A side impact test needs something to hit the car with, and that something has to represent a typical striking vehicle. The original barrier was built around a mid-2000s SUV: a particular weight, a particular bumper height.
The fleet moved on. Ride heights rose, kerb weights rose, and the barrier stopped representing anything on the road. IIHS rebuilt it heavier and drove it faster.
| Vehicles rated under both versions | Count |
|---|---|
| Rated Good on the original test | 338 |
| Still Good on the updated test | 106 |
| Dropped to Poor or Marginal | 133 |
Same cars, same structures, different barrier. A hundred and thirty-three of them went from the top rating to the bottom half.
Why weight matters more here than anywhere else
In a frontal crash the car has a crush zone in front of the occupant — a metre or more of structure designed to absorb energy before anything reaches the cabin.
In a side impact there are a few centimetres of door. Everything else has to be handled by the pillars, the seat structure and the curtain airbag, and none of those can absorb energy the way a crumple zone does. Add mass and speed to the striking object and the margin disappears quickly.
That is why this single change produced a collapse in ratings while the rear passenger update needed a whole new dummy to do the same thing.
It is being fixed, and quickly
| Model year | Rated | Poor or Marginal | Share |
|---|---|---|---|
| 2023 | 135 | 23 | 17% |
| 2024 | 163 | 16 | 10% |
| 2025 | 178 | 11 | 6% |
| 2026 | 167 | 9 | 5% |
From 17% to 5% in three model years. The fixes are structural — stronger B-pillars, better door beams, curtain airbags that cover more area for longer — and they are the kind of change that appears with a model update rather than mid-cycle.
The part nobody can engineer around
The reason the barrier was made heavier is that the vehicles hitting people got heavier, and that is a fleet problem rather than a design problem.
Our own specification records show the heaviest vehicle on sale weighs as much as five of the lightest. A test can measure how well a small car survives being struck by a large one. Nothing in the test changes the fact that it is being struck by a large one.
What to check on a used car
Any rating from before the update was measured against the old barrier. If you are comparing a 2020 and a 2024 of the same model, their side impact scores are not on the same scale, even though they look identical in a listing.
Look for whether the rating is the updated version specifically, not just the score. And on older vehicles, curtain airbag coverage is the single specification most worth confirming, because it is the component doing most of the work in exactly the crash this test recreates.
Frequently asked questions
What changed in the updated side impact test?
The striking barrier was made heavier and driven faster, to represent what vehicles on the road now actually weigh rather than a mid-2000s SUV. Of 338 vehicles rated Good under the original test, only 106 held that rating and 133 fell to Poor or Marginal.
Why does a heavier barrier cause so many failures?
A side impact has almost no crush zone. A frontal crash has a metre of structure to absorb energy before the cabin; a side impact has a few centimetres of door, with pillars, seat structure and curtain airbags doing the rest. Extra mass and speed erase that margin quickly.
Can I compare side impact ratings between model years?
Not directly across the update. A score from before the change was measured against the old barrier, so a 2020 and a 2024 rating are not on the same scale. Check whether the rating is the updated version, not just the grade.
Denis Kataev is a serial solopreneur and the founder of ForCar. With 15+ years in software engineering and 10 years in SEO, he builds data-driven products end to end — backed by a sharp eye for design. At ForCar he mines proprietary vehicle datasets, turning raw numbers into buying advice you can actually trust.
Keep reading
All articles →
Automatic Braking Is on Almost Every New Car. One in Five Systems Rates Poor or Marginal.
Two cars can both advertise automatic emergency braking and perform completely differently when tested against a pedestrian at night. The badge on the window sticker says nothing about which one you have.
A Third of Cars Have Headlights Rated Poor or Marginal. At Cadillac, Four in Five.
Crash stars measure what happens during an impact. Headlight ratings measure whether you saw the thing you hit. The two scores disagree more often than most buyers realise, and one of them almost never appears in a listing.
The Average Jeep Fails Three Safety Tests. The Average Mazda Fails None.
Any single safety test can be an outlier. Counting failures across six separate evaluations per vehicle removes that excuse, and the results separate the market far more sharply than any individual score.
Mark ForCar as a preferred source and our data shows up first in your Google results. One click, no account.