The Displaced Aggression Questionnaire

Journal of Personality and Social Psychology Copyright 2006 by the American Psychological Association
2006, Vol. 90, No. 6, 1032–1051 0022-3514/06/$12.00 DOI: 10.1037/0022-3514.90.6.1032
The Displaced Aggression Questionnaire
Thomas F. Denson William C. Pedersen

University of Southern California California State University, Long Beach
Norman Miller
University of Southern California
Previous measures of aggressive personality have focused on direct aggression (i.e., retaliation toward
the provoking agent). An original self-report measure of trait displaced aggression is presented.
Exploratory and confirmatory factor analyses provided support for a 3-factor conceptualization of the
construct. These analyses identified an affective dimension (angry rumination), a cognitive dimension
(revenge planning), and a behavioral dimension (general tendency to engage in displaced aggression).
The trait measure demonstrated good internal consistency and test–retest reliability as well as convergent
and discriminant construct validity. Unlike other related personality measures, trait displaced aggression
significantly predicted indirect indicators of real-world displaced aggression (i.e., self-reported domestic
abuse and road rage) as well as laboratory displaced aggression in 2 experiments.
Keywords: aggression, displaced aggression, rumination, aggressive personality, domestic abuse
Luis steps onto a crowded bus on his way to work one morning. and subsequently aggresses against a seemingly innocent target
As he is entering, another man bumps into him, spills hot coffee (Dollard, Doob, Miller, Mowrer, & Sears, 1939; Hovland & Sears,
onto Luis’s new shirt, and quickly darts away. On his 45-minute 1940). For instance, a man insults his wife for no apparent reason
commute to work, all Luis can think about is how angry he is after having been berated previously by his boss. A meta-analysis
feeling. He cannot stop thinking about how he would enjoy getting of laboratory studies on the construct— operationalized as aggres-
even with the coffee-sloshing provocateur. When he gets to work, sion directed toward a human target other than the source of initial
he is in a foul mood. A good-natured coworker comments jokingly provocation—reported a moderate effect size (d ⫽ .54; Marcus-
on Luis’s “fashion by Starbucks” appearance. Luis becomes furi- Newhall, Pedersen, Carlson, & Miller, 2000). Oftentimes, the
ous and proceeds to insult the coworker. target provides no justification or instigation to warrant a retalia-
The preceding anecdote illustrates a general phenomenon. When tory response from the aggressor. However, on other occasions the
some people are in a bad mood, they are likely to “take it out” on “innocent” target may provide a trivial and ambiguous instigation
innocent others. In the current research we argue that individual (“a trigger”) to aggress. For instance, in our opening example, Luis
differences in the tendency to exhibit displaced aggression may perceived the coworker’s comment as hostile and subsequently
contribute to such behavior. We describe next the development of displayed a disjunctively escalated aggressive response toward the
the Displaced Aggression Questionnaire (DAQ), the first self- witty coworker. This is an example of TDA (Aviles, Earleywine,
report measure designed to assess individual differences in the Pollock, Stratton, & Miller, 2005; Bushman, Bonacci, Pedersen,
tendency to displace aggression. Vasquez, & Miller, 2005; Miller, Pedersen, Earleywine, & Pol-
lock, 2003; Pedersen, Gonzales, & Miller, 2000; Vasquez, Denson,
Displaced Aggression and Triggered Displaced Pedersen, Stenstrom, & Miller, 2005). Such disjunctively escalated
Aggression (TDA) aggression refers to a level of aggression exceeding norms of
Displaced aggression occurs when a person is provoked, is reciprocity and tit-for-tat matching rules (Axelrod, 1984; Gould-
unwilling or unable to retaliate against the original provocateur, ner, 1960).
The current research represents the first investigation into indi-
Thomas F. Denson and Norman Miller, Department of Psychology, vidual differences in the tendency to displace aggression. Existing
University of Southern California; William C. Pedersen, Department of investigations of aggressive personality have been constrained to
Psychology, California State University, Long Beach. instances of direct aggression (i.e., retaliation toward the provok-
Portions of this article served as Thomas F. Denson’s master’s thesis ing agent). There is good reason to believe that personality differ-
and were presented at the American Psychological Society’s May 2005 ences exist in the tendency to engage in displaced aggression and
meeting in Los Angeles, CA. The current research was supported in part by that rumination plays an important role. For instance, in addition to
a John Randolph Haynes and Dora Haynes Foundation fellowship to laboratory studies on rumination and TDA (Bushman et al., 2005),
Thomas F. Denson and by Grant R21-AA013343 from the National Insti-
spousal or child abusers are characterized as harming innocent
tute on Alcohol Abuse and Alcoholism.
Correspondence concerning this article should be addressed to Thomas others when stressed, frustrated, or provoked. Understanding in-
F. Denson, Department of Psychology, University of Southern California, dividual differences in displaced aggression may have especially
Seeley G. Mudd Building, Room 501, Los Angeles, CA 90089-1061. important ramifications for society at large. Following a provoca-
E-mail: [email protected] tion, individuals high in this trait presumably are likely to aggress
1032
TRAIT DISPLACED AGGRESSION 1033
against undeserving others such as family members, fellow driv- placed aggression. It is likely that everyday stressors and provo-
ers, or coworkers. Assessment of such individual differences may cations prime aggression-related affect, arousal, and cognition,
be a starting point for reducing the harm associated with domestic which are maintained or increased by angry rumination and
abuse, road rage, and workplace aggression. thoughts of revenge. Such chronic priming effects might then
cause one to emotionally overreact in social encounters with close
Rumination others. Over time, an individual may learn to take out his or her
aggressive urges on others, a practice that can manifest itself in a
What happens between the initial provocation and the subse- long-term general behavioral tendency to harm those other than the
quent aggressive act? When an individual is confronted with a source of the initial provocation.
provocation, there are a number of emotion regulation strategies he
or she may use to cope with the aversive event. We hypothesize
that individuals high in trait displaced aggression predominantly Trait Displaced Aggression
use rumination to cope with life’s provocations. We know from
laboratory research that angry feelings resulting from a provoca- Researchers have explored a number of dimensions related to
tion are relatively short-lived for most people, usually dissipating aggressive personality. These include the tendency to engage in
within 10 min (Fridhandler & Averill, 1982; Tyson, 1998). How- verbal and physical aggression; the frequent experience of hostility
ever, there are many instances of real-world displaced aggression and anger (Buss & Perry, 1992); the chronic accessibility of
in which an aggressor harms innocent others when substantially aggressive constructs (Dill, Anderson, & Deuser, 1997), often
more than 10 minutes have elapsed since the provoking event (as resulting from exposure to violent media (C. A. Anderson &
was the case with Luis in our opening anecdote). Indeed, some- Bushman, 2001); gender (Bettencourt & Miller, 1996); anger
times days or weeks may pass. What accounts for these occur- expression (Spielberger, Reheiser, & Sydeman, 1995); narcissism
rences? As indicated, we hypothesize that those who take it out on and self-esteem (Bushman & Baumeister, 1998); and frontal elec-
others are likely to ruminate about the initial provocation (Miller et troencephalogram asymmetry and approach–withdrawal tenden-
al., 2003). Specifically, these individuals are likely to focus on cies (Harmon-Jones & Allen, 1998; Hewig, Hagemann, Seifert,
their angry mood and to plan retaliation. We hypothesize that this Naumann, & Bartussek, 2004).
ruminative activity maintains aggression-related affect, cognition, The first major goal of the current investigation was to identify
and arousal and thereby increases negative emotional reactions individual differences in displaced aggression. All prior research
toward those they subsequently encounter. on aggressive personality (with the possible exception of domestic
On the basis of Berkowitz’s (1989, 1990, 1993) cognitive neo- violence research) has focused on direct aggression (i.e., retaliation
associationistic model of aggression and consistent with the gen- toward the source of provocation). To our knowledge, there is no
eral aggression model (GAM; C. A. Anderson & Bushman, 2002), existing individual-difference measure to assess the general ten-
Miller et al. (2003) proposed a theory of TDA. They argued that dency to engage in displaced aggression. We note that individual
individuals may ruminate about the initial provocation and further differences in displaced aggression are not orthogonal to individ-
that rumination maintains aggressive affect and cognition by keep- ual differences in direct aggression. Many of the same processes
ing a cognitive representation of the provocation highly accessible. that influence general trait aggressiveness also apply to trait dis-
Rumination, which can be conceptualized within the GAM as placed aggression. For example, biological factors, social learning,
either a situational or a personality factor, is likely to increase emotional instability (e.g., neuroticism), frequent anger experi-
aggression-related cognition, affect, and arousal. This, in turn, ence, and hostile attributional styles remain important correlates of
increases the strength of association between aggression-related displaced aggression. Therefore, we expected moderate correla-
concepts in the network. Each time a person thinks about or relives tions between the DAQ and trait measures of direct aggression.
a provoking incident (or its accompanying negative affect), a new Consequently, the individual-difference measure that we devel-
activation spreads through the network, making its aggression- oped in the current research builds on this prior work in trait
related components more accessible and in turn increasing the aggression.
likelihood of subsequent aggressive behavior. Such rumination- Those high in trait displaced aggression differ in an important
induced priming has been implicated in marital conflict (Kacha- manner from those high in general trait aggressiveness. Specifi-
dourian, Fincham, & Davila, 2005). Relatedly, Bushman (2002) cally, unlike direct aggressors, individuals with a strong tendency
demonstrated that a laboratory manipulation of rumination about a to exhibit displaced aggression are hypothesized to be behaviorally
provocation increased aggression toward the source of that prov- inhibited when provoked. When exposed to a provocation, we
ocation. Similarly, Konečni (1974) found that preventing partici- hypothesize that individuals who are high in displaced aggression
pants from ruminating decreased direct aggression. are likely to inhibit direct aggression toward the provocateur.
Relevant to the current research, in a series of three laboratory Indeed, one unique aspect of our construct is its positive relation-
studies, rumination about a provocation increased aggression to- ship to behavioral inhibition. This is entirely novel, as previous
ward the source of a subsequent minor annoyance (viz., the trig- work on direct aggressiveness has revealed a positive relationship
gering agent in the TDA paradigm; Bushman et al., 2005). Also between anger, trait aggression, and the behavioral approach sys-
consistent with the GAM, laboratory manipulations of rumination tem (Harmon-Jones, 2003; Harmon-Jones & Allen, 1998;
increased cardiovascular arousal, as well as aggressive affect and Harmon-Jones & Sigelman, 2001; Hewig et al., 2004). Thus, we
cognition (Pedersen, Denson, Goss, Vasquez, & Miller, 2005; see hypothesize that when provoked, individuals high in trait displaced
also Rusting & Nolen-Hoeksema, 1998). In the current research, aggression are likely to have an activated “flight” system whereas
we argue that individual differences in rumination about an initial those high in direct aggression are likely to have an activated
provocation are intricately linked to behavioral displays of dis- “fight” system.
1034 DENSON, PEDERSEN, AND MILLER
A second goal of the current research is to provide clarification Among married couples, cognitive measures of rumination about
of the dimensions underlying trait displaced aggression. Past re- interpersonal transgressions decreased forgiveness (Kachadourian
search has focused primarily on cognitive or affective features of et al., 2005) and marital satisfaction (Paleari, Regalia, & Fincham,
the provocation-focused ruminative personality. In the current 2005). Still other researchers have failed to differentiate between
research, we hypothesized the existence of three related but dis- cognitive and affective components of provocation-focused rumi-
tinct components of trait displaced aggression: (a) an affective nation, thus combining anger-related aspects with thoughts of
component consisting of the tendency to focus on one’s anger revenge (e.g., the Rehearsal subscale of the Emotional Control
following a provocation (angry rumination), (b) a cognitive com- Questionnaire; Roger & Najarian, 1989). In conceptualizing trait
ponent referring to the tendency to hold a grudge for a prior displaced aggression, we hypothesized that both affective (angry
provocation and plan for retaliation (revenge planning), and (c) a rumination) and cognitive (revenge planning) components of
behavioral component referring to a general tendency to behave provocation-focused rumination would be a part of the construct.
aggressively toward those other than the source of the initial Rumination has also been conceptualized in a number of ways
provocation (behavioral displaced aggression). that are distinct from provocation-focused rumination. In the cur-
A small yet highly relevant body of research has concentrated rent research we compared our measure with measures based on
on conceptualizations of rumination that focus specifically on these other perspectives on rumination referred to as self-focused
responses to provocations. We believe such provocation-focused rumination. Such a theoretical perspective describes rumination as
rumination (e.g., thinking about and reliving a negative event or an “self-focused attention” or directing attention inward on the self,
angering incident, as was the case in the opening anecdote about particularly on one’s own negative emotions (Lyubomirsky &
Luis) more closely corresponds to the layperson’s definition. Nolen-Hoeksema, 1995; Nolen-Hoeksema & Morrow, 1991,
Provocation-focused rumination has been shown to increase anger 1993; Trapnell & Campbell, 1999). A large body of research
and direct aggression to a greater extent than self-focused rumi- demonstrates that self-focused rumination increases depressive
nation (Pedersen et al., 2005). symptoms and lengthens episodes of depressed mood (e.g., Carver,
Within the realm of provocation-focused rumination, research- Scheier, & Weintraub, 1989; Nolen-Hoeksema, Morrow, &
ers have concentrated on affective and cognitive components. The Fredrickson, 1993). Self-focused rumination also exacerbates an-
affective aspect of provocation-focused rumination consists of the ger, stress, anxiety, and worry (Morrison & O’Connor, 2005;
negative affect, especially anger, that results from a provocation Muris, Roelofs, Meesters, & Boomsma, 2004; Rusting & Nolen-
(Caprara, 1986; Sukhodolsky, Golub, & Cromwell, 2001). Sukh- Hoeksema, 1998; Segerstrom, Tsao, Alden, & Craske, 2000;
odolsky et al. (2001) viewed anger as a primary component of Watkins, 2004). Within the category of self-focused rumination,
rumination and developed the Anger Rumination Scale to assess researchers have distinguished between reflective and brooding
individual differences pertaining to the “tendency to focus on rumination (Trapnell & Campbell, 1999; Treynor, Gonzales, &
angry moods, recall past anger experiences, and think about the Nolen-Hoeksema, 2003; Watkins, 2004). Trapnell and Campbell
causes and consequences of anger episodes” (p. 689). The scale (1999) created the Rumination–Reflection Questionnaire to assess
predicted self-reported direct aggression among athletes (Maxwell, these two distinct types of rumination. Reflective rumination is
2004). Thus, it is likely that individuals who focus on angry believed to be an adaptive self-regulatory process in which indi-
feelings should maintain this negative affect (Miller et al., 2003). viduals are motivated to focus inward to satisfy curiosity or self-
A large body of research indicates that regardless of its source, understanding, whereas maladaptive brooding rumination is initi-
negative affect produces a readiness to aggress (C. A. Anderson, ated in response to “threats, losses, or injustices” (Trapnell &
2001; Berkowitz, 1993; Lindsay & Anderson, 2000). Such priming Campbell, 1999, p. 297). Other definitions of rumination have
effects increase the likelihood that ambiguous events will be included “a class of conscious thoughts that revolve around an
perceived aggressively. In this manner, individuals who ruminate instrumental theme and that recur in the absence of immediate
angrily may be likely to aggress against close others with whom environmental demands” (Martin & Tesser, 1996, p. 12). Martin
they have extended daily contact (e.g., family, coworkers, annoy- and Tesser viewed the occurrence of a blocked goal as the main
ing drivers). instigation to ruminate. Still others have suggested that self-focus
Most researchers have concentrated on the cognitive aspects of leads to comparison of the actual self with the desired self, which
provocation-focused rumination. Caprara (1986) conceptualized in turn may cause increased negative affect (e.g., Duval & Wick-
rumination as a continuum with those likely to rapidly abandon lund, 1972).
distress and vengefulness motivation on one end (dissipators) and In summary, trait displaced aggression shares some features
those likely to focus on negative affect and think about revenge on with trait measures of brooding self-focused rumination (e.g.,
the other end (ruminators); nevertheless, all but 2 of the 15 items unwanted repetitive thoughts, exacerbated negative affect) but not
that make up Caprara’s (1986) Dissipation–Rumination Scale as- reflective or goal-oriented self-focused rumination. In addition,
sess the tendency to remember past provocations or the motivation although positively correlated, trait displaced aggression differs
to retaliate. The scale has proven useful in predicting direct ag- qualitatively from general trait negative affect because it is spe-
gression when participants were given the opportunity to ruminate cifically concerned with affective, cognitive, and behavioral re-
(Collins & Bell, 1997). Similarly, in the forgiveness literature, sponses to provocations (e.g., Watson, Clark, & Tellegen, 1988).
researchers have conceptualized ruminative thought as the source Although previous research has investigated individual differ-
of desire for revenge (Mauger et al., 1992; McCullough, Bellah, ences in provocation-focused rumination (e.g., Caprara, 1986;
Kilpatrick, & Johnson, 2001; McCullough et al., 1998). Studies Sukhodolsky et al., 2001), the DAQ is novel in that it contains
using the Dissipation–Rumination Scale have found negative cor- these elements as well as an additional assessment of the behav-
relations between vengeful trait rumination and forgivingness ioral tendency to aggress against undeserving others when pro-
(Berry, Worthington, Parrott, O’Connor, & Wade, 2001, 2005). voked. Together, these three factors provide a detailed description
of the affective, cognitive, and behavioral elements that character- Aggression Questionnaire. The Aggression Questionnaire (Buss &
ize trait displaced aggression. Perry, 1992) is a 29-item measure consisting of four subscales: Hostility
(␣ ⫽ .83), Anger (␣ ⫽ .84), Physical Aggression (␣ ⫽ .85), and Verbal
Aggression (␣ ⫽ .82). This scale has proven useful in predicting laboratory
The Current Research and real-world aggression (Bushman & Wells, 1998; Buss & Perry, 1992).
In the five phases of research that follow, we report on the development The latter two subscales were included as a source of preliminary validity
of the DAQ. We first selected items and established the hypothesized information.
three-factor structure (i.e., Angry Rumination, Revenge Planning, and
Behavioral Displaced Aggression). We confirmed the factor structure of
the final 31-item scale in two separate samples. In a correlational study, we
Procedure
demonstrated concurrent discriminant and convergent construct validity.
Participants completed the measures at home as part of a multiple-page
We also present evidence on the temporal stability of the DAQ. Finally, we
packet at the beginning of the semester. All items were rated on bipolar
provide behavioral evidence of construct validity in two laboratory studies
7-point Likert-type scales ranging from 1 (extremely uncharacteristic of
in which the DAQ predicted displaced aggression whereas other theoreti-
me) to 7 (extremely characteristic of me). Four randomized versions of the
cally relevant measures did not.
questionnaire were administered to participants.
Phase 1: Initial Item Selection

Results
The purpose of Phase 1 was to develop an initial item pool for
subsequent data analysis. As other researchers have done (e.g., Items were reverse scored when necessary. Missing values were
Amirkhan, 1990; Mauger et al., 1992), we used preexisting per- replaced with the mean for that item. Missing data did not exceed
sonality measures as a starting point. Specifically, we administered 2.5% for any single item.1 To obtain an adequate starting point for
three measures of rumination as well as measures of trait irritabil- item selection and subsequent data collection, we conducted an
ity, anger, and hostility. We also administered measures of trait exploratory factor analysis with oblique rotation on the Dissipation–
verbal and physical aggression (Buss & Perry, 1992) to verify that Rumination Scale, Scott–McIntosh Rumination Inventory, ven-
the preliminary items were related to self-reported aggressive geance items, and Anger and Hostility subscales from the Aggres-
behavior. We examined our data with exploratory factor analysis sion Questionnaire. Analysis of items and a scree plot resulted in
to identify a subset of useful items. We then correlated these four interpretable factors, accounting for 40% of the variance in
factors with self-reported aggression. This subset of items was then responses. The eigenvalues of these four factors were 18.42, 3.63,
used in a second phase of item development. 2.96, and 2.46, respectively. An item was said to load on a
particular factor if the loading was .40 or greater. The four retained
Method factors consisted of 32 items. The first factor contained 4 items
related to the individual’s sensitivity to personal insult (Sensitivity
Participants to Insult; ␣ ⫽ .60). The second factor contained 10 items relevant
to experiencing anger and irritability (Anger/Irritability; ␣ ⫽ .90),
A total of 521 (71% female and 29% male) University of Southern
and the third factor contained 15 items related to holding grudges
California (USC) and California State University, Long Beach (CSULB)
and plotting revenge (Vengeance/Grudge Holding; ␣ ⫽ .90). Fi-
undergraduates enrolled in psychology courses completed the materials in
exchange for extra course credit. nally, the fourth factor was the 3-item Motivation subscale from
the Scott–McIntosh Rumination Inventory (␣ ⫽ .82).
To determine the association of the four retained factors with
Materials physical and verbal aggression, we created composites of each
Rumination measures. We administered three measures of trait rumi- factor and correlated them with the Aggression Questionnaire
nation. First was the 20-item (␣ ⫽ .88) Dissipation–Rumination Scale Physical and Verbal Aggression subscales. The first three factors,
(Caprara, 1986), which assesses individual differences along a continuum but not the Motivation subscale, were moderately associated with
ranging from the tendency to shrug off provocations and annoyances trait physical and verbal aggression (see Table 1).
(dissipation-oriented personality) to the tendency to extensively focus on
thoughts of revenge and to hold a grudge in response to provocations
(rumination-oriented personality). Second, the Scott–McIntosh Rumination Discussion
Inventory (Scott & McIntosh, 1999) is a 9-item measure that assesses three
distinct aspects of rumination: emotionality (␣ ⫽ .68), distraction (␣ ⫽ Phase 1 provided a starting point for the development of our
.48), and motivation (␣ ⫽ .82). The scale is theoretically consistent with new measure. An exploratory factor analysis suggested the pres-
Martin and Tesser’s (1996) emphasis on blocked goals as an antecedent to ence of four latent variables, which accounted for much of the
rumination and therefore focuses on reactions to failed goal attainment.
variance in the observed scores. Three of these factors (Sensitivity
Third, seven items (␣ ⫽ .84) from Mauger et al.’s (1992) Forgiveness of
Others Scale were used in the current study to assess trait vengefulness,
to Insult, Anger/Irritability, and Vengeance/Grudge Holding) were
which is defined as the motivation to seek revenge in response to provo- significantly related to reliable and valid trait measures of general
cation. These items were related to greater rumination about an offense, physical and verbal aggression (Buss & Perry, 1992). The 32 items
greater negative affect, and less forgiving (McCullough et al., 2001). associated with these factors were used in the second and final
Irritability Scale. Trait irritability was assessed with the 30-item (␣ ⫽ phase of item selection.
.90) Irritability Scale (Caprara, 1985). This scale assesses the tendency to
respond to situations offensively and to emotionally overreact to frustrating
1
situations. It also has predicted laboratory aggression (K. B. Anderson, This item was “I am often sulky” from the Dissipation–Rumination
Anderson, Dill, & Deuser, 1998). Scale (Caprara, 1986).
Table 1
Robust Correlations Among the Four Factors From Phase 1 and the Physical and Verbal
Aggression Subscales of the Aggression Questionnaire
Subscale and factor 1 2 3 4 5 6
1. Physical Aggression —
2. Verbal Aggression .41* —
3. Sensitivity to Insult .28* .31* —
4. Anger/Irritability .54* .44* .56* —
5. Vengeance/Grudge Holding .52* .43* .59* .59* —
6. Motivation subscale ⫺.02 .03 ⫺.07 ⫺.15* ⫺.11 —
* significant, controlling for familywise error rate (␣ ⫽ .05)
Phase 2: Item Selection and Factor Structure Finalized Verbal and physical aggression. The Physical (␣ ⫽ .83) and Verbal
Aggression (␣ ⫽ .77) subscales from the Aggression Questionnaire (Buss
The purpose of Phase 2 was to obtain the final item set for our & Perry, 1992) were also included to provide preliminary validation of our
new measure of displaced aggression and to finalize the factor measure.
structure of the scale. We used exploratory and confirmatory factor
analyses to identify the factor structure. We also provide prelim- Procedure
inary validity evidence. The data from Phase 1 served as a starting
point for this second and final wave of item selection. Specifically, As in Phase 1, participants completed the scales at home as part of a
in Phase 2, we not only administered the 32 items composing the multiple-page packet at the beginning of the semester. All items were rated
four factors from Phase 1 but in addition included two existing on bipolar 7-point Likert-type scales ranging from 1 (extremely uncharac-
rumination scales and several original items designed to assess teristic of me) to 7 (extremely characteristic of me). Participants were asked
to complete the measures honestly and were informed that their responses
stable individual differences in the behavioral tendency to engage
would remain anonymous. Four randomized versions of the items com-
in displaced aggression (i.e., aggression toward those other than
posing the questionnaire were administered to participants.
the source of a provocation). We hypothesized that a three-factor
structure consisting of affective, cognitive, and behavioral ele-
ments would emerge. We tested this hypothesis with exploratory Results
and confirmatory factor analysis.
Items were reverse scored when necessary. Missing values were
replaced with the mean for that item. Missing data did not exceed
Method 1.7% for any single item.2
An exploratory factor analysis was used as a starting point to
Participants
determine the factor structure of the current data. Two items were
A total of 471 USC and CSULB undergraduates enrolled in a psychol- excluded because they showed low variability (i.e., standard de-
ogy course completed the materials in exchange for extra course credit. viations less than 1.0).3 An exploratory factor analysis with max-
imum likelihood estimation and oblique rotation on all of the items
Materials except the Verbal and Physical Aggression subscales from the
Aggression Questionnaire was conducted. Analysis of a scree plot
Items from Study 1. We administered the 32-items composing the four
retained factors from Phase 1. These were Sensitivity to Insult (␣ ⫽ .52), and proportion of variance accounted for resulted in a four-factor
Impulsivity (␣ ⫽ .88), Vengeance/Grudge Holding (␣ ⫽ .89), and the solution, accounting for 39% of the variance. The eigenvalues for
Motivation subscale from the Scott–McIntosh Rumination Inventory (␣ ⫽ these four factors were 26.98, 5.12, 3.73, and 3.09, respectively.
.76). Additional factors did not account for more than 2% of the vari-
Anger Rumination Scale. Although one study found only a single ance. To help reduce the number of items in the final scale, we
factor (Maxwell, 2004), this 19-item scale (Sukhodolsky et al., 2001) considered an item to load on a particular factor if the loading was
assesses four broad aspects of ruminative responses to anger-provoking .50 or greater (vs. .40 or greater in Phase 1).
experiences: angry afterthoughts (␣ ⫽ .83); thoughts of revenge (␣ ⫽ .73); The first factor consisted of 10 items related to negative affect.
angry memories (␣ ⫽ .81); and understanding of causes (␣ ⫽ .68). The
Specifically, these items were related to anger-based rumination
scale has good internal consistency and good test–retest reliability over 1
resulting from provocations (e.g., “When angry, I tend to focus on
month (r ⫽ .77; Sukhodolsky et al., 2001).
Emotional Control Questionnaire. The 14-item Rehearsal subscale my thoughts and feelings for a long period of time”). This factor,
(Roger & Najarian, 1989) was included in the current study (␣ ⫽ .81). This which we named Angry Rumination, consisted primarily of items
subscale assesses ruminative, grudge-holding tendencies and thoughts of from the Anger Rumination Scale (Sukhodolsky et al., 2001), but
retaliation. The subscale has adequate internal consistency and good test–
retest reliability (r ⫽ .80) and is significantly related to trait measures of
2
aggression (Roger & Najarian, 1989). This item was “I keep thinking about events that angered me for a long
Displaced aggression. We included 24 original items (␣ ⫽ .90) de- time” from the Anger Rumination Scale.
3
signed to assess the trait tendency to harm innocent others when provoked The two excluded items were “When angry, I have harmed a pet” and
(e.g., “When someone or something makes me angry I am likely to take it “I was a bully in school.” These items were originally written to assess trait
out on another person”). displaced aggression.
also included two of our a priori items designed to assess displaced viously noted, because the fourth factor (i.e., the Motivation sub-
aggression (e.g., “Sometimes I can’t help thinking about times scale from the Scott–McIntosh Rumination Inventory) was unre-
when someone made me mad”) and an additional two items from lated to the Aggression Questionnaire Physical and Verbal
the Rehearsal subscale of the Emotional Control Questionnaire Aggression subscales in both studies, we decided to drop this
(e.g., “I often find myself thinking over and over about things that factor from the final scale. This is consistent with a previous study
have made me angry”; Roger & Najarian, 1989). that failed to find relationships between the Scott–McIntosh Ru-
The second factor consisted of 10 items concerned with general mination Inventory and these subscales (Brown & Phillips, 2005).
tendencies to engage in displaced aggression (e.g., “When some- Internal consistency for the final scale was high (␣ ⫽ .94).
one or something makes me angry I am likely to take it out on
another person”). We named this factor Behavioral Displaced Discussion
Aggression.
The third factor consisted of 11 items concerned with planning Phase 2 identified items for the final scale and confirmed the
for retaliation in response to provocations (e.g., “If somebody hypothesized three-factor structure in a college student sample.
harms me, I am not at peace until I can retaliate”). We named this Specifically, exploratory and confirmatory factor analyses identi-
factor Revenge Planning. fied an affective dimension (Angry Rumination), a cognitive di-
As in Phase 1, the fourth factor to emerge was the three-item mension (Revenge Planning), and a behavioral dimension (Behav-
Motivation subscale from the Scott–McIntosh Rumination Inven- ioral Displaced Aggression). Each of these three dimensions was
tory. We did not include this factor in subsequent analyses because moderately related to self-reported direct verbal and physical trait
it was uncorrelated with self-reported direct aggressiveness in aggressiveness (Buss & Perry, 1992). To obtain further confidence
Phase 1 and Phase 2 (see below). in our hypothesized factor structure, we attempted to replicate the
An initial confirmatory factor analysis with EQS 6.1 (Bentler, findings of Phase 2 in a national community sample.
2005) was conducted to confirm the factor structure suggested by
the exploratory method. We specified a hierarchical factor analytic Phase 3: Replication of Factor Structure in a National
model in which all three subscales loaded on a higher order trait Community Sample—Evidence of Concurrent Convergent
displaced aggression factor. Investigation of Mardia’s (1970) co- and Discriminant Construct Validity
efficient suggested a significant deviation from multivariate nor-
mality, normalized estimate ⫽ 71.01. We therefore relied on a Phase 2 provided confirmatory evidence concerning the three-
robust residual-based chi-square goodness-of-fit test (Yuan & factor structure underlying our displaced aggression personality
Bentler, 1998) as well as robust goodness-of-fit indices. Signifi- dimension in a college student sample. However, we would have
cance tests for factor loadings were calculated with robust standard greater confidence in our factor structure if we could observe the
errors. As is customary in structural equation modeling, conver- same structure in a new sample. In addition, the construct validity
gent evidence of good model fit from multiple tests is highly data in Phase 2 was limited to the Physical and Verbal Aggression
desired. Because fit indices are often highly correlated, Hu and subscales of the Aggression Questionnaire (Buss & Perry, 1992).
Bentler (1999) advocated reporting nonredundant fit indices, and Before proceeding, we deemed it necessary to establish more
we adopted this strategy. Evaluation of Lagrange multiplier and thorough convergent and discriminant construct validity. To these
Wald statistics suggested that four items loaded on more than one ends, we conducted a third phase of data collection with a fairly
factor. Thus, a new model was created with these four items large national community sample of Internet respondents. Our
removed from analysis.4 The chi-square goodness-of-fit test was goals were to (a) replicate the previously obtained three-factor
nonsignificant, ␹2(431, N ⫽ 471) ⫽ 453.90, p ⫽ .22, and other structure in a new sample, (b) provide thorough concurrent con-
goodness-of-fit indices revealed a good model fit, ␹2/df ⫽ 1.05, vergent and discriminant construct validity data by coadminister-
comparative fit index (CFI) ⫽ .93, root-mean-square error of ing a variety of existing theoretically relevant personality mea-
approximation (RMSEA) ⫽ .05. Guidelines for good model fit sures, and (c) obtain normative data for our measure. To control
suggest a chi-square to degrees-of-freedom ratio of less than 2.00 for acquiescence response set, we also report on the development
and a CFI in the mid .90s, whereas RMSEA (a residual-based of two alternative reverse-keyed forms of the DAQ.
index) values of .06 or lower are believed to indicate a good fit (Hu
& Bentler, 1999; Raykov & Marcoulides, 2000; Ullman, 2001). Method
Moreover, each of the three subscale factors loaded highly and
significantly on the higher order trait displaced aggression factor.
Participants
Thus, these preliminary data revealed adequate support for our A total of 1,013 Internet respondents (mean age ⫽ 39 years, SD ⫽ 12.31,
hypothesized factor structure. The final scale with confirmatory range ⫽ 18 to 83; 84% female, 16% male) completed the survey in
factor loadings is presented in Table 2. exchange for a chance to win $200. Participants were recruited through a
To obtain preliminary concurrent construct validity, we created
composites of the four separate factors and correlated them with 4
the Physical and Verbal Aggression subscales of the Aggression These four items were “After being irritated or annoyed, I am ex-
tremely short-tempered for the rest of the day” from our displaced aggres-
Questionnaire. Internal consistency reliability was high for the
sion items; “I easily fly off the handle with those who don’t listen or
Angry Rumination (␣ ⫽ .92), Behavioral Displaced Aggression understand” from the Irritability Scale (Caprara, 1985); “When someone
(␣ ⫽ .91), and Revenge Planning (␣ ⫽ .91) factors and was insults or hurts me, I think for hours about things I could have said or done
moderate for the fourth factor (␣ ⫽ .76). Table 3 presents corre- to get even” from the Forgiveness of Others Scale (Mauger et al., 1992);
lations between the four factors and the Verbal and Physical and “Sometimes I fly off the handle for no good reason” from the Aggres-
Aggression subscales from the Aggression Questionnaire. As pre- sion Questionnaire (Buss & Perry, 1992).
Table 2
Items From the Three Subscales of the Final Questionnaire
Subscale and item Source
Angry Rumination (.80, .84)

I keep thinking about events that angered me for a long time. (.85, .87) Anger Rumination Scale
(Sukhodolsky et al., 2001)
I get “worked up” just thinking about things that have upset me in the past. (.79, .77) Emotional Control Questionnaire
(Roger & Najarian, 1989)
I often find myself thinking over and over about things that have made me angry. (.79, .84) Emotional Control Questionnaire
Sometimes I can’t help thinking about times when someone made me mad. (.67, .74) Displaced Aggression (original item)
Whenever I experience anger, I keep thinking about it for a while. (.77, .74) Anger Rumination Scale
After an argument is over, I keep fighting with this person in my imagination. (.63, .66) Anger Rumination Scale
I re-enact the anger episode in my mind after it has happened. (.71, .73) Anger Rumination Scale
I feel angry about certain things in my life. (.54, .59) Anger Rumination Scale
I think about certain events from a long time ago and they still make me angry. (.75, .83) Anger Rumination Scale
When angry, I tend to focus on my thoughts and feelings for a long period of time. (.64, .71) Displaced Aggression
Revenge Planning (.85, .78)
When someone makes me angry I can’t stop thinking about how to get back at this person. (.81, .85) Anger Rumination Scale
If somebody harms me, I am not at peace until I can retaliate. (.72, .77) Dissipation-Rumination Scale
(Caprara, 1986)
I often daydream about situations where I’m getting my own back at people. (.71, .75) Emotional Control Questionnaire
I would get frustrated if I could not think of a way to get even with someone who deserves it. (.71, .80) Forgiveness of Others Scale
(Mauger et al., 1992)
I think about ways of getting back at people who have made me angry long after the event has happened. Emotional Control Questionnaire
(.75, .78)
If another person hurts you, it’s alright to get back at him or her. (.68, .74) Forgiveness of Others Scale
The more time that passes, the more satisfaction I get from revenge. (.65, .73) Dissipation-Rumination Scale
I have long living fantasies of revenge after the conflict is over. (.73, .77) Anger Rumination Scale
When somebody offends me, sooner or later I retaliate (.62, .76) Dissipation-Rumination Scale
If a person hurts you on purpose, you deserve to get whatever revenge you can. (.59, .77) Forgiveness of Others Scale
I never help those who do me wrong. (.48, .46) Dissipation-Rumination Scale
Displaced Aggression (.60, .71)
When someone or something makes me angry I am likely to take it out on another person. (.84, .81) Displaced Aggression
When feeling bad, I take it out on others. (.81, .84) Displaced Aggression
When angry, I have taken it out on people close to me. (.74, .78) Displaced Aggression
Sometimes I get upset with a friend or family member even though that person is not the cause of my Displaced Aggression
anger or frustration. (.72, .73)
I take my anger out on innocent others. (.71, .83) Displaced Aggression
When things don’t go the way I plan, I take out my frustration on the first person I see. (.71, .74) Displaced Aggression
If someone made me angry I would likely vent my anger on another person. (.68, .77) Displaced Aggression
Sometimes I get so upset by work or school that I become hostile toward family or friends. (.71, .76) Displaced Aggression
When I am angry, I don’t care who I lash out at. (.58, .68) Displaced Aggression
If I have had a hard day at work or school, I’m likely to make sure everyone knows about it. (.56, .56) Displaced Aggression
Note. Confirmatory factor loadings (in parentheses) are from Phases 2 and 3, respectively. Following each subscale are factor loadings on higher order
trait displaced aggression.
general purpose Web site (www.about.com). Although Whites were over- multiracial, 2.8% Latino, 2.4% Black, 1.4% Asian, 1% Native American,
represented, all major ethnic groups were present (87% White, 3.5% 0.4% Middle Eastern).
Materials
Table 3
Robust Correlations Among the Four Factors From Phase 2 Hypothesized three-factor scale. We administered the 31 items con-
and the Physical and Verbal Aggression Subscales of the stituting the final version of the DAQ that were obtained via confirmatory
Aggression Questionnaire (Buss & Perry, 1992) factor analysis in Phase 2 (see Table 2).
Anger and aggression. All four subscales of the Aggression Question-
Factor or subscale 1 2 3 4 5 6 naire were administered to assess individual differences in direct aggres-
sive personality. Spielberger’s (1998) measure of anger coping styles, the
1. Physical Aggression — State–Trait Anger Expression Inventory, was administered as well. Its
2. Verbal Aggression .42* — scales assess three distinct means of expressing angry feelings (Anger In,
3. Angry Rumination .39* .35* —
Anger Out, and Anger Control). The inventory has been evaluated factor
4. Revenge Planning .60* .44* .62* —
5. Displaced Aggression .43* .45* .55* .57* — analytically (Forgays, Forgays, & Spielberger, 1997) and has been used
6. Motivation subscale ⫺.03 .13 ⫺.04 ⫺.04 ⫺.04 — extensively in the anger literature.
Impulsivity. Because impulsivity has been studied extensively in rela-
* significant, controlling for familywise error rate (␣ ⫽ .05) tion to aggressive behavior and personality, the Barratt Impulsivity Scale
(Patton, Stanford, & Barratt, 1995) was administered to assess individual (RRQ; Trapnell & Campbell, 1999), which has two subscales of the same
differences in impulsivity. The measure has good internal consistency and names. Internal consistency is excellent for both scales. Because the
has been widely used in clinical and nonclinical samples. Because indi- Rumination subscale contains items concerned with negatively valenced,
viduals who tend to be high in displaced aggression do not immediately unwanted thoughts, we expected our displaced aggression measure to
“fly off the handle” when confronted with a provocation (although they positively correlate with it. However, we did not expect our scale to
may do so later), we did not expect that impulsivity would be highly correlate with the Reflection subscale, because these items are positively
correlated with our measure of displaced aggression. valenced and are concerned with volitional self-understanding.
Trait affect. To demonstrate the relationship between levels of trait Need for cognition. Cacioppo and Petty (1982) introduced a measure
affect and our displaced aggression measure, we included the Positive and of individual differences in the tendency to enjoy thoughtful, cognitively
Negative Affect Schedule (PANAS General; Watson et al., 1988). Because demanding activities. The scale taps a unitary construct and has good
our construct contains a considerable negative affect component, it was internal consistency and demonstrated convergent and discriminant valid-
hypothesized that our scale would be positively correlated with trait neg- ity. We included their Need for Cognition Scale (Cacioppo & Petty, 1982)
ative affect but negatively correlated with positive affect. The two affect to provide evidence that angry rumination and revenge planning are dis-
factors have good internal consistency and good 8-week test–retest reli- tinct from a more general preference for thoughtful cognitive activity.
ability (rs ⫽ .68 and .71; Watson et al., 1988). Therefore, we expected no reliable relation between the Need for Cognition
Big Five. One of the most robust findings in personality research is the Scale and the DAQ.
existence of the Big Five personality dimensions (Macrae & Allik, 2002). Social desirability. Finally, we included a short form of the Marlowe–
Goldberg’s (1990, 1992) 50-item inventory was used for the current study. Crowne social desirability measure (Marlowe & Crowne, 1961) to assess
We hypothesized that neuroticism would be positively correlated with our our construct’s relationship to social desirability (Strahan & Gerbasi,
scale whereas agreeableness and conscientiousness would be negatively 1972). The short form is highly correlated with the full version (with rs in
correlated with it. Extroversion and openness were expected to be uncor- the .90s). Owing to the sensitive nature of data on aggressive personality,
related with our displaced aggression measure. We derived these expecta- measures of social desirability and aggression tend to be moderately
tions from previous research on the relationship between the factors of the correlated (approximate r ⫽ ⫺.50; see Harris, 1997; Lange, Dehghani, &
Big Five and trait aggression (Ang et al., 2004; Tremblay & Ewart, 2005). de Beurs, 1995; Lange, Pahlich, et al., 1995; Morren & Meesters, 2002).
Internal consistency is good for all of the five dimensions. We therefore expected a moderate correlation between social desirability
Self-esteem. We assessed self-esteem with Rosenberg’s (1965) 10-item
and the DAQ.
measure. The scale has been in use for 40 years and has good psychometric
Domestic abuse and road rage. We also included two indirect indica-
properties. Previous research has demonstrated that high levels of self-
tors of displaced aggression. Because it is hypothesized that individuals
esteem (e.g., narcissism) are associated with increased direct aggressive-
high in displaced aggression tend to take it out on individuals close to them,
ness in response to personal insult (Baumeister, Bushman, & Campbell,
we included a measure of domestic abuse. The Abuse Within Intimate
2000; Bushman & Baumeister, 1998; Bushman, Bonacci, van Dijk, &
Relationships Scale (AWIRS; Borjesson, Aarons, & Dunn, 2003) is a
Baumeister, 2003). At the same time, however, Bushman and Baumeister
26-item self-report measure wherein participants indicate their abusive
(1998) found no relationship between laboratory displaced aggression and
behaviors on a scale ranging from never to more than once a day. Items
either the Rosenberg scale or a measure of narcissism (Bushman &
range from “criticized” and “belittled” to “physically attacked” and “used
Baumeister, 1998). Nevertheless, it was conceivable that our trait measure
an object to hit.” The scale has a five-factor structure assessing Emotional
would be related to self-esteem. Therefore, we included the Rosenberg
Abuse (e.g., insulting, belittling), Deception (e.g., lying, keeping secrets),
scale to determine the relationship of our trait measure to a standard
Verbal Abuse (e.g., using profanity, screaming), Overt Violence (e.g.,
measure of self-esteem.
using an object to hit, physically attacking), and Restrictive Violence (e.g.,
Norms of reciprocity. Recently Eisenberger, Lynch, Aselage, and
Rohdieck (2004) presented an individual-difference measure that discrim- grabbing arm, forcefully squeezing), with good internal consistency for
inates between norms of positive reciprocity (e.g., helping) and negative each of these subscales.
reciprocity (e.g., retaliation). The scales have good internal consistency and In addition, because it is believed that individuals high in displaced
good predictive validity. Because one of our factors is Revenge Planning, aggression may be aggressively primed owing to dwelling on angry feel-
we hypothesized that the norm of negative reciprocity would be highly ings and thoughts of revenge, we hypothesized that these individuals would
correlated with this subscale whereas the norm of positive reciprocity experience a high level of aggression during driving. To this end, we
would be negatively correlated with our full measure of displaced included the Driving Vengeance Questionnaire (Wiesenthal, Hennessy, &
aggression. Gibson, 2000). This 15-item measure presents a series of commonly
Behavioral approach and inhibition. We included the Behavioral Ap- encountered yet potentially annoying driving situations (e.g., “The car in
proach and Inhibition Scales (BIS/BAS; Carver & White, 1994) to assess front of you doesn’t proceed on an advanced green signal”). Respondents
individual differences in these behavioral orientations. The scales have were asked to indicate which among four behavioral options they would
good internal consistency and have demonstrated predictive validity. Re- engage in when confronting each situation. Ordinal options range from
cent research has demonstrated that anger is related to the behavioral nonaggressive (e.g., do nothing) to extremely aggressive (e.g., bump into
approach system (Harmon-Jones, 2003; Harmon-Jones & Allen, 1998; the other car) responses. The scale has good internal consistency. We
Harmon-Jones & Sigelman, 2001). However, we expected individuals high hypothesized that the DAQ would be a stronger predictor of these indirect
in displaced aggression to be conflicted regarding approach and avoidance indicators of displaced aggression than the Aggression Questionnaire.
tendencies. Specifically, whereas revenge planning represents an approach We consider these self-report measures of domestic abuse and road rage
orientation, the failure of those with high trait displaced aggression to to be indirect indicators of displaced aggression. However, because of the
respond immediately to provocations suggests a strong inhibition compo- self-report nature of these measures, we are not privy to whether prior
nent. Therefore, we expected our scale to be more strongly associated with provocation or immediate circumstances instigated the aggressive act. We
behavioral inhibition than behavioral approach. suspect that in many instances these individuals do aggress against their
Rumination–reflection. As discussed earlier, Trapnell and Campbell partners or fellow drivers in response to provoking or frustrating behaviors.
(1999) investigated the distinction between rumination and reflection. They However, we cannot decisively conclude this, and we recognize that we are
defined rumination as negative inward focus, whereas reflection is a making an assumption in considering these measures indicators of dis-
psychologically beneficial self-focus related to self-improvement and un- placed aggression. Rather, such self-reported behavior may reflect habitual
derstanding. We included their Rumination–Reflection Questionnaire modes of responding, not actions that are initiated by specific antecedent
provocations. (We specifically address this consideration in Phase 5 by Table 4

experimentally manipulating provocation.) Age Norms for the Angry Rumination, Revenge Planning, and
Displaced Aggression Subscales
Procedure Subscale Age group M SD
Participants completed the study from a location of their choosing (e.g., Angry Rumination 18–24 (n ⫽ 138) 3.98 1.39
home, work, school) as part of an Internet survey. All items were rated on 25–34 (n ⫽ 272) 3.98 1.41
bipolar 7-point Likert-type scales ranging from 1 (extremely uncharacter- 35–44 (n ⫽ 244) 3.73 1.41
istic of me) to 7 (extremely characteristic of me). Participants were asked 45–54 (n ⫽ 226) 3.75 1.50
to complete the measures honestly and were informed that their responses 55–64 (n ⫽ 96) 3.42 1.46
would remain anonymous. Owing to concerns about participant fatigue, 65–83 (n ⫽ 20) 3.27 1.42
each participant completed (a) the items composing the hypothesized Revenge Planning 18–24 (n ⫽ 138) 2.56 1.19
25–34 (n ⫽ 272) 2.51 1.22
three-factor scale and (b) a subset of the validity scales such that each
35–44 (n ⫽ 244) 2.28 1.15
participant completed approximately 115 items. To control for order ef- 45–54 (n ⫽ 226) 2.33 1.23
fects, with two exceptions, all items were randomized to create 10 versions 55–64 (n ⫽ 96) 2.04 1.01
of the questionnaire. The two exceptions were the Driving Vengeance 65–83 (n ⫽ 20) 2.23 1.18
Questionnaire and the AWIRS. Because these scales require unique re- Displaced Aggression 18–24 (n ⫽ 138) 3.03 1.24
sponse options, all of the items for these scales were presented consecu- 25–34 (n ⫽ 272) 2.97 1.24
tively (although the location of the scale placement remained random). 35–44 (n ⫽ 244) 2.57 1.22
45–54 (n ⫽ 226) 2.47 1.16
55–64 (n ⫽ 96) 2.27 1.06
Results and Discussion 65–83 (n ⫽ 20) 2.16 1.34
Items were reverse scored when necessary. Missing values were

replaced with the mean for that item. Among participants who Internal Consistency Reliability
received any single item, missing data did not exceed 2.9%.
Internal consistency reliability was high for the total scale
(Cronbach’s alpha ⫽ .95, Spearman–Brown split-half r ⫽ .86) and
Confirmatory Factor Analysis subscales (see Table 5).
To gain further confidence in the factor structure of our mea-
sure, we wished to replicate the results of the confirmatory factor Concurrent Convergent and Discriminant Construct
analysis from Phase 2. Therefore, an identical hierarchical model Validity
was specified. As in Phase 2, investigation of Mardia’s (1970) Table 5 reports the correlations between each of the three
coefficient suggested a significant deviation from multivariate subscales of our DAQ and each of the self-report measures. No
normality, normalized estimate ⫽ 89.71. We therefore relied on gender differences were observed after controlling for familywise
robust statistical methods. Although the chi-square goodness-of-fit error rate (␣ ⫽ .05). We briefly discuss some of the relationships
test (Yuan & Bentler, 1998) was significant (a common occurrence among each of the three subscales of our DAQ and the other
in large samples; see Ullman, 2001), ␹2(431, N ⫽ 1,103) ⫽ measures and provide further evidence of construct validity by
656.15, p ⬍ .001, other goodness-of-fit indices revealed good demonstrating the DAQ’s relationship to measures of domestic
model fit, ␹2/df ⫽ 1.52, CFI ⫽ .94, RMSEA ⫽ .05. In addition, all abuse and road rage.
three subscales loaded significantly on the higher order displaced Angry Rumination. In general, the Angry Rumination subscale
aggression factor (see Table 2). Thus, we successfully replicated of the DAQ correlated in the expected directions with the criterion
the factor structure observed in Phase 2. measures. For example, the subscale was positively correlated with
hostility, anger, negative affect, rumination, and neuroticism. Also,
Normative Data as expected, the correlation of this subscale with behavioral inhi-
bition was moderately positive, which is opposite to its generally
Age was weakly, albeit negatively, correlated with all three observed relationship with trait measures of anger and direct
displaced aggression subscales as well as the total score (Angry aggression (e.g., Harmon-Jones, 2003; Harmon-Jones & Allen,
Rumination, r ⫽ ⫺.11; Behavioral Displaced Aggression, r ⫽ 1998; Hewig et al., 2004). Additionally, Angry Rumination was
⫺.23; Revenge Planning, r ⫽ ⫺.13; and total score, r ⫽ ⫺.17; negatively correlated with positive affect and self-esteem, consci-
robust correlations, all ps ⬍ .05, controlling for familywise error entiousness, agreeableness, and social desirability. The subscale
rate). Comparisons of the youngest and oldest age groups for each was unrelated to extroversion, openness to experience, reflection,
of the three subscales revealed that Behavioral Displaced Aggres- and behavioral approach.
sion, T*y ⫽ 5.65, p ⬍ .001 (bootstrap Yuen’s robust t test; Wilcox, Revenge Planning. This scale correlated positively with direct
2005, p. 162), but not Angry Rumination or Revenge Planning, physical aggression, trait hostility, and, most strongly, with the
decreased with age. norm of negative reciprocity, while being negatively related to
Men and women did not differ on the Angry Rumination or anger control, agreeableness, conscientiousness, and social desir-
Behavioral Displaced Aggression subscales but did differ on Re- ability. The scale was unrelated to openness to experience and
venge Planning, T*y ⫽ 4.26, p ⬍ .001. Men rated themselves reflection.
higher in Revenge Planning than women (Ms ⫽ 2.76 vs. 2.29). Behavioral Displaced Aggression. This behavioral subscale
Table 4 displays the means and standard deviations by age group. correlated positively with trait anger, anger out, negative affect,
Table 5
Concurrent Construct Validity Data (Robust Correlations)
Angry Rumination Revenge Planning Displaced Aggression

Construct n ␣ (␣ ⫽ .927) (␣ ⫽ .930) (␣ ⫽ .926)
Physical Aggression 224 .81 .26* .50* .41*

Verbal Aggression 201 .70 .28* .36* .34*
Trait Anger 196 .78 .53* .54* .63*
Trait Hostility 182 .77 .58* .59* .49*
Anger In 196 .74 .57* .42* .32*
Anger Out 196 .80 .45* .55* .64*
Anger Control 196 .88 ⫺.41* ⫺.45* ⫺.59*
Impulsivity 210 .85 .38* .31* .43*
Negative Affect 224 .92 .60* .49* .53*
Positive Affect 210 .92 ⫺.40* ⫺.34* ⫺.28*
Neuroticism 224 .81 .62* .42* .54*
Extroversion 201 .89 ⫺.24 ⫺.17 ⫺.13
Openness 182 .78 ⫺.01 ⫺.11 ⫺.11
Agreeableness 196 .81 ⫺.53* ⫺.71* ⫺.60*
Conscientiousness 210 .87 ⫺.35* ⫺.27* ⫺.30*
Negative Reciprocity 210 .94 .47* .81* .50*
Positive Reciprocity 224 .84 .22 .26* .17
Behavioral Inhibition 182 .79 .58* .18 .38*
Behavioral Approach 182 .82 .11 .09 .10
Rumination 201 .86 .81* .36* .39*
Reflection 201 .93 .02 ⫺.04 ⫺.24
Need for Cognition 224 .91 ⫺.33* ⫺.32* ⫺.29*
Social Desirability 182 .66 ⫺.56* ⫺.58* ⫺.62*
Domestic Abuse 197 .92 .38* .36* .39*
Road Rage 170 .78 .28* .56* .36*
* significant, controlling for familywise error rate within each sample (␣ ⫽ .05).
neuroticism, and behavioral inhibition, while being negatively the variance inflation factors and tolerance (using the suggested
correlated with anger control, agreeableness, conscientiousness, formula: tolerance ⫽ 1 – R2; see Tabachnick & Fidell, 2001)
and social desirability. The subscale was unrelated to extroversion, revealed that these results could not be attributable to collinearity
openness to experience, and behavioral approach. among the predictors.
Road rage and domestic abuse. We suggested previously that Because much of domestic abuse can be considered verbal
the measures of road rage and domestic abuse could serve as aggression, we also examined the effects of the Aggression Ques-
indirect indicators of displaced aggression. If true, our displaced tionnaire Verbal Aggression subscale relative to the DAQ in
aggression measure (the DAQ) should predict these measures, predicting the three subscales of the AWIRS (Emotional Abuse,
whereas the measure of direct aggression (the Aggression Ques- Deception, and Verbal Abuse; n ⫽ 197). Our hypotheses were
tionnaire) should be unrelated to domestic abuse and road rage. To supported in all three of these regression analyses. The DAQ
test this hypothesis, we simultaneously regressed scores from the predicted Emotional Abuse (␤ ⫽ .40, p ⬍ .001), but the Aggres-
Driving Vengeance Questionnaire (Wiesenthal et al., 2000) and sion Questionnaire Verbal Aggression subscale did not (␤ ⫽ .04,
each of the five subscales from the AWIRS (Borjesson et al., 2003) p ⫽ .63). The DAQ also predicted Deception (␤ ⫽ .37, p ⬍ .001),
on both the composite scores from the Aggression Questionnaire whereas the Aggression Questionnaire Verbal Aggression subscale
and our displaced aggression measure. This Aggression Question- was negatively related to Deception (␤ ⫽ ⫺.21, p ⫽ .005). Finally,
naire composite reflects a separate subset of the sample (n ⫽ 182) the DAQ predicted verbal abuse (␤ ⫽ .43, p ⬍ .001), whereas the
who received the Aggression Questionnaire Anger and Hostility Aggression Questionnaire Verbal Aggression subscale did not
subscales. The DAQ predicted scores on the Driving Vengeance (␤ ⫽ ⫺.09, p ⫽ .23).
Questionnaire (␤ ⫽ .59, p ⬍ .001), but the Aggression Question- Additional analyses controlling for theoretically relevant mea-
naire did not (␤ ⫽ ⫺.09, ns). The DAQ also predicted scores on sures. As stated previously, not all participants received all of the
the Verbal Abuse subscale from the AWIRS (␤ ⫽ .35, p ⬍ .001), measures. We were therefore unable to test the effects of the DAQ
but the Aggression Questionnaire did not (␤ ⫽ .09, ns). Further- controlling for all possible theoretically relevant measures. How-
more, the DAQ predicted emotional abuse (␤ ⫽ .59, p ⬍ .001) and ever, we report on those that we were able to conduct. We
deception (␤ ⫽ .43, p ⬍ .001), whereas scores on the Aggression examined the effects of the DAQ on road rage while controlling
Questionnaire were unrelated to emotional abuse (␤ ⫽ .04, ns) and for social desirability (n ⫽ 169). As expected, the DAQ predicted
were negatively related to deception (e.g., keeping secrets, lying; road rage (␤ ⫽ .38, p ⬍ .001), whereas social desirability (i.e., the
␤ ⫽ ⫺.24, p ⫽ .005). Neither the DAQ nor the Aggression Marlowe–Crowne scale) was negatively related to road rage (␤ ⫽
Questionnaire predicted scores on the Overt Violence (e.g., using ⫺.21, p ⫽ .02). Moreover, these effects were not due to multicol-
an object to hit, physically attack) and Restrictive Violence (e.g., linearity. We also analyzed the effects of the DAQ on road rage
grabbing arm, forcefully squeezing) subscales; this lack of relation while controlling for openness, self-esteem, and BIS/BAS (n ⫽
was likely due to restriction of range. Furthermore, inspection of 170). All of these tests supported the construct validity of the
DAQ. In a simultaneous regression, the DAQ was a strong and Phase 4: Test–Retest Reliability
significant predictor of road rage (␤ ⫽ .60, p ⬍ .001), whereas
neither openness (␤ ⫽ ⫺.03, p ⫽ .72), self-esteem (␤ ⫽ .06, p ⫽ Two studies were conducted to obtain evidence of test–retest
.49), BIS (␤ ⫽ ⫺.14, p ⬍ .08), nor BAS (␤ ⫽ .03, p ⫽ .70) reliability. In the first study, 133 USC and CSULB undergraduates
significantly predicted road rage. participated in exchange for extra course credit. Participants were
In addition, we were able to assess the effects of the DAQ on the told that the study consisted of two sessions. At Time 1, they
three domestic abuse subscales while controlling for extroversion completed the 31 items from the DAQ via the Internet. Four weeks
and rumination–reflection (n ⫽ 197). Results from these three later, all participants were contacted via e-mail and asked to
simultaneous regression analyses supported the discriminant con- complete the second portion of the study. This 4-week test–retest
struct validity of the DAQ. The DAQ significantly predicted reliability coefficient for the total scale was acceptable (r ⫽ .77,
emotional abuse (␤ ⫽ .48, p ⬍ .001), whereas neither extroversion p ⬍ .001). The test–retest coefficients for the individual subscales
(␤ ⫽ .00, p ⫽ .95), rumination (␤ ⫽ ⫺.10, p ⫽ .32), nor reflection were also acceptable: Angry Rumination (r ⫽ .80, p ⬍ .01),
(␤ ⫽ ⫺.04, p ⫽ .54) was a significant predictor. The DAQ also Revenge Planning (r ⫽ .75, p ⬍ .01), and Behavioral Displaced
significantly predicted verbal abuse (␤ ⫽ .48, p ⬍ .001), whereas Aggression (r ⫽ .78, p ⬍ .01).
neither extroversion (␤ ⫽ .03, p ⫽ .64) nor rumination (␤ ⫽ ⫺.13, In the second study, 101 USC undergraduates completed the
p ⫽ .19) predicted verbal abuse. Reflection was negatively related study on the Internet in exchange for extra course credit during a
to verbal abuse (␤ ⫽ ⫺.16, p ⫽ .03). For the Deception subscale 3-week period at the beginning of the semester. During the last 3
of the AWIRS, the DAQ was a marginally significant predictor weeks of the semester, the participants were contacted via e-mail
(␤ ⫽ .18, p ⬍ .07). However, neither extroversion (␤ ⫽ .03, p ⫽ and asked to complete the second questionnaire (mean days fol-
.68), rumination (␤ ⫽ .17, p ⫽ .11), nor reflection (␤ ⫽ ⫺.04, p ⫽ lowing completion of first questionnaire ⫽ 77.20, SD ⫽ 8.57).
.57) was a significant predictor of deception. This 11-week test–retest reliability coefficient for the total scale
was excellent (r ⫽ .87, p ⬍ .001). Again, these test–retest coef-
ficients for the individual subscales were also acceptable: Angry
Reverse-Keyed Forms Rumination (r ⫽ .89, p ⬍ .01), Revenge Planning (r ⫽ .86, p ⬍
.01), and Behavioral Displaced Aggression (r ⫽ .78, p ⬍ .01).
Because all of the items in the DAQ are direct keyed, one may
be concerned about the potential confounding effect of acquies-
cence bias. To address this issue, we created two forms of the
Phase 5: Behavioral Evidence of Construct Validity
DAQ with reverse-keyed items. We wrote reverse-keyed original We conducted two experiments to establish further construct
items for each of the 31 DAQ items and administered them via the validity of the DAQ. In Experiment 1, participants served in the
Internet to a community sample (N ⫽ 205; 88% female, 12% male) four cells of the TDA paradigm (Pedersen et al., 2000). They were
along with three validity measures (neuroticism, agreeableness, exposed or not exposed to a provocation from the experimenter.
and behavioral inhibition). For example, the item “I reenact the Then, half of each group received and half did not receive a mild
anger episode in my mind after it has happened” was reversed to triggering event from a bogus participant. Participants were then
“I move on to other things after an anger episode has happened,” allowed to deliver a noxious physical stimulus to the bogus par-
and the item “If someone made me angry, I would likely vent my ticipant (i.e., displaced aggression). In Experiment 2, a situational
anger on another person” was changed to “If someone made me rumination condition was introduced into the TDA paradigm. We
angry, I would tell them how I feel.” Form 1 consisted of 15 induced participants to either ruminate about the provocation,
randomly chosen reverse-keyed items and the 16 direct-keyed ruminate about themselves, or distract themselves before engaging
items. Form 2 consisted of the remaining 16 reverse-keyed items in displaced aggression. In both studies, it was expected that the
and 15 direct-keyed items. These reverse-keyed forms (␣s ⫽ .91 DAQ would moderate the degree of laboratory displaced aggres-
and .94 for Forms 1 and 2, respectively) demonstrated good sion but that other related personality variables of interest would
internal consistency, although these estimates were slightly lower not moderate actual displaced aggression. Specifically, we ex-
than for the original direct-keyed form, especially for the sub- pected those high in trait displaced aggression to ruminate angrily
scales: Angry Rumination (␣ ⫽ .89), Behavioral Displaced Ag- and focus on thoughts of retaliation after being provoked by the
gression (␣ ⫽ .84), and Revenge Planning (␣ ⫽ .87). The direct- experimenter, whereas those low in trait displaced aggression were
keyed items correlated highly with the reverse-keyed items for not expected to ruminate angrily or plan revenge. Consistent with
both Form 1 (r ⫽ ⫺.70, p ⬍ .001) and Form 2 (r ⫽ ⫺.85, p ⬍ the GAM (C. A. Anderson & Bushman, 2002) and cognitive–
.001) (robust correlations; Wilcox, 2005, p. 407). Using the three affective models of personality (Mischel & Shoda, 1995), higher
DAQ subscales as indicators of the trait displaced aggression levels of trait displaced aggression should increase the accessibil-
factor, we constrained these factor loadings to those from the ity of aggression-related affect, cognition, and arousal, which in
Phase 3 data. This conservative test of the reverse-keyed factor turn should increase displaced aggression when one is given the
structure provided an excellent fit to the data, ␹2(3, N ⫽ 205) ⫽ opportunity to take it out on an undeserving other. In Experiment
2.78, p ⫽ .25, ␹2/df ⫽ 0.93, CFI ⫽ .96, RMSEA ⫽ .04. The total 2, we manipulated rumination and expected that this would have
scale also correlated in the expected directions with neuroticism an especially strong effect on those high in trait rumination be-
(r ⫽ .66), agreeableness (r ⫽ ⫺.67), and behavioral inhibition cause both situation and person inputs are expected to produce
(r ⫽ .40) (robust correlations, all ps ⬍ .05, controlling for fami- increases in displaced aggression according to the GAM.
lywise error rate). These reverse-keyed forms are available upon The goal of Experiment 1 was to obtain behavior-related con-
request, although the remainder of our studies used the direct- struct validity evidence for our Behavioral Displaced Aggression
keyed form because of its slightly higher reliability. subscale. In Phase 3, we obtained correlational evidence that the
full DAQ scale, as well as the individual subscales, were related to anteit ⫽ lieutenant). The experimenter informed the participant that he or
indirect indicators of displaced aggression (e.g., domestic abuse she would have 3.5 min to complete all 15 anagrams and left the room.
and road rage). If, in fact, our Behavioral Displaced Aggression When the 3.5 min had elapsed, the experimenter reentered, took the
subscale is a valid indicator of trait displaced aggression, we anagram answer sheet, and left the room ostensibly to score the partici-
should expect it to predict laboratory displaced aggression as well. pant’s performance.
To this end, participants completed our Behavioral Displaced Provocation manipulation. In the provocation condition, participants
were told that their performance was far below average compared with a
Aggression subscale and the Aggression Questionnaire. The Ag-
sample of engineering students. Furthermore, the experimenter insulted
gression Questionnaire is likely the most widely used instrument
participants in an irritated and exasperated tone of voice: “You really got
of aggressive personality in nonclinical populations. A Social a lot of these wrong. I should really give you another anagram task to do
Science Citation Index search revealed over 350 citations of the over again. However, to be perfectly honest with you, I don’t want to waste
Buss and Perry (1992) article in which the questionnaire was my time.” In the no provocation condition, participants were told that their
introduced. It is an excellent predictor of laboratory and real-world performance was average compared with a sample of engineering students
aggression (Bushman, 1995; Bushman & Wells, 1998; Tremblay and were not insulted.
& Belchevski, 2004). For example, Bushman and Wells (1998) Trigger manipulation. Participants were then asked to list desirable
found the Aggression Questionnaire Physical Aggression subscale traits in an astronaut (Bettencourt, Brewer, Croak, & Miller, 1992; Vasquez
to be a good predictor of time spent by hockey players in the et al., 2005). The experimenter then appeared and took the participant’s
penalty box. astronaut task sheet, ostensibly to give it to the other participant. Two
A second goal of Experiment 1 was to assess emotional reac- minutes later, the experimenter returned with the bogus astronaut task sheet
tivity to the source of a minor annoyance (i.e., the triggering and an evaluation form for the participant to fill out. Allegedly, it would be
agent). We believe that following a provocation, individuals high exchanged with the other participant. Similar to procedures used in prior
in trait displaced aggression maintain a negative affective state research (Pedersen et al., 2000, Study 2; Vasquez et al., 2005), this
exchange of evaluation forms served as the Time 2 trigger manipulation.
more intensely and for a longer duration than those low in trait
To implement the trigger conditions, participants received from the other
displaced aggression. Therefore, we also expected that our mea-
participant an evaluation of the degree to which his or her performance on
sure should predict reactions to the bogus participant, such that the the astronaut task exhibited originality, quality, effort, and variety among
Behavioral Displaced Aggression subscale would predict negative traits listed and the degree to which it made sense. In addition, an overall
emotional reactions to the other participant, whereas the Aggres- evaluation was provided. In the trigger condition the individual ratings and
sion Questionnaire Physical Aggression Subscale should not pre- overall evaluation were 3, 4, 3, 3, 4, and 4, respectively, on 7-point
dict these reactions. In addition, because the Behavioral Displaced Likert-type scales (1 ⫽ not good at all, 7 ⫽ extremely good). In addition,
Aggression subscale assesses post-provocation trait differences, space was available for participants to indicate additional comments. In this
we did not expect scores on this measure to be related to reactions space, the following statement was written: “The performance was not
to the initial provocation. great and I think a college student could do better.” In the no trigger
condition, participants received a neutral evaluation (6, 5, 6, 5, 5, and 5)
and the following statement: “My partner did a decent job. I think the task
Experiment 1: Method
was well done.”
Participants and Design Displaced aggression. After returning to the room, the experimenter
informed participants that the final task would examine how sensory
One hundred twenty CSULB undergraduates (79% female, 21% male) distraction affects a person’s decision-making and impression formation
enrolled in introductory psychology courses completed the study in ex- abilities. The experimenter then indicated that the participant and the other
change for extra course credit. The design was a 2 (provocation: yes or participant would receive different distraction tasks. Participants were told
no) ⫻ 2 (trigger: yes or no) between-subjects factorial. Participants were that they had been randomly assigned to the visual distraction condition
randomly assigned to one of the four experimental conditions. (watching a pleasant nature video) and that the other participant had been
assigned to the tactile distraction condition (placing their hand in painfully
Procedures cold water). Participants were then required to place their own hand in a
bucket of cold water (10 °C) for 5 s, ostensibly so that they could best
Participants took part in a TDA study (Miller et al., 2003). Similar
decide the length of distraction for the other participant. The participant
procedures have been described elsewhere (Pedersen et al., 2000; Vasquez
was also informed that the other participant was simultaneously previewing
et al., 2005) but are briefly presented here as well. As indicated, in the TDA
the nature video and would be making a similar decision. Next, participants
paradigm, participants receive a Time 1 provocation (or not), followed by
were instructed to circle on a sheet of paper the amount of time for which
a mild Time 2 triggering event (or not) from another participant. In the
absence of provocation, the triggering event does not increase aggression. the other participant should be distracted on a 9-point Likert-type scale
However, when previously provoked and exposed to a mild triggering starting at 1 (no distraction at all) and increasing by 10-s intervals to 9 (80
event, participants display disjunctively escalated aggression (Bushman et seconds/very strong distraction). This value served as the dependent mea-
al., 2005; Pedersen et al., 2000; Vasquez et al., 2005). Upon arrival to the sure. Participants were asked to slide the sheet under the door so that a
laboratory, participants were told that the study was investigating person- second research assistant could administer the task to the bogus participant.
ality, cognitive ability, and social impression formation. Specifically, par- Participants then completed the remaining dependent measures at their own
ticipants were told that they would complete personality measures and a pace.
test of general cognitive ability and then interact with another (bogus) Secondary dependent measures. To assess affect from the provocation,
participant in another room. we had participants complete a modified version of Mood Adjective
After providing informed consent, participants completed the Behavioral Checklist (Nowlis, 1965). Specifically, participants rated the degree to
Displaced Aggression subscale of the DAQ and the Aggression Question- which they experienced each of 26 emotions in relation to the provocation
naire (Buss & Perry, 1992). The experimenter then told the participants that (with these instructions: “Each of the following words describes feelings or
the first part of the study involved a test of cognitive ability. Specifically, moods. Please use the list to describe your feeling after finishing the
participants completed a sheet with 15 difficult anagrams (e.g., elun- anagram task you completed at the beginning of the study”). Each emo-
tional descriptor was rated on a 7-point Likert-type scale (1 ⫽ not at all, high levels of trait displaced aggression (one SD above the mean),
7 ⫽ extremely so). provocation significantly predicted displaced aggression (␤ ⫽ .30,
An additional five items assessed the emotional reaction to the bogus p ⫽ .01), but not at low levels of trait displaced aggression (one SD
participant. Specifically, participants were asked to rate how happy, below the mean; ␤ ⫽ ⫺.06, p ⫽ .60). The Behavioral Displaced
pleased, annoyed, irritated, and angered or upset they felt upon receiving
Aggression subscale did not interact with trigger condition (nor did
the evaluation from the bogus participant. Each item was rated on a 7-point
the Aggression Questionnaire), suggesting that the subscale as-
Likert-type scale (1 ⫽ not at all, 7 ⫽ extremely so).
sesses individual differences in one’s tendency to aggress against
innocent others when provoked regardless of specific situational
Experiment 1: Results features (e.g., receiving a second minor annoyance from the trig-
Displaced Aggression ger). Neither the DAQ ⫻ Provocation ⫻ Trigger interaction (␤ ⫽
.05, p ⫽ .79) nor the Aggression Questionnaire ⫻ Provocation ⫻
Data from 6 participants were removed owing to suspicion of Trigger interaction (␤ ⫽ ⫺.06, p ⫽ .63) was significant.
the study hypotheses. We first attempted to replicate previous
TDA findings, which show disjunctively escalated aggression
among provoked participants exposed to a minor annoyance. Be- Secondary Dependent Measures
cause several prior TDA studies have demonstrated disjunctively
The mood adjectives intended to assess emotional reactions to
escalated displaced aggression only in the yes provocation/yes
the provocation formed acceptably reliable composites of positive
trigger conditions (Pedersen et al., 2000; Vasquez et al., 2005), we
(␣ ⫽ .88) and negative affect (␣ ⫽ .72). Relative to the no-
used a planned contrast to test this condition against the other three
provocation condition, participants in the provocation condition
conditions (weights 3, ⫺1, ⫺1, ⫺1). As expected, a bootstrap
reported less positive affect (Ms ⫽ 9.68 vs. 15.63), T*y ⫽ ⫺3.79,
linear contrast on 20% trimmed means (Wilcox, 2005, p. 297)
p ⬍ .001, and more negative affect (Ms ⫽ 8.91 vs. 4.07), T*y ⫽
replicated the pattern of results found in previous research with the
8.27, p ⬍ .001. Thus, it appears that the provocation manipulation
TDA paradigm (␺ ⫽ 3.57, p ⬍ .001; see Table 6).5
was successful. Neither the Aggression Questionnaire Physical
To provide a high power test of our primary hypotheses, we
Aggression nor our Behavioral Displaced Aggression subscale
used regression analyses (vs. a dichotomous split) with the Behav-
predicted the reaction to the provocation. Because the DAQ is
ioral Displaced Aggression subscale (␣ ⫽ .91) and Aggression
concerned with individual differences following a provocation, we
Questionnaire Physical Aggression subscale (␣ ⫽ .85) entered as
expected that those high in trait displaced aggression would not
continuous variables and the provocation and trigger conditions
immediately experience the provocation as more aversive than
entered as dummy-coded variables (0 ⫽ no; 1 ⫽ yes).6 A hierar-
those low in trait displaced aggression. However, we expected that
chical regression analyses was conducted. At the first step, we
they would ruminate throughout the course of the experiment,
entered the provocation and trigger conditions as well as the
which would lead to the increased displaced aggression toward the
mean-centered Behavioral Displaced Aggression subscale and Ag-
bogus participant that we observed in this experiment.
gression Questionnaire Physical Aggression subscales. As ex-
The items designed to assess the reaction to the trigger also
pected, the Behavioral Displaced Aggression subscale predicted
formed a reliable composite (␣ ⫽ .94). Participants in the trigger
laboratory displaced aggression (␤ ⫽ .30, p ⫽ .001), but the
condition reported a more negative reaction toward the bogus
Aggression Questionnaire Physical Aggression subscale did not
participant than those in the no-trigger condition (Ms ⫽ 23.67 vs.
(␤ ⫽ .03, ns). At the second step, we entered interaction terms of
the Behavioral Displaced Aggression and Aggression Question-
naire Physical Aggression subscales with the provocation and 5
In the presence of even small violations of assumptions, traditional
trigger conditions. Only a significant Behavioral Displaced Ag- analysis of variance methods may poorly estimate mean differences be-
gression ⫻ Provocation interaction emerged (␤ ⫽ .23, p ⬍ .05). tween groups. Indeed, Wilcox and Keselman (2003) reviewed a substantial
Post hoc analyses (Aiken & West, 1991) revealed that among body of evidence suggesting that traditional methods of inferential statis-
provoked participants, the Behavioral Displaced Aggression sub- tics based on means perform poorly under most circumstances encountered
scale significantly predicted displaced aggression (␤ ⫽ .39, p ⬍ in psychological research (e.g., heavy tails, slight skewness, heteroscedas-
.03), but not among unprovoked participants (␤ ⫽ .06, ns). An ticity). These authors demonstrated that bootstrap methods and analyses
with trimmed means provide superior performance relative to traditional
alternative approach is to examine the effects of provocation at
procedures. Specifically, modern methods accurately control Type I error
high and low levels of trait displaced aggression. Specifically, at
rate, provide increased power, and tolerate violations of the homogeneity
and normality assumptions. The bootstrap linear contrast computes 600
bootstrap means (sampling with replacement) and, for each bootstrap
Table 6 mean, a 20% trimmed mean whereby the upper and lower 20% of the data
Aggression 20% Trimmed Means, Winsorized Standard have been removed. All of the robust statistics were computed using
Deviations, and Sample Sizes for Experiment 1 functions written by Rand Wilcox for the computer program R. R is
available for free at http://www.r-project.org, and the functions are also
Provocation No provocation available for free at http://www-rcf.usc.edu/⬃rwilcox/.
6
Because our main dependent variable was a physical aggression mea-
Condition M SD n M SD n sure, we focused our analyses on the Physical Aggression subscale of the
Aggression Questionnaire. The same pattern of results was obtained for all
Trigger 46.00 13.25 25 35.71 16.45 33
No trigger 23.81 12.26 33 26.00 8.39 23 three Aggression Questionnaire subscales as well as the Aggression Ques-
tionnaire total score (i.e., the Behavioral Displaced Aggression subscale
Note. The dependent measure is the length of time that the other (bogus) predicted the outcome of interest, but the Aggression Questionnaire did
participant must immerse his or her hand in painfully cold ice water. not).
9.36), T*y ⫽ 15.00, p ⬍ .001. Moreover, trait displaced aggression limited because participants did not complete the full 31-item
marginally predicted the reaction to the trigger (␤ ⫽ .24, p ⫽ .08), DAQ. We therefore conducted a second experiment to test the
but the Aggression Questionnaire Physical Aggression subscale construct validity of our full measure of trait displaced aggression.
did not (␤ ⫽ .06, ns). As expected, scores on the Behavioral A second weakness of Experiment 1 was that we did not directly
Displaced Aggression subscale were unrelated to the reaction to manipulate rumination. In Experiment 2, we included situational
the provocation but were associated with the emotional reaction to manipulations of provocation-focused and self-focused rumina-
the bogus participant. We also conducted mediation analyses tion. We expected an interaction with the DAQ such that a par-
(Baron & Kenny, 1986; MacKinnon, Lockwood, Hoffman, West, ticularly strong relationship would emerge between DAQ scores
& Sheets, 2002) to determine the mediating role of the affective and displaced aggression in the rumination conditions. In this
reaction to the bogus participant. The Behavioral Displaced Ag- second experiment, participants completed the entire DAQ, the
gression subscale predicted the emotional reaction to the trigger PANAS General (Watson et al., 1988), and the Rumination-
(␤ ⫽ .26, p ⬍ .01), as well as displaced aggression (␤ ⫽ .36, p ⬍ Reflection Questionnaire (RRQ; Trapnell & Campbell, 1999).
.001). The reaction to the trigger predicted displaced aggression Participants were provoked and subsequently engaged in a situa-
(␤ ⫽ .51, p ⬍ .001). Finally, when entered simultaneously with the tional rumination task prior to aggressing against a bogus partic-
Behavioral Displaced Aggression subscale, both this subscale ipant. We predicted that scores on the DAQ would moderate the
(␤ ⫽ .25, p ⬍ .01) and reaction to the trigger remained significant effects of situational rumination on displaced aggression but that
predictors of displaced aggression (␤ ⫽ .44, p ⬍ .001). A Sobel scores on the remaining personality measures would not predict
(1982) test also revealed that the reaction to the trigger mediated aggressive behavior.
the effects of the DAQ on laboratory displaced aggression (z ⫽
2.58, p ⬍ .01), as did an additional test of indirect effect (z ⫽ 2.61, Experiment 2: Method
p ⬍ .01). These results suggest that participants high in trait
displaced aggression perceived the other participant more nega- Participants and Design
tively than those low in trait displaced aggression (likely owing to
Eighty-seven CSULB undergraduates (89% female, 11% male) enrolled
rumination about the provocation), which in turn increased the
in introductory psychology courses completed the study in exchange for
likelihood of aggressive behavior toward the “innocent” other extra course credit. The design was a 3 (rumination type: provocation-
participant. In other words, the reaction to the trigger partially focused, self-focused, or distraction) ⫻ 2 (trigger: yes or no) between-
mediated the effects of trait displaced aggression on actual dis- subjects factorial. All participants were provoked by the experimenter in
placed aggression. A three-step hierarchical regression for provo- the same manner as in Experiment 1 (i.e., through negative feedback on the
cation, trigger, and the Behavioral Displaced Aggression subscale anagram task). Participants were randomly assigned to one of the six
did not reveal any significant interactive effects. experimental conditions.
Experiment 1: Discussion Materials and Procedures

In general the procedures for Experiment 2 were similar to those of
Experiment 1 provided behavioral evidence for the construct
Experiment 1. Participants took part in a modified TDA paradigm (Miller
validity of our Behavioral Displaced Aggression subscale. In a
et al., 2003) with the following three differences: (a) Participants filled out
laboratory TDA paradigm, our measure moderated the degree of all three subscales of the DAQ as well as additional individual-differences
actual displaced aggression and predicted the affective reaction to measures to obtain discriminant behavioral construct validity evidence, (b)
a bogus participant. Individuals who reported the general tendency all participants were provoked by the experimenter, and (c) participants
to aggress against undeserving others when in a negative affective completed one of two situational rumination manipulations or a distracting
state did indeed aggress against an undeserving participant to a control task. Upon arrival to the laboratory, participants were told that the
greater extent than those who did not endorse such statements. At study was investigating the relationship between an individual’s personal-
the same time, the Physical Aggression subscale of the Aggression ity and his or her ability to complete academic tasks. As in Experiment 1,
Questionnaire (Buss & Perry, 1992), a reliable and valid measure participants were told that they would interact with another (bogus) par-
of direct aggression, did not moderate displaced aggression or ticipant in an adjoining room. Participants completed a packet of person-
ality measures, were provoked, engaged in a rumination or control task,
predict reactions to the trigger. Thus, even in the short 15–20-
were triggered by the bogus participant (or not), and were given the
minute interval following the provocation and opportunity to ag- opportunity to aggress against the bogus participant.
gress, it appears that individuals high in trait displaced aggression Personality measures. After giving informed consent, participants
were likely to ruminate about the initial provocation, which in turn completed the personality measures. Specifically, participants completed
led to a more negative reaction to the bogus participant and all 31 items on the DAQ, the PANAS General (Watson et al., 1988), and
subsequently increased displaced aggression. Because we did not the RRQ (Trapnell & Campbell, 1999). We hypothesized that only the
directly manipulate rumination, we were not able to assess the DAQ would moderate the effects of situational rumination on displaced
unique situational effects of rumination directly. In the current aggression. To control for order effects, we administered six counterbal-
study, participants were free to attempt to control their ruminative anced versions of the measures.
thoughts (although this did not appear to work). Therefore, we Provocation manipulation. After participants completed the personal-
ity measures, the experimenter provoked them in the same manner as in
conducted a second experiment to directly examine the additive
Experiment 1 (i.e., through negative feedback on the anagram task).
effects of person (trait displaced aggression) and situation vari- Rumination manipulation. After being provoked by the experimenter,
ables (a rumination writing task) as conceptualized within the participants were told that they would complete a 20-min writing task that
GAM on laboratory displaced aggression. purportedly assessed their ability to write effectively. Participants were told
Although Experiment 1 provided solid evidence of construct that there were several different writing topics and that one had been
validity for the Behavioral Displaced Aggression subscale, it was randomly chosen for them. Those individuals in the provocation-focused
rumination condition were asked to write about what had occurred in the Table 7
experiment up to that point, including their actions, feelings, and interac- Aggression 20% Trimmed Means, Winsorized Standard
tions with other individuals. Similar to procedures used by Rusting and Deviations, and Sample Sizes for Experiment 2
Nolen-Hoeksema (1998), participants in the self-focused rumination con-
dition were given a packet with a phrase on each page. Each phrase in this Rumination condition
packet was internally focused and contained no mention of affect (e.g.,
“what kind of person you are,” “how you interact with people”). In fact, Provocation
judges had rated these phrases as affectively neutral (Rusting & Nolen- focused Self-focused Distraction
Hoeksema, 1998). Using the procedure employed in Bushman et al. (2005,
Study 1), participants were told to think about each phrase, spend 1 or 2 Condition M SD n M SD n M SD n
minutes writing any thoughts that came to mind on a pad of paper, and then Trigger 31.43 15.00 9 33.33 17.40 15 26.00 20.16 16
move on to the next page of the packet, continuing this same process for No trigger 30.00 8.45 15 15.56 11.26 13 19.00 12.69 14
20 minutes. Finally, in the distraction condition, participants were in-
structed to write about the layout of their college campus. Participants in all
three conditions were instructed not to worry about either spelling or
grammar. focused condition (␤ ⫽ .18, p ⫽ .30). No three-way interactions
Trigger manipulation and displaced aggression. These procedures emerged between trigger, rumination condition, and the personal-
were identical to those used in Experiment 1. ity variables.
Because the additional personality measures and 20-min rumination Finally, we conducted a series of parallel analyses to those
manipulation greatly lengthened the experiment, secondary dependent reported above, in which we examined the effects of each of the
measures were not included in Experiment 2. three DAQ subscales. Because the Angry Rumination and Re-
venge Planning subscales explicitly describe ruminative tenden-
Experiment 2: Results cies, we hypothesized that these subscales would interact with the
experimental rumination manipulations but not the Behavioral
Data from 5 participants were removed due to suspicion of the Displaced Aggression subscale. Overall, each subscale predicted
study hypotheses. As in Experiment 1, we used hierarchical re- actual displaced aggression: Angry Rumination (␤ ⫽ .26, p ⬍ .02),
gression analyses to test our hypotheses. At the first step, we Revenge Planning (␤ ⫽ .34, p ⬍ .005), and Behavioral Displaced
regressed laboratory displaced aggression on the rumination con- Aggression (␤ ⫽ .29, p ⬍ .01). In the regression analyses, we
ditions (referenced to the distraction control task) and the trigger entered each subscale, the condition variables, and their interaction
conditions, as well as the mean-centered PANAS Positive Affect terms in predicting actual displaced aggression. The two rumina-
(␣ ⫽ .85) and Negative Affect (␣ ⫽ .86) subscales, the Rumina- tive subscales reliably interacted with the rumination conditions:
tion (␣ ⫽ .86) and Reflection (␣ ⫽ .83) subscales from the RRQ, Angry Rumination (␤ ⫽ .36, p ⬍ .01) and Revenge Planning (␤ ⫽
and the DAQ (␣ ⫽ .93). As expected, participants in the .30, p ⬍ .02). However, the Behavioral Displaced Aggression
provocation-focused rumination (T*Y ⫽ 2.45, p ⬍ .05, MT ⫽ subscale did not (␤ ⫽ .18, p ⫽ .20), thus providing evidence of
32.50 s) and self-focused rumination conditions (T*Y ⫽ 2.46, p ⬍ convergent and discriminant predictive validity of the three sub-
.01, MT ⫽ 28.42 s) displayed significantly higher levels of dis- scales. Post hoc probing (Aiken & West, 1991) of these interac-
placed aggression than participants in the control condition (MT ⫽ tions revealed that Angry Rumination predicted actual displaced
17.06 s), irrespective of whether they were triggered. There was no aggression in the provocation-focused (␤ ⫽ .59, p ⬍ .005) and
main effect for trigger (␤ ⫽ ⫺.00, p ⫽ .99). Most important, self-focused conditions (␤ ⫽ .40, p ⬍ .02) but not in the control
among the individual-differences measures, the DAQ was the only condition (␤ ⫽ ⫺.09, p ⫽ .61). Similarly, Revenge Planning
significant predictor of displaced aggression (␤ ⫽ .52, p ⬍ .001). predicted actual displaced aggression in the provocation-focused
At the second step, we entered the interaction terms of the (␤ ⫽ .77, p ⬍ .005) and self-focused conditions (␤ ⫽ .49, p ⬍
personality measures with the condition variables as well as the .005) but not in the control condition (␤ ⫽ .10, p ⫽ .54). In
Trigger ⫻ Rumination Condition interaction term. Although the general, the effects of these two ruminative subscales were larger
regression coefficient for the Trigger ⫻ Rumination interaction in the provocation-focused conditions, consistent with the notion
term was not significant (␤ ⫽ .38, p ⫽ .12), a more powerful that individual differences on these two subscales are especially
bootstrap linear contrast (4, 3, 1, ⫺2.33, ⫺2.33, ⫺2.33) on 20% relevant to provocation-related rumination.
trimmed means provided support for the expected pattern of results
(␺ ⫽ 2.53, p ⫽ .03; see Table 7). We based this contrast on
Experiment 2: Discussion
previous studies from our laboratories (Bushman et al., 2005;
Pedersen et al., 2005). Two of the personality measures moderated Experiment 2 provided further behavioral evidence of construct
the effects of rumination condition on displaced aggression: a validity for the DAQ. Using the full 31-item scale, we replicated
DAQ ⫻ Rumination interaction (␤ ⫽ .44, p ⫽ .04) and a signif- the predictive effect of the DAQ on displaced aggression in the
icant PANAS Negative Affect ⫻ Rumination interaction (␤ ⫽ .37, laboratory. Discriminant validity evidence was also obtained such
p ⬍ .03). Post hoc tests (Aiken & West, 1991) revealed that DAQ that the DAQ predicted displaced aggression in the provocation-
scores predicted displaced aggression in the provocation-focused focused and self-focused rumination conditions, whereas a mea-
(␤ ⫽ .78, p ⬍ .001) and self-focused (␤ ⫽ .51, p ⬍ .001) sure of trait negative affect predicted aggression in the self-focused
rumination conditions but not in the distraction condition (␤ ⫽ .09, rumination condition. This latter finding is consistent with research
p ⫽ .61). The PANAS Negative Affect subscale predicted dis- demonstrating that situational manipulations of self-focused rumi-
placed aggression only in the self-focused rumination condition nation increase anger (Rusting & Nolen-Hoeksema, 1998) and
(␤ ⫽ .54, p ⬍ .02), marginally negatively in the distraction displaced aggression (Bushman et al., 2005). Therefore it is not
condition (␤ ⫽ ⫺.32, p ⫽ .06), and not at all in the provocation- surprising that individuals prone to experiencing negative affect
were strongly affected by the situational self-focused rumination Limitations of the DAQ
manipulation. This applies to both those high in trait displaced
aggression and those high in general negative affect (i.e., as There are several issues inherent in the current research that may
measured by the PANAS Negative Affect subscale). However, limit the applicability of the DAQ. First, across the five phases of
discriminant validity was demonstrated to the extent that the DAQ the current research, all of the samples contained primarily female
predicted displaced aggression in the provocation-focused rumi- respondents. Although no gender effects were found (except for
nation condition but the other measures did not. The relationship the Revenge Planning subscale), gender differences with the DAQ
between the DAQ and displaced aggression was particularly strong may yet be discovered in future samples with larger numbers of
in this condition, suggesting an additive effect of person and men. A second major concern is the self-report nature of the DAQ.
situation variables. This demonstrates that high scorers on the Particularly troubling is the moderate relationship between the
DAQ are especially likely to be affected by manipulations that DAQ and trait social desirability. Although this is a problem with
increase anger and negative affect following a provocation. These self-report measures of aggressive personality in general (Harris,
results appear consistent with our conceptualization of trait dis- 1997; Lange, Dehghani, & de Beurs, 1995, Lange, Palich, et al.,
placed aggression as being relevant to anger-inducing provoca- 1995; Morren & Meesters, 2002), future research could focus on
tions, not simply general negative mood. In addition, as in Exper- implicit assessment of aggression-related personality dimensions.
iment 1, we observed that the DAQ predicted displaced aggression For example, Uhlmann and Swanson (2004) used the Implicit
regardless of the trigger. Once provoked, these individuals simply Association Test (Greenwald, McGhee, & Schwartz, 1998) to
take out their aggression on undeserving others even if these others demonstrate increases in aggressive self-concept following expo-
have objectively done absolutely nothing to annoy them. sure to violent video games. It is conceivable that individuals high
Further support for this is evident in the finding that a general in trait displaced aggression could more rapidly pair stimuli related
measure of negative self-focus (i.e., the Rumination subscale of the to displaced aggression (e.g., scenes of domestic abuse) with
RRQ) failed to predict displaced aggression. This is not surprising positive adjectives than those low in trait displaced aggression.
given that the Rumination subscale of the RRQ consists of a more A third limitation is related to the methodological details of our
general set of items that are related to the self and do not specify experimental validity studies. In both studies, the status of the
specific types of situations (e.g., “I often reflect on episodes in my experimenter was slightly higher than that of the participant,
life that I should no longer concern myself with” and “Often I’m whereas the bogus participant and the actual participant were of
playing back over in my mind how I acted in a past situation”). The equal status. Although both participant and experimenter were
DAQ, on the other hand, refers specifically to rumination about undergraduates at the same university, in the experimental situa-
anger-inducing provocations and the general tendency to aggress tion, the experimenter was probably considered to be an expert on
against innocent others when provoked. Thus, the DAQ provided the experimental procedures and therefore of higher status in this
more information about displaced aggression under conditions of particular setting. This design feature of our paradigm may have
provocation-focused rumination than other theoretically relevant interacted with other unmeasured personality dimensions in which
measures. sensitivity to status is an essential feature. For example, individ-
uals with passive–aggressive personality disorder display status
sensitivity manifested as dislike and criticism of authority figures.
General Discussion This may have caused increased negative affect in these individ-
We have described the development of a self-report measure of uals when provoked by a high-status experimenter. However,
trait displaced aggression. Prior research on aggressive personality passive–aggressive individuals are also characterized by high lev-
has focused on direct responses to provocations. Thus, we pre- els of impulsivity, which displayed only small to moderate corre-
sented the DAQ as the first attempt at measurement of the dis- lations with the DAQ. Another personality dimension that may
placed aggression personality construct. We first developed a set have interacted with our status confound is right-wing authoritari-
of items based on prior rumination and aggression-related scales as anism (Altemeyer, 1996). Right-wing authoritarians display def-
well as original items designed to assess the general behavioral erence to high-status individuals and thus may have been overly
tendency to aggress against innocent others. Next, we identified a willing to “do their job” and aggress against the other participant.
subset of these items, identified the hypothesized three factors To our knowledge no data exist regarding passive–aggressive
(Angry Rumination, Revenge Planning, and Behavioral Displaced personality disorder or right-wing authoritarianism and displaced
Aggression), and demonstrated that the DAQ has high levels of aggression. Thus, it remains an empirical question as to whether
internal consistency and test–retest reliability. Moreover, the factor these individuals actually display more displaced aggression when
structure was confirmed in both a college student sample and a provoked by a high-status individual and allowed to subsequently
large (N ⬎ 1,000) national community sample of Internet respon- aggress against an equal status other. Future research should ad-
dents. We also demonstrated concurrent construct validity by dress the effects of additional personality dimensions on displaced
correlating the DAQ with theoretically relevant measures such as aggression as well as reverse the status of the experimenter and
neuroticism, conscientiousness, agreeableness, general trait ag- bogus participant.
gressiveness, anger expression, social desirability, a general mea-
sure of self-focused rumination, road rage, and domestic abuse. Relationship to Other Theoretical Models
Finally, we demonstrated that the DAQ is a good predictor of
displaced aggression in two laboratory studies. Together, these Our data are consistent with the GAM (C. A. Anderson &
findings provide support for the DAQ as a reliable and valid Bushman, 2002). The GAM is the most recent comprehensive
instrument for assessing individual differences in displaced social psychological model of aggressive behavior. Within the
aggression. GAM, when one is confronted with a social interaction, aggression
may be instigated by personality or situational factors, which in purpose dimension reflect searching for new perspectives on the
turn affect aggression-related cognition, affect, and arousal. These self (i.e., searching); scores at the other end indicate attempts to
latter internal states then bias appraisal and decision-making pro- gain closure on current emotional experiences (i.e., problem solv-
cesses, which in turn affect behavior (aggressive or otherwise). In ing). To the extent that angry rumination and revenge planning are
this view, trait displaced aggression represents a personality input attempts at restoring emotional balance and rectifying an injustice,
variable. Consistent with cognitive–affective perspectives on per- these types of rumination appear to lie toward the problem-solving
sonality (Mischel & Shoda, 1995), through repeated use of aggres- end of the dimension.
sive schemas and knowledge structures, individuals who score
high on the DAQ are likely to have acquired strong associations
between provocations and aggressive cognition, affect, and Future Research
arousal. The content of these internal processes is likely to be
angry rumination and a focus on retaliation. Moreover, those who Research on trait displaced aggression appears promising. One
score high on the DAQ are likely to maintain a negative affective area of potential future research concerns health implications as-
state for a long period of time. Evidence of this was obtained in our sociated with trait displaced aggression. A large body of evidence
two experiments. In Experiment 1, 15–20 minutes following a has revealed a moderate association between aggression-related
provocation, participants high in trait displaced aggression re- traits and negative health consequences. For instance, meta-
sponded more negatively to the bogus participant and aggressed analytic reviews have linked trait aggression, anger, and hostility
more than those low in trait displaced aggression. In Experiment 2, to increased occurrence of coronary heart disease (Booth-Kewley
the DAQ predicted displaced aggression when participants were & Friedman, 1987) and elevated blood pressure (Suls, Wan, &
induced to ruminate about a provocation or to focus on themselves Costa, 1995). In addition, a small but growing body of research
for 20 minutes. This negative, ruminative state is likely to affect suggests that rumination may also negatively impact physical
subsequent appraisal and decision-making processes, thus making health (Hogan & Linden, 2004; Thomsen, Mehlsen, Hokland, et
aggression toward the bogus participant more likely. This is con- al., 2004; Thomsen, Mehlsen, Olesen, et al., 2004). These findings
sistent with research demonstrating the effect of rumination on suggest that individuals high in trait displaced aggression may be
displaced aggression (Bushman et al., 2005). especially prone to experiencing physical health problems.
Another interesting feature of the DAQ is its relation to behav- Much remains to be done to identify moderating variables
ioral inhibition (e.g., Gray, 1987). Previous research has linked related to trait displaced aggression. Additional laboratory studies
anger and direct aggression to the behavioral approach system could allow opportunities for learning about direct aggression as
(Harmon-Jones, 2003; Harmon-Jones & Allen, 1998; Harmon- well as displaced aggression. Moreover, when individuals high in
Jones & Sigelman, 2001). The current research supports the notion trait displaced aggression are provoked publicly, they may be
that displaced aggressive personality is related to behavioral inhi- especially sensitive to inhibitory cues but become much angrier
bition. One key feature of behavioral inhibition that may account over time because of rumination than when they are provoked
for this relationship is the finding that inhibited individuals are privately. This increased ruminative intensity would likely in-
punishment aversive (Carver & White, 1994). It appears that when crease levels of displaced aggression. In addition, given the asso-
confronted with provocations, individuals scoring high on the ciation of the DAQ with behavioral inhibition, it is likely that when
DAQ are likely to initially inhibit retaliatory responses while provoked, individuals scoring high on the DAQ may experience
continuing to dwell angrily and plot revenge. Such data are con- heightened levels of withdrawal-related affect such as fear or
sistent with research linking behavioral inhibition with other forms anxiety. A future research agenda could incorporate recent find-
of rumination (Leen-Felder, Zvolensky, Feldner, & Lejuez, 2004) ings from social and affective neuroscience.
and negative affect (Gable, Reis, & Elliot, 2000). Our data linking A final yet highly important consideration for future research
trait displaced aggression with road rage, domestic abuse, and concerns potential interventions aimed at reducing displaced ag-
laboratory displaced aggression are consistent with the notion that gression among at-risk individuals. Given the relationship between
individuals high in trait displaced aggression may initially inhibit the DAQ and domestic abuse and road rage observed in Phase 3,
aggressive behavior when provoked yet subsequently take it out on interventions are likely to prove beneficial to society at large.
close others (in our case, spouses, fellow drivers, or fellow stu- Future work could focus on the interventions aimed at preventing
dents). In fact, these individuals may strongly endorse the catharsis marital aggression and its associated harmful effects on children
hypothesis: that aggressing toward others will make them feel (e.g., Margolin & Gordis, 2000). Specific cognitive strategies
better and less aggressive in the future. These research avenues aimed at reducing the priming effects of angry rumination and
remain to be explored. revenge planning might also prove helpful. In the distraction
How does the DAQ fit into recent empirical work on rumina- condition of Experiment 2, scores on the DAQ were unrelated to
tion? Part of our three-factor conceptualization of displaced ag- actual displaced aggression, suggesting that other forms of distrac-
gression consists of two types of provocation-focused ruminative tion or thought switching may prove beneficial when individuals
activity: angry rumination and revenge planning. In an attempt to high in trait displaced aggression are confronted with a
characterize various forms of repetitive thought, Segerstrom, Stan- provocation.
ton, Alden, and Shortridge (2003, Study 1) used multidimensional Establishing the DAQ as a mechanism to measure individual
scaling on relevant trait measures to identify two dimensions: differences in displaced aggression provides an initial foundation
content valence and purpose. Content valence refers to whether the for further exploration of this multifaceted personality dimension.
form of repetitive thought concerns positive or negative themes. It is hoped that additional research and effective intervention
Clearly, angry rumination and revenge planning are negatively strategies will provide us with the tools necessary to reduce the
valenced forms of repetitive thought. Scores at one end of the harm associated with trait displaced aggression.
References evidence for the validity of the Tendency to Forgive Scale. Personality
and Individual Differences, 38, 627– 638.
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and Bushman, B. J. (1995). Moderating role of trait aggressiveness in the
interpreting interactions. Thousand Oaks, CA: Sage. effects of violent media on aggression. Journal of Personality and Social
Altemeyer, B. (1996). The authoritarian specter. Cambridge, MA: Harvard Psychology, 69, 950 –960.
University Press. Bushman, B. J. (2002). Does venting anger feed or extinguish the flame?
Amirkhan, J. H. (1990). A factor analytically derived measure of coping: Catharsis, rumination, distraction, anger and aggressive responding.
The Coping Strategy Indicator. Journal of Personality and Social Psy- Personality and Social Psychology Bulletin, 28, 724 –731.
chology, 59, 1066 –1074. Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcis-
Anderson, C. A. (2001). Heat and violence. Current Directions in Psycho- sism, self-esteem, and direct and displaced aggression: Does self-love or
logical Science, 10, 33–38. self-hate lead to violence? Journal of Personality and Social Psychol-
Anderson, C. A., & Bushman, B. J. (2001). Effects of violent video games
ogy, 75, 219 –229.
on aggressive behavior, aggressive cognition, aggressive affect, physi-
Bushman, B. J., Bonacci, A. M., Pedersen, W. C., Vasquez, E. A., &
ological arousal, and prosocial behavior: A meta-analytic review of the
Miller, N. (2005). Chewing on it can chew you up: Effects of rumination
scientific literature. Psychological Science, 12, 353–359.
on triggered displaced aggression. Journal of Personality and Social
Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual
Psychology, 88, 969 –983.
Review of Psychology, 53, 27–51.
Bushman, B. J., Bonacci, A. M., van Dijk, M., & Baumeister, R. F. (2003).
Anderson, K. B., Anderson, C. A., Dill, K. E., & Deuser, W. E. (1998). The
Narcissism, sexual refusal, and aggression: Testing a narcissistic reac-
interactive relations between trait hostility, pain, and aggressive
thoughts. Aggressive Behavior, 24, 161–171. tance model of sexual coercion. Journal of Personality and Social
Ang, R. P., Ng, A.-K., Wong, S.-S., Lee, B.-O., Oei, T. P., & Leng, V. Psychology, 84, 1027–1040.
(2004). Relationship between Big Five traits and aggression: A compar- Bushman, B. J., & Wells, G. L. (1998). Trait aggressiveness and hockey
ison between undergraduates from Australia and Singapore. Journal of penalties: Predicting hot tempers on the ice. Journal of Applied Psychol-
Psychology in Chinese Societies, 5, 291–305. ogy, 83, 969 –974.
Aviles, F., Earleywine, M., Pollock, V., Stratton, J., & Miller, N. (2005). Buss, A. H., & Perry, M. (1992). The Aggression Questionnaire. Journal
Alcohol’s effect on triggered displaced aggression. Psychology of Ad- of Personality and Social Psychology, 63, 452– 459.
dictive Behaviors, 19, 108 –111. Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of
Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books. Personality and Social Psychology, 42, 116 –131.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable Caprara, G. V. (1985). Indicators of impulsive aggression: Present status of
distinction in social psychological research: Conceptual, strategic, and research on irritability and emotional susceptibility scales. Personality
statistical considerations. Journal of Personality and Social Psychology, and Individual Differences, 6, 665– 674.
51, 1173–1182. Caprara, G. V. (1986). Indicators of aggression: The Dissipation–
Baumeister, R. F., Bushman, B. J., & Campbell, W. K. (2000). Self-esteem, Rumination Scale. Personality and Individual Differences, 7, 763–769.
narcissism, and aggression: Does violence result from low self-esteem or Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping
from threatened egotism? Current Directions in Psychological Science, strategies: A theoretically based approach. Journal of Personality and
9, 26 –29. Social Psychology, 56, 267–283.
Bentler, P. M. (2005). EQS 6 structural equations program manual. Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral
Encino, CA: Multivariate Software. activation, and affective responses to impending reward and punishment:
Berkowitz, L. (1989). Frustration–aggression hypothesis: Examination and The BIS/BAS Scales. Journal of Personality and Social Psychology, 67,
reformulation. Psychological Bulletin, 106, 59 –73. 319 –333.
Berkowitz, L. (1990). On the formation and regulation of anger and Collins, K., & Bell, R. (1997). Personality and aggression: The
aggression: A cognitive–neoassociationistic analysis. American Psy- Dissipation–Rumination Scale. Personality and Individual Differences,
chologist, 45, 494 –503. 22, 751–755.
Berkowitz, L. (1993). Aggression: Its causes, consequences, and control. Dill, K. E., Anderson, C. A., & Deuser, W. E. (1997). Effects of aggressive
New York: McGraw-Hill. personality on social expectations and social perceptions. Journal of
Berry, J. W., Worthington, E. L., Parrott, L., O’Connor, L. E., & Wade,
Research in Personality, 31, 272–292.
N. G. (2001). Dispositional forgivingness: Development and construct
Dollard, J., Doob, L. W., Miller, N. E., Mowrer, O. H., & Sears, R. R.
validity of the Transgression Narrative Test of Forgivingness. Person-
(1939). Frustration and aggression. Oxford, England: Yale University
ality and Social Psychology Bulletin, 27, 1277–1290.
Press.
Berry, J. W., Worthington, E. L., Parrott, L., O’Connor, L. E., & Wade,
Duval, S., & Wicklund, R. A. (1972). A theory of objective self awareness.
N. G. (2005). Forgivingness, vengeful rumination, and affective traits.
Oxford, England: Academic Press.
Journal of Personality, 73, 183–225.
Bettencourt, B. A., Brewer, M. B., Croak, M. A., & Miller, N. (1992). Eisenberger, R., Lynch, P., Aselage, J., & Rohdieck, S. (2004). Who takes
Cooperation and the reduction of intergroup bias: The role of reward the most revenge? Individual differences in negative reciprocity norm
structure and social orientation. Journal of Experimental Social Psychol- endorsement. Personality and Social Psychology Bulletin, 30, 789 –799.
ogy, 28, 301–319. Forgays, D. G., Forgays, D. K., & Spielberger, C. D. (1997). Factor
Bettencourt, B. A., & Miller, N. (1996). Gender differences in aggression structure of the State–Trait Anger Expression Inventory. Journal of
as a function of provocation: A meta-analysis. Psychological Bulletin, Personality Assessment, 69, 497–507.
119, 422– 447. Fridhandler, B., & Averill, J. R. (1982). Temporal dimensions of anger: An
Booth-Kewley, S., & Friedman, H. S. (1987). Psychological predictors of exploration of time and emotion. In J. R. Averill (Ed.), Anger and
heart disease: A quantitative review. Psychological Bulletin, 101, 343– aggression: An essay on emotion (pp. 253–280). New York: Springer-
362. Verlag.
Borjesson, W. I., Aarons, G. A., & Dunn, M. E. (2003). Development and Gable, S. L., Reis, H. T., & Elliot, A. J. (2000). Behavioral activation and
confirmatory factor analysis of the Abuse Within Intimate Relationships inhibition in everyday life. Journal of Personality and Social Psychol-
Scale. Journal of Interpersonal Violence, 18, 295–309. ogy, 78, 1135–1149.
Brown, R. P., & Phillips, A. (2005). Letting bygones be bygones: Further Goldberg, L. R. (1990). An alternative “description of personality”: The
Big-Five factor structure. Journal of Personality and Social Psychology, Macrae, R. R., & Allik, J. (2002). The five-factor model of personality
59, 1216 –1229. across cultures. New York: Kluwer/Plenum Press.
Goldberg, L. R. (1992). The development of markers for the Big-Five Marcus-Newhall, A., Pedersen, W. C., Carlson, M., & Miller, N. (2000).
factor structure. Psychological Assessment, 4, 26 – 42. Displaced aggression: A meta-analytic review. Journal of Personality
Gouldner, A. W. (1960). The norm of reciprocity. American Sociological and Social Psychology, 78, 670 – 689.
Review, 25, 161–178. Mardia, K. V. (1970). Measures of multivariate skewness and kurtosis with
Gray, J. A. (1987). The neuropsychology of emotion and personality. In applications. Biometrika, 57, 519 –530.
S. M. Stahl & S. D. Iverson (Eds.), Cognitive neurochemistry (pp. Margolin, G., & Gordis, E. B. (2000). The effects of family and community
171–190). London: Oxford University Press. violence on children. Annual Review of Psychology, 51, 445– 479.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring Marlowe, D., & Crowne, D. P. (1961). Social desirability and response to
individual differences in implicit cognition: The Implicit Association perceived situational demands. Journal of Consulting Psychology, 25,
Test. Journal of Personality and Social Psychology, 74, 1464 –1480. 109 –115.
Harmon-Jones, E. (2003). Anger and the behavioral approach system. Martin, L. L., & Tesser, A. (1996). Some ruminative thoughts. Advances in
Personality and Individual Differences, 35, 995–1005. Social Cognition, 9, 1– 47.
Harmon-Jones, E., & Allen, J. B. (1998). Anger and frontal brain activity: Mauger, P. A., Perry, J. E., Freeman, T., Grove, D. C., McBride, A. G., &
EEG asymmetry consistent with approach motivation despite negative McKinney, K. E. (1992). The measurement of forgiveness: Preliminary
affective valence. Journal of Personality and Social Psychology, 74, research. Journal of Psychology & Christianity, 11, 170 –180.
1310 –1316. Maxwell, J. P. (2004). Anger rumination: An antecedent of athlete aggres-
Harmon-Jones, E., & Sigelman, J. (2001). State anger and prefrontal brain sion? Psychology of Sport and Exercise, 5, 279 –289.
activity: Evidence that insult-related relative left-prefrontal activation is McCullough, M. E., Bellah, C. G., Kilpatrick, S. D., & Johnson, J. L.
associated with experienced anger and aggression. Journal of Person- (2001). Vengefulness: Relationships with forgiveness, rumination, well-
ality and Social Psychology, 80, 797– 803. being, and the Big Five. Personality and Social Psychology Bulletin, 27,
Harris, J. A. (1997). A further evaluation of the Aggression Questionnaire: 601– 610.
Issues of validity and reliability. Behaviour Research & Therapy, 35, McCullough, M. E., Rachal, K. C., Sandage, S. J., Worthington, E. L.,
1047–1053. Brown, S. W., & Hight, T. L. (1998). Interpersonal forgiving in close
Hewig, J., Hagemann, D., Seifert, J., Naumann, E., & Bartussek, D. (2004). relationships: II. Theoretical elaboration and measurement. Journal of
On the selective relation of frontal cortical asymmetry and anger-out Personality and Social Psychology, 75, 1586 –1603.
versus anger-control. Journal of Personality and Social Psychology, 87, Miller, N., Pedersen, W. C., Earleywine, M., & Pollock, V. E. (2003). A
926 –939. theoretical model of triggered displaced aggression. Personality and
Hogan, B. E., & Linden, W. (2004). Anger response styles and blood Social Psychology Review, 7, 75–97.
pressure: At least don’t ruminate about it! Annals of Behavioral Medi- Mischel, W., & Shoda, Y. (1995). A cognitive–affective system theory of
cine, 27, 38 – 49. personality: Reconceptualizing situations, dispositions, dynamics, and
Hovland, C., & Sears, R. (1940). Minor studies in aggression: VI. Corre- invariance in personality structure. Psychological Review, 102, 246 –
lation of lynchings with economic indices. Journal of Psychology, 9, 268.
301–310. Morren, M., & Meesters, C. (2002). Validation of the Dutch version of the
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance Aggression Questionnaire in adolescent male offenders. Aggressive Be-
structure analysis: Conventional criteria versus new alternatives. Struc- havior, 28, 87–96.
tural Equation Modeling, 6, 1–55. Morrison, R., & O’Connor, R. C. (2005). Predicting psychological distress
International Personality Item Pool. (2001). A scientific collaboratory for in college students: The role of rumination and stress. Journal of Clinical
the development of advanced measures of personality traits and other Psychology, 61, 447– 460.
individual differences. Retrieved May 2004, from http://ipip.ori.org/ Muris, P., Roelofs, J., Meesters, C., & Boomsma, P. (2004). Rumination
Kachadourian, L. K., Fincham, F., & Davila (2005). Attitudinal ambiva- and worry in nonclinical adolescents. Cognitive Therapy and Research,
lence, rumination, and forgiveness of partner transgressions in marriage. 28, 539 –554.
Personality and Social Psychology Bulletin, 31, 334 –342. Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depres-
Konečni, V. J. (1974). Self-arousal, dissipation of anger, and aggression. sion and posttraumatic stress symptoms after a natural disaster: The
Personality and Social Psychology Bulletin, 1, 192–194. 1989 Loma Prieta earthquake. Journal of Personality and Social Psy-
Lange, A., Dehghani, B., & de Beurs, E. (1995). Validation of the Dutch chology, 61, 115–121.
adaptation of the Buss–Durkee Hostility Inventory. Behavior Research Nolen-Hoeksema, S., & Morrow, J. (1993). Effects of rumination and
and Therapy, 33, 229 –233. distraction on naturally occurring depressed mood. Cognition and Emo-
Lange, A., Pahlich, A., Sarucco, M., Smits, G., Dehghani, B., & Hanewald, tion, 7, 561–570.
G. (1995). Psychometric characteristics and validity of the Dutch adap- Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response
tation of the Buss–Durkee Hostility Inventory. Behavior Research and styles and the duration of episodes of depressed mood. Journal of
Therapy, 33, 223–227. Abnormal Psychology, 102, 20 –28.
Leen-Felder, E. W., Zvolensky, M. J., Feldner, M. T., & Lejuez, C. W. Nowlis, V. (1965). Research with the Mood Adjective Check List. In S. S.
(2004). Behavioral inhibition: Relation to negative emotion regulation Tompkins & C. E. Izard (Eds.), Affect, cognition, and personality:
and reactivity. Personality and Individual Differences, 36, 1235–1247. Empirical studies (pp. 352–389). Oxford, England: Springer.
Lindsay, J. J., & Anderson, C. A. (2000). From antecedent conditions to Paleari, F. G., Regalia, C., & Fincham, F. (2005). Marital quality, forgive-
violent actions: A general affective aggression model. Personality and ness, empathy, and rumination: A longitudinal analysis. Personality and
Social Psychology Bulletin, 26, 533–547. Social Psychology Bulletin, 31, 368 –378.
Lyubomirsky, S., & Nolen-Hoeksema, S. (1995). Effects of self-focused Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of
rumination on negative thinking and interpersonal problem solving. the Barratt Impulsiveness Scale. Journal of Clinical Psychology, 51,
Journal of Personality and Social Psychology, 69, 176 –190. 768 –774.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Pedersen, W. C., Denson, T. F., Goss, R. J., Vasquez, E. A., & Miller, N.
Sheets, V. (2002). A comparison of methods to test mediation and other (2005). The impact of rumination on aggressive thoughts, feelings,
intervening variable effects. Psychological Methods, 7, 83–104. behavior, and arousal. Manuscript submitted for publication.
Pedersen, W. C., Gonzales, C., & Miller, N. (2000). The moderating effect young and an elderly sample. Journal of Behavioral Medicine, 27,
of trivial triggering provocation on displaced aggression. Journal of 215–231.
Personality and Social Psychology, 78, 913–927. Trapnell, P. D., & Campbell, J. D. (1999). Private self-consciousness and
Raykov, T., & Marcoulides, G. A., (2000). A first course in structural the five-factor model of personality: Distinguishing rumination from
equation modeling. Mahwah, NJ: Erlbaum. reflection. Journal of Personality and Social Psychology, 76, 284 –304.
Roger, D., & Najarian, B. (1989). The construction and validation of a new Tremblay, P. F., & Belchevski, M. (2004). Did the instigator intend to
scale for measuring emotion control. Personality and Individual Differ- provoke? A key moderator in the relation between trait aggression and
ences, 10, 845– 853. aggressive behavior. Aggressive Behavior, 30, 409 – 424.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, Tremblay, P. F., & Ewart, L. A. (2005). The Buss and Perry Aggression
NJ: Princeton University Press. Questionnaire and its relation to values, the Big Five, provoking hypo-
Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regulating responses to thetical situations, alcohol consumption patterns, and alcohol expectan-
anger: Effects of rumination and distraction on angry mood. Journal of cies. Personality and Individual Differences, 38, 337–346.
Personality and Social Psychology, 74, 790 – 803. Treynor, W., Gonzales, R., & Nolen-Hoeksema, S. (2003). Rumination
Scott, V. B., & McIntosh, W. B. (1999). The development of a trait reconsidered: A psychometric analysis. Cognitive Therapy and Re-
measure of ruminative thought. Personality and Individual Differences, search, 27, 247–259.
26, 1045–1056. Tyson, P. D. (1998). Physiological arousal, reactive aggression, and the
Segerstrom, S. C., Stanton, A. L., Alden, L. E., & Shortridge, B. E. (2003). induction of an incompatible relaxation response. Aggression and Vio-
A multidimensional structure for repetitive thought: What’s on your lent Behavior, 3, 143–158.
mind, and how, and how much? Journal of Personality and Social Uhlmann, E., & Swanson, J. (2004). Exposure to violent video games
Psychology, 85, 909 –921. increases automatic aggressiveness. Journal of Adolescence, 27, 41–52.
Segerstrom, S. C., Tsao, J. C., Alden, L. E., & Craske, M. G. (2000). Worry Ullman, J. B. (2001). Structural equation modeling. In B. G. Tabachnick &
and rumination: Repetitive thought as a concomitant and predictor of
L. S. Fidell (Eds.), Using multivariate statistics (pp. 653–771). Needham
negative mood. Cognitive Therapy and Research, 24, 671– 688.
Heights, MA: Allyn & Bacon.
Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural
Vasquez, E. A., Denson, T. F., Pedersen, W. C., Stenstrom, D. A., &
equations models. In S. Leinhart (Ed.), Sociological methodology (pp.
Miller, N. (2005). The moderating effect of trigger intensity on triggered
290 –312). San Francisco: Jossey-Bass.
displaced aggression. Journal of Experimental Social Psychology, 41,
Spielberger, C. D. (1998). State–Trait Anger Expression Inventory: Pro-
61– 67.
fessional manual. Odessa, FL: Psychological Assessment Resources.
Watkins, E. (2004). Adaptive and maladaptive ruminative self-focus during
Spielberger, C. D., Reheiser, E. C., & Sydeman, S. J. (1995). Measuring
emotional processing. Behaviour Research and Therapy, 42, 1037–
the experience, expression, and control of anger. In H. Kassinove (Ed.),
1052.
Anger disorders: Definition, diagnosis, and treatment. Series in clinical
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali-
and community psychology (pp. 49 – 67). Philadelphia: Taylor & Francis.
Strahan, R., & Gerbasi, K. C. (1972). Short, homogeneous versions of the dation of brief measures of positive and negative affect: The PANAS
Marlowe–Crowne Social Desirability Scale. Journal of Clinical Psy- scales. Journal of Personality and Social Psychology, 54, 1063–1070.
chology, 28, 191–193. Wiesenthal, D. L., Hennessy, D., & Gibson, P. M. (2000). The Driving
Sukhodolsky, D. G., Golub, A., & Cromwell, E. N. (2001). Development Vengeance Questionnaire (DVQ): The development of a scale to mea-
and validation of the Anger Rumination Scale. Personality and Individ- sure deviant drivers’ attitudes. Violence and Victims, 15, 115–136.
ual Differences, 31, 689 –700. Wilcox, R. R. (2005). Introduction to robust estimation and hypothesis
Suls, J., Wan, C. K., & Costa, P. T. (1995). Relationship of trait anger to testing (2nd ed.). London, England: Elsevier.
resting blood pressure: A meta-analysis. Health Psychology, 14, 444 – Wilcox, R. R., & Keselman, H. J. (2003). Modern robust data analysis
456. methods: Measures of central tendency. Psychological Methods, 8, 254 –
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. 274.
Needham Heights, MA: Allyn & Bacon. Yuan, K.-H., & Bentler, P. M. (1998). Normal theory based test statistics
Thomsen, D. K., Mehlsen, M. Y., Hokland, M., Viidik, A., Olesen, F., in structural equation modeling. British Journal of Mathematical and
Avlund, K., et al. (2004). Negative thoughts and health: Associations Statistical Psychology, 51, 289 –309.
among rumination, immunity, and health care utilization in a young and
elderly sample. Psychosomatic Medicine, 66, 363–371.
Thomsen, D. K., Mehlsen, M. Y., Olesen, F., Hokland, M., Viidik, A., Received June 21, 2005
Avlund, K., & Zachariae, R. (2004). Is there an association between Revision received November 17, 2005
rumination and self-reported physical health? A one-year follow-up in a Accepted December 15, 2005 䡲
Instructions to Authors
For Instructions to Authors, please visit www.apa.org/journals/psp and click on the “Instructions to
Authors” link in the Journal Info box on the right.

The Displaced Aggression Questionnaire

Uploaded by

Copyright:

Available Formats

The Displaced Aggression Questionnaire

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Displaced Aggression Questionnaire

Uploaded by

Copyright:

Available Formats

Journal of Personality and Social Psychology Copyright 2006 by the American Psychological Association

2006, Vol. 90, No. 6, 1032–1051 0022-3514/06/$12.00 DOI: 10.1037/0022-3514.90.6.1032

The Displaced Aggression Questionnaire

Thomas F. Denson William C. Pedersen

Keywords: aggression, displaced aggression, rumination, aggressive personality, domestic abuse

Phase 1: Initial Item Selection

Subscale and factor 1 2 3 4 5 6

* significant, controlling for familywise error rate (␣ ⫽ .05)

Subscale and item Source

Angry Rumination (.80, .84)

provocations. (We specifically address this consideration in Phase 5 by Table 4

Items were reverse scored when necessary. Missing values were

Angry Rumination Revenge Planning Displaced Aggression

Physical Aggression 224 .81 .26* .50* .41*

Experiment 1: Discussion Materials and Procedures

You might also like