THE RELATIONSHIP BETWEEN ATTACHMENT AND EMOTIONAL EXPERIENCE
DURING SEMI-NATURAL AND THERAPY-LIKE COUPLE INTERACTIONS
By
Ryan B. Seedall

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Human Development and Family Studies
2011

ABSTRACT
THE RELATIONSHIP BETWEEN ATTACHMENT AND EMOTIONAL EXPERIENCE
DURING SEMI-NATURAL AND THERAPY-LIKE COUPLE INTERACTIONS
By
Ryan B. Seedall
Although a sizeable amount of research has investigated and demonstrated the
effectiveness of couple therapy, a substantial number of couples still do not receive the optimal
benefits of statistically and clinically significant change. This highlights the importance of
moving from if to how couple therapy is effective. More specifically, it is important to
understand client experience in therapy and those factors that influence both the course and
outcome of therapy. This would facilitate the ability to more effectively adapt treatment to the
needs of clients. Client-treatment matching research has attempted to accomplish this, but
results have been disappointing as it has primarily focused on matching treatment models to
specific diagnoses. A common factors perspective provides the insight that the core ingredients
of change that exist across treatment models may be confounding results. If this is the case, it
would be important to understand client-treatment matching from a more process-oriented
approach that looks at how specific client and treatment characteristics interact to influence
client experience and ultimately treatment outcome.
This study sought to begin looking at treatment process from a common factors
perspective by looking at how attachment as a client characteristic interacted with a low (seminatural) or high (therapy-like) structure interactional context to influence two indicators of
emotional experience, physiological arousal and interpersonal distress. Emotional experience, as
defined by emotional processing and regulation, ultimately influences emotional expression and
has been shown to influence therapeutic outcomes. Findings appear to indicate that attachment

influences emotional experience and provides preliminary evidence that taking into account
attachment may be useful in improving outcomes. Specific relationship process, clinical, and
research implications are discussed. Overall, it is hoped that this study will provide the impetus
for additional work that will continue to investigate how attachment and other client
characteristics interact with elements of therapy process to positively influence client experience
and outcomes. This will then facilitate the more effective adaptation of treatment to the specific
needs of clients.

ACKNOWLEDGMENTS

As I near the end of my doctoral degree, I would be remiss if I did not express gratitude
to those who had an integral part in helping me to reach this point. I would like to first
acknowledge the hand of God in giving me all that I have and making all of this possible and
giving me all that I have. I next want to thank Ruth for her continual patience and support. I am
so glad that I am on this journey of life with her. Her strength has meant more than she will ever
know. I also wish to express love and gratitude to Spencer, Madelyn, Eliza, and Benjamin.
They are sweet children and inspirations to me. So much of my family‘s lives have been
organized around my education. I am so grateful for their sacrifice and unconditional love. I
want to also thank my mother and family, Margarete and the late Jerry Seedall. They provided
me with a wonderful foundation. They taught me the value of work, something that served me
well throughout my graduate degrees. However, as important was their faith in me. Throughout
this process, God and my family gave me perspective and helped me never lose sight of those
things that are most important.
I also want to thank Karen Wampler for her tremendous mentoring. I do not believe I
could have hoped for a better mentoring experience. I always knew that she had my best
interests in mind and saw the very best in me. I also want to thank my committee members,
Richard Wampler, Adrian Blow, and Deborah Kashy, for their time, support, and helpful
feedback. My dissertation is much better because of their input and support. I also want to
acknowledge my first advisor, Mark Butler. His mentoring helped to put me on my current
academic path. Overall, I have been tremendously blessed with academic mentors who believed
in me when I did not, taught me the skills I needed to succeed, and were true friends throughout.

iv

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES .........................................................................................................................x
CHAPTER I
INTRODUCTION ...........................................................................................................................1
CHAPTER II
LITERATURE REVIEW ................................................................................................................8
Effectiveness Research in Couple Therapy .........................................................................8
Increasing Effectiveness of Couple Therapy through Client-Treatment Matching ...........11
Matching Type of Treatment to Client Characteristics..........................................12
A Common Factors Approach to Client-Treatment Matching ..............................14
Matching Treatment Process to Client Characteristics ..........................................16
Emotions in Couple Relationships and Therapy ................................................................19
An Overview of Attachment Theory and Close Relationships ..........................................22
Attachment in Adulthood ...................................................................................................24
Adult Attachment: Two Alternative Measurement Approaches ...........................25
Narrative/Discourse Attachment Measurement .....................................................27
The Adult Attachment Interview and the Strange Situation ......................28
Self-Report Attachment Measurement ..................................................................29
Attachment and Emotional Experience in Couple Relationships ..........................32
Adult Attachment and Physiological Arousal........................................................34
Narrative/Discourse Attachment ................................................................35
Self-Report Attachment .............................................................................35
Adult Attachment and Interpersonal Distress ........................................................37
Narrative/Discourse Attachment ................................................................39
Self-Report Attachment .............................................................................41
Attachment in Psychotherapy ................................................................................43
Narrative/Discourse Attachment ................................................................43
Self-Report Attachment .............................................................................45
Attachment in Clinical Work with Couples ...........................................................46
The Purpose of This Study .................................................................................................48
CHAPTER III
METHODS ....................................................................................................................................51
Design ................................................................................................................................51
Predictor Variables.................................................................................................52
Outcome Variables.................................................................................................54
Other Variables ......................................................................................................56
Participants .........................................................................................................................56
Couples ..................................................................................................................56

v

Participant Recruitment .............................................................................56
Participant Exclusion Criteria ....................................................................57
Participant Demographics ..........................................................................59
Therapists ...............................................................................................................61
Procedure ...........................................................................................................................64
Couple Data Collection Procedure.........................................................................64
Phase 1: Informed Consent and Adult Attachment Interview ...................64
Phase 2: Discussion Topic Development ...................................................65
Phase 3: Baseline, Semi-Natural, and Therapy-Like Interactions .............65
Phase 4: Follow-Up Session with Participants ..........................................67
Confirmation of Conditions ...................................................................................68
Occurrence of Each Condition ...................................................................68
Integrity of the Therapy-Like Condition Measured by Self-Report ..........69
Integrity of the Therapy-Like Condition Measured by Observational
Coding ........................................................................................................69
Measures ............................................................................................................................71
Continuous Measures .............................................................................................71
Physiological Arousal ................................................................................71
Interpersonal Distress.................................................................................73
Self-Report Measures.............................................................................................74
Attachment: The Experiences in Close Relationships Questionnaire........74
Emotional Tone of the Relationship: The Subjective Emotional
Experience Questionnaire ..........................................................................75
Intrapersonal Distress: Outcome Questionnaire-45.2 ................................76
Couple Conflict Experiences: Conflicts and Problem-Solving Scales ......77
Relationship Satisfaction: The Revised Dyadic Adjustment Scale ...........78
CHAPTER IV
RESULTS ......................................................................................................................................79
General Approach to Data Analysis ..................................................................................79
Preliminary Analyses .............................................................................................81
Missing Data ..............................................................................................81
Predictor Variables.....................................................................................81
Outcome Variables.....................................................................................82
Dyadic Data Preparation ............................................................................83
Demographic Variable Inclusion ...............................................................84
Decision Process for the Analysis of Demographic Variables, Covariates,
and Interactions ..........................................................................................87
Statistical Tests of Hypotheses ..........................................................................................89
Research Questions and Their Analyses ................................................................89
Research Question 1: Level of Physiological Arousal ..............................89
Research Question 1: Variability of Physiological Arousal ......................91
Research Question 1: Change in Physiological Arousal............................94
Research Question 1: Level of Moment-to-Moment Feelings towards
One‘s Partner .............................................................................................98
Research Question 1: Proximal Self-Report of the Level of Feelings

vi

towards Partner ........................................................................................101
Research Question 1: Variability of Moment-to-Moment Feelings towards
One‘s Partner ...........................................................................................103
Research Question 1: Change in Moment-to-Moment Feelings towards
One‘s Partner ...........................................................................................106
Research Question 2: Physiological Arousal ...........................................109
Research Question 2: Feelings towards One‘s Partner ............................110
Research Question 3: Congruence between Level of Physiological Arousal
and Feelings towards One‘s Partner ........................................................115
Research Question 3: Congruence between Variability of Physiological
Arousal and Variability of Feelings towards One‘s Partner ....................120
Research Question 3: Congruence between Change in Physiological
Arousal and Change in Feelings towards One‘s Partner .........................122
CHAPTER V
DISCUSSION ..............................................................................................................................124
Overview of the Study .....................................................................................................124
A Common Factors Perspective of Client-Treatment Matching .........................124
Attachment Theory as an Important Client Characteristic ..................................125
Research Questions and a Brief Summary of Findings .......................................127
Relationship Process Implications ...................................................................................131
Attachment Avoidance and Repressive Coping ..................................................131
Attachment Anxiety and Partner Suppression .....................................................134
Clinical Implications ........................................................................................................136
Emotional Experience during the Couple Interactions ........................................136
Attachment Anxiety and Clinical Process ...........................................................139
Attachment Avoidance and Clinical Process .......................................................140
Research Implications ......................................................................................................143
Process Research from a Common Factors Perspective ......................................143
Research on Attachment and Therapy .................................................................144
Conclusion .......................................................................................................................145
Limitations and Suggestions for Future Research ...............................................145
Summary ..............................................................................................................147
APPENDICES .............................................................................................................................148
Appendix A: Brief Screening Assessment .......................................................................149
Appendix B: Facilitating Emotional Content – Common Treatment Processes..............150
Appendix C: Research Associate Interaction with Participants ......................................156
Appendix D: Informed Consent .......................................................................................162
Appendix E: Therapist/Coach Content and Quality Check .............................................164
REFERENCES ............................................................................................................................166

vii

LIST OF TABLES

TABLE 1a. Descriptive Statistics of Categorical Demographic Variables ...................................60
TABLE 1b. Descriptive Statistics of Continuous Demographic Variables ...................................61
TABLE 2. Descriptive Statistics for Physiological Arousal and Interpersonal Distress ...............73
TABLE 3. Psychometric Properties of Predictor Variables and Covariates .................................75
TABLE 4. Bivariate Correlations between Predictor Variables and Covariates ...........................81
TABLE 5. Bivariate Correlations between Data for Men and Women .........................................84
TABLE 6a. One-Way ANOVA Omnibus F Values for Categorical Demographic Predictors of
Physiological Arousal ....................................................................................................................85
TABLE 6b. Linear Regression Standardized Beta Values for Continuous Demographic
Predictors of Physiological Arousal...............................................................................................86
TABLE 6c. One-Way ANOVA Omnibus F Values for Categorical Demographic Predictors of
Moment-to-Moment Feelings towards Partner ..............................................................................86
TABLE 6d. Linear Regression Standardized Beta Values for Continuous Demographic
Predictors of Moment-to-Moment Feelings towards Partner ........................................................87
TABLE 7. Research Question 1 – Level of Physiological Arousal ...............................................91
TABLE 8. Research Question 1 – Variability of Physiological Arousal ......................................94
TABLE 9. Research Question 1 – Change from Relaxation Baseline of Physiological Arousal ..97
TABLE 10. Research Question 1 – Level of Moment-to-Moment Feelings ...............................100
TABLE 11. Research Question 1 – Variability of Moment-to-Moment Feelings ......................105
TABLE 12. Research Question 1 – Change from Baseline in Moment-to-Moment Feelings ....107
TABLE 13. Research Question 2 – Condition Comparison of Physiological Level, Variability,
and Change...................................................................................................................................109
TABLE 14. Research Question 2 – Condition Comparison of Moment-to-Moment Level,
Variability, and Change in Feelings towards Partner ..................................................................113

viii

TABLE 15. Research Question 3 – Comparison of the Congruence between the Level of
Physiological Arousal and Level of Moment-to-Moment Feelings towards Partner ..................119
TABLE 16. Research Question 3 – Comparison of the Congruence between the Variability of
Physiological Arousal and the Variability of Moment-to-Moment Feelings towards Partner ....121
TABLE 17. Research Question 3 – Comparison of the Congruence between Change in
Physiological Arousal and Change in Moment-to-Moment Feelings towards Partner ...............123

ix

LIST OF FIGURES

FIGURE 1. Process Research from a Common Factor Perspective ..............................................16
FIGURE 2. The Actor-Partner Interdependence Model ................................................................80
FIGURE 3. General Approach to Data Analysis ...........................................................................88
FIGURE 4. Research Question 1 – The Effect of Attachment Anxiety and Gender on the
Variability of Physiological Arousal during the Therapy-Like Interaction ...................................93
FIGURE 5. Research Question 1 – The Effect of Attachment Anxiety and Gender on Change in
Physiological Arousal during the Baseline / Semi-Natural / Therapy-Like Interactions ..............96
FIGURE 6. Research Question 1 – The Effect of Attachment Avoidance and Age on
Physiological Arousal during the Baseline Interaction ..................................................................97
FIGURE 7. Research Question 1 – The Effect of Attachment Avoidance and Gender on
Proximal Feelings towards Partner for all Combined Interactions ..............................................101
FIGURE 8. Research Question 1 – The Effect of Attachment Anxiety and Gender on Proximal
Feelings towards Partner for all Combined Interactions..............................................................102
FIGURE 9. Research Question 1 – The Effect of Attachment Avoidance and Global Feelings
towards Partner on the Variability of Moment-to-Moment Feelings during the Baseline
Interaction ....................................................................................................................................104
FIGURE 10. Research Question 2 – The Effects of Condition and Attachment Anxiety on Level
of Moment-to-Moment Feelings towards Partner........................................................................111
FIGURE 11. Research Question 2 – The Effects of Condition and Attachment Anxiety on
Change from Baseline in Moment-to-Moment Feelings .............................................................112
FIGURE 12. Research Question 3 – Attachment Avoidance and the Congruence between Skin
Conductance and Moment-to-Moment Feelings during the Semi-Natural Interaction ...............116
FIGURE 13. Research Question 3 – Partner Attachment Avoidance and the Congruence between
Skin Conductance and Moment-to-Moment Feelings during the Semi-Natural Interaction .......117
FIGURE 14. Research Question 3 – Attachment Avoidance and the Congruence between Skin
Conductance and Moment-to-Moment Feelings during the Therapy-Like Interaction ...............118

x

CHAPTER I
INTRODUCTION
Although numerous studies and meta-analyses agree that psychotherapy and couple
therapy are effective (Lambert & Ogles, 2003; Shadish, Ragsdale, Glaser, & Montgomery,
1995), a substantial number of clients do not receive the full benefits of treatment (Clarkin &
Levy, 2003; Lambert & Ogles, 2003; Sexton, Alexander, & Mease, 2003). As a result, empirical
research and scholarly work in psychotherapy have sought to identify and define those therapist
and client variables associated with engagement in therapy as well as overall outcome (Beutler et
al., 2003; Clarkin & Levy, 2003). Although more research is needed to identify variables that
contribute to client dropout and poorer outcomes in psychotherapy, the need is much greater in
couple therapy, where only a handful of studies have sought to move beyond if couple therapy is
effective to how.
Existing client-treatment matching research clearly illustrates this issue. Client-treatment
matching offers potential for being able to enhance the effectiveness of therapy by identifying
not only what works but for whom it works (Paul, 1967). However, the primary focus of clienttreatment matching research has been on identifying which theoretical approach (e.g. behavioral
versus cognitive therapy) works best with individuals placed in specific diagnostic categories.
This oversimplifies the issue and loses sight of the numerous other client characteristics that may
be interacting with treatment to influence outcome. Of even greater concern is that these studies
have consistently lacked statistically significant results, even when great time, resources, and
methodological rigor have been employed to improve study outcome.
A common factors perspective has the potential to shed significant light on this issue.
Comparing the effectiveness of therapy models does not acknowledge those core ingredients of

1

change shared across models that are likely confounding results. As such, client-treatment
matching research needs to focus more on how to adapt those core ingredients to specific client
characteristics in order to improve effectiveness. In terms of what happens in therapy, many of
these core ingredients of change refer to treatment processes that influence the therapist-client
interaction (Greenberg & Pinsof, 1986) or common mechanisms that facilitate in-session change
events (Sprenkle & Blow, 2004).
Although a substantial body of treatment process research exists in psychotherapy, much
less research has addressed treatment process in couple therapy. However, one body of
theoretical and empirical work has identified enactments as a potential common change
mechanism for improving treatment process (Butler, Davis, & Seedall, 2008; Butler & Gardner,
2003; Butler & Wampler, 1999; Davis & Butler, 2004; Seedall & Butler, 2006). Enactments are
therapist-coached couple interaction with the primary goals of facilitating more productive
interaction patterns and overall couple self-reliance (Butler & Gardner, 2003; Butler & Wampler,
1999). Butler and Gardner (2003) developed a conceptual model that also demonstrates how to
adapt treatment process involving enactments to varying levels of couple emotionality,
reactivity, and overall readiness. This idea introduces an important issue of interactional
structuring that has rarely been discussed in couple therapy research or scholarly work but is
important to understanding and improving treatment process in the context of therapy with
multiple clients (Seedall & Butler, 2006).
In addition to helping structure couple interaction, enactments also offer the potential for
providing a different experience for couples. For example, emotionally focused therapy utilizes
enactments to deepen emotional experience and heighten a newly emergent way of emotional
relating for the couple (Johnson, 2004). Emotional experience has shown strong potential in

2

both psychotherapy and couple therapy as a treatment process that positively influences
outcomes. Emotions have been identified as a central organizing feature of individual (Frijda,
1986; Izard, 1991) and couple (Johnson & Greenberg, 1994) experience. Overall, emotions
inform meaning, motivation, and adaptation (Frijda, 1986; Izard, 1991).
Two essential aspects of emotion are processing and regulation. Processing refers to the
depth at which an individual can explore, cope with, and transform emotions (Klein, MathieuCoughlan, & Kiesler, 1986). Similarly, regulation refers to ―attempts individuals make to
influence which emotions they have, when they have them, and how these emotions are
experienced and expressed‖ (Rottenberg & Gross, 2007, p. 325). Individuals who struggle with
emotional processing typically cannot adequately regulate their emotional experience, thereby
over- or under-regulating their emotions and leading to disruptive emotional expression. As a
result, facilitating deeper emotional processing within psychotherapy offers great potential in
yielding more positive outcomes (Beutler et al., 2003; Bridges, 2006; Greenberg & PascualLeone, 2006; Wiser & Arnow, 2001). Within couple therapy, emotional processing and
regulation in some form has also been identified as an important change factor in several
empirically supported couple therapies, including emotionally focused therapy (Johnson, 2004),
enhanced cognitive behavioral therapy (Epstein & Baucom, 2002), behavioral marital therapy
(Jacobson & Margolin, 1979), and insight-oriented marital therapy (Snyder, 1999, 2002).
Interestingly, although clinical research has acknowledged the importance of
understanding treatment process as well as client-treatment matching, no research in either
psychotherapy or couple therapy has analyzed the interaction between treatment process and
client characteristics. Attachment theory represents one potential avenue for understanding how
to adapt treatment process to specific client characteristics and thereby enhance client experience

3

and improve outcomes. Attachment theory is a theory of close relationships, most often focusing
on early, formative relationships and the adaptive strategies infants learn regarding how to
respond when they are separated, threatened, or distressed (Bowlby, 1982). Findings have
emphasized the importance of the attachment figure‘s sensitivity and attunement in consistently
meeting the infant‘s needs (Ainsworth, Blehar, Waters, & Wall, 1978), thereby helping them to
develop a secure internal working model of self and others (Bowlby, 1973; Haft & Slade, 1989).
However, when their needs are not met on a regular basis, they develop coping strategies that
lead them to either shut down emotionally (deactivating) or demand that their needs be met in an
exaggerated and ultimately rejecting way (hyperactivating; Kobak, Cole, Ferenz-Gillies,
Fleming, & Gamble, 1993). These concepts closely parallel the ideas of under- and overregulation of emotion, but with an additional component explaining the reciprocal influence
between intrapersonal attachment strategies and close relationships.
More recently, attachment theory has been utilized to understand how relationship
models formed in infancy may be generalized to adulthood (Grossmann, Grossmann, & Waters,
2006; Mikulincer & Shaver, 2007; Sroufe, Egeland, Carlson, & Collins, 2005), with a variety of
informative findings complicated somewhat by the existence of two measurement techniques
with very little empirical overlap (Fortuna & Roisman, 2008). One approach, anchored in
developmental psychology, measures attachment according to a person‘s discourse or narrative
about childhood attachment experiences, while the other approach utilizes a self-report
measurement strategy typical of social and personality psychology.
Nonetheless, findings from a number of psychotherapy studies have shown important
associations between a person‘s general approach to attachment relationships (secure versus
insecure; hyperactivation versus deactivation) and two viable indicators of emotional experience:

4

physiological arousal and interpersonal distress. Physiological arousal refers to the unconscious
physical processes that accompany emotion and interpersonal distress refers to the influence that
emotions have on the conscious processes involved in relationships, including couples in a
committed romantic relationship, which are the focus of this study.
In terms of clinical work, psychotherapy research has found little evidence to support the
claim that therapy may be expected to consistently change attachment in terms of an individual‘s
global attachment category (Dozier, Manni, & Lindhiem, 2005). More evidence exists of the
potential for therapy to shift an individual‘s behavior towards increasing security during couple
interactions (Steele & Steele, 2008). Although this is an important area for research, of
potentially even greater importance is the question of how attachment influences a person‘s
experience in therapy and its overall outcome. A few articles have begun to address how selfreported attachment may moderate treatment outcome (McBride, Atkinson, Quilty, & Bagby,
2006; Meyer, Pilkonis, Proietti, Heape, & Egan, 2001) or how certain narrative/discourse
attachment characteristics may provide unique treatment challenges (Dozier, Lomax, Tyrrell, &
Lee, 2001; Korfmacher, Adam, Ogawa, & Egeland, 1997).
Although a relatively small but growing body of psychotherapy research has addressed
the role of attachment in adult couple relationships, no empirical work has been done in terms of
attachment and couple therapy, including how attachment models and strategies (secure versus
insecure; hyperactivating versus deactivating) influence core therapy processes such as
emotional experience. Emotionally focused therapy (EFT), an empirically-supported couple
therapy (Sexton et al., 2003), has effectively demonstrated how distressed relationships may be
conceptualized within an attachment framework. However, even though EFT has been shown to
be effective, results have not been able to demonstrate an empirical connection between its

5

underlying assumptions regarding attachment and couple therapy process and outcome.
Research is needed to draw more definitive conclusions about the role of attachment in
couple therapy and to begin identifying implications that therapists can use to adapt and improve
treatment process. For example, how might attachment influence an individual‘s experience in
couple therapy? Those clients who have consistently had their attachment needs met in
relationships (characteristic of secure attachment) may be more able to express themselves, ask
for their needs to be met, and then to reach out to their partner in therapy.
Conversely, individuals who have not experienced consistent responses in the past when
they were distressed may have greater difficulty in therapy. When attachment needs or threats
have been inconsistently met by attachment figures (characteristic of preoccupied/anxious
attachment), it may be difficult for individuals to appropriately request that their needs be met.
They may also struggle to regulate their attempts to meet their partners‘ needs. Those who have
been regularly rejected (characteristic of dismissing/avoidant attachment) may struggle to open
up emotionally, thereby making it difficult for them to request that their needs be met or to reach
out to their partner. These issues are made more complex by the varying attachment strategies
and expectations that partners may have for their close relationships.
As a result, if couple therapy is generally tailored toward distressed clients with secure
attachment strategies, then those clients with less secure strategies may not be receiving optimal
benefits from treatment. Research is needed to understand the attachment-related experiences of
clients in a therapeutic context so that treatment process may then be adapted to their needs and
experiences. The attachment strategies of clients potentially interact with specific elements of
treatment process to influence the emotional experience of clients. This study adapted
interpersonal process recall (Elliott, 1986; Kagan & Kagan, 1990; Kagan, Krathwohl, & Miller,

6

1963) to analyze how attachment within two contrasting interactional conditions (low structure,
semi-natural and high-structure, therapy-like) influences the emotional experience of couples in
terms of their physiological arousal and interpersonal distress.

7

CHAPTER II
LITERATURE REVIEW
Effectiveness Research in Couple Therapy
In addition to examining the effectiveness of individual psychotherapy (see Lambert &
Ogles, 2003), meta-analyses have studied the effectiveness of couple treatment.
Overwhelmingly, both behavioral and nonbehavioral couple approaches have been found more
effective than no-treatment control groups (Alexander, Holtzworth-Munroe, & Jameson, 1994;
Christensen & Heavey, 1999; Sexton et al., 2003). In fact, the most recent meta-analysis found
that, on average, 80% of distressed couples receiving treatment showed more improvement than
distressed couples in no-treatment control groups (overall mean effect size = .84; Shadish &
Baldwin, 2003).
However, Sexton et al. (2003) caution, ―Despite what look to be large effect size
differences, the clinical significance of couple therapy remains in question . . . . Even among the
most efficacious treatments, fewer than half of couples treated have clinically significant positive
outcomes‖ (p. 598; see also Christensen & Heavey, 1999, and Jacobson & Addis, 1993, for
similar conclusions). Even as effectiveness and efficacy research has continued and therapeutic
approaches such as emotionally focused therapy (EFT) have become evidence-based, the overall
level of clinical effectiveness has not improved since Jacobson and Addis (1993) first addressed
the issue by challenging couple researchers to address the following questions: ―Who benefits
from couple therapy? What are the active ingredients of successful couple therapy? What are
the pertinent change processes?‖ (pp. 86-87).
Couples who do not achieve clinically significant outcomes can be classified five ways:
(a) those whose condition worsens as a result of therapy; (b) those who remain unchanged; (c)

8

those who demonstrate improvement but do not move into the non-distressed range; (d) those
who do not maintain their gains at follow-up (i.e., relapse); or (e) those who drop out before
change can be measured or achieved. Thus, efforts are needed not only to improve the
effectiveness of therapy, but also the engagement of clients so that they remain in therapy to
receive the full potential benefits.
Dropping out of treatment, also known as attrition, premature termination, or unilateral
termination, is an area of great concern for psychotherapists. Most often, researchers have
defined dropout using one or a combination of the following methods: (a) failure to complete a
specified number of treatment sessions; (b) therapist rating the level of treatment completion;
and (c) discontinuation of therapy after the intake session (Masi, Miller, & Olson, 2003;
Wierzbicki & Pekarik, 1993). One meta-analysis of 125 psychotherapy studies using varying
definitions of dropout found an average dropout rate of 46.9% (Wierzbicki & Pekarik, 1993).
The mean dropout rate for individual therapy (47.5%) was comparable to the rates found for
studies of group, family, and/or couple therapy (45.7%).
Similarly, a study of 463 individuals in a university-based marriage and family therapy
(MFT) clinic found no significant differences in dropout rates between individual, couple, or
family therapy across any of the aforementioned dropout conditions (Masi et al., 2003).
However, a more recent study with 434,317 patients in an administrative health care system
(CIGNA) found that, although dropout, defined as discontinuing treatment after one session, was
less likely when the provider was an MFT as compared to other mental health professions,
patients were 33.2% more likely to dropout when participating in relational therapy than when
attending individual therapy (Hamilton, Moore, Crane, & Payne, 2010). Regardless, beyond an
initial understanding of dropout rates in couple or family therapy and how they compare to

9

individual therapy, relatively little is known theoretically or empirically, especially when
considering the much larger number of studies in general psychotherapy that have addressed the
topic of dropout (Bartle-Haring, Glebova, & Meyer, 2007; Bischoff & Sprenkle, 1993). For
example, it is not known how much of couple therapy dropout can be attributed to factors such as
initial client or couple distress, less effective treatment, or the lack of fit regarding the therapist‘s
approach to client characteristics (e.g., attachment) that might lead to difficulties such as
disagreement regarding the depth and breadth of change needed (Bischoff & Sprenkle, 1993;
Helmeke, Bischof, & Sori, 2002).
One review of couple research identified the dimensions involved in predicting therapy
dropout as client- and therapist-related characteristics and contextual factors, elements of
therapeutic process, and the techniques or interventions used in treatment (Bischoff & Sprenkle,
1993). Client-related factors associated with higher rates of dropout in couple therapy were
lower socio-economic status (Williams, Ketring, & Salts, 2005); therapist-client ethnic mismatch
(Williams et al., 2005); a larger number of children in the family (Allgood & Crane, 1991);
lower husband life happiness (Anderson, Atilano, Bergen, Russell, & Jurich, 1985); higher
husband anxiety (Allgood & Crane, 1991); and the attribution of blame on one individual for
relational problems (Allgood & Crane, 1991). These couple therapy findings mirror results from
psychotherapy in general that have found associations between outcome and a client‘s
interpersonal relatedness, psychological resources, and level of psychopathology (Clarkin &
Levy, 2003).
Therapist-related factors associated with higher dropout rates were being a man (Allgood
& Crane, 1991); lower levels of therapeutic rapport (Tryon & Kane, 1993); low therapist activity
(Anderson et al., 1985); high levels of therapist interpretation and talking (Hollis, 1968); less

10

treatment follow-up (Hollis, 1968); and therapist predictions of client discontinuation (Allgood
& Crane, 1991). Therapy process and intervention variables are much rarer in the research
literature, but there have been indications from both individual and couple therapy research that
the risk of dropout originates primarily during the first two sessions (Hollis, 1968; Phillips,
1985), and might decrease by focusing on understanding and meeting client expectations early.
Nonetheless, additional research needs to expand upon and replicate these preliminary findings
regarding dropout predictors.
In addition to the relatively small literature addressing dropout, no empirical research has
analyzed the potential for deterioration in couple therapy. Nonetheless, with approximately 20%
of participants in couple therapy meta-analyses no better off than no-treatment controls, client
decline is a very salient risk. If results parallel psychotherapy research reviews, approximately
5-10% of couples will actually get worse while in treatment (Lambert & Ogles, 2003). This is a
sobering statistic, especially considering that outcomes from routine clinical practice may be less
effective than those conducted in highly-controlled clinical trials due to the additional time,
energy, and resources often expended in empirical studies (Lambert & Ogles, 2003).
Increasing the Effectiveness of Couple Therapy through Client-Treatment Matching
Over 40 years ago, Paul (1967) encouraged researchers to focus beyond treatment
effectiveness to discover ―what treatment, by whom, is most effective for this individual with
that specific problem under which set of circumstances‖ (p. 111). For this to happen, increased
attention must be given to research that provides an in-depth look at the experiences of those at
risk for sub-optimal results from couple therapy. This includes understanding how their
experiences differ across client, therapist, intervention, and therapy-process factors from those
who demonstrate clinically significant outcomes. With respect to the importance of client

11

factors, Lambert and Ogles (2003) explained that ―carefully matching techniques to client
dispositions, personality traits, and other diagnostic differences‖ (p. 180) may ultimately be vital
to enhancing retention and expanding clinically significant outcomes. This section will
summarize the work that has been done on matching treatment to client characteristics.
Matching Type of Treatment to Client Characteristics
Client-treatment matching, also known as aptitude by treatment interaction (ATI) and
systematic treatment selection (STS) has been researched most commonly in the fields of
psychotherapy in general and substance abuse treatment. The breadth of attention in this area
has been substantial, with entire books (Nathan & Gorman, 2002; Roth & Fonagy, 2005) and
research projects (Elkin et al., 1989; Project MATCH Research Group, 1997a, 1997b, 1998)
dedicated to this subject. The most prevalent research approach to client-treatment matching has
focused on identifying the most effective treatment model in treating a specific diagnosis, with
some studies also looking at how additional intrapersonal characteristics mediate this
relationship. For example, one study within the Treatment for Depression Collaborative
Research Program (TDCRP) compared interpersonal therapy (IPT) and cognitive therapy (CT),
finding that IPT reduced depression more effectively for individuals who exhibited higher levels
of obsessiveness, while CT was more effective in treating depression for those higher in
avoidance (Barber & Muenz, 1996).
Perhaps the most expansive, methodologically sophisticated, and well-known study that
analyzed client-treatment interaction is Project MATCH, conducted at multiple sites by the
National Institute on Alcohol Abuse and Alcoholism (NIAAA; Project MATCH Research
Group, 1997a, 1997b, 1998). Project MATCH analyzed data from 952 alcohol-dependent
individuals in outpatient therapy and 774 participants in aftercare therapy who were randomly

12

assigned to one of three treatment conditions: cognitive behavioral coping skills therapy (12
sessions), motivational enhancement therapy (4 sessions), or 12-step facilitation therapy (12
sessions). Participants in all three modalities demonstrated statistically significant improvement
on a variety of drinking and general outcome measures. Client characteristics compared across
treatments were severity of alcohol use, cognitive impairment, client conceptual level, gender,
meaning seeking, motivational readiness to change, psychiatric severity, social support for
drinking versus abstinence, sociopathy, and typology (Project MATCH Research Group, 1997a).
Of those three modalities, only psychiatric severity was found to interact significantly
with treatment type during the full year following treatment (Project MATCH Research Group,
1997b). These disappointing results led to speculation that either client-treatment matching is
not a salient aspect of therapy or it is a much more complex and dynamic issue (Project MATCH
Research Group, 1998). Since then, research and scholarly work have sought to identify
potential methodological issues that limited the Project MATCH results (Buhringer, 2005; Lakey
& Ondersma, 2008; Project MATCH Research Group, 1998) and have continued testing clienttreatment interaction effects (Rychtarik et al., 2000; UKATT research team, 2008). Regardless,
recent results have done little to empirically substantiate the significance of client-treatment
matching, in spite of the overwhelming theoretical expectation of its importance.
Couple therapy adds additional complexities to understanding how to effectively match
treatment process to client characteristics. Not only are there two client-treatment interactions in
couple therapy, but there is also the relationship between both clients that interacts with
treatment. Unfortunately, couple therapy research has employed virtually the same method as
psychotherapy and substance abuse treatment research in attempting to demonstrate the
effectiveness of one approach over another in treating a specific diagnosis (see Snyder &

13

Whisman, 2003, and Sprenkle, 2002, for in-depth examples). In fact, no empirical examples
exist in couple therapy of matching treatment to client characteristics other than client diagnosis.
Although not an explicit example of client-treatment matching, Jacobson and Christensen (1996)
did find that incorporating an acceptance component into an integrative behavioral treatment
protocol for at-risk couples enhanced therapy effectiveness by adapting treatment to client needs.
Nonetheless, more work is needed to understand how the role of client-treatment matching in
couple therapy may improve therapy outcome.
A Common Factors Approach to Client-Treatment Matching
Although a number of justifications for the relatively small effects of client-treatment
matching have been given, another potentially viable explanation for the consistent lack of
significant findings may be derived from a common factors perspective. A common factors
approach to therapy was developed in partial response to meta-analytic findings that very few
substantive differences exist between treatment models in terms of achieved outcome when
controlling for confounding variables (Hubble, Duncan, & Miller, 1999; Smith, Glass, & Miller,
1980; Wampold, 2001). In fact, one meta-analysis suggested that techniques associated with
specific models accounted for only 8% of therapeutic outcome variance (Wampold, 2001).
Thus, a common factors framework has sought to shift the focus from differentiating between
models by focusing on identifying active ingredients, mechanisms, and processes of change that
are shared by effective therapeutic approaches as well as those intrapersonal and interpersonal
characteristics shared by effective therapists (Blow, Sprenkle, & Davis, 2007; Frank & Frank,
1991; Lambert, 1992).
As a result, a common factors perspective may yield crucial understanding as to why
existing client-treatment matching research has yielded few statistically significant results.

14

Interestingly, current client-treatment matching research may be considered an outgrowth of
efficacy/effectiveness research. The methodology employed by existing client-treatment
matching research focuses primarily on identifying which model most effectively treats a
specific diagnosis, but with the added dimension of looking at how additional client
characteristics influence the model-diagnosis interaction. Although this component does provide
further information about how treatment in general might be matched to client characteristics, it
oversimplifies the issue by assuming that models differ solely according to the types of clients
for whom they are effective. The lack of significant differences in outcome between models then
is attributed to unmeasured client characteristics that are confounding the results.
However, the consistent non-significant findings in client-treatment matching research
indicates that this may not be the case. Instead, the very models being tested actually may be
confounding results. From a common factors standpoint, these non-significant results point to
the fact that the tested models—despite being operationally distinct—share similar change
ingredients, mechanisms, and/or processes that lead to positive outcomes. The disappointing
findings regarding the interaction between treatment type and client characteristics led the
Project MATCH Research Group to a similar conclusion that future research needed to ―identify
the common and unique active ingredients of treatments, as well as provide a better
understanding of how these treatment variables lead to differential client outcomes‖ (Project
MATCH Research Group, 1997b, p. 1695). Thus, client-treatment matching research is needed
that goes beyond what models are effective to discovering how and for whom they are most
effective by investigating the interaction between treatment process and client characteristics
(see Figure 1).

15

Figure 1.

Process Research from a Common Factor Perspective

Client
Experience
and/or
Treatment
Outcomes

Matching Treatment Process to Client Characteristics
The concept of process has been defined in a variety of ways, but generally refers to how
change occurs. That being said, one potentially valuable distinction that needs to be made is the
difference between change process and treatment/therapy process (Orlinsky, Ronnestad, &
Willutzki, 2003). For purposes of simplicity, this study defines change process as those
intrapersonal and interpersonal factors that contribute to change in a person‘s life both inside and
outside of therapy. One important element of change process—especially in terms of therapy
and its effectiveness—is treatment process, which is defined here as those technical or
interpersonal conditions occurring in the context of therapy that lead to positive outcomes and
overall change. In this manner, treatment process is conceptualized here as consisting of (a) the
nature of the interaction between the therapist and client systems (Greenberg & Pinsof, 1986),
and (b) those core mechanisms that facilitate change events within the framework of therapy
(Sprenkle & Blow, 2004).
The relationship between therapy process and outcome has been a much-studied topic in

16

psychotherapy research. In a comprehensive review of hundreds of articles, Orlinsky et al.
(2003) found that some of the most robust research findings included a positive therapist-client
relationship, characterized by collaboration, communicative rapport, empathy, and mutual
affirmation. Therapist skillfulness in facilitating positive affect, expressiveness, emotional
processing, cooperation, and openness in clients, as opposed to increasing resistance and
defensiveness, were also associated with more positive outcomes (Orlinsky et al., 2003; see also
Beutler et al., 2003). In terms of specific techniques, a number of studies consistently found that
the ―timely and skillful application of potent interventions such as experiential confrontation,
interpretation, and paradoxical intention‖ (Orlinsky et al., 2003, p. 324) were also found to be
consistently effective.
In terms of couple therapy, much less treatment process research has been conducted.
Many of those studies that do exist involve couple change process (i.e., dynamics between the
couple that influence outcome) rather than treatment process, as defined above. Those that do
address treatment process independent of theoretical orientation have yielded results similar to
those from psychotherapy research in general. Couple therapy studies found that a strong
therapeutic alliance was associated with higher rates of retention (Raytek, McCrady, Epstein, &
Hirsch, 1999) and outcome (Quinn, Dotson, & Jordan, 1997). In addition, therapists were
expected to be active and directive (Estrada & Holmes, 1999) while demonstrating competence
(Raytek et al., 1999) and empathy (Bischoff & McBride, 1996) during the therapy process.
In terms of general couple interventions, one study found that clients perceived change
according to affective, cognitive, and communicative dimensions (Christensen, Russell, Miller,
& Peterson, 1998). Butler and Wampler (1999) also found that therapist behaviors that
facilitated couple-responsible process such as enactments, accommodation, and inductive

17

process were associated with higher levels of cooperation and lower levels of therapist-client
struggle, two conditions associated with positive outcomes. Enactments have been defined as
―therapist-facilitated, couple-focused interaction‖ (Butler & Wampler, 1999, p. 30) that adapts
therapist role, involvement, and interactional structure to varying levels of couple reactivity and
volatility (Butler & Gardner, 2003; Seedall & Butler, 2006). Enactments have been suggested as
a potential core change mechanism by virtue of their utility across theoretical orientations and
applicability for a variety of couple issues (Butler et al., 2008).
A number of studies specific to EFT have also sought to identify treatment processes that
influence outcome. These studies have found that a strong therapeutic bond (Johnson &
Talitman, 1997), emotional experiencing (Bradley & Furrow, 2004; Johnson & Greenberg,
1988), and positive emotional responses (Bradley & Furrow, 2004; Greenberg, James, & Conry,
1988; Johnson & Greenberg, 1988) are related strongly to perceived change in EFT and overall
success in achieving positive outcomes. In this manner, the process of EFT intends to allow
couples to process their emotions, express their vulnerable feelings and needs, and gain mutual
understanding while receiving support and validation from one another and the therapist. These
processes are intended to help partners soften towards one another and thereby restore trust and
repair past relationship traumas.
Despite yielding some important information about therapeutic change, none of these
studies investigated how treatment process might be adapted to varying client characteristics to
maximize outcome. Yet client-treatment matching from this perspective may be vital to
improving therapy outcomes. One qualitative study of three couples over 23 sessions noted that,
―Pivotal moments tended to be highly individualized accounts, with little overlap between
spouses and little overlap between therapist and client‖ (Helmeke & Sprenkle, 2000, p. 469).

18

Thus, because intrapersonal characteristics often vary between partners within the couple
relationship, it is even more imperative for empirical work in couple therapy to examine how to
systematically adapt therapy to specific intrapersonal and interpersonal client characteristics in
order to maximize positive therapeutic outcomes.
In their evaluation of current process research in couple therapy, Heatherington,
Friedlander, and Greenberg (2005) comment that much of research has focused on what clients
do in therapy through the use of observational coding. This understanding of interpersonal
behavior is an important component of understanding effective treatment. However,
Heatherington et al. (2005) also emphasize the need to pay attention to clients‘ intrapersonal
experiences in therapy. In other words, what is therapy really like for clients? This study
represents a preliminary effort to understand this by integrating client-treatment matching
concepts with process-outcome research. Specifically, it investigated how attachment as a
potential match variable influenced a person‘s emotional experience during low structure, seminaturalistic and high structure, therapy-like couple interactions.
Emotions in Couple Relationships and Therapy
The previous section addressed the importance of treatment process research in
identifying those factors that contribute to therapeutic change. This section discusses the role of
emotion in couple relationships and therapy. Emotional experience in some form (e.g., empathy,
positive emotions, emotional processing) has been identified consistently as a core therapeutic
process variable. More generally, emotion has been described as having semantic, motivational,
and adaptive functions (Frijda, 1986; Izard, 1991) in being ―involved with organizing, guiding,
and driving perception, thought, and action‖ (Izard, 1991, p. 51). In this manner, emotion may
be considered a driving force for many intrapersonal and interpersonal processes. These ideas

19

have led Johnson and Greenberg (1994) to the conclusion that emotion is the ―primary building
block of adult intimate relationships‖ (p. 10).
Because emotion is so instrumental in providing intra- and interpersonal meaning and
organizing how an individual responds to a situation, the processing of felt emotion becomes a
vital adaptive skill that allows for the appropriate regulation of emotion. During periods of
arousal, emotional processing allows individuals to focus on, cope with, and potentially expand
and/or transform emotional experience and expression (Klein et al., 1986). However, when
individuals cannot adequately cope with their emotions or the arousal that they might produce,
they often attempt to over- or under-regulate them, leading them towards more negative,
relationship-disrupting expressions of affect.
In terms of couple relationships, Gottman (1999) identified several of these negative
emotional responses that appear related to later relationship dissolution: criticism, contempt,
defensiveness, and stonewalling. More generally, Gottman and Levenson (2002) found that,
over a 14-year period, couples who divorced early demonstrated high expressivity, anger, and
wife negativity (i.e., under-regulation). Those who divorced later during the longitudinal study
were more likely to demonstrate prolonged periods of neutral affect during conflict characterized
by physiological evidence of emotional suppression (i.e., over-regulation). Regardless, it
appears that a person‘s emotional experience and expressions, measured by physiological means
and observed couple interaction, have strong implications for understanding couple relationships.
In addition to the relationship consequences, emotional suppression has been linked to greater
cardiovascular health risks, especially if it is part of a consistent relationship pattern (Mauss &
Gross, 2004).
In terms of psychotherapy, emotional processing and regulation have been described as

20

key factors that influence the facilitation of therapeutic change across models (Bridges, 2006;
Greenberg & Pascual-Leone, 2006). A deeper awareness and processing of emotions allows for
personal meaning to be explored (see also Wiser & Arnow, 2001) and impels clients to work for
change (see also Linehan, 1993). Beutler et al. (2003) cite several major reviews and conclude
that ―emotional arousal will increase positive outcomes, especially if applied early in the
treatment‖ (p. 263). However, Greenberg and Pascual-Leone (2006) point out that some
incorrectly equate emotional arousal with felt intensity instead of representing emotional
experience or the depth of emotional processing. This level of processing is achieved when, in
the context of an empathic and validating therapeutic relationship, the therapist helps clients to
create new meaning by experiencing, exploring, and restructuring their emotions (Greenberg &
Paivio, 2003).
Although the focus on emotions in couple therapy has steadily increased in recent years
across therapeutic approaches, the most salient example of utilizing emotions to facilitate change
in couple therapy is emotionally focused therapy (EFT), a systemic and experiential approach
that helps clients ―reprocess and reformulate their emotional responses‖ in the context of their
couple relationship and ultimately establish more secure emotional bonds (Johnson, 2003b, p.
16). This process centers on healing from an attachment trauma, where one partner reached out
to the other during a time of distress and did not have his/her needs met. This often leads to a
pattern similar to pursue-withdraw, where one partner demonstrates high levels of emotionality
and negativity and the other partner consistently delays and separates from the partner during
couple interaction. In this manner, part of treatment in EFT involves engaging the withdrawing
partner while also softening the pursuing/blaming partner (Bradley & Furrow, 2004).
One of the unique contributions of EFT is that it emphasizes the idea that emotional

21

processes within couple relationships are related to attachment and that understanding couple
relationships from an attachment perspective can inform clinical practice. However, EFT stops
short of conceptualizing the reciprocal relationship between emotional experience inside and
outside of therapy and a person‘s attachment model of self and other. The following sections
will provide an overview of attachment theory and its potential as an important client
characteristic in client-treatment matching in couple therapy. This will include how relationships
with caregivers as early as infancy contribute to how individuals view their adult close
relationships as well as how emotions are processed, regulated, and ultimately expressed within
those relationships.
An Overview of Attachment Theory and Close Relationships
The previous section explained how empirical research has shown that emotional
experience is an important aspect of treatment process and contributes to outcome. However,
little research has looked at how client characteristics (e.g., attachment) might influence
emotional experience. This section will provide an overview of attachment theory, its
relationship with emotional experience, and existing research that points towards attachment
being a potentially important match variable. Originating in the 1940‘s, attachment theory
emphasizes the innate human need to establish non-interchangeable, long-term relationships with
a few individuals, beginning during the first months of life (Ainsworth et al., 1978; Bowlby,
1982). Bowlby utilized an integrative, multi-disciplinary approach to conceptually validate the
idea that even infants‘ earliest experiences with primary caregivers instruct them about whom
―attachment figures are, where they may be found, and how they may be expected to respond‖
(Bowlby, 1973, p. 208).
Attachment figures serve as potential secure bases from which individuals explore the

22

environment and safe havens from which they seek comfort when feeling threatened or
distressed (Bowlby, 1973; Haft & Slade, 1989). Unexplained separation from these attachment
figures yields distress, while reunion brings relief. These interactions with attachment figures
contribute to an internal working model (IWM; Bowlby, 1973) of self and other that regulates a
person‘s emotional response by indicating (a) the level of safety and comfort found in close
relationships when confronted with separation, threat, or distress, (b) the probable availability
and responsiveness of attachment figures in attending to the person‘s distress, and (c) the
individual‘s own worthiness to receive sensitive, loving care from attachment figures.
Ainsworth, after working with Bowlby early in her career, developed a strong empirical
base for Bowlby‘s theory based upon meticulous observation of naturalistic caregiver-child
interaction. Because of Ainsworth‘s methodological rigor, she and her colleagues (Ainsworth et
al., 1978) developed a laboratory procedure, the Strange Situation (SS), during which they
identified distinct infant behavior patterns that occurred at various points of separation and
reunion. The SS is a 20 minute, eight-episode experiential paradigm designed to promote
attachment-seeking behaviors in the infant (Ainsworth & Bell, 1970). Infants who expressed
their needs for closeness and connection directly and were calmed easily (secure) were more
likely to have attuned, responsive mothers. Those who actively tried to ignore the caregiver
(avoidant) were more likely to have had their expression of needs responded to with rejection,
and those who expressed a form of direct anger while simultaneously seeking closeness
(ambivalent) were more likely to receive inconsistent responses to their expressions of need.
Later, another classification (disorganized) was added to describe those infants who
demonstrated odd or disorganized behaviors in the SS (Main & Hesse, 1990).

23

Dozier et al. (2001) offer a useful conceptualization of these varying attachment states of
mind: ―When secure or autonomous strategies fail to produce desired outcomes, alternative
secondary strategies are adopted‖ (p. 63). These secondary strategies exist on a continuum of
deactivation-hyperactivation of the attachment behavioral system (Kobak et al., 1993). When
infants‘ signals for closeness and safety from a caregiver during moments of distress have been
met consistently with rejection, they respond by shutting down or deactivating emotionally when
distressed or threatened. Others whose caregivers provided inconsistent responses during
distressing moments—consisting of a mix of loving, rejecting, and enmeshed caregiver
reactions—become hyper-aroused when distressed in an effort to maximize the chance for a
loving response. This hyperarousal is often an exaggerated solicitation of love and comfort that
often pushes away and rejects the attachment figure. Disorganized attachment, typically
manifested by anomalous oscillations between hyperactivating and deactivating infant behavior,
has been linked to frightened, threatening, or dissociative actions by the attachment figure (Hesse
& Main, 2006). In this manner, the attachment figure is the primary source of the infant‘s
distress rather than a secure base and safe haven from which to explore or seek comfort.
Attachment in Adulthood
Having briefly explained how a person‘s attachment model forms in infancy, this section
summarizes how attachment has been utilized to understand attachment relationships in
adulthood. The original focus on the formation of infant attachment in the early years of life and
its relationship to caregiver behavior generated a great deal of interest in attachment at other
developmental periods across the lifespan, including childhood and adolescence (Parkes,
Stevenson-Hinde, & Marris, 1991). The focus on the influence of attachment shifted onto
adulthood as researchers began to study parents‘ attachment experiences in addition to those of

24

their children. This began with conceptualizing adult bereavement (Bowlby & Parkes, 1970;
Parkes, 1972) and marital separation (Weiss, 1975) from an attachment perspective. Later, the
focus shifted to the functioning of marital/sexual relationships (Ainsworth, 1991; Weiss, 1991),
including how adult attachment patterns and representations (George, Kaplan, & Main, 2002)
influence their romantic relationships (Hazan & Shaver, 1987).
Adult Attachment: Two Alternative Measurement Approaches
Two distinct empirical traditions of studying attachment in adulthood have developed
with divergent strategies in conceptualizing and measuring attachment that yield substantially
different results (Crowell, Treboux, & Waters, 1999; Fortuna & Roisman, 2008; Roisman et al.,
2007). One approach is rooted in social and personality psychology and based primarily on selfreport; the other utilizes narrative/discourse measurement based on the Adult Attachment
Interview (George et al., 2002) which is more typical of developmental psychology. A social or
personality psychology approach operates from the belief that attachment can be measured from
individuals‘ conscious appraisals and attributions about themselves and their relationships
(Crowell, Fraley, & Shaver, 1999). In contrast, a developmental perspective presumes that the
expression of attachment is an unconscious process approximated through specific
representations occurring during the process of an interview or interpersonal interaction
(Roisman et al., 2007).
There are also conceptual differences between these two approaches to measuring adult
attachment (Crowell, Treboux, & Waters, 1999), with each emphasizing somewhat different
elements of both Bowlby‘s and Ainsworth‘s work. In addition, empirical studies have found, at
best, modest to moderate statistical associations (Bartholomew & Shaver, 1998; Shaver, Belsky,
& Brennan, 2000). In perhaps the most comprehensive analysis of the relationship between self-

25

report and narrative/discourse categorical dimensions of attachment to date, Roisman et al.
(2007) conducted a meta-analysis that found only trivial to small associations (mean r = .09,
meta-analytic N = 961). They also assessed for similarities between the two approaches across
personality traits and adult relationship functioning, again yielding few commonalities. Fortuna
and Roisman (2008) analyzed the relationship between the two adult attachment measurement
traditions across reports of psychiatric symptomatology and life stress. Results again showed
little overlap between the two approaches, providing additional evidence that these attachment
measures ―seem to tap different aspects of ‗security‘‖ (Fortuna & Roisman, 2008, p. 12).
A variety of potential explanations exist for these dissimilarities, many of which target
data that support the efficacy of the author‘s preferred approach (Fraley, 2002). Beyond the
debate over validity, there appear to be some legitimate differences between approaches in terms
of the type of relationship between attachment insecurity and variations in an individual‘s
psychopathology or close relationships. Recent research has found divergent implications of
attachment insecurity as measured by the two approaches in terms of psychopathology and
interactional quality. The primary difference rests in whether attachment insecurity represents a
general risk for problems in these areas (risk model) or if problems arise only under conditions of
high stress (diathesis-stress model; Fortuna & Roisman, 2008; Roisman et al., 2007). In other
words, the issue is not whether a link exists but how stress may moderate these relationships
differently for each measurement approach.
Thus far, no consistent pattern of results and implications for couple relationships has
emerged for either approach. Although self-report attachment security was related to
relationship quality when couple interactions were perceived as stressful (diathesis-stress model;
Roisman et al., 2007), self-report attachment insecurity was associated with psychopathology

26

under both high and low life stress (risk model; Fortuna & Roisman, 2008). Conversely, security
as measured by the Adult Attachment Interview (AAI) was associated with higher levels of
relationship quality under conditions of high and low life stress (risk model; Roisman et al.,
2007), while AAI insecurity was associated with psychopathology only under conditions of high
life stress (diathesis-stress model; Fortuna & Roisman, 2008).
Regardless of the reasons behind the differences between approaches, the trivial to small
overlap between them necessitates that clear distinctions be made when discussing the literature
and implications of research using each measurement approach. Too often generalizations are
incorrectly made across both approaches, as if the measurement approaches were equivalent (see
Johnson, 2003a, for an example). Because literature from each is relevant to the current
research, attachment will be discussed using both self-report and narrative/discourse measures.
The sections that follow summarize attachment-related measurement issues, relevant
intrapersonal and interpersonal findings, and clinical implications separately for each approach.
Narrative/Discourse Attachment Measurement
The Adult Attachment Interview (AAI; George et al., 2002) has become the primary
narrative/discourse method for measuring adult attachment. Developed by Mary Main, a student
of Ainsworth, the AAI is a semi-structured interview that assesses a person‘s current ―state of
mind with respect to attachment‖ (Hesse, 2008, p. 554). Consisting of 20 open-ended questions
with a variety of probes, the AAI asks respondents about their early childhood relationships with
their parents, including any issues of rejection, abuse, loss, and trauma. Coding the interview is
based upon the concept of ―narrative coherence‖ regarding childhood experiences that has been
linked to a secure state of mind related to attachment and is based upon criteria Grice (1975)
established for collaborative discourse: quality (consistent information and proper evidence);

27

quantity (appropriate length of talk-turns); relation (relevant speech); and manner (clear and
orderly speech).
Classification labels for the AAI (secure-F, dismissing-Ds, preoccupied-E, and
unresolved-U) are different from those assigned in the Strange Situation (secure-B, avoidant-A,
resistant/ambivalent-C, and disorganized-Ds), but they are conceptually parallel to one another in
terms of how individuals view attachment relationships and respond to distress. Secure
individuals—as measured by the AAI—present a coherent and balanced narrative of their
experiences characterized by an autonomous and collaborative way of relating (Hesse, 2008).
Dismissing individuals typically present an incomplete or inconsistent picture of their early
childhood characterized by extreme idealization or derogation of attachment figures and
experiences. They may also appear to block discourse by insisting on a lack of memory.
Individuals assigned the preoccupied classification on the AAI demonstrate involving anger
towards the attachment figure or passivity/vagueness of discourse. Both forms of preoccupation
are characterized by a sense of incoherence, irrelevance, and overall confusion. Unresolved
attachment is assigned when a person‘s monitoring of discourse or reasoning abruptly changes
during discussions of loss or abuse (Hesse & Main, 2006).
The Adult Attachment Interview (AAI) and the Strange Situation (SS). One of the most
important elements of the AAI is its link with SS findings. For example, an infant‘s avoidant
classification upon reunion in the SS is strongly related to his/her caregiver‘s dismissing state of
mind as measured in the AAI. A meta-analysis of 14 studies found a 75% correspondence
between two category (secure/insecure) parent-infant attachment and a 70% three category
match (ambivalent/preoccupied, avoidant/dismissing; secure/secure; van IJzendoorn, 1995).
Establishing an association between the AAI and the SS was important for questions of validity,

28

since SS findings were based upon naturalistic observations of the actual infant-caregiver
relationship. In this manner, the AAI has been able to predict differences in how parents behave
towards their offspring and how their infants respond.
To control for potential confounds, including infant characteristics that might influence
parenting behavior (e.g. temperament), several studies administered the AAI to mothers prior to
birth and compared those results to infant SS classifications at one year (Benoit & Parker, 1994;
Fonagy, Steele, & Steele, 1991; Ward & Carlson, 1995). Each of these studies yielded similar
results to those of the meta-analysis, finding at least a 75% two-group (secure/insecure) match.
Two of the studies (Benoit & Parker, 1994; Ward & Carlson, 1995) actually showed a 4-group
match between SS and AAI classifications of 68% (secure/secure, avoidant/dismissing,
ambivalent/preoccupied, and disorganized/unresolved). Interestingly, Benoit and Parker (1994)
also gathered information across three generations, with a grandmother-infant match of 75%
across three categories and 49% across four. Thus, there is a much stronger than chance
association between mother-infant attachment classifications.
Self-Report Attachment Measurement
Self-report attachment is an outgrowth of the effort to understand and effectively measure
adult romantic attachment from a social and personality psychology perspective. In particular,
the self-report tradition sought to explore ―the possibility that romantic love is an attachment
process‖ (Hazan & Shaver, 1987, p. 511). There are a variety of categorical or continuous selfreport measures from which to choose which represent the progression of attachment
measurement in the field of social and personality psychology. However, developing a reliable
and valid self-report measure of attachment has proven to be somewhat of a challenge. Whereas
substantial attention and effort have been given to anchor attachment as measured by the AAI

29

with Strange Situation results, no research as yet has sought to demonstrate a relationship
between self-report attachment and the Strange Situation. This fact, coupled with the evidence
of very little empirical overlap between AAI and self-report attachment results, has made it
difficult to definitively measure attachment via self-report methods, even though those methods
appear to conceptually parallel to Ainsworth et al.‘s (1978) coding and subscales. The remainder
of this section outlines the historical progression of self-report attachment instruments while also
summarizing the progress that has been made in this area.
In order to begin understanding the potential role of self-report attachment in adulthood,
Hazan and Shaver (1987) established a categorical, forced-choice measure with three paragraphs
based upon Strange Situation attachment categories. Responses ranged from relative ease in
forming and comfort with relationships (secure), to discomfort with closeness and difficulty
trusting others (avoidant), to fears that others would not get as emotionally close as desired and
fears of abandonment (anxious/ambivalent). Hazan and Shaver‘s (1987) initial study found
associations between attachment-style categories and love experiences, working models of self
and relationships, reports about childhood experiences with parents, and a measure of trait
loneliness.
Hazan and Shaver‘s (1987) measure was utilized in a number of additional studies before
continuous self-report measures were developed to enable measurement of the variation within
attachment categories, increase reliability, and decrease measurement error (Mikulincer &
Shaver, 2007). Two continuous measures were independently developed in the early 1990‘s to
address some of these issues by deconstructing Hazan and Shaver‘s measure into separate items.
The Adult Attachment Questionnaire (AAQ; Simpson, 1990) consists of 13 items and two
factors, attachment anxiety and avoidance, while the Adult Attachment Scale (AAS; Collins &

30

Read, 1990) is an 18-item scale that assesses respondent beliefs on three factors relating to
beliefs about closeness, depending on others, and potential abandonment.
A few years later, the Attachment Style Questionnaire (ASQ; Feeney, Noller, &
Hanrahan, 1994) was developed. Instead of its items being adapted from Hazan and Shaver‘s
measure, Feeney et al. (1994) constructed the ASQ directly from attachment theory principles.
The ASQ consists of 40 items with five factors relating to attachment anxiety or avoidance: selfand other-confidence, relationship anxiety, feelings about closeness, level of approval-seeking,
and priority given to relationships. These continuous measures of attachment, each primarily
addressing attachment anxiety and avoidance, had the desired effect by improving internal
consistency and test-retest reliability to moderate levels.
Bartholomew (1990) constructed a measure from an integration of attachment principles
relevant to both the AAI and Hazan and Shaver‘s (1987) self-report measure. In this manner, a
dimensional model was created that consisted of four quadrants representing the potential for
positive or negative models of self and other. This self-report measure is unique in that it used
the adapted AAI labels. Bartholomew conceptualized secure attachment as having a positive
model of self and other. Those with a positive model of self but a negative model of other were
classified as dismissing, while individuals with a positive model of other and a negative model of
self were classified as preoccupied. The final group represents Bartholomew‘s unique
contribution to the attachment categories. Individuals with both a negative model of both self
and other were considered fearful.
From these concepts, two attachment measures were developed: the Relationship
Questionnaire (RQ; Bartholomew & Horowitz, 1991), and the Relationship Scales Questionnaire
(RSQ; Griffin & Bartholomew, 1994). Adapted from Hazan and Shaver‘s (1987) measure, the

31

RQ is composed of four short paragraph prototypes of the attachment styles, with the additional
paragraph representing fearful attachment. An additional item allows respondents to rate each
paragraph on how much it is like them by using a 7-point Likert scale, thus making continuous
measurement possible. The RSQ was developed by deconstructing the RQ content into a 30item questionnaire intended to yield only a continuous measure of attachment.
Although Bartholomew‘s (1990) dimensional concept of attachment is valuable from a
conceptual standpoint, internal consistency has been rather low, bringing into question its general
utility. As a result, Brennan, Clark, and Shaver (1998) constructed a questionnaire incorporating
all the items from the existing continuous scales and administered the entire set to a large
undergraduate sample. A factor analysis led to the development of the Experiences in Close
Relationships scale (ECR; Brennan et al., 1998). The ECR consists of two, 18-item scales with
high reliability that are based on the two dimensions identified in the factor analysis (avoidance
and anxiety). As a result, Brennan et al. argue that the ECR scales more accurately parallel
Ainsworth et al.‘s (1978) attachment constructs and coding scales originally utilized for motherinfant interactions.
Attachment and Emotional Experience in Couple Relationships
Earlier sections emphasized the importance of effectiveness research in couple therapy,
including the need to understand how to decrease the number of clients who do not receive
optimal results from therapy. Matching treatment to client characteristics represents one
potentially useful way to enhance treatment effectiveness. However, the client-treatment
matching research has been limited in focus and scope, thereby limiting the ability to accurately
assess the possible impact of client characteristics. A common factors approach that focuses on
matching and adapting core elements of treatment process to client characteristics such as

32

attachment seemingly holds a great deal of promise. In addition, attachment has been suggested
as a more relevant client characteristic than the usual choice of diagnostic category. Aspects of
emotional experience (e.g., depth of emotional processing) have regularly been found to be
examples of such process mechanisms that have been linked to positive outcomes. How
individuals process emotion has a direct effect on how they attempt to regulate it. When
individuals have difficulty processing emotion, they often respond by over- or under-regulation.
Attachment theory sets forth parallel ideas within a relational framework. From a very
early age, people develop relationship models that inform them how caregivers will respond
when they are confronted with separation, threat, or distress (Bowlby, 1973, 1982). If their
needs are not consistently met with comfort, sensitivity, and closeness, they develop
hyperactivating or deactivating coping strategies that continue to be relevant throughout the
lifespan and influence their close adult relationships (Kobak et al., 1993). These attachment
ideas closely mirror the concepts of emotion processing and regulation (including emotional
over- and under-regulation), but add the relationship component that illustrates how individuals
develop an internal working model of self and other (Bowlby, 1973).
Distress (often evidenced through the emotions of fear, anger, or sadness) activates the
attachment behavioral system (Kobak & Madsen, 2008), which results from the internal model
of relationships with respect to the instrumentality and function of attachment figures in helping
regulate distress when separated or threatened. This model then organizes the cognitive,
behavioral, and emotional response (i.e., the depth of emotional processing and emotion
regulation). As such, attachment is an intrapersonal construct with intrapersonal and
interpersonal consequences and shows great potential for being a client characteristic with an
important theoretical connection to a core variable associated with positive treatment outcomes:

33

emotional experience. The following sections focus on empirical research that has begun to
investigate the relationship between attachment and two indicators of emotional experience:
physiological arousal and interpersonal distress.
Adult Attachment and Physiological Arousal
Signs of high levels of emotional intensity and the adverse emotional coping they
promote are subtle and may not be immediately demonstrated. However, physiological measures
have proven to be valuable indicators of emotional experience that only later manifest itself in
outward behavior. In research with couples, physiological measures (e.g., skin conductance and
cardiovascular) have been labeled as indicators of diffuse physiological arousal (DPA; Levenson
& Gottman, 1985), conceptualized as a general alarm mechanism found to be predictive of
couple dissatisfaction. DPA results from an inability to self-soothe in a stressful interpersonal
situation such as a recorded discussion of a current couple problem or concern. This emotional
flooding contributes to higher levels of interpersonal distress and dysfunctional interaction,
including a propensity towards demand-withdraw patterns (Gottman, 1998).
Gottman (1990) also found that DPA was associated with a person‘s decreased ability to
process information and a regression to more automatic patterns of behavior and cognition, thus
providing further evidence that lower levels of emotional processing lead to heightened
emotional intensity and lower regulation. In terms of attachment, physiological indicators
represent an important way to understand the potentially subtle ways that attachment relates to a
person‘s experience when distressed, one goal of this study. Sroufe and Waters (1977) first used
physiological data this way, finding that in several of their case studies, avoidant infants
demonstrated accelerated heart rate upon reunion in the Strange Situation, even though they
appeared unexpressive and unresponsive to their caregiver‘s presence and did not signal that they

34

were distressed and in need of comfort. The following sections summarize how researchers have
used physiological measures since that time to understand attachment in adulthood.
Narrative/discourse attachment. Relatively few studies have looked at the link between
AAI security versus insecurity and a person‘s physiological response to distress. Two of these
(Dozier & Kobak, 1992; Roisman, Tsai, & Chiang, 2004) utilized the Kobak (1989) q-set
method of coding the AAI to differentiate between attachment strategies and physiological
responses during the AAI. Dozier and Kobak (1992) found that individuals higher in
deactivating strategies showed greater changes from baseline (signs of emotional suppression) on
AAI questions about separation, rejection, threat, and relationship evaluation. However, changes
in skin conductance were not related to other patterns of insecurity. Roisman et al. (2004) later
replicated these findings and provided additional evidence that deactivation was ―uniquely
correlated with electrodermal activity‖ (p. 777) by showing that deactivation was not correlated
with cardiovascular activation.
One other study analyzed how attachment strategies were related to physiological
responses during couple conflict (Roisman, 2007). In seeking to minimize potential limitations
of other studies that focused primarily on dating couples, participants in this study were 40
engaged couples and 40 couples who had been married at least 15 years. Findings revealed that
during couple interaction about conflict, deactivating adults demonstrated higher deviations in
skin conductance from baseline levels, while hyperactivation was associated with increases in
heart rate. Interestingly, skin conductance levels for secure individuals remained near baseline
throughout the couple interaction.
Self-report attachment. The link between physiology and self-report attachment has been
analyzed by several different methods, with varying results. One study of dating couples found

35

an inverse relationship between anxious and avoidant attachment in terms of diastolic blood
pressure as well as the interaction between pulse rate and systolic blood pressure (Kim, 2006).
Although each of these indicators increased as perceived stress increased for those who reported
anxious attachment, they actually decreased for those who endorsed avoidant attachment.
Another study of women in a dating relationship confronted participants with a potentially
stressful situation on two different occasions, varying her partner‘s presence (Carpenter &
Kirkpatrick, 1996). Interestingly, anxious and avoidant individuals exhibited elevated heart rate
and blood pressure responses to the potential stressor only when the partner was present. These
divergent findings make it difficult to understand whether attachment insecurity yields an
increase or decrease in physiological indicators of stress as measured by heart rate and blood
pressure. An important question is whether those two indicators discriminate between
attachment styles in ways that are consistent with attachment theory.
Another study measured the skin conductance of 74 non-distressed, cohabiting couples
during an attachment-related discussion (Diamond, Hicks, & Otter-Henderson, 2006). Findings
for those who reported avoidant attachment were similar to those for individuals classified as
dismissing on the AAI. The elevated skin conductance for those reporting an avoidant
attachment style were typical of repressive coping, ―in which tasks that elicit negative thoughts
and feelings are accompanied by heightened and escalating sympathetic nervous system
reactivity in the absence of corresponding self-reported distress‖ (p. 221). Also similar were the
findings that attachment anxiety was not associated with skin conductance level or change.
Two other methods for measuring the relationship between attachment styles and
physiological reactivity have been vagal tone and cortisol levels. Vagal tone, which has been
identified as an indicator of emotion regulation, was found in one study to be negatively

36

associated with attachment anxiety and positively associated with perceptions of emotional
security (Diamond & Hicks, 2005). Interestingly, fluctuations in vagal tone were not related
with the presence of an individual‘s partner. In another study (Powers, Pietromonaco, Gunlicks,
& Sayer, 2006), cortisol levels were measured at seven points before and after 124 adolescent
couples engaged in a problem resolution task and varied according to gender. Although
attachment avoidance predicted cortisol levels in women, attachment anxiety was most
predictive for men.
In sum, it appears that skin conductance, at least in terms of deviation from baseline, has
been a reliable and valid indicator of emotional suppression that often accompanies attachment
deactivation as measured by both self-report and narrative/discourse traditions. However,
findings regarding potential indicators of attachment hyperactivation have been less conclusive,
with heart rate and vagal tone showing the most promise. However, heart rate is influenced by
both sympathetic and parasympathetic activity (Bradley & Lang, 2007). This can be problematic
because these systems at times operate in opposition to one another, with the sympathetic
nervous system varying with emotional intensity and the parasympathetic nervous system
activating during times of rest versus stimulation (Bradley & Lang, 2007). This makes it
difficult to reach definitive conclusions from heart rate results alone. As an indicator of emotion
regulation, vagal tone (Porges, Doussard-Roosevelt, & Maiti, 1994) may effectively measure
attachment hyperactivation, but more research is needed to demonstrate its reliability and
validity in both attachment measurement traditions.
Adult Attachment and Interpersonal Distress
The foregoing summarized research focused on the relationship between attachment and
an unconscious, intrapersonal indicator of emotional experience, physiological arousal. Another

37

area that is strongly influenced by emotional processing and regulation and that serves as an
important indicator of a person‘s experience is the interpersonal distress within a relationship.
Interpersonal distress has been conceptualized as a conscious, interpersonal construct that
reflects how emotions are both experienced and expressed within a relationship along a
continuum of softening, hostility, and withdrawal according to the following five dimensions
(adapted from Seedall and Butler, 2006): (a) level of emotional evenness and stability (calming);
(b) amount of clarity in understanding self, partner, and the problem (understanding); (c) self and
partner accessibility, receptivity, and responsiveness (closeness); (d) the ability to express one‘s
thoughts, feelings, and concerns (safety); and (e) the optimism for relationship improvement and
change (hope).
Interpersonal distress parallels how emotional intensity is manifested during couple
interaction. Emotion under-regulation (i.e., attachment hyperactivation) is likely to lead to
hostility, while over-regulation (i.e., attachment deactivation) is more predictive of withdrawal.
Both disrupt relationships and lead to relationship distress, dissatisfaction, and ultimately
dissolution (Gottman, 1993; Gottman & Levenson, 1992; Heavey, Christensen, & Malamuth,
1995; Weiss & Heyman, 1997). The alternative to withdrawal and hostility is softening. In
emotionally focused therapy (EFT) literature, softening is a bonding process contributing to
positive outcomes (Bradley & Furrow, 2004; Johnson & Greenberg, 1988), whereby ―a newly
vulnerable spouse reaches out to a now accessible and engaged partner and asks for his or her
attachment needs to be met‖ (Johnson, 2003a, p. 108).
Interestingly, attachment research has looked at interpersonal distress exclusively in
terms of how couple interaction is perceived by an outside observational coder rather than how it
is continuously experienced (as reported by each partner) during the couple interaction.

38

Nonetheless, much can be learned from assessing interpersonal distress from a variety of
methods, including alternative perspectives (global and continuous self-report as well as the
observational reports of couple behaviors or interaction quality by trained coders) and times
(immediate as well as long-term feedback). The following section will summarize the literature
on the relationship between attachment and interpersonal distress coded observationally.
Narrative/discourse attachment. Research on how attachment is associated with couple
interaction has generally found associations between AAI security and lower levels of
interpersonal distress. AAI security has been associated with higher observed marital quality
(Roisman, Collins, Sroufe, & Egeland, 2005), as well as greater couple synchrony and lower
dominance (Bouthillier, Julien, Dube, Belanger, & Hamelin, 2002). In terms of individual
functioning in the context of couple interaction, Bouthillier et al. (2002) found that security in
men was associated with statistically significant differences in terms of support, self-disclosure,
and engagement. Security in women was also associated with higher levels of support offered to
their partner. Another study specifically coded caregiving behaviors during a conflict resolution
task (Simpson, Winterheld, Rholes, & Orina, 2007). Secure individuals on the AAI were more
accepting of and calmed by their partners‘ emotional caregiving efforts, while dismissing
individuals were calmed by more instrumental caregiving.
Similarly, security in both spouses was associated with less conflict and better overall
functioning than married couples where at least one member was insecure on the AAI (Cohn,
Silver, Cowan, Cowan, & Pearson, 1992). The lone exception was when a secure man was
paired with an insecure woman, in which case their interaction did not qualitatively differ from
couples where both spouses were secure. Another study of 145 couples also found some
variation in results based on gender and whether security was associated with more positive or

39

simply fewer negative behaviors (Creasey, 2002). Specifically, there were no differences in
positive behaviors expressed by men when controlling for attachment. Rather, the only
significant difference was that couples with a man categorized as insecure were coded as more
negative than those couples with a secure man. Couples with a woman coded secure, however,
displayed significantly more positive and less negative behaviors than those couples with a
woman coded insecure. Thus, men and women may differ in the effects that their attachment
strategies have on their couple interactions as well as the partner in general.
Although most studies of attachment and couple interaction typically consist of
relationship-relevant partner distress stemming from discussing an area of disagreement, one
study looked at the potential role of attachment on relationship-irrelevant partner distress
(Simpson, Rholes, Orina, & Grich, 2002). Heterosexual couples who had been dating for at least
three months were observed while the man waited to perform a stressful task. Women classified
as secure appeared to be more attuned to their partners‘ needs, providing more support when it
was sought and less when it was not. Interestingly, attachment was not significantly associated
with the amount of support-seeking in men.
Most research analyzing the links between attachment and relationship-relevant distress
have had non-clinical couple samples. However, one study of distressed clinical couples found
an association between lower security on the AAI and difficulty in problem-solving (Wampler,
Shi, Nelson, & Kimball, 2003). This included ―more negative affect, less respect, less openness,
more avoidance, and less willingness to negotiate when interacting with their partner‖ (p. 497).
Wampler et al. (2003) also found that only negative affect was associated with less AAI security,
while the only predictor of positive affect was positive affect from the partner, indicating that
positive affect may be a reciprocal process while negative affect may be related to greater AAI

40

attachment insecurity. One other study of clinical couples indicated that insecurity is more
common among violent than maritally distressed men (Babcock, Jacobson, Gottman, &
Yerington, 2000). Dismissing, violent husbands were the most controlling and distancing, while
preoccupied, violent husbands were more provocatively confrontational.
Self-report attachment. A social and personality psychology perspective on attachment
has looked not only at the differential effects of attachment on observed behaviors that contribute
to interpersonal distress but also on the self-reported perceptions and attributions that may
influence those behaviors. For example, more secure individuals reported feeling less threatened
during conflict, thereby leading to less fighting and more effective problem resolution (Pistole &
Arricale, 2003). Those who reported greater attachment insecurity also reported lower levels of
validation, higher escalation, and more problematic conflict management strategies (Creasey &
Hesson-McInnis, 2001; Creasey, Kershaw, & Boston, 1999). Both avoidant and anxious
individuals reported fewer adaptive responses to positive and negative events (Campbell,
Simpson, Boldry, & Kashy, 2005; Collins, Ford, Guichard, & Allard, 2006) Specifically,
Campbell et al. (2005) found that attachment anxiety was associated with perceptions of conflict
as more severe, hurtful, and far-reaching, while attachment avoidance was related to viewing
supportive events as less positive.
Similar to AAI results, self-reported security appeared to be an asset during observed
couple conflict (Guerrero & Jones, 2005). Individuals who reported high levels of attachment
anxiety were viewed as less composed (Guerrero & Jones, 2005), more negative (Simpson,
Rholes, & Phillips, 1996) and more distressed following the couple interaction, regardless of
their partners‘ behaviors (Campbell et al., 2005). They were also more likely to overreact and
escalate conflict (Campbell et al., 2005). Attachment avoidance was associated with lower levels

41

of observed expressiveness and other-orientation (Guerrero & Jones, 2005) as well as fewer prorelationship behaviors (Collins, Cooper, Albino, & Allard, 2002).
Two other observational studies looked at how self-reported attachment differentiated
between behaviors when one partner experienced relationship-irrelevant distress. In each study,
couple behaviors were coded while they waited for the man (Simpson et al., 2002) or the woman
(Rholes, Simpson, & Orina, 1999) to engage in a stressful task. Significant results were
primarily related to attachment avoidance. When women scored as avoidant were faced with a
potentially stressful situation, they displayed greater anger in the context of their own distress or
when they perceived little partner support (Rholes et al., 1999). They also provided less support
when their partners were the ones faced with a potentially stressful situation, regardless of the
amount of support their partners sought (Simpson et al., 2002). When their partners were
experiencing distress, men scored as avoidant were more likely to display high levels of anger.
The only significant result for anxious (ambivalent) attachment was that women scored as
anxious behaved more negatively toward their partners in a five-minute period after they were
informed that they would not have to do the stressful task, especially if they had been distressed
and sought support (Rholes et al., 1999)
In sum, just as analyzing mother-infant interaction was critical to understanding infant
behavior in the SS and linking it to attachment, couple interaction research involving either
relationship-relevant or irrelevant distress has the potential to yield valuable insights into ways
that attachment behaviors are manifested in adult romantic relationships. In terms of observed
couple interaction, it appears that attachment security as measured by both traditions is an overall
asset, while insecurity is associated with higher levels of interpersonal distress and more negative
interaction patterns. However, one study using the AAI found that security was not associated

42

with more positive feelings about the relationship when assessed five years after the initial
assessment. In that study, all partner attachment pairings (secure-secure, insecure-secure,
insecure-insecure) were equally likely (20%) to have ended their relationship (Crowell &
Treboux, 2001), indicating that attachment may not predict later relationship dissolution.
Regardless, more research is needed to understand how attachment influences relationship and
therapy process and how it may differentiate between partner experiences in each.
Attachment in Psychotherapy
As mentioned previously, although research has begun to establish a link between
attachment and indicators of emotional experience in relationships (e.g., physiological arousal
and interpersonal distress), more research is needed to fully understand the differentiating effects
of attachment in couple relationships. Of perhaps even greater importance is the need for
analyzing how attachment relates to individuals‘ experiences in psychotherapy. Twenty years
ago, Bowlby (1988) expressed disappointment at the apparent disconnect between attachment
theory and clinical practice, even though he believed that both were inherently compatible. At
that time, he outlined five therapeutic tasks anchored in attachment theory: (a) create a secure
therapeutic environment, (b) assist clients to understand and evaluate how s/he engages current
relationships, (c) explore ways in which current behaviors are linked to past childhood
experiences with attachment figures, (d) use the therapeutic relationship as a model for
exploration, and (e) help clients evaluate current relationship images, develop alternatives, and
work towards restructuring them. Unfortunately, relatively little empirical work has been done
in any of these areas. This section summarizes existing literature that has addressed the
relationship between attachment and clinical work.
Narrative/discourse attachment. Although valuable work has begun to look at the

43

relevance of attachment-informed assessment and intervention and their influence on outcome
(see Steele & Steele, 2008), only a small number using narrative/discourse measures of
attachment have treated attachment as a client characteristic that influences therapeutic process.
Clients classified secure on the AAI were more compliant during treatment (Dozier, 1990) and
more committed to learning and putting what they learned into practice (Korfmacher et al., 1997)
than those with other attachment strategies. Clients with dismissing states of mind were judged
to be less invested in treatment (Dozier, 1990; Korfmacher et al., 1997) and less likely to selfdisclose (Dozier, 1990). However, a study of that analyzed the attachment state of mind of 54
clients with serious psychiatric disorders and their case managers found a complementary case
manager-client effect of deactivation (Tyrrell, Dozier, Teague, & Fallot, 1999). Highly
deactivating clients reported a higher working alliance with case managers who were lower in
deactivation, while clients low in deactivating tendencies reported a higher working alliance with
case managers who were higher in deactivation.
Also of note are the findings of an outcome study of 85 nonpsychotic inpatients that those
classified as dismissing on the AAI were most likely to demonstrate clinical improvement
(Fonagy et al., 1996). This led Meyer and Pilkonis (2001) to conclude,
Securely attached patients might engage productively in most therapy settings,
whereas those with dismissing attachment may require concentrated or targeted
interventions, helping them overcome their characteristic detachment. Once they
do connect emotionally with a therapist, however, improvement might be all the
more dramatic. (p. 467).
Although these findings support the idea that attachment does indeed influence therapeutic
process and outcome, more work is needed to understand how attachment relates to client

44

experiences, which in turn influence process and outcome. In other words, beyond knowing that
certain individuals may be easier or more difficult to treat, it is important to know what they may
be experiencing in therapy and how that may be facilitating or impeding positive therapy process
and how treatment could thereby be adapted to improve outcomes.
Self-report attachment. With respect to self-report attachment, much of the clinical
research has focused on the effects of attachment on the therapist-client working alliance.
However, results have varied. In one study using the RQ, security was significantly related to a
positive working alliance (Satterfield & Lyddon, 1998). Other studies have found a relationship
between a weaker working alliance and higher levels of avoidance (using the RQ; Parish &
Eagle, 2003) and anxiety (using the AAS; Mallinckrodt, Coble, et al., 1995). As one exception
to these findings, Reis and Grenyer (2004) found no significant relationship on the RQ between
attachment and working alliance. One other study measured the working alliance during the
beginning, middle, and end stages of therapy and found evidence of unique treatment trajectories
(Kanninen, Salo, & Punamaki, 2000). Although the working alliance remained relatively stable
at all three points for secure clients, it was more likely to be poor at the middle stage for
anxiously attached individuals before becoming much stronger at the end. This differed from
avoidant clients, who reported that the alliance was deteriorating at the end stage.
In addition to being associated with the working alliance, attachment has also been found
to relate to therapy process. Although not all studies found that attachment differentiated
between those clients who progressed and those that did not (Cyranowski et al., 2002), security
has been linked to higher levels of client-therapist agreement perceived by the therapist
regarding the tasks and goals of therapy (Dolan, Arnkoff, & Glass, 1993). Attachment avoidance
in one study was associated with more negative views about engaging in therapy (Lopez,

45

Melendez, Sauer, Berger, & Wyssmann, 1998), and higher concurrent levels of anxiety and
avoidance (conceptualized as fearful attachment) were inversely related to viewing the first
therapy session as smooth and valuable (Mohr, Gelso, & Hill, 2005).
Although findings indicate that attachment as measured by self-report is associated with
therapy process and outcome, more research is needed to ensure that these significant results are
not a product of method variance that may accompany using self-report attachment and outcome
measures. Regardless, additional research will enhance understanding of the process by which
attachment influences therapy. There is also some evidence that treatment approach can be
effectively matched to client attachment. One study found that cognitive behavioral therapy
(CBT) was more effective in treating major depression for individuals scored as avoidant than
interpersonal psychotherapy (IPT; McBride et al., 2006). However, beyond simply knowing that
one approach is better than the other in terms of attachment, it is important to discover what
elements and processes are making the difference in clients‘ experiences.
Attachment in Clinical Work with Couples
Unfortunately, only one research article has analyzed the relationship between attachment
and clinical work. Makinen and Johnson (2006) utilized the Experiences in Close Relationships
scale (ECR) to assess whether EFT led to categorical attachment change after 12 therapy
sessions. However, their study found no differences between attachment dimensions in terms of
therapeutic outcome, leading to hypotheses that they ―may be more enduring characteristics that
are not easily modified‖ (p. 1062), that attachment change may ―require a number of positive
experiences even after an optimal 'corrective' experience of new and reparative interactions in
therapy‖ (p. 1062), or that the ECR was ―not sensitive enough to detect changes in attachment‖
(p. 1063) over the 12-session timeframe.

46

Regardless of the reason, these non-significant outcomes bring up the need to further
investigate how couple therapy might influence attachment continuously, along a continuous
spectrum of security, rather than categorically (Dozier et al., 2005). In other words, rather than
focusing on helping individuals completely change longstanding attachment models, therapists
work to help clients to enhance the security of their interactions by (a) becoming more aware of
their propensities and automatic responses; (b) more consciously responding to their partner‘s
distress; and (c) learning to signal and express their needs more appropriately during times of
distress. In this manner, therapy focuses on setting the stage for clients to progress towards more
secure behavior (Bick & Dozier, 2008). To do this, couple therapists and researchers need to
understand what factors may contribute to positive attachment change and whether therapy helps
change attachment classification, mobilizes clients for future change, or primarily lessens
individual and couple distress.
However, not only can couple therapy potentially influence client attachment, an
understanding of attachment may also help therapists effectively adapt their clinical interventions
to their clients‘ attachment characteristics. Eagle (2006) points out, ―The challenge to
attachment theorists and researchers is to spell out as precisely as possible the ways in which
attachment theory contributes to the way one does psychotherapy and the specific interventions
one uses‖ (p. 1087). This may produce farther-reaching effects for therapists than demonstrating
that therapy changes or influences attachment.
Unfortunately, couple therapy has seldom attempted to address attachment and therapy
process beyond the attachment-based EFT model. EFT explains how attachment informs its
approach conceptually but is less clear about how treatment process may be adapted to fit
specific client characteristics. Adapting the timing and delivery of specific interventions to client

47

characteristics such as internal working model of attachment may enhance engagement in the
therapeutic process and thereby more effectively facilitate change. If couple therapy is to be
effectively adapted to client attachment, more research is needed that analyzes how varying
attachment strategies may influence (a) the treatment experiences of clients, (b) how clients
approach therapy and (c) how and what interventions may be effectively matched to clients.
The Purpose of This Study
This literature review summarizes how to enhance treatment effectiveness in couple
therapy by adapting treatment process to varying client characteristics. Attachment is an
intrapersonal characteristic that organizes how individuals approach close relationships in
general and the degree to which they seek support when they feel distressed or threatened. When
individuals do not trust that their needs will be met, they adopt hyperactivating or deactivating
strategies as a way of coping. In this manner, attachment represents a potentially important
client characteristic that could influence the clinical experiences of couples as well as the
ultimate outcome of therapy.
One of the primary objectives of couple therapy is to improve the process by which
partners interact with one another. Emotions and how they are experienced (processed and
regulated) have been identified as an integral part of couple interaction that influences process
and outcome. An attachment perspective adds a relational component to this process by
emphasizing the importance of partners becoming a secure base and safe haven for each other.
Yet expressing more vulnerable, primary emotions can be potentially threatening and impede
therapeutic change. For example, an individual who has consistently experienced rejection from
attachment figures and has developed a deactivating coping strategy may find it difficult to open
up emotionally. Others whose attachment figures inconsistently responded to their needs have

48

adopted hyperactivating strategies that include expressing more volatile, secondary emotions.
Thus, attachment patterns as well as the couple‘s history of interactions will likely influence how
well partners signal their own distress as well their responsiveness to each other‘s distress.
A core issue in couple therapy is how to appropriately handle emotional intensity that
often accompanies distress. Therapist-coached, direct couple interaction through enactments
holds great promise as a common change mechanism for altering interaction patterns, improving
treatment process, and ultimately facilitating long-term couple self-reliance. Yet enactments
need to be adapted throughout therapy to partners‘ willingness and readiness to discuss
emotionally-laden content with one another. Although low structured, semi-natural couple
interaction may be utilized initially to assess couple functioning, very often the couple‘s high
volatility needs to be buffered by filtering much of the interaction through the therapist, who
models healthy interaction. As the emotional experience of the couple begins to change and
healthier patterns emerge, the interactional structure shifts again from the standard therapy-like
interaction to more freeform direct interaction where the therapist is less active and functions
primarily as a process coach.
As a result, only limited conceptual work has addressed how the interactional structure
during therapy sessions may be adapted according to couple emotional reactivity. This
represents a small step in couple therapy toward understanding how to adapt treatment process to
varying client characteristics in order to improve outcome. In order to accomplish this objective
more fully, empirical research is needed to understand what influences emotional experience in
therapy. Because couple therapy often explores the more difficult aspects of a couple‘s
relationship, an understanding of attachment and how individuals and couples cope with distress
in a relationship context is also needed. Although some valuable psychotherapy research exists

49

that explains how attachment influences relationships, no research has addressed the role of
attachment in the context of couple therapy.
This study represents a first step in making some of these connections by studying how
the client characteristic of attachment interacts with two general interactional conditions (low
structure, semi-natural versus high structure, therapy-like) to influence an individual‘s emotional
experience (physiological arousal and self-reported interpersonal distress). Although both
narrative/discourse and self-report attachment were measured during data collection, results from
only the attachment self-report measurement tradition are reported as part of this dissertation.
In this manner, this study addressed the following research questions:
Research Question 1: Are higher levels of self-reported attachment avoidance and/or
anxiety associated with a more negative emotional experience for partners in each of the
interactional conditions?
Research Question 2: How do the effects of self-reported attachment account for the
differences in emotional experience between the semi-natural and therapy-like conditions?
Research Question 3: How will self-reported attachment influence the congruence
between the observed physiological arousal of participants and their moment-to-moment feelings
towards their partner?

50

Chapter III
METHODS
Design
This study investigated the relationship between attachment and emotional experience
(i.e., physiological arousal and interpersonal distress) in the context of moderately distressing
low structure, semi-natural, and high structure, therapy-like couple interactions. Data were
gathered during one 3 ½ -4 ½ hour experimental session consisting of four phases combining
self-report, continuous physiological measurement, and interpersonal process recall (Elliott,
1986; Kagan & Kagan, 1990; Kagan et al., 1963) designed to increase experimental validity by
measuring variables using a variety of methods (i.e., physiological data, observational coding,
interpersonal process recall, and self-report).
The original plan was to counterbalance the semi-natural and therapy-like conditions.
However, the option of establishing a fixed order of semi-natural interaction and then therapylike was chosen for two primary reasons. The potential limitations arising from confounds
associated with not counterbalancing the interactions (i.e., not being able to directly compare the
effectiveness of the semi-natural versus therapy-like interactions) were deemed less than those
that would likely result from couples processing their issues first with the therapist/coach before
then being able to discuss the issue directly between spouses. In other words, gaining
therapeutic resources before the semi-natural interaction would greatly reduce the likelihood of it
representing how participants interact at home. In addition, it made conceptual sense that
couples most often talk first about their issues alone before a therapist might intervene. In this
manner, this study contrasted the semi-natural condition with a therapy-like condition that was
preceded by a semi-natural interaction.

51

Predictor Variables
Because narrative/discourse attachment data were obtained using the Adult Attachment
Interview (AAI) but not analyzed during the dissertation portion of this study, two predictor
variables were used in this study relating to attachment: self-reported avoidance and anxiety as
measured by the Experiences in Close Relationships scale (ECR). Anxiety and avoidance
conceptually parallel the concepts of attachment hyperactivation and deactivation (Kobak et al.,
1993). As stated previously, client-treatment matching from a common factors perspective
provides a great deal of promise for understanding how to adapt treatment to specific client
characteristics. Yet nearly all of the studies have sought to differentiate which therapeutic
modalities can be matched to client characteristics, disregarding the ingredients of change shared
by treatments. A common factors approach allowed a conceptual shift from focusing on
modalities to those processes or interventions that might be can be adapted to client
characteristics, including attachment.
Attachment refers to a model of self and other that helps to organize exploratory
behaviors as well as the regulation of emotions when confronted with separation, threat, or
distress (Bowlby, 1973). When needs are not consistently met, individuals develop one of two
forms of emotional coping: hyperactivation or deactivation (Dozier et al., 2001). Although
perhaps adaptive in helping the infant‘s world be more predictable and avoid the pain of
rejection, these insecure strategies have been consistently found across two measurement
traditions to have negative intrapersonal and interpersonal consequences throughout the lifespan.
Yet very little research has been done to understand the relationship between attachment and
therapy. Much more is needed to understand how attachment relates to client experiences in
therapy in a way that informs how therapists adapt their approaches to specific clients.

52

One other predictor variable looked at in this study was the condition within which the
couple interaction occurred. Whereas individual psychotherapy focuses on therapist interaction
with the client, couple therapy provides an additional therapist-client interaction as well as clientclient interaction. With this added complexity, organizing and structuring the interactions is
paramount for positive outcomes to occur. Yet few empirical or theoretical articles have
addressed these issues in couple therapy. Recent work studying enactments has provided the
most impetus in this area. During an enactment, the therapist focuses on facilitating direct
couple interaction (Butler & Wampler, 1999). However, an enactment consists of more than
simply telling partners to talk to one another, something which many beginning therapists
errantly believe (Nichols & Fellenberg, 2000). This misconception leads to permissive,
inadequately-structured enactments where emotionality and volatility are not appropriately
buffered by the therapists, who are then more likely to abandon enactments entirely.
On the other extreme, some therapists tend to exert unnecessary control over sessions by
filtering all the interaction through themselves. The centrality of enactments in couple therapy
(i.e., facilitating direct couple interaction) has led to a developmental model (Butler & Gardner,
2003) that provides guidelines on how to adapt enactments to varying levels of emotionality and
volatility by adjusting the therapist role and overall structure of the interaction (see also Seedall
& Butler, 2006). During earlier stages of therapy, the interaction is much more structured and
filtered almost exclusively through the therapist. Later stages are more free-form, with the
couple interacting directly more often while the therapist coaches from the periphery. The
overall goal of the developmental enactment model is couple interactional self-reliance (Butler &
Gardner, 2003).
In addition to the structuring problems that may arise when facilitating an enactment,

53

therapists may also permit couples to interact solely on the surface by talking about superficial
issues. This may serve as a buffer and maintain couple reactivity and volatility at manageable
levels, but it does not facilitate a deeper emotional experience or move the couple to new
relationship patterns, something that models such as EFT emphasize (Johnson, 2004). Although
not explicitly tied to Butler and Gardner‘s (2003) stages of enactments or to any one therapy
model, the conditions used in this study (semi-natural and therapy-like) provided varying levels
of structuring (low versus high) in the form of the therapist/coach role and activity while also
encouraging partners to focus on their emotional experience and expression as they discussed a
recent time when they felt hurt, angry, or offended in the relationship.
In the low structure, semi-natural condition, the therapist was present but inactive as both
partners interacted directly. During the high structure, therapy-like condition, the therapist/coach
was much more active, with the majority of the dialogue filtered through him/her. Although a
number of studies have analyzed the relationship between attachment and emotional experience
during couple interaction, none has analyzed how that experience changes in the context of a
therapy-like interaction. In addition, conceptual and empirical work has discussed how the
interactional structure of therapy sessions (therapist role and type of involvement: therapist-client
or client-client interaction coached by the therapist) may be adapted according to couple distress,
reactivity, and volatility (Butler & Gardner, 2003; Seedall & Butler, 2006). However, no work
has compared how attachment strategies might influence therapist or client behaviors within
varying interactional contexts. This study facilitated a comparison of how attachment strategies
influence variations in partner experiences during both semi-natural and therapy-like interaction.
Outcome Variables
The purpose of this study was to begin to understand the relationship between attachment

54

and the emotional experiences of individuals during two types of couple interactions commonly
experienced in couple therapy. As a result, two outcome variables were measured, consisting of
two indicators of emotional experience (physiological arousal and interpersonal distress).
Accessing emotions and using them to create a positive emotional experience and a sense of
emotional support is an important vehicle for facilitating positive outcomes in EFT(Makinen &
Johnson, 2006). However, even outside of EFT, other evidence-based couple therapy
approaches that focus primarily on behaviors and cognitions have identified emotional
experience and expression as important pathways for intervention and indicators of outcome
(Epstein & Baucom, 2002; Jacobson & Margolin, 1979; Snyder, 2002).
Because emotions are often so complex and learned coping strategies can mask emotional
arousal, a physiological indicator, electrodermal activity (EDA) as measured by skin
conductance, was chosen as an indicator of emotional experience. EDA has been found used to
identify emotional arousal and its subsequent suppression (Dawson, Schell, & Filion, 2007). A
variety of strategies exist for measuring skin conductance (Andreassi, 2007; Dawson et al.,
2007), including skin conductance level (SCL) and response (SCR). SCL refers to the average
level of skin conductance over a period of time. SCR refers to the number, magnitude, and/or
length of changes in skin conductance. The primary methods of skin conductance measurement
employed in past attachment literature has been related to SCL as well as the difference between
SCL and a pre-established baseline. Measured in this manner, skin conductance appears to be a
strong indicator of attachment deactivation (Diamond et al., 2006; Dozier & Kobak, 1992;
Roisman et al., 2004). However, SCR indicators and their potential relationship to attachment
have not been previously analyzed in the literature.
Whereas physiological arousal as measured by skin conductance is an intrapersonal

55

indicator, interpersonal distress relates more to emotion expressed in a relational context. This
study focused on participant perceptions of their experience reported continuously while
observing video of their own couple interaction, in contrast to the usual approach of an outside
observer coding the interaction and the participants providing a global, self-report appraisal after
the interaction is complete. Global and self-report measures were also used according to
principles of interpersonal process recall (Elliott, 1986; Kagan & Kagan, 1990; Kagan et al.,
1963). Participants watched their own couple interaction while indicating via wireless
perception analyzer dials their feelings (positive versus negative) towards their partner during
each moment. Continuous self-report has not been used regularly in couple research but
represents an innovative way of understanding experience and change (Gardner, 2004; Gottman
& Levenson, 1985, 1992; Griffin, 1993; Levenson & Gottman, 1983). After reporting in each
moment their feelings towards their partner, participants provided a global assessment of their
overall experience and interpersonal distress across all interactional conditions. Having two
contrasting measures of self-report also provided further insight into whether attachment
strategies influence how interactions are perceived (globally or continuously).
Other Variables
To assess possible confounds, self-report data were gathered on several other variables
that may influence how couples interact and ultimately experience their interaction during seminatural and/or therapy-like interactions. These included participant gender, level of disruptive
couple conflict, global feelings towards one‘s partner, and individual symptom distress.
Participants
Couples
Participant recruitment. A total of 65 couples blind to the specific research questions

56

and study purposes participated in the research project. Participants were recruited over a seven
month period using flyers posted at various locations in the community, including grocery stores,
thrift stores, laundromats, doctor‘s offices, therapy offices, and Michigan Head Start Association
offices in Lansing, Michigan, and surrounding areas. Two online methods of recruitment were
also used: the Michigan State University Family Resource Center listserv and posts in the
―general community‖ section of www.craigslist.org. The study advertisement invited couples
who were engaged, living together, or married to participate in a 4 ½ hour research session.
They were then given a very brief procedural overview and told that couples chosen to
participate would receive $100.00. Those interested in participating contacted the primary
investigator and received a more detailed overview of the study and a brief telephone screening
(see Appendix A) designed to ensure that they met the following inclusion criteria: (a)
participants were 18 years or older; (b) partners were engaged, cohabiting, or married; (c) they
had been in the relationship for at least one year; and (d) they had a fairly recent couple
disagreement in which they felt hurt or angry with their partner.
Participant exclusion criteria. Couples were excluded from participation at both
extremes of the continuum of distress. Those who reported that they rarely felt hurt, angry, or
offended with their partner and had not had a fairly recent incident were screened out because of
an increased risk that they would not have a topic to discuss during the experimental session. At
the other extreme of couple distress, those couples where physical violence had occurred in the
past year that left marks or required a hospital visit were also screened out. Prospective
participants were also asked whether they considered themselves to be in a personal or
relationship crisis of any kind or if they were seriously considering ending their relationship.
Affirmative answers to any of those questions were probed further to ensure that participation in

57

the study would not compromise their mental or physical health and safety. Those chosen to
participate were also asked in-person whether there had been any recent instances of physical
violence since the telephone interview that left marks or required a hospital visit. If they
confirmed an incident of physical violence, the research session was discontinued.
The total number of couples who expressed interest in the study was 122. Of the 122
couples who made contact regarding the study, 65 actually participated, representing a
participation rate of 53.3%. Of those who did not participate, 22 (18.0%) chose not to participate
or did not make contact after the initial telephone screening, 9 (7.4%) did not show for their
scheduled research session, and 7 (5.7%) expressed interest in participating but had scheduling
conflicts. A total of 19 prospective participants (15.6%) were screened out. Reasons for this
were that they did not meet the relationship requirements (n = 2, 1.6%); they expressed an
interest in participating after data collection was complete (n = 4, 3.3%); they expressed
concerns about communicating in English, which was not their primary language (n = 4, 3.3%);
they indicated that they did not have any issues to discuss (n = 5, 4.1%), or they reported
experiencing physical violence in the previous year that had left marks or required a hospital
visit (n = 4, 3.3%).
In addition to those who did not participate, data from two couples were set aside during
the analyses. One research session was frequently disrupted by the presence of the couple‘s sixmonth-old infant, who became quite distressed during the therapy-like interaction, thereby
confounding the factors that contributed to the participants‘ emotional experience. The other
couple was the only same-sex couple who participated in the study. Although future research
should address the same research questions as this study with a sample of same-sex partners, no
real questions could be answered in this study with only one same-sex couple. In addition,

58

because the analyses in this study used gender as a distinguishing variable in order to understand
actor and partner effects, it was deemed appropriate to set this couple‘s data aside until a later
time.
Participant demographics. A summary of participant demographics may be found in
Tables 1a and 1b. For the 63 couples, participant age ranged from 18 to 69, with a median age of
30.0 years. Participants represented a wide range of racial/ethnic categories, including EuroAmerican, African American, Mexican-American/Hispanic, American Indian or Alaska Native,
Arab American, and more than one racial category. Overall, 41.3% (n = 52) of participants
identified themselves as part of a non-White, typically under-represented population, making this
a rather diverse sample. In terms of religious preference, participants identified themselves as
members of a specific Christian (n = 72; 59.0%) or non-Christian (n = 4; 3.3%) denomination as
well as non-denominational (n = 30; 24.6%) or non-religious (n = 16; 13.1%).
Couples were screened out during the telephone interview if they were not in some type
of committed relationship (e.g., married, living together, engaged to be married) or if they had
not been together for at least one year. However, one couple reported on the assessment packet
that they had only been together eight months. One of the primary reasons for the requirement
that participants be together at least one year and be married, engaged, or living together was to
ensure that couples were in an established relationship beyond that of semi-casual dating and
thereby ensure a higher likelihood that partners had actually reached the level of an attachment
figure with one another. Although it was realized after their research session that one couple had
been together less than a year, it was deemed appropriate in that instance to include them in the
study because they were living together and had been together almost one year.
A large majority of participants had received some form of education beyond high school

59

(n = 97; 77.0%). Although most couples reported that at least one of them was employed fulltime (n = 38; 60.3%), 11 couples (17.4%) reported that one or both partners were employed only
part-time, and 9 couples (14.3%) reported that both were unemployed. Four other couples
(6.3%) identified themselves as students, and one couple (1.6%) reported that both were retired.
In terms of couple income, there some indication of financial distress in a number of couples,
with 43.5% (n = 27) reporting reported a couple income of less than $40,000.
Table 1a
Descriptive Statistics of Categorical Demographic Variables
Variable Name
Race / Ethnicity
American Indian or Alaska Native
African American / Black
Caucasian / White
Mexican-American / Hispanic
Biracial
Other
TOTAL

n (%)
3 (2.4)
31 (24.6)
73 (57.9)
6 (4.8)
12 (9.5)
1 (.8)
126

White / Caucasian / Euro-American
Under-Represented Minority
TOTAL

74 (58.7)
52 (41.3)
126

Religious Preference
Catholic
Protestant / Non-Catholic Christian
Non-Denominational
Non-Christian
Non-Religious
TOTAL

24 (19.7)
48 (39.3)
30 (24.6)
4 (3.3)
16 (13.1)
122

Highest Level of Education Completed
High School Diploma / GED or less
Undergraduate College Experience
Advanced College Degrees
TOTAL

29 (23.0)
75 (59.5)
22 (17.5)
126

60

Table 1a (cont‘d)
Variable Name

n (%)

Employment Status
Full-Time
Less than Full-Time
Student
TOTAL

50 (39.7)
62 (49.2)
14 (11.1)
126

Annual Income
< $20,000
$20,000 - $49,999
> $50,000
TOTAL

63 (50.4)
42 (33.6)
20 (16.0)
125

Current Relationship Status
Committed, Non-Marriage Relationship
Marriage Relationship
TOTAL

54 (42.9)
72 (57.1)
126

Number of Children
No Children
1-2 Children
> 3 Children
TOTAL

34 (27.0)
64 (50.8)
28 (22.2)
126

Table 1b
Descriptive Statistics of Continuous Demographic Variables
Variable Name
Age

n
122

M (SD)
32.52 (9.95)

Range
18-69

Time in Relationship

126

7.27 (7.21)

.75 -41

Therapists
Five therapist/coaches (three women; two men) were recruited from the Michigan State
University Couple and Family Therapy Clinic. All therapists/coaches had completed their
master‘s degree and were pursuing their doctoral degrees in marriage and family therapy with at
least 700 hours of direct clinical experience. Training consisted of five sessions lasting

61

approximately two hours each, occurring over a period of three weeks. Each training session
included both didactic and experiential components. In addition, core principles were reviewed
and highlighted regularly for each therapist/coach based upon his/her individual questions,
concerns, and experiences with participants. Overall, four of the coaches facilitated at least 12
research sessions with participants, with the other therapist/coach completing only 4 due to
scheduling constraints.
Therapist/coaches were trained regarding their role and responsibility in the process, as
well as those necessary skills that are common across therapeutic modalities in facilitating and
processing emotional content (see Appendix B for the training manual containing a theoretical
and training overview of how to facilitate emotional content with couples). The primary focus of
therapist/coach training was not to introduce an entirely new set of therapeutic skills but to
emphasize the adaptation and use of the therapists‘ existing skill set to accessing and reprocessing emotion following the guidelines in the training manual. Therapist/coach roles and
responsibilities were distinguished according to two segments for each topic. For the sake of
clarity, the person whose issue was being discussed was referred to as the speaking partner,
while the person who did not introduce the issue was labeled the listening partner. The core
techniques the therapist/coach utilized during the interactions were several of those identified by
EFT for ―accessing and reformulating emotion‖ (Johnson, 2004, p. 78). These included
reflection, validation, evocative responding, heightening, empathic conjecture, enactment, and
promoting positive interaction.
Many of these skills are also utilized in some form by other empirically-supported couple
therapies but with slightly different labels. For example, enhanced cognitive behavioral therapy
(Epstein & Baucom, 2002) provides guidelines of how to alter ―both the experience and

62

expression of emotions‖ (p. 375), including accessing, processing, and expanding the range of
primary emotions that have previously been minimized or avoided. Interventions include
creating a safe environment; amplifying emotional experience through questions; helping the
client recognize emotional meaning beyond what is stated; helping clients understand the
relationship between emotions, cognitions, and behaviors; and heightening their experience.
Behavioral marital therapy (Jacobson & Margolin, 1979) focuses a great deal on
decreasing interpersonal distress by enhancing positivity and helping the couple learn problemsolving skills. This is done by coaching each partner to listen by expressing ―validation,
reflection, and generalized acceptance‖ (p. 206) to one another. In addition to learning the
importance of positive communication, clients also learn how to express their negative feelings
to their partners more appropriately. Insight-oriented marital therapy (Snyder, 2002) emphasizes
affective reconstruction in an effort to promote insight by ―enabling interventions of increasing
depth and emotional challenge‖ (Snyder, 1999, p. 350) that ultimately leads to increased
empathy and support (Snyder & Mitchell, 2008). Thus, each of these couple therapy approaches
utilize emotions in some way to help clients (a) access and understand their emotional
experience, (b) process, expand, and potentially transform that experience, and (c) utilize it to
then enact new patterns of interaction with their partners.
Significant time was spent during training to help therapist/coaches adapt the skills
required for this research study in coaching couple interaction into their existing therapeutic
approach, without the expectation that they were ―doing EFT‖ or any other specific treatment
model. Rather, during the training, therapist/coaches learned to identify how some of the
common change mechanisms across modalities involve the accessing and re-processing of
emotions. In addition, they were instructed that the primary goal for the high structure, therapy-

63

like interaction with the speaking partner was to emotionally process his/her experience using
primary, softer emotions so that the listening partner would be more likely to be engaged in and
really hear what the speaking partner was expressing. At the same time, the therapist/coach was
modeling healthy responding for the listening partner. The purpose of the interaction with the
listening partner was to check for and facilitate understanding and an empathic response to the
needs expressed by the speaking partner.
Training also focused on explaining the purpose of the low structure, semi-natural
interaction as well as the role of the therapist/coach during this condition. During the beginning
stages of therapy, clients may be asked to discuss an area of disagreement as they would at home
while the therapist shifts to the periphery and observes (Gottman, 1999). This allows the
therapist to assess interaction patterns and strategies that the couple uses when attempting to
resolve conflict. Training was designed to help therapist/coaches understand that their role
during the semi-natural interaction was to remain in the room while not interfering with the
couple interaction. They were asked to time the interaction and indicate to the participants when
they should switch topics. Therapist/coaches were present in the room during both interactions
in order to control for the influence a third person‘s presence might have on the experience.
Procedure
Couple Data Collection Procedure
Phase 1: Informed consent and Adult Attachment Interview (AAI). Data collection for
this study consisted of four phases occurring during a 3 ½ to 4 ½ hour experimental session.
Couples were randomly assigned to one of two sequences. Within the first sequence, phase one
(45-90 minutes) consisted of each participant meeting separately with a research associate (the
person administering the AAI for sequence one and the primary investigator for sequence two)

64

who summarized the session layout (see Appendix C for the research session procedure manual)
and discussed issues of informed consent, including assessing for violence and any other safety
issues (see Appendix D). If both partners provided written consent to participate and no violence
or safety issues arose, they remained in separate rooms and the Adult Attachment Interview
(AAI) was administered and digitally recorded for later transcription (Main, 1996). The AAI is a
20-item, semi-structured interview designed to explore early childhood experiences and
relationships with attachment figures (Hesse, 2008). Six graduate research associates were
trained didactically and experientially in the administration of the AAI over a period of two
weeks for approximately 10 hours, including administering at least one practice interview during
that time.
Phase 2: Discussion topic development. During phase two (30-45 minutes), the couple
completed a packet of self-report assessments. A research associate, who functioned as the
therapist/coach during the therapy-like interaction, also met individually with each partner for 510 minutes during this time. This meeting had two primary goals: (a) to begin to establish a
rapport with each person that would allow them to effectively interact with the couple; and (b) to
help the participant recall a recent time in the relationship when s/he felt hurt, angry, or offended
by his/her partner and about which s/he still had feelings (Gardner, 2004; Waldinger, Moore, &
Schulz, 2003). Once the participant chose an incident to discuss, the research associate verified
that the participant was willing and able to discuss it during the couple interaction without
feeling unsafe or too distressed. If either of those criteria was not met, another topic was chosen.
Phase 3: Baseline, semi-natural, and therapy-like interactions. After each partner met
with the research associate and chose a discussion topic, they entered the same room to prepare
for the baseline, semi-natural, and therapy-like interactions (phase three; 60 minutes). At this

65

point, a physiological measurement device was attached to the middle and ring fingers of each
participant‘s non-dominant hand for the duration of phase three. After approximately five
minutes that allowed for the electrode paste to establish an adequate connection with the each
participant‘s skin, they were instructed to clear their minds and focus on relaxing for three
minutes to establish a physiological baseline of skin conductance levels (SCL). After that time,
the therapist/coach entered the room and asked the participants to begin discussing how they met
for five minutes. This served as an interactional warm-up and established an interaction. It also
paralleled a portion of the Oral History Interview, which has been found predictive of
relationship outcome (Buehlman, Gottman, & Katz, 1992; Carrere, Buehlman, Gottman, Coan,
& Ruckstuhl, 2000).
At the conclusion of the five minute interactional warm-up, the therapist/coach explained
the process of the semi-natural couple interaction and identified which partner‘s issue had been
randomly chosen to be discussed first. The therapist/coach also explained that the goal of the
semi-natural interaction was to take steps towards a resolution of the topic. Even though the
couple had likely discussed their issue previously, they were encouraged to focus on processing
their experience—including their thoughts and feelings—so that both partners could gain a better
understanding of the incident and how it occurred (Gardner, 2004). After eight minutes, the
therapist/coach directed the couple to shift to the other partner‘s issue. Although the
therapist/coach remained present during the semi-natural interaction in order to control for the
presence of a third person, s/he functioned from the periphery and remained inactive during the
couple interaction process. After both semi-natural interactions, the therapist/coach facilitated
the therapy-like interaction, which consisted of discussing the same two topics with the
therapist/coach over four, 3-5 minute segments (two segments per issue). The goal of the

66

therapy-like interaction was to help facilitate emotional processing and expression in the
speaking partner as well as understanding and empathy in the listening partner.
The first segment of the therapy-like interaction began with the therapist/coach asking
about the speaking partner‘s experience of the situation. The primary goal was to validate,
explore, and deepen the experience of the person who introduced the issue while helping him/her
to express feelings in a non-threatening, softened way. During the second segment, the
therapist/coach then shifted to the partner who did not introduce the issue and checked for
understanding of both content and emotions while also helping him/her offer comfort to the
distressed partner, when possible. At the conclusion of each segment, the therapist/coach sought
to heighten the experience that both partners had by encouraging the participant who had been
interacting with the therapist/coach to share his/her primary emotions and attachment needs,
threats, or longings directly with his/her partner (i.e., an enactment). At the conclusion of the
second segment, the couple shifted to the other partner‘s topic and repeated the process. When
both partners‘ topics had been discussed, each participant spent 3-4 minutes completing an
adaptation of the couple version of the System for Observing Family Therapy Alliances
(SOFTA-S; Friedlander, Escudero, & Heatherington, 2006). This cool-down period also
provided an opportunity to approximate the physiological recovery towards baseline.
Phase 4: Follow-up session with participants. Following a 5-10 minute break, the
participants entered a room with a television and perception analyzer dials (continuous selfreport rating dials) where they reviewed all five interactions (discussion of how they met, seminatural/own issue; semi-natural/partner‘s issue; therapy-like/own issue; therapy-like/partner‘s
issue) in the order that they occurred and reflected on their interpersonal process (phase four; 60
minutes). Partners were in the same room but were separated by a portable screen that provided

67

each with privacy as they rated their experiences. They also wore earphones with noise
canceling capabilities in order to reduce the likelihood of them influencing one another with
verbal expressions. As they watched, participants rated on a scale of 0-100 how they felt in each
moment towards their partner during each interaction, with 0 being the most negative possible,
50 being neutral, and 100 being the most positive possible. After viewing all of the interactions,
participants provided a composite, proximal rating of their interpersonal distress by completing
the 30-item Subjective Emotional Experience Questionnaire (SEEQ; Seedall, 2008) for the time
period consisting of the baseline, semi-natural and therapy-like interactions.
Because it was unknown if or how participating in the AAI would influence the
interactions, two sequences were planned that counterbalanced the AAI with the interaction
segments. The first sequence (n = 31 couples) consisted of completing the informed consent,
administering the AAI, completing the assessment packet, participating in the baseline, seminatural, and therapy-like interactions, and then watching and reporting on each interaction. The
second sequence (n = 32 couples) consisted of completing the informed consent, completing the
assessment packet, participating in the baseline, semi-natural, and therapy-like interactions,
watching and reporting on each interaction, and then administering the AAI. In this manner, the
AAI was counterbalanced with the interaction phase in the two sequences by occurring
immediately following the informed consent or after participants had watched and reported on
their interactions, with the order for all other activities remaining the same.
Confirmation of Conditions
Occurrence of each condition. During each research session, the primary investigator
observed and confirmed that five interactions occurred (the interactional baseline, two seminatural and two therapy-like interactions). The criteria for the interactional baseline were (a) the

68

couple was interacting directly; (b) the therapist was present but inactive; and (c) the couple was
reminiscing about how they first met. The criteria for the semi-natural interaction were (a) the
couple was interacting directly; (b) the therapist was present but inactive; and (c) each partner
discussed a moderately distressing issue with the other partner. The criteria for the therapy-like
interaction were (a) a therapist/coach was present and facilitating emotional content; (b) each
partner interacted directly with the therapist/coach with the exception of an enactment at the
conclusion of each segment; and (c) each partner discussed his/her issue as well as his/her
partner‘s issue with the therapist/coach. Because identification of these conditions was
observationally straightforward, no reliability coding was deemed necessary.
Integrity of the therapy-like condition measured by self-report. Both participant selfreport and observational coding measured the overall quality of the therapist/coach‘s facilitation
of emotional content during the therapy-like interactions. Immediately following the therapylike interaction, participants completed an adapted, 12-item version of the System for Observing
Family Therapy Alliances scale (SOFTA-S) as an indicator of effectiveness in terms of the
alliance each participant felt with the therapist/coach across the following three dimensions:
engagement in the process, emotional connection, and safety (Friedlander, Escudero, Horvath,
Heatherington, Cabero, & Martens, 2006). Reliability for the revised, 12-item version in this
sample was (α = .81), similar to the reliability found for the original, 16-item version (α = .87;
Friedlander et al., 2006). Although theoretically ranging from 12 to 60, actual participant scores
ranged from 31 to 60 (M = 50.2; Mdn = 50.0; SD = 5.67), providing evidence in terms of
participant self-report of high quality intervention implementation by the therapist/coach.
Integrity of the therapy-like condition measured by observational coding. Eight
undergraduate research associates also rated the overall quality of the therapist/coach‘s

69

implementation of the intervention as well as the couple‘s response using a six-item, multiple
choice scale for each of the four therapy-like interaction segments (segment one: partner one
talks with the therapist/coach about own issue; segment two: partner two talks with the
therapist/coach about partner‘s issue; segment three: partner two talks with therapist/coach about
own issue; segment four: partner one talks with therapist/coach about partner‘s issue; see
Appendix E for more information). Specifically, coders used a four-point Likert scale to rate
how effectively the therapist/coach (a) explored emotional experience; (b) validated emotional
experience; (c) reframed feelings and experience in terms of primary emotions and attachment
needs; and (d) highlighted and encouraged positive behaviors through the use of an enactment
(i.e., direct couple interaction). They also rated the couple‘s response to the intervention in terms
of (a) softening; and (b) the overall effectiveness of the intervention in facilitating couple
progress and change.
Training for these coders occurred over a 2-3 week period, during which coders met for
3-5 hours with the primary investigator. Both didactic and experiential training occurred,
including (a) defining the key terms and indicators for each question; (b) coding practice
segments together; (c) talking about each item; and (d) answering any questions coders had
regarding the coding process. Each week of the training process, coders were assigned to code
five research sessions. Training ended only after reliability and validity were deemed sufficient
to begin coding actual research sessions. Weekly meetings lasting one to two hours continued
throughout the coding process to review and discuss difficult-to-code sessions and answer any
coder questions. Coding of each research session lasted approximately one hour. The coders
first watched the semi-natural interactions in order to gain a basic idea of the couple‘s typical
interaction process. Next, they watched each therapy-like interaction and rated the overall

70

intervention quality with the first four questions. They then used their knowledge of the seminatural interactions to rate the overall effect of the intervention on the couple with the final two
questions.
Only the first four questions were used as a measure of intervention integrity. For
reliability purposes, 31 cases (49.2%) were coded separately by two coders. Typical reliability
statistics were not possible because of the low variability between coder scores. As a result, only
percentages of exact matches, close matches, and non-matches are reported here. Exact matches
were those where coders did not disagree. Close matches were those where both coders agreed
the therapist was in the range of proficiency, but they disagreed on whether the therapist was
extremely proficient (4) or proficient (3). Non-matches occurred when one coder rated the
therapist as proficient (3-4), while the other coder rated the therapist as non-proficient (0-2). For
the 31 double-coded cases, coders matched exactly on 82.5% of the items (n = 409), closely on
15.5% (n = 77), and did not match on 1.6% (n = 8) of items (there were also two missing items).
Because of the high match, coder scores were averaged on each item, and the aggregate score
was used to calculate the integrity of the intervention delivery. Ratings for intervention integrity
were in the proficient-extremely proficient range (M = 3.85, SD = .39, n = 1006), and therapists
were rated as non-proficient on only 2.1% of items (n = 21). As a result, overall therapist
delivery of the intervention was deemed proficient in all cases.
Measures
Continuous Measures
Physiological arousal. Participant physiological experiencing was measured using skin
conductance (also known as galvanic skin response; GSR). Psychophysiological measures
represent measurement of unique processes not under conscious awareness or control, thereby

71

increasing the potential for a more valid, sensitive measurement of individual, moment-bymoment emotional arousal. Skin conductance was specifically chosen for this study because (a)
it is a valid indicator of stress levels in the autonomic nervous system (ANS) and is not affected
by the parasympathetic nervous system (Dawson et al., 2007); (b) it is a relatively unobtrusive
measure and would not impede couple interaction; (c) it has been linked to stressful experiencing
and attempts to regulate emotions (Katz & Gottman, 1995); and (d) research has already
identified it as a valuable indicator of how attachment (especially deactivation) relates to
physiological arousal (Diamond et al., 2006; Dozier & Kobak, 1992; Roisman et al., 2004).
The apparatus consisted of a digital biofeedback device (M150, Biopac, Goleta, CA) and
an amplifier module (GSR100C, Biopac, Goleta, CA) with the gain switch set to 5.0 micromhos,
a low pass filter set to 1.0 Hz, and other filters set to provide absolute readings from each
participant. A constant voltage (0.5 V) technique was applied through two disposable adhesive
disks containing isotonic gel and silver-silver chloride electrodes that were attached to the
palmar surface of each participant‘s distal phalanges on the middle and ring fingers of the nondominant hand.
Scores for skin conductance level (SCL) were recorded 10 times per second. At the
conclusion of each research session, mean SCLs were calculated for the relaxation baseline, as
well as the baseline, semi-natural, and therapy-like interactions (see Table 2). In addition to
mean SCLs, deviation scores from the relaxation baseline were calculated for each of the
interactions. As an indicator of skin conductance response (SCR), the standard deviation was
also calculated for each interaction because it provided information regarding the wave‘s
variability, with the assumption that higher magnitude waves also would yield higher standard
deviation scores.

72

Table 2
Descriptive Statistics for Physiological Arousal and Interpersonal Distress
Variable Name
Physiological Arousal – Skin Conductance
Relaxation Baseline
Baseline Interaction
Semi-Natural Interaction
Therapy-Like Interaction
Baseline Change from Relaxation
Semi-Natural Change from Relaxation
Therapy-Like Change from Relaxation

n

M (SD)

Range

126
124
124
124
124
124
124

8.46 (4.21)
9.29 (4.13)
9.51 (4.17)
9.27 (4.27)
1.01 (.99)
1.23 (1.61)
.99 (2.25)

.79 – 31.25
.85 – 32.31
.86 – 30.28
.83 – 29.40
-1.48 – 4.10
-1.85 – 10.33
-8.46 – 10.34

Interpersonal Distress – Perception Analyzer
Baseline Interaction
Semi-Natural Interaction
Therapy-Like Interaction
Semi-Natural Change from Baseline
Therapy-Like Change from Baseline

126
124
126
124
124

75.16 (19.45)
52.63 (19.08)
61.19 (16.94)
-22.47 (21.05)
-13.97 (19.55)

4.32 – 100.00
.11 – 100.00
.50 – 98.87
-78.36 – 27.65
-60.92 – 31.92

Interpersonal distress. Interpersonal distress in couple interaction research has been
primarily measured using either global self-report or global or continuous observational coding.
Although both are important ways to understand couple interaction, neither provides continuous
construct measurement of each partner‘s perceptions of the interaction. As a result, such
measures end up being a conglomerate of the entire interaction, making it difficult to understand
more subtle occurrences or nuances. Continuous self-report (Biocca, David, & West, 1994)
represents a reliable way to understand couple interaction through interpersonal process recall
(Gardner, 2004; Gottman & Levenson, 1985, 1992; Griffin, 1993; Levenson & Gottman, 1983).
In fact, Gottman and Levenson (1985) found that couples who reported on their interaction even
one week after it occurred appeared to physiologically relive their emotional experience. It also
provides another method in making distinctions between the effects of the different attachment

73

strategies on couple interaction patterns.
Moment-to-moment self-report of feelings of connection, support, and understanding
were measured using the Perception Analyzer system (Model IV, MS Interactive, Portland, OR),
consisting of a computer and a wireless handheld electronic dial by which the participants
continuously indicated their responses to the question, ―In each moment, how did you feel
towards your partner?‖ Numbers on the dial ranged from 0 (extremely negative feelings) to 100
(extremely positive feelings), with 50 being a neutral rating. The wireless handheld electronic
device sends signals to the computer once every second, which saves the data in a file format that
can be exported to statistical software. The Perception Analyzer has been used for marketing
research, assessment of teaching effectiveness, and nutrition education (Papakonstantinou,
Hargrove, Huang, Crawley, & Canolty, 2002), and represents a potentially powerful tool in
therapeutic process research.
The continuous self-report data was divided in a similar manner to the physiological data
for SCL, with the levels calculated for the baseline, semi-natural, and therapy-like interactions
(see Table 2). This allowed for parallel statistical comparisons to be made regarding the
influence of attachment on physiological arousal and interpersonal distress. Differences from the
baseline interaction were calculated for the semi-natural and therapy-like interactions. The
variability of the data was also calculated using the standard deviation for each interaction.
Lastly, the correlation between the physiological data and perception analyzer data was
calculated for each participant during all three interactions.
Self-Report Measures
Attachment: The Experiences in Close Relationships questionnaire (ECR). With a
variety of self-report measures available to choose from and little empirical knowledge of their

74

convergence or divergence from one another, a factor analysis was performed in the late 1990‘s
on all self-report measures of attachment that had been created that decade in order to establish a
valid, internally consistent measure (Brennan et al., 1998). Results led to the development of the
Experiences in Close Relationships scale (ECR; Brennan et al., 1998) consisting of two, 18-item
subscales. Each subscale represents one of the higher order factors common to the measures
analyzed: avoidance and anxiety. One of the greatest strengths of the ECR that sets it apart from
many of the other self-report attachment measures is its strong internal consistency (α > .90) in
numerous studies (Mikulincer & Shaver, 2007). See Table 3 for a summary of the psychometric
data for all covariates used in this study.
Table 3
Psychometric Properties of Predictor Variables and Covariates
Range
n

M

SD

α

Potential

Actual

Attachment Avoidance

126

3.18

1.08

.92

1.00 – 7.00

1.06 – 6.25

Attachment Anxiety

126

3.61

1.12

.91

1.00 – 7.00

1.56 – 6.89

Feelings towards Partner

126

-11.8

18.4

.94

-90.0 – 30.0

-61.0 – 25.0

Individual Symptom Distress

126

55.5

21.0

.93

0 – 180.0

15.0 – 134.0

Conflict Frequency / Severity

126

10.6

3.42

.81*

3.0 – 18.0

3.0 – 18.0

Disruptive Conflict

126

60.3

22.9

.81*

-18.0 – 192.0

-3.0 – 116.0

Conflict Resolution

126

38.9

6.00

.81*

13.0 – 52.0

20.0 – 50.0

0 – 69.0

20.0 – 63.0

Variable Name

Relationship Satisfaction
126
46.4
8.52
.84
* α reported is for the entire CPS rather than its subscales.

Emotional tone of the relationship: The Subjective Emotional Experience Questionnaire
(SEEQ). The Subjective Emotional Experience Questionnaire is a 30-item self-report measure

75

designed to assess participants‘ emotional experiences (15 items) as well as their perceptions of
their partners‘ emotional experience (15 items). Responses range along a continuum of softened,
hostile, and withdrawn emotions and represent an adaptation of Seedall and Butler‘s (2006)
conceptual dimensions: (1) level of emotional evenness and stability (calming); (2) amount of
comprehension about self, partner, and the problem (understanding); (3) perceptions of
accessibility, receptivity, and responsiveness (closeness); (4) how much partners feel open to
express their thoughts, feelings, and concerns (safety); and (5) felt optimism concerning
relationship improvement and change (hope). A recent study (Seedall, 2008) has found strong
indications of the SEEQ‘s reliability (α = .96) and validity in terms of two measures of marital
satisfaction (r = .73 - .74) and a measure of the level of conflict resolution (r = .74).
Total scores were calculated by adding each item (1-5 Likert scale) pertaining to both self
and partner softening and then subtracting scores for withdrawal and hostility. One of the
strengths of the SEEQ‘s construction is that the wording describes couple experience for any
length of time. As a result, couples completed the SEEQ as part of the assessment packet as a
report of their relationship‘s emotional tone for the week prior to the research session (distal
measure of emotional tone). They also completed the same questionnaire when reporting on
their overall emotional tone for the semi-natural and therapy-like interactions (proximal measure
of emotional tone). Participant scores ranged from –61.0 to 25.0 on the distal measure (M = 11.8; Mdn = -9.5; SD = 18.4) and –45.0 to 30.0 on the proximal measure (M = 3.30; Mdn = 8.0;
SD = 17.4). The correlation between the two time-point measures was .49, providing some
indication that the instrument is able to detect changes in emotional tone and experience over
varying time periods.
Intrapersonal distress: Outcome Questionnaire-45.2 (OQ). Prior to participating in the

76

couple interactions, participants completed the Outcome Questionnaire-45.2 (OQ). The OQ is a
45-question, general measure of global psychological distress with questions intended to access
three primary domains: symptomatic distress, interpersonal relations, and social role (Lambert,
Burlingame, Umphress, Hansen, Vermeersch, Clouse, & Yanchar, 1996; Mueller, Lambert, &
Burlingame, 1998). It has well-established validity, including concurrent (.51 to .88) and
construct, and good reliability, including test-retest (.83) and internal consistency (.93; Mueller et
al., 1998). It also discriminates between clinical and non-clinical samples (Lambert et al., 1996).
Mean scores for participants in this study were slightly below the established cut-off for distress
(63).
Couple conflict experiences: Conflicts and Problem-Solving Scales (CPS). General
couple conflict was also measured as part of the assessment packet using the Conflicts and
Problem-Solving Scales (CPS; Kerig, 1996). The CPS assesses the following dimensions of
conflict: frequency/severity, conflict strategies employed, and resolution. Conflict
frequency/severity is measured using a two-item, six-point ordinal scale with scores ranging
from once a year or less to just about every day in regards to how frequently each partner
perceived that the couple engaged in minor or major disagreements in the past year. The conflict
strategies dimension is measured by a 46-item, 4-point Likert scale where participants describe
how often they utilized cooperation, avoidance/capitulation, stonewalling, verbal aggression,
and/or physical aggression during their conflicts. They also complete the same items related to
their partner‘s behavior. Each participant‘s scores for self and partner are averaged to come up
with one measure for each sub-scale. In addition, to establish one measure of disruptive conflict
strategies, the averages of all sub-scales were added except cooperation, which was subtracted
from the total. Finally, a 13-item scale of conflict resolution measured couple conflict outcomes.

77

Measures of the psychometric properties of the CPS have yielded satisfactory test-retest
reliability as well as satisfactory convergent and divergent validity in terms of other measures of
couple conflict and marital satisfaction (Kerig, 1996).
Relationship satisfaction. The Revised Dyadic Adjustment Scale (RDAS). The RDAS is
a 14-item instrument with seven sub-scales designed to measure adjustment in both distressed
and non-distressed relationships (Busby, Crane, Larson, & Christensen, 1995). Items represent
three potential couple dimensions: consensus (decision-making, values, and affection),
satisfaction (stability and conflict), and cohesion (activities and discussion). All but one item has
a 0 to 5 rating scale with the other item being 0 to 4. The RDAS has demonstrated strong
construct validity, criterion validity, and internal consistency/split-half reliability (Busby et al.,
1995). Mean RDAS scores for this sample were slightly below the established cut-off of 48,
indicating some relationship distress in this sample.

78

CHAPTER IV
RESULTS
The overarching objective of this study was to increase the understanding of how
attachment influences a person‘s emotional experience in terms of the level, variability, and
change from baseline of moment-to-moment physiological arousal and feelings towards one‘s
partner. In addition, it was important to understand how the relationship between attachment and
emotional experience varied according to interactional conditions. In this case, this study
compared two conditions commonly used in research on couples: a semi-natural interaction,
where the couple discussed the problem as they would at home; and a therapy-like interaction
facilitated by a therapist/coach. The following sections outline the general approach to data
analysis, the preliminary analyses used to organize and structure the data, and the statistical
analyses used to answer the primary research questions.
General Approach to Data Analysis
This study used a dyadic approach to data analysis for statistical tests of the hypotheses.
Because partners in a committed relationship were participating together, their results were more
likely to be inter-related. After many years in which ―interpersonal concepts have been studied
by examining individuals in isolation‖ (Kenny, Kashy, & Cook, 2006, p. 3), dyadic analysis has
become a more commonly used approach. Treating the individual as the unit of analysis carries
the assumption that participant scores are independent from one another. This may be the case,
except when a relationship exists between two participants that might enhance the likelihood of
their scores being similar. Discounting the potential non-independence of participant scores
increases the likelihood for Type I or II errors, depending on the analytic context. Dyadic
analysis treats partner data as non-independent by taking into account the potential influence that

79

partners have on one another‘s relationship outcomes (Kenny et al., 2006).
Several conceptual and statistical models and designs exist within the framework of
dyadic analysis. One of these, the actor-partner interdependence model (APIM; Kenny et al.,
2006; Figure 2), allows for mixed variables (i.e., both within- and between-dyads) to be analyzed
through ―simultaneous estimation‖ (West, Popp, & Kenny, 2008, p. 325) of how much a person
predicts his/her own outcome (actor effects) as well as the amount that a partner contributes to
the person‘s outcome (partner effects). In this manner, the APIM allows researchers to
statistically analyze the role of mutual influence in predicting individual outcomes (Kenny et al.,
2006). This study used mixed predictor variables measured at interval levels and treated as
distinguishable by gender. A series of multilevel models using either restricted maximum
likelihood estimation (REML; research questions one and three) or maximum likelihood
estimation (ML; research question two) was used to test how the actor and partner effects for
attachment avoidance and anxiety influence the level, variability, and change associated with
physiological arousal and interpersonal distress.
Figure 2.

The Actor-Partner Interdependence Model (APIM; Kenny et al., 2006)

a

X1

Y1

E1

Y2

E2

p
p

X2
a

a1 p21 + a2 p12 + r(a1 a2 + p12 p21)

80

Preliminary Analyses
Missing data. In order to prepare the data for the statistical tests of hypotheses using
dyadic data analysis, a variety of preliminary analyses were undertaken. The issue of missing
data was a relatively minor one. None of the 11 demographic variables or 270 self-report items
had more than 5% of responses missing, with 81.8% of demographic variables (n = 9) and 81.5%
of self-report items (n = 220) having less than 1% missing. For all missing self-report data, the
SYSTAT EM Method was used to estimate values using maximum likelihood. Little‘s MCAR
test statistic for each series of imputed values used in this study‘s analyses was non-significant,
indicating that the null hypothesis of data missing completely at random was retained. No
demographic data were imputed.
Predictor variables. Preliminary analyses also helped examine the relationship between
attachment (ECR; avoidance and anxiety) and the other variables measured in this study (Table
4). Relationships were as expected, with self-reported attachment avoidance and anxiety
negatively correlated with conflict resolution strategies, relationship satisfaction (RDAS), and
global feelings towards one‘s partner (SEEQ). Attachment avoidance and anxiety were also both
positively correlated with disruptive conflict (CPS) and intrapersonal distress (OQ-45.2).
Table 4
Bivariate Correlations between Predictor Variables and Covariates (n = 126 for all)

Symp.
Distress
--

Freq/Sev

Conflict
Disrupt.

Resolve

--

Partner
Feelings
--

--

--

--

.12

--

--

--

--

--

--

-.45***

-.31***

--

--

--

--

--

Avoid.

Anx.

Avoidance

--

Anxiety
Partner
Feelings

81

Table 4 (cont‘d)

Avoid.

Symp.
Distress
--

Freq/Sev
--

--

--

Symptom
Distress

.45*** 53***

Partner
Feelings
-.44***

Conflict Freq.
/ Sev.

.35***

.39***

-.55***

.46***

--

--

--

Disruptive
Conflict

.37***

.28**

-.71***

.46***

.55***

--

--

Conflict
Resolution

-.40***

-.14

.72***

-.36***

-.53***

-.66***

--

-.31***

.76***

-.52***

-.54***

-.66***

.72***

Relationship
-.49***
Satisfaction
** p < .01 *** p < .001

Anx.

Conflict
Disrupt.

Avoid.

Outcome variables. This study focused on emotional experience in terms of moment-tomoment physiological arousal as well as feelings towards one‘s partner. Research questions
addressed emotional experience across two primary conditions: a semi-natural, low structure
condition where partners talked as they would at home and a therapy-like, high structure
condition facilitated by a therapist/coach. Emotional experience data were also measured for a
baseline interaction where partners reminisced with each other about how they met and how their
relationship progressed to greater commitment. A relaxation baseline was also established for
partners in terms of their physiological arousal. Because so many individual factors and
biological characteristics vary within and across individuals that influence the actual skin
conductance level (Dawson et al., 2007; Venables & Mitchell, 1996) and make meaningful
interpretation difficult, ―residualizing‖ scores by controlling for baseline skin conductance levels
in the variability and change from baseline analyses was especially important (Diamond et al.,
2006). As a result, the physiological relaxation baseline was used as a covariate in all variability

82

and change analyses with a physiological outcome variable. Although physiological arousal was
also measured during a cool-down period, it was not included in the primary analyses because it
did not involve structured couple interaction that had a parallel with the continuous self-report
data and it was a measure of return to baseline rather than a baseline condition.
Dyadic data preparation. To prepare for dyadic data analysis, a pairwise data set was
created, categorical variables were effect coded (e.g., gender was coded -1 for men and +1 for
women), and all continuous predictor variables without a meaningful zero-point were grand
mean centered (see Kenny et al., 2006). In addition, a bivariate correlation between men and
women for each variable was calculated as a measure of non-independence (Table 5). All but
one correlation (see Cohen, 1988) between men and women for attachment and physiological
arousal were small (.3 > r > .1) or trivial (r < .1). The one exception was a medium-sized (.5 > r
> .3) correlation between men and women for skin conductance change between the relaxation
baseline and baseline interactions. Correlations between men and women for the level of
moment-to-moment feelings towards partner were medium, ranging from .37 to .43. However,
the correlations for moment-to-moment change from the interaction baseline were small for both
the semi-natural (r = .18) and therapy-like (r = .22) conditions. Lastly, individual symptom
distress, couple conflict, conflict resolution, satisfaction, and feelings towards partner during the
communication segments yielded medium to large (r > .5) correlations ranging from .32 to .73.
Overall, results of these preliminary analyses indicated at least some non-independence of the
data, especially when considering the large number of dyads that are typically required to
generate sufficient power to detect non-independence (Kenny et al., 2006).

83

Table 5
Bivariate Correlations between Data for Men and Women
Variable Name
Attachment Avoidance

n (dyads)
63

r
.23

Attachment Anxiety

63

.01

Individual Symptom Distress

63

.32*

Conflict: Frequency / Severity

63

.48***

Conflict: Disruptive Conflict

63

.61***

Conflict: Conflict Resolution

63

.65***

Relationship Satisfaction

63

.73***

Global Feelings towards Partner

63

.62***

Proximal Feelings towards Partner

63

.64***

Physiological Arousal – Skin Conductance
Relaxation Baseline
Baseline Interaction
Semi-Natural Interaction
Therapy-Like Interaction
Baseline Change from Relaxation
Semi-Natural Change from Relaxation
Therapy-Like Change from Relaxation

63
62
62
62
62
62
62

.27*
.20
.21
.21
.37**
.22
.19

63
62
63
62
63

.37**
.39**
.43***
.18
.22

Interpersonal Distress – Perception Analyzer
Baseline Interaction
Semi-Natural Interaction
Therapy-Like Interaction
Semi-Natural Change from Baseline
Therapy-Like Change from Baseline
*p < .05 **p < .01 ***p < .001

Demographic variable inclusion. In order to get a sense of what demographic variables,
in addition to gender, might be influencing outcomes, demographic variables were grouped
conceptually and compared through a series of analyses (ANOVA or regression). The

84

Bonferroni method was used for post hoc comparisons in the ANOVAs. This made it possible to
select influential demographic variables that would be included as covariates in the analyses (see
Table 6 for ANOVA and regression results). As a result of the preliminary analyses, four
demographic variables were chosen that demonstrated the strongest relationships to the outcome
variables: education, age, race/ethnicity, and relationship status. Religion was not retained
because it did not significantly predict any of the outcome variables.
Lower education was associated with more negative outcomes and appeared to be more
predictive than income or employment. Taken as a whole, age also appeared to be slightly more
predictive than either relationship time or number of children with respect to skin conductance
level during all communication segments and the difference in interpersonal distress between the
therapy-like and semi-natural interactions, with higher ages being associated with lower skin
conductance levels and less difference between the therapy-like and semi-natural interactions.
Preliminary results for ethnicity and relationship status showed some significant associations
with the outcome variables, making it important to retain them. In order to increase the
meaningfulness of the results, both ethnicity (Caucasian/Euro-American: n = 74, 58.7%; underrepresented minority: n = 52, 41.3%) and relationship status (committed, non-marriage
relationship: n = 54, 42.9%; married: n = 72, 57.1%) were transformed into dichotomous
variables and effect coded (-1 and 1).
Table 6a
One-Way ANOVA Omnibus F Values for Categorical Demographic Predictors of Physiological
Arousal
+

Demographic Baseline SN
TL
Baseline
SN SD TL SD
Variable
Mean Mean
SD
Mean
Gender
.08
.15
.001 4.7*
9.5** 7.5**
Children
5.2**
5.7** 5.2** .13
.92
2.1

85

SN Baseline
.17
.45

TL Baseline
.48
.07

Table 6a (cont‘d)
+

Demographic Baseline SN
TL
Baseline
SN TL SN SD TL SD
Variable
Mean Mean
SD
Baseline Baseline
Mean
Religion
.43
.48
.55
.70
.69
.61
.15
.65
Income
3.6*
2.8
2.0
1.9
1.8
1.7
.46
.83
Education
3.4*
3.2*
2.7
.66
.01
.27
.27
.80
Employment
.03
.10
.37
.11
.16
1.6
.36
1.5
Relationship 5.4*
3.6
2.2
4.7*
.08
.32
1.9
2.9
Ethnicity
.68
1.2
1.3
1.3
.18
.04
1.1
.54
* p < .05 ** p < .01 *** p < .001
SN = Semi-Natural TL = Therapy-Like
+
The baseline referred to here is baseline interaction.
Table 6b
Linear Regression Standardized Beta Values for Continuous Demographic Predictors of
Physiological Arousal
+

Baseline
Mean
-.40***
-.31***

SN
Mean

TL
Mean

Baseline
SD

SN
SD

TL
SD

SN Baseline

TL Baseline

Age
-.39***
-.36***
-.19*
-.20* -.20*
.04
.07
Relate
-.31***
-.29**
-.16
-.15
-.15
-.01
.02
Time
* p < .05 ** p < .01 *** p < .001
SN = Semi-Natural TL = Therapy-Like
+
The baseline referred to here is baseline interaction.
Table 6c
One-Way ANOVA Omnibus F Values for Categorical Demographic Predictors of Moment-toMoment Feelings towards Partner
Demographic Baseline SN
TL
Variable
Mean
Mean Mean
Gender
4.9*
.29
1.8
Children
3.1*
4.6*
4.4*
Religion
.43
1.1
2.0
Income
4.7*
.76
.32
Education
4.6*
.51
.83
Employment
.04
.33
.52
Relationship
.88
4.1*
.42
Ethnicity
3.4
.001
.58
* p < .05 ** p < .01 *** p < .001

Baseline SN
TL
SN TL SD
SD
SD
Baseline
Baseline
.003
4.5* 4.4*
6.5*
12.1**
.89
.06
.81
4.0*
1.8
.57
.40
.60
.87
1.5
7.1**
.40 2.4
1.9
3.1*
6.2**
.12
.05
3.9*
7.0**
2.2
.08 2.7
.46
.32
5.9*
.17 3.6
.94
.14
10.6**
3.7
4.7*
2.9
1.3
SN = Semi-Natural TL = Therapy-Like

86

Table 6d
Linear Regression Standardized Beta Values for Continuous Demographic Predictors of
Moment-to-Moment Feelings towards Partner
+

Baseline
Mean
-.11
.11

Age
Relate
Time

SN
Mean
.08
.17

TL
Mean
-.09
-.05

Baseline
SD
-.04
-.12

* p < .05 ** p < .01 *** p < .001

SN SD

TL SD

-.09
-.01

-.07
-.09

SN = Semi-Natural

SN Baseline
.17
.05

TL Baseline
.03
-.15

TL = Therapy-Like

Decision process for the analysis of demographic variables, covariates, and interactions.
Each analysis initially included gender as a distinguishing variable as well as the aforementioned
demographic variables and covariates. The following decision-making process guided the
retention or removal of variables to increase parsimony and power (see Figure 3). If there were
significant interactions and no main effect involving gender, it was retained as a distinguishing
variable. However, if there was no significant interaction or main effect, then a deviance test
analyzed whether the additional complexity gender introduced as a distinguishing variable was
justified. The deviance test consisted of running the same analysis twice using maximum
likelihood estimation, with one version treating the data as distinguishable by gender and the
other treating the data as non-distinguishable. A chi-square test then compared the loglikelihood results for each, with a significant difference indicating that the complexity introduced
by gender as a distinguishing variable was justified. If there were no significant gender
interactions or a main effect, then gender was removed and the data treated as nondistinguishable. If the deviance tests indicated that the additional complexity introduced by
gender was justified, then only the non-significant gender interactions were removed from the
analysis.

87

Figure 3.

General Approach to Data Analysis
All demographic
variables, covariates, and
interactions included

NO
Is there a significant
interaction
involving gender?

Remove
gender
interactions

YES
Gender is
retained as a
distinguishing
variable

Significant gender
main effect?

Significant chisquare deviance
test comparing
log-likelihoods?

Remove all nonsignificant
covariates and
demographic
variables from
analysis

NO
Remove
gender
main
effect

NO
Remove gender as
distinguishing
variable

Disruptive conflict (CPS), global feelings towards one‘s partner (SEEQ), and individual
symptom distress (OQ) were the three covariates, in addition to the demographic variables,
included in the analyses. These three variables were chosen because they offered a broad

88

representation of factors that might influence, in addition to attachment avoidance and anxiety,
proximal measures of physiological arousal and interpersonal distress. Non-significant main
effects of both covariates and demographic variables were removed from the analyses. No
interactions between covariates or demographic variables and gender were analyzed. However,
if a main effect was found for any covariate or demographic variable, interactions between the
variable and attachment avoidance and anxiety were added to the analysis and tested.
Two methods helped further probe significant two-way interactions. If the interaction
involved gender and attachment, then a post hoc analysis was conducted within the context of the
APIM. Two variables (man, woman) with dummy codes for each (0 and 1) were created and
substituted in the analysis for gender. This generated interactions between man and woman
separately with actor and partner attachment avoidance and anxiety. For significant interactions
involving a continuous predictor and a different categorical variable or two continuous variables,
a high and low value for each predictor variable (+1 standard deviation above the mean and -1
standard deviation below the mean) was substituted into the original regression equation (Cohen,
Cohen, West, & Aiken, 2003; Keith, 2006; West, 1991). The simple slope of each resulting
regression line was then analyzed for significance within the same dyadic framework as the
original analysis.
Statistical Tests of Hypotheses
Research Questions and Their Analyses
Research Question 1: Are higher levels of self-reported attachment avoidance and/or
anxiety associated with a more negative emotional experience for partners in each of the
interactional conditions?
Level of physiological arousal (see Table 7). For the purposes of this study, the level of

89

physiological arousal refers to the moment-to-moment skin conductance occurring during each
interactional condition. Increases in skin conductance indicate higher stress or physiological
arousal. The actual level of skin conductance has been shown to vary according to a variety of
individual factors, including where the electrodes are placed on the hand, characteristics relating
to the hand‘s surface (e.g., cuts or calluses), ambient temperature, and time of day (Dawson et
al., 2007). An established way to ―residualize‖ results is to use a relaxation baseline as a
covariate in all analyses (Diamond et al., 2006). The analyses of skin conductance variability
and change from baseline used this method of ―residualizing‖ results, while raw level scores
were used in the analysis involving the level of physiological arousal. Accordingly, results will
be reported in this section for comparison purposes rather than for the purpose of developing
specific implications.
For each communication segment, there was no significant interaction or main effect
involving gender. In addition, the data were treated as non-distinguishable because deviance
tests revealed that the additional complexity introduced by distinguishing the data by gender was
not justified. The only significant main effects involving attachment were for partner attachment
anxiety during the semi-natural, b = .65, t(110.6) = 1.96, p = .05, and therapy-like, b = .81,
t(110.2) = 2.36, p = .02, communication segments, with higher skin conductance levels
associated with higher levels of attachment anxiety reported by that person‘s partner. There
were also significant age main effects for the baseline, b = -.16, t(62.9) = -4.51, p < .001; seminatural, b = -.16, t(63.4) = -4.25, p < .001; and therapy-like, b = -.15, t(63.5) = -3.85, p < .001
communication segments. Specifically, younger age was associated with higher skin
conductance levels. Lastly, no interaction between age and attachment was significant when
included in the analyses for any of the segments.

90

Thus, when raw level scores were used for skin conductance, one of the most salient
individual characteristics that influenced outcome was age, with increased age associated with
lower levels of physiological arousal. This may be related to the long-established finding that
thermoregulation changes as people age, including decreases in palmar sweat (Ferreira & Winter,
1965). It might also coincide with the finding that suppression as a regulatory strategy decreases
with age (John & Gross, 2004). In addition, partner attachment anxiety also related to
physiological arousal during the therapy-like communication segment. Higher levels of
attachment anxiety appeared linked to increases in physiological arousal for the other partner,
even though the couple interaction was being filtered through the therapist/coach.
Table 7
Research Question 1 – Level of Physiological Arousal (Non-Residualized)
Communication
Segment

+

b

Predictor Variables

Avoidance (A)
Avoidance (P)
Baseline
Anxiety (A)
Interaction
Anxiety (P)
Age
Avoidance (A)
Avoidance (P)
Semi-Natural
Anxiety (A)
Interaction
Anxiety (P)
Age
Avoidance (A)
Avoidance (P)
Therapy-Like
Anxiety (A)
Interaction
Anxiety (P)
Age
* p < .05 ** p < .01 *** p < .001

-.28
-.05
.27
.48
-.16***
-.35
-.06
.29
.65*
-.16***
-.42
-.22
.37
.81*
-.15***

SE

df

t

.35
.35
.33
.33
.04
.35
.35
.33
.33
.04
.36
.36
.34
.34
.04

112.1
112.2
111.5
111.7
62.9
112.9
113.0
110.3
110.6
63.4
113.1
113.2
109.9
110.2
63.5

-.81
-.15
.82
1.45
-4.51
-1.00
-.18
.86
1.96
-4.25
-1.17
-.61
1.08
2.36
-3.85

-.07
-.01
.07
.13
-.39
-.09
-.02
.08
.18
-.38
-.11
-.05
.10
.21
-.34

+

The following variables were removed from the final analysis for all communication
segments due to non-significance: Gender, Symptom Distress (A, P), Conflict (A, P), and Global
Partner Feelings (A, P), Race / Ethnicity, Relationship Status, and Education.
Variability of physiological arousal (see Table 8). Variability of skin conductance level

91

was measured in this study using the standard deviation scores. Past attachment literature has
analyzed physiological arousal using skin conductance level (SCL), which refers to levels and
their change from baseline. Skin conductance response (SCR), however, focuses primarily on
the fluctuations in skin conductance, including the number, magnitude, and/or length of changes
in skin conductance (Andreassi, 2007). Standard deviation represents an approximate indicator
of skin conductance response because it provides a value representing change in skin
conductance over the course of each interaction. In this manner, the standard deviation permitted
an analysis of whether attachment influenced skin conductance level as well as its fluctuations
Although there was no significant interaction or main effect involving attachment, there
were significant gender main effects for the baseline, b = .09, t(55.8) = 2.95, p < .01, and seminatural, b = .11, t(59.7) = 3.69, p < .001, communication segments. For each of these segments,
higher variability in physiological arousal was exhibited by women. There were also significant
main effects for disruptive couple conflict reported by one‘s partner, b = .01, t(71.1) = 3.79, p <
.001, and relationship status, b = -.10, t(59.2) = -3.36, p = .001, during the baseline
communication segment. Greater variability in an individual‘s physiological arousal was
associated with higher reported levels of disruptive conflict by that person‘s partner as well as
being in a committed, non-marriage relationship. There was no significant interaction between
attachment and relationship status or disruptive couple conflict reported by one‘s partner when
they were added to the model. There were also no significant main effects besides gender for the
semi-natural communication segment.
For the therapy-like communication segment, there was a significant interaction between
gender and partner attachment anxiety, b = .06, t(86.5) = 2.03, p < .05. A post hoc analysis that
created two variables (man, woman) with dummy codes for each was used to generate

92

interactions between attachment and each gender. Results showed a significant effect of partner
attachment anxiety for women, b = .11, t(123.2) = 2.20, p = .01, with more variability in
physiological arousal for women whose partners reported higher levels of attachment anxiety
(see Figure 4). However, the effect of attachment anxiety for men was not significant, b = -.02,
t(123.2) = -.51, p = .61. Although the results of this study did not link attachment and the
variability of physiological arousal during the baseline or semi-natural communication segments,
women whose partners reported higher levels of attachment anxiety demonstrated more
variability in their physiological arousal during the therapy-like segment. This indicates that for
women, having a male partner with a more anxious attachment strategy leads to greater
fluctuations in physiological arousal during a therapy-like setting.
Figure 4.

Research Question 1 – The Effect of Attachment Anxiety and Gender on the
Variability of Physiological Arousal during the Therapy-Like Interaction

Male
Attachment
Anxiety

Male‘s
Physiological
Variability

E1

Female‘s
Physiological
Variability

-.03

E2

.11**
-.02
Female
Attachment
Anxiety

.05

* p < .05 ** p < .01 *** p < .001

93

Table 8
Research Question 1 – Variability of Physiological Arousal
Communication
Segment

+

b

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Baseline
Anxiety (A)
Interaction
Anxiety (P)
Relaxation Baseline Level
Conflict (P)
Relationship Status
Gender
Avoidance (A)
Avoidance (P)
Semi-Natural Anxiety (A)
Interaction
Anxiety (P)
Relaxation Baseline Level
Conflict (P)
Relationship Status
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Therapy-Like Relaxation Baseline Level
Interaction
Conflict (P)
Relationship Status
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
* p < .05 ** p < .01 *** p < .001

.09**
.01
-.02
-.06
-.05
.03
.04***
-.10***
.11***
.01
.01
-.01
.01
.04***
--.10***
-.01
-.02
.01
.05
.03***
--.002
-.05
.04
.06*

SE

df

t

.03
.03
.03
.03
.03
.01
.001
.03
.03
.03
.03
.03
.02
.01
--.03
.03
.03
.03
.03
.01
--.03
.03
.03
.03

55.8
95.2
79.6
104.8
69.2
76.5
71.1
59.2
59.7
104.6
89.0
113.4
82.1
84.7
--54.1
107.9
95.2
112.3
86.2
87.8
--104.9
94.4
112.7
86.5

2.95
.50
-.87
-1.91
-1.82
4.14
3.79
-3.36
3.69
.28
.38
-.22
.57
5.51
--3.45
-.52
-.55
.50
1.63
5.00
--.06
-1.85
1.60
2.03

.23
.04
-.07
-.16
-.14
.32
.31
-.27
.29
.02
.03
-.02
.04
.41
--.28
-.04
-.05
.04
.15
.40
--.005
-.16
.14
.19

+

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A), Global Partner Feelings (A, P),
Age, Race / Ethnicity, and Education.
Change in physiological arousal (see Table 9). For this study, change in physiological
arousal refers to the moment-to-moment skin conductance level changes from the relaxation

94

baseline that occurred during each interactional condition. There was a significant interaction
between gender and partner attachment anxiety for the baseline, b = .23, t(82.3) = 2.73, p < .01;
semi-natural, b = .40, t(80.4) = 2.64, p = .01; and therapy-like, b = .49, t(81.8) = 2.29, p = .02,
communication segments. Post hoc analyses using man/woman dummy codes showed a
significant effect of partner attachment anxiety for women‘s changes in skin conductance during
the baseline, b = .35, t(108.1) = 2.83, p < .01; semi-natural, b = .78, t(119.7) = 3.52, p = .001;
and therapy-like, b = 1.01, t(120.3) = 3.27, p = .001, communication segments, with greater
increases in skin conductance levels for women whose partners reported higher levels of
attachment anxiety (see Figure 5). However, no significant effect of attachment for men was
found during the baseline, b = -.12, t(108.6) = -1.27, p = .21, semi-natural, b = -.05, t(119.9) = .29, p = .77; and therapy-like, b = .03, t(120.5) = .13, p = .90, communication segments. This
finding provides evidence that the increase in physiological arousal for women whose male
partners were higher in attachment anxiety occurs across interactional conditions, whether the
couple is discussing a positive issue or an emotionally charged issue by themselves or with a
therapist/coach.

95

Research Question 1 – The Effect of Attachment Anxiety and Gender on Changes
in Physiological Arousal during the Baseline / Semi-Natural / Therapy-Like
Interactions

Figure 5.

Male
Attachment
Anxiety

.17 / .12 / .14

Male‘s
Physiological
Change

E1

Female‘s
Physiological
Change

E2

.35**/ .78*** / 1.01***
-.12 / -.05 / .03
Female
Attachment
Anxiety

-.09 / -.03 / .19

* p < .05 ** p < .01 *** p < .001
For the baseline communication segment, there were also interactions between age and
attachment avoidance, b = .02, t(94.2) = 2.12, p = .04. A post hoc analysis revealed a significant
effect of attachment avoidance for young individuals (-1 SD), b = -.24, t(101.4) = -2.06, p = .04,
but not for older individuals (+1 SD), b = .10, t(101.6) = .91, p = .36 (see Figure 6). For younger
individuals, high levels of attachment avoidance were associated with smaller changes in skin
conductance levels from the relaxation baseline to the baseline communication segment. Thus,
when controlling for relaxation baseline levels of skin conductance, age did not have a
significant association with skin conductance, except during the baseline communication
segment when partners reminisced about how they met. During that segment, younger
individuals who were high in attachment avoidance evidenced smaller changes in skin
conductance.

96

Figure 6.

Research Question 1 – The Effect of Attachment Avoidance and Age on
Physiological Arousal during the Baseline Interaction

b = -.24, t(101.4) = -2.06, p = .04

b = .10, t(101.6) = .91, p = .36

Table 9
Research Question 1 – Change from Relaxation Baseline of Physiological Arousal
(Residualized)
Communication
Segment

Baseline
Interaction

+

b

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Relaxation Baseline Level
Age
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)

.24**
-.07
.07
.04
.12
.02
-.02*
-.12
.10
-.13
.23**

97

SE

df

.07
.07
.08
.08
.08
.02
.01
.08
.08
.08
.08

53.5
101.8
93.3
89.6
80.5
79.0
67.6
85.1
83.6
90.9
82.3

t
3.29
-.91
.84
.50
1.42
.73
-2.24
-1.51
1.11
-1.62
2.73

.24
-.08
.07
.05
.14
.06
-.22
-.14
.10
-.15
.26

Table 9 (cont‘d)
Communication
Segment

+

b

Predictor Variables

Age * Avoidance (A)
Age * Avoidance (P)
Age * Anxiety (A)
Age * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Semi-Natural
Relaxation Baseline Level
Interaction
Age
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Therapy-Like
Relaxation Baseline Level
Interaction
Age
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
* p < .05 ** p < .01 *** p < .001

.02*
-.003
-.02*
-.01
.34*
-.08
.01
.06
.37*
.97***
--.23
.03
-.08
.40**
.19
-.12
-.16
.17
.52*
.92***
--.33
-.11
.02
.49*

Baseline
Interaction
(cont‘d)

SE

df

t

.01
.01
.01
.01
.14
.14
.14
.14
.15
.03
-.14
.15
.13
.15
.19
.19
.20
.19
.21
.05
-.20
.21
.19
.21

94.2
81.7
99.0
95.4
56.8
103.9
90.1
107.0
79.9
83.5
-94.2
84.7
107.6
80.4
56.9
104.8
91.3
107.9
81.2
84.8
-95.5
86.0
108.5
81.8

2.12
-.36
-1.96
-1.83
2.48
-.60
.08
.41
2.44
29.7
--1.65
.18
-.57
2.64
1.01
-.65
-.78
.87
2.45
20.2
--1.66
-.55
.12
2.29

.18
-.03
-.17
-.17
.21
-.05
.01
.04
.26
-.09
--.15
.02
-.05
.28
.09
-.06
-.07
.08
.26
-.14
--.16
-.05
.01
.24

+

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A, P), Global Partner Feelings (A,
P), Race / Ethnicity, Relationship Status, and Education.
Level of moment-to-moment feelings towards one’s partner (see Table 10). For the
purposes of this study, the level of feeling towards one‘s partner refers to how much moment-tomoment positive versus negative feelings towards one‘s partner were felt by an individual.
Level is conceptually similar to the level of physiological arousal, with the primary difference
being that the level of feeling towards one‘s partner provides meaningful information without

98

controlling for baseline levels. As a result, baseline level of feelings towards one‘s partner was
used as a covariate only in the analysis of change in baseline with respect to feelings toward
one‘s partner.
In terms of the level of feeling towards one‘s partner during the baseline communication
segment, there were significant main effects for partner avoidance, b = -2.94, t(114.9) = -2.14, p
= .03; and actor anxiety, b = 3.16, t(88.6) = 2.43, p = .02. Higher levels of attachment avoidance
reported by an individual were associated with lower levels of positivity felt by that person‘s
partner, whereas higher levels of attachment anxiety reported by a person actually were
associated with more positivity felt by that person towards his/her partner during the baseline
segment. There were also significant main effects for gender, b = 3.45, t(61.5) = 2.46, p = .02;
education, b = 5.03, t(88.6) = 2.21, p = .03; and disruptive couple conflict, b = -.36, t(88.6) = 5.21, p < .001. Being a woman, higher levels of education, and lower levels of disruptive
conflict were associated with higher levels of positivity felt towards one‘s partner during the
baseline communication segment. There was no significant interaction between education or
conflict and attachment.
There was no significant interaction involving attachment for either segment, but there
was a significant main effect for attachment anxiety, b = 3.74, t(101.5) = 2.76, p < .01, during
the therapy-like communication segment. Similar to the baseline communication segment,
higher levels of attachment anxiety were associated with higher levels of positivity felt towards
one‘s partner. For both the semi-natural, b = -.38, t(112.4) = -4.68, p < .001, and therapy-like, b
= -.32, t(114.6) = -4.41, p < .001, communication segments, higher levels of disruptive conflict
were significantly associated with less positivity felt towards one‘s partner. However, there was
no significant interaction between conflict and attachment when tested as part of the model.

99

Thus, during positive communication and more highly structured, emotionally-focused
communication, attachment anxiety contributed to more positive feelings towards one‘s partner.
Attachment avoidance appeared to dampen positive feelings experienced by partners during
couple communication centered on a positive topic.
Table 10
Research Question 1 – Level of Moment-to-Moment Feelings
Communication
Segment

+

b

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Baseline
Anxiety (A)
Interaction
Anxiety (P)
Conflict (A)
Relationship Status
Education
Gender
Avoidance (A)
Avoidance (P)
Semi-Natural Anxiety (A)
Interaction
Anxiety (P)
Conflict (A)
Relationship Status
Education
Gender
Avoidance (A)
Avoidance (P)
Therapy-Like Anxiety (A)
Interaction
Anxiety (P)
Conflict (A)
Relationship Status
Education
* p < .05 ** p < .01 *** p < .001

3.45*
-2.36
-2.94*
3.16*
.41
-.36***
-5.03*
-.69
-1.58
1.39
1.26
-.38***
3.46
--.38
-.71
3.74**
1.11
-.32***
---

+

SE

df

t

1.40
1.38
1.37
1.30
1.37
.07
-2.28
-4.58
1.51
1.54
1.52
.08
1.81
--1.40
1.35
1.36
1.34
.07
---

61.5
101.3
114.9
88.6
114.7
88.6
-88.6
-110.8
110.6
95.7
95.0
112.4
75.9
--115.6
115.7
101.5
101.0
114.6
---

2.46
-1.71
-2.14
2.43
.30
-5.21
-2.21
-.44
-1.04
.90
.83
-4.68
1.91
--.27
-.53
2.76
.83
-4.41
---

.18
-.13
-.16
.18
.02
-.42
-.26
-.04
-.09
.08
.07
-.46
.18
--.02
-.05
.25
.07
-.43
---

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (P), Global Partner Feelings (A, P),
Age, and Race / Ethnicity.

100

Proximal self-report of the level of feelings towards partner. A proximal self-report
measure of feelings towards one‘s partner during all communication segments was also obtained
using the Subjective Emotional Experience Questionnaire (SEEQ). Although there was not a
significant gender main effect, there were significant gender interactions with actor avoidance, b
= -2.81, t(73.5) = -2.23, p = .03, and anxiety, b = -3.76, t(79.0) = -3.19, p < .01. A post hoc
analysis using man/woman dummy codes showed a significant effect for men and their own
attachment anxiety, b = 7.24, t(90.8) = 4.02, p < .001, and for women and their own attachment
avoidance, b = -4.39, t(93.5) = -2.86, p < .01 (see Figures 7 and 8). Higher levels of reported
attachment anxiety in men were associated with more positive feelings towards one‘s partner
during the communication segments. In addition, attachment avoidance in women was
associated with more negative feelings towards one‘s partner during the segments as a whole.
Figure 7.

Research Question 1 – The Effect of Attachment Avoidance and Gender on
Proximal Feelings towards Partner for all Combined Interactions

1.22

Male
Attachment
Avoidance

Male‘s Global
Partner
Feelings

E1

Female‘s
Global Partner
Feelings

E2

-.74
-4.71**
Female
Attachment
Avoidance

-4.39**

101

Figure 8.

Research Question 1 – The Effect of Attachment Anxiety and Gender on
Proximal Feelings towards Partner for all Combined Interactions

7.24***
Male‘s Global
Partner
Feelings

E1

Female‘s
Global Partner
Feelings

Male
Attachment
Anxiety

E2

3.08
2.66
Female
Attachment
Anxiety

-.29

There were also significant main effects for partner avoidance, b = -2.74, t(90.4) = -2.44,
p = .02, and partner anxiety, b = 2.84, t(79.4) = 2.37, p = .02. A person‘s attachment avoidance
was associated with more negative feelings, while attachment anxiety was associated with more
positive feelings towards one‘s partner. Lastly, there were significant main effects for
relationship status, b = 7.18, t(88.0) = 2.57, p = .01, as well as for disruptive couple conflict, b =
-.39, t(107.5) = -6.91, p < .001, with being in a marriage relationship and lower levels of couple
conflict associated with more emotional safety felt during the communication segments taken as
a whole. However, there were no significant interactions between any of these variables and
attachment when tested as part of the model. Overall, attachment-related results for a proximal
measure of feelings towards one‘s partner are similar to those reported moment-to-moment.
Attachment avoidance in women was linked to more negative feelings towards the man, whereas
attachment avoidance in an individual was associated with more negative feelings felt by his/her
partner. Conversely, attachment anxiety led to more positive feelings for anxious men, and
partners of those higher in anxiety reported more positive feelings when rating the

102

communication segments as a whole.
Variability of moment-to-moment feelings towards one’s partner (see Table 11). For
moment-to-moment feelings towards one‘s partner, the construct of variability is parallel to
variability of moment-to-moment physiological arousal. In other words, it refers to the amount
each person‘s responses deviate from his/her mean level of feelings towards one‘s partner (i.e.,
the standard deviation). There was no significant gender main effect or interaction for the
baseline communication segment, and the additional complexity introduced from distinguishing
the data by gender was not justified. As a result, gender was removed from the analysis.
Although there was no significant interaction or main effect involving attachment during the
original analysis, a significant interaction between partner attachment avoidance and global
feelings towards one‘s partner, b = .08, t(109.1) = 2.19, p = .03, was found when it was added to
the model.
A post hoc simple slopes analysis revealed a significant effect of partner feelings for
lower levels, b = -.24, t(109.8) = -3.50, p = .001, but not higher levels, b = -.07, t(104.5) = -1.19,
p = .24, of attachment avoidance (see Figure 9). In this manner, for those with lower levels of
attachment avoidance, more positive global feelings reported by their partner were associated
with less variability in their partner‘s moment-to-moment feelings during the baseline
communication segment. Although the relationship direction was similar for those with higher
attachment avoidance, the simple slope test was not statistically significant. In addition to global
feelings towards one‘s partner, there were significant main effects for ethnicity, b = 1.98, t(85.5)
= 2.47, p = .02, and education, b = -3.81, t(88.0) = -3.10, p < .01. Both identifying as a member
of a racially non-white minority group and lower education were associated with greater
variability in feelings towards one‘s partner during the baseline communication segment. There

103

was no significant interaction between ethnicity or education and attachment when analyzed.
Figure 9.

Research Question 1 – The Effect of Attachment Avoidance and Global Feelings
towards Partner on the Variability of Moment-to-Moment Feelings during the
Baseline Interaction

b = -.24, t(109.8) = -3.50, p = .001

b = -.07, t(104.5) = -1.19, p = .24

For the semi-natural communication segment, there was no significant main effect or
interaction involving attachment. However, there were significant main effects for gender, b =
1.59, t(60.7) = 2.18, p = .03, and ethnicity, b = 1.66, t(83.2) = 2.41, p = .02, with being a woman
and identifying as a member of a non-white minority group being associated with greater
variability in moment-to-moment feelings toward one‘s partner. For the therapy-like
communication segment, there was no significant interaction involving attachment, but there was
a significant main effect for attachment anxiety, b = 1.49, t(117.9) = 2.15, p = .03, with higher
levels of anxiety also associated with more variability. Similar to the results for the other two
segments, individuals who identified as members of a non-white minority group were more
likely to have greater variability in their feelings toward their partner during the therapy like
interaction, b = 2.00, t(80.6) = 2.82, p < .01. However, there were no significant interactions

104

between attachment and any demographic variable or covariate when added to the model.
In sum, variability in feelings towards one‘s partner differed by condition. For those
lower in attachment avoidance, more positive global feelings reported by their partner led to less
variability in the partner‘s feelings during the baseline segment. This provides some evidence
that having a partner lower in attachment avoidance leads to more consistent feelings towards
that partner during a neutral to positive relationship discussion. During the therapy-like
communication segment, higher levels of attachment anxiety were also associated with greater
variability, indicating that, even though attachment anxiety appeared related to more positive
feelings towards one‘s partner during the therapy-like communication segment, hyperactivating
tendencies are associated with more fluctuations in feelings towards one‘s partner.
Table 11
Research Question 1 – Variability of Moment-to-Moment Feelings
+

Segment

Baseline

SemiNatural

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Global Partner Feelings (A)
Global Partner Feelings (P)
Race / Ethnicity
Education
Global Feelings (P) * Avoidance (A)
Global Feelings (P) * Avoidance (P)
Global Feelings (P) * Anxiety (A)
Global Feelings (P) * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)

Table 11 (cont‘d)

105

b
--.42
-.20
-.72
-1.01
--.15**
1.98*
-3.81**
-.04
.08*
-.06
-.01
1.59*
-.63
-.03
.65
.33

SE
-.78
.80
.75
.75
-.05
.80
1.23
.04
.04
.04
.04
.83
.68
.66
.65
.62

df
-112.7
113.7
111.0
111.2
-91.5
85.5
88.0
113.9
109.1
113.8
113.5
60.7
112.8
104.6
115.4
101.1

t
--.54
-.25
-.97
-1.34
--3.03
2.47
-3.10
-.79
2.19
-1.72
-.21
2.18
-.93
-.05
1.01
.53

--.05
-.02
-.09
-.12
--.31
.22
-.42
-.08
.17
-.14
-.02
.21
-.09
-.004
.10
.05

Segment
SemiNatural
(cont‘d)

TherapyLike

* p < .05

+

b
--1.66*
---.54
-.17
1.49*
-.18
--2.00**
--

Predictor Variables
Global Partner Feelings (A)
Global Partner Feelings (P)
Race / Ethnicity
Education
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Global Partner Feelings (A)
Global Partner Feelings (P)
Race / Ethnicity
Education
** p < .01 *** p < .001

SE
--.69
--.77
.76
.69
.69
--.71
--

df
--83.2
--105.3
105.0
117.9
117.7
--80.6
--

t
--2.41
---.71
-.22
2.15
-.26
--2.82
--

--.22
---.07
-.02
.20
-.02
--.23
--

+

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A, P), Age, and Relationship Status.
Change in moment-to-moment feelings towards one’s partner (see Table 12). Similar to
the change analyses for skin conductance, baseline levels were included as covariates in the
analyses of the moment-to-moment changes in feelings towards one‘s partner. One primary
difference in the analyses of change in feelings towards one‘s partner is the baseline used in the
analysis. There was no baseline established for moment-to-moment feelings towards one‘s
partner prior to any couple communication (i.e., during the relaxation baseline). As a result, the
baseline used for comparison was the neutral- to positively-toned couple communication
segment about how the couple met. In this manner, change from baseline related to how much
of a drop in positive feelings towards one‘s partner was experienced during the semi-natural and
therapy-like communication segments when the couple was discussing a problem compared to
the segment when they were discussing how they met.
For the change between baseline and both the semi-natural and therapy-like
communication segments, there was no significant main effect or interaction involving
attachment or gender. The only significant main effect for the change between baseline and the
106

semi-natural segments was disruptive couple conflict, b = -.30, t(115.0) = -3.53, p = .01, with
lower levels of disruptive conflict associated with less of a decrease in positivity toward one‘s
partner from the baseline to semi-natural communication segment. However, there was no
significant interaction between conflict and attachment when added to the analysis. For the
change between the baseline and therapy-like segments, there was a significant gender main
effect, b = -4.04, t(60.1) = -3.46, p = .001, with men reporting less of a decrease in positivity
towards their partner.
Overall, the analyses involving research question one generated some interesting results.
In terms of physiological arousal, there was a significant effect of attachment avoidance for
younger individuals, with higher attachment avoidance being associated with smaller skin
conductance changes occurring from the relaxation baseline to the baseline interaction. In
addition, women whose male partners reported high attachment anxiety demonstrated increased
physiological arousal and subsequent suppression during each of the interaction conditions.
With respect to feelings towards partner, attachment avoidance was associated with less positive
feelings reported by the person‘s partner during the baseline interaction. Attachment anxiety was
also associated with more positive moment-to-moment feelings towards one‘s partner during the
baseline and therapy-like interactions but not during the semi-natural interaction.
Table 12
Research Question 1 – Change from Baseline in Moment-to-Moment Feelings
Communication
Segment
Semi-Natural
Interaction

+

b

Gender
Avoidance (A)
Avoidance (P)

Table 12 (cont‘d)

107

SE

df

-.98
-.89

Predictor Variables

-1.58
1.51

-111.3
112.0

t
-.62
-.59

-.05
-.05

Communication
Segment

+

b

Predictor Variables

Anxiety (A)
Semi-Natural Anxiety (P)
Interaction
Baseline Interaction Level
(cont‘d)
Conflict (A)
Education
Gender
Avoidance (A)
Avoidance (P)
Therapy-Like Anxiety (A)
Interaction
Anxiety (P)
Interaction Baseline Level
Conflict (A)
Education
* p < .05 ** p < .01 *** p < .001

-.17
.84
-.74***
-.30***
-4.60
-4.04***
.50
-.17
2.39
-.15
-.63***
---

SE

df

t

1.57
1.48
.09
.08
2.73
1.17
1.35
1.31
1.30
1.27
.08
---

100.1
96.2
107.9
115.0
112.6
60.1
116.1
110.6
101.5
94.4
110.5
---

-.11
.57
-8.10
-3.53
-1.69
-3.46
.37
-.13
1.84
-.12
-8.34
---

-.01
.04
-.69
-.32
-.22
-.21
.03
-.01
.14
-.01
-.62
---

+

The following variables were removed from the final analysis for all communication
segments due to non-significance: Symptom Distress (A, P), Conflict (P), Global Partner
Feelings (A, P), Age, Race / Ethnicity, and Relationship Status.
Research Question 2: How do the effects of self-reported attachment account for the
differences in emotional experience between the semi-natural and therapy-like conditions?
The analyses for this research question were similar to those for the first research
question (dyadic analysis using the APIM). The primary difference was that the data were
treated as non-distinguishable and interactional condition (semi-natural versus therapy-like) was
analyzed as a predictor variable in order to understand whether there was a main effect for
condition as well as whether it interacted with actor and partner attachment effects to influence
physiological arousal and feelings towards one‘s partner. Although the data were treated as nondistinguishable by gender, it should be noted that tests for gender main effects and interactions
were conducted and only removed from the analysis if non-significant. In addition, the same
covariates and demographic variables were included in the final analysis if they were significant.
Nonetheless, the primary focus of results for this question was on the main effect for condition or
any interactions involving condition and attachment. In this manner, it was important to

108

understand whether emotional experience in these segments varied not only according to
attachment characteristics of both actors and partners but also according to the interactional
condition.
Physiological arousal (see Table 13). In terms of the raw scores for skin conductance
level, there was no significant interaction between condition and attachment. However, there
was a significant main effect for condition, b = .12, t(60.0) = 2.30, p = .03, with higher skin
conductance more likely during the semi-natural communication segment. There was also no
condition main effect or interaction with attachment when considering physiological variability,
meaning that variability did not change from the semi-natural to the therapy-like segments. For
physiological change from the relaxation baseline there again was no significant interaction
between condition and attachment, but there was a significant main effect for condition, b = .12,
t(62.0) = 2.32, p = .02, with more change from the baseline to semi-natural segments. In this
manner, participants typically evidenced higher levels of physiological arousal when they were
talking about a difficult issue with their partner with no coaching from the therapist than when
talking about that issue with a therapist/coach while their partner was present.
Table 13
Research Question 2 – Condition Comparison of Physiological Level, Variability, and Change
Type

Level

+

b
.12*
.27
-.10
-.07
.11
.45**
.94***
--

Predictor Variables
Condition
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Relaxation Baseline Level
Global Partner Feelings (A)

Table 13 (cont‘d)

109

SE
.05
.15
.15
.16
.15
.17
.04
--

df
62.0
62.0
114.7
99.7
117.6
88.3
93.3
--

t
2.32
1.75
-.67
-.46
.73
2.64
25.9
--

.03
.06
-.03
-.02
.03
.12
.94
--

Type

+

b
--.28
-.04
-.03
.45**
.01
.10***
.01
.004
.01
.02
.04***
-.07
.12*
--.12
-.09
-.005
.21*
-.07*
-.01*
--

Predictor Variables
Education
Gender * Avoidance (A)
Level
Gender * Avoidance (P)
(cont‘d)
Gender * Anxiety (A)
Gender * Anxiety (P)
Condition
Gender
Avoidance (A)
Avoidance (P)
Variability
Anxiety (A)
Anxiety (P)
Relaxation Baseline Level
Relationship Status
Education
Condition
Gender
Avoidance (A)
Avoidance (P)
Change
Anxiety (A)
Anxiety (P)
Relaxation Baseline Level
Global Partner Feelings (A)
Education
* p < .05 ** p < .01 *** p < .001

SE
-.16
.16
.15
.17
.01
.03
.02
.02
.02
.02
.01
-.04
.05
-.12
.11
.11
.10
.03
.01
--

df
-103.7
93.5
118.3
89.0
62.0
62.0
117.6
102.8
122.4
92.1
96.6
-102.3
62.0
-102.8
91.8
118.8
87.6
85.8
90.6
--

t
--1.78
-.26
-.18
2.63
.92
3.78
.23
.19
.41
.77
6.10
-1.56
2.32
--1.06
-.85
-.04
2.15
-2.47
-2.06
--

--.07
-.01
-.01
.12
.03
.26
.02
.01
.03
.05
.43
-.18
.08
--.08
-.06
-.004
.15
-.18
-.18
--

+

The following variables were removed from the final analysis for all measurement types due to
non-significance: Symptom Distress (A, P), Conflict (A, P), Global Partner Feelings (P), Age,
Race / Ethnicity, and Relationship Status.
Feelings towards one’s partner (see Table 14). In terms of the level of feelings (positive
versus negative) towards one‘s partner during the communication segments, there was a
significant condition by attachment anxiety interaction, b = -1.58, t(113.0) = -3.02, p < .01. A
post hoc analysis indicated a significant simple slope for the therapy-like, b = 4.14, t(97.8) =
2.95, p = .004, but not for the semi-natural, b = 1.66, t(91.2) = 1.01, p = .32, condition (see
Figure 10). Specifically, higher levels of attachment anxiety were associated with more positive
feelings towards one‘s partner during the therapy-like condition. There was no condition main
effect or interaction with respect to the variability of feelings towards one‘s partner, indicating

110

that variability was not significantly different across conditions. However, when considering the
changes from baseline and controlling for original baseline levels, the two conditions differed
significantly. There was a significant interaction between condition and attachment anxiety, b =
-1.58, t(113.1) = -3.01, p < .01. A post hoc analysis revealed a significant slope for the therapylike, b = 3.30, t(101.0) = 2.48, p = .02, but not the semi-natural b = .86, t(93.3) = .53, p = .60,
communication segment (see Figure 11). Higher levels of attachment anxiety were associated
with less of a decrease in positive feelings felt towards one‘s partner from the baseline to the
therapy-like communication segments. As a result, attachment anxiety was linked to a more
positive emotional experience during the therapy-like communication segment, suggesting that
greater interactional structure focused on primary emotions and attachment needs may be
especially useful for those higher in attachment anxiety.
Figure 10.

Research Question 2 – The Effects of Condition and Attachment Anxiety on
Level of Moment-to-Moment Feelings towards Partner

b = 4.14, t(97.8) = 2.95, p = .004

b = 1.66, t(91.2) = 1.01, p = .32

In sum, several significant results were found when looking at the effects of attachment
and condition on emotional experience. There were no statistically significant interactions

111

between attachment and condition for physiological arousal. However, there was a significant
condition main effect, indicating that participants evidenced more physiological arousal and
subsequent suppression during the semi-natural communication segment. When considering
moment-to-moment feelings towards one‘s partner, there was a significant interaction between
attachment anxiety and condition, indicating that those higher in attachment anxiety felt more
positively toward their partner during the therapy-like rather than the semi-natural condition.

Figure 11.

Research Question 2 – The Effects of Condition and
Attachment Anxiety on Change from Baseline in Moment-to-Moment Feelings
towards Partner

b = 3.30, t(101.0) = 2.48, p = .02

b = .86, t(93.3) = .53, p = .60

Table 14
Research Question 2 – Condition Comparison of Moment-to-Moment Level, Variability, and
Change in Feelings towards Partner

112

Type

Level

Variability

Change

+

b
-4.27***
-1.40
.50
-.86
2.85*
1.12
-.33***
--.92
-.12
-.55
-1.58**
.02
-1.30
2.68*
-1.37
1.27
.26
1.55*
-1.02
-.34
.66
.08
--.08*
1.72**
-.06
.02
-.28
.10
-4.27***
-2.43*
1.39
.02
2.03
1.06
-.72***
-.23***

Predictor Variables
Condition
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Conflict (A)
Global Partner Feelings (A)
Race / Ethnicity
Condition * Gender
Condition * Avoidance (A)
Condition * Avoidance (P)
Condition * Anxiety (A)
Condition * Anxiety (P)
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
Condition
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Conflict (A)
Global Partner Feelings (A)
Race / Ethnicity
Condition * Avoidance (A)
Condition * Avoidance (P)
Condition * Anxiety (A)
Condition * Anxiety (P)
Condition
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Interaction Baseline Level
Conflict (A)

SE
.62
1.09
1.33
1.27
1.34
1.32
.07
--.49
.53
.53
.52
.52
1.35
1.35
1.30
1.29
.27
.72
.67
.64
.62
.59
-.04
.61
.26
.26
.25
.25
.62
1.10
1.28
1.22
1.28
1.24
.08
.07

df
62.4
63.7
119.2
119.1
104.3
104.4
119.4
--62.5
123.5
123.4
113.0
113.0
98.0
98.5
105.4
106.9
62.4
64.2
120.2
111.5
121.7
109.8
-92.2
86.9
122.5
122.5
121.7
121.7
62.5
65.0
121.2
122.3
108.1
107.2
114.0
124.5

t
-6.93
-1.29
.38
-.68
2.12
.85
-4.90
--1.90
-.23
-1.05
-3.02
.04
-.96
1.99
-1.06
.99
.98
2.16
-1.52
-.53
1.06
.13
--1.98
2.81
-.24
.09
-1.15
.41
-6.93
-2.20
1.09
.02
1.58
.86
-9.46
-3.36

-.23
-.08
.03
-.05
.17
.07
-.41
--.05
-.01
-.03
-.10
.001
-.08
.16
-.08
.08
.03
.19
-.14
-.04
.09
.01
--.17
.21
-.01
.003
-.04
.01
-.21
-.12
-.07
.001
.11
.06
-.67
-.25

SE
--

df
--

t
--

--

Table 14 (cont‘d)
Type
Change

+

b
--

Predictor Variables
Global Partner Feelings (A)
113

(cont‘d)

Race / Ethnicity
Condition * Gender
Condition * Avoidance (A)
Condition * Avoidance (P)
Condition * Anxiety (A)
Condition * Anxiety (P)
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
* p < .05 ** p < .01 *** p < .001

-.93
-.12
-.55
-1.58**
.02
-1.55
2.84*
-1.22
1.29

-.49
.53
.53
.52
.52
1.27
1.26
1.22
1.21

-62.6
123.5
123.5
113.1
113.1
101.1
101.5
108.1
109.9

-1.90
-.24
-1.05
-3.01
.04
-1.22
2.25
-1.00
1.06

-.04
-.01
-.03
-.09
.001
-.08
.15
-.07
.07

+

The following variables were removed from the final analysis for all measurement types due to
non-significance: Symptom Distress (A, P), Conflict (P), Global Partner Feelings (P), Age,
Relationship Status, and Education.
Research Question 3: How will self-reported attachment influence the congruence
between the observed physiological arousal of participants and their moment-to-moment feelings
towards their partner?
It was also valuable to use multilevel modeling and the APIM to understand whether
attachment influences the congruence between the physiological arousal a participant
experiences and their reported feelings towards their partner in each moment (in terms of level,
variability, and change from baseline). For example, would someone with a deactivating
attachment strategy demonstrate a weaker relationship between their self-report and
physiological data because they were more likely to deny their emotional experience? This
question was answered using similar analyses to those performed for previous questions, and the
same decision-making process was used to determine which covariates and demographic
variables remained in the analysis. The primary difference was that moment-to-moment feelings
towards one‘s partner were included in the analysis to predict the corresponding measure of
moment-to-moment physiological arousal (level, variability, and change). Covariates were used
in the same manner as previous analyses, with skin conductance level during the relaxation

114

baseline used as a covariate in each analysis. The results of interest were the main effect for the
physiological arousal variable and any interactions it had with attachment.
Congruence between level of physiological arousal and feelings towards one’s partner
(see Table 15). For the baseline communication segment, there was no significant main effect or
interaction, indicating a lack of congruence between physiological arousal and feelings towards
one‘s partner. There were, however, significant interactions for the semi-natural communication
segment between both actor, b = .03, t(79.7) = 4.13, p < .001, and partner, b = -.01, t(69.5) = 2.20, p = .03, avoidance and feelings towards one‘s partner. For actor avoidance, post hoc
simple slope analyses showed a significant effect of moment-to-moment feelings towards one‘s
partner for those who report lower levels, b = -.04, t(96.0) = -3.35, p = .001, as well as higher
levels, b = .02, t(103.6) = 2.25, p = .03, of attachment avoidance (see Figure 12). In other words,
those individuals who reported higher levels of attachment avoidance as well as more positive
moment-to-moment feelings towards their partner demonstrated higher levels of physiological
arousal. The converse was true for lower levels of attachment avoidance. Those who reported
more positive feelings towards their partner demonstrated lower levels of physiological arousal.

115

Figure 12.

Research Question 3 – The Effect of Attachment Avoidance on the Congruence
between the Level of Physiological Arousal and Moment-to-Moment Feelings
towards Partner during the Semi-Natural Interaction

b = -.04, t(96.0) = -3.35, p = .001

b = .02, t(103.6) = 2.25, p = .03

MTM Feelings towards Partner
Post hoc simple slope analyses also showed a significant effect of moment-to-moment
feelings towards one‘s partner for those whose partners were higher in attachment avoidance, b
= -.02, t(104.0) = -2.16, p = .03, but not for those whose partners were low in attachment
avoidance, b = .01, t(102.6) = .69, p = .49 (see Figure 13). When partner attachment avoidance
was higher, those who reported higher moment-to-moment feelings towards their partner
demonstrated lower levels of physiological arousal. This relationship shows that, unlike those
high in attachment avoidance, partners of those high in attachment avoidance demonstrate the
expected relationship between moment-to-moment feelings and physiological arousal, with more
positive feelings being associated with lower levels of physiological arousal.

116

Figure 13.

Research Question 3 – The Effect of Partner Attachment Avoidance on the
Congruence between Level of Physiological Arousal and Moment-to-Moment
Feelings towards Partner during the Semi-Natural Interaction

b = .01, t(102.6) = .69, p = .49

b = -.02, t(104.0) = -2.16, p = .03

There was also a significant interaction between attachment avoidance and feelings
towards one‘s partner for the therapy-like communication segment. Post hoc simple slope
analyses were similar to those for the semi-natural communication segment, with a significant
effect of moment-to-moment feelings towards one‘s partner for both high, b = .04; t(108.4) =
2.65; p = .01, and low, b = -.04; t(101.9) = -2.27; p = .03 (see Figure 14), levels of attachment
avoidance. As such, those high in attachment avoidance who reported more positive feelings
towards their partner also demonstrated higher physiological arousal. With respect to partner
effects of attachment avoidance during the therapy-like interaction, post hoc simple slope
analyses were not significant for high, b = -.02; t(87.6) = -1.27; p = .21, or low, b = .02; t(91.3) =
1.28; p = .20, attachment avoidance, although it should be noted that the slopes were in the same
direction as the partner effects during the semi-natural interaction.

117

Figure 14.

Research Question 3 – The Effect of Attachment Avoidance on the Congruence
between Physiological Arousal and Moment-to-Moment Feelings towards Partner
during the Therapy-Like Interaction

b = -.04; t(101.9) = -2.27; p = .03

b = .04; t(108.4) = 2.65; p = .01

Overall, there was little overall evidence of congruence between moment-to-moment
physiological arousal and moment-to-moment feelings towards one‘s partner. However, there
was congruence between both measures of emotional experience when considering attachment
avoidance during the semi-natural communication segment. Specifically, higher levels of
attachment avoidance coupled with moment-to-moment reports of positive feelings towards
one‘s partner were actually associated with higher levels of physiological arousal. Interestingly,
partners of those high in attachment avoidance actually demonstrated congruence similar to those
low in attachment avoidance. Those who indicated feeling more positively towards their partner
demonstrated lower levels of physiological arousal.

118

Table 15
Research Question 3 – Comparison of the Congruence between the Level of Physiological
Arousal (PH) and Level of Moment-to-Moment Feelings towards Partner (MTM)
Segment

+

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
MTM Baseline
PH Relaxation Baseline
Age
Baseline
Gender * MTM Baseline
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
MTM Semi-Natural
PH Relaxation Baseline
Age
Gender * MTM Semi-Natural
SemiGender * Avoidance (A)
Natural Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)
MTM Semi-Natural * Avoidance (A)
MTM Semi-Natural * Avoidance (P)
MTM Semi-Natural * Anxiety (A)
MTM Semi-Natural * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
TherapyAnxiety (P)
Like
MTM Therapy-Like
PH Relaxation Baseline
Age
MTM Therapy-Like * Avoidance (A)

119

b
.26***
-.09
.03
.02
.09
-.01
1.02***
-.03*
-.003
-.15
.07
-.10
.24**
.44***
-.07
-.06
.08
.38*
-.01
.98***
--.01
-.19
.06
-.14
.41**
.03***
-.01*
.01
.002
.16
.07
-.17
.16
.22
.001
.94***
-.04***

SE
.07
.08
.08
.08
.09
.005
.02
.01
.004
.09
.09
.08
.09
.13
.13
.14
.13
.15
.01
.03
-.01
.14
.14
.13
.15
.01
.01
.01
.01
.18
.18
.18
.19
.17
.01
.05
-.01

df
t
54.7
3.53
103.7 -1.06
96.4
.40
90.1
.21
83.4
1.09
98.4 -1.44
83.9
48.1
68.4 -2.60
75.8
-.76
89.4 -1.72
84.7
.76
92.5 -1.20
85.4
2.87
50.4
3.41
95.7
-.51
88.7
-.41
97.6
.62
73.8
2.61
91.3
-.87
77.1
31.0
--73.9
-.79
85.6 -1.39
80.8
.38
98.9 -1.06
76.0
2.77
79.7
4.13
69.5 -2.20
93.1
1.47
66.2
.35
61.2
.89
104.1
.39
91.6
-.97
106.2
.82
84.1
1.32
94.5
.14
85.7 20.7
--102.7
3.48

.06
-.02
.01
.005
.03
-.03
1.04
-.06
-.02
-.04
.02
-.03
.07
.10
-.02
-.01
.02
.10
-.03
.99
--.03
-.05
.01
-.04
.11
.13
-.06
.05
.01
.04
.02
-.04
.04
.06
.01
.93
-.16

Table 15 (cont‘d)
Segment
TherapyLike
(cont‘d)
* p < .05

+

Predictor Variables
MTM Therapy-Like * Avoidance (P)
MTM Therapy-Like * Anxiety (A)
MTM Therapy-Like * Anxiety (P)
** p < .01 *** p < .001

b
-.02*
.01
.001

SE
.01
.01
.01

df
73.5
98.7
86.5

t
-2.18
.89
.09

+

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A, P), Global Partner Feelings (A,
P), Race / Ethnicity, Relationship Status, and Education.
Congruence between variability of physiological arousal and variability of feelings
towards one’s partner (see Table 16). There was no significant main effect for feelings towards
one‘s partner or interaction between feelings towards one‘s partner and attachment avoidance or
anxiety for the baseline or therapy-like communication segments, suggesting no congruence
between the variability of moment-to-moment feelings towards one‘s partner and variability of
moment-to-moment physiological arousal. There was a significant interaction involving
attachment avoidance and the variability of feelings towards one‘s partner for the semi-natural
segment. However, post hoc simple slope analyses revealed no significant effect of the
variability in moment-to-moment feelings towards partner for higher, b = -.01, t(105.3) = -1.29,
p = .20, or lower, b = .01, t(98.8) = 1.55, p = .12, levels of attachment avoidance.

120

-.07
.05
.004

Table 16
Research Question 3 – Comparison of the Congruence between the Variability of Physiological
Arousal (PH) and Variability of Moment-to-Moment Feelings towards Partner (MTM)
Segment

+

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
Baseline
MTM Baseline
PH Relaxation Baseline
Conflict (P)
Relationship Status
Education
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
MTM Semi-Natural
PH Relaxation Baseline
SemiConflict (P)
Natural
Relationship Status
Education
Gender * MTM Semi-Natural
MTM Semi-Natural * Avoidance (A)
MTM Semi-Natural * Avoidance (P)
MTM Semi-Natural * Anxiety (A)
MTM Semi-Natural * Anxiety (P)
Gender
Avoidance (A)
Avoidance (P)
Anxiety (A)
Anxiety (P)
MTM Therapy-Like
PH Relaxation Baseline
TherapyConflict (P)
Like
Education
Gender * MTM Therapy-Like
Gender * Avoidance (A)
Gender * Avoidance (P)
Gender * Anxiety (A)
Gender * Anxiety (P)Relationship
Status

121

b
.09**
.02
-.02
-.05
-.05
-.003
.03***
.01***
-.11***
-.10***
.01
.0001
.02
.02
.001
.04***
----.004
-.01*
-.003
-.01
-.002
.10***
-.01
-.01
.02
.05
-.003
.04***
--.08
-.01*
-.002
-.04
.05
.06*

SE
.03
.03
.03
.03
.03
.003
.01
.001
.03
-.03
.03
.03
.03
.03
.004
.01
---.004
.004
.004
.004
.004
.03
.03
.03
.03
.03
.003
.01
--.05
.003
.03
.03
.03
.03

df
55.4
92.6
77.8
103.2
67.9
101.9
75.0
70.0
57.9
-54.4
98.1
93.9
107.5
80.4
105.3
82.6
---105.4
107.9
102.7
102.5
91.4
53.1
106.9
95.0
107.9
85.9
99.9
88.3
--88.4
105.7
102.5
95.1
108.6
86.1

t
2.94
.56
-.82
-1.88
-1.88
-1.18
4.19
3.96
-3.60
-3.40
.33
.01
.53
.86
.19
5.42
----.93
-2.00
-.74
-1.51
-.67
3.58
-.30
-.50
.69
1.84
-.95
5.35
--1.68
-2.18
-.08
-1.41
1.72
2.13

.23
.04
-.06
-.16
-.14
-.09
.32
.32
-.28
-.26
.03
.001
.05
.06
.02
.41
----.07
-.16
-.06
-.13
-.05
.29
-.02
-.04
.06
.17
-.07
.43
--.22
-.17
-.01
-.12
.15
.19

* p < .05 ** p < .01 *** p < .001
+

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A), Global Partner Feelings (A, P),
Age, and Race / Ethnicity.
Congruence between change in physiological arousal and change in feelings towards
one’s partner (see Table 17). There was no significant main effect or two-way interaction that
pointed to any congruence between change in physiological arousal and change in feelings
towards one‘s partner when considering the change from the baseline to the semi-natural
communication segment or from the baseline to therapy-like segment. This lack of statistically
significant results may be due to the fact that change in physiological arousal was measured
using baseline values at rest, while the change in feelings towards one‘s partner was measured
using baseline values during a typically positive interactional warm-up where couples talked
about how they met.
Overall, any congruence between physiological arousal and moment-to-moment feelings
towards one‘s partner involved attachment avoidance. Specifically, for those higher in
attachment avoidance, more positive reported moment-to-moment feelings towards one‘s partner
were associated with higher levels of physiological arousal and subsequent suppression during
both the semi-natural and therapy-like communication segments. Conversely, for those who
reported low levels of attachment avoidance, higher reported moment-to-moment positive
feelings towards one‘s partner were associated with decreases in physiological arousal. Partners
of those high in attachment avoidance evidenced a similar pattern as those low in attachment
avoidance, with more positive feelings being associated with less physiological arousal.

122

Table 17
Research Question 3 – Comparison of the Congruence between Change in Physiological Arousal
(PH) and Change in Moment-to-Moment Feelings towards Partner (MTM)
Communication
Segment

+

b

Predictor Variables

Gender
Avoidance (A)
Avoidance (P)
Semi-Natural Anxiety (A)
Interaction
Anxiety (P)
MTM Semi-Natural Change
PH Relaxation Baseline
Global Partner Feelings (A)
Gender
Avoidance (A)
Avoidance (P)
Therapy-Like Anxiety (A)
Interaction
Anxiety (P)
MTM Therapy-Like Change
PH Relaxation Baseline
Global Partner Feelings (A)
* p < .05 ** p < .01 *** p < .001
+

.12
-.14
-.03
-.05
.15
.005
-.05*
-.01*
-.01
-.19
-.16
.01
.26*
.01
-.09*
-.02

SE

df

t

.11
.10
.09
.10
.08
.004
.02
.01
.18
.16
.14
.16
.13
.01
.04
.01

62.1
90.4
80.1
102.5
72.9
76.8
74.5
80.9
67.3
97.8
84.5
108.7
78.4
87.3
79.3
83.9

1.11
-1.42
-.31
-.48
1.90
1.10
-2.10
-2.25
-.04
-1.15
-1.10
.06
1.99
1.08
-2.48
-1.78

.10
-.13
-.03
-.04
.15
.09
-.17
-.22
-.004
-.11
-.09
.01
.16
.09
-.20
-.17

The following variables were removed from the final analysis for all communication segments
due to non-significance: Symptom Distress (A, P), Conflict (A, P), Global Partner Feelings (P),
Age, Race / Ethnicity, Relationship Status, and Education.

123

CHAPTER V
DISCUSSION
Overview of the Study
A Common Factors Perspective of Client-Treatment Matching
A substantial body of research has shown that couple therapy is more effective than a notreatment control group for a large majority of clients (Alexander, Holtzworth-Munroe, &
Jameson, 1994; Christensen & Heavey, 1999; Lambert & Ogles, 2003; Sexton et al., 2003;
Shadish & Baldwin, 2003). Despite these promising results, there still are a substantial number
of couples who do not achieve ―clinically significant positive outcomes‖ (Sexton et al., 2003, p.
598). Although studying the effectiveness of couple therapy is important and gives the field
additional credibility, increased efforts are needed to understand more about how, why, and for
whom couple therapy is effective (Jacobson & Addis, 1993; Paul, 1967).
Two research avenues offer potential for adequately addressing these objectives: clienttreatment matching and process research. The goal of client-treatment matching is to identify
how treatment can be adapted to specific client characteristics. For the most part, the primary
focus of client-treatment matching research has been to identify those treatment modalities that
most effectively treat a specific diagnosis (Treatment for Depression Collaborative Research
Program-TDCRP, Elkin et al., 1989; Project MATCH Research Group, 1997a, 1997b, 1998).
Efforts in this area have been disappointing, however, with few indications of significant
differences among treatments. This lack of significant findings may be attributed to the fact that
the models tested in the studies likely have many of the same core ingredients that bring about
change, thereby confounding results.

124

In this manner, client-treatment matching with a common factors perspective may yield a
more comprehensive understanding of those factors that ultimately influence. In addition to
utilizing a common factors perspective to conceptualize client-treatment matching, process
research is necessary to identify those change mechanisms that occur during therapy to bring
about positive outcomes, independent of treatment model. By identifying those elements of
therapy process that contribute to positive outcomes, researchers can then work to understand
how various common factors (e.g., client characteristics) interact with therapy process.
Attachment Theory as an Important Client Characteristic
This study sought to understand how one client characteristic (self-reported attachment
avoidance or anxiety) influenced the emotional experience of participants (as measured by skin
conductance and self-reported, moment-to-moment feelings towards partner). Emotions are a
central organizing feature of intrapersonal and interpersonal experience (Frijda, 1986; Izard,
1991; Johnson & Greenberg, 1994). In addition, appropriate regulation and processing of
emotions have been linked to positive therapeutic outcomes (Beutler et al., 2003; Bridges, 2006;
Greenberg & Pascual-Leone, 2006; Wiser & Arnow, 2001). The effects of attachment on
emotional experience (i.e., the level, variability, and change from baseline for physiological
arousal and moment-to-moment feelings towards one‘s partner) were investigated within the
context of two contrasting conditions. During the semi-natural interaction, couples talked about
an issue as they would at home. The therapy-like interaction was designed to promote and
facilitate couple-responsible and couple-focused process, an intervention which has been
associated with more positive outcomes (Butler & Wampler, 1999). In this manner, the
therapist/coach worked to facilitate a new emotional experience and more softened couple
interaction by reframing secondary emotions in terms of primary emotions and attachment needs.

125

One of the core tenets of attachment theory is that humans are pre-programmed to seek
safety from close relationships (Bowlby, 1973, 1982). This assurance of a secure base provides
the foundation for exploration but also makes it possible for the individual to seek and receive
comfort when emotionally distressed. Unmet attachment needs often lead to the formation of
disruptive coping strategies in the form of deactivation or hyperactivation of the attachment
system (Kobak et al., 1993). These strategies also impede an individual‘s ability to clearly and
appropriately signal their attachment needs. Attachment deactivation forms in response to
consistent rejection of requests to have needs met and typically manifests itself as an inhibition
of thoughts and emotions related to attachment (Diamond et al., 2006). Although deactivating
individuals do not necessarily demonstrate outward distress, physiological measures of distress
provide evidence that attempts at suppression actually may increase stress (Sroufe & Waters,
1977). On the other hand, hyperactivation of the attachment system typically occurs in response
to inconsistency in responding to attachment needs, which makes it difficult for the individual to
predict when their needs will be met and leads to an exaggerated stress response where the
person‘s requests actually serve to push the attachment figure away.
Relatively few studies have empirically analyzed how self-report attachment influences a
person‘s physiological response to stress. Those that have addressed this issue have looked at
various physiological measures, including blood pressure, cortisol levels, heart rate, skin
conductance, and vagal tone (Carpenter & Kirkpatrick, 1996; Diamond & Hicks, 2005; Diamond
et al., 2006; Kim, 2006; Powers et al., 2006). Findings have varied, but one of the more
promising findings has been that attachment deactivation leads to higher levels of skin
conductance, a valid measure of emotional arousal and its suppression (Dawson et al., 2007).
Several studies addressing narrative/discourse (Dozier & Kobak, 1992; Roisman et al., 2004) as

126

well as self-report (Diamond et al., 2006) attachment have found a link between attachment
deactivation and greater change from baseline in terms of skin conductance level. At the same
time, studies of skin conductance have found no link to hyperactivating attachment strategies.
One purpose of this study was to further understand the link between self-report attachment and
physiological arousal as measured by skin conductance.
In addition, no previous study has analyzed how attachment influences moment-tomoment reported feelings towards one‘s partner. Although global and proximal self-report
measures have been used to understand participant experience, the potential for various biases in
these measures has limited the ability to generalize such findings. Observer ratings have often
been used but do not tap into actual participant experience, an understanding of which is also
important, especially in clinically-oriented research. Moment-to-moment self-report allows a
much richer and potentially valid understanding of participant experience to be gained. Another
unique element of this study is its incorporation of a therapeutic component. Most attachmentrelated studies have focused on relationship process rather than how it influences therapy
process. This study analyzed and compared both.
Research Questions and a Brief Summary of Findings
This study focused on three primary research questions. The first addressed how
attachment avoidance and/or anxiety would influence physiological arousal as well as momentto-moment feelings towards one‘s partner. For raw levels of skin conductance (not controlling
for baseline skin conductance levels), age was predictive of lower skin conductance for the
baseline, semi-natural, and therapy-like interactions. The only other significant result was that
partners of those high in attachment anxiety demonstrated higher physiological arousal and
subsequent suppression during the therapy-like interaction (Table 7). For physiological

127

variability, women whose partners reported higher levels of attachment anxiety exhibited more
physiological variability during the therapy-like interaction (Table 8). Women in general also
exhibited more variability during the baseline and semi-natural interactions. Being in a
committed, non-marriage relationship and partner reports of higher conflict were associated with
more variability during the baseline interaction. With respect to change from baseline, women
whose partners were high in attachment anxiety demonstrated increased physiological arousal
and subsequent suppression during all interaction segments (Table 9). In addition, during the
baseline interaction, attachment insecurity interacted with age. Older adults who were low in
avoidance or high in anxiety demonstrated less of an increase in physiological arousal from the
relaxation baseline to the baseline interaction. Conversely, younger adults in these two
attachment categories exhibited greater change from the relaxation baseline during the
interaction baseline (Figure 6).
In terms of moment-to-moment feelings towards one‘s partner, a few patterns emerged in
the analysis of research question one. Either self- or partner-reported couple conflict was
predictive of the level of feelings reported for each of the interaction segments (Table 10).
Attachment anxiety was actually associated with more positive feelings towards one‘s partner
during the baseline and therapy-like interactions. In addition, partner avoidance led to less
positive feelings towards one‘s partner during the baseline interaction, while being a woman and
having attained a higher education level were associated with more positive moment-to-moment
feelings. Attachment avoidance also interacted significantly with global feelings towards one‘s
partner. Higher attachment avoidance and more positive reported feelings towards one‘s partner
were associated with greater variability in moment-to-moment feelings towards one‘s partner
during the baseline interaction (Table 11). Participants also completed a measure of their overall

128

feelings towards their partner for all of the interaction segments combined. Partner avoidance
and anxiety were associated with less positive proximally-reported feelings towards one‘s
partner. Women who reported higher levels of attachment avoidance also felt less positively
towards their partner, while men who reported higher attachment anxiety actually reported more
positive feelings towards their partner. Lastly, higher reported levels of conflict and being in a
committed, non-marriage relationship were associated with less positive feelings.
Somewhat surprisingly, actor and partner attachment avoidance were not predictive of the
level, variability, or change in physiological arousal, with the exception of the previously
mentioned relationship between avoidance, age, and physiological change from baseline. In
addition, actor avoidance was not predictive of moment-to-moment feelings towards one‘s
partner, with the exception of interacting with global feelings towards one‘s partner to influence
moment-to-moment variability. It was also predictive of a proximal report of women‘s feelings
towards their male partners for all interactions combined. In terms of covariates and
demographic variables, only gender (variability) and age (level) demonstrated a pattern of
association for physiological arousal. For moment-to-moment feelings towards one‘s partner,
couple conflict (level), and ethnicity (variability) demonstrated a pattern of association across
interactional contexts.
The second research question addressed how attachment accounted for any differences
between participant emotional experience during the semi-natural and therapy-like interactions.
As stated previously, similar covariates, demographic variables, and interactions to research
question one were used in the analyses for research question two. However, of primary interest
was an interaction between condition and attachment or a main effect for condition. As a result,
only those findings are reported here. Although there were physiological level and change

129

differences between the semi-natural and therapy-like conditions, and participants were less
physiologically aroused during the therapy-like condition, these differences did not appear to be
the result of attachment anxiety or avoidance (Table 13). This finding that attachment did not
interact with condition to influence physiological arousal was somewhat unexpected. However,
there was a significant interaction between attachment and condition for moment-to-moment
feelings towards one‘s partner (Table 14). Those higher in attachment anxiety felt more
positively towards their partner during the therapy-like interaction. Interestingly, this analysis
did not find any difference in condition for those who reported high levels of attachment
avoidance.
The final research question looked at the congruence between participant physiological
arousal and the moment-to-moment feelings they reported toward their partner and how
attachment influenced that congruence. The covariates, demographic variables, and interactions
used in the analyses for research question three were the same as for the first two research
questions. However, the primary focus was on a potential main effect or interactions involving
moment-to-moment feelings towards one‘s partner on physiological arousal. Interestingly, all
evidence of congruence or a lack thereof involved attachment avoidance rather than attachment
anxiety. Results yielded no evidence of congruence for level during the baseline interaction.
However, those high in attachment avoidance who reported more positive moment-to-moment
reports of feelings towards their partner during the semi-natural (Figure 12) and therapy-like
(Figure 14) conditions were more likely to demonstrate higher levels of physiological arousal
and subsequent suppression. Conversely, those low in attachment avoidance who reported
positive moment-to-moment feelings towards their partner were more likely to demonstrate
lower levels of physiological arousal. Interestingly, results for partners of those high in

130

attachment avoidance were similar to those who were low in avoidance during the semi-natural
interaction (Figure 13). A similar effect was observed for partners of those high in attachment
avoidance during the therapy-like interaction, but the simple slope analysis was not significant.
In terms of variability, those individuals high in attachment avoidance who reported low
variability in their feelings towards their partner demonstrated higher variability in their
physiological arousal during the semi-natural interaction. For those low in attachment
avoidance, higher variability in their moment-to-moment feelings towards their partner were
associated with higher variability in physiological arousal (Table 17). There was no evidence of
congruence between physiological arousal and moment-to-moment feelings towards one‘s
partner in terms of the change from baseline, possibly because the baseline used for
physiological arousal was a relaxation baseline, while the baseline used for moment-to-moment
feelings was actually focused on reminiscing about what was likely to be a positive time in the
couple‘s relationship.
This section has briefly outlined all of the findings for each of this study‘s three research
questions. Overall, there were a number of significant findings regarding the influence of
attachment on emotional experience. The next sections will synthesize and discuss the most
salient implications of this study‘s results for couple relationship process, clinical work with
those who report higher levels of attachment anxiety or avoidance, and future directions for
research in this area.
Relationship Process Implications
Attachment Avoidance and Repressive Coping
Repressive coping has been defined as low reported anxiety coupled with high
defensiveness and the employment of distraction strategies to submerge and avoid negative

131

affect (Myers, 1998; Weinberger, 1990). As such, repressive coping represents the suppression
of unwanted thoughts and feelings in an effort to mask behavioral expressions of distress.
However, even though that distress is not reported or demonstrated outwardly, physiological
measures reveal ―heightened and escalating sympathetic nervous system reactivity‖ (Diamond et
al., 2006, p. 221). Most studies have used social desirability scales coupled with assessments of
anxiety to measure repressive coping (Vetere & Myers, 2002). However, physiological
measures are also important to actually verify that individuals are experiencing distress even in
the absence of their self-report of it.
Although several narrative/discourse attachment studies (Dozier & Kobak, 1992;
Roisman, 2007; Roisman, Tsai, & Chiang, 2004) have linked attachment deactivation and
emotional inhibition as measured by skin conductance, this study represents only the second
attempt to understand the influence of self-report attachment on physiological arousal as
measured by skin conductance and the first that did not screen out moderately distressed couples.
As a result, this section will outline findings and implications from research questions one and
three as they compare with Diamond et al. (2006). This is especially important because the
elevated sympathetic nervous system activity that accompanies repressive coping has been
linked to various health risks (Repetti, Taylor, & Seeman, 2002). In a similar vein, individuals in
a cardiac rehabilitation program were more likely than a control group to report avoiding
problem discussions with their partner (Denton, Burleson, & Brubaker, 2009).
Interestingly, the results of research question one did not find a direct link between
avoidance and elevated skin conductance levels from baseline. However, the findings from
research question three regarding congruence between moment-to-moment physiological arousal
and moment-to-moment feelings towards one‘s partner provide evidence for an association

132

between attachment avoidance and repressive coping. Specifically those higher in attachment
avoidance who reported more positive moment-to-moment feelings towards their partner during
the semi-natural and therapy-like interactions also showed elevated physiological arousal.
However, those at lower levels of attachment avoidance who reported more positive feelings
towards their partner demonstrated lower levels of physiological arousal. Interestingly, the same
phenomenon as those with low levels of attachment avoidance was also found for partners of
those higher in attachment avoidance. During both interactions, those who reported positive
feelings towards their partner also showed less physiological arousal.
In other words, reports of positive feelings from those high in attachment avoidance were
incongruent with data showing that they were actually physiologically aroused and working
toward emotional suppression. This was the case for both semi-natural and therapy-like
interactions. These results provide additional evidence of the link between attachment avoidance
and repressive coping. For those low in attachment avoidance, more moment-to-moment
positive feelings reported towards their partner were associated with lower levels of
physiological arousal. However, those high in attachment avoidance appeared to cope with
physiological arousal by reporting increases in positive feelings. This interpretation is consistent
with results from another study that indicates those high in repressive coping often distract
themselves from negative affect by recalling positive thoughts and memories (Boden &
Baumeister, 1997). If this is indeed the case, the short term effect in the context of couple
interaction would be more positive feelings towards one‘s partner. However, it also points
towards an unwillingness to address the more difficult issues that inevitably arise in
relationships, thereby putting long-term relationship satisfaction at risk. This is supported by the
fact that preliminary analyses in this study found avoidance to be negatively correlated with

133

conflict resolution, global relationship satisfaction, and positive feelings towards partner.
Attachment Anxiety and Partner Suppression
Because no studies of attachment (self-report or narrative/discourse measurement) had
identified a link between attachment anxiety and physiological arousal as measured by skin
conductance level (SCL), this study also sought to add to the current body of attachment and skin
conductance research by further investigating a potential link primarily through research
question one. It also looked at how attachment might influence the variability of skin
conductance (i.e., the standard deviation) as a potential indicator of skin conductance response
(SCR), a measure not examined in other research. Results did not reveal a predictable
relationship between attachment and the variability of skin conductance. Rather, gender was the
most consistent predictor across interactional conditions, with greater variability for women‘s
skin conductance.
However, a strong link was found between the influence of gender and partner
attachment anxiety on physiological arousal during each of the communication segments.
Specifically, women whose male partners reported higher levels of attachment anxiety showed
increased physiological arousal and subsequent emotional suppression during each of the
interactional conditions, including the baseline interaction where couples talked about how they
met. Conceptually, it makes sense that the intensification of emotion (both positive and
negative, see Feeney, 1999) that typically accompanies attachment anxiety may elicit a
complementary reaction from a person‘s partner in the form of emotional suppression. However,
this was only the case for man-woman dyads in which the man reported anxious attachment and
the woman was the person suppressing her emotion. That gender influenced this finding in the
direction it did was somewhat unexpected in part because the female partners did not necessarily

134

report higher tendencies towards avoidance, and women have been found to typically be more
outwardly expressive than men (see Kring & Gordon, 1998, for a summary)
In addition, this finding was not associated with more negative moment-to-moment
feelings reported by women or men towards their partners. In fact, anxious men actually
reported more positive feelings towards their partner when reporting on their interactions as a
whole. However, viewed from a systemic perspective, the intensifying of emotions that
accompanies the man‘s exaggerated attempts to have attachment needs met (Cassidy, 1994) may
be counter-productive and actually push the woman further away. There is also a possibility that
the woman‘s emotional suppression could be linked to an inhibitory behavioral response akin to
withdrawing, shutting down, or pulling away, thereby impeding the woman‘s ability to
emotionally engage and be responsive to the man‘s needs. As a result, although the suppression
did not appear to negatively influence the interaction or the partners‘ feelings towards each other
in this study, it is important to understand more about the potential reasons for this finding.
It may be valuable to consider relationship process as well as gender role expectations.
Although the results disconfirmed that this is a dynamic of relationships at the level of
attachment, it may be an element of relationship process at the level of interaction. In other
words, the emotional suppression in this case may be an adaptive response and complementary
process to the more intense positive or negative emotions exhibited by the partner who is higher
in attachment anxiety. That this only seems to be the case for women may be related to gender
role socialization. Specifically, women may modulate their responses according to the demands
of each situation according to specific display rules, defined as ―culture-specific prescriptions
about who can show which emotions, to whom, and when‖ (Ekman, 1993, p. 384; see also Buck,
Losow, Murphy, & Costanzo, 1992; Gross & John, 1997). It is possible that the woman‘s

135

emotional suppression represented an attempt to balance the overall emotionality of the couple
discussion by dampening the overall emotional intensity of the discussion. In addition, there
appears to be some indication that men are not as successful as women in interpreting nonverbal
cues of specific emotions, with the exception of being able to recognize anger in men more
readily (Rotter & Rotter, 1988; see also Zuckerman, Hall, DeFrank, & Rosenthal, 1976). If this
is indeed the case, it would offer one potentially viable explanation of why the woman‘s
emotional suppression did not influence the man‘s reported feelings.
Clinical Implications
Over twenty years ago, Bowlby (1988) emphasized the importance of informing clinical
work with attachment theory. However, only recently has more attention has been given to
attachment theory and its relationship to clinical work, including the development of attachmentbased interventions for children and their caregivers (e.g., Cooper, Hoffman, Powell, & Marvin,
2005; Dozier, Lindhiem, & Ackerman, 2005) as well as emotionally-focused therapy (EFT),
which utilizes attachment theory as a core part of its theoretical foundation and work with
couples and families. Each of these focuses on helping individuals learn to more appropriately
signal their attachment needs and longings while at the same time fostering responsiveness in the
attachment figure. Unfortunately, little empirical work has been done—especially in couple
therapy—to understand how attachment influences therapy process in general and how to
thereby adapt treatment process to attachment characteristics. This study sought to lay a
foundation for future research in this area by analyzing how attachment influences emotional
experience during contrasting interactional conditions.
Emotional Experience during the Semi-Natural and Therapy-Like Interactions
This section will further discuss the findings of research question two regarding the

136

differences in emotional experience during the semi-natural and therapy-like interactions and the
role of attachment in accounting for these differences. In terms of physiological arousal,
individuals in this study demonstrated less physiological arousal and subsequent suppression
during the therapy-like interaction. Attachment avoidance or anxiety did not influence this
relationship, suggesting that regardless of attachment, the presence of a therapist/coach who
focused on validating and reframing emotions as well as helping partners soften towards one
another was a positive experience for couples. At least from a physiological standpoint, the
presence of the therapist/coach and an emotionally-focused intervention served as a buffer
against physiological arousal and suppression rather than being threatening to those who reported
higher levels of attachment avoidance.
With respect to moment-to-moment feelings, there were relatively few significant
associations when considering avoidance. As stated previously, partners of those higher in
attachment avoidance reported less positive feelings during the baseline interaction. However,
results did not reveal any other statistically significant associations, including no significant
difference between the moment-to-moment feelings towards one‘s partner experienced during
the semi-natural and therapy-like interactions. As a result, it appears that even though they may
have experienced less physiological arousal and suppression during the therapy-like interaction,
they did not necessarily report more positive feelings towards their partner.
Conversely, those higher in attachment anxiety reported higher levels of positive feelings
towards their partner during the therapy-like interaction. More positive feelings towards partner
were also reported by those higher in attachment anxiety during the baseline interaction.
Attachment anxiety relates to worries about the availability and responsiveness of attachment
figures. Thus, it appears that reminiscing about positive memories and discussing relationship

137

issues from an attachment perspective with the help of a therapist/coach diminish fears about the
availability of attachment figures and provide reassurance that the partner will be responsive to
the individual‘s attachment needs. In sum, results provide considerable evidence that the
emotionally-focused intervention used in this study led to more effective outcomes than having
partners talk about their issues as they would at home. Positive physiological effects were
evident during the therapy-like interaction regardless of attachment, while moment-to-moment
feelings towards one‘s partner were more positive during the therapy-like interaction only for
those higher in attachment anxiety.
Nonetheless, because this was not a randomized control trial with the order of conditions
varied, it is not possible to make definitive conclusions that the intervention by itself will have
the same results as found in this study. Rather, the emotionally-focused intervention used in this
study was preceded by the semi-natural condition. As a result, it is more accurate to conclude
that the combination of the semi-natural interaction followed by the emotionally-focused
intervention led to lower levels of physiological arousal and more positive moment-to-moment
feelings towards one‘s partner for those higher in attachment anxiety than the semi-natural
condition alone. In addition, the primary goal of this study was not to test the effectiveness of
this specific intervention, but rather to begin developing an understanding of how attachment
influences emotional experience during a variety of interactional conditions, including an
emotionally-focused, therapy-like condition. As such, this study provides preliminary evidence
that attachment does indeed influence the emotional experience of individuals, thereby making it
necessary to identify how clinical work with couples can be further adapted according to
attachment in order to improve outcomes.

138

Attachment Anxiety and Clinical Process
This section will build upon the previous one by discussing specific implications for the
finding of research question two regarding the positive outcomes during the baseline and
therapy-like interactions for those higher in attachment anxiety. Butler and Gardner (2003)
explain how a developmental approach can be applied to couple interaction facilitated within
therapy by adapting the interaction to couple volatility and reactivity. Put another way, the
therapist structures the interaction according to varying levels of couple distress by filtering all of
the interaction through the therapist when distress and volatility are high, and shifting to the
periphery of the interaction and engaging in coaching when volatility is lower and the couple
demonstrates greater interactional self-reliance (see also Seedall & Butler, 2006). Although this
study was not an explicit test of the developmental enactment model, it did incorporate the two
interactional extremes into its methodology. During the semi-natural interaction, structure was
low and couples interacted as they would at home, with the only potential buffer being the
presence of the therapist in the room. During the therapy-like interaction, structure was much
higher and all of the interaction was filtered through the therapist, with the exception of a brief
enactment designed to highlight a softened response at the conclusion of each segment.
The results of this study point towards the importance of not only taking into account
couple distress and volatility when structuring the interaction, but also adapting to self-reported
attachment strategies. Specifically, those higher in attachment anxiety benefited from the
emotional containment provided by having a therapist/coach buffer the couple interaction.
Interestingly, this benefit from the increased structure was evident independent of each person‘s
globally reported feelings towards his/her partner or individual symptom distress and in
conjunction with reports of disruptive conflict. It should also be noted that the increased

139

structure was accompanied by an intervention focused on dampening disruptive conflict and
facilitating emotional exploration, validation, and reframing in an effort to help partners
experience each other in a softened, more emotionally responsive way.
Thus, it is important for a therapist/coach working with those high in attachment anxiety
to help them contain volatile emotions and learn to express and explore their issues in terms of
primary emotions and core attachment needs. Because there was at least some indication that
partners of those higher in attachment anxiety at times experienced greater physiological arousal
and subsequent suppression, failure to contain the expression of attachment anxieties (i.e.,
attachment hyperactivation) may risk overwhelming the partner and thereby decrease the
likelihood of responsiveness and having the individual‘s emotional needs met. This likely would
only serve to exacerbate the attachment insecurities. However, if clinicians can instead help
those higher in attachment anxiety express their issues in more contained ways by helping them
look below the surface of secondary emotions and become aware of their primary emotions and
attachment needs, their partners will likely see the person and their needs in new ways that help
them soften as well as be more responsive. This will then, in turn, offer much needed
reassurance to the person higher in attachment anxiety that their needs can and will be met if
they can signal them more appropriately.
Attachment Avoidance and Clinical Process
As mentioned previously, the results of research question three provided evidence that
during the semi-natural interaction, the more physiologically aroused those high in attachment
avoidance became, the more positively they indicated feeling toward their partner. This provides
additional evidence for the link between attachment avoidance and repressive coping, which
explains that the more distressed an individual becomes, the more they suppress that emotion in

140

an attempt to mask any outward signs of distress (Diamond et al., 2006). The limiting of any
outward expression of emotional distress is also accompanied by self-reports of non-distress,
thus evidencing substantial incongruence between what is experienced physiologically and what
is actually reported.
From an attachment perspective, there are conceptual and empirical links between
avoidance and repressive coping. Sroufe and Waters (1977) found that, although avoidant
infants did not show any outward signs of needing comfort from their caregiver after being
separated during the Strange Situation, physiological measures indicated that they were just as
distressed as other infants. When caregiver-infant interactions were observed in the home,
avoidant infants as classified in the Strange Situation were more likely to have their expressions
of need met with rejection rather than responsiveness (Ainsworth, 1963; Ainsworth et al., 1978).
These infants develop an internal working model that relationships are not a safe haven from
which to seek comfort when distressed. This results in a preemptive approach to emotion
regulation ―involving disengagement of attention from distressing experiences before negative
affect has been encoded and experienced‖ (Diamond et al., 2006; see also Fraley, Garner, &
Shaver, 2000). When also taking into account the finding that avoidant individuals lacked
awareness of their own physiological signs of anger (Mikulincer, 1998), it appears that the
incongruence of experience found in this study relates more to an adaptive lack of awareness
than conscious misrepresentation.
Thus, it appears that one important clinical process with individuals who report higher
levels of attachment avoidance would be helping them develop greater self- and other-awareness.
One prerequisite to doing this is emotional safety within the session. Because individuals higher
in avoidance have developed their attachment strategies in response to a lack of safety in their

141

attachment relationships, fostering this within session is essential. Although ways to foster
relationship safety were not explicitly tested as part of the study, one consideration might be to
normalize the difficulty that these individuals have with safety in relationships and the adaptive
response of preemptively suppressing more vulnerable emotions. This can begin the process of
helping both the individual and partner become aware of and understand their relationship and
interaction patterns.
In addition, emotional suppression is only one approach to decrease the behavioral
expression of negative emotion, and it is associated with impairment in memory,
communication, and problem-solving (Richards & Gross, 1999; Richards, 2004). However,
reappraising and reinterpreting one‘s situation also limits the outward expression of negative
emotion, but without the negative consequences. In addition, research has found that reappraisal
actually decreases subjective distress (Gross, 1998). However, reappraisal is only possible once
these processes are brought into conscious awareness. In this manner, helping those higher in
attachment avoidance become aware and then learn to reappraise and reinterpret their situations
may be an effective way to help individuals high in avoidance to cope with distress.
Results from this study also seem to indicate that an intervention focused on emotional
exploration, validation, reframing, and softening may need to incorporate some important initial
steps to increase awareness and guard against the preemptive, nearly-automatic disengagement
from difficult emotions and distressing experiences typical of avoidance. That is not to say that
an emotionally-focused intervention may not be effective with those higher in attachment
avoidance. Rather, the intervention may need to be adapted to normalize their experience, help
them become more emotionally aware, and facilitate reappraisal and reinterpretation of their
experiences, all while gradually incorporating emotionally-focused interventions that will help

142

them appropriately signal their attachment needs and also receive responsiveness from their
partners. Future research will need to empirically test these potential implications. Until that
time, these clinical considerations should be considered tentative.
Research Implications
Process Research from a Common Factors Perspective
Having just addressed some of the most salient implications of this research for couple
relationship process as well as clinical work with couples, this section will discuss specific
research implications and how future research can build upon the findings of this study. From a
broad perspective, this study demonstrates the importance of process research using a common
factors lens. This study looked at how attachment (a client characteristic) within the context of
two interactional conditions (low structure, semi-natural and high structure, therapy-like)
influenced emotional experience that has previously been linked to positive outcomes. Findings
yielded some important information regarding couple interaction process as well as how
treatment process may be adapted to improve client experience.
It is hoped that future research will continue to look beyond identifying the specific
treatment models that most effectively treat specific diagnoses to looking at how individual (e.g.,
hope, motivation, individual distress), relationship (e.g., couple conflict, relationship
satisfaction), and contextual characteristics (e.g., life events, stressors) interact with one another
to influence couple relationship process, experiences, and outcomes. In addition, those
characteristics may also be combined with therapist characteristics (e.g., attachment, therapist
skill) and the therapeutic context (e.g., process, techniques) to understand more about how to
improve and adapt treatment in therapy. In sum, there are a vast array of potential research
opportunities in terms of couple relationship and therapy process from a common factors

143

perspective that can ultimately improve couple relationships.
Research on Attachment and Therapy
In addition to process research generally, this study has provided important evidence for
incorporating attachment theory into how we understand couple relationships and especially how
we conceptualize cases in therapy. Very little research has looked at the relationship between
attachment and therapy. This includes understanding the effects of therapy on attachment as
well as how attachment influences therapy. As mentioned earlier in this paper, empirical
research that analyzes how attachment influences client experience as well as outcome in therapy
will allow clinicians to appropriately adapt their approaches to specific client characteristics and
needs. Some of the most salient research questions regarding attachment that need to be
addressed are understanding further how to adapt treatment process to attachment avoidance.
This study provided some initial indications, but much more work is needed to really understand
what adaptations will yield better outcomes for couples where at least one partner is high in
attachment avoidance.
In a similar vein, research is needed not only to understand how to adapt treatment to
individual attachment characteristics, but also to understand how to adapt treatment to various
dyadic attachment constellations. For example, treating a couple where both partners are low in
attachment avoidance and anxiety will likely be fundamentally different from treatment of a
couple where both partners report high attachment anxiety or where one reports high anxiety
while the other reports high avoidance. This will be especially useful in helping therapists know
how to adapt treatment process in ways that will avoid being effective for one partner but
ineffective for the other. One other consideration that was not addressed in this study but is
extremely important is how therapist attachment interacts with client attachment to influence

144

outcomes (see Tyrrell et al., 1999). There is a strong likelihood that the therapist‘s own response
to distress and negative affect will greatly influence how s/he attends to those issues in therapy.
As more research is done in this area, it will allow training programs to help therapists explicitly
address these issues and be more aware of them in the process of therapy. Overall, the
opportunities for attachment-related research in these areas are plentiful. Nonetheless, a
concerted effort needs to be made to begin to address these process questions and issues.
Conclusion
Limitations and Suggestions for Future Research
Although this study has a number of strengths that have been demonstrated throughout
this paper, it is not without its limitations that point to the importance of future research in this
area. For example, this study included a diverse sample across a variety of demographic
variables. Nonetheless, it is difficult to know whether findings can be generalized to the
population as a whole for several reasons. One of the most important considerations involves
study participation. Slightly more than 50% of people who expressed interest in the study
actually ended up participating. Reasons for non-participation were that they did not meet the
inclusion criteria or elected later not to participate. It is unclear whether other unknown factors
limited the participation of these individuals and whether their results would have changed the
findings in any way. In addition, it is possible that the incentive of $100 may have led to a
selection bias that influenced the demographics of participants. However, it should be noted that
incentives, or the lack thereof, in any study may influence the sample. The best way to address
these questions is for future research to ask similar questions and attempt to replicate findings.
In addition to the potential limitations regarding the sample, it should be noted that this
study did not use an experimental design with a comparison group or a quasi-experimental

145

design with counterbalanced conditions. This was actually intentional, as an experimental design
would have been useful had the primary purpose of this study been to explicitly test the
effectiveness of the intervention. Rather, the goal of this study was to understand how
attachment influenced emotional experience within two interaction conditions. The intervention
used in the therapy-like condition focused on exploring and reprocessing emotions and reframing
experience in terms of primary emotions and attachment needs. The core components of the
intervention were used across a variety of theoretical models that have been proven effective. In
this manner, the focus was not on the effectiveness of the intervention, but rather how attachment
influences the experience of that intervention.
With respect to not counterbalancing conditions, the rationale was provided previously.
It was deemed more appropriate for each couple to first talk about the issue as they would at
home and then explore the issue with a therapist/coach. Although this prevents being able to
make a distinct comparison of the results between the semi-natural and therapy-like conditions, it
was considered a more acceptable confound than having the couples talk about the issue first
with a therapist/coach and then trying to talk about the issue as they would at home. As a result,
because this was not a randomized control trial or the conditions were not counter-balanced, it is
not possible to state that the findings demonstrated that the therapy-like condition was more
effective than the semi-natural condition. Rather, it can be more accurately stated that the
therapy-like condition as it followed the semi-natural discussion was more effective than the
semi-natural condition alone.
As mentioned previously, this study‘s findings provide some valuable preliminary
information about how to adapt therapy process to varying attachment styles. However, the
clinical considerations outlined as a result of these findings should be considered tentative until

146

additional research specifically tests them. This is especially true regarding attachment
avoidance. For example, unlike those who reported higher attachment anxiety, individuals
higher in attachment avoidance did not necessarily feel more positively towards their partners
during the therapy-like interaction. This may indicate that some preliminary steps are necessary
in order to improve outcomes. Some clinical considerations, anchored in existing research, were
provided that might improve outcomes. However, future research will need to test whether those
steps actually represent useful adaptations for those higher in attachment avoidance.
Summary
Despite these potential limitations and the need for further research, this study represents
an important contribution to the existing self-report attachment and therapy process literatures in
couple therapy. This study analyzed how one important client characteristic, self-report
attachment, was related to the emotion experience of partners within the context of semi-natural
and therapy-like conditions. Findings revealed important relationships between attachment and
emotional experience in the context of both conditions and point toward some important clinical
considerations. It is hoped that future research will build upon this study‘s findings by seeking
to replicate its findings as well as continuing to analyze how specific individual and couple
characteristics influence an individual‘s experience in his/her relationship and in the context of
couple therapy. This represents an important foundation upon which future process research and
research regarding the effectiveness of specifically tailored interventions may build.

147

APPENDICES

148

APPENDIX A

Brief Screening Assessment
1. How often do you and your partner disagree?
a. Less than once a month
b. Monthly
c. Weekly
d. Daily
2. When was your last disagreement? ______________
3. How often do you or your partner feel hurt, angry, or offended in your relationship?
a. Less than once a month
b. Monthly
c. Weekly
d. Daily
4. When did you or your partner last feel hurt, angry, or offended in your relationship?
______________
5. During disagreements in the past year, how often have you or your partner
a. been physical with one another in any way? _______
b. been physical such that it left marks? _______
c. been physical such that one of you had to go to the hospital? _______
If they have been physical but it has not left marks or required a hospital visit:
5d. When was the last time you or your partner were physical?
_______
5e. Do you feel that if you talk about difficult issues, it will threaten your safety?
6. Do you consider yourself in a personal or relationship crisis of any kind?
a. Yes
b. No
7. Are you seriously considering ending your relationship?
a. Yes
b. No

149

APPENDIX B
Facilitating Emotional Content: Common Treatment Processes
Theoretical Overview: Emotions in Couple Relationships and Therapy
This section will discuss the role of emotion in couple relationships and therapy.
Emotional experience in some form (e.g. empathy, positive emotions, emotional processing) has
consistently been identified in research as a core therapeutic process variable. More generally,
emotion has been described as having semantic, motivational, and adaptive functions (Frijda,
1986; Izard, 1991) in being ―involved with organizing, guiding, and driving perception, thought,
and action‖ (Izard, 1991, p. 51). In this manner, emotion may be considered a driving force for
many intrapersonal and interpersonal processes. These ideas have led Johnson and Greenberg
(1994) to the conclusion that emotion is the ―primary building block of adult intimate
relationships‖ (Johnson & Greenberg, 1994, p. 10).
Because emotion is so instrumental in providing intra- and interpersonal meaning and
organizing how an individual responds to a situation, the processing of felt emotion becomes a
vital adaptive skill that allows for the appropriate regulation of emotion. During periods of
arousal, emotional processing allows individuals to focus on, cope with, and potentially expand
and/or transform emotional experience and expression (Klein et al., 1986). However, when
individuals cannot adequately cope with their emotions or the arousal that they might produce,
they often attempt to over- or under-regulate them, leading them towards more negative,
relationship-disrupting expressions of affect.
In terms of couple relationships, Gottman (1999) identified several of these negative
emotional responses that appear related to later relationship dissolution: criticism, contempt,
defensiveness, and stonewalling. More generally, Gottman and Levenson (2002) found that,
over a 14-year period, couples who divorced early demonstrated high expressivity, anger, and
wife negativity (i.e., under-regulation). Those who divorced later during the longitudinal study
were more likely to demonstrate prolonged periods of neutral affect during conflict characterized
by physiological evidence of emotional suppression and withdrawal (i.e., over-regulation). The
alternative to withdrawal and hostility is softening. In emotionally focused therapy (EFT)
literature, softening is a bonding process contributing to positive outcomes (Bradley & Furrow,
2004; Johnson & Greenberg, 1988), whereby ―a newly vulnerable spouse reaches out to a now
accessible and engaged partner and asks for his or her attachment needs to be met‖ (Johnson,
2003a, p. 108).
In terms of psychotherapy, emotional processing and regulation have been described as
key factors that influence the facilitation of therapeutic change across models (Bridges, 2006;
Greenberg & Pascual-Leone, 2006). A deeper awareness and processing of emotions allows for
personal meaning to be explored (see also Wiser & Arnow, 2001) and impels one to work for
change (see also Linehan, 1993). Beutler et al. (2003) cite several major reviews that conclude
that ―emotional arousal will increase positive outcomes, especially if applied early in the
treatment‖ (p. 263). However, Greenberg and Pascual-Leone (2006) point out that some
incorrectly equate emotional arousal with felt intensity instead of representing emotional
experience or the depth of emotional processing. This level of processing can be achieved when,
in the context of an empathic and validating therapeutic relationship, the therapist helps clients to
experience, explore, and restructure their emotions in an effort to create new meaning
(Greenberg & Paivio, 2003).

150

Training Overview
Four coaches will be recruited from the Michigan State University Family and Child
Clinic. All coaches will be pursuing master‘s or doctoral degrees from a COAMFTE accredited
MFT program and have at least 200 hours of direct clinical experience. Training is expected to
consist of five two-hour segments occurring over a period of five weeks, with one-hour
―booster‖ training sessions occurring at least bi-weekly throughout the study to ensure that core
principles remain salient for coaches. Each training session will include both didactic and
experiential components (e.g., vignettes and role plays). The primary focus of training is not to
introduce an entirely new set of therapeutic skills but to emphasize the adaptation and use of the
coaches‘ existing skill set to accessing and re-processing emotion. Rather, the goal is for
coaches to be able to facilitate emotional content in a way that focuses on primary, attachmentbased emotions that invite a softened response from both partners. Training will be completed
when each coach can demonstrate, through role play, the ability to facilitate emotional content in
the context of a wide variety of couple dynamics.
Theoretical Underpinnings of the Coach’s Role
Coaches will be trained regarding their role and responsibility in the process, as well as
those necessary skills that will help couples process emotional content and exist across
therapeutic modalities (see below). These roles and responsibilities will be distinguished
according to two segments for each topic. For the sake of clarity, the person whose issue is
being discussed will be referred to as the speaking partner, while the person who did not
originate the issue will be labeled the listening partner. The core techniques the coach will
utilize during the interactions will be several of those identified by EFT for ―accessing and
reformulating emotion‖ (Johnson, 2004, p. 78). These include reflection, validation, evocative
responding, heightening, empathic conjecture, enactment, and promoting positive interaction.
However, a significant part of training will be to help coaches adapt these skills into their
existing therapeutic approach, without the expectation that they will be ―doing EFT‖ or any other
specific approach. Rather, during the training, coaches will learn to identify how some of the
common change mechanisms across modalities involve the accessing and re-processing of
emotions. Many of these skills are also utilized in some form by other empirically supported
couple therapies but with slightly different labels. For example, enhanced cognitive behavioral
therapy (Epstein & Baucom, 2002) provides guidelines of how to alter ―both the experience and
expression of emotions‖ (p. 375), including accessing, processing, and expanding the range of
primary emotions that have previously been minimized or avoided. Their interventions include
creating a safe environment; amplifying emotional experience through questions; help the client
recognize emotional meaning beyond what is stated; helping clients understand the relationship
between emotions, cognitions, and behaviors; and heightening their emotional experiences.
Behavioral marital therapy (Jacobson & Margolin, 1979) focuses a great deal on
decreasing interpersonal distress by enhancing positivity and helping the couple learn problemsolving skills. They do this by coaching each partner to learn to listen by expressing ―validation,
reflection, and generalized acceptance‖ (p. 206) to one another. In addition to learning the
importance of positive communication, clients also learn how to more appropriately express their
negative feelings to their partners. Insight-oriented marital therapy (Snyder, 2002) emphasizes
affective reconstruction in an effort to promote insight by ―enabling interventions of increasing
depth and emotional challenge‖ (Snyder, 1999, p. 350) that ultimately leads to increased
empathy and support (Snyder & Mitchell, 2008). Thus, each of these couple therapy approaches

151

utilize emotions in some way to help clients (1) access and understand their emotional
experience, (2) process, expand, and potentially transform that experience, and (3) utilize it to
then enact new patterns of interaction with their partners.
Therapy-like Interaction
During the therapy-like interaction, partners will discuss the same topics as in the seminatural interaction over a period of two segments per topic (lasting 3-4 minutes each). The
research associate who met previously with both partners and helped them decide on an issue
will enter the room and function as a process coach. The first speaker will have been previously
assigned the experimenter and the person who introduced the topic will be encouraged to share
his/her experience of the situation. The overall goal will be to help partners be engaged in and
―really hear‖ what is being expressed. In this manner, they are more likely to soften towards one
another as they experience their relationship in a way that increases self- and otherunderstanding.
Introduction with Speaking Partner:
Now that you have had some time to talk about each of your issues with each other, you are
each going to talk for 3-4 minutes about your issue with me. _______ (speaking partner) will
begin, and I will ask _______ (listening partner) to simply focus on listening and trying to
understand his/her experience. (To the listening partner) His/her experience may not be your
experience, and you may disagree, but I’ll ask that you wait to respond. I’m most concerned
right now in hearing your (speaking partner) experience, and then I will shift to understand
your (listening partner) experience.
Initial Approach:
Tell me a little bit about your experience of the situation.
Development of the interaction:
1. Emotional Exploration and Processing
Explanation: The coach will begin by helping the speaking partner explore his/her
experience, with a particular focus on emotional exploration and processing. As the speaking
partner explains his/her experience, the coach models healthy interaction for the listening partner
by validating and seeking to understand the speaking partner‘s experience as s/he reports it.
Although this may mean occasionally validating secondary emotions (e.g., anger, frustration,
etc.), it is extremely important that during this time the coach look beyond expressed content for
primary emotions (e.g., sadness, hurt, etc.) and underlying attachment needs, threats, or longings
(the need to feel understood, safe, important, respected, etc.).
Explore their experience (emotional, cognitive, and behavioral)
What was that situation like for you?
What thoughts were going through your mind when that happened?
How did you respond?
Validate their experience as they report it.
That sounds really hard.
So you were really angry when s/he responded that way.

152

Deepen processing by shifting to and validating primary emotions and attachment
needs.
So when s/he responded that way, it seems like you might have felt
disrespected/unimportant/alone/unheard.
In your situation, I may have felt ___________ (primary emotion/attachment need)
So your experience is when you needed him/her most, s/he wasn’t there.
2. Emotion Regulation
Explanation: As emotional processing continues and those more vulnerable emotions
become clear, the coach briefly addresses emotion regulation by reframing the person‘s
experience and behaviors in terms of the newly developed emotional context. At times, it may
be appropriate to explore the level of congruence between the speaking partner‘s behaviors at the
time of the issue and his/her underlying, primary emotions. This helps to establish a link for
both partners between two seemingly contradictory experiences (i.e., primary emotions and
attachment needs versus reactive and volatile behavior).
Reframe motivation and behaviors in terms of primary emotions
How do you typically respond when you feel ____ (primary emotion/attachment need)?
So when you’re feeling _______, you most often respond by _______ to your partner.
Check the congruency between the experience of primary emotions and what is
typically expressed to the listening partner.
Do you think your partner knew you were feeling _________?
3. Enactment
Explanation: The final step is to facilitate a brief, emotionally-heightened experience for
the couple. Whereas all of the interaction to this point has been between the speaking partner
and the coach, this exchange is between the speaking and listening partners. After briefly
summarizing the speaking partner‘s experience in terms of primary emotions and ensuring that it
is accurate, the coach instructs the speaking partner to express the thought in his/her own words
directly to the listening partner. This creates a much deeper emotional experience than solely
expressing the idea to the coach.
Briefly summarize (in one sentence and in positive terms) the speaking partner’s
experience, focusing on the primary emotion and attachment needs.
So you really want to feel important to your partner, and ______ (experience) is one
way that helps you know s/he thinks you are important?
Check to ensure that your summary fits the speaking partner’s experience
Does that fit?
Direct the speaking partner to express that idea in their own words to the listening
partner.
Will you please turn to him/he and say that in your own words, please?

153

Interaction with the Listening Partner
After the expression from the speaking partner, the coach shifts to the listening partner
and understanding what the previous few minutes were like for him/her. It may also be useful to
allow the listening partner to express what s/he heard the speaking partner say. Nonetheless, the
overarching framework remains the same. The coach facilitates emotional processing by
validating and exploring the listening partner‘s experience (as it relates to the speaking partner‘s
experience). The coach then looks for primary emotions and underlying attachment needs in
order to reframe both partners‘ motivation and behaviors. After a brief coach summary of the
listening partner‘s emotional experience, the coach directs him/her to express those thoughts
directly to the speaking partner.
Introduction with Speaking Partner:
Now we’re going to shift and I’m going to talk for 3-4 minutes with _______ (previously the
listening partner) about his/her experience. So please (previously the speaking partner)
simply listen and try to understand his/her experience, even if you may not completely agree.
Initial Approach:
What were the previous few minutes like for you?
Please note that there are two primary paths that the listening partner may take: (1) S/he may discuss their
explicit experience of the previous 3-4 minutes; or (2) S/he may shift and talk about his/her experience
during the situation to which the speaking partner was referring. Either reaction is appropriate.

Development of the interaction:
1. Emotional Processing
Explore their experience (emotional, cognitive, and behavioral)
What thoughts were going through your mind when that happened?
How might you have responded if you had been at home?
Validate their experience as they report it.
So what s/he just said doesn’t really fit your experience.
So you feel like you’ve heard this all before.
So you feel kind of bad after hearing him/her express it in that way.
Deepen processing by shifting to and validating primary emotions and attachment
needs.
So it sounds like you are a bit worried that s/he’s not being sincere.
It seems like I might feel like I was not good enough if I heard those things from him/her.
2. Emotion Regulation
Reframe motivation and behaviors in terms of primary emotions
So when s/he typically feels _______ (primary emotion/attachment need), how do you
respond?
So when you’re feeling _______, you most often respond by _______ to your partner.
Check the congruency between the experience of primary emotions and what is

154

typically expressed to the listening partner.
Do you think your partner knew you were feeling _________?
How might you respond differently if you heard what s/he just expressed to you instead?
So even though you feel kind of bad, you’re afraid your partner really doesn’t feel you
are good enough, and you kind of shut down.
3. Enactment
Briefly summarize (in one sentence and in positive terms) the speaking partner’s
experience, focusing on the primary emotion and attachment needs.
It sounds like you really want to know that, even when she might be upset at you, she
still values and admires you.
Check to ensure that your summary fits the speaking partner’s experience
Does that fit?
Direct the speaking partner to express that idea in their own words to the listening
partner.
Will you please turn to him/he and say that in your own words, please?

155

APPENDIX C
Research Associate Interaction with Participants
PHASE 1:
Ryan and Interviewers:
1.
Couple arrives and is greeted.
2.
Tell/show the couple where the water bottle, snack packs, and bathrooms are located.
3.
Ask them if they are willing to turn their cell phones off, or if they would like they can
leave them with Ryan and they can remain on. He will answer them to ensure that there are
no emergency calls.
Interviewers:
4.
Each interviewer goes into a separate room with one participant.
5.
Informed Consent
The first part of this process is for you to read a basic overview of this study and provide
your consent to participate. You are welcome to read it silently, or I am happy to read it
for you. Do you have a preference?
Ask if s/he has any questions about the informed consent
I just want to highlight a few things that are really important. One thing is that what you
do here is completely confidential. You will be assigned a participant number, and no
names or identifying information will be attached to any of your materials, including the
video. In addition, everything will be locked up in a secure place that only the primary
investigators can access. Also, if at any time you feel unable to continue, please notify
any member of the research team, and we will let you have a break or end the research
session. Do you have any questions for me?
6.

You were asked this in the brief telephone survey, but I wanted to ask again: Has there
been a physical altercation of any kind between you and your partner since the telephone
interview? If yes, make sure it was not something that left marks or required a trip to the
hospital. If it was, then they cannot participate. If it was less severe, verify that the
participant considers it safe to participate in the research and discuss potentially difficult
issues.
7.
The research associate then administers the Adult Attachment Interview.
For the next hour or so, I am going to ask you some questions about your early
childhood, especially about your relationship with your parents. There are 20 questions
with a variety of additional questions that I will ask. It is likely that I will ask you a
question you have already answered earlier in the interview. The interview is very
structured, so I will have to ask you the question, but just know that I am listening, and
I’ll probably say something like, “I know you already kind of mentioned this, but…” Do
you have any questions for me before we begin?

156

8.

You will then start the recorder. The first step is to take it off of hold (on back). After
that, all you need to do is hit the record button (red) and make sure that it says it is recording
and begins to count up. You will also need to have made sure the batteries are not low.

Equipment needed for interviewers:
1.
Informed consent
2.
Digital voice recorder
3.
Paper/Pen
PHASE 2:
Coach:
1.
You will give both partners the assessment packet to complete (SEEQ, CPS, RDAS, OQ45.2).
2.
While one partner begins the packet, you will invite the other into room 1.
3.
Use this opportunity to talk briefly with the partner and establish some rapport.
4.
In room 1, you will say the following:
This meeting with me is to prepare for the interaction with your partner. I would like
you to think of a fairly recent time when you felt hurt, angry, or offended by your
partner, and about which you still have at least some feelings. What recent instance
comes to mind when you think about that?
Let them know that they can take their time and that they can brainstorm and find one
instance about which they would like to talk. After they come up with one, ask if another
one comes in mind—a back-up topic, so to speak. They need to leave this brief meeting
with you having at least one topic.
Ask them to rate their distress level (1-10) regarding the issue, with 10 extremely high
distress and 1 being very little distress. Make sure that the report a distress level of 37. If it falls outside of those ranges, explore whether a different topic would be better.
VERIFY WHETHER THEY FEEL SAFE ABOUT DISCUSSING THAT TOPIC WITH
THEIR PARTNER. If they do not, please help them choose another issue.
5.

Thank the partner for his/her time and take them back to where partner 2 is completing
the assessment packet. Then take partner 2 into room 1 and repeat the process.
6.
Tell the couple that you will be attaching a measure of skin conductance to their index
and middle fingers of their non-dominant hand and encourage them to wash their

hands prior to entering room 1.
Equipment needed for the therapist/coach:
1.
An assessment packet for each partner.
PHASE 3:
Ryan:
1.
Open up AcqKnowledge software (it should automatically detect the hardware: #000911)
2.
Make sure hardware is set to 5, 1.0 Hz, DC, and DC.
3.
Select MP150 > Set Up Channels

157

4.
5.
6.

Make sure both A1 and A2 channels are selected, with all boxes checked
Select ―Scaling‖ (You will need to do this separately for each channel)
Make sure that Cal1 ―Input‖ and ―Scale‖ are both set at 0. Cal2 ―Input‖ should be 1 and
―Scale‖ should be 5. The units are ―micromhos.‖
7.
Click on Cal1. It will change the input volts value slightly.
8.
Add the input volts for Cal1 to the value for Cal2 (1).
9.
Click on the mean value box, and then select settings. Select 10 as the number of
samples.
10.
Click on ―ok.‖
11.
Select MP150 > Set Up Acquisition
12.
It should read ―Record‖ and ―Append‖ to ―Disk.‖ The Sample Rate should be 10
samples per second.
13.
The total length of all interactions should be set for 40 minutes (I can start and stop
recording and it will put a small diamond at the top of the graph.
14.
Select MP150 > Set Up Stimulator
15.
Under ―Parameters,‖ change Seg #1 Level to 7 volts and Seg #1 Width to 2000 msec.
Then change the Levels of all the others to 0 volts (it won‘t matter what their corresponding
widths are then).
16.
Under ―Output‖, select A0 and under ―Duration,‖ select 1x. Make sure it says that it will
start with Acquisition and that the sampling rate is 200 samples/sec or higher.
17.
Let the couple know that they will be in this room for the next 45-50

minutes, and that they will not be able to answer their cell phone or go to the
bathroom. If they need their cell phone on, have them give it to Ryan.
18.

Partners will then both enter room 1 and the physiological equipment will be attached to
the index and middle fingers of their non-dominant hand.
19.
They will then have about five minutes of free time.
In order for the physiological equipment to work properly, we need to give it about five
minutes, so we’ll take a brief break and we’ll start back up in about five minutes.
20.
21.

Turn the preview monitor on and make sure the source is ―AV IN.‖
To see the couple interaction on the preview monitor, make sure the switch is on
―Camera,‖ and press ―Mode‖ twice and ―Fade‖ once on the CS-450.

Coach:
22.
They will then be asked to clear their minds and relax for about 3 minutes.
You’ve already talked about quite a few things today. Now what I’d like you to do is take
about three minutes and focus on relaxing and clearing your mind for the next three
minutes.
Ryan:
23.

Start recording physiological data.

158

Coach:
24.

At the conclusion of the three minutes, enter the room and begin recording with
both cameras (use remote).

Ryan:
25.

Stop recording physiological data when the coach enters. Count to 15.
Begin collecting physiological data again.

Coach:
26.
Instruct the couple that they will then talk for 5 minutes about how they met.
For the next five minutes, I’d like for you to talk with one another about how you met.
For example, you can talk about your first impressions, your dating experience, what
you valued in one another, and how you decided to stay together.

27.

28.

Let them know that you will be present in the room as they discuss how they met and
discuss an issue as they would at home.
Help them to know that you will simply be off to the periphery and will not be involved
in what they are doing.
Validate any potential anxiety they might feel at having you in the room.
Turn your body position away from the couple and do not look at them. Engage yourself
in something else. You will be timing their interaction.
If at any time they try to engage you. Let them know that they need to be talking with
one another, and that they will have the opportunity to talk to you in the next segment.
Begin recording with both cameras.
Coach:
They will then begin talking about one of the partner‘s issues (previously assigned) for 78 minutes).
Now I’d like (predetermined partner) to talk with your partner about your issue for about
7-8 minutes. Then you will talk for 7-8 minutes about (to partner) your issue. The goal
of talking with your partner is to talk about each others’ experience and to take steps
towards a resolution.

29.
30.

They will then shift and talk about the other partner‘s issue for 7-8 minutes.
You will then talk to one partner for 3-4 minutes about his/her issue, with a focus on
facilitating emotional content and a softening experience.
Now, I’m going to talk directly with each of you about the issues you have been
discussing for a few minutes.

159

31.

You will then shift and talk to the other partner for 3-4 minutes about his/her experience
of hearing his/her partner express those thoughts and feelings.
32.
You will then shift and do the same two segments for the other partner‘s issue.
33.
After you finish discussing both partners‘ issues, stop video recording.
34.
Thank them for their time and then excuse yourself. Let them know that Ryan will be in
shortly.
Ryan:

Stop recording physiological data when the coach exits. Begin collecting
physiological data again.
35.
36.

Enter the room and provide each partner with the SOFTA. Ask them to answer each
question according to the interaction they just had with the coach.
37.
After they have finished the SOFTA, remove the skin conductance.
38.
Let them know they can take a 7 minute break. Tell them that they will need to go to the
bathroom now or at the conclusion of the study.
Equipment needed for the therapist/coach:
1.
Timer
2.
Biopac skin conductance measure
3.
Two copies of the SOFTA-S (adapted)
4.
Pens
PHASE 4:
Ryan:
1.
Put up the divider sheet.
2.
Turn on headphones and amplifier.
3.
Turn on TV
4.
Make sure the computer monitor resolution is 1024*768
5.
Make sure the following are turned on:
PA Dials
DVD Recorder
6.
Make sure there is a DVD in the recorder.
7.
Select the ―PA Launch‖ icon
8.
Select ―Collect‖
9.
Select the appropriate session (or create one to represent the participant ID)
10.
Take a few moments and instruct the participants about the perception analyzer
and how it works.
11.
Select the ―What is your gender?‖ question.
12.
To see on the preview monitor what participants are seeing in terms of perception
analyzer questions, make sure the switch is on ―PA Questions‖ and press ―Mode‖ twice and
―Fade‖ once on the CS-450.
13.
Take a reading on the gender question by pressing F4 on the computer. The computer
screen should then switch to the results for that question.
14.
Select the ―Please look at the bottom of your dial and enter the dial number‖ question.
15.
After giving the participants a few seconds to answer it, take a reading by pressing F4 on
the computer. Again, the computer screen should show the results.

160

16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.

Select the moment-to-moment question.
Go to ―Questions‖
Select ―Subset Options.‖
Click on ―Select Subset‖
Choose ―What is your gender‖ as the question to use as the subset.
Select ―OK.‖
Select ―Camera‖ on the switcher.
Press ―Mode‖ twice (no ―Fade‖)
Using the ―tvone‖ remote, press ―Autoset‖
Then press ―Control F‖
Go into the room with the participants and let them know that they will be indicating, 0100, how they felt toward their partner in the couple interaction. They are not indicating
their overall feelings in their relationship, only what they felt during the interaction.
27.
Start the camera playback.
28.
Quickly go into the research room and watch the preview monitor for

the small LED light to turn on. When it turns on, press F4 on the computer
and “Record” on the DVD recorder.
29.
At the conclusion of the playback, press F4 on the computer and ―Stop‖ on the DVD
recorder.
30.
Enter room 1 and give the participants the SEEQ for the segments they just watched.
31.
Then go through the debriefing process and thank them for their time.
Equipment needed for the research associate:
1.
Divider screen
2.
Two sets of ear phones
3.
Two perception analyzer dials
4.
Two copies of the SEEQ
To switch to the overlay mode:
1. Change switcher to ―Camera‖
2. Press ―Mode‖ twice (no ―Fade‖)
3. Using the ―tvone‖ remote, press ―Autoset‖
4. Then press ―Control F‖
5. Begin the video, when the green light turns on, then press ―F4‖, and begin recording
DVD.
(NOT NEEDED) To switch to next moment-to-moment question:
1. Change switcher to ―PA Questions‖
2. Press ―Fade‖
3. Select the appropriate question.

161

APPENDIX D
Informed Consent
WHAT WE ARE DOING:
This is a study by Ryan Seedall and Dr. Karen Wampler of Michigan State University. We want
to understand what affects your emotional experience when you interact with your partner. We
expect to get this information by (1) having you fill out several surveys, (2) interviewing you
about your early childhood with your parents, and (3) having you talk with your partner.
WHY WE ARE DOING THIS:
We really want to understand more about how couples interact and learn about possible ways
that we can improve their experience when they come in for therapy.
WHAT WILL HAPPEN:
1. If you provide your consent to participate, you will be interviewed alone about your early
childhood experiences with your parents. This 45-90 minute interview has been used in many
studies and is called the Adult Attachment Interview. It will be digitally recorded and later
transcribed and studied by Ryan Seedall and his research associates.
2. After the interview, you and your partner will meet individually with a research associate who
will ask you to think about a recent time when you felt hurt, angry, or offended by your partner
and about which you still have feelings. You will also fill out some surveys about your
individual and couple functioning while your partner is talking with the research associate.
3. You and your partner will then enter the same room and have a painless measure of skin
conductance attached to your middle and ring fingers of one of your hands. You will first focus
on clearing your mind and relaxing for 3 minutes and then talk with your partner for 5 minutes
about how you met one another.
4. After that, you and your partner will talk about each of your issues in a couple of different
ways. One way you will do this is to talk directly with your partner for 8-10 minutes per issue.
A research associate will be in the room but will not participate in the conversation.
5. The other way is that you will talk about your issue with a research associate who will act as a
communication coach. When it is your issue, you and the coach will talk for 3-5 minutes about
your experience while your partner listens. The coach will then shift and talk to your partner for
3-5 minutes to help him/her better understand your experience. The same thing will happen
when you discuss your partner‘s issue.
6. Your conversations with your partner will be video recorded, and the last part of your
participation will be to watch yourself talking with your partner. As you do this, you will rate
with a dial how you felt towards your partner in each moment.
7. We will take breaks as needed throughout the 3-4 hour process.
CONFIDENTIALITY:
Your confidentiality will be protected to the maximum extent allowable by law. All of your
information will be stored in a locked file cabinet. An identification number will be used on all
of your materials so that your name will not be used. Only Ryan Seedall, Dr. Karen Wampler,
and their research associates will see any of your information. No one else, including your
partner, will have access to your personal information.

162

RISKS:
We do not expect any significant risks for participating in this research, but it is possible that you
may experience some uncomfortable or upsetting feelings when you talk about your early
childhood experiences (including potential abuse, loss, or trauma) as well as during your
conversation with your partner. You are free to take breaks or stop participating in the study at
any time. If you choose to stop, you will still receive full compensation. We may also stop the
research session if we feel that your overall emotional or physical well-being is in jeopardy.
At the conclusion of the research session, you will briefly speak with a marriage and family
therapist who will make sure that no emotional or relationship problems have occurred because
of your participation in this study. At that time, he will provide you with information regarding
potential resources if any upset feelings persist. This includes meeting with someone from the
Family and Child Clinic (517-432-2272) who can talk with you. We will also contact you within
two weeks of your participation to request feedback and to check on your well-being.
BENEFITS:
Although this research is not intended to be therapeutic or permanently improve your
relationship, it is possible that you may experience some positive effects from participating in
this research. For example, you may learn more about your own as well as your partner‘s
emotional experience from participating in this research. Each partner will receive a $50 gift
card prior to leaving today, even if you choose to end your participation early.
IF YOU NEED ANYTHING:
If you have any concerns or questions about this research study, such as scientific issues, how to
do any part of it, or if you believe you have been harmed because of this research, please contact
Karen Wampler (Room 7, Human Ecology Building, MSU, East Lansing, MI 48824;
kwampler@msu.edu; 517-355-0230) or Ryan Seedall (Room 1, Human Ecology Building, MSU,
East Lansing, MI 48824; seedallr@msu.edu; 517-432-0893).
If you have questions or concerns about your role and rights as a research participant, would like
to obtain information or offer input, or would like to register a complaint about this study, you
may contact, anonymously if you wish, the Michigan State University‘s Human Research
Protection Program at 517-355-2180, Fax 517-432-4503, or email irb@msu.edu or regular mail
at 202 Olds Hall, MSU; East Lansing, MI 48824.
CONSENT:
I have read this form and had my questions answered by a research associate.
I give my consent to participate in this project and have been given my own copy of the
consent form.
________________________________________________
Participant Name (Please print)
________________________________________________
Participant Signature

Date

________________________________________________
Karen S. Wampler, Project Director, or Authorized Rep.

Date

________________________________________________
Witness – Research Associate

Date

163

APPENDIX E
Therapist/Coach Content and Quality Check
Directions: Please watch all of the partner-coach segments and complete one sheet for each
partner. After you have finished a rating sheet for each partner, watch the partner-coach
segments one more time to verify your ratings.
1. When talking about this partner‘s issue, the coach explored his/her feelings and experience.
0: Not used or ineffectively used
1: Rarely used or for the most part ineffective
2: Used effectively in some instances but poorly in others
3: Something missing but used effectively for the most part
4: Used effectively in virtually all instances
2. When talking about this partner‘s issue, the coach communicated in a way that validated
his/her feelings and experience.
0: Not used or ineffectively used
1: Rarely used or for the most part ineffective
2: Used effectively in some instances but poorly in others
3: Something missing but used effectively for the most part
4: Used effectively in virtually all instances
3. When talking about this partner‘s issue, the coach helped to reframe his/her feelings and
experience in terms of primary emotions and attachment needs.
0: Not used or ineffectively used
1: Rarely used or for the most part ineffective
2: Used effectively in some instances but poorly in others
3: Something missing but used effectively for the most part
4: Used effectively in virtually all instances
4. When talking about this partner‘s issue, the coach utilized an enactment to heighten his/her
emotional experience.
0: Not used or ineffectively used
1: Rarely used or for the most part ineffective
2: Used effectively in some instances but poorly in others
3: Something missing but used effectively for the most part
4: Used effectively in virtually all instances
5. How much did this partner soften towards the other person during this segment?
0: This partner did not soften at all in either segment.
1: This partner softened very minimally towards the other person.
2: This partner had a balance of softening moments coupled with defensive or volatile
moments.
3: This partner softened moderately, with minimal defensiveness or volatility.
4: This partner softened substantially towards the other person.

164

6. How effective did the intervention seem for this partner during this segment?
0: The intervention had no effect or was ineffective.
1: The intervention had small moments of effectiveness but was mostly
ineffective.
2: The intervention had a balance of effective and ineffective moments.
3: The intervention had small moments of ineffectiveness but was mostly effective.
4: The intervention was very effective.

165

REFERENCES

166

REFERENCES

Ainsworth, M. D. S. (1963). The development of infant-mother interaction among the Ganda. In
B. M. Foss (Ed.). Determinants of infant behaviour, II (pp. 67-112). New York: John
Wiley & Sons.
Ainsworth, M. D. S. (1991). Attachments and other affectional bonds across the life cycle. In C.
M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment across the life cycle (pp.
33-51). London: Routledge.
Ainsworth, M. D. S., & Bell, S. M. (1970). Attachment, exploration, and separation: Illustrated
by the behavior of one-year-olds in a strange situation. Child Development, 41, 49-67.
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A
psychological study of the strange situation. Hillsdale, NJ: Erlbaum.
Alexander, J. F., Holtzworth-Munroe, A., & Jameson, P. (1994). The process and outcome of
marital and family therapy: Research review and evaluation. In A. E. Bergin & S. L.
Garfield (Eds.) Handbook of psychotherapy and behavior change (4th ed., pp. 595-630).
New York: Wiley & Sons.
Allgood, S. M., & Crane, D. R. (1991). Predicting marital therapy dropouts. Journal of Marital
and Family Therapy, 17, 73-79.
Anderson, S. A., Atilano, R. B., Bergen, L. P., Russell, C. S., & Jurich, A. P. (1985). Dropping
out of marriage and family therapy: Intervention strategies and spouses‘ perceptions. The
American Journal of Family Therapy, 13, 39-54.
Andreassi, J. L. (2007). Electrodermal Activity and Behavior. In J.L. Andreassi (Ed.).
Psychophysiology: Human behavior and physiological response. (5th ed., pp.259-288).
New Jersey: Lawrence Erlbaum Associates.
Babcock, J. C., Jacobson, N. S., Gottman, J. M., & Yerington, T. P. (2000). Attachment,
emotional regulation, and the function of marital violence: Differences between secure,
preoccupied, and dismissing violent and nonviolent husbands. Journal of Family
Violence, 15, 391-409.

167

Barber, J. P., & Muenz, L. R. (1996). The role of avoidance and obsessiveness in matching
patients to cognitive and interpersonal psychotherapy: Empirical findings from the
Treatment for Depression Collaborative Research Program. Journal of Consulting and
Clinical Psychology, 64, 951-958.
Bartholomew, K. (1990). Avoidance of intimacy: An attachment perspective. Journal of Social
and Personal Relationships, 7, 147-178.
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.
Bartholomew, K., & Shaver, P. R. (1998). Measures of attachment: Do they converge? In J. A.
Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 25-45).
New York: Guilford Press.
Bartle-Haring, S., Glebova, T., & Meyer, K. (2007). Premature termination in marriage and
family therapy within a Bowenian perspective. American Journal of Family Therapy, 35,
53-68.
Benoit, D., & Parker, K. C. H. (1994). Stability and transmission of attachment across three
generations. Child Development, 65, 1444-1456.
Beutler, L. E., Malik, M., Alimohamed, S., Harwood, T. M., Talebi, H., Noble, S., & Wong, E.
(2003). Therapist variables. In M. J. Lambert (Ed.), Bergin and Garfield’s handbook of
psychotherapy and behavior change (5th ed., pp. 227-306). New York: John Wiley &
Sons.
Bick, J., & Dozier, M. (2008). Helping foster parents change: The role of parental state of mind.
In H. Steele & M. Steele (Eds.), Clinical applications of the Adult Attachment Interview
(pp. 452-470). New York: Guilford Press.
Biocca, F., David, P., & West, M. (1994). Continuous response measurement (CRM): A
computerized tool for research on the cognitive processing of communication messages.
In A. Lang (Ed.), Measuring psychological responses to media messages (pp. 15-64).
Hillsdale, NJ: Lawrence Erlbaum.
Bischoff, R. J., & McBride, A. (1996). Client perceptions of couples and family therapy. The
American Journal of Family Therapy, 24, 117-128.

168

Bischoff, R. J., & Sprenkle, D. H. (1993). Dropping out of marriage and family therapy: A
critical review of research. Family Process, 32, 353-375.
Blow, A. J., Sprenkle, D. H., & Davis, S. D. (2007). Is who delivers the treatment more
important than the treatment itself? The role of the therapist in common factors. Journal
of Marital and Family Therapy, 33, 298-317.
Boden, J. M., & Baumeister, R. F. (1997). Repressive coping: Distraction using pleasant
thoughts and memories. Journal of Personality and Social Psychology, 73, 45-62.
Bouthillier, D., Julien, D., Dube, M., Belanger, I., & Hamelin, M. (2002). Predictive validity of
adult attachment measures in relation to emotion regulation behaviors in marital
interactions. Journal of Adult Development, 9, 291-305.
Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation. New York: Basic Books.
Bowlby, J. (1982). Attachment and loss: Vol. 1. Attachment, (2nd ed.). New York: Basic Books.
Bowlby, J. (1988). A secure base: Clinical applications of attachment theory. New York: Basic
Books.
Bowlby, J., & Parkes, C. M. (1970). Separation and loss within the family. In E. J. Anthony & C.
Koupernik (Eds.), The child in his family: International Yearbook of Child Psychiatry
and Allied Professions (pp. 197-216). New York: Wiley
Bradley, B., & Furrow, J. L. (2004). Toward a mini-theory of the blamer softening event:
Tracking the moment-by-moment process. Journal of Marital and Family Therapy, 30,
233-246.
Bradley, M. M., & Lang, P. J. (2007). Emotion and motivation. In J. T. Ciacioppo, L. G.
Tassinary, & G. G. Berntson (Eds.), Handbook of psychophysiology (pp. 581-607).
Cambridge: Cambridge University Press.
Brennan, K. A., Clark, C. L., & Shaver, P. R. (1998). Self-report measurement of adult
attachment: An integrative overview. In J. A. Simpson & W. S. Rholes (Eds.),
Attachment theory and close relationships (pp. 46-76). New York: Guilford Press.

169

Bridges, M. R. (2006). Activating the corrective emotional experience. Journal of Clinical
Psychology: In Session, 62, 551-568.
Buck, R., Losow, J. I., Murphy, M. M., & Costanzo, P. (1992). Social facilitation and inhibition
of emotional expression and communication. Journal of Personality and Social
Psychology, 63, 962-968.
Buehlman, K. T., Gottman, J. M., & Katz, L. F. (1992). How a couple views their past predicts
their future. Predicting divorce from an oral history interview. Journal of Family
Psychology, 5, 295-318.
Buhringer, G. (2005). Allocating treatment options to patient profiles: Clinical art or science?
Addiction, 101, 646-652.
Busby, D. M., Christensen, C., Crane, D. R., & Larson, J. H. (1995). A Revision of the dyadic
adjustment scale for use with distressed and nondistressed couples: Construct hierarchy and
multidimensional scales. Journal of Marital and Family Therapy, 21, 289–308.
Butler, M. H., Davis, S. D., & Seedall, R. B. (2008). Common pitfalls of beginning therapists
utilizing enactments. Journal of Marital and Family Therapy, 34, 329-352.
Butler, M. H., & Gardner, B. C. (2003). Adapting enactments to couple reactivity: Five
developmental stages. Journal of Marital and Family Therapy, 29, 311-327.
Butler, M. H., & Wampler, K. S. (1999). Couple-responsible therapy process: Positive proximal
outcomes. Family Process, 38, 27-54.
Campbell, L., Simpson, J. A., Boldry, J. & Kashy, D. A. (2005). Perceptions of conflict and
support in romantic relationships: The role of attachment anxiety. Journal of
Personality and Social Psychology, 3, 510-531.
Carpenter, E. M., & Kirkpatrick, L. A. (1996). Attachment style and presence of a romatic
partner as moderators of psychophysiological responses to a stressful laboratory
situation. Personal Relationships, 3, 351-367.
Carrere, S., Buehlman, K. T., Gottman, J. M., Coan, J. A., & Ruckstuhl, L. (2000). Predicting
marital stability and divorce in newlywed couples. Journal of Family Psychology, 14, 4258.

170

Cassidy, J. (1994). Emotion regulation: Influences of attachment relationships. Monographs of
the Society for Research in Child Development, 59, 228-283.
Byng-Hall, J. (1995). Creating a secure family base: Some implications of attachment theory for
family therapy. Family Process, 34, 45-58.
Byng-Hall, J. (2001). Attachment as a base for family and couple therapy. Child and Adolescent
Mental Health, 6, 31-36.
Byng-Hall, J., & Stevenson-Hinde, J. (1991). Attachment relationships within a family system.
Infant Mental Health Journal, 12, 187-200.
Clarkin, J. F., & Levy, K. N. (2003). The influence of client variables on psychotherapy. In M. J.
Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change
(5th ed., pp. 194-226). New York: John Wiley & Sons.
Christensen, A., & Heavey, C. L. (1999). Interventions for couples. Annual Review of
Psychology, 50, 165-190.
Christensen, L. L., Russell, C. S., Miller, R. B., & Peterson, C. M. (1998). The process of
change in couples therapy: A qualitative investigation. Journal of Marital and Family
Therapy, 24, 177-188.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (Rev. ed.). Hillsdale, NJ:
Erlbaum.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation
analysis for the behavioral sciences (3rd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Cohn, D. A., Silver, D. H., Cowan, C. P., Cowan, P. A., & Pearson, J. (1992). Working models
of childhood attachment and couple relationships. Journal of Family Issues, 13, 432449.
Collins, N. L., Cooper, M. L., Albino, A., & Allard, L. (2002). Psychosocial vulnerability from
adolescence to adulthood: A prospective study of attachment style differences in
relationship functioning and partner choice. Journal of Personality, 70, 965-1008.

171

Collins, N. L., Ford, M. B., Guichard, A. C., & Allard, L. M. (2006). Working models of
attachment and attribution processes in intimate relationships. Personality and Social
Psychology Bulletin, 32, 201-219.
Collins, N. L., & Read, S. J. (1990). Adult attachment, working models, and relationship quality
in dating couples. Journal of Personality and Social Psychology, 58, 644-633.
Cooper, G., Hoffman, K., Powell, B., & Marvin, R. (2005). The Circle of Security Intervention:
Differential diagnosis and differential treatment. In L. J. Berlin, Y. Ziv, L. AmayaJackson, & R. Marvin (Eds.), Enhancing early attachments: Theory, research,
intervention, and policy. (pp. 127-151). New York: Guilford Press.
Creasey, G. (2002). Associations between working models of attachment and conflict
management behavior in romantic couples. Journal of Counseling Psychology, 49, 365375.
Creasey, G., & Hesson-McInnis, M. (2001). Affective responses, cognitive appraisals, and
conflict tactics in late adolescent romantic relationships: Associations with attachment
orientations. Journal of Counseling Psychology, 48, 85-96.
Creasey, G., Kershaw, K., & Boston, A. (1999). Conflict management with friends and
romantic partners: The role of attachment and negative mood regulation expectancies.
Journal of Youth and Adolescence, 28, 523-543.
Crowell, J. A., Fraley, R. C., & Shaver, P. R. (1999). Measurement of individual differences in
adolescent and adult attachment. In J. Cassidy & P. R. Shaver (Eds.), Handbook of
attachment: Theory, research, and clinical applications (pp. 434-465). New York:
Guilford Press.
Crowell, J. A., Treboux, D. (2001). Attachment security and adult partnerships. In C. Clulow
(Ed.), Adult Attachment and couple psychotherapy: The “secure-base” in practice and
research (pp. 28-42). London: Brunner-Routledge.
Crowell, J. A., Treboux, D., & Waters, E. (1999). The Adult Attachment Interview and the
Relationship Questionnaire: Relations to reports of mothers and partners. Personal
Relationships, 6, 1-18.
Cyranowski, J. M., Bookwala, J., Feske, U., Houck, P., Pilkonis, P., Kostelnik, B., et al. (2002).
Adult attachment profiles, interpersonal difficulties, and response to interpersonal

172

psychotherapy in women with recurrent major depression. Journal of Social and Clinical
Psychology, 21, 191-217.
Davis, S. D., & Butler, M. H. (2004). Enacting relationships in marriage and family therapy: A
conceptual and operational definition of an enactment. Journal of Marital and Family
Therapy, 30, 319-333.
Dawson, Schell, & Filion, (2007). The Electrodermal System. In J.T. Cacioppo, L.
G.
Tassinary, & G. G. Berntson (Ed.). Handbook of Psychophyisology. (3rd ed., pp.159181). New York: Cambridge University Press.
Denollet, J., Martens, E. J., Nyklicek, I., Conraads, V. M., & de Gelder, B. (2008). Clinical
events in coronary patients who report low distress: Adverse effect of repressive coping.
Health Psychology, 27, 302-308.
Denton, W. H., Burleson, B. R., & Brubaker, P. H. (2009). Avoidance may be bad for the heart:
A comparison of dyadic initiator tendency in cardiac rehabilitation patients and matched
controls. Behavioral Medicine, 35, 135-142.
Diamond, L. M., & Hicks, A. M. (2005). Attachment style, current relationship security, and
negative emotions: The mediating role of physiological regulation. Journal of Social
and Personal Relationships, 22, 499-518.
Diamond, L. M., Hicks, A. M., & Otter-Henderson, K. (2006). Physiological evidence for
repressive coping among avoidantly attached adults. Journal of Social and Personal
Relationships, 23, 205-229.
Dolan, R. T., Arnkoff, D. B., & Glass, C. R. (1993). Client attachment style and the
psychotherapist‘s interpersonal stance. Psychotherapy, 30, 408-412.
Diamond, G. S. (2005). Attachment-based family therapy for depressed and anxious adolescents.
In J. L. Lebow (Ed.), Handbook of clinical family therapy (pp. 17-41). New York: Wiley.
Diamond, G., Siqueland, L., & Diamond, G. S. (2003). Attachment-based family therapy for
depressed adolescents: Programmatic treatment development. Clinical Child and Family
Psychology Review, 6, 107-127.

173

Diamond, G. S., & Stern, R. S. (2003). Attachment-based family therapy for depressed
adolescents: Repairing attachment failures. In S. M. Johnson, & V. E. Whiffen (Eds.),
Attachment processes in couple and family therapy (pp. 191-212). New York: Guilford.
Dozier, M. (1990). Attachment organization and treatment use for adults with serious
psychopathological disorders. Development and Psychopathology, 2, 47–60.
Dozier, M., & Kobak, R. R. (1992). Psychophysiology in attachment interviews: Converging
evidence for deactivating strategies. Child Development, 63, 1473-1480.
Dozier, M., Lindhiem, O., & Ackerman, J. P. (2005). Attachment and Behavioral Catch-Up: An
intervention targeting empirically identified needs of foster infants. In L. J. Berlin, Y.
Ziv, L. Amaya-Jackson, & R. Marvin (Eds.), Enhancing early attachments: Theory,
research, intervention, and policy. (pp. 178-194). New York: Guilford Press.
Dozier, M., Lomax, L., Tyrrell, C. L., & Lee, S. W. (2001). The challenge of treatment for
clients with dismissing states of mind. Attachment & Human Development, 3, 62-76.
Dozier, M., Manni, M., & Lindhiem, O. (2005). Lessons from the longitudinal studies of
attachment. In K. E. Grossmann & K. Grossmann (Eds.), Attachment from infancy to
adulthood: The major longitudinal studies. (pp. 305-320). New York: Guilford Press.
Eagle, M. N. (2006). Attachment, psychotherapy, and assessment: A commentary. Journal of
Counseling and Clinical Psychology, 74, 1086-1097.
Ekman, P. (1993). Facial expression and emotion. American Psychologist, 48, 384-385.
Elkin, I., Shea, M. T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins, J. F., Glass, D. R.,
Pilkonis, P. A., Leber, W. R., Docherty, J. P., Fiester, S. J. & Parloff, M. B. (1989).
National Institute of Mental Health Treatment of Depression Collaborative Research
Program: General effectiveness of treatments. Archives of General Psychiatry, 46, 971982.
Elliott, R. (1986). Interpersonal process recall (IPR) as a psychotherapy process research method. In L.
S. Greenberg & W. M. Pinsof, (Eds.), The psychotherapeutic process: A research handbook (pp.
503-528). New York: Guilford Press.
Epstein, N. B., & Baucom, D. H. (2002). Enhanced cognitive-behavioral therapy for couples.
Washington, DC: American Psychological Association.

174

Estrada, A. U., & Holmes, J. M. (1999). Couples‘ perceptions of effective and ineffective
ingredients of marital therapy. Journal of Sex and Marital Therapy, 25, 151-162.
Feeney, J. A. (1999). Adult attachment, emotional control, and marital satisfaction. Personal
Relationships, 6, 169-185.
Feeney, J. A., Noller, P., & Hanrahan, M. (1994). Assessing adult attachment. In M. B. Sperling
& W. H. Berman (Eds.), Attachment in adults: Clinical and developmental perspectives
(pp. 128-152). New York: Guilford Press.
Ferreira, A. J., & Winter, W. D. (1965). Age and sex differences in the palmar sweat print.
Psychosomatic Medicine, 27, 207-211.
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., Target, M., & Gerber, A.
(1996). The relation of attachment status, psychiatric classification, and response to
psychotherapy. Journal of Consulting and Clinical Psychology, 64, 22-31.
Fonagy, P., Steele, H., & Steele, M. (1991). Maternal representations of attachment during
pregnancy predict the organization of infant-mother attachment at one year of age. Child
Development, 62, 891-905.
Fortuna, K., & Roisman, G. I. (2008). Insecurity, stress, and symptoms of psychopathology:
Contrasting results from self-reports versus interviews of adult attachment. Attachment &
Human Development, 10, 11-28.
Fraley, R. C. (2002). Introduction to the special issue: The psychodynamics of adult
attachments. Bridging the gap between disparate research traditions. Attachment &
Human Development, 4, 131–132.
Fraley, R. C., Garner, J. P., & Shaver, P. R. (2000). Adult attachment and the defensive
regulation of attention and memory: Examining the role of preemptive and postemptive
defensive processes. Journal of Personality and Social Psychology, 79, 816-826.
Fraley, R. C., Waller, N. G., & Brennan, K. A. (2000). An item reesponse theory analysis of selfreport measures of adult attachment. Journal of Personality and Social Psychology, 78,
350-365.
Frank, J. D., & Frank, J. B. (1991). Persuasion and healing: A comparative study of
psychotherapy. (3rd ed.). Baltimore: Johns Hopkins University Press.

175

Friedlander, M. L., Escudero, V., & Heatherinton, L. (2006). Therapeutic alliances in couple and
family therapy: An empirically informed guide to practice. Washington, DC: American
Psychological Association.
Friedlander, M. L., Escudero, V., Horvath, A. O., Heatherington, L., Cabero, A., & Martens, M.
P. (2006). System for Observing Family Therapy Alliances: A tool for research and
practice. Journal of Counseling Psychology, 53, 214-224.
Frijda, N. H. (1986). The emotions. New York: Cambridge University Press.
Gardner, B. C. (2004). A dynamic systems exploration of affective self-organization processes
during marital interaction. Unpublished Doctoral Dissertation, Texas Tech University.
George, C., Kaplan, N., & Main, M. (2002). The Adult Attachment Interview, (3rd ed.).
Unpublished manuscript, University of California, Berkeley, CA.
Gottman, J. M. (1990). How marriages change. In G. R. Patterson (Ed.), Depression and
aggression in family interaction (pp. 110-135). Hillsdale, NJ: Erlbaum.
Gottman, J. M. (1993). The roles of conflict engagement, escalation, and avoidance in marital
interaction: A longitudinal view of five types of couples. Journal of Consulting and
Clinical Psychology, 61, 6-15.
Gottman, J. M. (1998). Psychology and the study of marital processes. Annual Review of
Psychology, 49, 169-197.
Gottman, J. M. (1999). The marriage clinic. New York: W.W. Norton.
Gottman, J. M., & Levenson, R. W. (1985). A valid procedure for obtaining self-report of affect
in marital interaction. Journal of Consulting and Clinical Psychology, 53, 151-160.
Gottman, J. M., & Levenson, R. W. (1992). Marital processes predictive of later dissolution:
Behavior, physiology, and health. Journal of Personality and Social Psychology, 63, 221233.
Gottman, J. M., & Levenson, R. W. (2002). A two-factor model for predicting when a couple
will divorce: Exploratory analyses using 14-year longitudinal data. Family Process, 41,
83-96.

176

Greenberg, L. S., James, P. S., & Conry, R. F. (1988). Perceived change processes in
emotionally focused couples therapy. Journal of Family Psychology, 2, 5-23.
Greenberg, L. S., Pascual-Leone, A. (2006). Emotion in psychotherapy: A practice-friendly
research review. Journal of Clinical Psychology: In Session, 62, 611-630.
Greenberg, L. S., & Paivio, S. C. (2003). Working with emotions in psychotherapy. New York:
Guilford Press.
Greenberg, L., & Pinsof, W. (1986). Process research: Current trends and future perspectives.
In L. Greenberg & W. Pinsof (Eds.), The psychotherapeutic process: A research
handbook (pp. 3-21). New York: Guilford Press.
Grice, P. (1975). Logic and conversation. In P. Cole & J. L. Moran (Eds.), Syntax and semantics:
Vol. 3. Speech acts (pp. 41-58). New York: Academic Press.
Griffin, D. W., & Bartholomew, K. (1994). The metaphysics of measurement: The case of adult
attachment. In K. Bartholomew & D. Perlman (Eds.), Advances in personal
relationships: Vol. 5. Attachment processes in adulthood (pp. 17-52). London: Jessica
Kingsley.
Griffin, W. A. (1993). Transitions from negative affect during marital interaction: Husband and
wife differences. Journal of Family Psychology, 6, 230-244.
Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent
consequences for experience, expression, and physiology. Journal of Personality and
Social Psychology, 74, 224-237.
Gross, J. J., & John, O. P. (1997). Revealing feelings: Facets of emotional expressivity in selfreports, peer ratings, and behavior. Journal of Personality and Social Psychology, 72,
435-448.
Gross, J. J., & John O. P. (2003). Individual differences in two emotion regulation processes:
Implications for affect, relationships, and well-being. Journal of Personality and Social
Psychology, 85, 348-362.
Grossmann, K. E., Grossmann, K., & Waters, E. (2006). Attachment from infancy to adulthood:
The major longitudinal studies. New York: Guilford Press.

177

Guerrero, L. K., & Jones, S. M. (2005). Differences in conversational skills as a function of
attachment style: A follow-up study. Communication Quarterly, 53, 305-321.
Haft, W. L., & Slade, A. (1989). Affect attunement and maternal attachment: A pilot study.
Infant Mental Health Journal, 10, 157-172.
Hamilton, S., Moore, A. M., Crane, D. R., & Payne, S. H. (in press). Psychotherapy dropouts:
Differences by modality, license, and DSM-IV diagnosis. Journal of Marital and Family
Therapy. doi: 10.1111/j.1752-0606.2010.00204.x
Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Journal
of Personality and Social Psychology, 52, 511-524.
Heatherington, L., Friedlander, M. L., & Greenberg, L. (2005). Change process research in
couple and family therapy: Methodological challenges and opportunities. Journal of
Family Psychology, 19, 18-27.
Heavey, C. L., Christensen, A., & Malamuth, N. M. (1995). The longitudinal impact of demand
and withdrawal during marital conflict. Journal of Consulting and Clinical Psychology,
63, 797-801.
Helmeke, K. B., Bischof, G. H., & Sori, C. E. F. (2002). Dropping out of couple therapy: A
qualitative case study. Journal of Couple & Relationship Therapy, 1, 51-73.
Helmeke, K. B. & Sprenkle, D. H. (2000). Clients‘ perceptions of pivotal moments in couples
therapy: A qualitative study of change in therapy. Journal of Marital and Family
Therapy, 26, 469-483.
Hesse, E. (2008). The Adult Attachment Interview: Protocol, method of analysis, and empirical
studies. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research,
and clinical applications. (2nd ed., pp. 552-598). New York: Guilford Press.
Hesse, E., & Main, M. (2006). Frightened, threatening, and dissociative parental behavior in
low-risk samples: Description, discussion, and interpretations. Development and
Psychopathology, 18, 309-343.
Hollis, F., (1968). Continuance and discontinuance in marital counseling and some observations
on joint interviews. Social Casework, 49, 167-174.

178

Hubble, M. A., Duncan, B. L., & Miller, S. (1999). The heart and soul of change: What works in
therapy. Washington, DC: The American Psychological Association.
Izard, C. E. (1991). The psychology of emotions. New York: Plenum Publishing.
Jacobson, N. S., & Christensen, A. (1996). Acceptance and change in couple therapy: A
therapist’s guide for transforming relationships. New York: Norton.
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, 61, 85-93.
Jacobson, N. S., & Margolin, G. (1979). Marital therapy: Strategies based on social learning
and behavior exchange principles. New York: Brunner/Mazel.
John, O. P., & Gross, J. J. (2007). Healthy and unhealthy emotion regulation: Personality
processes, individual differences, and life span development. Journal of Personality, 72,
1301-1334.
Johnson, S. M. (2004). The practice of Emotionally Focused Couple Therapy, (2nd ed.). New
York: Brunner-Routledge.
Johnson, S. M. (2003a). Attachment theory: A guide for couple therapy. In S. M. Johnson & V.
E. Whiffen (Eds.), Attachment processes in couple and family therapy (pp. 103-121).
New York: Guilford Press.
Johnson, S. M. (2003b). Let us keep emotion at the forefront: A reply to Roberts and Koval.
Journal of Couple & Relationship Therapy, 2, 15-20.
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., & Greenberg, L. S. (1994). Emotion in intimate relationships: Theory and
implications for therapy. In S. M. Johnson & L. S. Greenberg (Eds.), The heart of the
matter: Perspectives on emotion in marital therapy. (pp. 3-26). New York:
Brunner/Mazel.
Johnson, S. M., & Talitman, E. (1997). Predictors of success in emotionally focused marital
therapy. Journal of Marital and Family Therapy, 23, 135–152.

179

K Johnson, S. M., & Greenberg, L. S. (1995). The emotionally focused approach to problems in
adult attachment. In N. S. Jacobsen & A. S. Gutman, (Eds.), Clinical handbook of couple
therapy (pp. 121-141). New York: Guilford.
Johnson, S., & Sims, A. (2000). Attachment theory: A map for couples therapy. In T. M. Levy
(Ed.) Handbook of attachment interventions (pp. 167-191). San Diego, CA: Academic
Press.
Johnson, S., & Whiffen, V. (1999). Made to measure: Adapting emotionally focused couples
therapy to couples‘ attachment styles. Clinical Psychology: Science and Practice, 6, 366381.
Kagan, N. I., & Kagan, H. (1990). IPR—A validated model for the 1990s and beyond.
Counseling Psychologist, 18, 436–440.
Kagan, N., Krathwohl, D. R., & Miller, R. (1963). Stimulated recall in therapy using video
tape—A case study. Journal of Counseling Psychology, 10, 237–243.
Kanninen, K., Salo, J., & Punamaki, R. L. (2000). Attachment patterns and working alliance in
trauma therapy for victims of political violence. Psychotherapy Research, 10, 435-449.
Katz, L. F., & Gottman, J. M. (1995). Vagal tone protects children from marital conflict.
Development and Psychopathology, 7, 83–92.
Keith, T. Z. (2006). Multiple regression and beyond. Boston: Pearson Education.
Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York: Guilford
Press.
Kerig, P. K. (1996). Assessing the links between interparental conflict and child adjustment: The
Conflicts and Problem-Solving Scales. Journal of Family Psychology, 10, 454-573.
Kim, Y. (2006). Gender, attachment, and relationship duration on cardiovascular reactivity to
stress in a laboratory study of dating couples. Personal Relationships, 13, 103-114.
Klein, M. H., Mathieu-Coughlan, P. & Kiesler, D. (1986). The experiencing scales. In L. S.
Greenberg & W. Pinsof (Eds.), The psychotherapeutic process: A research handbook
(pp. 21-71). New York: Guilford Press.

180

Kobak, R. R. (1989). The attachment interview q-set. Unpublished document, University of
Virginia.
Kobak, R. R., Cole, H. E., Ferenz-Gillies, R., Fleming, W. S., & Gamble, W. (1993). Attachment
and emotion regulation during mother-teen problem solving: A control theory analysis.
Child Development, 64, 231-245.
Kobak, R. R., & Madsen, S. (2008). Disruptions in attachment bonds: Implications for theory,
research, and clinical intervention. In J. Cassidy & P. R. Shaver (Eds.), Handbook of
attachment: Theory, research, and clinical applications. (2nd ed., pp. 23-47). New York:
Guilford Press.
Korfmacher, J., Adam, E., Ogawa, J., & Egeland, B. (1997). Adult attachment: Implications for
the therapeutic process in a home visitation intervention. Applied Developmental Science,
1, 43−52.
Kring, A. M., & Gordon, A. H. (1998). Sex differences in emotion: Expression, experience, and
physiology. Journal of Personality and Social Psychology, 74, 686-703.
Lakey, B., & Ondersma, S. J. (2008). A new approach for detecting client-treatment matching in
psychological therapy. Journal of Social and Clinical Psychology, 27, 56-69.
Lambert, M. J. (1992) Psychotherapy outcome research: Implications for integrative and
eclectic therapists. In J. C. Norcross & M. R. Goldfried (Eds.), Handbook of
psychotherapy integration (pp. 94–129). New York: Basic Books.
Lambert, M. J., Burlingame, G. M., Umphress, V., Hansen, N. B., Vermeersch, D. A., Clouse, G.
C., & Yanchar, S. C. (1996). The reliability and validity of the Outcome Questionnaire.
Clinical Psychology and Psychotherapy, 3, 249-258.
Lambert, M. J., & Ogles, B. M. (2003). The efficacy and effectiveness of psychotherapy. In M. J.
Lambert (Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change
(5th ed., pp. 139-193). New York: John Wiley & Sons.
Levenson, R. W., & Gottman, J. M. (1983). Marital interaction: Physiological linkage and
affective exchange. Journal of Personality and Social Psychology, 45, 587-597.
Levenson, R. W., & Gottman, J. M. (1985). Physiological and affective predictors of change in
relationship satisfaction. Journal of Personality and Social Psychology, 49, 85-94.

181

Linehan, M.M. (1993). Cognitive–behavioral treatment of borderline personality disorder. New
York: Guilford Press.
Lopez, F. G., Melendez, M. C., Sauer, E. M., Berger, E., & Wyssmann, J. (1998). Internal
working models, self-reported problems, and help-seeking attitudes among college
students. Journal of Counseling Psychology, 49, 79-83.
Main, M. (1996). Recording and transcribing the Adult Attachment Interview. Unpublished
manuscript, University of California, Berkeley, CA.
Main, M., Goldwyn, R., & Hesse, E. (2002). Adult attachment scoring and classification systems
(Version 7.1) Unpublished manuscript, University of California, Berkeley, CA.
Main, M. & Hesse, E. (1990). Parents‘ unresolved traumatic experiences are related to infant
disorganized attachment status: Is frightened/frightening parental behavior the linking
mechanism? In M. Greenberg, D. Cicchetti, & M. Cummings (Eds.), Attachment in the
preschool years, (pp. 161-182). Chicago: University of Chicago Press.
Makinen, J. A., & Johnson, S. M. (2006). Resolving attachment injuries in couples using
Emotionally Focused Therapy: Steps toward forgiveness and reconciliation. Journal of
Consulting and Clinical Psychology, 74, 1055-1064.
Mallinckrodt, B., Coble, H. M., & Gantt, D. L. (1995). Working alliance, attachment memories,
and social competencies of women in brief therapy. Journal of Counseling Psychology,
42, 79-84.
Masi, M. V., Miller, R. B., Olson, M. M. (2003). Differences in dropout rates among individual,
couple, and family therapy clients. Contemporary Family Therapy, 25, 63-75.
Mauss, I. B., & Gross, J. J. (2004). Emotion suppression and cardiovascular disease: Is hiding
feelings bad for your heart? In L. R. Temoshok, A. Vingerhoets, & I. Nyklicek (Eds.),
The expression of emotion and health: Advances in theory, assessment, and clinical
applications (pp. 60-80). New York: Brunner-Routledge.
McBride, C., Atkinson, L., Quilty, L. C., & Bagby, R. M. (2006). Attachment as moderator of
treatment outcome in major depression: A randomized control trial of interpersonal
psychotherapy cognitive behavior therapy. Journal of Consulting and Clinical
Psychology, 74, 1041-1054.

182

Meyer, B., & Pilkonis, P. A. (2001). Attachment style. Psychotherapy, 38, 466-472.
Meyer, B., Pilkonis, P. A., Proietti, J. M., Heape, C. L., & Egan, M. (2001). Adult attachment
styles, personality disorders, and response to treatment. Journal of Personality Disorders,
15, 371-389.
Mikulincer, M. (1998). Adult attachment style and individual differences in functional versus
dysfunctional experiences of anger. Journal of Personality and Social Psychology, 74,
513-524.
Mikulincer, M., & Shaver, P. R. (2007). Attachment in adulthood: Structure, dynamics, and
change. New York: Guilford Press.
Mohr, J. J., Gelso, C. J., & Hill, C. E. (2005). Client and counselor trainee attachment as
predictors of session evaluation and countertransference behavior in first counseling
sessions. Journal of Counseling Psychology, 52, 298-309.
Mueller, R. M., Lambert, M. J., & Burlingame, G. M. (1998). Construct validity of the Outcome
Questionnaire: A confirmatory factor analysis. Journal of Personality Assessment, 70,
248-262.
Myers, L. B. (1998). Repressive coping, trait anxiety and reported avoidance of negative
thoughts. Personality and Individual Differences, 24, 299-303.
Myers, L. B. (2000). Identifying repressors: A methodological issue for health psychology.
Psychology and Health, 5, 205-214.
Nathan, P., & Gorman, J. M. (2002). A guide to treatments that work. New York: Oxford
University Press.
Nichols, M. P, & Fellenberg, S. (2000). The effective use of enactments in family therapy: A
discovery-oriented process study. Journal of Marital and Family Therapy, 26, 143-152.
Orlinsky, D. E., Ronnestad, M. H., & Willutzki, U. (2003). Fifty years of psychotherapy processoutcome research: Continuity and change. In M. J. Lambert (Ed.), Bergin and Garfield’s
handbook of psychotherapy and behavior change (5th ed., pp. 307-389). New York: John
Wiley & Sons.

183

Papakonstantinou, E., Hargrove, J. L., Huang, C. L., Crawley, C. C., & Canolty, N. L. (2002).
Assessment of perceptions of nutrition knowledge and disease using a group interactive
system: The perception analyzer®. Journal of the American Dietetic Association, 102,
1663-1668.
Parish, M., & Eagle, M. N. (2003). Attachment to the therapist. Psychoanalytic Psychology, 20,
271-286.
Parkes, C. M. (1972). Bereavement: Studies of grief in adult life. New York/Madison, CT:
International Universities Press.
Parkes, C. M., Stevenson-Hinde, J., & Marris, P. (1991). Attachment across the life cycle.
London: Routledge.
Paul, G. (1967). Strategy of outcome research in psychotherapy. Journal of Consulting and
Clinical Psychology, 31,109-118.
Phillips, L. E. (1985). Psychotherapy revised: New frontiers in research and practice. Mahwah,
NJ: Lawrence Erlbaum Associates.
Pistole, M. C., & Arricale, F. (2003). Understanding attachment: Beliefs about conflict. Journal
of Counseling & Development, 81, 318-328.
Porges, S. W., Doussard-Roosevelt, J. A., & Maiti, A. K. (1994). Vagal tone and the
physiological regulation of emotion. In N. Fox (Ed.), The development of emotion
regulation:Biological and behavioral considerations. Monographs of the Society for
Research in Child Development, 59, 167–186.
Powers, S. I., Pietromonaco, P. R., Gunlicks, M., & Sayer, A. (2006). Dating couples‘
attachment styles and patterns of cortisol reactivity and recovery in response to a
relationship conflict. Journal of Personality and Social Psychology, 90, 613-628.
Project MATCH Research Group. (1997a). Matching alcoholism treatments to client
heterogeneity: Project MATCH posttreatment drinking outcomes. Journal of Studies on
Alcohol, 58, 7–29.
Project MATCH Research Group. (1997b). Project MATCH secondary a priori hypotheses.
Addiction, 92, 1671–1698.

184

Project MATCH Research Group. (1998). Matching patients with alcohol disorders to
treatments: Clinical implications from Project MATCH. Journal of Mental Health, 7,
589-602.
Quinn, W. H., Dotson, D., & Jordan, K. (1997). Dimensions of therapeutic alliance and their
associations with outcome in family therapy. Psychotherapy Research, 7, 429-438.
Raytek, H. S., McCrady, B. S., Epstein, E. E., & Hirsch, L. S. (1999). Therapeutic alliance and
the retention of couples in conjoint alcoholism treatment. Addictive Behaviors, 24, 317330.
Reis, S., & Grenyer, B. F. S. (2004). Fear of intimacy in women: Relationship between
attachment styles and depressive symptoms. Psychopathology, 37, 299-303.
Repetti, R. L.,Taylor, S. E., & Seeman,T. E. (2002). Risky families: Family social environments
and the mental and physical health of offspring. Psychological Bulletin, 128, 330–366.
Rholes, W. S., Simpson, J. A., & Orina, M. M. (1999). Attachment and anger in an anxietyprovoking situation. Journal of Personality and Social Psychology, 76, 940-957.
Richards, J. M. (2004). The cognitive consequences of concealing feelings. Current Directions
in Psychological Science, 13, 131-134.
Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The cognitive consequences of
emotion suppression. Personality and Social Psychology Bulletin, 25, 1033-1044.
Roisman, G. I. (2007). The psychophysiology of adult attachment relationships: Autonomic
reactivity in marital and premarital interactions. Developmental Psychology, 43, 39-53.
Roisman, G. I., Collins, W. A., Sroufe, L. A., & Egeland, B. (2005). Predictors of young adults‘
representation of and behavior in their current romantic relationship: Prospective tests of
the prototype hypothesis. Attachment & Human Development, 7, 105-121.
Roisman, G. I., Fraley, R. C., & Belsky, J. (2007). A taxometric study of the Adult Attachment
Interview. Developmental Psychology, 43, 675-686.

185

Roisman, G. I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., & Clarke, A. (2007). The
Adult Attachment Interview and self-reports of attachment style: An empirical
rapprochement. Journal of Personality and Social Psychology, 92, 678-697.
Roisman, G. I., Tsai, J. L., & Chiang, K. H. S. (2004). The emotional integration of childhood
experience: Physiological, facial expressive, and self-reported emotional response
during the Adult Attachment Interview. Developmental Psychology, 40, 776-789.
Roth, A., & Fonagy, P. (2005). What works for whom? A critical review of psychotherapy
research. (2nd ed.). New York: Guilford Press.
Rottenberg, J., & Gross, J. J. (2007). Emotion and emotion regulation: A map for psychotherapy
researchers. Clinical Psychology: Science and Practice, 14, 323-328.
Rotter, N. G., & Rotter, G. S. (1988). Sex differences in the encoding and decoding of negative
facial emotions. Journal of Nonverbal Behavior, 12, 139-148.
Rychtarik, R. G., Connors, G. J., Whitney, R. B., McGillicuddy, N. B., Fitterling, J. M., & Wirtz,
P. W. (2000). Treatment settings for persons with alcoholism: Evidence for matching
clients to inpatient versus outpatient care. Journal of Consulting and Clinical Psychology,
68, 277-289.
Satterfield, W. A., & Lyddon, W. J. (1998). Client attachment and the working alliance.
Counseling Psychology Quarterly, 11, 407-415.
Seedall, R. B., (2008, November). The Subjective Emotional Experience Questionnaire (SEEQ):
A psychometric evaluation. Poster presented at the annual conference of the National
Council on Family Relations, Little Rock, AR.
Seedall, R. B., & Butler, M. H. (2006). The effect of proxy voice intervention on couple
softening in the context of enactments. Journal of Marital and Family Therapy, 32, 421437.
Sexton, T.L., Alexander, J. F., & Mease, A. L. (2003). Levels of evidence for the models and
mechanisms of therapeutic change in couple and family therapy. In M. Lambert (Ed).
Bergin and Garfield’s handbook of psychotherapy and behavior change. (5th ed., pp.
590-646). New York: John Wiley & Sons.

186

Shadish, W. R., & Baldwin, S. A. (2003). Meta-analysis of MFT interventions. Journal of
Marital and Family Therapy, 29, 547-570.
Shadish, W. R., Ragsdale, K., Glaser, R. R., & Montgomery, L. M. (1995). The efficacy and
effectiveness of marital and family therapy: A perspective from meta-analysis. Journal of
Marital and Family Therapy, 21, 345-360.
Shaver, P. R., Belsky, J., & Brennan, K. A. (2000). The adult attachment interview and selfreports of romantic attachment: Associations across domains and methods. Personal
Relationships, 7, 25-43.
Simpson, J. A. (1990). Influence of attachment styles on romantic relationships. Journal of
Personality and Social Psychology, 59, 971-980.
Simpson, J. A., Rholes, W. S., Orina, M. M., & Grich, J. (2002). Working models of
attachment, support giving, and support seeking in a stressful situation. Personality and
Social Psychology Bulletin, 28, 598-608.
Simpson, J. A., Rholes, W. S., & Phillips, D. (1996). Conflict in close relationships: An
attachment perspective. Journal of Personality and Social Psychology, 71, 899-914.
Simpson, J. A., Winterheld, H. A., Rholes, W. S., & Orina, M. M. (2007). Working models of
attachment and reactions to different forms of caregiving from romantic partners.
Journal of Personality and Social Psychology, 93, 466-477.
Smith, M. L., Glass, G. V., & Miller, T. I. (1980). The benefits of psychotherapy. Baltimore:
Johns Hopkins University Press.
Snyder, D. K. (2002). Integrating insight-oriented techniques into couple therapy. In J. H.
Harvey & A. E. Wenzel (Eds.), A clinicians guide to maintaining and enhancing close
relationships (pp. 259-275). Mahwah, NJ: Erlbaum.
Snyder, D. K. (1999). Affective reconstruction in the context of a pluralistic approach to couple
therapy. Clinical Psychology: Science and Practice, 6, 348-365.
Snyder, D. K., & Mitchell, A. E. (2008). Affective-reconstructive couple therapy: A pluralistic,
developmental approach. In A. S. Gurman (Ed.), Clinical handbook of couple therapy,
Fourth Edition (pp. 353-382). New York: Guilford Press.

187

Snyder, D. K., & Whisman, M. A. (2003). Treating difficult couples: Helping clients with
coexisting mental and relationship disorders. New York: Guilford Press.
Sprenkle, D. H. (2002). Effectiveness research in marriage and family therapy. Alexandria, VA:
American Association for Marriage and Family Therapy.
Sprenkle, D. H., & Blow, A. J. (2004). Common factors and our sacred models. Journal of
Marital and Family Therapy, 30, 113-129.
Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The development of the
person: The Minnesota study of risk and adaptation from birth to adulthood. New York:
Guilford Press.
Sroufe, L. A., & Waters, E. (1977). Heart rate as a convergent measure in clinical and
developmental research. Merrill-Palmer Quarterly, 23, 3-25.
Steele, H., & Steele, M. (2008). Clinical applications of the Adult Attachment Interview. New
York: Guilford Press.
Tryon, G. S., & Kane, A. S. (1993). Relationship of working alliance to mutual and unilateral
termination. Journal of Counseling Psychology, 40, 33-36.
Tyrrell, C. L., Dozier, M., Teague, G. B., & Fallot, R. D. (1999). Effective treatment
relationships for persons with serious psychiatric disorders: The importance of
attachment states of mind. Journal of Consulting and Clinical Psychology, 67, 725-733.
UKATT research team (2008). UK alcohol treatment trial: Client-treatment matching effects.
Addiction, 103, 228-238.
van IJzendoorn, M. H. (1995). Adult attachment representations, parental responsiveness, and
infant attachment: A meta-analysis on the predictive validity of the Adult Attachment
Interview. Psychological Bulletin, 117, 387-403.
Venables, P. H., & Mitchell, D. A. (1996). The effects of age, sex, and time of testing on skin
conductance activity. Biological Psychology, 43, 87-101.
Vetere, A., & Myers, L. B. (2002). Repressive coping style and adult romantic attachment style:
Is there a relationship? Personality and Individual Differences, 32, 799-807.

188

Waldinger, R. J., Moore, C., & Schulz, M. S. (2003, April). Child abuse and adult attachment:
Links with emotion during video recall of couple interactions. Workshop presented at the
annual conference of the Society for Research in Child Development, Tampa, FL.
Wampler, K. S., Shi, L., Nelson, B. S., & Kimball, T. G. (2003). The adult attachment interview
and observed couple interaction: Implications for an intergenerational perspective on
couple therapy. Family Process, 42, 497-515.
Wampold, B. E. (2001). The great psychotherapy debate: Models, methods, and findings.
Mahwah, NJ: Erlbaum.
Ward, M. J., & Carlson, E. A. (1995). Associations among adult attachment representations,
maternal sensitivity, and infant-mother attachment in a sample of adolescent mothers.
Child Development, 66, 69-79.
Weinberger, D. A. (1990). The construct validity of the repressive coping style. In J. L. Singer
(Ed.). Repression and dissociation (pp. 337-386). Chicago: University of Chicago Press.
Weiss, R. (1975). Marital Separation: Coping with the end of a marriage and the transition to
being single again. New York: Basic Books.
Weiss, R. (1991). The attachment bond in childhood and adulthood. In C. M. Parkes, J.
Stevenson-Hinde, & P. Marris (Eds.), Attachment across the life cycle (pp. 66-76).
London: Routledge.
Weiss, R.L., & Heyman, R.E. (1997). A clinical-research overview of couple interactions. In
W.K. Halford & H. Markman (Eds.), The clinical handbook of marriage and couples
interventions (pp. 13-41). Brisbane, Australia: Wiley.
Werzbicki, M., & Pekarik, G. (1993). A meta-analysis of psychotherapy dropout. Professional
Psychology: Research and Practice, 29, 190-195.
West, T. V., Popp, D., & Kenny, D. A. (2008). A guide for the estimation of gender and sexual
orientation effects in dyadic data: An actor-partner interdependence model approach.
Personality and Social Psychology Bulletin, 34, 321-336.
Williams, S., Ketring, S., & Salts, C. (2005). Premature termination as a function of intake data
based on ethnicity, gender, socioeconomic status, and income. Contemporary Family
Therapy, 27, 211-229.

189

Wiser, S. & Arnow, B. (2001). Emotional experiencing: To facilitate or regulate? Journal of
Clinical Psychology, 57, 157-168.
Zuckerman, M., Hall, J.A., DeFrank, R.S. and Rosenthal, R. (1976). Encoding and decoding of
spontaneous and posed facial expressions. Journal of Personality and Social Psychology,
34, 966-977.

190