Discussion
Background: TADRA’s enactment came after a 6-year period of intense legislative lobbying and community activism. The political pressure to do something about teen crash fatalities reached a crescendo in the fall of 1996 following a series of highly publicized, multiple-fatality teen car crashes in the Metropolitan Atlanta area. The first version of TADRA was signed into law by then Governor Zell Miller on March 14, 1997.21-22 The legislation was further strengthened in 2000 with extension of driving curfew hours, elimination of curfew exemptions, and additional limitations on the number of passengers that can ride in a vehicle driven by provisional license holder. These features were signed into law by then Governor Roy Barnes on April 11, 2001.25-26
The first examination of the impact of TADRA was conducted by Angelyn Rios and reported in the Georgia Highway Safety 1999 Fact Book.24 She noted an immediate and highly positive effect of the law, particularly among 16- and 17-year-old drivers (a decrease of 35.1% in fatal crashes involving 16- and 17-year-old drivers). These findings suggest that TADRA had an immediately favorable impact, but the evaluation interval was too short to reach a definitive conclusion about the law.
In 2001, Rios reexamined TADRA 18 months following enactment. At that point, the rate of fatal crashes involving Georgia drivers 16 and 17 years old was still 30 percent lower than it had been immediately prior to enactment.27 Although Rios’ findings were highly encouraging, Georgia’s experience was not compared to that of neighboring States, so it was possible that much of the law’s perceived benefit might be due to historical effects, such as stricter traffic enforcement or rising rates of seat belt use.
The Current Study: Our evaluation confirms Rios’ earlier findings and indicates that the effects of TADRA were sustained over time. During the 5-½-year follow-up period, the rate of fatal crashes involving 16-year-old Georgia drivers was 36.8 percent lower than during the final 5 ½ years before enactment. This reduction was substantially greater than that observed in the comparison States, and twice the decline noted among 16-year-olds nationwide during this time interval. This suggests that enactment of TADRA produced significant benefits above and beyond what might be attributed to historical effects.
We also found a statistically significant but more modest effect on fatal crashes involving 17-year-old drivers. No appreciable decrease in fatal crashes involving 18-year-olds was noted, but no increase was noted, either. This largely rules out displacement of crashes from younger to older age groups. If TADRA had simply encouraged young drivers to delay the onset of driving, or to reduce their frequency of driving until they were older, any net benefit from a decrease in fatal crashes involving 16- and perhaps 17-year-old drivers would be offset by higher rates of fatal crashes among 18-and-older drivers. This was not the case.
TADRA’s impact on fatal crash rates involving 16-year-old drivers was greater than that noted in a number of States where GDL legislation has been evaluated (Appendix 1). We believe TADRA’s impact is greater because it includes a number of supplemental provisions aimed at deterring specific high-risk driving behaviors, most notably excessive speeding. TADRA also differs from many other State GDL systems because it imposes a highly meaningful and inescapable sanction -- automatic license revocation – on those who violate its provisions (Appendix 2).
Analyzing 11 years of data allowed us to compare driver fatal crash rates by age group for 5 ½ years pre- and post-enactment. Most evaluations of GDL laws have had been limited by relatively short follow up intervals.11 Rather than limit our analysis to the youngest age groups of drivers, we examined changing rates of fatal crashes across the age spectrum. This allowed us to determine if TADRA simply displaced fatal crashes to older age groups. Comparing our data with that of three neighboring States allowed us to exclude potentially confounding historical effects. Adjusting for changes in population was necessary because Georgia, like other southern States, experienced significant growth during the study interval. Use of relatively lengthy pre- and post-enactment periods increased the power of our study to explore TADRA’s effects on specific types of hazardous driving behavior, such as excessive speeding, drunk driving, and late-night driving. It also enabled us to confirm that the law’s benefits were long-lasting.
Every study is subject to limitations. The data used in our analysis is derived from fatal crash reports that reflect the perspectives of the law enforcement officers that complete them as well as the fiscal limitations of the agencies that support these collections. Enforcement of laws against various high-risk driving behaviors, such as nonuse of seat belts, speeding, and drunk driving, vary from State to State and probably change over time.
We decided to calculate age-specific rates of driver fatal crashes using each State’s population, rather than the number of licensed drivers in each age group because the only national database that contains State and age-specific counts of licensed drivers (FHWA, Highway Statistics, DL-22) specifically excludes drivers holding a provisional or restricted license.28 In 1989, a FHWA committee decided that the enumeration of licensed drivers for this file should not include Class P permit holders (i.e., instructional permits) as well as drivers with limited-use or restricted licenses.
When this FHWA rule was implemented, it significantly reduced State counts of licensed drivers in the youngest age groups. Moreover, since different States interpreted the rule differently, it added a degree of variability to what had previously been a consistent process. Both effects make State-by-State comparisons based on current counts of licensed drivers highly problematic. Regardless of how officials in a particular State interpret the rule, it is clear that FHWA statistics significantly undercount drivers in youngest age groups, particularly drivers in GDL States where the issuance of Class P and various types of restricted licenses is quite common.
Because drivers with restricted licenses are frequently on the road, particularly during peak hours of the day and evening, using FHWA data to calculate fatal crash rates could seriously undercount the number of young drivers at risk. The practical impact of this error would be to artificially inflate fatal crash rates in this age group, thereby masking the beneficial effects of GDL legislation and other measures aimed at young drivers.
To avoid this error, we used standardized, age-specific population data as our denominator to calculate driver fatal crash rates. This approach offers the advantage of consistency, but it has its own limitations. It is possible that use of age-specific population data may overestimate the number of drivers at risk, since not all individuals choose to obtain driver licenses the moment they are eligible to do so. However, there is little evidence that large numbers of teens are deferring their decision to acquire driver licenses, beyond speculation based on misinterpretation of falling rates of licensure using FHWA statistics.
Confounding effects: Certain historical effects may explain some but clearly not all of the benefits we observed. Although seat belt use was not directly encouraged by TADRA, it increased in Georgia as it did elsewhere throughout the 11-year study interval.29 In Georgia, observed rates of seat belt usage in all age groups increased from 50.6 percent in 1992 to 77.0 percent in 2002.30-31 As recorded on the fatal crash reports filed by law enforcement officers, seat belt usage by occupants involved in fatal crashes also increased, from 47 percent in 1992 to 79.6 in 2002. To control for the historical effect of rising seat belt usage as well as other factors such as the proliferation of more crash worthy vehicles, we compared Georgia’s experience to that of three neighboring States that did not enact similarly comprehensive legislation during the study time frame. Nationwide, the fatal crash rate involving 16-year-old drivers decreased 14.8 percent during the 5 1/2 year period immediately following the Georgia General Assembly’s enactment of TADRA. This is less than half the decrease that was observed among Georgia’s 16-year-old drivers during the same time period. If the decrease was due to seat belt use, we would have expected it to be gradual and mirrored in all States. Instead, we noted an immediate, “step-wise” reduction in fatal crash rates involving 16-year-old drivers – an effect that was not mirrored in the comparison States (Appendix Four).
Why is TADRA so effective among 16-year-olds? At age 16, the vast majority of teenagers are inexperienced drivers who live at home and are therefore subject to their parents’ or guardians’ supervision.32-33 They are also dependent on their parents or guardians for economic support. By age 17, many teens have acquired enough driving experience to become overconfident; they may also be less fearful of the disapproval of their parents or guardians. By age 18, many teens have moved out of the house to go to college or start their first job, and are therefore living independently of their parents or guardians. They may also find it easier to acquire alcohol. In light of these developmental and social milestones, it is not surprising that TADRA had its greatest effect on 16-year-old drivers, and a diminishing but nonetheless valuable impact on 17- year-old drivers.
If TADRA merely discouraged 16- and 17-year-olds from driving, or delayed the onset of high-risk driving without fundamentally altering driver behavior, any benefits accrued among 16- and 17-year-olds might be offset by a compensatory increase in fatal crashes involving drivers 18 or older. This did not occur.
The provision in TADRA that calls for automatic revocation of the license of a teen driver traveling more that 24 miles per hour over the posted speed limit may be particularly effective. Before enactment of TADRA, unsafe or illegal speed-related fatal crashes accounted for more than one-third of all fatal crashes involving 16-year-old drivers. After TADRA went into effect, annual rates of speed-related fatal crashes involving 16-year-old Georgia drivers were cut nearly in half. In fact, the reduction of speed-related fatal crashes accounted for more than half of the law’s effect among 16-year-old drivers.
A Potential Cohort Effect: We found evidence that TADRA may impart lasting benefits to drivers who grow up under the law. The first cohort of Georgia drivers who reached age 21 in the era of TADRA were involved in fatal crashes at a rate fully 38 percent lower than that of Georgia drivers who turned 21 in 1997, the year that TADRA went into effect. Although some of this benefit is undoubtedly due to improvements in vehicle design and rising rates of seat belt use, the change is much greater than that noted on a national basis during this time frame. This suggests that Georgians who are reaching age 21 in the era of TADRA may be driving more responsibly than their predecessors. This is a fertile area for further research.
Comparisons to Neighboring Southeastern States: TADRA’s effects were not mirrored in three adjoining Southern States* that did not enact similar legislation during the study interval.34
- The three comparison States added partial graduated licensing provisions to their existing motor vehicle statutes during the TADRA study period. Legislation was passed in Alabama in 2003, South Carolina in 1998, and Tennessee in 2004. See 54 Alabama Law Review 1473 (Summer, 2003), 56 South Carolina Code Annotated 1 § 175,176,180 and 55 Tennessee Code Annotated 50 §301-304, 311,312, 322 as well as, 49 Tennessee Code Annotated 6 §3017.
The differences between Georgia and its neighboring States of Tennessee and South Carolina are particularly stark. In the third comparison State, Alabama, fatal crashes involving 16-year-old drivers declined to a greater degree than the national average, but the decrease was not as great as that noted in Georgia. Based on these comparisons, it is highly unlikely that the decreases we observed among 16- and 17-year-old Georgia drivers were exclusively due to historical factors. This suggests that other States may achieve impressive reductions in teen fatal crash rates if they enact legislation similar to Georgia’s TADRA law.
In Alabama, the rate of fatal crashes involving 16-year-old drivers decreased 22.5 percent during the post-enactment period. Alabama did not enact GDL legislation prior to 2002, and the State’s GDL law did not go into effect until September 1 of that year (See Appendix 2). No other major legislative changes were adopted by Alabama during the 11-year interval of our study. To determine if other public education or enforcement interventions were implemented in Alabama during this time, we contacted colleagues in Alabama to ask them why Alabama’s experience diverged from that of Tennessee and South Carolina. They could not identify a reason. While Alabama’s decline during the study interval is intriguing, it was only 7.5 percent greater than the overall decline in the United States during the same time period. Furthermore, Alabama’s decrease was 14.3 percent less than that observed in Georgia’s during the same time frame.
Final Observations: Although our findings are encouraging, it is important to keep TADRA in perspective. Before TADRA was enacted, the rate of fatal crashes involving 16-year-old drivers in Georgia was quite high. In the first 5 ½ years following enactment, fatal crashes involving 16-year-old Georgia drivers declined dramatically, but the post enactment rate was still 28.6 percent higher than the U.S. average among 16-year-olds (36.1 per 100,000 in Georgia versus 28 per 100,000 nationwide). Happily, this “mortality gap” is less than half what it was before TADRA was enacted (57 per 100,000 in Georgia versus 32.8 per 100,000 nationwide). Georgia has made progress, but the State still has a long way to go.
Nevertheless, TADRA appears to have had a favorable impact on highway safety in Georgia. One way to quantify this benefit is to use sensitivity analysis to project the rate of fatal crashes that might otherwise have occurred if the Georgia’s General Assembly had failed to enact TADRA in 1997, or the Governor had failed to sign it into law. If Georgia had experienced the more modest decline in driver fatal crashes that was noted in Alabama during the 5-½-year follow-up interval (the “best case scenario”), 221 additional drivers age 16 to 20 would have been involved in a fatal crash. However, if driver fatal crash rates in Georgia had risen as much as they did in South Carolina (the “worst case scenario”), 434 additional drivers age 16 to 20 would have been involved in a fatal crash. Since a “driver fatal crash” can claim more than one life, it is likely that TADRA saved a minimum of 220 to 430 lives during its first 5 ½ years of existence. It is also possible, if not likely, that TADRA prevented a large number of serious but nonfatal injuries.
Why did TADRA have more impact than many States’ GDL laws? We cannot answer this question with certainty, but we believe it is probably because TADRA links graduated driver licensing with supplemental provisions designed to deter high-risk driving behaviors by teens. Chief among these are measures designed to deter driving at excessive rates of speed, drag racing, and evading a police officer, and a measure specifically designed to deter alcohol-impaired driving. Provision for an immediate, believable, and highly meaningful sanction, automatic license revocation, probably strengthens the deterrent effect of TADRA. Teen drivers in Georgia quickly learned that their parents could not help them evade the legal consequences of violating the law. Anecdotally, the loud complaints of the first young drivers who lost their license constituted the best publicity TADRA received. It is important to acknowledge that because the many provisions of TADRA were enacted as a package, it is impossible to dissect which particular aspect had the greatest effect.
Nevertheless, it appears that enactment of TADRA resulted in an immediate and substantial decline in driver fatal crash rates in Georgia. The impact was greatest among 16-year-old drivers, but notable among 17-year-old drivers as well. Fatal crashes were not displaced to 18-year-olds and older age groups. Interestingly, there is preliminary evidence that drivers reaching age 21 under TADRA may be driving more safely than their predecessors did.
State governments interested in duplicating Georgia’s success should consider adopting a similarly comprehensive package of legislation. Successful replications in other States would provide compelling evidence that TADRA’s multi-faceted approach is effective. Georgia’s experience suggests that it is possible to facilitate a safe and orderly transition from adolescence to adult driving. Graduated licensing is a key component of that process, but States should consider strengthening it with additional provisions aimed at deterring high-speed driving, DUI, and other dangerous behaviors. Such laws, backed up by meaningful enforcement, and certain sanctions, may help more adolescent drivers and passengers reach adulthood.