Bicyclist injury severity classification using a random parameter logit model

Bicycling has been actively promoted as a clean and efficient mode of commute. Besides, due to the personal and societal benefits it provides, it has been adopted by many city dwellers for short-distance trips. Despite the integral role this active transport mode plays, it is unfortunately associated with a high risk of fatalities in the event of a traffic crash as they are not protected. Many studies have been conducted in several jurisdictions to examine the factors contributing to crashes involving these vulnerable road users. In the case of Louisiana which is currently experiencing increased cases of severe and fatal bicycle-involved crashes, less attention has been paid to investigating the critical factors influencing bicyclist injury severity outcomes using more detailed data and advanced econometric modeling frameworks to help propose adequate policies to improve the safety of riders. Against this background, this study examined the key contributing factors influencing bicyclist injuries by using more detailed roadway crash data spanning 2010–2016 obtained from the state of Louisiana. The study then applies an advanced random parameter logit modeling with heterogeneity in means and variances to address the unobserved heterogeneity issue associated with traffic crash data. To overcome the imbalanced data issue, three major crash injury levels were used instead of the conventional five crash injury levels. Besides, the data groups classified under each injury level were compared for the final variable selection. The study found that distracted drivers, elderly bicyclists, careless operations, and riding in dark conditions increase the probability of having severe injuries in vehicle-bicyclist crashes. Moreover, the variables for straight-level roadways and city streets decrease the odds of severe injuries. The straight-level roadway may provide better sight distance for both drivers and bicyclists, and complex environments like city streets discourage crashes with severe injuries.