Special Reports

The National Highway Traffic Safety Administration (NHTSA) has changed its methodology for determining which cars are most likely to roll over, and for the first time is able to assign rollover percentage risks for specific 2004 vehicle models in simulated single-vehicle crashes. Single-vehicle crashes, in which the vehicle flips over without being struck by another vehicle, are responsible for a full 85 percent of all rollover-related deaths.

Although rollovers are still relatively rare, they have a higher fatality rate than other kinds of crashes. According to the government agency, more than 10,000 people died in rollover crashes in 2002, accounting for nearly 33 percent of all deaths from passenger vehicle crashes. Between 2002 and 2003, SUV rollover deaths increased 6.8 percent, though passenger car and pickup truck rollover deaths decreased.

Previously, the NHTSA determined which vehicles are more likely to roll over based on how top-heavy they are — a simple mathematical calculation, based on vehicle width and height, known as the Static Stability Factor (SSF). This is why sedans and car-based SUVs, which have a lower center of gravity, tended to score better than trucks and truck-based SUVs. But the problem with the SSF calculation is that it is just that — a calculation. Only by putting a vehicle in a real-world driving situation can you determine with any reasonably accuracy its propensity to roll over during an evasive maneuver.

In order to address the obvious limitations of the SSF measurement, the NHTSA has developed a new rating system that combines the existing SSF with a new dynamic maneuvering test to gauge a vehicle's overall risk of an on-road, untripped rollover. ("Untripped" rollovers are those for which the only apparent cause is normal surface friction. Running into a curb or shallow ditch would be considered a "tripped" rollover. Roughly 95 percent of rollovers are tripped.) The dynamic test comes much closer to simulating real-world conditions: The vehicle is loaded up with weight, representing five occupants, and a full gas tank. Electronic Stability Control (ESC) was activated if more than 50 percent of a test model's sales had that option. The vehicle then simulates what a person might do while trying to avoid a high-speed collision — suddenly turning sharply in one direction, then the other (often called a "fishhook"). A remote-controlled steering system performs the maneuver assuring that each vehicle is given approximately the same inputs. Test instruments measure if the vehicle's tires lift off the pavement. If they lift off at least two inches, the vehicle is considered to have tipped up, the first stage of rolling over.

As with the earlier rating system, vehicles are classified as having between one and five stars, with five stars awarded to the vehicle least likely to roll over. Because most rollovers are tripped, though, the SSF is weighted far more heavily than the dynamic test in calculating 2004 star ratings. The star ratings for 2003 and earlier model years will remain the same, but consumers comparing pre-2004 with post-2004 models should be aware of the difference.

The new 2004 ratings reports confirm what we already knew about the risks of SUVs versus passenger cars. The SUV class performed worst overall, with 13 of the 36 SUVs tested tipping up during the sudden turns. The worst rating went to the two-wheel-drive Ford Explorer Sport Trac, which received just two stars and a whopping 34.8-percent chance of rollover. The best performer in this category was the all-wheel-drive Chrysler Pacifica, a vehicle we classify as a wagon due to its low ground clearance, but is classified as an SUV by the EPA. Among vehicles that we list as SUVs, the Nissan Murano 4x4 was the least likely to roll with its four-star rating and 15.1 rollover percentage. Like the Pacifica, the Murano is a car-based SUV that has a lower center of gravity than most truck-based sport-utilities.

As expected, passenger cars fared best, with the Mazda RX-8 leading the pack. It was the only model to receive five stars, with an 8-percent chance of rollover. Minivans also did well, with all other models tested receiving four stars. Results for individual vehicle tests, where available, can be viewed at www.safercar.gov.

It's interesting to note that, in the pickup truck category, rear-wheel-drive (2WD) vehicles outperformed their four-wheel-drive counterparts. But in the SUV category, the 4x4s outperformed the 4x2s. This held across the board, for every make and model pickup and SUV tested. Occasionally, the drivetrain even meant the difference between a vehicle that tipped and one that didn't. For example, the all-wheel-drive Mercury Mountaineer did not tip, and was given a 22.8-percent rollover probability. But the 2WD Mountaineer 4x2 did tip, and was given a 28.3-percent probability. Both, however, were still rated three stars overall. Why the difference? We speculate that it has to do with the fact that most four-wheel-drive and all-wheel-drive SUVs now come with full-time systems that automatically route engine power to all four wheels in lower traction situations — providing them with greater balance and grip than their rear-wheel-drive (2WD) counterparts. In contrast, most 4WD pickups still come with part-time four-wheel-drive systems, meaning that only the rear wheels get power unless you're driving off-road or in heavy snow. This fact, combined with their raised, off-road-ready suspensions, makes 4WD pickups less nimble than 2WD pickups, which weigh less and sit closer to the ground.

Blame it on record-high gasoline prices or the growing public awareness of rollover dangers, but large SUVs sales are taking it on the chin from their smaller cousins. Although SUVs still rule the road in general, sales of SUVs based on truck platforms, such as the Chevy Tahoe and Ford Excursion, are declining from their all-time high in 2002 to roughly their 1999 sales level. There are more dealer rebates and incentives on the large SUVs than there were previously, and these vehicles are sitting longer on dealer lots. Meanwhile, sales of SUVs based on car platforms (sometimes called "sport wagons" or "crossovers"), such as the Toyota Highlander, Honda CR-V and the best-selling Ford Escape, are skyrocketing. Sport wagon sales were up 18.5 percent in the first seven months of 2004, according to Automotive News Data Center.

The NHTSA has recently announced the 2005 vehicles it will test for rollover potential. It includes carry-overs and redesigns from previous model years, plus all-new models like the Buick LaCrosse, Chevrolet Cobalt, Ford Five Hundred, Ford Freestyle and Pontiac G6. All told, the upcoming rollover reports will offer safety information for approximately 73 percent of the 2005 passenger vehicle fleet.

Rules of the Road

NHTSA data suggests that over 90 percent of the vehicles in fatal single-vehicle rollover crashes were involved in routine driving maneuvers (going straight or negotiating a curve) at the time of the crash. So other than foregoing the purchase of a large SUV or other vehicle with a high center of gravity, how can you prevent a rollover when driving? There are some simple precautions the NHTSA recommends you take.

Lower Your Speed: Fatal rollover crashes are speed-related more often than fatal non-rollover crashes. Some 40 percent of fatal rollover crashes involved excessive speeding. Additionally, nearly three-quarters of fatal rollovers took place where the posted speed limit was 55 mph or higher.

Don't Drink and Drive: Nearly half of all fatal rollover crashes involve alcohol. Impairment can result from any blood alcohol concentration (BAC) above .00. Even a small amount of alcohol will negatively affect your judgment, muscular coordination and vision, making you more likely to lose control of your vehicle.

Use Extra Caution on Back Roads: Rural roads tend to be undivided and without barriers. When a vehicle goes off a rural road, it can overturn when it strikes a ditch or embankment, or is tripped by soft soil. Almost three-quarters of fatal rollovers occur in rural areas, where the posted speed limit is typically 55 mph or higher.

Buckle Up: Though it seems obvious, the best way to avoid dying in a rollover is the most simple…use your safety belt. Seventy-two percent of people killed in rollovers weren't wearing them at the time. Safety belts are even more helpful in rollovers than in other types of crashes because so many people are killed from being thrown from the vehicle.

Take Advantage of New Technologies: According to a new preliminary report by NHTSA, having electronic stability control (ESC) in your vehicle can reduce fatal, single-vehicle crashes by 30 percent, and 63 percent in SUVs. ESC, variable ride-height suspension and side curtain airbags with rollover sensors can all help you avoid or survive a rollover. If you're in the market for a new vehicle, check to see if these are available as standard or optional equipment.

Follow Manufacturers' Directions: Keep your tires properly inflated (check once a month when they are cold) and replace them when worn. (See "Tire Safety".) Only load your vehicle up according to your owner's manual. Any load placed on the roof will raise the center of gravity — and thus the likelihood of a rollover. (See "Cargo Weight and Rollovers".)

Know Proper Maneuvering: Avoid panicked steering. Many rollovers occur when drivers overcorrect their steering as a panic reaction to an emergency — sometimes to the point of a wheel going off the pavement's edge. At highway speeds, overcorrecting or excessive steering can cause the driver to lose control, which can force the vehicle to slide sideways and roll over. If your vehicle does leave the roadway, gradually reduce speed. Then ease the vehicle back onto the roadway.

Pay Attention to Weather Conditions: Bad weather equals higher accident rates. Use extra caution on icy roads, or avoid them altogether. Be most wary of wet roads when rain first starts as it tends to bring oils to the road surface resulting in less traction. Also, check your tire pressure on very hot days when tires are likely to heat up faster.