Content uploaded by Judy Makinen
Author content
All content in this area was uploaded by Judy Makinen
Content may be subject to copyright.
Journal of Couple & Relationship Therapy, 9:31–47, 2010
Copyright © Taylor & Francis Group, LLC
ISSN: 1533-2691 print / 1533-2683 online
DOI: 10.1080/15332690903473069
Resolving Attachment Injuries in Couples
Using Emotionally Focused Therapy:
A Three-Year Follow-Up
REBECCA E. HALCHUK
University of Ottawa, Ottawa, Ontario, Canada
JUDY A. MAKINEN
Royal Ottawa Mental Health Centre, Ottawa, Ontario, Canada,
and University of Ottawa, Ottawa, Ontario, Canada
SUSAN M. JOHNSON
University of Ottawa, Ottawa, Ontario, Canada; Alliant University, San Diego,
California, USA; and International Center for Excellence in Emotionally Focused
Therapy, Ottawa, Ontario, Canada
Couples who seek therapy for marital distress often do so because
they have suffered an attachment injury, characterized by an
abandonment or betrayal during a time of critical need. This
follow-up assessed the efficacy of the newly developed Attachment
Injury Resolution Model based in Emotionally Focused Therapy
(EFT). Twelve couples with attachment injuries who received EFT
were assessed to determine if the significant improvement in re-
lationship distress observed in resolved couples at post-treatment
would be maintained at 3-year follow-up. Results demonstrated
that improvements in dyadic adjustment, trust, and forgiveness,
as well as decreases in the severity of the attachment injury, were
maintained over time. This follow-up study provides initial evidence
of the long-term benefits of the Attachment Injury Resolution Model.
KEYWORDS couples, attachment, attachment injury, emotionally
focused therapy
Address correspondence to Judy A. Makinen, Ottawa Couple and Family Institute, 1869
Carling Avenue, Suite 201, Ottawa, Ontario, K2A 1E6, Canada. E-mail: judy.makinen@rohcg.
on.ca
31
32 R. E. Halchuk et al.
Couples seeking therapy have typically endured significant relationship dis-
tress. In some cases, this distress is precipitated by an attachment-related
incident in which one partner failed to support or betrayed the other in a
critical time of need. Recently this concept has been conceptualized into what
is known as an attachment injury (Johnson, Makinen, & Millikin, 2001). This
event may then become the standard for the offending partner’s dependabil-
ity, instead of the exception, and manifest as a clinically recurring theme that
blocks relationship repair. Indeed, such distress is often propagated, main-
tained, and exacerbated because partners are entrapped in a vicious cycle
of negative emotions and ways of interacting (i.e., blame-withdraw), which
undermines their desire to heal their relationship and reach out to each other
in more positive ways.
The attachment injury concept has roots in attachment theory, and also
arose from the theory and application of emotionally focused therapy (EFT).
EFT is one of today’s leading short-term interventions in couples’ therapy
(Baucom, Shoham, Mueser, Daiuto, & Stickle, 1998). EFT centers on chang-
ing attachment behaviors as a means to improve distressed relationships
(Johnson, 2004). It views relationship distress as resulting from a breakdown
in a couple’s ability to communicate their emotions and cope with feel-
ings of insecurity, which leads to negative interactional cycles. Its goals are
therefore to help the couple access underlying emotions and foster positive
interactions that promote accessibility and trust between partners. EFT has
been successfully applied to a broad range of populations, from those deal-
ing with posttraumatic stress disorder (Johnson & Williams-Keeler, 1998),
to couples dealing with chronic illness (Gordon-Walker, Johnson, Manion,
& Cloutier, 1996) and depression (Dessaulles, Johnson, & Denton, 2003).
Moreover, EFT has demonstrated relatively high treatment effects (Johnson,
Hunsley, Greenberg, & Schindler, 1999), in addition to stable recovery rates
(Gordon-Walker, Manion, & Cloutier, 1998).
With respect to adult romantic relationships, adult attachment theory
is based on the assumption that romantic partners will develop an affec-
tional bond and serve as the main attachment figure for the other (Hazan
& Shaver, 1987). The ensuing attachment patterns (i.e., secure, insecure)
are based on the quality of their affectional bond and are reflected in each
partner’s psychological well-being. Securely attached couples exhibit more
positive functioning, characterized by high levels of trust, commitment, and
satisfaction (Kobak & Hazan, 1991). Insecure attachment, resulting from a
partner’s lack of responsiveness or inaccessibility, can lead to relational dis-
tress. Further research has characterized distress as involving reciprocal neg-
ativity, ineffective communication, and holding negative relationship schema
(Halford, Kelly, & Markman, 1997). Such interactions cause impasses in re-
solving issues, as they tend to cycle, and ultimately inhibit positive interac-
tion. Rigid negative attributions then dominate how each partner tends to
view the other’s behavior and the relationship in general (Johnson, Makinen,
& Millikin, 2001).
Resolving Attachment Injuries 33
In couples suffering from an attachment injury, this violation or breach
of trust by their attachment figure indicates to the injured partner that the
other can no longer be counted on for caring and support when needed
(Johnson, Makinen., & Millikin, 2001). What is more, such incidents are
observed to have disproportionately damaging effects on close relationships,
which often then seem irreparable (Simpson & Rholes, 1994). Injuries may be
compounded if the injured partner approaches the offending partner about
the event and is met with responses such as dismissal or denial. With time,
a failure to remedy fractured bonds results in increased feelings of despair
and alienation (Johnson, 2004).
What constitutes an attachment injury is highly variable for every couple,
and may not necessarily mark the beginning of distress; rather a relationship
may be distressed long before an explicit event occurs (Makinen & Johnson,
2006). Particularly vulnerable periods include times of loss, physical illness,
and life transitions. What is important is how the injured partner views the
event, with respect to their expectations, perceptions, and beliefs (Jones &
Burdette, 1994). Correspondingly, therapy is not concerned with the actual
event itself, but instead focuses on the emotional effects that result from an
attachment injury.
Main and Hesse (1990) pointed out that relationships where the pri-
mary attachment figure serves as both the source and solution to pain are
especially difficult to endure and result in disorganization of the attachment
bond. The injured partner is caught in a cycle of hyper- and hypo-arousal
whereby they vacillate between seeking and withdrawing from the other
partner, who finds this type of interaction chaotic and aversive. To make
matters worse, the injured partner cannot accept or trust any comfort offered
by their partner (Schore, 1994). This makes it difficult to overcome the sense
of alienation, as open confiding and coping are impossible (Pennebaker,
1985). The psychological effects on couples experiencing such trauma in-
clude intensified negative affect and hypervigilance toward further acts of
betrayal (Johnson & Williams-Keeler, 1998). Couples that are successful in
recovering from betrayal refer to forgiveness as a mediating factor (Gordon,
Baucom, & Snyder, 2000).
The conceptualization of an attachment injury arose when clinicians
noted that certain couples did not repair their distress following EFT and
found that clear patterns emerged. Specifically, some couples would refer
back to specific incidents of betrayal or abandonment, during which they
would use the language of trauma (i.e. life and death terms) to describe
it. If the therapist could not succeed in helping the couple address and
deal with the breach in trust, then couples were not able to achieve new
positive ways of interacting and, hence, continued to suffer distress (Johnson
& Greenberg, 1988). Thus, this concept is crucial in understanding impasses
so that interventions may be tailored to confer long-lasting changes.
Recently, the process of resolving attachment injuries has been out-
lined and tested. Milliken (2000) collected exploratory data on three couples
34 R. E. Halchuk et al.
that successfully resolved their attachment injuries through 15 EFT sessions.
From this, the rational-empirical Attachment Injury Resolution Model was
developed. This model assumes that an attachment injury is a workable
concept that may be implemented at the beginning of EFT and continue
through to resolution, through four conceptual phases: identifying the at-
tachment injury marker, differentiating affect, re-engagement between part-
ners, and forgiveness and reconciliation. The purpose of the first phase is to
help couples understand the origin of the attachment injury and the negative
interactional cycle that has developed. Following this, the injured partner is
encouraged to share the full impact of the injury on their attachment bond
with the offending partner, who in turn is helped to hear and understand
its significance. Subsequently, partners begin to increase their emotional in-
volvement. In particular, the injured partner further explores and processes
the emotional pain of the loss of the attachment bond, and the offending
partner observes this vulnerability, acknowledges their role in its disruption,
and expresses their empathy, remorse, and regret. Finally, the injured part-
ner takes a risk and reaches to the other for comfort and care, who in turn
responds effectively, which ultimately repairs the trauma and restores the
attachment bond (Milliken, 2000).
In a subsequent study, Makinen and Johnson (2006) investigated the at-
tachment injury resolution model using a sample of 24 couples. The purpose
of this study was two-fold, to offer validation of the model using process
measures to compare resolved and nonresolved groups, and to relate the
process of change to distal outcomes. With respect to this second objec-
tive, it was found at post-treatment that resolved couples reported higher
levels of dyadic adjustment and forgiveness than nonresolved couples. The
two groups were not distinguished on their levels of avoidant and anxious
attachment or pain, although changes in these measures were found at post-
treatment. It was noted that there was a limited amount of therapy sessions
that could be offered to participants; the average length of therapy was 13
sessions. Also, most of the nonresolved couples had reported compound
attachment injuries, whereas the resolved couples reported only single
injuries.
At present, only one study has demonstrated the longer-term bene-
fits of EFT. This 2-year follow-up study by Cloutier, Manion, and Walker
(2002) demonstrated that improvements in marital functioning in couples
with chronically ill children showed not only maintenance, but also contin-
ued improvement. Therefore, there is a need for further evidence outlining
the enduring effects of EFT.
The current study was designed as a follow-up study to previous re-
search examining the verification phase of the attachment injury resolution
model. Specifically, it serves to determine if the significant improvement in
marital distress of resolved couples at post-treatment would be observable
3 years after EFT. With respect to the attachment injury resolution model
and its long-term outcome effects, it was predicted that at 3-year follow-up
Resolving Attachment Injuries 35
resolved couples would show stability in decreased marital distress, increased
relationship trust, decreased avoidant and anxious attachment, increased for-
giveness, decreased emotional pain, and decreased attachment injury sever-
ity, as exhibited at post-treatment.
METHOD
Participants
Participants included a volunteer sample of couples with an attachment in-
jury who had previously participated in an outcome study that assessed the
validity and effectiveness of the attachment injury resolution model in allevi-
ating relationship distress (Makinen & Johnson, 2006). To maintain a degree
of homogeneity in the sample, couples had to meet the following inclusion
criteria: (1) living together for at least 1 year; (2) one partner had to have
identified an incident of betrayal or loss of trust in the relationship; and
(3) scores on the Dyadic Adjustment Scale (DAS) had to be in the mild to
moderate distress range for at least one partner (i.e., scores between 80 and
97); scores lower than 97 indicate marital distress, yet scores lower than 80
indicate severe distress and scores 70 or below correspond to those cou-
ples seeking divorce (Spanier, 1976). Exclusion criteria included (1) drug
or alcohol problems; (2) psychiatric history; (3) a history of sexual abuse;
and (4) physical violence in the relationship. Of the 24 couples who partic-
ipated in the original study, 12 couples participated in the follow-up study
(i.e., 8 resolved couples and 4 nonresolved couples from post-treatment).
Of these couples, there were 9 injured wives and 3 injured husbands.
The types of attachment injuries reported were an actual abandonment
(n =2), perceived abandonment following a miscarriage (n =1), infidelity
(n =5), flirtation (n =1), friendship with opposite sex (n =2), and financial
deception/loss (n =1). The mean age was 36.5 years (SD =8.6). Couples
had lived together for a mean of 10.1 years (SD =7.1), and had a mean
of 1.5 children (SD =1.2). In terms of racial groups, all couples were Cau-
casian, excluding one nonresolved couple who were of East Indian descent.
Finally, with respect to socioeconomic status, the range of gross family in-
come ranged from $40,000 to $100,000 and $50,000 to $90,000 in resolved
and nonresolved couples, respectively.
SELF-REPORT MEASURES
The following self-report instruments were selected as they were used to
assess outcome measures in the original study and thus were essential for
comparison purposes. In addition, they have previously demonstrated the-
oretical relevance to EFT and an ability to predict outcome in distressed
couples (Johnson & Talitman, 1997; Millikin, 2000).
36 R. E. Halchuk et al.
DYADIC ADJUSTMENT SCALE (DAS)
The DAS (Spanier, 1976) is a 32-item self-report measure of marital satis-
faction that uses a Likert-type response format. Reliability for this scale has
been determined to be 0.96 (Spanier, 1976). The DAS is scored by summing
the weights of each fixed response. Scores can range from 0 to 151, with
lower scores indicative of more distress and lower adjustment. Mean total
scale scores of 114.8 are indicative of happily married couples and 70.7 for
divorced couples (Spanier, 1976). A couple’s mean score is obtained by aver-
aging the sum of each partner’s score. For the follow-up sample, Cronbach’s
alpha for the DAS was .97.
RELATIONSHIP TRUST SCALE (RTS)
The RTS (Holmes, Boon, & Adams, 1990) is a 30-item self-report inventory
designed to assess interpersonal trust using a Likert-type response format.
Reliability for this scale has been determined to be .89 (Holmes et al., 1990).
The theoretical range of scores is 30 to 210. High scores indicate a stronger
presence of trust between partners. A couple’s mean score is obtained by
averaging the sum of each partner’s score. For the follow-up sample, Cron-
bach’s alpha for the RTS was .98.
EXPERIENCES IN CLOSE RELATIONSHIPS (ECR)
The ECR (Brennan, Clark, & Shaver, 1998) is a 36-item measure consisting of
two 18-item self-report scales that assess individual differences with respect
to attachment-related anxiety and avoidance. Reliability coefficients for the
avoidance and anxiety scales were .94 and .91, respectively (Brennan et al.,
1998). Responses are recorded using a Likert-type response format. Each
subscale is individually summed, and scores may range from 18 to 126. High
scores indicate avoidance and anxiety. For the follow-up sample, Cronbach’s
alpha for avoidance and anxiety scales was .95 and .89, respectively.
INTERPERSONAL RELATIONSHIP RESOLUTION SCALE (IRRS)
The IRRS (Hargrave & Sells, 1997) is a 44-item measure consisting of two
scales designed to assess the extent to which a person continues to feel pain
as a result of an offense and has forgiven that person for the offence. Relia-
bility coefficients for the forgiveness and the emotional pain scale were .92
and .95, respectively (Hargrave & Sells, 1997). Injured partners were asked
to respond with yes or no to each item. High scores on the forgiveness
scale indicate that the individual has made little progress in the work to-
ward forgiveness. Conversely, high scores on the pain scale indicate that the
Resolving Attachment Injuries 37
individual has made considerable progress in dealing with their pain. For
the follow-up sample, Cronbach’s alpha for the forgiveness and emotional
pain scales was .78 and .80, respectively.
ATTACHMENT INJURY MEASURE (AIM)
The AIM (Millikin, 2000) is a modification of the single-item Target Com-
plaints Discomfort Box Scale (Battle et al., 1966). The measure was expanded
to four items designed to measure the current severity of the injury using a
5-point scale ranging from 1 (severe)to5(negligible). Items included: “How
do you rate the event that injured your relationship bond?” “How does it af-
fect the level of trust between you and your partner?” “How does the injury
interfere with your relationship now?” “When you talk to your partner about
the injury how much of a problem does it create?” The theoretical range of
scores is 4 to 20. High scores indicate greater resolution of the attachment
injury than low scores. A couple’s mean score is obtained by averaging the
sum of each partner’s score. For the follow-up sample, Cronbach’s alpha for
the AIM was .77.
Procedures
All couples who originally received EFT were contacted by telephone and
invited to participate in the 3-year follow-up study. Couples who agreed
to participate were sent a complete set of questionnaires corresponding
to those used in the original study along with postage-paid return en-
velopes. The questionnaires were self-report outcome measures: the DAS, the
RTS, the ECR, the IRRS, and the AIM. Along with the questionnaires, couples
were sent an information and consent form. This study was conducted in
compliance with the University of Ottawa’s Research Ethics Board.
RESULTS
To test the hypothesis that previously resolved couples would show stabil-
ity in prior treatment gains, repeated-measures analyses with resolved and
nonresolved groups as the between-subjects independent variable and time
(i.e., post-treatment and follow-up) and partner type (i.e., injured vs. offend-
ing) as the within-subjects independent variables were conducted. Separate
analyses were conducted on each dependent variable (i.e., dyadic adjust-
ment, relationship trust, avoidant attachment and anxious attachment, and
the attachment injury measure). As the IRRS forgiveness and emotional pain
scores existed only for the injured parties, the only within-subjects indepen-
dent variable used was time.
38 R. E. Halchuk et al.
TABLE 1 Means and Standard Deviations for Resolved and Nonresolved Groups on Dyadic
Adjustment at Post-treatment and Follow-up for Injured and Offending Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda114.38 (14.81) 110.63 (12.92) 100.25 (26.49) 103.38 (17.49)
Nonresolvedb78.25 (11.50) 84.50 (8.81) 67.25 (27.17) 81.25 (11.27)
Note. Values in parentheses represent standard deviations.
an=8 couples.
bn=4 couples.
Given that 12 couples that had participated in the original study were
not part of the follow-up study, this limited the type of analyses that could
be performed and what corrections were feasible. Due to the exploratory
nature of these data, Bonferroni corrections were not made, to reduce the
possibility of making a Type II error.
DYADIC ADJUSTMENT
There was a significant difference between groups on dyadic adjustment
from post-treatment to follow-up, F(1,10) =13.89, p<.005, ηp2=.58,
suggesting that treatment effects remained stable over time. See Table 1 for
means and standard deviations.
RELATIONSHIP TRUST
There was a two-way interaction for partner type by group on relationship
trust, F(1,10) =5.36, p<.05, ηp2=.35, and a main effect for partner
type, F(1,10) =21.15, p=.001, ηp2=.68. The injured partners’ level of
relationship trust was significantly lower than for the offending partners’
level of relationship trust at both times. There was also a significant effect
between groups, F(1,10) =11.99, p<.01, ηp2=.55. See Table 2 for means
and standard deviations.
TABLE 2 Means and Standard Deviations for Resolved and Nonresolved Groups on Rela-
tionship Trust at Post-treatment and Follow-up for Injured and Offending Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda158.38 (28.72) 167.00 (23.59) 146.75 (46.69) 155.38 (48.12)
Nonresolvedb92.00 (22.91) 110.75 (17.78) 88.50 (20.70) 122.00 (11.28)
Note. Values in parentheses represent standard deviations.
an=8 couples.
bn=4 couples.
Resolving Attachment Injuries 39
TABLE 3 Means and Standard Deviations for Resolved and Nonresolved Groups on Avoid-
ance at Post-treatment and Follow-up for Injured and Offending Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda49.08 (15.32) 44.00 (11.30) 50.88 (30.82) 47.00 (21.90)
Nonresolvedb45.25 (17.93) 61.25 (12.84) 68.50 (24.08) 58.75 (6.08)
Note. Values in parentheses represent standard deviations.
an=8 couples.
bn=4 couples.
AVOIDANT ATTACHMENT DIMENSION
There was a three-way interaction for partner type by time by group
F(1,10) =5.14, p<.05, ηp2=.34. Table 3 shows the means and stan-
dard deviations for the avoidance attachment dimension. The presence of a
three-way interaction suggests that there are differences over time between
both partner type and by group placement. Specifically, nonresolved injured
partners reported higher levels of avoidant attachment at follow-up.
ANXIOUS ATTACHMENT DIMENSION
There was no significant difference between groups in terms of anxious
attachment, F(1,10) =.85, p>.01, ηp2=.08. See Table 4 for means and
standard deviations.
EMOTIONAL PAIN AND FORGIVENESS
In terms of forgiveness, there was a significant difference between groups,
F(1,10) =9.55, p<.05, ηp2=.49. Resolved partners reported more forgive-
ness than nonresolved partners at follow-up. With respect to emotional pain,
there was a significant difference between groups, F(1,10) =10.54, p<.01,
ηp2=.51. Higher scores reported by resolved partners indicate that they
TABLE 4 Means and Standard Deviations for Resolved and Nonresolved Groups on Anxiety
at Post-treatment and Follow-up for Injured and Offending Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda60.13 (14.71) 71.02 (20.23) 49.12 (15.60) 57.00 (22.85)
Nonresolvedb60.75 (25.91) 72.00 (14.54) 70.00 (24.86) 59.00 (9.13)
Note. Values in parentheses represent standard deviations.
an=8 couples.
bn=4 couples.
40 R. E. Halchuk et al.
TABLE 5 Means and Standard Deviations for the Resolved and Nonresolved Groups on
Emotional Pain and Forgiveness at Post-treatment and Follow-up for Injured Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda40.21 (1.32) 24.36 (1.90) 42.00 (1.31) 26.00 (4.41)
Nonresolvedb37.25 (4.99) 31.00 (2.16) 37.75 (3.59) 30.75 (4.65)
Note. Values in parentheses represent standard deviations.
an=8 partners.
bn=4 partners.
TABLE 6 Means and Standard Deviations for Resolved and Nonresolved Groups on Attach-
ment Injury Resolution at Post-treatment and Follow-up for Injured and Offending Partners
Post-treatment Follow-up
Groups Injured Offending Injured Offending
Resolveda15.13 (2.17) 14.50 (2.00) 15.75 (3.06) 13.06 (2.48)
Nonresolvedb8.00 (4.24) 9.50 (3.11) 9.25 (2.06) 12.25 (3.95)
Note. Values in parentheses represent standard deviations.
an=8 couples.
bn=4 couples.
have made more progression than nonresolved partner in terms of dealing
with their emotional pain. See Table 5 for the means and standard deviations
for the resolved and nonresolved groups on emotional pain and forgiveness
at post-treatment and follow-up.
ATTACHMENT INJURY RESOLUTION
There was a two-way interaction between partner type by group, F(1,10) =
5.16, p<.05, ηp2=.34. In addition, there was a significant effect between
groups, F(1,10) =21.13, p=.001, ηp2=.68. The two-way interaction be-
tween group and partner is likely due to injured partners reporting higher
levels of attachment injury resolution in the resolved group, and lower levels
in the nonresolved. See Table 6 for means and standard deviations.
DISCUSSION
The purpose of this research was to investigate the long term effects of EFT
at three year follow-up, and to explore whether resolved and nonresolved
couples having undergone EFT for an attachment injury are discriminated
by distal outcome measures. This follow-up study speaks to the lack of
Resolving Attachment Injuries 41
long-term follow-up assessments in the area of marital research in the general
population (Jacobson & Addis, 1993).
The hypothesis that couples previously identified as having resolved
their attachment injury would show stability in their gains in dyadic adjust-
ment from post-treatment to follow-up was supported. This is a substantial
finding for EFT as the majority of research on long-term follow-up with
Behavioral Marital Therapy indicates that for the majority of couples, mari-
tal satisfaction regresses to pre-treatment levels 4 to 5 years after treatment
(Snyder, Wills, & Gradys-Fletcher, 1991). For the resolved couples, the mean
couple DAS score showed no significant changes from post-treatment to
follow-up, despite a slight decrease. The nonresolved couples also showed
nonsignificant changes in dyadic adjustment despite slight decreases from
post-treatment to follow-up with mean DAS scores remaining significantly
lower than resolved injured couples. The nonresolved couples’ mean DAS
scores indicate that they are still distressed. Moreover, injured nonresolved
partners reported a mean DAS score of 67 at follow-up, which is below
the average of 70.7 for divorced couples (Spanier, 1976). These results are
not surprising as nonresolved couples, specifically the injured partners, have
been living in a distressed relationship for several years, and without reso-
lution they are likely still entrapped in their cycles of being unable to reach
for each other and provide support (Johnson, 2004). As for the decreases in
scores for both groups, it should be noted that this may be due to increased
variability due to the small sample size. Nevertheless, there were still signif-
icant differences between the two groups indicating the strength of effects
conferred to resolved couples having undergone EFT.
Stability in post-treatment results at follow-up was also found among
resolved and nonresolved couples on the level of relationship trust. The
mean scores for resolved couples decreased slightly from post-treatment
to follow-up, but this was not significant. The nonresolved couples also
showed no significant difference from post-treatment to follow-up, although
offending partners showed a slight increase in trust over time.
The observed stability in the levels of dyadic adjustment and trust over
time, in both resolved and nonresolved couples, is consistent with the liter-
ature, which reports these two variables to be related (Holmes & Remple,
1989), and that positive scores on these variables are characteristic of se-
curely attached couples (Kobak & Hazan, 1991). This suggests that over time
couples who were able to resolve their attachment injury in therapy sessions,
were more likely to maintain high levels of marital satisfaction and greater
trust in their relationship than those couples who did not manage to resolve
their attachment injury in therapy.
The hypothesis that resolved couples would show significant decreases
in anxious and avoidant attachment at follow-up compared to nonresolved
couples was not supported by the data. Despite the nonsignificant findings
between groups, there was a significant three-way interaction effect between
42 R. E. Halchuk et al.
time, partner type, and group for the avoidance dimension. This suggests that
over time, injured and offending partners responded differentially depend-
ing on whether they had been designated as resolved or nonresolved with
respect to their attachment injury. Specifically, nonresolved injured partners
reported higher levels of avoidance in their relationship at follow-up than at
post-treatment. Again, this is likely attributable to the fact that injured part-
ners have spent 3 years in a relationship with an insecure attachment that
has not been remedied, and hence, have increased their avoidance of their
partner who serves as both the solution to and a source of distress to them
(Main & Hesse, 1990). The anxiety dimension did not show any significant
changes over time, with respect to partner type or between groups.
Consistent with previous results comparing pre- and post-treatment at-
tachment, these findings also show attachment anxiety and avoidance as
measured by this scale to be poor predictors of attachment injury resolution
(Makinen & Johnson, 2006). Moreover, as was suggested in the previous
study (Makinen & Johnson, 2006), the ECR scale was adjusted so that ques-
tions were worded to measure attachment behaviors exclusive to the couple,
in hopes of increasing sensitivity to any resulting changes. Previously this
measure was administered with ambiguous wording that left room to inter-
pret questions as being directed specifically towards one’s current partner,
or as a general trend arising from one’s previous romantic history. Unfor-
tunately, changes to the scale did not approach significance, as effect sizes
remained small for the avoidance and anxiety dimensions for the two groups.
The lack of significant findings between resolved and nonresolved
groups on the attachment measures may be due to several reasons. To begin,
it should be noted that couples’ attachment patterns could not be ascertained
prior to the injurious event. Thus, it cannot be precisely determined whether
the couples’ work on their attachment injury altered their long-standing at-
tachment patterns. Another possibility as suggested in the original study is
that this attachment measure is not sensitive enough to detect changes in
avoidance or anxiety resulting from treatment (Makinen & Johnson, 2006).
Despite changes made to the ECR in an attempt to increase this sensitiv-
ity, perhaps another scale would provide improved distinction over time,
across partners, and between groups. However, given that resolved cou-
ples reported increased marital satisfaction and improved trust over time, it
seems unlikely that there would not be a corresponding shift in attachment
patterns. As noted in the original study, perhaps EFT helps couples to in-
teract in more positive ways, rather than alter underlying attachment styles
(Makinen & Johnson, 2006).
Following from this, another possible explanation stems from previous
research, which has noted the extreme difficulty in altering attachment be-
haviors associated with specific patterns (Brennan et al., 1998; Keelan, Dion,
& Dion, 1994). Furthermore, attachment patterns have demonstrated sub-
stantial continuity over time when there is continuity in the environment
Resolving Attachment Injuries 43
(Baldwin & Fehr, 1995; Davila, Karney, & Bradbury, 1999). Perhaps the non-
resolved couples did not experience enough security in therapy to be able
to change their environments, which could be due to the limited amount of
therapy sessions and the fact that they had compounded attachment injuries.
Research has shown that security allows an individual to be open to new
experience (Bar-Haim, Sutton, Fox, & Marvin, 2000). A likely circumstance
is that the traumatic injury triggered attachment-related behaviors consistent
with the insecure attachment patterns. At present there exists a lack of re-
search examining the stability of attachment patterns in adults who have
experienced traumatic events (Scharfe, 2003).
Lastly, the nonsignificant findings may also be due to low power asso-
ciated with small sample size. A risk with any study with small nis that it
is inherently difficult to account for a lack of significance. Especially given
that EFT has been shown to be effective at reducing relationship distress
(Johnson et al., 1999), it is possible that the actual changes in the attachment
dimensions are small and cannot be detected with a small sample.
With respect to forgiveness, the hypothesis that resolved injured part-
ners would show stability with respect to increased forgiveness levels at
post-treatment was supported, as there were no changes with respect to
time. In terms of forgiveness, the lack of significant group differences over
time was as expected. Post-treatment data revealed that the resolved couples
made significant gains in terms of their level of forgiveness through therapy
(Makinen & Johnson, 2006), whereas no change was noted for the nonre-
solved couples. Therefore, finding no significance in differences over time
suggests that these gains were maintained over the last 3 years.
The hypothesis that there would be a decrease in pain levels over time
was not supported. Original post-treatment data reported that there was no
significant difference between resolved and nonresolved partners following
therapy (Makinen & Johnson, 2006). However, there were significant group
differences in terms of pain at follow-up. Specifically, resolved injured part-
ners reported slightly lower levels of pain than nonresolved injured partners.
It is possible that differences across time for the two groups were not
detected because of low levels of internal consistency for the forgiveness
and pain subscales. Resolved partners may in fact have experienced higher
levels of forgiveness and less pain, than nonresolved partners. Compared to
the literature, reliability analyses revealed low internal consistency for the
pain subscale at post-treatment (.73) and at follow-up (.80).
The prediction that couples identified as having resolved the attachment
injury at post-treatment would show stability in their gains of attachment
injury resolution from post-treatment to follow-up was supported. The mean
score for resolved couples was stable from post-treatment to follow-up.
The nonresolved couples also showed no significant difference from post-
treatment to follow-up, despite a slight increase in scores. There were signif-
icant differences between resolved and nonresolved couples at follow-up.
44 R. E. Halchuk et al.
The two-way interaction between partner type and group is due to the
fact that resolved injured partners report higher levels of attachment injury
resolution than nonresolved injured partners. This makes sense as resolved
injured partners experienced the injury as more severe than offending
partners prior to therapy (Makinen & Johnson, 2006). Therefore, it makes
intuitive sense that injured partners will report higher levels of resolution
than offending partners. Conversely, injured partners who have not resolved
their injury are likely still experiencing insecurity surrounding the injury and
correspondingly still view the injury as more severe than offending partners.
In general, then, hypotheses were supported except for insignificant
group differences found for the anxious and avoidant attachment dimensions
and reported pain levels. Significant differences were detected between cou-
ples on measures of dyadic attachment, trust, forgiveness, and attachment
injury severity.
There are several important limitations inherent to the study that deserve
consideration. The most striking of these is the small sample size. Unfortu-
nately but not unsurprisingly, half of the sample was lost at follow-up, which
resulted in there being little power available to determine significant differ-
ences between groups. Consequently, although results were determined us-
ing ANOVAs, our small sample size indicates that these must be interpreted
with caution and may only be considered preliminary support for the long-
term impact of EFT. Indeed, additional follow-up studies will be needed to
help further support findings.
Another limitation relates to the fact that there was no control group
used in the original study or the follow-up. This certainly poses a threat to
the internal validity of this research as only a single group was used and
no random assignment was possible. Unfortunately, the addition of a con-
trol group would have been difficult, expensive, and inefficient. In terms of
choosing a control group, this would have had to consist of another group of
distressed couples that did not receive therapy. From an ethical standpoint,
researchers would have had to provide the control group with therapy af-
terward, and this was simply beyond the means of the researchers to do
so. Also, as the original study was process oriented, where the focus was to
investigate the steps underlying the development of forgiveness and recon-
ciliation and to examine how this differs between resolved and nonresolved
couples, a control group was simply beyond the scope of the study.
A further limitation stems from the fact that partners’ scores were
summed to yield a mean couple score on all outcome variables of interest.
This was in keeping with the original article that also focused on explor-
ing resolved and nonresolved group differences. Unfortunately, this method
restricted our ability to examine gender differences within the sample and
across groups, but given our small sample size and limited power, the ex-
ploration of smaller subgroups would have been statistically unfeasible.
Resolving Attachment Injuries 45
Future research will endeavor to adopt a more classic experimental
design in which distressed couples are randomly assigned to either an
EFT treatment group or to a control group where couples would receive
educational classes on forgiveness. Also, it is recognized that more follow-up
studies are needed to confirm the lasting effects of EFT, and longitudinal
studies are being actively designed and implemented to address this. At
present, the most likely reason for the stability of results in this study
appears, in the light of clinical practice in EFT and the understanding
offered by attachment theory, to be that once a couple can resolve
these kinds of injuries and have a mutually accessible and responsive
dialogue, the bond between then becomes more secure. This security
buffers them against future stress and relationship distress, allowing them
to deal with stressors as a team and thus continually strengthens their
relationship.
In terms of clinical relevance, this study highlights the utility of EFT as
a clinical treatment method whose effects may endure with time. The ben-
efits conferred to couples who resolved their attachment injuries through
the attachment injury resolution model are stable over 3 years. In addi-
tion, the model continues to discriminate between resolved and nonresolved
couples on measures of dyadic adjustment, trust, forgiveness, and injury
severity.
REFERENCES
Baldwin, M. W., & Fehr, B. (1995). On the instability of attachment style ratings.
Personal Relationships,2, 247–261.
Bar-Haim, Y., Sutton, D. B., Fox, N. A., & Marvin, R. S. (2000). Stability and change
of attachment at 14, 24, and 58 months of age: Behaviour, representation, and
life events. Journal of Child Psychology and Psychiatry,41, 381–388.
Battle, C. C., Imber, S. D., Hoehn-Saric, R., Stone, A. R., Nash, E. R., & Frank,
J. D. (1966). Target complaints as criteria of improvement. American Journal of
Psychotherapy,20, 184–192.
Baucom, D., Shoham, V., Mueser, K., Daiuto, A., & Stickle, T. (1998). Empirically
supported couple and family interventions for marital distress and adult mental
health problems. Journal of Consulting and Clinical Psychology,66, 53–88.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult
attachment. In. J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and
close relationships (pp. 46–76). New York: The Guilford Press.
Davila, J., Karney, B. R., & Bradbury, T. N. (1999). Attachment changes processes
in the early years of marriage. Journal of Personality and Social Psychology,79,
783–802.
Dessaulles, A., Johnson, S. M., & Denton, J. H. (2003). Emotion-focused therapy for
couples in the treatment of depression: A pilot study. The American Journal of
Family Therapy,31, 345–353.
46 R. E. Halchuk et al.
Gordon, K. C., Baucom, D. H., & Snyder, D. K. (2000). The use of forgiveness in
marital therapy. In M. E. McCullough, K. I. Pargament, & C. E. Thoresen (Eds.),
Forgiveness: Theory, research, and practice (pp. 203–227). New York: Guilford
Press.
Gordon-Walker, J., Johnson, S., Manion, L., & Cloutier, P. (1996). Emotionally fo-
cused marital interventions for couples with chronically ill children. Journal of
Consulting and Clinical Psychology,64, 1029–1036.
Gordon-Walker, J., Manion, I., & Clothier, P. (1998). Emotionally focused interven-
tion for couples with chronically ill children. A two-year follow-up. Journal of
Marital and Family Therapy,28, 391–399.
Halford, W. K., Kelly, A., & Markman, H. J. (1997). The concept of a healthy marriage.
In W. K. Halford & H. J. Markman (Eds.), Clinical handbook of marriage and
couples interventions (pp. 3–12). New York: John Wiley & Sons.
Hargrave, T. D., & Sells, J. N. (1997). The development of a forgiveness scale. Journal
of Marital and Family Therapy,23, 41–62.
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment
process. Journal of Personality and Social Psychology,52, 511–524.
Holmes, J. G., Boon, S. D., & Adams, S. (1990). The Relationship Trust Scale. Un-
published manuscript. University of Waterloo, Waterloo, Canada.
Holmes, J. G., & Remple, J. K. (1989). Close relationships. Newbury Park, CA: Sage.
Jacobson, N. S., & Addis, M. E. (1993). Research on couples and couple therapy:
What do we know? Where are we going? Journal of Consulting and Clinical
Psychology,59, 547–557.
Johnson, S. M. (2004). The practice of emotionally focused couple therapy: Creating
connections, (2nd ed.). New York: Brunner/Mazel.
Johnson, S. M., & Greenberg, L. S. (1988). Relating process to outcome in marital
therapy. Journal of Marital and Family Therapy,14, 175–183.
Johnson, S. M., Hunsley, J., Greenberg, L. S., & Schindler, D. (1999). Emotionally
focused couples therapy: Status and challenges. Clinical Psychology: Science
and Practice,6, 67–79.
Johnson, S. M., Makinen, J. A., & Millikin, J. W. (2001). Attachment injuries in couple
relationships: A new perspective on impasses in couples therapy. Journal of
Marital and Family Therapy,27, 145–155.
Johnson, S. M., & Talitman, E. (1997). Predictors of success in emotionally focused
marital therapy. Journal of Marital and Family Therapy,23, 135–152.
Johnson, S. M., & Williams-Keeler, L. (1998). Creating healing relationships for cou-
ples dealing with trauma: The use of emotionally focused couples therapy.
Journal of Marital and Family Therapy,24, 25–40.
Jones, W. H., & Burdette, M. P. (1994). Betrayal in relationships. In A. L. Weber &
J. H. Harvey (Eds.), Perspectives on close relationships (pp. 243–262). Boston:
Allyn & Bacon.
Keelan, J. P. R., Dion, K. L., & Dion, K. K. (1994). Attachment styles and heterosexual
relationships among young adults: A short-term panel study. Journal of Social
and Personal Relationships,11, 201–214.
Resolving Attachment Injuries 47
Kobak, R., & Hazan, C. (1991). Attachment in marriage: Effects of security and
accuracy of working models. Journal of Personality and Social Psychology,60,
861–869.
Main, M., & Hesse, E. (1990). Parent’s unresolved traumatic experiences are related
to infant disorganized attachment status. In M. Greenberg & D. Cicchetti (Eds.).
Attachment in the pre-school years (pp. 152–176). Chicago: University of Chicago
Press.
Makinen, J. A., & Johnson, S. M. (2006). Resolving attachment injuries in couples us-
ing Emotionally Focused Therapy: Steps towards forgiveness and reconciliation.
Journal of Consulting and Clinical Psychology,74, 1055–1064.
Millikin, J. A. (2000). Resolving attachment injuries in couples using emotionally
focused therapy: A process study. Unpublished doctoral dissertation, Virginia
Polytechnic Institute and State University, Virginia.
Pennebaker, J. W. (1985). Traumatic experience and psychosomatic disease: Explor-
ing the psychology of behavioral inhibition, obsession and confiding. Canadian
Psychology,26, 82–95.
Scharfe, E. (2003). Stability and change of attachment representations from cradle to
grave. In S. M. Johnson & V. E. Whiffen (Eds.), Attachment processes in couple
and family therapy (pp. 64–84). New York: Guilford Press.
Schore, A. N. (1994). Affect regulation and the organization of self. Hillsdale, NJ:
Erlbaum.
Simpson, J., & Rholes, W. (1994). Stress and secure base relationships in adulthood.
In K. Batholomew & D. Perlman (Eds.), Attachment processes in adulthood (pp.
181–204). London: Jessica Kingsley.
Snyder, D. K., Wills, R. M., & Grady-Fletcher, A. (1991). Long-term effectiveness
of behavioural versus insight-oriented marital therapy: A four-year follow-up.
Journal of Consulting and Clinical Psychology,59, 138–141.
Spanier, G. (1976). Measuring dyadic adjustment. Journal of Marriage and the Fam-
ily,38, 15–28.
