Until the early 1980s, government regulations were the primary way safety features made it into vehicles. For automakers, that spawned a minimalist approach wherein they spent the least amount necessary to meet the minimum standards required by law. For the OEM'S, it was the perfect solution. Simple standards kept the playing field level, affordable and provided an easy pass/fail target. But to consumers, it drew attention to the fact that some automakers, particularly those from Europe, voluntarily designed to a higher standard.

By the mid-1980s, safety minded companies like Volvo, Saab, BMW and Mercedes-Benz began cashing in on their reputation for crashworthiness. And to the shock of the U.S.-based industry that defiantly said "you can't sell safety," consumers pushed domestic automakers to make safety a priority.

"We saw that as an opportunity to do things the government couldn't," says Brian O'Neill, president of the Arlington, Va-based IIHS. "We conceived of having our own crash test facility and test procedures that were different from the government's. We didn't want to compete, but government agencies, by design, are not flexible or fast responding. We knew we could promote safety improvements more rapidly."

One of the tests the IIHS would use to achieve that goal was a controversial offset, 40-mph frontal crash into a deformable barrier. The test was actually conceived by Mercedes-Benz and already under development in Europe. It was viewed as a more real world alternative to fiat barrier testing, which was and is the basis of today's Federal New Car Assessment Program (NCAP) standard.

Then, in 1992, when the Institute opened the Vehicle Research Center (VRC), a crash test facility located in Ruckersville, Va, it put into action its plan to raise the bar on safety. It performed its first low-speed crash in February of 1993 and subsequently shook the world of automotive safety.

Shaking Up the OEMs

From the very first crash test program conducted at IIHS, the results rattled automakers. Even O'Neill expresses surprise at their early findings.

"Our first program was with mid-size cars like Taurus," he recalls. "And of those cars, we expected the Saab 900 to do very well because of Saab's concern for safety. Well, it did terrible. It collapsed. We thought maybe our speed was too high, but then we did our second test, which was a Subaru Legacy. We didn't expect that car to do well at all and it did fine. At that point we knew we had a great program because we were as knowledgeable as anyone and we guessed completely wrong."

Initially, automakers cried "foul" over the tests because vehicles were tested to a standard they were not designed for. Body engineers in particular were being chastised for doing a bad engineering job.

"I feel for those guys," O'Neill laments. "They did their job, but they weren't given the right priorities -- that's the fault of the policymakers. I admire Toyota because it actually sent engineers to us who asked, 'What are you doing and how do you do it? We want to do well in these tests.' The company bought into this from the top down and now there isn't a single Toyota that doesn't get a "good" rating from us."

David Zuby, vice president at the VRC, says his experience with automakers proves that the Institute's tests help promote safety awareness.

"Engineers tell me this is what they need to do better," Zuby explains. "Too often they are just given the outside shape of the vehicle, told what kind of engine will be in it and then asked to eke safety out of what's left. They leverage the fact that they don't test well because it gives them more clout. They don't have to fight so hard for crush space."

The offset test is particularly important to safety engineers because it tests structural integrity. It challenges the safety cage that protects occupants. In contrast, the 35-mph NCAP test makes a good adjunct to an offset crash. It generates much higher deceleration forces because the vehicle stops abruptly, making it ideal for evaluating restraint systems.

Different Views on Safety

Not everyone agrees with the results or validity of the IIHS tests or even the approach of the Institute. For example, in the offset test, heavy vehicles need to have a stiff front end to pass. That makes them potentially dangerous when they strike other vehicles.

"It's a valid criticism," O'Neill responds. "That is part of the reason we have drawn the line and not tested the largest vehicles, like the Ford Expedition. There is a point where the vehicle is just too heavy for this test."

Another common criticism is that the results can be deceiving. Critics argue that when a small vehicle gets the same rating as a large vehicle, it falsely implies that they are equally safe. O'Neill counters that the limitation in all barrier tests is that the crash is always equivalent of a vehicle hitting itself. He adds that using a moving barrier might give an assessment independent of weight, but because half of the traffic deaths are in single vehicle crashes, it might inappropriately put emphasis where it doesn't belong.

"We already know enough to tell the public that if your first concern is safety -- avoid small vehicles," O'Neill advises. "You don't need the biggest vehicle, but get out of small because they don't offer much safety."

Some of the concerns over IIHS testing lie not in the process but in the interpretation of the results. Robert Lange, executive director of safety integration at General Motors Corp. says that at least with the NCAP test GM has been able to clearly identify the differences between vehicles that rank high and those that rank low.

"We haven't been able to make that assessment yet for Brian's test," Lange says. "They spend a lot of time trying to correlate it to real-world safety; but we are a little concerned with the subjectiveness of that evaluation procedure. It would also be good to have a correlation between that test and injury reduction, but it will be a while before there is enough data."

O'Neill responds that there is a correlation to real world safety and a common sense issue as well.

"You already know you're in trouble when the passenger compartment starts to collapse," O'Neill says. "Then the restraints are no longer operating in the envelope for which they were designed. We use the analogy of shipping china to illustrate a safe design. The first thing you need is a strong box that won't collapse. Then you put in bubble wrap. The box is the safety cage and the bubble wrap is the restraint system. Most vehicles fail because the structure fails -- not the restraints. There's nothing wrong with the bubble wrap."

Supplier Input Boosts Safety

"When you start designing a restraint system, you assume the structure is strong -- that intrusion into a vehicle will be minimized," says Craig White, chief technology officer for Breed Technologies. "Without that, there isn't a lot a restraint supplier can do. But assuming that, the very basic formula is to sense a crash and then develop countermeasures for it."

White says sensing the crash is the tricky part, particularly in designing a sensor that does well in both a barrier and an offset test.

"There's a big difference in sensing time between the two," he says. "If you deploy an airbag too slow, the occupant is already out of position and the airbag can hurt them; too fast and they can bottom out the airbag. We do well in both because we have an excellent algorithm. It actually defines the class of the crash, then determines the severity within that class."

O'Neill gives suppliers high marks for the job airbag systems are doing, but also draws the line on airbag proliferation. Pointing to proposals for thorax airbags, torso airbags and even external airbags, he says enough is enough.

"We have airbags everywhere," he chuckles. "Do we really need all that? No. That's the short answer."

His one exception to that is side airbags, which O'Neill feels are needed to protect the head in side impacts (see sidebar at right).

"Inflatable side curtain bags are coming into their own," says Pat Jarboe, director of corporate communications for Autoliv. "We're on Volvo and Mercedes with those products now, and with rollover regulations coming it will grow. We'll use sensors that know' if the vehicle will roll, even when there are still four wheels on the ground. It'll fire the seat belt pretensioners before the person moves out of position, inflate the bag, and it will stay inflated for five to seven seconds to prevent occupant ejection. Those are things we are researching now."

Other supplier approaches to safety include accident prevention technology, such as Delphi's Integrated Safety System (ISS). Dr. Glenn Widmann, chief engineer for ISS, says NHTSA studies looked at forward collision warning, one of the fundamental elements of ISS, and determined it could reduce accidents substantially