By on February 6, 2009

There are two main problems with debunking auto-related misconceptions. First, not everyone is ready, willing or able to confront the truth. Second, once you debunk something, it doesn’t stay debunked. TTAC’s Bob Elton dealt with the roof crush standard issue in his editorial “The Counterintuitive Truth About Roof Crush Standards” back in June 2006. He argued that increasing roof strength only increases the number of rollover accidents. Common sense: the higher a vehicle’s center of gravity, the more likely it will roll. Elton also revealed that “In 74% of cases, roof intrusion was not a factor. Rollover accidents are fatal because the occupants are usually ejected, or partially ejected, during the crash.” And that’s because… they’re not wearing their seat-belts. And yet, The Detroit News reports that “The Insurance Institute for Highway Safety [IIHS] said Wednesday it will require automakers to dramatically increase the strength of vehicle roofs to receive its top safety pick ratings.” The road to hell? You don’t know the half of it…

As we’ve pointed out numerous times, the National Highway Transportation Administration (NHTSA) is one of America’s best governmental agencies. When it comes to new regulations, they move slowly because they move carefully.

A great deal of road safety is counter-intuitive (e.g. “sobriety checkpoints” are less effective at removing drunk drivers from the the road than roving patrols). The NHTSA doesn’t rush to judgement to avoid exactly the kind of unintended consequences described above. And here goes the industry lobby group, once again, subverting science and pretending to stick-up for their customers’ safety.

In January, the National Highway Traffic Safety Administration unveiled a proposal to require a vehicle roof to withstand a force equal to 2.5 times the vehicle weight while at the same time maintaining sufficient head room for a buckled-in, average-size adult male to avoid being struck. That’s up from the current standard of withstanding a force equal to 1.5 times the vehicle weight. But NHTSA hasn’t finalized its regulation.

Lund said starting in the fall IIHS will require automakers to have a 4.0 rating to win a top safety pick.

“We see significant safety benefits in stronger vehicle roofs,” Lund said.

“The government is moving slowly and they are going to continue to move slowly.”

He said NHTSA has “clearly undercounted” the number of injuries and deaths that can be prevented by stronger roofs.

How? And how can The Detroit News repeat such a viscious slur without asking for proof? Show me the benefits.

You know me folks; I don’t blindly accept the word of any auto industry pressure groups of any stripe on anything. But on this one, I reckon the Auto Alliance has got it right. The IIHS’ roof crush standard would decrease fleet wide mpg and increase costs to the consumer– without adding significant demonstrable safety to motorists.

“Drivers and passengers are better served by a system of enhancements including improvements in vehicle stability, ejection mitigation, roof crush resistance as well as road improvement and behavioral strategies aimed at consumer education,” alliance spokesman Wade Newton said.

Get the latest TTAC e-Newsletter!

56 Comments on “IIHS Demands Increased Roof Crush Standards. Wrong Answer....”

  • avatar

    Are they coming after my convertible again? I ain’t giving up without a fight.

    Besides, has anyone ever figured out how many more people have been killed since A-pillars grew to the size of redwoods just to house air-bags? The hell with a back-up camera if this makes them even larger you will need a side view camera.

  • avatar

    Oh, jeez. A-pillars are gigantic enough as it is.

  • avatar

    Reporting the roof performance to consumers might be a good idea. Let consumers decide what they want. The only question with reporting is how to represent the numbers. Requiring 4x strength for a “good” rating seems like adding excessive weight (pun intended!) to that one statistic.

    As far as I remember reading, most rollovers are self-inflicted and are highly correlated with high center of gravity. If that is the case, then I see little reason to increase requirements or to even have requirements. Instead, NHTSA should concentrate on something consumers can’t choose for themselves: The amount of danger a vehicle imposes on other vehicles on the road. We could have a win-win situation if NHTSA required that vehicles operating on public roads had a low ride height and center of gravity; this would greatly improve safety for people in smaller vehicles while simultaneously improving fuel efficiency, handling, and rollover resistance for those vehicles that are forced into being lowered. Vehicles designed for combined on-road and off-road use could be built with an adjustable suspension that could raise the vehicle and lengthen the suspension travel it transitions from pubic roads to off-road. The improved safety and decreased cost to smaller vehicles could outweigh the increased cost to the small number of vehicles that need the extra off-road capability.

  • avatar

    Yes… but will this change keep automotive engineers employed?

  • avatar

    They are regulating the automobile out of existence just the way they did nuclear power.

  • avatar

    The IIHS has only one purpose, and that is to save auto insurance companies money.

    They don’t care if they waste people’s time with a 55 mph speed limit, cost manufacturers and consumers money or decrease fuel economy.

    They do care if they are causing more accidents, and therefore payouts.

    As a front for insurance agencies they have massive accident data. And supposedly some decent scientists and engineers to look at it.

    Possibly they are right?

  • avatar


    Anything is possible. But Bob Elton and I spent a long time researching this issue. We spoke to scientists and accident investigators. As far as I can tell, they are very, very wrong.

  • avatar
    Gardiner Westbound

    How do European roof crush standards compare to domestic requirements?

  • avatar

    Robert Farago:

    I usually assume spin lying to support their own bias when I hear something wrong from a corporate interest group, but stupidity is also a possibility.

  • avatar

    I only have experience with building codes instead of automotive regulations, but they work the same way.

    If I were writing the regulations, I’d make roof strength dependent on rollover propensity. The latter would be a simple calculation based on center of gravity and vehicle geometry, but they would have the option of using empirical data instead. Stability control would give them a reduction factor but not completely because they don’t prevent all rollovers.

    So if it’s a rollover-prone vehicle, its roof strength would have to be greater. It’d be a vicious circle… a common code technique to discourage doing something in the first place. And I think it would fit reality. I don’t need my Miata to have the same roof safety as a Suburban – I think it’s fair if I just have the same chance of dying in a rollover in either vehicle (odds of rolling over * odds of dying in a rollover).

  • avatar


    Thanks for bringing my old article back to light.

    This whole controversy about roof crush standards can be reduced to answering a simple question.

    Are the rollover deaths and inuries greater in cars with the current roof standard than in cars that have no roof at all (convertibles)?

    The answer, of course, is no. So if the complete lack of a roof doesn’t cause additional injuries, why would strengthening the existing roof improve things?

    When I was in junior high school, back in the dark ages, we learned the scientific method of looking at evidence. The above question is a direct application of ninth grade science.

    Everything else written on this issue is just blather.


  • avatar

    Can someone explain to me at which point I will not have to install a cage in my racecar because the thing is already a tank?

  • avatar

    Why not use high strength steel? Do pillars need to get bigger, if better steel is used?

    Wouldn’t stronger roofs yield other benefits to rigidity of the frame?

  • avatar

    Bob –

    You miss the point. We’re only talking about cars with a roof; convertibles are a different system subject to different variables.

    You ask: “So if the lack of a roof doesn’t increase rollover death, why would strengthening the existing roof improve things?”

    The answer is because they are completely unrelated. I’ve already decided that I want a roof; now should I get a strong roof or a weak one?

    To answer that question, you need to look at the # deaths in strong-roof cars vs. weak-roof cars.

  • avatar


    That’s like looking for your missing car keys under a street lamp because the light’s better.

  • avatar


    “It’d be a vicious circle… a common code technique to discourage doing something in the first place.”

    I think you have just solved the problem. Why would the IIHS want to do something that would cause tall SUVs to have more accidents, when that will only make its members have to pay out more claims?

    It doesn’t, it wants to discourage tall SUVs all together.

    Between the Rollover Rating System and the stronger roof requirements it will be very difficult to build a tall SUV, and from a clams perspective that’s the way the IIHS wants it.

    Ironic. The SUV won’t be killed by environmentalists. It will be killed by a corporate lobbying group.

  • avatar

    There is one other option besides larger or stronger pillars, and if that is the route they take with this well you will be seeing a lot more cars with dome shaped roofs like my New Beetle…

    Does anyone remember the (now old) new beetle TV commercial about how the dome shape of the beetle’s roof is one of the strongest shapes?

    This could be a whole new kinda ugly for cars sold in the U.S.

  • avatar

    With carbon fiber 5x as strong as steel you could, at some additional cost, make a roof structure with A,B and C pillars that are 1/2 the size and 2x as strong.

    It would be an interesting a design exercise to see what a car like a Mercedes CL or BMW 650 would look like, if the designers had the ability to go with an all carbon fiber greenhouse.

  • avatar


    A big no on the carbon fiber. It can take a HUGE initial hit. But the material literally disappears after that impact.

    Most car accidents, and certainly most rollovers, involve more than one impact. A CF roof would reduce survivability.

    [NB: Ford justified the use of steel in the GT supercar rather than CF on this basis.]

  • avatar

    So how can Volvo and SAAB have such thin roof pillars and be compliant since like 1967?

    Does this primarily affect heavy vehicles, such as pickups and SUVs?

  • avatar

    So, will this still mean they have to meet California emission standards too? Cars gonna get heavier. I would suspect the 300c look of high beltline will also become prominent. I would suspect the smaller the windows, the stronger the roof.

    And I’m sure you’re right about rollovers. But they were already successful in getting VSC/Stability Control installed in everyone’s cars. So that problem is already solved, right?

    On the other hand, its IIHS. Car companies can just choose if they wanna do this or not.

  • avatar

    RF –

    No, it’s like refusing to wear a helmet on your motorcycle and justifying your decision because it doesn’t reduce your risk of drowning while waterskiing.

    Look, clearly the rollover death risk in an F430 Spyder is much lower than in a 1991 Ford Explorer, despite it’s much stronger roof. That doesn’t mean roof strength is irrelevant, it just means one is a sports car and the other is an SUV.

    I completely agree that other options need to be considered (ESP, suspension changes, collision avoidance systems).

    But the convertible vs. non-convertible comparison is bogus.

  • avatar

    Carbon fiber wrapped around steel pillars would help. You still need steel under there – carbon fiber does poorly in compression, but wrapped around steel it would make it harder for the pillar to buckle. Big hoops at the B- and C-pillars that get in the way of big sunroofs would help too. Or just allow for more intrusion. It’s ridiculous how little headroom there is in some of the tallest vehicles for sale. I feel more cramped in a Navigator than in the average economy car.

  • avatar


    With enough money, anything’s possible. But affordability is a valid issue.

  • avatar
    Martin B

    My neck is stuck at a bit of an angle because I was a passenger in a car (Mitsubishi Galant) which skidded sideways and flipped onto its roof 20 years ago.

    I was held upside down by my seat belt and felt my neck bending more and more awkwardly as the roof slowly crumpled while we skidded upside down at 70 mph into the bush.

    Finally the car hit a low earth bank and flipped back on its wheels and we jumped out. I didn’t realise I’d damaged my neck so I didn’t get the proper treatment. We were out in open country far from a hospital anyway.

    So, yeah, you need decent roof strength. But I take the point it would mean extra vehicle weight, and there might be other areas on a vehicle where extra weight might be more effectively used.

  • avatar
    john m flores


    Firstly, I very much appreciate your site and the perspective that it offers. I do have, however, some questions about this story and the analytical conclusions you are drawing from some of the data:

    1 – Regarding the stat “In 74% of cases, roof intrusion was not a factor.” If that is true, then roof intrusion is a factor in 26% of cases, or approximately 2,600 rollover-related deaths.

    2- Regarding the stats on convertibles, namely this quote from, “Although concern has mounted over rollover deaths in SUVs, convertible rollover deaths are a nonissue for NHTSA, the insurance industry and some automakers. The 94 fatalities attributed to 87 convertibles that rolled over in 2004 accounted for fewer than 1 percent of about 10,000 U.S. rollover deaths last year.”

    This quote is stating that convertible deaths aren’t statistically significant enough to warrant concern, not that convertibles are every bit as safe as their hard-topped brethren. The fact that convertible don’t cause more rollover deaths may be attributed to many things, including their relatively low sales volumes and, most importantly, their low CofG and thus low likelihood of rolling over in the first place. And if you look at that number again, every time a convertible rolled over, one or more persons died. I doubt that their hard-topped brethren have the same ratio, but I could be wrong.

    3 – The notion that increased roof rigidity will raise the CoG significantly enough to cause more rollovers. I’d love to see some hard data on this.

    If my job was to reduce rollover deaths, and there was already a lot of activity addressing the 74% – seat belt programs like Click it or Ticket, etc… – then trying to address the 26% seems like a natural thing to do.

  • avatar

    I like Carlismo’s observation. IIHS is an industry group with some valuable data, but we should never forget that they are in the business of taking your insurance premiums and never paying them out (ideally). It does not matter to them if your car is as up-armored as an Abrams tank – and gets similar gas mileage – as long as it helps avoid injuries and payouts.

    I googled high strength automotive steel which led me to Acelor USIBOR1500, where that 1500 is the ultimate tensile strength in megaPascals. In Christian units, that’s about 218,000 psi. The yield is around 1100 MPa or 160,000psi. That is pretty darned strong – it doesn’t sound like you can achieve the increased roll-over strength by bumping the yield and ultimate strengths. Actual performance may be governed by a combination of buckling and yield. Whats left to do is thickness and cross section increases and therefore weight increases. Question – would that add more weight than a sunroof and attendant mechanism? The Acelor pdf I looked at had an A pillar stamping – ballpark of doubling its thickness was four cubic inches or one pound. Six pillars, six pounds? (Edit: make this a closed section from two stampings and its ballpark weight delta of eight cubic inches – two pounds. Six pillars, twelve pounds?)

  • avatar

    While SUVs have a high center of gravity relative to cars, not all of this due to heavy pillars. The bottoms of these things are much higher than most cars too. A lot of this is the BOF thing. It would be interesting to see the stats on CUVs. Given that CUVs are later to market than SUVs in general, CUVs will more likely have stability control and a full complement of air bags. Comparing CUV deaths to those from SUVs might be difficult.

    The Suburban has a rollover rating of 3 stars, the Traverse rates 4 stars. The 4 Runner gets 3 stars, the Sequoia 4 stars, the Highlander 4 stars and the FJ Cruiser gets 3 stars. I haven’t measured the pillars on any of these, but, other than the newly designed Sequoia, the BOF vehicles do poorly relative to the CUV class from what I can see.

  • avatar

    Instead of mandating roof standards why don’t we just try to improve driver’s education? It seems like people are willing to throw the keys at someone who can’t drive and then are justify it by saying it has side curtain air bags, roof that can survive nuclear Armageddon… From my experience it is hard to flip an SUV innless you are doing something really stupid.

  • avatar

    I was a back-seat passenger in a Ford Bronco that rolled when I was a teenager, and I emerged totally unscathed (no seatbelt, sleeping on a rear bench seat and a lot of luck). We crawled (walked out the windshield really) from the wreck and realized immediately that no other vehicle would have flipped over the guard rail at such low speeds (maybe 35mph). So there’s a lot to be said for low CoG in my experience, and this kind of standard would only have made other cars more likely to do the same.

    Also, I’ve noticed that most rollovers I’ve seen have occured on higher speed roads, where the car on it’s back obviously had a 70+mph impact with some obstacle before getting there. I’m seriously not comfortable with attributing all of these deaths to roof collapse. There are a whole lot of ways to die on your way to upside down.

    I also used to drive a Ram 2500 with awful A pillars. When I almost took out a kid in a crosswalk it taught me a serious lesson about patience in intersections, but it also made me realize that my own personal safety isn’t worth that risk. If you’d rather drive up a tree than hit a kid in the road you know what I mean.

  • avatar

    carlisimo, it’s interesting that you mention the Miata. I know a guy who rolled his. I know it’s hard to believe, but true. Don’t ask me how, but he managed. The funny thing, is that nothing happened to him or his passenger. The A-pillars took the impact amazingly.

  • avatar

    Perhaps I should have made the question more specific.

    If people in Mustang GT convertibles made in the model years 1989-1993 suffer no greater chance of injury in rollover accidents than occupants of Mustang GT coupes made in model years 1989-1993, why would making the roof of a Mustang GT coupe stronger make a difference?

    Substitute K-cars, Sebrings, or Corvettes, and you get the same result.

    Controlling for the variables is also part of the scientific method,also taught in ninth grade science class, at least bck in 1960 or so.


  • avatar

    A couple of years ago, a C2 Corvette convertible didn’t make a curve near my house and collided with a very low stone wall in my neighbors front yard. It flipped over the wall and landed upside down on the lawn. I thought the driver was going to be dead, but I was wrong. Although he spent some time in the hospital, he ended up surviving. I’m sure if he had been thrown from the vehicle it would have been a different story.

    How far do you go with vehicle safety? Are we all going to be driving versions of the Obama limo with our speed controlled by roadside transponders to keep us safe?

  • avatar
    john m flores

    @ relton

    If people in Mustang GT convertibles made in the model years 1989-1993 suffer no greater chance of injury in rollover accidents than occupants of Mustang GT coupes made in model years 1989-1993, why would making the roof of a Mustang GT coupe stronger make a difference?

    Will you please point me to the source of this statistic? I find it difficult to believe but will yield to facts.

    All that I have seen up to now points to the low volume of convertible rollover deaths, which is to be expected since there are fewer convertibles on the road.

    I’m looking for mortality as expressed as a percentage. In other words:

    x % of Mustang convertible rollovers result in a fatality


    y % of Mustang hard top rollovers result in a fatality.

    Please enlighten.

  • avatar

    “Drivers and passengers are better served by a system of enhancements including… behavioral strategies aimed at consumer education.”

    Tell me you’re joking!

    @ TEW “Instead of mandating roof standards why don’t we just try to improve driver’s education?”

    Easier said than done. We haven’t invented a course yet that keeps people from crashing! It’s like saying, “if only we could keep those drunks/speed freaks/young ‘uns/old folks/mental retards off the road.”

    BTW, current design vehicles tested by NHTSA that already meet the 4.0 strength-to-weight ratio are the Camry, Jetta, Volvo XC90, Tacoma, and Civic. Vehicles that almost meet it are the Ford Taurus, Nissan Frontier, and Subaru Tribeca. Not exactly rocket science then — and are any of these vehicles more prone to rollover than their class equivalents?

    Also, side curtain airbags for the most part are not housed inside A-pillars, just their tethers (if even that). I think thicker A-pillars are a consequence of laying back the windshield at ever steeper angles rather than for roof strength.

  • avatar

    John M flores

    NHTSA has done just these studies, back in the 80s and 90s when convertibles were making a comeback.

    Chrysler studied this to death inthe 80s. I contributed to that research. Chrysler wanted to be ready to defend convertibles, again, if there were lawsuits.

    The issue was originally studied in the 70s, when roof crush standards wereoriginally proposed. Part of the proposal would have eliminated convertibles. That issue went all the way to the Supreme Court. You can read the decision, and the reasoning associated with it.

    Roof crush, at least in modern cars, is simply a non-issue when it comes to injuries.


  • avatar

    He argued that increasing roof strength only increases the number of rollover accidents.

    I asked for proof of this in 2006, and got nothing. It’s contrary to the evidence that I can find. It was brought up by TTAC at least once in between then and now, and now here it is again.

    About to unsubscribe from my RSS feed.

  • avatar

    MBella, I have a Miata and I’ve heard of people rolling theirs… after all, a lot of us drive through hills where the good roads are. If you mess up and fall into a ditch or down a ravine, you have a good chance of rolling over. A few months ago it happened in the Berkeley hills and the passenger was killed, but they’re not always fatal. Aftermarket rollbars help, but then you risk cracking your head against it in a mere rear-ender.

    John Flores, that statement didn’t mean that there were only 87 convertible rollovers in a year… just that there were that many fatal ones.

  • avatar

    steronz, you didn’t get an answer because strengthening the roof (the pillars really) doesn’t necessarily mean raising the center of gravity of the vehicle.

    On the other hand, the IIHS research shows that the stronger the roof, the fewer the injuries in rollover crashes. There is no evidence showing that the stronger-roofed vehicles rolled over at a greater rate in the first place.

  • avatar

    I still like carlismo’s idea of a simple weight, trackwidth and cg formula determining the required upside down roof carrying capacity in order to earn the Good Housekeeping Seal of Approval. Why punish the rest of us for the inelegant vehicles of some of us? As for convertible drivers, they willingly assume extra risk of some level, same as motorcycle riders do. Insure at an appropriate real-world experience based level.

    210delray – you need to strengthen not only the pillars, but also roof perimeter and cross beams. Still don’t see it outweighing a powered glass sunroof or worse a dual/panoramic sunroof. I can’t see how it won’t increase the cg, just as sunroofs do. But is it significant?

  • avatar

    All the science goes out the window when the evening news shows a picture of a rolled over car with a crushed roof. “Those BASTARDS!!!” If only they had made the roof stronger!!!!

    Remember, if it bleeds, it leads.

    And engineering data, in English, in America, means the (formerly) deep-pocketed Big 3 are fat, juicy, lawsuit targets.

  • avatar

    According to AutoSafety, who quotes “The Federal Government”, one-in-four car crash deaths occur in an accident involving a rollover. You have to dig deeper than this statistic to determine if the accident was survivable at all and to what degree roof crush contributed to the death.

    Public Citizen and Center for Auto Safety say front and side fatalities have decreased 25% but rollover deaths have increased eight-fold. Without knowing the vehicle mix, the total number of accidents and miles traveled, we’re at a loss to properly digest this statistic.

    Both seem to be saying that if you are involved in a vehicle that goes greasy side up, your chances of ending up dead, or worse, paralyzed from the cheek bones down “eating” with a tube shoved up your nose, wearing a diaper continue to climb. Someone posted it’s 100% if you are in a convertible.

    That ain’t good.

    The current testing of 1.5 times the vehicle weight very gently placed on the roof is the result of highly-paid lobbyists sabotaging the testing requirements. They are at it again, offering to increase the weight to 2.5 times the vehicle weight. Just say No.

    The dynamic testing recommended by Joan Claybrook, the Jordan Rollover System (JRS), rates a No also.

    Static tests with weight placed gently on the roof or dynamic testing a full 90-degrees from the direction of travel is being pushed on us by either the lazy, sleazy, dishonest, or the truly ignorant. It doesn’t matter which.

    Roof Crush Testing must be dynamic, like the JRS, but it also must have about a 45mph forward component to it OR IT IS ALL BUT WORTHLESS.

    Unless the forces on that roof and those “A” pillars match typical accidents, you are looking for your keys under a streetlight because the light is better. And that’s about the level of safety improvement you’ll see.

  • avatar

    Roof Pillars have gotten fatter to add strength and to house side curtain air-bags, either in or on them.

    After several near-misses, I learned that in a 2007 Honda Accord, you drive differently or you will hit pedestrians in the crosswalk or stepping off the curb. You cannot just move your head side-to-side and look, you move your whole upper body as much as it takes to get a damn good view around the “A” pillars.

    The pillars also block the turn signal indicators used by the tiny few to indicate a pending lane change. If you are overlapped with the car beside you, you will not spot the blinking light until you drive right into their door as they make a signaled lane change.

    Some vehicles now have indicators on the mirrors and on the front fender behind the wheel. Bless ’em.

  • avatar

    Weak roof supports are just God’s way of getting rid of bad drivers. And no, I am not religious.

  • avatar

    Increased pillar and roof weight contribute to rollover accidents.

    So does raising the radio antenna, lifting that tiny chrome ball ever higher and higher into the air. Calculate the increased risk to me, please.

    As someone pointed out, the weight of the sunroofs and roof racks far outstrip any increase in additional rollover risk. How many pounds do you stuff in your Thume?

    I believe I read where if the Ford Explorer had a track of just 1.5 inches wider, the rollover risk would have been halved. The problem, according to Ford, was that it would cost $350,000,000 to widen it. They must be using AutoCAD instead of CATIA.

  • avatar

    IIHS does have an incentive to maximize safety by minimizing their potential payouts.

    However, IIHS also has an incentive to maximize vehicle MSRP. More expensive cars are generally more expensive to insure.

    Given the huge human stupidity element present in many roll over “problems”, I doubt better roof standards will help all that much.

  • avatar

    As someone mentioned above, Volvo has been advertising their strong roofs for decades. Surely there is data from those cars to back up the claim that stronger roofs result in more rollovers.

    As for convertibles, I’m under the impression, perhaps mistakenly, that the windshield frame on them is heavily reinforced compared to cars with roofs. Today’s sharply raked windshields means the top of the frame is positioned much like a rollbar over the front seats. I also get the impression that many or most of modern convertibles have a variety of fixed and pop-up roll protection devices. I am sure that convertibles are less likely to have rear seat occupants, if they even have rear seats. So probably there are fewer people in convertibles to begin with. And lastly, I’m under the impression that convertibles seem to have lower seats and/or higher beltlines than sedans. Usually all you can see of convertible occupants is their heads.

    So if all this is taken together, it severely compromises any attempt to compare rollover fatalities between sedans and convertibles, for the purpose of comparing tendency to roll over and weight of the roof structure.

    I think there should be a study to see if thick A pillars have resulted in more accidents to begin with, and how the rate of resulting injuries and fatalities compares to the amount of avoided harm resulting from stronger roofs. If the point of the article is that these sorts of things should be not be mandated on an ad-hoc basis, then I agree completely.

    Additional roof strength legislation will also give us more of those funny little A-pillar windows, since that allows bracing the A-pillar.

  • avatar

    Brandloyalty, it’s just the luxury convertibles that have pop-up stuff. Cheap ones like my Miata aren’t particularly strong. The A-pillars aren’t terrible – more for stiffness though, so the windshield stays intact after many years of flexing. Sometimes they mousetrap, other times they hold up just enough. A swoopier rake makes it very difficult to make it act as a roll bar.

    Tendency to roll over is THE reason convertibles are okay.

    Great point on A pillars in general. Not much of it is metal though – it’s mostly airbags, padding, and style too.

  • avatar

    As someone mentioned above, Volvo has been advertising their strong roofs for decades. Surely there is data from those cars to back up the claim that stronger roofs result in more rollovers.

    The data will show there is NO real correlation between stronger roofs and a meaningful increase in tendency to roll over. You have to look three places to the right of the decimal place to find a trace of it. Depending on the vehicle, less than 10 pounds of tin in the right place does the job.

    Task the engineers to design a proper testing procedure free of politics and lobbyists and they will do so. Task the engineers to design roofs with less tendency to collapse in a rollover crash at insignificant cost and weight penalty and they will do so.

    Also stop the automaker’s knee-jerk reaction to cage-fight all safety legislation.

    Patty Murray D-WA sponsored $1M in government funding to get us to JAS, now we need a few more to complete the job.

    As far as the Volvo strong roof claims, you must be remembering the Big Foot drive over the wagons and only the Volvo remaining up. Look closely and you will see a ton of 4×4 cross-bracing the full length of the roof. Volvo paid a nasty fine for that stunt.

  • avatar
    john m flores


    John Flores, that statement didn’t mean that there were only 87 convertible rollovers in a year… just that there were that many fatal ones.
    My bad. Good catch

  • avatar
    John Horner

    “He argued that increasing roof strength only increases the number of rollover accidents.”

    A little bit of fact wrapped up in a pile of conjecture.

    The auto industry safety lobbyists are typically looking for only two things.

    1) How do we reduce production costs enough to justify our lobbying fees and high executive pay.

    2) How do we make not-my-fault arguments when the inevitable lawsuits come our way. Bob Elton’s efforts on behalf of Chrysler in the 1970s seem to be of this paid-to-find-the-desired-answer variety.

    BTW, someone better tell all those race car makers to stop with the silly roll cages already. Don’t they know that those roll cages are going to get drivers killed?

  • avatar
    john m flores


    “NHTSA has done just these studies, back in the 80s and 90s when convertibles were making a comeback.

    Chrysler studied this to death inthe 80s. I contributed to that research. Chrysler wanted to be ready to defend convertibles, again, if there were lawsuits.

    The issue was originally studied in the 70s, when roof crush standards wereoriginally proposed. Part of the proposal would have eliminated convertibles. That issue went all the way to the Supreme Court. You can read the decision, and the reasoning associated with it.

    Roof crush, at least in modern cars, is simply a non-issue when it comes to injuries.”

    Bob, what about race cars is different that compel all race series require some sort of roll cage? Are they mistaken in their need for them, or the conditions of racing that different from the road?

  • avatar

    “He said NHTSA has “clearly undercounted” the
    number of injuries and deaths that can be
    prevented by stronger roofs.

    How? And how can The Detroit News repeat
    such a viscious slur without asking for
    proof? Show me the benefits.” farago

    (Following quotes are from the Insurance Institute for Highway safety’s Status report: Vol.44 No.3 arch 24, 2009)(bold emphasis added)

    “More than 10,000 people a year are killed in rollovers.” (according to NHTSA,an additional 24,000 people a year in the US are catastrophically injured in rollover accidents including brain injuries and spinal injuries and are the leading cause of preventable quadriplegia and paraplegia)

    “‘Our research shows that a strength-to-weight ratio of 4 reflects an estimated 50 percent reduction in the risk of serious and fatal injury in single vehicle rollover crashes compared with the current Federal standard of 1.5,’ Lund explains.”

    “About 25 percent of occupant deaths in crashes of cars and minivans involve rolling over. This portion jumps to 59 percent in SUVs.” (Rollover accidents are only 3% of all US motor vehicle accidents according to NHTSA)

    “When vehicles roll, their roofs hit the ground, deform and crush. Stronger roofs crush less, reducing the risk that people will be injured by the roof itself. Stronger roofs also can prevent occupants, especially those who aren’t using belts, from being ejected through windows, windshields, or doors that have broken or opened because the roof has deformed. Roofs that don’t collapse help keep people inside vehicles as they roll.”

    In specific response to Farago questioning the validity of the articles statement regarding NHTSA’s ‘undercounting’ the number of injuries and deaths that can be prevented with stronger roofs, the article state:

    “‘The agency made too many unproven assumptions about the relationship between roof strength and injury risk,’ say Institute president Adrian Lund. ‘For example, if someone was ejected during a rollover, NHTSA ignored the possibility that this wouldn’t have occurred if the roof had held up better. Based on this one flawed assumption , the agency eliminated more than a third of all deaths as potential beneficiaries of stronger roofs.\'” NHTSA assumed that stronger roofs wouldn’t help people who died in vehicles that didn’t roll beyond the first side, possibly because the roll was arrested by a tree or other object (14% of deaths) or in vehicles that were arrested later in the roll (12%), nor those that were unbelted (15%) or those not in front seats or under the age of 12 (3%)or if their most serioyus injuries weren’t related to roof crush directly over their heads (12%).” ”

    “It’s a long list,” Lund points out. “When NHTSA got to the end of it, hardly anyone was left who met the narrow criteria. The agency disgarded 94% of rollover deaths.”

  • avatar

    I’m new to this so I am not sure what “flaming the site” means.

    I read the lead article and responded with facts directly from the organization sighted. Not sure what the problem with that was but if if I didn’t follow protocol my apologies.

    I lost my only son in a roof crush incident 5 years ago and have devoted more than a thousand hours of research on the subject. I thought others who hadn’t read the article by the IIHS would benefit from there position on the questions posed by the author.

  • avatar

    Leading consumer advocate, Paula Lawlor of the People Safe in Rollovers Foundation ( found damaging internal documents from car companies and put them into an excellent expose report called “Roof Crush Intrusion, Deadly by Design” available for free on her web sight.

    The report documents the history of the roof crush issue from the ’50’s all the way up to current times. An excellent 24 page read for anyone with a serious interest in the subject matter.

    Included is detailed documentation of Volvo’s Industry leading design concepts in roof design.

  • avatar

    I totally welcome increased roof strengths. The Dodge Charger is at a bit more than 20,000 pounds for 5 inches of crush!
    Chevy Volt at about 16,000. Compact cars above 10,000.
    They are using high tech steeles to keep weight down I think and keep pillar thickness down but eventually we probably will need computer screens wrapped around the pillars so the pillars appear opaque.

Read all comments

Back to TopLeave a Reply

You must be logged in to post a comment.

Recent Comments

  • Kruser: My grandmother had an early ’70s Imperial coupe. From my perspective, it had “nicer” (in...
  • bullnuke: Who is talking about Thai? I made a comment on infotainment displays. One should never eat anything while...
  • DenverMike: It’s not relevant. Screens can be a pain, but we’re talking an annoyance. Playing with them...
  • Ol Shel: But will Dodge’s electric muscle cars still be the go-to for driving into crowds?
  • Ol Shel: Humans seek convenience and disengagement, and we fiercely resist the removal of those conveniences. Deaths...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Jo Borras
  • Mark Baruth
  • Ronnie Schreiber