Okay, great. So lets make a real comparison then instead of comparing the smart to "every other car" out there.
Lets pick a competitor to the smart...say, a Toyota Yaris. If' I'm not mistaken, people tend to cross shop these vehicles, right?
So lets make the same assumptions used for the smart, even the low miles driven. Oh wait, we have to know how many people have died in a Yaris. Well, I'm not sure how comprehensive Bob's research has been, but when I did a quick web search, I couldn't find that info for the Yaris or the smart. So how can we even make such an assumption in the first place? Hard data is just not there....it's mere speculation. Maybe someone who has access to insurance company data would be able to find real world numbers for us. Until then, the data given doesn't seem to have any real credibility.
I think perhaps a more accurate way to look at this (if it were even possible) is to try and determine total miles driven by specific vehicle and then calculate deaths per mile driven. Sure, that's what Bob is doing, but the problem is he's doing it for the smart only and then lumping every other vehicle together into one giant blob of data in order to prove that the smart is safer than the status quo. Maybe it's just my statistics background, but it sure seems like a severely flawed way to come up with numbers.
Okay, great. So lets make a real comparison then instead of comparing the smart to "every other car" out there.
Lets pick a competitor to the smart...say, a Toyota Yaris. If' I'm not mistaken, people tend to cross shop these vehicles, right?
So lets make the same assumptions used for the smart, even the low miles driven. Oh wait, we have to know how many people have died in a Yaris. Well, I'm not sure how comprehensive Bob's research has been, but when I did a quick web search, I couldn't find that info for the Yaris or the smart. So how can we even make such an assumption in the first place? Hard data is just not there....it's mere speculation. Maybe someone who has access to insurance company data would be able to find real world numbers for us. Until then, the data given doesn't seem to have any real credibility.
I think perhaps a more accurate way to look at this (if it were even possible) is to try and determine total miles driven by specific vehicle and then calculate deaths per mile driven. Sure, that's what Bob is doing, but the problem is he's doing it for the smart only and then lumping every other vehicle together into one giant blob of data in order to prove that the smart is safer than the status quo. Maybe it's just my statistics background, but it sure seems like a severely flawed way to come up with numbers.
Actually you can do a search of fatalities by vehicle make, model, year and a whole host of other variables by going to the FARS database site that Bob provided a link for above. That contains the federal government's NHTSA traffic safety database, the definitive source of national and state-by-state traffic statistics.
But I think you're missing Bob's point. If I'm not mistaken, I don't think Bob's goal was to compare the smart with each comparable vehicle out there for the purpose of showing, for example, that the smart is safer than the Yaris. I think his goal was simply to refute the idiots out there who claim that the smart is a deathtrap, plain and simple. Bob tries to refute this by showing that there have only been two deaths associated with the 35,000 smarts sold in the U.S. since 2008, and I tried to bolster his argument by showing the deaths per 100 million miles (the standard metric) and compared it to the national average. That's all the analysis was intended to do; we were not trying to show the smart is safer or less safe than a Yaris under real world conditions, but that the smart is not the death trap some ignoramuses (ignorami?) would have you believe.
Okay, great. So lets make a real comparison then instead of comparing the smart to "every other car" out there.
Lets pick a competitor to the smart...say, a Toyota Yaris. If' I'm not mistaken, people tend to cross shop these vehicles, right?
So lets make the same assumptions used for the smart, even the low miles driven. Oh wait, we have to know how many people have died in a Yaris. Well, I'm not sure how comprehensive Bob's research has been, but when I did a quick web search, I couldn't find that info for the Yaris or the smart. So how can we even make such an assumption in the first place? Hard data is just not there....it's mere speculation. Maybe someone who has access to insurance company data would be able to find real world numbers for us. Until then, the data given doesn't seem to have any real credibility.
I think perhaps a more accurate way to look at this (if it were even possible) is to try and determine total miles driven by specific vehicle and then calculate deaths per mile driven. Sure, that's what Bob is doing, but the problem is he's doing it for the smart only and then lumping every other vehicle together into one giant blob of data in order to prove that the smart is safer than the status quo. Maybe it's just my statistics background, but it sure seems like a severely flawed way to come up with numbers.
Hmm. Being that the two known deaths in smarts seem to have occurred in 2009, and the FARS database doesn't seem to post statistics until the end of the year, I guess we'll just have to wait until 2010 and then cross-reference the two models' statistics in the FARS db. We could also cross-reference other models like the mini, the fit, and the kia rio.
At any rate, being that every time a smart is crashed it is highly played up by the media, I think our stats for smart fatalities are not that far off. My guess is at worst the smart has an average number of deaths per number of cars and number of miles. If this is true, it is hardly the "death trap" so many make the car out to be.
Lets pick a competitor to the smart...say, a Toyota Yaris. If' I'm not mistaken, people tend to cross shop these vehicles, right?
I disagree with this, right off the bat. The Smart is a city car, an A-series car as I think they're called in Europe. It's got no direct competitors to compare it with until we get the Toyota iQ or Fiat 500.
And I'm really sad to see you're polluting another thread, btw.
Actually the smart is smaller than the A series (Mercedes A class, as an example.) I think Bob and smart enuf are on target. As far as comparisons to other cars - too many variable IMHO: road conditions, time of day, medical event, impaired driver, etc. Unless you can group all the accident fatalities by similar circumstances the comparisons are pretty much apples and oranges. The recently expired thread on the smart running the stop sign is a good example: smart driver dies in crash - since we aren't the local cops or the county coroner, we don't know what he died from - the crash itself or possibly a heart attack just prior to the intersection.
Last edited by jwight; 08-22-2009 at 09:04 PM.
Reason: inserted word
Let me explain how I was able to come up with the figure of "2 Smart Car Deaths":
Using Google (news), Google (web), Yahoo, and Bing, I did a key word search on "Smart Car", "Accident", ("died" or "Death" or "killed"). These are all the words we would expect to find in any account of an Smart Car accident with someone who was killed.
This took me days to sort through all the bad hits I got, but in the end I kept landing on the same two stories of real deaths. The coverage appears in both online news papers and TV Stations.
Now, one could make a case that maybe there was a death or deaths I missed. Possible, BUT given the dislike for the Smart Car and the thinking, "It's a Death Trap". I find it hard to believe that a Smart Car death story would be missed. In both cases of the Smart Car Deaths, the news appeared in several places; so I tend to believe that my search caught all the deaths. If anyone believes that there are deaths I missed, they are free to prove it and list the deaths with the links.
Potential Errors In The Model:
As a number have pointed out, the model isn't perfect. In my model, 8,378 cars = 1 death in 1 year. Granted that there are accidents where more than one person in the car dies as a result, but most of the time people are driving solo. With the Smart Car one could have either 1 or 2 people in the car, so up to 2 deaths are possible. With other cars: some sports cars are limited to 1 or 2 people, but most other cars could have 3, 4, or 5 people and some even more.
It would take a lot of research to make the corrections, but my feeling is that the difference would be minor. Again, if someone feels differently, prove it!
The issue of millage traveled in a Smart vs. other cars; we could also add the type of driving and the kind of drivers. This type of question is very difficult to resolve without very extensive and expensive research. However, if the average for all Smart Car owners was 1/2 the mileage of the other cars per year, the correction for the number of deaths would match the expected average for all other cars. At what point to we consider the number excessive, 6 deaths when there should have been only 4, Smart Car drivers would need to drive 1/3 the miles per year as other drivers. How about 8 deaths, with 1/4 the millage per year. As you can see, the numbers start to look silly:
Quote:
Keep in mind that cars typically accumulate about 12,000-15,000 miles per year.
1/2 = 6,000 - 7,500 miles
1/3 = 4,000 - 5,000 miles
1/4 = 3,000 - 3,750 miles
I'll be the first to admit that my model is not perfect, BUT it does give us a ballpark idea as to the general range of expected deaths vs. number of real deaths. Right now, it's coming in low, NOT high as the "Death Trap" supporters are saying.
Bob, first of all kudos for doing the work on this.
Second, it seems the real issue here is what Stephen Colbert calls "truthiness" - not what is real, but what "feels real." So, you're trying to use factual data to refute others' conclusions based on what they "feel" about the safety of the smart. It's small so it must be unsafe - just like those itty bitty Formula 1 cars.
So, many thanks for doing the research but I'm doubtful any of the closed minded will be motivated to "open their minds."
With so few smarts on the road just one or two deaths can really skew the statistics. It's similar to how Concorde went statistically from being one of the safest planes ever to a very dangerous one as the result of its only fatal crash back in 2000--and it had been flying for 30 years!
Our smart's small size has psychological implications. This is reflected by some of the coverage we see in the media and in the attitudes of people we encounter. How many times have you seen the photo of that Ford Escape that's pinned between two trucks that's portrayed, always incorrectly, as a smart car? I'd be willing to bet you'll see things like that 10 times more often than, say, that recent article/photo showing the smart holding up rather well against 9,000lbs of roof pressure in a recent test.
Last night I observed a couple of fellows eating outside at a restaurant near my smart car. Although a pane of glass separated us, based on body language and hand gestures, they seemed to be under the impression that driving such a car would be crazy. Yet, both guys cast admiring glances at some Harleys that rode past. Both rather heavy set men were also chain smoking during their entire conversation.
(For future reference, here's a helpful hint to any smart-bashing bubbas out there: Harleys, chain smoking, and pot bellies will probably ALL kill you far sooner than any smart car.)
In statistics, we learn that 2 is not a sample. Putting aside all of the humanity, the numbers are what we're talking about here. We simply don't have enough data to say the smart is safe or isn't based on death counts. If we had 200 people die, we'd still question the variables: seat belt, age, health, position, speeds involved, distracted driving, glare from someone's back window. This means that a sample size of two is more significantly uninformative based only on news reports. The only way to get the safety analysis is to have full data from each incident, and a sample size of several dozen or hundred. And I sincerely hope we never get that data.
Bob's analysis is appreciated - it helps answer my question that the car itself is not a death trap; if it were, I'd expect to see 50-100 deaths where the national average is 4.
Conclusion: The car's extremely safe for its size. It'll handle a rollover better than anything else on the road. It's nimble, and for any accident it doesn't avoid, it's got safety features that have proven themselves in dozens of nonfatal cases.
And reading the comments on that USA Today link - most of the posts use phrases like "I imagine", "looks like", and "it might". All armchair critics. Not a single one has done the research themselves. Nobody's driven one or been in an accident where the safety systems proved themselves.
People love to jump on the bandwagon of saying something sucks. "Look at that stupid little car." "My ____ would crush it." "Yeah, I bet it would blow over if a truck passed it." "Coors!" "Budweiser!" "Sam Adams!" "My dad can beat up your dad!" It doesn't matter what you're talking about, someone out there believes they're superior.
My own confidence while driving is unfased by the noise around me. I KNOW my car is safe and I bought it for ME.
The AutoGuide.com network consists of the largest network of enthusiast-owned enthusiast-operated automotive communities.
AutoGuide.com provides the latest car reviews, auto show coverage, new car prices, and automotive news. The AutoGuide network operates more than 100 automotive forums where our users consult peers for shopping information and advice, and share opinions as a community.
1976 Rover Mini
2003 Chevrolet Corvette Z06
2007 Mercedes-Benz E320 Bluetec (SOLD)
2009 smart Passion (Gone)
2011 smart Passion
Today
Linear Mode
