We often hear the statistic that “94% of all motor vehicle accidents are due to human error.” It’s an oft-repeated fact and the cornerstone of many arguments for getting more autonomous vehicles onto our roadways.

But is this “fact” true?

Short answer: No. Long answer: it’s complicated.

The “94%” does indeed come from real research. In a 2015 data report the National Highway Traffic Safety Administration (NHTSA) noted that the “critical reason [for a crash], which is the last event in the crash causal chain, was assigned to the driver in 94% of the crashes.” [emphasis added]

But it’s important to note that the NHTSA was careful to clarify that any auto accident is a complex set of interlocking circumstances. They never intended for anyone to assume that the “critical reason” for a crash was the sole cause of the accident. As they noted in their report, “although the critical reason is an important part of the description of events leading up to the crash, it is not intended to be interpreted as the cause of the crash.”’ The report went on to add “…[I]n none of these cases was the assignment intended to blame the driver for causing the crash.”

A more accurate summer of the NHTSA position would be “human error contributes to 94% of all traffic accidents.”

Unfortunately, misreading that statistic as “94% of all motor vehicle accidents hinge on human error”  has resulted in years spent trying to improve roadway safety almost solely through efforts to stop people from making bad decisions without considering other factors. For example, as this diagram illustrates, a skilled and attentive driver in 2022 is significantly more likely to clip a cyclist now than they were in 1980, simply because roadways offer a much narrower margin of error for the most popular vehicle types.

If "Share The Road" feels more difficult/dangerous than it used to, it's because it is. pic.twitter.com/4ioRTEMOAx

— @jdavey (@jdavey_2) June 6, 2022

Widening the scope of our inquiries into what additional factors contribute to an accident (and in what ways those factors contribute to crashes) promises to expand the tools we have to increase roadway safety overall.

Driving Research Finds Practical Solutions to Reduce Crashes

At Oregon State University Department David S. Hurwitz and his team are using networked vehicle simulators–including immersive bicycle, CMV (commercial motor vehicle), and car sims–to explore how a mix of users (cyclists, pedestrians, couriers, drivers, etc.) share the road in real-world situations. The end goal is to develop evidence-based design processes for safer vehicles, roadways, and traffic systems—whether it’s better understanding how cyclists behave in traffic, what makes teens more likely to drive distracted, or when drivers are most likely to run red lights.

How do Hurwitz’s labs conduct driving research to collect that data? Ever since our collaborative installation of OSU Driving and Bicycling Simulator Laboratory in 2010, they’ve used FAAC RTI (Realtime Technologies Inc) simulators to rapidly develop and deploy new studies. This driving research simulation solutions don’t simply run scenarios, but also track and coordinate data both from the sim itself, as well as third-party software and hardware (such audio/video recordings and ASL Mobile Eye Tracker hardware).

This kind of dynamic driving research uses simulation to extrapolate from observable data, in order to untangle the complex interaction of technology and infrastructure design on choice and human behavior. This holistic approach is especially timely, considering worrying post-pandemic traffic safety trends.