Every Vehicle Class Drinks at the Same Rate. Only One Class Walks Away.

According to the toxicology reports, and there are 490,736 of them, America's drivers are remarkably consistent about one thing: every vehicle class shows impairment rates between 18% and 22% in fatal crashes.^[1] Sports car drivers. Sedan drivers. Pickup drivers. Van drivers. All of them drink and use drugs at roughly the same rate. A 4.4 percentage-point spread across five vehicle classes. Behavioral uniformity that would make a statistician weep.

Then those same drivers crash. And the uniformity vaporizes.

Lethality, measured as deaths per crash involvement, ranges from 48.9% for pickups to 68.2% for sports cars. A 19.3 percentage-point spread.^[1] The behavioral input varies by 4.4 points. The engineering output varies by 19.3. That is a 4.4-to-1 amplification ratio. Same driver behavior in, wildly different body bags out.

Class-by-class breakdown

Sports cars: 22.5% impaired, 68.2% lethality. Highest impairment, highest lethality. No surprise there, but note the scale. A sports car driver who crashes is 1.4x more likely to die than a pickup driver who crashes, despite only being 1.1x more likely to be impaired.^[1]

Sedans: 20.4% impaired, 64.5% lethality. Nearly two-thirds of sedan occupants in fatal crash involvements die. That is 197,584 FARS records and 89,127 deaths.

SUVs: 19.5% impaired, 52.4% lethality. Twelve points better than sports cars on lethality. Two points better on impairment. Engineering doing the heavy lifting.

Vans: 18.1% impaired, 51.4% lethality. Lowest impairment, second-lowest lethality. Minivan drivers are the most sober class in America and crash survival reflects it, but mostly because vans weigh 4,500 pounds.

Pickups: 20.1% impaired, 48.9% lethality. Middle of the pack on impairment, lowest on lethality by a comfortable margin. A pickup driver in a fatal crash has better than coin-flip odds of walking away. A sports car driver does not.

Individual vehicles make it worse

Zoom from class averages to specific models and the gap explodes. A Chevrolet Cavalier driver: 78% sober, 85.7% lethality. A Ram 2500 driver: 80% sober, 20.5% lethality.^[1] Equivalent sobriety. Both drivers did their part. One is 4.2x more likely to die. Not because of behavior. Because of what they sat inside.

IIHS research on vehicle size and weight confirms the mechanism: in two-vehicle crashes, the lighter vehicle absorbs a disproportionate share of crash energy.^[2] A 3,000-pound sports car hitting a 6,000-pound pickup at equal speeds experiences roughly four times the deceleration. Physics does not grade on a curve.

Novel finding: the behavioral constant

No published analysis we could find cross-references FARS impairment rates by vehicle class against occupant lethality ratios. The individual datasets exist. NHTSA publishes impairment rates. Lethality can be computed from FARS fatal crash involvements and death counts. But the comparison, showing that driver behavior is functionally constant while engineering outcomes diverge by 4.4x as much, appears to be original.

This matters because it reframes the intervention target. If you assume impairment causes fatal crashes, the solution is behavioral: enforcement, education, ignition interlocks. If impairment rates are near-identical and lethality is what diverges, the lever is structural. Heavier vehicles. Better crumple zones. Mandated ESC.^[3] You cannot engineer away the drinking. You can engineer away some of the dying.

Strongest counterargument

Sports car drivers engage in riskier driving independent of substance use. Speed, aggressive lane changes, later braking points. FARS toxicology captures alcohol and drugs but not reckless driving as a standalone variable. A 25-year-old in a Corvette doing 95 mph while stone sober will produce a higher-severity crash than a 45-year-old in a Suburban doing 65 after two beers. Impairment rates being similar does not mean overall risk-taking is similar. Vehicle choice may correlate with driving style in ways toxicology cannot measure.

This is a legitimate confound. We cannot isolate vehicle engineering from driver behavior entirely using FARS alone, because FARS only captures fatal outcomes, not the full distribution of driving behavior that leads to those outcomes.

What you can do

If you are shopping for a vehicle and safety matters to you: weight saves lives, at least for the vehicle's own occupants. Check IIHS crashworthiness ratings at iihs.org/ratings and pay attention to the driver death rate data at iihs.org/topics/fatality-statistics. A midsize SUV or crossover in the 4,000 to 5,000-pound range offers the best balance of own-occupant survival and manageable risk to other road users. If you currently drive a lightweight sports car or aging compact sedan, understand that your crash survival margin is measurably worse than a heavier vehicle, even controlling for driver sobriety. Check your VIN for open safety recalls at nhtsa.gov/recalls.

Methodology

Impairment rates from NHTSA FARS toxicology data, 2014-2023: BAC > 0 or drug-positive test result classified as "any impaired." Vehicle class bins (sports car, sedan, SUV, pickup, van) follow FARS body type codes. Lethality ratio = deaths in vehicle class / fatal crash involvements for that class. Total sample: 490,736 drivers across five classes. Individual vehicle comparisons (Cavalier, Ram 2500) use model-level FARS data with minimum 50-death threshold. Amplification ratio (4.4x) computed as lethality spread (19.3pp) divided by impairment spread (4.4pp).

Limitations

FARS captures fatal crashes only, not all crashes or all injuries. A vehicle class with low fatal-crash lethality might still produce high injury rates in non-fatal crashes. Drug testing protocols vary dramatically by state: some test every fatal-crash driver, others test selectively. This means drug impairment is underreported and unevenly measured. Alcohol testing is more standardized but still incomplete. Lethality ratio treats all crash involvements equally regardless of single-vehicle vs. multi-vehicle configuration, crash speed, or road type. Vehicle class bins aggregate heterogeneous fleets: "sedan" includes both a 2,800-pound Civic and a 4,200-pound Charger. Class-level impairment similarity does not establish that individual-level risk behaviors are identical.