- Email: [email protected]
The expression of anger on the road
Safety Science 72 (2015) 153–159
Contents lists available at ScienceDirect
Safety Science journal homepage: www.elsevier.com/locate/ssci
The expression of anger on the road Mark J.M. Sullman Research Group, School of Engineering, Cranfield University, MK43 0AL, United Kingdom
a r t i c l e
i n f o
Article history: Received 10 February 2012 Received in revised form 28 August 2014 Accepted 28 August 2014
Keywords: Driving anger Anger expression Driving behaviour Angry drivers Risky driving
a b s t r a c t This research used the Driving Anger Expression Inventory (DAX) to investigate the expression of driving anger amongst a sample of New Zealand drivers. Confirmatory Factor Analysis found that the data fitted a three factor solution, which consisted of: Verbal Aggressive Expression; Use of a Vehicle to express anger, and an Adaptive/Constructive factor. The Personal Physical Aggressive Expression factor was not supported in the present sample. However, in line with previous findings, the present research found that the two aggressive types of anger expression were significantly related to potentially crash related conditions, such as: losing control of the vehicle, loss of concentration and near-misses. The two aggressive types of driving anger were also positively related to trait anger and driving anger, while the Adaptive/ Constructive factor was negatively related to both these variables. The present study also confirmed that those drivers who expressed their anger in an aggressive manner are more likely to be male and younger, while those who deal with their anger in a constructive manner were more likely to be older and female. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Anger is one emotion that is commonly experienced while driving (Underwood et al., 1999). Although research has found that individuals are not more likely to experience anger while driving, than in non-driving situations, those individuals who reported high levels of anger are more likely to be outwardly aggressive when driving (Lawton and Nutter, 2002). This is perhaps one of the reasons that the issue of driving anger appears to have gained popularity amongst researchers over the last 15 years (e.g. Deffenbacher et al., 2001b; Maxwell et al., 2005; Parker et al., 2002; Underwood et al., 1999; Sullman, 2006; Sullman et al., 2007; Sullman and Stephens; 2013). Another reason for the increasing popularity of this topic is that research has found that angry drivers engage more often in aggressive and dangerous driving behaviours (Stephens and Groeger, 2011), putting themselves and others at risk (Deffenbacher et al., 1994; Deffenbacher et al., 2001b; Lajunen et al., 1998; Maxwell et al., 2005; Underwood et al., 1999). Research has also found significant relationships to exist between driving anger and crash related conditions, such as losing control of their vehicle, losing concentration, speeding, tailgating, near misses, and moving violations (tickets) (Deffenbacher et al., 2001b; Deffenbacher et al., 2003a; Deffenbacher et al., 2003b; Sullman et al., 2013). Although a number of studies have investigated the situations which evoke anger amongst drivers (e.g. Deffenbacher et al., E-mail address: m.sullman@cranfield.ac.uk http://dx.doi.org/10.1016/j.ssci.2014.08.013 0925-7535/Ó 2014 Elsevier Ltd. All rights reserved.
1994; Lajunen et al., 1998; Parker et al., 2002; Stephens and Groeger, 2009; Sullman and Stephens, 2013), much less research has investigated how drivers react when angry. In order to measure the ways in which individuals respond to anger while driving Deffenbacher et al. (2002) developed the Driving Anger Expression Inventory (DAX). They found that drivers’ reactions could be placed into four categories: Verbal Aggressive Expression (VAE) – which measures people’s tendency to express their anger through verbally aggressive means (e.g. swearing at the other driver); Personal Physical Aggressive Expression (PAE) – which measures the ways in which the person uses themselves to express anger (e.g. shaking their fist); Use of Vehicle to express anger (UOV) – which measures how often drivers use their vehicle to express their anger (e.g. flashing their lights); Adaptive/Constructive Expression (A/C) – which measures constructive or adaptive behaviours the driver can make in potentially anger inducing situations (e.g. just ignore it). These four subscales have been found to have good internal reliability (range 0.80–0.90). Although the four DAX subscales have good internal reliability, little research has factor analysed the scale. As few peer-reviewed studies have reported factor analysing the 49-item version of the DAX, there is some degree of uncertainty regarding the scale’s underlying structure. Although the study which developed the DAX (Deffenbacher et al., 2002) subjected their data to factor analysis, this was a 62-item scale which was subsequently refined. Most of the following studies using the scale have used the item arrangement reported in the original research and have not subjected their data to any type of factor analysis (e.g. Dahlen and
154
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2003; Jovanovic´ et al., 2011; Moore and Dahlen, 2008), or have factor analysed a slightly different version of the DAX (e.g. HerreroFernández, 2011; Villieux and Delhomme, 2010). For example, Villieux and Delhomme (2010) removed all items from the Personal Physical Aggression subscale (11-items) before attempting to fit a three factor solution to their data using CFA. Furthermore, Herrero-Fernández (2011) used a 53-item version of the DAX and fitted a five factor model using CFA. Lastly, Sullman et al. (2013) performed a CFA on the 49-items using Turkish taxi drivers and confirmed the four factor solution, after removing one item from the model. Therefore, although previous research has broadly supported the factor structure of the DAX, the evidence is not extensive. Another issue surrounding research into the expression of driving anger is that most peer-reviewed research has focused mainly or exclusively on introductory psychology students (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2002; Deffenbacher et al., 2003; Deffenbacher et al., 2004; Deffenbacher et al., 2007; Esiyok et al., 2007; Herrero-Fernández, 2011; Moore and Dahlen, 2008; Villieux and Delhomme, 2010) who are not necessarily representative of the general population. Only two of the DAX studies did not rely on university students for their sample (Jovanovic´ et al., 2011; Sullman et al., 2013). Furthermore, most of the research has been conducted in the USA (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2002; Deffenbacher et al., 2003; Deffenbacher et al., 2004; Deffenbacher et al., 2007; Moore and Dahlen, 2008). There are a small number of studies which have been conducted outside of America, including France (Villieux and Delhomme, 2010), Spain (Herrero-Fernández, 2011), Turkey (Esiyok et al., 2007; Sullman et al., 2013) and Serbia (Jovanovic´ et al., 2011). As research has shown the level of driving anger differs between countries, as does the level of anger evoked by different types of situations (e.g. Deffenbacher et al., 1994; Lajunen et al., 1998; Sullman, 2006; Sullman et al., 2007), it is important to study additional countries. Given the fact that most research has largely relied on university students (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Esiyok et al., 2007; Herrerro-Fernandez, 2011; Villieux and Delhomme, 2010) it is surprising that most research has found the expression of driving anger to differ with age. For example, using a sample of Serbian motor vehicle owners, Jovanovic´ et al. (2011) found a significant negative relationship between age and the total of all the aggressive expression items. In terms of the individual DAX subscales, age was found to be negatively related to Personal Physical Aggressive Expression, Use of a Vehicle and Verbal Aggressive Expression, while Adaptive/Constructive expression was positively related to age (Deffenbacher et al., 2007). Furthermore, despite having a very narrow age range (18–25 years old), using a sample of French university students Villieux and Delhomme (2010) found age was negatively related to Using a Vehicle to express anger and positively related to the Adaptive/ Constructive subscale. In addition, in a sample consisting of almost 75% university students, Esiyok et al. (2007) found that age was a significant predictor of two of the aggressive subscales. Furthermore, despite having a sample where 75% of the participants were 30 years old or younger, Herrerro-Fernandez (2011) also found significant differences by age for all but the Adaptive/Constructive subscale. In contrast, research by Moore and Dahlen (2008) found age was not a predictor of any of the DAX subscales. Perhaps the reason for the lack of consistency in the age differences is due to the relatively narrow age spreads in most of the previous research. As with the research on driving anger, most research on the expression of driving anger has found sex differences. Several studies have found that females reported responding more often in an Adaptive/Constructive manner (e.g. Deffenbacher et al., 2004;
Esiyok et al., 2007; Jovanovic´ et al., 2011). Several studies have also found that males report more Personal Physical Aggressive Expression (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Esiyok et al., 2007). There were also some inconsistent findings, such as the fact that only Deffenbacher et al. (2004) found males reported Use of a Vehicle to express their anger more often than women and Dahlen and Ragan (2004) found that women reported more Verbal Aggressive Expression than men. Furthermore, there have also been a number of studies that have reported no sex differences (e.g. Deffenbacher et al., 2007; Moore and Dahlen, 2008; Villieux and Delhomme, 2010). In addition to demographic variables, the DAX has also been found to be related to other important variables. As would be expected, trait anger and driving anger have been found to be positively correlated with the three aggressive types of anger expression and negatively correlated with Adaptive/Constructive expression (Deffenbacher et al., 2001a). Furthermore, the three aggressive types of anger expression have been found to be positively correlated with risky and aggressive driving behaviour (e.g. Deffenbacher et al., 2002; Jovanovic´ et al., 2011). Verbal Aggressive Expression and Use of a Vehicle to express anger both had significant correlations with self-reported involvement in major accidents. Minor accidents, however, were most strongly correlated with Verbal Aggressive Expression of driving anger, as were close calls (Deffenbacher et al., 2001a). In addition, UOV and PAE were correlated with the most crash related to conditions, including risky behaviour and lost concentration (Deffenbacher et al., 2002). Given the relationships the DAX subscales have with variables important to the reduction of traffic crashes, the expression of driving anger appears to be an area worthy of further research. Therefore, the present study aimed to broaden the knowledge regarding the expression of driving anger in a number of ways. Firstly, the research explored the factor structure of the DAX amongst a sample taken from the general population of New Zealand, a country not previously studied. Furthermore, the relationship the resultant factors had with a number of descriptive variables (e.g. age, gender, experience, speed and annual mileage), the 14-item unidimensional version of the DAS, the TAS and a number of driving related outcomes (e.g. crash involvement, tickets) were also tested. Finally the descriptive variables and the DAS were also used to investigate whether they were significant predictors of the DAX factors.
2. Method 2.1. Materials The 49-item DAX was used to measure how drivers express their anger whilst driving (Deffenbacher et al., 2002). The scale presents 49 potential reactions to feeling angry while driving and respondents were asked to report how often they react in each of these ways on a four point scale (1 = Almost never, 4 = Almost always). The 14-item Driving Anger Scale (DAS) was used to provide an overall measure of driving anger (Deffenbacher et al., 1994). The scale presents 14 different situations and asks participants to rate how angry each situation makes them feel on a scale ranging from 1 (Not at all) to 5 (Very much). The 10-item Trait Anger Scale (TAS) was also used to measure general anger (Spielberger, 1999). The TAS measures how an individual generally responds when angry and is rated on a 4-point scale (1 = Almost never, 4 = Almost always). The eight violation items from the Driving Behaviour Questionnaire (DBQ; Reason et al., 1990) were used to measure risky driving
M.J.M. Sullman / Safety Science 72 (2015) 153–159
behaviour. Additional measures included: moving violations (tickets), loss of concentration, loss of control, close calls, minor accidents and major accidents in the last 3 months (Deffenbacher et al., 2002). Finally, participants were also asked to report the speed they would normally drive on five types of roads (motorway, highway, busy main street, winding country road and a residential road). The five items were standardised and combined to form a single speed item. Engagement in road rage behaviours and road rage victimisation were measured using Smart et al. (2005) scales. The first scale consisted of four questions which asked participants to report how frequently they had been the victims of four increasingly severe road rage behaviours over the past 12 months. The second scale used the same four items to ask how frequently the participants themselves had engaged in these road rage behaviours during the past 12 months. The four items from each scale were combined to form a road rage victimisation scale and a road rage perpetration scale, as per Smart et al. 2.2. Procedure Participants were selected at random from the electoral rolls in four major New Zealand cities (Auckland, Christchurch, Wellington & Palmerston North), meaning that mainly urban residents were selected for inclusion in the survey. A random number generator was used to select page numbers from which to obtain names and addresses and was again used to select five individuals from each selected page. Each selected individual was sent a package containing a cover letter, questionnaire and reply paid envelop. 2.3. Participants Of the 2300 questionnaires posted, 130 were returned (address unknown), while 677 completed questionnaires were returned, giving a response rate of 31.2%. Respondents had an average age of 49 and 46.6% were female. The average annual mileage was 13,600 km/year and the respondents had an average of 30 years driving experience. In total, 6.6% of the respondents had been involved in a crash during the last 3 months and 7.3% had been issued with a ticket of some kind (other than a parking ticket). 3. Results Table 1 shows the means and standard deviations for the 49 DAX items. This shows that the three most commonly reported responses were: I tell myself it is not worth getting involved in (2.98 ± 0.95); I decide not to stoop to their level (2.87 ± 1.02); I just try to accept that there are frustrating situations while driving (2.82 ± 0.91). In fact nine of the top ten items were from the Adaptive/Constructive subscale. The three least frequently reported responses were I try to get out of the car and have a physical fight (1.02 ± 0.17); I try to force the other driver to the side of the road (1.03 ± 0.19); I try to get out of my car to tell the other driver off (1.03 ± 0.25). These three were all from the Personal Physical Aggressive Expression subscale. Responses to the 49-items were subjected to Confirmatory factor analysis using EQS v 6.1 for windows (Bentler, 2005). The fourfactor structure by Deffenbacher et al. (2002) was tested using the Robust Maximum Likelihood (ML) method. This method accounts for non-normal multivariate distribution. Model fit was evaluated with the Santorra-Bentler Scaled Chi-Squared (S-Bv2), S-Bv2/df index, adjusted Comparative Fit Index (CFI) and Root Mean Square Error of Approximation (RMSEA). Acceptable model fit is traditionally indicated by df index < 5, an adjusted CFI of 0.90 or greater, an RMSEA of 0.06 or less and a confidence interval (C.I.) reporting a 90% interval surrounding the RMSEA acceptable level < 0.05
155
(Browne and Cudeck, 1993). As a large proportion of cases (25%) had one missing value, variable mean imputations were conducted prior to the CFA. Twenty-seven cases with 10% or more missing data were excluded from the analysis. Thus, the CFA was conducted on a total of 607 cases. The initial four-factor model showed poor fit to the data: S-Bv2(1121) = 3230.23, p < .001; S-Bv2/df = 2.88, CFI = .60, RMSEA = 0.06 (C.I. = 0.05–0.06). When the factors and factor loadings were examined, the personal physical factor was not significant and none of the items significantly loaded on to it. This may be explained by the low means for these items. For example, the average score for the factor was only 1.09 (on a scale from 1 to 4). Given that the personal physical aggressive factor did not contribute to the model, the model was re-run with these items removed. The three-factor model also demonstrated poor fit to the data S-Bv2(662) = 2681.46, p < .001; S-Bv2/df = 4.05, CFI = .65, RMSEA = 0.07 (C.I. = 0.07–0.07). Lagrange Multiplier test revealed that a number of correlated errors could improve the model fit. When the model was respecified with 14 error pairs covaried, the goodness of fit statistics were satisfactory (see Fig. 1), S-Bv2(648) = 1251.36, p < .001; S-Bv2/df = 1.93, CFI = 0.90, RMSEA = 0.04 (C.I. = 0.07–0.07). Fig. 1 shows the final 38-item three-factor model. The model had acceptable internal consistency with a Rho of 0.74. Composite reliability was also acceptable and ranged between 0.77 and 0.87. Within method convergent validity was shown with statistically significant regression co-efficients and 84% of the factor loadings were greater than 0.40. Fig. 1 also shows the correlations amongst the different forms of anger expression. The adaptive constructive category of anger expression was negatively correlated with the two other types of expression and there was also a positive correlation between Use of the Vehicle to express anger and Verbal Aggressive Expressions of anger. As shown by the moderate size of the correlation between the two aggressive responses to anger (.485), these factors are related but clearly different forms of anger expression. As there were no age differences by sex, simple t-tests were used to investigate sex differences in the three factors. This found that males were slightly more likely to engage in Verbal Aggressive Expression and to Use a Vehicle to express their anger, while females were more likely to respond in an Adaptive/Constructive manner (Table 2). However, if Bonferroni corrections are used to account for multiple comparisons, the sex differences on the Verbal Aggressive Expression is no longer significant. Furthermore, the effect size for Verbal Aggressive Expression was below Cohen’s (1988) criteria for small (0.2), indicating that this was not a meaningful effect. However, the other two effect sizes (0.27 and 0.34) were in the small to medium range. Table 3 shows the correlations the types of anger expression have with the other continuous variables. This shows that almost all of the correlations fall into the small to medium categories outlined by Cohen (1988): small (r = 0.10), medium (r = 0.30) and large (r = 0.50). The verbal aggressive factor had positive correlations with the DAS, trait anger, annual mileage, preferred driving speed, violations, loss of control, loss of concentration and near misses. This factor also had significant negative correlations with age and experience. Similarly, use of a vehicle to express anger had significant positive correlations with the DAS, trait anger, annual mileage, speed, violations, loss of control, loss of concentration, near misses and major crashes. As with the previous factor, UOV also had negative correlations with age and experience. As would be expected, the adaptive constructive factor was negatively correlated with the DAS, trait anger, mileage, speed and violations. In addition, more experienced and older drivers were more likely to respond to potentially anger inducing situations in an adaptive or constructive manner.
156
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Table 1 Means and standard deviations for the DAX items and factors. DAX items
M (SD)
Adaptive/Constructive expression 35 34 40 37 31 7 26 39 13 19 20 47 49 45 46
I I I I I I I I I I I I I I I
tell myself it is not worth getting involved in decide not to stoop to their level just try to accept that there are frustrating situations while driving just try to accept that there are bad drivers on the road tell myself it is not worth getting all mad about pay even closer attention to being a safe driver pay even closer attention to other’s driving to avoid accidents tell myself to ignore it think things through before I respond try to think of positive solutions to deal with the situation try to think of positive things to do do things like take deep breaths to calm down think about things that distract me from the frustration on the road think about things that distract me from thinking about the other driver turn on the radio or music to calm down
2.41 2.98 2.87 2.82 2.77 2.69 2.68 2.68 2.67 2.60 2.15 2.15 1.80 1.79 1.77 1.68
(.59) (.95) (1.02) (.91) (.95) (.94) (.97) (.96) (.96) (.94) (.89) (.93) (.83) (.81) (.84) (.85)
Use of a Vehicle 10 43 44 6 5 12 11 28 25 33 18
I drive a little faster than I was I flash my lights at the other driver I slow down to frustrate the other driver I purposefully block the other driver from doing what he/she wants to do Drive right up the other driver’s bumper I do to other drivers what they did to me I drive a lot faster than I was I speed up to try and frustrate the other driver I follow right behind the other driver for a long time I leave my lights on full beam when following the other driver I try to cut in front of the other driver
1.19 1.45 1.34 1.21 1.19 1.16 1.14 1.13 1.13 1.11 1.09 1.05
(.26) (.57) (.54) (.48) (.46) (.42) (.39) (.35) (.38) (.41) (.38) (.22)
Verbal Aggressive Expression 16 3 15 32 8 2 48 22 1 21 14 27
I I I I I I I I I I I I
shake my head at the other driver make negative comments about the other driver under my breath think things like ‘‘where did you get your license?’’ call the other drivers names under my breath glare at the other driver make negative comments about the other driver aloud swear at the other driver under my breath give the other driver dirty looks call the other driver names aloud swear at the other driver aloud yell questions like ‘‘Where did you get your license?’’ yell at the other driver
1.79 2.16 2.02 2.00 1.92 1.92 1.90 1.85 1.84 1.66 1.56 1.33 1.32
(.47) (.79) (.71) (.88) (.75) (.80) (.75) (.76) (.73) (.75) (.74) (.65) (.64)
Physical Aggressive Expression 23 24 30 36 29 41 38 42 9 4 17
I I I I I I I I I I I
give the other driver the finger(s) make hostile gestures other than the finger(s) roll down my window to help communicate my anger try to scare the other driver shake my fist at the other driver go crazy behind the wheel bump the other driver’s bumper with mine stick my tongue out at the other driver try to force the other driver to the side of the road try to get out of the car and tell the other driver off try to get out of the car and have a physical fight with the other driver
1.09 1.23 1.12 1.11 1.11 1.09 1.09 1.04 1.04 1.03 1.03 1.02
(.23) (.53) (.38) (.42) (.45) (.36) (.38) (.30) (.21) (.19) (.25) (.17)
The DAS and TAS were both related to road rage victimization and perpetration. Furthermore, road rage victimisation showed a significant positive relationship with both Use of a Vehicle and Verbal Aggressive Expression but no correlation with the Adaptive/Constructive factor. On the other hand, the road rage perpetration scale showed a significant negative correlation with the Adaptive/Constructive subscale. As would be expected, the road rage perpetration scale also had significant positive correlations with Use of a Vehicle and Verbal Aggressive Expression. In order to investigate whether any of the descriptive variables were able to predict the three types of driving anger expression, three multiple regressions were used. In each regression age, gender, annual mileage and reported speed were entered into the model in the first block and the DAS in the second to test whether the DAS was able to predict each of the DAX factors over and above that predicted by the descriptive variables.
Table 4 shows that age, sex and mileage were significant predictors of verbal aggressive expression. The addition of the DAS resulted in a significant improvement in the R2, meaning that the DAS remained a significant predictor of the verbal aggressive expression factor even once the contributions of the descriptive variables had been partialled out. The descriptive variables were also significant predictors of using a vehicle to express anger, but only age and sex were significant. The addition of the DAS again resulted in a significant improvement in R2, meaning that the DAS was a significant predictor of using a vehicle to express anger even after the contributions of the descriptive variables had been taken into account. The adaptive constructive expression was also predicted by the descriptive variables and again it was only age and sex which made significant contributions. However, the addition of the DAS factor did not result in a significant increase in the R2 meaning that the DAS was not able to predict the adaptive/
157
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Fig. 1. CFA model for the DAX three-factor model with standardised factor loadings (R2); CR = composite reliability. ⁄p < 0.001.
Table 2 Gender differences in the types of anger expression. Gender Measure
Verbal aggressive Expression Use of vehicle Adaptive Constructive
t
Men
p
d
0.05 0.00 0.00
.17 .34 .27
Women
M
SD
M
SD
1.84 (21.91) 1.22 (13.40) 2.33 (34.66)
0.48 (5.80) 0.26 (2.86) 0.56 (8.54)
1.76 (21.07) 1.14 (12.47) 2.48 (36.65)
0.46 (5.57) 0.21 (2.29) 0.57 (8.47)
1.95 4.52 3.11
Note: Figures in brackets are the totals, as calculated by Deffenbacher et al. (2002).
constructive expression of anger once the contributions of age and sex had been partialled out.
4. Discussion The present study investigated the factor structure of the DAX amongst a sample of New Zealand drivers. Using Confirmatory Factor Analysis this research supported a three factor solution which was broadly similar to the original four factors (Deffenbacher et al., 2002), but without the Personal Physical Aggressive Expression (PPAE) factor. The Adaptive/Constructive, Verbal Aggressive Expression and Use of a Vehicle factors found here replicated those factors found by Deffenbacher et al. However, the Physical Aggressive Expression factor was not supported amongst this sample of
New Zealand drivers. This is not the first study which has not supported the PPAE factor. For example, in a pilot study Villieux and Delhomme (2010) decided to remove the PPAE factor from their research due to French drivers reporting engagement in these behaviours at a very low level (M = 1.15) and that exploratory factor analysis did not replicate the four factor solution. The overall mean for the PPAE factor in the present study was even lower than that found in the French research (M = 1.09), with an item range of M = 1.02–M = 1.23. For example, item 17 (I try to get out of my car and have a physical fight) had a mean of 1.02 and a standard deviation of 0.17 (scale range 1–4). In contrast, research amongst male Turkish taxi drivers supported the original four factor solution and found a mean for the PPAE factor of 1.24. Therefore, perhaps this factor (and the items within this factor) is not useful when studying a sample from the general population.
158
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Table 3 Correlation of the forms of driving anger expression with the continuous variables. Measure
VAE
DAS-short TAS Age Mileage Experience Speed Violations Loss of concentration Loss of control Near misses Tickets Minor crash Major crash RR Victim RR Perp
UOV ⁄⁄⁄
.50 .45⁄⁄⁄ .35⁄⁄⁄ .23⁄⁄⁄ .29⁄⁄⁄ .27⁄⁄⁄ .41⁄⁄⁄ .10⁄ .14⁄⁄ .19⁄⁄ .00 .02 .02 .30⁄⁄⁄ .60⁄⁄⁄
AC ⁄⁄⁄
.27 .33⁄⁄⁄ .30⁄⁄⁄ .17⁄⁄ .25⁄⁄⁄ .21⁄⁄⁄ .45⁄⁄⁄ .13⁄⁄ .12⁄⁄ .12⁄⁄ .04 .03 .12⁄ .26⁄⁄⁄ .37⁄⁄⁄
DAS-short ⁄
.10 .12⁄⁄ .14⁄⁄⁄ .03 .11⁄⁄ .18⁄⁄⁄ .20⁄⁄⁄ .05 .05 .04 .03 .04 .031 .07 .26⁄⁄⁄
TAS
– .39⁄⁄⁄ .22⁄⁄⁄ .10⁄ .20⁄⁄⁄ .08⁄ .20⁄⁄⁄ .04 .06 .12⁄ .04 .02 .010 .17⁄⁄⁄ .34⁄⁄⁄
– .35⁄⁄⁄ .23⁄⁄⁄ .35⁄⁄⁄ .11⁄⁄ .33⁄⁄⁄ .24⁄⁄⁄ .19⁄⁄⁄ .23⁄⁄⁄ .07 .04 .008 .29⁄⁄⁄ .36⁄⁄⁄
VAE = Verbal aggressive expression; UOV = Use of Vehicle to express anger; AC = Adaptive constructive; DAS = Driving anger scale; TAS = Trait anger scale; RR Victim = Road rage victim, RR Perp = Road rage perpetrator. ⁄ p < 0.05. ⁄⁄ p < 0.01. ⁄⁄⁄ p < 0.001.
Table 4 Prediction of the driving anger expression factors. Step
Variable
R2
Verbal Aggressive Expression 1 Age .147 Sex Mileage 2
DAS
DAS
Adaptive Constructive 1 Age Gender Mileage 2 ⁄ ⁄⁄ ⁄⁄⁄
DAS
F Change
Beta
.147⁄⁄⁄
34.19⁄⁄⁄
.330⁄⁄⁄ .085⁄ .134⁄⁄
.186⁄⁄⁄
166.08⁄⁄⁄
.445⁄⁄⁄
.112⁄⁄⁄
25.07⁄⁄⁄
.279⁄⁄⁄ .181⁄⁄⁄ .040
.159
.047⁄⁄⁄
33.15⁄⁄⁄
.223⁄⁄⁄
.035
.035⁄⁄⁄
7.23⁄⁄⁄
.141⁄⁄ .135⁄⁄ .019
.039
.004
2.51
.066
.333
Use of a vehicle to express anger 1 Age .112 Sex Mileage 2
Change in R2
p < 0.05. p < 0.01. p < 0.001.
This study used a univariate measure of driving anger. However, research has shown there are between three and six different types of driving anger (Deffenbacher et al., 1994; Lajunen et al., 1998; Parker et al., 2002; Sullman et al., 2007). Furthermore, although Parker et al. (2002) found that some types of anger evoking situations were more likely than others to lead to some kind of reaction they did not use a scale as sophisticated as the DAX to measure reactions. Therefore, future research should use the 33-item version of the DAS to test whether any of the different types of anger evoking situations are more likely to lead to particular types of responses (e.g. is anger evoked by hostile gestures more likely to lead to Verbal Aggressive Expression?) and whether they are predictive of certain types of expression. Finding a significant relationship between road rage victimisation and the aggressive forms of driving anger expression is in agreement with Smart et al.’s (2005) research. This perhaps indicates that either those who react in an aggressive manner are more likely to become victims of road rage, or that road rage victims are more likely to engage in aggressive forms of anger expression. Surprisingly there was no relationship between road rage victimisation and the Adaptive/Constructive expression, meaning that
those who respond in this manner are no more or less likely to become the victims of road rage. There was, however, a significant negative relationship between the adaptive/constructive factor and road rage perpetration. This was to be expected, as those who deal with their anger in an adaptive/constructive way would be less likely to engage in road rage behaviours, which is a maladaptive form of driving anger expression. As would be expected, the relationships road rage perpetration had with the three types of anger expression were stronger than those for victimisation, as essentially they are measuring similar actions and behaviours. The significant relationships road rage perpetration has with Adaptive/Constructive, Verbal Aggressive Expression, and Use of a Vehicle to express anger provides additional evidence of convergent validity. The two aggressive forms of anger expression had positive correlations with mileage. This partially agrees with previous research which found mileage was only correlated with Verbal Aggressive Expression and total aggressive expression (Deffenbacher et al., 2002) or with Use of the Vehicle to express anger (Villieux and Delhomme, 2010). This also seems to suggest that driving more kilometres per year makes a driver more likely to respond in an aggressive manner, or that those who drive more often are more prone to responding in an aggressive manner. However, when entered into the regression equation with age and sex, mileage only remained a significant predictor of Using a Vehicle to express anger. This study also found a relationship between the expression of driving anger and age. Age had negative relationships with the two aggressive types of anger expression, but was positively correlated with the Adaptive/Constructive subscale. This is largely in agreement with the research from America (e.g. Deffenbacher et al., 2007), France (Villieux and Delhomme, 2010), Turkey (Esiyok et al., 2007), Serbia (Jovanovic´ et al., 2011) and Spain (HerreroFernández, 2011). Speed was also positively correlated with the two aggressive forms of anger expression, but was negatively correlated with the Adaptive/Constructive factor. This finding is in complete agreement with the only other research to report the relationship between the DAX factors and speed, in that they found preferred speed was positively correlated with Verbal Aggressive Expression and Use of a Vehicle, but was negatively related to Adaptive/Constructive expression (Sullman et al., 2013). In support of previous research (Dahlen and Ragan, 2004; Deffenbacher et al., 2002; Esiyok et al., 2007; Jovanovic´ et al., 2011; Villieux and Delhomme, 2010) is the fact that the two aggressive forms of driving anger expression were positively related to the DAS and trait anger. Also in agreement with previous research (Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Jovanovic´ et al., 2011) the two aggressive forms of expression were related to a number of the crash related conditions (e.g. loss of concentration, near misses, loss of control). The present study also found a significant relationship between major crashes and Use of a Vehicle to express anger. Although most research has found a relationship between crash involvement and the DAX subscales (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2002) it has not been consistently the same factors. For example, Dahlen and Ragan (2004) found that both major and minor accidents were positively related to Personal Physical Aggressive Expression and negatively related to the Adaptive/Constructive subscale. However, Deffenbacher et al. (2002) found major accidents to be correlated with Verbal Aggressive Expression, Use of a Vehicle and total aggressive expression. On the other hand, Deffenbacher et al. (2004) found it was only Personal Physical Aggressive Expression that was related to major accidents. However, clearly the expression of driving anger is in some way related to crash involvement. Perhaps the inconsistency is related to the relatively short periods of time across which accidents were measured (3 months) or the
M.J.M. Sullman / Safety Science 72 (2015) 153–159
mostly small samples, considering accidents are rare events, and the relatively homogeneous samples used in most of the research. This study also supports the finding that female drivers express driving anger in an Adaptive/Constructive manner more often than male drivers (Deffenbacher et al., 2004; Esiyok et al., 2007). Furthermore, Use of a Vehicle to express anger was significantly higher amongst male drivers, while Verbal Aggressive Expression did not significantly differ between male and female drivers (after Bonferroni corrections were made). Although this contrasts with Deffenbacher et al.’s (2002) findings, it is in line with Deffenbacher et al.’s (2004) findings. The current study has also shown that various descriptive variables were significant predictors of the forms of anger expression. Interestingly age was a significant predictor of all three factors. Therefore it seems that younger drivers are most likely to become angry (Sullman, 2006; Sullman et al., 2007) and to react in an aggressive manner, while the opposite is true for the Adaptive/ Constructive subscale. Sex was found to be a significant predictor of Use of a Vehicle to express anger and Adaptive/Constructive expression. Furthermore, mileage was also found to be a significant predictor of Verbal Aggressive Expression. In summary this research found that the expression of driving anger can be placed into three different categories (Adaptive/Constructive, Verbal Aggressive Expression and Use of a Vehicle). Although these three factors were precisely the same as those found in the original research, the PAE factor was not supported using data from New Zealand drivers. The present research found that all three types of anger expression could be predicted using descriptive variables. The Verbal Aggressive Expression was predicted by age, mileage and sex; Use of a Vehicle by age and sex; while Adaptive/Constructive expression was predicted by age and sex. Finally, the current study also confirmed a number of the finding of previous research using a broader sample of drivers than in most previous research. Largely the findings reported here are consistent with those in the five other countries which have investigated the expression of driving anger using the DAX (i.e., America, France, Serbia, Spain and Turkey), despite the different sampling methods. As these findings were made using a sample of drivers randomly selected from the general population, we can have more confidence in the external validity of the findings made here, than in the previous research which mostly relied on student samples. References Browne, M.W., Cudeck, R., 1993. Alternative ways of assessing model fit. In: Bollen, K.A., Long, J.S. (Eds.), Testing Structural Equation Models. Sage, Beverly Hills, CA, pp. 136–162. Cohen, J., 1988. Statistical power analysis for the behavioral sciences, second ed. Lawrence Erlbaum, New Jersey. Dahlen, E.R., Ragan, K.M., 2004. Validation of the propensity for angry driving scale. J. Saf. Res. 35, 557–563. Deffenbacher, J.L., Deffenbacher, D.M., Lynch, R.S., Richards, T.L., 2003a. Anger, aggression, and risky behaviour, a comparison of high and low anger drivers. Behav. Res. Ther. 41, 701–718.
159
Deffenbacher, J.L., Lynch, R.S., Deffenbacher, D.M., Oetting, E.R., 2001a. Further evidence of the reliability and validity for the Driving Anger Expression Inventory. Psychol. Rep. 89, 535–540. Deffenbacher, J.L., Lynch, R.S., Filetti, L.B., Dahlen, E.R., Oetting, E.R., 2003b. Anger, aggression, risky behavior, and crash-related outcomes in three groups of drivers. Behav. Res. Ther. 41, 333–349. Deffenbacher, J.L., Lynch, R.S., Oetting, E.R., Swaim, R.C., 2002. The Driving Anger Expression Inventory: a measure of how people express their anger on the road. Behav. Res. Ther. 40, 717–737. Deffenbacher, J.L., Lynch, R.S., Oetting, E.R., Yingling, D.A., 2001b. Driving anger: correlates of a test of state-trait theory. Personality Individ. Differ. 31, 1321– 1331. Deffenbacher, J.L., Oetting, E.R., Lynch, R.S., 1994. Development of a driver anger scale. Psychol. Rep. 74, 83–91. Deffenbacher, J.L., White, G.S., Lynch, R.S., 2004. Evaluation of two new scales assessing driving anger: The Driving Anger Expression Inventory and the Driver’s Angry Thoughts Questionnaire. J. Psychopathol. Behavior. Assess. 26 (2), 87–99. Deffenbacher, J.L., Kemper, C.C., Richards, T.L., 2007. The driving anger expression inventory: A validity study with community college student drivers. J. Psychopathol. Behavior. Assess. 29 (4), 220–230. Esiyok, B., Yasak, Y., Korkusuz, I., 2007. Anger expression on the road: validity and reliability of the driving anger expression inventory. Turkish J. Psychiatry 18 (3), 231–243. Herrero-Fernández, D., 2011. Psychometric adaptation of the Driving Anger Expression Inventory in a Spanish sample: differences by age and gender. Transp. Res. Part F 14, 324–329. Jovanovic´, D., Lipovac, K., Stanojevic´, P., Stanojevic´, D., 2011. The effects of personality traits on driving-related anger and aggressive behaviour in traffic among Serbian drivers. Transp. Res. Part F 14, 43–53. Lajunen, T., Parker, D., Stradling, S.G., 1998. Dimensions of driver anger, aggressive and highway code violations and their mediation by safety orientation in UK drivers. Transp. Res. Part F 1, 107–121. Lawton, R., Nutter, A., 2002. A comparison of reported levels and expression of anger in everyday and driving situations. Br. J. Psychol. 93, 407–423. Maxwell, J.P., Grant, S., Lipkin, S., 2005. Further validation of the propensity for angry driving scale in British drivers. Personality Individ. Differ. 38, 213–224. Moore, M., Dahlen, E.R., 2008. Forgiveness and consideration of future consequences in aggressive driving. Accid. Anal. Prev. 40, 1661–1666. Parker, D., Lajunen, T., Summala, H., 2002. Anger and aggression among drivers in three European countries. Accid. Anal. Prev. 34, 229–235. Reason, J., Manstead, A., Stradling, S., Baxter, J., Campbell, K., 1990. Errors and violations on the roads, a real distinction? Ergonomics 33, 1315–1332. Smart, R.G., Mann, R.E., Zhao, J., Stodutoa, G., 2005. Is road rage increasing? results of a repeated survey. J. Saf. Res. 36, 195–201. Spielberger, C.D., 1999. State-Trait Anger Expression Inventory, first ed. Psychological Assessment Resources, Odessa, FL. Stephens, A.N., Groeger, J.A., 2009. Situational specificity of trait influences on drivers’ appraisals and driving behaviour. Transp. Res. Part F 12, 29–39. Stephens, A.N., Groeger, J.A., 2011. Anger-congruent behaviour transfers across driving situations. Cogn. Emot. 25, 1423–1438. Sullman, M.J.M., 2006. Anger amongst New Zealand drivers. Transp. Res. Part F 9, 173–184. Sullman, M.J.M., Gras, M.E., Cunill, M.O., Planes, M., Font-Mayolas, S., 2007. Driving Anger in Spain. Personality Individ. Differ. 42, 701–713. Sullman, M.J.M., Stephens, A.N., 2013. A comparison of the Driving Anger Scale and the Propensity for Angry Driving Scale. Accid. Anal. Prev. 58, 88–96. http:// dx.doi.org/10.1016/j.aap.2013.05.002. Sullman, M.J.M., Stephens, A.N., Kuzu, D., 2013. The expression of driving anger amongst Turkish taxi drivers. Accid. Anal. Prev. 56, 42–50. http://dx.doi.org/ 10.1016/j.aap.2013.03.013. Underwood, G., Chapman, P., Wright, S., Crundall, D., 1999. Anger while driving. Transp. Res. Part F 2, 55–68. Villieux, A., Delhomme, P., 2010. Driving anger and its expressions: further evidence of validity and reliability for the Driving Anger Expression Inventory French adaptation. J. Saf. Res. 41, 417–422.
Contents lists available at ScienceDirect
Safety Science journal homepage: www.elsevier.com/locate/ssci
The expression of anger on the road Mark J.M. Sullman Research Group, School of Engineering, Cranfield University, MK43 0AL, United Kingdom
a r t i c l e
i n f o
Article history: Received 10 February 2012 Received in revised form 28 August 2014 Accepted 28 August 2014
Keywords: Driving anger Anger expression Driving behaviour Angry drivers Risky driving
a b s t r a c t This research used the Driving Anger Expression Inventory (DAX) to investigate the expression of driving anger amongst a sample of New Zealand drivers. Confirmatory Factor Analysis found that the data fitted a three factor solution, which consisted of: Verbal Aggressive Expression; Use of a Vehicle to express anger, and an Adaptive/Constructive factor. The Personal Physical Aggressive Expression factor was not supported in the present sample. However, in line with previous findings, the present research found that the two aggressive types of anger expression were significantly related to potentially crash related conditions, such as: losing control of the vehicle, loss of concentration and near-misses. The two aggressive types of driving anger were also positively related to trait anger and driving anger, while the Adaptive/ Constructive factor was negatively related to both these variables. The present study also confirmed that those drivers who expressed their anger in an aggressive manner are more likely to be male and younger, while those who deal with their anger in a constructive manner were more likely to be older and female. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Anger is one emotion that is commonly experienced while driving (Underwood et al., 1999). Although research has found that individuals are not more likely to experience anger while driving, than in non-driving situations, those individuals who reported high levels of anger are more likely to be outwardly aggressive when driving (Lawton and Nutter, 2002). This is perhaps one of the reasons that the issue of driving anger appears to have gained popularity amongst researchers over the last 15 years (e.g. Deffenbacher et al., 2001b; Maxwell et al., 2005; Parker et al., 2002; Underwood et al., 1999; Sullman, 2006; Sullman et al., 2007; Sullman and Stephens; 2013). Another reason for the increasing popularity of this topic is that research has found that angry drivers engage more often in aggressive and dangerous driving behaviours (Stephens and Groeger, 2011), putting themselves and others at risk (Deffenbacher et al., 1994; Deffenbacher et al., 2001b; Lajunen et al., 1998; Maxwell et al., 2005; Underwood et al., 1999). Research has also found significant relationships to exist between driving anger and crash related conditions, such as losing control of their vehicle, losing concentration, speeding, tailgating, near misses, and moving violations (tickets) (Deffenbacher et al., 2001b; Deffenbacher et al., 2003a; Deffenbacher et al., 2003b; Sullman et al., 2013). Although a number of studies have investigated the situations which evoke anger amongst drivers (e.g. Deffenbacher et al., E-mail address: m.sullman@cranfield.ac.uk http://dx.doi.org/10.1016/j.ssci.2014.08.013 0925-7535/Ó 2014 Elsevier Ltd. All rights reserved.
1994; Lajunen et al., 1998; Parker et al., 2002; Stephens and Groeger, 2009; Sullman and Stephens, 2013), much less research has investigated how drivers react when angry. In order to measure the ways in which individuals respond to anger while driving Deffenbacher et al. (2002) developed the Driving Anger Expression Inventory (DAX). They found that drivers’ reactions could be placed into four categories: Verbal Aggressive Expression (VAE) – which measures people’s tendency to express their anger through verbally aggressive means (e.g. swearing at the other driver); Personal Physical Aggressive Expression (PAE) – which measures the ways in which the person uses themselves to express anger (e.g. shaking their fist); Use of Vehicle to express anger (UOV) – which measures how often drivers use their vehicle to express their anger (e.g. flashing their lights); Adaptive/Constructive Expression (A/C) – which measures constructive or adaptive behaviours the driver can make in potentially anger inducing situations (e.g. just ignore it). These four subscales have been found to have good internal reliability (range 0.80–0.90). Although the four DAX subscales have good internal reliability, little research has factor analysed the scale. As few peer-reviewed studies have reported factor analysing the 49-item version of the DAX, there is some degree of uncertainty regarding the scale’s underlying structure. Although the study which developed the DAX (Deffenbacher et al., 2002) subjected their data to factor analysis, this was a 62-item scale which was subsequently refined. Most of the following studies using the scale have used the item arrangement reported in the original research and have not subjected their data to any type of factor analysis (e.g. Dahlen and
154
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2003; Jovanovic´ et al., 2011; Moore and Dahlen, 2008), or have factor analysed a slightly different version of the DAX (e.g. HerreroFernández, 2011; Villieux and Delhomme, 2010). For example, Villieux and Delhomme (2010) removed all items from the Personal Physical Aggression subscale (11-items) before attempting to fit a three factor solution to their data using CFA. Furthermore, Herrero-Fernández (2011) used a 53-item version of the DAX and fitted a five factor model using CFA. Lastly, Sullman et al. (2013) performed a CFA on the 49-items using Turkish taxi drivers and confirmed the four factor solution, after removing one item from the model. Therefore, although previous research has broadly supported the factor structure of the DAX, the evidence is not extensive. Another issue surrounding research into the expression of driving anger is that most peer-reviewed research has focused mainly or exclusively on introductory psychology students (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2002; Deffenbacher et al., 2003; Deffenbacher et al., 2004; Deffenbacher et al., 2007; Esiyok et al., 2007; Herrero-Fernández, 2011; Moore and Dahlen, 2008; Villieux and Delhomme, 2010) who are not necessarily representative of the general population. Only two of the DAX studies did not rely on university students for their sample (Jovanovic´ et al., 2011; Sullman et al., 2013). Furthermore, most of the research has been conducted in the USA (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2001a; Deffenbacher et al., 2002; Deffenbacher et al., 2003; Deffenbacher et al., 2004; Deffenbacher et al., 2007; Moore and Dahlen, 2008). There are a small number of studies which have been conducted outside of America, including France (Villieux and Delhomme, 2010), Spain (Herrero-Fernández, 2011), Turkey (Esiyok et al., 2007; Sullman et al., 2013) and Serbia (Jovanovic´ et al., 2011). As research has shown the level of driving anger differs between countries, as does the level of anger evoked by different types of situations (e.g. Deffenbacher et al., 1994; Lajunen et al., 1998; Sullman, 2006; Sullman et al., 2007), it is important to study additional countries. Given the fact that most research has largely relied on university students (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Esiyok et al., 2007; Herrerro-Fernandez, 2011; Villieux and Delhomme, 2010) it is surprising that most research has found the expression of driving anger to differ with age. For example, using a sample of Serbian motor vehicle owners, Jovanovic´ et al. (2011) found a significant negative relationship between age and the total of all the aggressive expression items. In terms of the individual DAX subscales, age was found to be negatively related to Personal Physical Aggressive Expression, Use of a Vehicle and Verbal Aggressive Expression, while Adaptive/Constructive expression was positively related to age (Deffenbacher et al., 2007). Furthermore, despite having a very narrow age range (18–25 years old), using a sample of French university students Villieux and Delhomme (2010) found age was negatively related to Using a Vehicle to express anger and positively related to the Adaptive/ Constructive subscale. In addition, in a sample consisting of almost 75% university students, Esiyok et al. (2007) found that age was a significant predictor of two of the aggressive subscales. Furthermore, despite having a sample where 75% of the participants were 30 years old or younger, Herrerro-Fernandez (2011) also found significant differences by age for all but the Adaptive/Constructive subscale. In contrast, research by Moore and Dahlen (2008) found age was not a predictor of any of the DAX subscales. Perhaps the reason for the lack of consistency in the age differences is due to the relatively narrow age spreads in most of the previous research. As with the research on driving anger, most research on the expression of driving anger has found sex differences. Several studies have found that females reported responding more often in an Adaptive/Constructive manner (e.g. Deffenbacher et al., 2004;
Esiyok et al., 2007; Jovanovic´ et al., 2011). Several studies have also found that males report more Personal Physical Aggressive Expression (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Esiyok et al., 2007). There were also some inconsistent findings, such as the fact that only Deffenbacher et al. (2004) found males reported Use of a Vehicle to express their anger more often than women and Dahlen and Ragan (2004) found that women reported more Verbal Aggressive Expression than men. Furthermore, there have also been a number of studies that have reported no sex differences (e.g. Deffenbacher et al., 2007; Moore and Dahlen, 2008; Villieux and Delhomme, 2010). In addition to demographic variables, the DAX has also been found to be related to other important variables. As would be expected, trait anger and driving anger have been found to be positively correlated with the three aggressive types of anger expression and negatively correlated with Adaptive/Constructive expression (Deffenbacher et al., 2001a). Furthermore, the three aggressive types of anger expression have been found to be positively correlated with risky and aggressive driving behaviour (e.g. Deffenbacher et al., 2002; Jovanovic´ et al., 2011). Verbal Aggressive Expression and Use of a Vehicle to express anger both had significant correlations with self-reported involvement in major accidents. Minor accidents, however, were most strongly correlated with Verbal Aggressive Expression of driving anger, as were close calls (Deffenbacher et al., 2001a). In addition, UOV and PAE were correlated with the most crash related to conditions, including risky behaviour and lost concentration (Deffenbacher et al., 2002). Given the relationships the DAX subscales have with variables important to the reduction of traffic crashes, the expression of driving anger appears to be an area worthy of further research. Therefore, the present study aimed to broaden the knowledge regarding the expression of driving anger in a number of ways. Firstly, the research explored the factor structure of the DAX amongst a sample taken from the general population of New Zealand, a country not previously studied. Furthermore, the relationship the resultant factors had with a number of descriptive variables (e.g. age, gender, experience, speed and annual mileage), the 14-item unidimensional version of the DAS, the TAS and a number of driving related outcomes (e.g. crash involvement, tickets) were also tested. Finally the descriptive variables and the DAS were also used to investigate whether they were significant predictors of the DAX factors.
2. Method 2.1. Materials The 49-item DAX was used to measure how drivers express their anger whilst driving (Deffenbacher et al., 2002). The scale presents 49 potential reactions to feeling angry while driving and respondents were asked to report how often they react in each of these ways on a four point scale (1 = Almost never, 4 = Almost always). The 14-item Driving Anger Scale (DAS) was used to provide an overall measure of driving anger (Deffenbacher et al., 1994). The scale presents 14 different situations and asks participants to rate how angry each situation makes them feel on a scale ranging from 1 (Not at all) to 5 (Very much). The 10-item Trait Anger Scale (TAS) was also used to measure general anger (Spielberger, 1999). The TAS measures how an individual generally responds when angry and is rated on a 4-point scale (1 = Almost never, 4 = Almost always). The eight violation items from the Driving Behaviour Questionnaire (DBQ; Reason et al., 1990) were used to measure risky driving
M.J.M. Sullman / Safety Science 72 (2015) 153–159
behaviour. Additional measures included: moving violations (tickets), loss of concentration, loss of control, close calls, minor accidents and major accidents in the last 3 months (Deffenbacher et al., 2002). Finally, participants were also asked to report the speed they would normally drive on five types of roads (motorway, highway, busy main street, winding country road and a residential road). The five items were standardised and combined to form a single speed item. Engagement in road rage behaviours and road rage victimisation were measured using Smart et al. (2005) scales. The first scale consisted of four questions which asked participants to report how frequently they had been the victims of four increasingly severe road rage behaviours over the past 12 months. The second scale used the same four items to ask how frequently the participants themselves had engaged in these road rage behaviours during the past 12 months. The four items from each scale were combined to form a road rage victimisation scale and a road rage perpetration scale, as per Smart et al. 2.2. Procedure Participants were selected at random from the electoral rolls in four major New Zealand cities (Auckland, Christchurch, Wellington & Palmerston North), meaning that mainly urban residents were selected for inclusion in the survey. A random number generator was used to select page numbers from which to obtain names and addresses and was again used to select five individuals from each selected page. Each selected individual was sent a package containing a cover letter, questionnaire and reply paid envelop. 2.3. Participants Of the 2300 questionnaires posted, 130 were returned (address unknown), while 677 completed questionnaires were returned, giving a response rate of 31.2%. Respondents had an average age of 49 and 46.6% were female. The average annual mileage was 13,600 km/year and the respondents had an average of 30 years driving experience. In total, 6.6% of the respondents had been involved in a crash during the last 3 months and 7.3% had been issued with a ticket of some kind (other than a parking ticket). 3. Results Table 1 shows the means and standard deviations for the 49 DAX items. This shows that the three most commonly reported responses were: I tell myself it is not worth getting involved in (2.98 ± 0.95); I decide not to stoop to their level (2.87 ± 1.02); I just try to accept that there are frustrating situations while driving (2.82 ± 0.91). In fact nine of the top ten items were from the Adaptive/Constructive subscale. The three least frequently reported responses were I try to get out of the car and have a physical fight (1.02 ± 0.17); I try to force the other driver to the side of the road (1.03 ± 0.19); I try to get out of my car to tell the other driver off (1.03 ± 0.25). These three were all from the Personal Physical Aggressive Expression subscale. Responses to the 49-items were subjected to Confirmatory factor analysis using EQS v 6.1 for windows (Bentler, 2005). The fourfactor structure by Deffenbacher et al. (2002) was tested using the Robust Maximum Likelihood (ML) method. This method accounts for non-normal multivariate distribution. Model fit was evaluated with the Santorra-Bentler Scaled Chi-Squared (S-Bv2), S-Bv2/df index, adjusted Comparative Fit Index (CFI) and Root Mean Square Error of Approximation (RMSEA). Acceptable model fit is traditionally indicated by df index < 5, an adjusted CFI of 0.90 or greater, an RMSEA of 0.06 or less and a confidence interval (C.I.) reporting a 90% interval surrounding the RMSEA acceptable level < 0.05
155
(Browne and Cudeck, 1993). As a large proportion of cases (25%) had one missing value, variable mean imputations were conducted prior to the CFA. Twenty-seven cases with 10% or more missing data were excluded from the analysis. Thus, the CFA was conducted on a total of 607 cases. The initial four-factor model showed poor fit to the data: S-Bv2(1121) = 3230.23, p < .001; S-Bv2/df = 2.88, CFI = .60, RMSEA = 0.06 (C.I. = 0.05–0.06). When the factors and factor loadings were examined, the personal physical factor was not significant and none of the items significantly loaded on to it. This may be explained by the low means for these items. For example, the average score for the factor was only 1.09 (on a scale from 1 to 4). Given that the personal physical aggressive factor did not contribute to the model, the model was re-run with these items removed. The three-factor model also demonstrated poor fit to the data S-Bv2(662) = 2681.46, p < .001; S-Bv2/df = 4.05, CFI = .65, RMSEA = 0.07 (C.I. = 0.07–0.07). Lagrange Multiplier test revealed that a number of correlated errors could improve the model fit. When the model was respecified with 14 error pairs covaried, the goodness of fit statistics were satisfactory (see Fig. 1), S-Bv2(648) = 1251.36, p < .001; S-Bv2/df = 1.93, CFI = 0.90, RMSEA = 0.04 (C.I. = 0.07–0.07). Fig. 1 shows the final 38-item three-factor model. The model had acceptable internal consistency with a Rho of 0.74. Composite reliability was also acceptable and ranged between 0.77 and 0.87. Within method convergent validity was shown with statistically significant regression co-efficients and 84% of the factor loadings were greater than 0.40. Fig. 1 also shows the correlations amongst the different forms of anger expression. The adaptive constructive category of anger expression was negatively correlated with the two other types of expression and there was also a positive correlation between Use of the Vehicle to express anger and Verbal Aggressive Expressions of anger. As shown by the moderate size of the correlation between the two aggressive responses to anger (.485), these factors are related but clearly different forms of anger expression. As there were no age differences by sex, simple t-tests were used to investigate sex differences in the three factors. This found that males were slightly more likely to engage in Verbal Aggressive Expression and to Use a Vehicle to express their anger, while females were more likely to respond in an Adaptive/Constructive manner (Table 2). However, if Bonferroni corrections are used to account for multiple comparisons, the sex differences on the Verbal Aggressive Expression is no longer significant. Furthermore, the effect size for Verbal Aggressive Expression was below Cohen’s (1988) criteria for small (0.2), indicating that this was not a meaningful effect. However, the other two effect sizes (0.27 and 0.34) were in the small to medium range. Table 3 shows the correlations the types of anger expression have with the other continuous variables. This shows that almost all of the correlations fall into the small to medium categories outlined by Cohen (1988): small (r = 0.10), medium (r = 0.30) and large (r = 0.50). The verbal aggressive factor had positive correlations with the DAS, trait anger, annual mileage, preferred driving speed, violations, loss of control, loss of concentration and near misses. This factor also had significant negative correlations with age and experience. Similarly, use of a vehicle to express anger had significant positive correlations with the DAS, trait anger, annual mileage, speed, violations, loss of control, loss of concentration, near misses and major crashes. As with the previous factor, UOV also had negative correlations with age and experience. As would be expected, the adaptive constructive factor was negatively correlated with the DAS, trait anger, mileage, speed and violations. In addition, more experienced and older drivers were more likely to respond to potentially anger inducing situations in an adaptive or constructive manner.
156
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Table 1 Means and standard deviations for the DAX items and factors. DAX items
M (SD)
Adaptive/Constructive expression 35 34 40 37 31 7 26 39 13 19 20 47 49 45 46
I I I I I I I I I I I I I I I
tell myself it is not worth getting involved in decide not to stoop to their level just try to accept that there are frustrating situations while driving just try to accept that there are bad drivers on the road tell myself it is not worth getting all mad about pay even closer attention to being a safe driver pay even closer attention to other’s driving to avoid accidents tell myself to ignore it think things through before I respond try to think of positive solutions to deal with the situation try to think of positive things to do do things like take deep breaths to calm down think about things that distract me from the frustration on the road think about things that distract me from thinking about the other driver turn on the radio or music to calm down
2.41 2.98 2.87 2.82 2.77 2.69 2.68 2.68 2.67 2.60 2.15 2.15 1.80 1.79 1.77 1.68
(.59) (.95) (1.02) (.91) (.95) (.94) (.97) (.96) (.96) (.94) (.89) (.93) (.83) (.81) (.84) (.85)
Use of a Vehicle 10 43 44 6 5 12 11 28 25 33 18
I drive a little faster than I was I flash my lights at the other driver I slow down to frustrate the other driver I purposefully block the other driver from doing what he/she wants to do Drive right up the other driver’s bumper I do to other drivers what they did to me I drive a lot faster than I was I speed up to try and frustrate the other driver I follow right behind the other driver for a long time I leave my lights on full beam when following the other driver I try to cut in front of the other driver
1.19 1.45 1.34 1.21 1.19 1.16 1.14 1.13 1.13 1.11 1.09 1.05
(.26) (.57) (.54) (.48) (.46) (.42) (.39) (.35) (.38) (.41) (.38) (.22)
Verbal Aggressive Expression 16 3 15 32 8 2 48 22 1 21 14 27
I I I I I I I I I I I I
shake my head at the other driver make negative comments about the other driver under my breath think things like ‘‘where did you get your license?’’ call the other drivers names under my breath glare at the other driver make negative comments about the other driver aloud swear at the other driver under my breath give the other driver dirty looks call the other driver names aloud swear at the other driver aloud yell questions like ‘‘Where did you get your license?’’ yell at the other driver
1.79 2.16 2.02 2.00 1.92 1.92 1.90 1.85 1.84 1.66 1.56 1.33 1.32
(.47) (.79) (.71) (.88) (.75) (.80) (.75) (.76) (.73) (.75) (.74) (.65) (.64)
Physical Aggressive Expression 23 24 30 36 29 41 38 42 9 4 17
I I I I I I I I I I I
give the other driver the finger(s) make hostile gestures other than the finger(s) roll down my window to help communicate my anger try to scare the other driver shake my fist at the other driver go crazy behind the wheel bump the other driver’s bumper with mine stick my tongue out at the other driver try to force the other driver to the side of the road try to get out of the car and tell the other driver off try to get out of the car and have a physical fight with the other driver
1.09 1.23 1.12 1.11 1.11 1.09 1.09 1.04 1.04 1.03 1.03 1.02
(.23) (.53) (.38) (.42) (.45) (.36) (.38) (.30) (.21) (.19) (.25) (.17)
The DAS and TAS were both related to road rage victimization and perpetration. Furthermore, road rage victimisation showed a significant positive relationship with both Use of a Vehicle and Verbal Aggressive Expression but no correlation with the Adaptive/Constructive factor. On the other hand, the road rage perpetration scale showed a significant negative correlation with the Adaptive/Constructive subscale. As would be expected, the road rage perpetration scale also had significant positive correlations with Use of a Vehicle and Verbal Aggressive Expression. In order to investigate whether any of the descriptive variables were able to predict the three types of driving anger expression, three multiple regressions were used. In each regression age, gender, annual mileage and reported speed were entered into the model in the first block and the DAS in the second to test whether the DAS was able to predict each of the DAX factors over and above that predicted by the descriptive variables.
Table 4 shows that age, sex and mileage were significant predictors of verbal aggressive expression. The addition of the DAS resulted in a significant improvement in the R2, meaning that the DAS remained a significant predictor of the verbal aggressive expression factor even once the contributions of the descriptive variables had been partialled out. The descriptive variables were also significant predictors of using a vehicle to express anger, but only age and sex were significant. The addition of the DAS again resulted in a significant improvement in R2, meaning that the DAS was a significant predictor of using a vehicle to express anger even after the contributions of the descriptive variables had been taken into account. The adaptive constructive expression was also predicted by the descriptive variables and again it was only age and sex which made significant contributions. However, the addition of the DAS factor did not result in a significant increase in the R2 meaning that the DAS was not able to predict the adaptive/
157
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Fig. 1. CFA model for the DAX three-factor model with standardised factor loadings (R2); CR = composite reliability. ⁄p < 0.001.
Table 2 Gender differences in the types of anger expression. Gender Measure
Verbal aggressive Expression Use of vehicle Adaptive Constructive
t
Men
p
d
0.05 0.00 0.00
.17 .34 .27
Women
M
SD
M
SD
1.84 (21.91) 1.22 (13.40) 2.33 (34.66)
0.48 (5.80) 0.26 (2.86) 0.56 (8.54)
1.76 (21.07) 1.14 (12.47) 2.48 (36.65)
0.46 (5.57) 0.21 (2.29) 0.57 (8.47)
1.95 4.52 3.11
Note: Figures in brackets are the totals, as calculated by Deffenbacher et al. (2002).
constructive expression of anger once the contributions of age and sex had been partialled out.
4. Discussion The present study investigated the factor structure of the DAX amongst a sample of New Zealand drivers. Using Confirmatory Factor Analysis this research supported a three factor solution which was broadly similar to the original four factors (Deffenbacher et al., 2002), but without the Personal Physical Aggressive Expression (PPAE) factor. The Adaptive/Constructive, Verbal Aggressive Expression and Use of a Vehicle factors found here replicated those factors found by Deffenbacher et al. However, the Physical Aggressive Expression factor was not supported amongst this sample of
New Zealand drivers. This is not the first study which has not supported the PPAE factor. For example, in a pilot study Villieux and Delhomme (2010) decided to remove the PPAE factor from their research due to French drivers reporting engagement in these behaviours at a very low level (M = 1.15) and that exploratory factor analysis did not replicate the four factor solution. The overall mean for the PPAE factor in the present study was even lower than that found in the French research (M = 1.09), with an item range of M = 1.02–M = 1.23. For example, item 17 (I try to get out of my car and have a physical fight) had a mean of 1.02 and a standard deviation of 0.17 (scale range 1–4). In contrast, research amongst male Turkish taxi drivers supported the original four factor solution and found a mean for the PPAE factor of 1.24. Therefore, perhaps this factor (and the items within this factor) is not useful when studying a sample from the general population.
158
M.J.M. Sullman / Safety Science 72 (2015) 153–159
Table 3 Correlation of the forms of driving anger expression with the continuous variables. Measure
VAE
DAS-short TAS Age Mileage Experience Speed Violations Loss of concentration Loss of control Near misses Tickets Minor crash Major crash RR Victim RR Perp
UOV ⁄⁄⁄
.50 .45⁄⁄⁄ .35⁄⁄⁄ .23⁄⁄⁄ .29⁄⁄⁄ .27⁄⁄⁄ .41⁄⁄⁄ .10⁄ .14⁄⁄ .19⁄⁄ .00 .02 .02 .30⁄⁄⁄ .60⁄⁄⁄
AC ⁄⁄⁄
.27 .33⁄⁄⁄ .30⁄⁄⁄ .17⁄⁄ .25⁄⁄⁄ .21⁄⁄⁄ .45⁄⁄⁄ .13⁄⁄ .12⁄⁄ .12⁄⁄ .04 .03 .12⁄ .26⁄⁄⁄ .37⁄⁄⁄
DAS-short ⁄
.10 .12⁄⁄ .14⁄⁄⁄ .03 .11⁄⁄ .18⁄⁄⁄ .20⁄⁄⁄ .05 .05 .04 .03 .04 .031 .07 .26⁄⁄⁄
TAS
– .39⁄⁄⁄ .22⁄⁄⁄ .10⁄ .20⁄⁄⁄ .08⁄ .20⁄⁄⁄ .04 .06 .12⁄ .04 .02 .010 .17⁄⁄⁄ .34⁄⁄⁄
– .35⁄⁄⁄ .23⁄⁄⁄ .35⁄⁄⁄ .11⁄⁄ .33⁄⁄⁄ .24⁄⁄⁄ .19⁄⁄⁄ .23⁄⁄⁄ .07 .04 .008 .29⁄⁄⁄ .36⁄⁄⁄
VAE = Verbal aggressive expression; UOV = Use of Vehicle to express anger; AC = Adaptive constructive; DAS = Driving anger scale; TAS = Trait anger scale; RR Victim = Road rage victim, RR Perp = Road rage perpetrator. ⁄ p < 0.05. ⁄⁄ p < 0.01. ⁄⁄⁄ p < 0.001.
Table 4 Prediction of the driving anger expression factors. Step
Variable
R2
Verbal Aggressive Expression 1 Age .147 Sex Mileage 2
DAS
DAS
Adaptive Constructive 1 Age Gender Mileage 2 ⁄ ⁄⁄ ⁄⁄⁄
DAS
F Change
Beta
.147⁄⁄⁄
34.19⁄⁄⁄
.330⁄⁄⁄ .085⁄ .134⁄⁄
.186⁄⁄⁄
166.08⁄⁄⁄
.445⁄⁄⁄
.112⁄⁄⁄
25.07⁄⁄⁄
.279⁄⁄⁄ .181⁄⁄⁄ .040
.159
.047⁄⁄⁄
33.15⁄⁄⁄
.223⁄⁄⁄
.035
.035⁄⁄⁄
7.23⁄⁄⁄
.141⁄⁄ .135⁄⁄ .019
.039
.004
2.51
.066
.333
Use of a vehicle to express anger 1 Age .112 Sex Mileage 2
Change in R2
p < 0.05. p < 0.01. p < 0.001.
This study used a univariate measure of driving anger. However, research has shown there are between three and six different types of driving anger (Deffenbacher et al., 1994; Lajunen et al., 1998; Parker et al., 2002; Sullman et al., 2007). Furthermore, although Parker et al. (2002) found that some types of anger evoking situations were more likely than others to lead to some kind of reaction they did not use a scale as sophisticated as the DAX to measure reactions. Therefore, future research should use the 33-item version of the DAS to test whether any of the different types of anger evoking situations are more likely to lead to particular types of responses (e.g. is anger evoked by hostile gestures more likely to lead to Verbal Aggressive Expression?) and whether they are predictive of certain types of expression. Finding a significant relationship between road rage victimisation and the aggressive forms of driving anger expression is in agreement with Smart et al.’s (2005) research. This perhaps indicates that either those who react in an aggressive manner are more likely to become victims of road rage, or that road rage victims are more likely to engage in aggressive forms of anger expression. Surprisingly there was no relationship between road rage victimisation and the Adaptive/Constructive expression, meaning that
those who respond in this manner are no more or less likely to become the victims of road rage. There was, however, a significant negative relationship between the adaptive/constructive factor and road rage perpetration. This was to be expected, as those who deal with their anger in an adaptive/constructive way would be less likely to engage in road rage behaviours, which is a maladaptive form of driving anger expression. As would be expected, the relationships road rage perpetration had with the three types of anger expression were stronger than those for victimisation, as essentially they are measuring similar actions and behaviours. The significant relationships road rage perpetration has with Adaptive/Constructive, Verbal Aggressive Expression, and Use of a Vehicle to express anger provides additional evidence of convergent validity. The two aggressive forms of anger expression had positive correlations with mileage. This partially agrees with previous research which found mileage was only correlated with Verbal Aggressive Expression and total aggressive expression (Deffenbacher et al., 2002) or with Use of the Vehicle to express anger (Villieux and Delhomme, 2010). This also seems to suggest that driving more kilometres per year makes a driver more likely to respond in an aggressive manner, or that those who drive more often are more prone to responding in an aggressive manner. However, when entered into the regression equation with age and sex, mileage only remained a significant predictor of Using a Vehicle to express anger. This study also found a relationship between the expression of driving anger and age. Age had negative relationships with the two aggressive types of anger expression, but was positively correlated with the Adaptive/Constructive subscale. This is largely in agreement with the research from America (e.g. Deffenbacher et al., 2007), France (Villieux and Delhomme, 2010), Turkey (Esiyok et al., 2007), Serbia (Jovanovic´ et al., 2011) and Spain (HerreroFernández, 2011). Speed was also positively correlated with the two aggressive forms of anger expression, but was negatively correlated with the Adaptive/Constructive factor. This finding is in complete agreement with the only other research to report the relationship between the DAX factors and speed, in that they found preferred speed was positively correlated with Verbal Aggressive Expression and Use of a Vehicle, but was negatively related to Adaptive/Constructive expression (Sullman et al., 2013). In support of previous research (Dahlen and Ragan, 2004; Deffenbacher et al., 2002; Esiyok et al., 2007; Jovanovic´ et al., 2011; Villieux and Delhomme, 2010) is the fact that the two aggressive forms of driving anger expression were positively related to the DAS and trait anger. Also in agreement with previous research (Dahlen and Ragan, 2004; Deffenbacher et al., 2004; Jovanovic´ et al., 2011) the two aggressive forms of expression were related to a number of the crash related conditions (e.g. loss of concentration, near misses, loss of control). The present study also found a significant relationship between major crashes and Use of a Vehicle to express anger. Although most research has found a relationship between crash involvement and the DAX subscales (e.g. Dahlen and Ragan, 2004; Deffenbacher et al., 2002) it has not been consistently the same factors. For example, Dahlen and Ragan (2004) found that both major and minor accidents were positively related to Personal Physical Aggressive Expression and negatively related to the Adaptive/Constructive subscale. However, Deffenbacher et al. (2002) found major accidents to be correlated with Verbal Aggressive Expression, Use of a Vehicle and total aggressive expression. On the other hand, Deffenbacher et al. (2004) found it was only Personal Physical Aggressive Expression that was related to major accidents. However, clearly the expression of driving anger is in some way related to crash involvement. Perhaps the inconsistency is related to the relatively short periods of time across which accidents were measured (3 months) or the
M.J.M. Sullman / Safety Science 72 (2015) 153–159
mostly small samples, considering accidents are rare events, and the relatively homogeneous samples used in most of the research. This study also supports the finding that female drivers express driving anger in an Adaptive/Constructive manner more often than male drivers (Deffenbacher et al., 2004; Esiyok et al., 2007). Furthermore, Use of a Vehicle to express anger was significantly higher amongst male drivers, while Verbal Aggressive Expression did not significantly differ between male and female drivers (after Bonferroni corrections were made). Although this contrasts with Deffenbacher et al.’s (2002) findings, it is in line with Deffenbacher et al.’s (2004) findings. The current study has also shown that various descriptive variables were significant predictors of the forms of anger expression. Interestingly age was a significant predictor of all three factors. Therefore it seems that younger drivers are most likely to become angry (Sullman, 2006; Sullman et al., 2007) and to react in an aggressive manner, while the opposite is true for the Adaptive/ Constructive subscale. Sex was found to be a significant predictor of Use of a Vehicle to express anger and Adaptive/Constructive expression. Furthermore, mileage was also found to be a significant predictor of Verbal Aggressive Expression. In summary this research found that the expression of driving anger can be placed into three different categories (Adaptive/Constructive, Verbal Aggressive Expression and Use of a Vehicle). Although these three factors were precisely the same as those found in the original research, the PAE factor was not supported using data from New Zealand drivers. The present research found that all three types of anger expression could be predicted using descriptive variables. The Verbal Aggressive Expression was predicted by age, mileage and sex; Use of a Vehicle by age and sex; while Adaptive/Constructive expression was predicted by age and sex. Finally, the current study also confirmed a number of the finding of previous research using a broader sample of drivers than in most previous research. Largely the findings reported here are consistent with those in the five other countries which have investigated the expression of driving anger using the DAX (i.e., America, France, Serbia, Spain and Turkey), despite the different sampling methods. As these findings were made using a sample of drivers randomly selected from the general population, we can have more confidence in the external validity of the findings made here, than in the previous research which mostly relied on student samples. References Browne, M.W., Cudeck, R., 1993. Alternative ways of assessing model fit. In: Bollen, K.A., Long, J.S. (Eds.), Testing Structural Equation Models. Sage, Beverly Hills, CA, pp. 136–162. Cohen, J., 1988. Statistical power analysis for the behavioral sciences, second ed. Lawrence Erlbaum, New Jersey. Dahlen, E.R., Ragan, K.M., 2004. Validation of the propensity for angry driving scale. J. Saf. Res. 35, 557–563. Deffenbacher, J.L., Deffenbacher, D.M., Lynch, R.S., Richards, T.L., 2003a. Anger, aggression, and risky behaviour, a comparison of high and low anger drivers. Behav. Res. Ther. 41, 701–718.
159
Deffenbacher, J.L., Lynch, R.S., Deffenbacher, D.M., Oetting, E.R., 2001a. Further evidence of the reliability and validity for the Driving Anger Expression Inventory. Psychol. Rep. 89, 535–540. Deffenbacher, J.L., Lynch, R.S., Filetti, L.B., Dahlen, E.R., Oetting, E.R., 2003b. Anger, aggression, risky behavior, and crash-related outcomes in three groups of drivers. Behav. Res. Ther. 41, 333–349. Deffenbacher, J.L., Lynch, R.S., Oetting, E.R., Swaim, R.C., 2002. The Driving Anger Expression Inventory: a measure of how people express their anger on the road. Behav. Res. Ther. 40, 717–737. Deffenbacher, J.L., Lynch, R.S., Oetting, E.R., Yingling, D.A., 2001b. Driving anger: correlates of a test of state-trait theory. Personality Individ. Differ. 31, 1321– 1331. Deffenbacher, J.L., Oetting, E.R., Lynch, R.S., 1994. Development of a driver anger scale. Psychol. Rep. 74, 83–91. Deffenbacher, J.L., White, G.S., Lynch, R.S., 2004. Evaluation of two new scales assessing driving anger: The Driving Anger Expression Inventory and the Driver’s Angry Thoughts Questionnaire. J. Psychopathol. Behavior. Assess. 26 (2), 87–99. Deffenbacher, J.L., Kemper, C.C., Richards, T.L., 2007. The driving anger expression inventory: A validity study with community college student drivers. J. Psychopathol. Behavior. Assess. 29 (4), 220–230. Esiyok, B., Yasak, Y., Korkusuz, I., 2007. Anger expression on the road: validity and reliability of the driving anger expression inventory. Turkish J. Psychiatry 18 (3), 231–243. Herrero-Fernández, D., 2011. Psychometric adaptation of the Driving Anger Expression Inventory in a Spanish sample: differences by age and gender. Transp. Res. Part F 14, 324–329. Jovanovic´, D., Lipovac, K., Stanojevic´, P., Stanojevic´, D., 2011. The effects of personality traits on driving-related anger and aggressive behaviour in traffic among Serbian drivers. Transp. Res. Part F 14, 43–53. Lajunen, T., Parker, D., Stradling, S.G., 1998. Dimensions of driver anger, aggressive and highway code violations and their mediation by safety orientation in UK drivers. Transp. Res. Part F 1, 107–121. Lawton, R., Nutter, A., 2002. A comparison of reported levels and expression of anger in everyday and driving situations. Br. J. Psychol. 93, 407–423. Maxwell, J.P., Grant, S., Lipkin, S., 2005. Further validation of the propensity for angry driving scale in British drivers. Personality Individ. Differ. 38, 213–224. Moore, M., Dahlen, E.R., 2008. Forgiveness and consideration of future consequences in aggressive driving. Accid. Anal. Prev. 40, 1661–1666. Parker, D., Lajunen, T., Summala, H., 2002. Anger and aggression among drivers in three European countries. Accid. Anal. Prev. 34, 229–235. Reason, J., Manstead, A., Stradling, S., Baxter, J., Campbell, K., 1990. Errors and violations on the roads, a real distinction? Ergonomics 33, 1315–1332. Smart, R.G., Mann, R.E., Zhao, J., Stodutoa, G., 2005. Is road rage increasing? results of a repeated survey. J. Saf. Res. 36, 195–201. Spielberger, C.D., 1999. State-Trait Anger Expression Inventory, first ed. Psychological Assessment Resources, Odessa, FL. Stephens, A.N., Groeger, J.A., 2009. Situational specificity of trait influences on drivers’ appraisals and driving behaviour. Transp. Res. Part F 12, 29–39. Stephens, A.N., Groeger, J.A., 2011. Anger-congruent behaviour transfers across driving situations. Cogn. Emot. 25, 1423–1438. Sullman, M.J.M., 2006. Anger amongst New Zealand drivers. Transp. Res. Part F 9, 173–184. Sullman, M.J.M., Gras, M.E., Cunill, M.O., Planes, M., Font-Mayolas, S., 2007. Driving Anger in Spain. Personality Individ. Differ. 42, 701–713. Sullman, M.J.M., Stephens, A.N., 2013. A comparison of the Driving Anger Scale and the Propensity for Angry Driving Scale. Accid. Anal. Prev. 58, 88–96. http:// dx.doi.org/10.1016/j.aap.2013.05.002. Sullman, M.J.M., Stephens, A.N., Kuzu, D., 2013. The expression of driving anger amongst Turkish taxi drivers. Accid. Anal. Prev. 56, 42–50. http://dx.doi.org/ 10.1016/j.aap.2013.03.013. Underwood, G., Chapman, P., Wright, S., Crundall, D., 1999. Anger while driving. Transp. Res. Part F 2, 55–68. Villieux, A., Delhomme, P., 2010. Driving anger and its expressions: further evidence of validity and reliability for the Driving Anger Expression Inventory French adaptation. J. Saf. Res. 41, 417–422.