Citation

Alhomaidat F, Abushattal M, Morgan Kwayu K, Kwigizile V. Transp. Res. Interdiscip. Persp. 2022; 14: e100612.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.trip.2022.100612

PMID

unavailable

Abstract

The present study examined State of Michigan Traffic Crash Reports filed between 2015 and 2019 to explore the interaction between age and liability for crashes at stop-sign-controlled intersections. A driver's liability for a crash was derived from the "Hazardous Action" field in each crash report. The likelihood of assigning liability to an elderly driver was examined in light of pre-crash actions defined in each report's "Actions Prior to Crash" field. Logistic regression was applied to calculate odds ratios used to explain the likelihood. Furthermore, Random Forest machine learning technique was used to predict driver liability based on pre-crash actions. Distraction, number of travel lanes, and driving under the influence of alcohol or drugs were significant predictors of the likelihood that an elderly driver was at-fault in a crash. A "going straight" pre-crash action by an elderly driver was the best indicator of liability, regardless of the pre-crash action by a young driver. For interaction scenarios, an elderly driver going straight at a stop-sign-controlled intersection was associated with a lower likelihood of being liable for a crash. Turning actions increased the likelihood of the elderly driver being liable for a crash. The results can be used to appraise countermeasures that improve the safety of elderly drivers at stop-controlled intersections.

Language: en

Keywords

At-fault crashes; Elderly crashes; Logistic regression; Random forest; Stop-sign-controlled intersection