Toyota Accelerator Data Skewed Toward Elderly 776
An anonymous reader passes along this discussion on the data for the Toyota accelerator problem, from a few weeks back. (Here's a Google spreadsheet of the data.) "Several things are striking. First, the age distribution really is extremely skewed. The overwhelming majority are over 55. Here's what else you notice: a slight majority of the incidents involved someone either parking, pulling out of a parking space, in stop and go traffic, at a light or stop sign... in other words, probably starting up from a complete stop."
not enough data (Score:4, Insightful)
27 data points is not enough to draw a strong conclusion.
So . . . (Score:5, Insightful)
parking, pulling out of a parking space, in stop and go traffic, at a light or stop sign... in other words, probably starting up from a complete stop
Or in other words, they take their foot off the pedal and put it on the wrong one.
I trust Woz (Score:2, Insightful)
Woz has already described the repro case.
Now, the iPad may not be the be all and end all of consumer devices, but I trust Woz when he talks.
Here's a question (Score:3, Insightful)
If the vehicle has that much computer controlled functionality, why doesn't the black box tell which pedals were pressed at the time of impact and for the moments before impact? The black box system is arguably an invasion of privacy, but in this case it would go a long way toward fixing the problem(s) and perhaps saving lives.
stastically significant doesn't mean all (Score:2, Insightful)
This assumes there is only 1 problem, not a half dozen different problems occuring in different situations. Yes, there are probably some that are putting their foot on the wrong pedal, that happens with every make and model of vehicle out there. Lets say statistically all cars have some percentage of elderly putting their foot on the wrong pedal, subtract them out and look at what's left. Serious electrical or mechanical issues can be lost in the noise.
Re:not enough data (Score:5, Insightful)
27 data points is not enough to draw a strong conclusion.
So why then should the court of public opinion concluded that it's Toyota's fault?
Re:I trust Woz (Score:3, Insightful)
Yes but the Woz case is possible bug in the cruise control software, not the accelerator.
Right, because Cruise Control Software is in no way related to acceleration, right?
Correlation != causation (Score:3, Insightful)
The real explanation could be as simple as "Those 55 and older are the ones who can afford to buy the cars in question".
Re:not enough data (Score:5, Insightful)
Because GM is owned by the government, and by far the easiest way to gain market share is to take down the leader.
Are old people technologically challenged? (Score:2, Insightful)
Re:So . . . (Score:4, Insightful)
But if you're starting up from a complete stop you're expecting to go.
Re:Correlation != causation (Score:3, Insightful)
Possibly, but that should be easy to answer. We need the data for ages for all owners of the affected cars.
The data is for fatalities, not accidents. (Score:5, Insightful)
The data in question catagorizes fatalities. Elderly people are often
killed by accidents that would only injure a young person. This could explain
the data skew regardless of whether or there is an actual accelerator defect.
Re:not enough data (Score:2, Insightful)
Both are concerning but I can forgive panicked drivers who were not expecting to deal with this, much more easily than I can forgive a company that sheltered profits rather than lives.
If in fact driver error was contributory to the accidents, then at least these incidents will serve to (re-)educate the public as to what to do in the event their engine starts racing uncontrollably.
And though I'm sure you know this, it's worth reminding others that turning off your car usually means you're killing your steering (and power brakes), too. So unless you're on a long stretch of straight road and have no other options, better to shift into neutral if you have the presence of mind, or if that's locked out for some strange reason, utilize the emergency brake if your regular brakes are inoperable or have faded because you've been riding them to try to stop.
Re:not enough data (Score:5, Insightful)
The spreadsheet shows 20 age 50+ and 15 age 0-50. That doesn't sound statistically significant, let alone "overwhelming."
And if a driver is 50, are they put into the 40-50 category, or the 50-60 category? Where's the data on Toyota model/year ownership by age, needed to even begin to make a valid comparison? Is 55 the median age for the owner's of the models/years involved in these accidents?
Seems like a poorly thought out attempt to make a case to me.
Statistics (Score:5, Insightful)
First thing you would need, if you really wanted to see if there was a correlation, would be the age distribution of Toyota drivers.
If, perhaps, the distribution looked just like this graph, it would mean nothing.
If, perhaps, the distribution of driver ages skewed to younger drivers, or showed a flat pattern, then you might have something.
Without that baseline, it isn't even worth coming up with theories.
Re:not enough data (Score:5, Insightful)
The easiest way is to make a superior product, which GM fails at. Isn't NUMI enough proof, it's not the workers, it's the managers? Toyota took a failing GM plant, slapped together Corollas, or Prism's... what ever you choose to call them. All in less then 6 months. Toyota is on the top because they build a quality product, stand by it and when it's bad due to design, they just f'n fix it and GM looks at it like an opportunity to rape their clients (yet again).
Context (Score:5, Insightful)
These numbers are meaningless without the proper context.
First of all, what is the percentage of ownership, by driver age. In other words: Do a disproportionate amount of older people buy these cars?
Secondly, what is the comparable accident percentage, by car manufacturer and driver age. In other words: Do older people have a problem with all manufacturers or only Toyota?
Lastly, 24 incidents is way too few to make any kind of sane inference. Once you break it down by age category you have some categories that only have one to three members. At that low an amount they could simply represent random chance and not some sort of trend.
When you have such a low number you have two choices: ignore the problem or dig deeper beyond these simple statistics. Given that people's lives (and Toyota's reputation) are at stake I'd say that Toyota is doing the right thing by dissecting the cars and chasing every possible problem. If they find something then they can fix it, if they don't find anything then at least they gave it their best and can honestly say that these incidents seem to be user error.
Re:And 1/2... (Score:3, Insightful)
It's simply a matter of economics, my friend. Gasoline (petrol) is currently running about $3 (2.25 Euro) per gallon here in the States vs. the average price of gasoline in Europe, which has been running about $6 (4.5 Euro) per hallon in Europe (on average).
Europeans pay double what we pay for gas, so it only makes sense that they'd be driving smaller, higher efficiency cars.
When gas prices pushed over $4 a gallon range last summer, hybrids were selling like hotcakes.
Re:not enough data (Score:3, Insightful)
This data is not very useful without knowing the age distribution of the drivers of Priuses. If 80% of the drivers are between the ages of 50 and 90, this data would mean that the frequency per age group is actually skewed low.
Re:Black Box... (Score:3, Insightful)
IOW, showing that the system is self-consistent doesn't prove that it is correct.
Re:not enough data (Score:4, Insightful)
Either the seat belts were not working/good, or you should have worn them.
Then why don't other manufacturers have problem? (Score:2, Insightful)
Re:not enough data (Score:3, Insightful)
(This statement in no way implies that I believe older people are more dangerous drivers.)
(( Though I do. ))
Look at the average Toyota Driver... (Score:4, Insightful)
According to TTAC, the number #1 vehicle for unintended acceleration is the Lincoln TownCar. The Ford Police cruiser is one of the lowest, however. Funny thing is that, mechanically-- they are the same car. The difference is the people who drive them-- one group being highly trained with fast reaction times, and the other group-- well not so much.
It is not just age distribution that they need to look at with Toyota, it is the complete demographic of the Toyota owner. Car enthusiasts do not usually buy Toyota's these days. Toyota's are incredibly boring in appearance and they handle like slugs. The are anti-exciting, right up there with a root canal. The average Toyota driver is the person in the fast lane doing 45mph and texting someone at the same time. For the average user, unintended acceleration happens everytime they touch that strange scary pedal on the right. When you add in that their brakes are likely shot because they drag them all the damn time while talking on their i-phone going down the road-- and never do routine maintenance on their vehicle: it is no wonder they can't stop.
Toyota's main problem is that they decided to make cars for idiots and got bit by that (granted that is a large market share, just ask Microsoft).
A couple of points. (Score:5, Insightful)
1) Older people have slower reflexes. A thirty-year-old is more likely to regain control of a runaway without incident than a seventy-year-old regardless of the cause.
2) Older people are not as strong. A twenty-year-old may be able to stop a runaway by hitting the brakes where a seventy-year-old can't.
3) Regardless of whether or not Toyota has a computer problem, some of the Toyota runaways are probably due to "wrong pedal syndrome". What is the age distribution for "runaway" accidents for all makes?
4) As others have pointed out, the elderly are more likely to die in accidents.
5) As others have pointed out, the sample is too small to justify any conclusions about age.
Re:And 1/2... (Score:4, Insightful)
No, that's not the reason. The reason is that most people around the world drive stick. There are only a few countries around the world where people drive mostly automatic transmission. In most of the world, if you only know how to drive automatic, you'll be restricted to an license that only allows you to drive automatic transmission.
Where I live (Argentina) virtually no one drives automatic transmission cars. We get the same models you do, but with manual transmission. This is true in most of South America. I once made the mistake of renting a car in the UK. Driving on the wrong side of the road was fucking difficult, but the car had manual transmission (In the US is almost impossible to find a rental car with manual transmission, I know from experience, but in the UK they gave me one automatically, and without asking).
If you had this issue in MT, it would be:
a) Trivial to just hold the clutch and disengage the gears.
b) On the AT model, pushing the accelerator would switch gears, while in MT you would still be in your current gear.
You have way more control. Also, the whole calculation done is probably different, I'm guessing even completely different, so, maybe the bug isn't present in those versions.
Re:Did you type this on a manual typewriter? (Score:5, Insightful)
Spoken like a true AT fan. Have you ever even tried driving stick? It's hard for about 3 hours, but once you get a feel for it you simply have so much more control over how the car behaves that it is actually hard to deal with not doing it. I feel like I'm going to die every time I pull into busy traffic in an automatic... they always seem to upshift too early, sacrificing torque for smoothness, which would be great if I didn't have some whacko barreling up behind me at 50 and I need to be going fast enough that he won't smash into me 5 seconds ago.
Oh, and when they flub going up steep hills, that's just terrific.
But you just go ahead and keep knocking people who are better at driving than their cars, I'm sure you know better than they do.
Re:Did you type this on a manual typewriter? (Score:2, Insightful)
Re:not enough data (Score:4, Insightful)
What I find far more concerning than people who can't tell their floor mat is pressing on their gas pedal are the vast numbers of people like you who think unsubstaniated assertions are a valid form of evidence against someone/something they don't like.
Re:Did you type this on a manual typewriter? (Score:4, Insightful)
I'd much rather switch gears myself. The car can't see the hill coming up, or spot the hole in traffic I need to merge into. I can, and having the ability to select gears for power or economy as I please makes handling those scenarios that much easier. The only place I'd prefer an automatic is when there's a string of stop signs on a hill, and there are morons behind me pulling right up to my bumper. I do sometimes roll back a hair, you know...
Re:Good god, please stop (Score:3, Insightful)
Not to mention, looking at the data indicates that "fatalities" are skewed toward the elderly, you read it wrong in that the Google spreadsheet provided was tracking incidents, not fatalities. Unless you meant the article linked, which indicated over 50% of fatalities involved drivers over 60, while the general fatality rate for all crashes has about 15% of crashes involving over 60s. So unless old people are 4 times as likely to drive a Prius as the general population, it still shows old people are over represented. Now all we need is rates for Prius mileage driven by over 60 vs under 60. If the rate is anywhere close to even, then the numbers show a distinct age related factor.
So, you are 100% right that we can't draw any real conclusions yet, but it does prove that either old people drive Priuses much more than everyone else, or there is a distinct age-related factor. So, since you seem to poo poo the idea of an age related factor, please present your proof that the elderly are more likely to drive a Prius. Any less than this, and your post was just an ill-thought out "I don't like that implication, so I'll just make a half-assed comment about the data being crap without ever having thought about the issue.
Or are you no less a moron than the people you complained about?
Oh, and if anyone wonders where you get good stats on traffic fatalities in the US, go to http://www-fars.nhtsa.dot.gov/Main/index.aspx [dot.gov]. Fatal Accident Reporting System, abbreviated FARS, then renamed Fatal Analysis Reporting System because the government hates "accident" because people associate "unavoidable" with accident, when 99% of crashes are avoidable driver error.
Re:The data is for fatalities, not accidents. (Score:3, Insightful)
Re:Non-issue (Score:3, Insightful)
It's easy to explain. "Unintended acceleration" from hitting the wrong pedal is a common cause of accidents, especially among the elderly, and it's easily accepted as driver error. As soon as there is a report of it being the fault of some specific car, it opens a way for everyone involved in the accident to avoid blame, and potentially collect more money.
Had an accident in a toyota? Now not only was it not your fault, but you might get money out of it!
Re:Did you type this on a manual typewriter? (Score:3, Insightful)
i can't say, never having seen you drive.
You're the one arguing automatics can't be controlled properly. I want to know in what ways.
i think you've never driven for fun.
I drive quite frequently for fun, though not as often as I'm on a motorbike, these days.
when, on the other hand, you don't need to shift so frequently, it becomes nice to have something to think about. here, automatics just take away the enjoyment of a long drive.
Long drives are boring, regardless. Having to regularly row through the gearbox just makes the whole experience more fatiguing - and the last thing you want on a long trip is more fatigue.
cruise control is usually liked by people who aren't bothered about driving and just want it to be as painless as possible. i find it quite boring.
Cruise control is liked by people who are experienced at driving distances and realise that it makes monitoring your speed one less thing you have to worry about, again reducing fatigue.
It blows my mind that anyone would want to drive for any non-trivial distance without cruise control.
Re:Did you type this on a manual typewriter? (Score:1, Insightful)
i think we may have`to be satisfied with having different opinions in this subject, since you can't even CONTEMPLATE how someone can not-like cc. just amazing!
-perryizgr8
/. not letting me post >25 posts in a day!!?
wtf??
Re:Non-issue (Score:3, Insightful)
How do you know they don't have wrong pedal crashes in other cars with the same frequency? This Toyota recall has pulled people out of the woodwork and drawn the media into a frenzy; how do you know the incidence rate isn't equivalent for other cars?
Re:one in a thousand odds not good enough? (Score:1, Insightful)
argument->kaboom
Re:Non-issue (Score:3, Insightful)
If the pedal spacing is causing excessive fatalities, then it is a design fault.
Re:Did you type this on a manual typewriter? (Score:3, Insightful)
Spoken like a true AT fan. Have you ever even tried driving stick?
I'm not GP, but allow me to chime in. I initially learned to drive manual - there is no other option in my home country. So that's what I had to prepare for, and that's what the practice exam was with.
But, as soon as I could, I got an AT car - and never regretted it. GP's comment about manual typewriters is spot on. For the sake of a flamebait, I'll add an even more apt one - driving manual is like running OpenBSD. Sure, you're in control, but do you really need to waste so much time and effort for so little benefit? Maybe, but for most people the answer is definitely "no".
Re:And 1/2... (Score:3, Insightful)
The problem isn't about how easy it is to do that. The problem is that, with AT, you almost never have to do it in the course of day-to-day driving; whereas with a stick, you do it so often it's practically instinctive.
Re:not enough data (Score:3, Insightful)
Rubbish.
In any reasonably mature market (which cars definitely are), it's damn difficult to build a superior product. Particularly when there are so many variables which can be tweaked and changing them impacts other aspects.
Make the car look better? How will that impact aerodynamics? Which will impact fuel economy...
Better build quality so bits don't fall off? OK, but that'll cost more money which will have to be either recouped in the sale price or swallowed by the manufacturer, resulting in lower per-vehicle profits (and it may take years for the general public to notice your improvements).
Better after-sales service? Possible, but difficult when the after-sales is dealt with by franchised dealerships you have limited control over. And most people shouldn't be having to rely on it anyway.
Easier to service? You'll win friends among drivers (who will surely appreciate lower servicing bills), and your warranty repairs will become cheaper because there will be less labour involved. But you'll also risk alienating your franchised dealers who make a lot of money from plugging in the Mysterious Box to the OBD port and getting more detailed diagnostics than a generic one will give. Furthermore, would any extra cost involved in development and manufacture outweigh this? For that matter, is it even possible to design a car that's a doddle to service while remaining in line with all the various bits of legislation in the markets you want to sell it?
Re:not enough data (Score:3, Insightful)
No.
It's easier to make lots of money in a shorter time with the "inferior product/blitz marketing" combo rather than a long slower trudge up "earning customer trust" hill with well-made products. And this suits today's "profits now -- fuck the long term viability of this company, I'll be gone by then" corporate mindset just fine. It's not easier to make the products themselves. Making cheaper products means a balancing act between a product that seems to be "good enough" quality for consumers, yet low enough quality to break/wear down to facilitate requiring replacement. There's less room for variances in manufacturing tolerances, because you have a much smaller margin between "decent" and "too little".
The company that makes higher quality products will be able to command higher prices for it's goods, but have to invest more energy into research and development (product improvement). And the added "quality margin" will not be as much revenue as repeat purchases of the inferior product (that's mostly because consumers look too closely at price and don't know enough about many goods to judge the quality to see which is really the better buy). A well made product will last longer and that company will have to find new ways to get consumers to purchase new models of the product if they still have a perfectly functioning earlier model. Market saturation will reduce revenue until this innovative new feature can be found.
I should have specified I was talking about engineering and manufacturing when I said I agreed it's easier to make the higher quality product.
Even if the product is cheap enough that the consumer gets discouraged and avoids the same manufacturer for the replacement, it still works out -- if all the other manufacturers of the same product are practicing the same deliberately shoddy product design. Because while John may swear not to get another Brand X DVD player and instead get a Company Y model, someone else is similarly disgusted with Company Y's goods and will buy a Brand X one instead. Neither one is aware really that both companies suck, because the over-marketing is drowning out any serious product reviews, that and the recent trend of threatening to sue any publication that says anything bad about a product helps that, too.
So the companies are instead now competing using advertising firms to out-dazzle each others' share of consumers. The difficulty in marketing crappy products year after year isn't counted in the company's own difficulty in making said products -- that's the advertising firm's issue. But it's worth pointing out that is another facet to the "make it cheap and sell a bunch" approach.
Sigh, that answer is so simple (Score:3, Insightful)
Because Toyota hasn't pointed it out. Don't you think that if these incidents occured across all cars, the Toyota would have pointed it out by now?
Usually the best indication that something is not a defense is that the defense ain't using it.
Re:Did you type this on a manual typewriter? (Score:3, Insightful)
It's not that women can't because they're women and they're small, stupid, whatever reason anyone may argue. Teenage girls, in particular, may also be 80 pounds (though probably not, especially in the US) but they aren't experienced truck drivers, and they *don't care* about controlling the car safely. They care about getting where they want to go... the most technical thing they care about is hooking up their ipod to the stereo.
Teenage girls are the least likely of all demographics to care enough to learn to understand technical things, and driving an 18-speed transmission sounds rather technical. Even a regular five-speed manual takes quite a bit of understanding and practice to learn, and to a teenage girl that time is 100% wasted if they can just drive an automatic instead. It's not that they *can't* learn, but they don't care and they don't want to learn.
There's nothing sexist about the argument you're responding to. It has nothing to do with whether women are capable of driving with manual gearboxes or not - obviously they are. And girls aren't conforming to any stereotype when they choose an automatic over a standard... just like the majority of guys (in the US), who also choose automatics! They've just got more important things to care about, and automatic transmissions are "good enough" for most - obviously not for truckers.
Actually, come to think of it, as a young guy in the US if I think about all the people I know well enough to know this about them, I think I know more girls who can drive a manual than guys.
Re:And 1/2... (Score:4, Insightful)
True. However many cars are drive by wire, meaning, there is no physical connection between the shifter and the transmission. So if there is a computer glitch, which may have happened with some of these cars, then you're SOL. Only option is to turn off the car -- unless it's a pushbutton start (like the Prius), then you're really up the creek ...
Re:And 1/2... (Score:3, Insightful)
It's not trivial to shift into neutral *in a car with an electronically controlled automatic*.
The more things that are drive by wire, the more complexity there is on paths where you really don't want more complexity.
It's fairly difficult to quantify complexity, but it's so easy to estimate it that there's really no call for this.
Read *any* of the last 25 year's issues of RISKS Digest [risks.org] for more on this.
Re:Did you type this on a manual typewriter? (Score:3, Insightful)
Thanks for an enlightening post. It clear up some confusion in my mind about AT vs Manual transmission.
I've never driven a truck of the sort you are talking about, but I can see how human intelligence can make the difference in alot of situations as you pointed out.
I got my license on an AT, and when I bought a car with a Manual transmission, I thought, hmm, I guess it's going to be harder to eat that Whopper now while I'm driving, other than that I didn't see any real advantages or disadvantages to it. Vaguely, I half remembered that race car drivers and car afficionados preferred manual transmissions, because they had more control over the car. I guess I agreed since starting off can be a little sluggish in an AT but then those times when you forget you are still in third and try to take off in third, and don't go anywhere that happen to me sometimes even after driving nothing but manual transmission vehicles for six years, sort of compensate for that.
Anyway, a couple of days ago on wired I read that all the new Ferrarris are coming out with AT. I don't know a damn thing about cars, but I was thinking HMM... I guess the afficionados have changed their collective minds.. I wonder why... I still don't know, supposedly the automatic transmissions were faster on some track tests, but I guess if you have a huge enough engine you'll take off just fine no matter what gear the transmission's electronics are in...
Then again, watching dang, I wonder the name of that show is where the celebrities try for the best time, Top Gear?, whatever, I always see them farting around with the electronics on even very expensive cars like ones that cost twice what my house does, and I'm thinking - WHAT A PIECE OF CRAP!
Electronics that you see, are a big minus. Note to electronic interface designers designing interfaces to computerized crap on cars - When accessed through the electronic interface, your car should respond like Mario does on SNES games. INSTANTLY, and without any thought. NOTHING of importance should require looking down at any controls. When controlling mario, you don't have ANY menus, you don't take your eyes off the screen. DRIVERS whose lives may depend on their cars responding to their input, and being able to see the road, DON'T want to be fracking around with electronics. Even stuff that ought to not be time sensitive ought not to take much time or thought. There should be no digital displays on a car. The radio's display can be digital and show the time, the station and the current mp3 track. AM/FM, Seek, favorites buttons, should be BIG. The volume and tuner should be the only two knobs. The tuner can double to select MP3 tracks read from a keyfob.
You are going to have your car for at least 5 years, and maybe 10. Any electronics you have are going to be obsolete after a two or at most three years. Any menus necessary to access features effectively subtract the feature as a selling point. Your car should be a car first and foremost, with electronics added only where absolutely necessary to make your car do it's car thing.
The only other exception to the no display except the clock radio rule is that the check engine light should have a display that is blank unless there is a problem. In that case it should display a human readable/understandable description of the problem with an error code. It might be too tempting for designers to use that display for something other than displaying error codes though. You DON'T want any displays you have to read to access funtionality of your car. accessing your car's functionality is otherwise known as driving, and that is dangerous when texting.
Re:Did you type this on a manual typewriter? (Score:3, Insightful)
Sure, you're in control, but do you really need to waste so much time and effort for so little benefit? Maybe, but for most people the answer is definitely "no".
When you're in control of half a ton of tempered steel traveling at roughly 60mph or more, then the answer is a definitive "yes." When we are talking about operating systems on home computers where a crash causes some headaches and a few days worth of inconvenience, you're right, you don't need that extra control. When we are talking about what is, essentially, a very powerful weapon that is supposed to be used for peaceful transport purposes (yes, that much directed energy is a weapon, like it or not) then that control is absolutely essential. Those folk who are too damned incompetent to deal with that level of control should stay the hell off the roads, just like those folk who are so old they can no longer tell the brake from the gas pedal.
Re:From the point of view of an end user and dev (Score:3, Insightful)
Whether it's a complex interaction fo systems that can't be reproduced in QA, the uncovering of a hardware of software bug; or if something as simple as the user consistently clicking the wrong button -- or pressing the wrong pedal, if that is what happened.
On the surface, yes - in some of those cases, the user does the "wrong" thing. But what that really means is "the user did not do what I said they should do". So is that user error, or interface design error? Why would they do it wrong in *this* case, but not in other cases? WHy did the same user never have this problem with any other car?
A bug doesn't mean only that code is broken. It can occur in any number of steps in the process -- code, interface, expectations we have set for the users, design, assumptions, hardware, etc.