 

:3 l...“ .37 a
.. . o...
it}. will
ban, was

.mwu
Jar. :.

 

52.. .. .5;
.... 434.31»?
2:5...
5H 9114»
3.2.,

4 I . ‘ ‘ ‘
é. .. $3.. ‘

 

Lama“

 

3 LIBRARY
aooq Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

A LONGITUDINAL EXAMINATION OF EMOTION
REGULATION ACROSS EARLY DEVELOPMENT:
INFANT ATTACHMENT AND MATERNAL PARENTING

IN THE CONTEXT OF DOMESTIC VIOLENCE

presented by

Carolyn Joy Dayton

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Psychology

 

 

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

 

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./ProgIAcc&Pres/ClRC/DaIeDue,indd

A LONGITUDINAL EXAMINATION OF EMOTION REGULATION ACROSS
EARLY DEVELOPMENT: INFANT ATTACHMENT AND MATERNAL
PARENTING IN THE CONTEXT OF DOMESTIC VIOLENCE
By

Carolyn Joy Dayton

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Psychology

2009

Copyright by
CAROLYN JOY DAYTON
2009

ABSTRACT
A LONGITUDINAL EXAMINATION OF EMOTION REGULATION ACROSS
EARLY DEVELOPMENT: INFANT ATTACHMENT AND MATERNAL
PARENTING IN THE CONTEXT OF DOMESTIC VIOLENCE
By
Carolyn Joy Dayton

Many clinical disorders can be understood as fundamentally related to deﬁcits in
the capacity for emotion regulation (Mullin & Hinshaw, 2007). Attachment theory argues
that regulation strategies develop within the context of the primary attachment
relationship in early childhood (Bowlby, 1969/1982). In addition, exogenous factors such
as trauma exposure may also inﬂuence the development of regulatory capacities
(Osofsky, 1999). Exposure to domestic violence (DV) may be particularly inﬂuential in
that it directly threatens the child’s wellbeing while simultaneously threatening the young
child’s most important regulatory mechanism: the attachment ﬁgure.

In a sample of heterogeneous-for-risk families relative to SES and DV, this study
examined the inﬂuence of infant attachment, cumulative DV exposure and concurrently-
assessed parenting behaviors on the emotional self-regulation and externalizing and
internalizing behaviors of four-year-Old children. Compared to the securely attached
group, children with ambivalent attachment histories demonstrated higher levels of
dysregulation and engaged in increased levels of other-focused regulatory strategies that
were both ineffective and developmentally regressed. Results were broadly consistent
with prior ﬁndings demonstrating the effects of infant attachment on preschool regulation
capacities at the dichotomous (secure/insecure) level of analysis (Gilliom, et al., 2002).

These results add to this literature by examining these constructs at the typological level.

Contrary to predictions, neither infant attachment nor child self-regulation
capacities inﬂuenced child psychopathology. It is argued that the concurrent assessment
of child regulation and psychopathology may have inﬂuenced these results. Speciﬁcally,
it may be that deﬁcits in self regulation only begin to inﬂuence externalizing and
internalizing behaviors as they interact with negative social feedback over time.

Consistent with prior work, maternal DV exposure negatively inﬂuenced positive
parenting (Levendosky & Graham-Bermann, 2000). The current study contributes to the
literature by examining the inﬂuence of cumulative DV across early parenting. In
contrast, child DV exposure did not directly inﬂuence child outcomes. A possible indirect
inﬂuence of DV on child outcomes via its impact on maternal parenting is hypothesized.

Finally, proximal positive parenting negatively inﬂuenced child externalizing
behaviors. In contrast, parenting did not inﬂuence child internalizing behaviors or child
regulatory capacities. Maternal preoccupation with the violence may be associated with a
diminished capacity of the DV-exposed women to attend to and report the subtle signs of
internalizing behaviors. In addition, the lack of association between parenting and child
regulatory capacities is understood as possibly related to the transitional nature of
parenting during the preschool period. Parenting may be transitioning to accommodate
the child’s increased psychosocial Skills such that earlier parenting is more inﬂuential on

the child’s internalized strategies for emotional self-regulation.

To my family.

My mother who loved me ﬁrst.
My father who believed in me always.
My Sister who brings me joy.
My husband who holds my heart.
My niece and nephew whose love feeds my soul.
My son who is my world.

ACKNOWLEDGMENTS

I would like to thank the families who have so generously agreed to participate in
the Mother Infant Study. I would also like to thank the staff of the Mother Infant Study —
both past and present. Without their hard work this project would not have been possible.

A huge thank you is due to my committee members. Dr. Ellen Whipple brought to
this project not only a vast amount of knowledge and expertise, but also a focus on the
Social Work values that she embodies and that I hope to live up to throughout my career.

Dr. Bill Davidson has been a source of guidance throughout the entirety of my
graduate work in Psychology. I will miss the ways in which Bill added both expertise and
good humor to our lab meetings.

It has been my great pleasure to work with Dr. Joel Nigg during my tenure as a
doctoral student. Joel integrates and models the best of Social Work and Psychology in
his work. He has stood up for me when I needed an advocate, pushed me to bring my
clinical and academic work to a higher level and, most importantly, demonstrated
compassion and a wonderful sense of humor in our work together.

Finally, I owe a large portion of my success as a graduate student to Dr. Alytia
Levendosky. My internalized representation Of Alytia has facilitated my professional
development for many years now. She is my mentor, my advisor, and my friend. I have
met very few people in this world who approach life with the level of enthusiasm,
competence, joy and love that she does. She has pushed me to improve my work and She
has encouraged me to remember my priorities, as she always seems to do. Thank you,

Alytia!

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................. ix
LIST OF FIGURES ................................................. x
INTRODUCTION .................................................. 1
CHAPTER 1
EMOTIONS AND EMOTIONAL SELF-REGULATION .................... 9
Emotions: Central to Human Survival and Experience: ............... 9
The Regulation of Emotion: Deﬁnitions and
Developmental Considerations .............................. 16
Emotional Self-Regulation: A Summary and Formulation ............. 23
Measurement of Emotional Self-Regulation in the Current Study ........ 26
CHAPTER 2 .......................................................
ATTACHMENT AND THE DEVELOPMENTAL ETIOLOGY
OF EMOTION REGULATION ........................................ 28
CHAPTER 3
EMOTIONAL SELF-REGULATION IN THE PRESCHOOL YEARS ......... 40
Attachment and Emotional Self-Regulation during the
Preschool Period .............................................. 40
Early Attachment and Later Psychopathology: Extemalizing and
Intemalizing Behavioral Symptoms ........................... 48

The Mediating Role of Child Regulation in the Relation
of Infant Attachment and Child Extemalizing
and Intemalizing Symptoms ................................. 55

CHAPTER 4
INFLUENCING DEVELOPMENTAL PATHWAYS:
THE IMPACT OF DOMESTIC VIOLENCE AND PARENTING BEHAVIORS. 59

The Inﬂuence Of Violence within the Home on Early Development ...... 59
The Inﬂuence of Proximal Parenting Behaviors on Child Outcomes ...... 68
The Impact of Domestic Violence on Parenting Behaviors ............. 77

vii

CHAPTER 5

HYPOTHESES AND RATIONALE .................................... 82
Hypotheses .................................................. 84
SEM Model: Hypotheses One to Five ............................. 85
Additional Analyses: Hypotheses Six and Seven .................... 88

CHAPTER 6

METHOD ......................................................... 90
Participants .................................................. 90
Procedures ................................................... 96
Measures .................................................... 99

CHAPTER 7

RESULTS ......................................................... 109
Preliminary Analyses .......................................... 109
Measurement Models .......................................... 114
Initial Exploration of the Relations of Main Effects Variables .......... 116
Hypotheses One to Five: Full Structural Model ..................... 118
Hypothesis Six ............................................... 121
Hypothesis Seven ............................................. 121

CHAPTER 8

DISCUSSION ...................................................... 125
The Inﬂuence of Infant Attachment on Preschool Self-Regulation

and Psychopathology ...................................... 126

Domestic Violence: Inﬂuences on Parenting and Child Outcomes ....... 139
Parenting and Child Outcomes ................................... 145
Limitations of the Current Study ................................. 147
Clinical and Research Implications ............................... 150
Summary of Study Contribution .................................. 154

APPENDIX ........................................................ 157

REFERENCES ..................................................... 180

viii

LIST OF TABLES

TABLE 1: Ethnicity of Family Members - percentages ..................... 93
TABLE 2: Mother and Father Educational Achievement data — percentages ...... 93
TABLE 3: Mother and Father Occupation Data — percentages ................. 93

TABLE 4: Socioeconomic Status Data by Wave of Data Collection: T1 - T4 ..... 94

TABLE 5: Maternal Rates of DV Exposure by Year (% abused) .............. 94

TABLE 6: Strange Situation Test Protocol ................................ 100
TABLE 7: Strange Situation Classiﬁcations — One-year Old children ........... 101
TABLE 8: Weighted Kappa Reliability Statistics for the Parenting Data ......... 104

TABLE 9: Weighted Kappa Reliability Statistics for the Child Regulation Data. . . 108
TABLE 10:Means and Standard Deviations Of Observed Variables in SEM Model. 110
TABLE 11:Correlation Table of Observed Continuous Variables ............... 111

TABLE 12:Point Biserial Correlation Table Of Observed Continuous Variables
and Categorical Attachment Variables (dummy—coded) ...................... 113

ix

LIST OF FIGURES

FIGURE 1: Study Model .............................................. 84

FIGURE 2: Child Positive Regulation (mean) by Attachment Category .......... 117
FIGURE 3: Full Structural Model ........................................ 120
FIGURE 4: Other-Focused Behaviors (mean) by Attachment Classiﬁcation ....... 123

A Longitudinal Examination of Emotion Regulation across Early Development:
Infant Attachment and Maternal Parenting in the Context of Domestic Violence
INTRODUCTION

The capacity for emotion regulation has been described as central to healthy
psychological functioning in both child and adult populations (Bradley, 2000; Gross,
2007). Relatedly, the theoretical relationship between emotional dysregulation in early
development and the development of psychopathology has recently received increased
empirical attention (Eisenberg etal., 2001; Hill, Degnan, Calkins, & Keane, 2006; Lewis
et al., 2008). Indeed, the most prevalent forms of psychopathology can generally be
subsumed under the higher order categories of internalizing and externalizing behaviors
(Lahey et al., 2008), and are increasingly being described as directly related to deﬁcits in
the capacity for emotion regulation (Campbell-Sills & Barlow, 2007; Mullin & Hinshaw,
2007)

Two of the primary theoretical orientations that have informed the exploration of
child regulation capacities beginning during the infancy period are the temperament and
attachment ﬁelds. Both of these literatures argue for causative links between early infant
regulation processes and subsequent self-regulation abilities later in child and adult
development (Calkins, 2004; Cassidy, 1994; Crittenden, 1995; Gunnar, Mangelsdorf,
Larson, & Hertsgaard, 1989; Rothbart, Posner, & Kieras, 2006; Seifer, 2000). While
empirical investigations examining the inﬂuence of early child temperament on later
regulatory and psychosocial functioning have been widely reported (for a review see

Nigg, 2006), the link between early attachment experiences and later child regulation

capacity has received almost no empirical attention (Calkins, 2004). The current study
ﬁlls this empirical gap by investigating the relationship between infant attachment
strategy and later capacities for emotional self-regulation during the preschool period. In
addition, the inﬂuence of self-regulatory deﬁcits on the development of externalizing and
internalizing behavioral symptoms is explored.

The present review examines the constructs of emotion and emotion regulation
from a multi-disciplinary and multi-modal (e. g., emotional, behavioral, cognitive)
perspective. The human emotional system, as well as the processes that regulate it, are
understood to be complex and dynamic. Consequently, deﬁnitional consensus both across
and within disciplines has been somewhat elusive. In a recent inter-disciplinary
integration, Gross and Thompson (2007) deﬁned emotion as an integrated multisystem
response to a person-situation transaction that both compels the attention of an individual
and to which the person ascribes particular meaning. Further, emotion regulation is
understood to occur when a person Shifts from a dysregulated emotional state (the
‘multisystem response’) to a regulated state, regardless of the process that facilitates this
shiﬁ. This process, for example, is evident from the beginning of life, when a newborn
cries due to hunger, is fed a bottle, and stops crying; the caretaker’s intervention results in
the infant moving from a dysregulated to a regulated emotional state. In contrast, the
process of emotional self-regulation does not become possible until later in development
(Rothbart et al., 2006; Sroufe, Egeland, Carlson, & Collins, 2005) AS deﬁned in this
review, and described in greater detail in chapter one, the deﬁnition of emotional self-
regulation involves the ability to: 1) m emotionally-salient stimuli, 2) m

adjust one’s reactions to emotionally-salient contextual stimuli such that, 3) one’s

experiences and behaviors are modiﬁed when necessary to meet one’sgoals. This process

 

can include both conscious (deliberate) and unconscious (automatic, learned) responses
to internal (cognitive) and external (environmental) stimuli. The emergence and
development of emotional self-regulatory capacities unfolds over the course of early
childhood. For example, there are discrete self-regulatory behaviors (e. g., gaze shifting)
that become available in infants as young as three to four months of age (Rothbart et al.,
2006). However, as described later in this review, a coordinated, integrated and
internalized system of emotional self-regulation does not become consolidated until
children reach preschool age (e. g., 3.5 to 4.5 years) (Kopp, 1989; Sroufe et al., 2005).
Fundamentally a theory of emotion regulation (Schore, 2003), attachment theory
posits that early interactions with the primary caretakers(s) are internalized in the young
child through the formation Of working models that are learned during infancy and, over
time, come to guide automatic behaviors (Bowlby, 1969/ 1982; Main, Kaplan, & Cassidy,
1985; P. Zimmermann, 1999). Initially driven by the infant’s instinctual fear of
separation from the caretaker, attachment theory argues that, by the ﬁrst year of life, the
infant has developed emotional and behavioral regulation strategies that are based on
repeated experiences with the caretaker and are designed to keep the caretaker physically
close (Ainsworth, 1993). In early infancy, when the child becomes upset, he or she
requires actual physical proximity to the caregiver in order for emotion regulation to
occur. That is, although temperamental qualities are evident at birth that result in
differential thresholds for becoming dysregulated (e. g., “fussy” versus “calm” babies),
the very young infant is not yet capable of independently adjusting his or her reactions to

meet his or her goals. Over time, as development unfolds, regulation capacity and

regulatory strategies are thought to become internal personality characteristics (e. g.,
schemas, templates, representations, or working models) of the developing child such that
they can be described as emotional self-regulation strategies. Thus, the attachment
behaviors that were initially driven by the infant’s fear of separation from the caretaker
are ultimately consolidated into an internalized and unconscious template of emotion
regulation that is generalized to other relationships and Situations (Bowlby, 1969/1982).
In this way, the regulation of emotion moves increasingly from a parent-mediated process
(6 - g -, the mother physically comforts the crying newborn) to a child-controlled process
(e - g -, the child distracts herself by playing with a toy when her mother leaves the room)
over the course of development (Kopp, 1989; Sroufe, 1995). Developed, in part, within
the context of the child’s attachment relationship(s) during infancy, the child’s regulation
Strategy manifests in a more complex manner later in development; a process that
becomes consolidated during the preschool years (e.g., approximately 3.5 to 4.5 years of
age) - This capacity emerges in the preschool-aged child partly as a function of the
81’0an ability for attentional control and cognitive ﬁmctioning (Crittenden, 1995; Kopp,
1 98 9; Rueda, Posner, & Rothbart, 2005; Sroufe et al., 2005).
While early secure attachment is believed to be important to the healthy social-
emotjonal development of the young child, many other factors are also inﬂuential. That
is, <3 onsistent with the framework of developmental psychopathology, a secure attachment
rel ati (Inship is understood here as one important and proximal characteristic which
Col1":I‘ibutes to adaptive regulation capacities in the young child (Greenberg, 1999).
H0 wever, it is primarily the conﬂuence of multiple risk and protective factors which best

Ma -
d 1 cts to child social-emotional outcomes (Sameroff, Lewis, & Miller, 2000). During

- 4

 

 

early childhood development, environmental factors that inﬂuence the primary
caretaker’s parenting capacities are likely to be especially inﬂuential on child
development outcomes.

In light of this, the current study examined regulatory capacities in a sample of
mothers and children, many Of whom have been exposed to domestic violence (DV)
across the ﬁrst 4 years of the child’s life. During the infancy and early childhood periods
of development, trauma perpetrated toward a child’s attachment ﬁgure is likely to have
unique and enduring social-emotional effects due to the centrality of this relationship to
the child’s early growth and development (Lieberman & Amaya-Jackson, 2005; Osofsky,
2004). Speciﬁcally, in the case of DV, the child is exposed to danger that can be expected
to activate his or her need for physical proximity to their caregiver, and yet, the
caregiver’s physical integrity is being actively threatened. In addition, maternal parenting
is likely to be negatively inﬂuenced by the violence perpetrated against her (Levendosky
& Graham-Bermann, 2001a), thereby further compromising the child’s development of
healthy and adaptive regulatory capacities (Davies & Cummings, 1994).

The current study examined the relationship between an infant’s attachment
category with the mother at one year of age, and the child’s later capacity for emotional
self”regulation during the preschool period, using a longitudinal design in a sample of

hetero geneous-for-risk children (e.g., relative to DV exposure and SES). Drawing from
an esllablished coding scheme, emotional self-regulation was measured using observer-
rated behavioral coding of regulatory behaviors in a laboratory environment when the
chi ld was four years of age (Whipple, Denburg, & Davies, 1993; Whipple, Fitzgerald, &

Z
”Ck": 1995). The relationship between the child’s observer-rated regulatory capacities

and parent-rated behavioral reports of externalizing and internalizing symptomatology
was also examined (Achenbach, 1991). In addition, the inﬂuence Of DV exposure and

current parenting behaviors on the child’s regulatory capacities and behavioral symptoms

of psychopathology was examined.
This study is unique in that it examined the direct inﬂuence of infant attachment

category on the child’s later, observer-rated capacity for emotional self-regulation.
Although this relationship has been clearly hypothesized within the theoretical literature
(Bowlby, 1969/1982; Calkins, 2004; Guttmann-Steinmetz & Crowell, 2006), it had not
been directly, empirically examined. Importantly, although the two organized insecure
attachment categories have been theoretically hypothesized to inﬂuence differential
t‘ e glllatory strategies (over- versus under-controlled), this hypothesized relationship does
r:l()t suggest that insecure attachment should necessarily result in later psychopathology
( Guttmann-Steinmetz & Crowell, 2006). Instead, from a developmental psychopathology
fi‘amework, insecure attachment is thought to function as a risk factor for later
P SD’Chopathology in the context of other environmental risk factors. This study examined
be Se relationships in a sample of mothers and children, many of whom are at further

p 83” ehosocial risk for regulatory deﬁcits due to their exposure to DV and lower SES. The

temﬁal Inﬂuences of DV on the child’s regulatory capacmes and on the mother’s

ﬁ=11ting behaviors have been well-described within the theoretical literature (Davies &

pit‘

C
Mugs, 1994; Levendosky & Graham-Bermann, 2001a; Lieberman & Amaya-

Iii-Q
keen, 2005; Osofsky, 2004). However, while the inﬂuence of DV on parenting
be
l:t-‘§.‘Iiors had been previously examined empirically, the hypothesized relationship

bet:
Ween DV exposure and child regulatory capacities had not. Finally, the current study

 

examined the mediating inﬂuence of regulation deﬁcits on the relationship of infant
attachment strategy and later symptoms of child psychopathology. In examining the
relationship between child regulatory capacity and psychopathology this study adds to the
growing number of empirical studies that have recently begun to examine this
relationship (Hill et al., 2006; Stieben et al., 2007). By furthering an understanding of the
development of regulatory capacities across the ﬁrst four years of life, as well as their
inﬂuence on the development of child psychopathology, this study facilitates a richer
theoretical and clinical formulation of how these processes unfold in the lives of DV-
exposed young children. In addition to contributing to the empirical literature, results
ﬁ-orn this study may help to guide future prevention and intervention efforts aimed at
reducing the impact of DV on the lives of infants and young children (Graham-Hermann,
Lynch, Banyard, DeVoe, & Halabu, 2007; Shavers, Levendosky, Dubay, Basu, & Jenei,
2 oo 5).
The current literature review is organized into chapters. Chapter One reviews the
Cl'I-e-Eoretical and empirical work examining the structure and function of human emotions
mg emotion regulation. Chapter Two describes the foundation of the current study which
Peroses to examine the inﬂuence of early attachment experiences on later child
Fa gulation capacities during the preschool period. In this chapter attachment theory and
mg relation to the etiology of emotional self-regulation is reviewed. Chapter Three
SLI'T‘IITIJTIarizes the extant research examining the inﬂuence of early attachment history on
111.5% . . . . . .
r Chlld regulation capacrtles and psychopathology. Chapter Four rev1ews the literature
ex: Ming the inﬂuence of DV exposure on both child regulatory capacities and maternal
lie“

§mting behaviors. In addition, the inﬂuence of concurrently assessed parenting

behaviors on child regulatory capacities is also reviewed. Chapter Five presents the
hypotheses and rationale of the current study. Chapter Six describes the methodology of
the study. Chapter Seven presents the results of the analyses, and Chapter Eight discusses
these results, draws conclusions, and outlines the clinical and research implications of the

ﬁndings.

CHAPTER 1
EMOTIONS AND EMOTIONAL SELF-REGULATION
The current study was concerned with investigating the developmental pathways
leading to emotional self-regulation capacities in a heterogeneous-for-risk sample of
preschool aged children. The constructs of emotion, emotion regulation and emotional
sel f-regulation are complex in nature and comprehensive deﬁnitions require a
multidisciplinary theoretical approach. The following review summarizes the leading
theories which have informed current conceptualizations. Following this, Chapter Two
outlines the developmental trajectory of the primary attachment relationship on the
capacity for emotion self-regulation across the ﬁrst four years of the child’s life.
motions: Central to Human Survival_ and Experience
An understanding of the regulation of emotion hinges on a clear conceptualization of
t he construct of emotion itself. In fact, since William James’ seminal paper, What is an
emOIion? (James, 1884), psychologists have been struggling to understand, deﬁne, and
S tudy this construct. From an evolutionary perspective, it is clear that emotions are
a"(:1‘a13tive in promoting the survival of the species (Darwin, 1872/1965; LeDoux, 1996).
The well-documented ‘ﬁght, ﬂight or freeze’ responses, for example, are dependent both
011 1116 ability to perceive a dangerous stimulus (e. g. a functioning sensory system) as well

as
ﬁle elicitation of the emotional experience of fear to cue the organism to react

(D
M011, 1929; Gray, 1988). Indeed, attachment theory is based on the importance of the

cw“

l intionary adaptability of emotions. That is, the ability to experience and express
fe§ l e

‘- lags of distress on the part of the infant (e. g. crying) serves an adaptive and protective

101
§ i 11 cueing the mother to feel a sense of fear for her offspring. In response to this

 

feeling, she is compelled to seek contact with, and thereby protect, her infant. Hoeksma
and colleagues (2004) have stated that the emotional system, “is inherent in the process
of promoting and maintaining life, and is always poised to prevent the loss of physical
and psychological integrity” (p. 355). However, despite the fundamental importance of

emotions in promoting human survival, deﬁnitional agreement across disciplines remains

elusive.
While disagreements persist, a review of the literature suggests that emotions are best

understood as dynamic processes which are rooted in the neurobiological makeup of
human beings and are interrelated with cognitive and behavioral processes (Cole, Martin,
& Dennis, 2004). In his work on fear reactions, LeDoux (1996; 2000) has argued for a
two —pronged approach to conceptualizing human emotions. Using primarily animal
models, LeDoux and his colleagues and contemporaries have used relatively simple
methodologies that employ basic behavioral strategies such as pairing methods to
condition a fear response to an otherwise neutral stimulus (for reviews see, LeDoux,
B 000; Maren, 2001). The biological and neurological implications Of these responses are
h en examined using established surgical techniques in order to identify the brain

h‘uctures that are associated with the animal’s behavioral and physiological responses.
I)“ . . . . . . . .
e to their srmplrcrty and the reduced ethical constraints of work wuh animals versus

h
mans, these studies have yielded valid and reliable results that have been replicated by
0%
Q1 research teams. Subsequent to the identiﬁcation of these neurological processes in

an! ‘
l ﬁnals, more complicated methodologies have been employed with human populations

Hg ~
1 O O O O C 0
mg methods such as tMRI Imagmg In order to begin to examrne these processes In

h
New (e.g., LaBar, Gatenby, Gore, LeDoux, & Phelps, 1998).

10

L

Based on this body of research, LeDoux has argued that, in the case of fear, emotions
can be understood as being manifest at two levels. The ﬁrst level can be described as
“bottom-up” in that it involves an immediate fear reaction that is processed outside of
consciousness by the amygdala and is not subject to voluntary control. This process is not
mediated by cognition and represents neurological mechanisms which evolved relatively
early in humans. An example is the immediate urge to freeze or run at the sight of a
snake. This urge is relatively automatic and does not require higher levels of mental
processing to take place before a person responds physiologically and even behaviorally.
Re search with human populations has provided consistent support for the role of the
amygdala in responding to the detection of emotionally salient (especially negative)
Stimuli (for a review see Phillips, Drevets, Ranch, & Lane, 2003). In particular,
methodological approaches that utilize functional neuroimaging procedures to examine
i nereases in blood ﬂow and activation within amygdala have corroborated ﬁndings from
Work with animals and have also yielded valid, reliable and well-replicated results.
Unlike the primary emotional response of the amygdala, the second fear reaction
mechanism proposed by LeDoux (1996; 2000) is closely tied to higher cognitive
amctioning and is hypothesized as being processed in the neocortex (e.g., “top-down”).
L'Q boux has argued that this process is under voluntary control, is tied to executive
11‘ hetioning abilities and is inﬂuenced by a person’s recollections of past experiences.

1‘11

re
m ization that one is in danger, the experience or ‘feeling’ of fear, followed by the

0%
gdtive processing of what one should do when one encounters a snake. Research with

115, for example, the secondary fear reaction to seeing a snake may include the

h
“T512111 subjects has recently begun to support this second part of LeDoux’s model. For

11

example, using f MRI methodology, Phan and colleagues (Phan et al., 2005) exposed
adult participants (n=l4) to pictures of aversive and arousing stimuli (e.g., pictures of
burn victims) and, in alternating conditions, asked participants to either suppress or
maintain their emotional response to the pictures. The visual stimuli utilized in this study
had been validated in prior research and were standardized with respect to their
presentation to the subjects. Subjects were screened to insure that their vision was normal
and that they did not endorse current or historical symptoms of psychiatric or neurologic
illness. To facilitate the suppress condition participants were initially taught how to use
tile technique of cognitive reappraisal of the stimulus in order to regulate their negative
emotional reaction (e. g., imagining that the burn victim was just an actor wearing
makeup). In this way, researchers documented changes in neurological functioning
d uring the process of emotion regulation. In support of LeDoux’s hypothesis, results
demonstrated that exposure to the negative stimuli across both conditions evoked
imrlwdiate activation of the amygdala. In addition, during the suppress condition several
aTorltal regions of the brain were recruited in the effort to regulate (e.g., reappraise) the
lie gative stimuli and the cognitive reappraisal strategy was effective in reducing the
lg SQciated negative affect. Consistent with these ﬁndings, the cognitive behavioral
hél‘apy (CBT) literature has demonstrated that negative emotional reactions to both
hx’ironmental stimuli and internal cognitions can be modiﬁed by the use of cognitive
r§
aibpraisal/restructuring and can then result in different reactions when one encounters
that . . . .
stimulus (external or internal) 1n the future (e.g., Dav1d-Ferdon & Kaslow, 2008).
Within the clinical psychology literature, the syndrome of alexithymia also provides a

u&§
:54] example of LeDoux’s proposed dichotomy. Krystal (1988) has described

12

alexithymia as a condition wherein a person is experiencing the neurobiological
correlates (primary emotion functioning) of a given emotion but is utterly unable to

cognitively process that emotion (secondary functioning). These individuals can describe

physical signs within their body (e. g. the awareness that tears are coming out of their

eyes) but they are unable to cognitively identify a feeling state and, therefore, have no
way of understanding, articulating, or controlling it. Alexithymic individuals may have
tlle capacity to respond in a physiologically congruent manner to a situation (e. g. crying
at a funeral) but their ability to cognitively process their feelings in a way that allows
tllern to internally make sense of them or to articulate their emotional experience is
ab sent. As a result, their ability to integrate the physiological and cognitive aspects of
‘tlleir emotional experiences is impaired and their reactions to emotionally disturbing
Stimuli tend to be processed as physiological (somatisized) manifestations but not as
cOnscious affective states. In this way, alexithymia essentially could represent a regressed
psychological state in adult populations which is presumably similar to the emotional
experiences of an infant. That is, at least on a physiological, but not necessarily a fully
e‘Ol'lscious level, the person perceives the stimulus (e.g., hears that someone has died), and
baa—cm to it in a manner that is congruent with the situation (e. g., cries). However, the

I arson cannot cognitively integrate or understand the relationship between the stimulus

‘Q their physiological reaction to it.
Growing empirical evidence has examined the alexithymic patient’s inability to

CD.
gntively process physiologic emotional reactions (Luminet, Rime, Bagby, & Taylor,
2C)
Q4; Wearden, Cook, & Vaughan-Jones, 2003). For example, results from one study

’6‘;-

‘ﬁaled that, in response to a sad movie, adults who received elevated scores on a

13

measure of alexithymia demonstrated elevated heart rates (Luminet et al., 2004).
However, self-report evaluations of their cognitive and affective processing of the movie
content revealed that they experienced the movie as less negative and less important than
other subjects and tended to think about (ruminate) the movie less than others. Thus,
while their primary emotion processing was elevated, their secondary functioning was
incongruent with their physiological response. This study utilized a standardized, well-
validated, self-report measure of alexithymia in addition to measures of physiological
reactions (e.g., heart rated, blood pressure) that were continuously recorded using
computer technology. Participants were recruited from classes for senior citizens (mean

age = 63.6). Thus, while the methodology utilized in this study was well-controlled,

re Slllts may not generalize to other, younger, groups of adults.
Within the child development literature, Cole and colleagues (Cole et al., 2004) have

SWaﬁzed several common assumptions about the deﬁnition of emotion shared by most

theorists and researchers in this area. First, and consistent with neo-Darwinian theory,
emotions are assumed to be biologically-based and adaptive from an evolutionary
peI‘Spective. Secondly, they constitute an appraisal system which allows for constant

1 Qtlitoring of the environment with the goal of cueing rapid context-speciﬁc reactions to

c O O O O I O O
13-. angmg Situations and Circumstances. They have described emotions as, “appraisal-
ae ‘I: ~ . . . . . .
l on readiness stances, a ﬂuid and complex progressron of orienting toward the ongomg

Str
‘é am of experience” (p 320). In this way, for example, the experience of positive

ellj
Qtions would be expected to cue the individual to continue their current behaviors

WEI
Qreas negative emotions would likely lead to strategies designed to effect situational

€11
\ge. Finally, they argued that emotion is best understood as a dynamic process that,

14

although it is always operating, does not always reach the level of conscious awareness
(e.g. as in the individual with alexithymia).
Hoeksma and colleagues (2004) have articulated a similar theoretical approach to
understanding human emotional systems as dynamic in nature. Citing the research which
demonstrates that the emotional system is rooted in the neurological structures of the
brain, they proposed that the positive and negative feedback loops (labeled emotion
circuits by LeDoux, 2000) which result from the reciprocal projections of these
neurological structures ﬁrnction constantly in order to facilitate continuous and dynamic
changes within the emotional system. Similar to the arguments made by Cole and
col leagues (2004), Hoeksma and colleagues (2004) argued that the healthy functioning of
tl'liS system results in the ability to perceive and appraise situations in terms of the degree
to which they promote the person’s goals (e. g. the goal to stay safe and not suffer a snake
bite). In addition, Hoeksma and colleagues (2004) identiﬁed the construct of feelings as
diStinct from emotions. Similar to LeDoux’s model, they deﬁned feelings as involving
hi gher order cognitive processing and serving the role of monitoring the dynamic and
e"et~cha.nging emotional system. They argued that feelings are, “the private mental
ex perience of emotions” (p. 355). Their model predicts that feelings are somewhat

id - . . . . . . . . .
‘- Osyncratlc to each 1nd1v1dual and that they allow 1nd1v1duals to make cognitive sense

a F
the dynamic, neurologically-based, and constantly operating emotional system.

“Taken together, these various conceptualizations are broadly consistent with the inter-
di
Q‘Q iplinary integration recently provided by Gross and Thompson (2007), who deﬁned

e111 Q‘ﬁon as, “A person-situation transaction that compels attention, has particular

In Q
ELI-ring to an individual, and gives rise to a coordinated yet ﬂexible multisystem

15

D
I—‘L

‘M.

P.
All“

response to the ongoing person-situation transaction” (p. 5). In other words, the

‘response’ must be related to an event, either internal or external, and is experienced by

the person as meaningful in some way. In the next section, the construct of the regulation

of emotion (e.g., regulation of the ‘response’) will be examined.

'I‘he Regulation of Emotion: Deﬁnitions and Developmental Considerations

Like emotions, emotional regulation processes also serve basic and adaptive goals

of the individual. As children move into the preschool years (e.g., 3.5 — 4.5 years of age),
behaviors reﬂecting emotional dysregulation, which are generally tolerated by the social
envrironment during the infancy and toddler periods, become the targets of social

intervention (Sroufe et al., 2005). To a much greater degree than toddlers, preschool-aged

children are both expected and capable of controlling their emotions in the service of

their own, and society’s goals.
Related to the continued debate over the nature, structure and function of

e1lilotions, agreement about the deﬁnition of emotion regulation has also been elusive
even within the child development literature. Further, the constructs of emotion and
el‘liflotion regulation are often used interchangeably and have proven difﬁcult to
di Sentangle both theoretically and empirically (Cole et al., 2004). In their attempt to

d e

'1 ﬁerentiate these constructs, Hoeksma and colleagues (2004) have used signal and
s
3"Pg1iem theory as an organizing framework. They described the difference between

€111

‘Otion and emotion regulation as similar to the difference between a physical dynamic
SS)—
§tem and the mechanism by which that system is controlled. Using this heuristic, they

0L1
t l 1 ned three deﬁning elements of emotion regulation. The ﬁrst involves the conditions

e ssary to trigger active control of the system. In the case of emotions, regulation

IleQ

l6

 

becomes necessary when the child’s feeling state (e. g., an infant’s fear that mother has

left the room) signals that the set goal of the emotional system, which is deﬁned as
maintaining physical and psychological integrity, is discordant with the current situation.
For example, the fact that the mother is physically absent is likely to result in upset
feelings on the part of the infant or toddler and thus, to be divergent from the set goal of
the emotional system. The second element of emotion regulation involves the goal of the
regulation process itself. Here the goal of emotion regulation is deﬁned as returning the
emotional system to a state of relative equilibrium (e. g., a reduction in the feeling of
fear). Finally, the process of emotion regulation is achieved by altering the input into the
emotional system. In the case of the infant or toddler left alone in the room this can be
accomplished, for example, by diverting her attention to some other activity (e. g., playing
With a toy) or actively seeking to reconnect with the mother (e. g., calling for her mother
to return), or, in the case of a preschool aged or older child, by using a cognitive strategy

Such as remembering that her mother always comes back when she has to leave for a

811011 period (e.g., using a form of reappraisal).
In a recent review, Cole, Martin and Dennis (2004) provided an important

1 S‘Uiinction between emotion and emotion regulation. They argued that emotions “infuse

a):
Deﬁence with meaning. . . Emotion allows us to evaluate steadily and quickly whether

it: —
l S in our interest to stay in the chair and to act instantly if we need to escape through the

de

e& Q are”
PF‘Q
eess of actively attempting to regulate one’s emotional state and to the process

Qt. Emotion regulation helps us stay in the chair even when we feel compelled to
(p. 318). They suggested that the term emotion regulation applies both to the

17

 

whereby an activated emotion (e.g., happiness) serves to regulate other systems (e.g.,

memory, learning) in an automatic (e.g., non-effortful) manner.

In contrast to this broad deﬁnition, Eisenberg and Spinrad (2004) have argued for
a more narrow view of emotion regulation. In their View the term emotion regulation as
used in child psychology research should be conﬁned to the process of emotion as
re g1_1_lated and not to the process of emotion as a regulator. In addition, they have used the
tem emotion-related self-regulation to highlight the fact that this construct refers to the
ongoing modulation of the emotional and physiological state of the individual as well as
to the modulation of overt behaviors that are related to emotional experience including
behaviors that are intended to inﬂuence the social context in an effort to regulate

emotion. These authors deﬁned emotion-related self-regulation as:

The process of initiating, avoiding, inhibiting, maintaining, or modulating the
occurrence, form, intensity or duration of internal feeling states, emotion-related
physiological, attentional processes, motivational states, and/or the behavioral
concomitants of emotion in the service of accomplishing affect-related biological
or social adaptation or achieving individual goals. (p. 338)
I‘llis formulation is critical in the distinction it makes between emotion regulation and

em Otional self-regulation. That is, the construct of emotion regulation does not
héeessarily specify the way in which emotion is regulated. For example, environmental
he tors that are completely beyond the person’s control can, nevertheless, result in a shift
burn a dysregulated to a regulated state (e. g., a tornado watch is called off and the person

Tb!
ﬁxes). However, most systems of emotion regulation that are proposed and discussed

Wi
min the child development literature refer to an active process on the part of the

18

individual child or adult. This process may be or become unconscious over the course of
time as with internal working models that are based on attachment experiences. They are
not, however, accidental processes. That is, whereas emotion regulation refers only to the
shift from a dysregulated to a regulated emotional state, emotional self-regulation refers
to a process (conscious or unconscious) that is active on the part of the individual.
Eisenberg’s (2002) conceptualization of emotional self-regulation was based, in
part, on basic tenets of attachment theory (Bowlby, 1969/1982) as well as the inﬂuential
work of Block and Block (1980) within the personality literature. To date, attachment
theory’s most important contribution to investigations of emotional self-regulation
involves a theoretical description of the ways in which early relationship experiences lead
to basic patterns of emotion regulation in the infant and young child through the
fomation of internal working models. Due to the centrality of the attachment construct

Within the current investigation, the relationship between attachment and emotional self-

l"'egulation is more fully explored in chapter two.
While they were less interested in the etiology of regulation capacities, Block and

B l Ock (1980), in their inﬂuential work on this topic, examined the personality
CleWielopment of young children. These investigators conducted comprehensive
as Sessments of a sample of primarily Caucasian, middle- to upper-class, children
I3'Eginning at three years of age and continuing across the subsequent four years of their

Q Velopment. Subjects were recruited at two universrty nursery schools. They reported an

R
0% retention rate across this four year period. Individual assessments as well as parent
it?!
thviews with both the mothers and fathers of these children were obtained. In addition,

"1
1‘ 1 tiple well-validated measures of child functioning were included in the protocol,

19

 

including measures of intelligence and receptive and expressive language (e.g., WPPSI,
PPVT). In addition, qualitative measures were developed and used that tapped both
teacher and parent report of child functioning. In addition, the fact that these children
were available for evaluation on a daily basis facilitated the collection of a wide range of
observational data. Overall, these researchers conducted comprehensive evaluations of
multiple domains of child functioning.

In their synthesis of the results of this study, Block and Block (1980) argued that
tile existing personality systems of individuals can be understood as related to two basic
ego structures: ego control and ego resiliency. Ego control represents the person’s modal

l evel of “impulse control and modulation” (p. 41). Within this model, two subtypes of
exaggerated ego control tendencies were identiﬁed. At one end of the spectrum are
Overcontrolled individuals who, irrespective of social context, tend to be emotionally and

behaviorally constrained and inhibited. On the other end, undercontrolled individuals
tend to be extremely behaviorally impulsive and to express their emotions in a manner
that demonstrates a high level of emotional lability. These investigators argued that,

‘ ‘extreme placement at either end of the ego-control continuum implies a constancy in
‘11 Ode of behavior that, given a varying world, can be expected to be adaptively
dysfunctional” (p. 44). Therefore, within non-clinical samples the majority of individuals

WQ 111d be expected to fall somewhere within the mid-range of this continuum and

dél’honstrate ego-control tendencies which reﬂect relative levels of over- or under-

cg
“Ed.

The concept of ego resiliency, in comparison, describes an individual’s capacity

it)
qD’namically modify their modal level of ego-control in order to adapt to changes

20

 

within the environmental context. Thus, the construct of ego-resiliency is most suggestive
of the capacity for healthy psychosocial functioning and is very similar to Eisenberg’s
current conceptualization of emotion-related self-regulation. In fact, Eisenberg and
colleagues (Eisenberg et al., 2001) have argued explicitly that children who are well-
regulated are neither over- nor under-controlled. Instead, consistent with the notion of
ego-resiliency these authors argued that the capacity for emotional self-regulation
ixlvolves the ability to adapt to contextual and situational changes in environment in a
ﬂexible and spontaneous manner and to alter one’s reactions when appropriate. Sroufe
and colleagues (Sroufe et al., 2005) have argued that this capacity becomes available
when children reach preschool age which they deﬁne as emergent at 3.5 years of age and
ﬁll)! consolidated by 4.5 years of age. At this point cognitive, attentional, and language
abilities are sufﬁciently developed to support the child’s own self-regulation of her
emotional system (Kopp, 1989). The developmental process of emotional self-regulation
as it manifests in the preschool period will be more fully described in chapter three.

In addition, Eisenberg (2002) contended that emotion-related self-regulation is
beSt understood as a dichotomous construct that includes both involuntary (reactive or
i1j-L'llnilsive; but not inactive/accidental) as well as voluntary regulation. Although she did

not reference his work, this model is reminiscent of LeDoux and colleagues’ (LeDoux,

l 9 96, 2000; Sotres-Bayon, Bush, & LeDoux, 2004) conceptualization of emotion as

in V0 lving two levels of neurological activation; primary emotional functioning controlled
by the amygdala (e.g., involuntary emotional control) and secondary functioning
méq i rated by the neocortex (e.g., voluntary emotional control). Eisenberg described

”01'
112:2:on control as overlapping with Rothbart and Bates’ (1998) notion of effortful

21

control: “the ability to inhibit a dominant response to perform a subdominant response”
(p. 137). However, Eisenberg’s deﬁnition was somewhat broader in scope as compared to
Rothbart’s. While both included attentional regulation and inhibitory control (in the
service of a goal), Eisenberg’s conceptualization included the ability to engage in
undesired behaviors to attain a desired goal (e.g., cleaning one’s room to earn an
allowance). Consistent with LeDoux’s formulation, Eisenberg (2002) drew on work by
Posner and DiGirolamo (2000) who have argued that the frontal structures in the cortex
are related to voluntary control, just as they are to executive functioning. Further, she
argued that involuntary (e.g., reactive, impulsive) control is associated with subcortical
systems such as the amygdala and reﬂect temperament constructs (which she deﬁnes as
based in physiological processes) such as fearful avoidance.
Although Eisenberg (2002) contended that her conceptualization of emotion-
related self-regulation was based, in part, on attachment theory, attachment theory does
not specify types of emotion regulation beyond the emotionally deactivated (avoidant;
o‘Iercontrolled) and emotionally hyper-activated (ambivalent; undercontrolled) categories
(described in Chapter 2). However, because voluntary control does not become fully
consolidated until the preschool years (Kopp, 1989), it is likely that the infant’s early
a-‘:‘:achment experiences are related in part to reactive processes that are dependent on the
£31'13rsiological makeup of the child such as the child’s temperamental characteristics
[Eelsky & Rovine, 1987; Calkins & Fox, 1992; Mangelsdorf, Gunnar, Kestenbaum,
L'ang, & Andreas, 1990; Seifer, 2000; Sroufe, 1985; vandenBoom, 1994). For example,
Kchbart and colleagues (Rothbart et al., 2006) have demonstrated that infants as young

a.
§ <1 months of age are capable of both deliberate and automated forms of self-regulation

22

.L

t
lip

'I»,

All.“

93

E11

through behaviors such as attentional shifting, oral stimulation and respiratory changes.
They argued that, “mechanisms used early [in infancy] to cope with negative emotion

may later be transferred to the control of cognition and behavior. (p. 341). These

individual self-regulation behaviors have been identiﬁed in early infancy and have been

demonstrated to be related to measurements of child temperament (Rothbart, Ellis,

Rueda, & Posner, 2003).

As the child matures, cognitive processes (e.g., attentional regulation, executive
functioning) become increasingly sophisticated and lead to a more fully consolidated
system of self control abilities that can be used in the service of maintaining an emotional
set goal (e.g., felt security, excited exploration of a toy). As is described in the next
chapter, this process does not necessarily require conscious awareness. That is,
internalized representations are hypothesized to inﬂuence behavior even though they are

not conscious (N. L. Collins & Read, 1994). As described by Kopp (1989), this proposed
evolution of the child’s growing and changing capacity for emotional self-regulation is
c(>lasistent with theories of development that have described the existence of latent

Constructs (e.g., regulatory abilities) as manifesting differently at different developmental

St<'=5I.ges.
E&r1otional Self-Regulation: A Summag and F ormulation

In sum, there are disagreements both across and within disciplines about the

 

s . . . . . .
iJ‘chture and function of human emotions and the definitions of emotion regulation and

emotional self-regulation. Consistent with Eisenberg and Spinrad’s (2004) deﬁnition,

h"‘E‘—vl:-e appears to be an important distinction between emotion regulation and emotional

23

self-regulation. A synthesis of the extant literature suggests the following requisite

abilities and deﬁnition of the process of emotional self-regulation:

First, it is clear that emotions can be elicited either by external (e.g.,
environmental) or internal (e. g., cognitive) stimuli and that the stimuli may manifest
either consciously or unconsciously to elicit an emotional reaction (Gross & Thompson,
2007). Thus, the capacity to experience emotions requires that the individual posses
either a functioning sensory system or a functioning cognitive system.

Secondly, the individual must have the ability to detect emotion-salient stimuli,
although the detection process need not be conscious. For example, alexithymic
individuals appear to meet this criterion due to the fact that their physiologic responses
are generally congruent with external stimuli (e.g., crying at a funeral), even though their

abilities to understand, articulate and cognitively process the stimulus are impaired or
absent. Furthermore, clinical examples of detection deﬁciencies reinforce this point. For
example, individuals who do not detect danger when it is present (e.g., some individuals
With autism), are exhibiting an underdeveloped ability to detect emotion-salient stimuli.
In contrast, individuals who demonstrate a hypersensitive reaction to environmental
StiI‘riuli (e.g., individuals suffering from certain phobias and some forms of obsessive
c OmpuISive disorders) react with fear to innocuous stimuli. In each of these cases, the
“appraisal-action readiness stance” described by Cole and colleagues (Cole et al., 2004),

3 l7|??lisattuned resulting in elther the under- or over-appralsal of emotion-salient

it: f .
Qrmation.

A third point involves responses to emotional stimuli and returns to the two-

p P’Qtjged approach described by both LeDoux ( 1996; 2000) and Eisenberg (2002).

24

 

Speciﬁcally, individuals are thought to react in one of two ways to the emotion-salient
stimuli they detect. The ﬁrst involves a physiological reaction that initially occurs outside
of conscious awareness and appears to be processed by the amygdala (e.g., seeing a snake
results in physiologic arousal). The second involves a learned response reaction based on
the individual’s lived experiences (e.g., hearing a favorite piece of music and
experiencing positive feelings). This too, may be processed outside of conscious
awareness but is likely processed in more recently developed areas of the cortex such as
the frontal regions (Phan et al., 2005).

Subsequent to detection and response, the process of emotional self-regulation,
per se, involves the modiﬁcation of one’s emotional reactions in the service of one’s
goals (Eisenberg, 2002). This process can occur on a conscious or unconscious level and
can involve both active and automatic processes. For example, Rothbart and Bates’

( l 998) notion of effortful control: “the ability to inhibit a dominant response to perform a
subdominant response” (p. 137), described a voluntary, active regulation process. An
example of this process is a child’s capacity to hold a small candy on the tongue for a
period of time without eating it in an effort to win a bigger prize (Murray & Kochanska,
2 002). A further example of a conscious emotional regulation process involves cognitive
techniques that teach individuals to stop particular thoughts (e.g., self deprecating
l:l"loughts) that are leading to painful emotional responses such as depression or anxiety
CEeck, 1995). Each of these processes relies heavily on the person’s capacity for
a)‘C'ecutive functioning. In contrast, attachment theory, more fully described in the next
c: l‘Etpter, posits the existence of unconscious regulation strategies that develop over the

c
Q Use of early infancy based on repeated experience with a primary caretaker and

25

 

become automatic templates from which the person understands and reacts to their

interpersonal environment.
Finally, based on the developmental trajectory that is described in the next

chapter, the emotion regulation strategies that were developed in the context of early
attachment experiences become more fully consolidated during the preschool period. At
this point in development, due to the more mature cognitive and attentional capacities of
the preschool-aged child, these regulation strategies are sufficiently internalized and
integrated across the domains of cognition, emotion and behavior such that they are
active within a range of social and interpersonal contexts and can be labeled emotional
self-regulation strategies. That is, the ways in which the child has been learning to think,

feel and behave in the context of the attachment relationship become the template from

which he or she approaches future experiences and relationships.

Measurement of Emotional Self-Regulgion in the Current Study

The measurement of child behaviors and developmental constructs within
longitudinal investigations of child development must be adapted at each assessment
pOint in order to adequately assess the changing capacities of the developing child. In
1i glit of this, the current investigation measured the child’s regulatory processes during
i ITllt‘ancy in the context of the mother-child relationship (e. g., within the attachment
paradigm). Subsequently, the child’s emotional self-regulation abilities and strategies
were measured during the preschool period apart from the physical presence of the
mother. Using their emotional self-regulation systems, children were expected to react to
guessful situations differentially based on their early (attachment) experiences with

e o o I o o
tsr-Tlotron regulation as well as their exposure to traumatic events (e.g., DV) and their

26

 

exposure to proximal maternal parenting behaviors (e.g., sensitivity, discipline). It was
hypothesized that well-regulated children would evince mid-levels of regulatory
capacities. That is, in the context of a stressful situation (e. g., measured during a mother-
child separation sequence in an unfamiliar laboratory setting), they were not expected to
be markedly over- or under-regulated.

Finally, the inﬂuence of these emotion regulatory capacities and deﬁcits on the
development of internalizing and externalizing behavioral symptoms was also examined.
Speciﬁcally, it was expected that, in the context of other psychosocial stressors (DV,

SES, insensitive parenting), children who had developed emotion regulation strategies
that involved over-control of their emotional experiences (e.g., a “shut-down” emotional
system stemming from an earlier avoidant attachment with the primary caregiver) would

demonstrate higher levels of internalizing behavioral symptoms. In addition, children

who had developed emotion regulation strategies that involved under-control of their

emotional experiences (e.g., a “ramped up” emotional regulation and expression style
Stemming from an earlier ambivalent attachment with the primary caregiver) were

expected to demonstrate higher levels of externalizing behavioral symptoms.

27

CHAPTER 2
ATTACHMENT AND THE DEVELOPMENTAL ETIOLOGY OF EMOTION
REGULATION
Originally developed by John Bowlby (1969/ 1982), attachment theory, along with
the empirical investigations it has generated over the last four decades, has been a
primary contributor to current theoretical and clinical conceptualizations of emotion
regulation throughout the lifespan (N. L. Collins & Read, 1994). Attachment theory
posits that the capacity for emotional self-regulation develops in the context of early
relationships and is intemalized by the developing child through internal representations
(e.g., working models) such that the child’s regulatory system eventually becomes a
relatively stable personality characteristic of the individual.

Drawing on the psychoanalytic, general systems, cognitive and ethological
theories of his day, Bowlby emphasized the importance of the primary caretaker (usually
the mother) to the survival and healthy social-emotional development of the infant and

y o ung child (Bowlby, 1969/1982, 1980). The theory emphasized the fact that human
iInfants, relative even to other mammals, exist for an extended period of time in a state of
utter dependency wherein proximity to a caretaker is essential for physical survival and
psychological health (Simpson, 1999). A contemporary of Bowlby coming from the
O bject—relations tradition, Winnicott underscored this point when he proclaimed, “There
i S no such thing as a baby” (Winnicott, 1965). With this provocative declaration he
S Q‘I.1ght to emphasize the fact that a human infant will simply not survive without the care

hid ministrations of an older caretaker. From this perspective, then, the development of

28

 

the capacity for emotional self-regulation in humans is fundamentally a relational

process.
In this way, the association between early relational experiences and the later

ability to regulate one’s own emotions and behaviors was central to Bowlby’s attachment
theory. The theory was informed by his work with delinquent teenage boys who, today,
would be described as demonstrating extreme externalizing behaviors suggestive of a
DSM-IV (American Psychiatric Association, 1994) conduct disorder or antisocial
personality disorder diagnosis (Bowlby, 1944). Bowlby argued that disrupted early
relationship experiences could partly account for the delinquent (externalizing) behaviors
these boys were demonstrating. Developed in the context of his study of child
psychopathology as well as ethology and evolutionary theory, Bowlby’s attachment
theory held that human infants and young children must remain in close physical
proximity to an adult caretaker in order to physically survive and for psychological health
(e- g. security, emotion regulation) to develop. Working within this theoretical
perspective, Sroufe and Waters (1977) later introduced the term felt security,
emphasizing the need for the psychological availability of the caretaker to the infant and
yo 1mg child within an organized attachment system in this process. In other words,
beyond simple physical proximity, the caretaker must remain psychologically available
841d responsive to the infant’s needs in order for the infant to experience a sense of safety
amid security which will allow for exploration of the environment and normal
qQVelopmental growth to occur. From this perspective, the adult, through her presence
mid responsiveness to the infant in a relational context, provides the physical and

g3'chological structure the infant needs to experience a sense of physio-emotional

29

regulation and safety (Bowlby, 1969/1982; Fonagy, Gergely, Jurist, & Target, 2004;
Schore, 2003; Weinﬁeld, Sroufe, Egeland, & Carlson, 1999). In fact, in the earliest weeks
and months of life, the infant requires an adult caretaker to help regulate even the most
basic of physiological processes including body temperature, hunger/satiation states and,
as some research has suggested, even heart and respiratory rates (Small, 1998).
As the infant progresses developmentally, the primary caretaker becomes critical
in assisting the infant in the regulation of states and processes beyond basic survival
needs (Cassidy, 1999; F uendeling, 1998; Sroufe, 1995; Stern, 1985). While the set goal
of proximity to or access to the caretaker remains the same, infants develop and employ
diverse strategies based on their increasing repertoires of cognitive and behavioral
abilities to attain this goal. In this way securely attached infants and young children use
their caretakers as a means by which to regulate their own emotions and feelings of
safety. For example, upon hearing a loud, unfamiliar noise, a pre-crawling infant may cry
and wave her arms in the direction of the mother whereas a walking infant will likely
walk toward the mother in order to attain physical and psychological closeness and
protection from her. Both behaviors serve the same set goal of proximity attainment to
the mother, but each child uses the most efﬁcient strategy she has available to achieve
ﬁllet goal. Furthermore, in each case the infant is actively using the mother to alleviate the

ClB’Sregulating experience of feeling frightened. Physical closeness with the mother helps

to bring the child back into a state of emotional equilibrium.

As development unfolds, the securely-attached child intemalizes these regulation

S‘tt‘eregies and therefore becomes increasingly capable of the capacity for emotional self-

gtrlation even when the mother 18 not present. That Is, the 1n1t1al ability of the parent to

30

regulate the infant’s emotional states (e. g., soothing her when she cries, etc.) becomes
internalized as the capacity for emotional self-regulation in the older child. In fact, as the
child matures, the comforting presence of the caregiver allows securely attached children
to engage in and explore their environment (Bowlby, 1988). In the process of exploration,
the child is likely to encounter situations wherein she can, in effect, practice her self-
regulation skills (Burgess, Marshall, Rubin, & Fox, 2003). There will inevitably be times,
for example, when the child can not immediately connect with (e.g., run to, look at) the
mother. In these situations, the child will be required to make an attempt to employ her
own emotional self-regulatory abilities. If the stressor is of sufficient intensity that the
Child is Unable to self-regulate, increased attempts to return to the caretaker will likely
ensue. However, if the stressor is less intense and the child ﬁnds herself capable of self-
regulation, she will begin to gain conﬁdence in her own abilities. Beyond the
consideration of these developmental differences in attaining a set-goal, however,
attaChl'l'lent theory also explicates differential sub-types of attachment relationships which
are essentially emotion regulation strategies that have developed within the context of the
caregiVer-infant relationship. Speciﬁcally, as infants gain interpersonal experience with
their Primary caretaker, they learn that certain behaviors are effective in keeping the
mother close, whereas other behaviors tend to be ineffective. As described below, and
consistent with the work done by Block and Block (1980), the two insecure organized
strategies involve either a strategy of inhibiting (overcontrolling) or amplifying
( under Controlling) one’s expressions of emotional upset. Further, these attachment

strategies (categories) have been shown to occur cross-culturally (vanIJzendoom & Sagi,

’/ 999).

31

 

Attachment theory owes its status as an empirically-based theory to Mary
Ainsworth (Ainsworth, Blehar, Waters, & Wall, 1978). A contemporary of Bowlby,
Ainsworth essentially operationalized the study of attachment relationships with the
development of the strange situation protocol (SSP, Ainsworth et al., 1978). Based on
her naturalistic observations of mother-infant dyads in Uganda (Ainsworth, 1967), she
hypothesized that in the context of an increasingly stressful situation, the infant’s

attachment strategy would be activated to ensure that proximity to the caretaker, and
consequently, feelings of security, would be maintained. In her laboratory observations of
the SSP, Ainsworth documented three distinct and organized strategies exhibited by one-
year 01d, white, middle-class, American infants; one secure type and two anxious types:
avoidant and ambivalent. Of equal, if not greater importance to attachment theory and
empirical work in this area was the naturalistic observation of maternal behavior
Ainsworth conducted in the home environments of these families. While her sample size
linking laboratory and home behavior was small (n = 23), her thorough documentation of
maternal behavior in the home environment suggested the mechanism through which the
infant’s attachment strategy developed. What emerged from these data was a striking
relationship between infant attachment strategy as observed in the laboratory and
maternal caretaking behavior as manifest in the home environment (Ainsworth et al.,
1973)- These data demonstrated that parenting mattered, and it mattered because the
infant’s emotional regulation strategy with regard to the mother seemed to be based
difeCﬂY on the infant’s relationship history with the mother. By one year of age the infant
had developed an integrated emotional, cognitive and behavioral regulation strategy in

._/der to effectively make and maintain a relationship with the mother. More recently,
1

32

 

/

Mikulincer, Shaver and Pereg (2003), have described attachment as, “the systematic
pattern of relations and expectations, emotions and behavior that results from
internalization of a particular history of attachment experiences and consequent reliance
on a particular attachment-related strategy of affect regulation” (p. 79).

In her groundbreaking work, Ainsworth and colleagues (1978) used both
naturalistic and laboratory assessments of mothers and infants. Her naturalistic
assessments included home observations of mother-child behaviors that spanned multiple
hours of recorded observations. One limitation of these data is that Ainsworth was the
ﬁrst to conduct naturalistic observations of mother-infant interactions and thus, did not
use standardized observer-rated measures. Although several, more current observer-rated
measures of parenting sensitivity are based on her early work (e. g., Biringen, Robinson,
& Emde, 2000) the home-visit portion of her study was never fully replicated.

In her early studies, Ainsworth and colleagues (Ainsworth et al., 1978) found that
infants demonstrating a secure pattern of attachment exhibited their negative feelings
OPCHIY and sought comfort from their mother following a stressful episode in the

1abm'atory. Speciﬁcally, securely attached infants were distinguished from the two
organized, insecure categories in that they 1) engaged in some form of contact (e.g.,
visual, verbal, or physical) with their mother upon reuniting with them and 2) were
effectively comforted by that contact. In addition, once comforted, securely attached
infants returned to exploratory play. In this way they openly and genuinely displayed
their emotions and, by using an active process of seeking contact with their mothers, they
used the mother in their efforts to regulate these emotions. In the home environment the

10mm of these infants were judged to be sensitive and tender in their caretaking

33

interactions and were responsive to a wide variety of their infant’s emotional displays.
Their caretaking behaviors were contingently responsive and attuned to the expressed
needs and desires of their infants. More recently, in their meta-analysis examining the
association of maternal sensitivity and infant attachment security, DeWolff and van
ljzendoom (1997) found a moderate positive association between these constructs (n=66
studies; r = .24). While these data supported the hypothesized relationship between
maternal sensitivity and infant attachment, they also suggested that other important
factors itlﬂuence the child’s early development of attachment-related behaviors. Relative
to the manifestation of the secure attachment strategy, Mikulincer and colleagues
(Mikulincer et al., 2003) have argued that, “Relatively secure individuals have learned
that acknowledgment and display of distress elicit supportive responses from others. .
[and that] . . their own actions are often able to reduce distress” (p. 83).
In contrast to the securely attached group, however, infants categorized as
anxious—avoidant in Ainsworth and colleagues’ original study (1978), demonstrated a
striking pattern of affective deactivation in the laboratory wherein they appeared not to
need the comfort of their mothers at all. They played independently and often seemed
thel'Vious to their mother’s presence or absence. At home, the mothers of these infants
tended to demonstrate a rejecting behavioral pattern, especially when their infants
expressed negative affect. For example, their responses to their infant’s crying included
behaViors such as leaving the infant alone or expressing anger toward the infant.
Additionally, they seemed to dislike and avoid physical contact with their infants. In the
case of these dyads, the infants’ behavioral strategies in the laboratory situation have

1 or e recelTitly been described as an effort to maintain proximity to the mother by

34

deactivating their emotional responses due to their mother’s inability to tolerate them
(Magai, 1999). Thus, through their interactions with their mothers, these infants appeared
to have learned that by inhibiting their outward emotional expressions, they were more
likely to reach the set-goal of staying within close physical proximity to their mothers.
Their emotional self-regulation strategy appeared to be organized around their experience
of interacting with a behaviorally rejecting primary caretaker.

Alternatively, anxious-ambivalent infants in Ainsworth and colleagues’ (1978)
original study demonstrated a hyper-activating strategy in the laboratory setting wherein
they desperately attempted to have contact with their mothers but were unable to be
soothed by them once contact was achieved. At home the mothers of these infants tended
to be inconsistent in providing sensitive caregiving and their interactions with their
infants were not contingently based on the infants’ cues. These infants have more
recently been described as employing a hyper-vigilant strategy wherein they focus their
attention exclusively on the mother at the expense of the ability to explore and enjoy the
enVirol'lzlnent (Magai, 1999). Having been inconsistently responded to by their mothers,

infants employing an ambivalent strategy appear unable to use their mother in an
OtgtmiZing or regulating manner but nevertheless are always attempting to elicit the warm
response which is provided by the mother on an inconsistent basis (Main, 1995). This
formulation is consistent with the known behavioral effects of a variable reinforcement
schedule (F erster & Skinner, 1957). That is, positive reinforcement of a behavior that is

offered inconsistently tends to result in enduring behavioral patterns that are very difﬁcult

to extingui sh. In this case, infants who have been positively reinforced for crying

35

behaviors on a variable schedule are likely to both intensify and continue these behaviors
in an effort to achieve subsequent reinforcements.
A fourth attachment classiﬁcation was later articulated by Main and Solomon

(1986), and labeled disorganized. At the time of their investigations, the strange situation
protocol had been empirically established as a valid and reliable measure. These authors
initially examined attachment in groups of mothers with established trauma histories and
found that the infants of these mothers often did not demonstrate an organized strategy at
all, but instead displayed multiple bizarre and uncoordinated behaviors in response to a
stressful situation (e. g. behavioral freezing, uncoordinated and conﬂicted attempts to gain
PIOXimity to the mother). Mothers of these infants have been found to display frightening
behaViOrs such as producing unusual and bizarre vocalizations and displaying sudden
inthiVe and threatening physical movements into their child’s personal space (Solomon
& Geol‘ge, 1999). In addition, mothers of disorganized infants have also been found to
display behaviors in which they seem to be themselves frightened by their infant’s
bwinders such as suddenly moving away from the infant in a fearful manner (Lyons-

Ruth & Jacobvitz, 1999).

Throughout early development, these repeated experiences in the context of the
primary attachment relationship(s) are thought to become internalized as internal working
”1061813 within the child such that, over time, they become more stable characteristics of
the Child and can be observed outside of the primary attachment relationship(s) (Main et

31': 1985). A child’s working model of relationships incorporates both cognitive and
affective Components (Crittenden, 1990) and has historically been conceptualized as an

/ﬂtemali2ed template of behavioral and affective expectations of self and other within

36

relationships. Over the course of early development in an ongoing and dynamic process,
the child begins to construct an internal mental model of what to expect in the context of
relationships. Once developed, internal working models are powerful psychological
constructs in that they not only reﬂect lived experience but also serve as a template or
schema from which to interpret the behaviors of others and to guide their own behavior.
Children with avoidant attachment histories come to expect rejection within the
context of relationships and are likely to withdraw emotionally from social experiences
and to play independently. Rather than seeking contact with other individuals (either
children or adults), these children are likely to focus on engagement with their physical
enVifonInent (e. g., exploring new toys, playing by themselves). Thus, they are thought to
have adopted an internalizing, or overcontrolled emotional self-regulatory style. In
addition, since their experiences of difﬁcult emotions cannot be outwardly expressed,
they 31‘ e hypothesized to be at risk for internalizing disorders. That is, their withdrawn
and deélctivating emotional regulation style places them at risk for internalizing
behm’ioral disorders.
In contrast, children with ambivalent attachment histories have the experience of
being attended to on an inconsistent basis. They have learned to behave in an over-
aroused, or undercontrolled manner in an attempt to garner the emotional warmth which
has been offered inconsistently. This strategy takes a toll on their ability to fully explore
their environment such that these children are likely to spend time clinging to adults and
seeking attention through the forceful outward expression of their thoughts and feelings.
Thus, the over-aroused and under-controlled emotional self-regulation strategies of these

fllﬂdren are thought to put them at risk for externalizing behavioral symptoms

37

(Guttrnann-Steinmetz & Crowell, 2006). Note that, by the preschool period of

development, a child who has developed an undercontrolled self regulatory style is likely
to self—regulate by seeking contact with others. That is, the preschool-aged child’s ‘other-
seeking’ behaviors are understood from the attachment perspective as reﬂecting their
intemalized emotional self-regulation strategy. This is fundamentally different from the
newborn who is dependent on the caretaker for physical comfort in order for the process
of emotional regulation (e.g., moving from a state of dysregulation to a state of
regulation) to occur. The preschool-aged child’s cognitive, attentional and behavioral
systems are sufficiently developed such that they are able to engage in emotional self-
regulation (conscious or unconscious). For some preschoolers this includes a pattern of
eliciting the ‘help’ of others (parents, teachers, etc.) in order to self regulate. Notably,
these behaviors in preschool-aged children (e. g., whining, crying, pleading) are often
resl’el'lded to by adults with anger and annoyance, thereby reinforcing the child’s
experience of never being fully satisﬁed within relational contexts.
In sum, as outlined here, the attachment strategy as measured in one-year-old
Children is understood as a manifestation of the child’s internalized regulation strategy
that has developed in the context of the infant’s relationship with the primary caretaker.
Attachment theory posits that, if this strategy does not undergo changes due to other
thifonmental inﬂuences over the course of early childhood (e.g., exposure to traumatic
events such as DV or changes in maternal levels of sensitivity), it will become an
increasingly stable personality characteristic of the child. In the present study attachment

was measured at 12 months of age using Ainsworth’s SSP and associations with

38

preschool emotional self-regulation capacities and behavioral externalizing and

internalizing symptoms were examined in a group of heterogeneous-for—risk children.

39

 

CHAPTER 3
EMOTIONAL SELF-REGULATION IN THE PRESCHOOL YEARS
As outlined in the previous chapter, emotion regulation during infancy is
conceptualized from the perspective of attachment theory as a dyadic process in which
the infant requires the physical presence of the caretaker(s) in order to achieve and
maintain a sense of physiological and emotional equilibrium. Over time, these early
relationships are internalized such that, beginning in the preschool period, the child’s
regulatory capacities become consolidated internal characteristics and, at this point in
development, can be accurately described as emotional self-regulation strategies.
Therefore, during the preschool period, the child’s emotional self-regulation abilities can
be measured outside of the dyadic relationship. Furthermore, one central thesis of this
paper is that deﬁcits in self-regulation capacities place children at risk for the
development of externalizing and internalizing behavioral symptoms. The following
review summarizes ﬁndings from the few empirical examinations of the relation between
attachment, child regulation capacities and the behavioral symptoms of internalizing and
externalizing disorders.

Went Ed EmotionﬁSelf-Regulation during the Preschool Period

While the theoretical links between attachment and subsequent emotional self-
r egulation capacity are clear and have been well-described, empirical investigations into
tl'lis relationship using direct observations of child behaviors have been surprisingly rare.
A few studies have examined associations among these constructs in the infancy and
t()Cldler periods (e.g., Kochanska, 2001). However, from an attachment perspective,

t-egl:llation capacity during these developmental stages is still fundamentally a dyadic

40

construct. The preschool period, in contrast, heralds an important developmental shift
With respect to the child’s capacity for emotional self-regulation (Kopp, 1989; Sroufe et
81., 2005). It is during this time that the child becomes more capable of regulation outside
01' the context of the dyadic relationship. In addition, it is during this stage of
development in Western culture when an increasing degree of societal pressure for self
regulation and control begins to be placed upon the child (Sroufe et al., 2005). Sroufe and
colleagues (2005) have argued that as a result of these developmental and societal shifts,
the measurement of child regulation capacities during the preschool period can be
considered evidence for the internalization of the child’s relationship history. They stated
tIIat, in the preschool period, “a new level of organization in the child apart from
care givers is apparent . . . [evinced by] the greater stability and stronger predictive power
0f individual variations from this time forward” (p. 121). They further stated that, “One
reason behavior is more coherent at this time is that the child has more cognitively
elaborated representations of self and others for guidance” (p. 121). Whereas the infant or
tOdd ler almost always requires that the caretaker be physically present in order to use
them as a regulatory mechanism, the preschool-aged child has, to a much greater degree,
interIlalized a sense of the caretaker such that the child is more often able to utilize self-
regl—llatory strategies during times of stress and is no longer solely dependent on close
physical proximity to the primary caretaker.
Consistent with this line of reasoning, the Minnesota longitudinal study (Sroufe et
al- ’ 2 005), shed considerable light on the emergence of self-regulatory behaviors during
the preschool period. In their recent compilation (Sroufe et al., 2005), these investigators

tQVieWed three decades of empirical results from their study of attachment and

41

 

Psychosocial risk in a high risk group of children and their families. Most relevant to the
current study are their assessments of the relationship between infant attachment and
Preschool development. Speciﬁcally, when the children were 31/2 years of age their ability
to persevere in an extremely challenging task was measured. In this commonly used and
standardized task, children were required to open a box in order to have access to several
attractive toys to play with. The box was nearly impossible to open. Child behaviors
during this task were coded using scales including agency, ﬂexibility, withdrawal,
creativity and ego control. Using cluster analysis methodology, children scoring similarly
on these behavioral codes were created. Findings demonstrated that, of the children with
secure infant attachment histories, 40% of them fell in the highest competent cluster
grouP. In addition, none of the children with avoidant or resistant attachment histories fell
in this group. Thus, the ability of these children to regulate their behaviors and emotions
in the service of an extremely challenging goal was linked in theoretically consistent
Ways to their prior attachment classiﬁcation.
Subsequently, when children reached the age of four and one-half to ﬁve years
their ego-resiliency capacity as deﬁned by Block and Block (1980) was assessed. Recall
that this construct is similar to the current deﬁnition of emotional self-regulation as
deﬁned by Eisenberg and Spinrad (2004), and is consistent with the over- and under-
reglllated emotional strategies (avoidant, ambivalent) posited by attachment theory.
Using composite variables based on behavioral responses to laboratory tasks as well as
Q‘sort data from preschool teachers, findings indicated that children who were securely
attached at one year of age had signiﬁcantly higher ego resiliency scores during the

QtesCl’lool period. That is, they were much more able to adapt their behaviors in a ﬂexible

42

manner in the face of varying situational demands. While these data were illustrative in
suggesting the relationship between early attachment and later self-regulation during the
preschool years, the data were primarily reported at the dichotomous insecure/ secure
level of analysis and, consequently, a closer examination of the differences between the
two insecure groups was not fully explicated (Sroufe et al., 2005).

In addition, despite their theoretical consistency, some of these data were never
published in a peer-reviewed journal. Instead, they were described in the recent
compilation from the Minnesota longitudinal study (Sroufe et al., 2005). In addition,
important information about these data were not reported such as the number of clusters
created from the box task and the total numbers of children within the different
attachment categories. Thus, there is no way to fully evaluate the methodological
strengths and weaknesses of these particular analyses.

Although other longitudinal projects have examined the continuity and change of
attachment systems throughout the lifespan (Grossmann, Grossmann, & Waters, 2005),
very few have examined the inﬂuence of early attachment on later observer-coded child
regulation capacities. In one notable exception Gilliom and colleagues (Gilliom, Shaw,
Beck, Schonberg, & Lukon, 2002) employed a longitudinal examination of the
relationship between infant attachment classiﬁcation and the expression and regulation of
the speciﬁc emotion of anger. In addition, a second study examined the link between
infant attachment and later emotional functioning in the context of peer interactions with
an unknown peer among preschool children (Booth, Rose-Krasnor, & Rubin, 1991).
Finally, two additional studies examined the constructs of attachment and emotion

regulation in middle childhood. Similar to the methodology typically used in the

43

adolescent and adult attachment literature, the ﬁrst study examined attachment status
concurrently with emotion regulation (Contreras, Kems, Weimer, Gentzler, & Tomich,
2000). It is therefore less relevant to the current investigation due to the lack of a
longitudinal analysis. In contrast, the second study did utilize a longitudinal design to
examine the relationship between infant attachment and later psychosomatic problems
and found that child negative emotionality mediated this relationship (Hagekull &
Bohlin, 2004). However, negative emotionality was measured through parent-report and
deﬁned as a temperament construct, making this study, too, less relevant to the current
investigation.

In a study that directly informs the current work, Gilliom, and colleagues (Gilliom
et al., 2002), examined infant attachment and later child expression and regulation of
anger using a longitudinal design. As part of a broader study on antisocial development in
lower SES boys, these investigators examined the inﬂuence of infant attachment security
as measured in the SSP (Ainsworth et al., 1978) on anger control abilities in preschool-
aged (3.5 years), disadvantaged boys (n=189). Regulatory strategies and expressed
negative emotions were measured during a waiting task wherein the boys were required
to wait for an undisclosed period of time to eat a cookie which their mother held while
she ﬁlled out paperwork. Regulation strategies were coded in lO-sec intervals and
included active self distraction, passive waiting, information gathering, physical comfort
seeking and focus on the object or on the waiting task itself (e. g., commenting to their
mother that this was taking a long time). Results from regression analyses demonstrated
that attachment security (coded dichotomously as insecure or secure), was related to type

of regulation strategy used such that securely attached boys used the strategies of self-

44

distraction, waiting quietly and information gathering more than insecurely attached
boys. In addition, the use of these strategies was associated with subsequent decreases in
expressed anger. The authors argued that these particular strategies reﬂect adaptive
regulation because they involve a higher level of cognitive, attentional and behavioral
sophistication that only becomes available during the preschool period. In comparison,
the strategies of comfort seeking and focusing on the aversive task do not involve higher
order regulatory skills. Instead, they keep the child preoccupied with the aversive
experience or reﬂect an earlier stage of development wherein close physical proximity to
the mother was necessary for effective regulation to occur. Thus, in this study, children
who had been securely attached in infancy demonstrated an increased ability to use age-
appropriate regulation strategies compared to children with histories of insecure
attachment. Further, these strategies were effective; they were associated with decreases
in expressed negative affect.

While this study is an important contribution to the literature documenting the
inﬂuence of attachment security on anger regulation, it does have some signiﬁcant
limitations. Most importantly, the sample was comprised only of boys. An understanding
of emotional self-regulation processes in boys is extremely important given, as the
authors stated, the fact that boys are much more likely than girls to develop externalizing
social-emotional disturbances such as conduct disorder during childhood. However, this
particular study leaves an understanding of these processes in girls unexarnined. Also, the
laboratory paradigm included the presence of the mother in the room with the child.
Although she was engaged in another task and, therefore, not necessarily engaging with

the child, this design allowed the possibility for the child to use maternal contact as a

45

regulatory mechanism and, therefore, did not inform us as to what that child would have
done instead, had the (presumed) primary attachment ﬁgure not been physically present.
In addition, by examining the attachment construct at the dichotomous level of security
versus insecurity, the analyses used in this study did not explicate the theoretical
relationship between the two insecure attachment subtypes and the deactivated (over-
controlled) versus the hyper-activated (under-controlled) emotional self-regulation
classiﬁcations.

An area of further study related to the current work involves investigations of
emotion expression and regulation in the context of peer interactions in school-aged
children. This area has received signiﬁcant attention in the literature (Contreras et al.,
2000; Kems, Klepac, & Cole, 1996; Wood, Emmerson, & Cowan, 2004). Often,
however, attachment is measured concurrently with the dependent variable in these
studies. In one notable exception using a preschool-aged sample, researchers used a
longitudinal design to examine the association of attachment history and emotion
regulation. In this study, Booth, Rose-Krasnor and Rubin (1991) examined the relation
between secure versus insecure attachment as measured in the SSP when the children
were 20 months of age (Ainsworth et al., 1978) and later social-emotional functioning
when the children were preschool aged (4 years) within the context of a structured,
laboratory, peer interaction protocol (n=62). The sample was comprised of a combination
of children from two distinct longitudinal studies and was approximately evenly split
between a high risk group (n=32) and a low-risk group (n=30). Social ﬁmctioning with an
unfamiliar, securely attached, same-gender peer was assessed using a structured play

activity (e. g. the children were asked to build a house together) and via a novel toy

46

sharing procedure (e. g., the children were required to share an attractive toy). Child
behaviors were coded from videotape and included expressed affect (positive, negative or
neutral) as well as child goal attainment strategies such as the use of physical or verbal
aggression and asking questions. ANOVA methodology revealed a main effect of
security status on negative affect and aggressivity such that children who had been rated
as insecurely attached with their mothers as infants were more likely to express negative
affect during interchanges with peers and demonstrated higher levels of aggressivity in
social interactions with peers. SES risk status did not moderate this relationship. Thus,
while the dichotomously scored attachment variable did not allow for an explication of
how the different insecure patterns may have affected these ﬁndings, the ﬁnding that
insecurity in infancy was related to later negative affectivity was theoretically consistent.
In addition, these analyses were confounded by the (empirically) unpredictable behaviors
of the other child(ren) and the fact that child gender was not controlled. However, despite
these constraints, the ﬁndings shed light on regulatory child outcomes, especially given
the inclusion within the research design of the potentially stressful situation of interaction
with an unknown peer.

In sum, although it is surprising given the rich theoretical literature in this area,
there remains a paucity of empirical investigations examining the relationship between
infant attachment strategy and later observer-rated emotional self-regulation during the
preschool period. The studies that have been conducted have not explicated the
relationship between attachment typology and regulatory outcomes. Instead, data from
these studies are examined at the secure/insecure level of analysis. These data suggest

that differences in preschool regulation capacity are related in theoretically consistent

47

ways to infant attachment security. However, examinations of these constructs at the
typological level of analysis are needed to further clarify these relationships. In addition
the investigations conducted to date have examined a relatively narrow range of
emotional expressions and emotional self-regulation capacities (e.g., anger regulation).
The current study extends these ﬁndings by examining child emotional self-regulation
capacities that tap multiple elements of this construct such as a range of emotional
expressions and the attentional as well as behavioral strategies children use in the service
of emotional self-regulation. In addition, the present study examined these behaviors
using a coding scheme which included positive and negative emotional expressions as
well as behavioral and attentional regulation abilities.
Early Attachmentﬁand Later PssLhopathologv: Extemalizing and Intemalizing Behavioral
Smptoms

As described herein, there are strong theoretical arguments for the proposed
relationship between the infant’s regulation style within the attachment relationship (e. g.,
attachment classiﬁcation), and his or her later emotional self-regulation style. These
differences, especially among the organized (e. g., A/B/C) categories, are likely to
represent differences along a relatively normative spectrum. However, for some infants,
extreme reliance on internal methods of regulation (e. g., avoidant, internalizing,
overcontrolled) or external methods of regulation (e.g., ambivalent, externalizing,
undercontrolled), especially in the context of other psychosocial stressors, may be
sufficient to lead to later psychopathology in the form of internalizing or externalizing
behavioral disorders. Due to the developmental and social shifts that occur during the

preschool period, this may be an ideal time to begin to understand how early attachment

48

experiences may develop in terms of possible pathological outcomes in children.
Speciﬁcally, the preschool period is unique not only in that children have increased
cognitive and self-regulatory capacities, but also because it is the beginning point in
development wherein children are expected to engage more fully in their social
environments. For example, many children begin preschool at this age and, within
cultural and religious groups, formal training (e.g., Sunday School) often begins.
Consequently, social expectations for behavioral control increase dramatically during this
period. Sroufe and colleagues (Sroufe et al., 2005) have deﬁned the preschool period as
emergent at approximately 3% years of age and fully consolidated by age 4% years. At
this point in development the child is both capable of and expected to engage successfully
in self-regulatory behaviors.

Very few studies have examined the hypothesized relationship between infant
attachment strategy and later child internalizing and externalizing behaviors in young
children. Results from the few studies that have explored these relationships have
primarily found differences only when the data were dichotomized into secure/insecure
or organized/disorganized group (Shaw, Keenan, Vondra, Delliquadri, & Giovannelli,
1997; Shaw, Owens, Vondra, Keenan, & Winslow, 1996). These ﬁndings appear to be
based, in part, on the fact that low sample sizes did not provide the statistical power
necessary for an examination at the typological level of analysis.

For example, Shaw and colleagues (Shaw et al., 1997; Shaw et al., 1996) reported
on two separate sets of analyses examining the effect of infant attachment category on
later child externalizing (study 1) and internalizing (study 2) behaviors. Their

longitudinal sample (n=100) was comprised of low income families who had been

49

recruited at a social services office. Infant attachment was measured at 12 months with
the SSP using a four-category coding system (A, B, C, D) and child internalizing and
externalizing behaviors were examined using maternally reported CBCL measures for
ages 4-16 (Achenbach, 1991) at ﬁve years of age. In the ﬁrst study (Shaw et al., 1996),
they examined the effects of early attachment on the broadband externalizing scale of the
CBCL as well as the narrowband aggression scale of this measure (n=82 for the analyses
in this study). Initial correlation analyses indicated that when attachment was
dichotomized into disorganized (D) versus organized (A, B, C) groups it was
signiﬁcantly, positively correlated with both the externalizing (r=.24) and aggression (r
=.34) scales such that children with disorganized (D) attachment classiﬁcations scored
higher on these constructs. In addition, when the attachment variable was dichotomized
into secure (B) and insecure (A, C, D) categories, insecurity was signiﬁcantly, positively
correlated with the aggression scale (r=.21). Child gender was not related to the outcome
variables. In chi-square analyses the 4-category attachment variable was signiﬁcant
indicating that 60% of children with disorganized attachment histories received CBCL
scores in the clinically elevated range on the aggression scale (clinical cut-off of t Z 63).
However, no signiﬁcant effects were found in the chi-square analyses for externalizing
behaviors. Similarly, in subsequent regression analyses, disorganized attachment was
signiﬁcant in predicting aggression scores but not externalizing scores.

The second study (Shaw et al., 1997) examined infant attachment and later
internalizing behaviors (n=86). The broadband internalizing scale, as well as the
narrowband withdrawal and depression/anxiety scales was used to assess these domains.

Initial correlation analyses indicated that when attachment was dichotomized into

50

disorganized (D) versus organized (A, B, C) groups it was signiﬁcantly, positively
correlated with the withdrawal (r=.28) and internalizing (r =.21) scores such that children
with D classiﬁcations scored higher on these constructs. However, when the attachment
variable was dichotomized into secure (B) and insecure (A, C, D) categories, it was not
signiﬁcantly correlated with the outcome variables. In addition, child gender was not
signiﬁcantly correlated with any of the outcome variables. Regression analyses revealed
that the disorganized (D) attachment category had a signiﬁcant positive effect on the
broadband internalizing scale, as well as the withdrawal scale. In addition, a ﬁnding
relevant to the current analysis was that child exposure to parental conﬂict was also a
signiﬁcant predictor of the internalizing and withdrawal scale scores. No effects of these
constructs were found for the depression/anxiety scale.

Overall, these results suggested that, in this high risk sample, a history of
disorganized attachment was associated with both internalizing and externalizing
behaviors in young children. The effect of the disorganized category on children’s
behaviors was most frequently found to be signiﬁcant. However, there was some
evidence that the organized insecure classiﬁcations also had an effect evidenced by the
signiﬁcant ﬁndings when attachment was dichotomized into secure versus insecure
categories. In addition, exposure to parental conﬂict also accounted for unique variance
in the prediction of internalizing behaviors suggesting that early exposure to conﬂict
between the primary attachment ﬁgure(s) inﬂuenced the child’s later development of
symptoms of psychopathology. While these studies offer a signiﬁcant contribution to the
longitudinal examination of these constructs in a high risk group of children, they were

nevertheless limited by sample size. In addition, possibly due to the small sample size,

51

the attachment construct was dichotomized in these analyses such that an understanding
of the inﬂuence of particular attachment classiﬁcations was not explicated.

In another, recent, longitudinal examination, Vondra and colleagues (Vondra,
Shaw, Swearingen, Cohen, & Owens, 2001) also examined infant attachment history in
relation to child externalizing and internalizing behaviors in the preschool period (3%
years). This sample was considerably larger and included 223 mother-child dyads.
Participants were drawn from a low-income, urban population, and were recruited from a
local social services agency. Attachment classiﬁcations were tested and coded in the SSP
at three time points: 12, 18 and 24 months. At the ﬁrst two time periods, attachment
category was rated using an A/B/C/D categorical system. At 24 months Crittenden’s
Preschool Assessment of Attachment (PAA, Crittenden, 1994) was used to measure child
attachment. This system yields a secure (B) category as well as two organized insecure
strategies (A, defensive and C, coercive). In addition, three atypical categories are
included in this system: Defended/Coercive (AC), Anxious Depressed (AD), and
Disorganized (D). Composite scores were calculated for attachment category at all time
periods which yielded a value for the frequency of each attachment category (A, B, C and
D/atypical) for each child (e.g., a child who was securely attached at each of the three
time points would have a score of three for the B category and a score of zero in each of
the other categories). Correlations with outcome variables yielded signiﬁcance for
externalizing behaviors for both the securely attached (B; r = -.29) and disorganized
(D/atypical; r=.21) children. In subsequent regression analyses both externalizing and
internalizing scores were regressed on the 24-month attachment score and on a combined

12/18 month attachment score (computed as described above) in separate analyses. When

52

both scores were entered into the regression, only the 24-month attachment score was
found to signiﬁcantly predict externalizing and internalizing behaviors and accounted for
11 and 8 percent of the variance, respectively. That is, 12/18 month attachment scores did
not predict to the outcome variables above and beyond the 24 month scores. At 24
months, scores on the externalizing scale were predicted by scores for each insecure
classiﬁcation (A, C, Atypical). However, scores for internalizing problems were
predicted only by the atypical classiﬁcation. Thus, while the secure and disorganized
categories provided the most robust predictions to later child psychopathology in this
sample, there was also evidence that the organized insecure categories (A, C, Atypical)
were predictive of externalizing behaviors. The larger sample size used in this study,
compared to most other studies that have examined these constructs, may account for the
signiﬁcant ﬁndings in their analyses using all of the typological categories.

In a further study that informs the current work, Lyons-Ruth and colleagues
(Lyons-Ruth, Easterbrooks, & Cibelli, 1997) examined the relationship between infant
attachment as measured by the SSP and later internalizing and externalizing behaviors at
age seven years as measured by mother- and teacher-rated CBCL/T RF in a sample of low
SES children and their families (n=50). The sample was a subset of an intervention study
(n=76) within which the attrition rate was relatively high (n=26). In addition, only three
children were classiﬁed during the SSP as having an ambivalent attachment and, due to
the low cell size of this group, these children were not included in the analyses. In
regression analyses avoidant attachment predicted to teacher-rated, but not to mother-
rated, CBCL/T RF internalizing scores. That is, children who were avoidantly attached as

infants received increased teacher-rated internalizing scores on the TRF. However, prior

53

attachment did not predict to clinical cut-off levels of internalizing behaviors. In addition,
infant attachment category was associated with clinical levels of teacher-rated but not
mother-rated CBCL externalizing scores. The fact that the attrition rate in the original
longitudinal sample was so high and the lack of a group of ambivalently attached children
in the analyses suggest that this study was, overall, somewhat methodologically weak.

Overall, there have been very few examinations of the longitudinal effects of
attachment classiﬁcations as measured during infancy and later child externalizing and
internalizing behavioral functioning during the preschool and early childhood periods in
high risk groups of children. The studies reviewed here have markedly mixed results.
Disorganized attachment was most consistently related to child behavioral outcomes.
However, there was evidence that the organized insecure strategies also held predictive
power. In fact, in the one study with a larger sample size, infant attachment as rated
categorically (versus dichotomously) was signiﬁcantly related to later child symptoms of
psychopathology (V ondra et al., 2001). This suggests that with increased statistical
power, categorical analyses are possible and potentially fruitful. In addition, these studies
consistently deﬁned high-risk status as low SES. The question as to whether these
relationships would continue to be signiﬁcant in groups experiencing other risk factors
such as family violence was leﬁ unanswered.

The present study examined the inﬂuence of early attachment on later
externalizing and internalizing behaviors in the child’s daily environment as measured by
maternal report on the CBCL (Achenbach, 1991). The current sample was comprised of a
heterogeneous-for-risk group of women and children, many of whom were relatively low

in reported SES. In addition, approximately half of the women endorsed DV exposure

54

when they were initially recruited into the longitudinal study (during their pregnancy with
the study child), and approximately 70% of the women reported DV exposure at some
point across the four years of data collection. Consistent with a developmental
psychopathology framework, the present analyses examined the inﬂuence of early
attachment on later child regulation strategies in the context of other risk factors in the
child’s life.
The Mediating Role of Child Regulation in the Relation of Infant Attachment and Child
Extemalizing and Intemalizing Smptoms

Rubin and Burgess (2002) have stated, “The inability to regulate one’s emotions
and, relatedly, to control one’s behavioral impulses places the child ‘at risk’ for
psychological dysfunction” (p. 388). However, while a child may indeed be at risk for
dysfunction as a result of insecure attachment (regulation) strategies, this does not imply
that pathological behavioral responses will necessarily develop. In other words, while
differences in emotion regulation strategies in the laboratory environment may be related
in theoretically consistent ways to earlier attachment strategies, they do not necessarily
reﬂect pathological (dysfunctional) processes as they manifest in the child’s daily living
experiences. Children with organized insecure attachment strategies may be more or less
regulated in a short laboratory procedure but these behavioral observations cannot fully
reﬂect the child’s overall functioning on a daily basis in the context of naturally shifting
psychosocial stress levels. As Shaw and colleagues (Shaw et al., 1997) have described,
“the functionalist view of emotionality implicitly suggests that emotional experience is
deﬁned through transactions with the environment” (p. 1761). The current study

examined the hypothesis that, once developed in the context of the early attachment

55

relationship, and in combination with psychosocial risk factors (e.g., DV and concurrent
parenting), deﬁcits within the child’s emotional self-regulatory capacities would make it
more likely that behavioral symptoms of internalizing and externalizing disorders will
develop.

More than 20 years ago in an explication and empirical investigation of the
inﬂuence of attachment relationships on child development, Erickson, Sroufe, & Byron
(1985) stated that,

Disturbances of the attachment relationship are the main cause of
psychopathology characterized by chronic anxiety or distrust, placing children
doubly at risk. First, they render the child less able to cope with later adverse
experiences, and, second, they increase the likelihood that the child will behave in
such a way as to bring about more adverse experiences. (p. 148)
This description aptly highlights the transactional nature of child development wherein
children are both inﬂuenced by and have an influence on, people in their interpersonal
environments (Sameroff, 1993). However, in the time since these early writings by
Erickson and colleagues from the Minnesota study (Erickson et al., 1985), there has been
signiﬁcant change in the ways in which child development is understood (Greenberg,
1999). Currently, one of the main tenets of developmental psychopathology is that
pathological developmental outcomes such as excessive externalizing and internalizing
behaviors are the product of multiple factors including child factors (temperament,
physical health and illness), relationship history (attachment, loss), and environmental
factors (poverty, family violence, community violence). In contrast to their earlier

statement, Sroufe and colleagues (Sroufe et al., 2005) have more recently stated that,

56

“disturbance is created by the interplay of multiple factors operating over time, and links
between antecedent conditions and disturbance are probabilistic and nonlinear” (p. 239).
The implication here is that development is comprised of both continuities and lawful
discontinuities and is best understood and predicted by the interaction of risk and
protective factors (Belsky, Fish, & Isabella, 1991). Similarly, Keller and colleagues
(Keller, Spieker, & Gilchrist, 2005) have argued that the child’s ability to adapt to
changing environmental demands is determined by the combination of past attachment
history and present circumstances. In fact, existing data on the relationship between early
attachment history and later child psychopathology supports this hypothesis.

Speciﬁcally, while investigations of this relationship in high-risk samples of
children have found associations between early attachment security and later
psychopathology, investigations with low-risk groups of children have generally failed to
ﬁnd associations between these constructs (for a review see Greenberg, 1999). Thus, as
predicted by tenets of developmental psychopathology, it appears to be the interplay of
risk and protective factors that best predicts child outcomes. One risk factor in isolation,
as with an organized insecure attachment strategy in a low risk environment, is unlikely
to predict well to later psychopathological outcomes.

As described above, the current study examined the inﬂuence of infant attachment
on both child emotional self-regulation capacities and symptoms child externalizing and
internalizing behaviors. In addition, there is now growing evidence that a child’s emotion
regulation capacities may inﬂuence their propensity to develop symptoms of both
externalizing and internalizing disorders (Campbell-Sills & Barlow, 2007; Mullin &

Hinshaw, 2007). Thus, the current study investigated whether a child’s emotional self-

57

regulation capacities mediated the relationship between infant attachment and later
symptoms of behavioral disorders. Further, consistent with the tenets of developmental
psychopathology, the present mediating model was examined in the context of other
psychosocial risk and protective factors. Speciﬁcally, the inﬂuence of DV exposure and

concurrent parenting behaviors were examined, as described in detail in Chapter 4.

58

CHAPTER 4
INFLUENCING DEVELOPMENTAL PATHWAYS:
THE IMPACT OF DOMESTIC VIOLENCE AND PARENTING BEHAVIORS
As argued here, attachment is understood as the foundation for the child’s early
ability to regulate emotions. However, as outlined by the developmental psychopathology
framework, it is primarily the interplay of risk and protective factors within a child’s life
which best predicts social-emotional outcomes. Further, as these factors ebb and ﬂow
throughout the lives of children, their developmental pathways can be expected to shift.
In light of this, the current investigation examined the inﬂuence of a) DV exposure and b)
current maternal parenting behaviors, on the relationship between early infant attachment
category and later child regulation capacities and behavioral symptoms of internalizing
and externalizing disorders. It was expected that increases in DV exposure would be
negatively related to the child’s emotion regulation capacity and positively associated
with increased levels of both internalizing and externalizing behaviors. In addition,
positive parenting behaviors were expected to be positively associated with increased
child emotion regulation capacities and negatively associated with symptoms of
internalizing and externalizing disorders.
The Inﬂuence of Violence within the Home on Early Development
Over the last several years, the effects of DV exposure on later social-emotional
outcomes in children have increasingly been the focus of empirical investigations (Jaffee,
Moffrtt, Caspi, Taylor, & Arseneault, 2002). However, deﬁnitions as to what constitutes
DV vary within the literature. Early deﬁnitions focused exclusively on physical abuse

(Pynoos & Eth, 1986). More recently, the inﬂuence of verbal and psychological abuse

59

(ﬁx! .

Bani

has been integrated into conceptualizations of DV (Morevvitz, 2004). Fantuzzo and Mohr
(1999) have argued for a comprehensive deﬁnition that includes the following elements:
“A pattern of assaultive and coercive behaviors, including physical, sexual, and
psychological attacks, as well as economic coercion, that adults or adolescents use
against their intimate partners” (p. 22). Based on these prior deﬁnitions, DV is deﬁned
here as comprised of the following elements: 1) intentionality on the part of the
perpetrator (e. g., not accidental), 2) that involves an attempt to harm, control, or
manipulate the victim, 3) using psychological, verbal, or physical means that are, 4)
perpetrated within the context of a romantic relationship. A romantic relationship is
deﬁned as a relationship between two adolescent or adult individuals that includes sexual
feelings or behaviors by both partners and is, or has been, consensual in nature over the
course of the relationship.

Due to their exposure to this type of violence, children living in homes where DV
is perpetrated are at higher risk for the development of a wide variety of psychosocial
problems including both externalizing and internalizing behaviors (J affee et al., 2002).
Investigations of the inﬂuence of DV on the psychosocial development of older children
and adolescents have increased in recent years (J affee et al., 2002). Empirical
examinations of the inﬂuence of DV during early development, however, have been less
prevalent. In addition, studies which examine these constructs within longitudinal
analyses have been even less common.

Trauma theory, with its emphasis on the predictable cognitive, affective and
physiological human responses to traumatic events, is helpful in understanding the

ubiquitous inﬂuence trauma exposure has on the emotional regulation capacities of

60

individuals across the lifespan (Herman, 1992). In the case of children, however,
exposure to violence within the home can be further understood from a developmental
and, speciﬁcally, an attachment perspective (Davies & Cummings, 1994). Attachment
theory posits that, children, and especially young children, depend on their caretakers for
assistance in regulating their affective and physiological states. Consequently, exposure
to violence in the home among infants and young children has the dual impact of directly
threatening the child’s wellbeing while simultaneously threatening the young child’s
most important regulatory mechanism - the attachment ﬁgure. Due to these direct and
indirect effects on the regulation capacities of young children, DV exposure is likely to
have unique effects on a child’s ability to self-regulate.

As outlined by attachment theory, early child development, in particular, is an
especially vulnerable developmental stage that has important effects on subsequent
developmental outcomes. The Diagnostic Classiﬁcation of Mental Health and
Developmental Disorders of Infancy and Early Childhood (National-Center—for-Clinical-
Infant-Programs, 1994), has deﬁned a traumatic stressor as the young child’s “direct
experience, witnessing, or confrontation with an event or events that involve actual or
threatened death or serious injury to the child or others, or a threat to the psychological or
physical integrity of the child or others” (p. 19). In addition, research within the
biobehavioral literature is beginning to document the physiological impact that exposure
to early trauma has on young children (Mohr & Fantuzzo, 2000; Saltzman, Holden, &
Holahan, 2005). Thus, it is likely that exposure to violence involving one of the child’s
attachment ﬁgures in the home environment of young children will have unique and

powerful effects on their psychosocial and physiological functioning.

61

331'

.1
.l‘nl

it};

Beyond the effect of threats to the primary attachment ﬁgure, Davies and
Cummings (Davies & Cummings, 1994) have argued that threats to the marital dyad
itself are disruptive to child development. This is especially salient for children when the
conﬂict is intense and unresolved as in the case of DV. In what they have termed the
emotional security hypothesis (ESH), they asserted that children hold a vested interest in
the continued maintenance of the marital relationship and that exposure to unresolved
marital conﬂict results in an increase in their feelings of emotional insecurity.

F urthermore, they argued that repeated child exposure to marital discord results in the
development of behavioral and emotional responses as a way of both protecting
themselves and intervening in the situation to reduce the discord. Within this framework,
children are thought to be affected by marital conﬂict (including DV) in three primary
ways: 1) their ability for emotional self-regulation is negatively affected by the
overwhelming experience of threats to both the primary attachment ﬁgure and threats of
dissolution of the marital relationship, 2) these threats to their emotional security serve as
motivators to attempt to inﬂuence their parents’ behaviors, and 3) a reduction in
emotional security negatively inﬂuences both their cognitive appraisals and their
representational models of their parents and family. This model suggests that the
inﬂuence of DV exposure will have unique effects on a child’s regulation capacities due,
in part, to the inﬂuence it has on the child’s internal representations of their parents.

Examinations of the inﬂuence of traumatic events on behavioral responses in
early infancy have suggested that infants as young as three months of age demonstrate
behavioral responses to trauma exposure consistent with Davies and Cummings’ (1994)

ESH model (Gaensbauer, 1995; Scheeringa & Gaensbauer, 2000; Scheeringa, Zeanah,

62

Drell, & Larrieu, 1995). In addition, two recent studies that have used subsamples of the
longitudinal sample under investigation in this study, have found effects of DV exposure
on infant behaviors that are consistent with this model. Speciﬁcally, using AN OVA
methodology, DeJonghe and colleagues (DeJonghe, Bogat, Levendosky, Eye, &
Davidson, 2005) found that, compared to non-exposed infants, one-year-old infants who
had been exposed to DV demonstrated heightened levels of observer-rated facial distress
in response to an experimenter simulated telephone argument when the infants were
alone in an unfamiliar room with the experimenter (n=89). These results suggested that
DV exposure during the ﬁrst year of life may have served to heighten an infant’s
emotional, attentional and behavioral sensitivity to adult conﬂict. However, no effects of
infant postural distress were detected, suggesting that behavioral manifestations of DV
exposure in infancy may be difficult to measure, especially given the dyadic nature of
emotion regulation during this period. For example, because their mothers were not
present in the room, these infants did not have an attachment ﬁgure to physically orient to
during the argument. This may account for the lack of ﬁndings with regard to postural
responses. In addition, the sample represents a relatively small subset of the entire
longitudinal study. These authors reported that many infants did not complete this
procedure due to maternal resistance to the necessary separation. It is possible that the
mothers who did not agree to the separation were sensitively attuned to their infant’s
developmental inability for emotional self-regulation during this task and refused
participation on this basis. In other words, these mothers may have been appropriately
protecting their children and, consequently, the infants of these sensitive mothers are not

included in the analyses, thereby biasing the results. The current study examined the

63

ad

Vll.

 

regulatory capacities of preschool-aged children who are developmentally capable of
emotional self-regulation using the entire longitudinal sample (3 cases from the
longitudinal sample were excluded because they did not participate in any of the waves of
data collection examined here; FIML estimation was employed on the rest of the
longitudinal sample, n = 203).
In a second study drawing from the current sample (n=48), Bogat and colleagues
(Bogat, DeJonghe, Levendosky, Davidson, & vonEye, 2006) found that mothers’ self-
reports of their own trauma symptoms resulting from DV exposure predicted infant
trauma Symlatoms, but only for infants who had witnessed severe levels of violence.
Infants Who witnessed milder forms of DV were less affected by their mother’s resultant
trauma Syntptoms. These results suggested that, in the face of more extreme levels of
abuse, the mother’s own pathology inﬂuenced the infant’s ability to cope with the trauma
to a greater degree. The sample used in this study is also small relative to the larger
longitudinal sample and suggests that participant self-selection bias may have inﬂuenced
the results. Inclusion criteria for this study were based on maternal reports of infant DV
CXposure as deﬁned by the mother’s report that the infant had witnessed the DV. It is
possible that mothers may have minimized their child’s direct exposure to abuse. In fact,
it may be that mothers who were most sensitive to their child’s experiences were the most
likely to be aware of, and therefore report, that their children were directly exposed to the
DV in the home. Alternatively, it may be that only the mothers of children who were
exposed to relatively higher levels of DV reported their children as being exposed. In

either case, reporter bias may be a factor in these results.

64

 

é" "

 

#3 ,
“c1

Findings from both of these studies are relevant to the current investigation in that
it is possible that these early behavioral symptoms of infant trauma may be precursors to
later behavioral and psychosocial problems in older children such as externalizing and
internalizing behaviors which are typically not measured until the preschool and early

childhood years. In addition, the current study examined DV longitudinally over the
course of development; that is, across the four year period from birth to age four. Thus,
the fact that infant trauma symptoms have been noted in some of the children within this
sample has direct bearing on the current investigation.

Later in child development, the effects of child DV exposure have frequently been
measured Using the CBCL externalizing and internalizing subscales. In the few studies
that have examined these constructs in preschool samples, results are contradictory with
“351)CCt t0 the inﬂuence of DV exposure on child behavior problems. This may be due, in
part, to the diverse sample characteristics of these studies (e. g., clinical vs. community
samples) and to relatively low sample sizes in most of the studies. For example, in a
community sample of women experiencing DV over the course of the previous year and
their children (n=62), Levendosky and colleagues (Levendosky, Huth-Bocks, Semel, &
Shapiro, 2002) found that preschool children (aged three to ﬁve years) exhibited elevated
levels 0f externalizing behaviors as measured by the CBCL. They reported that 42% of
the sample had T-scores of at least 60 and that 29% of the sample had scores within the
clinical l'allge. However, children’s scores on the internalizing scale were not elevated.
Child Pal'tiCipants in this study were selected based on current DV exposure; the

inﬂuence Of past exposure was not assessed. Consequently, some children who may have

65

experienced high levels of DV in the past were not included in the analyses. The current
study examined DV exposure over a three year period of time.

A second study by the same researchers (Levendosky, Huth-Bocks, Shapiro, &
Semel, 2003) examined both maternally rated child behavior problems in four-year-old
children using the externalizing and internalizing subscales of the CBCL as well as
observer rated child behaviors in a semi-structured, mother-child play activity (n=103).
The direct observation of these behaviors was an important strength of this study and is
similar to the methodology proposed in the current study. Results demonstrated that
higher levels of DV were related to observer ratings of increased negative and decreased
positive child behaviors during the mother-child play activity. However, in contrast to
ﬁndings from their previous study, DV was not associated with increases in either
externalizing or internalizing behaviors.

Additionally, in a recent examination of these constructs using a clinical sample
of preschool children who had been exposed to DV (n=85), Lieberman and colleagues
(Lieberman, VanHorn, & Ozer, 2005), found that, while the mean CBCL total problem
score fell below the clinical cut-off (t 270), CBCL scores were signiﬁcantly positively
correlated with increasing levels of DV exposure (r = .29). This ﬁnding is consistent with
the ﬁndings by Bogat and colleagues (Bogat et al., 2006) which indicated that more
extreme levels of DV are predictive of psychological symptomatology. However,
Lieberman and colleagues found that when maternal life stress, mother-child relationship
quality and maternal PTSD symptoms were controlled in a regression analysis, child DV
exposure failed to predict to child behavior problems. Thus, while correlation analyses

suggested that there was a relationship between DV exposure and child behavioral

66

problems, this relationship was not signiﬁcant when other contextual and environmental
variables were controlled. In addition, although the authors described their sample as a
preschool sample, child participants actually ranged in age from 25 to 59 months.
Developmentally, the manifestations of psychosocial problems in 2-year-old children and
5-year—old children are likely to be quite different. Further, the CBCL for four to 18 year
old children was used in this study for all children, calling into question the validity of
the ratings for the two and three year old subjects. Consequently, it is unclear as to
whether examinations of these relationships within a more controlled age range, using
valid measures, would have yielded different results.

In sum, Davies and Cummings (1994) have argued that due to its inﬂuence on both
the attachment relationship and the child’s relationship to the parental dyad, marital
conﬂict has a unique inﬂuence on a child’s capacity for emotional regulation. In the case
of DV, children are exposed to conﬂict which is, by deﬁnition, not resolved in a
psychologically healthy manner. In addition, analyses with subsets of the current sample
have demonstrated that children show elevated behavioral signs as a result of DV
exposure when compared to children who have not been exposed to DV or have
experienced relatively lower levels of exposure. Further, Davies and Cummings (1994)
have argued that DV exposure in early childhood will signiﬁcantly inﬂuence a child’s
representation of the primary attachment ﬁgure (usually the mother) as well as the
representation of the marital dyad. Given that the internalized attachment strategy is
thought to guide a child’s emotional self-regulation capacities during the preschool

period, in the current study it was hypothesized that DV exposure would be negatively

67

associated with the emotional self-regulation capacities of young children and positively
associated with their behavioral symptoms of internalizing and externalizing behaviors.
The Inﬂuence of Proximal Parenting Beh_aviors on Child Outcomes

As outlined by attachment theory, the mother’s parenting behaviors are critical to
the young child’s social emotional development due to their effects on the attachment
strategy of the infant. Moreover, the importance of parenting to child development does
not end in infancy; parenting behaviors inﬂuence development across the lifespan
(Bomstein, 2002). In fact, in one recent study, Belsky and Fearon (2002) found that,
when examined in the context of attachment history and parenting behaviors, parenting
behaviors which were more proximal to child outcome measurements predicted better to
later child social emotional functioning than early attachment history. In the present
study, current positive parenting was expected to be positively associated with child
regulation capacities and negatively associated with behavioral symptoms of
externalizing and internalizing disorders at four years of age.

The notion that good or optimal parenting will have positive effects on children’s
development is widely accepted (Bomstein, 2002). As described in this review,
attachment theory argues that the caretaker’s parenting behaviors become internalized in
the child such that, over time, they become an increasingly stable feature of the child’s
personality. Other theoretical approaches, too, have described the critical nature of
parenting to child development. Social learning theory, for example, posits that children
learn to behave in a manner that is consistent with their observations of parental
behaviors (Bandura & Walters, 1963). In this case, parents who exhibit high levels of

externalizing behaviors, for instance, are thought to put their children at risk for similar

68

kinds of behavioral difﬁculties. From many different theoretical perspectives it is clear
that parenting matters to a child’s development. Optimal parenting behaviors, however,
are by no means deﬁned similarly across differing developmental periods (e.g., infancy
vs. preschool). Neither is there complete agreement across disciplines or theoretical
orientations about how best to understand, deﬁne and operationalize parenting constructs.
This review highlights two conceptualizations of parenting that have been presented in
the literature.

One of the most widely cited conceptualizations of parenting behaviors comes
from Baurnrind (1971). Developed using a sample of preschool-aged children, this
investigator carefully and exhaustively examined the behaviors of Caucasian children and
their parents using a multi-method approach which included observational as well as
parent report techniques. She conceptualized parenting as the interaction of two primary
dimensions of parenting which included a warmth/responsiveness dimension and a
control/demandingness dimension. The ﬁrst dimension included parenting behaviors
which ranged from warm and sensitive to cold and hostile. The second dimension
included power-oriented parenting behaviors which ranged from ﬁrm control to lack of
supervision and neglect. In combination, these dimensions yielded four primary parenting
constructs which have subsequently been the focus of much developmental research (for
a review see Rubin & Burgess, 2002). These included: Authoritative parenting (high on
warmth and control), Authoritarian parenting (low warmth, high control), Indulgent-
Permissive parenting (high warmth, low control), and Indifferent-Uninvolved parenting
(low warmth, low control). This conceptualization of parenting is essentially a trait theory

which identiﬁes a categorical taxonomy of parenting, not unlike theories of temperament

69

or personality as applied to the psychological makeup of individuals. In their review and
meta-analysis, Holden and Miller (1999) have argued that the conceptualization of
parenting as a stable trait has been too readily accepted, and not empirically challenged
within the child development and parenting literatures. They stated that,

Another example of the proclivity of researchers to adopt a stable view of

parenting is the prominence of the trait approach to parenting. This orientation

toward similarity in child rearing has been so central to conceptions of child
rearing that it may have precluded reviews on the topic; we were unable to locate
any. . . . the issue of stability and change has long been recognized by
developmental psychologists to be the core issue of the discipline. However, that
discussion has been limited to only one side of the developing dyad — the

children. (p. 224)

Coming from a family systems perspective, Cusinato (1998) outlined a more
dimensional and ﬂuid conceptualization of parenting that included three essential
parenting factors that apply to parental behaviors across child deveIOpment. These
included warmth, control and consistency. Each of these factors was viewed as a
continuous construct such that numerous combinations are possible. Citing the work of
Rollins and Thomas (1979), Cusinato conceptualized warmth broadly as the overall
balance of supportive vs. non-supportive behaviors within the parent-child relationship.
In addition, again based on earlier work by Rollins and Thomas (1979), he described the
construct of control as including subcomponents of frequency and style. Within this
model, parents who were high in terms of their control ﬁequency, for example, would

tend to intervene and, at the extreme, interfere with their children’s behaviors frequently.

70

At the other extreme, parents who were extremely low on this factor, for example, would
be expected to demonstrate negligent behaviors. In contrast, parental style of control
reﬂected the ways in which a parent may intervene with respect to child behaviors. Thus,
style of control could range from coercive behaviors to egalitarian, reason-based attempts
to alter the child’s behaviors. Finally, the construct of consistency was related to the
degree to which the parent’s demands and evaluations of the child’s behaviors were
internally consistent. That is, the degree to which the parents’ overt communications to
the child were consistent with their underlying beliefs and intentions.

While both the conceptualizations of Baurnrind (1971) and Cusinato (1998)
captured important basic elements of parenting as manifest over the course of childhood,
it is equally true that, due to the nature of child development, parenting behaviors and
strategies must shift over time if they are to be effective. Reasoning with an infant, for
example, has no chance of success, whereas reasoning with a preschool-aged child may
be an appropriate strategy. Further, it is likely that the parents’ own constellations of
strengths and weaknesses will inﬂuence their parenting skills differently at each different
stage of development. In light of this, Holden and Miller (1999) have made important
distinctions between the concepts of absolute stability (e.g., does a parent hug their child
with same frequency during the preschool period and the adolescent period), and the
concept of relative stability which reﬂects the degree to which the parent’s relative
position in relation to other parents with respect to a given behavior remains constant. In
fact, in their meta-analysis of 87 longitudinal studies of parenting, they found evidence
for both stability and instability in parenting behaviors. In general, their data suggested

that there is evidence for relative stability but not absolute stability over time. In addition,

71

they found that parenting behaviors tended to be more stable as assessed across time with
older children as compared to infants and young children. These data suggested that
parenting behaviors were more likely to shift early in development than they were later in
development. In a subsequent study using the NICHD Study of Early Child Care (SECC)
dataset (n=l ,364), Dallaire and Weinraub (2005) examined parenting behaviors in a
relatively low-risk group of parents and replicated both of these results. That is, they
found evidence for relative parenting stability across time. In addition, consistent with
Holden and Miller’s (1999) results, they also found that stability in parenting behaviors
was much more likely with older children than with younger children (e. g., more stability
from four to six years versus from two to four years).

Furthermore, in a recent study which also used the NICHD SECC dataset, Belsky
and F earon (2002) sought to investigate the relative inﬂuence of attachment and matemal
sensitivity on child psychosocial outcomes. Attachment security was measured at 15
months and maternal sensitivity, as well as maternal levels of social and family stress,
was measured at 15 and 24 months. When children were three years of age, multiple
psychosocial outcomes including the CBCL for two to 3-year olds were measured. As
predicted, their analyses revealed that children with insecure attachment histories whose
mothers were insensitive at 24 months had the lowest scores on measures of psychosocial
health and ﬁmctioning, and children with secure attachment histories whose mothers
were later rated as sensitive had the highest scores. In addition, children with secure
attachment histories at 15 months whose mothers were insensitive at 24 months scored
lower on measures of psychosocial functioning compared to children with insecure

attachment histories whose mothers were sensitive later in development. This suggested

72

e."
.4
i..j-

A

it

 

that more proximal parenting behaviors were highly predictive of child outcomes, and
could even overcome early insecure attachment histories in some cases. That is, proximal
parenting that was insensitive appeared to have a dysregulating effect on the psychosocial
ﬁrnctioning of young children even in the context of a secure attachment history. This
ﬁnding lends support to the current hypothesis that concurrently assessed parenting
behaviors would signiﬁcantly inﬂuence child outcomes and would account for additional
variance in the model beyond the inﬂuence of infant attachment. While this prediction is
broadly consistent with Belsky and Fearon’s (2002) ﬁndings, it is also notable that, in
their study, effects on behavior problems as measured by the CBCL did not reach
statistical signiﬁcance when SES was controlled. A further ﬁnding indicated that there
was a direct effect of maternal psychosocial stress on maternal sensitivity such that
changes in sensitivity from 15-month to the 24-month evaluations were predicted by
either more or less maternal stress. This last ﬁnding is especially relevant to the current
investigation given that within the present sample some of the mothers were experiencing
elevated levels of psychosocial stress due to DV exposure and other psychosocial risk
factors (e.g., poverty, living in dangerous neighborhoods). Thus, it was predicted that at
the four year evaluation of parenting behaviors in the current study, many of the mothers
in the study would be experiencing high levels of environmental stress, and that this
stress may inﬂuence their parenting abilities in negative ways.
While some of the original work examining the inﬂuence of parenting behaviors on

child behavior was conducted with preschool samples (Baurnrind, 1971), subsequent
studies have focused primarily on either the infancy/toddler period of development or on

child outcomes in samples of school-aged children. In fact, in the ﬁve-volume

73

compilation of parenting topics edited by Bomstein (2002), there are separate chapters
for the infancy (birth to approximately 12 months), toddlerhood (approximately 12
months to approximately 36 months), and middle childhood (5 to 12 years) periods.
However, there is no chapter that reviews parenting during the preschool period.
Presumably, this deﬁcit is related to the relative paucity of research on preschool-aged
children in the extant empirical literature.

As reviewed previously in this paper, parenting during the infant and toddler
periods of development has historically focused on the construct of maternal sensitivity
and its inﬂuence on the formation of infant attachment strategies and internalized
representations (Bowlby, 1969/1982; Bretherton & Munholland, 1999; DeWolff &
Ijzendoom, 1997; Main et al., 1985). In contrast, parenting during the middle childhood
period has focused on parenting behaviors such as those outlined by both Baurnrind and
Cusinato (Baumrind, 1971; Cusinato, 1998). Given the transitional nature of the
preschool period, optimal parenting during this stage of development is likely best
understood as a combination of the parenting behaviors that are involved during both the
infant/toddler period and the middle childhood periods. As described earlier in this paper,
children during this stage of development begin to be capable of emotional self-regulation
(Kopp, 1989; Sroufe et al., 2005). Their regulatory capacities are no longer totally
dependent upon the presence of the attachment ﬁgure as they were earlier in
deve10pment. Concurrent with this shift, societal norms begin to require self-regulation
behaviors from the child outside of the dyadic context. However, due to its transitional
nature, it is likely that children during this period will need to return occasionally to

dyadic regulation strategies — especially when attempts at the developing skills of self-

74

regulation fail them. Thus, it is likely that parenting sensitivity as well as factors such as
parental control (e. g., the use of positive discipline strategies) will all inﬂuence the
preschool child’s ability for regulation during this period. Parenting which is highly
positive (as deﬁned as a combination of parenting skills from the earlier and later
periods) is likely to have regulating effects on child behaviors, whereas parenting which
does not include positive elements (e.g., sensitivity, warmth, appropriate control) is likely
to have dysregulating effects.

Consistent with these predictions, the few studies which have examined these
constructs within preschool samples, have found that parenting behaviors are associated
with the psychosocial and regulatory ﬁmctioning of these children (Deater-Deckard et al.,
2001; Javo, Ronning, Heyerdahl, & Rudmin, 2004; Keown & Woodward, 2006). For
instance, Keown and Woodward (2006) assessed parenting behaviors using a multi-
method approach similar to the current study. They assessed maternal parenting through
the use of maternal self-report of parenting behaviors as well as observer-coded maternal
parenting behaviors during a videotaped mother-child interaction segment. Children in
this sample included a group of four-year-old, Caucasian boys with pervasive
hyperactivity (n=33) as well as a group of comparison boys (n=34) in New Zealand.
Results revealed that mothers of boys in the hyperactive group self-reported higher levels
of lax parenting and higher tendencies to overreact with anger to child misbehaviors
when compared to the comparison group. In addition, during the videotaped segment,
mothers of boys in the hyperactive group were less responsive and less mutually focused
with their sons. Causality of effect can not be implied by these data. However, they do

suggest that parenting behaviors are associated with child regulation capacity during the

75

preschool period. In addition, a Norwegian study (J avo et al., 2004) examined parenting
behaviors as well as child externalizing and internalizing behaviors in a sample of 191
preschool-aged children. Effects of parenting behaviors on child outcomes were found for
girls only. They found that parental cuddling behaviors were negatively correlated with
both internalizing and externalizing behaviors. In addition, physical punishment was
positively correlated with both internalizing and externalizing behaviors and parental
teasing was positively correlated with externalizing problems. This study also suggested
an association with parenting behaviors and child regulation capacities, although this
association was evident for girls only. Finally, in an examination of parenting and child
outcome variables in a preschool sample, one research group found differences in
maternal behaviors of mothers of identical twins (n=62 pairs) (Deater-Deckard et al.,
2001). Psychosocial outcomes of the children were differentially related to their mother’s
parenting behaviors with them. Twins who received higher levels of supportive parenting
and lower levels of punitive parenting were rated as having higher levels of positive
mood and prosocial behavior, and lower levels of negative mood and behavior problems.
In sum, from an attachment perspective, parenting is thought to inﬂuence child
behaviors through its inﬂuence on the child’s internal representational models (Main et
al., 1985; P. Zimmermann, 1999). This process is understood as being most salient during
the early years of development and has been shown empirically to be related to the
mother’s overall ability to provide sensitive caretaking (DeWolff & Ijzendoom, 1997).
Beginning in the preschool years, parenting has been understood as involving several,
more complex dimensions beyond sensitivity (Baurnrind, 1971). For example, parenting

during the early childhood period that includes high levels of responsiveness and

76

appropriate parental control (e.g., discipline) has been associated with positive child
psychosocial outcomes (Rubin & Burgess, 2002). In fact, attachment theory would
predict that, even in the context of an insecure attachment history, positive and
responsive parenting during the preschool period could have a positive inﬂuence of the
child’s internalized representational models and, consequently, on the child’s capacity for
self-regulation.

Parenting during this period can also be understood from a social learning theory
perspective (Bandura, 1977) wherein, through the observation of the parent’s style of
responsive or non-responsive parenting behaviors, the child learns to respond in kind.
From either theoretical perspective it is likely that the experience of proximal positive
parenting is likely to positively inﬂuence the child’s ability for emotional self-regulation.
This study examined the inﬂuence of current maternal parenting behaviors on child self-
regulation capacities and behavioral symptoms of internalizing and externalizing
behaviors. Parenting behaviors were measured using a multimethod design that combined
direct observation of maternal behaviors and maternal self-reported parenting behaviors.
The inﬂuence of DV on parenting behaviors was also examined and is reviewed in the
following section.

The Impact of Domestic Violence on Parentirﬁehaviors

The impact of DV on a mother’s ability to parent her children has increasingly
been explored in the empirical literature (Levendosky & Graham-Bermann, 2001a). DV
may inﬂuence parenting behaviors due to the potentially enduring effects of trauma on
the basic psychosocial functioning of the woman. In addition, the impact of trauma

suffered at the hands of an intimate partner is likely to be uniquely damaging.

77

.94

- an ..I . -

x:
I.

I
15.
,‘IA.
.3»

g

Levendosky and Graham-Bermann (2001 a) have argued that, “Trauma perpetrated by
another person, as opposed to experiencing severe illness or natural disasters, is
simultaneously a psychological, physiological and relational event (p. 29).” In addition to
the relational element inherent in DV, trauma theory, as outlined by Herman (1992),
distinguishes between the effects of acute traumatic exposure versus chronic exposure.
She argued that chronic exposure exacts an additional toll on the woman’s psychosocial
functioning and often results in additional symptoms such as somatic disorders,
depression and dissociation. This increased vulnerability to psychosocial stress is likely
to negatively inﬂuence a woman’s ability to parent effectively.

Empirical examinations of this hypothesis, however, have largely been limited by
small samples, usually of school-aged children, and results have been mixed. In one of
the larger studies of this relationship, McCloskey, F igueredo, and K035 (1995) examined
a cross-sectional sample of 365 mothers of school-aged children (6 — 12 years of age)
using self- and child-report methodology. They found that DV exposed mothers reported
using less warmth with their children compared to non exposed women. In addition,
children’s perceptions of parental warmth and nurturance were negatively correlated with
DV but were not predictive of child mental health outcomes. The use of child-report as
the measure of parenting behaviors (e.g., warmth, nurturance) in this study, however,
raises concerns about reporter bias. Speciﬁcally, the ability of children to accurately
report past events has been contested within the child abuse literature due to evidence
suggesting that children’s ability to remember and report the behaviors of others may be

unreliable (Ceci, Kulkofsky, Klemfuss, Sweeney, & Bruck, 2007).

78

More recently, using maternal self-report, Margolin and Gordis (2003), found that
increases in DV in combination with increases in other life stressors were associated with
increases in harsh and abusive parenting. However, similar to the study by McCloskey
and colleagues (McCloskey et al., 1995), one important limitation of this study was that
observer-rated measures were not included in the study design. Evidence from qualitative
studies has suggested that mothers may experience feelings of remorse about exposing
their children to DV (DeVoe & Smith, 2002). This may result in their propensity to
underreport child exposure to DV as well as their own use of harsh discipline with their
children. Thus, research using observer-rated measures of parent and child behavioral
functioning is necessary to conﬁrm these prior ﬁndings.

In one study using observer-report methodology, Holden and Ritchie (1991),
reported ﬁndings suggesting that some women may be able to maintain positive parenting
behaviors even in the context of abuse (Holden & Ritchie, 1991; Levendosky & Graham-
Berrnann, 2000) In a sample of 37 mothers living in DV shelters and 37 community
matched mothers, these investigators failed to ﬁnd a direct effect of DV status on several
domains of observer-rated parenting including physical affection and punishment. These
researchers did, however, report group differences in the effect of parenting stress on
maternal demonstrations of physical affection. Speciﬁcally, they found that battered
women, rurlike non-battered women, did not demonstrate higher levels of physical
affection in response to lower levels of self-reported parenting stress. That is, battered
women demonstrated similar levels of physical affection (e. g., warmth) toward their
children regardless of whether they were experiencing high or low levels of parenting

stress. In explaining this ﬁnding, the authors hypothesized that mothers experiencing DV

79

T‘l .
USU 5

ﬁt.

UM

fl

:- .. -1
ulkk

2951.5

may have demonstrated a kind of pseudo-warmth toward their children which was
designed to alleviate the mother’s anxiety but was not likely to be beneﬁcial to the
children. They argued that whereas the capacity to demonstrate genuine warmth is likely
tied to levels of stress in a mother’s life and consequently the psychological resources the
mother has available, “pseudo-warmth” does not require the availability of such resources
and therefore is unaffected by parenting stress level. An alternative explanation of this
ﬁnding, however, may be that the coding system utilized by these researchers failed to
capture “warm ” and, instead, captured a different construct. In that case, a revision of
their coding scheme, or the use of an established coding scheme, may have yielded
different results.

To date, the majority of the literature documenting the relation between DV and
parenting behavior has used maternal self-report data as the primary parenting measure.
More recently researchers have begun to use established behavioral observation measures
in an effort to obtain a more objective assessment of parenting outcomes in DV
populations (Levendosky & Graham-Bermann, 2000). In one study using this
methodology, Levendosky and Graham-Bermann (2000) coded mother-child interactions
in 95 families with latency-aged children using an established coding scheme. In contrast
to the results reported by Holden and Ritchie (Holden & Ritchie, 1991), they found that
battered women demonstrated signiﬁcantly less warmth during a semi-structured
interaction task than their non-battered counterparts.

The current study examined the inﬂuence of DV on maternal parenting behaviors
using observer-rated parenting data that were coded using an established coding scheme

that has been previously reported in peer reviewed publications (Whipple et al., 1995).

80

if "
h m

Consistent with trauma theory, it was expected that increases in maternal exposure to DV
would result in decreases in positive parenting behaviors. The cumulative effects of DV

were measured over the course of the ﬁrst four years of the child’s life.

81

I751:

35 .I nil
uLirkiil

ii if,“

‘ﬁ~.h:l
r“

5

L‘.‘ t

CHAPTER 5
HYPOTHESES AND RATIONALE
Using a heterogeneous-for-risk sample of children, this study examined the main
effect of infant attachment category on child emotional self-regulation capacities and
behavioral symptoms of internalizing and externalizing disorders in preschool-aged
children using a longitudinal, prospective research design. In contrast to previous work in
this area that has focused on speciﬁc emotions, this study assessed a broad range of
emotions including positive affect, sadness and anger. It was hypothesized that infant
attachment strategy would predict to later emotional self-regulation capacities when
children were four years of age in the context of a laboratory setting and that these
behaviors would mediate the inﬂuence of infant attachment on child internalizing and
externalizing behaviors.

Although it was originally formulated in the context of her early observations of
relatively impoverished mother-infant dyads in Uganda (Ainsworth, 1967), Ainsworth’s
American studies testing the validity and reliability of the Strange Situation Procedure
were based exclusively on middle-class, Caucasian samples of mothers and infants
(Ainsworth et al., 1978). Subsequent studies, including the longitudinal study by Sroufe
and colleagues (Sroufe et al., 2005) have examined this construct using economically
disadvantaged and minority samples. However, given recent evidence suggesting that
exposure to violence has unique and speciﬁc effects on early child development (Bogat et
al., 2006; DeJonghe et al., 2005), research examining the construct of attachment
classiﬁcation and emotion regulation in children exposed to DV is an important next step

within this area of research. In addition, recent work in the area of early parenting has

82

suggested that proximal parenting behaviors may be more inﬂuential than previous
attachment history on child psychosocial outcomes (Belsky & Fearon, 2002). Therefore,
the present study examined the inﬂuence of the mother’s current parenting behaviors on
these two child outcomes as well.

The emotional self-regulation capacities of preschool children were captured in the
laboratory setting using observer-rated coding methodology. Children in this study
participated in the SSP at two time points: when they were one year and four years of
age. Their emotional self-regulation capacities were measured during their participation
in the SSP when they were four years of age. Only the SSP segments in which the mother
was not present in the room with the child were used to measure the child’s emotional
self-regulation. The methodological strategy of using SSP segments to code discrete child
behaviors (versus attachment category) has been described previously within the child
development literature (Anan & Barnett, 1999; Calkins & Fox, 1992; Dickstein,
Thompson, Estes, Malkin, & Lamb, 1984; F rodi & Thompson, 1985; Ganiban, Barnett,
& Ciccetti, 2000). In addition, behavioral symptoms of internalizing and externalizing
disorders were measured using maternal report methodology (e. g., Child Behavior
Checklist; CBCL). Speciﬁc hypotheses are described below. Figure 1 represents the

hypothesized study model.

83

 

 

4 Id 1
year 0 lﬁewﬁrl [s l

 

 

 

 

 

 

 

 

 

 

 

 

 

Maternal
Parenting
4 year
old CBCL
Extern.
1 year old 4 year old
Infant Child
Attachment Regulation
4year
old CBCL
Intern.
DV DV l
Mother Child l - _. ‘ F—i

Fi 1: Stud Model

M

It is clear that attachment theory and the research it has generated has inﬂuenced
our understanding of the development of emotion regulation processes (Main et al., 1985;
Sroufe et al., 2005; Vondra et al., 2001; P. Zimmermann, 1999). Attachment research has
taken a lifespan perspective in that the literature reﬂects the application of attachment
measures in every age-group of human development (Cassidy & Shaver, 1999). While
the adult attachment literature has experienced a dramatic increase in empirical work,

however, the examination of emotional self-regulation as it relates to attachment

84

categories in children, has experienced relatively less growth. Additionally, prior to the
current study, there were no studies which comprehensively examined the longitudinal
effects of early attachment classiﬁcation on later regulation capacities in preschool-aged
children. Therefore, the present study contributes to this area of research by explicating
this relationship in a heterogeneous-for—risk sample of children using a longitudinal
dataset which included infant attachment category measured at one year of age as well as
an observer-rated measure of child regulation and a parent-rated measure of child
externalizing and internalizing behaviors at four years of age. Analyses were conducted
using the 3-category attachment classiﬁcation systems. Hypotheses one through ﬁve were
tested within the full SEM model. Hypotheses six and seven represent separate sets of
analyses, as described below.
SEM Model: Hypotheses 1 - 5

Hypothesis 1: Infant attachment_category will predict to the child’s overall
Mity for emotional self-regulation at Hears of age in a laboratcmz environment.
Attachment theory posits that children who over- or under-regulate their emotions are
accommodating to a less sensitive parenting relationship. Thus, children who behave in
either an over- or under- regulated manner are demonstrating a pattern of regulation
which does not allow them to authentically experience and express their emotions. The
child who over-regulates is thought to internalize feelings of distress such that they result
in feelings such as sadness and shame which, while they are internally experienced as
painful, are not outwardly expressed. In essence, these children have learned to rely on
themselves for comfort and do not tend to actively seek comfort from others. Thus, in the

current study, infant attachment category was expected to inﬂuence observer-rated child

85

.
“a...

.
_~;‘-.M
-

Xfll'f

1'“:on

Ll}.

it.-
El» ll]

3C7? 31

F?
r

mo

H
’27
. .
A

scores

regulation capacities such that children who were avoidantly attached as infants would
receive high positive regulation scores. The child who under-regulates (e. g., becomes
dysregulated), in comparison, is thought to outwardly express strong emotions in a
manner which does not allow for the capacity for self-soothing or self-regulation to
develop. Instead, these children seek others to deal with their emotions but are never fully
able to internalize a sense of self-care and soothing. In the current study, children who
were ambivalently attached as infants were expected to have very low positive regulation
scores. Children who are able to experience and express a wide range of emotional
reactions can authentically engage in relationships without the need to alter their affects
or behaviors inauthentically in order to meet the emotional needs of another person.
Thus, children who were securely attached as infants were expected to receive mid-range
scores of positive emotional self- regulation.

Hypothesis 2: Infant ﬂachment categol will predict to behavioral symptoms of
internalizing and externalizing disorders at four years ofjgg Maternal ratings of
children’s externalizing and internalizing behaviors were expected to be related, in
theoretically consistent ways, to the infant attachment categories. That is, children with
secure attachment histories were expected to receive low CBCL scores on both the
externalizing and internalizing scales. In addition, children with ambivalent attachment
histories were expected to receive relatively higher scores on the externalizing CBCL
scale due to the externalizing symptoms that characterized them as infants. Finally,
children with avoidant attachment histories were expected to receive relatively higher
scores on the internalizing CBCL scale due to the highly regulated style that

characterized them as infants.

86

HYQOthesis 3: Child rggulatory capacities will mediate the relationship between
infant attachment category and later child externalizing and internalizing behaviors.
Given the growing evidence that child psychopathology is directly related to a child’s
capacity for emotion regulation (Mullin & Hinshaw, 2007), it was predicted that child
emotional self-regulation would mediate the effect of infant attachment strategy on later
child externalizing and internalizing symptoms. Children who were rated as under-
regulated in the laboratory environment were expected to evince higher levels of
externalizing behaviors whereas children who appeared over-regulated were expected to
evince higher levels of internalizing behaviors.

Hypothesis 4: Exposure to DV will inﬂuence maternal parentingimhaviors, child
ﬁgulation capacities. and symptoms of externalizing and internalizing disorders.
Examination of the emotional and behavioral sequelae of trauma exposure in children has
been a recent focus in the child development literature (Osofsky, 2004). There is
evidence that exposure to tramna may have a unique inﬂuence on the emotional and
behavioral systems of the developing child relative to its inﬂuence on the more stable
personality structure of the adult survivor. It was hypothesized that early DV exposure
would inﬂuence later child regulation capacities in predictable and theoretically
consistent ways such that increased DV exposure would lead to decreases in child self-
regulation capacities and increases in externalizing and internalizing behaviors.

Exposure to DV has also been shown, in some studies, to inﬂuence maternal
parenting behaviors (Levendosky & Graham-Bermann, 2001a; McCloskey et al., 1995),
although other studies have failed to ﬁnd these effects (Holden & Ritchie, 1991). Trauma

theory predicts that exposure to DV will negatively inﬂuence many aspects of a mother’s

87

psychosocial functioning, including her parenting (Herman, 1992; Levendosky &
Graham-Bermann, 2001b). In the current study it was predicted that cumulative exposure
to DV across the ﬁrst four years of the child’s life would negatively inﬂuence maternal
parenting behaviors when the child is four years of age.

Hypothesis 5: Current matepnal parenting behaviors will inﬂuence both child
ragulation capacities in the laboratory and symptoms of externalizing and internalizing
disorders. Just as children’s relationship representations and consequent regulation
capacities may be malleable during early childhood if children are exposed to high levels
of psychosocial stress, a parent’s ability to parent effectively may be somewhat ﬂuid
during the early parenting period depending on her own life experiences (e.g., DV
exposure, access to social support, etc.). These analyses examined whether the more
proximal variable of current parenting behaviors would inﬂuence the child’s regulation
capacities and behavioral symptoms of externalizing and internalizing disorders beyond
the effects of the early attachment relationship (which itself is based on early maternal
parenting behaviors). Speciﬁcally, it was predicted that exposure to proximal positive
parenting would be positively associated with child self-regulation capacities and
negatively associated with symptoms of internalizing and externalizing behaviors.
Additional Analyses: Hypotheses 6 - 7

Hypothesis 6: The presence of DV in the home will moderate the relationship
between infant attachment categog and internalizing and externalizing behaviors. Tenets
of developmental psychopathology and attachment theory predict that early attachment
will predict to later symptoms of psychopathology in the presence of other risk factors. In

the current study, it was predicted that children whose mothers reported DV exposure in

88

 

the home would evince higher levels of internalizing and externalizing behaviors in
preschool, compared to children whose mothers did not endorse DV exposure.
Hypothesis 7: Infant attachment category will predict differentially to speciﬁc
types of child regulation behaviors (e.g., self- versus other-focused behaviors). In the
current study, one facet of preschoolers’ emotional self-regulation strategies that will be
explored involves the speciﬁc types of behaviors children engage in when they are
attempting to cope with a stressful situation. It was expected that a child’s behaviors
within the stress-inducing SSP will vary based on his or her infant attachment category.
Speciﬁcally, it was predicted that in the laboratory setting children who were avoidantly
attached as infants would exhibit increased numbers of self-soothing techniques (e. g., self
stimulation, talking to self) in the context of a stressful situation and would not outwardly
express their feelings of distress. In contrast, it was predicted that children who were
ambivalently attached as infants would exhibit increased numbers of other-directed
regulation techniques (e.g., seeking contact with others). Finally, it was expected that
children who held secure attachments as infants would be ﬂexible in their use of
behavioral regulation strategies and will be able to simultaneously use a balance of both
self- and other-directed regulation techniques as needed to facilitate their emotional self-
regulation. These children were expected to exhibit some signs of distress and
dysregulation but also to utilize diverse mechanisms with which to calm themselves such

that they were not expected to become utterly dysregulated.

89

CHAPTER 6
METHOD
Participants

Participants of the current study included 203 women and children participating in
a longitudinal study examining child risk and protective factors in a group of mothers
approximately half of whom reported exposure to DV at the time of recruitment. Women
were initially recruited into the study when they were in their third trimester of
pregnancy. Recruitment efforts included the posting of ﬂyers at local public stores and
agencies (e.g., laundromats, grocery stores, Head Start offices, etc.) and medical clinics
(e.g., Obstetric/Gynecology clinics, public health clinics, etc.) in a medium-sized,
Midwestern city and surrounding areas. Two versions of recruitment ﬂyers were utilized.
Initially ﬂyers advertised a study about mother-infant relationships. Later, a more focused
attempt was made to recruit women experiencing DV and consequently ﬂyers were
distributed that invited women to participate in a study about the experience of DV
during pregnancy. Approximately 46% of women reported DV exposure when children
were one year of age (see Table 5). As Table 5 indicates, by the time the children were
four years of age, 70% of the women had experienced incidents of DV at some point over
the course of the ﬁrst four years of the child’s life.

The current sample represents a sample of convenience. Pregnant adult women
who were between 16 and 40 years of age and who responded to the ﬂyers were invited
to participate in the study. Potential participants were screened to insure that they had

been involved in a romantic relationship for at least six weeks during the pregnancy and

90

 

that they were able to speak and understand English well enough to complete the
questionnaires and participate in the interviews.

The longitudinal study, from which this sample is drawn, is comprised of 206
mother-child dyads. The longitudinal study began when the mothers were pregnant with
the study child. Three of the mothers from the longitudinal sample withdrew from the

study aﬁer the pregnancy interview and did not participate in any future waves of data

collection. The current study uses data beginning when the children were one year of age

 

(possible n = 203). For the purposes of the current study, the one-year wave of data
collection will be labeled TI. This study will include data that have been collected at four
time points: T1 was conducted when the children were one year of age; T2, T3 and T4
were conducted every year thereafter, around the date of the child’s birthday. For the
purposes of this study, missing data resulting from attrition within the longitudinal
sample will be estimated using full information maximum likelihood (FIML) such that
the current analyses will include 203 mother-child dyads. The current data meet the
criteria for use of FIML including the criterion that missingness on the outcome variable
(y) is not dependent on y itself (L. M. Collins, Schafer, & Kam, 2001). The FIML
approach ﬁts the model to all of the non-missing values for each observation and has all
of the strengths of multiple imputation (MI) including maximizing power by using all of
the participants in the analysis (Widaman, 2006). In addition, the M—plus program, which
is used in these analyses, employs test statistics in FIML that are robust to non-normally
distributed data (Allison, 2003; Muthen & Muthen, 2007).

Within the longitudinal sample (n = 203), child gender was approximately evenly

split (males = 51%). Participants represented a range of cultural and ethnic groups,

91

educational backgrounds and occupations (see Tables 1 - 3). Participants also represented
a range of socio-economic groups (see Table 4). At the time of recruitment into the study,
approximately 40% of the women were married, 50% were never married, 4.5% were
separated, 5% were divorced and 1 woman was widowed. Ninety-three percent of the
families in the longitudinal sample resided in Eaton and Ingham counties. The
demographic characteristics of the longitudinal sample were consistent with the census

data in these counties at the time of recruitment.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Caucasian / African Latina Multi- Other,
White American / Racial Unknown
Black or Missing
Mother 64.0% 25.1% 4.9% 3.9% 2.1%
Father 49.9% 34.7% 5.4% 5.0% 5.0%
Child 46.3% 24.6% 2.0% 23.6% 3.5%
Table 1: Ethnicity of Family Members — mrcentages (n = 203)
Less than High Some BA or BS Some Unknwn
High School College, Graduate or
School Education Associates school or Missing
education or GED Degree or Graduate
Trade Degree
School
Degree
Mother 16.3% 28.1% 39.4% 7.8% 5.4% 3.0%
Father 14.3% 36.0% 30.0% 10.3% 6.4% 3.0%

 

 

 

 

 

 

 

 

ﬁble 2: Mother and Father Educational Achievement data Lpercentages (n = 203)

 

 

 

 

 

 

 

 

 

 

 

menial unskilled semi- skilled manual clerical,
service workers skilled workers, sales
workers workers craftsmen

Mother 18.7% 18.3% 19.7% 10.4% 10.3%

Father 19.7% 13.8% 26.1% 11.3% 2.5%
technicians, semi- small administrators, executives, large
professionals business medium business owners

owner business owners

Mother 13.8% 3.9% 4.9% 0%

Father 7.9% 9.4% 5.4% 3.9%

 

 

 

Table 3: Mother and Father Occupation Data ;percentages (n = 203)

93

 

 

 

 

 

 

 

 

 

 

 

 

Income 2? Median Income Household Household
(sd) Income Range Size Size range
3? (sea
T1 — 1 yr olds $2,201 $1,500 $267 — 3.9 2 — 9
(n = 189) ($1,752) $10,000 (1 .4)
T2 — 2 yr olds $2,487 $2,000 0 - 4.0 2 — 9
(n = 186) ($2,199) $18,000 (1.4)
T3 — 3 yr olds $2,472 $1,700 0 - 4.1 2 —- 13
(n = 178L ($2,152) $17,000 (1.5)
T4 — 4 yr olds $2,552 $1,950 $184 - 4.2 2 - 14
n = 177) ($1,987) $13,000 (1.6)
% % % %
receiving receiving receiving receiving
public WIC Food Medicaid
ﬁnancial Stamps
assistance
T1 — 1 yr olds 7.9% 57.1% 26.5% 59.8%
(n = 189)
T2 - 2 yr olds 10.2% 43.5% 31.2% 54.3%
n = 186)
T3 - 3 yr olds 8.4% 43.8% 35.4% 56.2%
n = 178)
T4 - 4 yr olds 2.3% 39.0% 35.0% 51.4%
n = 177)

 

 

 

 

 

 

Table 4 — Socioeconomic Status Data by Wave of Data Collection: T1 -— T4

 

 

 

Tl-lyr T2—2yr T3—3yr T4—4yr %abusedatsome
olds olds olds olds point over the 4
year time period
45.6% 51.0% 40.8% 46.1% 70.0%

 

 

 

 

 

 

ﬂble 5 - Maternal Rates of DV Exposure by Year (% abused)

94

 

I f;‘_,
1119.. l

Data were collected at four time points:
Time 1 (n = 189): Infant attachment catagory (SSP); DV exposure (SVAWS)
The average age of the mother at the one-year interview was 26.74 years (sd =

5.00; range =19.21 — 41.55), and the average age of the child was 1.10 years (sd = .11;
range = .92 —- 1.10).
Time 2 (n = 186): DV exposure (SVAWS)

The average age of the mother at the two-year interview was 27.70 years (sd =
5.04; range =20.19 - 42.50), and the average age of the child was 2.03 years (sd = .07;
range = 1.77 — 2.40).
Time 3 (n = 178): DV exposure (SVAWS)

The average age of the mother at the three-year interview was 28.55 years (sd =
5.12; 21.21 —43.53), and the average age of the child was 3.01 years (sd = .08; range =
2.87 - 3.79).
Time 4 (n = 177): Observed child regulation behaviors (SSP separational'LParent report
of child- Intemalizing and externalizing behaviors (CBCL); Observed parentirrg
he;h_aviors G’CIT): DV exposure (SVAWS) ; Self report of parenting behmors (PBC)

The average age of the mother at the four-year interview was 29.58 years (sd =
4.99; range =21.42 — 44.58), and the average age of the child was 4.04 years (sd = .12;

range = 3.09 — 4.82).

95

.‘ “

Procedures

Initial Screening

Women contacted the project ofﬁce to inquire about participating in the study.
Initial contacts with potential research participants involved brief phone screenings that
were conducted by trained research assistants. The screening included an assessment of
the woman’s age, pregnancy status, length of time with current romantic partner and DV
status. Women were informed that the study was investigating women’s relationships
with the important people in their lives including partners, family members, and children.
They were informed that if they chose to participate in the study they would be asked to
talk about their thoughts and feelings about their relationships and their recent life events,
including DV, and that they would be paid for their participation in the study.

Subsequent to the recruitment of approximately 50% of the sample,
administration of the Conﬂict Tactics Scale (Straus, 1979) was implemented during the
initial screening contact. This measure was used to exclude women who had not
experienced DV during pregnancy in order to oversarnple women who were experiencing
DV. During this phase of participant recruitment, potential participants were informed
that they would be asked some questions about themselves and their relationships and
that, in an effort to ensure that the study included a representative sample of women from
the community, they may or may not be eligible to participate based on this information.
A total of 161 women contacted the project ofﬁce but were deemed ineligible to
participate because they did not meet age, relationship status, or battering experience
criteria. No demographic differences existed between the excluded women and the

research participants.

96

f Y-
ﬁ‘

ii Eli
t

ml

ml .-
Lr. .1:

Retention Procedures

Participant retention efforts were initiated approximately one week after the
infant’s projected due date. Phone calls were made to the women to conﬁrm each infant’s
date of birth. Subsequently and for the duration of the study, participants were contacted
by mail every 90 days between interviews. Mail correspondence included a letter and a
form which was completed by the participant with their current address, phone number,
and names and numbers of friends or family members who could be contacted in the
event that we could not reach the participant. A self-addressed and stamped envelope was
included in this mailing. In addition, a contract with the US Post Ofﬁce was established
such that they generated and sent a postcard with current addresses for those participants
who had moved and registered a new address. When participants did not return their
information sheet within three weeks, they were contacted by phone. In the event that the
participant could not be reached directly by phone, we contacted the friends and/or family
members who had been identiﬁed by the participant for recontact purposes. Participants
received ten dollars in monetary compensation for returning their information forms at
each recontact time point.
Time 1: Mothers and Their One-Year Old Children

When their infants were approximately one year of age, women were contacted to
schedule the Time 1 interview. Mothers and infants were interviewed at the laboratory.
The Strange Situation Protocol (SSP) (Ainsworth et al., 1978) was administered at this
time. In addition, the Severity of Violence Against Women Scales (SVAWS) (Marshall,

1992) were administered. Trained research assistants administered the maternal

97

tarts ‘
575$ and

Tit-e 3: ll

 

will

I l ‘
., ..\ . ..
l‘JEtil'idU

castle

“I:

interviews while the infants were cared for in a separate playroom. Women were paid
$75.00 and given a baby gift worth $8.00 after completion of the Time 1 interview.
Time 2: Mothers and Their Two-Year Old Children

When their children were approximately two years of age, women were contacted
to schedule the Time 2 interview. Children were not assessed directly during this wave of
data collection. Mothers were interviewed by trained research assistants in the woman’s
home, in the project ofﬁce or over the phone depending on the woman’s preferences and
her personal circumstances (e. g. women who had moved to another state were
interviewed over the phone). The SVAWS was administered at this time (Marshall,
1992). Women were paid $75 after the completion of the Time 2 interview.
Time 3: Mothers and Their Three-Year Old Children

When their children were approximately three years of age, women were contacted
to schedule the Time 2 interview. Children were not assessed directly during this wave of
data collection. Mothers were interviewed by trained research assistants in the woman’s
home, in the project office or over the phone depending on the woman’s preferences and
her personal circumstances. The SVAWS was administered at this time (Marshall, 1992).
Women were paid $90 after the completion of the Time 3 interview.
Time 4: Mothers and Their Four-Year Old Children

When their children were approximately four years of age, women were contacted

to schedule the Time 4 interview. Mothers and infants were interviewed at the laboratory
whenever possible. In a few cases, home visits were conducted at the mother’s insistence.
The Mgr: Situation Protocol (SSP) was administered at this time, as was the Egg;

Child Interaction Tasl_( (PCIT). In addition, the SVAWS (Marshall, 1992) and the Parent

98

 

Better 1
IBM-
arch a

tarps

mist

firearm:
‘

Stage .

lit-yea]

than

Behavior Checklist (PBC) (Fox, 1994) were administered. The Child Beﬂior Checglﬁ
(CBCL) (Achenbach, 1991) was also administered to the mother at this time. Trained
research assistants administered the maternal interviews while the children were cared for
in a separate playroom. Women were paid $150 after the completion of the Time 4
interview.

Measures
Assessment of Attachment Category
Strange Situation Protocol (Ainsworth et al., 197 8)

This is a 22-minute structured behavioral protocol wherein the mother and her
one-year-old infant are observed interacting in an unfamiliar environment (the
laboratory). There are eight episodes (described below), each designed to impose a small
but gradually increasing degree of psychological stress on the dyad (see Table 5). The
three mother-infant separations, in particular, are designed to activate the attachment
system of the infant. Four 7-point scales including proximity seeking, contact
maintaining, avoidance, and resistance and one 9-point scale capturing infant
disorganization were coded based on observed infant behaviors (Ainsworth et al., 1978;
Main & Solomon, 1990). Based on their behavioral patterns and responses to this
protocol infants were coded into one of four categories: secure, ambivalent, avoidant and
disorganized. A forced-choice A/B/C category was also assigned. The protocol was
videotaped and tapes were coded by two professional social workers at the University of
Washington who had received speciﬁc training and demonstrated reliability in coding
SSP protocols. Reliability was calculated using 11% of the sample and 90% agreement

was achieved on attachment classiﬁcations, yielding a kappa of .84 (p<.001). The Kappa

99

mi

'3:
- riff]

the i
an
Matti?

at due

Ambit;

statistic provides a chance-corrected index of categorical agreement (Cohen, 1968).
Differences in classiﬁcations were resolved by conferencing.

There were a total of 177 SSP procedures administered in the current study. Thus,
of the 189 families who otherwise completed this wave of data collection, there were 12
missing SSP administrations: 6 of the families lived out of state, 4 of the mothers had lost
custody of their children, 1 mother refused to be videotaped and, in one case, data were
lost due to technical errors that occurred in the videotaping process.

Data were coded at both the 4-category (A/B/C/D) and the 3-category (A/B/C)
levels. For the 3-category coding system, a forced classiﬁcation was used. Children in the
disorganized group and children who had been coded as “unclassiﬁable” were assigned to
one of the 3 categories. Of the children who were originally coded as disorganized
(n=l7), 3 received a forced classiﬁcation within the Avoidant group, 10 were classified
as Secure, and 4 were classiﬁed as Ambivalent. In addition, of the 9 infants who were
originally in the unclassiﬁable group, 2 were classiﬁed as Secure and 7 were classiﬁed as

Ambivalent. Sample sizes for the three-group categorization are reported in Table 6.

 

 

 

 

 

 

 

 

 

 

 

 

 

Episode Duration Description

1 1 minute Parent & Infant are introduced to room by
experimenter

2 3 minutes Parent and Infant are alone in room

3 3 minutes Stranger enters: sits quietly for 1 minute, engages
mother for 1 minute, engages infant/child for 1 minute

4 3 minutes Infant and Stranger alone in room

5 3 minutes Mother returns, stranger leaves

6 3 minutes Mother leaves infant alone in room

7 3 minutes Stranger returns: Infant and Stranger alone in room

8 3 minutes Mother returns, parent leaves

 

Table 6: Strange Situation Test Protocol

100

 

 

 

 

 

 

 

 

 

 

 

 

One year (T1) Strange Situation
Forced Choice Classiﬁcation
Avoidant Secure Ambivalent Total
Avoidant 31 0 0 31
_ a Secure o 96 0 96
a}: Ambivalent o 0 24 24
: g Disorganized 3 10 4 17
° 5 Unclassiﬁable o 2 7 9
Total 34 108 35 177

 

 

 

 

hble 7: Strange Situation Classiﬁcations — One-year old children

Assessment of Domestic Violence
Severity of Violence Against Women Scales — Maternal Exposure (SVAWS) (Marshall,
1992)

The SVAWS is a 46-item questionnaire designed to assess both violent behaviors
and threats the woman has experienced from her partner(s). The scale is composed of
nine categories of abuse including symbolic violence, threats of mild violence, threats of
minor violence, threats of moderate violence, threats of serious violence, mild violence,
minor violence, moderate violence, serious violence, and sexual violence. Examples of

9, ‘6

items include “destroyed something belonging to you, punched you,” and “demanded
sex whether you wanted to or not.” Respondents were instructed to rate their experiences
of abuse on a 4-point scale ranging from “Never” to “Many Times.” Scores were
calculated for each woman during each wave of data collection. Scores were summed for
women who reported exposure to abuse from multiple partners during the same one-year
period, yielding a composite score for each woman for each year. In addition, a

dichotomous DV variable was created for use in the moderation regression model

(hypothesis 6), by conducting a median split of the data. High internal consistency (a =

101

.97) has been reported for the full scale (Huth-Bocks, Levendosky, & Semel, 2001).
Analyses using Cronbach’s alpha suggested high internal consistency across the four time
periods in the current sample as well: Partner 1: a = .95 at age 1, a = .95 at age 2, a = .94
at age 3, a = .94 at age 4; Partner 2: a = .99 at age 1, a = .97 at age 2, or = .97 at age 3, a =
.96 at age 4.

Severity of Violence Against Women Scales — Child Witnessing (SVAWS) (Marshall,
1992)

Following the administration of the SVAWS to the mother, each item for which
she endorsed DV was followed up by asking her if her child had either visually witnessed
or overheard that particular experience. Scores were calculated in an identical manner to
the maternal experience of abuse; scores were summed and calculated for each child
during each wave of data collection. Analyses using Cronbach’s alpha suggested high
internal consistency across the four time periods in the current sample: Partner 1: a = .97
at age 1, a = .92 at age 2, a = .90 at age 3, a = .89 at age 4; Partner 2: or = .85 at age 1, or =
.98 at age 2, a = .93 at age 3, a = .96 at age 4.

Assessmentgf Parenting Behaviors
Parent-child interaction task, adapted (PCIT) (Eyberg & Robinson, 1982)

An adapted version of the PCI was utilized. This protocol involves the mother and
child sitting next to each other at a table with a variety of toys and activities. The adapted
protocol lasts a total of 14 minutes and is equally divided into two situations: child
directed play and mother directed play. The coding scheme used to score these data was
adapted from work done by Whipple and colleagues (Crandell, Fitzgerald, & Whipple,

1997; Whipple et al., 1993; Whipple et al., 1995) (see Appendix). This research team

102

 

 

based their coding system on work done by Belsky and colleagues (Belsky, Youngblade,
Rovine, & Volling, 1991). In addition, the sensitivity scale was based on Biringen’s
Emotional Availability Scales (Biringen, 2000; Biringen et al., 2000). There are 7, 5-
point, maternal scales which are scored in one-minute intervals: positive aﬂect and
aﬂection, negative aﬂect and feedback, positive or neutral feedback, facilitates self-
regulation, intrusive/overcontrolling, unresponsive/unavailable/undercontrolling, and
sensitivity. Scores for each domain were then averaged across all 7 one-minute segments
for each of the two protocol situations. This yielded a mean score which was used as the
ﬁnal score for that domain. Thus, each case received a ﬁnal average score for each
domain (e. g., positive affect). In the current analysis, the 2 domain scales were then
averaged, yielding one score for each code. The scale scores were then factor analyzed to
create a latent, positive parenting variable.

Coders were undergraduate students who received intensive training by a lead
graduate student and with assistance from the author of the current study (CD). Ongoing
supervision was provided by the lead graduate student who served as the gold standard
for reliability coding and each coder established and maintained reliability with this lead
graduate student. Consistent with other published work in the ﬁeld (Slade, Belsky, Aber,
& Phelps, 1999), weighted kappas were calculated for each scale (Schuster, 2004). The
Kappa statistic adjusts for chance agreement among coders (Cohen, 1968). Altman
(1991) identiﬁed strength of agreement as Very Good for kappa values within the .81-1 .0
range, Good within the 61-80 range, and Moderate within the .41-.60 range. Kappa
weights were assigned such that exact matches received a 1.0 weight, l-point differences

were weighted at .75, 2-point differences were weighted at .25, and remaining differences

103

 

 

received a 0 weight. Initial reliability was established between the lead graduate student
and each of the four coders on a random selection of 15% of the tapes. After establishing
initial reliability, double coding was conducted at regular intervals to minimize rater drift.

Final weighted kappa values ranged from .60 to 1.0:

 

 

 

 

 

 

 

 

Scale Coder l Coder 2
Positive Affect .97 1.0
Negative Affect .88 .95
Positive Feedback 1.0 1.0
Facilitates Self-Regulation .76 .82
Intrusive .60 .82
Unresponsive .82 .85
Sensitivity .89 .95

 

 

 

 

 

Table 8: Weighted Kappa Reliability Statistics for the Parentinglhta

Parent Behavior Checklist, Discipline subscale (Fox, 1994).

This is a lOO-item scale measuring mother’s self reports of their parenting
behaviors. The measure is comprised of three subscales: expectations, discipline, and
nurturing. Responses were scored on a 4-point Likert scale ranging from “Almost
Never/N ever” to “Almost Always/Always.” For the present study, the discipline (reverse
coded such that higher scores reﬂect positive discipline) was used. The discipline
subscale is comprised of 30 items (e.g., “I send my child to bed as a punishment”). High
internal consistency has been reported for the full scale (a = .93) and for the discipline
scale (0. = .91) (Fox, 1994). Internal consistency for the discipline subscale used in the
current sample was good (a = .86).

Assessment of Child Emotional Self-Regulation Capacities

Strange Situation Protocol (Ainsworth et al., 1978) —- Coding of Discrete Child Behaviors

104

This is a 22-minute structured behavioral protocol wherein the mother and her
four-year-old child are observed interacting in an unfamiliar environment (the
laboratory). Identical to the one-year old protocol, there are eight episodes (described
above), each designed to impose a small but gradually increasing degree of psychological
stress on the dyad. The three mother-child separations, in particular, place stress on the
regulatory capacity of the young child. Child behaviors during three of the 3-minute
separation episodes of the SSP were coded. The episodes include the two child/stranger-
together episodes in addition to the child-alone episode. Of the 177 families who
participated in the T4 wave of data collection, 20 families were unable to complete the
SSP, either because they lived out of state or they were unable to come to the laboratory
session (in which case the collection of the remaining data was completed during a home
visit or over the telephone). In addition, one child was unable to tolerate separations from
his mother and, therefore, there were no separation segments available for coding. Thus,
four-year-old SSP data were available for 156 families. Three segments were coded for
each child, yielding a total of 468 segments.

Seven child scales were scored in three-minute intervals: comfort/enthusiasm,
anger/frustration, sadness, emotional lability, activity level, attentional regulation and
self-regulation (see Appendix). All scales were scored using a ﬁve-point, interval,
anchored rating system. Scales were designed such that higher scores reﬂected higher
levels of that particular construct and were adapted from several sources:
comfort/enthusiasm, anger/frustration and sadness scales: (Miller, Gouley, Seifer,
Dickstein, & Shields, 2004; Whipple et al., 1993; Whipple et al., 1995); emotional

[ability scale: (Clark, 1985, 1999); activity level scale: (Seifer, Sameroff, Dickstein,

105

Schiller, & Hayden, 2004); attentional regulation scale: (Clark, 1985, 1999); self-
regulation scale: (Clark, 1985, 1999; Cole, Barrett, & Zahn-Waxler, 1992; Miller et al.,
2004).

The ﬁrst six scales captured speciﬁc elements of the child’s behaviors and
reactions in the stressful context. The seventh scale, self-regulation, was a more global
score which captured the child’s overall ability to regulate, irrespective of the way in
which regulation was, or was not, achieved. For example, a child may have received a
low score on self-regulation due to the fact that she was angry and pounding on the door
for the entire segment, or, she may have received a low score because she sat in the chair
and stared into space while whimpering quietly. In each case, the child was unable to
remain regulated in the sense that she was unable to engage in any focused activity; both
children would have received low scores on self-regulation. However, in the ﬁrst case the
child would have received high scores on anger and in the second case she would have
received high scores on sadness. In the current analysis scale scores were factor analyzed
to create a latent, positive child regulation variable.

In addition to the seven scales described above, ﬁve self-regulation strategy scales
were also coded including: contact seeking, attention to toy/self play, physical self
stimulation, talk/sing to self, and attempts to leave room (see Appendix). These scales
were also scored using a ﬁve-point, interval, anchored rating system and were designed
such that higher scores reﬂected higher levels of that particular construct. Conceptual
work by Stansbury and colleagues (Stansbury & Sigman, 2000; L. K. Zimmermann &
Stansbury, 2003) informed the creation of these codes in that these researchers have

argued for the importance of assessing the discrete behaviors children use to regulate

106

 

their emotions in stressful situations. These investigators have assessed child regulation
in the presence of the parent and their coding scheme assesses the broad domains of, self-
comforting, instrumental regulation, distraction and cognitive regulation. However, the
codes used in the present study were independently developed for the purposes of this
study and are theoretically grounded in attachment theory which posits differential styles
of self versus other behavior in children with differing attachment histories.

As with the observed parenting data described above, coders for this dataset were
undergraduate students who received intensive training and ongoing supervision by a
lead graduate student (in this case, C.D.) who served as the gold standard. Each coder
established and maintained reliability with the lead graduate student and weighted kappas
were calculated for each scale (Schuster, 2004). Consistent with the parenting data, kappa
weights were assigned such that exact matches received a 1.0 weight, l-point differences
were weighted at .7 5, 2-point differences were weighted at .25, and remaining differences
received a 0 weight. Initial reliability was established between the lead graduate student
and each of the four coders on a random selection of 22% of the tapes. Aﬁer establishing
initial reliability, double coding was conducted at regular intervals to minimize rater drift.

Final weighted kappa values ranged from .75 to 1.0:

107

 

 

 

 

 

 

 

 

 

 

 

 

 

Scale Coder 1 Coder 2 Coder 3 Coder
4
comfort/enthusiasm 1 .0 1 .0 .96 1 .0
anger/ﬁ'ustration .77 .70 .90 .92
sadness 1.0 .93 .95 .95
emotional lability 1 .0 1 .0 1.0 l .0
activity level .92 .93 .91 .86
attentional regulation .91 1.0 .92 .93
self-regulation l .0 l .0 1 .0 .87
contact seeking .79 .94 .82 .91
attention to toy/self play .79 .85 .79 .90
physical self stimulation 1.0 1.0 .93 1.0
talk/sing to self .78 .87 .75 .77
attempts to leave room 1.0 1.0 1.0 .92

 

 

 

 

 

 

 

ILble 9: Weighted Kappa Reliabihiy Statistics for the Child Regulation Data

Assessment of Child External izing and Intemalizing Behaviors
Child Behavior Checklist (C BCL) — Mother Report F arm (Achenbach, 1991)

This is a 112-item, parent-report instrument that measures the child’s social and
emotional functioning over the last six months. Eight subscales, two broad band
subscales of internalizing and externalizing behaviors and a total problem behavior score
were calculated. Participants rated behavioral descriptions of their child on a 3-point
scale from “Not True” to “Very True or Often True.” In the present analyses scores from
the externalizing and internalizing subscales during the T4 (four year old children) wave
of data collection were utilized. The T4 alpha for the full scale in the current sample is

.92

108

CHAPTER 7
RESULTS
Preliminary Analyses
Initial examination of the data (e.g., distributions, descriptive statistics,
ANOVAs.), as well as exploratory factor analyses (EFAs) were conducted using SPSS,
version 15.0 (SPSS, 2006) . Conﬁrmatory factor analyses (CFAs) and structural equation
modeling (SEM) analyses were conducted using M-plus, version 4.1 (Muthen & Muthen,
2007). As described above, CFA and SEM analyses used full information maximum
likelihood (FIML) estimation. See Tables 10, 11 and 12 for means, standard deviations
and correlations of the variables under study. The income and discipline variables were

standardized before inclusion in the analyses. The remaining variables were not

standardized.

109

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Observed Variables 2? sd
T4 CBCL Intemalizing 2.13 2.50
T4 CBCL Extemalizing 8.16 5.95
Tl SVAWS Mother DV 6.32 15.41
T2 SVAWS Mother DV 4.88 12.91
T3 SVAWS Mother DV 3.09 9.01
T4 SVAWS Mother DV 3.61 10.50
T1 SVAWS Child DV 2.38 8.98
T2 SVAWS Child DV 2.01 7.17
T3 SVAWS Child DV 1.19 4.32
T4 SVAWS Child DV 1.43 4.90
Child Self Regulation Score 12.54 2.79
Child Attentional Regulation Score 10.25 2.45
Child Attention to Toy Score 13.51 2.40
Child Low Anger Score 11.85 1.95
Child Low Sadness Score 11.12 2.59
Child Emotional (non) Lability Score 10.41 2.44
Maternal Parenting Positive Affect & Attention 2.62 0.48
Maternal Parenting Positive or Neutral Feedback 2.28 0.29
Mother Facilitates Self Regulation 3.95 0.65
Maternal Parenting Sensitivity 3.34 0.62
Maternal Parenting Positive Discipline 111.02 7.31
Family Income 2551.98 1986.82

 

 

 

 

 

Table 10: Meana and Standard Deviations of Observed Variables in SEM Model

Key to Table 11:

Intemalizing (1), Extemalizing (2), T1 DV Mother (3), T2 DV Mother (4), T3 DV
Mother (5), T4 DV Mother (6), T1 DV Child (7), T2 DV Child (8), T3 DV Child (9), T4
DV Child (10), Child Self Regulation (1 1), Child Attentional Regulation (12), Child Toy
or Self Play (13), Child Low Anger (14), Child Low Sad (15), Child Affective non-
Lability (16), Mother Positive Affect (17), Mother Positive Feedback (18), Mother
Facilitates Regulation (19), Mother Parenting Sensitivity (20), Mother Positive Discipline
(21), Family Income (22)

110

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 2 3 4 s 6 7 8 9 10

1 1.0

2 .62* 1.0

3 .04 .06 1.0

4 .01 .02 .28* 1.0

5 -.03 .06 .20* .23* 1.0

6 -.05 .00 .25* 32* .29* 1.0

7 -.07 .00 .48* .15* .14 .14 1.0

8 -.01 -.03 .33* .76* .29* .2s* .23* 1.0

9 -.04 .02 .19* .23* 84* .31* .1s* 39* 1.0

10 -.09 -.02 .21* .17* .14 .82* .10 .14 .16* 1.0
11 -.06 -.03 .04 -.05 .07 .02 .11 -.05 .04 -.06
12 .10 -.03 -.03 -.10 .07 -.06 .08 -.05 .02 -.11
13 -.07 -.07 -.02 -.07 .06 .00 .06 -.04 .02 -.05
14 .00 -.05 -.02 .01 .02 -.05 .08 -.06 -.04 -.15
15 -.09 -.06 -.02 -.10 -.12 -.03 .09 -.09 -.01 -.11
16 -.04 -.06 .01 -.02 .01 .00 .08 -.02 .02 -.12
17 -.20* -.17* -.09 -.19* .04 -.01 .03 -.18* .06 .08
18 -.16* 413 -.11 -.19* .01 -.13 .00 -.15 .04 -.11
19 -.07 -.09 -.16* -.27* -.07 -.24* .02 -.23* -.05 -.21
20 -.13 -.20* -.20* -.25* -.13 -.17* -.09 -.24* -.10 -.08
21 -.30* -.41* -.13 -.14 -.11 -.14 -.07 -.13 -.07 -.07
22 -.08 -.19* -.10 -.10 -.09 -.04 -.06 -.05 -.06 -.03

 

Table 11: Correlation Table of Observed Continuous Variables

lll

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11 12 13 14 15 16 17 18 19 20 21 22
1
2
3
4
5
6
7
8
9
10
11 1.0
12 .75* 1.0
13 .89* .72* 1.0
14 .75* .60* .68* 1.0
15 .84* .61* .78* 69* 1.0
16 .81* .60* .71* .80* .88* 1.0
17 -.10 -.03 -.07 -.08 .04 .03 1.0
18 .05 .04 .06 .08 .12 .12 .40* 1.0
19 .12 .15 .14 .16* .21* .14 .44* .53* 1.0
20 .05 .01 .06 .08 .14 .10 .55* .52* .83* 1.0
21 .05 -.07 .05 .00 .16* .09 .18* .20* .21* 30* 1.0
22 .01 -.04 .00 .06 .03 .07 -.02 .20* .12 .14 .24* 1.0

 

Table 11: (cont’d)

112

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CBCL Maternal DV Exposure

lntern- Extern- T1 DV T2 DV T3 DV T4 DV

alizing alizing Mother Mother Mother Mother
Avoidant .02 .08 .01 -.03 .00 -.08
Ambivalent -.05 .03 -.08 -.03 .11 .11
Secure .03 -.09 .05 .05 -.08 -.03

Child DV Witnessing

T1 DV T2 DV T3 DV T4 DV

Child Child Child Child
Avoidant .02 -.01 -.03 -.04
Ambivalent -.07 .01 .10 .13
Secure .04 .00 -.06 -.07

Child Positive Regulation

Self Attn. Toy Low Low non-

Reg. Reg. Play Anger Sad Labil.
Avoidant .05 .05 .07 .03 -.02 -.02
Ambivalent -.19* -.16 -.13 -.23* -.22* -.22*
Secure .12 .09 .05 .17* .20* .20*

Maternal Positive Parenting SES

P05. P05. Fac. Par. Pos. Family

Affect Feed. Reg. Sens. Disc Income
Avoidant .04 .03 .05 .07 -.02 -.09
Ambivalent .00 -.08 -.06 -.03 -.10 -.07
Secure -.03 .04 .01 -.03 .10 .13

 

Table 12: Point Biserial Correlation Table of Observed Continuous Variables and

Categorical Attachment Variables (dummy-codem

113

 

 

Measurement Models

Initially, Exploratory Factor Analyses (EFSs) were conducted using Principal
Component Analysis (PCA) with Varimax (orthogonal) rotation in SPSS to explore the
factor structures of the proposed latent variables. PCA analyzes all of the variance in the
observed variables. Scree plot solutions, an analysis of Eigenvalues, as well as theoretical
tenets and prior empirical ﬁndings were used to determine the most mathematically and
theoretically stable factor structure of each of the latent variables.

Subsequently, and prior to ﬁnal structural model testing, CFA’s were conducted
in M-plus, using F IML, to conﬁrm the factor structure of each latent variable. Residual
covariances were freed, as needed, when this lead to a signiﬁcantly better model ﬁt. All
CFA models were found to ﬁt the data well, as determined by a nonsigniﬁcant chi-square
value or a value that was less than two times the degrees of freedom, a root mean square
error (RMSEA) value of less than .05 or a RMSEA value falling within the 90%
conﬁdence interval (CI), a Tucker-Lewis index (TF1) and Comparative Fit index (CF I)
greater than .95.

Maternal Positive Parenting. The CFA for the 4-year old (T4) maternal parenting
measurement model consisted of 5 indicator variables. Four of the indicators were from
the observed parenting variables of the Parent Child Interaction Task (PCIT): Positive
Affect, Positive Feedback, Facilitation of Self-Regulation and Sensitivity. In addition, the
Discipline subscale of the Parent Behavior Checklist (PBC), was also an indicator of this
latent factor. The Discipline variable was standardized before inclusion in the model. All
variables were coded such that high levels of each construct indicated positive parenting

(e.g., high levels of the discipline variable indicated that the mother engaged in positive

114

discipline techniques). Model ﬁt was very good: X2 = 10.84, df=5, p=.06: CFI=.98;
TLI=.96; RMSEA=.08 (90% CI = .00 - .18), p=.42. All factor loadings were signiﬁcant.

Child Emotional Regulation. The CFA for the 4-year old (T4) child regulation
measurement model consisted of 6 indicator variables from the observed child regulation
coding of the Strange Situation Protocol (SSP): Self-Regulation/Organizational
Capacities, Attentional Regulation, Emotional Lability, Sadness, Anger and T oy/Self
Play. All variables were coded such that high levels of each construct indicated that the
child was emotionally and behaviorally regulated. The Emotional Lability, Sadness and
Anger variables were recoded such that high levels were indicative of successful
regulation (e.g., high levels of recoded anger indicated that the child did not demonstrate
anger during the session). Model ﬁt was very good: X2 = 11.52, dﬁ7, p=.l2: CF [=1 .0;
TLI=.99; RMSEA=.06 (90% CI = .00 - .13), p=.31. All factor loadings were signiﬁcant.

Cumulative Intimate Partner Violence — Mother Experience. The CFA for the
maternal experience of Domestic Violence (DV) measurement model consisted of 4
indicator variables from T1 through T4 (the ﬁrst 4 years of the child’s life) of the
Severity of Violence Against Women (SVAWS) measure. Thus, each indicator reﬂected
the woman’s scores from that year’s SVAWS administration. Model ﬁt was very good:
X2 = .56, dﬁ2, p=.76: CFI=.1.0; TLI=1.1; RMSEA=.00 (90% CI = .00 - .10), p=.84. All
factor loadings were signiﬁcant.

Cumulative Intimate Partner Violence — Child Exposure. The CFA for the child
exposure to DV measurement model consisted of the 4 indicator variables from T1
through T4 (the ﬁrst 4 years of the child’s life) of the mother’s endorsement of items on

the SVAWS that were directly observed by the child. Model ﬁt was very good: X’2 = .77,

115

df=2, p=.68: CFI=1.0; TLI=1.1; RMSEA=.00 (90% CI = .00 - .11),p=.79. All factor

loadings were signiﬁcant.

Initial Exploration of the Relations of Main Effects Variables

Prior to ﬁnal model testing, the attachment, child regulation and child
extemalizing/internalizing (CBCL) variables were explored with AN OVA methodology
using SPSS. A PCA factor score for the regulation variables was derived using the six
variables contained in the measurement model described above. As reported above, the
PCA revealed a single factor model. This factor had an Eigenvalue of 4.71, and explained
approximately 78% of the variance. Communalities ranged from .64 for Attentional
Regulation to .90 for Self-Regulation. The factor score was computed in SPSS for these
preliminary analyses using the regression method.

Infant Attachment Category and Child Regulation Score

Mow SSP (A/B/C): An examination of the relation of the 3 category SSP

variable and the child regulation score was conducted. The omnibus F was signiﬁcant for
child regulation [F (2, 144) = 3.63; p =.03]. A Bonferroni post-hoe analysis revealed that
children who had been securely attached as infants were signiﬁcantly better regulated
than children who had been ambivalently attached as infants (2? avoidant = .05, it secure =
.10, J? ambivalent = -.46). There were no signiﬁcant differences between the Avoidant

and Secure groups or between the Avoidant and Ambivalent groups. See ﬁgure 2.

116

 

 

Child Positive Regulation (mean) by Attachment Category

 

g 0.20 —
e
0
0'1
3
ii
“' 0.0 —
C
.2
H
(B
'5
DD
é’ 020—
g .
.5
o
a.
2
5
-0.40—

 

 

 

l i l
Avoidant Secure Ambivalent

Attachment Classification -- SSP

Figure 2: Child Positive Regulation (mean) by Attachment Category

Infant Attachment Category and Child Extemalizing and Intemalizing Scores

Mary SSP (A/B/C) a_n_c_l Extern_al_i_zﬂg: An examination of the relation of
the 3 category SSP variable and the CBCL externalizing score was conducted. The
omnibus F was not signiﬁcant for this model [F(2, 161) = .78; p = .46].

193mg! SSP (A_/B/C) and Intemalizing: An examination of the relation of the
3 category SSP variable and the CBCL internalizing score was conducted. The omnibus F

was not signiﬁcant for this model [F(2, 161) = .22; p = .81].

117

In summary, the 3 category SSP variable distinguished between the later child
regulation scores of the Secure and Ambivalent groups. However, it was not predictive of

later child externalizing or internalizing behaviors.

Hypotheses 1 — 5: SEM: Full Structural Model: Test of the ﬂoposed relations between
the predictors of infant attachment classiﬁcation, child self-regulation capacities. DV
exposure and parenting behaviors to the outcomes of child intemalizing__and externalizing
behaviors

The full structural model included 23 observed variable indicators and 4 latent
variables (see Figure 3). All analyses were conducted using the raw data, from which the
means and covariance matrix of the indicators were analyzed (Muthen & Muthen, 2007).
Consistent with the CFA analyses described above, all indicator variables for each of the
latent variables loaded signiﬁcantly onto that variable in the full model. Modiﬁcations
were made to the model based on theoretical considerations and model testing results. In
addition, due to the reported associations of income (Wadsworth & Achenbach, 2005)
and gender (Rescorla et al., 2007) with externalizing and internalizing behaviors these
variables were controlled in these analyses. The income variable was standardized prior
to inclusion in the full model. In the current longitudinal sample, the covariance coverage
of the data was very good, ranging from .73 to .93 for the observed variables in the
model. In addition, the covariance values were within the expected range.

The infant attachment variable (SSP) was coded using contrast codes that were
based, in part, on the results of the preliminary ANOVA analyses reported above. The

ﬁrst contrast variable examined the difference between the mean of the combined secure

118

and avoidant groups (coded at 1/3 each) and the ambivalent group (coded at -2/3). The
second contrast variable examined the difference between the secure (coded at 1/2) and
avoidant (coded at -1/2) groups.

The resultant structural model was found to ﬁt the data well (see ﬁgure 3).
Although it was signiﬁcant, X2 was well below two times the degrees of freedom of the
model (X2=295, df=240, p<.05). In addition RMSEA=.03 (90% CI = .02 - .05), p = .99,
CFI = .97, and TLI = .97. Four of the paths in the model were signiﬁcant:

1. The ﬁrst contrast code was signiﬁcant in the model, the second was not.
This indicated that, A) the combined group of children from the secure
and avoidant infant attachment classiﬁcations received higher emotion
regulation scores than children who were ambivalently attached; and B)
that there was no signiﬁcant difference between the secure and avoidant
groups;

2. Mothers who reported higher cumulative experiences of DV exposure
tended to receive lower positive parenting scores with their four-year-old
children;

3. Mothers who exhibited lower levels of positive parenting had children
who received higher externalizing scores on the CBCL;

4. Children from lower income families tended to receive higher scores on

the externalizing subscale of the CBCL.

119

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

+Aff. 0.27a 7 7
+ F8 16, 4year old gender l 555
F R ' Maternal _1_40¢ 1.46 -2_o3*
' ' 056‘ Parenting
Sens. 059* 3&6 4 ear old
6'78" 033* 03" lChild
7 701: . DlSC. . '0.38* Ext
DV2 ~ Birth — Age 4 -0.13 ern.
443* W 0.07
ms MOTHER _0 75
DV4 455* EXPERIENCE 4year old
Child
"' R l t'
1year old contra“ 1; 0.50 , egu a non
Infant Com-Vast 23“ .03 _0-14 -0.40
Attach. O 4
Re 2 72a ' 4' 0-12 4vearold
3' ' DV1 2.59a Child
Atn. 1.87 DV2 511,. Birth _ Age 4 -03 Intern.
TOY 2.17* DV
-1.49* DV3 230* CHILD {0.48 {-0.16
2.20* ' WITNESS l 1 ‘ 1
Sad Lab DV4 111* Gender 1,5155
- 139*
a o I
Fixed Estimate;

’p < .05. All factor loadings and path coefficients are standardized values. Signiﬁcant
paths are represented by bold black lines.

Figme 3: Full Structural Model

120

Hypothesis 6: Testing the hypothesized moderatig role of DV in the relation between
infant attachment catggory and preschool symptoms of externalizing and intemal_iz£1g
behaviors.
Hierarchical linear regression methodology was utilized to test this hypothesis. Two
separate regressions were conducted; one for the internalizing and one for the
externalizing outcome variables. DV was dichotomized using a median split. In addition,
two (e.g., k-I) dummy variables for the infant SSP (attachment category) data were
created, using the secure group as the reference group. Interaction terms were created by
multiplying the dichotomized DV variable with each of the dummy variables. The model
included the ﬁrst attachment dummy variable (avoidant = 1), the second attachment
dummy variable (ambivalent = 1), the dichotomized DV variable, followed by the ﬁrst,
and then the second, interaction variables.

The regression model for externalizing behaviors was signiﬁcant overall (F
(5,158) = 3.04, p < .05). However, the only signiﬁcant variable in the model was the
dichotomous DV variable (B= .21; p<.05). The regression model for internalizing

behaviors was not signiﬁcant (F (5,158) = 0.40, ns).

Hymthesis 7: Testing the proposed relation of infant attachment category to later child
use of speciﬁc regulatory behaviors (e. g., self-oriented versus other—oriented)
Principal Component Analysis

A Principal Component Analyses (PCA) with Varimax (orthogonal) rotation was
conducted in SPSS, version 15.0 (SPSS, 2006), to explore the factor structure of the

regulatory behavioral codes: Self Stimulation, Talk/Sing to Self Attention to Toy/Self

121

 

Play, Contact Seeking and Attempts to Leave Room. Scree plot solutions, an analysis of
Eigenvalues, as well as theoretical tenets and prior empirical ﬁndings were used to
determine the most mathematically and theoretically stable factor structure.

The resultant model suggested 2 factors: Factor 1 included the Contact Seeking
and Attempts to Leave Room variables (Eigenvalue = 1.56) and was labeled the chgr;
Focused Regglation factor; Factor 2 included the Self Stimulation and Talk/Sing to Self
variables (Eigenvalue = 1.27) and was labeled Self-Focused Regnlation. Attention to
Toy/Self Play did not load well onto either of these factors. Together, these factors
accounted for approximately 70% of the variance. Communalities ranged from .52 to .87.
The factor scores were computed in SPSS using the regression method.

ANO VA: Analyses of Infant Attachment Category and Self and Other Factor Scores

The Self-Focused and Other-Focused regulatory behavior factors of the 4-year old
children were explored in relation to their prior infant attachment category using
ANOVA methodology in SPSS.

Other-Focused Regulation Factor. An examination of the relation of the 3
category SSP variable and the Other-Focused Regulation factor score was conducted. The
Omnibus F was signiﬁcant [F(2, 144) = 7.26; p = .001]. A Bonferroni post-hoe analysis
revealed that children who had been ambivalently attached as infants had signiﬁcantly
higher Other-Focused regulation behaviors than children who had been securely attached
as infants (3': ambivalent = .60, f secure = -.19, i avoidant = .02). Although there was not
a signiﬁcant difference between the Avoidant and Ambivalent groups, there was a trend

in the data in the expected direction (p=.078; Ambivalent higher on Other-Focus than

Avoidant). See ﬁgure 4.

122

Other-Focused Behaviors (mean) by Attachment Classiﬁcation

 

0.60 —

0.40 ‘

0.20 —

0.00 '—

Mean of Other-Focused Factor Score

-o.2o‘

 

 

 

l l l
Avoidant Secure Ambivalent

Attachment Classification -- SSP

Figme 4: Other-Focused Behaviors (mean) by Attachment Classiﬁcation

Self-Focused Regulation Factor. An examination of the relation of the 3 category
SSP variable and the Self-Focused Regulation factor score was conducted. The omnibus
F was not signiﬁcant [F(2, 144) = .70;p = .50].

Ip_§_ur_n_, when examined at the 3 category level, children with prior ambivalent
infant attachment classiﬁcations tended to demonstrate signiﬁcantly higher levels of
Other-Focused regulatory behaviors compared with the secure group during separations

from their mother in a laboratory environment. In addition, there was a trend suggesting

123

that the Ambivalent group may also evince higher Other-Focused behaviors compared
with the Avoidant group. The model for Self-Focused regulatory behaviors was not
signiﬁcant, suggesting that there are no differences between the attachment groups
relative to their use of Self-Focused regulatory behaviors.

Post-hoc Agalyses

There is currently debate within the literature about whether or not parental report
of child witnessing of DV truly captures the extent to which DV in the home inﬂuences
child outcomes. For instance, in a recent meta-analysis of child DV exposure (n=118
studies), Kitzmann, and colleagues (Kitzmann, Gaylord, Holt, & Kenny, 2003) identiﬁed
the subset of studies which measured direct assessment of children’s witnessing (usually
by maternal report) of the DV versus those that used the broader deﬁnition of exposure.
Results demonstrated that witnessing the violence did not moderate the relationship
between DV presence in the home and later child social-emotional outcomes. They
interpret these results as suggesting that violence perpetrated within the home
environment of children results in child exposure.

Consequently, post-hoe analyses in the current study were undertaken to examine
the possible differential effects of maternal experience of DV versus maternal report of
child witnessing of DV on the outcome variables. The results of these analyses indicated
that replacing the child witnessing DV variable with the maternal experience of DV

variable did not improve model ﬁt.

124

 

CHAPTER 8
DISCUSSION
The current study examined the inﬂuence of infant attachment on later child self-

regulation capacities and psychopathology in a sample of children and their mothers who
were heterogeneous for risk relative to domestic violence (DV) status and SES. The
hypothesized mediating role of child regulatory capacities on the relationship between
infant attachment and child psychopathology (e. g., internalizing and externalizing) was
also examined. The SEM model ﬁt the data well; three of the hypothesized paths and one
of the covariates (SES) were signiﬁcant. The hypothesized moderating effect of DV
exposure on the relationship between infant attachment category and later symptoms of
internalizing and externalizing behaviors in preschool was also examined using
regression methodology and, overall, did not ﬁt the data well. In addition, the inﬂuence
of infant attachment on the differential expression of speciﬁc regulatory behaviors (e.g.,
self- versus other-focused) in preschool-aged children was examined using ANOVA
methodology, yielding signiﬁcant effects with regard to other-focused regulation
behaviors.

Results from these analyses are discussed in terms of three main groups of
ﬁndings. First, results suggested that matemal-infant attachment classiﬁcation was
related in theoretically consistent ways to overall child regulatory capacities as well as to
the speciﬁc types of regulatory behaviors that the children used (e. g., self- vs. other-
focused), In contrast, and contrary to predictions, neither infant attachment nor
concurrently-assessed child self-regulation capacities were found to inﬂuence child

externalizing and internalizing behaviors and DV was not a moderator of this

125

 

relationship. Secondly, the inﬂuence of DV on the mother’s parenting, the child’s self-
regulation capacities and behavioral outcomes were tested. Results suggested that
cumulative maternal exposure to DV had a negative inﬂuence on positive parenting
behaviors. However, cumulative child DV exposure did not inﬂuence child self-
regulation capacities or behavioral symptoms of internalizing or externalizing behaviors
in this sample. Thirdly, proximal maternal parenting inﬂuenced child externalizing
behaviors but not internalizing behaviors or child self-regulatory capacities.
The Inﬂuence of Infant Attachment on Preschool Self-Regulation and Psychopathology
One of the primary tenets of attachment theory as outlined by Bowlby (1969/1982)
is that the infant’s emotional system is shaped in fundamental ways by the developing
parent-infant relationship. The internalization of this early relationship is thought to
become generalized over the course of development such that it becomes a stable
personality characteristic of the individual. This process becomes consolidated during the
preschool period of development at which point it can be understood as the child’s
internalized style of emotional self-regulation (Kopp, 1989). However, there have been
very few empirical examinations of the relationship between early attachment and child
regulatory capacities, and the existing studies have examined discrete emotions such as
anger regulation, generally at the dichotomous (secure/insecure) level of analysis (e. g.,
Gilliom et al., 2002). The current study, therefore, contributes to our understanding of the
longitudinal effects of infant attachment on the development of emotional self-regulation
and behavioral outcomes across early child development by examining this relationship

at the typological level using a measure of child emotional self-regulation that is

126

 

comprised of multiple indicators of regulation including attentional and emotional
regulation, expression of negative emotions and emotional lability.

The ﬁndings reported here suggested that the child’s emotional self-regulation
capacities, as measured within a stress-inducing laboratory paradigm, are related to his or
her matemal-infant attachment relationship. Speciﬁcally, compared to children with
secure attachment histories, children who were ambivalently attached as infants were less
well-regulated as preschoolers, and demonstrated increased use of other-directed forms of
self-regulation (e. g., contact seeking). In infancy, ambivalently attached children can
generally be distinguished from the other two organized attachment groups, in part, by
their demonstration of high levels of dysregulation and negative affect expressed within
the context of their primary attachment relationship (Ainsworth et al., 1978; Magai,
1999). These children continuously seek contact with their caretakers but are rarely able
to be effectively soothed by them. Mothers of ambivalently attached infants tend to be
inconsistent in their provision of sensitive caretaking, thus providing a variable
reinforcement schedule of the infant’s dysregulated (upset) emotional expressions that
serves to strengthen the intensity and frequency of these expressions.

In early infancy the use of other-seeking behaviors (e. g., crying, fussing), is
adaptive in that it generally serves to elicit the attention of the caretaker. However, as
children mature, social expectations for emotional and behavioral self-regulation
increase. In the current sample, children with secure attachment histories appeared to

have met these developmental expectations. However, children with ambivalent
attachment histories were both dysregulated and overly reliant on developmentally

regressed behavioral regulation strategies. Given that the behavior of these children is

127

incongruent with social expectations, it is likely that their emotional self-regulation
strategies (e.g., crying, whining, clinging to adults), will elicit feelings of annoyance and
anger on the part of their adult caretakers (e.g., preschool teachers, daycare providers,
etc.). It is furthermore possible that, if these interpersonal transactions continue over time,
these children may be at increased risk for negative psychosocial outcomes in the future.
In addition to their theoretical consistency, the present results are also broadly
consistent with the few empirical studies that have examined the inﬂuence of early
attachment on child regulatory capacities in preschool. For example, using a dichotomous
attachment variable (secure/insecure), Gilliom and colleagues (Gilliom et al., 2002)
found that, compared with the secure group, children with insecure attachment histories
tended to demonstrate other-directed self-regulatory behaviors in the context of a
frustrating waiting task. In comparison, children with secure attachment histories tended
to use more sophisticated self-regulatory skills such as self distraction (e. g., playing with
a toy). The current study adds to these ﬁndings by examining these behaviors at the three-
category level of attachment. The results reported here suggest that there are differences
in the use of other-focused self-regulatory behaviors between the secure and ambivalent
groups, but not between the secure and avoidant groups. In addition, a trend in the data
suggested that there may also be differences between the avoidant and ambivalent groups
along this dimension such that the ambivalent group evinced higher rates of other-
focused behaviors. Thus, when analyzed at the three category level, it appears that the
secure and avoidant groups may use signiﬁcantly fewer other-focused emotional self-
regulation behaviors compared with the ambivalent group. This ﬁnding stresses the

importance of analyzing these data at the typological, versus the dichotomous level.

 

128

Further, it suggests the possibility that, in the case of the data reported by Gilliom and
colleagues (Gilliom et al., 2002), the ﬁnding that the insecure group demonstrated higher
rates of other-focused behaviors may have been driven by the variance accounted for by
the ambivalent group.

With regard to the group of children who were securely attached as infants, the
current ﬁnding that they demonstrated positive regulation capacities is also broadly
consistent with other studies that have examined these constructs. For example, using a
dichotomous attachment variable (secure/insecure), Sroufe and colleagues (Sroufe et al.,
2005) found that children who were securely attached as infants remained emotionally
regulated even in the face of an extremely challenging and frustrating task, whereas the
children with insecure attachment histories did not. In addition, in comparison with the
insecure group, the children with secure attachment histories were better able to adapt

their emotional and behavioral reactions in a ﬂexible and more sophisticated manner
when faced with environmental challenges. That is, their ego resiliency, as deﬁned by
Block and Block (1980), was found to be more sophisticated compared to children with
insecme attachment histories. Furthermore, the current study found that the variable that
captured the child’s propensity to engage in play behaviors (e. g., “toy or self play”)
during the stressful episode loaded onto the positive child emotional self-regulation factor
and not, as was originally hypothesized, onto the self-focused behavioral factor. This is
consistent with the results reported by Gilliom and colleagues (Gilliom et al., 2002), in
that they found that children with secure attachment histories tended to engage in
incl‘eased rates of toy play in a (usually successful) attempt to cope with a frustrating

waiting task. Thus, it appears that children with secure attachment histories are able to

129

 

more ﬂexibly adapt their emotional self-regulatory strategies in a manner that allows
them to remain emotionally regulated, whether that involves playing with the available
toys or continuing to work to solve a difﬁcult task. This ﬂexibility presumably allows
them to more effectively use the resources that are available in a given environment or
situation in order to remain regulated. In the context of changing environments that
provide different kinds of resources (e. g., home, preschool, social situations), their ability
to ﬂexibly adapt their emotional self-regulatory strategies may provide them with a
distinct advantage in overall psycho-social adaptation and development in comparison
with the ambivalent group.

Despite differences in emotional self-regulation capacities between the secure and
ambivalent groups in the present study, no differences were evident between the secure
and avoidant or the ambivalent and avoidant groups. Furthermore, contrary to the original
hYPOtl‘lesis, ANOVA analysis revealed that the children with secure attachment histories

did not demonstrate a mid level of regulatory capacities in the laboratory assessment. In
fact, Within the full SEM analysis, no differences were evident between the secure and
aWidant groups in terms of their self regulatory capacities and the combined mean of
these groups was found to be signiﬁcantly different from that of the ambivalent group,
suggesting that both the secure and avoidant groups of children are capable of relatively
etifective emotional self-regulation in stressful situations for short periods of time (e.g., 3
minute segments). This suggests that the development of either an avoidant or a secure
attachment relationship during infancy may result in the capacity for emotional self-
regulation during the preschool period. This is consistent with prior research that has

demonstrated poorer psychosocial outcomes for the ambivalent group compared with

130

 

m

it

both the avoidant and secure groups (for a review see, Greenberg, 1999). It also suggests
that children with secure attachment histories do not demonstrate a mid-range level of
regulatory capacities but, instead, are able to maintain high levels of emotional self-
regulation during short separations from their parent. Children at this age are often
expected to separate from their parents and function independently within larger social
contexts such as preschool environments. Usually, however, this expectation is applied to
environments that are familiar to the child. Thus, a child may cry when separating from
their parent the ﬁrst day or two of preschool but, over time, the child is expected to
acclimate to the new environment. The current study hypothesized that separation for a
short period of time in an unfamiliar environment, would elicit a mid-range level of
reglllatory capacities in preschool aged children with secure attachment histories. These
results suggest that this was not the case and that, at least during short separations, these

Children are capable of high levels of emotional self-regulation in unfamiliar

enVil‘O nments.

This ﬁnding calls into question basic tenets of attachment theory which argue that
diffe'l‘eznces in behavioral regulation capacities are inherently linked to differential styles
0f attachment relationships in infancy (Bowlby, 1969/1982). That is, the ﬁnding that both
Avoidantly and Securely attached infants ultimately develop adaptive regulatory
capacities later in development suggests an alternative hypothesis of the etiology of these
behaViors. Speciﬁcally, as early investigators of child temperament have contended, it is
POSSible that the infant behaviors which are observed in the SSP may be indicators of
infant temperamental qualities and not linked, necessarily, to the mother-infant

attachment relationship (Belsky & Rovine, 1987). Future research that tracks both

131

 

temperamental characteristics and indicators of the mother-child relationship at regular
intervals over the course of early development is necessary to fully explore this

hypothesis.

Alternative explanations are also possible. Speciﬁcally, the ﬁnding that children

with avoidant attachment histories were well regulated is theoretically consistent.
However, although it was not measured in this study, prior investigations of the
physiological functioning and regulation of infants in the SSP has demonstrated that
infants who are avoidantly attached do, in fact, evince elevated biological signs of stress
during the separation episodes (e. g., vagal tone, cortisol levels), even though they appear
behaviorally to be calm and well-regulated (e.g., Hill-Soderlund et al., 2008). Thus, it is
Possible that there are physiological differences across the secure and avoidant groups in
terms of their emotional self-regulation in preschool even though they appear,
behaviorally, to be quite similar. Future research is necessary to explore this possibility.
Although there were theoretically consistent ﬁndings in the relationship between
infant attachment and later child emotional self-regulation capacities, neither of these
variables predicted symptoms of psychopathology (e. g., internalizing and externalizing
bellaviors) in preschool children in the current study. Thus, these data suggest that early
attaJcliment experiences that are presumed to be based on the relationship(s) with the
Primary caretaker(s), do not inﬂuence psychopathology in later development. This
ﬁnding is important given the existing evidence that temperamental qualities in children
have been shown to be associated with symptoms of psychopathology (N igg, 2006).
ThEBI‘efore, whether or not infant behaviors within the SSP are indicative of relationship

qualities or of temperamental characteristics, the current data suggest that the attachment

132

 

categories that are derived from these behaviors are not predictive of later child
externalizing and internalizing behaviors in the same ways that child temperamental
qualities are.

However, developmental psychopathology holds that it is the conﬂuence of risk

and protective factors that best predicts to social-emotional outcomes such as symptoms
of psychopathology (Sameroff et al., 2000). Consistent with this framework, infant
attachment (as well as infant temperament) is understood as one factor within a child’s
broader psychosocial context that may inﬂuence later development. That is, while
attacbltlent theory predicts the internalization of differential styles of emotional self-
l‘egulation based on early attachment experiences, these are understood as affecting later
development within a normative range; there is no theoretical prediction from insecure
attachnmnt history to later psychopathology in low-risk groups. However, when other
Psychosocial risk factors are present, insecure attachment styles have been shown to
predict to later psychopathological outcomes. For example, Shaw and colleagues (Shaw
et al- 3 1997; Shaw et al., 1996) found that dichotomized infant attachment variables
(secure/insecure, organized/disorganized) were associated with both externalizing and
interIlalizing behaviors in a sample of preschool-aged children whose families were low-
income and receiving social services.

Consistent with this framework, the current study predicted that DV exposure
would moderate the relationship between infant attachment and later symptoms of
Psychopathology. Contrary to this prediction, however, results demonstrated that DV did
not moderate this relationship. Thus, the current results are inconsistent with these

ﬁndings and with theoretical predictions that, within high risk environments, insecure

133

attachment will be a signiﬁcant risk factor for poor psychosocial outcomes. This ﬁnding
lends support to the alternative hypothesis that the SSP measure of infant attachment
categories does not measure internalized relationship strategies, but, instead, captures
infant temperamental qualities that manifest early in development and are reﬂective of
the normal range of temperamental qualities in humans from reserved and organized to
outgoing and gregarious. Furthermore, from a theoretical standpoint, attachment theory
argues that DV exposure is likely to have a powerful inﬂuence on the early social-
emotional development of the infant and young child, and on the formation of early
infant relationships. That is, DV is expected to be a particularly inﬂuential risk factor
during early childhood. Thus, the ﬁnding that DV does not moderate this relationship
suggests that attachment relationships may not be associated with later psychopathology,
even in the context of other contextual risk factors.

An alternative explanation, however, is related to the use of a community sample
in the present study. Speciﬁcally, the lack of association between infant attachment
history and child symptoms of psychopathology (e. g., internalizing and externalizing
behaviors), in this heterogeneous-for-risk sample, may also suggest that exposure to DV
Within community samples may not constitute sufﬁcient psychosocial risk such that
insecure attachment leads to symptoms of psychopathology in young children. This

ﬁ“ding is partially consistent with other research examining the inﬂuence of DV on child
onto()mes in community samples. For example, using a community sample, Levendosky
and Colleagues (Levendosky et al., 2003) found increases in externalizing, but not
internalizing behaviors in DV-exposed preschool aged children. In addition, in a second

Study by the same researchers, DV exposure in a community sample was not associated

134

trimmers:
lrlrer. u
trim
are com
marred
ill moi
samples

herpsy

pi'Sitir

With increases in either internalizing or externalizing behaviors (Levendosky et al., 2002).
Further, using a sample of DV-exposed families, Lieberman and colleagues (Lieberman
et al., 2005) found that, when other variables such as mother-child relationship quality
were controlled, child DV exposure failed to predict to child behavior problems as
measured on the CBCL. Thus, the current study, which includes maternal parenting in the
ﬁll] model, lends support to the hypothesis that the levels of DV found in community
sainples are not sufﬁcient to place children with insecure early attachments at risk for
later psychopathology.

A further alternative explanation of these results is that the lack of association
between infant attachment and later child psychopathology within the current sample
may be due to methodological and contextual factors that are somewhat idiosyncratic to
DV-BXposed samples. Speciﬁcally, the variance and range of both the internalizing and

extﬁl‘llalizing subscales of the matemally-rated CBCL in this study were relatively

reStl‘icted and positively skewed. Further, none of the internalizing scores and only three

(’me externalizing scores within this dataset fell within the clinical range (t > 70). These
r“ults are consistent with the work by Lieberman and colleagues (Lieberman et al.,
2005) who found that the mean CBCL total problem score in their community sample of
DV‘exposed women and children fell below the clinical cut-off. Thus, it may be that
“1336 scores represent a pattern of maternal underreporting. There are two potential
teas0ns for this possibility. First, the negative psychosocial effects of DV exposure on
maternal parenting have been demonstrated in other samples (McCloskey et al., 1995),
and are evident in this sample as well. That is, as DV exposure increases, maternal

Positive parenting decreases. Positive parenting, as deﬁned here and elsewhere (Dix,

135

1991), requires the capacity to be attentive to the behaviors of one’s child. Trauma theory
predicts that traumatic exposure leads to cognitive and emotional preoccupation on the
part of the victim, making her less able to function effectively within other social
domains (Herman, 1992). Thus, it is possible that the low CBCL scores reported by the
mothers in this study represent a rater bias that is related, in part, to decreases in the
ability of these mothers to fully attend to and report on psychopathological symptoms in
their children. Secondly, feelings of guilt and shame on the part of the victim are
associated with DV exposure (DeVoe & Smith, 2002). For example, when asked about
how DV exposure inﬂuenced her relationship with her children, one mother in the current
study said, “When he [the assailant] was around and we were ﬁghting, I was stressed out.
And then I wouldn’t take it out on them [her children]. . . but I was short with them and I
could tell and then that would make me feel bad because it wasn’t their fault. It was my
fault, I felt, like from letting him be here. And then I would feel bad, I’m like god, you
know, it’s not their fault. I should get them out of here.” Consistent with this woman’s
self. report, prior qualitative work has documented that some women experience feelings
0f guilt and shame about their belief that they have intentionally placed their children in
hal‘l‘l'l’s way (DeVoe & Smith, 2002). Thus, there may be a form of cognitive dissonance
that is interfering with the ability of these mothers to accurately report on the symptoms
0f DSychopathology exhibited by their children. That is, the idea that their children are
exhibiting signs of psychopathology may be extremely difﬁcult for these mothers to

attend to and report due to their feelings of shame and guilt about “causing” their

Children’s problems.

136

In addition to the lack of an association between infant attachment and later child
internalizing and externalizing behaviors, the current study also failed to ﬁnd a
signiﬁcant association between the child’s observer-rated emotional self-regulation

capacities and concurrently-assessed maternal ratings of externalizing and internalizing
behaviors. This ﬁnding suggests that evaluations of the child’s emotional self-regulation
capacities as measured in an unfamiliar laboratory environment are not related to
behavioral symptoms of internalizing and externalizing behaviors as they manifest in the
child’s daily living environment. This suggests that other factors are inﬂuential in the
development of behavioral symptoms of psychOpathology and that a child’s observed
reEflllatory behaviors are not inﬂuential. Social learning theory, for example, predicts that
the Child’s behaviors are inﬂuenced by observations of behaviors of others within their
enViI‘onment. In fact, the current study found that maternal parenting behaviors were
Predictive of the child’s externalizing behaviors, lending support to this hypothesis. Other
faCtOrs that were not measured in this study are also likely to be inﬂuential in the
development of behavioral symptoms of psychopathology such as the genetic makeup of
the c>hild. Given these results, future research should examine the inﬂuence of parenting,
fatni1y and child characteristics that may be inﬂuencing the development of these
Syu'lptoms.

Alternatively, it is possible that the caretakers of the children with ambivalent
attaChment histories have adapted to their dysregulated style and are able to
acc0mmodate their need for extra support. For example, parents and teachers may have
adapted to the child’s deﬁcits in emotional self-regulation by remaining physically close

to the child more often than they would for other children. These may be the children

137

who hold the teacher’s hand at recess, are allowed to always sit in the front of the group

 

during ‘story time’ at school and who avoid playing with other neighborhood children
unless a parent or caretaker is physically present. If the child’s caretakers are able to
adapt to these kinds of behaviors, the needs of these children may be sufﬁciently met
within their social environment such that symptoms of psychopathology do not develop
despite their deﬁcits in emotional self-regulation. This study is the ﬁrst to examine this
relationship and, as such, contributes to our understanding of this relationship in
preschool-aged children.

An alternative explanation of the lack of a signiﬁcant association between the
child’s emotional self-regulation and symptoms of psychopathology may be that the
concurrent assessment accounts for the lack of signiﬁcant ﬁndings in this relationship.
Speciﬁcally, from a transactional perspective (Sameroff & Fiese, 2000), it is the
interaction of child, parent, and contextual factors across time that best predicts to child
outcomes. As argued here, social expectations for emotional self-regulation become
primary during the preschool period of development. It is at this point that there is likely
to be a shift from tolerance of the child’s dysregulation on the part of the child’s adult
caretakers to intolerance and anger. For example, “temper tantrums” that may be partially
tolerated during the toddler period of development, are unlikely to be tolerated during the
preschool period. In this study emotional self-regulation was measured at the point in
development when the child’s capacity for self-regulation is beginning to be more fully
expected. Thus, it may be that the negative consequences of emotional dysregulation will

only manifest later in development as a consequence of repeated negative interactions

with adults in the child’s life.

138

 

Domestic Violence: Inﬂuences on Parenting and Child Outcomes

Results from the current study demonstrated that increases in maternal exposure to
DV over the ﬁrst four years of the child’s life resulted in decreases in observer-rated
levels of positive maternal parenting measured when children were four years of age.
This ﬁnding is consistent with trauma theory which argues that traumatic exposure
inﬂuences the psychosocial functioning of women in ways that affect her functioning
across many life domains, including parenting (Herman, 1992; Levendosky & Graham-
Bermann, 2001b). Speciﬁcally, trauma theory predicts that exposure to DV will result in
a state of both emotional and cognitive hypervigilance on the part of the victim (Herman,
1992). Consistent with this theory, prior empirical research has demonstrated the negative
inﬂuence of DV on parenting behaviors (Bogat et al., 2006; DeVoe & Smith, 2002;
Levendosky, Lynch, & Graham-Bermann, 2000). Parenting may be impaired due to the
cognitive and emotional preoccupation that occurs in the context of traumatic
hypervigilance and this may decrease a woman’s ability to ﬂexibly adapt her parenting to
meet the changing needs of her children.

Speciﬁcally, exposure to DV may inhibit a mother’s ability to regulate her own
affects and behaviors which may then compromise her ability to engage in well-regulated
interactions with her child (Fonagy et al., 2004). Schore (2003) has argued that the
parent’s ability to regulate her own emotions is especially important with respect to the
parenting of young children given the centrality of the mother-child relationship in early
child development. He stated, “Infant research now suggests that the baby becomes

attached to the modulating caregiver who expands opportunities for positive affect and

139

‘-"“r In“ “'5
ll

 

 

minimizes negative affect” (p. 8). Consistent with this view, Dix’ (1991) empirical
review and formulation argued that sensitive parenting across child development is
dependent on the parent’s ability to provide an affectively organized response to the
child’s behaviors. He argued that parents who are able to regulate their emotions are
better able to attend to the child’s behaviors and then modify their own behaviors to meet
the physical and emotional needs of the child. More recently, the growing
neurobiological literature has shed light on the ways in which early, regulated, parent-
child interactions play a critical role in the neurological and psychosocial development of
young children (Fonagy et al., 2004). DV exposure may inhibit a mother’s ability to
provide this kind of sensitive, modulating care to her child. The majority of empirical
research in this area, however, has examined the inﬂuence of DV on parenting in school-
aged groups of children. The current study sheds light on the relationship between DV
and parenting during early childhood.

Prior work using data from the current longitudinal study has demonstrated both
that DV exposure predicts to a mother’s internal representation of her child (Huth Bocks,
Levendosky, Theran, & Bogat, 2004) and, furthermore, that both pre- and post-natal
representations predict to the mother’s parenting behaviors with that child (Dayton,
Levendosky, Davidson, & Bogat, in press; Theran, Levendosky, Bogat, & Huth-Bocks,
2005). Speciﬁcally, Huth-Bocks and colleagues found that exposure to DV during
pregnancy was associated with the prenatal balanced/non-balanced status of a woman’s
internal working model of her child (Huth Bocks et al., 2004); women who were
experiencing DV during pregnancy were more likely to develop non-balanced prenatal

representations of their infants such that their representations were either emotionally

140

 

 

dysregulated (e. g., distorted) or emotionally distanced (e. g., disengaged). Subsequently,
Dayton and colleagues (Dayton et al., in press.) found that prenatal maternal
representations inﬂuenced later parenting when children were one year of age such that
mothers who held balanced (secure) representations of their children were more likely to
engage in positive parenting behaviors when compared to the two insecure groups.
Additionally, mothers holding distorted representations were more likely to engage in
hostile behaviors with their children, and mothers holding disengaged representations
were more controlling with their children. The current study extends the prior work
regarding the impact of DV on parenting behaviors by demonstrating that post-natal DV
across early development is also inﬂuential in a mother’s ability to effectively parent her
child.

In contrast to the signiﬁcant effects of DV exposure on maternal parenting
behaviors, the hypothesized relationships between DV and child emotional self-
regulation and psychopathology were not signiﬁcant. This result is inconsistent with both
the theoretical and empirical trauma literature (Lieberman et al., 2005; Osofsky, 2004).
Especially relevant to the present investigation are two studies that used subsamples of
the current longitudinal study. Speciﬁcally, DeJonghe and colleagues (DeJonghe et al.,
2005) found that one-year old infants who were exposed to DV in the home were more
distressed by a simulated telephone argument in a laboratory setting than non-exposed
infants. Secondly, Bogat and colleagues (Bogat et al., 2006) found that maternal self
reports of their own DV-related trauma symptoms predicted maternal report of infant

trauma symptoms for infants who had witnessed severe levels of violence.

141

 

 

Thus, results from the current study are inconsistent with other work drawing
from the same longitudinal sample. Current results suggest that DV exposure across early
development may be less inﬂuential on later child outcomes than they are on outcomes in
infancy. This ﬁnding suggests that children are more resilient to the effects of DV
exposure than was previously thought. It may be that, although they demonstrate early
signs of negative reactions to DV exposure, children are able to adapt to their
environments effectively over the course of time. In fact, more recent published data
from the longitudinal sample have demonstrated that both risk and resiliency factors are
inﬂuential in accounting for the relationship between child DV exposure and
psychosocial outcomes in early childhood including internalizing and externalizing
behaviors (Martinez-Torteya, Bogat, Von Eye, & Levendosky, in press). Further, results
of these analyses demonstrated that cumulative exposure to DV was not a signiﬁcant
predictor of child psychosocial outcomes. Instead, they found that chronic exposure to
abuse, in contrast to intermittent exposure, best predicted outcomes. Thus, the current
ﬁndings are consistent with more recent work on the longitudinal study that failed to ﬁnd
a linear (e. g., dose-response) relationship between DV exposure and child psychosocial
outcomes.

One alternative explanation of this ﬁnding is that DV may inﬂuence child
outcomes primarily through the indirect effects of the abuse on the mother’s parenting
behaviors. Consistent with the current results, using an observer-rated measure of
parenting behaviors, Levendosky and colleagues (Levendosky et al., 2003), found an
indirect, but not a mediating, relationship wherein DV exposure was negatively related to

positive parenting and positive parenting was negatively associated with the externalizing

142

 

 

behaviors of preschool-aged children. Consistent with these ﬁndings, Katz and
Windecker-Nelson (2006) found that emotion coaching, a particular subskill of parenting
that is thought to promote children’s emotion regulation capacities, moderated the effect
of DV exposure on child aggression in a preschool sample (n = 130). However, other
studies that have examined these constructs in older children have failed to ﬁnd these
indirect effects (McCloskey et al., 1995).

The disparity among studies of preschool-aged versus school-aged children may
be due to the differential effects of parenting on younger (birth to preschool) versus older
(school-age through adolescence) children. That is, as argued by attachment theory,
parenting during early child development establishes a psychological foundation within
the child through the creation of internal representations. During early development, the
child’s basic understanding of his or her social world is ﬁltered through the parent in
important ways that, in healthy development, tend to decrease over time. Infants and
toddlers, for example, engage in the process of social referencing whereby, when faced
with an unknown stimulus, they look to their caretaker in an effort to gauge the emotional
salience of the event (Stern, 1985). If they perceive danger or anger on the part of the
adult, they are likely to become upset; if they read the parent’s face as reﬂecting humor
or warmth, they are likely to remain calm or even engage more fully with the stimulus.
Thus, it may be that, when mothers are able to engage in positive parenting with their
young children they effectively shield them from some of the possible trauma-related
consequences of exposure to the abuse. This process is not likely to occur as intensively

for older children. In fact, as children mature, they may be more likely to attempt to

143

 

in
pr;

ch

is

(I!
a—-

CI

intervene to protect the mother, thereby placing themselves even more at risk for
experiencing the direct effects of the violence.

This hypothesis may explain the difference between the current ﬁndings and
ﬁndings from the two prior studies drawn from the present longitudinal sample. In the
prior studies, DV exposure was found to inﬂuence child outcomes in one-year-old
children. Speciﬁcally, in the ﬁrst study, DeJonghe and colleagues (DeJonghe et al., 2005)
examined the infant’s reactions to an angry phone call when the mother was not in the
room. Thus, the infant was not able to visually reference the mother in order to gauge her
reaction to the event. These investigators found that DV exposed infants were more
reactive to the simulated angry event than non-exposed infants. It is possible, however,
that these results may have been different had the mother been present in the room. That
is, if the mothers were present, they would have likely reﬂected a sense of safety to the
children about the simulated anger being expressed and this may have resulted in the
children not reacting to the event. The second study by Bogat and colleagues also lends
support to the hypothesis that DV inﬂuences child outcomes primarily through its indirect
effect on maternal parenting. In that study, maternal reports of their own trauma
symptoms predicted infant trauma symptoms for infants who had witnessed severe levels
of violence. This suggests that, when the mother’s reactions to the abuse were extreme as
in PTSD, the infant may have read maternal cues as reﬂecting that the situation was, in
fact, dangerous. These infants then began to demonstrate PTSD symptoms of their own.

In summary, results ﬁ'om the current study suggest that DV may be most inﬂuential
in predicting child outcomes through its effect on maternal parenting. The direct

inﬂuence of DV exposure on child psychosocial outcomes in preschool-aged children

144

 

 

 

appears to be less inﬂuential. It is hypothesized that the inﬂuence of DV on parenting
may be especially important for young children. Further research is necessary to examine
the possible differential effects of DV exposure and parenting on children at varying
developmental stages.
Parenting and Child Outcomes

Despite the restricted variance within the child internalizing and externalizing
variables, the current investigation found that maternal positive parenting was
signiﬁcantly, negatively associated with concurrently assessed externalizing behaviors.
This ﬁnding is consistent with a substantial amount of empirical literature that has
demonstrated the effects of parenting behaviors on child psychosocial outcomes within
school-aged samples of children (W. A. Collins, Madsen, & Susman-Stillman, 2002). In
addition, it is consistent with Belsky and Fearon’s (2002) ﬁnding that concurrently
assessed parenting behaviors of toddlers signiﬁcantly inﬂuenced child behavioral
problems beyond the inﬂuence of infant attachment classiﬁcation. In contrast, a
signiﬁcant relationship between positive parenting behaviors and internalizing behavioral
symptoms was not found in this study. This suggests that parenting in DV-exposed
families is more inﬂuential on the development of child externalizing than internalizing
behaviors. Exposure to violence in the home may explain this distinction. That is, social
learning theory predicts that violence observed by children is learned and imitated
(Bandura, 1977). In this particular sample, however, given the possible trauma-related
effects of maternal preoccupation with the violence in the DV-exposed women, it is

possible that mothers are better able to attend to and report the externalizing behaviors of

145

 

 

 

Ll:

(IE1;
.- J

their children and less able to detect internalizing behaviors which require a greater
sensitivity to the subtle signs of the child’s social-emotional functioning.

The current study also failed to ﬁnd an effect of concurrently-assessed parenting on
the emotional self-regulatory capacities of preschool children. In contrast to the infant,
toddler and school-aged periods, parenting during the preschool period has not been well-
studied. During the infancy and toddler periods, the child’s attachment strategy is thought
to reﬂect, in part, the parenting behaviors of the caretaker(s). During the preschool period
of development, the emotional self-regulation capacities of children that have evolved in
the context of the parent-child relationship(s), are expected to be more fully internalized
and have become relatively stable characteristics of the child’s personality (Sroufe,
1995). At this point, parenting behaviors, like the child’s behaviors, are likely to shift as
the child’s regulation moves more fully from a dyadic to an independent process. In fact,
in their meta-analysis of the extant empirical examinations of parenting behaviors,
Holden and Miller (1999) found that parenting becomes increasingly stable across time
beginning when children are school-aged. Thus, parenting during the preschool period
may be in a period of transition and therefore less predictive of concurrently assessed
regulatory ﬁrnctioning. That is, earlier parenting, as reﬂected in the child’s infant
attachment category may still be more predictive of child regulatory capacities at this
point in development whereas concurrent parenting may be more inﬂuential in predicting
the child’s daily behavioral functioning (e. g., externalizing behaviors).

The differences in effective or optimal parenting during the preschool stage of
development compared with earlier parenting have not been fully examined. However,

from a theoretical, child development perspective, it may be that parenting during the

146

 

w- --.—_n— I—._

 

preschool period is less concerned with the “how” of regulation that is more reﬂective in
the child’s internalized attachment and regulatory strategy, and more concerned with the
“where, when and why” of regulation. The latter may be better captured by the CBCL
(Achenbach, 1991) in that this measure assesses the child’s behaviors as they manifest
within their day to day environments. For example, items that reﬂect the child’s
propensity to disobey and to argue with adults are captured by this measure and reﬂect
the child’s relative ability to ﬁmction within his or her social environment. Children
during this developmental period need adults to help them identify when and where it is
necessary to utilize their newly consolidated regulatory strategies (Sroufe, 1995), and
parenting may be shifting during this period to accommodate this developmental need.
For example, at this age it becomes important to sit still when the preschool teacher is
reading a book to the class, but it is equally important to understand that the demands on
emotional self-regulation shift during recess when it becomes important to engage in
activities that allow for the discharge of energy. Children need their parents, as well as
other caretaking adults in their lives, to help them understand the logistics of these
changing environmental demands. Future research on the differential skills necessary for
effective parenting in the preschool period compared to the infant and toddler periods is
necessary to more fully test these hypotheses.
Limitations of the current study

The current study has several important methodological limitations. For example,
although the use of observer rated maternal and child behaviors in this study is a
methodological strength, these data were gathered during relatively short behavioral

observations in the laboratory environment and not in the family’s home environment.

147

 

The nature of the sample made home observation potentially dangerous to both the 1
family and the researcher. However, especially in the case of maternal behaviors, it is
possible that women were able to exhibit their highest level of parental ﬁrnctioning in this
setting for this limited period of time. It may be that had observations been conducted for
longer periods of time in the home environments of these families, different results would
have emerged. Ainsworth’s (Ainsworth et al., 1978) original work, for example, used
home assessments of longer duration, and yielded signiﬁcant results despite a small
sample size (n = 23). Given the frequent use of relatively short-duration laboratory
assessments within the child development area of research, future work that establishes
the validity of these assessments compared to home observations will be an important
next step.

Another signiﬁcant methodological limitation of this study is the lack of temporal
precision in the measurement of domestic violence exposure. That is, the measures
utilized in these analyses assessed whether the woman had been exposed to abuse in the
year preceding the interview; the speciﬁc timing of the abuse was not assessed. Thus, one
woman could have been abused eleven months prior to the time of the data collection and
then left that relationship, whereas another woman may have been currently in an abusive
relationship. The measure captured the number and type of abusive events; it did not
assess its recency. Therefore, in the case of this example, if both of these women had
been exposed to the same level of abuse, they would have received the same domestic
violence score, even though it is likely that the impact of the abuse on current parenting
behaviors and child outcomes may have been stronger for the woman who remained in an

abusive relationship.

148

In addition, despite the signiﬁcant use of observer rated methodology for both
child and maternal behaviors, three of the measures rely solely on maternal report.
Speciﬁcally, measurement of both the mother and child DV exposure is reliant on
maternal self report. In both cases, practical considerations, including the safety of the
mother and child, make it difﬁcult to obtain third-party reports of the frequency and
intensity of the DV exposure. In some families, obtaining information from the assailant
could place the mother and her children at signiﬁcant risk of continued or intensiﬁed
abuse. A third party reporter such as a friend or family member of the mother may be one
way to obtain conﬁrmatory data about the abuse. This too, however, presents safety
issues in that, unlike the research team, family and friends are not bound by
conﬁdentiality and possible disclosures may compromise the mother’s safety.

As described within this paper, maternal report of child exposure to the DV is
especially problematic in that there are many potential reasons that a mother may
minimize her report of the child’s exposure, including her psychological need to
minimize the exposure in order to reduce her own sense of guilt. In addition, maternal
report of her child’s symptoms of internalizing and externalizing behavioral problems
may be minimized for similar reasons. Concurrent reports of these behaviors from other
signiﬁcant caretakers in the child’s life would have strengthened the methodology of the
current study. While the longitudinal study attempted to obtain teacher/caretaker reports
of these behaviors in children, many children were not in out-of-home care or preschool
at this time point, and only a minority of the available teachers and childcare providers
provided CBCL data. Thus, the results of this data were deemed to be of insufﬁcient

quantity for use in the current analyses.

149

 

Finally, one signiﬁcant limitation of the current study is that it utilized a sample
of convenience and, therefore, there are inherent threats to both internal and external
validity. For example, there were many variables that were not measured or accounted for
in the current study such as the child’s possible exposure to child abuse, or genetic
predisposition for psychopathology. These kinds of data were not collected and,
therefore, could not be controlled or manipulated in the current study. This leaves open
the possibility that the data examined here are inﬂuenced by other variables that may
account for the variance in both the independent and dependent variables. In addition,
threats to external validity include the fact that, since these women self-selected for
participation in the study, they may be unique in some way relative to the broader
population. It is possible, for instance, that the women who chose to participate were
experiencing lower levels of DV, even relative to their non-shelter, DV-exposed peers. In
this case, our sample would not represent the population of women who are exposed

community levels of DV.

Clinical & Research Implications

Results from the current study indicated that the children with ambivalent maternal
attachment histories evinced deﬁcits in emotional self-regulation capacities during the
preschool period of development. These children became dysregulated during a short
separation from their mother in an unfamiliar environment. However, emotional self-
regulation as measured in this environment was not associated with externalizing or
internalizing behavioral symptoms in the child’s everyday life. Although this suggests
that the child’s emotional self-regulation deﬁcits do not affect his or her global

functioning — at least in terms of the development of externalizing and internalizing

150

 

 

itll‘

 

behaviors, it is likely that there will be occasions wherein the child is required to endure
separations from parents or caretakers in unfamiliar environments. The results from this
study suggest that children with ambivalent attachment histories may have an especially
difﬁcult time in these situations. In the case of DV exposed families, there may be
separations due to the mother and children ﬂeeing the abusive situation and, as is often
the case in the cycle of abuse, returning after a period of time. Children in these families
are oﬁen placed with different caretakers at different times in this cycle. In what has been
termed “compassion fatigue”, clinical experience suggests that a woman’s friends and
relatives are often willing to provide them with assistance, including childcare support,
early on in the cycle of abuse, but may become weary after several cycles of the woman
leaving, and then returning, to the abusive relationship.

These kinds of disruptions may be especially difﬁcult for children with
ambivalent attachment histories. That is, due to their inability to use ﬂexible and age-
appropriate emotional self-regulation strategies, children with ambivalent attachment
histories are likely to struggle within these kinds of chaotic life styles. This suggests that
infant mental health (IMH) interventions that are designed to foster secure mother-infant
attachment relationships within DV exposed populations may be particularly helpful to
the children who would otherwise form an insecure-ambivalent attachment relationship.
A recent study using data from the current longitudinal sample suggests that child
outcomes in DV exposed families are best predicted by both risk and resiliency factors

(Martinez-Torteya et al., in press). Thus, IMH interventions that help infants develop

 

secure attachment relationships with their mothers may provide a protective factor for

151

children from the negative inﬂuence of the chaotic life style that is often associated with
DV exposure.

In a related ﬁnding, results from the current study also indicated that there is an
indirect effect of DV exposure on child externalizing behaviors via the impact of DV on
maternal parenting. This ﬁnding points to the need for intervention programs that are
aimed at improving parenting skills in DV exposed families. There are currently many
programs designed to assist women in leaving abusive relationships. Fewer interventions
are aimed speciﬁcally at improving the parenting skills of DV-exposed women. In one
important exception, Graham-Bermann (Graham-Bermann et al., 2007) has developed a
community-based intervention program for DV-exposed mothers and their children that
employs both peer support and skill building components and has produced empirical
data supporting its effectiveness (Shavers et al., 2005). Infant mental health theory argues
that parents are often willing to make signiﬁcant changes on behalf of their children that
they might be unable to make for their own wellbeing (Fraiberg, 1987). Thus, parenting
interventions with DV-exposed women may serve to decrease the possible cognitive
dissonance they experience with regard to minimizing the effect of the violence on their
children, and thereby increase their propensity to leave abusive relationships.

Results from the current study also suggest many avenues for ﬁxture research. For
instance, although results indicated that child emotional self-regulation capacities were
not associated with concurrently assessed internalizing and externalizing behaviors, it is
not known whether there may be an association with symptoms of behavioral disorders
later in development. Theoretical arguments from transactional theory and developmental

psychopathology suggest that the impact on these symptoms may not become evident

152

 

until later in development. This possibility is most clear for children with ambivalent
attachment histories who were found in this study to evince deﬁcits in their emotional
self-regulatory capacities. As social pressures for independent self-regulation continue to
increase, these children may encounter increasingly negative feedback from their adult
caretakers resulting in the establishment of externalizing behaviors later in development.

Another important step for future research within the domestic violence ﬁeld is the
development of temporal precision in the assessment of DV. One strength of the current
study was the use of quality tracking procedures to ensure a substantial retention rate over
several waves of data collection in a high risk sample. Speciﬁcally, women were
contacted every three months to conﬁrm and, when necessary, update their contact
information. One possible way in which to capture more precise measures of DV is to
link these tracking contacts with the assessment of DV. Employing this level of data
collection could improve the precision with which DV is measured and may also improve
the validity of the maternal report in that they would be asked to remember and report on
DV exposure over a much shorter period of time.

Similarly, the assessment of SES in this study was based fully on the family’s
income. There are many important indicators of a family’s SES including income,
parental educational level, access to concrete supports and services, among others. In
addition, the sociocultural and racial background of the family is also likely to hold
signiﬁcant consequences for the overall functioning of the family due to social processes
that often restrict access to resources for socially oppressed groups. Future research

should examine the inﬂuence of a broader range of factors that inﬂuence SES in an effort

153

 

 

to better understand the inﬂuence of particular variables, or combinations of variables,
that impact child regulation and symptoms of psychopathology.

Finally, the current study found signiﬁcant associations between infant attachment
classiﬁcation and later emotional self-regulation capacities in preschool aged children.
However, concurrently assessed, observer-rated parenting behaviors were not found to be
related to the child’s regulatory capacities. Future research should further examine the
parenting behaviors that are effective in helping preschool children negotiate the age-
appropriate developmental press to consolidate their emotional self-regulation capacities.
The ways in which effective parenting during this developmental period compared to
other periods of development is especially needed, given the relative paucity of
examinations of parenting during the preschool period.

Summary of Study Contribution

Despite explicit theoretical predictions linking infant attachment to emotional
self-regulation styles across the lifecourse (Bowlby, 1969/1982), this study is the ﬁrst to
examine the inﬂuence of infant attachment category on child regulatory capacities at the
typological level in preschool children. The preschool period is unique relative to the
child’s development of emotional self-regulation; it is at this point in development when
regulatory strategies become more fully consolidated resulting in increases in the
prediction of individual differences from this point forward (Sroufe et al., 2005). Results
from the current study suggested that, in a combined community (e. g., non-shelter)
sample of DV-exposed and non-exposed families, insecure infant attachment was, in fact,

associated with the child regulatory capacities of four year old children. These results

154

provide empirical support for the theoretical hypothesis that early attachment experiences

inﬂuence later emotional self-regulation capacities.
In addition, the current study contributes the growing body of literature examining

the effects of attachment on later child symptoms of psychopathology. Once considered a
fundamental deﬁcit associated with negative developmental outcomes (Erickson et al.,
1985), empirical work has demonstrated that insecure attachment is predictive of
psychopathology most robustly in the presence of other risk factors (Greenberg, 1999).
The current study contributes to this literature by examining this relationship in a sample
of mothers and children of whom seventy percent endorsed maternal DV exposure at
some point during the child’s ﬁrst four years of life. DV exposure in early childhood is
unique in that it poses threats to the child and to the child’s primary attachment ﬁgure at a
time in development when the child’s regulatory capacities are based most ﬁrlly on their
attachment relationship. Results indicated that DV exposure was not directly predictive of
child externalizing behaviors but, instead, was indirectly predictive through its impact on
maternal parenting. This ﬁnding is consistent with attachment research and theory which
highlights the centrality of the primary attachment relationship to the child’s early growth
and development. The ﬁndings reported here contribute to our understanding of the

complex dynamics that underlie the effects of DV in early child development.

155

APPENDIX

156

MATERNAL PARENTING CODING SCHEME

Mother Infant Study Coding Manual
4 Year Old Eyberg Protocol (adapted)

MATERNAL CODES
This adapted version of the Eyberg Protocol lasts a total of 14 minutes and is equally
divided into two situations: 1) child directed play and 2) mother directed play. There are
’7, 5-point, maternal scales which are scored in one-minute intervals. Scores for each
domain are then averaged across all 7 one-minute segments for each of the 2 protocol
situations. This yields a mean score which is used as the ﬁnal score for that domain in
tllat situation. Thus, each case receives a ﬁnal average score for each domain (e.g.,
positive affect) in both of the 2 protocol situations.

~I".>/Iatern_al Codes“
1 - Positive Affect and Affection
2 . Negative Affect and Feedback
3 - Positive or Neutral Feedback
4. Facilitates Self Regulation
5. Intrusive / Overcontrolling
6. Unresponsive, Unavailable, Undercontrolling
7. Sensitivity

*Adaptations:

The ﬁrst 6 codes included in this manual were adapted from the MS U Family Project
Second Revision of Belsky ’s Parent-Child Interaction Coding System, by Ellen E.
Whipple, Ph.D., Natalie L. Denburg, and W. Hobart Davies, Ph.D. Their version was
adapted from : Belsky, J., Youngblade, L., Rovine, M. & Volling, BA. (1991). Patterns
of Marital Change and Parent-Child Interaction. Journal of Marriage and the Family, 53,
487-498.

The sensitivity code is adapted from the Emotional Availability Scales developed by Dr.
Zeynep Biringen in collaboration with Drs. JoAnn L. Robinson and Robert N. Emde.

Biringen, Z. (2000). Emotional availability: Conceptualization and research ﬁndings.
American Journal of Orthopsychiatry, 70, 104-1 14.

Biringen, Z. in collaboration with J. L. Robinson and R. N. Emde, Emotional availability

scales, 3rd edition. Attachment and Human Development (special issue on emotional
availability and attachment), 2, 257-270.

157

General Coding Comments
0 Coding Procedures:

0 Reread the manual before you start coding. After you’ve coded a critical
mass of these you will be able to skip this part but in the beginning and
middle phases of this project you should do this every time

0 Watch the ﬁrst minute, stop the DVD and code each domain.

0 Rewatch the ﬁrst minute, stop the DVD and check your codes. Make any
changes and then repeat this process for each of the 14 minutes.

0 You may need to watch each minute more than two times —- especially at
the beginning.

0 Make sure that you ﬁll out the coding sheet completely (date, name, etc.).

0 For each of the scales, the baseline score is indicated with a B. This is to give the
coder a frame of reference as to which is the most frequently occurring rating for each
construct.

0 Coders should make note on the code sheet of any unusual incidents which occurred
during the segment. Examples may include the child leaving the room for a period of
time, problems with audio or video, mistakes in the administration of the protocol
such as timing, etc.

0 An important feature of this coding system is that it taps both qualitative (feeling,
intuition) and quantitative (behavioral frequency) aspects of the parent-child
interaction. This includes verbal / nonverbal communication and direct / indirect
factors. Additionally, this system is an effort to identify the way things are said (the
affective quality) in addition to the content of what is said.

0 Some of the scales are interrelated, in that if you give a certain score on one item,
other scales must receive especially high/low scores.

158

1. POSITIVE AFFECT & AF FECTION

This rating assesses the extent to which the parent displays warmth, nurturance,
and positive affection toward the child and enjoys interacting with the child. Extent is
deﬁned in terms of both frequency and intensity. Behaviors that evidence such an
orientation are numerous and diverse. Among others, they include affectionately touching
the child, smiling at and laughing with the child, as well as being enthusiastically
involved in what the child is saying or doing.

In comparison with the positive feedback rating, these behaviors do not have to be
contingent on child behavior. However, for positive affect to be coded as feedback, the
parent's response must be speciﬁc to a very explicit child behavior. The more general the
situation that evokes the parents' positive affect, the more likely it is that it should be
coded msitive affect instead of positive feedback. Positive affect that qualiﬁes as positive
feedback is not coded here (e. g., affect is a feeling , feedback is a behavior).

NOTE: All instances of positive affect should represent genuine warmth. Smiles and
laughter that are sarcastic or mocking should not be counted.

1.No or very few instances of warmth, affection, or enjoyment are observed. Parent may
be instructional or uninvolved.

2. B Parent is involved in what the child is doing, initiates talk with the child.

3. 1-2 instances of verbal or nonverbal warmth, affection, and enjoment are observed.
For example, parent laughs with the child. Parent smiles or appears genuinely happy,
or makes eye contact which conveys warmth.

4. 3 instances of verbal or nonverbal warmth, affection, and enjoment are observed.
Parent is involved and enthusiastic. Feeling is one of support and nurturance
consistent throughout the minute.

5. Instances of warmth, affection, and enjoyment are very frequent, or intense, exuberant
(4 or more instances).

159

2. NEGATIVE AFFECT & FEEDBACK

This rating assesses the extent to which the parent displays hostiligg, negative
affect, and displeasure or annoym toward the child, with extent deﬁned in terms of
both ﬁequency and intensity. Behaviors that evidence such an orientation are numerous
and diverse. These include annoyed or scomful facial expressions and posturing,
aggressive handling of the child, explicitly negative or scomful vocal tones, and clear
lack of enjoyment of the child in this situation. The parent behaviors rated in this scale do
not have to be contingent upon the child's behavior but they may be. The demonstration
of “ﬂat” affect is not coded here. A mother must actively demonstrate negative affect to
score here (note that this may come through in her body posture, tone of voice, etc.).

 

NOTE: This rating includes sarcasm. Sometimes it may be difﬁcult to distinguish
between positive and negative affect as the parent may smile in a way which on the
surface appears positive, but s/he is laughing at the child and/or ridiculing him/her.

NOTE: This rating also includes criticism.
NOTE: Negative affect can be expressed nonverbally.

NOTE: This code includes disappointment and displeasure with the child’s behavior even
if this disappointment seems justiﬁed. For example, the mother repeatedly expresses
displeasure because her child disobeys her when she asks him/her to switch tasks.

1. B No instances of hostility, negative affect, or displeasure are evident.

2. Parent's facial expression, posturing, or tone of voice is negative. However, there are
no explicit negative messages (e. g. verbal comments or physical gestures).

3. Some subtle instances of hostility, negative affect, or displeasure/annoyance occur,
but there is no escalation in intensity or loss of control.

4. 34 instances with some escalation or loss of control.

5. Instances of hostility, negative affect, and displeasure are frequent (>4) and/or

intense. Much of the time during the one-minute period is characterized by this kind
of behavior.

160

3. POSITIVE OR NEUTRAL FEEDBACK

This scale assesses the extent to which the parent provides contingent rewards and
praise to the child for his/her behavior via verbal and/or nonverbal feedback, with extent
deﬁned in terms of both frequency and intensity of feedback. Statements such as “that’s
good” would thus be weighted less heavily in this rating than more elaborate ones like
“that’s terriﬁc, you really worked hard at that ; I’m proud of youl”. Other examples of
positive feedback include statements like “you did a good job”, or a pat on the back or
clapping of hands in response to the child’s accomplishments. The observer must be able
to identify the contingency between child and parental behaviors in order to score this as
feedback as opposed to positive affect. The parent should be intentionally or explicitly
providing a positive response to particular child action.

NOTE: Emotional feedback should be credited here as well. Labeling the child’s feelings

or responding empathically (e.g., affect attunement) (e.g., “You really like doing this,
don’t you?”) should be scored >4.

NOTE: This scale is includes “neutral” feedback, as sometimes the parent provides

neutral feedback such as “there you go” or “that’s right”. Thus, the parent is providing
the child with some — as opposed to no — input.

1. No feedback is observed. Mother has to be fairly non-responsive to be coded here.

2. B Parent makes one or more neutral acknowledgements (e.g., OK, uh-huh) in
response to child’s verbal or nonverbal behavior or vocalizations. Thus, parents who
are really attentive are probably going to score at least a 2 on this scale.

3. Within any single minute parent says 1-2 times “that’s good” or “very good” without
elaborating on the feedback and/or without special enthusiasm. Other examples
include “there you go” or “nice job”.

4. Parent gives 3 feedbacks or fewer than 3 but with special enthusiasm (e. g.,
elaborating on the feedback verbally).

5. Positive feedback is frequent or intense and characterizes much of the way the parent
responds to child. To give a 5, it should seem as if there is continuous positive

feedback. The feedback should have a personal quality that validates the child’s sense
of self.

161

4. FACILITATES SELF-REGULATION

This scale assesses the extent to which the parent facilitates his/her child's ability
to control self and actively and positively engage the situation. That is, provides the
"scaffolding" which allows the child to direct/structure the play. This is based on the
presupposition that in order to facilitate the child’s attempts at self-regulation, the
parent must be able to take the child’s perspective in any given moment or
interaction and respond according to the child’s needs. Parents are rated in terms of
their ability to provide supportive assistance that facilitates the child’s competent
functioning, with extent being deﬁned in terms of the frequency and intensity of parent's
behavior. With a facilitative parent, you get the sense that she is involved and available to
the child and is "pre-digesting" the information for the child and making the task a bit
easier and more manageable. The interaction has a positive feeling with a facilitative
parent.

The evaluation of the parent on this scale is not dependent upon the success of
parent's facilitative acts; thus this scale assesses only the parent's skill effort and intent.
Also, one never lowers scores (except when considering a score of 5) because of
additional intrusive, unresponsive, or negative parental behavior.

For Child-Directed Play, it should be noted that if the parent sits close to child and
is attentive and available to him/her, the parent will receive a minimum score of 3. For
Egrent-Directed Play, parent and child working together on a chosen activity is scored a

minimum of 3. As more speciﬁc facilitative behavior is displayed, the parent’s score will
increase appropriately.

 

 

Examples:

A. During play, parent provides supportive presence. whether actively playing with
child or watching child play yet, "being there" for child.

B. The mvision of rationﬂ which offer information or appeal to positive or neutral
consequences in order to obtain compliance. For example, the parent’s rationale
might indicate that some child action will please the parent, will be in accordance
with a rule, or will lead to pleasant outcomes. Rationales that provide information
might involve statements like "Why don't you do X, because something interesting
might happen; because it works that way," or the like.

C. The moulation of materials in a way that improves child's chances for being
successful, yet does not involve doing the task or action for child. Such facilitative
assistance may take the form of giving verbal hints, repositioning a piece where
child will see it better, assisting child in doing something difﬁcult by helping

him/her manipulate the piece -- in essence, by providing a "scaffolding" for the
child to use.

D. The ppovision of a well-timed directions (but not too frequent or intrusive),
delivered in a pleasant or encouraging tone of voice which points child in the
"ri t" direction. For example, describing how to ﬁt the block together to achieve a

desired goal, or helping them write their name. Suggestions made by the parent lead
to more developed play.

162

The provision of well-timed interventionp that prevent child from becoming over-
aroused and disorganized. Such facilitative parenting may take the form of the
parent "steering" or "inducing" child away from a potential frustration, but not
before child him/herself has a chance to cope, unless even one effort to cope is
likely to overwhelm child.

Assistingchild in the expression of his/her thoughts or feelings that supports the
child’s desire to express and control them. Conversation or dialogue in the service
of maintaining child’s organization. More generally paraphrasing child’s feelings in
ways that facilitate the organization of a child's behavior and his/her coping is
considered here.

. No evidence of facilitative behavior is observed OR parent does not pay attention to
child in child-directed play OR parent plays by him/herself during parent-directed
play OR just orders and sits back observing.

 

. Echoing child's comments while sitting back OR has limited involvement with the
child.

. B 1 clear instance of facilitative behavior in addition to being available to child;
paying attention to what child is doing; involving child in play during gent directed
play.

 

. 2 clear instances of genuine facilitative behavior in addition to being attentive, where
the parent seems emotionally invested Must be explicit acts which have a genuine
quality of the parent trying to be helpful.

. Instances of facilitative behavior are frequent (3 or more) or especially salient and
characterize much of the parent's way of relating to the child.

163

5. INTRUSIVE / OVERCONTROLLING

This scale assesses the extent to which parental behavior is ill-timed, intrusive and
excessively and inappropriately controlling relative to what the child is doing. The
parent’s behavior may be ill-timed in the sense that it disrupts child’s own goals and
pursuits, or lacks empathy or synchrony with the child’s feelings and action, respectively,
and thus is psychologically intrusive. Intrusive behavior is likely to be dictated by a
parental agenda regarding what should be going on and disregarding child’s behavior. It
can be either verbal or nonverbal, direct or indirect.

NOTE: Intrusive and facilitative behaviors are on a continuum, where it can be hard to
draw a line between the two. For example, a parent may receive high facilitation scores
for several minutes because s/he is actively engaged in helping the child build a house
with Lincoln logs. However, a "line" may be crossed from facilitative to intrusive when
you feel the parent is over-involved in helping the child in a manner that is not promoting
his/her success/independence. In this example, the parent "tells" (directly or indirectly)
how to build the house in a manner that stiﬂes the child’s creativity/ growth.

Examples:
0 during play parent directs/structures play in a way that does notallow child

 

 

to exploreﬁand decide wlyat to play with. Parent often tells child what to play with or
in some way decides what parent/child will play with together, without regard to
child’s wishes (e.g., "We're going to play with this toy now" or "Here, you play with
these blocks now"). Parent interrupts or distracts child’s own play or conversation.
(this does not include the transition to parent-directed play)

0 the govision of constant verbal directions that are not timed according to
what child is doing and leaves child which little room for autonomous ﬁmctioning.

o "guizzing" child in an interfering wg ("What color is that? How about that?
Who is that?" etc). This is typically done in a repeated manner where the child is
forced to respond to the parent’s question.

0 intrusive manipulations of child’s body or materials to force child to behave

in a certain manner, (e.g., pushing child’s arm back and forth to "help" him do
something.)

0 interventions into child’s actions before childpas a real chgce for mastery -
- not timed to child’s degree of coping, but rather to parent's need to " get on with it"

0 while child is pretending/participating in pretend with parent, parent usurps
control by trying to force in literal explanations rather than going along with child's
non-literalness.

0 during parent-directed play, parent gives child no freedom within the chosen
activig; stiﬂes child's creativity or elaboration.

164

INTRUSIVE / OVERCONTROLLING (cont)

1. B No evidence of intrusiveness is observed

2. Unnecessarily dictatorial instructions OR subtle intrusions that don't necessarily
distract child. One verbal intrusion.

3. Some instances or intrusiveness are observed in a way which clearly interrupts the

ﬂow of the play. (2 instances of verbal intrusions or one intense physical intrusion).

4. 3 verbal intrusions or 2 physical intrusions.

5. Instances of intrusiveness are frequent (>3) or especially salient and characterize
much of the parent's way of relating to the child.

* Remember to code for intrusive behavior that occurs during the transition (the ﬁrst
minute of mother directed play).

165

6. UNRESPONSIVE, UNAVAILABLE, UNDERCONTROLLING

This scale assesses the extent to which the parent makes no attempt either to
control or to facilitate the child's behavior at a time when support, assistance, or
availability would be helpful to the child. Indeed, undercontrol should be seen as "doing
nothing," or token gestures made by the parent for the beneﬁt of the experimenter, but not
the child. This scale taps parental permissiveness. This scale also assesses the mother’s
overall level of involvement with the child. For example, a parent who stares into space
while the child plays happily would nevertheless receive a higher score on this scale.

Examples:

 

0 Parent does not attempt to engage child in any activity or makes a "token
gesture" for the beneﬁt of the experimenter, during periods when child could use
some support, guidance, etc. Parent allows child to direct/structure play activities, but
does not supply a supportive presence for the child, when it is clear that the child
could use some assistance. Parent may ignore bids for assistance from the child, or
mistarget or misinterpret child’s cues, or make perfunctory attempts at aiding child.

0 If child transgresses, parent will deliver thceprohibition in a vagpp,
haphazard m_anner, or simply not deliver it at all.

0 Parent does not respond to child’s comments or questions or explicit child
bids for engagement such as the child saying, “look at this.”.

0 Parent is playing independently of the child and appears self-absorbed. For
example, both parent and child are drawing but mother concentrates on her own
picture and makes no effort to connect with the child (e.g. conversation or smiles) or
observe how the child is doing (e. g. glances over at the child picture). For these
periods of time, the number of seconds should be counted exactly.

1. B No evidence of under-control or uninvolvement is observed.

2. A little evidence of under-control or uninvolvement. At least some (at least one clear
instance, lasting at least 5 seconds) of the minute spent uninvolved with child (sitting
back in chair OR looking around OR obviously bored OR not answering the child’s
questions).

 

3. A moderate amount of under-control or uninvolvement. Approximately 10 — 20
seconds of the minute.

4. About half of the minute spent in under-control or uninvolvement (21-30 seconds).

5. Most of the minute. Signiﬁcant lack of involvement with the child OR salient
individual instances of unresponsiveness as deﬁned above (e.g., not responding to
child's question). (Over 30 seconds)

166

7. SENSITIVITY

Sensitivity reﬂects the mother’s ability to perceive and accurately interpret the child’s
signals and to respond to them appropriately and promptly. Sensitivity requires not
only that the mother is accessible to the child, rather than ignoring or neglecting, but that
she is alert to subtle aspects of the child’s signals. Sensitive responses are well-timed,
reﬂect empathy with child’s needs and feelings, and even when limit setting is called
for, involve behavior that will enhance child’s security, comfort, and development. In
this manner, a sensitive mother provides a “supportive presence” for the child and
seems to be emotionally connected to the child.

0 CONTINGENCY is weighted heavily here. Mothers who respond contingently
to their child’s behaviors are scored higher.

0 NEGATIVE BEHAVIORS & AFFECT such as hostility and intrusiveness are
also important. Mothers who demonstrate several examples of negative behaviors
are rated lower.

0 POSITIVE BEHAVIORS & AFFECT such as warmth and joy are also
considered here. Mothers who demonstrate several examples of positive behaviors
are rated higher. In addition, mothers who demonstrate authentic and contingent
affect are scored higher.

0 ATTUNEMENT and AWARENESS (which are related to contingency) of child
behaviors and cues is critical to obtain a high score on this scale. Mothers who are
in tune with their child will often comment on the child’s behavior and extend the
meaning — “That’s right, that’s the red ball”. This may manifest behaviorally as
well as when a mother extends a behavioral game which her child has started or is
interested in. You have a red ball at home, don’t you.” Alternatively, mothers who
consistently respond to their child’s behavior with one script will score lower here
— e. g. mothers who say, “what?, what?, what?” to everything the child does or
comment only with one response over and over again.

0 RESOLUTION OF CONFLICT is also important and includes the ability to be
ﬂexible and sensitive in resolving conﬂicts with the child.

1. Highly Insensitive Extreme insensitivity to the child’s communications and little
apparent knowledge of crucial child-rearing techniques. Affective negativity (in the
form of either active harshness or passive disinterest/depression) is more extreme.
Mother does not take each child into account when initiating, prolonging, or turning
away from interactions. In fact, there may be little interaction for some of these
dyads. This relationship is very painful for an observer to observe. If the observer has
an impression of at least minimal positive experience for this child, the parent should
receive a higher rating.

2. Somewhat insensitive - Insensitivity is typically displayed in one of two general
ways, one being an active/harsh style (overly active and overbearing) and the other
being a passive/depressed/affectively ﬂat (noninteractive and silent) style. Still, there
are positives here. Both styles suggest unresponsiveness to child communications and
lack many of the features of sensitive interactions described earlier. The

167

 

 

active/harsh/volatile style involves facial expressions of disgust and anger and
harsh/abrasive/condescending tones of voice. The passive/depressed/aﬁ‘ectively ﬂat
style involves facial expressions that are depressed, and disinterested, and a vocal
tempo that is slow, lethargic, or simply unenthusiastic. Also often seen is a business-
like, matter-of-fact style that combines features of both abrasiveness and passivity.
Althougp such interactions m_ay lacka clear ﬁrn-likp, aynchronous gality. they
indicate that this parent has some notionﬂbout what is important for child rearing,
The observer feels somewhat uncomfortable or sad when watching this interaction,
but still sees some positive experiences provided by such a parent. If the parent has
only a couple of clearly non-optimal qualities, such as bland affect and an
unenthusiastic tone of voice, s/he should receive a slightly higher rating.

. Inconsistently Sensitive - The parent is sensitive in some ways, but the observer
ﬁnds it difﬁcult to give this relationship a clean bill of health. For example, the parent
may ﬂuctuate from being creative and joyful during play times to being preoccupied
with other concerns, or other questionable (though not clearly negative/insensitive)
behaviors. Mothers may “leak” inconsistencies of behavior; it may simply be too
stressful for some to maintain well-modulated positivity for long. A rating of “3” is
typically given when the observer notes some signs of sensitivity (e.g., positive
statements, smiling, and interest) but also notes some problems in these areas (e.g.,
positive statements said in a slightly bored tone, smiling that does not seem authentic
and genuine, or interest that is occasional or feigned).

. Generally Sensitive - This parent is very similar to a “5”,” except that there is a less
spectacular quality to these parent-child exchanges. This rating refers to a “good
enough” parent. Typically, very positive interactions get rated down to “4” for some
of the following reasons: the parent did not interact in a creative manner, although
s/he was affectively connected to the child and interactions were harmonious and
enjoyable; or the parent’s affect and behavioral style were extremely well suited to
this child, creating a generally lively and engaging climate, but at brief moments, s/he
displayed subtle preoccupation with his/her own thoughts, as if processing another
agenda; or the like.

. Highly Sensitive - Emotional communication and play interactions between parent
and child are positive, joyful, and creative. The highly sensitive parent displays much
gguine, authentic, and congruent interest, pleasure, and amusement with the child (as
opposed to performing these behaviors), as demonstrated by warm smiles and
giggles, interested eye contact, and comforting and playﬁrl physical contract. The
parent accurately reads the child’s signals, even subtle ones that may not be clear to
an outsider and reacts appropriately. She has a well-developed sense of timing and
rhythm during interactions with transitions between activities appearing smooth
rather than abrupt and enforced. Parental behavior appears ﬂexible and adaptable,
according to the demands of particular situations. Verbal and visual communications
between parent and child are ongoing but not constant or overwhelming. Conﬂict
situations do not lead to long breakdowns in the relationship; instead, they too are
handled smoothly and effectively.

168

 

 

CHILD REGULATION CODING SCHEME

Mother Infant Study Coding Manual
4 Year Old Child Coding Protocol

CHILD CODES

This protocol lasts a total of 12 minutes and includes 4 segments of the Strange Situation
Protocol: First Stranger & Child Segment, Child Alone Segment, Second Stranger and
Child Segment, Final Mother-Child Reunion Segment. There are 8 child scales which are
scored in three-minute intervals.

Affect Codes

1. Comfort / Enthusiasm (Whipple)

2. Anger / Frustration (Whipple)

3. Sadness (CJD — based on Anger from Whipple)

4. Emotional Lability (Clark / CJD)

"Was the session ended early because of child distress? YES / NO

Emotional Regulation Abilities
5. Self-Regulation / Organizational Capacities (Clark; Cole et a1 1992)
6. Regulation Strategies (5 domains) (CJD)

Attention & Activity
7. Attentional Regulation (CJD)
8. Activity Level (CJD)

The ﬁrst 3 codes included in this manual were adapted from the MSU Family Project
Second Revision of Belsky ’s Parent-Child Interaction Coding System, by Ellen E.
Whipple, Ph.D., Natalie L. Denburg, and W. Hobart Davies, Ph.D.

169

 

General Coding Comments
Coding Procedures:

0 Reread the manual before you start coding. After you’ve coded a critical
mass of these you will be able to skip this part but in the beginning and
middle phases of this project you should do this every time

0 Watch the ﬁrst segment, stop the DVD and code each domain, making
notes relative to each code on your coding sheet.

0 Rewatch the ﬁrst segment, stop the DVD and check your codes. Make any
changes and then repeat this process for each of the four segments.

0 You may need to watch each segment more than two times — especially at
the beginning.

0 Make sure that you ﬁll out the coding sheet completely (date, name, etc.).

Coders should make note on the code sheet of any unusual incidents which occurred
during the segment. Examples may include the child leaving the room for a period of
time, problems with audio or video, mistakes in the administration of the protocol
such as timing, etc.

An important feature of this coding system is that it taps both qualitative (feeling,
intuition) and quantitative (behavioral frequency) aspects of the child’s behaviors.
This includes verbal / nonverbal communication and direct / indirect factors.

Some of the scales are interrelated, in that if you give a certain score on one item,
other scales must receive especially high/low scores.

170

 

 

COMFORT and ENTHUSIASM

This scale assesses the comfortableness and enthusiasm of the child during the session.
Factors to consider when coding this scale include ease of movement, relaxed posture of
body, facial expressions, and/or behaviors. A child high on this scale would clearly be
having a good time, with spontaneous and zestful expressions of fun and delight and clear
interest in play and/or interaction. The child would also appear very relaxed, with a high
degree of ease of movement. A child low on this scale would look frozen or highly
tentative, with stiff body movements, and distressed facial expressions. An agitated, sad
or angry child is also low on this scale.

1. Very Uncomfortable: There is no enthusiasm, smiling or interest. Movement may be
either slow or agitated. Child may show extreme hesitancy to engage in play with
toys. Child may appear highly anxious and withdrawn, frozen, or highly distressed. In
sum, child seems very uncomfortable and apparently would rather not be here.

2. Neutral or Uncomfortable: Facial expressions most often are bland. Child may
appear to be intensely concentrating on toys, but there is no quality of real
engagement, excitement or fun. Body movements appear stiff and child shows little
evidence of enjoyment. Child may appear bored. Interaction with toys looks almost
obligatory as opposed to desirable. Further, interaction with toys will probably be
characterized by repetitive, "non-thoughtful", simple manipulations of the items. Or,
child may seem unhappy and/or agitated.

3. Modestly Comfortable: Brief moments of ﬂeeting pleasure, engagement or
enjoyment. Child looks relatively comfortable, but not fully at ease. Her movements
may be somewhat hesitant, but she is more likely to look neutral than totally self-
conﬁdent. Although child seems to be engaged in play or interaction, coder isn't
convinced that she would choose to come back if she had the choice. “Slow to warm
up” children who are hesitant at ﬁrst but then seem to be enjoying themselves by the
end of the session might be coded here. Note that there must be some sense of
enjoyment — even if the only evidence is that they are well engaged with the toys.

4. Comfortable: Some animation, engagement or enthusiasm is apparent during
segment. Facial expression shows interest and pleasure during play. Vocalizations
clearly indicate child is enjoying and/or highly engaged with playing or interacting.
Child appears comfortable and in charge. Child looks secure, and negative signs are
very rare. Child typically appears thorough and/or spontaneous in play, but some
tentativeness may be noted.

5. Enthusiastic: Child appears animated or delighted in terms of her interaction with
toys or with the adult. Child obviously feels good about him/herself, conﬁdent,
secure, comfortable. Child's activities appear spontaneous and zestful. He is clearly in
control of the situation, purposeful, engaged in play and/or interaction, and seems
glad to be here. Child is likely to be smiling, moving easily, "lost" in his play.

171

 

 

ANGER/FRUSTRATION

This scale assesses the extent to which the child shows anger, dislike, frustration or
hostility, with extent deﬁned in terms of frequency and intensity. Evidence of negatrvrty
may take the form of angry vocalizations, physical “acting-out” behaviors, etc.

Speciﬁc examples of negativity would be stomping feet, throwing toys, yelling in
unpleasant tones, making angry facial expressions, pulling angrily at the door knob, or
banging angrily on the toys or the door or the window. Negativistic behaviors directed
toward persons would include shouting "no", throwing a tantrum, or throwing an object.
Note that positive affect may be present (although not simultaneously) without
inﬂuencing this score.

1. No evidence of angry affective expressions is displayed.
2. A few mild instances observed for brief periods.

3. Child displays some or modest evidence of angry or frustrated expression. This
behavior should not escalate into an extensive display, i.e., behaviors are somewhat
isolated - 3 instances observed for brief instances each time OR one moderately
intense period.

4. Several instances are observed OR child displays angry behavior for extended
period(s) of time.

5. Child displays anger and/or hostility frequently and in long duration. Anger seems to
characterize this child’s affect for much of the session.

"Note: Usually, if child is so angry/upset that the session is terminated early then they
should receive a score of 5 on this code. However, there are cases where the child is not
totally upset but the session was ended early anyway (presumably due to the mother’s
wishes). In these cases, code based on the observed behaviors.

** Note: If the difference between anger and sadness is really unclear - you should split
the scores to represent an equal number of each.

172

 

 

SADNESS

This scale assesses the extent to which the child shows sadness with extent deﬁned in
terms of frequency and intensity. Sadness can include sad looking facial expressions such
as pouty lips, crying, whimpering, sad verbal content such as “I’m all alone in here,”
lethargic/spacey play, etc.

1. No evidence of sadness is displayed.
2. One to two mild instances observed for a brief periods.

3. Child displays some or modest evidence of sadness. This behavior should not
escalate into an extensive display, i.e., behaviors are somewhat isolated - 3
instances observed for brief instances each time OR one moderately intense
period

4. Child displays sadness frequently or in long duration. Sadness seems to
characterize this child’s affect for much of the session.

5. Child displays sadness frequently and in long duration. Sadness seems to
characterize this child’s affect for most of the session. For a child to get a “5” on
this scale (vs. a 4) you should not be able to identify a sigpiﬁcant period of
positive affect (happy) during the segment. Times of neutral affect may still be
present

"Note: Usually, if child is so sad/upset that the session is terminated early then they
should receive a score of 5 on this code. However, there are cases where the child is not
totally upset but the session was ended early anyway (presumably due to the mother’s
wishes). In these cases, code based on the observed behaviors.

** Note: If the difference between anger and sadness is really unclear — you should split
the scores to represent an equal number of each.

173

 

 

 

EMOTIONAL LABILITY

This item assesses the degree to which the child experiences identiﬁable
emotional/affective shifts. The frequency of the emotional shifts are considered here with
frequent transitional shifts resulting in higher scores on this scale. In addition smoothness
of transition is also considered in that “smooth” transitions are awarded lower scores and
“abrupt” transitions are awarded higher scores. Smooth transitions are deﬁned as a slow
move from one affective state to another. For example, a child may seem mildly sad and
then, over a period of time, their sadness may slowly increase until they are crying
outright. In contrast, an abrupt transition is exempliﬁed by a child who seems perfectly
happy and then, all of a sudden, begins to behave in an angry or sad manner. Note that
valence itself is not relevant here. That is, a child who is happy throughout the session
without shifts to other emotions would be scored the same as a child who is sad or angry
throughout the segment. However, it is more likely that a child who is demonstrating
strong negative emotions will have some shifts to more neutral emotions at some point(s)
during the segment as she tries to regulate herself. Therefore, it is easier for a child to
receive a “1” score if she is exhibiting relatively positive or neutral affect throughout.

1. Child’s affective state is relatively stable throughout. Very little evidence of
emotional lability. Note that children who seem mildly happy/comfortable and/or
neutral are coded here.

2. At least 2 distinct emotional qualities (e. g. happy, sad, angry, neutral) can be
detected in the course of the session but each emotion is likely to be relatively
mild and transitions from one to another are smooth.

3. Mild emotional lability is evident either because transitions are somewhat difﬁcult
or because the intensity of the emotions and the emotional shifts seem more
extreme.

4. High emotional lability is evident. Transitions seem difﬁcult and intense.
Emotions are expressed in an intense manner.

5. Extreme emotional lability. Child switches from one emotion to another quickly

and without a sense of smooth transition. You would describe this child as
“dysregulated”.

174

 

 

SELF-REGULATION/ORGANIZATIONAL CAPACITIES

This item assesses the extent to which the child is able to organize and regulate his/her
emotional state in response to internal or external stimulation or frustration (e.g. mother
leaving). At one extreme the child may appear disorganized (e. g. hyperactive, labile,
diffuse or unfocused) by the experience. On the other extreme the child is able to engage
in the world while continuing to maintain organization and state stability (e.g. shows
interest, focus, and an emotional resiliency to distress or frustration). The idea here is to
gauge the extent to which a child is able to regulate his/her emotions such that she can be
engaged with his/her environment (e. g. the toys in the playroom or the stranger or
mother). A dysregulated child may be throwing a temper tantrum (and receive a high
score on anger) or be crying intensely (and receive a high score on sadness) OR a child
may be wimpering softly but still be so “undone” by her emotions that he/she wanders
around the room aimlessly or sits in a chair without engaging fully in play, etc.

1. Child characteristically appears actively emotionally upset and disorganized OR
upset and disengaged from contact with the environment and unable to maintain
or regain focus or stable mood states. Child seems markedly unable to regulate
her emotional state.

2. Considerable difficulty regulating emotional state. Evidence of successful
attempts to stabilize mood on 1 or 2 occasions.

3. Moderate degree of difﬁculty maintaining self-regulation. Intense or prolonged
negative response to distress on several occasions; some evidence of attempts to
regulate affect.

4. Minimal difficulty. Child is able to quickly self regulate and/or refocus on most
occasions — or rarely becomes dysregulated.

5. Child characteristically maintains interest in his/her environment and appears very
resilient to stressful stimulation; maintains organized focus/approach.

175

 

REGULATION STRATEGIES

All children use various strategies to regulate themselves. Rate the degree to which the
child used each of these methods to self-regulate during this segment.

1 = not at all

2 = 1-2 times; or short duration — gpiplg attempts

3 = 3-4 times; or up to ‘/4 of the segment (= 45 seconds)

4 = 5-6 times; or up to 1/2 of the segment (= 90 seconds; 1’/2 minutes)
5 = 7 + times or more than V2 of the segment

A.

Contact seeking with stranger/mother. This includes looking at the adult, initiating
verbal conversations, going to the door to try to reunite with mother/ stranger,
looking in the window to see if she can see someone in there, etc.
0 Note that yelling for someone (most often yelling to mother) to come and
help or get them counts as contact seeking.

Attention to toy or self-play

Physical self stimulation (playing with hair, sucking thumb, rocking, physical
activity such as “pacing”, drinking from the baby bottle, tapping toys in a
repetitive manner with no obvious goal other than stimulation, manipulating toys
in hand in a non-functional manner)

. Talking or singing to self. This includes talking to the toys (e.g. comforting the

babydoll). All vocalizations count except calling for the adult in a manner which
indicates contact seeking.

Attempts to leave the room. This code captures the child’s attempts to physically
leave the room by getting out the door.

176

 

 

ATTENTIONAL REGULATION

This code captures the child’s ability to direct her attention to one stimulus and stay
focused on it without evidence of rapid shifting of her attention. A stimulus is deﬁned
here as an object (a toy, the water pipe in the room, etc.) or a person (stranger or mother),
or an activity (e. g. trying to get the childlock off of the doorknob so she can get out of the
room). Note that temper tantrums do NOT count as focused attention even though the
child may be engaging in the tantrum for the majority of the segment. During the
stranger/mother episodes we do NOT code shifts in attention which are the result of the
stranger/mother directing the child’s attention to a new toy or activity. Instead, we are
looking to capture the child’s own ability to stay focused on a task, stimulus, or
interaction without the need for frequent shifts. A child scoring low on this scale will
evince an ADHD behavioral quality. A child scoring high on this scale will be focused
and attentive to one stimulus for long periods of time.

Note that there is a quantitative aspect to attention shifts in that some shifts are
considered more major (e.g. moving from doll play with the doll house to putting
together the puzzle) and some are considered more minor (e. g. moving from putting
together the puzzle to trying to see if the puzzle pieces ﬁt in the hammer toy). Children
who are playing with a particular toy and attempt to draw the adult into their play (e.g.
“Look at this mom.”) are not considered to be engaging in a major shift of their attention.
However, a child who plays with a toy for a while and then turns to engage in a
conversation with the adult would be considered to evince a major switch of attention.

1. Child switches attention frequently throughout the segment. It seems as though he
can not settle on what he wants to do.

2. Child seems unfocused and unable to maintain attention on particular stimuli for
most of the segment. However, there are times - at lest 3-4 times; usually lasting
10 seconds or so per instance - when the child seems able to focus.

3. Child seems to be focused for about one half of the segment and relatively
unfocused for the other half.

4. Child is able to maintain attentional focus without difficulty for the majority of
the segment. There may be some switches in activity or attention given changes in
the environment or changes in the child’s interest but it is clear that the child can
maintain focus for extended periods of time.

5. Child maintains attentional focus on one activity for extended periods of the
segment. In fact, the child seems to be over-absorbed in one activity for most of
the session. Child is focused on one activity for the majority of the segment —
more than 2.5 minutes for a segment that lasts the entire 3 minutes.

177

 

ACTIVITY LEVEL

This code captures the child’s physical activity level throughout the segment. Gross
motor movements increase a child’s score on this scale more than ﬁne motor movements.
That is, a child who is running around the room is considered to be more active than a
child who is actively placing puzzle pieces in a puzzle. Thus, children who sit in one
place the entire time and only move to manipulate toys receive low scores on this scale.
Children who move around the room a great deal and/or attempt to leave the room
receive higher scores.

1.

Very Low activity level. Child seems somewhat lethargic and is not moving
around much. This child may be engaged in just one ﬂoor toy the entire time.
Child’s play is almost entirely ﬁne motor.

Low Activity level. Child is engaged in play but is mostly using ﬁne motor
movements.

Medium activity level. Play involving ﬁne motor about 1/2 of the time and gross
motor about '/2 of the time.

High activity level. This child’s actions mostly involve gross motor movements
although there may be some ﬁne motor or slower play activities as well.

Very High activity level. This child is actively moving around the room and
exhibiting lots of gross motor movement. His movements may seem forceful and
he may be physically manipulating toys, banging on the door, etc. You might
describe him as somewhat “hyper”.

178

 

 

REFERENCES

179

 

 

REFERENCES

Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-I8 and I 991
Profile. Burlington: University of Vermont, Department of Psychiatry.

Ainsworth, M. D. S. (1967). Infancy in Uganda: Infant care and the growth of love.
Baltimore: Johns Hopkins University Press.

Ainsworth, M. D. S. (1993). Attachment as related to mother-infant interaction. Advances
in Infancy Research, 8, 1-50.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of
attachment.“ A psychological study of the Strange Situation. Hillsdale, NJ:
Erlbaum.

Allison, P. D. (2003). Missing Data Techniques for Structural Equation Modeling.
Journal of Abnormal Psychology, 112(4), 545-557.

Altman, D. G. (1991 ). Practical statistics for medical research. London: Chapman and
Hall.

Anan, R. M., & Barnett, D. (1999). Perceived social support mediates between prior
attachment and subsequent adjustment: A study of urban African American
children. Developmental Psychology, 35(5), 1210-1222.

Association, A. P. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th
ed.). Washington, DC: American Psychiatric Association.

Bandura, A. (1977). Social Learning Theory. Englewood Cliffs, NJ : Prentice-Hall.

Bandura, A., & Walters, R. (1963). Social learning and personality development. New
York: Holt, Rinehart, and Winston, Inc.

Baurnrind, D. (1971). Current patterns of parental authority. Developmental Psychology
Monograph, 4(1), 1-103.

180

 

 

Beck, J. S. (1995). Cognitive Therapy: Basics and Beyond. New York: The Guilford
Press.

Belsky, J ., & F earon, R. M. P. (2002). Early attachment security, subsequent maternal
sensitivity, and later child development: Does continuity in development depend
upon continuity of caregiving? Attachment & Human Development, 4(3), 361-
387.

Belsky, J ., Fish, M., & Isabella, R. (1991). Continuity and discontinuity in infant negative
and positive emotionality: Family antecedents and attachment consequences.
Developmental Psychology, 2 7, 421-43 1.

Belsky, J ., & Rovine, M. (1987). Temperament and attachment security in the strange
situation: An empirical rapprochement. Child Development, 58, 787-795.

Belsky, J ., Youngblade, L., Rovine, M., & Volling, B. (1991). Patterns of marital change
and parent-child interaction. Journal of Marriage & the Family, 53, 487-498.

Biringen, Z. (2000). Emotional availability: Conceptualization and research ﬁndings.
American Journal of Orthopsychiatry, 70, 104-1 14.

Biringen, 2., Robinson, J. L., & Emde, R. N. (2000). Emotional Availability Scales, 3rd
edition. Attachment & Human Development, 2, 256-270.

Block, J ., & Block, J. (1980). The role of ego-control and ego-resiliency in the
organization of behavior. In W. A. Collins (Ed), Development of Cognition
Aﬁiact, and Social Relations (V 01. 13). Hillsdale, NJ: Lawrence Erlbaum.

Bogat, G. A., DeJonghe, E., Levendosky, A. A., Davidson, W. S., & vonEye, A. (2006).
Trauma symptoms among infants exposed to intimate partner violence. Child
Abuse & Neglect, 30, 109-125.

Booth, C. L., Rose-Krasnor, L., & Rubin, K. H. (1991). Relating preschoolers' social
competence and their mothers' parenting behaviors to early attachment security
and high-risk status. Journal of Social & Personal Relationships, 8, 363-382.

Bomstein, M. H. (2002). Handbook of Parenting (2nd ed.). Mahwah, NJ: Lawrence
Erlbaum Associates.

181

 

Bowlby, J. (1944). Forty-four juvenile thieves: Their character and home lives.
International Journal of Psycho-analysis, 25, 19-52.

Bowlby, J. (1969/1982). Attachment and loss, Vol. I : Attachment (2nd ed.). New York:
Basic Books.

Bowlby, J. (1980). Attachment and loss, Vol. 3: Loss, separation and depression. New
York: Basic Books.

Bowlby, J. (1988). A Secure Base. London: Basic Books.

Bradley, S. J. (2000). Affect Regulation and the Development of Psychopathology. New
York: Guilford Press.

Bretherton, I., & Munholland, K. (1999). Internal working models in attachment
relationships: A construct revisited. In J. Cassidy & P. R. Shaver (Eds),
Handbook of attachment (pp. 89-11 1). New York: The Guildford Press.

Burgess, K. B., Marshall, P. J ., Rubin, K. H., & Fox, N. A. (2003). Infant attachment and
temperament as predictors of subsequent externalizing problems and cardiac
physiology. Journal of Child Psychology and Psychiatry, 44(6), 819-831.

Calkins, S. D. (2004). Early Attachment Processes and the Development of Emotional
Self-Regulation. In R. F. Baumeister & K. D. Vohs (Eds), Handbook of Self-

Regulation: Research, Theory, and Applications (pp. 324-339). New York:
Guilford Press.

Calkins, S. D., & Fox, N. A. (1992). The relations among infant temperament, security of
attachment, and behavioral inhibition at twenty-four months. Child Development,
63, 1456-1472.

Campbell-Sills, L., & Barlow, D. H. (2007). Incorporating Emotion Regulation into
Conceptualizations and Treatments of Anxiety and Mood Disorders. In J. Gross,

J. (Ed.), Handbook of Emotion Regulation (pp. 542-559). New York: The
Guilford Press.

Cannon, W. (1929). Bodily changes in pain, hunger, ﬁear and rage: An account of recent
research into the function of emotional excitement (2nd ed.). New York:
Appleton-Century-Crof’ts.

182

 

 

Cassidy, J. (1994). Emotion Regulation: Inﬂuences of Attachment Relationships.
Monographs of the Society for Research in Child Development, 5 9(2/3), 228-249.

Cassidy, J. (1999). The nature of the child's ties. In J. Cassidy & P. Shaver (Eds),

Handbook of A ttachment: Theory, Research and Clinical Applications (pp. 3-20).

New York: Guilford Press.

Cassidy, J ., & Shaver, P. (Eds). (1999). Handbook of Attachment: Theory, Research,
and Clinical Applications. New York: Guilford Press.

Ceci, S. J., Kulkofsky, S., Klemfuss, J. 2., Sweeney, C. D., & Bruck, M. (2007).
Unwarranted assumptions about children's testimonial accuracy. Annual Review
of Clinical Psychology, 3, 311-328.

Clark, R. (1985). The Parent-Child Early Relational Assessment: Instrument and
Manual. Department of Psychiatry, University of Wisconsin Medical School,
Madison, WI.

Clark, R. (1999). The parent-child early relational assessment: A factorial validity study.
Educational and Psychological Measurement, 59(5), 821-846.

Cohen, J. (1968). Weighted kappa: Nominal scale agreement provision for scaled
disagreement or partial credit. Psychological Bulletin, 70(4), 213-220.

Cole, P., Barrett, K. C., & Zahn-Waxler, C. (1992). Emotion displays in two-year-olds
during mishaps. Child Development, 63(2), 314-324.

Cole, P., Martin, S. E., & Dennis, T. A. (2004). Emotion regulation as a scientiﬁc
construct: Methodological challenges and directions for child development
research. Child Development, 75(2), 317-333.

Collins, L. M., Schafer, J. L., & Karn, C.-M. (2001). A comparison of inclusive and

restrictive strategies in modern missing data procedures. Psychological Methods,
6(4), 330-351.

Collins, N. L., & Read, S. J. (1994). Cognitive representations of attachment: The
structure and function of working models. In K. Bartholomew & D. Perlman

(Eds), Attachment Processes in Adulthood (pp. 53-90). Bristol, PA: Jessica
Kingsley Publishers.

183

 

J

Collins, W. A., Madsen, S. D., & Susman-Stillman, A. (2002). Parenting during middle
childhood. In M. H. Bomstein (Ed), Handbook of Parenting (2nd ed., Vol. 1:
Children and Parenting). New Jersey: Lawrence Erlbaum Associates.

Contreras, J. M., Kems, K. A., Weimer, B. L., Gentzler, A. L., & Tomich, P. L. (2000).
Emotion regulation as a mediator of associations between mother-child

attachment and peer relationships in middle childhood. Journal of Family
Psychology, 14(1), 111-124.

Crandell, L. E., Fitzgerald, H. E., & Whipple, E. E. (1997). Dyadic synchrony in parent-
child interactions: A link with maternal representations of attachment
relationships. Infant Mental Health Journal, 18(3), 247-264.

Crittenden, P. M. (1990). Internal representational models of attachment relationships.
Infant Mental Health Journal, 11(3), 259-277.

Crittenden, P. M. (1994). The preschool Assessment of Attachment coding manual.
Miami, FL: Family Relations Institute.

Crittenden, P. M. (1995). Attachment and Psychopathology. In S. Goldberg, R. Muir & J.
Kerr (Eds), Attachment Theory: Social, Developmental, and Clinical
Perspectives (pp. 367-406). Hillsdale, NJ: Analytic Press.

Cusinato, M. (1998). Parenting styles and psychopathology. In L. L'Abate (Ed.), Family
Psychopathology: The Relational Roots of Dysfunctional Behavior (pp. 158-179).
New York: Guilford Press.

Dallaire, D. H., & Weinraub, M. (2005). The stability of parenting behaviors over the
ﬁrst 6 years of life. Early Childhood Research Quarterly, 20, 201-219.

Darwin, C. (1872/1965). The expression of the emotions in man and animals. Chicago:
University of Chicago Press.

David-Ferdon, C., & Kaslow, N. J. (2008). Evidence-Based Psychosocial Treatments for
Child and Adolescent Depression. Journal of Clinical Child and Adolescent
Psychology, 37(1), 62-104.

Davies, P. T., & Cummings, E. M. (1994). Marital conﬂict and child adjustment: An
emotional security hypothesis. Psychological Bulletin, 116(3), 387-411.

184

 

 

Dayton, C. J ., Levendosky, A. A., Davidson, W. S., & Bogat, G. A. (in press). The child
as held in the mind of the mother: The impact of prenatal maternal
representations on parenting behaviors. Infant Mental Health Journal.

Deater-Deckard, K., Pike, A., Petrill, S. A., Cutting, A. L., Hughes, C., & O'Connor, T.
G. (2001). Nonshared environmental processes in social-emotional development:
An observational study of identical twin differences in the preschool period.
Developmental Science, 4(2), F1-F6.

DeJonghe, E. S., Bogat, G. A., Levendosky, A. A., Eye, A. v., & Davidson, W. S. (2005).
Infant exposure to domestic violence predicts heightened sensitivity to adult
verbal conﬂict. Infant Mental Health Journal, 26(3), 268-281.

DeVoe, E., & Smith, E. L. (2002). The impact of domestic violence on urban preschool
children: Battered mothers' perspectives. Journal of Interpersonal Violence,
17(10), 1075-1101.

DeWolff, M., & Ijzendoom, M. H. v. (1997). Sensitivity and attachment: A meta-
analysis on parental antecedents of infant attachment. Child Development, 68(4),
571-591.

Dickstein, S., Thompson, R. A., Estes, D., Malkin, C., & Lamb, M. E. (1984). Social
referencing and the security of attachment. Infant Behavior and Development,
7(4), 507-516.

Dix, T. (1991). The affective organization of parenting: Adaptive and maladaptive
processes. Psychological Bulletin, 110(1), 3-25.

Eisenberg, N. (2002). Emotion-related regulation and its relation to quality of social
functioning. In W. Hartup & R. Weinberg (Eds), The Minnesota symposia on
child psychology. Child psychology in retrospect and prospect: In celebration of
the 75th anniversary of the Institute of Child Development. (V 01. 32, pp. 133-
171). Mahwah, NJ: Lawrence Erlbaum Associates.

Eisenberg, N., Cumberland, A., Spinrad, T. L., Fabes, R. A., Shepard, S. A., Reiser, M.,
et a1. (2001). The Relations of Regulation and Emotionality to Children's

Extemalizing and Intemalizing Problem Behavior. Child Development, 72(4),
1112-1134.

185

 

 

Eisenberg, N., & Spinrad, T. L. (2004). Emotion-Related Regulation: Sharpening the
Deﬁnition. Child Development, 75(2), 334-3 39.

Erickson, M. F., Sroufe, L. A., & Egeland, B. (1985). The relationship between quality of
attachment and behavior problems in preschool in a high-risk sample.
Monographs of the Society for Research in Child Development, 5 0(1/2), 147-166.

Eyberg, S. M., & Robinson, E. A. (1982). Parent-child interaction training: Effects on
family functioning. Journal of Clinical Child Psychology, 11(2), 130-137.

Fantuzzo, J. W., & Mohr, W. K. (1999). Prevalence and effects of child exposure to
domestic violence. The Future of Children, 9(3), 21-32.

Ferster, C. B., & Skinner, B. F. (1957). Schedules of reinforcement. New York:
Appleton-Century-Crofts.

Fonagy, P., Gergely, G., Jurist, E., & Target, M. (2004). Affect Regulation, Mentalization,
and the Development of the Self New York: Other Press.

Fox, R. A. (1994). Parent Behavior Checklist. Brandon, VT: Clinical Psychology Co.

Fraiberg, S. H. (1987). Selected writings of Selma Fraiberg. Columbus, OH: Ohio State
University Press.

Frodi, A., & Thompson, R. (1985). Infants’ affective responses in the strange situation:
Effects of prematurity and of quality of attachment. Child Development, 56(5),
1280-1290.

Fuendeling, J. M. (1998). Affect regulation as a stylistic process within adult attachment.
Journal of Social & Personal Relationships, 15(3), 291-322.

Gaensbauer, T. J. (1995). Trauma in the preverbal period: Symptoms, memories, and
developmental impact. Psychoanalytic Study of the Child, 50, 122-433.

Ganiban, J ., Barnett, D., & Ciccetti, D. (2000). Negative reactivity and attachment: Down
syndrome's contribution to the attachment-temperament debate. Development and
Psychopathology, 12, 1-21.

186

 

Gilliom, M., Shaw, D. 8., Beck, J. E., Schonberg, M. A., & Lukon, J. L. (2002). Anger
regulation in disadvantaged preschool boys: Strategies, antecedents, and the
development of self-control. Developmental Psychology, 38(2), 222-235.

Graham-Bermann, S. A., Lynch, 8., Banyard, V., DeVoe, E. R., & Halabu, H. (2007).
Community-based intervention for children exposed to intimate partner violence:
An efﬁcacy trial. Journal of Consulting and Clinical Psychology, 75(2), 199-209.

Gray, J. A. (1988). The psychology of fear and stress (2nd ed.). Cambridge: Cambridge
University Press.

Greenberg, M. T. (1999). Attachment and psychopathology in childhood. In J. Cassidy &
P. Shaver (Eds), Handbook of Attachment: Theory, Research, and Clinical
Applications (pp. 469-496). New York: Guilford Press.

Gross, J ., J. (Ed.). (2007). Hanbook of Emotion Regulation. New York: The Guilford
Press.

Gross, J ., J ., & Thompson, R. (2007). Emotion Regulation: Conceptual Foundations. In J.
Gross, J. (Ed.), Handbook of Emotion Regulation. New York: The Guilford Press.

Grossmann, K. E., Grossmann, K., & Waters, E. (2005). Attachment from Infancy to
Adulthood: The Major Longitudinal Studies. New York: Guilford Press.

Gunnar, M., Mangelsdorf, S., Larson, M., & Hertsgaard, L. (1989). Attachment,
Temperament, and Adrenocortical Activity in Infancy: A Study of
Psychoendocrine Regulation. Developmental Psychology, 25, 355-363.

Guttmann-Steinmetz, S., & Crowell, J. A. (2006). Attachment and externalizing

disorders: A developmental psychopathology perspective. Journal of the
American Academy of Child & Adolescent Psychiatry, 45(4), 440-451.

Hagekull, B., & Bohlin, G. (2004). Predictors of middle childhood psychosomatic
problems: An emotion regulation approach. Infant and Child Development, 13,
389-405.

Herman, J. L. (1992). Trauma and recovery. New York: Basic Books.

187

 

Hill-Soderlund, A. L., Mills-Koonce, W. R., Propper, C., Calkins, S. D., Granger, D. A.,
Moore, G. A., et al. (2008). Parasympathetic and sympathetic responses to the
strange situation in infants and motehrs from avoidant and securely attached
dyads. Developmental Psychobiology, 50(4), 361-376.

 

Hill, A. L., Degnan, K. A., Calkins, S. D., & Keane, S. P. (2006). Proﬁles of
externalizing behavior problems for boys and girls across preschool: The roles of
emotion regulation and inattention. Developmental Psychology, 42(5), 913-928.

Hoeksma, J. B., Oosterlaan, J ., & Schipper, E. M. (2004). Emotion regulation and the
dynamics of feelings: A conceptual and methodological framework. Child
Development, 75(2), 354-360.

Holden, G. W., & Miller, P. C. (1999). Enduring and different: A meta-analysis of the
similarity in parents' child rearing. Psychological Bulletin, 125(2), 223-254.

Holden, G. W., & Ritchie, K. L. (1991). Linking extreme marital discord, child rearing,
and child behavior problems: Evidence from battered women. Child
Development, 62(2), 31 1-327.

Huth Books, A. C., Levendosky, A. A., Theran, S. A., & Bogat, G. A. (2004). The impact
of domestic violence on mothers' prenatal representations of their infants, Infant
Mental Health Journal.

Jaﬁ‘ee, S. R., Mofﬁtt, T. E., Caspi, A., Taylor, A., & Arseneault, L. (2002). Inﬂuence of
adult domestic violence on children's internalizing and externalizing problems:
An environmentally informative twin study. Journal of the American Academy of
Child & Adolescent Psychiatry, 41(9), 1095-1 103.

James, W. (1884). What is an emotion? Mind, 9, 188-205.

Javo, C., Ronning, J. A., Heyerdahl, S., & Rudmin, F. W. (2004). Parenting correlates of
child behavior problems in a multiethnic community sample of preschool children
in northern Norway. European Child & Adolescent Psychiatry, 13(1), 8-18.

Katz, L. F., & Windecker-Nelson, B. (2006). Domestic violence, emotiona coaching, and
child adjustment. Journal of Family Psychology, 20(1), 56-67.

188

 

Keller, T. E., Spieker, S. J ., & Gilchrist, L. (2005). Patterns of risk and trajectories of
preschool problem behaviors: A person-oriented analysis of attachment in
context. Development and Psychopathology, 1 7, 349-3 84.

Keown, L. J ., & Woodward, L. J. (2006). Preschool boys with pervasive hyperactivity:
Early peer functioning and mother-child relationship inﬂuences. Social
Development, 15(1), 23-45.

Kems, K. A., Klepac, L., & Cole, A. (1996). Peer relationships and preadolescents'
perceptions of security in the child-mother relationship. Developmental
Psychology, 32, 457-466.

Kitzmann, K. M., Gaylord, N. K., Holt, A. R., & Kenny, E. D. (2003). Child witnesses to
domestic violence: A meta-analytic review. Journal of consulting and clinical
psychology, 71(2), 339-352.

Kochanska, G. (2001). Emotional development in children with different attachment
histories: The ﬁrst three years. Child Development, 72(2), 474-490.

Kopp, C. B. (1989). Regulation of distress and negative emotions: A developmental
view. Developmental Psychology, 25 , 343-354.

Krystal, H. (1988). Integration and Self-Healing: A ﬂeet, Trauma, Alexithymia. Hillsdale,
NJ: The Analytic Press.

LaBar, K. S., Gatenby, J. C., Gore, J. C., LeDoux, J. E., & Phelps, E. A. (1998). Human
amygdala activation during conditioned fear acquisition and extinction: A mixed-
trial ﬂVIRI study. Neuron, 20, 937-945.

Lahey, B. B., Rathouz, P. J ., Van Hulle, C., Urbano, R. C., Krueger, R. F., Applegate, B.,
et al. (2008). Testing structural models of DSM-IV symptoms of common forms

of child and adolescent psychopathology. Journal of Abnormal Child Psychology,
36(2), 187-206.

LeDoux, J. (1996). The Emotional Brain: The Mysterious Underpinnings of Emotional
Life. New York: Simon & Schuster.

LeDoux, J. (2000). Emotion Circuits in the Brain. Annual Review of Neuroscience, 23,
155-184.

189

 

 

 

Levendosky, A. A., & Graham-Bermann, S. A. (2000). Behavioral observations of
parenting in battered women. Journal of Family Psychology, 14(1), 1-15.

Levendosky, A. A., & Graham-Bermann, S. A. (2001a). Parenting in battered women:
The effects of domestic violence on women and their children. Journal of Family
Violence.

Levendosky, A. A., & Graham-Bermann, S. A. (2001b). Trauma and parenting in
battered women: An addition to an ecological model of parenting. journal of
Aggression, Maltreatment & Trauma, 3(1), 25-35.

Levendosky, A. A., Huth-Bocks, A. C., Semel, M. A., & Shapiro, D. L. (2002). Trauma
symptoms in preschool-age children exposed to domestic violence. Journal of
Interpersonal Violence, 17(2), 150-164.

Levendosky, A. A., Huth-Bocks, A. C., Shapiro, D. L., & Semel, M. A. (2003). The
impact of domestic violence on the matemal-child relationship and preschool-age
children's functioning. Journal of Family Psychology, 1 7(3), 275-287.

Levendosky, A. A., Lynch, S. L., & Graham-Bermann, S. A. (2000). Mothers'
perceptions of the impact of woman abuse on their parenting. Violence Against
Women, 6(3), 247-271.

Lewis, M. D., Granic, I., Lamm, C., Zelazo, P. D., Stieben, J ., Todd, R. M., et al. (2008).
Changes in the neural bases of emotion regulation associated with clinical

improvement in children with behavior problems. Development and
Psychopathology, 20(3), 913-939.

Lieberman, A. F ., & Amaya-Jackson, L. (2005). Reciprocal Inﬂuences of Attachment and
Trauma. In L. Berlin, Y. Ziv, L. Amaya—Jackson & M. T. Greenberg (Eds),
Enhansing Early Attachments (pp. 100-124). New York: The Guilford Press.

Lieberman, A. F., VanHom, P., & Ozer, E. J. (2005). Preschooler witnesses of marital
violence: Predictors and mediators of child behavior problems. Development and
Psychopathology, 1 7, 385-396.

Luminet, O., Rime, B., Bagby, R. M., & Taylor, G. J. (2004). A multimodal investigation
of emotional responding in alexithymia. Cognition & Emotion, 18(5), 741-766.

190

 

 

Lyons-Ruth, K., Easterbrooks, M. A., & Cibelli, C. D. (1997). Infant attachment
strategies, infant mental lag, and maternal depressive symptoms: Predictors of

internalizing and externalizing problems at age 7. Developmental Psychology,
33(4), 681-692.

Lyons-Ruth, K., & Jacobvitz, D. (1999). Attachment disorganization: Unresolved loss,
relational violence, and lapses in behavioral and attentional strategies. In J.
Cassidy & P. Shaver (Eds), Handbook of Attachment: Theory, Research, and
Clinical Applications (pp. 520-554). New York: Guilford Press.

Magai, C. (1999). Affect, imagery, and attachment: Working models of interpersonal
affect and the socialization of emotion. In J. Cassidy & P. Shaver (Eds),
Handbook of A ttachment: Theory, Research and Clinical Applications (pp. 787-
802). New York: Guilford Press.

Main, M. (1995). Recent studies in attachment: Overview with selected implications for
clinical work. In S. Goldberg, R. Muir & J. Kerr (Eds), Attachment Theory:
Social, Developmental, and Clinical Perspectives. Hillsdale, NJ: The Analytic
Press.

Main, M., Kaplan, K., & Cassidy, J. (1985). Security in infancy, childhood and
adulthood: A move to the level of representation. In I. Bretherton & E. Waters
(Eds), Growing points of attachment theory and research. Monographs of the
Society for Research in Child Development (Vol. 50, pp. 66-104).

Main, M., & Solomon, J. (1986). Discovery of a new, insecure-disorganized/disoriented
attachment pattern. In T. B. Brazelton & M. Q. Yogman (Eds), Aﬁéctive
Development in Infancy. Norwood, NJ: Ablex.

Main, M., & Solomon, J. (1990). Procedures for identifying infants as
disorganized/disoriented during the Ainsworth Strange Situation. In M. T.
Greenberg, D. Cicchetti & E. M. Cummings (Eds), Attachment in the Preschool
Years: Theory, Research and Intervention (pp. 121-160). Chicago: University of
Chicago Press.

Mangelsdorf, S., Gunnar, M., Kestenbaum, R., Lang, 8., & Andreas, D. (1990). Infant
proneness-to-distress temperament, maternal personality, and mother-infant
attachment: Associations and goodness of ﬁt. Child Development, 61, 820-831.

Maren, S. (2001). Neurobiology of Pavlovian fear conditioning. Annual Reveiw of
Neuroscience, 24, 897-931.

191

 

 

Margolin, G., & Gordis, E. (2003). Co-occurrence between marital aggression and
parents' child abuse potential: The impact of cumulative stress. Violence and
Victims, 18(3), 243-258.

Marshall, L. L. (1992). Development of the Severity of Violence Against Women Scales.
Journal of Family Violence, 7(2), 103-121.

Martinez-Torteya, C., Bogat, G. A., Von Eye, A., & Levendosky, A. A. (in press).
Resilience among children exposed to domestic violence: the role of risk and
protective factors. Child Development.

McCloskey, L. A., Figueredo, A. J ., & Koss, M. P. (1995). The effects of systemic family
violence on children's mental health. Chld Development, 66(5), 1239-1261.

Mikulincer, M., Shaver, P. R., & Pereg, D. (2003). Attachment theory and affect
regulation: The dynamics, development and cognitive consequences of
attachment-related strategies. Motivation & Emotion, 27(2), 77-102.

Miller, A. L., Gouley, K. K., Seifer, R., Dickstein, S., & Shields, A. (2004). Emotions
and behaviors in the Head Start classroom: Associations among observed

dysregulation, social competence, and preschool adjustment. Early Education &
Development, 15(2), 147-165.

Mohr, W. K., & Fantuzzo, J. W. (2000). The neglected variable of physiology in
domestic violence. Journal of Aggression, Maltreatment & Trauma, 3(1), 69-84.

Morewitz, S. J. (2004). Domestic Violence and Maternal and Child Health. New York:
Kluwer Academic/Plenum Publishers.

Mullin, B. C., & Hinshaw, S. P. (2007). Emotion regulation and externalizing disorders
in children and adolescents. In J. Gross, J. (Ed.), Handbook of Emotion
Regulation (pp. 523-541). New York: The Guilford Press.

Murray, K. T., & Kochanska, G. (2002). Effortful control: Factor structure and relation to

externalizing and internalizing behaviors. Journal of A bnormal Child Psychology,
30(5), 503-514.

Muthen, L. K., & Muthen, B. O. (2007). Mplus User's Guide, 4th edition (Version 4.1).
Los Angeles, CA: Muthen & Muthen.

192

 

National-Center-for-Clinical-Infant-Programs, Z.-t.-T. (1994). Diagnostic Classification
of Mental Health and Developmental Disorders of Infancy and Early Childhood.
Arlington, VA: Zero to Three: National Center for Clinical Infant Programs.

Nigg, J. T. (2006). Temperament and developmental psychopathology. Journal of Child
Psychology and Psychiatry, 3 7(3/4), 395-422.

Osofsky, J. D. (1999). The impact of violence on children. The Future of Children, 9(3),
33-49.

Osofsky, J. D. (Ed.). (2004). Young Children and Trauma. New York: Guilford Press.

Phan, K. L., Fitzgerald, D. A., Nathan, P. J ., Moore, G. J ., Uhde, T. W., & Tancer, M. E.
(2005). Neural substrates fro voluntary suppression of negative affect: A
functional magnetic resonance imaging study. Biological Psychiatry, 5 7, 210-219.

Phillips, M. L., Drevets, W. C., Rauch, S. L., & Lane, R. (2003). Neurobiology of
emotion perception I: The neural basis of normal emotion perception. Biological
Psychiatry, 54, 504-514.

Posner, M. I., & DiGirolamo, G. J. (2000). Cognitive neuroscience: Origins and promise.
Psychological Bulletin, 126, 873-889.

Pynoos, R. S., & Eth, S. (1986). Witness to violence: The child interview. Journal of the
American Academy of Child Psychiatry, 25(3), 306-319.

Rescorla, L., Achenbach, T., Ivanova, M. Y., Dmnenci, L., Almqvist, F., Bilenberg, N., et
al. (2007). Behavioral and emotional problems reported by parents of children
ages 6 to 16 in 31 societies. Journal of Emotional and Behavioral Disorders,
15(3), 130-142.

Rollins, B. C., & Thomas, D. L. (1979). Parental support, power, and control techniques
in the socialization of children. In W. R. Burr, R. Hill, F. I. Ney & I. L. Reiss

(Eds), Contemporary theories about the family (Vol. 1, pp. 317-364). New York:
Free Press.

Rothbart, M. K., & Bates, J. E. (1998). Temperament. In N. Eisenberg (Ed.), Handbook

of child Psychology: Social, emotional, and personality development (V 01. 3, pp.
105-1 76). New York: Wiley.

193

 

 

Rothbart, M. K., Ellis, L. K., Rueda, M. R., & Posner, M. I. (2003). Developing
mechanisms of temperamental effortful control. Journal of Personality and Social
Psychology, 71, 1 113-1143.

 

Rothbart, M. K., Posner, M. 1., & Kieras, J. (2006). Temperament, Attention, and the
Development of Self-Regulation. In K. McCartney & D. Phillips (Eds),
Blackwell handbook of early childhood development. (pp. 338-357). Malden, MA .
US: Blackwell Publishing.

Rubin, K. H., & Burgess, K. B. (2002). Parents of aggressive and withdrawn children. In
M. H. Bomstein (Ed.), Handbook of Parenting (2nd ed., Vol. 1: Children and
Parenting, pp. 383-418). Mahwah, New Jersey: Lawrence Erlbaum Associates.

Rueda, M. R., Posner, M. I., & Rothbart, M. K. (2005). The development of executive
attention: Contributions to the emergence of self-regulation. Developmental
Neuropsychology, 28(2), 573-594.

Saltzman, K. M., Holden, G. W., & Holahan, C. J. (2005). The psychobiology of children
exposed to marital violence. Journal of Clinical Child and Adolescent
Psychology, 34(1), 129-139.

Sameroff, A. J. (1993). Models of development and developmental risk. In C. H. Zeanah
(Ed.), Handbook of Infant Mental Health (pp. 3-13). New York: Guilford Press.

 

Sameroff, A. J ., & Fiese, B. H. (2000). Models of Development and Developmental Risk.
In C. Zeanah, H. (Ed.), Handbook of Infant Mental Health (2nd ed., pp. 3-19).
New York: The Guilford Press.

Sameroff, A. J ., Lewis, M., & Miller, S. M. (Eds). (2000). Handbook of Developmental
Psychopathology. New York: Kluwer Academic/Plenum Publishers.

Scheeringa, M. S., & Gaensbauer, T. J. (2000). Posttraumatic stress disorder. In C. H.

Zeanah (Ed.), Handbook of Infant Mental Health (pp. 369-381). New York: Basic
Books.

Scheeringa, M. S., Zeanah, C. H., Drell, M. J ., & Larrieu, J. A. (1995). Two approaches
to the diagnosis of posttraumatic stress disorder in infancy and early childhood.
Journal of the American Academy of Child & Adolescent Psychiatry, 34, 191-200.

194

Schore, A. N. (2003). Aﬁ’ect Dysregulation and Disorders of the Self New York: w. w.
Norton & Company.

Schuster, C. (2004). A Note on the Interpretation of Weighted Kappa and its Relations to
Other Rater Agreement Statistics for Metric Scales. Educational and
Psychological Measurement, 64(2), 243-253.

Seifer, R. (2000). Temperament and Goodness of Fit: Implications for Developmental
Psychopathology. In A. J. Sameroff, M. Lewis & S. M. Miller (Eds), Handbook
of Developmental Psychopathology (2nd ed., pp. 257-276). New York: Kluwer
Academic/Plenum Publishers.

Seifer, R., Sameroff, A. J ., Dickstein, S., Schiller, M., & Hayden, L. (2004). Your own
children are special: Clues tot he sources of reporting bias in temperament
assessments. Infant Behavior & Development, 27, 323-341.

Shavers, C. A., Levendosky, A. A., Dubay, S. M., Basu, A., & Jenei, J. (2005). Domestic
Violence Research: Methodological Issues Related to a Community-Based

Intervention With a Vulnerable Population. Journal of Applied Biobehavioral
Research, 10(1), 27-38.

Shaw, D. S., Keenan, K., Vondra, J. 1., Delliquadri, E., & Giovannelli, J. (1997).
Antecedents of preschool children's internalizing problems: A longitudinal study

of low-income families. Journal of the American Academy of Child & Adolescent
Psychiatry, 36(12), 1760- 1 767.

Shaw, D. S., Owens, E. B., Vondra, J. 1., Keenan, K., & Winslow, E. B. (1996). Early
risk factors and pathways in the development of early disruptive behavior
problems. Development and Psychopathology, 8, 679-699.

Simpson, J. A. (1999). Attachment theory in modern evolutionary perspective. In J.
Cassidy & P. Shaver (Eds), Handbook of Attachment: Theory, Research and
Clinical Applications. New York: Guilford Press.

Slade, A., Belsky, J ., Aber, J. L., & Phelps, J. L. (1999). Mothers' representations of their
relationships with their toddlers: Links to adult attachment and observed
mothering. Developmental Psychology, 35(3), 611-619.

Small, M. F. (1998). Our Babies, Ourselves. New York: Anchor Books.

195

 

Solomon, J ., & George, C. (1999). Attachment disorganization. New York: The Guilford
Press.

Sotres-Bayon, F., Bush, D. E. A., & LeDoux, J. E. (2004). Emotional perseveration: An
update on prefrontal-amygdala interactions in fear extinction. Learning &
Memory, 11(5), 525-535.

SPSS, I. (2006). SPSS (Version 15). Chicago: SPSS, Inc.

Sroufe, L. A. (1985). Attachment classiﬁcation from the perspective of infant-caregiver
relationships and infant temperament. Child Development, 56, 1-14.

Sroufe, L. A. (1995). Emotional development: The organization of emotional life in the
early years. New York: Cambridge University Press.

Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The Development of
the Person: The Minnesota Study of Risk and Adaptation ﬁ'om Birth to
Adulthood. New York: Guilford Press.

Sroufe, L. A., & Waters, E. (1977). Attachment as an organizational construct. Child
Development, 48, 1 184-1 199.

Stansbury, K., & Sigman, M. (2000). Responses of preschoolers in two frustrating

episodes: Emergence of complex strategies for emotion regulation. Journal of
Genetic Psychology, 16(2), 182-202.

Stern, D. N. (1985). The Interpersonal World of the Infant: A View ﬁ'om Psychoanalysis
and Developmental Psychology. New York: Basic Books.

Stieben, J ., Lewis, M. D., Granic, I., Zelazo, P. D., Segalowitz, S., & Pepler, D. (2007).
Neurophysiological mechanisms of emotion regulation for subtypes of
externalizing children. Development and Psychopathology, 19(2), 455-480.

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

196

 

 

 

Theran, S. A., Levendosky, A. A., Bogat, G. A., & Huth-Bocks, A. C. (2005). Stability
and change in mothers' internal representations of their infants over time.
Attachment & Human Development, 7(3), 253-268.

vandenBoom, D. C. (1994). The inﬂuence of temperament and mothering on attachment
and exploration: An experimental manipulation of sensitive responsiveness
among lower-class mothers with irritable infants. Child Development, 65, 1457-
1477.

vanlJzendoom, M. H., & Sagi, A. (1999). Cross-cultural patterns of attachment:
Universal and contextual dimensions. In J. Cassidy & P. Shaver (Eds), Handbook
of attachment: Theory, Research and Clinical Applications (pp. 713-734). New
York: Guilford Press.

Vondra, J. 1., Shaw, D. S., Swearingen, L., Cohen, M., & Owens, E. B. (2001).
Attachment stability and emotional and behavioral regulation from infancy to
preschool age. Development and Psychopathology, 13, 13-33.

Wadsworth, M. E., & Achenbach, T. M. (2005). Explaining the Link Between Low
Socioeconomic Status and Psychopathology: Testing Two Mechanisms of the

Social Causation Hypothesis. Journal of Consulting and Clinical Psychology,
73(6), 1146-1153.

Wearden, A., Cook, L., & Vaughan-Jones, J. (2003). Adult attachment, alexithymia,
symptom reporting, and health-related coping. Journal of Psychosomatic
Research, 55, 341-347.

Weinﬁeld, N. S., Sroufe, L. A., Egeland, B., & Carlson, E. A. (1999). The nature of
individual differences in infant-caregiver attachment. In J. Cassidy & P. Shaver
(Eds), Handbook of Attachment: Theory, Research, and Clinical Applications.
New York: Guilford Press.

Whipple, E. E., Denburg, N. L., & Davies, W. H. (1993). MSU Family Project Second
Revision of Belsky's Parent-Child Interaction Coding System. Unpublished
Coding Manual. Michigan State University.

Whipple, E. E., Fitzgerald, H. E., & Zucker, R. A. (1995). Parent-child interactions in
alcoholic and nonalcoholic families. American Journal of Orthopsychiatry, 65 ,
153-159.

197

 

Widaman, K. F. (2006). Missing Data: What to do with or without them. In K.
McCartney, M. Burchinal & K. L. Bub (Eds), Monograpsh of the Society for
Research in Child Development: Best Practices in Quantitative Methods for
Developmentalists (pp. 42-64).

Winnicott, D. W. (1965). The maturational processes and the facilitating environment:
Studies in the theory of emotional development. New York: International
Universities Press.

Wood, J. J ., Emmerson, N. A., & Cowan, P. A. (2004). Is early attachment security
carried forward into relationships with preschool peers? British Journal of
Developmental Psychology, 22, 245-253.

Zimmermann, L. K., & Stansbury, K. (2003). The inﬂuence of temperamental reactivity
and situational context on the emotion-regulatory abilities of 3-year-old children.
Journal of Genetic Psychology, 164(4), 389-409.

Zimmermann, P. (1999). Structure and functions of internal working models of
attachment and their role for emotion regulation. Attachment & Human
Development, 1(3), 291-306.

198

ATE U

llllllWH||lllllllllllllllllllllIES
03062 8022

1293

 

"11111111