • COVID-19 has mixed effects on pedestrian injury severity. • Pedestrian injury severity determinants vary between vehicle- and motorcycle-related crashes.