This article was downloaded by: [mohamed dhibi]

On: 19 July 2015, At: 16:37

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: 5 Howick Place,
London, SW1P 1WG

International Journal of Injury Control and Safety

Promotion
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/nics20

Dimensions of aberrant driving behaviours in Tunisia:

identifying the relation between Driver Behaviour
Questionnaire results and accident data
a b
Dhibi Mohamed & Belkacem Lotfi
a
Department of Quantitative Economics, Higher Institute of Management Sousse,
Laboratory Research for Economy, Management and Quantitative Finance, IHEC, University
of Sousse, Sousse, Tunisia
b
Research Laboratory for Economy, Management and Quantitative Finance, IHEC, University
of Sousse, Sousse, Tunisia
Click for updates
Published online: 15 Jul 2015.

To cite this article: Dhibi Mohamed & Belkacem Lotfi (2015): Dimensions of aberrant driving behaviours in Tunisia: identifying
the relation between Driver Behaviour Questionnaire results and accident data, International Journal of Injury Control and
Safety Promotion, DOI: 10.1080/17457300.2015.1061559

To link to this article: http://dx.doi.org/10.1080/17457300.2015.1061559

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 International Journal of Injury Control and Safety Promotion, 2015
http://dx.doi.org/10.1080/17457300.2015.1061559

Dimensions of aberrant driving behaviours in Tunisia: identifying the relation between Driver
Behaviour Questionnaire results and accident data
Dhibi Mohameda* and Belkacem Lotfib
a
Department of Quantitative Economics, Higher Institute of Management Sousse, Laboratory Research for Economy, Management and
Quantitative Finance, IHEC, University of Sousse, Sousse, Tunisia; bResearch Laboratory for Economy, Management and Quantitative
Finance, IHEC, University of Sousse, Sousse, Tunisia
(Received 11 January 2014; accepted 14 May 2015)

In this study, the Manchester Driver Behaviour Questionnaire (DBQ) was used to examine the self-reported driving
behaviours of a group of Tunisian drivers (N D 900) and to collect socio-demographic data, driver behaviours and DBQ
items. A sample of Tunisian drivers above 18 years was selected. The aim of the present study was to investigate the
factorial structure of the DBQ in Tunisia. The principal component analysis identified three factor solutions: inattention
errors, dangerous errors and dangerous violations. Logistic regression analysis showed that dangerous errors, dangerous
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

violations and speeding preference factors predicted crash involvement in Tunisia. Speeding is the most common form of
aberrant behaviour reported by drivers in the current sample. It remains one of the major road safety concerns.
Keywords: Driver Behaviour Questionnaire (DBQ); driving behaviours; dangerous violations; dangerous errors; speeding
preference; Tunisia

1. Introduction subtle effects, such as the amplification of other dangers

Tunisia has a very high road accident fatality rate; in in the traffic situation by relatively small increases in
2006, 11.7 persons per 10,000 vehicles were killed in speed, which are likely to be overlooked.
Tunisia (Ministry of Interior, 2006). The same figures Previous studies have shown that injury and fatality
were approximately 1.8 for Finland, 2.4 for France, 1.5 rates in Tunisia and Arab countries are much higher than
for United Kingdom and 1.9 for USA. According to the in the developing countries with comparable vehicle own-
official statistics of the Tunisian police, in 2008, total ership levels. The magnitude of the problem, therefore,
crashes numbered 10,073 making 1530 deaths and 14,085 dictates the need for more research into road accidents
injuries. Mean crash rate was 27.52 deaths per day. Mean and associated risk factors. In Tunisia, with its expansion
fatality rate was 4.18 deaths per day. Mean injury rate was of road construction and increase in the number of
38.48 injuries per day. According to a study carried out by vehicles, road traffic accidents are becoming an increas-
the World Bank in 1996, the economic and social costs of ingly serious public risk. The magnitude of the problem
the road accidents in Tunisia are estimated at 60,000 could be greatly reduced immediately if appropriate seri-
dinars for each death and 12,000 dinars for each injuries ous measures were taken concerning driving behaviour
(1 dinar D 0.83$). Excessive speed is reported to be an and driving competency evaluation. This problem needs
important contributory factor in many crashes (Ministry targeted research in order to identify methods of reducing
of Interior, 2008). Analyses of a number of large data accidents and fatalities.
bases in the United States indicated that speeding or The safety and mobility of roads have become a pre-
excessive speed contributed to approximately 12% of all occupation in the entire world. Various programmes
crashes reported to the police and to about one-third of should be developed in the aim of keeping drivers safe. It
fatal crashes (Bowie & Walz, 1991). In Australia, exces- is necessary to identify and evaluate the driving perfor-
sive speed is an important factor in approximately 20% of mance of Tunisian drivers. Safety interventions such as
fatal crashes (Haworth & Rechnitzer, 1993), and speeding collision warning and vehicle control system have consid-
is a probable or possible cause in 25% of rural crashes ered the limits of drivers’ perceptual and motor abilities
(Armour & Cinquegrana, 1990). Such figures are likely to that constrain their performance when responding to road-
underestimate the role of speeding in crashes because of way demands (Lee, 2006). Driver performance involving

*Email: [email protected]

Ó 2015 Taylor & Francis

 2 M. Dhibi and L. Belkacem

attitudes, goals and priorities has a significant influence on The DBQ has been used in different countries
driving safety (Evans, 1991). Attitudes and traits are iden- including Sweden (Aberg & Rimm€o, 1998), Greece (Kon-
tified by focusing on driver behaviour rather than on per- togiannis, Kossiavelou, & Marmaras, 2002), Finland, the
formance. Research in this area represents an important Netherlands (Lajunen, Parker, & Summala, 1999), Aus-
contribution to understanding driver crash involvement tralia (Blockey & Hartley, 1995; Dobson et al., 1999),
(Parker, Manstead, Stradling, & Reason, 1992). Driver Britain (Lajunen et al., 2004; Parker et al., 1995a;
behaviour has been extensively analysed using surveys, Reason et al., 1990), Turkey (Lajunen, 2002) and China
such as the Driver Behaviour Questionnaire (DBR), that (Stradling, Parker, Lajunen, Meadows, & Xie, 1998) on
identify types of drivers who are disproportionately likely target groups. Comparisons between results of these stud-
to be involved in crash. ies are difficult because of cultural differences and meth-
Research on driver behaviour and associated factor odological variations. The number of items has varied
analyses has shown that three distinct patterns of behav- considerably between versions used in different studies.
iour have a powerful influence on driver safety: errors, Sampling strategies, target populations and statistical
violations and lapses (Blockey & Hartley, 1995; Parker, analysis such as principal component analysis (PCA) with
West, Stradling, & Manstead, 1995b). Errors and viola- varimax rotation and PCA with oblimin rotation can also
tions are supposed to have different psychological origins be different. In general, the results have commonly shown
(Parker et al., 1992; Reason, Manstead, Stradling, Baxter, that DBQ factors have different patterns of correlations
& Campbell, 1990). Errors can be defined as a type of with driver demographic and descriptive variables. Males
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

driving mistake involving failures of observation or mis- of all ages report engaging in violations more often than
judgement, and include such behaviours as failing to females (Parker et al., 1995b; Parker et al., 1998; Rimm€ o
notice ‘stop’ or ‘give way’ sign or failing to check your & Hakamies-Blomqvist, 2002). Violations are generally
mirrors before pulling out or changing lanes (Parker, reported with higher frequency by younger drivers (Parker
Lajunen, & Stradling, 1998). Violations can be defined as et al., 1995b, Rimm€o & Hakamies-Blomqvist, 2002).
a deliberate deviation from those practices thought to be Drivers who report engaging in aggressive violations
necessary to safely operate a vehicle, and include such more frequently tend to be younger and male (Lawton,
behaviours as speeding and closely following others (Rea- Parker, Manstead, & Stradling, 1997; Mesken, Lajunen,
son et al., 1990). Violations have been distinguished as & Summala, 2002). However, errors are not normally
aggressive or ordinary by Lajunen, Parker and Summula associated with any particular demographic group (Parker
(2004). Lapses involve problems with attention and mem- et al., 1998; Sullman, Meadows, & Pajo, 2002). Lapses
ory and include such things as switching on one thing have also been found to be associated with age, with older
when meaning to switch on something else (Parker et al., drivers reporting more lapses than younger drivers (Parker
1998). Lapses have been divided into two types: errors et al., 1998; Rimm€o & Hakamies-Blomqvist, 2002).
due to inattention or inexperience by Rimm€ o and Haka- To explore the role of driving behaviour self-assess-
mies-Blomqvist (2002). ment on drivers’ decisions about driving, DBQ was used
Previous research has addressed the problems of to explain why Tunisian drivers commit aberrant driving
aggressive behaviour and driving offences, and their rela- behaviours. The aim of this study is threefold:
tion to traffic accidents. Many studies demonstrated that
approximately 90% of all crashes could be attributed to (1) Determine factors that affect driving behaviour
road user characteristics (Aberg & Rimm€ o, 1998; Bener and replicate the distinction between errors,
& Alwash, 2002; Bener, Breger, & Al-Falasi, 1994; lapses and violations in a sample of 900 Tunisian
Lajunen et al., 2004; Parker, Reason, Manstead, & Stra- drivers.
dling, 1995a). The vast amount of the existing literature (2) Examine the relationship between driver behav-
shows that design of roadways, physical environment, iour and DBQ scores.
road design signalling and traffic pattern issues also have (3) Investigate which demographics and general driv-
a huge role. Many other factors including tiredness, stress, ing self-assessment items provided the best pre-
state of the road, state of the vehicle and climatic factors dictors of DBQ factors.
are also important and decisive when the driver commit-
ted an accident.
It is important to examine what drivers are doing when
2. Method
driving. The behaviour adopted by the driver can have neg-
ative or positive consequences. To analyse the aberrant 2.1. Sample of participants
driving behaviour, the DBQ has been used in several inter- A total of 900 individual drivers volunteered to participate
national studies. The DBQ is designed to measure three in the study: 666 (74%) males and 244 (24%) females.
factors: lapses, errors and violations that describe the inat- The average age of the sample was 38 years (range
tentive and experience attitude of the driver while driving. 18 65 years).
 International Journal of Injury Control and Safety Promotion 3

Participants were located throughout Tunisia in both second part included some questions on driver behaviour:
urban and rural areas. Vehicles were reported to be sedans self-reported previous crashes during their lifetime, type
(86%), four wheel drivers (12%) or other (2%). Participants of crashes, cause of crashes, time of crashes, type and
were drivers who had driven at least once in the last six nature of injury, use of mobile phone, seatbelt use and
months. In total, just 900 questionnaires were distributed why the respondent–driver committed errors. The third
and 861 responses were received, giving a response rate of part included the Manchester DBQ (Reason, 1990). It was
just 95.6%. Sixty-seven per cent of respondents live in urban used with 50 items relating to aberrant behaviour, i.e.
areas and only 33% in rural areas. The sex ratio was errors or mistakes, violations and lapses.
(female/male) 0.39. The mean age of males was 37.9 § 9.9 All information were based on structured face-to-face
and for females 30.9 § 11. Most were young and 30.3% of interviews. Only background information, i.e. age, gender,
the male drivers were in the age group of 20 29 and 36.1% mileage, number of crashes, self-assessment of driving
of females were in the age group of 30 39 years. Statisti- ability and the DBQ, will be used.
cally, there was a significant difference observed between
males and females (p < 0.001). The males had significantly
higher driving experience in years than females (p < 2.3. Statistical analysis
0.001). There were no significant differences between males Analyses were performed using Statistical Package for
and females with respect to the use of seat belts. Social Sciences (SPSS) for the statistical software applica-
Regarding mileage, 12% of the subjects cover less tion. PCA with varimax rotation was used to examine the
factor structure of the DBQ in Tunisia. Cronbach’s alpha
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

than 10,000 km per year, 25% cover between 10,000 and

20,000 km, 30% between 20,000 km and 60,000 km, 22% reliability coefficients were also calculated for assessing
cover between 60,000 and 80,000 km and, lastly, 11% the internal consistency of the DBQ scale scores. The rela-
drove more than 80,000 km per year. Eleven per cent of tionships between crash involvement and the main varia-
the drivers have traffic experience of less than 2 years, bles were initially examined through the calculation of
40% were experienced for 3 6 years, 33% were experi- Pearson’s correlation coefficients, and through the calcu-
enced for 6 15 years and 16% were experienced for lation of inter-correlation Pearson coefficients. A direct
more than 15 years. As for the number of crashes involv- logistic regression was performed with crash involvement
ing physical injury during the last five years, 58% (500 as dependent and three factors with the sociodemographic
persons) of the respondents had no crashes, 28% (241 characteristics as independent. Variables having signifi-
drivers) had a single crash, 8.9% (67 drivers) had two and cant Pearson’s correlation coefficients (age, sex, annual
5.1% had three or more than three crashes. mileage, speeding preference) were entered into the model
to control their effect in the first step. Student’s t-test was
used to find the difference between means of two continu-
2.2. Questionnaire ous variables. The Chi-square test and Fisher’s exact test
The questionnaire used in this study was composed of ‘the were used to compare frequencies between two or more
driver’s behaviours’ questionnaire improved by Reason than two categories. Unless specifically noted, a 5% level
et al. (1990). All 50 items of the DBQ used by Reason of significance was used.
et al. (1990) has been translated into Arabic. A translated
Arabic version of the DBQ was revised, translated back
3. Results
by a bilingual expert and unacquainted with the original
English version. The items were translated simultaneously 3.1. Mean frequencies of DBQ items
and independently from English to Arabic. Next, a prelim- Mean scores for lapses, errors, violations, speed and thor-
inary DBQ made up of the translated items was informally oughness in decision-making were calculated for each
shown to a set of interviewees. Only one item was adapted participant by averaging over the relevant items. The
to the Tunisia Highway Code, because in Tunisia, the means and standard deviation for all 50 questionnaire
renewal of car insurance is automatic every year (Item 31 items are given in Table 1. The most frequently occurring
in the original DBQ (Reason et al., 1990)). This item has behaviour involves aberrant behaviours that have conse-
been replaced with ‘Contrary to Highway Code, you drive quences on the driver and others. The seven frequently
without seat belt’. For each question of the DBQ, respond- occurring behaviours were unknowingly speeding, mis-
ents were required to indicate, on a six-point scale ranging judge speed of oncoming vehicle, disregard speed at
from 0 D never to 5 D nearly all the time, the frequency night, hit something when reversing, racing, queuing,
with which they committed each type of aberrant behav- nearly hitting car in the front and risky overtaking. Table 1
iour. Participants were assured that the information given depicts the mean and standard deviation scores for the
will be kept confidential. three highest ranked items, which were unknowingly
The questionnaire was divided into three sections. The speeding (M D 2.45, SD D 1 .234), disregard speed at
first part was used to collect sociodemographic data. The night (M D 1.78, SD D 0.982) and misjudge speed of
 4 M. Dhibi and L. Belkacem

Table 1. Mean frequencies and standard deviation of DBQ items.

Number Behavioural type Item Mean SD

DBQ2 UV Unknowingly speeding 2.45 1.234

DBQ30 S Misjudge the speed of oncoming vehicle 2.16 1.082
DBQ21 V Disregard speed at night 1.78 0.982
DBQ31 M Hit something when reversing 1.61 0.775
DBQ47 V Racing 1.50 1.058
DBQ25 S Queuing, nearly hitting car in front 1.03 0.859
DBQ16 V Risky overtaking 1.02 0.867
DBQ29 V Illegal parking 0.99 0.764
DBQ22 V Drive without seat belt 0.97 1.046
DBQ36 V Cut corner turning left 0.96 0.67
DBQ45 S Only half-an-eye on the road 0.81 0.652
DBQ44 V Disregard traffic lights late on 0.72 0.518
DBQ35 V Overtake on right on motorway 0.64 0.522
DBQ42 S Try to pass vehicle turning left 0.63 0.581
DBQ27 V Have an aversion 0.59 0.872
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

DBQ7 V Closely following another driver 0.56 0.654

DBQ28 UV Fail to see pedestrian waiting 0.51 0.762
DBQ3 S Locked out of car 0.49 0.735
DBQ40 V Ignore give-way signs 0.47 0.653
DBQ17 S On usual route by mistake 0.44 0.722
DBQ4 V Overtake on the inside 0.42 0.685
DBQ13 S No recollection of recent road 0.41 0.756
DBQ11 M Turn right into vehicle’s path 0.40 0.644
DBQ50 M Misjudge interval turning left 0.39 0.566
DBQ1 S Attempt driving 0.38 0.391
DBQ26 V Drink and drive 0.37 0.837
DBQ39 V Fail to give way to bus 0.36 0.592
DBQ20 S Try to pass without using mirror 0.34 0.583
DBQ14 S Miss motorway exit 0.33 0.431
DBQ5 M Drive as fast on dipped lights as on main beam 0.32 0.213
DBQ33 M Plan route badly 0.29 0.387
DBQ46 S Fail to see pedestrians crossing 0.28 0.601
DBQ34 M Overtake queue 0.27 0.421
DBQ48 V Race for a gap 0.27 0.634
DBQ12 M Misjudge gap in car park 0.26 0.629
DBQ41 S Manoeuvre without checking mirror 0.24 0.486
DBQ9 S Distracted, have to brake hard 0.23 0.661
DBQ38 S Get into wrong lane at roundabout 0.21 0.712
DBQ15 S Forget which gear 0.20 0.841
DBQ49 M Brake too quickly 0.19 0.672
DBQ23 S Forget light on main beam 0.18 0.568
DBQ10 S Intend lights but switch on wipers 0.18 0.845
DBQ18 V Shoot lights 0.17 0.543
DBQ24 S Turning right, nearly hitting cyclist 0.16 0.243
DBQ43 V Drive wrong way down one-way street 0.15 0.312
DBQ37 M Get into wrong lane at roundabout 0.14 0.423
DBQ19 V Angry, give chase 0.14 0.875
DBQ6 S Try driving off without switching on 0.13 0.241
DBQ32 S Fail to see pedestrian stepping out 0.12 0.342
DBQ8 S Forget where car is 0.11 0.462
 International Journal of Injury Control and Safety Promotion 5

Table 2. Factor structure of DBQ items.

Factors

Items Inattention error Dangerous error Dangerous violation

DBQ37: Get into wrong lane at roundabout 0.802

DBQ31: Hit something when reversing 0.765
DBQ11: Turn right into vehicle’s path 0.734 0.311
DBQ14: Miss motorway exit 0.703
DBQ17: On usual route by mistake 0.667
DBQ33: Plan route badly 0.613
DBQ23: Forget light on main beam 0.522
DBQ12: Misjudge gap in car park 0.494
DBQ34: Overtake queue 0.435 0.442
DBQ25: Queuing, nearly hitting car in front 0.412 0.389
DBQ9: Distracted, have to brake hard 0.406
DBQ15: Forget which gear 0.385
DBQ3: Locked out of car 0.348
DBQ13: No recollection of recent road 0.324 0.305
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

DBQ38: Get into wrong lane at roundabout 0.318

DBQ10: Intend lights but switch on wipers 0.306
DBQ20: Try to pass without using mirror 0.781
DBQ30 :Misjudge speed of oncoming vehicle 0.742
DBQ50: Misjudge interval turning left 0.702 0.474
DBQ42: Try to pass vehicle turning left 0.698
DBQ40: Ignore give-way signs 0.688
DBQ41: Manoeuvre without checking mirror 0.656
DBQ49: Brake too quickly 0.321 0.639
DBQ16: Risky overtaking 0.603
DBQ24: Turning right, nearly hitting cyclist 0.582
DBQ18: Run a lights 0.542 0.422
DBQ28: Fail to see pedestrian waiting 0.511 0.482
DBQ39: Fail to give way to bus 0.467
DBQ48: Race for a gap 0.442
DBQ36: Cut corner turning left 0.406
DBQ46: Fail to see pedestrians crossing 0.352
DBQ1: Attempt driving off in third 0.322
DBQ21: Disregard speed at night 0.708
DBQ5: Speed up at light changes? 0.681
DBQ4: Overtake on the inside 0.643
DBQ35: Overtake on right on motorway 0.609
DBQ2 Unknowingly speeding 0.576
DBQ22: Drive without seat belt 0.546
DBQ26: Drink and drive 0.529
DBQ47: Racing 0.503
DBQ27: Have an aversion 0.476
DBQ45 Only half-an-eye on the road 0.449
DBQ44: Disregard traffic lights late on 0.421
DBQ7: Closely following driver in front 0.364
DBQ19: Angry, give chase 0.312

Notes: Extraction method; principal component analysis rotation method; varimax with Kaiser normalization.
Factor loadings of less than 0.3 are omitted for the sake of.
 6 M. Dhibi and L. Belkacem

Table 3. Correlations with number of reported crashes. Finally, the third factor ‘dangerous violations’ included
Variables Correlations
10 items. It is composed of violations and slips which
define a possible definite risk. This factor explained
Gender ¡0.0214 5.62% of the total variance. Reliability analyses for the
Age ¡0.138 three-factor solution of the DBQ revealed that the alpha
Mileage per year 0.413 reliabilities for each factor were 0.78 (inattention errors);
Year of experience driving ¡.089 0.77 (dangerous errors) and 0.72 (dangerous violations).
Speed preference 0.421
Inattention errors 0.077
Dangerous errors 0.122 3.3. Correlations
Dangerous violations 0.431 The relationships between crash involvement and the
main variables were initially examined through the calcu-
Note: Speed preference is a combination of the five speed items
(unknowingly speeding, misjudge speed of oncoming vehicle, disregard
lation of Pearson’s product-moment correlation coeffi-
speed at night, hit something when reversing, racing). cients (Table 3). Table 3 shows that drivers who reported

p < 0.05; p < 0.01; p < 0.001. having a crash in the previous two-year period were more
likely to be younger (p < 0.05), have less experience in
oncoming vehicle (M D 2.16, SD D 1.082). The driver driving (p < 0.01), reported a higher annual mileage (p <
behaviour in Tunisia compared with United Kingdom, 0.01), and tended to have higher scores on inattention
Finland and Australia demonstrated that there was a sig- errors (p < 0.01), dangerous errors (p < 0.05) and danger-
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

nificant difference. Previous research that has focused on ous violations (p < 0.01). Preferred speed has positive
college students (Bianchi & Summala, 2004), elderly relationship with crash involvement (p < 0.001). It shows
drivers (Parker, McDonald, Rabbitt, & Sutcliffe, 2000) the effect of speeding in accidents and their severities.
and professional drivers (Sullman et al., 2002; Xie & Table 4 presents the inter-correlation Pearson coeffi-
Parker, 2002) indicate that Tunisian drivers sampled have cients for participants’ driving exposure and DBQ factors.
higher means for their questionnaire items. Consistent with the previous research (Kontogiannis,
Kossiavelou, Marmaras, 2002; Stradling et al., 1998), age
and years of driving experience appear to have a signifi-
3.2. Factor structure of the DBQ in Tunisia cant negative relationship with dangerous errors, danger-
The factor structure of the DBQ (Table 2) was examined ous violations and inattention errors. One possible
using a PCA with a varimax rotation. The interpretation explanation is that as drivers gain more experience, they
of the screen plot suggested the three-factor solution to be are less likely to engage in aberrant driving behaviour on
the most interpretable one, whereas 17 factors had eigen- public roads. However, contrary to previous research
values greater than one. The three factors accounted for (Parker et al., 1995a; Sullman et al., 2002;Chliaoutakis,
34% of the total variance (33% in Reason et al., 1990). Koukouli, Lagunen, & Tzamlalouka, 2005; Lajunen,
The first factor included 16 items. It seems to reflect Parker, & Stradling, 1998), a positive relationship was
mostly ‘inattention errors’ on the road, explaining 20.34% notified between the numbers of kilometres driven each
of the total variance. This factor consisted of lapses and year and dangerous violations.
mistakes which posed mix risk to other road users.
The second factor ‘dangerous errors’ consisted of 14
items. This component explained 8.64% of the total vari- 3.4. Predictions
ance and consisted of lapses, mistakes and violation. This Finally, logistic regression was performed to identify
factor seems to describe a definite risk to other road users. which of the DBQ factor scores significantly improved
Table 4. Correlations among DBQ factors, demographics variables and driving self-assessment.

Variables 2 3 4 5 6 7 8

(1) Gender
(2) Age ¡0.138 1
(3) Mileage per year ¡0.413 ¡0.157 1
(4) Year of experience .089 ¡0.201 0.212 1
(5) Speed preference ¡0.121 0.051 0.177 0.025 1
(6) Inattention errors 0.008 ¡0.21 ¡1.98 0.068 1
(7) Dangerous errors ¡0.122 ¡0.01 0.081 ¡0.256 0.134 0 1
(8) Dangerous violations ¡0.431 ¡0.178 0.341 ¡0.015 0.211 0 0 1
 
Note: p < 0.05, p < 0.01.
 International Journal of Injury Control and Safety Promotion 7

Table 5. Logistic regression analysis with self-reported crash involvement as the dependent variable.

Variables (B) B SE Wald p Odds ratio 95% C.I. lower Exp upper

Step 1
Gender .16 .02 42.64 .0001 .721 .12 1.23
Age ¡.21 .13 1.41 .200 .811 .79 1.91
Model Chi-square 40.61 (df D 2)
Step 2
Gender .13 .03 23.83 .000 1.13 1.08 1.19
Age ¡.32 .17 .77 .308 .726 .92 1.18
Kilometre per year .11 .02 13.55 .005 1.11 1.08 1.19
Year of experience ¡.015 .07 .378 .724 .985 .84 1.12
Speed preference .016 .14 7.27 .007 1.17 .11 1.88
Inattention errors .05 .13 .15 .702 1.05 .81 1.35
Dangerous errors .13 .13 6.53 .317 1 .14 .89 1.46
Dangerous violations .127 .426 5.11 .000 1.135 .68 1.21

Model Chi-square 82.20 (df D 8)

Note: p < 0.05; p < 0.01; p < 0.001;

Downloaded by [mohamed dhibi] at 16:37 19 July 2015

the prediction of crashes beyond that afforded by knowl- structure of the DBQ used by Reason et al. (1990) and
edge of demographic and descriptive variables. Table 5 Blockey and Hartley (1995) with a Tunisian sample. This
depicts the variables in each model, the regression coeffi- is the first survey using the Manchester DBQ in Tunisia.
cients (B) and their standard deviation (SDB), as well as Using it has shown that Tunisian drivers have scored high
the Wald, odds ratio values and their confidence interval. on violations, errors and intentional violations. In fact, the
The demographic and descriptive variables were entered factor structure produced here was very similar to that
into the regression prior to the factor scores. Surprisingly, found in earlier studies (Reason and Blockey). The three
speed preference (beta D .016, Wald D 7.27, p < 0.01), factors are inattention errors, dangerous errors and dan-
dangerous errors (beta D .13, Wald D 6.53, p < 0.01) and gerous violations.
dangerous violations (beta D ¡.127, Wald D 5.11, p < First, reliability analysis of the DBQ indicated coeffi-
0.01) factors were significant predictors of crash involve- cients that were relatively robust and acceptable compared
ment. The odds ratio (Exp(B)) shows that for every one- to the other countries (United Arab Emirates, France,
unit increase in dangerous errors, the chances of being Switzerland and Australia). Second, examination of the
crash involved increased by around 14%. In addition, with above mean scores for the original DBQ factors revealed
every one-unit increase in the dangerous violation factor high scores explaining that participants reported engaging
score, the chances of being involved in a crash increased in a higher frequency of Highway Code violation, com-
by 13%. Every one-unit increase in the speed preference pared to Spain and France. Speeding is the most regularly
factor score increased the accident risk by 17%. To further reported aberrant driving behaviour on public roads.
investigate the relationship between the DBQ factors and Speeding violations are the most common form of aber-
crash involvement, logistic regressions were performed to rant behaviour both for males and females. This result
test which of the factors could predict active and passive may also reflect that speeding violations pose a serious
crash involvements. This found that after the demographic road safety risk. These results suggest that targeted strate-
and descriptive variables had been partialled out, it was gies are needed for reducing road crashes among Tunisian
the errors factor and speed preference factor which were drivers. These could focus on reducing speeding and
significant predictors of both active and passive crashes. violations.
In regards to the results, the speeding preference proved Third, an exploratory factor analysis of the full-scale
to be a significant predictor of crashes; as not surprisingly, DBQ was conducted to determine the consistency of tradi-
speed is the first cause of fatality in Tunisia. tional DBQ factors for the current sample of Tunisian
drivers. Similar to the previous research, a three-factor
solution was established from the DBQ. The sample
4. Discussion included young, middle-age and elderly drivers, the same-
DBQ was used to investigate driving behaviours of 900 factor structure found by Reason et al. (1990) and Blockey
Tunisian motorists and to understand the characteristic and Hartley (1995); however, different relative contribu-
features of them. DBQ scale with 50 items was imple- tions to the total variance existed. The factor explaining
mented. The first objective was to confirm the factor the majority of the total variance is ‘inattention errors’,
 8 M. Dhibi and L. Belkacem

i.e. ‘silly errors’ in Reason et al. (1990) or ‘general errors’ climatic factors are also important variables that influence
in Blockey and Hartley (1995), and ‘dangerous violations’ the drivers behaviours. These variables are not retained at
is the factor which contributes the least to the variance. the level of the DBQ and, thus, are difficult to analyse. In
This situation confirms the idea that two types of behav- Tunisia, 70% of severe accidents have one person’s death
iours have different psychological origins. This result to a minimum. Tiredness and stress are the most important
explains the different rates of fatalities between states and causes of injury severity. Twenty-five per cent of fatal
suggests that the psychological situation of drivers has an accidents have occurred in places and specific routes. This
important influence on driving behaviour, as well as play- failure at the DBQ level suggests by integrating another
ing a critical role in driving safety. questionnaire which takes into account the psychological
In line with the previous research (Aberg & Rimm€o, state of the driver and the condition of the vehicle used.
1998; Parker et al., 1995a; Reason et al., 1990, Sullman et In conclusion, the strength of this study is the over-
al., 2002), speeding on the motorway and in residential sampling drivers in Tunisia. This is the first survey using
areas was the most often reported behaviour by Tunisian the Manchester DBQ in Tunisia. Some methodological
drivers. The strategy of road safety adopted in Tunisia is limitations do exist. Sampling bias does exist, and the
not able to change the behaviours of drivers, because response rate of participants was not extremely high.
Tunisian crash rates are still increasing. Also, in agree- Previous research that has focused on driving behaviour
ment with the previous research (Parker et al., 1995b; has used the DBQ with less questionnaire items (Lajunen
Parker et al., 1998; Rimm€ o & Hakamies-Blomqvist, et al., 2004; Mesken et al., 2002). In this survey, the origi-
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

2002), males reported engaging more often in violations nal DBQ includes 50 items.
than females. However, in this study, younger drivers
report more inattention errors and aggressive violations
than older drivers. Age and years of driving experience Acknowledgements
were significantly negatively correlated with the number We would like to thank Prof. Leonard Evans (USA), Prof. Xav-
of crashes reported. Also, in agreement with the previous ier Godard (Inrets France)., for their valuable advice and infor-
mation, which have improved the presentation of this paper.
research was the positive correlation with mileage (Stra-
dling et al., 1998). When someone spends more time on
the road, their chances of being involved in some kind of
crash will increase. Disclosure statement
Regarding dangerous violations, elderly drivers commit No potential conflict of interest was reported by the author.
less violation than younger drivers. The DBQ is a good
predictor of the driving behaviours of Tunisian drivers.
References
Inattention and dangerous errors were significantly associ-
Aberg, L., & Rimm€ o, P.-A. (1998). Dimensions of aberrant
ated with errors by those who think they are bad drivers. driver behaviour. Ergonomics, 41(1), 39 56.
Inattention errors increase with the age. Dangerous errors Armour, M., Cinquegrana, C. (1990). Victorian study of single
that have the highest factor loading are specific errors of vehicle rural accidents. Proceedings of the 15th Australian
aged drivers. Elderly drivers commit fewer violations than Road Research Board Conference, Part 7, Australian Road
younger drivers, and their errors are very specific. Research Board, Melbourne, pp. 79 91.
Bener, A., & Alwash, R. (2002). A perspective on motor vehicle
The results relative to Tunisian sampled drivers have crash injuries and speeding in the United Arab Emirates.
shown that mean scores of inattention errors, dangerous Traffic Injury Prevention, 3, 61 64
errors and dangerous violations are high compared to Bener, A., Breger, E., Al-Falasi, A. (1994). Risk-taking behav-
European countries and Australia. This difference has iour in road traffic accidents. Journal of Traffic Medicine,
also shown in injury and fatality rates. Fatality rate in 23, 65 70.
Bianchi A, Summala H. (2004). The “genetics” of driving
Tunisia is higher than developing countries with compara- behavior: parents’ driving style predicts their children’s
ble vehicle ownership. The problem and the difference in driving style. Accident, Analysis and Prevention Journal, 36
fatality rates between states suggest that culture has an (4), 655 659.
important influence in driving behaviour, as well as play- Blockey, P.N., & Hartley, L.R. (1995). Aberrant driving behav-
ing a critical role in driving safety more generally (Lee, iour: Errors and violations. Ergonomics, 38(9), 759 1771.
Bowie, N., Walz, M. (1991). Data analysis of the speed-related
2006). Driving behaviours constitute an important risk crash issue. Paper presented to 13th International Technical
factor. The speeding preference proved to be a significant Conference on Experimental Safety Vehicles, Paris, 4 7
predictor of crashes, as not surprisingly, speed is the first Nov.
cause of fatal crashes in Tunisia. However, age, sex, driv- Chliaoutakis J., Koukouli S., Lagunen T., & Tzamlalouka G.
ing experience, education, tiredness, reaction time and (2005). Lifestyle traits as predictors of driving behaviour in
urban areas of Greece. Transport Research Part F, 8,
driving speed also play significant roles and need to be 413 428.
considered as associated risk factors in road accidents. Dobson, A., Brown, W., Ball, J., Powers, J., & McFadden, M.
Tiredness, stress, state of the road, state of vehicle and (1999). Women drivers’ behaviour, socio-demographic
 International Journal of Injury Control and Safety Promotion 9

characteristics and accidents. Accident Analysis & Preven- Ministry of Interior. (2006). Tunisian National Police report on
tion, 31, 525 535. road accidents, p. 60.
Evans, L. 1991. Traffic safety and the driver. New York, NY: Ministry of Interior. (2008). Tunisian National Police Report on
Van Nostrand Reinhold. Road Accident, p. 65.
Hakamies-Blomqvist, L. (1994). Compensation in older drivers Parker, D., Lajunen, T., & Stradling, S. (1998). Attitudinal pre-
as reflected in their fatal accidents. Accident Analysis and dictors of interpersonally aggressive violations on the road.
Prevention, 26(1), 107 112. Transport Research Part F: Traffic Psychology and Behav-
Haworth, N., Rechnitzer, G. (1993). Description of fatal crashes iour, 1(10), 11 24.
involving various causal variables. Department of Transport, Parker, D., Manstead, A.S.R., Stradling, S.G., & Reason, J.T.
Federal Office of Road Safety, CR 119, Canberra, Australia. (1992). Determinants of intention to commit driving viola-
Kontogiannis, T., Kossiavelou, Z., & Marmaras, N. (2002). Self- tions. Accident Analysis and Prevention, 24(2): 117 131.
reports of aberrant behaviour on the roads: errors and viola- Parker, D., McDonald, L., Rabbitt, P., & Sutcliffe, P. (2000).
tions in a sample of Greek drivers. Accident Analysis & Pre- Elderly drivers and their accidents: The Aging Driver
vention, 34(3), 381 399. Questionnaire. Accident Analysis and Prevention, 32(6),
Lajunen, T. (2002). Psycho-technicol driver assessment: a useful 751 759.
tool for detecting bad drivers or completewaste of money? Parker, D., Reason, J., Manstead, A., & Stradling, S. (1995a).
Uluslararasi Trafic veyol Guventigi Kongresi ve Fuari, Driving errors, driving violations and accident involvement.
Ankara, Turkey. Ergonomics, 38(5), 1036 1048.
Lajunen, T., Parker, D., & Stradling, S. (1998). Dimensions Parker, D., West, R., Stradling, S.G., & Manstead, A.S.R.,
of driver anger, aggressive and highway code violations (1995b). Behavioural characteristics and involvement in dif-
and their mediation by safety orientation. Transportation ferent types of traffic accident. Accident Analysis & Preven-
Research Part F, 1, 107 121. tion, 27, 571 581.
Downloaded by [mohamed dhibi] at 16:37 19 July 2015

Lajunen, T., Parker, D., & Summala, H. (1999). Does traffic con- Reason, J., Manstead, A., Stradling, S., Baxter, J., & Campbell,
gestion increase aggression? Transportation Research Part K.L. (1990). Errors and violations on the roads: A real dis-
F: Traffic Psychology & Behaviour, 2, 225 236. tinction? Ergonomics, 33(10/11), 1315 1332.
Lajunen, T., Parker, D., & Summala, H. (2004). The Manchester Reason, J.T., 1990. Human error. New York, NY: Cambridge
Driver Behaviour Questionnaire: A cross-cultural study. University Press.
Accident Analysis & Prevention, 36, 231 238. Rimm€ o, P.-A., & Hakamies-Blomqvist, L. (2002). Older drivers’
Lawton, R., Parker, D., Manstead, A.S.R., & Stradling, S. aberrant driving behaviour, impaired activity, and health as
(1997). The role of affect in predicting social behaviors: The reasons for self-imposed driving limitations. Transportation
case of road traffic violations. Journal of Applied Social Psy- Research Part F, 5, 47 62.
chology, 27, 1258 1276. Stradling, S.G., Parker, D., Lajunen, T., Meadows, M.L., & Xie,
Lee, J.D. 2006. Driving safety. In R.S. Nickerson (Ed.), Review C.Q. (1998). Normal behaviour and traffic safety: Viola-
of human factors. Santa Monica, CA: Human Factors and tions, errors, lapses and crashes. Retrieved from http://mille
Ergonomics Society. miglia.ce.unipr.it/ARGO/theysay/urban/
Mesken, J., Lajunen, T., & Summala, H. (2002). Interpersonal Sullman, M., Meadows, M., & Pajo, K. (2002). Aberrant driving
violations, speeding violations and their relation to accident behaviours amongst New Zealand truck drivers. Transporta-
involvement in Finland. Ergonomics, 45(7), 469 483. tion Research Part F, 5, 217 232.