30 August 2013

Refuting the "Big Car = Safe" Myth

It is a universally known "fact" that the bigger the vehicle you drive, the safer you are.
Even those who buy small vehicles know this, they just feel that the increase in risk is small, and the benefits to parking, mileage, and cost are worth it.
Like many other universally known things, it just happens to be wrong.
This is extremely easy to prove:  just look at the actual crash statistics, compiled by vehicle weight:

Inline image 1
(the NHTSA website is down, if / when it is restored, I'll post links to the original data)

At first glance this may appear to support the myth: Large vans are at the bottom, with the least crashes, and compact cars are at the top, with the most.
But look a little closer:
Subcompact cars are SAFER than compact cars.  They are even safer than small pickup trucks.
But subcompact cars weight LESS than small pick up trucks, as well as less than compacts.
Skip down a couple more lines: Full size cars are SAFER than full-size SUVs and standard pick-ups, even though on average they weigh substantially less.
But wait, there's more!  Midsize cars actually rank as safer than all sizes of truck, all sizes of SUV, and even safer than full-size cars!
So, if you were rationalizing that SUVs and trucks are only more dangerous than large cars due to roll-over risk, you still have to explain why midsize cars have fewer fatalities than large cars.
Here is similar data, with different presentation: a chart of risk relative to average (100=average) of several vehicle types


      Vehicle Class                Avg. Weight      Relative Fatality Risk
Subcompact (high-risk)             2,000lbs                    143
Sports Cars                              3,200lbs                    142
Compact Pick-ups                    3,500lbs                    123
1/2-ton pick-ups                       4,300lbs                    105
3/4-ton pick-ups                       5,400lbs                    101
1-ton pick-ups                          7,000lbs                    100
Compact cars                           2,500lbs                     96
Subcompact (low risk)              2,000lbs                     85
Truck based SUV                     5,400lbs                     82
Large Cars                               4,400lbs                     75
Mid Size Cars                          3,200lbs                     74
Full-size Vans                          5,000lbs                      52
Cross-over SUVs                    3,500lbs                      48
Minivans                                  4,500lbs                      40
Import Luxury cars                   4,000lbs                     35

http://energy.lbl.gov/ea/teepa/pdf/aps-ppt-wenzel.pdf
This data is from a different range of years, and formatted differently, so there isn't 100% agreement, but it shows the same trend - or rather, lack-there-of: there is absolutely no direct correlation between vehicle weight and risk of fatality.
Even within a single body type: cross-over SUVs weigh less than truck-based, yet have lower fatality rates.
Minivans weigh less than full-size vans, yet have lower fatality rates.
Mid-size cars weigh less than full-size, yet have lower fatality rates.
Notice that the authors of this study divided sub-compacts into two categories, because the range of data points was so wide.  Were they lumped together (as is often the case), the really bad ones would seemingly drag the safer ones down with them, making the entire category look bad, when its really a specific set of them. 
The lower risk subcompacts were found in real world crash statistics to have LOWER FATALITY RATES THAN TRUCKS OF ALL SIZES, up to and including the largest category of "passenger" truck, the 1-ton; which, despite the name, weigh in the range of 3-4 tons, up to 4 times as much as the sub-compacts that are safer than them.

Here is yet more data, in case you like graphs better than charts:
This graph is counting fatalities per crash, so its already assuming a crash occurs:
(from: http://www-nrd.nhtsa.dot.gov/Pubs/808570.pdf )

And this one, specifically for the type of crash where weight matters most: frontal collision with another vehicle


(from: http://energy.lbl.gov/ea/teepa/pdf/aps-ppt-wenzel.pdf )

You can download the original full reports if you want all the details, but the important thing to take away from these you can see at a glance: the dots are all over the place.  There is no trend for the lighter cars to have more fatalities, whether you look at per vehicle, per accident, or even per accident involving another car.


This should be enough.
Case closed.
The data is clear: heavier cars aren't safer.
But of course it isn't so simple.  Not because the facts are complicated, but because the human mind is complicated.
We aren't optimized to think in terms of statistics, we are optimized to think in anecdote.
And so when the most well-intentioned people attempt to study auto safety in order to improve it, even professional researchers fall victim to the same faulty reasoning and assumptions as the general public, generalizing things like "common sense" and "crash test data" to actual real-world risk.
And so, despite what the actual information about the real world clearly shows, even the Insurance Institute for Highway Safety (IIHS) claims

"All other things being equal, occupants in a bigger, heavier vehicle are better protected than those in a smaller, lighter vehicle." 

That sentence stands alone, as though it were universally factually accurate... but then soon after it is qualified by the important distinction that makes it accurate:

"Weight comes into play in a collision involving two vehicles. The bigger vehicle will push the lighter one backward during the impact. As a result, there will be less force on the occupants of the heavier vehicle and more on the people in the lighter vehicle."

(Then, to demonstrate this, they have a graph not of vehicle weight, but of vehicle size relative to risk (the bigger in size, the bigger the crumple zone))



This second sentence is actuate, and it is where all the confusion comes from.
The qualifier is nearly always neglected, but it absolutely completely 100% changes the context and meaning of the entire idea.
A heavier car is safer IN A HEAD-ON COLLISION WITH ANOTHER VEHICLE.
Here's the thing about that, though:
Most car accidents aren't head-on collisions with other vehicles.
In fact, the majority of them aren't.
In fact, the vast majority of them aren't.
Head-on collisions only make up 2% of all car crashes!!
(They make up 9% of fatalities, so even limiting to severe accidents, they are relatively insignificant - 89% of fatal accidents are not head-on)

In comparison, rear-end accidents make of 32% of all crashes.
Collision with fixed objects make up 33.5% of all crashes.
Rollovers accounts for 10%
In a collision with a solid object, like a concrete barrier or a tree, the total deceleration will be the same whether you are in a mini car or a land yacht: essentially whatever speed you were going to zero instantly.
In a rear-end accident, your car gets pushed forward, and instead of jerk, you just get acceleration.  The energy is absorbed by the movement of your vehicle.  You may get whiplash, but you don't get dead.
Consider the most extreme scenario: you are in a passenger car, and you get hit from behind by a 80,000 lbs semi-truck.
As long as you don't aren't pushed into the car ahead of you and get smushed (which you won't, if you leave proper following distances) the fatality rate is only 0.34%
Even if you get hit from behind by a vehicle that weighs 40 tons, about 20 times more than your car, you have a 9,966 in 10,000 chance of survival.

http://www.fmcsa.dot.gov/facts-research/research-technology/analysis/rear-end-crashes.htm

Least you think that's only because rear-end collisions only happen at low speed, when its the cars hitting the semi's, the fatality rate is about 4 times higher - for the car, running into the semi is almost as bad as running into a brick wall.  When its the semi hitting the car, the car gets pushed forward, the movement absorbs the kinetic injury, and everyone is happy and smiling.
This shows two things: 1) Rear-end accidents are rarely fatal regardless of size differential, and 2) the direction an impact comes from absolutely does change whether weight matters.
If the mass differential of a semi-truck - 80,000lbs - vs a car - 4000lbs - is so insignificant, what do you think the impact of mass is on a rear-end collision between a 2000lb car and a 4000lb car?  The answer is none.
In a side impact the situation is similar: the impact force in tangential to your momentum, and so the amount of momentum you have is irrelevant.  Simple thought experiment: whether you are speeding or at a stand still, getting broad-sided will impact you equally hard.  Momentum doesn't matter.  But then why would mass? 
In fact the IIHS themselves - the very people who state categorically on their consumer website that "heavy cars are safer", say explicitly elsewhere on their consumer site:

"Unlike frontal crash test ratings, side ratings can be compared across vehicle type and weight categories. This is because the kinetic energy involved in the side test depends on the weight and speed of the moving barrier, which are the same in every test."
In other words, in a side-impact crash, your car's weight makes absolutely no difference.
But side-impacts make up 23% of crashes, and 18% of fatalities (compared to 2% and 9% for frontal crashes), so this rather undermines their own claim about the impact of weight.
In roll-overs, too, weight does nothing to improve your chances.

Fixed objects, rear-ends, side-impacts, roll overs - in 98% of all crashes, extra weight does literally nothing to protect you.
How much sense does it make that simulated two-vehicle frontal impact tests are the standard for "crash testing", when it is one of the least common types of crash?

So what about those few times you are actually on a high-speed undivided back-country highway, and some drunk crosses over the double yellow lines?
Even then, weight is not the most important factor.

KE=½ mass X velocity²
Kinetic Energy= (1/2 of mass) X (Speed squared)
The impact of your relative speeds is squared (multiplied by itself).  The impact of weight is divided by two.  The speed you are going is overwhelmingly more important than how heavy your car is.


Ok, so...
If there is all this evidence that weight really doesn't matter that much to safety, then why does everyone - even people who's entire job is analyzing car crashes, keep repeating the same myth?
Perhaps for the same reason you, the reader, are still not convinced.
Because, on a purely intuitive level, this belief feels like it makes sense.
It is an extension of the (equally false) assumption that being in a car is safer than being on a motorcycle: "because the steel cage protects you".
How could you not feel safer in a nice strong cage than exposed to the world?
Here's the thing about that though: a steel cage does not "absorb" the crash energy.  It TRANSMITS it.  It transmits it through the steel structure of itself, and on to you.  The stronger it is, the more effectively that force is transmitted.  Think about the "Newton's cradle" desk toy:








5 steel balls on strings, the first one is given a swing, and when it hits the rest, the force travels right though them to the last one.  The last one takes just as much impact as it would if the first hit it directly, because the others are solid, and the force just goes right through them.  It doesn't even matter that the 3 in the center have a combined mass of 3 times as much as the two on the ends.  The one on the end is in no way "protected" by them, as they don't "absorb" any of the force, they simply transmit it.

Because it feels right intuitively, and because it is repeated as a given almost universally, no amount of text is going to help people understand the error of this belief.
They say a picture is worth 1000 words, so a moving picture has got to be worth even more:







Here is perhaps a less abstract way to think of it:
Imagine that, instead of being inside a car, you are standing in front of one that is parked.  It is parked in neutral, with no parking brake on, but it is on perfectly flat ground, so it doesn't move.  You are just standing there, minding your own business, when a truck comes along and runs into the car.  Imagine how this will affect you.  Is the mass of the car going to somehow magically absorb the impact energy and make it go away?  No, of course not, its going to start moving forward.  And then its going to hit you, with close to as much force as if the truck had just hit you directly.









But if the mass of the car doesn't do anything to protect you when it is fully between the truck and your body, why would it do anymore to protect you if you are inside of it?
The answer is that it doesn't.  Like the Newton's cradle, the steel frame of a car simply transmits the force of any impacts on to you.

This is why modern cars have crumple zones.  They are deliberately, by design, weaker than the solid steel tanks of the past. Of course they aren't arbitrarily weaker - the human containing cabin is made stiff, while the front and rear are made soft on purpose so they take the impact energy.
Take, for example, this crash test between an old tank of a car and its modern descendant (which, incidentally, is about 200lbs lighter)






http://youtu.be/mJ5PcWziXT0



But even with crumple zones, having a steel cage doesn't do much to keep you safe on its own.  Another major difference between the old and new cars is seat belts and airbags.
Seatbelts and airbags don't actually protect you from the car that crashes into you.  They protect you from you hitting the inside of your own car.  The entire reason for having seat belts and airbags is to protect you FROM the steel cage you are riding in.

Again, this may be easier to fully grasp in cartoon form:











Given that a car has seatbelts and airbags (and that you actually use them) to protect you from the steel and glass of your own car, having strategically placed crumple zones outside of a stronger solid frame around the passenger compartment creates infinitely more safety in a crash than increasing mass, as shown in the crash test above, where the slightly lighter car completely obliterates the poor crash test dummy in the older car.

But even with two misunderstandings corrected, we are still looking at the entire question the entirely wrong way!

Because we are still thinking in terms of how survivable the passenger compartment of our car is, in the event of a severe crash.
This means we are treating severe crashes as though they are inevitable. In the real world, of course, since about 98% of accidents are caused by human error, its fair to say that nearly all accidents are avoidable.  They shouldn't even be called "accidents", because it makes it sound like its just some random thing that happens.  Really the VAST majority of auto collisions are due directly to negligence, on the part of one or both drivers.  If everyone drove below the speed limit, left large following gaps, refrained from alcohol and drugs, avoided all electronic distractions, and focused on driving safely, the fatal accident rate would drop from the single largest cause of accidental death to fairly negligible.  Combined with proper maintenance, it would be barely above zero.
So if severe accidents aren't actually inevitable, maybe instead of just focusing on likelihood of surviving an accident, it would make more sense to factor in the risk of getting into an accident in the first place.
Ask yourself: Which would you rather do, crash and survive, or not crash in the first place?

So then you have to wonder, what factors might reduce the chances of getting into a crash?
Well, imagie you are going 60MPH in a 5,500lb Ford Expedition on a rural highway, and a truck pulls out from a cross street 140 feet ahead of you. If you instantly apply maximum brakes (ignoring reaction time, which is the same regardless of vehicle) you are going to slam into it at roughly 35mph, the same speed that crash tests are conducted at, and enough to cause very serious injury.
If, however, you were driving the 3000lb Ford Focus, and were in the exact same situation, you would be able to come to a full stop a full 26 feet in front of the truck.






All other things being equal, smaller cars tend to have better braking distances, more maneuverability, and frequently better 360 degree roadway visibility for the driver compared to a larger vehicle.
Comparing trucks and SUVs to cars, due to their higher clearance, are far more likely to roll over, an event with a higher risk of fatality than most accident types.
In addition to all those factors making them capable of avoiding accidents better, the lack of (false) perception of safety may encourage drivers of small cars to take fewer stupid risks (which are, ultimately, the cause of almost all accidents).  The very fact that people feel safe in big vehicles make them do more stupid stuff, like speeding and reckless driving, than the drivers of smaller vehicles.  It's called risk compensation - and its counter-productive when the assessment of risk is completely wrong.


Extra mass only comes into play in a helpful way in 2% of crashes.  In the other 98% it is neutral at best - but in some percent, it is almost certainly a contributing factor - not only because of worse braking distance and handling, but also by encouraging drivers to drive worse.  In that last 2% mass helps, but not nearly as much as people assume.
This myth has been a significant driver of the trend of average passenger vehicles on the road to get heavier and heavier, as consumers pick cars that feel "safe", fueled by crash tests ratings being treated interchangeably with "safety", and official proclamations by official agencies.  One thing that is shown consistently in the statistics is that heavier cars and trucks are definitely much more deadly on average to the people they hit.  So the net effect is more traffic fatalities overall.   This is more than just counterproductive.  It is tragic.
Every time you here this myth repeated, think about the cartoons above.  Think about the graphs and charts. 
And don't let the myth influence your next car purchase.

2 comments:

  1. Safety features (crumple zones, airbags, ...) are the biggest factor. The reason that small trucks are so dangerous is that these features have only very recently started to be implemented, many years later than normal cars. It also implies more generally that new cars tend to be safer than old cars.

    ReplyDelete
    Replies
    1. The driver is the biggest factor. The vast majority - at least 98% - of accidents could have been avoided if both drivers were driving safely, legally, and carefully. The single biggest one is speeding, followed by communication devices and alcohol/drug impairment.

      However, given an accident actually occurs, yes, you are correct, safety features which reduce the impact force are the biggest factor is reducing injury.

      Delete

If you ask a question, I will answer it.

NEW: Blogger finally put in a system to be notified of responses to your comments! Just check the box to the right, below, before you hit "publish"