When people compare vehicle safety, one statistic can sound especially compelling: how many drivers died in this vehicle? The Insurance Institute for Highway Safety publishes driver death rates by make and model, showing how often drivers of specific vehicles die in real-world crashes. The data is serious and worth understanding. But a high death rate does not automatically mean a vehicle is poorly engineered — and a low death rate does not automatically mean a vehicle protects every driver better in the same crash.
IIHS driver death rates are reported as driver deaths per million registered vehicle years — one vehicle registered for one full year. That normalization matters: a popular vehicle is not automatically penalized just because more of them are on the road. IIHS also adjusts the rates for driver age and gender, which corrects for the fact that younger male drivers appear disproportionately in crash statistics and tend to buy certain types of vehicles.
But the metric has important limits. IIHS is explicit that death rates are not adjusted for how fast people drive or how many miles they travel per year. Those factors can substantially influence outcomes.
A high-performance sports coupe may attract a different type of buyer than a minivan or family crossover. It may be driven faster, driven late at night more often, driven on different roads, and involved in different types of crashes. IIHS has acknowledged this directly — when discussing vehicles with high driver death rates among performance cars, IIHS has pointed to vehicle image and marketing as contributors to crash risk, not just the engineering. A vehicle's death rate can partly reflect the culture and driving patterns around the vehicle.
Real-world fatality data captures something controlled crash tests cannot: the physics of real roads. A subcompact car may perform well against a standardized barrier, but on real roads it collides with full-size pickups and large SUVs that weigh twice as much. That mismatch affects outcomes. This is why real-world death rates often show smaller, lighter vehicles with higher rates — it reflects actual road conditions, not engineering failure.
Controlled crash tests are designed to isolate vehicle performance from driver behavior. A standardized frontal barrier test at 35 mph runs the same conditions for every vehicle. The crash dummy experiences whatever forces the vehicle structure and restraint system produce. That allows a direct engineering comparison that real-world data cannot cleanly provide. SafeScore shows how much injury margin the dummy had — not what happened across thousands of real-world drivers with different behaviors and conditions.
Driver death rates are best used as a context signal rather than a final engineering verdict. They can reveal patterns that crash tests alone may not show. But a high death rate may be caused by weak crash protection, small size, lack of crash avoidance features, risky driving patterns among typical buyers, or some combination. Real-world death rates tell you what happened across many drivers. Crash-test injury data tells you what happened to the dummy in a controlled test. Neither is perfect. Together they give a fuller picture.
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.