By on June 17, 2006

rollover2.jpgEver since the United States began issuing safety and emission standards, regulations have led to better cars.  Emission standards forced automakers to develop electronic engine controls, creating modern cars’ power and drivability.  Safety standards– seat belts, airbags, etc.– have saved countless lives.  But there’s one standard that’s not only ineffective, but antithetical to its stated goal: the National Highway Traffic Safety Administration’s (NHTSA) roof crush standard. 

NHTSA enacted FMVSS 216 in the mid-70’s.  Before then, unbelted passengers were often ejected during a crash and killed or injured by “external factors.”  It was assumed that roof strength played an insignificant role during serious accidents.  As seat belts gained popularity, NHSTA created its new roof crush standard to ensure that vehicle roofs wouldn’t crush belted occupants during a rollover.     

Rollover test.jpgFMVSS 216 requires all light vehicles sold in the United States to withstand a force equal to 1.5 times the vehicle’s unladen weight or 5000lbs. (whichever is less), with no more than five inches intrusion.  NHTSA’s technicians apply a measured force to one side of the roof’s edge, at a 25-degree angle (sideways) and a five-degree pitch angle (forwards), then measure the deflection, or failure, of the roof structure.  As a result of this requirement, car makers have strengthened their vehicles’ roof structures.  It would be easy to posit that the standard was– and is– an excellent piece of federal legislation.  A closer examination of the evidence suggests otherwise. 

The legislation’s framers knew that convertibles would never be able to meet the new roof crush standard.  So they exempted them.  At the time that the standard was created, convertibles were becoming increasingly unpopular.  NHTSA figured that extinction would make the exemption a moot point.  When convertibles made a comeback, the exemption provided valuable statistical data.  Unfortunately for the rule-makers, real world experience didn’t validate the roof crush standard’s original premise.  Passengers in convertibles during rollover accidents were not killed in any greater numbers than occupants of cars that met NHTSA’s roof crush standard.    

This empirical data has been ignored in the face of the increased lethality of rollover accidents.  Although rollover crashes constitute just three percent of all American accidents, they account for almost half of all fatalities.  That’s an estimated 26,376 deaths annually. But the assumption that a weak roof collapsed and crushed the passengers is simply not supported by the data. 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. 

Illustration3.jpgThe best way to meet NHTSA’s roof crush standard: increase the strength (and weight) of components at the top of the car, well above the center of gravity.  This adds significantly to an SUV’s “top-heavy” character, and contributes greatly to their propensity to roll when subjected to large lateral forces.  In other words, the requirement to make roofs stronger to withstand rollovers may well be making vehicles more likely to rollover.  NHTSA’s roof crush standard is creating a vicious circle; people may be dying because of the law– not despite it.    

Safety campaigners like “Protecting Health, Safety & Democracy” are lobbying the federal government to revise roof crush standards “upwards.”  They want roofs to be “stronger,” so they can withstand greater impacts.  Although it’s theoretically possible to find light weight materials that would withstand insult better than steel, in the real world, more stringent legislation would result in even more weight up top.  And that would mean that vehicles already more prone to rollover accidents will be made even more prone to rollover accidents by the increased stringency of the rollover protection standard.  

SSF_web.jpgThe past three decades have shown the basic futility of the roof crush standard. Yes, rollovers are a deadly and common factor in American road deaths.  But regulating the parts of the car that don’t tend to cause death or serious injury has little effect on real-world results.  If rollover protection is the goal, there are better ways to save lives.  Enforcing seat belt laws would have a major depressing effect on rollover fatalities.  Electronic stability systems have proven enormously effective at preventing rollovers from occurring in the first place.  Suspension systems that automatically lower the car’s center of gravity also have great potential  

In any case, data gathered over the last three decades shows that it’s time to repeal NHTSA’s roof crush standard. The increasing trend towards taller vehicles, with a greater propensity to roll over, argues for the repeal of this standard completely, before it does even more harm.  While safety campaigners argue that the revised roof crush standards don’t go far enough, NHTSA should fully and scientifically investigate the likelihood that the exact opposite is true.  The rule-makers must admit the possibility that they’ve made a mistake, and find a new and more effective fix for a deadly problem.   

Get the latest TTAC e-Newsletter!

16 Comments on “The Counterintuitive Truth About Roof Crush Standards...”

  • avatar
    Jonny Lieberman

    Correct me if I am wrong (being a punk with a server and all…), but aren’t cars’ safety stars for rollover determined by sticking the vehicle on a platform and rotating said platform sideways until the thing tips?

    So, something like a Vette which if 46 inches off the ground would get a 5-star rating cause the thing never tips over?

  • avatar

    Click on the pics, Jonny.

  • avatar

    It seems to me there was a time when high center-of-gravity
    vehicles that we’d now call SUVs came with “roll bars” since
    they were *expected* to roll (they were also 4WD – most
    modern SUVs are 2WD, eliminating the _utile_ part of
    “SUV”). I’m guessing the only reason that’s not seen
    anymore is because these vehicles are now marketed to
    people who really need station wagons and who might be
    scared off by knowing that a high center of gravity carries a
    rollover risk. It follows that any problems automakers face
    from rollover-related regulations are creations of their own

  • avatar
    Bob Elton

    Stars smars. In real life, people aren’t being killed in rollover accidetns in Chrysler Sebring Convertibles any more than they are in Chrysler Sebring Coupes. The coupes meet the roof crush standard, the convertibles don’t. So why have the standard?

    Bob Elton

  • avatar
    C. Alan

    From a structural stand point, the amount of steel added to the roof pillars is more than likely very minimal. Most of the reinforcement are in the roof pillars, and the point of connection to the roof line. In a roll over situation, the position of the mass of the engine and driveline would have a much greater influence on the likly hood of the car overturning that the relitively small mass of a reinforced roof.

    I say keep the roof standards as they are, removing them are not going to make SUV’s any safer.

  • avatar

    Bob, I get this alarmist vibe from your stories. Where are you getting this data? The read any more authoritatively than “my friend’s uncle’s brother’s friend says …”.

    The best way to meet NHTSA’s roof crush standard: increase the strength (and weight) of components at the top of the car, well above the center of gravity. This adds significantly to an SUV’s “top-heavy”? character, and contributes greatly to their propensity to roll when subjected to large lateral forces. In other words, the requirement to make roofs stronger to withstand rollovers may well be making vehicles more likely to rollover. NHTSA’s roof crush standard is creating a vicious circle; people may be dying because of the law… not despite it.
    Bolded emphasis is mine. Where are the references to back this up? Are you a materials and structural engineer? You sound like one of the uninitiated commenters on this and other sites (like me) who predict the end of the automobile if the Democrats get elected.

    There’s not a single reference to hard data, or quote from a reliable, knowledgeable source, and without those, this smacks of irresponsible journalism.

  • avatar

    Bravo, blalor, for calling out Mr. Elton on this editorial garbage. As I read through it, I kept thinking, “shouldn’t there be hard evidence to back this up? And haven’t vehicles been getting safer and safer as far as rollovers go?” Indeed, a quick google check reveals that the NHTSA revised it’s rollover ratings to be more strict/accurate in 2004, and a few minutes searching by brand at doesn’t yield any vehicle by any of the major manufacturers whose rollover rating has gotten worse over time.

    Granted, more people seem to be dying in rollover accidents, but the increase could very well be attributed to an increase in SUV sales rather than those SUVs being less safe. I don’t know, but I would expect TTAC to do some research instead of just throwing out sensational editorials.

  • avatar

    I urge all readers to click on the photos for hard data. Meanwhile, here’s this from :

    “There has been a sharp increase in convertible offerings in the past decade, and existing roof-strength rules exempt convertibles. Several import automakers have added rollover safety equipment voluntarily. 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.

    “We’ve been asked the convertible question many times, and we don’t see a higher pattern of injury losses,” said Kim Hazelbaker, senior vice president of the Highway Loss Data Institute, a research organization for auto insurers.”

  • avatar

    My issue isn’t that the roof strength requirements are worthless, which they may very well be, but Mr. Elton’s assertion that the requirements may be causing more fatalities than they’re preventing. That’s an inflamatory statement that needs to be backed up with evidence.

    Furthermore, the quote that is shown when clicking on the last picture seems to confirm my suspicions that vehicles are becoming less likely to roll over as time goes on, despite added requirements for structural rigidity up top.

    The convertible issue is even more convoluted, since convertibles almost always have lower centers of gravity than their hard top counterparts due to a lack of a roof (obviously) and structural reinforcements lower in the car.

  • avatar
    Bob Elton

    The real issue with quoting NHTSA’s numbers is that they don’t answer the real question, and the basic premise of my article, which is, “How many people are injured by crushed roof during rollover accidetns?” The answer, if you look dep enough into the data, is very few. Rollover deaths are caused, largely, by unbelted passengers being ejected, and then often crushed by the car itself. If the roof is not the problem, and the low numbers of convertible fatalities indicate that that is the case, why spend resources making the roof stronger?

    For what it’s worth, the CG of the old Chrysler K-car convertibles, when I measured back in the 80s, was actually higher, by about 2 millimeters, than the coupe (with the top up). Casual observation ofthe Mustang convertible looks as though all of the heavy parts are added near or above the center of gravity. The convertible top is a heavy piece of equipment. Not the fabric, but all of the mechanisms, framework parts, hydraulic pump and motor, and so on.

    Old cars, like my 64 Imperial convertible with a frame like the Brooklyn Bridge, may have had lower CGs than coupes, but that’s ancient history.

  • avatar

    I find the entire issue of vehicle safety standards interesting. It seems rather odd that we can ride motorcycles that have no crumple zones (unless, like Ben Roethlisberger, you reserve your face for that purpose), no seat belts, no airbags, no side impact beams, yet the government(s) continue to insist on ever-increasing use of nanny devices to protect us in our cars. I suppose I should shut up before bikes are outlawed…

    The roof strength issue reminds of something I learned about industrial forklift trucks. Before rollover protection standards were implemented very few people were seriously injured in forklift tipping accidents. When the use of a rollover protection cage was required people started dying because their instinctive reaction was to jump out of the tipping truck, now to be crushed by the very device meant to save their skin. Of course they do serve as falling object protective devices, which is why they were renamed as such in Canada.

    I don’t know if Bob has presented evidence to justify repealing the standard, but I support him in his contention that it should not be beefed up.

  • avatar

    My question is, is the lethality of rollovers “increased”, or simply “higher than other accidents”?

    In other words, without disputing that a rollover is more likely to kill you than any other accident, is it so that they’re more lethal than they used to be?

    Or is the increased percentage of deaths due to rollovers (the rate) higher simply because other safety improvements have made the total number of deaths (total and/or per-passenger-mile) lower?

    It’s very important to not confuse the total and the rate, when doing such things, especially over time. The rate is what’s really important.

  • avatar
    Dave Skinner

    Sigivald asks a very important question- Are rollover deaths higher due to an increase in EVENTS, or simply an increase in PERCENTAGE.

    Bob Elton touched on this in the text of the article, but did not provide a detailed explanation:

    “This empirical data has been ignored in the face of the increased lethality of rollover accidents. Although rollover crashes constitute just three percent of all American accidents, they account for almost half of all fatalities”.

    Here’s what we know about the issue without cracking a book:

    1) Deaths per million miles traveled have been trending downward since the seventies.

    2) All automotive safety equipment is designed to protect occupants INSIDE the car.

    Based on those two facts, it seems that recent safety innovations work, protecting people who remain in the car, while those persons ejected receive NO protection.

    The same dangerous outside hazards present in 1960 present an equal hazard today (trees, concrete curbs, the rollover car itself, other cars…), so the death rate among persons ejected represent an ever-growing percentage of total fatalities, rather than an ever-growing number of people.

  • avatar

    A few pounds of metal re-inforcement is almost irrelevant compared to the amount of insulation installed in modern cars. European and Japanese cars are up about 25% compared to equivalent cars made ten to fifteen years ago. That is a love-handle of roughly a thousand pounds per car.

  • avatar

    Are increased safety standards why nothing is made with T-tops anymore? I really, really miss those…

    I would hate to see the laws eliminated as I am one of the people that’s alive because those standards were in place. My Geo Prizm retained most of its roof integrity when I decided to park it greasy side up about 10 years ago. Something tells me my old ’72 DeVille project car I owned at the same time wouldn’t have done that.

  • avatar

    And though tough to measure – might it not be that folks in a convertible are conscious that there is little (top up) or no (top down) roof structure over their heads and thus they tone down their driving style to save their necks?

    When I drive vintage cars I am VERY conscious of the fact that there is very little safety equipment in my vehicle. In my VW van I am very aware that only the bumper structure (VERY stout) and the steel dash structure is going to save me in a collision. In between those two parts is a thin sheetmetal skin that may spread the collision forces – but only if something doesn’t puncture it. I have seen alot of crashed VW van pics from around the web and they aren’t automatic death traps but they are much less safe than most of the other vehicles on the web.

    The problem with the SUVs is that people have been sold them on the premise that they are safer b/c they are larger.

    Because I’d like to see more of them go away (the large ones), let them be sold but make it clear to the buying public exactly what happens in hard turns and rollovers.

    Safety campaigns = people without a healthy respect for safety and understanding of risk.

    I truly think that the safety campaign has had a negative result on America. I grew up crashing my bike. Lots of skinned elbows and knees. I quickly grew to respect the physics involved in riding my bike and my awareness that my bicycle was a conscious being trying to kill me. VBG! Alot of kids today have been coddled through the period of their life where my generation was going to the emergency for stitches often and therefore do not have a healthy respect for risk. Same with those same kids when they reach driving ages.

Read all comments

Recent Comments

  • EBFlex: “ cleared ~ 3000 acres of forest FOREST next to the expressway. Guess what s going in -in that newly created...
  • jkross22: People are more angry and stressed than I’ve anecdotally seen in my lifetime. Add in selfishness,...
  • thornmark: >>The article is based on a false premise. There is no reason to believe that batteries that don’t...
  • thornmark: Why All Those EV-Battery ‘Breakthroughs’ You Hear About Aren’t Breaking Through In the superheated market...
  • Goatshadow: Can always count on this site for the delusional whackjobs to show up in the comments. Disappointed there...

New Car Research

Get a Free Dealer Quote

Who We Are

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