An empirical analysis on the effects of alcohol-impairment in injury severities of motorcyclists

Abstract

AbstractThis research study examined the factors contributing to the injury severity of alcohol-impaired motorcyclists (BAC ≥ 0.08), compared to non-alcohol-impaired motorcyclists (BAC < 0.08). This study modelled rider injury severity in single-motorcycle crashes using random parameter multinomial logit with heterogeneity in mean and variance. This approach accounts for possible unobserved heterogeneity in the data resulting from rider, roadway, motorcycle, traffic and/or environmental conditions. The model results show that rider characteristics, such as role (rider), demographics (age between 50 and 65 years), motorcycle type (cruiser), rider endorsement (endorsement required), riding behaviour (speed more than 10 mph over the speed limit), riding action (no distraction prior to crashes), no helmet usage, roadway characteristics (left curved alignment, road surface condition), temporal characteristics (Sundays), spatial characteristics (Florida Department of Transportation District 5) and crash characteristics (collision with roadside fixed object) interact in a complex way to affect riders’ injury severity. The estimated model results suggest that the likelihood of fatality was 3.1 times higher for harmful events of roadside fixed objects, 6.6 times higher for left-aligned roadway and 10.4 times higher for riders in single-motorcycle crashes involving alcohol-impairment (BAC ≥ 0.08) relative to non-alcohol impairment (BAC < 0.08). The results clearly indicate the elevated safety concerns of alcohol impairment beyond the legal limit (BAC ≥ 0.08) has on rider injury severities. Alcohol-impaired riding and its impact on risky riding behaviours needs to be prioritised and addressed through education, enforcement and engineering countermeasures.Keywords: Single-motorcycle crashesinjury severityblood alcohol content (BAC)unobserved heterogeneityalcohol impairment Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 Alcohol Test Type is variable, which is coded as ‘Test Not Given’, ‘Test Refused’, and ‘Unknown, if Tested’ according to Crash Analysis Reporting (CAR) system. However, the corresponding BAC test numbers are updated after some legal procedures (medical test and priority of police cases of such nature), with full information of BAC test, in the CAR database. Since updating of BAC related parameter is time-dependent procedure in the police-reported crash database, relatively older data could be more reliable (avoiding missing or under reporting issues) to use in the study of statistical analysis.2 See detail on age group of motorcyclists in Florida in [Citation76]. The effect of motorcyclists’ age on injury severities in single-motorcycle crashes with unobserved heterogeneity. Journal of Safety Research, 77, 125–138.3 Although there are many other variables in the crash dataset, Table 1 shows only the statistically significant variables. Otherwise, the list of variables and their descriptive statistics will be quite long.