Quirks and Quarks

Lethal memory fail: Why drivers see, and then forget, motorcyclists

Driving simulation tests suggest that human memory, rather than negligence, may be responsible for "looked but failed to see" collisions, a new study suggests.

Drivers see motorcyclists, but then their short-term memory can be overwritten by other things, study finds

CORTE MADERA, CA - OCTOBER 16:  A motorcyclist rides between cars in slow moving traffic on Highway 101 October 16, 2007 in Corte Madera, California. Motorcycle deaths are on the rise in California with 433 deaths in 2006, up from 275 in 2000. Officials estimate that deaths are up another 8 percent this year as sales of powerful motorcycle continue on an upward trend.   (Photo by Justin Sullivan/Getty Images)
'Looked but failed to see' errors occur when a driver pulls out into the path of an oncoming vehicle. Motorcyclists and bicyclists are most vulnerable to this type of crash. (Photo by Justin Sullivan/Getty Images)

Originally published on Sept. 28, 2019.

A new study from the United Kingdom offers a surprising explanation for frustrating and dangerous "looked but failed to see" collisions. These occur when the driver of a car pulls out into the path of an oncoming motorcycle or other vehicle, even though they looked and seemed to see them. 

These kinds of accidents may be quite common and lead to tens of thousands of deaths every year, according to new work by Peter Chapman, an expert in driving psychology. Surprisingly, he thinks driver negligence may not always be to blame.

His new research, published in the journal PLOS One, suggests our short-term memory is at fault here. 

"We've always assumed that [the drivers] just looked in the wrong place or were not paying attention," Chapman, who led the study, told Quirks & Quarks host Bob McDonald.

"But our research has suggested something different. What we think is that the car driver is actually seeing the bike but forgetting it by the time they pull out." 

What did you see? 

To understand what's causing drivers to make these critical errors, Chapman designed a series of experiments that made use of the high-fidelity driving simulator at the University of Nottingham. 

The simulator is a BMW Mini housed within a projection dome that provided the participants with a 360-degree view of the simulated driving environment.  

Peter Chapman testing the University of Nottingham’s high-fidelity driving simulator. (Peter Chapman)

He recruited 60 individuals for the study, who were asked to drive past the same road junction many times to measure their risk threshold  — how far an oncoming vehicle would have to be for the participant to pull out safely. 

To study how memory factored into the phenomenon, Chapman gave the participants a surprise memory test only during the trials when the driver decided to pull out in front of oncoming vehicles at their risk threshold — the critical distance between a car and an oncoming vehicle that could lead to a collision. 

Chapman always sent two vehicles from opposite sides of the road during those trials — either a car and a car, or a car and a motorcycle. Then, right after the participant pulled out, he would stop the test and ask them to use a laser pointer to point where the oncoming vehicle or vehicles had been on the screen.

The view to the right of the road junction as seen by the driver in the driving simulator, with the circle marking their point of fixation. (Peter Chapman)

What Chapman found was that, every once in a while, there were instances where the driver had failed to report seeing the motorcycle and the car, even though an eye-tracking device showed that they clearly looked at the oncoming vehicles. But he also found that participants were much more likely to miss the motorcyclist than the car driver. 

"We had three times as many cars as motorcyclists in our experiment," Chapman said. "And yet, we still had five times as many memory errors for motorcyclists as we did for the cars." 

Based on these results, Chapman concluded that many "looked but failed to see" crashes may actually be a case of "saw but forgot" errors, where memory lapse is the problem. 

'See bike, say bike'

Chapman said the results suggest that our short-term memory isn't as robust as we like to think. 

"All memories are lost in a really short time," he said. "We have this picture of the world like we're watching a movie in front of us, and we think we're paying attention to all of it, but we're not. We're just paying attention to one thing at a time, unless we actively remember it." 

He also found in the experiments that a driver is more likely to forget a motorcyclist the longer they spent looking at other things after seeing them, suggesting that our visual short-term memory is limited. 

"Other things [that the driver] looked at later have overwritten [their] memory of the motorcyclist," Chapman said.

Chapman theorizes that the reason that motorcyclists and bicyclists are especially vulnerable to these "saw but forgot" crashes may have to do with our limited peripheral vision. 

After a collision with a car in his early days of riding, Quirks & Quarks host Bob McDonald has switched over to a much bigger motorcycle to improve his visibility on the road. (Submitted by Bob McDonald)

Most of the time, our weak peripheral vision is able to detect an oncoming car that we may have forgotten in these "looked but fail to see" events because they're big. It provides us with an extra safety net in the event where we forgot about the car.

But that extra protection is not available to motorcyclists because they're too small to be reliably picked up by our peripheral vision.

To counter our cognitive shortfalls, Chapman has proposed an intervention that he calls the "see bike, say bike" rule. 

"If you're at a junction and you see a motorcyclist coming, just say 'bike,'" he said. "That puts it into an extra form of verbal short-term memory and you know you'll still remember it's there seconds later."