Embed Size (px)
Citation preview
A
K
prwrialta
tM
1d
Original article
Road Rage Victimization Among AdolescentsReginald G. Smart, Ph.D.a,*, Gina Stoduto, M.Ed.a, Edward M. Adlaf, Ph.D.a,b,
Robert E. Mann, Ph.D.a,b, and Justin M. Sharpley, B.A.aaSocial, Prevention and Health Policy Research, Centre for Addiction and Mental Health, Toronto, Ontario, Canada
bDepartment of Public Health Sciences, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
Manuscript received November 28, 2006; manuscript accepted April 16, 2007
bstract Purpose: Although much has been learned about road rage among adults, data on road rageexperiences among adolescents has not been available previously. We examine the prevalence anddemographic correlates of road rage victimization based on a population survey of Ontario students.Methods: Based on the 2005 Ontario Student Drug Use Survey, a self-administered survey ofOntario students in grades 7–12 (n � 7726), the contribution of demographic factors (gender,region, driver’s license status, grade, overall marks) to three road rage victimization measures wasexamined with logistic regression analysis.Results: Just over half of students (53.2%) had been victims of shouts, curses and rude gestures inthe past year, 8.9% were threatened with damage to their vehicle or personal injury and 6.2%experienced an attempt or actual damage to their vehicle or personal injury. Logistic regressionanalyses revealed that being a victim of shouting was significantly related to region, driver’s licensestatus, and grade. Victimization by threats was significantly related to gender, driver’s license,grade, and marks. Being a victim of attempts or actual vehicle damage or injury was significantlyrelated to region, driver’s license, and marks.Conclusion: This study provides the first indication of prevalence of road rage victimization amongadolescents. Road rage victimization in its milder form is common, involving just over half ofOntario students in grades 7–12. About 1 in 10 students were threatened with vehicle damage orpersonal injury, and about 1 in 20 were victims of attempts or actual damage or injury. © 2007Society for Adolescent Medicine. All rights reserved.
Journal of Adolescent Health 41 (2007) 277–282
eywords: Adolescence; Victimization; Road rage; Student population survey
ed[ft1rcsmvte
In recent years road rage has been identified as aublic concern [1– 4]. Among adults in North America,oad rage appears to be a relatively common experience,ith nearly half of drivers reporting experiencing road
age victimization or perpetration in the past year at leastn its milder versions [5,6]. Road rage is more commonmong males, younger persons, and people living inarger urban centers [5]. Reports of road rage victimiza-ion and perpetration are increased by traffic congestionnd stressful driving conditions [7]. As well, higher lev-
*Address correspondence to: Reginald Smart, Ph.D., Social, Preven-ion and Health Policy Research Department, Centre for Addiction and
ental Health, 33 Russell St., Toronto, ON, M5S 2S1.
cE-mail address: [email protected]054-139X/07/$ – see front matter © 2007 Society for Adolescent Medicine. Alloi:10.1016/j.jadohealth.2007.04.013
ls of road rage are associated with increased alcohol andrug involvement [8 –10] and psychiatric distress11,12]. Road rage perpetration, in its more seriousorms, appears to share similar epidemiology and symp-om patterns with Intermittent Explosive Disorder [13–5]. Recent studies have shown that the experience ofoad rage is associated with a significant increase inollision risk [6,13]. Victimization and perpetration aretrongly related; individuals who report being victims areuch more likely to report being perpetrators, and vice
ersa [16,17]. Experience as a victim, as well as perpe-rator, is associated with increased likelihood of collisionxperience [13].
Motor vehicle collisions are the leading cause of adoles-
ent death and serious injury in North America; for exam-rights reserved.
piuchuoceedthtsmla[fdabtse
visthTcoastoiia
M
DoAsu(7t9
RM
twimosw(dycponvolPsCT
(w6TOadicoasotadhlBs
fcotltoR
278 R. Smart et al. / Journal of Adolescent Health 41 (2007) 277–282
le, in 1997 collisions accounted for over 35% of all deathsn the 15- to 19-year age group in Canada [18]. Thus annderstanding of risk factors for collisions among adoles-ents must be considered a health priority. Several factorsave been identified as affecting collision risk in this pop-lation. Factors such as inexperience and use of alcohol andther psychoactive drugs, and distraction, which influenceollision risk among all drivers [19,20], may exert a largerffect on adolescents than on adult drivers [21–24]. Zhangt al [25] compared risk factors among adolescent and adultrivers, and found that adolescent drivers were more likelyo experience collision risk resulting from risk-taking be-aviors such as alcohol and drug use, speeding, and failureo wear seat belts, were more likely to crash during theummer, on weekends, and at night, and during overtakinganeuvers. Adolescent collision risk is low during the
earner period when driving is typically supervised by andult driver, but increases substantially right after licensure23]. Mayhew et al [26] observed that collision rates declineollowing the first 6 months of licensing, with largest re-uctions occurring in run-off-the-road, single-vehicle, nightnd weekend collisions. Adolescent collision risk appears toe enhanced by many factors and the propensities for risk-aking and exploration characteristic of this period by them-elves may enhance collision risk and as well contribute tonhanced risk due to other factors.
No studies have reported on the experience of road rageictimization among adolescents. Since road rage victim-zation and perpetration appear to be risk factors for colli-ion involvement among adults [6,13], it is possible thatheir impact could be even greater among adolescents, asas been seen for such factors as alcohol and drugs [21,23].he propensity of adolescent drivers for risk-taking [27]ould result in an even larger impact of road rage experiencen collision risk than seen in adult drivers, if, for example,dolescent drivers are more likely to respond in an aggres-ive or threatening manner to provocations while driving. Inhis paper we report prevalence estimates for the experiencef less severe and more severe forms of road rage victim-zation in a large representative sample of Ontario studentsn grades 7–12, and examine demographic factors associ-ted with that experience.
ethods
Data are based on the 2005 cycle of the Ontario Studentrug Use Survey (OSDUS) [28], a self-administered surveyf Ontario students (n � 7726), conducted by the Centre forddiction and Mental Health and administered by the In-
titute for Social Research, York University. The surveysed a full-probability two-stage (school, class), stratifiedregion and school type), cluster design sample in grades–12. The resulting sample represents about 975,000 On-ario students. The school and student participation rate was
0% and 72%, respectively. This study was approved by the nesearch Ethics Board of the Centre for Addiction andental Health.Road rage items were introduced with, “The next ques-
ions are about things that might have happened when youere in a car, motorcycle or other motor vehicle.” Three
tems asked students, “How many times in the last 12onths: (1) has someone in another vehicle shouted, cursed,
r made rude gestures at you or others with you? (2) hasomeone in another vehicle threatened to hurt you or othersith you, or threatened to damage the vehicle you were in?
3) has someone in another vehicle damaged or attempted toamage the vehicle you were in, or hurt or attempted to hurtou or others with you?” These items were designed toapture a broad victim experience whether it occurs as aassenger or driver. Response categories, which were none,ne, two, and three or more were binary coded to representever versus one or more times. The questions on road rageictimization appear near the end of the survey, followingther driving-related questions, such as type of driver’sicense, drinking and driving and cannabis use and driving.ercentages reported are based on weighted data and con-idered representative for the population surveyed. SPSS 14omplex Samples software (SPSS Inc., Chicago, IL) usingaylor series estimation was used in analyses.
Prevalence of road rage victimization by demographicsgender, age, region, driver’s license status, grade, marks)as examined through chi-square analyses (valid n values:811–7276). The regions involved in this study includedoronto, and the Northern, Western and Eastern regions ofntario. Toronto is a region with the largest city in Canada
nd its suburbs. The Eastern and Western regions are pre-ominantly rural and some smaller cities, whereas the Norths the largest region but the least populated with no largeities. The driver’s license status variable measures the typef license the student has, if any. In Ontario new driversged 16 years or more enter a graduated driver’s licensingystem. The G1 driver’s license (this stage lasts a minimumf 8 to 12 months, depending on whether a driving course isaken) has several driving restrictions including a zero bloodlcohol concentration (BAC), no driving without an adultriver in the front passenger seat, and no driving on majorighways or after midnight. The second stage, termed a G2icense (lasts a minimum of 12 months), only has the zeroAC restriction. A full driver’s license has no driving re-
trictions.Multivariable logistic regression analyses were per-
ormed with five dummy-coded independent variables (Be-ause of the high inter-correlation between age and grade,nly grade was included in the logistic regressions). Con-rasts for region were made to Toronto residents, for driver’sicense status were made to no license, for grade were madeo grade 7, and marks were made to 80–100%. To evaluateverall goodness of fit we used the Cox and Snell Pseudo2, which assesses whether the specified model differs sig-
ificantly from a perfect model. We are quite confident that
trwpvtdi
R
vMw
ttwdtcsrHovs.l
TP
TG
A
R
G
O
D
279R. Smart et al. / Journal of Adolescent Health 41 (2007) 277–282
he listwise loss of data should not bias our regressionesults. Bias due to listwise deletion is less likely to occurhen the fraction of complete cases is high and when therobability of being missing depends on the independentariables, but not on the dependent variables [29]. Addi-ional regression analyses indicated that values of the threeependent variables were unrelated to whether any of thendependent variables were also missing.
esults
Table 1 presents data on the prevalence of road rageictimization in the last 12 months by demographic factors.ore than half of the respondents (53.2%) reported they
able 1revalence of road rage victimization by demographic characteristics: 200
Someone shouted, cursedor made rude gestures atyou
N %
otal sample 6811 53.2enderMale 3297 53.1Female 3514 53.3
ge (y) ***�12 567 38.713 784 45.414 1031 49.015 1238 56.516 1359 58.117 1190 62.618 554 58.0�19 89 55.8
egion *Toronto 1016 48.4North 1081 47.0West 2523 56.4East 2191 53.1
rade ***7 799 40.18 821 43.69 1255 52.5
10 1289 57.311 1392 60.612 1255 61.4
verall marks80�100% 2790 54.467�79% 2954 51.860�66% 808 54.3�60% 246 55.1
river’s license ***No license 4478 49.9G1 1173 60.6G2 916 64.2Full license 85 60.8Don’t know 110 31.0
* p � .05; ** p � .01; *** p � .001.
ere shouted at, cursed at or had rude gestures directed at f
hem in the past year, 8.9% were threatened with damage toheir vehicle or personal injury, and 6.2% were victimshere someone in another vehicle damaged or attempted toamage the vehicle they were in, or hurt or attempted to hurthem or others with them. Victimization by others shouting,ursing or making rude gestures at them was related totudents’ age (p � .001), grade (p � .001), region ofesidence (p � .05), and driver’s license status (p � .001).owever, there was no significant difference by gender andverall marks. For victimization by threats of injury orehicle damage, there was a statistically significant relation-hip for all factors except region (gender p � .001, age p �001, grade p � .01, overall marks p � .001, and driver’sicense status p � .001). The least common and most severe
US data
omeone threatened to hurtou or damage yourehicle
Someone damaged orattempted to damage yourvehicle, or hurt orattempted to hurt you
% N %
280 8.9 7276 6.2***
483 11.0 3480 6.8797 6.7 3796 5.7
***603 4.2 605 5.6852 7.3 834 5.1121 9.4 1117 6.5334 9.1 1333 6.5432 10.3 1436 6.0266 9.1 1267 6.0581 11.4 593 7.891 23.6 91 13.7
085 9.3 1075 7.2164 9.2 1158 4.8685 9.7 2687 6.8346 7.7 2356 5.4
**854 6.3 846 5.9893 6.5 882 5.3370 9.8 1366 6.2374 10.1 1371 6.5470 9.1 1483 6.8319 10.9 1328 6.6
*** ***021 6.1 3126 4.4117 9.4 3122 6.9869 14.8 855 8.7255 16.4 255 12.3
*** ***825 7.8 4820 5.6237 10.3 1239 6.1961 11.6 967 8.489 22.2 88 19.8
112 11.2 112 5.6
5 OSD
Syv
N
7
33
1111
1122
1111
33
41
orm of road rage victimization, damage or attempts of
dpo
dTcstrdeteptsv(
a
ageSt(Gtachm
awlaeTh
TR
G
R
D
G
O
P
280 R. Smart et al. / Journal of Adolescent Health 41 (2007) 277–282
amage to their vehicle and hurting or attempts to hurteople, varied only by driver’s license status (p � .001) andverall marks (p � .001).
Multivariate analyses of the three road rage items withemographics as independent factors are presented in Table 2.he logistic regression analysis of victimization by shouts,urses and gestures revealed that region, driver’s licensetatus, and grade were significant contributors. Students inhe West had significantly greater odds (OR � 1.36) ofeporting victimization compared with Toronto students. G2river’s license holders had significantly greater risk ofxperiencing shouts (OR � 1.32) and those who don’t knowheir driver’s license status had about half the odds ofxperiencing shouts, curses and gestures (OR � .52) com-ared with those without a license. Students in grades 9hrough 12 had significantly greater odds compared withtudents in grade 7 of having experienced this type ofictimization, with grade 11 students having twice the oddsOR � 1.99).
Reporting that someone threatened to hurt them or dam-
able 2oad rage victimization by demographic characteristics: Logistic regressi
Someone shouted, cursed ormade rude gestures at you(N � 6750)
OR 95% CI
enderMale 1.00 0.89, 1.12Female (r.c.) — —
egion **Toronto (r.c.) — —North 0.94 0.67, 1.31West 1.36** 1.08, 1.70East 1.18 0.91, 1.55
river’s license **No license (r.c.) — —G1 1.13 0.90, 1.43G2 1.32* 1.05, 1.66Full license 1.29 0.74, 2.24Don’t know 0.52* 0.31, 0.86
rade ***7 (r.c.) — —8 1.11 0.90, 1.379 1.62*** 1.27, 2.07
10 1.89*** 1.47, 2.4311 1.99*** 1.52, 2.6112 1.91*** 1.50, 2.45
verall marks80–100% (r.c.) — —67–79% 0.92 0.81, 1.0660–66% 1.02 0.83, 1.24�60% 0.94 0.67, 1.32
seudo R2 .033
Pseudo R2: based on Cox and Snell.r.c. � Reference category.* p �.05; ** p � .01; *** p � .001.
ge their vehicle within the last 12 months was significantly 5
ssociated with students’ gender, driver’s license status,rade, and marks. Males had 50% greater odds of havingxperienced threats (OR � 1.52) compared with females.tudents with a full driver’s license had more than three
imes greater odds of reporting victimization of this kindOR � 3.25) compared with those without a license. G1 and2 license holders also had greater odds of experiencing
hreats (OR 1.43 and 1.7, respectively). Students in grades 9nd 10 had a 60% greater chance of experiencing threatsompared with those in grade 7, and those with lower marksad greater odds of experiencing threats than those witharks 80–100%.Experiencing someone damaging or attempting to dam-
ge their vehicle, hurt or attempting to hurt them or othersith them was significantly associated with region, driver’s
icense status, and marks. Students in the East (OR � .68)nd North (OR � .54) were least at risk of having experi-nced this type of severe road rage compared with those inoronto. G1 license holders had over twice the odds and G2olders had more than five times the odds (OR 2.31 and
yses, 2005 OSDUS data
eone threatened to hurtor damage your vehicle
7208)
Someone damaged or attemptedto damage your vehicle, or hurtor attempted to hurt you(N � 7210)
95% CI OR 95% CI
*** 1.24, 1.85 1.03 .82, 1.30— — —
*—
0.66, 1.28 0.54** 0.34, 0.840.77, 1.25 0.84 0.62, 1.140.59, 1.07 0.68* 0.48, 0.97
***— — —
* 1.01, 2.03 1.37 0.90, 2.09** 1.18, 2.44 2.31*** 1.47, 3.63** 1.65, 6.42 5.35*** 2.79, 10.22
0.64, 2.57 0.90 0.37, 2.22
— — —0.67, 1.72 0.96 0.60, 1.53
* 1.07, 2.44 1.02 0.66, 1.58* 1.11, 2.34 1.05 0.70, 1.58
0.75, 1.70 0.82 0.52, 1.270.82, 1.90 0.63 0.38, 1.04
***— — —
*** 1.30, 1.99 1.62*** 1.25, 2.12*** 1.88, 3.32 1.98*** 1.39, 2.82*** 1.83, 3.63 2.87*** 1.85, 4.462 .013
on anal
Somyou(N �
OR
1.52—
—0.920.980.79**—1.431.703.251.28*—1.081.621.611.131.25***—1.612.492.58
.02
.35, respectively) of reporting severe victimization than
tav
D
oatcio2do
eWaatsae7tamibsfitsatse
arrTarmhfioar
vr
gliteftdatmtovsIro[3mi
csodporlnmoilrlap
larratcav
281R. Smart et al. / Journal of Adolescent Health 41 (2007) 277–282
hose without a driver’s license. Students with lower markslso had significantly greater odds of experiencing severeictimization compared with those with marks of 80–100%.
iscussion
These results provide the first evidence on the prevalencef road rage victimization among adolescents. Among alldolescents (drivers and non drivers) road rage victimiza-ion, in its milder form of shouts, curses and gestures, wasommon, involving just over half of respondents. About onen 10 Ontario students were threatened with vehicle damager personal injury by someone in another vehicle, and one in0 students were victims where someone in another vehicleamaged or attempted to damage the vehicle they were in,r hurt or attempted to hurt them or others with them.
These data also suggest that road rage victimizationxperience is more common among adolescents than adults.hen we compare the prevalence of road rage victimization
mong Ontario adults based on 2005 Centre for Addictionnd Mental Health Monitor and our student data, we findhat 40.9% of adults versus 53.2% (CI � 51.0%, 55.4%) oftudents experienced shouts, curses and gestures, 5.3% ofdults versus 8.9% (CI � 8.1%, 9.9%) of students experi-nced threats and 3.9% of adults versus 6.2% (CI � 5.6%,.0%) of students experienced attempts or actual damage toheir vehicle or attempts to injure or actual injury. Amongdults in Ontario, Smart et al [5] found that road rage wasore common for males, younger persons, and people liv-
ng in Toronto. In our adolescent sample, males seemed toe more likely to experience road rage threats, but femaleseemed as likely as males to experience the more severeorm of victimization studied here. These differences arentriguing and potentially of significance. Adolescence is aime when driving behaviors are being formed, and expo-ure to victimization may act to predispose young people toggressive driving behaviors themselves. However, al-hough these differences between adolescents and adults areuggestive, more research is needed to confirm these differ-nces, identify their causes and explore their significance.
These data also suggest that students in the higher gradesre more likely to have experienced the milder form of roadage victimization, but not the more severe forms of roadage. Unlike adults in Ontario [5], adolescents who lived inoronto were not more likely to be victims of shouts, cursesnd gestures on the road but instead those in the Westernegion of the province were at greater risk of experiencingild road rage. However, students in the North and the East
ad much lower odds of having experienced the most severeorm of road rage victimization compared with those livingn Toronto. Reasons for these regional differences in thedds of experiencing road rage of various levels of severityre unclear at this time. It would be interesting for future
esearch to examine if those regions where adolescents are rictims of road rage have significantly higher levels of roadage perpetration by adults and adolescents.
For all forms of road rage victimization these data sug-est that as adolescents progress through the levels of theicensing system (G1 is the learner’s stage with many driv-ng restrictions; G2 has a smaller number of driving restric-ions; full license has no restrictions) they are more likely toxperience road rage victimization. This is especially trueor the more severe forms of road rage victimization, wherehe odds were 3 and 5 times greater for those with a fullriver’s license compared with those with no license. Asdolescents progress through the levels of driver’s licensinghey probably do more driving and therefore are exposed toore congested or more stressful driving conditions and
hey may take greater risks or make mistakes that frustratether drivers as they drive more often without adult super-ision. They may also be driving more aggressively and/orpeeding, which may provoke reactions from other drivers.n one study, 27% and 35% of 16- and 17-year-old drivers,espectively, were found to report speeding at least 10 mphver the posted speed limit [30]. O’Malley and Johnston31] found that among 12th graders in the United States,2% reported having received a ticket or a warning foroving violations such as speeding, running a stop light, or
mproper passing in the last year.Several limitations in the study must be taken into ac-
ount when interpreting the results. These data involveelf-report and thus are likely underestimates of the true ratef experiences. This sample excludes adolescents whoropped out of school, are in correctional facilities or hos-itals, or are homeless. Our observation that the experiencef road rage increased as marks decreased suggests that roadage experience could be substantially higher among ado-escent groups not included in this study. Finally, we wereot able to control for mileage driven. More driving, andore stressful driving, is associated with increasing reports
f victimization and perpetration among adults [7] and thencrease in road rage victimization associated with drivericensing, and with less restrictive licensing, seen here mayeflect a similar relationship in this sample. Keeping theseimitations in mind, the results are of substantial interest andre the first results describing road rage in any adolescentopulation.
These findings provide a first indication of the preva-ence of road rage victimization of various levels of severitynd the major demographic factors associated with roadage among adolescents. These data provide evidence thatoad rage is common among Ontario students. Amongdults, road rage victimization has been shown to be relatedo psychological distress, alcohol and drug misuse, andollisions [13,8,10–12]. Because adolescents in this sampleppear to be more likely than adults to report road rageictimization, including its more serious forms, additional
esearch on this issue is needed.
A
mpSCSn
R
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
282 R. Smart et al. / Journal of Adolescent Health 41 (2007) 277–282
cknowledgments
This research was supported by a grant from AUTO21, aember of the Networks of Centres of Excellence (NCE)
rogram, which is administered and funded by the Naturalciences and Engineering Research Council (NSERC), theanadian Institutes of Health Research (CIHR) and the Socialciences and Humanities Research Council (SSHRC), in part-ership with Industry Canada.
eferences
[1] Smart RG, Mann RE. Deaths and injuries from road rage: Cases inCanadian newspapers. Can Med Assoc J 2002;167:761–2.
[2] Harding RW, Morgan FH, Indermaur D, Ferrante AM, Blagg H. Roadrage and the epidemiology of violence: Something old, somethingnew. Stud Crime Prev 1998;7:221–38.
[3] Hennesy DA, Wiesenthal DI. The relationship between traffic con-gestion, driver stress and direct versus indirect coping behaviors.Ergonomics 1997;40:348–61.
[4] Joint M. Road Rage. Washington, DC: AAA Foundation for TrafficSafety, 1995.
[5] Smart RG, Mann RE, Stoduto G. The prevalence of road rage:Estimates from Ontario. Can J Public Health 2003;94:247–50.
[6] Wells-Parker E, Ceminsky J, Hallberg V, et al. An exploratory studyof the relationship between road rage and crash experience in arepresentative sample of U.S. drivers. Acc Anal Prev 2002;34:271–8.
[7] Smart RG, Stoduto G, Mann RE, Adlaf EM. Road rage experienceand behavior: Vehicle, exposure and driver factors. Traffic Inj Prev2004;5:343–8.
[8] Mann RE, Smart RG, Stoduto G, Adlaf EM, Ialomiteanu A. Alcoholconsumption and problems among road rage victims and perpetrators.J Stud Alcohol 2004;65:161–8.
[9] Yu J, Evans PC, Perfetti L. Road aggression among drinking drivers:Alcohol and non-alcohol effects on aggressive driving and road rage.J Criminal Justice 2004;32:421–30.
10] Butters J, Mann RE, Smart RG. Assessing road rage victimization andperpetration in the Ontario adult population: Illicit drug use andpsychiatric distress. Can J Public Health 2006;97:96–9.
11] Fong G, Frost D, Stansfeld S. Road rage: A psychiatric phenomenon?Soc Psychiat Psychiatric Epidem 2001;36:277–86.
12] Smart RG, Asbridge M, Mann RE, Adlaf EM. Psychiatric distressamong road rage victims and perpetrators. Can J Psychiat 2003;48:681–8.
13] Mann RE, Zhao J, Stoduto G, Adlaf EM, Smart RG, Donovan JE.Road rage and collision involvement. Am J Health Behav 2007;31:
384–91.14] Galovski T, Blanchard EB. Psychological characteristics of aggres-sive drivers with and without intermittent explosive disorder. BehavRes Ther 2002;40:1157–68.
15] Kessler RC, Coccaro EF, Fava M, Jaeger S, Jin R, Walters E. Theprevalence and correlates of DSM-IV Intermittent Explosive Disorderin the National Comorbidity Survey replication. Arch Gen Psychiat2006;63:669–78.
16] Asbridge M, Smart RG, Mann RE. The ‘homogamy’ of road rage:Understanding the relationship between victimization and offending.Violence Victims 2003;18:517–31.
17] Asbridge M, Mann RE, Smart RG. Can we prevent road rage?Trauma, Violence Abuse Rev J 2006;7:109–21.
18] Statistics Canada. Mortality—Summary List of Causes, 1997. Ot-tawa, ON: Statistics Canada, 1999.
19] Mann RE, Stoduto G, Macdonald S, Shaikh A, Bondy S, Jonah B.The effects of introducing or lowering legal per se blood alcohollimits for driving: An international review. Acc Anal Prev 2001;33:61–75.
20] Macdonald S, Mann RE, Chipman M, Anglin-Bodrug K. Collisionsand traffic violations of alcohol, cannabis and cocaine abuse clientsbefore and after treatment. Acc Anal Prev 2004;36:795–800.
21] Zhao J, Mann RE, Chipman M, Adlaf EM, Stoduto G, Smart RG. Theimpact of driver education on self-reported collisions among youngdrivers with a graduated licence. Acc Anal Prev 2006;38:35–42.
22] Simpson HM. Polydrug effects and traffic safety. Alcohol DrugsDriving 1985;1:17–44.
23] Williams AF. Teenage drivers: Patterns of risk. J Safety Res 2003;34:5–15.
24] Neyens DM, Boyle LN. The effect of distractions on the crash typesof teenage drivers. Acc Anal Prev 2007;39:206–12.
25] Zhang F, Fraser S, Lindsay J, Clarke K, Mao Y. Age-specific patternsof factors related to fatal motor vehicle traffic crashes: Focus onyoung and elderly drivers. Public Health 1998;112:289–95.
26] Mayhew DR, Simpson HM, Pak A. Changes in collision rates amongnovice drivers during the first months of driving. Acc Anal Prev2003;25:683–91.
27] Jonah BA. Sensation seeking and risky driving: A review and syn-thesis of the literature. Acc Anal Prev 1997;29:651–65.
28] Adlaf EM, Paglia-Boak A. Drug Use Among Ontario Students, 1977–2005: Detailed OSDUS Findings. Toronto: Centre for Addiction andMental Health, 2005. Available online at: http://www.camh.net/Research/osdus.html#2005osdusdrugusereport.
29] Little RJ, Rubin DB. Statistical Analysis with Missing Data. 2ndEdition. Hoboken, NJ: Wiley, 2002.
30] Husted DS, Gold MS, Frost-Pineda K, Ferguson MA, Yang MCK,Shapira NA. Is speeding a form of gambling in adolescents? J GamblStud 2006;22:209–19.
31] O’Malley PM, Johnston LD. Unsafe driving by high school seniors:National Trends from 1976 to 2001 in tickets and accidents after useof alcohol, marijuana and other illegal drugs. J Stud Alcohol 2003;
64:305–12.
