Results were similar in the rear tests. Reducing the damage required the bumpers to engage the barrier and absorb the energy of the impact, but this mostly didn't happen. A relatively good performer in the full-rear test was the Hyundai Sonata. Its bumper did engage the barrier, and most of the damage was limited to the bumper (minor repair of the rear body panel also was required). Total damage came to $739.

Good bumper performance requires not only engagement with the test barrier but also strength sufficient to absorb the energy of a low-speed crash. Hyundai engineers strengthened the Sonata's bumper after learning about the Institute's upcoming series of new tests.

In contrast to the Sonata, the bumpers on other cars did slide under the barrier, and damage was much worse. The Chrysler Sebring, Nissan Altima, Volkswagen Jetta, and AURA sustained more than $3,000 damage apiece in the full-rear test.

"The bumpers on the Altima and Sebring didn't stay engaged with the barrier at all. The bumper bars on these two cars escaped unscathed," Lund says, "which means they didn't do what they're supposed to do. They didn't absorb any of the crash energy. Making sure bumper bars line up better to engage other bumpers in crashes is the first step toward preventing damage in the kinds of low-speed collisions our tests represent."

All parts don't cost the same: The total cost to fix damage after a minor bump is influenced by more than the extent of the damage. Another issue is that the price of the same part -- a fender, hood, or other part -- varies from vehicle to vehicle. For example, the Toyota Camry needed a lot of repairs after the full-rear test, including repair of fenders and body panels. The trunk floor and unibody structure had to be straightened out. However, the total cost of these repairs was a relatively low $1,480, in part because Camry parts don't cost as much as those on some other cars.

Besides the cost of damage in low-speed collisions, there's the aggravation. Most people want to avoid having to get repair estimates, arrange for repairs, and then do without their cars while they're in the shop.

"Much of this could be avoided if car bumpers were better at doing their job of resisting damage," Lund says. "But instead we have to put up with so much damage that it can take days or weeks to fix."

Styling influences performance: The performances of three cars show how front-end styling can influence the amount and cost of damage that occurs in low-speed crashes. The AURA, G6, and Malibu are all General Motors cars built on the same platform. So they're similar cars, but the cost of repairing them isn't the same. The G6's front end slopes more, and its front bumper bar is lower than those on the other two cars. The result was that the G6's bumper didn't stay engaged with the barrier during the full-width test. Instead it slid down and under the barrier. Damage including a crumpled hood, buckled fenders, broken headlights, and a bent air conditioning condenser cost four times as much to fix as damage to the Malibu or AURA. On top of this, the G6's front unibody had to be straightened out. The bill for all of this topped $4,500.

Lund explains that automakers "don't have to sacrifice car style for function. There's empty space under the covers of the bumpers on all three of these cars that could be used to put in energy-absorbing materials. Engineers also could make bumper bars taller and extend them farther out to the corners of the car without changing the front-end styling."

Corners left unprotected: The bars that are part of most bumper systems often fail to extend far enough into vehicle corners. The result is a failure to protect vital and costly parts such as lights and fenders.

"Headlight assemblies on all 17 cars were damaged in the corner impacts," Lund points out. "The lights essentially served as the bumpers on these cars because the bumpers themselves didn't provide any protection. There's no excuse for this. Safety equipment like headlights shouldn't be damaged in impacts at a mere 3 mph."

The width of the bumper bars was a factor in rear-corner tests too. While the Honda Accord sustained about $600 damage in this test, damage to the Mazda 6 totaled twice as much. The difference was that the Accord's bumper bar is nearly 80 percent as wide as the car, while the Mazda 6's is only 58 percent as wide.