Bono 000

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/5784151
Forgiveness, Feeling Connected to Others, and Well-Being: Two Longitudinal

Studies
Article in Personality and Social Psychology Bulletin · March 2008

DOI: 10.1177/0146167207310025 · Source: PubMed
CITATIONS READS
299 2,124
3 authors:
Giacomo Bono Michael E Mccullough

California State University, Dominguez Hills University of Miami
34 PUBLICATIONS 2,366 CITATIONS 158 PUBLICATIONS 25,648 CITATIONS
SEE PROFILE SEE PROFILE
Lindsey Root Luna

Hope College
21 PUBLICATIONS 1,576 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Youth Gratitude Project View project
Gratitude in the Time of Covid-19 View project
All content following this page was uploaded by Michael E Mccullough on 02 June 2014.
The user has requested enhancement of the downloaded file.

Personality and Social Psychology Bulletin
http://psp.sagepub.com
Forgiveness, Feeling Connected to Others, and Well-Being: Two Longitudinal Studies

Giacomo Bono, Michael E. McCullough and Lindsey M. Root
Pers Soc Psychol Bull 2008; 34; 182 originally published online Dec 6, 2007;
DOI: 10.1177/0146167207310025
The online version of this article can be found at:

http://psp.sagepub.com/cgi/content/abstract/34/2/182
Published by:
http://www.sagepublications.com
On behalf of:
Society for Personality and Social Psychology, Inc.
Additional services and information for Personality and Social Psychology Bulletin can be found at:
Email Alerts: http://psp.sagepub.com/cgi/alerts
Subscriptions: http://psp.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations (this article cites 38 articles hosted on the

SAGE Journals Online and HighWire Press platforms):
http://psp.sagepub.com/cgi/content/refs/34/2/182
Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

© 2008 Society for Personality and Social Psychology, Inc.. All rights reserved. Not for commercial use or unauthorized distribution.
Forgiveness, Feeling Connected to Others,
and Well-Being: Two Longitudinal Studies
Giacomo Bono
Michael E. McCullough
Lindsey M. Root
University of Miami
In two studies, the authors investigated the associations we have conceptualized forgiveness as prosocial changes
between interpersonal forgiveness and psychological in people’s transgression-related interpersonal motiva-
well-being. Cross-sectional and prospective multilevel tions toward a transgressor. That is, when people for-
analyses demonstrated that increases in forgiveness (mea- give a transgressor, they become less motivated by
sured as fluctuations in individuals’ avoidance, revenge, revenge and avoidance and more motivated by benevo-
and benevolence motivations toward their transgressors) lence toward the transgressor (McCullough et al., 1998;
were related to within-persons increases in psychological McCullough, Worthington, & Rachal, 1997). In cross-
well-being (measured as more satisfaction with life, more sectional and longitudinal studies, forgiveness has been
positive mood, less negative mood, and fewer physical associated with improved interpersonal relationships
symptoms). Moreover, forgiveness was more strongly between the forgiver and the transgressor (Karremans
linked to well-being for people who reported being closer & Van Lange, 2004; McCullough et al., 1998; Tsang,
and more committed to their partners before the trans- McCullough, & Fincham, 2006). In addition to its
gression and for people who reported that their partners apparent relational benefits, forgiveness is also posi-
apologized and made amends for the transgression. tively associated with psychological well-being (Brown,
Evidence for the reverse causal model, that increases in 2003; Freedman & Enright, 1996; Karremans, Van
well-being were related to increases in forgiveness, was Lange, Ouwerkerk, & Kluwer, 2003; Poloma &
also found. However, changes in feelings of closeness Gallup, 1991). The link of forgiveness to these two sets
toward the partner appeared to account for the associa- of outcomes (relational and psychological) may be more
tions of forgiveness with well-being, but not vice versa. than coincidental: Perhaps forgiveness obtains its asso-
ciation with psychological well-being precisely because
Keywords: forgiveness; closeness; well-being; apology; forgiveness helps people maintain and restore close rela-
longitudinal tionships (Karremans et al., 2003).
Forgiveness, Well-Being, and Social Connectedness

W hen negotiating interpersonal relationships, people
experience many motives and preferences, some in
service of immediate self-interest and some in service of
Maintaining supportive relationships with kin and
nonkin is important for mental and physical health in
broader relationship concerns. Sometimes, though, the present day (Baumeister & Leary, 1995; House,
people’s self-interested motives and preferences may not
correspond to those of a given relationship partner (de
Waal, 2000; Kelley & Thibault, 1978; Rusbult, Verette, Authors’ Note: This research was generously supported by grants
Whitney, Slovik, & Lipkus, 1991). Thus, conflicts from the Campaign for Forgiveness Research and National Institute of
Mental Health R01MH071258. Correspondence regarding this arti-
inevitably arise and challenge relationships. By forgiving cle may be sent to Michael McCullough, Department of Psychology,
each other, it appears, people are able to overcome the University of Miami, P.O. Box 248185, Coral Gables, FL 33124-
negative effects conflict can have on their relationships 0751; e-mail: mikem@miami.edu.
(e.g., Fincham, 2000; McCullough et al., 1998). PSPB, Vol. 34 No. 2, February 2008 182-195
In our program of research (McCullough, Bono, & DOI: 10.1177/0146167207310025
Root, 2007; McCullough, Fincham, & Tsang, 2003), © 2008 by the Society for Personality and Social Psychology, Inc.
182

Bono et al. / FORGIVENESS AND WELL-BEING 183
Landis, & Umberson, 1988), and these salutary effects relationships that result) that should lead to the most
may mirror the crucial role that such relationships dramatic improvements in psychological well-being.
played in human evolution (Hamilton, 1964; Nesse, In this light, factors besides the victim’s degree of
2001; Trivers, 1971). The apparent link of forgiveness commitment to the transgressor are likely to be impor-
to psychological well-being may therefore be due to the tant also. One important factor to consider is the extent
fact that low psychological well-being can function as a to which transgressors apologize and make amends for
type of “information” that alerts people to relational their transgressions. Ample evidence suggests that apolo-
difficulties and creates motivation for improving gies can mitigate anger and yield constructive reactions
impaired social relations. This way of thinking about to harm and forgiveness (Hodgins & Liebeskind, 2003;
the relationship of forgiveness, psychological well- Ohbuchi, Kameda, & Agarie, 1989; Zechmeister,
being, and relational well-being is similar to the way Garcia, & Romero, 2004). Apologies are among the
that self-esteem can be conceptualized as an indicator of most robust facilitators of forgiveness (e.g., McCullough
lack of social acceptance (Leary, 2004), or that loneli- et al., 1997)—an effect that may be due in part to their
ness can be viewed as feedback designed to prompt ability to convince victims that their transgressors pos-
people to remediate the hedonic and cognitive conse- sess goodwill for them (despite having previously hurt
quences of social isolation (Cacioppo et al., 2006). Self- them). Thus, in the absence of an apology, forgiveness
determination theory provides a complementary (and may be less conducive for psychological well-being
more proximally focused) explanation: Forgiveness because the relationship may still be marked by uncer-
might obtain its association to well-being by influencing tainty about a transgressor’s intentions toward the vic-
the forgiver’s perceptions of his or her relationship with tim (and thus, their future value as a relationship
the transgressor because maintaining a proper degree of partner). As a result of this lingering uncertainty about
connectedness to others is a fundamental psychological a transgressor’s intentions toward the victim, forgive-
need (Reis, Sheldon, Gable, Roscoe, & Ryan, 2000). ness of an unapologetic transgressor would be less
Because forgiveness involves reductions of negative strongly related to psychological well-being.
motivations and the re-establishment of positive motiva-
tions toward transgressors, we posit that forgiveness The Current Investigation
helps people to regain the perception that their relation-
ships with their offenders are characterized by closeness We conducted two studies to investigate how tempo-
and commitment (Tsang et al., 2006). Therefore, we pro- rary changes in forgiveness—operationalized as transient
pose that forgiving a relationship partner causes people fluctuations in avoidance, revenge, and benevolence
to perceive that they have “reconnected” to an important motivations (McCullough et al., 2003; McCullough
source of social support and can again take advantage of et al., 2007)—are longitudinally related to well-being.
the material and emotional resources that supportive Using multilevel random coefficient models (Bryk &
social ties can confer. As a result, well-being improves. Raudenbush, 1992; Hedeker, 2004), we sought to extend
the existing insights on the relationship between forgive-
Forgiveness, Restored Closeness, and Well-Being: ness and psychological well-being in several ways. First,
A Function of Pretransgression Closeness and to date, all the work on the relationship between forgive-
Transgressor Apology? ness and well-being has focused on individual differences
between persons rather than intra-individual differences
The ability of forgiveness to facilitate psychological that occur within persons over time. In light of recent
well-being, however, may depend on the quality of rela- research demonstrating forgiveness to be a time-varying
tionship partner one is forgiving. Karremans and col- phenomenon (McCullough et al., 2003; McCullough
leagues (2003) found that forgiving was more strongly et al., 2007), use of evidence of within-persons (rather
associated with psychological well-being for people than between-persons) associations are required to best
who forgave a relationship partner to whom they were test whether forgiveness may lead to greater well-being.
highly committed than for people who forgave a rela- Thus, we used recent methods for studying the within-
tionship partner to whom they were not highly commit- persons associations among variables (Nezlek, 2001) to
ted. This comports well with the theory of reciprocal examine if psychological well-being is greater for people
altruism (Trivers, 1971) as well as with self-determination on days when they are also more forgiving than is typical
theory (Ryan & Deci, 2000), as it is relationships that for them and, insofar as it is, whether this association
are close and committed prior to the transgression that appears to be mediated by increased perceptions of close-
hold the most potential value to the forgiver. As a ness and commitment to the transgressor.
result, if our informational view of psychological well- Second, we examined whether these intra-individual
being here is accurate, then it is the forgiveness of close, relationships between forgiveness and psychological well-
committed relationship partners (and the improved being were stronger for people who felt relatively close

184 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
and committed to their relationship partners before the Benevolence subscale has 6 items measuring concilia-
transgression (Karremans et al., 2003). Third, because tory motivation toward the transgressor (e.g., “Despite
research has shown that apologies foster forgiveness and what he/she did, I want us to have a positive relation-
reconciliation, we also investigated whether individual ship again”). The subscales had adequate internal con-
differences in the degree to which transgressors were per- sistency and test-retest stability across all measurement
ceived to apologize and make amends for their transgres- occasions (i.e., alphas ranged from .85 to .95, and rs
sions moderated the within-person associations among ranged from .54 to .93). The scale has also shown good
forgiveness, closeness vis-à-vis the transgressor, and well- convergent and discriminant validity (McCullough,
being. In Study 1 we examined the concurrent associa- Bellah, Kilpatrick, & Johnson, 2001; McCullough
tions among these variables, and in Study 2 we conducted et al., 1998; McCullough et al., 2003; McCullough &
prospective, cross-lagged analyses to examine possible Hoyt, 2002).
causal relations among these variables.
Perceived transgression painfulness. Participants
completed a single item having them rate on a 7-point
STUDY 1 Likert-type scale how painful they perceived the trans-
gression to be (ranging from 0 = not painful at all to
METHOD 6 = worst pain I ever felt) in the initial survey.1
Participants and Procedure Psychological well-being. In this article, we define
Participants were 115 students in undergraduate psy- well-being in the hedonic tradition as a person’s subjec-
chology courses (91 women; M age = 19.76, SD = 2.61) tive assessment of the overall presence of pleasure and
at Southern Methodist University. We identified individ- absence of pain, commonly defined as happiness, and
uals who had experienced a serious interpersonal trans- viewed well-being as embodying satisfaction with life,
gression within the past 7 days through in-class experiencing positive emotions more and negative emo-
solicitations (and a screening instrument). Qualified par- tions less, and experiencing fewer physical symptoms
ticipants received preliminary questionnaires. After (Reis et al., 2000). Psychological well-being was mea-
returning these questionnaires, they were scheduled to sured with the combination of four scales in both initial
complete up to five follow-up questionnaires in the sec- and follow-up surveys. The extent to which participants
ond author’s laboratory, approximately every 2 weeks on felt satisfied with their lives was measured by the
the same weekday and time (ideally). They received extra Satisfaction With Life Scale (SWLS; Diener, Emmons,
course credit for participating and $20 for completing all Larsen, & Griffin, 1985). This scale has five items that
five assessments. All participants had experienced inter- measure subjective well-being and showed high internal
personal transgressions within the past 7 days (M = 4.04 consistency and moderate test-retest stability estimates
days, SD = 1.82). Of the 115 participants, 96 completed across all five measurement occasions (alphas ranged
all five visits, 5 completed four visits, 5 completed three from .89 to .96 and test-retest rs ranged from .71
visits, 5 completed two, and 4 completed only one. to .92).
Although aspects of these data have been analyzed and
reported elsewhere (McCullough et al. 2003, Study 2), Moods. Positive and negative mood states were mea-
this is the only article addressing the relationship of for- sured at initial and follow-up sessions with the Positive
giveness and psychological well-being. and Negative Affect Schedule (PANAS; Watson, Clark,
& Tellegen, 1988). The measure consists of 20 items
asking participants to rate on 5-point Likert-type scales
Measures
how much they experienced different emotions (e.g.,
Interpersonal forgiveness. Participants’ forgiveness “upset” or “proud”) in the past 2 weeks. Internal con-
toward a transgressor was measured in both the initial sistency and test-retest stability estimates across all five
follow-up surveys with an 18-item version of McCullough measurement occasions were high for both positive and
et al.’s (1998) Transgression-Related Interpersonal negative PANAS subscales (alphas ranged from .78 to
Motivations (TRIM) Inventory. Participants are asked to .93 and test-retest rs ranged from .44 to .76).
rate on 5-point Likert-type scales how much they agree or
disagree with each item. The Avoidance subscale has 7 Psychosomatic symptoms. A self-report inventory
items measuring motivation to avoid contact with a trans- was used to measure participants’ experience of 27
gressor (e.g., “I live as if he/she doesn’t exist, isn’t around”), physical symptoms (e.g., dizziness, aches and pains, and
the Revenge subscale has 5 items measuring motivation to poor appetite) using items drawn from other surveys
seek revenge (e.g., “I’ll make him/her pay”), and the (e.g., Bartone, Ursano, Wright, & Ingraham, 1989;

Emmons, 1992). These scales also showed high internal Data Analytic Strategy
consistency and test-retest stability estimates across all
Our data conformed to a two-level hierarchical
five measurement occasions (alphas ranged from .91 to
structure (repeated measures nested within individuals),
.94 and test-retest rs ranged from .20 to .51).
so we used multilevel random coefficient models (Bryk
Because these measures of well-being were moder-
& Raudenbush, 1992; Hedeker, 2004) using the HLM
ately intercorrelated, we constructed an overall well-
6.2 statistical software package (Raudenbush, Bryk,
being composite by standardizing them to render them
Cheong, & Congdon, 2000) to analyze how people’s
in the same scale, reverse scoring the negative indices
forgiveness and feelings of closeness to their transgres-
(i.e., negative affect and physical symptoms) to render
sors were related to within-person changes in the well-
them all in the same direction, and averaging them
being composite variable.
together into an overall well-being index (alphas across
First, we wished to know if people’s TRIMs (avoid-
all five measurement occasions ranged from .73 to .79
ance, revenge and benevolence) and feelings of closeness
and test-retest rs ranged from .62 to .88).
toward their transgressors were each related to psycho-
logical well-being. We tested these hypotheses with
Closeness to the partner before and after the trans-
within-person, or Level-1, models (for each TRIM and
gression. Participants’ feelings of closeness to their
for closeness) of the form:
transgressors were measured in the initial survey by
three self-report items: (a) “How close were you to the
Well-Beingij = β0j + β1j (Time)ij + β2j
person who hurt you before the transgression?” (rated
(TRIM or Feelings of Closeness)ij + rij (1)
on a Likert-type scale from 0 = not at all to 6 =
extremely close); (b) “How committed were you to the In Equation 1, person j’s well-being on Day i is modeled
person who hurt you before the transgression?” (rated as a function of person j’s intercept (i.e., expected well-
on a scale from 0 = not at all to 6 = extremely commit- being score when the other predictors have a value of 0),
ted); and (c) the Inclusion of Other in the Self Scale linear change in time since the transgression occurred for
(IOS; Aron, Aron, & Smollan, 1992).2 The IOS presents person j (i.e., number of weeks to the nearest 1/7th of a
seven pairs of circles (ranging from no overlap to week from the first assessment), one of the TRIM sub-
extreme overlap) and instructs participants to circle the scales (i.e., Avoidance, Revenge, or Benevolence motiva-
pair that best describes their relationship to the person tion) or feelings of closeness, and a residual (i.e., variance
before the transgression. Responses were coded on a 7- in well-being that cannot be accounted for by the inter-
point scale, with 1 indicating no overlap and 7 indicat- cept, linear change in well-being, or a person’s TRIM
ing extreme overlap. These three items were combined score on that specific measurement occasion). We predicted
to measure participants’ sense of closeness to their rela- between-persons differences in the within-persons associa-
tionship partners before the transgression (α = .87).3 tions with the following fixed (i.e., time-invariant) covariates
The above items were modified to measure partici- at Level-2:
pants’ closeness, commitment, and self–other overlap
with their relationship partners after the transgression
and included in follow-up assessments. They were like- β0j = γ00 + u0j; (2)
wise combined to measure participants’ feelings of
closeness after the transgression (alphas ranged from β1j = γ10 + u1j; (3)
.87 to .94 and test-retest rs ranged from .62 to .82).
β2j = γ20 + γ21 (Feelings of Closeness Before)j + γ22
Apology/making amends. How much participants (Apology)j + u2j. (4)
thought their transgressors apologized and made
amends for their transgressions was measured with the Thus, we modeled the intercept and linear change β
mean of two items: (a) “How apologetic was the coefficients in the Level-1 models as a function of a
offender toward you?” and (b) “To what extent did he grand mean and a person-specific residual, and we fur-
or she make amends for what he/she did to you?” (both ther modeled the TRIM (or feelings of closeness) β coef-
had Likert-type scales ranging from 0 = not at all to 6 = ficients as functions of feelings of closeness to the
completely). The measure had good internal consistency transgressor before the transgression and apology/
and test-retest stability across follow-up assessments amends on the part of the transgressor. These multilevel
(alphas ranged from .79 to .90 and rs ranged from .56 models allowed us to examine whether fluctuations in
to .70). An individual difference variable for overall well-being occurred as a function of transient fluctua-
apology/made amends was created by averaging all five tions in each TRIM (or in feelings of closeness) above
measurement occasions (α = .91). or below the values expected on the basis of people’s

initial status and linear change (over time) in well-being. Are Within-Person Fluctuations in Well-Being
Moreover, these analyses allowed us to test if between- Associated With Forgiveness?
persons differences in closeness before the transgression
We first examined whether people tended to experi-
and apology/making amends moderated the strength of
ence greater well-being than expected on any given day
the within-person associations between the TRIMs (or
(based on their initial status and linear change estimates
feelings of closeness) and well-being.
in well-being) when they were also more forgiving (i.e.,
Having evaluated the associations of forgiveness and
less avoidant and vengeful and more benevolent) toward
feelings of closeness with well-being, we proceeded to
their transgressors than was typical for them. Table 2
evaluate whether feelings of closeness mediated any
lists the parameter estimates for these models.
obtained relationships between the TRIMs and well-
Avoidance motivation covaried significantly and nega-
being. Thus, we first regressed feelings of closeness onto
tively with well-being, t(df = 112) = –2.48, p = .015,
each TRIM, and then we regressed well-being onto each
effect size r = –.23. For the average person, every one-
TRIM and feelings of closeness with the following
unit increase in avoidance above that person’s mean
Level-1 equations:
was associated with a .09 reduction in well-being. There
was also a cross-level interaction (of marginal statistical
Feelings of Closenessij = β0j + β1j(Time)ij + significance) with closeness of the relationship before
β2j(TRIM)ij + rij; (5) the transgression, t(df = 112) = –1.89, p = .06, r = –.18.
For each unit a person felt closer to the partner before
the transgression, that person tended to experience an
Well-Beingij = β0j + β1j(Time)ij + β2j(TRIM)ij +
additional .06 reduction in well-being for each one-unit
β3j(Feelings of Closeness)ij + rij. (6)
increase in avoidance motivation.
The association between within-persons changes in
We included between-person covariates for closeness revenge motivation and well-being was nonsignificant for
and apology in only the TRIM parameters. These mul- the average person in our sample, t(df = 112) = −1.15,
tilevel models allowed us to examine whether feelings of p = .25, effect size r = –.11. However, this association
closeness appeared to mediate the relationships between had a cross-level interaction with apology/amends on
the TRIMs and well-being. the part of the transgression, t(df = 112) = –2.22, p =
.03, r = –.20. For each unit increase in the degree to
RESULTS which people thought their transgressors apologized/
made amends, the association of revenge motivation
Descriptive Statistics and psychological well-being became .06 units more
negative. That is, the negative association between
Participants reported transgressions committed by
revenge motivation and well-being became stronger to
girlfriends or boyfriends (59%), friends of the same
the extent that participants perceived the transgressor
gender (19%), friends of the other gender (11%), rel-
had apologized.
atives (10%), husbands or wives (3%), and “others”
For the Benevolence subscale, the opposite pattern
(9%). One person did not report the type of relation-
emerged. Benevolence motivation covaried significantly
ship involved. Participants described several types of
and positively with well-being, t(df = 112) = 2.92, p =
transgressions, including betrayals of confidence or
.005, r = .26. For the average person, every one-unit
insults by a friend (28%), arguments or neglect by a
increase in benevolence above the mean was associated
romantic partner, spouse, or ex-romantic partner
with a .13 increase in well-being. Therefore, on occa-
(22%); infidelity by a romantic partner or spouse
sions when people reported being more benevolent than
(19%); rejection, neglect, or insult by a family member
typical toward their transgressors, they also tended to
(10%); termination of romantic relationship (11%);
report greater well-being.4
insults by people other than family or friends (3%);
and rejection or abandonment by a friend or prospec-
Are Within-Person Fluctuations in Well-Being
tive relationship partner (2%). Five participants did
Associated With Feeling Close to the
not describe their transgressions. Mean levels on the
Transgressor?
between-persons measures were as follows: closeness
before the transgression was M = 4.97, SD = 1.25, Next, we investigated whether people experienced
and perceived apology by the transgressor across all greater well-being than expected (based on their initial
five time points was M = 2.44, SD = 1.63. Table 1 dis- status and linear change estimates) on any given day
play means and standard deviations for the repeated when they also felt closer to their transgressors on that
measures. day. We found that fluctuations in feelings of closeness

TABLE 1: Means and Standard Deviations for Major Variables Across Observations for Study 1 and 2
Avoidance Revenge Benevolence Closeness Well-Being
Time Point M (SD) M (SD) M (SD) M (SD) M (SD)
Study 1
T1 3.11 (1.08) 1.83 (0.83) 3.05 (1.02) 2.38 (1.74) –0.35 (0.76)
T2 2.96 (1.10) 1.61 (0.80) 3.25 (1.00) 2.73 (1.80) –0.05 (0.77)
T3 2.82 (1.13) 1.53 (0.77) 3.31 (1.04) 2.58 (1.85) 0.11 (0.70)
T4 2.74 (1.10) 1.52 (0.79) 3.38 (1.04) 2.64 (1.89) 0.16 (0.75)
T5 2.70 (1.15) 1.48 (0.87) 3.39 (1.07) 2.54 (1.84) 0.18 (0.78)
Study 2
T1 3.52 (1.02) 2.28 (1.02) 2.51 (0.91) 2.51 (1.81) –0.61 (0.87)
T2 3.52 (1.07) 2.18 (1.02) 2.48 (0.91) 2.37 (1.72) –0.55 (0.87)
T3 3.40 (1.10) 2.19 (1.06) 2.54 (0.99) 2.35 (1.76) –0.41 (0.84)
T4 3.31 (1.15) 2.17 (1.06) 2.59 (1.08) 2.38 (1.80) –0.27 (0.87)
T5 3.26 (1.21) 2.08 (1.08) 2.64 (1.10) 2.38 (1.85) –0.15 (0.82)
T6 3.20 (1.22) 2.02 (1.04) 2.64 (1.07) 2.38 (1.85) –0.08 (0.82)
T7 3.18 (1.23) 2.05 (1.06) 2.69 (1.08) 2.47 (1.87) –0.05 (0.78)
T8 3.13 (1.24) 1.97 (1.04) 2.76 (1.13) 2.47 (1.94) –0.02 (0.80)
T9 3.05 (1.27) 1.91 (1.01) 2.79 (1.16) 2.46 (1.93) 0.04 (0.77)
T10 3.03 (1.22) 1.84 (0.95) 2.75 (1.09) 2.46 (1.92) 0.05 (0.76)
T11 3.03 (1.22) 1.85 (0.96) 2.81 (1.09) 2.50 (1.94) 0.06 (0.79)
T12 3.00 (1.19) 1.84 (0.95) 2.78 (1.10) 2.49 (1.97) 0.06 (0.76)
T13 2.99 (1.16) 1.87 (0.97) 2.82 (1.04) 2.46 (1.92) 0.12 (0.68)
T14 2.92 (1.19) 1.89 (0.95) 2.85 (1.05) 2.45 (1.87) 0.11 (0.68)
T15 2.85 (1.16) 1.84 (0.96) 2.93 (1.04) 2.53 (1.89) 0.09 (0.68)
T16 2.77 (1.17) 1.79 (0.93) 3.00 (1.06) 2.69 (1.91) 0.17 (0.69)
T17 2.75 (1.19) 1.81 (1.00) 2.99 (1.10) 2.73 (2.02) 0.22 (0.69)
T18 2.71 (1.21) 1.86 (1.03) 3.02 (1.10) 2.70 (1.73) 0.18 (0.70)
T19 2.67 (1.18) 1.79 (0.97) 3.06 (1.14) 2.77 (2.04) 0.24 (0.72)
T20 2.71 (1.23) 1.81 (0.94) 3.09 (1.18) 2.81 (2.06) 0.27 (0.68)
T21 2.64 (1.20) 1.75 (0.94) 3.16 (1.15) 2.93 (2.07) 0.28 (0.71)
TABLE 2: Linear Models of Longitudinal Change in Well-Being (WB) as a Function of Forgiveness, Study 1
Measure (Parameter) Coefficient SE t Value Effect Size r % VAFa χ2(1, N = 114)
Avoidance motivation as measure of forgiveness

Initial status in WB –0.32 0.07 — — 79.54 388.47§
Linear change in WBb 0.06 0.01 7.92§ 0.60 0.62 220.13§
Avoidance –0.09 0.04 –2.48*** –0.23 2.54 127.95**
Closeness beforec –0.06 0.03 –1.89* –0.18 — —
Apology/amendsc –0.02 0.01 –1.29NS –0.12 — —
Revenge motivation as measure of forgiveness

Linear change in WB 0.07 0.01 7.58§ 0.58 0.68 137.96§
Revenge –0.07 0.06 –1.15NS –0.11 6.44 103.62*
Closeness before –0.00 0.04 –0.08NS –0.01 — —
Apology/amends –0.06 0.03 –2.22** –0.20 — —
Benevolence motivation as measure of forgiveness

Benevolence 0.13 0.04 2.92**** 0.26 4.09 97.08NS
Closeness before 0.06 0.04 1.33NS 0.12 — —
Apology/amends 0.02 0.02 0.62NS 0.06 — —
a. % VAF = percentage of variance accounted for by parameters.

b. Linear change estimate is based on a 1-week unit.
c. All models included between-persons covariates for feelings of closeness before the transgression and perceived apology/amends from trans-
gressor in only the forgiveness parameters.
*p < .10. **p < .05. ***p < .02. ****p < .01. §p < .001.

TABLE 3: Mediation Models for Study 1 and Study 2a
Predictor(X) > Mediator(M) > Outcome(Y) β(yx) β(mx) β(ym.x) β(yx.m) Sobel’s t
Study 1
Avoidance > Closeness > Well-Being −.09*** −.70§ .06** −.04NS 2.21**
Revenge > Closeness > Well-Being −.07NS −.42§ N/A N/A N/A
Benevolence > Closeness > Well-Being .13**** .83§ .07** .08NS 2.62****
Study 2
Avoidance yesterday > Closeness > Well-Being −.04*** −.25§ .10§ −.01NS 4.38§
Revenge yesterday > Closeness > Well-Being −.04** −.16§ .10§ −.03NS 3.55§
Benevolence yesterday > Closeness > Well-Being .01NS .16§ N/A N/A N/A
NOTE: N/A = Nonapplicable results because the presumed predictor was not associated with the presumed outcome in the bivariate regression.
a. All Level-1 models included fixed Level-2 covariates (centered around grand means) representing feelings of closeness before the transgression
and perceived apology from transgressor in the Transgression-Related Interpersonal Motivations (TRIM) parameters of those models.
**p < .05. ***p < .02. ****p < .01. §p < .001.
to the transgressor covaried significantly and positively outcome (well-being). These conditions can be evalu-
with well-being, t(df = 112) = 2.09, p = .04, r = .19. For ated with Sobel’s (1982) test for mediation. Krull and
the average person, every one-unit increase in feelings of MacKinnon (2001) showed that these criteria can be
closeness (above his or her mean) was associated with a used to test mediation in multilevel models. We used
.06 increase in well-being. Moreover, the strength of Preacher and Leonardelli’s (2001) online calculation
this association was moderated by the extent to which tool for mediation tests to conduct the Sobel test.
that person perceived the transgressor apologized/made Results, so far, supported the first two conditions for
amends, t(df = 112) = 2.27, p = .04, r = .21. Thus, for mediation. The first two columns of data in Table 3,
each unit-increase in the extent to which someone per- labeled β(yx) and β(mx), display the coefficients pro-
ceived that his or her transgressor apologized/made duced from individually regressing outcome y (i.e., well-
amends, a one-unit increase in closeness was associated being) and presumed mediator m (i.e., feelings of
with an additional .04 increase in well-being above that closeness) onto the predictor variables x (i.e., each
person’s mean. TRIM). They show that although all three TRIMs were
associated with feelings of closeness, only avoidance
Do Feelings of Closeness Mediate and benevolence motivation were associated with well-
the Relationship Between Well-Being being. The third column, labeled β(ym.x), shows that
and Forgiveness? feelings of closeness remained significantly associated
with well-being when avoidance or benevolence moti-
Having found that well-being was associated with
vation were controlled. The fourth column, labeled
transient fluctuations in people’s TRIMs (viz.,
β(yx.m), shows that the associations well-being had
Avoidance and Benevolence) and transient fluctuations
with avoidance and benevolence motivation were no
in people’s feelings of closeness to their transgressors,
longer significant when feelings of closeness were con-
we proceeded to examine whether feelings of closeness
trolled. This pattern supports the hypothesis that feel-
mediated the associations of avoidance and benevolence
ings of closeness mediated the associations of avoidance
with well-being.
and benevolence with well-being. The Sobel tests,
According to Baron and Kenny (1986), to test
whose t values are in the fifth column, demonstrate that
whether a variable such as feelings of closeness mediates
the degree of mediation was statistically significant.5
the associations between the TRIMs and well-being,
several conditions must be satisfied. First, the predictor
variable (i.e., each TRIM) must be associated with the DISCUSSION
outcome variable (i.e., well-being). Second, the pre-
sumed mediator (i.e., feelings of closeness) must like- In Study 1, we examined whether the well-established
wise be associated with the outcome (i.e., well-being). correlation between measures of forgiveness and mea-
Third, the predictor (each TRIM) must be associated sures of well-being (e.g., Brown, 2003; Hebl & Enright,
with the presumed mediator (closeness). Lastly, upon 1993; Karremans et al., 2003; Krause & Ellison, 2003;
regressing the outcome of well-being onto both the pre- McCullough et al., 1998; Seybold, Hill, Neumann, &
dictor (each TRIM) and presumed mediator (closeness) Chi, 2001) existed at the within-persons level. This
simultaneously, the presumed mediator (closeness) must hypothesis was largely supported. Results also sup-
maintain a significant association with the presumed ported the notion that the relationship of forgiveness

and well-being can differ between people depending on to complete one each day, date and complete each
how close and committed people are to their transgres- honestly (without responding later to skipped days),
sors prior to the transgression and the degree to which and then return completed packets to the laboratory
people perceived their transgressors to have apologized (after approximately 21 days). Completion rates were
and made amends for their hurtful behavior. In general, very high, with 109 participants (66.9%) completing all
feelings of closeness to one’s transgressor prior to the 21 diary entries and only 9 (5.5%) completing fewer
transgressor and the perception that one’s transgressor than 10. Though we evaluated the links of forgiveness and
has apologized and made amends appear to strengthen well-being using this dataset in only the present article,
the association of forgiveness and well-being. This is other details of the data set can be obtained from
consistent with Karremans et al.’s (2003) research on the McCullough et al. (2007).
role of commitment as a moderator of the association of
forgiveness with psychological well-being. Measures
Evidence also supported the notion that when people
We used the same or shorter instruments (to reduce
felt closer and more committed than typical to their part-
burden) as in Study 1 to measure forgiveness, satisfac-
ners they also experienced elevations in well-being, and
tion with life, positive/negative mood (9-item version),
that these increases in feelings of closeness were greater
physical symptoms (16-item version), feelings of close-
for people whose transgressors apologized and made
ness to the transgressor before and after the transgres-
amends more. Moreover, evidence showed that increases
sion, and perceptions of transgressor apology/amends.
in feelings of closeness could explain the within-persons
Likewise, all instruments were included in initial and
associations between forgiveness and well-being.
follow-up surveys. The major difference was that par-
ticipants completed instruments daily (rather than
bimonthly) and wherever they chose (rather than in a
STUDY 2
laboratory). All measures had comparable internal con-
sistency and test-retest reliabilities to Study 1.
The purpose of Study 2 was to explore with greater
statistical rigor the possibility that the observed associ-
Statistical Models and Data Analyses
ations between forgiveness and well-being are causal in
nature. That is, we wanted to know if the covariance Our first aim was to replicate the findings from Study
structure among the TRIMs, feelings of closeness, and 1 using cross-lagged analyses. Thus, we first examined
well-being supported the hypotheses that (a) being rela- if people’s TRIM scores or feelings of closeness on any
tively forgiving on a given day is associated with greater given day were each prospectively associated with their
well-being on the successive day and (b) being relatively well-being on the next day, and then we examined if
forgiving on a given day is associated with better well- their TRIM scores were prospectively associated with
being on the successive day by way of increasing feel- feelings of closeness toward their transgressors on the
ings of closeness toward the transgressor on the next day, while controlling for their well-being and feel-
successive day. An additional aim of Study 2 was to test ings of closeness on the same day, to test whether feel-
the reverse causal model for the correlations obtained in ings of closeness served as a mediator between the
Study 1, namely, if well-being exerts a causal influence TRIMs and well-being. We computed the following
on forgiveness via feelings of closeness. HLM equations (separately for each TRIM):
Method Well-Beingij = β0j + β1j(Time)ij + β2j(Well-Being)i – 1,j +

β3j(TRIM)i – 1,j + β4j(Feelings of Closeness)i – 1,j + rij; (7)
Participants and Procedure
Participants were 165 students in undergraduate psy- Feelings of Closenessij = β0j + β1j(Time)ij +
chology courses (112 women) at the University of β2j(Well-Being)i – 1,j + β3j(TRIM)i – 1,j + β4j(Feelings
Miami. Again, all participants had experienced a trans- of Closeness)i – 1,j + rij. (8)
gression within the past 7 days (M = 4.37 days, SD =
1.85; M age = 19.61, SD = 3.82) at the time of enroll- To assess the possibility that forgiveness on a given
ment. They received course credit for participating and day led to increased well-being the next day, we eval-
$20 for completing the tasks described here (and a sep- uated the statistical significance of the β3j coefficients
arate laboratory session not described here). We visited for each TRIM. To assess the possibility that feelings
these courses throughout the semester, screened, then of closeness on a given day led to increased well-being
enrolled participants who qualified by giving them pre- on the next day, we computed a model like Equation
liminary and 21 follow-up surveys. They were instructed 7, but without a TRIM as a predictor variable, and

evaluated the statistical significance of the β coefficient termination of a romantic relationship (13%); neglect
for feelings of closeness on Day i – 1. Then we pro- by a romantic or ex-romantic partner or spouse (10%);
ceeded to examine if any of the TRIMs’ prospective rejection or abandonment by a friend or prospective
associations with well-being were mediated by feelings relationship partner (10%); and insults by others (5%).
of closeness. We did this by estimating a model akin to The between-persons measures had comparable means
Equation 7 separately for each TRIM, but with person to Study 1: Closeness before the transgression was M =
j’s feelings of closeness on Day i added as another pre- 4.90, SD = 1.29; overall perceived apology/amends was
dictor variable. These multilevel models allowed us to M = 2.39, SD = 1.82 (all on 7-point scales). Table 1 dis-
evaluate whether (a) forgiveness was associated with plays descriptive statistics for the repeated measures.
well-being on the next day, (b) feelings of closeness
were associated with well-being on the next day, (c)
forgiveness was associated with feelings of closeness Is Forgiveness on a Given Day Associated With
toward the transgressor on the next day, and (d) the Greater Well-Being on the Successive Day?
prospective association of forgiveness with well-being
Multilevel models showed that on any given Day i –
could be accounted for by increased feelings of close-
1, when people had higher levels of avoidance or
ness on the next day.
revenge motivation toward their transgressors than was
As in Study 1, we included between-persons covari-
typical for them, they tended to also have lower levels
ates representing closeness before the transgression and
of well-being on the following Day i than we could
apology/amends in only the TRIM on Day i-1 parame-
expect based on their well-being trajectories alone [for
ters (or in the feelings of closeness on Day i – 1 para-
avoidance t(df = 161) = –2.38, p = .02, effect size r =
meter). This allowed us to examine if any prospective links
–.18; for revenge t(df = 161) = –1.96, p = .05, effect size
of the TRIMs (or feelings of closeness) with well-being dif-
r = –.15], even after controlling for the covariation
fered as a function of prior closeness and apology/ amends.
among avoidance (or revenge), well-being, and feelings
Finally, our second aim in Study 2 was to examine if
of closeness on Day i – 1 (see Table 4). Moreover, we
well-being was prospectively related to the TRIMs and
found cross-level interactions with both of our individ-
whether feelings of closeness served as a mediator in
ual difference variables. That is, the prospective associ-
this reverse causal model. We did this by computing
ation of avoidance and well-being was stronger for
HLMs similar to those above for each TRIM (but using
people who were closer to their partners before the
no between-person covariates6):
transgression, t(df = 161) = –2.12, p = .04, effect size r
= –.16, and for people who thought that their trans-
gressors were more apologetic for the transgression t(df
TRIMij = β0j + β1j(Time)ij + β2j(Well-Being)i – 1,j +
= 161) = –2.33, p = .02, effect size r = –.18. The prospec-
β3j(TRIM)i – 1,j + β4j(Feelings of Closeness)i – 1,j + rij; (9)
tive association of revenge and well-being was also
stronger (with marginal statistical significance) for people
Feelings of Closenessij = β0j + β1j(Time)ij +
who were relatively close to their partners before the trans-
β2j(Well-Being)i – 1,j + β3j(TRIM)i – 1,j + β4j(Feelings
gression, t(df = 161) = –1.65, p = .10, effect size r = –.13.
of Closeness)i – 1,j + rij. (10)
The link between benevolence on any given Day i – 1,
and well-being on the following Day i, was nonsignifi-
cant for the average person in our sample, t(df = 161) =
RESULTS 0.63, p = .13, effect size r = .12. However, this associa-
tion was moderated by between-persons differences in
Descriptive Statistics
perceived apology/amends, t(df = 159) = 3.64, p = .001,
As in Study 1, our sample included many types of effect size r = .27. Together, these results suggest that
partners and transgressions. Participants reported trans- forgiveness may lead to later levels of well-being, partic-
gressions committed by girlfriends or boyfriends (50%), ularly when forgiveness is measured in terms of reduced
friends of like gender (19%), relatives (13%), friends of avoidance and revenge motivation (Finkel, 1995). As
the other gender (9%), husbands or wives (1%) and “oth- with Study 1, these results also provide evidence that the
ers” (8%). They also described several types of transgres- associations between forgiveness and well-being were
sions, including infidelity by a romantic partner or spouse stronger when people were forgiving someone to whom
(29%); friends’ insults or betrayals of confidence (20%); they felt closer before the transgression or someone who
rejection, neglect, or insult by a family member (13%); apologized/made amends more.

TABLE 4: Linear Models of Cross-Lagged Correlations Between Well-Being (WB) and Forgiveness, Study 2a
Measure (Parameter) Coefficient SE t Value Effect Size r % VAFb χ2(1, N = 164)
Avoidance motivation as measure of forgiveness

Linear change in WBc 0.02 0.00 8.60§ 0.53 0.04 244.42§
Avoidance yesterday –0.04 0.02 –2.38*** –0.18 0.96 183.97****
Closeness befored –0.03 0.01 –2.12** –0.16 — —
Apology/amendsd –0.02 0.01 –2.33*** –0.18 — —
Revenge motivation as measure of forgiveness

Revenge yesterday –0.04 0.02 –1.96** –0.15 2.10 184.92****
Closeness before –0.03 0.01 –1.65* –0.13 — —
Apology/amends –0.01 0.01 –0.91NS –0.07 — —
Benevolence motivation as measure of forgiveness

Benevolence yesterday 0.01 0.02 0.63NS 0.05 1.46 197.40****
Closeness before –0.01 0.09 –0.36NS –0.03 — —
Apology/amends 0.18 0.06 3.64**** 0.27 — —
a. All models also included Level-1 parameters for previous day’s feelings of closeness and well-being (estimates not shown here).
b. % VAF = percentage of variance accounted for by parameters.
c. Linear change estimate based on a 1-day unit.
d. All models included between-persons covariates for feelings of closeness before the transgression and perceived apology/amends from trans-
gressor in only the forgiveness parameters.
*p < .10. **p < .05. ***p < .02. ****p < .01. §p < .001.
Are Feelings of Closeness on a Given Day feelings of closeness toward the transgressor on Day i
Associated With Greater Well-Being on the and whether feelings of closeness on Day i mediated the
Successive Day? prospective associations of avoidance and revenge with
well-being, after controlling for the covariation
Next, we investigated whether people experienced
between these TRIMs, well-being, and feelings of close-
greater well-being than expected on any given day (on
ness on Day i – 1. The first column in Table 3 displays
the basis of their initial status and linear change estimates)
the coefficients produced from regressing the outcome
when they felt closer to their transgressors on the previ-
(well-being) on Day i upon the predictor variables
ous day. The link between feelings of closeness on any
(each TRIM) on Day i – 1. Column 2 shows the results
given Day i – 1 and well-being on the following Day i
of regressing the presumed mediator (feelings of close-
was nonsignificant for the average person in our sam-
ness) on Day i upon each TRIM on Day i – 1. These
ple, t(df = 161) = 0.89, p = .38, effect size r = .07.
coefficients indicate that avoidance, revenge, and
However, this association was moderated by between-
benevolence motivation on Day i – 1 were each associ-
persons differences in perceived apology/amends, t(df =
ated with feelings of closeness on Day i. The third and
161) = 2.99, p = .004, effect size r = .23. Thus, the rela-
fourth columns display the unique associations of well-
tionship of closeness and well-being was stronger
being on Day i with feelings of closeness on Day i and
among people whose transgressors had apologized and
with avoidance or revenge motivations on Day i – 1
made amends than it was among those whose trans-
(respectively), when predictor and mediator are entered
gressors had not done so.
simultaneously in regression models. These coefficients
indicate that feelings of closeness (our presumed medi-
Does Forgiveness Obtain Its Link to Well-Being by
ator) had unique and significant associations with well-
Way of Feelings of Closeness to the Transgressor?
being when avoidance or revenge motivations on Day i
Because increases in avoidance and revenge on any – 1 were statistically controlled. Together with the
given day were associated with reductions in well-being Sobel tests, whose t values are in the fifth column, this
the next day, we next examined whether people’s pattern supports the conclusion that the observed
TRIMs on Day i – 1 were associated with people’s increases in well-being on Day i associated with

from their avoidance or revenge trajectories alone. On

the other hand, when people had above-typical well-
being on any given Day i – 1, they also had higher lev-
els of benevolence, t(df = 163) = 2.75, p = .007, effect
size r = .21, on the next Day i than we could expect
from their benevolence trajectories alone. These find-
ings emerged, moreover, after controlling for the covari-
ation among well-being, each TRIM, and feelings of
closeness on Day i – 1. Thus, this provides evidence that
psychological well-being may also lead to increases in
forgiveness.
Next we examined whether people’s well-being on
any given Day i – 1 was associated with their feelings
of closeness toward the transgressor on the following
Day i, after controlling for the covariation between
each TRIM, well-being, and feelings of closeness on
Day i – 1. Results demonstrated that when people had
higher well-being than was typical for them on Day i –
1, they did not have significantly greater feelings of
closeness to their transgressors on the next Day i than
Figure 1 Models of the prospective associations of forgiveness (i.e.,
changes in Transgression-Related Interpersonal Motivations) could be expected from their closeness trajectories
with well-being alone, ts(df = 163) = –0.20, 0.41, and 0.28, ps = .85,
NOTE: The figure shows that for the average person in the sample, .68, and .78 (for models using the avoidance, revenge
feelings of closeness completely mediated the associations that both
avoidance (1A) and revenge (1B) had with later increases in well-being. or benevolence on Day i – 1 parameters, respectively).
*p < .05. **p < .01. ***p ≤ .001. Therefore, feelings of closeness did not operate as a
mediator in well-being’s prospective associations with
increases in the TRIMs. Taken together, these results
indicate that well-being may lead to increases in for-
giveness, but that this does not occur via changes in
reduced avoidance and revenge motivations on Day i –
feelings of closeness.
1 are mediated by increased feelings of closeness with
the transgressor on Day i (see Figures 1A and 1B).7
DISCUSSION
Is Well-Being on a Given Day Linked to
The results from Study 2 largely replicated the find-
Forgiveness or Feelings of Closeness on a
ings from Study 1. Fluctuations in all three motiva-
Successive Day?
tions underlying forgiveness were uniquely related to
Having found evidence that forgiveness is linked to subsequent psychological well-being. Specifically, we
increases in well-being in a cross-lagged fashion and that found that reductions in avoidance and revenge moti-
the association between the forgiveness and well-being in vations on any given day were related to greater well-
these analyses could still be explained by feelings of close- being on the next day for the average person; and that
ness (at least with respect to avoidance and revenge), our increases in benevolence, on the other hand, were
second aim for Study 2 was to further test the causal via- more strongly related to greater well-being on the next
bility of our mediational models. This required ruling out day the more a person perceived his or her transgres-
evidence that the covariance structure among the TRIMs, sor to have apologized or made amends. Results again
feelings of closeness, and well-being also supports close- supported the idea that forgiveness (in terms of
ness’ mediation of any prospective associations between reduced avoidance) may be more strongly related to
well-being and the TRIMs. well-being for people who reported greater close-
Multilevel models for the reverse causal models ness/commitment to the relationship before the trans-
showed that on any given Day i – 1, when people had gression. Finally, results from Study 2 replicated the
higher levels of psychological well-being than was typi- finding that the positive association between forgive-
cal for them, they also tended to have lower levels of ness and well-being can be accounted for by feelings of
avoidance, t(df = 163) = –2.91, p = .005, effect size r = closeness toward the transgressor.
–.22, and revenge, t(df = 163) = –4.03, p < .001, effect Study 2 also provided support for the proposition
size r = –.30, on the next Day i than we could expect that transgressors may play a productive role in helping

victims forgive (and, thereby, experience improved well- present studies therefore contribute to the research on
being), by apologizing and making amends for their forgiveness, relationship closeness, and psychological
transgressions. The fact that Study 2 replicated so many well-being by demonstrating that the between-persons
of the findings from Study 1—using a prospective relationships between these constructs, obtained in pre-
design that allowed us to test one important condition vious studies, also obtain at the within-persons level.
for causal conclusions (i.e., that changes in a putative x Moreover, by using time-lagged analyses in Study 2 to
variable precede the changes in a putative y variable)— demonstrate that forgiveness at any given point in time
yields some additional (though by no means conclusive) is associated with later levels of psychological well-
evidence to the notion that the links of forgiveness to being (even after controlling for previous levels of psy-
closeness and well-being may in fact be causal in nature. chological well-being), we found support for the notion
However, it is important to note that Study 2 results that the associations between forgiveness and well-
were not perfectly consistent with Study 1 results. For being are causal in nature.
example, in Study 1, avoidance and benevolence moti- Overall, the links between forgiveness and psycholog-
vation were both (cross-sectionally) associated with ical well-being were stronger in relationships character-
well-being, but in Study 2, avoidance, but not benevo- ized by two qualities: (a) greater closeness and
lence, was (prospectively) associated with well-being.8 In commitment to the relationship (at least from the vic-
Study 1, revenge motivation was not (cross-sectionally) tim’s perspective) and (b) a high degree of apology and
associated with well-being, but in Study 2, it was making amends from the transgressor following the
(prospectively) associated with well-being. Frankly, we transgression. If a relationship is personally valuable to
are at a loss to offer a coherent explanation for these the victim (e.g., if the transgressor was a close friend or
discrepancies and are tempted to conclude that they loved one), then the disruption of this relationship also
simply represent the vagaries of sampling error, but limits the victim’s access to the social-psychological
future work may reveal a more substantive explanation. resources that the relationship provided (e.g., emotional
Finally, Study 2 provided evidence that well-being is and material social support, love, a sense of social inclu-
also associated with subsequent increases in forgiveness sion, etc.), which leads to psychological distress. By forgiv-
and that, therefore, both causal orderings of the for- ing and thereby promoting the restoration of the
giveness and well-being relationship may be simultane- relationship, those social-psychological resources become
ously possible. Importantly, however, these final available again to the victim. Likewise, failures to forgive
analyses also demonstrated that well-being’s association an apologetic partner are more negatively associated with
with subsequent forgiveness could not be accounted for well-being than are failures to forgive an unapologetic one,
by feelings of closeness to the transgressor. Therefore, perhaps because apologies send signals that a relationship
Study 2 also indicated that it is only in the case of the is likely to possess value to the forgiver in the future,
association of forgiveness with later well-being that feel- whereas an unapologetic offender’s future relationship
ings of closeness appear to serve a mediating role. intentions remain more uncertain. The notion that it is the
close, committed relationships that are likely to have future
value to the forgiver are the ones in which forgiveness is
GENERAL DISCUSSION most closely linked to better well-being is highly consistent
with evolutionary accounts of the psychological processes
In the present article, we adduced data from two sep- underlying reciprocal altruism (Trivers, 1971).
arate studies to test the hypothesis that forgiveness is Finally, we found evidence that well-being was also
associated with psychological well-being. Furthermore, associated with increases in forgiveness, a notion that,
we hypothesized that it obtains this association by help- as far as we know, has not been examined empirically.
ing people restore a subjective sense of closeness and While this finding aligns well with the common belief
commitment to a transgressing relationship partner. that forgiveness is difficult to practice, it implies further
The two studies were quite consistent in their support of that one is more capable of practicing it when well-
these hypotheses, in keeping with other recent findings being is high. Nonetheless, this last finding suggests that
(Karremans & Van Lange, 2004; McCullough et al., future researchers should attend to the notion that for-
1998; Tsang et al., 2006). These results are largely con- giveness and well-being may be bidirectionally related
sistent with the idea that psychological well-being can when untangling the relationships between the two.
serve as an indicator of the availability of positive social
relations, that positive social relations are a crucial human Directions for Future Research
need (Ryan & Deci, 2000; Trivers, 1971), and that help-
ing to restore valuable social relations is how forgiveness Controlled experiments might be the logical next
obtains its positive association with well-being. The step for exploring the associations between forgiveness

and well-being in greater detail. Second, because our assessments in both studies. We used oblimin rotation (delta = 0). Four
factors emerged after nine iterations. The first factor accounted for
findings show that “reconnecting” with valued relation- 42.8% of the variance (6 Avoidance items), with pattern loadings >.59.
ship partners who transgress is important to well-being, The second factor accounted for 13.9% of the variance (all 5 Revenge
large longitudinal research programs on relationships or items), with pattern loadings >.75. The third factor accounted for
7.3% of the variance (all 3 Closeness items), with pattern loadings
well-being may benefit from incorporating a focus on >.79. The fourth factor accounted for 5% of the variance (4
forgiveness. This, moreover, would enable measure- Benevolence items), with pattern loadings >.47. One Avoidance item
ment of pretransgression closeness/commitment prior to and one Benevolence item did not load on their target or nontarget
factors with pattern loadings >.11, and another Benevolence item
a transgression, thereby avoiding a limitation for which loaded on the revenge factor (–.36). The Avoidance factor was corre-
our studies could be criticized. lated with the Revenge, Benevolence, and Closeness factors at r(N =
Given the importance of apology as an apparent 280) = .42 and –.44, and –.59, respectively; the Revenge factor was
correlated with the Closeness and Benevolence factors at r(N = 280)
moderator of the associations between forgiveness and = –.13 and –.30, respectively; and the Closeness factor was correlated
well-being (McCullough et al., 1997), further research with the Benevolence factor at r(N = 280) = .13.
is also needed to investigate how other conciliatory 4. Gender and age differences in the within-persons associations of
the TRIMs with well-being were consistently nonsignificant across
strategies on the part of the transgressor (e.g., compen- Studies 1 and 2.
sation, gifts, expressions of remorse, and nonverbal 5. The traditional Sobel test for mediation does not consider
expressions of distress, shame, etc.) might promote for- potential covariance between the a and b paths for lower level medi-
ation in random effects multilevel models (Kenny, Korchmaros, &
giveness, and thereby, psychological well-being. Research Bolger, 2003). This can lead to inaccurate estimations of mediated
involving both victims and transgressors may be partic- effects when lower level associations vary randomly at upper levels.
ularly useful for answering such questions. Third, it is We used Korchmaros and Kenny’s (2003) approach for calculating
percentage amounts of mediation after adjusting for the ab covariance
worth investigating whether the apparent associations in the avoidance model (the association of benevolence and well-being
of forgiveness with relational well-being and psycholog- did not vary randomly across upper level units). In Study 1, closeness
ical well-being have implications for understanding the still mediated avoidance’s association with well-being 100%.
6. Between-persons closeness before the transgression and apology
relationship of forgiveness with physiological function- were not included in the reverse causal model to simplify analyses.
ing and physical health (Lawler et al., 2003; Witvliet, Also, testing if these variables moderated well-being’s prediction of
Ludwig, & Vander Laan, 2001). closeness had less empirical rationale than did testing if they moder-
ated forgiveness’ prediction of closeness. Regardless, reverse models
In their recent interest in the concept of forgiveness, exploring these covariates did not alter results.
psychologists have tended to study its various dimen- 7. In Study 2, closeness still mediated the prospective associations
sions in isolation: Some studies have examined the rela- of avoidance and revenge with well-being 100%, respectively, after
using Korchmaros and Kenny’s (2003) ab covariance adjustment.
tionship causes and consequences of forgiveness, some 8. It is worth noting that we found benevolence and well-being to be
have focused on the social precursors of forgiveness, and cross-sectionally associated in both datasets (t = 7.99 in Study 2). The
some have focused on forgiveness’ associations with quick prospective decay of this effect suggests that conciliatory stances
following transgressions may only correspond to immediate, short-lived
health and well-being. In the present study, we have tried boosts in well-being, perhaps like bursts of self-empowerment.
to integrate these concerns to provide a richer and more
comprehensive picture of the basic links between for-
giveness, relationship processes, and well-being.
REFERENCES
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in

NOTES the Self Scale and the structure of interpersonal closeness. Journal
of Personality and Social Psychology, 63, 596-612.
1. It is interesting that individual differences in transgression Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator dis-
painfulness did not significantly moderate the Transgression-Related tinction in social psychological research: Conceptual, strategic,
Interpersonal Motivations’ (TRIM’s) associations with well-being in and statistical considerations. Journal of Personality and Social
either study, except for that between revenge and well-being in Study Psychology, 51, 1173-1182.
1 (effect size r = .18). Thus, it was not included in analyses. Bartone, P. T., Ursano, R. J., Wright, K. M., & Ingraham, L. H.
2. One limitation in our studies was that we measured pretrans- (1989). The impact of a military air disaster on the health of
gression closeness and commitment after the transgression had assistance workers. Journal of Nervous and Mental Disease, 177,
already occurred. Although there is no way to know whether this 317-328.
method introduced bias or error, it is unlikely that such artifacts Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire
would make results invalid. If it introduces systematic elevations or for interpersonal attachments as a fundamental human motiva-
deflations in participants’ reports, this would affect their absolute tion. Psychological Bulletin, 117, 497-529.
scores on the measure but not their rankings relative to other partici- Brown, R. P. (2003). Measuring individual differences in the tendency
pants. Thus, correlations with other variables would not be affected. to forgive: Construct validity and links with depression.
If the bias introduces unsystematic error, this would simply handicap Personality and Social Psychology Bulletin, 29, 759-771.
our ability to find any results—a limitation that would likely create Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models.
underestimates of the importance of pretransgression closeness as a Newbury Park: Sage.
moderator of the forgiveness/well-being association. Cacioppo, J. T., Hawkley, L. C., Ernst, J. M., Burleson, M., Berntson,
3. To confirm that the closeness scale was distinct from the TRIM G. G., Nouriani, B., et al. (2006). Loneliness within a nomologi-
subscales, we conducted a maximum likelihood factor analysis on the cal net: An evolutionary perspective. Journal of Research in
18 TRIM items and the 3 closeness items using data from the first Personality, 40, 1054-1085.

de Waal, F. B. M. (2000). Primates—A natural heritage of conflict McCullough, M. E., & Hoyt, W. T. (2002). Transgression-related
resolution. Science, 289, 586-590. motivational dispositions: Personality substrates of forgiveness
Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). and their links to the Big Five. Personality and Social Psychology
The Satisfaction with Life Scale. Journal of Personality Bulletin, 28, 1556-1573.
Assessment, 49, 71-75. McCullough, M. E., Rachal, K. C., Sandage, S. J., Worthington, E. L.,
Emmons, R. A. (1992). Abstract versus concrete goals: Personal striv- Jr., Brown, S. W., & Hight, T. L. (1998). Interpersonal forgiving in
ing level, physical illness, and psychological well-being. Journal of close relationships. II: Theoretical elaboration and measurement.
Personality & Social Psychology, 62, 292-300. Journal of Personality and Social Psychology, 75, 1586-1603.
Fincham, F. D. (2000). The kiss of the porcupines: From attributing McCullough, M. E., Worthington, E. L., Jr., & Rachal, K. C. (1997).
responsibility to forgiving. Personal Relationships, 7, 1-23. Interpersonal forgiving in close relationships. Journal of
Finkel, S. E. (1995). Causal analysis with panel data. Thousand Oaks, Personality and Social Psychology, 73, 321-336.
CA: Sage. Nesse, R. M. (2001). Natural selection and the capacity for commit-
Freedman, S. R., & Enright, R. D. (1996). Forgiveness as an inter- ment. In R. M. Nesse (Ed.), Evolution and the capacity for com-
vention goal with incest survivors. Journal of Consulting and mitment (pp. 1–36). New York: Russell Sage.
Clinical Psychology, 64, 983-992. Nezlek, J. B. (2001). Multilevel random coefficient analyses of event-
Hamilton, W. D. (1964). The genetical evolution of social behaviour. and interval-contingent data in social and personality psychology
Journal of Theoretical Biology, 7, 17-52. research. Personality and Social Psychology Bulletin, 27, 771-785.
Hebl, J. H., & Enright, R. D. (1993). Forgiveness as a psychothera- Ohbuchi, K., Kameda, M., & Agarie, N. (1989). Apology as aggres-
peutic goal with elderly females. Psychotherapy, 30, 658-667. sion control: Its role in mediating appraisal of and responses to
Hedeker, D. (2004). An introduction to growth modeling. In harm. Journal of Personality and Social Psychology, 56, 219-227.
D. Kaplan (Ed.), The Sage handbook of quantitative methodology Poloma, M. M., & Gallup, G. H. (1991). Unless you forgive others:
for the social sciences (pp. 215-234). Thousand Oaks, CA: Sage. Prayer and forgiveness. In M. M. Poloma & G. H. Gallup (Eds.),
Hodgins, H. S., & Liebeskind, E. (2003). Apology versus defense: Varieties of prayer (pp. 85-106). Philadelphia: Trinity Press.
Antecedents and consequences. Journal of Experimental Social Preacher, K. J., & Leonardelli, G. J. (2001, March). Calculation for the
Psychology, 39, 297-316. Sobel test: An interactive calculation tool for mediation tests
House, J. S., Landis, K. R., & Umberson, D. (1988). Social relation- [Computer software]. Available from http://www.unc.edu/~ preacher/
ships and health. Science, 241, 540-545. sobel/sobel.htm
Karremans, J., & Van Lange, P. A. M. (2004). Back to caring after Raudenbush, S. W., Bryk, A., Cheong, Y. F., & Congdon, R. (2000).
being hurt: The role of forgiveness. European Journal of Social HLM 5: Hierarchical linear and nonlinear modeling. Lincolnwood,
Psychology, 34, 207-227. IL: Scientific Software International.
Karremans, J. C., Van Lange, P. A. M., Ouwerkerk, J. W., & Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M.
Kluwer, E. S. (2003). When forgiving enhances psychological well- (2000). Daily well-being: The role of autonomy, competence, and
being: The role of interpersonal commitment. Journal of relatedness. Personality and Social Psychology Bulletin, 26(4),
Personality and Social Psychology, 84, 1011-1026. 419-435.
Kelley, H., & Thibault, J. W. (1978). Interpersonal relationships: A Rusbult, C. E., Verette, J., Whitney, G. A., Slovik, L. F., & Lipkus, I.
theory of interdependence. New York: John Wiley. (1991). Accommodation processes in close relationships: Theory
Kenny, D. A., Korchmaros, J. D., & Bolger, N. (2003). Lower level and preliminary empirical evidence. Journal of Personality and
mediation in multilevel models. Psychological Methods, 8, 115-128. Social Psychology, 60, 53-78.
Korchmaros, J. D., & Kenny, D. A. (2003). Step by step procedure for Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the
estimating lower-level mediation in random-effects multilevel facilitation of intrinsic motivation, social development, and well-
models using HLM5 .Unpublished manuscript. being. American Psychologist, 55, 68-78.
Krause, N., & Ellison, C. G. (2003). Forgiveness by god, forgiveness Seybold, K. S., Hill, P. C., Neumann, J. K., & Chi, D. S. (2001).
of others, and psychological well-being in late life. Journal of the Physiological and psychological correlates of forgiveness. Journal
Scientific Study of Religion, 42, 77-94. of Psychology and Christianity, 20(3), 250-259.
Krull, J. L., & MacKinnon, D. P. (2001). Multilevel modeling of indi- Sobel, M. E. (1982). Asymptotic intervals for indirect effects in struc-
vidual and group level mediated effects. Multivariate Behavioral tural equations models. In S. Leinhart (Ed.), Sociological method-
Research, 36, 249-277. ology 1982 (pp. 290-312). San Francisco: Jossey-Bass.
Lawler, K. A., Younger, J. W., Piferi, R. L., Billington, E., Jobe, R., Trivers, R. L. (1971). The evolution of reciprocal altruism. Quarterly
Edmondson, K., et al. (2003). A change of heart: Cardiovascular Review of Biology, 46, 35-57.
correlates of forgiveness in response to interpersonal conflict. Tsang, J., McCullough, M. E., & Fincham, F. (2006). The longitudinal
Journal of Behavioral Medicine, 26, 373-393. association between forgiveness and relationship closeness and com-
Leary, M. R. (2004). The sociometer, self-esteem, and the regulation mitment. Journal of Social and Clinical Psychology, 25, 448-472.
of interpersonal behavior. In R. F. Baumeister & K. D. Vohs Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and
(Eds.), Handbook of self-regulation: Research, theory, and appli- validation of brief measures of positive and negative affect: The
cations (pp. 373-391). New York: Guilford. PANAS scales. Journal of Personality and Social Psychology, 54,
McCullough, M. E., Bellah, C. G., Kilpatrick, S. D., & Johnson, J. L. 1063-1070.
(2001). Vengefulness: Relationships with forgiveness, rumination, Witvliet, C. v. O., Ludwig, T. E., & Vander Laan, K. L. (2001).
well-being, and the Big Five. Personality and Social Psychology Granting forgiveness or harboring grudges: Implications for emo-
Bulletin, 27, 601-610. tion, physiology, and health. Psychological Science, 12, 117-123.
McCullough, M. E., Bono, G., & Root, L. M. (2007). Rumination, Zechmeister, J. S., Garcia, S., & Romero, C. (2004). Don’t apologize
affect, and forgiveness: Three longitudinal studies. Journal of unless you mean it: A laboratory investigation of forgiveness and
Personality and Social Psychology, 92, 490-505. retaliation. Journal of Social and Clinical Psychology, 23, 232-564.
McCullough, M. E., Fincham, F. D., & Tsang, J. (2003). Forgiveness,
forbearance, and time: The temporal unfolding of transgression-
related interpersonal motivations. Journal of Personality and Received October 19, 2006
Social Psychology, 84, 540-557. Revision accepted July 15, 2007

View publication stats

Bono 000

Uploaded by

Copyright:

Available Formats

Bono 000

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bono 000

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Forgiveness, Feeling Connected to Others, and Well-Being: Two Longitudinal

Article in Personality and Social Psychology Bulletin · March 2008

Giacomo Bono Michael E Mccullough

SEE PROFILE SEE PROFILE

Lindsey Root Luna

Youth Gratitude Project View project

Gratitude in the Time of Covid-19 View project

The user has requested enhancement of the downloaded file.

Forgiveness, Feeling Connected to Others, and Well-Being: Two Longitudinal Studies

The online version of this article can be found at:

Society for Personality and Social Psychology, Inc.

Email Alerts: http://psp.sagepub.com/cgi/alerts

Citations (this article cites 38 articles hosted on the

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Forgiveness, Well-Being, and Social Connectedness

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Avoidance Revenge Benevolence Closeness Well-Being

Time Point M (SD) M (SD) M (SD) M (SD) M (SD)

Measure (Parameter) Coefficient SE t Value Effect Size r % VAFa χ2(1, N = 114)

Avoidance motivation as measure of forgiveness

Revenge motivation as measure of forgiveness

Benevolence motivation as measure of forgiveness

a. % VAF = percentage of variance accounted for by parameters.

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

TABLE 3: Mediation Models for Study 1 and Study 2a

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Method Well-Beingij = β0j + β1j(Time)ij + β2j(Well-Being)i – 1,j +

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Measure (Parameter) Coefficient SE t Value Effect Size r % VAFb χ2(1, N = 164)

Avoidance motivation as measure of forgiveness

Revenge motivation as measure of forgiveness

Benevolence motivation as measure of forgiveness

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

from their avoidance or revenge trajectories alone. On

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other in

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

Downloaded from http://psp.sagepub.com at UNIV OF RICHMOND on February 13, 2008

View publication stats

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.