Finding the silver lining

Abstract

The present research sought to identify a novel factor that reduces the emotional impact of relationship dissolution. Specifically, we examined individuals’ use of redemptive narratives—a form of narrative focused on positive outcomes in negative situations—in post-dissolution journal entries written over the course of a 4-day diary study. An interaction emerged, that is, individuals reported reduced emotional distress to the extent that they had used more, versus fewer, redemptive narratives the day before (a lagged effect), but only at the end, not the beginning, of the study. This effect was independent from similar effects predicted by greater cognitive processing. These findings suggest that redemptive narratives may reduce individuals’ post-dissolution distress.

Full text

Article
Finding the silver lining:
The relative roles of
redemptive narratives and
cognitive reappraisal in
individuals’ emotional
distress after the end of a
romantic relationship
Erica B. Slotter
Deborah E. Ward
Villanova University, USA
Abstract
The present research sought to identify a novel factor that reduces the emotional impact
of relationship dissolution. Specifically, we examined individuals’ use of redemptive
narratives—a form of narrative focused on positive outcomes in negative situations—in
post-dissolution journal entries written over the course of a 4-day diary study. An inter-
action emerged, that is, individuals reported reduced emotional distress to the extent
that they had used more, versus fewer, redemptive narratives the day before (a lagged
effect), but only at the end, not the beginning, of the study. This effect was independent
from similar effects predicted by greater cognitive processing. These findings suggest that
redemptive narratives may reduce individuals’ post-dissolution distress.
Keywords
Emotional distress, reappraisal, redemptive narratives, relationship dissolution
Corresponding author:
Erica B. Slotter, Villanova University, Tolentime Hall, 800 Lancaster Avenue, Villanova, PA 19085, USA.
Email: erica.slotter@villanova.edu
Journal of Social and
Personal Relationships
1–20
ªThe Author(s) 2014
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0265407514546978
spr.sagepub.com
J S P R
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

If clouds are blocking the sun, there will always be a silver lining that reminds me to keep on
trying. ~Matthew Quick, The Silver Linings Playbook: A Novel
The end of a romantic relationship is typically upsetting to individuals (e.g., Bowlby,
1980; Simpson, 1987). Divorce robustly predicts a variety of negative mental and
physical health outcomes ranging from emotional distress and depression to a greater
likelihood of various physical health problems (e.g., Lucas, 2005; Sbarra, Law, &
Portley, 2011). Even the end of dating relationships predicts detriments to well-being
(e.g., Boelen & Reijntjes, 2009; Davis, Shaver, & Vernon, 2003; Rhoades, Kamp Dush,
Atkins, Stanley, & Markman, 2011), with breakup being a robust predictor of major
depressive disorder in young adults (Monroe, Rohde, Seeley, & Lewinsohn, 1999).
Given the impact of relationship termination on well-being, in combination with the fact
that nearly 2 million adults in the US divorce every year and the dissolution of dating
relationships is even more common (e.g., Tejada-Vera & Sutton, 2010), it is hardly sur-
prising that many psychological traditions examine factors that influence individuals’
emotional distress after the end of a romantic relationship. In the present study, we
sought to further this investigation by examining whether the use of redemptive narra-
tives (e.g., McAdams, Reynolds, Lewis, Patten, & Bowman, 2001) when thinking about
the end of a relationship would predict emotional recovery. In short, redemptive narra-
tives entail seeking the silver lining in dark clouds, focusing on positive outcomes that
can emerge from negative circumstances.
Extensive research on relationship dissolution, both of dating relationships and of
marriages, has focused on facets of individuals’ cognition, emotions, and behavior that
may help alleviate post-dissolution distress. Factors that appear to reduce distress—
although certainly not an exhaustive list—include individual differences, such as attach-
ment security (e.g., Feeney & Noller, 1992), self-views such as self-compassion (Sbarra,
Smith, & Mehl, 2012), and various cognitive or emotional tendencies. One such as cog-
nitive or emotional tendency that we include in the present investigation is the use of
cognitive reappraisal and processing of emotional events (e.g., Gross, 2002).
Appraisal theories of emotion suggest that it is the subjective interpretation of an
event, rather than the event itself, that leads to specific emotional reactions (Lazarus &
Folkman, 1985). Research on appraisals and stress have found that people respond quite
differently to the same or similar stressful life events depending on their interpretation of
the situation. Cognitive reappraisal, which entails reframing an event in order to change
one’s emotional response to it, strives to aid individuals in reinterpreting emotional
events in ways that mitigate the associated distress (Finkel, Slotter, Luchies, Walton, &
Gross, 2013; Gross, 1998; Troy, Wilhelm, Shallcross, & Mauss, 2010). Unlike rumi-
nation, which encourages continued focus on the emotionality of an event and thus is
associated with prolonged distress, cognitive reappraisal encourages individuals to think
about an event rationally and distance themselves from it emotionally (e.g., Denson,
Moulds, & Grisham, 2012; Richards, Butler, & Gross, 2003; Sbarra, Boals, Mason,
Larson, & Mehl, 2013).
At its core, cognitive reappraisal involves construing emotion-eliciting occurrences in
ways that alter their meaning and focuses on engaging in greater rational cognitive pro-
cessing, rather than ‘‘hot’’ emotional processing, of an emotional occurrence (e.g.,
2Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Gross, 2002; Lazarus & Alfert, 1964). Research has demonstrated that reinterpreting, or
reappraising, negative emotional events in a variety of contexts improves subjective
well-being (e.g., Gross & John, 2003). For example, a 6-week intervention study of older
adults experiencing depression and life challenges pitted types of cognitive therapies
against one another (Watt & Cappeliez, 2000). In comparison to those in a social skills
control group, participants in cognitive therapeutic conditions showed a decrease in
depressive symptoms over a period of 3 months (Watt & Cappeliez, 2000). Further, par-
ticipants receiving ‘‘integrative therapy,’’ in which they had actively practiced cognitive
reappraisal, outperformed participants receiving ‘‘instrumental therapy,’’ in which they
had passively recalled positive coping strategies that could be applied to their current
negative life conditions (Watt & Cappeliez, 2000). These and other research findings
(e.g., Finkel et al., 2013; Gross, 1998; Gross, 2002; Lepore & Greenberg, 2002) suggest
that cognitive reappraisal can reduce emotional distress in response to life events, even
when the content being processed is negative in nature.
The present research sought to contrast cognitive reappraisal with a different cog-
nitive or emotional tendency that might also predict reduced emotional distress after the
end of a romantic relationship, namely redemptive narratives. Research on narrative
identities has demonstrated that, when contemplating their lives, individuals’ focus on
different aspects of important life events, and the patterning of what individuals choose
to focus on, versus not, serves as an important predictor of well-being (Bauer, McAdams,
& Pals, 2008). Redemptive narratives represent a particular pattern of storytelling that
individuals sometimes engage in when recounting events in their lives (McAdams
et al., 2001). At their core, redemptive narratives encompass the idea that negative life
events or circumstances can be meaningful points in individuals’ lives that result in
positive outcomes or silver linings. Greater redemptive narrative use when recounting
life events is associated with enhanced well-being, generativity in older adults, and
emotional resilience to various negative events across several different age-groups
(e.g., McLean & Pratt, 2006; Pals, 2006). Redemptive narratives differ from cognitive pro-
cessing in terms of the focus of interpretation. While cognitive reappraisal involves dis-
tancing oneself from an event emotionally, creation of a redemptive narrative involves
actively increasing and enhancing one’s positive emotional experiences derived from
an event by engaging in attempts to derive meaning from the event itself (Tugade, 2011).
One way of assessing both individuals’ cognitive reappraisals or processing and
redemptive narrative use regarding events in their lives is via coding of the linguistic
content of individuals’ verbal statements or writing samples (Pennebaker & Beall, 1986;
Pennebaker, Mehl, & Niederhoffer, 2003). Rating the statements that individuals make
regarding cognitive processes, such as insight and causality, has been shown to tap into
the extent to which individuals are processing a negative emotional event in a calm,
rational way (North, Meyerson, Brown, & Holahan, 2012; Richards, Butler, & Gross,
2003). Relevant to the present work, greater rational cognitive processing during jour-
naling tasks has been shown to be associated with lessened emotional impact after the
end of dating relationships (Boals & Klein, 2005; Lepore & Greenberg, 2002). Similarly,
redemptive narratives are traditionally coded from audiotapes, transcripts, or writing
samples of individuals describing particular events in their lives (McAdams et al., 2001).
Thus, coding individuals’ writing samples, such as journals, can provide important
Slotter and Ward 3
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

insight into the different processes that they may employ in their efforts to come to terms
with the loss of a romantic relationship. Furthermore, expressive writing (e.g., Penne-
baker & Beall, 1986) has been used as an intervention in previous research to aid
individuals in overcoming a variety of stressful life events, including the loss of a
romantic relationship (e.g., Lepore & Greenberg, 2002; Sbarra, et al., 2013). Thus,
simply writing about a relationship dissolution benefits individuals, and, as outlined
above, additional work has established that the qualities of individuals’ writing during
journaling tasks can help (e.g., cognitive processing or redemptive narrative use; e.g.,
Finkel et al., 2013; McLean & Pratt, 2006) or hinder (e.g., rumination; Sbarra et al.,
2013) individuals’ emotional recovery after relationship dissolution. The present
research sought to integrate the field’s knowledge of the factors that may aid in recovery
after the end of a romantic relationship and thus focused on the relative associations of
redemptive narrative use and cognitive reappraisal or processing with emotional distress
post-relationship dissolution.
The present research
The central aims of the research were threefold. First, the present work represents one of
the first attempts to examine the usefulness of redemptive narrative use, specifically, on
reducing emotional distress after a life event, and to focus on romantic relationship
dissolution as the life event of interest. Second, to the researchers’ knowledge, this is the
first effort to directly compare the effectiveness of redemptive narrative use and cog-
nitive reappraisal or processing within the same event context. Both redemptive narra-
tives and cognitive reappraisal are positively associated with well-being after a variety of
life events (e.g., McLean & Pratt, 2006; Watt & Cappeliez, 2000), but theoretically oper-
ate via distinct pathways. Redemptive narratives focus on finding emotional positivity in
negative events (e.g., Pals, 2006), whereas cognitive reappraisal focuses on rationally
examining the negative event while reducing the emotional charge associated with it
(e.g., Gross, 2002). Thus, the one process increases or maintains emotional focus, albeit
altering the valence of the emotions, whereas the other reduces emotional focus. The cur-
rent research sought to examine how these two types of event-based processing may pre-
dict emotional distress over time and may independently predict distress following
relationship dissolution. Thus, the second central goal of the present study was to exam-
ine two different psychological processes—redemptive narrative use and cognitive pro-
cessing—that could be beneficial for individuals to engage in after a negative life
event—the end of a romantic relationship—and to establish whether they represent sim-
ilar or independent psychological processes in predicting emotional distress. We first
examined the independent effects of redemptive narrative use and cognitive processing
in predicting emotional distress. We then pit these two effects against each other in a
simultaneous analysis in order to determine whether redemptive narrative use and cog-
nitive processing would exert independent predictive effects upon emotional distress.
Third, the present study sought to investigate the brief temporal effects of these two
different psychological processes—redemptive narrative use and cognitive processing—
in predicting reduced emotional distress. Previous work examining redemptive narrative
use and cognitive processing tends to focus on either a cross-sectional approach, in
4Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

which single narratives or writing samples are coded for the relevant processes and
linked to well-being (e.g., Pals, 2006), or a longitudinal approach, in which narratives
or writing samples assessed in previous months or years are coded for relevant processes
and linked to well-being (e.g., Sbarra et al., 2013). The present work took a different
approach. Specifically, we examined whether redemptive narrative use and cognitive
processes would predict reduced emotional distress over an extremely brief period of
time—a 4-day diary study. We selected this brief time period in order to investigate how
powerful redemptive narratives and cognitive processing were shifting individuals’ emo-
tional distress. Do these processes need months or years to be effective, or are they asso-
ciated with distress over shorter time periods? We also focused on how these processes
might interact with time to predict distress. We examined whether redemptive narrative
use or cognitive processing predicts immediate shifts in emotional distress or whether
these processes exert their hypothesized effects over time.
Our three primary aims were to examine via three central hypotheses:
Hypothesis 1: We predicted that, across a brief daily dairy study, emotional dis-
tress would be reduced to the extent that a person engaged in greater redemptive
narrative use over time. Specifically, we predicted an interactive slope effect such
that greater redemptive narrative use on the previous day would predict less cur-
rent emotional distress (a lagged effect) and that this effect would be strengthened
over the course of a 4-day diary study.
Hypothesis 2: A second focus of the present study was to examine whether emo-
tional distress would be reduced post-dissolution to the extent that a person
engaged in greater cognitive reappraisal over time. Specifically, we predicted an
interactive slope effect such that greater cognitive processing regarding the end
of the relationship on the previous day would predict less current emotional dis-
tress (a lagged effect) and that this effect would be strengthened over the course
of a 4-day diary study.
Hypothesis 3: Finally, we predicted that redemptive narrative use and cognitive
reappraisal would reduce post-dissolution distress independently of one another.
Specifically, we predicted that when used to simultaneously predict individuals’
post-dissolution emotional distress, the effects of lagged redemptive narrative use
over time and lagged cognitive processing over time would both remain significant
predictors. This would suggest that both redemptive narrative use and cognitive
reappraisal and processing can predict reduced distress after life events but would
demonstrate empirically that they operate via distinct pathways.
Method
Participants
One-hundred and six nonstudent adults took part in a 4-day online daily diary study via a
recruitment panel from Qualtrics.com (59 women, 46 men, 1 unreported; M
age
¼38.42
years, SD ¼12.28, range ¼19–64). In order to be included, participants must have
experienced a romantic relationship dissolution within the past 6 weeks (88.35%within
Slotter and Ward 5
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

4 weeks of study start; M
weekssincedissolution
¼2.94, SD ¼1.46, range ¼0–6;M
relationshipduration
¼40.04 months, SD ¼58.67, range ¼1–288). Recruited participants had reported that they
were not keeping a journal or diary of any kind upon entering the study.
Procedure
Participants completed a series of questionnaires at an intake session (Day 0). On Days
1–4, participants completed abbreviated versions of the questionnaires used at intake as
well as a journaling task in which they were asked to write about the end of their rela-
tionship for up to 10 min. During the journaling task, participants wrote for an average of
8.5 min across the 4 days (SD ¼2.3) and retention rates in the study were excellent
(Intake/Day 0 n¼106, Day 1 n¼105, Day 2 ¼92, Day 3 ¼100, and Day 4 ¼98).
Relevant to the current study, participants completed a questionnaire measure of
emotional distress and participants’ journal entries were later coded for cognitive pro-
cessing and redemptive narrative use. Pearson’s rbivariate correlations are among these
three key variables and all continuous participant characteristics are presented in Table 1.
For the journaling task, participants were actually given one of three sets of
instructions, that is, to journal from a self-focused perspective, to journal from a
relationship-focused perspective, or to write anything they wanted about the end of their
relationship. This manipulation did not predict any of the variables reported in the pres-
ent study or interact with any of the variables in the models.
Self-report measures
Emotional distress. Participants completed a measure of depressive symptomology at each
assessment. Specifically, participants completed the Center for Epidemiological Stud-
ies–Depression scale, which is a measure of the amount of emotional distress an
Table 1. Bivariate correlations among key variables and continuous participant characteristics.
Redemptive
Narrative
(RN)
Cognitive
Processing
(CP)
Daily
Distress
(DD)
Intake
Distress
(ID) Age
Relationship
Duration
Weeks
Since
Breakup
Essay
Word
Count
RN – – – – – – – –
CP .05 – – – – – – –
DD .22^ .05 – – – – – –
ID .21^ .06 .79^ – – – – –
Age .04 .11* .05 .002 – – – –
Rel.
duration
.06 .01 .02 .03 .01 – – –
Weeks
since
Breakup
.004 .04 .04 .07 .12* .02 – –
Word
count
.05 .12* .06 .02 .11* .14^ .08 –
***p< .10. **p¼.05. *p< .05. ^p< .01.
6Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

individual is experiencing (Radloff, 1977). The present sample was nonclinical and no
diagnoses were made as part of this study; thus, we refer to this measure as emotional
distress.
In the present study, the original CES-D was modified slightly. The original measure
asks individuals about how they felt in the past few weeks, given the brief nature of the
present study, the CES-D in the present research was reworded to ask participants how
they were presently feeling on each day of the study (i.e., ‘‘I am bothered by things that
don’t usually bother me.’’). Also differing from the original measure, participants were
asked to agree or disagree with each statement on a Likert-type scale, and participants’
responses were averaged to create a means of emotional distress for each day of the
study, with higher scores indicating greater emotional distress (1 ¼strongly disagree
to 7 ¼strongly agree;M
overall
¼3.55,SD¼1.41, Mrange
Days 2-5
¼3.39–3.58; a
overall
¼.95, arange
Days2–5
¼.95–.96).
Coded measures
After study completion, the linguistic content of participants’ journal entries (Days 1–4)
was coded using two techniques. First, Linguistic Inquiry and Word Count (LIWC;
Pennebaker, Francis, & Booth, 2001) analyzed each entry for the percentage of words
out of the total that expressed cognitive processing. Second, two objective coders rated
the extent to which each entry contained qualities of redemptive narratives. Complete
descriptions are as follows.
LIWC provides a benefit beyond traditional linguistic coding practices in that the
program is not prone to a variety of biases that can be present with human coders
(Pennebaker et al., 2001; Tausczik & Pennebaker, 2010). Furthermore, LIWC pro-
vides reliable coding of text samples across time and different programs of research as
it identifies the percentage of words, out of the total, in a text sample that fall into
particular, predefined categories. Calculating the percentage of words out of the total
also provides the added benefit of accounting for variation in the total word count
across different individuals’ writing samples as well as writing samples over time in
the present work. Previous work validating LIWC as a coding paradigm also has
demonstrated that, among others, the category used in the present research of
cognitive processing (described below) can be extracted from text samples by the
program with high levels of predictive validity (Tausczik & Pennebaker, 2010). LIWC
does possess a serious limitation, however, of a lack of contextual sensitivity
(Tausczik & Pennebaker, 2010). Thus, LIWC was not an ideal choice to code for our
key variable of redemptive narratives. As described below, human coders were
utilized instead.
Cognitive processing. LIWC coded cognitive processing from the present writing
samples using a default dictionary, defined by program creators, that identified words
in our text samples relevant to rational thought processes (M
overall
¼7.45%of total
word counts,SD¼7.02, Mrange ¼6.84–7.72). The program flags all words relevant
to a variety of forms of cognitive processing (312 total in the category; Pennebaker
et al., 2001). The dictionary for category cognitive processing includes words relevant
Slotter and Ward 7
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

to causation (e.g., because/cause), insight (e.g., think/know), certainty (e.g., always/
never), tentative thinking (e.g., maybe/perhaps), inhibitions (e.g., constrain/restrain),
and discrepancies (i.e., should/would). Not only does the dictionary include all
iterations of relevant words (e.g., cause, caused, etc.), it also includes all synonyms of
the target constructs. The full dictionary for cognitive processing is available in the
manual for the LIWC program (Pennebaker et al., 2001). The subcategories of cau-
sation, insight, certainty, tentative thinking, inhibitions, and discrepancies are the
names used by the program itself and reflect different types of cognitive mechanisms.
All of the subcategories of the cognitive processing construct were highly correlated
with each other in the text samples. Thus, we opted to conduct the analyses using the
overarching construct.
Higher levels of cognitive processing as coded by LIWC have been linked to
enhanced emotional and physical well-being in a variety of samples (e.g., Pennebaker
& Francis, 1996; Pennebaker et al., 2003). Importantly, greater use of cognitive pro-
cessing, as coded by LIWC, can reflect attempts to cognitively reappraise the meaning
of emotional events as cognitive processing indicates effort to find meaning in emo-
tional events and view them in calmer, more rational ways (e.g., Lepore & Greenberg,
2002; North et al., 2012). Use of the LIWC cognitive processing dictionary to assess
attempts to apply cognitive reappraisal events has been successfully employed in
research examining progression of physical illness and positive growth following trau-
matic experiences and adjustment to college (e.g., Ullrich & Lutgendorf, 2002; Pen-
nebaker & Francis, 1996). On average in the current sample, cognitive processing did
not systematically change over the course of the daily diary study, B¼.28, t(247) ¼
0.96, p¼.34.
Redemptive narrative use. Two objective human coders, blind to study hypothesis, also
rated the entire set of text samples for redemptive narrative use. Redemptive narratives
reflect a form of thought and expression that individuals sometimes use when thinking,
talking, or writing about events in their lives (McAdams Reynolds, Lewis, Patten, &
Bowman, 2001). Specifically, redemptive narrative use reflects a search for and focuses
on meaning making, or the lessons learned, and positive outcomes that can stem from
negative events. In the current text samples, coders rated participants’ explicit statements
with regard to their attempts to find positive meaning in the end of their relationships and
generated a single redemptive narrative use score for each participant on each journaling
day of the study (e.g., McAdams, Reynolds, Lewis, Patten, & Bowman, 2001; 0 ¼
redemptive narrative not present and 6 ¼redemptive narrative extremely present;
M
overall
¼3.33,SD¼1.09, Mrange ¼2.32–3.66; ICC
overall
¼.78, ICC range
Days 2–5
¼
.78–.79). The two coders’ rating of each individual’s diary on each day was averaged to
create an index of the extent to which the diary represented a redemptive narrative, with
higher scores indicating greater redemptive narrative use. For example, one of the
statements made by a participant who received an indexed score of 5.5 on Day 3 of the
study was, ‘‘I am really sad that we broke up, but maybe it’s for the best. I am better off
without somebody who doesn’t treat me right.’’ On average in the current sample,
redemptive narrative use did not systematically change over the course of the daily diary
study, B¼.04, t(207) ¼0.97, p¼.33.
8Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Results
For each individual i, we ran a series of multilevel growth curve analyses employing a
maximum likelihood approach (Singer & Willett, 2003). Multilevel modeling was
necessary to accommodate the nesting of responses over time within individuals, which
violates the assumption of independence inherent in traditional ordinary least squares
regression. In the current models, time twas coded 0–3 for the journaling Days 1–4 and
all other variables were standardized (M¼0, SD ¼1). In all models, we opted to
standardize at the level of the group, rather than within individuals, as our key pre-
dictions hinged on differences in emotional distress between individuals based on their
relative journaling tendencies over the course of the study, rather than variation in
journaling tendencies over the course of the study within individuals. All variables were
Level-1, or time-varying, predictors. In all models, time was treated randomly as differ-
ent individuals had started the study on different days and thus were on different time-
tables for study completion; our other key predicted was treated as fixed. Additionally,
an unrestricted covariance structure was allowed as this structure fit the data better than
an autoregressive covariance structure (Bayesian Information Criterion [BIC] ¼711.8,
compared to an autoregressive (1) covariance structure with BIC ¼731.8 when exam-
ining the intercept-only model).
After standardizing, lagged variables were created for both redemptive narrative use
and cognitive processing as our key predictions hinged on the effects of yesterday’s
redemptive narratives and/or cognitive processing predicting today’s emotional distress
over the course of our 5-day diary study. Thus, to test Hypothesis 1, redemptive narrative
use at Day 1 (the first day of journaling) in our analyses was lagged and used to predict
emotional distress at Day 2, redemptive narrative use at Day 2 to predict Day 3 distress,
and so forth. To test Hypothesis 2, the same procedure was employed with regard to
cognitive processing.
To test Hypothesis 1, lagged redemptive narrative, time, study time, and their inter-
action effect were entered into a growth curve model predicting current emotional dis-
tress.
1
The model also included the main effect of redemptive narrative use on the current
day and its interaction with time. The starting model for testing Hypothesis 1 was
2
:
Time-varying emotional distressit ¼p0iþp1iðlagged redemptive narrative useit1Þ
þp2iðTimetÞþ þp3iðlagged redemptive narrative useit 1TimetÞ
þp4iðcurrent redemptive narrative useitÞ
þp5iðcurrent redemptive narrative useit TimetÞ
þp6iðintake emotional distressiÞþeit:
For Hypothesis 1, our key prediction hinged on a significant slope effect such that
individuals’ emotional distress would be reduced to the extent that they engaged in
greater redemptive narrative on the previous day, but that this effect would become
greater over the course of study time.
The starting model for testing Hypothesis 2 was essentially the same as for
Hypothesis 1 but with lagged cognitive processing rather than lagged redemptive
narrative use as the key variable. The starting model for testing Hypothesis 2 was:
Slotter and Ward 9
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Time-varying emotional distressit ¼p0iþp1iðlagged cognitive processingit1Þ
þp2iðTimetÞþp3iðlagged cognitive processingit1TimetÞ
p4iðcurrent cognitive processingitÞþp5iðcurrent cognitive processingit TimetÞ
þp6iðintake emotional distressiÞþeit:
For Hypothesis 2, our key hypothesis also hinged on a slope effect such that indi-
viduals’ emotional distress would be reduced, with this effect becoming greater over the
course of study time, to the extent that they had engaged in greater cognitive processing
on the previous day.
The full model for Hypothesis 3 was a composite of the Hypothesis 1 and Hypothesis
2 models. The Hypothesis 1 and Hypothesis 2 models were entered simultaneously,
controlling for previous day emotional distress, into the regression testing Hypothesis 3.
The full model for Hypothesis 3 was:
Time-varying emotional distressit ¼p0iþp1iðlagged redemptive narrative useit1Þ
þp2iðTimetÞþp3iðlagged redemptive narrative useit1TimetÞ
þp4iðlagged cognitive processingit1Þ
þp5iðlagged cognitive processingit1TimetÞ
þp6iðcurrent redemptive narrative useitÞ
þp8iðcurrent redemptive narrative useit TimetÞ
þp9iðcurrent cognitive processingit TimetÞþ
þp10iðintake emotional distressiÞþeit:
For Hypothesis 3, our key hypothesis was that the two interaction effects examined in
Hypothesis 1 and Hypothesis 2 would remain significant controlling for the other, thus
establishing the two processes as independent in predicting emotional distress. We also
conducted an auxiliary version of the Hypothesis 3 model in which we controlled for the
main effects of participant age (Age
i
) and biological sex (Sex
i
), relationship duration
(Relationship Duration
i
), weeks since the relationship dissolution at intake (Weeks Since
Dissolution
i
), who terminated the relationship dissolution (Dissolution Initiator
i
), par-
ticipants’ initial assignment to the failed writing manipulation (Condition
i
), and writing
word count (Word Count
it
).
Testing Hypothesis 1
As predicted, the interaction between lagged redemptive narrative use and study time
emerged as significant predicting current emotional distress (Table 2), B¼.07, t(183)
¼2.00, p< .05. Supporting Hypothesis 1, greater lagged redemptive narrative use
predicted less emotional distress over time (þ1SD; Figure 1; Aiken & West, 1991),
B¼.07, t(183) ¼2.33, p¼.02, whereas less lagged redemptive narrative use did not
(1SD), B¼.01, t(183) ¼0.58, p¼.56.
We also examined our simple effects differently, by examining the effect of lagged
redemptive narrative use at the beginning versus end of our study, rather than looking at
10 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Table 2. Full growth curve models for Hypotheses 1, 2, and 3 with covariates
Hypothesis Parameter df B t
Hypothesis 1 Intercept 124 .06 0.98
BIC ¼465.2 Lagged redemptive narrative (LRN) 183 .02 0.61
Study Time (T) 183 .05 2.23*
LRN T 183 .05 2.00**
Current day RN (CRN) 183 .12 2.48*
CRN T 183 .03 1.36
Intake distress 124 .81 15.54^
Hypothesis 2 Intercept 143 .06 1.21
BIC ¼583.4 Lagged cognitive processing (LCP) 241 .04 1.17
T 241 .04 2.28*
LCP T 241 .04 2.00*
Current day CP (CCP) 241 .01 0.20
CCP T 241 .01 0.32
Intake distress 143 .78 16.57^
Hypothesis 3 Intercept 124 .05 0.80
LRN 177 .04 0.97
BIC ¼471.8 T 177 .03 1.86
LCP 177 .06 1.23
LRN T 177 .04 2.06*
LCP T 177 .04 1.84
CRN 177 .12 2.37*
CRN T 177 .05 1.92
CCP 177 .12 1.76
CCP T 177 .08 1.58
Intake distress 124 .80 15.27^
Hypothesis 3 Intercept 110 .19 1.32
With covariates LRN 168 .04 0.80
BIC ¼475.1 T 168 .03 1.61
LCP 168 .07 1.48
LRN T 168 .04 2.02*
LCP T 168 .01 1.94
CRN 168 .08 1.57
CRN T 168 .04 1.55
CCP 168 .14 2.02*
CCP T 168 .09 1.72
Intake distress 110 .79 14.00^
Age 110 .00 0.04
Sex 110 .01 0.14
Relationship duration 110 .01 0.10
Breakup initiator 110 .05 0.87
Weeks since breakup 110 .04 0.76
Word count 168 .00 1.73
Condition 110 .07 1.21
**p ¼.05. *p < .05. ^p <.01.
Slotter and Ward 11
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

the slope of distress over time as a function of lagged redemptive narrative use. Inter-
estingly, the effect of lagged redemptive narrative use was dependent on study time. At
the beginning of the study (Day 2, as journaling began on Day 1), lagged redemptive
narrative use did not predict emotional distress. B¼.003, t(183) ¼0.10, p¼.92;
however, at the end of the study (Day 4) greater lagged redemptive narrative use pre-
dicted less distress, B¼.06, t(183) ¼2.04, p¼.04.
Testing Hypothesis 2
As predicted, the interaction between lagged cognitive processing and study time
emerged as significant predicting current emotional distress (Table 2), B¼.04,
t(241) ¼2.00, p< .05. Supporting Hypothesis 2, greater lagged cognitive processing
predicted less emotional distress over time (þ1SD; Figure 2), B¼.08, t(241) ¼3.08,
p< .05, whereas less lagged cognitive processing did not (1SD), B¼.005,
t(241) ¼.18, p¼.85.
We also examined our simple effects differently, by examining the effect of lagged
cognitive processing at the beginning versus end of our study, rather than looking at the
slope of distress over time as a function of lagged cognitive processing. Interestingly, the
effect of lagged cognitive processing was dependent on study time. At the beginning of
the study (Day 2, as journaling began on Day 1), lagged cognitive processing did not
predict emotional distress, B¼.03, t(241) ¼0.95, p¼.34; however, at the end of the
study (Day 4) greater lagged cognitive processing predicted marginally less distress,
B¼.06, t(241) ¼1.86, p¼.06.
Testing Hypothesis 3
As predicted, the relevant interaction of lagged redemptive narrative use by time,
B¼.04, t(177) ¼2.06, p¼.04, emerged as significant. However, contrary to the
Figure 1. The effect of lagged redemptive narrative use over study time on emotional distress in
the Hypothesis 1 model.
12 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

predictions, the effect of lagged cognitive processing by time did not, B¼.04, t(177) ¼
1.84, p¼.06 (Table 2). Tests of simple effects revealed the same patterns of effects
and statistical significance levels reported above for lagged redemptive narrative use
(Figure 3). Given the nonsignificant effect of lagged cognitive processing by time, we
did not deconstruct this interaction.
Auxiliary covariate analyses
We conducted a series of auxiliary analyses on the Hypothesis 3 model. We present the
key lagged redemptive narrative use over time and lagged cognitive processing over time
effects and the main and moderating effects of each covariate for each of our auxiliary
Figure 2. The effect of lagged cognitive processing over study time on emotional distress in the
Hypothesis 2 model.
Figure 3. The effect of lagged redemptive narrative use over study time on emotional distress in
the Hypothesis 3 model.
Slotter and Ward 13
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

analyses in Table 2.
3
The interaction between lagged redemptive narrative use and study
time remained significant in this stringent test, p¼.04, and the patterns of effects
remained robust. Additionally, none of the covariates contributed significantly to the
model, p> .18. Finally, we examined whether lagged cognitive processing would
moderate the interaction among lagged redemptive narrative use and study time. It did
not, p¼.15. Lagged redemptive narrative use and lagged cognitive processing were
uncorrelated, r¼.05, p¼.34.
Interpreting the magnitude of our effects
When examining the effects of lagged redemptive narrative use or cognitive pro-
cessing over time, we sought to explicitly interpret the reduction in distress implied by
our three key models. For the model testing Hypothesis 1 (Figure 1), the magnitude of
participants’ distress reduction from Day 2 to Day 4 of the study if they were þ1SD
on lagged redemptive narrative use was a 0.26 SDs on the CES-D. For participants
who were 1SD on lagged redemptive narrative use that reduction was zero. Stated
differently, the difference in distress for participants þ1 versus 1SD from the mean
on lagged redemptive narrative use on Day 2 of the study was 0.02 SDs on the CES-
D—as noted above this difference was not statistically significant. On Day 4 of the
study, however, this difference was 0.20 SDs on the CES-D. Although statistically
significant, this difference may not appear large from a practical standpoint. That said,
if we extrapolate our model forward, the model implied difference in distress for par-
ticipants þ1 versus 1SD from the mean on lagged redemptive narrative use on Day
7 would be 0.48 SDs on the CES-D, and on Day 30 would be 2.78 SDs on the
CES-D.
Similarly, for the model testing Hypothesis 2 (Figure 2), the magnitude of partici-
pants’ distress reduction from Day 2 to Day 4 of the study if they were þ1SD on lagged
cognitive processing was a 0.16 SDs on the CES-D. For participants who were 1SD on
lagged redemptive narrative use that reduction was zero. Stated differently, the differ-
ence in distress for participants þ1 versus 1SD from the mean on lagged redemptive
narrative use on Day 2 of the study was 0.02 SDs on the CES-D—as noted above this
difference was not statistically significant. On Day 4 of the study, however, this differ-
ence was 0.16 SDs on the CES-D. Although statistically significant, this difference
may not appear large from a practical standpoint. That said, if we extrapolate our model
forward, the model implied difference in distress for participants þ1 versus 1SD from
the mean on lagged redemptive narrative use on Day 7 would be 0.40 SDs on the CES-
D, and on Day 30 would be 2.24 SDs on the CES-D.
Finally, for the model testing Hypothesis 3 (Figure 3), which was our most stringent
model, we examined the magnitude of our significant interaction between lagged
redemptive narrative use and time only. Again, the magnitude of participants’ distress
reduction from Day 2 to Day 4 of the study if they were þ1SD on lagged redemptive
narrative use was a 0.14 SDs on the CES-D. For participants who were 1SD on lagged
redemptive narrative use, there was actually a slight increase of 0.02. Stated differently,
the difference in distress for participants þ1 versus 1 SD from the mean on lagged
redemptive narrative use on Day 2 of the study was 0.02 SDs on the CES-D—as noted
14 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

above this difference was not statistically significant. On Day 4 of the study, however,
this difference was 0.16 SDs on the CES-D. Although statistically significant, this dif-
ference may not appear large from a practical standpoint. That said, if we extrapolate our
model forward, the model implied difference in distress for participants þ1 versus 1
SD from the mean on lagged redemptive narrative use on Day 7 would be 0.35 SDs
on the CES-D, and on Day 30 would be 1.96 SDs on the CES-D.
Across the tests of our three hypotheses, the effect of our predictors over time sug-
gests that the contribution of lagged redemptive narrative use or cognitive processing is
cumulative. The differences in distress between individuals high versus low in our
predictors start small but grow quickly. Although our brief diary study did not include
the later time points for any of our three models, the cumulative nature of our observed
effects suggests that small decreases in distress may build over time, thus making our
effects of potential practical significance for individuals who have recently experienced
the end of a romantic relationship.
Discussion
In the present study, we examined the lagged effects of redemptive narrative use and
cognitive processing in the writing samples of adults who had recently experienced the
end of a romantic relationship. We found that greater lagged redemptive narrative use
and greater lagged cognitive processing predicted less emotional distress over time.
Importantly, the effects of lagged redemptive narrative use and lagged cognitive pro-
cessing were dependent on study time. After only 1 day of journaling, lagged redemptive
narrative use and lagged cognitive processing did not predict reduced distress. After
4 days of journaling, greater redemptive narrative use predicted less distress and greater
cognitive processing predicted marginally less distress. Overall, the results from the
present study supported our hypotheses by demonstrating that—when examined inde-
pendently—greater, versus lesser, use of redemptive narratives over time predicted less
emotional distress after the end of a romantic relationship (Hypothesis 1), and greater,
versus lesser, cognitive processing over time predicted less emotional distress after the
end of a romantic relationship (Hypothesis 2). When examined simultaneously,
redemptive narrative use over time remained a significant predictor of reduced distress
and appears to operate independently of cognitive processing (Hypothesis 3). That said,
in our test of Hypothesis 3, cognitive processing over time did not remain a significant
predictor of reduced distress.
Theoretically, these findings add to our understanding of how individuals emotionally
recover after the loss of a romantic relationship as well as enhancing our understanding
of the distinction between two cognitive-based coping mechanisms for dealing with
potentially distressing life events. Although the present research could not determine the
specific psychological mechanisms by which redemptive narratives and cognitive
reappraisal function in the present study, the present research suggests that they are two
distinct cognitive processes. Both redemptive narrative use and cognitive processing
independently predicted reduced emotional distress in the current, even when analyzed
in the same statistical model. This independence argues for two distinct processes that
can aid individuals in their emotional recovery from a relationship loss. Perhaps viewing
Slotter and Ward 15
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

negative events in the context of their positive outcomes can aid individuals in reinter-
preting the meaning of these events through a similar, yet complementary, process to
cognitive reappraisal. Future research would benefit from further examining the links
between these two approaches to thinking and writing about significant life events.
The present research also contributed to our understanding of the temporal process
behind the association between redemptive narrative use and cognitive reappraisal and
reduced emotional distress after the end of a romantic relationship. Specifically, the
current research suggests that these processes can be beneficial for individuals even
when engaged in for short periods of time. The present study asked non-journaling
participants to write for up to 10 min a day for 4 days. In this time, greater redemptive
narrative use and cognitive processing predicted declining feelings of emotional distress.
Furthermore, the effects of these processes on distress appeared as though it might be
cumulative. No differences in distress emerged until the end of the present study. This
suggests that, although redemptive narrative use and cognitive processing can benefit the
individuals using them very quickly, they may take a few days to confer their benefits.
Practically, the findings from the present study suggest a potential point of intervention
for helping individuals cope with the loss of a romantic relationship. Although the
reductions in distressevidenced in our data started small, they appeared to accumulate over
time, suggesting not only statistically significant but also practically significant reductions
in distress. Additionally, our data evidenced reductions in distress on a well-validated clin-
ical measure of depression, the CES-D. We believe that reductions in distress on such a
measure, even within a nonclinical range, represent important fluctuations in individuals’
psychological well-being. Additionally, given that the loss of a romantic relationship is
typically a highly distressing event for individuals (e.g., Boelen & Reijntjes, 2009;
Rhoades et al., 2011), which is associated with major depression disorder (e.g., Monroe
et al., 1999), and depressive symptomology has broad reaching negative personal and soci-
etal consequences (National Institutes of Mental Health, 2014), any reduction in the
depression-related distress that individuals experience after the end of a relationship has
the potential to enhance overall well-being and functioning. We believe that the redemp-
tive narratives examined in the present work may present an opportunity to give individ-
uals a tool with which they can reduce their post-dissolution distress.
Future work should examine the precise mechanism behind why redemptive narrative use
confers benefits to well-being. This may occur because redemptive narratives help individuals
to ‘‘make meaning’’ out of a negative situation. According to the meaning-making model,
recovering from a stressful event involves reducing the discrepancy between the initial
appraised meaning of some event and the appraiser’sgeneral beliefs about the world and goals
within it (Park, 2010). The individuals in the present study began writing about their ex-
relationship at our request; none reported keeping a journal or diary prior to beginning our
study. Thus, the beneficial effects of redemptive narrative use and cognitive processing as
coded from their writing samples over just 4 days suggests that asking participants to write
about the end of their relationships, and perhaps other life events, could be a simple, cost-
effective way at aiding individuals in emotionally recovering. Indeed, the largest meta-
analysis to date of the impact expressive writing has on emotional distress revealed that
expressive writing generally is associated with moderate increases on self-reported and
objective markers of well-being—including reduced depressive symptomology, even
16 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

following negative events such as losing a job or a relationship (Frattaroli, 2006; Gortner,
Rude, & Pennebaker, 2006; Pennebaker & Chung, in press). Additionally, compared
to control groups, participants who engage in expressive writing tasks also have shown
improved functioning in their immune system as well as improvement on other biolo-
gical markers of stress (Gortner et al., 2006).
Although the present study did not experimentally encourage versus discourage partici-
pants from using redemptive narratives, future studies should investigate whether this specific
type of storytelling could be developed and honed as a way of treating post-dissolution dis-
tress. Future research also should look at boundaries of writing-focused interventions in
enhancing well-being after a negative event, especially in light of recent work demonstrating
that expressive writing tasks can enhance distress in certain individuals, specifically those
prone to engaging in ruminative thinking, and thus be detrimental to well-being following
marital separation (Sbarra et al., 2013). In the present study, participants were simply asked
to write about the end of their relationship, without specific instructions regarding emotionally
expressive writing or rumination versus reappraisal; however, investigating if/when redemp-
tive narrative use could actually predict negative well-being outcomes seems to be a crucial
investigation to pursue. It is possible that, for some individuals, attempts at redemptive nar-
rative use could induce rumination rather than a focus on positive outcomes.
Limitations and additional avenues for future work
Despite these strengths and implications of the work, the present study possessed some
limitations, and the prospect of addressing them presents exciting avenues for future
research. For example, the effect of lagged cognitive processing by time in the test of
Hypothesis 3 only emerged as marginally significant. This may be due to the differential
sensitivity betweenthe LIWC program and human coders.We think the LIWC may nothave
captured contextual sensitivity as well as the human coders captured the use of redemptive
narratives. Further research should consider using the same method of narrative coding
when comparing different types of cognitive processing in the context of narrative
sequences. Additionally, although we view the duration of the study as one of its strengths,
as it fills a gap in our understanding of the temporal effects of redemptive narrative use and
cognitive reappraisal, it is worth re-stating that the present study was only 4 days in duration.
Thus, future research should investigate the time course of the effects from the present
research. Finally, the present research did not find evidence in the data for the mechanism
by which redemptive narrative use or cognitive processing reduced distress. Identifying the
precise process by which redemptive narrative use and cognitive processing enhance well-
being are both important for determining the usefulness of this process for interventions.
Conclusions
Taken together, the present research demonstrated that finding the silver lining in the
dark cloud of relationship dissolution aids individuals in reducing their emotional
distress. That this effect emerged independent from the processing associated with cog-
nitive reappraisal suggests that redemptive narratives may provide a novel pathway in
helping individuals emotionally recover after the loss of a romantic relationship. Given the
Slotter and Ward 17
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

negative effects of relationship dissolution on individuals’ physical and psychological
well-being, identifying new ways of providing relief is a worthwhile endeavor.
Authors’ Note
This work was conducted in accordance with the APA standards. This work was supported by an
internal Faculty Development Grant to the corresponding author from Villanova University. Inqui-
ries regarding this research should be addressed to Erica B. Slotter at erica.slotter@villanova.edu.
Funding
This research received no specific grant from any funding agency in the public, commercial, or
not-for-profit sectors.
Notes
1. For the models examining Hypotheses 1 and 2, we also investigated the reverse pathway, that is,
whether lagged distress would predict either redemptive narrative use or cognitive processing
either as a main effect or over time. Neither lagged distress nor its interaction with time pre-
dicted either redemptive narrative use or cognitive processing.
2. For all three models, we also conducted our analyses without controlling for Intake Emotional
Distress. All key effects remained robust. We opted to include Intake Emotional Distress in our
final models as doing so improved model fit using a Bayesian Information Criterion (BIC) for
each model (Hypothesis 1: BIC without control ¼592.9, with control ¼462.2; Hypothesis 2:
BIC without control ¼730.6, with control ¼583.4; and Hypothesis 3: BIC without control ¼
598.1, with control ¼471.8).
3. Although notpresented in Table 2 for the sake of clarity, wealso examined whether ourkey effects
would be moderated by any of our covariates. No significant moderational effects emerged.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions.
Newbury Park: Sage.
Bauer, J. J., McAdams, D. P., & Pals, J. L. (2008). Narrative identity and eudaimonic well-being.
Journal of Happiness Studies,9, 81–104.
Boals, A., & Klein, K. (2005). Word use in emotional narratives about failed romantic relation-
ships and subsequent mental health. Journal of Language and Social Psychology,24, 252–268.
Boelen, P., & Reijntjes, A. (2009). Negative cognitions in emotional problems following romantic
relationship break-ups. Stress and Health,1, 11–19.
Bowlby, J. (1980). Attachment and loss: Volume 3. Loss: Sadness and depression. New York, NY:
Basic.
Davis, D., Shaver, P. R., & Vernon, M. L. (2003). Physical, emotional, and behavioral reactions to
breaking up: The roles of biological sex, age, emotional involvement, and attachment style.
Personality and Social Psychology Bulletin,29, 871–884.
Denson, T. F., Moulds, M. L., & Grisham, J.R. (2012). The effects of analytical rumination, reap-
praisal, and distraction on anger experience. Behavior Therapy,43, 355–364.
Feeney, J. A., & Noller, P. (1992). Attachment style and romantic love: Relationship dissolution.
Australian Journal of Psychology,44, 69–74.
18 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Finkel, E. J., Slotter, E. B., Luchies, L. B., Walton, G. M., & Gross, J. (2013). A brief intervention
to promote conflict reappraisal preserves marital quality over time. Psychological Science,24,
1595–1601.
Frattaroli, J. (2006). Experimental disclosure and its moderators: A meta-analysis. Psychological
Bulletin,132, 823–865.
Gortner, E., Rude, S. S., & Pennebaker, J. W. (2006). Benefits of expressive writing in lowering
rumination and depressive symptoms. Behavior Therapy,37, 292–303.
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. (2002). Emotion regulation: Affective, cognitive, and social consequences. Psychophy-
siology,39, 281–291.
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 Psy-
chology,85, 348–362.
Lazarus, R. S., & Alfert, E. (1964). Short-circuiting of threat by experimentally altering cognitive
appraisal. Journal of Abnormal and Social Psychology,69, 195–205.
Lazarus, R., & Folkman, S. (1985). Stress, appraisal, and coping. New York, NY: Springer.
Lepore, S. J., & Greenberg, M. A. (2002). Mending broken hearts: Effects of expressive writing on
mood, cognitive processing, social adjustment and health following a relationship breakup.
Psychology and Health,17, 547–560.
Lucas, R. E. (2005). Time does not heal all wounds: A longitudinal study of reaction and adapta-
tion to divorce. Psychological Science,16, 945–950.
McAdams, D. P., Reynolds, J., Lewis, M., Patten, A., & Bowman, P. J. (2001). When bad things
turn good and good things turn bad: Sequences of redemption and contamination in life narra-
tive, and their relation to psychosocial adaptation in midlife adults and in students. Personality
and Social Psychology Bulletin,27, 472–483.
McLean, K. C., & Pratt, M. W. (2006). Life’s little (and big) lessons: Identity statuses and
meaning-making in turning point narratives of emerging adults. Developmental Psychology,
42, 714–722.
Monroe, S. M., Rohde, P., Seeley, J. R., & Lewinsohn, P. M. (1999). Life events and depression in
adolescence: Relationship loss as a prospective risk factor for first onset of major depressive
disorder. Journal of Abnormal Psychology,108, 606–614.
National Institutes of Mental Health. (2014). The numbers count: Mental disorders in America.
Retrieved January 30, 2014 from http://www.nimh.nih.gov/health/publications/the-numbers-
count-mental-disorders-in-america/index.shtml.
North, R. J., Meyerson, R. L., Brown, D. N., & Holahan, C. J. (2012). The language of psycholo-
gical change: Decoding and expressive writing paradigm. Journal of Language and Social
Psychology,32, 142–161.
Pals, J. L. (2006). The narrative identity processing of difficult life experiences: Pathways of
personality development and positive self-transformation in adulthood. Journal of Personality,
74, 1079–1110.
Park, C. (2010). Making sense of the meaning literature: An integrative review of meaning mak-
ing and its effects on adjustment to stressful life events. Psychological Bulletin,136, 257–301.
Pennebaker, J. W., & Beall, S. K. (1986). Confronting a traumatic event: Toward and understand-
ing of inhibition and disease. Journal of Abnormal Psychology,95, 274–281.
Slotter and Ward 19
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from

Pennebaker, J. W., & Francis, M. E. (1996). Cognitive, emotional, and language processes in
disclosure. Cognition & Emotion,10, 601–626.
Pennebaker, J. W., Francis, M. E., & Booth, R. J. (2001). Linguistic Inquiry and Word Count
(LIWC): A computerized text analysis program. Mahwah, NJ: Erlbaum Publishers.
Pennebaker, J., Mehl, M., & Niederhoffer, K. (2003). Psychological aspects of natural language
use: Our words, our selves. Annual Review of Psychology,54, 547–577.
Pennebaker, J. W., & Chung, C. K. (2011). Expressive writing and its links to mental and physical
health. In H. S. Freidman (Ed.), Oxford handbook of health psychology (pp. 417–437).
New York, NY: Oxford University Press.
Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general
population. Applied Psychological Measurement,1, 385–401.
Rhoades, G. K., Kamp Dush, C. M., Atkins, D. C., Stanley, S. M., & Markman, H. J. (2011).
Breaking up is hard to do: The impact of unmarried relationship dissolution on mental health
and life satisfaction. Journal of Family Psychology,25, 366–374.
Richards, J. M., Butler, E. A., & Gross, J. J. (2003). Emotion regulation in romantic relationships:
The cognitive consequences of concealing feelings. Journal of Social and Personal Relation-
ships,20, 599–620.
Sbarra, D. A., Boals, A., Mason, A. E., Larson, G., & Mehl, M. R. (2013). Expressive writing can impede
emotional recovery following marital separation. Clinical Psychological Science,1, 120–134.
Sbarra, D. A., Law, R. W., & Portley, R. M. (2011). Divorce and death: A meta-analysis
and research agenda for clinical, social, and health psychology. Perspectives on Psychological
Science,6, 454–474.
Sbarra, D. A, Smith, H. L., & Mehl, M. R. (2012). When leaving your ex, love yourself: Observa-
tional ratings of self-compassion predict the course of emotional recovery following marital
separation. Psychological Science,23, 261–269.
Simpson, J. A. (1987). The dissolution of romantic relationships: Factors involved in relationship
stability and emotional distress. Journal of Personality and Social Psychology,53, 683–692.
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis. New York, NY: Oxford
University Press.
Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC and
computerized text analysis methods. Journal of Language and Social Psychology,29, 24–54.
Tejada-Vera, B., & Sutton, P. D. (2010). Births, marriages, divorces, and deaths: Provisional data
for 2009 (National Vital Statistics Reports No. 58). Hyattsville, MD: National Center for
Health Statistics.
Troy, A. S., Wilhelm, F. H., Shallcross, A. J., & Mauss, I. B. (2010). Seeing the silver lining: Cog-
nitive reappraisal moderates the relationship between stress and depressive symptoms. Emo-
tion,10, 783–795.
Tugade, M. M. (2011). Positive emotions and coping: Examining dual-process models of resili-
ence. In S. Folkman (Ed.), The Oxford handbook of stress, health, and coping (pp. 186–199).
New York, NY: Oxford University Press.
Ullrich, P. M., & Lutgendorf, S. K. (2002). Journaling about stressful events: Effects of cognitive
processing and emotional expression. Annals of Behavioral Medicine,24, 244–250.
Watt, L. M., & Cappeliez, P. (2000). Integrative and instrumental reminiscence therapies for
depression in older adults: Intervention strategies and treatment effectiveness. Aging & Mental
Health,4, 166–177.
20 Journal of Social and Personal Relationships
at VILLANOVA UNIV on January 8, 2015spr.sagepub.comDownloaded from