Re-conceptualising the reckless driving behaviour of young drivers

Re-conceptualising the reckless driving behaviour of young drivers

Brenton McNally* and Graham L. Bradley

School of Applied Psychology, Griffith University, Gold Coast, Australia, Behavioural Basis

of Health Research Centre, Griffith Health Institute, Queensland, Australia

Abstract

Reckless driving is a major contributing factor to road morbidity and mortality. While further

research into the nature and impact of reckless driving, particularly among young people, is

urgently needed, the measurement of reckless driving behaviour also requires increased

attention. Three major shortcomings apparent in established measures of driver behaviour are

that they do not target the full range of reckless driving behaviours, they measure

characteristics other than driving behaviours, and/or they fail to categorise and label reckless

driver behaviour based on characteristics of the behaviours themselves. To combat these

shortcomings, this paper reports the development and preliminary validation of a new

measure of reckless driving behaviour for young drivers. Exploratory factor analysis of self-

reported driving data revealed four, conceptually distinct categories of reckless driving

behaviour: those that increase crash-risk due to (a) distractions or deficits in perception,

attention or reaction time (labelled “distracted”), (b) driving under the influence of drugs or

alcohol (labelled “substance-use”), (c) placing the vehicle in an unsafe environment beyond

its design expectations (labelled “extreme”), and (d) speed and positioning of the vehicle

relative to other vehicles and objects (labelled “positioning”). Confirmatory factor analysis of

data collected from a separate, community sample confirmed this four-factor structure.

Multiple regression analyses found differences in the demographic and psychological

variables related to these four factors, suggesting that interventions in one reckless driving

domain may not be helpful in others.

Keywords: Reckless driving; young drivers; scale development; driver behaviour; factor

analysis; risky behaviour

Email address: [email protected] (B. McNally), [email protected] (G. L.

Bradley)

*Corresponding Author: Tel: +61 7 56788119.

1. Introduction

Research is increasingly suggesting that risk-taking is a multi-dimensional construct

(Gibbons et al., 2006). Consistent with this, several researchers have sought to categorise

driving behaviours into meaningful subtypes (Dula & Geller, 2003; Maslowsky et al., 2010;

Reason et al., 1990; Schmidt, 2012). Arnett (1992) distinguishes between risk and reckless

behaviours, defining risk behaviours as thrill-seeking activities that have mainstream social

approval, and defining reckless behaviours as those lacking such social approval, carrying

stronger connotations for negative consequences, and involving a failure to take available

precautions. Reckless driving behaviours satisfy all three of Arnett‟s criteria of recklessness:

they lack mainstream social approval and may even involve violations of the law (Jessor et

al., 1997; Maslowsky et al., 2010); they carry strong connotations of negative consequences

by placing drivers and/or their passengers at risk of morbidity, mortality, and other negative

outcomes (Patil et al., 2006); and, by definition, they involve deliberate deviations from safe

driving (Malta, 2004).The focus of the current research was thus on behaviours in which

there is a failure to take available precautions, while other behaviours, described elsewhere as

errors (Reason et al., 1990), were excluded from study.

We aimed to re-conceptualise reckless driving behaviours and thereby contribute to

the existing road safety literature by identifying coherent and relevant subtypes of behaviour

within this domain. Achievement of this goal will assist future research in uncovering the

antecedents of conceptually distinct categories of behaviour and designing interventions

suitable for each behavioural category. The focus of the current research was also on the

measurement of these behaviours in young drivers, a group requiring better informed road

safety interventions (Glendon et al., 2014).

1.1. Types of reckless driving behaviour

The literature provides many examples of what could be described as reckless driving

behaviour. Perhaps the most widely studied example is driving a vehicle at excessive speeds.

Speeding has been found to increase the likelihood of crash involvement, commonly

implicated in vehicle crashes (Blows et al., 2005; Rundmo & Iversen, 2002; Fernandes et al.,

2007; Parker, 2002; Simons-Morton et al., 2005; Smart et al., 2005). Smart and colleagues

(2005) found that 31% of a large sample of young drivers had been detected speeding at least

once, and 80% reported exceeding the speed limit by 10 km/hr in at least one of their last ten

trips. Of particular relevance to the current research, young drivers report speeding more

frequently than do older drivers (Clarke et al., 2005; Hatfield & Fernandes, 2009), with

speeding found to be the strongest crash-risk factor for those under the age of 25 years

(Blows et al., 2005).

Other examples of driving behaviours that could be described as reckless include

close passing and tailgating (Harris & Houston, 2010), running red lights (Fergusson et al.,

2003), changing lanes, and overtaking when unsafe to do so (Hartos et al., 2002). Becoming

increasingly common (Neighbors et al., 2002), these behaviours heighten the likelihood of

traffic violations and vehicle crashes (Hartos et al., 2002). Research also suggests high

prevalence of these reckless driving behaviours among younger drivers (Agerwala et al.,

2008; Krahe & Fenske, 2002; Shinar & Compton, 2004; Wickens et al., 2011).

Driving whilst using a cell phone, including dialing (Klauer et al., 2006) and texting

(Owens et al., 2011), driving whilst using MP3 players and other electronic devices

(Chisholm et al., 2008), driving whilst under the influence of alcohol and other psychoactive

substances, including cannabis (Calafat et al., 2009; Richer & Bergeron, 2009), and driving

whilst tired or fatigued (Smart et al., 2005; Smith et al., 2005), are additional examples of

driving behaviour that fall under the banner of reckless. These behaviours have been found to

compromise both driving performance and driving safety (Anderson & Baumberg, 2006;

Berghaus et al., 1995; Charlton, 2009; Deery & Lowe, 1996; Kircher et al., 2004; Lenne et

al., 2000; Owens et al., 2011; Potter, 2000; Raemakers et al., 2004; Sexton et al., 2000; Smith

et al., 2005; Strayer & Johnson, 2001). Research by McEvoy and colleagues suggests that

young drivers are more likely to have had a crash due to the distraction of a cell phone than

are older drivers (McEvoy et al., 2006). Armstrong and colleagues (2005) studied 331 young

drivers (mean age 24 years) and found that 26% reported driving under the influence of

psychoactive drugs at least once in their lifetime. Smart and colleagues (2005) found that

64% of young drivers reported recently driving whilst very tired, compounding the effect

demonstrated in other research that young drivers are more prone to the effects of fatigue

than are older drivers (Smith et al., 2005).

1.2. Measuring self-reported reckless driving behaviour

Many road safety researchers use established self-report scales to measure driving

behaviour. Arguably the most commonly used is the Driver Behaviour Questionnaire (DBQ)

(Reason et al., 1990). de Winter and Dodou (2010) identified 174 English-language studies

that have used the DBQ or a modified version thereof. However, recent evidence exists

asserting that the association of the DBQ with important outcome variables (e.g., motor

vehicle crashes) may be quite modest, with this relationship potentially inflated due to

common-method variance (af Wåhlberg, Dorn, & Kline, 2011). Since the influential research

of Reason and colleagues, where 50 items were used to differentiate between errors and

violations (reckless behaviours), many researchers have developed their own measures of

driver behaviour. Responses are typically required on either Likert or frequency scales.

Instruments vary in length from less than 12 to upwards of 70 items, and in structure, with the

number of factors or categories of behaviour ranging from one to upwards of eight. Table 1

provides an illustrative list of published research in which measures of self-reported driving

behaviour were either developed and/or refined. It must be noted that while several of these

measures have been used in studies of reckless driving behaviour, not one was designed

specifically for this purpose. With the current focus on reckless behaviour, studies were

included in this table if they targeted driving behaviour (and not related constructs like driver

stress, e.g., Driver Behaviour Inventory, Gulian et al., 1989), and if they measured more than

one type of reckless driving behaviour (including behaviours variously labeled as “risky”,

“violations”, and “aggressive”). As can be seen, the table specifies the number of factors or

subscales present in each measure, and separately indicates whether the scale items fit into

each of five different driving content categories (plus an “other” category). Note that, in some

studies, a single factor spans several of the Table 1 content categories, reflecting the fact that

the items comprising that single factor pertain to multiple types of behaviour.

[Table 1 around here.]

Most existing measures assess driver behaviour in general, and are thus not designed

to focus on reckless behaviours as herein defined. This concentration on the broad, rather

than the narrow is surprising given, as already noted, reckless behaviours contribute

disproportionately to driver crash-risk, especially young driver crash-risk. As such, driving

behaviours that are reckless warrant independent and urgent investigation. Existing

instruments are limited in their ability to measure reckless driving behaviours in three

important ways, although most do not possess more than one of these attributes. First, some

instruments are incomplete in that they do not include items that represent the diversity of

reckless driving behaviours (such as those discussed in the previous section). Second, some

are over inclusive which often leads to the measurement of mixed constructs and non-

behavioural phenomena rather than focusing solely on reckless behaviour. Third, some are

divided into factors or categories that overlap and are not independent. Many such scales

categorise and label reckless driver behaviour based on the assumed intentions and

motivations of the driver, as opposed to characteristics of the behaviours themselves.

Incompleteness

As can be seen in Table 1, many scales designed to measure reckless driving

behaviour do not measure the diversity that exists in this domain. Of particular import is the

insufficient focus on two categories of behaviours. First, there is a neglect of behaviours that

involve distraction, especially by technological devices. Although it is unrealistic to expect

twenty year-old scales to refer to behaviours that have emerged only recently as a function of

advancing technology, recent versions of the DBQ have rarely updated this particular aspect

(e.g., Kontogiannis et al., 2002; Lawton et al., 1997; Xie & Parker, 2002). A second reckless

behaviour that is commonly neglected, despite evidence supporting its importance to young

driver risk (Smith et al., 2005), is driving whilst fatigued or extremely tired. In order to be

comprehensive and remain relevant, the measurement of reckless driving behaviours must

maintain an updated repository of items that reflect the entire range of reckless driving

behaviours.

Over-Inclusiveness

A second limitation of some existing measures, and the classificatory schemes upon

which they are based, is the inclusion of targets that lie outside the domain of reckless driving

behaviours. Two particular aspects of over-inclusiveness exist: a) references to phenomena

that are either partially or entirely non-behavioural, and b) references to phenomena that are

behaviour-related, but are not considered reckless. Regarding the first of these examples of

over-inclusiveness, inspection of the instruments listed in Table 1 reveals that many items

refer, at least in part, to internal psychological states (e.g., goals, motives, feelings), rather

than behaviours. In reference to the second type of over-inclusiveness, some scales include

items that refer to in-vehicle behaviours that are not acts of reckless driving.

Whilst compiling this list, it was evident that many scales utilise items from a range of

non-behavioural domains. For example, the Dula Dangerous Driving Index (Dula & Ballard,

2003) includes attitudinal items similar to “I feel that most traffic „laws‟ could be considered

suggestions”, in addition to items measuring actual behaviour. Similarly, the Multi-

dimensional Driving Style Questionnaire (Taubman – Ben-Ari et al., 2004) includes items

measuring interests and personal preferences similar to “I like the thrill of flirting with death

or disaster”. Items in some instruments confound the measurement of overt behaviours with

their underlying motives. For example, the Risk Behaviour Scale (Ulleberg & Rundmo,

2002) includes items similar to “Drive fast to show that I am tough enough” which tap

motivations when assessing behaviour. Double-barreled items such as these place an

unnecessary burden on respondents and render the meaning of their response unclear.

Measures that embrace general driving behaviour are similarly over-inclusive in that

they index behaviour that falls outside of the definition of “reckless”. For example, the

Behaviour of Young Novice Drivers scale (BYNDS, Scott-Parker, et al., 2010), and the

Youth Domains of Risky Driving scale (Schmidt, 2012) both include behavioural items

similar to “You drove on the weekend” and “You made a fist at another driver” that are not in

themselves inherently reckless. Similarly, doubts may be raised about the content validity of

the DBQ items “Use your horn to indicate your annoyance to another road user” and “Get

angry at a certain type of driver and chase them with the intention of showing them how

angry you are.” While pertinent to their research topics, these non-reckless items render the

scales unsuitably multidimensional, thereby diluting the measurement of the construct of self-

reported reckless driving behaviour.

Overlapping Categories

Finally, while most of the scales listed in Table 1 include subtypes of reckless driving

behaviours, these types are often difficult to distinguish from one another. In terms of

categorising or labeling groups of reckless driving behaviours, most scales in the compiled

list are consistent with Reason and colleagues‟ (1990) ideation that this domain of behaviours

requires description in terms of social and motivational antecedents. This has lead to a failure

to place different reckless driving behaviours in homogenous conceptual categories based on

the characteristics of the behaviours themselves (Schmidt, 2012). Describing behaviour on

the basis of motivations becomes problematic when the same behaviour, for example

speeding, can be committed for a variety of reasons (e.g., to test the vehicle‟s capabilities, to

be aggressive, or for the instrumental goal of time-saving). By categorising by presumed

motivations, behaviours are capable of being allocated to multiple, different groups. For

example, Lawton and colleagues‟ (1997) initial category of “normal highway code

violations” includes items measuring a diversity of behaviours, such as “Risky over-taking”,

“Close following”, and “Drink and drive”. A second of the researchers‟ categories is labeled

“aggressive violations”, because it is assumed that individuals who perform them do so by

“acting aggressively toward other road users” (p. 1262). With the first of these categories

defined in legal terms and the second defined in terms of assumed underlying motivations, it

is logically possible for a single act to be coded into both categories. As evidenced by

subsequent analyses where new items were introduced, these two categories of behaviour

(normal highway code violations and aggressive violations) are not independent since items

from each category were found to load onto the same factor.

The current research sought to conceptualise and measure reckless driving behaviours

in a way that overcomes the three deficiencies evident in many existing approaches. In

particular, we sought to develop a measure that provides both focus on a specific category of

driving behaviours – reckless behaviours – and breadth of coverage of behaviours within this

category. If successful, practitioners and researchers will have access to a validated scale for

use in distinguishing specific types of driving recklessness, identifying the shared and unique

determinants of these behaviours, and developing interventions that are precisely targeted to

combat the most problematic driving behaviours. Study 1 aimed to measure a diversity of

reckless driving behaviours that conform to Arnett‟s (1992) definition. After the selection of

relevant behavioural items and collection of frequency data pertaining to them, exploratory

factor analysis was used to discover their underlying factorial structure and thereby identify

and differentiate specific subtypes of reckless driving behaviour. Using an independent

sample, Study 2 aimed to confirm the findings of Study 1, to provide additional evidence as

to the reliability and concurrent validity of the behaviour subtypes, and to determine whether

differences exist in how each of the four factors relate to relevant demographic and

psychological variables.

2. Study 1

2.1. Method

2.1.1. Participants

Study 1 used a convenience sample of 189 undergraduate psychology students

attending an Australian university who participated as part of their course program and to

receive course credit. Ages ranged from 17 to 25 years (M = 20.04, SD = 1.83), with 147 of

the participants female. Participants were required to have Australian citizenship or

residency, a current Australian driver‟s licence, and regular access to a motor vehicle they

were licensed to drive.

2.1.2. Materials and procedure

Participants were asked to indicate how many times they have engaged in each of 21

reckless driving behaviours in the preceding six months. Selection of these behaviours was

derived from a variety of sources, including the studies and scales listed in Table 1. Particular

emphasis was placed on selecting items that represented the diversity of behaviours

considered reckless, with all items worded so as to refer to a single type of reckless driving

behaviour. Items deliberately represented domains of speeding (e.g. “Driven at least 15km/hr

above the speed limit”), steering (e.g. “Changed lanes frequently on a multi-lane road”),

distracted driving (e.g. “Texted or looked for numbers on your phone or searched for songs

on your MP3 player whilst driving”), fatigued driving (“Driven whilst extremely tired or

exhausted”), and other, more extreme, behaviours (e.g. “Raced or chased another vehicle

driven by someone you do not know”). Participants were required to answer on an 11-point

scale ranging from 0 (“none”) to 10 (“100+ times (i.e. at least 4 times a week)”) in accord

with past research on reckless behaviour (Bradley & Wildman, 2002; Teese & Bradley,

2008). The anonymous, online, self-report questionnaire, which also included items not

related to the current study, took approximately 20-25 minutes to complete.

2.2. Results

The data were examined to ensure the appropriateness of an exploratory factor

analysis. All items correlated at least .3 with at least one other item, suggesting reasonable

factorability. Sampling adequacy was deemed appropriate, KMO = .86, and Bartlett‟s test of

sphericity was significant, 2

(210) = 1763.12, p < .001. Finally, all communalities were

above .3. Principal axis factoring extraction revealed four clear factors. Initial eigenvalues

indicated that the factors explained 34%, 11%, 7%, and 6% of the variance, respectively.

Parallel analysis was also performed to ensure the factor structure was appropriate. Applying

methods developed by O‟Connor (2000), including the use of 1000 raw data permutations for

multivariate non-normal data, the analyses reinforced the four factor structure. Following

oblique (direct oblimin.) rotation of the factor-loading matrix, two items (“Failed to stop at a

stop sign” and “Driven through an orange traffic light when easily avoidable or when it is

about to turn red”) were excluded due to loading below .4 on all factors. A second analysis

using the remaining 19 items yielded a clean solution with all items having primary loadings

over .4 and no items cross-loading in excess of .3. These results can be seen in Table 2.

[Table 2 around here.]

From the groupings of items with primary loadings on each factor, the first factor was

indicative of behaviours that increase the probability of a vehicle crash due to distraction or a

deficit in perception, attention or reaction time (labelled “distracted”). The second factor

represented behaviours that increase the probability of a vehicle crash due to driving under

the influence of drugs or alcohol (labelled “substance-use”). The third factor corresponded to

behaviours that increase the probability of a vehicle crash due to placing the vehicle in an

unsafe environment that places stress upon the vehicle close to, or beyond, its design

expectations (labelled “extreme”, in reference to the prevalence of these behaviours in

„extreme‟ motorsports). Finally, the fourth factor was indicative of behaviours that increase

the probability of a vehicle crash due to the speed and/or position of the vehicle relative to

other vehicles and/or objects (labelled “positioning”). This final factor includes both the

“speeding” and the “steering” behaviours that are represented separately in several existing

measures.

Composite scores were created for each of the four factors by averaging responses to

items that had their primary loadings on each factor. Ranging from 0 to 10, higher scores

indicate greater frequency of participation in each behaviour category. Correlations between

these composite scores can be seen in Table 2. Cronbach‟s alpha ranged from .70 (for the

two-item substance use factor) to .84, indicating satisfactory scale reliability.

2.3. Discussion

Factor analyses from Study 1 examined a set of 21 driving behaviours derived from

past research, all of which conformed to Arnett‟s (1992) definition of reckless behaviour.

Results indicated that 19 of these behaviours loaded onto four factors. The selection of a

diverse set of items, with the wording of all items referring to a single type of reckless driving

behaviour, and the categorisation of these behaviour items on empirical grounds, enabled the

development of a new way of classifying and measuring self-reported reckless driving

behaviours. Most importantly, this new conceptualisation and measurement approach

addresses three major concerns inherent in many previous instruments. First, the current

instrument assesses the full range of reckless driving behaviours, including several that are

not measured by many existing instruments. Second, by following Arnett‟s (1992) definition

of reckless behaviour when selecting items, the current instrument provides a content valid

measure of self-reported reckless driving behaviour, one that avoids references to extraneous

variables. Third, the exploratory nature of the analyses provided conceptually distinct

categories of reckless driving behaviour, labelled on the basis of characteristics of the

behaviours themselves.

Respondents reported engaging in distracted behaviours more often than the other

behaviour types. Interestingly, this subset of behaviours is absent from many existing

measures (see Table 1). Distracted behaviours were also strongly correlated with the subset of

behaviours with the second highest reported frequency, positioning behaviours. Despite all

items representing reckless driving behaviours, the correlations between subsets were, for the

majority, moderate, with variance shared by pairs of subscales never exceeding 50%.

As previously noted, two items were excluded from the final factor solution due to

low factor loadings on all factors. Review of these items and focus group discussions

suggested that the wording of these items lacked specificity and clarity. Wording changes

were made so that the items could be reassessed in Study 2. The revised items that were

included with the 19 used in Study 1 were “Failed to completely stop at a stop sign” and

“Driven through a traffic light when it is about to turn red or could turn red as you pass

through”.

3. Study 2

The aims of Study 2 were three-fold. First, the study aimed to determine whether the

reckless driving behaviour factor structure elicited in Study 1 adequately fits a data set

obtained from a separate, larger, and more diverse sample. Second, Study 2 aimed to obtain

additional evidence as to the scale‟s reliability and validity, and third, it sought to identify

any differences in the pattern of demographic and psychological variables that predict each of

the behaviour types. In pursuit of this third aim, the five predictors briefly described below

were selected for study based on their history of association with driving behaviour.

Gender has been found to be a strong predictor of reckless driving behaviours (Harre

et al., 2000; Turner & McClure, 2003), with males participating more than females (Bina et

al., 2006; Catchpole & Styles, 2005; Fergusson et al., 2003; McEvoy et al., 2006; Oltedal &

Rundmo, 2006). These gender differences have been reliably replicated (Lonczak et al.,

2007), and we thus expected that males would report performing reckless driving more often

than would their female counterparts. However, past studies have typically used driving

recklessness data that have been aggregated across multiple types of behaviour, raising the

possibility that reliable gender differences may not exist in all of several more narrowly-

defined categories of behaviour. Adding a further layer of complexity, Byrnes and colleagues

(1999), in a meta-analysis of 150 studies, observed that while males engage in greater risk

taking, the gap between the sexes is both context- and age-dependent. Together, these

observations suggest the need for greater understanding and considerable caution when

predicting gender differences in participation in specific reckless driving behaviours (Bina et

al., 2006).

Self-efficacy has been identified as important in a number of health-behaviour models

(Montano et al., 1997; Strecher & Rosenstock, 1997), with perceptions of ability to control

the vehicle and respond quickly and accurately to hazards (Groeger, 2006) found to predict

both safe and unsafe driving (George et al., 2007; Montag, 1989; Sarkar & Andeas, 2004;

Taubman- Ben-Ari et al., 2005; Victoir et al., 2005). Through this, the current research

expects stronger self-efficacy beliefs to be associated with increased self-reports for each

reckless driving behaviour.

Risk perception has been found to be related to self-reported participation in unsafe

driving behaviours (Jonah, 1986; Hatfield & Fernandes, 2009; Rundmo, 1995; Weinstein et

al., 1998). However, the psychological processes associated with risk judgements require

further investigation (McNally & Titchener, 2012; Price, 2001), especially in relation to the

situational influences on risk perception (Sjoberg et al., 2004) and whether young drivers

perceive less risk than older drivers (cf. Beyth-Marom et al., 1993; Hatfield & Fernandes,

2009; Williamson, 2003). Notwithstanding the need for this research, the current research

expects higher risk perception to be associated with decreased self-reports of each reckless

driving behaviour.

Risk willingness refers to the amount of risk an individual is prepared to accept in a

given situation (Lund & Rundmo, 2009). Deery (1999) found evidence to suggest young

drivers are more willing to accept risk than are older, more experienced drivers. Willingness

consistently explains variance in behaviour additional to intentions (Gibbons et al., 2003) and

has been found to be a better predictor of health risk behaviour than intentions, for

adolescents and emerging adults (Gibbons et al., 1998; Gibbons et al., 2004), including

unsafe driving behaviour (Gibbons et al., 1998). The current research thus expects higher

willingness to be associated with increased self-reports of each reckless driving behaviour.

Sensation-seeking, or the need for varied, novel and complex sensations and

experiences (Zuckerman, 1994), has been consistently found to predict unsafe driving

(Arnett, 1996; Deery & Fildes, 1999; Jonah, 1997), including driving under the influence of

alcohol and cannabis, speeding, racing and unsafe passing as well as a range of other reckless

driving behaviours (Armstrong et al., 2005; Arnett, 1996; Arnett et al., 1997; Burns & Wilde,

1995; Clement & Jonah, 1984; Dahlen et al., 2005; Furnham & Saipe, 1993; Greene et al.,

2000). Sensation-seeking is thought to explain the motivations underlying individuals taking

risks whilst driving because of the thrill these behaviours provide (Arnett, 1994). The current

research expects higher sensation-seeking to be associated with increased self-reports of each

reckless driving behaviour.

3.1. Method

3.1.1. Participants

Overall, 790 participants were sampled, consisting of 522 females, with ages ranging

from 17 - 25 years (M = 20.83, SD = 2.17). Participants were members of the general

Australian and New Zealand public as well as undergraduate psychology students who

completed the research in order to receive course credit. To ensure sampling adequacy, all

participants were required to have current Australian or New Zealand citizenship or

residency, hold a current Australian or New Zealand driver‟s license, and have regular access

to a motor vehicle they are licensed to drive. Of the sample, 469 participants had a

provisional licence (i.e., one that permits unsupervised but restricted driving).

3.1.2. Materials and procedure

Self-efficacy was assessed with a modified 6-item version of the Adelaide Driving

Self-Efficacy Scale (ADSES) (George et al., 2007). The scale asks participants to assess their

own driving ability by indicating how confident they are in performing six reckless driving

behaviours whilst maintaining the safety of themselves and those around them (regular

driving behaviours were the subject of the original version). Reckless behaviours included in

this scale were selected from each of the four factors from Study 1. Sample reckless driving

behaviours are “Driving when you suspect you are over the .05 blood alcohol limit”, and

“Turning, merging, or changing lanes without indicating”. Participants responded on a 7-

point scale, ranging from 1 (Not at All Confident) to 7 (Extremely Confident), with scores

summed such that higher scores indicate stronger self-efficacy beliefs. Pilot-testing (using the

Study 1 sample) revealed the modified scale to have high internal consistency (Cronbach‟s

alpha = .84), as well as construct and criterion-related validity.

To assess risk perception, participants were asked to indicate (a) the likelihood that

each of 16 reckless driving behaviours would result in vehicle crash and (b) the severity of

the subsequent, negative consequences (e.g., physical harm) of each behaviour. The 16

reckless behaviours were sampled from each of the four factors from Study 1. Sample items

include “Eating, drinking, or smoking whist driving”, ”Exceeding a decreased speed limit

(e.g., 40km/hr) in a road work zone by at least 15km/hr”, and “Racing or chasing another

vehicle driven by a friend or someone you know”. Participants responded on 5-point scales,

ranging from 1 (Not at All/ Minimal) to 5 (Very Much/ Substantial). In line with past practice

(cf. Lund & Rundmo, 2009; Sjoberg et al., 2004; van Gelder et al., 2009), responses to

corresponding likelihood and severity items were multiplied to obtain a measure of risk

perception for each behaviour, and these 16 scores were averaged such that higher scores

indicate higher total perceived risk.

In accordance with measures designed to assess behavioural willingness developed by

Gibbons and Gerrard (1995), willingness to take driving risks was assessed through four

hypothetical scenarios in which reckless driving may occur (e.g., “Suppose you are running

very late for an important job interview or some other important meeting but there is a

chance that if you travel quickly you may still be able to make it on time.”). Content of the

other scenarios varied from a night out with your friends, to a family member in hospital in a

critical condition, and evading a bushfire. After reading each hypothetical scenario,

participants indicated their willingness, in the given situation, to engage in each of six

reckless driving behaviours (thus, there was a total of 24 items over four scenarios).

Participants responded on a 5-point scale, ranging from 1 (Not at All Willing) to 5 (Extremely

Willing). Responses to all 24 items were summed so that higher scores indicated greater

willingness to participate in reckless driving behaviours.

Sensation-seeking was assessed using the Brief Sensation-Seeking Scale (BSSS)

(Hoyle et al., 2002). This scale consists of 8 items, adapted from previous measures of

sensation seeking (Zuckerman et al., 1978) and has been validated for adolescents and

emerging adults (Hoyle et al.). Internal consistency has been reported as moderate to strong

(Cronbach‟s alpha ranging from .74 -.79) (Hoyle et al.). Participants were asked to indicate

how well the statements (e.g., “I prefer friends who are excitingly unpredictable”) describe

themselves. Each of Zuckerman and colleagues‟ four primary dimensions of sensation-

seeking (experience seeking, boredom susceptibility, thrill and adventure seeking, and

disinhibition) are represented by two items, with responses indicated on a 5-point scale

ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). Scores were averaged so that a

high score reflected greater sensation-seeking tendencies.

Finally, participants were asked to indicate how many times they have engaged in the

21 reckless driving behaviours in the last six months. Response options and scoring were as

for Study 1, with responses to items in each factor averaged so that they varied from 0 to 10

and higher scores indicate greater frequency of participation in each behaviour category.

All data were collected through online, self-report, survey methods. Recruitment was

of undergraduate psychology students via a university run, online sign-up and notification

website (n = 123) as well as the general population through emailing lists (n = 586) and

notifications placed at Department of Transport and Main Roads customer service centres

located throughout Queensland, Australia (n = 81). Participants were provided a link to a

hosting website where the online questionnaire could be completed. Once they had accessed

the link, the anonymous questionnaire, which also included items not related to the current

study, took approximately 35 minutes to complete.

3.2. Results

3.2.1. Confirmatory factor analysis

SPSS and its associated Analysis of Moment Structures (AMOS) module version 19

were used to perform the following data analyses. A Confirmatory Factor Analysis (CFA)

was conducted to examine the fit of the four-factor, Study 1 model to the Study 2 data (n =

790). Assessment of normality revealed six items that departed from univariate normality

(standardised kurtosis index: β2 ranging from -1.10 to 59.80) with Mardia‟s (1970)

normalised estimate of multivariate kurtosis revealing severe positive kurtosis and

multivariate nonnormality (critical ratio = 151.20, p < .001). A review of Mahalanobis

distance scores revealed nine cases as distinctively different from all other cases. These cases

were discarded for the following analyses. With severe nonnormality remaining (critical ratio

= 124.55, p < .001), a Bollen-Stine bootstrap (Bollen & Stine, 1992), based on Nevitt and

Hancock‟s (2001) recommendation of 2000 bootstrap samples, was performed and its

associated test of overall model fit was examined in lieu of maximum likelihood methods that

assume multivariate normality.

Initial model fit was deemed unsatisfactory since the Bollen-Stine p-value for the

four-factor model was significant (p < .001), with traditional indices of model fit in

agreement, χ2 (183) = 864.56, p < .001; CFI = .87; TLI = .85; RMSEA = .07 (90% CI = .06-

.07); SRMR = .06. Exploration of standardised residuals and modification indices revealed

possible avenues to improve model fit, the most prominent of which was a non-specified

loading from the latent extreme behaviours factor to the item “Over-taken another vehicle

when not supposed to or with little visibility (e.g., on a double line, before a hill or crest)”.

Since the latent factor of positioning behaviours already loaded onto the item, for the sake of

model parsimony and purity of variance and covariance estimates, the removal of the item

was favoured over the addition of the cross-loading. This modification resulted in better

model fit, χ2 (164) = 679.89, p < .001; CFI = .90; TLI = .88; RMSEA = .06 (90% CI = .05-

.06); SRMR = .06, with the Bollen-Stine p-value remaining significant (p < .001).

Model fit was further improved by the addition of covariances between three pairs of

error terms, all of which pertained to items within a common latent factor. The first was

between the error variances of two items that had overlapping content pertaining to elements

of the unknown (“Driven on unsafe roads (e.g. when flooded; when cluttered with debris)”

and “Raced or chased another vehicle driven by someone you do not know”). The second was

between the error variances of two items that related to an internal subjective state (“Driven

whilst extremely tired or exhausted” and “Driven when extremely emotionally aroused (e.g.,

Angry)”). The third was between error variances of items that shared the content area of

speeding (“Driven at least 15km/hr above the speed limit” and “Exceeded a decreased speed

limit (e.g. 40km/hr) in a road work zone by at least 15km/hr”). Addition of each of these

covariance pathways resulted in improved model fit, with the final fit indices revealing

satisfactory model fit, χ2 (161) = 523.55, p < .001; CFI = .93; TLI = .92; RMSEA = .05 (90%

CI = .05-.06); SRMR = .05, despite the Bollen-Stine p-value remaining significant (p < .001).

Further modifications were deemed inappropriate in order to safeguard the distinctiveness of

the sub-scales and the parsimony of the overall model.

Three of the four latent variables were found to be reliable as reflected by their

multiple observed variables: Hancock and Mueller‟s (2001) coefficient H = .84 (distracted),

H = .84 (positioning), H = .47 (substance use), and H = .70 (extreme). Notably, the

substance-use factor reliability was lower than in Study 1. All factor loadings were

significantly different from zero (p < .001) and model fit was deemed satisfactory when

assessed separately for distracted, χ2 (8) = 28.82, p < .001; CFI = .99; TLI = .97; RMSEA =

.06 (90% CI = .04-.08); SRMR = .02, positioning, χ2 (19) = 56.75, p < .001; CFI = .98; TLI =

.97; RMSEA = .05 (90% CI = .04-.06); SRMR = .03, and extreme behaviours, χ2 (1) = 0.03,

p = .87; CFI = 1.00; TLI = 1.00; RMSEA = .01 (90% CI = .01-.05); SRMR = .01. Model fit

could not be determined for substance-use behaviours due to the model being unidentified.

From these indicators, most factors were deemed to display convergent validity.

Based on the moderate to strong covariances between the latent factors, in particular

the factors of distracted and positioning behaviours, four competing models were assessed

against the four-factor model. In comparison to both a one-factor solution (with covariances

between all factors constrained to one) and a three-factor solution (with the covariance

between distracted and positioning behaviours constrained to one), the four-factor solution

was found to have significantly better fit, Δχ2

= 900.25, Δdf = 6, p < .001 and Δχ2

= 116.16,

Δdf = 1, p < .001, respectively. Additionally, in comparison to both a three-factor solution

(with covariances between substance-use and extreme behaviours constrained to one) and a

five-factor solution (with the items “Driven when extremely emotionally aroused (e.g.,

angry)” and “Driven whilst extremely tired or exhausted” loading onto a fifth latent factor),

the four-factor solution was found to have significantly better model fit, Δχ2

= 375.24, Δdf =

1, p < .001 and Δχ2

= 58.79, Δdf = 3, p < .001, respectively. From this, the factors within the

four-factor solution were deemed to show discriminant validity.

Factorial invariance between males (n = 260) and females (n = 521) was assessed.

The configural model, with no constraints present, revealed satisfactory, gender-invariant fit

for the four-factor model, χ2 (322) = 713.00, p < .001; CFI = .92; TLI = .91; RMSEA = .04

(90% CI = .03-.04); SRMR = .06, with the Bollen-Stine p-value significant (p < .001). Next

the measurement model, where the female and male measurement weights (factor loadings)

were constrained to equality, was assessed for invariance. No significant change was found

between the configural and measurement models, Δχ2

= 26.11, Δdf = 16, p > .05, ΔCFI < .01,

suggesting measurement invariance in the model. Factorial invariance was also assessed

between provisional (n = 469) and open licence drivers (n = 321). The configural model, with

no constraints present, revealed satisfactory, licence type-invariant fit for the four-factor

model, χ2 (322) = 787.11, p < .001; CFI = .91; TLI = .90; RMSEA = .04 (90% CI = .04-.05);

SRMR = .05, with the Bollen-Stine p-value significant (p < .001). Next the measurement

model, where the provisional and open licence sample measurement weights (factor loadings)

were constrained to equality, was assessed for invariance. No significant change was found

between the configural and measurement models, Δχ2

= 23.29, Δdf = 16, p > .10, ΔCFI < .01,

suggesting measurement invariance across licence types.

Descriptive statistics and correlations can be seen in Table 3. In regards to

nomological validity, all driving factors were positively correlated with one another, as well

as being correlated in the expected directions with other risk related, psychological variables.

Thus, as expected, higher self-reported frequencies of all reckless driving behaviour subtypes

were associated with greater willingness to take driving risks, higher reckless driving self-

efficacy, higher sensation seeking tendencies, and lower perceptions of risk. Of interest was

the unexpectedly low correlation between “extreme” driving behaviours and willingness to

take risks. Importantly, the pattern of correlations between the driving behaviours and the

psychological variables varied between factors: for example, compared to the other two

driving behaviours, distracted and positioning behaviours were much more highly correlated

with risk willingness and self- efficacy, but they were not more highly correlated than were

the other behaviours with either sensation-seeking or risk perception. This provides further

validation for the differentiation between driving subtypes.

[Table 3 around here.]

Mann-Whitney U tests were chosen over independent group t-tests to assess gender

and licence type differences for each driving subtype due to significant skewness in each of

the outcome variables. Results revealed significant gender differences in positioning,

substance-use, and extreme driving behaviours, Mann-Whitney U = 56203.50, 61136, and

51756.50, all ps < .001, respectively. Inspection of the group means indicated that males (M

= 3.18, SD = 1.98; M = 0.50, SD = 1.15; M = 0.76, SD = 1.04) reported more frequent

positioning, substance-use, and extreme driving than did females (M = 2.60, SD = 1.79; M =

0.27, SD = 0.78; M = 0.39, SD = 0.72), respectively. No gender differences were observed for

distracted driving behaviours, Mann-Whitney U = 64620, p = .30. Within the current

restricted age range (17 to 25 years), there was a modest tendency for reporting of reckless

driving behaviours to increase with age (rs ranged from .07 to .14, all ps < .05), with the

exception of extreme behaviour (r = -.07, p > .05). Comparison between holders of

provisional licences and open licences revealed significant differences in positioning,

distracted, and substance-use driving behaviours, Mann-Whitney U = 63290.50, 62918,

65941.50, all ps = .001, respectively. No licence type differences were observed for extreme

driving behaviours, Mann-Whitney U = 69855, p = .20. Inspection of the group means

indicated that open licence holders (M = 3.08, SD = 1.95; M = 3.96, SD = 2.31; M = 0.30, SD

= 0.92) reported more frequent positioning, distracted, and substance-use driving than did

provisional licence holders (M = 2.60, SD = 1.79; M = 3.39, SD = 2.14; M = 0.41, SD = 0.93),

respectively.

3.2.2. Multiple regression analysis

To further validate the differentiation between driving subtypes, a standard multiple

regression analysis was conducted on each of the four reckless driving subtypes. Six variables

(gender [coded as females = 0 and males = 1], age, sensation-seeking, self-efficacy, risk

perception, and risk willingness) were entered at once to evaluate their unique contribution to

explaining the variance in each of the reckless driving behaviours. Licence type (provisional

or open) was not included in these analyses because of the strong relationship it has with age

(rpb = .73, p < .001) due to the graduated driver licensing systems present in Australia. In

response to a large amount of skew in the dependent variables of substance-use and extreme

reckless driving due to floor effects in the responses, these variables were transformed using

logarithmic procedures leading to significantly reduced skew. Coefficients from the four

standard multiple regressions can be seen in Table 4.

[Table 4 around here.]

Overall, 45% of the variance in distracted driving behaviours was explained, with all

six variables revealed as significant predictors. Coefficients for gender and age revealed a

tendency for females and older young drivers to report higher engagement in distracted

driving behaviours. Similarly, increased self-efficacy beliefs, sensation-seeking tendencies,

and willingness to drive recklessly were related to increased reported engagement in

distracted driving. Against expectation and contrary to the direction of the zero-order

correlation, perception of risk was also positively related to reports of engaging in distracted

driving behaviours. To explore the source of this incongruent regression coefficient, the

analysis was repeated using hierarchical procedures with gender, age and risk perception

entered in the first step of the model and sensation-seeking, self-efficacy, and risk willingness

entered separately in three successive steps. Results revealed that the coefficient for risk

perception was negative in the first two steps, turning positive with the inclusion of self-

efficacy and strengthened with the inclusion of risk willingness. This suggests that some

aspect of risk perception that does not overlap with self-efficacy, as well as only marginally

overlapping with risk willingness, has a positive relationship with distracted driving.

Overall, 17% of the variance in substance-use driving behaviours was explained, with

three variables revealed as significant predictors. Coefficients for self-efficacy beliefs,

sensation-seeking tendencies, and willingness to drive recklessly show a tendency for

increases in each to be related to increases in reported engagement in substance-use driving.

Gender, age, and risk perception were not significant predictors of substance-use driving

when accounting for the variance explained by the other variables.

Overall, 22% of the variance in extreme driving behaviours was explained, with five

variables significant. Coefficients for gender and age revealed a tendency for males and

younger drivers to report higher engagement in extreme driving behaviours. Similarly,

increased self-efficacy beliefs, sensation-seeking tendencies, and willingness to drive

recklessly were related to increased reported engagement in extreme driving. Risk perception

was not a significant predictor of extreme driving when accounting for the variance explained

by the other variables.

Overall, 48% of the variance in positioning driving behaviours was explained, with

two variables significant. Increases in self-efficacy beliefs and willingness to drive recklessly

were related to increases in reported engagement in positioning driving behaviour. Gender,

age, risk perception, and sensation-seeking were not significant unique predictors.

4. General Discussion

4.1. The Reckless Driver Behaviour Scale (RDBS)

Results from Study 2 confirmed those obtained in Study 1, with the four Study 1

factors of distracted, substance-use, extreme, and positioning reckless driving behaviours

replicated in a separate, larger sample of young drivers in Study 2. The 20-item instrument

developed in this research has been named the Reckless Driver Behaviour Scale (RDBS). Its

structure and content are in contrast to a majority of scales used in previous research.

Moreover, this new instrument for measuring self-reported reckless driving behaviours

addresses three of the major concerns inherent in many existing instruments.

Previous research has examined a diverse range of driving behaviours but has

typically not focused on reckless driving behaviours. Many studies that have examined

reckless behaviour have done so in the context of more general measurement, rather than

choosing or designing measures to capture just this subset of all driving behaviours. In

developing the RDBS, we aimed to improve the measurement of reckless driving behaviours

in terms of breadth, clarity, and relevance. Distracted behaviours, including distractions from

technology and fatigue, were of particular interest due to the limited attention given to these

behaviours in existing instruments. In fact, self-reports from both of the current studies reveal

these particular behaviours to have the highest reported engagement in this age group.

Despite a large range in the length of existing driver behaviour scales (see Table 1),

the number of items included in the RDBS approximates the average number of items used

previously. However, by following Arnett‟s (1992) definition, the RDBS provides a more

content valid measurement of self-reported reckless driving behaviour. To avoid confounding

reckless driving behaviours with that of related variables, the RDBS was created with the aim

of measuring reckless behaviours only. In this way, measurement of the antecedents,

correlates, and consequences of these behaviours is left to other instruments. The items

comprising the RDBS include minimal references to phenomena that are non-behavioural,

and the scale includes no items assessing behaviours that are not reckless.

The number of factors in the RDBS also closely approximates the average number in

existing instruments (Table 1). In fact whilst using slightly different methods (and criteria for

item inclusion), Schmidt (2012) also found a four-factor structure, with the current factors of

substance-use, distracted, and positioning behaviours mirroring those of Schmidt (termed

substance-use, distracted driving, and moving violation, respectively). Unfortunately

Schmidt‟s fourth factor, aggressive driving, in addition to including items not considered

reckless, follows previous research by describing the behaviours in terms of motivational

factors. Construction of the RDBS aimed to avoid this form of problematic categorisation by

the use of exploratory procedures in which different reckless driving behaviours were placed

in homogenous, conceptually distinct categories on the basis of characteristics of the

behaviours themselves, without mention of intentions or motivations. This procedure aimed

to prevent very different behaviours being categorised under the same generic label and

treated as similar in subsequent analyses, as well as avoiding the use of conceptually-

overlapping categories which could have resulted in a single behaviour being classified in

several different ways.

Multiple reliability estimates for each of the factors indicated that all but the

substance-use factor had adequate reliability. The factors were also assessed for both

convergent and discriminant validity. Of interest was the strong correlation found between

the factors of distracted and positioning behaviours. Due to these behaviours belonging to the

domain of reckless driving behaviours, it was indeed expected that each of the factors would

have a high level of covariation with the others. Model comparisons, however, found that the

relationship between these two factors was significantly different from unity. Also, the items

“Driven whilst extremely tired or exhausted” and “Driven when extremely emotionally

aroused (e.g., angry)” could be considered to represent a form of distraction that is different

from the other distracted driving items, however model fit was better when these items

loaded onto the distracted driving factor than when loading onto a separate factor. This makes

theoretical sense because both fatigue and extreme emotions can distract attention and deplete

perceptual and cognitive resources (Cai et al., 2007; Phillip et al., 2005).

The reliability of the item composite measuring the substance-use factor was

noticeably lower in Study 2 than was found in Study 1. This could be partly explained by the

more heterogeneous sample in Study 2. Although the two items comprising this factor do

correlate more with each other than with the other behaviours, very rarely does 2-item

measurement result in a stable and internally consistent factor (Green et al., 1977). Future

research may consider including additional items that represent the substance-use domain.

Schmidt (2012), for example, included items that measured passenger behaviour in the

domain of substance-use, however this procedure would not be appropriate if seeking to

measure driving behaviour. Quite possibly, the less reliable two-item subscale may be

preferred over a less valid, but more reliable, longer subscale.

4.2. Correlates of the different reckless driving behaviour sub-types

Differentiation between driving subtypes was identified with males reporting more

frequent positioning, substance-use, and extreme driving than did females, open licence

holders reporting more frequent positioning, distracted, and substance-use driving than did

provisional licence holders, and a tendency for reporting of reckless driving behaviours to

increase with age, with the exception of extreme behaviour. These findings, that gender-, age-

, and licence type-differences are not uniform across the reckless driving categories, support

the premise underlying the current research that reckless driving needs to be treated not as a

single homogenous group, but as a set of more narrowly-defined sub-types.

Multiple regression analyses confirmed this differentiation between the four

behaviour factors in terms of how they relate to relevant demographic and psychological

variables. The role of gender and age as predictors of each of the four subtypes of reckless

driving, whilst accounting for other variables, was particularly interesting. Whilst neither was

a significant unique predictor of substance-use nor positioning driving behaviours, females

and older drivers were more likely to report engagement in distracted driving, with males and

younger drivers more likely to report engagement in extreme driving behaviours. Also, the

predictive utility of willingness to take driving risks was shown to be much lower for

substance-use and extreme driving than for distracted and positioning driving. These findings

support previous research that highlights the varying correlates of specific reckless driving

behaviours (McNally & Titchener, 2012; Paleti et al., 2010; Thomas et al., 2007; Williams,

2003).

Another interesting finding was in relation to risk perception. While negatively

correlated with all four behaviour types, risk perception was positively related to distracted

driving in the regression analysis. This suggests that, when the influence of other predictors is

factored in, young drivers who perceive greater risk in performing reckless driving

behaviours are more likely to report engaging in distracted behaviours. Whilst counter-

intuitive, this finding has been found previously (Hallett et al., 2011; Walsh et al., 2008;

White et al., 2004). For example, Walsh and colleagues (2008) found that drivers who intend

to text message while driving are more aware of the inherent risks than those who do not. The

current multiple regression analyses suggest that a sub-component of risk perception that is

uncorrelated with both self-efficacy and risk willingness is positively correlated with

distracted driving. Although still open to interpretation and further investigation, this finding

challenges the commonly held belief that risk perception has consistent and entirely negative

effects on reckless driving (Jonah, 1986; Hatfield & Fernandes, 2009). As shown here, this is

not the case when separate, homogenous types of reckless driving behaviour are

independently assessed.

4.3. Implications, evaluation and recommendations for future research

These findings have implications for both intervention strategies and future research.

The finding of differentiation in the predictors of specific reckless driving categories suggests

that general-purpose interventions may have limited effectiveness. Rather, different types of

programs might be required for different forms of driver recklessness with interventions in

one reckless driving domain not necessarily helpful in others. This supports the notion that

prevention programs designed to modify driver behaviour, although rarely implemented, are

often too broad, failing to acknowledge the different antecedents of specific unsafe

behaviours (Schwebel et al., 2006; Sheehan et al., 2004). For example, our Study 2 analyses

suggest that interventions targeting willingness to take risks may be more appropriate for

young drivers who display distracted or positioning recklessness, than for those who engage

in other types of reckless driving behaviours. Knowledge of the reckless driving behaviour

subtypes found in the current research could inform both policy-makers and researchers as to

the types of behaviours that are conceptually and empirically proximate, and hence the

combinations of behaviours that are most and least likely to be tractable to a single

intervention.

Future research could aim at confirming the factor structure found in the current

research in samples drawn from other nations, as well as separately exploring the antecedents

of each reckless driving behaviour group. Following this, item and scale properties could be

further explored using Rasch analysis. This research would, in turn, improve understanding

of why young people engage in reckless driving behaviours, and would inform policy-makers

as to the most effective and efficient ways to construct interventions. Other research could

examine the predictive validity of self-reported driving recklessness, as measured by the

RDBS: do scores on each of the behaviour subscales taken at an earlier age predict crash and

offence involvement or other outcomes over the following months or years? Longitudinal

research would also shed light on the temporal relations between reckless behaviours, on the

one hand, and psychological variables like driving self-efficacy and reckless driving

willingness, on the other.

The current research should also be understood within a few limitations. First, it must

be acknowledged that self-reports of driving behaviours are subject to recall errors and

response biases and that the current research is no exception to this. Specifically, a six-month

recall period was used in the current research to ensure adequate variability and minimal floor

effects in the reckless driving data, but this period may have resulted in substantial recall

errors. To balance the advantages of both longer (e.g., six-month) and shorter (e.g., 4 weeks)

timeframes, future research should seek to verify findings across multiple recall periods.

Additionally, the current research relies heavily on self-report data, which may inflate

associations due to common-method variance. Like the DBQ and similar driving scales,

further validation of the RDBS requires evidence of associations with objective outcome

variables (af Wåhlberg et al., 2011).

Also, the samples used in the two studies were predominantly female. With much

evidence that males are more likely to engage in many types of reckless driving (Bina et al.,

2006; Catchpole & Styles, 2005; Fergusson et al., 2003; McEvoy et al., 2006; Oltedal &

Rundmo, 2006), different results could emerge from a more gender-balanced sample. The

predominant sampling of psychology students may have biased the results in unknown ways.

The RDBS was developed in a particular cultural and driving context and needs to be

validated elsewhere, with more heterogeneous samples. The meaning and relevance of the

scale items are also likely to change over time, suggesting the need for continual updating of

the scale. Minor wording changes may also improve the psychometric quality of the scale.

For example, the item “Driven on unsafe roads (e.g. when flooded; when cluttered with

debris)” could be more congruent with the extreme driving factor if amended to “Driven on

roads that you know are unsafe (e.g. when flooded; when cluttered with debris)”.

Additionally, a large amount of skew in the dependent variables of substance-use and

extreme reckless driving was found due to response floor effects. Even though

transformations were performed, and even though the current means and variances are similar

to those obtained in studies that have used the DBQ (e.g., Lawton et al., 1997; Xie & Parker,

2002), this could have had an influence on the results.

4.4. Conclusion

While the use of a common measurement instrument enables valid comparison

between studies, the need for uniformity must be weighed against the dangers associated with

continued use of potentially invalid, out-dated, or in other ways inappropriate tools. In the

case of reckless driving in young drivers, many existing instruments are lacking in focus,

breadth, and precision and thus may not accurately measure the full diversity of relevant

behaviours. The present study has introduced a new scale that enables the measurement of

four distinctive yet correlated types of reckless driving behaviour. Researchers can

investigate possible variations in the antecedents of these reckless driving types, and policy-

makers can use this evidence in the design and evaluation of interventions. By continuing to

re-conceptualise the reckless driving behaviour of young drivers, the relevance of research

and the impact of interventions can be further improved.

Acknowledgements

The researchers would like to thank the Queensland Department of Transport and

Main Roads for their assistance in the collection of the data used in this research. We also

thank the anonymous reviewers for their insightful feedback on earlier drafts of this paper.

References

af Wåhlberg, A. E., Dorn, L., & Kline T. (2011). The Manchester Driver Behaviour

Questionnaire as a predictor of road traffic accidents. Theoretical Issues in

Ergonomics, 12, 66-86.

Agerwala, S. M., Votta, A., Hogan, B., Yannocone, J., Samuels, S., & Chiffriller, S. (2008).

Aggressive driving in young motorists. International Journal of Humanities and

Social Sciences, 2, 182-185.

Anderson, P., & Baumberg, B. (2006). Alcohol in Europe. London: Institute of Alcohol

Studies.

Armstrong, K. A., Wills, A. R., & Watson, B. C. (2005). Psychosocial influences on drug

driving in young Australian drivers. In K. A. Armstrong, A. R. Wills, & B. C.

Watson, (Eds.) Australian road safety research policing education conference,

November, Wellington, New Zealand.

Arnett, J. (1992). Reckless behavior in adolescence: A developmental perspective.

Developmental review, 12, 339-373.

Arnett, J., Offer, D., & Fine, M. A. (1997). Reckless driving in adolescence: „State‟ and

„trait‟ factors. Accident Analysis and prevention, 29, 57-63.

Arnett, J. (1996). Sensation seeking, aggressiveness, and adolescent reckless behavior.

Personality and Individual Differences, 20, 693-702.

Arnett, J. (1994). Sensation seeking: A new conceptualization and a new scale. Personality

and Individual Differences, 16, 289-296.

Berghaus, G., Scheer, N., & Schmidt, P. (1995). Effects of cannabis on psychomotor skills

and driving performance: A meta-analysis of experimental studies. Paper presented at

the International Conference on Alcohol, Drugs and Traffic Safety, Adelaide.

Beyth-Marom, R., Austin, L., Fischoff, B., Palmgren, C., & Quadrel, M. J. (1993). Perceived

consequences of risky behaviors: Adults and adolescents. Developmental Psychology,

29, 549-563.

Bina, M., Graziano, F., & Bonino, S. (2006). Risky driving and lifestyles in adolescence.

Accident Analysis and Prevention, 38, 472-481.

Blows, S., Ameratunga, S., Ivers, R. Q., Lo, S. K., & Norton, R. (2005). Risky driving habits

and motor vehicle driver injury. Accident Analysis and Prevention, 37, 619-624.

Bollen, K. A., & Stein, R. A. (1992). Bootstrapping goodness-of-fit measures in structural

equation models. Sociological Methods Research, 21, 205-229.

Bradley, G., & Wildman, K. (2002). Psychosocial predictors of emerging adults‟ risk and

reckless behaviors. Journal of Youth and Adolescence, 31, 253-267.

Burns, P. C., & Wilde, G. J. S. (1995). Risk taking in male taxi drivers: Relationships among

personality, observational data and driver records. Personality and Individual

Differences, 18, 267-278.

Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differences in risk taking: A

meta-analysis. Psychological Bulletin, 125, 367-383.

Cai, H., Lin, Y., & Mourant, R. R. (2007, September). Study on driver emotion in driver-

vehicle-environment systems using multiple networked driving simulators. Paper

presented at the driving simulation conference, Iowa City, IA.

Calafat, A., Blay, N., Juan, M., Adrover, D., Bellis, M. A., Hughes, K., …Bohrn, K. (2009)

Traffic risk behaviors at nightlife: Drinking, taking drugs, driving, and use of public

transport by young people. Traffic Injury Prevention, 10, 162-169.

Catchpole, J., & Styles, T. (2005). Young drivers: Who takes risks and why? Vermont South,

VIC: ARRB Group Ltd.

Charlton, S. G. (2009). Driving while conversing: Cell phones that distract and passengers

who react. Accident Analysis and Prevention, 41, 160-173.

Chisholm, S. L., Caird, J. K., & Lockart, J. (2008). The effects of practice with MP3 players

on driving performance. Accident Analysis and Prevention, 40, 704-713.

Clarke, D. D., Ward, P., & Truman, W. (2005). Voluntary risk taking and skill deficits in

young driver accidents in the UK. Accident Analysis and Prevention, 37, 523-529.

Clement, R., & Jonah, B. A. (1984). Field dependence, sensation seeking and driving

behavior. Personality and Individual Differences, 5, 87-93.

Dahlen, E. R., Martin, R. C., Ragan, K., & Kuhlman, M. M. (2005). Driving anger, sensation

seeking, impulsiveness, and boredom proneness in the prediction of unsafe driving.

Accident Analysis and Prevention, 37, 341-348.

Dalziel, J., & Job, S. (1997). Motor vehicle accidents, fatigue and optimism bias in taxi

drivers. Accident Analysis and Prevention, 29, 489-494.

Deck, C., Lee, J., & Reyes, J. (2008). Risk attitudes in large stake gambles: Evidence from a

game show. Applied Economics, 40, 41-52.

Deery, H. A. (1999). Hazard and risk perception among young novice drivers. Journal of

Safety Research, 30, 225-236.

Deery, H. A., & Fildes, B. N. (1999). Young novice driver subtypes: Relationship to high-

risk behavior, traffic accident record, and simulator driving performance. Human

Factors, 41, 628-643.

Deery, H.A., & Love, A.W. (1996). The effect of a moderate dose of alcohol on the traffic

hazard perception profile of young drivers. Addiction, 91, 815-827.

Deffenbacher, J. L., Lynch, R. S., Oetting, E. U., & Swaim, R. C. (2002). The driving anger

expression inventory: A measure of how people express their anger on the road.

Behaviour Research and Therapy, 40, 717-737.

DePasquale, J. P., Geller, E. S., Clarke, S. W., & Littleton, L. C. (2001). Measuring road

rage: Development of the Propensity for Angry Driving Scale. Journal of Safety

Research, 32, 1-16.

De Winter, J. C. F., & Dodou, D. (2010). The Driver Behaviour Questionnaire as a predictor

of accidents: A meta-analysis. Journal of Safety Research, 41, 463-470.

Dula, C. S., & Ballard, M. E. (2003). Development and evaluation of a measure of dangerous,

aggressive, negative emotional, and risky driving. Journal of Applied Social

Psychology, 33, 263-282.

Dula, C. S., & Geller, E. S. (2003). Risky, aggressive, or emotional driving: Addressing the

need for consistent communication in research. Journal of Safety Research, 34, 559-

566.

Elliot, M. A., Baughan, C. J., & Sexton, B. F. (2007). Errors and violations in relation to

motorcyclists‟ crash risk. Accident Analysis and Prevention, 39, 491-499.

Fergusson, D. M., Swain-Campbell, N., & Horwood, J. (2003). Risky driving behaviour in

young people: Prevalence, personal characteristics and traffic accidents. Australian

and New Zealand Journal of Public Health, 27, 337-342.

Fernandes, R., Soames-Job, R. F., & Hatfield, J. (2007). A challenge to the assumed

generalizability of prediction and countermeasure for risky driving: Different factors

predict different risky driving behaviors. Journal of Safety Research, 38, 59-70.

French, D., West, R., Elander, J., & Wilding, J. (1993). Decision making style, driving style

and self-reported involvement in road traffic accidents. Ergonomics, 36, 627-644.

Furnham, A., & Saipe, J. (1993). Personality correlates of convicted drivers. Personality and

Individual Differences, 14, 326-329.

Garson, G. D. (2012). Cluster analysis. Asheboro, NC: Statistical Associates Publishers.

George, S., Clark, M., & Crotty, M. (2007). Development of the Adelaide driving self-

efficacy scale. Clinical Rehabilitation, 21, 56-61.

Gibbons, F. X., & Gerrard, M. (1995). Predicting young adults‟ health risk behaviour.

Journal of Personality and Social Psychology, 69, 505-517.

Gibbons, F. X., Gerrard, M., & Lane, D. J. (2003). A social reaction model of adolescent

health risk. In J. Suls, & K. A. Wallston (Eds.), Social psychological foundations of

health and illness (pp. 107-136). Malden, MA: Blackwell Publishers.

Gibbons, F. X., Gerrard, M., Ouellette, J. A., & Burzette, R. (1998). Cognitive antecedents to

adolescent health risk: Discriminating between behavioral intention and behavioral

willingness. Psychology and Health, 13, 319-339.

Gibbons, F. X., Gerrard, M., Vande Lune, L. S., Wills, T. A., Brody, G., & Conger, R. D.

(2004). Context and cognition: Environmental risk, social influence, and adolescent

substance use. Personality and Social Psychology Bulletin, 30, 1048-1061.

Glendon, A. I., McNally, B., Jarvis, A., Chalmers, S., & Salisbury, R. (in press). Evaluating a

novice driver and pre-driver road safety intervention. Accident Analysis and

Prevention. http://dx.doi.org/10.1016/j.aap.2013.11.017

Golias, I., & Karlaftis, M. G. (2002). An international comparative study of self-reported

driver behaviour. Transportation Research Part F, 4, 243-256.

Green, S. B., Lissitz, R. W., & Mulaik, S. A. (1977). Limitations of coefficient alpha an index

of test unidimensionality. Educational and Psychological Measurement, 37, 827-838.

Greene, K., Krcmar, M., Walters, L. H., Rubin, D. L., Hale, J., & Hale, L. (2000). Targeting

adolescent risk-taking behaviors: the contribution of egocentrism and sensation-

seeking. Journal of Adolescence, 23, 439-461.

Groeger, J. A. (2006). Youthfulness, inexperience, and sleep loss: The problems young

drivers face and they pose for us. Injury prevention, 12, i19-i24.

Gulian, E., Matthews, G., Glendon, A. I., Davies, D. R., & Debney, L. M. (1989).

Dimensions of driver stress. Ergonomics, 32, 585-602.

Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate data analysis

(7th ed.). Upper Saddle River, NJ: Pearson Prentice-Hall.

Hallett, C., Lambert, A., & Regan, M. A. (2011). Cell phone conversing while driving in

New Zealand: Prevalence, risk perception and legislation. Accident Analysis and

Prevention, 43, 862-869.

Hancock, G. R., & Mueller, R. O. (2001). Rethinking construct reliability within latent

variable systems. In R. Cudeck, S. du Toit, & D. Sörbom (Eds.), Structural Equation

Modeling: Present and Future – A Festschrift in Honor of Karl Jöreskog (pp.195-

216). Lincolnwood, IL: Scientific Software International Inc.

Harre, N., Brandt, T., & Dawe, M. (2000). The development of risky driving in adolescence.

Journal of Safety Research, 31, 185-194.

Harris, P. B., & Houston, J. M. (2010). Recklessness in context: Individual and situational

correlates to aggressive driving. Environment and Behaviour, 42, 44-60.

Hartos, J., Eitel, P., & Simons-Morton, B. (2002). Parenting practices and adolescent risky

driving: A three-month prospective study. Health Education Behaviour, 29, 194-206.

Hatfield, J., & Fernandes, R. (2009). The role of risk-propensity in the risky driving of young

drivers. Accident Analysis and Prevention, 41, 25-35.

Henley, N. (2004). Social marketing: „Selling‟ injury prevention. In R. McClure, M.

Stevenson, & S. McEvoy (Eds.), The scientific basis of injury prevention and control

(pp. 318-333). Melbourne: IP Communications.

Hoyle, R. H., Stephenson, M. T., Palmgreen, P., Lorch, E. P., & Donohew, R. L. (2002).

Reliability and validity of a brief measure of sensation seeking. Personality and

Individual Differences, 32, 401-414.

Jessor, R., Turbin, M. S., & Costa, F. M. (1997). Predicting developmental change in risky

driving: The transition to young adulthood. Applied Developmental Science, 1, 4-16.

Jonah, B. A. (1986). Accident risk and risk-taking behaviour among young drivers. Accident

Analysis and Prevention, 18, 255-271.

Jonah, B. A. (1997). Sensation seeking and risky driving: A review and synthesis of the

literature. Accident Analysis and Prevention, 29, 651-665.

Kircher, A., Vogel, K., Tornros, J., Bolling, A., Nilsson, L., Patten, C., …Ceci, R. (2004).

Mobile telephone simulator study. Linköping, Sweden: Swedish National Road and

Transport Research Institute.

Klauer, S. G., Dingus, T. A., Neale, V. L., Sudweeks, J. D., & Ramsey, D. J. (2006). The

impact of driver inattention on near-crash/crash risk: An analysis using the 100-car

naturalistic driving study data (DOT HS 810 594). Washington, DC: Department of

Transportation.

Kontogiannis, T., Kossiavelou, Z., & Marmaras, N. (2002). Self-reports of aberrant behaviour

on the roads: Errors and violations in a sample of Greek drivers. Accident Analysis

and Prevention, 34, 381-399.

Krahe, B., & Fenske, I. (2002). Predicting aggressive driving behaviour: The role of macho

personality, age, and power of car. Aggressive Behaviour, 28, 21-29.

Lawton, R., Parker, D., Manstead, A. S. R., & Stradling, S. G. (1997). The role of affect in

predicting social behaviors: The case of road traffic violations. Journal of Applied

Social Psychology, 27, 1258-1276.

Lenne, M. G., Dietz, G., Rumbold, G., Redman, J. R., & Triggs, T. J. (2000). Opioid

dependence and driving ability: A view in the context of proposed legislative change

in Victoria. Drug and Alcohol Review, 19, 427-439.

Lonczak, H. S., Neighbors, C., & Donovan, D. M. (2007). Predicting risky and angry driving

as a function of gender. Accident Analysis and Prevention, 39, 536-545.

Lund, I. O., & Rundmo, T. (2009). Cross-cultural comparisons of traffic safety, risk

perception, attitudes and behaviour. Safety Science, 47, 547-553.

Mardia, K. V. (1970). Measures of multivariate skewness and kurtosis with applications.

Biometrika, 57, 519-530.

Malta, L. S. (2004). (Doctoral dissertation). Predictors of aggressive driving in young adults.

Retrieved from ProQuest Dissertations and Theses database. (UMI No. 765833501)

Maslowsky, J., Keating, D. P., Monk, C. S., & Schulenberg, J. (2010). Planned versus

unplanned risks: Neurocognitive predictors of subtypes of adolescents‟ risk

behaviour. International Journal of Behavioural Development, 35, 152-160.

McEvoy, S. P., Stevenson, M. R., & Woodward, M. (2006). The impact of driver distraction

on road safety: Results from a representative survey in two Australian states. Injury

Prevention, 12, 242-247.

McNally, B. & Titchener, K. (2012). The role of affective processes on young drivers‟ risk

perceptions: A dual process model approach. Journal of Risk Research, 15, 39-51.

Montano, D., Thompson, B., Taylor, V. M., & Mahloch, J. (1997). Understanding

Mammography Intention and Utilization among Women in an Inner City Public

Hospital Clinic. Preventive Medicine, 26, 817-824.

Montag, I. (1989). Styles of thinking and driving. Paper presented at the international

conference of Changes and Innovation in Road Safety, Tel Aviv.

Neighbors, C., Vietor, N. A., & Knee, C. R. (2002). A motivational model of driving anger

and aggression. Personality and Social Psychology Bulletin, 28, 324-335.

Nevitt, J., & Hancock, G. R. (2001). Performance of bootstrapping approaches to model test

statistics and parameter standard error estimation in structural equation modeling.

Structural Equation Modeling, 8, 353-377.

O'Connor, B. P. (2000). SPSS and SAS programs for determining the number of components

using parallel analysis and Velicer's MAP test. Behavior Research Methods,

Instrumentation, and Computers, 32, 396-402.

Oltedal, S., & Rundmo, T. (2006). The effects of personality and gender on risky driving

behaviour and accident involvement. Safety Science, 44, 621-628.

Owens, J. M., McLaughlin, S. B., & Sudweeks, J. (2011). Driver performance while text

messaging using handheld and in-vehicle systems. Accident Analysis and Prevention,

43, 939-947.

Paleti, R., Eluru, N., & Bhat, C. (2010). Examining the influence of aggressive driving

behavior on driver injury severity in traffic crashes. Accident Analysis and Prevention

42, 1839-1854.

Parker, D. (2002). Changing drivers' attitudes to speeding using the Theory of Planned

Behaviour. In D. Rutter & L. Quine (Eds.), Changing Health Behaviour (pp. 138-

152), Buckingham: Open University Press.

Patil, S. M., Shope, J. T., Raghunathan, T. E., & Bingham, C. R. (2006). The role of

personality characteristics in young adult driving. Traffic Injury Prevention, 7, 328-

334.

Phillip, P., Sagaspe, P., Moore, N., Taillard, J., Charles, A., Guilleminault, C., & Bioulac, B.

(2005). Fatigue, sleep restriction and driving performance. Accident Analysis and

Prevention, 37, 473-478.

Potter, J. (2000). Drugs and driving in Australia. Sydney: Austroads.

Price, P. C. (2001). A group size effect on personal risk judgments: Implications for

unrealistic optimism. Memory and Cognition, 29, 578-586.

Raemakers, J. G., Berghaus, G., van Laar, M., & Drummer, O. H. (2004). Dose related risk of

motor vehicle crashes after cannabis use. Drug and Alcohol Dependence, 73, 109-

119.

Reason, J. T., Manstead, A., Stradling, S., Baxter, J. S., & Campbell, K. (1990). Errors and

violations on the roads: A real distinction? Ergonomics, 33, 1315-1332.

Richer, I., & Bergeron, J. (2009). Driving under the influence of cannabis: Links with

dangerous driving, psychological predictors, and accident involvement. Accident

Analysis and Prevention, 41, 299-307.

Rundmo, T. (1995). Perceived risk, safety status, and job stress among injured and non-

injured employees on offshore petroleum installations. Journal of Safety Research,

26, 87-97.

Rundmo, T., & Iversen, H. (2002). Personality, risky driving and accident involvement

among Norwegian drivers. Personality and Individual Differences, 33, 1251-1263.

Sarkar, S., & Andreas, M. (2004). Acceptance of and engagement in risky driving behaviors

by teenagers. Adolescence, 39, 687-700.

Schmidt, S. A. (2012). (Doctoral dissertation). Youth risky driving behaviours: Advancements

in measurement and theory. Retrieved from

https://atrium.lib.uoguelph.ca/xmlui/bitstream/handle/10214/6314/Dissertation.Sarah

%20Schmidt.pdf?sequence=1

Schwebel, D. C., Severson, J., Ball, K. K., & Rizzo, M. (2006). Individual differences factors

in risky driving: The roles of anger/hostility, conscientiousness, and sensation-

seeking. Accident Analysis and Prevention, 38, 801-810.

Scott-Parker, B., Watson, B. C., & King, M. J. The risky behaviour of young drivers:

Developing a measurement tool. 20th

Canadian Multidisciplinary Road Safety

Conference, Niagara Falls, Ontario, June, 2010.

Sexton, B. F., Tunbridge, R. J., Brooke-Carter, N., Jackson, P. G., Wright, K., Stark, M. M.,

& Englehart, K. (2000). The influence of cannabis on driving (TRL Report 477).

Crowthorne, UK: TRL Limited.

Sheehan, M., Siskind, V., & Schonfeld, C. (2004). A longitudinal study of adolescent drink

driving and other risk taking behaviors: Challenges for the change process. Paper

presented at the17th International Conference on Alcohol Drugs and Traffic Safety,

Glasgow, Scotland.

Shinar, D., & Compton, R. (2004). Aggressive driving: An observational study of driver,

vehicle, and situational variables. Accident Analysis and Prevention, 36, 429-437.

Simons-Morton, B., Lerner, N., & Singer, J. (2005). The observed effects of teenage

passengers on the risky driving behavior of teenage drivers. Accident Analysis and

Prevention, 37, 973-982.

Sjoberg, L., Moen, B. E., & Rundmo, T. (2004). Explaining risk perception. An evaluation of

the psychometric paradigm in risk perception research. Trondheim, Norway: Rotunde

Publications. Retrieved at

www.svt.ntnu.no/psy/Torbjorn.Rundmo/Psychometric_paradigm.pdf

Smart, D., Vassallo, S., Sanson, A., Cockfield, S., Harris, A., Harrison, W., & MacIntyre, A.

(2005). In the driver‟s seat: Understanding young adults‟ driving behaviour.

Canberra, ACT: Australian Institute of Family Studies.

Smith, S., Carrington, M., & Trinder, J. (2005). Subjective and predicted sleepiness while

driving in young adults. Accident Analysis and Prevention, 37, 1066-1073.

Strayer, D. L., & Johnston, W. A. (2001). Driven to distraction: Dual-task studies of

simulated driving and conversing on a cellular telephone. Psychological Science, 12,

462-466.

Strecher, V. J., & Rosenstock, I. M. (Eds). (1997). The health belief model. San Francisco,

CA: Jossey-Bass.

Taubman- Ben-Ari, O., Mikulincer, M., & Gillath, O. (2004). The multidimensional driving

style inventory: Scale construct and validation. Accident Analysis and Prevention, 36,

323-332.

Taubman- Ben-Ari, O., Mikulincer, M., & Gillath, O. (2005). From parents to children:

Similarity in parents and offspring driving styles. Transportation Research Part F:

Traffic Psychology and Behaviour, 8, 19-29.

Teese, R., & Bradley, G. L. (2008). Predicting recklessness in emerging adults: A test of a

psychosocial model. Journal of Social Psychology, 148, 105-126.

Thomas, J., Kavanagh, J., Tucker, H., Burchett, H., Tripney, J., & Oakley, A. (2007).

Accidental injury, risk-taking behaviour and the social circumstances in which young

people (aged 12-24) live: A systematic review. London: EPPI-Centre, Social Science

Research Unit, Institute of Education, University of London.

Turner, C., & McClure, R. (2003). Age and gender differences in risk-taking behaviour as an

explanation for high incidence of motor vehicle crashes as a driver in young males.

International Journal of Injury Control and Safety Promotion, 10, 123-130.

Ulleberg, P., & Rundmo, T. (2002). Risk-taking attitudes among young drivers: The

psychometric qualities and dimensionality of an instrument to measure young drivers'

risk-taking attitudes. Scandinavian Journal of Psychology, 43, 227-237.

van Gelder, J., de Vries, R. E., & van der Pligt, J. (2009). Evaluating a dual-process model of

risk: Affect and cognition as determinants of risky choice. Journal of Behavioral

Decision Making, 22, 45-61.

Victoir, A., Eertmans, A., van den Bergh, O., & van den Broucke, S. (2005). Learning to

drive safely: Social-cognitive responses are predictive of performance rated by novice

drivers and their instructors. Transportation Research Part F: Traffic Psychology and

Behaviour, 8, 59-74.

Walsh, S. P., White, K. M., Hyde, M. K., & Watson, B. C. (2008). Dialling and driving:

Factors influencing intentions to use a mobile phone while driving. Accident Analysis

and Prevention, 40, 1893-1900.

Weinstein, N. D., Rothman, A. J., & Nicolich, M. (1998). Use of correlational data to

examine the effects of risk perceptions on precautionary behavior. Psychology and

Health, 13, 479-501.

White, M. P., Eiser, J. R., & Harris, P. R. (2004). Risk perceptions of mobile phone use while

driving. Risk Analysis, 24, 323-334.

Wickens, C. M., Mann, R. E., Stoduto, G., & Ialomiteanu, A. (2011). Age group differences

in self-reported aggressive driving perpetration and vicitmisation. Transportation

Research Part F: Traffic Psychology and Behaviour, 14, 400-412.

Wiesenthal, D. L., Hennessy, D. A., & Gibson, P. (2000). The Driving Vengeance

Questionnaire (DVQ): The development of a scale to measure deviant drivers'

attitudes. Violence and Victims, 15, 115-136.

Williams, A. F. (2003). Teenage drivers: Patterns of risk. Journal of Safety Research, 34, 5-

15.

Williamson, A. (2003). Why are young drivers over represented in crashes? Summary of

issues. Update of literature reviews: Literature from 2000 to 2003. Sydney, NSW:

NSW Motor Accidents Authority.

Xie, C., & Parker, D. (2002). A social psychological approach to driving violations in two

Chinese cities. Transportation Research Part F, 5, 293-308.

Zuckerman, M. (1994). Behavioral Expressions and Biosocial Bases of Sensation Seeking.

Cambridge, UK: Cambridge University Press.

Zuckerman, M., Eysenck, S., & Eysenck, H. J. (1978). Sensation seeking in England and

America: Cross-cultural age and sex comparisons. Journal of Consulting and Clinical

Psychology, 46, 139-149.

Table 1.

Source, Structure, and Content of Selected Measures of Reckless Driving Behaviours.

Source Name/Basis of

Driving Scale

No. of

Items

No. of

Factors

Scale Item Contenta Details of the “other” driving

behaviours assessed

1 2 3 4 5 6

Reason et al., 1990 Driver Behaviour

Questionnaire (DBQ)

50 3 x x x x Racing/chasing

French et al., 1993 Driving Style

Questionnaire

15 6 x x

Dalziel & Job, 1997

Risk-taking measure

10 1 x x

Lawton et al., 1997 Selected DBQb +

additional

12 3 x x x x Racing/chasing

Jessor et al., 1997 within Young Adult

Driving Questionnaire

28 3 x x x

Wiesenthal et al., 2000 Driving Vengeance

Questionnaire

15 1 x x

DePasquale et al., 2001 Propensity for Angry

Driving scale

15 1 x

Golias & Karlaftis, 2001 SARTE (1998) survey

13 2 x x x

Xie & Parker, 2002 Selected DBQb +

additional

27 6 x x x x Racing/chasing

Source Name/Basis of

Driving Scale

No. of

Items

No. of

Factors

1 2 3 4 5 6 Details of the “other” driving

behaviours assessed

Kontogiannis et al., 2002 DBQb + additional 70 7 x x x x Racing/chasing, Driven on

unsafe roads

Ulleberg & Rundmo,

2002

Risk Behaviour Scale 15 3 x x

Deffenbacher et al., 2002 Driving Anger

Expression Inventory

39 5 x x

Dula & Ballard, 2003 Dula Dangerous

Driving Index

28 3 x x x Racing/chasing

Taubman – Ben-Ari et

al., 2004

Multidimensional

Driving Style

Inventory

44 8 x x

Elliott et al., 2007 Motorcycle Rider

Behaviour

Questionnaire

43 5 x x x x x Racing/chasing

Scott-Parker et al., 2010 The Behaviour of

Young Novice Drivers

Scale

44 5 x x x x x

Schmidt, 2012 Youth Domains of

Risky Driving

40 4 x x x x x Racing/chasing

Proposed Measure Reckless Driving

Behaviour Scale

21 4 x x x x x x Racing/chasing, Driven on

unsafe roads a1 = Speeding (10km/hr, 15km/hr, 30 km/hr over speed limit, etc), 2 = Position (unsafe lane changes, tailgating, rapid acceleration, running red

light, abrupt braking, rolling stop signs, etc), 3 = Technology/Distraction (calling/answering a phone, texting, searching for items in the car,

negative emotions, etc.), 4 = Fatigue (Tired driving), 5 = Substance-use (Alcohol-use, Cannabis-use, etc) 6 = Other (details given in this table). bBasis in DBQ items, however all original items not necessarily represented.

Table 2.

Loadings for the Factors Elicited from Study 1 Exploratory Factor Analysis of Self-Reported

Reckless Driving Behaviours (N = 189).

Items Factor

1

Factor

2

Factor

3

Factor

4

Texted or looked for numbers on your phone or searched

for songs on your MP3 player whilst driving.

.81

Ate, drank, or smoked whist driving. .75

Called or answered a hand-held phone whilst driving. .70

Searched for CDs or manually searched for stations on

your car radio whilst driving.

.59

Driven when extremely emotionally aroused (e.g.,

angry).

.59

Driven whilst extremely tired or exhausted. .52

Driven under the influence of cannabis or any other

psychoactive drug (e.g., ecstasy).

.75

Driven when you suspected you were over the .05 blood

alcohol limit.

.64

Raced or chased another vehicle driven by someone you

do not know.

.88

Performed burnouts or “doughnuts”. .85

Raced or chased another vehicle driven by a friend or

someone you know.

.61

Driven on unsafe roads (e.g., when flooded). .45

Driven at least 15km/hr above the speed limit. .64

“Tail-gated” or deliberately followed another vehicle at

an unsafe distance.

.64

Exceeded a decreased speed limit (e.g., 40km/hr) in a

road work zone by at least 15km/hr.

.62

Accelerated from a stationary position at an excessive

rate (e.g., from a red light).

.60

Changed lanes frequently on a multi-lane road. .45

Turned, merged, or changed lanes without indicating. .41

Over-taken another vehicle when unsafe to do so (e.g., on

a double line, before a hill or crest).

.40

M SD Mdn Range

Factor 1 (Distracted) 4.49 2.33 4.33 9.67 (.84) - - -

Factor 2 (Substance use) 0.97 1.26 1.00 6.50 .31* (.70) - -

Factor 3 (Extreme) 0.84 0.97 1.00 6.50 .33* .33* (.82) -

Factor 4 (Positioning) 3.04 1.78 2.71 9.29 .61* .30* .43* (.81)

Note. Factor loadings < .3 not shown, * denotes p < .001, Reliability estimates in italics.

Table 3

Descriptive Statistics, including Reliability Estimates and Correlations, for each Variable Included in Study 2 (N = 694).

Variable M SD Mdn Range 1 2 3 4 5 6 7 8

1. Distracted Behaviour 3.78 2.43 4.59 10.00 (.81) - - - - - - -

2. Substance-Use Behaviour 0.37 0.95 1.00 7.42 .29 (.45) - - - - - -

3. Extreme Behaviour 0.49 0.92 1.14 7.02 .30 .42 (.67) - - - - -

4. Positioning Behaviour 2.85 1.84 3.62 9.91 .65 .31 .44 (.83) - - - -

5. Risk Willingness 59.87 17.51 59.00 87.00 .63 .32 .28 .60 (.92) - - -

6. Self-Efficacy 17.32 7.26 17.00 36.00 .47 .30 .34 .55 .57 (.81) - -

7. Sensation-Seeking 2.81 0.81 2.88 4.00 .27 .26 .23 .27 .34 .36 (.79) -

8. Risk Perception 12.02 3.90 11.66 24.00 -.12 -.17 -.13 -.22 -.28 -.38 -.27 (.72)

Note. Reliability estimates in italics, p < .001 for all correlations.

Table 4.

Summary of Results from each Multiple Regression Analysis

Behaviour R2

Variable B SE B β sr2

Distracted .45 Gender -.74 .16 -.14*** .017

Age .06 .03 .06* .004

Self-Efficacy .08 .01 .24*** .032

Risk Perception .06 .02 .10** .008

Sensation-Seeking .20 .09 .07* .004

Willingness .07 .01 .50*** .160

Substance-

use

.17 Gender .01 .02 .03 .001

Age .01 .01 .06 .004

Self-Efficacy .01 .01 .12* .008

Risk Perception -.01 .01 -.02 .001

Sensation-seeking .04 .01 .17*** .026

Willingness .01 .01 .20*** .026

Extreme .22 Gender .05 .01 .12** .012

Age -.01 .01 -.16*** .023

Self-Efficacy .01 .01 .25*** .032

Risk Perception .01 .01 .05 .002

Sensation-seeking .03 .01 .11** .010

Willingness .01 .01 .16*** .017

Positioning .48 Gender -.01 .12 -.01 .000

Age -.03 .02 -.03 .001

Self-Efficacy .06 .01 .26*** .036

Risk Perception .01 .01 .03 .001

* p < .05. ** p < .01. *** p < .001.

Sensation-seeking .02 .07 .01 .000

Willingness .06 .01 .53*** .176