If you have put much faith in the crash ratings...

As an example. The best view is at :50 into the video: http://www.youtube.com/watch?v=NwtvhkRBxQM&feature=related

This test labels the car "Good" in a frontal impact. There are, however, some pretty scary things happening:

Note that the A-pillar bend in the impact, which causes the door to pop open.

The windshield also shatters, sending little bits of glass everywhere.

Also, look at the position of the dummy and how poorly it's motion into the airbag is controlled.

And on the final close-up, it looks like some piece of the interior trim goes flying (?) out of the car. Little pieces of the car formerly attached to the car turning into projectiles is generally frowned upon.

Compare the video above to this one, then tell me which car you'd rather be in (apologize for the annoying commentary). http://www.youtube.com/watch?v=yBmZF0pUv9c

As a comment, both of these cars are rated "Good" in this test. You can see how much variance there is just in the top rating...if anything, I'd say there needs to be a greater distinction between cars being tested.

Windshield popped because the passenger bag blew. If you look you can see the central location of the cracks centered around the passenger dash bag. All the cars I;ve seen both online and in life (I sell lots of crash parts) typically pop the windshield when the passenger bag blows.

The impact point was sheer luck. Driver lost control on black ice on a turn.

Broken windshield means pretty much nothing. There's a reason they use laminated glass. Yeah, you may have a few shards of glass to pick out of your hair, but that's not really a big deal. I mean, you could be dead instead.

That Civic in the video above is a larger car than the Elantra. The A pillar is also a lot closer to horizontal.

The door didn't open, the A pillar buckled. IIRC, an open door puts you at poor. Sure, that seal broke, but jeez, I don't think you're going to care too much about a draft when you would be dead if you were in a 40 year old car of the same size.

I agree that I'd rather have more training than more airbags. The problem is when the crash was caused by someone who had their first stroke behind the wheel. I'm not saying a billion airbags will keep me safe, but they help.

bravenrace wrote:

Zomby woof wrote: More importantly, does anybody (here) really buy a car based on crash ratings? I personally don't know anybody who does.

Then why are you making such a point of it?

The same reason I post anything.

To stir up some E36 M3.

I commute in a Swift, and not even a newer one. I wish my car was as safe as the previously mentioned, dangerous 95 Metro.

I think accidents are relatively rare, and serious ones even more so.

Zomby woof wrote: I think accidents are relatively rare, and serious ones even more so.

I used to subscribe to that line of thought...then I lost my mom in an accident. It can happen any time, anywhere to anyone. Period.

Derick Freese wrote: I agree that I'd rather have more training than more airbags. The problem is when the crash was caused by someone who had their first stroke behind the wheel. I'm not saying a billion airbags will keep me safe, but they help.

It doesn't take much to have a bad accident. Pilot error put me in the hospital for a month. i never thought i would have an accident, especially a bad one but it can happen to anyone.

I look at the ratings for the cars my wife and child will ride in...

Joe Gearin wrote: If our "leaders" were really concerned with saving lives, they would make drivers ed 1000 X more difficult. In many U.S. states drivers education is not needed to get a driver's license.

This.

Last I looked, Japan and Germany required a mandatory 2 or 3 week full on driving school. This costs two to three grand. Only serious applicants, please. Of course, it doesn't mean they don't have their share of "bottom of the class" and "fast and furious" superstars.

Crash test ratings are only a guide, just like Carfax and Consumer Reports reliability ratings. They should not be considered gospel.

The IIHS is not the happy sunshine and lollipops organization you may think that is interested in looking out for you. While they have done some good, they have an agenda (being funded by my employer and most all of their competition). The 'science' behind their tests isn't necessarily as high as you'd expect. The guy who runs the operation has a giant ego and lives to be on 60 Minutes every year, and he also believes driver training has no effect on safety.

The IIHS tests have simply created a scenario where manufacturers build cars that will perform well in their tests. Does it mean a safer car for you and me? Somewhat. Does it prove one is markedly safer than another? Not necessarily. There are so many variables in auto accidents you can't build a car (for it's occupants) to survive every impact. I can tell you that it is remarkable what people will walk away from these days, provided they are properly restrained inside the vehicle where the equipment can do it's job.

As far as the safety ratings, why does everyone keep referencing what the car looks like?

I thought the safety ratings indicated how well a passenger was protected in a crash, not how screwed the car looks afterward.

I'm sure someone will correct me if I'm wrong.

ddavidv wrote: Crash test ratings are only a guide, just like Carfax and Consumer Reports reliability ratings. They should not be considered gospel. The IIHS is not the happy sunshine and lollipops organization you may think that is interested in looking out for you. While they have done some good, they have an agenda (being funded by my employer and most all of their competition). The 'science' behind their tests isn't necessarily as high as you'd expect. The guy who runs the operation has a giant ego and lives to be on 60 Minutes every year, and he also believes driver training has no effect on safety. The IIHS tests have simply created a scenario where manufacturers build cars that will perform well in their tests. Does it mean a safer car for you and me? Somewhat. Does it prove one is markedly safer than another? Not necessarily. There are so many variables in auto accidents you can't build a car (for it's occupants) to survive every impact. I can tell you that it is remarkable what people will walk away from these days, provided they are properly restrained inside the vehicle where the equipment can do it's job.

Now that's some good info.

If you want to take the real-world into account for crash testing, you need to include driver demographics. For example, the Honda Civic and Toyota Corolla perform similarly when smashed into a barrier at 55 mph, but statistics will show that many more Civics are involved in crashes making them seem less safe.

We all know however, that only cautious old people drive Corollas and they never get up to 55, and Civics are always driven by reckless 16 year olds and never go below 75. With these real world statistics in mind, the cars are in fact equally safe.

ddavidv wrote: Crash test ratings are only a guide, just like Carfax and Consumer Reports reliability ratings. They should not be considered gospel. The IIHS is not the happy sunshine and lollipops organization you may think that is interested in looking out for you. While they have done some good, they have an agenda (being funded by my employer and most all of their competition). The 'science' behind their tests isn't necessarily as high as you'd expect. The guy who runs the operation has a giant ego and lives to be on 60 Minutes every year, and he also believes driver training has no effect on safety. The IIHS tests have simply created a scenario where manufacturers build cars that will perform well in their tests. Does it mean a safer car for you and me? Somewhat. Does it prove one is markedly safer than another? Not necessarily. There are so many variables in auto accidents you can't build a car (for it's occupants) to survive every impact. I can tell you that it is remarkable what people will walk away from these days, provided they are properly restrained inside the vehicle where the equipment can do it's job.

Agreed that any kind of lab testing is an approximation but as a guide it can be a good one. Yes, it's impossible to approximate the many variables in a real-world, crash, but I think IIHS and NCAP come a lot closer than NHTSA for instance. And just because a lab test isn't 100% real-world accurate doesn't mean it's of no or even little value.

Have manufacturers designed to the tests? In many cases, yes they have. Have cars become safer as a result? I would argue that they have. Heck, look at the classic Chevy Malibu vs. the newer car video on Youtube - there's proof right there that cars have changed for the BETTER over the last 20-30 years.

mistanfo wrote: Actually, the test is difficult to replicate, since the ram that moves the car towards the target must disengage before the barrier is hit (so as to not affect the results). No way to ensure that it hits the target at exactly the same point each time. Repeatability is nearly impossible. Of course, most of us don't go around running into IIHS approved barriers.

Having spent years working in test labs as a test engineer, I can tell you that repeatability is paramount and very much achievable. Test like these are easily and routinely duplicated.

pinchvalve wrote: If you want to take the real-world into account for crash testing, you need to include driver demographics. For example, the Honda Civic and Toyota Corolla perform similarly when smashed into a barrier at 55 mph, but statistics will show that many more Civics are involved in crashes making them seem less safe. We all know however, that only cautious old people drive Corollas and they never get up to 55, and Civics are often driven by reckless 16 year olds and never go below 75. With these real world statistics in mind, the cars are in fact equally safe.

Fixed.

foxtrapper wrote:

mistanfo wrote: Actually, the test is difficult to replicate, since the ram that moves the car towards the target must disengage before the barrier is hit (so as to not affect the results). No way to ensure that it hits the target at exactly the same point each time. Repeatability is nearly impossible. Of course, most of us don't go around running into IIHS approved barriers.

Having spent years working in test labs as a test engineer, I can tell you that repeatability is paramount and very much achievable. Test like these are easily and routinely duplicated.

Exactly.

speedblind wrote: ...Have cars become safer as a result? I would argue that they have...

59 Chevy vs 09 Chevy

Zomby woof wrote: More importantly, does anybody (here) really buy a car based on crash ratings? I personally don't know anybody who does.

No. My 60 Beetle is an absolute deathbox and I drive it on the street. Also ride motorcycles a lot, so no I don't pay attention to the ratings. I think most people that pay attention to them are Consumer Reports subscribers.....

I used to subscribe to the thought of it's better to have a good handling car to avoid accidents than a safety laden newer vehicle, but some accidents you just can't avoid. Just weeks ago we got rearended in our new to us 2005 Odyssey. We were stopped in our lane behind a non-moving car and I watched the guy drive right into us at 40mph. I couldn't go anywhere with the guy in front of us and really didn't have time anyway. My 3 year old and 3 month old were in the backseat, but they were fine. In fact I drove it home. When we got the van, I looked at the safety ratings to help convince my wife it was safe, but it wasn't the main reason we got it. Now I will make safety ratings my highest priority when buying future family haulers. I'm just thankful it wasn't one month earlier when we would have been in her 98 Accord as it definitely would not have been as safe in the same wreck. Also, my Dad died in a one car accident apparently trying to avoid an animal and he was an incredible driver in a very good handling B13 SE-R (he had autocrossed it earlier that week). He was thrown from the car (while wearing his seatbelt thanks to the door mounted seatbelts). To those who think it can't happen to them or their driving skills are all that matters, Good luck with that.

And just so you know where I'm coming from, I've commuted to work in a Miata for the last 8 years (until a few months ago). There are just too many bad drivers out there.

In reply to Clay:

Well for sure it would be best to have a car that is both safe and can avoid most accidents. Were your kids in the second or third row of the van?

Clay wrote: To those who think it can't happen to them or their driving skills are all that matters, Good luck with that. And just so you know where I'm coming from, I've commuted to work in a Miata for the last 8 years (until a few months ago). There are just too many bad drivers out there.

I realize it can happen to anyone, and have been in some pretty serious motorcycle accidents that left me hospitalized, but I enjoy riding and driving various vehicles, so accept the risk.

Zomby woof wrote: More importantly, does anybody (here) really buy a car based on crash ratings? I personally don't know anybody who does.

I wonder how many of us secretly hope we go out crashing some insane car we built for a $20XX Challenge. ;)

Of course, that crash would be on a deserted highway or test course and we wouldn't want to take anyone else out with us. But if we get to choose the way we go...

paanta wrote:

Zomby woof wrote: More importantly, does anybody (here) really buy a car based on crash ratings? I personally don't know anybody who does.

Solely on ratings? No. I wouldn't put my wife and kids in something that got bad ratings, though. Your lifetime odds of dying in a car accident are about 1:100. Buying a safe car is one of the better investments you can make. An '01 Passat has about 16 fatalities per million vehicles per year. A '95 Geo Metro has 200 fatalities per million registered vehicles per year. Even taking into account the differences between how an average Metro driver drives and average Passat driver drives that's huge. For what it's worth, the '01-0'04 Elantras fells at about 80-100 deaths per million vehicles. Lower middle of the pack. Not awful. Source

So sayeth the dude with a child seat in a Miata...;)

Joey