- Email: [email protected]
Learner driver experiences and crash risk as an unsupervised driver
Journal of Safety Research 46 (2013) 41–46
Contents lists available at SciVerse ScienceDirect
Journal of Safety Research journal homepage: www.elsevier.com/locate/jsr
Learner driver experiences and crash risk as an unsupervised driver P. Gulliver a,⁎, D. Begg a, R. Brookland a, S. Ameratunga b, J. Langley a a b
Injury Prevention Research Unit, Dunedin School of Medicine, PO Box 56, Dunedin, New Zealand Auckland School of Population Health, 261 Morrin Rd., Tamaki Campus, Glen Innes, Auckland, New Zealand
a r t i c l e
i n f o
Article history: Received 8 June 2012 Received in revised form 26 February 2013 Accepted 26 March 2013 Available online 9 April 2013 Keywords: Novice drivers crash involvement unsupervised driving
a b s t r a c t Objective: The purpose of this study was to describe the driving experiences of learner licensed drivers and examine the association between these driving experiences, associated factors, and on-road car crash involvement during the unsupervised restricted license stage. Methods: Data were drawn from a cohort investigation of newly licensed drivers. Information on demographic characteristics, personality, and risk behaviors was collected at the baseline interview. At the first follow-up interview (restricted license stage) study members were asked details about their experiences as a learner licensed driver: professional driving lessons, supervised driving, unsupervised driving, and driving courses in which they participated. During the second follow-up interview (full license stage), data were collected on crash involvement and driving exposure during the restricted license stage. Regression analysis was used to determine independent relationships between learner license driving experience variables and crash involvement. Results: After adjusting for demographic, personality factors, and driving exposure at the restricted license stage, increased time spent on the learner license was associated with a reduced risk of crash involvement during the unsupervised restricted license stage. Conclusion: Results presented in this paper suggest that learner drivers in New Zealand should be encouraged to spend more time on their learner license to enable them to gain skills and experience to help reduce their crash risk when they are allowed to drive unsupervised. Impact on Industry: Compared with novice drivers who are on their learner license for the least amount of time, those who spend the most amount of time on their learner license have reduced risk of on-road crash involvement as an unsupervised driver. Learner drivers and their supervisors need to be aware of the length of time required for practice in order to reduce the risks of crash involvement when they are able to drive unsupervised (O'Brien et al., 2012). The recently introduced increase in the minimum driver licensing age in NZ, tougher restricted license stage driving test (aimed at encouraging 120 hours of supervised driving), and the Safe Teen driver campaign (NZ Transport Agency, 2012) are all strategies targeted at improving the safety of learner drivers. These strategies need to be evaluated to ensure they are achieving their goals. © 2013 National Safety Council and Elsevier Ltd. All rights reserved.
1. Introduction In 1987, New Zealand (NZ) was the first country in the world to introduce a comprehensive Graduated Driver Licensing System (GDLS). Since that time, various versions of GDLS have been introduced in many countries, including the United States, Canada, and Australia. A typical GDLS includes progression through three stages of licensing: learner, provisional, and graduation to a full license (Hedlund, Shults, & Compton, 2003). Progression through the NZ GDLS involves first passing a theory and eyesight test to gain a learner license. At this ⁎ Corresponding author at: Family Violence Clearinghouse, School of Social and Community Health, Tamaki Campus, 261 Morrin Rd, Glen Innes, Auckland, New Zealand. Tel.: + 64 9 923 9216; fax: + 64 3 479 8337. E-mail addresses: [email protected] (P. Gulliver), [email protected] (D. Begg), [email protected] (R. Brookland), [email protected] (S. Ameratunga), [email protected] (J. Langley).
stage, the newly licensed driver is able to drive while supervised by a driver who has had their full license for a minimum of two years. Progression to the restricted (provisional) stage requires a minimum of 6 months on a learner license and for the applicant to pass a practical driving test. The restricted license stage allows the driver to drive unsupervised, but not between the hours of 10 p.m. and 5 a.m., and they are not allowed to carry passengers, unless they are supervised. The restricted license applies for 18 months, but this can be reduced to 12 months by completing an approved course (defensive driving or a Street Talk course). Progression to a full privilege license, with no conditions, requires the driver to pass another practical driving test. Although the introduction of the GDLS has been associated with reduced crash rates, heightened crash risk continues for young, learner drivers in the early stages of licensure due mainly to inexperience and immaturity (Begg & Stephenson, 2003; Lewis-Evans, 2010). One characteristic of GDLS, implemented to counter inexperience, is the requirement for learner drivers to only drive while supervised by a
0022-4375/$ – see front matter © 2013 National Safety Council and Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jsr.2013.03.007
42
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
fully licensed person (Hedlund et al., 2003). From an analysis of fatal crash involvement of 15 year old drivers from 33 states in America, Williams, Preusser, Ferguson, and Ulmer (1997) showed that driving without a learner’s permit, or with a learner’s permit but unsupervised, was associated with single vehicle night time crashes in which the 15 year old was legally culpable. In contrast, evaluation of the introduction of a minimum 30 hours supervised driving for learner drivers in Minnesota showed no impact on already reducing fatal and serious injury rates amongst drivers aged 16 and 17 years (O'Brien, Foss, Goodwin, & Masten, 2012). However, the authors also indicated that only one third of parents could correctly identify the minimum number of supervised hours required. To encourage more supervised, on-road driving experience before driving unsupervised, in 1993, the Swedish road authorities reduced the age of learner licensure from 17.5 years to 16 years. Evaluation of this change showed that, compared to those who obtained the typical level of supervised learner driving experience (41–47 hours), those who chose the extended supervised driving option and obtained the most supervised driving experience (mean = 118 hours) had significantly reduced post-license crash risk (Gregersen, 1997). 120 hours has since been adopted by some states in Australia (e.g., VicRoads, 2012) as the gold standard for supervised driving experience, while states in America range from no minimum requirement to 65 hours (of which 10 must be at night and 5 must be in inclement weather; Insurance Institute for Highway Safety, 2012). In NZ, apart from the time limit of at least six months, there is no legal minimum requirement for supervised driving experience during the learner license stage (New Zealand Transport Agency, 2010). Various options exist in NZ for gaining relatively safe driving experience as a learner driver, including professional instruction, supervised driving with a person who has held a full license for a minimum of two years, and learner driver courses. One nation-wide learner driver course, aimed at improving the safety of NZ learner drivers, is ‘Practice,’ developed by the NZ Transport Agency (NZTA) and the Accident Compensation Corporation (ACC; Accident Compensation Corporation & Land Transport Safety Authority, 2003). ‘Practice’ is a computer- and print-based driving program for 15 1 to 19 year olds and their driving supervisors, with an interactive CD-ROM (or internet based) to help learner drivers identify potential hazards and manage risks (Accident Compensation Corporation, & Land Transport Safety Authority, 2003). In the current format, ‘Practice’ participants are encouraged to practice driving with a supervisor for 120 hours, in a range of conditions, before they sit the restricted license test. All newly licensed drivers aged 15–19 years are posted a flyer advertising the program and inviting their participation. Preliminary evidence from the NZTA and ACC indicates that from 2007 to 2010, registration rates for ‘Practice’ had increased from 14% to 28% of eligible 15–19 year olds. The increase was considered to be the result of the re-development and re-launching of the program (Joel, 2010). Although there has been no independent evaluation of ‘Practice,’ the results of an internal unpublished evaluation of the re-development suggested that young drivers registered with ‘Practice’ had fewer crashes and traffic offenses than drivers who did not register (Andrew Joel, NZTA, personal communication, June 2011). In a recent review of investigations conducted to understand the effectiveness of the GDLS, Williams, Teft, and Grabowski (2012) acknowledged the increased availability of naturalistic studies of learner drivers (e.g., Lajunen & Summala, 2003). These studies have the potential to provide information about what learner drivers and their parents are actually doing on the roads, although the number participating in these investigations is generally quite small (e.g., O'Brien et al., 2012). Williams and colleagues also highlighted preliminary investigations on the effect of delaying licensure for one year (e.g., from 15 to 16 years; McCartt, Teoh, Fields, Braitman, & Hellinga, 2010). In the McCartt 1 At the time this cohort was recruited the minimum licensing age was 15 years. In 2011, the New Zealand government increased the minimum licensing age to 16 years.
investigation, fatal crash rates reduced by 13% with a one year delay in licensure (McCartt et al., 2010). In addition to this growing body of work, there have been recent studies that have described the key components of the GDLS for reducing crash involvement. For example, Jacobsohn, Garcis-Espana, Durbin, Erkoboni, and Winston (2012) described practice driving amongst American teens whose parents were involved in their driving education, and related this to parent characteristics (confidence in teaching their teen to drive, provision of support to teen) and the policy environment of the state in which they resided, while Scott-Parker, Watson, King, and Hyde (2011) described car ownership and driving exposure of novice drivers and how these relate to risky driving behaviors. However, as one of the concluding remarks in their review, Williams et al. (2012) suggested that “more information is needed about what transpires during the learner period and the policies and practices that will best prepare beginners for independent driving.” The aim of this work is to describe the driving experiences of learner licensed drivers and examine the association between these driving experiences and on-road car crash involvement as a driver during the unsupervised restricted license stage after controlling for a wide range of potential confounding factors. 2. Materials and methods The NZ Drivers Study (NZDS) is a longitudinal investigation of a cohort of 3,992 newly licensed novice drivers. A comprehensive description of study methods have been previously published (Begg et al., 2009). The main recruitment of cohort members was through face-to-face interviews at driver licensing centers and licensing courses throughout NZ. Recruitment regions were selected to ensure coverage of geographical locations (North and South Islands, rural and urban) and ethnicities (Māori, the indigenous people of NZ, and non-Māori). The eligible population was all drivers who passed the learner license test at the venues where recruitment was being undertaken, who had not previously held a car driver’s license. At recruitment, background information was collected. When the learner licensed driver passed the restricted license test, the first follow-up interview was undertaken. This was a telephone administered questionnaire to obtain data about the quality and quantity of driving experiences as a learner licensed driver and the supervision received. When the full license test was passed, a second follow-up interview was conducted to determine the quality and quantity of driving experience as well as traffic crash outcomes such as traffic offenses and crash involvement as a restricted license driver. 2.1. Data collection The data presented in this paper have been drawn from the baseline, first, and second follow-up interviews. As highlighted above, the baseline interviews were conducted during cohort recruitment (n = 3,992, 1 Feb 2006 to 31 Jan 2008). Data were extracted from the first follow-up stage on 17 September, 2010. At this point 2,474 had passed the restricted license test and 88% of them had completed the follow-up interview. Data were extracted from the second follow-up interview on 30 May 2011. At this point 1,470 study members had passed their full license test (full license drivers, FLDs), 93% of whom completed the NZDS second follow-up interview. There were 1,379 study members who had participated in both the first and second follow-up interviews and are included in the current analysis. The remaining 91 did not do both interviews either because they refused one or the other of the interviews, or could not be traced at one of the stages. 2.2. Driving experience variables At the first follow-up interview, study members were asked details about professional driving lessons (lessons paid for with a
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
professional driving instructor), supervised driving, unsupervised driving (illegal driving at the learner license stage), and driving courses in which they participated while on their learner license. For each professional driving lesson, and supervised or unsupervised driving experience, study participants were asked for the following information: • Driving experience: The amount of driving experience was estimated by the interviewer “walking” the study participant through a series of questions. 1. They were asked how frequently they drove: at least daily, or weekly, or fortnightly or monthly, and how many times they drove in a typical day, week, fortnight, or month. This number was later multiplied by the number of days, weeks, fortnights, or months they had been on their learner license (which was calculated from the driver license registry file) to give a total number of trips. 2. Next they were asked how far they usually drove on a typical trip. The response options were b11kms, 11-20 km, 21-50 km, 51-100 km or >100kms. The distance driven variable was calculated by multiplying the total number of trips they had driven (from (1) above) by the approximate midpoint for each distance category (5 km, 15 km, 35 km, 75 km, 150 km). 3. Finally, they were asked about the duration (in minutes) of a typical drive. An estimate of the duration driven (minutes) was calculated by multiplying the total number of trips they had driven (from (1) above) by the number of minutes per typical drive. An overall measure of total distance driven (kms) and total duration driven (minutes) was calculated by combining the distances and durations estimated for professional driving lessons, all supervised driving trips and unsupervised driving trips. Acknowledging the imprecision of the driving estimates, all of the driving exposure measures were categorized according to low (b25th percentile), medium (25-75th percentile), and high (>75th percentile) for analysis. In addition, a zero category was provided for those who did not drive unsupervised or who did not have professional driving lessons. • Driving conditions: To obtain an indication of the variety of conditions in which the learner driver had driven, the cohort members were asked “how many times did you drive on a public road (never, once, 2–5 times, 6–10 times, >10 times) during”: (1) conditions that made it difficult to see; (2) on icy, (3) wet or (4) gravel/unsealed roads; (5) in heavy city traffic; in a 100 km/h area (6) during daylight or (7) when it was dark; in a 50 km/h area (8) during daylight or (9) when it was dark; (10) on any long trips (greater than 50 km in distance). For the purposes of the current investigation, a score of ‘1’ was given for each of the 10 driving conditions described above that were experienced two or more times while on the learner license and ‘0’ for those conditions experienced fewer than two times. A summary score of 0 for this variable indicated that none of the conditions were experienced two or more times, while a score of 10 indicated that all of the conditions were experienced at least twice. 2.3. Confounding variables Potential confounding variables included in the analysis were identified from the literature. Age at learner licensure, gender, and residential location (rural vs. urban) were determined at baseline interview (Begg et al., 2009). Alcohol use was measured using the first three questions of the AUDIT (the Audit-C) (Barbor, de la Fuente, Saunders, et al., 1989) and categorized as established by alcohol researchers (Bradley et al., 2007; a cut-off of ≥4 was used to indicate high level alcohol use for males and ≥3 for females). The Zuckerman-Kuhlman Personality Questionnaire scales for Impulsivity-Sensation Seeking (ImpSS) and Aggression-Hostility (Agg-Host; Zuckerman, Kulman, Teta, Joireman, &
43
Kraft, 1993) were included to obtain a measure of impulsivity (8 items), sensation seeking (11 items), and aggression-hostility (17 items). Gender specific cut-off points were selected to classify each of the scales into three categories: low (lowest quartile), medium (middle quartiles), and high (highest quartile). In the second follow-up interview, cohort members were asked to report the duration and distance driven while on their restricted license. These data were collected using the same format as in the first follow-up interview described above. Restricted license driving exposure (distance driven) was included as a confounding variable. 2.4. Outcome The main outcome variable was whether or not a driver was in a car crash on a public road during the restricted license stage. Motor vehicle traffic crashes (those in which the cohort member was driving a car or similar vehicle) were obtained from the traffic crash reports (TCR) as recorded by the NZ Police. In NZ when a road traffic crash results in someone being injured it is required by law to be reported to the Police who complete a TCR for that crash (Ministry of Transport, 2011). Despite the legal requirement to report crashes, hospital based surveys have shown that only around two-thirds of injury crashes requiring admission to hospital are reported to, or recorded, by the Police. Furthermore there are biases in the type of crashes reported with crashes involving younger drivers being less likely to be reported than those involving older drivers (Alsop & Langley, 1998). Therefore, to complement the TCR crashes, and to provide a more complete picture of crash involvement, we included self-reported crashes. Data on self-reported crashes were collected at the second follow-up interview and included crashes that occurred on-road where someone was injured and/or there was vehicle or property damage. The combined crash outcome, with duplicates removed, was dichotomized (yes/no). Only crashes that occurred while cohort members were on their restricted license (i.e., when unsupervised driving was first allowed) were included in the analysis. 2.5. Analysis All data analysis was conducted in Stata12SE. The GLM procedure with Poisson distribution was used to estimate the relative risks of learner license variables with crash involvement. To adjust for variable length of time while on the restricted license stage, the number of days on a restricted license was included in the model. 3. Results Of the 1,379 FLDs, 339 (24.6%) reported a car crash on a public road where there was an injury or property damage while they were on their restricted license. Driving experience variables and their association with restricted license crash involvement are displayed in Table 1. Less than 10% of FLDs had taken part in a driving course when they were learner licensed drivers. Over 50% of FLDs did not have any professional driving instruction, and almost 20% engaged in unsupervised driving as a learner driver. Table 1 shows that the majority of driving experience for the FLDs was gained through informal supervised driving. The FLDs reported a high number of driving conditions experienced through supervised driving (>85% experienced at least 5 of the 10 listed driving conditions on at least two occasions). Almost 25% of FLDs reported that they drove for less than 27 hours while on their learner license, with the same proportion reporting that they drove for more than 117 hours. Almost 29% of the FLDs who were involved in a crash were on their learner license for 180–219 days, while almost 27% of those who were not involved in a crash were on their learner license for more than 438 days. Compared with those who did not drive unsupervised, relative risks for crash involvement were elevated for all of the unsupervised
44
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
Table 1 Learner licence driving exposure and experience and relationship with restricted licence crash involvement. Restricted licence crash No n Took part in driving course No 938 Yes 100
UL
%
75.52 74.07
304 35
24.48 25.93
1 1.06
0.78
1.43
26.37 22.91 20.75
1 0.87 0.79
0.72 0.46
1.05 1.35
27.12 19.74 24.29 26.11
1 0.73 0.9 0.96
0.56 0.71 0.73
0.95 1.13 1.26
27.03 23.72 22.77 23.56
1 0.88 0.84 0.87
0.66 0.68 0.64
1.16 1.05 1.18
19.89 25.21
1 1.25
0.92
1.71
23.30 22.78 29.04
1 0.98 1.24
0.77 0.97
1.24 1.61
25.00 21.99 29.46
0.88 1.18
0.7 0.93
1.1 1.5
23.39 30.57 30.00
1 1.31 1.28
1.01 0.9
1.69 1.82
23.03 29.81 32.33 26.32
1 1.29 1.4 1.14
0.94 1.07 0.77
1.77 1.84 1.69
23.03 30.77 31.29 26.67
1 1.33 1.36 1.16
0.96 1.04 0.78
1.85 1.77 1.71
UNSUPERVISED DRIVING Number of different driving experiences1 0 experiences 871 76.61 266 1-4 experiences 109 69.43 48 5-10 experiences 56 70.00 24 Unsupervised distance driven - LL 0 km 819 76.97 245 1-10 km 73 70.19 31 11-240 km 90 67.67 43 >240 km 56 73.68 20 Unsupervised duration driven - LL 0 hrs 819 76.97 245 1 min-9 mins 63 69.23 28 10 mins - 8 hrs 101 68.71 46 >8 hrs 55 73.33 20
1
LL
n
SUPERVISED DRIVING Number of different driving experiences1 1-4 experiences 145 80.11 36 5-10 experiences 887 74.79 299 Distance driven - LL 0-555 km 260 76.70 79 556-2925 km 522 77.22 154 >2925 km 237 70.96 97 Duration driven - LL 0-24.6 hrs 276 75.00 92 24.7-108.5 hrs 525 78.01 148 >108.5 hrs 237 70.54 99
206 536 242
RR
%
PROFESSIONAL LESSONS Number of different driving experiences1 0 experiences 525 73.63 188 1-4 experiences 471 77.09 140 5-10 experiences 42 79.25 11 Distance driven - LL 0 km 379 72.88 141 1-30 km 244 80.26 60 31-120 km 265 75.71 85 >120 km 150 73.89 53 Duration driven - LL 0 hrs 378 72.97 140 1 min-1.3 hrs 164 76.28 51 1.4-6 hrs 363 77.23 107 >6 hrs 133 76.44 41
TOTAL DRIVING Driving duration - LL b27 hrs 27-117 hrs >117 hrs Distance driven - LL b616 km 616-3075 km >3075 km Days on learner licence 180-219 days 220-290 days 291-437 days >438 days
95% CI
Yes
69.83 77.79 71.39
89 153 97
30.17 22.21 28.61
1 0.87 1.12
269 528 241
77.08 76.97 70.47
80 158 101
22.92 23.03 29.53
1 1 1.29
241 265 254 278
71.30 76.37 73.62 80.12
97 82 91 69
28.70 23.63 26.38 19.88
1 0.82 0.92 0.69
driving variables listed in Table 1. For example, compared with those who experienced no driving conditions as an unsupervised driver (or did not drive more than once in any of the conditions), those who experienced 1–4 conditions were 1.3 times more likely to be involved in a crash. Compared with those who did not drive unsupervised, those who drove 11–240 km unsupervised were 1.4 times more likely to be involved in a crash. There was also a relationship between total (supervised, unsupervised, and professional) distance driven on the learner license stage and crash involvement. In comparison with those whose total driving distance was less than 616 km during the learner license stage, those who drove >3,075 km were 1.3 times more likely to be involved in a crash. For those who were on their learner license for the least amount of time (180–219 days), there was a heightened risk of crash involvement compared with those who were on their learner license for the longest period (>438 days, Table 1). To identify learner license variables that were independently associated with crash involvement a fully specified regression model, including all described explanatory and confounding variables was initially attempted, however this model would not converge. We also investigated models that contained (1) only those variables describing the duration of driving experience or (2) the distance driven while on a learners, but again experienced problems with convergence. Finally, a more restricted model that included only those variables that showed an elevated relative risk (95% confidence interval for univariate RR not including 1) of crash involvement were included in the multivariate analysis. There was a high degree of multi-collinearity between the unsupervised driving variables, therefore a new variable was created that identified whether FLDs had driven unsupervised as a learner or not (Table 2). It is evident from the model that, after adjusting for age, gender, residential location, high level alcohol consumption, total distance driven during the restricted license stage, sensation seeking, impulsivity, aggression, and number of days on restricted license, compared with those who spent the least amount of time on their learner license, there was a reduced risk of crash involvement for those in the upper quartile of the length of time spent on a learner license. Those who drove unsupervised at the learner license stage had increased risk of crash involvement during the restricted license stage, although the confidence interval for the relative risk included 1. 4. Discussion The results presented indicate that, after adjusting for demographic, personality factors, and driving exposure, those who were in the upper quartile (greater than approximately 14 months) of the length of time they spent on their learner license were less likely to be involved in a crash during the first period of unsupervised driving (the restricted license stage) when compared with those who spent the least amount Table 2 Learner licence driving exposure and experience and relationship with restricted licence crash involvement – relative risk of traffic crash involvement as a restricted licence driver for selected learner driver experiences.⁎
0.69 0.88
0.79 1
0.64 0.72 0.53
1.09 1.43
1.27 1.66
1.06 1.17 0.91
Number of driving conditions experienced at least twice over the learner licence stage. Options were: (1) conditions that made it difficult to see; (2) on icy, (3) wet or (4) gravel/unsealed roads; (5) in heavy city traffic; in a 100 kph area (6) during daylight or (7) when it was dark; in a 50 kph area (8) during daylight or (9) when it was dark; (10) on any long trips (greater than 50 km in length).
RR
95% CI
Variable
LL
UL
z
p
Days on learner licence 0-219 days 1 220-290 days 0.77 291-437 days 0.94 >438 days 0.63
0.54 0.65 0.43
1.11 1.34 0.91
−1.38 −0.36 −2.42
0.17 0.72 0.02
Unsupervised driving No 1 Yes 1.29
0.95
1.76
1.66
0.20
⁎ Adjusted for age, gender, residential location, sensation seeking, impulsivity, aggression, hazardous alcohol consumption, distance driven on restricted licence, days on restricted licence.
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
of time on their learner license. In addition, those who had driven unsupervised as a learner driver were at heightened risk of crash involvement in the restricted license stage. The majority of experience gained by the NZDS cohort occurred in brief, informal supervised driving, with 65% reporting that the average duration of supervised driving episodes were 30 minutes or less (Gulliver, Begg, & Brookland, 2011). This being the case, if each cohort member spent 15 minutes driving each day for the duration of their learner license, and were not permitted to move on to the restricted license until they had acquired 120 hours of experience, 480 days would be required before this would occur. An alternative interpretation of the findings presented is that those who are less confident and who therefore spend longer on their learner license before opting to progress to their restricted license, may be more cautious drivers. Although we attempted to control for personality types by including sensation seeking, aggression, and hostility in the regression model, it is possible that the findings are a spurious association. However, our findings do support those of other researchers that more experience gained through either more months on the learner license or more hours driving, are associated with a reduced risk of crash involvement. It is of interest that the number of driving conditions experienced either through professional lessons, supervised driving, or unsupervised driving were not associated with crash involvement. In this investigation, there was no attempt to consider some types of driving conditions as more important than others. It is possible that some driving conditions would be more strongly related to crash involvement than others. For example, those cohort members who resided in the warmer parts of NZ would be less likely to be exposed to icy roads, reducing the chance of experiencing this driving condition as well as reducing the saliency of this driving condition for reducing their crash risk. However, further work is required to determine the relative importance of different types of experiences. In contrast with other investigations (Gregersen, 1997), the reported duration (number of hours) of driving exposure during the learner license stage was not related to crash involvement. To estimate the number of hours of driving exposure during the learner license stage, cohort members were stepped through a series of questions that have been shown to be a reliable method for obtaining travel exposure data (Lajunen & Ozkan, 2011). Also, the number of days, weeks, fortnights, or months they spent driving on a particular license was extracted from the driver licensing registry, and therefore not dependent on self-report. While this method of estimating driving exposure may introduce measurement error, it is likely that the measurement error was smaller than traditional methods of estimating exposure – by asking one single question concerning the number of hours being driven at the end of the license period. Two Australian studies have used travel journals to record duration of driving exposure with interesting results (Bates, Watson, & King, 2010; Scott-Parker, Bates, Watson, King, & Hyde, 2011). In a retrospective, cross-sectional study, Bates et al. (2010) compared the amount of driving experience reported by learner drivers in two Australian jurisdictions. Those from New South Wales (NSW) were legally required to complete 50 hours of supervised driving practice before attempting a practical driving test, while those from Queensland had no minimum level of driving practice required. Participants reported the amount of private, supervised driving and driving instruction at the end of the learner license period. Perhaps unsurprisingly, 99% of participants from NSW reported that they completed the 50 hours required experience, with the mean duration being 73.3 hours. In contrast, Queensland study members tended to report that they either participated in 26–50 hours or more than 100 hours supervised driving (mean 64 hours; Bates et al., 2010). We consider that the use of a stepwise method to estimate the distance and duration of driving exposure will produce more reliable estimates than a solitary question, or the use of travel diaries that would require a high level of motivation to complete regularly.
45
Less than 10% of the FLDs reported that they took part in a driving course. Although there are a number of driving courses available in NZ, of those who took part in a course, the majority took part in Practice (Gulliver et al., 2011). Despite the fact that all learner drivers aged 15–19 years are sent an invitation to participate in “Practice,” a free driving education program provided by the NZ government, only 8% of the cohort who participated in the first follow-up interview reported that they did so, while over 40% reported that they didn’t know a course was available (Gulliver et al., 2011). Contrary to preliminary evidence produced by NZTA and ACC, participation in a driving course, such as Practice, while on the learner license was not associated with a reduced risk of crash involvement (Table 1). Because of the nature of the investigation, to determine learner license experiences that were associated with crash involvement in the restricted license stage, only those cohort members who had progressed through to their full license were included. A previous investigation has shown that cohort members who did not progress from the learner license to restricted license stage were more likely to be female, tertiary students, older, and from more deprived areas (Langley et al., 2012). Non-progressors also had reduced risks of traffic offending, and the primary reasons for non-progression were “too lazy,” “busy,” or had “limited access to a car” (Langley et al.). In this investigation, compared with the remainder of the cohort, the FLDs were younger when they obtained their license, were predominantly non-Maori, males, and from less deprived areas. The FLDs were also exposed to a larger variety of driving conditions and drove more supervised distance. This suggests that the results presented are applicable to those who are motivated, and have the resources to gain the experience and supervision required to progress through the graduated license system. However, the Langley et al. investigation supports the results presented in this study (that those who were on their learner license for longer periods of time had reduced risks of crash involvement) and, as such, it is likely that the results presented underestimate the association of reduced crash involvement for those who were on their learner license for longer periods of time. 5. Conclusion Results presented in this paper suggest that learner drivers in NZ should be encouraged to spend more time on their learner license to enable them to gain skills and experience to help reduce their crash risk when they are allowed to drive unsupervised. Acknowledgments The New Zealand Drivers Study is funded by the Health Research Council of New Zealand (2005–2013), the Road Safety Trust (2005– 2013) and the Accident Compensation Corporation (2005–2009). No funding source had any involvement in the preparation of this paper. The authors wish to acknowledge Dr Rebbecca Lilley, Injury Prevention Research Unit, University of Otago, for comments on an earlier version of this paper. In addition, the authors wish to acknowledge the helpful comments of the reviewer on earlier drafts of this manuscript. References Accident Compensation Corporation, & Land Transport Safety Authority (2003). Your practice guide. Practice: The plan for learning to drive safely. Wellington, New Zealand: ACC, LTSA. Alsop, J. C., & Langley, J. D. (1998). A determination of biases in LTSA's Traffic Crash Report files with respect to serious occupant road crashes: a data linkage study. Dunedin: Injury Prevention Research Unit, University of Otago. Barbor, T. F., de la Fuente, J. R., Saunders, J., & Monteiro, M. G. (1989). The alcohol use disorders identification test: Guidelines for the use in primary health care. Geneva: World Health Organisation. Bates, L., Watson, B., & King, M. (2010). Required hours of practice for learner drivers: A comparison between two Australian jurisdictions. Journal of Safety Research, 41, 93–97.
46
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
Begg, D., Langley, J., Broughton, J., Brookland, R., Ameratunga, S., & McDowell, A. (2009). New Zealand Drivers Study: a follow-up study of newly licensed drivers. Injury Prevention, 15(4), e2–e11, http://dx.doi.org/10.1136/ip.2009.021998. Begg, D., & Stephenson, S. (2003). Graduated driver licensing: the New Zealand experience. Journal of Safety Research, 34(1), 99–105. Bradley, K. A., DeBenedetti, A. F., Volk, R. J., Williams, E. C., Frank, D., & Kivlahan, D. R. (2007). AUDIT-C as a brief screen for alcohol misuse in primary care. Alcoholism, Clinical and Experimental Research, 31(7), 1208–1217. Gregersen, N. P. (1997). Evaluation of 16-years age limit for driver training. First report. Linkoping, Sweden: VTI (Swedish National Road and Transport Research Institute). Gulliver, P., Begg, D., & Brookland, R. (November 2–4). Newly licensed drivers: How are they learning? Paper presented at the 10th National Conference on Injury Prevention and Safety Promotion. Brisbane Convention and Exhibition Centre. Hedlund, J., Shults, R. A., & Compton, R. (2003). What we know, what we don't know, and what we need to know about graduated driver licensing. Journal of Safety Research, 34, 107–115, http://dx.doi.org/10.1016/S0022-4375(02)00086-5. Insurance Institute for Highway Safety (2012). Young driver licensing systems in the U.S. Retrieved from. http://www.iihs.org/laws/gdl_learners.aspx. Jacobsohn, L., Garcis-Espana, J. F., Durbin, D. R., Erkoboni, D., & Winston, F. K. (2012). Adult supervised practice for adolescent learners: The current state and directions for interventions. Journal of Safety Research, 43, 21–28. Joel, A. (2010). Practice - selling safety to teen drivers. Paper presented at the Australasian Road Safety Research, Policing and Education Conference, Canberra, ACT. Lajunen, T., & Ozkan, T. (2011). Self-reported instruments and methods. In B. Porter (Ed.), Handbook of Traffic Psychology. Waltham, MA: Academic Press. Lajunen, T., & Summala, H. (2003). Can we trust self-reports of driving? Effects of impression management on driver behavior questionnaire responses. Transportation Research Part F: Traffic Psychology and Behavior, 6(2), 97–107. Langley, J., Begg, D., Brookland, R., Samaranayaka, A., Davie, G., & Jordon, H. (2012). Non-progression through graduated driver licensing: characteristics, traffic offending, and reasons for nonprogression. Traffic Injury Prevention, 13(1), 7–13. 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. McCartt, A. T., Teoh, E. R., Fields, M., Braitman, K. A., & Hellinga, L. A. (2010). Graduated licensing laws and fatal crashes of teenage drivers: a national study. Traffic Injury Prevention, 11, 240–248. Ministry of Transport (2011). Motor Vehicle Crashes in New Zealand. Retrieved from. http://www.transport.govt.nz/research/motorvehiclecrashesinnewzealand/. New Zealand Transport Agency (2010). The Official New Zealand Road Code. Retrieved from. http://www.nzta.govt.nz/resources/roadcode/road-code-index.html. NZ Transport Agency (2012). Safer journies for teen drivers: A toolkit for parents of teen drivers. Retrieved from. http://www.safeteendriver.co.nz/. O'Brien, N. P., Foss, R. D., Goodwin, A. H., & Masten, S. V. (2012). Supervised hours requirements in graduated driver licensing: Effectiveness and parental awareness. Accident Analysis and Prevention, http://dx.doi.org/10.1016/j.aap.2012.05.007. Scott-Parker, B. J., Bates, L., Watson, E. C., King, M. J., & Hyde, M. K. (2011). The impact of changes to the graduated driver licensing program in Queensland, Australia on the experiences of learner drivers. Accident Analysis and Prevention, 43, 1301–1308. Scott-Parker, B., Watson, B., King, M. J., & Hyde, M. K. (2011). Mileage, car ownership, experience of punishment avoidance, and the risky driving of young drivers. Traffic Injury Prevention, 12(6), 559–567.
VicRoads (2012). How to get your probationary license. Retrieved from. http://www. vicroads.vic.gov.au/Home/Licenses/GetYourPs/Howtogetyourprobationarycarlicense. htm. 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(1), 49–54. Williams, A. F., Teft, B. C., & Grabowski, J. G. (2012). Graduated driver licensing research, 2010-present. Journal of Safety Research, 43, 195–203. Zuckerman, M., Kulman, D. M., Teta, P., Joireman, J., & Kraft, M. (1993). A comparison of three structural models for personality: the big three, the big five and the alternative five. Journal of Personality and Social Psychology, 65, 757–768.
Pauline Gulliver has been employed as a Research Fellow at the Injury Prevention Research since 2001. She completed her PhD in Human Nutrition at Otago University in 1998 and has worked for the Cancer Society of New Zealand and Kings College London before returning to Otago University. She has worked on epidemiological investigations of childhood injuries in the home, the surveillance and description of sports injuries, and predictors of traffic crash involvement. Dorothy Begg is an injury epidemiologist and co-ordinates the IPRU road safety research. Her undergraduate training was in psychology and postgraduate training was in Public Health (epidemiology). Her principal research interest is young drivers, and in particular the graduated driver licensing system and its impact on young driver behaviour. Current and recent research focuses on young newly licensed drivers and the graduated driver licensing system, and motorcycling injury research. Rebecca Brookland is a Research Fellow with the Injury Prevention Research Unit, and has a Master of Arts, in Psychology from the University of Otago. Currently her research focuses on young and newly licensed drivers. For her PhD Rebecca is investigating parental influence on the driving experiences of young drivers. She is also Project Manager for the New Zealand Drivers Study, a large follow-up study of nearly 4000 newly licensed drivers. Shanthi Ameratunga is paediatrician and public health physician by background, Shanthi leads a multi-disciplinary research program focusing on trauma outcomes, injury prevention, disability and rehabilitation. She is the Project Director of the Traffic Related Injury in the Pacific (TRIP) Study, a collaboration with the Fiji School of Medicine funded by The Wellcome Trust and the Health Research Council of New Zealand. Other projects she contributes to include the national youth health surveys (Youth2000/Youth’07), the PREDICT project (cardiovascular disease prevention), and the ADAPT trial (mobile phone intervention for depression prevention in young people). John Langley has been involved in injury prevention research since 1978, and received his PhD from the University of Otago in 1985. He has written numerous articles in academic, professional and lay publications on a wide range of injury prevention issues. He tends to use quantitative research methods, especially epidemiological. He is an active advocate for improvements in injury prevention policy and practice. John’s current areas of research interest include: young drivers, outcomes of injury, alcoholrelated injury, and injury outcome indicators.
Contents lists available at SciVerse ScienceDirect
Journal of Safety Research journal homepage: www.elsevier.com/locate/jsr
Learner driver experiences and crash risk as an unsupervised driver P. Gulliver a,⁎, D. Begg a, R. Brookland a, S. Ameratunga b, J. Langley a a b
Injury Prevention Research Unit, Dunedin School of Medicine, PO Box 56, Dunedin, New Zealand Auckland School of Population Health, 261 Morrin Rd., Tamaki Campus, Glen Innes, Auckland, New Zealand
a r t i c l e
i n f o
Article history: Received 8 June 2012 Received in revised form 26 February 2013 Accepted 26 March 2013 Available online 9 April 2013 Keywords: Novice drivers crash involvement unsupervised driving
a b s t r a c t Objective: The purpose of this study was to describe the driving experiences of learner licensed drivers and examine the association between these driving experiences, associated factors, and on-road car crash involvement during the unsupervised restricted license stage. Methods: Data were drawn from a cohort investigation of newly licensed drivers. Information on demographic characteristics, personality, and risk behaviors was collected at the baseline interview. At the first follow-up interview (restricted license stage) study members were asked details about their experiences as a learner licensed driver: professional driving lessons, supervised driving, unsupervised driving, and driving courses in which they participated. During the second follow-up interview (full license stage), data were collected on crash involvement and driving exposure during the restricted license stage. Regression analysis was used to determine independent relationships between learner license driving experience variables and crash involvement. Results: After adjusting for demographic, personality factors, and driving exposure at the restricted license stage, increased time spent on the learner license was associated with a reduced risk of crash involvement during the unsupervised restricted license stage. Conclusion: Results presented in this paper suggest that learner drivers in New Zealand should be encouraged to spend more time on their learner license to enable them to gain skills and experience to help reduce their crash risk when they are allowed to drive unsupervised. Impact on Industry: Compared with novice drivers who are on their learner license for the least amount of time, those who spend the most amount of time on their learner license have reduced risk of on-road crash involvement as an unsupervised driver. Learner drivers and their supervisors need to be aware of the length of time required for practice in order to reduce the risks of crash involvement when they are able to drive unsupervised (O'Brien et al., 2012). The recently introduced increase in the minimum driver licensing age in NZ, tougher restricted license stage driving test (aimed at encouraging 120 hours of supervised driving), and the Safe Teen driver campaign (NZ Transport Agency, 2012) are all strategies targeted at improving the safety of learner drivers. These strategies need to be evaluated to ensure they are achieving their goals. © 2013 National Safety Council and Elsevier Ltd. All rights reserved.
1. Introduction In 1987, New Zealand (NZ) was the first country in the world to introduce a comprehensive Graduated Driver Licensing System (GDLS). Since that time, various versions of GDLS have been introduced in many countries, including the United States, Canada, and Australia. A typical GDLS includes progression through three stages of licensing: learner, provisional, and graduation to a full license (Hedlund, Shults, & Compton, 2003). Progression through the NZ GDLS involves first passing a theory and eyesight test to gain a learner license. At this ⁎ Corresponding author at: Family Violence Clearinghouse, School of Social and Community Health, Tamaki Campus, 261 Morrin Rd, Glen Innes, Auckland, New Zealand. Tel.: + 64 9 923 9216; fax: + 64 3 479 8337. E-mail addresses: [email protected] (P. Gulliver), [email protected] (D. Begg), [email protected] (R. Brookland), [email protected] (S. Ameratunga), [email protected] (J. Langley).
stage, the newly licensed driver is able to drive while supervised by a driver who has had their full license for a minimum of two years. Progression to the restricted (provisional) stage requires a minimum of 6 months on a learner license and for the applicant to pass a practical driving test. The restricted license stage allows the driver to drive unsupervised, but not between the hours of 10 p.m. and 5 a.m., and they are not allowed to carry passengers, unless they are supervised. The restricted license applies for 18 months, but this can be reduced to 12 months by completing an approved course (defensive driving or a Street Talk course). Progression to a full privilege license, with no conditions, requires the driver to pass another practical driving test. Although the introduction of the GDLS has been associated with reduced crash rates, heightened crash risk continues for young, learner drivers in the early stages of licensure due mainly to inexperience and immaturity (Begg & Stephenson, 2003; Lewis-Evans, 2010). One characteristic of GDLS, implemented to counter inexperience, is the requirement for learner drivers to only drive while supervised by a
0022-4375/$ – see front matter © 2013 National Safety Council and Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jsr.2013.03.007
42
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
fully licensed person (Hedlund et al., 2003). From an analysis of fatal crash involvement of 15 year old drivers from 33 states in America, Williams, Preusser, Ferguson, and Ulmer (1997) showed that driving without a learner’s permit, or with a learner’s permit but unsupervised, was associated with single vehicle night time crashes in which the 15 year old was legally culpable. In contrast, evaluation of the introduction of a minimum 30 hours supervised driving for learner drivers in Minnesota showed no impact on already reducing fatal and serious injury rates amongst drivers aged 16 and 17 years (O'Brien, Foss, Goodwin, & Masten, 2012). However, the authors also indicated that only one third of parents could correctly identify the minimum number of supervised hours required. To encourage more supervised, on-road driving experience before driving unsupervised, in 1993, the Swedish road authorities reduced the age of learner licensure from 17.5 years to 16 years. Evaluation of this change showed that, compared to those who obtained the typical level of supervised learner driving experience (41–47 hours), those who chose the extended supervised driving option and obtained the most supervised driving experience (mean = 118 hours) had significantly reduced post-license crash risk (Gregersen, 1997). 120 hours has since been adopted by some states in Australia (e.g., VicRoads, 2012) as the gold standard for supervised driving experience, while states in America range from no minimum requirement to 65 hours (of which 10 must be at night and 5 must be in inclement weather; Insurance Institute for Highway Safety, 2012). In NZ, apart from the time limit of at least six months, there is no legal minimum requirement for supervised driving experience during the learner license stage (New Zealand Transport Agency, 2010). Various options exist in NZ for gaining relatively safe driving experience as a learner driver, including professional instruction, supervised driving with a person who has held a full license for a minimum of two years, and learner driver courses. One nation-wide learner driver course, aimed at improving the safety of NZ learner drivers, is ‘Practice,’ developed by the NZ Transport Agency (NZTA) and the Accident Compensation Corporation (ACC; Accident Compensation Corporation & Land Transport Safety Authority, 2003). ‘Practice’ is a computer- and print-based driving program for 15 1 to 19 year olds and their driving supervisors, with an interactive CD-ROM (or internet based) to help learner drivers identify potential hazards and manage risks (Accident Compensation Corporation, & Land Transport Safety Authority, 2003). In the current format, ‘Practice’ participants are encouraged to practice driving with a supervisor for 120 hours, in a range of conditions, before they sit the restricted license test. All newly licensed drivers aged 15–19 years are posted a flyer advertising the program and inviting their participation. Preliminary evidence from the NZTA and ACC indicates that from 2007 to 2010, registration rates for ‘Practice’ had increased from 14% to 28% of eligible 15–19 year olds. The increase was considered to be the result of the re-development and re-launching of the program (Joel, 2010). Although there has been no independent evaluation of ‘Practice,’ the results of an internal unpublished evaluation of the re-development suggested that young drivers registered with ‘Practice’ had fewer crashes and traffic offenses than drivers who did not register (Andrew Joel, NZTA, personal communication, June 2011). In a recent review of investigations conducted to understand the effectiveness of the GDLS, Williams, Teft, and Grabowski (2012) acknowledged the increased availability of naturalistic studies of learner drivers (e.g., Lajunen & Summala, 2003). These studies have the potential to provide information about what learner drivers and their parents are actually doing on the roads, although the number participating in these investigations is generally quite small (e.g., O'Brien et al., 2012). Williams and colleagues also highlighted preliminary investigations on the effect of delaying licensure for one year (e.g., from 15 to 16 years; McCartt, Teoh, Fields, Braitman, & Hellinga, 2010). In the McCartt 1 At the time this cohort was recruited the minimum licensing age was 15 years. In 2011, the New Zealand government increased the minimum licensing age to 16 years.
investigation, fatal crash rates reduced by 13% with a one year delay in licensure (McCartt et al., 2010). In addition to this growing body of work, there have been recent studies that have described the key components of the GDLS for reducing crash involvement. For example, Jacobsohn, Garcis-Espana, Durbin, Erkoboni, and Winston (2012) described practice driving amongst American teens whose parents were involved in their driving education, and related this to parent characteristics (confidence in teaching their teen to drive, provision of support to teen) and the policy environment of the state in which they resided, while Scott-Parker, Watson, King, and Hyde (2011) described car ownership and driving exposure of novice drivers and how these relate to risky driving behaviors. However, as one of the concluding remarks in their review, Williams et al. (2012) suggested that “more information is needed about what transpires during the learner period and the policies and practices that will best prepare beginners for independent driving.” The aim of this work is to describe the driving experiences of learner licensed drivers and examine the association between these driving experiences and on-road car crash involvement as a driver during the unsupervised restricted license stage after controlling for a wide range of potential confounding factors. 2. Materials and methods The NZ Drivers Study (NZDS) is a longitudinal investigation of a cohort of 3,992 newly licensed novice drivers. A comprehensive description of study methods have been previously published (Begg et al., 2009). The main recruitment of cohort members was through face-to-face interviews at driver licensing centers and licensing courses throughout NZ. Recruitment regions were selected to ensure coverage of geographical locations (North and South Islands, rural and urban) and ethnicities (Māori, the indigenous people of NZ, and non-Māori). The eligible population was all drivers who passed the learner license test at the venues where recruitment was being undertaken, who had not previously held a car driver’s license. At recruitment, background information was collected. When the learner licensed driver passed the restricted license test, the first follow-up interview was undertaken. This was a telephone administered questionnaire to obtain data about the quality and quantity of driving experiences as a learner licensed driver and the supervision received. When the full license test was passed, a second follow-up interview was conducted to determine the quality and quantity of driving experience as well as traffic crash outcomes such as traffic offenses and crash involvement as a restricted license driver. 2.1. Data collection The data presented in this paper have been drawn from the baseline, first, and second follow-up interviews. As highlighted above, the baseline interviews were conducted during cohort recruitment (n = 3,992, 1 Feb 2006 to 31 Jan 2008). Data were extracted from the first follow-up stage on 17 September, 2010. At this point 2,474 had passed the restricted license test and 88% of them had completed the follow-up interview. Data were extracted from the second follow-up interview on 30 May 2011. At this point 1,470 study members had passed their full license test (full license drivers, FLDs), 93% of whom completed the NZDS second follow-up interview. There were 1,379 study members who had participated in both the first and second follow-up interviews and are included in the current analysis. The remaining 91 did not do both interviews either because they refused one or the other of the interviews, or could not be traced at one of the stages. 2.2. Driving experience variables At the first follow-up interview, study members were asked details about professional driving lessons (lessons paid for with a
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
professional driving instructor), supervised driving, unsupervised driving (illegal driving at the learner license stage), and driving courses in which they participated while on their learner license. For each professional driving lesson, and supervised or unsupervised driving experience, study participants were asked for the following information: • Driving experience: The amount of driving experience was estimated by the interviewer “walking” the study participant through a series of questions. 1. They were asked how frequently they drove: at least daily, or weekly, or fortnightly or monthly, and how many times they drove in a typical day, week, fortnight, or month. This number was later multiplied by the number of days, weeks, fortnights, or months they had been on their learner license (which was calculated from the driver license registry file) to give a total number of trips. 2. Next they were asked how far they usually drove on a typical trip. The response options were b11kms, 11-20 km, 21-50 km, 51-100 km or >100kms. The distance driven variable was calculated by multiplying the total number of trips they had driven (from (1) above) by the approximate midpoint for each distance category (5 km, 15 km, 35 km, 75 km, 150 km). 3. Finally, they were asked about the duration (in minutes) of a typical drive. An estimate of the duration driven (minutes) was calculated by multiplying the total number of trips they had driven (from (1) above) by the number of minutes per typical drive. An overall measure of total distance driven (kms) and total duration driven (minutes) was calculated by combining the distances and durations estimated for professional driving lessons, all supervised driving trips and unsupervised driving trips. Acknowledging the imprecision of the driving estimates, all of the driving exposure measures were categorized according to low (b25th percentile), medium (25-75th percentile), and high (>75th percentile) for analysis. In addition, a zero category was provided for those who did not drive unsupervised or who did not have professional driving lessons. • Driving conditions: To obtain an indication of the variety of conditions in which the learner driver had driven, the cohort members were asked “how many times did you drive on a public road (never, once, 2–5 times, 6–10 times, >10 times) during”: (1) conditions that made it difficult to see; (2) on icy, (3) wet or (4) gravel/unsealed roads; (5) in heavy city traffic; in a 100 km/h area (6) during daylight or (7) when it was dark; in a 50 km/h area (8) during daylight or (9) when it was dark; (10) on any long trips (greater than 50 km in distance). For the purposes of the current investigation, a score of ‘1’ was given for each of the 10 driving conditions described above that were experienced two or more times while on the learner license and ‘0’ for those conditions experienced fewer than two times. A summary score of 0 for this variable indicated that none of the conditions were experienced two or more times, while a score of 10 indicated that all of the conditions were experienced at least twice. 2.3. Confounding variables Potential confounding variables included in the analysis were identified from the literature. Age at learner licensure, gender, and residential location (rural vs. urban) were determined at baseline interview (Begg et al., 2009). Alcohol use was measured using the first three questions of the AUDIT (the Audit-C) (Barbor, de la Fuente, Saunders, et al., 1989) and categorized as established by alcohol researchers (Bradley et al., 2007; a cut-off of ≥4 was used to indicate high level alcohol use for males and ≥3 for females). The Zuckerman-Kuhlman Personality Questionnaire scales for Impulsivity-Sensation Seeking (ImpSS) and Aggression-Hostility (Agg-Host; Zuckerman, Kulman, Teta, Joireman, &
43
Kraft, 1993) were included to obtain a measure of impulsivity (8 items), sensation seeking (11 items), and aggression-hostility (17 items). Gender specific cut-off points were selected to classify each of the scales into three categories: low (lowest quartile), medium (middle quartiles), and high (highest quartile). In the second follow-up interview, cohort members were asked to report the duration and distance driven while on their restricted license. These data were collected using the same format as in the first follow-up interview described above. Restricted license driving exposure (distance driven) was included as a confounding variable. 2.4. Outcome The main outcome variable was whether or not a driver was in a car crash on a public road during the restricted license stage. Motor vehicle traffic crashes (those in which the cohort member was driving a car or similar vehicle) were obtained from the traffic crash reports (TCR) as recorded by the NZ Police. In NZ when a road traffic crash results in someone being injured it is required by law to be reported to the Police who complete a TCR for that crash (Ministry of Transport, 2011). Despite the legal requirement to report crashes, hospital based surveys have shown that only around two-thirds of injury crashes requiring admission to hospital are reported to, or recorded, by the Police. Furthermore there are biases in the type of crashes reported with crashes involving younger drivers being less likely to be reported than those involving older drivers (Alsop & Langley, 1998). Therefore, to complement the TCR crashes, and to provide a more complete picture of crash involvement, we included self-reported crashes. Data on self-reported crashes were collected at the second follow-up interview and included crashes that occurred on-road where someone was injured and/or there was vehicle or property damage. The combined crash outcome, with duplicates removed, was dichotomized (yes/no). Only crashes that occurred while cohort members were on their restricted license (i.e., when unsupervised driving was first allowed) were included in the analysis. 2.5. Analysis All data analysis was conducted in Stata12SE. The GLM procedure with Poisson distribution was used to estimate the relative risks of learner license variables with crash involvement. To adjust for variable length of time while on the restricted license stage, the number of days on a restricted license was included in the model. 3. Results Of the 1,379 FLDs, 339 (24.6%) reported a car crash on a public road where there was an injury or property damage while they were on their restricted license. Driving experience variables and their association with restricted license crash involvement are displayed in Table 1. Less than 10% of FLDs had taken part in a driving course when they were learner licensed drivers. Over 50% of FLDs did not have any professional driving instruction, and almost 20% engaged in unsupervised driving as a learner driver. Table 1 shows that the majority of driving experience for the FLDs was gained through informal supervised driving. The FLDs reported a high number of driving conditions experienced through supervised driving (>85% experienced at least 5 of the 10 listed driving conditions on at least two occasions). Almost 25% of FLDs reported that they drove for less than 27 hours while on their learner license, with the same proportion reporting that they drove for more than 117 hours. Almost 29% of the FLDs who were involved in a crash were on their learner license for 180–219 days, while almost 27% of those who were not involved in a crash were on their learner license for more than 438 days. Compared with those who did not drive unsupervised, relative risks for crash involvement were elevated for all of the unsupervised
44
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
Table 1 Learner licence driving exposure and experience and relationship with restricted licence crash involvement. Restricted licence crash No n Took part in driving course No 938 Yes 100
UL
%
75.52 74.07
304 35
24.48 25.93
1 1.06
0.78
1.43
26.37 22.91 20.75
1 0.87 0.79
0.72 0.46
1.05 1.35
27.12 19.74 24.29 26.11
1 0.73 0.9 0.96
0.56 0.71 0.73
0.95 1.13 1.26
27.03 23.72 22.77 23.56
1 0.88 0.84 0.87
0.66 0.68 0.64
1.16 1.05 1.18
19.89 25.21
1 1.25
0.92
1.71
23.30 22.78 29.04
1 0.98 1.24
0.77 0.97
1.24 1.61
25.00 21.99 29.46
0.88 1.18
0.7 0.93
1.1 1.5
23.39 30.57 30.00
1 1.31 1.28
1.01 0.9
1.69 1.82
23.03 29.81 32.33 26.32
1 1.29 1.4 1.14
0.94 1.07 0.77
1.77 1.84 1.69
23.03 30.77 31.29 26.67
1 1.33 1.36 1.16
0.96 1.04 0.78
1.85 1.77 1.71
UNSUPERVISED DRIVING Number of different driving experiences1 0 experiences 871 76.61 266 1-4 experiences 109 69.43 48 5-10 experiences 56 70.00 24 Unsupervised distance driven - LL 0 km 819 76.97 245 1-10 km 73 70.19 31 11-240 km 90 67.67 43 >240 km 56 73.68 20 Unsupervised duration driven - LL 0 hrs 819 76.97 245 1 min-9 mins 63 69.23 28 10 mins - 8 hrs 101 68.71 46 >8 hrs 55 73.33 20
1
LL
n
SUPERVISED DRIVING Number of different driving experiences1 1-4 experiences 145 80.11 36 5-10 experiences 887 74.79 299 Distance driven - LL 0-555 km 260 76.70 79 556-2925 km 522 77.22 154 >2925 km 237 70.96 97 Duration driven - LL 0-24.6 hrs 276 75.00 92 24.7-108.5 hrs 525 78.01 148 >108.5 hrs 237 70.54 99
206 536 242
RR
%
PROFESSIONAL LESSONS Number of different driving experiences1 0 experiences 525 73.63 188 1-4 experiences 471 77.09 140 5-10 experiences 42 79.25 11 Distance driven - LL 0 km 379 72.88 141 1-30 km 244 80.26 60 31-120 km 265 75.71 85 >120 km 150 73.89 53 Duration driven - LL 0 hrs 378 72.97 140 1 min-1.3 hrs 164 76.28 51 1.4-6 hrs 363 77.23 107 >6 hrs 133 76.44 41
TOTAL DRIVING Driving duration - LL b27 hrs 27-117 hrs >117 hrs Distance driven - LL b616 km 616-3075 km >3075 km Days on learner licence 180-219 days 220-290 days 291-437 days >438 days
95% CI
Yes
69.83 77.79 71.39
89 153 97
30.17 22.21 28.61
1 0.87 1.12
269 528 241
77.08 76.97 70.47
80 158 101
22.92 23.03 29.53
1 1 1.29
241 265 254 278
71.30 76.37 73.62 80.12
97 82 91 69
28.70 23.63 26.38 19.88
1 0.82 0.92 0.69
driving variables listed in Table 1. For example, compared with those who experienced no driving conditions as an unsupervised driver (or did not drive more than once in any of the conditions), those who experienced 1–4 conditions were 1.3 times more likely to be involved in a crash. Compared with those who did not drive unsupervised, those who drove 11–240 km unsupervised were 1.4 times more likely to be involved in a crash. There was also a relationship between total (supervised, unsupervised, and professional) distance driven on the learner license stage and crash involvement. In comparison with those whose total driving distance was less than 616 km during the learner license stage, those who drove >3,075 km were 1.3 times more likely to be involved in a crash. For those who were on their learner license for the least amount of time (180–219 days), there was a heightened risk of crash involvement compared with those who were on their learner license for the longest period (>438 days, Table 1). To identify learner license variables that were independently associated with crash involvement a fully specified regression model, including all described explanatory and confounding variables was initially attempted, however this model would not converge. We also investigated models that contained (1) only those variables describing the duration of driving experience or (2) the distance driven while on a learners, but again experienced problems with convergence. Finally, a more restricted model that included only those variables that showed an elevated relative risk (95% confidence interval for univariate RR not including 1) of crash involvement were included in the multivariate analysis. There was a high degree of multi-collinearity between the unsupervised driving variables, therefore a new variable was created that identified whether FLDs had driven unsupervised as a learner or not (Table 2). It is evident from the model that, after adjusting for age, gender, residential location, high level alcohol consumption, total distance driven during the restricted license stage, sensation seeking, impulsivity, aggression, and number of days on restricted license, compared with those who spent the least amount of time on their learner license, there was a reduced risk of crash involvement for those in the upper quartile of the length of time spent on a learner license. Those who drove unsupervised at the learner license stage had increased risk of crash involvement during the restricted license stage, although the confidence interval for the relative risk included 1. 4. Discussion The results presented indicate that, after adjusting for demographic, personality factors, and driving exposure, those who were in the upper quartile (greater than approximately 14 months) of the length of time they spent on their learner license were less likely to be involved in a crash during the first period of unsupervised driving (the restricted license stage) when compared with those who spent the least amount Table 2 Learner licence driving exposure and experience and relationship with restricted licence crash involvement – relative risk of traffic crash involvement as a restricted licence driver for selected learner driver experiences.⁎
0.69 0.88
0.79 1
0.64 0.72 0.53
1.09 1.43
1.27 1.66
1.06 1.17 0.91
Number of driving conditions experienced at least twice over the learner licence stage. Options were: (1) conditions that made it difficult to see; (2) on icy, (3) wet or (4) gravel/unsealed roads; (5) in heavy city traffic; in a 100 kph area (6) during daylight or (7) when it was dark; in a 50 kph area (8) during daylight or (9) when it was dark; (10) on any long trips (greater than 50 km in length).
RR
95% CI
Variable
LL
UL
z
p
Days on learner licence 0-219 days 1 220-290 days 0.77 291-437 days 0.94 >438 days 0.63
0.54 0.65 0.43
1.11 1.34 0.91
−1.38 −0.36 −2.42
0.17 0.72 0.02
Unsupervised driving No 1 Yes 1.29
0.95
1.76
1.66
0.20
⁎ Adjusted for age, gender, residential location, sensation seeking, impulsivity, aggression, hazardous alcohol consumption, distance driven on restricted licence, days on restricted licence.
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
of time on their learner license. In addition, those who had driven unsupervised as a learner driver were at heightened risk of crash involvement in the restricted license stage. The majority of experience gained by the NZDS cohort occurred in brief, informal supervised driving, with 65% reporting that the average duration of supervised driving episodes were 30 minutes or less (Gulliver, Begg, & Brookland, 2011). This being the case, if each cohort member spent 15 minutes driving each day for the duration of their learner license, and were not permitted to move on to the restricted license until they had acquired 120 hours of experience, 480 days would be required before this would occur. An alternative interpretation of the findings presented is that those who are less confident and who therefore spend longer on their learner license before opting to progress to their restricted license, may be more cautious drivers. Although we attempted to control for personality types by including sensation seeking, aggression, and hostility in the regression model, it is possible that the findings are a spurious association. However, our findings do support those of other researchers that more experience gained through either more months on the learner license or more hours driving, are associated with a reduced risk of crash involvement. It is of interest that the number of driving conditions experienced either through professional lessons, supervised driving, or unsupervised driving were not associated with crash involvement. In this investigation, there was no attempt to consider some types of driving conditions as more important than others. It is possible that some driving conditions would be more strongly related to crash involvement than others. For example, those cohort members who resided in the warmer parts of NZ would be less likely to be exposed to icy roads, reducing the chance of experiencing this driving condition as well as reducing the saliency of this driving condition for reducing their crash risk. However, further work is required to determine the relative importance of different types of experiences. In contrast with other investigations (Gregersen, 1997), the reported duration (number of hours) of driving exposure during the learner license stage was not related to crash involvement. To estimate the number of hours of driving exposure during the learner license stage, cohort members were stepped through a series of questions that have been shown to be a reliable method for obtaining travel exposure data (Lajunen & Ozkan, 2011). Also, the number of days, weeks, fortnights, or months they spent driving on a particular license was extracted from the driver licensing registry, and therefore not dependent on self-report. While this method of estimating driving exposure may introduce measurement error, it is likely that the measurement error was smaller than traditional methods of estimating exposure – by asking one single question concerning the number of hours being driven at the end of the license period. Two Australian studies have used travel journals to record duration of driving exposure with interesting results (Bates, Watson, & King, 2010; Scott-Parker, Bates, Watson, King, & Hyde, 2011). In a retrospective, cross-sectional study, Bates et al. (2010) compared the amount of driving experience reported by learner drivers in two Australian jurisdictions. Those from New South Wales (NSW) were legally required to complete 50 hours of supervised driving practice before attempting a practical driving test, while those from Queensland had no minimum level of driving practice required. Participants reported the amount of private, supervised driving and driving instruction at the end of the learner license period. Perhaps unsurprisingly, 99% of participants from NSW reported that they completed the 50 hours required experience, with the mean duration being 73.3 hours. In contrast, Queensland study members tended to report that they either participated in 26–50 hours or more than 100 hours supervised driving (mean 64 hours; Bates et al., 2010). We consider that the use of a stepwise method to estimate the distance and duration of driving exposure will produce more reliable estimates than a solitary question, or the use of travel diaries that would require a high level of motivation to complete regularly.
45
Less than 10% of the FLDs reported that they took part in a driving course. Although there are a number of driving courses available in NZ, of those who took part in a course, the majority took part in Practice (Gulliver et al., 2011). Despite the fact that all learner drivers aged 15–19 years are sent an invitation to participate in “Practice,” a free driving education program provided by the NZ government, only 8% of the cohort who participated in the first follow-up interview reported that they did so, while over 40% reported that they didn’t know a course was available (Gulliver et al., 2011). Contrary to preliminary evidence produced by NZTA and ACC, participation in a driving course, such as Practice, while on the learner license was not associated with a reduced risk of crash involvement (Table 1). Because of the nature of the investigation, to determine learner license experiences that were associated with crash involvement in the restricted license stage, only those cohort members who had progressed through to their full license were included. A previous investigation has shown that cohort members who did not progress from the learner license to restricted license stage were more likely to be female, tertiary students, older, and from more deprived areas (Langley et al., 2012). Non-progressors also had reduced risks of traffic offending, and the primary reasons for non-progression were “too lazy,” “busy,” or had “limited access to a car” (Langley et al.). In this investigation, compared with the remainder of the cohort, the FLDs were younger when they obtained their license, were predominantly non-Maori, males, and from less deprived areas. The FLDs were also exposed to a larger variety of driving conditions and drove more supervised distance. This suggests that the results presented are applicable to those who are motivated, and have the resources to gain the experience and supervision required to progress through the graduated license system. However, the Langley et al. investigation supports the results presented in this study (that those who were on their learner license for longer periods of time had reduced risks of crash involvement) and, as such, it is likely that the results presented underestimate the association of reduced crash involvement for those who were on their learner license for longer periods of time. 5. Conclusion Results presented in this paper suggest that learner drivers in NZ should be encouraged to spend more time on their learner license to enable them to gain skills and experience to help reduce their crash risk when they are allowed to drive unsupervised. Acknowledgments The New Zealand Drivers Study is funded by the Health Research Council of New Zealand (2005–2013), the Road Safety Trust (2005– 2013) and the Accident Compensation Corporation (2005–2009). No funding source had any involvement in the preparation of this paper. The authors wish to acknowledge Dr Rebbecca Lilley, Injury Prevention Research Unit, University of Otago, for comments on an earlier version of this paper. In addition, the authors wish to acknowledge the helpful comments of the reviewer on earlier drafts of this manuscript. References Accident Compensation Corporation, & Land Transport Safety Authority (2003). Your practice guide. Practice: The plan for learning to drive safely. Wellington, New Zealand: ACC, LTSA. Alsop, J. C., & Langley, J. D. (1998). A determination of biases in LTSA's Traffic Crash Report files with respect to serious occupant road crashes: a data linkage study. Dunedin: Injury Prevention Research Unit, University of Otago. Barbor, T. F., de la Fuente, J. R., Saunders, J., & Monteiro, M. G. (1989). The alcohol use disorders identification test: Guidelines for the use in primary health care. Geneva: World Health Organisation. Bates, L., Watson, B., & King, M. (2010). Required hours of practice for learner drivers: A comparison between two Australian jurisdictions. Journal of Safety Research, 41, 93–97.
46
P. Gulliver et al. / Journal of Safety Research 46 (2013) 41–46
Begg, D., Langley, J., Broughton, J., Brookland, R., Ameratunga, S., & McDowell, A. (2009). New Zealand Drivers Study: a follow-up study of newly licensed drivers. Injury Prevention, 15(4), e2–e11, http://dx.doi.org/10.1136/ip.2009.021998. Begg, D., & Stephenson, S. (2003). Graduated driver licensing: the New Zealand experience. Journal of Safety Research, 34(1), 99–105. Bradley, K. A., DeBenedetti, A. F., Volk, R. J., Williams, E. C., Frank, D., & Kivlahan, D. R. (2007). AUDIT-C as a brief screen for alcohol misuse in primary care. Alcoholism, Clinical and Experimental Research, 31(7), 1208–1217. Gregersen, N. P. (1997). Evaluation of 16-years age limit for driver training. First report. Linkoping, Sweden: VTI (Swedish National Road and Transport Research Institute). Gulliver, P., Begg, D., & Brookland, R. (November 2–4). Newly licensed drivers: How are they learning? Paper presented at the 10th National Conference on Injury Prevention and Safety Promotion. Brisbane Convention and Exhibition Centre. Hedlund, J., Shults, R. A., & Compton, R. (2003). What we know, what we don't know, and what we need to know about graduated driver licensing. Journal of Safety Research, 34, 107–115, http://dx.doi.org/10.1016/S0022-4375(02)00086-5. Insurance Institute for Highway Safety (2012). Young driver licensing systems in the U.S. Retrieved from. http://www.iihs.org/laws/gdl_learners.aspx. Jacobsohn, L., Garcis-Espana, J. F., Durbin, D. R., Erkoboni, D., & Winston, F. K. (2012). Adult supervised practice for adolescent learners: The current state and directions for interventions. Journal of Safety Research, 43, 21–28. Joel, A. (2010). Practice - selling safety to teen drivers. Paper presented at the Australasian Road Safety Research, Policing and Education Conference, Canberra, ACT. Lajunen, T., & Ozkan, T. (2011). Self-reported instruments and methods. In B. Porter (Ed.), Handbook of Traffic Psychology. Waltham, MA: Academic Press. Lajunen, T., & Summala, H. (2003). Can we trust self-reports of driving? Effects of impression management on driver behavior questionnaire responses. Transportation Research Part F: Traffic Psychology and Behavior, 6(2), 97–107. Langley, J., Begg, D., Brookland, R., Samaranayaka, A., Davie, G., & Jordon, H. (2012). Non-progression through graduated driver licensing: characteristics, traffic offending, and reasons for nonprogression. Traffic Injury Prevention, 13(1), 7–13. 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. McCartt, A. T., Teoh, E. R., Fields, M., Braitman, K. A., & Hellinga, L. A. (2010). Graduated licensing laws and fatal crashes of teenage drivers: a national study. Traffic Injury Prevention, 11, 240–248. Ministry of Transport (2011). Motor Vehicle Crashes in New Zealand. Retrieved from. http://www.transport.govt.nz/research/motorvehiclecrashesinnewzealand/. New Zealand Transport Agency (2010). The Official New Zealand Road Code. Retrieved from. http://www.nzta.govt.nz/resources/roadcode/road-code-index.html. NZ Transport Agency (2012). Safer journies for teen drivers: A toolkit for parents of teen drivers. Retrieved from. http://www.safeteendriver.co.nz/. O'Brien, N. P., Foss, R. D., Goodwin, A. H., & Masten, S. V. (2012). Supervised hours requirements in graduated driver licensing: Effectiveness and parental awareness. Accident Analysis and Prevention, http://dx.doi.org/10.1016/j.aap.2012.05.007. Scott-Parker, B. J., Bates, L., Watson, E. C., King, M. J., & Hyde, M. K. (2011). The impact of changes to the graduated driver licensing program in Queensland, Australia on the experiences of learner drivers. Accident Analysis and Prevention, 43, 1301–1308. Scott-Parker, B., Watson, B., King, M. J., & Hyde, M. K. (2011). Mileage, car ownership, experience of punishment avoidance, and the risky driving of young drivers. Traffic Injury Prevention, 12(6), 559–567.
VicRoads (2012). How to get your probationary license. Retrieved from. http://www. vicroads.vic.gov.au/Home/Licenses/GetYourPs/Howtogetyourprobationarycarlicense. htm. 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(1), 49–54. Williams, A. F., Teft, B. C., & Grabowski, J. G. (2012). Graduated driver licensing research, 2010-present. Journal of Safety Research, 43, 195–203. Zuckerman, M., Kulman, D. M., Teta, P., Joireman, J., & Kraft, M. (1993). A comparison of three structural models for personality: the big three, the big five and the alternative five. Journal of Personality and Social Psychology, 65, 757–768.
Pauline Gulliver has been employed as a Research Fellow at the Injury Prevention Research since 2001. She completed her PhD in Human Nutrition at Otago University in 1998 and has worked for the Cancer Society of New Zealand and Kings College London before returning to Otago University. She has worked on epidemiological investigations of childhood injuries in the home, the surveillance and description of sports injuries, and predictors of traffic crash involvement. Dorothy Begg is an injury epidemiologist and co-ordinates the IPRU road safety research. Her undergraduate training was in psychology and postgraduate training was in Public Health (epidemiology). Her principal research interest is young drivers, and in particular the graduated driver licensing system and its impact on young driver behaviour. Current and recent research focuses on young newly licensed drivers and the graduated driver licensing system, and motorcycling injury research. Rebecca Brookland is a Research Fellow with the Injury Prevention Research Unit, and has a Master of Arts, in Psychology from the University of Otago. Currently her research focuses on young and newly licensed drivers. For her PhD Rebecca is investigating parental influence on the driving experiences of young drivers. She is also Project Manager for the New Zealand Drivers Study, a large follow-up study of nearly 4000 newly licensed drivers. Shanthi Ameratunga is paediatrician and public health physician by background, Shanthi leads a multi-disciplinary research program focusing on trauma outcomes, injury prevention, disability and rehabilitation. She is the Project Director of the Traffic Related Injury in the Pacific (TRIP) Study, a collaboration with the Fiji School of Medicine funded by The Wellcome Trust and the Health Research Council of New Zealand. Other projects she contributes to include the national youth health surveys (Youth2000/Youth’07), the PREDICT project (cardiovascular disease prevention), and the ADAPT trial (mobile phone intervention for depression prevention in young people). John Langley has been involved in injury prevention research since 1978, and received his PhD from the University of Otago in 1985. He has written numerous articles in academic, professional and lay publications on a wide range of injury prevention issues. He tends to use quantitative research methods, especially epidemiological. He is an active advocate for improvements in injury prevention policy and practice. John’s current areas of research interest include: young drivers, outcomes of injury, alcoholrelated injury, and injury outcome indicators.