ArticlePDF Available

Physical, Emotional, and Behavioral Reactions to Breaking Up: The Roles of Gender, Age, Emotional Involvement, and Attachment Style

Authors:

Abstract

Associations between gender, age, emotional involvement, and attachment style and reactions to romantic relationship dissolution were studied in a survey of more than 5,000 Internet respondents. It was hypothesized that individual reactions to breakups would be congruent with characteristic attachment behaviors and affect-regulation strategies generally associated with attachment style. Attachment-related anxiety was associated with greater preoccupation with the lost partner, greater perseveration over the loss, more extreme physical and emotional distress, exaggerated attempts to reestablish the relationship, partner-related sexual motivation, angry and vengeful behavior, interference with exploratory activities, dysfunctional coping strategies, and disordered resolution. Attachment-related avoidance was weakly and negatively associated with most distress/proximity-seeking reactions to breakups and strongly and positively associated with avoidant and self-reliant coping strategies. Security (low scores on the anxiety and avoidance dimensions) was associated with social coping strategies (e.g., using friends and family as "safe havens"). Attachment insecurity, particularly anxiety, was associated with using drugs and alcohol to cope with loss.
10.1177/0146167203252884 ARTICLE
PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS
Physical, Emotional, and Behavioral Reactions
to Breaking Up: The Roles of Gender, Age,
Emotional Involvement, and Attachment Style
Deborah Davis
University of Nevada, Reno
Phillip R. Shaver
University of California, Davis
Michael L. Vernon
University of Nevada, Reno
Associations between gender, age, emotional involvement, and
attachment style and reactions to romantic relationship dissolu-
tion were studied in a survey of more than 5,000 Internet respon-
dents. It was hypothesized that individual reactions to breakups
would be congruent with characteristic attachment behaviors
and affect-regulation strategies generally associated with attach-
ment style. Attachment-related anxiety was associated with
greater preoccupation with the lost partner, greater perseveration
over the loss, more extreme physical and emotional distress, exag-
gerated attempts to reestablish the relationship, partner-related
sexual motivation, angry and vengeful behavior, interference
with exploratory activities, dysfunctional coping strategies, and
disordered resolution. Attachment-related avoidance was weakly
and negatively associated with most distress/proximity-seeking
reactions to breakups and strongly and positively associated with
avoidant and self-reliant coping strategies. Security (low scores
on the anxiety and avoidance dimensions) was associated with
social coping strategies (e.g., using friends and family as “safe
havens”). Attachment insecurity, particularly anxiety, was
associated with using drugs and alcohol to cope with loss.
Keywords: breaking up; relationships; attachment style
The dissolution of romantic relationships has been
empirically associated with a variety of negative physical
and emotional responses, ranging from anxiety, depres-
sion, psychopathology, loneliness, immune suppression,
fatal and nonfatal physical illness or accidents, and
decreased longevity to immediate death through suicide
or homicide (see reviews in Gottman, 1994; Kiecolt-
Glaser & Newton, 2001). Although the empirical associa-
tion of relationship dissolution with physical and emo-
tional distress is well established, there have been few
attempts to examine behavioral responses and fewer still
to examine their unique links to individual-difference
variables.
The primary goal of the present research was to exam-
ine the relationship of adult attachment style (Fraley &
Shaver, 2000; Hazan & Shaver, 1987) to a constellation of
reactions to breaking up that are theoretically and
empirically linked to attachment anxiety and/or avoid-
ance. To date, there have been relatively few studies of
attachment style and reactions to relationship dissolu-
tion (e.g., Feeney & Noller, 1992; Simpson, 1990;
Sprecher, Felmlee, Metts, Fehr, & Vanni, 1998).
Although informative, these studies have focused mostly
on degree of distress rather than theoretically predicted
specific emotional and behavioral reactions. Our
research was designed to examine three primary dys-
functional reactions to dissolution: (a) extreme distress
and preoccupation with the lost partner, (b) ambivalent
acting out (strenuous attempts to reestablish the rela-
tionship, including sexual contact, combined with
angry, hostile, or violent behavior), and (c) dysfunc-
tional coping and lack of resolution of the loss.
Of these reactions, we are particularly interested in
the violent storm of ambivalent acting out that we expect
871
Authors’ Note: Correspondence concerning this article should be ad-
dressed to Dr. Deborah Davis, Department of Psychology, University of
Nevada, Reno, Reno, NV 89557; e-mail: debdavis@unr.nevada.edu.
PSPB, Vol. 29 No. 7, July 2003 871-884
DOI: 10.1177/0146167203252884
© 2003 by the Society for Personality and Social Psychology, Inc.
to be associated with attachment anxiety. This pattern is
of particular interest because of the extremity of the
bereft person’s attempts to maintain or reestablish the
relationship while engaging in apparently contradictory
angry and hostile behavior toward the partner, despite
the contradictory and self-defeating effects of such
behaviors. The extremes of this ambivalent pattern can
be seen, for example, in the murder of romantic part-
ners, especially of women, which occurs primarily under
conditions in which the victim is leaving or threatening
to leave a relationship with a partner who wants to keep
the relationship intact (e.g., Bixenstine, 1999; Walker,
1979; Wilson & Daly, 1993).
Attachment theory provides a theoretical basis for
predicting this ambivalent mixture of hostility and
desire as well as other specific emotional and behavioral
responses. Thus, we will first provide a brief overview of
attachment theory and research and then turn to their
implications for emotion regulation and behavior fol-
lowing relationship dissolution.
Attachment Theory and the
Assessment of Attachment Style
Attachment theory was introduced by Bowlby (1973,
1980, 1969/1982) in a well-known series of volumes
titled Attachment and Loss. Empirical tests of the theory
were initially conducted by Ainsworth and her col-
leagues (Ainsworth, Blehar, Waters, & Wall, 1978) in
studies of infant-mother attachment and were later
extended to the domain of romantic and marital rela-
tionships by Hazan and Shaver (1987). According to the
theory, evolution has equipped human beings with a
number of behavioral systems that increase the likeli-
hood of survival and reproductive success. Among these
behavioral systems are an attachment system, an explor-
atory system, a sexual mating system, and a caregiving
system.
Beginning in infancy, most people form emotional
attachments to one or more caregivers on whom they
rely for protection, comfort, and support. A security-
enhancing caregiver is one who provides a “safe haven”
in times of danger or stress and a “secure base” of opera-
tions when exploration is undertaken. If a person’s
attachment figures are sufficiently sensitive and respon-
sive, she or he will develop what the theory refers to as
positive internal working models of self and relationship
partners. These models have been shown to provide a
foundation for healthy peer relations and personal com-
petence (see review by Weinfield, Sroufe, Egeland, &
Carlson, 1999). If one or more attachment figures are
generally insensitive or unresponsive, the individual who
is attached to them develops negative internal working
models of self, relationship partners, or both. Different
patterns of insecure attachment can be identified based
on anxious or avoidant behaviors in close relationships,
anxious or avoidant responses to self-report question-
naires, and conscious and unconscious anxious or
avoidant responses in laboratory situations (see Feeney,
1999; Mikulincer & Florian, 2001; Shaver & Clark, 1994,
for reviews).
In Ainsworth et al.’s (1978) early studies of the infant-
mother relationship, three different patterns of attach-
ment were delineated: secure, anxious (or anxious/
ambivalent), and avoidant. Subsequent studies of both
infant and adult attachment (e.g., Bartholomew &
Horowitz, 1991; Main & Solomon, 1986) expanded the
number of patterns to four, with the fourth pattern
incorporating features of both anxiety and avoidance. In
the domain of adult romantic and marital attachment,
Bartholomew and Horowitz (1991) called the four pat-
terns secure, preoccupied (with attachment), fearfully
avoidant, and dismissively avoidant. These authors also
showed that the four patterns could be arranged concep-
tually and empirically in a two-dimensional space.
Secure and dismissing adults differ from preoccupied
and fearful adults in having more positive models of self
and being less dependent on partners’ approval and less
anxious about abandonment. Secure and preoccupied
adults differ from both kinds of avoidant adults by being
more interested in and comfortable with closeness, inti-
macy, and interdependence. Brennan, Clark, and
Shaver (1998) constructed two highly reliable orthogo-
nal scales to assess dimensions similar to Bartholomew’s,
which Brennan et al. called attachment-related anxiety
and avoidance that we used in the present research.
Attachment and Regulation of Distress
Theoretically, the attachment system is activated by
any of three sources of distress: (a) threat to the person
(e.g., hunger, physical danger), (b) threat to a relation-
ship with an attachment figure (e.g., perceiving the fig-
ure as physically or psychologically unavailable), and (c)
challenging situations that motivate the person to use an
attachment figure as a secure base. When a person’s
attachment system is activated for any of these reasons,
the person will attempt to alleviate distress in ways char-
acteristic of his or her attachment style. The stronger the
activation of the attachment system, the more extreme
the characteristic behaviors are likely to be.
These characteristic differences in affect regulation
are the bases of our hypotheses regarding attachment-
style differences in reaction to relationship breakups. In
brief, three primary strategies are associated with attach-
ment style: (a) open, empathic communication and
negotiation of one’s needs and desires with the attach-
ment figure (the secure strategy), (b) suppression of
attachment-related distress combined with self-reliance
(the avoidant strategy), and (c) a coercive strategy (e.g.,
872 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Crittenden, 1997) involving alternation between angry
demands and rebukes and coy or flirtatious attempts to
elicit what one needs from a partner (the anxious strat-
egy). Each strategy is assumed to stem from past experi-
ences with parents or other caregivers during childhood
as well as later experiences with romantic partners.
The secure strategy. The secure strategy is believed to
stem from past experiences indicating that open expres-
sion of needs elicits love and support. Generally, secure
individuals enjoy greater communication skills and pro-
vide superior caregiving to others (see reviews by Feeney,
1999; George & Solomon, 1999). They are more likely to
provide comfort to others in distress and to seek comfort
from others when distressed themselves. Thus, in the
context of relationship dissolution, we would expect
secure individuals to express their feelings openly to
their partner and to use friends and family as beneficial
sources of comfort. They also should be better able to
understand their partner’s point of view regarding the
breakup, which should allow them to respond in a less
histrionic or angry fashion than less secure individuals.
The avoidant strategy. Theoretically, avoidant individu-
als have learned (a) that others are unlikely to satisfy
their needs and (b) that open expressions of need may
be ignored or punished. This is believed to be the reason
for their unexpressive and self-reliant stance (see review
by Fraley, Davis, & Shaver, 1998). In the context of break-
ing up, this characteristic affect-regulation strategy
should be reflected in fewer emotional expressions of all
kinds (including pleading, angry outbursts, and seeking
social support), greater emotional avoidance (including
avoiding the partner and staying clear of other remind-
ers that could activate attachment needs), and greater
self-reliance and use of nonsocial coping strategies (such
as drinking and taking drugs).
The anxious strategy. Anxious or preoccupied individu-
als are thought to have learned what Crittenden (1992,
1997) termed a “coercive strategy” for eliciting care.
According to Crittenden, this strategy is characteristic of
children whose attachment figures are inconsistently
sensitive and available, causing the children to believe
that pleas and demands are necessary to get a caregiver’s
attention. In childhood, the coercive strategy includes
erratic alternation between aggressive/threatening
behavior (crying, screaming, throwing a tantrum) and
coy/disarming behavior (e.g., glancing eye contact, an
open-mouth smile with teeth covered, meek and inno-
cent expressions, cocking the head to the side). To the
extent that this analysis applies to anxious adults whose
relationships are breaking up, both aggressive and coy or
seductive behaviors may be employed in attempts to
restore the relationship.
The Attachment Perspective on Loss
In his 1980 volume Loss, Bowlby developed an attach-
ment-theoretical account of the process of grieving,
including individual differences in grief reactions. He
proposed that reactions to loss of an adult romantic part-
ner parallel those of a child confronted with the pro-
longed or permanent loss of a primary attachment fig-
ure. These reactions can be viewed in terms of three
rough, overlapping, and sometimes recurring phases.
Protest. The bereaved adult first exhibits forceful pro-
test reactions, designed to deter the attachment figure’s
departure or reestablish contact. These can include
overt reactions such as pleading, crying, anger, aggres-
sion, and searching and psychological reactions such as
disbelief that the person is gone and sensing the person’s
invisible presence. Bowlby noted that such reactions,
although seemingly inappropriate when a partner has
died, make sense in evolutionary context (see also
Archer, 1999) in that they tend to promote survival by
assuring the proximity of infants to their caregivers.
There may be no evolutionary provision for turning off
the search process when an attachment figure leaves per-
manently. Thus, real or perceived threats to the availabil-
ity of an attachment figure will activate such deeply
ingrained patterns of response as to virtually compel the
person to search for the lost figure and try to reestablish
contact (Fraley & Shaver, 1999), even when success is
unlikely or impossible, as when the partner has died.
Despair. Eventually, if protest behaviors repeatedly fail
to establish contact with the attachment figure and the
attached individual realizes that the person will never
return, protest increasingly gives way to despair (depres-
sion, sadness, disorganization, and withdrawal). For
both bereaved adults and children, this phase is charac-
terized by disturbances of sleeping and eating, social
withdrawal, profound loneliness, and intense sorrow
(Fraley & Shaver, 1999). The degree of despair differs as
a function of context (e.g., foreknowledge of the
impending loss), degree of attachment to the partner,
and individual differences in attachment style.
Detachment/reorganization/integration. In Bowlby’s
(1973) early theorizing, he referred to a final phase of
detachment marked by apparent recovery and gradual
renewal of interest in other activities and relationships.
In later theorizing, Bowlby (1980) used the term reorgani-
zation to connote the reorganization of representations
of the self and lost attachment figure so that both a con-
tinuing (but altered) bond and adjustment to changed
circumstances are possible. This notion is similar in
some respects to other perspectives on loss or bereave-
ment that emphasize the importance of meaning (e.g.,
see reviews in Neimeyer, 2001) and continuing bonds
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 873
(Klas, Silverman, & Nickman, 1996). Extending this
analysis to the context of relationship dissolution, we can
expect that a relatively secure person may come to
believe that the relationship provided a learning experi-
ence, that the lost partner makes a better friend than
lover, or that the loss enhanced personal strength.
Individual differences in “disordered mourning.”
Mourning may be disordered or disorganized such that
either the duration of the protest and despair phases or
the nature of behaviors during these phases becomes
dysfunctional. Bowlby (1980) identified two disordered
patterns of mourning, chronic mourning and absence of
grief, that correspond roughly to the subsequent notions
of anxious and avoidant attachment. The former pattern
includes perseveration in the protest and/or despair
stages of mourning, and the latter involves rapid pro-
gression to the detachment phase. Those in chronic
mourning are perpetually preoccupied with the lost
partner and unable to function normally without him or
her; those exhibiting absence of grief continue their nor-
mal everyday activities without conspicuous disruption
or overt expressions of sorrow, anger, or distress.
HYPOTHESES
The preceding overview of attachment theory and
research provides the basis for a series of hypotheses
about behavior during and following romantic/marital
relationship breakups. The first set of hypotheses con-
cerns protest reactions and distress.
Protest and Distress Reactions
Research with children (e.g., Ainsworth et al., 1978)
has identified the roots of the ambivalent storm of pro-
test reactions, vacillating between desire for the lost part-
ner and attempts to reestablish the relationship, on one
hand, and seemingly contradictory angry and violent
behaviors on the other. Ainsworth et al. noted this pat-
tern among anxious/ambivalent children, who dis-
played the most extreme distress and protest reactions
toward their parents in the Strange Situation test proce-
dure, including indices of distress (such as crying), prox-
imity seeking (such as pleading, clinging), and hostility/
aggression toward the parent (which Bowlby, 1973, inter-
preted as retributive anger). Avoidant children, in con-
trast, were characterized by the least extreme reactions
of these kinds, with secure children in between.
Although studies of reactions to relationship dissolu-
tion among adults have focused on greater distress reac-
tions among anxious adults (e.g., Feeney & Noller, 1992;
Fraley & Shaver, 1997; Simpson, 1990), the literature
regarding loss through death has turned up some evi-
dence of the association of attachment anxiety with the
ambivalent bipolar reactions of pining for the relation-
ship and being angry at the deceased partner (e.g.,
Bonnano, Notarius, Gunzerath, Keltner, & Horowitz,
1995).
Thus, we expect the extremity of both distress and
protest reactions to be positively associated with attach-
ment anxiety and negatively associated with attachment
avoidance. These reactions should include indices of
distress such as physical symptoms, emotional
disregulation, self-blame or depression, and protest
reactions such as attempts to reestablish the relation-
ship, sexual desire, and attempts to reinvolve the partner
in sexual relations but also expressions of anger, blame,
and hostility toward the partner. Furthermore, we expect
the two contradictory poles of desire and hostility to be
positively associated with one another.
Preoccupation/Exploration
A second set of hypotheses concerns excessive preoc-
cupation with the attachment figure, which is theoreti-
cally associated with both attachment anxiety and avoid-
ance. Constant attention to or thinking about the person
and hypervigilance and sensitivity to cues regarding his
or her availability and responsiveness tend to be posi-
tively associated with attachment anxiety and negatively
associated with avoidance. For example, Mikulincer,
Gillath, and Shaver (2002) found that the mind gener-
ally turns automatically to mental representations of
attachment figures under conditions of threat. However,
more anxious adults seem to have these representations
active all the time, whereas more avoidant adults activate
them only under conditions of threat that are unrelated
to attachment (e.g., school failure) and actually inhibit
attachment-related representations when the threat has
to do with separation.
Thus, we expected relationship dissolution to cause
greater preoccupation with thoughts of the lost partner
among those higher in attachment-related anxiety. This
greater preoccupation should in turn lead to differences
in exploration of and engagement with the environ-
ment, just as anxious/ambivalent attachment in infancy
interferes with exploratory activities (Ainsworth et al.,
1978; Bowlby, 1973; Cassidy & Berlin, 1994).
Coping
Attachment style also is expected to affect the choice
of coping strategies during and following relationship
dissolution.
Social coping versus self-reliance. Coping with distress
through seeking contact with attachment figures is a fun-
damental feature of attachment behavior. Securely
attached infants express distress freely and both seek
and accept comfort from caregivers (e.g., Ainsworth
et al., 1978). In contrast, avoidant infants learn to sup-
press expression of distress and bids for support to avoid
874 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
alienating potentially rejecting attachment figures and
instead tend to become compulsively self-reliant. Simi-
larly, secure adults employ attachment figures (includ-
ing friends) as safe havens under conditions of threat
(e.g., Simpson, Rholes, & Nelligan, 1992), whereas
avoidant adults are more likely to pull away from part-
ners under threatening conditions. Thus, we expected
attachment insecurity, particularly avoidance, to be neg-
atively related to social coping strategies (such as talking
to friends) and positively related to self-reliance.
Self-medication. Previous research has indicated that
anxious and avoidant individuals are more likely than
secure individuals to use alcohol and drugs as a way of
regulating negative emotion (e.g., Brennan & Shaver,
1995; Cooper, Shaver, & Collins, 1998). Thus, we expect
such dysfunctional methods also to be used specifically
in the context of relationship dissolution.
Suppression/avoidance. Recall Mikulincer et al.’s
(2002) finding that avoidance is related to suppression
of thoughts of the attachment figure, particularly when
relationship threat is salient. These findings suggest that
avoidance will be related to attempts to suppress
thoughts or reminders of the lost partner; for example,
by avoiding encounters with the partner, dispensing with
objects and other reminders of him or her, attempting to
distract oneself through work or other activities, and per-
haps, for a time, staying away from related and similarly
painful situations such as seeking new relationship
partners.
Resolution
Whether a partner is lost through death or dissolu-
tion, one must eventually go on with life. Individuals dif-
fer, however, in both the manner and speed with which
they resolve losses.
Perseveration. Some people have extreme difficulty
resolving losses at all and persist in feeling distress, con-
tinuing desire to reestablish the relationship, and preoc-
cupation with the lost partner. Bereavement research
has shown that anxious individuals experience the most
distress over loss of a partner and recover less quickly, if
at all (see Fraley & Shaver, 1999, for review). Thus, we
expected that anxiety also would be associated with
perseveration in desire for or attempts to recover a part-
ner lost through relationship dissolution.
Integration. Several authors (e.g., Bowlby, 1980; Klas
et al., 1996; Schuchter & Zisook, 1993) have proposed
that successful resolution of bereavement (following the
death of an attachment figure) does not necessarily
involve detachment. Instead, an altered attachment
bond may persist, such as caring for the deceased person
and integrating thoughts, memories, and feelings about
the relationship into one’s self-concept and life while
being able to move on and live happily again. Such a con-
tinuing bond may serve important adaptive functions,
including continuity of identity, facilitation of coping,
and comfort and support during the transition to a new
life. This type of reorganization or integration should be
more likely for relatively secure individuals. Thus, we
expected those high in attachment anxiety or avoidance
to less often successfully transform the relationship into
a friendship or work relationship.
Disordered identity. Part of the integration process
involves reorganizing and redefining one’s conception
of self without (or in different relationship to) the other
person. Thus, we expected that to the extent integration
is difficult for those high in attachment anxiety, they
would report a greater sense of lack of identity following
loss of a partner.
Replacement. Paradoxically, we also expected attachment-
related anxiety to be associated with quicker attempts to
replace the partner. Because anxious individuals experi-
ence more distress and continuing attachment to their
lost partners, one might think they would be inhibited
from entering other relationships. There is substantial
evidence, however, that such people are highly moti-
vated to be in a relationship and that they experience
great distress when alone (e.g., Davis, 2000; Shaver &
Clark, 1994). They may therefore jump into a new rela-
tionship while still distressed about the loss of a previous
one. In contrast, as noted earlier, we expected avoidance
to be associated with avoiding new relationships for
some time after a breakup.
METHOD
Participants
Study participants were 1,868 male and 3,380 female
respondents (and 7 with unidentified gender) ranging
in age from 15 to 50 and distributed across the age
decades as follows: teens: 42.2%, 20s: 43.2%, 30s: 10.5%,
40 to 50: 4.1%. (We eliminated from the analyses all par-
ticipants who were younger than 15 or older than 50
years of age because age appeared to be related to many
variables and there were too few respondents in the very
young and older than 50 age ranges.) The majority were
Caucasian (77.7%), followed in order of frequency by
Other (6.5%), African American (6.2%), Hispanic
(4.3%), Asian (4.1%), and American Indian (.9%);
91.3% were heterosexual, 2.6% were homosexual, and
6.5% were bisexual.
Procedure
Our survey questionnaire was posted on the Internet
with the title “The Dating Survey IV: Breaking Up.” Par-
ticipants voluntarily followed links to the survey located
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 875
in three subcategories of the Yahoo search engine. Invi-
tations to visit the Internet site were phrased as follows:
“Dating Survey—Participate in the first study of Internet
singles.” Categories with links to the survey included
Dating (under the parent category “Society and Cul-
ture/Relationships”), Tests and Experiments (under the
parent category “Psychology/Research”), and Surveys
(also under the parent category “Society and Culture/
Relationships”). Participants were not actively solicited
in any way. The survey included assurances that
responses would be completely anonymous once trans-
mitted. However, it also included a warning that (like all
online communications) responses were not secure
until transmitted.
The survey was introduced as follows:
The purpose of the survey is to learn more about what
causes our relationships to break up, and how we cope
with breakups when they occur. To examine this issue we
will be asking you a few questions about yourself, and
then some that address various issues regarding break-
ing up.
The questionnaire included the questions, “Are you at
your computer alone?” and “Have you ever sent a re-
sponse to this survey before?” Those who were not alone
or who had responded before were excluded from the
analyses, as were respondents who described a breakup
that occurred more than 5 years ago, which might have
been beyond the reach of clear memory.
Measures
Attachment style. Attachment-related anxiety and
avoidance were measured by heterogeneous 9-item sub-
sets of the two 18-item scales that comprise the Experi-
ences in Close Relationships measure (Brennan et al.,
1998). Alphas for the two shortened scales were .90 and
.85, only slightly lower than the usual reliabilities for the
full scales. The correlation between the two scales, which
are meant to tap orthogonal dimensions, was close to
zero, r(4,958) = –.004.
Measures of reactions to breaking up. At the beginning of
the breakups section of the survey, participants were
instructed as follows:
To answer the questions in the following section, think of
the breakup of your last relationship that had lasted for
some time before the breakup or in which you were seri-
ously emotionally involved. Do not refer to a relation-
ship that broke up after a few dates.
Measures of the following reactions were included: pro-
test reactions, including (a) wanting/trying to get the
person back (11 items, α= .89), (b) sexual arousal/de-
sire (3 items, α= .74), (c) anger/hostility/revenge (10
items, α= .84), and (d) physically hurting partner (1
item); preoccupation (9 items, α= .91); interference
with exploration (3 items, α= .70); distress reactions, in-
cluding (a) physical and emotional distress (15 items, α
= .94), (b) lost interest in sex (1 item), (c) self-blame (2
items, α= .63), (d) guilt (2 items, α= .83), and (e) part-
ner blame (2 items, α= .81); coping strategies, including
(a) social coping (1 item), (b) self-reliant coping (1
item), (c) self-medication (increased use of drugs or al-
cohol—2 items, reported separately), (d) suppression/
avoidance (3 items, α= .63), and (e) moving or changing
jobs to get away from the person; and resolution, includ-
ing (a) perseveration in wanting the lost partner (3
items, α= .83), (b) integration/redefinition of the per-
son in different relationship (2 items, α= .57), (c) lost
sense of identity (1 item), (d) replacement of the lost
partner (1 item), and (e) avoiding new relationships for
a long time (1 item).
Respondents completed demographic questions first,
followed by the attachment measures. Third, they indi-
cated who had wanted to terminate the relationship
(self, partner, or both), how long the relationship had
lasted, how long ago the relationship broke up, and how
emotionally involved they were with the person at the
time of the breakup. This latter issue was raised because
emotional involvement might be a somewhat independ-
ent contributor to intensity of reactions to the breakup.
Measures of reactions to breaking up were completed
last.
RESULTS
Before turning to tests of the hypotheses, it is impor-
tant to examine potential associations between attach-
ment style and predissolution relationship variables.
First, 2 (gender) ×3 (who initiated the breakup) analyses
of variance were conducted on the two attachment scales
and the question assessing emotional involvement in the
relationship. Gender was unrelated to attachment anxi-
ety or avoidance. However, women were significantly
more emotionally involved with their partners prior to
the breakups, Ms = 7.42, 7.25; F(1, 4918) = 32.03, p< .001.
Furthermore, those who initiated the breakup were
lower in anxiety, Ms = 3.92, 4.07, 4.69; F(2, 4774) =
156.83, p< .001, and emotional involvement, Ms = 6.95,
7.12, 7.80; F(2, 4918) = 140.22, p< .001, and higher in
avoidance, Ms = 3.82, 3.71, 3.59; F(2, 4918) = 13.77, p<
.001, than those who mutually initiated the breakup or
whose partners initiated the breakup.
Anxiety was positively related to emotional involve-
ment at the time of the breakup, r(5,015) = .17, p< .001,
and negatively related, although minimally, to the
amount of time in the relationship, time since the
breakup, and age, rs(5,020-5,065) = –.06, –.04, –.07, ps<
.001.
876 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
In contrast, avoidance was negatively related to emo-
tional involvement at the time of the breakup, r(5,074) =
–.22, p< .001, and positively, although only weakly,
related to age, r(5,124) = .06, p< .001. Finally, emotional
involvement was positively related to time in the relation-
ship prior to breaking up, r(5,152) = .19, p< .001, and less
so, to age r(5,199) = .07, p< .001.
Tests of Hypotheses
Each reaction-to-breakup measure was subjected to
regression analyses in which independent variables were
entered in three blocks. The first block included age,
gender, time since breakup, person who initiated the
breakup, and emotional involvement in the relation-
ship. Measures of anxiety and avoidance were entered in
the second block and all two-way interactions in the
third. Time since breakup interacted with effects of anxi-
ety and/or avoidance for most analyses such that the
effects were slightly larger for more recent breakups.
However, comparing those who had broken up within
the past 3 months to the full sample, effects of anxiety or
avoidance were neither reduced to insignificance nor
inflated to significance, and regression coefficients
changed by .03 or less. Thus, time since breakup was
dropped from further analyses to simplify presentation
of the results.
Results are presented in Tables 1 through 3. Beta coef-
ficients and significance levels for main effects as well as
tests of significance for the overall regression equations
are included in each table—for the overall sample and
for both partner-initiated and self-initiated breakups.
Results for mutual breakups are not reported separately.
There were a number of small but significant interac-
tions where effect sizes were modified by a second vari-
able. These are not reported in the text except where the
effects of anxiety or avoidance are reduced to insignifi-
cance. We also will only selectively report effects of age or
gender where they are of particular interest, because
they are available in the tables but were neither pre-
dicted nor relevant to our hypotheses.
DISTRESS
We expected the degree of distress experienced to be
a function of emotional involvement at the time of the
breakup and anxiety. Table 1 summarizes the regression
analyses for four forms of distress: physical/emotional
distress, lost interest in sex, self-blame, and guilt. An
index of partner blame is included so that it can be con-
trasted with the self-blame measure.
As predicted, emotional involvement was significantly
associated with all indices of distress (βs ranged from .03
to .34, ps ranged from less than .05 to less than .001),
although the relationship was strongest for physical/
emotional distress and quite small for the index of self-
blame (see Table 1). Emotional involvement also was
associated with partner blame, β= .09, p< .001. Those
who initiated the breakup felt less distress, β= .25, p<
.001, but more guilt, β= –.28, p< .001, than those whose
partners initiated the breakup. As expected, attachment-
related anxiety was significantly associated with all indi-
ces of distress, βs = .11 to .34, all ps < .001, and with part-
ner blame, β= .09, p< .001. There were also small but sig-
nificant associations between attachment avoidance and
both self- and partner-blame such that avoidance was
associated with greater self-blame, β= .10, and less part-
ner blame, β= –.08.
PROTEST REACTIONS
The fifth through eighth sections of Table 1 summa-
rize regression coefficients for the four protest variables.
As expected, both emotional involvement, βs = .26, .11,
.10, .04, ps < .001, and attachment anxiety, βs = .20, .18,
.20, .06, ps < .001, were associated with the two proximity-
seeking variables of wanting/trying and sexual arousal
and the two hostile indices of anger/hostility/revenge
and physically hurting. Similar patterns were obtained
for both partner-initiated and self-initiated breakups
(see Table 1). The strongest associations were with want-
ing and attempting to regain the lost partner, whereas
the weakest associations were with reports of physically
hurting the lost partner, which was unusual in this sam-
ple (see below).
Two gender differences are of interest. First, women
were less likely, β= –.14, p< .001, to report sexual
arousal/desire. However, women were more likely to
report anger/hostility, β= .11, p< .001, and violence
(among those whose partners initiated the breakup, β=
.05, p< .05).
Because violence was an extremely low-base-rate
behavior in our sample, we performed a separate analy-
sis to examine differences in anxiety and avoidance lev-
els among those who reported violence and those who
did not. The 9-point ratings of “physically hurt him/her”
(not at all true to extremely true) were recoded such that not
at all true was coded as 0 and all other responses were
coded as 1. Then, 2 (physical violence, i.e., hurt or didn’t
hurt) ×3 (who initiated the breakup) analyses of vari-
ance, with anxiety and avoidance as dependent vari-
ables, were conducted.
As expected, attachment anxiety was higher among
those who hurt their partners than among those who did
not, Ms = 4.53 (yes) and 4.18 (no), F(1, 4652) = 56.37, p<
.001. The interaction of violence with who initiated the
breakup was not significant, indicating that regardless of
how the breakup was initiated, those who were physically
violent with their partners were higher in attachment
anxiety than those who were not. Avoidance did not dif-
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 877
fer between those who did and did not hurt their
partners.
ASSOCIATION BETWEEN DESIRE
AND ANGER/HOSTILITY
Although desire for the relationship to continue and
extreme anger/hostility toward the partner might seem
logically incompatible, we expected them to go hand in
hand, as explained by Bowlby (1973) in Attachment and
loss: Separation, Anxiety, and Anger. To examine this
hypothesis, we first examined the zero-order correla-
tions between the four protest variables. All proved to be
significantly positive (rs = 3814-4775) and ranged from
.07 to .48, all ps < .001). The correlations between want-
ing/trying to get the person back and sexual arousal
(.47), anger/hostility/revenge (.21), and physical vio-
lence (.11) were all significant. Similarly, those between
hostility/revenge and sexual arousal (.11) and physical
violence (.30) were significant. The relationship
between sexual arousal and physical violence was small
(.07).
Second, a regression analysis was performed to exam-
ine the relationship of gender, age, initiation of breakup,
878 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 1: Distress/Protest Reactions to Breaking Up: Regression Coefficients for Total, Partner-Initiated Breakup, and Self-Initiated Breakup
Samples
Emotional
Type of Reaction Gender Age Who Initiated Involvement Anxiety Avoidance
Distress reactions
Physical/emotional distress –.01 –.01 .25*** .34*** .34*** .03**
Partner initiated .01 .00 .37*** .41*** –.01
Self-initiated –.01 .02 .38*** .32*** .05*
Lost interest in sex .06*** .07*** .13*** .21*** .11*** .10***
Partner initiated .04 .08** .19*** .13*** .09***
Self-initiated .08** .11*** .21*** .05* .09**
Self-blame –.11*** –.10*** –.06*** .03* .11*** .10***
Partner initiated –.07*** –.08*** .08*** .18*** .09***
Self-initiated –.13*** –.13*** –.02 .01 .08***
Guilt –.11*** –.06*** –.28*** .17*** .14*** .01
Partner initiated –.15*** –.08*** .13*** .18*** .04
Self-initiated –.05 –.06* .19*** .09*** –.03
Partner blame .09*** .15*** –.03 .09*** .09*** –.08***
Partner initiated .09*** .15*** .03 .02 –.08***
Self-initiated .07** .15*** .12*** .09*** –.08***
Protest reactions
Want/try to get back –.03 –.10*** .38*** .26*** .20*** .01
Partner initiated .01 –.10*** .30*** .29*** –.03
Self-initiated –.07** –.10*** .31*** .12*** .03
Sexual arousal –.14*** –.07*** .14*** .11*** .18*** .05*
Partner initiated –.07** –.05 .06* .22*** –.01
Self-initiated –.20*** –.05 .18*** .12*** .09***
Anger/hostility/revenge .11*** –.02 .06*** .10*** .20*** .02
Partner initiated .14*** –.00 .09*** .17*** .05*
Self-initiated .08*** –.02 .11*** .19*** .00
Physically hurt .02 –.01 –.02 .04* .06*** .02
Partner initiated .05* .02 .01 .09*** .01
Self-initiated .02 –.03 .06* .04 .03
Preoccupation –.03* –.02 .33*** .37*** .27*** .08***
Partner initiated .00 –.03 .38*** .36*** .03
Self-initiated –.04 .02 .44*** .23*** .12***
Exploration/interference –.03* –.04** .14*** .25*** .20*** .00
Partner initiated –.02 –.05* .23*** .21*** –.02
Self-initiated –.01 –.02 .27*** .19*** .01
NOTE: Men were scored as 1, women as 2. Initiation of breakups was scored as 1 (self), 2 (both), and 3 (partner). Ns = 4484-4889 (total), 2018-2175
(partner-initiated), and 1682-1838 (self-initiated). For the variables of sexual arousal and lost sexual interest, respondents were selected based on
having had a relationship involving sexual intercourse. Distress reactions: Significance levels for the regression equations were as follows: Fs(6, 3258-
4883) = 523.90, 62.91, 31.92, 80.19, 40.31, ps < .001 (total); Fs(5, 1458-2170) = 206.60, 21.40, 26.55, 34.03, 13.48, ps < .001 (partner initiated); Fs(5,
1248-1833) = 122.90, 19.32, 13.46, 19.19, 21.73, ps < .001 (self-initiated). Protest reactions: Significance levels for the regression equations were as fol-
lows: Fs(6, 3228-4545) = 429.15, 299.75, 60.76, 4.59, 429.15, 159.52, ps < .001 (total); Fs(5, 1436-2056) = 109.83, 97.74, 26.54, 4.12, 177.07, 52.01, ps<
.001 (partner-initiated); Fs(5, 1234-1692) = 46.07, 21.93, 22.71, 2.48, 117.89, 45.97, ps < .001, except for the third from last analysis, where p=.03
(self-initiated).
*p< .05. **p< .01. ***p< .001.
emotional involvement, anxiety, avoidance, and want-
ing/trying to get the person back as predictors of angry/
hostile/vengeful behavior. In addition, a second analysis
examined predictors of wanting/trying to reestablish
the relationship, including hostility as a predictor.
The regression predicting desire was significant, F(7,
4516) = 379.37, p< .001. Desire was significantly pre-
dicted by all variables except avoidance, βs = –.04 to .37,
ps ranging from .01 to less than .001, including the
Anger/Hostility/Revenge measure, β= .19, p< .001.
However, although the overall regression was again
significant, F(7, 4516) = 59.96, p< .001, only gender, β=
.10, p< .003, attachment anxiety, β= .20, p< .001, and
wanting/trying to get the person back, β= .15, p< .001,
predicted angry/hostile/vengeful behavior. (Without
the desire variable in the equation, emotional involve-
ment was a significant predictor of the hostility measure;
but this relationship disappeared when the desire vari-
able was included.) Clearly, the theoretical link between
processes of proximity seeking/attempting to reestab-
lish the relationship and the experience/expression of
anger were supported in these analyses.
PREOCCUPATION/INTERFERENCE
Both emotional involvement, βs = .37, .25, ps < .001,
and attachment anxiety, βs = .27, .20, ps < .001, were asso-
ciated with stronger preoccupation with the lost partner
and interference with exploratory activities such as
school and work, as can be seen in the first two sections of
Table 2. Preoccupation and inter ference also were stron-
ger among respondents whose partners initiated the
breakup, βs = .33, .14, respectively, ps < .001.
COPING STRATEGIES
Table 2 summarizes regression analyses for social (vs.
self-reliant), avoidant, and self-medicating coping
strategies.
Social versus self-reliant coping. As predicted, anxiety was
positively and avoidance negatively associated with social
coping, βs = .14, –.17, ps < .001, whereas avoidance was
positively associated with self-reliant coping, β= .20, p<
.001. In addition, respondents whose partners initiated
the breakup and those higher in emotional involvement
were slightly more likely to use social, βs = .12, .22, ps<
.001, and less likely to use self-reliant coping, βs = –.04,
–.04, ps < .01.
Avoidant coping. The third and fourth sections of Table
2 summarize regression coefficients for coping by avoid-
ing the former partner and taking the extreme actions of
moving to another town or changing jobs to avoid the
person. For the latter measure, only those who reported
having a job before the breakup and being older than 21
(i.e., those who might be able to move if they desired)
were included.
The expected association of attachment avoidance
with the measure of avoiding the partner was small but
significant only for those whose partners initiated the
breakup, β= .07, p< .01. Attachment anxiety was associ-
ated with avoiding the person, but only for self-initiated
breakups, β= .11, p< .001. Those whose partners initi-
ated the breakup, β= –.09, p< .001, and those more emo-
tionally involved with their partners, β= –.10, p< .001,
were less likely to avoid former partners.
Results for the measure of moving or changing jobs to
avoid the partner were similar but even weaker (see
Table 2). However, in contrast to results for the more
general measure of avoidance, emotional involvement
was positively related to the tendency to move or change
jobs to avoid a former partner, β= .13, p< .001. Similar to
physical violence, moving and changing jobs are low-
base-rate behaviors. We therefore performed analyses of
variance to assess differences in emotional involvement
and attachment-related anxiety and avoidance between
those who did and those who did not move and those
who did and did not change jobs. The 9-point scales were
recoded such that not at all was coded as 0 and all other
responses were coded as 1.
Those who changed jobs to avoid their partner were
higher in attachment anxiety, Ms = 4.62, 4.23, F(1, 3235)
= 20.53, p< .001, and avoidance, Ms = 3.81, 3.69, F(1,
3267) = 3.93, p< .05, and emotional involvement, Ms=
7.68, 7.43, F(1, 3324) = 4.64, p< .05, than those who did
not. Similarly, those who moved to avoid the person were
higher in anxiety, Ms = 4.40, 4.18, F(1, 1997) = 7.72, p<
.006, and emotional involvement, Ms = 7.82, 7.45, F(1,
2056) = 11.78, p< .001, than those who did not. Avoid-
ance was not related to moving to avoid the person.
Self-medication. The final two sections of Table 2 con-
tain coefficients for coping through drugs or alcohol.
Only respondents who reported using alcohol before
the breakup were included in the alcohol analysis, and
only those who used drugs prior to the breakup were
included in the drug analysis. As expected, both anxiety,
βs = .13, .19, ps < .001, and avoidance, βs = .08, .14, ps<
.001, were associated with both alcohol and drug use fol-
lowing a breakup.
RESOLUTION
Table 3 presents regression results for five measures
of resolution of the loss.
Perseveration/lost sense of identity. As can be seen in the
top two sections of Table 3, both perseveration and lost
sense of identity were higher among those whose part-
ners initiated the breakup, βs = .29, .16, ps < .001, those
higher in emotional involvement, βs = .35, .30, ps < .001,
and those higher in attachment anxiety, βs = .13, .25, ps<
.001, as expected.
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 879
Integration. Also as expected, continuing to relate to
the partner, but in a different role (integration), was
negatively associated with both anxiety and avoidance,
although the associations were small, βs = –.05, –.06, ps<
.001.
Replacement. As expected, attachment anxiety was
associated with a tendency to jump immediately into a
new relationship, β= .17, p< .001. This tendency also was
greater among those who initiated the breakup, perhaps
because they were already beginning or contemplating
another relationship while they were breaking off the
old one, β= –.15, p< .001.
Avoidance of new relationships. As expected, attachment
avoidance was associated with the tendency to avoid new
relationships after the breakup, β= .15, p< .001. This ten-
dency also was associated with age, partner-initiated
breakups, emotional involvement, and attachment anxi-
ety, βs = .11, .14, .19, .07, ps < .001. Thus, it appears that
more anxious respondents tended either to swear off
relationships or immediately jump into a new one.
DISCUSSION
Although a few studies have shown that attachment
anxiety is associated with enhanced distress upon rela-
tionship dissolution (Feeney & Noller, 1992; Fraley &
Shaver, 1997; Simpson, 1990; Sprecher et al., 1998), the
present study provides the first demonstration of
attachment-related reactions to breakups ranging from
protest through coping to eventual resolution.
Distress. Bowlby (1969/1982) proposed that threat to
the availability of an attachment figure would result first
in distress, varying in magnitude partly as a function of
the degree of attachment to the attachment figure. Con-
sistent with this expectation, we found that people who
reported more emotional involvement with their part-
ners at the time of the breakup exhibited greater distress
of all kinds, including both physical and emotional dis-
tress, lost interest in sex, self-blame, and guilt. Similarly,
those whose partners initiated the breakup reported
greater physical/emotional distress and lost interest in
sex but less self-blame and guilt.
880 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 2: Coping Strategies: Regression Coefficients for Total, Partner-Initiated Breakup, and Self-Initiated Breakup Samples
Emotional
Type of Reaction Gender Age Who Initiated Involvement Anxiety Avoidance
Social: Talk friends/family .15*** –.02 .12*** .22*** .14*** –.17***
Partner initiated .20*** –.01 .12*** .13*** –.21***
Self-initiated .12*** –.03 .27*** .14*** –.14***
Self-reliant: Handle self without friends/family –.15*** .02 –.04** –.04** .01 .20***
Partner initiated –.17*** .01 –.01 .02 .17***
Self-initiated –.12*** .05 –.07** .01 .21***
Avoidant
Avoid person –.00 .10*** –.09*** –.10*** .08*** .05**
Partner initiated –.03 .11*** –.05* .00 .07**
Self-initiated –.02 .10*** –.15*** .11*** .04
Move/change jobs –.01 .01 –.06* .13*** .06* .06*
Partner initiated –.01 .01 .15*** .11** .06
Self-initiated .02 .01 .13** –.02 .05
Self-medication
Alcohol –.11*** –.06** .04* .20*** .13*** .08***
Partner initiated –.16*** –.05 .18*** .14*** .08**
Self-initiated –.06 –.08* .20*** .12*** .08*
Drugs –.15*** –.10** .03 .21*** .19*** .14***
Partner initiated –.13*** –.04 .19*** .20*** .16**
Self-initiated –.13*** –.15** .25*** .18** .18**
Drugs/alcohol –.11*** –.11*** .03 .20*** .13*** .10***
Partner initiated –.14*** –.11*** .18*** .14*** .11***
Self-initiated –.06 –.13*** .21*** .13*** .11***
NOTE: Men were scored as 1, women as 2. Initiation of breakups was scored as 1 (self), 2 (both), and 3 (partner). For the alcohol and drug ques-
tions, respondents were selected to report having drunk or used drugs prior to the breakup (Ns = 1704, 795, 622 for alcohol; 2391, 1096, 893 for
drugs). For the measure including changing jobs or moving to avoid the person, respondents were selected to be older than 21 and to have had a job
at the time of the breakup (Ns = 759, 339, 298). For other measures, Ns = 4541-4549 (total), 2018-2175 (partner-initiated), 1682-1838 (self-initi-
ated). Significance levels for the regression equations were as follows: Fs(6, 759-4883) = 162.65, 58.34, 26.13, 6.22, 36.95, 16.87, ps < .001 (total);
Fs(5, 334-2057) = 57.11, 26.01, 9.01, 6.20, 19.16, 7.74, ps < .001 (partner initiated); Fs(5, 293-1833) = 63.72, 28.23, 14.05, 2.08, 11.88, 7.74, ps < .001,
.001, .001, .07, .001,.001 (self-initiated).
*p< .05. **p< .01. ***p< .001.
Our results replicate previous findings of an associa-
tion between attachment anxiety and enhanced distress
due to relationship threats of all types, ranging from
temporary separation to permanent dissolution through
breaking up or death (Ainsworth et al., 1978; Feeney &
Noller, 1992; Fraley & Shaver, 1999; Simpson, 1990).
Unlike previous research, however, the present data
reveal a link between some distress reactions and avoid-
ance. In particular, avoidance was associated with
enhanced self-blame (and reduced partner-blame) and
lost interest in sex.
Anger and desire. Bowlby (1988) argued that both anxi-
ety and anger are natural and effective responses to
threat to any important relationship. Anger (in the right
amount, time, and place) can deter dangerous behavior,
disloyalty, or neglect and coerce the partner into exhibit-
ing the desired behavior. Bowlby further argued that
maladaptive family violence or aggression can be under-
stood as exaggerated versions of attachment behaviors
that (when more appropriate in form or level) are poten-
tially functional. Thus, anger or aggression derived from
relationship anxiety may be rooted in an attempt to pro-
tect the relationship but become so extreme as to
threaten it instead.
Crittenden’s (1997) analysis of the “coercive” strategy
suggests that angry/aggressive strategies for maintain-
ing contact or eliciting caregiving are particularly char-
acteristic of anxious/ambivalent/preoccupied individu-
als. Furthermore, her analysis of the association between
anxious attachment and the aggressive pole of the coer-
cive strategy has been clearly supported in studies of
both infants and adults. For example, Ainsworth and her
colleagues (Ainsworth et al., 1978) reported that both
anxious and avoidant attachment in infants was associ-
ated with greater displays of anger toward the parent
(although avoidant infants tended to display greater
anger at home but not in the Strange Situation). Adult
partner violence (physical and verbal), however, has
been associated with preoccupied and fearful attach-
ment (both involving high attachment anxiety) but not
with dismissing attachment (e.g., Dutton, 1998).
Furthermore, although not framing their research in
attachment terms, Downey, Feldman, and Ayduk (2000)
provided evidence that “rejection sensitivity” (defined as
the disposition to anxiously expect, readily perceive, and
intensely react to rejection by significant others) pre-
dicted dating violence among those who were highly
invested in the relationship. They also reviewed evidence
showing that rejection is often a trigger of male violence
toward romantic partners (see also Bixenstine, 1999;
Walker, 1979; Wilson & Daly, 1993).
The association between attachment anxiety and
angry responses to relationship threat has also been
clearly articulated in the bereavement literature. Bowlby
argued, for example, that ambivalence during bereave-
ment manifests itself as both yearning for the deceased
and anger over being abandoned. Some support for this
prediction was provided by Bonnano et al. (1995), who
showed that ambivalence regarding the lost partner was
associated with facial expressions of both anger and sad-
ness when a person was talking about his or her loss.
Unlike a dead partner, however, a live but unavailable
partner provides a ready target for angry responses.
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 881
TABLE 3: Resolution: Regression Coefficients for Total, Partner-Initiated Breakup, and Self-Initiated Breakup Samples
Emotional
Type of Reaction Gender Age Who Initiated Involvement Anxiety Avoidance
Perseveration –.07*** –.07*** .29*** .35*** .13*** .04**
Partner initiated –.04 –.08*** .31*** .20*** .02
Self-initiated –.07** –.03 .43*** .08*** .05*
Lost sense of identity –.03* –.07*** .16*** .30*** .25*** –.01
Partner initiated –.03 –.07*** .34*** .28*** –.05**
Self-initiated –.02 –.05* .30*** .24*** .01
Integration .02 –.13*** .05*** .06*** –.05*** –.06***
Partner initiated .04 –.16*** .04 .03 –.07**
Self-initiated .04 –.12*** .09*** –.08*** –.05*
Look for new partner –.06*** –.04* –.15*** –.08*** .17*** –.07***
Partner initiated –.03 –.01 –.05* .14*** –.01
Self-initiated –.09*** –.07** –.09*** .20*** –.11***
Avoid new relationships –.01 .11*** .14*** .19*** .07*** .15***
Partner initiated –.00 .14*** .18*** .09*** .09***
Self-initiated –.01 .13*** .18*** .03 .18***
NOTE: Ns = 4477-4554 (total), 2053-1838 (partner initiated), 1658-1698 (self-initiated). Significance levels for the regression equations were as fol-
lows: Fs(6, 4471-4548) = 322.68, 252.86, 21.93, 42.32, 87.67, ps < .001 (total); Fs(5, 2034-2054) = 71.39, 121.57, 16.22, 9.23, 30.73, ps < .001 (partner-
initiated); Fs(5, 1248-1833) = 73.79, 66.83, 11.04, 25.32, 25.91, ps < .001 (self-initiated).
*p< .05. **p< .01. ***p< .001.
The present results further supported the theoretical
link between anger and desire in that all positive and
negative protest reactions were positively associated with
one another. Those who most wanted their partner back
were also most hostile and aggressive—as further shown
by the association of reported emotional involvement
with the partner at the time of the breakup and the
strength of both angry/hostile response and desire for
the lost partner.
Preoccupation. Related to strength of desire for a lost
partner is strength of preoccupation in thinking about
him or her. Relationship anxiety is associated with
higher resting levels of preoccupation with an attach-
ment figure (Mikulincer et al., 2002) and with enhanced
preoccupation under conditions of relationship threat
(e.g., the temporary absence of mother in the Strange
Situation) (Ainsworth et al., 1978). Studies of bereave-
ment have also shown that relationship anxiety is associ-
ated with “chronic mourning” and preoccupation with
the dead partner (see review by Fraley & Shaver, 1999).
Consistent with these earlier findings, in the present
study, preoccupation was greater among those whose
partners initiated the breakup, those who were more
emotionally involved with their partners, and those high
in attachment anxiety.
Exploration. Studies of infant attachment identified
the link between preoccupation with the attachment fig-
ure and reduced exploration of the environment (e.g.,
Ainsworth et al., 1978; Bowlby, 1973; Cassidy & Berlin,
1994); later, Hazan and Shaver (1990) showed that anx-
ious attachment in adulthood is associated with reduced
exploration in the arena of work. Similarly, the present
research linked attachment anxiety to interference with
exploration in the form of school or work. Also as
expected, such interference was greater for those whose
partners initiated the breakup and those more emotion-
ally involved with their lost partners.
Clearly, the patterns of reactions to threat to an
attachment relationship characteristically associated
with attachment anxiety and avoidance appear in reac-
tions to the dissolution of romantic relationships. First,
the results provided support for the expected relation-
ship between attachment-related avoidance and self-reli-
ant, nonsocial coping strategies. Avoidant attachment
was associated with less use of friends and family and
greater self-reliance, as expected. Furthermore, avoid-
ance was associated with avoidance of the former part-
ner, even to the extent of changing jobs (although the
relationship was necessarily small, given the low base rate
of changing jobs). Unexpectedly, attachment anxiety
was also associated with avoidance of the partner, includ-
ing both moving and changing jobs to avoid the person.
Although apparently inconsistent with greater proximity
seeking and attempts to maintain the relationship
among those higher in attachment anxiety, the tendency
to avoid the partner may be characteristic of what
Bartholomew and Horowitz (1991) called fearful
avoidants, who are high in both anxiety and avoidance.
Coping. Generally, we expected both anxiety and
avoidance to be associated with dysfunctional coping
strategies. Perhaps the most dysfunctional strategies are
self-destructive strategies such as use of drugs or alcohol.
As expected, use of these strategies was associated with
both forms of attachment insecurity.
The use of all forms of coping might be more likely
and more intense among people who experience
greater distress. In line with this reasoning, those whose
partner terminated the relationship, those who were
more emotionally involved in the relationship, and
those high in attachment anxiety were more likely to
seek support from friends and family and to use drugs or
alcohol. Thus, it may be that those who experience the
most distress also must try harder, in any way compatible
with their characteristic distress-regulation strategies, to
soothe the distress of the breakup, including the use of
drugs or alcohol.
In contrast, although avoidance was not associated
with enhanced distress, it was associated with specific
characteristic coping strategies, all of them involving
self-reliance and avoidance. Both avoidance of former
partners and self-medication may be attempts to sup-
press attachment-related thoughts and feelings. In fact,
there is some evidence that avoidants (particularly those
low in anxiety) can successfully suppress attachment-
related distress—if they can avoid direct reminders of
the relationship (Fraley et al., 1998; Fraley & Shaver,
1997).
Resolution. As explained in the Introduction, in his
later theorizing, Bowlby (1980) renamed the final phase
of grief “reorganization” to convey that representations
of the self and the lost attachment figure are reorganized
in ways that may allow a continuing (but altered) emo-
tional bond in conjunction with adjustment to changed
circumstances. We expected persons high in anxiety to
be most apt to suffer disorganization of their own identi-
ties in the absence of the lost partner. Furthermore, we
expected both anxious and avoidant persons to be less
likely to integrate the ex-partner into their lives in an
altered form of attachment, such as friendship or work-
ing relationships.
The first expectation was clearly supported. Relation-
ship anxiety was strongly associated with reports of a lost
sense of identity without the former partner. The second
prediction received only weak support. That is, there
were small negative associations between both anxiety
882 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
and avoidance and integration of the lost partner into a
different role relationship.
Chronic mourning. Bowlby (1980) mentioned attachment-
related differences in chronic mourning, or
perseveration in the protest and/or despair stages of
mourning. Similar to Freud (1917/1957) and others
(e.g., Lazare, 1989; Sanders, 1993), Bowlby (1980)
argued that chronic mourning, or “complicated
bereavement,” derives from anxious/ambivalence
toward the lost partner. Thus, we expected attachment
anxiety to be associated with prolonged, exaggerated,
and dysfunctional protest and despair reactions (i.e.,
chronic mourning). And in fact, those who were higher
in anxiety and those who were more attached to the lost
partner (i.e., were higher in emotional involvement or
who did not initiate the breakup) reported greater
perseveration in desire for the lost partner.
Replacement. Consistent with their insecurity, however,
those higher in attachment anxiety also reported a
greater tendency to search immediately for a replace-
ment for the lost partner. Generally, those high in attach-
ment anxiety tend to feel uncomfortable when not in a
romantic relationship and report higher motivation to
be in a romantic relationship (e.g., Davis, 2000). It seems
likely that relationships formed under these desperate
conditions would be unusually troubled later on, which
may be a reason for the high breakup rate of relation-
ships formed by people high in attachment-related anxi-
ety (Shaver & Clark, 1994).
LIMITATIONS
Our Internet methodology allowed us to achieve a
greater number of respondents and greater sample
diversity than that of the more usual studies of college
students. The sample is not representative of any particu-
lar population, however; it was biased toward young
computer users and people interested in relationship
issues. Moreover, participants described memories of
past relationship breakups. However, although autobio-
graphical memory can be unreliable (e.g., Conway &
Pleydell-Pearce, 2000), and specifically memory for reac-
tions to loss can become distorted over time (e.g., Safer,
Bonanno, & Field, 2001), the present observed effects
remained significant and unaltered for breakups occur-
ring less than 3 months before the survey. Another con-
cern—that unmonitored participants may have been
frivolous while completing our questionnaire—can be
assuaged by noting that the coefficient alphas were high
for the multi-item measures and the results formed a the-
oretically predictable and sensible pattern.
Finally, it would clearly be desirable to examine simi-
lar measures prospectively to establish clear causal
effects of attachment style on reactions to breakups and
to rule out the reverse possibility that breakup processes
affected attachment style. However, although the possi-
bility that the nature of the breakup caused differences
in attachment style cannot currently be ruled out, there
is no foundation for predicting the full range of reverse
effects. For example, it is not clear how coping through
self-medication versus socially, becoming vengeful or
violent, or feeling a lost sense of identity would cause
changes in attachment style. Thus, the causal role of
attachment style in determining reactions to breaking
up remains highly plausible. Moreover, two previous
studies of attachment style and reactions to breaking up
were prospective in nature (Feeney & Noller, 1992;
Simpson, 1990). Similar to our study, they both found
predictable associations between anxiety, avoidance,
and distress. The second specifically found that breaking
up did not affect attachment style. Furthermore, at least
one study of stability in attachment style (Scharfe &
Bartholomew, 1994) showed that while positive interper-
sonal life events predicted change in attachment style,
negative interpersonal life events did not. Nevertheless,
our hypotheses certainly deserve to be tested longitudi-
nally, and the present study provides a rich guide for
such future research.
REFERENCES
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Pat-
terns of attachment. Hillsdale, NJ: Lawrence Erlbaum.
Archer, J. (1999). The nature of grief: The evolution and psychology of reac-
tions to loss. London: Routledge.
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among
young adults: A test of a four-category model. Journal of Personality
and Social Psychology,61, 226-244.
Bixenstine, V. E. (1999). Spousal homicide. In H. V. Hall (Ed.), Lethal
violence (pp. 231-257). New York: CRC Press.
Bonnano, G. A., Notarius, C. I., Gunzerath, L., Keltner, D., &
Horowitz, M. J. (1995). Interpersonal ambivalence, perceived
dyadic adjustment, and conjugal loss. Journal of Consulting and
Clinical Psychology,66, 1012-1022.
Bowlby, J. (1973). Attachment and loss: Separation, anxiety, and anger.
New York: Basic Books.
Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss: Sadness and depres-
sion. New York: Basic Books.
Bowlby, J. (1982). Attachment and loss: Attachment (rev. ed.). New York:
Basic Books. (Original work published 1969)
Bowlby, J. (1988). A secure base. New York: Basic Books.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report mea-
surement of adult attachment: An integrative overview. In J. A.
Simpson & W. S. Rholes (Eds.), Attachment theory and close relation-
ships (pp. 46-76). New York: Guilford.
Brennan, K. A., & Shaver, P. R. (1995). Dimensions of adult attach-
ment, affect regulation, and romantic relationship functioning.
Personality and Social Psychology Bulletin,21, 267-283.
Cassidy, J., & Berlin, L. (1994). The insecure/ambivalent pattern of
attachment: Theory and research. Child Development,65, 971-991.
Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of
autobiographical memories in the self-memory system. Psychologi-
cal Review,107, 261-288.
Cooper, M. L., Shaver, P. R., & Collins, N. L. (1998). Attachment
styles, emotion regulation, and adjustment in adolescence. Jour-
nal of Personality and Social Psychology,74, 1380-1397.
Davis et al. / ATTACHMENT AND REACTIONS TO BREAKUPS 883
Crittenden, P. M. (1992). Children’s strategies for coping with
adverse home environments: An interpretation using attachment
theory. Child Abuse & Neglect,16, 329-343.
Crittenden, P. M. (1997). Patterns of attachment and sexual behavior:
Risk of dysfunction versus opportunity for creative integration. In
L. Atkinson & K. J. Zucker (Eds.), Attachment and psychopathology
(pp. 47-96). New York: Guilford.
Davis, D. (2000, May). May to December: Relationship motivation and
behavior throughout the lifespan. Invited address, Psychology Depart-
ment, University of California, Davis.
Downey,G., Feldman, S., & Ayduk, O. (2000) Rejection sensitivity and
male violence in romantic relationships. Personal Relationships,7,
45-61.
Dutton, D. G. (1998). The abusive personality. New York: Guilford.
Feeney, J. A. (1999). Adult romantic attachment and couple relation-
ships. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment:
Theory, research, and clinical applications (pp. 355-377). New York:
Guilford.
Feeney, J. A., & Noller, P. (1992). Attachment style and romantic love:
Relationship dissolution. Australian Journal of Psychology,44, 69-74.
Fraley, R. C., Davis, K. E., & Shaver, P. R. (1998). Dismissing-avoidance
and the defensive organization of emotion, cognition, and behav-
ior. In W. S. Rholes & J. A. Simpson (Eds.), Attachment theory and
close relationships (pp. 249-279). New York: Guilford.
Fraley, R. C., & Shaver, P. R. (1997). Adult attachment and the sup-
pression of unwanted thoughts. Journal of Personality and Social Psy-
chology,73, 1080-1091.
Fraley, R. C., & Shaver, P. R. (1999). Loss and bereavement: Attach-
ment theory and recent controversies concerning “grief work”
and the nature of detachment. In J. Cassidy & P. R. Shaver (Eds.),
Handbook of attachment: Theory, research, and clinical applications
(pp. 735-759). New York: Guilford.
Fraley, R. C., & Shaver, P. R. (2000). Adult romantic attachment: The-
oretical developments, emerging controversies, and unanswered
questions. Review of General Psychology,4, 132-154.
Freud, S. (1957). Mourning and melancholia. In J. Strachey (Ed. and
Trans.), The standard edition of the complete psychological works of
Sigmund Freud (Vol. 14, pp. 237-260). New York: Basic Books.
(Original work published 1917).
George, C., & Solomon, J. (1999). Attachment and caregiving: The
caregiving behavioral system. In J. Cassidy & P. R. Shaver (Eds.),
Handbook of attachment: Theory, research, and clinical applications
(pp. 649-670). New York: Guilford.
Gottman, J. M. (1994). What predicts divorce? The relationship between
marital processes and marital outcomes. Hillsdale, NJ: Lawrence
Erlbaum.
Hazan, C., & Shaver,P. R. (1987). Romantic love conceptualized as an
attachment process. Journal of Personality and Social Psychology,59,
511-524.
Hazan, C., & Shaver, P. R. (1990). Love and work: An attachment-
theoretical perspective. Journal of Personality and Social Psychology,
59, 270-280.
Kiecolt-Glaser, J. K., & Newton, T. L. (2001). Marriage and health: His
and hers. Psychological Bulletin,127, 472-503.
Klas, D., Silverman, P. R., & Nickman, S. L. (Eds.). (1996). Continuing
bonds: New understandings of grief. Washington, DC: Taylor &
Francis.
Lazare, A. (1989). Bereavement and unresolved grief. In A. Lazare
(Ed.), Outpatient psychiatry: Diagnosis and treatment (2nd ed., pp.
381-397). Baltimore: Williams & Wilkins.
Main, M., & Solomon, J. (1986). Discovery of a new, insecure disorga-
nized/disoriented attachment pattern. In T. B. Brazelton & M.
Yogman (Eds.), Affective development in infancy (pp. 95-124).
Norwood, NJ: Ablex.
Mikulincer, M., & Florian, V. (2001). Attachment style and affect reg-
ulation: Implications for coping with stress and mental health. In
G. Fletcher & M. Clark (Eds.), Blackwell handbook of social psychol-
ogy: Interpersonal processes (pp. 537-557). Oxford, UK: Blackwell.
Mikulincer, M., Gillath, O., & Shaver, P. R. (2002). Activation of the
attachment system in adulthood: Threat-related primes increase
the accessibility of mental representations of attachment figures.
Journal of Personality and Social Psychology,83, 881-895.
Neimeyer, R. A. (Ed.). (2001). Meaning reconstruction and the experience
of loss. Washington, DC: American Psychological Association.
Safer, M. A., Bonanno, G. A., & Field, N. P. (2001). “It was never that
bad”: Biased recall of grief and long-term adjustment to the death
of a spouse. Memory,9, 195-204.
Sanders, C. M. (1993). Risk factors in bereavement outcome. In M. S.
Stroebe, W. Stroebe, & R. O. Hansson (Eds.), Handbook of bereave-
ment: Theory, research, and intervention (pp. 255-270). Cambridge,
UK: Cambridge University Press.
Scharfe, E., & Bartholomew, K. (1994). Reliability and stability of
adult attachment patterns. Personal Relationships,1, 23-43.
Schuchter, S. R., & Zisook, S. (1993). The course of normal grief. In
M. S. Stroebe, W. Stroebe, & R. O. Hansson (Eds.), Handbook of
bereavement: Theory, research, and intervention (pp. 23-43). Cam-
bridge, UK: Cambridge University Press.
Shaver, P. R., & Clark, C. L. (1994). The psychodynamics of adult
romantic attachment. In J. M. Masling & R. F. Bornstein (Eds.),
Empirical perspectives on object relations theories (pp. 105-156). Wash-
ington, DC: American Psychological Association.
Simpson, J. A. (1990). The influence of attachment styles on romantic
relationships. Journal of Personality and Social Psychology,59, 971-
980.
Simpson, J. A., Rholes, W. S., & Nelligan, J. S. (1992). Support seeking
and support giving within couples in an anxiety-provoking situa-
tion: The role of attachment styles. Journal of Personality and Social
Psychology,62, 434-446.
Sprecher, S., Felmlee, D., Metts, S., Fehr, B., & Vanni, D. (1998). Fac-
tors associated with distress following the breakup of a close rela-
tionship. Journal of Social and Personal Relationships,15, 791-809.
Walker, L. E. (1979). The battered woman. New York: Harper & Row.
Weinfield, N. S., Sroufe, L. A., Egeland, B., & Carlson, E. A. (1999).
The nature of individual differences in infant-caregiver attach-
ment. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment:
Theory, research, and clinical applications (pp. 69-88). New York:
Guilford.
Wilson, M., & Daly, M. (1993). Spousal homicide risk and estrange-
ment. Violence and Victims,8, 3-16.
Received July 9, 2002
Revision accepted July 29, 2002
884 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Request Permissions / Order Reprints
Request Permission or Order Reprints Instantly
Interested in copying, sharing, or the repurposing of this article? U.S. copyright law, in
most cases, directs you to first get permission from the article’s rightsholder before using
their content.
To lawfully obtain permission to reuse, or to order reprints of this article quickly and
efficiently, click on the “Request Permission/ Order Reprints” link below and follow the
instructions. For information on Fair Use limitations of U.S. copyright law, please visit
Stamford University Libraries, or for guidelines on Fair Use in the Classroom, please
refer to The Association of American Publishers’ (AAP).
All information and materials related to SAGE Publications are protected by the
copyright laws of the United States and other countries. SAGE Publications and the
SAGE logo are registered trademarks of SAGE Publications. Copyright © 2003, Sage
Publications, all rights reserved. Mention of other publishers, titles or services may be
registered trademarks of their respective companies. Please refer to our user help pages
for more details: http://www.sagepub.com/cc/faq/SageFAQ.htm
... On the one hand, some studies have found that initiator status was not associated with sadness or anger (Sbarra, 2006), distress (Eastwick et al., 2008;Tashiro & Frazier, 2003), emotional recovery (Sbarra & Emery, 2005), or growth (Tashiro & Frazier, 2003). However, other studies have demonstrated differences with recipients experiencing more distress than initiators (Boelen & Reijntjes, 2009;Davis et al., 2003;Field et al., 2009;Koessler et al., 2019a;Perilloux & Buss, 2008), especially when the breakup was unexpected (Boelen & Reijntjes, 2009;Field et al., 2009). Beyond distress, recipients experience more anger, confusion, and jealousy than initiators, who feel more happiness and guilt (Davis et al., 2003;Perilloux & Buss, 2008). ...
... However, other studies have demonstrated differences with recipients experiencing more distress than initiators (Boelen & Reijntjes, 2009;Davis et al., 2003;Field et al., 2009;Koessler et al., 2019a;Perilloux & Buss, 2008), especially when the breakup was unexpected (Boelen & Reijntjes, 2009;Field et al., 2009). Beyond distress, recipients experience more anger, confusion, and jealousy than initiators, who feel more happiness and guilt (Davis et al., 2003;Perilloux & Buss, 2008). Yet, Kansky and Allen (2018) found that initiators' peers reported that the initiators experienced more internalizing symptoms (e.g., anxiety and depression) than recipients. ...
... Not finding a difference in overall positive or negative valence of the accounts between ghosters and ghostees contrasts with prior research showing breakup recipients experience more distress than initiators (Boelen & Reijntjes, 2009;Davis et al., 2003;Field et al., 2009;Koessler et al., 2019a;Perilloux & Buss, 2008), but is consistent with other prior research showing no differences between initiators and recipients in breakups (Sbarra, 2006;Tashiro & Frazier, 2003). It also aligns with ostracism research suggesting that both parties experience emotional distress Zadro et al., 2016). ...
Article
Although ghosting (i.e., unilaterally ending a relationship by ceasing communication) has only recently entered the lexicon, it is a regularly used form of relationship dissolution. However, little research has examined the emotional experiences of ghosting, particularly the experiences of those on both sides of the ghosting process. In a multi-method study, participants who had both ghosted and been ghosted in previous romantic relationships (N = 80) provided narratives of their experiences and completed questionnaires. The narrative responses were analyzed by coders and by using LIWC. Ghosters and ghostees used similar overall levels of positively and negatively valenced words to describe their experiences, but ghosters were more likely to express guilt and relief, whereas ghostees were more likely to express sadness and hurt feelings. Ghostees also experienced more of a threat to their fundamental needs - control, self-esteem, belongingness, meaningful existence - than ghosters.
... Preliminary evidence indicates the role of each partner in the initiation of the divorce (i.e., "leaver" vs. "left") has relevant implications for their postdivorce adjustment. Specifically, findings suggest that those who initiate divorce have higher emotional and psychological wellbeing after divorce compared with noninitiators, particularly during the first few years postseparation (Baum, 2003;Davis et al., 2003;Hewitt & Turrell, 2011;Quinney, 2003;Steiner et al., 2015). Despite such findings, much of the available research on divorce initiation is based on a single-item, categorical assessment of initiator status (e.g., who initiated the divorce?). ...
... This understanding of divorce initiation as a complex, multifaceted phenomenon is critical for researchers and clinicians working in this area. Research evidence suggests that the role of each partner in the initiation of the divorce (i.e., "leaver" vs. "left") has implications for postdivorce functioning, in that the leaver often experiences better immediate and short-term outcomes (Baum, 2003;Davis et al., 2003;Hewitt & Turrell, 2011;Quinney, 2003;Steiner et al., 2015). It stands to reason that the processes undertaken by both the leaver and the left would have implications for mental health and well-being postdivorce. ...
Article
The goal of the present study was to determine the psychometric properties of an extended inventory that assesses the multiple facets of divorce initiation drawn from the existing research, the Revised Divorce Initiation Inventory (DII‐R). Previous research indicates the role of each partner in the initiation of divorce has relevant implications for emotional and psychological well‐being after the divorce. Much of this research was conducted with a single‐item, categorical assessment of initiator status (e.g., who initiated the divorce?). However, recent studies highlight the complexity of divorce initiation as a multidimensional concept. A confirmatory factor analysis (CFA) was conducted with a sample of 408 divorced individuals. Analyses to explore measurement invariance between men and women and construct validity were also conducted. Findings indicated a four‐factor model (i.e., Legal, Separation, Desire, Responsibility) of 12 items was a good fit to the data. Factor loadings and intercepts were equal across groups. Divorce initiation is a multifaceted construct that includes dimensions related to separation, legal, desire, and responsibility that should not be measured with a single‐item assessment. The DII‐R supports the comprehensive synthesis of ongoing divorce initiation research that may be used to inform clinical practice to intervene in specific areas shown to affect the decision to divorce.
... Specifically, we expect them to ruminate more frequently about the causes and consequences of the breakup (e.g., "What could I have done to prevent the breakup?") (e.g., Eisma & Stroebe, 2017;Fagundes, 2012;Saffrey & Ehrenberg, 2007) and to yearn more strongly for reunion with the ex-partner (e.g., "I imagine how wonderful we would be together right now") (e.g., Cope & Mattingly, 2021;Eisma et al., 2020). Second, based on the theoretical and empirical link between anxious and desired attachment and self-reported attempts to reunite with the expartner (Davis et al., 2003;Logan et al., 2019), we expect these attachment constructs to relate positively to automatic approach of the ex-partner assessed with an AAT. In line with prior theorizing and research, we further predict that all aforementioned variables relate positively to breakup distress. ...
... We also found that automatic approach tendencies toward the ex-partner, as well as yearning, related positively to breakup distress. This complements work suggesting positive relationships between insecure attachment and searching for the ex-partner, attempts to reunite with the ex-partner, and breakup distress (e.g., Weiss, 1988;Davis et al., 2003;Cope & Mattingly, 2021). ...
Article
Full-text available
Background and objectives Romantic relationship breakups can lead to severe emotional disturbances including major depression. Anxious attachment and desired attachment with the ex-partner are hypothesized to elicit repetitive thought about the breakup and the former partner and attempts to reunite with (i.e. approach) the ex-partner, which fuel breakup distress. Since prior research on this topic has mostly used survey methodology, the study aim was to examine the relations between above-mentioned variables employing a behavioral measure of approach of the ex-partner. Methods Automatic approach-avoidance tendencies toward the former partner were assessed with an Approach Avoidance Task (AAT). Sixty-two students (76% female) moved a manikin towards or away from stimuli pictures (ex-partner, matched stranger, landscape) as fast as possible based on the stimulus frame color (blue, yellow). Participants also completed questionnaires assessing anxious attachment, desired attachment, repetitive thought about the breakup (rumination) and the ex-partner (yearning), and breakup distress (prolonged grief symptoms). Results Anxious attachment related positively to rumination and breakup distress. Desired attachment related positively to yearning, automatic approach bias toward the ex-partner, and breakup distress. Both anxious and desired attachment, rumination, yearning, and approach bias related positively to breakup distress. Limitations The use of a student sample may limit generalizability. A correlational design precludes causal conclusions. Conclusions Together with prior work, results suggests anxious attachment hampers psychological adaptation to a breakup by increasing the use of ruminative coping. Desire to retain an attachment bond with the ex-partner, expressed in yearning and approach of the ex-partner, may also worsen breakup distress.
... In contrast, the consequences for initiators' health are more ambiguous: Depending on the outcome variable, their physical health either improves or remains unchanged after divorce whereas their mental health worsens or remains unchanged. While some additional studies analyze initiator status and postdivorce outcomes, these studies cannot inform us on any of the theoretical predictions made above, as they focus exclusively on a comparison of initiators and noninitiators in the period after the divorce, often at only one point in time (5)(6)(7)(8)(9)(10)(11)(12). Results from these studies either indicate that initiators do better than noninitiators after separation or indicate that there is no difference. ...
Article
Significance In economic and sociological theory of divorce, the link between divorce consequences and the decision to divorce is central: A couple divorces if at least one spouse expects to improve their life by initiating divorce. The present study provides empirical evidence in support of this theoretical link: Separation initiators become better off in terms of subjective well-being after a separation, whereas noninitiators become worse off, before they eventually experience a full recovery. Because separations are predominantly initiated by only one partner, this finding suggests that one partner typically benefits from the separation (the initiator), while the other is disadvantaged (the noninitiator). Accordingly, analyses of average divorce trajectories convey only limited information about the causal effects of divorce on individuals’ well-being.
... The skeptic may want to downplay the impact of romantic breakups by suggesting there is 'only' a change in relationship status. However, research shows (and some people may know from personal experience) that the end of a relationship may have severe physical, emotional and behavioral consequences (Davis, Shaver, and Vernon 2003). Slotter, Gardner, and Finkel (2010, 148) emphasize this point as well: 'individuals may [post-breakup] alter their appearance, social circles, activities, goals, or even their values or beliefs'. ...
... The findings display that physical health, race, and financial status were significant predictors of relational satisfaction. The results align with past research that states that positive experiences lead to higher relationship satisfaction while negative events lead to less relationship satisfaction (Davis et al., 2003;Frazier & Cook, 1993;Sbarra & Emery, 2005;Perlman, 2007;Reis et al., 2000;Roberts & Robins, 2000;Tan et al., 2018). In addition, given that relationship satisfaction has been found to vary by gender, this study highlights female partner specific sources of relationship satisfaction (Heiman et al., 2011). ...
Article
Relationship satisfaction is a subjective and global evaluation of a romantic relationship based on a sense of happiness, contentment, and fulfillment felt by a partner. Prior research demonstrates that relationship satisfaction can help mediate difficult life situations and stress and may contribute to a partner’s well-being and health. However, the literature examining partner-specific sources of relational satisfaction are lacking. Through a medical family therapy lens, this study examined women’s reported physical health’s impact on relational satisfaction. A total of 555 women in romantic relationships completed a survey with questions from the Medical Outcomes Study 20-Item Short-Form Health Survey and the Relationship Assessment Scale. A hierarchical linear regression indicated that physical health, race, and financial status were significant predictors of relational satisfaction. Findings highlight the importance of physical health’s impacts on relational satisfaction and provide clinical implications and future directions on strengthening couples’ relational satisfaction.
... State Worry Questionnaire (PSWQ; Meyer et al., 1990), Worry scale of Dimensional Clinical Personality Inventory (DCPI-2 Worry scale; Carvalho & Primi, 2019); (v) interpersonal rumination -Rumination on Interpersonal Offenses Scale (RIO; Wade et al., 2008), Relationship Preoccupation Scale (RPS;Davis et al., 2003). ...
Article
Objective The current quantitative review aims at comprehensively clarifying the role of rumination in borderline personality disorder (BPD) considering its relevance for several clinical models of the disorder. Method This meta-analysis included 29 independent studies assessing different forms of rumination—general tendency to engage in ruminative thinking patterns and four types of emotion-based rumination among both nonclinical subjects reporting BPD features and patients with BPD. Furthermore, the study tested whether rumination could be considered a widespread emotion-based cognitive vulnerability in BPD. Results Meta-analytic procedures were based on 46 r coefficients that showed large pooled effect sizes for all forms of rumination. With the exception of interpersonal rumination, the remaining forms of rumination were equally involved in BPD features. Conclusions Rumination should be considered a widespread emotion-based cognitive vulnerability in BPD. Future studies should provide longitudinal and contextual-based evaluations of rumination among treatment-seeking individuals with BPD.
... This cognitive framework of representation can determine whether a person responds to separation distress insecurely or in a mature, secure fashion. Drawing on research involving human attachment figures (Bartholomew & Shaver, 1998;Davis et al., 2003;Finzi et al., 2001), it is possible that insecurely attached individuals with an anxious working model will have a greater proclivity to respond to place attachment disruptions with protest, even when the threat to their attachment bond is relatively benign. Similarly, people characterized by an insecure-avoidant attachment style may be more inclined to avoid a place of attachment as a way of dealing with the threat of place attachment disruption. ...
Chapter
In Chapter 1, we introduced place as a relational object of attachment. This conceptualization of place allows us to consider what happens when a global pandemic threatens the bonds that people have with places of significance. It also provides a framework for understanding how separation from a place of significance can lead to a reparative process in which people either rebuild their sense of connection to the place of attachment that has been disrupted or replace that attachment bond with an alternative attachment object. Using the framework offered by Counted et al. (2021), this chapter will explore the three theorized response phases—protest, despair, and detachment—that follow place attachment disruption.
Article
The current research examined the phenomenon of fading affect bias – the tendency for affect associated with negative events to fade more than affect associated with positive events – within the context of romantic relationships. Participants recalled and evaluated positive and negative relationship-specific and non-relationship autobiographical events. Participants also completed measures of attachment avoidance and anxiety. Multi-level modeling demonstrated fading affect bias for relationship and non-relationship events, but that affect fade was shaped by attachment orientations. Specifically, higher attachment anxiety, and lower attachment avoidance predicted greater importance of relationship events which predicted lower fading of affective intensity of memories. Thus, attachment anxiety sustained, while attachment avoidance suppressed the affect of relational memories. We discuss implications of these findings for relationship maintenance.
Article
Attachment theory provides a useful framework for understanding emotional reactions to separation and loss and the process of adapting to these painful events. In this paper, we review adult attachment studies that have examined emotional reactions and adjustment to separation and loss in romantic and marital relationships. We begin with a brief account of attachment theory. Next, we review studies examining the emotional consequences of losing a relationship partner and the coping responses that can help a person adjust to this loss. Throughout the paper, we also summarize research documenting attachment-related individual differences in responses to separation and loss. (99 words)
Article
Three studies explored the effects of subliminal threat on the activation of representations of attachment figures. This accessibility was measured in a lexical decision task and a Stroop task following threat-or neutral-word primes, and was compared with the accessibility of representations of other close persons, known but not close persons, and unknown persons. Participants also reported on their attachment style. Threat primes led to increased accessibility of representations of attachment figures. This effect was specific to attachment figures and was replicated across tasks and experiments. Attachment anxiety heightened accessibility of representations of attachment figures even in neutral contexts, whereas attachment avoidance inhibited this activation when the threat prime was the word separation. These effects were not explained by trait anxiety. The discussion focuses on the dynamics of attachment-system activation in adulthood.
Article
This article explores the possibility that romantic love is an attachment process--a biosocial process by which affectional bonds are formed between adult lovers, just as affectional bonds are formed earlier in life between human infants and their parents. Key components of attachment theory, developed by Bowlby, Ainsworth, and others to explain the development of affectional bonds in infancy, were translated into terms appropriate to adult romantic love. The translation centered on the three major styles of attachment in infancy--secure, avoidant, and anxious/ambivalent--and on the notion that continuity of relationship style is due in part to mental models (Bowlby's "inner working models") of self and social life. These models, and hence a person's attachment style, are seen as determined in part by childhood relationships with parents. Two questionnaire studies indicated that relative prevalence of the three attachment styles is roughly the same in adulthood as in infancy, the three kinds of adults differ predictably in the way they experience romantic love, and attachment style is related in theoretically meaningful ways to mental models of self and social relationships and to relationship experiences with parents. Implications for theories of romantic love are discussed, as are measurement problems and other issues related to future tests of the attachment perspective.
Article
This study examined how adult attachment styles moderate spontaneous behavior between dating couples when 1 member of the dyad is confronted with an anxiety-provoking situation. Eighty-three dating couples were unobtrusively videotaped for 5 min in a waiting room while the woman waited to participate in an "activity" known to provoke anxiety in most people. Independent observers then evaluated each partner's behavior on several dimensions. Results revealed that persons with more secure attachment styles behaved differently than persons with more avoidant styles in terms of physical contact, supportive comments, and efforts to seek and give emotional support. Findings are discussed in the context of theory and research on attachment.