Some high-end cars do poorly on new crash tests

A new test that simulates a common and dangerous kind of front-end car crash shows major differences among high-end vehicles that would likely affect whether drivers walked away or suffered serious injuries in a kind of accident that kills thousands of U.S. motorists each year, according to the Insurance Institute for Highway Safety.

Among 11 models tested - all luxury or near-luxury 2012 midsize sedans - only three rated a score of "good" or "acceptable" in the new tests: an Acura TL, a Volvo S60, and an Infiniti G.

Four vehicles - an Audi A4, a Mercedes C-Class, a Lexus ES 350 and a Lexus IS 250/350 - earned an overall score of "poor," as crash dummies recorded forces sufficient to cause severe lower-body injuries and testers found other risks of serious head or chest injuries, the IIHS said in a report due to be released Tuesday morning.

IIHS officials said they hoped the new test - designed to simulate a crash where the front corner of a vehicle hits another car or, say, a tree or utility pole - would encourage design changes to make all cars safer by better protecting vehicles' passenger compartments.

"It's Packaging 101. If you ship a fragile item in a strong box, it's more likely to arrive at its destination without breaking," the institute's president, Adrian Lund, said in a report on the findings.

The new tests by the institute, which is funded by the insurance industry, won praise from Consumer Reports, which rates vehicles for safety based on its own evaluations as well as on crash tests peformed by IIHS and by the National Highway Traffic Safety Administration (NHTSA).

Jeff Bartlett, deputy editor at the magazine, said the tests should spur engineering advances.

"This is really a pop quiz. This is a test these cars were not engineered for," Bartlett said.

Russ Rader, a spokesman for the insurance institute, said the new test was designed to address a phenomenon that had long perplexed auto-safety experts: People were being seriously injured or killed while riding in vehicles that performed well on other crash tests.

Rader said investigators realized that carmakers might, in essence, be designing "to the test," much as schools will target curriculum to meet standards on which their students will be tested.

NHTSA crashes test vehicles at 35 m.p.h. into a stationary barrier, mimicking a head-on crash with another vehicle of similar weight. Since 1995, the IIHS has performed a "moderate overlap" test, in which 40 percent of a vehicle's front end crashes into a deformable barrier at 40 m.p.h. - a test that mimics a partial front-end crash with another car that also absorbs some of the impact.

The new "small overlap frontal crash test," also performed at 40 m.p.h., was designed to find what happens when just 25 percent of a car's front end - the front corner on the driver's side - hits a rigid barrier. It is similar to a real-world crash in which a driver drifts into oncoming traffic, or loses control and hits a tree or another car.

In vehicles that scored poorly, tests showed wheels being pushed into the well where a driver's foot is situated - a source of lower-body injuries. Some crashes also pushed windshield support pillars toward a dummy driver, or pushed away a steering-wheel-mounted airbag, lessening its protection, Rader said.

Earning a "marginal" overall score in the tests were an Acura TSX, a BMW 3 series, a Lincoln MKZ, and a Volkswagen CC. Worse scores mean more likelihood of serious injury.

"Most automakers haven't paid attention to this kind of crash scenario," Rader said. He said one exception was Volvo, whose S60 was one of two sedans, along with an Acura TL, to earn an overall "good" rating, an indication that injuries would likely be minor.

"Volvo has been looking at this kind of crash configuration since the 1980s," Rader said.