Accident Analysis and Prevention 109 (2017) 47–54

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Contributors to young drivers’ driving styles – A comparison between Israel MARK

and Queensland
⁎
Vera Skvirskya, Orit Taubman Ben-Aria, , Timothy J. Greenburyb, Carlo G. Pratob
a
The Louis and Gabi Weisfeld School of Social Work, Bar Ilan University, Ramat Gan, Israel
b
School of Civil Engineering, The University of Queensland, St. Lucia, 4072, Brisbane, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: Among the numerous factors that contribute to young novice drivers’ driving styles, personality characteristics,
Driving styles sociodemographic variables, family atmosphere, and friends’ norms are known to have an important impact.
Young drivers However, cross-cultural comparisons are relatively rare in the safety literature concerning young drivers. This
Separation-individuation study aimed at comparing young drivers from Israel and Queensland (Australia) and examining the contribution
Family climate for road safety
of personality, sociodemographic, family and friends’ aspects to their driving styles (reckless and careless; hostile
Safe driving climate among friends
and angry; anxious; patient and careful). More specifically, this study examined the associations between young
drivers’ driving style and their perceptions of separation-individuation, the family climate for road safety, and
the safe driving climate among friends. We also examined sociodemographic and driving history variables such
as gender, the marital status of parents, and personal exposure to traffic crashes. The study consisted of two
samples of male and female young drivers (age 17–22) from Israel (n = 160) and Queensland (n = 160), who
completed a set of valid and reliable self-report questionnaires. Findings indicate that in general, maladaptive
driving styles are associated with lower family tendency to engage in promoting road safety, higher pressure and
costs of driving with peers, and unhealthier separation-individuation aspects. The opposite is observed for the
patient and careful driving style that relates to higher engagement of the family in road safety, lower pressure
from friends, and healthier separation-individuation. Some differences were found regarding specific styles
between the two samples. In addition, women scored lower than men in the reckless and careless style, and
higher (in the Israeli sample) in the anxious as well as the patient and careful styles. Overall, similarities in the
associations between the study variables in the samples exceed the differences, and the importance of examining
variables on multi-levels when referring to young drivers’ driving styles, is confirmed. The findings attest to the
universal utility of the MDSI, together with the understanding that only a wider examination of personal and
environmental contributors enables true insights into the complex behavior of driving among young drivers.

1. Introduction
A recent review of the utility and uses of the Multidimensional
Driving Style Inventory (MDSI; Taubman-Ben-Ari et al., 2004) revealed
that it has been the subject of studies conducted around the world,
especially in Israel, where the instrument was originally designed
(Taubman-Ben-Ari and Skvirsky, 2016). The MDSI has been translated
into numerous languages (e.g., English, Italian, Russian, Arabic), and
used in various countries (e.g., England, Italy, Romania, the United
States, Argentina); however, direct cross-cultural comparisons had not
been undertaken thus far.
Previous cross-cultural studies tended to use the Driver Behaviour
Questionnaire (DBQ; Reason et al., 1990), showing differences in re-
ported driving behaviours of drivers in culturally different countries
(Bener et al., 2013; Özkan et al., 2006; Warner et al., 2011). In addition,
studies have indicated demographic as well as cultural factors related to
driver behavior or to involvement in traffic crashes (e.g., Nordfjarn
et al., 2014; Özkan et al., 2006). The current study aimed at comparing
young drivers from Israel and Queensland (Australia). Moreover, the
study looked into the role played by various contributors to the young
drivers’ driving styles: socio-economic and driving history variables,
their level of separation-individuation, their perceptions of the family
climate for road safety (Taubman-Ben-Ari and Katz-Ben-Ami, 2012,
2013), and their understanding of the safe driving climate among
friends (Guggenheim and Taubman-Ben-Ari, 2016).
The MDSI assesses four broad driving styles: (a) reckless and care-
less; (b) anxious; (c) angry and hostile; and (d) patient and careful. The
reckless and careless driving style refers to deliberate violations of safe

⁎
Corresponding author at: The Louis and Gabi Weisfeld School of Social Work, Bar Ilan University, Ramat Gan, 52900, Israel.
E-mail address: taubman@biu.ac.il (O.T. Ben-Ari).

http://dx.doi.org/10.1016/j.aap.2017.08.031
Received 12 April 2017; Received in revised form 25 August 2017; Accepted 29 August 2017
Available online 11 October 2017
0001-4575/ © 2017 Elsevier Ltd. All rights reserved.
V. Skvirsky et al.
driving norms and the seeking of sensations and thrills while driving. It
is characterized by driving at high speeds, passing in no-passing zones,
and driving while intoxicated. The anxious driving style reflects feelings
of alertness and tension, as well as ineffective engagement in relaxing
activities during driving. The angry and hostile driving style refers to
expressions of irritation, rage, and hostility while driving, along with a
tendency to act aggressively on the road, including cursing other dri-
vers, honking the horn, or flashing headlights. The patient and careful
driving style reflects well-adjusted driving behaviors, such as planning
ahead, paying full attention to the road, displaying patience, courtesy,
and calm behind the wheel, and obeying the traffic rules. These various
driving styles were found to be systematically related to socio-
demographic variables, driving history elements, personality char-
acteristics, as well as features of the environment, such as family and
peers’ safety-related attitudes and behavior (Taubman-Ben-Ari and
Skvirsky, 2016).
Separation-individuation is a concept which represents the extent
to which a teenager relinquishes the power of internalized parents in
order to attain greater intrapsychic autonomy (Blos, 1979), and is
considered an important sign of adolescent development. Importantly,
the disengagement from the internalized parents, should not come at
the expense of severing external family ties; both individuality and
connectedness in parent–adolescent relationships are important for
optimal adolescent development (Grotevant and Cooper, 1998). The
separation-individuation process enables one to be psychologically in-
dependent yet at the same time more capable of establishing deep in-
timate relationships with others outside of the family unit, such as
friends (Årseth et al., 2009).
More specifically, there appears to be two major components of the
individuation process: A risk component, reflected by separation or
emotional detachment from the family, and a protective component,
reflected by the development of autonomy while maintaining intimacy
and connection to the family (Baer and Bray, 1999). As such, the se-
paration-individuation concept entails alongside with healthy separa-
tion, also more problematical aspects, such as strong dependency needs,
separation anxiety, need denial, high degree of narcissism and self-
centeredness, and more (Levine, Green, and Millon 1986). Individua-
tion is considered a developmental process that occurs with all ado-
lescents regardless of ethnic background (Baer and Bray, 1999; Bray
et al., 2000).
It has been suggested that risky behaviors such as substance abuse,
are partially a result of problems with the individuation process, and
that emotional isolation and a lack of differentiation from parents in-
terfere with the adolescent's separation-individuation process.
Evaluations of this model with cross-sectional samples of a broad age
range of adolescents provide support for the role of individuation in
alcohol use (Baer and Bray, 1999; Bray et al., 2000; Bray et al., 2001),
so that higher healthy separation was related to lower alcohol con-
sumption. Although its importance has been proven in risky alcohol
use, to the best of our knowledge, this personal feature has never before
been tested in the context of driving styles, though its established as-
sociation with familial and social aspects may hint to its importance in
this context as well.
However, the personal level is not the only one that reflects on
young drivers’ driving style. One of the most important contributors,
investigated in recent years is the model set by the family (e.g., Hartos
et al., 2000; Shope and Bingham, 2008; Gil et al., 2016). To assess the
contribution of the family in the current study, we employed the Fa-
mily Climate for Road Safety (FCRS; Taubman-Ben-Ari and Katz-Ben-
Ami, 2012, 2013), a multi-dimensional construct referring to the va-
lues, perceptions, priorities, and practices of the family, and in parti-
cular – the parents, in regard to safe driving. The seven dimensions
comprising the FCRS refer to different angles of this concept, with
Modeling reflecting the model given by parents’ driving, parents’ safety
attitudes, as well as their tendency to obey traffic regulations; Feedback
referring to the degree to which parents tend to provide their offspring
48
Accident Analysis and Prevention 109 (2017) 47–54
with feedback and encouraging comments for safe and considerate
driving; Communication indicating direct and open communication
between parents and adolescents on driving behavior in general, and
especially on risky driving; Monitoring denoting the extent parents
supervise their offspring, and how much they insist on being informed
of where youngsters are taking the car, who is going with them, and
when they intend to be home; Commitment to safety relating to the
comprehensive commitment parents hold towards road safety; Mes-
sages referring to clarity of verbal safety messages from parents to
young drivers; and Limits indicating the magnitude of parental sys-
tematic and clear-cut limits setting on adolescents' driving behavior and
the degree to which they discipline them for traffic violations
(Taubman-Ben-Ari and Katz-Ben-Ami, 2013).
Studies indicate that positive dimensions of the family climate are
related to safe driving among young drivers. More specifically, when
parents are better role models, provide encouraging and empowering
feedback for safe driving, enable more open communication, convey
clearer messages regarding safe driving, monitor their driving to higher
extent, and tend to set clear limits on violating traffic laws, young
drivers tend to endorse a more patient and careful driving style and less
reckless and careless and angry and hostile styles. On the other hand,
young drivers who perceive their parents to be uncommitted to safety
report more endorsement of reckless and careless and angry and hostile
styles. Regarding the anxious driving style the results of previous stu-
dies are less unequivocal, showing that some positive aspects of FCRS
may relate to greater utilization of this driving style (Taubman-Ben-Ari,
2016; Taubman-Ben-Ari and Katz-Ben-Ami, 2012, 2013).
Notwithstanding the importance of the family, a third significant
aspect which should be considered when discussing young drivers’
driving styles is the central influence of their peers (e.g., Curry et al.,
2012; Horvath et al. 2012 ; Winston and Jacobsohn, 2010). An attempt
to comprehensively understand this aspect was undertaken recently by
Guggenheim and Taubman-Ben-Ari (2016) with the construction of a
new scale – the Safe Driving Climate among Friends (SDCaF), tapping
four dimensions of driving with friends: friends' pressure; social costs of
driving with friends; communication between friends about driving;
and shared commitment to safe driving. These dimensions have been
validated by significant associations, such as with self-disclosure (Miller
et al., 1983), resistance to peer influence (Steinberg and Monahan,
2007), the global tendency for peer pressure (Santor et al., 2000), and
personal commitment to safe driving (Taubman-Ben-Ari and Katz-Ben-
Ami, 2012).
When attempting to compare two countries with different licensing
systems, it is important to review similarities and differences in the laws
of both countries: 1) In Israel, teenagers can begin driving lessons at the
age of 16.5. In Queensland, teenagers can start their learning period at
16 years old, and in both countries this is depending on passing the
learner theory test; 2) In Israel, lessons are given by professional in-
structors on specially equipped vehicles; learners are permitted to drive
only during their lessons. In Queensland, there is no obligation of taking
lessons from professional instructors, but there is an obligation to log at
least 100 hours on a logbook (including 10 hours at night); 3) A driving
license is issued in Israel upon successfully passing an on-road driving
test. The road test cannot be taken until the learner has turned 16 and 9
months and has completed a minimum of 28 on-road driving lessons. In
Queensland, after a minimum of 1 year of accompanied driving, lear-
ners take a practical driving test that enables them to have a provisional
license where restrictions are maintained on mobile phone use and
high-powered vehicles. After another minimum 1 year, holder of pro-
visional license take a hazard recognition test that allows them to earn
an open license; 4) Israeli new young drivers (under the age of 24), have
to be accompanied by an experienced driver, someone over the age of
24 who has held a valid driving license for a minimum of five years, for
the first three months during day and night trips, and during the next
three months, only in night hours. New drivers under the age of 21 are
restricted from carrying more than two passengers for a period of two
V. Skvirsky et al.
years after licensure, unless accompanied by an experienced driver. In
Queensland, the accompanying person needs to have an open license
for the same vehicle for at least 1 year, and every hour supervised by a
professional instructor counts for 3 logged hours (with a maximum of
30 logged hours). Restrictions are posed to the use of mobile phones,
including hands-free, on both learners and passengers; 5) In Israel,
there are aggravations on traffic violations for novice young drivers. A
young novice driver who will be caught on violations such as driving
after alcohol consumption or without an accompanied driver, will lose
his or her driving license and will have to pass an on-road driving test
again. Notably, Queensland law imposes zero tolerance on alcohol for
all drivers until 25 years old (CARRS-Q, 2015; Israel Government
Portal- gov., 2017; Ministry of Transport State of Israel, 2016).
In view of the literature review, we hypothesized that higher reports
on the maladaptive driving styles (reckless and careless; angry and
hostile; anxious), and lower endorsement of the careful and patient
driving style, will be associated with: a) lower healthy separation in-
dividuation, and higher levels of problematic separation-individuation,
b) higher noncommitment of parents to safety and lower scores on the
positive aspects of the family climate for road safety, and c) lower re-
ports on the communication and shared commitment dimensions of
driving with friends, as well as higher reports of friends’ pressure and
social costs. Based on previous studies (Bener et al., 2013; Özkan et al.,
2006; Warner et al., 2011), we also examined exploratively the cross-
cultural differences in the MDSI factors, as well as in the associations
between the independent variables and the driving styles.
2. Methods
2.1. Participants and procedure
This study proposed a multidimensional analysis of the relations
between the four MDSI dimensions (Taubman-Ben-Ari et al., 2004) and
the effects of the relationships between young drivers and parents as
well as peers for two samples of young drivers between 17 and 22 years
old, who participated voluntarily in surveys in Israel and Queensland.
The two groups were administered the survey in their respective mother
tongue through an electronic data collection program, via social media,
and convenience sampling. They were asked questions concerning their
socio-demographic characteristics as well as about their driving styles,
separation-individuation dimensions, and family and peer climate of
road safety.
The two groups were similar in terms of size (Israeli sample
size = 160, Queensland sample size = 164) and gender distribution
(both the Israeli and the Queensland sample have 50.0% for male and
female young drivers). The two groups were different in terms of age
distribution (Israeli mean = 21.6 and st.dev. = 1.9 years old,
Queensland mean = 19.5 and st.dev. = 1.7), household size (Israeli
mean = 4.8 and st.dev. = 1.7, Queensland mean = 3.3 and
st.dev. = 1.4), parents’ marital status (15.6% of Israeli parents are se-
parated or divorced versus 34.8% of Queensland parents), and vehicle
ownership (40.6% for Israeli young drivers versus 62.2% for
Queensland). When looking at having experienced road crashes with at
least one light injury, similarities were reported in regard to whether
the young drivers had acquaintances who suffered injuries (56.3% for
Israelis and 56.7% for Queensland), although differences were men-
tioned in regard to whether the young drivers themselves have suffered
injuries (16.9% for Israelis and 22.6% for Queensland).
2.2. Instruments
The MDSI (Taubman-Ben-Ari et al., 2004) assesses the driving styles
of the young drivers through their answers to 44 statements about their
feelings and behaviors while driving. Responses are marked on a 6-
point scale ranging from “not at all” to “very much”. The inventory
consists of four driving styles: (i) reckless and careless driving that
49
Accident Analysis and Prevention 109 (2017) 47–54
implies violating deliberately road safety norms; (ii) anxious driving
that reflects a feeling of tension and distress behind the wheel; (iii)
angry and hostile driving that denotes irritability and aggression on the
road; (iv) patient and careful driving that indicates a well-adjusted
driving style. The Cronbach’s Alpha reliability coefficients were rea-
sonable for all four dimensions across both the Israeli and the
Queensland group, respectively for reckless and careless driving (0.806
and 0.865), anxious driving (0.829 and 0.861), angry and hostile
driving (0.790 and 0.762), and patient and careful driving (0.775 and
0.727). Accordingly, scores for the four dimensions of the MDSI were
computed for each young driver as the average over the items asso-
ciated with each dimension.
The Separation-Individuation Test of Adolescence (SITA, Levine et al.,
1986) was used to measure different dimensions of the second se-
paration-individuation process. This self-report instrument consists
originally on 103 items, which are assessed on a five-point scale, ran-
ging from “strongly agree” or “is always true of me” to “strongly dis-
agree” or “is never true of me.” For the current study, we used 43 items
from four dimensions with respective Cronbach’s Alpha measures of
reliability reasonable across the dimensions for both the Israeli and the
Queensland group: (i) practicing mirroring that relates to narcissistic
strivings (0.904 and 0.871); (ii) separation anxiety that links to sig-
nificant others experienced as abandoning (0.824 and 0.837); (iii)
nurturance seeking that links to strong caretaker attachment (0.716 and
0.717); (iv) healthy separation that reflects a balance of dependence
and independence strivings (0.824 and 0.791). Accordingly, the four
SITA dimensions were computed for each young driver as the average
scores over the items linked to each dimension.
The Family Climate for Road Safety Scale (FCRSS, Taubman-Ben-Ari
and Katz-Ben-Ami, 2012, 2013) scale assesses the degree to which the
young drivers consider 54 items to apply to their family on a 5-point
scale ranging from “not at all” to “very much”. The items allow eliciting
the perceptions of young drivers about their family attitudes and be-
haviour with respect to road safety. Seven dimensions of family climate
were defined and the respective Cronbach’s Alpha measures of relia-
bility were reasonable for the Israeli and the Queensland groups re-
spectively: modelling (0.844 and 0.842), feedback (0.941 and 0.873),
communication (0.895 and 0.869), monitoring (0.858 and 0.784), non-
commitment (0.637 and 0.789), messages (0.787 and 0.877), and limits
(0.803 and 0.795). Accordingly, scores for each FCRS dimension were
computed for each young driver as the average over the items be-
longing to each dimension.
The Safe Driving Climate among Friends (SDCaF) scale (Guggenheim
and Taubman-Ben-Ari, 2016) evaluates the effect of peers on the safety
of young drivers on four dimensions consisting on 19 items with a 5-
point scale ranging from “completely disagree” to “completely agree”.
The respective Cronbach’s Alpha measures of reliability were reason-
able across the dimensions for both the Israeli and the Queensland
group: (i) pressure from friends while driving (0.736 and 0.814), (ii)
social costs of driving with friends (0.868 and 0.818), (iii) commu-
nication about driving between friends (0.749 and 0.786), and (iv)
shared commitment to safe driving with friends (0.815 and 0.781).
Accordingly, scores for the four dimensions of the SDCaF scale were
calculated for each young driver as the average over the items asso-
ciated with each dimension.
2.3. Model
This study measured the association of the four MDSI dimensions
with the perceptions of the young driver about him/herself and the
safety climate of parents and peers. Accordingly, it should be noted that
the model selection should satisfy the following known criteria for
model formulation: (i) the model should allow for a multidimensional
analysis of the concurrent factors affecting the MDSI dimensions; (ii)
the model should allow for linear regression of the factors on the cal-
culated values of the MDSI dimensions; (iii) the model should capture
V. Skvirsky et al. Accident Analysis and Prevention 109 (2017) 47–54

the well-known correlations across the MDSI dimensions; (iv) the model
should be robust to the relatively small sample size for both the Israeli
and the Queensland groups.
The first three criteria indicated the need for a model where linear
regressions could be correlated, while the last criteria suggested the
need for Bayesian modelling that relaxes the very restrictive assumption
that frequentist modelling poses when sample sizes are small. The
model that satisfies all these four criteria is the seemingly unrelated
regression (SUR) model (Zellner, 1962). The SUR model enabled us to
express the linear variation of four dependent continuous variables
while accounting for their correlation, in light of previous findings
about the four MDSI dimensions (Taubman-Ben-Ari et al., 2004;
Taubman-Ben-Ari and Skvirsky, 2016). Moreover, the estimation of the
SUR model was performed with a Bayesian approach with a direct
Monte Carlo (DMC) approach (Zellner and Ando, 2010), given that the
procedure has demonstrated its ability and efficiency to retrieve sig-
nificant relations even with small sample sizes of 50 observations.
Lastly, the Bayesian approach removed any assumption on the dis-
tribution of the error terms, a desired feature when having small sample
sizes and non-normal dependent variables.
Given the average scores yj of each MDSI dimensions for each young
male driver n, the linear SUR model of the four scores formulates four
regression equations with correlated error terms having different var-
iances (Zellner and Ando, 2010):
y = Xβ + u with u ∼ N (0, Ω ⊗ I ) (1)
⎧ y1 = X1 β1 + u1
⎪ y2 = X2 β2 + u2
⎨ y3 = X3 β3 + u3
⎪
⎩ y4 = X 4 β4 + u 4
ωij I (i ≠ j )
with E [u1u'j] = ⎧ 2
⎨ ωi I (i = j ) (2)
⎩ 1946) were suitable highly non-informative priors for the parameters to
where yj are (n × 1) vectors of scores for the MDSI dimension j, Xj are
(n × vj) matrices for vj observed variables, βj are (vj × 1) vectors of
parameters to be estimated, and uj are (n × 1) vectors of error terms. It
should be noted that the four equations may have different independent
variables and error term variances, and that the four diagonal elements
as well as the six off-diagonal elements of the covariance matrix Ω are
parameters to be estimated (Zellner and Ando, 2010). It should also be
noted that two different SUR models were estimated separately for the
Israeli and the Queensland groups.
The specification of the model considered as observed variables the
seven dimensions of the FCRS, the four dimensions of the SDCaF, the
four dimensions of the SITA, and the socio-demographic characteristics
of the young drivers. The model estimation consisted of the DMC ap-
proach that allowed obtaining posterior marginal densities from prior
marginal densities for the parameters to be estimated, after re-
formulating the SUR models as follows (Zellner and Ando, 2010):
⎧ y1 = X1 β1 + e1 with u1 = e1
⎪ y2 = X2 β2 + ρ21 u1 + e2
⎨ y3 = X3 β3 + ρ31 u1 + ρ32 u2 + e3
⎪
⎩ y4 = X 4 β4 + ρ41 u1 + ρ42 u2 + ρ43 u3 + e4
0 (i ≠ j )
with E [e1e'j] = ⎧ 2 and Σ = diag [σ12, σ22, σ32, σ42]
⎨
⎩ iI
σ (i = j ) (3)
The DMC is applicable and the sampling in sequence is quite simple
to perform because the parameters of the second equation are now
related to the ones of the first one, the parameters of the third equation
are now related to the ones of the first two, and the parameters of the
fourth equation are now related to the ones of the previous three (for
details about the sampling procedure, see Zellner and Ando, 2010). The
DMC estimation procedure did not require prior knowledge on the
parameter distributions and hence Jeffrey’s invariant priors (Jeffreys,

Table 1
Sample characteristics.

Israel Queensland

Variable mean st.dev. min max mean st.dev. min max

MDSI − reckless and careless 2.446 0.756 1.000 5.090 3.203 0.979 1.000 5.091
MDSI − anxious 2.109 0.599 1.110 3.950 3.139 0.787 1.526 4.789
MDSI − angry and hostile 2.474 1.096 1.000 6.000 3.102 1.058 1.000 5.600
MDSI − patient and careful 4.567 0.752 2.110 5.890 4.039 0.699 2.111 5.889
FCRS − modelling 3.755 0.703 1.730 5.000 3.294 0.592 2.000 5.000
FCRS − feedback 3.200 1.255 1.000 5.000 3.224 0.779 1.400 5.000
FCRS − communication 3.389 0.971 1.000 5.000 3.348 0.656 1.889 4.889
FCRS − monitoring 2.690 1.034 1.000 4.860 3.146 0.779 1.143 5.000
FCRS − non commitment 1.934 0.581 1.000 3.880 2.666 0.624 1.000 3.875
FCRS − messages 4.361 0.637 2.000 5.000 3.541 0.665 2.250 5.000
FCRS − limits 3.038 0.976 1.000 5.000 3.297 0.716 1.333 5.000
SDCaF − friends' pressure 1.429 0.556 1.000 3.200 2.599 0.936 1.000 4.600
SDCaF − social cost of driving 1.721 0.836 1.000 4.600 2.617 0.842 1.000 4.600
SDCaF − communication 3.903 0.789 1.000 5.000 3.427 0.668 1.400 4.800
SDCaF − shared commitment 3.114 0.949 1.000 5.000 3.090 0.817 1.000 5.000
SITA − mirroring 2.993 0.664 1.000 4.670 2.969 0.631 1.000 4.800
SITA − anxiety 2.608 0.659 1.000 4.290 2.819 0.623 1.000 4.143
SITA − nurturance seeking 2.971 0.795 1.000 5.000 2.760 0.688 1.000 4.286
SITA − healthy separation 4.120 0.681 1.000 5.000 3.306 0.706 1.000 4.857
Female 0.500 0.502 0.000 1.000 0.500 0.502 0.000 1.000
Age 21.631 1.889 17.000 24.000 19.549 1.731 17.000 22.000
Household size 4.756 1.711 1.000 11.000 3.341 1.433 1.000 7.000
Household siblings 2.625 1.605 0.000 10.000 1.512 1.313 0.000 6.000
Parents married 0.794 0.406 0.000 1.000 0.530 0.501 0.000 1.000
Parents separated 0.006 0.079 0.000 1.000 0.177 0.383 0.000 1.000
Parents divorced 0.150 0.358 0.000 1.000 0.171 0.377 0.000 1.000
Single parent household 0.050 0.219 0.000 1.000 0.122 0.328 0.000 1.000
Own vehicle 0.406 0.493 0.000 1.000 0.622 0.486 0.000 1.000
Injured in a crash 0.169 0.376 0.000 1.000 0.226 0.419 0.000 1.000
Acquaintance injured in a crash 0.563 0.498 0.000 1.000 0.567 0.497 0.000 1.000

Note: MDSI scale 1–6; FCRS, SITA and SDCaF scales 1–5.

50
V. Skvirsky et al. Accident Analysis and Prevention 109 (2017) 47–54

be estimated. The DMC estimation procedure generated 10,000 pos- Table 2

terior samples for the computation of the means, standard deviations Bayesian estimates of the SUR models.
and confidence intervals at the 90% and 95% level for each element of
Variables Israel Queensland
the vectors βj as well as each element of the covariance matrix Ω. Lastly,
posterior simulations from the posterior marginal densities allowed mean st.dev. sig. mean st.dev. sig.
computing empirically vectors of predicted and residual values. Given
MDSI reckless and careless
the average var(res) of the residuals and the average var(pred) of the
FCRSS − feedback – – −0.133 0.086 *
predictions over the sample for each of the four equations in the two FCRSS − communication −0.147 0.074 **
−0.223 0.114 *

models, an empirical proportion of variance explained for each equa- FCRSS − non commitment 0.291 0.113 **
0.323 0.136 **

tion in each model was then calculated as 1–var(res)/var(pred) (Gelman FCRSS − limits −0.142 0.073 *
– –
and Pardoe, 2006). SDCaF − friends' pressure 0.249 0.104 **
0.498 0.078 **

SDCaF − social cost of 0.133 0.065 **

– –
driving
3. Results SDCaF − communication – – −0.199 0.096 **

SDCaF − shared −0.144 0.060 **

−0.200 0.077 **

3.1. Sample characteristics commitment

SITA − mirroring 0.165 0.083 **
0.331 0.090 **

SITA − nurturance seeking −0.208 0.113 *

−0.166 0.092 *
The two groups consisted of 160 Israeli and 160 Queensland young SITA − healthy separation −0.154 0.092 *
– –
drivers. When looking at their average scores on the self-reported Female −0.062 0.034 *
−0.062 0.037 *

measures and recalling that higher average scores on each dimension Parents separated or – – 0.195 0.126 *

reflect the young drivers’ perception and behaviour on that specific divorced
Own vehicle 0.245 0.101 **
– –
dimension, differences emerge between the two groups that are worth *
Constant 1.084 0.599 0.503 0.397
observing (Table 1).
MDSI anxious
The dependent variables, namely the four MDSI dimensions, are
FCRSS − communication −0.133 0.062 **
−0.222 0.097 **

significantly skewed for both groups with a fatter tail on the right hand FCRSS − non commitment 0.166 0.085 *
– –
side of the reckless and careless, anxious and angry and hostile style FCRSS − limits −0.148 0.055 **
−0.179 0.086 **

distributions, and a fatter tail on the left hand side of the patient and SDCaF − friends' pressure 0.234 0.078 **
0.303 0.060 **

careful style distribution. This demonstrates that the SUR model is SDCaF − social cost of 0.120 0.049 **
0.108 0.064 *

driving
appropriate for the task at hand of estimating the variation of mildly- SDCaF − communication – – −0.214 0.074 **

tailed distributed variables, given that this model does not assume the SITA − mirroring 0.133 0.062 **
0.162 0.069 **

error terms to be normally distributed and its performance is good also SITA − anxiety 0.326 0.063 **
0.145 0.088 *

with heavy-tailed error distributions (Ng, 2002). When comparing the SITA − nurturance seeking −0.088 0.051 *
−0.181 0.079 **

SITA − healthy separation −0.131 0.069 *

−0.105 0.067 *
scores on the MDSI dimensions, Queensland young drivers appear to
Female 0.204 0.075 **
– –
score higher in the reckless and careless, anxious and angry and hostile Parents separated or 0.192 0.102 *
– –
dimensions, while Israeli young drivers seem to score higher in the divorced
patient and careful dimension. Own vehicle – – 0.392 0.157 **

* **
The scores on the FCRS dimensions show comparable perceptions of Constant 1.072 0.560 0.793 0.304

young drivers belonging to the two groups, with difference emerging in MDSI angry and hostile
that Queensland young drivers feel that their parents are more active in FCRSS − feedback −0.146 0.088 *
−0.237 0.115 **

FCRSS − communication −0.152 0.091 *

−0.210 0.101 **
monitoring and setting limits to their driving, while Israeli young dri-
FCRSS − non commitment 0.257 0.125 **
0.195 0.109 *

vers feel that their parents are better role models and are more com- SDCaF − friends' pressure – – 0.307 0.093 **

mitted to discussing road safety within the family. Similarly, the scores SDCaF − social cost of 0.198 0.107 **
– –
on the SDCaF dimensions illustrate that Israeli young drivers experience driving
SDCaF − shared −0.228 0.100 **
−0.180 0.093 **
less pressure, associate less cost and communicate more about driving
commitment
with their friends, while the level of commitment is comparable to the SITA − mirroring 0.234 0.138 *
0.359 0.108 **

one experienced by the Queensland young drivers. Lastly, the scores on SITA − nurturance seeking −0.208 0.113 *
– –
the SITA dimensions suggest that Israeli young drivers live a relation SITA − healthy separation −0.332 0.152 **
−0.177 0.104 *

with their parents and peers characterised by lower anxiety, higher Parents separated or – – 0.245 0.134 *

divorced
level of nurturing, and healthier style of separation to reach in- **
Constant 1.273 0.992 2.056 0.476
dependence.
MDSI patient and careful
FCRSS − modelling 0.166 0.096 *
0.210 0.103 **
3.2. Model estimates FCRSS − feedback 0.090 0.051 *
– –
FCRSS − communication 0.230 0.079 **
0.209 0.098 **

The DMC estimation of the linear SUR model aimed at finding the FCRSS − non commitment −0.252 0.109 **
– –
best specification of the four equations representing the relation be- FCRSS − limits – – 0.212 0.087 **

SDCaF − friends' pressure −0.221 0.100 **

−0.118 0.060 **
tween the four MDSI dimensions and the scores of the young drivers on
SDCaF − social cost of −0.135 0.062 **
– –
the aforementioned instruments as well as the socio-demographic driving
characteristics of the young drivers. The best specification of the SDCaF − communication 0.136 0.074 *
– –
equations was found after testing for significance of the independent SDCaF − shared – – 0.121 0.060 **

commitment
variables in the four equations and comparing the deviance information
SITA − mirroring −0.143 0.080 *
−0.155 0.069 **
criterion (DIC) of the models (Spiegelhalter et al., 2002) for both the SITA − nurturance seeking 0.140 0.065 **
– –
Israeli and the Queensland samples. Namely, the independent variables SITA − healthy separation – – 0.122 0.067 *

** **
that are not presented in each equation were not significantly corre- Female 0.114 0.032 0.067 0.028
lated with the variation of the scores on the MDSI dimensions. Own vehicle −0.217 0.097 **
−0.328 0.090 **

** **
Constant 3.529 0.719 1.077 0.307
The best model specifications are presented in Table 2 for both
samples. The posterior distributions of the estimated parameters are covariance matrix of error terms
(continued on next page)
presented in terms of posterior means, standard deviations and

51
V. Skvirsky et al. Accident Analysis and Prevention 109 (2017) 47–54

Table 2 (continued) When looking at the specific styles, similarities emerge across the
two samples: (i) the reckless and careless style is associated with lack of
Variables Israel Queensland
communication and non-commitment to road safety in the family,
mean st.dev. sig. mean st.dev. sig. higher peer pressure and lower peer commitment to safe driving, as
well as higher narcissistic needs accompanied by higher nurturance
ω11 1.413 0.299 **
1.421 0.294 **
requirements; (ii) the anxious style is linked to lower tendency of the
ω22 1.203 0.286 **
1.243 0.271 **
family to communicate and provide boundaries to the young drivers,
ω33 1.243 0.285 **
1.230 0.274 **

ω44 1.531 0.299 **

1.491 0.300 ** higher peer pressure and perceived social costs of driving with friends,
ω12 0.205 0.210 0.673 0.208 ** higher anxiety in the relation with the parents, and lower search for
ω13 0.601 0.210 **
0.807 0.185 **
nurturing and healthy separation from the parents and friends; (iii) the
ω14 −0.668 0.214 **
−0.673 0.209 **
angry and hostile style is related to a lower level of communication and
ω23 0.356 0.219 *
0.555 0.186 **

ω24 −0.626 0.214 **

−0.655 0.185 **
ω34 −0.448 0.239 **
−0.451 0.182 **
* statistically significant difference from zero (90% credible set shows the same sign).
** statistically significant difference from zero (95% credible set shows the same sign).
significance levels for the constant α, the parameters β, and the terms
ωij of the covariance matrix. Moreover, the post-estimation analysis
estimated the proportion of variance explained for the four equations
across the two samples: for the Israeli sample, the proportion of var-
iance explained was estimated as equal to 58.8% for the reckless and
careless style, 66.0% for the anxious style, 54.1% for the angry and
hostile style, and 32.9% for the patient and careful style; for the
Queensland sample, the proportion of variance explains was estimated
as equal to 72.8% for the reckless and careless style, 75.9% for the
anxious style, 67.6% for the angry and hostile style, and 71.0% for the
patient and careful style.
It should be noted that the estimation of the covariance matrix al-
lows the computation of the correlation matrix across MDSI dimensions
for both samples. Table 3 presents the correlation across the dimen-
sions. It shows how unobserved correlation of the error terms of the
four equations is more pronounced for the Queensland sample. More
importantly, it shows that the correlation patterns are logical
(Taubman-Ben-Ari et al., 2004; Taubman-Ben-Ari and Skvirsky, 2016)
and significant, and hence support the formulation of the SUR model
with the four equations estimated simultaneously. It should be noted
that having taken care of the unobserved correlation between dimen-
sions makes the parameter estimates unbiased.
When looking at the estimates of the two SUR models, it seems that
similarities may be observed in the relations across the two samples, but
also differences may be highlighted in terms of what dimensions from
the instruments and which socio-demographic characteristics are sig-
nificantly related to the MDSI dimensions.
A general pattern is that the driving styles that relate to risky be-
haviour such as the reckless and careless, the anxious, and the angry
and hostile (Taubman-Ben-Ari et al., 2016), are associated with lower
family tendency to engage in promoting road safety, higher pressure
and cost of driving with peers, and unhealthier separation aspects. The
opposite is observed for the patient and careful driving style that relates
to higher engagement of the family in road safety, lower pressure from
friends, and healthier separation.
feedback provision by the family, a lack of commitment to safe driving
by both family and peers, higher narcissistic needs and lower healthy
independence from parents and friends; (iv) the patient and careful
style is connected to positive role modelling in the family, good com-
munication with the parents, lower peer pressure, and lower narcissistic
needs. Interestingly, for both samples the age of the survey participants
was not found related to the scores on the MDSI dimensions.
In addition, when looking at the specific styles, some differences
emerge in the instrument dimensions. The reckless and careless style is
associated negatively with the imposition of limits by the family in the
Israeli sample and the provision of feedback by the family in the
Queensland sample. The same style is also related to higher perceptions
of social costs of driving and lower levels of healthy separation in the
Israeli sample, and lower level of communication with friends and
higher nurturance seeking with parents and friends in the Queensland
sample. The lack of communication with friends is also related to the
anxious style and the peer pressure is associated with the angry and
hostile style in the Queensland sample, while perceptions of the social
cost of driving appear linked to the angry and hostile style for the Israeli
sample. The patient and careful driving style is related to higher pro-
vision of feedback, lower perception of social costs, higher commu-
nication with friends about safe driving, and nurturance seeking be-
haviour in the Israeli sample. The same driving style is related to the
definition of limits by the family and the commitment to safe driving
being shared with peers in the Queensland sample.
When looking at the socio-economic characteristics, males score
higher than females in the reckless and careless style while the opposite
applies to the anxious (only for the Israeli sample though) and patient
and careful styles. Owning a vehicle makes young drivers tend towards
more reckless and careless and less patient and careful driving in the
Israeli sample, and more anxious and less patient and careful driving in
the Queensland sample. Having separated or divorced parents is posi-
tively related to higher scores in the reckless and careless and anxious
styles for the Israeli sample, as well as higher scores in the anxious and
angry and hostile styles for the Queensland sample. Age was not sig-
nificant for.
4. Discussion and conclusions
This cross-cultural comparison between Israeli and Queensland
young drivers yielded an interesting result showing that Queensland
young drivers tend to report higher reckless, angry and anxious driving

Table 3
Correlation matrix of the error terms of the four equations in the SUR models.

MDSI Israel Queensland

Reckless and careless Anxious Angry and hostile Patient and careful Reckless and careless Anxious Angry and hostile Patient and careful

Reckless 1.000 1.000

Anxious 0.157 1.000 0.507 1.000
Angry 0.453 0.291 1.000 0.611 0.449 1.000
Careful −0.454 −0.461 −0.325 1.000 −0.463 −0.481 −0.333 1.000

Note: the correlation terms are statistically significantly different from zero at the 95% level.

52
V. Skvirsky et al.
styles, whereas Israeli young drivers seem to report higher patient and
careful driving. In addition, an examination of the FCRS dimensions
shows that Queensland young drivers experience their parents as more
active in monitoring and setting limits to their driving, whereas Israeli
young drivers perceive their parents as better role models and as more
committed to discussing road safety within the family. Similarly, the
scores on the SDCaF dimensions illustrate that Israeli young drivers
experience less pressure, associate less cost and communicate more
about driving with their friends, while the level of commitment is
comparable to the one experienced by the Queensland young drivers.
Lastly, the scores on the SITA dimensions suggest that Israeli young
drivers reached a higher level of separation-individuation. The picture
being drawn here is a coherent one, with those drivers who are more
mature and with tighter positive connections with parents and peers,
also the ones to report safer driving. It is also interesting to note that the
safe driving is not related in this group to stricter discipline such as
limit setting, but to better communication and mutual respect. In con-
trast, the more maladaptive driving styles may reflect the fact that more
Australians in this sample reported having divorced parents and owning
their own car, thus they may be less dependent on their parents in this
regard. The differences indicated here between the groups may well
represent cultural variations, as Israel is a more familistic and col-
lectivistic culture compared to other western countries, with more
close-knit ties, greater intergenerational interdependence and intimacy
among Israeli families than, for example among German ones
(Kagitcibasi et al., 2010)
More importantly, despite differences in magnitude of scores, the
associations between the study variables and the driving styles follow
previous findings (Taubman-Ben-Ari and Skvirsky, 2016), showing that
generally, maladaptive driving styles (the reckless and careless, an-
xious, and angry and hostile) are associated for both samples with lower
parental commitment to road safety and lower communication with
them in issues concerning safe vs. risky driving, resembling earlier
findings with the MDSI Israel (Taubman-Ben-Ari and Katz-Ben-Ami,
2012, 2013) and with a risky driving score in Belgium (Carpentier et al.,
2014). These styles are also connected to higher pressure from friends
and a higher perception of peers as being collectively uncommitted to
safe driving, echoing previous findings (e.g., Curry et al., 2012; Horvath
et al., 2012; Winston and Jacobsohn, 2010). Finally, the elements of
separation-individuation indicate that the maladaptive driving styles
are related to less mature connections between young drivers and their
family and friends, especially highlighting their association with the
narcissistic aspect of this concept, following previous findings regarding
the relation between narcissistic tendencies and risky behaviour
(Buelow and Brunell, 2014; Schreer, 2002). Such findings are also in
line with the claim that risky behaviors are partially a result of pro-
blems with the individuation process, and that emotional isolation and
a lack of differentiation from parents interfere with the adolescent's
separation-individuation process (e.g., Baer and Bray, 1999; Bray et al.,
2000; Bray et al., 2001).
The opposite was found for the patient and careful driving style.
This style was related to higher engagement of the family in promoting
road safety, lower pressure from friends, and healthier separation from
parents. These findings validate prior studies which found a link be-
tween these variables and less risk taking (Baer and Bray, 1999; Bray
et al., 2001; Taubman-Ben-Ari and Katz-Ben-Ami, 2012, 2013).
When looking at background variables, the most evidenced is the
association between gender and driving styles, with women endorsing
more patient and careful and less reckless and careless driving styles
than men. This is in accordance with other studies showing that men
tend to take more risks behind the wheel (Holland et al., 2010; Poó
et al., 2013; Taubman-Ben-Ari and Yehiel, 2012; Taubman-Ben-Ari and
Skvirsky, 2016). Overall, it seems that there are more similarities than
differences in the associations between the independent variables and
the four driving styles for both cultures, and that the “bigger picture”
seems to be quite the same.
53
Accident Analysis and Prevention 109 (2017) 47–54
Some limitations of the current study should be mentioned with the
most pronounced being the fact that the samples were not nationally
representative, and that all measures were based on self-reports.
However, studies have already attested to self-report measures as valid
instruments investigating risk taking behavior. More specifically, pre-
vious findings connected the MDSI scores with performance measures
collected in a driving simulator experiment (driving speed, number of
driving maneuvers, and passing gaps; Farah et al., 2007; Farah et al.,
2009) and using an In-Vehicle Data Recorder (IVDR; Taubman-Ben-Ari
et al., 2016). Nevertheless, the most important asset of the current
study is the ability to validate results using two culturally-different
samples. Future studies would do well to adopt this line of investiga-
tion, with more attention drawn to cross-cultural comparisons, re-
presenting diverse populations, and additional countries.
To conclude, the findings of the current study point to the im-
portance of cross-cultural comparisons as indicators of the universality
of various driving styles, as well as the factors that might contribute to
them. They also attest to the significance of looking at multi-dimen-
sional driving styles and not only on the uni-dimension aspect of risky
driving. The findings demonstrate the universal utility of the MDSI,
together with the understanding that only a wider examination of
personal and environmental contributors enables true insights into the
complex behavior of driving among young drivers. It highlights the fact
that both personal aspects, family and friends play a role in youngsters’
risk taking, pointing to the need to integratively consider these external
forces in any intervention aimed at moderating risks commonly taken
by young drivers. They also highpoint the positive aspects of shared
family and friends’ commitment to safety, which may be key elements
in such interventions. Such findings indicate that even if some cultural
differences in driving styles do exist in different countries, still, friends
and family are crucial elements in young drivers’ safety. This compre-
hension should set the grounds for campaigns rising awareness to safe
driving targeting both parents and young drivers.
References
Årseth, A.K., Kroger, J., Martinussen, M., Bakken, G., 2009. Intimacy status, attachment,
separation-individuation patterns, and identity status in female university students. J.
Soc. Pers. Relatsh. 26, 697–712.
Baer, P.E., Bray, J.H., 1999. Adolescent individuation and alcohol usage. J. Stud. Alcohol
13, 52–62.
Bener, A., Verjee, M., Dafeeah, E.E., Yousafzai, M.T., Mari, S., Hassib, A., Lajunen, T.,
2013. A cross ‘ethnical’ comparison of the Driver Behaviour Questionnaire (DBQ) in
an economically fast developing country. Glob. J. Health Sci. 5, 165–175.
Blos, P., 1979. The Adolescent Passage. International Universities Press, New York.
Bray, J.H., Baer, P.E., Getz, J.G., 2000. Adolescent individuation and alcohol use in
multiethnic youth. J. Stud. Alcohol 61, 588–597.
Bray, J.H., Adams, G.J., Getz, G., Baer, P.E., 2001. Developmental, family, and ethnic
influences on adolescent alcohol usage: a growth curve approach. J. Fam. Psychol.
15, 301–314.
Buelow, M.T., Brunell, A.B., 2014. Facets of grandiose narcissism predict involvement in
health-risk behaviors. Pers. Individ. Differ. 69, 193–198.
CARRS-Q, 2015. Graduated Driver Licensing. State of the Road. www.carrsq.qut.edu.au.
Carpentier, A., Brijs, K., Declercq, K., Brijs, T., Daniels, S., Wets, G., 2014. The effect of
family climate on risky driving of young novices: the moderating role of attitude and
locus of control. Accid. Anal. Prev. 73, 53–64.
Curry, A.E., Mirman, J.H., Kallan, M.J., Winston, F.K., Durbin, D.R., 2012. Peer passen-
gers: how do they affect teen crashes? J. Adolesc. Health 50, 588–594.
Farah, H., Polus, A., Bekhor, S., Toledo, T., 2007. Study of passing gap acceptance be-
havior using a driving simulator. Adv. Transp. Stud. Int. J. 23–31 (Special Issue).
Farah, H., Bekhor, S., Polus, A., Toledo, T., 2009. A passing gap acceptance model for
two- lane rural highways. Transportmetrica 5, 159–172.
Gelman, A., Pardoe, I., 2006. Bayesian measures of explained variance and pooling in
multilevel (hierarchical) models. Technometrics 48, 241–251.
Gil, S., Taubman-Ben-Ari, O., Toledo, T., 2016. A multidimensional intergenerational
model of young males’ driving styles. Accid. Anal. Prev. 97, 141–145.
Grotevant, H.D., Cooper, C.R., 1998. Individuality and connectedness in adolescent de-
velopment: review and prospects for research on identity, relationships, and context.
In: Skoe, E.E.A., von der Lippe, A.L. (Eds.), Personality Development in Adolescence:
A Cross National and Lifespan Perspective. Taylor & Francis/Routledge, Florence, KY,
pp. 3–32.
Guggenheim, N., Taubman-Ben-Ari, O., 2016. Safe Driving Climate Among Friends
(SDCaF): A New Scale. (Under Review).
Hartos, J.L., Eitel, P., Haynie, D.L., Simons-Morton, B.G., 2000. Can I take the car?: re-
lations among parenting practices and adolescent problem-driving practices. J.
V. Skvirsky et al.
Adolesc. Res. 15, 352–367.
Holland, C., Geraghty, J., Shah, K., 2010. Differential moderating effect of locus of control
on effect of driving experience in young male and female drivers. Pers. Individ. Differ.
48, 821–826.
Horvath, C., Lewis, I., Watson, B., 2012. Peer passenger identity and passenger pressure
on young drivers’ speeding intentions. Transp. Res. Part F: Traff. Psychol. Behav. 15,
52–64.
Israel Government Portal-gov. http://oldgov.gov.il/firstgov/english.
Jeffreys, H., 1946. An invariant form for the prior probability in estimation problems.
Proc. R. Soc. Lond. Ser. A: Math. Phys. Sci. 186, 453–461.
Kagitcibasi, C., Ataca, B., Diri, A., 2010. Intergenerational relationships in the family:
ethnic socioeconomic, and country variations in Germany, Israel, Palestine, and
Turkey. J. Cross-Cult. Psychol. 41, 652–670.
Levine, J., Green, C., Millon, T., 1986. The separation-individuation test of adolescence. J.
Pers. Assess. 50, 123–137.
Miller, L.C., Berg, J.H., Archer, R.L., 1983. Openers: Individuals who elicit intimate self-
disclosure. J. Personal. Social Psychol. 44, 1234–1244.
Ministry of Transport, State of Israel. http://rishuy.mot.gov.il/en/.
Ng, V.M., 2002. Robust Bayesian inference for seemingly unrelated regressions with el-
liptical errors. J. Multivar. Anal. 83, 409–414.
Nordfjarn, T., Simsekoglu, O., Rundmo, T., 2014. Culture related to road traffic safety: a
comparison of eight countries using two conceptualisations of culture. Accid. Anal.
Prev. 62, 319–328.
Özkan, T., Lajunen, T., Chliaoutakis, J., Parker, D., Summala, H., 2006. Cross-cultural
differences on driving behaviours: a comparison of six countries. Transp. Res. Part F
9, 227–242.
Poó, F.M., Taubman-Ben-Ari, O., Ledesma, R.D., Díaz-Lázaro, C.M., 2013. Reliability and
validity of a Spanish-language version of the multidimensional driving style in-
ventory. Transp. Res. Part F: Traff. Psychol. Behav. 17, 75–87.
Reason, J.T., Manstead, A.S.R., Stradling, S., Baxter, J., Campbell, K., 1990. Errors and
violations on the roads: a real distinction? Ergonomics 33, 1315–1332.
Santor, D.A., Messervey, D., Kusumakar, V., 2000. Measuring peer pressure, popularity,
and conformity in adolescent boys and girls: Predicting school performance, sexual
attitudes, and substance abuse. J. Youth Adolesc. 29, 163–182.
Schreer, G.E., 2002. Narcissism and aggression: is inflated self-esteem related to
54
Accident Analysis and Prevention 109 (2017) 47–54
aggressive driving? North Am. J. Psychol. 4, 333–342.
Shope, J.T., Bingham, C.R., 2008. Teen driving: motor-vehicle crashes and factors that
contribute. Am. J. Prev. Med. 35, 261–271.
Spiegelhalter, D., Best, N., Carling, B.P., Linde, A.V.D., 2002. Bayesian measures of model
complexity and fit. J. R. Stat. Soc. Ser. B (Stat. Methodol.) 64, 583–639.
Steinberg, L., Monahan, K.C., 2007. Age differences in resistance to peer influence. Devel.
Psy. 43, 1531–1543.
Taubman-Ben-Ari, O., Katz-Ben-Ami, L., 2012. The contribution of family climate for road
safety and aspects of the social environment to the reported driving behavior of
young drivers. Accid. Anal. Prev. 47, 1–10.
Taubman- Ben-Ari, O., Katz- Ben-Ami, L., 2013. Family climate for road safety: a new
concept and measure. Accid. Anal. Prev. 54, 1–14.
Taubman- Ben-Ari, O., Skvirsky, V., 2016. The multidimensional driving style inventory a
decade later: review of the literature and re-evaluation of the scale. Accid. Anal. Prev.
93, 179–188.
Taubman-Ben-Ari, O., Yehiel, D., 2012. Driving styles and their associations with per-
sonality and motivation. Accid. Anal. Prev. 45, 416–422.
Taubman-Ben-Ari, O., Mikulincer, M., Gillath, O., 2004. The multidimensional driving
style inventory -scale construct and validation. Accid. Anal. Prev. 36, 323–332.
Taubman-Ben-Ari, O., Eherenfreund-Hager, A., Prato, C.G., 2016. The value of self-report
measures as indicators of driving behaviors among young drivers. Transp. Res. Part F:
Traff. Psychol. Behav. 39, 33–42.
Taubman-Ben-Ari, O., 2016. Parents’ perceptions of the family climate for road safety:
associations with parents’ self-efficacy and attitudes toward accompanied driving,
and teens’ driving styles. Transp. Res. Part F: Traff. Psychol. Behav. 40, 14–22.
Warner, H.W., Özkan, T., Lajunen, T., Tzamalouka, G., 2011. Cross-cultural comparison
of drivers’ tendency to commit different aberrant driving behaviours. Transp. Res.
Part F: Traff. Psychol. Behav. 14, 390–399.
Winston, F.K., Jacobsohn, L., 2010. A practical approach for applying best practices in
behavioural interventions to injury prevention. Inj. Prev. 16, 107–112.
Zellner, A., 1962. An efficient method of estimating seemingly unrelated regression
equations and tests for aggregation bias. J. Am. Stat. Assoc. 57, 348–368.
Zellner, A., Ando, T., 2010. A direct Monte Carlo approach for Bayesian analysis of the
seemingly unrelated regression model. J. Econometr. 159, 33–45.