OK, I admit it. I wrote a rather harsh criticism of GM's recent study of its OnStar System (GM's hands-free cell-phone system for calling roadside assistance). The problem I had with the study was my incredulity at the finding that the probability of having an accident while talking on a cell phone was one in four million!
My concerns about the study were these:
1. That the one in four million number was misleading--if not explained. I attempted to explain it.
2. That the study was not representative of the "typical" use of cell phones.
3. That the media and the people responsible for legislating against cell phone use while driving would cite this number as proof that talking and driving was not a dangerous combination. This study does not prove that.
Since my little rant was written, the people at GM have been good enough to supply us with additional information about the study (along with a letter from its legal affairs department in which it suggested that some of the statements we made were inaccurate, unfair, and irresponsible.)
Not wishing to be guilty of any of these terrible vices, I've gone back to the drawing board, considered what GM had to say, and re-written the rant. I hereby admit that I may have gone a bit overboard in suggesting that GM's research was completely bogus and that GM intended to deceive the public. I take it back and apologize.
So, the new version (below) describes what they did and my reasons for the warning that we shouldn't be comfortable relying on this study to suggest that talking on a cell phone - hands-free or not - is as safe as the one in four million number might suggest.
Risk is One in Four Million? I Beg to Differ.
by Tom Magliozzi
GM has a system called OnStar which allows members to call GM via an "embedded" cell phone (it's built into the vehicle and is hands-free). You can use this device as a regular cell phone and call anyone you want to call. Or you can use it to call a service adviser. If you break down somewhere you can call and they will send a tow truck. If you get lost they will give you directions. We certainly don't dispute that this is a useful service when used to call a service advisor from the side of the road. (After all, it is called roadside assistance.)
Anyway, GM decided to analyze the phone calls that its received over the past few years - actually eight-million phone calls received over a five-year period. And out of all these phone calls, there were only two accidents. Two! So they publish their conclusion: two accidents, eight million phone calls. Therefore, according to the study, "the frequency of an OnStar call uniquely causing an airbag crash approaches zero airbag crashes per eight million embedded cell-phone calls." Sounds good doesn't it?
But maybe it's not quite as good as it looks. Here's why.
The GM report actually says that there may have been as many as eight accidents. But for the moment let's assume it was only two.
First, we have GM's interesting definition of "an accident." GM defined an accident as a crash in which the air bag deployed! Right away we know that this definition is going to be a problem - maybe that's how GM looked at it - but the way it defined "accident" leaves out most accidents. How big a difference might this make? Well, the air bag deploys in something like 5 percent of accidents (don't forget, it usually deploys only if a frontal accident occurs above approximately 20 to 30 mph). So if you sideswipe another car, or roll over and die, or drive into a crowd of people and kill them all, it's not an accident according to GM's study.
So if we straighten out that problem, the numbers go up by 20 times - to either 40 or 160 accidents, depending on whether you start with 2 or 8. Still not a huge number, but enough to make you pay attention. But that refers only to those who use the OnStar system. What about cell-phone use in general? That's when the numbers really make you think twice about ordering pizza, or calling your girlfriend, from the highway. Even if there were only two accidents in that sample of eight million calls, and that's a big "even if," let's just extrapolate to see how many total accidents it might represent.
Well, in the five-year period of the study, there were about 130-billion cell phone calls made in the U.S. 130 BILLION! We don't know how many were made from moving vehicles but let's say it's half of that, 65 billion. (This is my estimate, I don't think the real number is known.) So the eight-million calls are 0.012 percent of all the calls made from cars. So if we were to take GM's two OnStar "accidents" in a sample of eight million, that little number would extrapolate to 325,000 real accidents. If there were as many as eight accidents, this number goes up to 1.3-million accidents! Stay with me.
GM claims that a significant benefit of its study is that it, "covers on-road real-world (emphasis added by me) usage conditions." I beg to differ.
One of the first rules of research is that you have to use a "representative sample." That is, if you really want to extrapolate the findings to the REAL, real world (not just the OnStar real world), the phone calls you select to study should be as much as possible like ALL the phone calls in the real world (for example you wouldn't be dumb enough to draw conclusions about the attitudes of teenagers in America by studying old people in Japan. Or you wouldn't/shouldn't draw conclusions about the dangers of people driving and talking on cell phones if the people you studied weren't driving!) For lots of reasons GM's sample is far from representative of the REAL, real world.
Were any of the callers engrossed in an argument with their wives or husbands? No.
Were any of them getting read the riot act by their bosses? No.
They were all talking to a calm, well-trained service adviser.
Were these people even driving? Or had they pulled over to the side of the road? GM isn't really sure. (They report that they did a survey of 11 service advisors and concluded that the vehicles were in motion 70 percent of the time. This is explained in footnote 14 of the study, which I've read several times. If you can figure it out, call me.) But, was it the driver who called or a passenger? GM doesn't know - or doesn't say.
GM claims that its study is more representative of the real world than laboratory studies or other studies which do not have actual phone calls in the data. I agree that they have cell-phone calls in their data. But so what? The phone calls are representative of only a small percentage of real phone calls and the definition of an accident is so narrow that the conclusions are useless and irrelevant, in my humble opinion
So, what should we do? My suggestion is that we forget about this study and look at what more objective researchers without vested interests have already proven. And, the fact is, there's plenty of credible research out there on this topic. Here are few recent studies:
- "Driven to Distraction: Dual Task Studies of Simulated Driving and Conversing on a Cellular Phone"
- Interdependence of Non-overlapping Cortical Systems in Dual Cognitive Tasks
Both of these studies accurately quantify what many of us have expected to find all along. Your brain can't multitask. Effectively meaning that if you're talking on the phone, you can't pay close enough attention to the road.
If GM is trying to make the point that hands-free talking is safe, at least using OnStar, we don't agree. We don't actually know that it's not. We're just saying that this study doesn't show that it's safe.
Both of the studies above - as well as others - show that it is not safe for drivers to make cell-phone calls from moving cars.
So, if you're like us, you're left wondering, "Why would GM publish the results of this study?" Well, funny you should ask. It just so happens that GM is at the forefront of "telematics," the technology that will bring yet more distracting features to your car. They stand to make hundreds of millions of dollars with email, Web access, and cell-phone technology in its new cars.
In other words, don't mistake this study as being from an independent, disinterested source, because GM clearly isn't. In my humble opinion, GM's trying to lull us into a false sense of security. So far it hasn't suggested that you can surf the Internet and be a safe driver at the same time, but who knows what the next study will show?
They don't even mention the possibility that using OnStar, or some other cell phone, might mean you or a loved one could be driving a wheelchair the rest of your life because you were arguing with your boss instead of watching the curves in the road.
Please read its study and let me know if you agree. And thanks for listening.