The psychosocial characteristics and on-road behaviour of ...nature of unlicensed driving, in order to develop more effective countermeasures to the behaviour. Study One utilisedofficial

The psychosocial characteristics and on-road behaviour of

unlicensed drivers

Barry Watson

Bachelor of Arts (Honours), GradDip (SciSoc)

A thesis submitted as fulfillment for the Degree of Doctor of Philosophy

Queensland University of Technology

Centre for Accident Research and Road Safety – Queensland (CARRS-Q)

School of Psychology & Counselling

Brisbane, Australia.

2004

i

Key Words

Unlicensed, disqualified, suspended, revoked, road safety, driver licensing, deterrence

theory, social learning, Akers, theory testing.

ii

iii

Abstract

Unlicensed driving remains a serious problem for road safety, despite ongoing

improvements in traffic law enforcement practices and technology. While it does not

play a direct causative role in road crashes, unlicensed driving undermines the integrity

of the driver licensing system and is associated with a range of high-risk behaviours.

This thesis documents three studies that were undertaken to explore the scope and

nature of unlicensed driving, in order to develop more effective countermeasures to the

behaviour.

Study One utilised official road crash data from the Australian state of

Queensland to compare the crash involvement patterns of unlicensed drivers with those

of licensed drivers. The results confirmed that unlicensed driving is a relatively small,

but significant road safety problem. Unlicensed drivers represent over 6% of the drivers

involved in fatal crashes and 5% of those in serious injury crashes. Based on a quasi-

induced exposure method, unlicensed drivers were found to be almost three times as

likely to be involved in a crash than licensed drivers. In the event of a crash, those

involving unlicensed drivers were twice as likely to result in a fatality or serious injury.

Consistent with these results, the serious crashes involving unlicensed drivers were

more likely to feature risky driving behaviours, such as drink driving, speeding and

motorcycle use, than those involving licensed drivers.

Study Two involved a cross-sectional survey of 309 unlicensed driving offenders

who were recruited at the Brisbane Central Magistrates Court. The survey involved a

face-to-face interview that took approximately 25 minutes to complete and achieved a

response rate of 62.4%. A wide range of offenders participated in the study, including:

disqualified and suspended drivers; expired licence holders; drivers without a current or

appropriate licence; and those who had never been licensed. The results reinforced

concerns about the on-road behaviour of unlicensed drivers. Almost one quarter of all

the offenders reported driving unlicensed when they thought they might have been over

the alcohol limit. Similarly, 25% reported exceeding the speed limit by 10 km/h or more

on most or all occasions, while 15% admitted that they didn’t always wear their seat

belt. In addition, the results indicated that unlicensed drivers should not be viewed as a

homogeneous group. Significant differences were found between the offender types in

terms of their socio-demographic characteristics (age, education level, prior criminal

convictions); driving history (prior convictions for unlicensed driving and other traffic

iv

offences); whether they were aware of being unlicensed; the degree to which they

limited their driving while unlicensed; and their drink driving behaviour. In particular, a

more deviant sub-group of offenders was identified, that included the disqualified, not

currently licensed and never licensed drivers, who reported higher levels of prior

criminal offending, alcohol misuse and self- reported drink driving. The results of Study

Two also highlight the shortcomings of existing police enforcement practices. Almost

one-third of the sample reported that they continued to drive unlicensed after being

detected by the police (up until the time of the court hearing), while many offenders

reported experiences of punishment avoidance. For example, over one third of the

participants reported being pulled over by the police while driving unlicensed and not

having their licence checked.

Study Three involved the further analysis of the cross-sectional survey data to

explore the factors contributing to unlicensed driving. It examined the influence of

various personal, social and environmental factors on three aspects of the offenders’

behaviour: the frequency of their driving while unlicensed; whether they continued to

drive unlicensed after being detected; and their intentions to drive unlicensed in the

future. This study was also designed to assess the capacity of a number of different

theoretical perspectives to explain unlicensed driving behaviour, including deterrence

theory and Akers’ (1977) social learning theory. At an applied level, the results of Study

Three indicated that personal and social factors exert the strongest influence over

unlicensed driving behaviour. The main personal influences on unlicensed driving were:

the need to drive for work purposes; exposure to punishment avoidance; personal

attitudes to unlicensed driving; and anticipated punishments for the behaviour. The

main social influences reflected the social learning construct of differential association,

namely being exposed to significant others who both engage in unlicensed driving

(behavioural dimension) and hold positive attitudes to the behaviour (normative

dimension). At a theoretical level, the results of Study Three have two important

implications for traffic psychology and criminology. Firstly, they provided partial

support for Stafford and Warr’s (1993) reconceptualisation of deterrence theory by

demonstrating that the inclusion of punishment avoidance can improve the overall

predictive utility of the perspective. Secondly, they suggested that social learning theory

represents a more comprehensive framework for predicting illegal driving behaviours,

such as unlicensed driving. This is consistent with Akers’ (1977; 1990) assertion that

formal deterrence processes can be subsumed within social learning theory.

v

Together, the results of the three studies have important implications for road

safety. Most importantly, they question the common assumption that unlicensed drivers

drive in a more cautious manner to avoid detection. While the findings indicate that

many offenders reduce their overall driving exposure in order to avoid detection, this

does not appear to result in safer driving. While it remains possible that unlicensed

drivers tend to act more cautiously than they would otherwise, it appears that their

driving behaviour is primarily designed to reduce their chances of detection. In terms of

countermeasures, the research indicates that a multi-strategy approach is required to

address the problem of unlicensed driving. Unlicensed drivers do not represent a

homogeneous group who are likely to be influenced by the threat of punishment alone.

Rather, innovative strategies are required to address the wide range of factors that

appear to encourage or facilitate the behaviour. Foremost among these are punishment

avoidance and the need to drive for work purposes.

vi

vii

Table of Contents

Key Words ................................................................................................................. i

Abstract ...................................................................................................................... iii

Table of Contents....................................................................................................... vii

List of Figures ............................................................................................................ xiii

List of Tables ............................................................................................................. xv

Glossary of Terms and Acronyms ............................................................................. xxi

Statement of Original Authorship .............................................................................. xxv

Acknowledgements .................................................................................................... xxvii

Chapter One: Introduction.................................................................................... 1

1.1 Introductory comments.................................................................................................................. 3

1.2 Definition of unlicensed driving................................................................................................ 3

1.3 The rationale for the research....................................................................................................... 4

1.4 Theoretical framework for the research...................................................................................... 5

1.5 Research objectives........................................................................................................................ 7

1.6 Demarcation of Scope ................................................................................................................... 8

1.7 Outline of thesis .............................................................................................................................. 9

1.8 Chapter summary ........................................................................................................................... 11

Chapter Two: Literature review ........................................................................... 13


2.2 The prevalence of unlicensed driving......................................................................................... 15

2.2.1 Roadside licence check surveys .......................................................................................... 15

2.2.2 Observational studies ............................................................................................................ 16

2.2.3 Self-report surveys................................................................................................................. 17

2.2.3.1 Disqualified/suspended drivers................................................................................. 17

2.2.3.2 Other unlicensed drivers ............................................................................................ 18

2.2.4 Crash involvement of unlicensed drivers........................................................................... 18

2.3 The driving behaviour of unlicensed drivers ............................................................................. 20

2.3.1 The disqualified driver effect ............................................................................................... 20

2.3.2 Self-reported driving behaviour.......................................................................................... 20

2.3.3 The behaviour of unlicensed drivers involved in crashes............................................... 21

2.4 Differences among unlicensed drivers........................................................................................ 23

2.5 Factors contributing to unlicensed driving................................................................................. 24

2.6 Countermeasures to unlicensed driving...................................................................................... 26

2.6.1 Administrative and judicial processes............................................................................... 26

2.6.2 Police enforcement practices .............................................................................................. 28

2.6.3 Restricted licences ................................................................................................................ 29

2.6.4 Vehicle-based sanctions ...................................................................................................... 30

2.6.4.1 Alcohol ignition interlocks ........................................................................................ 30

2.6.4.2 Licence plate sanctions.............................................................................................. 31

viii

2.6.4.3 Vehicle immobilisation, impoundment and forfeiture .......................................... 31

2.6.4.4 Electronic licences....................................................................................................... 33

2.6.5 Rehabilitation of offenders ................................................................................................ 34

2.6.6 Mass media campaigns ........................................................................................................ 35

2.6.7 The likely road safety benefits of unlicensed driving countermeasures...................... 35

2.7 Research questions ......................................................................................................................... 36

2.8 Chapter summary ............................................................................................................................ 38

Chapter Three: Theoretical perspectives on unlicensed driving ........................ 39


3.2 Deterrence theory............................................................................................................................ 41

3.2.1 Classical deterrence theory................................................................................................ 42

3.2.1.1 Origins and overview................................................................................................ 42

3.2.1.2 Relevance to unlicensed driving................................................................................ 44

3.2.2 Criticisms of classical deterrence theory........................................................................... 44

3.2.3 Reconceptualisations of deterrence theory ....................................................................... 46

3.2.4 Deterrence-based models of unlicensed driving.............................................................. 48

3.3 Social learning theory .................................................................................................................... 50

3.3.1 Principles of social learning theory.................................................................................... 50

3.3.2 Akers’ social learning theory .............................................................................................. 51

3.3.2.1 Origins of Akers’ theory ............................................................................................. 51

3.3.2.2 Processes and constructs in Akers’ theory ............................................................... 52

3.3.2.3 Empirical support for Akers’ theory ......................................................................... 55

3.3.3 Application of Akers’ theory to unlicensed driving........................................................ 56

3.3.4 A social learning model of unlicensed driving ................................................................ 58

3.4 Sensation seeking............................................................................................................................ 59

3.5 Alcohol misuse................................................................................................................................ 61

3.6 Chapter summary ............................................................................................................................ 62

Chapter Four: The crash involvement of unlicensed drivers ............................. 65


4.2 Study aims and hypotheses ........................................................................................................... 67

4.3 Method.............................................................................................................................................. 69

4.3.1 Data characteristics ............................................................................................................... 69

4.3.2 Procedures .............................................................................................................................. 71

4.3.3 Statistical analyses ................................................................................................................ 73

4.4 Results .............................................................................................................................................. 74

4.4.1 Incidence and severity of unlicensed driver crashes....................................................... 74

4.4.2 Age and gender characteristics of drivers involved in crashes ..................................... 76

4.4.3 Crash circumstances ............................................................................................................. 78

4.4.4 Contributing factors to crashes ........................................................................................... 82

4.4.5 The crash risk of unlicensed drivers .................................................................................. 85

4.4.5.1 Risk of involvement in crashes................................................................................. 86

4.5.5.2 Risk of death and serious injury in the event of a crash ....................................... 91

4.5 Discussion........................................................................................................................................ 95

4.5.1 Study limitations ................................................................................................................... 95

ix

4.5.2 Support for study hypotheses.............................................................................................. 97

4.5.3 Implications for theory......................................................................................................... 102

4.5.4 Implications for road safety................................................................................................ 103

4.5.5 Future directions for research............................................................................................. 104

4.6 Chapter summary ........................................................................................................................... 105

Chapter Five: The self-reported behaviour of unlicensed drivers ...................... 107

5.1 Introductory comments.................................................................................................................. 109 5.2 Study aims and hypotheses ........................................................................................................... 109 5.3 Method ............................................................................................................................................. 112

5.3.1 General research strategy.................................................................................................... 112 5.3.2 Exploratory research ............................................................................................................ 113

5.3.2.1 Semi-structured qualitative interviews .................................................................... 113 5.3.2.2 Pilot quantitative interviews...................................................................................... 114

5.3.3 Main study............................................................................................................................. 115 5.3.3.1 Selection of survey location...................................................................................... 115 5.3.3.2 Participants................................................................................................................... 116 5.3.3.3 Materials ....................................................................................................................... 116 5.3.3.4 Procedure ...................................................................................................................... 118 5.3.3.5 Statistical analyses ...................................................................................................... 119

5.4 Results .............................................................................................................................................. 120 5.4.1 Sample characteristics.......................................................................................................... 120

5.4.1.1 Response rate............................................................................................................... 120 5.4.1.2 Type of offender.......................................................................................................... 122 5.4.1.3 Socio-demographic characteristics of the sample ................................................. 123 5.4.1.4 Driving and criminal history of offenders .............................................................. 125 5.4.1.5 Driving for work purposes......................................................................................... 127 5.4.1.6 Sensation seeking........................................................................................................ 127 5.4.1.7 Alcohol misuse............................................................................................................ 128

5.4.2 Circumstances of detection and outcome of court hearing............................................ 130 5.4.2.1 Reason stopped by Police .......................................................................................... 130 5.4.2.2 Reason for driving when detected............................................................................ 132 5.4.2.3 Vehicle driven when detected................................................................................... 132 5.4.2.4 Outcome of court hearing .......................................................................................... 134

5.4.3 Unlicensed driving behaviour............................................................................................. 135 5.4.3.1 Length of time driving unlicensed........................................................................... 135 5.4.3.2 Frequency of unlicensed driving .............................................................................. 136 5.4.3.3 On-road driving behaviour........................................................................................ 138

5.4.4 The impact of current administrative, enforcement and punishment processes................................................................................................................................ 144

5.4.4.1 Awareness of being unlicensed at time of detection............................................. 144 5.4.4.2 Possession of a photographic licence ...................................................................... 144 5.4.4.3 Unlicensed driving after detection ........................................................................... 145 5.4.4.4 Evasion of detection ................................................................................................... 146 5.4.4.5 Perceptions of enforcement and punishment processes ....................................... 148

x

5.4.4.6 Intention to drive unlicensed in the future .............................................................. 150 5.5 Discussion........................................................................................................................................ 151

5.5.1 Study limitations ................................................................................................................... 151 5.5.2 Support for study hypotheses .............................................................................................. 152 5.5.3 Implications for theory......................................................................................................... 156 5.5.4 Implications for road safety................................................................................................ 157

5.5.4.1 The extent and nature of unlicensed driving........................................................... 157 5.5.4.2 The effectiveness of current administrative, enforcement and

punishment processes ................................................................................................ 158 5.5.5 Future directions for research............................................................................................. 160 5.6 Chapter summary ............................................................................................................................ 161

Chapter Six: Factors contributing to unlicensed driving .................................... 163


6.2 Study aims and hypotheses ........................................................................................................... 166

6.3 Method.............................................................................................................................................. 168

6.3.1 Overview of method............................................................................................................. 168

6.3.2 Derivation of study variables .............................................................................................. 168

6.3.2.1 Independent variables ................................................................................................ 168

6.3.2.2 Dependent variables.................................................................................................... 170

6.3.3 Statistical analyses ................................................................................................................ 171

6.4 Results .............................................................................................................................................. 173

6.4.1 Socio-demographic variables .............................................................................................. 173

6.4.2 Sensation seeking.................................................................................................................. 175

6.4.3 Alcohol misuse...................................................................................................................... 176

6.4.4 Environmental facilitating factors...................................................................................... 177

6.4.5 Deterrence factors ................................................................................................................. 179

6.4.5.1 Classical deterrence variables ................................................................................... 179

6.4.5.2 Expanded deterrence variables.................................................................................. 181

6.4.5.3 Comparison of classical and expanded deterrence perspectives......................... 182

6.4.5.4 Summary of deterrence perspectives ....................................................................... 184

6.4.6 Social learning factors.......................................................................................................... 185

6.4.6.1 Imitation........................................................................................................................ 185

6.4.6.2 Differential association............................................................................................... 187

6.4.6.3 Personal attitudes......................................................................................................... 187

6.4.6.4 Differential reinforcement ......................................................................................... 188

6.4.6.5 Predictive role of social learning variables ............................................................. 190

6.4.7 Summary of contributing factors........................................................................................ 191

6.4.7.1 Comparison of the different theoretical perspectives............................................ 191

6.4.7.2 The effectiveness of deterrence and social learning theories in explaining deviant behaviour .................................................................................. 193

6.4.7.3 Extending the social learning explanation of unlicensed driving ....................... 195

6.5 Discussion........................................................................................................................................ 200

6.5.1 Study limitations ................................................................................................................... 200

6.5.2 Support for study hypotheses .............................................................................................. 201

6.5.3 Theoretical implications and directions for future research .......................................... 209

xi

6.5.3.1 Implications for deterrence theory ........................................................................... 209

6.5.3.2 Implications for social learning theory.................................................................... 211

6.5.4 Countermeasure implications............................................................................................. 215

6.6 Chapter summary ........................................................................................................................... 216

Chapter Seven: Discussion..................................................................................... 217


7.2 Review of findings......................................................................................................................... 219

7.2.1 Do unlicensed drivers engage in more risky driving than other drivers?.................. 219 7.2.2 Is unlicensed driving associated with a higher crash risk compared to legal driving?........................................................................................................................ 220 7.2.3 Do unlicensed drivers represent a homogeneous group, in terms of their psychosocial characteristics and on-road behaviour?.................................................... 222 7.2.4 How effective are current administrative, enforcement and punishment policies and processes in preventing unlicensed?.......................................................... 223 7.2.5 What are the personal, social and environmental factors contributing

to unlicensed driving?......................................................................................................... 224 7.3 Contribution to theory ................................................................................................................... 226 7.4 Implications for road safety.......................................................................................................... 228

7.4.1 The need to encourage participation in the driver licensing system............................ 228

7.4.2 Countermeasures suggestions............................................................................................. 230 7.4.2.1 Driver licensing and other administrative processes ............................................ 230

7.4.2.2 Traffic law enforcement practices ........................................................................... 231

7.4.2.3 Unlicensed driving sanctions and punishment processes .................................... 232 7.4.2.4 The need to target work-related driving................................................................ 234

7.5 Strengths and limitations of the research .................................................................................. 235

7.6 Suggestions for future research.................................................................................................... 239 7.7 Concluding remarks ....................................................................................................................... 241

References................................................................................................................. 243

Appendices................................................................................................................ 255

A Types of unlicensed driving under Queensland legislation..................................................... 257

B Results of semi -structured qualitative interviews conducted for Study Two/Three ........... 259

C Proforma for recording people approached to participate in Study Two/Three .................. 265

D Final questionnaire used for interviews in Study Two/Three................................................. 267

E Interviewer’s Guide used in Study Two/Three.......................................................................... 285

F Summary of scales used in Studies Two and Three................................................................ 289

G Reported reasons for non-participation in Study Two/Three................................................... 295

H Correlation matrix for dependent and independent variables used in Study Three ............. 297

I Supplementary tables from Study Three..................................................................................... 303

xii

xiii

List of Figures

Figure 3.1 Classical deterrence model of licensed/unlicensed driving................................48

Figure 3.2 Extended deterrence model of licensed/unlicensed driving ................................49

Figure 3.3 Social learning model of licensed/unlicensed driving................................59

Figure 4.1 Involvement rate (IR) for total crashes by driver type: 1994-98................................88

Figure 5.1 Reason cited by participants for being unlicensed ................................ 123

Figure 5.2 Reported reason for being stopped ................................................................130

Figure 5.3 Reported reason for driving when stopped ................................................................132

Figure 5.4 Owner of vehicle driven by offenders when detected ................................133

Figure 5.5

Reported frequency of driving 10 km/h or more over the

speed limit ................................................................................................141

Figure 5.6 Reported frequency of wearing a seat belt ................................................................142

Figure 5.7 General approach to drink driving ................................................................142

xiv

xv

List of Tables

Table 4.1

Number of drivers involved in crashes in Queensland between

1994-98, by year and crash severity ................................................................

70

Table 4.2

Licence status of drivers involved in crashes in Queensland:

1994-98 by crash severity.....................................................................................

75

Table 4.3 Unlicensed drivers involved in crashes in Queensland:

1994-98 by offender type......................................................................................

76

Table 4.4 Age and gender of drivers involved in serious casualty crashes

in Queensland: 1994-98 by licence status ............................................................

76

Table 4.5 Age and gender of unlicensed drivers involved in serious casualty

crashes in Queensland: 1994-98 by offender type................................

77

Table 4.6 Circumstances of crashes resulting in serious casualty crashes

in Queensland: 1994-98 by licence status ............................................................

79

Table 4.7 Circumstances of crashes resulting in serious casualty crashes

in Queensland: 1994-98 by unlicensed driver type ................................

81

Table 4.8 Contributing factors to serious casualty crashes in Queensland:

1994-98 by licence status ......................................................................................

83

Table 4.9 Contributing factors to serious casualty crashes in Queensland:

1994-98 by unlicensed driver type ................................................................

84

Table 4.10 Risk of involvement in a multi-vehicle crash by driver type

and crash severity for Queensland: 1994-98 ........................................................

86

Table 4.11 The involvement of drivers in multi-vehicle crashes compared

with all crashes, by severity for Queensland: 1994-98................................

89

Table 4.12 Percentage of at- fault drivers in single and multi-vehicle crashes

by licence status and severity of crash for Queensland: 1994-98.........................

90

Table 4.13

Risk of involvement in a serious casualty crash, relative to a minor

crash, for different licence categories in Queensland: 1994-98 ...........................

91

Table 4.14 Logistic regression analysis of the severity of crashes as a

function of selected driver-related variables.........................................................

93

Table 4.15 Logistic regression analysis of the severity of crashes as a function

of unlicensed driver types and selected driver-related variables ..........................

94

xvi

Table 5.1 Characteristics of all offenders approached to participate ................................121

Table 5.2 Types of unlicensed drivers ................................................................ 122

Table 5.3 Socio-demographic characteristics of participants by offender type 124

Table 5.4 Driving and criminal history of participants by offender type .............................126

Table 5.5 Needed to drive for work purposes while unlicensed by

offender type ................................................................................................127

Table 5.6 Sensation seeking scores by type of unlicensed driver................................128

Table 5.7 AUDIT scores by type of unlicensed driver .........................................................129

Table 5.8

Bivariate correlations between AUDIT scores and driving

offences, criminal offences and sensation seeking ................................ 130

Table 5.9 Reason stopped by police by offender type ..........................................................131

Table 5.10 Ownership of the vehicle driven by participants by offender type .......................133

Table 5.11 Penalties received for unlicensed/disqualified driving conviction

among offenders who were convicted of no other offences ................................135

Table 5.12 Length of time driving unlicensed by offender type................................136

Table 5.13 Frequency of driving while unlicensed by offender type ................................137

Table 5.14 Cautiousness of driving while unlicensed by offender type ................................139

Table 5.15 Drink driving behaviour while unlicensed by offender type ................................143

Table 5.16 Awareness of being unlicensed when detected, by offender type ........................144

Table 5.17 Possession of a photographic licence by offender type ................................145

Table 5.18 Continued driving unlicensed after detection by offender type ............................146

Table 5.19

Incidents where offenders evaded detection while driving

unlicensed................................................................................................147

Table 5.20

Perceived likelihood of being caught for illegal behaviours prior to

being detected for unlicensed driving ................................................................148

Table 5.21 Perceived severity, certainty and swiftness of punishment for

unlicensed driving ................................................................................................150

Table 5.22 Intention to drive unlicensed in the future by offender type................................151

Table 6.1 Bivariate correlations between dependent variables ................................171

Table 6.2

Bivariate correlations between dependent variables and

socio-demographic variables................................................................ 174

Table 6.3


sensation seeking................................................................................................175

xvii

Table 6.4

Bivariate correlations between dependent variables and AUDIT

scores ....................................................................................................................176

Table 6.5


environmental facilitating factors ................................................................177

Table 6.6


deterrence variables ..............................................................................................180

Table 6.7

Total number of unlicensed drivers known by participants

by type of offender................................................................................................186

Table 6.8

Bivariate correlations between dependent variables and social

learning variables ................................................................................................187

Table 6.9

Strength of association between the theoretical perspectives

and the dependent variables ................................................................ 192

Table 6.10

Strength of association between dependent variables and expanded

deterrence and social learning variables, by unlicensed driver sub-

group .....................................................................................................................194

Table 6.11

Standard multiple regression of social learning and other selected

variables on the frequency of unlicensed driving ................................ 196

Table 6.12

Logistic regression analysis of continued driving after detection as

a function of social learning and other selected variables ................................197

Table 6.13

Standard multiple regression of social learning and other selected

variables on intention to drive unlicensed in the future................................199

Table G1 Reported reasons for non-participation in Study Two/Three ...............................295

Table H1

Correlation matrix of dependent and independent variables

used in Study Three ..............................................................................................299

Table I1

Standard multiple regression of socio-demographic variables

on frequency of unlicensed driving ................................................................305

Table I2 Logistic regression analysis of continued driving after detection

as a function of socio-demographic variables ......................................................306

Table I3 Standard multiple regression of socio-demographic variables on

intent ion to drive unlicensed in the future ............................................................307

Table I4

Standard multiple regression of environmental facilitating factors

on frequency of unlicensed driving ................................................................308

xviii


as a function of environmental facilitating factors................................ 308

Table I6 Standard multiple regression of environmental facilitating factors

on intention to drive unlicensed in the future .......................................................309

Table I7

Standard multiple regression of classical deterrence variables

on frequency of unlicensed driving................................................................309

Table I8

Logistic regression analysis of continued driving after detection

as a function of classical deterrence variables ......................................................310

Table I9

Standard multiple regression of classical deterrence variables on

intention to drive unlicensed in the future ............................................................310

Table I10

Standard multiple regression of expanded deterrence variables

on frequency of unlicensed driving................................................................311

Table I11


as a function of expanded deterrence variables ....................................................312

Table I12


on intention to drive unlicensed in the future .......................................................313

Table I13

Hierarchical regression of deterrence variables on frequency

of unlicensed driving.............................................................................................314

Table I14

Sequential logistic regression analysis of continued driving after

detection as a function of deterrence variables .....................................................315

Table I15

Hierarchical regression of deterrence variables on intention

to drive unlicensed in the future................................................................316

Table I16

Standard multiple regression of social learning variables on

frequency of unlicensed driving................................................................317

Table I17


as a function of social learning variables..............................................................317

Table I18

Standard multiple regression of social learning variables on

intention to drive unlicensed in the future ............................................................318

Table I19


on frequency of unlicensed driving for disqualified, not currently

licensed and never licensed drivers................................................................319

xix

Table I20


on frequency of unlicensed driving for suspended, expired and

not appropriately licensed drivers................................................................320

Table I21


as a function of expanded deterrence variables for disqualified,

not currently licensed and never licensed drivers ................................ 321

Table I22


as a function of expanded deterrence variables for suspended,

expired and not appropriately licensed drivers .....................................................322

Table I23 Standard multiple regression of expanded deterrence variables

on intention to drive unlicensed in the future for disqualified,

not currently licensed and never licensed drivers ................................ 323

Table I24 Standard multiple regression of expanded deterrence variables

on intention to drive unlicensed in the future for suspended,

expired and not appropriately licensed drivers .....................................................324

Table I25 Standard multiple regression of social learning variables on

frequency of unlicensed driving for disqualified, not currently

licensed and never licensed drivers ................................................................325


frequency of unlicensed driving for suspended, expired and

not appropriately licensed drivers................................................................326


as a function of social learning variables for disqualified, not

currently licensed and never licensed drivers .......................................................327


as a function of social learning variables for suspended, expired

and not appropriately licensed drivers ................................................................327


intention to drive unlicensed in the future for disqualified, not

currently licensed and never licensed drivers .......................................................328


intention to drive unlicensed in the future for suspended, expired

and not appropriately licensed drivers ................................................................329

xx

xxi

Glossary of Terms and Acronyms

Alcohol ignition interlock

A device that prevents a vehicle being started if the breath specimen provided by the

driver indicates a blood alcohol concentration above a pre-set limit.

ATSB

Australian Transport Safety Bureau [formerly Federal Office of Road Safety (FORS)].

BAC

Blood alcohol concentration. A measurement of the proportion of alcohol in a person’s

blood, typically obtained using a breathalyser or by conducting a blood test. In Australia,

it is generally expressed as grams per 100ml of blood.

CARRS-Q

Centre for Accident Research and Road Safety - Queensland.

Cross-sectional method

A research method which involves studying a large group of subjects at one point in time,

as opposed to a longitudinal method which involves studying the behaviour of individual

subjects over an extended period of time (Reber, 1985; Strangor, 1998).

Demerit points

Points that are allocated to an individual's driving record for traffic offences. The

accumulation of the maximum number of demerit points over a set period results in

licence cancellation and a period of suspension or transfer to provisional licence status.

Deviant behaviour

A road safety perspective is adopted to the concept of deviance in this research. The term

is used to refer to behaviour that differs markedly from the accepted standards within

society, as generally reflected in the road rules.

xxii

Driver

The operator of a motorised vehicle including a car, truck, bus or motorcycle.

DUI

Driving under the influence of alcohol.

DWD/DWS

Driving-while-disqualified/Driving-while-suspended.

FARS

Fatal Accident Reporting System used in the USA.

Licence disqualification

The removal of a person’s authority to hold or obtain a driver’s licence, typically under a

court order.

Licence suspension

The administrative removal of a person’s authority to drive for reasons such as

accumulation of excess demerit points, being medically unfit or for unpaid fines.

Negative halo effect

In this study, the term is used to explain the possible tendency of police to assign blame

for a crash to an unlicensed driver due to their illegal licence status (from DeYoung et

al, 1997).

Psychosocial

Processes or factors that are both social and psychological in origin (Collins English

Dictionary, 1979) or “a grab bag term used freely to cover any situation where both

psychological and social factors are assumed to play a role” (Reber, 1985, p.620).

QLD

Australian state of Queensland.

xxiii

RBT

Random Breath Testing.

Road crash

A crash reported to the police that resulted from the movement of at least one road vehicle

(motorised or non-motorised) on a road and involving death or injury to any person, or

property damage.

- Fatal crash

A road crash resulting in the death of a person within 30 days of injuries sustained in

the crash.

- Serious injury crash (sometimes referred to as a hospitalisation crash)

A road crash resulting in the hospitalisation of a person due to injuries sustained in the

crash.

- Serious casualty crash

A road crash resulting in either the death (within 30 days) or hospitalisation of a

person due to injuries sustained in the crash (ie. a summation of fatal and serious

injury crashes).

- Minor injury crash

A road crash resulting in the injury, but not hospitalisation, of a person due to their

involvement in a crash.

- Property damage only (PDO) crash

A road crash where no one was injured but at least one vehicle is towed away or the

damage cost is greater than a predetermined level (eg. $2,500 in Queensland).

Unlicensed driver

A generic term used to refer to people who drive or ride a motor vehicle without a valid

driver's licence, including those who:

- have let their licence expire;

- have been disqualified or suspended from driving;

- hold an inappropriate licence for the class of vehicle they drive;

- drive outside the restrictions of a special licence;

- don’t currently hold a licence; or

- have never held a licence.

xxiv

xxv

Statement of Original Authorship

The work contained in this thesis has not been previously submitted for a degree or

diploma at any other higher education institution. To the best of my knowledge and

belief, this thesis contains no material previously published or written by another person

except where due reference is made.

Signed: .......................…………......…

Date: ...............................………….…

xxvi

xxvii

Acknowledgements

A large number of people have assisted with the development and conduct of this

program of research. While it is difficult to acknowledge all these people, I would

particularly like to thank:

§ My supervisors, Professor Mary Sheehan and Dr Vic Siskind, for their invaluable

support, advice and direction;

§ The Australian Transport Safety Bureau (ATSB) for their financial support of

Studies Two and Three;

§ Queensland Transport for providing the data used in Study One;

§ Clive Williams, Paul Dickson, Mike Smith and Aine Fitzpatrick for their interview

skills and untiring efforts in collecting the data for Studies Two and Three;

§ Dr Julie Hansen, Professor Clive Bean and Reyna Watson for their comments and

advice on the draft of this thesis;

§ My colleagues at CARRS-Q, particularly Cynthia Schonfeld for her ongoing

support, Anna Johnson for her assistance in locating relevant literature, and Diane

Jensen and Maxine Nott for their formatting skills;

§ Patricia Young for her assistance with data entry; and

§ the staff of the Brisbane Central Magistrates Court, particularly Ron Micola and

Nev Bawden.

Finally, and most importantly, I would like to thank my family, Reyna, Harry and

Emily for the untiring support and assistance over the last six years. It has been a long

haul that would not have been possible without them.

xxviii

The characteristics and on-road behaviour of unlicensed drivers 1

Chapter One: Introduction

1.1 Introductory comments ....................................................................................3

1.2 Definition of unlicensed driving ................................................................ 3

1.3 The rationale for the research...........................................................................4

1.4 Theoretical framework for the research...........................................................5

1.5 Research objectives..........................................................................................7

1.6 Demarcation of scope.......................................................................................8

1.7 Outline of thesis ...............................................................................................9

1.8 Chapter summary.............................................................................................11



1.1 Introductory comments

Unlicensed driving remains a serious problem in many countries, despite ongoing

improvements in traffic law enforcement practices and technology. In the United States,

over 10% of the drivers involved in fatal crashes do not hold a valid licence, while

approximately 20% of all fatal crashes involve at least one of these drivers (Griffin &

DeLaZerda, 2000; Scopatz, Hatch, Delucia & Tays, 2003). In Australia, unlicensed

drivers represent over 5% of the drivers involved in fatal crashes. The crashes involving

unlicensed drivers and riders account for almost 10% of the national road toll (FORS,

1997a).

Unlicensed driving represents a major problem for road safety in two respects.

Firstly, it serves to undermine the system used to monitor and manage driver behaviour.

Because they operate outside the licensing system, unlicensed drivers dramatically reduce

the ability of authorities to monitor and manage their behaviour through sanctions, such

as demerit points. In particular, it serves to undermine the effectiveness of licence

disqualification, which has otherwise been demonstrated to be a very effective deterrent

to illegal behaviour (Nichols & Ross, 1990; Vingilis, Mann, Gavin, Adlaf & Anglin, 1990;

Siskind, 1996). Secondly, there is a growing body of evidence linking unlicensed driving

to a cluster of high-risk behaviours, including drink driving, speeding, failure to wear seat

belts and motorcycle use (Harrison, 1997; FORS, 1997b; Watson, 1997 & 2000; Griffin

& DeLaZerda, 2000). Consistent with this, the crashes involving unlicensed drivers tend

to be more severe than those involving licensed drivers, resulting in higher rates of

fatality and serious injury (Watson, 2000).

Accordingly, there is a need to better understand the scope and nature of unlicensed

driving, in order to develop and implement more effective countermeasures to the

behaviour. This thesis documents a program of research undertaken for this purpose.

1.2 Definition of unlicensed driving

In the international literature, a variety of terms are used to describe the people who

choose to drive or ride a motor vehicle without a valid licence. Among the more common

terms used are unlicensed driver, unauthorised driver, disqualified driver, suspended

driver, revoked driver, cancelled driver and never licensed driver. Some of these terms

are used in a general sense, while others are used to describe particular sub-groups or

types of drivers. For example, terms like disqualified, suspended or revoked are generally

used to describe those drivers who have had their licence removed by a judicial or


administrative process. In Australia, the term unlicensed driver tends to be used as the

generic description for all those people who drive or ride a motor vehicle without a valid

licence (eg. Watson, 1998d; Travelsafe, 1998). While the term is also commonly used in

the USA, it is sometimes confined for use with those drivers who have never held a valid

licence (Scopatz et al, 2003).1 To avoid confusion, the Australian terminology will be

used throughout this research. Hence, the term unlicensed driver will be used to refer to

drivers who:

§ have let their licence expire;

§ have had their licence disqualified or suspended;

§ hold an inappropriate licence for the class of vehicle they drive;

§ drive outside the restrictions of a special licence;

§ don’t currently hold a licence; or

§ have never held a licence.

In Queensland, the legislation relating to unlicensed driving is covered in Section

78 of the Transport Operations (Road Use Management) Act 1995. Appendix A provides

a more detailed list of the ways in which a person can drive unlicensed in Queensland.

Unless specified otherwise, the term driver is used in this thesis to refer to both drivers

and riders of motorised vehicles.

1.3 Rationale for the research

A number of obstacles have hindered research into the issue of unlicensed driving

and the development of related countermeasures. Firstly, unlicensed drivers are not

necessarily a homogeneous group. A wide variety of people drive without a valid licence,

including those who: have let their licence expire; have had their licence disqualified or

suspended; drive a vehicle without an appropriate licence; or have never held a licence.

As such, the motives for a person being unlicensed and the associated driving behaviours

may vary greatly. In addition, the crash data indicate that those drivers who have let their

licence expire are less likely to be involved in serious crashes than those who have never

held a licence, have had it disqualified or hold an inappropriate class of licence (Watson,

1997). This suggests a possible link between the degree of risk-taking displayed by

different types of unlicensed drivers and the intentionality of their actions. Therefore,

from a theoretical perspective it may be more appropriate to view different types of

1. The term ‘never licensed’ is generally used in Australia to describe this particular group of drivers.


unlicensed driving as discrete behaviours (or even possibly a continuum of behaviours),

representing varying degrees of non-conformity or deviance2.

Secondly, minimal research has been conducted into the crash involvement patterns

of unlicensed drivers, both within Australia and internationally. While some studies have

been conducted in this area (eg. Harrison, 1997; FORS, 1997b; Watson, 1997 & 2000;

Griffin & DeLaZerda, 2000), they have tended to be general in approach and have not

specifically compared the crash involvement patterns of drivers unlicensed for different

reasons.

Finally, there is a lack of reliable data available relating to the behavioural

characteristics and perceptions of unlicensed drivers. Although some self-report surveys

have been conducted with unlicensed drivers (eg. Robinson, 1977; Williams, Hagen &

McConnell, 1984; Ross & Conzales, 1988; Smith & Maisey, 1990; Job, Lee & Prabhakar,

1994; Davies et al, 1999) the conclusions are constrained by low response rates.

Moreover, these studies have been largely descriptive, concent rating on self- reported

reasons for unlicensed driving, and have lacked a strong theoretical base to guide the

interpretation of results (Watson, 1998a).

Consequently, there is a need for further research into the attitudes, behaviour and

crash involvement of unlicensed drivers. In particular, there is a need to better understand

the extent and nature of unlicensed driving and the factors that contribute to the

behaviour, and whether differences exist among different groups of offenders.

1.4 Theoretical framework for the research

This thesis will adopt a multi-disciplinary approach to theory, drawing on

perspectives from psychology, sociology and criminology. The two main theoretical

perspectives that will be examined are deterrence theory and social learning theory.

While there are some similarities between these two perspectives, major differences exist

in the scope of factors upon which they focus. Deterrence theory is a criminological

perspective grounded in sociology, which is concerned with the influence of legal

sanctions on criminal behaviour (Gibbs, 1975). In contrast, social learning theory is a

psychologically-based perspective that is more concerned with the overall social setting

2. In general, a road safety perspective is adopted to the concept of deviance in this research. It will be

used to refer to behaviour that differs markedly from the accepted standards within society, as generally reflected in the road rules. However, it should be borne in mind that each of the theoretical perspectives utilised in the research features its own conception of deviance, grounded in its particular psychological or sociological traditions.


in which behaviours occur and the way in which they are differentia lly rewarded and

punished (Akers, 1990).

Deterrence theory focuses upon the perceived severity, certainty and swiftness of

legal sanctions and has been used extensively in Australia and other motorised countries

to guide the development of many road safety countermeasures, particularly in the area of

drink driving (eg. Ross, 1982; Homel, 1988). In Australia, it has underpinned the design

of traffic law enforcement programs such as Random Breath Testing (RBT) and speed

cameras (Homel, 1988; Cameron, Cavallo & Gilbert, 1992; Watson et al, 1996). Not

surprisingly, deterrence theory has also been used to explain the prevalence of unlicensed

driving. For example, researchers have suggested that the high level of unlicensed driving

in many jurisdictions is primarily a function of a low perceived risk of apprehension

(Nichols & Ross, 1990; Ross, 1991; Watson et al, 1996). Furthermore, studies undertaken

by Robinson and Kelso (1981) and Job et al (1994) have provided some support for a

deterrence-based explanation of unlicensed driving (see section 3.2.1.2 for further

discussion of this evidence).

However, deterrence theory has been criticised by some researchers for being too

narrow in scope to explain the wide range of factors (other than legal sanctions) which

can influence compliance with the law (eg. Vingilis, 1990). Indeed, Akers (1977, 1990)

has argued that deterrence theory is not a general or complete model of criminal

behaviour, but represents a sub-set of social learning theory. His central thesis is that “the

primary concepts and valid postulates of deterrence and rational choice are subsumable

under general social learning or behavioural principles” (Akers, 1990, p. 655).

Social learning theory appears to represent a more comprehensive perspective for

examining unlicensed driving. It provides a framework for incorporating a wide range of

psychological and sociological factors, including the influence of personal and social

rewards associated with the behaviour. The particular social learning perspective that will

be used in this research is Akers’ (1977, 1990) differential association-reinforcement

theory. This perspective draws on both sociological theory (Sutherland’s differential

association theory) and psychological theory (Skinner’s operant conditioning). While this

perspective has been successfully used to investigate a wide range of deviant or non-

conforming behaviours including alcohol and drug abuse, adolescent smoking,

delinquency, adolescent sexual behaviour and computer crime (eg. Akers et al, 1979;

Krohn, Skinner, Akers & Massey, 1985; DiBlasio & Benda, 1990; Akers & Lee, 1996;

Skinner & Fream, 1997), it has not been utilised widely in the road safety field. One


exception was a study by DiBlasio (1987), which showed that a social learning model

was a good predictor of adolescents’ choice to ride with a drinking driver.

Although social learning theory may represent a more comprehensive perspective

than deterrence theory, its capacity to better explain unlicensed driving remains unclear.

Indeed, deterrence theory may represent a more parsimonious approach for explaining the

behaviour and for guiding countermeasure design.

In addition, the empirical evidence suggests that other perspectives may be relevant

for understanding the behaviour of some unlicensed drivers. As already noted, there is a

growing body of evidence linking unlicensed driving to high-risk driving behaviours,

such as drink driving and speeding. This suggests that unlicensed driving may be

influenced by an individual’s general propensity to take risks on the road. In this regard, a

variety of studies have demonstrated a significant positive relationship between the

theoretical construct of sensation seeking and risky driving (Jonah, 1997). Furthermore,

other research suggests that some unlicensed drivers, particularly those who have

previously lost their licence for drink driving, may have an alcohol-related health

problem. This raises the possibility that the behaviour of some unlicensed drivers is

strongly influenced by alcohol misuse or dependence.

As Stafford and Warr (1993, p.133) note: “While no conception is right or wrong,

some are more useful than others”. Accordingly, a key aim of the current study is to

compare the capacity of different theoretical perspectives to explain unlicensed driving

behaviour. Besides providing insights into the factors that motivate the behaviour, this

will have major implications for the development of more effective prevention strategies.

1.5 Research objectives

Based on the empirical and theoretical considerations discussed above, the main

objectives of the research are to:

1. investigate the scale and nature of the unlicensed driving problem;

2. investigate whether there are any systematic differences between unlicensed driver

sub-groups, in terms of their psychosocial characteristics or on-road driving

behaviour;

3. examine the personal, social and environmental factors contributing to unlicensed

driving;

4. compare the capacity of different theoretical perspectives to explain the behaviour

of unlicensed drivers; and


5. identify improvements to current countermeasures and potential initiatives to

reduce the incidence of unlicensed driving.

The specific research questions examined in this thesis are outlined in section 2.7,

following a review of the literature relating to unlicensed driving.

1.6 Demarcation of scope

This thesis primarily examines unlicensed driving among drivers who, subject to

the satisfactory completion of any licence testing requirements or outstanding

disqualification/suspension, would have otherwise been eligible to obtain a driver’s

licence. As such, the primary focus is on those unlicensed drivers who have reached the

minimum age for obtaining a driver's licence3. While under-age drivers were included in

Study One, it was not possible to examine this group in Studies Two and Three. From a

legal perspective under-aged drivers are not dealt with as adults but are processed through

juvenile courts. As a result, it is considerably more difficult for researchers to obtain

approval to survey these drivers or access their driving records. Consequently, a more

age-appropriate methodology would have been required to recruit under-age drivers in

Studies Two and Three, which was beyond the scope of this research.

A second group of unlicensed drivers that were not specifically investigated in this

research was indigenous offenders. A recent national review of indigenous road user

safety (Edmonston et al, in press), reported that unlicensed driving was a major road

safety problem facing indigenous Australians. While acknowledging the data limitations

plaguing research in this area, it appears that unlicensed driving is a major contributing

factor to the higher rates of incarceration among indigenous people. For example, in

Queensland the rate of imprisonment among indigenous people is nearly 12 times that for

non- indigenous people and in more than half of the indigenous cases the index offence

was unlicensed driving or drink driving (National Crime Prevention Branch, 2000 cited in

Edmonston et al, in press).

A specific examination of unlicensed driving among indigenous people was beyond

the scope of this research. It would have required the use of a method that would have

ensured the participation of a sufficient number of indigenous offenders. Moreover,

research within indigenous communities needs to be conducted in a culturally-appropriate

manner, which includes obtaining informed consent from the appropriate figurehead in

the community, developing mutual rapport and acknowledging the community's

3. In Queensland, the minimum age for obtaining a provisional licence that allows unaccompanied

driving is 17 years.


ownership of the data (Edmonston et al, in press). Without the use of appropriate research

protocols, it is unlikely that research in this area will identify the unique cultural, access

and procedural barriers impacting on an indigenous person's ability to obtain and retain a

driver's licence.

Finally, this thesis is theoretically concerned with the factors that encourage or

motivate non-compliance with driver licensing laws. However, a robust theoretical

perspective should not only be able to explain why some drivers choose to drive

unlicensed, but also account for the fact that the majority of drivers currently comply with

driver licensing requirements. As noted by Meier and Johnson [in relation to deterrence

theory] (1977, p.295):

Indeed, crime rates may be logically irrelevant to the study of deterrence insofar as

they can only indicate who has not been deterred. Research on deterrence must

utilize observations of both compliance and non-compliance.

In this regard, the analysis of the crash data undertaken in this thesis (Study One)

compares the crash involvement patterns of unlicensed drivers with those of licensed

drivers. However, it was beyond the scope of the research to examine the attitudes,

perceptions and self-reported behaviour (Studies Two and Three) of both unlicensed and

licensed drivers.

1.7 Outline of thesis

The structure of this thesis reflects the specific tasks undertaken as part of the

research.

Chapter Two reviews the available evidence relating to unlicensed driving, drawing

on a variety of sources including crash studies, roadside surveys and self-report surveys.

The major issues examined are: the prevalence of unlicensed driving; the behaviour of

unlicensed drivers (both in general and among different groups of offenders); the

personal, socia l and environmental factors contributing to the behaviour; and the

effectiveness of current countermeasures to unlicensed driving. At the conclusion of the

chapter, a number of key research questions are identified to guide the research. Earlier

versions of this chapter appeared in two reports prepared by the author. The first was a

submission to the Queensland Parliamentary Road Safety Committee (Travelsafe)

entitled Submission to Travelsafe: Unlicensed driving in Queensland (Brisbane: Centre

for Accident Research and Road Safety – Queensland). The second was a report prepared

for the Australian Transport Safety Bureau (ATSB) entitled The road safety implications


of unlicensed driving: A survey of unlicensed drivers (Canberra: Australian Transport

Safety Bureau).

Chapter Three examines a number of theoretical perspectives relevant to unlicensed

driving, with a special focus on deterrence theory and social learning theory. Based on

this review, a number of alternative theoretical models of unlicensed driving are proposed

which inform Studies One and Two and are more directly examined in Study Three.

Chapter Four documents the first study undertaken as part of the research. This

study involved an analysis of the reported crashes involving unlicensed drivers in the

Australian state of Queensland. This study utilised the Queensland Road Crash Database

to compare the crash involvement of unlicensed drivers with licensed drivers and to

explore differences among unlicensed driver sub-groups involved in crashes. Earlier

versions of this chapter were published in two conference proceedings: Watson B.

(1997), The crash involvement of unlicensed drivers in Queensland, Proceedings of the

1997 Road Safety Research and Enforcement Conference (Hobart: Department of

Transport); and Watson B. (2000), The crash involvement of unlicensed drivers,

Abstracts from 17th Congress of the International Association of Automobile and Traffic

Medicine (AATM2000), Journal of Traffic Medicine, 28 (2S), 21.

Chapter Five documents Study Two, which involved a cross-sectional survey of

unlicensed driving offenders recruited at the Brisbane Central Magistrates Court. It

primarily focuses on the psychosocial characteristics of the offenders, the circumstances

of their detection, their driving behaviour while unlicensed, and the impact of current

administrative, enforcement and punishment processes on their behaviour. An earlier

version of this chapter was published in the peer-reviewed proceedings of the 2002 Road

Safety Research, Policing and Education Conference (Adelaide: Transport SA).

Chapter Six documents Study Three. It involved a further analysis of the survey

data collected in Study Two and explores the factors contributing to unlicensed driving

behaviour. In particular, the data were used to compare the explanatory power of a

number of different perspectives relevant to unlicensed driving, including: deterrence

theory, social learning theory, sensation seeking and alcohol misuse. Earlier versions of

both Chapters Five and Six were incorporated into the afore-mentioned report to the

Australian Transport Safety Bureau (ATSB).

Chapter Seven draws the findings of the three studies together and discusses the

theoretical and practical implications for managing unlicensed driving. The limitations of

the research are also discussed, along with suggestions for future research.


1.8 Chapter summary

This chapter has provided a brief overview of the unlicensed driving problem and

discussed some of the empirical and theoretical issues relevant to understanding the

behaviour. To date, limited research has been conducted into unlicensed driving, both

within Australia and internationally. Furthermore, research in the area has been plagued

by methodological problems associated with studying illegal behaviours. In particular,

many of the self-report studies conducted in the past have featured relatively low

response rates and have lacked a strong theoretical foundation, constraining the

generalisability of the findings. Accordingly, there is a need to obtain a better

understanding of the scope and nature of the unlicensed driving. Research in this area is

required to underpin the development and implementation of more effective

countermeasures.

The foundations for the current research program will be laid in Chapters Two and

Three, which will review the empirical and theoretical research literature relevant to the

study of unlicensed driving. The remaining chapters will present and discuss the findings

of three specific studies undertaken to improve the existing body of evidence relating to

unlicensed drivers and their on-road behaviour.



Chapter Two: Literature review

2.1 Introductory comments ............................................................................... 15

2.2 The prevalence of unlicensed driving ......................................................... 15

2.2.1 Roadside licence check surveys .......................................................... 15

2.2.2 Observa tional studies .......................................................................... 16

2.2.3 Self-report surveys .............................................................................. 17

2.2.3.1 Disqualified/suspended drivers ................................................... 17

2.2.3.2 Other unlicensed drivers.............................................................. 18

2.2.4 Crash involvement of unlicensed drivers............................................ 18

2.3 The driving behaviour of unlicensed drivers ............................................. 20

2.3.1 The disqualified driver effect .............................................................. 20

2.3.2 Self-reported driving behaviour .......................................................... 20

2.3.3 The behaviour of unlicensed drivers involved in crashes ................... 21

2.4 Differences among unlicensed drivers........................................................ 23

2.5 Factors contributing to unlicensed driving ................................................. 24

2.6 Countermeasures to unlicensed driving ...................................................... 26

2.6.1 Administrative and punishment processes .......................................... 26

2.6.2 Police enforcement practices .............................................................. 28

2.6.3 Restricted licences............................................................................... 29

2.6.4 Vehicle-based sanctions ...................................................................... 30

2.6.4.1 Alcohol ignition interlocks.......................................................... 30

2.6.4.2 Licence plate sanctions ................................................................ 31

2.6.4.3 Vehicle immobilisation, impoundment and forfeiture ................ 31

2.6.4.4 Electronic licences....................................................................... 33

2.6.5 Rehabilitation of offenders.................................................................. 34

2.6.6 Mass media campaigns ....................................................................... 35

2.6.7 The likely road safety benefits of unlicensed driving measures ......... 35

2.7 Research questions ...................................................................................... 36

2.8 Chapter summary........................................................................................ 39




The following chapter will review the available literature relating to the extent and

nature of unlicensed driving and the effectiveness of relevant countermeasures. Among

the key questions that will be explored are: how prevalent is unlicensed driving; is

unlicensed driving associated with higher levels of risk-taking on the roads and, if so,

does this increase the crash risk of offenders; are there differences in the characteristics

and on-road behaviour of drivers unlicensed for different reasons; what personal, social

and environmental factors contribute to unlicensed driving; and how effective are current

countermeasures in reducing unlicensed driving?

The focus of the chapter will be on consolidating the available research evidence

and identifying gaps in current knowledge relating to unlicensed driving. In this way, it

will lay a foundation for the subsequent program of research reported in this thesis.

2.2 The prevalence of unlicensed driving

It has proven difficult for road safety authorities to reliably estimate the

community-wide prevalence of unlicensed driving. Given that it is an illegal behaviour, it

is likely that some unlicensed drivers will attempt to conceal their actions from the

authorities and be reluctant to discuss their behaviour with researchers. As a

consequence, there is a lack of definitive evidence available relating to the extent and

nature of unlicensed driving. The following section reviews the data currently available

from both Australian and international studies.

2.2.1 Roadside licence check surveys

Due to various legal and logistical constraints, very few roadside licence check

surveys have been conducted in Australia or elsewhere in the world. Among these

constraints are difficulties associated with randomly sampling drivers and confirming the

validity of licences (Watson, 1998b).

In 1976, the police in Western Australia conducted a survey of 392 motorcyclists

that were intercepted as part of a weekend blitz on motorcycle safety. Based on

interviews and police radio checks, it was found that 12% of the riders were unlicensed

(Smith, 1976). However, this result should be treated with some caution due to: the

relatively small sample; the difficulty the police experienced in identifying the exact

licence status of some riders; and the fact that all riders did not have an equal probability

of being intercepted and interviewed (Smith, 1976).


In 1991, the New South Wales Police Service conducted a survey of unlicensed

driving in the northern suburbs of Sydney, as part of routine RBT operations (Carseldine,

Court & Graham, 1992). It was found that 2.4% of the 4,352 drivers surveyed did not

have a current valid driver's licence. The random nature of RBT should have made the

sample more representative of the general driving population than would be achieved

through a targeted enforcement approach. However, the researchers acknowledged the

difficulty in generalising their results beyond the area and times surveyed. In particular,

they noted that there was a lower representation of unlicensed drivers in crashes

occurring within the survey area than in other parts of the state (Carseldine et al, 1992).

More recently, Malenfant, Van Houton and Jonah (2002) conducted a roadside

survey of suspended drivers in the Greater Moncton area of New Brunswick, Canada.

This study found that 1.53% of the drivers stopped at roadside checkpoints (who resided

within the study area) were driving while suspended. At the same time, official records

showed that 2.7% of all drivers in the study area were suspended from driving. In other

words, the relative exposure of suspended drivers in the study area was 57% of that

which would be expected if all suspended drivers continued to drive. While these results

tend to confirm a general reduction in exposure among suspended drivers, the study

identified a number of concerns. Firstly, although the percentage of suspended drivers on

the road was higher after midnight, the police preferred to perform the checks during

daylight hours. Secondly, 91% of the suspended drivers identified in the study produced

an invalid driving permit at the time they were pulled over. This finding draws into

question the practice in New Brunswick of requiring suspended drivers to voluntarily

surrender their permits upon receiving notification in the mail. In particular, “there are no

provisions to contact suspended drivers who fail to surrender their driving permits”

(Malenfant et al, 2002, p.442). Finally, the official records indicated that the suspended

drivers in the study had four times more reportable crashes in the preceding five years,

than the average (licensed) New Brunswick driver.

2.2.2 Observational studies

An innovative observational study of suspended drivers was recently undertaken by

McCartt, Geary and Berning (2003). Surveillance professionals were recruited to

undertake systematic, unobtrusive observations of first-time drink driving offenders who

had been suspended from driving in two jurisdictions within the USA. The observations

were conducted for two four-hour periods during suspension and another two periods

after licence reinstatement. In total, 57 drivers were observed in the City of Milwaukee,


Wisconsin and 36 in Bergen County, New Jersey. The results indicated that the

prevalence of suspended driving was high, but varied substantially across the two

jurisdictions. For example, 88% of the Milwaukee subjects were observed driving

compared with 36% of the Bergen County drivers. Furthermore, only 5% of the

Milwaukee drivers renewed their licences compared with 78% of the Bergen County

drivers.

McCartt, Geary and Berning (2003) suggest that the findings may be partly

attributable to the tougher laws relating to suspended driving in New Jersey. In particular,

first time offenders in New Jersey received a mandatory hard suspension with no

provision for a restricted work licence (unlike Wisconsin where work licences were

available for eligible offenders). Substantial mandatory penalties for driving-while-

suspended and driving-while- impaired were also in place in New Jersey.

2.2.3 Self-report surveys

Due to the difficulties involved in conducting roadside licence checks and the costs

involved in surveillance, road safety authorities have generally preferred to rely on self-

reported survey methods to obtain data about the extent and nature of unlicensed driving.

For example, a number of studies have used official records to locate unlicensed drivers

and survey them by interview or self-administered mail questionnaires.

2.2.3.1 Disqualified/suspended drivers

The majority of the self-report surveys undertaken to date have focused on

disqualified/suspended drivers. These surveys suggest that driving while disqualified is

relatively common. For example, Australian surveys conducted in Victoria (Robinson,

1977) and Western Australia (Smith & Maisey, 1990) found that over 30% of

respondents admitted driving while disqualified. Similar surveys in the United Kingdom

(Mirrlees-Black, 1993; Corbett & Simon, 1992; Davies, Broughton, Clayton & Tunbridge,

1999) and the United States (Williams et al, 1984; Ross & Gonzales, 1988) have found

self-reported levels of disqualified/suspended driving ranging from 11% to almost 70%.

However, as noted by Davies et al (1999, p.18): “Naturally, respondents may well be

reluctant to report candidly the extent to which they have driven while disqualified”.

Moreover, these surveys tend to be characterised by low response rates and probably

under-estimate the full extent of the problem. For example, the interview surveys

conducted by Robinson (1977) and Mirrlees-Black (1993) obtained response rates of 23%

and 47% respectively, while the mail surveys conducted by Robinson (1977), Smith and


Maisey (1990) and Corbett and Simon (1992) had response rates of below 40%. These

low response rates reduce the likely representativeness of the data.

2.2.3.2 Other unlicensed drivers

One of the few self-report surveys conducted with a variety of unlicensed drivers is

reported by Job et al (1994). This study involved the distribution of a mail survey to all

drivers convicted of unlicensed driving in New South Wales during a six-month period in

1992-93. While many of the respondents indicated that they drove regularly while

unlicensed, the overall amount of driving undertaken across the sample was significantly

reduced compared to earlier periods of legal driving. However, similar to other studies,

this survey had a relatively low response rate of 19.6%. In part, this was due to the fact

that many of the drivers no longer resided at the address recorded in the official licensing

database. This finding is consistent with other studies (eg. Robinson, 1977; Mirrlees-

Black, 1993; Davies et al, 1999) and suggests that drivers convicted of unlicensed driving

are a relatively transient group, possibly reflecting a lack of social stability in their lives

(Mirrlees-Black, 1993; Job et al, 1994). In addition, Job et al (1994) suggest that there

may be a systematic respondent bias in these studies toward less serious offenders. This

could lead to an under-estimation of the extent and seriousness of the unlicensed driving

problem.

2.2.4 Crash involvement of unlicensed drivers

Due to the difficulties involved in surveying unlicensed drivers, road safety

authorities have tended to rely on the use of crash data as a surrogate measure of

unlicensed driving. In one of the first studies of this kind, Coppin and Van Oldenbeek

(1965) examined the crash and offence records of over 1,300 negligent drivers who had

their licences suspended or revoked as a result of offences committed in late 1955 and

early 1956. The records indicated that at least 33% of those who were suspended and

68% of those who were revoked drove during the sanction period.

More recently in the United States, Griffin and DeLaZerda (2000) examined the

involvement of unlicensed drivers in fatal crashes using five years of data (1993-97) from

the Fatal Accident Reporting System (FARS) database. This study found that 11.1% of

drivers involved in fatal crashes were unlicensed (and a further 2.7% were of unknown

licence status). In addition, 16.3% of fatal crashes involved at least one unlicensed driver.

These crashes resulted in over 34,000 fatalities during the period.


Similar results have been found in Australia. A study by the (then) Federal Office

of Road Safety (FORS) found that 5% of the drivers and 19% of the motorcyclists

involved in fatal crashes during 1992 and 1994 were unlicensed at the time (FORS,

1997a). Overall, the crashes involving unlicensed drivers accounted for almost 10% of

the national road toll for the two years. Almost half of the people killed in these crashes

were road users other than the unlicensed drivers (FORS, 1997a).

However, there are a number of limitations associated with using crash data as a

surrogate measure of unlicensed driving. Firstly, in many jurisdictions the crash records

contain a relatively high proportion of drivers with an unknown licence status; some of

whom may have been unlicensed. Secondly, the proportion of unlicensed drivers is not

necessarily uniform across different crash types. Crash data from both New South Wales

(Job et al, 1994) and Queensland (Watson et al, 1996) indicate that unlicensed drivers are

over-represented in fatal and serious injury crashes compared with minor injury and

property damage only crashes. For example, during the ten-year period 1986-1995,

unlicensed drivers represented 5.8% of all the drivers and motorcycle riders involved in

fatal crashes in Queensland. However, they only represented 3.1% of the drivers and

riders involved in minor injury crashes (Watson et al, 1996).

The above data illustrate the difficulties involved in estimating the community-wide

prevalence of unlicensed driving. In particular, it is unclear which figure is more

representative of unlicensed driving in general, since “it is not known whether

unauthorised (unlicensed) drivers are under represented, proportionately represented or

overrepresented in traffic offences or crashes” (Carseldine et al, 1992, p.2). As will be

discussed in the next section, the serious casualty crashes involving unlicensed drivers

appear to involve higher levels of risk taking than those involving licensed drivers. This

suggests that unlicensed drivers may be over-represented in these crashes. However, it is

also possible that unlicensed drivers are under-represented in minor crashes, because it is

easier for them to avoid reporting these crashes to the police than serious crashes

involving injury or major damage. That is, unlicensed drivers may be highly motivated to

avoid reporting crashes wherever possible, in order to avoid being punished. Indeed, it is

possible that the apparent over-representation of unlicensed drivers in serious casualty

crashes (compared with licensed drivers) is partly an artefact of the under-reporting of

minor crashes by these drivers (Watson, 1997).


2.3 The driving behaviour of unlicensed drivers

2.3.1 The disqualified driver effect

There is a common assumption in the literature that unlicensed drivers drive in a

more cautious manner to avoid detection. This assumption is based on the findings of two

types of studies. The first group of studies involve the examination of the driving offence

records of disqualified drivers and are discussed below. The second group involve self-

report surveys of drivers and are discussed in the next section.

In an often-cited article, Hurst (1980) was the first to propose the existence of a

disqualified driver effect. He suggested that disqualified and suspended drivers (and

presumably other unlicensed drivers) are rewarded for driving safely and inconspicuously

because it reduces the threat of detection. As a result, this can contribute to these drivers

learning defensive driving skills. To support his argument, Hurst (1980) cited research

that had shown improvements in the traffic offence records of disqualified drivers after

licence restoration, compared with offenders who had escaped disqualification. “The

lower crash involvement during the period of suspension may have partly, or entirely,

been due to less driving being done; but the improvement after licence restoration

suggests a learning effect” (Hurst, 1980, p.218). In a similar vein, Scopatz et al (2003)

suggest that the lower rates of drink driving reoffences among drivers who fail to have

their licence reinstated (compared with those who do become reinstated) may not

necessarily be due to these people giving up driving altogether. Rather, it may be the

product of continued driving which is less frequent and more cautious.

However, in a rejoinder to Hurst, Warren (1982) argued that the behaviour learned

while driving unlicensed may not actually be safer, but rather more oriented to avoiding

detection. He argued that the improved driving records of disqualified drivers after

restoration could be due to two other factors. Firstly, some of the reduction may be due to

offenders moving from one jurisdiction to another, in order to continue driving. Secondly,

it is possible that offenders become more skilful at avoiding apprehension, rather than

becoming safer drivers. “Indeed, their ‘safe’ driving behaviour may have deteriorated,

but this could be more than offset by improvements in their ability to avoid

apprehension” (Warren, 1982, p.170).

2.3.2 Self-reported driving behaviour

The second source of evidence relating to the disqualified driver effect is self-report

surveys. In various studies, 55% - 65% of disqualified drivers (who continue to drive)

have reported adopting strategies to reduce their risk of detection, including driving less


frequently and driving more cautiously (Williams et al, 1984; Smith & Maisey, 1990).

Similar findings were noted by Mirrlees-Black (1993):

Although disqualification had failed to keep all these offenders off the roads it was,

nevertheless, still effective as a method of restraint as the frequency of driving, and

the amount of dangerous driving, were probably reduced for the majority of those

that continued to drive (p.21).

In their survey of unlicensed drivers, Job et al (1994) found that between 40%-50%

of respondents reported driving more carefully in terms of complying with the speed

limit, traffic lights, stop signs, seat belt and drink driving laws. However, in light of

Warren’s (1982) concerns, it is difficult to assess whether these surveys provide reliable

evidence of more cautious driving, or simply reveal a pattern of behaviour more focused

on the avoidance of detection.

2.3.3 The behaviour of unlicensed drivers involved in crashes

The crash evidence draws into question the assumption that unlicensed drivers

drive in a more cautious manner, at least when compared to licensed drivers. While

unlicensed drivers may temper their driving more than they would otherwise, the crash

data suggest that they are more likely to engage in risky driving practices than licensed

drivers. As noted in section 2.2.4, Australian evidence indicates that unlicensed drivers

are over-represented in serious casualty crashes. The exact reasons for this over-

representation remain unclear. While it may in part be due to the under-reporting of

minor crashes by this group, the crash evidence suggests that it may also be a product of

differences in the behaviour of unlicensed drivers. In particular, there is a growing body

of evidence linking unlicensed driving to a cluster of high-risk behaviours, including

drink driving, speeding, failure to wear seat belts and motorcycle use (Healy & Harrison,

1986; FORS, 1997b; Harrison, 1997; Watson, 1997 & 2000, Griffin & DeLaZerda,

2000).

For example, a Victorian study found that disqualified drivers “are over-

represented in serious crashes and in crashes that suggest a pattern focused on

recreational road use and drink-driving” (Harrison, 1997, p.110). Disqualified drivers

were also over-represented in single vehicle crashes and in crashes involving loss of

control (eg. on curved section of roads). A FORS study found that unlicensed drivers and

riders involved in fatal crashes were more likely to be judged at fault for the crash and to

not wear seat belts or helmets than their licensed counterparts (FORS, 1997b). Similar

results were found by the ATSB (undated) in a study of never- licensed drivers. FORS


(1997a) also found that unlicensed driving was more prevalent among motorcycle riders

involved in fatal crashes than other drivers.

Historically, the link between motorcycling and unlicensed riding appears to be

quite persistent. A Queensland study of motorcycle crashes undertaken in the 1970s

found that over 40% of riders killed (where licence status was known) were effectively

unlicensed (Beggs & Siskind, 1978). After excluding overseas riders and the unknowns,

the present study indicates that 20% of the motorcycle riders involved in fatal crashes in

Queensland remain unlicensed.

In the USA, Griffin and DeLaZerda (2000) found that alcohol use was much more

common among unlicensed drivers involved in fatal crashes than licensed drivers. In

particular, 74.1% of the revoked driver and 50.7% of the expired licence holders involved

in fatal crashes had been drinking alcohol compared with 19.9% of the licensed drivers.

As noted earlier, Malenfant et al (2002) found that the suspended drivers detected in their

study (at roadside checkpoints) had four times more reportable crashes in the preceding

five years, than the average New Brunswick driver. Other data indicate that suspended

drivers in that province are over- involved in fatal crashes during their suspension period.

While the above data suggest that unlicensed drivers engage in more risky

behaviour than licensed drivers, it does not necessarily confirm that they have a higher

crash risk. This is because the crash data does not take into account possible differences

in the exposure of unlicensed drivers (Silcock, 2000). Given the difficulty in obtaining

accurate exposure data, DeYoung, Peck and Helander (1997) used a quasi- induced

exposure procedure to estimate the exposure and subsequent fatal crash rates of

unlicensed drivers in California. This procedure involves dividing the proportion of at-

fault drivers in a particular group by the proportion of innocent drivers to calculate an

estimated crash rate. It is based on the assumption that the proportion of crash- involved

innocent drivers should be indicative of their overall representation in the driving

population. While this approach has a number of limitations it provides a means to

compare different groups of drivers. Based on the method, DeYoung et al (1997)

estimated that suspended/revoked drivers and other unlicensed drivers were over-

involved in fatal crashes by a factor of 3.7:1 and 4.9:1, respectively, compared to licensed

drivers. (Further information about the quasi- induced exposure method is provided in

section 4.3.2.)

There are a number of possible explanations for the apparent contradiction between

the crash data and the self- reported behaviour of unlicensed drivers. Firstly, as noted

earlier, most of the self-report surveys feature relatively low response rates. As such, it is


possible that crash-involved unlicensed drivers tend to represent a sub-group who don’t

typically respond to self-report surveys. Secondly, some of the unlicensed drivers who

report driving more safely in general, may still engage in risky behaviours on some

occasions. Finally, it is possible that some of the behaviours that unlicensed drivers adopt

to evade detection may actually increase their crash risk. As noted by Job et al (1994, p.12):

“some behaviours such as driving only at night or using back streets rather than main

streets may reduce safety”. Further research is required to establish whether the behaviour

of crash-involved unlicensed drivers is indicative of unlicensed drivers as a whole, or

whether they represent a sub-group who are less concerned about the risks of detection and

punishment and may be less cautious in general.

2.4 Differences among unlicensed drivers

A driver may be unlicensed for a wide variety of reasons, ranging from those who

have inadvertently let their licence expire through to those who have never held a licence.

As such, the motives for driving without a valid licence and associated driving

behaviours may vary greatly, depending on the circumstances. Interestingly, some

research has found similarities among unlicensed drivers involved in crashes. Compared

with licensed drivers, unlicensed drivers involved in fatal crashes are more likely to be:

male; younger in age; motorcyclists; unemployed and, to a lesser extent, students or blue

collar workers; and involved in crashes in remote, rural areas (FORS, 1997a).

However, other research suggests that unlicensed drivers do not necessarily

represent a homogeneous group. Job et al (1994) found that those drivers who remained

cancelled or suspended at the time of their survey were more likely to be male, with less

education and to report less care when driving unlicensed than those drivers who were

eligible to obtain a licence. In addition, they found that those drivers with expired licences

and interstate licences were more likely to be unaware of their unlicensed status than other

offenders.

Griffin and DelaZerda (2000) also found differences among the unlicensed drivers

involved in fatal crashes in the USA. They compared the socio-demographic characteristics

and crash involvement patterns of the different groups of unlicensed drivers with valid

licence holders. The evidence suggested that the revoked drivers were the most divergent

from licensed drivers, with them being more likely to be under the age of 404, to be male

and to have three or more previous suspensions or revocations. In terms of their fatal crash


involvement, the revoked drivers were more likely to be driving while intoxicated, riding a

motorcycle, be involved in a single-vehicle crash and to crash at night time. The differences

among unlicensed drivers were particularly evident in the driving while intoxicated data.

After excluding the cases where alcohol use (in the opinion of the investigating officer) was

not reported or unknown, the percentage of drivers under the influence was 74.1% for the

revoked drivers, 56.7% for the suspended drivers, 50.7% for the expired drivers, 31.1% for

other unlicensed drivers and 19.9% for the valid licence holders.

The above findings are consistent with other evidence that some offenders,

particularly those who have previously lost their licence for drink driving, may have a

drinking problem. Persistent drink driving offenders tend to display numerous

psychological and behavioural characteristics that distinguish them from the general

driving population, including higher levels of aggression, hostility and sensation seeking

(Hedlund & Fell, 1995; Mayhew, Simpson & Beirness, 1997; see section 3.5 for a more

detailed discussion of this issue). In a similar vein, a recent Australian study examined

the driving behaviour of a sample of people who use illicit opiates, stimulants and

cannabis. The study involved a combination of focus groups (36 participants) and field

surveys (160 participants). It found that nearly 10% of the total sample was driving

unlicensed at the time (Aitken, Kerger & Crofts, 2000). Given the overlap between

unlicensed driving and alcohol and drug use, it is possible that drug misuse exerts a major

influence on the behaviour of some offenders.

Therefore, despite some similarities, the evidence suggests that unlicensed drivers

do not necessarily represent a uniform group. This highlights the need for research in the

area to examine both common and different factors contributing to unlicensed driving

among different groups of drivers.

2.5 Factors contributing to unlicensed driving

In order to better understand the factors contributing to unlicensed driving, it is

essential to obtain an insight into the motivations, attitudes and perceptions of offenders,

both in general and among different groups. This type of information is best obtained

through the use of focus groups, interviews or self-administered surveys.

4. Evans (2000) reports that the proportion of fatally injured drivers in the USA without a valid licence

declines from about the age of 30, but increases again among drivers in their 80s. He suggests that this is due to older drivers continuing to drive after their licences are revoked.


As noted earlier, a number of surveys conducted in Australia, the USA and the UK

have explored the factors contributing to unlicensed driving (albeit with low response

rates). The reasons most frequently cited by respondents have related to business or

employment commitments, family or social reasons and lack of public transport (Robinson,

1977; Ross & Gonzales, 1988; Smith & Maisey, 1990; Mirrlees-Black, 1993; Job et al,

1994; Corbett & Simon, 1992). A Victorian survey of over 1550 disqualified drivers

found that among those who continued to drive, many reported doing so only in

exceptional circumstances. Nevertheless, their responses suggested that "a considerable

number of exceptional circumstances presented themselves" (Robinson, 1977, p.75). While

employment reasons were most frequently cited in some surveys (Robinson, 1977; Ross &

Gonzales, 1988), family and social reasons were given equal or greater weight in others

(Smith & Maisey; Job et al, 1994; Corbett & Simon, 1992). Ross and Gonzales (1988)

found that driving while disqualified was more prevalent among those who were employed

and worked far from home, those who lived in households without another licence holder

and those who had access to a vehicle. A number of other studies have suggested that

owning a vehicle (or at least having easy access to one) represents a major temptation to

offenders (Mirrlees-Black, 1993; Corbett & Simon, 1992; Williamson, 1996).

While the majority of the surveys conducted to date have focused on disqualified or

suspended drivers, Job et al (1994) surveyed a cross-section of convicted unlicensed

drivers. In this NSW study, almost 30% of the 877 respondents reported that they were

not aware that they were unlicensed when they were detected. Not surprisingly, this was

most common among drivers whose licence had expired and those who held interstate

licences. Job et al (1994) also found a very low awareness of the penalty for unlicensed

driving among those offenders who admitted driving. However, over half of these people

reported that they would have probably driven even if they had known the penalty.

A more recent, smaller-scale survey of 50 unlicensed drivers in the UK found that

the respondents were more likely to agree with statements supporting aggressive driving

styles and had a high opinion of their driving skills (Silcock, Silcock, Sunter, van Lottum

& Beuret, 1999). The study found some evidence that difficulties involved in passing the

driving theory test and the costs involved with learning to drive could discourage

participation in the licensing system. In addition, the study highlighted the link between

unlicensed and uninsured driving. “The fact that with unlicensed driving the insurance

would be invalid did not seem to occur to most young people. Others were concerned

about this aspect and it was one of the major motives for eventually taking a driving test”

(Silcock et al, 1999, p.13).


Over and above these issues, researchers have suggested that a major contributing

factor to unlicensed driving is the low perceived risk of apprehension for the offence

(Nichols & Ross, 1990; Ross, 1991; Watson et al, 1996). A common perception among

both offenders and the general community is that you are unlikely to be caught for

unlicensed driving unless you draw attention to yourself (Ross, 1991; Mirrlees-Black,

1993; Watson et al, 1996). This reflects the difficulties that police have historically

experienced in identifying unlicensed drivers (see section 2.6.2).

While the studies undertaken to date provide some insight into the factors

contributing to unlicensed driving, they feature some common shortcomings. Firstly, as

already noted, most are characterised by relatively low response rates. Secondly, the

surveys have mainly focused on disqualified drivers, rather than unlicensed drivers in

general. Thirdly, most of the surveys are largely descriptive in nature providing little

insight into the underlying personal, social and environmental factors contributing to

unlicensed driving. As a result, it is generally unclear why some people find employment,

family or social reasons compelling enough to warrant driving without a valid licence.

This problem is symptomatic of a lack of theory being used to guide the research process.

As will be argued in the next chapter, a stronger theoretical framework is required for

research in this area to facilitate the broader interpretation of findings.

2.6 Countermeasures to unlicensed driving

A range of countermeasures are currently used in Australia and elsewhere in the

world to reduce the incidence of unlicensed driving. While some of these are specific to

unlicensed drivers, others target broader problem groups such as recidivist drink driving

offenders (who are often unlicensed). The following section reviews the available

evidence relating to these countermeasures.

2.6.1 Administrative and judicial processes

As noted earlier, Malenfant et al (2002) found that 91% of the suspended drivers

detected at roadside checkpoints in New Brunswick, Canada produced an invalid driving

permit when they were pulled over. This finding draws into question the practice of

requiring suspended drivers to voluntarily surrender their permits, as is the case in New

Brunswick. In this regard, Ross and Gonzales (1988) argue that it is important for

licensing authorities to recover the licences of those offenders who are disqualified from

driving. This should reduce the temptation among these people to attempt to evade

detection, even if the police intercept them.


The current practice in many parts of Australia, including Queensland, is to require

those offenders who are processed at court to surrender their licence at the time.

However, those offenders who lose their licence administratively (eg. for accumulation of

demerit points) are typically advised by mail to surrender their licence to a government

office. More recently, the policy in Queensland was changed for those drivers who lose

their licence for the accumulation of demerit points.5 These drivers are now advised by

mail that their licence has been suspended, but are no longer required to surrender it.

Little research appears to have been conducted into the effectiveness of traditional

penalties for unlicensed driving. In general, the available evidence suggests that very high

fines do not necessarily improve compliance with road rules, if the public perceive that

there is a low risk of apprehension (Nichols & Ross, 1990). At a minimum, however,

penalties for unlicensed driving should exceed the costs associated with participating in

the licensing system, otherwise the risk of remaining unlicensed may appear more

attractive (Watson, 1998a; Travelsafe, 1998).

Both Mirrlees-Black (1993) and Job et al (1994) have argued that improved

knowledge about the consequences of unlicensed driving could improve compliance.

Low cost measures in this area include ensuring that adequate information about penalties

is included in warning/notification letters about licence loss (Job et al, 1994) and that

offenders are warned about the consequences of non-compliance when they are sentenced

to disqualification in court (Mirrlees-Black, 1993).

It is also possible that some licensing processes may inadvertently encourage

unlicensed driving. For example, Job et al (1994) found some evidence that the cost of

rider training courses in New South Wales was contributing to the level of unlicensed

riding among motorcyclists. Similarly, Silcock et al (1999) report that the introduction of

a new written theory test in the UK appeared to have led to an increase in unlicensed

driving. This highlights the need for licensing authorities to avoid erecting barriers that

may discourage licensing, such as more complex driving tests or costly licensing

processes (Watson et al, 1996). In addition, a number of Australian jurisdictions,

including Queensland, have started cancelling driver's licences for non-driving related

offences, like non-payment of parking fines (eg. Carseldine et al, 1992). While this

approach offers certain administrative efficiencies, it has the potential to significantly

inflate the numbers of people driving unlicensed.

5. This policy change came into effect in December 2001, during the period in which this research was

being conducted.


2.6.2 Police enforcement practices

As already noted, a major contributing factor to unlicensed driving appears to be

the low perceived risk of apprehension associated with the offence (Nichols & Ross,

1990; Ross, 1991; Watson et al, 1996). This largely reflects the difficulties that police

have historically experienced in detecting unlicensed drivers.

The first difficulty for the police is that, in practice, many jurisdictions don’t

require drivers to carry their licence. For example, while the police have the power to

randomly check licences in Queensland, only learner and provisional licence holders are

actually required to carry their licence. Open licence holders are given a grace period of

48 hours to present their licence to a police station (Travelsafe, 1998). As a result, the

police are generally reluctant to systematically check licences. Within Australia, New

South Wales is the only jurisdiction that currently requires all drivers to carry their

licence, which facilitates the checking of licences at RBT operations in that state (Watson

et al, 1996). Consequently, many researchers have called for the introduction of

compulsory carriage of licence throughout Australia and the more widespread checking

of driver’s licences (Job et al, 1994; Watson et al, 1996; Staysafe, 1997; Watson, 1998d).

Indeed, the introduction of compulsory carriage of licence for all licence holders was an

action included in Queensland Transport’s Driver Safety and Education Strategy (Watson

et al, 1996, p.187).

The second difficulty faced by the police is confirming the validity of a licence,

even when it is carried by a driver. For example, without some means of checking the

available records, it can be difficult for the police to identify fraudulent licences or cases

where the licence has been cancelled or suspended. Consequently, researchers have

repeatedly noted the need to improve the roadside technology used by police to ensure

the rapid identification of drivers who are unlicensed (eg. Smith, 1976; Job et al, 1994).

Over recent years, some of the jurisdictions in Australia have responded to this

problem. In Queensland, for example, a computer system has been developed that

provides operational police with a computer link to Queensland Transport’s licensing and

registration databases, resulting in a 15 second turnaround time for inquiries. The

development of the system was prompted by the time delay experienced by police (15

minutes on average) when making roadside inquiries via their radio-based system.6

Evaluations of the system indicate that it has led to a substantial reduction in average

6. This system is known as Mobile Intelligence Data Access (MINDA) and was developed jointly by the

Queensland Police Service and Queensland Transport (Travelsafe, 1998).


response times and a fourfold increase in the detection of unlicensed driving, unregistered

vehicles and outstanding warrants (Watson et al, 1996).

While some important improvements have been made to police enforcement

practices, it is likely that other developments may have exacerbated the problem. For

example, the increasing reliance on traffic cameras to detect speed and red light offences

has reduced opportunities for the police to directly detain and interview drivers. This may

have contributed to a belief that it is possible to evade detection for unlicensed driving,

by persuading other drivers to accept responsibility for camera-detected offences. Over

and above this, there appears to be considerable scope to enhance the detection of

unlicensed driving.

2.6.3 Restricted licences

Job et al (1994) suggest that consideration be given to the use of restricted licences

as an alternative to full disqualification, to allow offenders to drive to and from work.

This suggestion was prompted by their finding that the most common reason cited for

unlicensed driving (among their sample of NSW offenders) was the need to undertake

employment. In Australia, only Queensland, Western Australia and the Australian Capital

Territory issue restricted licences to drink drivers on hardship grounds. In the USA,

however, the practice is far more widespread (Watson, 1998c).

Despite their intuitive appeal, the evidence suggests that the benefits of restricted

licences may be minimal. Smith and Maisey (1990) found that almost 30% of the

respondents, who had been granted a special licence, admitted driving outside the

conditions of the licence. A recent observational study undertaken by McCartt, Geary and

Berning (2003) provided some evidence that restricted licences can undermine the

specific deterrent effect of licence loss. They found that the prevalence of suspended

driving was substantially higher in Milwaukee, Wisconsin where work-related restricted

licences are available for eligible first-time offenders, than in Bergen County, New Jersey

where no such provisions exist. However, it is possible that the observed differences may

have been due to other factors, including the substantial mandatory penalties for driving-

while-suspended and driving-while- impaired in place in New Jersey.

Research undertaken in Queensland with drink driving offenders indicates that,

while restricted licences appear to reduce drink driving recidivism on a par with full

disqualification, they do not deliver the same reductions in overall offences and crashes

(Watson & Siskind, 1997; Watson, Siskind & King, 2000). Furthermore, it is possible

that the use of restricted licences may actually undermine the general deterrent effect of


licence disqualification, by creating the impression that licence loss is neither certain or

inevitable.

2.6.4 Vehicle-based sanctions

As noted earlier, a number of studies have found that unlicensed driving is more

prevalent among those people who have ready access to a vehicle (Ross & Conzales,

1988; Mirlees-Black, 1993; Williamson, 1996). Consistent with these findings, a number

of vehicle-based sanctions are increasingly being used throughout the world to reduce the

opportunity for offenders to drive without a valid licence. Typically, vehicle-based

sanctions are used to reduce recidivism among both drink drivers and

disqualified/suspended drivers. The strategy is particularly relevant for persistent

offenders, who appear undeterred by the threat of further punishment, and often drive

with high blood alcohol concentrations (BACs).

2.6.4.1 Alcohol ignition interlocks

Since the late 1980s, a number of jurisdictions in the USA and Canada have

implemented alcohol ignition interlock programs for drink driving offenders. In effect,

these devices will not allow a vehicle to be started until a breath test has been passed at a

pre-set BAC level (Watson et al, 1996). While these devices are primarily designed to

reduce drink driving recidivism, they are increasingly being used in conjunction with

shorter suspension periods (Voas, Marques, Tippetts & Beirness, 1999). As such, the use

of alcohol ignition interlocks represents an alternative sanctioning approach that, if

successful, can serve to reduce the overall level of suspended driving in the community.

The available evidence suggests that alcohol ignition interlock programs can reduce

drink driving recidivism, over and above conventional sanctions such as license

suspension, at least while the devices are fitted to the vehicles of offenders (Weinrath,

1997; Beck et al, 1997; Voas et al, 1999). However, the effect once the devices are

removed remains unclear. Indeed, there is some evidence that interlocks may only delay

recidivism, with reoffence rates returning to higher levels once they are removed

(Watson, 1998c; Voas et al, 1999). Moreover, the impact of interlocks on drink driving

recidivism is currently constrained by the low take-up rates (often below 10%) of the

devices among eligible offenders (Voas et al, 1999; Voas, Blackman, Tippetts &

Marques, 2002). In practice, most drivers prefer to receive traditional sanctions, such as

licence suspension, which in many cases are seen as less restrictive than interlocks.


In Australia, researchers have long advocated a trial of alcohol ignition interlocks

(Homel, 1988). Following trials in a number of states, interlock programs have been

established in South Australia, Queensland, Victoria and New South Wales. The

Queensland program involves the trialing of interlocks in conjunction with a

rehabilitation program, in an attempt to bring about longer-term changes in behaviour that

will persist after the interlock is removed (Sheehan et al, 2001).

2.6.4.2 Licence plate sanctions

A number of jurisdictions in the United States have implemented licence

(registration) plate sanctions for suspended drivers. In Oregon and Washington, the

police were empowered to place a zebra sticker on the registration plate of suspended

drivers when detected. If the driver was unable to show within 60 days that he/she had a

valid licence or that the vehicle was registered to another person, the vehicle's registration

was cancelled (Clayton, 1997). An evaluation of the program in Oregon suggested that it

was effective in reducing the level of moving violations and convictions for drink driving

and driving while suspended (Voas, Tippetts & Lange, 1997). Both these programs

operated for limited periods and have now lapsed (Clayton, 1997).

There is some evidence that impounding licence (registration) plates can reduce

recidivism among repeat drink driving offenders (Ross, Simon & Cleary, 1995; Clayton,

1997). This approach tends to present fewer practical difficulties to enforcement agencies

than impounding the vehicle itself (see below).

2.6.4.3 Vehicle immobilisation, impoundment and forfeiture

A number of countries have introduced legislative provisions that permit vehicle

immobilization, impoundment or forfeiture for driving under the influence of alcohol

(DUI) and/or driving while suspended (DWS) offences. Generally, impoundment

involves the vehicle being removed to a storage facility while immobilisation involves

the securing of the vehicle by a steering lock or wheel clamp (Clayton, 1997). To date, a

limited number of evaluations have been conducted of these sanctions.

In 1989, the Canadian province of Manitoba simultaneously introduced

administrative licence suspension for DUI offenders and vehicle impoundment for DWS

offenders. The impoundment period is 30 days for first offenders and 60 days for repeat

offenders. An evaluation by Beirness, Simpson and Mayhew (1997) was unable to isolate

the individual effects of the two measures. However, they did find evidence of a general

deterrent effect with a 12% reduction in alcohol- involved driver fatalities and a 26%


decrease in single vehicle night-time crashes. In addition, there appeared to be a specific

deterrent effect associated with the measures, evidenced by a 27% reduction in repeat

DWS offences within the first four years.

In 1993, the US state of Ohio enacted a law enabling the immobilization of vehicles

for DUI and DWS offences (Voas, Tippetts & Taylor, 1998). The legislation was broad in

nature allowing counties within the state to either immobilize or impound the vehicles of

offenders (30 days for the first DWS offence and 60 days for the second offence). Third

time DWS offenders were subject to vehicle forfeiture. Evaluations of both vehicle

immobilization (Franklin County) and vehicle impoundment (Hamilton County)

programs showed positive results, with recidivism rates reduced during both the sanction

period and after the vehicles were returned to offenders (Voas et al, 1998).

In early 1995, California enacted two laws to provide for the impoundment and

forfeiture of vehicles driven by drivers who were suspended/revoked or otherwise

unlicensed. The laws enabled the vehicles of offenders to be impounded for 30 days or to

be forfeited, if they had prior convictions for DUI or DWS. An evaluation of this

initiative by DeYoung (1997b) examined the impact of the laws on the subsequent

driving records of those who had their vehicles impounded. It found that first-time

offenders experienced 18.1% fewer traffic convictions and 24.7% fewer crashes in the

year following the sanction than a comparison group who hadn’t had their vehicles

impounded. Among repeat offenders the reductions were larger, with 34.2% fewer

convictions and 37.6% fewer crashes. As noted by Griffin and DeLaZerda (2000, p.31):

“Although the results of California’s vehicle impoundment program are impressive, it

should also be noted that . . . many unlicensed, suspended, and revoked drivers . . .

continued to be convicted of unlicensed driving . . . and continued to be involved in

crashes”.

More recently, Voas and DeYoung (2002) reviewed the evaluations of various

vehicle-based sanctions in California, Minnesota, New York, Ohio, Oregon and

Washington. They noted that all the programs demonstrated reductions in recidivism

associated with denying offenders the use of their vehicles for 1–6 months. At this stage,

they suggest that the evidence in favour of vehicle impoundment is more compelling than

that for license plate impoundment, license plate marking or vehicle forfeiture.

In May 1999, New Zealand introduced a range of countermeasures to reduce the

level of unlicensed driving, including photo driver licences, mandatory licence carriage

and vehicle impoundment (LTSA, undated). The vehicle impoundment provisions allow

the police to seize and impound (for 28 days) vehicles driven by drivers whose licence


has been disqualified, revoked or suspended. The provisions also apply to never licensed

drivers and those with expired licences, as long as they have previously been forbidden to

drive by the police until they obtain a valid licence. Because the three measures were

introduced together it has proven impossible to disentangle their individual effects.

Nonetheless, the preliminary evidence is encouraging. The proportion of crash involved

drivers who were disqualified or unlicensed decreased by approximately two percentage

points following the introduction of the measures. In addition, there was an overall 38%

reduction in the number of disqualified driving offences detected by the police in the first

three years of the new provisions. Interestingly, however, the number of vehicle

impoundments remained relatively high during the period. While the exact reasons for

this are unclear, it appears that an increasing number of unlicensed drivers were being

detected (LTSA, undated).

While the emerging evidence relating to vehicle impoundment/forfeiture appears

promising, a number of concerns have been raised about vehicle impoundment and

forfeiture. Firstly, it has been suggested that they are overly punitive in their effect on

offenders, their families and other involved persons. For example, an early evaluation of

the vehicle immobilisation program in Ohio indicated that judges sometimes failed to

apply the sanction uniformly, particularly when offenders were driving vehicles

belonging to other people (Stewart, Voas & Taylor, 1995). Secondly, it appears that

offenders can sometimes utilise a variety of strategies to avoid vehicle-based sanctions,

such as failing to advise authorities about changes in the ownership of vehicles (Clayton,

1997). As noted by Clayton (1997, p.31): “Despite the availability of these vehicle-based

sanctions, their usage appears to be low, largely because of administrative and practical

problems associated with their implementation.”

2.6.4.4 Electronic licences

In the medium to long-term, the development of electronic licences offers the

potential to prevent vehicles being operated by drivers without a valid licence. The

Swedish National Road Administration is currently field-testing an electronic driver's

licence (EDL) system (Goldberg, 1997). It features a smart-card that acts as an ignition

key to the car. The system can also be programmed to require an alcohol ignition

interlock to be operated in the vehicle. It has been claimed that: “With an alcohol

interlock in the car combined with the KitteLock electronic driving licence system it will

be very difficult, if not impossible, to bypass the system” (Goldberg, 1997, p.850). As the


technology improves, there is potential to require electronic licences to be installed in the

vehicles of recidivist offenders, and eventually in all new vehicles.

2.6.5 Rehabilitation of offenders

There is a growing body of evidence indicating that rehabilitation programs can be

effective in reducing alcohol- related offences and crashes (Wells-Parker et al, 1995;

Ferguson et al, 1999). In particular, the research suggests that the combined use of

licence actions and rehabilitation programs are most effective in reducing drink driving

recidivism among first and multiple offenders and deliver the best overall road safety

outcomes (McKnight & Voas, 1991; Peck, 1991; Sadler, Perrine & Peck, 1991;

DeYoung, 1997a). By addressing the factors underlying alcohol misuse, these programs

offer the potential to reduce the compulsion to drive illegally among some offenders.

Since 1993, a rehabilitation program for drink drivers, known as Under the Limit

(UTL), has been operating in Queensland. This program is implemented through the

court system, in order to complement the probationary orders and licence disqualification

applied to offenders. An evaluation indicated that completion of the program reduced the

risk of reoffence by 30% among repeat offenders (Sheehan et al, 1999). However, there

was a group of offenders with exceptionally high recidivism rates who were less likely to

complete the program. The researchers suggested that this is a group requiring additional

interventions or controls.

In addition, there may be value in developing rehabilitation programs specifically

targeting disqualified drivers. Bakker et al (1997) argue that some persistent disqualified

driving offenders experience a compulsion to drive that represents a maladaptive

response to stressful life events, such as interpersonal conflict. They suggest that the

problem with these offenders is often not related to alcohol per se, but is indicative of a

lack of self-control in their driving behaviour. Consequently, Bakker et al (1997) have

developed and trialed a relapse-prevention program for driving-while-disqualified

(DWD) offenders, which involves teaching more effective ways of solving interpersonal

problems and regulating negative emotional states. An evaluation of this intervention has

produced promising results (Bakker, Hudson, & Ward, 2000). The program reduced

DWD recidivism rates among a group (n=144) of convicted offenders, compared with a

matched comparison group. While the program did not affect subsequent drink driving

offences, there were some indications that subsequent criminal convictions were reduced.

Besides confirming the potential benefits of rehabilitation, the results suggest that

recidivist disqualified drivers do represent a distinct subgroup of driving offenders.


The Bakker et al (2000) study utilised participants from both the general

community and the prison population within New Zealand. The implementation of

traffic-oriented rehabilitation programs also appears relevant for prison populations in

Australia. A Victorian study examining the conviction records for a sample (n=105) of

reception prisoners found that 37% had outstanding drink driving convictions while 80%

were unlicensed (Marshall, undated).

2.6.6 Mass media campaigns

A far more controversial issue concerns the likely effectiveness of using mass

media publicity to heighten the perceived risk of apprehension for unlicensed driving.

Homel (1988) has argued that mass-media publicity can reinforce and heighten the

impact of RBT operations, contributing to their deterrent impact. However, Job et al

(1994, p.59) have argued against the use of mass media campaigns to target unlicensed

driving, due to a concern that it “may promote a perception that many unauthorised

drivers go undetected or it may simply raise awareness of the possibility”.

Elliott (1992) argues that mass media publicity in isolation is rarely sufficient to alter

entrenched behaviours. However, it can contribute to improved road user behaviour by

creating a climate of opinion supportive of other measures and by signposting the need for

behaviour change. This suggests that mass media publicity relating to unlicensed driving

should only be considered if real improvements are made in the probability of detection.

2.6.7 The likely road safety benefits of unlicensed driving countermeasures

As noted in section 2.3.3, unlicensed driving appears to be associated with a cluster

of high-risk behaviours, such as drink driving and speeding. However, it should be

acknowledged that it is not the cause of these behaviours. Hence, there is no guarantee

that unlicensed drivers involved in crashes would be any safer if they were actually

licensed. In other words, these drivers may still engage in higher levels of risk-taking,

irrespective of their licence status. Nonetheless, it is likely that more effective

countermeasures to unlicensed driving should have a positive effect on road safety by:

§ encouraging drivers who have never been licensed to participate in the licensing

system and thus be subject to processes such as graduated licensing and demerit point

systems;

§ deterring people from driving vehicles for which they do not have an appropriate

class of licence;


§ reducing the level of disqualified or suspended driving, thereby improving the

deterrent impact of these sanctions; and

§ exposing persistent offenders to rehabilitation programs that may assist them to

resolve the personal or social factors associated with their behaviour (Watson,

1998d).

2.7 Research questions

The preceding review of the literature indicates that there are some important

questions relating to unlicensed driving which remain unanswered. The main research

questions requiring further attention are discussed below.

1. Do unlicensed drivers engage in more risky driving than other drivers?

There is a growing body of evidence linking unlicensed driving with certain high-

risk road user behaviours, including drink driving, speeding, failure to wear seat

belts and motorcycle use. Consistent with this, the crashes involving unlicensed

drivers tend to be more severe than those involving licensed drivers resulting in

higher fatality and injury rates. However, further research is required to establish

the extent of these risk-taking behaviours and their prevalence among different sub-

groups of unlicensed drivers. In addition, the majority of the research undertaken to

date has focused on the driving behaviour of crash-involved unlicensed drivers. It is

critical to obtain a broader understanding of the on-road behaviours of unlicensed

drivers, particularly those not involved in crashes, in order to design and target

more effective countermeasures.

2. Is unlicensed driving associated with a higher crash risk compared to legal

driving?

There is a common assumption in the literature that unlicensed drivers drive in a

more caut ious manner than they would otherwise, in order to avoid detection. This

view, sometimes referred to as the disqualified driver effect, is based on studies that

have found reduced reoffence rates among drivers after periods of licence

disqualification and is largely consistent with the self- reported behaviour of

offenders. From a policy perspective, this assumption has sometimes been used as a

rationale to avoid devoting more resources to the problem of unlicensed driving or

to justify policies that may inadvertently exacerbate the problem. However, the

validity of the ‘disqualified driver effect’ has been strongly questioned by other

researchers who argue that the avoidance of detection does not necessarily result in


safer driving. This position is consistent with the growing body of evidence linking

unlicensed driving to unsafe driving behaviours, like drink driving and speeding.

Hence, there is a need for further research to resolve the apparent contradiction in

the evidence. In particular, it is essential to establish whether unlicensed driving is

associated with any higher crash risk than legal driving. The resolution of this issue

is critical to determine the appropriate level of resources that should be devoted to

unlicensed driving.

3. Do unlicensed drivers represent a homogeneous group, in terms of their

psychosocial characteristics and on-road behaviour?

Given the varied nature of the traffic offences involved in unlicensed driving, it is

reasonable to assume that there may be differences among offenders. To date,

however, there has been limited research into this issue. Consequently, it is unclear

how extensive these differences may be, if indeed they do exist. This is a critical

issue for the design of more effective countermeasures. If there are few differences

among offenders it is likely that countermeasure efforts can be broadly directed. In

contrast, if unlicensed drivers do not represent a homogeneous group in terms of

either their motives for unlicensed driving or their on-road behaviour, it is likely

that multi-strategy approaches will be required to address the problem.

4. How effective are current administrative, enforcement and punishment policies and

processes in preventing unlicensed driving?

The available evidence suggests that the level of unlicensed driving in a jurisdiction

is influenced by the prevailing driver licensing and traffic law enforcement policies.

However, more research is required to identify the particular policies that serve to

reduce the level of unlicensed driving and those that may exacerbate the problem.

As a first step, there is a need to examine the effectiveness of the current policies

and practices typically used in Australia to manage unlicensed drivers.

5. What are the personal, social and environmental factors contributing to unlicensed

driving behaviour?

In order to design more effective countermeasures there is a need to better

understand the factors that contribute to the behaviour. To date, there has been little

research into this issue despite ongoing concerns about the scale of the unlicensed

driving problem. It is important that research in this area not only consider the


effect of formal sanctions, but consider the broad range of personal, social factors

or environmental factors that may act to encourage or facilitate the behaviour.

2.8 Chapter summary

Although it is difficult to reliably estimate the extent of unlicensed driving, it

appears to represent a relatively small but significant problem for road safety. Evidence

from both Australia and the USA indicates that somewhere between 10%-20% of fatal

crashes involve at least one unlicensed driver. In addition, there is a growing body of

evidence linking unlicensed driving to a cluster of high-risk behaviours. Consequently,

there has been growing effo rts in many jurisdictions to develop new approaches to reduce

unlicensed driving, particularly among disqualified drivers. This has been prompted by

the apparent failure of traditional traffic law enforcement practices to control unlicensed

driving. A common theme that emerges in the literature is the low perceived risk of

apprehension for unlicensed driving in many jurisdictions. Some new countermeasures to

unlicensed driving appear to offer potential to reduce the problem, particularly those

involving improved enforcement practices and related technologies and the use of

vehicle-based sanctions.

While many of these new countermeasures have been evaluated, little research has

been conducted into the personal and social characteristics of unlicensed drivers. As a

consequence, little is known about the factors contributing to the behaviour or whether

unlicensed drivers represent a homogeneous group or not. Without research of this nature

it will be difficult to design more effective countermeasures. To assist in this process, this

chapter has identified five key research questions that require more attention. They will

be used to guide the current research program and inform the study hypotheses.

In addition, much of the research in this area has been descriptive and lacked a

strong theoretical foundation. This is not uncommon in road safety, despite the evidence

that psychological and sociological theory have made a major contribution to

understanding driver behaviour (Brown, 1997; Grayson, 1997). Consequently, it is

important that the current research program utilises a sound theoretical framework. To

assist in this process, the following chapter will examine a range of theoretical

perspectives relevant to unlicensed driving, which will be used to guide and inform the

research program.


Chapter Three: Theoretical perspectives on unlicensed driving

3.1 Introductory comments ................................................................................. 41

3.2 Deterrence theory.......................................................................................... 41

3.2.1 Classical deterrence theory ................................................................... 42

3.2.1.1 Origins and overview ..................................................................... 42

3.2.1.2 Relevance to unlicensed driving .................................................... 44

3.2.2 Criticisms of classical deterrence theory .............................................. 44

3.2.3 Reconceptualisations of deterrence theory ........................................... 46

3.2.4 Deterrence-based models of unlicensed driving ................................... 48

3.3 Social learning theory................................................................................... 50

3.3.1 Principles of social learning theory....................................................... 50

3.3.2 Akers’ social learning theory................................................................ 51

3.3.2.1 Origins of Akers’ theory ................................................................ 51

3.3.2.2 Processes and constructs in Akers’ theory..................................... 52

3.3.2.3 Empirical support for Akers’ theory .............................................. 55

3.3.3 Application of Akers’ theory to unlicensed driving ............................. 56

3.3.4 A social learning model of unlicensed driving ..................................... 58

3.4 Sensation seeking.......................................................................................... 59

3.5 Alcohol misuse ............................................................................................. 61

3.6 Chapter summary.......................................................................................... 62




The aim of the chapter is to develop a theoretical framework for examining the

psychosocia l characteristics of unlicensed drivers and the related factors contributing to

their behaviour. This is critical for informing the research hypotheses and to provide an

analytical framework for interpreting the significance of the subsequent findings. As

noted by (Grayson, 1995, p.95, citing both Helmholtz and Lewin): “there is nothing so

practical as a good theory”.

This research program will adopt a multi-disciplinary approach to theory, drawing

on perspectives from psychology, sociology and criminology. This broad approach is

required due to the spectrum of behaviour involved in unlicensed driving. As already

noted, for some offenders unlicensed driving may represent an administrative oversight

(eg. failure to renew a licence), while for others it may represent more deviant behaviour

(eg. persistent driving while disqualified).

Consistent with this broad approach, this chapter will review a variety of theoretical

perspectives that have been widely used to explain illegal or high-risk behaviours,

including driver behaviour. The first two perspectives, deterrence theory and social

learning theory, focus on the social factors influencing behaviour, particularly the role of

legal (formal) and social (informal) sanctions. The last two perspectives, sensation

seeking and alcohol misuse, are not systematic theories per se but are conceptual

frameworks for explaining how personality-related factors can predispose people to risk-

taking.

While the perspectives reviewed in this chapter will inform all aspects of the

research, they are particularly relevant to Study Three. This study will examine the

factors contributing to unlicensed driving and will directly assess the predictive utility of

the different perspectives.

3.2 Deterrence theory

Deterrence theory is a criminological perspective that has been used extensively in

Australia and other countries to guide the development of many road safety

countermeasures, particularly in the area of drink driving (eg. Ross, 1982; Homel, 1988).

It has underpinned the design of traffic law enforcement programs such as RBT and speed

cameras (Homel, 1988; Cameron, Cavallo & Gilbert, 1992; Watson et al, 1996). Indeed,

South (1998) has argued:


The reduction in the road toll . . . has arguably been the most successful example of

public action to minimise a social problem in Australia, and there is solid evidence

that general deterrence programs have played a major role (p.76).

While Harrison (1998) has questioned South’s (1998) contention that there is

‘solid’ evidence supporting the role of general deterrence, there is no doubt that

deterrence principles have played a pre-eminent role in road safety policy making.

Accordingly, deterrence theory represents an important perspective from which to

examine unlicensed driving.

3.2.1 Classical deterrence theory

3.2.1.1 Origins and overview

Deterrence theory focuses on explaining the conditions under which criminal acts

are omitted or curtailed in response to the perceived risk and fear of legal punishment

(Gibbs, 1975; Homel, 1986). The theory has its origins in the writings of the 18th century

social philosophers, Cesare Beccaria and Jeremy Bentham (Gibbs, 1977; Homel, 1986;

Akers, 1994). The traditional or classical form of this theory asserts that the effectiveness

of a legal threat is related to the perceived certainty, severity and swiftness of punishment

(Homel, 1986; Vingilis, 1990). Both Beccaria and Bentham were utilitarian philosophers

primarily concerned with penal and legal reform rather than formulating a comprehensive

theory of criminal behaviour (Akers, 1994). As a consequence, Meier and Johnson (1977)

argue that classical deterrence theory actually represents a doctrine, rather than a

systematic criminological theory.

The deterrence doctrine, as formulated within criminology, is strikingly

atheoretical both in its philosophical origins and its historic inattention to

developments in the social sciences. As Gibbs (1975: 5) notes, deterrence is more a

doctrine than theory, ‘a vague congery of ideas with no unifying factor other than

their being the legacies of two major figures in moral philosophy (p. 293).7

Although the study of deterrence theory has primarily been the domain of

criminologists and sociologists, it has strong links to psychology. As argued by Erickson,

Gibbs and Jensen (1977) deterrence theory should not only be concerned with the objective

7. The status of deterrence as a doctrine, rather than a systematic theory, is evident in the difficulties

involved in disproving its central propositions. For example, evidence that a person has not been deterred by a legal sanction can be countered by the argument that the sanction must not have been perceived as sufficiently certain, severe or swift.


properties of legal sanctions. The perceptions of individuals towards sanctions, which are

psychological phenomena, are of central importance. As such, deterrence theory is

primarily concerned with the manner in which legal sanctions deter criminal acts through

the mechanisms of fear and perceived risk. “At the heart of the arguments for deterrence as

a tool for social control is the belief that the behaviour of human beings can be modified by

making them fearful of the consequences of committing illegal acts” (Homel, 1986, p.22).

At a practical level, this means that research into the process of deterrence needs to

consider both the objective properties of legal sanctions and the perceptions of individuals

towards these sanctions (Erickson et al, 1977). At a theoretical level, it means that the other

ways in which punishment can influence behaviour, such as incapacitation, are generally

excluded from definitions of deterrence (Erickson et al, 1977; Homel, 1986).

While deterrence theory may appear to be narrowly focused on the role of legal

sanctions, an important distinction is made between specific and general deterrence

(Homel, 1986; Akers, 1994). Traditionally, specific deterrence has been conceptualised as

the process by which an offender is deterred from reoffending through direct exposure to

sanctions, while general deterrence concerns the deterring of the general community

through the threat of sanctions (Homel, 1986). Consequently, through the process of

general deterrence, it is proposed that legal sanctions have the capacity to influence

community-wide behaviour.

While a full discussion of the empirical evidence relating to deterrence theory is

beyond the scope of this thesis, it is important to note that mixed results have been

obtained from a variety of fields, including road safety. In terms of specific deterrence,

road safety research has tended to indicate that policies based on increasing the certainty

and swiftness of punishment have more frequently proved effective than those based on

increasing the severity of punishment (Nichols & Ross, 1990; Watson et al, 1996). For

example, a study by Vingilis et al (1990) comparing the specific deterrent effect of

different drink driving penalties, found that licence suspensions were consistently related

to road safety benefits. In contrast, more severe penalties in the form of higher fines (for

first offenders) and more days in jail and not being placed in a temporary absence

program (for multiple offenders) were associated with more crashes and convictions. The

evidence relating to general deterrence appears more promising. Nichols and Ross (1990,

p.52) concluded that the most effective sanctions for deterring drink driving at a

population-wide level were ‘swift and sure’ licence actions. In addition, the success of

RBT in Australia is generally attributed to its general deterrent effect, principally


achieved through increasing the perceived risk of apprehension for drink driving (Homel,

1986; Watson et al, 1996).

3.2.1.2 Relevance to unlicensed driving

Given the pre-eminent role of deterrence theory in road safety policy, it is not

surprising that it has been used to explain the prevalence of unlicensed driving. As noted

in section 2.5, researchers have suggested that the high level of unlicensed driving in

many jurisdictions is primarily a function of the low perceived risk of apprehension

(Nichols & Ross, 1990; Ross, 1991). For example, research in Queensland has suggested

that the public’s perceived risk of apprehension for drink driving and speeding is much

higher than for unlicensed driving. There appears to be a common perception that you

are unlikely to be caught for unlicensed driving if you do not draw attention to yourself

(Watson et al, 1996). As noted by Ross (1991, p. 65):

... the experience of driving while unlicensed teaches that participation in the

licensing system is unnecessary if one takes precautions in the amount and nature of

driving. Given this belief, disincentives to rejoining the system may operate to keep

drivers from seeking reentry.

According to deterrence theory, the decision to drive unlicensed should be mainly

influenced by a person's perceptions of the risk of apprehension and the certainty,

swiftness and severity of punishment. Partial support for this explanation was obtained in

a study conducted by Robinson and Kelso (1981). Utilising the responses obtained from

Robinson's (1977) survey of disqualified drivers, they found that anxiety towards

apprehension was significantly related to the decision to drive. In other words, the

respondents who admitted driving were more likely to rate their anxiety about possible

apprehension as lower than those who reported that they didn't drive. A multiple

regression analysis indicated that the strongest predictor of apprehension anxiety was the

perceived risk of apprehension. Interestingly, awareness of the penalty for disqualified

driving added little to the predictive power of the analysis. Similarly, regression analyses

undertaken by Job et al (1994) suggested that the operation of RBT and associated

licence checking in NSW was an important factor in deterring unlicensed driving.

3.2.2 Criticisms of classical deterrence theory

Classical deterrence theory has been criticised in the literature for a variety of

reasons. Firstly, it has been criticised for being too narrow in scope and failing to account

for the wide range of factors that can influence social conformity (Meier & Johnson,


1977; Vingilis, 1990). It is argued that by focusing solely on legal sanctions, deterrence

theory ignores the important role that informal sanctions (sometimes referred to as

extralegal factors) can play in preserving social control and promoting conformity. In this

regard, a number of studies comparing the statistical contributions of formal (legal) and

informal (extralegal or social) sanctions to behaviours such as marijuana use and drink

driving have found that the informal sanctions have a greater predictive power (Meier &

Johnson, 1977; Anderson, Chiricos & Waldo, 1977; Berger & Snortum, 1986).

Drawing on the work of Webb (1980) and Gibbs (1975), Vingilis (1990) identifies

a range of informal sanctions that serve to promote conformity in society, including:

§ Criminal self image – the degree to which a person perceives himself/herself as a

criminal;

§ Criminal life organisation – the extent to which a person’s life is organised around

the offending behaviour;

§ Group support – the degree to which the person is supported or reinforced by their

support group;

§ Differential association – the learning of skills and rationalisations from a criminal

referent group;

§ Moral commitment – the extent of a person’s moral commitment to the law;

§ Opportunity for crime commission – the ease with which someone can commit a

crime; and

§ Social labelling – the stigma associated with criminal activity in the community.

The second major criticism of deterrence theory is that certain types of offenders

are less likely to be influenced by formal sanctions, particularly those that commit

offences for impulsive or compulsive reasons (Vingilis, 1990). In the case of unlicensed

drivers, this is quite relevant for recidivist offenders. For example, Bakker et al (1997,

p.30) cite evidence suggesting that many disqualified drivers continue to offend, "despite

strong expectations that they will be caught". Mirrlees-Black (1993) has argued that some

disqualified drivers appear to have a compulsion to drive, which cannot be satisfied by

alternative forms of transport. Bakker et al (1997) suggest that this compulsion to drive

represents a maladaptive response to stressful life events and that a therapeutic approach

is a more appropriate method to reduce unlicensed driving offences among persistent

offenders. As noted later, social learning perspectives appear better able to explain

unlicensed driving for compulsive or impulsive reasons (see section 3.3.3).

Two further criticisms of classical deterrence theory made by Stafford and Warr

(1993) are also quite relevant to understanding illegal driving behaviour. Firstly, they


argue that deterrence theory does not adequately account for the effect of punishment

avoidance on behaviour.

To illustrate, it is possible that punishment avoidance does more to encourage

crime than punishment does to discourage it. Offenders whose experience is limited

largely to avoiding punishment may come to believe that they are immune from

punishment, even in the face of occasional evidence to the contrary (Stafford &

Warr, 1993, p.125).

Secondly, Stafford and Warr (1993) argue that it is important to not only consider

the effect of a person’s direct experience with punishment and punishment avoidance, but

also their indirect experiences obtained through contact with their peer group. To support

their argument, they draw on the distinction in social learning theory between experiential

learning (via direct experience) and observational/vicarious learning (via indirect

experience). A more detailed discussion of vicarious learning is provided in section 3.3.1.

3.2.3 Reconceptualisations of deterrence theory

Since the 1970s, more sophisticated models of deterrence have been developed and

tested by researchers. A number of these have been used to explain illegal driving

behaviours, particularly drink driving. For example, Homel (1986; 1988) developed a

model of deterrence to explain the introduction of RBT in NSW that incorporated both

legal sanctions and informal sanctions, such as peer pressure and the internalisation of

norms relating to drink driving. Drawing on rational choice theory and prospect theory,

he suggested that the decision to drive after drinking is based on an individual’s

evaluation of the potential losses involved. While Homel’s results generally supported his

deterrence model, they suggested that deterrence is an unstable process: “. . . whether a

deterrent effect is maintained or not is essentially an outcome of a delicate balance, over

time, between the forces maintaining and those tending to erode perceptions of arrest for

drinking and driving as a likely event” (Homel, 1986, p.136).

In many respects, Homel’s model of deterrence addresses the concerns raised by

researchers such as Meier and Johnson (1977) and Vingilis (1990). By incorporating

informal sanctions and normative factors, Homel conceptualises deterrence within the

broader context of social conformity. Indeed, by focusing on the potential losses

associated with drink driving, he arguably broadens the conceptualisation even further.

Homel (1986) implies that the decision to drink and drive represents an assessment or

balancing of the various outcomes that could arise from the action:


In summary, the decision whether or not to drink and drive seems best framed as a

choice between losses. There are two kinds of certain losses associated with not

drinking and driving: the costs and inconveniences entailed in finding alternative

transport, and one’s portrayal as incompetent in one’s own eyes and in the eyes of

one’s peer. On the other side of the coin, feelings of guilt, to the extent to which

they occur, may be viewed as a sure loss entailed in the decision to drink and drive

(p.32).

Indeed, the process described above by Homel appears very similar to the concept

of differential reinforcement, incorporated within Akers’ social learning theory (see

section 3.3.2). Therefore, although Homel (1986) primarily characterised his model of

drink driving as a deterrence model, it is arguably consistent with broader

conceptualisations of social conformity and has elements not dissimilar to Akers’ social

learning theory.

More recently, Stafford and Warr (1993) have proposed a reconceptualisation of

deterrence theory that incorporates both personal and vicarious experiences with

punishment, as well as the concept of punishment avoidance. Traditionally, specific

deterrence has been conceptualised as the process by which an offender is deterred from

reoffending through direct exposure to punishment, while general deterrence concerns

the deterring of the general community through the threat of punishment (Homel, 1986).

Stafford and Warr (1993) argue that specific deterrence should be reconceptualised as the

direct effect on an individual of punishment and punishment avoidance. General

deterrence can then be used as a concept to cover an individual’s indirect or vicarious

experience of these contingencies. The advantage of this perspective is that specific and

general deterrence no longer become mutually exclusive processes operating on different

populations (as is the case in classical deterrence theory), but can operate conjointly on

individuals.

Support for Stafford and Warr’s perspective was obtained by Piquero and

Paternoster (1998) in a study examining drink driving behaviour. They found that

intentions to drink and drive were affected by both personal and vicarious experiences, as

well as experience of punishment and punishment avoidance. They also found strong

effects for legal and informal sanctions. Further research by Piquero and Pogarsky (2000)

has confirmed the link between personal and vicarious experiences of punishment

avoidance and drink driving intentions. However, they found that personal and vicarious


experiences of punishment appear to encourage offending rather than deter it. (This

anomaly is further discussed in section 6.5.3.1).

3.2.4 Deterrence-based models of unlicensed driving

Figure 3.1 illustrates a model of unlicensed driving based on classical deterrence

theory. According to this model, the key determinant of an individual’s decision to drive

unlicensed or not will be their perceived risk of punishment. This will be a function of

two contingencies: their perceived risk of apprehension and the perceived certainty,

severity and swiftness of legal sanctions. These perceptions will, in turn, be shaped by a

number of key life experiences. The first of these relates to the degree to which an

individual has been exposed to traffic law enforcement processes focusing on unlicensed

driving and other similar offences. The relevance of other offences relates to the

likelihood that “people may estimate the certainty and severity of punishment for a

particular type of crime by reference to crimes in general or at least similar types of

offenses . . . rather than from information that is crime-specific” (Stafford & Warr, 1993,

p.127). This component of the deterrence process has been traditionally referred to as

specific deterrence.

Figure 3.1: Classical deterrence model of licensed/unlicensed driving The other two major influences on punishment perceptions will be ongoing

exposure to traffic law enforcement operations and knowledge about enforcement

strategies and the sanctions applied to unlicensed driving. The role of these two factors

has traditionally been identified as general deterrence, since they influence potential

offenders as well as past offenders (Stafford & Warr, 1993). The types of enforcement

Past punishment experiences - for unlicensed driving - for other similar offences

Exposure to enforcement - RBT operations - licence checks

Knowledge of enforcement practices and sanctions - media - other sources

Perceived risk of punishment for unlicensed driving - perceived risk of apprehension - perceived certainty, severity & swiftness of sanctions

Decision to drive licensed or unlicensed


operations that will have the most relevance for deterring unlicensed driving are those

that result (or could result) in drivers having their licence checked by police. These

operations would include: RBT, being pulled over for a traffic offence or being involved

in a crash. Knowledge relating to enforcement operations and sanctions is primarily

obtained from the media (eg. television advertising) or indirectly from a person’s social

network.

Figure 3.2 illustrates a more comprehensive deterrence-based model of unlicensed

driving. It incorporates the concepts proposed by Stafford and Warr (1993) and is based

on a model developed by Paternoster and Piquero (1995) to examine drink driving

behaviour.

Figure 3.2: Expanded deterrence model of licensed/unlicensed driving

……………………………………….

Personal experiences with punishment

Perceived risk of punishment for self

Personal experiences with punishment avoidance

Personal exposure to enforcement

Personal knowledge of enforcement practices and sanctions

S P E C I F I C

D E T E RR E NC E

Decision to drive licensed or unlicensed

Vicarious experiences with punishment

Vicarious exposure to enforcement

Vicarious knowledge of enforcement practices and sanctions

G E N E R A L

D E T E RR E NC E

Vicarious experiences with punishment avoidance Perceived risk of

punishment for other people


As can be seen in Figure 3.2, this model distinguishes between direct (personal) and

indirect (vicarious) influences on an individual’s perceived risk of punishment. The top

half of the model illustrates the process that Stafford and Warr (1993) refer to as specific

deterrence. This part of the model is similar to classical deterrence model, but includes

personal experiences relating to punishment avoidance. These relate to episodes where an

individual evades detection and/or punishment for unlicensed driving. In practice, all

successful episodes of unlicensed driving represent instances of punishment avo idance.

Of particular relevance to this research, however, are episodes where offenders come into

direct contact with the police (or other authorities) but manage to evade detection.

The bottom half of the model represents the general deterrence process. It covers

the various ways that an individual can learn from others about the various contingencies

associated with unlicensed driving. It will include instances where an individual learns

about:

§ the strategies that the police use to detect unlicensed driving and other illegal

driving behaviours;

§ cases where people were able to evade detection and/or punishment for unlicensed

driving;

§ strategies that can be used to reduce the likelihood of detection and/or punishment;

and

§ the sanctions that are applied to unlicensed driving offenders.

Based on these vicarious experiences, an individual will form a perception about

the likelihood or risk of others being apprehended and punished for unlicensed driving. In

turn, this will influence an individual’s perception of his or her own risk.

While substantially broader than classical deterrence theory, this model retains a

strong focus on the influence of legal sanctions and related enforcement contingencies.

For example, no extralegal factors (ie. informal/social sanctions) have been explicitly

incorporated into the model. Similarly, no attempt was made to incorporate Homel’s

(1986) concept of balancing losses into the model. Rather, these additional factors will be

conceptualised within a social learning framework, described in the following section.

3.3 Social learning theory

3.3.1 Principles of social learning theory

Akers (1977; 1990) has argued that deterrence theory is not a general or complete

model of criminal behaviour, but represents a sub-set of social learning theory. His


central thesis is that “the primary concepts and valid postulates of deterrence and

rational choice are subsumable under general social learning or behavioral principles”

(Akers, 1990, p. 655). Whereas deterrence theory is concerned with the influence of legal

sanctions on criminal behaviour, social learning theory is more concerned with the

overall social setting in which behaviours occur and the way in which they are

differentially rewarded and punished (Akers, 1990).

Social learning theory is a broad descriptor used in psychology to encompass a set

of theoretical perspectives with some common elements. One of the key characteristics of

social learning theories is the proposition that behaviour is primarily learned and

reinforced through social interaction. Bandura (1977), who is generally recognised as the

pioneer of social learning theory, argued that behaviour is a product of the “continuous

reciprocal interaction between cognitive, behavioural, and environmental determinants”

(p.vii). Within this framework of reciprocal determinism, the most rudimentary mode of

learning is through direct experience and the positive and negative effects that arise from

different actions.

When people deal with everyday events, some of their responses prove successful,

while others have no effect or result in punishing outcomes. Through this process of

differential reinforcement, successful forms of behaviour are eventually selected

and ineffectual ones are discarded (Bandura, 1977, p.17).

Another key principle of social learning theory is that behaviour can be learnt

vicariously, through observing others and the outcomes of their actions. This process is

generally referred to as modelling or imitation (Bandura, 1969, 1977). Importantly, the

process of modelling not only involves the learning of new behaviours, but provides

coded information that “serves as a guide for action” (Bandura, 1977, p.22).

3.3.2 Akers’ social learning theory

3.3.2.1 Origins of the theory

In the mid-1960s, Ronald Akers in collaboration with Robert Burgess, developed a

form of social learning theory specifically designed to explain criminal or deviant

behaviour (Akers, 1977; 1990; 1994). The theory drew on elements of both sociological

theory (Sutherland’s Differential Association theory) and psychological theory (Skinner’s

operant conditioning and Bandura’s social learning concepts) and, consequently, is

sometimes referred to as differential association-reinforcement theory. While the theory

has been refined over time, the basic concepts remain consistent with its origins.


Sutherland’s theory, originally developed in the late 1940s, proposed that the

primary causal process underpinning criminal behaviour is differential association. This

concept refers to the degree to which individual’s are exposed to others who engage in

illegal/deviant behaviour or not (Akers, 1994). The central proposition is that

illegal/deviant behaviour is learned through association (interaction and communication)

with social groups, particularly significant groups such as family or peers (DiBlasio &

Benda, 1990). Sutherland articulated his theory in nine propositions. Two of these require

special mention:

8. The process of learning criminal behaviour by association with criminal and

anti-criminal patterns involves all the mechanisms that are involved in any

other learning.

9. Although criminal behaviour is an expression of general needs and values, it is

not explained by those general needs and values, because noncriminal

behaviour is an expression of the same needs and values (Sutherland, 1947

cited in Akers, 1994, p.93).

The first of the above propositions highlights the central role that learning played in

Sutherland’s theory. However, Sutherland did not attempt to explain the actual

mechanisms involved in this learning process. This was the aspect of the theory that

Akers and Burgess strengthened and extended through applying the learning principles of

operant and respondent conditioning (Akers, 1977, 1994). The second of the above

propositions illustrates how Sutherland’s perspective on criminal behaviour is consistent

with normative sociological theories of deviance. Rather than propose that crime is a

product of person-related factors like biological or psychological abnormality, or social

factors like poverty, he argues that it is a behaviour learned through interaction with

intimate social groups. “In this sense, there is no deviance in Sutherland’s conception,

because everybody conforms to some set of expectations. It is expectations which conflict

with the law” (Anleu, 1995, p.25).

3.3.2.2 Processes and constructs in Akers’ theory

As noted above, Akers and Burgess reformulated Sutherland’s theory utilising

learning principles from operant conditioning and social learning. Akers subsequently

developed this differential association-reinforcement theory over a number of years,

“most often labelling it ‘social learning’ and applying it to criminal, delinquent, and

deviant behaviour in general” (Akers, 1994, p.94). Consistent with other social learning

perspectives, it is proposed that social behaviour is acquired either directly through


conditioning or indirectly through imitation or modelling of others’ behaviour. A

person’s behaviour is strengthened (or reinforced) through rewards and avoidance of

punishment, while it is weakened (or punished) through aversive stimuli and loss of

rewards.

Whether deviant or conforming behaviour is acquired or persists depends on the

past and present rewards or punishments for the behaviour and the rewards and

punishments attached to alternative behaviour – differential reinforcement (Akers

et al, 1979, p.638).

There are four major theoretical constructs in Akers’ social learning theory:

differential association, imitation, definitions and differential reinforcement (Akers,

1977, 1994). As with Sutherland’s theory, the central construct in Akers’ theory is

differential association. This refers to the patterns of interaction between a person and

other individuals and groups with whom they identify. Most importantly, this relates to

interaction with primary groups such as friends and family, but also encompasses

secondary groups such as work colleagues (Capece & Akers, 1995). It is through

differential association that a person is “exposed to and learns definitions, is exposed to

behavioural models, and receives social reinforcement or punishment for taking or

refraining from some action” (Capece & Akers, 1995, p.345). Consequently, a person’s

behaviour tends to be congruent with the behaviour of those with whom they associate.

Differential association has both a behavioural and normative dimension (Krohn et

al, 1985; Capece & Akers, 1995). The behavioural dimension relates to the degree of

interaction a person has with different groups and the models, definitions and reinforcers

to which they are exposed. The normative dimension relates to the normative or

evaluative climate found in these groups toward different behaviours. This will be

particularly evident in the content of group-shared definitions (Akers, 1996).

Imitation refers to the process by which an individual models their actions on the

behaviour of significant others. The primary source of behavioural models is salient

social groups, such as family and peers. Other sources of imitation can include the media

and other reference groups (Skinner & Fream, 1997). While imitation is particularly

important during the initiation phase of a new behaviour, it is less relevant during the

maintenance or cessation phase (Akers et al, 1979).

Definitions are attitudes, beliefs or orientations that individuals hold toward

different behaviours. They are learned through imitation and reinforced through

interaction with significant groups of individuals. The definitions are evaluative in nature


and define for an individual what behaviour is appropriate or inappropriate in different

situations. Importantly, definitions do not function as direct motivators, but rather as

facilitative or inhibitory cues that signal whether a particular behaviour is likely to be

rewarded or punished in any given situation (Akers, 1996). “The more individuals define

the behaviour as good (positive definition) or at least justified (neutralizing definition)

rather than as undesirable (negative definition), the more likely they are to engage in it”

(Akers et al, 1979, p.638). Akers (1996) distinguishes two related but distinct classes of

definitions in his theory. The first relates to group-shared definitions that are exhibited or

expressed by the salient groups with which a person associates. The second relates to an

individual’s own internalised or professed definitions. While an individual’s own

definitions will generally be positively correlated with those held by the groups with

which they associate, there need not be a perfect relationship. A discrepancy is more

likely in cases where there is variability in the definitions and social responses exhibited

by the different groups that a person is exposed to.

Differential reinforcement relates to the balance of reinforcement (rewarding or

desired outcomes) and punishment (negative or undesirable consequences) that an

individual anticipates in relation to different actions (Capece & Akers, 1995). In any

given situation, an individual is more likely to perform (and repeat) the behaviour for

which they anticipate the most desirable balance of reinforcement ie. the behaviour that

maximises the degree of desirable outcomes they will experience over negative

outcomes. “Behaviour (whether deviant or conforming) results from greater

reinforcement, on balance, over punishing contingencies for the same behaviour and the

reinforcing-punishing contingencies on alternative behaviour” (Akers et al, 1979, p.638).

The reinforcers (rewards) and punishments can be either social or non-social in

nature. However, the theory posits that the most important are social reinforcers (Akers et

al, 1979). These relate to the approval or disapproval that an individual anticipates

receiving from their peers or significant others for performing (or not performing) a

particular behaviour. Importantly, social reinforcement is not limited to the immediate

(actual or anticipated) reactions of others, but includes both tangible and intangible

rewards valued by society and particular subgroups (Akers, 1994).

Although less important, non-social reinforcers relate to direct physiological

rewards or costs associated with certain behaviours, such as the physiological effects of

drugs (Akers et al, 1979).

In summary, Akers’ theory would predict that the probability of an individual

engaging in an illegal or deviant behaviour increases when he or she:


§ associates with significant others who hold favourable (or at least neutral)

definitions toward the behaviour in question and tend to engage in it themselves (ie.

provide models of the behaviour);

§ holds favourable definitions towards the behaviour; and

§ are differentially reinforced for the behaviour (ie. anticipate more rewards than

punishers arising from it) (Akers et al, 1979).

3.3.2.3 Empirical support for Akers’ theory

Akers’ social learning theory has been used to investigate a wide range of deviant

or non-conforming behaviours including alcohol and drug abuse, adolescent smoking,

delinquency, adolescent sexual behaviour and computer crime (eg. Akers et al, 1979;

Krohn et al, 1985; DiBlasio & Benda, 1990; Akers & Lee, 1996; Skinner & Fream,

1997). These studies have demonstrated extensive support for the theory with both

adolescent and adult populations. However, the theory has not been utilised widely in the

road safety field. An exception was the use by DiBlasio (1987) of Akers’ theory to

investigate why young adolescents choose to ride with a drinking driver. In this study, the

social learning model was able to explain almost half of the variance in riding behaviour

(R2=.49).

However, while many studies have found a moderate to strong relationship between

conforming/deviant behaviour and the degree of association with conforming or deviant

groups, there has been considerable debate over the causal direction of this relationship

(Thornbery et al, 1994; Smith & Brame, 1994; Akers, 1996). In Akers’ theory,

differential association performs a causal role in the development of deviant behaviour. It

is the initial factor that influences the initiation of conforming or deviant behaviour, by

providing the social environment in which a person is exposed to models, definitions and

social reinforcement supportive or not supportive of various behaviours. While imitation

will be less relevant after a behaviour is learned, differential association remains

important by providing the social reinforcers which reward or punish the behaviour and

its alternatives.

In contrast, some researchers have argued that differential association is only a

coincidental rather than causal factor in the development of deviant behaviours (Glueck

& Glueck, 1950; Gottfredson & Hirschi, 1990). In keeping with the Gluecks’ (1950)

assertion that birds of a feather flock together, it is argued that people with deviant

tendencies seek out each other for companionship (Thornberry et al, 1994). In other


words, differential association is a product of a predisposition to deviance rather than a

cause of it.

Warr (1993, cited in Akers & Lee, 1996) rejects this criticism of social learning. He

cites the considerable body of evidence indicating that peer associa tions more often

precede the development of deviant patterns of behaviour, rather than the other way

around. Instead, Warr (1993 cited in Akers & Lee, 1996) and others such as Thornbery et

al (1994) and Smith and Brame (1994) propose a more interactionist perspective, with the

differential association and deviance being related in a complex, reciprocal manner.

“Adolescents are commonly introduced to delinquency by their friends and subsequently

become more selective in their choice of friends. The ‘feathering’ and ‘flocking’ . . . are

not mutually exclusive and may instead be part of a unified process” (Warr, 1993 cited in

Akers & Lee, 1996, p.320).

In response to these arguments, Akers has asserted that his theory acknowledges the

complex relationship between differential association and deviance by incorporating

feedback effects into the social learning process (Akers & Lee, 1996). The feedback

process is built into differential reinforcement, which acknowledges that the

consequences of any behaviour (both rewarding and punishing contingencies) will

influence the reoccurrence of that behaviour. Therefore, while differential association

will be the primary influence on the initiation of a deviant behaviour, the resulting

consequences may in turn alter associa tional patterns. In this way, “further interaction

with others is based, at least in part, on whether they too are involved in the deviant

activity and to what degree” (Akers, 1985, p.60).

3.3.3 Application of Akers’ theory to unlicensed driving

While Akers’ social learning theory has not been used to examine unlicensed

driving to date, it appears to offer a number of heuristic advantages over deterrence-based

theories. Firstly, it provides a comprehensive means of addressing a range of important

factors including formal and informal sanctions, direct and indirect experiences and

punishment and punishment avoidance. Secondly, as noted in section 3.2.3, the concept

of differential reinforcement appears to encapsulate the weighing-up of potential losses

that Homel (1986) argues is central to the process of deciding to drink and drive. Thirdly,

this approach appears better equipped to explain compulsive behaviours, characteristic of

some recidivist offenders particularly those who repeatedly drive after having their

licence disqualified (Mirrlees-Black, 1993; Bakker et al, 1997). Such compulsive

behaviours arguably represent cases where the anticipated non-social rewards associated


with driving exert a powerful influence on behaviour. Indeed, Bakker et al (1997) have

developed and trialed a relapse-prevention program for disqualified driving offenders

(based on social learning theory) which involves teaching more effective ways of solving

interpersonal problems and regulating negative emotional states.

Based on the constructs and processes proposed by Akers, the initial decision to

drive unlicensed or not will be determined through differential association. Depending on

the significant groups that an individual interacts with, he or she will be exposed to a

social environment that is on balance favourable, neutral or opposed to unlicensed driving

(in general and possibly to different forms of the behaviour). The mechanisms through

which this will occur will be the salient models, definitions and social reinforcement

(rewards and punishments) operating in the individual’s environment. Through imitation

and reinforcement, individuals will learn definitions that will act as cues to the

appropriateness of unlicensed driving. These definitions in interaction with the behaviour

of salient models and the anticipated balance of reinforcement from their social

environment will encourage compliance with the law or produce the initial decision to

drive unlicensed. After someone has engaged in unlicensed driving, initiation would

become less important while his or her definitions may be modified by the experience of

the behaviour. Most importantly, it will be the actual consequences of unlicensed driving

(ie. the balance of social reinforcers and punishers) experienced by an individual that

will determine the likelihood of them continuing to drive unlicensed and to what extent

(Akers et al, 1979).

Drawing on the available qualitative evidence (eg. Mirrlees-Black, 1993; Bakker et

al, 1997; Williamson, 1996), the anticipated and actual social rewards for unlicensed

driving could include:

§ social approval (respect, praise, encouragement) from significant others;

§ the opportunity to participate in important social activities;

§ the thrill of breaking the law;

§ the sense of empowerment or autonomy experienced while driving; or

§ the personal freedom provided by driving.

Conversely, anticipated or actual social punishers could include:

§ social disapproval from significant others;

§ social sanctions (eg. loss of friends, social stigma, loss of self image);

§ legal sanctions (eg. loss of freedom or income due to fines, licence loss or

imprisonment); or

§ anxiety or guilt associated with apprehension and punishment via legal sanctions.


As noted earlier, nonsocial reinforcement in Akers’ model is confined to the direct

physiological rewards and punishments associated with various behaviours. Therefore,

unlike many other deviant behaviours (such as alcohol and drug abuse, adolescent

smoking or adolescent sexual behaviour), the nonsocial reinforcement directly associated

with unlicensed driving would appear minimal. In particular, the rewards would appear

limited to the physiological effects of driving, such as the adrenalin-rush associated with

fast driving.

At any point in time, an individual’s anticipated balance of reinforcement will be

determined by their perceived probability of rewards versus the perceived probability of

punishments arising from the act of unlicensed driving. These perceptions will be

strongly influenced by both direct and indirect experiences of (i) rewards for unlicensed

driving and instances of punishment avoidance; and (ii) punishment and loss of rewards

associated with unlicensed driving.

Based on the summary provided earlier, Akers’ theory would predict that the

probability of unlicensed driving would increase when an individual associates with

significant others who hold favourable (or at least neutral) definitions toward unlicensed

driving and tend to engage in the behaviour (ie. provide models of the behaviour); and

when he or she holds favourable definitions toward unlicensed driving and are

differentially reinforced for the behaviour (ie. anticipate more rewards than punishers for

driving while unlicensed) (Akers et al, 1979).

3.3.4 A social learning model of unlicensed driving

Figure 3.3 illustrates a social learning model of unlicensed driving, based on the

constructs and processes outlined in Akers’ theory. The constructs operationalised in the

social learning model include:

1. Differential association - degree of association with significant others who -

a. engage in the behaviour or not (behavioural dimension),

b. hold attitudes8 (definitions) favourable, neutral or unfavourable to the

behaviour (normative dimension);

2. Imitation - overall exposure to models who engage in the behaviour;

8. When operationalising the social learning model, the term attitudes was used to describe the construct

of definitions. This was prompted by the more common use of the term attitudes in road safety and traffic psychology research. However, the construct retains its meaning as intended by Akers ie. it refers to attitudes, beliefs or orientations that individuals hold toward different behaviours.


3. Personal attitudes (definitions) – favourable, neutral or unfavourable to -

a. unlicensed driving,

b. alternative behaviours, such as using public transport; and

4. Differential reinforcement – overall anticipated balance of reinforcement for a

behaviour -

a. anticipated social and non-social rewards for the behaviour,

b. anticipated social and non-social punishments for the behaviour,

c. overall balance of reinforcers (social and non-social).

3.4 Sensation seeking

As noted in section 2.3.3, there is a growing body of evidence linking unlicensed

driving to other high-risk driving behaviours, such as drink driving and speeding. This

suggests that unlicensed driving may be influenced by an individual’s general propensity

to take risks on the road.

There has been considerable research into the link between personality-related

factors, risky driving and crash involvement. In one of the first studies in the area,

Tillman and Hobbs (1949, cited in Evans, 1993) found that taxi drivers involved in

crashes were more likely than crash-free drivers to have a history of involvement with

criminal courts, social service, public health agencies and credit bureaus. Based on these

results, Tillman and Hobbs (1949, cited in Evans, 1993) argued that: “a man drives as he

lives”. A range of subsequent studies with other driver populations has found a link

between crash involvement and antisocial behaviour, although it is difficult to determine

Figure 3.3: Social learning model of unlicensed driving

Differential reinforcement - anticipated balance of rewards & punishments for unlicensed driving and alternatives

Decision to drive licensed or

unlicensed

Differential association - driving behaviour of significant others - attitudes of significant others toward unlicensed driving -

Imitation - overall exposure to people (models) who engage in unlicensed driving or not

Personal attitudes - to unlicensed driving - to alternative behaviours


whether this relationship is coincidental or causal (Evans, 1993). Overall, the relationship

between personality factors and crash involvement remains unclear.

Nonetheless, there are some personality-related factors that appear to be more

strongly related to risky driving than others. For example, many studies have

demonstrated a significant positive relationship between the construct of sensation

seeking and risky driving (Jonah, 1997). This construct has been defined as "a trait

defined by the seeking of varied, novel, complex, and intense sensations and experiences,

and a willingness to take physical, social, legal, and financial risks for the sake of such

experience" (Zuckerman, 1994 p. 27). Sensation seeking is generally measured in terms

of scores on the Sensation Seeking Scale (SSS), first published by Zuckerman et al (1964,

cited in Zuckerman, Eysenck & Eysenck, 1978). Using this scale, various studies have

shown higher levels of sensation seeking to be linked to a range of risky behaviours and

to be more common among males and young people (Zuckerman, 1994). Among the

driving behaviours that have been shown to be significantly associated with sensation

seeking are drink driving, speeding and following too closely (Jonah, 1997). The studies

have generally found a stronger relationship between sensation seeking and self- reported

driving behaviour, rather than with traffic offences or crash involvement. However, this

may be due to weaknesses in the crash measures used to date.

Risk-taking behaviour is not a variable explicitly included within Akers’ social

learning theory. However, it is reasonable to assume that the non-social rewards that an

individual expects to achieve from engaging in risky behaviours may be moderated by

their propensity for risk taking. For example, high sensation seekers may be more likely

to perceive risky behaviours as intrinsically rewarding than low sensation seekers. As

such, high sensation seekers may be more likely to perceive unlicensed driving as being

rewarding, due to the thrill and excitement associated with breaking the law.

Alternatively, among unlicensed drivers, the act of driving may simply facilitate other

behaviours such as drink driving and speeding that are rewarding to high sensation

seekers.

There are other theoretical reasons for examining the role of sensation seeking

within a social learning framework. One of the major alternative criminological

perspectives to social learning theory is social control theory (or self-control theory) (eg.

Gottfredson & Hirschi, 1990). This perspective is often contrasted with social learning

since it contends that deviant behaviour is not the product of any learning. Rather, social

control theory posits that deviant behaviour is a product of poorly developed self-control

and exposure to criminal opportunities (Smith & Brame, 1994; Winfree & Bernat, 1998).


The major determinant of self-control is the effectiveness of parental management and

child rearing practices. Among the characteristics of individuals with poor self-control is

a tendency or preference for risk-taking (Winfree & Bernat, 1998).

A recent study by Winfree and Bernat (1998) directly compared the ability of social

learning and social control theories to explain adolescent substance abuse. They found

that constructs within both theories were able to predict substance abuse. However, while

a number of the social learning theory variables were significant, the only social control

variable to perform consistently well was propensity for risk-taking. This suggests that an

important test for social learning theory is its ability to account for the significant

relationship between risk-taking and deviant behaviours.

3.5 Alcohol misuse

Among disqualified drivers, there appears to be a hard-core of persistent offenders

who continue to drive despite previous convictions (Bakker et al, 1997). In many cases,

these convictions occur in conjunction with other offences, such as drink driving,

refusing a breath test or reckless driving (NRMA, 1991). Australian research has

confirmed that there is a higher incidence of alcohol impairment among unlicensed

drivers involved in serious crashes, compared to licensed drivers (Harrison, 1997;

Watson, 1997; FORS, 1997b). This is most pronounced among disqualified drivers,

"perhaps reflecting a high representation of recidivist drink drivers in that group"

(FORS, 1997b). A Queensland study found that unlicensed driving was an independent

predictor of reoffence among drivers convicted of drink driving (Siskind, Schonfeld &

Sheehan, 2000). Similarly, international research has indicated that drivers involved in

fatal crashes with high blood alcohol concentrations (BACs) are more likely than other

groups to have a history of previous licence suspensions and to have been driving without

a valid licence at the time (Simpson & Mayhew, 1991).

Persistent drink driving offenders tend to display numerous psychological and

behavioural characteristics that distinguish them from the general driving population,

including higher levels of aggression, hostility and sensation seeking. They are also more

likely to have a criminal history, to use drugs, to have a poor driving history, to consume

large amounts of alcohol more frequently and to experience alcohol-related problems

(Mayhew et al, 1997; Bailey & Bailey, 2000). Hence, it is possible that the behaviour of

many unlicensed drivers, particularly those who have previously lost their licence for


drink driving, is strongly influenced by alcohol misuse.9 From a social learning

perspective, driving may represent an important means of accessing the social settings in

which alcohol is consumed. Hence, the anticipated social and non-social rewards

associated with driving may be intensified by alcohol misuse.

However, some caveats need to be placed on this conclusion. Firstly, while

persistent drink driving offenders appear to be different to the general driving population,

they are not a homogeneous group (Hedlund & Fell, 1995; Mayhew et al, 1997). "They

are diverse, with different backgrounds, problems, and most likely different reasons for

engaging in DWI (driving while impaired) behaviour" (Mayhew et al, 1997, p.794).

Secondly, there is evidence that drink driving offenders and high-risk problem drivers are

substantially overlapping populations, sharing many common characteristics (Wilson,

1991; Donovan et al, 1983 cited in Bakker et al, 1997). Indeed, Bakker et al (1997, p.29)

have argued that:

... while it is clear that most disqualified drivers have a number of offences for

alcohol-impaired driving, DWD (driving-while-disqualified) offences are, for most

of these individuals, more numerous, strongly suggesting that it is a significant

problem in its own right.

3.6 Chapter summary

The aim of this chapter has been to provide a theoretical foundation for the current

program of research into unlicensed driving. It has reviewed perspectives from

psychology, sociology and criminology to identify personal and social factors that appear

relevant to understanding the behaviour. This broad approach is warranted by the

spectrum of behaviour involved in unlicensed driving.

At the personal level, the review focused on the potential role of sensation seeking

and alcohol misuse in unlicensed driving. While there is strong evidence that these

psychological characteristics can influence the nature of on-road driving behaviour, it is

9. Alcohol misuse in this research is examined from the general perspective of problem drinking, rather

than focusing solely on alcoholism or alcohol dependence. While these terms are sometimes used interchangeably, it is important to distinguish between the concepts. Compared to alcoholics, problem drinkers have “a shorter problem-drinking history, greater social and economic stability, and greater personal resources. They typically have not experienced major losses because of their drinking and have not exhibited severe withdrawal symptoms, such as seizure or delirium tremens, upon cessation of past drinking” (Walitzer & Connors, 1999, pp.138-39). Hence, problem drinkers can include people who drink heavily but do not necessarily have a history of severe physical dependence on alcohol. From a road safety perspective, this may include people whose alcohol misuse only occasionally impinges on their driving behaviour. Consistent with this perspective, the Alcohol Use Disorders Identification Test (AUDIT) was selected to measure alcohol misuse in Study Two (see section 5.3.3.3). This test is designed to detect problem drinking that may be considered hazardous through to levels that suggest psychological or physical dependence (Early Intervention Unit, 1993).


unclear how much they may directly motivate unlicensed driving. At the social level, the

review has focused on the relative strengths of deterrence theory and social learning

theory. While deterrence theory has been criticised by some as being too narrowly

focused on persona l experiences of legal sanctions, a recent reconceptualisation has

broadened the scope of the theory to encompass punishment avoidance and vicarious

learning experiences. Nonetheless, social learning theory appears to represent a more

comprehensive framework for understanding the broad range of factors that can either

encourage or discourage illegal behaviours, such as unlicensed driving.

These various perspectives will be used to guide the formulation of the hypotheses

in the forthcoming studies. In addition, Study Three will compare the predictive utility of

the different perspectives, along with certain environmental factors that appear to

facilitate unlicensed driving. This is critical for identifying strategies to discourage the

behaviour. Although social learning theory may represent a more comprehensive

perspective than deterrence theory, its capacity to better explain unlicensed driving

remains unclear. Indeed, deterrence theory may represent a more parsimonious approach

for explaining the behaviour and for guiding countermeasure design.



Chapter Four: The crash involvement of unlicensed drivers

4.1 Introductory comments................................................................................ 67

4.2 Study aims and hypotheses.......................................................................... 67

4.3 Method......................................................................................................... 69

4.3.1 Data characteristics .............................................................................. 69

4.3.2 Procedure ............................................................................................. 71

4.3.3 Statistical analyses ............................................................................... 73

4.4 Results ......................................................................................................... 74

4.4.1 Incidence and severity of unlicensed driver crashes............................ 74

4.4.2 Age and gender characteristics of drivers involved in crashes ............ 76

4.4.3 Crash circumstances ............................................................................ 78

4.4.4 Contributing factors to crashes ............................................................ 82

4.4.5 The crash risk of unlicensed drivers .................................................... 85

4.4.5.1 Risk of involvement in crashes .................................................... 86

4.4.5.2 Risk of death and serious injury in the event of a crash .............. 91

4.5 Discussion.................................................................................................... 95

4.5.1 Study limitations .................................................................................. 95

4.5.2 Support for study hypotheses............................................................... 97

4.5.3 Implications for theory ........................................................................ 102

4.5.4 Implications for countermeasure development.................................... 103

4.5.5 Future directions for research.............................................................. 104

4.6 Chapter summary......................................................................................... 105




This chapter documents the first study undertaken as part of the research program.

It involves an analysis of the crash involvement patterns of unlicensed drivers in the

Australian state of Queensland. Much of what is currently known about unlicensed

driving has been derived from road crash statistics. This study is designed to both confirm

and extend this knowledge by undertaking a more thorough examination of the crashes

involving unlicensed drivers.

Two main lines of investigation are pursued throughout this study. The first entails

a comparison of the crashes involving unlicensed drivers with those involving licensed

drivers. This is designed to establish whether unlicensed drivers represent a special group

requiring specific attention from road safety policy-makers. The second line of

investigation involves a comparison of the crash involvement patterns of the different

types of unlicensed drivers. This will provide an insight into whether unlicensed drivers

represent a homogeneous group, who can be effectively targeted through generic

countermeasures.

While the primary aim of this study to address the specific research questions

identified below, it will also provide a foundation for the next two studies. These later

studies will investigate whether the characteristics of unlicensed drivers involved in

crashes are representative of those offenders not involved in crashes.

4.2 Study aims and hypotheses

This study will examine the first three research questions identified in section 2.7.

The hypotheses relating to each research question are detailed below along with a brief

rationale.


H1 Unlicensed driver crashes will be more likely to involve alcohol, speeding,

inexperience and motorcycle use than those involving licensed drivers.

H2 Unlicensed driver crashes will be more likely to occur at recreational

times than those involving licensed drivers.

H3 Unlicensed drivers will be more likely to be considered at fault by the

police for the crashes in which they are involved compared with licensed

drivers.

This group of hypotheses is based on the emerging evidence indicating that

unlicensed driving is associated with a cluster of high-risk driving behaviours. To date


this research has mainly focused on the over-representation of unlicensed drivers in

crashes involving alcohol (and related issues such as the prevalence of these crashes at

recreational times). Based on the evidence presented in Chapter 2, it is likely that

unlicensed driver crashes will also feature an over-representation of motorcycles,

inexperience and a range of other inappropriate actions for which they will be considered

at fault by the police. These hypotheses also reflect the possible role of alcohol misuse

and sensation seeking in the behaviour of unlicensed drivers, as discussed in the previous

chapter.

2. Is unlicensed driving associated with a higher crash risk compared to legal driving?

H4 Based on quasi-induced exposure measures, unlicensed drivers will be at a

higher risk of being involved in a road crash than licensed drivers.

H5 The crashes involving unlicensed drivers will result in significantly higher

levels of death and serious injury than those involving licensed drivers.

These two hypotheses are designed to confirm findings from other

jurisdictions that suggest that unlicensed driving is associated with a higher crash

risk than legal driving. As outlined in Chapter 2, there is evidence that unlicensed

drivers are at a higher risk of crashing in general, and that the crashes in which they

are involved tend to be more severe than those involving licensed drivers.



H6 Significant differences will be found between the unlicensed driver

sub-groups involved in crashes, in terms of age and gender.

H7 Significant differences will be found between the unlicensed driver

sub-groups in terms of the circumstances of the crashes in which they are

involved, the associated contributing factors and the severity of the crashes.

These two hypotheses are based on an awareness of the spectrum of

offences involved in unlicensed driving. As previously noted, unlicensed driving

ranges from offences that are essentially administrative in nature, such as driving

with an exp ired licence, through to more deviant behaviours, such as persistent

disqualified driving. Accordingly, it is reasonable to assume that the characteristics

and motives of offenders may vary greatly, and be reflected in their on-road

behaviour. In particular, any differences in risk-taking behaviour among the

unlicensed driver sub-groups should be evident in the types and severity of the

crashes in which they are involved. These hypotheses also have important


implications for the theory. They will confirm whether theoretical explanations of

unlicensed driving need to account for a lack of homogeneity among offenders.

4.3 Method

4.3.1 Data characteristics

The data used in this study was extracted from Queensland Transport’s road crash

database for the years 1994-98. This database contains records for all crashes reported to

the police in the state. During the period, drivers in Queensland were required to report

all crashes to the police that “resulted from the movement of at least one road vehicle on

a road and involving death or injury to any person, or property damage” (Queensland

Transport, 1999, p.A1-1). The requirement for reporting property damage only (PDO)

crashes was that at least one vehicle was towed away or the damage cost was greater than

$2,500.

Among the information provided for each crash was:

§ the age (in 12 categories), gender and licence status for all the controllers of

motorised vehicles (including cars, car derivatives, trucks, buses, motorcycles and

tractors) involved in the crash;10

§ details of the circumstances of the crash including the day, time, location,

prevailing road and traffic conditions, and type of vehicles involved; and

§ the contributing factors to the crash as cited by the attending police.

The licence classification used by Queensland Transport comprises various groups

of licensed and unlicensed drivers, as well as those holding overseas licences, and those

of unknown licence status. The unlicensed driver categories included:

§ drivers with an expired licence;

§ disqualified/suspended11 drivers;

§ drivers with an inappropriate class of licence;

§ drivers who have never held a licence; and

§ other unlicensed drivers.

Unfortunately, the crash database does not distinguish between those drivers who

have been disqualified (by a court order) and those who have had their licence suspended

(generally for accumulation of demerit points). Consequently, these two categories of

10. The term ‘driver’ will be generally used in this study to refer to all controllers of motor vehicles.

Therefore, unless specified otherwise, motorcycle riders will be identified as drivers. 11. In the crash database, the term ‘cancelled’ was used instead of ‘suspended’. In more recent years,

however, Queensland Transport has standardised the use of the term suspended to describe those drivers who have had their licence cancelled. Therefore, to ensure consistency with the later studies, the terms suspended is used in this study.


unlicensed drivers are combined in this particular study. In addition, those drivers who

had failed to renew their licence after a period of disqualification or after letting it lapse

(sometimes referred to as not currently licensed) are not specifically identified. The other

unlicensed category is used by the police in cases where drivers are detected contravening

special licence restrictions (eg. driving outside the hours permitted by a restricted licence)

or when they are satisfied that the driver does not hold a valid licence but the exact

circumstances are unclear. Although it is undesirable to have an ‘other’ category tha t is

relatively large (see section 4.4.1), it was impossible to re-allocate the drivers within this

group to any of the other categories.

Five years of data was analysed to ensure that general trends were identified and to

provide sufficient numbers to permit meaningful comparisons among the different groups

of unlicensed drivers. The main unit of analysis was the drivers involved in crashes

during the period, rather than crashes per se. The total number of crash- involved drivers

available for analysis is shown in Table 4.1, broken down by year and the severity of the

crash. In Queensland, a crash is classed as fatal if it results in the death of a person within

30 days of injuries sustained in the crash. A serious injury crash is one that results in the

hospitalisation of a person due to injuries sustained in the crash. The other injury category

includes cases where people are treated at a hospital accident and emergency department or

receive medical attention at the scene of the crash. A PDO crash is where no one was

injured but at least one vehicle was towed away or the damage cost exceeded $2,500

(Queensland Transport, 1999).

Table 4.1

Number of drivers involved in crashes in Queensland between 1994-98, by year and

crash severity

Severity of Crash L Year

Fatal Serious injury

Other injury

Property damage only Total

1994 542 5375 11796 17630 35343

1995 591 5431 12862 16887 35771

1996 486 5257 13172 16148 35063

1997 468 4887 12617 14239 32211

1998 356 5107 12318 14334 32115

Total 2443 26057 62765 79238 170503

Source: Queensland Road Crash Database, Queensland Transport


4.3.2 Procedure

Three main procedures were used in this study. The first involved comparing the

crashes involving unlicensed drivers with those involving licensed drivers. This required

the combining of the various sub-groups of unlicensed and licensed drivers into two

composite groups, and the exclusion of those drivers with an international licence or

unknown licence. In the crash database, each party involved in a crash is assigned a

particular unit number. During the study period, the standard practice was for the police

to assign unit number one to the party that they considered most at fault (QPS, 1993).

However, it was decided to include all licensed and unlicensed drivers in the initial

analyses, irrespective of whether they were judged by the police to be at fault for the

crash or not. This ensured that the factors contributing to overall crash involvement were

assessed and it avoided any biases that may have been introduced due to police reporting

or prosecution practices. For example, there may be a greater tendency for the police to

attribute fault for a crash to an unlicensed driver who has already demonstrated a

propensity to break the law.

The second main procedure used in this study was to compare the crash

involvement patterns of the different types of unlicensed drivers. Once again, these

analyses generally included all the unlicensed drivers involved in crashes, irrespective of

whether they were considered at fault or not.

The third main procedure used in this study involved estimating the crash risk of

unlicensed drivers (and licensed drivers) through the use of quasi- induced exposure

measures. This method was modelled on the work of DeYoung et al (1997) who used the

approach to estimate the exposure and fatal crash risk of unlicensed drivers in California.

The approach is designed to overcome the difficulties associated with measuring the

number of unlicensed drivers on the road (see section 2.2 for a discussion of these

difficulties). The quasi- induced exposure method attempts to overcome this problem by

using crash data to estimate exposure. It is based on the assumption that “innocent

drivers, or those not responsible in a multi-vehicle crash, represent a statistical sample of

the exposed population” (DeYoung et al 1997, p.18). In other words, it assumes that

those drivers who are judged to be innocent in a crash are largely incidental to the event

and hence represent a random sample of drivers.12 Expressed another way, it assumes

12. To some extent, this assumption is verifiable by examining the distribution of drivers within the

innocent group across the different types of drivers. While DeYoung et al (1997) found some evidence that this assumption was not met in their data, they suggest that their small sample sizes accounted for the discrepancies.


that “at-fault drivers ‘choose’ their innocent victims at random from all drivers present”

(DeYoung, 1997, p.20).

Based on this assumption, DeYoung et al (1997) derive two rates:

Involvement rate (IR) = % at fault Crash (ratio) rate = IR for unlicensed drivers % innocent IR for licensed drivers

An involvement rate (IR) is calculated for each licence class by dividing the

percentage of drivers considered to be at fault in multi-vehicle crashes with the

percentage who were innocent. A crash (ratio) rate can then be derived for unlicensed

drivers (or particular types of unlicensed drivers) by dividing their IR with that calculated

for licensed drivers. As such, the crash rate can be considered an odds ratio [although

DeYoung et al (1997) do not specifically refer to it in this way].

A major advantage of the quasi- induced exposure approach is that it overcomes

some of the potential problems associated with the under-reporting of crashes by

unlicensed drivers. Due to the illegal nature of their driving, it would be in the best

interests of an unlicensed driver to avoid reporting their involvement in a crash, wherever

possible. This would be more feasible in the case of a single vehicle crash, rather than

one in which another party is involved. Hence, the use of multi-vehicle crashes should

reduce the likelihood of under-reporting. Furthermore, it is arguable that any potential

bias associated with under-reporting should apply to all types of crashes, not just those in

which the driver believes they may be found at fault. Hence, the bias should affect both

the denominator and numerator in the calculation of the involvement rate.

However, the method used by DeYoung et al (1997) introduces a number of other

potential sources of bias into the data. The first relates to the need to exclude single

vehicle crashes. It is possible that the representation of unlicensed drivers in multi-vehicle

crashes is not sufficiently similar to their involvement in total crashes. After checking,

DeYoung et al (1997, p.18) concluded that the representation of unlicensed drivers in

fatal two-vehicle crashes was sufficiently similar to all fatal crashes to not “pose a special

problem”. The second sources of potential bias relates to the possibility that unlicensed

drivers considered innocent by the police may represent a more crash-prone group than

the innocent licensed drivers because many of them “are young, male, risk takers”

(DeYoung et al, 1997, p.20). In other words, the actions of unlicensed drivers may be

more likely to contribute to a crash, even in cases where they are considered the innocent

party. This would in effect overestimate the number of unlicensed drivers on the road

(because some of the at-fault unlicensed drivers would have been misclassified as


innocent drivers), resulting in an IR rate that is too low for the group. The final source of

potential bias relates to the reporting practices of the police. As already mentioned, the

police may be more likely to find an unlicensed driver at fault for a crash, not specifically

due to their driving behaviour, but because they have already broken the law by driving

without a valid licence.13 This would tend to inflate the percentage of unlicensed drivers

considered at fault and their corresponding IR rate. Consequently, while DeYoung et al

(1997) suggest that these two latter biases may mitigate each other, they note that:

Due to the lack of reliable data, it is not clear whether the crash rates presented

here are underestimates (crash proneness bias), overestimates (negative halo bias),

or whether the two biases cancel each other out and thus produce relatively

unbiased (relative to these two bias issues) estimates (DeYoung et al, 1997, p.20).

In addition, a number of manipulations were required to apply the quasi- induced

exposure method to the dataset used in this study. The first necessity was to exclude

single vehicle crashes from the analyses. This could not be done directly due to the unit-

based nature of the data provided (ie. each case in the dataset corresponded to a unit

involved in a crash, rather than a specific crash). However, for each unit in the dataset,

information was available about the nature of the relevant crash and, in particular, the

vehicle movements that occurred. Consequently, it was possible to identify all the drivers

involved in multi-vehicle crashes due to them being either an angle, head-on, rear-end or

side-swipe crash. These are the same crash types used by Queensland Transport (1999,

p.29) to describe multi-vehicle crashes. As already mentioned, it was possible to identify

the drivers considered at fault for the crash by the unit number they were assigned in the

dataset.

Finally, the above procedures involve the analysis of a variety of different crash

types including fatal, serious injury, minor injury and PDO crashes. In addition, a serious

casualty crash category was created by combining those crashes that resulted in either a

fatality or a hospitalisation. This category offers the advantage of being larger in size than

fatal crashes, but still indicative of more serious crashes.

4.3.3 Statistical analyses

As already noted, five years of data was analysed to ensure that general trends were

identified and to provide sufficient numbers to permit meaningful comparisons between

different groups. Consequently, the sample size used in some of the analyses is very

13. DeYoung et al (1997) refer to this potential bias as a negative halo effect.


large. In light of this and the multiple analyses undertaken, it was decided to set the

significance level (α) for the statistical tests at .01. In addition, a lower alpha level would

not have compromised the power of the statistical tests due to the large sample size.

The categorical data were mainly analysed using Chi-square (χ2) tests for

independence. Where necessary, post-hoc analyses were undertaken within each variable

using an adjusted standardised residual statistic (ê). The adjusted standardized residual

indicates the relative difference between the observed and expected frequencies for a

particular cell, adjusted for row and column totals. This statistic can be used to identify

those cells with observed frequencies significantly higher or lower than expected.

Adjusted standardized residuals are approximately normally distributed with a mean of 0

and a standard deviation of 1, and can be interpreted as Z-scores (Haberman, 1978). The

strength of association between categorical variables was measured using either the phi

(φ) coefficient (for 2 × 2 tables) or Cramer’s Phi (φc) coefficient (for tables greater than 2 ×

2). Cohen (1988, cited in Aron & Aron, 1999) has suggested that a phi coefficient of .10

represents a small effect size, .30 a medium effect size and .50 a large effect size. The

effect size conventions for Cramer’s coefficient vary with the degrees of freedom of the

contingency table. The phi coefficient (but not Cramer’s phi) can be squared to obtain a

squared correlation co-efficient indicating the amount of variance that can be accounted

for by one of the variables in the other (Smithson, 2000).

As noted earlier, the crash ratios produced by the quasi- induced exposure method

represent odds ratios. Consequently, in order to extend the analytical methods used by

DeYoung et al (1997), 99% confidence intervals were derived for the crash ratios to

assess their significance. In addition, logistic regression was used to calculate odds ratios

measuring the involvement of licensed and unlicensed drivers in serious casualty crashes

relative to minor crashes. Age, gender and vehicle type were used as control variables in

these analyses. All statistical tests were undertaken using the Statistical Package for the

Social Science (SPSS) Version 10.0.5.

4.4 Results

4.4.1 Incidence and severity of unlicensed driver crashes

Table 4.2 provides a breakdown of the licence status for all drivers (including

motorcycle riders) involved in crashes in Queensland between 1994 and 1998. During the

period, 6.3% of the drivers involved in fatal crashes and 5.1% of those involved in

serious injury crashes were unlicensed, compared with 2.6% for minor injury crashes and

2.5% for property damage only crashes. The higher relative involvement of unlicensed


drivers in serious casualty crashes (ie. fatal and serious injury crashes) is consistent with

the findings of other studies reviewed in Chapter 2 (see section 2.2.4). On average, there

were 30 unlicensed drivers involved in fatal crashes each year and almost 270 in serious

injury crashes. In total, over 1000 unlicensed drivers are involved in a recorded crash

each year in Queensland.

Table 4.2

Licence status of drivers involved in crashes in Queensland: 1994-98 by severity

Severity of Crash Licence type


Other Injury

Property Damage Total

No. % No % No. % No. % No. %

Licensed 2148 87.9 23560 90.4 58235 92.9 73602 92.9 157635 92.5

Unlicensed 154 6.3 1330 5.1 1626 2.6 1966 2.5 5076 3.0

International 24 1.0 314 1.2 675 1.1 960 1.2 1973 1.2

Unknown 117 4.8 851 3.3 2134 3.4 2693 3.4 5795 3.4

Total 2443 100.0 26055 100.0 62760 100.0 79221 100.0 170479 100.


Table 4.3 provides a further breakdown of the unlicensed drivers involved in

crashes during the period. The largest category of unlicensed drivers was the other

unlicensed group, who represented approximately one-third of the offenders involved in

crashes. The next two categories were the disqualified/suspended and the never licensed

drivers, who represented around one-quarter and one-fifth of the unlicensed drivers

involved in crashes, respectively. Together, the drivers with the expired and

inappropriate class of licence made up another one-fifth of the drivers. The proportion of

drivers at each crash severity was reasonably stable across all the groups, with the

exception of those with an inappropriate class of licence. This group was over-

represented in fatal and serious injury crashes. As will be discussed later, this appears to

be a product of the large number of motorcycle riders in that group (see section 4.4.3).


Table 4.3

Unlicensed drivers involved in crashes in Queensland: 1994-98 by offender type

Severity of Crash Licence type


Other Injury


No. % No. % No. % No. % No. %

Expired 20 13.0 162 12.2 240 14.8 282 14.3 704 13.9

Inapprop-riate class

20 13.0 147 11.1 119 7.3 30 1.5 316 6.2

Disqualified/ suspended

49 31.8 339 25.5 421 25.9 513 26.1 1322 26.0

Never licensed

25 16.2 278 20.9 264 16.2 404 20.5 971 19.1

Other unlicensed

40 26.0 404 30.4 582 35.8 737 37.5 1763 34.7

Total 154 100.0 1330 100.0 1626 100.0 1966 100.0 5076 100.0

Source: Queensland Road Crash Database, Queensland Transport 4.4.2 Age and gender characteristics of drivers involved in crashes

Due to privacy constraints, the only socio-demographic information provided in the

dataset about the drivers involved in crashes was their gender and age. Table 4.4

compares these characteristics for unlicensed drivers involved in serious casualty crashes

with those who had a valid licence.

Table 4.4

Age and gender of drivers involved in serious casualty crashes in Queensland: 1994-98

by licence status

Licence status Variable Licensed

(%) Unlicensed

(%) Significance level1

Gender (n=25707) (n=1483)

Males 69.1 83.3 χ2 (df1) = 134.2, p < .001, Females 30.9 16.7 φ = .07

Age (n=25707) (n=1483)

Under 17 0.3 12.9 χ2 (df3) = 2434.5, p < .001, 17- 24 29.0 42.8 φc? = .30

25 – 59 59.6 42.7 60 and over 11.1 1.6

1. The cells with significant (p<.01) adjusted standardised residuals are bolded. Source: Queensland Road Crash Database, Queensland Transport


Compared with the licensed drivers, unlicensed drivers were more likely to be male

and younger in age. In fact, over four-fifths [83.3%] of the unlicensed drivers involved in

serious casualty crashes during the period were male, while over half [55.7%] were under

the age of 25. It is interesting to note that the unlicensed drivers under the age of 17

[12.9%] represent underage drivers, since the legal age for unaccompanied driving in

Queensland is 17. The small group of licensed drivers under this age would have been

either drivers on Learners Licences or those with a special exemption.

To further explore these differences, Table 4.5 provides a breakdown of the age and

gender characteristics of the unlicensed drivers involved in serious casualty crashes by

offender type. As can be seen, there was a significant overall difference between the

offenders in terms of gender. An inspection of the standardised adjusted residuals

indicated that there were four significant differences. Compared with unlicensed drivers

as a whole, males were relatively over-represented in the inappropriate class of licence

and disqualified/suspended driver groups, but under-represented in the never licensed and

expired groups. (Nonetheless, the proportion of males in the never licensed and expired

groups was still higher than was the case for licensed drivers - see Table 4.4).

Table 4.5

Age and gender of unlicensed drivers involved in serious casualty crashes in

Queensland: 1994-98 by offender type

Unlicensed Driver Type

Variable Dis- qualified/ suspended

%

Never licensed

%

Inappropriate class

%

Expired

%

Other

%

Total

%

Significance level1

Gender n=388 n=302 n=167 n=182 n=444 n=1483

Males 90.5 76.5 96.4 74.7 80.2 83.3 χ2 (df4) = 57.7,

Females 9.5 23.5 3.6 25.3 19.8 16.7 p < .001, φc? =.20

Age n=388 n=302 n=167 n=182 n=444 n=1483

Under 17 0.5 55.6 0.6 0.0 4.5 12.9 χ2 (df12) = 666.0,

17- 24 47.4 33.1 47.9 39.0 45.0 42.8 p < .001, φc? =.39

25 – 59 51.5 10.9 51.5 57.7 47.1 42.7

60 and over 0.5 0.3 0.0 3.3 3.4 1.6


There was also significant age difference among the unlicensed drivers. As would

be expected, the under 17 year old drivers were relatively over-represented among the

never licensed drivers. In addition, there was a higher representation of 25-59 year old


drivers among the disqualified/suspended and expired licence holders. This probably

reflects the longer time periods that these licences would have been held and hence the

greater opportunity for the drivers to be disqualified/suspended or fail to renew their

licence. It is also interesting that there was an over-representation of 60 years of age and

over drivers in the other unlicensed group. This may in part be due to the detection of

people driving outside the conditions of a restricted licence granted on medical grounds.

4.4.3 Crash circumstances

Table 4.6 compares the circumstances of the crashes involving unlicensed drivers

with those involving licensed drivers. The first variable in Table 4.6 relates to the vehicle

driven at the time of the crash. As can be seen, the majority of the unlicensed drivers

[68.7%] were driving cars (or car derivatives) at the time they crashed. However, there

was a significantly higher representation of motorcycle riders among the unlicensed

drivers than among those with a valid licence. Almost one-third [29.7%] of the

unlicensed drivers were riding a motorcycle compared with only 9.9% of the licensed

drivers.

The next four variables in Table 4.6 relate to the time and location of the crash.

Significant differences were found between the licensed and unlicensed drivers in terms

of both the time of the crash and the day of the week it occurred. In particular, there was

an over-representation of unlicensed drivers in crashes at night-time and on the

weekends, times generally associated with recreational driving. There are a number of

possible explanations to account for these results. Firstly, it may the case that unlicensed

drivers are relatively more likely to drive at these times than licensed drivers. Secondly, it

is possible that unlicensed drivers are more likely to engage in risk-taking during

recreational driving, effectively increasing their crash risk at these times. Finally, it is

possible that these two factors could be working together to increase the crash

involvement of unlicensed drivers at night and on weekends.


Table 4.6

Circumstances of crashes resulting in serious casualty crashes in Queensland: 1994-98

by licence status


(%) Unlicensed

(%)

Significance level1

Vehicle type (n=25708) (n=1484)

Car (and derivatives) 82.7 68.7 χ2 (df2) = 603.5, p < .001, Motorcycles 9.9 29.7 φc? = .15

Trucks and buses 7.4 1.5 Time of day (n=25708) (n=1484)

Day (6:00am - 5:59pm) 70.5 52.3 χ2 (df1) = 220.8, p < .001

Night (6:00pm – 5: 59am) 29.5 47.7 φ = -.09

Day of week (n=25708) (n=1484)

Monday-Friday 71.9 64.3 χ2 (df1) = 40.3, p < .001

Saturday-Sunday 28.1 35.7 φ = -.04

Prevailing speed zone (n=25708) (n=1484)

60 km/h or less 60.1 59.4 χ2 (df2) = 0.4, p > .05

70 – 90 km/h 12.0 12.1 φc? = .00 100/110 km/h 27.9 28.6

Traffic control (n=25708) (n=1484)

Traffic control present 29.0 16.4 χ2 (df1) = 109.5, p < .001 No traffic control 71.0 83.6 φ = -.06

Vehicle involvement (n=23235) (n=1354)

Single vehicle crash 28.8 55.3 χ2 (df1) = 425.6, p < .001 Multi-vehicle crash 71.2 44.7 φ = .13

Hit pedestrian crash (n=25708) (n=1484)

Yes 7.7 3.7 χ2 (df1) = 32.5, p < .001 No 92.3 96.3 φ = -.04


Interestingly, there was no significant difference between the licensed and

unlicensed drivers in terms of the prevailing speed limit at the location where the crash

occurred. This suggests that both groups are equally exposed across the road network.

However, there was a significant difference between licensed and unlicensed driver crashes

in terms of the prevailing traffic conditions. As shown in Table 4.5, the crashes involving

unlicensed drivers were less likely to occur at locations where there was some form of

traffic control in place. This variable relates to the presence of controls such as traffic


lights, stop and give way signs and pedestrian crossings. It tends to suggest that unlicensed

drivers are more likely to lose control of their vehicle in general driving situations, without

any apparent conflicts.

This interpretation is consistent with the over-representation of unlicensed drivers in

single vehicle crashes. Over half of the serious casualty crashes involving unlicensed

drivers [55.3%] were single vehicle crashes, compared with only 28.8% of those

involving licensed drivers. Once again, this suggests that unlicensed drivers are more

likely to lose control of their vehicles in non-conflict situations; a type of behaviour often

associated with impairment or speeding. It is somewhat encouraging, however, that

unlicensed drivers were significantly less likely to hit a pedestrian than a licensed driver.

This may be indicative of a greater amount of night-time driving by unlicensed drivers.

Table 4.7 examines differences in the circumstances of the crashes involving the

various types of unlicensed drivers. As can be seen, there was a significant difference

between the groups in relation to the vehicle driven, with motorcycles particularly over-

represented among those offenders with an inappropriate class of licence. Indeed, this

group was almost exclusively [92.8%] made up of motorcycle riders. In contrast, there

was an over-representation of car drivers among the disqualified/suspended, expired and

other unlicensed drivers (which was similar to the overall proportion of car drivers

among licensed drivers – see Table 4.6).

The offenders with an inappropriate class of licence who were involved in crashes

riding motorcycles could have been unlicensed for two reasons. Firstly, they could have

been illegally riding because they only held a licence for another type of vehicle, such as

a car. Alternatively, they could have held a provisional motorcycle licence (RE) but been

riding a larger motorcycle (>250mls) than permitted on this licence. It is impossible to

distinguish between these two groups in the database. However, a separate analysis of the

155 motorcycle riders who held an inappropriate class of licence indicated that they were

mainly male [97.4%], with just over half [50.3%] being 25 years of age or over.


Table 4.7

Circumstances of crashes resulting in serious casualty crashes in Queensland: 1994-98

by unlicensed driver type

Unlicensed Driver Type Variable Dis-

qualified/ suspended

%

Never licensed

%

Inappropriate class

%

Expired

%

Other

%

Total

%

Significance level1

Vehicle type n=388 n=303 n=167 n=182 n=444 n=1484

Car (and derivatives)

75.5 72.9 4.2 84.1 77.9 68.7 χ2 (df8) = 386.5,

Motorcycles 21.9 27.1 92.8 13.7 21.2 29.7 p < .001, φc? =.36

Trucks/buses 2.6 0.0 3.0 2.2 0.9 1.5

Time of day n=388 n=303 n=167 n=182 n=444 n=1484

Day (6am–5:59pm)

45.6 47.5 62.3 59.3 54.7 53.3 χ2 (df4) = 21.0,

Night (6pm–5: 59am)

54.4 52.5 37.7 40.7 45.3 47.7 p < .001, φ =.12

Day of week n=388 n=303 n=167 n=182 n=444 n=1484

Mon.-Fri. 62.1 57.8 67.7 68.1 67.8 64.3 χ2 (df4) = 10.8,

Sat.-Sun. 37.9 42.2 32.3 31.9 32.2 35.7 p > .01, φ =.09

Prevailing speed zone

n=388 n=303 n=167 n=182 n=444 n=1484

≤ 60 km/h 53.9 58.7 62.3 64.8 61.3 59.4 χ2 (df8) = 8.9,

70 – 90 km/h 14.4 12.2 10.2 10.4 11.3 12.1 p > .05, φc? =.08

100/110km/h 31.7 29.0 27.5 24.7 27.5 28.6

Traffic control n=388 n=303 n=167 n=182 n=444 n=1484

Present 14.2 10.2 15.6 19.2 21.6 16.4 χ2 (df4) = 19.8,

Not present 85.8 89.8 84.4 80.8 78.4 83.6 p = .001, φ =.12

Vehicle involvement

n=388 n=303 n=167 n=182 n=444 n=1484

Single vehicle 57.3 70.2 32.1 45.6 55.0 55.3 χ2 (df4) = 61.7,

Multi-vehicle 42.7 29.8 67.9 54.4 45.0 44.7 p < .001, φ =.21

Hit pedestrian n=388 n=303 n=167 n=182 n=444 n=1484

Yes 3.4 3.0 2.4 4.4 4.7 3.7 χ2 (df4) = 2.9,

No 96.6 97.0 97.6 95.6 95.3 96.3 p < .001, φ =.05


Returning to Table 4.7, it can be seen that there was an overall significant

difference among the unlicensed driver sub-groups in terms of the time of their crashes.

The drivers with an inappropriate class of licence were relatively over-represented in

crashes during the day (although less so than licensed drivers), while the

disqualified/suspended drivers were over-represented at night. Although not significant,


the never licensed drivers were also relatively over-represented in night-time crashes.

However, there was no significant differences between the groups in relation to whether

the crashes occurred on a weekday or weekend.

Similarly, there were no significant differences in relation to the prevailing speed

limit where the crash occurred. However, there was a significant difference for the

presence of traffic control devices. The never licensed drivers were over-represented in

cases where the devices were not present (compared with both other unlicensed drivers

and licensed drivers). This suggests that these drivers experience difficulties in general

(non-conflict) driving conditions, which may be indicative of a lack of driving skills. In a

similar vein, there was a significantly higher representation of the never licensed drivers

in single vehicle crashes. This may once again be indicative of less driving skill or,

alternatively, a higher degree of impairment or speeding (see below). In contrast, the

drivers with an expired or inappropriate class of licence were relatively under-

represented in single vehicle crashes (although less so than licensed drivers). In the case

of the inappropriate class group, this may reflect the tendency for motorcyc les to come

into conflict with cars. There were no significant differences between the various types of

unlicensed drivers in terms of hitting pedestrians.

4.4.4 Contributing factors to crashes

Table 4.8 provides a summary of some of the key contributing factors cited by the

police in the serious casualty crashes involving licensed and unlicensed drivers. Some

care should be taken when interpreting this data since it is based on the assessment of the

attending police, most of whom would have had no special training in crash investigation.

While the data concerning alcohol and other drugs would have been corroborated by a

breath or blood test, much of the other data would be based on the testimony of the

involved parties and the subjective judgement of the attending police.

As can be seen, there was a significantly higher involvement of alcohol and other

drugs in the crashes involving the unlicensed drivers. Indeed, alcohol and other drugs

were cited as a factor in the crashes involving over one quarter [25.9%] of the unlicensed

drivers. The speeding variable is a composite of two factors cited by the police:

exceeding the speed limit and excessive speed for the conditions. Once again, the

unlicensed drivers were over-represented on this variable, with speeding being implicated

as a factor in the crashes of 14.7% of the unlicensed drivers compared with only 3.5% of

the licensed drivers.


Table 4.8

Contributing factors to serious casualty crashes in Queensland: 1994-98 by licence

status


(%) Unlicensed

(%)

Significance level1

Alcohol/drugs (n=25698) (n=1484)

Yes 5.8 25.9 χ2 (df1) = 886.6, p < .001 No 94.2 74.1 φ = .18

Speeding (n=25698) (n=1484)

Yes 3.5 14.7 χ2 (df1) = 443.2, p < .001 No 96.5 85.3 φ = .13

Fatigue (n=25698) (n=1484) χ2 (df1) = 0.1, p > .05

Yes 2.2 2.4 φ = .00 No 97.8 97.6

Inattention (n=25708) (n=1484) χ2 (df1) = 1.0, p > .05

Yes 0.1 0.2 φ = .01 No 99.9 99.8

Inexperience (n=25708) (n=1484)

Yes 13.4 26.2 χ2 (df1) = 190.2, p < .001

No 86.6 73.8 φ = .08

At fault (n=25708) (n=1484)

Yes 56.7 82.1 χ2 (df1) = 374.0, p < .001

No 43.3 17.9 φ = .12 1. The cells with significant (p<.01) adjusted standardised residuals are bolded.


No significant differences were found between the licensed and unlicensed drivers

for either driver fatigue or inattention. However, the unlicensed drivers were over-

represented in terms of inexperience, with it being implicated in the crashes of over one

quarter [26.2%] of the unlicensed drivers. Finally, unlicensed drivers were much more

likely to be found at fault for the serious casualty crashes in which they were involved.

Over four-fifths [82.1%] of the unlicensed drivers were considered to be at fault

compared with only 56.7% of the licensed drivers. This may in part reflect the negative-

halo effect referred to by DeYoung et al (1997). However, it is also consistent with the

greater involvement of factors like alcohol/drugs, speeding and inexperience in the

crashes involving unlicensed drivers.


Table 4.9 compares the contributing factors to the crashes involving the different

types of unlicensed drivers. As shown, there was a significant difference among the

offenders in relation to the presence of alcohol and other drugs. The

disqualified/suspended drivers were over-represented with almost one-third [32.5%]

having alcohol or drugs present at the time of the crash. In contrast, this factor was under-

represented [16.2%] among the inappropriate class of licence. However, even among this

group, the proportion of drivers with alcohol/drugs present was still considerably higher

than that for licensed drivers [5.8%] (see Table 4.8). Therefore, while drink/drug driving

appears to be more common among the disqualified/suspended drivers, it represents a

serious problem for all the different types of unlicensed drivers.

Table 4.9

Contributing factors to serious casualty crashes in Queensland: 1994-98 by unlicensed

driver type


qualified/ suspended

%

Never licensed

%

Inappropriate class

%

Expired

%

Other

%

Total

%

Significance level1

Alcohol/drugs n=388 n=303 n=167 n=182 n=444 n=1484

Yes 32.5 21.8 16.2 22.5 27.9 25.9 χ2 (df4) = 21.7,

No 67.5 78.2 83.8 77.5 72.1 74.1 p < .001, φ =.12

Speeding n=388 n=303 n=167 n=182 n=444 n=1484

Yes 13.7 17.8 10.8 11.0 16.4 14.7 χ2 (df4) = 7.8,

No 86.3 82.2 89.2 89.0 83.6 85.3 p > .05, φ =.07

Fatigue n=388 n=303 n=167 n=182 n=444 n=1484

Yes 3.9 1.7 0.0 1.6 2.7 2.4 χ2 (df4) = 9.2,

No 96.1 98.3 100.0 98.4 97.3 97.6 p > .05, φ =.08

Inattention2 n=388 n=303 n=167 n=182 n=444 n=1484

Yes 0.0 0.7 0.6 0.0 0.0 0.2 χ2 (df4) = 6.5,

No 100.0 99.3 99.4 100.0 100.0 99.8 p > .05, φ =.07

Inexperience n=388 n=303 n=167 n=182 n=444 n=1484

Yes 12.9 62.4 27.5 16.5 16.7 26.2 χ2 (df4) = 270.5,

No 87.1 37.6 72.5 83.5 83.3 73.8 p < .001, φ =.43

At fault n=388 n=303 n=167 n=182 n=444 n=1484

Yes 85.6 92.4 71.3 69.2 81.5 82.1 χ2 (df4) = 59.2,

No 14.4 7.6 28.7 30.8 18.5 17.9 p < .001, φ =.20

1. The cells with significant (p<.01) adjusted standardised residuals are bolded. 2. This test violated recommendations relating to the minimum number of cells allowed with an expected value of less

than 5. However, inspection of the data suggests that no significant difference was evident. Source: Queensland Road Crash Database, Queensland Transport


No significant differences were found among the unlicensed drivers in terms of the

involvement of speed, fatigue or inattention. However, a significant difference was found

in relation to the role of inexperience. Not surprisingly, the never licensed drivers were

over-represented on this factor. These results are consistent with those cited in Table 4.7,

relating to the role of traffic control devices and single vehicles in the crashes involving

never licensed drivers.

A significant difference was also found between the groups in relation to the

proportion of drivers considered at fault for the crash. Once again, the never licensed

drivers were over-represented on this factor with 92.4% considered at fault. This result

would partly be a product of the higher proportion of single vehicle crashes among this

group (see Table 4.7).14 However, it is also consistent with other findings indicating their

lack of experience and tendency to crash in general traffic conditions (ie. with no traffic

devices present). Interestingly, the drivers with an expired or inappropriate class of

licence were significantly less likely to be judged at fault than the other unlicensed driver

types. Nonetheless, the proportion at fault in these two sub-groups was still relatively

higher than that among the licensed drivers [56.7%]. This may be indicative of a

negative-halo effect or, alternatively, that intentional risk-taking may still represent a

problem among these groups.

4.4.5 The crash risk of unlicensed drivers

As mentioned in the method section, the lack of reliable exposure data for

unlicensed drivers makes it difficult to compare their crash risk with that of licensed

drivers. However, this is an important issue for determining the most appropriate

allocation of resources to the unlicensed driver problem. Consequently, the following

section attempts to address this issue through the use of two different approaches. The

first approach is based on quasi- induced exposure procedures and examines the risk of

unlicensed drivers being involved in a crash (relative to licensed drivers). The second

approach examines the outcomes of the crashes in which unlicensed and licensed drivers

are involved. This is designed to establish the risk of death and serious injury in the event

of a crash associated with unlicensed driving.

14. It should be noted that drivers involved in single-vehicle crashes are not automatically deemed to be at

fault. It is possible that fault may be attributed to a non-motorised road user involved in the crash, such as a pedestrian, cyclist, skate-board rider or a ridden animal. However, the vast majority of single vehicle crashes do not involve non-motorised road users.


4.4.5.1 Risk of involvement in crashes

Table 4.10 summarises the risk of involvement in a multi-vehicle crash for variety

of different driver groups, based on the quasi- induced exposure method (see section

4.3.2). The risk of being involved in a crash is expressed as an odds ratio with the

licensed drivers represent ing the primary reference category. Odds ratios are provided for

each of the different crash severities, as well as for total crashes, with the 99% confidence

limits shown in brackets. Due to the small number of unlicensed drivers involved in fatal

multi-vehicle crashes (n = 47), a fatal crash odds ratio was only calculated for unlicensed

drivers as a whole rather than for each of the sub-groups.

Table 4.10

Risk of involvement in a multi-vehicle crash by driver type and crash severity for

Queensland: 1994-98 (99% CI)

Risk of involvement in a crash

Driver group Fatal crash

Serious injury crash

Other injury crash PDO crash Total

crashes

Licensed drivers1

1.00

- n=1191

1.00 -

n=15348

1.00 -

n=46801

1.00 -

n=58955

1.00 -

n=122295

All unlicensed drivers

2.72

(0.53 – 13.97) n=47

2.03 (1.29 – 3.20)

n=558

2.75 (1.91 – 3.97)

n=925

3.80 (2.61 – 5.53)

n=1030

2.90 (2.32 – 3.62)

n=2560

Never licensed drivers

5.43

(1.24 – 23.72) n=81

3.93 (1.29 – 12.00)

n=117

7.52 (2.17 – 26.12)

n=142

5.38 (2.63 – 10.99)

n=343

Disqualified/ suspended drivers

2.73

(1.06 – 7.05) n=138

3.38 (1.58 – 7.23)

n=235

5.48 (2.42 – 12.39)

n=267

3.84 (2.39 – 6.16)

n=654

Other unlicensed drivers

1.86

(0.83 – 4.17) n=173

2.72 (1.49 – 4.98)

n=339

3.24 (1.83 – 5.73)

n=414

2.73 (1.90 – 3.93)

n=938 Expired licence drivers

1.26

(0.42 –3.79) n=87

2.74 (1.14 – 6.61)

n=161

2.33 (1.06 – 5.12)

n=188

2.13 (1.28 – 3.56)

n=442

Inappropriate class of licence drivers

1.14

(0.36 – 3.61) n=79

1.02 (0.30 – 3.46)

n=73

4.62 (0.25 – 84.20)

n=19

1.33 (0.62 – 2.84)

n=183 1. Primary reference category. The significant (p < .01) crash odds ratios are shown in bold. Source: Queensland Road Crash Database, Queensland Transport


Overall, unlicensed drivers were significantly more likely to be involved in a crash

than licensed drivers, at various crash severities. The highest risk related to PDO crashes

[3.8:1], while the lowest was for serious injury crashes [2.03:1]. Although the risk of an

unlicensed driver being involved in a fatal crash [2.72:1] was not significant, it was very

similar to that for total crashes [2.9:1]. As will be discussed later, these results are not that

dissimilar to those obtained by DeYoung et al (1997).

In addition, the pattern of results for the unlicensed driver sub-groups was relatively

stable across the other crash severities. While there is some movement up and down in

the ratios, the ordering of the sub-groups generally remains consistent, with the never

licensed drivers having the highest risk followed by the disqualified/suspended drivers. In

all but one case, the lowest ratios were found for the drivers with an inappropriate class

of licence followed by those with an expired licence. Although not significant, one

exception to this pattern is the ratio for the drivers with inappropriate class of licence

involved in PDO crashes [4.6:1]. This ratio is substantially higher than the other ones for

this group and may in part reflect the relatively small size of the group [n=19]. It should

also be borne in mind that this group mainly consists of motorcycles. Hence, it is possible

that motorcycle riders may be more likely to be found at fault in non-injury crashes (or

conversely, that they are less likely to be found at fault for crashes in which they are

injured).

While the general stability of the results obtained supports the validity of the

approach, it is important to explore potential biases or limitations of the quasi- induced

exposure method. Figure 4.1 examines the total crash involvement rate for each of the

groups of drivers by individual year.


0

2

4

6

8

10

12

14

1994 1995 1996 1997 1998

Licensed drivers Unlicensed

Never licensed Innappropriate class

Disqualified/suspended Expired

Other unlicensed

Figure 4.1: Involvement rate (IR) for total crashes by driver type: 1994-98

As can be seen in Figure 4.1, the involvement rates for most of the groups were

relatively stable over time. The major exception was for the never licensed drivers who

recorded much higher rates in 1997 and 1998. There is no obvious reason for this

fluctuation, other than the smaller numbers in this group (which was indicative of their

under-representation in multi-vehicle crashes). It is also possible that the fluctuation may

reflect some unknown change in reporting practices. As such, some care should be taken

when interpreting the results for the never licensed drivers.

Another potential source of bias relates to the possib ility that the representation of

unlicensed drivers in multi-vehicle crashes is not sufficiently similar to their involvement

in total crashes. Table 4.11 compares the involvement of various driver groups in all

crashes with their involvement in multi-vehicle crashes alone, for the different crash

severities.


Table 4.11

The involvement of drivers in multi-vehicle crashes compared with all crashes, by

severity for Queensland: 1994-98

Severity of Crash


Other Injury


Licence Type1 Multi-

vehicle All

crashes Multi-

vehicle All

crashes Multi-

vehicle All

crashes Multi-

vehicle All

crashes Multi-

vehicle All

crashes

Licensed 96.2 93.3 96.5 94.7 98.1 97.3 98.3 97.4 97.9 96.9

All unlicensed 3.8 6.7 3.5 5.3 1.9 2.7 1.7 2.6 2.1 3.1

Expired 0.5 0.9 0.5 0.7 0.3 0.4 0.3 0.4 0.4 0.4

Inapprop-riate class

1.0 0.9 0.5 0.6 0.2 0.2 0.0 0.0 0.1 0.2

Disqualified/ suspended

1.1 2.1 0.9 1.4 0.5 0.7 0.4 0.7 0.5 0.8

Never licensed

0.2 1.1 0.5 1.1 0.2 0.4 0.2 0.5 0.3 0.6

Other unlicensed

1.0 1.7 1.1 1.6 0.7 1.0 0.7 1.0 0.8 1.1

1. Drivers with an international licence or of unknown licence status are excluded. Source: Queensland Road Crash Database, Queensland Transport

Overall, the unlicensed drivers are under-represented in multi-vehicle crashes

compared with all crashes. This applies at all the crash severities but appears most

pronounced for fatal crashes in general, and among the never licensed and

disqualified/suspended drivers in particular. (Interestingly, the one exception concerns

the involvement of the inappropriate licence drivers in fatal crashes. This group has a

slightly higher involvement in multi-vehicle crashes than all crashes. This most likely

reflects the large proportion of motorcyclists in this group, who are particularly

vulnerable when hit by another vehicle). The under-representation of unlicensed drivers

in multi-vehicle crashes is a direct product of their over-involvement in single vehicle

crashes (see Table 4.6). Furthermore, it was the never licensed and disqualified/

suspended drivers who were most over-represented in single vehicle serious casualty

crashes (see Table 4.7).

These results raise some important questions about the validity of the quasi- induced

exposure method. In particular, it introduces a potential source of bias into the calculation

of the involvement rate for the unlicensed drivers. As shown in Table 4.12, both licensed

and unlicensed drivers are more likely to be considered at fault in single-vehicle crashes


than multi-vehicle crashes.15 Consequently, an involvement rate based on multi-vehicle

crashes will include a higher representation of innocent drivers than would be the case if

the single vehicle crashes were included. Furthermore, unlicensed drivers are more likely

to be considered at fault in single vehicle crashes than licensed drivers (see Table 4.12).

As such, the involvement rate for the unlicensed drivers in multi-vehicle crashes will tend

to underestimate the overall extent of their at- fault driving, compared to the licensed

drivers. This suggests that the odds ratios obtained for the unlicensed drivers may

actually represent an under-estimate of their crash risk. It is possible that this bias is

countered by the negative halo effect (ie. the tendency for police to assign blame to the

unlicensed drivers), since it would tend to inflate the involvement rates for the unlicensed

drivers, even in multi-vehicle crashes. However, the net effect of these biases is difficult

to assess. Therefore, the reliance of the quasi- induced method on multi-vehicle crashes

appears to introduce potential sources of bias that are problematic to resolve.

Table 4.12

Percentage of at -fault drivers in single and multi-vehicle crashes by licence status and

severity of crash for Queensland: 1994-98

Severity of Crash


Other Injury

Property Damage Total Licence

Type Single vehicle

Multi-vehicle

Single vehicle

Multi-vehicle

Single vehicle

Multi-vehicle

Single vehicle

Multi-vehicle

Single vehicle

Multi-vehicle

Licensed 90.7 44.0 94.6 44.8 93.5 44.7 94.5 44.8 94.1 44.8

All unlicensed 100.0 68.1 99.1 62.3 99.0 69.0 99.2 75.5 99.2 70.1

Source: Queensland Road Crash Database, Queensland Transport Finally, it was not possible with the dataset provided to identify the particular

drivers involved in crashes with each other. As a result, it was not possible to check

whether the at-fault drivers were any more likely to collide with another type of driver or

not (ie. after adjusting for the representation of the different groups in crashes). This

made it impossible to check the validity of the assumption that “at-fault drivers ‘choose’

their innocent victims at random from all drivers present” (DeYoung, 1997, p.20).

15. As noted earlier, drivers involved in single-vehicle crashes are not automatically deemed to be at fault,

since blame could be attributed to a non-motorised road user (see Footnote 14 on page 86). However, the majority of single-vehicle crashes do not involve non-motorised road users. Accordingly, the over-representation of at-fault drivers in single vehicle crashes is primarily due to the fewer road users involved.


4.4.5.2 Risk of death and serious injury in the event of a crash

Another method of examining the risks associated with unlicensed driving is to

compare the outcomes of crashes involving this group with those for licensed drivers.

For example, it is possible to measure whether the crashes involving these two groups are

equally likely to result in a fatality or serious injury (ie. a serious casualty) as opposed to

a minor injury or property damage only. Table 4.13 compares the risk of being involved

in a serious casualty crash (relative to a minor crash) for licensed and unlicensed drivers.

The relative risk is estimated by the odds ratio with 99% confidence intervals. Unlicensed

drivers as a whole and each of the sub-groups were compared with all licensed drivers.

Table 4.13

Risk of involvement in a serious casualty crash, relative to a minor crash, for different

licence categories in Queensland: 1994-98 (99% CI)

Type of driver Odds ratio risk 99% CI

All licensed drivers1 1.00 ----

All unlicensed drivers 2.12 1.79 – 2.51

Inappropriate class 5.75 3.21 – 10.29

Never licensed 2.33 1.63 – 3.34

Disqualified/suspended 2.13 1.56 – 2.92

Expired 1.80 1.15 – 2.81

Other unlicensed 1.73 1.30 – 2.31

1. Primary reference category. Source: Queensland Road Crash Database, Queensland Transport

As can be seen, the risk of being involved in a more severe crash was significantly

higher for all the groups of unlicensed drivers, compared to licensed drivers. As a whole,

unlicensed drivers are 2.12 times more likely to be involved in a serious casualty (relative

to a minor crash) than licensed drivers. The group most at risk were the drivers with an

inappropriate class of licence, followed by the never licensed and the

disqualified/suspended drivers. The groups with the lowest risk were the other

unlicensed drivers and those with expired licences.

It is likely that the result for the inappropriate class of licence reflects the high

proportion of motorcyclists in this group. Due to their unprotected nature, motorcyclists

are particularly vulnerable to death and serious injury in the event of a crash. In addition,

it is possible that the under-reporting of minor crashes is more extensive among this

group. Consequently, the high risk experienced by those with an inappropriate class of


licence may in part be an artefact of the under-reporting of minor crashes among this

group.

While the results suggest that the crashes involving unlicensed drivers are more

likely to result in a death or serious injury than those involving licensed drivers, the exact

reasons for this remain unclear. Firstly, it is possible that the results partly reflect the

under-reporting of minor crashes among unlicensed drivers. This is plausible, given the

illegal nature of the behaviour, particularly in the case of single-vehicle crashes.

Secondly, it is possible that the results are more indicative of the characteristics of the

people who decide to drive without a licence and the types of vehicle they drive (eg. a

motorcycle), rather than the nature of unlicensed driving per se. To explore this issue, a

logistic regression was conducted to examine the contribution of these factors to the

severity of road crashes. The dependent variable in this analysis was a dichotomous

variable measuring whether the crash resulted in a serious casualty or minor injury/PDO.

The independent variables were: gender; age (measured using dummy variables

corresponding to the age categories 12-20, 21-29, 30-49, 50-69, 70 and over); vehicle

type (motorcycle vs. car/truck/bus) and licence status (unlicensed vs. licensed).

The results of the logistic regression are reported in Table 4.14. Gender was a

significant predictor of crash severity with females having a lower risk than males of

being involved in a serious casualty crash. The relationship between age and crash

severity appeared to be linear, with older drivers having a greater risk of being involved

in a more severe crash. (Indeed, a separate analysis using age as a continuous variable,

based on the midpoints for the age categories, produced similar results.) The higher risk

experienced by the drivers aged 70 and over [1.52] is consistent with the effects of frailty

on injury outcomes. The type of vehicle driven appears to exert a particularly strong

influence on the severity of crashes. Riding a motorcycle was associated with four and

half times [4.58] the risk of a crash resulting in a serious casualty rather than a minor

injury or PDO. Finally, being unlicensed was associated with an increased likelihood

[1.82 times] of being involved in a serious casualty crash. Therefore, accounting for the

influence of gender, age and vehicle type, reduced the odds of an unlicensed driver being

involved in a more severe crash from 2.12 (as reported in Table 4.13) to 1.82.

Nonetheless, it remains a concern that being unlicensed is associated with almost double

the risk of being involved in a more severe road crash.


Table 4.14

Logistic regression analysis of the severity of crashes as a function of unlicensed driving

and selected driver-related variables (n=162,618)

99% CI for Odds ratio Variables B Std.

error Wald test

Odds Ratio

Upper Lower

Gender

Male 0 1.00

Female -.08 .02 31.6* 0.92 0.87 0.96

Age

12-20 0 1.00

21-29 .02 .02 0.7 1.02 0.96 1.07

30-49 .06 .02 8.0 1.06 1.01 1.11

50-69 .14 .02 34.5* 1.15 1.08 1.23

70 and over .42 .03 150.1* 1.52 1.39 1.66

Vehicle type

Car/truck/bus 0 1.00

Motorcycle 1.52 .03 3258.3* 4.58 4.28 4.91

Licence status

Licensed 0 1.00

Unlicensed .60 .03 321.5* 1.82 1.67 1.98

Full model vs. constant-only model: χ2 (df1) = 3754.8, p < .001; Nagelkerke R2 = .04. * p < .001

Table 4.15 reports the results of a second logistic regression conducted to examine

the influence of the different unlicensed driver types on crash severity. To facilitate this

analysis a series of dummy variables were created reflecting the involvement or not of the

various unlicensed driver types. As can be seen, the unlicensed driver types with the

highest odds ratios were the never licensed drivers [2.01] and the disqualified/suspended

drivers [1.96]. Interestingly, after taking the driver-related variables into account, the

odds ratio for the inappropriate class of licence group fell to 1.88 (from 5.75 in Table

4.13). This tends to confirm that it was the motorcycle use within this group that exerted

a strong influence on the severity of the crashes in which they were involved.


Table 4.15

Logistic regression analysis of the severity of crashes as a function of unlicensed driver

types and selected driver-related variables (n=162,168)


error Wald test

Odds Ratio

Upper Lower

Gender

Male 0 1.00

Female -.08 .02 31.2* 0.92 0.88 0.96

Age

12-20 0 1.00

21-29 .02 .02 1.0 1.02 0.97 1.08

30-49 .06 .02 8.8 1.06 1.01 1.12

50-69 .14 .02 35.7* 1.16 1.09 1.23

70 and over .42 .03 151.6* 1.53 1.40 1.67

Vehicle type

Car/truck/bus 0 1.00

Motorcycle 1.52 .03 3163.6* 4.57 4.27 4.90

Licence status

Licensed 0 1.00

Never licensed .70 .07 87.9* 2.01 1.66 2.43

Disqualified/suspended .67 .06 115.1* 1.96 1.67 2.31

Inappropriate class .63 .12 27.7* 1.88 1.38 2.56

Expired .57 .09 41.4* 1.76 1.41 2.21

Other unlicensed .49 .06 73.8* 1.63 1.41 1.89

Full model vs. constant-only model: χ2 (df11) = 3762.0, p < .001; Nagelkerke R2 = .04 * p < .001

Together, these logistic regressions suggest that being unlicensed increases the

likelihood of being involved in a serious casualty crash (rather than a minor crash) even

after taking account of key driver characteristics like gender, age and type of vehicle

driven. This is most evident among the never licensed and disqualified/suspended drivers,

who both have a risk approximately double that of licensed drivers. These results are

consistent with the evidence presented in section 4.44, indicating that unlicensed drivers

are more likely to engage in risky on-road behaviours (such as drink driving and

speeding) and be involved in the types of crashes which are indicative of loss of control

and inexperience (eg. single vehicle crashes). These are all factors that tend to increase

the severity of crashes. Moreover, the pattern of results is largely consistent with those

obtained from the quasi- induced exposure method. In particular, they suggest that the


never licensed and disqualified/suspended drivers are particularly at risk of involvement

in a crash and that their crashes tend to be more severe.

4.5 Discussion

4.5.1 Study limitations

The use of official crash statistics provides a firm foundation for research into

unlicensed drivers. It represents a means of investigating both the extent to which

unlicensed drivers are involved in road crashes and the nature of these crashes. This

provides an important insight into the road safety implications of unlicensed driving.

From a traffic psychology point of view, it provides an insight into the types of

behaviours that unlicensed drivers engage in, compared to licensed drivers. Nonetheless,

the use of official crash statistics is associated with certain constraints that may affect the

quality of the data.

Firstly, the data included in mass-crash databases like Queensland’s is limited to

crashes reported to the police. While it is unlikely that many serious crashes go

unreported, drivers are only required to report property damage only crashes where the

damage exceeds $2,500 or a vehicle is towed away (Queensland Transport, 1999).

Consequently, the possibility exists for drivers to avoid reporting crashes, particularly

where the crash in relatively minor in nature and/or no other road users are involved. As

will be discussed later, this has important implications for comparing the severity of

crashes involving licensed and unlicensed drivers.

Secondly, there were some constraints placed on the comparison of the different

types of unlicensed drivers by the reporting practices of the police. As noted in the

method section, the crash database does not distinguish between those drivers who have

been disqualified (by a court order) from driving and those who are suspended (generally

for accumulation of demerit points). Consequently, these two categories of unlicensed

drivers were combined in this particular study. In addition, the not currently licensed

drivers were not specifically identified, while there was a relatively large category of

other unlicensed drivers who could not be re-allocated to any of the other categories.

These reporting practices reduced the precision with which different types of unlicensed

drivers could be identified in the study.

Thirdly, much of the information in the crash database rela ting to the contributing

factors to crashes (eg. the involvement of speed, fatigue, inexperience etc) and the

attribution of fault is based on the judgement of the attending police. Moreover, it is

possible that some of these judgements may be influenced by the discovery that a driver


involved in a crash is unlicensed. For example, it is possible that the police may be more

likely to find an unlicensed driver at fault for the crash (what DeYoung et al, 1997 refer

to as a negative halo effect) and to attribute this to certain factors, such as inexperience.

To minimise this potential bias, the analyses generally considered all drivers involved in

crashes, irrespective of whether they were judged by the police to be at fault for the crash

or not. However, the data relating to at- fault drivers was an important requirement of the

quasi- induced exposure method, utilised to calculate the drivers’ risk of involvement in a

crash.

The need to use the quasi- induced exposure method reflects another major

limitation of this study. Unfortunately, there is a lack of reliable information available

relating to the exposure patterns of unlicensed driving. This makes it difficult to estimate

the prevalence of unlicensed driving and the crash risk associated with the behaviour.

While it is possible to use crash data to provide a surrogate measures of exposure, this

introduces potential biases associated with police reporting practices and the under-

reporting of crashes by drivers (discussed above). In addition, the validity of the quasi-

induced exposure method remains unclear. There appears to be a number of potential

biases that could arise from the reliance on multi-vehicle crashes and the attribution of

fault by the police. These will be discussed in greater detail in the next section.

Fourthly, the data utilised in this study was drawn exclusively from one jurisdiction

within Australia. Hence, it is unclear how representative the data is of the crash

involvement of unlicensed drivers in other parts of Australia or the world. However, it is

important to note that the conventions for reporting crashes are fairly uniform across

Australia, particularly for serious crashes (eg. ATSB, 2002). Furthermore, there are

strong similarities between the findings of the current study and those obtained in other

Australian and international research. For example, Job, Lee and Prabhakar (1994) cite

very similar results for the overall involvement of unlicensed drivers in fatal [5.9%],

serious injury [5.3%] and total crashes [2.5%] in NSW. Harrison (1997) reports that

disqualified drivers (which included those who had been suspended, cancelled or

disqualified) constituted 2.4% of those involved in fatal crashes and 1% of all crash-

involved drivers in Victoria in 1994. As already noted, the crash involvement rates

calculated for drivers in this study are reasonably similar to those obtained by DeYoung

et al (1997) in California. While the similarity of these findings suggests that the

Queensland crash data may be reasonably representative of unlicensed driving in other

jurisdictions, this would require further research to confirm.


The final limitation associated with using crash data is that the information

provided might not be representative of the general population of drivers not involved in

crashes. For example, it is possible that the higher levels of risk-taking evident among the

unlicensed drivers in this study (as reflected in the higher involvement of alcohol,

speeding, motorcycle use, single vehicle crashes etc) may not be indicative of unlicensed

drivers in general, but only a particular sub-group who are more likely to be involved in

crashes (as a consequence of their risk-taking). This highlights the need to explore the

factors identified in this study with a more general sample of unlicensed drivers (as

undertaken in Study Two).

4.5.2 Support for study hypotheses

The following section discusses the results of the study in light of the hypotheses

outlined in section 4.2. The first group of hypotheses to be discussed are H1 – H3.

Unlicensed driver crashes will be more likely to involve alcohol, speeding,

inexperience and motorcycle use than those involving licensed drivers (H1)

Unlicensed driver crashes will be more likely to occur at recreational times than

those involving licensed drivers (H2)

Unlicensed drivers will be more likely to be considered at fault by the police for the

crashes in which they are involved compared with licensed drivers (H3)

Strong support was obtained for these three hypotheses. Firstly, compared to the

serious casualty crashes involving licensed drivers, unlicensed driver crashes were more

likely to involve younger males, motorcycles, alcohol/drug impairment, speeding and

inexperience. These results are consistent with previous research (Harrison, 1997; FORS,

1997a; ATSB, undated; Griffin & DelaZerda; 2000) and tend to confirm a link between

unlicensed driving and risk-taking behaviour. The results are also consistent with the

over-representation of unlicensed drivers in single vehicle crashes and ones where no

traffic control was present. These types of crashes are typically associated with loss of

control and running off the road, which is consistent with speeding, driver impairment

and inexperience.

Secondly, the serious casualty crashes involving unlicensed drive rs were more

likely to occur at night and on the weekend – times generally associated with recreational

driving. At first glance, this finding is somewhat at odds with other evidence suggesting

that a lot of unlicensed driving occurs for work-related reasons (see section 5.4.3).

However, it may only indicate that the driving unlicensed drivers undertake at


recreational times tends to be more risky than work-related driving (which traditionally

occurs during daylight hours through the week). This interpretation is consistent with

other evidence confirming a link between risk-taking and driving during recreational

hours. In Queensland, a higher proportion of fatal single vehicle crashes occur after dark,

while alcohol-related crashes are more likely to occur at night-time and on weekends

(Queensland Transport, 1998). In Victoria, Harrison (1997, p.110) found that the crashes

involving disqualified drivers suggested: “a pattern focused on recreational road use and

drink driving”.

Finally, unlicensed drivers invo lved in crashes in Queensland are more likely to be

considered at fault by the police than licensed drivers, at all levels of crash severity. It is

possible that this may in part reflect a negative halo effect (DeYoung et al, 1997),

whereby the police are more likely to find someone at fault who has already broken the

law. However, the greater attribution of fault to unlicensed drivers is consistent with the

higher level of risk-taking evident in their driving.

Based on the quasi-induced exposure method, unlicensed drivers will be at a higher

risk of being involved in a road crash than licensed drivers (H4)

This hypothesis was supported by the data from this study. The quasi- induced

exposure method indicated that unlicensed drivers were two to three times more likely to

be involved in crashes of various severities than licensed drivers. Overall, the risk of an

unlicensed driver being involved in a crash of any severity was almost three times [2.9:1]

higher than that for licensed drivers.

However, care needs to be exercised when comparing the crash ratios obtained in

this study with those obtained by DeYoung et al (1997) in California. Only fatal crashes

were examined in that study and a different definition was used for unlicensed driving.16

Consequently, the only groups that appear directly comparable are the suspended/revoked

group from the DeYoung et al (1997) study and the disqualified/suspended drivers in the

current study. Given this, it is interesting to note that the fatal crash ratio obtained for the

suspended/revoked drivers [3.7:1] in the Californian study was quite similar to the total

crash ratio calculated for the disqualified/suspended drivers in this study [3.84:1].

The quasi- induced exposure method appears to offer a number of strengths. Firstly,

it overcomes the difficulties involved in directly measuring the exposure of unlicensed

drivers. Secondly, by focusing on multi-vehicle crashes, the method would likely reduce

16. In the DeYoung et al (1997) study the revoked/suspended drivers were treated as a different group to

the unlicensed drivers. Although not clearly defined, the unlicensed drivers appeared to represent those drivers who did not currently have a licence or never had a licence.


the level of under-reporting of crashes among unlicensed drivers. In other words, it would

generally be more difficult for an unlicensed driver to conceal their involvement in a

multi-vehicle crash (as opposed to a single vehicle crash). Furthermore, it is arguable that

any remaining bias associated with under-reporting should apply to all types of crashes,

not just those in which the driver believes they may be found at fault. Hence, the bias

should affect both the denominator and numerator in the calculation of the involvement

rate, a key component of the quasi- induced exposure method.

Despite these strengths and the similarity of the Queensland and Californian results,

some potential problems exist with the quasi- induced method. Firstly, the method is open

to the bias associated with a negative halo effect (DeYoung et al, 1997). If the police are

more likely to find unlicensed drivers at fault for reasons other than their immediate

driving behaviour, it would tend to inflate their involvement rate and, hence, their crash

ratio when compared to licensed drivers.

On the other hand, the use of multi-vehicle crashes appears to introduce another

potential bias that serves to deflate the crash ratio for unlicensed drivers. Compared with

licensed drivers, unlicensed drivers are under-represented in multi-vehicle crashes. In

addition, they are more likely to be considered at-fault in the single vehicle crashes in

which they are involved. Consequently, an involvement rate based on multi-vehicle

crashes would likely underestimate the full extent of the at- fault driving undertaken by

unlicensed drivers. This suggests that the crash ratio obtained for unlicensed drivers may

actually underestimate their crash risk.

Finally, while the overall involvement rate for unlicensed drivers was reasonably

stable over time, there were some fluctuations for some of the subgroups, particularly the

never licensed drivers. This suggests that some of the factors influencing the

determination of at- fault driving may not be stable over time, at least for the smaller

groups of unlicensed drivers.

In summary, although the quasi- induced exposure method “has its limitations . . .

it is perhaps the best method we have now for estimating overinvolvement that corrects

for exposure, especially for unlicensed drivers” (Scopatz et al, 2003, p.17). Nonetheless,

the problems inherent in the approach suggest that the results obtained through its use

should be treated with some caution. Over and above this, there is a need to develop

better methods of estimating the exposure of unlicensed drivers, such as implementing

periodic roadside stopping surveys (see section 4.5.5). This would enable the crash risk of

unlicensed drivers to be estimated through more direct methods. This would also provide


a benchmark with which to assess the validity and reliability of the quasi- induced

exposure method.

The crashes involving unlicensed drivers will result in significantly higher levels of

death and serious injury than those involving licensed drivers (H5)

Strong support was obtained for this hypothesis. In the event of a crash, unlicensed

drivers were more than twice [2.12] as likely than the licensed drivers to be involved in a

serious casualty crash (ie. one resulting in a fatality or serious injury) compared with a

minor crash. However, it appears that some of this increased risk is partly due to the age

and gender of the people who tend to drive unlicensed and the types of vehicles they

drive, rather than unlicensed driving per se. In particular, a logistic regression analysis

indicated that riding a motorcycle was associated with a four and half times [4.58] the

risk of being involved in a serious casualty crash. Nonetheless, even after accounting for

the influence of age, gender and vehicle type, being unlicensed was associated with

almost twice [1.82] the risk of the crash resulting in a serious casualty.

It is possible that the over-representation of unlicensed drivers in serious crashes is

partly a produce of the under-reporting of minor crashes by these drivers. However, the

results are consistent with the higher involvement of risky behaviours in the crashes

involving unlicensed drivers. In particular, the crashes involving speeding and

alcohol/drug impairment tend to be more severe than many other types of crashes. For

example, in Queensland during the study period, alcohol/drugs and speeding were

implicated in 34% and 12% of fatal crashes respectively, but only 9% and 4% of total

crashes. Similarly, single vehicle crashes (in which unlicensed drivers are over-

represented) tend to be more severe than multi-vehicle crashes (Queensland Transport,

1999).

Significant differences will be found between the unlicensed driver sub-groups

involved in crashes, in terms of age and gender (H6)

The data from this study confirmed that there are significant gender and age

differences between unlicensed and licensed drivers. Compared with the licensed drivers,

unlicensed drivers were more likely to be male and younger in age. This evidence is

consistent with the findings of a number of previous studies (Job et al, 1994; FORS;

1997a; Griffin & DelaZerda; 2000).

In addition, significant gender and age differences were found among the different

sub-groups of unlicensed drivers. Males were relatively over-represented in the


inappropriate class of licence and disqualified/suspended driver groups, but under-

represented in the never licensed and expired groups. As would be expected, the under 17

year old drivers were over-represented among the never licensed drivers but under-

represented among the other types of offenders. There was also higher representation of 25-

59 year old drivers among the disqualified/suspended and expired licence holders. This

probably reflects the longer time periods that these licences would have been held.

Significant differences will be found between the unlicensed driver sub-groups in

terms of the circumstances of the crashes in which they are involved, the associated

contributing factors and the severity of the crashes (H7)

Strong support was obtained for this hypothesis, with a range of significant

differences emerging among the unlicensed driver types. A summary of the key

differences found in the serious casualty crashes involving unlicensed drivers is provided

below.

Differences in crash circumstances

§ The drivers with an inappropriate class of licence were more likely to be riding

motorcycles and be involved in multi-vehicle crashes.

§ The never licensed drivers were the group most likely to be involved in single

vehicle crashes and where no traffic control was present.

§ The disqualified/suspended drivers were the most likely to be involved in a crash at

night time and one of the groups most likely to be driving a car.

Differences in contributing factors

§ The disqualified/suspended drivers were significantly more likely to be involved in

a crash involving alcohol/drugs, while the inappropriate class of licence were the

least likely.

§ Inexperience was more likely to be cited as a contributing factor in the crashes

involving the never licensed drivers.

§ The never licensed drivers were the most likely to be considered at fault for the

crashes in which they were involved, followed by the disqualified/suspended

drivers.

Differences in the severity of crashes

§ In the event of a crash, those involving drivers with an inappropriate class of

licence were the most likely to result in a serious casualty. This appears to reflect


the high proportion of motorcyclists in this group who are particularly vulnerable to

injury in road crashes.

§ After accounting for age, gender and type of vehicle driven, the never licensed and

disqualified/suspended drivers were at the highest risk of being involved in a

serious casualty crash. Both groups were at approximately twice the risk compared

to the other drivers, with odds ratios of 2.01 and 1.96, respectively.

Together, these results suggest that there are significant differences in the driving

behaviour of the different types of unlicensed drivers. In particular, the never licensed

and disqualified/suspended drivers emerge as the groups most likely to engage in risk

taking behaviour and be involved in severe crashes. This pattern of results is also

consistent with the crash ratios derived from the quasi- induced exposure method. The

never licensed and disqualified/suspended drivers had the highest estimated crash risk for

most of the different crash severities and for total crashes. Indeed, the risk of involvement

in a crash was 5.43:1 for the never licensed drivers and 3.84:1 for the

disqualified/suspended drivers, compared with 2.9:1 for all unlicensed drivers. As noted

in section 4.4.5.1, the crash ratios for the never licensed drivers should be treated with

some caution due to the instability in their involvement rates. Nonetheless, the results

obtained for this group using the quasi- induced exposure method are consistent with the

other findings.

4.5.3 Implications for theory

The results from this study have some important implications for the various

theoretical perspectives outlined in Chapter 3. Firstly, the greater involvement of

alcohol/drugs, speeding and inexperience in the crashes involving unlicensed drivers

suggests that many of them may have a strong propensity for sensation seeking. In

addition, the role of alcohol/drugs suggests that alcohol misuse may exert a strong

influence on their driving behaviour. While these characteristics may be more common

among unlicensed drivers in general, it appears that they influence some offenders more

than others. For example, the involvement of alcohol/drugs was particularly high among

the disqualified/suspended drivers.

The wide range of differences among offenders also has some important

implications for theory. It suggests that unlicensed drivers do not represent a

homogeneous group and, hence, there may be important differences in their motives for

driving without a valid licence. In particular, two problem groups emerged in this study:

the never licensed and the disqualified/suspended drivers. Both of these groups appear the


most likely to be involved in crashes as a result of risk-taking or inexperience. This is

reflected in both a higher risk of being involved in a crash and the greater likelihood that

these crashes will result in a fatality or serious injury.

From a criminological perspective, the results suggest that the never licensed and

the disqualified/suspended drivers may represent a more deviant sub-group of offenders.

The behaviour of these drivers tends to represent a more flagrant breaking of the road

rules, since they have decided to drive either without a licence or in contravention of a

driving ban. In contrast, it is arguable that the offences committed by the drivers with an

expired or inappropriate class of licence are more administrative in nature.

Accordingly, a robust theoretical explanation of unlicensed driving will need to

account for the behaviour among a wide range of offenders. In particular, it will need to

account for a wide range of potential motives for the behaviour, some of which may be

more deviant than others.

4.5.4 Implications for road safety

The findings of this study have important implications for road safety and the

development of more effective approaches to counter unlicensed driving. Firstly, the

results confirm that unlicensed driving represents a relatively small, but significant road

safety problem. Unlicensed drivers represent over 6% of the drivers involved in fatal

crashes and 5% of those in serious injury crashes. Moreover, based on the quasi- induced

exposure method, unlicensed drivers appear to be two to three times more likely to be

involved in crashes of varying severity than licensed drivers. In the event of a crash, those

involving unlicensed drivers are twice as likely to result in a fatality or serious injury.

In addition, there was little firm evidence that unlicensed drivers restrict their

driving in order to avoid detection. It is possible that the higher involvement of

unlicensed drivers in night-time crashes is, in part, a reflection of greater exposure at this

time. In other words, unlicensed drivers may prefer to drive at night in the belief that they

are less likely to be detected than they would be during the day (although little support

for this was found in Study Two – see section 5.4.3.4). Moreover, there were no

significant differences between unlicensed and licensed drivers in terms of the speed

zones in which their crashes occurred. This suggests that both groups are equally exposed

across the road network.


Accordingly, unlicensed drivers represent a problem group requiring more

attention. Even if they do drive somewhat more cautiously to avoid detection, they appear

to pose more of a road safety risk than licensed drivers. This has two clear implications

for road safety. Firstly, it indicates that more effective approaches are required to reduce

the level of unlicensed driving. Secondly, there is a need to review policies that may be

inadvertently exacerbating the problem.

Finally, the findings of this study suggest that unlicensed drivers do not represent a

homogeneous group. There are important differences in the characteristics and behaviour

of different types of offenders. This suggests that countermeasures in this area may need

to be multi-strategy in nature. This issue is further explored in the next study.

4.5.5 Future directions for research

This study has highlighted a number of important issues requiring further research.

Most importantly, it remains unclear whether the characteristics and on-road behaviour of

unlicensed drivers identified in this study are only indicative of those involved in crashes,

or are representative of unlicensed drivers in general. This has important implications for

the scope of the countermeasures required to address the problem. Accordingly, there is a

need to investigate these issues with a more general sample of unlicensed drivers. This is

the main focus of Study Two.

This study has also highlighted a number of issues requiring further attention that

are beyond the scope of the current research program. Firstly, while the quasi- induced

exposure method offers certain advantages and warrants replication in other jurisdictions,

there is a need to develop better methods of estimating the exposure of unlicensed

drivers. This is required to better estimate the crash risk associated with unlicensed

driving and to act as a benchmark for evaluating the effectiveness of future

countermeasures. To this end, there is a need to evaluate the cost-effectiveness of

different methodologies such as periodic roadside stopping surveys, the sampling of

driver’s licences at RBT and the surveillance of unlicensed drivers (see sections 2.2.1 and

2.2.2). Secondly, there is a need for further research into the issue of underage driving.

Almost 13% of the unlicensed drivers involved in serious casualty crashes were under the

age of 17. Research with this group will require the use of special age-appropriate

methodologies, due to these drivers being technically minors. These issues are further

explored in section 7.6.


4.6 Chapter summary

This study was designed to explore the first three research questions identified as

part of this program of research. In doing so, it has both confirmed and extended the

available evidence relating to the crash involvement of unlicensed drivers. The results

indicate that unlicensed driving is a relatively small, but significant road safety problem.

Unlicensed drivers represent over 6% of the drivers involved in fatal crashes and 5% of

those in serious injury crashes. Based on a quasi- induced exposure method, unlicensed

drivers were found to be almost three times as likely to be involved in a crash than

licensed drivers. In the event of a crash, those involving unlicensed drivers were twice as

likely to result in a fatality or serious injury. Consistent with these results, the serious

crashes involving unlicensed drivers were more likely to feature risky driving behaviours,

such as drink driving, speeding and motorcycle use, than those involving licensed drivers.

However, it remains unclear whether the behaviour of unlicensed drivers involved

in crashes is representative of these drivers in general. In other words, it remains possible

that crash- involved unlicensed drivers represent a special sub-set of offenders who are less

concerned about the risks of detection and punishment. Furthermore, future research needs

to better distinguish between different types of offenders. In particular, there is a need to

explore potential differences between the disqualified and suspended drivers (which was

not possible in the crash data) and to avoid a large grouping of other unlicensed drivers.

Therefore, while this study has provided some important insights into unlicensed

driving behaviour, there is a need to further explore the research questions with a more

general sample of unlicensed drivers. This is required to establish how representative the

findings from this study are, and to examine certain issues not possible with crash-

involved drivers. In particular, there is a need to more specifically explore the

psychosocial characteristics of offenders and their perceptions toward current

enforcement and punishment processes. These issues form the main focus of Study Two,

reported in the next chapter.



Chapter Five: The self-reported behaviour of unlicensed drivers

5.1 Introductory comments............................................................................ 109

5.2 Study aims and hypotheses...................................................................... 109

5.3 Method..................................................................................................... 112

5.3.1 General research strategy................................................................. 112

5.3.2 Exploratory research........................................................................ 113

5.3.2.1 Semi-structured qualitative interviews .................................... 113

5.3.2.2 Pilot quantitative interviews .................................................... 114

5.3.3 Main study ....................................................................................... 115

5.3.3.1 Selection of survey location..................................................... 115

5.3.3.2 Participants............................................................................... 116

5.3.3.3 Materials ................................................................................... 116

5.3.3.4 Procedure ................................................................................. 118

5.3.3.5 Statistical analyses ................................................................... 119

5.4 Results ..................................................................................................... 120

5.4.1 Sample characteristics ..................................................................... 120

5.4.1.1 Response rate ........................................................................... 120

5.4.1.2 Type of offender....................................................................... 122

5.4.1.3 Socio-demographic characteristics of the sample .................... 123

5.4.1.4 Driving and criminal history of offenders ............................... 125

5.4.1.5 Driving for work purposes ....................................................... 127

5.4.1.4 Sensation seeking..................................................................... 127

5.4.1.5 Alcohol misuse......................................................................... 128

5.4.2 Circumstances of detection and outcome of court hearing.............. 130

5.4.2.1 Reason stopped by police......................................................... 130

5.4.2.2 Reason for driving when detected............................................ 132

5.4.2.3 Vehicle driven when detected .................................................. 132

5.4.2.4 Outcome of court hearing ........................................................ 134

5.4.3 Unlicensed driving behaviour .......................................................... 135

5.4.3.1 Length of time driving unlicensed ........................................... 135

5.4.3.2 Frequency of unlicensed driving................................................ 136


5.4.3.3 On-road driving behaviour ........................................................ 138

5.4.4 The impact of current administrative, enforcement and punishment processes ........................................................................ 144

5.4.4.1 Awareness of being unlicensed at time of detection ................. 144

5.4.4.2 Possession of a photographic licence ........................................ 144

5.4.4.3 Unlicensed driving after detection............................................. 145

5.4.4.4 Evasion of detection .................................................................. 146

5.4.4.5 Perceptions of enforcement and punishment processes............. 148

5.4.4.6 Intention to drive unlicensed in the future ................................. 150

5.5 Discussion ................................................................................................. 151

5.5.1 Study limitations ................................................................................ 151

5.5.2 Support for study hypotheses ............................................................ 152

5.5.3 Implications for theory...................................................................... 156

5.5.4 Implications for road safety............................................................... 157

5.5.4.1 The extent and nature of unlicensed driving.............................. 157

5.5.4.2 The effectiveness of current administrative, enforcement and punishment processes......................................................... 158

5.5.5 Future directions for research............................................................ 160

5.6 Chapter summary ...................................................................................... 161



This chapter will document the second study undertaken as part of this program of

research. The aims of the study were two-fold. The first aim was to confirm and extend

the findings obtained in Study One with a more general sample of unlicensed drivers.

This involved further examination of research questions one and three, relating to the

risk-taking behaviour of unlicensed drivers and the degree of homogeneity among

offenders. In particular, the study featured a strong focus on the psychosocial

characteristics of unlicensed drivers and their self- reported driving behaviour. The second

aim of the study was to investigate research question four: How effective are current

administrative, enforcement and punishment policies and processes in preventing

unlicensed driving?

The study involved a cross-sectional survey of unlicensed driving offenders

recruited at the Brisbane Central Magistrates Court. This recruitment strategy was

adopted to increase the likely response rate to the survey, a major shortcoming of

previous research in the area. Care was taken to ensure the appropriate identification of

different types of offenders, to overcome the some of the problems experienced in Study

One. The survey was designed to examine the offenders’:

§ psychosocial characteristics;

§ driving behaviour while unlicensed and the circumstances of their detection;

§ exposure to police traffic law enforcement activities;

§ experiences of punishment and punishment avoidance; and

§ perceptions toward enforcement and punishment processes.

The study hypotheses were informed by both the findings of Study One and the

theoretical perspectives discussed in Chapter 3. They are discussed in more detail below.


As noted above, the main aim of this study was to examine three particular research

questions identified in section 2.7, utilising a cross-sectional design. The research

questions examined and the related hypotheses are discussed below.


H8 Unlicensed drivers will report more frequent drink driving, speeding and

failure to wear a seat belt than general drivers.

H9 The self-reported drink driving behaviour of unlicensed drivers will be

positively associated with alcohol misuse and sensation seeking, while self-


reported speeding and non-seat belt use will be positively associated with

sensation seeking.

The first of these hypotheses builds on the findings of Study One, which

indicated that unlicensed drivers were over-represented in crashes involving alcohol

and speeding compared with licensed drivers. Although seat belt wearing was not

specifically examined in that study, there is other evidence that suggests that

unlicensed drivers are less likely to wear a seat belt (FORS, 1997b). The second

hypothesis was informed by the review of theoretical perspectives undertaken in

Chapter 3. In particular, it was hypothesised that more extensive drink driving

would be associated with higher levels of alcohol misuse and sensation seeking. In

addition, it was hypothesised that other risky behaviours, such as speeding and

failure to wear a seat belt, would be positively associated with higher levels of

sensation seeking among the offenders.



H10 Significant differences will be found between the unlicensed driver types in

terms of their socio-demographic characteristics.

H11 The disqualified, suspended and never licensed drivers will report higher

levels of sensation seeking than the other offenders.


levels of alcohol misuse than the other offenders.


levels of drink-driving, speeding and failure to wear a seat belt than the

other offenders.

Once again, these four hypotheses are informed by the findings of Study One

and the theoretical perspectives reviewed in Chapter 3. It is expected that the age

and gender differences found among crash-involved unlicensed drivers will be

reflected in other socio-demographic characteristics of the offenders. It is also

expected that the higher estimated crash risk for the disqualified, suspended and

never licensed drivers (found in Study One) will be reflected in the on-road

behaviour of these offenders, and be linked to the theoretical constructs of sensation

seeking and alcohol misuse. In other words, these sub-groups of offenders will be

more likely to report higher levels of sensation seeking and alcohol misuse and to


engage in risk-taking behaviours like drink driving, speeding and failure to wear a

seat belt.

4. How effective are current administrative, enforcement and punishment processes in

preventing unlicensed driving?

H14 The perceived risk of apprehension for unlicensed driving will be lower than

that for other illegal road user behaviours, such as drink-driving and

speeding.

H15 Instances of punishment avoidance will be common among unlicensed

driving offenders.

H16 Continued driving after detection will be common among unlicensed driving

offenders.

This group of hypotheses are designed to examine the effectiveness of

current countermeasures to unlicensed driving, from the perspective of deterrence

theory. As discussed in section 3.2.1, the central tenet of classical deterrence theory

is that the effectiveness of legal sanctions will be determined by the perceived

severity, certainty and swiftness of punishment. A key determinant of these

perceptions is the perceived risk of apprehension, which in the case of unlicensed

driving appears to be relatively low in many jurisdictions, including Queensland

(see section 3.2.1.2). To examine the tenability of this view, it is hypothesised that

the perceived risk of apprehension for unlicensed driving will be significantly lower

than it is for other illegal behaviours that are more are intensively enforced (than

unlicensed driving), such as drink driving and speeding. The second hypothesis,

relating to the role of punishment avoidance, is based on the historical difficulties

associated with detecting unlicensed driving (see section 2.6.2) and the theoretical

work of Stafford and Warr (1993) (see section 3.2.3). Stafford and Warr have

suggested that the experience of punishment avoidance can serve to encourage

illegal behaviour, particularly in cases where detection and punishment are

relatively rare. To explore the relevance of this theoretical construct, it is

hypothesised that many unlicensed driving offenders will experience instances of

punishment avoidance. The final hypothesis acknowledges the apparent difficulties

involved in deterring unlicensed driving using current enforcement and punishment

processes.


5.3 Method 5.3.1 General research strategy

A number of methodological problems, common to research dealing with illegal or

deviant behaviours, are encountered when surveying unlicensed drivers. Firstly, for a

number of reasons it is difficult to obtain a random, representative sample of these drivers.

The use of population-based survey methods do not usually represent a cost-effective

option, since a large number of people would need to be approached in order to obtain

sufficient participants who admit to driving unlicensed. (In addition, it would have been

difficult to ensure an appropriate representation of the different types of unlicensed

drivers.) To overcome this problem, some studies have used snowball-sampling procedures

to recruit unlicensed drivers within the general community (eg. Williamson, 1996). While

this technique is relatively cost-effective, the representativeness of such samples remains

unclear. Consequently, most studies in the area have relied on official records to target

drivers who have either been disqualified from driving or detected driving unlicensed.

However, the representativeness of these samples also remains unclear. As already noted, a

number of studies that have used official records to recruit subjects have found that many

no longer reside at the address provided (Robinson, 1977; Mirrlees-Black, 1993; Job et al,

1994; Davies et al, 1999). In addition, surveys in this area generally feature high refusal

rates – which is not surprising given the illegal nature of the behaviour. Together, these

factors have contributed to the low response rates of self-report surveys using both

interview (Robinson, 1977; Mirrlees-Black, 1993) and mail questionnaire (Robinson,

1977; Smith & Maisey, 1990; Job et al, 1994) methods.

A second problem encountered when surveying people about illegal behaviours, such

as unlicensed driving, relates to the quality of self-report data. Generally it is difficult to

corroborate the accuracy of this data, since there are few objective measures of the

behaviour available. This highlights the need to minimise any systematic bias produced by

low response rates and to utilise a survey method that encourages more truthful responses

from the subjects.

In order to obtain a reasonably representative sample of unlicensed drivers, it was

decided to utilise the court system to directly access offenders. In Queensland, there are

two types of offences relating to unlicensed driving: Disqualified Driving and Unlicensed

Driving.


At the time the survey was conducted, both of these offender types were required to

attend court (Travelsafe, 1998)17. In effect, this acted as a filter or bottleneck through which

all alleged offenders had to pass – representing an ideal time to recruit subjects. Ideally, it

provided a means of accessing all the different types of unlicensed drivers. In addition, a

recent survey of drink drivers conducted in central Queensland courts achieved a

reasonably high response rate (61%) (Ferguson et al, 2000).

Consequently, approval was obtained from the Registrar of the Brisbane Central

Magistrates Court to survey unlicensed driving offenders either before or after their court

hearing. (The rationale for selecting this location is explained in section 5.3.3.1.) This

approval was based on a number of provisos, including that the survey did not interfere

with Court business or case flow management, and that the interviewers did not provide

any advice to offenders. Approval for the study was also obtained from QUT’s Human

Research Ethics Committee.

The likely representativeness of the sample obtained through this method is discussed

in section 5.5.1. However, it is important to acknowledge at the outset one major constraint

of the sample. By focusing on offenders, the sample was limited to those unlicensed drivers

who had been detected by the police. In this regard, it is possible that unlicensed drivers

who remain undetected are generally more cautious (and possibly safer) than those

caught by the police. However, as will be documented in the results section, many of the

offenders in the sample were detected through random enforcement processes and had

been driving unlicensed for quite long periods of time before they were detected.

5.3.2 Exploratory research

A two-phase preliminary study was conducted to explore the perceptions, attitudes

and self- reported behaviour of unlicensed driving offenders. In both phases, eligible

offenders were offered $25 to participate in the study.

5.3.2.1 Semi-structured qualitative interviews

A semi-structured interview questionnaire was developed to explore a range of issues

related to unlicensed driving and the factors contributing to the behaviour. This qualitative

phase was designed to explore the social and legal factors that either encourage or

discourage the behaviour, and to inform the concepts and language to be used in the

17. In Queensland, a person is charged with Disqualified Driving if they drive contrary to a court-ordered

licence disqualification. Other unlicensed drivers are charged with Unlicensed Driving . In November 2002 (after the end of the data collection period), a Traffic Offence Notice (TON) was introduced for first-time Unlicensed Driving offenders, who were otherwise eligible to hold a licence. As a result, these offenders were no longer required to attend court.


subsequent quantitative pilot. The results confirmed the viability of the survey procedure

with a total of 15 respondents agreeing to participate from the 23 eligible offenders

approached (ie. 65% response rate). It also highlighted some driving behaviours associated

with unlicensed driving which had not previously been reported in the literature. For

example, some participants who had been caught driving on an expired or cancelled

Learner’s Licence reported that they had purposefully refrained from obtaining a

Provisional Licence (the first licence which permits solo driving). The reason cited for this

was that the current penalty for Unaccompanied Driving on a Learners Licence ($30) was

lower than the costs associated with obtaining a Provisional Licence. A summary of the

findings from this phase is included as Appendix B. (The results are presented in a similar

format to the questionnaire that was utilised.)

5.3.2.2 Pilot quantitative interviews

Based on the findings of the qualitative phase and the available literature, a

questionnaire was developed for piloting. The two aims of this phase were to identify

potential problems with the instrument and to compare the relative efficacy of an interview

versus a self-administered version of the questionnaire. A total of 19 respondents

participated in this phase from the 28 eligible offenders approached (ie. 68% response rate).

Some basic information was collected about all the eligible offenders approached

(irrespective of whether they agreed to participate or not) including their gender, the

offence with which they were charged and, where applicable, the reason cited for non-

participation (see Appendix C for a copy of the proforma used for this purpose). The main

findings of the quantitative pilot are discussed below.

§ The majority of the eligible offenders approached were male (86%). However, there

did not appear to be any major difference in the response rate among males and

females, with 71% of males agreeing to participate compared with 67% of females.

§ The majority of the eligible offenders approached were charged with Unlicensed

Driving (82%). The unlicensed drivers also appeared more likely to agree to

participate (70%) than the disqualified drivers (60%).

§ The most common reason cited by the 9 offenders who refused to participate was that

they were in a rush to go elsewhere (n=7). In particular, three participants specifically

mentioned that they needed to go to work. This raises the possibility that offenders

who were employed may have been less likely to agree to participate in the survey,

due to either time constraints or the fact that the incentive payment appeared less

attractive to them.


§ The self-administered version of the questionnaire had much higher levels of missing

data than the interview-administered version. Hence, while more costly, it appeared

that better quality information would be obtained by adopting an interview-based

approach for the main study.

§ It became apparent that a wider range of questions was required to assess the on-road

driving behaviour of the participants. While many reported modifying their behaviour

when driving unlicensed, there appeared to be strong differences among participants

in terms of the caution they exercised. Accordingly, it was decided to include some

questions relating to compliance with key road safety rules (namely speeding,

restraint use and drink driving) for the main survey.

§ The Sensation Seeking Scale items in the pilot questionnaire were measured on a

Likert scale, as suggested by Rimmo and Aberg (1999) in their research on drivers.

However, it became apparent that some participants found the rating scale

ambiguous. Consequently, it was decided to utilise the fixed-choice format (as

originally proposed by Zuckerman et al, 1978) in the main survey questionnaire.

5.3.3 Main survey

5.3.3.1 Selection of survey location

The main survey was conducted by face-to-face interviews in the Brisbane

Magistrates Court between June 2001 and April 2002. This location was selected for two

reasons. Firstly, at the Brisbane Court specific times are allocated for traffic-related

matters, including drink driving, speeding, disqua lified driving and unlicensed driving

charges. These sessions are typically held in the same courtroom each morning. This

practice made it possible to allocate interviewers to the court sessions when they were

most likely to encounter unlicensed and disqua lified drivers. Secondly, the Brisbane

Court processes more traffic offenders each year than any other court in Queensland,

accounting for over 10% of all unlicensed and disqualified driving offenders (Micola,

2002). Consequently, concentrating on this court represented the most cost-effective

method of obtaining a sample that was both reasonably large and representative of

offenders detected in a metropolitan setting. Nonetheless, it should be acknowledged that

the court primarily processes offenders who are detected in the inner city and suburban

areas of Brisbane.

In addition, it was originally intended to interview a sample of offenders in a non-

metropolitan setting. However, difficulties arose in identifying suitable courts that

processed sufficient numbers of unlicensed drivers in a concentrated manner. An attempt


was made to recruit participants at the Townsville Magistrate’s Court, since it conducts

specific traffic-related court sessions (albeit on a fortnightly basis only). In practice, this

did not prove an effective research strategy, with only 13 participants being recruited at

this court during the study period. Due to the small number of these participants, they

were excluded from the main study.

5.3.3.2 Participants

The sample for the main study consisted of 309 offenders. Information relating to the

survey response rate and the characteristics of the offenders who agreed to participate

compared with those who declined is provided in section 5.4.1. Offenders who participated

in the two exploratory studies were excluded from the main study.

5.3.3.3 Materials

The questionnaire and related information sheet used in the study are reproduced as

Appendix D. The questionnaire drew on a variety of items and scales from other studies,

as well as ones specifically developed for this study. A summary of the key issues and

related items examined in the study is provided below. Seven-point scales were generally

used, based on the findings of studies examining optimal scale divisions (eg. Diefenbach,

Weinstein & O’Reilly, 1993). It should be noted that there were other items included in

the questionnaire for the purposes of Study Three. These items are described in section

6.3.2.

Socio-demographic characteristics

The socio-demographic information collected from the participants included:

gender; age; marital status; level of education attained; employment status/occupation;

income; and driving experience (Q. 1–7). Items were also included to explore the

previous involvement of the participants in traffic offences (Q. 21-22) and criminal

offences (Q. 61).

Sensation seeking

Due to the prohibitive length of the full 40- item Sensation Seeking Scale (SSS),

only the Thrill and Adventure Seeking (TAS) subscale (10 items) was used (see Q. 54).

Previous driving-related studies that have used the SSS suggest that the TAS subscale has

the strongest relationship to risky driving (Jonah, 1997). The TAS subscale was drawn

from Form V of the SSS, since this is the version most widely used to date in driving

behaviour studies (Jonah, 1997; Rimmo & Aberg, 1999). This form of the scale uses a

fixed-choice format (Zuckerman et al, 1978).


Alcohol use

To explore the degree to which unlicensed driving may be associated with alcohol

misuse, the questionnaire included the Alcohol Use Disorders Identification Test

(AUDIT) (Saunders, Aasland, Babor, de la Fuente & Grant, 1993). This AUDIT is a brief

screening instrument developed for the World Health Organisation (WHO) for the early

detection of hazardous and harmful alcohol consumption (which is consistent with the

perspective adopted to alcohol misuse outlined in section 3.5). It has been used in a

variety of clinical and community settings and has been reported to be more accurate than

traditional screening questionnaires (Degenhardt, Conigrave, Wutzke & Saunders, 2001).

The 10 items making up the AUDIT are reproduced as Questions 57–59 in the study

questionnaire.

Circumstances/process of detection and punishment

A range of questions were included relating to the offence the participants were

charged with, the circumstances of their detection and the outcomes of the court hearing

(Q. 8–20).

Unlicensed driving behaviour

A number of aspects of the participants’ driving behaviour were examined,

including:

§ The amount of unlicensed driving undertaken by the offenders during their lifetime,

since their previous conviction (if applicable) and since their detection by the police

(Q. 23–28). The categories used in this section were modelled on the Job et al

(1994) survey of unlicensed drivers;

§ The degree of access to motor vehicles and whether they gave lifts to other people

(Q. 30–31);

§ Compliance with speeding, seat belt and drink driving laws (Q. 37–39). These items

were adapted from the annual survey of community attitudes to road safety

conducted by the Australian Transport Safety Bureau (ATSB, 2000);

§ Compliance with a range of road rules, modelled on Job et al (1994) (Q. 41);

§ Exposure to police enforcement and licence checking (Q. 44–48); and

§ The participants’ intention to drive unlicensed in the future (Q. 60).

Availability of a motor vehicle or alternative transport

A number of items were included to examine the influence of transport availability

on unlicensed driving. The particular issues addressed were:


§ access to a motor vehicle (Q. 30);

§ ownership of a motor vehicle (Q. 14 and 30);

§ possession of an old (invalid) photographic licence (Q. 32); and

§ access to public transport or other alternative transport (items in Q. 53).

Perceptions relating to enforcement and punishment processes

A number of items were included to examine the knowledge and perceptions of

participants toward current enforcement and punishment processes. These included:

§ perceived risk of apprehension for unlicensed driving compared with a selection of

other illegal behaviours (Q. 33);

§ knowledge of the fines for unlicensed/disqualified driving (Q. 15); and

§ perceived severity, certainty and swiftness of punishment for

unlicensed/disqualified driving (Q. 52a, 52i and 52o).

5.3.3.4 Procedure

In addition to the author, four people with psychology/social science degrees were

employed to interview the participants. Prior to commencing the survey, the interviewers

were given a training session with the questionnaire and provided with an Interviewer’s

Guide that specified the procedure to be followed (see Appendix E).

The interviewers approached people as they left the courts and explained that they

were conducting an anonymous, voluntary survey on the topic of unlicensed driving.

Only people who were charged with Disqualified Driving or Unlicensed Driving, and

appeared to understand English, were invited to participate in the study. The offence

category was primarily determined from information presented in the court hearing and/or

published on notice boards at the Court. Once potential participants were identified, they

were given a brief explanation of the survey and offered $25 to participate in the study.

As with the pilot studies, some basic information was collected from all offenders

approached to participate in the survey, in order to characterise the non-respondents.

During the course of the main study it became apparent that the presence of others (eg.

family or friends) appeared to increase the likelihood of an offender refusing to

participate in the study. Consequently, the interviewers were requested to record whether

the offenders were accompanied or not. Because this procedure was introduced during the

course of the study, the relevant data was only collected for the later offenders

approached.


Under normal circumstances the interview took approximately 25 minutes to

complete. However, the approval to conduct the survey had been granted on the proviso

that the interviews did not interfere with the normal workings of the court or case flow

management. Consequently, on some occasions the interviews were suspended for a

period while the participants attended to court-related business.

5.3.3.5 Statistical analyses

The data collected from the survey was analysed using the Statistical Package for

the Social Sciences (SPSS) Version 10.0.5. The level of missing data was minimal, given

the interview method adopted for the administration of the survey. Consequently, cases

with missing values were generally excluded from the relevant analyses since it had a

minimal impact on the sample size. Unless otherwise specified, the significance level (α)

for the main statistical tests was set at .05. A more stringent significance level

(α = .01) was used for post-hoc comparisons, to protect against inflating the Type 1 error

rate.

The categorical data were analysed using a variety of non-parametric tests. Chi-

square (χ2) tests were used to test for the independence of categorical variables. Where

necessary, post-hoc analyses were undertaken within each variable using the adjusted

standardised residual statistic (ê). As noted in section 4.3.3, this statistic indicates the

relative difference between the observed and expected frequencies for a particular cell,

adjusted for row and column totals, and can be used to identify those cells with observed

frequencies significantly higher or lower than expected. Adjusted standardized residuals

are approximately normally distributed with a mean of 0 and a standard deviation of 1,

and can be interpreted as Z-scores (Haberman, 1978).

One of the problems encountered with some of the Chi-square tests was the small

number of drivers in the sample with an inappropriate class of licence (n=10). Small cell

sizes can create problems by reducing the degree to which the chi-square distribution

provides an accurate approximation of the chi-square statistic. A common rule of thumb

is to ensure that no cell has an expected frequency of less than 1 and that no more than

20% of cells have an expected frequency less than 5 (Cohen, 1996). To avoid this

problem, the drivers with an inappropriate class of licence were sometimes excluded from

the analyses, reducing the sample size to approximately 300 and the degrees of freedom

to 4. However, this would not have overly impacted on the power of the analyses. Aron

and Aron (1999) report that a sample size of 133 is sufficient to detect a medium effect

size (φ =.30) with 4 degrees of freedom at .05 significance while maintaining 80% power.


As in Study One, the strength of association between categorical variables was

measured using either the phi (φ) coefficient (for 2 × 2 tables) or Cramer’s Phi (φc)

coefficient (for tables greater than 2 × 2). (Further information relating to these two

statistics can be found in section 4.3.3.) Other non-parametric methods, such as the

Kruskal-Wallis (H) test, were used to analyse interval data where the assumptions of

normality or homogeneity of variance were sufficiently violated.

Although not strictly interval data, the data collected by Likert scale were analysed

using parametric methods (in cases where there was no normality or homogeneity of

variance problems). The strength of association between continuous and dichotomous

variables was measured using the point-biserial correlation coefficient (rpb). The strength

of association between continuous dependent variables and discrete independent variable

was measured using eta squared (η2). Reliability analyses were undertaken on the scales

used in the study. A summary of the scales and their Cronbach’s alpha is provided in

Appendix F.

5.4 Results

5.4.1 Sample characteristics

5.4.1.1 Response rate

A total of 309 participants agreed to participate in the main survey from 495 eligible

offenders approached, representing a response rate of 62.4%. Table 5.1 compares the

offenders who agreed to participate with those who refused in terms of gender, their

offence, the interviewer who approached them and whether they were accompanied or not.

As can be seen, there was a significant difference between males and females in their

preparedness to participate in the study, with females (74.3%) being more likely to agree

than males (60.5%). There was no significant difference between the participants and

those who refused in terms of their offence. Although there was some variation in the

response rate achieved by the five interviewers used in the study, these differences were

not significant.


Table 5.1

Characteristics of all offenders approached to participate

Agreed or Not Offender Characteristics Agreed to

participate Refused to participate

Significance level

Gender n=309 n=186

Males 257 (60.5%) 168 (39.5%) χ2 (df1) = 4.89, p < .05, Females 52 (74.3%) 18 (25.7%) φ = -.10 Offence type n=309 n=186

Unlicensed driving 257 (62.4%) 155 (37.6) χ2 (df1) = 0.00, p > .05 Disqualified driving 52 (62.7%) 31 (37.3) φ = .00 Interviewer n=309 n=186

One 39 (52.0%) 36 (48.0%) χ2 (df4) = 5.67, p > .05 Two 32 (72.7%) 12 (27.3%) φc = .11 Three 68 (63.6%) 39 (36.4%) Four 63 (64.3%) 35 (35.7%) Five 107 (62.6%) 64 (37.4%) Accompanied n=88 n=39

Yes 31 (58.5%) 22 (41.5%) χ2 (df4) = 4.99, p < .05 No 57 (77.0%) 17 (23.0%) φ = -.20

The data relating to whether the offenders were accompanied or not was only

collected during the latter phase of the study (see section 5.3.3), accounting for the

smaller sample size. As can be seen, a significant difference was found between the

offenders with those who were unaccompanied (77.0%) being more likely to agree to

participate than those who were accompanied (58.5%). While this introduced a potential

source of bias, it did not appear to vary systematically across the sample. There were

similar representations of males/females (88.7/11.3% vs. 86.5/13.5%) and

Disqualified/Unlicensed Drivers (79.2/20.8% vs. 85.1/14.9%) among those who were

accompanied and those who weren’t. In addition, no significant differences were found

between the participants who were accompanied and those who weren’t in terms of

gender [χ2 (df1, n=88) = 0.99, p > .05, φ = .11], age [χ2 (df3, n=88) = 4.36, p > .05, φ c =

.22] or offence category [χ2 (df1, n=88) = 1.57, p > .05, φ = -.13].

As with the pilot study, the most common reason cited by the offenders who

refused to participate in the survey related to ‘being in a rush’ or ‘having no time’ (n=35).

A further 14 offenders specifically mentioned that they couldn’t participate due to work-


related commitments, while 9 cited other commitments. A breakdown of the reasons cited

for non-participation is provided in Appendix G.

5.4.1.2 Type of offender

The participants in the survey were charged with the offence of either Unlicensed

Driving or Disqualified Driving. However, within these two offence categories, there

were major differences relating to the reason that offenders were driving without a valid

licence. Table 5.2 provides a breakdown of the sample in terms of the relevant offence

and the reason for being unlicensed.

Table 5.2

Types of unlicensed drivers

Offence Reason for being without valid licence No. %

of offence %

of total Disqualified for drink driving 27 51.9 8.7

Disqualified for unlicensed driving 18 34.6 5.8

Other 7 13.5 2.3

Disqualified Driving

Sub-total 52 100.0

Cancelled (suspended) licence 109 42.4 35.3

Expired licence 91 35.4 29.4

Not currently licensed 21 8.2 6.8

Never licensed 26 10.1 8.4

Inappropriate licence 10 3.9 3.2

Unlicensed Driving

Sub-total 257 100.0

Total 309 100.0 The majority of the offenders charged with Disqualified Driving had previously lost

their licence for either drink driving (51.9%) or unlicensed driving (34.6%). The Other

category was mainly made up of offenders who had been disqualified as a result of a

Dangerous Driving conviction. Among the offenders charged with Unlicensed Driving,

two-fifths had had their licence cancelled as a result of the accumulation of demerit

points (42.4%), henceforth referred to as suspended drivers.18 The next two largest

groups were those who were driving on an expired licence (35.4%) and those who had

never held a licence (10.1%).

18. Queensland Transport refers to offenders whose licence they cancel administratively as suspended

drivers.


The next category of offender, the not currently licensed, is a group who are not

specifically identified in the Queensland crash statistics (see section 4.3.1). Of the 21

(8.2%) offenders in this category, 13 had a prior conviction for drink driving. Hence, it is

likely that many of these had failed to renew their licence following a period of

disqualification for drink driving. Accordingly, it is likely that the characteristics of this

group will be similar to the disqualified drivers. The inappropriate licence category

(3.9%) consisted of offenders who were driving a vehicle for which they were not

specifically licensed (eg. riding a motorcycle with an engine capacity exceeding 250ml

while on a provisional licence), were driving outside the conditions of a special licence or

were driving on an overseas licence.

In order to facilitate later comparisons between types of offenders it was decided to

collapse the disqualified drivers into one category. This served to increase the size of this

sub-group, while maintaining their unique status as offenders who were driving despite a

court-ordered ban (as opposed to licence suspension which is administratively imposed).

Based on this, Figure 5.1 provides a summary of the sample in terms of the reason for

which they were unlicensed.

Suspended35.3%

Disqualified16.8%

Expired29.4%

Not currently licensed

6.8%

Inappropriatelicence3.2%

Never licensed

8.4%

5.4.1.3 Socio-demographic characteristics of the sample

Table 5.3 provides a breakdown of the socio-demographic characteristics of the

sample by type of unlicensed driver. As can be seen, the overall sample was

predominantly male (83.5%). Although there was no significant difference between the

unlicensed driver subgroups, there were a higher proportion of females among the

offenders with expired licences.

Figure 5.1: Reason cited by participants for being unlicensed (n=309)


Table 5.3

Socio-demographic characteristics of participants by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level1

Gender n=52 n=109 n=91 n=21 n=26 n=10 n=3

09

Males 90.4 84.4 74.7 90.5 84.6 100.0 83.5 χ2 (df4) = 7.44,

Females 9.6 15.6 25.3 9.5 15.4 0.0 16.5 p > .05, φc? = .16

Age n=52 n=109 n=91 n=21 n=26 n=10 n=3

09

17-20 19.2 30.3 6.6 19.0 26.9 30.0 20.4 χ2 (df12) = 41.77,

21-25 42.3 41.3 26.4 38.1 19.2 20.0 34.3 p < .001,?? φc? =.22

26-39 28.8 26.6 59.3 33.3 46.2 50.0 39.5

40 and over 9.6 1.8 7.7 9.5 7.7 0.0 5.8

Marital status n=52 n=109 n=91 n=21 n=26 n=10 n=3

09

Single 65.4 78.0 67.0 52.4 69.2 80.0 70.2 χ2 (df8) = 10.73,

Married/defacto 23.1 18.3 22.0 38.1 26.9 20.0 22.3 p > .05, φc??=.13

Divorced or separated

11.5 3.7 11.0 9.5 3.8 0.0 7.4

Education level n=52 n=109 n=91 n=21 n=26 n=10 n=3

09

Grade 10 or < 63.5 33.9 37.4 76.2 84.6 10.0 46.3 χ2 (df12) = 44.74,

Grade 12 13.5 38.5 29.7 14.3 11.5 60.0 28.5 p < .001, φc? =.22

TAFE/Tech./ apprentice

13.5 12.8 13.3 9.5 0.0 10.0 12.0

University/CAE 9.6 14.7 18.7 0.0 3.8 20.0 13.3

Employed at time of court hearing

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Yes 59.2 70.6 66.3 63.2 53.8 70.0 65.6 χ2 (df4) = 2.83,

No 40.8 29.4 33.7 36.8 46.2 30.0 34.4 p > .05, φc =.10

Annual income n=52 n=109 n=91 n=21 n=25 n=9 n=3

07

Below $10,000 34.6 17.4 16.5 23.8 48.0 22.2 23.2 χ2 (df8) = 19.75,

$11,000-$30,000 42.3 51.4 48.4 52.4 44.0 44.4 48.1 p < .05, φc =.18

$31,000 or more 23.1 31.2 35.2 26.3 8.0 33.3 28.7

1. The Inappropriate licence category was excluded from the Chi-Square tests to ensure sufficient cell sizes. The cells with significant (p<.01) adjusted standardised residuals are bolded.

The overall sample was relatively young, with over half (54.7%) of the offenders

being 25 years of age or younger. There was a significant difference between the

unlicensed driver types in terms of age. An examination of the adjusted standardised

residuals highlights two issues of interest. The suspended drivers were predominantly

young with over three-quarters (71.6%) being under the age of 25. In contrast, those with


expired licences tended to be older with two-thirds (67.0%) being 26 years of age or

older.

Consistent with the youthful nature of the sample, most of the offenders were single

(70.2%) and had been educated to Grade 12 or less (74.8%). While there was no

significant difference between the unlicensed driver types in terms of marital status there

was for education level attained. In particular, there was a high representation of the not

currently licensed and never licensed drivers in the Grade 10 or less education category

(and an under-representation of suspended and expired drivers). The majority of the

sample were employed (65.0%) and earned between $11,000-30,000 per year (48.2%).

Although there was no significant difference between the offenders in terms of

employment, there was for annual income. The standardised adjusted residuals indicated

that the never licensed drivers were over-represented in the less than $10,000 category.

This is consistent with the high proportion of these offenders who were educated to

Grade 10 or less.

5.4.1.4 Driving and criminal history of offenders

Table 5.4 summarises the self-reported driving and criminal history of the offenders

by type of unlicensed driver. The driving history variables include years of driving

experience, past convictions for unlicensed/disqualified driving, and other traffic offences.

The item relating to driving experience specifically measured both legal and illegal

driving (ie. included periods of unlicensed driving). In light of this, the offenders in the

sample had considerable driving experience. The overall mean driving experience of the

offenders was 9.1 years. Due to vio lations of normality and homogeneity of variance, the

data were analysed using a Kruskal-Wallis (H) test. As can be seen, there was an overall

significant difference in the amount of driving reported by the different unlicensed driver

types. In particular, the suspended drivers had the least experience (M= 6.7 years),

consistent with the younger age distribution of this group. It is worth noting that the never

licensed group had a mean driving experience of 9.3 years. This highlights the extent of

the illegal driving undertaken by this group, given that these offenders would never have

had a valid licence.


Table 5.4

Driving and criminal history of participants by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Driving experience in years

n=52 n=107 n=89 n=21 n=26 n=10 n=305

Mean 9.9 6.7 10.8 10.6 9.3 11.1 9.1 H (df5)= 18.54,

Median 7.0 5.0 10.0 6.5 8.0 13.5 7.0 p < .01, η =.25

Std. deviation 8.1 5.0 7.7 9.6 8.2 7.1 7.3

Mean rank 160.1 126.1 177.9 158.3 149.6 180.7

Minimum 0.5 0.8 1.0 1.0 0.0 2.0 0.0

Maximum 35.0 25.0 42.0 38.0 30.0 19.0 42.0

Prior conviction for unlicensed/ disqualified driving1

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Yes (%) 71.2 29.4 23.1 61.4 61.5 0.0 39.2 χ2 (df5) = 57.70,

No (%) 28.8 70.6 76.9 28.6 38.5 100.0 60.8 p < .001, φc =.43

Other traffic offences1

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Yes (%) 82.7 88.1 69.2 85.7 65.4 70.0 79.0 χ2 (df5) = 15.02,

No (%) 17.3 11.9 30.8 14.3 34.6 30.0 21.0 p = .01, φc =.22

Prior criminal conviction1

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Yes (%) 65.4 30.3 28.6 42.9 65.4 10.0 38.8 χ2 (df5) = 34.19,

No (%) 34.6 69.7 71.4 57.1 34.6 90.0 61.2 p < .001, φc =.33

1. The cells with significant (p<.01) adjusted standardised residuals are bolded.

Overall, over one third (39.2%) of the sample reported having a prior conviction for

unlicensed or disqualified driving. Moreover, there was a significant difference among the

unlicensed driver types with three sub-groups being particularly likely to have a prior

conviction: the disqualified drivers (71.2%), the never licensed drivers (61.5%) and the not

currently licensed drivers (61.4%). In the case of the disqualified drivers this is not

surprising, given that over a third of these offenders had originally been disqualified for

unlicensed driving. However, the proportion of never licensed drivers reporting prior

convictions for unlicensed driving is somewhat surprising. It highlights that many of these

offenders were prepared to continue to drive unlicensed despite being previously detected

and punished.


Over one-third of the offenders (38.8%) reported having a prior criminal offence.

Similar to the findings relating to prior conviction for unlicensed/disqualified driving, there

was a much higher proportion of prior criminal convictions among the disqualified (65.4%)

and never licensed (65.4%) drivers. Indeed, a significant association was found between the

likelihood of prior unlicensed/disqualified driving offences and criminal offences [φ =.29, p

< .001], suggesting a strong relationship between the two variables.

5.4.1.5 Driving for work purposes

The participants were asked: “While you were unlicensed/disqualified, did you

need to drive as part of your job?” (Q. 29). Table 5.5 reports the results for this question

by offender type. Overall, over two-fifths (41.7%) reported that they did need to drive for

work purposes while unlicensed. While there were no significant differences between the

offenders, it is a concern that a substantial proportion of the disqualified, suspended and not

currently licensed drivers were continuing to drive for work. In some of these cases, the

participants would have been self-employed. However, in other cases the participants were

presumably able to avoid their employer becoming aware that they had lost their licence. It

is also worth noting that four of the never licensed drivers were driving for work purposes.

These findings have important implications for work-related road safety initiatives (see

section 7.4.2.4).

Table 5.5

Needed to drive for work purposes while unlicensed by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Needed to drive for work while unlicensed

n=51 n=109 n=91 n=21 n=25 n=10 n=307

Yes 43.1 46.8 42.9 38.1 16.0 40.0 58.3 χ2 (df5) = 8.17,

No 56.9 53.2 57.1 61.9 84.0 60.0 41.7 p > .05, φc =.16

5.4.1.6 Sensation seeking

As noted in section 5.3.3.3, the 10 sensation seeking items included in the

questionnaire comprised the Thrill and Adventure Seeking (TAS) subscale of the

Sensation Seeking Scale (SSS). As such, the ten items were summated to produce a scale

with a Cronbach’s alpha of .71 (see F1 in Appendix F). Table 5.6 reports the mean

sensation seeking scores for the different types of unlicensed drivers. The mean for the


sample (M= 6.8) was higher than that obtained for a sample of general drivers (M= 5.9) in

a recent Queensland study that used the same sensation seeking subscale (Tay,

Champness & Watson, 2003).19 This suggests that unlicensed drivers may be relatively

high sensation seekers. However, no significant differences were found among the

different types of offenders, suggesting that the trait is relatively uniform among

unlicensed drivers.

Table 5.6

Sensation seeking scores by type of unlicensed driver


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Sensation seeking score n=51 n=107 n=91 n=21 n=25 n=8 n=303

Mean 6.8 7.1 6.5 6.3 6.4 7.1 6.8 F (5,297) = 1.01,

Median 8.0 8.0 7.0 6.0 7.0 7.5 7.0 p > .05, η2 = .02

Std. deviation 2.5 2.4 2.6 2.1 2.0 2.4 2.4

Minimum 0.0 0.0 0.0 3.0 2.0 3.0 0.0

Maximum 10.0 10.0 10.0 10.0 10.0 10.0 10.0

Further analyses indicated that the participants’ sensation seeking score was not

significantly associated with either prior conviction for unlicensed driving [rpb = .02, p

>.05] or a conviction for another type of traffic offence [rpb = .09, p >.05]. However, it

was significantly correlated (albeit weakly) with prior criminal conviction [rpb = .11, p

<.05]. As will be reported in section 5.4.3.4, there was also a significant correlation

between sensation seeking and self- reported speeding.

5.4.1.7 Alcohol misuse

Table 5.7 provides a breakdown of the participants’ scores on the Alcohol Use

Disorders Identification Test (AUDIT). The Cronbach’s alpha for the ten items

comprising this test was .80 (see F2 in Appendix F). While the total possible score on the

test is 40, a score of 8 or more suggests that a person is likely to have a hazardous or

harmful level of alcohol consumption (Early Intervention Unit, 1993).

19. The sample in this study had a higher proportion of females and older drivers than the current study.

This would have in part contributed to the lower mean sensation seeking score obtained.


Table 5.7

AUDIT scores by type of unlicensed driver


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

AUDIT score n=52 n=109 n=91 n=21 n=26 n=10 n=309

Mean 11.7 8.6 7.7 13.4 12.7 6.9 9.5 F (5,303) = 4.66,

Median 12.0 7.0 7.0 13.0 12.0 7.5 9.0 p < .001, η2 = .07

Std. deviation 8.2 7.1 7.0 8.3 8.7 5.9 7.7

Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Maximum 37.0 37.0 30.0 33.0 31.0 16.0 37.0

% of sample with score of 8 or more

69.2 49.5 40.7 81.0 76.9 50.0 54.7

As shown in Table 5.7, over half (54.7%) of the participants in the study met the

criteria for hazardous/harmful alcohol consumption. In particular, four of the unlicensed

driver types had mean scores exceeding eight: the not currently licensed (M=13.4), the

never licensed (M=12.7), the disqualified (M=11.7) and the suspended (M=8.6) drivers.

The difference among the unlicensed driver types was significant [F (5, 303) = 4.66, p <

.001, η2=.07]. A Tukey’s HSD post-hoc test showed that the not currently licensed

[p=.02], the never licensed [p=.03], and the disqualified [p=.03] drivers all had

significantly higher mean scores than the expired drivers.

Table 5.8 examines the relationship between the AUDIT score and various

indicators of risk taking. Significant associations were found between the participant’s

AUDIT score and the likelihood of having: a prior unlicensed driving offence [rpb=.17, p

<.01]; a conviction for another traffic offence [rpb=.18, p =.001]; and a prior criminal

offence [rpb=.22, p <.001]. Interestingly, there was also a weak positive correlation

between the AUDIT and sensation seeking scores that approached significance [r =.11, p

=.05]. The relationship between the AUDIT score and self- reported drink driving is

examined in section 5.4.3.4.


Table 5.8

Bivariate correlations between AUDIT scores and driving offences, criminal offences and

sensation seeking

Variables Prior conviction

for unlicensed driving

Prior convi ction for another

traffic offence

Prior criminal conviction

Sensation seeking

AUDIT score .17** .18*** .22*** .11

* p < .05 ** p < .01 *** p ≤ .001

5.4.2 Circumstances of detection and outcome of court hearing

5.4.2.1 Reason stopped by the police

Figure 5.2 summarises the participants’ reported reasons for being stopped by the

police, when detected. The most common reason for the offenders being stopped was that

they had committed a traffic offence. Within this category, the most common offences

were: speeding (44.2%); vehicle registration offences (eg. expired registration, no

registration plates, obscured plates) (16.3%); and seat belt offences (relating to the driver

or passengers) (10.6%).

Involved in crash3.9%

Traffic offence33.7%

RBT22.0%

No reason/ unsure18.4%

Other12.6%

Licence/rego check

9.4%

The second most common reason that offenders reported being stopped by the

police was as a result of a random breath test. It is interesting to note that the police do

not routinely check licences at RBT operations in Queensland (Watson, 1998d). (Indeed,

data presented in section 5.4.3.5, indicate that almost one third of the participants were

pulled over by RBT when driving unlicensed and did not have their licence checked.) Not

surprisingly, it is likely that of the 68 offenders actually detected by RBT at least 20

(29.4%) registered a positive breath test at the time. This is based on the fact that these 20

offenders were convicted of drink driving on the same day they appeared in court for

their unlicensed/disqualified driving charge.

Figure 5.2: Reported reason for being stopped by police (n=309)


Of the remaining offenders in the sample:

• 18.4% were either unsure or could provide no reason as to why they were stopped

by the police;

• 12.6% cited a range of other reasons, mainly relating to the assumption that their

vehicle or driving had attracted the attention of the police;

• 9.4% specifically mentioned that they were pulled over for a routine or random

licence check or registration check; and

• 3.9% were identified as a result of being involved in a crash.

Based on the above information, the reasons that the police stopped offenders were

re-classified into four broad categories:

1. illegal driving (comprising those who were caught for committing a traffic offence

or being involved in a crash);

2. RBT (comprising those detected through RBT, irrespective of whether they were

subsequently charged with drink driving or not);

3. a targeted check (comprising cases where offenders were stopped for a licence or

registration check, either randomly or due to the nature of their driving behaviour or

the characteristics of their vehicle); or

4. an other category (comprising cases where the reason was unclear).

Table 5.9 provides a breakdown of the reasons offenders were stopped by type of

unlicensed driver, using the above categories. As can be seen, 37.9% of the offenders

were detected as a result of an illegal behaviour while a further 48.9% were detected

through random or targeted enforcement. There was no significant difference between the

offenders in terms of the method of their detection.

Table 5.9

Reason stopped by police by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level1

Reason stopped n=52 n=109 n=91 n=21 n=26 n=10 n=30

9

Illegal behavior 32.7 38.5 42.9 28.6 30.8 50.0 37.9 χ2 (df12) = 14.44,

RBT 17.3 22.0 23.1 28.6 30.8 0.0 22.0 p > .05, φc =.13

Targeted check 28.8 25.7 29.7 19.0 23.1 30.0 26.9

Other 21.2 13.8 4.4 23.8 15.4 20.0 13.3

1. The Inappropriate licence category was excluded from the Chi-Square test to ensure sufficient cell sizes.


5.4.2.2 Reason for driving when detected

Figure 5.3 shows the participants’ reported reasons for driving on the occasion that

they were detected by the police. The most common group of reasons related to

social/recreational activities (53.9%), followed by work-related (26.0%) and family

(15.9%) reasons. The other category mainly involved personal-related activities, such as

shopping. There was no significant difference between the various types of unlicensed

drivers and their reason for driving when detected [χ2 (df8, n=286) = 7.10, p > .05, φc

=.11]20.

Work-related26.0%

Family-related15.9%

Other4.2%

Social/ recreat-ional

53.9%

The predominance of driving for social/recreational and family-related driving

suggests that many unlicensed driving trips are discretionary in nature. A possible

exception to this relates to the behaviour of the offenders who were driving for work-

related reasons. It is possible that the primary motivation of these drivers was to retain

their job. In this regard, Question 29 asked the participants whether they needed to drive

as part of their job during the period in which they were unlicensed. A significant

association was found between this variable and whether the offenders were caught driving

for work-related purposes or not [φ =.25, p < .001].

5.4.2.3 Vehicle driven when detected

The majority of the offenders (92.6%) were detected driving a car. A further 5.2%

were riding a motorcycle while the remaining offenders were driving a truck or bus.

Unfortunately, the small number of offenders who were caught riding a motorcycle

makes it difficult to conduct separate analyses on this group. Nonetheless, it worth noting

that of the 16 motorcycle riders interviewed, 6 were suspended, 5 were inappropriately

licensed (2 only had a car licence while the other 3 were riding motorcycles with an

20. This analysis excluded the ‘Inappropriate licence’ category of offender and the ‘Other’ category of

reason for driving to ensure sufficient cell sizes.

Figure 5.3: Reported reason for driving when stopped by police (n=308)


engine capacity >250mls on a provisional licence), 3 were disqualified and 2 did not

currently hold licences.

Figure 5.4 provides a breakdown of who owned the vehicle driven by the offenders

at the time they were detected. It excludes 29 offenders for whom vehicle ownership was

unknown. The majority of the offenders were driving a vehicle owned by themselves

(62.5%), followed by a friend (21.4%) or family member (11.4%). The other category

(4.6%) was mainly made up of work vehicles.

Family member11.4%

Friend21.4%

Other4.6%

Respondent62.5%

As shown in Table 5.10, there was a significant difference among the offenders in

terms of vehicle ownership. In particular, the suspended drivers were more likely to be

driving a vehicle that they owned (72.2%), whereas the least likely were those who had

never been licensed (32.0%). This latter result is still surprising, since it indicates that

almost one-third of the never licensed drivers actually owned a vehicle.

Table 5.10

Ownership of the vehicle driven by participants by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level1

Vehicle ownership

n=49 n=97 n=82 n=20 n=25 n=7 n=280

Respondent 59.2 72.2 64.6 45.0 32.0 85.7 62.5 χ2 (df5) = 18.40,

Other 40.8 27.8 35.4 55.0 68.0 14.3 37.5 p < .01, φc =.26

1. The cells with significant (p<.01) adjusted standardised residuals are bolded.

Figure 5.4: Owner of vehicle driven by offenders when detected


5.4.2.4 Outcome of court hearing

Conviction for unlicensed/disqualified driving

The majority of the offenders (86.4%) were convicted of the

unlicensed/disqualified driving offence with which they were charged. Among the 42

offenders who weren’t convicted, 37 had the matter adjourned, 4 had no conviction

recorded and 1 was acquitted. A significant difference was found between the offenders

in terms of whether they were convicted of the offence on the day [χ2 (df4, n=298) = 9.60,

p < .05, φc =.18].21 In particular, the disqualified drivers (25.5%) were more likely to have

the matter adjourned than the other offenders. This would appear to relate to the more

serious nature of the offence. In many cases the offenders were granted an adjournment in

order to obtain further legal advice.

Penalties for unlicensed/disqualified driving

In cases where offenders were found guilty of more than one traffic offence, the

Magistrates would generally combine the relevant fines (and disqualification period if

applicable) into one penalty. As a consequence, to obtain an accurate indication of the

range of penalties applied to unlicensed/disqualified drivers it was necessary to exclude

the offenders with multiple offences from the analysis.

Table 5.11 provides a breakdown of the penalties received for Disqualified and

Unlicensed driving for those offenders who were not convicted of any additional

offences. Due to the exclusion of these latter offenders, the sample size for some of the

groups is relatively small. The fines and disqualification periods applied to the

disqualified drivers were typically much higher than for the other offenders. The average

fine for the disqualified drivers was almost $900 and the majority received an Absolute

Disqualification. This requires offenders to serve a disqualification period of at least two

years before being eligible to apply for the return of their licence from a court. Among

the other offenders, the lowest average penalties were applied to the drivers with expired

licences, in-keeping with the more administrative nature of this offence.

21. This analysis excluded the ‘Inappropriate licence’ category of offender t o ensure sufficient cell sizes.


Table 5.11

Penalties received for unlicensed/disqualified driving conviction among offenders who

were convicted of no other offences

Unlicensed Driver Type Variable

Disqualified Suspended Expired Not

currently licensed

Never licensed

Inapp. licence

Fine ($) n=19 n=62 n=53 n=12 n=12 n=5

Minimum 200 75 35 140 100 100

Maximum 1500 1400 420 500 800 400

Average 895 339 198 268 271 240

Disqualification period (Months)

Minimum 3 0 0 0 0 0

Maximum Absol.1 24 10 6 10 6

Average 22.8 2 2.7 0.7 0.7 2.8 3.0

1. Absolute disqualification. 2. Based on the minimum requirement for offenders who are Absolutely Disqualified to serve at least two years

disqualification. 5.4.3 Unlicensed driving behaviour

5.4.3.1 Length of time driving unlicensed

The reported length of time (in years) that offenders had driven unlicensed is

summarised in Table 5.12. Separate breakdowns are provided for all offenders, first

offenders and repeat offenders. In the case of the repeat offenders, the data relates to the

total length of time they reported driving unlicensed during their driving career. Kruskal-

Wallis (H) tests were performed on the data due to violations of normality and

homogeneity of variance.

As expected, the mean time driving unlicensed was much lower among first

offenders (Mean = 0.88 years; Median = 0.25 years). Although there was no overall

significant difference between offenders, it is interesting to note some of the highest and

lowest values among the first offenders. An offender who was riding a motorcycle on an

inappropriate licence reported riding over a 19–year period before being caught, while a

never licensed driver reported driving over a 15–year period. In contrast, a number of

offenders reported being detected on the first occasion that they had driven unlicensed.


Table 5.12

Length of time driving unlicensed by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

All offenders

Years driving unlicensed1

n=47 n=102 n=88 n=18 n=26 n=10 n=291

Mean 3.23 1.05 1.37 4.69 8.10 2.39 2.40 H (df5)= 37.83,

Median 0.75 0.33 0.33 2.75 5.75 0.46 0.50 p < .001, η = .43

Std. deviation 5.32 2.23 3.22 5.28 8.68 5.86 4.78

Minimum 0.003 0.001 0.001 0.083 0.001 0.058 0.001

Maximum 27.0 14.0 24.0 18.0 30.0 19.0 30.0

First offenders


n=14 n=77 n=70 n=6 n=10 n=10 n=187

Mean 0.69 0.51 0.58 2.54 3.48 2.39 0.88 H (df5)= 7.04,

Median 0.25 0.21 0.17 0.83 0.79 0.46 0.25 p > .05, η = .38

Std. deviation 1.34 0.69 1.11 3.22 5.15 5.86 2.1

Minimum 0.003 0.001 0.001 0.167 0.001 0.058 0.001

Maximum 5.0 3.5 8.0 8.0 15.0 19.0 19.0

Repeat offenders


n=33 n=25 n=18 n=12 n=16 n=104

Mean 4.30 2.69 4.42 5.72 10.99 5.13 H (df4) = 14.06,

Median 2.00 1.00 1.83 4.25 8.5 2.00 p < .01, η = .40

Std. deviation 6.00 3.04 5.98 5.92 9.31 6.63

Minimum 0.003 0.005 0.083 0.083 0.333 0.005

Maximum 27.0 14.0 24.0 18.0 30.0 30.0

Among the repeat offenders, the mean length of time driving unlicensed was 5.13

years (median = 2 years). While these data highlight the long periods over which many

offenders had driven unlicensed (particularly repeat offenders), it provides little insight

into the extent of their driving. This is examined further in the next section.

5.4.3.2 Frequency of unlicensed driving

The offenders were asked how many times a week they had driven prior to getting

caught, for both work-related and family/recreational reasons. They were advised to

count going somewhere and returning home as different trips. The relevant findings are

summarised in Table 5.13.


Table 5.13

Frequency of driving while unlicensed by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Work-related reasons

Trips per week1 n=52 n=108 n=91 n=21 n=26 n=10 n=308

Mean 5.4 8.1 7.3 5.8 2.4 6.4 6.7 H (df5)= 19.47,

Median 5.0 9.0 10.0 6.0 0.0 6.0 6.0 p = .01, η = .24

Std. deviation 5.2 7.8 6.9 5.4 3.8 5.8 6.8

Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Maximum 20.0 50.0 30.0 14.0 12.0 14.0 50.0

Social/recreational reasons


Mean 6.0 7.6 7.0 10.6 6.7 7.0 7.3 H (df5)= 4.73,

Median 2.0 6.0 6.0 4.0 4.0 4.0 6.0 p > .05, η = .12

Std. deviation 7.9 8.1 6.9 15.4 8.9 9.4 8.5

Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Maximum 30.0 50.0 30.0 56.0 42.0 30.0 56.0 All reasons


Mean 11.4 15.7 14.3 16.4 9.1 13.4 14.0 H (df5)= 12.03,

Median 10.0 14.0 14.0 14.0 7.0 12.0 12.0 p < .05, η = .18

Std. deviation 10.7 13.5 11.3 15.6 9.8 11.7 12.4

Minimum 0.0 0.0 0.0 0.0 0.0 0.0 0.0

Maximum 42.0 80.0 60.0 56.0 42.0 30.0 80.0

1. Note that some offenders reported that they didn’t drive on a weekly basis and were assigned a score of zero. The mean number of trips per week for work-related reasons was 6.7 (Median =

6.0). This relatively low mean was strongly influenced by the large number of offenders

who reported that they did not make any trips for work-related purposes (105 offenders

representing 34.0% of the sample). The maximum value of 50 trips was achieved by a

suspended driver who was working as a courier. (Incidentally, this participant reported

that he subsequently lost his job as a result of being detected.)

Due to the positively skewed nature of the distribution, a Kruskal-Wallis test was

used to test for differences among the unlicensed driver types. As can be seen in Table

5.13, this test found a significant overall difference among the offenders. Inspection of

the medians suggested that this difference was due to the low number of trips reported by


the never licensed drivers. This is consistent with the previous finding that this group of

offenders had the highest level of unemployment (46.2%) in the sample (see Table 5.3).

The mean number of trips per week for family/recreational reasons was slightly

higher at 7.3 (Median = 6.0). Once again, there were a large number of offenders who

reported that they did not make any trips for family/recreational reasons (72 offenders

representing 23.3% of the sample). A Wilcoxon Matched-Pairs Signed-Ranks test was

conducted to ascertain whether the higher level of reported trips for family/recreational

reasons (compared with work-related trips) was significant. This test found no significant

difference between the two types of trips [T = -0.16, p > .05]. In addition, post-hoc

comparisons of work-related and family/recreational trips were performed for each of the

offender types using a more stringent alpha (.01). The only comparison approaching

significance was for the never licensed drivers [T = -2.53, p = .011] who reported more

trips for social/recreational reasons.

When the two reasons for driving were combined, only 31 offenders (10.0%)

reported that they didn’t undertake at least one trip per week. The overall mean number

of trips was 14.0 per week (Median = 12.0). Interestingly, this represents at least one

return trip per offender each day of the week. The offender types with the highest

means/medians were the suspended, not currently licensed and expired drivers, while the

lowest was the never licensed drivers.

5.4.3.3 On-road driving behaviour

Cautiousness when driving

Question 41 asked the participants to rate on a 7-point Likert scale how much more

careful they were in obeying a range of road rules during the time they were driving

unlicensed. The items related to obeying: the speed limit, traffic lights, Stop and Give

Way signs, drink driving laws, seat belt laws and other traffic rules. There was a higher

level of missing data than usual with this question (8.4%), particularly among the

offenders who claimed that they were unaware of being unlicensed. Among the

participants who responded to all of the items, the Cronbach’s alpha was quite high (.87),

so the scores were combined to create a Care in obeying the road rule scale (see F3 in

Appendix F). Table 5.14 provides a breakdown of the scores on this scale by offender

type.


Table 5.14

Cautiousness of driving while unlicensed by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Care in obeying road rules scale

n=49 n=99 n=81 n=20 n=25 n=9 n=283

Mean 36.9 35.0 35.2 35.2 36.8 34.4 35.5 H (df5)= 4.73,

Median 38.0 37.0 37.0 35.5 39.0 36.0 37.0 p > .05, η = .12

Std. deviation 4.9 7.2 6.6 6.0 8.2 6.5 6.7

Minimum 22 6 21 22 10 24 6

Maximum 42 42 42 42 42 42 42

Limited driving (all respondents)1 n=49 n=104 n=82 n=20 n=25 n=10 n=290

Yes (%) 69.4 49.0 24.4 70.0 56.0 70.0 48.3 χ2 (df5) = 33.77,

No 30.6 51.0 75.6 30.0 44.0 30.0 51.7 p < .001, φc =.34,

Limited driving (those aware of being unlicensed)1

n=15 n=64 n=42 n=19 n=25 n=9 n=174

Yes (%) 73.3 56.3 38.1 73.7 56.0 77.8 56.3 χ2 (df5) = 11.45,

No 26.7 43.8 61.9 26.3 44.0 22.2 43.7 p < .05, φc =.26

Types of roads used1,2

n=49 n=103 n=87 n=21 n=23 n=10 n=293

Back roads (%) 14.3 12.6 10.3 9.5 13.0 10.0 11.9 χ2 (df5) = 6.45,

Back roads and main roads

65.3 61.2 60.9 81.0 73.9 50.0 63.8 p > .05, φc =.11

Main roads 20.4 26.2 28.7 9.5 13.0 40.0 24.2

1. The cells with significant (p<.01) adjusted standardised residuals are bolded. 2. The Inappropriate licence category was excluded from the Chi-Square test to ensure sufficient cell sizes.

As can be seen, the overall mean score on the scale was relatively high (35.5 from a

total possible score of 42). This represents an average of almost 6 for each item in the

scale (where a score of 7 equated to being ‘much more careful’). As such, the participants

generally considered themselves more careful obeying the road rules during the period in

which they were unlicensed, than they would otherwise have been. There was no

significant difference across the offender types. In addition, there was no significant

association between the need to driver for work while unlicensed and care in obeying the

road rules [rpb= -.06, p >.05].

Question 42 asked the participants whether they limited or altered their driving in

anyway in terms of when or where they drove while unlicensed. As shown in Table 5.14,

among the total offenders responding to this question, only 48.3% indicated that they did

limit their driving in some way. There was a significant difference across the offender


types, with the disqualified drivers being more likely to limit their driving and the expired

drivers being the least likely. However, many of the expired and suspended drivers

reported that they were unaware of being unlicensed at the time they were detected (see

section 5.4.4.1). Consequently, a further analysis was undertaken excluding those

offenders who were unaware of being unlicensed. Once this was done, the proportion of

offenders indicating that they did alter their driving rose to 56.3% and only the expired

drivers were significantly less likely to limit their driving. Among the respondents who

reported limiting their driving, the main methods cited were: restricting the overall

amount of driving (47.1%); driving on back streets/avoiding main roads (10.0%); and

only driving during the day (10.0%). Interestingly, three respondents said that they only

drove during peak periods, while two others reported avoiding these times.

Question 43 related to the types of roads on which the participants drove while

unlicensed. The responses were collapsed into three categories and are shown in Table

5.14. Among all the offenders responding to the question, only 11.9% reported driving

‘only on back roads’ or ‘mainly on back roads’. The majority (63.8%) reported driving on

‘both back roads and main roads’. There was no overall significant difference across

offender types as to where they reported driving.

Speeding behaviour

Question 37 asked the participants: “While you were driving unlicensed/

disqualified, how often did you drive at 10 km/h or more over the speed limit?” As noted

in the Method section, this question was directly modelled on an item regularly used in

the ATSB’s Community Attitudes to Road Safety telephone survey (ATSB, 2000). Figure

5.5 compares the responses obtained in the current study with those obtained in the most

recent ATSB (2000) survey, which featured a sample size of over 1400 licensed drivers

who had driven within the last two years.22

As can be seen, 25% of the unlicensed drivers reported that they always, nearly

always or on most occasions exceeded the speed limit by 10 km/h or more, compared

with only 10% of the ATSB respondents. While it is problematical to compare responses

across surveys that feature different methodologies, these results suggest that the level of

self-reported speeding among the unlicensed drivers may be relatively high. Interestingly,

there was no significant difference in self-reported speeding among the different types of

unlicensed drivers [H (df5, n=286) = 6.36, p > .05, η = .13]. However, the item was

significantly (albeit weakly) correlated with sensation seeking [rpb = .12, p <.05].

22. The results are shown as whole percents to enable direct comparison with the ATSB (2000) survey

results.


6%2%

6%3%

13%5%

22%20%

30%

49%23%

20%

0 10 20 30 40 50

Always

Nearly always

Most occasions

Sometimes

Just occasionally

Never

Unlicensed drivers ATSB survey

Figure 5.5: Reported frequency of driving 10km/h or more over the speed limit

(n= 286/1430)

Seat belt wearing

Question 38 asked the participants: “...... while you were driving unlicensed/

disqualified how often did you wear a seat belt?” Once again, this question was modelled

on one regularly used in the ATSB’s community attitudes to road safety telephone survey

(ATSB, 2000). Figure 5.6 compares the responses obtained in the two surveys. As can be

seen, 85% of the unlicensed drivers reported that they always wore their seat belt,

compared with 96% of the respondents in the ATSB survey. Bearing in mind the afore–

mentioned problems in comparing surveys, these results suggest that the level of self-

reported seat belt wearing among unlicensed drivers, although very high, is lower than

that reported by general drivers. Once again, there was no significant difference in the

responses to this question among the different types of unlicensed drivers [H (df5, n=283)

= 2.38, p > .05, η = .12]. There was also no significant association between seat belt

wearing and sensation seeking [r = .01, p >.05].


85%

96%6%

3%

3%0%

3%1%

2%1%1%1%

0 20 40 60 80 100

Always

Nearly always

Most occasions

Sometimes

Just occasionally

Never


Figure 5.6: Reported frequency of wearing a seat belt as driver (n=283/1593)

Drink driving behaviour

Question 39 related to the participants’ general approach to drinking and driving

during the time they were unlicensed (see Figure 5.7). The majority of the unlicensed

driving respondents indicated that they either didn’t drink at any time or didn’t drink if

they were driving (67%). While a further 26% indicated that they restricted their drinking

if they were driving, 7% reported that they didn’t restrict their drinking at all. While the

ATSB (2000) survey found a lower proportion of general drivers reporting that they

either didn’t drink at any time or didn’t drink if they were driving (58%), less than 1% of

their sample admitted to not restricting their drinking when they were driving.

45%18%

22%40%

26%42%

7%

0%

0 10 20 30 40 50

Didn't drink

Didn't drink if driving

Restricted drinking if driving

Didn't restrict drinking when driving


Figure 5.7: General approach to drink driving (n=288/1453)


Table 5.15 examines differences in the reported drink driving behaviour of the

unlicensed drivers by offender type. In the case of the participant’s general approach to

drink driving, the two categories of ‘Didn’t drink at any time’ and ‘Didn’t drink if

driving’ were collapsed into one category (Don’t drink and drive), to facilitate the

Chi-square analysis. As shown, there was a significant difference between the offenders

on this question. The never licensed drivers were the most likely to report that they didn’t

restrict their drinking when driving, while the not currently licensed drivers were the least

likely to report that they didn’t drink and drive at all.

Table 5.15

Drink driving behaviour while unlicensed by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

General approach to drink

1,2

n=48 n=101 n=86 n=19 n=24 n=10 n=288

Don’t drink and drive

60.4 78.2 67.4 31.6 58.3 70.0 67.0 χ2 (df8) = 34.51

Restricted drinking if driving

25.0 18.8 29.1 57.9 16.7 30.0 25.7 p < .001, φc =.25

Didn’t restrict drinking if driving

14.6 3.0 3.5 10.5 25.0 0.0 7.3

Drove when they thought they were over limit1

n=48 n=102 n=86 n=19 n=24 n=10 n=289

Yes 37.5 18.6 17.4 36.8 28.0 22.2 23.5 χ2 (df5) = 10.71,

No 62.5 81.5 82.6 63.2 72.0 77.8 76.5 p = .058, φc =.19

1. The cells with significant (p<.01) adjusted standardised residuals are bolded. 2. The Inappropriate licence category was excluded from the Chi-Square test to ensure sufficient cell sizes.

Question 50 concerned whether the participants had ever driven when they thought

they might have been over the legal limit, during the time they were unlicensed. As

shown in Table 5.15, 23.5% of the drivers admitted to this. The differences among the

offenders were approaching significance (p=.058), with the disqualified (37.5%), not

currently licensed (36.8%) and never licensed (29.2%) drivers the most likely to admit to

driving when they thought they may have been over the limit. In addition, a strong

positive correlation was found between this item and the participants’ AUDIT scores

[rpb=.49, p <.001]. The strength of this relationship tends to confirm the validity of this


item and reflects the link between alcohol misuse and drink driving behaviour. However,

there was no significant association between reported drink driving and sensation seeking

[rpb = .01, p >.05].

5.4.4 The impact of current administrative, enforcement and punishment processes

5.4.4.1 Awareness of being unlicensed at time of detection

In total, 100 (36.0%) participants claimed that they were unaware, or at least

unsure, that they were unlicensed at the time they were detected. Table 5.16 provides a

breakdown of these participants by type of unlicensed driver. The disqualified and never

licensed drivers were excluded, due to the very small numbers of these offenders who

reported being unaware of their invalid licence status.

Table 5.16

Awareness of being unlicensed when detected, by offender type

Unlicensed Driver Type

Variable Suspended

%

Expired

%


%

Inapp. licence

%

Total

%

Significance level1

Aware of being unlicensed

n=109 n=91 n=21 n=10 n=231

Yes 59.6 46.2 90.5 90.0 58.4 χ2 (df3) = 18.70,

No/unsure 40.4 53.8 9.5 10.0 41.6 p < .001, φc =.28

1. The cells with significant (p<.01) adjusted standardised residuals are bolded. As can be seen, a substantial proportion of the offenders with expired licences

(53.8%) claimed that they were unaware of being unlicensed. While almost one-fifth

(18.7%) of these drivers claimed that they did not receive a renewal notice in the mail,

most of the others admitted that they had either overlooked or forgotten to renew their

licence. Among the 44 suspended drivers who were unaware, 26 (59.1%) claimed that

they hadn’t received any notification in the mail (although some acknowledged that they

had changed address).

5.4.4.2 Possession of a photographic licence

Almost half of the offenders (49.0%) reported that they still had their photographic

licence when they were driving unlicensed. Table 5.17 provides a breakdown of these

participants by type of unlicensed driver. Not surprisingly, none of the never licensed

drivers reported that they had a photographic licence (since these drivers would never


have been officially issued with a licence). Consequently, these drivers were excluded

from the Chi-square test.

Table 5.17

Possession of a photographic licence by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level1

Still had photo licence

n=52 n=109 n=91 n=21 n=25 n=10 n=308

Yes 13.5 54.1 75.8 38.1 0.0 80.0 49.0 χ2 (df4) = 56.58,

No 86.5 45.9 24.2 61.9 100.0 20.0 51.0 p < .001, φc =.45

1. The never licensed drivers were excluded from the analysis. The cells with significant (p<.01) adjusted standardised residuals are bolded.

As shown in Table 5.17, there was a significant difference among the remaining

offenders with the expired drivers being the most likely to still have their photographic

licence and the disqualified drivers being the least likely. The results are not at all

surprising in the case of the expired drivers, since there is no mechanism in place to

recover the licences of drivers who fail to renew their licence on time. However, it

remains a concern that there were many suspended, disqualified and not currently

licensed drivers who still had their photographic licences. Many of these participants

claimed that they did not realise that they were meant to surrender their licence or that no

one had requested them to do so. Consistent with this, the participants who retained their

photographic licences were less likely to be aware of being unlicensed [χ2 (df1, n=277) =

60.12, p < .001; φ =.47]. Nonetheless, a total of 15 offenders acknowledged that they had

held onto their licence for identification purposes, while one admitted that they had done

it to “deceive the police”.

5.4.4.3 Unlicensed driving after detection

The offenders were asked whether they continued to drive unlicensed after being

detected by the police (ie. prior to the court hearing). For offenders with prior convictions

the question was phrased to refer to their most recent detection. As shown in Table 5.18,

almost one-third of the sample (30.5%) admitted that they did continue driving. However,

no significant difference was found between the offenders on this variable.


Table 5.18

Continued driving unlicensed after detection by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Continued to drive after detection

n=52 n=109 n=91 n=21 n=25 n=10 n=308

Yes 28.8 35.8 24.2 42.9 20.0 40.0 30.5 χ2 (df5) = 6.45,

No 71.2 64.2 75.8 57.1 80.0 60.0 69.5 p > .05, φc =.15

Among the offenders who admitted continuing to drive, the process of detection at

least appeared to temper the frequency of their driving. For example, the mean number of

trips per week among these drivers fell from 18.3 per week to 16.1 per week after being

detected by the police. A Wilcoxon Matched-Pairs Signed-Ranks test found that this

reduction in trips was significant [T = -2.91, p < .01]. Interestingly, although offenders

with a prior conviction for unlicensed driving were slightly more likely to report driving

after detection (33.1% cf. 28.9%), this difference was not significant [χ2 (df1, n=308) =

0.61, p > .05, φ =.04].

5.4.4.4 Evasion of detection

As detailed in section 5.4.2.1, many of the offenders in the sample were detected by

the police as a result of either RBT or due to an illegal driving behaviour, such as

committing a traffic offence or being involved in a crash. However, many of the offenders

reported incidents where they were pulled over by the police and didn’t have their licence

checked or were able to avoid the matter coming to the attention of the authorities. Table

5.19 provides a breakdown of these incidents. It should be noted that the data relate to the

most recent period of unlicensed driving among those offenders with prior convictions for

unlicensed driving.


Table 5.19

Incidents where offenders evaded detection while driving unlicensed

Type of enforcement/incident Variable

RBT Speeding offence

Other offence

Traffic crash

Speed camera ticket

All methods

Total number of offenders exposed to incident 164 71 51 23 24 225

% of total sample 53.1 23.0 16.5 7.4 7.8 72.8

Total number of offenders whose licence wasn’t checked 97 8 11 8 11 113

% of total sample 31.4 2.6 3.4 2.6 3.6 36.6 Number of offenders whose licence was not checked on one occasion

39 7 9 6 8 46

% of total sample 12.6 2.3 2.9 1.9 2.6 14.9 Number of offenders whose licence was not checked on two or more occasions

58 1 2 2 3 67

% of total sample 18.8 0.3 0.6 0.6 1.0 21.7

A total of 164 offenders were pulled over by a RBT operation at least once during

the time they were driving unlicensed, representing 53.1% of the sample. However, 97

(31.4% of total sample) of these offenders reported that they didn’t have their licence

checked on one or more occasions. Indeed, of these offenders, 58 (18.8% of total) failed

to have their licence checked on two or more occasions. In addition, a small number of

offenders also cited cases where they were pulled over for speeding or another offence

and did not have their licence checked (8 and 11 offenders, respectively). Another 8

offenders reported that they were involved in a traffic crash but were able to evade

detection. In these cases the crashes were either minor in nature and the police weren’t

called or they fled the scene. Finally, 11 offenders reported that they were able to evade a

speed camera ticket for which they were responsible. In these cases the offenders were

driving either another person’s car or a work vehicle, and hence were able to avoid the

penalty. In two of these instances, the offenders reported that another person lost their

licence as a result of the speeding offence(s) they committed.

In total, 113 offenders (representing 36.6% of the sample) reported one or more

instances where the police did not detect them when they could otherwise have been

identified. Of these offenders, 67 (21.7% of sample) evaded detection on two or more

occasions. These findings are very interesting in light of Stafford and Warr’s (1993)

arguments about the role of punishment avoidance in encouraging illegal behaviour (see


section 3.2.3). The influence of punishment avoidance on unlicensed driving behaviour is

examined in section 6.4.5.

5.4.4.5 Perceptions of enforcement and punishment processes

Perceived risk of apprehension

The participants were asked to rate (on a seven-point Likert scale) how likely they

thought they were to be caught for a variety illegal behaviours, including unlicensed

driving, prior to being detected for the offence (Question 33). As shown in Table 5.20,

the other behaviours included: being random breath tested; being caught speeding by

either radar or a speed camera, being caught if not wearing a seat belt; and being involved

in a crash.

Table 5.20

Perceived likelihood of being caught for illegal behaviours prior to being detected for

unlicensed driving

Type of event

Being caught driving

unlicensed

Being random breath tested

Being caught

speeding by radar

Being caught speeding by a speed camera

Being caught if not wearing a

seat belt

Being involved

in a crash

Significance level

Perceived likelihood of event

n=307 n=307 n=307 n=307 n=305 n=307

Mean 3.3 4.0 3.8 3.8 2.3 2.4 F (4,1154) = 61.56,

Median 3.0 4.0 4.0 4.0 1.0 2.0 p < .001, η2 = .17

Std.deviation 1.8 1.8 2.1 2.1 2.0 1.6

Minimum 1.0 1.0 1.0 1.0 1.0 1.0

Maximum 7.0 7.0 7.0 7.0 7.0 7.0

The mean score for the likelihood of being caught driving unlicensed driving was

3.3 while the median was 3.0 (where a score of 4 represents equally likely/unlikely).

Hence, the majority of the offenders thought it was unlikely that they would be caught for

unlicensed driving. In addition, a repeated measures ANOVA found an overall significant

difference between the participant’s perceptions. A series of paired-sample t tests

(utilising a more stringent alpha rate of .01) indicated that the perceived likelihood of

being detected for unlicensed driving was significantly:


• lower than the perceived likelihood of being random breath tested [t (306) = -5.55, p

< .001], being caught by a speed camera if speeding [t (306) = 3.08, p < .01] and

being caught by a radar if speeding [t (306) = -3.21, p = .001]; but

• higher than the likelihood of being involved in a crash [t (306) = 7.40, p < .001] and

being caught if not wearing a seat belt [t (304) = 6.93, p < .001].

To some degree, these questions would have been influenced by participant

perceptions toward the likelihood of performing the behaviour, not just their risk of

detection. For example, many participants commented that they always wore their seat

belt so they were unlikely to be caught for not wearing one (even though the question

specifically related to the likelihood of detection if they were unbelted). Nonetheless, the

results indicate that the perceived risk of detection for unlicensed driving in Queensland

is lower than that for drink driving or speeding. Interestingly, no significant difference

was found between the unlicensed driver types in terms of their perceived risk of being

caught for driving unlicensed (prior to detection) [F (5, 301) = 1.95, p > .05, η2=.03].

The participants were also asked to rate: “if you were to drive

unlicensed/disqualified in the future, how likely do you now think your chances of getting

caught are” (Question 34). Not surprisingly, the responses to this question showed a

significant increase in the participant’s perceived risk of apprehension (compared with

their reported perceived risk prior to getting caught), with the mean rating increasing

from M = 3.3 to M = 4.6 [t (306) = -9.37, p < .001]. As before, no significant difference

was found between the offenders on this question.

Knowledge of penalties

The participants were asked whether they knew what the fine for unlicensed/

disqualified driving was prior to getting caught (Q. 16). Overall, the level of knowledge

was relatively poor with only 42 (14.0%) participants reporting that they were aware of

the fine. There was no significant difference among the offenders on this question [χ2

(df4, n=291) = 7.15, p > .05, φc =.13].23

Perceived severity, certainty and swiftness of punishment

Three items were used to measure the perceived severity, certainty and swiftness of

punishment for unlicensed driving:

• Severity – “The penalties for unlicensed driving are very tough” (Q. 52a);

• Certainty – “You can sometimes avoid getting punished if you get caught for

unlicensed/disqualified driving” (Q. 52i – reversed scored); and

23. This analysis excluded the ‘Inappropriate licence’ category of offender t o ensure sufficient cell sizes.


• Swiftness – “You are likely to be punished quickly if you get caught for

unlicensed/disqualified driving” (Q. 52o).

As shown in Table 5.21, there was a significant difference between the participants’

responses to the three items. The lowest mean was obtained for the perceived severity of

punishment (M = 4.6) and the highest for the perceived certainty of punishment (M =

5.4). While all three means appear relatively high, it should be borne in mind that the

majority of the participants completed the questionnaire shortly after being sentenced in

court. This may have served to inflate the means. No significant difference was found

between the unlicensed driver types in terms of the perceived severity [F (5, 302) = 1.84,

p > .05, η2=.03], certainty [F (5, 303) = 2.16, p > .05, η2=.03] or swiftness [F (5, 301) =

1.46, p > .05, η2=.02] of punishment.

Table 5.21

Perceived severity, certainty and swiftness of punishment for unlicensed driving

Variable

Perceived severity

Perceived certainty

Perceived swiftness

Significance level

n=308 n=309 n=307

Mean 4.6 5.4 5.2 F (2,610) = 17.04,

Median 4.0 6.0 6.0 p < .001, η2 = .05

Std.deviation. 1.8 1.9 1.8

Minimum 1.0 1.0 1.0

Maximum 7.0 7.0 7.0

5.4.4.6 Intention to drive unlicensed in the future

The participants were asked to rate on a 7-point scale (1 – very unlikely to 7 – very

likely) how likely they were to drive without a licence sometime in the future. As shown

in Table 5.22, the overall mean was 2.7 indicating that offenders generally thought it was

unlikely that they would drive unlicensed in the future. While the expired drivers reported

the lowest intentions to drive unlicensed in the future, there was no overall significant

difference among the offender types.


Table 5.22

Intention to drive unlicensed in the future by offender type


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. licence

%

Total

%

Significance level

Intention to drive unlicensed1

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Mean 3.1 2.8 2.1 3.1 2.9 2.9 2.7 H (df5)= 10.22,

Median 2.0 2.0 1.0 3.0 2.0 2.0 1.0 p > .05, η = .17

Std. deviation 2.5 2.2 1.9 2.3 2.2 2.2 2.2

Minimum 1.0 1.0 1.0 1.0 1.0 1.0 1.0

Maximum 7.0 7.0 7.0 7.0 7.0 7.0 7.0

5.5 Discussion 5.5.1 Study limitations

The 62.4% response rate achieved in this study is relatively high when compared to

previous surveys of unlicensed drivers. In addition, it was very similar to the response

rate (61%) achieved in a survey of drink drivers conducted in central Queensland using a

similar methodology (Ferguson et al, 2000). The strategy to recruit participants through

the court system appeared to overcome the problems previously experienced in locating

unlicensed driving offenders. Along with the decision to offer a payment to participants,

these factors should have enhanced the representativeness of the sample. Nonetheless, it

is important to acknowledge the limitations of the current sample.

Firstly, over one-third of the eligible offenders approached to participate in the

survey refused. While there did not appear to be any systematic bias among those who

refused in terms of either offence category or interviewer, female offenders and those

unaccompanied by family or friends were more likely to agree to participate. Given that

an incentive payment was offered, it is possible that there was a higher refusal rate among

those who were currently employed. However, only 11 of the offenders who refused

specifically mentioned that they couldn’t participate due to work-related commitments. In

addition, some offenders would have been omitted from the study because they failed to

appear at court as directed. Among the serious offenders (such as those facing additional

drink driving or dangerous driving charges) this is a relatively rare event, since it will

generally represent a breach of bail conditions and result in the issuing of a warrant for

the arrest (Micola, 2002).


Secondly, the scope of the sample was limited in a number of respects. While the

Brisbane Court processes the largest number of traffic offenders in Queensland each year

(Micola, 2002), it primarily processes offenders who are detected in the inner city and

suburban area of Brisbane. Consequently, the degree to which the findings can be

generalised to other metropolitan and rural areas remains to be confirmed. Some of the

offender groups (particularly the inappropriately licensed drivers) were relatively small.

In addition, the sample did not include any under-age drivers since they are processed

through juvenile courts (which are considerably more difficult for researchers to access).

Therefore, caution should be exercised when generalising the findings to all unlicensed

drivers.

Finally, it is unclear to what extent the behaviour of the sample is indicative of

unlicensed drivers who have not been detected by the police. It is possible that offenders

who remain undetected are generally more cautious (and possibly safer) than those

caught by the police. However, it is important to note that many of the offenders in the

sample were detected through random enforcement processes. Indeed, only 37.9% of the

offenders were detected as a result of committing an illegal behaviour. Furthermore,

many of the offenders had been driving unlicensed for quite long periods of time before

they were detected. For example, among the first offenders the mean length of time

driving unlicensed prior to detection was almost one year [0.88 years]. Together, this data

suggests that many of the offenders had proven quite successful in evading detection until

caught by random enforcement processes.

5.5.2 Support for hypotheses


outlined in section 5.2.

Unlicensed drivers will report more frequent drink-driving, speeding and failure to

wear a seat belt than general drivers (H8)

A comparison of the data obtained in this survey with that from the ATSB’s (2000)

community telephone survey provides good support for this hypothesis. Seven percent of

the unlicensed drivers reported that they didn’t restrict their drinking when driving,

compared with less than 1% of the general drivers. Indeed, almost 25% of all offenders

(and over a third of the disqualified, not currently licensed and never licensed drivers)

reported driving unlicensed when they thought they might have been over the legal

alcohol limit. Similarly, 25% of the unlicensed drivers reported exceeding the speed limit


by 10 km/h or more on (at least) most occasions, compared with only 10% of the ATSB’s

respondents. Finally, 15% of the unlicensed drivers admitted that they didn’t always wear

their seat belt, compared with only 4% of the general drivers.

Caution should be exercised when comparing the results from surveys that use

different methodologies. However, the relevant items used in this study were directly

modelled on those in the ATSB survey. In addition, the results are consistent with those

obtained from crash-based studies, including Study One. Hence, while there is a need to

replicate these findings, they appear valid.

The self-reported drink driving behaviour of unlicensed drivers will be positively

associated with alcohol misuse and sensation seeking, while self-reported speeding

and non-seat belt use will be positively associated with sensation seeking (H9)

Partial support was obtained for this hypothesis. A significant positive correlation

was found between the participants’ self-reported drink driving and their AUDIT scores

[rpb=.49, p <.001]. However, there was no significant association between reported drink

driving and sensation seeking [rpb = .01, p >.05]. Similarly, there was a significant (albeit

weak) positive correlation between self-reported speeding and sensation seeking [rpb =

.12, p <.05]. However, there was no significant association between seat belt wearing and

sensation seeking [r = .01, p >.05]. Hence, the data provide some support for the

influence of alcohol misuse on drink driving behaviour. However, the influence of

sensation seeking (at least as it was measured in this study) on risky driving behaviour is

at best weak.

Significant differences will be found between the unlicensed driver types in terms of

their socio-demographic characteristics (H10)

Strong support was found for this hypothesis. The results clearly suggest that

unlicensed drivers should not be viewed as a homogeneous group. Significant differences

were found between the unlicensed driver types in terms of age, education level, prior

criminal convictions, prior convictions for unlicensed driving, and whether they were

aware of being unlicensed. In particular, the higher levels of prior criminal convictions

among the disqualified, not currently licensed and never licensed drivers, combined with

the self-reported drinking and drink driving data (discussed below), suggests that

unlicensed driving may be associated with a more general pattern of deviance and risk-

taking among these offenders. This conclusion is consistent with the findings from Study

One, which indicated that the never licensed and disqualified/suspended drivers had the


highest risk of being involved in a crash and for those crashes to be more severe (see

section 4.4.5).

Interestingly, the findings from this study tend to suggest that the suspended drivers

are less likely to have a history of deviant behaviour than the disqualified drivers. This

suggests that the grouping of these two types of drivers in the crash statistics may have

obscured some important differences between the two. This conclusion is supported by

other findings discussed below.

The disqualified, suspended and never licensed drivers will report higher levels of

sensation seeking than the other offenders (H11)

This hypothesis was not supported by the data. No significant differences were

found among the different types of unlicensed drivers in terms of their sensation seeking

scores. There was some evidence that the mean score for the unlicensed drivers was

higher than what would be expected in a general sample of drivers. This suggests that

unlicensed drivers as a whole may be typically high sensation seekers.


alcohol misuse than the other offenders (H12)

Partial support was obtained for this hypothesis. A significant difference was found

between the offenders on the AUDIT test. However, considerably higher mean scores

were obtained for the not currently licensed [M = 13.4], never licensed [M = 12.7] and

disqualified [M = 11.7] drivers than the suspended [M = 8.6] drivers. In addition, a post-

hoc test indicated that only the first three groups of drivers differed significantly from the

expired drivers [M = 6.5]. Hence, the hypothesis is supported for the disqualified, not

currently licensed and never licensed drivers, but not for the suspended drivers.


drink-driving, speeding and failure to wear a seat belt than the other offenders

(H13)

Minimal support was obtained for this hypothesis. While unlicensed drivers report

engaging in higher levels of risky driving than general drivers, there appear to be minimal

differences between offenders. No significant differences were found between the

participants in terms of either self-reported speeding or seat belt use. However, in the

case of drink driving, the differences were approaching significance (p=.06), with the

disqualified (37.5%), not currently licensed (36.8%) and never licensed (29.2%) drivers

the most likely to admit to driving when they thought they may have been over the limit.


This latter result is consistent with a pattern of deviant behaviour among the

disqualified, not currently licensed and never licensed drivers, evident in their prior

criminal histories and self-reported drinking behaviour. However, this pattern of

behaviour does not appear to hold for the suspended drivers. Therefore, as already noted,

this suggests that the grouping of the suspended drivers with the disqualified drivers in

the crash statistics obscures important differences between these two types of offenders.

The perceived risk of apprehension for unlicensed driving will be lower than that

for other illegal road user behaviours, such as drink driving and speeding (H14)

This hypothesis was largely supported by the data. The perceived risk of

apprehension for unlicensed driving was significantly lower than it was for being random

breath tested or for being caught speeding by either a speed camera or radar. This finding

lends support to a deterrence theory explanation of unlicensed driving. Consistent with

classical deterrence theory, the majority of the offenders considered that they were

unlikely to be caught driving unlicensed, prior to being detected. However, following

detection and punishment the offenders reappraised (upwards) their perceived risk of

apprehension.

However, it is interesting to note that the perceived risk of apprehension for

unlicensed driving was significantly higher than that for being involved in a crash or

being caught for not wearing a seat belt. The first of these contrary findings is consistent

with other research that indicates that drivers generally have a low perceived risk of

crashing (Job, 1999). In the case of seat belt wearing, the findings may reflect the

relatively high seat belt wearing rates reported by the offenders (even though these rates

appeared to be lower than the general driving population).

Instances of punishment avoidance will be common among unlicensed driving

offenders (H15)

The data suggest that the experience of punishment avoidance is relatively common

among unlicensed driving offenders. Many offenders reported incidents where they were

either pulled over by the police and didn’t have their licence checked or were able to

evade detection by the authorities. The most common case of punishment avoidance

involved RBT. Almost one-third of the sample [31.4%] reported at least once instance

when they were pulled over for a random breath test but didn’t have their licence

checked. In addition, a small number of offenders cited cases where they were pulled

over for speeding or another offence and did not have their licence checked, were able to

evade a speed camera ticket or were involved in a traffic crash but were able to evade


detection. In total, 113 offenders (representing 36.6% of the sample) were able to evade

detection from the police on one or more occasions when they could otherwise have been

identified. Of these offenders, 67 (21.7% of sample) evaded detection on two or more

occasions.

These findings have particular relevance for Stafford & Warr’s (1993)

reconceptualisation of deterrence theory. They argue that punishment avoidance can

serve to encourage illegal behaviours in cases where detection is a relatively rare event.

Given the extent of the punishment avoidance reported in this study, this theoretical

construct warrants further attention in unlicensed driving research.

Continued driving after detection will be common among unlicensed driving

offenders (H16)

The data suggest that continued driving after detection is also relatively common

among unlicensed driving offenders. Almost one-third of the sample (30.5%) admitted

that they did continue to drive unlicensed after being detected, at least up until the time of

the court hearing (when they were interviewed). Among the offenders who admitted

continuing to drive, the process of detection at least appeared to temper the frequency of

their driving somewhat, resulting in a reduction in the mean number of trips per week

from 18.3 to 16.1. Nonetheless, the extent of this behaviour raises serious concerns about

the effectiveness of current processes used to deter unlicensed driving.

5.5.3 Implications for theory

At a theoretical level, the results of this study both reinforce and extend the findings

of Study One. Most importantly, they provide further evidence that unlicensed drivers do

not represent a homogeneous group. A range of differences were found among the

participants in terms of their psychosocial characteristics and self- reported on-road

behaviour. Similar to Study One, a particular cluster of participants emerged as a more

deviant group of offenders: the disqualified, not currently licensed and never licensed

drivers. These offenders reported higher levels of prior criminal offending, alcohol

misuse and self-reported drink driving. In contrast to Study One (where the suspended

drivers were grouped with the disqualified drivers), the suspended drivers did not appear

to share the same characteristics as this more deviant group. These results confirm that, in

order to be robust, a theoretical model of unlicensed driving would need to account for

the behaviour among a wide range of offenders. In particular, it would need to account


for a wide range of potential motives for the behaviour, some of which may be more

deviant than others.

In addition, the findings of the study provide further insights into the relevance of

the various theoretical perspectives reviewed in Chapter 3. Firstly, while some evidence

was obtained confirming a link between sensation seeking and self-reported speeding, a

relatively strong relationship was found between alcohol misuse and reported drink

driving. This raises the possibility that alcohol misuse exerts a major influence on the

driving behaviour of unlicensed drivers, particularly those who are disqualified, not

currently licensed or never licensed. Secondly, the results of the study suggest that

deterrence theory may represent a useful means of explaining unlicensed driving.

Consistent with classical deterrence theory, the participants’ perceived risk of

apprehension for unlicensed driving was relatively low (particularly when compared with

drink driving and speeding) and their knowledge of penalties was poor. Furthermore, the

relatively high incidence of punishment avoidance found in the study has important

implications for Stafford and Warr’s (1993) reconceptualisation of deterrence theory.

Based on their theory, it would be hypothesised that the experience of evading detection

while driving unlicensed would play a major role in encouraging the behaviour. These

theoretical issues form the major focus of Study Three, reported in the next chapter.

5.5.4 Implications for road safety

5.5.4.1 The extent and nature of unlicensed driving

The results of this study highlight the long periods of time during which offenders

can remain undetected. Among the first offenders in the sample, the mean reported length

of time driving unlicensed was 0.88 years, while for the repeat offenders it was 5.13

years. In both cases, there were individuals who reported driving for a considerable

period of time before being detected. For example, 25 offenders (8.1%) reported driving

unlicensed for 10 or more years.

More encouragingly, other data suggest that many of the offenders restricted the

amount of driving they undertook while unlicensed. Firstly, the overall mean number of

trips reported per week was 14.0. While this represents at least one return trip per

offender each day of the week, 31 offenders (10.0%) reported driving less frequently than

once a week. Secondly, 48.3% of the sample reported that they limited their driving while

unlicensed. When the offenders who claimed that they were unaware of being unlicensed

were excluded, this percentage increased to 56.3%. The offenders reported a variety of

strategies to reduce their driving exposure, and hence their chances of detection,


including restricting their overall amount of driving, driving on back streets/avoiding

main roads and driving during daylight hours only.

As noted in section 2.3.1, there is a common assumption made in the literature that

unlicensed drivers drive in a more cautious manner to avoid detection (Hurst, 1980;

Williams et al, 1984; Ross & Conzales, 1988; Mirrlees-Black, 1993, Job et al, 1994).

While the evidence from this study tends to confirm that many offenders reduce their

overall driving exposure in order to avoid detection, it is unclear whether this results in

more cautious driving. While the offenders reported relatively high levels of care in

obeying the road rules, they also reported more frequent drink driving, speeding and

failure to wear a seat belt than a sample of general drivers responding to the ATSB’s

(2000) Community Attitudes to Road Safety Survey. The somewhat paradoxical nature of

these findings is further discussed in section 7.2.2.

It is also interesting that almost two-thirds (62.5%) of the sample were detected

driving a vehicle that they owned. Even among the never licensed drivers, 32.0% were

detected driving their own vehicle. Therefore, while many offenders appeared to limit the

amount of driving they undertook, most had easy access to a vehicle. As will be noted

later, this highlights the potential value of vehicle-based sanctions to reduce the level of

unlicensed driving.

5.5.4.2 The effectiveness of current administrative, enforcement and punishment

processes

The majority of the offenders indicated that it was relatively unlikely that they

would drive unlicensed in the future. Nonetheless, the findings have identified a range of

shortcomings in the operation of current policies and practices relating to the

management of unlicensed driving.

Administrative processes

A number of potential problems with current administrative processes were

identified in the study. Over one-third (36%) of offenders claimed that they were

unaware, or at least unsure, of being unlicensed at the time they were detected. Not

surprisingly, this was more common among the expired (53.8%) and suspended (40.4%)

drivers. While this result may in part reflect the tendency of partic ipants to provide more

socially acceptable responses, it raises questions about the effectiveness of current

methods used to inform drivers about the expiry and cancellation of licences.

Almost half of the sample (49%) reported that they still had their photographic

licence when driving unlicensed. While this is not surprising among the expired drivers, it


remains a concern that many suspended, disqualified and not currently licensed drivers

still had their photographic licence. In the case of the disqualified and not currently

licensed drivers, it suggests that the processes for surrendering licences at court may not

always be observed. In the case of the suspended drivers, the results may in part reflect

the change in policy introduced in Queensland in December 2002, that no longer required

drivers who have their licence cancelled for accumulation of demerit points to surrender

their licences. This issue is further discussed in section 7.4.3.1.

Enforcement processes

The results tend to suggest that probability of detection for unlicensed driving in

Queensland is relatively low. The most common reason for the participants being

detected was due to them either committing a traffic offence or being involved in a crash

(37.9%). While 22% of offenders were detected at RBT operations, many of these appear

to have been initially detected for drink driving. The remaining offenders were detected

through either a targeted check (26.9%) or for reasons unclear to the participants (13.3%).

Therefore, while many offenders were detected through proactive policing initiatives

(such as RBT or targeted checks), many others only came to the attention of the police

because of illegal behaviour.

Moreover, many of the offenders reported incidents where they were either pulled

over by the police and didn’t have their licence checked or were able to avoid the matter

coming to the attention of the authorities. In total, 113 offenders (representing 36.6% of

the sample) were able to evade detection from the police on one or more occasions when

they could otherwise have been identified.

The low probability of detection for unlicensed driving was also reflected in the

behaviour and perceptions of the survey participants. As already noted, many offenders

had been driving unlicensed for quite long periods of time. Furthermore, almost one-third

(30.5%) of the sample admitted that they continued to drive illegally after they were

detected. In terms of perceptions, the perceived risk of apprehension for unlicensed

driving (prior to detection) among the participants was significantly lower than that for

being random breath tested or being caught for speeding by either a speed camera or

radar.

Punishment processes

The majority of the participants in the sample were either convicted of the offence

on the day they were interviewed (86.4%) or had the matter adjourned (12.0%). Only four


participants had no conviction recorded, while one was acquitted. In this respect, the

penalties for unlicensed driving were applied with a very high degree of certainty.

However, other evidence suggests that the participants were not particularly

concerned about the punishment associated with unlicensed driving. Firstly, only 42

(14.0%) of the participants reported that they knew what the penalties for unlicensed

driving were prior to being detected. Secondly, while the participants rated the certainty

(M = 5.4 on a seven-point scale) and swiftness (M = 5.2) of current punishment processes

as reasonably high, this was less so for the severity of penalties (M = 4.6). This finding is

somewhat surprising, given that the majority of offenders were interviewed shortly after

being sentenced in court.

5.5.5 Future directions for research

This study has highlighted a number of important issues requiring further research.

Most importantly, the main motives or factors contributing to unlicensed driving remain

unclear. A range of psychosocial factors have been highlighted in this study that appear to

vary across offenders and may directly, or indirectly, influence their behaviour. These

include:

§ socio-demographic characteristics like age, gender, education, prior criminal

history and the need to drive for work;

§ psychological characteristics like sensation seeking and alcohol use;

§ perceptions toward traffic law enforcement and punishment processes; and

§ experiences of punishment and punishment avoidance for unlicensed driving.

Further research is required to ascertain which of these factors directly contribute to

the decision to drive unlicensed and influence the extent of the behaviour. Ideally, this

research should explore the utility of relevant theoretical perspectives. This is the main

focus of Study Three.

This study has also highlighted a number of research priorities beyond the scope of

the current research program. In particular, there is a need for further research into the

effectiveness of current approaches used to manage unlicensed driving offenders. This

study has highlighted a range of apparent shortcomings in current administrative,

enforcement and punishment processes. Further research is required into the extent of

these problems and strategies to discourage or at least control unlicensed driving. The

main research priorities in this area are discussed in section 7.6.


5.6 Chapter summary

This study was primarily designed to examine the psychosocial characteristics and

self-reported behaviour of unlicensed drivers. In doing so, it explored many similar

themes to Study One, but with a sample of non crash-involved offenders. Compared with

previous surveys of offenders, the 62.4% response rate achieved in this study represents a

major strength of the research. Nonetheless, it is important to acknowledge the limitations

of the sample. Firstly, it was drawn from a metropolitan setting with a bias toward

offenders detected in an inner city and suburban area. Secondly, it is unclear to what

extent the sample is representative of unlicensed drivers in general, particularly those

who have never been detected by the police. Within these constraints, the findings of the

study have important implications for understanding and countering unlicensed driving.

The results reinforced concerns about the on-road behaviour of unlicensed drivers.

Almost one quarter of all the offenders reported driving unlicensed when they thought

they might have been over the alcohol limit. Similarly, 25% reported exceeding the speed

limit by 10 km/h or more on most or all occasions, while 15% admitted that they didn’t

always wear their seat belt. In addition, the results provided further evidence that

unlicensed drivers should not be viewed as a homogeneous group. Significant differences

were found between the offender types in terms of their socio-demographic

characteristics (age, education level, prior criminal convictions); driving history (prior

convictions for unlicensed driving and other traffic offences); whether they were aware of

being unlicensed; the degree to which they limited their driving while unlicensed; and

their drink driving behaviour. In particular, a more deviant sub-group of offenders was

identified, that included the disqualified, not currently licensed and never licensed

drivers, who reported higher levels of prior criminal offending, alcohol misuse and self-

reported drink driving. The results of the study also highlight the shortcomings of

existing police enforcement practices. Almost one-third of the sample reported that they

continued to drive unlicensed after being detected by the police (up until the time of the

court hearing), while many offenders reported experiences of punishment avoidance.

While this study has provided further insights into unlicensed drivers and their

behaviour, a major issue remains unexplored. There is a need to identify the main factors

that contribute to or motivate unlicensed driving, in order to develop more effective

countermeasures to the behaviour. This study has provided a good foundation for such

research by examining the relevance of various theoretical perspectives. While some of

the results were mixed, links were found between the behaviour of the offenders and both

sensation seeking and alcohol misuse. In addition, support was provided for a deterrence


theory based explanation of unlicensed driving. Together, these results suggest that the

theoretical perspectives reviewed in Chapter 3 represent a useful framework for

examining the factors contributing to unlicensed driving. An examination of these factors

forms the main focus of Study Three, reported in the next chapter.


Chapter Six: Factors contributing to unlicensed driving

6.1 Introductory comments ....................................................................................... 165

6.2 Study aims and hypotheses ................................................................................. 166

6.3 Method ................................................................................................................ 168

6.3.1 Overview of method ..................................................................................... 168

6.3.2 Derivation of independent and dependent variables ................................... 168

6.3.2.1 Independent variables ....................................................................... 168

6.3.2.2 Dependent variables .......................................................................... 170

6.3.3 Statistical analyses....................................................................................... 171

6.4 Results ................................................................................................................. 173

6.4.1 Socio-demographic factors.......................................................................... 173

6.4.2 Sensation seeking ........................................................................................ 175

6.4.3 Alcohol misuse ............................................................................................ 176

6.4.4 Environmental facilitating factors ............................................................... 177

6.4.5 Deterrence factors ....................................................................................... 179

6.4.5.1 Classical deterrence variables ........................................................... 179

6.4.5.2 Expanded deterrence variables .......................................................... 181

6.4.5.3 Comparison of classical and expanded deterrence perspectives ....... 182

6.4.5.4 Summary of deterrence perspectives................................................. 184

6.4.6 Social learning factors ................................................................................. 185

6.4.6.1 Imitation........................................................................................... 185

6.4.6.2 Differential association .................................................................... 187

6.4.6.3 Personal attitudes.............................................................................. 187

6.4.6.4 Differential reinforcement ................................................................ 188

6.4.6.5 Predictive role of social learning variables ...................................... 190

6.4.7 Summary of contributing factors................................................................. 191

6.4.7.1 Comparison of the different theoretical perspectives ..................... 191

6.4.7.2 The effectiveness of deterrence and social learning theories in explaining deviant behaviour ......................................................... 193

6.4.7.3 Extending the social learning explanation of unlicensed driving ...........................................................................................

195

6.5 Discussion........................................................................................................... 200


6.5.1 Study limitations ......................................................................................... 200

6.5.2 Support for study hypotheses ...................................................................... 201

6.5.3 Theoretical implications and directions for future research........................ 209

6.5.3.1 Implications for deterrence theory ................................................. 209

6.5.3.2 Implications for social learning theory........................................... 211

6.5.4 Countermeasure implications ...................................................................... 215

6.6 Chapter summary................................................................................................ 216



This chapter documents the third study undertaken as part of this research. The

primary role of this study was to examine the personal, social and environmental factors

contributing to unlicensed driving behaviour. In doing so it explores in more depth many

of the issues shown to be associated with unlicensed driving in the two previous studies.

A secondary aim of the study was to compare the predictive value of a number of

different perspectives commonly used to explain illegal driving behaviour, namely

sensation seeking, alcohol misuse, deterrence theory and social learning theory.

Consequently, the focus of this study is more theoretical in orientation than either of the

two previous studies.

The research builds on Study Two by using the cross-sectional survey data

collected from offenders at the Brisbane Central Magistrates Court. It examines three

aspects of the offenders’ behaviour, in order to obtain a better insight into the factors that

motivate unlicensed driving. The three particular aspects that were selected as the

dependent variables in the study were: the frequency of unlicensed driving prior to

detection; whether the participants continued to drive unlicensed after they were detected;

and their intention to drive unlicensed in the future. (A rationale for the selection of these

three as dependent variables is provided in 6.3.2.2.)

The hypotheses for this Study were informed by the previous two studies and the

empirical and theoretical literature reviewed in Chapters 2 and 3. The hypotheses fell into

three groups, largely reflecting Bandura’s (1977) principle of reciprocal determinism (see

section 3.3.1). The first group examined the influence of person-related factors on

unlicensed driving behaviour, including socio-demographic factors, sensation seeking

and alcohol misuse. The second group of hypotheses examined the role of certain

facilitating factors within the environment of the offenders, which were identified in

Study 2, including access to motor vehicles and possession of a photographic licence. The

final group of hypotheses examined the influence of various social factors on unlicensed

driving behaviour, such as exposure to formal (legal) and informal (social) sanctions,

behavioural models and social reinforcement. In particular, these hypotheses examined

the relative capacity of deterrence theory and social learning theory to explain unlicensed

driving behaviour.



As noted above, the primary aim of this study was to examine the last research

question identified in section 2.7, namely: What are the personal, social and

environmental factors contributing to unlicensed driving behaviour? The hypotheses that

were used to guide this process are identified below, along with a brief rationale for each.

The role of person-related factors

H17 Among unlicensed driving offenders, the behaviour will be more extensive among

those who are male and younger.

This hypothesis was based on the crash data reviewed in Study 1 (see section

4.4.2) that indicated that there was an over-representation of males and young

drivers in the crashes involving unlicensed drivers.

H18 Unlicensed driving will be positively associated with prior criminal offending.

In Study Two, it was found that 38.8% of the participants had a prior

conviction for a criminal offence. This and the evidence relating to risk-taking

reviewed in Chapter 3 suggests that a link exists between unlicensed driving and

more deviant behaviour.

H19 Unlicensed driving will be positively associated with the need to drive for work

purposes.

The need to drive for work has previously been identified as one of the main

factors contributing to unlicensed driving (eg. Robinson, 1977; Ross & Conzales,

1988; Mirlees-Black, 1993; Job et al, 1994). In addition, in Study Two it was found

that a lot of the trips reported by the participants were for work-related purposes.

This suggests that the need to drive for work purposes encourages unlicensed

driving behaviour.

H20 Unlicensed driving will be positively associated with higher levels of sensation

seeking.

A significant, but weak, positive relationship was found in Study Two

between sensation seeking and self-reported speeding. Given other evidence linking

sensation seeking with risky driving, it is possible that unlicensed driving may also

be influenced by a propensity for sensation seeking.


H21 Unlicensed driving will be positively associated with alcohol misuse.

As reviewed in Chapter 3, there is considerable evidence that alcohol misuse

is a major influence on the behaviour of hard-core drink drivers, many of whom

continue to drive without a licence after disqualification. In addition, in Study Two

a relatively strong, positive relationship was found between alcohol misuse and

self-reported drink driving. This raises the possibility that alcohol misuse may be a

major catalyst for unlicensed driving, by facilitating access to social settings in

which alcohol is consumed.

The role of environmental facilitating factors

H22 Unlicensed driving will be positively associated with the availability of motor

vehicles and continued possession of a photographic licence.

Previous research has suggested that access to a vehicle is an important factor

contributing to the prevalence of unlicensed driving (Ross and Gonzales, 1988;

Mirrlees-Black; 1993; Williamson, 1996). The results of Study Two also confirmed

the potential role of vehicle ownership and possession of a photo licence as factors

that may facilitate unlicensed driving.

The role of social factors

H23 An expanded theory of deterrence, incorporating the constructs of punishment

avoidance and vicarious learning, will better predict unlicensed driving than

classical deterrence theory.

H24 Social learning theory will better predict unlicensed driving than either the

classical deterrence or expanded deterrence theories.

H25 Both expanded deterrence theory and social learning theory will better predict

unlicensed driving among more deviant offenders.

These three hypotheses were based on the theoretical arguments presented in

Chapter 3. The first reflects the proposition by Stafford and Warr (1993) that their

reconceptualisation of deterrence theory, incorporating punishment avoidance and

vicarious learning processes, represents a better predictor of illegal behaviour than

classical deterrence theory. The second is based on the proposition of Akers (1977;

1990) that deterrence theory can be subsumed within a social learning framework.

As such, social learning theory represents a more comprehensive perspective for

explaining illegal and deviant behaviours. Finally, given that both deterrence theory

and Akers’ social learning theory were primarily developed to explain deviant


behaviour, it is hypothesised that both theories will better predict the behaviour of

the more deviant offenders in the sample.

6.3 Method

6.3.1 Overview of method

This study involves a further analysis of the cross-sectional survey data collected

from unlicensed driving offenders at the Brisbane Central Magistrates Court. The

participants and the procedure used to collect the data are fully described in section 5.3.

The study used the same questionnaire as for Study Two (see Append ix D). The

particular items from the questionnaire that were used in this study are detailed below.

6.3.2 Derivation of study variables

6.3.2.1 Independent variables

This study was informed by the review of relevant theoretical perspectives reported

in Chapter 3. In addition to operationalising the main perspectives under consideration

(deterrence theory, social learning theory, sensation seeking, alcohol misuse), a range of

socio-demographic and facilitating factors were measured. The independent variables

(and related items) used in the study are detailed below.

Socio-demographic variables

The socio-demographic variables included:

§ gender (Q. 1);

§ age (Q. 2);

§ marital status (Q. 3);

§ level of education attained (Q. 4);

§ employment status at time of court appearance (Q. 5);

§ whether the participant needed to drive for work when unlicensed (Q. 29);

§ income (Q. 1); and

§ prior conviction for a criminal offence (Q. 61).

Sensation seeking and alcohol misuse

These variables were based on the two scales used to measure sensation seeking

and alcohol misuse in Study Two (see section 5.3.3.3 for a description).


Environmental facilitating factors

A number of variables were operationalised to measure the effect of certain factors

that may facilitate unlicensed driving including:

§ awareness of being unlicensed (Q. 11);

§ access to a motor vehicle (Q. 30);

§ ownership of a motor vehicle (Q. 14);

§ possession of an old (invalid) photographic licence (Q. 32); and

§ the availability of public transport (Q. 53h).

Deterrence variables

The deterrence variables measured in the questionnaire drew on both classical

deterrence theory and Stafford and Warr’s (1993) reconceptualisation of deterrence

theory. The two models used to guide the selection of variables are described in section

3.2.4. The variables operationalised in the study were:

§ perceived risk of apprehension prior to detection (Q. 33);

§ perceived risk of apprehension after detection (Q. 34);

§ knowledge of the fines for unlicensed/disqualified driving (Q. 15);

§ perceived severity of punishment for unlicensed/disqualified driving (Q. 52a);

§ perceived certainty of punishment for unlicensed/disqualified driving (Q. 52i);

§ perceived swiftness of punishment for unlicensed/disqualified driving (Q. 52o);

§ prior conviction for unlicensed driving (Q. 21);

§ direct exposure to enforcement (Q. 44 – 48);

§ direct exposure to punishment avoidance (Q. 44-48);

§ vicarious exposure to enforcement (Q. 49-50); and

§ vicarious exposure to punishment avoidance (Q. 51).

Social learning variables

The social learning variables operationalised in the questionnaire were based on the

theoretical model developed by Akers (1977; 1990; 1994) and drew on studies conducted

by Akers et al (1979), Krohn et al (1985), DiBlasio and Benda (1990), Akers and Lee

(1996) and Skinner and Fream (1997) (see section 3.3). The specific social learning

model of unlicensed driving that was used to inform the selection and operationalisation

of variables is described in section 3.3.4. The pilot testing reported in section 5.3.2 also

informed the particular items used in the questionnaire. The relevant variables and related

items were:


§ imitation (Q. 49–50);

§ personal attitudes (definitions) to (i) unlicensed driving, and (ii) alternative

behaviours (various items in Q. 52 – see Appendix F for specific items);

§ differential association (various items in Q. 49, 50 and 52 – see Appendix F for

specific items); and

§ differential reinforcement arising from (i) anticipated social and non-social rewards

for unlicensed driving, and (ii) anticipated punishments for unlicensed driving,

including informal social sanctions as well as formal legal sanctions (various items

in Q. 52 – see Appendix F for specific items).

6.3.2.2 Dependent variables

As noted in the introduction to this chapter, the nature of the dataset used in this

study did not make it possible to directly examine the factors contributing to the decision

to drive unlicensed or not. Consequently, it was necessary to select for investigation a

variable, or group of variables, that reflected important aspects or dimensions of

unlicensed driving behaviour. Three variables were selected to act as dependent variables

for this purpose. These variables were drawn from Study Two and are:

§ the frequency of unlicensed driving trips per week (see section 5.4.3.2);

§ whether the offenders continued to drive unlicensed after (their most recent)

detection (see section 5.4.4.3); and

§ intention to drive unlicensed in the future (see section 5.4.4.6).

The frequency of unlicensed driving was selected to reflect the extent of the

behaviour. This variable was considered a more valid and reliable measure of recent

behaviour than the total length of time driving unlicensed. This was because total length

of time did not necessarily represent continuous periods of time and would be subject to

more recall biases and the influence of random detection practices.

The continued driving after detection variable was selected to reflect the offender’s

commitment to (or reliance on) the behaviour. In other words, it was assumed that those

drivers who continued to drive after being detected (and prior to the court hearing) were

highly motivated to knowingly engage in the behaviour.

The intention to drive unlicensed variable was selected to provide an insight into

the psychological processes underpinning the behaviour. A variety of social

psychological theories incorporate the concept of intentions as a key predictor of

behaviour (Fishbein et al, 1991). While intentions are far from a perfect predictor of

future behaviour, they indicate a preparedness to engage in the behaviour.


All three dependent variables also have important theoretical and practical

implications. The first two (the frequency of unlicensed driving and continued driving

after detection) are indicative of the extent of law breaking and should reflect the impact

of current enforcement practices on the behaviour. Among repeat offenders they may also

reflect the impact of past punishment processes. While the third dependent variable

(future intentions) is not directly indicative of behaviour, it should reflect the impact of

both current enforcement and punishment practices.

The three dependent variables were moderately correlated with one another (see

Table 6.1). The strongest correlation was between the continued driving after detection

variable and intention to drive unlicensed in the future [rpb=.45, p <.001]. Hence, it was

expected that there would be similarities in the results obtained for each dependent

variable, particularly for the latter two. (A full correlation matrix for all the dependent

and independent variables used in the study is provided in Appendix H).

Table 6.1

Bivariate correlations between dependent variables

Variables Frequency of

unlicensed driving1


Intention to drive unlicensed

in the future1

Frequency of unlicensed driving − .30*** .20**

Continued driving after detection − .45***

Intentions to drive unlicensed in the future −

1. Logarithmically transformed. * p < .05 ** p < .01 *** p < .001

6.3.3 Statistical analyses

The two main aims of the study were to identify the factors that predict unlicensed

driving behaviour and compare the predictive utility of different theoretical perspectives,

rather than explore the structural nature of these perspectives. Accordingly, it was

decided to utilise regression-based techniques to analyse the data, rather than structural

analytic techniques like path analysis. Multiple regression was used to examine the

prediction of the two continuous dependent variables ie. the frequency of unlicensed

driving and intention to drive unlicensed in the future. (As in Study Two, data collected

by Likert scale was treated as interval in nature to facilitate the use of parametric

methods.) However, logistic regression was used to examine the factors that predicted

continued driving after detection, since it was a dichotomous variable. In most cases, the


sets of predictor variables that were examined were independent of one another.

Accordingly, standard multiple regression and direct logistic regression were primarily

used to analyse each set of predictors. However, in the case of the two deterrence

perspectives, the classical deterrence variables represented a subset of the expanded

deterrence variables. As such, it was possible to use hierarchical/sequential regression

analyses to compare the predictive utility of the two sets of deterrence variables.

The rule of thumb that was used wherever possible to maintain the power of the

multiple regression analyses was to ensure that N ≥ 50 +8m (where m was the number of

IVs) when testing the multiple correlation and N ≥ 104 + 8m for testing individual

predictors. Tabachnick and Fidell (1996, p.132) report that this is appropriate for

detecting a medium effect-size with a significance level of .05 and 80% power (ie. ß =

.20). The two analyses that violated this requirement are identified in the relevant section

of the results (see section 6.4.7.2).

In the case of the logistic regressions, the Nagelkerke R2 was used to measure the

strength of association for each model and to act as an analogue to R2 in the multiple

regressions. While this statistic does not have the same variance interpretation as R2 in

linear regression, it is designed to approximate it (Tabachnick & Fidell, 2001, p.545).

The transformation and recoding of certain data was required to facilitate the

regression analyses. Two of the dependent variables featured positively skewed

distributions with univariate outliers. The variables in question were the frequency of

unlicensed driving and intention to drive unlicensed. To overcome these problems, both

variables were transformed using logarithmic transformations. In the case of the

independent variables, the original age and income categories were collapsed and dummy

coded. The age categories used were: 17 – 20; 21 – 25; 26 –39; and 40 or over. These

categories reflected the youthful nature of the sample and common age groupings used in

road safety analyses. The annual income categories used were: less than $10,000;

$11,000 - $30,000; and $31,000 or more. In addition, the marital status variable was

recoded into a dichotomous variable distinguishing between those participants who were

single (including those who were separated or divorced) and those who were married or

in a de facto relationship. Similarly, the education variable was recoded into a

dichotomous variable distinguishing between those participants who were educated to

Year 10 or less and those with a higher level of education.


Due to the number of linear and logistic regressions performed on each dependent

variable, a more stringent significance level (α = .01) was adopted to protect against

inflating the experiment-wise error rate. This more stringent significance level was used

for the testing of overall regression models and individual predictors, as well as for the

correlations. However, given the exploratory nature of the study, results that were

significant at α = .05 are also noted.

Reliability analyses were undertaken on the various scales used in the study using

Cronbach’s alpha. A summary of the scales and their Cronbach’s alpha is provided in

Appendix F.

6.4 Results

6.4.1 Socio-demographic factors

Table 6.2 summarises the bivariate correlations between the three dependent

variables and the socio-demographic variables. (A full correlation matrix is provided in

Appendix H). As can be seen, the only variables that were significantly correlated at the p

< .01 level were the frequency of unlicensed driving and the need to drive for work while

unlicensed. The positive nature of this relationship indicates that those participants who

needed to drive for work while unlicensed reported more frequent unlicensed driving. In

addition, there was a weak positive relationship (p < .05) between being employed at the

time of the court hearing and the reported frequency of unlicensed driving. The age of the

participants was also weakly (p < .05) related to both continued driving after detection

and intention to drive unlicensed in the future. The negative nature of this relationship

indicates that the younger participants were more likely to report driving after detection

and had a stronger intention to drive unlicensed in the future. Finally, there was a weak

correlation between marital status and intentions, with the single participants reporting a

stronger intention to drive unlicensed in the future than those who were married or in de

facto relationships.


Table 6.2

Bivariate correlations between dependent variables and socio-demographic variables

Dependent variables

Variable Frequency of unlicensed driving1



in the future1

Gender .02 -.01 .10

Age2 -.04 -.13* -.12*

Marital status (Single vs. married/defacto) -.04 .09 -.12*

Educational level (≤ Year 10 vs. > Year 10) .09 .04 -.08

Employed at time of court hearing .12* -.09 .06

Needed to drive for work when unlicensed .28*** .09 .07

Annual income2 .11 .06 .00

Prior criminal conviction .10 .02 .06

1. Logarithmically transformed. 2. Correlations calculated using midpoints of categories. * p < .05 ** p < .01 *** p < .001

To further examine the relationship between the socio-demographic variables and

the dependent variables a series of regression analyses were undertaken. Table I1 in

Appendix I reports the results of a standard multiple regression examining the influence

of the socio-demographic variables on the frequency of unlicensed driving. The overall

regression model was significant [F (11, 292) = 3.38, p < .001] accounting for 8% of the

adjusted variance [R2 = .11, AdjR2 = .08] in the frequency of unlicensed driving. Two of

the variables were significant predictors at the p < .01 level: needed to drive for work

while unlicensed [ß = .26; p < .001; sr2 = .06;] and prior criminal conviction [ß = .17; p <

.01; sr2 = .02]. In addition, educational level [ß = .13; p < .05; sr2 = .01] was significant

at a less stringent level. The results indicate that more frequent unlicensed driving was

positively associated with the need to drive for work while unlicensed and having a prior

criminal conviction and, possibly, with having attained a higher level of education (than

Grade 10).

Table I2 in Appendix I reports the results of a logistic regression examining the

influence of the socio-demographic variables on whether the participants continued to

drive unlicensed after detection. Although the overall model was not significant [χ2 (df11,

n=304) = 18.56, p > .05], it is worth noting that the two employment related variables


were significant at the p < .05 level ie. needed to drive for work while unlicensed [odds

ratio = 1.96, CI = 1.12 – 3.42] and employed at the time of court hearing [odds ratio =

.46, CI = .24 - .88]. Therefore, while being employed at the time of the court hearing was

associated with a lower likelihood of continued driving after detection, needing to drive

for work while unlicensed increased the likelihood of continued driving by almost

double. These results suggest that being employed, except in those cases where it

necessitates driving, may tend to act as a protective factor against continued offending.

For example, employed offenders who don’t necessarily need to drive for work may be

concerned about the long-term impact of further penalties on their capacity to travel to

and from work.

Table I3 in Appendix I reports the results of a standard multiple regression

conducted to examine the influence of the socio-demographic variables on intentions to

drive unlicensed in the future. The overall regression model was only significant at the

less stringent level [F (11, 293) = 1.98, p < .05] and accounted for 3% of the adjusted

variance [R2 = .07, AdjR2 = .03] in the transformed intentions variable. The only

significant predictor was marital status [ß = -.15; p < .01; sr2 = .02]. The negative ß

indicates that single people reported stronger intentions to drive unlicensed in the future.

6.4.2 Sensation seeking

The Thrill and Adventure Seeking (TAS) subscale of the Sensation Seeking Scale

(SSS) was used to measure the participants’ propensity for sensation seeking (see section

5.2.3.3). The Cronbach’s alpha for the ten- item scale was .71 (see F1 in Appendix F). As

shown in Table 6.3, the sensation seeking score was not significantly correlated with any

of the dependent variables.

Table 6.3

Bivariate correlations between dependent variables and sensation seeking score

Dependent variables




in the future1

Sensation seeking .08 .03 .02



There are a number of possible explanations for the apparent lack of association

between sensation seeking and the various measures of unlicensed driving behaviour. The

first possibility is that unlicensed drivers are typically high sensation seekers and, hence,

there was a lack of variability in the data. This is consistent with the evidence presented

in Study Two (see section 5.4.1.6). An alternative explanation is that, although sensation

seeking is associated with risk-taking per se, it has little direct impact on unlicensed

driving. This is also consistent with the evidence from Study Two, since sensation

seeking was found to be significantly correlated with only self-reported speeding and

prior criminal conviction. Therefore, on balance it appears that sensation seeking may

contribute to certain risky behaviours that can lead to a person losing their licence (eg.

speeding), but not directly to unlicensed driving. Consequently, sensation seeking was

not used in any of the subsequent regression analyses.

6.4.3 Alcohol misuse

Alcohol misuse was measured using the Alcohol Use Disorders Identification Test

(AUDIT) (see section 5.2.3.3). The Cronbach’s alpha for this ten- item test was .80 (see

F2 in Appendix F). As shown in Table 6.4, the AUDIT score was not significantly

correlated with any of the dependent variables. However, in Study Two the AUDIT had

been found to be significantly correlated with both self- reported drink driving and the

likelihood of having a prior conviction for unlicensed driving. Hence, similar to sensation

seeking, although alcohol misuse may have contributed to some of the respondents losing

their licence in the first place, it appears to have little direct influence on the nature of

unlicensed driving behaviour. Consequently, the AUDIT score was not used in any

subsequent regression analyses.

Table 6.4

Bivariate correlations between dependent variables and AUDIT score

Dependent variables




in the future1

AUDIT score .07 .03 .04



6.4.4 Environmental facilitating factors

Table 6.5 reports the bivariate correlations between the dependent variables and a

number of potential environmental facilitating factors, including whether the offenders

were aware of being unlicensed, whether they owned a vehicle or at least were able to

access one while unlicensed; and whether they still had their photographic licence. While

awareness of being unlicensed could have been conceptualised as a person-related factor,

it appears to reflect the nature of the processes used to advise offenders of their licence

loss. As noted in Study Two, over one-third of offenders claimed that they were unaware

of being unlicensed at the time they were detected (see section 5.4.4.1). Consequently, it

was conceptualised as an environmental factor reflecting an apparent failure of current

licence loss notification processes.

Table 6.5

Bivariate correlations between dependent variables and environmental facilitating

factors

Dependent variables




in the future1

Unaware of being unlicensed .17** .10 -.24***

Able to access vehicle while unlicensed .17** .11 .12*

Owned a vehicle .15** .03 -.07

Still had photographic licence (all participants)

.22*** -.09 -.17**

Still had photographic licence (only participants who were aware of being unlicensed)

.18* -.02 -.02


While a number of the coefficients in Table 6.5 are significant, they are relatively

weak with only two exceeding .20. There was a significant correlation between

awareness of being unlicensed and the frequency of the behaviour. This indicates that

those drivers who were unaware of being unlicensed drove more frequently. Not

surprisingly, there was no significant correlation between awareness of being unlicensed

and continued driving after detection. However, the variable was negatively correlated

with intention to drive unlicensed, indicating that those participants who were unaware of


being unlicensed when detected were less likely to intend to drive illegally in the future.

In other words, those offenders who had knowingly driven unlicensed in the past were

more likely to report an intention to do so again.

Access to a vehicle was significantly related to a higher frequency of unlicensed

driving and, to a lesser extent, stronger intentions to drive unlicensed in the future, but

was not related to continued driving after detection. The participants were also asked who

owned the vehicle(s) they drove when unlicensed. Owning a vehicle was significantly

related to the frequency of unlicensed driving but not to the two other dependent

variables.

As noted in Study Two (see section 5.4.4.2), almost half the offenders in the sample

reported that they were still in possession of a photographic licence when they were

driving unlicensed. Interestingly, there was a significant positive correlation between

possession of a photographic licence and the frequency of unlicensed driving but a

negative correlation between it and intention to drive unlicensed in the future. These

results tend to suggest that some of the offenders who still had their photographic licence

were originally unaware that they were unlicensed. As such, while they may have

unknowingly driven frequently when unlicensed, they did not intend to do so in the

future. This interpretation is supported by the significant association found between being

in possession of a photographic licence and being unaware of being unlicensed [φ = .49,

p < .001]. To further examine the issue, an analysis was undertaken excluding the

participants who were unaware/unsure whether they were unlicensed (shown separately

in Table 6.5). Even among the offenders who were aware of being unlicensed, there was

still a weak positive correlation between the possession of a photographic licence and

more frequent unlicensed driving [rpb = .18, p < .05]. However, the correlation between

possession of a licence and future intentions was no longer significant after the exclusion

of those who were unaware/unsure whether they were unlicensed.

To further explore the relationships between the facilitating factors and the

dependent variables, a series of regression analyses were undertaken. Table I4 in

Appendix I reports the results of a standard multiple regression conducted to examine the

influence of the facilitating factors on the frequency of unlicensed driving. The overall

regression model was significant [F (4, 294) = 6.24, p < .001] accounting for 7% of the

adjusted variance [R2 = .08, AdjR2 = .07] in the frequency of unlicensed driving. None of

the predictors were significant at the p < .01 level. However, still had a photographic

licence [ß = .16; p < .05; sr2 = .02] was significant at a less stringent level.



influence of the facilitating factors on whether the participants continued to drive

unlicensed after detection. Neither the overall model nor any of the variables were

significant [χ2 (df4, n=299) = 7.53, p > .05].

Table I6 reports the results of a standard multiple regression conducted examining

the influence of the facilitating factors on intentions to drive unlicensed in the future. The

overall regression model was significant [F (4, 294) = 7.36, p < .001] accounting for 8%

of the adjusted variance [R2 = .09, AdjR2 = .08] in the intentions variable. Two of the

variables were significant predictors: unaware of being unlicensed [ß = -.21; p < .01; sr2

= .03] and able to access a vehicle [ß = .18; p < .01; sr2 = .03].

6.4.5 Deterrence factors

The bivariate correlations between the dependent variables and the deterrence

variables are shown in Table 6.6. The first group of variables in the table are those drawn

from classical deterrence theory, as outlined in Chapter 3 (see section 3.2.1). The second

group of variables are those drawn from the expanded theory of deterrence, as proposed

by Stafford and Warr (1993) (see section 3.2.3).

6.4.5.1 Classical deterrence variables

As shown in Table 6.6, there was a weak negative correlation between the

participants’ perceived risk of apprehension (prior to detection) and the frequency of their

unlicensed driving [r = -.13, p < .05]. In other words, a lower perceived risk of

apprehension was associated with more frequent unlicensed driving. In addition, there

was a significant negative correlation between the participants’ perceived risk of

apprehension (after detection) and whether they continued to drive after detection [rpb = -

.17, p <.01] and their intention to drive unlicensed in the future [r = -.18, p <.01]. Once

again, the negative relationships indicate that a lower perceived risk of apprehension was

associated with continued driving after detection and a stronger intention to drive

unlicensed in the future.

Consistent with deterrence theory, the perceived severity, certainty and swiftness of

punishment measures were all negatively correlated with the dependent variables

(although some of the correlations were close to zero). In other words, the more that

participants perceived current penalties to be severe, certain and swift, the less they

engaged in unlicensed driving (before and after detection) and the less they intended to

do so in the future. However, none of these correlations were significant at p < .01.


Table 6.6

Bivariate correlations between dependent variables and deterrence variables

Dependent variables




in the future1

Classical deterrence variables

Perceived risk of apprehension (prior to detection) -.13* - -

Perceived risk of apprehension (after detection) - -.17** -.18**

Knew what the fine for unlicensed driving was, prior to court -.04 -.01 .06

Perceived severity of punishment -.07 -.06 -.01

Perceived certainty of punishment -.04 -.13* -.07

Perceived swiftness of punishment -.07 -.04 -.09

Prior conviction for unlicensed driving -.03 .04 .20***

Exposure to enforcement -.14* .02 -.05

Expanded deterrence variables

Punishment avoidance .31*** .09 .17**

Vicarious exposure to punishment .09 .20*** .18***

Vicarious exposure to punishment avoidance .02 .08 .14*


As reported in Study Two (section 5.4.1.4), over one third (39.2%) of the sample

reported having a prior conviction for unlicensed (or disqualified) driving. As shown in

Table 6.6, this variable was not significantly related to either the reported frequency of

unlicensed driving or whether the participants continued to drive after detection. However,

there was a significant positive correlation between prior conviction and the participant’s

intention to drive unlicensed in the future [rpb = .20, p <.001]; those participants who had a

prior conviction for unlicensed driving were more likely to report an intention to do so in

the future. Contrary to classical deterrence theory, this suggests that previous exposure to

punishment has had no deterrent impact on the intentions of the participants (see section

6.5.3.1 for further discussion of this issue).


As noted in Study Two, many of the offenders were exposed to some form of traffic

law enforcement during the time they were driving unlicensed (see section 5.4.4.4). In

many of these incidents the participants were detected. However, 113 (36.6% of total

sample) were able to evade detection from the police on one or more occasions when they

could otherwise have been identified. Two dichotomous variables were created to

measure these two contingencies: i) whether the participants were exposed to

enforcement, irrespective of whether they were detected (ie. exposure to enforcement);

and ii) whether the participants were successful in avoiding detection (ie. punishment

avoidance). The first of these variables is traditionally included within classical

deterrence theory. As can be seen in Table 6.6, exposure to enforcement was weakly

related to the frequency of unlicensed driving [rpb = -.14, p < .05]. The negative

relationship between this variable and both the frequency of unlicensed driving and

intention to drive unlicensed is consistent with deterrence theory, since it indicates that

exposure to enforcement discourages illegal behaviour.

6.4.5.2 Expanded deterrence variables

It should be noted that the punishment avoidance variable described above

represents a special case of the construct proposed by Stafford and Warr (1993). In

practice, all successful episodes of unlicensed driving represent instances of punishment

avoidance (see section 3.2.4). However, in order to distinguish this construct from the

frequency of unlicensed driving and continued driving after detection (since they were

being used as dependent variables in this study), it was necessary to focus on particular

episodes where offenders come into direct contact with the police (or other authorities)

but managed to evade detection.

As can be seen in Table 6.6, the punishment avoidance variable was positively

associated with both the frequency of unlicensed driving [rpb = .31, p < .001] and

intention to drive unlicensed in the future [rpb = .17, p < .01]. In other words, evading

detection was associated with more frequent unlicensed driving prior to detection and a

stronger intention to drive unlicensed in the future. However, the causal direction of the

relationship between punishment avoidance and frequency of unlicensed driving is

unclear. While Stafford and Warr’s (1993) reconceptualisation of deterrence theory

would suggest that evading detection encourages more frequent unlicensed driving, the

results may only indicate that those people who drive more frequently have more

opportunities to evade detection. However, other evidence supports the view that the

experience of punishment avoidance encourages driving. For example, the punishment


avoidance variable was significantly correlated (in the expected direction) with a number

of items reflecting the probability of detection, including:

§ “. . if you were to drive unlicensed/disqualified in the future, how likely do you think

your chances of getting caught are?” (Q.34) [rpb = -.11, p < .05];

§ “You were lucky not to have been caught earlier for driving unlicensed/disqualified”

(Q.52p) [rpb = .13, p < .05]; and

§ “The police generally check driver’s licences when they conduct RBT ” (Q.53a) [rpb =

-.41, p < .001].

Secondly, if the findings concerning punishment avoidance were merely a product

of increased driving exposure, a similar relationship should have been found between

unlicensed driving and exposure to enforcement. In other words, exposure to enforcement

should also been associated with more frequent unlicensed driving. However, this was

not found to be the case. Thirdly, the relationship between punishment avoidance and

frequency of unlicensed driving is consistent with the participants’ future intention to

drive unlicensed. In this respect, it is reasonable to conclude that the experience of

avoiding punishment for unlicensed driving directly contributes to a stronger intention to

drive unlicensed in the future.

The remaining two variables included in Table 6.6 are also based on the work of

Stafford and Warr (1993). The vicarious exposure to punishment was based on whether

the participants were aware of a family member or friend who had been convicted of

unlicensed driving in the past. As shown, this variable was not significantly associated

with the frequency of unlicensed driving. However, it was significantly related to whether

the participants continued to drive after detection [φ =.20, p < .001] and with intentions to

drive unlicensed [rpb = .19, p ≤ .001]. These latter two findings are inconsistent with

deterrence theory, which would suggest that vicarious exposure to punishment should act

to deter unlicensed driving. The measure of the participant’s vicarious exposure to

punishment avoidance was based on whether they knew an unlicensed driver who had

evaded detection (ie. had not had their licence checked at some time). Similar to direct

punishment avoidance, vicarious exposure to punishment avoidance was positively

(albeit weakly) associated with intention to drive unlicensed [rpb = .14, p <.05].

6.4.5.3 Comparison of classical and expanded deterrence perspectives

Two sets of regression analyses were undertaken to compare the relative predictive

utility of the classical and expanded deterrence theories. Table I7 –I9 in Appendix I


report the analyses using the classical deterrence theory variables alone. Tables I10- I12

report the results for the expanded deterrence theory variables.

A comparison of Tables I7 and I10 indicates that the expanded deterrence variables

were more effective in predicting the frequency of unlicensed driving than the classical

deterrence variables alone. The regression model using the classical deterrence variables

was not significant [F (7, 288) = 1.82, p > .05]. In contrast, the model using the expanded

deterrence variables was significant [F (10, 281) = 3.99, p < .001] and accounted for 9%

of the adjusted variance [R2 = .12, AdjR2 = .09]. The strongest predictor in the expanded

model was punishment avoidance [ß = .30; p < .001; sr2 = .07].

In addition, a hierarchical regression was undertaken to examine the predictive

utility of the three expanded deterrence variables, over and above the classical deterrence

variables (see Table I13 in Appendix I). This indicated that the additional variables

accounted for significantly more variance in the frequency of unlicensed driving, as

indicated by a significant change in R2 [F (3, 281) = 8.66, p < .001].

Tables I8 and I11 report the logistic regressions examining the influence of the

classical and expanded deterrence variables on whether the participants continued to

drive unlicensed after detection. The model using the classical deterrence variables was

significant, but at the less stringent level [χ2 (df7, n=298) = 15.00, p < .05]. The only

predictor that was significant at p < .01 was the perceived risk of apprehension [odds

ratio = 0.84], while the perceived certainty of punishment [odds ratio = 0.86] was

significant at p < .05. The odds ratios indicate that both of these variables were associated

with a lower likelihood of continued driving after detection. In contrast, the model using

the expanded deterrence variables [χ2 (df10, n=294) = 25.40, p < .01] was significant at

the more stringent level. In addition, the Nagelkerke R2 for the expanded model was

marginally higher than that using the classical deterrence variables [12% vs. 11%),

indicating a slightly stronger association between the expanded deterrence variables and

the dependent variables. While the perceived risk of apprehension remained a significant

predictor (at p < .05) in the expanded model, the perceived certainty of punishment was

no longer significant. Instead, one of the new variables, vicarious exposure to

punishment, was significant at p < .01 [odds ratio = 2.20]. As noted earlier, however, the

direction of this result is inconsistent with deterrence theory. Rather than acting to

discourage illegal behaviour, the vicarious exposure to punishment variable was

associated with a higher likelihood of continued driving after detection.


As before, a sequential/hierarchical logistic regression was undertaken to examine

the predictive utility of the three expanded deterrence variables, over and above the

classical deterrence variables (see Table I14 in Appendix I). This analysis indicated that

the three additional variables significantly added to the prediction of continued driving

after detection [χ2 (df3, n=294) = 11.94, p < .01].

Tables I9 and I12 report standard multiple regressions examining the influence of

the classical and expanded deterrence variables on intention to drive unlicensed in the

future. The model using the classical deterrence variables was significant [F (7, 291) =

4.56, p < .001] accounting for 8% [R2 = .10, AdjR2 = .08] of the variance in the intentions

variable. Two of the variables were significant predictors: prior conviction for unlicensed

driving [ß = .21; p < .001; sr2 = .04] and perceived risk of apprehension (after detection)

[ß = -.21; p < .001; sr2 = .03]. As expected, this latter variable was negatively associated

with future intentions ie. a higher perceived risk of apprehension was associated with

lower intentions to drive unlicensed. However, the prior conviction for unlicensed driving

variable was positively associated with future intentions. As noted earlier, this result is

contrary to classical deterrence theory, since it suggests that previous exposure to

punishment has no deterrent impact on the intentions of the participants.

Once again, the model using the expanded deterrence variables was a better

predictor [F (10, 284) = 4.42, p < .001] accounting for 10% [R2 = .14, AdjR2 = .10] of the

variance in the intentions variable. In this model, the punishment avoidance variable also

became a significant predictor, but at a less stringent level [ß = .13; p < .05; sr2 = .01].

The added predictive power of the expanded deterrence variables was confirmed by a

hierarchical regression analysis (see Table I15 Appendix I), which found that the addition

of these variables produced a significant change in R2 [F (3, 284) = 4.45, p < .01].

6.4.5.4 Summary of deterrence perspectives

In the case of all three dependent variables, the expanded deterrence variables

proved more effective in explaining unlicensed driving behaviour than the model using

the classical deterrence variables alone. This was confirmed by the hierarchical

regression analyses that indicated that the three expanded deterrence variables

significantly added to the prediction of the dependent variables, over and above the

classical deterrence va riables. While the difference was relatively large in the case of the

frequency of unlicensed driving (with the explained variance increasing from 2% to 9%),

it was marginal for the other two dependent variables. For both the frequency of

unlicensed driving and intention to drive unlicensed in the future, the key additional


variable, punishment avoidance, was a significant predictor in both instances (albeit at p

< .05 in the case of future intentions). In the case of continued driving after detection, the

significant addition was vicarious exposure to punishment. However, this variable did not

appear to influence the participant’s behaviour in a manner consistent with deterrence

theory. This is a key finding that is further discussed in section 6.5.3.1.

6.4.6 Social learning factors

6.4.6.1 Imitation

In social learning theory, imitation refers to the process by which individuals model

their behaviour on the actions of others. As explained in Chapter 3, the primary source of

behavioural models is salient social groups, such as family and peers. However, other

reference groups can also provide a source of models. Consequently, a composite

variable was created to measure the participants’ overall exposure to unlicensed driving

models. This variable combined the number of family and friends known by the

participants to have driven unlicensed with the number of other people known to have

driven unlicensed.

Table 6.7 provides a breakdown of the total unlicensed driving models known by

the participants, by type of offender. The sub-groups with the highest means were the

inappropriate licence, not currently licensed and never licensed drivers. However,

caution should be exercised when interpreting the results for the inappropriate licence

holders, given the low number of drivers in this group and the influence of one particular

participant who reported knowing 41 other people who had driven unlicensed. (This

participant had been charged with riding a motorcycle exceeding the engine capacity

required by his class of licence. Hence, it is possible that many of the other unlicensed

people he knew were other similar riders.)

The bivariate correlations between the composite models variable and the

dependent variables are shown in Table 6.8. As can be seen, the only significant

correlation was between the total models known and intention to drive unlicensed in the

future [r = .23, p <.01].


Table 6.7

Total number of unlicensed drivers known by participants by type of offender


qualified

%

Suspended

%

Expired

%


%

Never licensed

%

Inapp. Licence

%

Total

%

Significance level

Total no. of unlicensed drivers (models) known by participants

n=52 n=109 n=91 n=21 n=26 n=10 n=309

Mean 4.8 3.5 3.6 7.1 6.3 9.1 4.4 H (df5)= 15.12,

Median 2.5 2.0 1.0 4.0 5.0 3.5 2.0 p = .01, η = .21

Std. deviation 6.1 4.5 5.7 7.0 6.3 13.1 6.0

Minimum 0 0 0 0 0 0 0

Maximum 25 26 30 22 23 41 41

Table 6.8

Bivariate correlations between dependent variables and social learning variables

Dependent variables




in the future1

Imitation

Total models who have driven unlicensed .03 .07 .23**

Differential association

Behavioural dimension .10 .20*** .27***

Normative dimension .15* .27*** .38***

Personal attitudes (definitions)

Attitudes to unlicensed driving .10 .27*** .48***

Attitudes to alternatives -.22*** -.16** -.22***

Balance of reinforcement

Anticipated rewards .06 .19** .17**

Anticipated punishments -.08 -.20*** -.38***

1. Logarithmically transformed. * p < .05 ** p < .01 *** p ≤ .001


6.4.6.2 Differential association

Differential association refers to the patterns of interaction between a person and

other individuals and groups with whom they identify, particularly family and friends.

The construct has both a behavioural and normative dimension. The behavioural

dimension relates to the degree of interaction a person has with significant others and the

resulting behavioural models to whom they are exposed. This dimension was

operationalised by a dichotomous variable relating to whether the participants had any

family and friends who currently engaged in unlicensed driving or not. As shown in

Table 6.8, this variable was significantly correlated with both continued driving after

detection and intention to drive unlicensed in the future. No significant differences were

found between the unlicensed driver types in relation to the behavioural dimension of

differential association [χ2 (df5, n=309) = 10.82, p > .05, φ c = .19].

The normative dimension of differential association relates to the normative or

evaluative climate toward different behaviours found in the groups of significant others.

This dimension was measured by four items representing favourable or neutral attitudes

of family and friends toward unlicensed driving (see F4 in Appendix F for a list of the

items). The scale created from these four items had a Cronbach’s alpha of .76 and was

significantly correlated with all three dependent variables (see Table 6.8). The strongest

relationship was with intention to drive unlicensed in the future [r = .38, p <.001]. No

significant differences were found between the unlicensed driver types in relation to the

normative dimension of differential association variable [F (5, 301) = 0.24, p > .05,

η2=.004].

6.4.6.3 Personal attitudes

Consistent with Akers’ (1977; 1990) social learning theory, the personal attitudes of

the participants were operationalised by creating two scales measuring (i) attitudes to

unlicensed driving, and (ii) attitudes to alternative behaviours. The personal attitudes to

unlicensed driving scale consisted of 12 items (7 favourable or neutral to unlicensed

driving and 5 unfavourable to the behaviour) measured on a seven-point Likert scale (see

F5 in Appendix F for a list of the items in the scale). The Cronbach’s alpha for the scale

was .73. As shown in Table 6.8, the attitudes to unlicensed driving scale was significantly

correlated with continued driving after detection [rpb = .27, p <.001] and intention to

drive unlicensed in the future [r = .48, p <.001]. There was no significant difference

between the unlicensed driver types in terms of their attitudes to unlicensed driving [F (5,

302) = 0.98, p > .05, η2=.02].


The personal attitudes to alternative transport scale consisted of five items (3

favourable/neutral to alternative transport and 2 unfavourable) (see F6 in Appendix F).

The Cronbach’s alpha for the scale was a somewhat low .66. Although it is ideal for the

Cronbach’s alpha to be at least .7, alphas of .6 can be considered adequate (Aron & Aron,

1999, p. 527). As shown in Table 6.8, the scale was negatively (and significantly)

correlated with all three dependent variables. In other words, more extensive unlicensed

driving was associated with less favourable attitudes to alternative transport. Once again,

there was no significant difference between the unlicensed driver types on this variable [F

(5, 303) = 2.01, p > .05, η2=.03].

Consideration was given to creating a composite balance of attitudes from the two

relevant scales. However, it was decided not to do this, in order to preserve the

information provided by the two scales. In other words, retaining both scales in the

analyses allowed an insight to be obtained into the relative influence of the two sets of

attitudes on unlicensed driving.24

6.4.6.4 Differential reinforcement

Differential reinforcement relates to the balance of reinforcement (rewarding or

desired outcomes) and punishment (negative or desirable consequences) that an

individual anticipates in relation to different actions. The reinforcers (rewards) and

punishments can be either social or non-social in nature. Two scales were used to

operationalise the differential reinforcement construct.

The first scale was anticipated rewards for unlicensed driving which was measured

using six items, five of which measured the potential social rewards for the behaviour

(eg. praise from family or friends) and one which measured the non-social rewards

(‘unlicensed driving generally makes you feel good’). The Cronbach’s alpha for this scale

was .74 (see F7 in Appendix F for a list of the items). As expected, this variable was

positively correlated with all three dependent variables (see Table 6.8). However, the

relationship was not significant in the case of the frequency of unlicensed driving.

Interestingly, there was a significant difference between the unlicensed driver types in

terms of their anticipated rewards for unlicensed driving scores [F (5, 302) = 2.64, p <

.05, η2=.04]. The unlicensed driver types with the highest means were the never licensed

24. To act as a check, the regression analyses reported later in this section were undertaken using a

‘balance of attitudes’ composite variable instead of the two separate attitude variables. These analyses produced similar results with slightly lower R2s.


drivers [M = 12.3], the not currently licensed [M = 10.8] and disqualified drivers [M =

10.0], compared with the cancelled [M = 9.3] and expired [M = 8.9] drivers.

The second scale used to measure the balance of reinforcement was punishments

for unlicensed driving, which was measured using eleven items. Consistent with Akers’

(1977; 1990) proposition that social learning theory subsumes formal deterrence

processes, the potential social punishments included both informal and formal social

sanctions. Four items were used to measure informal social sanctions (such as social

disapproval from family and friends, losing your job) and three were used to measure the

formal (legal) social sanctions (perceived severity, certainty and swiftness of

punishment).25 In addition, three items were included to measure potential non-social

punishers (feelings of guilt and worry, and perceived danger) and one item measured

overall perceived risks of driving unlicensed. The Cronbach’s alpha for the anticipated

punishments for unlicensed driving scale was relatively low at .64 (see F8 in Appendix

F). 26

As expected, the anticipated punishments variable was negatively correlated with

all three dependent variables, indicating that more extensive unlicensed driving is

associated with lower anticipated punishments for the behaviour. However, the

correlations were only significant for continued driving after detection and intention to

drive unlicensed (see Table 6.8). In contrast to the anticipated rewards variable, no

significant differences in the anticipated punishment scores were found between the

unlicensed driver types [F (5, 294) = 1.29, p > .05, η2=.02].

Some previous studies using Akers’ theory have operationalised a balance of

reinforcement variable by either summing the anticipated rewards and punishers or

including an item/scale that measured the participants’ overall perception of rewards and

punishers (eg. Akers et al, 1979; DiBlasio & Benda, 1990; Akers & Lee, 1996). The first

of these strategies was decided against in this study, in order to preserve information

about the relative contribution of anticipated rewards and punishments to unlicensed

driving behaviour.

25. Some studies using Akers’ model also incorporate a measure of the perceived risk of apprehension

into the ‘punishments’ construct. That was problematic in this study due to there being two measures of this construct (ie. prior to and after detection). In addition, the previous analyses had already suggested that perceptions regarding punishment avoidance might play a more important role in unlicensed driving than perceptions towards apprehension. Accordingly, it was decided to examine the influence of these perceptions on the social learning model at a later stage (see section 6.5.3.2).

26. It should be noted that a slightly higher Cronbach’s alpha (.68) was obtained if the three items relating to formal social (legal) sanctions were excluded from the scale. However, as noted earlier, the inclusion of these items is an important comp onent of Akers’ model.


The second strategy was explored by the inclusion of the following item in the

study: ‘Overall, more good things are likely to come from unlicensed/disqualified driving

than bad’. Subsequent regression analyses indicated that the inclusion of this item in the

model only increased the overall R2 marginally (in the order of 1%). Accordingly, in the

interests of parsimony it was decided to exclude this item from the analyses.

6.4.6.5 Predictive role of social learning variables


conducted to examine the influence of the social learning variables on the frequency of

unlicensed driving. The model was significant [F (7, 289) = 4.65, p < .001], accounting

for 8% of the adjusted variance [R2 = .10, AdjR2 = .08]. The attitudes to alternative

behaviours [ß = -.26; p < .001; sr2 = .06] variable was a significant predictor at the p <

.01 level, while attitudes to unlicensed driving [ß = -.17; p < .05; sr2 = .01] and

anticipated punishments [ß = -.15; p < .05; sr2 = .02] were significant at a less stringent

level.

It could be argued that social learning theory will be of little relevance in cases

where participants were unaware that they were driving unlicensed. Accordingly, a

further analysis was undertaken excluding these participants (resulting in a sample of 200

participants). This model proved to be a better predictor [F (7, 192) = 4.72, p < .001],

accounting for 12% [R2 = .15, AdjR2 = .12] of the variance in the frequency of unlicensed

driving. In this model, only attitudes to alternative behaviours and anticipated

punishments remained significant predictors.


influence of the social learning variables on whether the participants continued to drive

unlicensed after detection. The model was significant [χ2 (7, 297) = 44.30, p < .001], with

both the behavioural dimension [odds ratio = 2.38] and normative dimension [odds ratio

= 1.09] of differential association being significant predictors at the p < .01 level. As

indicated by the odds ratios, both of these variables were associated with a higher

likelihood of continued driving after detection. The Attitudes to alternative behaviours

variable was a significant predictor [odds ratio = 0.95] at the p < .05 level and was

associated with a lower likelihood of continued driving.


conducted to examine the influence of the social learning variables on intention to drive

unlicensed in the future. This regression model was significant [F (7, 290) = 18.05, p <

.001], accounting for 29% of the adjusted variance [R2 = .30, AdjR2 = .29]. Three of the


variables were significant predictors at the p < .01 level: attitudes to unlicensed driving [ß

= .26; p < .001; sr2 = .03]; the behavioural dimension of differential association [ß = .17;

p < .01; sr2 = .02]; and anticipated punishments [ß = -176; p < .01; sr2 = .02]. In addition,

attitudes to alternative behaviours [ß = -10; p < .05; sr2 = .01] was significant at the less

stringent level. The first two of these variables were positively associated with the

dependent variable, suggesting that more favourable attitudes to unlicensed driving and

exposure to family and friends who engage in the behaviour serve to encourage intentions

to drive unlicensed. In contrast, the other two variables were negatively related,

suggesting that anticipated punishments and more favourable attitudes to alternative

behaviours can serve to discourage intentions to drive unlicensed in the future.

6.4.7 Summary of contributing factors

6.4.7.1 Comparison of the different theoretical perspectives

Table 6.9 provides a summary of the strength of association between the dependent

variables and the various theoretical perspectives examined in the previous sections.27

Before proceeding, an important caveat needs to be placed on the discussion. In the case

of both the frequency of unlicensed driving and intentions to drive unlicensed in the

future, the strength of association is represented by the adjusted R2 derived from linear

regression. As such, this figure can be directly interpreted as the amount of variance

explained in the dependent variables by the independent variables. However, the same

variance interpretation cannot be applied to the Nagelkerke R2 derived from the logistic

regressions on the continued driving after detection dependent variable. As explained in

section 6.3.3, while the Nagelkerke R2 has been proposed as an analogue to R2 in linear

regression, it only acts as an approximation for it (Tabachnick & Fidell, 2001, p.545).

This has three important implications. Firstly, the Nagelkerke R2 should only be

interpreted as an indicator of the strength of association between the dependent variable

and the independent variables, rather than as a measure of the amount of variance

explained by the independent variables. Secondly, it would be inappropriate to compare

the Nagelkerke R2 and the linear regression R2 obtained for any particular theoretical

perspective. Finally, additional caution should be exercised when interpreting the

Nagelkerke R2, since this measure has been less widely studied than linear R2.

27. For the sake of completeness, non-significant results are included. While the R2 for the sensation

seeking and AUDIT variables were not reported earlier they can be calculated by squaring the correlation co-efficient.


Table 6.9

Strength of association between the theoretical perspectives and the dependent variables

Dependent variables

Variable Frequency of unlicensed driving1,2

Continued to drive after detection3


in the future1,2

Social-demographic variables .08*** .08 . 03*

Sensation seeking .01 .00 .00

Alcohol misuse .01 .00 .00

Environmental facilitating factors .07*** .04 .08***

Classical deterrence theory variables .02 .07* .08***

Expanded deterrence theory variables .09*** .12** .10***

Social learning theory variables .08*** .20*** .29***

1. Variables logarithmically transformed. 2. Adjusted R2 shown (except for sensation seeking and alcohol misuse variables). 3. Nagelkerke R2 shown (except for sensation seeking and alcohol misuse variables) with significance relating to

overall logistic regression models. * p < .05 ** p < .01 *** p < .001

As can be seen in Table 6.9, social learning theory appeared to perform the best

across the three dependent variables. This result was most clearly evident in the case of

intention to drive unlicensed in the future, where the social learning variables accounted

for a substantially greater amount of variance than the other perspectives. Similarly, in

the case of continued driving after detection, the results indicate a stronger association

with the social learning variables than for any of the other perspectives. The pattern of

results relating the Nagelkerke R2 is also reflected in the significance levels of the logistic

regression models.

The predictive utility of the social learning variables was less evident in the case of

the frequency of unlicensed driving. In this case, social learning theory only appeared on

a par with a number of the other perspectives. As will be argued later, these results

suggest that social learning may represent a better explanation of the factors influencing

the decision to break the law (ie. to drive unlicensed), while other more immediate factors

influence the frequency of the law breaking (see section 6.5.2).

While the results suggest that social learning represents a more comprehensive

theoretical perspective than deterrence theory, they highlight that there are contributing

factors to unlicensed driving that remain unexplained. Even in the best of the cases (ie.


intention to drive unlicensed), over two-thirds of the variance remains unaccounted for.

This raises the possibility of developing a more consolidated theoretical perspective

relevant to unlicensed driving. This issue is further explored in the next two sections.

6.4.7.2 The effectiveness of deterrence and social learning theories in explaining

deviant behaviours

The particular model of social learning theory used in this study was primarily

developed as a theory of deviance (Akers 1977; 1990). Similarly, deterrence theory is

generally characterised as a criminological theory of deviance. In this respect, it is

possible that these two perspectives are better suited to explaining more deviant forms of

behaviour. To explore this issue, the sample of unlicensed drivers was divided into two

groups. The first group comprised the disqualified, not currently licensed and the never

licensed drivers. The results of both this study and Study Two suggest that these drivers

represent a more deviant sub-group, as evidenced by their higher reported levels of: prior

convictions for unlicensed driving and criminal offences (see section 5.4.1.4); drink

driving behaviour (see section 5.4.3.4); and alcohol misuse (see section 6.4.3). The

second group comprised the suspended, expired and not appropriately licensed drivers.

While these offenders had failed to comply with the law, their offences were in some

instances of an administrative nature (eg. failing to renew their licence) and may have

been committed unknowingly (see section 5.4.4.1).

Table 6.10 summarises the results from a series of regression analyses undertaken

to examine the predictive utility of the expanded deterrence and social learning theories

with the two subgroups of unlicensed drivers identified above (full details of the analyses

are shown in Tables I19-I30 in Appendix I). Some caution should be exercised when

interpreting the results of the two multiple regressions involving the disqualified, not

currently licensed and never licensed drivers. The size of this sub-group was

approximately 93 for each analysis, which resulted in a ratio of cases to independent

variables that was below the optimum recommended for detecting a medium effect-size

while maintaining statistical power (Tabachnick & Fidell, 1996). This may have

contributed to the fact that three of the models were not significant at the p < .01 level. In

addition, the afore-mentioned interpretative issues relating to the Nagelkerke R2 should

be borne in mind when comparing the logistic regression models.


Table 6.10

Strength of association between dependent variables and expanded deterrence and social

learning variables, by unlicensed driver sub-group

Dependent variables

Variable Frequency of unlicensed driving1,2

Continued to drive after detection3


in the future1,2

Disqualified, not currently licensed and never licensed drivers (n ≈ 93)

Expanded deterrence theory variables .16** .26* .27***

Social learning theory variables .08* .20 .41***

Suspended, expired and not appropriately licensed drivers (n ≈ 200)

Expanded deterrence theory variables .07** .13* .09***

Social learning theory variables .05* .21*** .26***

1. Variables logarithmically transformed. 2. Adjusted R2 shown. 3. Nagelkerke R2 shown with significance relating to overall logistic regression models. * p < .05 ** p < .01 *** p < .001

A number of key points emerge from the results presented in Table 6.10:

§ both the expanded deterrence and social learning theory variables appear to better

explain unlicensed driving behaviour among the more deviant sub-group of drivers

(ie. the disqualified, not currently licensed and never licensed drivers);

§ with both sub-groups of drivers, the expanded deterrence theory variables better

predict the frequency of unlicensed driving than the social learning variables;

§ the results for the more deviant sub-group of drivers were mixed, with the expanded

deterrence theory variables proving superior in the case of the frequency of

unlicensed driving and continued driving after detection but the social learning

variables proving very effective in predicting intention to drive unlicensed [AdjR2 =

.41]; and

§ on balance the social learning theory variables appear better able to explain

unlicensed driving behaviour among the less deviant group of drivers (ie. the

suspended, expired and not appropriately licensed drivers).


As will be discussed later, these findings suggest that a major strength of social

learning theory may be its capacity to explain unlicensed driving among a wider range of

offenders than the expanded deterrence theory. Nonetheless, they also indicate that there

may be scope to enhance the application of social learning theory to unlicensed driving,

particularly among more deviant offenders.

6.4.7.3 Extending the social learning explanation of unlicensed driving

This section explores the key factors that appear to contribute to unlicensed driving

behaviour, over and above the particular social learning variables operationalised in this

study. It was prompted by an awareness that a number of the variables examined in

previous sections are not inconsistent with social learning theory. For example, although

punishment avoidance was principally conceptualised (and tested) as a deterrence

variable in this study, it is consistent with the tenets of social learning theory (see section

6.5.2). Similarly, there is a strong argument for incorporating the need to drive for work

purposes into a more comprehensive social learning explanation of unlicensed driving.

Three consolidated regression analyses were conducted for this purpose. They

examined the influence of the social learning variables, along with other variables that

had previously been identified as significant predictors of the three independent

variables.28

The frequency of unlicensed driving

Table 6.11 reports the results of the first consolidated regression analysis. The

overall model was significant accounting for 24% of the adjusted variance [R2 = .27,

AdjR2 = .24] in the frequency of unlicensed driving.

28. Given the exploratory nature of this study, those variables that had only proven significant at the less

stringent level of p < .05 were also included in the consolidated regression analyses. The deterrence variables selected were those that were significant in the expanded deterrence regression models.


Table 6.11

Standard multiple regression of social learning and other selected variables on frequency

of unlicensed driving (n=292)

Variables Mean Std. dev

B Std. error

ß sr2 R2 Adj R2

Frequency of unlicensed driving1 .98 .47

Total unlicensed driving models 4.36 5.96 -.00 .01 -.07

Differential association (behavioural dimension) 1.23 .42 .14* .07 .13 .01

Differential association (normative dimension) 10.61 5.57 .01* .01 .17 .02

Attitudes to unlicensed driving 37.59 12.08 -.01* .003 -.16 .01

Attitudes to alternative behaviours 18.80 7.09 -.01* .004 -.14 .02

Anticipated rewards 9.60 5.29 .00 .01 .02

Anticipated punishments 53.94 10.22 -.01 .003 -.13

Needed to drive for work when unlicensed .42 .50 .21*** .05 .22 .04

Educational level 1.54 .50 .08 .06 .08

Prior criminal conviction .40 .49 .17** .06 .18 .03

Still had photo licence .49 .50 .18** .05 .19 .03

Perceived risk of apprehension (prior to detection)

3.29 1.81 -.02 .01 -.07

Punishment avoidance .36 .48 .20*** .05 .20 .04

.27*** .24

1. Logarithmically transformed. Model: F (13, 278) = 7.88, p < .001 Unique variability = .20; shared variability = .07 * p < .05 ** p < .01 *** p < .001

As shown in Table 6.11, the predictors significant at the p < .01 level were:

§ needed to drive for work while unlicensed [ß = .22; p < .001; sr2 = .04];

§ punishment avoidance [ß = .20; p < .001; sr2 = .04];

§ still had photographic licence [ß = .19; p < .01; sr2 = .03]; and

§ prior criminal conviction [ß = .18; p < .01; sr2 = .03].

All four of these predictors were positively associated with the frequency of

unlicensed driving. This suggests that more frequent unlicensed driving is associated

with: needing to drive for work; being exposed to incidents of punishment avoidance;

being in possession of a photographic licence; and having a prior criminal conviction. In

addition, a further four variables were significant at the p < .05 level:


§ normative dimension of differential association [ß = .17; p < .05; sr2 = .02];

§ attitudes to unlicensed driving [ß = -.16; p < .05; sr2 = .01]

§ attitudes to alternative behaviours [ß = -.14; p < .05; sr2 = .02]; and

§ behavioural dimension of differential association [ß = .13; p < .05; sr2 = .01].

Greater caution should be exercised towards these findings. Nonetheless, they

suggest that differential association and attitudinal factors may play a role encouraging

more frequent unlicensed driving.

Continued unlicensed driving after detection

Table 6.12 reports the results of a logistic regression conducted on the factors

associated with continued driving after detection.

Table 6.12

Logistic regression analysis of continued driving after detection as a function of social

learning and other selected variables (n=292)


error Wald test

Odds Ratio Upper Lower

Total unlicensed driving models -.02 .03 .44 .98 .93 1.04

Differential association (behavioural dimension) .83 .35 5.45* 2.28 1.14 4.56

Differential association (normative dimension) .10 .04 7.76** 1.10 1.03 1.18

Attitudes to unlicensed driving .00 .02 .02 1.00 .97 1.04

Attitudes to alternative behaviours -.04 .02 3.31 .96 .92 1.00

Anticipated rewards -.01 .03 .03 .99 .94 1.06

Anticipated punishments -.02 .02 1.53 .98 .94 1.01

Age1

21 - 25 -.35 .39 .82 .70 .33 1.51

26 - 39 -.67 .40 2.76 .51 .23 1.13

40 or over -.38 .79 .23 .69 .15 3.23

Needed to drive for work when unlicensed

.68 .33 4.36* 1.98 1.04 3.75

Employed at the time of court hearing

-.68 .33 4.25* .51 .27 .97

Perceived risk of apprehension (after detection)

-.13 .08 2.86 .88 .76 1.02

Vicarious exposure to punishment .06 .33 .03 1.06 .56 2.02

1. Reference category is 17 – 20. Full model vs. constant-only model: χ2 (df14, n=292) = 52.97, p < .001; Nagelkerke R2 = .24 * p < .05 ** p < .01 *** p < .001


As shown in Table 6.12, the overall logistic regression model was significant, with

one predictor significant at the p < .01 level: the normative dimension of differential

association [odds ratio = 1.10]. As shown by the odds ratio, being exposed to family and

friends who hold favourable attitudes to unlicensed driving was associated with a modest

increase in the likelihood of continued driving after detection.

In addition, three predictors were significant at the p < .05 level:

§ needed to drive for work while unlicensed [odds ratio = 1.98];

§ employed at the time of court hearing [odds ratio = .51]; and

§ the behavioural dimension of differential association [odds ratio = 2.28].

Once again, more caution should be exercised towards these results. However, they

suggest that needing to drive for work while unlicensed and being exposed to family and

friends who engage in unlicensed driving may encourage continued driving after

detection. In contrast, being employed at the time of the court hearing may reduce the

likelihood of continued driving after detection.

Intention to drive unlicensed in the future

Table 6.13 reports the consolidated standard multiple regression for the factors

associated with intention to drive unlicensed in the future. The regression model was

significant and accounted for 32% of the adjusted variance [R2 = .35, AdjR2 = .32] in

intentions to drive unlicensed.


Table 6.13

Standard multiple regression of selected variables on intention to drive unlicensed in the

future (n=296)


B Std. error

ß sr2 R2 Adj R2

Intention to drive unlicensed in the future1 .69 .77

Total unlicensed driving models 4.32 5.93 .01 .01 .06


Differential association (normative dimension) 10.58 5.57 .02* .01 .13 .01

Attitudes to unlicensed driving 37.70 12.04 .01** .004 .20 .02


Anticipated rewards 9.59 5.30 -.01 .01 -.10

Anticipated punishments 53.93 10.23 -.01** .01 -.16 .02

Marital status 1.22 .42 -.14 .09 -.08

Unaware of being unlicensed

1.33 .47 -.21* .09 -.13 .01

Able to access vehicle .95 .21 .20 .18 .06


4.60 2.01 -.03 .02 -.08

Prior conviction for unlicensed driving .39 .49 .19* .08 .12 .01

Punishment avoidance .36 .48 .11 .08 .07

.35*** .32

1. Logarithmically transformed. Model: F (13, 282) = 11.85, p < .001 Unique variability = .09; shared variability = .26 * p < .05 ** p < .01 *** p < .001

Two of the variables were significant predictors: attitudes to unlicensed driving [ß

= .20; p < .01; sr2 = .02]; and anticipated punishments [ß = -.16; p < .01; sr2 = .02]. As

expected, the regression coefficients indicate that more favourable attitudes to unlicensed

driving are associated with stronger intentions while greater anticipated punishments is

associated with weaker intentions to drive unlicensed in the future.


In addition, five variables were significant at the p < .05 level:

§ behavioural dimension of differential association [ß = .14; p < .05; sr2 = .01];

§ unaware of being unlicensed [ß = -.13; p < .05; sr2 = .01];

§ normative dimension of differential association [ß = .13; p < .05; sr2 = .01];

§ prior conviction for unlicensed driving [ß = .12; p < .05; sr2 = .01]; and

§ attitudes to alternative behaviours [ß = -.11; p < .05; sr2 = .01].

These results suggest that intentions to drive unlicensed in the future may be

weaker among those who were originally unaware of driving unlicensed and hold

favourable attitudes to alternative behaviours. Conversely, the intentions of participants

may be stronger when they: hold favourable attitudes to unlicensed driving; are exposed

to family and friends who engage in the behaviour and hold favourable attitudes towards

it; and have a prior conviction for unlicensed driving.

Summary of predictors

Many of the significant predictors identified in the above three regression

analyses were the social learning variables operationalised in this study. However, a

number of other variables emerged that need closer consideration, in order to establish a

more consolidated perspective on unlicensed driving. In particular, needing to drive for

work purposes while unlicensed emerged as the strongest predictor of the frequency of

unlicensed driving. In addition, although only significant at p < .05, it was also the third

strongest predictor of continued driving after detection. Other significant influences on

the frequency of unlicensed driving were exposure to punishment avoidance and

possession of a photographic licence. The social learning variables were most effective in

accounting for intention to drive unlicensed in the future. Nonetheless, the addition of

variables relating to the participants’ original awareness of being unlicensed and whether

they had a prior conviction for unlicensed driving improved the predictive utility of the

regression model. The relevance of the above variables for better understanding

unlicensed driving within a more comprehensive social learning framework is discussed

in section 6.5.3.2.

6.5 Discussion

6.5.1 Study limitations

Given that the data for this study was collected in the same way as that for Study

Two, it is subject to the same limitations as outlined in section 5.5.1. Firstly, the data was

exclusively drawn from a metropolitan setting with a bias toward offenders detected in an


inner city and suburban area. Secondly, males were significantly less likely to agree to

participate than females and some of the offender groups were relatively small (eg. the

inappropriate licence category). Thirdly, it is unclear to what extent the behaviour of the

sample is indicative of unlicensed drivers who have not been detected by the police. It is

possible that offenders who remain undetected are generally more cautious (and possibly

safer) than those caught by the police. However, as noted in section 5.5.1, the majority of

the offenders in the sample were detected through random enforcement processes.

Furthermore, many of the offenders had been driving unlicensed for quite long periods of

time before they were detected. As such, it is arguable that many of the offenders would

have remained undetected except for the operation of RBT and other random licence

checking processes.

Moreover, due to the design of the study, it was not possible to directly compare

people who had driven unlicensed with those who hadn’t. As such, it was not possible to

directly examine the factors contributing to the decision to drive unlicensed or not.

Instead, a number of variables were selected for investigation tha t were intended to

reflect important theoretical and practical aspects of unlicensed driving behaviour. Hence,

the validity of the analyses undertaken in the study is largely a product of the

meaningfulness of these variables. In this respect, it was encouraging to note that the

results obtained for the various theoretical perspectives were largely consistent across the

three dependent variables, particularly for continued driving after detection and future

intentions. Nonetheless, it would be important for future research to replicate this study

with a more general sample of drivers, using other dependent variables as indicators of

unlicensed driving. This issue is further discussed in section 7.6.

Finally, the statistical methods used in the study were selected to examine the

factors that predict unlicensed driving behaviour and compare the predictive utility of

different theoretical perspectives, rather than explore the structural nature of these

perspectives. As such, few conclusions can be drawn about the structural nature of either

the deterrence or social learning perspectives utilised in this research. This would have

required the use of structural analytic techniques (such as path analysis or structural

equation modelling) that were beyond the scope of this research, and not directly relevant

to the research questions being examined.

6.5.2 Support for study hypotheses


outlined in section 6.2.


Among unlicensed driving offenders, the behaviour will be more extensive among

those who are male and younger (H17)

Limited evidence was found to support this hypothesis. Gender was not

significantly correlated with any of the dependent variables. Hence, while males may be

over-represented in the crashes involving unlicensed drivers (see section 4.4.2) and

among unlicensed driving offenders (see section 5.4.1.3), this would not appear to be

reflected in either the amount of unlicensed driving they undertake or a tendency to

continue driving after detection. It is more likely that their overinvolvement in road

crashes is indicative of other risk-taking behaviour, such as drink driving and speeding.

Age was weakly correlated with both continued driving after detection [rpb = -.13,

p <.05] and intention to drive unlicensed in the future [r = -.12, p <.05]. These

correlations were negative, indicating that younger offenders were more likely to

continue driving after detection and report an intention to do so. However, age did not

prove a significant predictor of any of the dependent variables in the consolidated

regression analyses.

Unlicensed driving will be positively associated with prior criminal offending (H18)

In Study Two it was found that 38.8% of the participants had a prior conviction for

a criminal offence, suggesting a possible link between unlicensed driving and more

deviant behaviour. While having a prior conviction for a criminal offence was not

significantly correlated with any of the dependent variables, it did emerge as a significant

predictor of the frequency of unlicensed driving in the consolidated regression analysis [ß

= .18; p < .01; sr2 = .03]. This suggests that unlicensed drivers who are prepared to

engage in criminal behaviour may be less concerned about the threat of detection and

punishment and hence drive more frequently.

Unlicensed driving will be positively associated with the need to drive for work

purposes (H19)

In Study Two it was found that over a quarter [26.0%] of the trips reported by the

participants were for work-related purposes. Hence, it was hypothesised that needing to

drive for work may be one of the major influences on unlicensed driving behaviour.

Strong support was found for this hypothesis. The need to drive for work when

unlicensed was significantly correlated with the frequency of unlicensed driving [r = .28,

p < .001]. In addition, the consolidated regression analyses undertaken in section 6.4.7.3

indicated that this variable was the strongest predictor of the frequency of unlicensed


driving [ß = .22; p < .001; sr2 = .04] and, although only significant at p < .05, it was the

second strongest predictor of continued driving after detection [odds ratio = 1.98].

The need to drive for work has previously been identified as one of the main factors

contributing to unlicensed driving (eg. Robinson, 1977; Ross & Conzales, 1988; Mirlees-

Black, 1993; Job et al, 1994). Besides acting as a major motivation for the behaviour, it

serves to increase the overall exposure of offenders.

These findings have important practical and theoretical implications. At a practical

level, it highlights the need to develop countermeasures that can target this issue (see

section 7.4.2.4). At a theoretical level, it highlights the need to acknowledge the

important indirect benefits or rewards that can arise from unlicensed driving. As will be

argued later, this is possible within a social learning framework.

Unlicensed driving will be positively associated with higher levels of sensation

seeking (H20)

The findings of Study Two suggested that the participants were relatively high

sensation seekers and that there was an association between sensation seeking and self-

reported speeding while unlicensed (see section 5.4.1.6). However, there was no evidence

in this study that sensation seeking directly influenced the extent of unlicensed driving

behaviour or future intentions. Therefore, while sensation seeking may in part contribute

to a person losing his/her licence, its influence appears to be limited to the types of

actions (such as speeding) that they are likely to engage in while unlicensed. At a

theoretical level, this suggests that sensation seekers do not find the experience of

unlicensed driving intrinsically rewarding. While this conclusion is consistent with the

evidence from this study, it needs to be confirmed with a more diverse sample of

unlicensed drivers.

Unlicensed driving will be positively associated with alcohol misuse (H21)

Alcohol misuse was reasonably common among the participants, with over half

(54.7%) reporting consumption levels that could be considered hazardous or harmful. As

with sensation seeking, however, it appears that alcohol misuse does not directly

influence either the extent of unlicensed driving or future intentions since no significant

correlation was found between it and the three dependent variables. However, alcohol

misuse was significantly associated with a range of other outcomes including prior

criminal and unlicensed driving convictions and self- reported drink driving while

unlicensed (see section 5.4.1.7). Therefore, while alcohol misuse appears to be associated


with drink driving among some unlicensed drivers, it does not appear to contribute to

more extensive unlicensed driving.

Unlicensed driving will be positively associated with the availability of motor

vehicles and continued possession of a photographic licence (H22)

Some support was obtained for this hypothesis. The ability of participants to access

a vehicle while unlicensed was significantly correlated with the frequency of unlicensed

driving [r = .17, p < .01] and, to a lesser extent, with intention to drive unlicensed [r = .12,

p < .05]. Owning a vehicle was also significantly related to the frequency of unlicensed

driving [r = .15, p < .01]. However, neither of these factors proved significant predictors

in the consolidated regression analyses undertaken in section 6.4.7.3.

Stronger evidence was obtained in relation to being in possession of a photo licence

while driving unlicensed. This variable was the third strongest predictor of the frequency

of unlicensed driving [ß = .19; p < .01; sr2 = .03] in the consolidated regression analysis.

As explained in section 6.4.4, however, the significance of this result is difficult to

interpret. Certain evidence suggests that some of the offenders who still had their photo

licence were not aware that they were unlicensed and, as such, would not have made any

attempt to restrict their driving. This interpretation is supported by the significant

association found between being in possession of a photographic licence and being

unaware of being unlicensed [φ =.49, p < .001]. Furthermore, while being in possession

of a photo licence was positively associated with the frequency of unlicensed driving it

was negatively correlated with future intentions. However, even after the participants

who were unaware of being unlicensed were excluded, there was still a weak, but positive

correlation [rpb = .18, p <.05] between the possession of a photographic licence and more

frequent unlicensed driving. This raises the possibility that some offenders may be

tempted to drive more regularly when they are in possession of a photo licence, in the

belief that it may assist them to avoid detection. This has implications for policies and

practices relating to the recovery of licences from drivers who are disqualified or

suspended from driving (see section 7.3.4.1).

An expanded theory of deterrence, incorporating the constructs of punishment

avoidance and vicarious learning, will better predict unlicensed driving than

classical deterrence theory (H23)

The results provided good support for this hypothesis. The expanded deterrence

theory variables appeared better able to explain all three dependent variables than the


classical deterrence variables alone. This was confirmed by hierarchical regression

analyses that indicated that the three expanded deterrence variables significantly added to

the prediction of the dependent variables, over and above the classical deterrence

variables. The additional variance explained by the expanded deterrence variables was

relatively large in the case of the frequency of unlicensed driving [with AdjR2 increasing

from .02 to .09], but more modest for intention to drive unlicensed [with AdjR2 increasing

from .08 to .10]. In both these cases, the key additional variable appeared to be

punishment avoidance, which was a significant predictor in both instances (albeit at p <

.05 in the case of future intentions). Similarly, the logistic regressions indicated that there

was a slightly stronger association between continued driving after detection and the

expanded deterrence variables compared with the classical deterrence variables [with the

Nagelkerke R2 increasing from .07 to .12]. Moreover, vicarious exposure to punishment

proved the strongest predictor among the expanded deterrence variables.

However, some of the above results are not straightforward. Although there was a

strong positive relationship between punishment avoidance and the frequency of

unlicensed driving [r = .31, p < .001], the causal direction of this relationship is unclear. It

is possible that the successful evasion of detection may serve to encourage more frequent

unlicensed driving. This is consistent with Stafford and Warr’s (1993, p.125) assertion

that offenders “whose experience is limited largely to avoiding punishment may come to

believe that they are immune from punishment, even in the face of occasional evidence to

the contrary”. Alternatively, the findings may only indicate that those offenders who

drive more frequently have more opportunities to evade detection.

Overall, the evidence appears to provide more support for the view that punishment

avoidance contributes to more frequent unlicensed driving. Firstly, if the findings

concerning punishment avoidance were merely a product of increased driving exposure, a

similar relationship should also have been found between unlicensed driving and

exposure to enforcement. In other words, exposure to enforcement should also have been

associated with more frequent unlicensed driving. However, this was not the case;

exposure to enforcement was negatively correlated [rpb = -.14, p <.05] with the frequency

of driving. Secondly, a positive relationship was also found between punishment

avoidance and intentions to drive unlicensed in the future [r = .17, p < .01]. In this

respect, it is reasonable to conclude that the experience of avoiding punishment for

unlicensed driving directly contributes to a stronger intention to drive unlicensed in the

future. Thirdly, as would be expected, the punishment avoidance variable was negatively

correlated with the perceived risk of apprehension in the future [rpb = -.11, p < .05] and


positively correlated with the item: “You were lucky not to have got caught earlier for

driving unlicensed/disqualified” [rpb = .13, p < .05]. This suggests that the experience of

punishment avoidance had impacted on the perceptions of the participants relating to the

likelihood of detection.

As noted above, the inclusion of the vicarious exposure to punishment variable

enhanced the capacity of the expanded deterrence theory to explain continued driving

after detection. However, this variable did not appear to influence the participant’s

behaviour in a manner consistent with deterrence theory. According to Stafford and Warr

(1993), vicarious exposure to punishment should have a similar effect as direct exposure,

namely to deter future offending. On the contrary, in this study there was a significant

positive relationship between vicarious exposure to punishment and unlicensed driving

(as measured by continued driving after detection and intention to drive unlicensed). This

result raises major conceptual problems for Stafford and Warr’s (1993) position that will

be further discussed in section 6.5.3.1. Moreover, an argument will be presented below

that the findings relating to vicarious exposure to punishment can be better explained

from a social learning perspective.

Social learning theory will better predict unlicensed driving than either the

classical deterrence or expanded deterrence theories (H24)

The findings of this study largely support this hypothesis. The social learning

variables accounted for substantially more variance in intention to drive unlicensed

[AdjR2 = .29] than either of the deterrence theory sets of variables [AdjR2 = .08 and AdjR2

= .10]. Similarly, the results of the logistic regressions suggested that there was a stronger

association between continued driving after detection and the social learning variables

[Nagelkerke R2 = .20], than for either of the deterrence theory variables [Nagelkerke R2 =

.07 and .12]. While the predictive utility of the social learning variables [AdjR2 = .08] was

less evident in the case of the frequency of unlicensed driving, the expanded deterrence

variables proved only marginally better [AdjR2 = .09]. Together, these results suggest tha t

social learning theory represents a more comprehensive explanation of unlicensed driving

than either deterrence theory.

However, the pattern of the results also suggest that social learning may represent a

better explanation of the factors influencing the decision to break the law (ie. to drive

unlicensed), while more immediate factors (such as the availability of a vehicle, the need

to drive for work purposes, the perceived likelihood of apprehension or the perceived

chances of avoiding detection) influence the frequency of the law breaking. As such, the


results highlight the need to distinguish between factors that contribute to a propensity to

break the law and those which influence the frequency of the law breaking.

Three other key results from the study tend to support the utility of a social learning

perspective. Firstly, as already noted, the need to drive for work while unlicensed was the

strongest predictor of the frequency of unlicensed driving and the second strongest

predictor of continued driving after detection. The role of this variable appears to be

better explained by a social learning perspective than either of the deterrence perspectives

used in the study. The need to drive for work represents a powerful motivation to drive

unlicensed, since it facilitates the obtaining of personal rewards (eg. income, social

status) and reduces the potential costs associated with not driving (ie. potential loss of

employment). Consequently, it is not surprising that those offenders who drive unlicensed

for work-related reasons would perceive the benefits of the behaviour to outweigh the

potential costs. In this respect, the strength of social learning theory is that it considers

both the perceived benefits and costs associated with illegal (and other) behaviours,

whereas deterrence theories traditionally focus on the perceived costs of illegal

behaviours.

A second finding that appears better explained by a social learning perspective than

deterrence theory relates to the influence of vicarious exposure to punishment. As noted

above, the findings relating to the role of this variable do not appear consistent with

deterrence theory. Indeed, it could be argued that the influence of vicarious exposure to

punishment is consistent with the effects of differential association, as proposed within

social learning theory. In other words, knowing others who have been previously

convicted of unlicensed driving represents exposure to models who have been actively

engaged in the behaviour (behavioural dimension). In addition, these models may hold

favourable attitudes to the behaviour (normative dimension), particularly if they weren’t

overly inconvenienced by the penalties they received. As such, the learning that occurs as

a result of vicarious exposure to punishment may in some circumstances act to encourage

the behaviour, rather than deter it. This explanation is consistent with the data obtained

from this study, particularly for continued driving after detection and intention to drive

unlicensed in the future.29

Finally, although punishment avoidance was principally conceptualised (and tested)

as a deterrence variable in this study, it is consistent with the tenets of social learning

29. Similarly, direct exposure to punishment would have a minimal deterrent impact in cases where

offenders do not perceive the penalties they receive to be overly severe or certain and where their behaviour is socially rewarded (or at least not punished) within their social reference group. This issue is discussed further in section 6.5.3.1.


theory (Akers et al, 1979). Indeed, as will be argued in section 6.5.3.1, there is a strong

argument for incorporating punishment avoidance into a social learning explanation of

unlicensed driving.

Both expanded deterrence theory and social learning theory will better predict

unlicensed driving among more deviant offenders (H25)

The results provided good support for this hypothesis. In five out of the six relevant

regression analyses, the expanded deterrence and social learning variables appeared to

better explain unlicensed driving among the more deviant sub-group of drivers within the

sample (ie. the disqualified, not currently licensed and never licensed drivers). In

particular, both perspectives accounted for substantially more variance in the future

intentions of the more deviant sub-group compared with the other offenders. Therefore,

these results tend to confirm that both theories are better able to explain non-compliance

with the law among more persistent, deviant offenders. This result may in part reflect the

fact than many of the expired and suspended drivers claimed that they were unaware of

being unlicensed (ie. their behaviour was not deliberate). While this may have influenced

the results for the frequency of unlicensed driving, it is unclear how it would have

impacted on the other two dependent variables. Over and above this, the results have

important implications for the ongoing development of criminological theory in general,

and traffic psychology theory in particular (see section 7.3 for further discussion of this

issue).

A number of other key results emerged from these analyses. Firstly, among the

more deviant offenders, mixed results were obtained for the two theories. The expanded

deterrence theory variables proved better at predicting the frequency of unlicensed driving

[AdjR2 = .16 cf. AdjR2 = .08], but the social learning variables proved very effective in

predicting intention to drive unlicensed in the future [AdjR2 = .41 cf. AdjR2 = .27].

Secondly, on balance the social learning theory variables appear better able to predict

unlicensed driving behaviour among the less deviant group of drivers (ie. the suspended,

expired and not appropriately licensed drivers). With this group of offenders, the social

learning variables accounted for more variance in intention to drive unlicensed than the

expanded deterrence variables [AdjR2 = .26 cf. AdjR2 = .09]. Similarly, there was a

stronger association between the social learning variables and continued driving after

detection, than was the case for the expanded deterrence variables [Nagelkerke R2 = .21***

cf. .21***]. While the results were reversed for the frequency of unlicensed driving, the R2

in both cases was relatively low (around 10%).


6.5.3 Theoretical implications

6.5.3.1 Implications for deterrence theory

Classical deterrence theory

Classical deterrence theory suggests that drivers will be deterred from driving

unlicensed if they perceive a high likelihood of apprehension, and if the resulting

penalties are perceived to be sufficiently certain, severe and swift. The evidence from this

study tends to suggest that these conditions are not currently being achieved for many

offenders. The classical deterrence variables accounted for minimal variance in the

frequency of unlicensed driving and intentions to drive unlicensed (at best 8%). The most

encouraging results were obtained for the perceived risk of apprehension, which was

found to be correlated (albeit weakly) with the three dependent variables used in the

study. However, it did not prove a significant predictor in any of the consolidated

regression analyses. In addition, the knowledge of current penalties among offenders was

relatively low and no significant association was found between the perceived severity or

swiftness of current penalties and the three dependent variables. While a weak correlation

[rpb = -.13, p < .05] was found between the perceived certainty of punishment and

continued driving after detection, it did not prove a significant predictor of the behaviour.

Finally, rather than act as a specific deterrent, prior conviction for unlicensed driving was

positively correlated [rpb = .20, p < .001] with intention to drive unlicensed in the future

and a marginal predictor [ß = .12; p < .05; sr2 = .01] of this variable in the consolidated

regression analysis. Although these latter findings are inconsistent with deterrence theory,

they are consistent with findings from other recent studies [see the quote from Piquero

and Pogarsky (2002) below].

On a positive note, there may have been one particular group of offenders who

were deterred (to some degree) by the process of detection and punishment. In one of the

consolidated regression analyses, being unaware of being unlicensed emerged as a

marginal negative predictor of intention to drive unlicensed [ß = -.13; p < .05; sr2 = .01].

In other words, the intentions of the participants to drive unlicensed in the future were

weaker when they had originally been unaware of driving unlicensed. However, the

degree to which this result reflects the process of deterrence is unclear. In many cases, it

may be more indicative of the fact that the participants’ behaviour was anomalous in the

first instance. It may also reflect a need on the part of some participants to reassert their

moral commitment to the law.

The above findings do not necessarily invalidate classical deterrence theory. Rather,

it could be countered that they merely indicate that the necessary conditions for deterring


unlicensed driving are not currently being achieved. However, the results of this study

indicate that other theoretical perspectives provide a more comprehensive explanation of

unlicensed driving. Moreover, the results suggest that classical deterrence theory does not

represent an adequate framework on which to base countermeasure development.

Expanded deterrence theory

The results of this study provide partial support for Stafford and Warr’s (1993)

reconceptualisation of deterrence theory. The expanded deterrence variables better

explained all three dependent variables than the classical deterrence variables alone. In

particular, the inclusion of punishment avoidance improved the utility of the expanded

deterrence perspective. This variable was significantly correlated with both the frequency

of unlicensed driving [rpb = .31, p < .001] and intention to drive unlicensed [rpb = .17, p <

.001]. Moreover, it was the second strongest predictor [ß = .20; p < .001; sr2 = .04] of the

frequency of unlicensed driving in the consolidated regression analysis (which included

the social learning variables).

As already noted, however, the way in which punishment avoidance was

operationalised in this study represents a special case of the construct proposed by

Stafford and Warr (1993) (see section 6.4.5.2). Due to the nature of this study, it was

necessary to operationalise punishment avoidance in terms of direct or active evasion of

detection to distinguish it from the frequency of unlicensed driving. As such, it is

arguable that these episodes would have represented an intensified instance of

punishment avoidance.

In contrast, the results relating to vicarious learning processes were less supportive

of Stafford and Warr’s reconceptualisation. Firstly, while vicarious exposure to

punishment avoidance was positively correlated with all three independent variables (as

expected), these relationships were all weak. Consequently, this variable did not prove a

significant predictor of any of the dependent variables. Secondly, the results relating to

vicarious exposure to punishment were inconsistent with Stafford and Warr’s (1993)

propositions. As already discussed, rather than deter offending, vicarious exposure to

punishment was positively correlated with both continued driving after detection [φ =.20,

p < .001] and future intentions [rpb = .18, p < .001].

These findings largely accord with previous tests of Stafford and Warr’s (1993)

reconceptualisation (Piquero & Pogarsky, 2002). It is worth noting the comments of these

authors in full:


Contrary to Stafford and Warr’s (1993) reconceptualisation and the overall logic

of deterrence, however, punishment experiences appear to encourage future

offending. Our findings reflect this “emboldening effect” in several ways. First, the

bivariate correlations relating offending to both personal and vicarious punishment

are positive. Second, in most multivariate specifications, the coefficient for

vicarious punishment is positive and statistically significant. Third, congruently

substantial personal and vicarious punishment experiences associate positively

with offending. Although contrary to Stafford and Warr (1993), these findings

largely accord with both prior tests of their theory . . . (Piquero & Pogarsky, 2002,

p.178).

Piquero and Pogarsky (2002) explore a number of possible explanations to account

for these anomalies including the effects of defiance and self-serving bias. While these

have yet to be empirically explored, there are other possible explanations for the results.

At their simplest level, they may simply indicate that exposure to punishment that is not

perceived as being particularly severe may only serve to encourage illegal behaviour.

However, as already argued, the results relating to vicarious exposure to punishment may

actually reflect the process of differential association within social learning theory.

6.5.3.2 Implications for social learning theory

The findings from this study provide strong support for social learning theory. The

social learning variables operationalised in this study proved particularly effective in

explaining continued driving after detection and intention to drive unlicensed in the

future (both of which are arguably more reflective of the decision to drive unlicensed

than the frequency of unlicensed driving). In addition, the social learning variables

appeared on a par with the expanded deterrence variables in predicting the behaviour of

the more deviant offenders in the sample and superior for the remainder of the sample.

Hence, a major strength of social learning theory appears to be its capacity to more

effectively explain unlicensed driving among a wider range of offenders than the

expanded deterrence theory.

The social learning construct that appeared most relevant for explaining the

behaviour of unlicensed drivers was differential association. The normative dimension of

differential association was a significant predictor of continued driving after detection

[odds ratio = 1.10; p < .01 ] in the consolidated logistic regression. It was also a marginal

predictor of both the frequency of unlicensed driving [ß = .17; p < .05; sr2 = .02] and

intention to drive unlicensed [ß = .13; p < .05; sr2 = .01]. Similarly, the behavioural


dimension of differential association was a predictor of all three dependent variables at

the p < .05 level. These results are consistent with a range of other studies that have

found differential association to be a key variable in explaining illegal or high-risk

behaviours such as: adolescent alcohol and drug use (Akers et al, 1979); cigarette

smoking (Krohn et al, 1985); drink driving (DiBlasio, 1986); adolescent sexual behaviour

(DiBlasio & Benda, 1990); and computer crime (Skinner & Fream, 1997).

All the other social learning variables were significant predictors (p < .01) of at

least one of the independent variables except exposure to models and anticipated

rewards. In the case of exposure to models, this finding is not necessarily surprising. As

noted by Akers et al (1979, p.647):

. . . imitation in social learning theory is considered to have its greatest effect in the

first acquisition stages of behaviour while the associational, reinforcement, and

definitional variables are more important in the maintenance of a behavioural

pattern.

Similarly, DiBlasio (1986, p.14) in a study of drink drivers found that the: “effect of

modelling is most influential during initial learning periods, then becomes slightly less

important than other variables (i.e. normative definitions) during later years”. In this

study, almost 40% of the participants had a prior conviction for unlicensed or disqualified

driving (see section 5.4.1.4). In addition, even among the first offenders the mean length

of time driving unlicensed was .88 years (ie. almost 11 months). Consequently, it was

highly likely that many of the offenders were beyond the initiation phase of the

behaviour.

The fact that anticipated rewards did not prove a significant predictor in any of the

consolidated regression analyses may reflect a shortcoming in the way the variable was

operationalised in the study. The variable was based on a scale that measured perceptions

about a range of social and non-social benefits of unlicensed driving. While the items in

the scale were drawn from previous research and pilot tested, there was no specific

item(s) that measured the anticipated benefits of driving for work while unlicensed.30

This would appear an important weakness of the scale, given the other results obtained

relating to needing to drive for work purposes. Hence, although the anticipated rewards

variable was significantly correlated with both continued driving after detection [rpb =

.19, p < .01] and intention to drive unlicensed [r = .17, p < .01], it may have proved

30. In contrast, two items were included in the anticipated punishments scale relating to potential work-

related punishments. These related to the potential for unlicensed driving to lead to social disapproval from workmates or loss of employment.


redundant in the consolidated analyses due to the influence of the needed to drive for

work while unlicensed variable.

In summary, while the findings of this study provide strong support for social

learning theory, they suggest that there is scope to enhance the application of the theory

to unlicensed driving. At its best, the social learning variables were able to account for

almost one-third of the variance in intention to drive unlicensed among the total sample

and over two-thirds among the more deviant participants. Nonetheless, this indicates that

there are important contributing factors to unlicensed driving that remain unaccounted

for. In this regard, some of the other key predictors of unlicensed driving identified in this

study appear consistent with social learning theory. These factors are discussed below.

The role of punishment avoidance

Based on the work of Stafford and Warr (1993), punishment avoidance was

principally conceptualised (and tested) as a deterrence variable in this study. Although

Akers (1977; 1990) has argued that deterrence theory can be subsumed within social

learning theory, the studies that have tested his social learning model have primarily

focused on perceptions toward punishment, rather than the experience (and related

perceptions) of punishment avoidance. Consequently, these studies have typically

measured formal deterrence in terms of the probability/certainty of detection and the

perceived severity of punishment (eg. Akers et al, 1979; DiBlasio, 1987). Accordingly, a

similar approach was adopted in this study. However, the concept of punishment

avoidance is consistent with the tenets of social learning theory. Indeed, as noted by

Akers et al (1979, p.638) when explaining the learning mechanisms underpinning social

learning theory:

Social behaviour is acquired through direct conditioning and through imitation or

modelling of others’ behaviour. Behaviour is strengthened through reward (positive

reinforcement) and avoidance of punishment (negative reinforcement) or

weakened by aversive stimuli (positive punishment) and loss of reward (negative

punishment). (Bolding added).

In the current study, the classification of punishment avoidance as a social learning

variable would have significantly improved the capacity of the theory to predict the

frequency of unlicensed driving and marginally improved the prediction of intention to

drive unlicensed. Consequently, there is a strong case for incorporating the concept of

punishment avoidance into future research utilising social learning, as well as deterrence,

perspectives. In this regard, punishment avoidance was operationalised in this study in


terms of actual experiences, rather than as a perceptual phenomenon. Piquero and

Pogarsky (2002) have provided evidence that both personal and vicarious experiences

affect behaviour by influencing sanction risk perceptions. This is an issue that requires

more investigation from a social learning perspective. While there was a positive

relationship between punishment avoidance and the perceived risk of apprehension (after

detection), the correlation was relatively weak [rpb = -.11, p < .05]. Hence, further

research is required into the mechanisms by which the experience of punishment

avoidance affects sanction risk perceptions.

The role of employment as an influence on unlicensed driving

The results clearly indicate the need to better acknowledge the role of employment

as a key factor influencing unlicensed driving behaviour. Firstly, the need to drive for

work while unlicensed was the strongest predictor of the frequency of unlicensed driving

and the second strongest predictor (albeit at p < .05) of continued driving after detection.

From a social learning perspective, this suggests that driving facilitates important work-

related benefits or rewards that outweigh potential punishments for some offenders.

In contrast, there was other evidence that being employed (at least at the time of the

court hearing) was found to reduce the odds of continued driving after detection. This

suggests that being employed, except in cases where it necessitates driving, may tend to

act as a protective factor against continued offending. For example, employed offenders

may be concerned about the long-term impact of a further penalty on their capacity to

travel to and from work. Alternatively, the results may be more indicative of the

characteristics of those people who tend to be employed. Consequently, future research in

this area should examine the social learning principles underpinning the influence of

employment on unlicensed driving.

The role of vicarious punishment in social learning

The results of this study suggest that vicarious exposure to punishment is positively

associated with illegal behaviour and future intentions. While this finding represents a

challenge for deterrence theory, it is consistent with the effects of differential association,

as proposed within social learning theory. As noted earlier, being exposed to others who

have broken the law not only provides a potential source of models for imitation, but also

may convey normative beliefs that are favourable (or at least neutral) to the behaviour.

However, further research is required into the way in which vicarious exposure to

punishment influences a person’s sanction risk perceptions, and hence their anticipated

balance of reinforcement. In particular, it is unclear to what degree the perceptions of a


vicarious party will be influenced by their own beliefs, compared with the observed

reactions of those who are punished. In this respect, Piquero and Pogarsky (2002) suggest

that a systematic perceptual bias may exist whereby individuals tend to minimise the

impact of other people’s punishment experiences. They explore a number of possible

reasons for this apparent emboldening effect including the effects of defiance and self-

serving bias. This latter phenomenon relates to the “tendency for individuals to shade

judgements in a manner favourable to themselves” (Piquero & Pogarsky, 2002, p.179)

and is similar to the concept of optimism bias which has received considerable attention

in the traffic psychology literature (eg. Job, 1999).

6.5.4 Countermeasure implications

This study has a number of important implications for the design of more effective

countermeasures to unlicensed driving. At a general level, the strong support obtained for

social learning theory in this study suggests that countermeasures should not solely focus

on increasing the costs associated with unlicensed driving, but attempt to reduce the

potential benefits associated with the behaviour (such as being able to drive for work

purposes). At a more specific level, a number of key issues emerged which need to be

given more consideration in the process of countermeasure development including:

§ the low perceived risk of apprehension for unlicensed driving behaviour and the

role of punishment avoidance indicates that enforcement practices need to be

modified to maximise detection rates, while maintaining high levels of enforcement

activity;

§ given that current penalties appear to have minimal deterrent impact, there is a need

to review their adequacy and the strategies used to educate offenders about the

consequences of re-offence;

§ the positive association between the frequency of unlicensed driving and owning a

vehicle suggests that there may be some value in implementing vehicle-based

sanctions;

§ although the relationship between the possession of a photo licence and unlicensed

driving is complex, it was a significant predictor of the frequency of unlicensed

driving and hence requires more investigation;

§ the need to drive for work purposes represents a major influence on unlicensed

driving behaviour that needs to be better targeted;


§ educational strategies targeting unlicensed driving should not only emphasise the

likely costs associated with unlicensed driving, but try to dispel beliefs about some

of the apparent rewards (such as driving for work purposes); and

§ driver licensing policy makers need to consider strategies to better reward

participation in the licensing system.

6.6 Chapter summary

This study represents a culmination of the program of research undertaken into

unlicensed driving. It has pursued many of the themes identified in the two previous

studies and attempted to answer some of the questions they raised. A central aim of the

study has been to identify the factors that contribute to the decision to drive unlicensed.

While this question could not be answered directly, it was possible to examine the factors

that contribute to a number of different aspects of the behaviour.

This study has served two important functions in the overall program of research.

At a theoretical level, it has examined the predictive utility of a number of different

perspectives that have received attention in other areas of driver behaviour research.

While strong support was obtained for social learning theory, it has also highlighted a

number of conceptual issues that require further attention to improve the prediction of

unlicensed driving behaviour. Of particular importance is the findings relating to the

influence of punishment avoidance. This has important implications for the development

of more comprehensive deterrence and social learning models of illegal behaviour. At a

practical level, the study has important implications for the design of more effective

unlicensed driving countermeasures. At a global level it suggests that countermeasures

should not solely focus on increasing the costs associated with unlicensed driving. Rather

they should attempt to reduce the potential benefits associated with the behaviour and

increase the rewards associated with participating in the licensing system. Some of the

specific findings of the study offer direction on how this might be achieved.

The following chapter will draw together the findings of all three studies. It will

attempt to synthesise the implications of the research for better understanding the

problem of unlicensed driving and developing more effective countermeasures.


Chapter Seven: Discussion

7.1 Introductory comments .................................................................................. 219

7.2 Review of findings ......................................................................................... 219

7.2.1 Do unlicensed drivers engage in more risky driving than other drivers?.................................................................................................. 219

7.2.2 Is unlicensed driving associated with a higher crash risk compared to legal driving? .................................................................................... 220

7.2.3 Do unlicensed drivers represent a homogeneous group, in terms of their psychosocial characteristics and on-road behaviour? .............. 222 7.2.4 How effective are current administrative, enforcement and punishment policies and processes in preventing unlicensed driving? .................................................................................................

223

7.2.5 What are the personal, social and environmental factors contributing to unlicensed driving? ...................................................... 224 7.3 Contribution to theory.................................................................................... 226

7.4 Implications for road safety ........................................................................... 228

7.4.1 The need to encourage participation in the driver licensing system.................................................................................................... 228 7.4.2 Countermeasure suggestions ................................................................. 230

7.4.2.1 Driver licensing and other administrative processes ................. 230

7.4.2.2 Traffic law enforcement practices ............................................. 231

7.4.2.3 Unlicensed driving sanctions and punishment processes .......... 232

7.4.2.4 The need to target work-related driving .................................... 234

7.5 Strengths and limitations of the research ...................................................... 235

7.6 Suggestions for future research...................................................................... 239

7.7 Concluding remarks ....................................................................................... 241




This program of research has explored the psychosocial characteristics and on-road

behaviour of unlicensed drivers. This has served two important purposes. Firstly, it has

provided an insight into the road safety implications of unlicensed driving in general, and

more particularly among different types of offenders. Secondly, the research has provided

important theoretical and practical insights into the factors that appear to contribute to the

behaviour.

This final chapter will draw together the findings from the three studies and discuss

the theoretical and practical implications for road safety and traffic psychology. From a

theoretical perspective, the research has important implications for the on-going

development of deterrence theory and social learning theory, both generally and

particularly in the area of driver behaviour. In terms of policy, the chapter will review the

likely effectiveness of different countermeasures and identify preventative strategies

requiring further development and evaluation. The limitations of the research will also be

discussed, along with suggestions for future research.

To provide a foundation for the chapter, the next section will review the main

findings of the three studies. They will be discussed under headings corresponding to the

research questions examined.

7.2 Review of findings

7.2.1 Do unlicensed drivers engage in more risky driving than other drivers?

The findings of Studies One and Two provide strong evidence that unlicensed

drivers do engage in more risky driving than general drivers. In Study One, it was found

that the serious casualty crashes involving unlicensed drivers were more likely to involve

alcohol/drugs, speeding and inexperience, than those involving licensed drivers. This

pattern of risk-taking was evident in the higher proportion of single vehicle, motorcycle

and at- fault crashes involving unlicensed drivers. In addition, the serious casualty crashes

involving unlicensed drivers were more likely to occur at night and on the weekends,

suggesting a link between crash involvement and recreational driving.

The data from Study Two suggest that this pattern of behaviour is not just limited to

crash- involved drivers, but is typical of unlicensed driving offenders. The offenders in the

study reported considerably higher levels of risky driving than respondents in the ATSB’s

(2000) community telephone survey. For example, among the unlicensed drivers, 25%

reported exceeding the speed limit by 10 km/h or more on most occasions, 15% admitted

that they didn’t always wear their seat belt, and 7% reported that they didn’t restrict their


drinking when driving. In comparison, among the drivers responding to the ATSB’s

(2000) telephone survey, only 10% reported exceeding the speed limit as frequently, 4%

admitted that they didn’t always wear their seat belt and only 1% reported that they didn’t

restrict their drinking when driving. Furthermore, almost 25% of the unlicensed drivers

reported driving (while unlicensed) when they thought they might have been over the

legal alcohol limit.

While caution should be exercised when comparing the results of different types of

surveys, the relevant items used in Study Two were directly modelled on those in the

ATSB survey. In addition, the results are consistent with the findings of crash-based

studies, including Study One. Therefore, while there is a need to replicate these findings,

the comparison appears valid.

In addition, Study Two examined two of the psychological factors that have been

shown to influence risky driving. A relatively strong, positive relationship [rpb=.49, p

<.001] was found between alcohol misuse and self- reported drink driving. This is

consistent with international research that has demonstrated a link between alcohol

misuse, drink driving and disqualified driving (Simpson & Mayhew, 1991; Mayhew,

Simpson & Beirness, 1997). In contrast, the association between sensation seeking and

risky driving behaviour was, at best, weak and limited to speeding [rpb = .12, p <.05].

This finding is somewhat surprising given the numerous studies that have shown a link

between sensation seeking and risky driving (Jonah, 1997). It is possible that the

construct was poorly operationalised in this study. However, this seems unlikely given

that the particular scale used (the Thrill and Adventure Subscale of the Sensation Seeking

Scale) has been shown to be associated with risky driving in many other studies (Jonah,

1997). Alternatively, there was some evidence that the sample consisted of a relatively

uniform group of high sensation seekers. Hence, the results may be indicative of the lack

of variability in the offenders’ scores.

7.2.2 Is unlicensed driving associated with a higher crash risk compared to legal

driving?

Consistent with the above evidence, the results of Study One indicate that

unlicensed drivers do have a higher crash risk than licensed drivers. Based on the quasi-

induced exposure method, unlicensed drivers in Queensland were found to be almost

three times [2.90:1] more likely to be involved in all crashes, compared with licensed

drivers. It is not possible to directly compare these results with those obtained by

DeYoung et al (1997) in California, due to differences in the way that unlicensed driving


was defined in the two studies. However, it is interesting to note that the fatal crash ratio

obtained for the suspended/revoked drivers [3.7:1] in the Californian study was quite

similar to the total crash ratio calculated for the disqualified/suspended drivers in this

study [3.84:1].

The similarity of these results points to the robustness of the quasi- induced

exposure method. However, there are a number of potential problems with the method.

On the one hand, the distinction between at- fault and innocent drivers (which is implicit

in the method) is open to a negative halo bias, whereby the police may be more likely to

find an unlicensed driver at fault for a crash than a licensed driver. This bias would tend

to inflate the involvement rates for the unlicensed drivers and, hence, their crash rates. On

the other hand, the method’s use of multi-vehicle crashes introduces another potential

bias because it is likely to underestimate the full extent of at- fault driving among

unlicensed drivers. This would serve to deflate the crash rates for this group. While these

biases would tend to mitigate each other, it highlights that the results obtained using the

quasi- induced exposure method should be treated with some caution. It would be ideal to

replicate the calculations in other jurisdictions. More importantly, there is a need to

develop better methods of estimating the exposure of unlicensed drivers, which would

enable their crash risk to be estimated through more direct methods (see section 7.6). This

would also provide a benchmark to assess the validity and reliability of the quasi- induced

exposure method.

Other findings from Study Two highlight the risks associated with unlicensed

driving. In the event of a crash, unlicensed drivers were more than twice [2.12:1] as likely

than licensed drivers to be involved in a serious casualty crash (ie. one resulting in a

fatality or serious injury) compared with a minor crash. It appears that some of this

increased risk is partly due to the age and gender of the people who drive unlicensed and

the types of vehicles they drive, rather than unlicensed driving per se. In particular, riding

a motorcycle was associated with four and half times [4.58:1] the risk of being involved

in a serious casualty crash. However, after accounting for the influence of these factors,

unlicensed driving was still associated with almost twice [1.82:1] the risk of being

involved in a serious casualty crash.

Together, the results from Study One strongly question the existence of a

disqualified driver effect, whereby unlicensed drivers drive in a more cautious manner to

avoid detection (Hurst, 1980). While there was some evidence that unlicensed drivers

reduce their overall driving exposure, this does not seem to make them any safer than

licensed drivers. On the contrary, they appear to be less safe.


There are a number of possible explanations to account for these results. Firstly, it

is possible that among some (possibly many) unlicensed drivers the desire to avoid

detection tends to result in more cautious driving. Secondly, it is possible that even

among those unlicensed drivers who admit regularly breaking road rules, their driving is

more cautious than it would otherwise be (even though they are not as cautious as general

drivers). Finally, as suggested by Warren (1982), it is possible that the behaviour learned

while driving unlicensed may not actually be safer, but rather more oriented to avoiding

detection. Further research is required into this issue. However, the results of this study

do question the common assumption that unlicensed drivers drive in a more cautious

manner, if general community behaviour is adopted as the yardstick.

7.2.3 Do unlicensed drivers represent a homogeneous group, in terms of their


The results from all three studies provide strong evidence that unlicensed drivers do

not represent a homogeneous group. Study One indicated that there were significant

differences in the crash involvement patterns of the different types of unlicensed drivers.

These differences related to the age and gender of the drivers; the circumstances of the

crashes in which they were involved (including the types of vehicles involved, the time of

the crash and the number of other vehicles involved); and the contributing factors to these

crashes (including the role of alcohol/drugs and inexperience and being at-fault). In

particular, two problem groups of drivers emerged: the never licensed and the

disqualified/suspended drivers. These drivers were the most likely to be at- fault for the

crashes in which they were invo lved and had the highest risk of being involved in a crash

and for that crash to be more severe.

Study Two largely confirmed these results. A range of differences were found

between the offenders in terms of their psychosocial characteristics (age, education,

income, alcohol misuse) and self-reported on-road behaviour (limiting of exposure, drink

driving). As in Study One, a particular cluster of participants emerged who appeared to

be a more deviant group of offenders: the disqualified; not currently licensed; and never

licensed drivers. These offenders reported higher levels of prior criminal offending,

alcohol misuse and self- reported drink driving. Importantly, the suspended drivers did not

appear to share the same characteristics as the disqualified drivers. This suggests that the

grouping of the disqualified and suspended drivers in the crash statistics obscured some

important differences between these types of offenders in Study One.


Study Three provided further evidence of the differences between the offenders. A

series of regression analyses were undertaken to examine how well the expanded

deterrence and social learning perspectives could predict different aspects of unlicensed

driving among the more deviant group of offenders (ie. the disqualified, not currently

licensed and never licensed drivers) compared with the remaining offenders (ie. the

suspended, expired and inappropriate licence drivers). In almost all the cases, both

perspectives proved better at predicting unlicensed driving among the more deviant

offenders. This suggests that there are key differences between the two groups in terms of

the factors that contribute to unlicensed driving.

Together, these results have important implications for the design of future

countermeasures to unlicensed driving. They suggest that the motivations for driving

unlicensed may vary considerably among different types of offenders and that a narrow

range of countermeasures is unlikely to adequately address the problem. Rather, multi-

strategy approaches will be required that target different psychological and social

characteristics of offenders and their social networks. These issues are further discussed

in section 7.4.1. and 7.4.2.

7.2.4 How effective are current administrative, enforcement and punishment policies and

processes in preventing unlicensed driving?

At the time they were interviewed, the majority of the offenders in Study Two

indicated that it was relatively unlikely that they would drive unlicensed in the future.

This suggests that the process of detection and punishment appears to have had a salutary

effect on many offenders, at least in the short term. However, other findings from Studies

Two and Three suggest that there are a range of shortcomings in the operation of current

policies and practices. In particular, it does not appear that the necessary conditions to

deter offending are currently being achieved. The perceived risk of apprehension for

unlicensed driving is significantly lower than it is for being random breath tested or being

caught for speeding. In general, there is limited knowledge of the penalties for unlicensed

driving, with only 14% of participants reporting that they were aware of the fine prior to

the court hearing. Instances of punishment avoidance were relatively common, with over

36% of the sample reporting that they were able to evade detection from the police on

one or more occasions when they could otherwise have been identified. Almost one-third

(30.5%) of the sample admitted that they continued to drive illegally after they were

detected. Finally, prior conviction for unlicensed driving was found to be a positive

predictor of intention to drive unlicensed in the future.


In addition, a number of potential problems with current administrative processes

were identified. Over one-third (36%) of offenders claimed that they were unaware, or at

least unsure, of being unlicensed at the time they were detected. This was more common

among the expired (53.8%) and suspended (40.4%) drivers. While this result may in part

reflect the influence of social desirability, it raises questions about the effectiveness of

current methods used to inform drivers about the expiry and cancellation of licences.

Interestingly, Job et al (1994) reported a similar proportion (29.4%) of unaware

partic ipants in their postal survey of unlicensed drivers in New South Wales. This

suggests that this issue may be a problem across Australia.

Almost half of the sample (49%) reported that they still possessed their

photographic licence when driving unlicensed. While this is not surprising among the

expired drivers, it remains a concern that many suspended, disqualified and not currently

licensed drivers still had their photographic licence. In the case of the disqualified and not

currently licensed drivers, it suggests that the processes for surrendering licences at court

may not always be observed. In the case of the suspended drivers, the results may in part

reflect the change in policy introduced in Queensland in December 2002, that no longer

required drivers who have their licence cancelled for accumulation of demerit points to

surrender their licences.

Interestingly, possession of a photographic licence was also found to be a

significant predictor of more frequent driving in Study Three. The nature of this

relationship is complex and may be more indicative of the fact that those who were

unaware of being unlicensed were the most likely to still have their licence. Nonetheless,

the result raises the possibility that some offenders may be tempted to drive more

regularly if they retain their licence, in the belief that it will assist them to evade

detection. It is for this reason that Ross and Gonzales (1988) argued that it is important

for licensing authorities to recover licences from offenders.

Strategies to improve the various problems identified above are discussed further in

section 7.4.2.

7.2.5 What are the personal, social and environmental factors contributing to

unlicensed driving behaviour?

In order to design more effective countermeasures there is a need to better

understand the factors that contribute to the behaviour. To date, there has been little

research into this issue despite ongoing concerns about the scale of the unlicensed driving

problem. Due to the design of this research, it was not possible to directly examine the


factors that contribute to the decision to drive unlicensed or not in the first instance.

However, in Study Three it was possible to examine the contributing factors to three

important aspects of unlicensed driving: the frequency of the behaviour; whether

offenders continued to drive illegally after detection; and their future intentions.

Little evidence was obtained that either sensation seeking or alcohol misuse directly

influence unlicensed driving per se. Neither variable was significantly associated with the

three aspects of unlicensed driving examined. As already noted however, alcohol misuse

was strongly related to reported drink driving while unlicensed, and sensation seeking

was weakly related to reported speeding. Hence, it appears that the influence of these

factors is limited to the types of driving that unlicensed drivers engage in, rather than the

extent to which they engage in unlicensed driving.

In contrast, there was evidence that other personal, social and environmental factors

influenced unlicensed driving behaviour. A consolidated regression analysis indicated

that there were four significant (p < .01) predictors of the frequency of unlicensed driving.

Of these, three were personal factors (needing to drive for work purposes, being exposed

to incidents of punishment avoidance, and having a prior criminal conviction) and one

was an environmental factor (being in possession of a photographic licence). Four other

variables were significant at p < .05, suggesting the potential influence of other personal

factors (the attitudes of the offenders to unlicensed driving and alternative behaviours)

and social factors (being exposed to family and friends who engage in unlicensed driving

and who hold more positive attitudes toward the behaviour). The two strongest predictors

of the frequency of unlicensed driving were the need to drive for work purposes and

exposure to punishment avoidance.

The strongest predictor of continued driving after detection was being exposed to

family and friends who hold favourable attitudes to unlicensed driving, which was

significant at p < .01. In addition to this social factor, two personal factors were

significant at p < .05: the need to drive for work purposes; and being employed at the time

of the court hearing.

The two strongest predictors of intentions to drive unlicensed in the future were

both personal in nature (attitudes to unlicensed driving and anticipated punishments).

Five other variables were significant at p < .05 and included personal factors (prior

conviction for unlicensed driving); social factors (being exposed to family and friends

who engage in unlicensed driving and who hold more positive attitudes toward the

behaviour); and environmental factors (awareness of being unlicensed).


Together, these results suggest that personal and social factors exert the strongest

influence over unlicensed driving behaviour. At the personal level, these are: the need to

drive for work purposes; exposure to punishment avoidance; personal attitudes to

unlicensed driving; and anticipated punishments for the behaviour. At a social level, the

strongest factors relate to the social learning construct of differential association, namely:

being exposed to significant others who both engage in unlicensed driving (behavioural

dimension) and who hold positive attitudes to the behaviour (normative dimension).

7.3 Contribution to theory

Unlicensed driving represents a major challenge for theory in the area of the

behavioural/social sciences. Although it is generally conceptualised as a discrete

behaviour (driving without a valid licence), in many respects it represents a spectrum of

behaviours. This program of research has clearly shown that unlicensed drivers do not

represent a homogeneous group and that different factors appear to contribute to the

behaviour among different types of offenders. Therefore, in order to be robust, any

theoretical explanation of unlicensed driving will need to account for the behaviour

among a wide range of offenders.

In this respect, unlicensed driving represents an ideal behaviour with which to

compare the predictive utility of different theoretical perspectives. A specific objective of

this research was to compare the utility of a variety of perspectives from psychology,

criminology and sociology that are commonly used to explain illegal driving behaviour.

As such, the findings of this study were not only designed to contribute to a better

theoretical understanding of unlicensed driving, but also contribute to the on-going

development of theory in the field of traffic psychology.

A detailed discussion of the theoretical implications of Study Three is provided in

section 6.5.3. This includes a review of the relative strengths and weaknesses of the three

main perspectives that were used to explore the factors contributing to unlicensed driving

behaviour. It should be reiterated that the primary theoretical aim of this research was to

compare the predictive utility of the different perspectives, rather than confirm their

structural nature. This latter task was beyond the scope of this research program, but

should be investigated in the future.

In summary, this research has made five important contributions to criminological

theory in general, and traffic psychology theory in particular.


1. Support has been provided for Stafford and Warr’s (1993) reconceptualisation of

deterrence theory. However, the added predictive utility of this perspective (over

and above classical deterrence theory) appears to be mainly related to the inclusion

of the punishment avoidance construct (as operationalised in this study). Indeed, the

results obtained for vicarious exposure to punishment were contrary to the

predictions of the theory and more consistent with social learning theory.

2. Consistent with the propositions of Akers (1977; 1990), the results suggest that

social learning theory represents a more comprehensive perspective for predicting

illegal behaviour than deterrence theory. Two key differences between the theories

appear to account for this. Firstly, social learning theory considers both the

perceived benefits and costs associated with illegal (and alternative behaviours),

whereas deterrence theories traditionally focus on the perceived costs of illegal

behaviours. Secondly, social learning theory incorporates both normative and

personal attitudinal constructs that are absent in deterrence theories. This

recognises the important role that the attitudes of significant others can play in the

formation of personal attitudes towards a behaviour (and its alternatives), and how

these can act to strengthen or weaken behavioural options.

3. The research has identified two important constructs that need to be more

specifically operationalised in social learning theory models: experiences of

punishment avoidance and vicarious exposure to punishment (and punishment

avoidance). These constructs are consistent with the principles of social learning

and are particularly relevant in areas where enforcement is difficult and episodes of

successful punishment avoidance are likely to be common (which is the case with

many illegal driving behaviours).

4. Both the expanded deterrence and social learning perspectives proved more

effective at predicting unlicensed driving with more deviant offenders. However, a

robust theoretical model should not only be able to explain why some drivers

choose to drive unlicensed, but also account for the fact that the majority of drivers

currently comply with driver licensing requirements. Consequently, there is a need

for further research to test how well both deterrence theory and social learning can

explain compliance with licensing laws, both among the general driving population

and with those drivers who lose their licence but choose to comply with the ban.

5. The results of Study Three suggest that social learning theory represents a better

explanation of the factors influencing the decision to break the law (ie. to drive

unlicensed), but more immediate factors (such as the availability of a vehicle, the


need to drive for work purposes, the perceived likelihood of apprehension or the

perceived chances of avoiding detection) exert a stronger influence on the

frequency of the law breaking. As such, the results highlight the need to distinguish

between factors that contribute to a propensity to break the law and those which

influence the frequency of the law breaking. This is an issue requiring more

attention in future theory development and testing.

7.4 Implications for road safety

This program of research has confirmed that unlicensed driving is a relatively

small, but significant road safety problem. In particular, the findings seriously question

the existence of the disqualified driver effect and the use of this as a rationale by policy-

makers to avoid devoting more resources to the problem of unlicensed driving or to

justify policies that may inadvertently exacerbate the problem. Consequently, this

research has two clear implications for road safety. Firstly, it indicates that more effective

approaches are required to reduce the level of unlicensed driving. Secondly, there is a

need to review policies that may be inadvertently exacerbating the problem.

7.4.1 The need to encourage participation in the driver licensing system

Before reviewing countermeasure options, there is a need to discuss a ma jor

practical problem for road safety highlighted by this research. During the exploratory

research undertaken for Studies Two and Three (see section 5.3.2), it became apparent

that there were very few rewards in place to encourage participation in the licensing

system. From a social learning perspective, this means that the potential benefits

associated with driving unlicensed (such as retaining employment, being able to

participate in social activities etc) can only be offset by the anticipated punishment s for

the behaviour (such as being fined), rather than by potential rewards for complying with

the driver licensing laws.

This problem is indicative of road safety in general. The traditional approach to

behaviour change in road safety has been punitive and coercive in nature (Elliott, 1992;

Watson et al, 1996). Consequently, many researchers have argued that there is a need for

the greater use of incentives or rewards to establish and sustain higher compliance with

road rules (eg. Harano & Hubert, 1974; Wilde, 1985). Despite the intrinsic attractiveness

of incentive strategies, very few incentive programs have been implemented for drivers.

Furthermore, those that have been evaluated have proven inconclusive (Watson et al,

1996).


The difficulties associated with implementing incentives or rewards in road safety

were well recognised by Hurst (1980) in the same article in which he proposed the

existence of the disqualified driver effect (see section 2.3.2). He argued that there are two

fundamental difficulties involved:

The first is that it is hard to use reward to teach a person not to do something (like

speeding or drinking before driving) just as it is hard to use reward to train your

dog not to chase cats. . . The second is that . . . for our primary target group,

speeding and drinking driving must enjoy some peer group approval (since they are

done in a largely social context); strict observance of the road rules is scarcely

going to win admiration. Authorities are viewed with hostility. How, then, can the

authorities administer socially meaningful rewards, when they are viewed as the

enemy? (Hurst, 1980, p.271).

These comments are somewhat pessimistic in the case of the general driving

population. However, they do appear particularly apt for traffic offenders who may

already believe that they have a good reason for being hostile towards the authorities.

Based on his analysis, Hurst (1980) argued that rewards could only be successfully

administered through indirect means. Using the behaviour of disqualified drivers as an

example, he argued that there was a need to create situations where drivers were

motivated to drive safely through fear of being detected for driving unlawfully. Based on

this logic, he recommended the introduction of restrictions on young drivers (such as a

night driving ban) in the belief that those drivers who failed to comply would be

motivated to drive safely to avoid detection.

Ironically, Hurst’s (1980) recommendation has since been implemented in many

countries in the form of graduated driver licensing (although the rationale for these

schemes is quite different). However, the logic of his argument appears flawed. As noted

in section 2.3.1, Warren (1982) argues that the behaviour learnt while driving unlicensed

is not necessarily safe, but simply oriented towards avoiding detection. This is largely

supported by the findings of this research. Nonetheless, Warren (1982, p.172) suggested

that:

Reward systems are still quite plausible, but it would seem that the second

necessary pre-condition for the introduction of meaningful reward systems would be

identification of a delivery mechanism which would be viewed by the public as having a

vested interest in the promotion of safety (rather than an excessive preoccupation with

punishment).


Where does this leave the possibility of implementing incentives to encourage

compliance with driver licensing laws? Clearly, this is a difficult task that appears outside

the scope of current driver licensing and road safety policies. Indeed, the current polices

and practices in this area, such as demerit points schemes and licence

disqualification/cancellation, are punishment oriented rather than incentive-based.

Consequently, this is a critical issue requiring further research and development.

However, the current research has some important implications for re-engineering

the system in a way that would facilitate the use of more incentives. At the moment, it is

difficult to offer meaningful incentives to comply with driver licensing laws when there

are significant potential benefits associated with driving unlicensed, such as retaining

employment, participating in social activities etc. In particular, this research has

highlighted the apparent benefits associated with continuing to drive for work purposes

(since this was one of the strongest predictors of unlicensed driving in Study Three).

Therefore, a key implication of this research is the need to reduce or, at least, offset some

of the perceived benefits of driving unlicensed for work purposes. There currently

appears to be three key ways in which this could be achieved:

1. reducing the perceived benefits associated with driving for work purposes by

encouraging employers to more regularly check the licences of their workers and

take action against those who are driving illegally;

2. providing opportunities for offenders to drive legally for work purposes only,

through the use of restricted licences; and

3. providing incentives for offenders to participate in alcohol ignition interlock (or

even speed interlock) programs.

Options for implementing these strategies, along with other countermeasure

suggestions, are discussed below.

7.4.2 Countermeasure suggestions

7.4.2.1 Driver licensing and other administrative processes

The findings of this research have a number of implications for driver licensing and

related administrative processes.

Notification processes

The relatively high proportion (36%) of offenders who claimed that they were

unaware of being unlicensed raises questions about the methods currently used to inform

drivers of the expiry or cancellation of licences. At the moment in Queensland, a letter is

sent to drivers six weeks before the expiry of their licence. Given the relatively low costs


involved in sending letters, consideration may need to be given to providing an additional

reminder letter to those drivers who fail to renew their licence. However, this strategy

will have little impact in cases where drivers have failed to notify the authorities of a

change in their address. Therefore, as suggested by Job et al (1994, p.59): “Promotion of

the need to notify changes of address (with the message that it may avoid failing to

receive reminder notices) may help with this problem”.

Licence surrender processes

Almost half of the sample (49%) in this study still had a photographic licence when

driving unlicensed. Consequently, there is a need for licensing authorities to examine the

processes used for managing the surrender/retrieval of driver’s licences. In particular,

jurisdictions need to ensure that they have processes in place to monitor whether

offenders have surrendered their licences as required. More particularly, there is a need in

Queensland to examine the impact of the change in policy that no longer requires drivers

to surrender their licence when it is cancelled for accumulation of demerit points. In

particular, it is possible that this change may have increased the temptation for suspended

drivers to continue driving and may have reduced the impact of licence loss in terms of

deterrence and convenience (eg. being able to retain a licence for identification purposes).

Licence loss as a penalty for non-safety offences

Many jurisdictions have now introduced licence loss as a penalty for non-road

safety related offences, such as non-payment of fines. This has the potential to increase

the level of unlicensed driving and reduce the integrity of licence loss as a road safety

measure. Hence, research is required to examine the impact of these practices on driver

perceptions and the level of unlicensed driving.

7.4.2.2 Traffic law enforcement practices

As noted in the literature review, researchers have repeatedly identified the need to

improve the roadside technology used by police to ensure the rapid identification of

drivers who are unlicensed (eg. Smith, 1976; Job et al, 1994). In Queensland, major

developments have occurred in this area over recent years. Many police vehicles are now

equipped with a computer link to Queensland Transport’s licensing and registration

databases (Travelsafe, 1998). Continued development and implementation of this

technology is required to ensure that the police have the capacity to quickly verify the

validity of licences by the roadside.


However, the results of this survey highlight the need for more widespread

checking of driver’s licences. It is a concern that many offenders are not being detected

when they come into contact with the police. For example, 97 (31.4%) of the participants

in the study reported that they didn’t have their licence checked at an RBT operation

during the time they were driving unlicensed. As noted earlier, punishment avoidance

was found to be positively associated with both the frequency of unlicensed driving and

intention to drive unlicensed in the future. In addition, the perceived risk of apprehension

for unlicensed driving is significantly lower than that for drink driving or speeding.

A major impediment to more widespread licence checking in Australia is the lack

of compulsory carriage of licence laws. For example, while the police have the power to

randomly check licences in Queensland, it is difficult for them to do so on a systematic

basis because open licence holders are not, in practice, required to carry their licence

(Travelsafe, 1998). New South Wales is the only state that currently requires all drivers to

carry their licence, which facilitates the checking of licences at RBT operations in that

state (Watson et al, 1996).

Consequently, a strong argument exists for the national adoption of compulsory

carriage of licence and for the police to conduct more widespread, random checking of

driver’s licences (eg. at RBT and specific licence checking operations). Without these

initiatives, it will remain very difficult to meaningfully improve the detection of

unlicensed driving, and hence to heighten drivers’ perceived risk of apprehension. It is

also interesting to note that a community survey conducted by the then Federal Office of

Road Safety (FORS, 1996) found that 54% of Queensland respondents already believed

that it was compulsory to carry their licence. In addition, 78% reported that they

approved of compulsory licence carriage.

Although a strong case exists for the more systematic checking of driver’s licences

at RBT operations, it would be important to maintain the overall level of breath testing

currently being performed. Accordingly, further work is required to assess the resource

implications of such a change, in order to maintain the overall effectiveness of RBT.

7.4.2.3 Unlicensed driving sanctions and punishment processes

Penalties for unlicensed driving

The findings of the study suggest that there is a need to review the adequacy of

current penalties and punishment processes for unlicensed driving. Current penalties appear

to have a minimal deterrent impact on offenders. Levels of knowledge about the penalties

are limited and the offenders did not rate them as particularly severe. While there was


some evidence that the current penalties are perceived to be certain, their impact on

unlicensed driving behaviour appears limited. Most importantly, prior conviction for

unlicensed driving did not appear to have any significant impact on either the subsequent

frequency of unlicensed driving, continued driving after detection or future intention to

drive unlicensed.

The penalties for other offences related to unlicensed driving may also need to be

examined. For example, evidence emerged during the qualitative pilot that some

participants preferred to remain on their Learner’s Licence because the current penalty for

Unaccompanied Driving was only $30, which was lower than the costs associated with

obtaining a Provisional Licence (see section 3.2.1).

Vehicle-based sanctions

Consideration needs to be given to the use of vehicle-based sanctions to deter or

constrain unlicensed driving. In Study Two it was found that the majority of the offenders

(62.5%) were detected driving a vehicle that they owned. This highlights the potential

value of vehicle-based sanctions, such as alcohol ignition interlocks, vehicle or

registration plate confiscation/ impoundment and electronic licences.

In particular, efforts need to be directed at increasing the take-up rate of alcohol

ignition interlocks among offenders (see section 2.6.4.1). Given the findings of the

current study, strategies need to be developed to encourage drink driving offenders who

need to drive for work to participate in interlock programs, rather than have them drive

while disqualified. Such strategies may include reducing the length of disqualification

periods or making participation an alternative to more restrictive penalties like gaol or

electronically monitored house arrest (Voas et al, 1999; 2002). Further research is

required to determine whether shortening disqualification periods (in return for

participation in an interlock program) would undermine the general deterrent effect of

licence loss for drink driving.

In the past, it has also been suggested that repeat speeding offenders be required to

have their vehicles fitted with speed interlocks (ie. speed governors) (Queensland

Transport, 1993). This could also be used as a requirement to allow offenders to reduce

the length of their licence suspension, in order to return to driving for work (or other)

purposes. This would provide another means of offsetting the perceived benefits of

driving unlicensed for work.

Other vehicle-based sanctions that require consideration within the Australian

context are vehicle impoundment/immobilisation programs and the blocking of the


registration of vehicles owned by unlicensed drivers. This latter initiative was

recommended by Scopatz et al (2003) as both a deterrent to unlicensed driving and a

means of strengthening the application of vehicle-based sanctions like impoundment.

Over and above these potential sanctions, there is a need for governments to fast

track the implementation of electronic licences. Irrespective of whether offenders own a

vehicle, it is generally easy for them to access another person’s vehicle. Therefore, the

best long-term solution would appear to be to make it difficult for offenders to start a

vehicle unless they have a valid licence.

Rehabilitation

The results highlight the need to incorporate information relating to unlicensed

driving into existing drink driving rehabilitation programs. This information should focus

on the crash risks associated with unlicensed driving and the legal ramifications of the

behaviour. In addition, there is a need to consider the trialing of rehabilitation programs

specifically targeting recidivist unlicensed driving offenders. This is particularly relevant

for the more deviant sub-group of offenders identified in Study Three (ie. disqualified,

not currently licensed and never licensed drivers). While alcohol misuse appears to be a

central factor influencing the on-road behaviour of these offenders, there may be other

psychosocial factors that contribute to their unlicensed driving, such as poor impulse

control (see section 2.6.5).

7.4.2.4 The need to target work -related unlicensed driving

The need to drive for work appears to act as a major motivation for unlicensed

driving. As noted above, one way to address this problem is to encourage offenders that

need to drive for work purposes to participate in alcohol ignition interlock (or even speed

interlock) programs. Two other strategies identified in section 7.4.1 require further

attention.

The use of restricted licences

Job et al (1994) suggested that consideration be given to the use of restricted

licences as an alternative to full disqualification, to allow offenders to drive to and from

work. However, the evidence suggests that the benefits of restricted licences appear to be

minimal despite their intuitive appeal. Compliance with such licences is difficult to

enforce and they do not tend to reduce overall offences and crashes as much as full

disqualification (Watson & Siskind, 1997; Watson et al, 2000). It is also possible that the

use of restricted licences may actually undermine the general deterrent effect of licence


disqualification, by creating the impression that licence loss is neither certain nor

inevitable.

Nonetheless, the results of this research suggest that further consideration needs to

be given to the use of restricted licences. Research and development in this area should

concentrate on two issues. Firstly, there is a need to explore ways in which technology

can be used to improve the integrity of restricted licences. For example, consideration

could be given to requiring offenders on restricted licences to have timelocks (possibly as

a feature of an alcohol ignition interlock) fitted to the vehicles that they drive. These

timelocks could restrict their driving to working hours (or other preset hours). Although

this would involve additional costs for the offenders, this may be considered worthwhile

by many offenders to allow them to drive legally for work. Secondly, there is a need for

further research to determine whether the availability of restricted licences does

undermine the perceived deterrent value of licence loss. Queensland represents an ideal

location for this research, since it is one of the few Australian jurisdictions that still

makes restricted licences available for certain drink driving offenders.

Work-related driver safety initiatives

Another way to target the problem of work-related unlicensed driving is to

encourage employers to more actively monitor the licence status of their drivers. If

drivers believed that their employers were likely to check their licences, the perceived

benefits of unlicensed driving may be diminished. While this practice would have limited

impact in the case of self-employed people, many government and non-government

organisations are adopting comprehensive fleet safety policies and programs that could

include this practice. It would also provide a means of ensuring that employers were

meeting their workplace health and safety obligations and protect them against claims of

vicarious liability.

7.5 Strengths and limitations of the research

This program of research featured a number of strengths that enhanced the quality

of the information obtained. The first of these concerned the variety of methods that were

used to examine the behaviour of unlicensed drivers. The methods used in Study One

were primarily epidemiological in nature. This facilitated the analysis of the crash

involvement patterns of unlicensed drivers and how these compared to licensed drivers.

While much of the research into unlicensed driving has used this approach, it has tended

to focus on fatal crashes alone and/or has failed to distinguish between different types of

offenders (eg. FORS, 1997a,b; Griffin & DeLaZerda, 2000). To address this problem,


Study One examined the involvement of different types of offenders in a range of crash

types (including fatal crashes, serious casualty crashes and total reported crashes). While

the analysis of crash data provides important information, it remains unclear whether the

trends identified are representative of drivers as a whole or only those involved in

crashes. Consequently, a different methodology was used in Studies Two and Three. The

cross-sectional survey design used in these two studies made it possible to examine the

behaviour of a more general sample of unlicensed drivers. This facilitated the use of

social science research methods to further explore the issues raised in Study One and to

investigate other aspects of unlicensed driver behaviour.

The second strength of the research was the manner in which the cross-sectional

survey data was collected. Previous surveys of unlicensed drivers have been characterised

by a number of methodological problems, common to research dealing with illegal or

deviant behaviours. Firstly, a number of studies that have used official records to recruit

subjects have found that many no longer reside at the address provided (Robinson, 1977;

Mirrlees-Black, 1993; Job et al, 1994). Secondly, surveys in this area generally feature

high refusal rates – which is not surprising given the illegal nature of the behaviour.

Together, these factors have contributed to the low response rates (typically under 40%)

for surveys using both interview (Robinson, 1977; Mirrlees-Black, 1993) and mail

questionnaire (Robinson, 1977; Smith & Maisey, 1990; Job et al, 1994) methods. To

overcome these problems, it was decided to interview offenders at the time they attended

court and to offer an incentive payment for participation. While this approach was

considerably more laborious and costly than other methods, such as a mail survey, it

resulted in a much higher response rate [62.4%] than previous studies. In addition,

extensive piloting of the method was undertaken to minimise missing data and to ensure

that offender types were appropriately identified.

A third strength of this program of research was its strong theoretical foundation.

Previous research in the area of unlicensed driving has tended to be primarily descriptive

in nature, lacking a strong theoretical base to guide the interpretation of results (Watson,

1998a). To overcome this problem, a multi-disciplinary approach to theory was adopted

in this research, drawing on perspectives from psychology, sociology and criminology.

This not only provided a strong base for the research, but also facilitated a comparison of

the different perspectives. This broadened the scope of the research program and

increased its relevance to other areas of traffic psychology and criminology. Most

importantly, the theoretical framework used in the study enhanced the practical

application of the findings. For example, the social learning model provided a very useful


tool for understanding the practical significance of needing to drive for work purposes. In

this sense, it confirms the adage that: “there is nothing so practical as a good theory”

(Grayson, 1995, p.95).

The final strength of the research program has been its strong applied focus. A

major objective of the research was to ‘identify improvements to current countermeasures

and potential initiatives to reduce the incidence of unlicensed driving’. This was a theme

explored in all three studies and resulted in a range of options being identified to improve

current countermeasures, along with priorities for future research and development.

While recognising the strengths of the research program, it is also important to

acknowledge its limitations. The major limitation is that the data for all three studies were

drawn exclusively from one jurisdiction within Australia. Consequently, caution needs to

be exercised when generalising the results to other parts of Australia or the world.

Moreover, while the crash data were drawn from the whole of Queensland, the survey

data was collected from one particular location. Although the Brisbane Court processes

the largest number of traffic offenders in Queensland each year (Micola, 2002), it

primarily processes offenders who are detected in the inner city and suburban areas of

Brisbane. Consequently, the degree to which the findings can be generalised to other

metropolitan and rural areas remains to be confirmed.

Nonetheless, it is worth noting that there were important similarities in the findings

obtained from this program of research and previous studies, which to some degree

attests to its external validity. For example, key findings from Study One were consistent

with results cited by Job, Lee and Prabhakar (1994), Harrison (1997), FORS (1997a;b),

DeYoung et al (1997) and Griffin and DeLaZerda (2000). The reasons cited by offenders

for driving unlicensed in Study Two were very similar to those reported in other surveys

(Robinson, 1977; Ross & Gonzales, 1988; Smith & Maisey, 1990; Mirrlees-Black, 1993;

Job et al, 1994). Similarly, the links between alcohol misuse, drink driving and disqualified

driving found in Study Two are well established in the international literature (Simpson &

Mayhew, 1991; Mayhew, Simpson & Beirness, 1997). Despite these similarities, caution

should be exercised when generalising many of the other findings of the research until

they are replicated in other jurisdictions.

In addition, each of the studies featured specific limitations that need to be

acknowledged. While a more detailed discussion of these is provided in each of the

relevant chapters, the key limitations are summarised below.

Limitations of Study One


While the use of an official crash database in Study One enhanced the integrity of

the data, it did not overcome all potential problems. For example, the data is still open to

influence by the under-reporting of crashes and the subjective judgements of the police

attending crashes. Similarly, some of the unlicensed driver types were not separately

identified in the database, constraining the comparisons that could be made. Moreover,

there is a need to acknowledge the potential problems associated with the use of the

quasi- induced exposure method in Study One. This method was used to estimate the

crash risk of the unlicensed drivers due to the lack of reliable exposure data. While the

results obtained using this method were similar to those previously obtained by De

Young et al (1997), it appears to introduce a range of biases that are difficult to assess. In

this regard, Scopatz et al (2003, p.17) argue that the quasi- induced exposure method:

“has its limitations, most notably the need to establish the identity of the driver who is at

fault in fatal crashes. However, it is perhaps the best method we have now for estimating

overinvolvement that corrects for exposure, especially for unlicensed drivers”.

Over and above these considerations, the need to rely on the quasi- induced

exposure method highlights a core limitation of the crash-based research undertaken in

Study One. It was necessary to use this method because of the lack of reliable exposure

data available for unlicensed drivers. Whilst this remains the case, it will be very difficult

to estimate the community-wide prevalence of unlicensed driving and to accurately

quantify the crash risk associated with the behaviour. Accordingly, there is an ongoing

need to develop more direct methods of estimating the exposure of unlicensed drivers,

such as roadside licence checking (see section 7.6).

Limitations of Studies Two and Three

As already noted, the major limitation of Studies Two and Three was that the data

were exclusively drawn from a metropolitan setting with a bias toward offenders detected

in an inner city and suburban area. In addition, it is unclear whether the results obtained

in these studies are indicative of unlicensed drivers in general or only those that have

been detected by the police. It is possible that offenders who remain undetected are

generally more cautious (and possibly safer) than those caught by the police.

Nonetheless, it is important to acknowledge that the majority of the offenders in the

sample were detected through random enforcement processes (as opposed to committing

an offence) and many of them had been driving unlicensed for long periods of time prior

to being detected.


Study Three also featured two other important limitations. Firstly, due to the design

of the study, it was not possible to directly compare people who had driven unlicensed

with those who hadn’t. Instead, the factors contributing to unlicensed driving were

investigated using three surrogate variables, selected to reflect different aspects of the

behaviour. Hence, the validity of these findings is largely a product of the meaningfulness

of these variables. Consequently, it would be important for future research to replicate

this study with other samples of drivers, using other dependent variables as indicators of

unlicensed driving.

Secondly, the statistical methods used in Study Three were selected to examine the

factors that predict unlicensed driving behaviour and compare the predictive utility of

different theoretical perspectives, rather than explore the structural nature of these

perspectives. As such, few conclusions can be drawn about the structural nature of either

the deterrence or social learning perspectives utilised in this research. This would have

required the use of structural analytic techniques that were not directly relevant to the

research questions being examined and, hence, beyond the scope of this research.

7.6 Suggestions for future research

All three studies have highlighted a range of theoretical and applied issues that

require further research. This research is required to better explain aspects of unlicensed

driving behaviour, to establish the impact of various policies or countermeasures on the

behaviour, and to inform the design of new countermeasures.

Theoretical research issues

A thorough discussion of the theoretical research opportunities arising from Study

Three is provided in section 6.5.3. The three key research directions that emerged are

summarised below.

§ There is a need to replicate the study with other groups of drivers, including general

drivers and those who have lost their licence but have chosen to comply with the

ban. This would provide an opportunity to ascertain whether the factors that predict

unlicensed driving among offenders can also account for compliance with the law

among other drivers.

§ Further research is required to explore the influence of different types of

punishment avoidance experiences on illegal behaviour, particularly those that

involve less direct evasion of detection. This research should also examine the

mechanisms by which punishment avoidance affects sanction risk perceptions.


§ It would be useful to more specifically examine the structural nature of social

learning theory. This would provide a means of exploring how the constructs of

punishment avoidance and vicarious exposure to punishment could be integrated

within a social learning framework.

The need for better methods to estimate exposure

Study One highlighted the need to develop better methods of estimating the

exposure of unlicensed drivers. While the use of the quasi- induced exposure method

offered some advantages, it appears to introduce a range of biases that are difficult to

assess. Consequently, there is a need to evaluate the cost-effectiveness of different

methodologies for estimating the exposure of unlicensed drivers such as periodic roadside

stopping surveys, the sampling of driver’s licences at RBT and the surveillance of

unlicensed drivers (see sections 2.2.1 and 2.2.2). Improved methods are required in this

area to better estimate the crash risk associated with unlicensed driving (which will

facilitate the verification of the quasi- induced exposure method) and to act as a

benchmark for evaluating the effectiveness of future countermeasures.

Incentives for participating in the licensing system

Research is required into ways of offering more tangible incentives or rewards for

drivers to participate in the driver licensing system. This research should not only focus

on the types of incentives that could be offered, but on appropriate mechanisms for their

delivery. In addition, research is required into strategies to reduce or, at least, offset some

of the perceived benefits of driving unlicensed for work purposes (see section 7.4.2).

Underage driving

There is a need for further research into the issue of underage driving. As noted in

Study One, almost 13% of the unlicensed drivers involved in serious casualty crashes

were under the age of 17. To date, limited research has been undertaken in this area. In

Western Australia, Palamara et al (1999) found that young people who admitted driving

or riding unlicensed prior to obtaining a Learner’s Permit were three times more likely to

report drinking and driving. Not surprisingly, more underage driving was reported by

males, particularly in rural areas (Palamara, 2002). In addition, Williams (2002) has

recently cited evidence indicating a link between underage driving and opportunistic car

theft. This highlights that research into underage driving will need to adopt a broad

perspective, acknowledging the social context in which the behaviour occurs and the

variety of other risk-taking behaviours that may accompany it.


Unlicensed driving among Indigenous drivers

Although not directly addressed in this program of research, there is a growing

concern about the level of unlicensed driving in Indigenous communities (Edmonston et al,

in press). As noted in section 1.7, unlicensed and drink driving offences play an important

role in the over-representation of indigenous people in Queensland prisons. It is essential

that future research in this area is conducted in a culturally appropriate manner

(Edmonston et al, in press). Without the use of appropriate research protocols, it will be

difficult to identify the unique cultural, access and procedural barriers impacting on

indigenous people's ability to obtain and retain a driver's licence.

7.7 Concluding remarks

This program of research has questioned a number of key assumptions about

unlicensed driving. Foremost among these is the notion that unlicensed drivers drive in a

more cautious manner to avoid detection (the disqualified driver effect). While the

findings indicate that many offenders reduce their overall driving exposure in order to

avoid detection, this does not appear to result in safer driving. The data suggest that

unlicensed drivers are more likely to be involved in a crash than licensed drive rs and that

their crashes are more likely to result in a fatality or serious injury. While it remains

possible that unlicensed drivers tend to act more cautiously than they would otherwise, it

appears that their driving behaviour is primarily designed to reduce their chances of

detection and that they are not as safe on the road as licensed drivers.

The second main assumption examined by this program of research is that

unlicensed drivers represent a relatively uniform group. Once again, the findings

strongly question this assumption. Major differences were found between the different

types of unlicensed drivers in terms of their psychosocial characteristics and their on-road

behaviour. In particular, a more deviant sub-group of offenders was identified that

included the disqualified, not currently licensed and never licensed drivers. This indicates

that a multi-strategy approach is required to address the problem of unlicensed driving.

However, it should be acknowledged that reducing the level of unlicensed driving

may not automatically improve road safety. Many drivers who would otherwise drive

unlicensed may still engage in higher levels of risk-taking, irrespective of their licence

status. Nonetheless, it is likely that more effective countermeasures to unlicensed driving

should have a positive effect on road safety. For example, reducing the proportion of

drivers on the road who have never been licensed, along with those driving a vehicle

(particularly a motorcycle) with an inappropriate class of licence, should contribute to a


reduction in crashes involving inexperience. Reductions in the overall level of disqualified

and suspended driving would strengthen both the specific and general deterrent impact of

these sanctions. Finally, identifying and channelling the more deviant offenders into

rehabilitation programs offers the potential to reduce recidivism rates for both drink driving

and unlicensed driving.


References

Aitken C., Kerger M. and Crofts N. (2000). Drivers who use illicit drugs: behaviour and

perceived risks. Drugs: education, prevention and policy, 7 (1), 39-50.

Akers R.L. (1977). Deviant behaviour: A social learning approach (Second edition).

Belmont, California: Wadsworth Publishing Company.

Akers R.L. (1990). Rational choice, deterrence, and social learning theory in

criminology: The path not taken. The Journal of Criminal law and Criminology, 81

(3), 653-676.

Akers R.L. (1994). Criminological theories: Introduction and evaluation. Los Angeles:

Roxbury Publishing Company.

Akers R.L., Kron M.D., Lanza-Kaduce L. and Radosevich M. (1979). Social learning and

deviant behaviour: A specific test of a general theory. American Sociological

Review, 44, 636-655.

Akers R.L. and Lee G. (1996). A longitudinal test of social learning theory: Adolescent

smoking. Journal of Drug Issues, 26 (2), 317-342.

Anderson L.S., Chiricos T.G. and Waldo G.P. (1977). Formal and informal sanctions: A

comparison of deterrent effects. Social Problems, 25, 103-114.

Anleau S.L.R. (1995). Deviance, conformity and control. 2nd Edition. Melboune:

Longman.

Aron A. and Aron E.N. (1991). Statistics for psychology. Second edition. Upper Saddle

River, New Jersey: Prentice Hall.

ATSB (2000). Community attitudes to road safety: Community attitudes survey Wave 13,

CR197. Prepared by P. Mitchell-Taverner. Canberra: Australian Transport Safety

Bureau (ATSB).

ATSB (2002). Benchmarking road safety – the 1999 report . Canberra: Australian Transport

Safety Bureau (ATSB).

ATSB (undated). Never-licensed motorists: Fatal crash involvement of never-licensed

motorists. Monograph 1. Canberra: Australian Transport Safety Bureau (ATSB).

Bailey J.P.M. and Bailey M.L. (2000). Hard-core drink drivers in New Zealand. In Laurell

and Schlyter (Eds.), Proceedings of the 15th International Conference on Alcohol,

Drugs and Traffic Safety (T2000) (Paper No.102). Stockholm: International Council

on Alcohol, Drugs and Traffic Safety (ICADTS).


Bakker L., Ward T., Cryer M. and Hudson S.M. (1997). Out of the rut: A cognitive

behavioural treatment program for driving-while-disqualified offenders. Behaviour

Change, 14 (1), 29-38.

Bakker L., Hudson S.M. and Ward T. (2000). Reducing recidivism in driving while

disqualified: A treatment evaluation. Criminal Justice & Behaviour, 27 (5), 531-

560.

Bandura A. (1969). Principles of behaviour modification. New York: Holt, Rinehart and

Winston.

Bandura A. (1977). Social learning theory. Englewood Cliffs, N.J.: Prentice Hall.

Beck K.H., Rauch W.J. and Baker E.A. (1997). The effects of alcohol interlock license

restrictions on multiple alcohol offenders: A randomised trial in Maryland. In

C.Mercier-Guyon (Ed.), 14th International Conference on Alcohol, Drugs and

Traffic Safety - T'97, Vol.1 (pp.177-183). Annecy, France: Centre d'Etudes et de

Recherches en Medecine du Trafic.

Berger D.E. and Snortum J.R. (1986). A structural model of drinking and driving: Alcohol

consumption, social norms, and moral commitments. Criminology, 24 (1), 139-153.

Beggs W.T. and Siskind V. (1979). Motorcycle fatalities in Queensland – 1975-76.

Research Report 1/79. Brisbane: Road Safety Council of Queensland, Department of

Transport.

Beirness D.J., Simpson H.M. and Mayhew D.R. (1997). Evaluation of administrative

licence suspension and vehicle impoundment programs in Manitoba. Ottawa,

Ontario: Transport Canada.

Brown I.D. (1997). How traffic and transport systems can benefit from psychology. In T.

Rothengatter and E. Carbonell Vaya (Eds.). Traffic and transport psychology:

theory and application. Amsterdam: Pergamon.

Cameron M., Cavallo A. and Gilbert A. (1992). Crash-based evaluation of the speed

camera program in Victoria 1990-1991, Phase 1: General Effects Phase 2: Effects

of program mechanisms. Report No. 42. Melbourne: Monash University Accident

Research Centre.

Carseldine D., Court R. and Graham A. (1992). Unlicensed drivers in the northern

metropolitan area of Sydney: July-August 1991. Research Note RN 5/92. Sydney:

Roads and Traffic Authority of New South Wales.

Catchpole J.E., Macdonald W.A. and Bowland L. (1994). Young driver research program:

The influence of age-related and experience-related factors on reported driving

behaviour and crashes. CR 143. Canberra: Federal Office of Road Safety.


Clayton A. (1997). What way forward? A review of drink driving countermeasures in

selected countries world-wide. London: The Portman Group.

Cohen B.H. (1996). Explaining psychological statistics. Pacific Grove, California:

Brooks/Cole Publishing.

Collins English Dictionary (1979). Collins Dictionary of the English language. Sydney:

William Collins Sons & Co. Ltd.

Coppin R.S. and Van Oldenbeek G. (1965). Driving under suspension and revocation: A

study of suspended and revoked drivers classified as negligent operators.

Sacramento, California: Department of Motor Vehicles.

Corbett C. and Simon F. (1992). Unlawful driving behaviour: A criminological perspective.

TRL Contractor Report 301. Crowthorne, Berkshire: Transport Research Laboratory.

Davies G.P., Broughton J. Clayton A. and Tunbridge R. (1999). The high risk offender

scheme for drink drivers. TRL Contractor Report 394. Crowthorne, Berkshire:

Transport Research Laboratory.

Degenhardt L.J., Conigrave K.M., Wutzke S.E. and Saunders J.B. (2001). The validity for

an Australian modification of the AUDIT questionnaire. Drug and Alcohol Review,

20, 143-154.

DeYoung D.J. (1997a). An evaluation of the effectiveness of alcohol treatment, driver

license actions and jail terms in reducing drunk driving recidivism in California.

Addiction, 92 (8), 989-997.

DeYoung D.J. (1997b). An evaluation of the specific deterrent effect of vehicle

impoundment on suspended, revoked and unlicensed drivers in California.

Sacramento, California: California Department of Motor Vehicles.

DeYoung D.J., Peck R.C. and Helander C.J. (1997). Estimating the exposure and fatal

crash rates of suspended/revoked and unlicensed drivers in California. Accident

Analysis and Prevention, 29 (1), 17-23.

DiBlasio F.A. (1987). Predriving riders and drinking drivers. Journal of Studies on

Alcohol, 49 (1), 11-15.

DiBlasio F.A. and Benda B.B. (1990). Adolescent sexual behaviour: Multivariate

analysis of a social learning model. Journal of Adolescent Research, 5 (4), 449-466.

Diefenbach M.A., Weinstein N.D. and O’Reilly J. (1993). Scales for assessing

perceptions of health hazard susceptibility. Health Education Research, 8, 2, 181-

192.


Early Intervention Unit (1993). AUDIT: the Alcohol Use Disorders Identification Test.

Information Sheet No.1/93. Sydney: Early Intervention Unit, Centre for Drug and

Alcohol Studies, Royal Prince Alfred Hospital and The University of Sydney.

Edmonston C.J., Macaulay J., Thomas R., Mabbott N., Sheehan M. and Schonfeld S. (in

press). Australian Indigenous road safety. Canberra: Australian Transport Safety

Bureau (ATSB).

Elliott B. (1992). Achieving high levels of compliance with road safety laws: A review of

road user behaviour modification. Brisbane: Legislative Assembly of Queensland,

Parliamentary Travelsafe Committee.

Erickson M.L., Gibbs J.P. and Jensen G.F. (1977). The deterrence doctrine and the

perceived certainty of legal punishments. American Sociological Review, 42, 305-

317.

Evans L. (2000). Risks older drivers face themselves and threats they pose to other road

users. International Journal of Epidemiology, 29, 315-322.

Ferguson M., Sheehan M., Davey J. and Watson B. (1999). Drink Driving Rehabilitation:

the Present Context. Canberra: Australian Transport Safety Bureau (ATSB).

Ferguson M., Schonfeld C., Sheehan M. and Siskind S. (2000). The impact of the ‘Under

the Limit’ drink driving rehabilitation program on the lifestyle and behaviour of

offenders. CR187. Canberra: Australian Transport Safety Bureau (ATSB).

Fishbein, M., Bandura, A., Triandis, H., Kanfer, F., Becker, M. and Middlesadt, S.

(1992). Factors influencing behaviour and behaviour change: Final report.

Rockville, MD: National Institute of Mental Health.

FORS (1996). Community attitudes to road safety: Community attitudes survey Wave 9.

CR167. Canberra: Federal Office of Road Safety (FORS).

FORS (1997a). Profile of unlicensed motorists in fatal crashes. Monograph 20. Canberra:

Federal Office of Road Safety (FORS).

FORS (1997b). Road behaviour of unlicensed motorists involved in fatal crashes,

Monograph 21. Canberra: Federal Office of Road Safety (FORS).

Gibbs J. (1975). Crime, punishment and deterrence. New York: Elsevier.

Griffin L.I. and DeLaZerda S. (2000). Unlicensed to kill. Washington, DC: AAA

Foundation for Traffic Safety.

Goldberg F. (1997). Swedish Government is now field-testing the electronic driving

licence. In C.Mercier-Guyon (Ed.), 14th International Conference on Alcohol,

Drugs and Traffic Safety - T'97, Vol.2 (pp.849-853). Annecy, France: Centre

d'Etudes et de Recherches en Medecine du Trafic.


Grayson G.B. (1997). Theory and models in traffic psychology - a contrary view. In T.

Rothengatter and E. Carbonell Vaya (Eds.). Traffic and transport psychology:

theory and application. Amsterdam: Pergamon.

Haberman S.J. (1978). Analysis of qualitative data. Vol.1: Introductory topics. New

York: Academic Press.

Harano R.M. and Hubert D.E. (1974). An evaluation of California’s ‘Good driver’

incentive program. Sacremento: California Department of Motor Vehicles.

Harrison W.A. (1997). An exploratory investigation of the crash involvement of

disqualified drivers and motorcyclists. Journal of Safety Research, 28 (2), 105-111.

Harrison W.A. (1998). Applying psychology to a reluctant road safety: A comment on

South (1998). Australian Psychologist, 33, 3, 238-240.

Healy B. and Harrison W. (1986). The driving behaviour of motorists disqualified from

driving. Report RN 86/1. Melbourne: Road Traffic Authority of Victoria.

Hedlund J. and Fell J. (1995). Repeat offenders and persistent drinking drivers in the US.

In C.N. Kloeden and A.J. McLean (Eds.), 13th International Conference on

Alcohol, Drugs and Traffic Safety - T'95, Vol.2 (pp.596-604). Adelaide: NHMRC

Road Accident Research Unit, The University of Adelaide.

Homel R. (1986). Policing the drinking driver: random breath testing and the process of

deterrence. Canberra: Federal Office of Road Safety.

Homel R. (1988). Policing and punishing the drinking driver: A study of specific and

general deterrence. Springer-Verlag, New York.

Homel R. (1993). Random breath testing in Australia: Getting it to work according to

specifications. Addiction, 88, 27S-33S.

Homel R., Carseldine D. and Kearns I. (1987). Drink-drive countermeasures in Australia.

Alcohol, Drugs and Driving, 4 (2), 113-144.

Hurst P.M. (1980). Can anyone reward safe driving? Accident Analysis and Prevention,

12, 217-220.

Job S.R.F. (1999). The driver: the psychology of road safety. In J. Clark (Ed.), Safe and

mobile: Introductory studies in road safety. Armidale: University of New England

Press.

Job R.F.S., Lee V.S.H. and Prabhakar T. (1994). Unauthorised driving in New South

Wales. DLR 1. Sydney: Roads and Traffic Authority of New South Wales.

Jonah B.A. (1997). Sensation seeking and risky driving: A review and synthesis of the

literature. Accident Analysis and Prevention, 29 (5), 651-665.


Krohn M.D., Skinner W.F., Akers R.L. and Massey J.L. (1985). Social learning theory and

adolescent cigarette smoking: A longitudinal study. Social Problems, 32(5), 455-473.

LTSA (undated). Photo driver licences, mandatory licence carriage and vehicle

impoundment. Wellington, NZ: Land Transport & Safety Authority (LTSA).

Malenfant J.E.L., Van Houton R. and Jonah B. (2002). A study to measure the incidence

of driving under suspension in the Greater Moncton area. Accident Analysis &

Prevention, 34, 439-447.

Marshall P. (undated). Drug of choice and driving: Drink drive programs in Victorian

prisons. Scaffer Library of Drug Policy. Retrieved 20 June 2002, from:

http://www.druglibrary.org/schaffer/Misc/driving/s28p2.htr

Mayhew D.R., Simpson H.M. and Beirness D.J. (1997). The hard core drinking driver

revisited. In C.Mercier-Guyon (Ed.), 14th International Conference on Alcohol,

Drugs and Traffic Safety - T'97, Vol.2 (pp.791-796). Annecy, France: Centre

d'Etudes et de Recherches en Medecine du Trafic.

McCartt A.T., Geary L.L. and Berning A. (2003). Observational study of the extent of

driving while suspended for alcohol impaired driving. Injury Prevention, 9, 133-37.

McKnight A. J. and VoaS R.B. (1991). The effect of licence suspension upon DWI

recidivism. Alcohol, Drugs and Driving, 7 (1), 43-54.

Meier R.F. and Johnson W.T. (1977). Deterrence as social control: the legal and

extralegal production of conformity. American Sociological Review, 42, 292-304.

Micola R. (2002). Personal communications with Ron Micola. Registrar, Brisbane

Central Courts. December 2002.

Mirrlees-Black C. (1993). Disqualification from driving: an effective penalty? Paper 74.

London: Research and Planning Unit, Home Office.

Nichols J.L. and Ross H.L. (1990). The effectiveness of legal sanctions in dealing with

drinking Drivers. Alcohol, Drugs and Driving, 6 (2), 33-60.

NRMA (1991). The role of legal and illegal drugs in road crashes, NRMA Submission to

the NSW Standing Committee on Road Safety, Part 11. Sydney: National Roads and

Motorists' Association (NRMA).

Palamara P., Stevenson M.R., Morrison D. and Ryan G.A. (1999). Risk factors associated

with drink-driving behaviour among young Western Australian drivers.

Proceedings of the 1999 Road Safety Research, Policing and Education

Conference. (pp.655-664). Vol.2, Canberra: Australian Transport Safety Bureau.

Palamara P. (2002). Personal communication with Peter Palamara, Road Accident

Prevention Unit, University of Western Australia, July 2002.


Piquero A. and Paternoster R. (1998). An application of Stafford and Warr’s

reconceptualization of deterrence to drinking and driving. Journal of Research in

Crime and Delinquency, 35 (1), 3-39.

Piquero A.R. and Pogarsky G. (2002). Beyond Stafford and Warr’s reconceptualization

of deterrence: Personal and vicarious experiences, impulsivity and offending

behaviour. Journal of Research in Crime and Delinquency, 39 (2), 153-186.

Peck R.C. (1991). The general and specific deterrent effects of DUI sanctions: A review

of California's experience. Alcohol, Drugs and Driving, 7 (1), 13-42.

QPS (1993). Traffic Incident Reporting System User Guide. Brisbane: Information

System Branch, Queensland Police Service.

Queensland Transport (1999). Road traffic crashes in Queensland. Brisbane: Queensland

Transport.

Queensland Transport (1993). Speed management in Queensland – Issues paper.

Brisbane: Queensland Transport.

Rimmo P. and Aberg L. (1999). On the distinction between violations and errors:

sensation seeking associations. Transportation Research Part F, 2, 151-166.

Reber A.S. (1995). Dictionary of Psychology. 2nd Edition. London, England: Penguin

Books.Robinson C.D. (1977). The operation of driver licence disqualification as a

sanction. Melbourne: Department of Criminology, University of Melbourne.

Robinson C.D. and Kelso G.I. (1981). Determinants of the decision to drive while

disqualified. Law and Human Behaviour, 5 (4), 291-297.

Ross H.L. (1982). Deterring the drinking driver: Legal policy and social control.

Lexington, MA: Lexington Book.

Ross H.L. (1991). Licence deprivation as a drunk-driver sanction. Alcohol, Drugs and

Driving, 7 (1), 63-69.

Ross H.L. and Gonzales (1988). Effects of licence revocation on drunk-driving offenders.

Accident Analysis & Prevention, 20 (5), 379-391.

Ross H.L., Simin S. and Cleary J. (1995). License plate impoundment for multiple

offender drunk drivers. In C.N. Kloeden and A.J. McLean (Eds.), 13th International

Conference on Alcohol, Drugs and Traffic Safety - T'95, Vol.2 (pp.610-612).

Adelaide: NHMRC Road Accident Research Unit, The University of Adelaide.

Sadler D.D, Perrine M.W. and Peck R.C. (1991). The long-term traffic safety impact of a

pilot alcohol abuse treatment as an alternative to license suspensions. Accident



Saunders J.B., Aasland O.G., Babor T.F., de la Fuente J.R. and Grant M. (1993).

Development of the alcohol use disorders identification test (AUDIT): WHO

collaborative project on early detection of persons with harmful alcohol

consumption – II. Addiction, 88, 791-804.

Scopatz R.A., Hatch C.E., DeLucia B.H. and Tays K.A. (2003). Unlicensed to kill: The

sequel. Washington, DC: AAA Foundation for Traffic Safety.

Sheehan M., Siskind V., Schonfeld C., Ferguson M., and Davey J. (1999). Outcome

evaluation of a drink driving rehabilitation program. Handbook/Abstracts of the

Third National Conference on Injury Prevention and Control (p.35). Brisbane:

University of Queensland.

Sheehan M., Schonfeld C., Watson B., King M. and Siskind V. (2001). Developing a

Model for a Randomised Trial of Alcohol Ignition Interlocks in Queensland: Final

Report. Canberra: Australian Transport Safety Bureau.

Silcock D., Sunter A., van Lottum C. and Beuret K. (1999). Unlicensed driving: a

scoping study to identify potential areas for further research. England: AA

Foundation for Road Safety Research.

Silcock R. (2000). Research into unlicensed driving: Literature review. Report to the

Department of the Environment, Transport and the Regions. London, England:

Babtie Group Limited.

Simpson H.M. and Mayhew D.R. (1991). The hard core drinking driver. Ontario: Traffic

Injury Research Foundation of Canada.

Siskind V. (1996). Does licence disqualification reduce reoffence rates? Accident


Siskind V., Schonfeld C. and Sheehan M. (2000). Predictors of re-offence among

Australian drink drivers. In Laurell and Schlyter (Eds.), Proceedings of the 15th

International Conference on Alcohol, Drugs and Traffic Safety (T2000) (Paper

No.130). Stockholm: International Council on Alcohol, Drugs and Traffic Safety

(ICADTS).

Skinner W.F. and Fream A.M. (1997). A social learning theory analysis of computer

crime among college students. Journal of Research in Crime and Delinquency,

34(4), 495-518.

Smith D.I. (1976). Survey of riding by unlicensed motorcyclists. Report No.3. Perth:

Research & Statistics Division, Road Traffic Authority.

Smith D.I. and Maisey G.E. (1990). Survey of driving by disqualified and suspended

drivers in Western Australia. CR 94. Canberra: Federal Office of Road Safety.


Smithson M. (2000). Statistics with confidence. London: Sage Publications.

South D. (1998). General deterrence and behaviour change: A comment on the Australian

Psychological Society position paper on punishment and behaviour change.

Australian Psychologist, 33, 1, 76-78.

Stafford M.C. and Warr M. (1993). A reconceptualization of general and specific

deterrence. Journal of Research in Crime and Delinquency. 30 (2), 123-135.

Staysafe (1997). Driver licensing in New South Wales - First entry into the driver

licensing system (Staysafe 37). Report 10/51. Sydney: Parliament of New South

Wales Joint Standing Committee on Road Safety.

Stewart K., Voas R. and Taylor E. (1995). The implementation of an immobilization

sanction for impaired driving offenders. In C.N. Kloeden and A.J. McLean (Eds.),

13th International Conference on Alcohol, Drugs and Traffic Safety - T'95, Vol.2

(pp.613-616). Adelaide: NHMRC Road Accident Research Unit, The University of

Adelaide.

Strangor C. (1998). Research methods for the behavioural sciences. Boston: Houghton

Mifflin Company.

Tabachnick B.G. and Fidell L.S. (1996). Using multivariate statistics. Third Edition. New

York: HarperCollins College Publishers.

Tabachnick B.G. and Fidell L.S. (2001). Using multivariate statistics. Fourth Edition.

New York: HarperCollins College Publishers.

Tay R., Champness P. and Watson B. (2003). Personality and Speeding: Some Policy

Implications. IATSS Research, 27 (1), 68-74.

Travelsafe (1998). The road safety implications of unlicensed driving and the driving of

unregistered vehicles in Queensland. Issues Paper No.2. Brisbane: Parliamentary

Travelsafe Committee, Queensland Legislative Assembly.

Vic.RSC (1994). Inquiry into the Demerit Points Scheme. Road Safety Committee of the

Parliament of Victoria. Melbourne: L.V. North, Government Printer.

Vingilis E.R. (1990). A New Look at Deterrence. In R.J. Wilson & R.E. Mann (eds.)

Drinking and driving: Advances in research and prevention. New York: Guilford

Press.

Vingilis E.R., Mann R.E., Gavin D., Adlaf E. and Anglin L. (1990). Effects of sentence

severity on drinking driving offenders. Alcohol, Drugs and Driving, 6 (3-4), 189-197.

Voas R.B., Tippetts A.S. and Lange J.E. (1997). Evaluation of a method for reducing

unlicensed driving: The Washington and Oregon license pla te sticker laws.

Accident Analysis & Prevention, 29 (5), 627-634.


Voas R.B., Tippetts A.S. and Taylor E. (1998). Temporary vehicle impoundment in Ohio:

A replication and confirmation. Accident Analysis & Prevention, 30 (5), 651-655.

Voas R.B., Marques P.R., Tippetts A.S., and Beirness D.J. (1999). The Alberta Interlock

Program: the evaluation of a province-wide program on DUI recidivism. Addiction,

94, 1849-1859.

Voas R.B. and DeYoung D.J. (2002). Vehicle action: Effective policy for controlling

drunk and other high-risk drivers? Accident Analysis and Prevention, 34, 263-270.

Voas R.B., Blackman K.O., Tippetts A.S. and Marques P.R. (2002). Evaluation of a

program to motivate impaired driving offenders to install ignition interlocks.

Accident Analysis and Prevention, 34, 449-455.

Walitzer K.S. and Connors G.J. (1999). Treating problem drinking. Alcohol Research &

Health, 23, 2, 138-143.

Warren R.A. (1982). Rewards for unsafe driving? A rejoinder to P.M. Hurst. Accident


Watson B. (1997). The crash involvement of unlicensed drivers in Queensland.

Proceedings of the 1997 Road Safety Research and Enforcement Conference

(Session 2.3, Paper 3). Hobart: Department of Transport.

Watson B. (1998a). Factors contributing to unlicensed driving. Proceedings of the 1998

Local Government Road Safety Conference (Paper 13). Sydney: Local Government

Association of New South Wales.

Watson B. (1998b). Methodological problems associated with surveying unlicensed

drivers. Proceedings of the 1998 Road Safety Research, Policing and Education

Conference, Vol.2 (pp.98-102). Wellington, New Zealand: Land Transport Safety

Authority.

Watson B. (1998c). The effectiveness of drink driving licence actions, remedial programs

and vehicle-based sanctions. Proceedings of the 19th ARRB Research Conference

(pp.66-87). Melbourne: Australian Road Research Board.

Watson B. (1998d). Submission to Travelsafe: Unlicensed driving in Queensland.

Brisbane: Centre for Accident Research and Road Safety (CARRS-Q).

Watson B. (2000). The crash involvement of unlicensed drivers. Abstracts from 17th

Congress of the International Association of Automobile and Traffic Medicine

(AATM2000). Journal of Traffic Medicine, 28 (2S), 21.

Watson B. (2002). A survey of unlicensed driving offenders. 2002 Road Safety Research,

Policing and Education Conference Proceedings (pp.181-190). Adelaide: Transport

SA.


Watson B. (forthcoming). The road safety implications of unlicensed driving: A survey of

unlicensed drivers. Canberra: Australian Transport Safety Bureau.

Watson B., Fresta J., Whan H., McDonald J., Dray R., Bauermann C. and Churchward R.

(1996). Enhancing driver management in Queensland. Brisbane: Land Transport &

Safety Division, Queensland Transport.

Watson B. and Siskind V. (1997). The effectiveness of licence restriction for drink drivers.

Proceedings of the 1997 Road Safety Research and Enforcement Conference

(Session 1.1, Paper 4). Hobart: Department of Transport.

Watson B., Siskind V. and King M. (2000). The effectiveness of licence restriction for

drink drivers: the Australian experience. In Laurell and Schlyter (Eds.), Proceedings

of the 15th International Conference on Alcohol, Drugs and Traffic Safety (T2000)

(Paper No.408). Stockholm: International Council on Alcohol, Drugs and Traffic

Safety (ICADTS).

Weinrath M. (1997). The Ignition Interlock Program for Drunk Drivers: A Multivariate

Test. Crime & Delinquency, 43 (1), 42-59.

Wells-parker E., Bangert-Downs R., McMillen R. and Williams M. (1995). Final results

from a meta-analysis of remedial interventions with drink/drive offenders. Addiction,

90, 907-926.

Wilde G. J.S. (1985). The use of incentives for the promotion of accident- free driving.

Journal of Studies on Alcohol, Supplement, 10, 161-167.

Williams C. (2002). Masculinity, delinquency and the risky road use behaviour of young

male offenders. 2002 Road Safety Research, Policing and Education Conference

Proceedings (p. 305). Adelaide, Australia: Transport SA.

Williams R.L., Hagen R.E. and McConnell E.J. (1984). A survey of suspension and

revocation effects on the drinking-driving offender. Accident Analysis and

Prevention, 16 (5-6), 339-350.

Williamson S. (1996). The effectiveness of licence disqualification in deterring convicted

drink drivers. Unpublished Honours thesis. Hobart: Department of Geography and

Environmental Studies, University of Tasmania.

Wilson R.J. (1991). Subtypes of DWIs and high risk drivers: Implications for differential

intervention. Alcohol, Drugs and Driving, 7 (1), 1-12.

Zuckerman M. (1994). Behavioural expressions and biosocial bases of sensation-seeking.

New York: Cambridge University Press.


Zuckerman M., Eysenck S. and Eysenck H.J. (1978). Sensation seeking in England and

America: Cross-cultural, age, and sex comparisons. Journal of Consulting and

Clinical Psychology, 46 (1), 139-149.


Appendices

A Types of unlicensed driving under Queensland legislation....................... 257

B Results of semi-structured qualitative interviews conducted for Study Two/Three ....................................................................................... 259 C Proforma for recording people approached to participate in Study Two/Three ....................................................................................... 265 D Final questionnaire used for interviews in Study Two/Three.................... 267

E Interviewer’s Guide used in Study Two/Three.......................................... 285

F Summary of scales used in Studies Two and Three ................................... 289

G Reported reasons for non-participation in Study Two/Three .................... 295

H Correlation matrix for dependent and independent variables used in Study Three ................................................................................... 297 I Supplementary tables from Study Three .................................................... 303



Appendix A:

Types of Unlicensed Driving under Queensland Legislation

Under Section 78 of Queensland’s Transport Operations (Road Use Mangement)

ACT 199, unlicensed driving includes cases where:

§ a person’s driver licence has expired not more than 1 year before the offence was

committed;

§ a person’s driver licence has expired not more than 1 year but less than 5 years

before the offence was committed;

§ a person’s driver licence has expired more than 5 years before the offence was

committed;

§ a person’s authority to drive in Queensland under their non-Queensland driver

licence has been withdrawn because of residency considerations;

§ a person’s previous driver licence was cancelled because of a court disqualification

but at the end of the disqualification period they did not obtain a further driver

licence before driving;

§ a person’s driver licence had been surrendered;

§ a person holds a driver licence but is not authorised to drive, or learn to drive, the

particular class of vehicle;

§ a person’s driver licence has been cancelled because they are medically unfit to

drive safely;


licence was suspended because they were medically unfit to drive safely;

§ a person has never held a driver licence;

§ a person was disqualified from holding or obtaining a driver licence because of

unpaid fines;


licence was suspended because of unpaid fines;

§ a person was disqualified from holding or obtaining a driver licence because of the

accumulation of excess demerit points;


licence was suspended because of the accumulation of excess demerit points;

§ a person was disqualified from holding or obtaining a driver licence under a court

order.



Appendix B:

Results of semi-structured qualitative interviews conducted for Study Two/Three

Response rate: A total of 15 respondents agreed to participated from 23 eligible offenders approached ie. 65% response rate. All 8 offenders who declined to be interviewed said that they didn’t have time to participate. There were no apparent socio-economic differences among those who agreed to participate and those who declined. Sample characteristics Males = 11 Females = 4 Mean age = 22.5 years Reason for being unlicensed: Cancelled licence (due to accumulation of points) = 6 Disqualified = 2 Never licensed/not licensed = 3 Inappropriate class of licence = 2 Expired licence = 2 Prevalence of unlicensed/disqualified driving :

How much unlicensed/disqualified driving did you do before you were caught?

Ranged from 1st occasion through to 10 years (driving since 12 years of age). Places (eg. where to/where from; what roads)?

Typically, all roads in the city. Times?

Some avoided driving late at night. Purpose of trips?

Work = 7 Social/shopping = 8

What types of vehicles? (eg. car, motorcycle, truck)

Primarily cars.


Whose vehicles?

Some theirs; some friends. Passengers?

Mainly friends; some family & friends. How much unlicensed/disqualified driving did you do after you were caught? (ie.before court hearing)

Most none. One reported driving after 1st conviction, but not after being caught the 2nd time.

Perceived risk of apprehension and punishment Before you got caught for unlicensed/disqualified driving, what did you think your chances were of getting caught by the Police?

Quite variable eg. “Pretty slim”, “Below average”, “50-50, “Pretty high”. Do you think you are more likely to get caught for unlicensed/disqualified driving than: Getting caught by a speed camera? Mainly No. Getting caught for speeding by radar? ... Yes/No. Getting Random Breath Tested? Yes/No (8 respondents reported that they had either experienced or were aware of licences not being checked at RBT) Getting caught for not wearing a seat belt? Yes/No. Had you ever been caught for unlicensed/disqualified driving before (this time)? 7 had prior convictions. Before you got caught (the last time), how harsh did you think the penalties for unlicensed/disqualified driving were? Divided opinion, on this issues, including some who said they weren’t sure. How do the penalties for unlicensed/disqualified driving compare to other offences like speeding or drink driving?

Variable, but unlicensed/disqualified penalties generally seen as harsher than speeding.


Have you been caught for any other traffic offences in the past? (eg. drink driving, speeding)

12 had previous other offences, mainly for speeding. Police detection methods How do you think the Police usually catch unlicensed/disqualified drivers?

Most believe it is a random process, but that drivers can attract the attention of the Police by their behaviour or their vehicle (eg. if hotted-up or “known” to the police). A number of respondents believed that being caught was simply “bad luck”, since they were not been driving badly or unsafely.

Driving behaviour while unlicensed/disqualified Are there things you can do while driving to cut down your chances of getting caught for unlicensed/disqualified driving?

Obey the road rules; drive cautiously. Punishment avoidance Have you ever been in a situation where the Police should have found out that you were unlicensed/disqualified, but didn’t?

After prompting, 8 respondents reported being pulled over by the police and not having the licences checked. In most of these occasions they were pulled over by RBT.

Do you know of anyone else getting away with unlicensed driving when they should have been caught?

Many reported “heaps”, but it was unclear whether this related to the fact that the people evaded detection in general or were actually pulled over and not checked.

Exposure to models Do you know of many people who drive while unlicensed/disqualified?

10 respondents reported knowing people who drove unlicensed. Many of these reported knowing “heaps” who drove unlicensed. They were typically friends who drove unlicensed, but in some cases family and in one it was his “boss”. Most of the 10 reported that the people who drove unlicensed had (eventually) been caught and fined. However, some reported that these people had been able to avoid detection.


Personal attitudes (definitions) Do you think there is anything wrong with unlicensed/disqualified driving?

6 respondents reported in an unqualified manner that there was something wrong with unlicensed/disqualified driving. Among the remaining 9 respondents there was a variety of responses including: - most commonly, that it depended on the situation – for example, if the

person was driving safely and not breaking any road rules compared with driving dangerously or drink driving;

- justified by the need to get to work or survive; and - nothing wrong with it.

How important do you think it is to have a licence?

Most thought that it was very important for reasons such as: insurance; to check driving standards; to “get from A to B”. One person reported that licensing was more of a “Big brother” thing designed to “keep tabs on people”.

Normative qualities of family and friends Do your family or friends think there is anything wrong with unlicensed/disqualified driving?

Variable. Most of the respondents reported that their family do think there is something wrong with unlicensed driving. However, some reported that family did not or that they were only concerned if you got caught.

Social rewards for unlicensed/disqualified driving What good things can come from unlicensed/disqualified driving?

9 respondents reported nothing, but this may have been influenced by the experience of attending caught. The remainder mainly reported instrumental benefits of unlicensed driving, such as: “can’t have your licence suspended”; being a “lot more cautious”; “convenience”; “keep my lifestyle”; won’t lose job; and “don’t need to rely on public transport”.


How do you feel when you drive unlicensed/disqualified? (Do they get a bit of a thrill from driving unlicensed)

Most reported feeling “worried”, “scared”, or “paranoia”. Some reported feeling scared at first, but getting use to it over time and just feeling a bit apprehensive. No respondents reported getting a thrill from unlicensed driving

What do your family think of you driving unlicensed/disqualified?

Variable. Many reported that their parents did not know or that they wouldn’t be pleased eg. “my mother thinks I’m retarded”. Others reported that their family didn’t think there was anything wrong with unlicensed driving.

What do your friends think of you driving unlicensed/disqualified?

Friends were generally more indifferent about it or in a number of cases supportive.

Social punishments for unlicensed/disqualified driving What bad things can happen when you drive unlicensed/disqualified?

Many were concerned about the financial implications of being caught (eg. fines) or being in a crash (invalid insurance). A few explicitly mentioned concerns about going to court. A few also mentioned concerns about being in a crash in general.

What did your family think when you got caught?

Many had not told their family about the incident, while the remainder generally reported that their family was disappointed or angry about it. A couple said their family didn’t care all that they were unlucky.

What did your friends think when you got caught?

Most reported that their friends were sympathetic (eg. “that’s bad luck”) and, in some cases, thought that it was humorous (eg. “they just laughed”). However, some reported that they hadn’t told their friends or that they were disappointed (eg. “they said I was stupid”).


Balance of re inforcement Overall, do you think more good things come from unlicensed/disqualified driving than bad?

The majority reported that more bad things come from unlicensed driving than good. Among the four who reported that more good things could come, one specifically mentioned that it enabled them to get to work. Another one of the people who reported that more good could come from it commented that they still thought the law should be enforced.

What do you think were the main reasons you drove without a licence?

There were three main reasons reported: some claimed that they were unaware of being licensed; another group reported driving mainly for work reasons; while another group reported driving for convenience.

Attitudes/experiences to relicensing How easy do you think it would be for you to get a licence in the future?

Most reported that it would be easy. The exception was three offenders who had been Absolutely Disqualified or were serving long disqualifications. One of reported that they had experienced difficulties in passing the test in the past.

Attitudes/experiences to alternative transport How easy was it for you to use public transport while you were unlicensed?

Most reported that they found it difficult use public transport. One person mentioned that they needed to be much more organised to use public transport. However, a small group reported that public transport was easy to access.

How easy was it for you to get lifts from family and friends?

There was divided opinion on this issue. Some reported that it was relatively easy to get lifts from friends, while others reported that it was either difficult, or only generally available for short distances or difficult to organize. One person mentioned that they felt bad needing to rely on others for a lift.


Appendix C:

Proforma for recording persons approached to participate in Study Two/Three

No. Date Gender Charge (U=Unlicensed D= Disqualfiied)

Agree to participate (Yes/No)

Comments (eg. whether unaccompanied)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20



Appendix D:

Final questionnaire used for interviews in Study Two/Three

Chief investigator: Mr Barry Watson Telephone: 3864 4955 E-mail: [email protected]

Factors contributing to unlicensed driving

Information sheet

You are invited to take part in a survey about unlicensed and disqualified driving. If you agree to take part, you will be asked a range of questions about your driving and other aspects of your life. The survey is completely anonymous. You will not be asked for your name and none of your answers will be passed onto either the Police or the Court. The survey is voluntary and you are able to stop at any time if you feel uncomfortable about the questions. If you do find any of the questions distressing, please feel free to contact the QUT Counselling and Health Services, by phoning 3864 4539. They have been told about the survey and will provide you with counselling, free of charge. The survey should take about 25 minutes to complete and you will be paid $25 for taking part. If you have any questions about the survey you can contact the Chief Investigator on the phone number given above. You may also contact the secretary of the QUT Research Human Ethics Committee by phoning 3864 2902, if you have any concerns. Thank you.



UNLICENSED DRIVER INTERVIEW We’ll begin the interview now. The first few questions are just about you. I won’t be able to identify you from this information, but it will help me to see if I’m talking to a wide variety of people.

1. Gender Male…………………………………………………………….….1 Female……………………………………………………………..2

Circle - don’t ask

2. Could you look at Card A and tell me which age group you fall into? 17-20 ........................................................................................1 21-25 ........................................................................................2 26-29 ........................................................................................3 30-39 ........................................................................................4 40-49 ........................................................................................5 50-59 ........................................................................................6 60-69 ........................................................................................7 70 or more.................................................................................8

Show Card A

3. What about your marital status, are you: Single .......................................................................................1 Married.....................................................................................2 De facto/have a partner ..............................................................3 Divorced...................................................................................4 Widowed ..................................................................................5 Separated..................................................................................6

Read categories

4. What is the highest level of education you have finished? Primary.....................................................................................1 Junior (Grade 10).......................................................................2 Senior (Grade 12) ......................................................................3 TAFE/Tech College/Apprenticeship ...........................................4 CAE/University .........................................................................5 Other (Please Specify________________________________).....6

Don’t read categories Code the highest level they’ve actually completed.

5. Do you have a job at the moment? Yes...........................................................................................1 No………………………………………………………………… 2

If yes, what do you do? ___________________________________

If more than one job - ask about the position in which they work the most hours.

6. Could you look at Card B and tell me how much you earn in a year? Less than $10,000 ......................................................................1 $11,000 – $20,000 .....................................................................2 $21,000 – $30,000 .....................................................................3 $31,000 – $40,000 .....................................................................4 $41,000 – $50,000 .....................................................................5 $51,000 – $60,000 .....................................................................6 More than $60,000.....................................................................7

Show Card B

7. How long, in total, have you been driving a car or riding a motorcycle, with or without a licence?

Years _________ Months__________

Record months if less than 1 year


In this next section I would like to ask you some questions about the unlicensed driving or disqualified driving charge that has brought you to court today.

8. Were you charged with unlicensed driving or disqualified driving? Unlicensed driving…………………………………………………...1 Disqualified driving …………………………………………………2

When did the offence occur? ______________ (Record date) (If Unlicensed go to Qu.10)

After recording date of offence, go onto Qu.9 if charged with disqualified driving and Qu.10 if unlicensed

9. Why had you been disqualified from driving? (Record reason) _______________________________________________________________ What type of licence did you hold before you were disqualified? Learner’s licence....................................................................... 1 Provisional licence.................................................................... 2 Open licence............................................................................. 3 Other (Please Specify________________________________) .... 4 (Go to Qu.12)

Record reason and go to Qu.12

10. What was the reason for you being charged for unlicensed driving? Expired licence Learner’s licence had expired..................................................... 1 Provisional licence had expired.................................................. 2 Open licence had expired........................................................... 3 Cancelled licence Learner’s licence cancelled for accumulation of demerit points .... 4 Provisional licence cancelled for accumulation of demerit points . 5 Open licence cancelled for accumulation of demerit points .......... 6 Inappropriate class of licence Riding motorcycle without motorcycle licence............................ 7 Riding motorcycle >250ml with RE licence only ......................... 8 Driving manual car with automatic licence only .......................... 9 Outside conditions of a special licence Driving outside restrictions on licence.......................................10 No licence Don’t currently hold a licence...................................................11 Never held a licence.................................................................12 Other Other (Please Specify________________________________) ...13

Don’t read categories Categorise response - probe if necessary

11. At the time you were caught, did you know that you were unlicensed? Yes .......................................................................................... 1 No............................................................................................ 2 Unsure...................................................................................... 3

If no/unsure: Why? ___________________________________________________ (Record reason ___________________________________________________

Explore reasons eg. change of address


12. When you were charged with unlicensed/disqualified driving, why were you stopped by

the police? Involved in a road crash .............................................................1 Pulled over for a Random Breath Test (RBT)..............................2 Caught for speeding ...................................................................3 Caught for another offence (Please Specify________________) ...4 Pulled over for no particular reason.............................................5 Other (Please Specify________________________________).....6 Did the police check your licence at the time? Yes................................................................................. 1 (Go to Qu.13) No .................................................................................. 2

Read for rest of questions as either unlicensed or disqualified, depending on charge

If no, how did the police find out you were unlicensed/disqualified? _______________ _____________________________________________________________________

13. When you were charged with unlicensed/disqualified driving, what was your reason for driving?

Work related reasons..................................................................1 Family reasons...........................................................................2 Social/recreation activities..........................................................3 Other (Please Specify________________________________).....4

Read categories

14. When you were charged with unlicensed/disqualified driving, what type of vehicle were you driving?

Car............................................................................................1 Truck........................................................................................2 Motorcycle ................................................................................3 Bus ...........................................................................................4 Other (Please Specify________________________________).....5

Who owned the vehicle? ______________________________________________

Record owner as ‘Respondent’ if owned by interviewee

15. Were you convicted of unlicensed/disqualified driving today? Yes...........................................................................................1 No ............................................................................................2

If yes: What penalty did you receive for driving while unlicensed/disqualified? Fine: ________________ Disqualification (if applicable): _________________ (Record amount) (Record no. of months)

If no: Why weren’t you convicted? (Record reason)________________________________

Confirm conviction Record as adjourned where applicable

16. Did you know what the fine for unlicensed/disqualified driving was before you were caught?

Yes...............................................................................1 (Go to Qu.17) No ................................................................................2 Unsure ..........................................................................3

If no or unsure: Would you have still have driven while unlicensed/disqualified if you knew what the fine was?

Yes...............................................................................1 Yes, but less often..........................................................2 No ................................................................................3 Unsure ..........................................................................4


17. Were you charged/convicted of any other offences today? Yes...............................................................................1 No ................................................................................2 (Go to Qu.18)

If yes: What offences? What penalty did you receive? ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________

Read as charged if matters adjourned today or convicted if finalised Record all offences and related penalty

18. Did you have any legal representation today (ie. a solicitor)? Yes................................................................................1 (Go to Qu.19) No .................................................................................2

If no: Why? __________________________________________________________

19. Did you tell your family that you had been charged with unlicensed/disqualified driving? Yes................................................................................. 1 No .................................................................................. 2

If yes: How did they react? _______________________________________________

If no: Why? __________________________________________________________ __________________________________________________________

20. Did you tell your friends that you had been charged with unlicensed/disqualified driving? Yes................................................................................. 1 No .................................................................................. 2

If yes: How did they react? _______________________________________________

If no: Why? __________________________________________________________ __________________________________________________________

21. Have you ever been convicted of unlicensed or disqualified driving in the past? Yes....................................................................................1 No .....................................................................................2 (Go to Q.22)

If yes: How many times? ___________________________

What penalties did you receive (e.g. fines / community service)? ___________________________________________________ ___________________________________________________

22. Have you ever been convicted of any other driving relating offences in the past? Yes....................................................................................1 No .....................................................................................2 (Go to Q.23)

If yes: What offences? What penalty did you receive? ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________

Include offences that were dealt with by Traffic Offence Notice, as well as by Court


In this next section, I would like to ask some questions about the times you have driven while unlicensed/disqualified. If the participant has not had a previous conviction for unlicensed or disqualified driving, go directly to Question 25. 23. Since you started driving, how long in total would you have driven either unlicensed or while disqualified? Years __________ Months ____________ (Record amount of time)

For participants with a previous conviction

24. How long were you driving unlicensed/disqualified between your last conviction and when you were recently caught? Years __________ Months ____________ (Record amount of time) (Go to Qu.26)

For participants who have had no previous convictions for unlicensed/disqualified driving. 25. Prior to getting caught by the Police, how long had you been driving

unlicensed/disqualified? Years __________ Months ____________ (Record amount of time)

For participants with no previous conviction

For all participants 26. Prior to getting caught, how many times a week would you have driven while

unlicensed/disqualified for work-related reasons? (Count going somewhere and returning home as different trips)

Number of times __________ (Record number of trips)

Record actual number of trips, not range

27. How many times a week would you have driven while unlicensed/disqualified for family or recreational reasons? (Count going somewhere and returning home as different trips)

Number of times __________ (Record number of trips)

Record actual number of trips, not range

28. After being caught by the police (the last time, for those with prior convictions), did you continue to drive unlicensed/disqualified up until the court date?

Yes.............................................................................1 No ..............................................................................2 (Go to Qu.29) If yes: For how many months? __________________ (Record amount) How many times a week did you drive for work-related reasons? ____ (Record no. trips)

(Count going somewhere and returning home as different trips) How many times a week did you drive for family or recreational reasons?_____ (Record) (Count going somewhere and returning home as different trips)

29. While you were unlicensed/disqualified, did you need to drive as part of your job? Yes......................................................................................1 No. ......................................................................................2


30. While you were unlicensed/disqualified, were you able to get hold of a car or motorcycle

when you needed to? Yes....................................................................................... 1 No........................................................................................ 2 (Go to Q.31)

If yes: What type of vehicle(s)? Who owned the vehicle(s)? ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________

Record owner as ‘Respondent’ if owned by interviewee

31. While you were unlicensed/disqualified, did you ever give lifts to other people? Yes.........................................................................................1 No..........................................................................................2 (Go to Q.32)

If yes: Were your passengers generally aware that you were unlicensed/disqualified? Yes.........................................................................................1 No..........................................................................................2 (Go to Q.32)

Who were your passengers? _____________________________________________ ____________________________________________________________________

32. While you were unlicensed/disqualified, did you still have your photographic driver’s licence?

Yes.........................................................................................1 No..........................................................................................2 (Go to Q.33) If Yes: Why? ______________________________________________________

33. Could you use Card C to answer the next questions. Before you got caught for unlicensed/ disqualified driving, how likely did you think the following things were? Remember you can give an answer from ‘1’ Very unlikely to ‘7’ Very likely.

Very Very unlikely likely Getting caught for unlicensed/ disqualified driving 1....... 2 ...... 3 .......4....... 5 .......6....... 7 Getting Random Breath Tested 1....... 2 ...... 3 .......4....... 5 .......6....... 7 Being involved in a car crash 1....... 2 ...... 3 .......4....... 5 .......6....... 7 Getting caught if you were not wearing a seat belt 1 ....... 2 ...... 3 .......4....... 5 .......6....... 7 Getting caught if you were speeding by a speed camera 1....... 2 ...... 3 .......4....... 5 .......6....... 7 Getting caught if you were speeding by radar 1 ....... 2 ...... 3 .......4....... 5 .......6....... 7

Show Card C

34. Using Card C again, if you were to drive unlicensed/disqualified in the future, how likely do you now think your chances of getting caught are?

Very Very unlikely likely

1 ....... 2 ...... 3 .......4....... 5 .......6....... 7


The next two questions are for those participants who had held a Learner’s Licence immediately prior to driving unlicensed/disqualified. In other words, only ask for participants who were unlicensed because their Learner’s Licence had expired; it had been cancelled for accumulation of points, or they had been disqualified while on their Learners. Go to question 37 for all other participants. 35. Why did you only hold a Learner’s Licence at the time you became unlicensed /

disqualified? _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________

36. Have you ever attempted the driving test to obtain a Provisional Licence? Yes......................................................................................1 No .......................................................................................2 (Go to Q.37) If Yes: What happened (each time)? ________________________________________ _______________________________________________________________________ _______________________________________________________________________

The next questions relate to your driving while you were unlicensed/disqualified (the most recent time for those with previous convictions). 37. Could you look at Card D to answer the next two questions. While you were driving

unlicensed/ disqualified, how often did you drive at 10 km/hr or more over the speed limit?

Always Nearly Most Sometimes Just occasionally Never always (90%) occasions (20% or less) 1.................2.....................3....................4.........................5......................6

Explain that the questions relate to the most recent period of unlicensed driving for those with previous convictions. Show Card D

38. Looking at Card D again, while you were driving unlicensed/disqualified how often did you wear a seat belt?

Always Nearly Most Sometimes Just occasionally Never always (90%) occasions (20% or less) 1.................2.....................3....................4.........................5......................6

39. Looking at Card E, which of the statements best describes your drink driving behaviour while you were driving unlicensed/disqualified?

I didn't drink at any time .............................................................1 If I was driving, I didn't drink .....................................................2 If I was driving, I'd restrict what I drank......................................3 If I was driving, I did not restrict what I drank .............................4

Show Card E

40. While you were unlicensed/disqualified, did you ever drive when you thought you may have been over the legal alcohol limit?

Yes........................................................................................ 1 No ......................................................................................... 2 (Go to Q.41) If Yes: How many times? ______________ (Record number of occasions)


41. Using Card F, could you tell me if you were more careful than usual doing the following

things when you were driving unlicensed/disqualified? Remember you can give an answer from ‘1’ Much less careful to ‘7’ Much more careful.

Much less Much more careful careful Were you more careful: Obeying the speed limit 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Obeying traffic lights 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Obeying Stop and Give Way signs 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Obeying drink driving laws 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Obeying seat belt laws 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Obeying other traffic rules 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7

Show Card F

42 While you were unlicensed/disqualified, did you limit or alter your driving in anyway - for example, in terms of when or where you drove?

Yes.........................................................................................1 No..........................................................................................2 (Go to Q.43) If Yes: In what ways? _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________

43. Using Card G, could you tell me where you did most of your driving while you were unlicensed/disqualified?

Only back roads .......................................................................1 Mainly back roads ....................................................................2 Both back roads and main roads ................................................3 Mainly main roads ....................................................................4 Only main roads .......................................................................5 Other (Please Specify_________________________).................6

Show Card G

44. While you were driving unlicensed/disqualified (the last time), were you random breath tested by the Police?

Yes......................................................................................... 1 No.......................................................................................... 2 (Go to Q.45) If yes: How many times? ______________ Was your licence checked on each occasion? Yes......................................................................................... 1 No.......................................................................................... 2 If no, how many times wasn’t it checked? ___________________ What happened on the times it was checked? ________________________________ _____________________________________________________________________

Only read “the last time” for participants with previous convictions


45. While you were driving unlicensed/disqualified (the last time), were you pulled over

for speeding by the Police? Yes .....................................................................................1 No .....................................................................................2 (Go to Q.46)

If yes: How many times? ______________

Was your licence checked on each occasion? Yes......................................................................................... 1 No .......................................................................................... 2

If no, how many times wasn’t it checked? ____________________

What happened on the times it was checked? ____________________________________________________________________ ____________________________________________________________________


46. While you were driving unlicensed/disqualified (the last time), were you pulled over by the Police for another offence?

Yes...............................................................................1 No ................................................................................2 (Go to Q.47)


Was your licence checked on each occasion? Yes...............................................................................1 No ................................................................................2


What happened on the times it was checked? __________________________________________________________________ __________________________________________________________________


47. While you were driving unlicensed/disqualified (the last time), were you involved in a traffic crash?

Yes...............................................................................1 No ................................................................................2 (Go to Q.48)


Was your licence checked on each occasion? Yes...............................................................................1 No ................................................................................2



What happened on the times it was checked? ____________________________________________________________________ ____________________________________________________________________

48. While you were driving unlicensed/disqualified (the last time), did you get caught for speeding by a speed camera?

Yes...............................................................................1 No ................................................................................2 (Go to Q.49)


What happened as a result of the tickets? __________________________________ ____________________________________________________________________ ____________________________________________________________________



In this section, I would like to ask you some general questions about other people’s driving?

49 Has any of your family or friends ever driven while unlicensed or disqualified? Yes.............................................................................1 No..............................................................................2 (Go to Q.50)

If yes: How many? ______________________________

How many of these people have been convicted of unlicensed or disqualified driving? ___

What penalties did they receive (e.g. fines / community service)? __________________________________________________________ __________________________________________________________

How many of your family and friends currently drive unlicensed or disqualified? _______

50. Do you know of anyone else who has driven while unlicensed or disqualified at sometime? Yes.............................................................................1 No..............................................................................2 (Go to Q.51)

If yes: How many? ______________________________

How many of these people have been convicted of unlicensed or disqualified driving? ___ What penalties did they receive (e.g. fines / community service)? __________________________________________________________ __________________________________________________________

How many of these people currently drive unlicensed or disqualified? _____________

51. Do you know of anyone who was pulled over by the police when they were driving unlicensed or disqualified and didn’t have their licence checked?

Yes......................................................................................... 1 No.......................................................................................... 2 (Go to Q.52)

If yes: What were the circumstances? ________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________ _______________________________________________________________________

52. The following statements are about unlicensed/disqualified driving. Using Card H, could you te ll me how much you agree or disagree with each statement? Remember you can give an answer from ‘1’ Strongly Disagree to ‘7’ Strongly Agree. There is no right or wrong answer. Strongly Strongly disagree agree The penalties for unlicensed/disqualif ied driving are very tough 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 It’s OK to drive unlicensed/disqualified as long as you don’t get caught 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Everybody drives unlicensed once in a while 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Your family would think you were really Stupid for driving unlicensed/disqualified 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7 Your friends would think you were really stupid for driving unlicensed/disqualified 1 .......2....... 3 ...... 4 ....... 5 .......6....... 7

Show Card H Emphasise bolded words


There is no excuse for unlicensed/disqualified driving 1........2.......3....... 4 .......5....... 6 .......7 Your family doesn’t care about unlicensed/ disqualified, as long as you don’t get caught 1........2.......3....... 4 .......5....... 6 .......7 Your friends don’t care about unlicensed/ disqualified, as long as you don’t get caught 1........2.......3....... 4 .......5....... 6 .......7 You can sometimes avoid getting punished if you get caught for unlicensed/disqualified driving 1........2.......3....... 4 .......5....... 6 .......7 The Police spend too much time hassling unlicensed/disqualified drivers 1........2.......3....... 4 .......5....... 6 .......7 Unlicensed/disqualified drivers are generally more careful on the road 1........2.......3....... 4 .......5....... 6 .......7 We need tougher penalties for unlicensed/ disqualified driving 1........2.......3....... 4 .......5....... 6 .......7 Most of your family think it’s OK to drive unlicensed or disqualified 1........2.......3....... 4 .......5....... 6 .......7 Most of your friends think it’s OK to drive unlicensed or disqualified 1........2.......3....... 4 .......5....... 6 .......7 You are likely to be punished quickly if you get caught for unlicensed/disqualified driving 1........2.......3....... 4 .......5....... 6 .......7 You were lucky not to have got caught earlier for driving unlicensed/disqualified 1........2.......3....... 4 .......5....... 6 .......7 You feel guilty about driving unlicensed/ Disqualified 1........2.......3....... 4 .......5....... 6 .......7 The chances of being caught for unlicensed/ disqualified driving are over-rated 1........2.......3....... 4 .......5....... 6 .......7

It’s OK to drive unlicensed/disqualified as long as you don’t do it too much 1........2.......3....... 4 .......5....... 6 .......7 Unlicensed/disqualified driving gives you the freedom to lead a normal life 1........2.......3....... 4 .......5....... 6 .......7 Most of your family would praise you for driving unlicensed/disqualified 1........2.......3....... 4 .......5....... 6 .......7 Most of your friends would praise you for driving unlicensed/disqualified 1........2.......3....... 4 .......5....... 6 .......7 Unlicensed/disqualified driving gives you a thrill 1........2.......3....... 4 .......5....... 6 .......7 It would be easy for you to obtain a valid driver’s licence in the future 1........2.......3....... 4 .......5....... 6 .......7


Unlicensed/disqualified driving gives you a feeling of power 1........2.......3....... 4 .......5....... 6 .......7

Unlicensed/disqualified driving is generally not worth the risks 1........2.......3....... 4 .......5....... 6 .......7

You wouldn’t like your workmates to know you had been driving without a licence 1........2.......3....... 4 .......5....... 6 .......7

Unlicensed/disqualified driving generally makes you feel good 1........2.......3....... 4 .......5....... 6 .......7

Unlicensed/disqualified driving is potentially dangerous 1........2.......3....... 4 .......5....... 6 .......7

Unlicensed/disqualified driving gives you a feeling of being in control 1........2.......3....... 4 .......5....... 6 .......7

Unlicensed/disqualified driving makes you feel worried 1........2.......3....... 4 .......5....... 6 .......7

You could lose your job if your boss found out that you had been driving without a licence 1........2.......3....... 4 .......5....... 6 .......7

Overall, more good things are likely to come from unlicensed/disqualified driving than bad 1........2.......3....... 4 .......5....... 6 .......7

53. The following statements are about driving in general. Using Card H, could you tell me how much you agree or disagree with each statement? Remember you can give an answer from ‘1’ Strongly disagree to ‘7’ Strongly agree.

Strongly Strongly disagree agree

The Police generally check driver’s licences when they conduct RBT 1........2.......3....... 4 .......5....... 6 .......7

It is really important to have a valid driver’s licence 1........2.......3....... 4 .......5....... 6 .......7

It costs too much money to obtain a driver’s licence 1........2.......3....... 4 .......5....... 6 .......7

You find it possible to do most things by using public transport 1........2.......3....... 4 .......5....... 6 .......7

You can generally get a lift from family or friends when you need one 1........2.......3....... 4 .......5....... 6 .......7

The whole process of getting and renewing a driver’s licence is a hassle 1........2.......3....... 4 .......5....... 6 .......7

The Police generally check driver’s licences when they pull someone over 1........2.......3....... 4 .......5....... 6 .......7

There is not much public transport available in the area where you live 1........2.......3....... 4 .......5....... 6 .......7

You can’t always rely on your family or friends for lifts 1........2.......3....... 4 .......5....... 6 .......7 It is easy to get out of a speed camera ticket 1........2.......3....... 4 .......5....... 6 .......7 It costs too much to use taxis regularly 1........2.......3....... 4 .......5....... 6 .......7 You could get by without driving if you really had to 1........2.......3....... 4 .......5....... 6 .......7

Show Card H


54 The following statements are about things you like to do in your general life. For each of

the following pairs of statements, please choose the one that most describes your likes or the way you feel. We are interested only in your likes or feelings, not in how others feel or how one is supposed to feel. There is no right or wrong answers as with the previous questions.

a. I often wish I could be a mountain climber .............................................................. 1 b. I can’t understand people who risk their necks climbing mountains ........................... 2

a. A sensible person avoids activities that are dangerous .............................................. 1 b. I sometimes like to do things which are a little frightening ....................................... 2

a. I would like to take up the sport of water skiing....................................................... 1 b. I would not like to take up water skiing................................................................... 2

a. I would like to try surfboard riding ......................................................................... 1 b. I would not like to try surfboard riding .................................................................... 2

a. I would not like to learn to fly an airplane ............................................................... 1 b. I would like to learn to fly an airplane ..................................................................... 2

a. I prefer the surface of the water to the depths........................................................... 1 b. I would like to go scuba diving............................................................................... 2

a. I would like to try parachute jumping ...................................................................... 1 b. I would never want to try jumping out of a plane with or without a parachute ............ 2

a. I would like to dive off the high board..................................................................... 1 b. I don’t like the feeling I get standing on the high board or I don’t go near it at all....... 2

a. Sailing long distances in small sailing crafts is foolish.............................................. 1 b. I would like to sail a long distance in a small but seaworthy sailing craft................... 2

a. Skiing fast down a high mountain slope is a good way to end up on crutches............. 1 b. I think I would enjoy the sensations of skiing very fast down a high mountain slope .. 2

Circle the statement in each pair selected by the respondent

In this next section, I would like to ask some general questions about your drinking behaviour. 55. Looking at Card I, how often do you have a drink containing alcohol? Never...................................................................................1 (Go to Q.60) Monthly or less.....................................................................2 2 to 4 times a month..............................................................3 2 to 3 times a week ...............................................................4 4 or more times a week..........................................................5

Show Card I

Can you look at Card J. This tells you what a standard drink is. It basically just says that a standard drink is a pot of beer, or a nip of spirits, or a glass of wine, or a can of light beer, or a glass of port. I would now like to ask you some personal questions about your drinking and would appreciate your honesty. 56. Looking at Card J, how many ‘standard’ drinks containing alcohol do you have on a

typical day when you drink? 1 or 2 ...................................................................................... 1 3 or 4 ...................................................................................... 2 5 or 6 ...................................................................................... 3 7 to 9....................................................................................... 4 10 or more............................................................................... 5

Show Card J


57. For the next few questions I would like you to use Card K. This card has answer categories ranging from never to less than monthly, monthly, weekly, and daily or almost daily.

Never Less Than Monthly Weekly Daily Or Monthly Almost Daily How often do you have six or more drinks on one occasion? 1.............. 2 ............3..........4............5

How often during the last year have you found that you were not able to stop drinking once you had started? 1.............. 2 ............3..........4............5

How often during the last year have you failed to do what was normally expected from you because of drinking? 1.............. 2 ............3..........4............5

How often during the last year have you needed a drink in the morning to get yourself going after a heavy drinking session? 1 2 3 4 5

How often during the last year have you had a feeling of guilt or remorse after drinking? 1.............. 2 ............3..........4............5

How often during the last year have you been unable to remember what happened the night before because you had been drinking? 1.............. 2 ............3..........4............5

Show Card K

58. Have you or someone else been injured as a result of your drinking? Yes......................................................................................1 No .......................................................................................2 (Go to Q.59)

If yes: Was it in the last year? Yes (during last year)............................................................1 No (not in last year) ..............................................................2

Prompt for yes/no response

59. Has a relative, a friend, a doctor or other health worker been concerned about your drinking or suggested you cut down?

Yes......................................................................................1 No .......................................................................................2 (Go to Q.60)

If yes: Was it in the last year? Yes (during last year)............................................................1 No (not in last year) ..............................................................2

I would now like to ask you a question about the future.

60. Using Card L, could you tell me how likely it is that you will drive without a licence sometime in the future?

Very Very unlikely likely

1........2.......3....... 4 .......5....... 6 .......7 Why? (Probe if necessary) ___________________________________________________________________ ___________________________________________________________________ ___________________________________________________________________ ___________________________________________________________________

Show Card L


This is the last question I have today. 61. Have you ever been convicted of any criminal offences (other than driving offences)? Yes .......................................................................................... 1 No............................................................................................ 2 If yes: What offences? What penalty did you receive? ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________ ___________________________ ____________________________________

Thank you for participating in the survey.

Record the date of interview: __________________



Appendix E:

Interviewer’s Guide used in Study Two/Three

General information Recruitment process The interviewers will approach people outside the courts and explain that they are conducting an anonymous, voluntary survey on the topic of unlicensed driving. In particular, they will ask whether the person is attending court in relation to a charge of unlicensed driving or disqualified driving. In some cases, it may be necessary to clarify whether unlicensed driving was one of a number of charges, including drink driving etc. Only people who appear to understand English should be requested to participate in the study. (If it later becomes apparent that the participant is unable to understand the questions or communicate their responses, the interview should be terminated.) Once potential participants are identified, they should be given a brief explanation of the survey and offered a $25 incentive payment to participate in the study. If they agree, they should be given the Information sheet and ‘walked’ through its contents. The interview should be commenced once the person consents to participate. Interaction with Court staff and processes Permission has been obtained from the Acting Registrar of the Brisbane Court to conduct the survey. This approval is conditional on all interviews/surveys being conducted in a manner that “does not interfere with the normal workings of the court, case flow management and wishes of the persons attending court and/or their legal representatives”. Consequently, in the event that a participant has commenced an interview prior to their hearing but is required to attend to court business, the interviewer will immediately terminate the interview and only continue after the participant becomes available (eg. after their hearing). In addition, the interviewers will need to: § carry a copy of the letter from the Acting Registrar that confirms that they have

approval to conduct the survey; and § avoid providing any legal advice or opinions to the participants (or to appear to be

doing so to others). Managing participants Interviews should be discontinued if participants become distressed or aggressive. The availability of QUT’s Counselling Service should be reiterated to any participant who appears distressed by he survey questions or process. Any problems experienced with participants should be brought to the attention of the Chief Investigator.


Recording details of offenders approached to participate Brief details should be recorded for all eligible offenders (ie. those charged with Unlicensed or Disqualified Driving) approached to participate in the study. This information should be recorded on the form entitled: “Unlicensed Driving Survey (Quantitative): Record of persons approached to participate”. Most importantly, record whether the offender was charged with Unlicensed or Disqualified Driving and whether they agree to participate in the study. This information will be used to determine the overall response rate for the survey. In the comments section, record any issues of interest relating to the offender. In addition, please record whether the offender was accompanied by someone else when approached, since this sometimes has an influence on whether they agree to participate. Specific information Where relevant, special instructions for the questions are provided on the interview form (in the right column). Based on experience to date, some issues requiring special attention are detailed below. Qu.7 Record the total time period that the participants have been driving a car or riding a motorcycle, counting periods of both licensed and unlicensed driving. Qu.8 Record whether they were charged with unlicensed driving or disqualified driving and remember to record when the offence occurred (or at least how long ago if the participant is unsure of the exact date). Qu.9 Remember to record the reason why the participant had been disqualified from driving eg. drink driving offence, unlicensed driving offence or dangerous driving. Qu.10 Some probing may be required to find out the exact reason why the participant was driving unlicensed. In cases where they had failed to obtain a new licence after a period of licence disqualification, classify them as ‘Don’t currently hold a licence’. Qu.12 onwards From Question 12 onwards, all questions referring to ‘unlicensed/disqualified’ driving should be read in a way that reflects the offence the participant was charged with. In other words, if they were charged with Unlicensed Driving then the ‘unlicensed/disqualified’ questions should be read as ‘unlicensed’ only. In contrast, the questions should be read as ‘disqualified’ if the participant was charged with Disqualified Driving.


Qu.14 Remember to record who owned the vehicle that the offender was driving when they were caught and charged. Record the owner as ‘Respondent’ if the vehicle is owned by the person being interviewed. Qu.15 Record the fine received by the participant, along with any licence disqualification they received. In some cases, offenders are given a combined fine covering a number of offences, including the unlicensed driving offence. If the specific fine for the unlicensed driving offence is not known, record the total fine but note that it covers other offences as well. This should also be noted in Question17 where the other relevant offences will be listed. Qu.s 19 & 20 Record the reactions of family and friends verbatim, if reported. Qu.s 23, 24 & 25 Questions 23 and 24 should only be completed for those participants who have had a previous conviction for unlicensed driving. For those participants without a previous conviction, Question 25 should be used. Qu.s 26, 27 & 28 These questions concern the number of times the participants drove per week (for different reasons) while unlicensed. The number of times should be interpreted in terms of the number of trips undertaken during the week. As noted on the questionnaire, the participants should count going somewhere and returning home as different trips. It is important to record the estimated number of trips, rather than a range. Hence, prompt for a number if the participants initially provide a range. Qu.s 35 & 36 These questions are only applicable for those participants who only held a Learner’s Licence immediately prior to the period when they drove unlicensed or disqualified. Qu.s 37 to 43 These questions relate to the period during which the participant drove unlicensed or disqualified. For those with previous convictions, it should be explained that they relate to the most recent period of unlicensed/disqualified driving. The questions are going to be less relevant for those participants who were unaware that they were unlicensed at the time they were caught. This is particularly the case for Questions 41 and 42. These participants should still be encouraged to complete all the questions. However, Questions 41 and 42 can be omitted for those participants who have obvious difficulties answering them.


Qu.s 44 to 48 These questions also relate to the most recent period during which the participant drove unlicensed or disqualified. Therefore, for those with previous convictions, the words ‘the last time’ should be incorporated into the questions where shown. Qu.s 52 & 53 On occasions, participants may provide an answer to one of the items that does not appear consistent with an earlier response. If you suspect that the participant has misunderstood the question or confused the direction of their response, repeat the question. However, the participant should not be prompted to change their response. Qu.60 Probe for the main reasons that will either encourage or discourage the participant from driving without a licence sometime in the future. Try to record the participant’s response verbatim. After completing interview, thank the participant for participating and record the date of the interview. Also remember to fill out the “Unlicensed Driving Survey (Quantitative): Record of persons approached to participate” form.


Appendix F:

Summary of Scales used in Studies Two and Three

F1. Thrill and Adventure subscale of Sensation Seeking Scale (Form V)

Item: For each of the following pairs of statements, please choose the one that

most describes your likes or the way you feel. (Sensation seeking option shown in

italic)

a. I often wish I could be a mountain climber

b. I can’t understand people who risk their necks climbing mountains

a. A sensible person avoids activities that are dangerous

b. I sometimes like to do things which are a little frightening

a. I would like to take up the sport of water skiing

b. I would not like to take up water skiing

a. I would like to try surfboard riding

b. I would not like to try surfboard riding

a. I would not like to learn to fly an airplane

b. I would like to learn to fly an airplane

a. I prefer the surface of the water to the depths

b. I would like to go scuba diving

a. I would like to try parachute jumping

b. I would never want to try jumping out of a plane with or without parachute

a. I would like to dive off the high board

b. I don’t like the feeling I get standing on the high board or I don’t go near it

at all

a. Sailing long distances in small sailing crafts is foolish

b. I would like to sail a long distance in a small but seaworthy sailing craft

a. Skiing fast down a high mountain slope is a good way to end up on crutches

b. I think I would enjoy the sensations of skiing very fast down a high

mountain slope

Cronbach’s alpha = .71


F2. The Alcohol Use Disorders Identification Test (AUDIT)

The test consists of ten items. The first eight are measured on a 5-point scale

(Never, monthly or less, 2 to 4 times a month, 2 to 3 times a week, 4 or more

times a week – scored 0 to 4)

How often do you have six or more drinks on one occasion?

How often during the last year have you found that you were not able to stop

drinking once you had started?

How often during the last year have you failed to do what was normally

expected from you because of drinking?

How often during the last year have you needed a drink in the morning to get

yourself going after a heavy drinking session?

How often during the last year have you had a feeling of guilt or remorse

after drinking?

How often during the last year have you been unable to remember what

happened the night before because you had been drinking?

The following two times are measured using the following response categories:

No; Yes, but not in last year; and Yes, during the last year (scored 0,2 and 4)

Have you or someone else been injured as a result of your drinking?

Has a relative, a friend, a doctor or other health worker been concerned

about your drinking or suggested you cut down?


F3. Cautiousness while driving scale

Items: . . . were more careful than usual doing the following things when you

were driving unlicensed/disqualified? (Measured on 7-point Likert scale)

Obeying the speed limit

Obeying traffic lights

Obeying Stop and Give Way signs

Obeying drink driving laws

Obeying seat belt laws

Obeying other traffic rules



F4. Differential association (normative dimension) scale

Items: . . . how much do you agree or disagree with each statement?

(Measured on 7-point Likert scale)

Most of your family think it’s OK to drive unlicensed/disqualified

Most of your friends think it’s OK to drive unlicensed/disqualified

Your family doesn’t care about unlicensed disqualified, as long as you don’t

get caught

Your friends don’t care about unlicensed/disqualified, as long as you don’t

get caught


F5. Attitudes to unlicensed driving scale



The police spend too much time hassling unlicensed/disqualified drivers

Unlicensed/disqualified driving gives you the freedom to lead a normal life

It costs too much money to obtain a driver’s licence

The whole process of getting and renewing a driver’s licence is a hassle

Everybody drives unlicensed once in a while

Unlicensed/disqualified driving is generally not worth the risks (reversed

scored)

Its OK to drive unlicensed/disqualified as long as you don’t get caught

Its OK to drive unlicensed/disqualified as long as you don’t do it too much

There is no excuse for unlicensed/disqualified driving (reversed scored)

We need tougher penalties for unlicensed/disqualified driving

(reversed scored)

Unlicensed/disqualified driving is potentially dangerous (reversed scored)

It is really important to have a valid driver’s licence (reversed scored)



F6. Attitudes to alternative transport scale



You find it possible to do most things by using public transport

You can generally get a lift from family of friends when you need one

There is not much public transport available in the area where you live

(reverse scored)

You can’t always rely on your family or friends for lifts (reverse scored)

You could get by without driving if you really had to


F7. Rewards for unlicensed driving scale



Most of your family would praise you for driving unlicensed/disqualified

Most of your friends would praise you for driving unlicensed/disqualified

Unlicensed/disqualified driving gives you a thrill

Unlicensed/disqualified driving gives you a feeling of power

Unlicensed/disqualified driving gives you a feeling of being in control

Unlicensed/disqualified driving generally makes you feel good


F8. Punishments for unlicensed driving scale



Your family would think you were really stupid for driving

unlicensed/disqualified

Your friends would think you were really stupid for driving

unlicensed/disqualified


You wouldn’t like your workmates to know you had been driving without a

licence

You could lose your job if your boss found out that you had been driving

without a licence

The penalties for unlicensed driving are very tough

You can sometimes avoid getting punished if you get caught for

unlicensed/disqualified driving (reversed scored)

You are likely to be punished quickly if you get caught for

unlicensed/disqualified driving

You feel guilty about driving unlicensed/disqualified

Unlicensed/disqualified driving makes you feel worried

Unlicensed driving is potentially dangerous

Unlicensed/disqualified driving is generally not worth the risks




Appendix G:

Reported reasons for non-participation in Study Two/Three

Table G1

Reported reasons for non-participation in Study Two/Three

Reported reason No. %

No reason provided 110 59.1

In a rush/no time 35 18.8

Work-related commitment 14 7.5

Other commitment 9 4.8

Upset/angry 7 3.7

Not interested 5 2.9

Language/literacy problems 3 1.6

Accompanied by children 2 1.1

Claimed they were not guilty 1 0.5

Total 186 100.0



Appendix H:

Correlation matrix for dependent and independent variables used in Study Three



Table H1 Correlation matrix of dependent and independent variables used in Study Three (Page 1) Variable 1 2 3 4 5 6 7 8 9 10 11 12

Frequency of unlicensed driving1 1 1.00 .30*** .20*** .02 -.03 -.04 .09 .12* .28*** .11 .10 .08

Continued driving after detection 2 1.00 .45*** -.01 -.13* -.09 -.04 -.09 .09 -.06 .02 .03

Intention to drive unlicensed in the future1 3 1.00 .10 -.12* -.12* -.08 .06 .07 .00 .09 .02

Gender 4 1.00 -.08 .07 -.22*** .08 .11 .08 .16** .29***

Age2 5 1.00 .11 -.04 -.07 .02 .18** .07 -.17**

Marital status 6 1.00 -.03 .07 .03 .17** .00 -.04

Educational level 7 1.00 .11 .02 .13* -.37*** .07

Employed at time of court hearing 8 1.00 .31*** .51*** -.17** .11

Needed to drive for work 9 1.00 .29*** -.01 .01

Annual income2 10 1.00 -.13* .05

Prior criminal conviction 11 1.00 .11*

Sensation seeking score 12 1.00

AUDIT score 13

Unaware of being unlicensed 14

Able to access vehicle 15

Owned a vehicle 16

Still had photographic licence 17

Perceived risk of apprehension (prior) 18

Perceived risk of apprehension (after)

19

Knew what fine for unlicensed driving was 20

Perceived severity of punishment

21

Perceived certainty of punishment 22

Perceived swiftness of punishment 23

Prior conviction for unlicensed driving 24

Exposure to enforcement 25

Punishment avoidance 26

Vicarious exposure to punishment 27

Vicarious exposure to punishment avoidance 28

Imitation 29

Differential association Behavioural dimension

30

Differential association Normative dimension 31

Attitudes to unlicensed driving

32

Attitudes to alternatives 33

Anticipated rewards 34

Anticipated punishments 35

1. Logarithmically transformed 2. Correlations calculated usin g midpoints of categories * p < .05 ** p < .01 *** p < .001


Table H1 Correlation matrix of dependent and independent variables used in Study Three (Page 2) Variable 13 14 15 16 17 18 19 20 21 22 23 24

Frequency of unlicensed driving1 1 .07 .17** .17** .15** .22*** -.13* -.16** -.04 -.07 -.04 -.07 -.03

Continued driving after detection 2 .03 -.10 .11 -.03 -.09 -.02 -.17** -.01 .06 -.13* -.04 .04

Intention to drive unlicensed in the future1 3 .04 -.24*** .12* -.07 -.17** .00 -.18** .06 -.01 -.07 -.09 .20***

Gender 4 .22 -.05 -.01 -.08 -.11 .05 -.04 .04 -.04 -.06 .03 .14*

Age2 5 -.13* .04 .04 .05 .06 .04 .12* -.03 .05 .05 .10 .05

Marital status 6 .03 .03 -.03 -.02 .00 .02 .04 -.05 .01 -.01 .02 .05

Educational level 7 -.11 .18 .08 .14* .20*** -.15** -.15** -.08 -.05 .01 -.13* -.35***

Employed at time of court hearing 8 -.04 .09 .10 .11 .17** -.08 -.03 -.06 .02 -.02 .05 .07

Needed to drive for work 9 .00 .16** .06 .14* .18** -.03 .04 -.08 .04 -.06 -.01 .00

Annual income2 10 -.04 .21*** .06 .14* .21*** -.08 .00 -.06 .09 .03 .05 -.09

Prior criminal conviction 11 .22*** -.07 .01 -.06 -.16** .02 .07 .03 .02 .00 .03 .29***

Sensation seeking score 12 .11 -.05 .14* .04 -.02 .00 -.07 -.04 -.09 .02 -.03 .02

AUDIT score 13 1.00 -.25*** -.14* -.21*** -.10 .13* .12* .06 -.03 .00 .06 .17**

Unaware of being unlicensed 14 1.00 .12* .16** .49*** -.25*** -.13* -.17** -.01 .04 -.03 -.27***

Able to access vehicle 15 1.00 .25*** .14* -.05 -.08 -.09 -.02 -.02 -.10 .05

Owned a vehicle 16 1.00 .18** -.01 .03 -.11 .02 .07 -.06 -.15*

Still had photographic licence 17 1.00 -.12* -.12* -.17** -.06 -.01 -.10 -.28**

Perceived risk of apprehension (prior) 18 1.00 .24*** .12* .01 .07 -.05 .17**

Perceived risk of apprehension (after) 19 1.00 .05 .04 .09 .17** .08

Knew what fine for unlicensed driving was 20 1.00 .01 -.05 .01 .25***

Perceived severity of punishment 21 1.00 .06 .14* .08

Perceived certainty of punishment 22 1.00 .13* -.02

Perceived swiftness of punishment

23 1.00 -.01

Prior conviction for unlicensed driving 24 1.00

Exposure to enforcement 25

Punishment avoidance 26

Vicarious exposure to punishment 27

Vicarious exposure to punishment avoidance 28

Imitation 29 Differential association Behavioural dimension 30

Differential association Normative dimension 31

Attitudes to unlicensed driving 32

Attitudes to alternatives 33

Anticipated rewards 34

Anticipated punishments 35

1. Logarithmically transformed 2. Correlations calculated using midpoints of categories * p < .05 ** p < .01 *** p < .001


Table H1 Correlation matrix of dependent and independent variables used in Study Three (Page 3) Variable 25 26 27 28 29 30 31 32 33 34 35

Frequency of unlicensed driving1 1 -.14* .31*** .09 .02 .03 .10 .15* .10 -.22*** .06 -.15**

Continued driving after detection 2 .02 .09 .20*** .08 .07 .20*** .27*** .27*** -.16** .19** -.20***

Intention to drive unlicensed in the future1 3 -.05 .17** .18** .14* .23*** .27*** .38*** .48*** -.22*** .17** -.38***

Gender 4 -.06 .14* .13* .08 -.06 .10 .05 .01 .16** -.02

Age2 5 -.04 -.01 -.09 -.05 .05 -.11* -.18** -.14* -.03 -.17** .16**

Marital status 6 .02 .00 .09 .04 .04 .01 -.10 -.08 .01 -.03 .07

Educational level 7 .07 -.08 -.12* -.12* -.18** -.06 -.11 -.07 -.05 -.22*** -.09

Employed at time of court hearing 8 .01 .04 .02 -.04 -.03 -.06 -.06 -.05 -.05 -.14* -.06

Needed to drive for work 9 -.08 .14* -.04 -.04 -.03 .03 -.12* -.06 -.22*** -.11 .11

Annual income2 10 -.01 .05 .03 -.08 .00 -.09 -.15* -.17** -.12* -.23*** .07

Prior criminal conviction 11 -.05 .07 .05 .12* .18** .10 .08 .14* -.07 .11 .03

Sensation seeking score 12 .06 .00 .05 .05 -.01 .08 .09 .12* -.03 .06 -.09

AUDIT score 13 -.02 .01 .12* .10 .18** .12* .15** .05 .09 .21*** .03

Unaware of being unlicensed 14 .06 -.01 -.18** -.21*** -.26*** -.19** -.18** -.20*** -.05 -.16** .11

Able to access vehicle 15 .06 .01 .00 .00 .00 .01 .09 .17** -.02 .07 -.19**

Owned a vehicle 16 .10 -.07 -.18** -.06 -.17** -.17** -.06 -.03 -.09 -.15** .07

Still had photographic licence 17 .01 .03 -.18** -.11 -.13 -.25 -.08 -.06 -.04 -.12* -.02

Perceived risk of apprehension (prior) 18 .00 .01 .00 .02 .01 .07 .03 -.03 .09 -.04 .03

Perceived risk of apprehension (after) 19 .02 -.11 -.10 .06 -.01 -.06 -.12* -.21*** .09 -.08 .23***

Knew what fine for unlicensed driving was 20 .00 .01 .11 .12* .06 .06 -.04 -.02 .04 .00 -.03

Perceived severity of punishment 21 -.03 -.01 -.01 -.03 .00 -.01 .01 .09 -.12* -.10 .25***

Perceived certainty of punishment 22 .01 -.08 -.01 -.06 -.14* -.17** -.13* -.11 -.03 -.13* .26***

Perceived swiftness of punishment 23 -.06 .01 .03 -.04 .07 -.04 -.05 -.13* .06 -.05 .40***

Prior conviction for unlicensed driving 24 -.12* .05 .15** .12* .09 .13* .06 .15** -.03 .17** -.02

Exposure to enforcement 25 1.00 -.46*** .02 -.04 -.03 -.07 -.02 -.03 .03 -.06 .00

Punishment avoidance 26 1.00 .10 .07 .06 .11 .06 .11 -.15** .03 -.12*

Vicarious exposure to punishment

27 1.00 .24*** .29*** .43*** .22*** .16** -.04 .21** -.10

Vicarious exposure to punishment avoidance 28 1.00 .33*** .20*** .22*** .18** -.04 .19*** -.15**

Imitation 29 1.00 .35*** .23*** .17** -.09 .16** -.05

Differential association Behavioural dimension 30 1.00 .18** .19** -.02 .22*** -.14*

Differential association Normative dimension 31 1.00 .58*** -.08 .44*** -.46***

Attitudes to unlicensed driving 32 1.00 -.28*** .34*** -.55***

Attitudes to alternatives 33 1.00 .03 .06

Anticipated rewards 34 1.00 -.31***

Anticipated punishments 35 1.00

1. Logarithmically transformed 2. Correlations calculated using midpoints of categories * p < .05 ** p < .01 *** p < .001



Appendix I:

Supplementary tables from Study Three



Table I1

Standard multiple regression of socio-demographic variables on frequency of

unlicensed driving (n=304)


B Std. error

ß sr2 R2 Adj R2

Frequency of unlicensed driving1

.99 .47

Gender .84 .37 -.03 .08 -.02

Age2

21 - 25 .35 .48 -.02 .08 -.02

26 - 39 .40 .49 -.08 .08 -.08

40 or over .06 .23 -.06 .13 -.03

Marital status 1.22 .42 -.05 .06 -.04

Educational level 1.54 .50 .13* .06 .13 .01


.42 .49 .25*** .06 .26 .06


.65 .48 .02 .07 .02

Annual income3

$11,000 - $30,000 .48 .50 .01 .07 .01

$31,000 or more .29 .45 .07 .09 .07

Prior criminal conviction .38 .49 .17** .06 .17 .02

.11*** .08

1. Logarithmically transformed. 2. Reference category is 17 – 20. 3. Reference category is Less than $10,000. Model: F(11, 292) = 3.38, p < .001 Unique variability = .09; shared variability = .02. * p < .05 ** p < .01 *** p < .001


Table I2

Logistic regression analysis of continued driving after detection as a function of socio-

demographic variables (n=304)


error Wald test

Odds Ratio

Upper Lower

Gender -.25 .37 .46 .78 .38 1.61

Age1

21 - 25 -.24 .36 .45 .79 .39 1.59

26 - 39 -.73 .37 3.95* .48 .24 .99

40 or over -1.26 .71 3.19 .28 .07 1.13

Marital status -.47 .34 1.94 .63 .32 1.21

Educational level -.15 .28 .30 .86 .49 1.49

Needed to drive for work when unlicensed .67 .29 5.55* 1.96 1.12 3.42

Employed at the time of court hearing -.79 .33 5.59* .46 .24 .88

Annual income2

$11,000 - $30,000 -.09 .34 .07 .91 .47 1.79

$31,000 or more .25 .43 .34 1.29 .55 3.01

Prior criminal conviction -.01 .30 .00 .99 .55 1.77

1. Reference category is 17 – 20. 2. Reference category is Less than $10,000. Full model vs constant-only model: χ2 (11, 304) = 18.56, p > .05; Nagelkerke R2 = .08. * p < .05 ** p < .01 *** p < .001


Table I3

Standard multiple regression of socio-demographic variables on intention to drive

unlicensed in the future (n=305)


B Std. error

ß sr2 R2 Adj R2

Intention to drive unlicensed in future1

.68 .77

Gender .84 .37 .19 .13 .09

Age2

21 - 25 .35 .40 -.20 .13 -.13

26 - 39 .06 .24 -.19 .13 -.12

40 or over .35 .40 -.35 .21 -.11

Marital status 1.22 .41 -.28** .11 -.15 .02

Educational level 1.54 .50 -.06 .10 -.04


.42 .49 .09 .09 .06


.65 .48 .11 .11 .07

Annual income3

$11,000 - $30,000 .49 .50 -.14 .12 -.09

$31,000 or more .29 .45 -.07 .15 -.04

Prior criminal conviction .39 .49 .14 .10 .09

.07* .03

1. Logarithmically transformed. 2. Reference category is 17 – 20. 3. Reference category is Less than $10,000. Model: F (11, 293) = 1.98, p < .05 Unique variability = .02; shared variability = .05. * p < .05 ** p < .01 *** p < .001


Table I4

Standard multiple regression of environmental facilitating factors on frequency of



B Std. error

ß sr2 R2 Adj R2


.99 .47


1.32 .47 .08 .06 .08

Able to access vehicle .96 .19 .28 .15 .11

Owned a vehicle .63 .48 .08 .06 .09

Still had a photo licence .49 .50 .15* .06 .16 .02

.08*** .07

1. Logarithmically transformed. Model: F (4, 294) = 6.24, p < .001 Unique variability = .02; shared variability = .06. * p < .05 ** p < .01 *** p < .001

Table I5

Logistic regression analysis of continued driving after detection as a function of

environmental facilitating factors (n=299)


error Wald test

Odds Ratio

Upper Lower


-.33 .32 1.11 .72 .39 1.33

Able to access vehicle 1.81 1.07 2.85 6.08 .75 49.3

Owned a vehicle -.18 .27 .44 .84 .50 1.42

Still had a photo licence -.25 .29 .74 .78 .45 1.37

Full model vs constant-only model: χ2 (4, 299) = 7.53, p > .05; Nagelkerke R2 = .04. * p < .05 ** p < .01 *** p < .001


Table I6

Standard multiple regression of environmental facilitating factors on intention to drive



B Std. error

ß sr2 R2 Adj R2


.69 .77


1.32 .47 -.35** .11 -.21 .03

Able to access vehicle .96 .19 .72** .24 .18 .03

Owned a vehicle .63 .48 -.11 .09 -.07

Still had a photo licence .49 .50 -.12 .10 -.08

.09*** .08

1. Logarithmically transformed. Model: F(4, 294) = 7.36, p < .001 Unique variability = .06; shared variability = .03. * p < .05 ** p < .01 *** p < .001

Table I7

Standard multiple regression of classical deterrence variables on frequency of



B Std. error

ß sr2 R2 Adj R2


.98 .47


3.32 1.82 -.02 .02 -.11

Knew fine for unlicensed driving

.14 .35 -.02 .08 -.01


4.56 1.84 -.02 .02 -.07

Perceived certainty of punishment

5.36 1.92 .00 .01 .01


5.17 1.79 -.02 .02 -.06

Prior conviction for unlicensed driving

.39 .49 -.02 .06 -.02

Exposure to enforcement .26 .44 -.16* .06 -.15 .02

.04 .02

1. Logarithmically transformed. Model: F(7, 288) = 1.82, p > .05 Unique variability = .02; shared variability = .02. * p < .05 ** p < .01 *** p < .001


Table I8


classical deterrence variables (n=298)


error Wald test

Odds Ratio

Upper Lower


-.18 .07 7.57** .84 .74 .95


-.15 .39 .14 .86 .40 1.86


.07 .07 .98 1.08 .93 1.24


-.14 .07 4.38* .87 .76 .99


-.01 .08 .00 1.00 .86 1.15


.29 .28 1.13 1.34 .78 2.31

Exposure to enforcement .23 .29 .63 1.26 .71 2.24

Full model vs constant-only model: χ2 (7, 298) = 15.00, p < .05; Nagelkerke R2 = .07. * p < .05 ** p < .01 *** p < .001

Table I9

Standard multiple regression of classical deterrence variables on intention to drive



B Std. error

ß sr2 R2 Adj R2


.69 .77


4.58 2.01 -.08*** .02 -.21 .04


.14 .34 .07 .13 .03


4.57 1.83 -.01 .02 -.02


5.38 1.92 -.02 .02 -.06


5.18 1.78 -.03 .03 -.07


.39 .49 .33*** .09 .21 .04

Exposure to enforcement .26 .44 -.04 .10 -.02

.10*** .08



Table I10

Standard multiple regression of expanded deterrence variables on frequency of



B Std. error

ß sr2 R2 Adj R2


.97 .47


3.33 1.83 -.03* .02 -.11 .01


.14 .34 -.03 .08 -.03


4.55 1.84 -.02 .02 -.06


5.37 1.92 .00 .01 .01


5.17 1.79 -.02 .02 -.06


.39 .49 -.02 .06 -.02




.29 .45 .01 .06 .08

Vicarious exposure to punishment avoidance

.36 .48 -.01 .06 -.01

.12*** .09



Table I11


expanded deterrence variables (n=294)


error Wald test

Odds Ratio

Upper Lower Perceived risk of

apprehension (after detection)

-.14 .07 4.49* .87 .76 .99


-.35 .42 .73 .70 .31 1.58


.09 .08 1.31 1.09 .94 1.27


-.13 .07 3.31 .88 .77 1.01

Perceived swiftness of punis hment

-.02 .08 .05 .98 .84 1.15


.15 .29 .28 1.16 .66 2.05

Exposure to enforcement .49 .35 2.00 1.64 .83 3.25

Punishment avoidance .47 .32 2.14 1.60 .85 3.00


.79 .30 6.99** 2.20 1.23 3.93


.18 .29 .39 1.20 .68 2.10



Table I12

Standard multiple regression of expanded deterrence variables on intention to drive



B Std. error

ß sr2 R2 Adj R2


.68 .77


4.59 2.00 -.07** .02 -.18 .03


.14 .34 .02 .13 .01


4.57 1.84 -.01 .02 -.02


5.38 1.91 -.01 .02 -.03


5.18 1.78 -.03 .03 -.08


.39 .49 .30*** .09 .19 .03

Exposure to enforcement .26 .44 .07 .11 .04

Punishment avoidance .37 .48 .21* .10 .13 .01


.29 .45 .16 .10 .10


.36 .48 .67 .09 .10

.14*** .10

1. Logarithmically transformed. Model: F (10, 284) = 4.42, p < .001 Unique variability = .07; shared variability = .07. * p < .05 ** p < .01 *** p ≤ .001


Table I13

Hierarchical regression of deterrence variables on frequency of unlicensed driving

(n=292)


B Std. error

ß R2 Adj R2 ∆R2

Step one


3.33 1.83 -.03* .02 -.11


.14 .34 -.03 .08 -.02


4.55 1.84 -.02 .02 -.07


5.37 1.92 .00 .02 -.01


5.17 1.79 -.02 .02 -.06


.39 .49 -.02 .06 -.02

Exposure to enforcement .26 .44 -.16* .06 -.15

.04 .02

Step two

Punishment avoidance .37 .48 .29*** .06 .30


.29 .45 .07 .06 .08


.36 .48 -.01 .06 -.01

.12*** .09 .08***

* p < .05 ** p < .01 *** p < .001


Table I14

Sequential logistic regression analysis of continued driving after detection as a function

of deterrence variables (n=294)


error Wald test

Odds Ratio

Upper Lower

Block 11


-.17 .07 6.72* .84 .74 .96

Knew fine for unlicensed driving -.22 .40 .30 .80 .37 1.76

Perceived severity of punishment .07 .07 .92 1.07 .93 1.24

Perceived certainty of punishment -.13 .07 3.61 .88 .77 1.00

Perceived swiftness of punishment -.01 .08 .00 1.00 .86 1.15

Prior conviction for unlicensed driving .28 .28 1.03 1.33 .77 2.29

Exposure to enforcement .26 .29 .78 1.30 .73 2.31

Block 22



.79 .30 6.99** 2.20 1.23 3.93


.18 .29 .39 1.20 .68 2.10

1. χ2 (7, 294) = 13.46, p > .05; Nagelkerke R2 = .06. 2. χ2 (3, 294) = 11.94, p < .01; Nagelkerke R2 = .12. * p < .05 ** p < .01 *** p < .001


Table I15

Hierarchical regression of deterrence variables on intention to drive unlicensed in the

future (n=295)


B Std. error

ß R2 Adj R2 ∆R2

Step one


4.59 2.00 -.08** .02 -.20


.14 .34 .06 .13 .03


4.57 1.84 -.01 .02 -.02


5.38 1.91 -.02 .02 -.04


5.18 1.78 -.03 .03 -.08


.39 .49 .33*** .09 .21


.09*** .07

Step two

Punishment avoidance .37 .48 .21* .10 .13


.29 .45 .16 .10 .10


.36 .48 .67 .09 .10

.14*** .10 .04**

* p < .05 ** p < .01 *** p < .001


Table I16

Standard multiple regression of social learning variables on frequency of unlicensed

driving (n=297)


B Std. error

ß sr2 R2 Adj R2


.98 .47

Total unlicensed driving models

4.36 5.94 -.01 .01 -.06

Differential association (behavioural dimension)

1.23 .42 .13 .07 .11

Differential association (normative dimension)

10.57 5.57 .01 .01 .14


37.61 12.09 -.01* .00 -.17 .01

Attitudes to alternative behaviours

18.79 7.13 - .02*** .00 -.26 .06


Anticipated punishments 54.01 10.29 -.01* .00 -.15 .02

.10*** .08

1. Logarithmically transformed. Model: F (7, 289) = 4.65, p < .001 Unique variability = .09; shared variability = .01 * p < .05 ** p < .01 *** p < .001 Table I17


learning variables (n=297)


error Wald test



-.02 .03 .61 .98 .94 1.03


.87 .33 7.03** 2.38 1.26 4.53


.09 .03 7.36** 1.09 1.02 1.16


.01 .02 .15 1.01 .98 1.04


-.05 ,02 5.56* .95 .91 .99

Anticipated rewards .02 .03 .46 1.02 .97 1.08

Anticipated punishments -.02 .02 .93 .98 .95 1.02



Table I18

Standard multiple regression of social learning variables on intention to drive



B Std. error

ß sr2 R2 Adj R2


.68 .77


Differential association (behavioural dimension) 1.23 .42 .30** .10 .17 .02

Differential association (normative dimension) 10.56 5.57 .02 .01 .13

Attitudes to unlicensed driving 37.60 12.07 .02*** .00 .26 .03



Anticipated punishments 54.05 10.30 .01** .01 -.17 .02

.30*** .29



Table I19


unlicensed driving for disqualified, not currently licensed and never licensed drivers

(n=94)


B Std. error

ß sr2 R2 Adj R2



3.73 1.75 .01 .03 .03

Knew fine for unlicensed driving .19 .40 -.16 .13 -.12

Perceived severity of punishment 4.84 1.83 .02 .03 .08

Perceived certainty of punishment 5.59 1.90 -.02* .03 -.07 .04

Perceived swiftness of punishment 5.59 1.53 -.07 .03 -.21

Prior conviction for unlicensed driving .68 .47 .15 .11 .14


Punishment avoidance .36 .48 .31** .12 .29 .06

Vicarious exposure to punishment .34 .48 .12 .11 .11

Vicarious exposure to punishment avoidance .46 .50 .04 .10 .04

.25** .16



Table I20


unlicensed driving for suspended, expired and not appropriately licensed drivers

(n=198)


B Std. error

ß sr2 R2 Adj R2

Frequency of unlicensed driving1 1.02 .44


3.14 1.84 -.02 .02 -.10

Knew fine for unlicensed driving .11 .32 .06 .10 .04

Perceived severity of punishment 4.42 1.83 -.03 .02 -.11

Perceived certainty of punishment 5.26 1.92 .02 .02 .08

Perceived swiftness of punishment 4.97 1.87 .00 .02 .02

Prior conviction for unlicensed driving .25 .43 -.06 .08 -.07



Vicarious exposure to punishment .26 .44 .06 .07 .06

Vicarious exposure to punishment avoidance .31 .46 -.03 .07 -.03

.12** .07



Table I21


expanded deterrence variables for disqualified, not currently licensed and never

licensed drivers (n=94)


error Wald test



-.22 .14 2.30 .81 .61 1.07


-1.75 .86 4.19* .17 .03 .93


.21 .16 1.76 1.24 .90 1.70


-.04 .15 .08 .96 .72 1.28


-.23 .17 1.71 .80 .57 1.12


.99 .62 2.51 2.68 .79 9.09




.21 .59 .13 1.23 .39 3.88


.50 .55 .83 1.64 .57 4.78

Full model vs constant-only model: χ2 (10, .94) = 19.04, p < .05; Nagelkerke R2 = .26. * p < .05 ** p < .01 *** p < .001


Table I22


expanded deterrence variables for suspended, expired and not appropriately licensed

drivers (n=200)


error Wald test



-.11 .08 1.85 .90 .77 1.05


.20 .53 .14 1.22 .43 3.40


.04 .09 .148 1.04 .87 1.24


-.16 .08 3.54 .85 .73 1.01


.06 .09 .37 1.06 .88 1.26


-.13 .40 .11 .88 .40 1.92




1.05 .37 7.97** 2.87 1.38 5.97


.01 .36 .00 1.01 .50 2.06



Table I23


unlicensed in the future for disqualified, not currently licensed and never licensed

drivers (n=94)


B Std. error

ß sr2 R2 Adj R2

Intention to drive unlicensed in future1 .79 .79


4.79 1.90 -.07 .04 -.16

Knew fine for unlicensed driving .19 .39 -.33 .18 -.17

Perceived severity of punishment 4.84 1.82 .03 .04 .06

Perceived certainty of punishment 5.60 1.89 -.06 .04 -.14

Perceived swiftness of punishment 5.60 1.53 -.15** .05 -.29 .07

Prior conviction for unlicensed driving .67 .47 .41* .16 .25 .05



Vicarious exposure to punishment .35 .48 -.19 .16 -.12


.35*** .27



Table I24


unlicensed in the future for suspended, expired and not appropriately licensed drivers

(n=200)


B Std. error

ß sr2 R2 Adj R2



4.50 2.05 -.07** .03 -.18 .03

Knew fine for unlicensed driving .11 .31 .22 .17 .09

Perceived severity of punishment 4.43 1.83 -.02 .03 -.06

Perceived certainty of punishment 5.28 1.92 .00 .03 .01

Perceived swiftness of punishment 4.99 1.86 .00 .03 .00

Prior conviction for unlicensed driving .26 .44 .14 .12 .08 .



Vicarious exposure to punishment .26 .44 .33** .13 .19 .03


.14*** .09

1. Logarithmically transformed. Model: F (10, 189) = 3.02, p = .001 Unique variability = .06; shared variability = .08. * p < .05 ** p ≤ .01 *** p ≤ .001


Table I25


driving for disqualified, not currently licensed and never licensed drivers (n=93)


B Std. error

ß sr2 R2 Adj R2



Differential association (behavioural dimension) 1.32 .47 .19 .13 .18


Attitudes to unlicensed driving 39.09 11.60 -.01 .01 -.04


Anticipated rewards 10.81 6.18 .01 .01 .07


.15* .08



Table I26


driving for suspended, expired and not appropriately licensed drivers (n=204)


B Std. error

ß sr2 R2 Adj R2

Frequency of unlicensed driving1 1.03 .44




Attitudes to unlicensed driving 36.93 12.28 -.01 .00 -.19

Attitudes to alternative behaviours 18.48 6.99 -.01*** .01 -.26 .06



.08* .05



Table I27


learning variables for disqualified, not currently licensed and never licensed drivers

(n=92)


error Wald test


Total unlicensed driving models -.00 .04 .01 1.00 .92 1.09

Differential association (behavioural dimension) .64 .58 1.19 1.89 .60 5.92

Differential association (normative dimension) .08 .06 1.94 1.09 .97 1.22

Attitudes to unlicensed driving .01 .03 .05 1.01 .95 1.07

Attitudes to alternative behaviours -.08 .04 3.90 .92 .85 1.00


Anticipated punishments .01 .03 .05 1.01 .95 1.07

Full model vs constant-only model: χ2 (7, 92) = 13.67, p > .05; Nagelkerke R2 = .20 * p < .05 ** p < .01 *** p < .001 Table I28


learning variables for suspended, expired and not appropriately licensed drivers

(n=205)


error Wald test

Odds Ratio

Upper Lower


-.02 .03 .42 .98 .92 1.04


1.05 .41 6.48 2.85 1.27 6.39


.08 .04 4.69 1.09 1.01 1.17


.01 .02 .21 1.01 .97 1.05


-.04 .03 2.13 .96 .92 1.01


Anticipated punishments -.02 .02 .99 .98 .94 1.02

Full model vs constant-only model: χ2 (7, 205) = 33.89, p < .001; Nagelkerke R2 = .21 * p < .05 ** p < .01 *** p < .001


Table I29


unlicensed in the future for disqualified, not currently licensed and never licensed

drivers (n=93)


B Std. error

ß sr2 R2 Adj R2




Differential association (normative dimension) 10.80 5.59 -.02 .02 -.16

Attitudes to unlicensed driving 39.09 11.60 .02** .01 .35 .07

Attitudes to alternative behaviours 19.46 7.41 -.01 .01 -.12


Anticipated punishments 54.98 9.94 -.03*** .01 -.39 .10

.46*** .41



Table I30


unlicensed in the future for suspended, expired and not appropriately licensed drivers

(n=205)


B Std. error

ß sr2 R2 Adj R2





Attitudes to unlicensed driving 36.93 12.25 .03*** .01 .20 .04

Attitudes to alternative behaviours 18.43 7.01 .01* .01 -.10 .02


Anticipated punishments 53.63 10.46 .01 .01 -.11

.29*** .26

1. Logarithmically transformed. Model: F (7, 197) = 11.39, p < .001 Unique variability = .06; shared variability = .23. * p < .05 ** p < .01 *** p ≤ .001

Documents

The psychosocial characteristics and on-road behaviour of ...nature of unlicensed driving, in order to develop more effective countermeasures to the behaviour. Study One utilisedofficial