The National Highway Traffic Safety Administration (NHTSA) conducts on-road dynamic rollover tests on most new SUVs, pickups, and minivans. The tests score rollover propensity in part by putting vehicles through a handling maneuver called a fishhook, a quick left-right turn, at increasing speeds from 35 to 50 mph. This simulates a driver overcorrecting the vehicle's steering, something that can happen in an emergency.

If the vehicle lifts two wheels off the ground it's considered a "tip-up" and testing stops. If the vehicle slides out or gets through the test without incident at 50 mph, it passes.

Before 2004, NHTSA rated rollover propensity with a static (non-moving) measurement of a vehicle's shape and weight distribution. Called the Static Stability Factor (SSF), it's derived from a formula that compares a vehicle's track width with its center of gravity height. But starting with 2004 models, NHTSA added the dynamic (vehicle in motion) test, and the agency now uses those results to augment its rollover ratings.

NHTSA has compiled rollover ratings for scores of vehicles, including separate ratings for two- and four-wheel drive versions of SUVs. So far, no car or minivan has tipped up. In fact, NHTSA conducts on-road rollover testing on only two sedans per year and assigns star ratings to cars based on their SSF alone, as it did with 2003 and earlier models.


The Scores

NHTSA combines the SSF and dynamic test to assign a rollover-resistance score of one to five stars. Five stars represents rollover likelihood in a single-vehicle crash at 10 percent or less; one star predicts a rollover likelihood of 40 percent or more. The SSFs underlying the star ratings vary from about 1.0 to 1.5. (The higher that number, the better.) SUVs usually measure out at 1.0 to 1.3, and cars normally fall in the range of 1.3 to 1.5.

"Our analysis indicates that the two tests together correlate with what we expect in the real world about 90 percent of the time," says NHTSA spokesman Rae Tyson.


What the results Mean

Many SUVs have tipped up in NHTSA's testing--understandable since the dynamic test is designed to provoke a tip-up if a steering maneuver alone can produce one. What is surprising is that a tip-up doesn't affect the star rating much. That's because the government's rollover ratings give much more weight to the static measurement than to the on-road test. NHTSA's rationale for this is that the dynamic test looks for an untripped rollover tendency, which NHTSA believes are rare in real life.

"We take the position that if a vehicle can tip up in a steering maneuver, without impacting anything else first, then that's a good reason to look for a less tip-prone alternative," says David Champion, director of auto testing for Consumers Union.

"Even supposing that the ratings are good at predicting rollover propensity, you really need to go beyond the stars," says Champion. "There's a big range between the best and worst within each star category. CU will not recommend any vehicle that tips onto two wheels in this test."

NHTSA's rollover ratings can be found at www.safercar.gov. For specific information about a vehicle's star rating, click on the vehicle class, such as SUV, then its year, and the vehicle's name. Scroll down to the heading Rollover, and a chart there will tell you whether the vehicle tipped, and also its likelihood of rollover expressed as an exact percentage rather than a star.

You can also search for a specific vehicle from the main safercars.gov page by selecting its class, year, make, and model.