What Can A Hazard Function Teach Us About Drivers’ Perception of Hazards?

Hazard perception (HP) is typically defined as the ability to read the road and anticipate hazardous situations. Several studies have shown that HP is a driving skill that correlates with traffic crashes. Measuring HP differences between various groups of drivers typically involves a paradigm in which participants observe short videos of real-world traffic scenes taken from a driver’s or a pedestrian’s perspective and press a response button each time they identify a hazard. Young, inexperienced drivers are considered to have poor HP skills compared to experienced drivers, as evident by their slower response times (RTs) to road hazards. Nevertheless, though several studies report RT differences between young, inexperienced and experienced drivers, other studies did not find such differences. The authors have already suggested that these contradictory findings may be attributed to how cases of no response—that is, a situation where a participant did not respond to a hazard—are being treated. Specifically, the authors showed that though survival analysis handles cases of no response appropriately, common practices fail to do so. These methods often replace a case of no response with the mean RT of those who responded or any other central tendency parameters. The present work aims to show that treating cases of no response appropriately as well as selecting a distribution that fits the RT data is more than just a technical phase in the analysis. This work used simulation of predefined distributions and real-world data. It was demonstrated that selecting the appropriate distribution and treating nonresponse cases appropriately affect the shape and characteristics of the density, survival, and hazard functions. The suggested process has the ability to provide researchers with additional information regarding the nature of the traffic scenes that enables differentiating between various hazardous situations and between various users with different characteristics such as age or experience.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01706055
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Apr 18 2019 3:01PM