«by Johnathon P. Ehsani A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Health Behavior ...»
The supervised driving requirement mandates novice teen drivers to practice during the learner license period. Effortful, deliberate, and guided practice is an important component of expertise acquisition (Ericsson 2006; Keating and HalpernFelsher 2008). This places the onus on the supervisor to structure driving practice in a productively. In its current form the supervised driving requirement of GDL does not require those who are supervising novice drives to structure the practice in a particular way, nor does it provide any guidelines for driving supervisors, and may explain why we found no significant reduction in fatal crashes following the introduction of supervised driving requirements. Further research is necessary to understand how teens learn to drive and how safe driving can be most effectively taught (Keating and Halpern-Felsher 2008).
As novice drivers progress through each stage of GDL, they experience increasing complexity in the driving environment. Beginning with an extended period of supervised driving, followed by independent driving with restrictions limiting their exposure to the highest risk driving conditions (driving at night or driving with passengers), and finally to driving independently with no restrictions, this progression is consistent with research that suggests expertise is best acquired when there is a gradual progression from simple to complex conditions (Gagne and Paradise 1961). Within this context, specific driving skills are embedded into routines, and become increasingly automatic as teens gain practice. Conceptually, passenger and nighttime driving restrictions should be effective in reducing fatal crash rates among 16- and 17-year-old drivers. The absence of significant reductions in 16- and 17-year-old drivers’ fatal crash rates following the introduction of nighttime driving restrictions merits further investigation. As discussed in Chapter 3, the sample used for the evaluation of the nighttime driving restriction was limited to two small states where driving was restricted between the hours of 12 midnight and 6 a.m. The total number of fatal crashes occurring during that time was small, and may have been insufficient to yield statistically significant changes.
Learning to drive safely not only requires the acquisition of specific knowledge and skills needed for driving, but also mature functioning of a broader set of selfregulatory capacities that promote reliable self-control over behavior (Dahl 2008).
Emerging evidence indicates that changes in the prefrontal cortex (PFC) of the brain, correspond to changes in decision making, and self-regulatory capacities during adolescence (Keating 2004). Research suggests that on average, adolescents are not cognitively mature enough to fully execute safe driving skills, with particular risks arising from regulatory challenges that occur in complex and distracting contexts (Keating and Halpern-Felsher 2008). Limiting the number of passengers in the vehicle, which represent a source of potential distraction and social influence, the passenger restriction eliminates an element of driving complexity that teen drivers may be ill equipped to handle. This is reflected in our finding that the introduction of a strict passenger restriction was followed by a reduction in 16- and 17-year-old drivers’ fatal passenger crashes.
The results of Chapters 2 and 3 indicate that when individual GDL components were examined independently, reductions in 16- and 17-year-old drivers’ fatal crashes were modest and in most instances not significant. This is in stark contrast with overall evaluations of GDL that have shown significant reductions in teen drivers’ fatal crashes of 20 percent or more (Shope and Molnar 2003; Shope 2007). These findings suggest there may be a non-linear relationship between the number of components in a GDL policy and the degree of reduction in 16- and 17-year-old drivers’ fatal crashes, and that GDL exerts its effect operating as a single system rather than as an aggregation of individually effective components. Previous evaluations had assumed the total effect of a GDL program was the cumulative or additive contribution of its individual components (Chen, Baker et al. 2006; Morrisey, Grabowski et al. 2006; Trempel 2009; KaracaMandic and Ridgeway 2010; McCartt and Teoh 2011). In contrast, our findings suggest that while some components show a safety effect when examined in isolation these effects to not sum to equal the overall observed effect of GDL. This supports the conclusion that GDL operates most effectively as a single integrated system.
One mechanism that might explain the differences in crash reductions of comprehensive GDL systems relative to single components is the role of parental involvement in monitoring and enforcing GDL. While differences in the risk perceptions among parents or individual preferences towards components may limit the effectiveness of any single components, a comprehensive GDL program provides parents with more tools at their disposal for teen driver monitoring and management (Simons-Morton and Ouimet 2006).
An alternative explanation for the modest impact of individual GDL components relative to a comprehensive GDL system is the role and influence of driving norms and culture. Risk management literature suggests that the modifications to structures and policies are powerful instruments for cultural and behavioral change (Reason 1998).
Through modifying the driving environment and regulating when, and with whom newlylicensed teens drive, GDL may be effectively changing the driving culture among teens (Moeckli and Lee 2007). Any shift in newly-licensed teens’ driving culture is likely to be more pronounced following the introduction of a comprehensive GDL system relative to when one of two GDL components are in effect. Further research is necessary to understand the impact of driving culture as a potential mechanism for explaining the non-linear relationship between the number of components in a GDL policy and the degree of reduction in 16- and 17-year-old drivers’ fatal crashes.
It should be noted that the study sample for Chapters 2 and 3 was limited to fatal crashes. Some have argued that fatal crashes represent a small subset of all crashes, and the etiology of fatal crashes differs from that of less serious crashes, where high-risk behaviors are more common among drivers involved in fatal crashes (Lam 2003). Future research examining the effect GDL components on teen drivers’ crash rates should be extended to include all crash types. While only a limited number of states make crash data available to researchers (National Highway Traffic Safety Administration 2011), these data would allow an examination of the differential effects of GDL on crash severity.
Furthermore, the study sample was constrained to those instances where a single GDL component implemented independently was identified during the period 1990 to 2009. In several instances, the component of interest was the first element of a GDL system to be implemented, allowing us to quantify an independent effect. However, several states had existing GDL components in place when the new component was implemented, introducing the possibility of order effects. For example, in every instance where a passenger restriction was introduced, it was nested within an established GDL system. Similarly, we identified no cases where a nighttime driving restriction was introduced as a standalone restriction. We did not control for order effects, and this may be a potential source of bias in our findings.
Chapter 4 addressed the second aim of the dissertation: to determine the effect of GDL on 18-year-old drivers’ crash rates and shed light on a possible mechanism responsible for any increase in crash rates. Using three levels of crash severity as outcomes, we exploited a natural experiment resulting from GDL policy implementation to compare the states where GDL provisions apply exclusively to 15- to 17-year-old drivers (Florida, Michigan) to a state where GDL applies to novice drivers of all ages (Maryland). Prior to this research, evidence of the effect of GDL on 18-year-old drivers’ crashes was mixed, and little was understood about the mechanisms responsible for a change in crash rates in this population (Males 2007; Trempel 2009; McCartt, Teoh et al.
2010; Masten, Foss et al. 2011).
We found that the introduction of GDL was followed by a significant increase in possible-injury/property-damage-only crashes among 18-year-old drivers in Michigan (where GDL applies to 16- and 17-year-old drivers) and by a significant decrease in possible-injury/property-damage-only crashes among 18-year-old drivers’ rates in Maryland (where GDL applies to novice drivers of all ages). This result established support for the presence of an offset effect, where a proportion of teens are not licensed until age 18, and for these individuals the risk associated with inexperience is delayed and manifests as elevated possible-injury/property-damage-only crashes among 18-year-old drivers.
Beginning drivers of any age go through a learning process in which more errors are made in the early stages than later. In their current form, most GDL systems in the U.S. do not apply beyond the age of 17. Therefore, novice drivers aged 18 years and above are not afforded an opportunity to develop expertise in the safer driving conditions that GDL systems create, and are effectively unprotected from their own inexperience.
More research is required to understand the characteristics of teens who choose to be licensed at age 18. Currently, very little is known about the differential effects of GDL by racial or socio-economic group, but some evidence exists that GDL exerts a greater safety effect on certain populations, relative to others. A recent study reported reductions in 15- to 17-year-old drivers’ fatal crashes following the introduction of GDL were largest for young White drivers, followed by African-Americans and Asians, however there was no significant reductions in Hispanic drivers’ crashes (Romano, Fell et al. 2011).
Whether this is due to differing average ages of licensure between racial and ethnic groups is unknown.
Requiring some form of graduated licensing program to be completed by all novice drivers is recommended. Several international jurisdictions necessitate that all novice drivers complete a form of GDL prior to receiving a regular license (Organisation for Economic Co-operation and Development and European Conference of Ministers of Transport 2006; Senserrick 2007); however, further research is necessary to identify the elements that would be required in the U.S. context.
The body of literature examining the effects of GDL has focused on 16- and 17year-old drivers. This population is distinct from drivers aged 18 years and older on multiple dimensions. For example, GDL requirements that may be relatively straightforward for 16- and 17-year-old drivers to fulfill, such as completing a minimum number of supervised driving hours with a parent, may be more difficult for those aged 18 years or older, who are at a different life stage and may no longer reside with their parents. Maryland’s GDL program, which requires an extended learner permit period for all novice drivers, represents one approach that could be considered. This research demonstrated significant reductions in 18-year-old drivers’ possible injury/PDO crashes in that state following the introduction of GDL; however, it is unknown whether additional requirements or restrictions would result in the same or a larger crash reduction.
Longitudinal state crash and licensure records that allow the linkage of an individual’s crashes with age at licensure and GDL stage would permit the identification of crash types for which drivers who have not completed GDL (i.e., were licensed at age 18 or older) are at greater risk. Comparing 18-year-old drivers’ crash involvement between teens who are licensed through GDL and teens in the same state not licensed until age 18 following implementation of GDL, will identify the unique crash risks for novice drivers licensed for the first time at age 18. Based on these data, GDL programs for novice drivers aged 18 years and older could be designed, implemented and evaluated.
As a public health intervention, GDL forgoes efforts to provide specific, detailed training to individuals directly, but exerts its influence through modifying the environment by restructuring of the driver licensing system. The principle of environmental modification is not new to the field of public health or injury prevention, and can be found in the design of motor vehicles and roadways (Hemenway 2009) and beyond the field of road safety, to the prevention of unintentional poisoning through package and label design (Schneider 1977). Unlike changes in engineering and design, GDL modifies the driving environment through redesigning the licensing system to make the initial learning period safer for inexperienced drivers by specifying the conditions wherein that experience can be gained.
Extending the application of this principle to mitigating the elevated crash risk of older drivers has recently been proposed where the licensing environment for drivers above a certain age (e.g., 70 years) could be modified to accommodate declining cognitive and physical capabilities (Redelmeier and Stanbrook 2012). With this approach, a full driver’s license defaults automatically to a restricted license that allows seniors to continue to drive only under relatively favorable driving conditions. Under this system, older drivers’ continued licensure is contingent on demonstrating competency to drive. Moreover, an opt-out rather than opt-in approach to licensing is congruent with existing safety policies for aviators, vaccination programs for children and commercial driver testing.