Safety ratings are usually read one at a time. Headlights here, crash structure there, automatic braking somewhere else, each a separate line on a separate page.
Read that way, a single poor score looks like an outlier. Counted together, they stop looking like outliers.
We took the 305 vehicles rated on at least four of six IIHS evaluations — headlights, child seat anchors, belt reminders, automatic braking, the updated moderate overlap test and the updated side impact test — and counted how many each one failed.
| Tests failed | Vehicles |
|---|---|
| None | 145 |
| One | 89 |
| Two | 27 |
| Three | 35 |
| Four | 6 |
| Five | 3 |
Just under half fail nothing at all. At the other end, 44 vehicles fail three or more, and three of them fail five out of six.
Failures cluster by brand, not by price
| Brand | Vehicles | Average tests failed |
|---|---|---|
| Jeep | 14 | 3.07 |
| Chevrolet | 12 | 1.83 |
| Ford | 14 | 1.71 |
| Audi | 16 | 1.56 |
| Kia | 17 | 1.24 |
| Hyundai | 27 | 0.44 |
| Honda | 22 | 0.41 |
| Mazda | 25 | 0.08 |
Mazda averages 0.08 failures across 25 vehicles — two failures in total, across the entire range. Jeep averages just over three.
Audi in the middle of that table is the point worth pausing on. Price does not predict this. A brand can average a perfect five stars on the ratings everyone quotes and still lose points on headlights and pedestrian braking — the same way five-star SUVs turn out to score three for rollover.
Why clustering happens
Six evaluations, six unrelated engineering problems. A brand failing several is not unlucky — it reflects when the platform was designed and what it was designed against.
A vehicle architecture developed before an evaluation existed fails it, and a manufacturer running a long product cycle carries that across the whole range at once. The brands at the top of this table all sell platforms with long lives and heavy shared parts between models.
The brands at the bottom replace platforms more often, which is expensive and is the actual reason their scores look the way they do.
How to use it when choosing
Look at the count, not the individual scores. One Marginal on a car that is otherwise clean tells you very little. Three failures across unrelated systems tells you about the generation of the design.
And check the specific model year, because the gap between two years of the same nameplate can exceed the gap between two brands — most of these evaluations are recent, and cars redesigned after they appeared behave completely differently from cars redesigned before.
Frequently asked questions
How many safety tests does a typical car fail?
Of 305 vehicles rated on at least four of six IIHS evaluations, 145 fail none and 89 fail one. Forty-four fail three or more, and three vehicles fail five of the six.
Which brands fail the most safety evaluations?
By average failures per vehicle: Jeep at 3.07, Chevrolet at 1.83, Ford at 1.71 and Audi at 1.56. Mazda averages 0.08 across 25 vehicles, with Honda at 0.41 and Hyundai at 0.44.
Does a more expensive car score better?
Not reliably. Audi sits mid-table alongside mainstream brands. What predicts results better is platform age: a design developed before an evaluation existed tends to fail it, and long product cycles spread that across a whole range.
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 →
95% of Dealers Sell One Brand. The Exceptions Are Almost All Stellantis.
The franchised dealer model is usually described as one brand, one building. It mostly is — and the places where it breaks explain a distortion that shows up in every count of dealers by brand.
One GMC Sierra Weighs 2,000kg. Another, Same Year, Weighs 3,919.
A model name is treated as if it described a vehicle. On the specification sheets it frequently describes a range, and the range on some models is wider than the gap between entire vehicle classes.
A Large Sedan Has More Cabin Than a Small SUV. And a Third of the Boot.
The reason people moved from sedans to crossovers is usually described as space. Split the measurement into cabin and boot and the sedan wins one half outright — which means the trade was never about room for people.
Mark ForCar as a preferred source and our data shows up first in your Google results. One click, no account.