From Driver Stress to Workplace Aggression
Dwight A. Hennessy
Department of Psychology
BuffaloState College
Paper Presented at the 111th Annual American Psychological Association Convention
August 7-10, 2003
TorontoOntarioCanada
From Driver Stress to Workplace Aggression
Dwight A. Hennessy, Ph.D.
BuffaloState College
Paper Presented at the 111th Annual American Psychological Association Convention
Following their regular commute to work, part time employment completed a questionnaire battery containing a state driver stress questionnaire, trait aggressiveness, demographic and vehicle information. Then, following completion of work, they completed a state version of the Workplace Aggression Scale (Neuman & Baron, 1998). Expressed hostility and obstructionism were both predicted by the interaction of gender X state driver stress. As state driver stress increased, the reported frequency of both forms of aggression increased (independently) among male employees only. In contrast, overt aggression was predicted by the interaction of gender X trait physical aggressiveness. Specifically, overt aggression was greatest among males that were higher in physical aggressiveness as a general trait characteristic. The present study highlights the interactive nature of the traffic environment, in that negative experiences in the traffic environment, for some individuals, may “spillover” and influence non-driving interactions.
Workplace aggression has been defined as the purposeful infliction of physical or psychological harm to current or previous co-workers, or to organizations in which an individual is currently employed or were previously employed (Baron & Neuman, 1996). While workplace aggression can take many forms, Neuman and Baron (1998) have proposed a three factor structure which distinguishes between expressed hostility, obstructionism, and overt aggression. Expressed hostility typically involves the verbal and symbolic release of angry feelings, discontent, or negative attitudes towards others (e.g. yelling, ostracism, and character assassination). Obstructionism includes obscure actions intended to intentionally impede another’s performance, largely to damage the image or reputation of specific employees or the organization (e.g. work slowdowns, inaction, and employee sabotage). Acts of overt aggression entail predominantly physical actions directed toward an individual or their personal and corporate property (e.g. physical assault and property damage). Despite the fact that the media has emphasized the dilemma of overt aggression and fatal violence in the workplace, the vast majority of aggressive such incidents are verbal, passive, or indirect actions, in large part due to their ambiguous nature which make it more difficult for victims, bystanders, and employers to identify true aggressive motives (Baron & Neuman, 1996; Glomb 2002).
Previous research has uncovered a number of personal (e.g. Type A, vengefulness, anger, stress) and organizational (e.g. workforce diversity, overload, unfair treatment) factors that interactively heighten aggression and violence in the workplace (see Barling, 1996; Douglas & Martinko, 2001; Folger & Skarlicki, 1998; Neuman & Baron, 1998). However, recent attention has also focused on the issue of “spillover” in which workplace attitudes and behavior are influenced by non-work factors (Leiter & Durup, 1996). The notion is that unresolved daily hassles from one domain can persist, even when no longer in conscious awareness, and add to the pressures of subsequent hassles in other life contexts (Kohn & Macdonald, 1992; Lazarus, 1981; Taylor,1991). These "after-effects" can continue to do psychological and physiological damage, and may intensify over time as they accumulate with other previously unresolved stress reactions (Glass & Singer, 1972). While there is greater evidence that workplace stress increases stress, social withdrawal, and conflict at home (Koizumi, Sugawara, & Kitamura, 2001; Repetti, 1989; Thompson, Kirk-Brown, & Brown, 2001), negative experiences at home have been found to magnify job overload, occupational stress, psychological distress, and arguments/conflict in the workplace (Barnett, Marshall, & Pleck, 1992; Behson, 2002; Bolger, DeLongis, Kessler, & Wethington, 1989; Jenkins, 1996; Jones & Fletcher, 1996; Leiter & Durup, 1996). The latter issue is particularly pressing given that numerous studies have independently identified conflict and workplace stress as prevalent antecedents to workplace aggression and violence (Denenberg & Braverman, 1999; Glomb, 2002; Jenkins, 1996; Neuman & Baron, 1998; Rai, 2002).
Notwithstanding the importance and prevalence of the home environment in an individual’s total life space, there is currently little information on the potential impact of other domains on workplace reactions, such as daily commuting. The driving environment is a common context that has been found to elicit elevated levels of stress and arousal, particularly in highly frustrating conditions (Hennessy & Wiesenthal, 1997, 1999; Hennessy, Wiesenthal, & Kohn, 2000; Novaco, Stokols, and Milanesi, 1990; Rasmussen, Knapp, & Garner, 2000; Stokols & Novaco, 1981; Wiesenthal, Hennessy, & Totten, 2000). Within any driving encounter, there exists a multitude of stimuli that may be perceived or interpreted as stressful, such as bad weather, time pressures, slow moving vehicles, or traffic congestion. According to Broome (1985), stress typically occurs when goals, such as getting to work on time or travelling at a desired pace, are blocked. Most regular commuters report facing numerous frustrations or irritations that can lead to driver stress (Novaco et al., 1990; Gulian, Matthews, Glendon, Davies, & Debney, 1989b). In fact, a recent workplace survey found that traffic congestion was the most prevalent source of daily stress among U.K. drivers (BBC News, 2000).
This issue is particularly pressing in that 60% of those over the age of 16 in the U.S. are part of the labor force and, of these, 88% commute to work via personal vehicle (76% alone) with a median commute time of approximately 20 minutes in each direction (U.S. Census, 2000). While the consequences of commuter stress are individual and varied, such conditions have been linked to a number of negative outcomes, such as increased blood pressure and heart rate (Novaco, Stokols, Campbell, & Stokols, 1979; Stokols, Novaco, Stokols, & Campbell, 1978), negative mood (Gulian, Debney, Glendon, Davies & Matthews, 1989a), emotional arousal (Hennessy & Wiesenthal, 1997), poor concentration levels (Matthews, Dorn, & Glondon, 1991), driving errors, lapses, violations (Westerman & Haigney, 2000), traffic offences (Matthews, Tsuda, Xin, & Ozeki, 1999), traffic collisions (Legree, Heffner, Psotka, Martin, & Medsker, 2003; Selzer & Vinokur, 1974), and aggression (Hennessy & Wiesenthal, 1999).
Stress experienced while driving is embedded in, and interacts with, the total life settings of home, work, and leisure (Gulian et al., 1989b; Novaco et al., 1990). There is growing evidence driver stress is influenced by stressors experienced in other life contexts that accumulate and carry forward to the driving environment (Gulian et al., 89b; Hennessy et al., 2000). Unresolved hassles such poor weather, sleep problems, or conflicts at work and home may exacerbate situational problems faced in the traffic environment, and increase the potential for driver stress and related pathologies (Gulian et al., 1989b; Gulian, Glendon, Matthews, Davies, & Debney, 1990; Hennessy et al., 2000). However, much less is known regarding the potential impact of the traffic environment on subsequent domains such as the workplace. Hence the present study was designed to examine the possible “spillover” effect of traffic stress to the workplace environment. Specifically, it was predicted that drivers experiencing greater state levels of driver stress would exhibit greater expressed hostility and obstructionism during that work day.
Method
Participants
Participants consisted of 70 female and 44 male students from Buffalo State College that worked part time and commuted to work. The age range was from 18 to 56 years, with an average of 19.95 years. Participants were obtained through course recruitment, word of mouth, or campus advertisements.
Measures
The Aggression Questionnaire (AQ) is a 29 items self report measure of an overall trait aggressiveness (Buss & Perry, 1992). The AQ consists of 4 subscales that measure a trait physical aggressiveness (9 items), trait verbal aggressiveness (5 items), trait anger proneness (7 items), and trait hostility (8 items). Participants are asked to rate how characteristic each item is of them in general using a 1-5 Likert scale (1 = “extremely uncharacteristic” and 5 = “extremely characteristic”) . The AQ demonstrates high internal consistency (alpha = .92, Williams, Boyd, Cascardi, & Poythress, 1996) and good test-retest reliability (r = .80) over a 9 week period. Subscale scoring consisted of summing across items, with higher scores representing greater trait aggressiveness.
Driver Stress was measured using a variation of the Driving Behaviour Inventory—General (DBI-Gen) driver stress questionnaire (Gulian et al., 1989b). The DBI-Gen consists of 16 items designed to measure a susceptibility to stress while driving. Each item represents a situation that is often experienced by highly stressed drivers. Consistent with Hennessy and Wiesenthal (1997), items were reworded to represent state experiences from their current commute; for example, “Trying but failing to overtake was frustrating me”, was used rather than “Trying but failing to overtake frustrates me”. Participants were asked to rate how much they agree with each item, using a 0 – 5 Likert scale (0 = “totally disagree” and 5 = “totally agree”). Hennessy and Wiesenthal (1997) have found similar revisions to predict actual stress levels measured in both low and high traffic congestion. Scoring consisted of the mean response to the 16 items, with higher scores indicating greater trait driver stress susceptibility.
The Short Form Driving Anger Scale (DAS) includes 14 items that measure a single driver anger dimension (Deffenbacher, Oetting, & Lynch, 1994). Items present common driving scenarios that have been found to elicit varying degrees of anger. For the purpose of the present study, participants were asked to imagine that each scenario had just happened during their commute and rate the level of anger they would have experienced using a 1- 5 Likert scale (1 = “not at all” and 5 = “very much”). The short form of the DAS has been found to demonstrate good reliability (alpha = .80) and to correlate highly with the long form of the DAS (r = .95) (Deffenbacher, Huff, Lynch, Oetting, & Salvatore, 2000). Scoring consisted of summing the responses to the 14 items, with higher scores representing greater driver anger.
The Workplace Aggression Scale (WPAS) was developed as an overall measure of the self report prevalence of workplace aggression (Neuman & Baron, 1998). The WPAS consists of three factors representing unique categories of aggression: 1) Expressed Hostility involves 15 items representing verbal and symbolic release of feelings of anger, discontent, or negative attitudes towards others (e.g. staring, obscene gestures, belittling); 2) Obstructionism involves 9 items based on actions intended to impede another’s performance (e.g. failing to return phone calls, work slowdowns, showing up late for meetings); and 3) Overt Aggression involves 8 items that focus on physical actions toward an individual or their property (e.g. attack with a weapon, threats of physical violence, sabotaging company property needed by target). The present study revised the WPAS to represent a more “state” measure of the prevalence of aggression during the past work day. Participants are asked to rate the extent that they personally initiated each form of aggression during the target work day, using a 0-5 Likert scale (0 = “not at all” and 5 = “nearly all the time”). Subscale scoring consisted of summing across items, with higher scores representing greater aggression during that work day.
Procedure
The researchers initially met with participants to gain informed consent and to fully explain the research procedure. During this meeting, participants completed the Aggression Questionnaire to measure their trait aggressiveness and were provided the state questionnaire batteries to be completed following their next commute to work. They were instructed to drive to work as usual and upon arrival in the parking lot, complete the State Stress and Driver Anger scales. Subsequently, following that work day they were to complete the state version of the Workplace Aggression scale.
Results
Table 1 contains all bivariate correlations, means, and alpha reliabilities for all scale and subscales. Overall, all measures demonstrated moderate to high reliability (α = .61 - .91), although trait physical aggression (α = .67) and trait anger (α = .61) showed the weakest consistency. Significant relationships were found between the state driver measures (state stress and anger), between the workplace aggression subscales (expressed hostility, obstructionism, and overt aggression), and between trait aggression subscales (physical, verbal, anger, and hostility). In addition, the state driving indices, workplace aggression subscales, and trait aggression subscales all were significantly related to one another.
Table 1. Intercorrelations, Means, Standard Deviations, and Reliabilities for State Driver Stress, Driver Anger, Workplace Aggression, and Trait Aggression
1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 91. Driver Stress / — / .66** / .41** / .32** / .23* / .31** / .66** / .44** / .49**
2. Driver Anger / — / — / .43** / .29** / .21* / .27* / .53** / .38** / .41**
3. Expressed Hostility / — / — / — / .72** / .52** / .40** / .50** / .42** / .51**
4. Obstructionism / — / — / — / — / .51** / .29** / .32** / .43** / .38**
5. Overt Aggression / — / — / — / — / — / .41** / .33** / .35** / .38**
6. Physical Aggression / — / — / — / — / — / — / .60** / .58** / .40**
7. Verbal Aggression / — / — / — / — / — / — / — / .78** / .76**
8. Anger / — / — / — / — / — / — / — / — / .68*
9. Hostility / — / — / — / — / — / — / — / — / —
Mean / 37.47 / 43.78 / 16.47 / 6.32 / 1.60 / 17.13 / 29.17 / 19.76 / 18.38
SD / 13.7 / 10.40 / 11.30 / 6.49 / 3.37 / 5.28 / 8.71 / 5.39 / 6.36
Alpha / .86 / .87 / ..89 / .87 / .85 / .67 / .85 / .61 / .85
Note: n=114; *p<.05; **p<.01
In order to examine the influences of state driver stress on workplace aggression, separate hierarchical entry multiple regressions were conducted for expressed hostility, obstructionism, and overt aggression. The procedure was to enter demographics (age and gender) in the first step, aggressive personality aspects (physical aggression, verbal aggression, anger, hostility) in the second step, state driving factors (stress and anger) in the third step, and to add all cross product interactions stepwise in the fourth step.
Both expressed hostility (R2=0.532, F(9,61)=7.70, p<.01; see Table 2) and obstructionism (R2=0.427, F(9,61)=5.05, p<.01; see Table 3) were predicted by the interaction of gender X driver stress. Specifically, both expressed hostility and obstruction were more prevalent among male drivers that experienced elevated stress during their commute to work. Overt aggression was predicted by the interaction of gender X trait physical aggressiveness (R2=0.389, F(9,61)=4.31, p<.01; see Table 4), where overt aggression was greatest among males with an elevated trait physical aggressiveness.
Table 2. Regression Model For Expressed Hostility
b / t / ∆R2 / ∆FIntercept / 18.795
Demographics / .251 / 11.41**
Age / -.345 / -.92
Gender / -10.488 / -1.44
Trait Aggression / .198 / 5.76**
Hostility / .454 / 1.64
Anger / -.322 / -.82
Physical Aggression / .447 / 7.79
Verbal Aggression / .094 / .39
State Driving / .032 / 1.90
Driver Anger / .117 / .78
Driver Stress / -.626 / -2.37*
Interaction Effects / .050 / 6.57*
Gender X Driver Stress / .484 / 2.56*
Table 3. Regression Model For Obstructionism
b / t / ∆R2 / ∆FIntercept / 10.084
Demographics / .172 / 7.06**
Age / -.066 / -.29
Gender / -7.368 / -1.66
Trait Aggression / .184 / 4.57**
Hostility / .135 / .80
Anger / .288 / 1.21
Physical Aggression / .333 / 2.18
Verbal Aggression / -.236 / -1.62
State Driving / .008 / 0.40
Driver Anger / -.077 / .78
Driver Stress / -.308 / -1.91
Interaction Effects / .063 / 6.72*
Gender X Driver Stress / .298 / 2.59*
Table 4. Regression Model For Overt Aggression
b / t / ∆R2 / ∆FIntercept / 5.418
Demographics / .136 / 5.33**
Age / -.046 / -.40
Gender / -4.844 / -1.89
Trait Aggression / .171 / 3.95**
Hostility / .113 / 1.33
Anger / .027 / .23
Physical Aggression / -.371 / -1.56
Verbal Aggression / -.056 / -.76
State Driving / .014 / 0.63
Driver Anger / .001 / .02
Driver Stress / .003 / .07
Interaction Effects / .068 / 6.79*
Gender X Driver Stress / .364 / 2.60*
In order to investigate potential gender differences on key factors identified in the hierarchical regression, an ANOVA was conducted using gender as the IV and state driver stress, state driver anger, expressed hostility, obstructionism, and overt aggression as DV (see Table 5). Men reported greater state driver anger, expressed hostility, obstructionism, and overt aggression than women, while no gender difference was found for state driver stress.
Table 5. ANOVA Results Comparing State Driving Measures and Workplace Aggression Factors Across Gender
MalesMean (SD) / Females
Mean (SD)
Driver Stress / F(1,110) = .896, p<.ns / 39.04 (12.67) / 36.50 (14.31)
Driver Anger / F(1,110) = 5.18, p<.05 / 2.15 (1.04) / 1.70 (.99)
Expressed Hostility / F(1,86) = 17.69, p<.01 / 22.05 (12.38) / 12.61 (8.68)
Obstructionism / F(1,86) = 17.98, p<.01 / 9.55 (6.67) / 4.09 (5.36)
Overt Aggression / F(1,101) = 21.04, p<.01 / 3.35 (4.67) / 0.49 (1.30)
Discussion
As expected, state driver aggression was related to subsequent acts of workplace aggression, but only for males and only in the form of expressed hostility and obstructionism. This suggests that, some form of spillover from the traffic environment to the workplace environment is possible. Commuter stress is a serious issue that can have significant physical and psychological consequences. But of particular note in this realm is the fact that it may influence the cognitive processes involved in attention, information processing, and appraisal that are essential to effective coping (Matthews, 2002). It is possible that the efforts to deal with traffic stress may exhaust coping resources needed to deal with subsequent stressors in the workplace environment, while simultaneously priming individuals to construe workplace strains as more demanding and stressful; consequently, increasing the possibility of aggressive responses. According to Viitasara and Menckel (2002), such aggression and violence are strongly influenced by immediate contextual factors, but broader situational factors help shape the nature and outcome of those events. In the present study, it is possible that the broader situational factors for male commuters include previously experienced traffic difficulties.
Interestingly, driver anger did not predict workplace aggression, despite the well established general link between anger and aggression (see Berkowitz, 1993) and the strong relationship between state driver stress and state driver anger in the present study. One potential reason may be that anger is an affective state based on the interpretation of arousing conditions (see Geen, 1990). In the absence of the source of arousal, the affective reaction of anger may diminish. This is consistent with Van Rooy and Rotton (2003) who found that arousal and negative affect increased during a commute, but returned to baseline levels relatively quickly following its completion. However, as with other sources of daily hassles, it is possible that any unresolved issues in the traffic environment may have continued to influence and intensify subsequent workplace reactions to potential stressors, even though their immediate, or “state”, effects may have become indistinguishable in conscious awareness. This notion is the crux of the spillover effect; that unresolved difficulties from one domain accumulate and influence the interpretation and intensify negative reaction to stressors in subsequent domains (Leiter & Durup, 1996).
Another finding of note is the fact that the spillover effects from traffic stress to workplace aggression appeared only for men. Previous research has provided some evidence of gender differences in spillover, in that the sources and reactions to work and family stressors and their ultimate outcomes across domains sometimes differ between men and women, although the exact nature of this gender effect is mixed and not fully clear (Grzywacz & Marks, 2000; Maume, 2001; Thompson et al., 2001). However, Bolger et al. (1989) did find that, for men only, conflicts at home were proceeded by greater work stress the next day. Further, according to Almeida, Wethington, & Kessler (2002), men are more prone than women to report workplace stress as a major daily hassle, and, consistent with the present study, workplace aggression is typically committed by men rather than women (Neuman & Baron, 1998). In this respect, men may be more inclined to accept workplace aggression as an appropriate indirect coping response, particularly if the consequences of previous acts of aggression have been “successful” in solving disputes or reducing stress. Douglas and Martinko (2001) have stressed the importance of understanding personal factors in understanding workplace aggression. While this is not to minimize the impact of social and organizational antecedents, it is individuals that must interpret these experiences and their coping responses. Two such factors include anger and hostile attribution biases in that anger prone and vengeful individuals are more prone to interpret unpleasant events as stressful and demonstrate workplace aggression (Douglas & Martinko, 2001). State driver anger was found to be greater among men than women in the present study, and previous research has consistently found that men are more prone to hold a vengeful attitude in general (Stuckless & Goronson, 1992) and while driving (Wiesenthal, Hennessy, & Gibson, 2000; Hennessy, 1999). As a result, the present study may have uncovered a spillover effect only among men due to their greater tendency to interpret negative workplace events as stressful or provocative and select aggression as an appropriate coping strategy under the cumulative effects of heightened life hassles.