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
2.1 Introductory comments.................................................................................................................. 15
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
3.1 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
Chapter Four: The crash involvement of unlicensed drivers ............................. 65
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 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
6.1 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 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
7.1 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?.......................................................... 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
Bivariate correlations between dependent variables and
sensation seeking................................................................................................175
xvii
Table 6.4
Bivariate correlations between dependent variables and AUDIT
scores ....................................................................................................................176
Table 6.5
Bivariate correlations between dependent variables and
environmental facilitating factors ................................................................177
Table 6.6
Bivariate correlations between dependent variables and
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
Table I5 Logistic regression analysis of continued driving after detection
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
Logistic regression analysis of continued driving after detection
as a function of expanded deterrence variables ....................................................312
Table I12
Standard multiple regression of expanded deterrence variables
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
Logistic regression analysis of continued driving after detection
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
Standard multiple regression of expanded deterrence variables
on frequency of unlicensed driving for disqualified, not currently
licensed and never licensed drivers................................................................319
xix
Table I20
Standard multiple regression of expanded deterrence variables
on frequency of unlicensed driving for suspended, expired and
not appropriately licensed drivers................................................................320
Table I21
Logistic regression analysis of continued driving after detection
as a function of expanded deterrence variables for disqualified,
not currently licensed and never licensed drivers ................................ 321
Table I22
Logistic regression analysis of continued driving after detection
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
Table I26 Standard multiple regression of social learning variables on
frequency of unlicensed driving for suspended, expired and
not appropriately licensed drivers................................................................326
Table I27 Logistic regression analysis of continued driving after detection
as a function of social learning variables for disqualified, not
currently licensed and never licensed drivers .......................................................327
Table I28 Logistic regression analysis of continued driving after detection
as a function of social learning variables for suspended, expired
and not appropriately licensed drivers ................................................................327
Table I29 Standard multiple regression of social learning variables on
intention to drive unlicensed in the future for disqualified, not
currently licensed and never licensed drivers .......................................................328
Table I30 Standard multiple regression of social learning variables on
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
The characteristics and on-road behaviour of unlicensed drivers 2
The characteristics and on-road behaviour of unlicensed drivers 3
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
The characteristics and on-road behaviour of unlicensed drivers 4
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.
The characteristics and on-road behaviour of unlicensed drivers 5
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.
The characteristics and on-road behaviour of unlicensed drivers 6
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
The characteristics and on-road behaviour of unlicensed drivers 7
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
The characteristics and on-road behaviour of unlicensed drivers 8
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.
The characteristics and on-road behaviour of unlicensed drivers 9
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
The characteristics and on-road behaviour of unlicensed drivers 10
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.
The characteristics and on-road behaviour of unlicensed drivers 11
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.
The characteristics and on-road behaviour of unlicensed drivers 12
The characteristics and on-road behaviour of unlicensed drivers 13
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 characteristics and on-road behaviour of unlicensed drivers 14
The characteristics and on-road behaviour of unlicensed drivers 15
2.1 Introductory comments
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).
The characteristics and on-road behaviour of unlicensed drivers 16
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,
The characteristics and on-road behaviour of unlicensed drivers 17
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
The characteristics and on-road behaviour of unlicensed drivers 18
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.
The characteristics and on-road behaviour of unlicensed drivers 19
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).
The characteristics and on-road behaviour of unlicensed drivers 20
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
The characteristics and on-road behaviour of unlicensed drivers 21
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
The characteristics and on-road behaviour of unlicensed drivers 22
(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
The characteristics and on-road behaviour of unlicensed drivers 23
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
The characteristics and on-road behaviour of unlicensed drivers 24
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.
The characteristics and on-road behaviour of unlicensed drivers 25
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).
The characteristics and on-road behaviour of unlicensed drivers 26
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 characteristics and on-road behaviour of unlicensed drivers 27
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.
The characteristics and on-road behaviour of unlicensed drivers 28
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).
The characteristics and on-road behaviour of unlicensed drivers 29
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
The characteristics and on-road behaviour of unlicensed drivers 30
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.
The characteristics and on-road behaviour of unlicensed drivers 31
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%
The characteristics and on-road behaviour of unlicensed drivers 32
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
The characteristics and on-road behaviour of unlicensed drivers 33
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
The characteristics and on-road behaviour of unlicensed drivers 34
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 characteristics and on-road behaviour of unlicensed drivers 35
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;
The characteristics and on-road behaviour of unlicensed drivers 36
§ 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
The characteristics and on-road behaviour of unlicensed drivers 37
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
The characteristics and on-road behaviour of unlicensed drivers 38
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.
The characteristics and on-road behaviour of unlicensed drivers 39
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 characteristics and on-road behaviour of unlicensed drivers 40
The characteristics and on-road behaviour of unlicensed drivers 41
3.1 Introductory comments
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 characteristics and on-road behaviour of unlicensed drivers 42
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.
The characteristics and on-road behaviour of unlicensed drivers 43
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
The characteristics and on-road behaviour of unlicensed drivers 44
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,
The characteristics and on-road behaviour of unlicensed drivers 45
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
The characteristics and on-road behaviour of unlicensed drivers 46
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:
The characteristics and on-road behaviour of unlicensed drivers 47
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
The characteristics and on-road behaviour of unlicensed drivers 48
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
The characteristics and on-road behaviour of unlicensed drivers 49
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
The characteristics and on-road behaviour of unlicensed drivers 50
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
The characteristics and on-road behaviour of unlicensed drivers 51
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.
The characteristics and on-road behaviour of unlicensed drivers 52
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
The characteristics and on-road behaviour of unlicensed drivers 53
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
The characteristics and on-road behaviour of unlicensed drivers 54
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:
The characteristics and on-road behaviour of unlicensed drivers 55
§ 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
The characteristics and on-road behaviour of unlicensed drivers 56
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
The characteristics and on-road behaviour of unlicensed drivers 57
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.
The characteristics and on-road behaviour of unlicensed drivers 58
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.
The characteristics and on-road behaviour of unlicensed drivers 59
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
The characteristics and on-road behaviour of unlicensed drivers 60
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 characteristics and on-road behaviour of unlicensed drivers 61
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
The characteristics and on-road behaviour of unlicensed drivers 62
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).
The characteristics and on-road behaviour of unlicensed drivers 63
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.
The characteristics and on-road behaviour of unlicensed drivers 64
The characteristics and on-road behaviour of unlicensed drivers 65
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
The characteristics and on-road behaviour of unlicensed drivers 66
The characteristics and on-road behaviour of unlicensed drivers 67
4.1 Introductory comments
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.
1. Do unlicensed drivers engage in more risky driving than other drivers?
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
The characteristics and on-road behaviour of unlicensed drivers 68
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.
3. Do unlicensed drivers represent a homogeneous group, in terms of their
psychosocial characteristics and on-road behaviour?
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
The characteristics and on-road behaviour of unlicensed drivers 69
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.
The characteristics and on-road behaviour of unlicensed drivers 70
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
The characteristics and on-road behaviour of unlicensed drivers 71
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.
The characteristics and on-road behaviour of unlicensed drivers 72
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
The characteristics and on-road behaviour of unlicensed drivers 73
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.
The characteristics and on-road behaviour of unlicensed drivers 74
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
The characteristics and on-road behaviour of unlicensed drivers 75
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
Fatal Serious injury
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.
Source: Queensland Road Crash Database, Queensland Transport
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).
The characteristics and on-road behaviour of unlicensed drivers 76
Table 4.3
Unlicensed drivers involved in crashes in Queensland: 1994-98 by offender type
Severity of Crash Licence type
Fatal Serious injury
Other Injury
Property Damage Total
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
The characteristics and on-road behaviour of unlicensed drivers 77
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
1. The cells with significant (p<.01) adjusted standardised residuals are bolded. Source: Queensland Road Crash Database, Queensland Transport
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
The characteristics and on-road behaviour of unlicensed drivers 78
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.
The characteristics and on-road behaviour of unlicensed drivers 79
Table 4.6
Circumstances of crashes resulting in serious casualty crashes in Queensland: 1994-98
by licence status
Licence status Variable Licensed
(%) 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
1. The cells with significant (p<.01) adjusted standardised residuals are bolded. Source: Queensland Road Crash Database, Queensland Transport
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
The characteristics and on-road behaviour of unlicensed drivers 80
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.
The characteristics and on-road behaviour of unlicensed drivers 81
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
1. The cells with significant (p<.01) adjusted standardised residuals are bolded. Source: Queensland Road Crash Database, Queensland Transport
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 characteristics and on-road behaviour of unlicensed drivers 82
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.
The characteristics and on-road behaviour of unlicensed drivers 83
Table 4.8
Contributing factors to serious casualty crashes in Queensland: 1994-98 by licence
status
Licence status Variable Licensed
(%) 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.
Source: Queensland Road Crash Database, Queensland Transport
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.
The characteristics and on-road behaviour of unlicensed drivers 84
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
Unlicensed Driver Type Variable Dis-
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
The characteristics and on-road behaviour of unlicensed drivers 85
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.
The characteristics and on-road behaviour of unlicensed drivers 86
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
The characteristics and on-road behaviour of unlicensed drivers 87
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.
The characteristics and on-road behaviour of unlicensed drivers 88
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.
The characteristics and on-road behaviour of unlicensed drivers 89
Table 4.11
The involvement of drivers in multi-vehicle crashes compared with all crashes, by
severity for Queensland: 1994-98
Severity of Crash
Fatal Serious injury
Other Injury
Property Damage Total
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
The characteristics and on-road behaviour of unlicensed drivers 90
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
Fatal Serious injury
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.
The characteristics and on-road behaviour of unlicensed drivers 91
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
The characteristics and on-road behaviour of unlicensed drivers 92
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.
The characteristics and on-road behaviour of unlicensed drivers 93
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.
The characteristics and on-road behaviour of unlicensed drivers 94
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)
99% CI for Odds ratio Variables B Std.
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
The characteristics and on-road behaviour of unlicensed drivers 95
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
The characteristics and on-road behaviour of unlicensed drivers 96
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 characteristics and on-road behaviour of unlicensed drivers 97
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
The characteristics and on-road behaviour of unlicensed drivers 98
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 characteristics and on-road behaviour of unlicensed drivers 99
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
The characteristics and on-road behaviour of unlicensed drivers 100
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
The characteristics and on-road behaviour of unlicensed drivers 101
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 characteristics and on-road behaviour of unlicensed drivers 102
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
The characteristics and on-road behaviour of unlicensed drivers 103
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.
The characteristics and on-road behaviour of unlicensed drivers 104
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.
The characteristics and on-road behaviour of unlicensed drivers 105
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.
The characteristics and on-road behaviour of unlicensed drivers 106
The characteristics and on-road behaviour of unlicensed drivers 107
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
The characteristics and on-road behaviour of unlicensed drivers 108
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
The characteristics and on-road behaviour of unlicensed drivers 109
5.1 Introductory comments
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.
5.2 Study aims and hypotheses
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.
1. Do unlicensed drivers engage in more risky driving than other drivers?
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-
The characteristics and on-road behaviour of unlicensed drivers 110
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.
3. Do unlicensed drivers represent a homogeneous group, in terms of their
psychosocial characteristics and on-road behaviour?
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.
H12 The disqualified, suspended and never licensed drivers will report higher
levels of alcohol misuse than the other offenders.
H13 The disqualified, suspended and never licensed drivers will report higher
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
The characteristics and on-road behaviour of unlicensed drivers 111
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.
The characteristics and on-road behaviour of unlicensed drivers 112
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.
The characteristics and on-road behaviour of unlicensed drivers 113
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.
The characteristics and on-road behaviour of unlicensed drivers 114
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 characteristics and on-road behaviour of unlicensed drivers 115
§ 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
The characteristics and on-road behaviour of unlicensed drivers 116
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).
The characteristics and on-road behaviour of unlicensed drivers 117
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:
The characteristics and on-road behaviour of unlicensed drivers 118
§ 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.
The characteristics and on-road behaviour of unlicensed drivers 119
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.
The characteristics and on-road behaviour of unlicensed drivers 120
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.
The characteristics and on-road behaviour of unlicensed drivers 121
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-
The characteristics and on-road behaviour of unlicensed drivers 122
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 characteristics and on-road behaviour of unlicensed drivers 123
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)
The characteristics and on-road behaviour of unlicensed drivers 124
Table 5.3
Socio-demographic characteristics of participants by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 125
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.
The characteristics and on-road behaviour of unlicensed drivers 126
Table 5.4
Driving and criminal history of participants by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 127
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 128
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 129
Table 5.7
AUDIT scores by type of unlicensed driver
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 130
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)
The characteristics and on-road behaviour of unlicensed drivers 131
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 132
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)
The characteristics and on-road behaviour of unlicensed drivers 133
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 134
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 $900 and the majority received an Absolute
Disqualification. This requires offenders to serve a disqualification period of at least two
years before being eligible to apply for the return of their licence from a court. Among
the other offenders, the lowest average penalties were applied to the drivers with expired
licences, in-keeping with the more administrative nature of this offence.
21. This analysis excluded the ‘Inappropriate licence’ category of offender t o ensure sufficient cell sizes.
The characteristics and on-road behaviour of unlicensed drivers 135
Table 5.11
Penalties received for unlicensed/disqualified driving conviction among offenders who
were convicted of no other offences
Unlicensed Driver Type Variable
Disqualified Suspended Expired Not
currently licensed
Never licensed
Inapp. licence
Fine ($) 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.
The characteristics and on-road behaviour of unlicensed drivers 136
Table 5.12
Length of time driving unlicensed by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
Years driving unlicensed1
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
Years driving unlicensed1
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.
The characteristics and on-road behaviour of unlicensed drivers 137
Table 5.13
Frequency of driving while unlicensed by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
Trips per week1 n=52 n=108 n=91 n=21 n=26 n=10 n=308
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
Trips per week1 n=52 n=108 n=91 n=21 n=26 n=10 n=308
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 characteristics and on-road behaviour of unlicensed drivers 138
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.
The characteristics and on-road behaviour of unlicensed drivers 139
Table 5.14
Cautiousness of driving while unlicensed by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 140
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.
The characteristics and on-road behaviour of unlicensed drivers 141
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].
The characteristics and on-road behaviour of unlicensed drivers 142
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
Unlicensed drivers ATSB survey
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
Unlicensed drivers ATSB survey
Figure 5.7: General approach to drink driving (n=288/1453)
The characteristics and on-road behaviour of unlicensed drivers 143
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 144
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
%
Not currently licensed
%
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
The characteristics and on-road behaviour of unlicensed drivers 145
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
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 146
Table 5.18
Continued driving unlicensed after detection by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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.
The characteristics and on-road behaviour of unlicensed drivers 147
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
The characteristics and on-road behaviour of unlicensed drivers 148
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:
The characteristics and on-road behaviour of unlicensed drivers 149
• 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.
The characteristics and on-road behaviour of unlicensed drivers 150
• 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.
The characteristics and on-road behaviour of unlicensed drivers 151
Table 5.22
Intention to drive unlicensed in the future by offender type
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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).
The characteristics and on-road behaviour of unlicensed drivers 152
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
The following section discusses the results of the study in light of the 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
The characteristics and on-road behaviour of unlicensed drivers 153
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
The characteristics and on-road behaviour of unlicensed drivers 154
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.
The disqualified, suspended and never licensed drivers will report higher levels of
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.
The disqualified, suspended and never licensed drivers will report higher levels of
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.
The characteristics and on-road behaviour of unlicensed drivers 155
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
The characteristics and on-road behaviour of unlicensed drivers 156
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
The characteristics and on-road behaviour of unlicensed drivers 157
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,
The characteristics and on-road behaviour of unlicensed drivers 158
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
The characteristics and on-road behaviour of unlicensed drivers 159
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
The characteristics and on-road behaviour of unlicensed drivers 160
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.
The characteristics and on-road behaviour of unlicensed drivers 161
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
The characteristics and on-road behaviour of unlicensed drivers 162
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.
The characteristics and on-road behaviour of unlicensed drivers 163
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
The characteristics and on-road behaviour of unlicensed drivers 164
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
The characteristics and on-road behaviour of unlicensed drivers 165
6.1 Introductory comments
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.
The characteristics and on-road behaviour of unlicensed drivers 166
6.2 Study aims and hypotheses
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.
The characteristics and on-road behaviour of unlicensed drivers 167
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
The characteristics and on-road behaviour of unlicensed drivers 168
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).
The characteristics and on-road behaviour of unlicensed drivers 169
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:
The characteristics and on-road behaviour of unlicensed drivers 170
§ 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.
The characteristics and on-road behaviour of unlicensed drivers 171
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
Continued to drive after detection
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
The characteristics and on-road behaviour of unlicensed drivers 172
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.
The characteristics and on-road behaviour of unlicensed drivers 173
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.
The characteristics and on-road behaviour of unlicensed drivers 174
Table 6.2
Bivariate correlations between dependent variables and socio-demographic variables
Dependent variables
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
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
The characteristics and on-road behaviour of unlicensed drivers 175
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
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
in the future1
Sensation seeking .08 .03 .02
1. Logarithmically transformed. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 176
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
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
in the future1
AUDIT score .07 .03 .04
1. Logarithmically transformed. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 177
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
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
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
1. Logarithmically transformed. * p < .05 ** p < .01 *** p < .001
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
The characteristics and on-road behaviour of unlicensed drivers 178
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.
The characteristics and on-road behaviour of unlicensed drivers 179
Table I5 in Appendix I reports the results of a logistic regression examining the
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.
The characteristics and on-road behaviour of unlicensed drivers 180
Table 6.6
Bivariate correlations between dependent variables and deterrence variables
Dependent variables
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
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*
1. Logarithmically transformed. * p < .05 ** p < .01 *** p < .001
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).
The characteristics and on-road behaviour of unlicensed drivers 181
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
The characteristics and on-road behaviour of unlicensed drivers 182
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
The characteristics and on-road behaviour of unlicensed drivers 183
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.
The characteristics and on-road behaviour of unlicensed drivers 184
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
The characteristics and on-road behaviour of unlicensed drivers 185
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].
The characteristics and on-road behaviour of unlicensed drivers 186
Table 6.7
Total number of unlicensed drivers known by participants by type of offender
Unlicensed Driver Type Variable Dis-
qualified
%
Suspended
%
Expired
%
Not currently licensed
%
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
Variable Frequency of unlicensed driving1
Continued to drive after detection
Intention to drive unlicensed
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
The characteristics and on-road behaviour of unlicensed drivers 187
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 characteristics and on-road behaviour of unlicensed drivers 188
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.
The characteristics and on-road behaviour of unlicensed drivers 189
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 characteristics and on-road behaviour of unlicensed drivers 190
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
Table I16 in Appendix I reports the results of a standard multiple regression
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.
Table I17 in Appendix I reports the results of a logistic regression examining the
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.
Table I18 in Appendix I reports the results of a standard multiple regression
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
The characteristics and on-road behaviour of unlicensed drivers 191
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.
The characteristics and on-road behaviour of unlicensed drivers 192
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
Intention to drive unlicensed
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.
The characteristics and on-road behaviour of unlicensed drivers 193
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.
The characteristics and on-road behaviour of unlicensed drivers 194
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
Intention to drive unlicensed
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).
The characteristics and on-road behaviour of unlicensed drivers 195
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.
The characteristics and on-road behaviour of unlicensed drivers 196
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:
The characteristics and on-road behaviour of unlicensed drivers 197
§ 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)
95% CI for Odds ratio Variables B Std.
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
The characteristics and on-road behaviour of unlicensed drivers 198
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.
The characteristics and on-road behaviour of unlicensed drivers 199
Table 6.13
Standard multiple regression of selected variables on intention to drive unlicensed in the
future (n=296)
Variables Mean Std. dev
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 (behavioural dimension) 1.23 .42 .25* .10 .14 .01
Differential association (normative dimension) 10.58 5.57 .02* .01 .13 .01
Attitudes to unlicensed driving 37.70 12.04 .01** .004 .20 .02
Attitudes to alternative behaviours 18.78 7.15 -.01* .01 -.11 .01
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
Perceived risk of apprehension (after detection)
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.
The characteristics and on-road behaviour of unlicensed drivers 200
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
The characteristics and on-road behaviour of unlicensed drivers 201
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
The following section discusses the results of the study in light of the hypotheses
outlined in section 6.2.
The characteristics and on-road behaviour of unlicensed drivers 202
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 over- involvement 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
The characteristics and on-road behaviour of unlicensed drivers 203
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
The characteristics and on-road behaviour of unlicensed drivers 204
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
The characteristics and on-road behaviour of unlicensed drivers 205
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
The characteristics and on-road behaviour of unlicensed drivers 206
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
The characteristics and on-road behaviour of unlicensed drivers 207
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.
The characteristics and on-road behaviour of unlicensed drivers 208
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%).
The characteristics and on-road behaviour of unlicensed drivers 209
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
The characteristics and on-road behaviour of unlicensed drivers 210
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:
The characteristics and on-road behaviour of unlicensed drivers 211
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
The characteristics and on-road behaviour of unlicensed drivers 212
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.
The characteristics and on-road behaviour of unlicensed drivers 213
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
The characteristics and on-road behaviour of unlicensed drivers 214
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
The characteristics and on-road behaviour of unlicensed drivers 215
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;
The characteristics and on-road behaviour of unlicensed drivers 216
§ 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.
The characteristics and on-road behaviour of unlicensed drivers 217
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
The characteristics and on-road behaviour of unlicensed drivers 218
The characteristics and on-road behaviour of unlicensed drivers 219
7.1 Introductory comments
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
The characteristics and on-road behaviour of unlicensed drivers 220
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
The characteristics and on-road behaviour of unlicensed drivers 221
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.
The characteristics and on-road behaviour of unlicensed drivers 222
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
psychosocial characteristics and on-road behaviour?
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.
The characteristics and on-road behaviour of unlicensed drivers 223
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.
The characteristics and on-road behaviour of unlicensed drivers 224
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
The characteristics and on-road behaviour of unlicensed drivers 225
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).
The characteristics and on-road behaviour of unlicensed drivers 226
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.
The characteristics and on-road behaviour of unlicensed drivers 227
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
The characteristics and on-road behaviour of unlicensed drivers 228
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 characteristics and on-road behaviour of unlicensed drivers 229
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).
The characteristics and on-road behaviour of unlicensed drivers 230
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
The characteristics and on-road behaviour of unlicensed drivers 231
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.
The characteristics and on-road behaviour of unlicensed drivers 232
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
The characteristics and on-road behaviour of unlicensed drivers 233
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
The characteristics and on-road behaviour of unlicensed drivers 234
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
The characteristics and on-road behaviour of unlicensed drivers 235
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,
The characteristics and on-road behaviour of unlicensed drivers 236
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
The characteristics and on-road behaviour of unlicensed drivers 237
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
The characteristics and on-road behaviour of unlicensed drivers 238
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.
The characteristics and on-road behaviour of unlicensed drivers 239
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.
The characteristics and on-road behaviour of unlicensed drivers 240
§ 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.
The characteristics and on-road behaviour of unlicensed drivers 241
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
The characteristics and on-road behaviour of unlicensed drivers 242
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.
The characteristics and on-road behaviour of unlicensed drivers 243
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).
The characteristics and on-road behaviour of unlicensed drivers 244
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.
The characteristics and on-road behaviour of unlicensed drivers 245
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.
The characteristics and on-road behaviour of unlicensed drivers 246
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.
The characteristics and on-road behaviour of unlicensed drivers 247
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.
The characteristics and on-road behaviour of unlicensed drivers 248
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.
The characteristics and on-road behaviour of unlicensed drivers 249
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
Analysis and Prevention, 23 (4), 203-224.
The characteristics and on-road behaviour of unlicensed drivers 250
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
Analysis and Prevention, 28 (4), 519-524.
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.
The characteristics and on-road behaviour of unlicensed drivers 251
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.
The characteristics and on-road behaviour of unlicensed drivers 252
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
Analysis and Prevention, 14 (3), 1169-172.
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.
The characteristics and on-road behaviour of unlicensed drivers 253
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.
The characteristics and on-road behaviour of unlicensed drivers 254
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.
The characteristics and on-road behaviour of unlicensed drivers 255
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
The characteristics and on-road behaviour of unlicensed drivers 256
The characteristics and on-road behaviour of unlicensed drivers 257
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;
§ a person’s authority to drive in Queensland under their non-Queensland driver
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;
§ a person’s authority to drive in Queensland under their non-Queensland driver
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;
§ a person’s authority to drive in Queensland under their non-Queensland driver
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.
The characteristics and on-road behaviour of unlicensed drivers 258
The characteristics and on-road behaviour of unlicensed drivers 259
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.
The characteristics and on-road behaviour of unlicensed drivers 260
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.
The characteristics and on-road behaviour of unlicensed drivers 261
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.
The characteristics and on-road behaviour of unlicensed drivers 262
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”.
The characteristics and on-road behaviour of unlicensed drivers 263
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”).
The characteristics and on-road behaviour of unlicensed drivers 264
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.
The characteristics and on-road behaviour of unlicensed drivers 265
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
The characteristics and on-road behaviour of unlicensed drivers 266
The characteristics and on-road behaviour of unlicensed drivers 267
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.
The characteristics and on-road behaviour of unlicensed drivers 268
The characteristics and on-road behaviour of unlicensed drivers 269
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
The characteristics and on-road behaviour of unlicensed drivers 270
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
The characteristics and on-road behaviour of unlicensed drivers 271
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
The characteristics and on-road behaviour of unlicensed drivers 272
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
The characteristics and on-road behaviour of unlicensed drivers 273
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
The characteristics and on-road behaviour of unlicensed drivers 274
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 characteristics and on-road behaviour of unlicensed drivers 275
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)
The characteristics and on-road behaviour of unlicensed drivers 276
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
The characteristics and on-road behaviour of unlicensed drivers 277
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? ____________________________________________________________________ ____________________________________________________________________
Only read “the last time” for participants with previous convictions
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)
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
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)
If yes: How many times? ______________
Was your licence checked on each occasion? Yes...............................................................................1 No ................................................................................2
Only read “the last time” for participants with previous convictions
If no, how many times wasn’t it checked? ____________________
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)
If yes: How many times? ______________
What happened as a result of the tickets? __________________________________ ____________________________________________________________________ ____________________________________________________________________
Only read “the last time” for participants with previous convictions
The characteristics and on-road behaviour of unlicensed drivers 278
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
The characteristics and on-road behaviour of unlicensed drivers 279
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
The characteristics and on-road behaviour of unlicensed drivers 280
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
The characteristics and on-road behaviour of unlicensed drivers 281
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
The characteristics and on-road behaviour of unlicensed drivers 282
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 your- self 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
The characteristics and on-road behaviour of unlicensed drivers 283
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: __________________
The characteristics and on-road behaviour of unlicensed drivers 284
The characteristics and on-road behaviour of unlicensed drivers 285
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.
The characteristics and on-road behaviour of unlicensed drivers 286
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.
The characteristics and on-road behaviour of unlicensed drivers 287
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.
The characteristics and on-road behaviour of unlicensed drivers 288
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.
The characteristics and on-road behaviour of unlicensed drivers 289
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
The characteristics and on-road behaviour of unlicensed drivers 290
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?
Cronbach’s alpha = .80
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
Cronbach’s alpha = .87
The characteristics and on-road behaviour of unlicensed drivers 291
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
Cronbach’s alpha = .76
F5. Attitudes to unlicensed driving scale
Items: . . . how much do you agree or disagree with each statement?
(Measured on 7-point Likert 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)
Cronbach’s alpha = .73
The characteristics and on-road behaviour of unlicensed drivers 292
F6. Attitudes to alternative transport scale
Items: . . . how much do you agree or disagree with each statement?
(Measured on 7-point Likert 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
Cronbach’s alpha = .66
F7. Rewards for unlicensed driving scale
Items: . . . how much do you agree or disagree with each statement?
(Measured on 7-point Likert 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
Cronbach’s alpha = .77
F8. Punishments for unlicensed driving scale
Items: . . . how much do you agree or disagree with each statement?
(Measured on 7-point Likert 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
The characteristics and on-road behaviour of unlicensed drivers 293
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
Cronbach’s alpha = .64
The characteristics and on-road behaviour of unlicensed drivers 294
The characteristics and on-road behaviour of unlicensed drivers 295
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
The characteristics and on-road behaviour of unlicensed drivers 296
The characteristics and on-road behaviour of unlicensed drivers 297
Appendix H:
Correlation matrix for dependent and independent variables used in Study Three
The characteristics and on-road behaviour of unlicensed drivers 298
The characteristics and on-road behaviour of unlicensed drivers 299
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
The characteristics and on-road behaviour of unlicensed drivers 300
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
The characteristics and on-road behaviour of unlicensed drivers 301
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
The characteristics and on-road behaviour of unlicensed drivers 302
The characteristics and on-road behaviour of unlicensed drivers 303
Appendix I:
Supplementary tables from Study Three
The characteristics and on-road behaviour of unlicensed drivers 304
The characteristics and on-road behaviour of unlicensed drivers 305
Table I1
Standard multiple regression of socio-demographic variables on frequency of
unlicensed driving (n=304)
Variables Mean Std. dev
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
Needed to drive for work when unlicensed
.42 .49 .25*** .06 .26 .06
Employed at the time of court hearing
.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
The characteristics and on-road behaviour of unlicensed drivers 306
Table I2
Logistic regression analysis of continued driving after detection as a function of socio-
demographic variables (n=304)
95% CI for Odds ratio Variables B Std.
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
The characteristics and on-road behaviour of unlicensed drivers 307
Table I3
Standard multiple regression of socio-demographic variables on intention to drive
unlicensed in the future (n=305)
Variables Mean Std. dev
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
Needed to drive for work when unlicensed
.42 .49 .09 .09 .06
Employed at the time of court hearing
.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
The characteristics and on-road behaviour of unlicensed drivers 308
Table I4
Standard multiple regression of environmental facilitating factors on frequency of
unlicensed driving (n=299)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1
.99 .47
Unaware of being unlicensed
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)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio
Upper Lower
Unaware of being unlicensed
-.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
The characteristics and on-road behaviour of unlicensed drivers 309
Table I6
Standard multiple regression of environmental facilitating factors on intention to drive
unlicensed in the future (n=299)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1
.69 .77
Unaware of being unlicensed
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
unlicensed driving (n=296)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1
.98 .47
Perceived risk of apprehension (prior to detection)
3.32 1.82 -.02 .02 -.11
Knew fine for unlicensed driving
.14 .35 -.02 .08 -.01
Perceived severity of punishment
4.56 1.84 -.02 .02 -.07
Perceived certainty of punishment
5.36 1.92 .00 .01 .01
Perceived swiftness of punishment
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
The characteristics and on-road behaviour of unlicensed drivers 310
Table I8
Logistic regression analysis of continued driving after detection as a function of
classical deterrence variables (n=298)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio
Upper Lower
Perceived risk of apprehension (after detection)
-.18 .07 7.57** .84 .74 .95
Knew fine for unlicensed driving
-.15 .39 .14 .86 .40 1.86
Perceived severity of punishment
.07 .07 .98 1.08 .93 1.24
Perceived certainty of punishment
-.14 .07 4.38* .87 .76 .99
Perceived swiftness of punishment
-.01 .08 .00 1.00 .86 1.15
Prior conviction for unlicensed driving
.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
unlicensed in the future (n=299)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1
.69 .77
Perceived risk of apprehension (after detection)
4.58 2.01 -.08*** .02 -.21 .04
Knew fine for unlicensed driving
.14 .34 .07 .13 .03
Perceived severity of punishment
4.57 1.83 -.01 .02 -.02
Perceived certainty of punishment
5.38 1.92 -.02 .02 -.06
Perceived swiftness of punishment
5.18 1.78 -.03 .03 -.07
Prior conviction for unlicensed driving
.39 .49 .33*** .09 .21 .04
Exposure to enforcement .26 .44 -.04 .10 -.02
.10*** .08
1. Logarithmically transformed. Model: F (7, 291) = 4.56, p < .001 Unique variability = .08; shared variability = .02. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 311
Table I10
Standard multiple regression of expanded deterrence variables on frequency of
unlicensed driving (n=292)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1
.97 .47
Perceived risk of apprehension (prior to detection)
3.33 1.83 -.03* .02 -.11 .01
Knew fine for unlicensed driving
.14 .34 -.03 .08 -.03
Perceived severity of punishment
4.55 1.84 -.02 .02 -.06
Perceived certainty of punishment
5.37 1.92 .00 .01 .01
Perceived swiftness of punishment
5.17 1.79 -.02 .02 -.06
Prior conviction for unlicensed driving
.39 .49 -.02 .06 -.02
Exposure to enforcement .26 .44 -.02 .07 -.02
Punishment avoidance .37 .48 .29*** .06 .30 .07
Vicarious exposure to punishment
.29 .45 .01 .06 .08
Vicarious exposure to punishment avoidance
.36 .48 -.01 .06 -.01
.12*** .09
1. Logarithmically transformed. Model: F (10, 281) = 3.99, p < .001 Unique variability = .08; shared variability = .04. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 312
Table I11
Logistic regression analysis of continued driving after detection as a function of
expanded deterrence variables (n=294)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio
Upper Lower Perceived risk of
apprehension (after detection)
-.14 .07 4.49* .87 .76 .99
Knew fine for unlicensed driving
-.35 .42 .73 .70 .31 1.58
Perceived severity of punishment
.09 .08 1.31 1.09 .94 1.27
Perceived certainty of punishment
-.13 .07 3.31 .88 .77 1.01
Perceived swiftness of punis hment
-.02 .08 .05 .98 .84 1.15
Prior conviction for unlicensed driving
.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
Vicarious exposure to punishment
.79 .30 6.99** 2.20 1.23 3.93
Vicarious exposure to punishment avoidance
.18 .29 .39 1.20 .68 2.10
Full model vs constant-only model: χ2 (10, 294) = 25.40, p < .01; Nagelkerke R2 = .12. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 313
Table I12
Standard multiple regression of expanded deterrence variables on intention to drive
unlicensed in the future (n=295)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1
.68 .77
Perceived risk of apprehension (after detection)
4.59 2.00 -.07** .02 -.18 .03
Knew fine for unlicensed driving
.14 .34 .02 .13 .01
Perceived severity of punishment
4.57 1.84 -.01 .02 -.02
Perceived certainty of punishment
5.38 1.91 -.01 .02 -.03
Perceived swiftness of punishment
5.18 1.78 -.03 .03 -.08
Prior conviction for unlicensed driving
.39 .49 .30*** .09 .19 .03
Exposure to enforcement .26 .44 .07 .11 .04
Punishment avoidance .37 .48 .21* .10 .13 .01
Vicarious exposure to punishment
.29 .45 .16 .10 .10
Vicarious exposure to punishment avoidance
.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
The characteristics and on-road behaviour of unlicensed drivers 314
Table I13
Hierarchical regression of deterrence variables on frequency of unlicensed driving
(n=292)
Variables Mean Std. dev
B Std. error
ß R2 Adj R2 ∆R2
Step one
Perceived risk of apprehension (prior to detection)
3.33 1.83 -.03* .02 -.11
Knew fine for unlicensed driving
.14 .34 -.03 .08 -.02
Perceived severity of punishment
4.55 1.84 -.02 .02 -.07
Perceived certainty of punishment
5.37 1.92 .00 .02 -.01
Perceived swiftness of punishment
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
.04 .02
Step two
Punishment avoidance .37 .48 .29*** .06 .30
Vicarious exposure to punishment
.29 .45 .07 .06 .08
Vicarious exposure to punishment avoidance
.36 .48 -.01 .06 -.01
.12*** .09 .08***
* p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 315
Table I14
Sequential logistic regression analysis of continued driving after detection as a function
of deterrence variables (n=294)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio
Upper Lower
Block 11
Perceived risk of apprehension (after detection)
-.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
Punishment avoidance .47 .32 2.14 1.60 .85 3.00
Vicarious exposure to punishment
.79 .30 6.99** 2.20 1.23 3.93
Vicarious exposure to punishment avoidance
.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
The characteristics and on-road behaviour of unlicensed drivers 316
Table I15
Hierarchical regression of deterrence variables on intention to drive unlicensed in the
future (n=295)
Variables Mean Std. dev
B Std. error
ß R2 Adj R2 ∆R2
Step one
Perceived risk of apprehension (after detection)
4.59 2.00 -.08** .02 -.20
Knew fine for unlicensed driving
.14 .34 .06 .13 .03
Perceived severity of punishment
4.57 1.84 -.01 .02 -.02
Perceived certainty of punishment
5.38 1.91 -.02 .02 -.04
Perceived swiftness of punishment
5.18 1.78 -.03 .03 -.08
Prior conviction for unlicensed driving
.39 .49 .33*** .09 .21
Exposure to enforcement .26 .44 -.03 .10 -.02
.09*** .07
Step two
Punishment avoidance .37 .48 .21* .10 .13
Vicarious exposure to punishment
.29 .45 .16 .10 .10
Vicarious exposure to punishment avoidance
.36 .48 .67 .09 .10
.14*** .10 .04**
* p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 317
Table I16
Standard multiple regression of social learning variables on frequency of unlicensed
driving (n=297)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1
.98 .47
Total unlicensed driving models
4.36 5.94 -.01 .01 -.06
Differential association (behavioural dimension)
1.23 .42 .13 .07 .11
Differential association (normative dimension)
10.57 5.57 .01 .01 .14
Attitudes to unlicensed driving
37.61 12.09 -.01* .00 -.17 .01
Attitudes to alternative behaviours
18.79 7.13 - .02*** .00 -.26 .06
Anticipated rewards 9.58 5.29 -.00 .01 -.01
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
Logistic regression analysis of continued driving after detection as a function of social
learning variables (n=297)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio Upper Lower
Total unlicensed driving models
-.02 .03 .61 .98 .94 1.03
Differential association (behavioural dimension)
.87 .33 7.03** 2.38 1.26 4.53
Differential association (normative dimension)
.09 .03 7.36** 1.09 1.02 1.16
Attitudes to unlicensed driving
.01 .02 .15 1.01 .98 1.04
Attitudes to alternative behaviours
-.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
Full model vs constant-only model: χ2 (7, 297) = 44.30, p < .001; Nagelkerke R2 = .20. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 318
Table I18
Standard multiple regression of social learning variables on intention to drive
unlicensed in the future (n=298)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1
.68 .77
Total unlicensed driving models 4.35 5.93 .01 .01 .08
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
Attitudes to alternative behaviours 18.76 7.14 -.01* .01 -.10 .01
Anticipated rewards 9.57 5.28 -.01 .01 -.08
Anticipated punishments 54.05 10.30 .01** .01 -.17 .02
.30*** .29
1. Logarithmically transformed. Model: F (7, 290) = 18.05, p < .001 Unique variability = .08; shared variability = .22. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 319
Table I19
Standard multiple regression of expanded deterrence variables on frequency of
unlicensed driving for disqualified, not currently licensed and never licensed drivers
(n=94)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1 .87 .52
Perceived risk of apprehension (prior to detection)
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
Exposure to enforcement .26 .44 -.13 .13 -.11
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
1. Logarithmically transformed. Model: F (10, 83) = 2.75, p < .01 Unique variability = .10; shared variability = .15. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 320
Table I20
Standard multiple regression of expanded deterrence variables on frequency of
unlicensed driving for suspended, expired and not appropriately licensed drivers
(n=198)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1 1.02 .44
Perceived risk of apprehension (prior to detection)
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
Exposure to enforcement .26 .44 .02 .08 .02
Punishment avoidance .37 .48 .28*** .07 .31 .07
Vicarious exposure to punishment .26 .44 .06 .07 .06
Vicarious exposure to punishment avoidance .31 .46 -.03 .07 -.03
.12** .07
1. Logarithmically transformed. Model: F (10, 187) = 2.50, p < .01 Unique variability = .07; shared variability = .05. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 321
Table I21
Logistic regression analysis of continued driving after detection as a function of
expanded deterrence variables for disqualified, not currently licensed and never
licensed drivers (n=94)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio Upper Lower
Perceived risk of apprehension (after detection)
-.22 .14 2.30 .81 .61 1.07
Knew fine for unlicensed driving
-1.75 .86 4.19* .17 .03 .93
Perceived severity of punishment
.21 .16 1.76 1.24 .90 1.70
Perceived certainty of punishment
-.04 .15 .08 .96 .72 1.28
Perceived swiftness of punishment
-.23 .17 1.71 .80 .57 1.12
Prior conviction for unlicensed driving
.99 .62 2.51 2.68 .79 9.09
Exposure to enforcement .73 .68 1.17 2.08 .55 7.85
Punishment avoidance .70 .63 1.23 2.01 .58 6.94
Vicarious exposure to punishment
.21 .59 .13 1.23 .39 3.88
Vicarious exposure to punishment avoidance
.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
The characteristics and on-road behaviour of unlicensed drivers 322
Table I22
Logistic regression analysis of continued driving after detection as a function of
expanded deterrence variables for suspended, expired and not appropriately licensed
drivers (n=200)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio Upper Lower
Perceived risk of apprehension (after detection)
-.11 .08 1.85 .90 .77 1.05
Knew fine for unlicensed driving
.20 .53 .14 1.22 .43 3.40
Perceived severity of punishment
.04 .09 .148 1.04 .87 1.24
Perceived certainty of punishment
-.16 .08 3.54 .85 .73 1.01
Perceived swiftness of punishment
.06 .09 .37 1.06 .88 1.26
Prior conviction for unlicensed driving
-.13 .40 .11 .88 .40 1.92
Exposure to enforcement .48 .43 1.29 1.62 .70 3.73
Punishment avoidance .45 .39 1.36 1.57 .74 3.35
Vicarious exposure to punishment
1.05 .37 7.97** 2.87 1.38 5.97
Vicarious exposure to punishment avoidance
.01 .36 .00 1.01 .50 2.06
Full model vs constant-only model: χ2 (10, 200) = 18.70, p < .05; Nagelkerke R2 = .13. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 323
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 (n=94)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1 .79 .79
Perceived risk of apprehension (after detection)
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
Exposure to enforcement .26 .44 .34 .18 .19
Punishment avoidance .36 .48 .22 .17 .13
Vicarious exposure to punishment .35 .48 -.19 .16 -.12
Vicarious exposure to punishment avoidance .45 .50 .29 .15 .18
.35*** .27
1. Logarithmically transformed. Model: F (10, 84) = 4.48, p < .001 Unique variability = .12; shared variability = .23. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 324
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
(n=200)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1 .63 .76
Perceived risk of apprehension (after detection)
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 .
Exposure to enforcement .27 .44 -.06 .13 -.03
Punishment avoidance .37 .48 .21 .12 .13
Vicarious exposure to punishment .26 .44 .33** .13 .19 .03
Vicarious exposure to punishment avoidance .31 .46 .07 .12 .05
.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
The characteristics and on-road behaviour of unlicensed drivers 325
Table I25
Standard multiple regression of social learning variables on frequency of unlicensed
driving for disqualified, not currently licensed and never licensed drivers (n=93)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1 .86 .51
Total unlicensed driving models 5.83 6.40 -.00 .01 -.05
Differential association (behavioural dimension) 1.32 .47 .19 .13 .18
Differential association (normative dimension) 10.80 5.59 .01 .01 .12
Attitudes to unlicensed driving 39.09 11.60 -.01 .01 -.04
Attitudes to alternative behaviours 19.46 7.41 -.01* .01 -.22 .04
Anticipated rewards 10.81 6.18 .01 .01 .07
Anticipated punishments 54.98 9.94 -.01 .01 -.13
.15* .08
1. Logarithmically transformed. Model: F (7, 85) = 2.20, p < .05 Unique variability = .04; shared variability = .11. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 326
Table I26
Standard multiple regression of social learning variables on frequency of unlicensed
driving for suspended, expired and not appropriately licensed drivers (n=204)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Frequency of unlicensed driving1 1.03 .44
Total unlicensed driving models 3.69 5.60 -.00 .01 -.03
Differential association (behavioural dimension) 1.19 .39 .09 .08 .08
Differential association (normative dimension) 10.46 5.58 .01 .01 .14
Attitudes to unlicensed driving 36.93 12.28 -.01 .00 -.19
Attitudes to alternative behaviours 18.48 6.99 -.01*** .01 -.26 .06
Anticipated rewards 9.02 4.74 -.00 .01 -.04
Anticipated punishments 53.57 10.44 -.01 .00 -.13
.08* .05
1. Logarithmically transformed. Model: F (7, 196) = 2.46, p < .05 Unique variability = .06; shared variability = .02. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 327
Table I27
Logistic regression analysis of continued driving after detection as a function of social
learning variables for disqualified, not currently licensed and never licensed drivers
(n=92)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio Upper Lower
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 rewards .06 .05 .52 1.04 .94 1.14
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
Logistic regression analysis of continued driving after detection as a function of social
learning variables for suspended, expired and not appropriately licensed drivers
(n=205)
95% CI for Odds ratio Variables B Std.
error Wald test
Odds Ratio
Upper Lower
Total unlicensed driving models
-.02 .03 .42 .98 .92 1.04
Differential association (behavioural dimension)
1.05 .41 6.48 2.85 1.27 6.39
Differential association (normative dimension)
.08 .04 4.69 1.09 1.01 1.17
Attitudes to unlicensed driving
.01 .02 .21 1.01 .97 1.05
Attitudes to alternative behaviours
-.04 .03 2.13 .96 .92 1.01
Anticipated rewards .03 .04 .70 1.03 .96 1.11
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
The characteristics and on-road behaviour of unlicensed drivers 328
Table I29
Standard multiple regression of social learning variables on intention to drive
unlicensed in the future for disqualified, not currently licensed and never licensed
drivers (n=93)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1 .81 .78
Total unlicensed driving models 5.83 6.40 .02 .01 .13
Differential association (behavioural dimension) 1.32 .47 .39* .15 .23 .04
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 rewards 10.81 6.18 -.02 .01 -.19
Anticipated punishments 54.98 9.94 -.03*** .01 -.39 .10
.46*** .41
1. Logarithmically transformed. Model: F (7, 85) = 10.12, p < .001 Unique variability = .21; shared variability = .25. * p < .05 ** p < .01 *** p < .001
The characteristics and on-road behaviour of unlicensed drivers 329
Table I30
Standard multiple regression of social learning variables on intention to drive
unlicensed in the future for suspended, expired and not appropriately licensed drivers
(n=205)
Variables Mean Std. dev
B Std. error
ß sr2 R2 Adj R2
Intention to drive unlicensed in future1 .63 .76
Total unlicensed driving models 3.67 5.59 .01 .01 .04
Differential association (behavioural dimension) 1.19 .39 .20 .12 .10
Differential association (normative dimension) 10.45 5.56 .02 .01 .27
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 rewards 9.01 4.73 -.01 .01 -.05
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