Crash Scores Improved Everywhere Except Rollovers

NHTSA scores every vehicle on three things: how it protects you in a frontal crash, how it protects you in a side impact, and how likely it is to roll over. Those three get averaged into the star rating on the advert.

I split them back apart across 2,338 ratings and grouped by era. Two of the three lines go up. One doesn't move at all.

Side impact is now effectively a solved problem — 4.98 out of 5 means almost everything scores full marks. Frontal climbed steadily. Rollover went from 4.11 to 4.08, which is to say nowhere.

The third test measures something different

This is the part worth understanding, because it isn't a story about lazy engineering.

Frontal and side ratings measure protection — what happens to a body inside the cabin when something hits it. That's an engineering problem, and engineers have been extraordinarily good at it: stronger cages, more airbags, structures designed to fold in a chosen order. Twelve years of steady progress, exactly as you'd hope.

Rollover measures likelihood. It's calculated from the static stability factor — track width against centre of gravity height — plus a dynamic manoeuvre test. It doesn't ask how well the car protects you. It asks how easy it is to tip over.

And that's geometry. A tall vehicle has a high centre of gravity because it's tall. No amount of airbag development changes the shape of the thing.

What actually moved the needle — once

Rollover scores did improve dramatically, but before this table starts. Electronic stability control became mandatory on US vehicles from the 2012 model year, and it is the single most effective anti-rollover technology ever fitted to a car. It catches the slide that precedes most rollovers.

That win is already banked in the 4.11 of the 2011–2014 column. Since then there's been nothing comparable, because the remaining problem is the shape of the vehicle — and the market has been buying taller vehicles every year, which quietly pushes the average the other way.

Why it matters more than three-versus-five suggests

Rollovers are rare and disproportionately lethal. They make up a small share of crashes and a large share of occupant deaths, because a car folding onto its roof does something a side impact doesn't: it can collapse the structure above your head, and it defeats how a seatbelt is designed to hold you.

So a vehicle with five stars frontal, five side and three rollover isn't a five-star car with one weak spot. It's a car that protects you very well in the crashes it's likely to have, and less well in the one it's more prone to than a saloon. We listed the models where that gap is widest — the Escalade, Tahoe, Navigator, Expedition and Grand Wagoneer all sit there.

What to do with this

Look up the three scores separately for the exact model and year rather than reading the averaged star rating. NHTSA publishes them individually, and the overall figure hides precisely the number that varies most.

Take loading seriously on anything tall. Roof boxes and rear-heavy loads raise the centre of gravity on a vehicle that already sits high — that's the same factor the test measures, and you're making it worse.

And on a used vehicle, confirm stability control is fitted and working. Pre-2012 cars may not have it, and it's the one intervention that genuinely changed these numbers. A free VIN check will show any open recalls on the stability and restraint systems, which are exactly the campaigns that go unclaimed when a car changes hands.