And Then You're Dead

The Autopsy / Case File · Vehicular & Transit

Good, Fair, Marginal, Poor

Every state now has some version of a graduated driver licensing law for teenagers. Almost none of them have the same one. A national crash study found that the difference between a law rated "good" and a law rated "poor" — measured on a simple point system built from six specific provisions — predicts a 30 percent gap in how many 15-to-17-year-olds die in car crashes in that state. A dozen states still sit closer to the poor end than the good one.

A peer-reviewed national study of all 50 states' crash data from 1996 through 2007 found that states with graduated licensing laws rated "good" under a point-based IIHS scoring system had fatal crash rates 30 percent lower for drivers ages 15-17 than states rated "poor" — 44 percent lower for 15-year-olds specifically, and 41 percent lower for 16-year-olds. The same analysis isolated which individual provisions were doing the work: a ban on any teen passengers cut the fatal crash rate 21 percent compared with no restriction; a night-driving curfew starting at 8 p.m. cut it 20 percent; delaying the licensing age by a full year cut it 13 percent. IIHS's current research finds a dozen states could still halve their teen fatal crash rate by adopting the strongest version of these provisions — worth an estimated 500-plus lives and 9,500-plus collisions prevented nationally each year.

And Then You're Dead · September 2026 · 14 min read

Join the discussion →
A row of hybrid cars marked 'DRIVER EDUCATION' with a college logo, parked in a lot outside a driver's education building
Driver's education vehicles at Portland Community College's Sylvania campus. What a teenager learns before licensure matters less to fatal-crash outcomes, national data shows, than the specific legal restrictions placed on them for months afterward. Photo: Ian Poellet, CC BY-SA 4.0, via Wikimedia Commons.

Every state and the District of Columbia now requires beginning drivers to pass through a graduated driver licensing system before getting a full, unrestricted license — a learner's permit stage under adult supervision, an intermediate stage with specific restrictions once a teenager passes their road test, and finally an unrestricted license. That basic three-stage structure has been close to universal for two decades. What isn't universal, and what turns out to matter far more than whether a state has "a GDL law" in any generic sense, is which specific restrictions the intermediate stage actually contains, how strict they are, and how long they last. A national research team quantified exactly how much that variation matters, using federal crash data and a scoring system built by the Insurance Institute for Highway Safety — and the answer is that it matters more than almost anything else states have tried in traffic safety.

A Six-Line Scorecard

IIHS's rating system, developed in 2000 and updated in 2005, assigns points across six specific provisions. In the learner's phase: a minimum permit age of 16 or older (1 point), a permit-holding period of six months or more (2 points), and a requirement of 30 or more supervised practice hours (1 point). In the intermediate phase: a nighttime driving restriction starting at 10 p.m. or earlier (2 points), a restriction limiting teenage passengers to zero or one (2 points), and each of those two restrictions lasting 12 months or more from the minimum licensing age (1 point each). Sum the points and a state lands in one of four tiers: good (6 or more points), fair (4-5), marginal (2-3), or poor (fewer than 2) — with one hard floor built in regardless of point total: no state can rate above marginal if its intermediate licensing age is younger than 16.

0
lower fatal crash rate for 15-17 year-olds in states with laws rated good vs. poor, per a national crash-data study
0
reduction in fatal crash rate from a zero-passenger restriction alone, compared with no passenger limit

What the Scorecard Actually Predicts

The definitive test of that scorecard came from a research team led by Anne T. McCartt, published in Traffic Injury Prevention in 2010, using Poisson regression on quarterly counts of every state's teenage drivers involved in fatal crashes from 1996 through 2007 — controlling for unrelated state and time trends using the crash rate for drivers ages 30-59, a group no graduated licensing law could have affected. The result: IIHS's law ratings lined up almost linearly with outcomes. For the combined 15-17 age group, laws rated good were associated with a 30 percent lower fatal crash rate than laws rated poor, and laws rated fair were associated with an 11 percent reduction. Broken out by single year of age, the effect was even sharper for the youngest, least experienced drivers: 15-year-olds in "good"-rated states had fatal crash rates 44 percent lower than in "poor"-rated states, and 16-year-olds had rates 41 percent lower. Every one of those figures was statistically significant. And critically, the researchers found no evidence that stronger laws were simply shifting deaths to older teenagers who delayed licensure to dodge the restrictions — a concern raised by earlier, smaller studies — with 18- and 19-year-old drivers in good-rated states showing modest additional reductions of their own, not increases.

A 15-year-old in a state with a "good" rating faces a fatal crash rate 44 percent lower than a 15-year-old one state over with a "poor" one.

Which Specific Lines on the Scorecard Are Doing the Work

The same study went further and isolated the individual provisions, and the results reorder which parts of graduated licensing actually deserve the credit. A restriction banning any teenage passengers produced a 21 percent reduction in the fatal crash rate for 15-17 year-olds, compared with allowing two or more; allowing exactly one passenger produced a smaller but still real 7 percent reduction. Nighttime driving restrictions scaled directly with how early they started: an 8 p.m. cutoff was associated with a 20 percent reduction, 9 p.m. with 18 percent, 10 p.m. with 16 percent, 11 p.m. with 14 percent, midnight with 12 percent, and 1 a.m. with 9 percent — each additional restricted hour buying a measurable additional safety margin. Delaying the age at which a teenager could get a learner's permit, and separately delaying the age at which they could get a full intermediate license, each independently reduced fatal crash rates by about 13 percent per full year of delay, or 7 percent for a six-month delay. What did not show a significant independent effect, once those other factors were accounted for: lengthening the minimum permit-holding period beyond what was already required for the age delay, and increasing the number of mandated practice-driving hours. Those two provisions matter for building a beginner's supervised experience, but the crash data didn't credit them with an independent life-saving effect the way it credited the passenger and nighttime rules.

From Zero States Rated Good to Most States Rated Good — and Still Not Enough

The scale of the shift since graduated licensing began is its own story. Florida implemented the first three-stage system in July 1996. Applying the current IIHS rating retroactively to every state's law as it stood at the end of that year, not one state qualified for a "good" rating — two rated fair, nine rated marginal, and 39 states plus D.C. rated poor. By April 2009, that had inverted almost completely: 31 states plus D.C. rated good, 12 rated fair, 7 rated marginal, and zero states rated poor. National teenage fatal crash rates fell right alongside that legal shift — declining 55 percent for 15-year-olds and 50 percent for 16-year-olds between 1996 and 2007, compared with a 19 percent decline over the same period for drivers ages 30-59, the unaffected comparison group. That gap between the teen decline and the adult-driver decline is the closest thing to a natural before-and-after measurement of graduated licensing's total effect. Even so, the study's authors noted that as of their analysis, no state fulfilled every single one of IIHS's criteria for what it defined as a truly optimal system — meaning even the best-rated states were still leaving some measurable safety margin on the table.

The States Still Closest to "Poor"

That gap hasn't closed. IIHS's current research identifies specific states where adopting the strongest version of every provision would still produce dramatic reductions — evidence that the ratings variation from 2009 persists today in a different form. South Dakota could cut its teen fatal crash rate by an estimated 63 percent by adopting optimal provisions, the largest remaining gap of any state; North Dakota could cut 56 percent, Iowa 55 percent, Montana 53 percent, Arkansas 50 percent, Idaho 49 percent, Mississippi 48 percent, New Mexico 47 percent, Kansas 46 percent, and South Carolina 45 percent. The gaps aren't always where they'd be expected, either: Vermont, despite already banning all teenage passengers, could still cut its fatal crash rate an estimated 20 percent simply by adding an 8 p.m. night-driving restriction it currently lacks, and New Jersey — the only state that licenses at age 17, the single strongest individual provision in the entire system — could still cut fatal crashes an estimated 16 percent by closing an exception in its own passenger-restriction law. Nationally, IIHS estimates a dozen states could halve or more than halve their teen fatal crash rate by matching the strongest provisions already in effect elsewhere, worth more than 500 additional lives and more than 9,500 additional collisions prevented every year.

The Number That Hasn't Moved

Two thousand, eight hundred and ninety-nine teenagers died in U.S. motor vehicle crashes in 2024 — down from 5,819 in 1996, a genuine 50 percent decline that graduated licensing helped drive. But the remaining deaths concentrate in exactly the circumstances the strongest provisions target: 34 percent of fatally injured drivers ages 16-19 in 2024 had a positive blood alcohol concentration, and 35 percent of teen driving fatalities occurred between 9 p.m. and 3 a.m., with well over half occurring between 6 p.m. and 3 a.m. — precisely the hours a strong nighttime restriction is built to remove from a new driver's first year behind the wheel. A rating system built in 2000 from six lines on a scorecard still maps almost exactly onto where today's remaining teen crash deaths cluster.

And Then You're Dead

A crash doesn't check which state a teenager was licensed in. The data does, and it says the difference between a "good" law and a "poor" one — a curfew that starts two hours earlier, a passenger ban instead of a passenger allowance, a license delayed by six months — is worth roughly a third of a state's teenage traffic deaths, year after year, for as long as that gap in the law stays open.

That's it. That's the whole thing.

Comments

Loading…

Comments are anonymous and moderated.

Sources