ABSTRACT
Title of Document: THE MODERATING EFFECT OF AN
INDIVIDUAL?S ROMANTIC ATTACHMENT
STYLE ON INTENT TO LEAVE A
HETEROSEXUAL CLINCAL COUPLE
RELATIONSHIP IN WHICH HE OR SHE IS
THE VICTIM OF EMOTIONAL ABUSE
Mark J. Treimel, Master of Science, 2006
Directed By: Professor Sandra Hofferth
Family Studies Department
This study examined the impact of emotional abuse on an individual?s steps
taken toward leaving a relationship and how individuals with different styles of
attachment to a romantic partner differed with regard to taking such steps. Analyses
of participant responses on the Multi-dimensional Emotional Abuse Scale (MDEAS),
the Relationship Questionnaire (RQ), and the Marital Status Inventory-Revised (MSI-
R) were conducted using analysis of variance. It was determined that abuse and
being female were significantly associated with taking greater steps toward
relationship dissolution. Finally, the interaction between abuse and attachment in
steps taken toward leaving was significant among women, but not among men.
Although attachment was associated with steps taken toward leaving an abusive
relationship was significant for women, differences between attachment styles were
not consistent with the hypotheses.
THE MODERATING EFFECT OF AN INDIVIDUAL?S ROMANTIC
ATTACHMENT STYLE ON INTENT TO LEAVE A RELATIONSHIP IN WHICH
HE OR SHE IS THE VICTIM OF EMOTIONAL ABUSE IN HETEROSEXUAL
CLINICAL COUPLES
By
Mark J. Treimel
Thesis submitted to the Faculty of the Graduate School of the
University of Maryland, College Park, in partial fulfillment
of the requirements for the degree of
Master of Science
2006
Advisory Committee:
Professor Sandra Hofferth, Chair
Professor Norman Epstein
Instructor Carol Werlinich
? Copyright by
Mark J. Treimel
2006
ii
Acknowledgements
I would like to thank the following people, for without them, I could have never
completed this project:
Thank you to Dr. Sandra Hofferth, my advisor, for all of the time, effort, and words
of encouragement that you provided me with. I can not thank you enough for your
patience as I worked on finishing this paper while living in Mexico.
Thank you to Dr. Norman Epstein and Dr. Carol Werlinich, my committee members,
for going far and above the call of duty as you worked to help me complete this
project before leaving the country, as well as for providing me with such excellent
professional examples to emulate during my time in the Marriage and Family
Therapy program.
Thank you to my classmates from the class of 2004 for all of the love and support
during our time together in the program. I also need to especially thank Rebecca Frey
and Rachel Alexander (twice!) for helping me with getting figures to fit within the
margins.
Thank you to two of my ?Seniors? from the class of 2003, Alexandra Goerl for
helping me with a variety of formatting issues and Lindsey Hoskins for sharing her
expertise in running SPSS.
Thank you to Mrs. Sherry Starr for countless words of encouragement and
understanding.
Thank you to my parents, Joe and Karen, as well as my sister, Ellen, for all of the
love, encouragement, and support. Most of all, thank you for always believing in me,
no matter what.
Thank you to my fianc?e, Veronica, there are no words to describe how your patience
and love has gotten me through the most difficult of times.
iii
Table of Contents
Acknowledgements........................................................................................................ ii
Table of Contents .........................................................................................................iii
List of Tables................................................................................................................ vi
List of Figures............................................................................................................. vii
Chapter 1: Introduction ............................................................................................ 1
Statement of the Problem...................................................................................... 1
Purpose.................................................................................................................. 2
Review of Literature ............................................................................................. 3
Relationship Dissolution................................................................................... 3
Gender and Relationship Dissolution............................................................... 6
What is Emotional Abuse?................................................................................ 7
Emotional Abuse and Potential for Relationship Dissolution .......................... 9
Symbolic Interaction Theory........................................................................... 12
Attachment Theory.......................................................................................... 14
Different Attachment Styles and Relationship Dissolution Potential ............. 19
Implications of Attachment Theory for Whether Individuals.......................... 20
Hypotheses.......................................................................................................... 23
Hypothesis 1: Positive Association between Psychological Abuse and Steps
Taken Toward Leaving a Relationship ........................................................... 24
Hypothesis 2: At Each Level of Emotional Abuse Women will Take Greater
Steps toward Relationship Dissolution than Men........................................... 24
Hypothesis 3: Different Individual Attachment Styles Will be Linked
Differently to Steps Taken to Leave a Relationship........................................ 25
Hypothesis 4: The Relationship Between Steps Taken Toward Leaving a
Relationship and Abuse will Differ Depending on an Individual?s Attachment
Style................................................................................................................. 25
Hypothesis 4.1: Individuals with a Secure Style of Attachment Demonstrate
a Positive Association Between Experienced Emotional Abuse and Steps
Taken Toward Leaving the Relationship .................................................... 25
Hypothesis 4.2: Individuals with a Dismissing Style of Attachment
Demonstrate a Positive Association Between Experienced Emotional Abuse
and Steps Taken Toward Leaving the Relationship.................................... 26
Hypothesis 4.3: Individuals with a Preoccupied Style of Attachment would
Not Demonstrate any Correlation Between Experienced Emotional Abuse
and Steps Taken Toward Leaving the Relationship.................................... 26
Hypothesis 4.4: Individuals with a Fearful Style of Attachment would
Demonstrate a Strong Positive Association Between Experienced
Emotional Abuse and Steps Taken Toward Leaving the Relationship. ...... 27
Definition of Variables ....................................................................................... 28
Independent Variables .................................................................................... 28
Emotional Abuse ......................................................................................... 28
Romantic Attachment Style ......................................................................... 28
Secure:..................................................................................................... 28
iv
Dismissing:.............................................................................................. 29
Preoccupied: ........................................................................................... 29
Fearful: ................................................................................................... 29
Gender......................................................................................................... 29
Dependent Variable ........................................................................................ 30
Steps Taken Toward Leaving a Relationship.............................................. 30
Control Variables............................................................................................ 30
Socioeconomic Status.................................................................................. 30
Length of Relationship ................................................................................ 30
Chapter 2: Methods ................................................................................................ 31
Participants.......................................................................................................... 31
Procedures........................................................................................................... 31
Measures ............................................................................................................. 32
Emotional Abuse ............................................................................................. 32
Attachment Style.............................................................................................. 34
Steps Taken Toward Relationship Dissolution ............................................... 35
Chapter 3: Results................................................................................................... 37
Study Participants ............................................................................................... 37
Univariate and Bivariate Analyses...................................................................... 38
Hypothesis Testing.............................................................................................. 39
Hypothesis 1: Positive Correlation between Psychological Abuse and Steps
Taken Toward Leaving a Relationship ........................................................... 39
Hypothesis 2: In the Presence of Abuse, women will Take Greater Steps
Toward Leaving an Intimate Relationship than men...................................... 41
Hypothesis 3: Depending on Attachment Style, Individuals Will Take Different
Degrees of Steps Toward Leaving a Relationship .......................................... 43
Hypothesis 4: The Relationship Between Steps Taken Toward Leaving a
Relationship and Abuse will Differ Depending on an Individual?s Attachment
Style................................................................................................................. 52
Chapter 4: Discussion............................................................................................. 61
There Will be a Positive Association Between Experienced Emotional Abuse
and Steps Taken to Leave a Relationship ........................................................... 62
In the Presence of Abuse, Women will Take Greater Steps Toward Leaving a
Relationship than Men ........................................................................................ 63
Depending on Attachment Style, Individuals will Take Different Steps Toward
Leaving a Relationship ....................................................................................... 63
Individuals with Different Attachment Styles will Take Different Steps Toward
Leaving a Relationship in the Presence of Emotional Abuse ............................. 65
Limitations of Current Study .............................................................................. 69
Recommendations for Future Research.............................................................. 70
Clinical Implications........................................................................................... 72
Conclusion .......................................................................................................... 73
Appendices: Measures ................................................................................................ 75
Appendix A......................................................................................................... 75
Appendix B ......................................................................................................... 77
Appendix C ......................................................................................................... 78
v
References................................................................................................................... 79
vi
LIST OF TABLES
Table 1: Demographics by Gender ............................................................................. 37
Table 2: Correlation Matrix ........................................................................................ 38
Table 3: Mean MSI-R Scores by Abuse Level - Women and Men............................ 40
Table 4: Bonferroni Comparison - Total for Women and Men.................................. 40
Table 5: Tests of Between-Subjects Effects ............................................................... 42
Table 6: Mean MSI-R Scores by Attachment Style.................................................... 43
Table 7: Bonferroni Comparison - Attachment Style................................................. 46
Table 8: Mean MSI-R Scores by Attachment Style by Gender.................................. 47
Table 9: Tests of Between-Subjects Effects ............................................................... 47
Table 10: Bonferroni Comparison (Women).............................................................. 51
Table 11: Bonferroni Comparison (Men) ................................................................... 51
Table 12: Tests of Between-Subjects Effects ............................................................. 53
Table 13: Tests of Between-Subjects Effects (Women) ............................................. 54
Table 14: Tests of Between-Subjects Effects (Men) .................................................. 54
Table 15: Women's Mean MSI-R Scores at Different Levels of Abuse by Attachment
Style ............................................................................................................................ 56
Table 16: Bonferroni Comparison (Women).............................................................. 58
Table 17: Bonferroni Comparison .............................................................................. 59
Table 18: Hypotheses and Results .............................................................................. 62
vii
LIST OF FIGURES
Figure 1. Model of Adult Attachment, (Bartholomew & Horowitz, 1991)................ 18
Figure 2. Expected results among individuals? intent to leave in the presence of
emotional abuse .......................................................................................................... 24
Figure 3: The Moderating Effect of Romantic Attachment Styles on Steps Taken to
Leave an Abusive Relationship .................................................................................. 28
Figure 4: Plot: Difference between Men and Women?s Mean MSI-R Scores at Each
Level of Abuse............................................................................................................ 41
Figure 5: Plot: Mean MSI-R Scores by Attachment Style.......................................... 45
Figure 6: Plot: Women?s Mean MSI-R Scores by Attachment Style ......................... 48
Figure 7: Plot: Men?s Mean MSI-R Scores by Attachment Style .............................. 49
Figure 8: Plot: Women?s and Men?s Mean MSI-R Scores by Attachment Style ....... 50
Figure 9: Plot: Women?s Mean MSI-R Scores by Attachment Style ......................... 57
1
CHAPTER 1: INTRODUCTION
Statement of the Problem
Why do some abused individuals remain in an abusive relationship and others
leave? Answering this question would not only be of interest to researchers but
would also be helpful for clinical professionals. In particular, what are the
characteristics of individuals or circumstances under which a relationship will survive
or a relationship will fail when one partner becomes abusive to the other?
Understanding these circumstances will help clinicians better counsel their clients in
such relationships. In addition, in many abusive relationships, emotional abuse is a
precursor to physical abuse (Murphy & O?Leary, 1989; Russell & Hulson, 1992).
The study of emotional abuse as an area of research may further the treatment and
prevention of physical abuse (O?Leary, 1999).
Recent research suggests that an individual?s personal or temperamental
characteristics, romantic attachment styles in particular, impact couple relationships
and may affect whether that individual remains in or leaves an abusive relationship.
It has been established that within intimate relationships, attachment styles are linked
to the perpetration of physical and emotional abuse (Dutton, Saunders, Starzomski, &
Bartholomew, 1994), but little research has been conducted in order to examine how
attachment styles affect the experience of abuse. Adult attachment styles play an
important role with regard to how both positive and negative behaviors are perceived
and reacted to by partners (Gallo & Smith, 2001). As defined by Bartholomew and
Horowitz (1991), there are four types of adult attachment: secure, dismissing,
preoccupied, and fearful. Research suggests that there is a strong link between
2
attachment and the types of attributions individuals make about their partners
(Collins, Ford, Guichard, & Allard, 2006).
Purpose
This study used data collected from couples who have participated in the
Couples Abuse Prevention Program, a study presently being conducted at the
University of Maryland?s Family Service Center. This current study examined
whether individuals? intentions to leave their abusive relationships are moderated by
those individuals? attachment styles. In order to conduct this study, data that were
collected from both members of a couple during an initial assessment at the Family
Service Center were utilized. This assessment is conducted before the couple is
invited to participate in the Couples Abuse Prevention Program as well as prior to the
beginning of therapy. These data were derived from the Multi Dimensional
Emotional Abuse Scale (MDEAS), the Relationship Questionnaire (RQ), and the
Marital Status Inventory Revised (MSI-R) which are filled out separately and
individually by both members of the couples seen at the Family Service Center.
Specifically, through the use of secondary data analysis, this study examined whether
the steps taken toward leaving a relationship in the presence of emotional abuse are
affected by the attachment style of the individual who is abused.
This study fits within the current literature as well as addresses gaps in the
research on couples, attachment, and abusive relationships. First, it furthers the work
of Henderson, Bartholomew, and Dutton (1997) who established that attachment
styles impact how individuals determine whether or not to leave physically abusive
relationships. This study examined the moderating effect of attachment styles on
3
determining to leave emotionally abusive relationships. Second, a significant amount
of the current research on couple abuse focuses on physical abuse. This current study
extended the current research on emotional abuse, which has been shown to have a
powerful impact on those receiving it (Herbert, Silver, & Ellard, 1991). Third, this
study will also address the dearth of clinical literature identified by Wampler, Riggs,
and Kimball (2004) on how attachment plays a role within relationships.
Additionally, this study examined both members of couples, thus addressing a
neglected area, namely the impact of abuse on men (Hines & Malley-Morrison,
2001). The results of this study not only contribute to research, but also ideally are
beneficial to clinicians. The members of the study?s sample are all individuals who
have at least completed an assessment in a clinical setting. Determining how
attachment styles moderate emotional abuse?s impact on potential relationship
dissolution, clinicians could better evaluate the likelihood that their clients may break
up and thus better determine an appropriate treatment plan. With the proper research
on which to base their claims, clinicians could better educate their clients on how
emotionally abusive behaviors may impact their relationship. Primarily they would
be able to explain to clients who are emotionally abusive toward their partners the
likelihood that their relationships will fail if they choose to continue to engage in
these behaviors.
Review of Literature
Relationship Dissolution
More than half of all first marriages end in either separation or divorce
(Bramlett & Mosher, 2002). Numerous studies have been conducted in order to
4
examine what factors contribute to relationship dissolution, primarily in the realm of
divorce. Research has been conducted to examine whether demographics play a role
in explaining divorce and a number of conclusions have been realized (Booth,
Johnson, White, & Edwards, 1986; Burns, 1984; Kitson & Sussman, 1982; South &
Spitze, 1986). In recently married couples, factors such as low family income, young
age at the time of marriage, and a wife?s unemployment have been found to be
important (Booth et al., 1986). In addition, the duration of the marriage,
socioeconomic status, religion, parental approval, and number of children (Burns,
1984) have been found to contribute to the likelihood of divorce. Low family
income/socioeconomic status and lack of employment contribute to the likelihood of
divorce in that the marriage does not provide financial security for the partners.
Longer marriages and children can have a negative impact on the likelihood of
divorce in that the couple has more invested in the relationship and are thus more
likely to work toward saving their relationship. Another study with similar findings
has also determined that marital partners? education impacts the likelihood of divorce.
Specifically, higher levels of education have been shown to have a positive impact on
marital quality, thus reducing the potential of divorce (South & Spitze, 1986).
Studies have also investigated whether beliefs about marriage can serve as a
predictor of divorce, and indeed, husbands? and wives? attitudes and beliefs have
been shown to be a means of predicting marital dissolution. When wives have more
traditional attitudes toward marriage than their husbands; i.e., expectations that the
wife will act more as a mother than a financial provider, there is a lesser likelihood of
divorce. The authors speculate that in such situations, the wives may be more likely
5
to accept inequalities in the distribution of housework and other responsibilities. In
contrast, when husbands have more traditional attitudes toward marriage than their
wives, the likelihood of divorce is increased. The authors suggest that the wives in
these situations may be unlikely to accept pressure to fulfill a single role as
homemaker or accept an unequal distribution of household responsibilities which the
husbands in these situations are likely to expect. Additionally, more traditional
husbands are considered less likely to share financial responsibilities with their wives
(Sanchez & Gager, 2000).
Gender differences in expectations have also been examined as a factor in
divorce. Studies have demonstrated that men and women differ on the factors they
cite as the reasons why they divorced. In their longitudinal study of divorced couples,
Amato and Previti (2003) found that women were far more likely to cite specific
behavioral problems on the part of their husbands as the reasons why they desired
divorce. These negative behaviors included physical and emotional abuse, both of
which women were far more likely than men to report as a reason for divorce. In
contrast, they found men were more likely to cite ?communication problems? as the
primary reason for divorce. The authors believed this was the result of men being
less likely to discuss specifically why their marriages failed, thus they would cite a
general reason. Additionally, Amato and Previti found that the majority of men and
women reported that it was the wife who wanted the divorce more than the husband.
Numerous studies have examined whether specific marital problems provide
insight into why certain relationships end. These problems typically vary depending
upon the demographic traits of the complainant (Kitson & Sussman, 1982). The most
6
significant problems leading to relationship dissolution were extramarital affairs, poor
financial management, substance abuse, jealousy, irritability, and irritating habits.
Among those problems, actions categorized as jealous, domineering, and critical were
found to be the strongest predictors of divorce (Amato & Rogers, 1997). Amato and
Rogers found that wives were more likely to report negative aspects of their
husbands? behavior than husbands were to make complaints about their wives?
behavior.
Amato and Rogers? finding regarding behaviors understood as being jealous,
domineering, and critical being strong predictors of divorce is consistent with
research on hostility and aggression?s impact on marital dissolution potential (Newton
& Kiecolt-Glaser, 1995; Rogge & Bradbury, 1996; Rogge, Bradbury, Hahlweg, Engl,
& Thurmaier, 2006). Marital aggression has been shown to be a stronger predictor of
marital dissolution than positive communication as a predictor of marital stability
(Rogge & Bradbury, 1996). Over a five year period, hostility (volatile temper,
destruction of property, threats of harm) has been shown to be the primary difference
between couples that remained married and those that either separated or divorced,
suggesting that it has a powerful impact on marital stability (Rogge et al., 2006).
Furthermore, these types of behaviors, often classified as emotionally abusive, have
been cited in not only quantitative studies, but in qualitative studies as being a
primary factor in a spouse?s decision to leave a relationship. In cases where physical
abuse was also present, the emotional abuse was cited as having a greater impact on
the decision to leave the relationship than the physical abuse (Ulrich, 1991).
Gender and Relationship Dissolution
7
Gager and Sanchez (2003) examined the role gender played in the dissolution
of couple relationships. Specifically, through the study of couples? responses to their
measures they examined the level of agreement between husbands and wives as to the
reasons why their relationships ended. In order to conduct their study, Gager and
Sanchez utilized data from the 1987-1988 National Survey of Families and
Households to find potential study participants. They then contacted potential
participants by phone and interviewed them and their partners. In all their sample
consisted of 2,035 respondents. They interviewed the participants on topics of
marital dissolution, time spent together, and perceptions of household work fairness,
marital happiness, marital trouble in the past year, current marital trouble, and
perceived chances of divorce. They ultimately found that there were significant
gender differences in the relationship between the evaluations of the marriage and the
potential for marital dissolution. They found that women had a higher likelihood of
reporting marital dissatisfaction as well as were more likely to cite specific reasons
for marital dissatisfaction.
What is Emotional Abuse?
Although the term ?emotional abuse,? is used in this thesis, it is important to
keep in mind when discussing emotional abuse that there are elements that target the
victim?s thoughts, not just emotions. Within the literature focused upon emotional
abuse, the problem of definition typically lies within whether or not to treat the term
as synonymous with psychological abuse (O?Hagan, 1995). O?Hagan states that it
does not make sense to define emotional and psychological abuse in the same ways,
as the two words have entirely separate meanings. In O?Hagan?s explanation, if one
8
were to make a statement using the word ?emotional,? the word ?psychological?
could not be used instead without changing the meaning of the statement. However,
O?Hagan also concedes that, ?[e]motionally abusive behavior is nearly always
psychologically abusive and vice versa,? (pg. 451). This is clear when comparing
definitions of emotional and psychological abuse provided by different sources.
One definition of emotional abuse explains it as any behaviors that, without
the use of physical violence, ?reduce the victim?s status and render the victim more
easily controlled by the abuser? (O?Hearn & Davis, 1997, p.376). Psychological
abuse can be conceptualized as, ?any behavior that is harmful or intended to be
harmful to the well-being of a spouse [in a non-physical way]? (Tolman, 1992, p.
292). Murphy and Hoover (2001) make little distinction between psychological
abuse and emotional abuse as they state, ?psychologically abusive behaviors are
directed at the target?s emotional well-being and sense of self? (p. 30).
However, regardless of the term being used, the specific behaviors cited by
authors writing about emotional or psychological abuse are typically the same.
O?Hearn and Davis (1997) write that emotionally abusive behaviors include,
?humiliation, degradation, threats of abandonment, persistent ridicule, and threats of
physical harm,? (p. 376). Behaviors that are considered psychologically abusive
include any that intimidate the victim, isolate the victim, cause him or her to feel
guilty or to blame him or herself for relationship problems, instill fear in the victim,
or cause the victim to feel powerless or helpless (Andersen, Boulette, & Schwartz,
1991). Murphy and Hoover (2001) point out that psychological abuse is directed at
the victim?s ?emotional well-being or sense of self,? and that they ?produce fear,
9
increase dependency, or damage the self-concept of the recipient,? (p. 30).
Additionally, all of the authors agree that in most relationships there may be
occurrences of behaviors that on their own could be classified as abusive. As such,
these behaviors must not be unique occurrences but rather must be part of a larger
pattern of behaviors. Unfortunately, there is little consensus on where the line is
between a unique occurrence and an abusive pattern of behavior when discussing
psychological abuse (Follingstad, Helff, Binford, Runge, & White, 2004).
Emotional Abuse and Potential for Relationship Dissolution
Researchers point to a myriad of reasons to explain why individuals remain in
abusive relationships. Cavanagh (1996) determined that there are six reasons that can
serve as explanations for why an individual would remain in an abusive relationship.
These reasons typically stem from a variety of cognitions. The first reason Cavanagh
provides is that individuals do not perceive the abuse as abusive, instead attributing
the behavior to other factors. Cavanagh?s second reason is that individuals only
consider physical abuse under the definition of abuse. As a result of this, in the
presence of solely psychologically abusive behaviors, these individuals do not
consider abuse as an issue within their relationships. Third, individuals may
rationalize the presence of abusive behaviors in order to protect themselves and their
relationships. In such cases, the victim of abuse concludes that the partner?s
behaviors are the result of some outside influence that is not his or her fault. Fourth,
individuals may have unconscious motives that keep them in abusive relationships.
Cavanagh writes that these motives include the belief that one deserves to be
unhappy, comfort with dysfunction due to being raised in an abusive home, an
10
unconscious association of pain and love, a desire to mute their own inadequacies
through the presence of a significant other with greater inadequacies, the belief that a
partner?s inability to control him or herself reduces or negates blame for abusive
behaviors, and a fear of closeness. Cavanagh?s fifth reason is that remaining in an
abusive relationship is considered to be a better alternative than being alone. In such
cases, an individual may fear being alone, consider him or herself a failure when not
in a relationship, believe that not raising children in a two-parent home would be
detrimental to their development, fear financial insecurity or believe that leaving the
relationship would be the equivalent of admitting he or she had made a mistake in
choosing a partner. Finally, Cavanagh suggests that an individual may remain in an
abusive relationship because he or she has developed a tolerance for the abuse.
Other research supports Cavanagh?s examples of how individuals? cognitions
play a major role in remaining in an abusive relationship. It has been determined that
physically abused women often underestimate the level of risk of abuse within their
relationships, as well as overestimate the level of difficulty they may face if they
choose to leave their relationships (Martin, Berenson, Griffing, Sage, Madry,
Bingham, & Primm, 2000). Herbert, Silver, and Ellard (1991) conducted a study of
abused women from the area around Ontario, Canada in order to examine why they
remained in their relationship. These women were recruited through announcements
on television, the radio, and in newspapers. They found among the respondents in
their study that these women remained in abusive relationships through the
engagement of a number of cognitive strategies that assisted them in perceiving their
relationship in a positive manner. However, Herbert et al. noted that an exception to
11
their findings was the case of verbal abuse. The presence of verbal abuse tended to
mitigate the ability of the women in their study to perceive their relationships in a
positive light and minimize the impact of abuse. They concluded that it was likely
that the greater frequency of verbal abuse may lead to it having a greater impact than
physical abuse on the women they studied.
Furthermore, research has demonstrated that psychological abuse is incredibly
damaging to the recipient in several ways (Lewis, Griffing, Chu, Jospitre, Sage,
Madry, & Primm, 2006; Straight, Harper, & Arias, 2003). Straight et al. found that
individuals who were the victims of psychological abuse experienced poorer physical
health and were more likely to engage in illegal drug use as a coping strategy. In
their study, Lewis et al. found that victims of psychological abuse often demonstrate
depression and greater vulnerability to further psychological abuse. Lewis et al. also
found that in relationships in which physical violence was present in addition to
psychological abuse, the frequency of physical violence was not related to the
victims? distress. This finding is supported by qualitative research involving victims
of abuse in which physical and emotional abuse was present. The victims described
emotional abuse as the worse form of abuse (Follingstad, Rutledge, Berg, Hause, &
Polek, 1990).
Under these circumstances, one would expect the relationship to end.
However, this may take a long time. Findings presently suggest that leaving an
abusive relationship is part of a process rather than a sudden decision. Typically, an
abused woman must reassess her situation as abusive before this process can begin. It
has been found that this often occurs after a ?turning point? that changes the way the
12
relationship is perceived. Ultimately, once a decision has been reached to leave an
abusive relationship, women are effective at taking the necessary measures to leave
(Rosen & Stith, 1997).
Studies have been conducted in which emotional abuse has been examined as
a determinant of divorce. More than one study has found that emotional abuse is
often a major cause of marital failure (Rokach, Cohen, & Dreman, 2004; Kincaid &
Caldwell, 1995). However, it does not always end a relationship. In order to
understand to understand this process we turn to symbolic interaction theory.
Symbolic Interaction Theory
This study was examined through the lens provided by symbolic interaction
theory. Ingoldsby, Smith and Miller (2004) identify three major themes within
symbolic interactionism theory.
The first theme is that how humans behave is related to the meaning they posit
in every facet of the world around them with which they interact. Within this theme,
Ingoldsby et al. include the assumption that the meaning individuals have for things
directly influences how they will react to them. Additionally, meaning is learned
through how individuals relate with other individuals. Based on these assumptions,
different individuals may perceive the same behaviors in completely different ways.
For example, two women are married to men who ask for detailed accounts of their
whereabouts when they are apart. To the first woman, this may mean that her
husband does not trust her and is a jealous individual. To the second woman,
however, this may mean that her husband is concerned about her welfare, likes to
make sure that she has been safe, and is quite attentive. The implication for this study
13
is that the meaning individuals attach to behavior and circumstances varies across
individuals.
The second theme is that individual self-concept plays a large role in how
each individual reacts to his or her experiences. Within this theme, Ingoldsby et al.
places the assumption that an individual?s sense of who he or she is will dictate his or
her behaviors. Ingoldsby et al. posits that all individuals develop their sense of self
through their experiences. This theme can be linked to a primary dimension of
Bowlby?s theory of attachment. Bowlby (1973) explains that an individual?s
attachment style is strongly linked to his or her sense of self. This is the first
dimension of attachment used in my thesis. Based on the idea that self concept is a
primary aspect of attachment, it would follow that an individual?s self-concept could
be linked to whether or not he or she looks to others as a means of validation. As
such, this would be a primary determinant as to whether an individual would be
comfortable in or out of emotionally close relationships. In comparison to individuals
with strong self concepts, individuals with weaker self concepts who look to intimate
relationships as a means of validation may thus be less likely to leave a relationship in
which emotional abuse is present. This ties to one of the primary concepts of
attachment theory, discussed later in this thesis, that how an individual perceives him
or herself is connected to his or her style of attachment.
Symbolic interaction theory?s third theme is that society plays a role in how
individuals determine the meaning of what they experience. The assumption placed
by Ingoldsby et al. within this theme is that society influences the meaning
individuals place upon their experiences. As such, individuals will internalize the
14
values and norms of their society. The couples that will be evaluated in this study are
members of American society, and thus it follows that they will have internalized the
values and norms of the United States. One of those norms is that individuals work
toward marriage and raising a family. The value placed by society on this norm is
that it is good to be in an intimate relationship in order to work toward the goal of
marriage and child rearing. As such, individuals who have completely internalized
this value may be less likely to feel comfortable outside of a couple relationship. In
the case of couple relationships in which abuse is present, the meaning an individual
places not only upon his or her partner?s behaviors, but also on the meaning he or she
places upon relationships themselves, will have powerful influences over his or her
desire to remain in the relationship. This is the second dimension of attachment used
in this thesis.
Attachment Theory
Attachment theory was first conceptualized by John Bowlby as he sought to
explain infants? reactions when separated from their primary caregivers. Bowlby
theorized that the infants? negative reactions, typically crying or screaming, that
resulted from this separation were due to an attachment bond they had with their
primary caregiver, or attachment figure (Bowlby, 1969). Bowlby also theorized that
the attachment bond children formed with their primary caregivers would determine
how these children conceptualized future relationships (1973). Furthermore, Bowlby
(1973) also theorized that an individual?s confidence in an attachment figure hinged
upon both, ?(a) whether or not the attachment figure is judged to be the sort of person
who in general responds to calls for support and protection; [and] (b) whether or not
15
the self is judged to be the sort of person toward whom anyone, and the attachment
figure in particular, is likely to respond in a helpful way,? (p. 204).
Later studies would support Bowlby?s theory as it applied to how children
behaved in relationships with primary caregivers in their pre-school years
(Ainsworth, Blehar, Waters, & Wall, 1978). Furthermore, Ainsworth et al.
determined that three types of attachment existed. These three types are typically
identified as secure, anxious/ambivalent, and avoidant. Based on which type of
attachment a child exhibited, reactions to a primary caregiver could be predicted.
According to Ainsworth et al., the secure type was typified by the welcoming of a
caregiver?s affection after his or her absence, anxious/ambivalent types were difficult
to comfort when reunited with an absent caregiver, and avoidant types avoided
interaction with a previously absent caregiver
Hazan and Shaver (1987) took the work of Bowlby as well as Ainsworth et al.
a step further, theorizing that adult romantic relationships could be understood within
the same context as the attachment bonds between children and parents. Basing their
study on Bowlby?s theory of attachment and using the framework of definitions
provided by Ainsworth et al., Hazan and Shaver found strong evidence to support
their belief that attachment theory could be applied to adult romantic relationships.
Through the use of two questionnaire studies conducted with two separate samples,
the first a sample of 620 respondents to a newspaper ad, and the second a sample of
108 college students, they determined that the type of attachment an individual
demonstrated would impact upon their experiences of romantic relationships,
individuals? attachment styles stemmed from mental models of self as well as of
16
social relationships and relationships with parents, and that individuals with different
attachment styles had different beliefs about romantic love.
The work of Bowlby, Ainsworth et al. and Hazan and Shaver was furthered by
Bartholomew and Horowitz (1991). They conceptualized a model of attachment
styles that included four groups, expanding upon the three group model that had been
the standard up to that point. Bartholomew and Horowitz based their four types on
Bowlby?s (1973) explanation of how an individual?s view of self and view of an
attachment figure determined how a relationship was perceived. They identified that
an individual believes he or she is worthy of love and affection, which coincides with
high sense of self and low levels of dependency, or believes he or she is not worthy of
love and affection, which coincides with low sense of self and high levels of
dependency. They also identified that an individual either expects others to be
trustworthy and available to care for him or her, which coincides with low levels of
avoidance or an individual does not expect others to be trustworthy and available to
care for him or her, which coincides with high levels of avoidance. Through different
combinations of the models of self and models of others Bartholomew and Horowitz
identified four different types of attachment styles which they labeled as secure,
preoccupied, dismissing, and fearful. This expanded on the previous three category
model in that the avoidant style of attachment was split into two styles that
Bartholomew and Horowitz identified as dismissing and fearful.
The secure style of attachment results from a combination of the belief that
one is worthy of love and affection and the belief that others are trustworthy and
available to care. This attachment style is characterized by comfort with intimate
17
relationships as well as personal independence. The preoccupied style of attachment
results from a combination of the belief that one is not worthy of love and affection
and the belief that others are trustworthy and available to care. This attachment style
is characterized by, ?an overinvolvement in close relationships, a dependence on
other people?s acceptance for a sense of personal well-being, a tendency to idealize
other people, and incoherence and exaggerated emotionality in discussing
relationships? (p. 228). The dismissing style of attachment results from the belief that
one is worthy of love and affection and the belief that others are not trustworthy and
available to care. This attachment style is characterized by, ?a downplaying of the
importance of close relationships, restricted emotionality, an emphasis on
independence and self-reliance, and a lack of clarity or credibility in discussing
relationships,? (p. 228). The fearful style of attachment results from the belief that
one is not worthy of love and affection and the belief that others are not trustworthy
or available to care. This attachment style is characterized by, ?avoidance of close
relationships because of a fear of rejection, a sense of personal insecurity, and a
distrust of others,? (p. 228). Figure 1 demonstrates how an individual?s sense of self
and sense of others intersect to form his or her attachment style.
Model of Self
(Dependence)
Positive Negative
(Low) (High)
18
SECURE PREOCCUPIED
Positive Comfortable with intimacy Preoccupied with
(Low) and autonomy relationships
Model of Other CELL I CELL II
(Avoidance)
Negative DISMISSING FEARFUL
(High) Dismissing of intimacy Fearful of intimacy
Counter-dependent Socially avoidant
CELL III CELL IV
Figure 1. Model of Adult Attachment, (Bartholomew & Horowitz, 1991)
Bartholomew and Horowitz (1991) conducted a pair of studies in order to
determine whether or not the beliefs of individuals categorized by each of their four
styles of attachment were consistent with their definitions. In their first study they
utilized a sample of forty female and thirty-seven male college students participating
in an introduction to psychology course. These students answered two sets of
questionnaires about themselves as well as their partners in order to investigate their
friendships and romantic relationships. Their second study was designed to replicate
the first study as well as include information about the subjects? families of origin.
For the second study they utilized a sample of 33 female college students and 36 male
college students from an introduction to psychology course. These subjects were
each interviewed twice. During the first interview they discussed their relationships
with family and friends. During the second interview they completed a questionnaire
about their relationships. Bartholomew and Horowitz found strong evidence that
these attachment styles were, indeed, separate categories and the beliefs and
behaviors of these individuals fit within those they associated with each style of
attachment.
19
Different Attachment Styles and Relationship Dissolution Potential
Only one study was discovered that evaluated how attachment affected an
abused woman?s decision to leave a physically abusive relationship (Henderson,
Bartholomew, & Dutton, 1997). This study was conducted over the course of six
months using a sample of 59 women who had recently left an abusive relationship.
The women initially answered questions through a semi-structured interview as well
as completed measures that tested received and inflicted physical and psychological
abuse. The women then completed questionnaires about their separation after two
months and six months. The research found that attachment security was associated
with neither receiving less abuse within relationships nor greater success in leaving an
abusive relationship. The authors did find limited evidence that individuals who can
be classified as having a preoccupied style of attachment were less likely to resolve to
leave an abusive relationship as well as less successful in leaving once they had made
the resolution to leave. The authors suggest that this may be accounted for by other
environmental factors that they did not examine. Their study is limited in a few
areas. First, the study, though it included a measure of emotional abuse, primarily
focused on physical abuse or the combination of abuse, but not solely on the
emotional abuse. Second, the sample contained an overrepresentation of the
preoccupied and fearful attachment styles. Third, the authors themselves report that
their research was limited by the fact that it only focused on the female partners and
had no information on the male partners in abusive relationships. The present study
was able to look at both males and females and at emotional abuse rather than
physical abuse.
20
Implications of Attachment Theory for Whether Individuals
Will Leave Abusive Relationships
Several studies to be described below have examined an individual?s
attachment style and how he or she operates within a relationship. Specifically, these
studies all demonstrate that an individual?s perceptions of his or her relationship as
well as how he or she reacts to these perceptions are affected by his or her attachment
style. Based on the findings of these studies, it stands to reason that an individual?s
attachment style will affect how he or she experiences emotional abuse, and that
experience or interpretation should affect his or her desire to leave an abusive
relationship.
Research supports that there is a strong link between attachment and the types
of attributions (meaning placed upon behaviors) individuals make about their partners
(Collins, Ford, Guichard, & Allard, 2006). Collins et al. concluded that individuals
with a preoccupied style of attachment felt greater anxiety as a result of perceived
partner transgressions as well as made more attributions that their relationship was
threatened by partner behaviors and engaged in more conflict-inducing behaviors than
individuals with other styles of attachment. Attachment styles have a profound
impact on marital quality. A secure style of attachment is associated with the
development of a strong marriage with resilient qualities that allow for managing life
stressors. Conversely, individuals with any of the three insecure attachment styles-
preoccupied, dismissing, or fearful- typically find themselves in unstable
relationships that have greater vulnerability to stress (Hollist & Miller, 2005). Within
the context of intimate relationships, attachment styles also play a role in how
21
partners perceive levels of equity (Grau & Doll, 2003). Grau and Doll found that
securely attached individuals primarily described their relationships as equitable,
meaning that they gave and took equally in comparison to their partners. In
comparison, anxiously (preoccupied) attached individuals indicated that they gave
more and took less than their partners. Avoidant (dismissing and fearful) individuals
tended to perceive equity in their relationships, but in comparison to secure
individuals, indicated that they gave and took less. Furthermore, it has been
determined that perceptions of and beliefs about relationships are a primary
determinant of whether an individual will entertain leaving an abusive relationship
(Byrne & Arias, 2004). As prior research has shown, attachment styles are closely
linked with individuals? perceptions concerning their partners? behaviors and their
relationships as well as how they react to these perceptions. Therefore, attachment
styles may be linked to whether an individual responds to partner abuse by making
moves toward leaving the relationship.
How individuals perceive their partners? actions as well as their beliefs about
their partners and how they believe they should respond to these actions with regards
to their relationship differs depending on an individual?s attachment style (Fincham,
2001). A preoccupied attachment style has also been demonstrated as having an
impact on an individual?s perception of the level of conflict experienced in his or her
relationship, how that conflict is responded to, and perceptions of support within the
relationship (Campbell, Simpson, Boldry, & Kashy, 2005). Campbell et al. found
that such individuals were more likely than others to perceive greater conflict within
their relationships as well as greater conflict escalation. Additionally, these
22
individuals showed a greater range of perceptions of their relationship stability.
Specifically, in the presence of conflict, preoccupied individuals perceived their
relationships in a more negative light than others. In the presence of positive
behaviors these individuals made more positive appraisals of their relationship
stability and the future of their relationship than others.
Fishtein, Pietromonaco, and Barrett (1999) determined that attachment style
plays a role in how an individual perceives conflict within his or her relationship.
They found that individuals identified as having a preoccupied style of attachment
were significantly more likely to perceive positive outcomes due to conflict whereas
individuals identified with the other three styles of attachment did not. Specifically,
preoccupied individuals believed that conflict brought them closer to their partners.
This is consistent with such individuals? preoccupation with closeness. As such, they
perceived conflict as positive in that it increased intimacy due to conflict forcing their
partners to engage them.
Guerrero (1998) found several differences in how individual attachment styles
affected the experience and expression of jealousy caused by perceived threats to
their romantic relationships. Findings were consistent with the author?s expectations
based on how each attachment style is composed. Individuals with attachment styles
that include negative self-models, preoccupied and fearful, were more likely to
experience jealousy than those who were identified as secure or dismissing. Those
with negative models of others, dismissing and fearful, were less likely to attempt to
maintain their relationships or engage in jealous behaviors. Preoccupied individuals
were more likely to engage in surveillance behaviors as well as relationship
23
maintaining behaviors than individuals identifying with the other three models.
Finally, Guerrero found that individuals with the dismissing attachment style were the
least likely to experience fear or sadness along with the experience of jealousy in
comparison to individuals with the other three attachment styles.
Davis, Shaver, and Vernon (2003) found that attachment style affects how an
individual reacts to relationship dissolution. The behaviors engaged in by individuals
identifying with each attachment style were consistent with their types of attachment.
Secure individuals were most likely to engage in coping strategies by engaging
friends and family for support. Dismissing and fearful individuals? behaviors were
weakly or negatively associated with relationship maintaining behaviors and
positively associated with both avoidant and self-reliant coping behaviors.
Preoccupied individuals experienced the greatest levels of physical and emotional
distress due to the loss of their relationships. Additionally, preoccupied individuals
were most likely to engage in behaviors directed toward saving their relationships as
well as negative coping strategies such as drug and alcohol use.
Hypotheses
Based on prior research, it was expected that the degree of experience of
emotional abuse would be associated with an individual?s intent to leave an abusive
relationship. It was expected that in the presence of equal levels of emotional abuse,
women would take greater steps toward leaving a relationship than men. It was
expected that independent of abuse, individuals with different attachment styles
would take different steps toward leaving a relationship. Finally, it was expected that
an individual?s style of romantic attachment will determine how powerful the effect
24
of emotional abuse is on that individual?s intent to leave an abusive relationship. This
study has four major hypotheses.
Hypothesis 1: Positive Association between Psychological Abuse and Steps Taken
Toward Leaving a Relationship
It was hypothesized that there would be a positive association between
emotional abuse and intent to leave a relationship. Specifically, as abuse increased,
steps taken toward leaving a relationship would increase as well. Figure 2
demonstrates the expected outcome for individuals? desire to leave in the presence of
emotional abuse.
Steps Taken to Leave
Low Medium High
Low X
Level
of Medium X
Abuse
High X
Figure 2. Expected results among individuals? intent to leave in the presence of emotional abuse
Hypothesis 2: At Each Level of Emotional Abuse Women will Take Greater Steps
toward Relationship Dissolution than Men
It was hypothesized that at each level of emotional abuse, women would be
more likely to have taken further steps toward marital dissolution than men. This is
based on recent research that has determined that wives? negative evaluations of a
relationship, in the presence of specific examples, are a stronger predictor of divorce
than men?s (Gager, C.T. & Sanchez, L., 2003).
25
Hypothesis 3: Different Individual Attachment Styles Will be Linked
Differently to Steps Taken to Leave a Relationship
It was hypothesized that attachment styles would be differentially associated
with steps taken to leave a relationship. Specifically, a secure style of attachment
would show no correlation with steps taken to leave a relationship, dismissing and
fearful styles of attachment will show a positive correlation with steps taken to leave
a relationship, and a preoccupied style of attachment would show a negative
correlation with steps taken to leave a relationship.
Hypothesis 4: The Relationship Between Steps Taken Toward Leaving a
Relationship and Abuse will Differ Depending on an Individual?s Attachment Style
It was hypothesized that an individual?s romantic attachment style would have
a moderating effect on his or her desire to leave an emotionally abusive relationship.
Depending on the attachment style of the abused individual, it was expected that he or
she would be more or less likely to have intent toward leaving an abusive
relationship. Specifically, an individual who identifies him or herself as having a
secure, dismissing, or fearful style of attachment have greater intent toward leaving
his or her relationship when emotional abuse is present, while an individual who
identifies him or herself as having a preoccupied style of attachment demonstrate less
intent toward leaving his or her relationship when emotional abuse is present.
Hypothesis 4.1: Individuals with a Secure Style of Attachment Demonstrate a
Positive Association Between Experienced Emotional Abuse and Steps Taken Toward
Leaving the Relationship
26
It was hypothesized that, for individuals with a secure romantic attachment
style, emotional abuse would be positively associated with intent to leave a
relationship. This is because individuals who identify themselves as having a secure
style of attachment perceive both themselves and others in a positive manner and
demonstrate a low level of dependence on relationships (see Figure 1, Cell I). Due to
this, secure individuals would not likely to tolerate being treated in an emotionally
abusive manner.
Hypothesis 4.2: Individuals with a Dismissing Style of Attachment Demonstrate a
Positive Association Between Experienced Emotional Abuse and Steps Taken Toward
Leaving the Relationship.
It was hypothesized that, for individuals who identify themselves as having a
dismissing style of attachment, emotional abuse would be positively associated with
intent to leave a relationship. This is because individuals who identify themselves as
having a dismissing style of attachment place a high level of value upon themselves
but low value on others (see Figure 1, Cell III). Dismissing individuals are fiercely
independent to a degree that they place greater worth on independence over intimacy.
As such, they are likely to be less tolerant of emotionally abusive behaviors. The
principal difference between these two styles of attachment and the preoccupied style
is that individuals with a preoccupied style of attachment value others more than they
value themselves.
Hypothesis 4.3: Individuals with a Preoccupied Style of Attachment would Not
Demonstrate any Correlation Between Experienced Emotional Abuse and Steps
Taken Toward Leaving the Relationship.
27
It was hypothesized that individuals with a preoccupied style of attachment
would not demonstrate a strong desire to leave their relationships regardless of the
level of emotional abuse present. Preoccupied individuals place a higher value on
others than they do on themselves as well as a high level of dependence on
relationships (see Figure 1, Cell II). A preoccupied individual may remain in an
abusive relationship for a variety of reasons. Such an individual may believe that he
or she is better off in an abusive relationship rather than alone. It is also possible that,
because he or she does not value him or herself as much as the partner, he or she may
be more vulnerable to internalizing the emotional abuse. Consequently, his or her
self-worth would continue to decline, further reinforcing the belief that the partner is
superior in the relationship.
Hypothesis 4.4: Individuals with a Fearful Style of Attachment would Demonstrate a
Strong Positive Association Between Experienced Emotional Abuse and Steps Taken
Toward Leaving the Relationship.
It was hypothesized that for individuals who identify themselves as having a
fearful style of attachment, emotional abuse would be associated with greater intent to
leave a relationship. Although a fearful individual also does not value him or herself,
he or she also does not value others, and as such avoids close relationships (see
Figure 1, Cell IV). Due to this, a fearful individual would be likely to seize upon
abuse as a reason to get out of a close relationship.
28
Figure 3: The Moderating Effect of Romantic Attachment Styles on Steps Taken to
Leave an Abusive Relationship
Secure Attachment
(+)
(+)
Dismissing
Attachment
(+)
(+)
Emotional Abuse Intent to Leave
Preoccupied
Attachment
(0)
0
Fearful Attachment
(+)
(+)
Definition of Variables
Independent Variables
Emotional Abuse
Any non-physically injurious behavior that may intimidate the victim, isolate
the victim, cause him or her to feel guilty or to blame him or herself for relationship
problems, instill fear in the victim, or cause the victim to feel powerless or helpless
(Andersen et al., 1991).
Romantic Attachment Style
An individual?s identified style of attachment. Based on Bartholomew and
Horowitz (1991) there are four specific attachment styles.
Secure:
29
An individual with a secure style of attachment values intimacy, is capable of
preserving independence although at the same time maintaining an intimate
relationship, and demonstrates logic in handling relationship issues.
Dismissing:
An individual with a dismissing style of attachment places a higher value
upon independence over intimate relationships, minimizes the importance of close
relationships, demonstrates a limited range of emotions, and has difficulty discussing
relationships in a cogent manner.
Preoccupied:
An individual with a preoccupied style of attachment demonstrates excessive
involvement in intimate relationships, and a need for the acceptance of others as a
means of determining his or her own self-worth, puts others on a pedestal, has
difficulty discussing relationships in a cogent manner and displays exaggerated
emotionality.
Fearful:
An individual with a fearful style of attachment avoids intimate relationships
due to a fear of rejection, distrusts others, and is generally insecure.
Gender
An individual?s identified sex, either male or female.
30
Dependent Variable
Steps Taken Toward Leaving a Relationship
This variable is defined as the number of actions taken by an individual in
order to end his or her intimate relationship. These actions range from thoughts of
separation to actually filing for divorce.
Control Variables
Because it was learned from the review of previous research that certain other
variables are strongly linked to with an individual?s intent to leave a relationship, they
were included as controls within this study. Specifically in this study, socioeconomic
status and length of relationship will be used as controls.
Socioeconomic Status
Respondents? socioeconomic status was controlled due to the fact that a
higher socioeconomic status can reduce the likelihood of divorce (Booth et al., 1986).
As such, it stands to reason that it may also negatively impact an individual?s desire
to leave a relationship. This variable will be measured by an individual?s self-reported
gross income.
Length of Relationship
Individuals who have been in longer relationships have more invested in their
relationship, thus they will be less likely to desire terminating their relationship
(Burns, 1984). This factor was also controlled within the present study.
31
CHAPTER 2: METHODS
Participants
For the study, secondary analyses were used to examine data collected by an
ongoing study focusing on the treatment of abuse in couple relationships being
conducted at the Family Service Center located on the campus of the University of
Maryland, College Park. Participants for the larger study were couples who
voluntarily contacted the Family Service Center in order to receive couple therapy.
Prior to the onset of therapy services, all couples completed a set of self-report and
behavioral measures. The sole requirement for participating in the first stage of the
study was to contact the Family Service Center and set up an appointment for
assessment prior to being seen by a therapist for couple therapy.
Procedures
The data used by this study were collected between November, 2001 and
February, 2005 as part of an ongoing study of couple abuse by Dr. Norman Epstein
and Dr. Carol Werlinich through the Family Service Center at the University of
Maryland, College Park. Data were collected by the therapist interns who staff the
Family Service Center. The couples from whom the data were collected voluntarily
contacted the Family Service Center seeking treatment. After completing an intake
over the phone, each couple would be contacted by a pair of therapist interns who
would serve as their therapists. An initial appointment to complete the assessment
instruments would be set up prior to the implementation of therapeutic services. Prior
to beginning work on the assessment instruments, the couple is informed that the
Family Service Center is a research facility and the measures they are completing
32
would assist the therapists in their work with the couple. The couple is then separated
into two rooms so they can complete their individual sets of measures privately. The
therapist interns check on each member of the couple approximately every twenty
minutes in order to make sure they are not having difficulty completing the measures
as well as to answer any questions they may have regarding the instruments.
Additionally, the instruments are checked for completeness prior to ending the
assessment appointment.
Measures
Emotional Abuse
Experienced emotional abuse within the relationship was measured using the
Multi-Dimensional Emotional Abuse Scale (MDEAS; Murphy, Hoover, & Taft,
1999). This 28 item self-report tool, presented in Appendix A, has an overall score
on emotional abuse as well as scores on four subscales. These four subscales include:
Restrictive Engulfment, measured by items 1-7 (e.g., ?Tried to stop the other person
from seeing certain friends or family members?); Denigration, measured by items 8-
14 (e.g., ?Said or implied that the other person was stupid?); Hostile Withdrawal,
items 15-21 (e.g., ?Refused to have any discussion of a problem?); and
Dominance/Intimidation, items 22-28 (e.g., ?Stood or hovered over the other person
during a conflict or disagreement?).
Each of the 28 items asks the individual to rate how often he or she as well as
his or her partner engaged in a specific behavior on a scale from 0 (never) to 6 (more
than 20 times in the past 4 months). For the purposes of this study, only the
individual?s rating of his or her partner?s behaviors was used for scoring whether he
33
or she experienced emotional abuse. Subscales are scored from a minimum of 0 to a
maximum of 42. An overall score ranges from a minimum of 0 to a maximum of
168. Low scores indicate low levels of emotional abuse experienced by an individual
while higher scores indicate higher levels of experienced emotional abuse.
The MDEAS was created as a means of assessing emotional abuse as a
multifactorial construct based on prior definitions of different types of emotional
abuse. Based on the literature, the authors of the MDEAS decided upon a 4-factor
model of emotional abuse which became the four subscales on the MDEAS.
Preliminary research on the scale was conducted by administering a 34-item set to
160 students in dating relationships. Items with low response rates, low item-scale
correlations, or poor differential correlation were removed from the measure. Other
items were added to the scale based on the clinical experience of the authors. This
resulted in a 54-item measure that was given to a second sample of college students in
dating relationships. Further analysis of this scale supported the authors? four factor
model of emotional abuse. Based on factor analysis, items were removed from the
54-item measure, resulting in the present 28-item, four subscale measure (Murphy &
Hoover, 2001).
For the purposes of this study participants? total score on the MDEAS was
utilized. This score was determined by totaling each participant?s responses
regarding his or her partner?s behaviors. In order to use this score with analysis of
variance tests the MDEAS scores were divided into three equal groups representing
low, moderate, and high levels of emotional abuse.
34
Attachment Style
An individual?s identified attachment style was measured using the
Relationship Questionnaire (RQ, Bartholomew & Horowitz, 1991). Based on prior
attachment research on children, the authors determined that four specific types of
adult attachment exist: secure, dismissing, preoccupied, and fearful.
The RQ, presented in Appendix B, includes four paragraphs, each of which
illustrates one of the four attachment styles: secure (e.g., ?It is relatively easy for me
to be emotionally close to others. I am comfortable depending on others and having
others depend on me. I don?t worry about being alone or having others accept me);
dismissing (e.g., ?I am somewhat uncomfortable getting close to others. I want
emotionally close relationships, but I find it difficult to trust others completely, or to
depend on them. I sometimes worry that I will be hurt if I allow myself to become
too close to others); preoccupied (e.g., ?I want to be completely emotionally intimate
with others, but I often find that others are reluctant to get as close as I would like. I
am uncomfortable being without close relationships, and I sometimes worry that
others don?t value me as I value them); and fearful (e.g., ?I am comfortable without
close relationships. It is very important to me to feel independent and self-sufficient,
and I prefer not to depend on others or have others depend on me). Individuals score
each paragraph as it relates to themselves on a scale of 0 (Not at all like me) to 7
(Very much like me). Additionally, participants are asked to select which of the four
paragraphs describes them best overall. In the event of a tie score between items, this
selection is used to determine the individual?s attachment style. For the purpose of
this study, each participant?s identified attachment style was determined based upon
35
which of the four paragraphs he or she selected as being most representative of him or
herself. This was determined based on the paragraph being used as a tie breaker, and
as such, could be considered most representative regardless of scores as this study is
not examining how strongly individual?s identify with their attachment styles.
This measure has demonstrated criterion validity as each of the four types of
attachment style were associated with different variables. Furthermore, this measure
demonstrates construct validity through its association with a semi-structured
interview utilizing the same four-category model of attachment (Bartholomew &
Horowitz, 1991).
Steps Taken Toward Relationship Dissolution
Actions taken by an individual toward ending his or her relationship were
measured using the Marital Status Inventory-Revised (MSI-R, Epstein & Werlinich,
2001). This measure, presented in Appendix C, includes 18 items that are responded
to in a ?yes? or ?no? format. Questions range from, ?Had frequent thoughts about
separating from your partner, as much as once a week or so,? to ?Filed for divorce or
ended the relationship.? The measure is scored from 0 to 18 based on the number of
?yes? responses given. A lower score indicates low fewer steps taken toward leaving
a relationship while a higher score indicates greater greater steps taken toward
relationship dissolution. Epstein and Werlinich modified the original MSI to include
neutral language that would apply to all intimate couples rather than only legally
married heterosexual couples, resulting in the MSI-R. This study will determine
participants? total MSI-R scores based on the total number of ?yes? responses
submitted on the MSI-R.
36
The original MSI was a 14 item self-report measure designed to determine the
level of intent an individual shows toward leaving a relationship (Weiss & Cerreto,
1980). As the MSI-R has been used in little research up to this point, normative data
are not yet available on mean scores in clinical or community samples.
37
CHAPTER 3: RESULTS
Study Participants
The sample for the study consisted of 288 heterosexual couples who
completed the first portion of measures associated with the study. Slightly greater
than half (55.2%) of the sample couples were currently married and living together
and nearly three quarters (70.9%) of the sample couples were living together at the
time they began therapy (please see Table 1 for demographic information of the
sample). The mean length of relationship across the sample was 7.08 years. The
mean age of participants was 32.86 years. Among women the mean age was 32.09
years and among men the mean age was 33.63 years. The mean annual income of the
sample was $29,583. Among women the mean annual income was $24, 886. Among
men the mean annual income was $34,352.
Table 1
Demographics by Gender (in means or percentages)
Variable Females Males
Total 288 288
Mean Age (in years)
a
32.09 (8.99) 33.63 (9.32)
Minimum to Maximum Age (in years)
b
17 ? 65 (48) 19 ? 82 (63)
Mean Length of Relationship (in years)
a
7.14 (6.826) 7.01 (6.814)
Minimum to Maximum Length of
Relationship (in years)
b
0 ? 41 (41) 0 ? 41 (41)
Mean Income (in dollars)
a
24,886 (19,225) 34,352 (28,597)
Attachment Style
Secure
Fearful
Preoccupied
Dismissing
100
107
29
27
118
72
40
35
Mean MDEAS Score
a
40.13 (30.59) 36.38 (29.82)
Mean MSI-R Score
a
7.08 (4.20) 5.44 (4.20)
a
Standard Deviations Listed in Parentheses;
b
Range Listed in Parentheses
38
Univariate and Bivariate Analyses
Prior to hypothesis testing correlations between the continuous variables were
run in order to examine basic relationships. This was of particular interest for
initially determining the relationship between the planned control variables, length of
relationship and socioeconomic status, and the continuous independent and dependent
variables, the total abuse score determined by the MDEAS and the total score
determining steps taken toward leaving as determined by the MSI-R, respectively.
Table 2 shows the correlation matrix.
Table 2
Correlation Matrix
Variables Income Length of
Relationship
MDEAS Score MSI-R Score
Income
Length of
Relationship
.207
**
MDEAS Score -.066 .049
MSI-R Score -.032 .042 .428
**
**
Correlation is significant at the .01 level (2-tailed)
As shown by the table, the only significant relationships were between income
and length of relationship and between the MDEAS score of total experienced
emotional abuse and the MSI-R score of steps taken toward leaving the relationship.
Counter to initial predictions, there was no significant relationship between the
control variables and the dependent variable. The correlation between income and
length of relationship, though significant, is only weakly positive. Based on these
findings it was determined that there was no need to control for these factors during
39
further hypothesis testing. There is, however, a moderate to moderately strong
positive correlation between the level of experienced emotional abuse and the steps
taken to leave the relationship. This supports the first hypothesis of this paper, that as
experienced abuse increases an individual will take greater steps toward leaving the
relationship.
Hypothesis Testing
Hypothesis 1: Positive Correlation between Psychological Abuse and Steps Taken
Toward Leaving a Relationship
In order to examine the relationship between emotional abuse and leaving a
relationship, an analysis of variance was conducted. In order to run an ANOVA, the
abuse scores had to be broken down into three proportional groups. Through the use
of the SPSS program the MDEAS scores were broken into three equal groups,
signifying low, moderate, and high levels of experienced emotional abuse. Total
scores of 21 or less were considered to be low levels of emotional abuse. Total scores
from 22 to 46 were considered to be moderate levels of emotional abuse. Scores of
47 or greater were considered to be high levels of emotional abuse. These levels were
determined by dividing the sample into three separate groups with each group
representing a third of the sample. Table 3 shows the mean MSI-R scores associated
with each of these groups as well as the number of women and men that fell within
each group. Additionally, post hoc comparisons were conducted using Bonferroni?s
comparison test in order to reduce the likelihood of a Type I error occurring. The
results of this test confirmed that there were strong significant differences (p < .000)
in MSI-R scores between each group across the entire sample. The results of the
40
Bonferroni comparison are shown in Table 4. Based on the ANOVA test we can
conclude that steps taken to leave a relationship do increase as abuse increases. The
mean MSI-R score for the sample at low levels of abuse was 4.41, at moderate levels
of abuse the mean MSI-R score for the sample was 6.07, and at high levels of abuse
the mean MSI-R score for the sample was 8.47. Furthermore, the Bonferroni
comparison confirms that the number of steps taken to leave between increasing
abuse levels are significantly different.
Table 3
Mean MSI-R Scores by Abuse Level ? Women and Men
Variable Abuse Level Mean Std. Deviation N
Female Low <= 21
Moderate 22 ? 46
High 47+
5.50
6.37
9.24
4.26
3.68
3.73
89
96
93
Male Low <= 21
Moderate 22 ? 46
High 47+
3.46
5.72
7.62
3.74
3.76
4.04
103
83
85
Total Low <= 21
Moderate 22 ? 46
High 47+
4.41
6.07
8.47
4.11
3.72
3.95
192
179
178
Dependent Variable: MSI-R 
Table 4
Bonferroni Comparison ? Total for Women and Men
Variable (I) Abuse (J) Abuse Mean
Difference
(I-J)
Std.
Error
Sig.
Bonferroni
Low <=21 Moderate 22 ? 46
High 47+
-1.66
*
-4.06
*
.409
.410
.000
.000
Moderate 22 ? 46 Low <= 21
High 47+
1.66
*
-2.40
*
.409
.417
.000
.000
High 47+ Low <= 21
Moderate 22 - 46
4.06
*
2.40
*
.410
.417
.000
.000
*
The mean difference is significant at the .05 level.
41
Hypothesis 2: In the Presence of Abuse, women will Take Greater Steps Toward
Leaving an Intimate Relationship than men
As experienced abuse increases, steps taken toward leaving a relationship
increase for both men and women. Figure 4 shows the plots of mean MSI-R scores
for both genders and demonstrates graphically the difference in mean scores shown in
Table 3.
Figure 4: Plot: Difference between Men and Women?s Mean MSI-R Scores at Each
Level of Abuse
42
At each level of abuse women had a higher mean MSI-R score than men. At
low levels of abuse women?s mean scores were 2.04 higher than men?s. At moderate
levels of abuse the mean scores between gender were somewhat closer with women?s
mean scores 0.65 higher than men?s. At high levels of abuse, the difference between
mean scores became greater again, with women?s mean MSI-R scores 1.6238 higher
than men?s. It is interesting to note, however, that although tests of between-subjects
effects demonstrated that gender has a significant effect upon steps taken to leave a
relationship (p < .000) and that experienced abuse has a significant effect upon steps
taken to leave a relationship (p < .000), the interaction of gender and abuse were not
demonstrated as having a significant effect upon steps taken toward leaving a
relationship (p = .215). Table 5 shows the results of tests of between-subjects effects.
Table 5
Tests of Between-Subjects Effects
Source Type III
Sum of
Squares
df Mean
Square
F Sig. Partial
Eta
Squared
Corrected
Model
1867.832
a
5 373.57 24.89 .000 .186
Intercept 21834.345 1 21834.34 1455.04 .000 .728
Gender 282.486 1 282.49 18.83 .000 .034
Abuse 1449.571 2 724.79 48.30 .000 .151
Gender*Abuse 46.277 2 23.14 1.54 .215 .006
a. R Squared = .186 (Adjusted R Squared = .179)
Additionally, though gender is shown to have a statistically significant effect,
based on the partial eta squared for gender, its effect only accounts for 3.4% of the
variance in the total model. Based on the differences of mean scores, there is support
for the hypothesis that women will take greater steps toward leaving a relationship
43
than men; however, based on the interaction effect (p = .215) women are not any
more likely than men to leave an abusive relationship.
Hypothesis 3: Depending on Attachment Style, Individuals Will Take Different
Degrees of Steps Toward Leaving a Relationship
It was hypothesized that individuals with different attachment styles (secure,
dismissing, preoccupied, and fearful) would take different steps toward leaving a
relationship, independent of abuse. Specifically, it was expected that a secure
attachment would demonstrate no effect upon steps taken to leave a relationship,
dismissing and fearful attachment would have a strong positive effect on steps taken
to leave a relationship, and that a preoccupied attachment style would have a negative
effect on steps taken to leave a relationship. Table 6 shows the mean MSI-R scores
for each attachment group.
Table 6
Mean MSI-R Scores by Attachment Style
Variable N Mean Std. Deviation Std. Error
Secure 218 5.29 4.30 .291
Fearful 178 6.88 4.13 .310
Preoccupied 69 6.71 3.80 .457
Dismissing 62 7.42 4.43 .563
Dependent Variable: MSI-R 
 
As shown in Table 6, the sample size is not equally distributed across the four
attachment styles. Based on the fact that this sample was comprised of couples who
had sought therapy it was expected that there would be greater numbers of secure and
preoccupied individuals based on their higher expectations of others. It was thus
expected that there would be fewer dismissing and fearful individuals based on their
44
lower interest in others. This, however, was not the case. Most individuals identified
themselves as secure (N = 218), which was within expectations. Additionally, the
fewest individuals identified themselves as dismissing (N = 62), which was also
within expectations. What was not expected was the large number of individuals who
identified themselves as fearful (N = 178) as well as the low number of individuals
who identified themselves as preoccupied (N = 69). Although there were differences
in the mean scores between the groups (see figure 5 for plots of mean scores by
attachment style), further examination of post hoc testing revealed that the differences
between groups were not always significant. A Bonferroni comparison was
conducted (please see Table 7 for results) and revealed that for the entire sample the
only group that differed significantly from the others was the secure group. That
group only differed significantly from the fearful and dismissing groups.
Furthermore, post hoc testing revealed that the preoccupied group did not differ
significantly from the other three groups with regards to steps taken toward leaving a
relationship. This was also counter to expectations as it was hypothesized that
preoccupied individuals would take significantly fewer steps toward leaving a
relationship than individuals identified with any of the other three styles of
attachment. It should be noted here that although the Bonferroni comparison is less
powerful than other tests, it was chosen due to its conservative nature in order to
protect against Type I errors from occurring because there were fewer than 100
individuals within the preoccupied and dismissing groups.
45
Figure 5: Plot: Mean MSI-R Scores by Attachment Style
(A = Secure; B = Fearful; C = Preoccupied; D = Dismissing)
46
Table 7
Bonferroni Comparison ? Attachment Style
Variable (I) Attachment
Style
(J) Attachment
Style
Mean
Difference
(I-J)
Std.
Error
Sig.
Bonferroni
Secure Fearful
Preoccupied
Dismissing
-1.59
*
-1.42
-2.13
*
.42
.58
.60
.001
.087
.003
Fearful Secure
Preoccupied
Dismissing
1.59
*
.172
-.537
.42
.60
.62
.001
1.00
1.00
Preoccupied Secure
Fearful
Dismissing
1.42
-.17
-.71
.58
.60
.73
.087
1.00
1.00
Dismissing Secure
Fearful
Preoccupied
2.13
*
.54
.71
.60
.62
.73
.003
1.00
1.00
*
The mean difference is significant at the .05 level.
Gender differences within this hypothesis were examined through the use of
ANOVA as well. Women?s mean MSI-R scores within each attachment group were
higher than men?s with the exception of the preoccupied group. Table 8 shows the
mean scores for women and men within each attachment style. There was a
difference of 2.13 between secure women?s mean MSI-R scores (6.44) and men?s
(4.31). There was a difference of 1.01 between fearful women?s mean MSI-R scores
(7.29) and men?s (6.28). There was a difference of 2.87 between dismissing women?s
mean MSI-R scores (9.04) and men?s (6.17). Preoccupied men, however, did have
slightly higher mean MSI-R scores (6.95) than women (6.38) for a difference of .57
between them. Tests of between-subjects effects (please see Table 9 for test results)
revealed that both gender (p = .001) and attachment style (p < .001) had a significant
effect upon steps taken to leave a relationship. In addition, there was a nearly
significant (p = .052) interaction effect between gender and attachment style when
47
explaining steps taken to leave a relationship. Therefore, separate ANOVAs were
conducted for each gender in order to further examine this interaction.
Table 8
Mean MSI-R Scores by Attachment Style by Gender
Variable Attachment
Style
Mean Std. Deviation N
Female Secure
Fearful
Preoccupied
Dismissing
6.44
7.29
6.38
9.04
4.20
4.30
3.48
4.20
100
106
29
27
Male Secure
Fearful
Preoccupied
Dismissing
4.31
6.28
6.95
6.17
4.15
3.82
4.04
4.25
118
72
40
35
Dependent Variable: MSI-R 
 
Table 9
Tests of Between-Subjects Effects
Source Type III
Sum of
Squares
df Mean
Square
F Sig. Partial
Eta
Squared
Corrected Model 791.259
a
7 113.04 6.68 < .001 .083
Intercept 16743.770 1 16743.77 988.73 < .001 .656
Gender 177.067 1 177.08 10.46 .001 .020
Attachment 337.243 3 112.41 6.64 < .001 .037
Gender*Attachment 131.843 3 43.95 2.50 .052 .015
a. R Squared = .083 (Adjusted R Squared = .070)
Separate ANOVAs were run to test the difference between attachment groups
on steps taken to leave a relationship by gender (please see Figure 6 for plotted MSI-
R scores of women by attachment style, Figure 7 for plotted MSI-R scores of men by
attachment style, and Figure 8 for plotted MSI-R scores of men and women by
attachment style). Post hoc testing using Bonferroni?s comparison revealed a
significant difference (p = .026) between the mean scores of secure women (6.44) and
48
dismissing women (9.04) (please see Table 10) whereas secure men (4.31) differed
significantly (p = .008) from fearful men (6.28) and from (p = .003) preoccupied men
(6.95) (please see Table 11).
Figure 6: Plot: Women?s Mean MSI-R Scores by Attachment Style
A = Secure; B = Fearful; C = Preoccupied; D = Dismissing
49
Figure 7: Plot: Men?s Mean MSI-R Scores by Attachment Style
A = Secure; B = Fearful; C = Preoccupied; D = Dismissing
50
Figure 8: Plot: Women?s and Men?s Mean MSI-R Scores by Attachment Style
A = Secure; B = Fearful; C = Preoccupied; D = Dismissing
51
Table 10
Bonferroni Comparison (Women)
Variable (I) Attachment
Style
(J) Attachment
Style
Mean
Difference
(I-J)
Std.
Error
Sig.
Bonferroni
Secure Fearful
Preoccupied
Dismissing
-.85
-.06
-2.60
*
.58
.88
.90
.861
1.00
.026
Fearful Secure
Preoccupied
Dismissing
.85
.91
-1.74
.58
.87
.90
.861
1.00
.320
Preoccupied Secure
Fearful
Dismissing
-.06
-.91
-.2.66
.88
.87
1.11
1.00
1.00
.107
Dismissing Secure
Fearful
Preoccupied
2.60
*
1.74
2.66
.90
.90
1.11
.026
.320
.107
*
The mean difference is significant at the .05 level.
Table 11
Bonferroni Comparison (Men)
Variable (I) Attachment
Style
(J) Attachment
Style
Mean
Difference
(I-J)
Std.
Error
Sig.
Bonferroni
Secure Fearful
Preoccupied
Dismissing
-1.96*
-2.64*
-1.86
.61
.74
.78
.008
.003
.109
Fearful Secure
Preoccupied
Dismissing
1.96*
-.67
.11
.61
.80
.84
.008
1.00
1.00
Preoccupied Secure
Fearful
Dismissing
2.64*
.67
.78
.74
.80
.94
.003
1.00
1.00
Dismissing Secure
Fearful
Preoccupied
1.86
-.11
-.78
.78
.84
.94
.109
1.00
1.00
*
The mean difference is significant at the .05 level.
52
Hypothesis 4: The Relationship Between Steps Taken Toward Leaving a Relationship
and Abuse will Differ Depending on an Individual?s Attachment Style
It was hypothesized that, in the presence of abuse, secure and
dismissing individuals would take steps toward leaving a relationship and fearful
individuals would take the greatest number of steps toward leaving a relationship
whereas preoccupied individuals would take the fewest steps toward leaving a
relationship among the four styles of attachment. While it was initially planned to
control for income and length of relationship, due to the findings during correlation
testing that there was no significant relationship between steps taken to leave a
relationship and income or length of relationship it was determined that it was
unnecessary to control for these factors. An ANOVA was conducted in order to
examine mean MSI-R scores of individuals as a function of abuse levels and different
attachment styles. As demonstrated by previous univariate analyses, tests of
between-subjects effects confirmed that the effects of emotional abuse (p < .001) and
attachment (p < .001) were significantly related to steps taken toward leaving a
relationship. The interaction effect of attachment and abuse was near the trend level
(p = .066). However, the interaction effect of attachment, abuse, and gender was
found to be significant (p = .029). Table 12 shows the results of the tests of between-
subjects effects.
53
Table 12
Tests of Between-Subjects Effects
Dependent Variable: MSI-R Marital Status Inventory
Source
Type III
Sum of
Squares df
Mean
Square F Sig.
Partial
Eta
Squared
Corrected Model 2364.372
a
23 102.80 7.17 < .001 .249
Intercept 16320.172 1 16320.17 1137.59 .000 .696
Attachment 274.918 3 91.64 6.39 .000 .037
Abuse 550.754 2 275.38 19.20 .000 .072
Gender 169.158 1 169.16 11.79 .001 .023
Attachment *
Abuse
170.893 6 28.48 1.99 .066 .023
Attachment *
Gender
84.634 3 28.21 1.97 .118 .012
Abuse * Gender 61.709 2 30.85 2.15 .117 .009
Attachment *
Abuse * Gender
203.501 6 33.92 2.36 .029 .028
a. R Squared = .249 (Adjusted R Squared = .214)
Based upon the presence of significance when examining the interaction
between attachment, abuse, and gender, it was determined that separate analyses
should be conducted for each gender in order to determine whether men or women
were responsible for the lack of significant interaction effect. This proved to be
important as tests of between-subjects effects for women showed a highly significant
association (p = .003) for steps taken to leave a relationship as a function of the
interaction between attachment styles and abuse levels. In contrast, for men there
was no significant relationship (p = .434) for steps taken to leave a relationship as a
function of the interaction of attachment and abuse. Furthermore, when testing the
differences in mean MSI-R scores among men based upon attachment style and level
of emotional abuse, a Levene?s test of equality of error variances resulted in a high
level of significance (p = .016), indicating that equal variances among the different
54
groups of men cannot be assumed, thus violating one of the primary assumptions
necessary for conducting analysis of variance. Tables 13 and 14 demonstrate the
results of the tests of between-subjects effects for women and men.
Table 13
Tests of Between-Subjects Effects (Women)
Dependent Variable: MSI-R Marital Status Inventory
Source
Type III
Sum of
Squares df
Mean
Square F Sig.
Partial
Eta
Squared
Corrected Model 1065.002
a
11 96.89 6.78 .000 .230
Intercept 9090.039 1 9090.04 636.54 .000 .719
Attachment 194.231 3 64.74 4.53 .004 .052
Abuse 228.776 2 114.39 9.01 .000 .060
Attachment *
Abuse
296.691 6 49.23 3.46 .003 .077
a. R Squared = .230 (Adjusted R Squared = .196)
Table 14
Tests of Between-Subjects Effects (Men)
Dependent Variable: MSI-R Marital Status Inventory
Source
Type III
Sum of
Squares df
Mean
Square F Sig.
Partial
Eta
Squared
Corrected Model 992.668
a
11 90.24 6.26 .000 .217
Intercept 7232.511 1 7232.51 501.82 .000 .669
Attachment 155.839 3 51.95 3.60 .014 .042
Abuse 400.653 2 200.33 13.90 .000 .101
Attachment *
Abuse
85.386 6 14.23 .99 .434 .023
a. R Squared = .249 (Adjusted R Squared = .214)
55
In support of the initial hypothesis, among secure women there was a positive
trend between abuse and steps taken to leave a relationship, with mean MSI-R scores
increasing as the level of abuse increased. Fearful women?s mean MSI-R scores
partially supported the initial hypothesis. Fearful women had the lowest overall mean
MSI-R score (4.81) at the low level of abuse among the four attachment types, which
does not support the initial hypothesis. However, fearful women had the highest
mean MSI-R score (10.73) at the high level of abuse among the four attachment
types. However, in contrast to this, dismissing women had the highest of all mean
MSI-R scores at the low level of abuse (11.00), whereas it was expected that fearful
individuals would have the highest mean scores at each level of abuse. Counter to
what was initially hypothesized; dismissing women had a higher mean MSI-R score
at the low level of abuse (11.00) than at the moderate (7.57) and high (8.77) levels of
abuse. The mean MSI-R scores of preoccupied women appear to only partially
support the initial hypothesis. Preoccupied women had the lowest mean MSI-R score
at the moderate level of abuse (4.82) of all the mean MSI-R scores at any level of
abuse. Preoccupied women also had the lowest mean MSI-R score at the high level
of abuse (7.88) among the four attachment types at that level of abuse. However,
preoccupied women had a higher mean score (6.90) at the low level of abuse than
both secure women (4.94) and fearful women (4.81). Table 15 demonstrates the
mean MSI-R for each attachment style at each level of abuse. Figure 9 shows the
plotted women?s mean MSI-R scores at each level of abuse by attachment style.
56
Table 15
Women?s Mean MSI-R Scores at Different Levels of Abuse by Attachment Style
Dependent Variable: MSI-R Marital Status Inventory
Attachment
Style Level of Abuse Mean
Std.
Deviation N
Low <= 21.00 4.94 4.22693 32
Mod. 22.00 - 46.00 5.57 3.69233 30
High 47.00+ 8.30 3.87918 37
Secure
Total 6.38 4.18137 99
Low <= 21.00 4.81 4.03428 36
Mod. 22.00 - 46.00 6.95 3.55867 40
High 47.00+ 10.73 3.18329 30
Fearful
Total 7.29 4.29832 106
Low <= 21.00 6.90 3.24722 10
Mod. 22.00 - 46.00 4.82 3.81623 11
High 47.00+ 7.88 2.69590 8
Preoccupied
Total 6.38 3.47865 29
Low <= 21.00 11.00 3.87298 7
Mod. 22.00 - 46.00 7.57 3.64496 7
High 47.00+ 8.77 4.53052 13
Dismissing
Total 9.04 4.20148 27
Low <= 21.00 5.61 4.31800 85
Mod. 22.00 - 46.00 6.26 3.68749 88
High 47.00+ 9.16 3.79015 88
Total
Total 7.03 4.21573 261
57
Figure 9: Plot: Women?s Mean MSI-R Scores by Attachment Style
A = Secure; B = Fearful; C = Preoccupied; D = Dismissing
A post hoc pairwise comparison of women?s mean MSI-R scores at the three
levels of abuse was conducted through the use of a Bonferroni comparison. Within
attachment styles, only the secure and fearful styles of attachment showed any
significant difference between scores at each level of abuse. At low levels of abuse,
secure women?s mean MSI-R score differed significantly (p = .019) from their mean
MSI-R score at a high level of abuse. Fearful women?s mean MSI-R scores differed
in the same fashion as secure women?s, with a significant difference (p < .001)
between mean MSI-R scores at low and high levels of abuse, but also with a
significant difference (p =.003) between mean MSI-R scores at moderate and high
58
levels of abuse as well. Table 16 demonstrates the differences in mean MSI-R scores
at each level of abuse within each attachment style.
Table 16
Bonferroni Comparison (Women)
Dependent Variable: MSI-R Marital Status Inventory
Variable (I) Attachment
Style and
Abuse Level
(J) Attachment
Style and
Abuse Level
Mean
Difference
(I-J)
Std.
Error
Sig.
Bonferroni
Secure and Low
Abuse
Secure & Mod.
Secure & High
-.63
-3.36
*
.96
.91
1.00
.019
Secure and
Moderate Abuse
Secure & Low
Secure & High
.629
-2.73
.96
.93
1.00
.236
Secure and High
Abuse
Secure & Low
Secure & Mod.
3.37
*
2.73
.91
.93
.019
.236
Fearful and Low
Abuse
Fearful & Mod.
Fearful & High
-2.14
-5.93
*
.87
.93
.936
.000
Fearful and
Moderate Abuse
Fearful & Low
Fearful & High
2.14
-3.78
*
.87
.91
.936
.003
Fearful and High
Abuse
Fearful & Low
Fearful & Mod.
5.93
*
3.78
.93
.91
.000
.003
Preoccupied and
Low Abuse
Preocc. & Mod.
Preocc. & High
2.08
-.98
1.65
.79
1.00
1.00
Preoccupied and
Moderate Abuse
Preocc. & Low
Preocc. & High
-2.08
-3.06
1.65
1.76
1.00
1.00
Preoccupied and
High Abuse
Preocc. & Low
Preocc. & Mod.
.98
3.06
1.79
1.76
1.00
1.00
Dismissing and
Low Abuse
Dismiss & Mod.
Dismiss & High
3.43
2.23
2.02
1.77
1.00
1.00
Dismissing and
Moderate Abuse
Dismiss & Low
Dismiss & High
-3.43
-1.20
2.02
1.77
1.00
1.00
Dismissing and
High Abuse
Dismiss & Low
Dismiss & Mod.
-2.23
1.20
1.77
1.77
1.00
1.00
*
The mean difference is significant at the .05 level.
Further examination of the post hoc analysis shows that at each level of abuse
mean scores do not necessarily differ significantly among the different attachment
styles. At low levels of abuse, securely attached women?s mean MSI-R scores only
differed significantly (p = .01) from women with a dismissing style of attachment.
59
The same was true of fearfully attached women?s mean MSI-R scores at the low level
of abuse. They only differed significantly (p = .006) from dismissing women.
Preoccupied women?s mean MSI-R scores at the low level of abuse did not differ
significantly from any of the other three attachment types? mean MSI-R scores. At
the moderate and high levels of abuse there were no significant differences among the
mean MSI-R scores among the four styles of attachment. Table 17 demonstrates the
Bonferroni comparisons among attachment styles at the three levels of abuse.
Table 17
Bonferroni Comparison
Dependent Variable: MSI-R Marital Status Inventory
(I) Attachment*Abuse (J) Attachment*Abuse
Mean
Difference
(I-J) Std. Error Sig.
Secure at Low Abuse Fearful at Low Abuse .13 .92 1.000
Preoccupied at Low
Abuse
-1.96 1.37 1.000
Dismissing at Low
Abuse
-6.06(*) 1.58 .010
Secure at Moderate
Abuse
Fearful at Moderate
Abuse
-1.38 .91 1.000
Preoccupied at
Moderate Abuse
.75 1.33 1.000
Dismissing at
Moderate Abuse
-2.00 1.59 1.000
Secure at High Abuse Fearful at High Abuse
-2.44 .93 .609
Fearful at High Abuse
-2.44 .93 .609
Dismissing at High
Abuse
-.47 1.22 1.000
Fearful at Low Abuse Secure at Low Abuse
-.13 .92 1.000
Preoccupied at Low
Abuse
-2.09 1.35 1.000
Dismissing at Low
Abuse
-6.19* 1.56 .006
Feaful at Moderate
Abuse
Secure at Moderate
Abuse
1.38 .91 1.000
60
Preoccupied at
Moderate Abuse
2.13 1.29 1.000
Dismissing at
Moderate Abuse
-.62 1.55 1.000
Fearful at High Abuse Secure at High Abuse 2.43 .93 .609
Preoccupied at High
Abuse
2.86 1.50 1.000
Dismissing at High
Abuse
1.96 1.25 1.000
Preoccupied at Low
Abuse
Secure at Low Abuse
1.96 1.37 1.000
Fearful at Low Abuse
2.09 1.35 1.000
Dismissing at Low
Abuse
-4.10 1.86 1.000
Preoccupied at
Moderate Abuse
Secure at Moderate
Abuse
-.75 1.33 1.000
Feaful at Moderate
Abuse
-2.13 1.29 1.000
Dismissing at
Moderate Abuse
-2.75 1.83 1.000
Preoccupied at High
Abuse
Secure at High Abuse
-.42 1.47 1.000
Fearful at High Abuse -2.86 1.50 1.000
Dismissing at High
Abuse
-.89 1.70 1.000
Dismissing at Low
Abuse
Secure at Low Abuse
6.06* 1.58 .010
Fearful at Low Abuse 6.19* 1.56 .006
Preoccupied at Low
Abuse
4.10 1.86 1.000
Dismissing at
Moderate Abuse
Secure at Moderate
Abuse
2.00 1.59 1.000
Feaful at Moderate
Abuse
.62 1.55 1.000
Preoccupied at
Moderate Abuse
2.75 1.83 1.000
Dismissing at High
Abuse
Secure at High Abuse
.47 1.22 1.000
Fearful at High Abuse
-1.96 1.25 1.000
Preoccupied at High
Abuse
.89 1.70 1.000
61
CHAPTER 4: DISCUSSION
The primary goal of this study was to examine whether or not an individual?s
attachment style moderated the impact of emotional abuse on his or her intent to leave
the relationship in which the abuse occurred. In doing so, the main effect of
emotional abuse on steps taken to leave a relationship and the main effect of
attachment style on steps taken to leave the relationship and their interaction were
examined, as well as whether gender differences existed in these effects. The
following table (Table 18) summarizes the four main hypotheses of this study as well
as whether or not the findings of the study supported these hypotheses.
62
Table 18
Hypotheses and Results
Hypothesis Supported?
1 ? There will be a positive association
between the level of experienced
emotional abuse and steps taken toward
leaving an intimate relationship.
Yes. Tests demonstrated a moderately
positive association between experienced
emotional abuse and steps taken toward
leaving a relationship.
2 ? In the presence of abuse women will
take greater steps toward leaving a
relationship than men.
Yes. Tests demonstrated that women had
higher mean MSI-R scores than men at
each level of abuse.
3 ? Depending on attachment style,
individuals will take different degrees of
steps toward leaving a relationship.
Partially. Only secure individuals
differed significantly from the other
styles of attachment insofar as secure
individuals took fewer steps toward
leaving their relationships than fearful
and dismissing individuals. Additionally,
expectations for how attachment would
specifically affect steps taken toward
leaving a relationship were not met.
4 ? Depending on attachment style,
individuals will take different degrees of
steps toward leaving a relationship at
low, moderate, and high levels of abuse.
Partially. The actual interaction effect of
attachment style and abuse was only
significant among the study?s female
participants. Scores within each
attachment style did not always differ
significantly. Scores also did not always
differ significantly between each style of
attachment at the same level of abuse and
differences were not as expected.
There Will be a Positive Association Between Experienced Emotional Abuse and
Steps Taken to Leave a Relationship
As expected, a correlation test demonstrated a moderate to moderately-strong
positive relationship between experienced emotional abuse and steps taken toward
leaving a relationship. This was further supported by the results of an analysis of
variance that demonstrated an increase of mean MSI-R scores at each higher level of
abuse. Additionally, a post hoc analysis demonstrated significant differences between
63
mean MSI-R scores at each level of abuse. At its most basic level, this study would
seem to support prior research that has demonstrated the effects of emotional abuse
on individuals? desire to leave their relationships.
In the Presence of Abuse, Women will Take Greater Steps Toward Leaving a
Relationship than Men
As hypothesized, women had higher mean MSI-R scores at each level of
abuse than men. However, upon closer examination of tests of between-subjects
effects, it became apparent that steps taken toward leaving a relationship were not
significantly affected by the interaction of gender and abuse. Ultimately, women
were not significantly more likely to leave an abusive relationship than men.
Depending on Attachment Style, Individuals will Take Different Steps Toward
Leaving a Relationship
This hypothesis was only partially supported by this study?s findings.
Foremost, it was expected that preoccupied individuals would demonstrate the lowest
mean MSI-R scores among all four of the attachment groups. This did not prove to
be the case as members of the secure group had significantly lower mean MSI-R
scores than members of the fearful and dismissing groups whereas the preoccupied
group not only scored higher than the secure group (though not significantly so), but
it was not significantly lower than the mean scores of the fearful and dismissing
groups. There are a few possible explanations for this result. The first is that
securely attached individuals may be more likely to have higher levels of social
support available from friends and family. This would make sense when one
considers that attachment does not only affect romantic relationships, but all intimate
64
relationships. As such, a secure individual would likely have more secure
relationships to rely on when facing relationship distress. With greater outside
support secure individuals may be able to take a more measured pace toward exiting a
relationship. Furthermore, this sample is drawn from couples who have entered
therapy with the goal of improving their relationships. Based on the theory that
secure individuals are able to evaluate their relationships from a more rational
standpoint than otherwise attached individuals, another possible explanation is that
secure individuals may be waiting to determine whether therapy has been effective
before moving toward ending a relationship.
The size of the sample may also have affected the results in testing this
hypothesis. Whereas there were 218 secure individuals included in this study, there
were only 69 preoccupied individuals. Prior to running frequency tests on the
sample, it was initially expected that there would be greater numbers of secure and
preoccupied individuals within the sample. This did not turn out to be the case.
Initial expectations were correct that the greatest number of participants identified
themselves as secure. This would seem to make sense in light of the assumption that
secure individuals are able to take a more rational approach toward their relationship
difficulties, and as such, would likely consider therapy a viable option for working on
their relationship problems. Perhaps the lack of preoccupied individuals could be
explained as the result of these individuals easily feeling threatened in their
relationships. As such, they may consider entering therapy as an admission that
something is wrong with their relationships that could lead to those relationships
ending.
65
Along with the low number of preoccupied participants, the high numbers of
fearful participants were also unexpected. Prior to examining the group frequencies
within the study it was assumed that there would be few fearful individuals as they
would most likely seize upon any relationship problem as a reason to leave. This was
due to fearful individuals? belief that they will likely be hurt in an intimate
relationship, thus when this fear is proven true, they would exit the relationship rather
than work to improve it. However, after secure individuals, the next highest number
of individuals (178) identified themselves as fearful. This lead to having to
reconsider what the fearful style of attachment meant. Perhaps the high number of
fearful individuals could be explained by their fear of being hurt. As such, they may
be willing to turn to therapy as a means of potentially preventing any further pain that
would result from intimacy figuring if they do not improve on the current relationship
they are just as likely to be hurt in any subsequent relationship.
The low number of dismissing participants was within expectations. This was
due to their identified independent nature. It was expected that due to this, dismissing
individuals would not likely pursue couples therapy on their own. Additionally,
dismissing individuals would probably not consent to entering therapy either, even if
suggested by his or her significant other.
Individuals with Different Attachment Styles will Take Different Steps
Toward Leaving a Relationship in the Presence of Emotional Abuse
This hypothesis was only partially supported. Based on tests of between-
subjects effects, among men the interaction of attachment style and abuse was not
66
demonstrated as significantly affecting the steps they take toward leaving a
relationship. Each of these variables on its own was demonstrated to have a
significant effect upon steps men take toward leaving a relationship. As such it could
possibly be concluded that while men?s attachment style does significantly affect how
they determine to leave their relationship, and while this is also true of emotional
abuse, among men, attachment style does not affect how they perceive emotional
abuse when considering whether or not to leave a relationship. At this point it can
only be concluded that if something does moderate the impact of emotional abuse
upon the steps taken to leave a relationship, it is not attachment style.
Among women, however, the interaction of attachment and abuse was
significant with regards to their steps taken toward leaving a relationship. The more
specific findings, however, only partially support the initial hypothesis. As abuse
increased, secure women?s mean MSI-R scores did indeed increase, which was
initially hypothesized. However, secure women?s scores did not increase
significantly between the low and moderate as well as the moderate and high levels of
abuse. Only between low and high levels of abuse were women?s mean MSI-R
scores significantly different. Fearful women?s scores also increased as abuse
increased, and their mean MSI-R scores at low and moderate levels of abuse both
differed significantly from their mean MSI-R scores at a high level of emotional
abuse. There was not, however, a significant difference between their mean MSI-R
scores at low and moderate levels of abuse. From this, we may be able to conclude
that among secure and fearful women, the steps taken toward leaving a relationship
do not actually begin in earnest until emotional abuse has reached a high level.
67
Among preoccupied and dismissing women there was no significant
difference among MSI-R scores at each level of abuse. Interestingly though, among
preoccupied women the mean scores actually decreased between low and moderate
levels of abuse before peaking at the high level of abuse. Similar to preoccupied
women, dismissing women?s mean MSI-R scores decreased between the low and
moderate levels of abuse and then increased between the moderate and high levels of
abuse. In contrast to all the groups, however, dismissing women?s highest mean
MSI-R score was at the low level of abuse. It is difficult, however, to consider these
results significant at all when one considers that in the study sample only 29 women
identified themselves as preoccupied and 27 women identified themselves as
dismissing. With such small samples of preoccupied and dismissing within the study
it can only be concluded that results are inconclusive.
At each level of abuse, there were few differences among the mean MSI-R
scores for each attachment type. Securely attached women?s mean MSI-R scores
only differed significantly from dismissing women at a low level of abuse. This may
be explained by secure individuals? comfort within relationships versus dismissing
individuals? insistence upon independence. This also may be the result of a small
sample of dismissing women being tested. At a low level of abuse, fearful women
also differed from dismissing women, but not from the other two attachment types.
This may be the result of fearful individuals not identifying strongly with a need for
independence as dismissing individuals do. This also may be the result of a small
sample of dismissing women.
68
Preoccupied women did not differ from any other attachment types at low
levels of abuse. In fact, they did not differ from the other attachment styles at any
level of abuse. This is difficult to explain in light of preoccupied individuals?
identified reliance on intimate relationships as well as their link between intimate
relationships and sense of self. Based on this it was hypothesized that not only would
preoccupied individuals have significantly different MSI-R scores at each level of
abuse, but that they would be significantly lower. Results of this study demonstrate
this is not the case. One possible conclusion is that with regards to taking steps
toward leaving a relationship, preoccupied individuals do not experience emotional
abuse any differently than individuals with other styles of attachment. Another
possible explanation is that whereas the MSI-R tests steps taken toward leaving an
intimate relationship, this study does not have information on how many of these
individuals ultimately left their intimate relationships, and furthermore, whether or
not they returned to these relationships later on. These results also may be due to the
small sample of preoccupied women within the overall study sample.
As mentioned earlier, dismissing women?s mean MSI-R scores differed
significantly from the mean MSI-R scores of secure and women at the low abuse
level. Otherwise, there were no significant differences at the other levels of abuse or
between each level of abuse within the dismissing group. Although it was not
deemed significantly different through the data analysis process, it was interesting
that of the three groups only dismissing women had the highest mean MSI-R score at
the low level of abuse as opposed to at the high level of abuse. Among the four
attachment types, dismissing women also had the highest overall score at the
69
moderate level of abuse. Again, this was not deemed a statistically significant
difference by the data analysis. That aside however, it would seem to make sense that
dismissing women would score higher in general at each level of abuse due to the
fiercely independent nature of individuals identified as dismissing. As such, it would
fit that they would be less tolerant of emotionally abusive behaviors which often
target an individual?s sense of worth and independence. It is important to again note
that the sample size of dismissing women was quite small and as such this may have
also profoundly impacted how reliable these findings can be considered to be.
Limitations of Current Study
There were several limitations to the current study that must be considered
when evaluating its findings. The actual sample must be considered as a limitation
for two reasons. First, this study aimed to examine the differences between
individuals who identify with different types of attachment styles. However, within
the actual sample, there were significantly fewer preoccupied and dismissing
individuals in comparison to the number of secure and fearful participants.
Additionally, the attachment styles were distributed differently by each gender as
well. This could have possibly led to incorrect conclusions regarding how individuals
with different attachment types differ from each other in the face of emotional abuse
with regards to leaving a relationship as well as the conclusions regarding gender.
Second, this sample was drawn from individuals who have sought out couple therapy.
As a result, it makes sense to conclude that they will on a whole have taken fewer
steps toward leaving their relationships as they are entering therapy in a likely attempt
70
to improve upon or save their relationships. As a result, overall MSI-R scores from
this sample may not necessarily generalize to the population.
The MSI-R measures steps taken toward leaving a relationship but does not
necessarily reflect whether or not it has been completely terminated (only a score of
18 can mean certain termination; however, some earlier steps may not be
acknowledged while still having filed for divorce, thus a low score could be attained
from an individual who has ended his or her relationship). As such, it only measures
steps taken toward leaving a relationship and not whether or not a relationship has
absolutely ended. Having follow up information about which individuals did end
their relationships would likely provide greater insight into how attachment may
moderate emotional abuse?s impact on whether a relationship ends.
Finally, this study did not include how strongly an individual identifies with
his or her style of attachment. The RQ includes a scale on which the user rates how
well he or she is identified by each paragraph. As such, an individual who strongly
identifies with the preoccupied style of attachment may demonstrate greater
differences from other strongly identified styles of attachment. This study did not
allow for the nuances inherent to attachment style as well as how it can be measured.
Future studies may address this by dividing the sample by not only attachment style
but how strongly an individual identifies with an attachment style as well.
Recommendations for Future Research
Based on the findings and limitations of the current study there are several
recommendations for future research projects. First, it would be of interest to conduct
71
this study again with a sample that included sufficiently large numbers of individuals
who identify with each style of attachment. Furthermore, this study should be
conducted with a sample that has not volunteered for therapy and which may be more
representative of the general population. This would correct two of the major
limitations of the current study.
To build upon the current study further, it would also be of interest to conduct
a longitudinal study in order to determine in the long term whether individuals with
different attachment styles are more or less likely to end an intimate relationship, and
under what conditions the relationships ended. This would be of interest with regards
to better understanding intimate relationships in general as well as what factors are
important with regards to why intimate relationships end.
This study demonstrated that attachment is significantly associated with steps
taken toward leaving a relationship among men, but not how it was significant.
Future research should attempt to address how attachment is significant among men
as well as whether it is different from how it is significant among women. This could
likely be best addressed through some form of qualitative study as this does not
appear to be something that can be captured by quantitative measures.
It is also important that further research on abuse focus on men. As
demonstrated by this study, emotional abuse plays a significant role in the steps taken
by men toward leaving an intimate relationship; however, research on abused men
remains far behind research on abused women. It will be important for future
researchers to address this.
72
The final recommendation would likely be mostly of interest to clinicians. It
would be interesting to examine the attachment styles of both members of couples in
order to determine whether one style of individual is more or less likely to enter into a
relationship with another type of attached individual. This could then be carried
further in order to investigate dyadic satisfaction, prevalence of abuse, or any number
of other factors. Armed with this information, clinicians may be able to better predict
the behaviors of the couples they work with and tailor clinical interventions
specifically for them based on this information.
Clinical Implications
The first clinical implication that can be drawn from this study is knowledge
about the lack of preoccupied and dismissing individuals who are engaging in couple
therapy. It remains unknown as to whether on a whole there are fewer such
individuals among the entire population. It is important to consider this in light of the
fact that these types of attachment do not necessarily lend themselves to healthy
relationship functioning. At this juncture, it may be that these individuals are less
likely to trust the process of therapy. If this is true, then clinicians must consider how
to reach out to these individuals, as they cannot be helped if they do not engage in
therapy. Furthermore, clinicians must consider that if such individuals do engage in
therapy, they may not demonstrate much patience with therapy?s often lengthy nature.
As such, these clients may be best served by the clinician making great efforts to
specifically engage them early in therapy or by the clinician engaging in brief therapy
strategies in order to provide these individuals with positive results early in the
therapeutic process until they feel thoroughly invested.
73
Second, based on this study?s findings, it is important for clinicians to note
that attachment style appears to apply differently to men and women. At least on a
basic level it can be concluded that attachment plays a role in how women perceive
emotional abuse within their intimate relationships whereas this is not the case with
men. However, attachment plays some role among men as they determine whether or
not to leave their relationships. It remains unclear at this juncture how men?s
attachment style affects their relationships as well as what else it may affect among
women. This speaks to not assuming that client behaviors can be easily predicted by
their style of attachment. This is especially true in the case of abused clients.
Although it would be easy to assume (as this study hypothesized) that certain types of
attached individuals would be more likely to accept being abused by an intimate
partner, that does not appear to be the case based on the study?s conclusions.
Conclusion
In conclusion, this study demonstrated that among men and women, emotional
abuse plays a crucial role in moving toward leaving an intimate relationship.
Furthermore, although women had higher mean scores on a measure of leaving a
relationship, abuse was demonstrated as being a more important factor than gender in
taking steps toward leaving an intimate relationship. This study also demonstrated
that individual attachment style has a significant effect upon the steps an individual
takes toward leaving a relationship. It also raised a number of important questions
regarding what role attachment style plays in the decisions individuals make
regarding leaving a relationship in which emotional abuse is present. Ideally future
researchers will further examine this important aspect of how individuals experience
74
intimate relationships.
75
APPENDICES: MEASURES
Appendix A
MDEAS (ASSESSMENT)
__________________________________________________________________________________
No matter how well a couple gets along, there are times when they disagree, get annoyed with the
other person, want different things from each other, or just have spats or fights because they are in a
bad mood, are tired, or for some other reason. Couples also have many different ways of trying to
settle their differences. This is a list of things that might happen when you have differences. Please
circle how many times you did each of these things IN THE PAST 4 MONTHS, and how many times
your partner did them in the IN THE PAST 4 MONTHS. If you or your partner did not do one of
these things in the past 4 months, but it happened before that, circle 0.
(0) Not in the past four months, but it did happen before (3) 3-5 times (6) More than 20 times
(1) Once (4) 6-10 times (9) This has never happened
(2) Twice (5) 11-20 times
How Often in the last 4 months?
1. Asked the other person where s/he had been or who s/he
was with in a suspicious manner.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
2. Secretly searched through the other person?s belongings. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
3. Tried to stop the other person from seeing certain friends
or family members.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
4. Complained that the other person spends too much time
with friends.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
5. Got angry because the other person went somewhere
without telling him/her.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
6. Tried to make the other person feel guilty for not
spending enough time together.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
7. Checked up on the other person by asking friends where
s/he was or who s/he was with.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
8. Said or implied that the other person was stupid. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
9. Called the other person worthless. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
10. Called the other person ugly. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
11. Criticized the other person?s appearance. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
12. Called the other person a loser, failure, or similar term. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
76
How Often in the last 4 months?
13. Belittled the other person in front of other people. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
14. Said that someone else would be a better girlfriend or
boyfriend.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
15. Became so angry that s/he was unable or unwilling
To talk.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
16. Acted cold or distant when angry. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
17. Refused to have any discussion of a problem. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
18. Changed the subject on purpose when the other
person was trying to discuss a problem.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
19. Refused to acknowledge a problem that the
other felt was important.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
20. Sulked or refused to talk about an issue. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
21. Intentionally avoided the other person during a
conflict or disagreement.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
22. Became angry enough to frighten the other person. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
23. Put her/his face right in front of the other person?s
face to make a point more forcefully.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
24. Threatened to hit the other person. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
25. Threaten to throw something at the other person. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
26. Threw, smashed, hit, or kicked something in
front of the other person.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
27. Drove recklessly to frighten the other person. You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
28. Stood or hovered over the other person during
a conflict or disagreement.
You: 0 1 2 3 4 5 6
Your partner: 0 1 2 3 4 5 6
9
9
77
Appendix B
RQ (Assessment)
1. The following are descriptions of four general relationship styles that people often
report. Please circle the letter corresponding to the style that best describes you or
is closest to the way you are in your relationships with PEOPLE IN GENERAL.
A. It is relatively easy for me to be emotionally close to others. I am comfortable
depending on others and having others depend on me. I don?t worry about
being alone or having others not accept me.
B. I am somewhat uncomfortable getting close to others. I want emotionally
close relationships, but I find it difficult to trust others completely, or to
depend on them. I sometimes worry that I will be hurt if I allow myself to
become too close to others.
C. I want to be completely emotionally intimate with others, but I often find that
others are reluctant to get as close as I would like. I am uncomfortable being
without close relationships, and I sometimes worry that others don?t value me
as I value them.
D. I am comfortable without close relationships. It is very important to me to
feel independent and self-sufficient, and I prefer not to depend on others or
have others depend on me.
2. Now please rate each of the relationship styles above according to the extent to
which you think each description corresponds to your general relationship styles.
Very Much Somewhat Not at All
like me like me like me
Style A. 1 2 3 4 5 6 7
Style B. 1 2 3 4 5 6 7
Style C. 1 2 3 4 5 6 7
Style D. 1 2 3 4 5 6 7
78
Appendix C
MSI-R (assessment)
We would like to get an idea of how your relationship stands right now. Within
the past four months have you?
Yes __ No__ 1. Had frequent thoughts about separating from your partner, as much
as once a week or so.
Yes __ No__ 2. Occasionally thought about separation or divorce, usually after an
argument.
Yes __ No__ 3. Thought specifically about separation, for example how to divide
belongings, where to live, or who would get the children.
Yes __ No__ 4. Seriously thought about the costs and benefits of ending the
relationship.
Yes __ No__ 5. Considered a divorce or separation a few times other than during or
shortly after a fight, but only in general terms.
Yes __ No__ 6. Made specific plans to discuss separation with your partner, for
example what you would say.
Yes __ No__ 7. Discussed separation (or divorce) with someone other than your
partner (trusted friend, minister, counselor, relative).
Yes __ No__ 8. Discussed plans for moving out with friends or relatives.
Yes __ No__ 9. As a preparation for living on your own, set up an independent bank
account in your own name to protect your interest.
Yes __ No__ 10. Suggested to your partner that you wish to have a separation.
Yes __ No__ 11. Discussed separation (or divorce) seriously with your partner.
Yes __ No__ 12. Your partner moved furniture or belongings to another residence.
Yes __ No__ 13. Consulted an attorney about legal separation, a stay away order, or
divorce.
Yes __ No__ 14. Separated from your partner with plans to end the relationship.
Yes __ No__ 15. Separated from your partner, but with plans to get back together.
Yes __ No__ 16. File for a legal separation.
Yes __ No__ 17. Reached final decision on child custody, visitation, and division of
property.
Yes __ No__ 18. Filed for divorce or ended the relationship.
79
REFERENCES
Ainsworth, M.D., Blehar, M.C., Waters, E., & Wall, S. (1978). Patterns of
attachment: A psychological study of the strange situation. Hillsdale:
Lawrence Erlbaum.
Amato, P.R., & Previti, D. (2003). People?s reasons for divorcing: Gender, social
class, the life course, and adjustment. Journal of Family Issues, 24, 602-626.
Amato, P.R., & Togers, S.J. (1997). A longitudinal study of marital problems and
subsequent divorce. Journal of Marriage and the Family, 59, 612-624.
Anderson, S.M., Boulette, T.R., & Schwartz, A.H. (1991). Psychological
maltreatment of spouses. In R.T. Ammerman & M. Hersen (Eds.), Case
studies in family violence. New York: Plenium Press.
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.
Booth, A., Johnson, D.R., White, L.K., & Edwards, J.N. (1987). Divorce and marital
instability over the life course. Journal of Family Issues, 7, 421-442.
Bowlby, J. (1969). Attachment and loss: Vol.1. Attachment. New York: Basic Books.
Bowlby, J. (1973). Attachment and loss: Vol.2. Separation. New York: Basic Books.
Bramlett, M.D., & Mosher, W.D. (2002). Cohabitation, marriage, divorce, and
remarriage in the United States. National Center for Health Statistics, Vital
Health Statistics, 23, (2), 1-93.
80
Burns, A. (1984). Perceived causes of marriage breakdown and conditions of life.
Journal of Marriage and the Family, 46, 551-560.
Byrne, C.A., & Arias, I. (2004). Predicting women?s intentions to leave abusive
relationships: An application of the theory of planned behavior. Journal of
Applied Social Psychology, 34, 2586-2601.
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, 88, 510-531.
Cavanagh, M.E. (1996). Why do people remain in abusive relationships?. Pastoral
Psychology, 44, 285-293.
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, 32, 201-219.
Davis, D., Shaver, P.R., & Vernon, M.L. (2003). Physical, emotional, and behavioral
reactions to breaking up: The roles of gender, age, emotional involvement,
and attachment style. Personality and Social Psychology Bulletin, 29, 871-
884.
Dutton, D.G., Saunders, K., Starzomski, A., & Bartholomew, K. (1994). Intimacy-
anger and insecure attachment as precursors of abuse in intimate relationships.
Journal of Applied Social Psychology, 24, 1367-1386.
Epstein, N.B., & Werlinich, C.A. (2001). Marital status inventory-revised.
Department of Family Studies, University of Maryland College Park:
Unpublished scale.
81
Fincham, F.D. (2001). Attributions in close relationships: From balkanization to
integration. In G.J.O. Fletcher & M.S. Clark (Eds.), Blackwell handbook of
social psychology: Interpersonal processes (pp. 3-31). Oxford, England:
Blackwell.
Fishtein, J., Pietromonaco, P.R., & Barrett, L.F. (1999). The contribution of
attachment style and relationship conflict to the complexity of relationship
knowledge. Social Cognition, 17, 228-244.
Follingstand, D.R., Helff, C.M., Binford, R.V., Runge, M.M., & White, J.D. (2004).
Lay persons? versus psychologists judgments of psychologically aggressive
actions by a husband and wife. Journal of Interpersonal Violence, 19, 916-
942.
Follingstand, D.R., Rutledge, L.L., Berg, B.J., Hause, E.S., & Polek, D.S. (1990). The
role of emotional abuse in physically abusive relationships. Journal of Family
Violence, 5, 107-120.
Gager, C.T., & Sanchez, L. (2003). Two as one? Couples? perceptions of time spent
together, marital quality, and risk of divorce. Journal of Family Issues, 24, 21-
50.
Gallo, L.C., & Smith, T.W. (2001). Attachment style in marriage: Adjustment and
responses to interaction. Journal of Social and Personal Relationships, 18,
263-289.
Grau, I., & Doll, J. (2003). Effects of attachment styles on the experience of equity in
heterosexual couples relationships. Experimental Psychology, 50, 298-310.
82
Guerrero, L.K. (1998). Attachment-style differences in the experience and expression
of romantic jealousy. Personal Relationships, 5, 273-291.
Hazan, C., & Shaver, S. (1987). Romantic love conceptualized as an attachment
process. Journal of Personality and Social Psychology, 52, 511-524.
Henderson, A.J.Z., Bartholomew, K., & Dutton, D.G. (1997). He loves me; He loves
me not: Attachment and separation resolution of abused women. Journal of
Family Violence, 12, 169-191.
Herbert, T.B., Silver, R.C., & Ellard, J.H. (1991). Coping with an abusive
relationship: How and why do women stay?. Journal of Marriage and the
Family, 53, 311-325.
Hines, D.A., & Malley-Morrison, K. (2001). Psychological effects of partner abuse
against men: A neglected research area. Psychology of Men & Masculinity, 2,
75-85.
Hollist, C.S., & Miller, R.B. (2005). Perceptions of attachment style and marital
quality in midlife marriage. Family Relations, 54, 46-57.
Ingoldsby, B.B., Smith, S.R., & Miller, J.E. (2004). Exploring family theories
(pp.81-102). Los Angeles, CA: Roxbury Publishing Company.
Kincaid, S.B., & Caldwell, R.A. (1995). Marital separation: Causes, coping, and
consequences. Journal of Divorce & Remarriage, 22, 109-128.
Kitson, G.C., & Sussman, M.B. (1982). Marital complaints, demographic
characteristics, and symptoms of mental distress in divorce. Journal of
Marriage and the Family, 44, 87-101.
83
Lewis, C.S., Griffing, S., Chu, M., Jospitre, T., Sage, R.E., Madry, L., & Primm, B.J.
(2006). Coping and violence exposure as predictors of psychological
functioning in domestic violence survivors. Violence Against Women, 12, 340-
354.
Martin, A.J., Berenson, K.R., Griffing, S., Sage, R.E., Madry, L., Bingham, L.E., &
Primm, B.J. (2000). The process of leaving an abusive relationship: The role
of risk assessments and decision-certainty. Journal of Family Violence, 15,
109-122.
Murphy, C.M., & Hoover, S.A. (2001). Measuring emotional abuse in dating
relationships as a multifactorial construct. In D.K. O?Leary & R.D. Maiuro
(Eds.), Psychological abuse in violent domestic relations. New York:
Springer.
Murphy, C.M., Hoover, S.A., & Taft, C.T. (1999). Multi-dimensional Emotional
Abuse Scale.
Murphy, C.M., & O?Leary, K.D. (1989). Psychological aggression predicts physical
aggression in early marriage. Journal of Consulting and Clinical Psychology,
57, 579-582.
Newton, T.L., & Kiecolt-Glaser, J.K. (1995). Hostility and erosion of marital quality
during early marriage. Journal of Behavioral Medicine, 18, 601-619.
O?Hagan, K.P. (1995). Emotional and psychological abuse: Problems of definition.
Child Abuse & Neglect, 4, 449-461.
84
O?Hearn, R.E., & Davis, K.E. (1997). Women?s experience of giving and receiving
emotional abuse: An attachment perspective. Journal of Interpersonal
Violence, 12, 375-391.
O?Leary, K.D. (1999). Psychological abuse: A variable deserving critical attention in
domestic violence. Violence and Victims, 14, 3-23.
Rogge, R.D., & Bradbury, T.N. (1999). Till violence does us part: The differing roles
of communication and aggression in predicting adverse marital outcomes.
Journal of Consulting and Clinical Psychology, 67, 340-351.
Rogge, R.D., Bradbury, T.N., Hahlweg, K., Engl, J., & Thurmaier, F. (2006).
Predicting marital distress and dissolution: Refining the two-factor hypothesis.
Journal of Family Psychology, 20, 156-159.
Rokach, R., Cohen, O., & Dreman, S. (2004). Triggers and fuses in late divorce: The
role of short term crises vs. ongoing frustration on marital break-up. Journal
of Divorce & Remarriage, 40, (3/4), 41-60.
Rosen, K.H., & Stith, S.M. (1997). Surviving abusive dating relationships: Processes
of leaving, healing, and moving on. In G.K. Kantor & J.L. Jasinski (Eds.), Out
of the darkness: Contemporary perspectives on family violence. Thousand
Oaks: Sage Publications.
Russell, R.J.H., & Hulson, B. (1992). Physical and psychological abuse of
heterosexual partners. Personality and Individual Differences, 13, 457-473.
Sanchez, L., & Gager, C.T. (2000). Hard living, perceived entitlement to a great
marriage, and marital dissolution. Journal of Marriage and the Family, 62,
708-722.
85
South, S.J., & Spitze, G. (1986). Determinants of divorce over the marital life course.
American Sociological Review, 51, 583-590.
Straight, E.S., Harper, F.W.K., & Arias, I. (2003). The impact of partner
psychological abuse on health behaviors and health status in college women.
Journal of Interpersonal Violence, 18, 1035-1054.
Tolman, R.M. (1992). Psychological abuse of women. In R.T. Ammerman & M.
Hersen (Eds.), Assessment of family violence. New York: Wiley Interscience.
Ulrich, Y.C. (1991). Women?s reasons for leaving abusive spouses. Health Care for
Women International, 12, 465-473.
Wampler, K.S., Riggs, B., & Kimball, T.G. (2004). Observing attachment behavior in
couples: The adult attachment behavior q-set (AABQ). Family Process, 43,
315-335.
Weiss, R.L., & Cerrato, M.C. (1980). The Marital Status Inventory: Development of a
measure of dissolution potential. American Journal of Family Therapy, 8, (2),
80-85.