Methodology

Model-year coverage

Coverage starts at model year 2010 and never rolls forward. This matches NHTSA's 15-year window for VIN-level recall lookups, so URLs stay stable and cumulative rather than retiring older years.

VIN lookups and the 15-year window

NHTSA's recall API returns open recalls for a vehicle going back roughly 15 model years. Older vehicles may have had recalls that no longer show up in a VIN lookup, and pages note that where it matters. A VIN check shows open recalls, meaning the campaigns that have not yet been repaired on that specific vehicle.

Cleaning up the make and model names

The raw NHTSA recall files list makes and models exactly as filed, which produces duplicates and trim-level noise. One model can appear under dozens of spellings and trim codes, and Mercedes alone files over 300 distinct model strings. A deterministic engine maps these to a stable set of consumer makes and clean model names. It matches make spellings to a fixed list of passenger-vehicle brands, drops commercial trucks, buses, RVs, and motorcycles, and folds trim, drivetrain, and powertrain variants into one model. Every run flags any string it could not place and any model added or removed since the last update, so coverage stays current and nothing is guessed.

How we count vehicles affected

NHTSA's "potentially affected" figure covers the whole recall campaign and is repeated for every vehicle line it touches. We never add these up across a model's recalls, because that would count the same vehicle many times over. Instead we show the size of the single largest recall, plus the affected count for each individual campaign.

How the safety score works

Each make and model carries a 0 to 100 safety score and a letter grade (version 1). It grades how serious a vehicle's recalls are, not how many of them there have been. Recall data carries no sales figures, so a score based on the raw count would just rank cars by popularity. A best-seller racks up more recalls because millions of them are on the road, not because it is less safe.

The score looks at how serious and how recent the recalls are, using proportions instead of raw counts. It weighs the average severity of each recall, where a do-not-drive advisory counts most, then park-outside and fire risk, then critical safety systems such as airbags, brakes, steering, and fuel. It also factors in how often recalls carry a do-not-drive warning and how many landed in the last three years. A small nod to recall frequency only breaks ties between vehicles with a similar severity profile. Every vehicle is then ranked against the rest, so a higher score means a lower recall-severity risk than that share of vehicles. A make's score is the recall-weighted average of its models. The score is one lens on public recall data, not a verdict on overall vehicle safety, and not legal or purchasing advice.

Corrections

All figures trace back to official NHTSA records. See the data sources page for the specific files and APIs used.