Could Mindfulness Decrease Anger, Hostility, and Aggression by Decreasing Rumination?

Abstract

Research suggests that rumination increases anger and aggression. Mindfulness, or present-focused and intentional awareness, may counteract rumination. Using structural equation modeling, we examined the relations between mindfulness, rumination, and aggression. In a pair of studies, we found a pattern of correlations consistent with rumination partially mediating a causal link between mindfulness and hostility, anger, and verbal aggression. The pattern was not consistent with rumination mediating the association between mindfulness and physical aggression. Although it is impossible with the current nonexperimental data to test causal mediation, these correlations support the idea that mindfulness could reduce rumination, which in turn could reduce aggression. These results suggest that longitudinal work and experimental manipulations mindfulness would be worthwhile approaches for further study of rumination and aggression. We discuss possible implications of these results.

Full text

AGGRESSIVE BEHAVIOR
Volume 36, pages 28–44 (2010)
Could Mindfulness Decrease Anger, Hostility,
and Aggression by Decreasing Rumination?
Ashley Borders
1
, Mitch Earleywine
2
, and Archana Jajodia
3
1
The College of New Jersey, Ewing, New Jersey
2
University at Albany, State University of New York, Albany, New York
3
Veterans Affairs San Diego Healthcare System, University of California at San Diego, San Diego, California
:::::::::::::::::::::::::::::::::::::::::
Research suggests that rumination increases anger and aggression. Mindfulness, or present-focused and intentional awareness,
may counteract rumination. Using structural equation modeling, we examined the relations between mindfulness, rumination, and
aggression. In a pair of studies, we found a pattern of correlations consistent with rumination partially mediating a causal link
between mindfulness and hostility, anger, and verbal aggression. The pattern was not consistent with rumination mediating the
association between mindfulness and physical aggression. Although it is impossible with the current nonexperimental data to test
causal mediation, these correlations support the idea that mindfulness could reduce rumination, which in turn could reduce
aggression. These results suggest that longitudinal work and experimental manipulations mindfulness would be worthwhile
approaches for further study of rumination and aggression. We discuss possible implications of these results. Aggr. Behav.
36:28–44, 2010.
r
2009 Wiley-Liss, Inc.
:::::::::::::::::::::::::::::::::::::::::
Keywords: aggression; anger; hostility; rumination; mindfulness
INTRODUCTION
Recent research documents important associa-
tions between rumination, hostility, anger, and
aggression. Defined as repetitive, uncontrollable
thoughts about negative internal or external experi-
ences [Ingram, 1990; Martin and Tesser, 1996],
rumination involves harping on something negative,
seemingly without end or control. These repetitive
thoughts often focus on current feelings, related
causes, consequences, and potential solutions
[Lyubomirsky and Nolen-Hoeksema, 1995]. For
instance, individuals might become fixated on
thoughts about why they cannot cope with things,
why their partners treated them badly, or how they
will perform in an upcoming speech. We will first
briefly review the evidence suggesting that rumination
exacerbates and leads to anger, hostility, and aggres-
sion. We will then turn to a discussion of how
mindfulness may counteract the effects of rumination.
Rumination and Aggression-Related Variables
Nonexperimental and experimental studies alike
have found robust links between rumination, anger,
hostility, and aggression. Examining the influence of
trait rumination, for instance, Collins and Bell [1997]
randomly assigned low ruminators and high rumi-
nators to provocation (negative feedback about
performance) or control conditions. Participants
then competed with an ‘‘opponent’’ and could
choose to increase their own points, deduct points
from their opponent, or deliver a loud white noise to
the opponent. High ruminators in the provocation
condition exhibited more aggression (e.g., total
number of noise blasts) than low ruminators in the
same condition. Similarly, Caprara [1986] and
Caprara et al. [1987] found that high dispositional
ruminators exhibited more hostility, deliberated
more over thoughts of retaliation, and delivered
stronger shocks following hostile experiences than
did low ruminators. Hostile rumination also longi-
tudinally predicts self-reported aggressive and delin-
quent behavior in adolescents [Caprara et al., 2007].
Published online 22 October 2009 in Wiley InterScience (www.
interscience.wiley.com). DOI: 10.1002/ab.20327
Received 5 October 2008; Accepted 20 July 2009

Correspondence to: Ashley Borders, The College of New Jersey,
Social Sciences Building 121, 2000 Pennington Rd., Ewing 08628,
New Jersey. E-mail: borders@tcnj.edu
r
2009 Wiley-Liss, Inc.

In addition, measures of rumination correlate with
increased vengefulness, or the disposition to seek
revenge after suffering an interpersonal offense
[Barber et al., 2005; McCullough et al., 1998, 2001].
Experimental manipulations of rumination show
similar links to anger and aggression. Bushman
[2002] instructed participants to hit a punching bag
as many times as they wanted (venting) and at the
same time to think about a partner who criticized
their work (rumination). By contrast, participants in
the distraction condition focused their thoughts on
becoming physically fit while they hit the punching
bag, and participants in the control condition sat
quietly for several minutes. Results indicated that
rumination while venting increased subsequent self-
reported anger and aggressive behavior (again in the
form of delivered noise blasts). Rusting and Nolen-
Hoeksema [1998] manipulated rumination by asking
angered participants to focus their attention on
thoughts that were emotion- and self-focused but
not explicitly about anger (e.g., ‘‘why people treat
you the way they do’’). In the distraction condition,
participants focused their attention on nonemo-
tional thoughts (e.g., ‘‘the layout of the local post
office’’). Following these manipulations, partici-
pants completed stories that were later rated for
anger content. In all experiments, rumination
increased anger, whereas distraction decreased or
had no effect on pre-existing anger.
Researchers have also investigated the impact of
rumination on triggered displaced aggression, or
instances in which a person who is provoked but
cannot retaliate directly against the source of
provocation subsequently becomes aggressive with
a person who merely provides a trivial annoyance
[Pedersen et al., 2000]. In two studies [Bushman
et al., 2005], provoked participants who ruminated
were more aggressive after a trivial triggering event
than were those in a distraction condition. Rumina-
tion did not increase aggression in the absence of a
trigger. Finally, provocation-induced negative affect
correlated with aggression only for participants in
the rumination condition.
In sum, research suggests that rumination exacer-
bates angry mood, increases hostility, interacts with
perceived provocation to elicit more aggressive
behaviors, and may increase physiological arousal.
The processes by which rumination causes aggres-
sion are still being explored. Network models
propose that mood-congruent information is orga-
nized in long-term memory around central nodes
[Berkowitz, 1990; Bower, 1981]. These associative
networks link related memories, thoughts, feelings,
and behavioral tendencies. Activation of one part of
an associative network (e.g., a particular negative
emotion) can activate other items stored in that
network. Miller et al. [2003] suggested that repetitive
thoughts and elaborations about a provocation may
maintain the activation of anger-related associative
networks over time, making angry mood and hostile
thoughts more accessible and subsequent aggressive
behavior more likely. Some evidence now supports
this theory. Angered participants made to ruminate
became angrier and exhibited more aggressive
responses following a minor trigger than did
participants who distracted themselves [Bushman
et al., 2005; Rusting and Nolen-Hoeksema, 1998].
Thus, rumination about perceived insults may
activate aggressive associative networks, exacerbat-
ing existing angry mood and hostile cognitions and
resulting in aggressive behavior. Moreover, these
results reinforce the idea that anger, hostility, and
aggression are unique constructs that can influence
each other [e.g., Bryant and Smith, 2001] and should
therefore be examined separately.
The above discussion strongly implies that coun-
teracting rumination might decrease aggression,
anger, and hostility. Research suggests that the
active components of rumination are out-of-control
repetitive thoughts [Segerstrom et al., 2003] and
negative judgments about current distress [Rude
et al., 2007]. Moreover, Martin and Tesser [1996]
theorized that rumination can be either about past
events or future concerns. Several researchers have
recently theorized that mindfulness may constitute a
potential antidote to this kind of thought process.
Mindfulness and its Impact
Extensive research suggests that mindfulness is
associated with a range of positive outcomes, from
improved life satisfaction and positive mood to
decreased stress [Brown and Ryan, 2003; Shapiro
et al., 2008]. Clinically, mindfulness interventions
have effectively treated depression, anxiety, eating
disorders, substance abuse, and borderline person-
ality disorder [for reviews, see Baer, 2003; Grossman
et al., 2004]. Only recently studies have examined the
mechanisms of mindfulness and its association with
rumination and aggression. In this section, we will
first discuss the concept of mindfulness and then
differentiate it from rumination and related thought
processes. We will also review the brief literature
examining associations between mindfulness and
aggression-related variables.
Born out of traditional Eastern meditation prac-
tices, mindfulness involves a unique way of paying
attention [Baer, 2003; Shapiro et al., 2006]. First, it
29Mindfulness and Rumination
Aggr. Behav.

involves bringing one’s full attention to the present
moment, rather than dwelling on past events or
future possibilities. In fact, mindfulness may be seen
in part as a form of attention regulation, in that
individuals who are mindful are able to consciously
redirect their attention away from past negative
memories or future worries and back to present
sensations [Baer, 2003; Bishop et al., 2004].
Mindfulness is also characterized by an attitude of
nonjudgmental acceptance [Shapiro et al., 2006,
2008]. People practicing mindfulness can notice and
experience current internal and external events in an
accepting way. Teasdale et al. [1995, 2003] argue
that this acceptance permits people to stop brooding
about their unpleasant thoughts and feelings. Rather
than labeling pain as ‘‘bad’’ and something to avoid,
mindful people view pain as just one of many
sensations to experience. Hayes et al. [1999] suggests
that individuals who are mindful are more willing to
experience negative internal and external events. In
one study, participants engaged in mindful breath-
ing showed a greater willingness to view pictures
with negative emotional content [Arch and Craske,
2006]. Thompson and Waltz [2008] also found links
between mindfulness and experiential avoidance.
Mindful individuals were less likely to view negative
feelings and unpleasant events as scary or unaccep-
table. This attitude of equanimity toward all kinds
of emotions and thoughts may decrease people’s
perceived need for rumination, as current distress is
not judged so negatively.
Distinguishing Between Mindfulness,
Rumination, and Related Constructs
The intentional ‘‘here and now’’ focus of attention
contrasts with the seemingly uncontrollable mental
circles people experience when they are ruminating.
Some researchers have in fact conceptualized
rumination as maladaptive self-focus [Ingram,
1990; Mor and Winquist, 2002], a process char-
acterized by internal, sustained, and rigid attention
to perceived discrepancies between current and
desired states. This rigid and repetitive self-focus
elicits negative affect, particularly when a negative
discrepancy exists and when one fails to resolve such
a discrepancy [Carver and Scheier, 1981]. By
contrast, mindfulness is thought to be characterized
by cognitive flexibility, so that it allows people to
disengage from rigid mental cycles [Bishop et al.,
2004; Shapiro et al., 2006]. Mindful people may
notice negative affect in one moment but will
inevitably notice other sensations or events in
subsequent moments, so they do not get swept up
in analyzing negative feelings. Because there is so
much to observe in any given moment, this
intentional attention to the present moment
[Shapiro et al., 2006] necessitates flexibility and
prevents rigid self-absorption. In support of this
theory, mindfulness is associated with increased
cognitive flexibility [Davis and Nolen-Hoeksema,
2000], whereas rumination is associated with decreased
flexibility on neuropsychological tests of executive
functioning [Philippot and Brutoux, 2007]. Thus, the
intentional and flexible present-focus of mindfulness
contrasts sharply with the more rigid dwelling in the
past or future that characterizes rumination.
Research provides support for the negative
associations between mindfulness and rumination.
Brown and Ryan [2003] found that their mind-
fulness measure correlated negatively with rumina-
tion. Coffey and Hartman [2008] found a pattern of
correlations suggesting that rumination might med-
iate the association between mindfulness and psy-
chological distress (as measured by a latent factor
combining depressive and anxious symptoms).
Recent experiments have examined the effect of
mindfulness meditation by novice meditators on
rumination. Chambers et al. [2008] had 20 partici-
pants engaged in a 10-day intensive mindfulness
meditation retreat. Compared with a control group
that did not attend the retreat, these participants
reported decreased depressive symptoms and reflec-
tive rumination. Similarly, two studies [Ramel et al.,
2004; Shapiro et al., 2008] found that people who
participated in 8-week mindfulness courses showed
decreased rumination. Moreover, a change in
reported rumination accounted for decreases in
depressive and anxious symptoms [Ramel et al.,
2004]. In sum, initial evidence suggests that heigh-
tened mindfulness can decrease rumination, which in
turn results in less depression and anxiety. Thus,
decreased rumination appears to be one of the
mechanisms by which mindfulness leads to greater
well-being and less suffering.
Theory and initial research suggest that mind-
fulness is a separate construct from distraction,
which has traditionally been seen as a form of
avoidant coping [Sexton and Dugas, 2008] and is a
common alternative to rumination manipulations
[Bushman, 2002; Rusting and Nolen-Hoeksema,
1998]. Distraction is a coping skill that involves
pushing away unpleasant thoughts and feelings,
often by focusing on other thoughts or activities.
Although moderate forms of distraction (such as
focusing on pleasant events or neutral thoughts)
may be beneficial [Morrow and Nolen-Hoeksema,
1990; Nolen-Hoeksema et al., 1993], this strategy
30 Borders et al.
Aggr. Behav.

only works for a short time and under certain
conditions [Hamilton and Ingram, 2001]. Moreover,
attempting to push specific thoughts and feelings out
of awareness can paradoxically increase the fre-
quency of negative cognitions and produce greater
emotional reactions [Wegner and Zanakos, 1994;
Wenzlaff and Wegner, 2000]. Mindfulness, by
contrast, constitutes an alternative to distraction.
Present-centered awareness should in fact engender
a deeper experience of one’s thoughts and emotions
[Shapiro et al., 2006], rather than avoidance. Baer
[2003] suggests that mindfulness might function in
part by exposing individuals to their pain, resulting
in subsequent desensitization and decreased negative
reactions to pain. Other researchers suggest that
mindfulness changes one’s relationship to thoughts
rather than changing the content of thought [Hayes
et al., 1999]. Unlike distraction, which temporarily
pushes away unpleasant cognitions or replaces them
with other thoughts, mindfulness should foster the
ability to fully experience negative feelings and
thoughts without being as emotionally upset by
them. Thus, even though experimental distraction
manipulations can decrease negative affect and
aggression [Bushman, 2002; Rusting and Nolen-
Hoeksema, 1998], mindfulness may provide a better
alternative. Broderick [2005] provided some initial
support for this hypothesis. Comparing 8-minute
rumination and distraction inductions to 8 minutes
of mindfulness meditation, they found that rumina-
tion exacerbated sad mood, whereas distraction and
mindfulness meditation decreased sad mood.
Importantly, however, mindfulness meditation re-
sulted in less sad mood than did distraction. Mind-
fulness therefore appears to be a different construct
than distraction and may have greater potential as an
alternative to ruminative thought processes.
Another interesting question is whether mind-
fulness or rumination, or both, are associated with
increased self-awareness. Self-awareness generally
decreases physical and verbal aggression [Kinney
et al., 2001; Prentice-Dunn and Rogers, 1982;
Scheier et al., 1974]. These studies typically induced
self-awareness by having participants look at
themselves in the mirror, take a self-esteem measure,
or focus primarily on themselves rather than others.
Carver [1975] suggested that self-awareness makes
people more aware of their internal rules about
aggression, thus preventing the deindividuation that
might allow for aggressive actions that violate social
norms [see also Prentice-Dunn and Rogers, 1982].
Interestingly, rumination may not be associated
with self-awareness. Trapnell and Campbell [1999]
found that rumination correlated with self-reflection
(‘‘I’m trying to figure myself out’’) but not with
awareness of internal states (‘‘I’m generally attentive
to my inner feelings’’), as measured by the Private
Self-Consciousness scale [Fenigstein et al., 1975]. By
contrast, mindfulness is associated with greater
awareness of internal states but not with increased
self-reflection [Brown and Ryan, 2003]. Moreover,
mindfulness correlated with increased attention to
and clarity of feelings, as well as value-concordant
behavior [Brown and Ryan, 2003]. Thus, although
rumination involves repetitive and rigid thoughts
about negative thoughts and emotions, it is not
apparently associated with greater self-awareness.
Mindfulness, by contrast, does appear to foster
more self-awareness of internal states and personal
values. As such, it should theoretically be associated
with less aggression.
Mindfulness and Aggression
Mindfulness covaries with aggression-related vari-
ables. Using their self-report measure of mind-
fulness, Brown and Ryan [2003] found that greater
mindfulness was associated with less impulsiveness
and hostility. Heppner et al. [2008] likewise found
that greater mindfulness correlated with less anger,
hostility, and verbal aggression. In a second study,
they had some participants engaged in a mindful
eating exercise before receiving a social rejection.
Participants in the mindful condition showed less
subsequent aggression than did participants who
only received a rejection but did not engage in the
mindfulness exercise. In a pair of multiple baseline
studies, Singh et al. [2007a,b] had three adults with
mental illness and three adolescents with conduct
disorder participate in a 4-week course in mind-
fulness training. They were then asked to continue
their mindfulness practice on their own. Incidents of
aggressive and delinquent behavior were recorded
for at least 25 weeks. For all participants, aggression
decreased or was eliminated. Although these results
are promising, no data have been collected on the
mechanisms by which mindfulness decreases aggres-
sive behavior. The above discussion strongly sug-
gests that rumination may be one such mechanism.
The Current Studies
In sum, although previous research suggests that
mindfulness decreases rumination and that mind-
fulness may decrease aggression, anger, and hosti-
lity, no studies have examined the relationships
among all of these variables. We conducted two
survey studies to examine whether rumination might
be one mechanism by which mindfulness is
31Mindfulness and Rumination
Aggr. Behav.

associated with less aggression, anger, and hostility.
Although nonexperimental research cannot prove a
causal mediational model, we expected to find a
pattern of correlations consistent with rumination
mediating the association between mindfulness and
these aggression-related variables. These results
would provide an encouraging first step toward
establishing a rumination-mediated link between
mindfulness and aggression.
STUDY 1
Our first study examined the associations between
mindfulness, rumination, and four aggression-related
variables in an undergraduate sample. Specifically,
we examined the subscales of the Aggression
Questionnaire (AQ) [Buss and Perry, 1992] sepa-
rately to yield indices of physical aggression, verbal
aggression, anger, and hostility. Given previous
studies [Brown and Ryan, 2003; Heppner et al.,
2008], we expected that mindfulness would be
strongly associated with hostility, anger, and verbal
aggression. Although one previous study failed to
find significant correlations between mindfulness and
physical aggression using the BPAQ [Heppner et al.,
2008], mindfulness-based training has been shown to
reduce physical aggression [Singh et al., 2007a,b].
Thus, we had no clear hypothesis about the link
between mindfulness and physical aggression.
We also used the Rumination and Reflection Scale
[Trapnell and Campbell, 1999], which was designed
to differentiate between two types of private self-
consciousness [Fenigstein et al., 1975]. Reflection
taps both self-reflection and awareness of internal
states and is related to openness to experience and
need for cognition [Trapnell and Campbell, 1999].
Rumination taps only self-reflection and is associated
with neuroticism, anxiety, depression, and negative
affect. We used this measure because it allowed us to
compare the associations between the two subscales
and other study variables. Consistent with prior
work [Brown and Ryan, 2003], we expected mind-
fulness to correlate with rumination but not with
reflection. Therefore, comparing the rumination and
reflection subscales allowed us to further establish
the construct validity of mindfulness.
We also selected this measure because it constitu-
tes a general measure of rumination, rather than
symptom-specific rumination. There are several
measures of anger-related rumination. Caprara
[1986] created a measure of ‘‘dissipation–rumina-
tion,’’ or the storing in memory of various experi-
ences of provocation, expectations, judgments, and
thoughts of retaliation. More recently, Sukhodolsky
et al. [2001] created a measure specifically about
‘‘anger rumination,’’ which assesses memories of
past anger experiences, attention to immediate anger
experiences, and thoughts about the antecedents and
consequences of an anger experience. Unfortu-
nately, these anger-rumination measures confound
process with outcome. We would expect more
shared variance between anger-rumination measures
and aggression than between general measures of
rumination and aggression. Therefore, the strongest
evidence for the links between rumination and
aggression-related variables will come from studies
that use measures of general rumination. Given the
impressive existing literature, we expected rumina-
tion to correlate positively with all four aggression
subscales.
Method
Participants and procedure. We collected
data from 464 undergraduate students at a major
university in Los Angeles. They completed these
three measures as part of a larger questionnaire
given to students enrolled in psychology courses.
From information reported on the larger question-
naire packet, two-thirds of the undergraduate
participants were female. Just over 50% of partici-
pants were Caucasian-American, 19% were Asian-
American, 9% were Hispanic-American, 7% were
African-American, and 13% rated themselves as
‘‘other.’’ Participants’ mean age was 19.7. All
responses were anonymous and informed consent
was obtained before completing the questionnaire
packet.
Measures
Mindfulness. We measured trait mindfulness
using the Mindful Attention Awareness Scale
[MAAS; Brown and Ryan, 2003]. This 15-item
questionnaire inquires about participants’ typical
level of present awareness and attention (e.g., ‘‘I
rush through activities without being really attentive
to them’’ and ‘‘I drive places on ‘automatic pilot’
and then wonder why I went there’’). Participants
indicate how often they have these experiences on a
6-point Likert scale from ‘‘almost always’’ to
‘‘almost never.’’ Higher scores reflect more mind-
fulness. Brown and Ryan [2003] showed that the
MAAS correlates with other mindfulness scales, as
well as with emotional intelligence, openness to
experience, and internal state awareness. The
MAAS also reliably distinguished between medita-
tion practitioners and control participants [Brown
and Ryan, 2003]. One item from the MAAS (‘‘I find
32 Borders et al.
Aggr. Behav.

myself preoccupied with the future or the past’’)
seemed to overlap conceptually with the rumination
construct. This item also showed low factor loadings
during measure construction [l5.28, Brown and
Ryan, 2003]. Therefore, this item was left out of our
analyses. The other 14 items were used to create a
latent variable of mindfulness (see below). Internal
consistency of these 14 items was good, a5.86.
Rumination. Participants completed the Rumi-
nation and Reflection Questionnaire [Trapnell and
Campbell, 1999]. This 24-item measure assesses both
rumination (e.g., ‘‘I always seem to be rehashing in
my mind recent things I’ve said or done’’) and
reflection (e.g., ‘‘I love exploring my ‘inner self’’).
Participants indicate their level of agreement with
each item on a 5-point Likert scale from ‘‘strongly
disagree’’ to ‘‘strongly agree.’’ Higher subscale
scores indicate more rumination and reflection.
Trapnell and Campbell [1999] found that the
rumination subscale correlated highly with neuroti-
cism, whereas the reflection subscale correlated with
measures of openness to experience and need for
cognition. The rumination and reflection subscales
were minimally but significantly correlated in this
sample, r5.18, Po.05. These results accord with
small correlations found by the Trapnell and
Campbell [1999], suggesting that these subscales
measure independent tendencies. Moreover, the
reflection subscale did not correlate significantly
with mindfulness or any aggression subscale.
1
Thus,
only the 12 rumination items were used in testing the
measurement and mediational models (see below).
Internal consistency of this measure was also good
in this sample, a5.90.
Aggression, anger, and hostility. Finally, parti-
cipants completed the AQ [Buss and Perry, 1992].
This 29-item measure assesses physical aggression,
verbal aggression, hostility, and anger. Participants
indicate how much each item describes them on a
5-point Likert scale from ‘‘extremely uncharacter-
istic of me’’ to ‘‘extremely characteristic of me.’’
Higher scores reflect more aggression. This measure
reliably differentiates between violent and nonvio-
lent participants and correlates with responses to
provoking scenarios, peer reports of aggression, and
other self-reports of aggression [Harris, 1997;
Morren and Meesters, 2002; O’Connor et al.,
2001]. Although some research confirms the original
factor structure of this measure [e.g., Harris, 1995],
other results suggest that the original 29 items do
not load sufficiently onto four factors, especially
across different populations [Bryant and Smith,
2001; Vigil-Colet et al., 2005; Williams et al.,
1996]. Bryant and Smith [2001] proposed a reduced
version of the AQ that maintains the four original
factors but includes only the three items with the
highest loadings on each factor. They found that
across five diverse samples, this 12-item AQ fit the
data better than the original 29-item measure.
Therefore, we used only three items to create each
of the four latent variables of physical aggression
(‘‘Given enough provocation, I may hit another
person,’’ ‘‘There are people who pushed me so far
that we came to blows,’’ ‘‘I have threatened people I
know’’), verbal aggression (‘‘I often find myself
disagreeing with people,’’ ‘‘I can’t help getting into
arguments when people disagree with me,’’ ‘‘My
friends say that I’m somewhat argumentative’’),
anger (‘‘I flare up quickly but get over it quickly,’’
‘‘Sometimes I fly off the handle for no good reason,’’
‘‘I have trouble controlling my temper’’), and
hostility (‘‘At times I feel I have gotten a raw deal
out of life,’’ ‘‘Other people always seem to get the
breaks,’’ ‘‘I wonder why sometimes I feel so bitter
about things’’). Internal consistency values for these
items of physical aggression (a5.72), verbal aggres-
sion (a5.78), hostility (a5.75), and anger (a5.70)
were adequate.
Plan of Analyses
Structural equation modeling (SEM) was used to
test whether it is plausible that rumination could
mediate the association between mindfulness and
aggression. We used the MPlus program 5.0
[Muthe
´n and Muthe
´n, 1998–2007] with the missing
at random assumption [Little and Rubin, 1987] and
maximum likelihood estimation.
We first tested a measurement model to confirm
the relations of the observed variables with their
latent constructs [Anderson and Gerbing, 1988].
Based on CFAs done during construction of the
MAAS [Brown and Ryan, 2003] and the RRS
[Trapnell and Campbell, 1999] and the high
Cronbach’s as obtained in these samples, we tested
single-factor models for the mindfulness and rumi-
nation items. Based on Bryant and Smith’s [2001]
findings, we tested a four-factor (physical aggres-
sion, verbal aggression, hostility, anger) model of
the refined 12-item AQ. The measurement model
1
We initially tested a measurement model that included the reflection
subscale, as well as rumination, mindfulness, and the four aggression
variables. Allowing these seven factors to intercorrelate, the model fit
the data well, w
2
(1154) 51845.53, CFI 5.85, RMSEA 5.051. The
12 reflection items all had high loadings (ls from .58 to .78).
However, the reflection scale showed small and nonsignificant
correlations with all factors (r’so.10) other than rumination
(r5.18, Po.05).
33Mindfulness and Rumination
Aggr. Behav.

allowed these six latent variables to intercorrelate
freely.
To better confirm this measurement model, we
compared the above six-factor model to several
alternative models: (a) a five-factor model in which
the mindfulness and rumination items loaded onto a
single factor, (b) a three-factor model in which all 12
aggression items loaded onto a single factor, (c) a
two-factor model in which the mindfulness and
rumination items loaded onto one factor and all
aggression items loaded onto one factor, and (d) a
single-factor model with all measured variables
loading onto one factor. w
2
difference tests com-
pared each alternative model to the proposed six-
factor model. In these tests, if the difference in w
2
was significant for the number of degrees of freedom
difference, then the model fit significantly less well
than the proposed model.
We next used path models to test the plausibility
of mediation from trait mindfulness to the four
aggression factors via rumination. First, we fit a
model where the aggression subscales were regressed
on trait mindfulness (see Fig. 1A). We then
introduced the rumination factor as a mediator
(see Fig. 1B).
We looked for evidence suggesting mediation in a
few ways. Traditionally, a test of mediation requires
three separate regression analyses [Baron and Kenny,
1986]. First, the independent variable (IV; mind-
fulness) must predict the dependent variable (DV;
aggression). This first analysis establishes that a
relationship in fact exists to be explained. Second, the
IV must predict the proposed mediator (rumination).
Third, the DV is regressed on both the IV and the
mediator. The mediator must predict the DV, even
with the IV also entered into the equation. If these
three tests are all significant, then the coefficient for
the IV in this third equation is examined. Statistically,
mediation occurs when the IV no longer affects the
DV after inclusion of the mediator in the regression
equation. Conceptually, this result suggests that the
original effect of the IV is in fact explained by the
mediating variable. In other words, the IV has no
direct effect on the DV, after accounting for the effect
of the mediator. True tests of mediation require
experimental or longitudinal designs in order to
demonstrate that the IV in fact causes the mediator,
which in turn causes the DV. Nonexperimental
studies like this one cannot determine the actual
temporal association between variables. Rather,
statistical tests of mediation with these data suggest
that the proposed mediation model is plausible.
Researchers have designed ways to quantify the
effect of mediation and subsequently test for
significance [for a review, see MacKinnon et al.,
2002]. One common approach is to create an
estimate of the ‘‘indirect effect,’’ or the effect of
the mediator [e.g., Sobel, 1982]. The indirect effect is
defined as a product of two regression coefficients:
(1) the effect of the IV on the mediator and (2) the
effect of the mediator on the DV after controlling
for the IV. A significance test of mediation roughly
determines whether the indirect effect differs sig-
nificantly from zero. Traditional significance tests
involve dividing the indirect effect by an estimate of
its standard error and comparing the resulting value
to the standard normal distribution [see MacKinnon
et al., 2002; e.g., Sobel, 1982]. However, analyses
indicate that estimates of indirect effects are not
necessarily normally distributed [Bollen and Stine,
1990; MacKinnon et al., 2002]. Accordingly, these
traditional tests generally have low power and may
lead to incorrect conclusions.
One promising new method of testing indirect
effects without making assumptions about normal-
ity uses bootstrapping methodology [Bollen and
Stine, 1990; MacKinnon et al., 2004; Preacher and
Hayes, 2004]. A ‘‘bootstrap sample’’ consists of N
individuals sampled randomly with replacement
from the original data set, where Nis the size of
the original data set. Five hundred bootstrap
samples are created in this way, yielding an
empirical sampling distribution from which 95%
confidence intervals can be estimated [Bollen and
Stine, 1990; MacKinnon et al., 2004]. In simulations
in which the distributions of indirect effects are
skewed, these resampling methods have more power
than traditional tests that assume normality [Bollen
and Stine, 1990; MacKinnon et al., 2004; Shrout and
Bolger, 2002]. Recent simulation studies suggest that
the bias-corrected bootstrap method produces the
most accurate confidence intervals and has greater
power and more accurate Type I error rates than
other bootstrap methods [Cheung and Lau, 2008;
MacKinnon et al., 2004]. Therefore, our analyses
below report both the indirect effects of mediation
as well as bias-corrected bootstrap confidence
intervals.
Results and Discussion
Measurement model. A measurement model
was used to confirm the factor structures of trait
mindfulness, rumination, and the four [shortened;
Bryant and Smith, 2001] aggression subscales. The
model fit the data well, w
2
(650) 51388.60, CFI 5.89,
RMSEA 5.049.
34 Borders et al.
Aggr. Behav.

As expected, the model confirmed a single latent
factor structure for trait mindfulness. The factor
loadings for the individual mindfulness items ranged
from .30 to .78. Items like ‘‘I could be experiencing
some emotion and not be conscious of it until some
time later’’ and ‘‘I tend not to notice feelings of
physical tension or discomfort until they really grab
my attention’’ had the lowest factor loadings. Items
like ‘‘It seems I am running on automatic, without
much awareness of what I’m doing’’ and ‘‘I find
myself doing things without paying attention’’ had
the highest factor loadings. These results are in
accordance with earlier factor analyses of this
measure [Brown and Ryan, 2003]. We computed a
composite reliability coefficient by the method
suggested by Werts et al. [1974] using the factor
loadings of individual items and residual variances
[see also Lattin et al., 2003]. The measure showed
good composite reliability, r
c
5.86.
The model also confirmed a single latent factor
structure for rumination. The factor loadings for the
individual rumination items were high, ranging from
Fig. 1. (A) Path model with mindfulness predicting aggression variables and (B) mediation model with rumination mediating the association between
mindfulness and the aggression variables.
35Mindfulness and Rumination
Aggr. Behav.

.51 to .77. The composite reliability for this measure
in this sample was .90.
The model confirmed a four-factor model of
aggression, using only 12 items from the AQ. Factor
loadings and composite reliabilities were good for
physical aggression (ls from .59–.73, r
c
5.73),
verbal aggression (ls from .71–.77, r
c
5.78), anger
(ls from .49–.79, r
c
5.71), and hostility (ls from
.69–.76, r
c
5.76).
To better confirm this measurement model, we
compared the above six-factor model to several
alternative models. Table I shows model fits for all
of these models. A w
2
difference test revealed that
the six-factor model provided a significantly better
fit than combining rumination and mindfulness into
a single factor, supporting the discriminant validity
of rumination and mindfulness as distinct con-
structs. Combining the aggression factors into a
single factor also created a significantly worse fit,
confirming previous work on the discriminant
validity of these separate facets of aggression. The
six-factor model also fit significantly better than
treating a two-factor model of rumination/mind-
fulness and aggression. In addition, a single-factor
model fit significantly less well than the proposed
six-factor model, revealing that the distinctions
among these constructs account for more data than
a single factor of self-report. Thus, theoretically
tenable alternative models fit the data less well than
the proposed six-factor model.
Table II shows variable means, standard devia-
tions, and ranges. Participants on average endorsed
a moderate level of trait rumination and trait
mindfulness. Participants in this sample also
reported moderate levels of verbal aggression,
hostility, and anger, but they reported lower levels
of physical aggression.
Table III presents intercorrelations between all
study variables. Rumination and trait mindfulness
showed a strong negative correlation. The more
mindful people were, the less they ruminated.
Higher mindfulness was related to less aggression
on all aggression subscales. Rumination showed
strong positive correlations with the anger and
hostility subscales and weaker or nonsignificant
associations with physical (r5.08, ns) and verbal
aggression (r5.12, Po.05). Finally, the aggression
subscales showed strong positive intercorrelations,
as expected [Bryant and Smith, 2001; Buss and
Perry, 1992].
Mediation models. A path model with the
aggression subscales regressed on trait mindfulness
was fitted first. This model fit the data adequately,
w
2
(289) 5610.19, Po.001, RMSEA 5.049, CFI 5.91.
Trait mindfulness was a significant predictor of all
four aggression subscales (see Table IV).
Next, we added rumination as a mediator
variable. The model fit was adequate, w
2
(650) 5
1388.59, Po.001, RMSEA 5.049, CFI 5.89. Trait
mindfulness significantly predicted rumination,
b5.46, SE 5.04, Po.001. Table IV shows all
TABLE I. Alternative Models to the Proposed six-factor Measurement Model
Model tested w
2
(df) RMSEA CFI Dw
2
(Ddf)
Study 1
Proposed six-factor model: mindfulness factor, rumination factor, four aggression factors 1388.60 (650) .049 .89 –
Five-factor model: one rumination/mindfulness factor, four aggression factors 2553.89 (655) .079 .71 1165.29 (5)

Three-factor model: rumination factor, mindfulness factor, one aggression factor 1973.57 (662) .065 .80 584.97 (12)

Two-factor model: one rumination/mindfulness factor, one aggression factor 3131.88 (664) .089 .62 1743.28 (14)

One-factor model: single rumination/mindfulness/aggression factor 3803.67 (665) .101 .52 2415.07 (15)

Study 2
Proposed six-factor model: mindfulness factor, rumination factor, four aggression factors 1069.17 (650) .059 .87 –
Five-factor model: one rumination/mindfulness factor, four aggression factors 1573.89 (655) .087 .71 504.72 (5)

Three-factor model: rumination factor, mindfulness factor, one aggression factor 1317.37 (662) .073 .79 248.20 (12)

Two-factor model: one rumination/mindfulness factor, one aggression factor 1819.33 (664) .097 .64 750.16 (14)

One-factor model: single rumination/mindfulness/aggression factor 2014.74 (665) .105 .58 945.57 (15)


Po.05;

Po.01.
TABLE II. Means, Standard Deviations, and Ranges for All
Study Variables
Study 1 Study 2
Variable Mean SD Range Mean SD Range
Mindfulness 4.05 .76 1.2–5.8 4.09 .70 2.2–5.9
Rumination 3.55 .71 1.4–5 3.19 .83 1.2–5
Physical aggression 1.81 .92 1–5 1.39 .66 1–4.3
Verbal aggression 2.54 .94 1–5 2.42 .83 1–4.7
Hostility 2.54 .98 1–5 2.03 1.03 1–5
Anger 2.21 .85 1–5 2.06 .85 1–4.7
36 Borders et al.
Aggr. Behav.

other betas in the mediational model. Mindfulness
still significantly predicted all aggression subscales.
Rumination only significantly predicted hostility
and anger.
To look for whether the pattern of relations was
consistent with a mediating role for rumination, we
first examined whether the bs between mindfulness
and the aggression subscales decreased with the
inclusion of rumination in the model. The b
estimates for the links between mindfulness and
hostility dropped from .38 to .16 with the
inclusion of rumination, consistent with partial
mediation. Likewise, the bs between mindfulness
and anger dropped from .37 to .31. The bs from
mindfulness to physical and verbal aggression barely
changed with the inclusion of rumination (for both,
bs dropped from .02 to .01), indicating that
rumination did not account for these associations.
We also calculated indirect effect estimates and bias-
corrected bootstrap estimates for the 95% confi-
dence intervals (see Table IV). Confidence intervals
that do not include zero indicate a significant
indirect effect. Only the indirect effects for hostility
and anger were meaningfully different from zero.
These results suggest that our proposed mediation
model is plausible and that rumination may partially
mediate the associations between mindfulness and
anger and hostility.
In sum, Study 1 found support for many of our
hypotheses. Specifically, greater mindfulness was
associated with less rumination. Mindfulness was
also negatively associated with all four aggression
subscales. Finally, the pattern of correlations was
consistent with rumination partially mediating the
links between mindfulness and both hostility and
anger. Thus, it appears that more mindful people
may be less angry and hostile in part because they
ruminate less. This finding constitutes the first
evidence that rumination may be one potential
mechanism by which mindfulness is associated with
anger and hostility.
The absence of meaningful associations between
rumination and physical and verbal aggression was
surprising, given previous evidence that rumination
produces increased behavioral aggression [Bushman,
2002; Bushman et al., 2005; Caprara et al., 1987].
The absence of an observed association between
rumination and aggression could have resulted
from low levels of particularly physical aggression
reported in this undergraduate sample. The presence
of a floor effect would make any existing relation-
ships difficult to find. Replication with a different
TABLE III. Zero-Order Correlations for All Study Variables
Variable 12345 6
1. Mindfulness – .48

.24

.25

.43

.23

2. Rumination .46

– .22

.36

.63

.43

3. Physical aggression .14

.08 – .36

.40

.48

4. Verbal aggression .22

.12

.46

– .40

.59

5. Hostility .38

.56

.35

.30

– .50

6. Anger .37

.28

.51

.62

.52

–
Values below the diagonal are from Study 1. Values above the diagonal are from Study 2.

Po.05;

Po.01;

Po.001.
TABLE IV. Betas, Standard Errors, Indirect Path Estimates, and Bias-Corrected Bootstrap CIs for Mediational Models
Model tested PA b(SE) VA b(SE) Hostility b(SE) Anger b(SE)
Study 1
Mindfulness to aggression (SE) .14 (.06)

.22 (.05)

.38 (.05)

.37 (.05)

Rumination to aggression (SE) .02 (.06) .02 (.06) .48 (.05)

.14 (.06)

Mindfulness to aggression, via rumination (SE) .13 (.06)

.21 (.06)

.16 (.06)

.31 (.06)

Indirect paths (SE) .01 (.03) .01 (.03) .22 (.03)
a
.06 (.03)
a
Bias-corrected bootstrap 95% confidence intervals (.18, .16) (.16, 11) (.91, .38) (.22, .01)
Study 2
Mindfulness to aggression (SE) .21 (.09)

.24 (.09)

.41 (.08)

.23 (.08)

Rumination to aggression (SE) .05 (.10) .33 (.10)

.54 (.08)

.41 (.09)

Mindfulness to aggression, via rumination (SE) .19 (.10) .08 (.10) .16 (.08) .03 (.10)
Indirect paths (SE) .02 (.05) .15 (.05)
a
.25 (.05)
a
.19 (.05)
a
Bias-corrected bootstrap 95% confidence intervals (.23, .11) (.51, .08) (1.13, .57) (.67, .11)
a
Zero is not included in the 95% confidence interval, indicating that the indirect path is significantly different from zero.

Po.05;

Po.01;

Po.001.
37Mindfulness and Rumination
Aggr. Behav.

population would help clarify the association
between rumination and behavioral aggression.
Moreover, because we did not have access to
demographic data with this sample, we were unable
to determine whether the associations between
mindfulness, rumination, and aggression reflected
any effects of age or gender. We addressed these
limitations in our second study.
STUDY 2
Study 2 was a replication study using a nonunder-
graduate population. The vast majority of research
on links between rumination and aggression has
employed undergraduate samples. Finding similar
results in a different population will help generalize
these important associations to people of different
ages and life experiences. Moreover, we wanted to
find a sample with increased exposure to and
knowledge about mindfulness. Grossman [2008]
suggests that because of their general inexperience
with mindfulness and related concepts, undergrad-
uates’ understanding of and subsequent reporting of
‘‘mindfulness’’ likely differs from the construct that
authors of these measures intended. Therefore, we
targeted adults who were presumably more knowl-
edgeable about the construct of mindfulness, as
evidenced by their membership in mindfulness-
related listservs. The meaning of mindfulness in this
population should be more in line with the
conceptual basis of mindfulness measures.
Method
Participants and procedure. To find partici-
pants, we created an Internet survey and posted the
survey link on several mindfulness-related listservs.
People were invited to participate through an
introductory email. Because of this method of
recruitment, we cannot know what percentage of
people responded to our request. All participants
had to read and agree to an informed consent before
participation. Study 2 had 211 participants from 26
different states and 6 countries. The participants
ranged in age from 15 to 72 (mean 531.8,
SD 511.8), and 60% were female. On average,
these participants had graduated from college and
taken some graduate courses and earned an average
annual income of $40,000–60,000. Over 80% of
these participants were Caucasian-American, 3%
were Hispanic-American, and 5% each were Asian-
American, African-American, and mixed. In this
sample, 57% of participants reported meditating at
least once in their lives, and 22% currently had
regular meditation practices. Of the participants
who meditated regularly, 7% meditated a few times
a month, 45% meditated a few times a week to once
a week, and 48% meditated once a day or more. We
included the same measures of mindfulness, rumina-
tion, and aggression as in Study 1.
Results and Discussion
Measurement models. Measurement models
were again used to confirm the factor structure of
the mindfulness, rumination, and aggression vari-
ables.
2
The measurement model with trait mind-
fulness, rumination, and aggression fit adequately,
w
2
(650) 51069.17, CFI 5.87, RMSEA 5.059. The
factor loadings for the individual trait mindfulness
items again showed a wide range (from .25 to .82)
and a similar pattern as in Study 1. The factor
loadings for the rumination items ranged from .62 to
.86. Aggression items loaded onto their subscales
similarly to Study 1 (ls from .43 to .90). The
composite reliabilities for trait mindfulness
(r
c
5.87) and rumination (r
c
5.94) were good.
The composite reliabilities for physical aggression
(.74), verbal aggression (.71), anger (.75), and
hostility (.83) were also adequate.
We again compared the six-factor model to several
alternative models (see Table I). A w
2
difference test
revealed that the six-factor model provided a
significantly better fit than combining rumination
and mindfulness into a single factor, supporting the
discriminant validity of rumination and mindfulness
as distinct constructs. Combining the aggression
factors into a single factor also created a signifi-
cantly worse fit, confirming previous work on the
discriminant validity of these separate facets of
aggression. The six-factor model also fit significantly
better than treating a two-factor model of rumina-
tion/mindfulness and aggression. In addition, a
single factor model fit significantly less well than
the proposed six-factor model, revealing that the
distinctions among these constructs account for
more data than a single factor of self-report. Thus,
theoretically tenable alternative models fit the data
less well than the proposed six-factor model.
2
We again initially tested a measurement model that included the
reflection subscale, as well as rumination, mindfulness, and the four
aggression variables. Allowing these seven factors to intercorrelate,
the model fit the data well, w
2
(1154) 51865.41, CFI 5.84,
RMSEA 5.058. The twelve reflection items all had high loadings
(ls from .57 to .80). However, the reflection scale showed small and
nonsignificant correlations with all factors (rso.15), including
rumination (r5.06, ns). Therefore, the reflection subscale was
not included in subsequent models.
38 Borders et al.
Aggr. Behav.

Participants again endorsed a moderate level of
rumination and mindfulness (see Table II). This
sample showed a smaller range on the mindfulness
measure. Specifically, no one in this sample averaged
below three on the 6-point scale, whereas under-
graduates in the first sample ranged from 1.2 to 5.8.
With this sample, physical aggression, anger and,
hostility averages were significantly lower than
values in Study 1 (using one-sample t-tests,
Pso.05); only verbal aggression was endorsed at a
similar level as in Study 1. The ranges of the
aggression subscales and rumination measure were
similar across both samples. Rumination showed
strong negative correlations with trait mindfulness
(see Table III). Both rumination and trait mind-
fulness showed strong correlations with all four
aggression subscales.
Mediation models. The path model with the
aggression subscales regressed on trait mindfulness fit
the data adequately, w
2
(288) 5468.88, Po.001,
RMSEA 5.060, CFI 5.89. Specifically, trait mind-
fulness was a significant predictor of all four
aggression subscales. When we added rumination as
a potential mediator variable, the model fit was again
adequate, w
2
(649) 51047.85, Po.001, RMSEA 5
.058, CFI 5.88. Finally, we added age and gender
to the models as covariates. The initial model with
mindfulness and aggression fit adequately, w
2
(330) 5
538.22, Po.001, RMSEA 5.055, CFI 5.87. The
mediation model also fit the data adequately,
w
2
(713) 51152.37, Po.001, RMSEA 5.054, CFI 5.87.
Trait mindfulness significantly predicted rumina-
tion, b5.46, SE 5.06, Po.001. Rumination
significantly predicted verbal aggression, hostility,
and anger, but not physical aggression (see
Table IV). Trait mindfulness no longer significantly
predicted any aggression subscale. Gender did
not show any significant associations with mind-
fulness, rumination, or the aggression subscales
(all bso.13). Older participants reported less rumina-
tion (b5.22, Po.01) and physical aggression
(b5.25, Po.01).
To look for evidence consistent with a mediating
role for rumination, we first examined whether the
bs between mindfulness and the aggression subscales
decreased with the inclusion of rumination in the
model. The inclusion of rumination in the model
caused bestimates to decrease to nonsignificance
for the associations between trait mindfulness
and hostility (bdropped from .41 to .16), anger
(adropped from .23 to .03), and verbal
aggression (bdropped from .24 to .08). As
before, we also calculated indirect effect estimates
and bias-corrected bootstrap estimates for the 95%
confidence intervals (see Table IV). These data are
consistent with the idea that rumination mediates
the links between mindfulness and verbal aggres-
sion, hostility, and anger.
In sum, this study served as a replication of
Study 1 with a nonundergraduate sample. Even after
including age and gender as covariates in the model,
the pattern of correlations was consistent with
rumination again partially mediating the associa-
tions between mindfulness and both anger and
hostility. This time, analyses also suggested that
rumination may partially mediate the link between
mindfulness and verbal aggression. Once again, the
pattern was not consistent with rumination mediat-
ing the relationship between mindfulness and
physical aggression. Implications of these results
will be discussed below.
Why rumination correlated with physical and
verbal aggression in this sample and not in the
undergraduate sample is unclear. This particular
sample showed a slightly larger variance on the
rumination measure, which might explain the larger
correlations. Alternatively, perhaps other factors
than rumination best predict tendencies for aggres-
sive behavior in younger, undergraduate popula-
tions. For example, alcohol consumption and
emotional vulnerability are associated with in-
creased aggressive behavior [Caprara et al., 1987;
Chermack and Giancola, 1997]. These constructs
may influence college-aged students more than
adults who are older and working. Although data
on aggression in children are plentiful, age-related
differences in predictors of aggression have not been
identified as frequently in adults [e.g., Hennessy and
Wiesenthal, 2004]. Future research should investi-
gate the predictors of behavioral aggression in
adults of different ages.
GENERAL DISCUSSION
This pair of studies explored whether rumination
might potentially mediate the association between
mindfulness and several aggression-related vari-
ables. SEM analyses allowed us to confirm the
factor structures of rumination and mindfulness and
to ensure that they are in fact distinct constructs.
Moreover, we examined physical aggression, verbal
aggression, anger, and hostility separately, rather
than as one composite measure of aggression. These
studies therefore yield important new information
about the potential of mindfulness to decrease both
rumination and the behavioral, emotional, and
cognitive components of aggression.
39Mindfulness and Rumination
Aggr. Behav.

With two different samples, we found that people
who are mindful are less angry and hostile. More-
over, our results are consistent with the proposed
model that the relationships between mindfulness
and anger and hostility are partly mediated by lower
levels of ruminative thinking. Only two previous
studies have examined the links between mind-
fulness, hostility, and anger [Brown and Ryan, 2003;
Heppner et al., 2008]. The current studies therefore
provide important initial evidence that mindfulness
may contribute to decreased anger and hostility.
Moreover, this is the first study we know of to
suggest that mindfulness might reduce anger and
hostility through its effect on rumination. Only one
other study has found that decreased rumination
following mindfulness training accounted for de-
creased depressive and anxious symptoms [Ramel
et al., 2004]. Therefore, our results provide necessary
additional evidence that rumination may be one
mechanism of mindfulness.
Mindfulness was also associated with less physical
and verbal aggression in both of our studies. Two
previous case studies found that teaching mind-
fulness decreased aggressive behavior [Singh et al.,
2007a,b]. Our studies therefore provide needed
statistical support for the relationship between
mindfulness and behavioral aggression. Experimen-
tal work should be a focus of future studies. The
potential role of rumination in these relationships,
however, is less clear. Results were consistent with
rumination mediating the association between mind-
fulness and verbal aggression only for our non-
undergraduate sample. Moreover, contrary to
predictions, the pattern of correlations did not
suggest that rumination might mediate the relation-
ship between mindfulness and physical aggression.
In the undergraduate sample, the lack of support for
mediation might be due to weak and nonsignificant
correlations between rumination and overt aggres-
sion. However, rumination was associated with both
physical and verbal aggression in the nonunder-
graduate sample. Despite this, the results did not
support the proposed model that rumination would
mediate the link between mindfulness and physical
aggression. Therefore, it seems that rumination may
not constitute a major mechanism through which
mindfulness affects verbal and especially physical
aggression.
Several other proposed mechanisms for mindfulness
may influence behavioral aggression. Relaxation is
one common outcome of mindfulness practice [for a
review, see Baer, 2003]. Studies have also found that
relaxation can decrease aggression, presumably be-
cause relaxation decreases physiological arousal
[Lopata, 2003; Tyson, 1998]. Thus, mindfulness
practice may decrease aggression partially through
making individuals more relaxed.
Another possible mechanism between mindfulness
and behavioral aggression is emotion regulation, or
the ability to influence the experience and expression
of emotions [Gross, 1998]. Common emotion
regulation strategies include altering thoughts or
behavior in order to better cope with negative
emotions. Two studies have found that mindfulness
correlates with the ability to regulate negative affect
[Brown and Ryan, 2003; Coffey and Hartman,
2008]. Moreover, Coffey and Hartman [2008] found
that emotion regulation mediated the association
between mindfulness and depressive and anxious
symptoms. Likewise, research has found links
between aggressive behavior and poor emotion
regulation [e.g., Izard et al., 2008; McNulty and
Hellmuth, 2008]. Thus, mindfulness may decrease
aggressive behavior in part because it promotes the
ability to cope with and even change negative
internal emotions.
Mindfulness may also promote better cognitive
functioning and flexibility. As discussed above,
mindfulness is associated with increased cognitive
flexibility [Davis and Nolen-Hoeksema, 2000].
A recent study also showed that participating in
mindfulness training led to increases in working
memory and sustained attention [Chambers et al.,
2008]. Some researchers in fact suggest that mind-
fulness is a form of ‘‘mental training’’ that involves
self-regulation of attention and cognitive inhibition
[Bishop et al., 2004]. On the other hand, aggression is
associated with decreased cognitive performance,
especially on tests of planning, attentional control,
and goal-directed behavior [Giancola, 2000]. Thus,
mindfulness may actually promote better executive
functioning, which may prevent some of the cogni-
tive errors that contribute to aggressive behavior.
Relatedly, mindfulness may decrease aggression
because it decreases behaving in automatic or
impulsive ways. Mindfulness should help make
people aware of feelings and thoughts when they
arise, rather than after they are already acted upon.
This may be especially important when emotional
association networks get activated, as in the case of
a perceived provocation. As Ryan and Deci [2000]
suggest, an open awareness is necessary for choosing
actions that coincide with one’s values, needs, and
best interest. Initial evidence in fact suggests that
mindfulness is positively associated with value-
concordant behavior and negatively associated
with impulsivity [Brown and Ryan, 2003]. Future
studies should explore these and other potential
40 Borders et al.
Aggr. Behav.

mechanisms of the relationship between mindfulness
and behavioral aggression.
The current studies also suggest important informa-
tion about the construct of mindfulness. Our use of
SEM analyses allowed us to confirm the single-factor
structure of the MAAS. In both samples, the mind-
fulness items assessing awareness of internal states
showed the weakest factor loadings, whereas items
tapping intentional attention had the strongest factor
loadings. These findings concur with a recent item
response analysis suggesting that the MAAS most
clearly measures automaticity [Van Dam et al., 2009;
see also MacKillop and Anderson, 2007]. Addition-
ally, mindfulness shared no variance with reflection,
suggesting that mindfulness is not related to self-
analysis or philosophical introspection [see also Brown
and Ryan, 2003]. Interestingly, both samples showed
the same mean level of mindfulness and similar
correlations between mindfulness and other study
variables. These similarities across samples suggest
that the MAAS functions similarly across populations
with and without extensive knowledge about mind-
fulness. Brown and Ryan [2004] in fact suggest that
mindfulness is an inherent, individual capacity, rather
than a product of any behavioral practice like
meditation. Future research should investigate the
meaning of self-reported mindfulness and the associa-
tions between mindfulness, spiritual beliefs, and
meditation experience in different populations.
This study possessed some methodological limita-
tions that should be addressed in future research.
As stated above, statistical tests of mediation
with cross-sectional data cannot illuminate the
temporal associations between study variables.
Thus, although we proposed that mindfulness
contributes to decreased rumination, which in turn
contributes to decreased anger, hostility, and ag-
gression, our results cannot rule out other possibi-
lities. For instance, high levels of rumination could
make people less mindful, which might lead to
increased subsequent anger, hostility, and aggres-
sion. It does seem theoretically plausible that people
who abstain from the repetitive and inflexible
process of rumination have more available attention
for the present moment. However, we know of no
experimental studies that have examined this possi-
bility. Moreover, experimental studies suggest that
increased mindfulness causes reductions in rumina-
tion [Ramel et al., 2004; Shapiro et al., 2008],
indicating the same directionality that we proposed.
Future research could examine the temporal asso-
ciation between mindfulness and rumination.
Alternatively, our results cannot rule out the
possibility that low aggressive tendencies might cause
individuals to engage in high levels of mindfulness
and low levels of rumination. However, substantial
experimental evidence suggests that higher levels of
rumination in fact cause increased behavioral
aggression, compared with lower levels of rumina-
tion [Bushman, 2002; Bushman et al., 2005; Rusting
and Nolen-Hoeksema, 1998]. We know of no studies
showing that manipulated aggression increases sub-
sequent rumination. Likewise, meditation training
leads to decreased aggressive behavior [Heppner
et al., 2008; Singh et al., 2007a,b]. Proposed
mechanisms of mindfulness include intentional
attention, cognitive and behavioral flexibility, self-
regulation, and values clarification [Shapiro et al.,
2006]. It seems much more likely that these mechan-
isms result in low aggressive tendencies than the
other way around. Therefore, both empirical and
theoretical work strongly suggest that aggression-
related variables should be the DVs in our model.
However, only longitudinal or experimental studies
can rule out alternative possibilities.
In addition, we included only self-report measures,
which allowed the possibility of reporter bias. As
with any study of aggression, participants may
underreport aggressive attitudes or behaviors due
to social norms. Both of our samples reported less
physical aggression than in other published research
[e.g., Buss and Perry, 1982]. Given that under-
reporting tendencies would diminish the likelihood
of finding correlations, these data may actually
underestimate true effects. Future studies should
therefore include behavioral measures of aggression.
However, this limitation does not pose a barrier to
interpretation of our data.
In sum, our studies contribute important new
information about how mindfulness relates to
rumination and aggression-related variables and
suggest several hypotheses for future research. First,
our results suggest theoretical hypotheses for how
mindfulness might help a person respond to a
provocation or triggering event. Specifically, pre-
sent-focused and intentional attention may prevent
the destructive cognitive and emotional cycles that
occur following a perceived provocation. Associa-
tive network theories suggest that when a particular
provoking trigger is experienced, related emotions,
beliefs, and memories are brought to mind
[Anderson and Bushman, 2002; Berkowitz, 1990].
Some research suggests that anger in particular
produces a high amount of arousal and cognitive
processing [Baumeister et al., 1990]. It is likely that
people with a tendency to ruminate will harp on
perceived provocations and possible revenge, thus
keeping their aggressive association networks active
41Mindfulness and Rumination
Aggr. Behav.

and increasing their likelihood of reporting hostile
thoughts and angry feelings [Berkowitz, 1990;
Rusting and Nolen-Hoeksema, 1998]. Mindfulness
may help stop or even prevent this activation by
keeping a person attentive to the present moment
and preventing rumination about past provocations.
Therefore, a mindful person may actually experience
less anger and hostility following a triggering event,
in part because of decreased rumination. Although
the current studies suggest that rumination does not
mediate the link between mindfulness and disposi-
tional physical aggression, it is possible that
rumination might be more related to situationally
induced aggression. Thus, perhaps a mindful person
would respond to an acute provocation with less
rumination and subsequently less behavioral aggres-
sion. Alternatively, mindfulness might help a person
respond to a provocation with less behavioral
aggression through other mechanisms like relaxa-
tion or self-regulation. Many more experimental
studies are needed to test these specific hypotheses.
Our nonexperimental findings also have important
implications for future anger management treat-
ments. Interventions for anger problems typically
include contingency management, stimulus control,
or social skills training [Deffenbacher et al., 2002;
Edmondson and Conger, 1996]. Mindfulness train-
ing may provide a good complement to these other
important skills. By increasing awareness of current
emotions, thoughts, and situations, mindfulness
may help promote acceptance, distress tolerance,
and intentional attention. These emotion regulation
skills provide a solid foundation for skills training.
Recent clinical interventions have incorporated
mindfulness training in treatments for depressive
relapse, chronic pain, anxiety, borderline personality
disorder, and relapse prevention in substance
abusers [see Baer, 2003 for a review]. The non-
experimental results presented here suggest the
hypothesis that mindfulness training may also
enhance treatments for anger problems. Future
research studies should empirically test the benefit
of adding mindfulness training to anger manage-
ment interventions.
REFERENCES
Anderson CA, Bushman BJ. 2002. Human aggression. Annu Rev
Psychol 53:27–51.
Anderson JC, Gerbing DW. 1988. Structural equation modeling in
practice: A review and recommended two-step approach. Psychol
Bull 103:411–423.
Arch JJ, Craske MG. 2006. Mechanisms of mindfulness: Emotion
regulation following a focused breathing induction. Behav Res
Ther 44:1849–1858.
Baer RA. 2003. Mindfulness training as clinical intervention:
A conceptual and empirical review. Clin Psychol Sci Pract
10:125–143.
Barber L, Maltby J, Macaskill A. 2005. Angry memories and
thoughts of revenge: The relationship between forgiveness and
anger rumination. Pers Individ Differences 39:253–262.
Baron RM, Kenny DA. 1986. The moderator-mediator variable
distinction in social psychological research: Conceptual,
strategic, and statistical considerations. J Pers Soc Psychol
51:1173–1182.
Baumeister RF, Stillwell A, Wotman SR. 1990. Victim and
perpetrator accounts of interpersonal conflict: Autobiographical
narratives about anger. J Pers Soc Psychol 59:994–1005.
Berkowitz L. 1990. On the formation and regulation of anger and
aggression: A cognitive-neoassociationistic analysis. Am Psychol
45:494–503.
Bishop SR, Lau M, Shapiro S, Carlson L, Anderson ND, Carmody J,
Segal ZV, Abbey S, Speca M, Velting D, Devins G. 2004.
Mindfulness: A proposed operational definition. Clin Psychol Sci
Pract 11:230–241.
Bollen KA, Stine R. 1990. Direct and indirect effects: Classical and
bootstrap estimates of variability. Sociol Methodol 20:115–140.
Bower GH. 1981. Mood and memory. Am Psychol 36:129–136.
Broderick PC. 2005. Mindfulness and coping with dysphoric
mood: Contrasts with rumination and distraction. Cogn Ther
Res 29:501–510.
Brown KW, Ryan RM. 2003. The benefits of being present:
Mindfulness and its role in psychological well-being. J Pers Soc
Psychol 84:822–848.
Brown KW, Ryan RM. 2004. Perils and promise in defining and
measuring mindfulness: Observations from experience. Clin
Psychol Sci Pract 11:242–248.
Bryant FB, Smith BD. 2001. Refining the architecture of aggression:
A measurement model for the Buss–Perry Aggression Ques-
tionnaire. J Res Pers 35:138–167.
Bushman BJ. 2002. Does venting anger feed or extinguish the flame?
Catharsis, rumination, distraction, anger and aggressive respond-
ing. Pers Soc Psychol Bull 28:724–731.
Bushman BJ, Pederson WC, Vasquez EA, Bonacci AM, Miller N.
2005. Chewing on it can chew you up: Effects of rumination on
triggered displaced aggression. J Pers Soc Psychol 38:969–983.
Buss AH, Perry M. 1992. The aggression questionnaire. J Pers Soc
Psychol 63:452–459.
Caprara GV. 1986. Indicators of aggression: The dissipation-
rumination scale. Pers Individ Differences 7:763–769.
Caprara GV, Gargaro BT, Pastorelli C, Prezza M, Renzi P, Zelli A.
1987. Individual differences and measures of aggression in
laboratory studies. Pers Individ Differences 8:885–893.
Caprara GV, Paciello M, Gerbino M, Cugini C. 2007. Individual
differences conducive to aggression and violence: Trajectories
and correlates of irritability and hostile rumination through
adolescence. Aggr Behav 33:359–374.
Carver CS. 1975. Physical aggression as a function of objective self-
awareness and attitudes toward punishment. J Exp Soc Psychol
11:510–519.
Carver CS, Scheier MF. 1981. Attention and Self-Regulation:
A Control Theory Approach to Human Behavior. New York:
Springer.
Chambers R, Lo BCY, Allen NB. 2008. The impact of intensive
mindfulness training on attentional control, cognitive style, and
affect. Cogn Ther Res 32:303–322.
42 Borders et al.
Aggr. Behav.

Chermack ST, Giancola PR. 1997. The relation between alcohol and
aggression: An integrated biopsychosocial conceptualization.
Clin Psychol Rev. Special Issue: Biopsychosoc Conceptualiza-
tions of Hum Aggression 17:621–649.
Cheung G, Lau R. 2008. Testing mediation and suppression effects
of latent variables: Bootstrapping with structural equation
models. Organ Res Methods 11:296–325.
Coffey KA, Hartman M. 2008. Mechanisms of action in the inverse
relationship between mindfulness and psychological distress.
Complement Health Pract Rev 13:79–91.
Collins K, Bell R. 1997. Personality and aggression: The dissipation-
rumination scale. Pers Individ Differences 22:751–755.
Davis RN, Nolen-Hoeksema S. 2000. Cognitive inflexibility among
ruminators and non-ruminators. Cogn Ther Res 24:669–711.
Deffenbacher JL, Oetting ER, DiGiuseppe RA. 2002. Principles of
empirically supported interventions applied to anger manage-
ment. Couns Psychol 30:262–280.
Edmondson CB, Conger JC. 1996. A review of treatment efficacy for
individuals with anger problems: Conceptual, assessment and
methodological issues. Clin Psychol Rev 16:251–275.
Fenigstein A, Scheier MF, Buss AH. 1975. Public and private self-
consciousness: Assessment and theory. J Consult Clin Psychol
43:522–527.
Giancola PR. 2000. Executive functioning: A conceptual framework
for alcohol-related aggression. Exp Clin Psychopharmacol
8:576–597.
Gross JJ. 1998. The emerging field of emotion regulation: An
integrative review. Rev Gen Psychol. Special Issue: New Dir Res
Emotion 2:271–299.
Grossman P. 2008. On measuring mindfulness in psychosomatic and
psychological research. J Psychosom Res 64:405–408.
Grossman P, Niemann L, Schmidt S, Wallach H. 2004. Mindfulness-
based stress reduction and health benefits: A meta-analysis.
J Psychosom Res 57:35–43.
Hamilton NA, Ingram RE. 2001. Self-focused attention and coping:
Attending to the right things. In: Snyder CR (ed.). Coping
With Stress: Effective People and Processes. London: Oxford
University Press, pp 178–195.
Harris JA. 1995. Confirmatory factor analysis of the aggression
questionnaire. Behav Res Ther 33:991–993.
Harris JA. 1997. A further evaluation of the aggression questionnaire:
Issues of validity and reliability. Behav Res Ther 35:1047–1053.
Hayes SC, Strosahl KD, Wilson KG. 1999. Acceptance and
Commitment Therapy: An Experiential Approach to Behavior
Change. New York: Guilford Press.
Hennessy D, Wiesenthal D. 2004. Age and vengeance as predictors
of mild driver aggression. Violence Vict 19:469–477.
Heppner WL, Kernis MH, Lakey CE, Campbell WK, Goldman BM,
Davis PJ, Cascio EV. 2008. Mindfulness as a means of reducing
aggressive behavior: Dispositional and situational evidence. Aggr
Behav 34:486–496.
Ingram RE. 1990. Self-focused attention in clinical disorders: Review
and a conceptual model. Psychol Bull 107:156–176.
Izard CE, King KA, Trentacosta CJ, Morgan JK, Laurenceau J,
Krauthamer-Ewing ES, Finlon KJ. 2008. Accelerating the
development of emotion competence in Head Start children:
Effects on adaptive and maladaptive behavior. Dev Psychopathol
20:369–397.
Kinney TA, Smith BA, Donzella B. 2001. The influence of sex,
gender, self-discrepancies, and self-awareness on anger and verbal
aggressiveness among U.S. college students. J Soc Psychol
141:245–275.
Lattin J, Carroll JD, Green PE. 2003. Analyzing Multivariate Data.
Pacific Grove, CA: Brooks/Cole-Thomson Learning.
Little RJA, Rubin DB. 1987. Statistical Analysis With Missing Data.
New York: Wiley.
Lopata C. 2003. Progressive muscle relaxation and aggression among
elementary students with emotional or behavioral disorders.
Behav Disord 28:162–172.
Lyubomirsky S, Nolen-Hoeksema S. 1995. Effects of self-focused
rumination on negative thinking and interpersonal problem-
solving. J Pers Soc Psychol 69:176–190.
MacKillop J, Anderson EJ. 2007. Further psychometric validation of
the Mindful Attention Awareness Scale (MAAS). J Psychopathol
Behav Assess 29:289–293.
MacKinnon DP, Lockwood CM, Hoffman JM, West SG, Sheets V.
2002. A comparison of methods to test mediation and other
intervening variable effects. Psychol Methods 7:83–104.
MacKinnon DP, Lockwood CM, Williams CM. 2004. Confidence
limits for the indirect effect: Distribution of the product and
resampling methods. Multivariate Behav Res 39:99–128.
Martin LL, Tesser A. 1996. Some ruminative thoughts. In:
Wyer Jr RS (ed.). Ruminative Thoughts, Advances in Social
Cognition, Vol. IX.New Jersey: Lawrence Erlbaum Associates,
pp 1–48.
McCullough ME, Rachal KC, Sandage SJ, Worthington EL,
Brown SW, Height TL. 1998. Interpersonal forgiving in close
relationships: II. Theoretical elaboration and measurement.
J Pers Soc Psychol 75:1586–1603.
McCullough ME, Bellah CG, Kilpatrick SD, Johnson JL. 2001.
Vengefulness: Relationships with forgiveness, rumination, well-
being, and the big five. Pers Soc Psychol Bull 27:601–610.
McNulty JK, Hellmuth JC. 2008. Emotion regulation and intimate
partner violence in newlyweds. J Fam Psychol 22:794–797.
Miller N, Pederson WC, Earleywine M, Pollock VE. 2003.
A theoretical model of triggered displaced aggression. Pers Soc
Psychol Rev 7:75–97.
Mor N, Winquest J. 2002. Self-focused attention and negative affect:
A meta-analysis. Psychol Bull 128:638–662.
Morren M, Meesters C. 2002. Validation of the Dutch version of the
aggression questionnaire in adolescent male offenders. Aggr
Behav 28:87–96.
Morrow J, Nolen-Hoeksema S. 1990. Effects of responses to
depression on the remediation of depressive affect. J Pers Soc
Psychol 58:519–527.
Muthe
´n LK, Muthe
´n BO. 1998–2007. Mplus User’s Guide, 5th
edition. Los Angeles, CA: Muthe
´n & Muthe
´n.
Nolen-Hoeksema S, Morrow J, Fredrickson BL. 1993. Response
styles and the duration of episodes of depressed mood. J Abnorm
Psychol 102:20–28.
O’Connor DB, Archer J, Wu FWC. 2001. Measuring aggression:
Self-reports, partner reports and responses to provoking scenar-
ios. Aggr Behav 27:79–101.
Philippot P, Brutoux F. 2007. Induced rumination dampens
executive processes in dysphoric young adults. J Behav Ther
Exp Psychiatry 39:219–227.
Preacher KJ, Hayes AF. 2004. SPSS and SAS procedures for
estimating indirect effects in simple mediation models. Behav Res
Methods Instrum Comput 36:717–731.
Prentice-Dunn S, Rogers RW. 1982. Effects of public and private
self-awareness on deindividuation and aggression. J Pers Soc
Psychol 43:503–513.
Ramel W, Goldin PR, Carmona PE, McQuaid JR. 2004. The
effects of mindfulness meditation on cognitive processes and
affect in patients with past depression. Cogn Ther Res 28:
433–455.
Rude SS, Maestas KL, Neff K. 2007. Paying attention to distress:
What’s wrong with rumination? Cogn Emotion 21:843–864.
43Mindfulness and Rumination
Aggr. Behav.

Rusting CL, Nolen-Hoeksema S. 1998. Regulating responses to
anger: Effects of rumination and distraction on angry mood.
J Pers Soc Psychol 74:790–803.
Ryan EL, Deci RM. 2000. Self-determination theory and the
facilitation of intrinsic motivation, social development, and
well-being. Am Psychol 55:68–78.
Scheier MF, Fenigstein A, Buss AH. 1974. Self-awareness and
physical aggression. J Exp Soc Psychol 10:264–273.
Segerstrom SC, Shortridge BE, Stanton AL, Alden LE. 2003.
A multidimensional structure for repetitive thought: What’s on
your mind, and how, and how much? J Pers Soc Psychol
85:909–921.
Sexton KA, Dugas MJ. 2008. The Cognitive Avoidance Question-
naire: Validation of the English translation. J Anxiety Disord
22:355–370.
Shapiro SL, Carlson LE, Astin JA, Freedman B. 2006. Mechanisms
of mindfulness. J Clin Psychol 62:373–386.
Shapiro SL, Oman D, Thoresen CE, Plante TG, Flinders T. 2008.
Cultivating mindfulness: Effects on well-being. J Clin Psychol
64:840–862.
Shrout PE, Bolger N. 2002. Mediation in experimental and
nonexperimental studies: New procedures and recommendations.
Psychol Methods 7:422–445.
Singh NN, Lancioni GE, Singh Joy SD, Winton ASW, Sabaawi M,
Wahler RG, Singh J. 2007a. Adolescents with conduct disorder
can be mindful of their aggressive behavior. J Emotional Behav
Disord 15:56–63.
Singh NN, Lancioni GE, Winton ASW, Adkins AD, Wahler RG,
Sabaawi M, Singh J. 2007b. Individuals with mental illness can
control their aggressive behavior through mindfulness training.
Behav Mod 31:313–328.
Sobel ME. 1982. Asymptotic intervals for indirect effects in
structural equations models. In: Leinhart S (ed.). Sociological
Methodology. San Francisco: Jossey-Bass. pp 290–312.
Sukhodolsky DG, Golub A, Cromwell EN. 2001. Development and
validation of the anger rumination scale. Pers Individ Differences
31:689–700.
Teasdale JD, Segal Z, Williams JMG. 1995. How does cognitive
therapy prevent depressive relapse and why should attentional
control (mindfulness) training help? Behav Res Ther 33:25–39.
Teasdale JD, Segal ZV, Williams MG. 2003. Mindfulness
training and problem formulation. Clin Psychol Sci Pract 10:
157–160.
Thompson BL, Waltz JA. 2008. Mindfulness, self-esteem, and
unconditional self-acceptance. J Rat Emotive Cogn Behav Ther
26:119–126.
Trapnell PD, Campbell JD. 1999. Private self-consciousness and the
five-factor model of personality: Distinguishing rumination from
reflection. J Pers Soc Psychol 76:284–304.
Tyson P. 1998. Physiological arousal, reactive aggression, and the
induction of an incompatible relaxation response. Aggr Violent
Behav 3:143–158.
Van Dam NT, Earleywine M, Borders A. 2009. Item response
theory analysis of the Mindful Attention Awareness Scale; under
review.
Vigil-Colet A, Lorenzo-Seva U, Codorniu-Raga MJ, Morales F.
2005. Factor structure of the Buss-Perry aggression questionnaire
in different samples and languages. Aggr Behav 31:601–608.
Wegner DM, Zanakos S. 1994. Chronic thought suppression. J Pers
62:615–640.
Wenzlaff RM, Wegner DM. 2000. Thought suppression. Annu Rev
Psychol 51:59–91.
Werts CE, Linn RL, Joreskog KG. 1974. Intraclass reliability
estimates: Testing structural assumptions. Educ Psychol Meas
34:25–33.
Williams TY, Boyd JC, Cascardi MA, Poythress N. 1996. Factor
structure and convergent validity of the Aggression Question-
naire in an offender population. Psychol Assess 8:398–403.
44 Borders et al.
Aggr. Behav.