How we label recall severity

Why we label severity at all

NHTSA recall data doesn't come with a built-in severity score. Each recall has a manufacturer-written description of the defect and the hazard it creates, but that text isn't structured for quick scanning across thousands of campaigns. So we built a mapping.

We borrowed the Class I, II, III framework from FDA-style recall classification because it's a familiar shorthand. It's not something NHTSA assigns. We apply it ourselves, after the fact, to make our pages easier to scan. Think of it as a sorting tool, not an official rating.

What the classes mean here

• Class I: The hazard description points to a risk of serious injury or death. Steering, brakes, airbags, fuel leaks, fire risk, that kind of thing.
• Class II: The hazard could cause injury but the language suggests lower severity or lower likelihood. Lighting failures, some electrical faults, minor structural issues.
• Class III: The issue is largely about compliance or a defect unlikely to cause injury on its own. Missing labels, paperwork mismatches, some emissions items.

These are buckets, not a spectrum with fine gradations. A recall lands in Class I or it doesn't. We don't publish sub-scores or weighted risk numbers.

Why the mapping is approximate

Vehicle recalls and drug or device recalls aren't the same animal. FDA classification was built for products with dosing, contamination levels, and clinical outcomes. Cars don't work that way. A single component defect can behave differently depending on how the car is driven, how old the part is, or what else is going on with the vehicle at the time.

On top of that, manufacturer hazard descriptions vary a lot in how they're written. Some are blunt about worst-case outcomes. Others are cautious and legalistic even when the underlying defect is serious. We read the description and assign a class based on the language and the component involved, but this is a judgment call applied at scale, not a lab measurement.

That means two recalls that feel similar to a reader might land in different classes, or the reverse. We'd rather be consistent in method than pretend to a precision the data doesn't support.

What the label is not

A severity class is not a probability. It doesn't tell you how likely your specific car is to experience the defect. It doesn't tell you how many units failed. It's not a substitute for reading the actual hazard description, which explains the mechanism: what part fails, what happens when it does, and under what conditions.

If you're deciding whether to act on a recall, read the description. It's short, usually a few sentences, and it tells you the thing our class label can only summarize.

What to actually do with a recall

The severity label is a sorting aid on our pages. It's not a reason to worry about a car you haven't checked. Recalls apply to specific VIN ranges within a model year, not every car built that year.

To find out if your car is covered:

• Go to nhtsa.gov/recalls and enter your VIN. It's free.
• If your VIN is listed, the repair is free at any franchised dealer for that make.
• If it's not listed, the recall doesn't apply to your car, regardless of what class we gave it.

We built this severity system to help you scan faster. Use it that way, then go read the description before you decide anything.