← The Crash Report
By The Numbers

Half of American Drivers Think Obeying the Speed Limit Is the Dangerous Part

Night highway with light trails from speeding traffic

The AAA Foundation for Traffic Safety spent last month asking American drivers how they really feel about speeding, and the answers read like a group therapy transcript for a nation that has quietly legalized it. Forty-nine percent agreed with the statement: "It's safer for me to drive with the flow of traffic than obey the speed limit."[1] Not "more convenient." Safer. Half the country has decided the law is the hazard.

49%
of U.S. drivers say it is safer to speed with traffic than to obey the posted limit

It gets better. Another 49% agreed that "going a little bit over the speed limit isn't really speeding," which is a bold redefinition of a word with a legal meaning, and 44% endorsed the deeply scientific proposition that "when I do speed, I don't go so fast that I worry about dying if I get in an accident."[1] Researchers call this moral disengagement. The rest of us call it Tuesday on I-95.

Strip away the psychology jargon and the survey describes a functioning social contract. Drivers report a shared threshold: about 10 mph over the posted limit is acceptable and unlikely to draw enforcement, with tickets expected to start around 10 to 12 over depending on the road, and speeds up to 25 over deemed acceptable "in certain circumstances."[1] Read that plainly and the speed limit on a 65-mph highway is not 65. It is 75, ratified by every cruiser that waves you past at 74. Everybody knows the real number. Including the cops.

Now the physics receipt, which nobody in the coverage bothered to run. Crash energy scales with the square of velocity, so 10 mph over a 65 limit is not 15% worse. At 75 in a 65, your car carries (75/65)² = 1.33 times the kinetic energy: 33% more energy to dissipate through sheet metal and sternum, and 33% longer braking distance. At 77, the edge of the enforcement tolerance band, it is 40% more. At 90 in a 65, the "acceptable in certain circumstances" special, it is 92% more. Nearly double. Your vehicle does not care what the culture voted for.

The study sorted drivers into five mindsets, and only 21% landed in "Strict Compliance Advocates."[1] Everyone else, all 79%, carries a personal philosophy of speeding, from the Principled Pragmatists (26%, rules matter but so does context) to the Autonomy-First crowd (16%, my foot, my choice). Most revealing is the deterrence hierarchy: asked why they slow down, drivers ranked avoiding a ticket (79%) above avoiding a crash (72%).[4] Enforcement beats physics in the driver's mind, every time. Which is worth remembering the next time someone argues cameras don't change behavior. Nearly three in four drivers support automated enforcement in areas with high crash potential, and 88% support more police presence there.[1] The public has pre-approved the intervention. We just haven't installed it.

Here is the part where intellectual honesty requires me to say the 49% are not entirely wrong, because the strongest version of their argument deserves daylight. Speed variance kills. Research going back to the Solomon curve in 1964 shows crash risk climbs as you deviate from the mean speed of traffic, so doing 55 in a 75-mph flow is genuinely dangerous, and speed limits set well below the 85th-percentile speed invite exactly the contempt this survey documents. The report itself flags speed-limit credibility as a core theme.[1] "Obey harder" was never going to fix a coordination problem. Uniform compliance would, which is precisely what advisory intelligent speed assistance and automated enforcement produce, and both poll better than the scolding. Drivers even prefer the nanny that warns over the one that restrains: support for in-vehicle speed tech runs from about a quarter to over half depending on how invasive it gets.[1]

So what do you do with this? First, know the band and its price: the tolerated zone tops out around 10 to 12 over, and every mph inside it is still physics. If you must surf it, 9-over costs you roughly 30% more crash energy, and you should know you are paying it. Second, retire "flow of traffic" from your vocabulary as a defense. It fails in court, it fails in physics, and the correct response to a speeding pack is distance, not membership. Third, if your car offers advisory intelligent speed assistance or an adjustable limiter, switch it on; it is the rare countermeasure drivers say they actually want. Fourth, when your city debates speed cameras at the intersection where people keep dying, show up: about 70% of drivers nationally already support them there.[4] Speeding was a factor in nearly 30% of 2023 traffic fatalities.[1] That is not a fringe behavior problem. It is the mainstream, doing 75 in a 65, convinced the sign is the dangerous part.

Sources & References

  1. AAA Foundation for Traffic Safety (Steinbach, McDonough, Hungund, Kasha & Bai), Attitudes Toward Speeding: Insights from Focus Groups and a National Survey of U.S. Drivers, Technical Report, August 2026. aaafoundation.org (PDF)
  2. AAA Foundation for Traffic Safety, 2025 Traffic Safety Culture Index : Report Summary, June 2026. aaafoundation.org (PDF)
  3. Rochester Democrat and Chronicle, “How do drivers justify speeding? New study gives clues,” August 19, 2026. democratandchronicle.com
  4. Insurance Journal, “Despite Risk, Speeding Becoming Normalized, AAA Study Finds,” August 24, 2026. insurancejournal.com
  5. NHTSA, Fatality Analysis Reporting System (FARS) speeding-related fatality estimates, via AAA Foundation (2026). nhtsa.gov

Source: AAA Foundation for Traffic Safety, August 2026 national survey + focus groups. Method: kinetic-energy ratios computed as (v/limit)² from the survey's reported social thresholds (10, 12, and 25 mph over a 65-mph limit). Limitations: survey measures self-reported attitudes, not observed behavior; the moral-disengagement scale was purpose-built for this study; the survey does not link attitudes to crash outcomes; NHTSA's speeding-fatality share is officer-coded at the crash scene. See methodology for caveats.