Michigan's graduated driver licensing program: Evaluation of the first four years
Introduction
Motor-vehicle crashes are the leading cause of death among teenagers in the United States (Insurance Institute for Highway Safety [IIHS], 2003). Crash rates among young drivers age 16-19, per mile driven, are the highest of any group and the crash rate among 16-17-year-old drivers is almost three times as high as that among 18-19-year-old drivers (IIHS, 2002a). Driving at night is especially risky for young drivers; about 40% of 16-17-year-old driver deaths occur at night, although only 15% of miles driven by these young drivers are at night (Williams & Preusser, 1997). Transporting other teenagers also adds to the crash risk of young drivers, with risk increasing exponentially with one, two, and three or more passengers (Williams, 2001). Yet, until recently, obtaining full, unrestricted driving privileges in the United States had been allowed at young ages and after only minimal classroom education and behind-the-wheel training.
Graduated driver licensing (GDL) allows young novice drivers to gain experience and maturity under conditions of low risk before progressing to more risky driving situations. A comprehensive approach to GDL for young novice drivers was first proposed in the 1970s (e.g., Croke & Wilson, 1977, Waller, 1977, Waller, 2003). Recommended elements include: three stages of licensure (learner's, intermediate, and full-privilege) to phase drivers into on-the-road driving experiences; adequate supervised practice in the learner's stage; restrictions on such risky activities as driving at night and with teenage passengers in the intermediate independent driving stage; and demonstrated competence before passage through each stage, using an integrated two-phase driver education program, written and road tests, and delays in progress if traffic violations or at-fault crashes occur National Highway Traffic Safety Administration [NHTSA], 1998, Williams & Mayhew, 2003.
Although GDL was implemented in New Zealand in 1987, and soon after in Australia and Canada, comprehensive U.S. GDL initiatives did not generally begin until the late 1990s. Nearly all states in the United States and the District of Columbia have now adopted one or more elements of GDL (Williams & Mayhew, 2003) and a good majority of states actually have a three-level GDL system for young novice drivers (IIHS, 2002b). However, the specific provisions of these GDL systems vary considerably. Several relevant articles are included in two special issues on the topics of young drivers and GDL (Injury Prevention, Volume 8, Supplement II, 2002 and The Journal of Safety Research, Volume 34, 2003).
Michigan's GDL program was implemented on April 1, 1997 and requires three levels of licensure for young novice drivers (see Table 1): Level 1 (the learner's stage) including extensive supervised practice; Level 2 (the intermediate stage) including a night driving restriction; and Level 3 (full licensure) still under the state's existing three-year probationary period for all new licensed drivers. Michigan's GDL program does not yet include a passenger restriction.
Early effects of Michigan's GDL program on 16-year-old drivers were assessed in two published studies. In the first (Shope, Molnar, Elliott, & Waller, 2001), population-based crash risks from pre-GDL data (1996) were compared with risks from post-GDL data (1998 and 1999). The overall crude risk of being involved in a crash in 1999 was 28% lower than in 1996. Statistically significant reductions in risk were also found for non-fatal injury crashes (33%) and combined fatal/non-fatal-injury crashes (33%); 5 a.m. - 8:59 p.m. (27%), 9 - 11:59 p.m. (26%), and midnight - 4:59 a.m. crashes (55%); and single-vehicle (34%) and multi-vehicle crashes (26%). Controlling for more general population-wide trends (specifically, among drivers age 25 years and older) tempered the reductions by 3-9 percentage points; the adjusted reduction in overall crash risk between 1996 and 1999 was 25%.
In the second Michigan evaluation (Elliott & Shope, 2003), a linear changepoint model was used to examine whether the pre-post comparisons in the first evaluation may have been confounded by underlying trends independent of GDL (that the absence of a control group made it difficult to sort out). The model took into account the uncertainty regarding when GDL effects, if any, began and ended, and whether a “rebound” in crash-involvement rates may have occurred. Findings generally supported the pre-post comparisons, with GDL appearing more likely than not to have had a substantial effect, dropping the overall crash-involvement rate among Michigan 16-year-olds by approximately 25% during the 12-18 months following its implementation.
Evaluations of GDL programs in other jurisdictions have provided additional support for GDL's effectiveness in reducing young driver risk, although comparison of evaluation results is difficult because of differences in pre-GDL licensing, differences in GDL programs, and differences in evaluation methodologies. A recent examination by Shope and Molnar (2003) of the six U.S. states that had a comprehensive GDL program in place long enough to evaluate, and for which at least preliminary evaluation results were available, found that all six states identified some type of crash reduction among young novice drivers following the implementation of GDL (for individual state reports, see Automobile Club of Southern California, 2000, Automobile Club of Southern California, 2001, Commonwealth of Pennsylvania Department of Transportation, 2002, Elliott & Shope, 2003, Foss et al., 2001, Nissley, 2001, Office of the Governor's Highway Safety Representative, 2001, Shope et al., 2001, Ulmer et al., 2000).
In jurisdictions outside the United States, where GDL laws have been in place longer, evaluations have reported declines among young novice drivers ranging from 7% to 37% (e.g., Langley et al., 1996, Mayhew et al., 1999). Results from a recent evaluation update of GDL in New Zealand suggest that the impact of GDL has not diminished over time, with the combined fatal/serious injury crash rate among young drivers remaining substantially below the pre-GDL level (Begg & Stephenson, 2003).
The majority of evaluations have examined GDL's impact overall, because it is difficult to separate out the individual effects of each component (with the exception of the night restriction, for which night crash pre-post comparisons have been reported). Thus, most evaluation results have focused on the combined effects of all GDL components, including requirements for practice, time in each stage of licensure, driver education, clean driving record, and night driving restrictions on the crash risk of young novice drivers. Although it should be possible to separate out passenger effects, most of the passenger restrictions in the U.S. systems are too new to have had their impact evaluated. New Zealand's passenger restriction, which has been in effect since 1987, has been shown to reduce crashes involving passengers among young novice drivers (Begg, Stephenson, Alsop, & Langley, 2001).
The purpose of the study reported here is to update Michigan's early evaluation results Elliott & Shope, 2003, Shope et al., 2001, using two additional years of crash data and updated population estimates not available at the time of the early evaluations. Extending the analyses of 16-year-old driver crash data through 2001 makes it possible to determine whether the initial reductions in crash risk among 16-year-old drivers following the introduction of GDL have been maintained. Additionally, this report provides information on crash risk by sex, as well as during the evening hours preceding the night restriction, and by passenger involvement. Such information could be useful in determining whether driving restrictions during the intermediate stage need to be strengthened in Michigan's GDL program to further enhance the safety of young novice drivers.
Section snippets
Method
Data on motor-vehicle crashes in Michigan in 1994-1996 (three full years before GDL) and 1998-2001 (four full years after GDL and the most recent data available to date) were obtained from the Michigan State Police. The police-reported crash files in Michigan for those years contained records for all reportable crashes (those involving a personal injury or property damage of $400 or more), including demographic data on drivers (age and sex). Injury severity data are coded in Michigan crash
Results
Table 2 presents numbers of crash involvements of 16-year-old drivers, population-based crash-involvement rates, and relative risks for 1996 versus 1998, 1999, 2000, and 2001, for all crashes and for selected crash subtypes. The reductions in crash risks among 16-year-olds previously found in 1998 and 1999 were generally maintained in 2000 and 2001. Examining just the comparisons of 2001 with 1996, significant reductions in risk of crash involvement were found for all crashes (RR, 0.71; CI,
Discussion
The significant post-GDL reductions in crash risk among Michigan's 16-year-olds first seen in 1998 and 1999 were maintained in 2000 and 2001. Reductions in crash risk among 16-year-olds between 1996 and 2001 were 29% for all crashes, 44% for fatal crashes, 38% each for nonfatal injury crashes and fatal plus nonfatal injury crashes, 32% for day crashes, 31% for evening crashes, 59% for night crashes, 32% for single-vehicle crashes, and 28% for multi-vehicle crashes. Even after adjusting for
Acknowledgements
This work was partially supported by the National Highway Traffic Safety Administration. We gratefully acknowledge the staff of the Library of Michigan, the University of Michigan Transportation Research Institute Transportation Data Center, and the Michigan Department of State for their assistance in obtaining the data; and Mary E. Chico for her assistance in manuscript preparation.
Jean Shope is Research Professor and Director of Social and Behavioral Analysis at the University of Michigan Transportation Research Institute (UMTRI), where she has been working since 1991. She is also on the faculty of the Department of Health Behavior and Health Education in the University of Michigan School of Public Health. Dr. Shope's PhD from Wayne State University is in the Theoretical and Behavioral Foundations of Education. She also completed a postdoctoral fellowship in Health
References (35)
- et al.
Graduated driver licensing: the New Zealand experience
Journal of Safety Research
(2003) - et al.
Enhancing the effectiveness of graduated driver licensing legislation
Journal of Safety Research
(2003) - et al.
What we know, what we don't know, and what we need to know about graduated licensing
Journal of Safety Research
(2003) - et al.
An evaluation of the New Zealand graduated driver licensing system
Accident Analysis and Prevention
(1996) - et al.
Provisional license
Journal of Safety Research
(2003) - et al.
Graduated driver licensing in the United States: Evaluation results from the early programs
Journal of Safety Research
(2003) - et al.
Effect of Florida's graduated licensing program on the crashes of teenage drivers
Accident Analysis and Prevention
(2000) The genesis of GDL
Journal of Safety Research
(2003)Teenage drivers: patterns of risk
Journal of Safety Research
(2003)- et al.
Deaths of teenagers as motor vehicle passengers
Journal of Safety Research
(1995)
Impact of graduated driver licensing restrictions on crashes involving young drivers in New Zealand
Injury Prevention
Carrying passengers as a risk factor for crashes fatal to 16- and 17-year-old drivers
JAMA: the Journal of the American Medical Association
Model for Provisional (Graduated) Licensing of Young Novice Drivers. Final report
Use of a Bayesian changepoint model to estimate effects of a graduated driver's licensing program
Journal of Data Science
Cited by (0)
Jean Shope is Research Professor and Director of Social and Behavioral Analysis at the University of Michigan Transportation Research Institute (UMTRI), where she has been working since 1991. She is also on the faculty of the Department of Health Behavior and Health Education in the University of Michigan School of Public Health. Dr. Shope's PhD from Wayne State University is in the Theoretical and Behavioral Foundations of Education. She also completed a postdoctoral fellowship in Health Behavior and Health Education in the University of Michigan School of Public Health. Dr. Shope's UMTRI research involves the driving behavior of adolescents, young adults, and older drivers. Special interests include at-risk drinking and drink/driving, as well as graduated driver licensing.
Lisa Molnar is a Senior Research Associate with the Social and Behavioral Analysis Division of UMTRI where she has worked since 1986. She holds a BA in Sociology from Michigan State University and an MHSA in Public Health Policy and Administration from the University of Michigan School of Public Health. Her primary areas of interest are traffic safety and driver behavior. Ms. Molnar has worked on a variety of projects focusing on: adolescent driving behavior; older driver safety and mobility; development and evaluation of traffic safety laws, policies, and programs; prevention of alcohol-impaired driving; and use and misuse of safety belts and child safety seats.