Drivers of vehicles that earn ratings of four stars in the European New Car Assessment Program (EuroNCAP) are about 30 percent less likely to be severely injured in a crash than drivers of cars with one-star ratings. This is the main finding of a Swedish study that compares real-world crash outcomes with crash test results.
EuroNCAP rates cars based on two tests, a 40 mph frontal offset test like the Institute's plus a side impact. New cars in Europe earn one to five stars based on performance in these tests. Four stars typically is the highest rating. To earn a fifth star, a car must pass an additional side-into-rigid-pole test (see "Euro NCAP results spur improvements in crashworthiness," Feb. 9, 2002). None of the cars assessed in the Swedish study earned five stars.
The researchers examined police-reported injuries to 12,214 drivers in car-to-car crashes in Sweden over six years, comparing the results with vehicles' EuroNCAP ratings. With each decrease in rating stars, the severe injury risk increased significantly.
"These findings lend additional validity to programs that rate vehicle crashworthiness based on the results of crash tests," says Institute president Brian O'Neill.
The study was conducted by researchers at the Swedish National Road Administration and Australia's Monash University. Similar results are reported in another Monash study conducted with the Institute. The researchers compared injury data reported from real crashes with results of both the Institute's frontal offset crash tests and the federal government's full-front tests. Using crash data from three states (Florida, Ohio, and Pennsylvania), the researchers found fewer deaths and police-reported serious injuries in vehicles with good test ratings.
However, the results were not robust, and the strongest correlation between crash test and real-world crash performance was for nonfrontal crashes. "These weak relationships likely reflect the lack of detail and fundamental difference in injury information in police crash reports compared to that used in deriving crashworthiness ratings," the researchers say.
An advantage of the Swedish data was that severe injuries were defined as those requiring hospitalization. In the U.S. study, they were defined as injuries recorded by police as serious. Studies have shown that most police-reported serious injuries actually are minor, so the Swedish researchers were better able to identify serious injuries of the kind measured in the crash tests. However, even the Swedish data didn't include information about the types or severities of specific injuries.
O'Neill explains that "detailed data on crash and injury severities are available only for a limited number of crashes for which special investigations are conducted — for example, crashes included in the National Automotive Sampling System. But the sample sizes in these datasets are inadequate for studies relating to individual vehicle models, so they don't allow us to say whether a specific model that earns good crash test results also does a good job of protecting its occupants in real crashes. On the other hand, the much larger samples of police-reported data have their own limitations. They typically include few if any reliable details on injury or crash severity."