Driving lessons can be a gauntlet of fire — for parents and children. As a parent and an expert in driving risk factors, Feng Guo has strong opinions about when and how kids should be driving: as early as safely possible with many, many hours under supervision

“Operating a vehicle can be overwhelming at first,” said Guo, Patricia Caldwell Faculty Fellow and professor in the Department of Statistics. “They need an experienced driver in the car to remind them of critical details.”
In addition to his experience parenting young drivers, Guo formed this expert opinion after 17 years researching traffic safety as the lead data scientist at the Virginia Tech Transportation Institute.

In 2014 and 2017, he coauthored papers in the New England Journal of Medicine and the Proceedings for the National Academy of Sciences, respectively, that provided quantitative proof that using a phone while driving increases crash risk, especially for young drivers.

A decade later, phone usage seems like an obvious risk factor, but back then, there was very little data to support this claim. Guo and his team leveraged the Virginia Tech Transportation Institute’s naturalistic driving dataset, 30+ years of data collected from in-vehicle cameras and sensors, to evaluate risk factors during the seconds leading up to a crash.

Guo has plied this dataset to answer questions such as how to measure driver distraction (eye glance is a good indicator), how long you can look away without substantially increasing crash risk (about two seconds in a 15-second period), and the ebbs and flows of risk throughout a driver’s life.

man stand in auto garage with car and equipment in background
Feng Guo. Photo by Virginia Tech.

In his more recent work, Guo is conducting statistical analyses to evaluate the safety of automatic driving systems, incorporating machine learning and artificial intelligence models to process much larger quantities of real-world data.

As a statistician, Guo doesn’t make any assumptions about the safety of a new technology until he can make sense of the implications. 

“That’s why statistics is so important,” Guo said. “You can have new technologies, new concepts but ultimately, we need to quantitatively verify that claim or hypothesis — which can be a lengthy, complicated process.”

But there are some questions to which Guo has no qualms providing a quick, decisive answer — such as ‘should the driving age start after age 18?’

“Absolutely not,” Guo said. “Kids should be getting hundreds of supervised hours behind the wheel. After 18, that opportunity is no longer guaranteed.”

Contact:

Share this story