\Wl

§

  

WW”WINIWUMMWW|\\l‘W|H\\\

——|
p
N

166

 

LIBRARY
Michigan State

llni'. In “0:4":
VI II'U'UI‘]

 

 

 

This is to certify that the
dissertation entitled

A LONGITUDINAL INVESTIGATION OF ECOLOGICAL FACTORS
RELATING TO THE MENTAL HEALTH AND WELL-BEING OF
SURVIVORS OF INTIMATE PARTNER VIOLENCE

presented by

Marisa Lynn Beeble

has been accepted towards fulfillment
of the requirements for the

PhD. degree in Psychology

4w ;.-§Jz ,9

Major Professor’s Signature
1’0 ' 7 ' 0 7'

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

 

 

. -.-.-.-o-.- u

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProj/Aoc&Pres/CIRCIDateDue.indd

 

 

A LONGITUDINAL INVESTIGATION OF ECOLOGICAL FACTORS
RELATING TO THE MENTAL HEALTH AND WELL-BEING OF
SURVIVORS OF INTIMATE PARTNER VIOLENCE
By

Marisa Lynn Beeble

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSPHY
Psychology

2009

ABSTRACT
A LONGITUDINAL INVESTIGATION OF ECOLOGICAL FACTORS
RELATING TO THE MENTAL HEALTH AND WELL-BEIN G OF
SURVIVORS OF INTIMATE PARTNER VIOLENCE
By
Marisa Lynn Beeble

Intimate partner violence (IPV) is a pervasive social problem impacting the
physical and psychological well-being of millions of women in the United States each
year. Much of the extant literature draws our attention to the devastating effects of IPV
on survivors’ mental health, but often overlooks how broader ecological factors may help
to further explain survivors’ psychological well-being. This dissertation comprises two
related studies that explored ecological factors as additional predictors of survivors’
psychological well-being over time. Study I examined the impact of resource constraints
on the mental health and well-being of survivors of IPV, specifically testing whether
difficulty obtaining resources is one mechanism that partially explains the relationship
between IPV and women’s depression and quality of life. The second study examined
how disadvantage within women’s neighborhood contexts may contribute to their
compromised well-being, in addition to the abuse they have experienced. Across these
two studies, both between-women and within-woman effects were examined.

Study 1, entitled Understanding the Impact of Resource Constraints on the
Mental Health and Well-Being of Survivors of Intimate Partner Violence Over Time,
revealed that resource constraints serve as one pathway by which abuse affects survivors’
psychological well-being. Although within-woman changes in resource constraints were

not found to mediate the relationship between change in physical abuse and change in

quality of life or depression, change in women’s difficulties obtaining resources fully
explained the relationship between change in psychological abuse and change in
depression and quality of life over time. That is, survivors’ resource constraints were
fully responsible for the significant mental health consequences women experienced
following psychological abuse. Between-women differences in initial difficulties women
reported obtaining community-based resources were also examined; however, they were
not found to mediate the relationship between women’s earlier experiences of physical or
psychological abuse and change in depression or quality of life over time.

In study 2, An Examination of the Impact of Neighborhood Factors on the Well-
Being of Survivors of Intimate Partner Violence Over Time, neighborhood disorder and
related feelings of fear of victimization had a significant impact on survivors’
psychological well-being, over and above the abuse women experienced. Although
between-women effects of neighborhood disorder and fear of victimization were
unrelated to changes in women’s depression or quality of life over time, significant
within-woman effects were detected. Specifically, change in women’s perceptions of
neighborhood disorder was negatively associated with change in quality of life over time,
and the relationship between survivors’ perceptions of neighborhood disorder and quality
of life was fully mediated by neighborhood-related fear of victimization. Interestingly,
however, no direct relationship between change in neighborhood disorder and survivors’
depression was detected. Instead, perceived neighborhood disorder was found to
indirectly relate to depression, through survivors’ feelings of fear of victimization.

Several implications for future research and practice from these studies are discussed.

For my mom and dad, and in memory of my friend, Kristin Blumenstock, who taught me

the importance of leaving this world a better place than I found it.

iv

ACKNOWLEDGMENTS

This dissertation reflects the efforts of many to whom I am grateful. First and
foremost I would like to thank my committee. A special acknowledgment goes to my
chair, Dr. Cris Sullivan for her continued guidance and support, her many words of
encouragement, and for the numerous opportunities she afforded me that helped me to
acquire the requisite knowledge and skills to successfully complete this dissertation. I am
also grateful to Dr. Deb Bybee for her commitment to my success, her patience,
generosity, and kindness, and most of all for sharing her knowledge, experience, and
words of wisdom countless times over. I would also like to thank Dr. Zaje Harrell for her
thoughtful feedback, the support and encouragement she provided, and for the time she
committed to seeing me through this process. I am also grateful to Dr. Sheryl Pimlott
Kubiak for her continued dedication to my success, the mentorship she provided to me
over the years, and for helping me to stay connected with the research I love most.

This dissertation would have never been possible without the emotional support
and encouragement provided to me by my dissertation support group, Dr. Adrienne
Adams, (soon to be Dr.) Lauren Lichty, and Dr. Carolyn McAllister. Together, we
struggled, we laughed, we commiserated, and through it all, we thrived. I will always
remember and cherish our times together.

To my friends and family, thank you for your patience and for always believing in
me; you gave me the strength and courage to pursue my dreams. A very special thanks
goes to my mother and father who, through their encouragement, love, and support taught

me to believe in myself and my abilities. I also thank my closest friends and my long-

distance supporters, Amy Galarneau, Michelle Gorglione, Stephanie Hoffman-
O’Connell, Bob Norton, Kelly Quackenbush—Cochran, Mujeer Qureshi, Jessica Ruddy,
and Tenicia (T.) Trask for being there for me every step of the way. I also would like to
express my gratitude to my ‘Eco’ colleagues and friends (Giannina Cabral, Charlie
Collins, Jason Fomey, Megan Greeson, Tiffeny Jimenez, Kevin Khamarko, Dr. Sheila
Lahousse, Kristen Law, Mercedes Morales-Aleman, Steve Pierce, and Nancy Sheldon)
for their support, and for making my extended East Lansing vacation most enjoyable. A
special thanks goes to all of the ‘Eco’ faculty, especially Dr. Rebecca Campbell, for
being a source of support to me during my graduate student career. Finally, I owe much
gratitude to Drs. Patricia O’Connor and Marion Terenzio, who through their guidance

and support, inspired me to pursue my Ph.D.

vi

TABLE OF CONTENTS

LIST OF TABLES .............................................................................
LIST OF FIGURES ...........................................................................

CHAPTER 1
Introduction .....................................................................................

CHAPTER 2
Study 1: Understanding the Impact of Resource Constraints on the Mental Health
and Well-Being of Survivors of Intimate Partner Violence Over Time ...............
Overview of the Literature ...............................................................
The Impact of Intimate Partner Violence on Survivors’ Mental Health and
Well-Being ...........................................................................
Resource Constraints in the Lives of Women with Abusive Partners ........
The Current Study .................................................................... ' ......
Method .......................................................................................
Participants ...........................................................................
Procedure .............................................................................
The Advocacy Intervention .................................................
Interviews and Interviewer Training .......................................
Retention Strategies ..........................................................
Participant Demographics .........................................................
Measures .............................................................................
Control Variables ............................................................
Psychological Abuse .........................................................
Physical Abuse ...............................................................
Depression .....................................................................
Quality of Life ................................................................
Resource Constraints .........................................................
Data Analysis ........................................................................
Results .......................................................................................
Predicting Depression ...............................................................
The Unconditional Model ...................................................
Control Variables Model ....................................................
Abuse Model ..................................................................
Difficulty Obtaining Resources Model ....................................
Testing Difficulty Obtaining Resources as 3 Mediator ..................
Predicting Quality of Life ..........................................................
The Unconditional Model ...................................................
Control Variables Model ....................................................
Abuse Model ..................................................................
Difficulty Obtaining Resources Model ....................................

vii

ix

X

1

6
6

7

14
19
23
23
24
25
25
27
28
29
29
30
31
31
32
32
33
36
39
39
39
39

4o
40

41
41
44
44
45

Testing Difficulty Obtaining Resources as a Mediator .................. 45

Discussion ................................................................................... 48
CHAPTER 3
Study 2: An Examination of the Impact of Neighborhood Factors on the Well-
Being of Survivors of Intimate Partner Violence Over Time ........................... 57
Overview of the Literature ............................................................... 57
Intimate Partner Violence and Psychological Well-Being ..................... 58
Intimate Partner Violence within Disadvantaged Neighborhoods ............ 60
Neighborhood Disorder and Psychological Well-Being ....................... 61
Cumulative Adversity: Women’s Experiences of Intimate Partner
Violence and Neighborhood Disorder ............................................. 67
The Current Study ......................................................................... 70
Method ....................................................................................... 75
Participants ........................................................................... 75
Procedure .............................................................................. 75
Participant Demographics .......................................................... 76
Measures ............................................................................... 77
Control Variables ............................................................. 77
Psychological Abuse ......................................................... 78
Physical Abuse ................................................................ 79
Depression ...................................................................... 79
Quality of Life ................................................................. 80
Ecological Context ............................................................ 80
Data Analysis ........................................................................ 82
Results ....................................................................................... 85
Predicting Depression ............................................................... 88
The Unconditional Model ................................................... 88
Control Variables Model .................................................... 88
Abuse Model .................................................................. 88
Perceived Neighborhood Disorder Model ................................. 89
Neighborhood Fear of Victimization Model .............................. 89
Testing Neighborhood Fear of Victimization as a Mediator ............ 92
Predicting Quality of Life .......................................................... 92
The Unconditional Model ................................................... 92
Control Variables Model .................................................... 92
Abuse Model .................................................................. 93
Perceived Neighborhood Disorder Model ................................. 93
Neighborhood Fear of Victimization Model .............................. 94
Testing Neighborhood Fear of Victimization as a Mediator ............ 94
Discussion. . . . . ............................................................................... 97
CHAPTER 4
Conclusion ...................................................................................... 107
REFERENCES ................................................................................. 112

viii

LIST OF TABLES

Table 1: Study 1 Variable Means and Standard Deviations across All Time
Points ........................................................................................... 37

Table 2: Study 1 Descriptive Statistics and Frequencies for Control Variables
across All Time Points ........................................................................ 38

Table 3: Study 1 Fixed Effects across All Sequentially Nested Models Predicting
Depression ...................................................................................... 42

Table 4: Study 1 Random Effects and Model Comparisons across All
Sequentially Nested Models Predicting Depression ...................................... 43

Table 5: Study 1 Fixed Effects across All Sequentially Nested Models Predicting
Quality of Life.. ... 46

Table 6: Study 1 Random Effects and Model Comparisons across All
Sequentially Nested Models Predicting Quality of Life ................................. 47

Table 7: Study 2 Variable Means and Standard Deviations across All Time
Points... ..................................................................................................................... 86

Table 8: Study 2 Descriptive Statistics and Frequencies for Control Variables
across All Time Points ....................................................................... 87

Table 9: Study 2 Fixed Effects across All Sequentially Nested Models Predicting
Depression ..................................................................................... 90

Table 10: Study 2 Random Effects and Model Comparisons across All
Sequentially Nested Models Predicting Depression ..................................... 91

Table 11: Study 2 Fixed Effects across All Sequentially Nested Models
Predicting Quality of Life .................................................................... 95

Table 12: Study 2 Random Effects and Model Comparisons across All
Sequentially Nested Models Predicting Quality of Life ................................. 96

ix

LIST OF FIGURES

Figure 1: A conceptual model depicting the mediating role of resource
constraints on the relationship between intimate partner violence and
psychological well-being over time ....................................................... 22

Figure 2: A conceptual model depicting the impact of neighborhood disorder and
the mediating role of neighborhood fear of victimization on the psychological
well-being of survivors of intimate partner violence over time ........................ 74

Chapter 1

INTRODUCTION

Violence against women is a widespread, pervasive social issue that compromises
the physical, psychological, and emotional health and well-being of women worldwide
(World Health Organization, 2008). Most commonly women fall victim to violence at the
hands of their intimate partners (Garcia-Moreno, Jansen, Ellsberg, Heise, & Watts, 2006).
According to the Centers for Disease Control and Prevention, intimate partner violence
(IPV) is a pattern of behavior involving threatened or actual physical, sexual, and/or
emotional harm perpetrated by a current or former intimate partner (Saltzman, Fanslow,
McMahon, & Shelley, 2002) to dominate or control their current or ex-spouses’, or dating
partners’, actions, thoughts, and feelings (Gordon, 2000; Pence & Paymar, 1993).

IPV is widespread, impacting women of all ages, and across all ethnic, racial, and
socioeconomic strata (Menard, 2001). Within the United States, prevalence rates of male-
initiated partner violence against women are astounding. In a national population-based
random sample, 25% of women married or cohabitating with a male partner reported
being physically and/or sexually assaulted by their current or former partner(s) during
their lifetime, and 1.5% reported current (i.e., past year) victimization (Tjaden &
Thoennes, 2000). Based on these findings, Tjaden and Thoennes estimate that 1.5 million
women are victims of physical and/or sexual IPV each year. Prevalence rates are often
higher (e. g., 29% - lifetime) when considering the impact of psychological or emotional

abuse in addition to other forms of violence (Coker et al., 2002).

Women with abusive partners often suffer devastating injuries as well as
additional short- and long-term physical and psychological health consequences from the
abuse they experience (see Campbell, 2002 for a review). For instance, survivors of IPV
are at significant risk for anxiety, depression, post-traumatic stress disorder, suicidal
ideation, and poor overall quality of life (Adeodato, Carvalho, de Siqueira, & de Matos e
Souza, 2005; Dutton, Kaltman, Goodman, Weinfurt, & Vankos, 2005; Golding, 1999;
Leiner, Compton, Houry, & Kaslow, 2008). Generally, the more severe and frequent the
abuse, the more significant the impact that IPV has on survivors’ overall psychological
well-being (Campbell, Kub, Belknap, & Templin, 1997; Dutton et al., 2005). The actual
tactics of abuse employed by assailants against their partners also have important
implications for survivors’ psychological sequelae. Research finds that cumulative
exposure to multiple forms of violence (e. g., physical, sexual, and psychological abuse)
appears to be most damaging to the well-being of women with abusive partners (Houry,
Kemball, Rhodes, & Kaslow, 2006).

While IPV has clearly been shown to relate to poor mental health outcomes (see
Dutton et al., 2006 and Golding, 1999 for a review), far more empirical attention is
needed to understand the process by which physical and psychological abuse
independently and collectively predispose women to pernicious psychological
symptomatology and compromise their overall well-being. To date, research has found
that women’s cognitive appraisals of the abuse they have experienced and their coping
strategies may play an important role in their overall well-being (e. g., Calvete, Corral, &
Estevez, 2008; Calvete, Estevez, & Corral, 2007). However, much of this research has

emphasized individual-level factors. It is essential to examine potential intervening

factors that go beyond qualities intrinsic to survivors. Broader contextual factors may
also help explain or contribute to the development of negative mental health sequelae in
the lives of women with abusive partners. Understanding the impact of contextually-
based additive factors and intervening mechanisms that explain or modify the abuse-
distress relationship is essential to informing intervention strategies which can possibly
assist women in alleviating or completely overcoming the deleterious mental health
consequences associated with abuse. Empirical attention to ecologically-based factors
that may alleviate, or potentially exacerbate, the relationship between IPV and mental
health will provide further insight into how to assist survivors following their experiences
of abuse.

Limited research has been conducted to examine broader ecological factors that
may be particularly salient to survivors’ experiences and their mental health trajectories
over time, following IPV. Past longitudinal research has found that social support and
secondary stressors (e. g., family responsibilities, loss of income) in addition to IPV have
important implications for survivors’ mental health (Anderson, Saunders, Yoshihama,
Bybee, & Sullivan, 2003; Beeble, Bybee, Sullivan, & Adams, 2009). Two additional
factors that may have important implications for survivors’ psychological well-being are
women’s access to community-based resources and the neighborhood environments
within which they reside.

External resources, whether tangible (e. g., food, clothing, shelter) or intangible
(e.g., social support), are important for all individuals’ autonomy and broader well-being.
Difiiculty obtaining resources, or a loss of resources, may pose significant stress to those

attempting to meet their basic needs, or seeking to broaden their social and/or economic

capital. Within abusive relationships, assailants may restrict women’s access to external
resources and supports as a mechanism of control and domination (Dobash & Dobash,
1998; Raphael, 1999). The broader literature examining threatened or actual material loss
or hardship finds this to be a significant correlate of poor psychological well-being (e.g.,
Hobfoll, Johnson, Ennis, & Jackson, 2003). Within the context of IPV, resource
constraints may have significant implications for women’s mental health and quality of
life, as women may be unable to gain access to the requisite resources and supports to
independently provide for themselves and their children, as a result of the abuse they
experience. For survivors, this in turn may foster further dependency on their assailants.
Limited empirical attention exists examining if resource constraints may partially mediate
the relationship between IPV and survivors’ psychological well-being.

Another area significantly understudied is how survivors’ contextual
environments may influence their broader psychological well-being. The neighborhood
context in which individuals live has important implications for residents’ mental health.
Disadvantaged neighborhoods perceived to be high in social and physical incivilities
have been found to be detrimental to residents’ psychological well-being (see Truong &
Ma, 2006 for a review). Prevalence rates of IPV within disadvantaged neighborhoods are
also alarmingly high (Benson, Fox, DeMaris, & Van Wyk, 2003; Miles-Doan, 1998).
However, limited attention has been directed to understanding the cumulative stress
survivors of IPV experience while residing in disadvantaged neighborhoods.
Understanding how resource constraints and survivors’ broader contextual environments
may impact their mental health has important implications for women’s future well-

being. This dissertation comprises two separate but related longitudinal studies that

examined the extent to which these mechanisms impacted the mental health and overall
quality of life of women who had experienced abuse over time.

Study 1, entitled Understanding the Impact of Resource Constraints on the
Mental Health and Well-Being of Survivors of Intimate Partner Violence Over Time, was
conducted to examine if women’s difficulty obtaining requisite resources may partially
be to blame for their compromised mental health and well-being. Study 2, An
Examination of the Impact of Neighborhood Factors on the Well-Being of Survivors of
Intimate Partner Violence Over Time, examined whether perceived neighborhood
disorder impacted survivors’ psychological well-being in addition to the abuse they
experienced, and tested whether fear of victimization within the community was a
mediating factor between women’s perceived neighborhood disorder and their overall
well-being. A final concluding chapter is presented to discuss how these studies together
enhance our understanding of the ecological factors impacting survivors’ psychological

well-being and to summarize the broad implications of this research.

Chapter 2
STUDY 1:
UNDERSTANDING THE IMPACT OF RESOURCE CONSTRAINTS ON
THE MENTAL HEALTH AND WELL-BEING OF SURVIVORS
OF INTIMATE PARTNER VIOLENCE OVER TIME
Overview of the Literature

Intimate partner violence (IPV) is among the most serious of social problems
impacting women’s lives in the US. today. It is estimated that 1.5 million women suffer
the devastating consequences of IPV each year (Tj aden & Thoennes, 2000). The
implications of IPV are significant and extensive, and for most women, extend far
beyond sustained physical injuries. Abuse also has substantial negative implications for
survivors’ overall mental health and well-being (Campbell, 2002; Golding, 1999).

To date, limited research has explored the processes and various mechanisms
through which abuse impacts women’s well-being. While some explanatory mechanisms
have been unveiled (e. g., cognitive appraisal and schemas, coping strategies, social
support; Beeble, Bybee, Sullivan, & Adams, 2009; Calvete, Corral, & Estevez, 2008;
Calvete, Estevez, & Corral, 2007), many of these studies have largely focused on
qualities intrinsic to survivors (e. g., Calvete et al., 2007, 2008), thereby overlooking
external, ecologically-based mechanisms. Additionally, most studies have been cross-
sectional, precluding our ability to delineate the temporal ordering of events.

One historically overlooked, and particularly important factor to examine in
accordance with the abuse-distress relationship is women’s access to community
resources, particularly the difficulties and constraints women experience when attempting

to acquire resources. Previous research has demonstrated that access to resources has

important implications for survivors’ mental health and well-being (e.g., Bybee &
Sullivan, 2002); however, it still remains unclear how abuse impacts women’s resource
constraints, and how these resource constraints in turn relate to survivors’ well-being
over time. Therefore, the current study investigated whether women’s difficulty obtaining
resources is one such mechanism through which abuse impacts survivors’ psychological
well-being, using a longitudinal design with a community-based sample of women with
abusive partners.
The Impact of Intimate Partner Violence on Survivors’ Mental Health and Well-Being

An abundance of research exists examining the psychological well-being of
survivors of IPV. Collectively this work demonstrates the serious and detrimental
psychological impact of exposure to any and all forms of IPV (e.g., physical, sexual,
and/or psychological abuse). Women experiencing abuse have been found to suffer a
wide range of mental health impacts, including but not limited to depression, anxiety,
post-traumatic stress disorder (PTSD), and suicidal ideation (Cascardi, O’Leary, &
Schlee, 1999; Dutton, Kaltman, Goodman, Weinfurt, & Vankos, 2005; Golding, 1999;
Helfrich, Fujiura, & Rutkowsli-Kmitta, 2008; Leiner, Compton, Houry, & Kaslow, 2008;
Pico-Alfonso et al., 2006). Among the most common mental health consequences
experienced by survivors are depression and PTSD (Campbell, 2002; Golding, 1999), and
for some women these disorders co-occur (Cascardi et al., 1999; Nixon, Resick, &
Nishith, 2004).

Rates of psychological distress and mental health disorders among women with
abusive partners consistently far exceed comparative groups of non-abused women

(Coker, Smith, Bethea, King, & McKeown, 2000; Houry, Kemball, Rhodes, & Kaslow,

2006; O’Campo, Woods, Jones, Dienemann, & Campbell, 2006; Theran, Sullivan, Bogat,
& Stewart, 2006), and women in the general population (Cascardi et al., 1999; Golding,
1999; Helfi'ich et al., 2008). Survivors of IPV are estimated to be three to four times more
likely than non-abused women to experience depression, suicidal ideation, and PTSD
(Golding, 1999), with the concomitance of multiple forms of IPV, and extent and severity
of exposure to abuse directly associated with the severity of mental health
symptomatology (Campbell, Kub, Belknap, & Templin, 1997; Dienemann et al., 2000;
Dutton et al., 2005; Golding, 1999; Houry et al., 2006; Pico-Alfonso et al., 2006).
Beyond specific mental health symptomatology or illness, some empirical
evidence exists which suggests that IPV also has significant negative implications for
women’s overall quality of life (Adeodato, Carvalho, de Siqueira, & de Matos e Souza,
2005; Alsaker, Moen, & Kristoffersen, 2008; Leung, Leung, Ng, & Ho, 2005). For
instance, when investigating life quality among a sample of 1,614 medical patients,
Leung and colleagues (2005) found that women who experienced IPV had significantly
lower scores across several life domains (physical and psychological well-being; social
relationships and fimctioning; overall satisfaction with one’s environment) when
compared to non-abused women. However, to date, when examining the effects of IPV,
researchers have almost exclusively focused on specific psychological symptomatology,
largely overlooking the impact of IPV on women’s broader well-being. For some
survivors, abuse may not manifest itself explicitly as mental health symptomatology;
rather, women may express more general discontentment with their lives. To gain more

insight into the impact of abuse on women's lives, research must attend to both

psychological symptomatology as well as women’s broader perceptions of their life
quality.

While any psychological distress or illness, or diminished perceptions of life
quality may pose significant difficulties and challenges to those affected, depression in its
most severe state is among the most debilitating of mental health outcomes, and is
associated with significant functional impairment (Kessler et al., 2003) and suicidal
thoughts and behaviors (Houry, Kaslow, & Thompson, 2005; Kessler, Borges, &
Walters, 1999). Several studies examining prevalence rates of depression among
survivors have been conducted (e. g., Campbell et al., 1997; Cascardi et al., 1999; Nixon
et al., 2004; O’Campo et al., 2006). While rates have been found to differ substantially
depending upon sampling approach, measurement criteria, and overall methodology, they
remain alarmingly high across all samples. In a recent study of 74 women in a domestic
violence shelter, Helfrich and colleagues (2008) found that 51.4% of women met 12-
month criteria for major depression, compared to only 2.4% of women in the general
population. Similarly, in an extensive review of 18 studies investigating the impact of
physical abuse on women’s mental health, Golding (1999) found the mean prevalence
rate of depression across all studies was 47.6%. Women who were physically battered
were 3.80 times more likely to be depressed than were women who were not physically
assaulted by their partners. However, many studies examining rates of depression and
other concurrent or post-abuse mental health sequelae among survivors have only
examined the effects of physical battering on women’s mental health.

While the impact of physical abuse is paramount to our understanding of the

etiology of survivors’ mental health and well-being, other non-physical tactics commonly

used by assailants to terrorize or demoralize their partners must also be at the forefront of
our investigations. Recently researchers have called attention to the need to examine the
multidimensional nature of abuse and its impact on women’s psychological well-being
(Mechanic, Weaver, & Resick, 2008; O’Leary, 1999; Stark, 2007).

Historically, the impact of psychological abuse has been overlooked in the
literature (O’Leary, 1999) despite women often rating this form of abuse as more
damaging to their mental health than physical violence (Follingstad, Rutledge, Berg,
Hause, & Polek, 1990). Several studies have found that psychological abuse uniquely
impacts survivors’ well-being (Dutton, Goodman, & Bennett, 1999; Sackett & Saunders,
1999; Marshall, 1999; Mechanic et al., 2008; of. Campbell et al., 1997). For instance,
investigating the impact of multiple forms of IPV (physical violence, sexual coercion,
psychological abuse, and stalking) on depression among a sample of413 recently-
battered women from community-based programs, Mechanic and colleagues (2008)
found that psychological abuse significantly predicted depression, over and above
women’s experiences of physical abuse, sexual abuse, and sustained injuries. When
reversing the order of the predictors, physical and sexual abuse did not contribute to the
prediction of depression, after accounting for psychological abuse.

In another study, Pico-Alfonso and colleagues (2006) compared depressive
symptomatology among two groups of abused women (those experiencing both physical
and psychological abuse, and those only experiencing psychological abuse) to a group of
non-abused women. Incidence and severity of depressive symptomatology was
significantly higher among abused women than non-abused women, as expected.

However, no differences were found among the two groups of abused women, suggesting

10

that psychological abuse alone may be just as damaging to women’s mental health as the
co-occurrence of physical and psychological abuse. These findings speak to the
importance of examining survivors’ experiences of psychological abuse, in addition to
physical violence, when attempting to understand their overall well-being.

To date, the majority of research investigating the relationship between IPV and
psychological well-being has been cross-sectional. While cross-sectional research
suggests that recency and continuity of abuse are particularly salient to changes in
survivors’ mental health (e.g., Bogat, Levendosky, Theran, von Eye, & Davidson, 2003;
Bonomi et al., 2006), we must draw upon longitudinal research to guide our
understanding of the long-term impact of abuse on women’s lives, their quality of life,
and on the development and manifestation of mental health symptomatology. There is a
dearth of research investigating the impact of IPV on women’s psychological well-being
over time and results are somewhat equivocal. In some studies examining depression, as
violence decreases over time, depression also decreases (e.g., Anderson, Saunders,
Yoshihama, Bybee, & Sullivan, 2003; Campbell, J. & Soeken, 1999; Campbell, R.,
Sullivan, & Davidson 1995), while in others, depression remains steady, despite a drop in
abuse (e.g., von Eye & Bogat, 2006).

To investigate this relationship, J. Campbell and Soeken (1999) interviewed a
cormnunity sample of 141 physically battered women three times over 3.5 years,
retaining 98 women for the final interview. Women who experienced abuse at all three
time points had significantly higher depression scores at time three than women for
whom the abuse stopped after the first or second interview. However, at time three,

depression levels for women only exposed to abuse at times one or two rose to their

11

original level, signifying mild and moderate depression, despite the absence of recent
abuse.

Similarly, R. Campbell and colleagues (1995) examined changes in depression at
three time points (upon shelter exit, 10 weeks thereafter, and again 6 months later) among
a sample of 139 women exiting a domestic violence shelter. For all women, depression
significantly declined from shelter exit (when 83% of women were at least mildly
depressed) to the lO-week follow-up (58%), and then remained steady through the 6-
month follow-up (58%). Women for whom physical abuse continued at lO-week and 6-
month follow-ups had significantly higher rates of depression, compared to women who
were no longer experiencing violence, suggesting that depression is most acute at times
of violence, and may lessen in intensity or completely subside when abuse stops.
However, still a substantial percentage of women not physically abused at lO-week and
6-month follow-ups reported at least mild depressive symptomatology (50% and 49%,
respectively). For some women still experiencing depressive symptomatology, in the
absence of physical violence, it is possible that psychological abuse (not measured in this
study) may have continued to exert an impact on women’s well-being over time.

One study specifically examined how cessation of all abuse (physical, sexual, and
psychological), versus some forms of abuse, impacts women’s trajectories of depression
over 2 years among a sample of 448 women who had either had police or court
intervention (Kernic, Holt, Stoner, Wolf, & Rivara, 2003). The largest decline in
depression over time was seen among women with a history of physical, sexual, and

psychological abuse for whom all abuse stopped. Those women, for whom some abuse

12

continued, particularly psychological abuse, experienced less of a decline in depression
over time.

Unlike the aforementioned studies, von Eye and Bogat (2006) found that, among
a sample of parenting women, as violence decreased over time, depression did not. The
authors concluded that earlier experiences of violence may in fact have long-lasting
mental health implications, and that attention is needed to examine factors in addition to
physical abuse that may impact the etiology of depression among survivors over time.
When examining long-term depression trends among a sample of 94 sheltered women
separated from their abusive partners, Anderson and colleagues (2003) also explored the
impact of women’s secondary stress. levels upon shelter exit. They found that depression
declined over time for women experiencing lower levels of violence and secondary stress
(e. g., family responsibilities, income loss) at shelter-exit. However, among women
experiencing high violence and stress at shelter-exit, depression remained stable over
time or increased slightly.

Longitudinal investigations of the impact of abuse on survivors’ quality of life are
even more limited, despite quality of life having important long-term implications for
survivors’ safety (Bybee & Sullivan, 2002, 2005). In a nationally-based prospective
study, Zlotnick, Johnson, and Kohn (2006) interviewed a representative sample of
married and cohabiting women twice over 5 years. They found that experience of
physical violence at Time 1 significantly predicted women’s quality of life at Time 2,
five years later; however, this analysis did not account for physical violence that had
occurred after the initial time point, precluding our understanding of how ongoing or new

abuse may impact women’s broader life satisfaction. In another earlier study, using the

13

same community-based sample of women with abusive partners as the current study,
Beeble and colleagues (2009) examined the collective impact of physical and
psychological abuse on women’s quality of life over two years. They found that initial
levels of abuse were not predictive of quality of life over time; rather, change in
psychological abuse relative to women’s initial interview was predictive of change in
quality of life. At times when women’s psychological abuse was higher, their quality of
life was lower, suggesting that abuse may have a more immediate impact on women’s
well-being.

These findings, taken together, suggest that while the abatement of abuse may
largely alleviate the devastating, immediate mental health consequences experienced by
survivors, residual impacts of abuse on mental health and well-being may remain over
time for some women. It is important to recognize that cessation of one formof abuse
does not necessarily mean all forms of abuse have stopped, and that some lasting mental
health consequences may be the result of more subtle forms of abuse (e.g., psychological
abuse). As recommended by von Eye and Bogat (2006), and employed by Anderson and
colleagues (2003), it is also important to consider additional factors and stressors that
may further explain, or perhaps, either alleviate or exacerbate women’s psychological
distress, as these other factors may be key elements in understanding the abuse-distress
relationship.

Resource Constraints in the Lives of Women with Abusive Partners

One important factor to consider that may impact or partially explain the abuse-

distress relationship for survivors is their access to community resources and supports.

Research has found that assailants use a variety of strategies which may directly or

14

indirectly limit women’s access to external resources, opportunities, and supports
(Dobash & Dobash, 1998; Raphael, 1999; Riger & Staggs, 2004; Swanberg & Logan,
2005; Tjaden & Thoennes, 2000). Perpetrators employ a variety of tactics to dominate
and control their partners, including physical threats, ridicule, jealousy, threatening to end
their relationships, restricting women’s activities and access to resources, and damaging
women’s property (F ollingstad etal., 1990; Stark, 2007). These tactics of psychological
abuse and coercive control create and exploit vulnerabilities among women, facilitate
attachment and emotional dependency on their abusers, and wear down women’s
resistance to assailants’ demands (Dutton & Goodman, 2005).

According to Dutton and Goodman (2005), women’s ability to resist coercion is
facilitated by, and in part, reliant upon access to tangible (e. g., economic resources,
housing, transportation), social (e. g., social support), and personal resources (e.g.,
determination). The inverse relationship between abuse and access to resources has been
supported both cross-sectionally and longitudinally (Bosch & Schumm, 2004; Bybee &
Sullivan, 2002, 2005; Goodman, Dutton, Vankos, & Weinfurt, 2005; Sullivan & Bybee,
1999). For instance, among a sample of women recruited from a domestic violence
shelter, Sullivan and colleagues found that access to community-based resources (e.g.,
housing, education, employment, health care, child care, social support, transportation)
protected women from experiencing future abuse. In another longitudinal study of 406
African American help-seeking survivors, Goodman and colleagues (2005) found that a
lack of material resources, specifically employment and independent residency, predicted
revictimization; however, after accounting for social support and women’s resistance

strategies, material resources were no longer significant predictors of reabuse.

15

Given the protective nature of resources and supports in women’s lives, depleting
or restricting women’s access to resources and opportunities allows assailants to maintain
control, leaving their partners with few alternatives but to comply with their demands.
The manipulative and controlling actions used by assailants have far-reaching, serious
implications for survivors’ immediate and future livelihood. Abusers often sabotage
women’s access to educational or employment opportunities either directly (e.g.,
preventing women from attending work or school) or indirectly (e. g., preventing women
from being able to work because of injuries sustained from the abuse), and deprive
women of access to additional economic resources, such as their own personal or family
income (Adams, Sullivan, Bybee, & Greeson, 2008; Leone, Johnson, Cohan, & Lloyd,
2004; Ptacek, 1997; Raphael, 1996; Swanberg & Logan, 2005). Assailants may also
directly isolate survivors from their friends and family (Goodkind, Gillum, Bybee, &
Sullivan, 2003; Murphy & Cascardi, 1999), fostering further emotional dependency and
leaving women without requsite support networks to cope with the devastating impacts of
IPV.

While the direct actions of assailants are devastating enough to survivors, the
implications of abuse extend much further, exacerbating the difficulty women experience
in accessing resources they need to escape from abuse and to restore their physical and
emotional well-being. For instance, restricting survivors’ access to economic resources
precludes women from meeting their financial obligations, often causing additional
adversities. Insufficient access to food, unmet nutritional needs, restricted access to health
care (Weinbaum et al., 2001), and loss of stable housing (Menard, 2001) are significant

barriers to self-sufficiency for survivors of IPV. Additionally, as women attempt to leave

16

their abusive relationships the difficulties they face are often insurmountable. For
example, the lack of affordable housing options available to survivors (Menard, 2001)
may lead to homelessness (Browne, & Bassuk, 1997), or may force women to return to
their abusive partners (Davies, Lyon, & Mionti-Catania, 1998). Therefore, as a result of
the abuse they experience, survivors may be restricted in their abilities to reach out to
fi'iends and family or professionals for help, find adequate housing for themselves and
their children, obtain requisite medical and mental health services, and/or engage in
educational or employment opportunities to enhance their skills (Dutton & Goodman,
2005; Lipsky & Caetano, 2007; Menard, 2001; Plichta & Falik, 2001).

Access to resources has been identified as an important correlate for individuals’
overall well-being (Diener & Fujita, 1995). As such, loss of or difficulty obtaining
requisite resources has been found to be associated with negative mental health sequelae
and poor overall quality of life across a variety of populations (Casey et al., 2004;
Hobfoll, Johnson, Ennis, & Jackson, 2003; Schumm, Hobfoll, & Keogh, 2004; Wells,
Hobfoll, & Lavin, 1999). For instance, in a study investigating threatened or actual
material resource loss (e. g., food, money, transportation, children’s essentials) among
inner-city women over nine months, Hobfoll and colleagues (2003) found that women
who reported increased material loss over time had significantly higher depression scores
at study follow—up than did women with stable levels of resources over time.

Resource constraints are a significant source of additional stress and compromised
well-being for survivors of IPV (Bybee & Sullivan, 2002). To understand the potential
connection between IPV, resource constraints, and deleterious mental health outcomes, I

draw upon conservation of resources theory (COR; Hobfoll, 1988, 1989; Hobfoll & Lilly,

17

1993). According to COR, individuals strive to acquire, strengthen, and protect the
resources they have in their lives, which enhances individuals’ well-being. Resources are
defined broadly as characteristics, conditions, or objects of value (Hobfoll, 1989) which
are mobilized during times of stress to alleviate the negative impact of a stressor
(Johnson, Palmieri, Jackson, & Hobfoll, 2007). Resources are vital to an individual’s
ability to cope with stress; therefore, when threatened or actual resource loss occurs,
serious negative implications follow for the individual’s psychological well-being.
Hobfoll (1989) contends that environmental circumstances are often the cause of actual
or threatened resource depletion. In addition, the unequal distribution of resources among
individuals in society further exacerbates vulnerability for loss among those who have the
fewest of resources (Dohrenwend, 1978). Loss or lack of resources sets in motion what
Hobfoll refers to as ‘loss spirals,’ whereby individuals who lack resources or experience
resource depletion then suffer distress, which in turn further exacerbates loss (Johnson et
al., 2007).

Women with abusive partners suffer significant resource loss as a result of the
demands imposed by their partners. Therefore, in the context of IPV, COR theory
suggests that the negative mental health sequelae and poor quality of life of survivors
may be due, in part, to the difficulty they have obtaining requisite resources, as a result of
the abuse they have experienced. Any loss or hardship in obtaining necessary resources
may be the direct result of the assailant wearing down women’s resistance strategies
(Dutton & Goodman, 2005), leaving women without the requisite resources to cope with

the abuse they experience, and in turn causing women significant distress. According to

18

COR theory, the negative mental health implications for survivors would then cause
further resource loss, starting a downward spiral (Johnson et al., 2007).

A paucity of research exists examining survivors’ access to resources as an
explanatory factor for their compromised mental health and well-being. In a recent study,
F ord-Gilboe et a1. (2009) tested whether the relationship between severity of abuse and
mental health was mediated by access to personal (women’s knowledge, skills, beliefs,
and behaviors), social (emotional and tangible support, and social conflict), and economic
(financial strain) resources and constraints cross-sectionally among a community-based
sample of 309 abused women. Severity of IPV was only found to significantly relate to a
reduction in economic resources, which mediated the relationship between abuse and
mental health. An important limitation of this study was the wide range in time since
women had left their abusive partners (3 - 40 months) at the time the study was
conducted, given the hypothesized relationship between abuse and its impact on women’s
access to resources. Longitudinal research is needed to explore this relationship,
emphasizing additional ecological factors that extend beyond the economic strain
experienced by survivors.

The Current Study

The current study investigated whether difficulty obtaining resources served as a
mechanism which partially explained, or mediated, the relationship between experience
of physical and psychological abuse and depression and quality of life among a sample of
recently abused women over time. The purpose of the current study was to expand our
understanding of the processes through which IPV impacts women’s mental health and

well-being, by investigating variables that extend beyond the individual level of analysis

19

to consider ecologically-based contextual factors. Identification of women’s difficulty
obtaining resources as an explanatory factor of their mental health and well-being has
important implications for future outreach, intervention, and service delivery.

A significant strength of this study, over prior research, is its ability to examine
these relationships longitudinally, while separating the between-women effects (i.e., how
women’s experiences differ from one another across time) from the within-woman
effects (i.e., how individual women’s trajectories change over time, relative to their own
past experiences). Simultaneously modeling within- and between-person effects allows
for the examination of variables that may predict individuals’ trajectories of change over
time (Singer & Willet, 2003). In addition, this study attended to the broader well-being of
survivors of IPV by examining quality of life as a consequence of abuse, in addition to
depression. Quality of life is a construct warranting further investigation as researchers
and interventionists strive to improve the lives of women with abusive partners.

As depicted in Figure 1, this study investigated whether the experiences of
physical and psychological abuse threatened survivors’ abilities to obtain requisite
resources, and whether this, in turn, may have negatively impacted their psychological
well-being. Specifically, the following over-arching research question was tested: Does
women ’s reported difficulty obtaining community resources partially mediate, or explain,
the relationship between intimate partner violence and psychological well-being
(depression and quality of lifia) overtime? As seen in Figure 1, this model allowed for the
simultaneous examination of between-women and within-woman effects. The hypotheses
that follow are divided so that the expected between-women and within-woman effects

are distinguishable.

20

A) Between-Women Differences Predicting Change over Time
H1: Higher levels of baseline physical and psychological abuse will be associated
with women’s greater initial difficulty obtaining resources.
H2: Higher levels of initial difficulty obtaining resources will be associated with
an increased trajectory in depression and decreased trajectory in quality of life
over time.
H3: Higher levels of baseline physical and psychological abuse will be associated
with an increased trajectory in depression and decreased trajectory in quality of
life over time.
H4: The relationship between baseline levels of physical and psychological abuse
and change in psychological well-being (depression and quality of life) over time
will be partially mediated by women’s initial difficulty obtaining resources.

B) Within-Woman Trajectories of Change over Time
H5: Within-woman increases in physical and psychological abuse will be
associated with increased difficulty obtaining resources over time.
H6: Within-woman increases in difficulty obtaining resources will be
accompanied by increases in depression and decreases in quality of life over time.
H7: Within-woman increases in physical and psychological abuse will be
accompanied by increases in depression and decreases in quality of life over time.
H8: The relationship between changes in physical and psychological abuse and
psychological well-being (depression and quality of life) over time will be

partially mediated by changes in difficulty obtaining resources over time.

21

Figure 1. A conceptual model depicting the mediating role of resource constraints on the

relationship between intimate partner violence and psychological well-being over time.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:9

8

g r

5 Initial

<8: Resource

3. J Constraints

t:

0

° Baseline

E

5 Abuse

Change in

K Psychological
r Well-being

g Change in ‘

t Abuse

m

g i Change in

g Resource

é Constraints

:5: K

3

 

 

 

22

Method

The current study utilized data from a larger longitudinal investigation of the lives
of 160 battered women and their children over two years who were residing in a mid-
sized Midwestern city. In this larger study an experimental design was employed to test
the effectiveness of an intensive community-based advocacy intervention lasting four
months. Families were randomly assigned to either receive advocacy services or a
‘service as usual’ condition. Women and their children, regardless of their placement in
the experimental advocacy or control condition, were interviewed six times over two
years for the larger study. However, because women’s resource constraints were not
examined during their first two interviews, the current study was limited to examining the
experiences women reported during their last four interviews from the larger study.
Participants

The current study investigated the subsample of 138 of the 160 female survivors
of IPV who participated in the first interview which inquired about the difficulties
women experience obtaining community-based resources (8 months after their initial
entry into the research study). All women who participated in the current study met
several eligibility criteria set for the larger study. First, all women had experienced
physical violence that was perpetrated by an intimate partner or ex-partner within the four
months prior to their study involvement. Second, women had to agree to be randomly
assigned to either the experimental or control condition and had to plan to stay in the
surrounding area for the four months following their first, baseline interview. Finally,
women had to have at least one child living with them between the ages of 5 and 12 who

was also interested in participating in the larger study.

23

Procedure

Participants were recruited from one of three sources that serve women who had
experienced recent violence by a current or former intimate partner, including: (a) a
community-based agency providing short-term services and supports to survivors
following a recent police response to a violent encounter (n = 59; 43%), (b) two
community-based domestic violence programs providing residential and nonresidential
services to women and their children (n = 40; 29%), or (c) the county prosecutor’s
personal protection order office (n = 39; 28%). Two research team members worked with
the aforementioned agencies to recruit women into the study.

After discussing the study with research staff and agreeing to participate, women
provided written informed consent. Baseline interviews were conducted with each
woman within 2 to 3 weeks following recruitment. For several reasons, women recruited
from residential shelter programs were not interviewed until they had exited the shelter.
First, it was important to avoid interviewing families while they may have been in crisis.
Second, the intervention was designed to help women and their children acquire requisite
resources and supports in their ‘naturally-occurring’ environments to ensure that the
resources would be sustainable over time, or able to be obtained independently, should
the need arise again for women in the future. Next, it would have been difficult for
advocates to work in the shelter setting, providing support to some women while
withholding it from others. Finally, waiting until women exited the shelter minimized the
risk of experimental contamination, had women residing in the shelter receiving
advocacy services shared information and resources with other women who were in the

control condition. All of the women recruited from residential programs had exited the

24

shelters within one week of being recruited, and had their baseline interviews within the
subsequent two to three weeks.

Immediately following their baseline interviews, families were randomly assigned
to either the experimental advocacy intervention (n =59; 43%) or the control condition (n
= 79; 57%), whereby families received ‘services as usual’ within the community.
Interviewers were blind to which group families would be assigned. Within two weeks,
families assigned to the experimental condition were provided a trained paraprofessional
advocate who began assisting them in obtaining resources that women had identified
would be helpful to their families.

The advocacy intervention. The advocacy intervention used a family-centered
model which focuses on the strengths, competencies, and broader unmet needs of
families (Dunst, Johanson, Trivette, & Hamby, 1991). According to Dunst and
colleagues, this model of intervention is built upon a consumer-oriented foundation,
whereby all service provision and resource allocation is solely based upon the needs
identified by the families. Advocates in this context are viewed as paraprofessionals who
serve as family agents to identify and procure requisite resources, while also assisting
families to increase their capacity to build broader networks of support and resources for
the future. Advocates worked with families on a number of areas such as legal assistance,
housing, health care, child care, child support, education, employment, obtaining a
restraining order, etc. Each advocate spent between 6 to 8 hours each week working with
and on behalf of the family for a four month period.

Interviews and interviewer training. As part of the larger study, participants were

interviewed six times over two years; as indicated, the baseline interview took place

25

shortly after recruitment, and the second interview occurred four months later, post-
intervention; however, the first and second interviews were excluded from the current
study as they did not ask about resource constraints. All subsequent interviews occurred
8, 12, 18, and 24 months after the baseline interview. The current study only utilized
levels of abuse at baseline and data from the 8, 12, 18, and 24-month follow-up
interviews, which are referred to as Time 1, 2, 3, and 4, respectively, hereafter.

All interviews were administered face-to-face, audio-taped, and conducted in a
space determined to be safe, comfortable, convenient, and private for the women and
their children. In most cases, interviews took place within the women’s homes.
Interviews across all four time points ranged from 30 minutes to 4.75 hours (M = 104.67
minutes; SD = 34.75). All women received $20 for their baseline interview, $60 for the
post-intervention interview (both of which were not included in the current study), $70
for their 8 month follow-up (Time 1) interview, $80 for their 12 month follow-up (Time
2) interview, $90 for their 18 month (Time 3) follow-up interview, and $100 for their 24
month (Time 4) follow-up interview, as remuneration.

Interviews were conducted by female undergraduate students attending a large
Midwestern university who received college credit for their work on this project.
Students underwent extensive training over an 8-week period, learning about the
theoretical and applied tenets of conducting community-based research, particularly when
working with families affected by IPV. Training provided interviewers with the skills
needed (e.g., developing rapport, empathetic response) to effectively handle a variety of
issues that they could have encountered during their interviews. Mock interviews were

conducted between students, until response coding reached at least 95% agreement.

26

Supervisors closely monitored student interviewers in 3-hour weekly student meetings,
and by listening to each audio-taped interview. Interviews were also randomly assigned
to the pool of student interviewers to be checked for consistency in coding over time.

Retention strategies. Relative to the full sample from the larger study, retention
rates at Times 1, 2, 3, and 4 were 86%, 88%, 88%, and 86%, respectively. Women who
missed one or more interviews were not found to differ by demographic characteristics,
or by level of abuse experienced.

Several strategies were employed over the course of this study to maximize
participant retention over time. Specifically, a three-phase protocol was used: (a) Phase I
- setting the stage for future contact, (b) Phase II - utilizing participant-oriented strategies
to locate participants, and (0) Phase III - reaching out to social and community—oriented
networks to locate participants (Sullivan, Rumptz, Campbell, Eby, & Davidson, 1996).

Phase I occurred at the time of the baseline interview, and involved several
strategies to lay the groundwork for later success in contacting participants. First,
emphasis was placed on building trust with participants by extensively informing women
about the purpose of the study, and the mechanisms in place to ensure their
confidentiality. During phase I interviewers collected detailed information about the most
effective strategies to use to contact them (e. g., the best days and times to call, unsafe
days and times to call, alternative contacts). Interviewees provided names, addresses, and
phone numbers of several people (e.g., family members, friends, neighbors, employers)
or government agencies providing individual or family assistance (e. g., local social
service office) who would likely know how to contact them for subsequent interviews,

should the research team have difficulties. For each alternative contact provided, women

27

signed a Release of Information Form that allowed these individuals to provide
information about women’s whereabouts to the research team. Additionally, women were
provided with an information card containing study contact information, the date and
time of their next interview, the amount women would be paid for their next interview,
and a request to follow up with project staff if they were to move.

Phase 11 involved utilizing participant-oriented strategies to contact women.
Often, the strategies employed in this phase differed depending upon women’s
preferences and safety needs. However, generally this phase began with attempting to
telephone women, or to visit their residences when women did not have phones.
Unsuccessful attempts to contact women with the aforementioned strategies were
followed up with a mailed letter to their last known address, or a phone call or letter to
their employer. When all Phase 11 efforts were unsuccessful, Phase [11 strategies were
implemented. Phase III involved utilizing phone calls, in-person visits, and/or letters to
the alternative contacts (both personal contacts and governmental agencies) provided by
the women. To ensure participants’ safety and confidentiality, no reference was made to
the nature of the research study, nor the reason for participant involvement when
contacting the alternative sources provided by the women.
Participant Demographics

At Time 1, women ranged from 22 to 49 years of age (M = 32.74; SD = 6.02).
Women were somewhat racially diverse, with slightly fewer than half identifying as non-
Hispanic White (n = 64; 46%), more than one third as African American (n = 49; 36%),
and the remaining women as multiracial (n = 11; 8%), Hispanic/Latina (n = 10; 7%),

Native American (n = 2; 1%), or Sudanese (n = 1; < 1%). One woman did not disclose

28

her racial/ethnic identity (< 1%). Women reported having between zero and eight Of their
children residing with them in their homes when their Time 1 interview took place (M =
2.72; SD = 1.32).

Women’s educational levels varied tremendously at Time 1. Nearly 19% (n = 26)
had not yet completed high school, 30% (n = 42) had completed high school or trade
school, 38% (n = 52) had attended college, and the remaining 13% (n = 18) had earned
their associates or bachelors degrees. Nearly 12% (n = 16) of all women reported that
they were attending school at the time of their interview to firrther their education.

More than half of the women (n = 85; 62%) were employed when their Time 1
interview took place. Women’s reported monthly family income ranged from $563 to
$4,400, with a mean income of $1,884 (SD = $879). Women reported that their family
income supported between two and ten people (M = 4.04; SD = 1.40). Nearly three-
quarters of the women in this sample (n = 100; 73%) received some form of
governmental assistance (e.g., welfare, food stamps, SSI/SSD).

Measures

Control variables. Several variables were statistically controlled for in this study,
including the experimental condition to which women were assigned, women’s
involvement with their assailant(s), women’s level of education, employment status,
adjusted monthly income, and race. Due to the nature of the advocacy intervention, it was
necessary to control for condition, as some women in the advocacy group post-
intervention may have been more knowledgeable or equipped in obtaining requisite
resources than women in the control group. In addition, women’s involvement with their

assailants varied over the course of the study; some women continued their relationships

29

over time while for others, their involvement subsided completely or fluctuated over
time. Therefore, variation in women’s involvement with their assailants was also
controlled. A dichotomous indicator of involvement with the assailant was derived using
items, such as “What is your relationship now?” and “Did you end your relationship in
the last 4 months or before that?”

In addition, it was important to control for several demographic characteristics
(i.e., education, employment status, income, and race) which often relate to the variables
under study, particularly women’s well-being and resource constraints. Women’s level of
education was assessed using three dummy codes: (a) less than high school, (b) high
school or trade school, and (c) some college and above, which served as the reference
category to which the other groups were compared. Employment status was represented
with a dichotomous indicator of employment (part or full-time) versus no employment.
Women’s average monthly income was adjusted to account for the number of individuals
this income supported, and finally, race was represented by a dichotomous variable
comparing non-Hispanic Whites to all remaining racial groups.

Psychological abuse. Psychological abuse was assessed using the short version of
the Index of Psychological Abuse scale (IPA), developed by Sullivan, Tan, Basta,
Rumptz, and Davidson (1992). This scale contains 24 items that measure the extent to
which assailants have emotionally or psychologically abused their partners. The original
33-item version was modified by the original authors and shortened to 24 items following
psychometric evaluation. Women reported the frequency of psychological abuse they
experienced by their partners or ex-partners in the four months prior to their baseline

interview and in between each interview. Sample items included, “How often has he

30

accused you of having or wanting other sexual relationships?” “How often has he tried to
humiliate you?” and “How often has he broken or destroyed something important to
you?” Responses were based on a 4-point Likert-type scale (1 = Never to 4 = Often). All
items were averaged to produce scale scores. Cronbach’s alpha across all time points
ranged from .89 to .95.

Physical abuse. Physical partner violence was measured using a 24-item,
modified version of the Conflict Tactics Scale (CTS; Straus, 1979). The CTS is among
the most widely used measures to assess IPV and has been shown to be valid and reliable
across multiple samples (e.g., Lucente, Pals-Stewart, Richards, & Goscha, 2001). Women
were asked to report the frequency with which various types of physical abuse had been
perpetrated against them by their partners or ex-partners in the four months prior to their
baseline interview and in between each interview. Women reported how frequently each
of the forms of physical abuse occurred on a 7-point Likert-type scale, with scores
ranging from 1 (never) to 7 (more than four times a week). Women were asked questions
such as, “How often in the last four months has he pushed or shoved you?” “How often in
the last four months has he thrown something at you?” and “How often in the last four
months has he choked or strangled you?” Women’s responses were averaged to compute
scale scores. Cronbach’s alpha across all time points ranged fi'om .86 to .95.

Depression. Depression was measured using the Center for Epidemiological
Studies-Depression Scale (CES-D; Radloff, 1977), a 20-item self-report measure used to
detect psychological distress in the general population. Women were asked to report the
extent to which they had experienced a range of depressive symptoms during the week

prior to each of their interviews. Sample items include, “I felt like everything I did was an

31

effort,” “I felt hopeful about the future,” and “I had crying spells.” The occurrence of
symptomatology was measured on a 4-point Likert-type scale, with 1 representing Rarely
or Never and 4 representing Most or All the Time. Several items were reverse-scored and
then all items were averaged to compute a scale score. Cronbach’s alpha across all time
points ranged from .91 to .93.

Quality of life. Overall well-being was measured using a nine-item scale adapted
from the Perceived Quality of Life scale originally developed by Andrews and Whitney
(1976). This scale is intended to subjectively measure global life quality and well-being.
Women were asked to report how they were feeling overall during the time period prior
to each interview on questions such as, “How do you feel about what you are
accomplishing in your life?” “How do you feel about the way you spend your spare
time?” and “In general, how do you feel about yourself?” Women responded to each item
with their reported level of satisfaction, on a 7-point Likert-type scale (1 = Terrible to 7 =
Extremely Pleased). Women’s responses were averaged to produce scale scores.
Cronbach’s alpha across all time points ranged from .89 to .92.

Resource constraints. Women’s perceived difficulties in acquiring requisite
resources to address their needs was measured using the Difficulty Obtaining Resources
scale developed by Sullivan and colleagues and used in prior work with women with
abusive partners (Bybee & Sullivan, 2002, 2005; Sullivan & Bybee, 1999). This 13-item
measure assesses women’s difficulties acquiring resources across multiple domains (e. g.,
housing, employment, education, child care, counseling, transportation, medical care),
and the perceived difficulties women would expect to encounter if they had not attempted

to obtain resources in a given domain since their prior interview. Sample items include,

32

“How much of a problem has dealing with child care or babysitting been for you in the
last 4 months?” “How much of a problem was getting health care for your child (ren)?”
and “How much of a problem was it to get a job?” Women responded on a 4-point
Likert-type scale, ranging from 1 (not a problem at all) to 4 (very much a problem).
Cronbach’s alpha ranged from .74 to .77 across all time points.

Data Analysis

Two-level multilevel models were run using HLM software (Raudenbush, Bryk,
& Congdon, 2005) to examine the longitudinal relationships among physical and
psychological abuse, resource constraints, and each of two outcome variables, quality of
life and depression. Repeated measurements (i.e., time) were modeled at level one, and
were nested within participants, modeled at level two.

Multilevel modeling (MLM) offers advantages over other approaches to
longitudinal data analysis (e. g., repeated measures AN OVA), as it allows for differential
time intervals between measurement points, as in the current study, and offers flexibility
in handling missing cases due to variability in participant retention over time (Singer &
Willlet, 2003; Snidjers, 1996). MLM is capable of using any available data across all
time points at level one to model change patterns (Luke, 2004); albeit data at level two
must be complete (Heck & Thomas, 2000). Therefore prior to running analyses, race was
imputed for one woman using the expectation-maximization algorithm in SPSS, which is
an iterative approach used to produce maximum likelihood estimates for missing data
based on available data (Allison, 2002).

MLM was used to examine both between-women effects (i.e., how women may

differ from one another) and within-woman effects (i.e., how women may change over

33

time relative to their own past experiences). Specifically the following effects were
examined in this study: (a) between-women associations of baseline abuse and Time 1
resource constraints and Time 1 depression and quality of life scores, (b) between-women
effects of baseline abuse and Time 1 resource constraints on changes in women’s
depression and quality of life over time, and (c) within-woman effects of change in abuse
and resource constraints on changes in depression and quality of life over time.

To assess within-woman effects, Singer and Willet (2003) recommend centering
time-varying covariates, producing deviation scores to reflect the extent of change
women report, relative to a meaningful constant, such as the score from the earliest time
point available for each measure. Deviation scores were computed by subtracting each
woman’s baseline physical and psychological abuse scores from her scores at each of the
subsequent time points. Deviation scores reflecting the extent of change in women’s
difficulty obtaining resources were computed by subtracting women’s Time 1 interview
scores from their scores at each of the three subsequent time points. The earliest possible
time point for which each measure was available served as an appropriate benchmark
from which to compare subsequent change, because these time points were closest to
when women experienced the most severe physical violence.

Full maximum likelihood estimation (MLE) was used for all analyses to estimate
model parameters to allow for statistical comparisons between sequentially nested
models. MLE permits the use of the chi-square difference test to determine if the addition
of each predictor variable significantly improves the variance accounted for in the
outcome variables. All statistical tests used robust standard errors to minimize the

influence of extreme scores on results.

34

Initially, for all models, intercept (Time 1) and slope terms (trajectories of change
over time) for quality of life and depression were estimated as random effects, allowed to
vary from woman to woman. However, there was no significant (or trend level) variation
from woman to woman in the slope of depression over time, and therefore in all
subsequent models predicting depression the slope of time was fixed. Due to the nature of
the sample recruitment criteria (having experienced physical violence prior to their study
involvement) and many women’s initial intentions to discontinue their involvement with
their assailant(s), the effects of abuse and resource constraints were assumed to be similar
across women and were fixed (i.e., not allowed to vary over time).

Four sequentially nested models were run for each outcome variable. First,
unconditional models, including only intercept and slope terms, were run to assess
variation in the dependent variables, QOL and depression over time, while controlling for
experimental condition at level two. In the second model, all control variables were
added, with level of education, employment status, and women’s adjusted monthly
income modeled at level one (modeled as time-varying covariates), and race modeled at
level two. Next, physical and psychological abuse and women’s involvement with their
assailants (as a control variable at level one) were added to the overall models. Finally,
women’s difficulty obtaining resources was added to each of the final models.

Following this, tests for mediation were conducted. A mediator, as defined by
Baron and Kenny (1986), is an additional variable through which an independent variable
impacts a dependent variable. To test for mediation, the product of coefficients method
recommended by MacKinnon, Lockwood, Hoffman, West, and Sheets (2002) was used.

Using this method, two regression equations were run, one which regressed the mediator

35

on the independent variable, and the latter which regressed the dependent variable on the
mediator and the independent variable. A Sobel test (Sobel, 1982) was then conducted
using the unstandardized coefficients and standard errors from these regression analyses
to determine if the mediated or indirect effects were statistically significant. Full
mediation was characterized by the absence of a direct association between the
independent and dependent variables, after accounting for the mediator; partial mediation
was characterized by a significant association between the independent and dependent
variables, after accounting for the mediator; and an indirect effect was characterized by
the absence of a direct association between the independent and dependent variables,
prior to the inclusion of the mediator in the model (Holmbeck, 1997).
Results

In comparison to women’s average scores at baseline, experiences of physical and
psychological abuse were lower at each follow-up time point. Similarly, perceptions of
difficulty obtaining community-based resources were also lower at all time points, when
compared to scores at Time 1. Women’s average depression scores were also slightly
lower at all three time points following Time 1, and women’s quality of life scores were
slightly higher at each follow-up time point. The means and standard deviations for all
independent variables, dependent variables, and the hypothesized mediator, difficulty
obtaining resources, are presented in Table 1, and descriptive statistics and frequencies
for time-varying control variables are presented in Table 2. Results of the four
sequentially nested models are presented separately for depression (in Tables 3 and 4)
and quality of life (in Tables 5 and 6). Specifically, model fixed effects are presented in

Tables 3 and 5, and random effects are in Tables 4 and 6.

36

.BEBU we? :08 not? Sou «Eon

2:: 2: Bob 03%.? some 5 owcmno 09226 05 Soto.— mouoOm c2833 :82 H8582 mew—850 bio—.96 .8 _ 08E.
8 0332 use 63% RES—238 can 32%;.“ ..om 05.3.8 8 0358 cater—9 5595-553 80:8 888 =2§>oo a

dos—S, 8:082 88052 888 3.5a .8393 use ngo—onoxma com 8255c 9356 use 288 _ 25,—. a

 

 

 

33 So- as 3.? 38 8d- cod cod smooaoauwaesno £886 Hansen
6.0 3o- 3o who- So Nod. 23 85 ESBESHEQ Basso
m2 5o- o3 wed- mg 33- 8.0 86 93328936? BESQ
2: New A: w? m: 8H m3 m3. encobsgo
Nod 43 3o 8N Bo 8a a3 3N cassava
who m2 m2 8H ”no 3; So 8d 805823258 £885
owe a: mg m: moo N: Go A: was“ Saba
m2 GB 8.0 a3 3o MS a3 2% mafia aomweofiaa
mm 2 a. 2 S. 2 a. 2

92 u 5 32 u 5 d: u S @215

4 use A 05:. m use _ as: 83%; 382

 

22.8% 235 2V 36.83 anotoEoQ 33328.0 more .333 fleets; N beam.

fl 2an

37

Coho mafia—.5.— eo Sec 36%.: 8 26 $2: 8 83. Ho: .38 mowfieooba oEom a

 

 

8.2. g 8.3 n: on? me See 2: 02
SN 8 8.: am 43A mm. 3.8 a new
a Bazaars 5T5 “snag—9,5
8.: 8 3mm 2 a: 2 Ron 2 30% ea owees 28m
on: 9. 3% am :._m a. 3.8 Na soaaoeaesoofimfim
3.8 a e..: 8 3.: mm 3:: em 3202? Heafi
«cosmozcm
2.3 GA 9.3 mm 8.3 8 $3 3 333388 oz
8.: a a: 3 RS 8 come a: QEEESES
«Eocene—mam

.x. a .x. a as 2 as s

 

$.03 mmdmm oofimm hwdom mndmm afivmm Emmom deom oEooErAEEoE Emsmnaa

 

Qm SN QM. 3 9w 3 QM. 3N

 

A5 n5 a2 us a: is 3215
e 08:. m can. N as: _ use mozea> 382

 

.eSom 3&5 3x emote 835.8; Nohtob Ro\§.~o§zw§k ~38 SENSE 3.23.8qu N 393%
N 033.

38

Predicting Depression

The unconditional model. An initial model was run containing intercept and slope
terms, both of which were estimated as random effects. There was no significant
variation from woman to woman in the slope of depression over time; therefore, the slope
term was fixed (in this model and in all subsequent models) which did not significantly

change model fit (LR 12 (2) = .01, ns). On average, across all women, no linear change in

depression was detected over time (y = -.003, ns), while controlling for the effect of the
experimental condition on baseline depression (y = .078, ns) and change in depression
over time (y = -.005, ns). Therefore, a quadratic term was added to the model to assess
for non-linear change in depression, and this effect was also non-significant.

Control variables model. Adding women’s race at level two and adjusted monthly
income, employment status, and level of education at level one marginally improved

model fit over the unconditional model (LR 12 (6) = 11.83, p = .07). When controlling for

all other predictors, race was unrelated to Time 1 depression (y = .062, ns) or to change
in depression over time (y = -.004, ns) in this model and in all subsequent models.
Within-woman variation in income (y = .000, ns), education (less than high school = y =
.163, ns; high school or trade school = y = .025, ns), and employment status (y = -.060,
ns) was unrelated to changes in women’s depression over time in this model and in all
models hereafter.

Abuse model. The addition of abuse and women’s involvement with their
assailants significantly improved model fit over the control variables model (LR x2 (7) =
30.62, p < .001). When controlling for experimental condition and women’s

demographics, neither baseline physical (y = .164, p = .07) nor psychological abuse (y =

39

.055, ns) significantly predicted Time 1 depression; however, the effect of physical abuse
approached statistical significance. Baseline physical (y = .000, ns) and psychological
abuse (y = .005, ns) were also unrelated to change in depression over time. Within-
woman change in physical abuse was unrelated to change in depression over time (y =
.060, ns). However, change in psychological abuse was positively related to change in
depression (p < .01); on average, a l-unit increase in psychological abuse, relative to
baseline, was associated with a .129-unit increase in depression. Variation in women’s
involvement with their assailants was unrelated to change in depression over time in this
model and in the subsequent model (y = -.008, ns).

Difliculty obtaining resources model. The addition of difficulty obtaining
resources significantly improved model fit, over the abuse model, with LR )(2 (3) = 62.04,

p < .001. While controlling for all aforementioned predictors, between-women effects of
variability in Time 1 difficulty obtaining resources was positively associated with Time 1
depression (p < .001), but was unrelated to change in depression over time (7 = -.002,
ns). At Time 1, a l-unit increase in perceived difficulty was associated with a .543-unit
increase in depression. Within-woman variation in difficulty obtaining resources
predicted change in depression over time (p < .001); on average, a l-unit increase in
perceptions of difficulty obtaining resources, relative to Time 1, was associated with a
.338-unit increase in depression. 1

Testing difi‘iculty obtaining resources as a mediator. Time 1 perceptions of
difficulty obtaining resources was not found to mediate the relationship between physical
abuse at baseline and depression at Time 1 (y = .055, ns); however, a trend-level indirect

effect was detected between baseline psychological abuse and Time 1 depression through

40

Time 1 difficulty obtaining resources (indirect y = .070, p = .09). Time 1 difficulty
obtaining resources did not mediate the relationship between baseline levels of physical
(y = -.000, ns) or psychological abuse (y = -.000, ns) and change in depression over time.
While within-woman variation in difficulty obtaining resources did not mediate the
relationship between changes in physical abuse and depression over time (y = .002, ns),
the time-varying effects of difficulty obtaining resources firlly mediated the relationship
between changes in psychological abuse and depression over time (y = .048, p < .01).
That is, the prior direct significant relationship between within-woman variation in
psychological abuse and change in depression over time became non-significant after the
addition of the time-varying effect of difficulty obtaining resources to the model.
Predicting Quality of Life

The unconditional model. An unconditional model containing random intercept
and slope terms was run. The random effect of the slope for time, assessing variation
from woman to woman in change in quality of life over time, was marginally significant
(p = .08); therefore, this random effect was retained in all subsequent models. However,
on average, across all women, no significant linear change in quality of life was detected
over time (y = .005, ns), when controlling for the effects of experimental condition on
Time 1 quality of life (y = -.O24, ns) and on change in quality of life over time (y = -.006,
ns), which were not significant in this model or any model hereafter. A quadratic term
was added to the model to assess for non-linearity in quality of life over time; this effect

was not significant and therefore not included in subsequent models.

41

.ooo. v a ...: .S. v a ...... .o_. v a s ”N :33 a afieefia o: u z x .23 a £25.88 mom "2 s32

 

 

 

ooo. ...:omm. 80582 9:588 9355
ooo. ooo. ooo. ooo. 33% Saba
2o. .ooo. 3o. toe. 33o aomwoafiao
ooo. ooo.- ooo. ooo.- sesame SE 2080235
ooo. ooo. to. :o. no. mo. 623 use .5 623. SE aeaooom
ooo. coo _. o2. ooo. o2. m2. 6&8 Ema 55 a3 aeaeem
ooo. ooo. ooo. ooo. ooo. ooo. @885 A388 333.
So. ooo.- So. ooo.- ooo. ooo.- . was... 52335
2:: $5 commmoaov E owano co mooobo Swanéfiza
moo. ooo.- 32:82 SEES £856 _ 08:.
ooo. ooo. ooo. ooo. 33o Saba 0:20am
ooo. 3o. moo. ooo. Soon aoaeofiao 8:33
ooo. ooo.- ooo. ooo.- ooo. ooo.- 8am
moo. ooo.- ooo. ooo.- moo. ooo.- moo. ooo.- 8228
ooo. moo. moo. ooo. ooo. ooo.- moo. ooo.- as: as «new 325 omega.
:ommmoaow mo Ono—m Ron: 2: :o moooto :oEoBéooEom
to. the. $0582 mango £856 _ use
ooo. o2. ooo. .42. 08% samba 2:33
so. ooo.- ooo. ooo. SBA aoaeofiao 8:33
So. So. moo. ooo. ooo. moo. 8am
woo. woo. woo. ooo. ooo. ooo. ooo. ooo. 52250
So. ......LmoN oo_. ...:oooa o2. ...:oooa ooo. ......ooo.~ €853 owes;
Sammy—mow — 08:. :0 308mm coEo3rcoo>Bom
mm. o H. no H. s mm or
mooSOmDMm @80me onmm
@338
bgoamfl omnn< flobcoo HGOBEEOOGD

 

tofimogRoQ Matomeoxm £38: 3282 \Sotzozuom. Ex emote Sootmwm hmbk N SEW

m 2an

42

..oo.vq:. .o_

. v Q o “N 354 on SSEBES mm: H 2 x 651.8 838E030 Nmm n 2 .202

 

6on Sta 3 9538

 

 

 

 

1* 39% *1, 293 o 3%: -- 79 8:80th 296?:5 «a
maomEoQEoU $.82

-- 0.2. -- v2. -- o2. -- NE. 8593 132m“: _ 3R5
r: so ...: mom. ...: oom. t; ox. 35333 a8§5
823.5» oo§tw> ooqoto> oo§cw> mooobm 8098M
38:83“

@858

£385 822 $8280 12.232833

22.23me MEBNEER

fluke: E282 332:»:me 5x .2986 Ecutamseb 33: 38 Sofiwm 53282 N 365%

v 2an

43

Control variables model. The addition of women’s race at level two and adjusted
monthly income, employment status, and level of education at level one did not

significantly improve model fit over the unconditional model, when predicting quality of

life (LR x2 (6) = 9.26, ns). When controlling for all other predictors, race was not related

to women’s quality of life at baseline (7 = -.039, ns) nor to change in quality of life over
time (y = .001 , ns) in this model or in all subsequent models. Within-woman changes in
employment (7 = .164, p = .09), education (less than high school = y = .166, ns; high
school or trade school = y = .116, ns) and adjusted monthly income (y = .000, ns) were
unrelated to changes in women’s quality of life in this model, although the effect of
change in women’s employment status approached statistical significance levels.

Abuse model. The addition of abuse and women’s involvement with their

assailants significantly improved model fit, over the control variables model (LR x2 (7) =

57.64, p < .001). When controlling for condition and all aforementioned demographic
variables, baseline levels of physical abuse were unrelated to Time 1 quality of life (7 = -
.190, ns), and the effect of baseline psychological abuse on Time 1 quality of life
approached statistical significance (y = -.310, p = .06). Between-women effects in
baseline physical abuse predicted change in quality of life over time (p < .05); a l-unit
higher physical abuse score was associated with a .027-unit lower quality of life score.
The effect of baseline psychological abuse on change in quality of life was not significant
(y = .006, ns). Within-woman change in psychological abuse was associated with change
in quality of life over time (p < .05); on average, a l-unit increase in psychological abuse,
relative to baseline, was related to a .193-unit decline in quality of life. Change in

physical abuse was unrelated to change in quality of life over time (y = -.099, ns).

44

Within-woman change in involvement with the assailant was unrelated to change in
quality of life over time in this model (7 = -.113, ns) and in the following model.
Difficulty obtaining resources model. The addition of difficulty obtaining

resources significantly improved model fit, over the abuse model, with LR x2 (3) =

109. 12, p < .001. When controlling for all the aforementioned predictors, women’s
reported Time 1 difficulty obtaining resources was significantly associated with Time 1
quality of life (p < .001), but was unrelated to change in quality of life over time (7 =
.005, ns). At Time 1, on average, a l-unit higher difficulty obtaining resource score was
associated with a 1.125-unit lower quality of life score. Within-woman variability in
difficulty obtaining resources was significantly associated with change in quality of life;
on average, a l-unit increase in difficulty obtaining resources, relative to Time 1, was
associated with a .716-unit decrease in women’s quality of life (p < .001).

Testing difficulty obtaining resources as a mediator. Time 1 difficulty obtaining
resources did not mediate the relationship between baseline physical abuse and Time 1
quality of life (y = -.114, ns); however, there was a trend level indirect effect detected
between baseline psychological abuse and Time 1 quality of life through Time 1
difficulty obtaining resources (indirect y = -. 145, p = .09). Time 1 resource constraints
did not mediate the relationship between baseline physical (y = .001, ns) or psychological
abuse (y = .001, ns) and change in quality of life over time. Although the time-varying
effects of difficulty obtaining resources did not mediate the relationship between change
in physical abuse and quality of life over time (y = -.OO4, ns), within-woman variation in
perceived difficulty obtaining resources fully mediated the relationship between time-

varying psychological abuse and change in quality of life over time (y = -.103, p < .001).

45

.ooo. v a ...: .3. v o. .. .2. v a . mm 33 a efiofifio o: u 2 x .33 a 223288 mm H z .soz

 

 

 

_:. .2...er 38:88 we??? .3355
mm _. no _ .- mm _. ooo.- 33m EommbE
who. moor to. ... mo ~ .- omsnm Eomwgonommm
.1 _. S 7 03. 2 ~... 28:88 :23 2080235
ooo. no. we. or. me. 2 _. 328 coed s 323 :me ”seesaw
v3. om o. o5. comm. of. of. 80:8 nwE 55 $3 “cosmosom
ooo. ooo.- ooo. ooo. ooo. ooo. 0802: >262: Bow: €<
ooo. ...R o. moo. .8 o. ooo. o 42. Show anoanm
mo: mo £35: 5 omega so Saute 5803-:E:>?
ooo. moo. 32:88 wEEoBo b_soE% _ oEEt
NS. ...omor m 5. {Nov 35% Rumba 25on
So. woo. :o. ooo. 83m Homeofixma cam—03m
woo. moo. woo. woo. ooo. So. 8mm
ooo. o _o.- ooo. ooo.- ooo. ooo.- ooo. ooo.- :oEocoU
woo. voor woo. o_o.- woo. woo. woo. moo. 3:6 .56 one—m 30:5 owfio><
no: .«o 5:95 oo one? .82.: 05 so Somme :oEOBéooBBm
2 o. ***mm_._- 82:88 wcmfifino bioEE _ 08:.
o2. boo... mm _. ooo.- 03% Roofing 25on
2:. $2.- Sm Ho_ m.- 82% HomwoBsozmm 250QO
3:. who; on m moor o3. omoc mood
on _. 2 _. oo _. 3o. on _. Soc E _. vmo... 52980
9L. frivobé EA. 52.23thv R _. ....Ih—wcw 3 ~. ...:hwocn Aumoouowfiv owwuo><
co: «0 3:36 _ 08:. co moooto :oEo3-:oo§om
mo. of mm o mm of mm »
mooSomom Boobm team
@2330
.3355 om=n< £8800 3529883

 

@3st bang mfitmoogm €303 3282. amozzmzoom 3x .3683 38% 383% N buxom.
m 2an

46

.25. v a ...: .3. v m ... .2. v a H a 325 a afiaogg w: u 2 2 .23 a 23320on mm M 2 .232

 

688 SE 8 3:38

... ... ... o3 .og ... ... ... omofim momo -- “m8 oocouobmo 0833-20 ~3

 

mnemgafiov E52

 

 

 

-- SN. -- 3N. -- Rm. -- «NM. 8532.832 o :33
.. 8c. .. o8. ooo. . ooo. 28:5 2.2.... 25
...: E. ...: So. ...: ax. ...: Em. Aoéogeaoeas
8593/ 352:; oo§§> 028mb; mooobm Sousa/m
mooSOmoM

@2630

£355 822 flobcov 1.832023

$3M.» 33$

MSBNEEQ. “.33: E882 \SSEmaomw 3x $6.83 atomtcmfiob 3302 BS SumKQ 53382 N 335m.
o 2nt

47

Discussion

This study examined the relationship between IPV, resource constraints, and
women’s psychological well-being over time. In an effort to better understand the
processes through which abuse impacts women’s psychological well—being, resource
constraints was examined as a partial explanatory factor of survivors’ compromised
mental health and quality of life. The results provide evidence to support prior research
concerning the negative relationship between IPV and psychological well-being, and
expand upon the existing literature by broadening our understanding of one of the many
mechanisms through which abuse contributes to the development of pernicious
psychological outcomes.

When examining the long-term effects of abuse on well-being, results revealed
that survivors’ experiences of physical and psychological abuse at baseline were
unrelated to changes in depression over time. Abuse appears to have relatively short-term
effects on women’s trajectories of depression. Generally, when abuse subsides,
depressive symptomatology also subsides, a finding congruent with several prior studies
examining women’s depression longitudinally (Anderson et al., 2003; Campbell, J. &
Soeken, 1999; Campbell, R. et al., 1995). Interestingly, however, earlier experiences of
physical abuse seem to exert a long-term impact on women’s quality of life. It is possible
that another intervening variable may be at play here. For instance, severe physical abuse
that women endure may give rise to significant long-term injuries, which in turn may
affect women’s quality of life over time, as injuries may preclude women from living the
kind of life they had hoped for and engaging in the type of activities they value and

enjoy.

48

In addition to exploring the long-term effects of baseline physical and
psychological abuse on well-being, time-varying effects were also examined. Ongoing
physical and psychological abuse were not found to uniquely predict women’s quality of
life or depression, as hypothesized. Rather, when both physical and psychological abuse
were included as predictors, only psychological abuse continued to exert an impact on
women’s psychological well-being over the course of the study. At times when
psychological abuse was high, depression was also high, and quality of life was low. The
lack of significant influence of physical abuse on women’s psychological well-being may
reflect limited variability in physical abuse over the course of the study. Recent abuse
was a condition of study involvement; therefore, women’s experiences of physical and
psychological abuse were consistently high at baseline, and then after baseline, physical
abuse largely declined. Psychological abuse, however, remained more frequent and more
variable over the course of the study. It is also possible that psychological abuse may
have been far more damaging to women’s psychological well-being than physical
violence (Follingstad et al., 1990), as some studies suggest that psychological abuse is a
stronger predictor of well-being (Mechanic et al., 2008; Pico-Alfonso et al., 2006). This
speaks to the importance of examining the multi-dimensional nature of abuse when
attempting to understand survivors’ psychological well-being.

To unravel the complexity of post-abuse sequelae, it is paramount that researchers
attend to the processes and mechanisms which underlie the abuse-distress relationship.
The difficulty survivors reported obtaining community-based resources was an important
explanatory mechanism to consider. Notably, women’s initial resource constraints did not

mediate the relationship between baseline levels of physical or psychological abuse and

49

change in psychological well-being over time. However, this study offers preliminary
evidence to suggest that women’s earlier experiences of psychological abuse may
contribute to their subsequent difficulties obtaining resources; albeit the effect of baseline
psychological abuse on women’s resource constraints eight months later did not reach
conventional levels of significance. It is likely that earlier effects of psychological abuse
may diminish in intensity as time passes, suggesting that smaller time intervals between
measurement points are needed in future studies. It also appears as if women’s resource
constraints have rather immediate effects on their mental health and quality of life and
that once these constraints are lifted, women return to their prior level of psychological
well-being.

To add to our understanding of how these relationships play out over time, this
study also investigated the time-varying effects of abuse, resource constraints, and
women’s psychological well-being. At times when psychological abuse was high,
survivors’ reported difficulties obtaining resources was also high, lending support for the
extant literature which contends that assailants deplete women’s access to resources using
a variety of abusive strategies (Adams et al., 2008; Goodkind et al., 2003; Leone et al.,
2004; Ptacek, 1997; Raphael, 1996; Swanberg & Logan, 2005). Further, ongoing
difficulties acquiring community-based resources were associated with negative
psychological sequelae over time as found in prior research (Bybee & Sullivan, 2002;
Schumm et al., 2004), and as expected, the relationship between ongoing psychological
abuse and change in quality of life and depression was fully explained by ongoing

resource constraints.

50

The primary motivation of assailants’ abusive actions is to dominate and control
their partners’ thoughts, feelings, and behaviors. Tactics of psychological abuse (threats
of physical abuse, ridicule, isolation, intimidation, among many others; Follingstad et al.,
1990; Stark, 2007) appear to be the primary mechanisms of control employed by
assailants to restrict women’s activities and limit their access to community-based
resources. As Dutton and Goodman (2005) contend, assailants use tactics of
psychological abuse and coercive control to create and exploit vulnerabilities in women,
and wear down survivors’ abilities to resist their demands by depleting their tangible,
social, and personal resources. Survivors’ access to resources and community supports
that would facilitate their independence and protect them from reabuse (Bybee &
Sullivan, 2002, 2005) likely threaten assailants’ sense of control and counter their goals
of dominating their partners’ thoughts, feelings, and behaviors. Assailants, therefore,
appear to utilize psychological abuse, which women often report as most damaging
(Follingstad et al., 1990), to maintain power over their partners by limiting their access to
resources.

In accordance with COR theory (Hobfoll, 1988, 1989; Hobfoll & Lilly, 1993),
taken together, these findings suggest that survivors experience actual or threatened
resource loss as a result of the psychological abuse they endure, and these resource
constraints limit survivors’ abilities to cope with or escape the abuse, which in turn brings
about significant distress. Without the adequate resources to mobilize during times of
stress, individuals’ psychological well-being may be compromised, and as suggested by
COR theory the distress individuals experience may exacerbate future resource loss

(Johnson et al., 2007).

51

It is noteworthy that only psychological abuse was related to survivors’ reported
difficulties obtaining community-based resources. Again, it is possible that limited
variability in physical abuse over time may explain the lack of association between
physical violence and resource constraints. Alternatively, psychologically abusive control
tactics used by assailants may be a more powerful tool than physical violence in limiting
survivors’ access to community-based resources. Whether it is primarily psychological
abuse or a host of abusive strategies collectively that restrict survivors’ access to
resources, this study brings to light the serious far-reaching implications of abuse and
demonstrates the need for advocates, researchers, and practitioners to identify strategies
to reconnect survivors with the resources necessary to restore and promote their
psychological well-being.

Though the findings from this study enhance our understanding of the role
resource constraints play in the lives of survivors of IPV, several limitations must be
considered. First, the nature of the sample used in the current study limits generalizability
to all women with abusive partners. Second, the sample was of limited size, likely
affecting power to detect significant effects on the slopes of depression and quality of life
over time, which are typically small (Snijders & Boskar, 1999). Future research efforts
should replicate this study using a larger and more diverse sample of women, while also
retaining the current study strengths by using a longitudinal design and recruiting women
from diverse settings.

Third, the analytic approach utilized was limited in its ability to infer causality
among the time-varying covariates under study. More specifically, within each time point

it was not possible to determine the temporal ordering of events; therefore, it is unclear if

52

abuse caused the difficulties survivors reported in acquiring resources, and if these
difficulties led to increased depressive symptomatology and reduced quality of life.
Despite this limitation, this study did demonstrate how these variables changed together
over time, that is, how change in one variable was accompanied by change in another.

Finally, this study did not account for additional covariates that may have had an
impact on survivors’ abilities to acquire community-based resources and on their levels
of distress over the course of the study. For instance, it is possible that exposure to other
traumatic events, outside one’s interpersonal relationship, one’s sense of self-efficacy, or
even preexisting mental health conditions, may have impacted one’s ability to procure
requisite resources, and may have also affected reports of depressive symptomatology
throughout the study. The difficulties women reported obtaining resources may also
reflect exosystem influences (Bronfenbrenner, 1979), as communities may have limited
resources available for survivors to draw upon. Future research should expand upon the
current study by exploring these additional variables as predictors of survivors’
psychological well-being.

Despite these limitations, this study offers several important implications for
future research and practice. Future research is needed to improve upon our
understanding of the relationship between abuse and the difficulties survivors have
acquiring resources. For instance, more attention could be directed to examining the
direct actions and tactics of psychological abuse employed by assailants (e. g., sabotage
efforts used to limit women’s access to employment and educational opportunities,
strategies that isolate women from their social support networks) that are most damaging

to survivors’ abilities to procure resources. Understanding the exact barriers women are

53

faced with as they attempt to acquire resources will greatly inform future intervention
efforts.

In addition, because physical abuse became so infrequent after baseline, this study
should be replicated with another sample of women with abusive partners over time to
determine if physical abuse may in fact also explain the difficulties women report
acquiring resources. In addition to examining physical abuse, additional research must
also account for the injuries women suffer as a result of the abuse they experience, as
injuries may also limit women’s abilities to acquire resources.

Further, because it appears as if psychological abuse may continue to exert an
impact on survivors’ access to resources over time, this study should be replicated using
shorter time intervals between measurement points that allow for a more sensitive
assessment of the effects under study. Replication efforts of this nature should also
increase the number of measurement points utilized to allow for a lagged design in which
the independent variable precedes the mediator, which precedes the dependent
variable(s). Finally, more attention should be directed to examining additional ecological
factors that may help to explain the process by which abuse impacts survivors’
psychological well-being over time. For instance, abuse may lead to survivors’ economic
dependency on their assailants or even to poverty; therefore, this is one potential
mechanism, among many, that should be investigated to further our understanding of the
pathways through which abuse affects survivors’ psychological well-being.

Violence against women advocates, researchers, and providers are dedicated to
helping women overcome the physical, psychological, and material hardships caused by

abusive partners. It appears from this study that women’s psychological well-being is

54

closely tied to their material well-being by way of resources. Thus, to help restore
survivors’ mental health and overall quality of life, collective efforts on the part of
community-based providers, advocacy organizations, and clinicians are needed to
improve women’s access and availability to community-based resources. Additionally,
outreach efforts are needed to increase survivors’ awareness of resources and supports
available within their communities. Outreach efforts must extend beyond those
organizations and agencies already providing services to survivors to ensure that those
women who have not sought help are also aware of available resources. For instance,
large-scale media campaigns may be used to inform community members of available
resources. Additionally, information should be made available at community health
providers, local religious organizations, and other community venues frequently utilized
or visited by survivors.

Service providers must also work together to ensure that care is holistic, enabling
survivors to acquire all the resources (e. g., housing, employment, medical care, mental
health care, child care, transportation) necessary to rebuild their lives and become free
from their abusive relationships. Fragmented care will only prolong survivors’ distress
and may even force women to return to their assailants to fulfill their most basic needs.
For instance, helping women find housing is insufficient assistance if they do not have
the financial means necessary to sustain this place of residence. Providers must work with
survivors to comprehensively assess their needs, identify sustainable resources to meet
these needs, and see them through until they have acquired all the resources necessary to
reestablish their lives, independent of their abusive partners. In addition, while initiating

and providing care, practitioners must be cognizant and sensitive to the difficulties

55

survivors may be experiencing immediately following their abusive relationship.
Survivors must be provided adequate time, counseling services, and the resources
necessary to overcome the deleterious mental health consequences of abuse.

This study highlights the importance of reconnecting survivors with requisite
resources that may have been lost as a result of their abusive relationships, in an effort to
restore their psychological well-being. While this study unveiled one of many processes
underlying the relationship between abuse and survivors’ psychological distress, far more
empirical attention is needed to identify additional ecologically-based explanatory
mechanisms that may aid in the development of community-based interventions to assist
survivors in overcoming the negative psychological impacts of abuse. Research of this
nature is paramount to ensure the most appropriate and applicable intervention
approaches are utilized to facilitate survivors’ recovery from the devastating effects of

abuse.

56

Chapter 3
STUDY 2:

AN EXAMINATION OF THE IMPACT OF NEIGHBORHOOD FACTORS
ON THE WELL-BEING OF SURVIVORS OF INTIMATE
PARTNER VIOLENCE OVER TIME
Overview of the Literature

Millions of women are subjected to violence at the hands of their intimate
partners each year (Tjaden & Thoennes, 2000). Survivors of intimate partner violence
(IPV) often experience psychological abuse, and actual or threatened physical or sexual
violence (Saltzman, F anslow, McMahon, & Shelley, 2002) which may significantly
compromise their overall physical health and psychological well-being (Campbell, 2002;
Golding, 1999). Researchers attending to the precipitating factors or effects of IPV have
largely emphasized individual or interpersonal characteristics, in lieu of more
contextually-based ecological factors (Bogat et al., 2005; Miles-Doan, 1998; Raghavan,
Mennerich, Sexton, & James, 2006).

The ecological framework underlying the current study emphasizes the
importance of understanding individuals’ functioning and development within the context
of their surrounding environments (Bronfenbrenner, 1977, 1979). According to
Bronfenbrenner’s theoretical model, individuals are nested within several larger social
systems (e. g., within families, neighborhoods, communities, and broader societies), each
of which exerts influence over individuals’ lives. Embedding survivors’ experiences of
abuse within the broader context within which they live may aid and advance our

understanding of the effects that commonly follow IPV, including serious negative

psychological sequelae (e.g., Golding, 1999). These broader ecological factors are

57

significant to our understanding of survivors’ experiences as they may directly contribute
to individuals’ psychological distress or may inhibit individuals’ abilities to overcome
significant distress (e.g., Nicolaidis, Curry, McFarland, & Gerrity, 2004; Nurius et al.,
2003)

To date, limited research has examined survivors’ broader contextual
environments as contributing factors to their compromised mental health and well-being.
Using a longitudinal design, the current study examined whether perceived neighborhood
disorder and related fear of victimization, due to the surrounding environment,
contributed to survivors’ negative psychological sequelae, in addition to the abuse they
had experienced.

Intimate Partner Violence and Psychological Well-Being

An extensive amount of research exists linking IPV to deleterious mental health
outcomes, such as depression, anxiety, post-traumatic stress disorder (PTSD), and
suicidal ideation (Cascardi, O’Leary, & Schlee, 1999; Dutton, Kaltman, Goodman,
Weinfurt, & Vankos, 2005; Golding, 1999; Helfrich, Fujiura, & Rutkowsli-Kmitta, 2008;
Leiner, Compton, Houry, & Kaslow, 2008; Pico-Alfonso et al., 2006). Research
consistently finds that prevalence rates of mental health disorders among survivors are
substantially higher than those of non-abused women (Coker, Smith, Bethea, King, &
McKeown, 2000; Houry, Kemball, Rhodes, & Kaslow, 2006; Theran, Sullivan, Bogat, &
Stewart, 2006). Additionally, while much less frequently studied, IPV has also been
found to have significant negative implications for women’s overall quality of life
(Adeodato, Carvalho, de Siqueira, & de Matos e Souza, 2005; Leung, Leung, Ng, & Ho,

2005)

58

Generally, the more severe and frequent the abuse, the more significant the impact
on women’s well-being (Campbell, Kub, Belknap, & Templin, 1997; Dienemann et al.,
2000; Dutton et al., 2005; Golding, 1999; Houry et al., 2006; Pico-Alfonso et al., 2006).
Within abusive relationships, assailants use multiple forms of violence (e. g., physical,
psychological), each of which may contribute to women’s compromised well-being
(Dutton, Goodman, & Bennett, 1999; Marshall, 1999; Mechanic, Weaver, & Resick,
2008; Sackett & Saunders, 1999). Cumulative exposure to multiple forms of violence,
however, appears to be most damaging to women’s overall mental health (Houry et al.,
2006; Ludermir, Schraiber, D’Oliveira, Franca-Junior, & Jansen, 2008). For instance,
Houry and colleagues (2006) compared women who had experienced physical, sexual,
and emotional abuse concurrently, with those who had either experienced no abuse, 1
form of abuse, or 2 forms of abuse. Concurrent exposure to all forms of IPV was
associated with the greatest odds of manifesting symptoms of depression, PTSD, and
suicidal ideation.

The majority of research investigating the psychological well-being of survivors
of IPV has been cross-sectional. This cross-sectional work suggests that recency and
continuity of abuse are particularly harmful for survivors’ mental health (Bogat,
Levendosky, Theran, von Eye, & Davidson, 2003; Bonomi et al., 2006). However,
findings from longitudinal research are somewhat more equivocal. In some studies, as
violence decreases over time, well-being improves for most women (Campbell, R.,
Sullivan, & Davidson 1995; Kemic, Holt, Stoner, Wolf, & Rivara, 2003), while in others

the negative psychological implications of abuse remain steady over time (von Eye &

59

Bogat, 2006), or initially decline and later return, despite an overall drop in abuse
(Campbell, J. & Soeken, 1999).

One important consideration, often overlooked in existing longitudinal studies, is
the extent to which women experience additional stressors over the abuse, and how these
factors may help explain survivors’ psychological well-being. In addition to the level of
abuse women experienced, Anderson, Saunders, Yoshihama, Bybee, and Sullivan (2003)
examined women’s secondary stress levels (e. g., family responsibilities, loss of income)
on two-year depression trends among a sample of sheltered women who were separated
from their abusive partners. In this study, depression declined over time for women who
experienced low levels of violence and secondary stress when exiting the shelter;
however, for women who experienced high levels of violence and secondary stress at
shelter-exit, depression remained stable over time, or for some women, even increased.

Additional stressors may play important roles in the lives of women with abusive
partners, as these factors may cumulatively contribute to the development or continuation
of psychological distress. Empirical attention to the multitude of ecologically-based
stressors women experience may be essential to our understanding of how the
relationship between abuse and mental health develops and changes over time.

Intimate Partner Violence within Disadvantaged Neighborhoods

IPV is a widespread social problem impacting women of all ages, and across all
racial, ethnic, and socioeconomic strata (Menard, 2001). However, several studies have
found that prevalence rates of IPV are higher in resource-deprived impoverished
neighborhoods, and among families falling within low socioeconomic stratum (Benson,

Fox, DeMaris, & Van Wyk, 2003; Breiding, Black, & Ryan, 2008; Cunradi, Caetano,

60

Clark, & Schafer, 2000; Miles-Doan, 1998; Thompson et al., 2006). It remains unclear
whether economic deprivation or neighborhood deterioration is directly responsible for
higher rates of IPV, if women’s likelihood to report IPV may differ by their broader
position within society (Benson et al., 2003; McCue, 2007), or if the most commonly
researched form of abuse in studies of IPV (physical violence) is more frequently or
severely used by batterers against low-income versus middle- to upper-class women (e. g.,
Holtzworth-Munroe, Smutzler, & Bates, 1997).

Although IPV is certainly not exclusive to impoverished families, the elevated
prevalence rates among low-income women give rise to significant concern, given the
additional sources of stress individuals in disadvantaged neighborhoods are plagued with
in their day-to-day lives. Disadvantaged neighborhoods are characterized by extreme
rates of poverty, crime, unemployment, substance abuse, residential instability,
dilapidated structural conditions, and a lack of economic and social resources (Boardman,
Finch, Ellison, Williams, & Jackson, 2001; Hill & Angel, 2005; Krivo & Peterson, 1996;
Latkin & Curry, 2003; Raghavan et al., 2006; Steptoe & Feldman, 2001; Zule et al.,
2008). Many of the aforementioned indicators of neighborhood disadvantage are
significantly associated with negative psychological sequelae and low overall quality of
life among residents (Boardman et al., 2001; Hill & Angel, 2005; Matheson et al., 2006;
Mulvey, 2002; Ross, 2000; Silver, Mulvey, & Swanson, 2002; Steptoe & Feldman, 2001;
see Truong & Ma, 2006 for a review).

Neighborhood Disorder and Psychological Well-Being
Within disadvantaged neighborhoods, multiple mechanisms are at work to impact

individuals’ psychological health and well-being. One such mechanism is the way in

61

which individuals perceive their environments (Christie-Mizell, Steelman, & Stewart,
2003; Curry, Latkin, & Davey-Rothwell, 2008; Haney, 2007; Ross, 2000; Ross &
Mirowsky, 1999), and how these perceptions are manifested psychologically. Perceptions
may be shaped in multiple ways, and influenced by the complexity of the processes
inherent in the context in which individuals are embedded. For instance, inadequate
social and economic capital within deprived communities weakens the governing
structures and social norms pivotal to maintaining social order, civility, and residential
safety (Cutrona et al., 2005). This breakdown of social control is manifested through
various social ills and indicators of societal unrest which are visible to residents and
influence their perceptions about their environments (Ross & Mirowsky, 1999; Ross,
Reynolds, & Geis, 2000). These collective visual indicators or social cues are termed
neighborhood disorder, and are thought to signify to residents a breakdown of social
order (Ross, 2000; Ross & Mirowsky, 1999; Skogan, 1990; Skogan & Maxfield, 1981)
and the potential for threat or danger (Hill & Angel, 2005; Xu, Fiedler, & Flaming,
2005).

Neighborhood disorder encompasses both social and physical incivilities (Skogan,
1990; Skogan & Maxfield, 1981). Generally, social disorder refers to behavioral
indicators of community deterioration (Skogan, 1990), such as public drinking or
intoxication, fighting, loitering, gang-related activity, panhandling, drug use and sales,
neighbor-to-neighbor conflict, prostitution, and harassment (Ross & Mirowsky, 1999;
Skogan, 1990; Zule et al., 2008). Physical disorder refers to signs of degradation within
communities, including abandoned structures, deteriorated homes, vandalism, graffiti,

abandoned cars, vacant lots, and uncontained litter (Franklin, Franklin & Fearn, 2008;

62

Haney, 2007; Ross & Mirowsky, 1999). While conceptually distinct, researchers find that
social and physical incivilities generally underlie the same construct and can be measured
uni-dimensionally (Franklin et al., 2008; Ross & Mirowsky, 1999).

Neighborhood disorder is a significant factor to be considered when attempting to
understand the psychological distress of residents in disadvantaged areas. However, the
impact of neighborhood disorder on residential well-being, particularly among adults, has
been understudied, only gaining empirical attention in recent years (Cutrona, Russell,
Hessling, Brown & Murry, 2000; Cutrona et al., 2005). These studies, while mostly
cross-sectional, have found strong support for an inverse relationship between
perceptions of neighborhood disorder and individuals’ overall quality of life and
psychological well—being while controlling for individual-level factors, such as
socioeconomic status (Gary, Stark, & LaVeist, 2007; Haney, 2007; Hill, Ross, & Angel,
2005; Ross, 2000; Ross et al., 2000; Steptoe & Feldman, 2001; Xu et al., 2005).

Only a very limited amount of research has investigated the relationship between
perceived neighborhood disorder and psychological well-being over time, and studies
have generally been limited to two time points (e. g., Christie-Mizell et al., 2003; Curry et
al., 2008; Latkin & Curry, 2003). However, the aforementioned cross-sectional findings
are substantiated in these studies. For instance, when investigating the relationship
between perceived neighborhood disorder and depression over nine months among 818
adults involved in a health-based educational intervention, Latkin and Curry (2003) found
that, after controlling for baseline depression scores, perceived neighborhood disorder at

baseline predicted depression nine months later.

63

Research finds that perceptions of neighborhood disorder may be even more
influential than objective neighborhood factors (e. g., location, group level indicators of
disadvantage) in understanding residents’ psychological well-being (Christie-Mizell et
al., 2003; Curry et al., 2008; Hadley-Ives, Stiffman, Elze, Johnson, & Dore, 2000; Ross,
2000; Ross et al., 2000). For instance, Curry and colleagues (2008) examined the
longitudinal impact of community violence (using objective crime data obtained from
police records and geocoded at the block level) among inner-city residents on the
development of depression. They found no direct relationship between community
violence and depression; rather, an indirect effect of community violence on depression
was revealed through perceived neighborhood disorder.

In another study investigating neighborhood location, perceived levels of
disorder, and depression among a sample of 2,204 parenting women, Christie-Mizell and
colleagues (2003) found that while objective location does impact perceptions of
disorder, only perceived disorder was predictive of depression, even after controlling for
socioeconomic status, race, health, and family structure (e. g., relationship status, number
of children). Similarly, Ross and colleagues (Ross, 2000; Ross et al., 2000) examined the
relationship among objective and subjective neighborhood factors, and mental health
among a probability sample of 2,482 Illinois residents. They found that the addition of
perceived neighborhood disorder to the prediction of depression and overall
psychological distress, while controlling for individual-level factors (e.g., age, gender,
race, socioeconomic status), rendered objective neighborhood characteristics (based on

census data) insignificant, despite a significant effect in an earlier model.

64

While objective indicators of neighborhood disadvantage are certainly not
inconsequential to individuals’ psychological well-being, residents’ subjective
assessments of their neighborhoods seem to be more salient to understanding their mental
health and quality of life. This may be because objective accounts of neighborhood
characteristics are often based upon neighborhood block-level data obtained from the
US. census, observational data coded at the neighborhood-level by external raters, or
aggregated residential experiences or perceptions across a pre-specified geographical
region, each of which happen to mask unique individual experiences. Implicit in these
types of data are the assumptions that: (a) residential experiences are only influenced by
the characteristics within the arbitrarily-delineated geographical boundaries of one’s
neighborhood (Dupere & Perkins, 2007), and (b) experiences within, and perceptions of
one’s neighborhood are universal across all residents. Researchers find that individuals’
perceptions not only vary within neighborhoods (Taylor, 2001), but in some instances are
unrelated to or only moderately correlated with objective indicators of neighborhood
disadvantage and disorder (Perkins, Wandersman, Rich, & Taylor, 1993; Wen, Hawkley,
& Cacioppo, 2006; c.f., Sampson & Raudenbush, 2004; Schieman & Pearlin, 2006).
Given the limitations of objective measures, it is important that research attends to
residents’ subjective assessments of their environments, as they “. . .go directly to the
heart of how people visualize their own circumstances” (Christie-Mizell et al., 2003,
p.405).

Residents’ perceptions of their broader environments are essential to
understanding the link between neighborhood disadvantage and mental health. These

perceptions are thought to be linked with mental health outcomes, through one’s

65

emotional response to his or her environment. Conceptually, perceived neighborhood
disorder is thought to signify to residents a breakdown in social order (Ross, 2000;
Skogan, 1990), which in turn leaves individuals feeling threatened or fearful about their
environments (Hill & Angel, 2005; Xu et al., 2005). A substantial amount of evidence
links neighborhood disorder with perceived risk and fear of both crime and victimization,
particularly among women (Kanan & Pruitt, 2002; LaGrange, Ferraro, & Supancic, 1992;
Perkins & Taylor, 1996; Ross et al., 2000; Ross & Mirowsky, 2001; Schafer, Huebner, &
Bynum, 2006; Wyant, 2008; Xu et al., 2005).

Three dominant theoretical models have been proposed to explain citizens’ fear of
crime: (a) the vulnerability model, (b) the disorder/incivilities model, and (c) the social
integration model (see Franklin et al., 2008 and Wyant, 2008). The vulnerability model is
based upon the theoretical tenet that individuals’ perceptions of personal vulnerability
(including physical vulnerability to assault, as well as social vulnerability via a lack of
material, social, and economic resources) may lead to fear of crime. The disorder and
incivilities model purports that manifestations of neighborhood disorder (physical and/or
social incivilities) give rise to feelings of fear and increased risk for victimization.
Finally, the social integration model posits that fear of crime is inhibited when residents
are socially integrated within their neighborhoods and feel a sense of belonging and
attachment to their communities. Franklin and colleagues (2008) tested each of these
theoretical models among a sample of 2,599 residents across 21 cities and found
perceived physical and social disorder (the disorder/incivilities model) to be the most

powerful determinants of citizen fear.

66

Fear of crime and victimization, and overall perceptions of safety, have important
implications for citizens’ psychological well-being, as each have beenrfound to relate to a
range of mental health disorders (e. g., anxiety, depression) and poor overall quality of life
(Green, Gilbertson, & Grimsley, 2002; Hill & Herman-Stahl, 2002; Kruger, Reischl, &
Gee, 2007; Stafford, Chandola, & Marmot, 2007; Xu et al., 2005; Zule et al., 2008).
Examining the impact of actual crime, neighborhood disorder, and fear of crime on
quality of life, Xu and colleagues (2005) found the only significant direct effects were
between citizens’ quality of life, and both perceived neighborhood disorder and fear of
crime. While not directly tested in this study, fear of crime may potentially serve as one
of several factors likely mediating the relationship between neighborhood disorder and
residents’ compromised mental health and quality of life.

The ecological context is an important determinant in understanding residents’
mental health and well-being. Prior cross-sectional and limited longitudinal research
suggest that perceptions of neighborhood disorder, and related fear of crime and
victimization, may be two factors contributing to the compromised mental health and
well-being of individuals within low-income disadvantaged neighborhoods.

Cumulative Adversity: Women ’s Experiences of Intimate Partner Violence and
Neighborhood Disorder

Low-income women with abusive partners may face additional adversity when
exposed to neighborhood disorder, given that IPV already places survivors at substantial
risk for negative psychological sequelae. Neither of these life stressors are generally
discrete incidents; rather, each represents a significant and chronic pathway of stress

which wears down individuals’ resistance and resources over time (Wheaton, 1999).

67

Therefore, concurrent or cumulative exposure to both'IPV and neighborhood
disadvantage may significantly undermine survivors’ psychological health and overall
well-being.

Stress theorists emphasize the importance of understanding overall well-being
within the broader context of individuals’ environments, considering individual-level
factors simultaneously with the influence of broader social systems (Lin & Ensel, 1989;
Pearlin, 1989). Given that both IPV and neighborhood disorder have important
implications for individuals’ well-being, it is important to assess these effects
simultaneously, and examine how these factors may cumulatively affect survivors’
psychological health.

Within the cumulative adversity theoretical framework, individuals’ well-being is
perceived as the byproduct of unfavorable persistent and/or additive circumstances
evolving over time (Hatch, 2005). When multiple stressors are at play, some may exert
influence on others, or may change the magnitude of the effect of others (Sandler, Braver,
& Gensheimer, 2000). The accumulation of multiple stressors may have additive or
multiplicative effects on individuals’ well-being, as the buildup of many stressors may
give rise to significantly greater consequences than those of the individual stressors
themselves (Mirowsky & Ross, 2003; Turner & Lloyd, 1995).

Within the lives of survivors of IPV, the additional stress brought on by perceived
neighborhood disorder and feelings of fear may cause significant further impairment in
psychological functioning, or may inhibit women’s abilities to overcome the distress they
experience as a result of IPV. To date, limited empirical attention has been directed to

understanding the impact of women’s experiences of IPV and the effects of their

68

neighborhoods on their overall psychological well-being, and even this work has
overlooked the impact of perceptions of neighborhood disorder.

Studies having bridged the gap between neighborhood effects and survivors’ well-
being have largely emphasized women’s exposure to community violence, over
perceptions of disorder. For instance, Bogat and colleagues (2005) investigated the
relationships among IPV, objective indicators of community violence (specifically
aggravated assault and disorderly conduct offenses occurring at the census block level),
and women’s mental health (anxiety, depression, and PTSD) among a community-based
sample of 94 women. Overall, women with no past-year history of IPV reported no
adverse mental health sequelae, regardless of the occurrence or frequency of past-year
community violence in their neighborhood(s). The co-occurrence of IPV and community
violence, however, was associated with anxiety, depression, and PTSD-related
symptomatology. The absence of a community-level effect on survivors’ mental health
likely reflects the way in which community violence was assessed. Aggregation of census
data to represent community-level experiences of residents does not directly reflect
individuals’ exposure to, nor perceived severity of, violence occurring in their
neighborhood(s).

Another cross-sectional study conducted by Clark and colleagues (2008)
examined community violence exposure among a sample of 386 women in an urban
community health center. After controlling for the impact of individual-level factors and
experiences of IPV, Clark and colleagues found that exposure to community violence
significantly predicted women’s depression and anxiety. Finally, among a community-

based sample of 90 low-income Afi'ican American women, Brown, Hill, and Lambert

69

(2005) examined the additive and multiplicative impact of community violence exposure
and IPV on the development of trauma-related symptoms cross-sectionally. They found
that community violence exposure significantly predicted trauma symptomatology, after
controlling for IPV. The authors then reversed the order of the predictors to assess the
impact of IPV, over the effect of community violence exposure, and based on these
findings concluded that both community and partner violence uniquely (thus, additively)
contribute to the development of trauma-related symptomatology, after controlling for the
other.

These findings, taken together, suggest that exposure to neighborhood violence,
as reported by residents, has important implications for individuals’ overall well-being, in
addition to personal victimization by an intimate partner. While community violence
exposure does not directly indicate perceptions of neighborhood disorder, researchers
have found a positive relationship between community crime and neighborhood disorder
(O’Shea, 2006; Skogan, 1990; Xu et al., 2005). Further investigation of salient
ecologically-based contextual factors within studies of IPV is warranted, as these factors
are likely to advance our understanding of how the effect of one’s contextual
environment may be implicated in the development of negative psychological sequelae
commonly experienced by survivors of IPV.

The Current Study

The current study investigated the concurrent and cumulative effects of IPV
victimization and neighborhood disorder on survivors’ psychological well-being over
time. In addition, to understand the process through which women’s perceptions of their

neighborhood contexts may impact their psychological well-being, fear of victimization

7O

was explored as a mediating factor. This study extends previous research in a number of
ways. First, little is known about how the neighborhood context, as a source of additional
stress, may impact survivors’ broader well-being, after accounting for women’s
experiences of abuse. This is particularly important to investigate among survivors with
limited economic resources who often reside in neighborhoods perceived to be high in
disorder. Identifying neighborhood-level effects, based upon women’s perceptions, has
important implications for future intervention strategies focused on enhancing women’s
broader resource and support networks (e. g., linking women with additional community-
based resources, enhancing women’s safety planning efforts, building cohesion across
neighborhoods), and is salient to our overall understanding of how best to assist survivors
in overcoming the deleterious mental health consequences they may experience.

In addition, this study examined the process through which neighborhood disorder
impacts survivors’ overall psychological well-being. While evidence exists suggesting a
link between perceptions of neighborhood disorder and fear of crime, and between fear of
crime and mental health, no study to date has examined fear of crime as a partial
mediating factor among a sample of women with abusive partners. Finally, one of the
most significant contributions of this study is the ability to delineate how these variables
may relate to each other over time.

The conceptual model underlying the current study is presented in Figure 2.
Specifically, two research questions were investigated: (a) Does perceived neighborhood
disorder impact women’s psychological well-being (depression and quality of life), over
and above their experiences of intimate partner violence .7, and (b) Does neighborhood

fear of victimization partially mediate, or explain, the relationship between perceived

71

neighborhood disorder and women’s psychological well-being (depression and quality of
life)? This study investigated both differences between women (between-women effects)
as well as how women’s individual experiences changed over time, relative to their own
past experiences (within-woman effects). The hypotheses below distinguish the between-
women and within-woman effects that were tested.
A) Between-Women Differences Predicting Change over Time
H1: Higher levels of baseline physical and psychological abuse will be associated
with an increased trajectory in depression and decreased trajectory in quality of
life over time.
H2: Higher levels of baseline perceptions of neighborhood disorder will be
associated with an increased trajectory in depression and decreased trajectory in
quality of life over time, over and above the impact of abuse at baseline.
H3: Higher levels of baseline perceptions of neighborhood disorder will be
associated with higher levels of baseline fear of victimization, over and above the
impact of abuse.
H4: Higher levels of baseline fear of victimization will be associated with an
increased trajectory in depression and decreased trajectory in quality of life over
time, over and above the impact of abuse at baseline.
H5: The relationship between baseline perceptions of neighborhood disorder and
changes in psychological well-being (depression and quality of life) over time
will be partially mediated by baseline levels of neighborhood fear of

victimization, over and above the impact of baseline abuse.

72

B) Within-Woman Trajectories of Change over Time
H6: Within-woman increases in physical and psychological abuse will be
accompanied by increases in depression and decreases in quality of life over time.
H7: Within-woman increases in perceived neighborhood disorder will be
accompanied by increases in depression and decreases in quality of life over time,
over and above the impact of abuse.
H8: Within-woman increases in perceived neighborhood disorder will be
accompanied by increases in neighborhood fear of victimization over time, over
and above the impact of abuse.
H9: Within-woman increases in neighborhood fear of victimization will be
accompanied by increases in depression and decreases in quality of life over time,
over and above the impact of abuse.
H10: The relationship between changes in perceived neighborhood disorder and
psychological well-being (depression and quality of life) over time will be
partially mediated by changes in neighborhood fear of victimization, over and

above the impact of abuse over time. '

73

Figure 2. A conceptual model depicting the impact of neighborhood disorder and the
mediating role of neighborhood fear of victimization on the psychological well-being of

survivors of intimate partner violence over time.

 

 

 

 

 

  

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:3 Baseline
@ Abuse
in
r:
0
E
o
. <
§ Baseline
5 Neighborhood
fi Baseline Fear 0'
Neighborhood Victimization
K Disorder - Change in
6 Psychological
Change in - Well-being
Neighborhood , .
a Disorder Change in
8 Neighborhood
E Fear of
5 Wctimization
s < ~
3
.5
35 Change in
3 Abuse

 

 

 

74

Method

Participants

The current study investigated the lives of 160 female survivors of IPV over time,
all of whom participated in a larger research study designed to test the efficacy of an
experimental family-based advocacy intervention for women and their children (see
Study 1 for a detailed description of the advocacy intervention). Several criteria were
used to determine women’s eligibility for this study. All women had to have been
physically victimized by a current or ex-intimate partner in the four months preceding
their involvement in the study; had to agree to be interviewed six times over two years
and to be randomly assigned to an experimental advocacy intervention group, or a control
condition, where women received ‘services as usual;’ and finally, had to have at least one
child between the ages of 5 and 12 living with them who was also interested in
participating in the larger study.
Procedure

Survivors were recruited from several sites: (a) a community-based agency
providing short-term supportive services to survivors following police intervention (n =
67; 42%), (b) two community-based programs that provide both residential and
nonresidential services to families experiencing IPV (n = 51; 32%), and (c) the county
prosecutor’s office through which women obtain personal protection orders (n = 42;
26%). All women provided written informed consent to participate prior to their first (i.e.,
baseline) interview, which was conducted within two to three weeks after recruitment.

Interviews with women in residential shelter programs were not conducted until all

75

families transitioned out of the shelter to avoid interviewing individuals while in crisis;
however, all women had exited the shelter within one week post-recruitment.

Following their baseline interviews, women were randomly assigned to the
experimental (n = 67; 42%) or control condition (n = 93; 58%). Subsequent interviews
were conducted 4, 8, 12, 18, and 24 months after the baseline interview. Women were
paid $20 for their baseline interview, and $60, $70, $80, $90, and $100 respectively, for
each follow-up interview thereafter. Interviewee retention rates were 84% or higher for
all five follow-up time points (84%, 86%, 88%, 88%, and 86%; see Sullivan, Rumptz,
Campbell, Eby, & Davidson, 1996 for a detailed description of the retention strategies
employed in this study), and women who missed one or more interviews were not found
to differ by demographic characteristics, or by exposure to abuse. Interviews across all
six time points ranged from 30 minutes to 6.5 hours (M = 115.70 minutes; SD = 43.55).
All interviews were conducted by female undergraduate students who underwent an
intensive 8-week training program. Interviews were administered face-to-face, audio-
taped, and conducted in a private, safe, and convenient space to ensure women’s
confidentiality and comfort.
Participant Demographics

Women’s average age at the time of their baseline interviews ranged from 21 to
49 years (M = 32.10; SD = 6.09). Nearly half the sample identified as non-Hispanic
White (n = 72; 45%), slightly over one third as African American (n = 59; 37%), and the
remaining women as multiracial (n = 14; 9%), Hispanic/Latina (n = 11; 7%), Native
American (n = 2; 1%), or Sudanese (n = l; < 1%). One woman did not report her

racial/ethnic identity (< 1%).

76

Primarily this sample was comprised of low-income women. Monthly family
income ranged from $232 to $6,800 at baseline (M = $1,632; SD = $889). However, after
adjusting women’s monthly income to account for the number of family members this
income supported (range = 2 to 10 individuals; M = 4.07; SD = 1.39), monthly income
dropped between $58 and $1,638, with M = $425, and SD = $241. The majority of
women in this sample (n = 106; 66%) received some form of government assistance (e. g.,
food assistance, welfare, SSI) at the time of their baseline interview.

Only about half the women were employed at baseline (n = 87; 54%), with 59 of
these women having full-time positions. Education levels varied substantially across this
sample; 12% (n = 19) of women had an associates or bachelors degree, 39% (n = 63) had
attended some college, 28% (n = 45) had completed high school or trade school, and the
remaining 21% (n = 33) had not yet completed high school.

The majority of women were legally separated or divorced from their assailants (n
= 65; 41%), or were never married to them, but had ended their relationships (n = 62;
39%) by the time of their baseline interview. The remaining women were continuing their
marriage (n = 11; 7%), romantic partnership (n = 17; 11%), or platonic relationship (n =
5; 3%) with their assailants at baseline. The majority of women (n = 132; 83%), however,
involved or not with their assailants, were no longer planning on continuing a relationship
with them after their baseline interview.

Measures

Control variables. Several variables were statistically controlled for in the current

analysis. First, because some women were exposed to additional resources and supports

through their involvement in the advocacy intervention, it was necessary to control for

77

experimental condition. Additionally, women’s involvement with their assailant(s) varied
across the study and therefore was controlled. A dichotomous indicator of women’s
involvement with their assailant(s) at each time point was derived using several items
assessing women’s relationship involvement in the time preceding each interview (e. g.,
“What is your relationship now?” “Did you end your relationship in the last 4 months or
before that?”).

It is especially important when examining contextual environments in relation to
individuals’ psychological well-being to control for individual-level characteristics such
as socioeconomic status and race, each of which may exert its own influence on the
relationship between these two variables (Truong & Ma, 2006). Therefore, in the current
analysis the following variables were statistically controlled: women’s level of education,
employment status, adjusted monthly income, and race. Women’s level of education was
represented by three dummy categories: (a) less than high school, (b) high school or trade
school, and (c) some college and above (which served as the reference group to which
other groups were compared). Employment status was represented by a dichotomous
indicator of no employment versus part- or full-time employment. Adjusted monthly
income was derived by dividing women’s reported monthly income by the number of
individuals supported by this income. Race was included as a dichotomous indicator,
comparing non-Hispanic Whites to the remaining racial groups.

Psychological abuse. The 24-item Index of Psychological Abuse (IPA; Sullivan,
Tan, Basta, Rumptz, & Davidson, 1992) was used to assess the extent to which women
reported experiencing emotional or psychological abuse from their partners or ex-partners

in the time preceding each interview (e. g., “How often has he told you about other sexual

78

relationships he wanted or was having in order to hurt you?” “How often has he
threatened to hurt your family or friends?” and “How often has he threatened to end the
relationship if you didn’t do what he wanted?”). Women’s responses were based on a 4-
point Likert-type scale, with l = Never and 4 = Often. All items were averaged to
produce scale scores, and Cronbach’s alpha for this scale ranged from .88 to .95 across all
time points.

Physical abuse. Physical abuse was measured using a modified version of the
Conflict Tactics Scale (CTS; Straus, 1979; Sullivan & Bybee, 1999). The CTS was
previously modified by Sullivan and Bybee by partitioning composite items into several
single items to assess the frequency of a variety of abusive tactics which may or may not
co-occur. Further, additional items were added to reflect other forms of abuse commonly
identified within the literature (e.g., sexual assault). This 24-item scale assessed the
frequency with which a variety of physically abusive tactics were used against
participants by their partners or ex-partners during the months preceding each interview
(e. g., “How often in the last 4 months has he slapped you with an open hand?” “How
ofien in the last 4 months has he kicked you?” and “How often in the last 4 months has he
smothered you?”). Participants responded to these items on a 7-point Likert-type scale (1
= Never to 7 = More than 4 times a week), and all items were averaged to produce scale 3
scores. Cronbach’s alpha, across all time points, ranged from .88 to .95.

Depression. The 20-item Center for Epidemiological Studies-Depression Scale
(CES-D; Radloff, 1977) was used to detect current depressive symptoms and overall
distress among the study participants. Women were asked to report the frequency they

experienced a range of depressive symptoms during the week prior to each interview. For

79

instance, women responded to items such as “I felt that everything I did was an effort,” “I
was bothered by things that usually don’t bother me,” and “I felt that I could not shake
off the blues even with the help from my family or friends” on a 4-point Likert-type scale
(1 = Rarely or never to 4 = Most or all of the time). All items were averaged to create
scale scores. Cronbach’s alpha ranged from .90 to .93 across all time points.

Quality of life. Subjective global quality of life and well-being was assessed using
nine items derived from the Perceived Quality of Life Scale (Andrews & Whitney, 1976).
Women reported their overall feelings across several domains of their lives since their
preceding interview (e. g., “How do you feel about what you are accomplishing in your
life?” “How do you feel about the way you spend your spare time?” and “How do you
feel about the responsibilities you have for members of your family?”). Women’s overall
level of satisfaction was recorded using a 7-point Likert-type scale, ranging from Terrible
(1) to Extremely Pleased (7). All items were averaged to produce scale scores.
Cronbach’s alpha ranged from .85 to .92 across all time points.

Ecological context. Ecological factors, including perceptions of neighborhood
disorder and related fear of victimization due to one’s neighborhood context, were
measured using two subscales from the Neighborhood Stability and Safety measure
created by Coulton, Korbin, and Su (1996). The various domains incorporated into this
measure (e. g., Disorder, Victimization, Facility Availability, Facility Quality) were
derived from prior ethnographic work by the authors (Korbin & Coulton, 1994).
Psychometric evaluation by Coulton and colleagues revealed that the subscales were

distinct, demonstrated strong reliability (e. g., Disorder subscale a = .95 and Victimization

80

subscale a = .94) and validity, and were found to discriminate between high and low risk
neighborhoods.

Perceived neighborhood disorder was measured using 15 items which inquired
about women’s perceptions of the conditions surrounding their homes. The original
subscale containing 14 items was supplemented with one additional item to assess the
frequency of racial slurs or attacks within women’s neighborhoods. Items from the
original subscale inquired about the frequency with which various neighborhood

,9 66

conditions exist, such as “litter or trash on the sidewalks and streets, graffiti on
buildings and walls,” and “gang activity.” Women responded on a 10-point scale, which
ranged from 1 (Rarely) to 10 (Frequently). Mean scores across all items were computed
for each woman. Cronbach’s alpha for this subscale ranged from .92 to .95 across all time
points.

The fear of victimization subscale measured the extent to which residents are
worried about becoming victims of crime and violence within their neighborhoods. The
original 14-item subscale was supplemented with an additional item to assess women’s
concerns about being harassed by persons of other racial or ethnic groups within their
neighborhoods. Items from the original subscale ask women to report how worried they
are about things, such as “being mugged or beaten up in your neighborhood,” “being
sexually assaulted or raped in your neighborhood,” and “walking alone after dark in your
neighborhood.” Women were asked to respond to each item on a 10-point scale;
responses ranged from 1 (Not worried at all) to 10 (Very worried). All responses were

averaged to create scale scores. Cronbach’s alpha for this subscale ranged from .92 to .94

across all time points.

81

Data Analysis

Multilevel modeling (MLM) was employed using HLM software (Raudenbush,
Bryk, & Congdon, 2005) to examine the relationship among experiences of IPV,
women’s neighborhood contexts, and their psychological well-being over time. MLM
allows for flexibility in handling missing data due to differential participant attrition over
time, as it uses any available data to model patterns of change (Luke, 2004).

Two-level hierarchical models were run to examine the effects of IPV, perceived
neighborhood disorder, fear of victimization, and each of two outcome variables,
including survivors’ quality of life and depression. Within these models, time or repeated
measurements are nested within participants. That is, at level one, MLM allows us to
explore how each individual woman may change over the course of the study, relative to
her own past experiences (within-person effects), while at level two, MLM allows us to
examine how these changes may differ across study participants (between-person effects)
(Bryk & Raudenbush, 1987; Singer & Willett, 2003). Baseline intercept levels and time
slopes for each outcome variable were estimated as random effects, therefore allowed to
vary from woman to woman, while the effects of abuse and neighborhood context were
estimated as fixed, as they were assumed to be similar across women.

HLM allows for the examination of three separate effects: (a) the cross-sectional
between-women associations between baseline abuse and neighborhood-related variables
and baseline depression and quality of life scores, (b) the between-women effects of
baseline abuse and neighborhood-related factors on changes in women’s depression and
quality of life over time, and (c) the within-woman effects of change in abuse and

neighborhood-related factors on change in depression and quality of life over time. To

82

model the within-woman effects, Singer and Willett (2003) recommend centering time-
varying covariates around a meaningful constant. Due to the larger study eligibility
criteria, it was meaningful to compare women’s trajectories of change relative to their
scores at baseline, when all women were exposed to physical violence in the four months
preceding their initial interview. Centering these time-varying covariates around
women’s baseline scores allowed for a comparison of women’s change to their position
at the beginning of the study. Centering involved subtracting women’s baseline scores
from their subsequent interview scores for physical abuse, psychological abuse, perceived
neighborhood disorder, and fear of victimization. This produced deviation scores which
reflected the extent of change women reported at each interview, relative to their scores
at the time of their first interview.

The expectation-maximization algorithm (EM) was used in SPSS to impute
missing values on the perceived neighborhood disorder and fear of victimization scales
for two women at baseline to ensure that deviation scores could be computed for each of
their subsequent interviews. In addition, while HLM can handle missing data at level one,
all cases must be complete at level two (Heck & Thomas, 2000); therefore, race was also
imputed for one woman using EM. EM is an iterative process used to produce maximum
likelihood estimates (i.e., values that maximize the probability of capturing what would
have been observed) based on available data to replace missing data (Allison, 2002).

Five sequentially nested models were run separately for each outcome variable.
To allow for statistical comparisons across these models, full maximum likelihood
estimation (MLE) was used. MLE allowed for comparisons across models using the chi-

square difference test to determine if the addition of predictor variables significantly

83

improved model fit. All statistical tests used robust standard errors to limit the influence
of extreme scores among variables with moderate levels of skew.

Unconditional models which included only intercept and slope terms were run
first to assess change in each of the dependent variables over time, while controlling for
the level two effect of the experimental condition. Next, several control variables,
including education, employment, and women’s adjusted monthly income at level one
(modeled as time-varying covariates), as well as race at level two were added to the
models. Third, physical and psychological abuse were added to each of the models, in
addition to women’s involvement with their assailant(s) which was modeled at level one,
as a control variable allowed to vary over time. Next, perceived neighborhood disorder
was added as a predictor to the each of the models, and fear of victimization was added to
the final models.

Tests for mediation were run using the product of coefficients method endorsed
by MacKinnon, Lockwood, Hoffman, West, and Sheets (2002). According to Baron and
Kenny (1986) a mediator is a variable through which an independent variable exerts
influence over a dependent variable. To formally test for mediation, two regression
equations were run, the first of which regressed the mediator (neighborhood fear of
victimization) on the independent variable (perceived neighborhood disorder), while
controlling for physical and psychological abuse and all additional aforementioned
control variables. The second model regressed the dependent variable (quality of life and
depression, each run separately) on the mediator (neighborhood fear of victimization) and
the independent variable (perceived neighborhood disorder). Using the unstandardized

regression coefficients and standard errors from these models, Sobel tests (Sobel, 1982)

84

were then conducted to determine if the mediated or indirect effects were significant,
using a conventional level of statistical significance (p g .05).

When the Sobel tests were significant, the following criteria were used to
decipher full mediation from partial mediation and indirect effects. Full mediation was
characterized by the absence of a direct association previously present between the
independent and dependent variables, after the addition of the mediator to the model,
while partial mediation was characterized by a significant relationship that remains
between the independent and dependent variables, after the addition of the mediator to
the model. Finally, indirect effects were characterized by an absence of a direct
relationship between the independent and dependent variables, before accounting for the
mediator in the model (Holmbeck, 1997).

Results

Comparative to women’s average scores at baseline, experiences of physical and
psychological abuse were lower at each subsequent time point, as were perceptions of
neighborhood disorder and neighborhood-related fear of victimization. Women’s average
depression scores were also lower at all follow-up interviews, in comparison to the
average score at baseline, and quality of life scores were higher at all time points. The
means and standard deviations for all variables (excluding control variables) across all six
time points are presented in Table 7. Descriptive statistics and frequencies for time-
varying control variables are presented in Table 8. The results of the five sequentially
nested multilevel level models follow and are presented separately for depression (in
Tables 9 and 10) and quality of life (in Tables 11 and 12). Specifically the fixed model

effects are presented in Tables 9 and 11, and random effects are in Tables 10 and 12.

85

.2080;

=m $88 3:38 EB... omega owflozw 2: 80:2 888 conflict :82 .2583 9 3:20.. .8535? $825.55; “ooze $58 3:330 e

 

 

 

New one- N; 8.? in So- :.m 25. SN 23- 85 85 saouameceio as Benson
SN 85. con 53- SN 85. 84 mg- :.N N2- cod 85 .5286 oooeoéwmoa 83:8
3o 85. Be who- 85 $5- So who- Bo m2- cod 85 somoofiaeafi 83:5
m; 85. Ed who- NS n3- :8 25. o; 8.? cod 85 “use aoaoefiao Bean
2: 8a m: woe N2 86 m: 3.4 3.0 m3. of N3. 5:5 530
$5 an $5 8:0. 86 SN one 3.“ 85 in one an Season
in w; an woe. an 2.». cum 3% NS 9% in m; ooufiaeomioaou
2: OZ of SN NE SN 2: SN :.~ mod SN 3m 528% oooeofiwmoz
v.8 o: as o: no N: 35 8; So o: 85 w: oeoaaooba
So R; 85 $4 35 a: $5 :4 3.5 a: 85 5mm oases 33235
9. 2 E. 2 9. 3 mm 2 E. 2 a. S

d: u 5 a: u 5 a: n 5 an n 5 A5 u 5 82.16

0 25 m 2:: v oafi m 85 N Eu. a 2:: mosofi> 382

 

52.8% 2&5 3.x emote ESBEmQ Bxehtenm has 25»: Seats; N 535%

n 033.

86

coho wfiuge 8 8% wEmmE. 3 26 $2: 3 83 no: >2: momma—522— oEom a

 

 

N5? 2: 22K o: 8.3 2: 5.05 2: m5? 2: cod c 02

coda mm 3. fl N Z” oodm mm ovdm In mwém 2m code— oo_ mo>
a 28:88 53> 2050335

mm. _ m K 2 .mm 3. film an mndm E. S . _ m on mm. _ m mw 30% was owe—.8 050m

3.3 m2 3.3 we film we 3.0m NV ESN mm Sam m2 .0023 09850058 :wE

3.3 mm No.2 em 3 .5 cm 23: on 3.2 mm 3.3 mm 30:8 ...wE 52: 984
o 2223:3—

255 mm 8.5 om swam ow 29$ vm ooém mm 2.3 Q 25:28—an 02

ends 8 in? NE 3.: 2: cmdn E: 3.3. .8— 3.: m: BEE—E no .33
a EoEzoEEm

..\o 2 o\.. 2 Axe 2 X. 2 o\o 2 ..\o 2

 

mmdom Sodom omdmm $.23 36mm ofimmm vfimom 3.0% 2.32 deE oodvm 92va 0585.25.86 vomit/x

 

 

a. 2 S. 2 9m. 2 E. E a E E. E
s: "5 93.15 a: "5 a: n5 9: n5 8215
o 25 m 2:: v 2:: m 25 N 25 _ 2:: 83%; .252

 

942%an MEWN NNV MMOLQQ ww~®~u~k~v§ NQLEKO’U LO\%N~U§Q3WNLK fig hbwuwmufium. m>-$x0wwg N «933%.
w 052,—.

87

Predicting Depression

The unconditional model. Results revealed a significant linear decline in
depression over time (7 = -.014, p < .001), while controlling for experimental condition,
which was not a significant predictor of baseline depression (y = .072, ns) or change in
depression (y = -.004, ns) in this model or any model hereafier. The correlation between
the intercept (baseline levels) and slope (change over time) terms for depression was not
significant (tau r = -.037, ns).

Control variables model. The addition of women’s race at level two and adjusted
monthly income, employment status, and level of education at level one only marginally

improved model fit over the unconditional model, when predicting depression (i.e., the

differences between models did not reach conventional significance levels), with LR )(2

(6) = 11.90, p = .06. Race was unrelated to women’s depression at baseline (7 = .096, ns)
and to changes in depression over time (y = -.004, ns) in this model and all models
hereafter. Similarly, within-woman changes in employment status (y = -.006, ns),
education (less than high school = y = .103, ns; high school or trade school = y = .007,
ns), and adjusted monthly income (y = -.000, ns) were unrelated to changes in women’s
depression in this model and in all subsequent models.

Abuse model. Afier controlling for the aforementioned demographics, the addition

of physical and psychological abuse, and women’s involvement with the assailant
significantly improved model fit, with LR x2 (7) = 89.55, p < .001. Between-women

effects of baseline physical and psychological abuse significantly predicted baseline
depression; however, were unrelated to change in depression over time (7 = .000, ns, and

y = -.002, ns, respectively). At baseline, a l-unit increase in physical and psychological

88

abuse was associated with a .193 (p < .01) and .221-unit (p < .01) increase in depression,
respectively. Within-woman changes in physical and psychological abuse were also
associated with changes in women’s depression over time. On average, a 1-unit increase
in physical and psychological abuse, relative to baseline, was associated with a .125 (p <
.05) and .157-unit (p < .001) increase in depression, respectively. Within-woman
variability in involvement with the assailant was unrelated to change in depression over
time (y = .038, ns) in this model and all models hereafter.

Perceived neighborhood disorder model. The addition of perceived neighborhood
disorder to the abuse model significantly improved model fit, with LR 12(3) = 12.65, p <
.01 . While controlling for abuse and all aforementioned demographics, between-women
variability in baseline perceptions of neighborhood disorder was significantly related to
women’s baseline depression, but was unrelated to change in depression over time (7 = -
.001, ns). Specifically, a l-unit increase in perceived neighborhood disorder at baseline
was associated with a .066-unit increase in women’s depression at baseline (p < .01).
Within-woman variability in perceptions of neighborhood disorder was unrelated to
change in depression over time (y = .020, ns).

Neighborhood fear of victimization model. The addition of neighborhood fear of
victimization did not significantly improve model fit, over the perceived neighborhood
disorder model, LR X2 (3) = 5.73, ns. Baseline fear of victimization was not predictive of
baseline depression (y = .028, ns) nor of changes in depression over time (7 = .001, ns);
however, within-woman effects of fear of victimization was significantly associated with
changes in depression over time. A l-unit increase in fear of victimization was associated

with a .028-unit increase in depression (p = .05 ).

89

.25. v 2 ...: .5. v 2 .. .3. w 2 .. a 353 3 232255 52 u z 2 332 3 3:23:85 83 u z .352

 

 

 

m 5. .38. 232255 5 322
:5. 25. 25. 52.. 2283 3823.232
58. :2. $5. .32. 38. .32. 823 323.22
58. :52. 25. ...-Na. 25. $2.2. 9.53 32322335
25. 58. 52.. we. 52.. 38. 33:33 :23 3330285
38. :5. «8. :5. 32. 58. $5. 25. 52.3. 23-5 .5 .258 52: Hc2223
25. 2 _. 25. N2. 25. 52. so. 2:. .823. .35 35 33 32388
25. 25.- 25. 25.- 25. 25.- 25. 25.- 2:82 25.53 3323....
58. 25.- 55. 25.- 52.. 25.- $5. 25.- 3:33 3022255
02:092. 8 033—0.. 6:5 .53 533520 E uwSEo :o flue-to 53595-5532
:5. 25. 32332835 32 2.03m
25. :5.- 25. :5.- 3283 3823.33: 233m
25. :5. 25. :5. 25. 25. 3.53 32355 233m
25. 25.- 25. .25.- 25. 25.- 9.23 3032228 2.33m
25. 25.- 25. 25.- 25. 25.- 25. 25.- 832
25. 25.- 25. 25.- 25. 25.- 25. 25.- 25. 25.- 322250
25. 25.- 25. 25.- 25. 25.- 25. ...: 5.- 25. 2.25.- as: 3% 2.2m 302: 3233..
2252225 mo ego—m .802: 05 :o Eco-co soEOBéoozzom
NS. 38. 3233252., 5 32 2.33m
85. .32.. 58. :55. 3285 3828.33: 233m
58. 5:. M55. :22. 55. :32. 3.23 3225 32332
25. .22. 25. .52. 25. .... am. .323 23322322. 2.33
:5. 38. 25. $5. :5. 2:. 25. 25. 832
25. $5. 25. 35. 25. 55. 25. :5. 25. NE. 322250
aoo. ...... LEN-N cc _ . ... ... emmmd 00¢. ... ......womd 3o. 3:59 _ .N cmo. ... 1.0va Once-3E5 owwho><
5232225 02:83 so ace-co 5803-:ooéom
.2. > 25. 2 S. > 25. 2 mm. 2 282m 82 a
3:35.56”: > SE85
Mo .2302 woo—13562 83¢. £2250 3.5228023

 

22.23%.»Q MSBNBEK 5.322: 2232 b~222§mmm 22 5.8.82 28%-“N 282K N 222%
o 033-

90

.25. v a ...: .5. v a .... .2. v a 5 ”N 3:: 3 2.32228 52 n 2 2 33 3 3:23:88 23 u 2 332

 

3on BE 9 9628

and ...... _mo.m~ .12.. fimmdw 5 8%: -- 58 vocab-£5 9853-30 «a

mnemtquoU $82

 

 

-- 52. -- 52. -- 5:. -- N5. -- 52. 83323235232 2 .23
......- 25. ......- 25. ...-... 25. ...-... 25. ...-... 25. 2332552322
......- 2:. ......- :2. ...-... 52. ...-... 52. ...-... 25. 3233250222
oo§§> oo§t3> 359$ oo§§> cos—35> mwootm :8ch
22532585 5285
mo Sum veer—Spams Z 822 $0.280 320352025

 

38.3-8ka Mat-239$ .53th 3582 2.232233me 22 58.83 ESIQREQU Eng: ~28 flanks-w ESSSQ m 2.23m.
2 033.“-

91

Testing neighborhood fear of victimization as a mediator. Between-women
variability in baseline fear of victimization was not found to mediate either the
relationship between baseline neighborhood disorder and depression (y = .018, ns), or the
relationship between baseline neighborhood disorder and change in depression over time
(y = .001, ns). However, within-woman changes in fear of victimization did explain the
effects of perceived neighborhood disorder on changes in women’s depression over time.
Specifically, there was a significant indirect effect of change in neighborhood disorder on
depression through change in fear of victimization (indirect y = .017, p = .05).

Predicting Quality of Life

The unconditional model. A significant linear increase in quality of life was
detected over time (y = .021, p <.001), while controlling for condition, which was not a
significant predictor of baseline quality of life (y = -.127, ns) or change in quality of life
(y = .003, ns) in this model, or any model hereafter. The tau correlation (tau r = -.010, ns)
between baseline levels (intercept) and the slope term signifying change over time for
quality of life was not significant.

Control variables model. The addition of several control variables (race at level

two and employment, income, and education at level one) significantly improved model

fit, LR xz (6) = 27.58, p < .001, over the unconditional model. Race was unrelated to

baseline levels of quality of life (y = —.202, ns) or to change in quality of life over time (y
= .005, ns). Change in women’s level of education (less than high school = y = .052, ns;
high school or trade school = y = .112, ns) was unrelated to change in women’s quality of
life over time in this model and all subsequent models. Within-woman variation in

employment status was marginally related to changes in women’s quality of life (y =

92

.148, p = .08) and change in adjusted monthly income was significantly associated with
women’s perceived quality of life over time; specifically, a l-unit increase in income was
associated with a .0003-unit increase in women’s quality of life (p < .05).

Abuse model. Controlling for all the aforementioned variables, the addition of
physical and psychological abuse, and women’s involvement with their assailants,

significantly improved model fit over the control variable model, with LR 12(7) = 171.05,

p < .001. At baseline, psychological abuse significantly predicted women’s quality of
life; however, physical abuse did not (7 = -.105, ns). Specifically a l-unit higher
psychological abuse score was associated with a .454 unit-lower quality of life score (p <
.01). Baseline levels of physical (y = -.015, p = .07) and psychological abuse (7 = -.000,
ns) did not predict change in quality of life over time; however, the effect of physical
abuse approached statistical significance. Within-woman change in psychological abuse
was related to changes in quality of life; however, change in physical abuse was not (y =
-.113, ns). A l-unit change in psychological abuse, relative to baseline, was associated
with a -.401-unit change in quality of life (p < .001). The time-varying effect of women’s
involvement with their assailants was also predictive of change in quality of life over
time; at times when women were involved with their assailants, on average, women’s
quality of life was .219 units lower than when women were not involved (p < .01).
Perceived neighborhood disorder model. The addition of perceived neighborhood

disorder did not significantly improve upon the prediction of women’s quality of life,

over the abuse model, with LR 12 (3) = 6.22, ns. At baseline, the effect of neighborhood

disorder on quality of life only approached statistical significance (y = -.058, p = .08),

and baseline perceptions of neighborhood disorder were unrelated to changes in quality

93

of life over time (y = -.000, ns). However, within-woman variability in neighborhood
disorder was significantly associated with changes in women’s quality of life over time.
On average, a l-unit higher neighborhood disorder score was associated with a .037-unit
lower quality of life score (p < .05).

Neighborhood fear of victimization model. The addition of neighborhood-related
fear of victimization significantly improved model fit, over the perceived neighborhood

disorder model, while controlling for all demographic variables, physical and

psychological abuse, and perceived neighborhood disorder (LR )(2 (3) = 9.64, p < .05).

Although baseline fear of victimization was unrelated to baseline quality of life (y = .014,
ns) or to change in quality of life over time (y = .001, ns), within-woman change in fear
of victimization was related to changes in quality of life over time. Specifically, a l-unit
higher score on the fear of victimization scale, relative to baseline, was associated with a
.050-unit lower score on quality of life (p < .05).

Testing neighborhood fear of victimization as a mediator. Baseline fear of
victimization did not mediate the relationship between baseline perceptions of
neighborhood disorder and baseline quality of life (7 = .009, ns) or between baseline
perceptions of neighborhood disorder and change in quality of life over time (7 = .001,
ns). However, within-woman changes in fear of victimization was found to fully mediate,
or explain, the effects of perceived neighborhood disorder on changes in women’s quality
of life over time (y = -.03 l, p < .05). That is, the prior direct relationship between
changes in perceived neighborhood disorder and change in quality of life over time
became non-significant after accounting for within-woman variability in fear of

victimization.

94

 

.8. v a ...: ..o. v a .. .8. w a .. .2. v a . .m .23 ... 3590...... ...... u 2 .. .23 ... 25.5.8... an. n 2 as...

 

 

 

... . o. .53.- 8.3.5.83... .8...
N8. 8...- a... :8.- .....8... 82.8.3.2
a... m. ..- e8. .... ..- 8... m. ..- 3.3.. 32%....
a... ......SM- 3.. rig..- NS. :19..- 33. 30.3.2.0...
.8. :08.- E... tam.- cwo. .35.. .5......3 ...... 2023.9...
.3... 5.. a... as. 8... c8. ...... m. .. .85. 2...... .o .85.. .....m H............_...-..
.3. m8. ...... .8. .9... NS. 3 .. N8. 32.8 ...... 5.... a3 “cease...
8o. .8... 8o. 8... 8o. .8... 8o. .8... 0...8... ......5... 8.3:...
So. . mm _. m8. m2. mac. .NE. 33. .m3. 3.8m EoEzoEEm
2.283 8 2620.. 6...: .53 e..: ..o bzasc E owcmno co woo-to §EOB-E£_>2
So. .8. ..o..§.e..o.> .o .8. 8:83.
8o. .2..- So. So.- .o...8... 82.82%: 05.8....
.3... La.- 8... .29- 8... .m z..- 82... .83.... ......oam
8... .8. 8o. 8...- 8... 8...- 8...... seesaw... "......a...
8... 8... 8... So. 8... 8o. 8... 8o. 85.
S... 8... So. 2.... So. 8... 8... 8o. 8... 8o. 8......50
8... 3.... m..... 2....- So. 2....- 8... ...: .... 3... ...... . 8. e..... .26 25.. ...-...... owe»?
om: ..o .052... ..o vac—m Soc: 05 :o acute :oEoB-cooiom
So. ...... 8.3.5.3... .o .8. 8.33m
...... 3...- MS. .3..- ......8... 82.3%.... 2.3....
2.. 8..- mm .. 8..- N... .. 3..- 8...... .85»... 05.33
...... .....a....- ...... :2..- .2 3%...- 3...... 5.3.2.0.... .....aam
...... m8.- m.... omn- N.... .48.- a... 8m. 83.
a... ....- a... 8..- ..3. so..- 5. .....- 3.. R..- 8......50
8.. :33... ..m .. ......5... «2. tie... 3.. its... .3... ...:3... 38.8.... Omani
om: ..o .925... 2.20an :o acute :oEoB-coozcom
...-..- .. m... > mm > $- .. m...- » astm 8......
533.833 > .0985
«0 80m @0918st Z 83¢. 20950 12822025

 

sh HE 6:35 w:.:...fi~.£ @3303 38mm? $3.3»:me :V ...-8...... flowfim hub-K N 3%
_ _ 033-

95

..8. v 2 ...: .8. w 2 .. ”N .23 .. 388.25.. 8. n 2 M. .23 .. Sonata... 8.. u 2 .82

 

Econ: Sta 9 03.22

... Nvod mmmd .12.. wvo. _: .12.. vwmfim .. .8. oocobtzu 2.9.3-30 ~3

 

3858.800 632

 

 

 

-- Sm. - Sm. -- 2m. - 8m. -- 8m. 88...... 8:28.23...
...... .8. ...... .8. ...... .8. ...... .8. ...... .8. .82....08388
...... 8m. ...... 8m. ...... 8m. ...... 88. ...... .8. 35.083.88.35
oo§5> oo§MS> 85.5; won-33> oo§E> Eon-mm :8ng

gouging; .0285
.0 Sum coon-.336 Z 32?. 28:80 12.2.6583

.3

\c 253% MRESER .233: .3232 223.253qu 3x 86.28 anew-232530 3.8: .38 58wa Souzax N 333%.
Q 2an

96

Discussion

Guided by an ecological perspective, this longitudinal study sought to enhance
our understanding of how the psychological well-being of survivors of IPV may be
affected by their experiences of abuse as well as their neighborhood environments. This
study expands upon the extant literature examining community violence exposure among
survivors of IPV by considering the long—term effect of subjective perceptions of
neighborhood disorder while also attempting to unveil the mechanism that underlies the
relationship between these perceptions and survivors’ overall mental health and well-
being. This study also furthers our understanding of how exposure to abuse and societal
ills may not only affect survivors’ mental health, but also their general life satisfaction.

The findings from this study provide further evidence for the link between abuse
and negative psychological sequelae identified in previous research (Adeodato et al.,
2005; Cascardi et al., 1999; Golding, 1999; Helfrich et al., 2008; Leung et al., 2005), and
generally support findings from prior longitudinal studies as well, suggesting that the
negative mental health implications of abuse are most severe at times when abuse is high
and are generally short-lived once the abuse ends (Anderson et al., 2003; Campbell et al.,
1995; Kernic et al., 2003). In support of this contention, women’s baseline experiences of
physical and psychological abuse were unrelated to changes in depression over time,
which is likely due to the fact that after baseline (when women’s abuse was highest)
abuse dropped off, as many women ended their relationships. Notably, however, the
relationship between baseline physical abuse and change in quality of life over time
approached statistical significance, suggesting that for some women, physical abuse may

continue to impact their life satisfaction over time. This may be especially true for

97

 

women who have suffered significant injuries as a result of the violence they endured, as
prior research has found a significant association between injuries and women’s reported
quality of life (Sutherland, Bybee, & Sullivan, 2002).

In further support of the aforementioned literature, ongoing abuse continued to
exert an impact on women’s psychological well-being over the course of the study. At
times when physical and psychological abuse were high, depression was also high.
However, findings from prior research examining the collective impact of both physical
and psychological abuse are somewhat inconclusive regarding whether each form of
abuse independently contributes to survivors’ depressive symptomatology. In some
studies, when both physical and psychological abuse are assessed, psychological abuse
appears to be so strongly related to depression that the effect of physical abuse is often
rendered insignificant (Mechanic et al., 2008; Pico-Alfonso et al., 2006); while in other
research, physical abuse is found to be the primary contributing factor to women’s
depression (Campbell et al., 1997; Sackett & Saunders, 1999).

Far less is known about the impact of abuse on women’s quality of life over time.
Interestingly, after accounting for the significant time-varying effect of psychological
abuse on quality of life, change in physical abuse was not significantly associated with
change in quality of life over time. Overall life satisfaction seems to primarily depend
upon the extent to which one is exposed to psychological abuse, over and above physical
violence. This is not particularly surprising that psychological abuse appears to be a
stronger predictor of women’s overall life satisfaction. Quality of life was assessed by
items concerning general enjoyment with one’s life, contentment with how one spends

one’s spare time, feelings of independence and freedom, feelings of personal safety,

98

feelings about accomplishments in life, etc. These are several of the very aspects of
women’s lives targeted by assailants’ isolating and controlling tactics of psychological
abuse (Marshall, 1999), and many women find psychological abuse to be more damaging
than physical abuse (Follingstad, Rutledge, Berg, Hause, & Polek, 1990).

The cumulative effects of chronic stress brought on by physical and psychological
abuse and deleterious environmental contexts are salient to our understanding of
survivors’ psychological well-being. Prior research has been limited to examining only
community violence exposure among survivors of IPV (Bogat et al., 2005; Brown et al.,
2005; Clark et al., 2008), overlooking women’s subjective perceptions of their
environments. Therefore, in addition to abuse, the effects of perceived neighborhood
disorder and neighborhood-related fear of victimization were examined to determine if
they further explained aspects of survivors’ psychological well-being. Baseline
perceptions of neighborhood disorder and feelings of fear of victimization were not
related to survivors’ quality of life or depression over time, as was expected. It is likely
that perceptions of neighborhood disorder, and the emotions that follow, are most
influential in the short-term, and it is quite likely that these perceptions and feelings from
an earlier time point, perhaps when women may have been residing in another location,
would not have long-term implications for their well-being.

Perceptions of neighborhood disorder and related feelings of fear of victimization
over the course of the study did enhance our understanding of how survivors’
psychological well-being changes over time. Changes in neighborhood disorder were
found to both directly and indirectly relate to survivors’ psychological well-being. At

times when perceived neighborhood disorder was high, women’s quality of life was low,

99

“X" I-‘—:— in..—

and the relationship between survivors’ perceptions of the social and physical incivilities
within their neighborhoods and their overall life satisfaction was fully explained by
survivors’ fear of victimization. As these findings suggest, general life satisfaction may
be affected by deteriorated structural conditions and community members’ disorderly
behavior that preclude residents from engaging in activities they enjoy outside their
homes because of fear. Additionally, these feelings of fear regarding one’s neighborhood
are likely to have direct impacts on other aspects of life satisfaction as well, such as
limiting one’s ability to carry out family responsibilities, and diminishing one’s feelings
of personal safety, freedom, and independence.

Notably, depression was not directly predicted by perceptions of neighborhood
disorder, despite support for this association in prior research (Christie-Mizell et al.,
2003; Latkin & Curry, 2003; Ross, 2000; Ross et al., 2000). It is possible that the
experience of abuse unique to this sample was so detrimental to women’s mental health
that exposure to physical and social incivilities was not significant enough a stressor to
manifest additional depressive symptomatology. Interestingly, however, perceived
neighborhood disorder was found to indirectly relate to depression through feelings of
fear. Therefore, afier accounting for abuse, in order for survivors’ subjective accounts of
their neighborhood environments to have an effect on their mental health, particularly
depression, they had to be fearful of their surroundings.

The relationship between survivors’ perceptions and feelings about their
neighborhoods and their psychological well-being is supported by previous research
conducted with several non-abused samples (Gary et al., 2007; Haney, 2007; Hill et al.,

2005; Xu et al., 2005; Zule et al., 2008). More importantly, however, these findings

100

advance our understanding of how perceptions of neighborhood disorder and feelings of
fear may affect women who are already experiencing a significant chronic life stressor
(e.g., IPV). Even above and beyond the devastating effects of abuse, the experiences of
‘neighborhood’ for women with abusive partners appear to be consistent with that of
individuals from prior studies, signifying that individuals’ environmental contexts may be
a substantial source of distress. The theoretically-driven disorder/incivilities model
(Franklin et al., 2008) supported in several prior studies (LaGrange et a1, 1992; Ross et
al., 2000; Ross & Mirowsky, 2001; Schafer et al., 2006; Xu et al., 2005) also holds true
for women who have suffered the devastating impacts of IPV. This model purports that
indicators of social and physical incivilities give rise to feelings of fear and increased risk
for victimization among residents, and these feelings of fear and risk of victimization
appear to be the mechanism through which survivors’ perceptions of their neighborhoods
impact their overall mental health and well-being. Findings from prior research have
supported the link between feelings of fear of crime and general perceptions of safety,
and residents’ mental health and overall quality of life (Green et al, 2002; Hill &
Herman-Stahl, 2002; Kruger et al., 2007; Xu et al., 2005; Zule et al., 2008); however, to
date, this is the only known study to have examined this meditational pathway over time,
and with survivors of IPV.

The findings presented must be considered in light of several limitations. First,
this sample was comprised of primarily low-income mothers who self-identified as
survivors of IPV. Therefore, it is unclear how these results would generalize to other
samples of women with abusive partners. For instance, it is possible that women with

children may have different experiences than those without; especially if their assailants

101

 

play a parenting role in the lives of their children. The continued involvement of an

assailant in a woman’s life may have significant long-term implications for her well-

being. Additional research is needed to attend to issues of generalizability using more

diverse samples. However, despite this limitation, this sample was rather unique, in that

women were recruited from multiple sources, which diversified the sample beyond other

studies of IPV which commonly rely solely on women recruited from domestic violence

shelters. ‘

Second, though these findings contribute to our understanding of the effects of
survivors’ environmental contexts on their well-being, this study should be replicated
using a larger sample. Generally, effects on slopes are typically small (Snijders & Boskar,
1999); therefore, a small sample limits statistical power to find significant predictors of
change in psychological well-being over time. Replication with a larger sample would
increase power to detect significant effects that may have been missed in the current
study.

Third, there was no clear definition of ‘neighborhood’ provided to survivors. This,
however, can be perceived as both a limitation and strength of this study. While some
women may have responded to interview items by reflecting on the immediate area
surrounding their home or apartment, others may have more broadly conceived their
neighborhood as their residential block, or even a section of the city in which they reside.
Therefore, there is likely a lack of consistency in the geographical area on which women
reported. Research, however, has found that individuals’ perceptions of their

neighborhoods are shaped not only by their immediate environment, but also by the

102

neighborhoods surrounding their own (Wyant, 2008); given this, it was best not to limit
one’s conceptualization of neighborhood.

A final limitation concerns the inability to disentangle temporal ordering among
events within each time point. That is, within the four- or six-month time period prior to
each interview it was not possible to determine if neighborhood disorder preceded fear of
victimization or if fear of victimization preceded women’s reported assessments of
depression and quality of life. The same holds true for the effects of abuse on
psychological well-being. The use of longitudinal data did, however, allow for the
assessment of how the study variables changed together across time, which enhanced our
understanding of the co-occurrence of events among this sample.

Despite these limitations, this study offers several important implications for
future research, intervention, and practice. Given that this study was not able to tease
apart the temporal ordering of events within each time point, additional research is
needed to examine the relationship between these variables over time using a lagged
design, in which survivors’ perceptions of neighborhood disorder are assessed prior to
feelings of fear, and fear in turn, is assessed prior to both depression and quality of life.
Additionally, because the time interval between women’s perceptions of neighborhood
disorder and feelings of fear that may follow is relatively short, interviews should be
spaced more closely together.

Given the importance of survivors’ perceptions of their contextual environments,
far more research is needed to understand how other aspects of individuals’
neighborhoods, assessed subjectively, may be implicated in the development of negative

psychological sequelae so that intervention efforts may be targeted to these areas. For

103

 

instance, it is clear from the current study that signs of neighborhood disorder may
directly or indirectly affect survivors’ psychological well-being; however, it would be
beneficial to understand how these indicators of disorder, coupled with women’s reports
of witnessing community violence, or even personal victimization outside one’s intimate
relationship may impact well—being. In addition, it is important to examine possible
protective mechanisms that may moderate the relationship between survivors’
perceptions of neighborhood disorder and their psychological well-being. For instance, it
is quite possible that residents residing in neighborhoods with strong community ties do
not have negative mental health implications, despite high instances of disorder within
their neighborhoods.

Community-based interventions should be designed to work with survivors and
their children to ensure they have safe, affordable housing and access to requisite medical
and mental health services to facilitate their physical and emotional recovery from the
devastating impacts of abuse. Interventions aimed at enhancing social support may also
have positive implications for women’s psychological well-being. The negative
psychological sequelae resulting from the isolation and degradation women endure as a
result of the tactics of psychological abuse employed by their partners may be partially
combated by efforts aimed at reconnecting women with positive social support networks.
Interventionists could work directly with survivors to reestablish bonds with preexisting
networks and work to develop additional supportive communities of women who can
share their stories with others who have had similar experiences.

Intervention efforts should also be directed toward reducing the social and

physical incivilities within low-income neighborhoods, as they too are associated with

104

diminished quality of life and increased depressive symptomatology. One potential
strategy is to work directly with neighborhood residents to build collective efficacy.
Collective efficacy and capacity to diminish incivilities may enhance residents’ sense of
community and recognition of a shared concern for community safety and stability
(Schafer et al., 2006). Together with a shared mission, residents may work together to
organize collective community clean-up projects to reduce the signs of neighborhood
deterioration; develop or enhance Neighborhood Watch programs to increase surveillance
and deter social and physical incivilities; and develop neighborhood or community
associations to advocate for increased police patrol and other resources that may aid in
restoring residents’ feelings of safety and security. More directly, local government and
law enforcement officials must prioritize efforts to control incivilities, perhaps by
modifying their general operating procedures to promote the safety and well-being of
their community members.

Finally, service providers and practitioners must be cognizant of the cumulative
effects of stress in the lives of survivors of IPV and should attend to the difficulties this
stress causes survivors who are attempting to rebuild their lives. First and foremost, when
women engage with community-based providers immediately following an abusive
relationship (particularly domestic violence shelter programs which often offer time-
limited assistance), it is likely that they are still dealing with depressive symptomatology
and therefore may have difficulties meeting the expectations or demands of program or
shelter staff (e. g., identifying and procuring requisite resources to rebuild their lives
within a limited amount of time). The effects of having had resided in a disadvantaged

neighborhood as well as other stressors may only exacerbate negative mental health

105

sequelae, further limiting women’s abilities to immediately deal with their most pressing
needs. Therefore, providers and practitioners must ensure that survivors are provided
adequate time, counseling services, and general support to overcome the devastating
mental health impacts of abuse and other additive stressors, before being expected to
actively engage in the processes/activities required to meet their expectations.

This study demonstrates the importance of context when attempting to understand
the complexity of survivors’ experiences and the negative psychological sequelae they
endure. Future research must attend to circumstances beyond the abuse survivors
experience to identify additional ecological factors that may contribute to their
compromised well-being. Placing emphasis on the effects of abuse, while also attending
to additional contextual barriers to survivors’ well—being, may greatly improve upon our
ability to provide survivors with the care, support, and resources they need to overcome

the adversities they have faced.

106

Chapter 4

CONCLUSION

Prior researchers have long contended that survivors’ experiences of physical and
psychological abuse are associated with devastating psychological sequelae (see Golding,
1999 for a review). The findings from this dissertation support and extend the existing
literature in a number of ways. First and foremost, the majority of prior research
examining the mental health implications of abuse has been cross-sectional. Examining
survivors’ psychological well-being over time allows us to determine optimal points of
intervention to assist women in overcoming the deleterious effects of abuse. Across these
two studies, depression appears to be a rather immediate consequence of abuse,
signifying the importance of providing survivors with counseling services, working to
establish and/or restore their support networks, and offering other salient resources to
promote their psychological well-being shortly following their abusive relationships.

It is noteworthy, however, that the effect of within-woman changes in physical
abuse on change in depression over time, across these two studies, was somewhat
inconsistent. In study 1, Understanding the Impact of Resource Constraints on the Mental
Health and Well—Being of Survivors of Intimate Partner Violence Over Time, changes in
physical abuse, relative to baseline, were found to be unrelated to changes in depression
over time; albeit, within-woman changes in physical abuse in study 2, An Examination of
the Impact of Neighborhood Factors on the Well-Being of Survivors of Intimate Partner
Violence Over Time, predicted change in depression over time. This inconsistency can be

attributed to the methodological differences between the studies. Specifically, due to the

107

nature of my research questions, study 1 was restricted to examining fewer time points,
using a smaller sample. Therefore, it is likely that the findings from study 2 are more
robust; however, additional research focused on examining the unique and collective
impacts of physical and psychological abuse over time is warranted.

Second, this dissertation expands upon the previous literature by defining and
measuring psychological health more broadly than the presence or absence of mental
illness. That is, in addition to examining depressive symptomatology, which is said to be
one of the most common mental health consequences of abuse (Campbell, 2002; Golding,
1999), this dissertation is one of few longitudinal studies drawing our attention to broader
aspects of well-being by examining survivors’ quality of life. Quality of life has
significant long-term implications for survivors’ safety (Bybee & Sullivan, 2002, 2005);
therefore, our efforts must not only focus on helping survivors overcome significant
psychological distress, but also on assisting women to restore their sense of satisfaction
with their lives. When examining survivors’ quality of life across these two studies, the
necessity to provide long-term intervention to women who have experienced severe
physical abuse within their relationships became clear. Physical abuse appears to have
long-standing implications for survivors’ quality of life, albeit depressive
symptomatology subsides once the abuse ends. Consequently, long-term interventions
aimed at restoring survivors’ overall life satisfaction must supplement more acute care
provided to assist women in overcoming the immediate, deleterious impacts of abuse.
Comprehensively addressing the short— and long-term needs of survivors will likely

facilitate their recovery from abuse and assist them in rebuilding their lives.

108

Third, and among the most significant of contributions of this dissertation, is the
attention afforded to ecological factors that further explain survivors’ psychological well-
being. Understanding the many pathways by which abuse affects survivors’ mental health
and quality of life is paramount to our ability to develop interventions to alleviate the
significant distress experienced by women as a result of the abuse they endured. Further,
attending to contextually-based factors that expose survivors to cumulative stress
provides supplementary points of intervention, thereby offering additional opportunities
to link survivors to needed resources.

More specifically, study 1 illustrated that survivors’ resource constraints are fully
responsible for the significant mental health consequences they experience, following
psychological abuse. Resource constraints represent a significant pathway by which
abuse impacts survivors’ psychological well-being, and this pathway illustrates multiple
opportunities for intervention. For instance, community-wide efforts must be directed to
increase awareness of available resources that exist within communities, and to improve
survivors’ access to these resources. Together, these strategies could help survivors break
free from their abusive relationships, and prevent or alleviate the distress that is otherwise
inevitable following abuse. Community providers must also work collaboratively to
ensure that care is coordinated and addresses all the needs survivors present with to
increase the likelihood that service provision is successful. It is essential that providers
see survivors through the process of procuring resources until their lives have become
stable, and the resources they need are independently sustainable over time. This

comprehensive approach to care requires patience, flexibility, and dedication on the part

109

of providers to ensure that survivors’ unique needs are addressed and the likelihood of
their future success is maximized.

Survivors’ neighborhood contexts, examined in study 2, also proved to be an
important factor to consider when attempting to understand women’s psychological well-
being. The cumulative stress survivors experience as a result of abuse and deleterious
neighborhood conditions only exacerbates negative psychological sequelae. In addition to
providing survivors with the resources and supports they need to effectively deal with the
abuse they have experienced, we must begin to help disadvantaged communities build
their capacity to address the social and physical incivilities that are responsible for the
compromised well-being of their residents.

Community leaders, providers, and advocates must take action to organize
residents to combat the signs of disorder and decay responsible for bringing about
feelings of fear and distress. Efforts and incentives to motivate and empower residents
must be provided to help build capacity and foster resident cohesion. Residents may then
collectively address neighborhood disorder by employing several strategies, such as
initiating community-wide clean up initiatives, instituting Neighborhood Watch
programs, developing neighborhood associations, and implementing zero-tolerance
policies for disorderly activities. Leaders and activists may also advocate on behalf of
disadvantaged communities to ensure that community officials remain aware of the
conditions that exist and the resources that are needed to overcome the threats to
community members’ safety. In addition to the direct benefit of addressing the source of
environmental stress for survivors of IPV, building collective efficacy within women’s

neighborhoods may also have indirect benefits for their psychological well-being. That is,

110

even when survivors have not self-identified as such, the efforts put forth to address
deleterious neighborhood conditions may indirectly help women with abusive partners to
connect with neighbors and other community-based support systems.

Our understanding of survivors’ experiences and the negative psychological
sequelae they present with is significantly enhanced when we consider additional
ecologically-based contextual factors. Future researchers must broaden their scope to
gain a comprehensive understanding of the complexity of survivors’ circumstances. That
is, in addition to abuse, contextual or situational factors must be at the forefront of our
investigations so that interventions may be developed to comprehensively address
survivors’ needs. Research of this nature is critical for determining how best to invest our
increasingly scarce intervention dollars. In addition, we must continue to expand upon
our understanding of survivors’ psychological well-being by investigating ecological
factors that promote women’s quality of life, in addition to examining those variables that
primarily lead to pernicious psychological outcomes. Together, with these goals in mind,
future research and practice will continue to expand our current capacity to assist
survivors in restoring their happiness, independence, and most importantly, their physical

and psychological well-being.

111

REFERENCES

Adams, A. E., Sullivan, C. M., Bybee, D., & Greeson, M. R. (2008). Development of the
scale of economic abuse. Violence Against Women, I 4(5), 563-588.

Adeodato, V. G., Carvalho, R. dos R., de Siqueira, V. R., & de Matos e Souza, F. G.
(2005). Quality of life and depression in women abused by their partners. Revista
de Saude Publica, 39(1), 108-113.

Allison, P. D. (2002). Missing data. Series: Quantitative Applications in the Social
Sciences (No. 136). CA: Sage.

Alsaker, K., Moen, B. E., & Kristoffersen, K. (2008). Health-related quality of life
among abused women one year after leaving a violent partner. Social Indicators
Research, 86(3), 497-509.

Anderson, D. K., Saunders, D. G., Yoshihama, M., Bybee, D. I., & Sullivan, C. M.
(2003). Long-term trends in depression among women separated from abusive
partners. Violence Against Women, 9(7), 807-838.

Andrews, F ., & Withey, S. (1976). Social indicators of well-being: Americans’
perceptions of life quality. New York: Plenum Press.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1 173-1 182.

Beeble, M. L., Bybee, D., Sullivan, C. M., & Adams. A. E. (2009). Main, mediating, and
moderating effects of social support on the well-being of survivors of intimate

partner violence across two years. Journal of Consulting and Clinical Psychology,
77(4), 718-729.

Benson, M. L., Fox, G. L., DeMaris, A., & Van Wyk, J. (2003). Neighborhood
disadvantage, individual economic distress and violence against women in
intimate relationships. Journal of Quantitative Criminology, 19(3), 207-235.

Boardman, J. D., Finch, B. K., Ellison, C. G., Williams, D. R., & Jackson, J. S. (2001).
Neighborhood disadvantage, stress, and drug use among adults. Journal of Health
and Social Behavior, 42(2), 151-165.

Bogat, G. A., Leahy, K., von Eye, A., Maxwell, C., Levendosky, A. A., & Davidson, W.
S. (2005). The influence of community violence on the functioning of women

experiencing domestic violence. American Journal of Community Psychology,
36(1/2), 123-132.

112

Bogat, G. A., Levendosky, A. A., Theran, S., von Eye, A., & Davidson, W. S. (2003).
Predicting the psychosocial effects of interpersonal partner violence (IPV). How
much does a woman’s history of IPV matter? Journal of Interpersonal Violence,
18(11), 1271-1291.

Bonomi, A. E., Thompson, R. S., Anderson, M., Reid, R. J ., Carrell, D., Dimer, J. A., et
a1. (2006). Intimate partner violence and women’s physical, mental, and social
functioning. American Journal of Preventative Medicine, 30(6), 45 8-466.

Bosch, K., & Schumm, W. R. (2004). Accessibility to resources: Helping rural women in
abusive partner relationships become free from abuse. Journal of Sex and Marital
Therapy. 30(5), 357-370.

Breiding, M. J ., Black, M. C., & Ryan, G. W. (2008). Prevalence and risk factors of
intimate partner violence in eighteen u.s. states/territories, 2005. American
Journal of Preventative Medicine, 34(2), 112-118.

Bronfenbrenner, U. (1977). Toward an experimental ecology of human development.
American Psychologist, 32(7), 513-531.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature
and design. Cambridge, MA: Harvard University Press.

Brown, J. R., Hill, H. M., & Lambert, S. F. (2005). Traumatic stress symptoms in women
exposed to community and partner violence. Journal of Interpersonal Violence,
20(11), 1478-1494.

Browne, A., & Bassuk, S. S. (1997). Intimate violence in the lives of homeless and poor
housed women. American Journal of Orthopsychiatry, 67(2), 261-278.

Bryk, A. S., & Raudenbush, S. W. (1987). Application of hierarchical linear models to
assessing change. Psychological Bulletin, 101(11), 147-158.

Bybee, D. I., & Sullivan, C. M. (2002). The process through which an advocacy
intervention resulted in positive change for battered women over time. American
Journal of Community Psychology, 30(1), 103-132.

Bybee, D., & Sullivan, C. M. (2005). Predicting re-victimization of battered women 3
years after exiting a shelter program. American Journal of Community
Psychology, 36(1/2), 85-96.

Calvete, E., Corral, S., & Estevez, A. (2008). Coping as a mediator and moderator
between intimate partner violence and symptoms of anxiety and depression.
Violence Against Women, 14(8), 886-904.

113

Calvete, E., Estevez, A., & Corral, S. (2007). Intimate partner violence and depressive
symptoms in women: Cognitive schemas as moderators and mediators. Behaviour
Research and Therapy, 45(4), 791-804.

Campbell, J. C. (2002). Health consequences of intimate partner violence. The Lancet,
359(9314), 1331-1336.

Campbell, J. C., Kub, J ., Belknap, R. A., & Templin, T. N. (1997). Predictors of
depression in battered women. Violence Against Women, 3(3), 271-293.

Campbell, J. C., & Soeken, K. L. (1999). Women’s responses to battering over time: An
analysis of change. Journal of Interpersonal Violence, 14(1), 21-40.

Campbell, R., Sullivan, C. M., & Davidson, W. S. (1995). Women who use domestic
violence shelters. Changes in depression over time. Psychology of Women
Quarterly, 19(2), 237-255.

Cascardi, M., O’Leary, K. D., & Schlee, K. A. (1999). Co-occurrence and correlates of
posttraumatic stress disorder and major depression in physically abused women.
Journal of Family Violence, 14(3), 227-249.

Casey, P., Goolsby, S., Berkowitz, C., Frank, D., Cook, J ., Cutts, D., et a1. (2004).
Maternal depression, changing public assistance, food security, and child health
status. Pediatrics, 113(2), 298-304.

Christie-Mizell, C. A., Steelman, L. C., & Stewart, J. (2003). Seeing their surroundings:
The effects of neighborhood setting and race on maternal distress. Social Science
Research, 32(3), 402-428.

Clark, C., Ryan, L., Kawachi, 1., Canner, M. J ., Berkman, L., & Wright, R. J. (2008).
Witnessing community violence in residential neighborhoods: A mental health
hazard for urban women. Journal of Urban Health, 85(1), 22-38.

Coker, A. L., Davis, K. E., Arias, I., Desai, S., Sanderson, M., Brandt, H. M., et a1.
(2002). Physical and mental health effects of intimate partner violence for men
and women. American Journal of Preventative Medicine, 23(4), 260-268.

Coker, A. L., Smith, P. H., Bethea, L., King, M. R., & McKeown, R. E. (2000). Physical
health consequences of physical and psychological intimate partner violence.
Archives of Family Medicine, 9(5), 451-457.

Coulton, C. J ., Korbin, J. E., & Su, M. (1996). Measuring neighborhood context for

young children in an urban area. American Journal of Community Psychology,
24(1), 5-32.

114

Cunradi, C. B., Caetano, R., Clark, C., & Schafer, J. (2000). Neighborhood poverty as a
predictor of intimate partner violence among white, black, and hispanic couples in
the united states: A multilevel analysis. Annals of Epidemiology, 10(5), 297-308.

Curry, A., Latkin, C., & Davey-Rothwell, M. (2008). Pathways to depression: The impact
of neighborhood violent crime on inner-city residents in baltimore, maryland, usa.
Social Science & Medicine, 67(1), 23-30.

Cutrona, C. E., Russell, D. W., Brown, P. A., Clark, L. A., Hessling, R. M., & Gardner,
K. A. (2005). Neighborhood context, personality, and stressful life events as
predictors of depression among african american women. Journal of Abnormal
Psychology, 114(1), 3-15.

Cutrona, C. E., Russell, D. W., Hessling, R. M., Brown, A. & Murry, V. (2000). Direct
and moderating effects of community context on the psychological well-being of
african american women. Journal of Personality and Social Psychology, 79(6),
1088-1101.

Davies, J ., Lyon, E., & Monti-Catania, D. (1998). Safety planning with battered women:
Complex lives, diflicult choices. Thousand Oaks, CA: Sage.

Dienemann, J ., Boyle, E., Baker, D., Resnick, W., Widerhom, N., & Campbell, J. (2000).
Intimate partner abuse among women diagnosed with depression. Issues in
Mental Health Nursing, 21(5), 499-513.

Diener, E., & Fujita, F. (1995). Resources, personal strivings, and subjective well-being:
A nomothetic and idiographic approach. Journal of Personality and Social
Psychology, 68(5), 926-935.

Dobash, R., & Dobash, R. (1998). Violent men and violent contexts. In R. Dobash & R.
Dobash (Eds.). Rethinking violence against women (pp. 141-168). Thousand
Oaks, CA: Sage.

Dohrenwend, B. S. (1978). Social status and responsibility for stressfiil life events. In C.
D. Spielberger & I. G. Sarason (Eds), Stress and anxiety (Vol. 5., pp.25-42). New
York: Wiley.

Dunst, C. J ., Johanson, C., Trivette, C. M., & Hamby, D. (1991). Family-oriented early
intervention policies and practices: Family-centered or not. Exceptional Children,
58(2), 115-126.

Dupere, V., & Perkins, D. D. (2007). Community types and mental health: A multilevel

study of local environmental stress and coping. American Journal of Community
Psychology, 3 9(1/2),107-l 19.

115

Dutton, M. A., & Goodman, L. A. (2005). Coercion in intimate partner violence: Toward
a new conceptualization. Sex Roles, 52(11/12), 743-756.

Dutton, M. A., Goodman, L. A., & Bennett, L. (1999). Court-involved battered women’s
responses to violence: The role of psychological, physical, and sexual abuse.
Violence and Victims, 14(1), 89-104.

Dutton, M. A., Green, B. L., Kaltman, S. I., Roesch, D. M., Zeffiro, T. A., & Krause, E.
D. (2006). Intimate partner violence, ptsd, and adverse health outcomes. Journal
of Interpersonal violence, 21(7), 955-968.

Dutton, M. A., Kaltman, S., Goodman, L. A., Weinfurt, K., & Vankos, N. (2005).
Patterns of intimate partner violence: Correlates and outcomes. Violence and
Victims, 20(5), 483-497.

Follingstad, 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(2), 107-120.

Ford-Gilboe, M., Wuest, J ., Varcoe, C., Davies, L., Merritt-Gray, M., Campbell, J ., et al.
(2009). Modeling the effects of intimate partner violence and access to resources
on women’s health in the early years after leaving an abusive partner. Social

Science & Medicine, 68, 1021-1029.

Franklin, T. W., Franklin, C. A., & F earn, N. E. (2008). A multilevel analysis of the
vulnerability, disorder, and social integration models of fear of crime. Social
Justice Research, 21(2), 204-227.

Garcia-Moreno, C., Jansen, H. A. F. M., Ellsberg, M., Heise, L., & Watts, CH. (2006).
Prevalence of intimate partner violence: Findings from the who multi-country

study on women’s health and domestic violence. The Lancet, 368(9543), 1260-
1269.

Gary, T. L., Stark, S. A., & LaVeist, T. A. (2007). Neighborhood characteristics and
mental health among african americans and whites living in a racially integrated
urban community. Health & Place, 13(2), 569-575.

Golding, J. M. (1999). Intimate partner violence as a risk factor for mental health
disorders: A meta-analysis. Journal of Family Violence, 14(2), 99-132.

Goodkind, J. R., Gillum, T. L., Bybee, D. I., & Sullivan, C. M. (2003). The impact of

family and fiiends’ reactions on the well-being of women with abusive partners.
Violence Against Women, 9(3), 347-373.

116

Goodman, L., Dutton, M. A., Vankos, N., & Weinfurt, K. (2005). Women’s resources
and use of strategies as risk and protective factors for reabuse over time. Violence
Against Women, 11(3), 311-336.

Gordon, M. (2000). Definitional issues in violence against women: Surveillance and
research from a violence research perspective. Violence Against Women, 6(7),
747-783.

Green, G., Gilbertson, J. M., & Grimsley, M. F. J. (2002). Fear of crime and health in
residential tower blocks. A case study in liverpool, uk. The European Journal of
Public Health, 12(1), 10-15.

Hadley-Ives, E., Stiffman, A. R., Elze, D., Johnson, S. D., & Dore, P. (2000). Measuring
neighborhood and school environments: Perceptual and aggregate approaches.
Journal of Human Behavior in the Social Environment, 3(1), 1-28.

Haney, T. J. (2007). “Broken windows” and self-esteem: Subjective understandings of
neighborhood poverty and disorder. Social Science Research, 36(3), 968-994.

Hatch, S. L. (2005). Conceptualizing and identifying cumulative adversity and protective
resources: Implications for understanding health inequalities. Journals of
Gerontology, 6OB(Special Issue), S130—S134.

Heck, R. H., & Thomas, S. L. (2000). An introduction to multilevel modeling techniques.
Mahwah, N.J.: Lawrence Erlbaum Associates, Inc.

Helfiich, C. A., Fujiura, G. T., & Rutkowski-Kmitta, V. (2008). Mental health disorders
and functioning of women in domestic violence shelters. Journal of Interpersonal
Violence, 23(4), 437-453.

Hill, N. E., & Herman-Stahl, M. A. (2002). Neighborhood safety and social involvement:
Associations with parenting behaviors and depressive symptoms among african

american and euro-american mothers. Journal of Family Psychology, 16(2), 209-
219.

Hill, T. D., & Angel, R. J. (2005). Neighborhood disorder, psychological distress, and
heavy drinking. Social Science & Medicine, 61(5), 965-975.

Hill, T. D., Ross, C. E., & Angel, R. J. (2005). Neighborhood disorder,
psychophysiological distress, and health. Journal of Health and Social Behavior,
46(2), 170-186.

Hobfoll, S. E. (1988). The ecology of stress. New York: Hemisphere.

Hobfoll, S. E. (1989). Conservation of resources. A new attempt at conceptualizing
stress. American Psychologist, 44(3), 513-524.

117

Hobfoll, S. E., Johnson, R. J ., Ennis, N., & Jackson, A. P. (2003). Resource loss, resource
gain, and emotional outcomes among inner city women. Journal of Personality
and Social Psychology, 84(3), 632-643.

Hobfoll, S. E., & Lilly, R. S. (1993). Resource conservation as a strategy for community
psychology. Journal of Community Psychology, 21 (2), 128-148.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical clarity in the
study of mediators and moderators: Examples from the child-clinical and pediatric
psychology literatures. Journal of Consulting and Clinical Psychology, 65(4),
599-610.

Holtzworth-Munroe, A., Smutzler, N., & Bates, L. (1997). A brief review of the research
on husband violence. Part III: Sociodemographic factors, relationship factors, and
differing consequences of husband and wife violence. Aggression and Violent
Behavior, 2(3), 285-307.

Houry, D., Kaslow, N. J ., & Thompson, M. P. (2005). Depressive symptoms in women
experiencing intimate partner violence. Journal of Interpersonal Violence, 20(11),
1467-1477.

Houry, D., Kemball, R., Rhodes, K. V., & Kaslow, N. J. (2006). Intimate partner violence
and mental health symptoms in afiican american female ed patients. American
Journal of Emergency Medicine, 24(4), 444-450.

Johnson, D. M., Palmieri, P. A., Jackson, A. P., & Hobfoll, S. E. (2007). Emotional
numbing weakens abused inner-city women’s resiliency resources. Journal of
Traumatic Stress, 20(2), 197-206.

Kanan, J. W., & Pruitt, M. V. (2002). Modeling fear of crime and perceived victimization
risk: The (in)significance of neighborhood integration. Sociological Inquiry,
72(4), 527-548.

Kemic, M. A., Holt, V. L., Stoner, J. A., Wolf, M. E., & Rivara, F. P. (2003). Resolution
of depression among victims of intimate partner violence: Is cessation of violence
enough? Violence and Victims, 18(2), 115-129.

Kessler, R. C., Berglund, P., Demler, 0., Jin, R., Koretz, D., Merikangas, K. R., et al.
(2003). The epidemiology of major depressive disorder. Results from the national
comorbidity survey replication (N CS-R). Journal of the American Medical
Association, 289(23), 3095-3105.

Kessler, R. C., Borges, G., & Walters, E. E. (1999). Prevalence of and risk factors for

lifetime suicide attempts in the national comorbidity survey. Archives of General
Psychiatry, 56(7), 617-626.

118

Korbin, J ., & Coulton, C. (1994). Neighborhood impact on child abuse and neglect. Final
Report of Grant No. 90-CA 1494. Washington, DC: National Center on Child
Abuse and Neglect, Department of Health and Human Services.

Krivo, L., & Peterson, R. D. (1996). Extremely disadvantaged neighborhood and urban
crime. Social Forces, 75(2), 619-648.

Kruger, D. J ., Reischl, T. M., & Gee, G. C. (2007). Neighborhood social conditions
mediate the association between physical deterioration and mental health.
American Journal of Community Psychology, 40(3/4), 261-271.

LaGrange, R. L., Ferraro, K. F., & Supancic, M. (1992). Perceived risk and fear of crime:
Role of social and physical incivilities. Journal of Research in Crime and
Delinquency, 29(3), 31 1-334.

Latkin, C. A., & Curry, A. D. (2003). Stressful neighborhood and depression: A
prospective study of the impact of neighborhood disorder. Journal of Health and
Social Behavior, 44(1), 34-43.

Leiner, A. S., Compton, M. T., Houry, D., & Kaslow, N. J. (2008). Intimate partner
violence, psychological distress, and suicidality: A path model using data from
afiican american women seeking care in an urban emergency department. Journal

of Family Violence, 23(6), 473-481.

Leone, J. M., Johnson, M. P., Cohan, C., & Lloyd, S. E. (2004). Consequences of male
partner violence for low-income minority women. Journal of Marriage and
Family, 66(2), 472-490.

Leung, T. W., Leung, W. C., Ng., E. H. Y., & Ho., P. C. (2005). Quality of life of victims
of intimate partner violence. International Journal of Gynecology and Obstetrics,
90(3), 258-262.

Lin, N., & Ensel, W. M. (1989). Life stress and health: Stressors and resources. American
Sociological Review, 54(3), 382-399.

Lipsky, S., & Caetano, R. (2007). Impact of intimate partner violence on unmet need for
mental health care: Results from the NSDUH. Psychiatric Services, 58(6), 822-
829.

Lucente, S. W., Pals-Stewart, W., Richards, H. J ., & Goscha, J. (2001). Factor structure
and reliability of the revised conflict tactics scales for incarcerated female
substance abusers. Journal of Family Violence, 16(4), 43 7-450.

Ludermir, A. B., Schraiber, L. B., D’Oliveira, A. F., F ranca-Junior, I., & Jansen, H. A.

(2008). Violence against women by their intimate partner and common mental
disorders. Social Science & Medicine, 66(4), 1008-1018.

119

Luke, D. (2004). Multilevel modeling. Series: Quantitative Applications in the Social
Sciences, (No. 143). CA: Sage.

MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002).
A comparison of methods to test mediation and other intervening variable effects.
Psychological Methods, 7(1), 83-104.

Marshall, L. L. (1999). Effects of men’s subtle and overt psychological abuse on low-
income women. Violence and Victims, 14(1), 69-88.

Matheson, F. I., Moineddin, R., Dunn, J. R., Creatore, M. I., Gozdyra, P., & Glazier, R.
H. (2006). Urban neighborhoods, chronic stress, gender and depression. Social
Science & Medicine, 63(10), 2604-2616.

McCue, L. M. (2007). Domestic violence (2“d Rev. ed.). Santa Barbara, CA: ABC-CLIO.

Mechanic, M. B., Weaver, T. L., & Resick, P. A. (2008). Mental health consequences of
intimate partner abuse: A multidimensional assessment of four different forms of
abuse. Violence Against Women, 14(6), 634-654.

Menard, A. (2001). Domestic violence and housing: Key policy and program challenges.
Violence Against Women, 7(6), 707-720.

Miles-Doan, R. (1998). Violence between spouses and intimates: Does neighborhood
context matter? Social Forces, 77(2), 623-645.

Mirowsky, J ., & Ross, C. (2003). Education, social status, and education. New York:
Aldine de Gruyter, Inc.

Mulvey, A. (2002). Gender, economic context, perceptions of safety, and quality of life:
A case study in lowell, massachusetts (u.s.a.), 1982-96. American Journal of
Community Psychology, 30(5), 655-679.

Murphy, C. M., & Cascardi, M. (1999). Psychological abuse in marriage and dating
relationships. In R.L. Hampton (Ed.), Family violence prevention and treatment
(2nd ed.). Beverly Hills, CA: Sage.

Nicolaidis, C., Curry, M. A., McFarland, B., & Gerrity, M. (2004). Violence, mental
health, and physical symptoms in an academic internal medicine practice. Journal
of General Internal Medicine, 19(8), 819-827.

Nixon, R. D. V., Resick, P. A., & Nishith, P. (2004). An exploration of comorbid

depression among female victims of intimate partner violence with posttraumatic
stress disorder. Journal of Affective Disorders, 82(2), 315-320.

120

Nurius, P. S., Macy, R. J ., Bhuyan, R., Holt, V. L., Kemic, M. A., & Rivara, F. P. (2003).
Contextualizing depression and physical functioning in battered women: Adding
vulnerability and resources to the analysis. Journal of Interpersonal Violence,
18(12), 1411-1431.

O’Carnpo, P., Kub, J ., Woods, A., Garza, M., Jones, A. S., Gielen, A.C., et a1. (2006).
Depression, ptsd, and comorbidity related to intimate partner violence in civilian
and military women. Brief Treatment and Crisis Intervention, 6(2), 99-110.

O’Leary, K. D. (1999). Psychological abuse: A variable deserving critical attention in
domestic violence. Violence and Victims, 14(1), 3-23.

O’Shea, T. C. (2006). Physical deterioration, disorder, and crime. Criminal Justice Policy
Review, 17(2), 173-187.

Pearlin, L. I. (1989). The sociological study of stress. Journal of Health and Social
Behavior, 30(3), 241-256.

Pence, E., & Paymar, M. (1993). Education groups for men who batter. New York:
Springer.

Perkins, D., & Taylor, R. B. (1996). Ecological assessments of community disorder:
Their relationship to fear of crime and theoretical implications. American Journal
of Community Psychology, 24(1), 63-107.

Perkins, D. D., Wandersman, A., Rich, R. C., & Taylor, R. B. (1993). The physical
environment of street crime: Defensible space, territoriality and incivilities.
Journal of Environmental Psychology, 13, 29-49.

Pico-Alfonso, M. A., Garcia-Linares, M. I., Celda-Navarro, N., Blasco-Ros, C.,
Echeburua, E., & Martinez, M. (2006). The impact of physical, psychological,
and sexual intimate male partner violence on women’s mental health: Depressive
symptoms, posttraumatic stress disorder, state anxiety, and suicide. Journal of

Women ’s Health, 15(5), 599—611.

Plichta, S. B., & F alik, M. (2001). Prevalence of violence and its implications for
women’s health. Women’s Health Issues, 11(3), 244-258.

Ptacek, J. (1997). The tactics and strategies of men who batter: Testimony from women
seeking restraining orders. In A. Cardarelli (Ed.), Violence between intimate

partners: Patterns, causes, and eflects (pp. 104-123). Boston: Allyn & Bacon.

Radloff, L. S. (1977). The CES-D scale: A self report depression scale for research in the
general population. Applied Psychological Measurement, 1(3), 385-401.

121

Raghavan, C., Mennerich, A., Sexton, E., & James, S. E. (2006). Community violence
and its direct, indirect, and mediating effects on intimate partner violence.
Violence Against Women, 12(12), 1132-1149.

Raphael, J. (1996). Prisoners of abuse: domestic violence and welfare receipt. A second
report of the Women, Welfare and Abuse Project. Chicago, IL: Taylor Institute.

Raphael, J. (1999). Keeping women poor: How domestic violence prevents women from
leaving welfare and entering the world of work. In R. Brandwein (Ed.), Battered
women, children, and welfare reform: The ties that bind (pp. 31-43). Thousand
Oaks, CA: Sage.

Raudenbush S., Bryk A. S., & Congdon R. (2005). HLM Version 6.03 Hierarchical
Linear and Nonlinear Modeling. Scientific Software International.

Riger, S., & Staggs, S. L. (2004). Welfare reform, domestic violence and employment:
What do we know, and what do we need to know? Violence Against Women,
10(9), 961-990.

Ross, C. E. (2000). Neighborhood disadvantage and adult depression. Journal of Health
and Social Behavior, 41(2), 177-187.

Ross, C. E., & Mirowsky, J. (1999). Disorder and decay. The concept and measurement
of perceived neighborhood disorder. Urban Affairs Review, 34(3), 412-432.

Ross, C. E., & Mirowsky, J. (2001). Neighborhood disadvantage, disorder, and health.
Journal of Health and Social Behavior, 42(3), 258-276.

Ross, C. E., Reynolds, J. R., & Geis, K. J. (2000). The contingent meaning of
neighborhood stability for residents’ psychological well-being. American
Sociological Review, 65, 581-587.

Sackett, L. A., & Saunders, D. G. (1999). The impact of different forms of psychological
abuse on battered women. Violence and Victims, 14(1), 105-117.

Saltzman, L. E., Fanslow, J. L., McMahon, P. M., & Shelley, G. A. (2002). Intimate
partner violence surveillance: Uniform definitions and recommended data
elements, Version 1.0. Atlanta, GA: National Center for Injury Prevention and
Control, CDC.

Sampson, R. J ., & Raudenbush, S. W. (2004). Seeing disorder: Neighborhood stigma and

the social construction of “broken windows.” Social Psychology Quarterly, 67(4),
319-342.

122

Sandler, I. N., Braver, S., & Gensheimer, L. (2000). Stress: Theory, research, and action.
In J. Rappaport, & E. Seidman (Eds), Handbook of community psychology
(pp. 1 87-23 1). New York: Kluwer.

Schafer, J. A., Huebner, B. M., & Bynum, T. S. (2006). Fear of crime and criminal
victimization: Gender-based contrasts. Journal of Criminal Justice, 34(3), 285-
301.

Schieman, S., & Pearlin, L. I. (2006). Neighborhood disadvantage, social comparisons,
and the subjective assessment of ambient problems among older adults. Social
Psychology Quarterly, 69(3), 253-269.

Schumm, J. A., Hobfoll, S. E., & Keogh, N. J. (2004). Revictimization and interpersonal
resource loss predicts ptsd among women in substance-use treatment. Journal of
Traumatic Stress, 1 7(2), 173-181.

Silver, E., Mulvey, E. P., & Swanson, J. W. (2002). Neighborhood structural
characteristics and mental disorder: Faris and dunham revisited. Social Science &
Medicine, 55(8), 1457-1470.

Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis. Modeling
change and event occurrence. New York, NY: Oxford University Press, Inc.

Skogan, W. G. (1990). Disorder and decline: Crime and the spiral of decay in american
neighborhoods. New York: Free Press.

Skogan, W. G., & Maxfield, M. G. (1981). Coping with crime: Individual and
neighborhood reactions. Beverly Hills, CA: Sage.

Snijders, T. (1996). Analysis of longitudinal data using the hierarchical linear model.
Quality & Quantity, 30(4), 405-426.

Snijders, T., & Bosker, R. (1999). Multilevel analysis: An introduction to basic and
advanced multilevel modeling. London: Sage.

Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural
equation models. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290-312).
Washington, DC: American Sociological Association.

Stafford, M., Chandola, T., & Marmot, M. (2007). Association between fear of crime and

mental health and physical functioning. American Journal of Public Health,
97(11), 2076-2081.

Stark, E. (2007). Coercive control: How men entrap women in personal life. New York:
Oxford University Press, Inc.

123

Steptoe, A., & Feldman, P. J. (2001). Neighborhood problems as sources of chronic
stress: Development of a measure of neighborhood problems, and associations
with socioeconomic status and health. Annals of Behavioral Medicine, 23(3 ), 177-
185.

Straus, M. A. (1979). Measuring intrafamily conflict and violence: The Conflict Tactics
(CT) Scales. Journal of Marriage and the Family, 41(1), 75-88.

Sullivan, C. M., & Bybee, D. I. (1999). Reducing violence using community-based
advocacy for women with abusive partners. Journal of Consulting and Clinical
Psychology, 6 7(1), 43-53.

Sullivan, C. M., Rumptz, M. H., Campbell, R., Eby, K. K., & Davidson, W. S. (1996).
Retaining participants in longitudinal community research: A comprehensive
protocol. The Journal of Applied Behavioral Science, 32(3), 262-276.

Sullivan, C. M., Tan, C., Basta, J ., Rumptz, M., & Davidson, W. S. (1992). An advocacy
intervention program for women with abusive partners: Initial evaluation.
American Journal of Community Psychology, 20(3), 309-322.

Sutherland, C. A., Bybee, D. I., & Sullivan, C. M. (2002). Beyond bruises and broken
bones: The joint effects of stress and injuries on battered women’s health.
American Journal of Community Psychology, 3 0(5), 609-636.

Swanberg, J. E. & Logan, T. K. (2005). Domestic violence and employment: A
qualitative study. Journal of Occupational Health Psychology, 10(1), 3-17.

Taylor, R. B. (2001). Breaking away from broken windows: Baltimore neighborhoods
and the nationwide fight against crime, grime, fear, and decline. Boulder:
Westview.

Theran, S. A., Sullivan, C. M., Bogat, G. A., & Stewart, C. S. (2006). Abusive partners
and ex-partners. Understanding the effects of relationship to the abuser on
women’s well-being. Violence Against Women, 12(10), 950-969.

Thompson, R. S., Bonomi, A. B, Anderson, M., Reid, R. J ., Dimer, J. A., Carrell, D., et
a1. (2006). Intimate partner violence. Prevalence, types, and chronicity in adult
women. American Journal of Preventative Medicine, 3 0(6), 447-457.

Tjaden, P., & Thoennes, N. (2000). Extent, nature, and consequences of intimate partner
violence: Findings fi'om the national violence against women survey. Washington,
DC : US. Department of Justice, Office of Justice Programs. (Report No. NCJ
1 81 867).

124

Truong, K. D., & Ma, S. (2006) A systematic review of relations between neighborhoods
and mental health. The Journal of Mental Health Policy and Economics, 9(3),
137-154.

Turner, R. J ., & Lloyd, D. A. (1995). Lifetime traumas and mental health: The
significance of cumulative adversity. Journal of Health and Social Behavior,
36(4), 360-376.

von Eye, A., & Bogat, G. A. (2006). Mental health in women experiencing intimate
partner violence as the efficiency goal of social welfare functions. International
Journal of Social Welfare, 15(Suppl. 1), $31-$40.

Weinbaum, Z., Stratton, T. L., Chavez, G., Motylewski-Link, C., Barrera, N., &
Courtney, J .G. (2001). Female victims of intimate partner physical domestic
violence (ipp—dv), califomia 1998. American Journal of Preventative Medicine,
21(4), 313-319.

Wells, J. D., Hobfoll, S. E., & Lavin, J. (1999). When it rains, it pours: The greater
impact of resource loss compared to gain on psychological distress. Personality
and Social Psychology Bulletin, 25(9), 1172-1182.

Wen, M., Hawkley, L. C., & Cacioppo, J. T. (2006). Objective and perceived
neighborhood environment, individual ses and psychosocial factors, and self-rated
health: An analysis of older adults in cook county, illinois. Social Science &
Medicine, 63(10), 2575-2590.

Wheaton, B. (1999). Social stress. In C. S. Aneshensel, & J. C. Phelan (Eds), Handbook
of the sociology of mental health (pp.277-300). New York, NY: Springer Science
+ Business Media, LLC.

World Health Organization. (2008, November). Violence against women. Retrieved
February 14, 2009, from http://www.who.int/mediacentre/factsheets/st39/en/

Wyant, B. R. (2008). Multilevel impacts of perceived incivilities and perceptions of
crime risk on fear of crime. Isolating endogenous impacts. Journal of Research in
Crime and Delinquency, 45(1), 39-64.

Xu, Y., Fiedler, M. L., & Flaming, K. H. (2005). Discovering the impact of community
policing: The broken windows thesis, collective efficacy, and citizens’ judgment.
Journal of Research in Crime and Delinquency, 42(2), 147-186.

Zlotnick, C., Johnson, D. M., & Kohn, R. (2006). Intimate partner violence and long-term

psychosocial functioning in a national sample of american women. Journal of
Interpersonal Violence, 21 (2), 262-275.

125

Zule, W. A., Morgan-Lopez, A. A., Lam, W. K., Wechsberg, W. M., Luseno, W. K., &
Young, S. K. (2008). Perceived neighborhood safety and depressive symptoms
among afiican american crack users. Substance Use & Misuse, 43(3), 445-468.

126