Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Association between leisure activity and risky driving behaviour in young drivers in Canada Mehrnoush, Vahid 2020

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2021_may_mehrnoush_vahid.pdf [ 4.45MB ]
Metadata
JSON: 24-1.0395576.json
JSON-LD: 24-1.0395576-ld.json
RDF/XML (Pretty): 24-1.0395576-rdf.xml
RDF/JSON: 24-1.0395576-rdf.json
Turtle: 24-1.0395576-turtle.txt
N-Triples: 24-1.0395576-rdf-ntriples.txt
Original Record: 24-1.0395576-source.json
Full Text
24-1.0395576-fulltext.txt
Citation
24-1.0395576.ris

Full Text

     Association between leisure activity and risky driving behaviour in young drivers in Canada  by  Vahid Mehrnoush    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (EXPERIMENTAL MEDICINE)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2020  © Vahid Mehrnoush, 2020    The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the thesis entitled: Association between leisure activity and risky driving behaviour in young drivers in Canada  submitted by Vahid Mehrnoush in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  Examining Committee: Dr. Jeff Brubacher, Emergency Medicine, Faculty of Medicine, UBC Supervisor  Dr. Herb Chan, Emergency Medicine, Faculty of Medicine, UBC Supervisory Committee Member  Dr. Mariana Brussoni, School of Population and Public Health, Faculty of Medicine, UBC Supervisory Committee Member Dr. Roy Purssell, Emergency Medicine, Faculty of Medicine, UBC Additional Examiner   iii  Abstract Background: There are myriad risk factors for risky driving behaviour in youth. Perceived environment which is defined as the perception of driving risk and norms is the most complex factor. Leisure activities are a central part of youth’s daily lives that can substantially shape the driving perceived environment by providing the platform for interaction with peers, family, society, and media. However, the potential relationship between leisure activities and risky driving behaviour has seldom been studied. The purpose of this study was to examine the relationship between various leisure activities and risky driving behaviour among young drivers in Canada. Methods: An online survey-based cross-sectional study was conducted. Participants aged 16–24 years were approached through Facebook advertisements. The survey comprised of four questionnaires, namely, sociodemographic, personality trait (Mini-IPIP), leisure activities, and Behaviour Young Novice Driver Scale (BYNDS). Chi-square test examined differences between the driver group and proportional odds logistic regression was used to determine the relationship between the predictor variables and risky driving behaviour. Results: Participants (n=964), aged 18.34±2.31, were grouped into high risk (46.9%), medium risk (32.4%), and low risk (20.7%) drivers. Those with higher levels of drug engagement (OR=2.09, CI 95%=1.21-3.71), time with friends (OR=1.98, CI 95%=1.46-2.68), social media engagement (OR=1.83, CI 95%=1.34-2.49), and movie watching engagement (OR=1.52, CI 95%=1.00-2.31) tended to manifest more risky driving behaviour. In contrast, those with high levels of reading/writing engagement (OR=0.60, CI 95%= 0.42-0.85), volunteering engagement (OR=0.60, CI 95%=0.36-0.96), and playing video game engagement (OR=0.56, CI 95%=0.38-0.81) were less likely to perform risky driving behaviour. Other factors such as owning a car (OR=3.01, CI 95% 2.21-4.11), being male (OR=2.52, CI 95%=1.85-3.42), being simultaneously employed and a student, (OR=1.58, CI 95%=1.16-2.16), high driving exposure (OR=2.58, CI 95%=1.54-4.41), high neuroticism (OR=1.83, CI 95%=1.23-2.73), high extroversion (OR=1.60, CI 95%=1.09-2.35), and low imagination (OR=1.53, CI 95%=1.01-2.34) increased the likelihood of risky driving behaviour. Conclusions:  This study provides new insight and explores the association between leisure activities and risky driving behaviour. Results from this study could be used to further explore leisure activities as a potential determinant of risky driving behaviour in future injury prevention research. iv  Lay Summary This research sought to understand the association between leisure activities and risky driving behaviours among young drivers (aged 16-24) in Canada. When it comes to leisure activities, high-risk drivers, compared to low-risk drivers, tended to spend more time with friends. They were also more likely to be engaged in music, movie watching, social media, and activities involving alcohol and drugs. Those who were highly engaged in activities such as volunteering and reading or writing showed less risky driving behaviours. With respect to personality trait, our findings also showed that high-risk drivers were more likely to be those with a high level of neuroticism, high level of extroversion, and low levels of imagination. It is important to understand the relationship between leisure activities and risky driving behaviours as it could help identify factors that influence unsafe driving practices in young drivers and aid in the design of future injury prevention programs.             v  Preface  This dissertation is original, unpublished and independent work by the author Vahid Mehrnoush. I came up with, nurtured, and implemented the idea, drafted the protocol, recruited the youth advisory committee (24 young folks), held focus groups, developed and validated leisure activity questionnaire, deployed the survey on Facebook, collated and analyzed the data, and drafted the final manuscript. The interim data analysis and final result were presented by the author, Vahid Mehrnoush, at Trauma Association of Canada (TAC) 2019 and 2020, respectively. The fieldwork reported in Chapter 2 was covered by University of British Columbia Behavioural Research Ethics certificate number (H17-00086).        vi  Table of Contents  ABSTRACT ........................................................................................................................................................................ iii LAY SUMMARY ............................................................................................................................................................... iv Preface ................................................................................................................................................................................. v Table of Contents ................................................................................................................................................................ vi List of Tables .................................................................................................................................................................... viii List of Figures ...................................................................................................................................................................... x List of Abbreviations .......................................................................................................................................................... xi Acknowledgements ............................................................................................................................................................ xii Dedication ......................................................................................................................................................................... xiii 1. INTRODUCTION ........................................................................................................................................................... 1 1.1 THE HEALTH AND ECONOMIC BURDEN OF MVCS AMONG YOUTH IN CANADA 1 1.1.1 Mortality burden ................................................................................................................................................ 1 1.1.2 Economic burden: Emergency department visits/Hospitalization .................................................................... 1 1.2 MVCS CONTRIBUTING FACTORS 2 1.2.1 Environmental factors ....................................................................................................................................... 4 1.2.2 Vehicle factors .................................................................................................................................................. 4 1.2.3 Human factors: The leading cause of MVCs .................................................................................................... 4 1.2.3.1 Risky driving behaviours as the most common human factor .................................................................. 5 1.2.3.2 Definition of risky driving behaviour ........................................................................................................ 6 1.2.3.3 Physiology of risk taking in youth ............................................................................................................ 6 1.2.3.4 Contributing factors to risky driving behaviour in youth .......................................................................... 6 1.2.4 Psychosocial theories for risky driving behaviour in youth .............................................................................. 8 1.2.4.1 Problem Behaviour Theory ....................................................................................................................... 8 1.2.4.2 Social Learning Theory and Social Cognitive Theory ............................................................................ 10 1.2.4.3 Theory of Planned Behaviour .................................................................................................................. 10 1.2.4.4 Evidence on how social norms may lead to risky driving behaviour in youth ....................................... 11 1.2.4.5 Evidence on risk perception and risky driving behaviour in youth ......................................................... 12 1.2.5 Theoretical Framework for the current study- Why Leisure Activity matters ................................................ 13 1.3 LEISURE ACTIVITIES AND GENERAL RISK-TAKING BEHAVIOUR 17 1.3.1 Leisure activity definition ............................................................................................................................... 17 1.3.2 Influence of Leisure Activities on Behaviour ................................................................................................. 17 1.3.2.1 General influence of leisure activities on youth ...................................................................................... 18 1.3.2.2 Role of leisure activities in instigating or controlling risk-taking behaviours ........................................ 18 1.3.3 Leisure activities and risky driving behaviour ................................................................................................ 19 1.3.3.1 Spending time with parents (Parental influence on driving behaviour) .................................................. 19 1.3.3.2 Peer Influence on Driving Behaviour ...................................................................................................... 20 1.3.3.3 Playing video games ................................................................................................................................ 20 1.3.3.4 Watching Movies and Television (TV) ................................................................................................... 21 1.3.3.5 Alcohol- or drug- involved activities ...................................................................................................... 21 1.3.3.6 Listening to music ................................................................................................................................... 22 1.3.3.7 Social media ............................................................................................................................................ 23 1.3.3.8 General lifestyle ...................................................................................................................................... 23 1.4 PERSONALITY TRAIT AS A MAJOR CONFOUNDER AFFECTING BOTH DRIVING BEHAVIOUR AND LEISURE ACTIVITIES 24 1.4.1 Role of personality traits in risky driving behaviour ...................................................................................... 24 1.4.1.1 Personality trait models (other than Big Five Model) and risky driving behaviour ................................ 25 1.4.1.2 Big Five Model of Personality traits and risky driving behaviour .......................................................... 25 1.4.2 Personality trait and leisure activity ................................................................................................................ 26 1.4.2.1 Personality trait models (other than Big Five Model) and leisure activities ........................................... 26 vii  1.4.2.2 Big Five Model of Personality traits and leisure activities ..................................................................... 26 1.5 SUMMARY AND KNOWLEDGE GAP 27 1.6 STUDY AIM, OBJECTIVES, AND HYPOTHESIS 28 1.6.1 Aim .................................................................................................................................................................. 28 1.6.2 Objectives ........................................................................................................................................................ 28 1.6.3 Hypotheses ...................................................................................................................................................... 28 1.6.4 Study questions ............................................................................................................................................... 28 2. METHODOLOGY ........................................................................................................................................................ 30 2.1 STUDY DESIGN AND TARGET POPULATION 30 2.2 ENGAGING A YOUTH ADVISORY COMMITTEE (YAC) 30 2.3 SURVEY INSTRUMENTS 31 2.3.1 Basic sociodemographic questionnaire ........................................................................................................... 31 2.3.2 Behaviour in Young Novice Driver Scale (BYNDS) ..................................................................................... 33 2.3.3 Leisure time activity questionnaire ................................................................................................................. 35 2.3.4 Mini-IPIP Questionnaire, measuring personality traits ................................................................................... 38 2.3.5 Canada Young Driver Survey (the entire survey) ........................................................................................... 40 2.4 SURVEY DEPLOYMENT 40 2.4.1 Data management and Ethical Consideration ................................................................................................. 41 2.4.1.1 Data collection ......................................................................................................................................... 41 2.4.1.2 Data confidentiality ................................................................................................................................. 42 2. 5 ETHICAL CONSIDERATION 42 2.6 DATA ANALYSIS 42 2.6.1 Variables ......................................................................................................................................................... 43 2.6.2 Statistical analysis plan ................................................................................................................................... 43 2.6.2.1 Descriptive Analysis ................................................................................................................................ 44 2.6.2.2 Proportional Odds Logistic Regression analysis ..................................................................................... 44 2.7 SAMPLE SIZE CONSIDERATIONS 46 3. RESULTS ...................................................................................................................................................................... 47 3.1 GENERAL DESCRIPTION OF STUDY POPULATION 47 3.1.1 Sociodemographic characteristics in study participants ................................................................................. 47 3.1.2 Personality traits in the study population ........................................................................................................ 48 3.1.3 Level of engagement in leisure activities in the study participants ................................................................. 48 3.1.4 Driving behaviour profile in study population ................................................................................................ 48 3.2 CHARACTERISTICS OF DRIVING BEHAVIOURS AND CORRESPONDING DIFFERENCES AMONG DRIVER GROUPS 49 3.2.1 Sociodemographic characteristics in driver categories ................................................................................... 49 3.2.1.1 Association between sociodemographic and driving behaviours ............................................................ 49 3.2.2 Personality trait in driver categories ............................................................................................................... 51 3.2.2.1 Association between personality traits and driving behaviour ................................................................ 51 3.2.3 Level of engagement in leisure activities in the driving categories ................................................................ 52 3.2.3.1 Association between leisure activities and driving behaviour ................................................................ 52 3.3 PROPORTIONAL ODDS LOGISTIC REGRESSION (POLR) ANALYSIS 54 3.3.1 Full Model ....................................................................................................................................................... 54 3.3.1.1 Predictive variables in positive correlation with risky driving behavior- Full Model ............................ 54 3.3.1.2 Predictive variables with protective effect on risky driving behavior- Full Model ................................ 55 3.3.1.3 Final Full Model Equation ...................................................................................................................... 56 3.3.2 Parsimonious model (Stepwise AIC selection model) .................................................................................... 57 3.3.2.1 Predictive variables in positive correlation with risky driving behavior- Parsimonious Model ............. 57 3.3.2.2 Predictive variables with protective effect on risky driving behavior- Parsimonious Model ................. 58 3.3.2.4 Final Parsimonious Model Equation ....................................................................................................... 59 4. DISCUSSION ................................................................................................................................................................ 62 4.1 SOCIODEMOGRAPHIC FACTORS AND DRIVING BEHAVIOUR 62 4.1.2 Age and driving behaviour .............................................................................................................................. 62 4.1.2 Gender and driving behaviour ......................................................................................................................... 64 viii  4.1.3 Employment and driving behaviour ................................................................................................................ 66 4.1.4 Living status and driving behaviour ................................................................................................................ 66 4.1.5 Education and driving behaviour .................................................................................................................... 68 4.1.6 Driving License class (with or without restriction) and driving behaviour .................................................... 69 4.1.7 Duration of driving independently and driving behaviour .............................................................................. 70 4.1.8 Driving exposure and driving behaviour ......................................................................................................... 71 4.1.9 Driver car status and driving behaviour .......................................................................................................... 72 4.1.10 Socioeconomic status and driving behaviour ................................................................................................ 73 4.2 PERSONALITY TRAITS AND DRIVING BEHAVIOUR 74 4.2.1 Extroversion and driving behaviour ................................................................................................................ 74 4.2.2 Neuroticism and driving behaviour ................................................................................................................. 76 4.2.3 Agreeableness and driving behaviour ............................................................................................................. 77 4.2.4 Conscientiousness and driving behaviour ....................................................................................................... 78 4.2.5 Imagination (Openness) and driving behaviour .............................................................................................. 80 4.3 LEISURE ACTIVITY AND DRIVING BEHAVIOUR 81 4.3.1 Video game and Driving Behaviour ............................................................................................................... 82 4.3.2 Sport and Driving Behaviour .......................................................................................................................... 84 4.3.3 Social media and Driving Behaviour .............................................................................................................. 85 4.3.4 Watching Movie engagement and Driving Behaviour .................................................................................... 87 4.3.5 Family-involved activity and Driving Behaviour ........................................................................................... 88 4.3.6 Friend-involved activity and Driving Behaviour ............................................................................................ 90 4.3.7 Alcohol-involved activity and driving behaviour ........................................................................................... 91 4.3.8 Drug-involved activity and driving behaviour ................................................................................................ 93 4.3.9 Art and driving behaviour ............................................................................................................................... 94 4.3.10 Reading/Writing and driving behaviour ........................................................................................................ 95 4.3.11 Volunteering and Driving Behaviour ............................................................................................................ 96 4.3.12 Music and Driving Behaviour ....................................................................................................................... 97 4.3.13 TV and driving behaviour ............................................................................................................................. 98 5. CONCLUSION .............................................................................................................................................................. 99 5.1 LIMITATIONS                        99 5.2 IMPLICATIONS 102 5.3 FUTURE RESEARCH 103 BIBLIOGRAPHY ............................................................................................................................................................ 105 APPENDIX A .................................................................................................................................................................. 121 APPENDIX B .................................................................................................................................................................. 126 APPENDIX C .................................................................................................................................................................. 131 APPENDIX D (CANADA YOUNG DRIVER SURVEY- SAMPLE) ........................................................................... 137            ix  List of Tables Table 1.1 MVC-related fatalities, serious injuries, and total injuries in Canada, 2015 .................. 1 Table 1.2 Leisure activity categories ............................................................................................ 16 Table 2.1 Sociodemographic variables, options in the survey, and final categories .................... 32 Table 2.2 Driving behaviour categories based on the composite score of BYNNDS .................. 34 Table 2.3 Questionnaire Changes on Young Novice Drivers Scale (BYNDS) ............................ 34 Table 2.4 Leisure activities and respective level of engagement ................................................. 37 Table 2.5 Mini-IPIP measures five personality traits ................................................................... 39 Table 2.6 Canadian Young Driver Survey .................................................................................... 40 Table 2.7 Variables and categories ............................................................................................... 42 Table 3.1 Socio-demographic characteristics of study population ............................................... 50 Table 3.2 Distribution personality traits in the study population ................................................. 51 Table 3.3 Distribution of study population according to Leisure activity .................................... 53 Table 3.4 Factors with significant influence on risky driving behaviour- Full model ................. 55 Table 3.5 Factors with protective influence against risky driving behaviour- Full model ........... 56 Table 3.6 Factors with significant influence on risky driving behaviour- Stepwise AIC ............ 58 Table 3.7 Factors with protective influence on risky driving behaviour- Stepwise AIC ............. 58 Table A.1 Distribution of driving license classes in driving behaviour categories .................... 117 Table A.2 Leisure activity subtypes ........................................................................................... 119 Table A.3 Distribution of study population according to leisure activity (Sub type)………... ..120 Table B.1 Bivariate Correlation ………………………………………………………………  121  Table B.2 Full model variable .................................................................................................... 122 Table B.3 Stepwise AIC selection model ................................................................................... 123 Table B.4 Full versus Stepwise selection Model ........................................................................ 124  x  List of Figures Figure 1.1 The System Approach- World Health Organization’s framework for the elements of road and transport system ………………………………………………………………………….…..3 Figure 1.2 Factors influencing youthful driving behaviour……………………………………… 7  Figure 1.3 Problem Behaviour Theory ........................................................................................... 9 Figure 1.4 Detailed conceptual theory. ......................................................................................... 14 Figure 1.5 Theoretical framework of project ................................................................................ 15 Figure 1.6 Simple conceptual diagram of project. ........................................................................ 28 Figure 2.1 Construct relationship diagram (Analysis perspective) ............................................... 42 Figure 3.1 Survey participation rate ............................................................................................. 47 Figure 3.2 Driver category distribution in the study population ................................................... 48 Figure 3.3 Percentage Distribution of high engagement in leisure activities in participants         52 Figure 3.4 Forest plot - Full vs. Stepwise selection Models. ........................................................ 60 Figure C.1 Percentage Distribution of Social Media Subtypes in drivers’ groups ..................... 126 Figure C.2 Percentage Distribution of Family Activity Subtypes in drivers’ groups ................. 127 Figure C.3 Percentage Distribution of Art Activities’ Subtypes in drivers’ groups ................... 129 Figure C.4  Percentage Distribution of Sociodemographic Characteristics  .............................. 130 Figure C.5 Percentage Distribution of high level of Personality Traits in drivers’ groups ........ 131    xi  List of Abbreviations AB: Alberta  BAC: Blood Alcohol Concentration BC: British Columbia BYNSD: Behaviour of Young Novice Drivers Scale GDL: Graduate Driving Licencing  IPIP: International Personality Item Pool LRT: Likelihood Ratio Test MB: Manitoba MVCs: Motor Vehicle Crashes NB: New Brunswick  NL: Newfoundland and Labrador  NT: Northwest Territories  NS: Nova Scotia  NU: Nunavut  ON: Ontario  PE: Prince Edward Island  QC: Quebec RSS: Road Safety Strategy  SK: Saskatchewan  UBC: University of British Columbia WHO: World Health Organization YT: Yukon           xii  Acknowledgements  Firstly, I would like to express my sincere gratitude to my mentor and supervisor Dr. Jeff Brubacher who has been accompanying me like friend throughout this project for the continuous support, his patience, inspiration, and immense knowledge that incented me to broaden my perspective and think outside the box.  I would also like to thank Dr Herb Chan, my supervisory committee member, who always advised me to be focused, for his insightful comments and open-to-question attitude which did never let me feel unsupported.  I would like to render my thanks to Dr. Mariana Brussoni, my supervisory committee member, for challenging me in a constructive fashion which encouraged me to form a comprehensive and objective critique and to extend my learning to more depth. With many thanks to Shannon Erdelyi, our team statistician, for welcoming my questions and simplifying the statistics points I had.   My sincere thanks also go to Youth Advisory Committee members (below) for their collaboration with the project. Without their valuable contribution, this project would have not never been possible to be conducted in such an impressive method. I would also like to thank Mr. Mehrdad Abkenari and Allen Najafi for filming and editing the advertisement video. I cannot thank enough my lovely family, i.e., my parents, sister, brothers, and my beautiful niece, for unconditional, relentless emotional and spiritual support all through my master’s period and whole life.  Last but not least, I would like to send my heartfelt thanks to my precious friends Dr. Fatemeh Darsareh and Dr. Mina Salehi for their incredible generosity of helping me with this accomplishment. Moreover, I would like to thank the following treasured friends Dr. Shahrzad Keramati, Dr. Ediriweera desapriya, and Ali Tafvizian for their invaluable support.   Youth Advisory Committee members: Amitis Ebrahimi, Omid Ghahremani, Arshia Hamedi, Shadi Meraji, Fatemeh Kiannejad, Aki Ediriweera, Amir Sarrafan, Ali Hosseinzadeh, Vania Rezaei, Kimia Danesh, Yasmin Rajwani, Nicholas Huen, Ava Keshavarz, Hossein Khazaie, Arash Khazaie, Saman Darabian, Daniel Almasizadeh, Amirali Mallakzadeh, Ayda Danesh, Kelvin Chan, Reihaneh Jamalifar, Anita Borhani, Lale Hpm. and Bita Jokar. xiii  Dedication I dedicate this work to the next generations of drivers, all young drivers, all over the world.  I hope that no family or individual suffers from any burdens of motor vehicle crash in near future.                             1  1. Introduction  1.1 The health and economic burden of MVCs among youth in Canada 1.1.1 Mortality burden   Globally, MVCs are the leading cause of death in 15-29 year-olds (1). Forty young people die of MVCs each hour, and many more sustain disabilities or emotional scars which they will carry for the rest of their lives (2). In Canada, MVCs are among the leading causes of hospitalization, premature death and potential years of life lost for youth (3). Young drivers between 15 and 24 years old in Canada are overrepresented in fatalities and serious injuries involving vehicle crashes (4,5). In 2015, this group accounted for 18.5% of driving fatalities and 21.7% of serious injuries, despite being only 12.3% of all licensed drivers in Canada. (Table 1.1)   Table 1.1 MVC-related fatalities, serious injuries, and total injuries in 15-24 years old and all ages, Canada, 2015 Age Licensed Drivers (%) Fatalities (%) Serious Injuries (%) Total Injuries (%) 15-24 3,103,514 (12.3) 343 (18.5) 2,228 (21.7) 33,922 (21) All ages 25,272,915 (100) 1858 (100) 10280 (100) 161902 (100) Transport Canada Annual Report: 2015 (4)  1.1.2 Economic burden: Emergency department visits/Hospitalization In addition to personal injury health burden, the economic burden associated with MVC-related injuries weighs heavily on Canada’s health care system (6,7). In 2012 MVCs caused 165,172 injuries costing billions of dollars, particularly related to healthcare cost and productivity losses due to serious injuries requiring hospital admission (7). Parachute Canada (4) reported that in 2010, transport related incidents, after falls, and were the second most common cause of injury-related emergency room visit, hospitalization, and disability. Of transport incidents, motor vehicle collisions constitute the majority of deaths (43%), hospitalized treatment (51%), emergency room visits (56%), and partial (i.e., a permanent disability with partial recovery), and total (i.e., complete and permanent loss of earning power) disabilities (49% and 52%, respectively). Transport injuries ranked third in direct costs of injury (i.e., health care costs arising from injuries) at $2.1 billion (13%), after falls and other unintentional injuries. In addition, transport incidents, regardless of the mode of transport, were the 2  major cause of indirect costs of injury (i.e., costs attributed to reduced productivity, disability, and premature death) in 2010, accounting for $2.1 billion or 20% of total indirect all-injury costs. Motor vehicle collisions imposed the greatest single societal cost (included both economic burden and valuation of quality of life lost due to MVCs) totalling $2.2 billion. They also made up 50% of all transport-related incidents. In addition, motor vehicle collisions had the highest costs when considering both direct and indirect costs, $1.2 billion and $987 million, respectively. In terms of the health burden, MVC-related injury in young people aged 15-19 years old and 20-24 years old accounted for the highest rates of transport-related hospitalization. Young people (ages 15-19 and 20-24 years old) had the first and second highest rates of transport-related emergency room visits and transport-related permanent partial disabilities (4).   Need for root cause analysis Canada’s Road Safety Strategies (RSS) 2025 identifies young and novice drivers (under the age of 25 years) as one of the key risk groups that should be targeted to improve road safety. RSS 2025 highlighted the fact that recognizing MVC contributing factors is critically important in order to set effective initiatives to improve the safety of this target group and enhance road safety in general (8). Factors such as human initiated (e.g., impaired driving, speed and aggressive driving), vehicular (e.g., occupant protection measures), and environmental (roadway configuration, roadway construction, road surface condition, road and roadside design, weather conditions) (8). 1.2 MVCs contributing factors A variety of frameworks and approaches have been applied to identify the risk factors for road traffic injuries (9-17), including:  a) Safe System Approach (SSA) is an Australian proposed framework to help countries to devise effective road safety strategies to eliminate fatalities and severe injuries. This approach is grounded on three principles: 1) Despite all measures, human fallibility of a driver is acceptable to happen but not at considerable cost to other road users’ lives or to the system. 2) Roads, roadsides and vehicles need to be designed to minimise crashes. 3) Road safety requires shared accountability and responsibility. Accordingly, some target driver groups especially young or novice drivers and some key contributing factors (e.g., speed and aggressive driving, distracted driving, drug and alcohol impaired driving, fatigue impaired driving) were prioritized as topics to address in SSA (12).  3  b) Systems Approach is adopted by WHO conceptualizes the road traffic system as a whole, designs the basic elements of transportation system, and interlinks the major risk factors in transport and road system. This approach helps the road traffic injury prevention strategies with identifying the potential contributing factors for risky driving behaviour and their interactions (13) (Fig 1.1). c) Public Health Approach provides a four-step process that helps to guide how to approach health-related issues like MVCs:1) Identify the size of the health burden, 2) Identify risk factors and causes, 3) Discern the preventive measures, and then 4) Implement prevention programs on a large scale and evaluate these programs. This approach is promoted by the WHO (14) and the U.S. National Center for Injury Prevention and Control, Atlanta (15). d) Haddon Matrix is a commonly used conceptual framework in injury prevention which serves to identify the influencing factors at different phases of an injury (Pre-event, event, and post-event) (16).  Figure 1.1 Schematic representation of the system approach (adapted from the World Health Organization’s framework for the elements of road and transport systems).  Risk factors for road traffic crashes fall under three major categories: human factors, vehicle factors, and environmental factors. Mobility for different reasons, such as leisure, is a desired output but it can interact with crash factors. The original WHO draft included three components (desired outputs, undesired outputs. In our modified version we removed the “other” component which is irrelevant to this topic (12).                                  Road and EnvironmentRoad UsersVehicleMobility Road traffic crashes Work Others School Leisure Shopping Road and environmental factors Human factors Vehicle factors Undesired outputs Desired outputs Purpose of trips Crashes and crash factors 4  According to these frameworks, and based on literature (18,19), contributing factors for MVCs broadly fall into three categories: (1) Environment factors; (2) Vehicle factors; (3) Human factors 1.2.1 Environmental factors Physical environmental factors include road infrastructure (such as undivided roadways, road curvature and inclination, road surface condition, merge and exit points, and bridge structures), lighting, weather conditions, and visibility of objects. The role of safety-oriented road engineering in reducing the frequency and severity of road traffic crashes is undeniable. Conversely, unsafe road networks can contribute to crashes because of unsafe road design. For example, misleading road environment elements (i.e., they are designed in a way that the driver is not aware of what is expected of them and does not behave appropriately) or signs which prompt error, or where some changes to the road would have attenuated the risk of crash. Moreover, the road environment may also determine how road users perceive their environment which, in turn, affects the crash risk (20). That is to say, road signs and traffic controls guide road users’ actions. Therefore, road safety engineering measures also work through influencing human behaviour. 1.2.2 Vehicle factors  Although the crashworthiness of private vehicles has improved, further advancements are possible. Vehicle factors related to design (e.g., crash avoidance, crashworthiness), maintenance, recalls, aftermarket vehicle modifications, commercial vehicles, unusual vehicles, automated vehicles, and new and emerging vehicle technologies play crucial roles in not only preventing MVCs but also in determining the severity of MVC-related injuries (i.e., occupant contact with the vehicle’s interior, or other inadequate vehicle protection safety standards) (21).  1.2.3 Human factors: The leading cause of MVCs Human factors include the driver’s mental and physical capacity, amount of driving training and level of experience, adopted driving style and driving errors. Social circumstances like cultural and social norms, socioeconomic status, and peer pressure to take risks can also influence driver behavior. Generally, human factors fall into the human condition (such as alcohol or drug involvement; extreme fatigue), and human actions (behaviours) (e.g. driving without due care; speeding; failing to signal; improper passing (22-24). Driving behaviours are the major constituent of the human factors. Rumar et al. stated in his book that human factors, such as driving behaviour, are involved in 90 percent of 5  crashes (25). Despite the focus on vehicle safety in traffic crashes, human factors such as risky driving behaviours are more important (21,26). Road trauma prevention strategies that utilize knowledge about human factors in conjunction with engineering and enforcement policy can improve crash reduction effectiveness (20).  Nabi et al., in a study on 13,447 French drivers found that MVCs decreased by changing drivers’ negative attitudes towards driving laws by including posted speed limits or warning signs regarding impaired driving on roads (27). The International Traffic Safety Data and Analysis Group (IRTAD) annual report (2017), which contains the most up-to-date road safety data, proposed measures, and strategies for 40 countries (28). For Canada, identified behaviours of road users included speeding, impaired driving, distracted driving, driver fatigue, and a lack of seatbelt use as significant contributing factors to road crash fatalities all of which can be proposed as focus points for injury prevention strategies (29).  The major topics identified in Canada’s Road Safety Strategy are the key risk groups (e.g., young drivers, high risk drivers, vulnerable road users, and medically-at-risk drivers) and the key contributing factors (i.e. speed and aggressive driving, alcohol and drug impaired driving, distracted driving and fatigue impaired driving) (6). Collectively these topics are aimed at capturing the human factors especially associated with risky driving behaviour (6). MVCs can result from human factors alone or in combination with environmental and/or vehicle factors. It is unclear what proportion of total MVCs are caused by each factor (i.e. environmental, vehicle, and human). However, human factors have been deemed to be the leading determinants both in the general population and among young drivers (21-23). In addition, it is worthwhile considering that even simple behavioural modifications such as increasing knowledge about simple traffic rules (commitment to the speed limit, compliance with seat belt use, not use of a mobile phone, and respect for traffic signals) can impact road safety (4). For instance, a 1% reduction in speed results in reducing the likelihood of a fatal collision by 5% (9). 1.2.3.1 Risky driving behaviours as the most common human factor  Risky driving behaviour are considered to be among the more significant human factors resulting in MVCs (22,30). Fergusen et al. stated that risky driving behaviours are generally more prevalent in younger drivers (16-24 years old) and attributed most MVCs in youth to reckless and risky driving behaviour (30). Rhodes et al., in a phone survey of 504 drivers aged 16-20 years old and 409 drivers aged 25-45 years old showed that the younger drivers were more risky drivers (31). In Canada, risky 6  driving behaviour is the leading cause of MVC related injuries and deaths in youth (5,32). In other words, risky driving behaviours (e.g., speeding, driving under influence of alcohol and/or drugs, and distracted driving) are the leading contributing factors in MVCs especially in young drivers. Identifying the factors influencing risky driving behaviours can help understand root cause. 1.2.3.2 Definition of risky driving behaviour By definition, risky driving behaviours are those actions undertaken by drivers which expose themselves, their passenger(s), and other road users such as pedestrians, cyclists, drivers, and passengers in other vehicles to hazardous situations, harm, or fatal injury (33). Examples include traffic law infringements such as speeding, driving under the influence of alcohol or drugs, cellphone distracted driving, tailgating, wearing no seatbelt, improper passing, inappropriate lane usage, right-of-way violation, illegal turns, and control sign violations. Risky driving behaviour can be either deliberate, regulations violations, or unintentional, such as errors and distractions (34). 1.2.3.3 Physiology of risk taking in youth As noted above, in addition to a limited knowledge of regulation rules and a lack of driving experience (and the associated acquired skills), driver physiology and development also plays role. From a psycho-physiology perspective, incremental changes in physical, cognitive, emotional and social development, experimenting, and risk-taking typify adolescence. Generally, changes in patterns of dopaminergic activity around the time of puberty are linked to increases in sensation seeking followed by an increase in risk-taking in adolescence (35). Along with understanding the physiological foundation, a full understanding of specific risky behaviours leading to injury or harm is critical. Gielen et al. highlighted that the behavioural approach, combined with environmental modifications, is integral to successful injury prevention strategies. To that end, they recommend that injury prevention researchers apply a conceptual framework that addresses a specific situation (e.g. motor vehicle crash), specific target group (e.g. young population), and the setting and the characteristics of the target behaviour (e.g. risky driving) (36). 1.2.3.4 Contributing factors to risky driving behaviour in youth To come up with a solid approach to risky driving behaviour in youth, Shope et al., proposed a framework that included the aforementioned constructs and determinants influencing driving behaviour in young people, namely, perceived environment (i.e. parents norm, peer norm, community 7  norm, cultural norm, media and risk perception) personality characteristics (i.e. risk taking propensity, aggressiveness, susceptibility to peer pressure, and tolerance of deviance), driving ability (i.e. knowledge, skill, and experience), driving environment (i.e. weather, vehicle, and passenger), demographic factors (age, sex, education, employment status, SES), and developmental factors (i.e. physical, psychological, behavioural) (Figure 1.2). They defined the perceived environment, as the driver’s perception of driving behaviour norms, the driver’s attitude, and the driver’s perception of risk in the driving environment. According to Shope, the perceived environment, is perhaps the most complex category and plays a crucial role in determining risky driving behaviour in young people (23). Perception of driving behaviour norms and risk perception in a driving setting (i.e. driving tasks and conditions) is formed through interaction with the environment i.e. parents, peers, culture, community, and media throughout life. In other words, the perceptions of what driving behaviour is normal or risky, or what driving behaviour is socially disapproved or accepted, spring from major reference groups including parents, peers, and broader sources such as community, culture, and media (37). A body of psychosocial theories can be employed to uphold the influence of the perceived environment on driving behaviour (38-42).    Figure 1.2 Factors influencing youthful driving behaviour All contributing factors in risky driving behaviour in youth fall under the following themes: developmental factors, personality characteristics, driving ability, demographic factors, driving environment, and perceived environment. Shope et al., believed that the “Perceived environment” is perhaps the most complex one (23).  8  1.2.4 Psychosocial theories for risky driving behaviour in youth  Risky driving in youth can be approached by different psychosocial theories such as problem behaviour theory, social cognitive theory, social learning theory, and the theory of planned behaviour on which a great deal of psychosocial and behaviour-related literature have been founded.   1.2.4.1 Problem Behaviour Theory  Problem Behaviour Theory is a multifaceted psychosocial conceptual framework. This framework explains those specific variations in adolescents' behaviours which are considered problematic and undesirable by the norms of conventional society (38). These problematic and undesirable behaviours elicit some form of social sanctions and disapproval in society. Succinctly, this theory postulates that all behaviours result from the interaction between individuals and their environment. Problem behaviour arises from interaction within and across the following five major systems: (Figure 1.3)  • Perceived environment system: indicates that friends’ and/or parents’ expectations, norms, and behaviours can influence youth. Moreover, parents’ and peer’s control and support models can prohibit or promote problem behaviours in youth. Accordingly, an adolescent is more likely to engage in a problem behaviour when they perceive their parents’ and/or peer’s norms, expectations, and behaviours are in keeping with and underpin that problem behaviour. In the case that the foregoing parents’ and peer’s factors are incompatible, i.e. one supports while the other discourages that problem behaviour, youth acknowledge greater influence and support from their peers relative to their parents.  • Personality System: suggests that intrinsic factors within an individual can contribute to problematic behavior. The personality system includes personal values, beliefs, expectations of achievement, attitude and level of intolerance towards deviant behaviours, self or social orientation, self-efficacy, locus of control (individual’s belief about the extent they have control over what occurs to them) which are mainly rooted in social learning and developmental experience.  • Behaviour System:  includes both behaviors that contribute to problem behavior (unconventional behavior) and socially approved behaviours (conventional behaviour). This system specifies that an adolescent who engages in one risky behaviour will be likely to engage in more risky behaviours. Influential behaviours include experimental or regular use of substances, risky sexual activity, and irresponsible driving 9  • Social environment: includes some variables such as parental ideology (maternal or paternal traditional beliefs and religiosity), home climate, peer influence, and media influence. • Biology/Genetics: includes gender differences, family history of risky behaviours such as parents who regularly overuse alcohol.  Each system includes variables categorized as instigations (factors that increase the likelihood of engaging in problem behaviours) and controls (factors that decrease the likelihood engaging in problem behaviours). The overall balance between instigations and controls within and across all five systems drives an individual to perform or to avoid a problematic behaviour. Moreover, this theory states that adolescence is a period of time when they are prone to form risky habits. In this period, also called “transition proneness”, adolescents mimic adults’ behaviours, which are typically not appropriate for adolescence, such as unprotected sex, smoking, substance abuse, and binge alcohol drinking, and driving under the influence of alcohol or drugs. Adolescents perceive that performing such “adult” behaviour would transition their status from “young” or “less mature” to “adult” or “more mature”. In other words, adolescents engage in these activities to show others that they are not “children” or “immature” anymore (38,39).  Figure 1.3 Problem Behaviour Theory.  According to this theory, a problem behaviour springs from the interaction of five major systems. Each system consists of risk factors which promotes the problem behaviour and protective factors which help to avoid that particular behaviour (39).             Social environment Personality Behaviour Adolescent risk behaviour Problem behaviour Health-related behaviour School behaviour   Biology /Genetics Perceived environment 10  1.2.4.2 Social Learning Theory and Social Cognitive Theory  Social Cognitive Theory (SCT) which evolved from Social Learning Theory (SLT) and has been among the most commonly used theories in the health behavior intervention literature. SCT and SLT hold that both psychological factors and the social environment can play a role in the development of behaviour (40). The theories posit that behaviour results from a learning process which occurs in a social context.  SLT and SCT have five common communal constructs, while one more construct, i.e. self-efficacy was added to SLT as it evolved from SCT. Reciprocal Determinism: The main concept of the theories which refers to the mutual interaction and influence of personal factors (e.g. cognitive skills, attitude, a set of learned experience) and social environment.  1- Behavioural capability (self-control): An individual has to know what and how to do it. It asserts the actual ability to perform a behaviour 2- Observational Learning: An individual can reproduce a behaviour by observing or witnessing it performed by others in their social environment 3- Reinforcements: The internal or external responses to a behaviour that can prompt a person to continue or refrain from it. In other words, reinforcement can be positive or negative.  4- Expectation and outcome expectancy: The perception of others’ expectation and the belief about the perceived consequences and outcomes of a behaviour are what an individual takes into account prior to engaging in that behaviour.  5- Self-efficacy: The individual’s belief about or level of confidence in one’s ability to successfully perform or master a behaviour (40).  1.2.4.3 Theory of Planned Behaviour   The Theory of Planned Behaviour by Icek Ajzan (1985) (40) is a psychosocial framework designed to predict and explain individual’s behaviour in particular context. This theory posits that the major determinant of a behaviour is the intention to perform that behaviour which is under the influence of and tightly related to: • Attitude, which is an individual belief of a certain act or behaviour. That is to say, attitude reflects on an individual’s belief about the consequence and how they evaluate that outcome which, in turn, leads to approval or disapproval of that behaviour by themselves. 11  • Subjective norm, which focuses on the individual’s social network group such as family, peers, culture, community, and society, and what those reference groups consider and appropriate, and what leads to approval or disapproval.  • Perceived behavioural control, which expresses a person’s belief on how easy or hard it is to display certain behaviours. It reflects an individual’s perception of their skills related to whether they could handle or manage that behaviour.  Ajzen believed that it is essential to take these themes into account to understand a behaviour and to implement any intervention to change a behaviour (40).  1.2.4.4 Evidence on how social norms may lead to risky driving behaviour in youth The foregoing psychosocial theories spotlight the role of social environments, especially the reference groups (i.e. important others). The main types of social norms that have been distinguished in the literature are descriptive norms and injunctive norms (41,42). Descriptive norms refer to perceptions of how commonly a behaviour is performed in the corresponding social environment, thereby demonstrating what behaviours are perceived as effective and adaptive. Injunctive norms refer to perceptions of what other people, especially reference groups, commonly approve or disapprove of, which can motivate to acquire/maintain or withhold or withdraw a particular behaviour through the promise of social approval/blending or sanctions, respectively (43). Of reference groups, parents and peers are two important social sources of influences on youth. A substantial body of literature provides evidence of social norm influences on youth driving behaviours (44-49). According to some studies, the influence of parent’s behaviour such as parental monitoring and restrictions on teen driving were inversely associated with risky driving behaviours among teens. Their results showed that those who exhibited higher levels of risky driving were more likely to report low parental monitoring and low parental restrictions on their driving (45,46). This can be justified by the control model of parents in the perceived environment system of problem behaviour theory or influence of subjective norm (disapproval) on behaviour according to the theory of planned behaviour (above). Carter et al., found that teens who perceived their parents to engage in driver distraction more frequently reported higher engagement in driver distraction themselves. Also, parents’ self-reported driver distraction engagement was positively associated with that of their teens (46). This can be explained by not only social cognitive theory, i.e. the construct of observational learning and mutual interactions of teen and social environment, but also by the subjective and descriptive norms influence 12  and support modeling of the perceived environment in theory of planned behaviour and problem behaviour theory, respectively.  Merrikhpour et al., surveyed 71 teens (aged 17-19 years)-parent dyads in Toronto, Canada, to understand the role of parental and peer norms in teen driver distraction, and to investigate the existence of social norm misperceptions among teens.  The results showed that teens’ perceptions of parent and peer norms were predictive of their self-reported distraction engagement. However, they found that the teens misperceived, i.e. overestimated, their parents’ and peer’s norms of distracted driving. Upon further experimental investigation, they showed that an intervention on the social norm correcting the corresponding misperception can positively affect driving behaviour in teens (48). Simons-Morton et al., showed that teens who reported to have more risk-taking friends (those who smoke cigarettes, drink alcohol, get drunk at least once a week, use marijuana, drive after having two or more drinks in the previous hour, exceed speed limits, and do not use safety belts) had significantly higher rates of crashes or near crashes and risky driving (19). Moreover, Allen et al., postulated that in the presence of peer passengers, teens were concerned with maintaining and strengthening their relationships with their peers rather than solely trying to drive safely (49).  Community, culture, and media as wider social sources can influence the perception of driving by presenting the modeling of how risky a driver can be; how likely it is to be injured or killed in a crash; how likely it is to get a ticket or be fined or even jailed for a particular driving rule infringement (23). The social environment system of Problem behaviour theory (above) can add support to such findings. 1.2.4.5 Evidence on risk perception and risky driving behaviour in youth Risk perception is a well-studied component of the perceived environment in youth. The level of risk perception varies from one individual to another and depends on factors such as past experience, driving training, age, and other personal and sociocultural factors (50,51).  Many previous studies found that “low” risk perception of driving is a significant predictor of risky driving behavior (45,52-55). However, evidence remains inconsistent. In a telephone survey study, Carter et al., questioned403 drivers aged 16-18 years old and their parents about their risk perception and distracted driving behaviour to examine the effect of social norms and risk perception on the youth’s driving behaviour. They found that low risk perception in youth, distracted driving behaviour in their parents, their perception of parents’ norm of distracted driving behaviour (i.e., youth thought 13  parent frequently drove while distracted), and perceived peer norms of distracted driving behaviour were all predictive of their own distracted driving behaviour (46). In a survey study on the effect of drivers’ risk perception and perception of driving tasks on road safety attitude., Ram et al., found that the risk perception was highly correlated with the perception of a specific driving task (e.g. speed adaptation according to the flow of traffic and judging the appropriate speed required when executing a safe pass manoeuvre according to the positions of following and upcoming vehicles. Both risk perception and perception of driving tasks were also correlated with the driver’s attitude toward traffic rules, driving responsibility (e.g. stopping the vehicle prior to a crosswalk, following in a traffic lane), and distracted driving. They concluded that the higher risk perception an individual had, the more likely they were to adopt safe driving behaviours (51). Likewise, Rhodes et al., surveyed 504 young adults (aged 16-20) and found that risk perception was a predictor of risky driving behaviour (31).  However, there is a fairly robust study contradicting the foregoing findings. Ulleberg et al., investigated the direct and indirect effects of personality trait, attitude and risk perception toward risky driving behaviour in 3942 Norwegian adolescents (52). They found that attitude had a significant direct impact on risky driving behaviour, personality trait exerted an indirect effect mediated through attitude, and risk perception was only a weak predictor of risky driving behaviour (54).  In conclusion, perception of normative driving behaviour and risk perception in driving can promote or discourage risky driving behaviours in young people (23). Moreover, attitudes towards traffic safety which arise from the knowledge of traffic rules (cognition), personal characteristics like respect and feeling towards other road users (affective), and the manner of driving (behavioural) can influence an individual’s driving behaviours that lead to a higher or lower tendency to engage in risky driving behaviours (52). All these factors (risk perception, perception of norms in driving setting, and attitude towards road safety) are greatly influenced by youths’ environments, which can influence acceptance of risky driving behaviours and lead to a higher or lower tendency to engage in risky driving behaviours (55). 1.2.5 Theoretical Framework for the current study- Why Leisure Activity matters As discussed above, family, peer groups, community, culture, and media can all influence risk perception, expectations and outcome expectancy, social (subjective and injunctive) norms, and attitude towards risky driving in young drivers. Approaches that consider the role of attitude, norms, and behavioural change are recommended to be incorporated in injury prevention strategies (23). 14  There is a need to understand the perceived environment more thoroughly given its contributing role in risky driving behaviour. To better understand how perceived environment influences driving behaviours in young drivers in Canada, it is useful to know how and to what extent (level of engagement) youth interact with their social environment (i.e., how they spend their daily life). According to the Statistics Canada General Social Survey (2010), leisure time related activities constitute almost 25.5 percent of young Canadians daily activity. Canadians aged 15-24 years old spend 6 hours and 8 minutes per day on leisure activities including “socializing”, “passive leisure” (watching TV, reading books, other passive leisure), and “Active leisure” (active sport, computer use, video games, other active leisure) (56). Thus, leisure time provides youth with the opportunity to be exposed to various elements of the perceived environment and is also a source for acquiring their perception of norms and risks in a driving setting (by interacting with family, peers, community, culture, society, and media). Figure 1.4 illustrates the conceptual theory describing the relationship between perceived environment and risky driving behavior in young drivers.  It is apparent that leisure time activities have both the sources (i.e, family, friends, community, and media) and the time (according to the aforementioned Statistics Canada report) (55,56) to influence driving behaviour of young drivers. Figure 1.5 illustrates the theoretical framework of current study explaining why and how leisure activities may influence driving behaviour. This framework is informed by the psychosocial theories described in section 1.2.4. From psychosocial standpoint, determinants of a behaviour are attitude, subjective norms, and behavioural control which are influenced by other factors such as behavioural belief, control belief, and normative belief. This framework shows the interaction of all factors and also where leisure activity (main independent variable of this study) can hypothetically exert its effect on risky driving behaviour. Leisure activities hypothetically influence risk perception, descriptive norms, social norms, and emotional self-regulation as described in Figure 1.5. According to Ulleberg et al., effects of personality traits are generally mediated through attitudes. Direct effect of attitude on risky driving behaviour is significant. Altruism is the only personality trait that has direct effect on risky driving behaviour. Risk perception has no direct effect on risky driving behaviour (52). (Fig 1.5) The following sections describe in detail the relationship between leisure activities and risky driving behaviours taking into consideration confounders such as socioeconomic status (SES), driving status, and personality traits.  15  Figure 1.4. Leisure activities play a role in perceived environment and risky driving behaviour.    Human factors are the primary causes in most in MVCs.  Risky driving behaviour, a human factor, is most often associated with young drivers. Perceived environment, i.e., behavioural norm and risk perception in driving setting, is a complex factor that elicits risky driving behaviour in youth through interactions with their parent, peers, culture, community and media. These interactions occur through leisure activities of youth which explain the link between the type of leisure activities and risky driving behaviour.                             Leisure Activities Perceived Environment Interaction with:  Peer, Family Community, Culture, Media,  Risky Driving Behaviours Supportive Theories  Problem Behaviour Theory Theory of Planned Behaviour   Social Learning Theory/ Social Cognitive Theory    16   Figure 1.5 Theoretical framework of the project   1 Leisure activity can exert its effect on risk perception, social norms, and emotional self-regulation whereby it can influence risky driving behaviour    ⇄  Personality Trait Attitude Behavioural control Subjective Norms Leisure activity Exposes adolescents and youth to the norms of reference groups (family, peer, community, culture, and society) Provides opportunities to learn how to respond with experience to a range of emotions in a social tolerable and sufficiently flexible manner, spontaneous reaction or delayed spontaneous reaction Risk perception Descriptive norms Social norms Emotional self-regulation    Risky Driving Behaviour:  Violation  Error Lapse Demographic variables Control beliefs Perceived power ⇄  11 Behavioural beliefs Evaluation of behavioural outcomes Normative beliefs Motivation to comply  17  1.3 Leisure activities and general risk-taking behaviour  1.3.1 Leisure activity definition Leisure activity refers to pleasurable and personally purposeful activities that one engages in during free time. Free time is defined as unobligated time, outside of work, school, or self-care activities usually in the evenings and on weekends. Leisure is often related to a sense of freedom and internal motivation (i.e., you engage in activities that you want to, not because you have to) (57). Youth engage in a variety of leisure activities such as individual and team sports, spending time with family and friends, socializing, volunteering, playing video games, watching TV and movies, social media, and so on. Leisure activities fall in two broad categories: structured and unstructured (Table 1.2). They can also be categorized into: (1) social activities (e.g, spending time with friends); (2) creative or expressive activities (e.g., artistic pursuits); (3) cognitive activities (e.g., reading); (4) spiritual activities (e.g., meditation) and (5) physical activities (e.g., walking, gardening) (58).   Table 1.2 Leisure activity categories.  Structured Unstructured Leisure or recreation activities that are typically deeply engaging (e.g., require an investment of attention and effort) and support personal expression “Doing nothing” or passive forms of activity that require low levels of engagement and often occur outside of organized recreation or leisure context Examples: volunteering, sports or club activities. This is sometimes also referred to as “active leisure.” Example: hanging out, watching television, listening to music, going to the mall, watching movies  1.3.2 Influence of Leisure Activities on Behaviour  Normative behaviours and expectations of parents and peers, cultural and community norms, and the media-transferred norms can be very influential in enabling youth to adopt or change behaviours and also establish or reform their perception of norms and risk. Youth spend a substantial amount of time on leisure activities and socialize with family, peers, and community members.  etc. (Section 1.2.4). They observe behaviour and perform similar behaviour or adjust their own behaviour to others’ (41,59).    18  1.3.2.1 General influence of leisure activities on youth  Leisure activities usually have positive impact on youth such as an increase in their social skills (60), they promote positive moods and emotions (61), enhance stress management (62), and help to develop psychosocial maturity and adjustment (63). An inherent characteristic of adolescence is being amenable to adopt or change a behaviour under the influence of leisure activities. Studies show that structured leisure activities (64) and level of satisfaction from leisure activities (65) affect physical health and mental well-being. Freire et al., found that engagement in leisure activities improved self-esteem, satisfaction with life, and psychological well-being in youth by helping them acquire and learn several developmental skills such as emotional control strategies and emotional self-regulation, i.e., the ability to respond to experience with socially acceptable and adequately flexible emotions (65). These strategies help individuals to apply cognitive skills to cope with stressful situations by reassessing their initial perception and attitude. As a result, individuals can learn how to re-interpret, modify, and reframe their perception of a situation or stimulus that elicits a behaviour reflecting negative emotional reaction (66). 1.3.2.2 Role of leisure activities in instigating or controlling risk-taking behaviours  Based on the psychosocial theories (Sections 1.2.4 and 1.2.5), certain leisure activities may either promote or prevent risk-taking behaviours. A great deal of literature has demonstrated the influence, either positive or negative, of different leisure activities on behaviour in young individuals (67-70). Darling et al., in a cross-sectional study on 3761 high school students in California, found that school-based extracurricular activities reduce the risk of substance abuse and increase grades and positive attitudes towards school (65). Studies on 228 Canadian and 3368 Norwegian youth (12-18 years old) found that organized leisure activities decrease problematic alcohol use (66). Budra et al., studied 10279 students aged 11 to 15 in 2017 and found that structured leisure activities such as sports (either individual or team) decreased risky behaviour such as smoking and drinking (68). Other studies show that young people who participate in pro-social activities such as volunteering and attending church were less prone to risk-taking behaviour like drinking alcohol (71-73).  On the other hand, there is also research suggesting that some leisure activities may adversely affect youth and promote risky behaviour. Carnagey et al., in a review of literature revealed that playing violent video games not only increased aggression but also desensitized youth to violence in real life and decreased pro-social behaviour (72). Contrary to the majority of literature concerning the 19  advantages of engaging in sports, Leichliter et al., studied 51483 college students and found that the athletes had more alcohol consumption per week, got more involved in binge drinking and substance abuse than those who were not involved in sports. This effect was stronger for males than for females. Leichliter et al., found that athletes were more likely to experience adverse consequences of drinking and substance abuse (such as driving under the influence), being injured, being involved in an argument or fight, or missing a class (73). Similarly, Lorente et al., studied 815 high school students and reported that sport involvement, in particular involvement in group sports, is related to increased alcohol consumption (74). Brenner et al., also found similar findings for team sports, when they studied 720 college athletes (75). Peretti et al., surveyed 10807 adolescents (14-19 years old) in France and found that, after controlling for confounding factors, repeated alcohol use (defined as to have drunk alcohol 10 times or more during the past 30 days) was associated with regular involvement in individual sports other than team sports, and was more frequent among intensive sportswomen (defined as 8 or more hours per week). Regular involvement in individual sport was also linked to cannabis use in boys (76). Wichstrom et al., conducted a 13-year follow up study on 3251 Norwegian students aged 13-19 years and found that involvement in organized sports not only predicted alcohol intoxication, but also was associated with tobacco and cannabis use (77). Moreover, the results of a study by Marten et al. showed an association between the sport discipline and drinking alcohol. Those involved in swimming and diving reported the highest rate of alcohol consumption compared to other sports (78).  1.3.3 Leisure activities and risky driving behaviour  Like other behaviours, driving behaviour can also be positively or negatively influenced by leisure activities (See theoretical framework - Figure 1.5). Limited leisure activities have been examined with respect to their association with driving behaviour.  1.3.3.1 Spending time with parents (Parental influence on driving behaviour)  There is clear evidence that the driving behaviour of parents and other family members has a strong influence on children’s subsequent driving behaviour (79-81). Parents act as role models for their children, therefore, as they go about their daily life, they establish norms for their children about what is considered acceptable driving behaviour and what is considered to be risky driving behaviour. (82,83,84,85). Moreover, spending time with parents can influence driving behaviour through parent-20  child socialization, both directly (e.g., verbal persuasion) and indirectly (e.g., vicarious experience) (84,86). Youth’s driving behaviour is positively impacted by parental monitoring (87) Loubean et al., showed that parental influence can even be used as a teaching tool for improving driving behaviours in young individuals (86). Importantly, driving violations and driving styles of young adults were found to be associated with those of their parents (87,88). Ferguson et al., in a study in the USA reported that driving records and crash rates of young drivers were related to those of their parents (47), and Miller et al., in a study in Israel showed that, in addition to crash rate and driving record, young individuals' driving style was also similar to that of their parents. (89). Simon-Morton et al., in the USA revealed that the graduated driving licence system (GDL) allows new drivers to gradually acquire driving skills and experience over time) (90). In another study in the USA, Simon-Morton et al., found that risky driving behaviour in young drivers decreased when parents were present but increased when peers were present (19).  1.3.3.2 Peer Influence on Driving Behaviour Individuals who perceive that their peers engage in problem behaviour(s) are also more likely to engage in that behaviour themselves (36-40,91). For example, youth who observe and perceive their friends having unsafe sex (92,93) or using alcohol and drugs (94) are or will be more likely to engage in these behaviours themselves. Peers exert direct and/or indirect pressure on young drivers’ behaviour while driving and spending time with them. Young drivers may change their speed or driving behaviour to fulfill their peers’ expectations, especially when peers are present as passengers (95). Furthermore, having multiple teenage passengers, a common occurrence in youth leisure time, is a risk factor for property damage, injury, and fatal crashes in teenage drivers (96). 1.3.3.3 Playing video games  In terms of the type and intensity level of a video game, two studies found that young people who played a racing video game tended to take more risks in critical driving situations in real life than those who played a neutral game (characterized as having no competitive component). In addition, the higher the intensity level of the racing game, the higher the risk-taking tendency while driving (97,98). However, there is mixed evidence when it comes to the effect of video games including car racing on real world driving behaviour. While some literature has found that playing video games has positive impact on driving behaviour through improving visuomotor control (99), more evidence shows that playing racing video games has a negative influence on actual driving behaviour, such as 21  several lane changes or speeding (100,101). Beullens et al., surveyed 589 drivers aged 18-24 years and found that those who played video games containing reckless driving were more likely (2-4 times) to drive without a license. Moreover, playing such video games seemed to be a better predictor of unlicensed driving than age, risk perception, and sensation seeking (102). Greitemeyer et al., in a study on 103 students at Austrian universities evaluated the influence of media content on driving behaviour. The results showed that youth who played a pro-social video games drove less recklessly compared to those who had played a neutral one (103). Moreover, a multi-wave longitudinal study conducted via a phone survey reported that playing mature-rated video games was related to increased risky driving behaviours and sensation seeking (101). Fischer et al. found that “risk-glorifying” media (e.g., reckless driving in video games, and drinking and smoking in movies) can change people’s self-concept and increase their risk-related cognition which can lead to increased risky behaviours such as risky driving (104). 1.3.3.4 Watching Movies and Television (TV)  Media (e.g., TV and movies) may affect the perception of reality, risk and norms in young people (103-105). A study that surveyed 1178 young drivers revealed that young adults who watched movies containing reckless driving scenes were more likely to demonstrate risky driving behaviour in real life (106). Beullens et al., in another survey-based study on 462 young drivers found that watching news was a negative predictor while watching action programs on TV was positively associated with risky driving in youth (107).  1.3.3.5 Alcohol- or drug- involved activities  The role of activities involving alcohol and drugs on driving behaviour has been extensively researched. In addition to the impairment influence on driving skills, balance, and coordination; alcohol and drugs also alter risk perception resulting in changes to driving behaviour and increase risky driving behaviours (108). Notably, driving after alcohol use is highly associated with binge drinking (109). Stevely et al., in a systematic review on the association of the drinking context with acute alcohol-related harm found the drink driving was associated with the following contextual characteristics: place (i.e. licensed premises), timing (i.e. Fridays, weekends, holidays, and evening while mid-week for students), psychological status (i.e. in students, higher objective and lower subjective intoxication), and drink type (i.e. beer which is commonly consumed by binge drinkers and young people) (110).  22  According to data from the 2012 Canadian Community Health Survey, compared to other age groups, cannabis use is more prevalent among 15-24-year-olds (111). Capler et al., in a scoping review on cannabis use in the driving context indicated that 4-6% of drivers drove within two hours after cannabis use according to the roadside testing results from 2016. They highlighted young male high school students who drove within one hour after cannabis use, which is associated with reckless driving. Some psychosocial factors were shown to be in association with driving after cannabis use, including demographic characteristics, poly drug use and dependency, driving styles and attitudes towards risk. In the driving settings, cannabis use has been found to negatively influence reaction time, motor coordination, and short-term memory, divided attention tasks, making decisions under pressure and complex situations (112). In a recent unpublished meta-analysis including 81 experimental driving studies, Simmons et al., demonstrated that cannabis (as the most popular drug in the world), like alcohol, leads to a negative impact on driving performance, e.g., lateral control impairment. The combination of both drugs can aggravate the impairment in driving performance resulting from either drug in isolation (113). Moreover, drug usage is associated with increased injury risk in driving such as not fastening seat belts (114). 1.3.3.6 Listening to music  Listening to music while driving is a very common occurrence. Based on a study of 1780 British participants, driving performance was influenced when not only listening to music but also by the type of music played during the driving session. This influence was confirmed when it was found that drivers with four or more years of accident-free record were those who preferred to not listen to music while driving. Youth, aged 18-29, with less than four years’ accident-free record tended to listen to dance and house music with a fast tempo and at high volumes. Characteristics of music such as tempo and volume are worth considering when it comes to the influence of music on driving behaviour (115). Based on another survey of 2000 public people in 2011 in the UK, Williamson et al. found that those who listen to drum and bass and heavy metal music (who are mostly young drivers), self-reported more aggressive driving behaviours when compared with classical music listeners. Jazz listeners tended to report receiving more speeding fines than other musical genre listeners. Reggae music listeners were more likely to experience near misses when driving. The survey results also showed music containing lyrics, louder, more complex music, and unfamiliar music were all more likely to be distracting. Some findings, however, might be confounded with age (116). Based on an experiment 23  with 50 drivers using driving simulators, Brodsky et al., found that violent-content and energetic types of music prompted the drivers to perform some risky driving behaviours, such as exceeding the speed limit resulting in a loss of lateral control (117). 1.3.3.7 Social media  Social media can influence behavioural norms and risk taking behaviour in youth (118,119). Abi-Jaoude et al., stated, in a published podcast at Canadian Medical Association (CMAJ), that social media can negatively affect youth mental health (i.e., self-view, interpersonal relationship, self-harms, and suicidality), despite the fact that digital interactions are the norm and social media use can also be beneficial (e.g. creativity, self-expression, on-line access to many resources, etc.) (119). Driving behaviour is one of the least examined risky behaviour in term of its relationship with social media (118).  Fast-evolving technology (e.g., the smart phones and increasing internet accessibility) has influenced people, especially youth, to increasingly conduct their social interactions online. Checking phones, texting, and calling while driving are areas of concern in the actions of young drivers which indirectly lead to more risky driving behaviours, such as distracted driving (120). A subgroup analysis of a population-based survey of 1133 students in Ontario demonstrated that 36% of student drivers 16 years and older texted while driving (121). Other social media, such as YouTube, can influence driving behaviour in youth through the streaming of top-viewed and top-rated videos that include high speed and risky driving behaviours (122,123), for example ’sidewalk skiing’ and ‘ghost riding the whip’ (59). As noted, it seems that the influence of social media on driving behaviour can be explained by the theories (section 1.2.4) that attitude and perception of risky driving behaviour viewed on social media tend to be normalized, learned, and adopted by the younger population. 1.3.3.8 General lifestyle   Gregersen et al., evaluated the association between lifestyle and car crashes. They found that the high-risk groups in MVCs were generally characterized by an average or below average engagement in sports, drank more often, had a little more “out and about” lifestyle (defined as having an interest in driving with friends, rock concerts, records, parties, disco, and movies), had a higher than average interest in cars, and had more ulterior motives in their driving (such as showing off, pleasure, sensation seeking). Intriguingly, the high-risk groups were also engaged more in cultural activities (classical music, theatre, and playing instruments) (124). Bina et al., reported that risky 24  driving is positively associated with non-organized activities with friends (such as meeting friends, driving around, going to a club, staying in a pub, going to friends’ homes), and non-organized activities at home (such as talking on the telephone, listening to music alone, just sitting doing nothing), especially for boys. Bina et al. also found that risky driving was negatively correlated with leisure activities such as reading for fun, physical exercise (e.g., running or biking), artistic activities, spending time in a library for boys only, and time spent with family. They concluded that antisocial and non-organized leisure activities with friends and at home such as drinking can be a predictor of risky driving behaviour in male youth. The study suggested that the level of engagement in risky driving is associated with the time spent outside the home with friends to a greater extent than with the type of activity (organized or non-organized). Bina et al. highlighted the important role of peers for behaviour normalization and for forming the perception of norms and risks in young drivers (125).  Overall, the association of different leisure activities and risky driving behaviour has not been extensively researched as indicated above by limited studies available.  1.4 Personality trait as a major confounder affecting both driving behaviour and leisure activities According to problem behaviour theory and Shope’s conceptual framework (23), personality traits can also play a role in the risky driving behaviour in youth. There is a large body of literature describing the association of different personality traits with risky driving behaviour and MVCs. In addition, the selection of leisure activities and the level of engagement in them can be influenced by personality traits (126,127). Previous studies investigated the association between leisure activities and driving behavior or the association between personality traits and driving behaviour, but no prior study has examined personality as a confounder in the association between leisure activities and driving behaviour.   1.4.1 Role of personality traits in risky driving behaviour Despite a few studies that found no significant differences in driving behaviour among people with different personality traits (128,129), other research concluded that personality traits can influence the drivers’ approach to keep safe on roads (130,131). The following sections describe how personality traits influence risky driving behaviour.   25  1.4.1.1 Personality trait models (other than Big Five Model) and risky driving behaviour Risky driving behaviour in adolescence has been associated with impulsivity and aggression (132), social deviance (defined as the recognized violation of cultural norms) (133), developmental needs for sensation seeking (134), feelings of invulnerability (135), and underestimation of risk (136). The following traits are shown to be closely associated with MVC related risky driving behaviours such as aggressiveness, sensation or thrill seeking, lower inhibitory control, lower attitudinal intolerance of deviance (i.e., cannot tolerate an action or behaviour that violates social norms), impulsiveness, aggression, being prone to boredom, impatience, and inattentiveness. (132-136).  1.4.1.2 Big Five Model of Personality traits and risky driving behaviour In this thesis project, the Big Five model (or Five Factor Model) was chosen to study risky driving behaviour as it is one of the most common models found in literature. There are five personality traits in the Big Five model: extroversion, agreeableness, openness, conscientiousness, and neuroticism (137-139).  The literature shows that extroversion, defined as the pleasure of interacting with others, and the tendency to be self-assured, outgoing, sociable, energetic (138,140), is associated with crashes, traffic fatalities, traffic violations, impaired driving and physical aggression (141-144).   Neuroticism, characterized by a tendency to be tense, anxious, intolerant to stress and challenges (140), is also positively associated with reckless driving, aggressive driving, verbal and physical aggression in driving, traffic violation, and can posse a higher risk for MVCs and traffic fatalities (145,146).  Conscientiousness, characterized by order, organization, self-discipline, pre-planning, and problem solving (140), is inversely associated with verbal and physical aggression while driving, at-fault crashes, total crashes, and driving violation tickets (147,149).  Individuals with agreeableness are characterized as trusting others, more forgiving, exhibiting altruistic behaviour, empathic, tolerant, and being generous and gentle (140). They also exhibit a careful and less aggressive driving behaviour and drive in a less hostile, reckless, anxious, and furious manner (149,150).  The relationship between risky driving behaviour and openness, characterized by aesthetic appreciation, values, idea acceptance, self-actualization, personal growth, and development (140) is inconclusive (120). A few studies found a positive association between openness and aggressive 26  driving behaviour (149) and at-fault crashes (148), while other studies show a positive relationship between openness and careful driving style and less reckless driving (147,151).  1.4.2 Personality trait and leisure activity  Participating in organized leisure activities is strongly associated with psychological well-being.   According to social control theory, problematic behaviour tends to occur more often when bonds among young people and conventional societal institutions (e.g., family, school, church) are severed. Participating in unstructured activities can increase problematic behaviours by decreasing young people’s bonds with conventional societal institutions (151). McGuiggan et al., found that some components of leisure activities are better explained by personality attributes (planning, follow through, seeking variety, people, and pace of activities), or by demographics (participation in household tasks, team sports and modernity) or by both (risk characteristics of activities). Neither personality nor demographics determined the level of engagement in leisure activities (127). 1.4.2.1 Personality trait models (other than Big Five Model) and leisure activities According to the three-dimension (i.e., Psychoticism, Extroversion, and Neuroticism), Eysenck personality model combative leisure activities (such as martial arts and boxing) are associated with psychoticism and competitive leisure activities (such as sport clubs and team sports) are associated with extroversion personality traits (152,153). Professional members of musical bands have been ranked higher on neuroticism, extroversion, and psychoticism scores (154,155), although Hills et al., found no correlation between personality types and participation in amateur musical bands (153). Interestingly, even TV program watchers vary in personality traits. Extroverted people and women were regular watchers of TV soap operas (156) while male undergraduate students with high neuroticism scores typically avoided comedy and adventure programs (157). 1.4.2.2 Big Five Model of Personality traits and leisure activities Researchers using the Big Five model found that extroversion and openness were associated with more physical activity and less inactivity (e.g., TV watching) (158). In addition, engaging in social media is more common in extroverted and open individuals (159-161). Lower neuroticism and higher conscientiousness were associated with more physical activity, less inactivity, and less sedentary behaviour (158). Literature has shown that people with an openness personality tend to participate in developmental activities and embrace technology use. But they disliked religious activities (162) and 27  less challenging activities such as watching soap operas, or reading romantic fiction, while agreeable, conscientious, and neurotic people tended to participate in these activities (163). People with the agreeableness personality trait disliked crafts, physical activities, and developmental activities (i.e. reading books as part of a job, attending public lectures, taking courses at a university, creative writing, going to the library, studying a foreign language, on-the-job-training, and attending movies), yet liked watching TV, religious activities, experiential activities, and social-public activities (162). 1.5 Summary and Knowledge gap MVCs are one of the leading causes of hospitalization and premature death in young Canadians (3). Despite the recent improvements in vehicles roadway design, and road safety strategies (22,164), MVCs remain a major public health problem, especially in youth. A variety of frameworks and approaches have been applied to identify the risk factors for MVCs (9-14). The contributing factors for MVCs broadly fall into three categories: environmental, vehicular, and human factors (15,16). Among these factors, the human factor is a major contributor to MVCs. In youth, risky driving behaviour, such as speeding, distracted driving, impaired driving, and driving with fatigue, all fall under the human factor category (20-24).  By definition, risky driving behaviours are those actions undertaken by drivers which expose drivers, their passenger(s), and other road users (such as pedestrians, cyclists, and drivers and passengers in other vehicles) to hazardous driving situations (33). Shope et al., proposed the following emerging constructs that contribute to risky driving behaviour: developmental factors, personality characteristics, driving ability, demographic factors, driving environment, and perceived environment. Shope et al., stipulated that the “perceived environment”, defined as the driver’s perception of norms, risk, and attitude, in a driving context, is perhaps the most complex contributing factor for risky driving behaviour in youth (23) There has been growing interest in the lifestyles of people who tend to be involved in risky driving behaviours (23,37-42).  Leisure activities are a major component of daily activity in youth and can expose young individuals to various elements of a perceived environment (165). Leisure activities is one of the main norms and a behavioural exchange platforms where youth can observe learn (166,167). Youth can be immensely affected by the norms and behaviours through interactions with parents, peers, culture, community and media, any of which may immensely affect their driving behaviours (166,168). The majority of literature has investigated the effect of one or a few leisure activities on a single or a 28  particular risky driving behaviour (e.g., aggressive driving). However, the relationship between leisure activities and driving behaviour has seldom been studied, particularly involving Canadian youth. In addition, evidence was mixed with respect to the effects of some leisure activities on driving behaviours and cannot be used for developing targeted safety and preventive strategies.  Moreover, the problem behaviour theories (38-40) have utilized an overarching approach to problem behaviour analysis focused on contributing factors (e.g., personality trait and sociodemographic factors) as being causal for problem behaviours such as risky driving behaviours (168-170).  Given the lack of evidence around the role of leisure activities in risky driving behaviours and the significance of a holistic approach, the current study was designed to primarily investigate the potential influence of leisure activities on driving behaviour while also taking the confounders of this association (personality trait and sociodemographic entities) into consideration (Fig. 1.6). 1.6 Study Aim, Objectives, and Hypothesis 1.6.1 Aim  Investigate the association between different types of leisure activities and risky driving behaviour among youth living in Canada. 1.6.2 Objectives The objectives of the study were to: 1. Investigate the association between leisure activities and risky driving behaviour adjusted for personality and sociodemographic factors 2. Describe the characteristics of driver risk-groups in young drivers in the context of leisure activities, personality, and sociodemographic factors 1.6.3 Hypotheses  1. Leisure activities are associated with the risky driving behaviour of youth living in Canada. 2. Personality traits are associated with risky driving behaviour of youth living in Canada. 3. Sociodemographic factors are associated with risky driving behaviour of youth living in Canada. 1.6.4 Study questions Are leisure activities associated with risky driving behaviour in Canadian youth?  29  What types of leisure activities are associated with risky driving behaviour among youth living in Canada?   Figure 1.6 The association of leisure activities and risky driving behaviour The association of leisure activities and risky driving behaviour is the main objective of this study. Sociodemographic factors and personality trait are confounding factors affecting both leisure activities and driving behaviour that are also incorporated in the study design. This framework depicts the role of leisure activities on risky driving behaviours and how personality traits and demographic factors may influence this association.     Demographic variables  Risky Driving Behaviour e.g. Speeding, Distracted driving, Drink driving, Drug driving, Aggressive driving   Personality Traits Leisure activities (e.g., Sports, Movies, Video games, Family activities)  30  2. Methodology 2.1 Study design and target population This is a cross-sectional study with data collected through an online survey of young drivers in Canada conducted between March and May 2018. Participants who were 16 -24 years old living in Canada with a valid driving license and had been driving at least 2 hours per week in previous 3 months at the time of survey were invited to participate. 2.2 Engaging a youth advisory committee (YAC) Following the patient-oriented research strategy of Canadian Institutes of Health Research to transform the role of participants from a passive role to a proactive one, 24 youth (12 males, 12 females), aged 16-24, were engaged as a Youth Advisory Committee (YAC). These youth were from diverse geographic and demographic backgrounds: 20 from Metro Vancouver, three from Toronto, and one from Montreal.  The first few young people were recruited by word of mouth (asking colleagues and other research teams). The remaining members were recruited by snowball sampling, with initial members assigned to approach young individuals from diverse backgrounds, i.e. different cities in Canada, cultures, ages (ranged 16-24 years) and genders, as well as different fields of interest in their social media, communities, clubs, and schools. YAC actively collaborated with the project from an early stage and contributed to the development of the study design, survey, data collection strategy and knowledge translation activities. Following YAC member preference, a Facebook group was created to allow interactive communication. The YAC was also consulted through interactive in-person meetings with refreshments held approximately twice per month for 15 months in the Research Pavilion at Vancouver General Hospital. Three members from Toronto joined the meetings via free online communication tools such as Skype or Face-Time services. Attendance of more than half of the members (twelve or above) was a requirement to convene the sessions. Those who were not able to attend the sessions were updated by posting meeting notes on the Facebook group. YAC member opinions were sought on road safety and its priority, developing the Leisure Activity and Sociodemographic Questionnaires, and reviewing the validated questionnaires for personality traits and driving behaviour to ensure they were understandable when presented to youth. 31  2.3 Survey instruments The online survey consisted of a cover letter (explaining the study and obtaining consent) (Appendix D), a sociodemographic questionnaire (SDQ), a leisure activity questionnaire (LAQ), and a driving behaviour questionnaire (using the modified version of validated Behaviour of Young Novice Drivers Scale (BYNDS), and a personality trait questionnaire (through a valid Mini International Personality Item Pool, i.e. Mini-IPIP) The order of questions in the survey followed the “warm up - cool down” approach  in order to achieve a more completely answered survey. 2.3.1 Basic sociodemographic questionnaire  The basic sociodemographic inventory consisted of twelve items which were developed with feedback from the YAC. This inventory (Appendix D) included the participant’s location by province, age, gender, drivers’ license status, amount of driving hours per week, number of cars in the family, personal car ownership status, education, employment status, living status (with whom), duration of driving independently (years), and family home ownership (Table 2.1). It was decided not to enquire about elements (such as family income) that warrant parental consent. Hence, socioeconomic status questions included questions on “home ownership” and “number of cars each family owns” which were used as proxies for socioeconomic status. Literature shows family structure affects risky behaviours due to the fact that single parent and both parents living situations can be different. According to Health Behavior in School-aged Children (HBSC), living arrangements of Canadian students: more than two thirds of Canadian young people (71%) live in home with both parents. The remaining youth live in single-family homes (with mother 14%, with father 3%), with a parent and stepparent (8%), or in another kind of arrangement including with grandparents, extended family, friends, or in a foster care situation (4%) (171).   32   Table 2.1 Sociodemographic variables, options in the survey, and final categories *Final categories are the ones used for analysis Variables Survey inquiry Answer Options in Survey Final Categories Gender I identify as Male/Female/Other  Male/Not Male Age What is your age in years? Age in years from 14 to 25  16-19: Beginner Driver  20-24: Young adult driver 1 Employment status What is your employment status?  Employed Student Both (Employed and student) Other Employed  Student  Both (Employed and student)  Unemployed Living status Who do you live with? Family (both parents) Family (single parent) With friend(s) Partner Alone Other  Living with both parents  Living with a single parent  Living with a spouse or partner or child  Living alone or with friends, roommates, or with other relatives 2 Socioeconomic status How many vehicles does your family own?  None ≤ 2 >2 Low = No home, No Car,  Moderate = ≤ 2 cars and No home, High= Home AND/OR car>2 Does your family own their home? Yes No Education status Please indicate the highest level of your education.  Less than high school diploma Graduated from high school Some postsecondary Postsecondary diploma or certificate University degree 3 Other  Less than high school diploma  High (Secondary) school diploma or equivalent Some post-secondary education 2  Province  Which province/territory do you reside in? AB, BC, MB, NB, NL, NT, NS, NU, ON, PE, QC, SK, YT 4 AB, BC, MB, NB, NL, NT, NS,  NU, ON, PE, QC, SK, YT 4 Driving profile:     Driver’s license status What is your driving license status? 3 refer to Table A.1  With restriction  Without restriction 3 Drive independently duration How long have you been able to drive independently? <1 year 1-3 years 3-5 years >5 years < 1 year  1-3 years >3 years  Driving exposure In the last three months, how many hours a week, on average, did you drive? (Text required) Time in hours from 0 to 168  Low: <11 hours/week Moderate: 11-20 hours/week High:  >20 hours/week 4 Driver’s car Do you own your own car (e.g. truck, motorcycle, van, SUV, etc.)? Yes No  Yes No 1 Categories based on the data available from:  https://www.madd.ca/english/research/youth_and_impaired_driving_2006.pdf https://www.tc.gc.ca/eng/motorvehiclesafety/tp-tp15145-1201.htm http://tirf.ca/wpcontent/uploads/2017/02/RSM_YD_Backgrounder.pdf 2 The options in the survey were based on Statistics Canada (174).  For the analysis, the last two options merged to reduce the number of categories. Of note, the research team (supervisory committee and I) agreed that merging would not significantly affect the results.  3 Due to having ten separate provinces and three territories, breaking the driving license to these two categories was the only way to make them united 4 These categories were defined based on the responses from participants and the transport to and from activity from General Social Survey (GSS) 2015 conducted every 5 years that Average driving for age 15-24= 1.5 hr/d (172).33  2.3.2 Behaviour in Young Novice Driver Scale (BYNDS)  In this study, drivers’ profile was measured using the Behaviour of Young Novice Drivers Scale (BYNDS) which is a self-report measurement tool for risky driving behaviours. BYNDS was developed by Scott-Parker et al. and validated with drivers aged 17-25 years in Australia and 16-24 years in New Zealand (173,174).  The BYNDS consists of 44 items comprising five subscales:   1. Transient rule violations (13 items, a=0.89) defined as “risky behaviours that can be performed multiple times throughout the journey, such as speeding” - Questions 1 to 13 2. Fixed rule violations (10 items, a=0.73) defined as “behaviours that are stable throughout the journey, such as driving after using illicit drugs” - Questions 14 to 23 3. Misjudgement (9 items, a=0.76) which reflects “driving errors, such as misjudging the required stopping distance” - Questions 24 to 32 4. Risky driving exposure (9 items, a=0.81) which reflects “conditions which place the novice at greater risk of crash, such as driving at night with friends as passengers” - Questions 33 to 41 5. Driver mood (3 items, a=0.87) defined as “the driver’s emotional driving, such as driving faster if they were in a bad mood” - Questions 42 to 44 The 44 questions are scored on a 5-item Likert scale, from 1 (never) to 5 (nearly all the time). The total score is calculated by adding up the score for each question. The composite score has a very high internal consistency measure (Cronbach’s alpha= 0.97). The composite score was weakly correlated with self-reported crashes (rs=0.17) and moderately correlated with future offences (rs=0.21), while more strongly correlated with the participants intentions to break traffic rules (rs=0.44) (175). The composite score of BYNDS was also used to categorize the drivers into three main risky driving behaviour groups: high, moderate and low risk (176,177). (Table 2.2)  Young drivers who score high have more tendency to engage in risky driving when there is a chance of not being observed by police. It is noteworthy that BYNDS covers all risky driving behaviours in young drivers such as speeding, distracted driving, etc. As the BYNDS was developed in Australia, the wording was modified with consultation from the YAC to reflect the Canadian driving context (e.g. driving on right side of the road versus on left side), and to replace Australian vernacular with Canadian terminology (e.g. “peak time” was changed to “rush hour”, “mates” was changed to “friends”) (Table 2.3). It should be noted that the BYNDS was used in a study in North 34  America which aimed to assess the association of risky driving behaviour and mental health symptoms in novice adolescent drivers (aged 16-17 years) in Pennsylvania (178).The psychometric properties of the BYNDS have not been validated in Canadian youth.    Table 2.2 Driving behaviour categories based on the composite score of BYNDS  Variable Categories (based on the total score) Composite score range (44-220) Risky Driving Behaviour  Low risk= Composite score <71 *  Moderate risk = Composite score= 71-91  High risk= Composite score >=92   *The cut-off values were adopted from the literature (179)   Table 2.3 Questionnaire Changes on Young Novice Drivers Scale (BYNDS) instrument Canadian BYNDS Australian BYNDS You drove over the speed limit in areas where you were unlikely to get caught  You drove over the speed limit in areas where it was unlikely there was a radar or speed camera You deliberately sped when passing another vehicle You deliberately sped when overtaking You travelled in the left lane on multilane highways You travelled in the right lane on multilane highways You passed a car on the right You overtook a car on the left You used spoke/texted on a handheld cell phone You spoke on a mobile that you held in your hands You misjudged the gap in traffic when you were turning left You misjudged the gap when you were turning right You turned left into the path of another vehicle You turned right into the path of another vehicle  You misjudged the gap in traffic when you were passing another vehicle  You misjudged the gap when you were overtaking another vehicle You cut off another vehicle when entering the road You entered the road in front of another vehicle You didn't always signal when you were changing lanes You didn’t always indicate when you were changing lanes You drove during morning and afternoon rush hour You drove at peak times in the morning and afternoon You went for a drive with your friends giving directions to where they wanted to go You went for a drive with your mates giving directions to where they wanted to go    For the entire questionnaire, refer to Appendix C 35  2.3.3 Leisure time activity questionnaire  The leisure time activity questionnaire was designed and developed after an extensive review of the literature (180-184) and revised based on feedback from the YAC. There is no instrument compatible with the young (aged 16-24 years) in the Canadian context that would inclusively and reasonably measure leisure activity. Therefore, a comprehensive list of leisure activities was abstracted from literature (185). For example at the time of developing the survey (2018), according to the most up-to-date statistics provided on the website of “BooknetCanada.ca” (a non-profit organization that develops technology, standards, and education to serve the Canadian book industry) , Canadians spent their leisure time on browsing the internet (33%), spending time with family (32%), watching TV (31%), watching a movie (23%), and reading (22%). The distributions of these leisure activities were fairly similar to those found in the previous three years (187). Of note, the website updated the statistics in April 2019 which lists the following as popular daily leisure activities among Canadians, browsing social media and/or the web daily (60%), watching videos/TV/movies (56%), listening to music (51%), cooking (47%), and spending time with family (40%) (186). This update did not influence the survey.  Each major category of leisure activity was divided into subtypes using classifications from the literature (when available) and input from the YAC. For example, in 2001 Rentfrow categorized musical styles as 1. Classical, 2. Pop, alternative, rock, heavy metal, 3. Jazz, blues, 4. Traditional, folk, country, ethnic, 5. Rap, electronic, and 6. Other (181). From the perspectives of YAC members and two musician friends who I reached out to, these categories, however, did not seem reasonable due to the fact that very different genres fell under the same categories. Therefore, for this study music genre categories were developed in consultation with the musicians and then obtained acceptance from YAC. For the movie, book, and magazine types of entertainment, the categories were adapted from two studies which evaluated the association of these activities’ preferences with demographic factors and personality traits (181,185). An appropriate classification of educational categories was adopted from the Council of Minister of Education of Canada and Canadian degree qualification framework and degree categories (182). The sub-types of volunteer activities were drawn from the Statistics Canada questionnaire (187). For the final questionnaire, we included examples of activities from each sub-type (Appendix A).  36  With respect to the level of engagement in each leisure activity, the existing evidence was limited or non-existent for most activities and was not applicable to young Canadians. Therefore, the insight of the YAC members was sought with regard to how to enquire regarding the general types and subtypes of leisure activities, and the level of engagement in each leisure activity. Accordingly, level of engagement was categorized into 5 groups with varied hours/times per week depending on the leisure activity types (Table 2.4).  Pilot testing the Leisure Activity Questionnaire Each YAC member sought input from 3-5 young individuals with a different age, culture, and field of interest (different individuals than those who had previously been involved helping with the survey development). They found volunteer youth from their schools, social media, and friends contact lists. The total sample of 75 young individuals from Toronto, Vancouver, and Montreal were asked to read and answer the survey, and to share their insights on the following fields: types of leisure activity (i.e. whether there were any missing leisure activities), subtypes (whether participants could find their choice without confusion, whether there were any missing leisure activity sub-types), the cut off values on the five-item scale for level of engagement (whether they think that the cut off values of 5-item scale made sense to them), wording, and understandability. The majority of participants (73 youth) approved all aforementioned fields without any comment. There were a couple of comments from two of participants regarding the cut off points for the leisure activities and subtypes which were not approved in the YAC meeting. Therefore, the face validity of the questionnaire was verified. With respect to the content validity, we chose not to calculate internal consistency or reliability for the Leisure Activity Questions due to their factual nature. The leisure activity questionnaire is a categorization of possible leisure activities. If a youth spends most of their time doing one type of activity, then they likely do not spend as much time on the others. When Cronbach’s Alpha is used to calculate internal consistency, it is meant to demonstrate consistency between different questions on a survey that measure the same construct. This is not applicable to the leisure activity questionnaire because each question was loaded on one construct, i.e., 13 questions measured 13 distinct leisure activities. Moreover, each question which represented one construct influenced one manifest variable. Therefore, even Explanatory Factor Analysis was not a reasonable option.  For leisure activity correlations, there can sometimes be issues with multicollinearity if two variables are too similar. However, there was not a problem with that in this case. When the 37  correlations were examined between high/low engagement in the 13 activities, the highest correlation was rho=0.45 (for movies and music) which was not high enough to cause multicollinearity issues. As noted above the leisure activity questionnaire consisted of thirteen types of activities which were evaluated in the current study based on the level of engagement and the sub-type of each leisure activity. For the analysis, the level of engagement collected based on the five-item scale was categorized into two groups: high or low engagement (Table 2.4). In addition, participants were asked for their top three choices from the sub-types of leisure activity within each category (if they had indicated that they partook in that specific leisure activity) (Appendix A, Table 2). Table 2.4 Leisure activities and level of engagement. The level of engagement was collected on a 5-item Likert scale (below-Primary category). Then, for the analysis, the level of engagement was stratified as Low or High engagement (below-Final categories) Variable Survey inquiry Primary categories (in the survey)  Final categories (For analysis)  Video game engagement In the last 3 months, how many hours a week did you play video games? 1. None 2. ≤ 4 hours per week 3. 4 - 9 hours per week 4. 9 - 14 hours per week 5. 14+ hours per week 1, 2, or 3: Low engagement  4 or 5: High engagement  ( appl ied to all rows) Sport engagement In the last 3 months, how many hours a week did you play sports? 1. None  2.  ≤ 1 hour per week  3. 1 - 3 hours per week  4. 3 - 7 hours per week  5. 7+ hours per week Social media engagement In the last 3 months, how many minutes a day did you spend on social media? 1. None 2.  ≤ 30 minutes per day 3. 30 - 60 minutes per day 4. 60 - 120 minutes per day 5. 120+ minutes per day Movie engagement In the last 3 months, how many hours a week did you spend watching movies? 1. None 2.  ≤ 3 hours per week 3. 3 - 6 hours per week 4. 6 - 10 hours per week 5. 10+ hours per week Family engagement In the last 3 months, how many times a week have you participated in an activity with your family? (N.B. It does not include the time you spent together at home doing your own tasks) 1. Never 2. Once a week 3. 2-3 times per week 4. 4-5 times per week 5. 5+ times per week Friend engagement In the last 3 months, how many times a week have you participated 1. Never 2. Once a week 38  in an activity with your friends? (N.B. It does not include the school time activities) 3. 2-3 times per week 4. 4-5 times per week 5. 5+ times per week Alcohol engagement In the last 3 months, how many times a week did you drink alcohol?   1. Never 2. Once a week 3. 2-3 times per week 4. 4-5 times per week 5. 5+ times per week Drug engagement In the last three months, how many times a week have you used recreational drugs?   1. Never 2. Once a week 3. 2-3 times per week 4. 4-5 times per week 5. 5+ times per week Art engagement In the last three months, how many hours a week did you engage in art activities? 1. Never 2. ≤ 1 hours per week 3. 1 - 3 hours per week 4. 3 - 7 hours per week 5. 7+ hours per week 1, 2, 3: Low engagement 4, 5: High engagement  ( a pplied to all rows)  Reading/Writing engagement In the last three months, how many hours a week did you spend on reading/writing something? (N.B. It does include reading/writing for a leisure activity and not for school or work) 1. Never 2. ≤ 1 hours per week 3. 1 - 3 hours per week 4. 3 - 7 hours per week 5. 7+ hours per week Volunteering engagement In the last three months, how many hours a week did you act as a volunteer?   1. Never 2. ≤ 2 hours per week 3. 2 - 4 hours per week 4. 4 - 6 hours per week 5. 6+ hours per week Music engagement In the last three months, how many hours a week did you listen to music?  1. Never 2. ≤ 15 hours per week 3. 15 - 30 hours per week 4. 30 - 45 hours per week 5. 45+ hours per week TV engagement In the last 3 months, how many hours a week did you spend watching TV? 1. None 2. ≤ 3 hours per week 3. 3 - 6 hours per week 4. 6 - 10 hours per week 5. 10+ hours per week  2.3.4 Mini-IPIP Questionnaire, measuring personality traits  Mini-International Personality Item Pool (Mini-IPIP) (Goldberg, 1999), is a validated questionnaire which measures the five personality traits based on the Big Five model, i.e. extroversion, agreeableness, conscientiousness, neuroticism, and imagination (openness) (189) (Appendix D). Donnellan et al., in a study on North American college and university students provide 39  evidence that Mini-IPIP is an effective, validated, and reliable questionnaire when researchers are seeking a fast method of assessment for Big Five Model. It has consistent and acceptable internal consistencies across five studies (a at or well above 0.60) (190).  The Mini-IPIP questionnaire consists of 20 questions designed as “+” keyed and “-” keyed questions on 5-item Likert scale to assess the five personality traits. For “+” keyed items (i.e. questions 1, 2, 3, 4, 5, 11, 12, 13, 14), the answer "Very Inaccurate", "Moderately Inaccurate", "Neither Inaccurate nor Accurate", "Moderately Accurate", and "Very Accurate" are assigned values of 1, 2, 3, 4, and 5, respectively. For “-” keyed items (i.e. questions 6, 7, 8, 9, 10, 15, 16, 17, 18, 19, 20), the scoring is reversed, i.e. the answer "Very Inaccurate" is scored a value of 5, "Moderately Inaccurate" a value of 4, "Neither Inaccurate nor Accurate" a value of 3, "Moderately Accurate" a value of 2, and "Very Accurate" a value of 1. A total scale score is calculated by just summing all the values for that scale (191). This instrument has been used to study personality traits of on 15,471 adults (mean age=29.1, SD=1.75) in North Carolina, USA. Of note, they were also interviewed while they were in grade 7-12 (ages 11-19) (192).  However, it has not been applied to young people in Canada. I enquired about the stratification of MINI-IPIP score (from the original website http://ipip.ori.org/), and was advised by Dr. John A. Johnson, Professor Emeritus of Psychology, Pennsylvania State University, to use local norms if possible, and report individual scores as below average, average, or above average. Accordingly, for the current study, we used local norms (from the participants) and opted for 3-category stratification. Based on the mean and standard deviation of the score in the sample, participants fell into three categories for each trait, i.e., low, average, and high (191,192). (Table 2.5)   Table 2.5 Mini-IPIP measures five personality traits. The total score for each personality trait were rated as low, average, and high based on the Mean and Standard of Deviation (SD) of the sample in the current study.   Variables Categories Extroversion Agreeableness Conscientiousness Neuroticism Imagination Low: Total score < (Mean-SD*)  Average: (Mean-SD) ≤ Total score ≤ (Mean + SD)  High: Total score > (Mean + SD) *Mean and Standard of Deviation calculated from the sample in the current study.   40  2.3.5 Canada Young Driver Survey (the entire survey) The entire “Canada Young Driver Survey” (Appendix D), including all the aforementioned instruments (Table 2.6), was reviewed by two English instructors and two high-school teachers to assess grammar and punctuation, as well as wording and comment for any ambiguities. They had no comments and found the survey clear and ready for use by youths. Thereafter, 40 participants (approached by YAC members) pilot-tested the whole survey before deployment to finalize the survey.  The YAC ensured that these participants were not those who had helped develop the survey previously (e.g., reviewing the leisure activity questionnaire) and this was the first time they were exposed to the survey. There was no change required afterwards.   Table 2.6 Canadian Young Driver Survey  The survey used in the current study consisted of four instruments  Instrument  Purpose Developed by Validation Basic sociodemographic  Sociodemographic factors General driving profile  Research team and YAC Research team and YAC (most variables’ categories adopted from valid resources)  The entire survey was reviewed by an additional 40 youth Behaviour in Young Novice Driver Scale (176,177)  Risky Driving behaviour Scott-Parker et al  Scott-Parker et al  Modified by Research team and YAC Leisure time activity questionnaire 12 Leisure activities (level of engagement) Research team and YAC Reviewed by 75 youth   Mini-IPIP questionnaire (190) Personality trait  Donnellan et al.  Donnellan et al.   2.4 Survey deployment Study data was collected and managed using REDCap electronic data capture tools hosted at the University of British Columbia (193). In consultation from YAC, it was decided to deploy a link to the survey via advertisement on popular social media networks including Facebook and Instagram. The promotional advertisement was a short video linked to the survey. The video was filmed by a team from YAC and their friends after obtaining verbal informed consent.  YAC was consulted on, and approved, the content and the text prompts contained in the video. To ensure that it would not affect participants’ answers, the video was scripted in a neutral way that would not elicit any positive or negative feelings in viewers before participating in the 41  survey which can be viewed at the following link: (https://www.facebook.com/samplepageresearch/videos/1476408445814697/).  Prior to publishing the link to the survey, which was embedded in the advertisement, a filter for age, location, and field of interest of viewers was set for age as 12-28 years, location as Canada, and field of interest as “all the options”.  The rationale behind extending the age range from 12-28 years, which was wider than the age inclusion criteria, i.e. 16-24 years, was that young people use invalid ages in their Facebook accounts (younger or older). Moreover, in some provinces/territories, youth are eligible for driving licenses at ages lower than 16. The YAC members randomly checked 100 of their Facebook contacts and found all ages were correct. The link to the survey, embedded in the advertisement, was published on Facebook and Instagram for three months: March 1st to May 31, 2018. The YAC members also posted the advertisement on other social media and encouraged their friends and families to disseminate the link as well.  Given that incentives (such as altruism, payment, recognition and visual props) may lead to increased response and engagement of the youth population for social research needs (194), the participants were entered into a draw for a chance to win one of fifty available $10 gift cards (e.g., Amazon or Starbucks). The option of incentive (i.e., a gift card) was suggested by the YAC. To ensure anonymity, the winners were provided with an online link via emails to claim their gift cards.  2.4.1 Data management and Ethical Consideration 2.4.1.1 Data collection REDCap (Research Electronic Data Capture) is a secure, web-based application designed to support data capture for research studies, providing 1) an intuitive interface for validated data entry, 2) audit trials for tracking data manipulation and export procedures, 3) automated export procedures for seamless data downloads to common statistical packages, 4) Canada-based server, and 5) procedures for importing data from external sources (193). REDCAp was supported by the Centre for Clinical Epidemiology and Evaluation’s (C2E2). To access data from the C2E2 instance of REDCap required research team members to use their UBC CWL (University of British Columbia Campus-Wide Login). The UBC IT department managed usernames and passwords for CWL, and the C2E2 REDCap Administrator supported access to the accounts. All data was stored on a secure institutional server at UBC.  42  2.4.1.2 Data confidentiality Survey data included no personal information or identifier except the email addresses of those who were interested in the draw. There was no link between the retained email addresses and the corresponding survey responses, hence anonymity was maintained. In addition, the email addresses were stored in a separate dataset, unlinked from Record IDs, and destroyed as soon as the winners of the draw were notified. However, the emails of the fifty winners of the draw were kept, as required by UBC finance, for 5 years after publication. All gift cards were provided via email to an online E Gift Card link; no physical mailing addresses were collected. 2. 5 Ethical Consideration The study was approved by the Behavioural Research Ethics Board of the University of British Columbia (H17-00086). 2.6 Data Analysis The diagram below illustrates the relationship of the constructs of the current study which can aid in understanding the analysis.  Figure 2.1 Construct relationship Diagram (Analysis perspective) The association between the leisure time activities (primary predictor variables) and risky driving behaviour (outcome variable), shown as a thick arrow (primary outcome variables), was studied. This association was adjusted for personality traits and sociodemographic factors (both are other confounders and also predictor variables) (thin arrows). The adjusted association was investigated with proportional odds logistic regression (POLR) analysis.   43  2.6.1 Variables The outcome (dependent) variable in the current study is risky driving behaviour and the predictive (independent) variables include leisure activities engagement level, personality traits, basic sociodemographics factors, and general driving profile.   Table 2.7 Variables and categories  Variable Name Measurement tool Classification Outcome Risky driving behaviour BYNDS questionnaire High risk (score ≥ 92) * Moderate risk (71≤ score ≤ 91)  Low risk (score ≤ 70) Predictor Leisure activity engagement level (13 leisure activities) Leisure activity questionnaire  High engagement **  Low engagement  Predictor Personality trait Mini-IPIP questionnaire Low (<Mean - SD) *** Average (Mean+/- SD)  High (> Mean + SD)  Predictor Sociodemographics General driving profile Sociodemographic questionnaire Refer to Final categories in Table 2.1 * The cut-off points are adopted from the literature (Table 2.2) ** The stratification in Low and High was explained in Table 2.4. *** Each of five personality traits was ranked as Low, Average, or High based on the Mean and Standard of deviation (SD) of the current sample (Table 2.5)      2.6.2 Statistical analysis plan The theoretical framework (Figure 1.5) illustrates possible ways that leisure activity can influence driving behaviour. In addition, it shows that personality and demographic factors are also able to affect driving behaviour. The Canadian Young Driver Survey (Table 2.6) collected all these variables to help shed more light on the possible association of leisure activity with driving behaviour while also considering personality and demographic factors. Raw data were exported from REDCap to Excel. The data was examined for completeness, cleaned, and coded for predetermined categorization of outcome and predictor variables; then, transferred to SPSS.  To investigate the association between leisure activity and risky driving behaviour in young drivers in Canada, the following analyses were conducted.  44  2.6.2.1 Descriptive Analysis Descriptive analyses were conducted to examine the characteristics of participants in terms of leisure activities, personality traits, demographic, and driving behaviours. All continuous variables were reported as means and standard deviations, and categorical variables as proportions.  Bivariate analyses with Chi-square statistics were used to examine the differences in leisure activities, personality traits, and demographics among the three groups of risky driving behaviours.  Along with the chi-square tests, the contingency table of counts and proportions of the characteristics of the survey participants was presented. Also, each group of outcome variables, i.e. risky driving behaviour (high, moderate, low) was presented Although the Chi square would not impose any assumptions about the functional form of that relationship, it was appropriate for the purpose of testing to determine whether or not a relationship or an association existed between each factor and risky driving behaviour.  Thereafter, Spearman rank-order correlation test was applied to examine, first, the correlations between predictor variables (leisure activities, demographic factors and personality traits) and outcome variable (risky driving behaviour); then, the correlation among all the predictor variables themselves, to identify whether there were any highly correlated variables. Generally, it was possible to infer the direction of association by examining the descriptive statistics after Chi square and Spearman correlation; however, the formal test of association was left to the regression analysis. 2.6.2.2 Proportional Odds Logistic Regression analysis The next step was the Proportional Odds Logistic Regression (POLR) analysis, since the outcome variable was ordinal, low, moderate, or high risky driving behaviour, was applied to formally test these associations with estimates of the magnitude and direction of association in the form of an odds ratio. The main advantage of this approach is that all explanatory factors are modelled simultaneously so that any confounding effects that may be present can be controlled. The model also considers the ordinal nature of the data. To avoid multicollinearity, highly correlated variables were first identified and then the most appropriate variables fitted the model were singled out. Highly correlated variables might not only hurt the precision of the estimate coefficient in the model but also undermine the power of the model. In other words, having highly correlated variables in the model might lead to decreased level of significance for the real influential predictors. In turn, the actual non-correlated variables might happen to have higher coefficient magnitudes compared to the actual correlated 45  variables and, in a backward elimination procedure, be selected for the model instead. Given that the predictors are categorical, the correlation was assessed with the Spearman's rank statistic (rho). The cut-off value for deciding which predictors were too correlated was randomly set to rho > 0.77. In the case of two highly correlated variables, the goodness of fit test would be applied to identify which one is, statistically, the better fit for the model. Given that there were no highly correlated variables were found according to the rho cut-off value, goodness of fit processing was not required (Table B.1, Appendix B).  In analysis, no highly correlated variables were found. Consequently, goodness of fit processing was not applied (Table B.1, Appendix B). First, all the potential predictors from personality traits, leisure activities and sociodemographic factors were examined in a POLR to construct a full model for risky driving behaviour. Afterwards AIC stepwise selection (combination of forward and backward elimination) was applied to identify the best model, i.e., the most parsimonious model, which includes the variables that best fit the regression predictive model. Odds ratios (ORs) with 95% confidence intervals were reported. Stepwise regression (or stepwise selection) yields a predictive model which consists of the best subset of predictor variables selected by iteratively adding/removing predictors to/from the predictive model. This selection procedure ends up with the best performing model, i.e., Parsimonious model, which is a model that lowers prediction error. There are three strategies for stepwise regression (195,196): • Forward selection, which starts with no predictors in the model, iteratively adds the most contributive predictors, and stops when the improvement is no longer statistically significant. • Backward selection (or backward elimination), which starts with all predictors in the model (full model), iteratively removes the least contributive predictors, and stops when you have a model where all predictors are statistically significant. • Stepwise selection (or sequential replacement), which is a combination of forward and backward selections. This method starts with no predictors, then sequentially adds the most contributive predictors (like forward selection). After adding each new variable, any variables that no longer provide an improvement in the model fit will be removed (like backward selection). The current study practised fitting the final selected model through backward elimination to determine the best predictive variables with statistically significant influence on driving behaviour. Of note, this analysis was exploratory, so there was no adjustment for multiple comparisons. A single regression model was applied to formally test the associations. 46  2.7 Sample size considerations The required sample size depended on the prevalence of the outcome in high risk driving (composite score ≥ 92 in BYNDS), and the prevalence of the exposure (either leisure activity or personality trait). Unfortunately, there are no data on the prevalence of risky driving behaviour, especially based on the composite BYNDS scores, in our population; although there is a prevalence for each risky driving behaviours such as not fastening seatbelts, fatigue, speeding, and impaired driving (113). There is also no information on the prevalence of different leisure activities among Canadian youth. Thus, for the sample size calculations, the following prevalence was conservatively assumed: a 10% prevalence of high risk driving and 60% prevalence for high-engagement leisure activity.  Of note, low outcome prevalence is harder to detect, thus a conservative assumption was made that the prevalence of high risk driving would only be 10%. Accordingly, minimum sample size of 812 survey participants was calculated to detect a relative risk (RR) of 2 (2 was used as the benchmark since it represents a significant RR, i.e., a doubling of risk), with a power of 80%, and at a significance level of 0.0045 (adjusted for multiple comparison of 13 leisure activities). This calculation is based on a simple comparison of independent proportions, assuming a dichotomous outcome and exposure. The planned statistical analysis considered multiple predictors simultaneously.  In summary, the sample size (n=812) definitely expressed the RR=2 as an OR based on the assumptions below: • Prevalence of outcome, i.e., High-risk driver (overall, in exposed and unexposed) = 0.1 • Prevalence of exposure, i.e., High engagement in leisure activity = 0.6 • Prevalence of outcome in exposed = 0.125  • Prevalence of outcome in unexposed = 0.0625 • RR = 0.125/0.0625 = 2 • OR = (0.125 / (1- 0.125)) / (0.0625 / (1- 0.0625)) = 2.14 (note that when the outcome is rare, the OR is very close to the RR) Thus, this study required a minimum of 812 participants in order to detect an OR of 2.14 (or equivalently an RR of 2) with 80% power.  47   3. Results On the survey closing day (date), 1,756 individuals responded to the survey and 964 respondents fully completed the survey and had their responses included in the statistical analysis.  (Figure 3.1)  Figure 3.1 Survey response rate.   The diagram depicts the completion rates at each section of the survey. The order of the instruments in the survey was: Cover letter (consent form), Sociodemographic Questionnaire, Leisure Activity Questionnaire, Behaviour of Young Novice Driver Scale (BYNDS), Personality Trait questionnaire (Mini IPIP). 3.1 General description of study population  This section presents a descriptive summary of sociodemographic, personality trait and level of leisure activity information from the 964 surveyed drivers. 3.1.1 Sociodemographic characteristics in study participants  Most of the 964 participants, aged 18.34±2.31, were male, beginner drivers (i.e., 16-19 years old), living with both parents and had less than a high school diploma. They were predominantly both employed and a student at the same time and had high family socioeconomic status. Most had a 1765 Participants showed interest and clicked through the survey286 Participants did not consent/complete the sociodemographic questioannaire1470 completed sociodemographic questioannaire303 Participants did not complete leisure activity questionnaire1167 completed leisure activity questionnaire104 Participants did not complete BYNDS1063 completed BYNDS99 Participants did not complete Mini IPIP964 completed Canada Young Driver Survey48  driving license with restriction, owned their cars, drove independently for less than a year, and had low driving exposure. The sociodemographic characteristics and the comparison between the three groups of risky driving behaviour are presented in Table 3.1.  3.1.2 Personality traits in the study population  Table 3.2 represents the personality trait of the 964 participants and the comparison between the three driver groups. For each of the five personality traits considered in this research the mean scores and standard deviations were calculated. On average, participants scored highest on the personality dimension of imagination (14.94±3.12) followed by agreeableness (14.46±3.04), conscientious (12.75±3.09) and neuroticism (11.39±3.84). 3.1.3 Level of engagement in leisure activities in the study participants The majority of participants in all three groups (low, moderate and high risk groups) had low engagement levels in playing musical instruments and/or listening to music, video games, sports, movies, spending time with friends, spending time with family, alcohol, drugs, art, TV, writing/reading, and volunteering, whereas most participants (72.2%) had high engagement in social media (Table 3.3). 3.1.4 Driving behaviour profile in study population  Based on the total score of the BYNDS questionnaire, drivers were grouped into three risky driving categories: low risk (score <71), moderate risk (71-91), and high risk (>91). In this study, the majority of respondents were categorized as high risk drivers (46.9%), followed by moderate risk drivers (32.4%).  (Figure 3.2) Figure 3.2 Driver category distribution in the study population   20.7 %32.4  %46.9 %0.0010.0020.0030.0040.0050.00Low Risk (n=200) Moderate Risk (n=312) High Risk (n=452)49  3.2 Characteristics of driving behaviours and corresponding differences among driver groups 3.2.1 Sociodemographic characteristics in driver categories Low risk drivers, aged 17.38±2.11 years were mostly found to be categorized as:  “other” (i.e. non-males) gender category, beginner drivers, living with both parents, high family socioeconomic status, less than a high school diploma, and a “student”. With respect to driving profile, they predominately held a driving license with restriction, did not own a car, and drove independently for less than a year with low driving exposure.   The majority of moderate risk drivers aged 18.51±2.39 years were male beginner drivers, living with both parents, had less than a high school diploma, were “both employed and student”, and had high family socioeconomic status. Their driving profile shows that they predominantly held a driving license with restriction, had their own car, and drove independently for less than 1 year with low driving exposure.   High risk drivers, 18.75±2.23 years old, were mostly male beginner drivers, were living with both parents, had high family socioeconomic status, had some post-secondary education, and were “both employed and student”. They predominately had a driving license with restriction, owned their car, drove independently for 1-3 years, and had low driving exposure. (Table 3.1)  3.2.1.1 Association between sociodemographic and driving behaviours Chi-square test was used to explore the relationships of sociodemographic factors and risky driving behaviors. Most sociodemographic factors including gender, age, living status, education level, employment status, independent driving duration, vehicle ownership, driving exposure, and driver license status were statistically significant in association with driving behaviours with a p-value of <0.0001, which was consistent with the hypothesis that there is an association between sociodemographic factors and risky driving behaviours. Socioeconomic status was the only sociodemographic factor with no significant association with risky driving behaviour.  As is shown in table 3.1 high risk drivers were more likely than low risk drivers to be male, young adult drivers, have “some post-secondary education”, and were “both employed and a student”. Moreover, they were more likely to live alone or with friends/roommates compared to the low risk drivers.  High risk drivers tended to have been driving for more than 3 years, while low risk drivers 50  tended to drive for less than a year. High exposure to driving for more than 20 hours per week was more prevalent time among high risk drivers compared to low risk drivers. In comparison with low risk drivers, high risk drivers were more likely to have their own cars. Table 3.1 Socio-demographic characteristics of study population and driving behaviour categories   Variables Group P-Value*  All participants (n=964) Low Risk (n=200)  Moderate Risk (n=312)  High Risk (n=452)  Gender Male Other**  86 (43.0%) 114(57.0%)  180 (57.7%) 132 (42.3%)  331 (73.2%) 121 (26.8%)  <0.0001   597 (61.9%) 367 (38.1%) Driving Status (Based on age) Beginner (16-19 years old) Young Adult (20-24 years old)  176 (88.0%) 24 (12.0%)  222 (71.2%) 90 (28.8%)  297 (65.7%) 155 (34.3%) <0.0001  695 (72.1%) 269 (27.9%) Living Status Family (Both parents) Family (Single parent) Friend(s) or Alone Partner  145 (72.5%) 28 (14.0%) 20 (10.0%) 7 (3.5%)  202 (64.7%) 51 (16.3%) 39 (12.5%) 20 (6.4%)  260 (57.5%) 83 (18.4%) 81 (17.9%) 28 (6.2%) 0.0125   607 (63.0%) 162 (16.8%) 140 (14.5%) 55 (5.7%) Educational level Less than high school diploma High school graduated Some postsecondary  146 (73.0%) 15 (7.5%) 39 (19.5%)  142 (45.5%) 66 (21.2%) 104 (33.3%)  168 (37.2%) 99 (21.9%) 185 (40.9%) <0.0001  456 (47.3%) 180 (18.7%) 328 (34.0%) Employment Status Employee Student Both (Employee and student) Unemployed  11 (5.5%) 109 (54.5%) 78 (39.0%) 2 (1.0%)  58 (18.6%) 105 (33.7%) 147 (47.1%) 2 (0.6%)  116 (25.7%) 100 (22.1%) 233 (51.5%) 3 (0.7%) <0.0001  185 (19.2%) 314 (32.6%) 458 (47.5%) 7 (0.7%) Driving Independently Duration <1 year 1-3 years >3 years  156 (78.0%) 35 (17.5%) 9 (4.5%)  138 (44.2%) 112 (35.9%) 62 (19.9%)  144 (31.9%) 172 (38.1%) 136 (30.1%) <0.0001   438 (45.4%) 319 (33.1%) 207 (21.5%) Socioeconomic Status *** Low Moderate High  2 (1%) 30 (15%) 168 (84%)  3 (1%) 35 (11.2%) 274 (87.8%)  2 (0.4%) 54 (11.9%) 396 (87.6%) 0.6063  7 (0.7%) 119 (12.3%) 838 (86.9%) Vehicle owned by driver Yes No  39 (19.5%) 161 (80.5%)  164 (52.6%) 148 (47.4%)  346 (76.5%) 106 (23.5%) <0.0001  549 (57%) 415 (43%) Driving Exposure (Hour/Week) Low (<11) Moderate (11-20) High (>20)  193 (96.5%) 4 (2%) 3 (1.5%)  248 (79.5%) 38 (12.2%) 26 (8.3%)  277 (61.3%) 111 (24.6%) 64 (14.2%) <0.0001  718 (74.5%) 153 (15.9%) 93 (9.6%) Driver license class With restriction Without restriction  181 (90.5%) 19 (9.5%)  224 (71.8%) 88 (28.2%)  284 (62.8%) 168 (37.2%) <0.0001  689 (71.5%) 275 (28.5%) * Based on qui square      **Other” includes female and transgender spectrum    ***Low SES= No home and No Car; Moderate SES= ≤ 2 cars and No home; High SES= Home AND/OR car>2  51  3.2.2 Personality trait in driver categories The majority of participants in all three groups (low, moderate, and high risk drivers) scored average in all five personality trait categories, i.e. extroversion, agreeableness, conscientiousness, neuroticism, and imagination. (Table 3.2) 3.2.2.1 Association between personality traits and driving behaviour Extroversion was the only personality trait associated with driving behaviours (p-value < 0.0001). Although low, moderate and high-risk drivers were predominately scored average on all personality traits, high risk drivers tended to have higher scores on extroversion compared to low risk drivers (23.9% vs. 11.0% respectively) (Table 3.2). No statistically significant association was found between other personality traits (i.e., neuroticism, agreeableness, conscientiousness, and imagination) and driving behaviours. However, there was a trend for “high risk” drivers to score lower in agreeableness, conscientiousness, neuroticism, and imagination (Table 3.2). Table 3.2 Distribution personality traits in the study population and driving behaviour categories  Category Group p-value  All participants (n=964) Low Risk  (n= 200)  Moderate Risk  (n= 312)  High Risk  (n= 452)  Extroversion trait score (Mean±SD) 11.27±3.89 12.13±3.88 12.87±4.15  12.30±4.06 Extroversion trait category Low  Average  High  57 (28.5%) 121 (60.5%) 22 (11.0%)  67 (21.5%) 196 (62.85) 49 (15.7%)  74 (16.4%) 270 (59.7%) 108 (23.9%) <0.0001  198 (20.5%) 587 (60.9%) 179 (18.6%) Agreeableness trait Score (Mean±SD) 14.77±2.81 14.58±2.97 14.24±3.18  14.46±3.04 Agreeableness trait category Low  Average  High  22 (11.0%) 145 (72.5%) 33 (16.5%)  44 (14.1%) 218 (69.9%) 50 (16.0%)  86 (19.0%) 295 (65.3%) 71 (15.7%) 0.1003  152 (15.8%) 658 (68.3%) 154 (16%) Conscientiousness trait Score (Mean±SD) 13.01±3.06 12.89±2.92 12.54±3.20  12.75±3.09 Conscientiousness trait category Low  Average  High  26 (13.0%) 132 (66.0%) 42 (21.0%)  42 (13.5%) 205 (65.7%) 65 (20.8%)  85 (18.8%) 277 (61.3%) 90 (19.9%) 0.2375  153 (15.9%) 614 (63.7%) 197 (20.4%) Neuroticism trait Score  (Mean±SD) 11.72±3.78 11.36±3.78 11.27±3.92  11.39±3.84 Neuroticism trait category Low  Average  High  28 (14.0%) 142 (71.0%) 30 (15.0%)  45 (14.4%) 222 (71.2%) 45 (14.4%)  79 (17.5%) 293 (64.8%) 80 (17.7%) 0.3554  152 (15.8%) 657 (68.2%) 155 (16.1%) Imagination trait Score  (Mean±SD) 15.22±3.05 15.12±3.06 14.70±3.18  14.94±3.12 Imagination trait category Low  Average  High  22 (11.0%) 148 (74.0%) 30 (15.0%)  36 (11.5%) 235 (75.3%) 41 (13.1%)  70 (15.5%) 327 (72.3%) 55 (12.2%) 0.3755  128 (13.3%) 710 (73.7%) 126 (13.1%) 52  3.2.3 Level of engagement in leisure activities in the driving categories As is shown in Figure 3.3, participants in all three groups (high, moderate or low risk) spent time on social media more than any other activity. Sixty eight percent of low risk drivers, 68.6% of moderate risk drivers, and 72.2% of high risk drivers spent more than 60 minutes of their daily time on social media. The majority of participants in all three groups had low engagement levels in playing musical instruments and/or listening to music, video games, sports, movies, spending time with friends, spending time with family, alcohol, drugs, art, TV, writing/reading, and volunteering. However, it appears like slightly more low risk drivers engaged in art, reading/writing, family, and volunteering compared to high risk drivers, while slightly more high risk drivers engaged in watching movies, alcohol, drug, music, sport, and TV compared to low risk drivers (Figure 3.3).  Figure 3.3 Percentage Distribution of high engagement in leisure Activities in driving behaviour groups  3.2.3.1 Association between leisure activities and driving behaviour Among all leisure activities, the engagement level of music, video games, social media, movies, family, friends, drug- and alcohol involvement, and volunteering, were all found to have statistically significant association with driving behaviours (p <0.05). Compared to low risk drivers, high-risk drivers were more likely to be highly engaged in music (45.4% vs. 37.5%), movies (15.3% vs. 7.0%), social media (76.5% vs. 68.0%), leisure activities involving alcohol (8.0% vs. 1.0%) or drugs (16.0% vs. 2.5%), and spending time with friends (46.7% vs. 25.0%) while they tended to have low engagement in art (88.0% vs. 79.5%) , reading/writing (88.7% vs. 76.0%), volunteering (94.7% vs. 0102030405060708090Low Risk Moderate Risk High Risk OverallLeisure Activity -High Level of EngagementVideo GameSportSocial MediaMovieFamilyFriendsAlcoholDrugArtWriting/Reading53  91.5%), and spending time with family (87.6% vs. 79.5%). (Table 3.3) These results support the hypothesis that there is an association between leisure activities and risky driving behaviour. It is therefore worthwhile to look at the possible correlations and independent influence of leisure activities on driving behaviour in more detail. To that end, Proportional Odds Logistic Regression (POLR) analysis was conducted which will be discussed in next section. Table 3.3 Distribution of study population according to Leisure activity  Variables Group P-value*  All participants (n=964) Low Risk  (n= 200)  Moderate Risk  (n= 312)  High Risk  (n= 452)  Video game engagement level Low  High   171 (85.5%) 29 (14.5%)  244 (78.2%) 68 (21.8%)  383 (84.7%) 69 (15.3%) 0.0329  798 (82.8%) 166 (17.2%) Social media engagement level Low High  64 (32.0%) 136 (68.0%)  98 (31.4%) 214 (68.6%)  106 (23.5%) 346 (76.5%) 0.0179  268 (27.8%) 696 (72.2%) Movie engagement level Low High  186 (93%) 14 (7%)  267 (85.6%) 45 (14.4%)  383 (84.7%) 69 (15.3%) 0.0126  836 (86.7%) 128 (13.3%) Family engagement level** Low High  159 (79.5%) 41 (20.5%)  289 (84.4%) 52 (15.2%)  396 (87.6%) 56 (12.4%) 0.0275  820 (85.1%) 144 (14.9%) Friends engagement level*** Low High  150 (75.0%) 50 (25.0%)  228 (73.1%) 84 (26.9)  241 (53.3%) 211 (46.7%) <0.0001  619 (64.2%) 345 (35.8%) Alcohol engagement level Low High  198 (99.0%) 2(1.0%)  303 (97.1%) 9 (2.9%)  416 (92.0%) 36 (8.0%) <0.0001  917 (95.1%) 47 (4.9%) Drug engagement level Low High  195 (97.5%) 5 (2.5%)  294 (97.1%) 18 (2.9%)  380 (84.1%) 72 (15.9%) <0.0001  869 (90.1%) 95 (9.9%) Art engagement level Low High  159 (79.5%) 41 (20.5%)  262 (84%) 50 (16%)  452 (100%) 0 (0.0%) 0.0159  873 (90.6%) 91 (9.4%) Writing/Reading engagement level Low High  152 (76.0%) 48 (24.0%)  240 (76.9%) 72 (23.1%)  401 (88.7%) 51 (11.3%) <0.0001  793 (82.3%) 171 (17.7%) Volunteering engagement level Low High  183 (91.5%) 17 (8.5%)  279 (89.4%) 33 (10.6%)  428 (94.7%) 24 (5.3%) 0.0239  890 (92.3%) 74 (7.7%) Music engagement level Low High  125(62.5%) 75 (37.5%)  207 (66.3%) 105 (33.7%)  247 (54.6%) 205(45.4%) 0.0038  579 (60.1%) 385 (39.9%) Sport engagement level Low High  140 (70.0%) 60 (30.0%)  210 (67.3%) 102 (32.7%)  311 (68.8%) 141 (31.2%) 0.8057  661 (68.6%) 303 (31.4%) TV engagement level Low High  156 (78.0%) 44 (22.0%)  237 (76.0%) 75 (24.0%)  333 (73.7%) 119 (26.3%) 0.472  726 (75.3%) 238 (24.7%) * Based on qui square     ** Engagement levels did not include time spent with family at home doing their own tasks *** Engagement levels did not include the activities done with friends during school time   54  3.3 Proportional Odds Logistic Regression (POLR) analysis As noted in section 2.6.2.2, we applied the Proportional Odds Logistic Regression analysis to formally test the aforementioned associations with estimates of the magnitude and direction of association in the form of an odds ratio. Given that no highly correlated variables were found, goodness of fit processing was not applied. (Refer to tables B.1, Appendix B). Therefore, first, all the potential predictors from personality traits, leisure activities and sociodemographic factors were examined in a POLR, i.e. a full model (below, Section 3.3.1).  Afterwards AIC stepwise selection was applied to identify the best model, i.e. a parsimonious model that is neither an over-fit nor under-fit (See Section 3.3.2, below).  Of note, in both models, ORs for the predictive variables show the probability of change in the ranking order from low-risk driving behaviour to moderate- and high- risk driving behaviour. 3.3.1 Full Model In the Full model, those predictor variables which significantly improve the model are reported ( i.e. those that when removed are found to result in a significant decrease in the model fit, are reported) Other predictors or variables with no significant influence on risky driving behaviour were also included which helps to get more accurate coefficients in the models and discriminate between the target (high- and moderate- risk driver) and reference category (low-risk driver) and against other important variables.  3.3.1.1 Predictive variables in positive correlation with risky driving behavior- Full Model As shown in table 3.4, car ownership, high /moderate driving exposure, male gender, driving for more than 3 years, being both student and employee, high neuroticism, highly engaged in drug-involved leisure activity, spending time with friends, and spending time on social media were all positively correlated with risky driving behaviours. For those who had their own cars, the odds of being a high and moderate risk versus low risk driver were 3.1 times higher than for those who did not own their cars (OR=3.12). Driving for more than 20 hours per week put the individual 2.5 times more at risk of risky driving behaviour than those who drove less than 11 hours per week (OR=2.49). Male drivers were 2.5 times more likely to be high risk drivers than other genders (OR=2.47). When it comes to employment status, the odds of being a high risk driver among those who were students and employed at the same time was 1.6 times as large as the odds for those who were just a student 55  (0R=1.56). In term of personality trait, risky driving behaviour was 78% more likely in those with high neuroticism than those with average neuroticism (OR=1.78). Among leisure activity, risky driving behaviour was 2.2 times more common in those with high drug engagement than those with low engagement (OR=2.23). Spending time with friends more than 4 times a week also was a significant predictor for risky driving behaviour (OR=1.94). And lastly being highly engaged on social media for more than 60 minutes daily put individuals 1.8 times more at risk of risky driving behaviour (OR=1.82).  Table 3.4 Factors with significant influence on risky driving behaviour- Full model Variable OR OR Range Reference Variable Sociodemographic characteristics    Car ownership (Yes) 3.12 2.28- 4.27 Driver Car Status (No) Driving Exposure (High) 2.49 1.48- 4.28 Driving Exposure (Low) Gender (Male) 2.47 1.80- 3.40 Gender (Other) Driving Exposure (Mod) 2.44 1.58- 3.81 Driving Exposure (Low) Duration of driving independently >3 years 1.93 1.09- 3.43 Duration of driving independently 1-3 years Employment status (Both employed and student) 1.56 1.13- 2.13 Employment status (Student) Personality Traits    Neuroticism Trait (High) 1.78 1.19- 2.68 Neuroticism Trait (Average) Leisure activities    Drug Engagement Level (High engagement) 2.23 1.29- 3.10 Drug Engagement Level (Low engagement) Friend Engagement Level (High engagement) 1.94 1.43- 2.65                   Friend Engagement Level (Low engagement) Social Media Engagement Level (High engagement) 1.82  1.33-2.48 Social Media Engagement Level (Low engagement)                 *variables ordered from highest to lowest odds ratio in each category (CI=95%) (P-value < 0.05)   3.3.1.2 Predictive variables with protective effect on risky driving behavior- Full Model Educational level “less than high school”, driving independently for less than a year, highly engaged in reading/writing, volunteering, and playing video games were protective factors for risky driving behaviour. In other words, the drivers with less than high school education were 35% less likely to engage in risky driving compared to those with a high or secondary school diploma (OR=0.65). Risky driving behaviour was 44% less likely in those who drove independently for less than a year compared to those who drove above one year (OR=0.56). Drivers with high engagement in reading and writing had 38% lower odds of being risky drivers than those with lower engagement 56  (OR=0.62). High engagement in volunteering activities put an individual at a 40% lower likelihood of risky driving behaviour (OR=0.60). Finally, those who played video games more than 9 hours per week had odds 0.56 times as great to be a risky driver than those who played less than 9 hours (Table 3.5).  Table 3.5 Factors with significant protective influence against risky driving behaviour- Full model Variable OR OR Range Reference Variable Sociodemographic characteristics    Education Status  (Less than high school) 0.65 0.42- 0.98 Education Status (High /secondary school diploma or equivalent)  Duration of driving independently (<1 year) 0.56 0.38- 0.81 Duration of driving independently (1-3 years) Leisure activities    Reading/Writing Engagement Level (High) 0.62 0.44- 0.88 Reading/Writing Engagement Level (Low) Volunteering Engagement Level (High) 0.60 0.37- 0.98 Volunteering Engagement Level (Low) Video Game Engagement Level (High) 0.56 0.38- 0.82 Video Game Engagement Level (Low)           *variables ordered from highest to lowest odds ratio in each category (CI=95%) (P-value < 0.05)  3.3.1.3 Final Full Model Equation Y= 3.12 X1 + 2.49 X2 + 2.47 X3 + 2.44 X4 + 2.23 X5 + 1.94 X6 + 1.93 X7 + 1.82 X8 + 1.79 X9 + 1.56 X10 – 0.65 X11 – 0.62 X12 – 0.60 X13 – 0.56 X14 – 0.56 X15 + (Other Variables) * (Predictive variables’ order is based on strength of association)  Outcome variable Y= Risky driving behaviour  Predictive variables  Sociodemographic  X1= Car ownership, OR= 3.12 (CI 95%, 2.28-4.27) X2= Driving Exposure (High), OR = 2.49 (CI 95%, 1.47-4.28) X3= Gender (Male), OR= 2.47(CI 95%, 1.79-3.4) X4= Driving Exposure (Moderate), OR=2.44 (CI 95%, 1.58-3.81) X7= Duration of driving independently (>3 years), OR= 1.93 (CI 95%, 1.09-3.42) X10= Employment status (Both Employed and student), OR= 1.56 (CI 95%, 1.13-2.13) 57  X11= Education Status (Less than High school), OR= 0.65 (CI 95%, 0.42-0.98) X15= Duration of driving independently <1 year, OR= 0.56 (CI 95%, 0.38-0.81) Leisure activities engagement level  X5= Drug Engagement Level (High engagement), OR= 2.23 (CI 95%, 1.28-3.99) X6= Friend Engagement Level (High engagement), OR= 1.94 (CI 95%, 1.42-2.65) X8= Social Media Engagement Level (High engagement), OR= 1.82 (CI 95%, 1.33-2.48) X12= Reading/Writing Engagement Level (High engagement), OR= 0.62 (CI 95%, 0.43-0.88) X13= Volunteering Engagement Level (High engagement), OR= 0.60 (CI 95%, 0.36-0.98) X14= Video Game Engagement Level (High engagement), OR= 0.56 (CI 95%, 0.38-0.81) Personality trait X9= Neuroticism trait (High), OR= 1.79 (CI 95%, 1.19-2.68) *Other Variables are non-highlighted variables found in the Tables in Appendix B. 3.3.2 Parsimonious model (Stepwise AIC selection model) The current study practised fitting the final selected model through backward elimination to determine the best predictive variables with statistically significant influence on driving behaviour.   3.3.2.1 Predictive variables in positive correlation with risky driving behavior- Parsimonious Model From the leisure activity perspective, participants who were highly engaged in drug-involved leisure activity, spending time with friends, on social media, and watching movies were respectively 2.0, 1.9, 1.8, and 1.5 times more likely to be self-reported as high risk and moderate risk drivers, . In term of personality traits, in addition to neuroticism, high extroversion and low imagination also became significant risk factors for risky driving behaviour in the stepwise model. For sociodemographic characteristics, the stepwise model included no new coefficients compared to the Full Model (see 3.3.1.1) (Table 3.6). 58  3.3.2.2 Predictive variables with protective effect on risky driving behavior- Parsimonious Model Protective variables were similar to those in the Full model. Among leisure activities, high engagement in reading/writing, volunteering, and playing video games were found to be protective factors against risky driving behaviour with virtually similar coefficients and interpretations (Table 3.7). Table 3.6 Factors with significant influence on risky driving behaviour- Parsimonious model Variable OR OR Range Reference Variable Sociodemographic characteristics    Car ownership (Yes) 3.01 2.21- 4.11 Car ownership (No)  Driving Exposure (High)  2.58  1.54- 4.42  Driving Exposure (Low)  Driving Exposure (Moderate)  2.57  1.69- 3.97  Driving Exposure (Low)  Gender (Male)  2.51 1.85- 3.42 Gender (Other) Duration of driving independently (>3 years)  Employment status (Both employed and student) 2.11  1.58 1.26- 3.57  1.16-2.16 Duration of driving independently (1-3 years)  Employment status (Student) Personality trait     Neuroticism Trait (High)  1.83 1.23- 2.74 Neuroticism Trait (Average) Extroversion Trait (High)  1.60 1.09- 2.36 Extroversion Trait (Average) Imagination Trait (Low) 1.53 1.02- 2.34 Imagination Trait (Average) Leisure activities    Drug Engagement Level (High)  2.09 1.22- 3.72 Drug Engagement Level (Low) Friend Engagement Level (High)  1.98 1.47- 2.69 Friend Engagement Level (Low) Social Media Engagement Level (High)  1.83 1.35- 2.49 Social Media Engagement Level (Low) Movie Engagement Level (High) 1.52 1.01- 2.31 Movie Engagement Level (Low)           *variables ordered from highest to lowest odds ratio in each category (CI=95%) (P-value < 0.05)    Table 3.7 Factors with significant protective influence on risky driving behaviour- Parsimonious model Variable OR OR Range Reference Variable Sociodemographic characteristics    Education Status (Less than High school) 0.64 0.42- 0.97 Education Status (High (secondary) school diploma or equivalent) Duration of driving independently <1 year 0.55 0.38- 0.79 Duration of driving independently 1-3 years 59  Leisure activities    Reading/Writing Engagement Level (High) 0.61 0.43- 0.86 Reading/Writing Engagement Level (Low) Volunteering Engagement Level (High) 0.59 0.37- 0.97 Volunteering Engagement Level (Low) Video Game Engagement Level (High) 0.57 0.39- 0.81 Video Game Engagement Level (Low)          *variables ordered from highest to lowest odds ratio in each category (CI=95%) (P-value < 0.05)   3.3.2.4 Final Parsimonious Model Equation Y= 3.01 X1 + 2.58 X2 + 2.56 X3 + 2.51 X4 + 2.11 X5 + 2.09 X6 + 1.98 X7 + 1.83 X8 + 1.83 X9 + 1.60 X10 + 1.58 X11 + 1.53 X12 + 1.52 X13 – 0.64 X14 – 0.61 X15 – 0.60 X16 – 0.56 X17 – 0.55 X18   Outcome variable Y= Risky driving behaviour  Predictive variables Sociodemographic    X1= Car ownership (Yes), OR=3.01 (CI 95%, 2.21-4.11) X2= Driving Exposure (High), OR=2.58 (CI 95%, 1.54-4.41) X3= Driving Exposure (Moderate), OR=2.56 (CI 95%, 1.68-3.96) X4= Gender (male), OR=2.51 (CI 95%, 1.85-3.42) X5= Duration of driving independently (>3 years), OR=2.11 (CI 95%, 1.25-3.57)  X11= Employment status (Both employed and student), OR=1.58 (CI 95%, 1.16-2.16) X14= Education Status (Less than high school), OR=0.64 (CI 95%, 0.42-0.97) X18= Duration of driving independently <1year, OR=0.55 (CI 95%, 0.38-0.7) Leisure activities X6= Drug Engagement Level (High engagement), OR=2.09 (CI 95%, 1.21-3.71) X7= Friend Engagement Level (High engagement), OR=1.98 (CI 95%, 1.46-2.68) X8= Social Media Engagement Level (High engagement), OR=1.83 (CI 95%, 1.34-2.49) X13= Movie Engagement (High engagement),), OR=1.52 (CI 95%, 1- 2.31) X15= Reading/Writing Engagement Level (High engagement), OR=0.61 (CI 95%, 0.42-0.85) X16= Volunteering Engagement Level (High engagement), OR=0.60 (CI 95%, 0.36-0.96) X17= Video Game Engagement Level (High engagement), OR=0.56 (CI 95%, 0.38-0.81) Personality traits 60  X9= Neuroticism Trait (High), OR=1.83 (CI 95%, 1.23-2.73) X10= Extroversion Trait (High), OR=1.60 (CI 95%, 1.09-2.35) X12= Imaging Trait (Low), OR=1.53, (CI 95%, 1.01-2.34)  To investigate the correlation between the different variables (sociodemographics, personality traits and leisure activities) and risky driving behaviours, both the full and stepwise selection models were performed. Figure 3.4 illustrates the full model and stepwise model side by side for the sake of comparison and some differences were noted. Among all variables, high engagement in movies, low imagination, and high extroversion were correlated with risky driving behaviours according to the stepwise selection model while no significant corresponding correlation was found based on the full model. 61  Figure 3.4 Forest plot - Full vs. Stepwise selection Models.  Red line: significant correlation (P-value < 0.05).   62   4. Discussion  In many Western countries, adolescent risk behaviors are among the top public health priorities and numerous prevention programs have been developed to reduce young people’s engagement in risky behaviours (197). As youth grow and reach their developmental competencies, there are contextual variables that promote or hinder the process. These are frequently referred to as protective and risk factors. Various risk and protective factors influence young people’s attitudes and behaviours. By definition, risk factors are any factors associated with the increased likelihood of an undesired behaviour or outcome, while protective factors reduce the likelihood of a bad outcome or help individuals avoid potentially harmful behaviour, and/or promote an alternative pathway (198). A growing body of cross-cultural evidence indicates that various psychological, social, and behavioural risk and protective factors affect individual’s health, especially during adolescence (199,200). This study employed a survey-based cross-sectional design to determine the driver characteristics associated with risky driving, i.e. the linkages between risky driving behaviours and various leisure activities, and the potential predictors of risky driving behaviour among young drivers in Canada. The findings showed that 46.9% respondents in this study self-reported with high risk driving behaviour which is concerning. Based on results, some of the sociodemographic variables, personality traits, and leisure activities were risk factors for risky driving behaviour and will be discussed individually in next sections. 4.1 Sociodemographic factors and driving behaviour  The first objective of this study explored the relationship between sociodemographic factors and risky driving behavior among young drivers in Canada. The findings supported the study hypothesis that sociodemographic factors affect driving behaviour in both positive and negative ways. Ten sociodemographic variables considered for evaluation will be discussed individually below.  4.1.2 Age and driving behaviour In the unadjusted analyses, young adult drivers (aged 20-24 years) had a higher likelihood of risky driving behaviour than beginner drivers (aged 16-19 years). However, in the adjusted analyses (i.e. the full model and the parsimonious models), age was not found to be an independent predictor for risky driving behaviour.   63  The majority of previous studies have either evaluated crash rates (not necessarily risky driving behaviour) in this age group, or compared all young drivers (16-24 years of age) with older drivers (25 years or older). Few studies compared very young or beginner drivers (16-19) with young adult drivers (20-24 years of age).  The previous studies revealed that risky driving is more prevalent among young drivers (< 25 years) than in older drivers (≥ 25 years). Young drivers are more likely to speed, run red lights, make illegal turns, ride with an intoxicated or a drinking driver, and drive after drinking alcohol (150,201). Tubman-Ben-Ari et al., studied 320 participants aged 18-60 years and showed that the younger the drivers were, the more recklessly and aggressively they would drive (150). Oppenheim et al., studied 527 Israeli drivers aged 20-55 years and found that the driving violation rate decreased as driver age increased (201). A meta-analysis by Zhang et al., also showed that, across studies that included different age ranges, there was a negative correlation between driving violations and age (i.e. as age increases, violation rates decrease) (202). The Organization for Economic Co-operation and Development, in its 2014 and 2017 reports stated that crash and near-crash rates are several times higher in younger drivers (203,204). Simon-Morton et al., studied 279 adolescent drivers and their parents and found that, during their first 18 months of holding a license, adolescent drivers experienced more crashes and near-crashes in comparison to their parents (19). A study, based on data from New Jersey Traffic Safety Outcomes data warehouse, found that older novice drivers were involved in fewer accidents than younger novice drivers in the short term (within 3 months of obtaining their license), but more accidents in the mid-term (within 3 years of obtaining their license). This supports the theory that the higher rate of accidents among younger novice drivers is potentially due to behavioral factors, not driving ability (205). Interestingly, although most risky drivers are young people, they are not new drivers. This suggests that driver inexperience alone does not account for driving or crash risk. Lack of driving experience combined with a penchant for risky driving behaviors, a tendency to underestimate the dangerous consequences of such behaviors, and a tendency to overestimate driving skill, contributes to the high crash rate among young drivers (206).  This creates a discrepancy as age is a significant predictor of crash risk (or risky driving behaviour) in previous studies, whereas the findings of this study showed that age did not play a role in affecting driving behaviour within this young driver population). This may arise from the fact that comparison group participants from previous studies included a broader age spectrum (young, middle, and/or old age) while the current study focused on a narrower age range (16-24 years old). Moreover, the current study included multiple confounders and variables that could refine the results in this age range.   64  A Canadian study of 47,356 drivers aged 16 years and older with driving experience showing that risky driving behaviours (such as cell phone distraction, aggressive behaviour, fatigue while driving and not wearing a seat belt) were more common among 20-24 years old drivers in comparison to those younger than 20 years of age (109).  There was no explanation by the author. The discrepancy can simply be attributed to the fact that, looking at age in isolation, high risk behaviours are more prevalent in the 20-24 years age group than in the16-19 years age group. When considering other factors (current study), age did not play a role in predicting risky behaviour 4.1.2 Gender and driving behaviour In this study, low-risk drivers were mostly non-male participants (57%). Moderate- and high-risk drivers were mostly men (57.7 % and 73.2 % respectively). In the adjusted models, male gender was still a strong, and independent predictor of risky driving behaviour in young drivers. The results of this study are consistent with many previous studies that also found male gender to be strongly associated with risky driving (150,201,209).  Tubman Ben-Ari et al., studied 320 Israeli drivers aged 18-60 years old, and found that male drivers drove more recklessly and angrily (150). Oppenheim et al., evaluated 527 online questionnaires from Israeli drivers of 20-55 years old and found that drivers who identified as male had more driving violations (201). Similar findings were mentioned in a study of 1907 Romanian drivers (210). Another study by Renner et al., compared 98 juvenile traffic offenders with 149 driving licence applicants and found a higher prevalence of male gender in offenders (209). There is evidence of a gender difference when it comes to attitudes towards risky driving. A study by Glendon et al., of 133 high school students aged 16-17 years showed that male drivers had more risky driving attitudes (210). Nabi et al., studied 13,447 French drivers and found that males had more negative attitudes towards enforced driving laws (27). Generally, males tend to engage in high-risk activities more than females (211). Some studies found that young male drivers are more prone to engage in risky driving behaviors such as driving fast and committing more violations (212,213). This has been explained in the literature by biological theories especially by the effect of testosterone (male sex hormone) which has been related to sensation seeking (214). Recently, some studies have explored the influence of the on risk taking behaviours of different adult gender groups. That is social expectations in terms of behaviors depending on the individual gender group (215). Evidence also showed that motivation can even change drivers’ intention towards speeding and overtaking while driving. For example, young men 65  drive over the speed limits more often in the presence of peers to “show-off” (117,216). Peers and society can establish norms and perceived behaviour which influence behaviours directly in young people (217).  Taubman-Ben-Ari et al., in a study measuring the effect of family climate on driving safety in children found that female drivers tend to listen to their parental advice regarding safe driving (85). This may also come as no surprise since male drivers have less risk perception and enjoy risky driving behaviour more than females. In addition, gender norms about masculinity already influential during adolescence do not favour safe driving (212). The combination of attitude, perception of social norms, behavioural control (over-confidence), and emotion can lead to more risky behaviours in males, according to the Theory of Planned Behaviour. This difference can be partly explained by social expectations and norms, which play a significant role in shaping gender stereotypes. According to social role theory, men and women behave as per normative expectations (217). Moreover, risky behaviour in presence of peers could also be described by Normative Social Behavior theory and Evolutionary Psychology (218). The former explains the role of peers’ influence and settled norms in behaviour formation and the latter talks about males learning to perform specific behaviours to attract more attention and appear masculine (see also Background, Section 1.2.4.) Other gender-related characteristics may also explain why males are prone to risky driving. Generally, in this age group, young females drive fewer hours per day, prefer not to drive recklessly or ride with unsafe drivers, prefer not to joyride as much as men do, and are more concerned with not hurting others (15). Moreover, it has been shown that females in the age range of current study, typically mature ahead of male peers which lead to a higher level of differentiation of self (i.e. can express their emotions while being able to control their impulses, able to devise a well-defined plan, and make a planned decision). However, it does not mean that young female drivers are necessarily safe. In addition, more reckless driving in young male drivers can be attributed to a higher sensation seeking (90). Based on evidence, young male drivers have higher perceptual-motor skills and less safety skills in driving compared to young female drivers which results in a higher rate of reckless driving in young males. Stereotypical male machismo attitudes and behaviors that are prevalent across nationalities and cultures are one of the leading causes of higher accident risk in males (215). 66  4.1.3 Employment and driving behaviour  In the current study, while most low-risk drivers were full time students (54.5%), moderate- and high-risk drivers were mostly students and employed at the same time (47.1% and 51.5% respectively). The adjusted and non-adjusted analysis showed that there was a significant correlation between driving behaviour and employment status. Being “both employed and student” was found to be a strong independent predictor of risky driving behaviour among the participants than can be attributed to work-related fatigue and stress, and more inflation of their general over-confidence and self-efficacy in performing risky behaviours including risky driving behaviours. Hughes et al. studied 47,356 drivers aged 16 years and older in Canada and showed that employed drivers were more distracted by their cellphones, drove while fatigued, and drove faster and more aggressively than unemployed drivers. There was no explanation for the potential reason but a suggestion that some intervention in the workplace like stress reduction, expectation modification to avoid work on the road, and a change in drinking culture of post-work meetings with clients or colleagues may mitigate the aforementioned risky driving behaviours. However, in that study, the association between employment status and risky driving behaviour was not statistically significant (109).  Another aspect of employment which could be considered in young adults is being involved in work at a young age. The majority of our young participants (aged 18.34±2.31 years) were both employed and students at the same time. Being involved in work may have other benefits for youth: income, interest in gaining independence, social activity. This situation may reflect the ‘transition proneness’ concept in Problem Behaviour Theory (216,217) which, as mentioned before, explains risky behaviours in youth as a way for them to show that they are grown-ups. When people start working at an early age, they try to mimic adults’ behaviours such as driving behaviour and they are influenced by their work environment (218). As noted above, it should be borne in mind that being a student and employee at the same time may cause driver fatigue due to lack of sleep, prolonged mental activity, and long periods of stress and anxiety as well as more driving exposure, (i.e. work or school trips), which may also influence risky driving behaviour (109).   4.1.4 Living status and driving behaviour  The participants in this study were mostly living with both their parents.  There was a significant correlation between living status and driving behaviour indicating that high risk drivers tended to live alone or with a roommate, which was expected as these risky drivers were also older. However, after 67  adjustment for other factors in regression analysis, this variable had no independent effect on risky driving behaviour.  It is noteworthy that the living status variable in the current study emphasized family structure. In other words, in the current study the “living status” variable indirectly evaluated the influence of family structure on youth behaviour. Living status or family structure is among the least examined sociodemographic factors in the literature in terms of its relation to driving behaviours. Aligning with our finding, a longitudinal study by Reeder et al., investigated the impact of psychological and social factors, including living status, on driving behaviours. They followed a New Zealand cohort every two years from birth until the age of 15, then at the ages 18 and 21 years. The analysis of 840 participants revealed that youth living in a single parent family at age 15 did not differ significantly with respect to traffic convictions compared to those who were living with both parents (219). However, in another longitudinal study following participants from age 13 to 17 years (first year of driving performance), Shope et al., assessed sociodemographic predictors of driving behaviour in 865 American adolescents and found that living situation “other than living with both parents” (e.g., single parents, relatives, etc.) was a predictor of crashes and traffic offences in young males (220).  The discrepancy may be attributed to the fact that Reeder, like the current study, examined several factors such as health-risk behaviour, family and background, social competence, and other psychological and social measures and adjusted their findings, while Shope et al., did not.  Family structure has the potential to influence driving behaviour, as it can also positively or negatively affect the quality of the relationships between parents and children due to supporting, monitoring, and role modelling functions.  According to psychosocial theories (e.g., Normative Social Behavior Theory and Social Cognitive Theory, Problem Behaviour theory), young people are directly influenced by their parents, and the quality of interaction is key. The parent-child interaction can positively or negatively affect young people’s driving through parental monitoring, reward/punishment feedback from parents, and norm acquisition from observing parents. Most previous research has supported the idea that increased parental support and control are strongly associated with lower levels of adolescent risk behavior (218). Tubman-Ben-Ari et al., reported that young drivers who had less risky driver parents were less likely to be risky drivers (150). Schmidt et al., studied 432 participants aged 17-22 years from Ontario, Canada and found that youth were influenced by their parents’ aggressive driving, distraction, and speeding (221). Scott-Parker et al., studied 165 drivers aging 17-24 years old and found that parents’ norm and anticipated parental 68  rewards for avoiding risky driving were positively associated with risky driving in young drivers whereas anticipated parent punishment was negatively associated with risky driving (81). A study by Brookland et al., on 3,992 newly licenced drivers aged 15-17 years old also showed that, adolescents who experienced ‘Active delay’ in getting driving license by their parents had lower crash rates (87). Carpentier et al., describes how parental monitoring can exert positive influence on their children’s driving behaviour (80). There is a huge body of literature on the effects of parents’ role modeling and monitoring on the driving behaviour (like other aspects of life) of their youths whereby they can exert a positive influence on the driving behaviour of their children (80-82). Although the interaction is obviously a by-product of family structure (i.e., living status), the effect of family structure on child’s behaviours may be mediated by the quality of parent-child relationship. In other words, the quality of interaction as interpersonal behaviour is the factor that really matters as it exerts a direct influence on young people. Examining the effect of parental role modelling or monitoring, is beyond the scope of this study. It is noteworthy that the peer influence may prevail over parent effect in this age group. Further investigation of young drivers in Canada with a more robust design is warranted to study the role of different living structures, quality of parent-child interaction, and influence of mother and/or father on youth driving behaviour. 4.1.5 Education and driving behaviour  The current study found that higher education is positively associated with risky driving. In this study, most low- and moderate-risk drivers had less than a high school diploma which may be due to their average age. In contrast, most high-risk drivers had some post-secondary education. The regression analysis showed that having less than a high school diploma education had an independent protective effect on risky driving behaviour.  Consistent with results, Hughes et al., found that young drivers (age 16 and above) in Canada with some post-secondary education drove more often while fatigued, drove faster, drove more aggressively, and were more distracted with cellphones while driving compared to young drivers with less than post-secondary education (109). Atombo et al., reported mixed findings: for some risky driving behaviours (such as speeding), higher education had a protective effect, while for others (like dangerous overtaking), higher education increased the tendency to take more risk (222). A study by the Center for Mental Health in Schools at UCLA highlighted the role of schools in adolescents’ risk-taking behaviours (such as sexual health and risky driving). This study found that there was a 69  reduction in risk taking behaviours in students when schools punished risky behaviour and/or offered programs and services designed to respond to risk-taking behaviours (223). Therefore, the risky behaviour monitoring programs (e.g., educate and/or focus more on punishment in reaction to risk-taking behaviours) at schools may be protective and be a reason that those with less than high school education tend less likely to be involved in risky driving behaviours. High school students with less self-confidence to attempt risky driving due to less experience and less knowledge compared to post-secondary students can be another plausible reason why individuals with higher education are more liable to take risks (224). According to the Theory of Planned Behaviour, less confidence (less behavioural control) along with a high perception of the consequences of risky driving (leading to a negative attitude toward risky driving) may prompt youth to “follow the rules” and be less involved in risky driving behaviour. In addition to driver age, it is not surprising that education has an independent association with risky driving behaviour given the fact that education plays a fundamental role in human, social and economic development of youth. The extent to which education is linked to health-related behaviours is subject to ongoing debate (223).  4.1.6 Driving License class (with or without restriction) and driving behaviour  The findings demonstrated that most participants in the current sample (71.5%) and also in each driver group had a restricted driving license. Moreover, there was a much higher proportion of unrestricted drivers falling under the high risk group (37%) compared to low risk drivers (9%); however, after adjusting for other factors, there was no significant, independent effect of driving licence type (with or without restriction) on risky driving behaviour.  Several recent studies assessed the importance of graduated driving licensing (GDL) programs and found that GDL programs decrease crash and violation risks in young novice drivers (227-229). A study in Nova Scotia, Canada, showed both a beneficial short term (first year) impact (i.e., decreasing crash rates in drivers aged 16 – 18 years) after starting a GDL system in the 1990s as well as a positive long-term effect (accident reduction) almost a decade after adopting GDL programs in all provinces and territories in Canada (230). The British Columbia Injury Research and Prevention Unit reported that BC’s GDL program reduced the crash rate in young drivers by targeting some risky driving behaviour like cell phones, drug/alcohol impaired driving, and driving with multiple peer-aged passengers (228). Ehsani et al., studied the crash rate difference in young drivers (aged 16-18 years) who drove independently from the beginning versus those who experienced GDL before being 70  permitted to drive independently at the same age. Despite driving independently for a longer period of time, the result showed more risky driving pattern in the former group (229). Needless to say, GDL imposes certain restriction and limitation on novice drivers making them behave as less risky drivers due to fear of losing their license which, in turn, leads to reduced crash rate in young drivers. GDL, like other protocols and programs, have to be dynamic and need to be revised and updated to match the most recent societal needs in terms of new restrictions and/or method of curbing law breaking habits (231,232). The plausible reasons for the findings in the current study (no significant association between license status and risky driving behaviour) could be due to the fact that the drivers with unrestricted licenses in the current study had also all gone through GDL programs. Therefore, it is unclear how risky their driving would have been if they had not completed these programs. Another reason could be the distribution rate of unrestricted and restricted license in the total sample and the driver groups which may warrant a more specific study to investigate this topic.  4.1.7 Duration of driving independently and driving behaviour  In the current study, the duration of driving independently was significantly correlated with risky driving behaviour. Those who drove independently more than three years were more likely to fall under the high-risk driver group. After adjusting for other factors, the association of driving independently for more than three years remained positively correlated with a higher tendency of high-risk driving behaviour. Moreover, driving independently for less than a year, had a protective effect on risky driving behaviour. A survey-based study on 511 Chinese drivers with different age groups demonstrated that driving experience (assessed by the duration of holding a driving license and the amount of daily driving hours) directly predicted accident risk, i.e. the more experienced drivers were more likely to be involved in traffic crashes and receive tickets (233). These results may reflect on the point that in young drivers, driving independently for a longer period of time may negatively affect their driving behaviours in that the more they drive independently, the more they may drive recklessly.  It is plausible that the more experienced drivers may perceive some risky driving behaviours, such as speeding, less hazardous. In other words, the more experience a young driver has with driving, especially when nothing bad happens, the more likely they are to reduce their assessment of the risk involved leading to less risk perception (233). Normative Social Behavior theory also posits the 71  influence of settled norms by family and friends on a person’s behaviours. Transitioning from driving with restrictions to driving independently occurs simultaneously with a reduced parental influence and an increased influence from peers. Accordingly, there tends to be more incidents of risky driving behaviour when driving independently (234). Moreover, the more a young driver drives independently, the riskier the driving behaviours become. The perception is that they have more experience and confidence in managing risky behaviours and managing possible hazardous outcomes (233). According to the Theory of Planned Behaviour (235), the weight of the subjective norm (increased influence of peers coupled with decreased parental presence) and an increase in behavioural control may over time account for why young drivers who drive for more than three years tend to be more risky drivers. Of note, when young drivers become adult drivers, the elements of Theory of Planned Behaviour may be weighted differently.  This finding may be important in guiding prevention programs to address youth driving independently for more than three years. However, it does not mean that the less experienced drivers should be neglected.  4.1.8 Driving exposure and driving behaviour The majority of drivers in the low-, moderate- and high-risk groups ranked low in their mileage driven exposure. This might be attributed to the timing of the survey during the school year, i.e., they would have had less time to drive than during the summer. However, high exposure, defined as driving more than 20 hours per week, was more prevalent among high risk drivers compared to low risk drivers. Driving exposure (moderate and high) was found to be a strong predictor for risky driving behaviour. Scott-Parker et al., in a study of 476 drivers (aged 17-25 years) reported that more driving experience was significantly associated with more risky driving (175).  Evidence in different age ranges also showed congruent results. Mekonnen et al., studied driving exposure (driving hours per week and kilometers per year) and driving behaviour in 361 professional drivers who had a driving professional license (mean age 34 ± 7.97 years) and concluded that drivers with more driving exposure (20,000 km versus 10,000 km per year) were twice as likely to engage in risky driving behaviours. The bivariate analysis of this study also showed that those who drove 10-20 hours per week were more likely to perform risky driving behaviours compared to those with fewer hours (236). Oppenheim et al., surveyed 527 drivers (mean age 29.2±7.5 years) about traffic violence and predictive factors (including driving exposure) and found that greater driving exposure (annual 72  kilometers driven) was associated with higher rates of traffic violations (201). Tao et al., in a study on 511 Chinese drivers (mean age 34.2±8.8) showed that more experienced drivers (years of driving and driving hours per day) were involved in more traffic accidents and received more tickets. However, after adjusting for other factors, there was no significant effect of driving exposure on risky driving behaviours (234). The literature suggests that driving exposure is positively associated with risky driving in all age groups. Consistent with this literature, the current study provides evidence of an association between driving exposure and risky driving behaviours among young drivers. This association can be explained by the Theory of Planned Behaviour. The more driving experience youth acquire, the more confidence they feel in controlling the outcome (increased behavioural control). Further, the more often youth have their peers (most influential reference group at this age) in their cars while driving (subjective norms), the less they perceive the hazardous outcome of risky behaviours (positive attitude towards risky driving behaviours). Therefore, more driving exposure may prompt young drivers to perform more risky driving behaviour. It is important to consider that this increase in confidence and propensity to take risks while driving may persist into adulthood. As such, it may be beneficial to continuously monitor risky driving behaviours for all age groups through self-reporting and objective measures.  This could potentially lead to implementation of promotion programs (such as speeding, aggressive driving, distracted driving) for experienced drivers in order to improve overall road safety. 4.1.9 Driver car status and driving behaviour In the current study, most high-risk drivers (76.5%) owned their cars while most low-risk drivers (80.5%) did not. Further analysis showed that car ownership was an independent risk factor for risky driving behaviours. The finding is consistent with the results of a study by Brookland et al., that surveyed 3992 newly licenced drivers (aged 15-17 years) and found that drivers who owned their cars tended to be riskier drivers and experience more crashes than those who drove their parents’ vehicles (87). Fewer legal restrictions, less parental control, and peer influence make drivers who own their cars more prone to risky driving behaviours. Moreover, car ownership may give young drivers a feeling of independence and increases confidence (rooted in the nature of ownership) which, in turn, leads to imitating adult problem behaviours in driving (237).  However, a study conducted by Mousavi Bazzaz et al., on 514 drivers aged 18-85 years old showed that more expensive cars (i.e. > $30,000 USD) were associated with lower accident rates, perhaps because drivers who own these vehicles 73  drive more cautiously in order to avoid car damage and financial loss (238). Parents should seriously consider that the result of granting vehicle ownership privileges to their children could be a higher level of risky driving behaviour. Parents could adopt   some preventive measures (e.g. provide more education and discussion about the outcome of risky behaviours) in advance, before providing access to the family vehicle. For the young drivers who own their vehicle,  parents could also provide more parental guidance and actively monitor attitudes to prevent risky driving behaviors from developing.  4.1.10 Socioeconomic status and driving behaviour Most participants in this study, across all driving behavior groups, scored high for socioeconomic status and the current study did not find a significant relationship between socioeconomic status and driving behavior even after adjusting for other factors.  Hansen et al., surveyed 360 students (aged 12-16 years) and found that a lower socioeconomic status (including factors such as family income and parents’ employment status), was linked to risky behaviours (like criminal behaviours) in youth, regardless of their gender (105). In another study, Hughes et al., used a national survey database to study 47,356 drivers (aged above 16 years) who had driven a motor vehicle in the previous 12 months and found that high household income was associated with aggressive driving (111). Another study, by Hoskin et al., showed that drivers in all age groups with high and medium socioeconomic status reported more risky driving behaviour, intention to speed, and positive attitudes toward speeding than drivers of low socioeconomic status (239). Machado-Leon et al., studied 492 Spanish drivers from all range groups (above 18 years) and found that the higher income per household that drivers had, the more likely they did not respect speed limits, safe distancing, and passing rules (236). The attitudes of high socioeconomic drivers toward risky driving might be a reflection of their perception of their wealth, a sense of entitlement that comes from being part of a privileged group.  A qualitative study interviewing people from a variety of professions in Iran showed that individuals with higher income may buy sophisticated high-powered vehicles as a symbol of prestige and this could lead to more positive attitudes toward speeding, risky driving and related behaviors (240,241). The present study found no association between socioeconomic status and risky driving behaviour. This discrepancy could be due to the difference in driver age groups (young novice drivers compared to drivers in all age groups). Another plausible reason may be due to the limitations of the socioeconomic status scale used in this study, which was designed to avoid the need for parental consent. Another possible reason (if the accuracy and validity 74  of the scale is relevant) may be that most participants in the current study had high socioeconomic status and there was insufficient range in socioeconomic status to show a difference. Alternatively, the findings from the current study may indicate socioeconomic status may not play an independent role in predicting risky driving behaviour in this age group (after adjusting for other factors such as car ownership).  4.2 Personality traits and driving behaviour The second objective of this study explored the relationship between personality traits and risky driving behavior in young drivers in Canada. The level of risky driving behaviour varied across the big five personality traits. Overall, the findings indicated that different personality traits can influence driving behaviour. Young people with high extroversion, high neuroticism, and low imagination tended to perform more risky driving behaviours. Each of the five personality traits dimension will be discussed below. 4.2.1 Extroversion and driving behaviour  Extroverted people are characterized by being excitable, sociable, talkative, and assertive, with high amounts of emotional expressiveness. These outgoing people gain energy in social situations (140-143). Most participants in all driving behaviour groups (low to high risk drivers) scored average on extroversion.  However, extroversion was significantly associated with driving behaviour. After adjusting for other factors in the proportional regression analysis, highly extroverted drivers were 1.6 times more likely to perform risky driving behaviour compared to drivers with average scores for extroversion.  Previous studies on risky driving also found that drivers who scored high on extroversion (extroverts) reported higher levels of risky driving behaviour than drivers who scored low on extroversion (introverts) (242-245). Stephens et al., surveyed 101 British drivers (17-78 years of age) and found that extroverted drivers were more likely to report high rates of traffic violations (intentional deviation from traffic rules) and accident rates. Of note, violations predominated in younger drivers aged 17-25 years, while errors and lapses of judgement showed no significant difference between age groups.  (242).  Wang et al., found that extroversion was associated with maladaptive driving behaviour, a term used for an anxious, unsafe, reckless and angry driving style, based on a study of 296 Chinese drivers, aged 20-56. In Wang study, drivers who reported more 75  violations were also more extroverted. (243). Renner et al., compared 95 young traffic offenders (in a driver improvement training program) with 78 young “control group” drivers (both groups aged 18-30 years) and found a significant relation between extroversion and risky behaviour in young male offenders compared to male controls (244). Anitei et al., interviewed 100 18-25 year old undergraduate students and found that the enjoyment of violence (a construct in the driving behaviour questionnaire assessing the intentional and deliberate aggressive behaviour to cause harm to others, such as sudden braking being tailgated) was predicted by high extroversion and low agreeableness (245).  Not all studies report a relationship between extroversion and risky driving. Tao et al., interviewed 511 Chinese participants (mean age 35 years) and found no relationship between risky driving behaviour and extroversion (233). Moreover, Shen et al., in a study that aimed to validate the Positive Driving Behaviour Scale on 421 Chinese drivers (20-60 years of age) found that extroverted personality was associated with positive driving behaviour such as not blocking or obstructing other drivers, as they would “avoid using left lane to facilitate the speed of traffic flow”. However, it should be mentioned that Shen did not specifically evaluate the association of extroversion with risky driving behaviour (246). Recently, Akbari et al., conducted a meta-analysis that included 22 studies with a total of 11211 participants to assess the correlation of personality characteristics (five personality traits, sensation seeking, and driving anger) with risky driving behaviour. They concluded that risky driving behaviour was not correlated with extroversion (247). The inconsistent results in literature may be explained by differences in methodology, different age groups, use of different tools to measure personality, different definitions of risky driving, and failure to adjust for other variables.  It is unclear why extroversion is associated with risky driving behavior. Assertiveness, one of the main characteristics of the extroversion trait, involves social ascendancy and forcefulness of expression so may be of importance. Assertive people tend to take charge, seek to influence others, take control and try to lead others (248). It may be that such dispositions contribute to lower the extroverted driver’s threshold for aggression, at least in traffic situations (249).  Extroversion concerns the quantity and intensity of energy directed outwards into the social world (250). Having a high baseline energy level may increase the risk of becoming easily frustrated, which, in turn, may increase the amount of aggressive behavior. Extroverts also tend to seek self-stimulation more rapidly and intensively than introverts, which may increase aggressive behavior, insofar as aggressive driving is a form of self-stimulation (251). Another possibility is “activity” (i.e., the person’s pace of living), another facet of extroversion. Active people are always busy and “on the go” and may attempt to 76  manage several tasks simultaneously (252). The frustration-aggression theory postulates aggression is often a result of frustration. As such, it is conceivable that extroverted drivers finding themselves in time pressured situations may be more likely to feel that their personal mobility needs are frustrated, making them prone to react with aggression (253-255). In summary, the current study contributes to the literature by showing that extroversion is an independent risk factor for risky driving behaviour in young drivers in Canada. 4.2.2 Neuroticism and driving behaviour  Individuals who score high on neuroticism are more likely to be moody and experience such feelings as anxiety, worry, fear, anger, frustration, envy, jealousy, guilt, depressed mood, and loneliness (138-141). In the current study, the participants mostly scored average in neuroticism personality trait. Although the result of chi square test showed no significant association between neuroticism and risky driving behaviour, the regression model revealed a positive independent effect of neuroticism on driving behaviour. Young drivers with high levels of neuroticism were 1.8 times more likely to perform aberrant driving behaviour.  Most previous research is consistent with the current finding. A self-report survey-based study on 101 British drivers aged 17-78 years found that drivers with higher scores for neuroticism had significantly more lapses in driving judgment, which was attributed to prompt reactions to the external stressors (e.g., anxiety, fatigue) and attenuation of cognitive functions (e.g. reaction time and memory) (242). An anxious driving pattern was significantly correlated with high neuroticism in a study of 100 undergraduate students aged 18-25 years (245). Another study among 312 college students (median age 19 years), found that the loss of concentration while driving and loss of vehicle control was more often reported in neurotic drivers (146). Of 296 Chinese drivers aged 20-56 years, drivers with high neuroticism tended to have a maladaptive driving style defined as an anxious, unsafe, reckless, and angry driving style.  The risky driving behaviour in anxious drivers could be due to a reduced level of confidence in both their own driving ability as well as that of other road users, which reinforces the driver’s stress. Moreover, neurotic drivers have higher thrill-seeking, anger and aggression, and distress and frustration intolerance (243). A recent meta-analysis reported that neuroticism have a significant, positive influence on risky driving behaviour (247).  However, not all research concludes that neurotic drivers are prone to risky driving.  77  Renner et al., compared the personality traits of ninety-eight juvenile traffic offenders (16.8% female) with a control group of 149 applicants for driving licences (43.6% female) and found no relationship between neuroticism and risky driving behaviour (245). Plausible reasons for this disagreement with the findings of the present study and literature may arise from the selection bias that the control group  (recruited from a driving school and who were applicants for driving license) was not an appropriate comparison group, second the female to male distribution difference in the case (16/79) and control (33/44 ) group.  The present findings that neuroticism is positively associated with risky driving behaviour can be explained by the personal characteristics of neurotic individuals. First, individuals who score high on neuroticism are generally prone to psychological distress and are particularly vulnerable to stress. Neuroticism has been identified as a strong predictor of driver stress (256) possibly because it is associated with ineffective coping strategies (257). Reactions to stress include decreased cognitive and performance capacities (258), and highly neurotic individuals may be more stress reactive than others in traffic (259). Second, high scores on neuroticism may be linked to anger. Moodiness and being temperamental are central aspects of neuroticism, and emotionally unstable individuals are more prone to anger than emotionally stable individuals (260), and anger is closely linked to aggressive driving (261). Third, neurotic individuals may frequently experience impatience, tension, nervousness and irritation (262). In a traffic context, impatience and elevated irritation may lower the driver's threshold for aggressive behavior. To deal with drivers’ impatience and irritation and aggressive driving behaviour, injury prevention stakeholders could develop training that promotes emotion regulation in driving setting for drivers who commit relevant offences.  4.2.3 Agreeableness and driving behaviour  The personality trait of agreeableness has characteristics such as altruism, trust, affection, kindness, less “showing-off”, and other prosocial behaviours (138-141). The current study found no significant correlation between agreeableness and driving behaviour. On the contrary, a wealth of literature has found a negative association between agreeableness and risky driving behaviour. Stephens et al., in a study of 101 British drivers aged 17-78 years stated that low levels of driving violations was linked to higher scores for conscientiousness and agreeableness (242).  Wang et al., in a study on 296 Chinese drivers aged 20-56 years, found that drivers with low agreeableness were more likely to be involved in risky driving behaviour (243). Anitei et al., indicated that, among one 78  hundred second year undergraduate psychology students aged 18-25 years, drivers with low agreeableness experienced more “enjoyment of violence” (i.e. aggressive behaviours performed intentionally that cause harm to other road users) in their driving behaviour (245). Tubman-Ben-Ari et al., assessed the correlation between four driving styles (i.e. reckless, anxious, angry, and careful) and the big five personality traits on 320 Israelis aged 18-60 years. They found that agreeableness was inversely correlated with angry and reckless driving styles. Moreover, the higher the agreeableness level, the more likely drivers were to adopt a careful driving style (150). A study in China, by Shen et al., on 421 drivers (20-60 years), showed that the agreeableness personality trait was associated with less risky driving behaviours (246). Dahlen et al., studied 312 college students (median age 19 years) and reported that reduced agreeableness was associated with the loss of vehicle control (146). Akbari’s meta-analysis (2019) of 22 studies investigating the correlation between personality characteristics and driving behaviour found that agreeableness was inversely correlated with risky driving (247). A possible explanation why altruism is inversely associated with risky driving is that drivers scoring higher on altruism are expected to consider the interests of others and show more active concern for others in traffic and thus reduce risky driving behavior (260). The inconsistency between current findings and the majority of literature may be largely due to the narrow range of scores. However, the current finding may draw attention to personality traits other than agreeableness and conscientiousness when it comes to risky driving in young individuals. The incongruence between the current findings and the literature may be attributed to differences in the age range of participants and to the selection bias. Participants in previous studies were recruited from a broader age range, e.g., 18-60 years, than the current study.  In the studies restricted to a smaller age range, like Anitei et al., participants were recruited from a certain population subgroup such as “college students”, or “second year undergraduate psychology students” which limits the generalizability of their results. In contrast, the current study included a less selected population of young drivers (i.e., those who had access to the internet and Facebook). Moreover, most other studies did not adjust for the potential confounding effect of other variables such as sociodemographic factors.  4.2.4 Conscientiousness and driving behaviour The conscientiousness personality trait has high levels of thoughtfulness, with good impulse control and goal-directed behaviours. These people are also organized and mindful of details (138-141). 79  In the present study there was no significant correlation between conscientiousness and driving behaviour.  Consistent with this finding, Dahlen et al. surveyed 312 US college students and also reported that conscientiousness was not significantly associated with risky driving behaviour (146). Similarly, risky driving behaviour was not associated with consciousness in the Akbari et al.’s   meta-analysis (247).  However, as with agreeableness, a great deal of literature showed that more conscientious drivers reported less risky driving behaviours. Shen et al. used a validated questionnaire to study 421 Chinese drivers (20-60 years) and showed that conscientiousness was associated with positive driving behaviours such as not blocking an oncoming car or slowing down to help a driver trying to overtake (246). Stephens et al., in a study of 101 British drivers aged 17-78 years, found that low levels of driving violations (intentional deviation from traffic rules) was associated with higher scores for conscientiousness and agreeableness. Moreover, the more conscientious a driver was, the lower the tendency to experience lapses (attention or memory failure) and errors (unintentional failure of achieving desired outcome through planned behaviour) (242). Wang et al.’s study of 296 Chinese drivers aged 20-56, found that conscientiousness negatively predicted angry and risky driving styles and positively predicted careful driving style (243). Anitei et al. studied the association of personality traits (according to the Big Five model) and risky driving behaviour (using the Aggressive Driving Behaviour assessment tool) among 100 second year undergraduate psychology students aged 18-25 years. They found that low conscientiousness was a predictor of anger in driving behaviour (manifesting angry behaviour in certain traffic situations or towards other road users’ behaviours) (245). Tubman-Ben-Ari et al. examined the association between four driving styles (i.e. reckless, anxious, angry, and careful) and the big five personality traits in 320 Israelis aged 18-60 years. They found that the lower the conscientiousness score drivers had, the more tendency they would have reckless, angry, and anxious driving styles (150). Also, the less likely they were to driver carefully (150).  The dissimilarity between current findings and some studies mentioned above may be largely due to the narrow range of scores.  In addition, given that the role of personality traits may vary by age, focusing on specific age group in the current study may lead to this incongruent finding. The nature of conscientiousness (i.e., impulse control, planning ahead, thoughtfulness) and its effect on behaviour may be subject to changes over time due to the acquisition of experience and skills with age (263).   80  4.2.5 Imagination (Openness) and driving behaviour People with strong imagination are adventurous and creative (138-141), imagination is also associated with sensation-seeking (242). Close to 70% reported average imagination in each risk group.  In the current study, low imagination was associated with an increased likelihood of risky driving.  The odds of risky driving behaviour in young drivers in Canada with a low imagination was found to be 1.54 times higher than those in drivers with an average imagination. This finding is consistent with a series of previous studies that demonstrated that the “openness to experience” personality is negatively correlated with risky driving behavior and positively associated with a more positive driving behavior. Anitei et al., in a study of 100 second year undergraduate students (mean age 20.68) demonstrated that low openness was a predictor of negativism in driving defined as “refusal to drive in a cooperative and understanding manner” (e.g., extra caution while passing a bicycle, promptly stopping for a pedestrian waiting to cross the street) as measured by the Aggressive Driving Scale (245). Using the 35-item Driving Survey, Dahlen et al., in a study of 312 college students with median age of 19 years old found that low openness was significantly associated with risky driving behaviour such as driving without a seatbelt, and unsafe passing (144). Monteiro, et al., interviewed 311 drivers aged 29.8 ± 9.81 years in Brazil and also found that openness was inversely correlated with risky driving (262). Wang et al. surveyed 296 drivers aged 20-56 years (mean age 35) to assess the relationship between personality traits (Big Five Inventory), driving styles (Careful, Anxious, Risky, and Angry or High-velocity), and driving behaviour (Ordinary Violations, Aggressive Violations, Errors, Lapses, and Positive Behavior). After adjusting for other factors, Wang et al., found that openness was correlated positively with a risky driving style and, at the same time, was inversely correlated with angry and high-velocity driving styles. Overall, Wang et al., concluded that openness had an independent positive influence on positive driving behaviours (243). Employing the same tools, Shen et al., in a study on 421 Chinese drivers aged 20-60 years (mean age=40.34) also concluded that openness was associated inversely with some dimensions of risky driving behaviour (Ordinary Violations, Aggressive Violations, Errors, Lapses) and was positively related with positive driving behaviour (246).    Based on literature different studies seems to show inconsistent results – some showing a positive and some a negative relationship between openness with risky driving. A study by Stephens et al., on 101 UK drivers aged 17-78 years, found no association between openness and driving errors, lapses 81  or violations (242). Akbari et al., (discussed in previous sections) showed no significant relationship between openness with risky driving behaviour (247) but in their discussions could not provide any plausible reasons for this finding.   In summary, this study lends more support to the positive influence of openness on positive driving behaviour that may be attributed to some of the characteristics of open people, e.g., focusing on tackling new things, and high creativity and imagination skills. Youth with high openness to experience might opt for other positive outlets like leisure activities and avoid negative or health related risky behaviours (like risky driving) to tame their adventure-seeking motivations. Future research can investigate this association in more depth to examine which aspects of openness are associated with driving behaviour.  Overall, a person’s personality includes several personality traits. Therefore, it is noteworthy to conduct further detailed research projects to evaluate a person’s personality as a whole and its interaction and association with risky driving behaviour.   4.3 Leisure activity and Driving Behaviour The main objective of this study explores the relationship between leisure activity and risky driving behavior among young drivers in Canada. Before listing the influence of leisure activities on driving behaviour, it’s noteworthy to mention that there may be an association between leisure activities and other predictive factors such as sociodemographic factors and personality. For example, in a survey of 1,719 participants (mean age 34.57 years), Ardahan et al., found a relationship exists betweensport and outdoor activities, and education status, age, income and employment status (264). Therefore, the current study investigated the influence of leisure activities while adjusting for important confounders such as sociodemographic factors and personality characteristics.  In this study, most participants had a low level of engagement in leisure activities, which might be due to the time period when the survey was deployed (i.e., March to May – during the school year). Of note, levels of engagement were defined by the number of hours or times a week that were allocated to a leisure activity – with unique cut-off values for each leisure activity as described in the method section. The study found that leisure activities can influence driving behaviour in positive and negative ways. Young people who were highly engaged in leisure activities that involved drug use, spending time with friends, watching movies, and social media use tended to display a high level of risky driving behaviour, while those more engaged in reading and writing, playing video games, 82  and volunteering displayed less risky driving behaviour. The 13 types of leisure activity considered for evaluation are discussed individually below.  4.3.1 Video game and Driving Behaviour With the popularity of high-tech devices (computer, tablet, and smartphone) and ever improving internet access, playing online or offline games has become a popular activity in recent years, especially among young people. People usually play video games for entertainment, excitement, when seeking a challenge, when coping with emotions, and to escape from reality (265).  The video games provide a means to fulfill gamers unsatisfied needs or motivations, virtually (265). Although some studies have demonstrated beneficial effects of playing video games on psychological and physical health (266,267), most research on video games has focused on the negative effects on gamers. It has been suggested that excessive video gaming is associated with reduced sleep time, insomnia, displacement ofhealthy leisure activities (269,270), attention problems, poor academic performance, lower self-esteem, lower satisfaction with daily life (270), anxiety, depressive symptoms, deterioration of interpersonal relationships, family conflicts, along with youth’s risky behaviour, violence or crimes (271,272). The present study showed that playing video games was not a widespread activity among youth in Canada at the time of the survey deployment. Most participants scored low (≤9 hours per week) for video game engagement in all three driver risk groups. This level of engagement is lower than reports from other countries such as United States, China, Germany and Saudi Arabia (274-277). For example, 41.6 percent of respondents in the U.S. admitted to playing video games for more than 20 hours a week, while another 11.4 percent claimed to spend between 12 and 20 hours a week gaming (277).  This might be due to the time of survey deployment in relation to the school year and scheduled examinations in Canada. Another explanation might be that different scales were used to measure video-game engagement in each study. Surprisingly, in the current study, playing video games was inversely associated with risky driving behavior: After adjustment for other factors, participants with high engagement in playing video games (≥9 hours per week) were less likely to be risky drivers than those who played less often (adjusted OR=0.56).  There is some evidence that action video games may improve reaction time and visuomotor-control tasks in real driving situations (with a reduction in crash rate) (97).  Although this could affect driving behaviour by boosting the confidence of control and driving ability, the evidence of positive influence of video games on driving behaviour is per se very limited.  An observational study of 103 83  drivers (mean age = 23.3 years) showed that those who played prosocial video games (e.g., ‘Emergency 4’ game in which an emergency physician has to manage accident scenarios) were less likely to drive recklessly and more likely to care more about other drivers on the road (108). Conversely, Fischer et al., in a study that included three related experiments, examined the correlative and causal effect of playing racing video games and risky driving behaviours (categorized as competitive, stunt/obtrusive, and cautious driving behaviours). First, their interview-based study of 190 German drivers aged 16-45 years (Mean age = 23.43 ± 4.65 years) indicated that playing racing games was positively associated with increased competitive driving and stunt/obtrusive driving, and negatively associated with cautious driving behaviour (277). An experimental study of 83 drivers aged 19-42 years (M=23.86 ± 4.40 years), found that participants who played racing games for 20 minutes before being tested in a driving simulator had a higher level of arousal/excitement and priming accessibility to risk-promoting thoughts than those who played a neutral game like soccer. Lastly, they studied 69 participants (age 19-35; mean age = 23.6 ± 4.14 years) during 15 computer-simulated critical driving situations and found that playing racing games was significantly associated with risk taking behaviour in critical traffic situations in male, but not female, participants (278). Aljied et al., conducted two experiments on Canadian students enrolled in a psychology course at the University of Ottawa. First, on 52 participants (mean age 19.25 ±2.17 years), they found no significant correlation between playing videogames and overall driving performance in simulated driving situations. Second, to assess driving skills (such as road information processing and vehicle handling when lane changing, tailgating, lateral control, and passing other cars) they studied 39 drivers aged 18-24 years and found that playing videogames was associated with risky driving behaviour such as speeding and poor performance in lateral control maneuvering (101).  The current study found that video game playing was not associated with risky driving. Evidence is mounting that playing video games may be one way for people to sharpen a number of cognitive skills (2792281). The present findings could be attributed to the fact that video games especially action games (mostly played among the participants) can act as an outlet for youth to express their anger, aggressiveness, stress, sensation seeking (284,283). Thanks to the current technology, video games simulate real-world driving closely enough that they can prompt gamers to interpret and analyze real world traffic situations more quickly. Playing video games obliges young drivers to react suddenly and take risks. Despite the fact that video game driving can translates the real-world driving skills, a dilemma still remains. On the one hand, the resulting improved technical skills might cause young 84  drivers (who now believe that they drive better than average drivers) to be more willing to take risks due to due to their excessive confidence and overestimation of driving self-efficacy. On the other hand, the learned skills can help a driver respond with appropriate reactions, even if risky, when presented with real world unexpected hazardous events. It is worth mentioning that the current study also investigated all types of video games (not just racing games). In analyzing the distribution of driver groups among the various video games types, the high risk driver numbers seemed to significantly trend up in race, action, simulation, and sport types while conversely trend down in adventure and strategy types. The results suggest that more detailed or stratified analyses are warranted for the association between risky driving behaviour and the various types of videogames. 4.3.2 Sport and Driving Behaviour  In the current study, the majority of participants in the three risky driving groups scored low for sports engagement and no association was found between driving behaviour and sport engagement. Although there was no association between sport and driving behaviour in the current study, it is well known that physical activity in childhood and adolescence lead to healthy lifestyles and cognitive and executive function improvements. Interestingly, physical activity also has a positive effect on perceptive, cognitive, and physical abilities that are considered important for driving performance and safety. For instance, aerobic fitness is associated with improved cognitive function (especially executive control, i.e., ignore conflicting information and focus on the desired task) (284), and people with good physical fitness show greater efficiency in information processing (285,286), enhanced attention capacity in dual-task situations (287), and better performance on tasks that require visual-spatial processing (288).  A great deal of evidence has investigated the association of physical activity/sport engagement with other health-related risky behaviours such as alcohol consumption, smoking, drug use, and risky sexual behaviour. The findings are mixed. In a review of 108 studies, Sallis et al., (289) found no clear associations between alcohol consumption and physical activity. Some studies have reported that the least physically active adolescents are at risk of drinking more alcohol (290) or using more cannabis (291). Conversely, other studies have found that physically active adolescents drink more alcohol (292) and use “snuff” more commonly (293) than less physically active adolescents. Few studies examined the effect of physical activity engagement on driving. Wayne et al., conducted an experimental study to understand the impact of organized sports on the driving skills 85  of 100 novice drivers (age 15-29 years) under real-world driving conditions in the Westside of Los Angeles. The main outcome was that practicing organized sports of any kind, solo or team sports, either in the past or present, was associated with enhanced driving skills (282). Another experimental study on 24 drivers (50% male; mean age 20.4±1.93 years; mean driving experience 4.7 years) evaluated whether the spatiotemporal skills acquired through high level sports can transfer to driving skills. Using an emergency-braking test, they showed that athletes (i.e., those participating in their university athletic team) achieved significantly longer collision time lags (lapse of time between when the lead car braked and the time-to-collision with the lead car) in simulated driving situations. This advantage of athletes was attributed to both their body movements and reaction times as well as their ability to respond with a desirable performance in a given situation. (294). Matos et al., conducted an experimental study of 32 novice drivers (mean age 19.1±1.9 years) divided into “team sports players” (at least three years of sports practice) and “non-team sports players” to study the association of physical activity with the capacity of novice drivers to detect peripheral lights on the left or right of the front panel. Using this Peripheral Detection Task during on the road driving sessions, they found that team sports players scored higher than non-team sports players and concluded that people who play sports have better peripheral vision than those not engaged in sports (295). Thus far, it seems the existing evidence has resulted from examining the effect of sport on driving skills but not driving behaviour.  Enhancing and improving driving skills has the potential to positively or negatively affect driving behaviour. In a driving setting, these skills can either elicit over confidence by tempting the driver to multitask (using cellphone) and take risks (speeding and driving aggressively) or help young drivers to plan ahead and process road information more efficiently in order to avoid taking risks. The current finding, indicating no association between sports engagement and driving behaviour, is novel. 4.3.3 Social media and Driving Behaviour Social media sites such as Facebook, Twitter, and YouTube offer opportunities for users to interact and share information with friends and family and with people who have similar interests. Over recent years the number of people using such sites has increased dramatically. Young adults (ages 18 to 29 years) are the group most likely to use social media – fully 90% do (296).  Along with the benefits such as improved socialization, communication and enhanced learning opportunities, the use of social media can introduce risks (297). 86  In our study, social media was the only activity among all leisure activities with a high level of engagement in all three risky driving groups. Given that recruitment for the study occurred through social media, the high engagement rates may be due to selection bias. Participants reported that they spend at least 60 minutes a day on social media. Confirming the stated hypothesis, the current study found a significant association between risky driving behaviour and social media engagement. After adjustment for other predictive factors, the current study found that participants who are highly engaged on social media (more than 60 minutes daily) were more likely to report risky driving behaviour (OR = 1.8). Existing evidence is lacking in studying the direct association of social media with risky driving behaviours. A great deal of literature is limited to texting and driving. However, recently, the cellphone has become the prevailing platform for entertainment and information. Social media apps which are accessible on cell phones are considered to be an important cause of distraction while driving (298). A substantial amount of evidence (298-303) has accumulate from studies investigating distracted driving by cellphone (especially texting but not specifically social media) which is beyond the scope of this section. A study on 935 college students (aged 18-62 years, 74% female) investigated Social Learning theoretical factors and cell phone use behind the wheel. They showed that, while driving, 25% and 39% of participants ‘sometimes’ used their cellphone to check their social media (Twitter, Instagram, Snapchat) mostly while moving and stopped, respectively despite the fact that almost all (98%) of respondents agreed or strongly agreed that using a cellphone is risky while driving (304).  Social media presents youth with different behaviours and role-models, even when they are not directly exposed to an environment (305). Exposure to risk content in media affects adolescents' behavior (306) but does not affect all youth in the same way. Many adolescents who use social media experience identity formation that may impact the way they perceive and behave (309).  Theories, such as social cognitive theory, evolutionary psychology, and identity theory, clarify how social media can change youth’s behaviour. Social cognitive theory explains how children adjust their behaviours to match that of others through watching and imitating their behaviours (308).  Evolutionary psychology shows how a behaviour presented and introduced by media can attract a huge audience when found attractive by youth and even when it is a risky behaviour, youth want to emulate the behaviour in order to attract more attention (309). Identity theory explains that a person’s self- identity can determine the way they express themselves through their behaviours and the social group(s) they belong to. As such, social media can dictate a specific type of behaviour to youth as the 87  characteristic behaviour of special social groups they identify with, e.g., road racers (310). Media frequently presents models of risk taking and aggressive behaviours.  Social Learning Theory and Cultivation Theory both suggests that the more an adolescent or emerging adult observes people engaging in risky driving behaviors in the media, the more likely they are to be influenced by and emulate that behaviour since they identify with and want to belong to the group portrayed in the media (311). “The primary proposition of cultivation theory states that the more time people spend 'living' in the television world, the more likely they are to believe social reality aligns with reality portrayed on television” (312).  A growing body of research has shown that risk-glorifying portrayals in the media have a significant impact on risk taking behaviour (272,298). A study by Havernearam et al., showed that passive media content and passive media exposure (such as watching a social media commercial) have an impact on risky and aggressive driving and normalize violation of traffic laws. Drivers who view positive media messages report lower levels of aggression and risk-taking behaviour. In contrast, the tendency to violate norms increases when being presented with negative media content (312). Therefore, understanding the influence of mass media content upon driver’s behaviour is essential when developing safe driving strategies.  4.3.4 Watching Movie engagement and Driving Behaviour The majority of participants in the three driving risk groups scored low on their watching movie engagement levels. The current study found an association between driving behaviour and movie engagement in that participants with a high level of movie high engagement had an increased likelihood of risky driving (OR = 1.5).  There is limited evidence in the literature on the association of movie engagement with risky driving behaviour. Most evidence concerns the influence of the type of movie on behaviour. Chrysalis et al examined the impact of watching movies that depicted dangerous driving on the perception of driving risk and driving behaviors by assessing 1356 college students (aged 17 to 25 years). They found that watching movies depicting dangerous driving resulted in more positive attitudes towards risky driving behaviors (104). Another study on 150 drivers (aged 19 to 26 years) showed a strong correlation between risk driving and media exposure (including movies). The drivers reported a decrease in risky and aggressive driving after viewing positive media content about driving while tending to drive more aggressively and engage in risky driving when presented with a negative media message content (314). There are several reasons why watching movies can influence youth driving 88  behaviour. The Werther effect states that negative information can lead to increased expression of maladaptive behavior (304). For example, a rise in suicide rate ensues from an increased frequency of exposure to media presenting a suicidal attempt or complete suicide, within a week after the exposure (315). Moreover, according to the General Aggression Model, media exposure has an impact on cognition, emotion, and behavior because of its priming effect (whereby exposure to certain influences subconsciously stimuli a response to a subsequent stimulus), and its capability to provide role models for society, especially youth (304). Normative Social Behaviour theory, Social Cognitive Theory, Evolutionary Psychology, and Identity Theory can all help explain the findings of the present study. Normative Social Behaviour theory posits that youth develop behavioural norms according to the behaviour of people around them (or from media role models) and this, in turn, influences their future behaviour (40). Based on Social Cognitive Theory, youth watch and render others’ behaviours and adjust their own behaviours accordingly. Media can present youth with role models for driving behaviour. Movies often portray men as more attractive for women based on their behaviours, including driving. Adolescents who identify with a movie characters may be particularly prone to engage in the same risky behaviour as portrayed by that character. This role modeling can influence a youths’ subsequent behaviours according to evolutionary psychology (316). Finally, according to identity theory, each person gives him/herself a personal identity and also belongs to certain groups. Consequently, people try to behave in a way that expresses the group he/she belongs to and the identity he/she gives him/herself. Movies can play an important role in demonstrating how a specific identity or group is expected to behave (43). It remains important to undertake causal research to evaluate the effect of types of movies and duration of watching movies, and the corresponding effect size on risky driving behaviour.  4.3.5 Family-involved activity and Driving Behaviour  In the current study family engagement was inversely associated with risky driving behaviour. Those who spend less time with families tended to show more risky driving behaviour. However, after adjusting for other factors, family engagement did not have a significant independent influence on driving behaviour. One contextual factor that may have great influence on adolescents is family. The parent-adolescent relationship is particularly important in behavioural science. Parental support of basic psychological needs is essential for positive psychosocial adaptation (i.e., maintaining oneself in harmony with the changing psychosocial circumstances that leads to a sense of well-being) 89  in children and adolescents (317). Family activities afford an opportunity for parents to connect with their children emotionally, to monitor their social and academic activity and to convey values and expectations (318).  Research shows that young drivers, who have strong parental role models with open lines of communication that provide positive feedback about safe driving and convey specific and clear messages/limits about traffic safety, report less risk taking, less aggressive driving and more commitment to safe driving (47). In addition to parental driving style, considerable research has revealed the relationship between parent-child conflict and negative consequences for adolescents. For example, it has been found that conflicts between parents and adolescents are related to antisocial behaviors (319). The parent-adolescent relationship has an important impact on adolescent risk-taking behaviors. The more conflict between parents and adolescents, the riskier taking behaviors adolescents will take in the future (320). Similar to other health-related behaviors, adolescent driving is strongly influenced by parenting styles and models of behavior (90,321).  Wilson et.al have shown in a study in British Columbia, Canada, that parents have substantial influence on their children’s driving behaviour. They found that parental driving records from the four years prior to their child’s licensure can predict child’s driving behaviour in the first three years after licensure (88). Miller et al., studied 130 drivers (age 17-24 years) first in driving class and again after a year of driving. They found that children’s driving behaviour is more influenced by their mother’s speeding offences than by their father’s speeding. In contrast, children (especially sons) are more influenced by their father’s angry driving and daughters are more influenced by anxious driving patterns of both parents (90). Ferguson et al., compared driving records of young drivers (18-21 years) with that of their parents. They found that when both parents have collisions, the effect is synergistic and multiplicative on children’s risky driving, explained by the fact that these children have no protective influence from either of their parents (47). Moreover, Jafarpour et al. also mentioned in his narrative study that youth become less risky drivers in the presence of their parents, meaning that the presence of parents can influence youth’s driving behaviour (15).  Family activities may give youth the opportunity to observe how their parents drive (role-modelling effect) or to drive while in the presence of their parents (controlling effect). Parental role-modeling can be described by Normative Social Behaviour Theory and by Social Cognitive Theory. According to Normative Social Behaviour Theory, parents influence their children by establishing norms, especially through their own behaviours (See background, section 1.2.4). Social Cognitive 90  Theory explains that behaviour in adolescents and youths is formed through watching how others (e.g., parents) act and adjusting their own behaviour accordingly (See background, section 1.3.2).  The discrepancy in the findings of the current study may be due to the fact that the survey did not enquiring about detailed information like time spent in a vehicle with parents. Another possible reason that the current study did not find an independent association between family activities and risky driving could be because it is a cross sectional study reflecting a single point in time, whereas parental role-modeling begins in childhood. Moreover, the parental controlling and monitoring influence may be restricted to the time when parents are present, because peers are the prevailing reference group for adolescents and young people. The current finding also supports the negative influence of friends on driving behaviour (refer to 4.3.6). In other words, since the principal reference group for adolescents and young adults is friends and peers (rather than parents and family), those groups may be expected to exert more influence on driving behaviour in a cross-sectional study of youth. Despite the null finding in this study, it is still important to note the importance of continuous parental role modelling and influence on children’s driving behaviours, beginning in childhood (even though parental influence may be reduced in adolescents and young adults). For example, Loubean et al., studied 31 seventh and eighth grade students and demonstrated the importance of parental influence on wearing bicycle helmets (86).  4.3.6 Friend-involved activity and Driving Behaviour  The question of whether and how peers influence an individual’s behaviour has been widely investigated. A large body of evidence suggests that individuals who are physically or socially close to a subject influence his/her behaviour and choices (322-325). In particular, the peers and friends with whom adolescents spend time have a significant role in shaping their risky behaviours (326-329). Peer effects have often been mentioned as a leading explanation for why people engage in risky or unhealthy activities such as smoking (327) and drug/alcohol use (327,328). The current study showed that risky driving behaviour in young drivers in Canada was significantly influenced by their friends. Participants who spent time with their friends more than 4 times a week were more likely to report risky driving behaviour than those who spent less time with friends (OR = 1.9). Consistent with these findings, Horvath et al studied 398 participants (17-24 years old) and found that pressure from peers, either direct or indirect, can influence young drivers’ behaviours such as speeding. Direct pressure, in this study, referred to distractions, disturbances or even arguments among driver and passengers, 91  while indirect pressure described the driver’s perception of his/her passengers’ expectations in terms of driving behaviour (329). McNabb et al., studied the scenario when friends drive to the same destination in separate vehicles with a leader vehicle and some followers. They found that the drivers following engaged in riskier driving behaviours in comparison to their usual driving pattern (e.g., driving faster, ignoring pedestrians, etc.). Of note, the lead vehicle was driving safely and at speed limit and was not exposed to the potentials for risky behaviours. Therefore, the risky driving behaviours in the follower vehicles were not attributed to the social contagion effect (330). An interview-based study (32 interviews) showed that the way friends relate to each other can affect youth driving behaviours. For instance, if friends feel equal and have a sense of responsibility towards each other, their driving becomes less risky. In contrast, when the goal of their relationship is leisure and pleasure, they take more risks when driving (331). Adolescents are highly susceptible to peer influences, where perceived and actual peer behaviors can influence risk behaviors (332). Prior research demonstrates that having friends who engage in risky driving predicts future risky driving for newly licensed adolescent drivers (17) and impaired driving among adolescents (333). Further, according to a study of 71 teen-parent groups (teens were 17-19 years old) by Merrikhpour et al., when teens are “hanging around”, those who were passengers can distract the driver r. They also note that teen drivers tend to believe that his/her peers are more likely to get distracted than himself/herself (49). According to Identity Theory, when children find themselves belonging to a group, they define a specific social identity for themselves which, in turn, influences their behaviour (334). In addition, friends (peers) influence each other, by establishing norms (normative social behaviour theory - section 1.2.4), acting as role models (social cognitive theory - section 1.3.2), and when an adolescent adapts certain behaviour to show their friends that they are “grown-up” (problem behaviour theory - section 1.2.4). Given this theoretical background, the conclusion of the current study, that spending time with friends has a negative influence on driving behaviours, is not surprising. This finding spotlights the role of peers in risky driving behaviour in youth and is worth considering when developing preventive strategies and educational programs.  4.3.7 Alcohol-involved activity and driving behaviour  In the current study, the frequency of alcohol consumption was low (less than 4 times per week) in all groups of participants. There was a significant association between alcohol consumption and risky driving behaviour in that those with a high frequency of alcohol consumption were more 92  involved in risky driving behaviour. However, after adjusting for other factors, the current study found that alcohol had no independent effect on driving. A large body of literature demonstrates that alcohol adversely affects driving behaviour (335-342), although the effect of overall drinking habits on driving behaviour has received less attention.  Many studies show that alcohol significantly impairs driving skills such as braking, steering, distance calculation, overtaking ability, and choosing an appropriate speed for the conditions. Lane position, line crossing, number of crashes, speed deviation and time at maximum speed are also impaired by alcohol (337). Drunk driving increases the risk of serious accidents. Compared to sober drivers, drinking drivers are twice as likely to be involved in traffic crashes, even after consuming relatively small amounts of alcohol (338). The latest findings from the Centers for Disease Control and Prevention (CDC, 2016, USA) indicate that a quarter of young adults involved in fatal MVCs had been drinking alcohol (339). A systematic review of cohort studies that followed up the general population of 15 to 19 years old for at least three years and assessed the outcomes at aged 20 or above revealed that late adolescent drinking was associated with early adulthood physical and mental health, and social consequences like early death among men, mainly due to motor vehicle crash (343). Because of the strong association between alcohol consumption and collision risk, driving after alcohol drinking is forbidden by law in most countries (344). Legal strategies such as making drunk driving a criminal offence and laws that make it illegal to drive with blood alcohol content above a certain limit (legal limit) can be effective deterrents of drinking and driving when they are enforced.  Despite laws against drinking before late adolescence, youth might start drinking earlier due to reference group role modeling, intrinsic factors, or coping with underlying mental health issues. Youth with more frequent alcohol use may be inclined to incur more driving violations (345). Part of a secondary use of data of a longitudinal birth cohort conducted by MacArthur et al.,  was to assess the association between the number of drinks consumed during the last 6 months by 15 year old participants and the prevalence and distribution of multiple risk behaviours, i.e. substance use, injuring self, unsafe sexual behaviour, and vehicle-related risk (including Car passenger risk, being a passenger and  either knowing that the driver had consumed alcohol or the driver did not have a licence to drive without supervision; Scooter risk, riding a motorbike or scooter off road or without a license on a public road; and cycling without a helmet) once the participants reached the age of 16 years, in 5067 young participants.  MacArthur et al., found that greater alcohol consumption at age 15 was significantly associated with higher prevalence of engagement in vehicle related risk, except 93  cycling without a helmet. (346). A large study of young Australians (15 to 21 years) found that regular drinking was consistently associated with a range of other health risk behaviours including injuries under the influence of alcohol, tobacco smoking in higher doses, cannabis use and sexual risk-taking (347). In contrast to previous research, the current study did not find that frequent alcohol use was an independent risk factor for risky driving behaviour. The failure to find association may be due to the narrow range in the frequency of alcohol consumption among the study participants (i.e., lack of power to detect a true difference) or the cut-off values in the survey for low and high engagement was not accurate.  4.3.8 Drug-involved activity and driving behaviour  In terms of drug engagement, 30.2% participants in the current study used recreational drugs (mostly marijuana) at least once a week and up to 5 times a week, which is concerning.  Most participants (91.4%) consumed drugs three times a week or less. The current study found a significant independent effect of drug engagement on risky driving behaviour. Those with high engagement in drug-involved activities were 2.2 times more likely to be risky drivers.  The association between risky driving and the pattern of recreational drug use in daily life has not been as widely studied as the acute effect of drugs on driving performance. In particular, the frequency of risky driving behaviours in youth with different patterns of drug use is poorly studied. The findings of the current study are consistent with a previous study by Bina et al., who evaluated the association between risky driving and lifestyle in 645 Italian adolescents (aged 14-17 years). They found that those who reported risky driving practices were more likely to also report a lifestyle characterized by tobacco smoking and drug involvement (125). Similarly, a nationally representative telephone survey of 900 young drivers (age 14-22 years) from the United States, (found that those who used marijuana in the last 30 days had 36% greater odds of involvement in risky driving behaviours leading to crashes (348).  Given the risks associated with substance use and driving, there is a pressing need to understand young people’s views about these behaviours. Theories of reasoned action and planned behaviour (40) emphasize attitudes and perceived norms as predictors of engagement in health-related behaviour. In line with these frameworks, research has demonstrated that individuals who perceive drug use as more dangerous and as increasing the level of crash risk are less likely to drive after drug use, whereas 94  users who believe their driving ability is not impaired by drug use are more likely to drive frequently after using the drug (349). Drug use among youth has been a major issue for decades and the issue of marijuana use among youth in Canada, and the impact of cannabis on road safety, is particularly important given its recent legalization (350). Discouraging drugged driving, particularly after the marijuana legalization, remains a public health priority especially for young adults. Successful education and translation of this knowledge to young people, families and schools is a key strategy to raise awareness, which may result in decreased vehicle crashes among younger drivers. Public education pertaining to the use of cannabis and driving must be based upon current research in order for it to be effective and relevant. The current findings, along with the other evidence, lend support to concerns of a significant adverse influence of drug related activities on risky driving behaviours and of the social norms that may send falsely positive messages regarding the safety of driving after consuming marijuana.  4.3.9 Art and driving behaviour  In this study, the majority of drivers from all risk groups scored low for art engagement as a leisure activity, although low risk drivers were more engaged in art than high risk drivers (20.5% and 12% respectively). After adjusting for other factors there was no independent effect of art on driving behaviour. No study on the relation between art engagement and risky driving behaviour was found to date. However, there are studies on the association of art engagement and alcohol consumption or other risky behaviours. Elpus et al., found that studying art can decrease risk of suspension from school, decrease alcohol and marijuana use, and make children more optimistic (351). Moreover, McCamy et al. (2014) reported on a preventative programteaching art to children and teens (9-16 years old) that was associated with decreased risk taking behaviours (such as biking without helmet) (352). Another study suggested that pro-social art decreased risky and antisocial behaviour in at-risk children (students with risk factors such as a low socioeconomic status, living with a dysfunctional family, experiencing language barriers, living in a disadvantaged neighborhood or struggling in school) (353).  These protective effects of art may be because young people who engage in art activities may be better able to control their violence and impulsivity. For instance, the Ministry of Children, Community and Social Services of Ontario, Canada states that participating in art can decrease aggressive behaviours and violence (354).  95  The current findings are novel as no previous study examined the effect of engagement in art activities on driving behaviour. However, this finding would not obviate a more robust research to evaluate specifically the effect of art activities on risky driving behaviour.    4.3.10 Reading/Writing and driving behaviour The majority of participants in the current study had low levels of engagement in reading/writing but reading/writing had a significant independent protective effect on risky driving behaviour. Drivers with high engagement in reading/writing had 38% lower odds of being risky drivers than those with lower engagement (OR = 0.62). No previous study that specifically investigated the relation between risky driving and reading or writing activity found any evidence to link them. However, other beneficial effects of reading and writing are demonstrated in the literature (356-358). Reading for pleasure and writing in a journal can have a positive impact on health and well-being. Reading for pleasure lowers stress and reduces depression. A survey-based study found that people who read for pleasure (but not necessarily people who read for school or work purposes) had fewer feelings of stress than non-readers. People who read for pleasure found the relaxing effects of enjoying a good book were greater than if they watched television, flipped through social media feeds, or read other types of material, such as newspapers or textbooks. A growing body of research indicates that reading changes the mind and the brain connections (356).   In 2009, a group of researchers measured the effects of yoga, humor, and reading on the stress levels of students in demanding health science programs in the United States. The study found that 30 minutes of reading lowered blood pressure, heart rate, and feelings of psychological distress just as effectively as yoga and humor did (357).  According to Riordan and Wilson, bibliotherapy is “the guided reading of written materials to gain understanding or to solve problems relevant to a person’s therapeutic needs.” The phenomenon of bibliotherapy (from the Greek words: books and therapy) was coined by Samuel Crothers in 1916 when he recognized the therapeutic benefits of literature. As an adjunct to psychotherapy, bibliotherapy allows people reading or listening to stories to identify with the significant characters. Readers may experience an emotional catharsis as the characters in the story express themselves. Other benefits of bibliotherapy include instilling moral values, shaping behavior, improving and enabling the growth of critical thinking skills, and overall strengthening of personal character (359). There is evidence that reading can also decrease anxiety. Olsen et al., point out that bibliotherapy can 96  be employed by nearly every helping profession, with almost every age group and population (360). However, it is noteworthy that the content of reading matters. Stevens et al., studied 134 undergraduate students and found that reading violent stories significantly increased aggressive thoughts (361). Writing, particularly expressive writing, can also have a positive effect on mental health. Pennebaker and Beall’s seminal study demonstrated that writing about a stressful experience improves indicators of physical health (352). A meta-analysis evaluating more than 400 studies that tested the effects of expressive writing on various outcomes in different populations identified a beneficial effect on physical, psychological and overall functioning outcomes including work-related outcomes, social relationships, cognitive functioning, and school performance (363).   In summary, the current finding is novel in terms of showing the influence of reading and writing activities on driving behaviour. Plausible reasons for a protective effect of reading and writing on risky driving behaviour may be explained through its ability to reduce stress and anxiety, improve mental well-being, and to shape social values and behaviours. This reasoning might suggest that other creative leisure activities, like art, should also have an effect on driving behaviour. The reason the current study found a significant effect for reading/writing but not for art engagement (previous section) might be because of the relatively small amount of time attributed to art engagement in the study group, or the fact that leisure activities may influence behaviour to different extents during different eras of life.   4.3.11 Volunteering and Driving Behaviour  As with reading/writing, the majority of participants in this study had low engagement in volunteering activities. Volunteering was found to be an independent protective factor against risky driving behaviour. Participants who were highly engaged in volunteering activities were less likely to report risky driving behaviour (OR = 0.60).  No research was uncovered on the relationship between volunteering and driving behaviour. Existing studies focused on other risky behaviours. Pro-social activities such as donating to charity or volunteering service to the community may act as protective factors against risky behaviors (364). A review of literature found evidence that prosocial activities such as volunteering decrease alcohol and substance use (71,365), illegal behaviour and crime involvement (366), and improve social behaviour (367).  A longitudinal study surveyed 531 rural teens in grades 10-12, and again in early adulthood 97  and found that rural teens who volunteered frequently and helped others were less likely to engage in substance use in young adulthood (368). The National Longitudinal Study of Adolescent to Adult Health recruited grade 7–12 students in 1994–1995 (n = 20,745) and followed up with them in 2001–2002 (n = 14,322) and in 2008–2009 (n = 12,288). In 2000–2001, participants were retrospectively asked about their volunteering experience from 12 to 18 years of age. Non-volunteers, Self-volunteers, Adult-required volunteers, and Court-ordered volunteers were compared for rates of illegal behavior, arrest, and convictions in adulthood (>18 years of age). The study found that relative to non-volunteers, self-volunteers reported 11% fewer illegal behaviors, 31% fewer arrests, and 39% fewer convictions by age 18–28 years, and 28% fewer illegal behaviors, 53% fewer arrests, and 36% fewer convictions by age 24–34 (367). The altruistic nature of volunteering prompts adolescents to feel compassion for people and society which translates into them avoiding risky behaviour which could harm other people or themselves (369). Social control theory describes the possible reason for this beneficial effect of volunteering. Based on this theory, there is a bond between children and conventional societal institutions (including schools, churches and volunteering groups). When this bond weakens, the probability of problematic behaviours in children increases (370). Moreover, when children volunteer, they give themselves a social identity of belonging to a community that cares about others. This social identity can form their future behaviour (371), like driving. What is more, volunteering provides children with an opportunity to observe prosocial adult behaviour and learn to adjust their own behaviour accordingly or even adapt prosocial norms from the adults. These effects are described by social cognitive theory and normative social behaviour theory respectively (365).  The current finding, of a beneficial effect of volunteering on driving behaviour, is new and extends previous research on the protective influence of volunteering on risky behaviour.  4.3.12 Music and Driving Behaviour  Low engagement in music leisure activities was reported by the majority of participants in this study. There was an association between music engagement and driving behaviour, however after adjusting for other factors, no significant, independent effect was detected.  Previous literature on the association of music and driving mainly concerns a music genre or the influence of listening to music while driving. Brodskey et al., studied the influence of different genres of music on driving behaviour in 85 novice drivers (17-18 years) and found that listening to music 98  with hostile or violent content prompted drivers to increase cruising speeds and spend a higher percentage of time exceeding speed limits. Listening to “energetic music was associated with decreased lateral control, increased excursions from the lane, and an increased tendency to stray onto the shoulder” (372). Greitemeyer et al., studied 103 students aged 18-20 years and reported that male drivers who preferred heavy metal music in general were more prone to reckless behaviour such as not fastening their seat belt, driving fast and risky driving. Conversely drivers who preferred music with prosocial lyrics had less risky driving than those who listened to neutral songs (103).  Although people with high engagement in music-related activity might end up listening to music while driving (373, 376), no study on the association between general level of engagement in music and risky driving behaviour was found. The findings of the current study may be because music engagement as a lifestyle variable was investigated and not the setting, such as in-vehicle music.  4.3.13 TV and driving behaviour  In this study, level of television engagement was low among all participants and there was no association between television engagement and driving behaviour. To the best of our knowledge, there is no research on the relationship between driving behaviour and time spent watching television. However, some associations between television viewing and other risky behaviours have been shown in the literature (377-381). Zilka et al stated that the influence television viewing (alone or with friends) on behavioural reactions is mediated by identification with the TV characters. TV programs with violent-behaviour can prompt adolescents to demonstrate negative and violent reactions (381). Possible reasons for these effects are well explained by normative social behaviour theory: television characters can establish behavioural norms for children, including driving behaviour norms (382). Further, according to social cognitive theory, children who watch TV characters will adjust their own behaviours to model the behaviour of the TV characters (383). Finally, Evolutionary Psychology explains that children who watch TV learn what specific behaviours are attractive for each gender and they adapt these behaviours and express them as their own (384).  There is an interaction between television viewing and personality traits. Weaver et al studied 119 undergraduate students and found that personality trait and media preferences have a strong relationship. Those who scored high for neuroticism preferring information/news and dramatic TV programs, whereas those who scored high for extroversion preferred comedy and adventure movies (385). However, as noted above, there is no research on the overall level of engagement in TV viewing 99  and risky driving. The current study failed to show any significant association between the level of engagement in watching TV and risky driving behaviours.  Leisure activities Summary   Based on psychosocial theories, many types of leisure activities can influence driving behaviour through role modeling, peer effect, a changing of the subjective norm, establishing a social identity, and improving mental health (by offering an outlet for stress, anxiety, and aggression). However, the current finding revealed that few leisure activities independently affected risky driving behaviour. There are several plausible reasons why different leisure activities which could potentially affect driving behaviors according to psychosocial theories, appear to exert a varying influence on driving behaviours. In this study, some leisure activities were more prevalent than others and the effect of more common leisure activities may significantly outweigh the dominant effects of other less common activities. Different leisure activities may have different effects on driving behaviour because of the differences in the nature of the activities and their depth and strength of influence on perception, attitude, values, and motivation of youth. Further, as noted previously, this study was conducted during the school year which likely restricted leisure time, potentially reducing the influence of leisure activity on driving behaviour. A more robust design, such as a longitudinal study (cohort) or experimental (intervention) design which can monitor participants since early childhood to young adult would provide more insight on these relationships. This proposed study can adopt an overarching approach considering leisure activity, factors related to family, peers, and society, long term personality traits, and risky driving behaviour at different ages.   5. Conclusion Road trauma is a significant public health issue as well as a burden to the Canadian health system. Driver behaviour is one of the main contributors to road trauma. An in-depth exploration of factors related to negative driving behaviours is helpful to inform effective countermeasures. This thesis investigated the association between leisure activity and risky driving behaviour and expanded our knowledge of the determinants of risky driving among young people in Canada. The results suggest that risky driving behaviours in young people are common and perhaps more importantly, a substantial minority of young people admit to high risk behaviours such as street racing and running red lights. The presence of these risky driving behaviours clearly suggests the 100  need for continued driver education, persuasion and law enforcement to minimize the frequency of such behaviours  The current study of adolescents and young adults showed that many sociodemographic factors (gender, age, living status, educational status, duration of driving independently, driver car status, employment status, driving exposure, driving license class) had a statistically significant association with driving behaviour.  Among leisure activities, sport engagement and TV viewing were not associated with driving behaviour, while other leisure activities, such as music, playing video games, social media, watching movies, family activities, spending time with friend, drug involvement activity, alcohol involvement activity, and volunteering were associated with driving behaviour.  Despite its limitations (mentioned in next section), the current study provides interesting results concerning sociodemographics, leisure activity, personality traits, and risky driving. These findings have both theoretical and practical implications. They confirm the important influence of lifestyle on risky driving in a young population. Moreover, this study showed that risky driving is associated with different personality traits. It is hoped that the key findings of this thesis may provide other researchers, road transport planners, decision makers, regulators, and organizations data for developing future research and road safety strategies. 5.1 Limitations  The present study has some methodological limitations that should be considered. As a cross-sectional study, this study revealed associations but not a causal relationship between the outcome and predictors, i.e., the associations identified may be difficult to interpret. Thus, this study represents a step forward which could prompt researchers to design research that would allow for causal inference.  The participants for this survey may not be representative of the general population of young drivers. That is, this study may have a selection bias due to the web-based recruitment approach (using Facebook) and the fact that the survey was only available in English. Moreover, only those who use Facebook and Instagram and were interested in the topic likely participated (i.e., sampling bias). The Facebook approach was opted by the YAC in order to provide a larger catchment area. To minimize the “field of interest” selection bias that can occur with web-based research, the Facebook link was disseminated to all young people living in Canada without filtering against their field of 101  interest. However, this approach was not entirely successful as social media advertisements filters can also be set by the “end-user” according to their field of interest.  To mitigate the effect of sampling bias and improve the response rate, we deployed our survey through social media (Facebook and Instagram), and attached the link of the survey to a video advertisement that was developed in consultation with our youth advisory committee to attract youths’ attention. Note that the content of the video was scripted in a way to mitigate the possible effect of video on participants’ responses. That effect was likely to be minor according to the youth advisory committee’s opinion. Other potential types of bias in this study include non-response bias, which is introduced when respondents differ from non-respondents, and “under-coverage bias”, which occurs when some members of the population are under-represented in the survey making the results less generalizable and potentially resulting in higher variance for some subgroup estimates because of small sample size (386,387). To avoid non-response bias, YAC members were engaged to make the study more acceptable to the target group; the survey was posted on Facebook/Instagram for three months and there were frequent reminders; and YAC members disseminated the advertisement through other social media pages. Moreover, an incentive was provided for participation which, in turn, may have resulted in a large sample size. The large sample size, in turn, may have lessened the possibility of there being a significant difference between those who responded and who did not.  Reporting bias is another limitation of this study. This study is survey-based and participant responses to the questions may not necessarily reflect their true actions (e.g., driving behaviour, amount of driving, type of leisure, personality). Thus, this study suffers from the commonly reported limitations associated with measures of behaviours based upon self-reporting. Social desirability may cause unreliable responses. This bias was minimized by using an anonymous survey and reassuring participants about confidentiality in the cover letter of the survey. To diminish the effect of recall bias and misclassification, participants were asked to recall their average experience (e.g., average hours of driving in a week, average leisure activities engagement in a week) in the past three months prior to participation.  There was the possibility that survey fraud, multiple submissions by participants, could skew the results. Survey fraud could occur if respondents completed the survey merely for the sake of getting the incentive, not with a desire to contribute to the study. In this case, they may not answer honestly nor accurately. The UBC survey approved tool (REDCap) provided privacy and confidentiality but could not restrict responses from the same IP address. The monetary incentive for 102  completing the survey was relatively small (a chance at a gift card of $10) and I believe that multiple responses by the same participant occurred infrequently. Moreover, participants had to provide an email address for the incentive that would tend to limit multiple submission by the same person unless they had several different email addresses. There were no repeat email addresses. The large sample size achieved would also attenuate the possible effect of survey fraud.  Abandonment, an issue with lengthy surveys such as this one, may cause lower response rate. To mitigate this effect, the survey was also designed to be “mobile-compatible” and user friendly with the scaling row frozen while scrolling down to diminish frustration when required to move up and down to be reminded of scaling options. Therefore, tapping to answer the question and scrolling down were all the actions required. Moreover, to mitigate the abandonment bias, the questions and instruments in the survey followed general survey design guidelines such that non-threatening questions (sociodemographics and leisure activities) were first (“warm-up”), followed by the risky driving behaviour questionnaire, which was the most interesting but also the most sensitive section according to YAC’s opinion (as “Peak”), and then personality trait (as “Cool down”).  Satisficing bias occurs when respondents select answer options without much thought, using the least mental activity, in order to satisfy the question requirements, rather than reflecting on answers that best represent their opinion (388). To minimize satisficing bias, the survey was made as short as possible by employing accurate and reliable short forms of questionnaires for personality trait (Mini-IPIP), driving behaviour, and socioeconomic status. Moreover, the survey was designed to have a compelling and appealing interface to engage the participants. Given that we did not test sensitivity analysis, the selection of cut off points for categories in Leisure Activity questionnaire may limit the accurate interpretation of the result. However, the Leisure Activity questionnaire and cut-off points were validated by pilot test prior to the deployment.  5.2 Implications Youth and Parents Potential interventions to prevent risky driving behavior in youth can target the driving behavior directly or target one or more of the factors that influence the driving behavior. Non-modifiable factors should not be the target of interventions but can be used to guide and inform interventions to prevent youthful risky driving. Most demographic factors are not possible to change, but they could be considered in decisions about readiness to drive independently. Personality is not 103  usually considered to be modifiable but is another factor that could be considered by parents when making decisions about their teens' readiness to drive independently. Risk perception can be changed, thus providing opportunities for intervention. Several things could be done to target young people's perceptions about driving, gleaned from the environment in which they find themselves. Parents could be taught the importance of setting a good driving example and being good role models themselves. Parents can also be taught / encouraged to become more involved with their teens' driving and to set realistic and effective restrictions on their novice teen drivers to enhance their safety.  Other stakeholders Other stakeholders like schoolboards, policymakers, prevention units, media, vehicle manufacturers/dealers, community groups and safety campaigns organization can also be alerted, educated and become involved.  Creative programs that provide opportunities for youth to engage in leisure activities beneficial to driving behaviour are also needed. Incorporating in school curriculums and partnering with other society resources to facilitate and provide incentive (as external motivation) may help to encourage the beneficial leisure activities like reading and writing. Programs can also engage peers and partners to promote safer driving by young people. Public sectors and funding agencies can take on public education and provide grants and financial resource to support the program and research. The media, both advertising and entertainment, could provide good avenues for interventions, since a high level of engagement was observed in target group. Social media platforms can be used as outlets to promote safe driving behaviour.  Public awareness campaigns are major sources of information and should be effectively used to deliver road safety messages and promote the short and long-term benefits of compliance with traffic laws in improving road safety. The communication messages should be tailored to consider cultural and social factors in order to achieve higher success. 5.3 Future Research Future research can investigate leisure activities in greater detail and adopt more robust designs like cohort (longitudinal) or experimental. Accordingly, the effect magnitude of each leisure activity can be evaluated more accurately. The leisure activities with the higher magnitude of effect (higher odds ratio) in the current study can be placed on top of the priority list. Experimental research 104  can include a variety of interventions, such as educational sessions, youth clubs, public health measures, engaging parents and stakeholders like campaigns and examined the corresponding effects.  Another research project can be validation of the current model and, afterwards, apply to driving licensure to identify potential high-risk drivers. Accordingly, the public authority and insurance companies can provide some recommendation to high risk drivers and their families.   105  Bibliography   1. WHO (World Health Organization). 2015a. Global Status Report on Road Safety 2015. Geneva: WHO. 2. WHO (World Health Organization). 2015b. Fact Sheet on Road Safety. WHO, Geneva. 3. Statistics Canada. Ottawa (ON): 2017. Vital Statistics - Death database 2015.  4. Parachute. Toronto (ON): The Cost of Injury in Canada Report [Internet] Available from: http://www.parachutecanada.org/costofinjury.  5. Canadian Motor Vehicle Traffic Collision Statistics 2015. Available from: http://www.tc.gc.ca/vehicle Collisions. 6. Road safety in Canada [Internet]. Transport Canada; 2011. Available from: https: //www .tc.gc.ca/eng/motorvehiclesafety/tp-tp15145-1201.htm 7. Government of Canada PHA of C. The Cost of Injury in Canada [Internet]. Available from: https://www.  canada.ca/ en/public-health/services/injury-prevention/cost-injury-canada.html 8. Larsson P, Tingvall C. The Safe System Approach – A Road Safety Strategy Based on Human Factors Principles. Engineering Psychology and Cognitive Ergonomics. Applications and Services. 2013;8:10-20. Springer, Berlin. 9. WHO (World Health Organization). 2015. Risk factors for road traffic injuries. Available from: https://www.who.int/violence_injury_prevention/road_traffic/ activities/roadsafety_training_manual_ unit_2.pdf 10. Petridou E, Moustaki M. Human factors in the causation of road traffic crashes. Eur J Epidemiol. 2000;16(9):819-26. 11. Mohan D, Tiwari G, Khayesi M, Nafukho FM. Road traffic injury prevention training manual. WHO. 2006; 21–41. 12. Magnitude and impact of road traffic injuries. Available from: https://www.who.int/violence_injury _prevention/ road_traffic/activities/roadsafety_training_manual_unit_1.pdf 13. Krug EG, Sharma GK, Lozano R. The global burden of injuries. Am J Public Health. 2000 Apr;90(4):523-6.  14. Mercy JA, Rosenberg ML, Powell KE, Broome CV, Roper WL. Public health policy for preventing violence. Health Aff (Millwood). 1993 Winter;12(4):7-29. 15. Implementing road injury prevention programs. National Center for Injury Prevention and Control, Atlanta. Available from:  https://www.cdc.gov/injury/index.html 16. Haddon W Jr. Advances in the epidemiology of injuries as a basis for public policy. Public Health Rep. 1980 Sep-Oct;95(5):411-21.  17. Jafarpour S, Rahimi-Movaghar V. Determinants of risky driving behavior: a narrative review. Med J Islam Repub Iran. 2014 Dec 6;28:142.  18. Munteanu PL, Rosu M, Panaitescu V, Pungǎ A. Human and environmental factors contributing to fatal road accidents in a Romanian population. Romanian J of Legal Med. 2014; 22(2):97–100. 19. Simons-Morton BG, Ouimet MC, Zhang Z, Klauer SE, Lee SE, Wang J, et al. The Effect of Passengers and Risk-Taking Friends on Risky Driving and Crashes/Near Crashes Among Novice Teenagers. J Adol Heal. 2011; 49: 587–93. 20. Thiffault P. Addressing human factors in the motor carrier industry in Canada. Canadian Council of Motor Transport Administrators (CCMTA); 2011. 21. Evans L. Risky driving related to driver and vehicle characteristics. Accident Analysis & Prevention, 1983; 15(2): 121-136. 22. Peden M. World report on road traffic injury prevention. Geneva: World Health Organization; 2004. 217 p. 23. Shope JT. Influences on youthful driving behavior and their potential for guiding interventions to reduce crashes. Inj Prev. 2006; 12:9–14. 24. Ward NJ, Keller SN. Traffic Safety Culture [Internet]. 2012. Available from:https://www.emerald.com /insight/publication /doi/10.1108/9781787146174 25. Rumar K. The Role of Perceptual and Cognitive Filters in Observed Behavior. In: Evans L., Schwing R.C.(eds) Human Behavior and Traffic Safety. Springer, Boston, MA; 1985. 26. Mao X, Yuan C, Gan J, Zhang S. Risk Factors Affecting Traffic Accidents at Urban Weaving Sections: Evidence from China. Int J Environ Res Public Health. 2019 May 1;16(9):1542. 27. Nabi H, Rachid Salmi L, Lafont S, Chiron M, Zins M, Lagarde E. Attitudes associated with behavioral predictors of serious road traffic crashes: results from the GAZEL cohort. Inj Prev. 2007 Feb;13(1):26-31. International Traffic Safety Data and Analysis. Available from: https://www.itf-oecd.org/road 106  28. OECD. Road Safety Annual Report 2017. OECD publisher; 2017. 1–581. 29. Motor Vehicle Collisions: Quick facts and stats - BC Injury Research and Prevention Unit [Internet]. Available from: https://www.injuryresearch.bc.ca/quick-facts/motor-vehicle/ 30. Fergusson D, Swain-Campbell N, Horwood J. Risky driving behaviour in young people: prevalence, personal characteristics and traffic accidents. Aust N Z J Public Health. 2003;27(3):337-42.  31. Rhodes N, Pivik K. Age and gender differences in risky driving: the roles of positive affect and risk perception. Accid Anal Prev. 2011 May;43(3):923-31. 32. Scott-parker B. A comprehensive investigation of the risky driving behaviour of young novice drivers. Queensland University of Technology; 2012. 33. Dula CS, Geller ES. Risky, aggressive, or emotional driving: addressing the need for consistent communication in research. J Safety Res. 2003;34(5):559-66.  34. Steinberg L. A Social Neuroscience Perspective on Adolescent Risk-Taking. Dev Rev. 2008 Mar;28(1):78-106. 35. Gielen AC, Sleet D. Application of behavior-change theories and methods to injury prevention. Epidemiol Rev. 2003;25:65-76. 36. Carter PM, Bingham CR, Zakrajsek JS, Shope JT, Sayer TB. Social norms and risk perception: predictors of distracted driving behavior among novice adolescent drivers. J Adolesc Health. 2014 May;54(5 Suppl):S32-41. 37. Jessor R The Origins and Development of Problem Behavior Theory. Advancing Responsible Adolescent Development. 2016 38. Jessor R. Risk behavior in adolescence: a psychosocial framework for understanding and action. J Adolesc Health. 1991 Dec;12(8):597-605.  39. Rotter JB. Social learning and clinical psychology. New York: Prentice-Hall;1954. 40. Ajzen I. From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior. Berlin, Heidelber, New York: Springer-Verlag; 1985:11-39. 41. Merton RK. Social theory and social structure (rev. ed.). New York: Free Press; 1957. 42. The Handbook of Evolutionary Psychology, edited by David M. Buss, John Wiley & Sons, Inc., 2005 43. Christie D, Viner R. Adolescent development. BMJ. 2005 Feb 5;330(7486):301-4.  44. Carter PM, Bingham CR, Zakrajsek JS, Shope JT, Sayer TB. Social norms and risk perception: predictors of distracted driving behavior among novice adolescent drivers. J Adolesc Health. 2014 May;54(5 Suppl):S32-41.  45. Bianchi A, Summala H. The "genetics" of driving behavior: parents' driving style predicts their children's driving style. Accid Anal Prev. 2004 Jul;36(4):655-9. 46. Taubman - Ben-Ari O, Mikulincer M, Gillath O. From parents to children - Similarity in parents and offspring driving styles. Transportation Research Part F Traffic Psychology and Behaviour. 2005; 8(1):19–29. 47. Ferguson SA, Williams AF, Chapline JF, Reinfurt DW, De Leonardis DM. Relationship of parent driving records to the driving records of their children. Accid Anal Prev. 2001 Mar;33(2):229-34 48. Merrikhpour M. Utilizing the social norms theory for mitigating teen driver distraction. University of Toronto; 2017. 49. Allen JP, Brown BB. Adolescents, peers, and motor vehicles: the perfect storm? Am J Prev Med. 2008 Sep;35(3 Suppl):S289-93. 50. Rosenbloom T, Shahar A, Elharar A, Danino O. Risk perception of driving as a function of advanced training aimed at recognizing and handling risks in demanding driving situations. Accid Anal Prev. 2008 Mar;40(2):697-703.  51. Ram T, Chand K. Effect of drivers’ risk perception and perception of driving tasks on road safety attitude. Transportation Research Part F Traffic Psychology and Behaviour. 2016; 42:162–76.  52. Ulleberg P, Rundmo T. Personality, attitudes and risk perception as predictors of risky driving behaviour among young drivers. Safety Science.2003; 41(5):427-443.  53. Hu T, Xie X, Li J. Negative or positive ? The effect of emotion and mood on risky driving. Transportation Research Part F Traffic Psychology and Behaviour. 2013; 16:29–40.  54. Statistics Canada. Table 45-10-0014-01 Daily average time spent in hours on various activities by age group and sex, 15 years and over, Canada and provinces [Internet]. 2015. Available from:https://www150.statcan. gc.ca/t1/tbl1/en/tv.action?pid=4510001401 55. Bradley GL, Inglis BC. Adolescent leisure dimensions, psychosocial adjustment, and gender effects. Journal of Adolescence. 2012; 35: 1167–76. 56. Hutchinson SL. Physical Activity, Recreation, Leisure, and Sport: Essential Pieces of the Mental Health and Well-being Puzzle Report prepared for: Recreation Nova Scotia and Nova Scotia Department of Health and 107  Wellness Physical Activity, Sport and Recreation Responsibi [Internet]. 2011. Available from: http://benefitshub.ca/entry/physical-activity-recreation-leisure-and-sport-essential-pieces-of-the-ment/ 57. Vingilis E, Yıldırım-Yenier Z, Vingilis-Jaremko L, Wickens C, Seeley J, Fleiter J, et al. Literature review on risky driving videos on YouTube: Unknown effects and areas for concern? Traffic Injury & Prevention. 2017; 18(6):606–15. 58. Westland C. Leisure and Mental Health. World Leisure Recreation. 1992; 34(3):14–9.  59. Han J, Patterson I. An Analysis of the Influence That Leisure Experiences Have on a Person’s Mood State, Health and Wellbeing. Annals of leisure research. 2007; 10(3–4):328-351. 60. Iwasaki Y, Mackay KJ, Mactavish JB, Ristock J, Bartlett J. Voices from the Margins: Stress, Active Living, and Leisure as a Contributor to Coping with Stress. Leisure Science. 2006; 28(2):163–80. 61. Ahonen A. Psychosocial well-being of schoolchildren in the Barents Region : a comparison from the northem parts of Norway, Sweden and Finland and Northwest Russia. Lapland University Press; 2010. 218 p 62. Hertting K, Kostenius C. Organized Leisure Activities and Well-being: Children Getting it Just Right! LARNet. The Cyber Journal of Applied. Leisure and Recreation Research. 2012; 15(2):13-28. 63. Freire T, Teixeira A. The influence of leisure attitudes and leisure satisfaction on adolescents’ positive functioning: The role of emotion regulation. Frontier Psychology. 2018; 9:1-12. 64. Goldin, Philippe & Jazaieri, Hooria & Gross, James. Emotion Regulation in Social Anxiety Disorder. Social Anxiety: Clinical, Developmental, and Social Perspectives: Third Edition. 2014;511-529.p 65. Darling N. Participation in extracurricular activities and adolescent adjustment: Cross-sectional and longitudinal findings. Journal Youth & Adolescent. 2005; 34(5):493–505. 66. Viau A, Denault AS, Poulin F. Organized Activities During High School and Adjustment One Year Post High School: Identifying Social Mediators. Journal Youth & Adolescent. 2015; 44(8):1638–51. 67. Hellandsjø Bu ET, Watten RG, Foxcroft DR, Ingebrigtsen JE, Relling G. Teenage alcohol and intoxication debut: the impact of family socialization factors, living area and participation in organized sports. Alcohol Alcohol. 2002; 37(1):74–80. 68. Badura P,  Sigmundova D, Sigmund E, Geckova M, P J, Reijneveld A, et al. Participation in organized leisure-time activities and risk behaviors in Czech adolescents. International Journal of Public Health. 2017; 62(3):387–96. 69. Hallingberg BE. Adolescent alcohol use and participation in organised activities: A mixed methods study of British young people. Cardiff University; 2014. 70. Zamboanga BL, Carlo G, Raffaelli M. Problem behavior theory: An examination of the behavior structure system in latino and non-latino college students. Interamerican Journal of Psychology. 2004; 38(2):253–62. 71. Alida B, Paul S, Herbert S. A psychometric study of adolescent risk perception. Journal of Adolescence. 1993; 16:153–68.  72. Carnagey NL, Anderson CA. Violent video game exposure and aggression: A literature review. Minerva Psichiatr. 2004; 45(1):1–18. 73. Leichliter JS, Meilman PW, Presley C a, Cashin JR. Alcohol use and related consequences among students with varying levels of involvement in college athletics. Journal of American College Health. 1998; 46:257-262. 74. Lorente FO, Souville M, Griffet J, Grélot L. Participation in sports and alcohol consumption among French adolescents. Addictive Behaviours. 2004; 29(5):941-6. 75. Brenner J, Swanik K. High-risk drinking characteristics in collegiate athletes. Journal of American College Health. 2007; 56(3):267–72. 76. Peretti-Watel P, Beck F, Legleye S. Beyond the U-curve: The relationship between sport and alcohol, cigarette and cannabis use in adolescents. Addiction. 2002; 97(6):707–16. 77. Wichstrom T, Wichstrom L. Does sports participation during adolescence prevent later alcohol, tobacco and cannabis use? Addiction. 2009; 104(1):138–49. 78. Marten MP, Watson JC, Beck NC. Sport-type differences in alcohol use among intercollegiate athletes. Journal of Applied Sport Psychology. 2006; 18(2):136–50. 79. De Looze M, van den Eijnden R, Verdurmen J, et al. Parenting practices and adolescent risk behavior: rules on smoking and drinking also predict cannabis use and early sexual debut. Prevention Science. 2012; 13(6):594-604. 80. Carpentier A, Brijs K, Declercq K, Brijs T, Daniels S, Wets G. The effect of family climate on risky driving of young novices: The moderating role of attitude and locus of control. Accident Analysis & Prevention. 2014; 73:53–64. 108  81. Scott-Parker B, Watson B, King M. Exploring how Parents and Peers Influence the Behaviour of Young Drivers Centre for Accident Research and Road Safety – Queensland, Queensland University of Technology. 2009; (November):809–16. 82. Raymond Bingham C, Zakrajsek JS, Almani F, Shope JT, Sayer TB. Do as I say, not as I do: Distracted driving behavior of teens and their parents. Journal of Safety Research. 2015; 55:21–9. 83. Lam LT. Factors associated with parental safe road behaviour as a pedestrian with young children in metropolitan New South Wales, Australia. Accident Analysis & Prevention. 2001; 33(2):203–10. 84. Taubman-Ben-Ari O, Katz-Ben-Ami L. Family climate for road safety: A new concept and measure. Accident Analysis & Prevention. 2013; 54:1–14. 85. Mayberry M. Adolescent risk behaviors: considering the influences and interactions of multiple contexts [Internet]. University of Illinois; 2008. Available from: https://search.proquest.com/docview/304603645?pq-origsite=gscholar 86. Loubeau PR. Exploration of the barriers to bicycle helmet use among 12 and 13 year old children. Accident Analysis & Prevention. 2000; 32(1):111–5. 87. Brookland R, Begg D, Langley J, Ameratunga S. Parental influence on adolescent compliance with graduated driver licensing conditions and crashes as a restricted licensed driver: New Zealand Drivers Study. 2014; 69:30–9. 88. Wilson RJ, Meckle W, Wiggins S, Cooper PJ. Young driver risk in relation to parents’ retrospective driving record. Journal of Safety Research. 2006; 37(4):325–32. 89. Miller G, Taubman - Ben-Ari O. Driving styles among young novice drivers-The contribution of parental driving styles and personal characteristics. Accident Analysis & Prevention. 2010; 42(2):558–70. 90. Simons-Morton B. Parent involvement in novice teen driving: Rationale, evidence of effects, and potential for enhancing graduated driver licensing effectiveness. 2007; 38(2):193–202. 91. Keating DP, Halpern-Felsher BL. Adolescent Drivers A Developmental Perspective on Risk, Proficiency, and Safety. American Journal of Preventive Medicine. 2008; 35(3S):272–7. 92. Romer D, Stanton BF. Feelings about risk and the epidemic diffusion of adolescent sexual behavior. Prevention Science. 2003; 4(1):39–53. 93. French DC, Dishion TJ. Predictors of early initiation of sexual intercourse among high-risk adolescents. Journal of Early Adolescent. 2003; 23(3):295–315. 94. Aseltine RH, Jr. A Reconsideration of Parental and Peer Influences on Adolescent Deviance. Journal of Health and Social Behavior. 1995 Jun; 36(2):103. 95. Gheorghiu A, Delhomme P, Line M. Peer pressure and risk taking in young drivers’ speeding behavior. Transportation Research Part F Traffic Psychology and Behaviour. 2015;35:101–11 96. Williams AF. Teenage drivers: Patterns of risk. Journal of  Safety Research. 2003; 34(1):5–15. 97. Vingilis E, Seeley J, Wiesenthal DL, Wickens CM, Fischer P, Mann RE. Street racing video games and risk-taking driving: An Internet survey of automobile enthusiasts. Accident Analysis & Prevention. 2013; 50:1–7. 98. Hull JG, Draghici AM, Sargent JD. A longitudinal study of risk-glorifying video games and reckless driving. Psychology of Popular Media Culture. 2012; 1(4):244–53. 99. Li L, Chen R, Chen J. Playing Action Video Games Improves Visuomotor Control. Psychological Science. 2016; 27(8).1092-108. 100. Aljied R, Gagnon S, Kadulina Y, Stinchcombe A, Lemieux C. Driving is not a game: Video game experience is associated with risk-taking behaviours in the driving simulator. Computers in Human Behavior. 2016; 69:415–20.   101. Deng M, Chan AHS, Wu F, Wang J. Effects of racing games on risky driving behaviour, and the significance of personality and physiological data. Injury Prevention. 2015; 21(4):238–44. 102. Beullens K, Roe K, Van den Bulck J. Driving game playing as a predictor of adolescents’ unlicensed driving in Flanders. Journal of Children and Media. 2013; 7:307–18. 103. Greitemeyer T. Exposure to media with prosocial content reduces the propensity for reckless and risky driving. Journal of Risk Research. 2013; 16: 583–94. 104. Fischer P, Vingilis E, Greitemeyer T, Vogrincic C. Risk-taking and the media. Risk Anal. 2011; 31(5):699–705. 105. Chrysalis L, Wright KS. Media influence on perception of driving risk and behaviors of adolescents and emerging adults,Transportation Research Part F: Traffic Psychology and Behaviour. 2018;54: 290-98. 106. Silberman K. Media influence on risky driving behaviours among adolescent and emerging adults. Conference: Association for Psychological Science. May 2014. 109  107. Beullens K, Roe K, Van den Bulck J. The Impact of Adolescents’ News and Action Movie Viewing on Risky Driving Behavior: A Longitudinal Study. Vol. 37, Human Communication Research. 2011. p. 488–508. 108. Rezaee-Zavareh MS, Salamati P, Ramezani-Binabaj M, Saeidnejad M, Rousta M, Shokraneh F, Rahimi-Movaghar V. Alcohol consumption for simulated driving performance: A systematic review. Chin J Traumatol. 2017 Jun;20(3):166-172.  109. Hughes S. Risky driving behaviours among Canadian adults – A Cluster Analysis. Psychology; 2016. 110. Stevely AK, Holmes J, Meier PS. Contextual characteristics of adults' drinking occasions and their association with levels of alcohol consumption and acute alcohol-related harm: a mapping review. Addiction. 2020 Feb;115(2):218-229.  111. Mahmood B, Bhatti JA, Leon A, Gotay C. Leisure Time Physical Activity Levels in Immigrants by Ethnicity and Time Since Immigration to Canada: Findings from the 2011-2012 Canadian Community Health Survey. J Immigr Minor Health. 2019 Aug;21(4):801-810.  112. Capler R, Bilsker D, Van pelt K. Cannabis use and driving: Evidence Review. Canadian Drug Policy Coalition (CDPC) Simon Fraser University.May 2017. 113. Simmons, S. M. The effects of cannabis and alcohol on driving performance and driver behaviour: a systematic review and meta-analysis (Unpublished doctoral thesis). University of Calgary, Calgary, AB.2020. 114. Hayley AC, et al. Risky driving behaviours among stimulant drug users and the role of aggression: Findings from a national survey. Journal Summaries in Internal Medicine. 2019;29:145-151. 115. Dibben N, Williamson VJ. An exploratory survey of in-vehicle music listening. Psychology of Music. 2007;35(4): 571–589.  116. Williamson VJ. Music and driving. MUSIC & CONSUMER BEHAVIOURMUSIC PSYCHOLOGY. Dec 2011. 117. Brodsky W, Olivieri D, Chekaluk E. Music genre induced driver aggression: A case of media delinquency and risk-promoting popular culture. Music & Science. 2018;1:10-15. 118. Tien A, Chu P, Tremblay L. Prevalence of risky driving behaviours on popular television series. Can J Surg. 2018 Oct 1;61(5):355-356.  119. Abi-Jaoude E, Naylor KT, Pignatiello A. Smartphones, social media use and youth mental health. CMAJ. 2020 Feb 10;192(6):E136-E141.  120. Choudhary P, Velaga NR. Analysis of vehicle-based lateral performance measures during distracted driving due to phone use. Transportation Research Part F Traffic Psychology and Behaviour. 2017; 44:120–33.  121. The prevalence and correlates of texting while driving among a population-based sample of Ontario students [Internet]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/30010424 122. Penetration of leading social networks in Canada as of the 3rd quarter of 2017. Available from: https://www.statista.com/statistics/284426/canada-social-network-penetration/ 123. Young male drivers’ perceptions of and experiences with YouTube videos of risky driving behaviours [Internet]. Available from: https://www.ncbi.nlm.nih.gov/pubmed/30086437 124. Gregersen P, Berg HY. Lifestyle and Accidents among Young Drivers. Accident Analysis & prevention. 1994; 26(228):297–303. 125. Bina M, Graziano F, Bonino S. Risky driving and lifestyles in adolescence. Accident Analysis & prevention. 2006; 38:472–81. 126. Lu L, Hu CH. Personality, leisure experiences and happiness. Journal of Happiness Studies. 2005; 6(3):325–42. 127. McGuiggan R. Determining Leisure Preference, Personality or Demographics? ACR Eur Adv. 1999; E-04. 128. Herzberg PY. Beyond “accident-proneness”: Using Five-Factor Model prototypes to predict driving behavior. Journal of research in personality. 2009; 43(6):1096–100. 129. Miles DE, Johnson GL. Aggressive driving behaviors: Are there psychological and attitudinal predictors? Transportation Research Part F Traffic Psychology and Behaviour. 2003; 6(2):147–61. 130. Dahlen ER, Martin RC, Ragan K, Kuhlman MM. Driving anger, sensation seeking, impulsiveness, and boredom proneness in the prediction of unsafe driving. Accident Analysis & prevention. 2005; 37(2):341–8. 131. Mallia L, Lazuras L, Violani C, Lucidi F. Crash risk and aberrant driving behaviors among bus drivers: The role of personality and attitudes towards traffic safety. Accident Analysis & prevention. 2015; 7 9:145–51. 132. Deffenbacher JL, Lynch RS, Oetting ER, Yingling DA. Driving anger: Correlates and a test of state-trait theory. Personality and Individual Differences. 2001; 31(8):1321–1131. 133. Zuckerman M. Behavioral Expression and Biosocial Bases of Sensations Seeking. Cambridge ; New York : Cambridge University Press; 1994. 463 p. 110  134. Grennee K, Krcmar M, Walters LH, Rubin DL, Jerold, Hale L. Targeting adolescent risk-taking behaviors: the contributions of egocentrism and sensation-seeking. Journal of Adolescence. 2000; 23(4):439–61.. 135. Moore S, Parsons J. A research agenda for adolescent risk-taking: where do we go from here? Journal of Adolescence. 2000; 23:371–6. 136. Parsons JT, Siegel  W, Cousins JH. Late adolescent risk-taking: effects of perceived benefits and perceived risks on behavioral intentions and behavioral change. Journal of Adolescence. 1997; 20(4):381–92. 137. John OP, Srivastava S. The Big-Five Trait Taxonomy: History, Measurement, and Theoretical Perspectives. In: Handbook of personality: Theory and research. 1999. 138. Cherry K, Steven G. The Big Five Personality Traits; 5 Major Factors of Personality. Personality and Individual Differences. 205; 39(3):557-567. 139. McCrae RR, John OP. An Introduction to the Five-Factor Model and Its Applications. Journal of Personality. 1992; 60(2):175–215. 140. Fine BJ. Introversion–Extroversion and motor vehicle driver behavior. Perceptual and Motor Skills. 1963; 95–100. 141. Eysenck HJ, Hans J. The structure of human personality. Methuen; 1970. 476 p. 142. Smith DI, Kirkham RW. Relationship between some personality characteristics and driving record. British Journal of Social Psychology. 1981; 20:229–231. 143. Lajunen T. Personality and accident liability: Are extroversion, neuroticism and psychoticism related to traffic and occupational fatalities? Personality and Individual Differences. 2001; 31(8):1365–73. 144. Dahlen ER, White RP. The Big Five factors, sensation seeking, and driving anger in the prediction of unsafe driving. Personality and Individual Differences. 2006; 41(5):903–15. 145. Bettencourt BA, Talley A, Benjamin AJ, Valentine J. Personality and aggressive behavior under provoking and neutral conditions: a meta-analytic review. Psychological Bulletin. 2006; 132(5):751–77. 146. Dahlen ER, Edwards BD, Tubré T, Zyphur MJ, Warren CR. Taking a look behind the wheel: An investigation into the personality predictors of aggressive driving. Accident Analysis & Prevention. 2012; 45:1–9. 147. Arthur W, Doverspike D. Predicting motor vehicle crash involvement from a personality measure and a driving knowledge test. Journal of Prevention & Intervention in the Community. 2001; 22(March 2015):35–42. 148. Arthur W, Graziano WG. The five-factor model, conscientiousness, and driving accident involvement. Journal of Personality. 1996; 64(3):593–618. 149. Jovanović D, Lipovac K, Stanojević P, Stanojević D. The effects of personality traits on driving-related anger and aggressive behaviour in traffic among Serbian drivers. Transportation Research Part F Traffic Psychology and Behaviour. 2011; 14(1):43–53. 150. Taubman-Ben-Ari O, Yehiel D. Driving styles and their associations with personality and motivation. Accident Analysis & Prevention. 2012; 45:416–22. 151. Hirschi, T. Causes of delinquency. Berkeley: University of California Press; 1969. 152. Frunham A. Personality and activity preference. Br J Soc Psychol. 1981; 29:57–68. 153. Hills P, Argyle M. Musical and religious experiences and their relationship to happiness. Personality and Individual Differences. 1998; 25:91–102. 154. Dyce J a., O’Connor BP. The Personalities of Popular Musicians. Vol. 22, Psychology of Music. 1994. p. 168–73. 155. Wills GI. A personality study of musicians working in the popular field. Personality and Individual Differences. 1984; 5:359–61. 156. Lu L, Argyle M. TV watching, soap opera and happiness. Journal of Medical Science. 1993; 9:350–60. 157. Weaver JB. Exploring the Links between Personalities. Personality and Individual Differences.1991; 12(12):1293–9. 158. Sutin AR, Stephan Y, Luchetti M, Artese A, Oshio A, Terracciano A. The five-factor model of personality and physical inactivity: A meta-analysis of 16 samples. Journal of Research in Personality. 2016; 63:22–8. 159. Correa T, Hinsley AW, de Zúñiga HG. Who interacts on the Web? The intersection of users’ personality and social media use. Computers in Human Behavior. 2010; 26(2):247–53. 160. Ebeling-Witte S, Frank ML, Lester D. Shyness, Internet use, and personality. CyberPsychology & Behavior. 2007; 10(5):713-716. 161. Ehrenberg A, Juckes S, White KM, Walsh SP. Personality and Self-Esteem as Predictors of Young People’s Technology Use. CyberPsychology & Behavior. 2008 Dec; 11(6):739–41. 162. Daniell H. NIH Public Access. 2012; 76:211–20. 111  163. Kraaykamp G, van Eijck K. Personality, media preferences, and cultural participation. Personality and Individual Differences. 2005; 38(7):1675–88. 164. Shope JT, Bingham CR. Teen Driving. Motor-Vehicle Crashes and Factors That Contribute. Am J Prev Med. 2008;35(3 SUPPL.):S261-71. 165. Stankowich T, Blumstein DT. Fear in animals: a meta-analysis and review of risk assessment. R Soc. 2005;272(1581):2626–34. 166. Statistics Canada. How do Canadians spend their time? 2015. Available from: https:// www. 150. Statcan .gc.ca/n1/en//pub/75-006-x/2018001/article/54947-eng.pdf 167. Jessor R. Problem Behavior Theory and the Social Context. Vol. 3. 2017. 168. Green LW, Kreuter MW. Health promotion planning : an educational and ecological approach. Mayfield Pub. Co; 1999. 621 p. 169. McGinnis JM, Williams-Russo P, Knickman JR. The case for more active policy attention to health promotion. Health Affairs. 2002; 21(2):78–93. 170. WHO. Health Behaviour in School-aged Children (HBSC). 2016. 1–7 p. 171. Classification of highest education in Canada. Available from: https://www23.statcan.gc.ca/imdb/p3VD.pl?Function=getVD&TVD=306182&CVD=306183&CLV=0&MLV=4&D=1 172. Statistics Canada. Census in Brief Young adults living with their parents in Canada in 2016. Stat Canada Cat no 98-200-X2016008 [Internet]. 2017; (98). Available from: http://www12.statcan.gc.ca/census-recensement/2016/as-sa/98-200-x/2016008/98-200-x2016008-eng.cfm 173. Scott-Parker B, Watson B, King MJ. The risky behaviour of young drivers: Developing a measurement tool. Proceedings of the 24th Canadian Multidisciplinary Road Safety Conference, Niagara Falls, Canada, June 6-9, 2010. 174. Bridie Scott-Parker CP. Validation of the Behaviour of Young Novice Drivers Scale (BYNDS) in a New Zealand young driver population [Internet]. 2014. Available from: https://www.semanticscholar.org/ paper/Validation-of-the-Behaviour-of-Young-Novice-Drivers-Scott-Parker-Proffitt/ff0ee867e77177a9014c67a 4719caca1a2415fb8 175. Scott-Parker B. A comprehensive investigation of the risky driving behaviour of young novice drivers. PhD by Publication, Queensland University of Technology; 2012. 176. Scott-Parker B, Watson BC, King MJ. The Behaviour of Young Novice Drivers Scale (BYNDS). 2010. 177. Center for injury research and prevention. The Behaviour of Young Novice Drivers Scale (BYNDS). Philadelphia; 2010. 178. McDonald CC, Sommers MS, Fargo JD, Seacrist T, Power T. Simulated Driving Performance, Self-Reported Driving Behaviors, and Mental Health Symptoms in Adolescent Novice Drivers. Nursing Research. 2018;67(3):202-211. 179. Scott-Parker B Revisiting the concept of the ‘problem young driver’ within the context of the ‘young driver problem’: Who are they? Accident Analysis & Prevention. 2013;59: 144-152. 180. Leisure-time activity survey questions. Available from: https://www.questionpro.com/survey-templates/leisure-time-activities/ 181. Rentfrow PJ. Listening, Watching, and Reading: The Structure and Correlates of Entertainment Preferences. 2012; 79(2):1–28. 182. Council of ministry of Education C. Education in Canada. Vol. 309. 1984 183. Youth at Leisure- How the Youth of the World Spend Their Time. Available from: https://www.un.org /esa/socdev/unyin/documents/ch08.pdf 184. CMEC. Ministerial Statement on Quality Assurance of Degree Education in Canada. 2007; 1–14. 185. Duuren D van. The relationship between personality and preference for either art house or mainstream movies. University of Twente; 2008. 186. Daily leisure activities among Canadians. Available from https://www.booknetcanada.ca/blog/2017 /4/20/canadians-and-their-leisure-time-during-2017-leisure-amp-reading-study. 187. Goldberg LR. A broad-bandwidth, public domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.). Personality psychology in Europ. Tilburg, Holland: Tilburg University Press. 1999;7:7-28. 188. Donnellan MB, Oswald FL, Baird BM, Lucas RE. The Mini-IPIP Scales: Tiny-yet-effective measures of the Big Five Factors of Personality. Psychological Assessment. 2006; 18(2):192–203.  112  189. Aldasaro R, Hanahan MIJS, Auer DAJB. Psychometric Properties of the Mini-IPIP in a Large, Nationally Representative Sample of Young Adults. 2013; 95(1):74–84. 190. The Mini-IPIP questionnaire. Available from https://ipip.ori.org/newScoringInstructions.htm 191. Converting IPIP Item Responses to Scale Scores [Internet]. Available from: https://ipip.ori.org/new Scoring Instructions.htm 192. Interpreting Individual IPIP Scale Scores [Internet]. Available from: https://ipip.ori.org/ Interpreting Individual IPIP ScaleScores.html 193. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap) A metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics. 2008; 42:377–81. 194. Seymour K. Using incentives: Encouraging and recognising participation in youth research. Youth Studies Australia. 2012; 31(3):51–9.  195. Gareth J, Witten D, Hastie T, Tibshirani R. An Introduction to Statistical Learning: With Applications in R. Springer Publishing Company, Incorporated.2014. 196. Bruce P, Bruce A. Practical Statistics for Data Scientists. O’Reilly Media; 2017. 197. World Health Organization: Save lives: a road safety technical package. 2017. Availbale at: https://apps. who.int/iris/bitstream/handle/10665/255199/9789241511704-eng.pdf?sequence=1 198. Otto C, Haller AC, Klasen F, et al. Risk and protective factors of health-related quality of life in children and adolescents: Results of the longitudinal BELLA study. PLoS One. 2017; 12(12):341-45. 199. Wille N, Bettge S, Ravens-Sieberer U, BELLA study group. Risk and protective factors for children's and adolescents' mental health: results of the BELLA study. European Child and Adolescent Psychiatry. 2008; 17:133–147. 200. Hölling H, Schlack R, Dippelhofer A, Kurth BM. [Personal, familial and social resources and health-related quality of life in children and adolescents with chronic conditions]. Bundesgesundheitsblatt, Gesundheitsforschung, Gesundheitsschutz. 2008; 51:606–620. German.  201. Oppenheim I, Oron-Gilad T, Parmet Y, Shinar D. Can traffic violations be traced to gender-role, sensation seeking, demographics and driving exposure? Transportation Research Part F Traffic Psychology and Behaviour. 2016; 4 3:387–95.  202. Zhang B, Winter JV, Silvia F, Happee R, & Martens M. Determinants of take-over time from automated driving: A meta-analysis of 129 studies. Transportation Research Part F: Traffic Psychology and Behaviour. 2019; 64:345-58. 203. OECD. Road Safety Annual Report 2017. OECD publisher; 2017. 1–581 p. 204. OECD, ITF. Road Safety Annual Report 2014 [Internet]. 2014. Available from: https://www.oecd-ilibrary.org/transport/road-safety-annual-report-2014_irtad-2014-en 205. Curry AE, Metzger KB, Williams AF, Tefft BC. Comparison of older and younger novice driver crash rates: Informing the need for extended Graduated Driver Licensing restrictions. Accident Analysis and Prevention. 2017; 108: 66. 206. Walsh SP, White KM, Hyde MK, Watson B. Dialling and driving: Factors influencing intentions to use a mobile phone while driving. Accident Analysis and Prevention. 2008; 40(6):1893–900. 207. Efrat K, Shoham A. The theory of planned behavior, materialism, and aggressive driving. Accident Analysis and Prevention. 2013; 59:459-465. 208. Măirean C, Havârneanu GM, Popușoi SA, Havârneanu CE. Traffic locus of control scale – Romanian version: Psychometric properties and relations to the driver’s personality, risk perception, and driving behavior. Transportation Research Part F: Traffic Psychology and Behaviour. 2017; 45:131–46. 209. Renner W, Anderle FG. Venturesomeness and extroversion as correlates of juvenile drivers’ traffic violations. Accid Anal Prev. 2000;32(5):673–8. 210. Glendon AI, Mcnally B, Jarvis A, Chalmers SL, Salisbury RL. Evaluating a novice driver and pre-driver road safety intervention. Accid Anal Prev [Internet]. 2014;64:100–10. Available from: http://dx.doi.org/10.1016/j.aap.2013.11.017 211. Byrnes, J. P., D. C. Miller, and W. D. Schafer. Gender differences in risk taking: A meta-analysis. Psychological Bulletin, Vol. 125, No. 3, May 1999, pp. 367–383. 212. Vogel DL, Wester SR, Heesacker M, Madon S. Confirming gender stereotypes: A social role perspective. Sex Roles. 2003; 48(11-12) : 519 -28 113  213. Chris G. Sibley, Niki Harré. The impact of different styles of traffic safety advertisement on young drivers’ explicit and implicit self-enhancement biases, Transportation Research Part F: Traffic Psychology and Behaviour. 2009; 12(2):159-167.  214. Jonah BA. Age differences in risky driving. Health Education Research,1990;5(2):139–149. 215. Ozkan T and Lajunen T. What causes the differences in driving between young men and women? The effects of gender roles and sex on young drivers’ driving behaviour and self-assessment of skills. Transportation Research Part F: Traffic Psychology and Behaviour. 2006; 9: 269–277. 216. Leandro M. Young drivers and speed selection: A model guided by the Theory of Planned Behavior. Transportation Research Part F: Traffic Psychology and Behaviour. 2012; 15(3):219–32. 217. Donovan JE. Problem-Behavior Theory and the Explanation of Adolescent Marijuana Use. Journal of Drug Issues. 1996; 26(2): 379–404. 218. Jessor R. Problem Behavior Theory,A Half-Century of Research. In: The Origins and Development of Problem Behavior Theory. 1st ed. Springer International Publishing; 2016. p 239–54. 219. Reeder AI, Begg DJ, Alsop JC, McLaren RL, Nada-Raja S. A longitudinal investigation of psychological and social predictors of traffic convictions among young New Zealand drivers. Transp Res Part F Traffic Psychol Behav. 2002;1(1):25–45. 220. Shope J, Waller P, Lang S. Alcohol-related predictors of adolescent driving: gender differences in crashes and oenses. Accident Analysis and Prevention. 1996; 28:755-764. 221. Schmidt SA. Youth Risky Driving Behaviours: Advancements in Measurement and Theory. Guelph; 2012. 222. Atombo C, Wu C, Zhong M, Zhang H. Investigating the motivational factors influencing drivers intentions to unsafe driving behaviours: Speeding and overtaking violations. Transportation Research Part F: Traffic Psychology and Behaviour. 2016; 43:104–21.  223. Brief PA. Youth Risk Taking Behavior : The Role of Schools. 2007; 1563(310). 224. Cutler DM, Lleras-Muney L. Understanding differences in health behaviors by education. Journal of Health Economics 2010; 29 (1): 1-28. 225. Juurges H, Reinhold C, Salm M. Does schooling affect health behavior? Evidence from the educational expansion in Western Germany. Economics of Education Review. 2011; 30 (5): 862-872. 226. Caragata Nasvadi G, MA, Wister A. Do Restricted Driver's Licenses Lower Crash Risk Among Older Drivers? A Survival Analysis of Insurance Data From British Columbia, The Gerontologist, Volume 49, Issue 4, August 2009, Pages 474–484, https://doi.org/10.1093/geront/gnp039 227. Juarez P, Schlundt DG, Goldzweig I, Stinson N. A conceptual framework for reducing risky teen driving behaviors among minority youth. Injury Prevention. 2006; 12(SUPPL. 1):49–55. 228. Injury Insight. 2014; (November):1–3. Available from: http://www.injuryresearch.bc.ca 229. Ehsani JP, Raymond Bingham C, Shope JT. Graduated driver licensing for new drivers effects of three states’ policies on crash rates among teenagers. American Journal of Preventive Medicine. 2013; 45(1):9–18.  230. Mayhew DR, Simpson HM, Williams AF, Desmond K. Scotia’s Graduated Licensing Program. 2002; (August). 231. Mayhew DR, Simpson HM, Singhal D. Best practices for graduated driver licensing in Canada. Traffic Inj Res Found [Internet]. 2005; Available from: http://assets.ibc.ca/Documents/AutoInsurance/road_ safety/Best_ Practices_Driver_Distraction_Study_Oct2005.pdf 232. Jimenez N. Child death review panel and recommendation. Burnaby; 2018. 233. Tao D, Zhang R, Qu X. The role of personality traits and driving experience in self-reported risky driving behaviors and accident risk among Chinese drivers. Accident Analysis & Prevention. 2017; 99:228–35. 234. Kunnawee K, Ketphat M, Jiwattanakulpaisarn P.  Application of the theory of planned behaviour to predict young drivers' speeding behaviour.  Injury Prevention 2012; 18:A200. 235. Pengfei Li, Jianjun Shi, Xiaoming Liu, Haizhong Wang. The Theory of Planned Behavior and Competitive Driving in China, Procedia Engineering. 2016; 137:362-371, 236. Mekonnen TH, Tesfaye YA, Moges HG, Gebremedin RB. Factors associated with risky driving behaviors for road traffic crashes among professional car drivers in Bahirdar city, northwest Ethiopia, 2016: A cross-sectional study. Environmental Health and Preventive Medicine. 2019; 24(1):1–9. 237. Chen HYW, Donmez B, Hoekstra-Atwood L, Marulanda S. Self-reported engagement in driver distraction: An application of the Theory of Planned Behaviour. Transportation Research Part F: Traffic Psychology and Behaviour. 2016; 38:151–63.  238. Mousavi Bazzaz M, Zarifian A, Emadzadeh M, Vakili V. Driving Behaviors in Iran: A Descriptive Study Among Drivers of Mashhad City in 2014. Global Journal of Health Science. 2015; 7(7 Spec No):39-45. 114  239. Hosking J, Ameratunga S, Exeter D, Stewart J, Bell A. Ethnic, socioeconomic and geographical inequalities in road traffic injury rates in the Auckland region. Australian and New Zealand Journal of Public Health.2013; 37: 162–167. 240. José Luis Machado-León, Juan de Oña, Rocío de Oña, Laura Eboli, Gabriella Mazzulla, Socio-economic and driving experience factors affecting drivers’ perceptions of traffic crash risk, Transportation Research Part F: Traffic Psychology and Behaviour. 2016;37: 41-51, 241. Banakar R, Fard SN. Driving dangerously: Law, culture and driving habits in Iran. British Journal of Middle Eastern Studies. 2012; 39: 241–257. 242. Stephens V, Haigney D. Self-reported risky driver behaviour and the Big Five personality traits. Psychology Journal of UK. 2011; (May):1–28. 243. Wang Y, Qu W, Ge Y, Sun X, Zhang K. Effect of personality traits on driving style: Psychometric adaption of the multidimensional driving style inventory in a Chinese sample. PLoS One. 2018; 13(9):1–17. 244. Renner W, Anderle FG. Venturesomeness and extroversion as correlates of juvenile drivers’ traffic violations. Accident Analysis & Prevention. 2000; 32(5):673–8. 245. Aniței M, Chraif M, Burtăverde V, Mihaila T. The Big Five personality factors in the prediction of aggressive driving behavior among Romanian youngsters. International Journal of Traffic and Transportation Psychology. 2014; 2(1):7–19.  246. Shen B, Qu W, Ge Y, Sun X, Zhang K. The relationship between personalities and self-report positive driving behavior in a Chinese sample. PLoS One. 2018; 13(1):1–16. 247. Akbari M, Lankarani K, Heydari S, Motevalian S, Tabrizi R, Asadi-Shekari Z, & Sullman, M. Meta-analysis of the correlation between personality characteristics and risky driving behaviors. Journal of Injury and Violence Research. 2019; 11(2): 107-122. 248. Costa PT, Terracciano A, McCrae RR. Gender differences in personality traits across cultures: Robust and surprising findings. Journal of Personality and Social Psychology. 2001; 82(2): 322-331.  249. Deffenbacher JL, Deffenbacher DM, Lynch RK, Richards TL. Anger, aggression, and risky behavior: A comparison of high and low anger drivers. Behavior Research and Therapy.2018; 41(6), 701-718.  250. Dollard J, Doob LW, Miller NP, Mowrer OH, Sears RR. Frustration and aggression. New Haven, CT: Yale University Press; 1939. 251. Barenbaum NB, Winter DG. History of modern personality theory and research. Handbook of personality. Theory and research (3rd ed). New York, NY: Guilford; 2008.3-25P. 252. Deffenbacher JL, Huff ME, Lynch RS, Oetting ER, Salvatore NF. Characteristics and treatment of high-anger drivers. Journal of Counselling Psychology, 2000; 47(1): 5-17. 253. DeVellis RF. Scale development: Theory and applications (2nd ed.). Thousand Oaks, CA: Sage; 2003. 254. Harris PB, Houston JM, Vazquez JA, Smither JA, Harms A, Dahlke JA, et al. The Prosocial and Aggressive Driving Inventory (PADI): A self-report measure of safe and unsafe driving behaviors. Accident Analysis & Prevention. 2014; 72:1-8.  255. Berdoulat E, Vavassori D, Sastre MTM. Driving anger, emotional and instrumental aggressiveness, and impulsiveness in the prediction of aggressive and transgressive driving. Accident Analysis & Prevention.2013; 50:758-767.  256. Lahey BB. Public health significance of neuroticism. Am Psychol. 2009;64(4):241-256. doi:10.1037/a0015309 257. Clarke, S, & Robertson IT. A meta-analytic review of the Big Five personality factors and accident involvement in occupational and non-occupational settings. Journal of Occupational and Organizational Psychology, 2017; 78(3): 355-376.  258. Bolger N, Zuckerman A. A framework for studying personality in the stress process. Journal of Personality and Social Psychology. 1995;69:890–902. 259. Bone SA, Mowen JC. Identifying the traits of aggressive and distracted drivers: A hierarchical trait model approach. Journal of Consumer Behavior.2006; 5(5): 454-464.  260. Kong J, Zhang K, Chen X. Personality and Attitudes as Predictors of Risky Driving Behavior: Evidence from Beijing Drivers. In: Duffy V.G. (eds) Digital Human Modeling and Applications in Health, Safety, Ergonomics, and Risk Management. Healthcare and Safety of the Environment and Transport. DHM 2013. Lecture Notes in Computer Science, vol 8025. Springer, Berlin, Heidelberg 261. Booth-Kewley, Vickers. Association between major domains of health behaviour. Development. (September 1994). 115  262. Monteiro RP, Coelho GL de H, Hanel PHP, Pimentel CE, Gouveia V V. Personality, dangerous driving, and involvement in accidents: Testing a contextual mediated model. Transp Res Part F Traffic Psychol Behav. 2018; 58:106–14.  263. Eisenberg N, Duckworth AL, Spinrad TL, Valiente C. Conscientiousness: origins in childhood?. Dev Psychol. 2014;50(5):1331-1349. doi:10.1037/a0030977 264. Ardahan F, Mert M. the Factors Affecting Individuals’ Participation in Cycling, Mountaineering and Trekking Activities: an Application of Probit Analysis for Turkey Case. Pamukkale J Sport Sci [Internet]. 2014;5(1):128–50. Available from: http://pjss.pau.edu.tr 265. Wan CS, W. Chiou W .Why are adolescents addicted to online gaming? An interview study in Taiwan. Cyberpsychology and Behavior. 2006; 9(6):762-766. 266. Cole H, Griffiths, MD. Social interactions in massively multiplayer online role-playing gamers. Cyberpsychology and Behavior. 2007; 10(4):575-583. 267. Blumberg FC, Altschuler EA, Almonte DE, Mileaf MI. The impact of recreational video game play on children's and adolescents' cognition. New Directions for Child and Adolescent Development. 2013;139:41-50 268. Rehbein F, Psych G, Kleimann M, Mediasci G, and Mößle T. Prevalence and risk factors of video game dependency in adolescence: results of a German nationwide survey. Cyberpsychology, Behavior, and Social Networking .2010; 13(3):269–277. 269. Ko C, Yen J, Chen C, Chen S, Yen C. Gender differences and related factors affecting online gaming addiction among Taiwanese adolescents. Journal of Nervous and Mental Disease. 2005; 193(4):273-277. 270. Gentile D. Pathological video-game use among youth ages 8 to 18: a national study: Research article. Psychological Science. 2009; 20(5):594–602, 2009. 271. Ipsos MediaCT. The essential facts about the computer and video game industry. Retrieved from: http://www.theesa.com/wp-content/uploads/2014/10/ESA_EF_2014.pdf 272. Schou C, Billieux J, Griffiths MD, Kuss DJ, Demetrovics Z, Mazzoni E, Pallesen S. The relationship between addictive use of social media and video games and symptoms of psychiatric disorders: A large-scale cross-sectional study. Psychology of Addictive Behaviors. 2016; 30(2):252–262. 273. Rehbein F, Kliem S, Baier D, Mößle T, Petry N.M. Prevalence of Internet gaming in German adolescents: Diagnostic contribution of the nine DSM-5 criteria in a state-wide representative sample. Addiction. 2015; 110(5):842–851. 274. Mei S, Yau YH, Chai J, Guo J, Potenza MN. Internet use, well-being, self-esteem and self-control: Data from a high-school survey in China. Addictive Behaviors. 2016; 61:74–79. 275. Saquib N, Saquib J, Wahid A, et al. Video game engagment and psychological distress among expatriate adolescents in Saudi Arabia. Addict Behaviour Reports. 2017; 6:112-117.  276. Wang CW, Chan CL, Mak KK, Ho SY, Wong PW, Ho RT. Prevalence and correlates of video and Internet gaming addiction among Hong Kong adolescents: A pilot study. Scientific World Journal. 2014; 87:46-48. 277. Age breakdown of video game players in the United States in 2019. Available from; https://www.statista.com/statistics/189582/age-of-us-video-game-players-since-2010/  278. Fischer P, Kubitzki J, Guter S, Frey D. Virtual Driving and Risk Taking: Do Racing Games Increase Risk-Taking Cognitions, Affect, and Behaviors? Journal of experimental psychology. Applied. 2007; 13: 22-31.  279. Do Videogames Cause Stress in Kids? It’s Complicated [Internet]. Available from: http://mediashift.org/2013/03/do-videogames-cause-stress-in-kids-its-complicated074/ 280. Gentile DA, Choo H, Liau A, Sim T, Li D, Fung D, Khoo A. Pathological video game use among youth: A two-year longitudinal study. Pediatrics. 2011; 127(2):e319–329. 281. Unsworth G, Devilly GJ, Ward T. The effect of playing violent video games on adolescents: should parents be quaking in their boots? Psychology Crime and Law. 2007; 13:383–394. 282. Wayne NL, Miller GA. Impact of gender, organized athletics, and video gaming on driving skills in novice drivers. PLoS One. 2018; 13(1):e0190885. Published 2018 Jan 24. doi:10.1371/journal.pone.0190885. 283. Tremblay MS, Chaput JP, Adamo KB. et al. Canadian 24-Hour Movement Guidelines for the Early Years (0–4 years): An Integration of Physical Activity, Sedentary Behaviour, and Sleep. BMC Public Health.2017; 17: 870-874.  284. Kramer AF, Colcombe S, Erickson K, Belopolsky A, McAuley E, Cohen NJ, Webb A, Jerome GJ, Marquez DX, Wszalek TM. analysis. Journal of Sport & Exercise Effects of aerobic fitness training on human cortical function: a proposal. Journal of Molecular Neuroscience. 2002; 19:227–231. 285. Hawkins HL, Kramer AF, Capaldi D. Aging, exercise, and attention. Psychology Aging. 1992; 7:643–653. 116  286. Shay KA, Roth DL. Association between aerobic fitness and visuospatial performance in healthy older adults. Psychology Aging. 1992; 7:15–24. 287. Foy HJ, Runham P, Chapman P (2016) Prefrontal Cortex Activation and Young Driver Behaviour: A fNIRS Study. PLoS ONE 2016; 11(5):e0156512. https://doi.org/10.1371/journal.pone.0156512 PMID: 27227990 288. Lai SK, Costigan S, Morgan P, et al. Do school-based interventions focusing on physical activity, fitness, or fundamental movement skill competency produce a sustained impact in these outcomes in children and adolescents? A systematic review of follow-up studies. Sports Medicine. 2014;44:67-79.  289. Sallis JF, Prochaska JJ, Taylor WC. A review of correlates of physical activity of children and adolescents. Medicine and Science in Sports and Exercise. 2000; 32:963–75. 290. Aarnio M, Kujala UM, Kaprio J. Associations of health-related behaviors, school type and health status to physical activity patterns in 16 year old boys and girls. Scandinavian journal of social medicine.1997; 25:156–67 291. Thorlindsson, T. Sport, peers and substance use. In: Bouroncle, A, Rauhamäki, S editors. Sport and substance use in the Nordic countries. Nad publication no. 45. Helsinki: Hakapaino Oy; 2005. pp. 133–58. 292. Tao FB, Xu ML, Kim SD, Sun Y, Su PY, Huang K. Physical activity might not be the protective factor for health risk behaviours and psychopathological symptoms in adolescents. Journal of Paediatrics & Child Health 2007; 43:762–7. 293. Rainey CJ, McKeown RE, Sargent RG, Valois RF. Patterns of tobacco and alcohol use among sedentary, exercising, nonathletic, and athletic youth. Journal of School Health.1996; 66:27–32. 294. Hancock PA, Kane MJ, Scallen S, Albinson CB. Effects of Gender and Athletic Participation on Driving Capability. International Journal of Occupational Safety and Ergonomics. 2002; 8: 281–292. 295. Matos R and Godinho M. Useful Field of Vision and Peripheral Reaction Time in novice driversÐTransfer to a real driving situation after a Perceptual-Motor Training Program. Adv Transport Stud. 2009; 19:77-84. 296. Social Media Usage: 2005-2015. Available OCTOBER 8, 2015 at https://www.pewresearch.org/internet /2015/10/08/social-networking-usage-2005-2015/ 297. Branley DB, Covey J. Is exposure to online content depicting risky behavior related to viewers' own risky behavior offline? Computers in Human Behavior. 2017; 75: 283-287. 298. Nielson V. How teens use media: A Nielson report on the myths and realities of teen media trends. 2009Retrievedfromhttp://blog.nielsen.com/nielsenwire/reports/nielsen_howteensusemedia_june09.pdf. 299. Strasburger VC, Jordan AB, Donnerstein E .Health effects of media on children and adolescents. Pediatrics. 2010; 125(4):756-67. 300. Haque M, Washington S. A parametric duration model of the reaction times of drivers distracted by mobile phone conversations. Accident Analysis & Prevention 2014;62: 42-53. 301. Zhou R, Yu M, Wang X. Why do drivers use mobile phones while driving? The contribution of compensatory beliefs. PLoS One. 2016; 11(8):3–9. 302. He J, Chaparro A, Crandall J, Turner C, Turner K, Ellis J. Mutual interferences of driving and texting performance. Human Factors and Ergonomics Society. 2014; Jan: 2151–5. 303. Dunlap J, Murtagh J. Face-To-Face Interaction: Emoticons and Electronically Mediated Communication. Emotions, Technology, Design, and Learning, 2016.  304. Havarnearnu C, Popusoi S. Does mass-media message influence aggressive and risky driving? A pilot study. Conference: International Conference Psychology and the Realities of the Contemporary World - 6th edition - PSIWORLD 2015 305. Hamer AH, Konijn EA. Adolescents' media exposure may increase their cyberbullying behavior: a longitudinal study. J Adolesc Health. 2015;56(2):203-208.  306. Jamieson P, Romer D. The Changing Portrayal of Adolescents in the Media since 1950. Oxford university press; 2008. 307. Bandura A. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs: Prentice-Hall; 1986. 308. Short SE, Mollborn S. Social determinants and health behaviors: conceptual frames and empirical advances. Current Opinion in Psychology. 2015; 5:78–84.  309. Pachana N, Jetten J, Gustafson L, Liddle J.  To be or not to be (an older driver): Social identity theory and driving cessation in later life. Ageing and Society. 2017; 37(8): 1597-1608.  310. Tontodonato P & Drinkard. Social Learning and Distracted Driving among Young Adults. American Journal of Criminal Justice. 2020:1: 1-23. 117  311. Riddle K. Cultivation Theory Revisited: The Impact of Childhood Television Viewing Levels on Social Reality Beliefs and Construct Accessibility in Adulthood (Conference Papers). International Communication Association. 2009. pp. 1–29. 312. Havarnearnu C, Popusoi S. Does mass-media message influence aggressive and risky driving? A pilot study. Conference: International Conference Psychology and the Realities of the Contemporary World - 6th edition - PSIWORLD 2015 313. Anderson C, Bushman BJ. Human aggression. Psychology. 2002;53 (1): 27. 314. Guardia JG. Developing who I am: A self-determination theory approach to the establishment of healthy identities. Educational Psychologist. 2009; 44:90–104.  315. Niederkrotenthaler T., Voracek M., Herberth A., Till B., Strauss M, Etzersdorfer E,  Sonneck G. Role of media reports in completed and prevented suicide: Werther v. Papageno effects. The British Journal of Psychiatry.2010;197(3): 234-243. 316. Siti H, Binti AH, Azmawati B, Mohammed N. Family Characteristics Associate with Risk Taking Behaviour among Urban and Rural Adolescents in Two Districts in Selangor. A Comparative Study. Procedia - Social and Behavioral Sciences. 2013; 19: 581-587. 317. Machteld H, Machteld SD, Judith S, Wilma RM. The relationship between parenting and delinquency: A meta-Analysis Journal of Abnormal Child Psychology. 2009; 37: 749-775. 318. Simons-Morton BG, Ouimet MC, Catalano RF. Parenting and the young driver problem. American Journal of  Preventive Medicine. 2008; 35(3):294-303 319. Allahverdipour H, Abbasi-Ghahramanloo A, Mohammadpoorasl A, Nowzari P. Cigarette Smoking and its Relationship with Perceived Familial Support and Religiosity of University Students in Tabriz. Iran J Psychiatry. 2015; 10(3):136-143. 320. Paul B. Stretesky and Michael J. The Relationship between Lead and Crime. Journal of Health and Social Behavior 2004; 45(2): 214-229. 321. Imberman SA, Kugler AD, Sacerdote BI. Katrina’s children: Evidence on the structure of peer effects from hurricane evacuees. American Economic Review. 2012; 102(5):2048–2082.  322. Carrell SE, Malmstrom FV, West JE. Peer effects in academic cheating. Journal of Human Resources. 2008; 43(1):173–207.  323. Alexander C, Piazza M, Mekos D, Valente T. Peers, schools, and adolescent cigarette smoking. Journal of Adolescent Health. 200; 29(1):22-30. 324. Clark AE, Lohéac Y. "It wasn't me, it was them!" social influence in risky behavior by adolescents. Journal of Health & Economy. 2007; 26(4):763-84.  325. Robalino JD, Macy M. Peer effects on adolescent smoking: Are popular teens more influential?. PLoS One. 2018;13(7):134-145. 326. Ennett ST, Bauman KE, Hussong A, Faris R, Foshee VA, Cai L, et al. The Peer Context of Adolescent Substance Use: Findings from Social Network Analysis. Journal of Research on Adolescence. 2006;16(2):159–186. 327. Borsari B, Carey KB. How the quality of peer relationships influences college alcohol use. Drug Alcohol Rev. 2006;25(4):361-370.  328. Leather N. Risk-taking behaviour in adolescence: A literature review for professionals working with children in the hospital and community. Journal of child health care. 200913: 295-304. 329. Horvath C, Lewis I, Watson B. Peer passenger identity and passenger pressure on young drivers’ speeding intentions. Transportation Research Part F Traffic Psychology and Behaviour. 2012; 15(1):52–64. 330. Mcnabb J, Kuzel M, Gray R. I’ll Show You the Way : Risky Driver Behavior When “Following a Friend. Frontiers in Psychology; 2017; 8:1–6. 331. Can friendship serve as an impetus for safe driving among young drivers? [Internet]. Available from: https://www.sciencedirect.com/science/article/abs/pii/S1369847815000406 332. Steinberg L, Ann N Y. Risk taking in adolescence: what changes, and why? Academy of Sciences. 2004; 1021:51-8. 333. Fernandes R, Hatfield J, Soames RF. A systematic investigation of the differential predictors for speeding, drink-driving, driving while fatigued, and not wearing a seat belt, among young drivers. Transportation Research Part F Traffic Psychology and Behaviour. 2010;13:179–96. 334. Dayan J, Bernard A, Olliac B, Mailhes A, Kermarrec S. Journal of Physiology - Paris Adolescent brain development, risk-taking and vulnerability to addiction. Journal of Physiology. 2010; 104(5):279–86. 335. NHTSA, “A review of the literature on the effects of low doses of alcohol on driving-related skills,” Tech. Rep. HS-809 028, National Highway Traffic Safety Administration, Washington, DC, USA, 2000. 118  336. Marczinski CA, Harrison EL, Fillmore MT. Effects of alcohol on simulated driving and perceived driving impairment in binge drinkers. Alcoholism: Clinical and Experimental Research. 2008; 32:1329–1337. 337. Veldstra JL, Brookhuis KA, de Waard D. Effects of alcohol (BAC 0.5) and ecstasy (MDMA 100 mg) on simulated driving performance and traffic safety. Psychopharmacology (Berl) 2012; 222:377–390. 338. Arnedt A, Wilde W, Munt M, Maclean M. Simulated driving performance following prolonged wakefulness and alcohol consumption: separate and combined contributions to impairment. Journal of Sleep Research. 2000; 9(3):233-241. 339. Teen drivers: Get the facts. Centers for Disease Control and Prevention. Atlanta, 2016. Retrieved from https://www.cdc.gov/motorvehiclesafety/teen_drivers/teendrivers_factsheet.html 340. Rezaee-Zavareh MS, Salamati P, Ramezani-Binabaj M, et al. Alcohol consumption for simulated driving performance: A systematic review. Chinese Journal of Traumatology. 2017; 20(3):166-172.  341. Martin TL, Solbeck PA, Mayers DJ, Langille RM, Buczek Y, Pelletier MR. A Review of Alcohol-Impaired Driving: The Role of Blood Alcohol Concentration and Complexity of the Driving Task. Journal of Forensic Sciences. 2013;58:1238–1250 342. Marczinski CA, Fillmore MT. Acute alcohol tolerance on subjective intoxication and simulated driving performance in binge drinkers. Psychology of Addictive Behaviors. 2009; 23:238–247. 343. McCambridge J, McAlaney J, Rowe R. Adult Consequences of Late Adolescent Alcohol Consumption: A Systematic Review of Cohort Studies. PLoS Med. 2011;8(2): e1000  344. Drink driving penalties. Centre for Road Safety. 2019. Available from: https://roadsafety.transport.nsw.gov.au/stayingsafe/alcoholdrugs/drinkdriving/reforms/index.html 345. Alcohol and Adolescence. Available from: https://www.albertahealthservices.ca/assets/info/amh/if-amh-alcohol-and-adolescents.pdf 346. MacArthur GJ, Smith MC, Melotti R, et al. Patterns of alcohol use and multiple risk behaviour by gender during early and late adolescence: the ALSPAC cohort. J Public Health (Oxf). 2012;34 Suppl 1(Suppl 1):i20-i30.  347. Wells E, Horwood J, Fergusson DM. Drinking patterns in mid adolescence and psychosocial outcomes in late adolescence and early adulthood. Addiction. 2004;99(12):1529–41. 348. Compton RP, Berning A. Drug and Alcohol Crash Risk. Washington, DC: National Highway Traffic Safety Administration; 2015. DOT HA 812 117. 349. Moncaleano Juan David, Brands Bruna. PERCEPTION OF RISK AND DRIVING UNDER THE EFFECTS OF ALCOHOL AND MARIJUANA ON UNIVERSITY STUDENTS IN A MULTICENTER STUDY: COLOMBIA. Texto contexto - enferm.  [Internet]. 2019  [cited  2020  July  02] ;  28( spe ): e2428. Available from: http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0104-07072019000600325&lng=en.  Epub Aug 26, 2019.  https://doi.org/10.1590/1980-265x-tce-cicad-24-28. 350. Cannabis laws and regulations. Available from: https://www.canada.ca/en/health-canada/services/drugs medication/cannabis/laws-regulations.html 351. Elpus K. Arts Education and Positive Youth Development. Art Work arts.gov [Internet]. 2013; Available from: https://www.arts.gov/sites/default/files/Research-Art-Works-Maryland.pdf 352. McCamy TM. The Role of Youth Arts : Providing Opportunity and Intervention for At-Risk Populations. East Tennessee State University; 2014. 353. Stinnett CSW. Art for at-risk students: a study and recommendation for teaching art to at-risk students. The University Of Florida; 2012. 354. Review of the Roots of Youth Violence: Literature Reviews. Available from: http://www.children.gov.on.ca /htdocs/English/professionals/oyap/roots/volume5/chapter10_media_violence.a px 355. Reading health benefits [Internet]. Available from: https://doctor.ndtv.com/living-healthy/reading-health-benefits-reading-benefits-mental-health-reduces-depression-anxiety-1920782. 356. Berns GS, Blaine K, Prietula MJ, Pye BE. Short- and long-term effects of a novel on connectivity in the brain. Brain Connection. 2013; 3(6):590-600. 357. Rizzolo Z, Zipp R, Genevieve S, Simpkins G, Susan S, Stiskal D, Doreen S. Stress Management Strategies For Students: The Immediate Effects Of Yoga, Humor, And Reading On Stress. Journal of College Teaching and Learning. 2009; 6:79-88.  358. Losing Yourself in a Fictional Character Can Affect Your Real Life, Press Release, The Ohio State University, Research and Innovation Communications https://letterpile.com/personal-essays/The-Health-Benefits-of-Reading-and-Writing 119  359. William T. O'Donohue, Nicholas A. Bibliotherapy as an adjunctive treatment: read all about it in practical Resources for the Mental Health Professional. Evidence-Based Adjunctive Treatments. Academic Press, 2008:7-39. 360. Olsen MA. Bibliotherapy: School Psychologists' Report of Use and Efficacy Brigham Young University. Dissertations; 2015. P.P:78-79. 361. Stevens MR. The Forgotten Medium : The Impact of Reading Violent Literature on Aggressive Thoughts. Brigham Young University; 2012. 362. Pennebaker JW, Beall SK. Confronting a traumatic event: toward an understanding of inhibition and disease. Journal of Abnormal Psychology. 1986; 95(3):274-81. 363. Frattaroli J. Experimental disclosure and its moderators: a meta-analysis. Psychology Bulletin. 2006; 132(6):823-65. 364. Volunteering and Helping Others Reduces Risk of Substance Use in Rural Adolescents. 2011. Available at: https://drugfree.org/learn/drug-and-alcohol-news/volunteering-and-helping-others-reduces-risk-of-substance-use-in-rural-adolescents/ 365. Volunteering Reduces Risk of Teen Addiction. 2020: Available at: https://www.teensavers.com/post/volunteering-reduces-risk-of-teen-addiction 366. Cronje EM. The Effects of Volunteering on Student Volunteers. Progress South African Journal Open Distance Learn Practice. 2015; 37(2):1–20. 367. Ranapurwala SI, Casteel C, Peek-Asa C. Volunteering in adolescence and young adulthood crime involvement: a longitudinal analysis from the add health study. International journal of Epidemiology. 2016; 3(1):312-321. 368. Latham M. Young Volunteers : The Benefits of Community Service. University of  Birgam; 2000; available at https://www.yumpu.com/en/document/view/27530657/young-volunteers-the-benefits-of-community-service-university 369. Brannigan A, Gemmell W, Pevalin D, Wade T. Self-control and social control in childhood misconduct and aggression: The role of family structure, hyperactivity, and hostile parenting. Canadian Journal of Criminology. 2002; 44(2):119−142. 370. Grönlund H. Identity and Volunteering Intertwined: Reflections on the Values of Young Adults. International Journal of Voluntary and Nonprofit Organizations. 2011;11(4):852–874.  371. Nur H, Azizan A, Ismi AI. Theoretical Framework of Predictors Volunteering Behavior. International Journal of Academic Research in Business and Social Sciences. 2017; Special Issue - 4th International Conference on Educational Research and Practice 2017. 372. Brodskey W, Slor Z. Background music as a risk factor for distraction among young-novice drivers. Accident Analaysis & Prevention. 2013; 59:382–93.  373. Brodsky W, Olivieri D, Chekaluk E. Music genre induced driver aggression. Music Science. 2018; 1: 201-206. 374. Unal AB.Please dont stop the music: The influence of music and radio on cognitive processes, arousal and driving performance. Groningen, The Netherlands: University of Groningen; 2013. 375. Unal AB, Waard D, Epstude K, Steg L. Driving with music: Effects on arousal and performance. Transportation Research, Part F: Traffic Psychology and Behaviour. 2013; 21:52–65. 376. Unal AB, Steg L, Epstude K. The influence of music on mental effort and driving performance. Accident Analysis and Prevention. 2012; 48: 271–278. 377. Wiley-Blackwell. "Risky Behaviors On TV May Be Modeled By Inexperienced Viewers." ScienceDaily. ScienceDaily, 29 September 2008. 378. Akintayo J. Western Entertainment Television Programmes: A Catalyst for Behaviourial Tendencies among Students of Babcock and Covenant Universities New Media and Mass. 2015;42: 211-216. 379. Achila, Isabelle Achieng(2007). Influence of Television on sexual behavior : The case of Nairobi Youth. URI: http://erepository.uonbi.ac.ke:8080/xmlui/handle/123456789/20605  380. Baya SY. The Impact of Television Viewing in Influencing Adolescents Sexual Behaviour. International Journal of Scientific and Research Publications. 2011;4(5): 126-132. ISSN 2250-315. 381. Zilka G, Romi S. Viewing Habits and Identification with Television Characters among At-Risk and Normative Children and Adolescents. Int J Child, Youth Fam Stud. 2018; 9(3):47. 382. Rimal RN, Real K. How Behaviors are Influenced by Perceived Norms: A Test of the Theory of Normative Social Behavior. Communication Research. 2005; 32(3), 389–414. 383. Pajares F. Social Cognitive Theory and Mass Media Effects. Wiliam & Mary school work; 2009. 384. Buss D. Evolutionary Psychology: The New Science of the Mind. Psychology. 2015. 120  385. Weaver J. Exploring the links between personality and media preferences. Personality & Individual Differences. 1991; 12(12):1293–9. 386. What is Non Response Bias? [Internet]. Available from: https://www.statisticshowto.datasciencecentral. com/ non-response-bias/ 387. Survey sampling bias [Internet]. Available from: https://stattrek.com/survey-research/survey-bias.aspx 388. Satisficing bias [Internet]. Available from:https://www.surveymonkey.com/curiosity/satisficing-learn-to-defeat-the-subtle-menace-in-your-survey-data/                             121  Appendix A Table A.1 Distribution of driving license classes in driving behaviour categories *  Variables Group All participants  n=964 Low Risk** n=200 Moderate Risk** n=312 High Risk** n=452 DL (AB) Class 7 (Learner) Class 5 (GDL) Class 5 (Full)* Other  12 (41.1%) 14 (48.3%) 2 (6.9%) 1 (3.4%)  5 (6.5%) 63 (81.8%) 5 (6.5%) 4 (6.2%)  4 (3.3%) 92 (76%) 23 (19%) 2 (1.7%)  26 (8.4%) 228 (73.8%) 45 (14.6%) 10 (3.2%) DL (BC) Class 7 (Learner) Class 7 (Novice) Class 5 (Full)* Other  42 (71.2%) 14 (23.7%) 3 (5.1%) 0  13 (15.7%) 44 (53%) 24 (28.9%) 2 (2.4%)  3 (2.6%) 78 (66.7%) 31 (26.5%) 5 (4.3%)  77 (21.8%) 191 (54.1%) 77 (21.8%) 8 (2.3%) DL (NB) Class 7 (Level 1) Class 7 (Level 2) Class 5 (Full)* Other  2 (66.7%) 1 (33.3%) 0 0  1 (12.5%) 3 (37.5%) 4 (50%) 0  5 (45.5%) 0 6 (54.5%) 0  4 (13.3%) 13 (43.3%) 12 (40%) 1 (3.3%) DL (NF) Class 5 (Level1) Class 5 (Level2) Class 5 (Full)*  1 (50%) 0 1 (50%)  0 0 3 (100%)  6 (46.2%) 0 7 (53.8%)  3 (12%) 7 (28%) 15 (60%) DL (NT) Class 5 (Full)*  0  0  1 (100%)  1 (100%) DL (MB) Class 5 L (Learner) Class 5 I (Intermediate) Class 5 (Full)* Other  2 (22.2%) 5 (55.6%) 2 (22.2%) 0  4 (21.1%) 9 (47.4%) 4 (21.1%) 2 (10.5%)  0 12 (63.2%) 7 (36.8%) 0  8 (13.3%) 36 (60%) 14 (23.3%) 2 (3.3%) DL (NS) Class 7 (Learner) Class 5 N Class 5 R Class 5 (Full)* Other  1 (9.1%) 7 (63.6%) 1 (9.1%) 1 (9.1%) 1 (9.1%)  0 9 (75%) 1 (8.3%) 2 (16.7%) 0  1 (5.6%) 6 (33.3%) 6 (33.3%) 3 (16.7%) 2 (11.1%)  5 (8.8%) 29 (50.9%) 11 (19.3%) 9 (15.8%) 3 (5.3%) DL (ON) G1 (Learner's Permit) G2 (Probationary license) G (Full License) * Other  55 (57.9%) 29 (30.5%) 8 (8.4%) 3 (3.2%)  14 (13.9%) 54 (53.5%) 32 (31.7%) 1 (1%)  13 (8.7%) 64 (43%) 69 (46.3%) 3 (2%)  116 (23.1%) 221 (44%) 157 (31.3%) 8 (1.6%) DL (NU) Class 5 (Full)*  0  0  0  1 (100%) DL (PE) Class 5G (Stage 2) Class 5G (Stage 3) Class 5 (Full) *  0 0 0  1 (50%) 0 1 (50%)  2 (40%) 3 (60%) 0  4 (44.4%) 1 (11.1%) 4 (44.4%) DL (QC) Class 5 (Learner stage) Class 5 (Probationary) Class 5 (Full) * Other  2 (50%) 1 (25%) 1 (25%) 0  1 (12.5%) 4 (50%) 3 (37.5%) 0  0 7 (63.6%) 3 (27.3%) 1 (9.1%)  3 (9.7%) 14 (45.2%) 13 (41.9%) 1 (3.2%) DL (SK) Class 7 (Learner) Class 7 (Novice 1) Class 7 (Novice 2) Class 5 (Full) * Other  3 (23.1%) 1 (7.7%) 4 (30.8%) 5 (38.5%) 0  2 (7.4%) 3 (11.1%) 11 (40.7%) 10 (37%) 1 (3.7%)  0 5 (13.9%) 11 (30.6%) 19 (52.8%) 1 (2.8%)  6 (6.5%) 13 (14.1%) 29 (31.5%) 42 (45.7%) 2 (2.2%) DL (YT) Class 5- Full * Other  0 0  0 1 (100%)  1 (100%) 0  3 (75%) 1 (25%) Total time of driving (hours per week) 4.50±9.88 9.79±16.42 16.75±31.41 12.53±24.35 * Class 5 (Full) license is known as a license without restriction, while other classes are licenses with restrictions. ** Total number in our sample: Low risk (n=200), Mod risk (n=312), high risk (n=452); however, in this table, numbers in the respective  columns are specific for each province *** All participant in our sample (n=964); however, in this table, this column shows the total number of participants in each province.    122      Table A.2 Leisure activity subtypes * Variable Subtypes Video game type -Racing games (e.g. Need for Speed™) -Action games (e.g. First-person shooter, GTA™) -Adventure (e.g. Fable™) -Simulation games (e.g. Sims™) -Strategy (e.g. Strong Hold™) -Sports (e.g. FIFA™) -Other Sport type -Collision (hockey, football, martial arts, etc.)  -Contact (soccer, basketball, etc.)  -Limited contact (baseball, volleyball, etc.)  -No contact (fitness, table tennis, etc.) Social media type -YouTube -Facebook -Twitter -Snapchat -Instagram -Other (Reddit, WhatsApp, LinkedIn, Pinterest, etc.) Movie type -Action, Adventure -Horror, Thriller, Mystery -Romance, Drama, Musical -Fantasy, Sci-Fi, Animation -Comedy -Biography, Documentary, History -Other Family activity type -Entertainment (e.g. concerts, games) -Physical recreation (e.g. hiking) -Social events (e.g. parties) -Spiritual activities (e.g. going to church) -Indoor (e.g. family gathering) Friend activity type -Entertainment (e.g. movies, games) -Physical activity (e.g. Sports) -Social activity (e.g. parties) -Spiritual activity (e.g. going to church) -Unstructured activity (e.g. Hanging out) Alcohol type -Wine -Beer, coolers, mixed drinks -Soft liquor (Under 20% alcohol) -Hard liquor (Greater than or equal to 20% alcohol) Drug type -Marijuana (weed, pot) -Opioid (Heroin, Fentanyl) -Stimulants (cocaine, crack, crystal, meth) -Psychedelics (hallucinogens e.g. LSD, mushroom) -Prescription drugs (other than as prescribed) -Other Art type -Visual arts (e.g. drawing, photography, architecture) -Performing arts (e.g. Music, Dancing, Theatre) -Crafting arts (e.g. Pottery, Sculpture) - Applied arts (e.g. Fashion/interior designing) - Other Writing type -Fiction (romance, mystery, sci-) -Non-fiction (poetry, history) -Articles/news -Philosophy (religion, self-help) -Other Volunteering type -Health (e.g. hospitals) -Art and culture (e.g. art galleries) -Sport and recreation (e.g. swimming pools) -Education and research (e.g. school programs) -Community and social services (e.g. community centers) -Religion (e.g. places of worship) -Other Music type -Classical -Pop, Rock, Alternative, Heavy metal -Jazz, Blues -Traditional, Country, Folk, Ethnic -Rap, Electronic -Other TV type -Animation, Cartoon, Sci-Fi -Comedy, Stand-up comedy, Sitcom, Prank -Documentary, Surviving show, Reality show -Sport, Game, Quiz show -News, Talk-show -Action, Adventure -Horror, Thriller, Mystery -Romance, Drama, Musical -Other or Combined    *The participants were asked about their first, second and third choices of each leisure activity subtypes  123   Table A.3 Distribution of study population according to Leisure activity (Sub type)  Variables Group P-value*  All participants (n=964) Low Risk  (n= 200)  Moderate Risk  (n= 312)  High Risk  (n= 452)  Video game engagement level Low  High   171 (85.5%) 29 (14.5%)  244 (78.2%) 68 (21.8%)  383 (84.7%) 69 (15.3%) 0.0329  798 (82.8%) 166 (17.2%) Video game genre** Racing  Action  Adventure Simulation  Strategy  Sports  Other  12 (10.0%) 49 (40.8%) 18 (15.0%) 16 (13.3%) 19 (15.8%) 5 (4.2%) 1 (0.8%)  22 (11.0%) 99 (49.5%) 24 (12.0%) 23 (11.5%) 22 (11.0%) 10 (5%) 0 (0.0%)  57 (18.2%) 169 (54%) 20 (6.4%) 32 (10.2%) 21 (6.7%) 11 (3.5%) 3 (1.0%)   91 (14.4%) 317 (50.1%) 62 (9.8%) 71 (11.2%) 62 (9.8%) 26 (4.1%) 4 (0.6%) Sport engagement level Low High  140 (70.0%) 60 (30.0%)  210 (67.3%) 102 (32.7%)  311 (68.8%) 141 (31.2%) 0.8057  661 (68.6%) 303 (31.4%) Type of sports** Collision Contact Limited contact No contact  15 (11.2%) 29 (21.6%) 24 (17.9%) 66 (49.3%)  57 (27.9%) 44 (21.6%) 25 (12.3%) 78 (38.2%)  100 (32.9%) 48 (15.8%) 41 (13.5%) 115 (37.8%)   172 (26.8%) 121 (18.8%) 90 (14.0%) 259 (40.4%) Social media engagement level Low High  64 (32.0%) 136 (68.0%)  98 (31.4%) 214 (68.6%)  106 (23.5%) 346 (76.5%) 0.0179  268 (27.8%) 696 (72.2%) Social Media Platform** YouTube Facebook Twitter Snapchat Instagram Other  37 (18.5%) 23 (11.5%) 3 (1.5%) 57 (28.5%) 68 (34.0%) 12 (6.0%)  78 (25.1%) 58 (18.6%) 2 (0.6%) 72 (23.2%) 85 (27.3%) 16 (5.1%)  97 (21.6%) 81 (18%) 9 (2.0%) 142 (31.6%) 101 (22.4%) 20 (4.4%)   212 (22.1%) 162 (16.9%) 14 (1.5%) 271 (28.2%) 254 (26.4%) 48 (4.9%) Movie engagement level Low High  186 (93%) 14 (7%)  267 (85.6%) 45 (14.4%)  383 (84.7%) 69 (15.3%) 0.0126  836 (86.7%) 128 (13.3%) Movie genre** Action, Adventure Horror, Thriller, Mystery Romance, Drama, Musical Fantasy, Sci-Fi, Animation Comedy Biography, Documentary, History Other  54 (32.0%) 17 (10.1%) 27 (16.0%) 17 (10.1%) 39 (23.1%) 10 (5.9%) 5 (3.0%)  111 (39.4%) 30 (10.6%) 35 (12.4%) 34 (12.1%) 58 (20.6%) 12 (4.3%) 2 (0.7%)  159 (41.0%) 37 (9.5%) 42 (10.8%) 22 (5.7%) 101 (26.0%) 26 (6.7%) 1 (0.3%)   324 (38.6%) 84 (10.0 %) 104 (12.4%) 73 (8.7%) 198 (23.6%) 48 (5.7%) 8 (1.0%) Family engagement level*** Low High  159 (79.5%) 41 (20.5%)  289 (84.4%) 52 (15.2%)  396 (87.6%) 56 (12.4%) 0.0275  820 (85.1%) 144 (14.9%) Family event** Entertainment Physical recreation Social events Spiritual activities Indoor activities  56 (32.7%) 22 (12.9%) 14 (8.2%) 15 (8.8%) 64 (37.4%)  84 (32.4%) 34 (13.1%) 44 (17%) 16 (6.2%) 81 (31.3%)  125 (35.6%) 36 (10.3%) 72 (20.5%) 18 (5.1%) 100 (28.5%)   265 (33.9%) 92 (11.8%) 130 (16.6%) 49 (6.3%) 245 (31.4%) Friends engagement level**** Low High  150 (75.0%) 50 (25.0%)  228 (73.1%) 84 (26.9)  241 (53.3%) 211 (46.7%) <0.0001  619 (64.2%) 345 (35.8%) Friends event** Entertainment Physical recreation Social events Spiritual activities Unstructured activities  38 (20.8%) 16 (8.7%) 22 (12.0%) 4 (2.2%) 103 (56.3%)  45 (15.3%) 41 (13.9%) 59 (20.1%) 6 (2.0%) 143 (48.6%)  80 (18.3%) 52 (11.9%) 99 (22.7%) 0 (0.0%) 205 (47.0%)   163 (17.9%) 109 (11.9%) 180 (19.7%) 10 (1.1%) 451 (49.4 %) 124  Alcohol engagement level Low High  198 (99.0%) 2(1.0%)  303 (97.1%) 9 (2.9%)  416 (92.0%) 36 (8.0%) <0.0001  917 (95.1%) 47 (4.9%) Type of alcohol drinks** Wine Bear, coolers, mixed drinks Soft liquor Hard liquor  8 (8.9%) 46 (51.1%) 14 (15.6%) 22 (24.4%)  18 (9.3%) 115 (59.3%) 8 (4.1%) 53 (27.3%)  15 (4.1%) 212 (57.9%) 15 (4.1%) 124 (33.9%)   41 (6.3%) 373 (57.4%) 37 (5.7%) 199 (30.6%) Drug engagement level Low High  195 (97.5%) 5 (2.5%)  294 (97.1%) 18 (2.9%)  380 (84.1%) 72 (15.9%) <0.0001  869 (90.1%) 95 (9.9%) Type of drug** Marijuana Stimulants Psychedelics Prescription  Other  28 (93.3%) 2 (6.7%) 0 (0.0%) 0 (0.0%) 0 (0.0%)  65 (97%) 0 (0.0%) 1 (1.5%) 1 (1.5%) 0 (0.0%)  173 (89.2%) 11 (5.7%) 4 (2.1%) 6 (3.1%) 0 (0.0%)   266 (91.4%) 13 (4.5%) 5 (1.7%) 7 (2.4%) 0 (0.0%) Art engagement level Low High  159 (79.5%) 41 (20.5%)  262 (84%) 50 (16%)  452 (100%) 0 (0.0%) 0.0159  873 (90.6%) 91 (9.4%) Type of art** Visual  Performing  Crafting  Applied arts Other  49 (40.8%) 61 (50.8%) 7 (5.8%) 2 (1.7%) 1 (0.8%)  76 (47.2%) 64 (39.8%) 12 (7.5%) 8 (5.0%) 1 (0.6%)  62 (38%) 75 (46%) 15 (9.2%) 11 (6.7%) 0 (0.0%)   187 (42.1%) 200 (45.0%) 34 (7.7%) 21 (4.7%) 2 (0.5%) Writing/Reading engagement level Low High  152 (76.0%) 48 (24.0%)  240 (76.9%) 72 (23.1%)  401 (88.7%) 51 (11.3%) <0.0001  793 (82.3%) 171 (17.7%) Writing/Reading genre** Fiction  Non-fiction  Articles/news  Philosophy  Other  75 (52.8%) 20 (14.1%) 34 (23.9%) 11 (7.7%) 2 (1.4%)  90 (39.3%) 42 (18.3%) 74 (32.3%) 16 (7.0%) 7 (3.1%)  96 (38.2%) 35 (13.9%) 91 (36.3%) 22 (8.8%) 7 (2.8%)   261 (42.0%) 97 (15.6%) 199 (32.0%) 49 (7.9%) 16 (2.5%) Volunteering engagement level Low High  183 (91.5%) 17 (8.5%)  279 (89.4%) 33 (10.6%)  428 (94.7%) 24 (5.3%) 0.0239  890 (92.3%) 74 (7.7%) Type of volunteering activities** Health Art and culture Sport and recreation Education and research Community and social services Religion Other  13 (13.1%) 6 (6.1%) 10 (10.1%) 19 (19.2%) 39 (39.4%) 12 (12.1%) 0  17 (11.8%) 4 (2.8%) 26 (18.1%) 25 (17.4%) 57 (39.6%) 15 (10.4%) 0  17 (13.1%) 4 (3.1%) 35 (26.9%) 14 (10.8%) 46 (35.4%) 12 (9.2%) 2 (1.5%)   47 (12.6%) 14 (3.8%) 71 (19.0%) 58 (15.5%) 142 (38.1%) 39 (10.5%) 2 (0.5%) Music engagement level Low High  125(62.5%) 75 (37.5%)  207 (66.3%) 105 (33.7%)  247 (54.6%) 205(45.4%) 0.0038  579 (60.1%) 385 (39.9%) Type of music** Classical  Pop, Alternative, Rock, Heavy metal  Jazz, Blues  Traditional, Folk, Country, Ethnic  Rap, Electronic  Other  7 (3.6%) 112 (58%) 4 (2.1%) 10 (5.2%) 48 (24.9%) 12 (6.2%)  5 (1.6%) 160 (51.9%) 3 (1.0%) 42 (13.6%) 94 (30.5%) 4 (1.3%)  2 (0.4%) 191 (42.8%) 2 (0.4%) 65 (14.6%) 181 (40.6%) 5 (1.1%)   14 (1.5%) 463 (48.9%) 9 (1.0%) 117 (12.3%) 323 (34.1%) 21 (2.2%) TV engagement level Low High  156 (78.0%) 44 (22.0%)  237 (76.0%) 75 (24.0%)  333 (73.7%) 119 (26.3%) 0.472  726 (75.3%) 238 (24.7%) Type of TV show** Animation, Cartoon  Comedy, Stand-up comedy, Sitcom, Prank  Documentary, Surviving show, Reality show  Sport, Game, Quiz show  News, Talk show   9 (5.8%) 68 (44.2%) 39 (25.3%) 11 (7.1%) 6 (3.9%)  31 (13.3%) 104 (43.9%) 49 (20.7%) 22 (9.3%) 14 (5.9%)  39 (11.1%) 160 (45.7%) 80 (22.9%) 29 (8.3%) 20 (5.7%)   79 (10.7%) 332 (44.8%) 168 (22.7%) 62 (8.4 %) 40 (5.4%) 125  Other 21 (13.6%) 17 (7.2%) 22 (6.3%) 60 (8.0%) * Based on qui square ** The sum of the numbers in each cell may not be consistent with the total number in that category (low, moderate, high and all participants).  This is due to the fact that the questions concerning subtypes of leisure activity were optional and participants had right to skip that question. *** Engagement levels did not include time spent with family at home doing their own tasks **** Engagement levels did not include the activities done with friends during school time  Table A.4 BYNDS subscales’ scores in driving behaviour groups Subscales  Score range (Min-Max) Score (Mean±SD)  All participants (n=964) Low Risk (n=200)  Moderate Risk (n=312)  High Risk (n=452)  P-value* Transient violations (13-65) 30.8 ± 11.2 17.2 ± 3.5 26.5 ± 5.0 39.8 ± 8.2 <0.000001 Fixed violations (10-50) 12.8 ± 4.1 10.4 ± 0.9 11.1 ± 1.7 15.0 ± 4.9 <0.000001 Misjudgement (9-45) 12.9 ± 3.5 11.2 ± 2.2 12.2 ± 2.5 14.1 ± 4.0 <0.000001 Risky driving exposure (9-45) 28.9 ± 8.8 17.0 ± 5.1 27.4 ± 5.5 35.1 ± 5.4 <0.000001 Driving mood (3-15) 5.6 ± 2.8 3.6 ± 1.1 4.7 ± 1.8 7.1 ± 3.1 <0.000001 Total BYNDS Score 90.9 ± 23.9 59.4 ± 8.2 81.8 ± 5.8 111.2 ± 16.0 <0.000001 *P-value shows significant Mean difference in each subscale by all comparisons in the groups (High risk vs. Low risk, High risk vs. Moderate risk, Moderate risk vs. Low risk)                                 126  Appendix B Table B.1 Bivariate correlation Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 1 Age 1                            2 Gender -0.156 1                           3 Employment Status    1                          4 Living Status 0.401 -0.062  1                         5 Education Status 0.603 -0.12  0.359 1                        6 License Class 0.648 -0.18  0.312 0.57 1                       7 Driving Independently 0.669 -0.197  0.351 0.669 0.687 1                      8 Driving Exposure 0.14 -0.169  0.053 0.187 0.18 0.197 1                     9 Car Status 0.221 -0.216  0.14 0.248 0.259 0.374 0.276 1                    10 SES -0.095 0.011  -0.257 -0.013 -0.024 -0.015 -0.009 0.042 1                   11 Extroversion -0.039 -0.023  -0.047 -0.02 0.024 -0 0.064 0.12 0.082 1                  12 Agreeableness -0.109 0.213  -0.051 -0.032 -0.072 -0.099 -0.045 -0.064 -0.016 0.239 1                 13 Conscientiousness 0.061 -0.014  0.02 0.1 0.121 0.104 0.028 0.053 0.051 0.048 0.04 1                14 Neuroticism -0.09 0.276  -0.008 -0.128 -0.089 -0.125 -0.089 -0.103 -0.036 -0.08 0.013 -0.137 1               15 Imagination 0.034 -0.051  0.042 0.069 0.034 0.029 -0.021 0.009 -0.008 0.103 0.126 0.001 -0.122 1              16 Video Game 0.046 -0.221  0.053 0.015 0.014 0.019 0.087 -0.021 -0.035 -0.105 -0.046 -0.149 -0.056 0.034 1             127  17 Sport -0.124 -0.026  -0.07 -0.165 -0.099 -0.134 -0.03 -0.052 0.081 0.139 0.025 0.016 -0.107 0.038 -0.013 1            18 Social Media -0.081 0.17  -0.052 -0.038 -0.105 -0.106 0.007 -0.05 0.018 0.047 0.06 -0.084 0.139 -0.025 0.057 -0.032 1           19 Movie 0.101 0.08  0.038 0.078 0.038 0.084 0.107 0.024 -0.032 -0.012 0.047 -0.03 0.014 -0.046 0.012 -0.035 0.126 1          20 Family -0.018 0.025  -0.097 -0.056 -0.032 -0.051 0.025 -0.034 0.022 0.028 0.091 0.011 -0.054 0.024 -0.036 0.033 -0.003 0.152 1         21 Friend -0.07 -0.071  -0.006 0.011 -0.05 0.013 0.118 0.092 0.072 0.193 0.074 -0.048 -0.042 0.058 0.034 0.104 0.083 -0.007 0.014 1        22 Alcohol 0.126 -0.151  0.146 0.132 0.125 0.159 0.084 0.122 0.003 0.068 -0.078 -0.017 -0.035 0.001 0.005 -0.051 -0.037 0.052 -0.019 0.082 1       23 Drug 0.085 -0.123  0.109 0.096 0.094 0.153 0.153 0.136 -0.001 0.021 -0.044 -0.014 -0.008 0.042 0.077 -0.08 0.019 0.042 -0.062 0.155 0.061 1      24 Art -0.109 0.123  -0.063 -0.145 -0.132 -0.119 -0.087 -0.104 0.018 0.003 0.071 -0.114 0.013 0.16 -0.05 0.063 0.039 -0.062 0.004 0.051 -0.051 0.014 1     25 Reading/Writing 0  0.083  0.04 -0.024 -0.02 -0.03 -0.039 -0.088 -0.04 -0.038 0.013 0.041 0.046 0.134 0.011 0.069 -0.057 0.033 0.08 0.029 -0.055 -0.021 0.125 1    26 Volunteering -0.014 0.035  -0.071 0.002 0.024 -0.006 -0.013 -0.055 0.006 0.04 0.047 0.037 -0.002 -0.006 -0.023 0.04 0.035 0.022 0.098 0.053 -0.036 -0.047 0.034 0.094 1   27 Music -0.01 -0.011  0.075 -0.046 -0.007 -0.033 0.158 0.021 -0.028 0.069 0.023 -0.114 0.026 0.078 0.085 0.082 0.094 0.089 0.023 0.194 0.073 0.085 0.082 0.078 -0.025 1  28 TV 0.039 0.117  0.011 0.04 0.015 0.038 0.057 -0.042 0.023 -0.017 0.075 -0.095 0.082 -0.026 0.004 -0.037 0.136 0.356 0.097 -0.001 0.018 0.035 0.047 -0.013 0.01 0.001 1                                     128  Table B.2 Full model variable (CI=95%)  Variable est SE t p-value OR OR (low) OR (high) Driver Car Status (Yes) 1.138 0.16 7.123 0 3.121 2.284 4.275 Driving Exposure (High) 0.91 0.271 3.359 0.001 2.486 1.476 4.282 Gender (Male) 0.904 0.163 5.549 0 2.47 1.797 3.405 Driving Exposure (Moderate) 0.892 0.224 3.985 0 2.439 1.584 3.813 Drug Engagement Level (High engagement) 0.803 0.288 2.79 0.005 2.231 1.289 3.996 Friend Engagement Level (High engagement) 0.665 0.158 4.215 0 1.944 1.429 2.653 Duration of driving independently >3 years 0.658 0.291 2.259 0.024 1.931 1.092 3.426 Social Media Engagement Level (High engagement) 0.598 0.159 3.767 0 1.819 1.333 2.485 Neuroticism Trait (High) 0.58 0.206 2.811 0.005 1.786 1.195 2.683 Employment status (Both Employed and student) 0.442 0.16 2.757 0.006 1.556 1.136 2.131 Education Status (Less than High school) -0.437 0.215 -2.034 0.042 0.646 0.423 0.983 Reading/Writing Engagement Level (High engagement) -0.478 0.179 -2.673 0.008 0.62 0.436 0.88 Volunteering Engagement Level (High engagement) -0.51 0.25 -2.04 0.041 0.6 0.367 0.981 Video Game Engagement Level (High engagement) -0.577 0.192 -3.009 0.003 0.561 0.385 0.818 Duration of driving independently <1 year -0.581 0.19 -3.061 0.002 0.559 0.385 0.811 Orange: OR>1 (P-value < 0.05) Green: OR <1 (P-Value <0.05) 129  Table B.3 Stepwise AIC selection model (CI=95%)  Variable Estimation SE t p-value OR OR (low) OR (high) Driver Car Status (Yes) 1.103 0.158 6.987 0 3.014 2.214 4.113 Driving Exposure  (High) 0.949 0.267 3.549 0 2.583 1.545 4.418 Driving Exposure  (Moderate) 0.942 0.218 4.327 0 2.566 1.687 3.967 Gender (Male) 0.922 0.157 5.891 0 2.515 1.853 3.425 Duration of driving independently >3 years 0.748 0.266 2.813 0.005 2.114 1.257 3.571 Drug Engagement Level (High engagement) 0.738 0.284 2.602 0.009 2.091 1.219 3.718 Friend Engagement Level (High engagement) 0.684 0.154 4.429 0 1.981 1.466 2.686 Social Media Engagement Level (High engagement) 0.606 0.157 3.871 0 1.833 1.349 2.494 Neuroticism Trait (High) 0.604 0.204 2.966 0.003 1.83 1.231 2.737 Extroversion Trait (High) 0.47 0.196 2.402 0.016 1.6 1.094 2.358 Employment status (Both employed and student) 0.461 0.158 2.917 0.004 1.585 1.163 2.16 Imagination Trait (Low) 0.43 0.212 2.023 0.043 1.537 1.017 2.341 Movie Engagement Level (High engagement) 0.418 0.212 1.977 0.048 1.52 1.007 2.312 Education Status (Less than High school) -0.441 0.21 -2.097 0.036 0.643 0.425 0.971 Reading/Writing Engagement Level (High engagement) -0.5 0.176 -2.835 0.005 0.606 0.429 0.857 Volunteering Engagement Level (High engagement) -0.518 0.247 -2.093 0.036 0.596 0.367 0.968 Video Game Engagement Level (High engagement) -0.58 0.189 -3.062 0.002 0.56 0.386 0.812 Duration of driving independently <1 year -0.6 0.185 -3.248 0.001 0.549 0.382 0.788 Orange: OR>1 (P-value < 0.05) Green: OR <1 (P-Value <0.05)   130  Table B.4 Full versus Stepwise selection Model  Variable Full model  Stepwise AIC OR Lower Upper p-value OR Lower Upper p-value Gender (Male) 2.47 1.797 3.405 0  2.515 1.853 3.425 0 Education Status (Less than high school) 0.646 0.423 0.983 0.042  0.643 0.425 0.971 0.036 Duration of driving independently <1 year 0.559 0.385 0.811 0.002  0.549 0.382 0.788 0.001 Duration of driving independently >3 years 1.931 1.092 3.426 0.024  2.114 1.257 3.571 0.005 Driver Car Status (Yes) 3.121 2.284 4.275 0  3.014 2.214 4.113 0 Driving Exposure (High) 2.486 1.476 4.282 0.001  2.583 1.545 4.418 0 Driving Exposure (Moderate) 2.439 1.584 3.813 0  2.566 1.687 3.967 0 Employment status (Both Employed and student) 1.556 1.136 2.131 0.006  1.585 1.163 2.16 0.004 Video Game Engagement Level (High engagement) 0.561 0.385 0.818 0.003  0.56 0.386 0.812 0.002 Social Media Engagement Level (High engagement) 1.819 1.333 2.485 0  1.833 1.349 2.494 0 Movie Engagement Level High engagement 1.396 0.902 2.173 0.137  1.52 1.007 2.312 0.048 Friend Engagement Level (High engagement) 1.944 1.429 2.653 0  1.981 1.466 2.686 0 Drug Engagement Level (High engagement) 2.231 1.289 3.996 0.005  2.091 1.219 3.718 0.009 Reading/Writing Engagement Level (High engagement) 0.62 0.436 0.88 0.008  0.606 0.429 0.857 0.005 Volunteering Engagement Level (High engagement) 0.6 0.367 0.981 0.041  0.596 0.367 0.968 0.036 Neuroticism Trait (High) 1.786 1.195 2.683 0.005  1.83 1.231 2.737 0.003 Imagination Trait (Low) 1.474 0.969 2.258 0.072  1.537 1.017 2.341 0.043 Extroversion Trait (High) 1.48 0.997 2.211 0.053  1.6 1.094 2.358 0.016 Orange: significant variables in both models Blue: significant variables in just stepwise AIC        131    Appendix C  Figure C.1 Percentage Distribution of Social Media Subtypes in drivers’ groups 11.518.6181.50.6228.523.231.63427.322.465.14.40% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Low RiskModerate RiskHigh RiskSocial Media Subtypes EngagementFacebookTwitterSnapchatInstagramOther132     Figure C.2 Percentage Distribution of Family Activity Subtypes in drivers’ groups    32.732.435.612.913.110.38.21720.58.86.25.137.431.328.50% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Low RiskModerate RiskHigh RiskFamily Activity Subtypes EngagementEntertainmentPhysical RecreationSocial EventsSpiritual ActivitiesIndoor Activities133     Figure C.3 Percentage Distribution of Art Activities’ Subtypes in drivers’ groups  40.847.23850.839.8465.87.59.21.756.70.80.600% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Low RiskModerate RiskHigh RiskArt Subtypes EngagementVisualPerformingCraftingAppliedOther134  Figure C.4 Percentage Distribution of Sociodemographic Characteristics    90.58872.596.571.871.264.779.562.865.757.561.30 20 40 60 80 100 120Driving License With RestrictionBeginner DriverLiving with Both ParentsLow Driving ExposureSociodemographic CharacteristicsHigh RiskModerate RiskLow Risk135     4319.57857.752.644.273.276.531.90 10 20 30 40 50 60 70 80 90MaleOwning Their CarDriving Independently <1 ySociodemographic Characteristics High RiskModerate RiskLow Risk136  Figure C.5 Percentage Distribution of high level of Personality Traits in drivers’ groups     051015202530Low Risk Moderate Risk High RiskPersonality Traits Subtypes-High LevelExtraversionAgreeblenessConscientiousnessNeuroticismImagination137  Appendix D (Canada Young Driver Survey- Sample)   4@ CEB=86GE8764CBE:	"."%"!/5.(2*6&2526&9FFB6<4G<BA58GJ88A%8<FHE86G<I<GL4A7+<F>LE<I<A:8;4I<BHE<ALBHA:4A47<4AF"8??B#9LBH4E8HA78EG;84:8B9?<I8<A4A4744A7;4I847E<I<A:?<68AF8LBH4E8<AI<G87GBG4>8C4EG<AG;<FBA?<A8FHEI8L-;<FFGH7L<FC4EGB94@4FG8ENFG;8F<F<AG;88C4EG@8AGB9KC8E<@8AG4?&87<6<A8G;8.A<I8EF<GLB9E<G<F;B?H@5<40;L<FG;<FFHEI8L<@CBEG4AG&BGBEI8;<6?86E4F;8F&/F4E8G;8?847<A:64HF8B9@4=BE<A=HEL4A7784G;<ALBHA:4A47<4AF+<F>L7E<I<A:BA8B9G;8@4=BE64HF8FB964E6E4F;8F<FG;BH:;GGB586?BF8?L4FFB6<4G87J<G;?8<FHE8G<@846G<I<G<8F4A7J<G;C8EFBA4?<GLGE4<GF0;4G<FG;84<@B9G;8FHEI8L084E86BA7H6G<A:4A4ABAL@BHFBA?<A8FHEI8LB9LBHA:4A47<4A7E<I8EFGB8K4@<A8G;8E8?4G<BAF;<C58GJ88A?8<FHE8G<@846G<I<G<8FC8EFBA4?<GLGE4<GF4A77E<I<A:58;4I<BE0;4G4E8G;8DH8FG<BAF45BHG-;8FHEI8LDH8FG<BAF4F>45BHG7<998E8AG4FC86GFB9LBHEC8EFBA4?<GL7E<I<A:;45<GF4A7?8<FHE8G<@846G<I<G<8F-;8E8<FABE<:;GBEJEBA:4AFJ8E)?84F84AFJ8E4F58FG4FLBH64A"BJ?BA:J<??G;8FHEI8LG4>8	@<AHG8F0;4G;4CC8AF<9#JBH?7?<>8GB?84I8G;8FHEI8L)4EG<6<C4G<BA<AG;<FFHEI8L<F6B@C?8G8?LIB?HAG4EL1BH64A?84I8G;8FHEI8L4G4ALG<@8;BJ8I8EBA68LBHF4I8LBHEE8FCBAF8FJ8;4I8ABJ4LGBE8@BI8G;8@??<A9BE@4G<BAJ<??58>8CGFGE<6G?L6BA9<78AG<4?4A764AABG58?<A>87546>GBLBHC8EFBA4??L138    139   140  141  142  143  144  145  146  147  148  149  150  151  152   

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0395576/manifest

Comment

Related Items