III. What Happened?

In the United States in 1982, there were 10,270 drivers under the age of 21 involved in fatal crashes. Forty-three percent (4,393) of these drivers were deemed to have been drinking prior to their crashes. In 1998, the number of under 21 year old drivers in fatal crashes was 8,128 with 21 percent (1,714) of these determined to have been drinking. Comparing 1998 with 1982, the number of youthful drivers involved in fatal crashes declined by 21 percent and the number who had been drinking declined by 61 percent.

These basic trends for young drivers in fatal crashes have been evident for some time. In particular, annual NHTSA reports (Wright, 2000) document the substantial drop in youth alcohol-related traffic fatalities and rates and compare these to alcohol-related traffic fatalities and rates for older persons. This chapter begins by presenting these national trends and disaggregating them by region and state. Next, national and regional youth drinking data and trends are analyzed and drinking trends are compared to drinking and driving trends. Finally, evidence is presented on youth drinking and driving behavior and how it has changed since 1982.

A. Young Drinking Drivers in Fatal Crashes - 
National Trends

Young Drivers in Fatal Crashes

Figure 1 shows the trend in the numbers of young drinking drivers involved in fatal crashes over the 1982-1998 period. The figure shows that there was a general downward trend until about 1993 with little change since then.

Figure 1. 
Young Drinking Drivers in Fatal Crashes
Graph of Figure 1. Young Drinking Drivers in Fatal Crashes

The trend for not-drinking young drivers in fatal crashes, shown in Figure 2, is quite different: increasing from 1982 to 1988, decreasing until 1992, then increasing again.

Figure 2. 
Young Drinking and Not-Drinking Drivers in Fatal Crashes
Graph of Figure 2. Young Drinking and Not-Drinking Drivers in Fatal Crashes

Figure 3 summarizes the two trends as the percentage of young drivers in fatal crashes with a positive BAC. This has decreased steadily from 43 percent in 1982 to 21 percent in 1998.

Figure 3. 
Percent of Young Drivers in Fatal Crashes with Positive BAC

Graph of Figure 3. Percent of Young Drivers in Fatal Crashes with Positive BAC

It is very clear from these figures that the number of young drinking drivers in fatal crashes has dropped spectacularly since 1982. It's also clear that this is not just due to a general reduction in young driver crash involvement, since the number of non-drinking drivers actually increased. This evidence suggests that overall drinking and driving by youth has decreased substantially since 1982.

Comparisons by Age

Figure 4 shows the number of young drinking drivers in fatal crashes for each individual year of age. The figure shows that in virtually every calendar year the number of drinking driver involvements increases with each year of age: there are more drinking drivers age 16 than age under 16, more age 17 than 16, etc. The figure also shows that drinking driver involvement decreases followed the same pattern for each age. Overall, drinking drivers in fatal crashes decreased between 59 percent and 65 percent for each age, 16 through 20 from 1982 to 1998.

Figure 4. 
Young Drinking Drivers in Fatal Crashes by Age
Graph of Figure 4.  Young Drinking Drivers in Fatal Crashes by Age

The virtually identical trends by age illustrated in Figure 4 suggest an important conclusion. Drivers under 21 differ substantially by age in where they live, what they do, how much and why they drive. Most youth aged 16 attend high school and live at home with at least one parent. By the age of 20, most youth are considerably more independent, attend college or have a full-time job, and many do not live with their parents for much of the year. Yet the drinking driver decrease pattern was the same for 16- and for 20-year-olds. The causes of this decrease appear to have influenced youth of all ages.

Figure 5 compares the trends in drinking drivers in fatal crashes for three age groups: under 21, 21-24, and 25 and above. Driver involvements in the two younger age groups decreased steadily throughout the period, while the older age group's decrease was concentrated between 1990 and 1994.

Figure 5. 
Drinking Drivers in Fatal Crashes by Age Group
Graph of Figure 5. Drinking Drivers in Fatal Crashes by Age Group

Figure 6 plots the three trends from a base of 1982 = 100 percent. Thus, in 1998 the number of drinking drivers under age 21 was 39 percent of the 1982 level, a decrease of 61 percent. Similarly, in 1998 drinking drivers aged 21-24 were 44 percent of their 1982 level, and drinking drivers age 25 and above were 75 percent. Put another way, drinking drivers in fatal crashes aged 16-20 decreased 61 percent from 1982 to 1998; drivers aged 21-24 decreased 56 percent; and drivers aged 25 and above decreased 25 percent.

Figure 6. 
Drinking Drivers in Fatal Crashes by Age Group
Percentage Change, 1982-1998
Graph of Figure 6. Drinking Drivers in Fatal Crashes by Age Group Percentage Change, 1982-1998

Two conclusions are clear: the number of young drinking drivers in fatal crashes dropped faster than the number of older drinking drivers, but the number of drinking drivers aged 21-24 decreased almost as much.

Involvement Rates by Age

Changes in the US population age distribution clearly affect driver involvements in fatal crashes. In the past 20 years the number of young persons of driving age has decreased while the number of older persons has increased substantially. More precisely, between 1982 and 1998 the US population age 15 to 20 decreased by 4 percent, the population aged 21 to 24 decreased by 20 percent, and the population aged 25 to 54 increased 31 percent.

Figure 7 accounts for these population changes by plotting the number of drinking drivers in fatal crashes per 100,000 population.

Figure 7. 
Drinking Drivers in Fatal Crashes per 100,000 Population
Graph of Figure 7. Drinking Drivers in Fatal Crashes per 100,000 Population

The difference from the absolute numbers of Figure 5 is striking. Drivers aged 21-24 have the highest involvement rates, followed by drivers under 21, then by drivers aged 25 and above. Involvement rates for all three age groups have decreased quite steadily since 1982.

Figure 8 plots the three trends of Figure 7 from a base of 1982 = 100 percent. The involvement rate for drivers age 16-20 has decreased the most - 59 percent; drinking driver involvements for age 21-24 decreased 46 percent; and involvements for age 25-54 decreased 43 percent.

Figure 8. 
Drinking Drivers in Fatal Crashes per 100,000 Population
Percentage Change, 1982-1998
Graph of Figure 8. Drinking Drivers in Fatal Crashes per 100,000 Population Percentage Change, 1982-1998

Table 1 summarizes the reductions illustrated in Figures 6 and 8. For drivers under 21, decreases in the number of drinking driver involvements and the involvement rate per population were very similar, at about 60 percent. For drivers aged 21-24, the 56 percent decrease in the number of driver involvements translates to a 46 percent decrease in involvement rate. Thus, some of their decrease in driver involvements was due to a decrease in the number of drivers on the road, not to a change in driver behavior. In contrast, due to the substantial increase in the population aged 25-54, the 24 percent decrease in their driver involvements grew to a 43 percent decrease in the involvement rate.

Table 1. 
Change in Drinking Drivers in Fatal Crashes
Age Number of Drivers Change, 1982 to 1998 
(Figure 6)
Drivers per Population Change, 1982 to 1998 
(Figure 8)
Under 21 -61% -59%
21-24 -56% -46%
25-54 -24% -43%

The driver involvement rates of Figure 8 and Table 1 provide the best measure of behavior change that can be deduced from FARS data. They show that drivers of all ages reduced their drinking and driving; young drivers reduced their drinking and driving about 37 percent more than older drivers (59 percent compared to 43 percent), with drivers aged 21-24 falling in between. Thus, population changes account for some of the difference between young and older drivers suggested by a simple comparison of the number of drinking driver involvements in fatal crashes.

Conclusions

The national data examined so far suggest several conclusions.