Think there's not much difference? There is when it comes to driving.

At 16 1/2, he looks about the same as he did 6 months ago, when he got his driver's license. But there's a substantial change when it comes to the likelihood of a crash. According to two new studies, it's immediately after teenagers get their full-privilege licenses and begin to drive without supervision that their crash rates are highest. Then the rates go down during subsequent months so that, by the time beginners have been driving unsupervised for 6 months, their crash rates are 40 percent lower than when they first got their licenses. This is the main finding of a new study conducted in Nova Scotia.

It's not that 16 1/2 is a magic age. It's that getting a few months of experience behind the wheel helps a lot. Then the decline in crash risk continues more gradually over the next year and a half.

Supervised drivers have few crashes:

The big problem isn't among the least experienced drivers on the road, those with learner's permits. Other drivers may give them a wide berth, but learning drivers who are supervised by licensed adults have very few crashes. It's when the beginners start to drive unsupervised that crash rates go up.

"The low-risk learner's period followed by a spike in crash rates when drivers first get licenses indicates the need for a graduated system," says Dan Mayhew of Canada's Traffic Injury Research Foundation, lead author of the Nova Scotia study. He's referring to the graduated licensing systems for young beginning drivers that are being adopted in the United States and Canada.

"There's a powerful argument for extending the period of supervision and then phasing in full driving privileges, keeping new license holders out of the highest risk driving situations." This is what graduated licensing does by extending the process of earning full privileges.

U.S. Study Shows First Month Is Worst:

For another study conducted in the United States, researchers looked at the crash rates of young drivers in relation to the number of months and miles driven since licensure. They found that the risk of a crash during the first month of licensure is substantially higher than during any of the next 11 months. The likelihood of a citation also is higher during the first month than in any month the rest of the year.

When viewed as a function of cumulative miles driven, the risk of a crash or a citation is highest during the first 500 miles after getting a full-privilege license to drive without supervision.

"The findings show teenagers face a substantially higher crash risk during the first weeks and the first miles of licensure, when most teens begin to drive independently for the first time," says Anne McCartt of the Preusser Research Group, lead author of the U.S. study.

Findings of the U.S. and Nova Scotia studies aren't confined to North America. A 1998 study of new drivers in Norway, where the licensing age is 18, reported a sharp decline in crash risk per kilometer driven during the first few months of licensure. Australian researchers have reported similar results.

Taken together, these studies add scientific evidence to the common-sense case for graduated licensing (see Status Report, Dec. 4, 1999; on the web at http://www.highwaysafety.org). Such programs restrict driving during the first year of licensure, creating an interim period that leads up to full unrestricted licensure. The idea is that beginning drivers will improve with on-road experience, which initially should be acquired in low-risk circumstances--that is, during daylight hours and without teenage passengers in the car.

Details of Canadian and U.S. Studies:

The Nova Scotia study focuses on a sample of drivers who obtained their learner's permits during 1990-93, before a provincial graduated licensing program was introduced. Driver records and police reports were used to compare month-to-month changes in the crash rates of newly licensed drivers.

Their crash rates went from 123 collisions per 10,000 drivers during the first month of licensure to 73 in the seventh month, a decrease of 41 percent. The crash rate continued to decline gradually over the two years of the study until it was 60 percent lower at the end of the study period compared with the first month the teenagers were licensed.

For the U.S. study, a total of 911 teenagers in Connecticut, Delaware, New Jersey, and New York were surveyed. Crash involvements and citations were examined for the first year of licensure and for the first 3,500 miles driven. The crash rate per 100 drivers was 5.9 in the first month after licensure, dropping to 3.4 in the second month and ranging from 1.3 to 3.0 during the subsequent 10 months. Per 10,000 miles driven, the crash rate was 3.2 for the first 250 miles, 1.8 for the second 250 miles, and 1.3 for the third.

The per-mile crash rate also was calculated for each of the first 12 months of licensure, falling from 2.3 crashes per 10,000 miles in the first month to 1.1 per 10,000 in the second month. Declines continued in subsequent months.