Thread by @KostelecPlan

We have to get bureaucracies off their deadly fixation that 94% of crashes are caused by human error and thus the reason why fixing human behavior is the answer to traffic fatalities. It has created a barrier to real safety interventions in the US. Let's go threading...

Since the 1930s we have used the "human error is the cause of crashes, thus the cause of deaths on road." This has been used to deflect attention away from safe systems approaches.

(Evans, L. 2004. Traffic safety, Bloomfield, Mich: Science Serving Society.)

This fixation with perfecting human behavior because of this 90% statistic was recognized as false and misleading by those at the forefront of #VisionZero in Europe nearly 50 years ago.

In the US, however, it has been a leading motivator of normalizing all crashes as just one data point among many since now we say 94% of crashes are caused by human error. No reason to differentiate between fender benders and deaths because it's all human error, not design.

It's also intellectual laziness on the part of state DOTs and state traffic safety offices. "Studies show that deficiencies in the traffic environment and vehicle system are the main cause of approximately 63% of deaths (Stigson, Krafft, & Tingvall, 2008)

Well now, that finding shows that the management of kinetic energy in crashes is the key to saving lives. Human error does not influence the amount of kinetic energy transferred to the human body when roads are designed for high speeds.

"This finding implies that a different road and vehicle design, which improves the tolerance of human beings to external violence, would mean that at least 63% of all deaths could be avoided." Deaths aren't reduced by reducing human error. They are reduced by safe system design.

Where do we know this works? Roundabouts. Roundabouts reduce kinetic energy by systematically managing the speed of traffic entering and exiting them. MnDOT found they drastically cut fatalities with roundabouts even though property damage crashes rose.

Oooh, minor property damage crashes rose. So what? That's what we want. It reveals that our fixation on reducing crashes overall and blaming them on human error is exactly the opposite of what we need to be doing.

Mother Nature is a great systems engineer for road safety when she teams up with Old Man Winter. In snow, crashes go up but fatalities go down. That's #VisionZero. Our state DOTs readily admit this (example below from one state crash summary report) but can't connect the dots

Our fixation on human error has permeated our legal and justice systems without much reconsideration of the validity of it. Highway agencies can't admit that their design might be leading to more deaths or they become liable. It's much easier to use human error as a shield.

This latest tragedy on I-84 in Boise reveals the institutional biases that contribute to the false human error narrative. Like in many bike/ped crashes, the truck driver is blamed for suspected distraction even though he's dead & can't defend himself. idahopress.com/news/local/rep…

Of course, it can't be the fault of the road agency who was in charge of the construction zone. Why not? Because they "followed Federal standards."

NEWS FLASH: Just because you followed the standard does not mean it was safe.

I'm working on confirming with the reporter to see if the police investigating this crash have eyewitness accounts or video to indicate that the driver was "inattentive." Or is it just another instance of them saying "it can't be the road, so we must blame the driver."

And that's exactly how we get to blaming human error for 94% of crashes. This man's death becomes a statistic that only adds to that false narrative without any real examination of the systematic failures that might have occurred that led to it. Bureaucracies carry on.

Foundational elements of this thread came from this. Worth a read if you care about real #VisionZero and not falling for the false promises that correcting human behavior is how we'll get "Toward Zero Deaths" tandfonline.com/doi/full/10.10…