Crash and traffic violation rates before and after licensure for novice California drivers subject to different driver licensing requirements.

Journal of Safety Research 50 (2014) 125–138

Contents lists available at ScienceDirect

Journal of Safety Research journal homepage: www.elsevier.com/locate/jsr

Crash and traffic violation rates before and after licensure for novice California drivers subject to different driver licensing requirements Eric A. Chapman a, Scott V. Masten a,⁎, Kelly K. Browning b a b

California Department of Motor Vehicles, Research and Development Branch, 2570 24th Street, MS H-126, Sacramento, CA 95818-2606, USA Impact Teen Drivers, P.O. Box 1611209, Sacramento, CA 95816-1209, USA

a r t i c l e

i n f o

Article history: Received 5 April 2014 Accepted 28 May 2014 Available online 14 June 2014 Keywords: GDL Graduated driver licensing Novice drivers Crashes Traffic violations

a b s t r a c t Introduction: How do crash and traffic violation rates for novice 16–17-year-old drivers change over the months before and after licensure under a graduated driver licensing (GDL) program relative to those for older novices who are not subject to GDL? Method: Plots and Poisson regression comparing overall rates and subtypes of crashes and traffic violations among California novice drivers ages 16 to 35 years over time before and after unsupervised licensure. Results: Majorities of 16-year-olds (57%) and 17-year-olds (73%) actually hold their learner permits longer than the required 6 months; majorities (67%–81%) of age 18 or older novices hold their learner permits less than 6 months. Crash rates of novice 16- and 17-year-olds—as well as most other age groups—are highest almost immediately after they are licensed to drive unsupervised, after which their rates decline quickly during their first year of licensure and at a slower rate for the second and third years. Novice 16- and 17-year-olds' traffic violation rates reach their zenith long after their total crash rates peak and decline, whereas violation rates for older novices peak during their first year of licensure. Over 70% of 16- and 17-year-old novices are crash-free for the first 3 years of licensure. Conclusions: While novice 16- and 17-year-olds' highest crash rates occur almost immediately after they are licensed, their peak traffic violation rates are delayed until around the time they turn age 18. Both pre-licensure crash rates and post-licensure crash peaks were more pronounced for some older age groups of novices than was the case for 16–17-year-olds. Practical Applications: Extending learner permit holding periods for 16–17-year-old novices appears consistent with their actual behavior; requiring older novices—particularly those ages 18 to 20— to hold permits for minimum periods may reduce their initial crash rates. © 2014 National Safety Council and Elsevier Ltd. All rights reserved.

1. Introduction Supervised driving while on a learner permit is a relatively safe type of driving exposure that allows novices to gain driving experience under conditions of reduced risk (Mayhew, Simpson, & Pak, 2003; Williams, 2003; Williams, Preusser, Ferguson, & Ulmer, 1997). However, once novices are allowed to drive unsupervised their crashes increase radically—with crash rates per driver that are 3 to 7 times higher for at least the first 6 months of unsupervised driving (Lewis-Evans, 2010; Mayhew et al., 2003; VicRoads, 2008). These higher crash rates are likely due to increased driving exposure, but they also reflect errors made during the shift from co-driving under supervision to self-regulated solo driving (Forsyth, Maycock, & Sexton, 1995; Masten, Foss, & Marshall, 2011). Fortunately, the crash rates of newly licensed drivers decrease dramatically within the first months of licensure, after which the decline continues for years at a less steep rate; a learning curve that persists even among novices licensed under graduated driver licensing (GDL), which requires them to hold a learner permit for a minimum time and limits their exposure to high risk conditions (i.e., nighttime ⁎ Corresponding author. Tel.: +1 11 1 916 657 6638; fax: +1 11 1 916 657 8589. E-mail address: [email protected] (S.V. Masten).

http://dx.doi.org/10.1016/j.jsr.2014.05.005 0022-4375/© 2014 National Safety Council and Elsevier Ltd. All rights reserved.

driving and transporting young passengers) when they initially begin to drive independently (Lewis-Evans, 2010; Masten & Foss, 2010; Mayhew et al., 2003; Twisk & Stacey, 2007; VicRoads, 2008). The types of driving errors that result in crashes during the initial months of independent licensure by teen novices—who comprise the majority of novice drivers—do not seem to be due to “youthfulness,” deliberate risk-taking, or overconfidence; rather they seem to reflect skill and understanding deficiencies associated with being inexperienced drivers (Braitman, Kirley, McCartt, & Chaudhary, 2008; Curry, Hafetz, Kallan, Winston, & Durbin, 2011, Curry, McDonald, Kandadai, Sommers, & Winston, 2013; Foss, Martell, Goodwin, & O'Brien, 2011; Irwin, 1996; McCartt, 2013; McKnight & McKnight, 2003). Although inexperience appears to initially be the most important factor for explaining the high crash rates of teen novices upon licensure, driver age is also an important factor given that their crash rates remain elevated relative to older novice drivers for years subsequent (Mayhew, 2007; Mayhew et al., 2003; McCartt, Mayhew, Braitman, Ferguson, & Simpson, 2009). Even with the same level of experience and driving exposure, younger novices tend to have higher crash rates than do older novices, with even just a 1-year increase in age at licensure being associated with having fewer crashes (Mayhew et al., 2003; Waller, Elliott, Shope, Raghunathan, & Little, 2001). However, older novices do appear

126

E.A. Chapman et al. / Journal of Safety Research 50 (2014) 125–138

to have higher traffic violation rates per driver—to some extent due to their greater driving exposure—which makes finding older age of licensure to be associated with fewer crashes even more remarkable (Ferdun, Coppin, & Peck, 1965; U.S. Department of Transportation & Federal Highway Administration, 2009; Waller et al., 2001). The youngest teen drivers—those ages 16 and 17 in California and most other U.S. states—typically have the highest crash rates due to their combination of both inexperience and young age (Ferdun et al., 1965; Williams, 2003). While both inexperience and young age are important factors for elevating 16–17-year-old crash rates, their relative influences change as a function of time licensed (McCartt et al., 2009; McKnight & McKnight, 2003). After novice 16–17-year-olds master basic vehicle handling skills and also gain confidence driving without adult supervision during the initial months of licensure, inexperience would be expected to become less of a causal factor for their crashes (Waller, 2003; Waller et al., 2001). Perhaps their crash rates remain elevated relative to other age groups of drivers because they deliberately begin to test the limits of their driving abilities in order to optimize their skills and gain further mastery (Ericsson, Krampe, & Tesch-Römer, 1993). If so, this shift would be expected to be expressed as changes in the types of crashes in which 16–17-year-old drivers are involved and the types of traffic violations that they receive over the course of early licensure—with higher rates of inexperience-related crashes and violations during initial unsupervised driving, and higher rates of overconfidence-related crashes and violations subsequently (McKnight & McKnight, 2003; Waller et al., 2001). The primary goal of this study was to describe how the monthly overall crash and traffic violation rates of novice 16–17-year-old drivers licensed under a GDL program change as a function of time; also presented are the rates for novice drivers age 18 and older who were not licensed under GDL (Williams, Tefft, & Grabowski, 2012). Because traffic violations are more common driving events than crashes, and because crashes that result from disobeying traffic laws often result in traffic violations, studying both outcomes together provides a more robust picture of the actual driving behavior of novices over the course of initial licensure than does studying only crashes (Gebers & Peck, 2003; McKnight & McKnight, 2003). Percentages of novice drivers by age who remained crash-free after 3 years of licensure are also presented to provide a normative context to help interpret their absolute crash risk. Another goal of this study was to characterize how the types of traffic violations for which novice 16–17-year-old drivers licensed under a GDL program are convicted vary during the course of licensure. Specifically, the goal was to determine whether they initially tend to receive traffic violations for actions associated with inexperience and a lack of driving skill—such as violations for disobeying traffic signs/ signals, making improper maneuvers, or failing to maintain proper lane position—whereas the traffic violations that they receive after gaining some independent driving experience tend to be more reflective of overconfidence and limit-testing—such as violations for speeding, disobeying license restrictions, using cell phones, being unbelted, and alcohol/drug-related driving (McKnight & McKnight, 2003; Waller et al., 2001). Again the rates for novice age 18 and older drivers who were not licensed under GDL are also presented for comparison purposes. 2. Method 2.1. The California licensure process Novice drivers licensed at ages 16 or 17 are subject to the California GDL program, which is a three-stage specialized licensing system for persons under age 18. The GDL program includes a 6-month minimum learner permit holding period prior to licensure (Stage 1) during which young teens must drive under adult supervision for at least 50 h. To obtain this permit they must be at least age 15½, have completed a driver education course, and be enrolled in a behind-the-wheel driver training

course. After holding the learner permit for at least 6 months and completing driver training, they may obtain a provisional license (Stage 2) with which they are allowed to drive unsupervised; however, they are restricted from driving during 11 PM to 5 AM and from transporting any passenger younger than age 20. The provisional license period lasts for 12 months or until they turn age 18, after which they may drive under all conditions (Stage 3). Note that the passenger restriction only applied to the first 6 months of driving and the nighttime restriction start time was 12 AM for those licensed before January 1, 2006. All persons in the sample licensed at age 16 would be subject to one or both of the provisional licensing restrictions for 12 full months. Persons in the sample licensed at age 17 would only be subject to the provisional restrictions until they turned age 18; only about 60% of the 17year-olds in the sample were subject to these restrictions for 6 months or longer. Hence, the crash and violation rates of 16-year-old drivers during their first year of unsupervised licensure are likely to be influenced more by the GDL provisional restrictions than are those of 17year-old drivers. Novice drivers licensed at ages 18 or older are not subject to the GDL program. They are also not required to complete driver education or driver training courses. Although they are issued learner permits valid for 12 months prior to licensure when they pass the written knowledge test, they are not required to hold the permits for a minimum length of time or practice a minimum number of hours with a supervisor before being allowed to take the on-road driving test. Furthermore, they are not subject to nighttime or passenger driving restrictions (i.e., provisional licensure) after they are licensed to drive unsupervised upon passing the driving test. 2.2. Data source and coding procedures All persons who applied for and obtained a novice California noncommercial driver license at ages 16 or 17 from January 1, 2001 to December 31, 2007 were identified from the Provisional Licensing Database of the California Department of Motor Vehicles (DMV). For comparison purposes a 10% random sample of all persons ages 18 to 35 who obtained a novice California non-commercial driver license during this same time period was extracted from the typical licensing DMV database. This time period was chosen because the Provisional Licensing Database, which was used to identify large numbers of novice 16 and 17-year-old drivers, was not established until 2001. Furthermore, this time period resulted in a sample of drivers who all had at least 3-years of post-licensing driver records for comparison purposes. Drivers with evidence of licensure in a prior jurisdiction were excluded so that the sample consisted of novices with no prior unsupervised driving experience. Drivers who subsequently obtained commercial licenses were also excluded because original license data elements in DMV's database are overwritten for such drivers. The final sample of 1,709,342 novices was classified according to the age in years at which the drivers obtained their original California driver license into the following age groups: 16, 17, 18, 19, 20, 21–24, and 25–35 years-old. Crash and traffic violation records for the sample were extracted from the California DMV database for the 3-year period following their dates of original license issuance. This database contains information on all motor vehicle crashes in California reported by law enforcement, insurance companies, and drivers. Crashes are required to be reported to the DMV if they involve an injury or death, or at least $750 in property damage. In addition to total crashes, which include all injury severity levels and those involving only property damage, the crashes were also aggregated separately into those in which at least one person was killed or injured (fatal/injury crashes) and those in which the novice driver was specifically found to be at fault by law enforcement (atfault crashes). The DMV database also contains data on all court-reported traffic violations in California for which a driver is convicted, failed-to-appear, or had the citation dismissed after attending traffic violator school. These


VicRoads, 2008), the crash and violation outcomes were aggregated for analysis purposes into those occurring during the 12 months immediately prior to licensure and five post-licensure time periods that differentiated these early high risk months from the lower-risk second and third years of licensure: (a) 1–3 months, (b) 4–6 months, (c) 7–12 months, (d) 13–24 months, and (e) 25–36 months subsequent to licensure. However, monthly crude rates per 10,000 drivers are presented in the figures to show finer detail of changes over time.

Table 1 Categorization and distribution of traffic violations occurring during the first 3 years after novice licensure, California 2001–2007. Violation category

Violations

%

Overall Inexperience-related violations Disobeyed traffic sign/signal Improper maneuver Poor lane position Miscellaneous inexperience Overconfidence-related violations Exceeded maximum speed limit Unsafe speed for conditions Unbelted driver or passenger Driving with suspended/revoked license Alcohol/drug-related driving Use of cellular device Reckless driving (other than speeding) GDL restriction violation Miscellaneous overconfidence Other types of violations Failure to appear for courta No evidence of insurance Vehicle registration related Miscellaneous other

5,673,467 1,067,680 528,161 340,749 141,979 56,791 2,975,873 1,113,113 817,797 318,032 198,770 193,091 176,054 85,187 45,433 28,396 1,629,914 863,230 198,770 124,941 442,973

100.0 18.8 9.3 6.0 2.5 1.0 52.5 19.6 14.4 5.6 3.5 3.4 3.1 1.5 0.8 0.5 28.7 15.2 3.5 2.2 7.8

127

2.3. Analysis method Poisson regression models were used to estimate adjusted rate ratios comparing age 16, 17, 18, 19, 20, and 21–24 novices to those ages 25–35 during each time interval. Higher percentages of females than males were licensed at ages 19 or older, so driver gender was used as a covariate in the models. Calendar month of licensure, calendar year of licensure, and their interaction, were included as categorical covariates in the models to account for seasonal effects and historical trends. The covariates were interacted with licensing age to allow for the effects of gender, calendar year of licensure, and month of licensure to vary by novice driver age. The natural logarithm of the inverse of 10,000 was used as an offset term in the models to create rates per 10,000 licensed drivers (McCullagh & Nelder, 1989). Generalized estimating equations with a first-order autoregressive correlation matrix and robust (empirical) variance were used in the models to account for any correlation due to repeated measurements of the same drivers over time (Liang & Zeger, 1986). Mean monthly adjusted rates per 10,000 drivers and crude cumulative percentages of novices experiencing at least one outcome event (pre-licensure and also post-licensure) were computed for descriptive purposes. SAS software version 9.2 was used to conduct the analyses. Rate ratios were considered to be statistically no different from the null if their 95% confidence interval included 1.0, a procedure that is equivalent to two-sided hypothesis test using an alpha level of 0.05.

Note. GDL = graduated driver licensing program (age 16 and 17 drivers only); violations of the nighttime and passenger restrictions during provisional licensure. Categories are based on prior schemes from McKnight and McKnight (2003) and Waller et al. (2001). About 51.3% of novices were convicted of at least one violation during follow-up; 24.4% were convicted of violations that occurred on two or more occasions. a Includes other categories, but the specific vehicle code section violated is unknown.

are collectively called “traffic violations” for study purposes and were accumulated as of the date of the traffic violation—not the date of conviction—so as to reflect the date of the behavior in question rather than delays due to adjudication. Total traffic violations include both minor violations (e.g., failure to yield, lane positioning, or vehicle control violations) and major violations (e.g., reckless driving, alcohol/ drug-related driving, or driving on a suspended/revoked license). The various types of traffic violations were categorized using a taxonomy (Table 1) based on prior categorization schemes developed by McKnight and McKnight (2003) for crash causal factors and Waller et al. (2001) for traffic violations as being indicative of: (a) inexperience or a lack of driving skill (e.g. hazard recognition, visual search, and vehicle control violations); (b) overconfidence, volitional acts of risk-taking, distracted driving, or testing limits (e.g., excessive speed, cell phone use, GDL restriction violations, and alcohol/drug-related violations); or (c) other types of violations (e.g., equipment violations). In addition to comparing rates of inexperience-related and overconfidence-related violations, those involving alcohol/drugs use or violations of major traffic laws were also considered separately. Because crash rates for newly licensed teens decrease remarkably within the first months of unsupervised licensure, after which the decline continues for years at a less steep rate (Lewis-Evans, 2010; Masten & Foss, 2010; Mayhew et al., 2003; Twisk & Stacey, 2007;

3. Results 3.1. Length of learner permit holding by age at novice licensure While all 16–17-year-old novices held their learner permits for at least 6 months in accordance with the GDL requirement to do so, majorities of 16-year-olds (56.8%) and 17-year-olds (73.2%) actually held their learner permits for longer than the 6 month minimum (Table 2). In fact, 30.7% of 16-year-olds and 48.5% of 17-year-olds held their learner permits for 9 months or longer. On the contrary, majorities (ranging from 67.1% to 81.1%) of age 18 or older novices held their learner permits for less than 6 months, with the percentages who held their permit for 1 month or less increasing as a function of licensure age. Hence, the crash and violation rates of 16–17-year-old novices presented later reflect the effects of 6 month or longer periods of learner permit holding,

Table 2 Percentages of novice drivers holding learner permits for different lengths of time by age of licensure, California 2001–2007. Age

Drivers

161 171 18 19 20 21–24 25–35

485,283 183,682 32,106 12,256 7844 23,049 37,042

Months learner permit held 0

1

2

3

4

5

6

7

8

9

10

11

12+

0.0 0.0 23.6 25.8 30.3 43.2 45.9

0.0 0.0 14.5 12.7 13.3 12.4 12.2

0.0 0.0 10.6 9.4 9.1 8.2 7.8

0.0 0.0 9.2 7.2 7.1 6.1 6.2

0.0 0.0 7.9 6.4 5.9 4.9 4.7

0.0 0.0 7.7 5.6 5.1 4.4 4.3

43.2 26.8 6.8 5.2 4.6 3.8 3.7

15.9 14.3 4.7 4.7 4.4 3.1 3.1

10.2 10.4 3.8 4.4 3.7 2.9 2.5

8.2 8.4 3.3 4.1 3.4 2.6 2.2

8.1 8.5 3.0 4.6 4.2 2.7 2.5

9.9 14.0 4.6 9.5 8.2 5.4 4.4

4.5 17.6 0.3 0.7 0.8 0.4 0.6

Note. Figures are estimates based on the 781,262 novices (45.7% of the original sample) who had not yet renewed, upgraded, or added endorsements to their licenses. Lengths learner permits were held differed by age, χ2(72, N = 781,262) = 653,402.84, p b .001. 1 Drivers licensed at ages 16 or 17 are required to hold their learner permits for at least 6 months under the graduated driver licensing program.

128


while those for age 18 or older novices generally reflect very short periods of learner permit holding. 3.2. Total crash and traffic violation rates of novices as a function of time 3.2.1. Total crashes The crude total crash rates of novice 16- and 17-year-olds—as well as most other age groups—were highest almost immediately after they

were licensed to drive unsupervised, after which their rates declined quickly during their first year of licensure and at a slower rate for the second and third years (Fig. 1, upper). These patterns over time held after adjusting for differences in driver gender and calendar year of licensure (Table 3). Note that the highest crash rates for most age groups occurred the second month after licensure rather than during the first month. Pre-licensure crash rates for all age groups peaked immediately prior to licensure, were much lower than at any point after licensure, and those for 16–17-year-olds were notably lower than those for most

Fig. 1. Crude total crash rates (upper) and traffic violation rates (lower) per 10,000 novice drivers by age and months before and after licensure.


129

Table 3 Average monthly adjusted total crash and traffic violation rates per 10,000 novice drivers and cumulative percentages of novices who had at least one outcome event before and after licensure by age, California 2001–2007. Novice age (drivers)

Total crashes Pre-licensure2

161 (1,013,801) Rate Cumulative % 171 (426,246) Rate Cumulative % 18 (80,375) Rate Cumulative % 19 (30,192) Rate Cumulative % 20 (19,001) Rate Cumulative % 21–24 (53,593) Rate Cumulative % 25–35 (86,134) Rate Cumulative %

Total traffic violations Pre-licensure2

Months after licensure 1–3

4–6

7–12

13–24

25–36

7.8 0.7

160.4 3.7

147.7 7.1

138.5 12.9

120.7 22.0

94.3 28.3

12.1 1.1

166.1 3.8

155.8 7.4

145.7 13.5

120.1 22.4

13.9 1.4

166.0 4.2

141.8 7.8

119.8 13.3

18.6 1.8

141.9 3.4

126.9 6.4

27.9 2.0

180.6 3.1

22.1 2.0 18.1 1.8


4–6

7–12

13–24

25–36

15.4 1.2

242.2 4.6

299.5 9.9

333.4 20.0

394.2 38.0

437.1 51.3

93.7 28.6

42.9 3.1

297.3 5.5

386.2 11.8

457.2 24.0

506.1 42.7

470.4 54.1

95.6 21.3

78.5 27.2

89.0 6.6

491.0 8.9

566.5 17.5

573.8 30.9

553.8 48.6

495.2 59.1

114.4 11.4

88.4 18.7

72.6 24.0

123.3 8.3

442.2 7.7

500.9 15.0

489.1 26.5

476.4 42.7

425.6 52.9

173.4 6.1

141.3 10.7

115.6 17.4

96.5 22.6

127.7 8.9

406.9 7.3

434.2 13.8

440.4 25.0

421.7 40.1

376.3 49.8

104.5 2.4

95.0 4.6

88.3 8.5

74.6 14.5

62.9 19.0

119.7 9.2

310.3 6.2

324.6 11.9

320.6 21.4

301.6 34.6

276.0 43.6

60.8 1.6

62.0 3.1

59.0 6.0

50.6 10.5

46.5 14.4

98.6 7.3

215.3 4.2

227.7 8.3

223.6 15.2

216.6 26.0

207.1 34.0

Note. Rates were adjusted for driver gender, calendar month of licensure, and calendar year of licensure, and the interaction of calendar year and month of licensure. 1Licensed under the graduated driver licensing program; otherwise not licensed under GDL. 2During the 12 months prior to licensure.

older novices (Table 4). Also noteworthy is the fact that over 70% of 16and 17-year-old novices were crash-free for the first 3 years of unsupervised licensure. 3.2.2. Total traffic violations The crude total traffic violation rates of novice 16- and 17-year-olds were not highest immediately after licensure; they reached their apex long after their total crash rates had peaked and declined (Fig. 1, lower). In fact, the traffic conviction rates of 16- and 17-year-olds increased steadily after licensure until around the time they turned age 18, and then almost immediately began to decline. This pattern over time remained after adjusting for confounders. The traffic violation rates for all other age groups of novices were highest during their first year of

unsupervised driving, typically only several months after their peak crash rates had declined. Violation rates for all age groups were lower during their first month of licensure than during the immediately subsequent months. Pre-licensure total traffic violation rates of all age groups peaked immediately prior to licensure, with the lowest rates for 16- and 17-yearolds. The violation rates immediately prior to licensure for 19-, 20-, and 21–24-year-old novices were higher or about the same as those immediately afterwards; upon further inquiry it was found that about one-third of the pre-licensure violations for these drivers involved driving without a license or a related violation. Only about 45% of 16- and 17-year-old novices were traffic violation-free for their first 3 years of unsupervised licensure, which reinforces the fact that traffic violations are much more common driving events than are crashes.

Table 4 Adjusted rate ratios for total crashes and traffic violations comparing teen and young adult novices to those aged 25–35. Outcome Novice age

Total crashes 161 171 18 19 20 21–24 25–35 Total traffic violations 161 171 18 19 20 21–24 25–35

Adjusted rate ratio (95% confidence interval) Pre-licensure2


4–6

7–12

13–24

25–36

0.43 (0.37–0.50) 0.67 (0.58–0.78) 0.77 (0.64–0.93) 1.03 (0.81–1.31) 1.55 (1.19–2.01) 1.22 (1.00–1.5) 1.00 (ref)

2.64 (2.29–3.04) 2.73 (2.36–3.16) 2.73 (2.27–3.27) 2.33 (1.84–2.95) 2.97 (2.30–3.83) 1.72 (1.40–2.11) 1.00 (ref)

2.38 (2.07–2.74) 2.51 (2.17–2.91) 2.29 (1.91–2.75) 2.05 (1.62–2.60) 2.80 (2.17–3.61) 1.53 (1.25–1.88) 1.00 (ref)

2.35 (2.05–2.69) 2.47 (2.14–2.85) 2.03 (1.70–2.43) 1.94 (1.54–2.44) 2.40 (1.87–3.06) 1.50 (1.23–1.83) 1.00 (ref)

2.39 (2.09–2.73) 2.38 (2.07–2.73) 1.89 (1.59–2.25) 1.75 (1.39–2.19) 2.29 (1.8–2.91) 1.48 (1.21–1.79) 1.00 (ref)

2.03 (1.78–2.32) 2.02 (1.75–2.32) 1.69 (1.42–2.01) 1.56 (1.24–1.96) 2.08 (1.63–2.65) 1.35 (1.11–1.65) 1.00 (ref)

0.16 (0.14–0.17) 0.43 (0.40–0.48) 0.90 (0.81–1.01) 1.25 (1.09–1.44) 1.29 (1.10–1.53) 1.21 (1.07–1.38) 1.00 (ref)

1.12 (1.03–1.23) 1.38 (1.26–1.52) 2.28 (2.04–2.55) 2.05 (1.78–2.37) 1.89 (1.60–2.24) 1.44 (1.27–1.64) 1.00 (ref)

1.32 (1.21–1.43) 1.70 (1.55–1.86) 2.49 (2.23–2.78) 2.20 (1.91–2.53) 1.91 (1.61–2.26) 1.43 (1.25–1.62) 1.00 (ref)

1.49 (1.37–1.62) 2.04 (1.87–2.23) 2.57 (2.31–2.85) 2.19 (1.91–2.51) 1.97 (1.67–2.32) 1.43 (1.27–1.62) 1.00 (ref)

1.82 (1.68–1.97) 2.34 (2.14–2.55) 2.56 (2.30–2.83) 2.20 (1.92–2.51) 1.95 (1.66–2.28) 1.39 (1.23–1.57) 1.00 (ref)

2.11 (1.95–2.29) 2.27 (2.08–2.48) 2.39 (2.16–2.65) 2.05 (1.80–2.35) 1.82 (1.55–2.13) 1.33 (1.18–1.51) 1.00 (ref)

Note. Rates were adjusted for driver gender, calendar month of licensure, and calendar year of licensure, and the interaction of calendar year and month of licensure; all covariates were interacted with age of licensure. 1 Licensed under the graduated driver licensing program; otherwise not licensed under GDL. 2During the 12 months prior to licensure.

130


3.3. Fatal/injury and at-fault crash rates of novices as a function of time 3.3.1. Fatal/injury crashes The crude fatal/injury crash rates of novice 16- and 17-year-olds peaked soon after they were licensed, again with the second month after licensure being the highest (Fig. 2, upper). Their fatal/injury crash rates also declined quickly during their first year of licensure and at a slower rate for the second and third years. Note the “bump” in the 16-year-old fatal/injury crash rate at 7-months, which corresponded with the end of the 6-month passenger restriction of the provisional license stage for most of these novices. The passenger restriction only applied to the first 6 months of driving and the nighttime restriction start time was 12 AM for those licensed before January 1, 2006. This includes the majority of the study sample because it consisted of persons who applied for and obtained a novice California non-commercial driver license at ages 16 or 17 from January 1, 2001 to December 31, 2007. For those licensed on or after January 1, 2006 the passenger restriction lasts for 12 months and the nighttime restriction began at 11 PM, though they both end early if the novices turn age 18. This temporary increase is not apparent among 17-year-olds, perhaps because only about 60% of the 17-year-olds in the sample were subject to the GDL restrictions for 6 months or longer. The fatal/injury crash rates for age 18 and older novices were also highest during their first year of unsupervised licensure, with dramatic reductions in the rates of 18–24-year-olds during the first 3 months of licensure, even after adjusting for confounders (Table 5). The pre-licensure fatal/injury crash rates for all age groups were generally very low—particularly for 16- and 17-year-olds—with all ages peaking immediately prior to licensure (Table 6). Over 90% of 16- and 17-year-old novices were not involved in a fatal/injury crash during their first 3 years of unsupervised licensure.

3.3.2. At-fault crashes The patterns of crude at-fault crash rates of the novice drivers were similar to those for their total crashes; the rates were highest starting the second month after licensure, with steep declines during the first year, and smaller reductions over subsequent years (Fig. 2, lower). While the at-fault crash rates of 18-year-old novices seem particularly high relative to other novices, their rates were only reliably different from those of novices ages 21 or older. Pre-licensure at-fault crashes also peaked immediately prior to licensure, with particularly low rates for 16-year-olds, and also to some extent 17-year-olds. Over 87% of 16- and 17-year-olds were not involved in an at-fault crash within 3 years of unsupervised licensure.

3.4. Inexperience-related and overconfidence-related traffic violation rates of novices as a function of time

3.4.2. Over-related traffic violations Overconfidence-related traffic violations of 16- and 17-year-old novices also peaked around the time these drivers turned age 18— again a year or more after their crash rates were highest—although the increases during their first year of licensure were more dramatic for overconfidence-related violations than was the case for their inexperience-related violations (Fig. 3, lower). In contrast, the overconfidence-related violation rates for age 18 and older novices were highest during their first year of licensure—the year their crash rates were also highest—although their peak overconfidence-related violation rates occurred later during this first year than did their peak rates of crashes or inexperience-related violations. Pre-licensure overconfidence-related violation rates mostly mirrored those for inexperience-related violations, with peaks immediately prior to licensure and particularly low rates for 16- and 17-year-old novices. During their first 3 years of licensure, over 57% of 16- and 17-year-old novices were not convicted of an overconfidence-related traffic violation. 3.5. Major and alcohol/drug-related traffic violation rates of novices as a function of time 3.5.1. Major traffic violations The crude rates of major traffic violations increased over the first year of licensure for all age groups of novices, and did not peak until at least the third year after licensure for all those under age 20 (Fig. 4, upper). The rate increases were notably large for 16- and 17-year-olds during the year leading up to when they turned age 18 (Table 9). Prelicensure major violation rates were extremely low for 16- and 17year-olds, as were their rates immediately after licensure (Table 10). Pre-licensure major violation rates for all other novices noticeably peaked immediately before licensure—with the rate for age 21–24year-old novices actually exceeding their rate immediately after licensure. Over 95% of 16- and 17-year-old novices were not convicted of a major traffic violation during their first 3 years of licensure. 3.5.2. Alcohol/drug-related traffic violations Alcohol/drug-related traffic violations of 16- and 17-year-olds increased throughout their first 3 years of licensure, with the most striking increases again occurring during the year immediately prior to when they turned age 18 (Fig. 4, lower). The rates of alcohol/drug violations among 16- and 17-year-old novices were extremely low relative to other novices both before licensure and for a year or more afterwards. Alcohol/drug-related violation rates among novices ages 18 to 24 started out high and remained that way for at least 3 years after licensure relative to novices age 25 or older. During their first 3 years of licensure, over 97% of 16- and 17-year-old novices were not convicted of an alcohol/drug-related traffic violation. 4. Discussion 4.1. General discussion of findings

3.4.1. Inexperience-related traffic violations The crude rates of inexperience-related traffic violations for novice 16- and 17-year-olds peaked around the time these drivers were age 18, after which they slowly declined (Fig. 3, upper). This was unexpected given that their crash rates—which would be expected to initially be elevated due to inexperience—were highest almost immediately after licensure. The inexperience-related traffic violation rates for all other age groups of novices were highest during their first year of licensure (Table 7), when their crash rates also peaked. Pre-licensure rates of inexperience-related violations were highest immediately prior to licensure, with the rates for 16- and 17-yearolds being lower than those for all other age groups of novices (Table 8). Over 76% of 16- and 17-year-old novices were not convicted of an inexperience-related traffic violation within their first 3 years of licensure.

The crash rates of 16- and 17-year-old California novice drivers—as well as those of most novice drivers age 18 or older—are highest during the first few months of unsupervised licensure, after which their rates decline quickly during the first year and continue to decline slowly during subsequent years. This general pattern is consistent with the distributions of young novice crash rates found in other jurisdictions (Lewis-Evans, 2010; Masten & Foss, 2010; Mayhew et al., 2003; Twisk & Stacey, 2007; VicRoads, 2008; Waller et al., 2001). Because California issues over 11% of all U.S. driver licenses (U.S. Federal Highway Administration, 2012), large enough samples were also available to describe the driver learning curves of several groups of adult novices. The shapes of the various novice driver learning curves—which are similar across total, fatal/injury, and at-fault crashes—are comparable to those for other complex tasks in which learners make more errors early in


131

Fig. 2. Crude fatal/injury crash rates (upper) and at-fault crash rates (lower) per 10,000 novice drivers by age and months before and after licensure.

the learning process, and exponentially fewer errors as experience is acquired (Anderson, 1993; Anderson, Fincham, & Douglass, 1999). Although errors committed while learning to drive can produce crashes, the majority of California novice drivers of all ages do not crash within their first 3 years of unsupervised driving—including more than twothirds of 16- and 17-year-old novices (Simons-Morton, Hartos, Leaf, & Preusser, 2006). The traffic violation rates of 16- and 17-year-old California novice drivers are distributed differently than their crashes, and differently than the violation rates for all other age groups of novices (Underwood, 2013). Violations for 16- and 17-year-olds increase steadily after licensure until around the time they turn age 18, which is a year or more after their crash rates reach their apex and begin to decline. In contrast, the traffic violation rates for all other age groups of

novices are highest during their first year of unsupervised driving, typically only months after their peak crash rates. This pattern among 16–17-year-old novices was not unexpected—45 years earlier Ferdun et al. (1965) reported that older California teen drivers had higher traffic violation rates than those who were younger. More recently Waller et al. (2001) noted that traffic violation rates among young Michigan novices rose after their first year of licensure—even while their crash rates decreased—with a peak violation rate around 2–3 years postlicensure. The findings presented here tie together these prior results in that they suggest that the peak in violations 2–3 years after licensure observed by Waller et al. (2001) is likely related to the fact that novices licensed at age 18 have exceptionally high violation rates compared to novices licensed at any other age, and also because the violation rates of 16–17-year-old novices peak when they turn age 18.

132


Table 5 Average monthly adjusted fatal/injury and at-fault crash rates per 10,000 novice drivers and cumulative percentages of novices who had at least one crash before and after licensure by age, California 2001–2007. Novice age (drivers)

Fatal/injury crashes Pre-licensure2


At-fault crashes Pre-licensure2


4–6

7–12

13–24

25–36

2.0 0.2

36.9 0.9

33.4 1.7

32.4 3.3

28.1 5.9

22.0 7.9

3.7 0.3

44.7 1.0

41.4 2.0

39.5 3.7

32.8 6.5

4.6 0.4

52.0 1.2

45.6 2.3

37.9 4.0

6.2 0.5

45.9 1.0

43.9 1.9

8.4 0.6

57.5 1.1

5.9 0.6 4.5 0.5


4–6

7–12

13–24

25–36

3.0 0.3

72.2 1.7

62.1 3.1

54.8 5.5

45.6 9.4

34.6 12.1

25.7 8.7

5.8 0.5

76.0 1.8

67.8 3.4

60.6 6.1

48.0 10.1

34.7 12.9

30.5 6.7

24.7 8.9

9.1 0.8

94.3 2.1

79.6 3.9

61.2 6.4

47.9 10.4

35.9 13.1

38.6 3.5

29.4 5.9

23.9 7.9

8.0 0.9

56.8 1.6

51.6 3.0

42.1 5.3

32.1 8.6

24.7 11.0

45.7 1.9

39.1 3.4

33.1 5.8

25.7 7.6

14.6 1.1

82.2 1.5

72.6 2.9

56.1 4.9

43.8 8.0

33.9 10.2

28.1 0.7

24.1 1.4

22.3 2.5

19.6 4.4

16.7 6.0

10.4 0.9

48.1 1.0

42.2 2.0

36.9 3.5

29.1 5.9

22.1 7.6

16.3 0.4

15.6 0.9

15.5 1.8

12.6 3.1

11.6 4.3

8.0 0.7

25.0 0.5

24.1 1.1

22.6 2.0

18.1 3.5

15.3 4.8

Note. Rates were adjusted for driver gender, calendar month of licensure, and calendar year of licensure, and the interaction of calendar year and month of licensure. 1Licensed under graduated driver licensing program; otherwise not licensed under GDL. 2During the 12 months prior to licensure.

The high violation rates of 18-year-olds likely reflect increases in driving exposure resulting from 17-year-olds shedding GDL license restrictions when their provisional licensing phase ends, along with reduced parental controls as they age and become legal adults, and a greater need for young adults to drive for practical reasons such as for employment or to attend college (Simons-Morton & Ouimet, 2006). Cross-sectional estimates of miles driven by age indicate that mileage among 18-year-old drivers is more than double that of drivers ages 16–17 (U.S. Department of Transportation & Federal Highway Administration, 2009). However, additional research is needed that specifically measures longitudinal changes in driving exposure among novices over the first several years of licensure, because existing research indicating that the mileage of novice 16year-olds changes little over the first 18 months of driving may have failed to detect an increase in mileage starting at age 18 due to insufficient follow-up time (Klauer et al., 2011; Lee, Simons-Morton, Klauer, Ouimet, & Dingus, 2011).

It was hypothesized that 16–17-year-old's inexperience-related violations would be highest immediately after licensure while they gain experience driving without the benefit of a supervisor, and that their rates of overconfidence-related violations would be higher later, perhaps because they begin to test the limits of their driving abilities in order to optimize their skills and gain further mastery (Ericsson et al., 1993). In reality, both inexperience-related and overconfidence-related traffic violations of 16–17-year-old novices peaked around the same time—when the drivers turned age 18. The predicted pattern was unexpectedly more apparent among the rates of all other age groups of novices, with their highest overconfidence-related violation rates occurring later during their first year of driving than their peak rates of crashes or inexperience-related violations. A similar hypothesis by Waller et al. (2001) predicting that driving experience would result in fewer minor traffic violation over time, but with less of an effect on major violations, was likewise not supported in their study. The

Table 6 Adjusted rate ratios for fatal/injury and at-fault crashes comparing teen and young adult novices to novices aged 25–35. Outcome Novice age Fatal/injury crashes 161 171 18 19 20 21–24 25–35 At-fault crashes 161 171 18 19 20 21–24 25–35

Pre-licensure2


4–6

7–12

13–24

25–36

0.44 (0.34–0.58) 0.83 (0.62–1.10) 1.04 (0.73–1.46) 1.38 (0.89–2.13) 1.87 (1.16–3.02) 1.32 (0.89–1.96) 1.00 (ref)

2.26 (1.72–2.98) 2.74 (2.07–3.64) 3.19 (2.27–4.47) 2.82 (1.84–4.30) 3.53 (2.21–5.63) 1.72 (1.16–2.56) 1.00 (ref)

2.13 (1.62–2.81) 2.64 (1.99–3.52) 2.92 (2.08–4.09) 2.81 (1.84–4.29) 2.92 (1.82–4.68) 1.54 (1.04–2.29) 1.00 (ref)

2.09 (1.61–2.72) 2.55 (1.94–3.36) 2.45 (1.76–3.40) 2.49 (1.65–3.76) 2.53 (1.60–3.99) 1.44 (0.98–2.11) 1.00 (ref)

2.24 (1.72–2.90) 2.61 (1.99–3.42) 2.43 (1.75–3.36) 2.34 (1.56–3.51) 2.63 (1.68–4.12) 1.56 (1.07–2.28) 1.00 (ref)

1.89 (1.45–2.45) 2.21 (1.68–2.89) 2.12 (1.53–2.93) 2.05 (1.37–3.09) 2.21 (1.41–3.47) 1.44 (0.99–2.10) 1.00 (ref)

0.37 (0.29–0.46) 0.72 (0.56–0.91) 1.13 (0.84–1.51) 1.00 (0.67–1.49) 1.82 (1.20–2.77) 1.30 (0.92–1.81) 1.00 (ref)

2.89 (2.29–3.66) 3.05 (2.38–3.89) 3.78 (2.84–5.04) 2.28 (1.53–3.38) 3.29 (2.18–4.98) 1.93 (1.38–2.70) 1.00 (ref)

2.58 (2.04–3.27) 2.82 (2.20–3.60) 3.31 (2.48–4.42) 2.14 (1.44–3.19) 3.02 (1.99–4.57) 1.75 (1.25–2.46) 1.00 (ref)

2.43 (1.93–3.05) 2.68 (2.12–3.40) 2.71 (2.05–3.58) 1.87 (1.26–2.75) 2.49 (1.66–3.73) 1.64 (1.18–2.27) 1.00 (ref)

2.51 (2.01–3.14) 2.65 (2.09–3.34) 2.64 (2.00–3.48) 1.77 (1.21–2.60) 2.41 (1.62–3.59) 1.60 (1.16–2.21) 1.00 (ref)

2.26 (1.81–2.83) 2.28 (1.80–2.88) 2.35 (1.78–3.10) 1.62 (1.10–2.39) 2.22 (1.49–3.32) 1.45 (1.04–2.00) 1.00 (ref)

Note. Rates were adjusted for driver gender, calendar month of licensure, and calendar year of licensure, and the interaction of calendar year and month of licensure; all covariates were interacted with age of licensure. Values in parentheses are 95% confidence intervals. 1 Licensed under the graduated driver licensing program; otherwise not licensed under GDL. 2During the 12 months prior to licensure.


133

Fig. 3. Crude inexperience-related (upper) and overconfidence-related (lower) traffic violation rates per 10,000 novice drivers by age and months before and after licensure.

compelling theoretical model supporting these hypotheses (Ericsson et al., 1993), paired with the existence of the predicted pattern among age 18–19-year-old novices in the present study—who tend to have more stable levels of driving exposure—leads us to suspect that this pattern probably also exists for 16–17-year-old novices, but could not be

detected because of confounding due to changes in driving exposure (Ferdun et al., 1965). Longitudinal research measuring 16–17-year-old monthly changes in driving exposure and rates of inexperiencerelated and overconfidence-related traffic violations over the first several years of licensure may yet show this hypothesis to be supported.

134


Table 7 Average monthly adjusted inexperience-related and overconfidence-related traffic violation rates per 10,000 novice drivers and cumulative percentages of novices who had at least one violation before and after licensure by age, California 2001–2007. Novice age (drivers)

Inexperience traffic violations

Overconfidence traffic violations

Pre-licensure2

Pre-licensure2



4–6

7–12

13–24

25–36


4–6

7–12

13–24

25–36

3.4 0.3

67.4 1.5

75.6 3.2

86.7 6.8

97.0 14.2

94.7 20.4

11.6 0.9

167.7 3.3

212.7 7.3

227.0 14.8

263.7 29.2

296.7 41.2

9.4 0.8

90.4 1.9

105.4 4.2

123.0 8.9

123.9 17.4

110.1 23.9

33.9 2.5

192.5 3.6

252.5 8.1

290.9 17.0

320.9 32.1

295.2 42.2

18.7 1.6

144.6 3.1

146.6 6.2

147.5 11.6

132.4 20.3

113.8 26.8

67.0 5.1

264.4 5.1

326.6 10.9

333.1 20.7

324.9 35.3

293.6 45.0

24.0 2.1

118.6 2.6

128.7 5.4

118.3 10.0

110.6 17.6

99.3 23.6

90.6 6.1

233.4 4.1

272.8 8.6

281.4 16.5

281.5 29.0

256.7 37.9

26.5 2.4

116.6 2.6

110.1 5.1

110.0 9.5

97.7 16.6

89.4 22.3

86.2 6.4

194.8 3.8

221.1 7.8

230.8 15.0

227.9 26.5

203.3 34.4

29.3 2.6

101.9 2.3

98.8 4.4

93.7 8.2

88.0 14.8

79.6 19.9

81.8 6.4

156.9 3.3

173.0 6.7

174.7 12.7

164.9 21.9

155.6 29.1

26.1 2.4

80.0 1.8

80.4 3.7

73.9 6.8

69.3 12.2

66.4 16.9

67.4 4.9

104.8 2.0

118.3 4.2

121.4 8.2

119.3 15.1

113.6 20.5


The post-licensure crash peaks were more pronounced for some older age groups of novices than was the case for 16–17-year-old novices—most notably those for 18-year-olds, but also those for 19- and 20-year-olds with regard to fatal/injury crashes. Early differences between novice 16–17-year-old crash rates and those for 18–20-yearolds are likely accentuated by the fact that 16–17-year-olds are required to complete a minimum of 6 months of learner driving on a learner permit prior to licensure and are restricted from driving during late night hours and from transporting passengers during their first year of unsupervised driving as part of the GDL program. Young teens licensed under GDL also face novice driver learning curves, but their learning curves may be lower overall than those for their same-aged peers who are not licensed under GDL (Lewis-Evans, 2010; Masten & Foss, 2010). Whereas majorities of 16–17-year-olds actually hold their learner permits for longer than the minimum 6 months required by the GDL program, majorities of novices age 18 or older hold their learner permits

for less than 6 months. In fact, one-third or more of the novices in each adult age cohort hold their learner permits for 1 month or less. Novice drivers who spend minimal time on their learner permits have higher crash risk after they are licensed to drive unsupervised (Gulliver, Begg, Brookland, Ameratung, & Langley, 2013). Perhaps the post-licensure crash peaks of young adult novice drivers would be dampened if they too were required to hold learner permits for a minimum time period prior to licensure—an approach already adopted for all novice drivers in Connecticut and Maryland—or alternatively if they were licensed under a three-stage GDL program, as is the case for all novices under age 21 in New Jersey (Williams, Chaudhary, Tefft, & Tison, 2010). Two emerging findings about novice crash rates observed in the present study differ from descriptions of young novice crash rates in other jurisdictions (Masten & Foss, 2010; Mayhew et al., 2003; McCartt, Shabanova, & Leaf, 2003; Twisk & Stacey, 2007; VicRoads,

Table 8 Adjusted rate ratios for inexperience-related and overconfidence-related traffic violations comparing teen and young adult novices to novices aged 25–35. Outcome Novice age Inexperience violations 161 171 18 19 20 21–24 25–35 Overconfidence violations 161 171 18 19 20 21–24 25–35

Pre-licensure2


4–6

7–12

13–24

25–36

0.13 (0.11–0.15) 0.36 (0.31–0.41) 0.72 (0.60–0.86) 0.92 (0.72–1.17) 1.01 (0.76–1.35) 1.12 (0.92–1.37) 1.00 (ref)

0.84 (0.74–0.96) 1.13 (0.98–1.30) 1.81 (1.52–2.16) 1.48 (1.17–1.88) 1.46 (1.09–1.94) 1.27 (1.04–1.56) 1.00 (ref)

0.94 (0.82–1.08) 1.31 (1.14–1.51) 1.82 (1.53–2.18) 1.60 (1.26–2.03) 1.37 (1.03–1.83) 1.23 (1.00–1.50) 1.00 (ref)

1.17 (1.03–1.33) 1.66 (1.45–1.91) 2.00 (1.68–2.37) 1.60 (1.27–2.01) 1.49 (1.13–1.97) 1.27 (1.04–1.54) 1.00 (ref)

1.40 (1.23–1.59) 1.79 (1.56–2.04) 1.91 (1.61–2.26) 1.59 (1.27–2.00) 1.41 (1.07–1.85) 1.27 (1.05–1.54) 1.00 (ref)

1.43 (1.26–1.62) 1.66 (1.45–1.90) 1.71 (1.45–2.03) 1.49 (1.19–1.87) 1.35 (1.02–1.77) 1.20 (0.99–1.45) 1.00 (ref)

0.17 (0.15–0.19) 0.50 (0.45–0.56) 0.99 (0.87–1.13) 1.34 (1.13–1.59) 1.28 (1.04–1.57) 1.21 (1.04–1.42) 1.00 (ref)

1.60 (1.43–1.79) 1.84 (1.63–2.07) 2.52 (2.20–2.90) 2.23 (1.87–2.66) 1.86 (1.49–2.31) 1.50 (1.27–1.76) 1.00 (ref)

1.80 (1.61–2.01) 2.13 (1.90–2.40) 2.76 (2.41–3.17) 2.31 (1.94–2.75) 1.87 (1.51–2.32) 1.46 (1.24–1.72) 1.00 (ref)

1.87 (1.68–2.08) 2.40 (2.14–2.68) 2.74 (2.41–3.13) 2.32 (1.96–2.74) 1.90 (1.55–2.34) 1.44 (1.23–1.68) 1.00 (ref)

2.21 (1.99–2.45) 2.69 (2.41–3.00) 2.72 (2.40–3.10) 2.36 (2.00–2.78) 1.91 (1.56–2.34) 1.38 (1.19–1.61) 1.00 (ref)

2.61 (2.35–2.90) 2.60 (2.33–2.90) 2.58 (2.27–2.94) 2.26 (1.92–2.66) 1.79 (1.46–2.19) 1.37 (1.17–1.60) 1.00 (ref)



135

Fig. 4. Crude major traffic violation rates (upper) and alcohol/drug-related traffic violation rates (lower) per 10,000 novice drivers by age and months before and after licensure.

2008), though they were both apparent to some extent in a study of novices in New Zealand (Lewis-Evans, 2010). These findings may refine our understanding how novice crash rates change over time and suggest additional ways to reduce their crash risk, should they be verified in future studies.

First, although it is considered to be “well established” that novice crash rates are the highest during the first month of unsupervised licensure (e.g., Williams et al., 2012, p. 197), the peak crash rates for most age groups of novices in the current study—including 16- and 17-year-olds— did not occur until the second or third months after licensure. Crash

136


Table 9 Average monthly adjusted major and alcohol/drug-related traffic violation rates per 10,000 novice drivers and cumulative percentages of novices who had at least one violation before and after licensure by age, California 2001–2007. Novice age (drivers)


Major traffic violations

Alcohol/drug-related traffic violations

Pre- licensure2

Pre-licensure2


4–6

0.3 0.0

5.2 0.1

6.3 0.2

0.8 0.1

8.2 0.1

2.6 0.2

7–12

13–24

25–36

7.6 0.4

15.0 1.4

25.2 2.8

12.7 0.3

19.9 1.0

29.2 2.6

29.2 0.5

32.8 1.0

40.0 2.2

4.1 0.3

21.4 0.4

25.4 0.9

3.4 0.4

13.3 0.4

10.3 0.7 9.5 0.6


4–6

7–12

13–24

0.0 0.0

1.7 0.0

32.1 4.3

0.1 0.0

45.7 4.5

51.4 6.9

29.3 1.9

34.8 4.1

17.0 0.9

24.8 2.2

23.2 0.4

28.9 0.8

12.0 0.2

16.5 0.5

25–36

2.0 0.1

2.7 0.2

5.8 0.7

10.7 1.6

3.0 0.1

4.9 0.2

7.7 0.5

11.6 1.4

12.6 2.4

1.1 0.1

17.5 0.2

22.8 0.6

22.8 1.2

23.2 2.5

25.5 3.8

36.5 6.2

1.9 0.1

16.7 0.3

19.3 0.7

17.5 1.3

17.2 2.5

18.4 3.7

25.8 4.6

24.5 6.5

1.6 0.1

16.3 0.3

19.6 0.6

21.2 1.3

21.3 2.7

19.1 3.9

31.0 1.8

30.7 3.4

30.6 4.9

2.4 0.2

14.1 0.3

18.4 0.6

17.3 1.3

13.9 2.2

12.7 3.1

18.0 1.0

16.4 1.9

15.2 2.7

2.6 0.2

7.4 0.2

9.2 0.4

7.9 0.7

6.7 1.3

5.9 1.8


rates were also found to peak during the second month of licensure among some New Zealand novices, albeit only those who sought swift full licensure (Lewis-Evans, 2010). Finding that the traffic violation rates of all age groups in the current study were low during the first month of licensure, but greatly increased during the second month, suggests that finding the highest novice crash rates to be during the second or third months of licensure may be a real phenomenon resulting from increased driving exposure rather than error. The delay in peak crashes may reflect the time required for some novices to obtain or otherwise be provided greater access to vehicles; activities that are associated with increased risky driving and crashes among young teens (Cammisa, Williams, & Leaf, 1999). Policies that encourage parents to delay or reduce young novices' primary access to vehicles could potentially become another component of GDL in the future. Second, although crash rates during the 12 months before licensure are low compared to those subsequent—suggesting that driving under supervision on a learner permit is indeed low risk (e.g., Williams et al.,

2012)—the pre-licensure crash rates for all age groups increase notably during the last few months immediately preceding licensure. This pattern was also found among New Zealand novices (Lewis-Evans, 2010), though not necessarily among novices in other jurisdictions (e.g., Mayhew et al., 2003; VicRoads, 2008). Again the corresponding increases in traffic violation rates during the months immediately preceding licensure suggest increased driving exposure is the likely cause, probably reflecting additional practice among novices in anticipation of their approaching driving tests. However, the crash rates for all age 18 and older novices during the months immediately preceding licensure are higher than those for age 16–17 novices. This could indicate that older novices attempt to learn to drive in a very short time period rather than distributed over many months—a practice that is less efficient for learning (Cepeda, Pashler, Vul, Wixted, & Rohrer, 2006; Dempster, 1988). Perhaps this finding also supports requiring minimum learner permit periods for all novices in order to increase the probability that driving practice will occur over an extended period (Waller, 2003),

Table 10 Adjusted rate ratios for major and alcohol/drug-related traffic violations comparing teen and young adult novices to novices aged 25–35. Outcome Novice age Major violations 161 171 18 19 20 21–24 25–35 Alcohol/drug violations 161 171 18 19 20 21–24 25–35

Pre-licensure2


4–6

7–12

13–24

25–36

0.03 (0.02–0.04) 0.08 (0.06–0.11) 0.28 (0.19–0.41) 0.43 (0.25–0.73) 0.35 (0.18–0.70) 1.08 (0.71–1.65) 1.00 (ref)

0.43 (0.31–0.61) 0.69 (0.48–0.98) 2.43 (1.63–3.62) 1.78 (1.04–3.04) 1.11 (0.55–2.24) 1.94 (1.23–3.03) 1.00 (ref)

0.38 (0.27–0.52) 0.77 (0.55–1.08) 1.99 (1.35–2.93) 1.54 (0.91–2.60) 1.03 (0.52–2.05) 1.75 (1.13–2.71) 1.00 (ref)

0.42 (0.31–0.57) 1.11 (0.80–1.53) 2.23 (1.54–3.22) 1.63 (0.99–2.69) 1.38 (0.71–2.67) 1.73 (1.14–2.62) 1.00 (ref)

0.91 (0.68–1.23) 1.78 (1.3–2.44) 2.79 (1.94–4.00) 2.12 (1.30–3.46) 1.58 (0.82–3.03) 1.87 (1.24–2.81) 1.00 (ref)

1.66 (1.23–2.23) 2.11 (1.54–2.89) 3.38 (2.35–4.85) 2.40 (1.47–3.91) 1.61 (0.84–3.09) 2.01 (1.34–3.03) 1.00 (ref)

0.02 (0.01–0.03) 0.05 (0.03–0.09) 0.44 (0.24–0.80) 0.72 (0.34–1.53) 0.63 (0.26–1.52) 0.92 (0.48–1.77) 1.00 (ref)

0.22 (0.14–0.37) 0.41 (0.24–0.69) 2.37 (1.36–4.15) 2.27 (1.09–4.70) 2.21 (0.97–5.03) 1.91 (1.00–3.65) 1.00 (ref)

0.22 (0.14–0.37) 0.54 (0.32–0.90) 2.48 (1.43–4.30) 2.11 (1.02–4.34) 2.13 (0.94–4.81) 2.00 (1.06–3.78) 1.00 (ref)

0.35 (0.22–0.56) 0.97 (0.59–1.60) 2.89 (1.69–4.94) 2.22 (1.10–4.49) 2.68 (1.22–5.90) 2.19 (1.18–4.08) 1.00 (ref)

0.88 (0.55–1.40) 1.74 (1.07–2.84) 3.49 (2.06–5.91) 2.59 (1.29–5.18) 3.21 (1.48–6.96) 2.08 (1.13–3.86) 1.00 (ref)

1.83 (1.14–2.94) 2.16 (1.32–3.52) 4.36 (2.57–7.40) 3.14 (1.57–6.29) 3.27 (1.50–7.10) 2.17 (1.17–4.02) 1.00 (ref)



which is an approach adopted for all novices in three U.S. states that seems to be supported by emerging findings (Gulliver et al., 2013). 4.2. Study limitations The results are based on novice drivers in only one U.S. state—albeit one that contains over 12% of the entire U.S. population—but nonetheless reflect the idiosyncrasies associated with the demography, laws, and driving environment of that state (U.S. Bureau of the Census, 2013). For example, relatively few young teens are licensed in California (less than 15% of 16-year-olds and less than 33% of 17-year-olds are licensed), which may limit the extent to which the results for these age groups are generalizable to young teens in other jurisdictions where their licensure rates are higher. Only traffic violations for which drivers were actually convicted are included in this study, because violations that are dismissed before conviction are never reported to the DMV. However, violations dismissed after the drivers attended traffic violator school were included in the rates. There may be systematic underreporting of minor violations in favor of those that are more serious given limited law enforcement resources. The extent to which this occurs, or whether it would bias the traffic violation rates of some age groups more than others, remains unknown. The classification scheme used to categorize traffic violations as inexperience-related versus overconfidence-related was based on prior schemes used by McKnight and McKnight (2003) and Waller et al. (2001), but there was nonetheless some subjectivity regarding how to classify particular violations. For example, it was not clear how to classify convictions for leaving the scene of a crash in which the driver was involved (i.e., hit-and-run violations) given that such behavior could be a result of both inexperience and overconfidence. Although violations for which the classification was ambiguous were few in number, using a different scheme to classify violations as inexperiencerelated or overconfidence-related may result in different findings than those presented here. Although not presented here, we also replicated the major versus minor violation classification scheme used by Waller et al. (2001) and found that violation rates for both types increased among 16–17-year-olds up until the time they turned age 18, consistent with the results found for inexperience-related versus overconfidencerelated violations. A word of caution seems warranted regarding the appropriateness of different techniques for analyzing traffic convictions versus crashes. We originally intended to use Cox proportional hazards survival analysis for both outcomes, but switched to Poisson regression upon realizing that the former underestimated the various traffic violation rates because of repeat offenses. Specifically, about 24% of novices across all ages were convicted of traffic violations on two or more occasions, which conventional Cox regression does not easily handle. Event time models exist for multiple events or events of different types, but the generalized estimating equations for Poisson regressions that we used account for potential correlation caused by multiple events for the same driver are appropriate for analyzing crash and violation rates. In addition, drivers were often cited for multiple violations on the same occasion—sometimes, but not always falling into the same classification category—which made it necessary to have the violation counts represent receiving one or more violations during a particular law enforcement traffic stop to maintain independence of counts. Finally, crashes and traffic violations were not adjusted for differences in driving exposure (mileage), which varies according to both driver age and years of driving experience (Forsyth et al., 1995; Peck, 1993; Williams, 2003; Williams, Lund, & Preusser, 1985). Drivers age 20 or older tend to drive more miles per year on average than those who are younger, and older teens tend to drive more miles per year than younger teens (U.S. Department of Transportation and Federal Highway Administration, 2009). Estimates of annual mileage by individual year of age for California drivers can be calculated from the

137

National Household Travel Survey, but it is not possible to differentiate novice from experienced drivers using these data. Hence, their use requires the assumption that all drivers (novice and experienced) of the same age drive the same number of miles. These annualized averages are inappropriate for interpolating monthly estimates given evidence that within any particular age, drivers who are licensed longer tend to drive more miles (Forsyth et al., 1995; Williams et al., 1985), although for young novices there is little evidence of a difference over the first 18 months of licensure (Klauer et al., 2011; Lee et al., 2011). While it was not possible to control for differences in exposure over time within each age group or for gross differences in exposure between age groups given limitations of the available data, the findings in this report are certainly influenced by and to some extent reflective of differences in driving exposure, so the results should therefore be interpreted with this limitation in mind. 5. Conclusion and implications While novice 16- and 17-year-olds' highest crash rates occur almost immediately after they are licensed, their peak traffic violation rates are delayed until around the time they turn age 18. Both pre-licensure crash rates and post-licensure crash peaks were more pronounced for some older age groups of novices than was the case for 16–17-year-olds. Extending learner permit holding periods for 16–17-year-old novices appears consistent with their actual behavior; requiring older novices— particularly those ages 18 to 20—to hold permits for minimum periods may reduce their initial crash rates. Acknowledgment This study was funded by the National Highway Traffic Safety Administration (CONTRACT NUMBER: DTNH22-08-H-00240) in cooperation with the National Safety Council and the California Department of Motor Vehicles. The opinions, findings, and conclusions expressed in this report are those of the authors and not necessarily those of the National Highway Traffic Safety Administration, the National Safety Council, or the State of California. References Anderson, J. R. (1993). Problem solving and learning. American Psychologist, 48, 35–44. Anderson, J. R., Fincham, J. M., & Douglass, S. (1999). Practice and retention: A unifying analysis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 25, 1120–1136. Braitman, K. A., Kirley, B. B., McCartt, A. T., & Chaudhary, N. K. (2008). Crashes of novice teenage drivers: Characteristics and contributing factors. Journal of Safety Research, 39, 47–54. Cammisa, M. X., Williams, A. F., & Leaf, W. A. (1999). Vehicles driven by teenagers in four states. Journal of Safety Research, 30, 25–30. Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132, 354–380. Curry, A. E., Hafetz, J., Kallan, M. J., Winston, F. K., & Durbin, D. R. (2011). Prevalence of teen driver errors leading to serious motor vehicle crashes. Accident Analysis and Prevention, 43, 1285–1290. Curry, A. E., McDonald, C. C., Kandadai, V., Sommers, M. S., & Winston, F. K. (2013). Comparison of adult and teen driver crash scenarios in a nationally representative sample of serious crashes. 57th Annual Proceedings of the Association for the Advancement of Automotive Medicine, Chicago, IL (September). Dempster, F. N. (1988). The spacing effect: A case study in the failure to apply the results of psychological research. American Psychologist, 43, 627–634. Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). The role of deliberate practice in the acquisition of expert performance. Psychological Review, 100(3), 363–406. Ferdun, G. S., Coppin, R. S., & Peck, R. C. (1965). The teen-aged driver: an evaluation of age, experience, driving exposure and driver training as they related to driving record (Report No. 21). Sacramento, CA: California Department of Motor Vehicles. Forsyth, E., Maycock, G., & Sexton, B. (1995). Cohort study of learner and novice drivers. Part 3: Accidents, offences and driving experience in the first three years of driving, Project Report PR111. Crowthorne: Transport Research Laboratory. Foss, R. D., Martell, C. A., Goodwin, A. H., & O'Brien, N. P. (2011). Measuring changes in teenage driver crashes during the early months of driving. Washington, DC: AAA Foundation for Traffic Safety. Gebers, M. A., & Peck, R. C. (2003). Using traffic conviction correlates to identify high accident-risk drivers. Accident Analysis and Prevention, 35, 903–912.

138


Gulliver, P., Begg, D., Brookland, R., Ameratung, S., & Langley, J. (2013). Learner driver experiences and crash risk as an unsupervised driver. Journal of Safety Research, 46, 41–46. Irwin, C. (1996). Adolescent development and driving. In H. Simpson (Ed.), New to the road: Reducing the risks for young motorists (pp. 51–70). Los Angeles, CA: University of California, Youth Enhancement Service. Klauer, S. E., Simons-Morton, B. G., Lee, S. E., Ouimet, M. C., Howard, E. H., & Dingus, T. A. (2011). Novice drivers' exposure to known risk factors during the first 18 months of licensure: The effect of vehicle ownership. Traffic Injury Prevention, 12(2), 159–168. Lee, S. E., Simons-Morton, B. G., Klauer, S. E., Ouimet, M. C., & Dingus, T. A. (2011). Naturalistic assessment of novice teenage crash experience. Accident Analysis and Prevention, 43, 1472–1479. Lewis-Evans, B. (2010). Crash involvement during the different phases of the New Zealand graduated driver licensing system (GDLS). Journal of Safety Research, 41, 359–365. Liang, K. Y., & Zeger, S. L. (1986). Longitudinal data analysis using generalized linear models. Biometrika, 73(1), 13–22. Masten, S. V., & Foss, R. (2010). Long-term effect of the North Carolina graduated driver licensing system on licensed driver crash incidence: A 5-year survival analysis. Accident Analysis and Prevention, 42, 1647–1652. Masten, S. V., Foss, R. D., & Marshall, S. W. (2011). Graduated driver licensing and fatal crashes involving 16- to 19-year-old drivers. Journal of the American Medical Association, 306(10), 1098–1103. Mayhew, D. R. (2007). Driver education and graduated licensing in North America: Past, present, and future. Journal of Safety Research, 38, 229–235. Mayhew, D. R., Simpson, H. M., & Pak, A. (2003). Changes in collision rates among novice drivers during the first months of driving. Accident Analysis and Prevention, 35, 683–691. McCartt, A. T. (2013). Rounding the next curve on the road toward reducing teen drivers' crash risk. Journal of Adolescent Health, 53, 35. McCartt, A. T., Mayhew, D. R., Braitman, K. A., Ferguson, S. A., & Simpson, H. M. (2009). Effects of age and experience on young driver crashes: A review of recent literature. Traffic Injury Prevention, 10, 209–219. McCartt, A. T., Shabanova, V. I., & Leaf, W. A. (2003). Driving experience, crashes, and traffic citations of teenage beginning drivers. Accident Analysis and Prevention, 35, 311–320. McCullagh, P., & Nelder, J. A. (1989). Generalized linear models (2nd Ed.). London, England: Chapman & Hall, 204–208 (254–255). McKnight, A. J., & McKnight, A. S. (2003). Young novice drivers: Careless or clueless? Accident Analysis and Prevention, 35, 921–925. Peck, R. C. (1993). The identification of multiple accident correlates in high risk drivers with specific emphasis on the role of ages, experience and prior traffic violation frequency. Alcohol, Drugs and Driving, 9(4), 145–166. Simons-Morton, B. G., Hartos, J., Leaf, W. A., & Preusser, D. F. (2006). Do recommended driving limits affect teen-reported traffic violations and crashes during the first 12 months of independent driving? Traffic Injury Prevention, 7, 238–247. Simons-Morton, B. G., & Ouimet, M. C. (2006). Parent involvement in novice teen driving: A review of the literature. Injury Prevention, 12(Suppl. I), i30–i37. Twisk, D. A. M., & Stacey, C. (2007). Trends in young driver risk and countermeasures in European countries. Journal of Safety Research, 38, 245–257. U.S. Bureau of the Census (2013). California QuickFacts from the US Census Bureau: 2012. Washington, DC: Bureau of the Census (http://quickfacts.census.gov/qfd/states/ 06000.html. Accessed August 14, 2013).

U.S. Department of Transportation, Federal Highway Administration (2009). National Household Travel Survey. URL: http://nhts.ornl.gov U.S. Federal Highway Administration (2012). Highway statistics series, Table DL-1C: Licensed drivers by sex and ratio to population-2011. Available at: http://www. fhwa.dot.gov/policyinformation/statistics/2011/ (Retrieved August 13, 2013) Underwood, G. (2013). On-road behavior of younger and older novices during the first six months of driving. Accident Analysis and Prevention, 58, 235–243. VicRoads (2008). Victoria's graduated licensing system evaluation. Available from: http://www.vicroads.vic.gov.au/Home/SafetyAndRules/SaferDrivers/ YoungAndNewDrivers/VictoriasNewGraduatedLicensingSystem.htm Waller, P. F. (2003). The genesis of GDL. Journal of Safety Research, 34, 17–23. Waller, P. F., Elliott, M. R., Shope, J. T., Raghunathan, T. E., & Little, R. J. A. (2001). Changes in young adult offense and crash patterns over time. Accident Analysis and Prevention, 33, 117–128. Williams, A. F. (2003). Teenage drivers: Patterns of risk. Journal of Safety Research, 34, 5–15. Williams, A. F., Chaudhary, N. K., Tefft, B. C., & Tison, J. (2010). Evaluation of New Jersey's graduated driver licensing program. Traffic Injury Prevention, 11, 1–7. Williams, A. F., Lund, A. K., & Preusser, D. F. (1985). Driving behavior of licensed and unlicensed teenagers. Journal of Public Health Policy, 6(3), 379–393. Williams, A. F., Preusser, D. F., Ferguson, S. A., & Ulmer, R. G. (1997). Analysis of the fatal crash involvements of 15-year-old drivers. Journal of Safety Research, 28, 49–54. Williams, A. F., Tefft, B. C., & Grabowski, J. G. (2012). Graduated driver licensing research, 2010–present. Journal of Safety Research, 43(3), 195–203. Eric A. Chapman is an applied traffic safety researcher in the Alcohol and Drugs Unit of the Research and Development Branch of the California Department of Motor Vehicles. He earned his MA in psychology from California State University, Sacramento. His recent work includes studies on how graduated driver licensing laws affect licensing and crashes of older teens. He is currently working on an evaluation of mandatory ignition interlock devices for DUI offenders in four California counties. Scott V. Masten is research manager of the Alcohol and Drugs Unit in the Research and Development Branch of the California Department of Motor Vehicles. He earned his Ph.D. in epidemiology from the University of North Carolina, Chapel Hill, where he worked as a traffic safety researcher at the University of North Carolina Highway Safety Research Center. His recent work includes studies on how graduated driver licensing laws affect licensing and crashes of older teens, increasing motorcycle licensure among improperly licensed owners, and an evaluation of mandatory ignition interlock devices for DUI offenders in four California counties. He also tirelessly pursues his lifelong goals of starting a band and finding preeminent burritos. Kelly K. Browning has been in leadership positions promoting awareness and education through youth programs over the past 15 years. She received her formal education from Minnesota State University Moorhead, the University of Central Florida, and her Ph.D. in Criminology from the University of South Florida. She joined the Impact Teen Drivers team as Executive Director in November 2008. Her current work with Impact Teen Drivers merges law enforcement, education, the insurance industry, transportation agencies, and other corporate and non-profit organizations to successfully promote education and awareness of the dangers of distracted driving and poor decision making.

Long-term changes in crash rates after introduction of a Graduated Driver Licensing decal provision.

Driver license renewal policies and fatal crash involvement rates of older drivers, United States, 1986-2011.

Driver education and fatal crash involvement of teenaged drivers.

The association of graduated driver licensing with miles driven and fatal crash rates per miles driven among adolescents.

Can a video-based hazard perception test used for driver licensing predict crash involvement?

Motorcycle licensure, ownership, and injury crash involvement.

Crash rates of quebec drivers with medical conditions.

Evaluating a novice driver and pre-driver road safety intervention.

A system for traffic violation detection.

Using naturalistic driving data to explore the association between traffic safety-related events and crash risk at driver level.

Novice drivers' individual trajectories of driver behavior over the first three years of driving.

Trends in older driver crash involvement rates and survivability in the United States: an update.

Pregnancy and risk of a traffic crash.

Higher crash and near-crash rates in teenaged drivers with lower cortisol response: an 18-month longitudinal, naturalistic study.

International aspects of driver licensing.

Pregnancy and risk of a traffic crash.

Conscientious personality and young drivers' crash risk.

Driver education and teen crashes and traffic violations in the first two years of driving in a graduated licensing system.

Association of Graduated Driver Licensing With Driver, Non-Driver, and Total Fatalities Among Adolescents.

Pregnancy and risk of a traffic crash.

A new access density definition and its correlation with crash rates by microscopic traffic simulation method.

Crash risk and aberrant driving behaviors among bus drivers: the role of personality and attitudes towards traffic safety.

Extending the value of police crash reports for traffic safety research: collecting supplemental data via surveys of drivers.

Graduated Driver Licensing: An international review.