PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES
All in the Mind’s Eye? Anger Rumination and Reappraisal
Rebecca D. Ray
Frank H. Wilhelm
University of Basel
James J. Gross
Research on rumination has demonstrated that compared with distraction, rumination intensifies and
prolongs negative emotion. However, rumination and distraction differ both in what one thinks about and
how one thinks about it. Do the negative outcomes of rumination result from how people think about
negative events or simply that they think about them at all? To address this question, participants in 2
studies recalled a recent anger-provoking event and then thought about it in 1 of 2 ways: by ruminating
or by reappraising. The authors examined the impact of these strategies on subsequent ratings of anger
experience (Study 1) as well as on perseverative thinking and physiological responding over time (Study
2). Relative to reappraisal, rumination led to greater anger experience, more cognitive perseveration, and
greater sympathetic nervous system activation. These findings provide compelling new evidence that
how one thinks about an emotional event can shape the emotional response one has.
Keywords: rumination, reappraisal, anger, emotion regulation, autonomic nervous system
Emily and Bob stand on the corner waiting for their coworker
Jeff, who is driving the van pool to work. Nervous about her
unfinished presentation, Emily waits impatiently and mentally
rehearses her main points. Bob glances anxiously at his watch, not
wanting to miss an important phone call. The van finally arrives,
though late enough to leave Emily no time to finish her presenta-
tion and to make Bob miss his call. Jeff, the driver, apologizes for
being late again. Throughout the ride to work, Emily ruminates on
Jeff’s tardiness and gets angrier by the minute. By contrast, Bob
thinks about how Jeff, though not punctual, is a fine coworker
overall. So Bob decides to propose to the van pool that they take
turns covering Jeff’s driving duties.
An age-old conviction holds that how we think about an event
determines how we feel about it. This proposition is the corner-
stone of appraisal theories of emotion (Arnold, 1960; Lazarus,
1991; Schachter & Singer, 1962; Scherer, Schorr, & Johnstone,
2001). Appraisal theories postulate that the appraisals (or evalua-
tions) an individual makes in a given situation determine both the
type and intensity of emotion that a person will have in that
particular situation. One important corollary of this view is that
modifying how an individual appraises an event should alter the
emotional responses to that event. Despite the intuitive appeal of
this proposition, it has been very difficult to disentangle the effects
of thinking about an upsetting event per se from the effects of
using one way versus another of thinking about that event.
One literature area concerned with the link between thinking
and feeling is the literature on rumination, which may be broadly
defined as repetitive thoughts around a common theme, absent
immediate environmental demands for those thoughts (Martin &
Tesser, 1996). Findings suggest that, compared with distraction,
thinking repeatedly about an emotional event maintains and/or
amplifies the emotional response (Morrow & Nolen-Hoeksema,
1990). Furthermore, individual differences in rumination predict
both the diagnosis (Nolen-Hoeksema, 2000) and severity of de-
pression (Just & Alloy, 1997). One important limitation of this
literature is that it is not yet clear whether rumination produces
negative emotional outcomes solely because the negative material
is called to mind or because of the way that material is called to
To address this issue, rumination must be compared not with
distraction, which involves thinking about something unrelated,
but with another cognitive strategy for processing the same emo-
tional event. In the following sections, we first review the literature
on rumination and its consequences and then distinguish rumina-
tion from reappraisal, the process of reinterpreting the meaning of
an upsetting event. We then present two studies that examine how
these two different ways of thinking about a personally relevant
Rebecca D. Ray and James J. Gross, Department of Psychology, Stan-
ford University; Frank H. Wilhelm, Department of Psychology, University
of Basel, Basel, Switzerland.
We thank Ozlem Ayduk, Carl Frankel, Cendri Hutcherson, Kelly
McGonigal, and Kateri McRae for their helpful comments on an earlier
version of this article.
Correspondence concerning this article should be addressed to Rebecca
D. Ray, Department of Psychology, Stanford University, Stanford, CA
94305. E-mail: email@example.com
Journal of Personality and Social Psychology, 2008, Vol. 94, No. 1, 133–145
Copyright 2008 by the American Psychological Association 0022-3514/08/$12.00 DOI: 10.1037/0022-35220.127.116.11
anger episode predict different emotional responses, as assessed by
self-reported emotion experience and by physiological responding.
Defining rumination as repetitive self-focused attention to sad or
angry thoughts and feelings, Nolen-Hoeksema and her colleagues
(Nolen-Hoeksema & Morrow, 1991) have shown rumination to be
associated with several negative consequences. Individuals who
ruminate report higher levels of negative emotion (Segerstrom,
Tsao, Alden, & Craske, 2000) and less control over intrusive
thoughts (Watkins, 2004). In addition, a tendency to ruminate is
associated with longer lasting depressive symptoms (Nolen-
Hoeksema, Morrow, & Fredrickson, 1993) and an increase in
occurrence of depressive episodes (Nolen-Hoeksema, 2000). Ru-
mination has been linked to physiological indicators of negative
emotion, including increased and/or prolonged amygdalar activa-
tion in response to negative pictures and negative words (Ray et
al., 2005; Siegle, Steinhauer, Thase, Stenger, & Carter, 2002) and
increased cortisol levels after a stressor (Roger & Jamieson, 1988).
Alternatives to Rumination: Distraction and Cognitive
Experimental studies have focused on one alternative to rumi-
nation, namely, distraction—moving one’s attention away from
the upsetting event and onto unrelated neutral contents. Compared
with distraction, rumination increases or maintains sad moods for
depressed individuals (Morrow & Nolen-Hoeksema, 1990) as well
as increasing or maintaining anger and aggressive behavior (Bush-
man, Bonacci, Pedersen, Vasquez, & Miller, 2005; Rusting &
Nolen-Hoeksema, 1998). However, this comparison has several
important limitations, the most important of which is the fact that
distraction by definition involves not thinking about the emotion-
generating event. Comparisons of rumination to distraction cannot
resolve whether adverse consequences are guaranteed just by
thinking about a negative event or whether the way one thinks
about the negative event affects the outcome.
Investigators have started to explore alternatives to rumination
that, unlike distraction, keep the focus on the upsetting situation.
One alternative to rumination is cognitive reappraisal, which
entails keeping the emotion-eliciting event in mind and actively
seeking alternate interpretations of the meaning or self-relevance
of the event (Gross, 2001). Reappraisal as an emotion regulation
strategy was first specified in the stress and coping work of
Lazarus and colleagues (Lazarus & Alfert, 1964) and in appraisal
theories of emotion (Lazarus, 1991; Schachter & Singer, 1962).
We conceptualize cognitive reappraisal as a rich class of cognitive
strategies that alter a person’s initial appraisals so that an emo-
tional stimulus comes to have a different emotional impact.
Among others, reappraisal can involve considering new informa-
tion about the stimulus and/or taking a different perspective on the
stimulus’ relevance to one’s goals.
Previous studies have documented the efficacy of reappraisal.
The proclivity to reappraise is correlated with fewer depressive
symptoms (Garnefski & Kraaij, 2006) and with both expressions
of greater positive emotion and better interpersonal functioning
(Gross & John, 2003). Reappraisal has also been linked to de-
creased startle response and to decreased amygdalar activation in
negative slide-viewing paradigms (Jackson, Malmstadt, Larson, &
Davidson, 2000; Ochsner, Bunge, Gross, & Gabrieli, 2002;
Ochsner et al., 2004) as well as to decreased cardiovascular re-
sponse to disgusting stimuli and to lesser blood pressure increases
in response to negative slides (Richards & Gross, 1999). Work on
reappraisal is augmented by Kross, Ayduk, and Mischel (2005),
who reported that creating distance—by visualizing increased
physical distance—is insufficient to attenuate negative emotions.
Reinterpreting or gaining new insight seems to be necessary for
such visualizations to work.
Another relevant line of work focuses on changing the social
perspective from a first-person to a third-person perspective. This
strategy draws attention to the broader meaning or consequences
of events (Libby & Eibach, 2002), expanding one’s own aperture,
to consider additional situational features to be important. This can
create a more nuanced or altogether different appraisal of events,
which, in turn, can change the character and the intensity of the
person’s emotional response to the event (Lazarus, 1991). Such
changes in cognitions have been suggested as the mechanism by
which cognitive therapy works to decrease the distress in psycho-
pathology (Clark, Beck, & Alford, 1999). Although clinical trials
have demonstrated that cognitive therapy can reduce distress, few
studies to date have tested the acute effects of the mechanism by
which it is thought to work.
The Present Studies
There is growing support for the notion that an upsetting event
may be called to mind in more or less disruptive ways. However,
there are two important gaps in the research literature. First, there
has thus far been no direct experimental comparison of rumination
to a different and less disruptive way of continuing to think about
negative events (such as reappraisal). Second, the effectiveness of
reappraisal has not yet been demonstrated in the context of potent
self-generated negative autobiographical events.
To address these important gaps, we conducted two experimen-
tal studies within the context of recent anger-inducing autobio-
graphical events and directly contrasted rumination and reappraisal
as ways of thinking about negative life events. The first study
contrasts the emotional and cognitive consequences of ruminating
or reappraising while recalling a recent anger-eliciting event. The
second study extends the first by examining how rumination and
reappraisal differentially affect multiple channels of emotional
responding, including physiology, over time. In both studies, our
general expectation was that reappraisal (compared with rumina-
tion) would produce less anger experience and fewer thoughts
about the anger-eliciting event. To control the reappraisal strategy
participants deployed, these studies focus on a specific reappraisal
strategy—viewing an upsetting autobiographical event from the
perspective of an uninvolved observer or a mediator. This strategy
maintains the negative event as the focus of attention, while
allowing for the possibility of alternate interpretations of the
meaning and/or relevance of the event. This form of reappraisal
thus provides a strong contrast to rumination.
Study 1: Anger Rumination Versus Reappraisal
In Study 1, we asked participants to call to mind a recent,
personally relevant, anger-eliciting event. We chose anger because
RAY, WILHELM, AND GROSS
prior work on rumination has demonstrated the importance of
anger rumination in maintaining angry moods, increasing the like-
lihood of aggressive acts, and adversely affecting physiological
recovery (e.g., Bushman et al., 2005; Glynn, Christenfeld, &
Gerin, 2002). To ensure a clear contrast between rumination and
reappraisal, we then asked participants either to ruminate about
this event (review it repeatedly from a first-person perspective
focusing on their own feelings) or to reappraise this event (e.g.,
review it from an objective third-person perspective and how the
third person might feel).
To examine the acute effects of rumination and reappraisal on
emotional responding, we assessed self-reports of emotion expe-
rience at three time points: (a) at baseline, (b) after individuals
called the event to mind without other instructions, and (c) again
after individuals called the event to mind with specific between-
groups experimental instructions to ruminate or reappraise. The
uninstructed period, in which participants thought about the event
without yet having specific instructions to ruminate or reappraise,
allowed us to evaluate, within subjects, how calling and holding an
event in mind differs from rumination or reappraisal.
On the basis of the prior literature, we expected rumination and
reappraisal to differ both during the acute emotional response and
during the recovery period. Specifically, we expected that reap-
praisal would produce low levels of anger experience during the
recall period and would facilitate recovery during the recovery
period, whereas rumination would perpetuate angry feelings and
thoughts about the anger-eliciting event both during the recall and
Eighty-two female participants aged 18–27 years (M ? 19.3,
SD ? 1.7) were recruited from an introductory psychology course
and given course credit. Only female participants were used to
reduce variability in anger response (Linden et al., 2003). Partic-
ipants identified their ethnicity as Caucasian (50%), African Amer-
ican (4.9%), Asian American (26.9%), Latina/Hispanic (10.9%),
and Native American (1.2%). Five participants’ data were ex-
cluded, 3 due to high ratings of dysphoria at baseline, and 2 for not
following instructions, leaving usable data for 77 participants.
A female experimenter conducted individual experimental ses-
sions. Participants began by rating the amount of positive and
negative emotions they currently felt (baseline rating). Each par-
ticipant then identified an unresolved event from the preceding 2
weeks in which she became very angry with another person
(angering event). The participant then rated how unresolved the
event was and the amount of anger she felt during the event as well
as other negative and positive emotions. This personal, unresolved
anger event was then used as the focal content in each of four
subsequent task periods.
In the first period, the participant freely recalled the event and
kept it in mind for 2 min with no other specific instructions
(free-recall phase). Immediately after this period, the participant
rated her current anger and other emotions. The participant also
indicated what percentage of the period she had spent thinking
about the event and the extent to which she viewed the event from
her own perspective. This was followed by a period in which the
participant was told to “Rest quietly” for 2 min. At the end of the
period, the participant rated her current emotions and again indi-
cated what percentage of the previous rest period was spent think-
ing about the event that had previously made her angry.
The participant was then randomly assigned to either a rumina-
tion condition (n ? 34) or a reappraisal condition (n ? 43). In each
condition, the participant was instructed to think about the angry
event for 2 min. In the rumination condition, participants also were
instructed to “think about [the event] from your own perspective,
and turn it over and over in your mind. Focus on those things that
initially made you feel and respond the way that you did.” In the
reappraisal condition, participants were instructed to “think about
[the event] from a different perspective from the one you used
earlier. For example, you might try to see this event from the
perspective of an impartial observer.” At the end of the period, the
participant again provided ratings of anger as well as other emo-
tions, an estimate of the actual time spent thinking about the
anger-inducing event during the recall period, and ratings of the
extent to which the participant viewed the event from her own
perspective. In the final period, the participant was again instructed
to “Rest quietly,” followed by a final rating of anger and other
emotions as well as an estimate of time spent thinking about the
Instruction manipulation check.
to which they thought about the event from their own perspective
and, separately, from another person’s perspective. Ratings were
made using a 5-point Likert scale (0 ? not at all, 4 ? a lot).
Time thinking about anger episode.
percentage of time they spent thinking about the emotional event
(0%–100%). This allowed us to test whether people were thinking
about the event when asked to and also assess the extent to which
participants continued to think about the event during the rest
periods without being instructed to do so.
The experience of anger, positive emotion, and negative emo-
Participants rated the degree to which they felt anger as well
as five positive emotions (happy, amused, proud, hopeful, affec-
tionate; ?? .82) and six negative emotions (anxious, jealous, sad,
rejected, guilty, confused; ? ? .81). Ratings were made on a
5-point Likert scale (0 ? not at all, 4 ? a lot).
Participants wrote about the perspective from
which they thought about the event. Raters categorized narratives
into one of five different perspectives. These perspectives varied in
distance from the participant and the event. The perspective closest
was one’s own perspective; then the perspective of the other
person with whom the participant was angry; the perspective of a
good friend involved in the anger-eliciting event; the perspective
of a good friend not involved in the anger-eliciting event; and the
most distant perspective was that of a stranger not involved in the
situation. Interrater reliabilities were high (K ? .96).
Participants rated the extent
Participants estimated the
To examine anger experience, we subtracted the anger at base-
line (i.e., before thinking about the anger-eliciting event) from
RUMINATION AND REAPPRAISAL
subsequent anger ratings. Nested repeated measures analyses of
variance (ANOVAs) were conducted, with condition (ruminate vs.
reappraise) as a between-subjects factor and time (recall before
manipulation and recall after manipulation) as a within-subjects
factor, with task (recall and rest) nested within time. Significant
main effects and interactions were followed up with independent t
tests and paired t tests. Greenhouse–Geisser corrections for het-
eroscedasticity were applied where necessary (Keselman, Algina,
& Kowalchuk, 2001). Effect sizes are reported in terms of partial
eta squared, with the notation ?p
aspect of the event recalled,1nor did they differ in the amount of
time spent thinking about the anger-eliciting event during both the
free-recall phase and instructed recall phase, as evidenced by the
lack of main effect of condition, F(1, 75) ? 0.62, p ? .43. As
expected, those instructed to ruminate viewed the event from their
own perspective (M ? 3.30 ? 0.84) more than those who were
instructed to reappraise (M ? 1.14 ? 0.86), t(75) ? 11.28, p ?
.001. Also, those assigned to reappraise viewed the event from
another person’s perspective (M ? 2.59 ? 1.34) more than those
who were assigned to ruminate (M ? 0.76 ? 0.92), t(75)? ?6.80,
p ? .001.
When participants called the anger-eliciting
event to mind prior to group assignment instructions, they reported
significantly greater levels of anger as compared with baseline
(Mbaseline? 0.12 ? 0.56, Mfree recall l?1.78 ? 0.95), t(75) ?
15.01, p ? .001. During the rest period immediately following,
participants reported substantial decreases in their anger (Mrest1?
0.46 ? 0.73), t(75) ? 11.78, p ? .001. Success of random
assignment was reflected by an absence of any differences be-
tween experimental groups during these periods (all ps ? .58).
The two groups did not differ in any
Main Analyses: Emotion Experience
To test whether reappraisal and rumination differed in their conse-
quences for anger, a nested repeated measures ANOVA was per-
the two tasks after the instructions (instructed recall, Rest 2), with
condition (ruminate, reappraise) as a between-subjects variable and
time (prior to group assignment instructions, after group assignment
instructions) and task (recall, rest) as within-subjects variables. This
analysis revealed a three-way Time ? Task ? Condition interaction,
F(1, 75) ? 7.71, p ? .007, ?p
Condition interaction, F(1, 75) ? 5.60, p ? .021, ?p
To follow-up on the three-way interaction, a repeated measures
ANOVA was conducted separately for each condition, with task and
time as within-subjects variables.
For those in the rumination condition, there was a main effect of
task only, with the periods of both free and instructed recall
producing more self-reported anger than the rest periods (free
recall and instructed recall ? 1.77 ? 1.06, Rest 1 and Rest 2 ?
0.41 ? 0.81), F(1, 75) ? 221.66, p ? .001, ?p
no anger differences between the periods before and after the
2? .09, as well as a two-way Task ?
2? .07 (see
2? .75. There were
manipulation for those assigned to ruminate, nor any interaction
with task (ps ? .40). With respect to level of anger invoked,
simply calling to mind an unresolved angry event is not distin-
guishable from rumination.
By contrast, for those assigned to the reappraisal condition, the
repeated measures ANOVA showed an interaction between Task
and Time, F(1, 75) ? 12.65, p ? .001, ?p
effects of task and time. The periods of recall produced more anger
than the periods of rest (free recall and instructed recall ? 1.28 ?
0.96, Rest 1 and Rest 2 ? 0.30 ? 0.56), F(1, 75) ? 142.02, p ?
after the reappraisal manipulation instructions than before the
reappraisal manipulation instructions (free recall and Rest 1 ?
0.95 ? 0.79, instructed recall and Rest 2 ? 0.63 ? 0.72), F(1,
75) ? 16.32, p ? .001, ?p
recall and rest periods for those in the reappraisal condition
showed that those assigned to reappraise reported much less anger
after the manipulation in the instructed recall phase than in the
free-recall phase, t(75) ? 2.14, p ? .018 (free recall ? 1.57 ?
0.97, instructed recall ? 0.95 ? 0.94); however, there were no
differences between the two rest periods, t(75) ? 0.14. Addition-
ally, independent t tests between the groups confirmed that par-
ticipants in the rumination condition reported more anger in the
instructed recall phase than those in the reappraisal condition,
t(75) ? 2.90, p ? .002 (rumination ? 1.79 ? 1.31, reappraisal ?
0.95 ? 0.94). There were no differences in self-reported anger
between the conditions during the final rest period, t(75) ? 0.22,
p ? .4, indicating that reappraisal did not facilitate anger recovery
over and above the instruction to rest.
2? .14, as well as main
2? .65, and participants reported less anger in the periods
2? .18. Follow-up paired t tests on both
and reappraisal were specific to anger, a nested repeated measures
ANOVA was performed on the composite of negative emotions
(apart from anger) reported after each period. Time and task were
within-subjects variables, and condition was a between-subjects
To test whether the effects of rumination
1The events that participants reported were assessed along several
dimensions to ensure that random assignment was successful in producing
groups with events that were comparable. Participants in the two conditions
did not differ either in the type of event called to mind or in the intensity
of the initial emotional response.
Change from Baseline
trials in Study 1.
Anger experience (change from baseline) in recall and rest
RAY, WILHELM, AND GROSS
variable. Like the anger analysis, this analysis revealed a signifi-
cant three-way interaction of Time ? Task ? Condition, F(1,
75) ? 10.45, p ? .002, ?p
of Task ? Condition, F(1, 75) ? 4.30, p ? .042, ?p
a two-way interaction of Time ? Condition, F(1, 75) ? 5.13, p ?
and time. To follow-up on the three-way interaction, a repeated
measures ANOVA was conducted separately for each condition,
with task and time as within-subjects variables. For those in the
rumination condition, there was a main effect of task only, with the
periods of both free and instructed recall producing more self-
reported negative emotion than the rest periods (free recall and
instructed recall ? 1.29 ? 0.75, Rest 1 and Rest 2 ? 0.49 ? 0.42),
F(1, 75) ? 55.05, p ? .001, ?p
in negative emotion between the periods before and after the
manipulation for those assigned to ruminate, nor any interaction
with task (ps ? .63). With respect to amount of negative emotion
invoked, as with anger invoked, simply calling to mind an angry
event is not distinguishable from rumination.
For those assigned to the reappraisal condition, the repeated
measures ANOVA showed an interaction between task and time,
F(1, 75) ? 32.03, p ? .001, ?p
task and time. The periods of recall produced more negative
emotion than the periods of rest (free recall and instructed recall ?
0.88 ? 0.61, Rest 1 and Rest 2 ? 0.33 ? 0.39), F(1, 75) ? 70.61,
p ? .001, ?p
emotion in the periods after the reappraisal manipulation instruc-
tions than before the reappraisal manipulation instructions (free
recall and Rest 1? 0.67 ? 0.51, instructed recall and Rest 2 ?
0.54 ? 0.25), F(1, 75) ? 43.00, p ? .002, ?p
paired t tests on both recall and rest periods for those in the
reappraisal condition showed that those assigned to reappraise
reported much less negative emotion after the manipulation in the
instructed recall phase than in the free-recall phase, t(75) ? 9.20,
p ? .001 (free recall ? 1.03 ? 0.61, instructed recall ? 0.73 ?
0.61); however, there were no differences between the two rest
periods, t(75) ? ?0.91. Additionally, independent t tests between
the groups confirmed that participants in the rumination condition
reported more negative emotion in the instructed recall phase than
those in the reappraisal condition, t(75) ? 17.97, p ? .001 (rumi-
nation ? 1.32 ? 0.83, reappraisal ? 0.73 ? 0.61). There was a
significant difference in self-reported negative emotion between
the conditions during the final rest period, t(75) ? 2.36, p ? .020
(rumination ? 0.48 ? 0.49, reappraisal ? 0.34 ? 0.39), suggest-
ing that reappraisal did facilitate negative emotional recovery over
and above the instruction to rest.
One possible explanation for our results is that
the reappraisal group increased their positive affect to decrease their
anger. However, a repeated measures ANOVA performed on the
free-recall and instructed recall periods, with condition as the group
factor and time as the repeated measure, showed that there were no
differences between the groups in positive emotion at any time point
(p ? .21). In addition, a repeated measures ANOVA performed on
the free-recall and instructed recall periods, with positive emotion
partialled out of the anger reports, demonstrates that controlling pos-
itive emotion does not change the pattern of anger reports (Time ?
Condition interaction for anger partialling out positive emotion), F(1,
75) ? 10.85, p ? .002, ?p
2? .13, as well as a two-way interaction
2? .06, and
2? .07. Main effects were also evident for condition, task,
2? .64. There were no differences
2? .43, as well as main effects of
2? .62, and participants reported less negative
2? .20. Follow-up
2? .19, compared with the previous
interaction of Time ? Condition for anger, F(1, 75) ? 10.55, p ?
Amount of time thinking.
To confirm that the difference in
anger between the groups was not due to the amount of time spent
thinking about the anger-eliciting event, an ANOVA was per-
formed on the anger ratings for the instructed recall period, par-
tialling out the percentage of time spent thinking about the anger-
eliciting event. The anger ratings between the conditions remained
significantly different, F(1, 75) ? 10.44, p ? .002, ?p
those assigned to ruminate experiencing more anger than those
assigned to reappraise (rumination residual ? 0.44 ? 1.12, reap-
praisal residual ? ?0.33 ? 0.95).
If reappraisal success depends on the participant
taking a more socially distanced perspective, then we would expect
perspectives closer to the self to result in higher anger scores and
perspectives farther away, such as the perspectives of a stranger to
produce lower anger scores. To test this possibility, an ANOVA
was performed on the anger ratings of those assigned to reappraise
using the coded perspective as the independent variable. This
analysis demonstrated no differences between the reappraisal per-
spectives in anger scores during the instructed recall, F(4, 40) ?
0.18, p ? .950. Anger ratings decreased similarly, whether one
reappraised from one’s own perspective or from the perspective of
an uninvolved, distant stranger. This indicates that the mechanism
responsible for the change in anger ratings was not the degree of
social distance of the perspective taken.
2? .12, with
In Study 1, we compared two common ways of attending to and
thinking about emotionally evocative autobiographical events. Re-
appraising the event from an impartial observer’s perspective
produced less self-reported anger. This was true both within sub-
jects, compared with the prior free recall of the event, and also
between groups, compared with ruminating about the event. By
contrast, ruminating about the event yielded the same level of
anger as during the prior free recall. From these findings, we
conclude that calling an angry event to mind and maintaining focus
on it behaves much like rumination, in invoking and maintaining
anger. By contrast, reappraising an event actively alters the emo-
tional trajectory, decreasing anger.
One possible explanation for reappraisal’s effects on anger
might be that reappraisal focuses on positive aspects of the event
and increases positive emotion. However, this does not appear to
be the case because the groups did not differ in reported positive
emotion. Nor were reappraisal’s effects on anger a result of social
distance from the anger-eliciting event. Instead, these findings
suggest that reappraisal produces different emotional conse-
quences from rumination by reinterpreting the meaning of the
anger-provoking event as less intensely negative. However, in
order to understand the nature of any changes in interpretation,
more information is needed about how reappraising participants
thought about the anger-eliciting event.
Beyond the acute effects, there were also lingering effects of
rumination and reappraisal. Those assigned to ruminate continued
to think more about the anger-eliciting event after they were told
they could stop, compared with those assigned to reappraise.
Interestingly, those assigned to reappraise felt less overall negative
emotion during the rest period, but not less anger, contrary to our
RUMINATION AND REAPPRAISAL
initial hypothesis. This may be due to the instruction to rest being
taken as an imperative to clear the mind and impose calm, leaving
just lingering negative affect not specific to anger. Thus, to test
whether rumination and reappraisal differ in recovery profiles, the
instruction to rest must be replaced with a truly uninstructed
recovery period. In the same vein, one might expect stronger
effects over time, if the effects of rumination and reappraisal are
cumulative. The instructed rest period after the instructed thinking
period may have not only prevented accumulation of effects but
also increased attention on recovery as a goal. In order to test this
hypothesis of cumulative effects, in a second study (described
below), rumination and reappraisal were assessed across multiple
rumination and reappraisal cycles.
Study 2: Physiological Consequences of Rumination
Given the results of Study 1, we thought it important, in a
second study, to test directly the hypothesis that rumination and
reappraisal affect not only self-reported emotion but also addi-
tional channels of emotional responding (e.g., physiological chan-
nels) and that such differences may unfold over time. Measure-
ment of the sympathetic branch of the autonomic nervous system
provides a reliable measure of responding that is unaffected by
demand characteristics (see Bradley, 2000, for a review).
On the basis of our analysis of rumination and reappraisal, we
expected different patterns of physiological responding. Specifi-
cally, the sympathetic branch of the autonomic nervous system is
sensitive to changes in anger (Cacioppo, Berntson, Larsen, Poehl-
mann, & Ito, 2000; Stemmler, Heldmann, Pauls, & Scherer, 2001).
We thus expected rumination to be associated with greater in-
creases in sympathetic responding than reappraisal. In addition,
parasympathetic reactivity has been associated with perseverative
thinking (Brosschot, Gerin, & Thayer, 2006) and recovery trajec-
tories for depression (Rottenberg, Salomon, Gross, & Gotlib,
2005). We thus expected that rumination would be associated with
lesser parasympathetic reactivity than reappraisal. To test these
predictions, we assessed both branches of the autonomic nervous
system. To confirm that any physiological effects were not due to
effort, questions about perceived effort were added.
Previous studies have not captured the unfolding of rumina-
tion’s cumulative effects in the laboratory. In Study 2, we wanted
to examine the effects of cyclical rumination over time to better
understand how the short-term effects of rumination build. We
expected such effects to persist in subsequent uninstructed trials.
As in Study 1, we expected rumination to lead to greater amounts
of thinking about the event during uninstructed periods.
A third important goal of Study 2 was to explore the possible
mechanisms by which reappraisal changes affective responding.
Because the social distance embedded in the reappraisal strategy
had no effect in Study 1, the reappraisal perspective was made the
same for all participants. Also, to assess changes in participants’
interpretation of the recalled event, we asked how participants
thought about the anger-eliciting event during the instructed peri-
ods. On the basis of work by Kross et al. (2005) and Pennebaker,
Mehl, and Niederhoffer (2003), we expected that those assigned to
reappraise would use more change and insight words, reflecting
the reinterpretation used to decrease their anger.
One hundred seventeen female participants aged 18–28 years
(M ? 20.0, SD ? 1.9) were recruited from an introductory psy-
chology course. Participants were paid $40 for their participation
in a larger study, of which this was one element. Participants
self-identified as Caucasian (47.9%), African American (17.1%),
Asian American (24.8%), Latina/Hispanic (9.4%) and Native
American (0.8%). One participant’s data were removed due to her
failure to complete the study.
As in Study 1, participants called to mind recent personally
relevant anger-eliciting events. One important change was that we
then immediately asked them either to ruminate about these events
or to reappraise these events, without an extended period of free
recall. To increase the salience of the objective nature of the
third-person’s perspective and goal structure, we suggested the
perspective of a mediator who is not the advocate for either party
but who wants the best outcome for all. Also, because repeated
focus on a reappraised event might attenuate its negative affective
load, and with it, the personal importance of the event, we short-
ened the trial length to 1.5 min to ensure that reappraisal partici-
pants continued to focus on the event for the same amount of time
as those assigned to ruminate. Finally, we continuously assessed
physiological responding throughout the experiment, and after
each trial, we assessed how angry the participants felt and the
extent to which they were thinking about the anger-making event.
Participants were tested individually. A female experimenter
attached physiological sensors for continuous monitoring of phys-
iological responses. Participants viewed a 5-min neutral film,2
providing a physiological baseline, and completed an emotion
rating form. Participants identified an unresolved angry event and
rated it as they did in Study 1.
Participants were randomly assigned to either a rumination (n ?
61) or reappraisal (n ? 55) condition. In either condition, they
were asked to think about the event that made them angry for 90 s.
In the rumination condition, participants were instructed to “Think
about it (the event) from your own perspective, and turn it over and
over in your mind. Focus on those things that initially made you
feel and respond the way that you did.” In the reappraisal condi-
tion, participants were instructed to “Think about it (the event)
from the perspective of a third party—such as a counselor or
mediator—(someone) who wants the best for all involved. Focus
on how this person might frame the event and find the good that
could come from it.” Participants saw the instructions on the
television screen as they heard them.
After the 90-s instructed recall phase, participants provided anger
ratings and an estimate of the actual time spent thinking about the
anger-eliciting event during the recall phase. Participants were in-
structed to wait until they received the next instructions (90-s delay),
2The neutral film was a 5-min segment from the documentary “Denali,”
depicting nature scenes previously rated by judges as not inducing negative
emotion and inducing only marginal amounts of positive emotion (Rotten-
berg, Ray, & Gross, 2007).
RAY, WILHELM, AND GROSS
followed by another anger rating and an estimate of time spent
thinking about the event during the delay. The participants completed
retrospective questions about the extent to which they thought about
the event from their own perspective and the effortfulness of the task.
The set of tasks (instructed anger recall, delay) were completed two
more times with a short verbal working memory filler task, to mark
iterations, for a total of three repetitions. After the last repetition,
participants were instructed to write a description of the anger-
eliciting event and describe what and how they thought about the
event during the instructed recall and delay periods.
The experience of anger, positive emotion, and negative emo-
The same ratings as those in Study 1 were used. These
positive and negative emotions, not including anger, were used to
create positive (? ? .81) and negative (? ? .77) emotion com-
Cognitive components contributing to reappraisal.
tive ratings of how effortful it was to think about the event in the
instructed manner were collected for each instructed recall trial,
using the same 5-point Likert scale ranging from 0 (not effortful at
all) to 4 (extremely effortful).
Physiological responding was assessed
using an SA 12-channel bioamplifier and a Pentium PC. Measures
used to assess the sympathetic nervous system response included a
central measure, preejection period (PEP), and a peripheral composite
measure (Gross & Levenson, 1997). A measure of motor movement
was included by measuring somatic activity. Respiratory sinus ar-
rhythmia (RSA) evaluated parasympathetic activation and was de-
rived using heart rate and respiration. All physiological measures,
with the exception of the PEP, were based on signals sampled at 400
to perform waveform transformation, feature detection, graphic dis-
play, and artifact removal for each physiological channel and derived
parameter offline (Wilhelm, Grossman, & Roth, 1999). After artifact
control, period averages were created for each measure.
We assessed the PEP using impedance cardiography (HIC-2000,
Instrumentation for Medicine, Inc., Chapel Hill, NC), ensembled
by electrocardiographic (ECG) R-waves (Sherwood et al., 1990).
The PEP was calculated as the time (in milliseconds) between the
Q-point in the ECG signal and the B-point in the derived imped-
ance signal. Four spot electrodes were used in a configuration,
suggested by Sramek (1982), thereby decreasing the effects of any
movement artifacts and increasing the signal-to-noise ratio (Qu,
Zhang, Webster, & Tompkins, 1986).
A second measure of sympathetic activation of the cardiovascular
system used a composite score that included measures indicating
transit times and amplitudes, as well as ear pulse transit time; Gross &
Levenson, 1997). A plethysmograph transducer (Model 1020 photo-
plethysmograph; UFI, Morro Bay, CA) attached to the tip of the
participant’s second finger provided finger pulse amplitude and finger
pulse transit time, whereas a second plethysmograph transducer at-
tached to the ear lobe provided ear pulse transit time. A thermistor
attached to the palmar surface of the distal phalange of the fourth
finger recorded finger temperature in degrees fahrenheit. The com-
posite score was constructed by transforming change scores to z
scores and reversing their polarity so that increased sympathetic
responses would be positive and decreased sympathetic responses
would be negative. The alpha for those measures across tasks was
between .78 and .85.
We measured RSA as an index of parasympathetic activation.
RSA was estimated by converting the respiratory rate (RR) inter-
val series from the ECG to a time series of instantaneous RR
intervals with a resolution of 4 Hz. The transfer function was
determined relating RR interval oscillations to lung volume oscil-
lations (resampled to 4 Hz) at the peak respiratory frequency,
adjusting RSA for tidal volume changes. The peak respiratory
frequency was automatically detected as the greatest local maxi-
mum in the lung volume power spectral density. All of the epochs
met inclusion criteria for spectral coherence being greater or equal
to 0.5 and peak respiratory frequency being above 0.15 Hz. Heart
rate (HR) was obtained using Beckman miniature electrodes
(Beckman Coulter, Inc., Fullerton, CA) placed in a bipolar con-
figuration on opposite sides of the participant’s chest. The interval
(in milliseconds) between successive R-waves in the ECG was
converted to instantaneous HR after editing RR interval outliers
due to movement artifacts or ectopic myocardial activity. RR was
measured using bellow devices placed around the abdomen and
chest. For each individual, signals were calibrated using a fixed
volume bag. RR was calculated by counting the breath-to-breath
cycle at breath onsets per minute.
At the end of the session, each participant wrote a
free-form description of the emotional event during the instructed
anger-recall periods and how they thought about it in both the in-
structed and uninstructed periods. Independent coders rated the seri-
ousness of the initial event recalled. Interrater reliabilities for these
ratings were excellent (? ? .94). This information was used to test
whether randomization was successful. Independent coders rated the
extent to which people reported reinterpreting the event during the
periods of instructed recall. The scale was from “?1,” the participant
expressed a change in interpretation of the event to be more negative,
to “1,” meaning a change in interpretation to be more positive. There
was no change with a rating of “0”. Interrater reliabilities for these
that was thought to reflect a change of thinking when used in a
sentence about one’s thoughts. This list included words such as
different, change, started, stopped, more, less, realized, discovered,
begin, end, and move. Raters then counted the number of insight or
change words that the participants used in their descriptions of how
they thought about the event. Interrater reliabilities for both of these
ratings were good (? ? .89).
For emotion experience and physiological responding, changes
from baseline were calculated. For thoughts of the anger episodes,
raw scores were used. Nested repeated measures ANOVAs were
conducted, with condition (ruminate vs. reappraise) as a between-
subjects factor and time (Block 1–3) as a within-subjects factor,
with task (instructed anger recall and delay) nested within time.3
3Analyses in which hierarchical linear modeling was used produced the
same results. Because there were few missing data, a repeated measures
ANOVA was preferred because it was both conservative and paralleled the
analyses in Study 1.
RUMINATION AND REAPPRAISAL
Significant interactions were followed up with t tests. Greenhouse–
Geisser adjustments for heteroscedasticity were applied, as indi-
cated by degrees of freedom.
The groups did not differ in any aspect of the initial event.4They
also did not differ in how much they thought about the anger-
eliciting event, as evidenced by the repeated measures ANOVA,
with condition (reappraise vs. ruminate) as the between-subjects
factor and time (instructed recall for Time 1–3) as the repeated
within-subjects factor performed on the percentage of time think-
ing about the event. This analysis produced neither a main effect
of condition (MRuminate? 83.78% ? 14.83%, MReappraise?
80.43% ? 17.43%), F(1, 113) ? 1.97, p ? .16, nor a Condition ?
Time interaction, F(2, 226) ? 1.99, p ? .15. The main effect of
time was significant, F(2, 226) ? 17.74, p ? .001, ?p
post hoc contrasts showed that thinking about the event made
participants angry less in the last instructed recall trial than in the
previous two (instructed recall Time 1 and 2 ? 85.20% ? 13.19%,
instructed recall Time 3 ? 76.23% ? 22.52%), F(1, 226) ? 25.51,
p ? .001, ?p
As in Study 1, participants instructed to ruminate reported
viewing the anger-eliciting event more from their own perspective
than those instructed to reappraise. This was demonstrated by a
repeated measures ANOVA, with condition as a between-groups
measure and time (repetition Times 1–3) as the repeated within-
groups measure, performed on the first perspective ratings of
viewing the anger-eliciting event, which produced a main effect
for condition and no interaction with time, F(1, 113) ? 62.78, p ?
2? .14, and
would maintain anger, whereas reappraisal would decrease anger,
a nested repeated measures ANOVA was performed. This analysis
revealed a main effect of condition, such that those assigned to
ruminate reported more anger than those assigned to reappraise at
all time points (ruminate ? 1.44 ? 0.96, reappraise ? 0.87 ?
0.80), F(1, 113) ? 16.80, p ? .001, ?p
well as a significant Time ? Task ? Condition interaction, F(2,
226) ? 3.13, p ? .046, ?p
To test the hypothesis that rumination
2? .13; see Figure 2), as
2? .03 Follow-up nested repeated
measures ANOVAs on anger ratings for each condition indicated
that rumination participants did not show signs of waning anger
over time, as evidenced by the lack of a significant main effect of
time or Task ? Time interaction (p ? .1). By contrast, those
assigned to reappraise experienced a linear decrease in anger over
time during the repeated instructed recall periods. This linear
decline was exhibited by a main effect for time, F(2, 226) ? 18.65,
p ? .001, ?p
effect of time for the instructed anger-recall task, Flinear(2, 226) ?
36.78, p ? .001, ?p
t tests (Time 1–Time 2 ? 0.24 ? 1.01), t(226) ? 1.10, p ? .10,
(Time 1–Time 3 ? 0.36 ? 0.80), t(226) ? 1.88, p ? .03, and a
Time ? Task interaction, F(2, 226) ? 6.36, p ? .002, ?p
Follow-up t tests demonstrated that reappraisal participants re-
ported less anger in the instructed recall periods over time while
actively focusing on and reappraising an anger-eliciting event.
During the instructed recall tasks, participants’ reported anger held
steady from Time 1 to Time 2, t(226) ? 1.21, p ? .113, and then
decreased between Time 2 and Time 3, t(226) ? 2.80, p ? .003.
To test the hypothesis that rumina-
tion would perpetuate thinking about the anger-eliciting event
more than reappraisal, a nested repeated measures ANOVA was
performed. Condition was the between-groups factor, and time
(repetition of reported percentage of time thinking about the anger-
eliciting event during the delay period for Times 1–3) was the
repeated factor, and it revealed a main effect of condition. Rumi-
nation participants continued to think about the event more than
the reappraisal participants during the uninstructed delay periods,
F(1, 113) ? 8.55, p ? .005, ?p
indicated that over time, all participants reported thinking less
exclusively about the event that made them angry during the delay
period, F(2, 226) ? 19.85, p ? .001, ?p
Condition interaction was not significant. Not only did ruminators
continue to think about the event more when they were not
instructed to, they also thought about it in the same way. Indepen-
dent t tests on the amount the participant reported changing her
interpretation of the event shows that those assigned to ruminate
were less likely to change how they thought about the event,
whereas those assigned to reappraise expressed change in their
interpretation of the event (ruminate ? 0.17 ? 0.35, reappraise ?
0.46 ? 0.48), t(87.48) ? 3.52, p ? .001.
differed from reappraisal in terms of physiological responding,
nested repeated measures ANOVAs were performed. As predicted,
rumination participants showed greater physiological activation in
both measures of sympathetic responding than reappraisal partic-
ipants. A nested repeated measures ANOVA performed on the
peripheral sympathetic composite score (change from baseline)
detected a main effect of condition: Those assigned to ruminate
showed greater increases in peripheral sympathetic responding
2? .14, and by the follow-up contrasts to the main
2? .14, as well as suggested by further paired
2? .07. Also, a main effect of time
2? .15. The Time ?
To examine whether rumination
4The events that individuals reported in Study 2 were assessed along
several dimensions to ensure that random assignment was successful in
producing groups with equally anger-provoking events. Participants in the
two conditions did not differ either in the type of event called to mind or
in the intensity of the initial emotional response. These events originally
elicited moderate levels of anger and were not yet resolved. The partici-
pants also did not differ in physiological activation during baseline.
Time 1Time 2Time 3
Change from Baseline
recall and delay trials in Study 2.
Anger experience (change from baseline) in instructed anger-
RAY, WILHELM, AND GROSS
than those assigned to reappraise (ruminate ? 0.99 ? 1.28, reap-
praise ? 0.14 ? 1.38), F(1, 105) ? 12.53, p ? .001, ?p
(see Figure 3A). Condition did not interact with either time or task,
nor was there a three-way interaction of Task ? Condition ?
Time, F(1.71, 179.23) ? 1.06, ns. The nested repeated measures
ANOVA did, however, reveal a Task ? Time interaction, F(1.71,
179.23) ? 30.75, p ? .001, ?p
105) ? 49.48, p ? .001, ?p
increased peripheral sympathetic responding more than the delay,
and a main effect of time, F(1.48, 155.85) ? 45.94, p ? .001,
ing increased over time. Follow-up repeated measures ANOVAs
on peripheral sympathetic composite scores for each task sepa-
rately and follow-up linear contrasts confirmed that peripheral
sympathetic responding increased from baseline linearly over time,
Finstructed anger recall(1.86, 179.23) ? 8.02, p ? .001, ?p
Fdelay(1.67, 179.23) ? 110.08, p ? .001, ?p
Similarly, a nested repeated measures ANOVA performed on
the central sympathetic measure, PEP, revealed a main effect of
condition: Those assigned to ruminate showed larger decreases in
PEP from baseline than those assigned to reappraise (ruminate ?
?4.19 ? 5.59, reappraise ? ?1.95 ? 3.34), F(1, 104) ? 5.74, p ?
2? .23. A main effect of task, F(1,
2? .32, showed that instructed recall
2? .30, suggested that with each iteration, sympathetic respond-
2? .05 (see Figure 3B). A decrease in PEP indicates
increased central sympathetic response, with those assigned to
reappraise experiencing less central sympathetic response than
those assigned to ruminate. Additionally, a main effect of task
suggests that participants’ PEP was shorter during the delay task
than during the instructed anger-recall task (delay ? ?3.29 ?
5.42, anger recall ? ?2.86 ? 5.76), F(1, 104) ? 5.35, p ? .023,
findings demonstrate that rumination produces greater sympathetic
responding than does reappraisal.
Written response of cognitive change.
that reappraisal instructions would encourage reinterpretation of
the anger-eliciting event, participants’ written descriptions of how
they thought about the event were coded for change in interpreta-
tion of the event. A chi-square analysis tested the association
between condition and the three ratings. Those assigned to the
reappraisal group were significantly more likely to reinterpret the
anger-eliciting event more positively, ?2(2) ? 19.23, p ? .001,
ruminate ? .24, reappraise ? .41, whereas those assigned to
ruminate were equally likely to interpret the anger-eliciting event
more negatively (negative ? .23) or more positively (positive ?
.24). Another test for change in interpretations was performed on
the number of change or insight words that participants used when
describing how they thought about the anger-eliciting event. An
independent t test on the number of change or insight words
confirmed that those assigned to reappraise reported more insight
or change words than those assigned to ruminate, t(111) ? 2.43,
p ? .017, ?p
1.00). However, the number of change words does not fully
account for the differences between the conditions in anger reports.
An independent t test on anger reports controlling for the amount
of change words finds that the conditions still differ significantly,
t(111) ? 4.17, p ? .001, ?p
2? .05. As with the composite sympathetic measure, the PEP
To test the hypothesis
2? .05 (Mruminate? 0.28 ? 0.90, Mreappraise? 0.72 ?
cognitive effort, and this might translate into increased sympa-
thetic responses. To confirm that the sympathetic differences were
not due to differences in effort, a repeated measures ANOVA was
performed on retrospective ratings of effort after each cycle of
instructed recall and delay. The repeated measures ANOVA re-
vealed a main effect of time, F(2, 226) ? 5.89, p ? .005, ?p
.05, whereas the main effect of condition and interaction with
condition were not significant, FCondition(1, 113) ? 0.004, p ?
.985; FTime ? Condition(2, 226) ? 1.09, p ? .244. Moreover,
repeated measures ANOVAs on both the central and peripheral
sympathetic composite controlling for effort ratings did not sub-
stantially change the main effect for condition: central sympathetic
responding, F(1, 103) ? 4.34, p ? .032, ?p
103) ? 5.22, p ? .024, ?p
ing, F(1, 104) ? 12.45, p ? .001, ?p
12.64, p ? .001, ?p
autonomic nervous system can be influenced by the parasympa-
thetic branch (Berntson, Cacioppo, & Quigley, 1991) and by
somatic activity. Consequently, secondary analyses examined (a)
the impact of rumination and reappraisal on the parasympathetic
branch of the autonomic nervous system and (b) the role of
somatic activity in determining physiological responding. A re-
Rumination and reappraisal both require
2? .04, from F(1,
2? .05; peripheral sympathetic respond-
2? .11, from F(1, 105) ?
The sympathetic branch of the
Standard Units Change from Baseline
Time 2Time 3
Milliseconds (Change from Baseline)
in instructed anger-recall and delay periods in Study 2, scaled so that
greater activation is up. B: Preejection period (change from baseline) in
instructed anger-recall and delay periods in Study 2, scaled so that greater
activation is up.
A: Peripheral sympathetic responding (change from baseline)
RUMINATION AND REAPPRAISAL
peated measures ANOVA using RSA scores failed to detect a
significant difference between those assigned to ruminate and
those assigned to reappraise, F(1, 100) ? 1.43, p ? .232; however,
the means were in the direction of those assigned to ruminate
experiencing smaller increases in RSA (M ? 0.001 ? 0.54) than
those assigned to reappraise (M ? 0.094 ? 0.59). These null
findings suggest that in this task, sympathetic and parasympathetic
activity is not tightly coupled, and consequently, each is differently
affected by rumination and reappraisal (Berntson et al., 1991).
Study 2 is the first to demonstrate affective, cognitive, and
physiological differences between rumination and reappraisal.
These differences were found not only during the task but also
during the recovery period. This study provides strong evidence
that reappraisal can contribute to emotional, cognitive, and phys-
iological recovery from anger-inducing events.
As in Study 1, Study 2 showed that those instructed to ruminate
reported more anger than those instructed to reappraise despite
spending the same amount of time thinking about the anger-
inducing event. Reappraising allowed individuals to think about
the event without experiencing the intensity of anger that is a
consequence of rumination. One mechanism suggested by our
results is the increased insight reported by those who reappraised.
Study 2 also demonstrated the differential effects of rumination
and reappraisal over time. Those assigned to ruminate about their
anger-eliciting event maintained their anger each time they
brought the event to mind, whereas those assigned to reappraise
showed decreases in anger with each repetition. This finding,
together with the finding from Study 1 that free recall produces
levels of anger indistinguishable from ruminating, provides strong
evidence that reappraisal works to attenuate anger. In addition, by
using delay instead of “rest” periods, Study 2 revealed that, com-
pared with reappraisal, rumination perpetuates anger and perse-
verative thinking about the event, even after participants had been
told they could stop.
Study 2 showed that rumination and reappraisal differed not
only in self-reported emotion but also in physiological responding.
Those assigned to ruminate had greater central and peripheral
sympathetic responses than those assigned to reappraise, both
during the instructed anger-recall periods and in the delay periods.
This difference between the groups increased over time, with those
assigned to ruminate maintaining or increasing their level of acti-
vation and those assigned to reappraise decreasing their activation.
These physiological findings may be interpreted as evidence
that ruminators continue to refresh the arousing properties of
events that make them angry, whereas reappraisers were able over
time to cognitively transform the event so that their physiology
could recover to baseline. Furthermore, increases in sympathetic
responding for ruminators is consistent with other findings of
decreased autonomic recovery following rumination (Glynn et al.,
2002; Vickers & Vogeltanz-Holm, 2003). This is also consistent
with the association between hostility and increases in measures of
central PEP sympathetic activation (Neumann, Waldstein, Sollers,
Thayer, & Sorkin, 2004). This pattern in the anger rumination
components (anger experience, thoughts about the angry event,
and increased central cardiovascular sympathetic activation) sup-
ports the notion that chronic anger rumination creates a vulnera-
bility for cardiovascular disease (Brosschot et al., 2006).
According to appraisal theories of emotion, the key determinant
of emotional responses is the way one thinks about (or appraises)
an event (Lazarus, 1991; Scherer et al., 2001). Anger-eliciting
situations provide an ideal context within which to investigate the
relationship between thinking and feeling. In daily life, we en-
counter many opportunities to have our plans and hopes thwarted,
by accident or by the incompetence or even malice of others. Some
people, like Emily in the van pool, ruminate over these events,
mentally replaying upsetting appraisals and focusing on how it
makes them feel. They are emotionally the worse for it (Morrow &
Nolen-Hoeksema, 1990). Others, like Bob, manage mentally to
review and focus on upsetting events in ways that help them
integrate these experiences, leaving them emotionally recovered
from the event (Pennebaker et al., 2003).
Rumination and Reappraisal Have Different
In the present studies, we contrasted rumination and reappraisal
by having participants either ruminate about or reappraise recent,
autobiographical, unresolved anger-eliciting events. Despite
spending equal amounts of time thinking about the events, those
who ruminated about the event (vs. those who reappraised) felt
angrier when they recalled the event, and they were more likely to
continue to think about the event (and feel angrier) even when told
they could stop.
In line with their increased anger and preoccupation with the
anger-eliciting event, those who ruminated showed greater central
and peripheral sympathetic responses, a difference that actually
increased over time. These effects between groups were not due to
increases in positive emotion on the part of reappraisers, increases
in social distance, or differential effort. This indicates that rumi-
nation leads to more sympathetic nervous system activation, with
little or no habituation to the angry thoughts. Reappraisal, how-
ever, makes further thinking about the anger-eliciting event less
likely and leads to decreases in self-reported anger and central and
peripheral sympathetic activation, as well as increases in reports of
insight and positive reinterpretation, suggesting that reappraisal
makes upsetting material less arousing over time.
Implications for Emotion Regulation Processes
The findings presented here address a crucial gap in the litera-
ture on rumination by showing that even when one holds relatively
constant what a person is thinking about, the way a person thinks
about the event shapes his or her emotional responses. The present
studies provide converging evidence for the work of Ayduk,
Mischel, and Downey (2002), which has shown that cognitive
representations of an event influence emotion experience reports.
These studies are the first to show that changing how one thinks
about life events can penetrate not only self-report but also phys-
iological responding. It thus serves as a crucial link in understand-
ing how different approaches to life events may have profound
consequences for physical health.
RAY, WILHELM, AND GROSS
There have been growing calls for a distinction between differ-
ent ways of bringing upsetting content to mind (Trapnell & Camp-
bell, 1999; Treynor, Gonzales, & Nolen-Hoeksema, 2003). Several
of the possible alternatives to rumination have not focused on how
one thinks about the particulars of an upsetting event, but instead
operate on a level of abstraction above the emotion cognition
interaction so as to drain events of their meaningfulness. For
example, decentering (Watkins, Teasdale, & Williams, 2000) by-
passes the situated appraisals causing a mood and works instead to
recast one’s mood or emotions in light of the trans-situational
transience of any moment in one’s life. Similarly, diffusion (Ma-
suda, Hayes, Sackett, & Twohig, 2004) does not attempt to change
one’s thoughts about events, but instead attempts to change the
metacognitive environment, repeating thoughts until they lose their
meaning, so as to dissociate the part of the self that is experiencing
events and emotions from the part that is evaluating the event. By
clear contrast, reappraisal directly changes one’s appraisals so as to
change one’s affective response. The present study provides evi-
dence that reappraisal can be effective in decreasing the intensity
of one’s negative affect.
Another strategy that has been proposed in the context of anger
is called “constructive anger verbal behavior” (Davidson, MacGre-
gor, Stuhr, Dixon, & MacLean, 2000). This strategy involves
interacting with another person to explore more nuanced causes for
the problem and its solutions in such a way as to lead to decreased
negative emotion. It is assessed by questionnaire, and use of
constructive anger verbal behavior has been associated with lower
resting blood pressure (Davidson et al., 2000). We suspect that
constructive anger verbal behavior is a social species of reappraisal
because items in the questionnaire index the extent to which
discussion increases one’s appreciation for the complexity of sit-
uations and thus may lead to different interpretations. Reappraisal
generally, and constructive anger verbal behavior specifically, can
create opportunities to integrate upsetting material and produc-
tively solve problems by maintaining a problem-solving atten-
tional focus on the upsetting material while decreasing the inten-
sity of the negative affect.
The present findings extend prior reappraisal research in the
emerging field of emotion regulation (Gross, 1998, 2007). Re-
searchers interested in reappraisal have directed their efforts at
understanding how reappraisal changes physiological aspects of
the emotional response. These studies, however, generally have
been limited to standardized stimuli. The present study is the first
to show that reappraisal can lessen anger experience and sympa-
thetic activation in the ecologically central context of powerful,
socially complex, and personally relevant anger episodes that are
not yet resolved. Additionally, this study extends the finding by
Gross and John (2003) that those who tend to reappraise experi-
ence better outcomes during stressful events by showing that such
individuals are able to decrease the affective, cognitive, and phys-
iological effects of negative emotional events.
Limitations and Future Directions
The present studies were limited in comparing rumination
with one type of cognitive reappraisal in women who were
angry about an unresolved interpersonal event. Although rumi-
nation about interpersonal events is more common in women
(Mezulis, Abramson, & Hyde, 2002), much of the work has
been done on depressive rumination rather than on angry rumi-
nation, with women reporting using a wider range of coping
strategies such as social support to deal with their anger (Linden
et al., 2003). However, work by Rusting and Nolen-Hoeksema
(1998) suggests that men may engage in anger rumination more
often than women. Future studies should thus examine the
differential effects of rumination and reappraisal in men as well
as women, in the context of other negative and positive emo-
tional states, and in the context of other, equally ecologically
generalizable emotion elicitors.
Our focus here has been the immediate consequences of rumi-
nation and reappraisal on emotion, cognition, and physiology, not
on functional outcomes for problem solving and interpersonal
relationships. However, based on the present findings as well as on
the prior literature on anger rumination and aggression and marital
satisfaction, we predict that anger rumination leads to increases in
hostile interpersonal interactions and decreases in productive prob-
lem solving (Bushman et al., 2005). By contrast, reappraisal, in
addition to decreasing anger experience, has been associated with
positive social outcomes (Gross & John, 2003) and, compared with
expressive suppression, improved memory for emotionally tense
discussions with dating partners (Richards, Butler, & Gross, 2003).
Distraction has been associated with decreased anger reports
(Bushman et al., 2005; Rusting & Nolen Hoeksema, 1998) and
better problem solving than rumination (Lyubomirsky, Tucker,
Caldwell, & Berg, 1999). Consequently, we expect that the kind of
cognitive reappraisal in which our participants engaged will lead to
better interpersonal interactions than rumination or distraction and
will better facilitate problem solving and resolution than rumina-
tion. We regard this as an important direction for research.
Finally, work by Pennebaker and colleagues (2003) suggests
that differences in the way one cognitively processes upsetting
events can have longer term health consequences. We captured,
in miniature within the laboratory setting, aspects of the tem-
porally extended, repetitive nature of rumination. Like any
laboratory study, ours is limited in its ability to capture longer
and more naturalistic rumination and reappraisal over time.
Nonetheless, the pattern of emotional, cognitive, and physio-
logical effects of anger rumination demonstrated in these stud-
ies corroborates theorizing by Brosschot, Gerin, and Thayer
(2006) that rehearsal and amplification of distress over the long
term may create vulnerabilities to cardiovascular disease. For
example, chronic anger rumination is associated with hostility,
which, in turn, has been associated with poorer survival in
women with suspected coronary heart disease (Olson et al.,
2005). Future research is needed to demonstrate how acute
increases in sympathetic activation seen in anger rumination
manifests into increases in vulnerability to cardiovascular dis-
Albeit speculative, the present studies are consistent with the
idea that reappraisal may be a useful tool to mitigate some of the
negative effects on emotional and physical health caused by ru-
minating on anger-inducing events. More generally, our results
suggest the importance of investigating, over long periods of time,
the cumulative effects on psychological, relational, and physical
health of different primary and/or chronic strategies for thinking
about life’s problems.
RUMINATION AND REAPPRAISAL
Arnold, M. (1960). Emotion and personality. New York: Columbia Uni-
Ayduk, O., Mischel, W., & Downey, G. (2002). Attentional mechanisms
linking rejection to hostile reactivity: The role of “hot” versus “cool”
focus. Psychological Science, 13, 443–448.
Berntson, G. G., Cacioppo, J. T., & Quigley, K. S. (1991). Autonomic
determinism: The modes of autonomic control, the doctrine of auto-
nomic space, and the laws of autonomic constraint. Psychological Re-
view, 98, 459–487.
Bradley, M. M. (2000). Emotion and motivation. In J. T. Cacioppo, L. G.
Tassinary, & G. G. Berntson (Eds.), Handbook of psychophysiology (2nd
ed., pp. 602–642). Cambridge, England: Cambridge University Press.
Brosschot, J. F., Gerin, W., & Thayer, J. F. (2006). The perseverative
cognition hypothesis: A review of worry, prolonged stress-related phys-
iological activation and health. Journal of Psychosomatic Research, 60,
Bushman, B. J., Bonacci, A. M., Pedersen, W. C., Vasquez, E. A., &
Miller, N. (2005). Chewing on it can chew you up: Effects of rumination
on triggered displaced aggression. Journal of Personality and Social
Psychology, 88, 969–983.
Cacioppo, J. T., Berntson, G. C., Larsen, J. T., Poehlmann, K. M., & Ito,
T. A. (2000). The psychophysiology of emotion. In M. Lewis & J. M.
Haviland-Jones (Eds.), Handbook of emotions (2nd ed., pp. 173–191).
New York: Guilford Press.
Clark, D. A., Beck, A. T., & Alford, B. A. (1999). Scientific foundations of
cognitive theory and therapy of depression. Hoboken, NJ: Wiley.
Davidson, K., MacGregor, M. W., Stuhr, J., Dixon, K., & MacLean, D.
(2000). Constructive anger verbal behavior predicts blood pressure in a
population-based sample. Health Psychology, 19, 55–64.
Garnefski, N., & Kraaij, V. (2006). Relationships between cognitive emo-
tion regulation strategies and depressive symptoms: A comparative
study of five specific samples. Personality and Individual Differences,
Glynn, L. M., Christenfeld, N., & Gerin, W. (2002). The role of rumination
in recovery from reactivity: Cardiovascular consequences of emotional
states. Psychosomatic Medicine, 64, 714–726.
Gross, J. J. (1998). Antecedent- and response-focused emotion regulation:
Divergent consequences for experience, expression, and physiology.
Journal of Personality and Social Psychology, 74, 224–237.
Gross, J. J. (2001). Emotion regulation in adulthood: Timing is everything.
Current Directions in Psychological Science, 10, 214–219.
Gross, J. J. (Ed.). (2007). Handbook of emotion regulation. New York:
Gross, J. J., & John, O. P. (2003). Individual differences in two emotion
regulation processes: Implications for affect, relationships, and well-
being. Journal of Personality and Social Psychology, 85, 348–362.
Gross, J. J., & Levenson, R. W. (1997). Hiding feelings: The acute effects
of inhibiting negative and positive emotion. Journal of Abnormal Psy-
chology, 106, 95–103.
Jackson, D. C., Malmstadt, J. R., Larson, C. L., & Davidson, R. J. (2000).
Suppression and enhancement of emotional responses to negative pic-
tures. Psychophysiology, 37, 515–522.
Just, N., & Alloy, L. B. (1997). The response styles theory of depression:
Tests and an extension of the theory. Journal of Abnormal Psychology,
Keselman, H. J., Algina, J., & Kowalchuk, R. K. (2001). The analysis of
repeated measures designs: A review. British Journal of Mathematical
and Statistical Psychology, 54, 1–20.
Kross, E., Ayduk, O., & Mischel, W. (2005). When asking “why” does not
hurt. Psychological Science, 16, 709–715.
Lazarus, R. S. (1991). Cognition and motivation in emotion. American
Psychologist, 46, 352–367.
Lazarus, R. S., & Alfert, E. (1964). The short-circuiting of threat. Journal
of Abnormal and Social Psychology, 69, 195–205.
Libby, L. K., & Eibach, R. P. (2002). Looking back in time: Self-concept
change affects visual perspective in autobiographical memory. Journal
of Personality and Social Psychology, 82, 167–179.
Linden, W., Hogan, B. E., Rutledge, T., Chawla, A., Lenz, J. W., & Leung,
D. (2003). There is more to anger coping than “in” or “out.” Emotion, 3,
Lyubomirsky, S., Tucker, K. L., Caldwell, N. D., & Berg, K. (1999). Why
ruminators are poor problem solvers: Clues from the phenomenology of
dysphoric rumination. Journal of Personality and Social Psychology, 77,
Martin, L. L., & Tesser, A. (1996). Some ruminative thoughts. In R. S.
Wyer, Jr. (Ed.), Ruminative thoughts: Advances in social cognition (pp.
1–47). Hillsdale, NJ: Erlbaum.
Masuda, A., Hayes, S. C., Sackett, C. F., & Twohig, M. P. (2004).
Cognitive defusion of self-relevant negative thoughts: Examining the
impact of a ninety year old technique. Behaviour Research and Therapy,
Mezulis, A. H., Abramson, L. Y., & Hyde, J. S. (2002). Domain specificity
of gender differences in rumination. Journal of Cognitive Psychother-
apy: An International Quarterly, 16, 421–434.
Morrow, J., & Nolen-Hoeksema, S. (1990). Effects of responses to depres-
sion on the remediation of depressive affect. Journal of Personality and
Social Psychology, 58, 519–527.
Neumann, S. A., Waldstein, S. R., Sollers, J. J., Thayer, J. F., & Sorkin,
J. D. (2004). The relation of hostility, rumination and distraction to
cardiovascular reactivity and recovery responses to anger. Annals of
Behavioral Medicine, 43, 140.
Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders
and mixed anxiety/depressive symptoms. Journal of Abnormal Psychol-
ogy, 109, 504–511.
Nolen-Hoeksema, S., & Morrow, J. (1991). A prospective study of depres-
sion and posttraumatic stress symptoms after a natural disaster: The
1989 Loma Prieta earthquake. Journal of Personality and Social Psy-
chology, 61, 115–121.
Nolen-Hoeksema, S., Morrow, J., & Fredrickson, B. L. (1993). Response
styles and the duration of episodes of depressed mood. Journal of
Abnormal Psychology, 102, 20–28.
Ochsner, K. N., Bunge, S. A., Gross, J. J., & Gabrieli, J. D. E. (2002).
Rethinking feelings: An fMRI study of the cognitive regulation of
emotion. Journal of Cognitive Neuroscience, 14, 1215–1229.
Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S.,
Gabrieli, J. D. E., & Gross, J. J. (2004). For better or for worse: Neural
systems supporting the down- and up-regulation of negative emotion.
NeuroImage, 23, 483–499.
Olson, M. B., Krantz, D. S., Kelsey, S. F., Pepine, C. J., Sopko, G.,
Handberg, E., et al. (2005). Hostility scores are associated with increased
risk of cardiovascular events in women undergoing coronary angiogra-
phy: A report from the NHLBI-sponsored WISE study. Psychosomatic
Medicine, 67, 546–552.
Pennebaker, J. W., Mehl, M. R., & Niederhoffer, K. G. (2003). Psycho-
logical aspects of natural language use: Our words, our selves. Annual
Review of Psychology, 54, 547–577.
Qu, M. H., Zhang, Y. J., Webster, J. G., & Tompkins, W. J. (1986). Motion
artifact from spot and band electrodes during impedance cardiography.
IEEE Transactions on Biomedical Engineering, 33, 1029–1036.
Ray, R. D., Ochsner, K. N., Cooper, J. C., Robertson, E. R., Gabrieli,
J. D. E., & Gross, J. J. (2005). Individual differences in trait rumination
and the neural systems supporting cognitive reappraisal. Cognitive,
Affective, & Behavioral Neuroscience, 5, 156–168.
Richards, J. M., Butler, E. A., & Gross, J. J. (2003). Emotion regulation in
romantic relationships: The cognitive consequences of concealing feel-
ings. Journal of Social and Personal Relationships, 20, 599–620.
RAY, WILHELM, AND GROSS
Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The
cognitive consequences of emotion suppression. Personality and Social
Psychology Bulletin, 25, 1033–1044.
Roger, D., & Jamieson, J. (1988). Individual differences in delayed heart-
rate recovery following stress: The role of extraversion, neuroticism, and
emotional control. Personality and Individual Differences, 9, 721–726.
Rottenberg, J., Ray, R. D., & Gross, J. J. (2007). Emotion elicitation using
films. In J. A. Coan & J. J. B. Allen (Eds.), The handbook of emotion
elicitation and assessment (pp. 9–28). New York: Oxford University
Rottenberg, J., Salomon, K., Gross, J. J., & Gotlib, I. H. (2005). Vagal
withdrawal to a sad film predicts subsequent recovery from depression.
Psychophysiology, 42, 277–281.
Rusting, C. L., & Nolen-Hoeksema, S. (1998). Regulating responses to
anger: Effects of rumination and distraction on angry mood. Journal of
Personality and Social Psychology, 74, 790–803.
Schachter, S., & Singer, J. (1962). Cognitive, social, and physiological
determinants of emotional state. Psychological Review, 69, 379–399.
Scherer, K. R., Schorr, A., & Johnstone, T. (2001). Appraisal processes in
emotion. New York: Oxford University Press.
Segerstrom, S. C., Tsao, J. C. I., Alden, L. E., & Craske, M. G. (2000).
Worry and rumination: Repetitive thought as a concomitant and predic-
tor of negative mood. Cognitive Therapy and Research, 24, 671–688.
Sherwood, A., Allen, M. T., Fahrenberg, J., Kelsey, R. M., Lovallo, W. R.,
& van Doornen, L. J. (1990). Methodological guidelines for impedance
cardiography. Psychophysiology, 27, 1–23.
Siegle, G. J., Steinhauer, S. R., Thase, M. E., Stenger, A., & Carter, C. S.
(2002). Can’t shake that feeling: An event-related fMRI assessment of
sustained amygdala activity in response to emotional information in
depressed individuals. Biological Psychiatry, 51, 693–707.
Sramek, B. B. (1982). Cardiac output by electrical impedance. Medical
Electronics, 13, 93–97.
Stemmler, G., Heldmann, M., Pauls, C. A., & Scherer, T. (2001). Con-
straints for emotion specificity in fear and anger: The context counts.
Psychophysiology, 38, 275–291.
Trapnell, P. D., & Campbell, J. D. (1999). Private self-consciousness and
the five-factor model of personality: Distinguishing rumination from
reflection. Journal of Personality and Social Psychology, 76, 284–304.
Treynor, W., Gonzales, R., & Nolen-Hoeksema, S. (2003). Rumination
reconsidered: A psychometric analysis. Cognitive Therapy and Re-
search, 27, 247–259.
Vickers, K. S., & Vogeltanz-Holm, N. D. (2003). The effects of rumination
and distraction tasks on psychophysiological responses and mood of
dysphoric and nondysphoric individuals. Cognitive Therapy and Re-
search, 27, 331–348.
Watkins, E. (2004). Appraisals and strategies associated with rumination
and worry. Personality and Individual Differences, 37, 679–694.
Watkins, E., Teasdale, J. D., & Williams, R. M. (2000). Decentering and
distraction reduce overgeneral autobiographical memory in depression.
Psychological Medicine, 30, 911–920.
Wilhelm, F. H., Grossman, P., & Roth, W. T. (1999). Analysis of cardio-
vascular regulation. Biomedical Sciences Instrumentation, 35, 135–140.
Received April 25, 2006
Revision received July 18, 2007
Accepted July 18, 2007 ?
RUMINATION AND REAPPRAISAL