Psychology and Health, 2002, Vol. 17, No. 5, pp. 547–560
MENDING BROKEN HEARTS: EFFECTS OF
EXPRESSIVE WRITING ON MOOD, COGNITIVE
PROCESSING, SOCIAL ADJUSTMENT
AND HEALTH FOLLOWING
A RELATIONSHIP BREAKUP
STEPHEN J. LEPORE
* and MELANIE A. GREENBERG
Department of Health and Behavior Studies, Teachers College, Columbia University,
Thorndike Hall, Box 114, 525 West 120th Street, New York, NY 10027, USA
Clinical Psychology, Alliant International University, 10455 Pomerado Rd., San Diego,
CA 92131, USA
Seventy-two male and 73 female undergraduates were randomly assigned to an experimental group, in which
they wrote expressively about a relationship breakup, or to a control group, in which they wrote in a non-
emotional manner about impersonal relationship topics. Control participants reported short-term increases
in upper respiratory illness (URI) symptoms, tension and fatigue, whereas experimental participants did
not. Further, higher levels of intrusive thoughts and avoidance were associated with short-term increases in
URI symptoms in the control group, but were unrelated to URI symptoms in the experimental group.
Finally, there was a trend ( p<0.06) suggesting that experimental participants were more likely to reunite
with their ex-partner than were control participants. These findings indicate that expressive writing has a
wide range of social, emotional, and physical health benefits for individuals coping with stressful events, par-
ticularly if they are experiencing ongoing intrusive thoughts and avoidance responses related to the stressor.
Keywords: Expressive writing; Emotional expression; Intrusive thoughts; Avoidance; Social adjustment;
Upper respiratory illness
The ability to freely express stress-related thoughts and feelings appears to reduce the
negative mental and physical health effects of stressful life events (Smyth, 1998).
While individuals often choose to disclose to significant others (Rime
´, 1995), for vari-
ous reasons this is not always feasible or advised (Lepore et al., 1996). Non-social
modes of expression, such as expressive writing, may be particularly useful for individ-
uals who feel constrained in disclosing to members of their social network. Expressive
writing also provides a convenient method for individuals to confront and work
through unresolved feelings and thoughts related to stressful events. In this article,
*Corresponding author. E-mail: email@example.com
ISSN 0887-0446 print: ISSN 1476-8321 online ß2002 Taylor & Francis Ltd
we examine the mental and physical health benefits of expressive writing, and we
attempt to identify some mechanisms of action to advance theorizing and research in
Pennebaker and Beall (1986) have developed a brief written emotional expression
intervention that appears to help individuals to cognitively confront and process
their reactions to stressful life events. Participants write about their deepest thoughts
and feelings associated with a specified or self-selected stressful event for 20 min per
day, typically for a 3-day period. Over a decade of research suggests the efficacy of
this intervention in improving physical health and psychosocial adjustment. For
instance, written expression interventions have reduced physician visits for illness
(Pennebaker and Beall, 1986; Pennebaker et al., 1990; Greenberg et al., 1996; King
2000), reduced self-reports of illness symptoms (Pennebaker and Beall, 1986;
Greenberg and Stone, 1992), reduced levels of negative mood and post-traumatic
stress symptoms (Donnelly and Murray, 1991; Lepore, 1997; Lange et al., 2000),
enhanced immune functioning (Pennebaker et al., 1988; Petrie et al., 1995), and
enhanced role and physical functioning (Spera et al., 1994; Cameron and Nicholls,
1998; Smyth et al., 1999).
Early theoretical work suggested that expressive writing produced physiological
release, thus reducing health risks and physiological strain associated with prolonged
inhibition of the desire to disclose (Pennebaker, 1989). According to this approach,
past traumas, especially those involving shame or social stigma, are the most appropri-
ate subject for disclosure because they are most likely to be inhibited. More recently,
cognitive interpretations of the effects of disclosure have gained theoretical prominence
(Tait and Silver, 1989; Clark, 1993; Pennebaker, 1995; Greenberg et al., 1996; Lepore
et al., 1996; Lepore, 1997; Kliewer et al., 1998; Lepore and Helgeson, 1998;
Kennedy-Moore and Watson, 1999; Lutgendorf and Antoni, 1999; Major and
Gramzow, 1999; Lepore et al., 2000).
According to cognitive models of disclosure, stressful events contain novel informa-
tion that is difficult to reconcile with prior assumptions about the self and the world.
Thus, these events are initially stored in active, short-term memory as cognitively dispa-
rate fragments. Because active memory has a tendency to repeat its contents, distressing
trauma-related thoughts, images, dreams, or feelings intrude on awareness (Horowitz,
1986). Intrusions result from the uncomfortable discrepancy between the trauma and
existing schemas, leading individuals to deny, avoid, or suppress these responses.
Cognitive processing of the event involves alternating between intrusive and avoidant
reactions, gradually reappraising the event or modifying extant schemas to reduce the
discrepancy. Some individuals, however, are unable to make this cognitive transforma-
tion. Their continued awareness of cognitive discrepancy leads to prolonged, intense
intrusive states and/or rigid dysfunctional avoidance. These intrusive thoughts can be
an ongoing source of internal stress (Baum et al., 1993), which may be implicated in
mental and physical health problems. Investigators also have suggested that avoidance
can contribute to unfavorable health consequences (e.g., Lepore and Helgeson, 1998).
Pennebaker and colleagues (Pennebaker, 1989; Pennebaker et al., 1990) argue that
expressive writing can facilitate cognitive processing by changing the meaning or signif-
icance of the trauma to make it more consistent with existing self- and world-views.
Consistent with this perspective, Pennebaker et al. (1997) found that an increase in
the use of causal and insight words over writing sessions predicted subsequent health
benefits. Writing also might facilitate organizing the event into a coherent narrative
548 S.J. LEPORE AND M.A. GREENBERG
or story, which allows new perspectives, problem definitions, or coping strategies to
emerge (Smyth and Greenberg, 2000).
Adopting a cognitive processing perspective naturally leads to a focus on recent
events for which adjustment is ongoing. If one can intervene early in the cycle and
help individuals to cognitively process events as they are unfolding, this should prevent
further emotional or physiological damage later on. This can help to prevent a potential
negative cycle in which intrusions both provoke and are triggered by chronic physiolo-
gical arousal and emotional distress (McFarlane, 1992). Memories and emotional reac-
tions relating to recent events should be particularly salient and easily accessible, thus
lending themselves to cognitive modification. Furthermore, reappraisals of the stressor
may lead to new coping strategies or adaptive behaviors. With ongoing events, there is
an opportunity to initiate actions that resolve the problem. Thus, expressive writing
may be especially powerful for these types of events.
The goal of this study is to evaluate the impact of expressive writing on adjustment to
a recent event, the breakup of a romantic relationship. For young adults, such relation-
ships can be an important aspect of identity and a source of intimacy, social status, and
emotional security. The breakup of a relationship may involve dealing with rejection,
loneliness, or guilt, and can result in emotional distress and grief responses
(Kaczmarek et al., 1990). Loss of romantic relationships may be a good arena for
exploring the cognitive effects of expressive writing because research shows a relation-
ship between cognitive variables (e.g., attachment styles, cognitive self-complexity,
mood regulation expectancies) and distress reactions to these events (Smith and
Cohen, 1993; Sprecher et al., 1998).
The major focus of this study was on how expressive writing affects cognitive
processing. Pennebaker (1989) argues that expressive writing brings about cognitive
assimilation, organization, and new perspectives, thus reducing the frequency of
intrusions and avoidance. Lepore (1997), on the other hand, suggests that expressive
writing reduces the negative emotional and physiological impact of intrusive thoughts.
This argument is based on evidence that stressor-related intrusive thoughts lose their
emotional sting when individuals are able to express themselves in a supportive and
non-constraining social context (Lepore et al., 1996; Lepore, 1997; Kliewer et al.,
1998; Lepore and Helgeson, 1998; Major and Gramzow, 1999; Manne, 1999; Lepore,
2001). In a similar vein, Greenberg and colleagues (1996) suggest that expressive writing
may promote affective regulation by increasing individuals’ tolerance for negative
affect, enhancing their perceptions of control and self-efficacy over negative affect,
and promoting self-empathy and acceptance of their own emotional reactions.
Others, too, have argued that expressive writing might promote habituation to stress-
related stimuli (Bootzin, 1997).
Whether writing actually reduces the quantity of intrusions and avoidance in individ-
uals experiencing ongoing stressors is still at issue. In one expressive writing study,
Greenberg et al. (1996) found increased avoidance in a trauma group, relative to con-
trols, but interpretation is complicated because traumas were diverse and only one writ-
ing session was used. In Lepore’s (1997) study, expressive writing did not reduce the
frequency of intrusions over time. However, writing condition moderated the effects
of intrusions, such that high intrusions predicted increases in depressive symptoms in
the control group, but not in the expressive writing group. This is consistent with the
desensitizing or blunting effect of writing on intrusive thoughts. Participants in
Lepore’s study wrote about their anticipatory reactions to an upcoming examination.
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 549
It is not clear if the same processes would occur in relation to a recent stressor involving
loss, rather than challenge. In the present study, using relationship breakup as the stres-
sor, we assessed intrusive thoughts and avoidance at 3 weeks and 15 weeks following
expressive writing and examined intrusions and avoidance as both dependent variables
and moderators to obtain a comprehensive assessment of this issue.
An implication of the cognitive processing perspective is that expressive writing
should change people’s views of the breakup to be more consistent with prior positive
schemas of self and world. For example, people may reevaluate a breakup as providing
opportunities for growth and meaning making (Greenberg, 1995). The relatively few
studies that have assessed the impact of expressive writing on appraisals of a stressor
have demonstrated beneficial cognitive and emotional changes, such as new perspec-
tives on the stressor, increases in positive feelings about the stressor, and increases in
self-esteem and adaptive behavior (Donnelly and Murray, 1991; Murray et al., 1989).
In these studies, cognitive and affective appraisals were assessed short-term either via
content analysis of the essays or immediate post-experimental questionnaires. Thus,
it is unclear if expressive writing has more enduring effects on stressor-related cogni-
tions and emotions. Further, participants in these studies wrote about diverse events,
which precluded a more detailed examination of the specific types of cognitive changes.
The current study adds to the research on cognitive and emotional effects of written
disclosure by assessing participants’ moods and their feelings and attitudes towards
their ex-partner. We examined several specific moods, including negative moods (e.g.,
anger, depression) and physiological arousal/activation (e.g., tension, fatigue). We pre-
dicted that expressive writing would improve negative mood and reduce tension and
fatigue. Based on previous work (e.g., Lepore, 1997), we expected that individuals evi-
dencing incomplete cognitive processing (i.e., high intrusions and avoidance) would
experience the greatest mood benefits through expressive writing. We also examined
perceptions of resent toward the ex-partner, positive feelings toward the ex-partner,
and guilt over the breakup. We predicted that expressive writing would lead to positive
reappraisal of one’s role in the breakup (i.e., less guilt), its impact on one’s life (i.e.,
reduced resent) and feelings toward the ex-partner (e.g., restoration of positive regard).
We also predicted that expressive writing would facilitate social adjustment. Once
individuals express and deal with their emotional reactions, this should free up cogni-
tive resources and emotional energy, which they could devote to planning and imple-
menting restoration activities (e.g., finding other relationships, getting the ex-partner
back). In Spera et al.’s (1994) study, unemployed professionals who were able to pro-
cess their anger about the layoff through writing were more likely to be reemployed
at follow-up, relative to non-writing controls. Freshmen who processed their feelings
about adjusting to college via written disclosure had higher grade point averages
than controls at the end of the semester (Pennebaker et al., 1990, Cameron and
Nicholls, 1998). In addition, writing may help some individuals to see more clearly
their own contribution to relationship problems or to have more empathy for their
ex-partner’s perspective, which could lead to a change of heart and a resumption of
The current randomized, prospective study sought to replicate the beneficial health
effects of written disclosure found in previous studies by comparing the effects of writ-
ing about relationship breakups to writing about non-emotional relationship topics. We
assessed physical health effects using a reliable, validated measure of upper respiratory
symptoms. Based on previous research (e.g., Pennebaker et al., 1990), we predicted that
550 S.J. LEPORE AND M.A. GREENBERG
expressive writing would reduce individuals’ risk for developing upper respiratory
symptoms. As with the mood outcomes, we expected that individuals evidencing incom-
plete cognitive processing would experience the greatest health benefits from expressive
We identified eligible participants via a questionnaire administered in introductory-
level psychology courses. We recruited all students who indicated that they had a
breakup in the prior year. Of the 152 eligible participants enrolled, 145 (72 males; 73
females) completed all phases of the study. Six recruits could not be contacted at
follow-up and one was dropped because of language problems. Attrition was not
related to condition or the dependent variables.
We used a two-group, repeated measures design. Female interviewers collected data
during three structured, telephone interviews with participants. The baseline interview
was conducted approximately 7 months (mean days ¼192.23, SD ¼111.59) after the
breakup. Participants received and completed their writing assignments (see below)
within the next week. The second interview was conducted approximately 2 weeks
(mean days ¼18.07, SD ¼6.40) after participants completed their writing assignments.
The final interview was conducted approximately 15 weeks (mean days ¼102.57,
SD ¼13.65) after participants completed their writing assignments. Demographic and
background variables were measured once; the remaining variables (see below) were
measured in all three interviews. After the first interview, we randomly assigned partic-
ipants to an experimental or control writing condition.
We mailed participants a packet that included instructions for writing in either the
experimental or control condition. As recommended by Pennebaker (1989),
participants were instructed to write one 20-min essay on each of three consecutive
days, and to write in a private, quiet, and comfortable spot.
Participants in the experimental group received the following daily instructions: ‘‘We
want you to let go and write about your deepest thoughts and feelings about the rela-
tionship. You can write about your thoughts and feelings regarding the relationship,
how the relationship affected your life when you were in it, or the effect of the relation-
ship on your life in the present. The important thing is that you dig down into your
deepest emotions and explore them in your writing. Do not worry about grammar
and spelling.’’ To help individuals to form a coherent story about the breakup, we pro-
vided additional directions for writing on each day. On day one, they were instructed
to: ‘‘Write as much as you can remember about what your relationship was like
before you broke-up with your romantic partner.’’ On day two, they were instructed
to: ‘‘Write down the events and factors that you think lead up to your breakup and
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 551
about the actual breakup.’’ On day three, they were instructed to: ‘‘Write about the
aftermath of the breakup.’’
Participants in the control group wrote about impersonal relationship topics. They
received the following daily instructions: ‘‘Try to develop rational, or logical, argu-
ments and do not express your feelings or emotional reactions to this issue. Do not
worry about grammar and spelling.’’ Participants also received additional instructions
for each day. On day one, they wrote about the following problem: ‘‘Should universities
promote ‘safe sex’ materials, even though this may offend some students’ religious
views?’’ On day two, they wrote about this problem: ‘‘Should men and women be
allowed to cohabitate in the same dormitory or dormitory room?’’ On day three,
they wrote about this problem: ‘‘Should college students and professors be allowed
Manipulation Check In the second interview, participants answered several questions
about their essays. Using a 9-point scale (1 ¼not at all, 9 ¼very much), they indicated
the extent to which their essay was ‘‘meaningful,’’ ‘‘personal,’’ and ‘‘revealing of your
Upper Respiratory Symptoms We assessed physical symptoms of illness using the
Upper Respiratory Symptoms Scale (URSS; Vickers and Hervig, 1988). Respondents
indicated on a 5-point scale (1 ¼not at all/absent, 5 ¼extremely) how severe each symp-
tom (e.g., sore throat, sneezing) was for them in the prior seven days. The scale had
good reliability (alpha ¼0.86).
Mood We used the shortened version of the Profile of Mood States (POMS-SF;
Shacham, 1983) to assess five dimensions of mood. Respondents indicated on a
5-point scale (1 ¼not at all, 5 ¼extremely) how much they felt each mood in the prior
seven days. The total summary mood scale had good reliability (alpha ¼0.90), as did
the subscales: depression (alpha ¼0.92), tension (alpha ¼0.88), anger (alpha ¼0.82),
vigor (alpha ¼0.87), and fatigue (alpha ¼0.90). These subscales tap both negative
mood (depression, anger) and physiological arousal/activation (tension, vigor, fatigue).
The intercorrelations between the subscales ranged from 0.27 to 0.69 (average r¼0.49).
These results justified separate analyses of each of the mood subscales. Principal com-
ponents analysis confirmed the existence of the five underlying mood factors.
Cognitive Processing We adapted the Impact of Events Scale (IES; Horowitz et al.,
1979) to assess intrusive thoughts and avoidance related to the breakup. High intrusion
and avoidance reactions are indicative of incomplete or unsuccessful cognitive pro-
cessing. The IES can apply to any stressful event, referred to in the items as ‘‘it.’’ We
substituted the term ‘‘breakup’’ for ‘‘it.’’ A sample intrusion item is: ‘‘had thoughts
about the breakup when you didn’t meant to.’’ A sample avoidance item is: ‘‘tried
not to think about the breakup’’. In the present study, the intrusion and avoidance sub-
scales were very highly correlated (r¼0.77) and preliminary analyses revealed that the
subscales were indistinguishable in terms of their association with other variables.
Therefore, we just used the total score in subsequent analyses. Respondents indicated
552 S.J. LEPORE AND M.A. GREENBERG
on a 5-point scale (1 ¼never, 5 ¼very often) how true each statement was for them in
the prior 7 days. The scale had good reliability (alpha ¼0.90).
Relationship Status,Feelings and Attitudes about the Ex-partner At each wave, respon-
dents described their current relationship status (e.g., in a new relationship, reunited
with ex-partner). We also assessed their views about their ex-partner and the breakup
using the Feelings and Attitudes Toward Ex (FATE; Ahrons, 1997) measure. We
used five of the FATE subscales: regret over the breakup, distance from ex-partner,
anger toward ex-partner, positive feelings toward ex-partner, and blame ex-partner
for breakup. Respondents indicated on a 5-point scale (1 ¼strongly disagree,
5¼strongly agree) how much they agreed or disagreed with each statement about
their ex-partner and their personal role in the breakup. Preliminary analyses revealed
that some of the subscales had very low reliability (alpha<0.60), so we conducted a
principal components analysis on all of the items. This analysis revealed three reliable
underlying factors, which we have labeled caring/positive regard for ex-partner
(alpha ¼0.85), resent/anger toward ex-partner (alpha ¼0.83), and guilt over the
breakup (alpha ¼0.81). A sample caring item is: ‘‘I care about my ex’s welfare.’’ A
sample resent item is: ‘‘I feel angry for the hurt I have gone through.’’ A sample
guilt item is: ‘‘I feel guilty about the breakup.’’
As shown in Table I, participants in the expressive writing condition rated their essays
as significantly more meaningful, personal, and revealing of their emotions than did
participants in the control condition.
Effects of Expressive Writing on Upper-Respiratory Symptoms and Mood
Figure 1 shows participants’ mean levels of upper respiratory symptoms as a function
of writing condition at Time 1 (T1), Time 2 (T2) and Time 3 (T3). This figure suggests
that the control group experienced a short-term (T1toT2) increase in symptoms fol-
lowed by a return to baseline, whereas the expressive writing group experienced no
change in symptoms over time. Change scores (T2T1andT3T1) were calculated
to analyze the change in symptoms over time as a function of writing condition.
Table I. Manipulation Check on Essay Contents (n¼145)
Control group Experimental group
M SE M SE t (143)
Rating of essay as revealing of emotions 5.28 0.22 6.99 0.14 6.70***
Rating of essay as personal 5.04 0.24 7.77 0.14 9.79***
Rating of essay as meaningful 5.81 0.19 7.04 0.18 4.75***
Note: All rating scales range from 1 (not at all) to 9 (very much). ***p<0:001.
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 553
The change scores were subjected to a 2 (condition) 2 (period) repeated measures
analysis of variance (ANOVA). There was no significant main effect of time or con-
dition, but there was a significant Condition Period interaction, F(1, 143) ¼4.69,
p<0.05. Simple effect analyses confirmed that the locus of this interaction was the
increase in symptoms in the control group from T1toT2(p<0.05). There were no dif-
ferences between groups in level of symptoms at T1andT3.
Change scores (T2T1andT3T1) were calculated to analyze changes in mood
states over time as a function of writing condition. We first analyzed changes in the
overall POMS summary scores using a 2 (condition) 2 (period) ANOVA. We
found no significant main or interactive effects. Next, we looked for effects on specific
moods by analyzing changes in participants’ scores on the five mood subscales.
The multivariate tests revealed no significant main effects of condition or time,
but there was a significant Condition Period interaction, F(6, 138) ¼2.55, p<0.05.
The univariate tests revealed significant Condition Period interactions on tension
scores, F(1, 143) ¼5.56, p<0.05, and fatigue scores, F(1, 143) ¼7.76, p<0.05, both
of which reflect physiological arousal/activation. Simple effect analyses on the tension
and fatigue scores showed that the locus of the interaction was the increase in tension
and fatigue in the control group from T1toT2(p< 0.05). There were no differences
between groups in level of tension and fatigue at T1 and T3. As shown in Fig. 2, these
effects parallel the effect on upper respiratory symptoms.
Effects of Expressive Writing on Relationship Status, Feelings and
Attitudes Toward Ex-Partner and Cognitive Processing
Over the course of the study, six (8.2%) of the participants in the expressive
writing condition were reunited with their ex-partner, while only one (1.4%) of the
FIGURE 1 Mean upper respiratory illness symptoms as a function of writing condition and epoch. The
experimental group expressed their deepest thoughts and feelings about the breakup in their writing, whereas
the control group wrote dispassionately about impersonal relationship topics. Bars represent the standard
error of the mean. Time 1, 2 and 3 refer, respectively, to approximately 7-, 8- and 11-months after the
breakup. The writing manipulation was administered in the week following the Time 1 measurement period.
554 S.J. LEPORE AND M.A. GREENBERG
participants in the control condition was reunited. This association between writing and
reuniting was marginally significant [
(1, 145) ¼3.68, p<0.06]. Writing condition
did not influence whether participants engaged in a relationship with a new
Change scores (T2T1 and T3T1) were calculated to analyze changes in feelings
and attitudes toward the ex-partner and level of cognitive processing. We analyzed the
change scores using 2 (condition) 2 (period) ANOVAs and found that expressive
writing had no influence on participants’ feelings and attitudes toward their ex-partner
or their level of intrusive thoughts and avoidance. However, mean scores on these
variables did change over time, revealing a pattern of increased emotional detach-
ment from the ex-partner and the breakup. Specifically, participants reported sig-
nificant decreases in resentment toward their ex-partner, caring for their ex-partner,
guilt over the breakup, and symptoms of intrusions and avoidance (all p’s <0.05; see
Table II. Changes in Feelings and Attitudes Toward Ex-Partner, Coping, and Cognitive Processing Over
Time 1 Time 2 Time 3
Feeling & attitudes toward ex-partnera
Resent toward ex-partner 1.66 0.05 1.58 0.05 1.54 0.05 4.02*
Caring for ex-partner 3.38 0.07 3.27 0.07 3.13 0.08 8.87***
Guilt about break-up 2.69 0.08 2.46 0.09 2.41 0.08 8.63***
Intrusions/avoidance 2.11 0.06 1.79 0.05 1.53 0.05 49.78***
Scales ranged from1 (strongly disagree) to 5 (strongly agree).
Scale ranged from 1 (never) to 5 (very often).
* p<.05. ** p<.01. ***P<.001.
FIGURE 2 Mean tension (left) and fatigue (right) as a function of writing condition and epoch. Bars
represent the standard error of the mean. See the caption accompanying Fig. 1 for details on the manipulation
and timing of measurements.
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 555
Interactive Effects of Expressive Writing and Cognitive Processing Variables on
Upper Respiratory Symptoms and Mood
While expressive writing did not have a main effect on cognitive processing over time,
we were interested in whether it interacted with cognitive processing to influence upper
respiratory symptoms and mood. We limited analyses to the outcomes that were
affected by expressive writing: changes from T1toT2 in upper respiratory symptoms,
tension and fatigue. Moderated regression analyses revealed a significant interaction of
Condition Cognitive Processing (Intrusions/Avoidance), Unstandardized B ¼0.55,
p<0.01,on changes in upper respiratory symptoms. As shown in the plot in Fig. 3,
higher levels of intrusive thoughts and avoidance were associated with increases
in upper respiratory symptoms among participants in the control group but not
among participants in the expressive writing group. There were no significant interac-
tive effects of condition and cognitive processing on the mood outcomes.
The present findings suggest that expressive writing has beneficial effects on a wide
range of outcomes, including mood, physical health, and social functioning. In the
aftermath of a relationship breakup, experimental participants, who wrote expressive
essays about their breakup, reported no increases in upper respiratory symptoms, ten-
sion, or fatigue. In contrast, control participants, who wrote about impersonal topics,
reported short-term increases in upper respiratory symptoms, tension and fatigue over
time. There also was a trend ( p<0.06) suggesting that participants in the experimental
group were more likely to reunite with their ex-partner than were participants in the
control group. Expressive writing did not affect participants’ probability of starting a
new relationship, their level of intrusions and avoidance, or their feelings and attitudes
toward their ex-partner.
By analyzing the interaction between writing condition and cognitive processing, we
had hoped to learn more about the mechanisms linking expressive writing to mood and
FIGURE 3 Slope of the relation between level of intrusions/avoidance at Time 1 and changes in upper
respiratory illness (URI) symptoms from Time 1 to Time 2 (T2T1) as a function of writing condition.
IES ¼Impact of Event Scale score. See the caption accompanying Fig. 1 for details on the manipulation and
timing of measurements.
556 S.J. LEPORE AND M.A. GREENBERG
health outcomes. These analyses partially supported our predictions. Expressive writing
appeared to attenuate the effects of incomplete cognitive processing on upper respira-
tory symptoms but not on mood. As shown in Fig. 3, individuals with incomplete
cognitive processing, as represented by high levels of intrusions and avoidance, had
short-term increases in upper respiratory symptoms only if they did not engage in
expressive writing. If they did engage in expressive writing, incomplete cognitive
processing was not associated with changes in upper respiratory symptoms. Our data
do not allow us to identify the precise reason for this buffering effect, but several
explanations seem plausible.
One explanation is that expressive writing allowed individuals to become habituated
to stressful stimuli (cf. Bootzin, 1997; Lepore, 1997), such as intrusive thoughts. From
this perspective, expressive writing is a form of exposure therapy, in which negative
responses to stressful stimuli become extinguished through repeated exposure to the
stimuli in a safe context. Another explanation for the buffering effect is that expressive
writing enhanced individuals’ control, or perceived control, over their responses to
stressful stimuli (see Greenberg et al., 1996). Alternatively, expressive writing could
have enhanced self-regulation by altering how individuals coped with stress-related
thoughts. For instance, in contemplating their breakup in the expressive writing task,
individuals might have learned to deal with confusing and uncomfortable thoughts
and memories by using distraction, problem solving, or relaxation. Yet another expla-
nation of the buffering effects is that the quality of participants’ intrusive thoughts and
avoidant behaviors in the experimental group changed over time, even though the fre-
quency of these responses was unaffected by expressive writing. For instance, in the
experimental group, intrusions might have become less stressful because they now fit
into a coherent, organized story of the failed relationship, and were not diffuse and
unprocessed thoughts, memories, and images. It is also possible that expressive writing
changed the quality of individuals’ responses to intrusive thoughts through reappraisal
processes. That is, confronting stress-related thoughts and feelings through writing
might have enabled participants to form benign appraisals about previously threatening
stimuli, such as intrusive thoughts.
In order to account for the pattern of risk for infectious illness among participants in
the current study, the above processes must be linked to biological processes of pathol-
ogy. We believe that the stress of incomplete cognitive processing could have increased
risk for infectious illness in the control group by directly activating neuroendocrine and
physiological stress responses that suppress immune system functioning. In this com-
promised state, individuals in the control group who came into contact with viruses
in the university environment would have a heightened risk for developing an illness.
In the experimental group, the various mechanisms noted above (e.g., habituation,
reappraisal, self-regulation), could have directly dampened neuroendocrine responding
to the stress of incomplete cognitive processing, thus short-circuiting physiological
stress response and concomitant changes in immune functioning that influence risk
for infectious illnesses. While these speculations are not implausible, they await further
empirical evidence. Several investigators have found beneficial effects of expressive
writing on immune system functioning, but the clinical significance of these effects is
unknown (see review by Petrie et al., 1995). Further, there is a lack of data on the
underlying neural, physiological, and hormonal mechanisms accounting for the effects
of expressive writing on immune functioning.
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 557
The lack of interactive effects of expressive writing and cognitive processing on mood
in the present study is inconsistent with Lepore’s (1997) study, as is the lack of main
effects of expressive writing on several mood outcomes (see Smyth, 1998). Prior
work, however, suggests that such findings are not unprecedented. Several investigators
have reported that the effects of expressive writing are more robust with physical out-
comes than with emotional outcomes (see reviews by Pennebaker, 1989; Smyth, 1998).
Even in the current study, the few main effects of expressive writing on mood were lim-
ited to those measures that tapped physiological arousal (tension, fatigue). These find-
ings reinforce the notion that stress, and stress-reduction techniques, have somewhat
independent effects on emotional and physical outcomes (see Lepore et al., 2002).
These results also parallel findings in the broader literature on stress, which show
that while stress increases both negative emotions and physiological activation (e.g.,
increased autonomic arousal, increased cortisol production), the association between
negative emotions and physiological stress responses is quite small (Feldman et al.,
In addition to the interaction hypotheses, we predicted that expressive writing would
facilitate adjustment to the breakup by altering participants’ feelings and attitudes
toward their ex-partner. However, the null effects of expressive writing on feelings of
guilt, positive regard for the ex, and resent toward the ex are inconsistent with this pre-
diction. Thus, we have no evidence of expressive writing resulting in schema revision. It
is possible that our measures were not sensitive enough or did not tap the appropriate
schemas. For instance, expressive writing might have influenced individuals’ future
expectations for relationships rather than their beliefs about the failed relationship.
A related possibility is that expressive writing altered global beliefs about relationships
(e.g., whether to trust others), but not beliefs about the failed relationship. It is also
possible that expressive writing influenced beliefs about the failed relationship that
we did not assess. For instance, expressive writing might have influenced feelings of
control over the breakup. The fact that experimental participants were more likely to
reunite with the ex-partner is consistent with this latter hypothesis.
In summary, in the aftermath of a failed relationship, expressive writing reduced indi-
viduals’ risk for developing symptoms of upper respiratory illness, tension and fatigue.
Incomplete cognitive processing, which was indicated by high levels of intrusive
thoughts and avoidance, was associated with an increased risk of developing symptoms
of upper respiratory illness in the control group, but this association was absent in the
expressive writing group. Thus, it appears that expressive writing buffers individuals
from the inimical health effects of intrusion and avoidance reactions to stressors.
Further, we found that expressive writing influenced social adjustment, as individuals
in the experimental group were more likely to re-unite with their ex-partner. Future
research is needed to identify the precise mechanisms through which expressive writing
enhances social adjustment and mitigates the negative health effects of incomplete
This work was supported in part by a grant from The City University of New York
PSC-CUNY Research Award Program. In addition, we are grateful for the excellent
558 S.J. LEPORE AND M.A. GREENBERG
assistance of the following individuals: Tracy Bach, Michelle Bruno, Shannon Gibson,
Rivkah Kaufmann, Jodi Kurtz, Elizabeth Legowski and Melissa Sloan.
Ahrons, C. (1997). Feelings and Attitudes Toward Ex-Partner (FATE). Personal communication.
Baum, A., Cohen, L. and Hall, M. (1993). Control and intrusive memories as possible determinants of chronic
stress. Psychosomatic Medicine,55, 274–286.
Bootzin, R.R. (1997). Examining the theory and clinical utility of writing about emotional experiences.
Psychological Science,8, 167–169.
Clark, L.F. (1993). Stress and the cognitive-conversational benefits of social interaction. Journal of Social and
Clinical Psychology,12, 25–55.
Cameron, L.D. and Nicholls, G. (1998). Expression of stressful experiences through writing: effects of a self-
regulation manipulation for pessimists and optimists. Health Psychology,17, 84–92.
Donnelly, D.A. and Murray, E.J. (1991). Cognitive and emotional changes in written essays and therapy
interviews. Journal of Social and Clinical Psychology,10, 334–350.
Feldman, P., Cohen, S., Lepore, S.J., Matthews, K., Kamarck, T. and Marsland, A. (1999). The impact of
negative emotions on acute physiological responses to stress. Annals of Behavioral Medicine,21, 800–805.
Gold, D.B. and Wegner, D.M. (1995). Origins of ruminative thought: trauma, incompleteness, nondisclosure,
and suppression. Journal of Applied Social Psychology,25, 1245–1261.
Greenberg, M.A. (1995). Cognitive processing of traumas: the role of intrusive thoughts and reappraisals.
Journal of Applied Social Psychology,25, 1262–1296.
Greenberg, M.A. and Stone, A.A. (1992). Emotional disclosure about traumas and its relation to health:
effects of previous disclosure and trauma severity. Journal of Personality and Social Psychology,63,
Greenberg, M.A., Wortman, C.B. and Stone, A.A. (1996). Emotional expression and physical health: revising
traumatic memories or fostering self-regulation?. Journal of Personality and Social Psychology,71,
Horowitz, M.J. (1986). Stress Response Syndromes, 2nd Edn. Jason Aronson, New York.
Horowitz, M.J., Wilner, N. and Alvarez, W. (1979). The Impacts of Events Scale: a measure of subjective
stress. Psychosomatic Medicine,41, 209–218.
Kaczmarek, P., Backlund, B. and Biemer, P. (1990). The dynamics of ending a romantic relationship: an
empirical assessment of grief in college students. Journal of College Student Development,31, 319–324.
Kennedy-Moore, E. and Watson, J.C. (1999). Expressing Emotion. The Guilford Press, New York.
King, L.A. (2002). Gain without pain? Expressive writing and self-regulation. In: Lepore, S.J. and Smyth,
J.M. (Eds.), The Writing Cure: How Expressive Writing Promotes Health and Emotional Well-Being,
pp. 119–134. American Psychological Association, Washington, D.C.
Kliewer, W.L., Lepore, S.J., Oskin, D. and Johnson, P.D. (1998). The role of social and cognitive processes in
children’s adjustment to community violence. Journal of Consulting and Clinical Psychology,66, 199–209.
Lange, A., van de Ven, J-P.Q.R., Schrieken, B.A.L., Bredeweg, B. and Emmelkamp, P.M.G. (2000). Internet-
mediated, protocol driven treatment of psychological dysfunction. Journal of Telemedicine and Telecare,
Lepore, S.J. (1997). Expressive writing moderates the relation between intrusive thoughts and depressive
symptoms. Journal of Personality and Social Psychology,73, 1030–1037.
Lepore, S.J., Greenberg, M.A., Bruno, M., and Smyth, J.M. (2002). Expressive writing and health: Self-
regulation of emotion-related experience, physiology, and behavior. In: Lepore, S.J. and Smyth, J.M.
(Eds.), The Writing Cure: How Expressive Writing Promotes Health and Emotional Well-Being, pp. 99–
118. American Psychological Association, Washington, D.C.
Lepore, S.J. (2001). A social-cognitive processing model of emotional adjustment to cancer. In: Baum, A. and
Andersen, B. (Eds.), Psychosocial Interventions for Cancer, pp. 99–118. American Psychological
Association, Washington, DC.
Lepore, S.J. and Helgeson, V. (1998). Social constraints moderate the relation between intrusive thoughts and
mental health in prostate cancer survivors. Journal of Social and Clinical Psychology,17, 89–106.
Lepore, S.J., Ragan, J.D. and Jones, S. (2000). Talking facilitates cognitive-emotional processes of adaptation
to an acute stressor. Journal of Personality & Social Psychology,78, 499–508.
Lepore, S.J., Silver, R.C., Wortman, C.B. and Wayment, H.A. (1996). Social constraints, intrusive thoughts,
and depressive symptoms among bereaved mothers. Journal of Personality and Social Psychology,70,
Lutgendorf, S.K. and Antoni, M.H. (1999). Emotional and cognitive processing in a trauma disclosure para-
digm. Cognitive Therapy and Research,23, 423–440.
EXPRESSIVE WRITING AND RELATIONSHIP BREAKUPS 559
Major, B. and Gramzow, R.H. (1999). Abortion as stigma: cognitive and emotional implications of conceal-
ment. Journal of Personality and Social Psychology,77, 735–745.
Manne, S.L. (1999). Intrusive thoughts and psychological distress among cancer patients: the role of spouse
avoidance and criticism. Journal of Consulting and Clinical Psychology,67, 539–546.
McFarlane, A. (1992). Avoidance and intrusion in posttraumatic stress disorder. Journal of Nervous and
Mental Disease,180, 439–445.
Murray, E.J., Lamnin, A.D. and Carver, C.S. (1989). Emotional expression in written essays and psychother-
apy. Journal of Social and Clinical Psychology,8, 414–429.
Pennebaker, J.W. (1989). Confession, inhibition, and disease. In: Berkowitz, L. (Ed.), Advances in
Experimental Social Psychology, Vol. 22, pp. 211–244. Academic Press, Orlando, FL.
Pennebaker, J.W. (1995). Emotion,Disclosure,and Health. American Psychological Association, Washington,
Pennebaker, J.W. and Beall, S.K. (1986). Confronting a traumatic event: toward an understanding of inhibi-
tion and disease. Journal of Abnormal Psychology,95, 274–281.
Pennebaker, J.W., Colder, M. and Sharp, L.K. (1990). Accelerating the coping process. Journal of Personality
and Social Psychology,58, 528–537.
Pennebaker, J.W., Kiecolt-Glaser, J.K. and Glaser, R. (1988). Disclosure of traumas and immune function:
health implications for psychotherapy. Journal of Consulting and Clinical Psychology,56, 239–245.
Pennebaker, J., Mayne, T. and Francis, M. (1997). Linguistic predictors of adaptive bereavement. Journal of
Personality and Social Psychology,72, 863–871.
Petrie, K.J., Booth, R.J., Pennebaker, J.W., Davison, K.P. and Thomas, M.G. (1995). Disclosure of trauma
and immune response to hepatitis B vaccination program. Journal of Consulting and Clinical Psychology,
´, B. (1995). Mental rumination, social sharing, and the recovery from emotional exposure. In:
Pennebaker, J.W. (Ed.), Emotion, Disclosure & Health, pp. 271–292. American Psychological
Association, Washington, DC.
Shacham, S. (1983). A shortened version of the profile of mood states. Journal of Personality Assessment,47,
Smith, H.S. and Cohen, L.H. (1993). Self-complexity and reactions to a relationship breakup. Journal of
Social and Clinical Psychology,12, 367–384.
Smyth, J.M. (1998). Written emotional expression: effect sizes, outcome types, and moderating variables.
Journal of Consulting and Clinical Psychology,66, 174–184.
Smyth, J.M. and Greenberg, M.A. (2000). Scriptotherapy: the effects of writing about traumatic
events. In: Masling, J. and Duberstein, P. (Eds.), Empirical Studies in Psychoanalytic Theories
Vol. 9, Psychoanalytic Perspectives on Health Psychology, pp. 121–164. American Psychological
Association, Washington, DC.
Smyth, J.M., Stone, A.A., Hurewitz, A. and Kaell, A. (1999). Effects of writing about stressful experiences on
symptom reduction in patients with asthma or rheumatoid arthritis. Journal of the American Medical
Spera, S.P., Buhrfeind, E.D. and Pennebaker, J.W. (1994). Expressive writing and coping with job loss.
Academy of Management Journal,37, 722–733.
Sprecher, S., Felmlee, D., Metts, S., Fehr, B. and Vanni, D. (1998). Factors associated with distress following
the breakup of a close relationship. Journal of Social and Personal Relationships,15, 791–809.
Tait, R. and Silver, R.C. (1989). Coming to terms with major negative life events. In: Uleman, J.S. and Bargh,
J.A. (Eds.), Unintended Thought, pp. 351–382. Guilford, New York.
Vickers, R., Jr. and Hervig, L.K. (1988). Psychosocial Risk Factors for Upper Respiratory Infections:
Assessment of Upper Respiratory Illness During Basic Training Tech., Rep. No. 88–23. Naval Health
Research Center, San Diego.
560 S.J. LEPORE AND M.A. GREENBERG