Development and preliminary validation of a scale of driving moral disengagement as a tool in the exploration of driving aggression

Transportation Research Part F 46 (2017) 124–136

Contents lists available at ScienceDirect

Transportation Research Part F journal homepage: www.elsevier.com/locate/trf

Development and preliminary validation of a scale of driving moral disengagement as a tool in the exploration of driving aggression Alison Swann, Alexia Lennon ⇑, Jasmine Cleary Centre for Accident Research and Road Safety, Queensland (CARRS-Q), Queensland University of Technology, Australia

a r t i c l e

i n f o

Article history: Received 24 April 2016 Received in revised form 5 September 2016 Accepted 24 January 2017

Keywords: Driving aggression Driving anger Moral disengagement Scale development

a b s t r a c t Aggressive driving has been found to result in road collisions which are a major cause of injury, fatality and financial cost in motorised countries. Qualitative and survey based studies suggest that drivers use justifications or explanations of their aggressive driving that bear strong resemblance to Bandura’s mechanisms of moral disengagement. The aim of the current study was to explore the applicability of moral disengagement to the driving context using a purpose-adapted scale, the Driving Moral Disengagement Scale. A convenience sample of general drivers (N = 294) responded to an on-line survey comprised of measures of trait anger, driving anger (DAX revised), moral disengagement and driving moral disengagement. Factor analysis allowed for reduction of the new scale from 23 items to 13 items, and this shortened Driving Moral Disengagement Scale (DMDS) had good internal reliability (Cronbach’s alpha = .83). Scree plot criteria indicated a one factor solution accounting for 34.34% of the variance. Bivariate correlations on the shortened DMDS revealed significant and positive relationships with measures of driving aggression, moral disengagement, trait anger and driving anger, r = .28–.55. Moreover the strength of the association between driving aggression and moral disengagement was greater than that with driving anger. Hierarchical regression revealed driving moral disengagement as the strongest significant predictor of driving aggression, accounting for 18% of the unique variation in the DV, and suggesting this may be a more useful predictor than driving anger. In addition, significant differences between participants’ mean scores for moral disengagement in everyday situations and their driving moral disengagement scores support the interpretation that drivers may behave differently from their ‘usual’ selves when driving, and that the driving context may encourage both greater moral disengagement and greater tendency towards aggressive responses. Chi square analysis indicated that those who scored high on driving moral disengagement were significantly more likely to report aggressive responses to driving situations than those with low driving moral disengagement scores (with a large effect size, u = .42). This suggests that the DMDS may be useful for future driving aggression research. Implications for intervention are that aiming to alert drivers to their usual self-censure mechanisms or to prevent the tendency to moral disengagement while driving may be effective in reducing driving aggression and the risky or dangerous responses associated with it on road. Ó 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Centre for Accident Research and Road Safety (CARRS-Q), Queensland University of Technology, 130 Victoria Park Road, Kelvin Grove, Queensland 4053, Australia. E-mail address: [email protected] (A. Lennon). http://dx.doi.org/10.1016/j.trf.2017.01.011 1369-8478/Ó 2017 Elsevier Ltd. All rights reserved.

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

125

1. Introduction Aggressive driving may have important implications for road safety as it has been found to contribute to road collisions which can result in injuries, fatalities, damage to vehicles and financial cost (Wickens, Mann, & Wiesental, 2013). While the more extreme, road violence events are rare, non-violent driving aggression appears to be much more common, and may also be of concern, especially where drivers choose to display riskier behaviours such as following too closely or attempting to intimidate others with the vehicle (Chliaoutakis et al., 2002; Wells-Parker et al., 2002). For this discussion, driving aggression can be regarded as any behaviour that is directed at other road users with an intention to cause a negative physical or psychological effect on the other in an attempt to achieve a goal (Soole, Lennon, Watson, & Bingham, 2011). Some situations have been found to be more likely to evoke aggression than others. For example: environments involving alcohol (Leonard, Quigley, & Collins, 2003), social pressure (Warley, 2009) or uncomfortably warm temperatures (Carlsmith & Anderson, 1979). Similarly, results from survey studies have suggested that aggression is more common in the driving environment than in everyday situations. Self-report surveys have found that up to 82% of drivers claimed to have been a victim of driving aggression (AAMI, 2007), and as much as half of drivers admit to either initiating an aggressive driving action or responding to another driver’s action in an aggressive manner. In addition, one multicountry survey found that people who reported high levels of trait anger also reported that they were more likely to express anger aggressively when on the road than in non-driving situations (Lawton & Nutter, 2002). While both trait anger and trait aggression have been shown to be positively associated with driving anger and aggression (Deffenbacher, Deffenbacher, Lynch, & Richards, 2003; Deffenbacher, Filetti, Richards, Lynch, & Oetting, 2003; HerreroFernández, 2013; Lajunen & Parker, 2001; Stephens & Sullman, 2014; Wickens, Wiesenthal, Flora, & Flett, 2011) it appears that these factors do not fully account for driving aggression, and that other situational or person-related factors such as age and gender (Hennessy & Wiesenthal, 2001, 2004), narcissism (Britt & Garrity, 2006), sensation seeking (Jonah, Thiessen, & Au-Yeung, 2001) or stress tolerance (Beck, Ali, & Daughters, 2014) as well as cognitive factors (Vallieres, Bergeron, & Vallerand, 2005; Wickens, Roseborough, Hall, & Wiesenthal, 2013; Yagil, 2001), may influence whether a driver expresses his or her anger in the driving environment. One factor that may explain why people may be more likely to express their anger in an aggressive manner on-road compared to their responses in everyday situations, is that the driving environment may encourage drivers to feel more anonymous (Ellison-Potter, Bell, & Deffenbacher, 2001; Ellison, Govern, Petri, & Figler, 1995), less responsible (Wickens et al., 2011) and less connected to their fellow road users. Anonymity may also lead to deindividuation (Diener, Fraser, Beaman, & Kelem, 1976) where a person becomes unselfconscious in relation to his/her behaviour and loses concern for the evaluations of others. In the driving context, deindividuation may render drivers more likely to engage in behaviours that violate social norms (Ellison et al., 1995; Ellison-Potter et al., 2001; Hennessy & Wiesenthal, 1999) and to greater beliefs that the probability of being held to account is reduced (Diener et al., 1976; Wickens et al., 2011) and that the vehicle provides safety and an easy escape (Hennessy & Wiesenthal, 1999). Thus perceived anonymity while driving may result in individuals being less courteous, more likely to display anger or aggression, and less inclined to fear sanction or accountability for their behaviour. A second, and potentially related factor, is that the lack of connection with other drivers that is inherent in the driving environment may provoke the activation of psychological mechanisms that promote individuals’ disengagement from their typical ways of thinking about themselves in relation to others, in particular, their moral values about appropriate behaviour. This has been termed ‘moral disengagement’ (Bandura, Barbaranelli, Caprara, & Pastorelli, 1996). Moral disengagement comes from a broader self-theory which looks at moral functioning and is based on the idea that people normally act in accord with their moral standards, and that behaviour is constrained by activation of self-regulatory mechanisms (eg; self-blame and self-pride). However, sometimes people behave in ways that are inconsistent with their normal morals, even to the point of engaging in behaviour that is generally regarded as immoral. Bandura (2002) proposed that for this to occur, individuals must deactivate their self-regulatory mechanisms, and disengage from their normal moral code. Bandura (2002) termed this moral disengagement, and proposed that eight psychological mechanisms underlie the

Mechanisms of moral disengagement Moral justification: reframing an act to be for the greater good Advantageous comparison: comparing the act to more harmful acts Displacement of responsibility: mentally shifting responsibility for the harmful act to someone else Diffusion of responsibility: mentally distributing the responsibility for the act across a group Distortion of consequences: convincing the self that the act did not in fact cause harm Dehumanisation: adopting the view of the victim as an object or animal to weaken empathy Attribution of blame: arguing that the victim enticed the harmful act Fig. 1. Mechanisms of moral disengagement (adapted from Bandura, 1999, 2002; Caprara et al., 2014).

126

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

process of disengaging from one’s internal moral standards (see Fig. 1). These mechanisms permit engagement in immoral, harmful or illegal behaviour while still maintaining perceptions of the self as moral. In addition they allow avoidance of selfcensure and the accompanying unpleasant emotions that are normally aroused during self-regulation (Bandura, 2002; Caprara et al., 2014). 1.1. Moral disengagement and driving aggression Qualitative and survey-based evidence from Australian and international studies on driving aggression suggests that drivers may use justifications or explanations of their aggressive driving behaviours that bear strong resemblance to Bandura’s mechanisms (AAMI, 2007; Lawton & Nutter, 2002; Lennon & Watson, 2011, 2012). That is, when explaining their behaviour, some drivers have claimed that their aggressive driving did not actually have a negative effect on the target or that the target in fact harmed them instead (Lennon & Watson, 2011). Others have described their intentions as essentially honourable or socially valuable: they were teaching the target of their driving actions a valuable lesson. Some claimed that their retaliation was justified by the intentions or impact of the other’s behaviour (AAMI, 2007; Lawton & Nutter, 2002). Finally, some drivers who behaved in retaliatory or aggressive ways towards others used examples of others’ more dangerous, risky or aggressive behaviours in order to make their own behaviours seem mild in comparison (AAMI, 2007; Lawton & Nutter, 2002). Such descriptions of self-behaviour or motives can be argued as being examples of distortions of consequences, attributions of blame, moral justifications and advantageous comparisons as described above, and support the contention that elements of the driving environment may facilitate moral disengagement. In turn, these may explain why people self-report behaving more aggressively on the road than in everyday circumstances, despite feeling equally high levels of anger in both situations (Lawton & Nutter, 2002). Previous studies have demonstrated a clear link between moral disengagement and general aggression in US adolescents (Gini, Pozzoli, & Bussey, 2015); with peer reported bullying in Danish 12 year olds (Obermann, 2011); and with violence in Italian adolescents (Caprara et al., 2014; Gabbiadini, Riva, Andrighetto, Volpato, & Bushman, 2014). The concept has also been extended into studies of other behaviour such as unethical decision-making in organisational employees (Detert, Trevino, & Sweitzer, 2008). One Scottish study found that juvenile offenders scored higher on moral disengagement measures than the general community (Kiriakidis, 2008). Moreover, in that study, moral disengagement was reported to be unrelated to social, family, school, employment, legal or lifestyle factors, suggesting that moral disengagement is a strong and unique predictor of aggressive or illegal behaviour (Kiriakidis, 2008). 1.2. Mechanisms of moral disengagement scale The mechanisms of moral disengagement scale was developed by Bandura to measure moral disengagement in children and adolescents (Bandura et al., 1996; Bandura, Caprara, Barbaranelli, & Pastorelli, 2001; Pelton, Gound, Forehand, & Brody, 2004). The scale contains 32 items aimed at measuring a child’s readiness to use the different moral disengagement mechanisms in different moral scenarios such as lying (‘‘It is ok to tell small lies because they don’t really do any harm”). Responses are measured on a child-friendly three point scale (1 = disagree, 3 = agree). Although the scale was designed to assess the eight posited mechanisms of moral disengagement, Bandura and colleagues reported the results of a factor analysis which only supported a single factor solution (Cronbach’s alpha = .82, N = 799) (1996). Moral disengagement scores were found to be positively associated with self, peer, teacher and parent-reported aggression and negatively associated with self, peer, teacher and parent-reported pro-social behaviour, indicating sound construct validity (Bandura et al., 1996). Bandura’s scale has subsequently been adapted by others for specific contexts or to extend beyond children. Examples include: Moral Disengagement in Sport Scale (Boardley & Kavussanu, 2007), Civic Moral Disengagement Scale (Caprara, Fida, Vecchione, Tramontano, & Barbaranelli, 2009) and a general Moral Disengagement Scale (Detert et al., 2008). In their extension of Bandura’s work, Detert and colleagues adapted the wording of items in the Moral Disengagement Scale to better fit an adult population (Detert et al., 2008) and incorporated a 5 point response scale (1 = strongly disagree, 5 = strongly agree) in order to allow for varying degrees of perception of self-behaviour. The overall moral disengagement scale was found to have adequate internal consistency (Cronbach’s alpha = .87, N = 424) (Detert et al., 2008). Factor analysis using the adapted scale supported the existence of eight factors (as suggested by the original theory) but with a reduced number of items (to 32 from 24). The final version of the adapted scale (Detert et al., 2008) has been used in multiple subsequent studies including research targeting the relationship between moral disengagement: aggression and revenge (White-Ajmani & Bursik, 2014); social support and acceptance (Chugh, Kern, Zhu, & Lee, 2014); and cynicism and consumer ethics (Chowdhury & Fernando, 2014). 1.2.1. Current study The aim of the current study was to explore the utility of the construct of moral disengagement in the driving context, particularly in accounting for driving aggression. To examine this, a context-specific adaptation of the adult version of the Moral Disengagement Scale (Detert et al., 2008), suitable for driving situations, was created.

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

127

2. Method 2.1. Participants Drivers (N = 294, 154 = females, 140 = male) were recruited from two different sources. A community sample was recruited via email invitation to the employees and associates of two large private organisations in Australia (n = approximately 268). A small student sample was also recruited using the student participant pool of a large publicly funded university in the Australian state of Queensland (n = 26). Community respondents were eligible to enter a random draw for one of six $50 AUD gift vouchers while students received course credit in exchange for their participation. Participant eligibility criteria were: over 18 years old, hold a current Australian drivers’ licence, and drive on average at least two hours per week. In this study, the majority of drivers were aged 25–34 (44%), received an annual income of $50,000–$80,000 (29%), drove 10 h per week or less (51%) and held a tertiary qualification of some kind (85%). 2.2. Measures 2.2.1. Trait anger Trait anger was measured using the 7 item anger subscale from the Aggression Questionnaire (Buss & Perry, 1992). The Aggression Questionnaire measures verbal aggression, physical aggression, hostility and anger. While the full 29 item scale was administered in the online survey, only the anger subscale has been used in the following analyses. The anger subscale of the AQ has been used in previous driving research and found to be adequately correlated with other driving anger and aggression variables (Lajunen & Parker, 2001). Participants respond to items on the AQ by indicating the extent to which the statements are characteristic of the self a five point scale ranging from 1 (extremely uncharacteristic of me) to 5 (extremely characteristic of me). The AQ anger sub scale has been found to have adequate internal consistency, a = .83, adequate stability, test-retest correlation coefficient = .72, and adequate construct validity, as measured by the correlations between self-report and peer nomination coefficient = .29 (Buss & Perry, 1992). 2.2.2. Driving anger Driving anger was measured using the UK Driving Anger Scale (Lajunen & Parker, 2001), an adaption of the Driving Anger Scale (DAS) (Deffenbacher, Oetting, & Lynch, 1994). Previous validation studies of the DAS have shown it to have good internal reliability (.90) (Deffenbacher et al., 1994) and criterion validity, positively correlating with self reported trait anger, risky driving behaviour, near-misses (Sullman & Stephens, 2013) and all subscales of the DAX (Stephens & Sullman, 2014). Convergent validity has also been reported as good, with strong correlations between the DAS and the Trait Anger Scale (Deffenbacher, Lynch, Deffenbacher, & Oetting, 2001). The 21 item adaptation was formed by omitting two subscales: ‘police presence’ and ‘traffic obstruction’ from the original 33 item DAS scale as these did not provoke appreciable anger in UK drivers. The UK Driving Anger Scale and subscales have been found to have adequate internal reliability (.76–.90) and criterion validity, correlating strongly with self-reported measures of angry driving reactions, driving aggression and general aggression (Lajunen & Parker, 2001). Previous driving research using the DAS has found that Australian drivers are more similar to their UK counterparts than to US drivers in terms of driving anger, particularly on the ‘police presence’ subscale (O’Brien, Tay, & Watson, 2002). Accordingly, the UK Driving Anger Scale was chosen for this study. Responses were measured using a five point scale ranging from 1 (not at all angry) to 5 (extremely angry). The subscales have been found to have adequate internal consistency with Cronbach’s alphas ranging from .76 to .90 (Lajunen & Parker, 2001). 2.2.3. Driving aggression The revised Driving Anger Expression inventory (DAX-r) (Stephens & Sullman, 2014) was selected as the measure of driving aggression in order to reduce questionnaire length. This 25 item scale was adapted from the original 49 item Driving Anger Expression inventory, (Deffenbacher et al., 2001), one of the most commonly used measures of driving aggression (Sullman, 2015; Sârbescu, Costea, & Rusu, 2012), in order to allow for a shorter version that is more easily combined with other scales, but which still retained four underlying factors. The revised scale measures four different ways that people express anger on the road: use of vehicle, e.g. ‘‘Follow right behind for a long time”; verbal aggression, e.g. ‘‘Swear at the other driver aloud”; personal physical aggression, e.g. ‘‘Try to get out of the car and tell the other driver off”; and adaptive constructive, ‘‘Tell myself it’s not worth getting mad at.” Validation analyses (Stephens & Sullman, 2014) revealed adequate convergent validity of the revised scale: that is, participants who reported the initiation of the more extreme aggressive interactions and the involvement in recent crash related conditions had higher scores on the revised scale when compared with those of other participants. Internal consistency was acceptable, and factors were also found to have acceptable reliability, with Cronbach’s alphas ranging between .74 and .91 across the principal components and confirmatory factor analyses (Stephens & Sullman, 2014). Response options are on a four point scale (‘Almost never’, Rarely, ‘Sometimes’, ‘Almost always’). In the current study, responses to items were coded 1 = ‘Almost never’ to 4 = ‘Almost always’. An overall driving aggression score for each driver was then calculated by summing values for responses to all items, excluding those on the Adaptive/ Constructive sub-scale, since the interest was in aggressive responses only. Unfortunately, an error in the conversion of the

128

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

survey to the online form resulted in one verbal aggression item being omitted: ‘‘make negative comments about the driver aloud”. Thus the driving aggression scores were derived from responses to the remaining 14 aggressive behavioural items (out of the full 25 items) of the Revised Driver Anger Expression inventory. 2.2.4. Moral disengagement Moral disengagement was measured using Detert and colleagues’ Moral Disengagement Scale (MDS) (Detert et al., 2008), itself an adaptation of Bandura’s original scale devised for children (2002). In adapting Bandura’s original scale to an adult population, Detert and colleagues altered wording of the items. For example: ‘‘Kids who get mistreated usually do things that deserve it” became ‘‘People who are mistreated have usually done things to deserve it”; and ‘‘Children are not at fault for misbehaving if their parents force them too much” was altered to ‘‘People are not at fault for misbehaving at work if their managers mistreat them.” In addition the MDS uses a five (rather than three) point response scale ranging from 1 (strongly disagree) to 5 (strongly agree). 2.2.5. Driving moral disengagement Driving moral disengagement was measured using the Driving Moral Disengagement Scale (DMDS), a driving-context, author-adaptation of the Moral Disengagement Scale (MDS, Detert et al., 2008) (described above). Adapting the scale consisted of first developing a list of common provocative driving situations based on literature in the driving aggression area (Soole et al., 2011). Next, items were reworded to fit the driving context, using content from the list. As with Detert’s scale, the revised scale items were designed to preserve the eight mechanisms of the moral disengagement construct, and context details were devised to match the level of response and seriousness of the behaviour from the original item. For example: ‘‘It is alright to fight to protect your friends”, a relatively strong response to perceived need, was changed to ‘‘It’s ok to yell things at drivers who put the lives of your passengers at risk” which also implies a relatively strong response and perceived threat, though it is verbally rather than physically aggressive. ‘‘Stealing some money is not too serious compared to those who steal a lot of money” was changed to ‘‘Speeding a little over the speed limit is not too serious compared to those that speed a lot over the limit” (full wording of the purpose-designed revised scale is displayed in Table 2 below). One item did not translate well into the driving environment so the resulting scale contained 23 items (compared to the original 24). For the resulting Driving Moral Disengagement Scale, participants were instructed: ‘The following statements refer to your thoughts and beliefs about different driving situations. Indicate how much you agree or disagree with each statement.’ Responses were measured using Detert et al. (2008) five point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Although participants also responded to the Anger Rumination Scale (Sukhodolsky, Golub, & Cromwell, 2001) and the Empathy Scale (Goldberg, 2001), this data is not reported in the current paper. 2.3. Procedure An online questionnaire (30 min duration) was developed and posted on line using the Key Survey program. Items consisted of the previously validated scales (described above) intended to measure driving aggression, trait anger, driving anger, and moral disengagement. In addition, a purpose-designed scale intended to measure driving moral disengagement was included. Demographic details included age bracket, income band, hours driven per week and highest level of education. After piloting and minor subsequent changes to the wording of some items in the questionnaire, the final version was posted online. Ethics approval for the study was granted from the university Human Ethics Research Committee (Approval No.: 1500000494). 2.4. Analyses Data analysis was carried out using IBM Statistical Package for the Social Sciences software Version 21 (IBM SPSS Statistics for Windows., 2012). Responses to the measures of interest were treated as continuous variables for the purposes of analysis. Principal axis factor analysis (PAF) was used to explore the psychometric properties of the purpose-designed driving moral disengagement scale and determine the most appropriate factor solution (i.e. one-factor or eight-factor). PAF was also conducted on the Moral Disengagement Scale for comparative purposes. Pearson correlations, Chi-square and t-test procedures were also used to explore driver responses to the measures and examine the associations between the variables of interest. 3. Results Table 1 displays descriptive data (mean, standard deviation, Cronbach’s Alpha) for each scale. Overall, item mean scores for all measures were below midpoint (M = 1.54–2.24) with the exception of driving anger, (M = 2.97), where the item mean score was equivalent to moderate levels of anger in driving contexts. Scores for driving anger (M = 2.97) were somewhat higher than for trait anger (2.24). For driving aggression, inspection of the distribution of item mean scores revealed that around 15% of the sample had scores equivalent to frequently engaging in many of listed aggressive driving behaviours. Ten participants (3.4%) stated that they never participated in any of the driving aggression behaviours, which may be more

129

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

Table 1 Descriptive characteristics (Mean, Standard Deviation and Cronbach’s Alpha) of measures of trait anger, driving anger, moral disengagement, driving moral disengagement and driving aggression. Scale

M

SD

Cronbach’s a

No. of items

Trait anger (AQa) Driving anger (UK DAS) Moral disengagement (MDS) Driving moral disengagement (23 items) Driving aggression (Revised DAX)

2.24 2.97 1.99 2.19 1.59

0.77 0.66 0.45 0.51 0.39

.83 .91 .88 .87 .81

7 21 24 23 14

Note. Driving Aggression calculated using a 4-point response scale. Remaining scales used a 5-point response scale.

Table 2 Gender differences (t-tests) on measures of trait anger, driving anger, moral disengagement (everyday situation, driving contexts) and driving aggression (with 95% Bias Corrected and Accelerated Confidence Intervals and effect sizes).

Measure

Male M (SD) n = 139

Female M (SD) n = 154

Mean difference [CI]

t

p

g2

Trait anger (AQa) Driving anger (UK DAS) Moral disengagement (everyday, MD) Driving moral disengagement scale (full version) Driving aggression (DAX-r)

2.12 2.90 2.29 2.30 1.58

2.36 3.04 2.10 2.09 1.59


0.24 [
0.41,
0.50]
0.15 [
0.30, 0.01] 0.21 [0.11, 0.31] 0.21 [0.10, 0.32]
0.01 [
0.10, 0.08]


2.71
1.92 4.14 3.58
0.13

.010 .052 .001 .001 .887

0.02

(0.73) (0.68) (0.51) (0.51) (0.41)

(0.78) (0.64) (0.42) (0.49) (0.38)

0.06 0.04

Note. Confidence intervals based on 1000 bootstrapped samples.

a reflection of socially desirable responding. However, if so, this suggests that the overall levels of endorsement of aggressive behaviours is probably an underestimate of real behaviour and that our results may be somewhat conservative. The pattern of results is consistent with previous studies, where low mean scores for driving aggression have been reported (Suhr & Nesbit, 2013; Sullman, 2015). 3.1. Gender Table 2 displays results independent samples t-tests to examine whether there were gender differences on any variable. Bias corrected accelerated (BCa) bootstrapping with 1000 samples were performed in order to compensate for violation of assumptions of normality in distributions. All variables met the assumption for homogeneity of variance. Results revealed gender differences for measures of moral disengagement (MDS), driving-context moral disengagement (Driving Moral Disengagement Scale) and trait anger (AQ anger subscale) such that men had significantly higher scores for moral disengagement (driving and everyday situations) than women while women had significantly higher trait anger than men. However, effect sizes for trait anger and driving moral disengagement were small, while that for moral disengagement in everyday situations was moderate. 3.2. Driving moral disengagement and moral disengagement A repeated measures (paired-samples) t-test was conducted to assess if there was a meaningful difference between participant scores on the Moral Disengagement Scale compared to the Driving Moral Disengagement Scale. Assumptions of normality were breached and thus, bias corrected accelerated (BCa) bootstrapping with 1000 samples was performed. On average, participant scores were significantly higher (
0.20, BCa 95% CI [
.25,
.15], t (293) =
7.30, p < .001), for moral disengagement in the driving context (M = 2.19, SE = 0.04) compared with everyday situations (M = 1.99, SE = 0.03), and this represented a large sized effect (g2 = 0.15) (Cohen, 1988). 3.3. Factor analyses In order to assess the performance of the context-adapted version of the moral disengagement measure for driving, the Driving Moral Disengagement Scale (23 item version), a factor analysis of the participant responses was conducted. Mean scores and standard deviations for the responses to each item (overall sample, N = 294) can be seen in Table 3. An initial factor analysis indicated poor loadings for some items. In order to address this and to develop a shorter scale, items with a mean score of less than 2.0 were deleted. As this score is equivalent to participant responses of ‘strongly disagree’ or ‘disagree’, it was reasoned that the removed items were less likely to reflect perceptions of acceptable on-road behaviours for this Australian driver sample. Removal of items on this basis is also not an uncommon practise in scale development, with other authors similarly electing to delete items with low means due to their apparent irrelevance to the sample population on the measures being examined (Lajunen & Parker, 2001; Marchegiani & Phau, 2013). Applying this criterion

130

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

Table 3 Driving Moral Disengagement Scale (23 item version): item wording, subscales, item mean scores, standard deviations, factor loadings (Principal Axis Factor Analysis) for items retained in the shortened version (13 item, bold type), and Cronbach’s Alpha if item excluded. Driving Moral Disengagement Scale: item wording, subscale membership, ordering (N = 294) (a = 0.87 full scale; a = 0.83 shortened scale –bold items) Moral justification 1. It’s alright to deliberately hold someone up by going slow if it’s for their own good 7.

It’s ok to tailgate if it gets people to realise they are doing the wrong thing

14.

It’s ok to yell at other drivers who put the lives of your passengers at risk

Euphemistic labelling 2. Honking the horn loudly is just a way of letting off frustration 8.

Preventing others from passing is just part of the game

15.

Following too closely or cutting someone off, is just a way of teaching someone a lesson they need

Advantageous comparison 3. Tailgating is no big deal when you consider other people are deliberately running red lights 9.

Yelling at other drivers is pretty tame when compared to people that attack other drivers

20.

Speeding a little over the limit is not too serious compared to those that speed a lot over the limit

Displacement of responsibility 10. If a driver is pushed into being rude to other drivers they shouldn’t be blamed for it 21.

People can’t be blamed for intimidating another driver if their friend pressured them into it

Diffusion of responsibility 4. You can’t blame a single driver for going through an amber light if a whole group does it 5.

It’s ok to go over the speed limit if it means you are keeping up with the rest of the traffic

16.

It’s unfair to blame a driver for driving in the bus lane if everyone is doing it

Distortion of consequences 11. Drivers don’t mind being honked at because they know it just means ‘hurry up’ 17.

Flashing headlights to get someone to move over, doesn’t really hurt anyone

22.

Tailgating other vehicles when the traffic is heavy isn’t really dangerous

Attribution of blame 13. If you are getting honked at while driving you probably deserve it 19 23.

Overly cautious drivers who are tailgated, deserve it because they are a risk to everyone on the road People who don’t know how to drive, provoke bad driving in others

Dehumanisation 6. It’s alright to abuse drivers who are behaving like ‘‘knobs” 12.

A driver who is inconsiderate doesn’t deserve to be treated like a normal person

18.

Some drivers deserve to be treated like the idiots they are

M (SD)

PAF loading

Alpha if item deleted

0.62

.82

0.45

.83

0.63

.82

0.45

.83

0.57

.82

0.36

.83

2.08 (0.86) 2.88 (1.16) 1.60 (0.76)

0.44

.83

0.73

.83

2.63 (0.99) 1.79 (0.85) 2.91 (1.17)

0.42

.83

0.43

.83

2.21 (1.08) 2.22 (0.99) 2.45 (1.13)

0.73

.81

0.53

.82

0.73

.81

2.35 (1.06) 1.65 (0.84) 3.07 (1.15) 2.22 (1.09) 1.49 (0.75) 1.58 (0.73) 1.39 (0.66) 2.50 (1.19) 3.12 (1.16) 2.03 (0.94) 1.39 (0.53) 2.44 (1.11) 2.52 (1.15) 1.84 (0.80)

reduced the 23 item Driving Moral Disengagement Scale to 15 items with mean scores (overall sample) ranging from 2.03 to 3.12. Principal Axis Factoring with oblique rotation was conducted on the remaining 15 item Driving Moral Disengagement Scale. Bartlett’s test of Sphericity was significant for the 15 item Driving Moral Disengagement Scale (c2 (105) = 1097.44, p < .001) indicating that data was factorable and adequately correlated.

131

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

Using Kaiser’s criterion of eigenvalues greater than one, the initial analysis identified a three-factor solution for the scale, which together accounted for 41.12% of the variance. However, inspection of the scree plot identified only one factor of significance. Statistical advice suggests that if sample size exceeds 250, but the average communality is less than 0.60, then it is appropriate to use the more conservative scree plot than Kaiser’s criterion (Field, 2013; Stevens, 2002). Inspection of the communalities revealed that the average value after extraction for the 15 item Driving Moral Disengagement Scale was .36 (N = 294). Hence the scree plot criterion was used, resulting in a one factor solution which accounted for 26.08% of the variance. Items 1 and 4 were removed due to having factor loadings under .4 (.22 and .31 respectively) and no sufficiently large (P.3) inter-correlations with other items. This resulted in a final revised 13 item Driving Moral Disengagement Scale (short version) with adequate reliability (Cronbach’s alpha = .83). The revised version accounted for 34.34% of the variance, an improvement on the previous solution. Factor loadings for the final Driving Moral Disengagement Scale and alpha if item deleted are presented in Table 4. For comparison purposes, a factor analysis was conducted on participant responses to the 24 item Moral Disengagement Scale (Detert et al., 2008). Principal Axis Factoring with oblique rotation (Direct Oblimin) was conducted with extraction based on Eigenvalues greater than one. Initial analysis of the scale identified six factors accounting for 43.42% of the variance. However, as the average communality was again less than 0.60 a one factor solution was deemed the most appropriate for the scale, as identified by the more conservative scree plot. Thus it would appear that the results for the current study are more consistent with those of Bandura and colleagues’ (Bandura et al., 1996) original development of their Mechanisms of Moral Disengagement Scale, where similarly a one factor solution was deemed appropriate, than with Detert and colleagues’ later work, which reported an eight factor solution for the adult version of the scale (Detert et al., 2008). For all subsequent analyses, the final 13 item version of the Driving Moral Disengagement Scale (short version, DMDS) has been used. Values for the factor loadings of the thirteen item DMDS are displayed in Table 3. As found for the full scale, men had significantly higher scores on the shortened driving moral disengagement (DMDS) than women (M = 2.67, SE = 0.05, M = 2.40, SE = 0.05 respectively), t (291) = 3.81, p < .001, BCa95% CI [0.13, 0.41], but this was a small sized effect, g2 = .01. 3.4. Driving moral disengagement and driving aggression Intercorrelations (Pearson’s r) were used to explore the associations between the variables of interest. With the exception of the correlation between driving anger and moral disengagement, all variables were significantly and positively associated (Table 4). In particular, driving moral disengagement as measured by the DMDS was significantly and moderately positively correlated with the MDS (r = .48) suggesting that the two measures are assessing related, but somewhat different, underlying constructs. Similarly, trait anger and driving anger were moderately correlated, suggesting an overlap in the constructs being captured. Consistent with the premise of the study, the strongest obtained relationship was between driving aggression scores and driving moral disengagement scores, r = . 55. As anticipated, the association between the driving moral disengagement variable and the driving aggression variable was stronger than that between driving aggression and moral disengagement in everyday situations, supporting the contention that the driving context elicits somewhat different behaviour from that of other everyday situations. Moreover, the association between driving moral disengagement and driving aggression was stronger than the associations with the anger measures (trait anger and driving anger). Consistent with past research, age was significantly and negatively related to all variables. However correlation values were small, ranging between r =
.15 and r =
.21. In order to explore the extent to which driving moral disengagement explains the variation in driving aggression, a hierarchical regression was performed. Age and driving anger were entered earlier in the regression as these variables have been found to be strong predictors of driving aggression in previous studies. Because of the moderate correlation between the measures of trait and driving anger, only driving anger was entered in this analysis. Moral disengagement in everyday situations (MD) was also entered in the analysis, despite the moderate level of correlation with driving moral disengagement,

Table 4 Intercorrelations (Pearson’s r) between measures of driving aggression, Driving Behaviour, Moral Disengagement and Anger Variables as well as Demographic Variables (N = 294). 1 1. 2. 3. 4. 5. 6.

Age Trait anger (AQ-a) Driving anger (UK DAS) Moral disengagement (MDS) Driving moral disengagement (DMDS short version) Driving aggression (DAX-r)

Two tailed significance. * p < .05. ** p < .01. *** p 6 .001.

2

3

4

5

6


.19***


.21*** .34***


.19*** .23*** .08


.19*** .31*** .28*** .20***


.15* .42*** .39** .28*** .55***

132

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

Table 5 Hierarchical regression for driving aggression (dependent variable) and age, driving anger, moral disengagement and driving moral disengagement. Variable

B

SE

Beta

t

p

sr2

Step 1 Age


.049

.019


.149*


2.547

.011

.02

Step 2 Age Moral disengagement Step 3 Age Moral disengagement Driving anger Step 4 Age Moral disengagement Driving anger Driving moral disengagement * ** ***


.038 .153


.015 .140 .212


.004
.057 .141 .319

.019 .051

.018 .047 .032

.016 .047 .030 .035


.114 .176**


.044* .162** .356***


.013
.066 .238*** .507***


1.958 3.028


.788 2.969 6.521


.261
1.211 4.760 9.035

.051 .003

.431 .003 .000

.794 .227 .000 .000

R

R2

R2adj

DR 2

.147

.022

.018

.022*

.227

.052

.045**

.0302**

.416

.173

.164***

.121***

.596

.355

.346***

.182***

.03

.02 .12

.05 .18

p < .05. p < .01. p < .001.

in order to permit separate conclusions about the contribution of each of these variables to explanation of variation in the DV. Results of the overall hierarchical regression models (Table 5) were significant for each step, with the final model (Step 4) accounting for 34.4% (adjusted) of the variance in driving aggression, F(4, 289) = 39.774, p < .001 (Table 5). Although age and moral disengagement in everyday situations (MD) were significant predictors in the early models, with the addition of the other variables (Steps 3 and 4) these were no longer significant in the final model. Driving anger and driving moral disengagement were the only significant predictors of driving aggression in the final model. Inspection of the beta weights revealed that driving moral disengagement was the strongest predictor, beta = .507, accounting for 18% of the unique variance in the DV. Moreover, driving moral disengagement was more than twice as strong as driving anger (beta = .238), after controlling for the explanation from age and moral disengagement in everyday situations. Driving anger accounted for a more modest 5% of the unique variance in driving aggression. As shown above, on average, participant scores for driving aggression were low (M = 1.59, SD = 0.39), and for driving moral disengagement were only somewhat higher, at just above the mid-point of the response scale (M = 2.53, SD = 0.63). These relatively low mean scores led to a decision to categorise drivers into high and low score groups on driving moral disengagement and on driving aggression in order to better explore the relationship between the two. Low scores on driving moral disengagement were defined conservatively as item mean scores on the DMDS of 63 (that is, overall mean responses equivalent to ‘Unsure’, ‘Disagree’ or ‘Strongly disagree’) (n = 227). High scores on this measure were defined as item mean scores >3 (n = 67). For driving aggression, low scores were defined as item mean scores <2 on the Revised DAX, equivalent to mean responses of ‘Almost never’ or ‘Rarely’ (n = 251). High scores were defined as item mean scores P2 (n = 43) (Table 5). Chi square testing revealed a significant, moderate to large sized effect of driving moral disengagement on driving aggression (v2 = 51.28 (1), p = .0005, Fisher’s Exact Test, u = .42) (Cohen, 1988). Whilst only consisting of 15% of the overall sample, the majority of drivers categorised as high in driving aggression were also categorised as high in driving moral disengagement (65%). In comparison, only 16% of drivers categorised as low in driving aggression category were categorised as high in driving moral disengagement (see Table 6).

Table 6 Number of participants (%) categorised as high or low scores for driving moral disengagement and driving aggression (N = 294). Driving moral disengagement (DMDS) Low n (% total sample) Driving aggression (DAX-r)

y

Low High Totals

212 (72.1) 15 (5.1)y 227 (77.2)

High n (% total sample) y

39 (13.3) 28 (9.5)y 67 (22.8)

These cells had a major contribution to the chi-square result (standardised residuals greater than ±1.96).

Totals 251 (85.4) 43 (14.6) 294 (100)

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

133

4. Discussion The purpose of the current study was to design and validate a tool, the Driving Moral Disengagement scale, to effectively measure moral disengagement in the driving context in order to explore the relationship between moral disengagement and driving aggression. To the researchers’ knowledge, this is the first study to attempt to apply the construct of moral disengagement to the driving environment. As noted above, previous research on driving aggression has reported that features of the driving environment are likely to increase the sense of one’s anonymity and potentially deindividuation (Ellison et al., 1995; Ellison-Potter et al., 2001; Hennessy & Wiesenthal, 1999; Wickens et al., 2011). These studies have found that under such environmental circumstances, driver tendencies to behave in ways that might contravene social norms or invite censure from others are increased. That is, factors external to the individual influence driver behaviour towards more aggressive responses because these disconnect the driver from social control mechanisms such as negative evaluation by others. The current study extends on this finding by exploring the influence of factors internal to the driver. Driving, because it involves fleeting interactions with unknown others, in an environment where drivers are likely think of themselves as anonymous, may be a particularly strong activator of psychological processes that disengage the self from others or from the impact of one’s actions on others. Moral disengagement mechanisms may also be more likely to be activated under such circumstances, producing disconnection from self-evaluation as well as disconnection from evaluation by others. Processes which disable self-sanctions, as in moral disengagement, may therefore help explain claims that drivers are more likely to express anger via aggressive behaviours while driving than they are in everyday situations. 4.1. Scale development By removing poorly performing items a shorter, more parsimonious 13 item Driving Moral Disengagement scale, with a single factor solution, and acceptable internal reliability (Cronbach’s alpha = .83), was created. Adequate convergent validity (r = .48) with everyday moral disengagement, as measured in this study by the Moral Disengagement Scale, was found for the new scale. Associations between the Driving Moral Disengagement scale and measures of theoretically related constructs, such as driving aggression (r = .55), moral disengagement (r = .48), trait anger (r = .31) and driving anger (r = .28) were positive and significant, as anticipated, suggesting that the new scale may be useful for future driving research. A single factor structure for the Driving Moral Disengagement scale, while being at odds with Detert and colleagues’ eight factor structure for moral disengagement, is consistent with Bandura’s earlier validation work on the Mechanisms of Moral Disengagement Scale. Moreover, the results of the comparative factor analysis of the moral disengagement measure (MD) also had a single factor structure. Bandura et al., 1996 earlier validation work on moral disengagement supported a single factor structure despite re-writing and exclusion of items to encourage sub factors for each mechanism to emerge. While Detert et al. (2008) analysis of the adult revision of the Mechanisms of Moral Disengagement Scale reported eight individual factors, these authors then decided to treat the resulting refined instrument as a single-factor scale in their subsequent analyses. In the current study, the single factor solution for the new driving-related measure of moral disengagement may be partly due to the constraints of the driving context: that is, the reduced quality of interpersonal interaction and the reduced time available for more complex reasoning in coming to decisions about which behavioural responses are the most appropriate or desired. It also seems reasonable to speculate that some of the mechanisms of moral disengagement in the original and adult versions of the scales are less applicable to the types of decisions or behaviours involved in driving. For instance, in development of the driving-related items for the new scale, it was challenging to generate valid and common examples for some of the mechanisms (e.g. diffusion or displacement of responsibility). 4.2. Moral disengagement in everyday versus driving contexts This driving sample appeared to morally disengage more in the driving context than in other everyday situations. Moreover, additional t-testing revealed that the difference in mean responses on the measures of disengagement in the two contexts was greater when the shortened version (13 item) was used than for the 23 item version, t (294) =
15.949, df = 294, p < .000 CI [
.60,
.47]. The effect size was also larger (g2 = .46 compared to g2 = .15 for 23 item scale) (Cohen, 1988), providing support for the contention that people behave differently from their ‘usual’ selves when driving, though it should be noted that the mean responses on the MD scale were low (corresponding to low tendency to moral disengagement). Further research should examine whether this provides a plausible and strong explanation for past findings that people report more aggressive responses to provocations encountered on-road than for non-driving situations (Lawton & Nutter, 2002). 4.3. Driving moral disengagement and driving aggression Results of the hierarchical regression revealed that driving moral disengagement was the strongest of the predictors in explaining driving aggression, with beta weights suggesting that this variable is over twice as strong as driving anger, a variable that is normally a strong predictor in research on driving aggression. Moreover, driving moral disengagement explained a high proportion (18%) of the unique variance in driving aggression, suggesting it may be a very useful predictor in future research and in intervention design.

134

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

Drivers categorised as high in driving moral disengagement were more likely than those categorised as low in driving moral disengagement to report high levels of driving aggression, and this was a medium to large effect size. Moreover, almost two thirds of drivers high in moral disengagement reported greater tendency towards driving aggression, suggesting that moral disengagement may be a mediator of driving aggression. Although the resulting proportion of drivers high in both driving moral disengagement and driving aggression was only 9.5%, this is a substantial minority of drivers whose behaviour on-road may represent a safety threat to themselves and others. This supports the main premise of the study and complements previous research in relation to the way that drivers justify their driving aggression (AAMI, 2007; Lawton & Nutter, 2002; Lennon & Watson, 2011). 4.4. Strengths and limitations A strength of the study was in the preliminary validation of a new scale that is more strongly related to driving aggression than previous measures and thus potentially very useful in future driving aggression research. In addition, the regression results point to a driving moral disengagement as a strong predictor for driving aggression. Conservative measures in the determination of high and low scoring driver groups in this study allows greater confidence that the significant differences that were detected between these drivers also represents a meaningful difference. One potential limitation in relation to the results here is that the purpose-designed DMDS contained some items that might be regarded as examples of driving aggression, and thus may be measuring something similar to the DAX-r. However, the items that were retained in the shortened DMDS are more general in nature, and access attitudes and cognitions in relation to other drivers, with the more extreme types of driving behaviour (e.g. intimidating others) having been dropped in the final scale. In contrast, the DAX-r is more behaviourally-focussed, and asks about frequency of particular responses. As such, we would argue that the DMDS adds to the tools of driving aggression researchers in that it may permit the detection of modifiable ways of approaching driving interactions, and thus the potential to reduce on-road aggression. Future research could endeavour to use more objective ways of exploring the relationship between moral disengagement in the driving context and actual driving responses via observational means (e.g. driving simulator studies). Although the current study examined convergent validity in relation to the shortened version of the DMDS, the issue of discriminant validity has not been addressed in this study, largely because of constraints on the length of the survey. However such issues are important. Moore, Detert, Trevino, Baker, & Mayer, 2012 demonstrated discriminant validity between their measure of moral disengagement and a measure of socially desirable responding, a competing potentially explanatory construct. Future research could thus attempt to explore discriminant validity in the DMDS using such measures. As this study relied on self-report measures, a further limitation is that the results may not be a true reflection of drivers’ beliefs, attitudes or behaviours, or that socially desirable responding may have resulted in drivers being reluctant to admit to thinking about other drivers in morally disengaged ways. However, this would tend to make the results reported here more conservative than for real driving, suggesting that levels of moral disengagement and driving aggression are actually higher. 4.5. Implications for practice In terms of intervening with drivers to increase safety, the results of this study suggest that if drivers can be supported or encouraged to remain engaged with their usual morals and self-censure mechanisms, driving aggression may be reduced. Some studies of moral disengagement in everyday situations have produced promising results in this respect. For instance, when hotel guests were cued as to the ethical dimensions of their decisions (e.g. inclusion of the information ‘There are very important ethical and moral aspects to this situation’), tendency to moral disengagement was reduced, which in turn reduced aggression as measured by likelihood of vindictive or exaggeratedly negative reviews of the hotel (He & Harris, 2014). In a simulated employment context, Kish-Gephart, Detert, Trevino, Baker, and Martin (2014) demonstrated that informing participants about the negative potential impact of poor behaviour on others altered the extent to which individuals engaged in unethical behaviour or utilised morally disengaged reasoning, albeit not when levels of personal gain were very high. Studies of ethical behaviour in organisations have reported that conscientiousness (Kish-Gephart et al., 2014) and empathy (Moore et al., 2012) act as mediators of the relationship between ethical behaviour and moral disengagement, suggestive of the potential usefulness of these factors in reducing moral disengagement. For driving, these findings suggest that targeted campaigns that alert drivers to the moral implications of aggressive behaviours or the direct impact on others with whom they might identify, may minimise driving moral disengagement and thus reduce the potential for driving aggression. Alternatively, interventions may be effective if they attempt to engage or strengthen drivers’ self-regulatory mechanisms, such as self-pride, self-blame and anticipated regret, which might normally be suppressed during moral disengagement (Bandura, 2002). Future research could examine the impact of encouraging self-pride and self-blame towards one’s own driving on driving aggression. References AAMI (2007). It’s all about me: Selfish attitudes set motorists on crash course: Crash Index 13 [press release] Available at:. Accessed March 10, 2015 .

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

135

Bandura, A. (1999). Selective moral disengagement in the exercise of moral agency. Journal of Moral Education, 31(2), 101–119. http://dx.doi.org/10.1080/ 0305724022014322. Bandura, A. (2002). Selective moral disengagement in the exercise of moral agency. Journal of Moral Education, 31(2), 101–119. http://dx.doi.org/10.1080/ 0305724022014322. Bandura, A., Caprara, G., Barbaranelli, C., & Pastorelli, C. (2001). Sociocognitive self-regulatory mechanisms governing transgressive behavior. Journal of Personality and Social Psychology, 80(1), 125–135. http://dx.doi.org/10.1037/0022-3514.80.1.125. Bandura, A., Barbaranelli, C., Caprara, G., & Pastorelli, C. (1996). Mechanisms of moral disengagement in the exercise of moral agency. Journal of Personality and Social Psychology, 71(2), 364–374. http://dx.doi.org/10.1037/0022-3514.71.2.364. Beck, K., Ali, B., & Daughters, S. (2014). Distress tolerance as a predictor of risky and aggressive driving. Traffic Injury Prevention, 15(4), 349–354. http://dx. doi.org/10.1080/15389588.2013.829569. Boardley, I., & Kavussanu, M. (2007). Development and validation of the moral disengagement in sport scale. Journal of Sport and Exercise Psychology, 29(5), 608–628. Britt, T. W., & Garrity, M. J. (2006). Attributions and personality as predictors of the road rage response. British Journal of Social Psychology, 45, 127–147. Buss, A., & Perry, M. (1992). The aggression questionnaire. Journal of Personality and Social Psychology, 63(2), 452–459. http://dx.doi.org/10.1037/00223514.63.3.452. Caprara, G., Fida, R., Vecchione, M., Tramontano, C., & Barbaranelli, C. (2009). Assessing civic moral disengagement: Dimensionality and construct validity. Journal of Individual Differences, 47(5), 504–509. http://dx.doi.org/10.1016/j.paid.2009.04.027. Caprara, G., Tisak, M., Alessandri, G., Fontaine, R., Fida, R., & Paciello, M. (2014). The contribution of moral disengagement in mediating individual tendencies toward aggression and violence. Developmental Psychology, 50(1), 71–85. http://dx.doi.org/10.1037/a0034488. Carlsmith, M., & Anderson, C. (1979). Ambient temperature and the occurrence of collective violence: A new analysis. Journal of Personality and Social Psychology, 37(3), 337–344. http://dx.doi.org/10.1037/0022-3514.37.3.337. Chliaoutakis, J. E., Demakakos, P., Tzamalouka, G., Bakou, V., Koumaki, M., & Darviri, C. (2002). Aggressive behaviour while driving as predictor of selfreported car crashes. Journal of Safety Research, 33(4), 431–443. Chowdhury, R., & Fernando, M. (2014). The relationships of empathy, moral identity and cynicism with consumers’ ethical beliefs: The mediating role of moral disengagement. Journal of Business Ethics, 124(4), 677–694. http://dx.doi.org/10.1007/s10551-013-1896-7. Chugh, D., Kern, M., Zhu, Z., & Lee, S. (2014). Withstanding moral disengagement: Attachment security as an ethical intervention. Journal of Experimental Social Psychology, 51, 88–93. http://dx.doi.org/10.1016/j.jesp.2013.11.005. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Routledge. Deffenbacher, J., Deffenbacher, D., Lynch, R., & Richards, T. (2003). Anger, aggression, and risky behavior: A comparison of high and low anger drivers. Behaviour Research and Therapy, 41(6), 701–718. http://dx.doi.org/10.1016/S0005-7967(02)00046-3. Deffenbacher, J. L., Filetti, L. B., Richards, T. L., Lynch, R. S., & Oetting, E. R. (2003). Characteristics of two groups of angry drivers. Journal of Counseling Psychology, 50(2), 123–132. Deffenbacher, J., Lynch, R., Deffenbacher, D., & Oetting, E. (2001). Further evidence of reliability and validity for the driving anger expression inventory. Psychological Reports, 89(3), 535–540. http://dx.doi.org/10.2466/pr0.2001.89.3.535. Deffenbacher, J., Oetting, E., & Lynch, R. (1994). Development of a driving anger scale. Psychological Reports, 74(1). http://dx.doi.org/10.2466/ pr0.1994.74.1.83. 83–81. Detert, J., Trevino, L., & Sweitzer, V. (2008). Moral disengagement in ethical decision making: A study of antecedents and outcomes. Journal of Applied Psychology, 93(2), 374–391. http://dx.doi.org/10.1037/0021-9010.93.2.374. Diener, E., Fraser, S. C., Beaman, A. L., & Kelem, R. T. (1976). Effects of deindividuation variables on stealing among Halloween trick-or-treaters. Journal of Personality and Social Psychology, 33(2), 178–183. Ellison-Potter, P., Bell, P., & Deffenbacher, J. (2001). The effects of trait driving anger, anonymity, and aggressive stimuli on aggressive driving behavior. Journal of Applied Psychology, 31(2), 431–443. http://dx.doi.org/10.1111/j.1559-1816.2001.tb00204.x. Ellison, P. A., Govern, J. M., Petri, H. L., & Figler, M. H. (1995). Anonymity and aggressive driving behavior: A field study. Journal of Social Behavior and Personality, 10(1), 265–272. Field, A. (2013). Discovering statistics using IBM SPSS statistics (3rd ed.). London, England: SAGE Publications Ltd. Gabbiadini, A., Riva, P., Andrighetto, L., Volpato, C., & Bushman, B. (2014). Interactive effect of moral disengagement and violent video games on self-control, cheating, and aggression. Social Psychological and Personality Science, 5(4), 451–458. http://dx.doi.org/10.1177/1948550613509286. Gini, G., Pozzoli, T., & Bussey, K. (2015). The role of individual and collective moral disengagement in peer aggression and bystanding: A multilevel analysis. Journal of Abnormal Child Psychology, 43(3), 441–452. http://dx.doi.org/10.1007/s10802-014-9920-7. Goldberg, L. (2001). International Personality Item Pool: A scientific collaboratory for the development of advanced measure of personality traits and other individual differences Available at:. Accessed March 10, 2015 . He, H., & Harris, L. (2014). Moral disengagement of hotel guest negative WOM: Moral identity centrality, moral awareness, and anger. Annals of Tourism Research, 45, 132–151. http://dx.doi.org/10.1016/j.annals.2013.10.002. Hennessy, D. A., & Wiesenthal, D. L. (1999). Traffic congestion, driver stress, and driver aggression. Aggressive Behavior, 25(6), 409–423. Hennessy, D. A., & Wiesenthal, D. L. (2001). Gender, driver aggression, and driver violence: An applied evaluation. Sex Roles, 44(11), 661–676. Hennessy, D. A., & Wiesenthal, D. L. (2004). Age and vengeance as predictors of mild driver aggression. Violence and Victims, 19(4), 469–477. Herrero-Fernández, D. (2013). Do people change behind the wheel? A comparison of anger and aggression on and off the road. Transportation Research Part F: Traffic Psychology and Behaviour, 21, 66–74. http://dx.doi.org/10.1016/j.trf.2013.09.008. IBM SPSS Statistics for Windows. (2012) [computer program]. Version 21.0. Armonk, NY: IBM Corp. Jonah, B. A., Thiessen, R., & Au-Yeung, E. (2001). Sensation seeking, risky driving and behavioral adaptation. Accident Analysis and Prevention, 33(5), 679–684. Kiriakidis, S. (2008). Moral disengagement relation to delinquency and independence from indices of social dysfunction. International Journal of Offender Therapy and Comparative Criminology, 52(5), 571–583. http://dx.doi.org/10.1177/0306624X07309063. Kish-Gephart, J., Detert, J., Trevino, L., Baker, V., & Martin, S. (2014). Situational moral disengagement: Can effects of self-interest be mitigated? Journal of Business Ethics, 125, 267–285. http://dx.doi.org/10.1007/s10551-013-1909-6. Lajunen, T., & Parker, D. (2001). Are aggressive people aggressive drivers? A study of the relationship between self-reported general aggressiveness, driver anger and aggressive driving. Accident Analysis and Prevention, 33(2), 243–255. http://dx.doi.org/10.1016/S0001-4575(00)00039-7. Lawton, R., & Nutter, A. (2002). A comparison of reported levels and expression of anger in everyday and driving situations. British Journal of Psychology, 93 (3), 407–423. http://dx.doi.org/10.1348/000712602760146521. Lennon, A., & Watson, B. (2011). ‘‘Teaching them a lesson?” A qualitative exploration of underlying motivations for driver aggression. Accident Analysis and Prevention, 43(6), 2200–2208. http://dx.doi.org/10.1016/j.aap.2011.06.015. Lennon, A., & Watson, B. (2012). Coping with aggressive driving: Driver accounts of how they manage themselves and others to reduce conflict on the road. In Proceedings of XXII Canadian multidisciplinary road safety conference. CARSP; Canada. Leonard, K., Quigley, B., & Collins, L. (2003). Drinking, personality, and bar environmental characteristics as predictors of involvement in barroom aggression. Addictive Behaviors, 28(9), 1681–1700. http://dx.doi.org/10.1016/j.addbeh.2003.08.042. Marchegiani, C., & Phau, I. (2013). Development and validation of the personal nostalgia scale. Journal of Marketing Communications, 19(1), 22–43. http://dx. doi.org/10.1080/13527266.2010.542075. Moore, C., Detert, J., Trevino, L., Baker, V., & Mayer, D. (2012). Why employees do bad things: Moral disengagement and unethical organizational behaviour. Personnel Psychology, 2012(65), 1–48.

136

A. Swann et al. / Transportation Research Part F 46 (2017) 124–136

O’Brien, S., Tay, R., & Watson, B. (2002). An exploration of Australian driving anger. Paper Presented at: Road Safety Research, Policing and Education Conference; August 4–5, Adelaide, Australia. Obermann, M. (2011). Moral disengagement in self-reported and peer-nominated school bullying. Aggressive Behavior, 37(2), 133–144. http://dx.doi.org/ 10.1002/ab.20378. Pelton, J., Gound, M., Forehand, R., & Brody, G. (2004). The moral disengagement scale: Extension with an American minority sample. Journal of Psychopathological and Behavioral Assessment, 26(1), 31–39. http://dx.doi.org/10.1023/B:JOBA.0000007454.34707.a5. Sârbescu, P., Costea, I., & Rusu, S. (2012). Using the alternative Five Factor Personality Model to explain driving anger expression. Procedia-Social and Behavioral Sciences, 33, 34–36. http://dx.doi.org/10.1016/j.sbspro.2012.01.126. Soole, D., Lennon, A., Watson, B., & Bingham, R. (2011). Towards a comprehensive model of driver aggression: A review of the literature and directions for the future. In C. Ferraro (Ed.), Traffic safety. Hauppauge, NY: Nova Science Publishers. Stephens, A., & Sullman, M. (2014). Development of a short form of the driving anger expression inventory. Accident Analysis and Prevention, 72, 169–176. http://dx.doi.org/10.1016/j.aap.2014.06.021. Stevens, J. (2002). Applied multivariate statistics for the social sciences (4th ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Suhr, K., & Nesbit, S. (2013). Dwelling on ‘road rage’: The effects of trait rumination on aggressive driving. Transportation Research Part F: Traffic Psychology and Behaviour, 21, 207–218. http://dx.doi.org/10.1016/j.trf.2013.10.001. Sukhodolsky, D., Golub, A., & Cromwell, E. (2001). Development and validation of the anger rumination scale. Personality and Individual Differences, 31(5), 689–700. http://dx.doi.org/10.1016/S0191-8869(00)00171-9. Sullman, M. (2015). The expression of anger on the road. Safety Science, 72, 153–159. http://dx.doi.org/10.1016/j.ssci.2014.08.013. Sullman, M., & Stephens, A. (2013). A comparison of the driving anger scale and the propensity for angry driving scale. Accident Analysis and Prevention, 58, 88–96. http://dx.doi.org/10.1016/j.aap.2013.05.002. Vallieres, E. F., Bergeron, J., & Vallerand, R. J. (2005). The role of attributions and anger in aggressive driving behaviours. Paper presented at the international conference of traffic and transport psychology. Warley, R. (2009). Situational predictors of adolescent homicide: A secondary analysis [dissertation]. New York, NY: City University of New York. Wells-Parker, E., Ceminsky, J., Hallberg, V., Snow, R. W., Dunaway, G., Guiling, S., ... Anderson, B. (2002). An exploratory study of the relationship between road rage and crash experience in a representative sample of US drivers. Accident Analysis and Prevention, 34, 271–278. White-Ajmani, M., & Bursik, K. (2014). Situational context moderates the relationship between moral disengagement and aggression. Psychology of Violence, 4(1), 90–100. http://dx.doi.org/10.1037/a0031728. Wickens, C., Mann, R., & Wiesental, D. (2013). Addressing driver aggression: Contributions from psychological science. Current Directions in Psychological Science, 22(5), 386–391. http://dx.doi.org/10.1177/0963721413486986. Wickens, C., Roseborough, J., Hall, A., & Wiesenthal, D. (2013). Anger-provoking events in driver diaries: A content analysis. Transportation Research Part F: Traffic Psychology and Behaviour, 19, 108–120. Wickens, C., Wiesenthal, D., Flora, D., & Flett, G. (2011). Understanding driver anger and aggression: Attributional theory in the driving environment. Journal of Experimental Psychology, 17(4), 354–370. http://dx.doi.org/10.1037/a0025815. Yagil, D. (2001). Interpersonal antecedents of drivers’ aggression. Transportation Research Part F: Traffic Psychology and Behaviour, 4(2), 119–131.