You are on page 1of 12

Accident Analysis and Prevention 38 (2006) 913–924

Patterns of driver stress and coping strategies in a Greek sample and their
relationship to aberrant behaviors and traffic accidents
Tom Kontogiannis
Department of Production Engineering and Management, Technical University of Crete, University campus, Chania 73100, Crete, Greece
Received 12 July 2005; received in revised form 20 February 2006; accepted 14 March 2006

Abstract
This paper presents an investigation into the relationship between driver stress, coping strategies and aberrant driving of a Greek sample of
company employees (N = 714). The results supported the main factor structures of the Driver Behavior Inventory (Matthews, G., Tsuda, A., Xin,
G., Ozeki, Y., 1999. Individual differences in driver stress vulnerability in a Japanese sample. Ergonomics 42, 401–415) and Driver Behavior
Questionnaire (Kontogiannis, T., Kossiavelou, Z., Marmaras, N., 2002. Self-reports of aberrant behavior on the roads: errors and violations in
a sample of Greek drivers. Accident Anal. Prev. 34, 391–399). An inventory of coping strategies was also examined in terms of self reports.
Confrontive coping was characteristic of drivers high on aggression who also had higher rates of mistakes and violations. Coping in terms of
self-criticism was exhibited by drivers high on dislike of driving (i.e., anxiety) who reported a higher incidence of mistakes. Task-focus coping was
characteristic of alertness and had a weak correlation with confidence. Drivers high in confidence reported fewer mistakes and violations which, in
association with low perception of risk, was a particularly worrying aspect of driving. A contextual model of accident involvement was tested with
LISREL in which violations yielded a direct effect whilst aggression yielded an indirect effect mediated by violations. Alertness and confidence
were both related to safety orientation but failed to predict accident rates and speeding convictions.
© 2006 Elsevier Ltd. All rights reserved.

Keywords: Driver stress vulnerability; Coping strategies; Aberrant behaviors; Traffic accidents

1. Introduction becomes an overloaded system with higher frequencies of con-


gestion and other daily hassles (e.g., tail gaiting, driver hostility),
Untangling the role of human behavior in the complex chain the relationship between driver stress and aberrant behaviors
of events leading to traffic accidents has been a worthwhile task emerges as a challenging research issue.
in traffic safety. A promising line of research (e.g., Reason et The majority of the research on driver stress and aberrant
al., 1990; Aberg and Rimmo, 1998) focused on understanding behaviors has used self-reports of drivers’ habits because of the
the varieties of aberrant driving behaviors and their association wealth of information that can be gleaned from such accounts in
to the Risk of Traffic Accidents (RTA). Aberrant behaviors have comparison to official accident reports. Self-reporting of human
been broadly classified into two categories: ‘errors’ defined as behavior, however, can be vulnerable to biases and there is a
deviations from the right path in achieving goals and ‘violations’ requirement that patterns of driving behaviors are replicated in
defined as deliberate circumventions of traffic rules and socially different studies both within and between countries. The moti-
approved codes of behavior. Both errors and violations have been vation of the present study has been to explore patterns of driver
found to be associated with high levels of stress experienced stress and coping strategies in a Greek sample of company
in the traffic environment. Performance mechanisms linking employees and examine how driver stress relates to several forms
driver stress to aberrant driving may include: impaired psy- of errors and violations. This objective is particularly important
chomotor control, poor hazard detection, increased risk-taking since Greece has the worst record in traffic accidents in EEC
and task interferences arising from managing emotions and dis- that is approximately twice the average rate of constituent states
tress (Matthews et al., 1998). Given that the traffic environment (Chliaoutakis et al., 1999).
Errors and violations are examined with the use of the Driver
Behavior Questionnaire (Parker et al., 1995; Aberg and Rimmo,
E-mail address: konto@dpem.tuc.gr. 1998) and a later version applied to a sample of Greek drivers

0001-4575/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2006.03.002
914 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

(Kontogiannis et al., 2002). On the other hand, driver stress confrontive behaviors. A study of a public transport authority
is addressed with the use of the Driver Behavior Inventory (Kuhlmann, 1990) found that bus and tram drivers reported a
(Matthews et al., 1999; Westerman and Haigney, 2000). The higher frequency of problem-oriented behaviors whilst a study
following sections provide an introduction on the psychometric of commuter drivers (Hennessy and Wiesenthal, 1997) found no
and theoretical underpinnings of driver stress, coping strategies differences in their respective frequencies.
and aberrant behaviors. Further work by Matthews et al. (1997) established a consis-
tent relationship between chronic stress (DBI scale) and coping
1.1. Driver stress and coping strategies behaviors. Drivers expressing anger, frustration and impatience
(i.e., the aggression scale) tended to react to stress through con-
The Driver Behavior Inventory (DBI) has been chosen as frontive coping. Drivers occupied by anxiety and fear (i.e., the
a measure of driver stress because it adopts a transactional dislike of driving scale) were prone to negative emotional coping
approach that views driver stress and performance as a func- that diverts attention from the driving task. The third dimension
tion of assessing traffic demands, appraising personal com- of alertness seemed to relate to a problem-focused coping, char-
petence and selecting strategies to cope with stress (Lazarus acteristic of active hazard monitoring. Alertness predicted speed
and Folkman, 1984). The DBI dimensions can be conceptual- of discrimination of roadside pedestrians in simulated driving
ized as behavioral dispositions related to the accumulation of (Matthews et al., 1998). These studies provide useful insights
stress in the driving context. In recent reviews of DBI appli- how to refine the distinction of problem-focused versus emotion-
cations (Matthews et al., 1991, 1997; Westerman and Haigney, regulation strategies into more elaborate types of coping skills.
2000), three major aspects of driver stress vulnerability emerged, Furthermore, it is necessary to examine how drivers respond to
namely: ‘aggression’ (i.e., anger, impatience, and risk taking), specific road scenarios rather than ‘general driving experience
‘dislike of driving’ (i.e., anxiety, self-blame and lack of confi- that is difficult and upsetting’.
dence) and ‘alertness’ (i.e., awareness of risk and active search
for road hazards). 1.2. Aberrant driving behaviors
Previous research suggested several refinements of driver
stress, as briefly presented here. The original study of DBI Aberrant behaviors have been extensively studied over the
(Gulian et al., 1989) suggested two additional forms of driver last 15 years across a range of different driver populations.
stress referred to as ‘irritation when overtaken’ and ‘frustra- Reason et al. (1990) developed the Driver Behavior Question-
tion when overtaking’. Lajunen and Summala (1995) recovered naire (DBQ) and administered it to a sample of 520 British
only the three largest factors (i.e., aggression, dislike of driving drivers. Factor analysis of the original 50-item questionnaire
and alertness) while the two overtaking factors loaded predomi- produced a three-factor structure comprising violations, mis-
nantly on aggression. Westerman and Haigney (2000) suggested takes (i.e., errors of judgement) and lapses (i.e., memory and
two new ‘situation-specific’ factors in a five-factor solution. attentional failures). A shorter version of the DBQ comprising
‘Situation-specific’ tension correlated with aggression (r = 0.39) 24 items administered to a larger sample (Parker et al., 1995)
and included two overtaking items whilst ‘situation-specific’ replicated the three-factor solution. Refinements of the DBQ
concentration correlated with alertness (r = 0.30) but consisted structure have been proposed by Aberg and Rimmo (1998) who
of only two items. Finally, Matthews et al. (1999) administered identified a four-factor solution in a sample of Swedish drivers
an extensive questionnaire which included a new ‘confidence whereby ‘lapses’ were split into ‘inattention’ and ‘inexperience’
factor’ that was positively related to perception of control. The errors.
‘confidence factor’ is further explored in this study as anecdotal There is some evidence to support a more fine-grained
evidence suggests that Greeks tend to overestimate their driving description of driving aberrations. An earlier study of a sam-
skills. An association between confidence and violations that ple of Greek drivers (N = 1400) used a 112-item inventory
lead to accidents has been established in a study of Australian (Kontogiannis et al., 2002) and produced a seven-factor solution.
drivers which showed that individuals with a high record of vio- Refinements of violations were suggested in terms of ‘aggres-
lations rated themselves higher in terms of confidence in vehicle sive or emotional’ violations, ‘highway-code’ violations and
control (Hartley and Hassani, 1994). ‘parking’ violations (i.e., arising from traffic constraints). The
Following the interpretation of a traffic situation as stress- structures of mistakes and lapses were similar to the previous
ful, drivers should select appropriate actions to cope with it. studies, but the other two classes of aberrant behaviors could
Although several classification schemes have been proposed not be understood in terms of the error – violation distinction.
for coping behaviors (Lazarus and Folkman, 1984; Newman One class of safety critical omissions referred to a state of low
and Beehr, 1979) two broad categories are ‘problem-focused’ preparedness (e.g., driving after having some drinks, or being
and ‘emotion-oriented’ coping. Drivers who focus on the prob- very tired) and a state of ‘negligence’ for maintaining safety fea-
lem appear to direct their efforts towards the task itself by tures of the car. The other class referred to ‘social disregard’ for
adopting rational strategies such as information seeking, taking other road users by failing to communicate intentions or creating
precautions, and making plans. In contrast, emotion-oriented inconvenient situations. Therefore, it is worth re-assessing the
drivers direct their efforts at regulating emotions and reducing DBQ structure of the Greek questionnaire although probably in
discomfort; they may become passive by expressing avoid- a more concise format as well as exploring how driver stress
ance and wishful thinking, or alternatively, they may resort to relates to several forms of errors and violations.
T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924 915

1.3. Aims of the study workplace and the road environment. For this reason, the focus
has been on the ‘average working person’ rather than the whole
The first aim of the present study was to investigate the factor population of Greece. Questionnaires were mailed to compa-
structure of driver stress in a Greek sample and its relationship nies (i.e., public sector services, banks, and private industries)
to coping strategies. The poor traffic environment may increase that were randomly selected from the ICAP directory of Greek
frustration on the roads contributing to high levels of stress. In companies. The data reported in this study were provided by
addition, Greek drivers may differ from other populations in respondents from eighteen cities all over Greece. From the 1500
their perception of self-confidence, alertness and coping skills. questionnaires that were sent out, approximately 714 question-
For instance, there is anecdotal evidence that Greeks tend to naires were completed and returned for further analysis. The
underestimate the risk of being involved in an accident and first section of the questionnaire contained general items con-
over-rate their driving skills. A high degree of confidence in cerning the driver’s age, gender, marital status, type of license
vehicle control may lessen driver stress but, on the other hand, held, speeding convictions, and number of accidents involved
may lead drivers to undertake driving maneuvers for which they since qualified as drivers. Information was also requested about
have inadequate skills. It is interesting therefore to examine how their profession, the amount of driving spent at work, their
differences in the traffic environment and socio-cultural back- annual mileage, and their driving experience (i.e., number of
ground across countries affect the factor structure of driver stress active years in driving since passing the test). Subsequent sec-
and its relationship to coping skills. tions included the items from the DBI that measure driver stress,
The second aim was to examine the nature of the association the items measuring coping behaviors, the items from the DBQ
between driver stress and errors. For routine and well-practiced measuring aberrant behaviors and measures of driver skills (i.e.,
tasks, performance decrements and lapses may relate to fail- perceptual-motor and safety-oriented skills).
ures of memory and attention because of the changing task
demands or other thought intrusions (Wickens, 1992; Reason, 2.1. Participants
1990). For resource-demanding tasks, stress may result in mis-
takes due to reduction of processing capacity (Kahneman, 1973), A third of the respondents were female drivers and another
or changes of attentional selectivity and narrowing of attentional third were single. With regard to their education level, 52%
focus (Stokes and Kite, 1994). On the other hand, the significant graduated from universities and technical colleges, 38% went
correlation may imply that errors could result in additional stress, to high school and 10% went to primary school. The jobs most
an argument consistent with the transactional stress approach. commonly reported were administration staff in the industry,
The relationship between stress and violations is less easy to technicians in the industry and clerks in the private sector.
predict. A positive relationship seems to be more likely on the The sample also included a small group of professional drivers
grounds that drivers high in stress are maladjusted to their envi- (N = 70) working in the industry. Other demographic charac-
ronment, hence conforming less often to the traffic rules. On teristics of the respondents are presented in Table 1. Although
the other hand, a negative relationship could be hypothesized participants were equally split into the five age bands, the sam-
since violations lead to a temporal reduction of workload and ple had a rather small percentage of drivers younger than 25
emotional release. Empirical evidence tends to support the for-
mer position since violations were found to be positively related
Table 1
to aggression (Matthews et al., 1997; Westerman and Haigney, Summary of demographic details of the study (N = 714)
2000). Given that stress and errors are multi-facetted, it is con-
ceivable that different components of driver stress may exert Percentage of respondents by age (in years)
and gender
different influences on forms of errors and violations. Relating
<29 29–34 35–40 41–46 >46
driver stress to the new forms of aberrant behaviors found in
Female 9.1 8.3 7.0 5.3 2.6
an earlier study of Greek drivers (Kontogiannis et al., 2002) Male 11.0 14.0 12.6 13.3 16.8
becomes an interesting theme for research. All 20.1 22.3 19.6 18.6 19.4
The third aim of this study was to investigate how driver
stress contributes to traffic accidents either directly or indirectly Percentage of respondents annual
mileage (×1000 km)
through the adoption of certain aberrant behaviors. A model of
accident causation is proposed in which distal factors (i.e., driver <6 6–10 11–20 21–30 31–50 >50
8.1 20.7 44.1 13.9 6.4 6.8
stress, demographic variables, and attitudes or beliefs) create a
disposition or general tendency in drivers to have high levels of Percentage of respondents by years
risky driving behaviors (i.e. proximal factors such as speeding, of active driving experience
errors and violations) which, in turn, predict the actual accident <6 6–10 11–20 21–30 >30
involvement. 14.9 22.7 37.2 20.7 4.5

Percentage of respondents number of


2. Methods
accidents (since qualification)
0 1 2 3 4 5 >5
The present study has been mainly concerned with traffic 42.4 27.9 18.5 5.7 2.5 1.8 12
campaigns that companies undertake to promote safety at their
916 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

years old (i.e., 7.6%) and older than 55 years old (i.e., 5.2%). et al. (1997): ‘confrontive coping’ (e.g., relieving one’s feel-
Because the sample consisted of company employees, certain ings through risk taking), ‘reappraisal’ (e.g., viewing drive as a
population groups (i.e., drivers younger than 25 years) were learning experience), ‘emotion-focus’ (e.g., criticizing oneself
under-represented or, in other cases, not represented at all (e.g., for mistakes), ‘avoidance’ (e.g., trying to suppress negative feel-
the unemployed). This sample would be representative of the ings) and ‘task-focus’ (e.g., making an effort to drive safely). For
‘average working person’ but not of the Greek population. each road scenario, participants indicated on a five-point Likert
scale (that was similar to driver stress) how often they resorted
2.2. Measures of driver stress to each of the coping strategies specified. Participants were also
asked to rate the amount of frustration and irritation generated
The Driving Behavior Inventory (Matthews et al., 1999) was by each road scenario on a five-point Likert scale (0 = not at all,
translated into Greek and adapted to the Greek culture. Most 1 = a little, 2 = some, 3 = much, 4 = very much).
of the items loading on aggression, alertness, confidence and
dislike of driving have been included in the questionnaire. The 2.4. Measures of aberrant driving behaviors
‘overtaking tension’ (Matthews et al., 1999) consisted mainly
of items related to ‘irritation when overtaken’ and ‘frustration The present study used a reduced version of the DBQ ques-
in overtaking’ which, however, were found to load predomi- tionnaire administered by Kontogiannis et al. (2002) in an earlier
nantly on aggression in a previous study (Lajunen and Summala, sample of 1400 Greek drivers. Approximately half of the 70
1995). For this reason, ‘overtaking tension’ was excluded from items of the earlier study were retained in the reduced version,
the present questionnaire. Participants were asked to indicate on representing violations (i.e., high-code and aggressive viola-
a five-point Likert scale (0 = never, 1 = seldom, 2 = sometimes, tions), errors (i.e., mistakes and lapses), social disregard and
3 = frequently, 4 = almost continually) how often they had the negligence. The present questionnaire did not contain any items
feelings and dispositions described in an inventory of 19 DBI related to ‘parking’ violations but included seven new items,
items. forming an inventory of 44 items. Participants were asked to
indicate how often they committed these behaviors in a Likert
2.3. Measures of coping behaviors scale similar to one used in driver stress.

Two road scenarios were used as follows: (i) ‘how would you 2.5. Measures of driver skills
react in a situation where someone in front of your car is mov-
ing very slowly and this makes it difficult for you to overtake Participants were also asked to rate their own level of driving
?’ (the slow-driving scenario) and (ii) ‘how would you react skill on a reduced version of the Driver Skill Inventory (Lajunen
in a situation where someone is driving very close to your rear and Summala, 1995) distinguishing between perceptual-motor
bumper and beeps at you because s/he cannot overtake you ?’ skills (seven items) and safety-orientation (seven items). Safety-
(the discourtesy scenario). An inventory of five coping strate- oriented skills are interesting to examine because they may
gies comprising 18 items was developed on the basis of the relate to certain driver stress vulnerabilities and aberrant behav-
transactional model of stress and a previous study by Matthews iors which otherwise fail to show any negative relationships to

Table 2
Driver stress items forming four factors (DBI) using oblique rotation: aggression (F1), alertness (F2), dislike of driving (F3) and confidence (F4)
No. Description of DBI items F1 F2 F3 F4

1 I become irritated during rush hour 0.77


2 It annoys me to drive behind slow-moving vehicles 0.77
3 When I try, but fail to overtake, I am usually frustrated 0.71
4 In general, I mind being overtaken 0.68
5 When irritated, I drive aggressively paying no proper attention 0.64
6 I tend to overtake vehicles whenever possible 0.58
7 I become tense and concentrate on heavy traffic 0.78
8 I usually make an effort to look for potential hazards 0.78
9 Accidents can always happen due to other’s poor judgement 0.71
10 I increase concentration on unfamiliar roads to be on alert 0.67 0.31
11 In general, I do not enjoy driving 0.85
12 Driving usually makes me feel frustrated 0.81
13 Driving usually does not make me happy 0.79
14 In general, too much driving is a waste of time 0.67
15 It worries me to drive in bad weather or heavy traffic 0.39 0.38 0.46 −0.30
16 I feel confident in my ability to avoid an accident 0.80
17 I feel in command of the situation when overtaking 0.77
18 I find it easy to control temper when driving −0.30 0.70
19 I am always ready to react to others’ unexpected manoeuvres 0.51 0.65
T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924 917

accidents rates. Participants assessed their abilities relative to a a feeling of anxiety related to a difficult driving condition. Factor
hypothetical ‘average Greek driver’ on a five point Likert scale 3 is labeled ‘dislike of driving’ and includes feelings of anxiety,
(from 0 = much worse to 4 = much better). tension and negative appraisals of driving. Factor 4 resembles
the ‘confidence’ factor of the Japanese and Australian samples
3. Results (Matthews et al., 1999; Hartley and Hassani, 1994).
Following the factor analysis, factor scores were computed
3.1. Dimensions of driver stress for each participant by calculating the means of the marker
items that comprised each of the corresponding factor (all
The factor structure of the DBI was examined using prin- alpha reliability coefficients were approximately 0.75). Item
cipal components analysis with oblique rotation since Gulian 15 was not included in the dislike scale because it loaded
et al. (1989) and Glendon et al. (1993) assumed that the DBI high on all other scales. The most frequent stress driver fac-
dimensions were related. The scree test suggested that four fac- tors were confidence (mean = 3.44, S.D. = 0.56) and alertness
tors should be extracted. All factor intercorrelations were less (mean = 3.32, S.D. = 0.71) followed by aggression (mean = 1.43,
than 0.2 with the exception of correlation between factors 2 and S.D. = 0.73) and dislike of driving (mean = 1.14, S.D. = 0.89).
4 that reached 0.32. Because of the relative weak correlations, Because the distributions of confidence and alertness were neg-
a varimax rotation produced similar results and accounted for atively skewed, a square transformation was applied to both
56.3% of the variance. factors.
Table 2 shows the principal loadings on the four factors in the
oblique solution that was finally retained for further analysis (all 3.2. Dimensions of coping strategies
eigenvalues were greater than 1.5). In the oblique solution, the
first factor contributed 20.2% to the total variance, the second The factor structure of coping behaviors was examined sep-
factor contributed 16.7%, the third 11.5% and the fourth 7.9 %. arately for the two road scenarios, using principal components
Factor 1 is equivalent to ‘aggression’ of other studies and analysis with varimax rotation. The scree tests for the two road
relates to feelings of anger, hostility and impatience. Factor scenarios suggested that five factors should be extracted. For
2 is labeled ‘alertness’ and relates to hazard monitoring and both scenarios, it appeared that seven factor correlations were
increased concentration; however, one item (#7) also indicated weak while three correlations were moderate 0.25–0.50. A sec-

Table 3
Coping behavior items forming five orthogonal factors; slow-driving and discourtesy scenarios are in bold and italics fonts
Description of coping behavior items F1 F2 F3 F4 F5

Factor 1: confrontive coping


I get upset and drive very close to his or her rear bumper 0.82
I become impatient and flash my light beams on other drivers 0.76
I get upset and start overtaking to show my determination to get on 0.74
I become impatient and beep at the other driver 0.81 0.74
I get upset, lose my temper and make obscene gestures 0.72 0.70
I get upset and slow down to teach him or her a lesson 0.82
I get upset and find ways to make the other car abort its effort 0.79
Factor 2: re-appraisal
Remind myself how much worse things could be 0.82 0.84
I try to relax and tell myself things that help me feel better 0.79 0.81 0.25 0.32
I watch out that the same situation does not happen again 0.77 0.74
I look on the bright side of the situation and maintain my temper 0.69 0.77 0.30 0.30
Factor 3: self-criticism
I feel hopeless and think that all my attempts would fail 0.81 0.84
I criticize or blame myself for the situation 0.69 0.72
I tell myself I won’t make it because I am not a good driver 0.69 0.69
Factor 4: time facilitation (supression)
I do something (smoke, listen to music) to pass my time 0.28 0.31 0.77 0.79
I am day dreaming to take my mind off the situation 0.35 0.35 0.75 0.78
Factor 5: task-focus
I wait until the road becomes wider or the car ahead changes lane 0.79
I try to avoid creating a risky overtaking scenario 0.35 0.48 0.62
I make a plan to escape the situation (e.g., change route) 0.69 0.58
I try to find the right time to move to the right or change lanes −0.30 0.65
I drive faster to avoid holding up the car behind me 0.31 0.59
Variables excluded from the analysis (factor loadings <0.4)
I wish somehow the other car changes direction or pull over 0.29 0.28 0.33 0.55 0.33 0.26
918 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

ond analysis with oblique rotation did not show any significant tor 2 includes attempts to learn from the past and attempts to
changes and the varimax solution was retained for further anal- feel better and maintain temper (‘reappraisal’). Factor 3 reflects
ysis. Table 3 shows the principal loadings on the five factors in resignation from the task and blaming oneself for the situation
the varimax solution for the two road scenarios separately. The (‘self-criticism’). Factor 4 refers to activities aimed at facilitat-
five factors explained approximately 59% of the total variance ing the situation rather than avoidance, suggesting that ‘time
in each scenario. Factor contributions to the total variance were facilitation’ could be the appropriate term. Finally, factor 5
(19.3%, 17.9%, 8.9%, 6.6%, and 6.2%) for the slow-driving includes attempts at developing plans to cope actively or escape
scenario and (16.3%, 22.2%, 8.3%, 5.2%, and 6.9%) for the the situation and restraining one’s self from giving rise to RTA
discourtesy scenario. (‘task-focus’). The item related to ‘wishful thinking’ loaded on
Items loading high on factor 1 express confrontation and different scales in the two scenarios and was excluded from the
relief of anger through risk taking (‘confronting coping’). Fac- analysis.

Table 4
Aberrant behavior items forming five factors (DBQ scale) using oblique rotation: social disregard (F1), lapses (F2) and negligence (F3), violations (F4), and mistakes
(F5)
Description of DBQ items F1 F2 F3 F4 F5 Mean Referencea

1 Fail to drive slowly at ‘right turn’ signs 0.59 0.67 NEW


2 Temporarily park car in a way that obstructs traffic 0.50 0.55 NEW
3 Forget to put the hazards lights on, when pulling over 0.55 0.43 SD
4 Forget to signal when overtaking 0.52 0.30 0.25 0.76 SD
5 Forget to signal when changing lanes 0.50 0.26 0.50 SD
6 Fail to give way to pedestrians ready to cross 0.40 0.75 NEW
7 Pull out from a parking place without signalling 0.39 0.66 NEW
8 Shift into the wrong gear at traffic lights 0.63 0.35 LAP
9 Wrong gear when reversing 0.60 0.21 LAP
10 Forget what gear is in and check with hand 0.52 0.41 LAP
11 Suddenly realize obstacle on the road 0.43 0.28 1.06 MIS
12 Forget lights on full beam, dazzling other drivers 0.35 0.36 SD
13 Drive with a beam lamp out of order 0.89 0.35 NEG
14 Drive with signal lamps out of order 0.88 0.29 NEG
15 Drive with brake lamps out of order 0.86 0.36 NEG
16 Drive with icy and misty windows 0.36 0.39 NEG
17 Drive without wheel chains on an icy road 0.33 0.32 NEG
18 Showing hostility to other drivers 0.64 1.00 AGV
19 Giving chase to other drivers 0.63 0.33 AGV
20 Tail gaiting to force others to give way 0.58 0.39 AGV
21 Become impatient and overtake on the inside 0.56 0.72 AGV
22 Overtake slow vehicle in a risky manner 0.51 0.68 AGV
23 Accelerating when lights turn green to impress 0.48 0.37 AGV
24 Cross solid line when changing lane 0.46 0.26 1.60 HCV
25 Exceed speed limit during low traffic 0.46 0.35 1.64 HCV
26 Overtake expecting others to drive in their right lane 0.32 0.44 0.55 AGV
27 Overtake even when close to turning 0.41 0.29 AGV
28 Close following, ignoring safe gap 0.31 0.40 0.70 AGV
29 Cross on lights that have just turned red 0.30 0.36 0.67 HCV
30 Drive fast even when visibility is poor 0.35 0.31 NEW
31 Drive after having some drinks 0.27 0.33 0.69 NEG
32 Not noticing a ’give way’ sign behind trees 0.64 1.09 MIS
33 Fail to look for ’stop’ signs in a crossroad 0.54 0.28 NEW
34 Miss road exit because failed to change lanes 0.26 0.45 0.89 MIS
35 On turning right nearly hit a cyclist 0.43 0.34 MIS
36 Stop longer than expected when braking 0.30 0.42 0.70 MIS
37 Wrong lane at crossing 0.42 0.25 MIS
38 Fail to see pedestrian stepping out 0.30 0.35 0.81 MIS
39 Realize late that vehicle in front has slowed down 0.29 0.34 0.76 MIS
Variables excluded from analysis (loadings <0.33)
40 Near to catch a nap during extensive driving 0.31 0.46 NEG
41 Fail to give priority to vehicles on the right 0.32 0.30 0.51 NEW
42 Manoeuvre without checking mirror 0.28 0.29 SD
43 Disregard red lights at night 0.29 0.31 0.37 HCV
44 Not wear safety belt in urban areas 0.27 2.02 NEG
aKontogiannis et al. (2002); AGV: aggressive violations; HCV: highway code violations; MIS: mistakes; LAP: lapses; SD: social disregard; NEG: negligence;
NEW: items added in present study.
T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924 919

In the two road scenarios, alpha coefficients were above 0.75 Following the factor analysis, factor scores were computed
for confrontive coping, reappraisal and time facilitation. How- for each participant by calculating the means of the marker items
ever, alpha coefficients for task-focus coping were rather low that comprised each of the corresponding factor. Alpha coeffi-
(0.50) for both scenarios. Self-criticism had alpha coefficients cients for negligence and violations were approximately 0.83
0.60 and 0.67 for the slow-driving and discourtesy scenarios. whilst for social disregard and mistakes were 0.68. The alpha
Factor scores were computed, for each participant, by calcu- coefficient for lapses was rather weak, that is 0.60. Lapses, social
lating the mean of the marker items that comprised each of disregard and negligence were positively skewed and a square
the corresponding factor. For the slow-driving scenario, the root transformation was applied.
most frequent coping behaviors were task-focus (mean = 2.63,
S.D. = 0.85) and reappraisal (mean = 2.27, S.D. = 0.99), fol- 3.4. Correlations of driver stress, coping behaviors and
lowed by time facilitation (mean = 1.78, S.D. = 1.08). The aberrant behaviors
least frequent behaviors were confrontive coping (mean = 0.96,
S.D. = 0.74) and self-criticism (mean = 0.45, S.D. = 0.58). A Table 5 shows gender and age differences in driver stress and
similar pattern emerged for the discourtesy scenario where cop- coping behaviors. Two-tailed t-tests were used to examine the
ing behaviors could be ranked as follows in terms of frequen- effects of gender (females coded as 0 and males as 1). Zero-
cies of use: task-focus (mean = 2.46, S.D. = 0.76), reappraisal order correlations and partial correlations were used for age,
(mean = 1.99, S.D. = 1.13), time facilitation (mean = 1.50, controlling for the effects of driving experience. Gender seemed
S.D. = 1.14), confrontive coping (mean = 0.45, S.D. = 0.61) and to affect confidence and alertness with men feeling more confi-
self-criticism (mean = 0.35, S.D. = 0.56). dent and alert than women. Age influenced all aspects of driver
stress vulnerability but its effects were confounded by driving
3.3. Dimensions of aberrant behaviors experience. Confidence and alertness increased with age but this
effect could equally be due to driving experience. Aggression
Principal components analysis with oblique rotation was used was decreased with age even when controlling for the effects of
because dimensions of aberrant behaviors were found to be driving experience. Finally, dislike of driving increased with age
highly correlated (Kontogiannis et al., 2002). The scree plot (r = 0.16, partialling out experience) and decreased with driving
suggested a five factor solution which had rather strong inter- experience (r = −0.12, partialling out age).
correlations ranging from 0.25 to 0.53. Factor contributions to With regard to coping behaviors, gender differences were
the total variance were as follows: 19.6%, 6.5%, 5.1%, 3.5% rather weak. Women scored higher than men in reappraisal
and 3.2% for each factor, respectively. Table 4 shows the prin- for the slow-driving scenario and time-facilitation for the dis-
cipal loadings on the five factors (all eigenvalues were greater courtesy scenario. Women also tended to score higher in self-
than 1.4) together with their mean values while the last column criticism but this tendency just failed significance for both road
presents the classification of each item according to an earlier
study (Kontogiannis et al., 2002).
The first factor is identified as ‘social disregard’ to other road Table 5
users by failing to communicate intentions. The second factor is Gender and age differences in driver stress and coping behaviors
identified as ‘lapses’ (i.e., failures of memory and attention) and Driver stress Gender, Age, zero-order Age, controlled for
the third factor is identified as ‘negligence’ (i.e., maintain sev- and coping t-values correlations experience, partial
eral safety features of ones’ own car). The fourth factor includes correlations
‘ordinary’ violations associated with circumventions of the high- Driver stress
way code and ‘aggressive’ violations associated with venting out Aggression −0.55 −0.13** −0.10*
emotions and showing hostility. The fifth factor is identified as Dislike of driving 0.69 0.10* 0.16***
Confidence 5.33*** 0.19*** 0.05
‘mistakes’ referring to failures of judgment and regulation of Alertness 3.49** 0.15*** 0.05
attention. Finally, the last five items failed to satisfy the crite-
Coping scale (slow-driving scenario)
rion (i.e., their loadings were lower than 0.33) and were excluded
Confrontive coping 2.48** −0.03 −0.10*
from further analysis. Reappraisal −2.05* 0.19*** 0.18***
In the present study, the social disregard factor included Self-criticism −1.83 0.04 0.15***
items that indicated, not only failure to communicate intentions, Task-focus 1.60 0.12** 0.10*
but also ‘indifference’ causing inconvenience and embarrass- Time facilitation −1.69 0.03 0.02
ment to pedestrians and other drivers (see items #1, #2, #6 Coping scale (discourtesy scenario)
and #7 in Table 4). On the other hand, the negligence factor Confrontive coping −1.82 −0.13** −0.00
was more constrained in the present study referring only to Reappraisal −0.41 0.25*** 0.19***
Self-criticism −1.77 0.04 0.13**
omissions in keeping safety features of the car in good condi- Task-focus 1.41 0.17*** 0.14***
tion. The previous study (Kontogiannis et al., 2002) included Time facilitation −2.34* 0.10* 0.10*
aspects of negligence in keeping oneself fit for driving (see
Females coded as 0 and males as 1.
items, #31, #40 and #44) that are associated with a low state * p < 0.05.
of driver preparation about the possibility of causing a traffic ** p < 0.01.

accident. *** p < 0.001.


920 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

scenarios (P = 0.07–0.08). On the other hand, men scored higher Table 7


in confrontive coping only for the slow-driving scenario. For Partial correlations between driver stress and aberrant driving (controlling for
gender, age, driving experience and mileage collectively)
both road scenarios, reappraisal, task-focus and self-criticism
seemed to increase with age (partialling out experience, Table 5). Aberrant Driver stress vulnerability (DBI scale)
The results also showed that self-criticism seemed to decrease driving (DBQ)
Aggression Dislike Confidence Alertness
with driving experience (r = −0.13 to 0.15, partialling out age)
Violations 0.62*** 0.02 −0.19*** −0.05
for the discourtesy and slow-driving scenarios. Mistakes 0.26*** 0.19*** −0.28*** −0.08
A similar analysis was done for aberrant behaviors which Lapses 0.24*** 0.15** −0.26*** −0.01
showed that men reported more violations (t = 4.45, P < 0.001) Social disregard 0.31*** 0.14** −0.16** −0.10*
and social disregard (t = 3.13, P < 0.01) while women reported Negligence 0.23*** 0.16** −0.23*** −0.13*
higher frequencies of lapses (t = −5.06, P < 0.001). Viola- * p < 0.05.
tions and social disregard seemed to decrease with age ** p < 0.01.
(r = −0.20, P < 0.001; r = −0.16, P < 0.001) but increased as *** p < 0.001.
drivers acquired more expertise (r = 0.16, P < 0.001; r = 0.14,
P < 0.01). On the other hand, driving experience had a benefi- more errors and, surprisingly, greater scores in social disregard
cial effect in reducing lapses (r = −0.11, P < 0.01). and negligence.
Partial correlations between driver stress and coping can
be seen in Table 6, controlling for the effects of gender, age, 3.5. Patterns of coping as a function of frustration
driving experience and annual mileage collectively. Aggression experienced
was positively related to confrontive coping but surprisingly
it was also positively related to self-criticism. Confidence was In a public transport study, Kuhlmann (1990) found that
positively related to reappraisal and negatively related to self- professional drivers had a constellation of different strategies
criticism. For the slow-driving scenario, confidence was corre- comprising elements from both problem-focused and emotion-
lated with task-focus as well. Alertness had a stronger positive regulation strategies. This constellation of coping strategies
correlation with task-focus for both road scenarios and a pos- may change as a function of the level of frustration and anger
itive correlation with reappraisal in the slow-driving scenario. experienced in several road scenarios. An investigation exam-
Finally, dislike of driving was positively related to self-criticism ined whether some coping strategies were used more frequently
as expected but it also seemed to be positively related to con- depending on the intensity of frustration felt in the slow-driving
frontive coping for the discourtesy scenario. and discourtesy scenarios. For this reason, three levels of frustra-
Partial correlations between driver stress and aberrant behav- tion were examined on the basis of the 33rd and 67th percentiles
iors (Table 7) showed that aggression appeared to be positively of the frustration and anger felt. The reader is reminded that par-
related to all types of aberrant behavior and predominantly to ticipants were also asked to rate the level of frustration felt at
violations. It is worth noting that confidence was negatively the slow-driving and discourtesy scenarios. A two-way ANOVA
related to all types of aberrant behaviors but alertness reached was performed for each road scenario with ‘type of coping’ (i.e.,
a moderate relationship to social disregard and negligence only. five levels) as a within-subjects variable and ‘frustration level’
On the other hand, people high in dislike of driving reported (i.e., three levels) as a between-subjects variable.
For the slow-driving scenario, there were significant main
effects of type of coping (F(1,701) = 20.1, P < 0.001) and frus-
Table 6
tration level (F(3,701) = 7.9, P < 0.001). However, comparisons
Partial correlations between driver stress and coping behaviors (controlling for
gender, age, driving experience and mileage collectively) were made in terms of simple effects because the interaction was
significant (F(3,701) = 7.9, P < 0.001). For the slow-driving sce-
Coping behaviors Driver stress vulnerability (DBI scale)
nario, a post hoc Tukey test found that drivers who experienced
Aggression Dislike Confidence Alertness high levels of frustration resorted more frequently to confrontive
Slow-driving scenario coping than drivers who experienced medium and low levels of
Confrontive coping 0.56*** 0.07 −0.06 0.03 frustration. For the discourtesy scenario, there were no signifi-
Reappraisal −0.13** −0.03 0.16*** 0.11* cant main effects of frustration level on coping strategies.
Self-criticism 0.25*** 0.26 *** −0.20*** −0.02
Task-focus 0.07 −0.01 0.14** 0.22***
3.6. Self-perception of motor skills and safety orientation
Time facilitation 0.02 0.01 0.037 0.04
Discourtesy scenario To investigate the effects of driver stress on the perception of
Confrontive coping 0.35*** 0.13** −0.07 −0.01
Reappraisal −0.15*** 0.01 0.15*** 0.06
perceptual-motor skills and safety-oriented skills, hierarchical
Self-criticism 0.18*** 0.31*** −0.16*** −0.04 regression analyses were performed. An interesting result was
Task-focus 0.07 0.02 0.07 0.20*** that driver stress accounted for the greatest variance (Change
Time facilitation 0.01 0.01 0.03 −0.06 in R2 = 0.33) in safety skills. Confidence, alertness and dislike
* p < 0.05. of driving were positively related to safety-orientation while
** p < 0.01. aggression was negatively related to safety-orientation. In a sim-
*** p < 0.001. ilar regression analysis for perceptual-motor skills, driver stress
T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924 921

Fig. 1. Model of relationships (std B coefficients) between age, gender, mileage, driving experience, aggression and violations. For the sake of clarity, only latent
variables are presented.

again accounted for the greatest variance (Change in R2 = 0.19) P = 0.33; RMSEA = 0.017; confidence interval for RMSEA
mainly because of the strong positive correlation with confidence 0.00–0.06; P-value for test of close of fit [RMSEA < .05] = 0.88).
and the modest negative correlation with dislike of driving. According to the structural model, driving experience had the
most profound effect on accident rates followed by violations
3.7. Predictors of accident involvement and mileage. Aggression had an indirect influence on accident
rates through violations but an insignificant direct influence.
Another aim of this study was to investigate a contextual Although violations could also be seen as a causal factor of driver
model of accident involvement with driver stress acting as distal stress (i.e., aggression), the data in this study seem to suggest
contextual factors and aberrant behaviors as proximal factors. that violations should be rather seen as the result of accumulated
Initially, two hierarchical regression analyses were performed aggression and frustration in the explanation of accident rates. A
on annualized accident rates (i.e., the total number of accidents large portion of variance in violations (R2 = 0.93) was explained
divided by the driving experience) separately for driver stress by age, gender, experience and predominantly by aggression.
and aberrant behaviors. Driving experience had a strong nega- In the present study, approximately 20% of the sample of
tive impact upon accident rates and accidents increased with the Greek drivers had previously convicted for speeding. A logis-
number of mileage covered each year. Violations and aggression tic regression was performed where gender, age, experience,
were good predictors of accident rates in the two models. mileage and driver stress vulnerability were entered in a sin-
The results of the regression analyses have been used to gle step. Overall, three quarters of the responses were pre-
construct a structural model of accident rates using LISREL dicted (x2 = 65.9, df = 8, P < 0.0001). Significant background
8 (Joreskog and Sorbom, 1993). First, age, gender, driving predictors were gender (Wald statistic = 19.9, P < 0.001) and
experience and annual mileage were entered as exogenous vari- annual mileage (Wald statistic = 5.7, P < 0.05). Speeding convic-
ables vis-à-vis violations, aggression and accident rate. Second, tions were positively related to aggression (Wald statistic = 14.2,
aggression was entered as a disposition variable explaining vio- P < 0.001) and negatively related to dislike of driving (Wald
lations. Finally, violations and aggression were entered as vari- statistic = 12.8, P < 0.001).
ables explaining accident rates. One measurement variable was
attached to each latent variance which, however, is not shown 4. Discussion of results
in Fig. 1 for the sake of clarity. For the DBI and DBQ scales,
reliability estimates were coefficient alpha found in this study. 4.1. Factor structures of measuring scales
Fig. 1 shows the structural model of accident rates expressed
in the form of standardized path coefficients and variances In general, the findings of this study seem to support the
explained (R2 ) for aggression, violations and accident rates. The factor structures of the measuring scales of driver stress, coping
model meets an exact fit criterion to the data (x2 = 8,02; df = 7; strategies and aberrant behaviors.
922 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

The main factor structure of driver stress in terms of aggres- have negative feelings towards driving as implied by the posi-
sion, dislike of driving and alertness was also obtained in the tive correlation between negligence and dislike of driving. With
present study. Confidence was an additional dimension of driver regard to social disregard, impatience and aggression was asso-
stress that comprised items similar to the ones found in Matthews ciated with actions exhibiting disregard for other road users and
et al. (1999) and Hartley and Hassani (1994). Also the structure pedestrians. However, the modest positive correlation between
of coping behaviors had many similarities with the one found social disregard and dislike of driving was difficult to predict.
in Matthews et al. (1997) although their ‘avoidance’ scale was When dislike of driving was individually correlated with all
rather different from the ‘time facilitation’ scale found in the items of social disregard it appeared that the positive correlation
present study. was held significant only for the dimension of communication
The pattern of age correlations in driver stress was somewhat errors. No significant correlations were found between dislike of
different to those previously reported. There was good agree- driving and the dimension of ‘indifference’ or ‘inconvenience’
ment that aggression decreased with age while dislike of driving of social disregard.
increased as expected. An interesting finding was that alertness
increased with age, pointing out the supportive role of strate- 4.3. The relationship between stress and accident
gic behavior as drivers acquire more experience. This contrasts involvement
the finding of Westerman and Haigney (2000) that alertness is
decreased with age due to the age-related decline in cognitive Dimensions of driver stress differed in terms of safety-
capacity. Hence, strategy and capacity may be two factors with orientation, accident involvement and speeding convictions.
contrasting effects on alertness. Drivers high in aggression were more likely to be caught over-
The factor structure of aberrant behaviors was again fairly speeding and had more accidents whilst drivers high in dislike
similar to the one found in an earlier sample of Greek drivers of driving may try to reduce the demands of driving by slow-
(Kontogiannis et al., 2002). The present study used a reduced ing down as indicated by their negative reaction to speeding.
version of the previous questionnaire and, as a result, aggressive Alertness has been found to contribute to safe driving although
violations seemed to merge with highway-code violations as a no direct link was established with accident rates or speeding
single factor. The main difference between the two studies was convictions. The positive correlation between confidence and
in the structure of social disregard and negligence factors. safe driving was less easy to interpret in terms of causality. It is
likely that drivers who are confident in responding to unexpected
4.2. The relationship between stress, coping patterns and events and control their temper, may drive safely. The reverse
aberrant behaviors is also plausible, that is, drivers who emphasize safe driving
may expose themselves to fewer risks and have less accidents
The pattern of correlations between driver stress and coping which may increase their confidence. This could explain the fact
was somewhat different from the one found in Matthews et al. that no direct link was found between confidence and accident
(1997). There was good agreement that drivers high in aggres- rates.
sion were more likely to engage in confrontive coping and that The different correlations of the confidence factor in com-
dislike of driving was associated with self-criticism as expected. parison to other studies deserve further attention. In Hartley and
However, the positive correlation found between aggression and Hassani (1994), confidence was much higher in young drivers
self-criticism may suggest that aggressive drivers could be aware while there was a positive relation to accident rates. In Matthews
of their behavior but unable to take any preventive actions. This et al. (1999), confidence was not characteristic of young drivers
may have implications for traffic safety campaigns aiming at but it was strongly related to speeding convictions. In the present
reducing aggression on the roads. On the other hand, the negative study, confidence was characteristic of older drivers. Although
correlation between alertness and confrontive coping reported high confidence was related to a higher perception of motor
in Matthews et al. (1997) was not replicated in the present study. competence and lower perception of risk, no correlations were
This may imply that even alert drivers, who generally focus on found with accident rates nor speeding convictions. Equally
reappraisal and task coping, may become confrontive in certain well, drivers high in confidence reported fewer mistakes and
road situations. violations. In general, this study has not found any indications
The results also indicated that high stress was associated with that high confidence relates to risky driving and higher risk of
higher levels of errors and violations. Specifically, aggression accidents.
and dislike of driving were positively correlated to mistakes This finding could be attributed to the demographic variables
and lapses whilst confidence was negatively correlated to these of this sample, the method of measuring confidence and the cul-
errors. On the other hand, aggression was positively related to tural background of the driving population. First, the present
violations whilst confidence was negatively related to violations. sample had a lower percentage of young and old drivers in
More difficult to predict was the association between stress and comparison to other studies. It is likely that greater sampling
social disregard or negligence. of young and inexperienced drivers may have produced higher
Driver stress (indexed as aggression) seems to influence perceptions of confidence and higher accidents, especially for
drivers who neglect safety-features of their cars probably the 19–25 yrs group of age. Second, measuring confidence on
because their behavior is characterized by a sense of impatience a Likert scale may be less effective than using a 100 mm visual
and time urgency. Some drivers high in negligence also seem to analogue scale similar to one employed by earlier studies of
T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924 923

DBI. Finally, cross-cultural differences may explain the correla- disregard may lead to automaticity. Hence, drivers may get into
tion pattern of the confidence scale. The high confidence values these unsafe habits without realizing the long-term consequence
obtained in this sample may be explained by the fact that the in terms of near-miss and traffic accidents.
Greek culture is optimistic and self-reassuring, hence encour- There are two practical implications from this study in terms
aging people not to undervalue their skills. This may work well of defensive driving and stress-management training. First, some
for safety-oriented drivers but may create some problems for kind of re-fresher training may be needed addressing aspects of
drivers who feel confident to undertake difficult maneuvers or preparedness, alertness and driver communication since both
take risks. Further research should look at finer measures of violations and social disregard increased with experience and
risk perception and self-confidence in order to gain a deeper exposure to traffic (see Section 3.4). Second, aberrant behav-
insight. iors are likely to be fostered by driver stress as indexed mainly
by aggression (see Fig. 1). As the traffic environment becomes
5. Concluding remarks an overloaded system with a high frequency of frustrations, an
attractive counter-measure is the training of drivers in tolerating
The study reported here has some methodological limitations. road frustration and managing aggression.
First, most of the participants were company employees who
may have been particularly interested in traffic safety. This fact,
in joint with the relatively low response rate, may explain the References
high rates of alertness and confidence obtained in this study.
Second, there are well-known methodological difficulties with Aberg, L., Rimmo, P.A., 1998. Dimensions of aberrant driver behavior.
Ergonomics 41, 39–56.
self-reports of behavior and emotions such as reporting biases Chliaoutakis, J.E., Darviri, C., Demakakos, P., 1999. The impact of young
and difficulties in controlling for potential confounds (Hale and drivers’ lifestyle on their road traffic accident risk in the greater Athens
Glendon, 1987). Finally, coping behaviors have been preselected area. Accident Anal. Prev. 31, 771–780.
and this could have prevented respondents from identifying Hale, A.R., Glendon, A.I., 1987. Individual Behavior in the Control of Dan-
aspects of copings that were omitted in this research. A better ger. Elsevier, Amsterdam.
Hartley, L.R., Hassani, J.E., 1994. Stress, violations and accidents. Appl.
understanding of coping may have been possible by allowing Ergon. 25, 221–230.
open-ended answers. Hennessy, D.A., Wiesenthal, D.L., 1997. The relationship between traffic
In conclusion this study has found that, as it was the case congestion, driver stress and direct versus indirect coping behaviors.
with other countries, driver stress in the form of aggression was Ergonomics 40, 348–361.
associated with unsafe behaviors and traffic accidents. The data Glendon, A.I., Dorn, L., Matthews, G., Gulian, E., Davies, D.R., Debney,
L.M., 1993. Reliability of the driving behavior inventory. Ergonomics 36,
showed that the factor structure of the Greek version of DBI was 719–726.
broadly similar to those found in other populations. It may be Gulian, E., Matthews, G., Glendon, A.I., Davies, D.R., Debney, L.M., 1989.
worthwhile for companies to screen drivers for aggression since Dimensions of driver stress. Ergonomics 32, 585–602.
this sort of stress vulnerability may fuel even higher levels of Joreskog, K.G., Sorbom, D., 1993. LISREL 8:Structural Equation Modeling
frustration and anger on the roads. Future work in Greek samples with the SIMPLIS Command Language. Scientific Software International,
Chicago.
may explore more elaborate measures of confidence and self- Kahneman, D., 1973. Attention and Effort. Prentice Hall, Englewood Cliffs,
control and relate them to accident and conviction rates. The NJ.
analysis of coping strategies indicated that individuals activate Kontogiannis, T., Kossiavelou, Z., Marmaras, N., 2002. Self-reports of aber-
a variety of coping strategies during a stress scenario rather than rant behavior on the roads: errors and violations in a sample of Greek
a single strategy. Future work is therefore needed to understand drivers. Accident Anal. Prev. 34, 391–399.
Kuhlmann, T.M., 1990. Coping with occupational stress among urban bus
these combinations of coping strategies and the way that these and tram drivers. J. Occup. Psychol. 63, 89–96.
support or hinder each other. Lajunen, T.L., Summala, H., 1995. Driving experience, personality, and skill
Understanding the cognitive, attitudinal and social context of and safety-motive dimensions in driver’s self-assessments. Pers. Indiv.
driving errors is very important for devising appropriate counter- Differ. 19, 307–318.
measures to reduce accident rates. The present study has tested Lazarus, R.S., Folkman, S., 1984. Stress, Appraisal and Coping. Springer,
New York.
and confirmed the distinction between errors and violations in Matthews, G., Dorn, L., Glendon, A.I., 1991. Personality correlates of driver
the Greek transportation environment where attitudes, alertness stress. Pers. Indiv. Differ. 12, 535–549.
and preparedness differ from other European and North Amer- Matthews, G., Desmond, P.A., Joyner, L., Carcacy, B., Gilliland, K., 1997.
ican countries. New forms of aberrant behaviors have emerged A comprehensible questionnaire measure of driver stress and affect. In:
in this study referring to a mental state of low preparedness Rothengatter, T., Vaya, E.C. (Eds.), Traffic and Transport Psychology:
Theory and Application. Pergammon, Amsterdam.
regarding potential safeguards against errors (i.e., negligence) Matthews, G., Dorn, L., Hoyes, T.W., Davies, D.R., Glendon, A.I., Taylor,
and communication errors with other road users (i.e., social dis- R.G., 1998. Driver stress and performance on a driving simulator. Hum.
regard). The new scope of the social disregard scale suggests that Factors 40, 136–149.
this aberrant behavior should be regarded as a special type of Matthews, G., Tsuda, A., Xin, G., Ozeki, Y., 1999. Individual differences
violation referring to safety-critical omissions in communicat- in driver stress vulnerability in a Japanese sample. Ergonomics 42, 401–
415.
ing intentions to other road users. The fact that these two factors Newman, J.E., Beehr, T.A., 1979. Personal and organizational strategies for
were negatively correlated with alertness implies that there is handling job stress: a review of research and opinion. Pers. Psychol. 32,
a danger that the frequent repetition of negligence and social 1–43.
924 T. Kontogiannis / Accident Analysis and Prevention 38 (2006) 913–924

Parker, D., Reason, J.T., Manstead, S.R., Stradling, S.G., 1995. Driving errors, Stokes, A., Kite, K., 1994. Flight Stress: Stress, Fatigue and Performance in
driving violations and accident involvement. Ergonomics 38, 1036–1048. Aviation. Ashgate Publishing Company, Vermont.
Reason, J., 1990. Human Error. Cambridge University Press, Cambridge. Westerman, S.J., Haigney, D., 2000. Individual differences in driver stress,
Reason, J.T., Manstead, A., Stradling, S., Baxter, J., Campbell, K., 1990. error and violation. Pers. Indiv. Differ. 29, 981–998.
Errors and violations on the roads: a real distinction? Ergonomics 33, Wickens, C.D., 1992. Engineering Psychology and Human Performance, 2nd
1315–1332. ed. Harper Collins, New York.

You might also like