American safety regulators are expected to start putting sport-utility vehicles through actual rollover tests instead of mathematical calculations as concern about the safety of the vehicles grows.

American safety regulators are expected to start putting sport-utility vehicles through actual rollover tests instead of mathematical calculations as concern about the safety of the vehicles grows.

According to the , the rollover scores from the “static stability” formula are calculated by dividing half of a vehicle’s width by the height of its centre of gravity to determine a rollover result.

The move towards an actual test comes after the Firestone and Ford Explorer recall in 2000 due to a high number of rollover accidents. American research in 2001 also found that an SUV occupant is three times more likely to die as a result of a rollover than an occupant in a car.

David Champion, director of automobile testing for Consumer Reports, said the new rollover test is likely to include various manoeuvres and turns.

Manufacturers, however, are concerned that the test will not be accurate as real-life accidents depend on so many different scenarios. Said Eron Shosteck, a spokesman for the Alliance of Automobile Manufacturers: "It's very difficult to replicate the real-world conditions that affect a vehicle's propensity to roll over. There are two dozen factors."

"We hope that any test is repeatable," said General Motors spokesman Mike Morrissey. "There's an issue of variability in a driver-dependent test."

The National Highway Traffic Safety Administration would not say when the tests could change, but the newspaper speculated as early as September. The problem is working out a way to factor in driver behaviour, road conditions and vehicle characteristics.