Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

IIHS Picks Safest Vehicles



  • Statistically speaking, based on driver death rates compiled by IIHS, the answer is the Mercedes E Class. See:
  • carlisimocarlisimo Posts: 1,280
    Woah, great find.

    So SUVs do tend to be significantly better in multiple vehicle crashes, but "make up" for it in single vehicle and rollover accidents. Midsized and large 4-wheel drive SUVs do very well overall (on par with minivans), whereas small and very large 4WD SUVs, small and midsized 2WD SUVs, and most pickups have death rates very close to those of small sedans (~100 to 110 deaths per million registered years).

    Also interesting: "In almost every size group of two-door and four-door cars, for example, the death rate of the worst vehicle was at least twice the death rate for the best vehicle."
    (Different story among midsized SUVs: 12 for the 4Runner, 119 for the Land Rover Discovery II, and 134 for the 2-door Explorer. I didn't leave a digit out of the 4Runner's rate.)

    It also sounds like they DID adjust the rates to account for driver behavior, by using 25-64 year old women as a reference point to normalize to. The Pontiac Firebird is still one of the worst despite that.

    Ooh, here's a tidbit that'll make good bait for our Townhall arguments, given how often people bring up the Civic as a danger to the driver and all of humanity in general:

    The Civic coupe had a lower death rate than the Excursion and 2WD Tahoe. It also beat the Taurus, Malibu, Grand Am, Sebring, S-type, and Caravan. And the Volvo S40 as well (!?). The sedan was a little better!
  • Thanks - glad you got some use out of it. You're right about SUV's although there are some interesting exceptions to the trend - notably the 4Runner, Lexus RX300 and a few others. I have some firsthand experience here - my wife's previoius vehicle, a Ford Expedition, rolled with my family inside. Fortunately it was a very low speed accident and noone was seriously injured. The roof, however, experienced major crushing. This led me into car safety research, especially as it relates to SUV's. There is some interesting material out there on roof crush resistance of SUV's. The current US federal standard on roof crush resistance is believed by many to be sadly inadequate. Alot of lawsuits going on: Obviously, Toyota, Lexus, and Volvo have figured out the importance of combining active safety features such as stability control with passive devices like roll-triggered curtain airbags and greater strength in the roof pillars; Chevy and Ford - I'm not so sure. . . (even Ford and it's Volvo subsidiary have allegedly had some disagreements on this).
  • carlisimocarlisimo Posts: 1,280
    The best SUVs seem to be car-based, but not without exception. I'm a little confused about the 4WD vs 2WD designations... 4WD is 2WD almost all the time, isn't it? I don't think a lot of those larger SUVs have full-time AWD, so the difference in rates surprises me.

    The domestic cars made a poor showing for these years (2000-2003). They'd better improve by the next survey.
  • calhoncalhon Posts: 87
    The report is generally useless for comparing individual vehicles for two reasons.

    First, there isn't enough data for the individual vehicles to give statistically valid results, i.e., the margin of error is too large. For example, the range of possible values for the E-class ("top" of it's class) is 0 - 22, while the range of the "bottom" Jaguar S-type is 14 - 123. The true values could be 16 for the E-class and 15 for the Jaguar.

    Second, demographics (e.g., differences in driver age, driver behavior and driving conditions) have not been taken into account.

    The problems are readily apparent in the dramatically different results for essentially identical vehicles. Examples:

    Pontiac Grand Prix 73
    Oldsmobile Intrigue 93 (+27%)

    Toyota Corolla 93
    Chevrolet Prizm 128 (+38%)

    Ford Crown Victoria 53
    Mercury Marquis 83 (+57%)

    Buick LeSabre 60 ("best" large car)
    Pontiac Bonneville 97 (+62%, "worst" large car)
  • You're right - again. As you probably know, the car-based SUV's have lower centers of gravity combined with features like stability and traction control that the older truck-based SUV designs lack. Yes, 4WD is 2WD unless "switched" into 4WD mode. The difference really is surprising, I agree. I guess even 4WD makes a difference, even though it has to be deliberately engaged. However, It does look like they combined 4WD and AWD into a single category.

    I agree, the domestic makers are lagging behind in safety. When it comes to safety, you often get what you pay for. Stories abound of car makers asked to cut costs - and what do they cut? Metal - from the frame and roof pillars. Using lesser (and thinner) grades of steel is a simple cost-cutting measure that most consumers will never notice - unless of course the roof crushes in on you. Some of the foreign makers test their vehicles to higher standards - especially in the case of roof crush strength.
  • Your point about the stats is a good one. However, I don't find the report useless for comparisons - only to be used with caution. You're right, some comparisons are tricky due to statistical overlap, but others are more clear cut - where the confidence intervals don't intersect.

    Consider the Chevy blazer 2dr with a calculated rate of 308 and lower 95% confidence interval of 190 (and an upper one of 426!!). Certainly one can compare this vehicle, as an example, to a Toyota 4Runner with an upper 95% confidence bound of 29. It is statistically established with a very high degree of confidence the 4Runner is a safer vehicle. I certainly wouldn't put my family in a 2 dr blazer.

    Also, the ranges are based on a 95% CI. Selection of a different CI, such as 90%, will cause the intervals to diverge further. As I'm sure you know, statistics rarely prove anything - they simply provide a probability-measure of how likely a hypothesis is to be true.

    Demographics were taken into account. See the note on page 10. The data were normalized based on the death rates of women aged 25-64. I don't know how they did this so I really can't defend it - but at least in some fashion it was done.
  • calhoncalhon Posts: 87
    The adjustment for women is just about the only demographic variable taken into consideration. There are several other factors that should be considered, such as driver age, geographic distribution (relevant to weather/road/traffic condition, laws and enforcement which in turn affect driver behavior such as speeding and seat belt usage), miles driven, etc.

    Those very important factors have been ignored. There is no greater proof of their importance than the results for rebadged cars. How else could the same car (LeSabre/Bonneville) have both the lowest and highest death rate in its class? The bottom line is we cannot tell whether the observed differences in death rates are due to demographics or the cars themselves.

    Hence the report is generally useless (i.e., utterly unreliable) for comparing individual cars. It can be used though for comparing aggregates, such as large cars versus small cars. Such aggregates help to cancel out demographic differences and also reduce the statistical margin of error.
  • The IIHS acknowledges the influence of demographics and so do I - but it is their judgment, presumably based on their closeness to the data and study design, that the observed differences are due in significant part to vehicle differences as opposed to demographics. I would also argue that the death rates appear to match up quite well with actual crash test results - which would de-emphasize the impact of demographics. Would you really attribute the blazer vs. 4Runner differences predominantly to demographics? Or better yet, compare the 2 dr vs. 4 dr blazer - striking death rate differences but no logical demographic explanation. What a difference a few inches of wheelbase apparently make. Again, I think you can still glean useful information from the numbers in spite of the demographic question.

    As far as the rebadged cars are concerned, I agree - I wouldn't put too much stock in the comparison. But, the vehicle differences may be more than meets the eye. Here are a couple of possibilities, aside from demographics, that may explain the death rate differences:
    1. Are the option packages (i.e., side and curtain airbags, stability control) identical and are they sold in identical proportion for the model as a whole. This is a potentially critical difference.
    2. Are the vehicles sold with identical tires? If not, there may be some influence depending on the relative difference in safety/traction performance.

    Your point is well taken, particularly for the statistically-close vehicles, and thanks for making me look hard at the numbers, but would you really choose a Blazer over a 4Runner?
  • carlisimocarlisimo Posts: 1,280
    Given that the statistics are real, hard numbers, I'd say there has to be an explanation. It could be luck and coincidence, but that's only likely for cars that crashed in small numbers.

    You're right that we'd be jumping to conclusions if we said one car was safer than other car that was just slightly worse in the statistics. That would be the reader's fault, not the statistician's. So let's work on coming up with the explanations.

    We've all seen the crash test photos of the F-150 Crew Cab cabin collapse. That one shows that a variant of a vehicle can be much less safe than its other variants. The Mercedes A-class was prone to rollovers until stability control was added - that could account for differences in rebadges, especially where one brand is more upscale and likely to include stability control or side airbags.

    Besides, we're supersticious in nature. If someone told me I could have either a LeSabre or a Bonneville, and more people had died between 2000-2003 in the Bonneville, I'd take the LeSabre...
  • calhoncalhon Posts: 87
    There are at least three components to the differences in death rates: differences in the vehicles themselves, demographic differences and chance. In a minority of cases, such as the 4Runner vs. Blazer, the confidence intervals do not overlap, so from a purely statistical point of view we can say the difference is not due entirely to chance.

    However, chance aside, we are still left with the question of how much of the difference is due to demograhics versus the vehicles themselves. I would say the 4Runner is probably better than the Blazer, but how much better? .... I cannot tell. So, I agree, the data is not totally useless for individual vehicle comparisons.

    As for the LeSabre/Bonneville, ABS, side airbags and DRL were standard and ESC optional on both. Note that differences in take rates on optional features is a function of demographics. A potential buyer would want to compare the safety of a LeSabre with (without) ESC to a similarly equipped Bonneville.

    I would say the wear condition, proper inflation and rotation (i.e., demograhics-related items) have a much greater effect than the specific type of all-season tires originally sold on the LeSabre/Bonneville. Besides, many of the original tires had already been replaced. Replacement tire choice is yet another demographics-related item.
  • Agreed. It will be interesting to see what IIHS does next in terms of factoring in demographics. I've enjoyed the discussion and have to ask out of curiosity based on your knowledge, do you work in the automobile industry? Few people outside the auto or insurance industry would give the IIHS report such critical thought. I don't work in the industry; for me this is just a personal interest, although my degrees and work experience are in mechanical engineering.
  • prosaprosa Posts: 280
    There are at least three components to the differences in death rates: differences in the vehicles themselves, demographic differences and chance. In a minority of cases, such as the 4Runner vs. Blazer, the confidence intervals do not overlap, so from a purely statistical point of view we can say the difference is not due entirely to chance.
    However, chance aside, we are still left with the question of how much of the difference is due to demograhics versus the vehicles themselves.

    Which brings up the question of whether there actually are significant demographic differences among the buyers of comparable vehicles. Is the information publicly available detailed enough to allow for these determinations?
  • calhoncalhon Posts: 87
    I have no connection to the automobile industry, but I work in a field involving customer behavior, demographics and statistics.

    Thanks for the great discussion.
  • calhoncalhon Posts: 87
    There are undisputed demographic differences because the vehicles appeal to different tastes and dealers are distributed differently around the country.

    Camry and Mazda6 buyers differ in age and driving style. Volvo and BMW owners differ in age, driving style and geographic distribution. The distribution of Japanese vehicles is more skewed towards the coasts and metropolitan areas than domestic vehicles.

    Vehicle registration data provides some of the information needed for these determinations. Other differences, such as driving style (aggressiveness, risk-taking, etc.) and maintenance levels are more difficult, even impossible, to quantify. However, that doesn't diminish their importance.
  • rshollandrsholland Posts: 19,788
    This article first appeared in Firehouse magazine, which is geared towards emergency responders. It has been reprinted with permission in the summer 2006 issue of DRIVE, which is the Subaru owner's magazine.

    Roof cave-ins of SUVs and pickups first came to my attention with the Ford Explorer tire fiasco of a couple of years ago. At that time images of rolled Explorers with severely crushed roofs were plastered all over the news. As it turns out this roof-strength issue is not just a Ford problem, but can be found on many vehicles, but it is especially evident with pickups and SUVs.

    Subaru has shown here that having a strong roof is not cost-prohibitive and that you don't have to have roof pillars as thick as building girders in order to be safe.

  • jim314jim314 Posts: 491
    The higher crash frequency of 2-dr vs. 4-dr versions of the same model is a long observed phenomenon. It's got a clear demographic basis. At least in the past, younger, less experienced, and more reckless drivers choose the 2-dr style.
  • prosaprosa Posts: 280
    Caliber: Front - good, side - marginal
    Camry: both good
    Optima: front - good, side - not tested
    Zephyr - both acceptable
    RAV4 - both good
    Tuscon/Sportage - both acceptable

    None got the "best pick" awards.

    My comments: this is bad news for Lincoln. A newly designed near-luxury vehicle should've gotten "good" in both categories. No excuses. Also, with all the money and effort Chrysler put into designing the Caliber, the "marginal" in the side test has to be viewed as a major failure.
  • 210delray210delray Posts: 4,722
    Agree totally on the Zephyr. The Caliber's side score could likely be improved once the optional side torso airbags are finally made available. (These should have been available AND standard from the start.)

    At least the Caliber is way better than the Neon it replaced, which was marginal in the frontal test and poor in the side test, without side airbags.
  • jeffyscottjeffyscott Posts: 3,855
    The Caliber's side score could likely be improved once the optional side torso airbags are finally made available.

    It was marginal with the side air bags.
  • 210delray210delray Posts: 4,722
    It had side curtain airbags (standard) that did a good job protecting the front and rear dummies' heads.

    The problem was that the driver's torso was pummeled - no side thorax bags were in the tested car. Very few cars the IIHS has tested do well if thorax bags for the driver aren't included -- the Chevy Impala is one of these exceptions.
  • jeffyscottjeffyscott Posts: 3,855
    Ahh, I see. Thanks for clarifying that.

    I had not noticed that detail in IIHS ratings before. I had only ever noted if they have side air bags or not, without really paying attention to the type(s) of bags.
  • backybacky Twin CitiesPosts: 18,907
    BUT Sentra has more standard safety features (same ones the Renault Megane uses in Europe - and it got 5 out of 5 stars in EuroNCap crash tests for both front/lateral) than both of them...

    Untrue. Rabbit/Jetta are class-leading in standard safety features, which include 6 airbags (2 more are optional), ABS with EBD and 4-wheel disc brakes, traction control, and active front head restraints. As noted, ESC is optional, very rare in this class (Corolla has optional VSC, but it's hard to find). Also, both the Civic and Elantra have more standard safety features than the Sentra. Civic and Elantra have standard ABS with EBD in all trim lines. Sentra only has ABS standard on the highest trim line. Additionally, Elantra has standard 4-wheel disc brakes on all trim lines. All these cars have 6 standard airbags and standard active front headrests.

    I found it hard to believe that C/D rated the new Sentra last, behind even the old Corolla, in its comparo for December's issue. I can't wait to get it to find out why they would do that.
  • v_dv_d Posts: 89
    Rabbit/Jetta are class-leading in standard safety features, which include 6 airbags (2 more are optional), ABS with EBD and 4-wheel disc brakes, traction control, and active front head restraints

    Isn`t it the same as a SL Sentra?

    And we might not get ESP on the Sentra here in Canada, but well...

    LAST? Well, it`s people`s tastes that are going to decide.
  • Are you freakin kidding me? the rabbit is outstanding if someone else is gonna pay for it. the maintenance is ridiculous. and of course it has 57 airbags, because if you get in an accident you will need all 57. the sentra has a ton of safety features, whether they are standard or not, it's not much to add. not to mention it has CVT, which in it's self makes it better in every aspect to drive which is the part that really matters. i'm begging you not to buy a rabbit or an accent, or an elantra. The civic is also a very nice vehicle, And the one thing you will get with the Sentra and the civic is residual value. which is a joke with the other manufacturer's.

    ps...... and yes it is true that the sentra has more standard safety features
  • backybacky Twin CitiesPosts: 18,907
    No, I'm not kidding you.

    The Jetta and Rabbit have excellent crash-safety scores--even without all 57 airbags. The new Sentra is untested. Have you found out how much it costs to add ABS to a Sentra 2.0S? Go configure one and see for yourself. It's not cheap.

    Your saying that the Sentra has nore standard safety features doesn't make it true. It's not. It doesn't lead in optional safety features either. And it's easy to demonstrate those facts to yourself. If you care about the truth.

    BTW, why in the world would you BEG someone not to buy a car like a Rabbit or Accent or Elantra? Do you get a cut on every Sentra sold (maybe on Civics too)? :surprise:
  • v_dv_d Posts: 89
    Yeah, someone on a Rabbit forum here actually had a list of vehicles that got the exact same scores the Rabbit did, and if I recall right, there were a couple of Mercedes', Lexus', Volvo's and other luxury brands. I think the Rabbit was actually the single car under $50k out there. So if you want safety, I`ll go with the Rabbit.

    ABS (not a Pack - standalone) in Canada is standard on the 2.0S model, but it costs $500 on the 2.0. In the US it`s $600 but you get 16" + ABS. And yes that`s a lot for something that should be standard on every vehicle by now.

    A base Rabbit vs a base Sentra in airbags will come 6-6. True the Rabbit has 2 additional rear side airbags, so that comes to 8. Pretty amazing!

    Coming to buying... no I would not buy an Accent, or a Civic, Elantra or even Sentra, I`d take the Rabbit for it`s driving dynamics and german engineering. But again, that`s just me!
  • backybacky Twin CitiesPosts: 18,907
    Actually the Civic gets better IIHS crash scores than the Rabbit--the Civic is an IIHS "Gold" award winner, the Jetta and 4-door Rabbit "Silver." Still very good though. If the Rabbit had better fuel economy and reliability, it would probably be my top choice in this class (in 2-door form).

    When I configured a Sentra 2.0S with ABS, not only did I need to select the ABS+alloys package, I also had to select the Convenience package. So adding ABS was about a $1200 bump--of course you get some extra stuff too.
This discussion has been closed.