SafeCarCompare does one thing that car safety sites rarely make easy: translating crash-test injury measurements into plain English. Two vehicles can both score 5 stars while producing very different forces on the crash dummy inside — one leaving 40% of the injury budget unused, another using 95% of it. SafeScore makes that margin visible. Every score is built entirely from U.S. government crash-test data, is fully reproducible, and is never sold or licensed from a private organization.
All SafeCarCompare data is sourced from U.S. government databases. Government data is public domain — it belongs to the public, not to a private organization.
SafeScore is based on three ATD sensor measurements from the NHTSA frontal crash test, each matched against a federal injury threshold established under FMVSS 208:
Each measurement is expressed as a percentage of its federal injury threshold, the three percentages are averaged, and the result is subtracted from 100. Equal weighting is used because each measurement is already normalized — no invented medical weights.
Formula: SafeScore = 100 − mean( HIC15÷700×100, chest_mm÷63×100, NIJ÷1.0×100 ). A score of 80 means the vehicle used an average of 20% of the federal injury threshold — leaving 80% of the margin unused. A score of 55 means it used 45% of the threshold on average, leaving less margin.
Score scale: 70–100 = Excellent injury margin. 65–69 = Strong. 58–64 = Good. 50–57 = Limited. Below 50 = Review carefully.
The front-row rule: when a test includes both driver and passenger dummies, SafeScore uses the lower of the two scores — never the average. A car should protect both seats, not just the better one.
Not every vehicle has complete ATD sensor data for all three measurements. When one or more measurements are unavailable, SafeScore is calculated from whichever measurements are present — null values are excluded, never treated as zero or penalized. A vehicle is not disadvantaged for data NHTSA has not published. The vehicle page notes which measurements were available for each score.
NHTSA does not retest every vehicle every year. When a vehicle year has not been independently tested, SafeCarCompare checks whether NHTSA has associated it with a previous tested year in the same vehicle generation. When fallback data is used, the vehicle page clearly shows the tested year the data was drawn from so users can assess its recency. Three data tiers are possible:
Separate from SafeScore, the Prevention Score rates how well-equipped a vehicle is to avoid crashes before they happen. It is based entirely on NHTSA-reported safety technology data for the tested trim. Feature weights reflect the strength of evidence for each feature's crash-reduction effectiveness:
Prevention Score = sum of confirmed feature points ÷ sum of known feature points × 100. Unknown features are excluded — never treated as absent. A vehicle is not penalized for data NHTSA has not reported.
Base-trim caveat: NHTSA tests a specific vehicle variant — often a base or fleet trim — which may not include optional safety technology available on higher trims. Where data suggests a base trim was tested, SafeCarCompare displays a note. Always verify your specific trim's feature list with the manufacturer.
The Insurance Institute for Highway Safety (IIHS) publishes valuable crash test research, including Top Safety Pick awards and driver death rate studies. IIHS and its sister organization HLDI are private non-profits, and their compiled datasets are subject to usage terms that restrict commercial redistribution without permission. To keep SafeScore transparent, reproducible, and fully public-domain, SafeCarCompare builds exclusively on NHTSA data. The same inputs always produce the same score, and every underlying measurement can be verified directly at api.nhtsa.gov. IIHS findings are referenced in editorial context where appropriate but are not incorporated into SafeScore.
Crash tests simulate one specific scenario at one speed. A vehicle that performs well at 35 mph frontal impact may behave differently at higher speeds, in a rollover, or against a significantly heavier vehicle. NHTSA tests a specific trim — often a base variant — and higher trims with additional airbags or structural reinforcement may perform differently. SafeScore does not account for seatbelt use, road conditions, pre-crash speed, or the weight of the other vehicle. Physics favors the heavier vehicle in a multi-vehicle crash regardless of crash test performance.
Independence: SafeScore is calculated by SafeCarCompare and is not issued, endorsed, or approved by NHTSA, IIHS, or any automaker. All underlying data is sourced from U.S. government databases and is independently verifiable.
SafeCarCompare shows injury margins from NHTSA crash-test data — beyond star ratings. Enter any two vehicles to see head, chest, and neck injury margins side by side.