Your Car Was Crash-Tested With a Man’s Body. 8,600 Women Paid for That With Their Lives.

Every vehicle sold in America was crash-tested using a dummy modeled on a 170-pound male. The so-called “female” dummy, a Hybrid III 5th-percentile unit dating to 1988, stands 4’11” and weighs 108 pounds.^[1] Ninety-five percent of actual American women are taller and heavier than this thing. It has no accurate pelvis, no representative chest geometry, and no ability to measure the spinal injuries that disproportionately kill women in frontal collisions. Women face 20% higher fatality risk in comparable crashes, even after controlling for vehicle type, airbag deployment, and crash severity.^[2]

Doing the Grim Math

IIHS reports approximately 53,315 female passenger vehicle driver deaths between 2014 and 2023.^[3] That PMC relative risk of 1.20 means the equalized death count would have been about 44,429, which leaves roughly 8,886 excess female driver deaths over the decade. But NHTSA found that newer vehicles (model year 2010 and later) shrank the gender fatality gap from 18% to 6.3%,^[4] so the blended fleet-wide excess sits closer to 12%. Apply that conservative figure across the full female driver cohort: 53,315 minus 47,603 equals 5,712 excess driver deaths. Add roughly 27,000 female passenger deaths over the same window, apply the same 12% correction, and the total excess female occupant deaths reaches approximately 8,600 across the decade.

One woman every 10.2 hours, for ten years, killed in part because the dummy strapped into the car during its safety certification did not represent her body.

The Fix Exists. Nobody Mandated It.

NHTSA approved the THOR-05F in November 2025.^[5] It costs a million dollars per unit, carries 150 sensors, and for the first time in federal crash testing history, it actually models female anatomy. Terrific. But mandating its use in the New Car Assessment Program? That won't happen before 2027 or 2028 at the earliest, according to NHTSA's own rulemaking timeline. Meanwhile TU Graz published data in May 2026 confirming women are 1.6 times more likely to be injured in any car crash, with the gap more than doubling at low speeds and widening further for occupants over 50.^[6]

Newer cars are narrowing the gap on their own, which is genuinely encouraging, and also entirely insufficient. A 6.3% excess risk in the newest vehicles still compounds across millions of female occupants every year, and the average vehicle on American roads is 12.6 years old, meaning most women are still driving cars designed around a male body with a rubber prosthetic bolted on.

What You Should Do

Choose vehicles rated IIHS Top Safety Pick+ from model year 2018 or later, because those designs reflect post-ESC and post-AEB maturity, and the gender gap is measurably smaller in that cohort. Wear your seatbelt with the shoulder strap crossing your collarbone and the lap belt low across your pelvis. No reclined positions. No blankets or coats under the belt; compressed material during a crash is absorbed slack your restraint system cannot recover. If you're shopping for your teenage daughter's first car, skip the used 2012 Corolla and stretch for something with current-generation side curtain airbags and a front-center airbag. Run your VIN at nhtsa.gov/recalls before you take delivery.

Strongest Counterargument

NHTSA's own data shows the gap is shrinking. In newer vehicles the excess risk drops to 6.3%, down from 18% in older designs, and the fleet is modernizing steadily. Automakers are already voluntarily incorporating female crash test protocols beyond federal requirements, and some manufacturers now run internal tests with biofidelic female anthropomorphic test devices before NHTSA's mandate takes effect. Give the system another decade of fleet turnover and the problem partially solves itself through incremental engineering.

That counterargument is correct on trajectory and wrong on urgency. Self-correction at the pace of fleet turnover means another eight to twelve years of compounding excess deaths. The THOR-05F is ready now. NHTSA has authority under 49 U.S.C. § 30111 to set motor vehicle safety standards, and the Administrative Procedure Act allows agencies to bypass the standard notice-and-comment process when “good cause” exists under 5 U.S.C. § 553(b)(B). Choosing not to expedite is a policy decision, not a technical constraint.

Limitations

Our excess-death calculation rests on a blended 12% figure interpolated between NHTSA's 18% (pre-2010 fleet) and 6.3% (post-2010 fleet) benchmarks; the actual fleet-age distribution shifted annually during the 2014-2023 window, and we assumed a linear blend rather than modeling yearly fleet composition. FARS captures only fatal crashes, which constitute roughly 0.5% of all reported crashes; the excess injury toll, which TU Graz data suggests is substantially larger, is invisible in this analysis. Some portion of the gender disparity traces to biological differences in bone density and ligament structure rather than crashtest design bias, though the measurable improvement in newer vehicles indicates engineering is the primary addressable lever. Passenger death data was estimated from IIHS aggregate statistics rather than individual-record FARS queries, introducing additional approximation error in the total 8,600 figure.

Data sourced from NHTSA FARS 2014–2023, IIHS fatality statistics, PMC peer-reviewed research, Stanford Gendered Innovations, and TU Graz. Excess death estimates use blended fleet-age interpolation with acknowledged approximation error; consult the methodology section for additional caveats on rate calculations and fleet-size estimates.