Another salvo in the war over America's SUVs was fired recently when Jeffrey Runge, a doctor who just so happens to head the Bush administration's National Highway Traffic Safety Administration, declared that SUVs pose an "astounding" threat to their owners due to their proclivity to roll over and that strict new regulations may be in the offing.
"The thing that I don't understand is people, when they choose to buy a vehicle, they might go sit in it and say, 'Gee, I feel safe,' " said Dr. Runge. "Well, sorry, but you know gut instinct is great for a lot of stuff, but it's not very good for buying a safe automobile." Well, medical doctors may be great at a lot of stuff, but they're not necessarily very good at assessing non-medical data. The public, in fact, is right and Dr. Runge, in fact, is wrong.
It is true that SUVs are more dangerous to be in should they roll over than are most passenger vehicles. But only 3 percent of all accidents involve roll-overs.
If you're driving an SUV and get into an accident, most of the time it will involve hitting (or getting hit by) something. Accordingly, drivers are right not to worry too much about rolling over in their vehicles, particularly because it can be avoided simply by eschewing NASCAR racing practices when making sharp turns.
The "gut instinct" of SUV owners that increased safety in one- or two-car collisions more than offsets the risk of roll-overs was validated last October in a remarkable study published by the National Bureau of Economic Research and authored by University of Michigan economist Michelle White. Miss White managed to secure data regarding each and every automotive accident reported to the police between 1995-99. She examined three types of crashes: those involving two vehicles, those involving a single vehicle, and those involving a vehicle striking a pedestrian or bicyclist.
Vehicles were divided into five categories: cars, SUVs, pickups and minivans, large trucks, and buses. She then performed a regression analysis of the data, controlling for seatbelt use, urban and rural conditions, weather, time of day, negligence, age of the drivers, road type, speed, and number of vehicular occupants.
And what do you know? The analysis found SUVs were saving a net of between 1,023 and 1,225 lives every single year. Moreover, the study found no statistically significant evidence that you are more likely to die if your passenger car got into a collision with an SUV than if your passenger car got into a collision with another passenger car.
Interestingly enough, Professor White found that light trucks as a class were responsible for an unnecessary 2,260 deaths on the road every single year. Apparently, it's the pickups and minivans — not the SUVs — that are the problem.
What makes this study remarkable is that it's the first time that actual case-by-case crash data were used to examine SUV safety. Earlier studies used aggregated data that prevented analysts from controlling for all the relevant factors that might contaminate the findings. Any statistician will tell you that micro data are much preferred over macro data for this very reason. And Miss White is the first analyst to put the relevant micro data through the paces.
"Well," a skeptic might reply, "The National Research Council of the National Academy of Sciences concluded last July that SUVs were responsible for an unnecessary 2,000 deaths a year, and I'll believe the consensus of experts over some (probably on the take) economist."
The problem, of course, is that the NRC study was based on a review of the published literature, and that review did not (indeed, could not) include the White study published three months later. The NRC simply assessed the findings of various studies that used aggregated macro data studies that are now far less persuasive given the far better data set used to produce the findings reported by Miss White.
And not that it really matters any, but it was the Institute of Civil Justice at Rand — not the auto industry — that supported Miss White's study.
So there you have it. The anti-SUV jihad may continue to roll on, but it cannot credibly do so with an anti-safety argument in tow. Dr. Runge should check the facts before he jumps on this rickety bandwagon.