By this time, most people know that a sport utility vehicle has a greater chance of rolling over than a normal passenger car and that rollovers have a higher fatality rate than other crashes. The federal government tells how to find which SUVs are most likely to roll over, but critics say it should be easier.

Type of vehicle is, of course, not the only cause of rollovers. Speed, driver distraction and alcohol also figure in the mix. And rural driving, on roads with no guardrails and sometimes no center lines, is another factor, especially in our part of Maine. Many factors played in the tragic rollover two months ago, in which two mothers, their four children and a third woman were killed on Interstate 95 near Carmel. Still, vehicle type is worth considering, and narrow frames with high centers of gravity are the worst offenders.

To find on the Internet whether a particular vehicle has a tendency to tip up on two wheels in a sudden turn, go to the National Highway Traffic Safety site, www.safercar.gov. Click on SUV or pickup or some other class of vehicle, click on the model you want to know about, scroll down to the bottom to find, under rollovers, and see whether it says “tip” or “no-tip.” If it says “tip,” the vehicle went up on two wheels when a test driver, going 35 to 50 miles an hour, turned sharply to the right and then overcorrected to the left.

Consumer groups complain that some SUVs that tip are still given three stars in the NHTSA scale, up in some cases from two stars before tipping tests began. Three stars mean a 20-to-30 percent chance of a rollover in a single-vehicle crash. The best rating, five stars, means less than a 10 percent chance of rolling over. No passenger cars tested received less than four stars, meaning a 10-to-20 percent chance of a rollover.

Critics say the rollover ratings still rely too heavily on a vehicle’s dimensions and not enough on the rollover test. But NHTSA is resisting demands to give more weight to the road test, since the test measures vehicle behavior in only an “untrippable” rollover. The agency has not yet devised a test for the far more likely crashes in which a vehicle is “tripped” by something like a curb, soft mud or a guardrail.

So the rollover ratings will never be perfect, but they can help in selecting a vehicle. It is just as important to remember that rollovers account for nearly one-third of all deaths from passenger vehicle crashes. Ten thousand people died in rollover crashes in 2002 alone, and 72 percent of them were not wearing safety belts.