 

 

LIBRARY

Michigan State

University

 

 

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

 

“DEC 1 4 2005

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuepGS-p. 15

 

MOTHER-INFANT ATTACHMENT: THE IMPACT OF MATERNAL
REPRESENTATIONS DURING PREGNANCY, MATERNAL RISK FACTORS,
AND SOCIAL SUPPORT
By

Alissa Christine Huth-Bocks

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Psychology

2002

ABSTRACT
MOTHER-INFANT ATTACHMENT: THE IMPACT OF MATERNAL
REPRESENTATIONS DURING PREGNANCY, MATERNAL RISK FACTORS,
AND SOCIAL SUPPORT
By

Alissa Christine Huth-Bocks

The mother-child relationship during infancy is considered to be one of the
most important determinants of a child’s social and emotional development by
clinical and developmental theorists. Because mother-infant attachment is so central to
children’s well-being, it is critical to understand the mechanisms through which mother-
infant attachment is formed. The theoretical framework for this study is based on both
attachment theory and Daniel Stern’s (1995) recent formulation about the transition to
motherhood. It examines the processes through which mother-infant attachment is
formed, by investigating the effects of both individual and environmental factors, in a
diverse and high-risk sample using a longitudinal research design. Two hundred and
seven women were interviewed during their last trimester of pregnancy, two months alter
birth, and one year after birth Participants were recruited throughout several counties in
mid-Michigan. During pregnancy, mothers’ own experiences with caregivers during
childhood were assessed, as were mothers’ thoughts and feelings about themselves as
mothers and their soon-to-be infants. Mothers’ perceived level of social support and
experiences with violence, depression, and poverty were assessed at multiple time points.
Mother-infant attachment was assessed at 1 year postpartum using a well-known

laboratory procedure. Structural equation modeling results indicated that individual and

environmental variables measured during pregnancy and after birth had a signiﬁcant
impact on mother-infant attachment. The most important predictors were mothers’ own
attachment experiences during childhood, representations of the infant-to-be during
pregnancy, and psychosocial risk factors. The results from this study expand the
current level of understanding regarding the development of different patterns of
mother-infant attachment. The results also have important clinical implications for

the treatment of mothers and young children at risk for psychopathology.

Dedicated to the mothers and babies who participated in this study and taught me so
much, to the little ones in my family who have given me so much joy, and to my

husband who has loved and supported me throughout this work.

iv

ACKNOWLEDGEMENTS

This dissertation would not Imve been possible without the care and assistance of
many, many people. I would ﬁrst like to express my deepest appreciation for my mentor,
Alytia Levendosky, who has given generously in all ways during my six years of
graduate school. In addition to speciﬁc guidance and support for this dissertation, she has
done many things to help me grow professionally and personally—far too many to list
here. I could not have asked for a better teacher, supervisor, and ﬁ'iend, and I am very
thankful to have spent so much time with her over these years. It has really been a
wonderful experience for me.

I would also like to thank the rest of my dissertation committee, who have all
contributed in unique and important ways. I thank Anne Bogat for her support in my
professional aspirations and her careful, meticulous way of carrying out our research (not
to mention her wonderful cookingl), Alex von Eye for his patience throughout my
statistical analyses and his funny sense of humor, and Hiram Fitzgerald for sharing his
incredible knowledge about infant mental health and early childhood development and
for challenging me to keep improving. I feel very fortunate to have learned from all of
them. I am also grateful to Alytia, Anne, Alex, and Bill Davidson for their generous
ﬁnancial support of this dissertation, which was made possible through their many hours
of grant writing.

In addition, I want to thank my main partners in this research endeavor: Sally
Theran, Robin Weatherill, and Jennifer Trotter. I could not have completed this without

their daily emotional support and practical help. They are great ﬁ'iends and I will always

remember the ﬁm we had together, even during highly stressful times. I will really miss
them. Many, many thanks also go to the large number of dedicated, caring research
assistants who have worked on this project and made this dissertation possrble. They
spent endless hours recruiting subjects, interviewing families, and coding, scoring, and
entering data. They too provided me with many much-needed laughs along the way.
Finally, I would like to thank my family for their continued support,
encouragement, conﬁdence, and enthusiasm for my work. I have greatly appreciated

their interest in what I do and their sharing in my achievements.

TABLE OF CONTENTS

List of Tables ......................................................................................... ix
List of Figures ......................................................................................... x
Introduction ............................................................................................ 1
Chapter 1: Theoretical Background ................................................................ 4
Development of Internal Working Models ...................................... 4
Categories of Attachment in Infancy ............................................. 8
Attachment Representations and Categories of Attachment in Adults. . . .. l 2
Chapter 2: Maternal Representations ............................................................ l6
Maternal Representations of Attachment and Mother-Inﬁnt
Attachment ......................................................................... 16
Maternal Representations of the Infant ......................................... 27
Maternal Representations of Self-As-Mother ................................. 40
Chapter 3: Ecological Factors Related to Attachment ......................................... 49
Maternal Social Support and Mother-Infant Attachment .................... 49
Risk Factors and Mother-Infant Attachment .................................. 56
Chapter 4: Hypotheses of the Present Study .................................................... 61
Chapter 5: Method .................................................................................. 67
Participants ........................................................................ 67
Procedures ......................................................................... 68
Measures ........................................................................... 74
Chapter 6: Results .................................................................................. 89
Missing Data ....................................................................... 89
Data Reduction .................................................................... 90
Measurement Models ............................................................ 94
Full Structural and Measurement Model ....................................... 97
Exploratory Analyses: Concordance Between Attachment Groups ...... 101
Chapter 7: Discussion ............................................................................. 104
Representational Processes ..................................................... 104
Social Support and Mother-Infant Attachment .............................. 111
The Impact of Risk Factors ..................................................... 113
Exploratory Analyses ............................................................ 115
Summary of Results ............................................................. 120
Directions for Future Research ................................................ 124

Clinical Implications ............................................................. 126

Conclusions ....................................................................... 129
Appendices .......................................................................................... 130
Appendix A: F lyers for Recruitment .......................................... 131
Appendix B: Time 1 Consent Form ........................................... 133
Appendix C: Time 2 Verbal Assent Form .................................... 134
Appendix D: Tracking Letter and Information Sheet ....................... 135
Appendix E: Time 3 Consent Form ........................................... 137
Appendix F: Perceptions of Adult Attachment Questionnaire ............ 138
Appendix G: Working Model of the Child Interview ...................... 143
Appendix H: Maternal Eﬂicacy Questionnaire .............................. 150
Appendix I: The Strange Situation Laboratory Procedures ................ 152
Appendix J: Social Support Questionnaire ................................... 153
Appendix K: Postnatal Pregnancy Support Items ........................... 158
Appendix L: Demographic Questionnaire .................................... 159
Appendix M: Beck Depression Inventory .................................... 163
Appendix N: Severity of Violence Against Women Scales Time 1 ...... 166
Appendix 0: Severity of Violence Against Women Scales Time 3 ...... 168
Appendix P: Parent Behavior Checklist ...................................... 169
Appendix Q: Physical and Emotional Neglect Scale ........................ 170
Appendix R: Correlation Matrix of Original Model Variables ............ 171
Appendix S: PAAQ Measurement Model .................................... 173
Appendix T: WMCI Measurement Model .................................... 174
Appendix U: Time 1 Social Support Measurement Model ................ 175
Appendix V: Time 3 Social Support Measurement Model ................ 176
Appendix W: Risk Factors Measurement Model ........................... 177
Appendix X: Revised Risk Factors Measurement Model .................. 178
Appendix Y: Correlation Matrix of Final Model Variables ............... 179
Appendix Z: Standardized Residual Values for Observed and Latent
Variable ........................................................... 181
Appendix AA: Lisrel 8.5 Syntax for Final Model .......................... 182
Appendix BB: Goodness of Fit Statistics .................................... 183
Appendix CC: Q-Plot of Standardized Residuals ........................... 184
References .......................................................................................... 186

LIST OF TABLES

Table 1. Demographic Characteristics of Study Participants ................................. 69
Table 2. Descriptions of Original Attachment Classiﬁcation by Subgroup ................ 82
Table 3. Descriptive Data for Measures in the Hypothesized Model .................... 91 -92
Table 4. Breakdown of Participants Experiencing Demographic Risks ..................... 93
Table 5. Breakdown of Participants Experiencing Depression .............................. 93
Table 6. Breakdown of Participants Experiencing Domestic Violence ..................... 93
Table 7. Breakdown of Participants with Child Abuse/Neglect Risk ....................... 93
Table 8. Breakdown of Participants Experiencing Depression and Domestic Violence

at Time 1 .................................................................................... 96
Table 9. Concordance Between Three WMCI and SS Classiﬁcations .................... 102
Table 10. Concordance Between Two WMCI and SS Classiﬁcations .................... 103

LIST OF FIGURES

Figure 1. Hypothesized Model of Prenatal and Postnatal Predictors of Mother-Infant
Attachment at 12 Months ............................................................... 64

Figure 2. Best Fitting Model Depicting Prenatal and Postnatal Predictors of Mother-
Infant Attachment at 12 Months ..................................................... 100

INTRODUCTION

Beginning with the work of John Bowlby (1944, 1969) and Mary Ainsworth
(Ainsworth, Blehar, Waters, & Wall, 1978), attachment theorists and researchers have
been studying the nature and quality of mother-infant attachment, including the
precursors and sequelae of different patterns of attachment. Empirical work has largely
focused on the infant’siperspective of attachment with his/her mother by studying inﬁInt
behaviors in stressful and non-stressful situations, from which attachment quality is
inferred. More recently, theorists and researchers have begun to examine mother-infant
attachment ﬁ'om the parental perspective. In addition, there has been a trend to “move to
the level of representation” rather than to approach the study of attachment solely at the
behavioral level (Main, Kaplan, & Cassidy, 1985; Oppenheim & Waters, 1995).

Both clinical reports (e.g., Frarherg, Adelson, & Shapiro, 1975) and recent
empirical research have indicated that there is a signiﬁcant relationship between mothers’
representations of their own childhood attachment and mothers’ attachment to their own
infants (Benoit & Parker, 1994; Fonagy, Steele, Moran, Steele, & Higgitt, 1993; Fonagy,
Steele, & Steele, 1991; Levine, Tuber, Slade, & Ward 1991; Slade & Cohen , 1996; Ward
& Carlson, 1995). However, few empirical studies have been conducted to help clarify
this continuity of attachment within families. Existing studies have suggested that
maternal representations of the infant and of the self-as-mother might explain the
relationship between mothers’ own attachment during childhood and the attachment
formed with their infants (Ammaniti, 1991; Bretherton, Biringen, Ridgeway, Masline, &

Sherman, 1989; Fava Vizziello, Antonioli, Cocci, & Invernizzi, 1993; Zeanah, Benoit,

Hirshberg, Barton, & Regan, 1994). Unfortunately, most of these studies have been
limited by small, primarily Caucasian and middle-class samples, as well as the use of
different measures with no known psychometric properties.

In addition to the impact of maternal representations on mother-infant attachment,
other researchers have suggested that attachment is inﬂuenced by external factors such as
the amount and type of maternal social support. Empirical studies have shown
inconsistent results (Crockenberg, 1981; Zeanah et al, 1993); however, ﬁndings ﬁom
different studies suggest that social support may have a more indirect effect on mother-
infant attachment through factors such as maternal behaviors or maternal self-efﬁcacy
(Crockenberg, 1988).

The current study examined possible mechanisms through which mothers’ own
attachment experiences could impact the attachment relationship with their infants, as
well as the impact of environmental factors on this relationship, within a large, diverse,
and high-risk sample. This study used a longitudinal research design and previously
established measures, and was conducted in conjunction with a larger study examining
the impact of domestic violence during pregnancy on women and infants. Maternal
representations of attachment, of the infant, and of the self-as-mother were assessed
during pregnancy, while self-efﬁcacy and mother-infant attachment were measured at 12
months after birth In addition, social support and several risk factors were measured
both prenatally and postnatally. It was hypothesized that characteristics of both the
mother and the mother’s environment would predict the quality of attachment formed

with her infant.

Because this study was largely based on attachment theory, an overview of this
theory will be presented ﬁrst in order to provide a ﬁamework for understanding the
hypothesized model. Next, studies that have examined the relationship between mothers’
own attachment during childhood and mothers’ attachment to their infants will be
reviewed. Although continuity of attachment has repeatedly been observed, possible
mechanisms through which attachment is transmitted have generally been missing in the
literature. Two possible mechanisms appear to be mothers’ representations of their
infants and representations of themselves as mothers, and the existing literature in these
areas will be reviewed and critiqued. Since the present study also examined the
relationship between social support and attachment, the few studies that have examined
this relationship will be reviewed and their mixed ﬁndings will be discussed. Finally,
because this study examined factors related to mother-infant attachment within a high-
risk sample, literature that describes the relationship between a number of potential risk

factors and mother-infant attachment will be explored.

CHAPTER 1: THEORETICAL BACKGROUND
Attachment Theory and the Mother-Infant Relationship

The mother-infant relationship has been of interest to psychoanalytic theorists
since its inception. Sigmund Freud initially recognized the importance of the mother as
an “Object” that gratiﬁes the infant’s instinctual drives and ﬁequently noted the profound
effect that early relationships had on later ﬁmctioning. It was not until the development
of object relations theory, by key ﬁgures such as Donald Winnicott, William Fairbairn,
and John Bowlby, that this relationship became a more central focus of inquiry. Unlike
previous analytic theorists, object relations theorists maintained that the infant is
primarily “driven” to establish relatedness and interactions with others as an end in itself.
They purported that infants come “wired” or prepared for immediate human interactions,
which form the foundation for the infant’s developing sense of selfand future
relationships. In addition, clinical investigators such as Rene Spitz helped clarify the
infant’s role as an active participant in dyadic relationships, which furthered our
understanding of self-object differentiation.

Drawing ﬁ'om evolutionary biology, cognitive psychology, and ethology, in
addition to object relations theory, John Bowlby (1969, 1988) developed attachment
theory to further explain the mother-infant relationship. Bowlby stated that the infant
comes into the world with an innate predisposition toward social interaction as a means
of survival and adaptation. Thus, the infant is driven to seek proximity with the primary
caregiver, particularly when threat or danger is perceived, because s/he is unable to
protect him/herself. According to Bowlby (1969), inﬁmts are born with a number of

attachment behaviors that are used to regulate the attachment relationship with the

mother, including sucking, smiling, crying, and clinging, the rmin purpose of which are
to attain or maintain proximity to a caregiver who is conceived as better able to cope with
the world.

Bowlby proposed that attachment behaviors are part of a behavioral system that
help maintain a sense of emotional homeostasis for the infant. Within this system, infant
behaviors are activated when a distance threshold between the infant and caregiver is
exceeded or another threat is perceived (Carlson & Sroufe, 1995). In a well-ﬁrnctioning
system, parents contribute to this homeostasis by showing complementary attachment
behaviors, which are of the caregiving type. While the infant behaves in ways that seek
out help and protection, mothers behave in ways that protect and secure the infant. When
proximity is achieved, infant attachment behaviors are deactivated. As will be described
further, parents and infants can also develop more maladaptive attachment patterns
during which infant security is compromised.

Mother-infant attachment appears to develop within the ﬁrst few months of life
(Bowlby, 1988), possibly because it is during this time that infant regulation (or
dysregulation) has long-term effects on neural system development (Carlson & Sroufe,
1995). During the ﬁrst 3 months of life, a synchrony related to physiological needs sets
the tone for later psychological regulation. Between 3 and 6 months, the caregiver helps
the infant maintain organized behaviors and aﬁect regulation in the face of stress and
arousal by making stress tolerable. This enables the infant to feel secure even when faced
with negative feelings and to gain conﬁdence that s/he can inﬂuence what lappens to
him/her. Around the ages of6 to 9 months, the infant assumes a more active role in self-

regulation and self-initiated activities due to increased development of motor and

cognitive abilities. With the onset of locomotion, additional attachment behaviors are
added to the infant’s repertoire such as purposeful approach to the mother, following the
mother, and use of the mother as a secure base from which to explore (Ainsworth, 1967),
all of which continue to serve the function of maintaining proximity to the mother when
needed.

Development of Internal Workgg' Models

By the end of the ﬁrst year, Bowlby (1969, 1980) proposed that infants begin to
develop “internal working models,” which are mental representations of self and others
based on early experiences with caregivers (a more elaborated version of what other
psychoanalysts called “internal objects” or “internal world”). Internal working models
are formed ﬁ‘om the history of the infant’s actions, infant-parent interactions, and the
caregiver’s responses to the infant, rather than ﬁom an objective picture of the parent
(Main et al., 1985), and working models of self and other are believed to be separate, but
complementary. Although internal working models are thought to consist of generalized
representations, Bowlby preferred the term “internal working model” because it
emphasizes the dynamic and ﬁrnctional aspects of representations (Bretherton, Ridgeway,
& Cassidy, 1990).

Bowlby stated that the function of internal working models is to help the infant
interpret and anticipate others’ behaviors and to guide his/her own behaviors in
relationships, which is adaptive to the individual. For example, the infant can operate
internally by making a plan am! selecting behaviors to achieve a goal before action is
needed (Marvin & Britner, 1999). Although early internal working models are not

necessarily ﬁxed, they tend to show stability and resistance to change because a) they are

partly unconscious (Bowlby, 1973, 1980; Bretherton, 1990; Cassidy, 1990), b) they bias
what individuals expect and perceive, leading to some distortion of incoming information
(Bretherton, 1990; Bretherton & Munholland, 1999), and c) reconstruction requires
sigm'ﬂcantly different emotional experiences that do not ﬁt with the current working
model (Bretherton et al., 1990; Ricks, 1985).

Bretherton and colleagues (1990, 1999) recently extended Bowlby’s ideas about
working models by considering advances in the cognitive sciences (e.g., memory
research). She proposed that internal working models are best conceptualized as
hierarchically organized schemas (or representations), with lower levels consisting of
speciﬁc, interactional schemas and higher levels becoming more and more general as
they subsume lower-level schemas. For example, beginning at the lowest level and
proceeding to the highest level, one part of a working model might be: “When I fall down
and cry, my mother picks me up,” “When I am hurt, my mother comforts me,” “My
mother is there when I need her,” and ﬁnally “My mother is a loving person.”

Bretherton ( 1990) also stated that internal working models of self and others are
probably several inter-linked hierarchies of representations, rather than a single hierarchy.
In addition, she noted that there are individual diﬁerences in the degree of organization
and consistency within and across levels, which have an effect on a person’s
interpretation of events and subsequent behaviors. For example, individuals who have
representations that are not linked properly or are dissociated from one another will
misinterpret signals from the environment and will be more likely to show contradictory
behaviors. Further, incoming information may not be generalized throughout the

hierarchies, which may lead to overly rigid working models. Individuals with such

distorted, “ill-organized” hierarchies are more likely to be insecurely attached than
individuals with more well-organized and consistent representations.

Internal working models of the self are similar to other working models, i.e., of
the attachment ﬁgure, as they are mental constructions based on actual experiences that
guide appraisals, feelings, and behaviors (Cassidy, 1990). Working models of the self
develop out of events and interactions and are believed to be generalized representations
of the self. Further, working models of the self contain both cognitive and aﬂ‘ective
components, that is, ideas about self-image (“I am a hard worker”) and self-esteem (“I am
lovable”) (Bowlby, 1979; Cassidy, 1990).

Based on internal working models of the self and of others, infants interact in
certain ways with their mothers during times of perceived threat. Under stressful
conditions, infants either have their attachment needs met or they ﬁnd other ways to cope
with their unmet needs, i.e., by exaggerating signals or suppressing their needs (Barnett
& Vondra, 1999). Therefore, based on early experiences with caregivers, individual
differences in attachment develop. Differences in attachment are believed to be ﬁnite and
can be classiﬁed into several central patterns, which are “organized” in the sense that
they reﬂect a consistent, coherent strategy for behaving in relation to the caregiver
(Ainsworth et al., 1978; Main et al., 1985). These categories of attachment will be
summarized next.

Categories of Attachment in Infancy

Mary Ainsworth, a student of John Bowlby’s, made substantial contributions to

attachment theory through her early observations of mother-infant dyads (Ainsworth,

1967). As a result of her work, she created a procedure to empirically study mother-

infant attachment called the Strange Situation (Ainsworth et al., 1978). This brief, mildly
stressful, laboratory procedure consists of a series of interactions between the infant, the
mother, and a stranger, during which the infant is separated and reunited with the mother
twice. The infant’s behaviors during reunion with the mother are believed to be indicative
of the infant’s internal working model of the relationship (Main et al., 1985), and
therefore, quality of attachment is inferred ﬁom Observed behaviors.

A Secure Attachment (Type B) is characterized by the infant’s conﬁdence that a
caregiver will be available and responsive when in need. As attachment theory would
suggest, research has shown that a secure attachment is associated with a history of
consistent, sensitive, responsive care, which presumably gives rise to an internal working
model of others, and the world in general, as dependable, comforting, and protective.
Because these infants can rely on their caregivers, they are able to safely explore the
environment and are likely to develop a sense of mastery and control. At the same time,
securely attached children develop an internal working model of the self as valuable,
worthy of love, and self-reliant. Based on these internal working models, securely
attached children show unique behaviors in the Strange Situation. They are able to use
the caregiver as a secure base from which to explore the room. Upon separation, these
infants may be distressed and may be comforted by a stranger, but will show a clear
preference for the mother and will seek contact with her upon reunion. In addition, these
children will be readily comforted by the mother and often return to exploration and play
(see Carlson & Sroufe, 1995; Weinﬁeld, Sroufe, Egeland, & Carlson, 1999 for reviews).
Research has shown that approximately 60-75% of low-risk infants display a secure

pattern of attachment (van Ijzendoom & Kroonenberg, 1988).

The internal working model of an infant Showing an Insecure-A voidant
Attachment (Type A) is characterized by a lack of conﬁdence in the caregiver and
expectations of rejection. The caregiving history of these children include overt
rejection, intrusive and controlling behaviors, as well as disengagement by the caregiver.
Internal working models of the self likely include a view of the self as unworthy of care,
but also possibly an extreme view of self-reliance in the context of a relationship in
which the infant receives little care and comfort. Infants displaying this pattern of
attachment in the Strange Situation show little distress at separation, no preference for the
mother, and often actively avoid the mother upon reunion by ignoring her, looking or
turning away, and resisting her initiation of contact. In fact, these infants show almost no
aﬁ‘ective responses at all and may appear indifferent. Avoidant behavior is believed to
reﬂect a defensive strategy (namely repression) against emotioml pain associated with
prior parental rejection, as well as a strategy to prevent the direct expression of anger that
might put the child at risk for alienating the attachment ﬁgure (Bowlby, 1980; Cassidy &
Kobak, 1988). Evidence in support for this notion comes from studies showing a
signiﬁcantly higher level of reactivity in avoidant infants as measured by cortisol levels
and cardiac ﬁrnctioning (e.g., Gunner, Mangelsdorf, Larson, & Hertsgaard, 1989).
Approximately 15-25% of infants in the general population have an avoidant attachment.

Insecure-Ambivalent Attachment (Type C), seen in about 10-15% of low-risk
samples, occurs when the infant is uncertain whether a parent will be available and
responsive. Insecure-ambivalent infants generally lack conﬁdence in the world because
caregivers have been unresponsive and inconsistently available for meeting their

children’s attachment needs. As a result of such experiences, parallel internal working

10

models of selfas unworthy and/or incompetent develop. During the Strange Situation,
these children seem preoccupied with their caregivers and Show a paucity of exploration
and play. Upon separation, these infants show a high level of overt distress and are
especially wary of strangers. Upon reunion, ambivalent infants are not able to be soothed
and display anger and ambivalence toward the caregiver by seeking and resisting contact
at the same time. Unlike avoidant infants who may suppress attachment behaviors and
needs, ambivalent infants are believed to exaggerate attachment needs and signals in an
effort to get a response ﬁom the caregiver and to ensure the mother’s availability (see
Cassidy & Berlin, 1994 for a review).

Aﬁer a number of studies discovered that about 10-15% of infants could not be
reliably classiﬁed into the A, B, or C patterns of attachment, a fourth pattern was
identiﬁed by Main and Solomon (1990), which they termed Disorganized-Disoriented
Attachment (Type D). Disorganized infants are believed to experience an unusually high
degree of fear within the attachment relationship, and research has shown that these
caregivers are often abusive, neglectﬁrl, and/or traumtized by their own experiences of
loss or violence (Main & Hesse, 1990; Main & Solomon, 1990; Lyons Ruth, Bronfrnan,
& Parsons, 1999). Unlike the three traditional groups, infants with this attachment
pattern appear to lack a coherent strategy for coping with stress and interacting with their
caregivers. Thus, their behaviors in the Strange Situation are idiosyncratic and may
include apprehensive or depressed behavior, alterations between approach and avoidance
toward the caregiver, strange behaviors such as ﬁ'eezing or stereotypies, and
disorientation reﬂected by wandering, confusion, and rapid changes of affect (Main &

Solomon, 1990).

11

Attachment Representations and Categories of Attachment in Adults

Although attachment has historically been studied at the behavioral level with
mother-infant dyads, as in the Strange Situation, recent advances have made it possible to
study attachment at the “level of representation” with older individuals (Main et al.,
1985). Main et al. proposed that internal working models formed during infancy function
as a “template” through which experiences are ﬁltered. Thus, internal working models of
self and others tend to be fairly consistent throughout childhood and into adulthood. On
the basis of these aspects of attachment theory, Main and colleagues (George, Kaplan, &
Main, 1985) developed the Adult Attachment Interview (AAI) to assess adults’
representations of attachment or internal working models. These researchers suggested
that internal working models inﬂuence feelings and behaviors, as well as attention,
memory, and thinking processes as they relate to attachment. Therefore, individual
diﬂ'erences in attachment representations will manifest themselves in the way one speaks
about attachment experiences and the way in which one constructs narratives about early
experiences with caregivers. Although most research in this area has used the AAI to
measure adults’ representations of attachment, it should be noted that several
questionnaires have also been developed to measure adults’ attachment experiences
during childhood (e.g., Epstein, 1983; Lichtenstein & Cassidy, 1991).

Based on responses to the AAI, researchers are able to infer a person’s overall
state of mind with respect to his/her own attachment experiences. Emphasis is placed on
the way in which a person recalls memories and experiences ﬁom childhood (the
narrative process), rather than on what they recall (the content). Therefore, special

attention is paid to the degree of coherence and integration of a person’s narrative. The

12

AAI coding system, developed by Main and Goldwyn (1984a, 1998), allows people to be
classiﬁed into one of four attachment categories.

Adults classiﬁed as Autonomous tend to produce consistent and coherent
narratives and are able to integrate positive and negative aspects of feelings and
experiences. In addition, autonomous adults tend to value attachment relationships and
believe that they inﬂuence current behavior. Dismissing adults are likely to minimize
attachment-related experiences and tend to give short, inconsistent responses. Although
these adults often report highly positive experiences, they are unable to provide examples
to support their favorable representations. It is not uncommon for dismissing adults to
have trouble remembering aspects of their childhood, and some researchers have
suggested that the dismissing style and tendency to idealize may reﬂect a defensive
strategy against painful memories (Cassidy & Kobak, 1988). Adults classiﬁed as
Preoccupied also display inconsistent and/or incoherent interviews; however, they seem
preoccupied with past attachment experiences. They appear to be greatly inﬂuenced by
earlier experiences in a rather negative manner (i.e., they display much anger or fear).
Adults can also be classiﬁed as Unresolved if they fail to maintain an organized narrative
or show striking lapses in reasoning or strange use of speech, particularly when
discussing traumatic events (George et al., 1985; Hesse, 1999). Finally, these four groups
are thought to parallel the four types of attachment observed in infants during the Strange
Situation (secure, insecure-avoidant, insecure-ambivalent, and disorganized,
respectively).

In sum, attachment theory suggests that infants develop internal working models,

or representations of self and others, based on early interactions with caregivers.

13

Individual differences in working models exist, and subsequently, infants behave
differently toward their mothers and display different patterns of attachment to their
mothers, which may be more readily seen during times of stress. Attachment theory
suggests that internal working models are relatively stable through childhood and into
adulthood, and thus, adults maintain representations of attachment based on early
experiences with caregivers. Some studies have ﬁrrther suggested that mothers
“transmit” their own attachment patterns to their infants, demonstrating a stability of
attachment across generations.

Based largely on attachment theory, Daniel Stern (1985, 1995) has recently
developed a unique framework for understanding the process through which mother-
infant attachment is formed and possibly transmitted intergenerationally. Stern (1995)
proposes that mothers develop a unique, psychic organization during pregnancy, which
he terms “the motherhood constellation,” composed of a new set of fantasies, fears,
action tendencies, and wishes never before experienced and directly due to becoming a
mother. Stern believes that this psychological change begins during pregnancy,
particularly after the ﬁrst trimester of pregnancy when the fetus begins to move, and the
reality of the baby becomes more apparent.

In addition to the mother’s representations of self and other, which become
activated and reworked during pregnancy, Stern (1995) suggests that the infant’s
developing representations are an important component of the mother-infant relationship.
Like Bowlby and other attachment theorists, Stern emphasizes that the infant’s
representations are constructed from interactive experiences with the caregiver.

However, he notes that rather than viewing representations as internalized events or

14

Objects, representations are built ﬁom within, “from the experience of being with
another.” Similar to Bretherton’s notion of multiple, hierarchically organized internal
working models, Stern (1995) suggests that repeated interpersonal experiences (or
“emergent moments”) give rise to a network of schemas, which form an overall schema-
of-being-with. Furthermore, he notes that schemas-of-being-with that are tied together
by a common theme may be considered a ‘Tepresentation-of-being-with.” Thus, Stern
suggests that schemas-of-being-with are hierarchically ordered ﬁom actual interactions to
more generalized internal working models, the latter of which ultimately guide behavior.

Consistent with attachment theory, Stern (1995) indicates that representations-of-
being-with and internal working models are relatively stable throughout childhood and
adulthood. Like other researchers, Stern also suggests that mothers’ representations of
attachment may be transmitted or passed on to their own infants, resulting in a continuity
of attachment across generations. Empirical studies that have investigated the
relationship between mothers’ own representations of attachment and their attachment
with their infants will be reviewed next, followed by a review of the literature examining
the possible mdiating role of other types of maternal representations, such as

representations of the infant and of the self.

15

CHAPTER 2: MATERNAL REPRESENTATIONS
Maternal Representations of Attachment and Mother-Infant Attachment

Psychoanalytic theory and clinical papers have long suggested that there is an
intergenerational pattern of parent-child relationships, that is, that history repeats itself
(e.g., Freud, 1940; Bowlby, 1969, 1980; Fraiberg et al., 1975). Although Fraiberg et al.
stated “history is not [necessarily] destiny,” they convincingly described clinical cases
that demonstrated how the conﬂicted past of the parents interfered with the relationship
with their children. That in fact, a mother’s own past experiences as a child were being
repeated in striking detail with her own child. Early writings such as these, however, had
little empirical work to support their assertions. No systematic investigations had been
conducted to determine the factors that led to the repetition of the past or (in fewer cases)
that blocked the repetition.

With the development of the AAI, researchers have been able to empirically
examine the intergenerational patterns of relationships and the impact of mothers’ own
representations of attachment on mother-infant attachment. In several early studies,
mothers’ responses on the AAI and AAI attachment classiﬁcations were compared to
infants’ behaviors and attachment security observed during the Strange Situation
(Grossman, Fremmer-Bombik, Rudolph, & Grossman, 1988; Main & Goldwyn, 1984b;
Main et al., 1985). The Strange Situation was administered to middle-class, mother-
inﬁtnt dyads when infants were 12 months old (N ranged ﬁom 40 to 65 in different
studies), and the AAI was later administered to the same mothers when their children
were 5 to 6 years old. Main et a1. (1985) found that mothers’ own attachment security

was signiﬁcantly related to their infants’ attachment security. Further, dismissive

16

mothers were more likely to be parents of avoidant infants, while preoccupied mothers
were more likely to have ambivalent infants. Main and Goldwyn (1984b) also noted a
striking similarity between a mother’s own experiences of rejection, her rejection of her
infant, and the infant’s subsequent avoidant behaviors, suggesting that speciﬁc behaviors
may also be repeated.

Grossman et a1. (1988), using their own coding system for the AAI made up of
four scales that were believed to reﬂect a parent’s attachment representations, reported
that mothers with a representation of their parents as supportive and mothers who were
able to focus on attachment experiences in a non-defensive way (even if these
experiences were negative) were signiﬁcantly more likely to have secure infants, while
mothers who presented defensive representations (i.e., unable to discuss experiences or to
recall memories) were more likely to have insecurely attached infants. Thus, overall,
these studies indicated that mother-infant attachment patterns tended to replicate mothers’
own attachment experiences and representations.

The generalizability of results ﬁ'om these studies are limited by the predominately
upper-middle class, educated, Caucasian sample. In addition, because the AAI was
administered several years after the Strange Situation, it is more difﬁcult to conclude that
maternal representations led to infant attachment style. That is, it is possible that having
children for a number of years somehow changed a mother’s own representations of her
relationship with a caregiver. However, attachment theory suggests that intemal working
models are fairly stable over time, and therefore, mothers’ representations of attachment

were presumably simihr when infants were 12 months old.

17

Mm RgnesengtionsAQd Infant Attach_n_r_ent Measured Concurrenm

Several other studies have examined the continuity of attachment by measuring
both mothers’ representations of attachment and infant attachment concurrently
(Pederson, Gleason, Moran, & Bento, 1998; Zeanah et al., 1993). Researchers in both
studies administered the AA] and Strange Situation to a sample of 60 middle-class,
mostly married, Caucasian mothers with 1 year-old infants. Using the original AAI
coding system, both studies reported a signiﬁcant degree of concordance between
mothers’ own attachment representations and their infants’ attachment categories (73%
agreement with a kappa = .56 and 75% agreement with a kappa = .62, respectively). By
collapsing insecure types, Pederson et al. found that the concordance rate between
autonomous-secure and nonautonomous-insecure groups increased to 80% (kappa = .60).
However, when examining speciﬁc types of attachment, the researchers found that
although there was a signiﬁcant concordance between dismissive mothers and avoidant
infants (69% and 94%, respectively) and between autonomous mothers and secure infants
(81% and 87%, respectively), there was no signiﬁcant relationship between preoccupied
mothers and ambivalent infants.

These studies suggest that the degree of concordance between mothers’ and
infants’ patterns of attachment may be unique to each type of attachment classiﬁcation,
with more stability for the dismissive-avoidant and autonomous-secure groups and less
stability for the preoccupied-ambivalent groups, at least among a Caucasian, middle-class
sample. Pederson et a1. (1998) attempted to examine the process through which
attachment was transmitted by assessing maternal sensitivity as a possible mediator.

They discovered that although matemal sensitivity was signiﬁcantly related to mothers’

18

own security of attachment and to infant security of attachment, sensitivity accounted for
less than 25% of the association between mother and infant attachment. Therefore, it
seems important to continue to assess possrble mechanisms that might help explain the
intergenerational continuity of attachment patterns, as well as to examine these
relationships in larger, more diverse samples.

Finally, in discussing the social transmission of parenting and attachment, Ricks
(1985) examined the relationship between mothers’ recollections of relationships with
parents and current mother-infant attachment in a sample of 24 middle-class mother-
infant dyads. Ricks used the Mother-Father-Peer Scale (Epstein, 1983), which is a
questionnaire assessing the degree of parental acceptance/rejection,
independence/overprotection, and idealization of parents, rather than the AAI, as the
measure of matemal attachment. As expected, mothers of securely attached infants
reported more positive recollections of childhood relationships including more
acceptance, encouragement, and less rejection than mothers of insecurely attached
infants. However, there were no diﬁ‘erences between mothers on amount of idealization.
Unfortunately, concordance rates between attachment categories were unable to be
examined, as seen in other studies, due to the type of measure used in this study.

Existing studies that have assessed maternal representations of attachment
concurrently or years after the assessment of mother-infant attachment have provided
fairly strong empirical support for the continuity of attachment across generations.
Furthermore, a few studies have shown a signiﬁcant concurrent relationship between
maternal representations based on the AA] and preschooler attachment security,

suggesting that maternal representations may continue to inﬂuence young children’s

l9

developing representations of attachment after infancy (Bus & van Ijzendoom, 1992;
Eiden, Teti, & Corns, 1995; Posada, Waters, Crowell, & Lay, 1995). However, a
growing number of research studies have utilized a prospective research design, by
examining maternal representations of attachment before the infant is even born and then
assessing mother-infant attachment at 1 year postpartum. In contrast to cross-sectional
studies, a prospective design beginning in pregnancy increases conﬁdence in the
conclusion that maternal representations contribute to the development of mother-infant

attachment rather than sinme being related correlationally.

Maternal Representations and Infant Attachment Measured Prom

 

One such study administered the AAI to a sample of 96 pregnant, Caucasian
women ﬁom middle to upper-middle class backgrounds and then administered the
Strange Situation when infants were 1 year old (Benoit & Parker, 1994). Using the four-
classiﬁcation system, results revealed a signiﬁcant degree of concordance between
mothers’ representations of attachment during pregnancy and mother-infant attachment at
1 year (68% match, kappa = .46). Using three groups only (by reclassifying the
unresolved/disorganized type into one of the other insecure groups), the authors reported
an 81% concordance rate (kappa = .55). However, consistency across generations
secured to diﬂer for secure and insecure groups; 83% of autonomous mothers had secure
infants, while only 22% of dismissing mothers had avoidant infants, 57% of preoccupied
mothers had ambivalent infants, and 58% of unresolved mothers had disorganized
infants. Thus, the authors concluded that the highly signiﬁcant rate of stability across
generations was primarily due to the autonomous-secure groups while insecure groups

seemed to be signiﬁcantly less predictable over time.

20

Interestingly, Benoit and Parker (1994) also administered the AAI postnatally and
found that AAI responses were signiﬁcantly stable over time (77% match rate using the
four groups and 90% using three groups), suggesting that mothers’ states of mind about
attachment experiences are highly stable. The authors also reported that concordance
rates were slightly higher for the AAI measured concurrently with the Strange Situation
(74% using four groups and 82% using three groups), compared to pregnancy AAIs,
although this was again primarily due to the autonomous-secure groups. Finally, this
study was the ﬁrst to examine patterns of attachment across three generations by also
administering the AAI to the pregnant women’s own mothers. Results ﬁ'om these
analyses revealed an overall concordance rate of 65% using the three classiﬁcation
groups across all three generations. Although the ﬁndings from this study are impressive
in a number of ways (e.g., prospective design, sample size, and data ﬁ'om three
generations), the results must be interpreted with caution due to the skewed distribution
towards attachment security (76% of infants were secure using three groups) and the
limited generalizability ﬁom a non-representative sample.

Fonagy and colleagues have also extensively studied the intergenerational
continuity of attachment beginning with maternal representations during pregnancy
(Fonagy et al., 1991; Fonagy, Steele, Steele, Moran, &. Higgitt, 1991; Fonagy et al., 1993;
Steele, Steele, & Fonagy, 1996). In an early study, they administered the AAI during the
last trimester of pregnancy to a fairly large sample (N = 100) of Caucasian, well educated
and mostly married women and then administered the Strange Situation when infants
were 1 year old (Fonagy et al., 1991; Fonagy et al., 1991). They reported an overall

match between the three original classiﬁcations of 66% (kappa = .38) and a 75% match

21

rate (kappa = .48) when two-way comparisons were made (secure-insecure). When
examining concordance between speciﬁc groups, they found that 75% of autonomous
mothers had secure infants, 68% of dismissive mothers had avoidant infants, but only
20% of preoccupied mothers had ambivalent infants. These ﬁndings are similar to those
in other studies (Pederson et al., 1998; Zeanah et al., 1993), which suggest less stability
for the preoccupied/ambivalent groups, and more generally, for insecure types of
attachment.

In an effort to understand the discontinuity between preoccupied mothers and
ambivalent infants, Fonagy et al. (1991) discovered that a number of mothers who later
had insecure-ambivalent infants had a unique interview style during pregnancy. They
found that many of these mothers initially presented an overall impression of security in
their representations of attachment (and therefore were rated as being autonomous), but
upon closer examination, tended to have a somewhat exaggerated sense of security in
combination with difﬁculties adjusting to their new maternal role (i.e., less conﬁdence in
themselves). On the other hand, one-third of preoccupied mothers went on to have a
secure attachment pattern with their infants, suggesting tlmt positive changes in
attachment representations are possible as well, perhaps even made more likely by
becoming a parent. These are interesting hypotheses that need to be further examined in
order to better understand the continuity and discontinuity of different attachment
patterns.

Finally, Fonagy and colleagues have demonstrated that a parent’s capacity for
understanding the mental states of self and others (which they term reﬂective-self

function) is related to both maternal and infant security (Fonagy et al., 1991; Fonagy et

22

al., 1993). In order to test this, they created a Reﬂective-Self Function scale to be coded
ﬁ'om the AAI. As expected, autonomous mothers (as measured by the AAI during
pregnancy) had signiﬁcantly higher reﬂective-self scores than dismissive and

preoccupied mothers, and mothers with higher reﬂective-self scores were also more
likely to have securely attached infants at 1 year. Thus, a mother’s ability to anticipate
her infant’s needs by being in tune with her infant’s mental world presumably leads to
more sensitive caregiving and more secure attachment.

Just as some mothers are aware of and in tune with their own and their infants’
internal world, other mothers are unaware of such internal experiences. Fonagy et al.
(1993) reported that mothers’ incoherence in telling narratives about their childhood
experiences on the AAI was the clearest predictor of infant insecurity 12 months later.
The authors explained that incoherence was indicative of certain defenses, and indeed,
there was also evidence of idealization, repression of affect, and splitting in these
mothers’ interviews. The authors concluded that the defensive strategies used by some
mothers may explain an unconscious mechanism underlying the transmission of insecure
attachment. In other words, a mother’s own defenses against acknowledging certain
experiences in her life make it difﬁcult to respond empathically to her infant, which in
turn, leads to the development of similar defensive strategies in the child. The studies
conducted by Fonagy and colleagues provide strong empirical support for the clinical
writings of Fraiberg et al. (1975), who noted early on that mother-child dyads ﬁequently
and unknowingly repeat “scenes” ﬁ'om the mothers’ own childhood experiences with
caregivers. Moreover, studies by F onagy and colleagues support the notion of continuity

of attachment across generations, especially for secure and dismissing/avoidant patterns.

23

The majority of studies examining the link between maternal representations of
attachment and mother-infant attachment have been conducted with primarily Caucasian,
well educated, and economically advantaged samples. This is problematic because of the
limited generalizability of results and possible skewed distributions toward attachment
security (including higher rates of concordance across generations due to the greater
stability of security). However, two known studies have investigated the transmission of
attachment in more high-risk samples (Levine et al., 1991; Ward & Carlson, 1995).

In these studies, the AAI was administered during pregnancy to adolescent
mothers (N = 42 and 72, respectively) from diverse, ethnic backgrounds (majority were
Aﬁ'ican-American) and later conducted the Strange Situation when infants were 15
months old. As expected, mothers had much higher rates of insecure representations of
attachment (e.g., 21% autonomous, 48% dismissive, 7% preoccupied, 24% unresolved;
Levine et al., 1991) compared to studies using low-risk samples. However, consistent
with prior studies, results demonstrated a highly signiﬁcant relationship between
maternal representations of attachment and later mother-infant attachment. Levine et al.
reported a 62% concordance rate using the four categories of attachment, and an 83% rate
when all three insecure groups were combined for a two-group comparison, while Ward
and Carlson (1995) reported a 78% concordance when using two-groups. Further
analyses revealed a signiﬁcant degree of concordance between autonomous mothers and
secure infants (100% and 86%, respectively), dismissive mothers and avoidant infants
(60% and 73%, respectively), and preoccupied mothers and ambivalent infants (66% and

60%, respectively). The studies did not ﬁnd signiﬁcant continuity of

24

unresolved/disorganized attachment across generations in these adolescent samples, with
concordance rates of only 30% and 43%, respectively.

Taken together, preliminary results ﬁom studies with high-risk samples (e. g.,
adolescent mothers) are consistent with results ﬁ'om middle-class, well educated samples.
They provide ﬁrrther support for the link between maternal attachment representations
and mother-infant attachment as posited by attachment theory and clinical observations,
as well as further evidence that there is relatively less stability (but still signiﬁcant in
some cases) among certain insecure groups. However, these results must be interpreted
with caution because of the srmll sample sizes (e.g., only three mothers in the
preoccupied group for one study). In addition, there is a clear need for more studies with
different types of high-risk samples, i.e., low income or depressed mothers, since
adolescent motherhood is qualitatively distinct from other risk factors.

In sum, it appears that maternal representations of attachment are a strong
predictor of the types of attachment mothers form with their own infants. According to a
recent meta-analysis of 18 studies and 854 subjects (van IJzendoorn, 1995), the effect
size of this relationship was 1.06, which is comparable to a concordance rate of 75%
(kappa = .49), using secure and insecure classiﬁcations. Interestingly, when the
unresolved group was not used in studies, the dismissing and preoccupied classiﬁcations
both predicted their respective infant groups (combined gs = .45 and .42, respectively);
however, the preoccupied classiﬁcation was only marginally related to the ambivalent
group in infancy when the unresolved group was included in analyses. van Ijzendoom
(1995) also reported that only about 12% of the variation in maternal sensitivity, believed

to be a possible mechanism through which attachment is transmitted, is explained by

25

maternal representations of attachment. Thus, other processes are likely occurring and
further research needs to be conducted to more fully explain the continuities and
discontinuities of attachment. That is, possible mechanisms through which attachment is
(or is not) transmitted ﬁom one generation to the next must be measured and examined
empirically, which unfortunately has been largely missing in the current literature.

The few mechanisms that have been suggested include maternal sensitivity and
responsiveness, the degree of disorganization among a mother’s internal working model
which gives rise to contradictory behavior and ﬁtulty communication (Bretherton, 1990),
and a mother’s defensive strategies, evidenced by an inability to integrate past
experiences and a tendency to distort current relationships, which may force an infant to
rely on similar maladaptive defensive strategies (Cassidy & Kobak, 1988; Fonagy et al.,
1993; Grossman et al., 1988; Main & Goldwyn, 1984b; Slade & Cohen, 1996). Other
possible mechanisms of transmission that are consistent with attachment theory and
appear to be promising explanations based on the available literature are rmtemal
representations of the infant and representations of the self-as-mother.

Just as working models of the self develop along with working models of others
during childhood, representations of the infant and representations of the self-as-mother
develop together during pregnancy. A number of attachment researchers have shown that
these representations are similar in both content and organization (e.g., Ammaniti et al.,
1992). That is, the way mothers conceptualize themselves as mothers is similar to the
way mothers think about their infants. Although some researchers believe
representations of selfand infant are distinct (Ammaniti et al., 1992; Stern, 1995), other

attachment researchers conceptualize them as integrated “caregiving” representations

26

(George & Solomon, 1989, 1996). The current study proposes to examine
representations of self-as-mother and infant as complementary but distinct, just as
representations of self and other are complementary but distinct during childhood.
Therefore, the possible role of each type of representation as a mediator between
mothers’ own attachment and mother-infant attachment will be considered separately,
and existing studies in each area will be reviewed next.
Maternal Representations of the Infant

Although not as prevalent as the literature on maternal representations of
attachment to caregivers, some theorists and researchers have written about mothers’
representations of their infants, particularly during pregnancy. Bibring, Dwyer,
Huntington, and Valenstein (1961) were among the ﬁrst to discuss the psychological
processes that occur during pregnancy including processes related to the “earliest mother-
child relationship.” They noted that pregnancy seems to revive old psychological
conﬂicts, reorganize the new mother’s relationship with her own mother, and causes the
mother to develop attitudes and representations of her developing infant. As the infant
begins to move and grow inside the mother, there is a shift ﬁ'om a focus on the self
(“enhanced narcissism”) to a focus on the infant as a separate object. Leifer (1977) also
reported that women gradually develop rich and speciﬁc representations of their infants
as pregnancy progresses, most often beginning in the second trimester, which prepares
them psychologically for motherhood. Similarly, Lumley (1982) reported that women in
their ﬁrst trimester of pregnancy have difﬁculty imagining what the fetus might be like
and are more focused on their own bodily changes, but they increasingly develop

representations of their infants as the pregnancy progresses. For example, only 30% of

27

mothers in her sample considered the fetus “a person” in the ﬁrst trimester, but 63% and
92% of women in the second and third trimesters perceived them as separate individuals,
giving vivid descriptions of their babies.

Some researchers in the health professions, such as nursing, have also described
the process through which pregnant women begin to interact with their infants by
developing ideas and feelings for the unborn children. The majority of these studies have
assessed “prenatal attachment” through use of questionnaires and have found that
attachment (deﬁned here as positive feelings toward the infant) increases during
pregnancy, with larger increases following quickening (see Muller, 1992 for a review). It
has been suggested by these researchers that early feelings toward the infant form the
basis for the mother’s relationship with her child after birth

Zeanah and colleagues have examined mothers’ perceptions of their infants
during pregnancy in a number of studies with both low- and high-risk groups of women.
In one study, they administered infant temperament questionnaires and an interview
about maternal perceptions to 35 pregnant, Caucasian women from middle-class
backgrounds (Zeanah, Keener, & Anders, 1986a; Zeanah, Keener, Stewart, & Anders,
1985; Zeanah, Zeanah, & Stewart, 1990). They found that mothers had fairly stable
perceptions of their inﬁants’ temperament throughout pregnancy, which also predicted
their perceptions postnatally. They noted that women were able to give vivid, abstract
descriptions of their babies’ personalities before birth, and perceptions during pregnancy
seemed to be strikingly positive. These authors replicated their ﬁndings in a sample of
adolescent, mostly Hispanic girls (N=24), by showing that maternal perceptions of infant

temperament were signiﬁcantly stable ﬁom 32 weeks to 36 weeks gestation to 1 month

28

postnatally (Zeanah, Keener, & Anders, 1986b; Zeanah, Keener, Anders, & Vieira-Baker,
1987). Zeanah et al. (1986b) concluded that adolescent mothers also develop detailed
perceptions of their infants, which likely impact the developing relationships with their
children.

Similar to the work by Zeanah and colleagues, another group of researchers
examined maternal perceptions of infants during pregnancy by administering an infant
temperament questionnaire during the second and third trimesters, as well as postnatally,
to 41 middle-class women (Mebert, 1989; Mebert & Kalinowski, 1986). Their results
also indicated that mothers form stable perceptions of their infants during pregnancy and
continuing after the baby is born. In addition, they found that ﬁrst-time parents had
signiﬁcantly more negative views of their infants than multiparous parents.

In a number of reports, Zeanah and colleagues (Zeanah et al., 1986b; Zeanah et
al., 1987; Zeanah & Carr, 1990) indicated that fetal movements contribute to the
formation of mothers’ initial perceptions of their infants which then develop into stable,
detailed representations. Further, they suggested that fetal movements may act as a
“projective stimulus,” from which mothers form elaborate representations that are not
simply made up randomly, but likely constructed ﬁom the mothers’ own experiences in
relationships. Since a mother’s own relationship with her mother appears to be
reactivated during pregnancy, it seems probable that this particular relationship may have
an impact on the mother’s developing representations of her infant.

Although the studies by Zeanah et al. (1985, 1986a, 1986b, 1987) and Mebert et
al. (1986, 1989) described above provide interesting information about mothers’

perceptions of their infants during pregnancy, they are limited by the sole focus on infant

29

temperament, the bias of self-report measures, and the atheoretical nature of their studies.
For example, they do not address why mothers form vivid perceptions of their infants
before birth, why these perceptions are so stable, or how mothers understand and
experience their infants more generally. In contrast, more recent work on maternal '
perceptions during pregnancy has been guided by both attachment and psychoanalytic
theories. In fact, Zeanah and colleagues were among the ﬁrst to reﬁame earlier ﬁndings
according to attachment theory (Zeanah & Anders, 1987; Zeanah & Barton, 1989), by
suggesting that mothers form internal working models of their infants during pregnancy
which, like other internal working models, are relatively stable and guide perceptions of
their infants and behaviors toward them. Since that time, several groups of researchers
have attempted to develop more theory-driven, qualitative measures to assess rmternal
representations of infants.
Maternal Rapresentations of Attachment and Representations of the Infant

Ammaniti and colleagues (Ammaniti, 1991; Ammaniti et al., 1992) hypothesized '
that during pregnancy, representations of self and other are modiﬁed and reworked,
particularly of self-as-mother and the “imaginary” child. Like others have suggested
(Leifer, 1977; Zeanah et al., 1987), representational processes accelerate after the ﬁrst 3
months when the infant becomes more subjectively real due to growth and movement. In
order to investigate these psychological processes, they developed the Interview of
Maternal Representations During Pregnancy (IRMA G), a semi-structured interview
about parents’ reactions to and experiences of pregnancy, perceptions of their infants, and
their own personal experiences during childhood. Similar to the AAI, they developed a

coding system that emplmsized content-ﬂee dimensions such as the coherence and

30

quality of the narrative (or how parents talked about their experiences). The authors
administered this interview and several adjective checklists (asking mothers to rate
themselves in various roles and their infants) to 23 middle-class, Italian women in their
last trimester of pregnancy.

They found that women developed rich representations of themselves as mothers
and of their babies during pregnancy, and that individual differences existed between
women. In addition, content-ﬂee dimensions of their interviews (i.e., coherence,
openness, richness) were highly related between representations of selfand of inﬁmts.
That is, the organization of women’s narratives was consistent across domains of self and
other, as would be expected based on attachment theory. Women also tended to rate their
infants similarly to themselves on a number of characteristics measured by adjective
checklists (although they were signiﬁcantly different on some characteristics like
sociability and independence). Overall, the authors concluded that women develop clear
representations of themselves and their infants during pregnancy, and that these are
complementary in both organization and content. Further, through a series of case
studies, Ammaniti (1991) indicated that maternal representations of self and infant during
pregnancy were related to mothers’ own representations of attachment (according to the
AAI) and to mother-infant attachment in the Strange Situation. These ﬁndings imply that
prenatal representations may be one mechanism through which attachment is transmitted
across generations, and it would be useful for future research to examine this with a
larger, more heterogeneous sample.

In a similar study (Fava Vizziello et al., 1993), 51 pregnant women from middle

to upper-middle class backgrounds were administered an interview about representations

31

of the self, their infants, and their own childhood experiences, as well as asked to make
ratings along a number of bipolar-adjective lists. The content and organization of
women’s narratives produced during the interview were coded for several themes
including: a) a desire to repair or compensate for previous experiences (not necessarily
considered pathological), b) a defensive effort to maintain current ﬁmctioning, c) intense
fear stemming ﬁom matemal anxiety about the health and development of the child, and
d) a lack of organization or consistent theme during the interview.

First, in analyzing the stability of representations over time, the authors reported
that representations of women’s own mothers were most stable, while representations of
self-as-mother and of the infant were only somewhat stable ﬁ-om pregnancy to
postpartum (according to adjective ratings). During pregnancy, fear themes seemed to
dominate women’s representations during interviews as compared to other themes; most
mothers displayed conﬂicted fantasies of life and death and showed great concern for the
babies’ health Women who displayed themes of a desire to repair or compensate or
those unable to organize themes in the interview during pregnancy were signiﬁcantly
more likely to show similar themes after the birth Like the ﬁndings reported by
Ammaniti et al. (1991 , 1992), these results suggest that women’s representations during
pregnancy are detailed and rich and relatively stable in organization. In addition, F ava
Vizziello et al. (1993) found that representations of attachment ﬁom childhood predicted
mothers’ representations of themselves as mothers, but not representations of their
infants. Again, it seems important to replicate these ﬁndings in other samples, as well as
to validate these different interview techniques that are designed to assess maternal

representations.

32

In an effort to better understand the role of prenatal representations in the
transmission of attachment, Slade and Cohen (1996) developed and administered the
Pregnancy Interview, a semi-structured interview for assessing women’s internal working
models of their infants, to 66 married, middle-class women in their last trimester of
pregnancy. They also administered the AAI during pregnancy to assess matemal
representations of attachment. Although analyses ﬁ‘om the total sample were
unfortunately not presented, the authors described the results ﬁ‘om a number of case
studies, which revealed a strong association between mothers’ representations of
attachment based on childhood experiences and their representations of their infants.
Mothers tended to conceptualize and organize their relationship with their babies during
pregnancy in ways that were similar to the way they conceptualized their relationships
with their own mothers, evidenced by the degree of ﬂexibility, coherence, and richness in
their representations. In other words, mothers were beginning to repeat their own
experiences during childhood with their infants (at the representational level) before the
infants were even born. Although these ﬁndings are consistent with attachment theory
and with previous ﬁndings (Ammaniti et al., 1992), the results should be interpreted with
caution since they are based only on case studies using a different, unvalidated interview.

Another study empirically examined the relationship between maternal
representations of attachment and representations of children in 32 middle-class,
Caucasian mothers (George & Solomon, 1996). When children were 6 years old,
mothers were administered the AAI and the Experiences of Caregiving Interview,
designed to assess mothers’ representations of themselves as parents and representations

of their children (combined to form a “caregiving” representation). Scales were

33

developed to assess both the content and organization of mothers’ responses, which
resulted in general categories that were analogous to the four AAI categories. Results
revealed a signiﬁcant concordance between representations of attachment and
representations of caregiving (69% match with a kappa = .58). There was a 77% match
between categories for the three traditional attachment groups (autonomous, dismissing,
preoccupied). Although this study examined internal working models among mothers of
older children rather than infants who are not yet born, it makes an important contribution
to the literature by demonstrating a statistically signiﬁcant relationship between mothers’
representations of attachment and representations of their children. Thus, it is important
to assess these relationships among pregnant women and their infants in more diverse
samples, in order to better understand how patterns of attachment are repeated through
representational processes.

Overall, the current body of literature shows an association, albeit preliminary,
between mothers’ representations of attachment and representations of their own infants
or children and also suggests that representations of the infant may be a mechanism
through which attachment patterns are transmitted. Other researchers have focused on
the second pathway in this process: the relationship between representations of the infant
and mother-infant attachment. These studies will be reviewed next.

Maternal Representations) ofth_e Infant and Mother-Infant Attachment

 

In one such study, Bretherton et al. (1989) administered the Parent Attachment
Interview, developed by the authors to assess mothers’ thoughts and feelings about their
infants and their relationships with their infants, as well as the Strange Situation to

mothers (N = 37) of 2 year olds. One year later, they also administered an attachment

34

story completion task to measure attachment at age 3. Results revealed that mothers who
demonstrated sensitivity and insight into their relationships with their children based on
interview responses (and therefore, presumably had more “secure” representations of
their children) were more likely to have securely attached children at age 2 and age 3.

Similarly, George and Solomon (1989, 1996) reported that mothers who displayed
more secure representations of their 6 year-old children (deﬁned as an understanding of
the child’s needs, as well as coherent organization of thought processes related to these
issues) had more securely attached children during a separation-reunion procedure. In
fact, secure maternal representations accounted for more than 50% of the variance in
child security. Furthermore, the concordance between classiﬁcations for caregiving
representations and attachment classiﬁcations was highly signiﬁcant at 81% (kappa =
.75). The authors concluded that the way in which mothers thought about their children
and their relationships with their children had a signiﬁcant association with the
attachment pattern their children developed.

Although the results reported by Bretherton et al. (1989) and George and
Solomon (1989, 1996) are consistent with attachment theory and provide empirical
support for a link between maternal representations of caregiving and mother-child
attachment, they are not necessarily generalizable to pregnant women or parents with
young infants or to high-risk samples. In addition, like most other research in this area,
they are limited by small samples and varying methods of assessment. Because each
group of researchers tend to develop their own interviews to assess maternal
representations of the child, it is difﬁcult to compare results across studies because

different constructs are being coded and analyzed. In addition, existing measures are not

35

being validated across different sarrrples, making it unclear how useful these measures
are.

In contrast, Zeanah and colleagues (Zeanah et al., 1994) have recently developed
an interview that is being used more extensively, i.e., across different samples of both
non-clinical and clinical groups, to measure mothers’ representations of their infants.
The Working Model of the Child Interview (WMCI) is a 1 hour structured interview that
inquires about a mother’s perceptions and subjective experiences of her infant and
relationship with her infant. Interview responses are rated along a number of scales that
assess qualitative (i.e., richness of representations, coherence, and Openness to change),
content (i.e., infant difﬁculty), and affective features (i.e., joy, pride, anger) of maternal
representations. Coders then assign an overall classiﬁcation to the narratives provided by
the mothers based on the scale proﬁles. This classiﬁcation system is consistent with
attachment theory and with AAI and Strange Situation categories.

Balanced narratives include both positive and negative characteristics of the
infant. They convey value for the infant’s individuality and appreciation for the infant’s
subjective experience. The caregiver’s perceptions are open to change and are at least
moderately rich in detail about the infant and the caregiving experiences. Disengaged
representations are characterized by emotional distance or indifference towards the
infant. Caregivers are unable to recognize the infant’s individuality, and if it is
recognized, it is not valued. Details about the infant or parenting experience lack
richness, are not ﬂexible or open to new experience, and are emotionally unintegrated.
Finally, Distorted representations reﬂect general inconsistencies. Caregivers rmy be

preoccupied or overwhelmed by the infant or may have unrealistic expectations about the

36

infant. Unlike disengaged parents, distorted parents do not deny their impact on their
infants. However, they Often do not recognize how their behaviors may be detrimental to
the infants. Much feeling is expressed toward the infants, but these emotions lack a sense
of modulation or meaning.

In their original study, Zeanah et al. (1994) administered the WMCI to middle-
class mothers of 12-month old infants (N = 45), as well as the Strange Situation to all
mother-infant dyads. They found that WMCI categories were signiﬁcantly related to
Strange Situation categories, with an overall concordance rate of 69% (kappa = .50).
Seventy-four percent of mothers with balanced representations of their infants had secure
attachments according to the Strange Situation, while 73% of disengaged mothers had
avoidant infants and 55% of distorted mothers had ambivalent infants. In addition,
mothers of secure infants scored signiﬁcantly higher on a number of WMCI subscales
including richness of perceptions, openness to change, coherence, and caregiving
sensitivity than mothers of insecure infants. Consistent with other interviews such as the
AAI, content-free dimensions were especially important for distinguishing mothers of
secure and insecure infants.

In a later study, these researchers administered the WMCI to 96 Caucasian, upper-
middle class pregnant women and then readministered the WMCI along with the Strange
Situation when infants were approximately 1 year old (Benoit, Parker, & Zeanah, 1997).
First, they reported that WMCI categories were signiﬁcantly stable ﬁom pregnancy to
postpartum (80% agreement, kappa = .59), although this was accounted for by the
stability of balanced (89%) and distorted (85%) mothers. The strong stability of working

models over time is consistent with attachment theory, although results also suggested

37

that certain types of working models may be less stable across time. Results also
revealed a 73% concordance rate between concurrent WMCI and Strange Situation
categories, including an 88% match for balanced-secure groups, 100% match for
disengaged-avoidant groups, but only 40% match for distorted-ambivalent groups.
Similar to Zeanah et al. (1994), the high overall concordance was primarily due to the
balanced and disengaged groups.

Perhaps most interestingly, Benoit et al. (1997) reported a signiﬁcant concordance
between mothers’ representations of their infants during pregnancy and Strange Situation
classiﬁcations at 1 year (overall match rate = 74%, kappa = .44), although this association
was rminly explained by the balanced-secure groups (91% concordance) rather than the
disengaged (30% concordance) and distorted groups (50%). One possible explanation for
the lower concordance rates for disengaged and distorted groups is that continuity may
not be speciﬁc to type of insecurity. That is, disengaged mothers may have ambivalent
infants and vice versa Therefore, it may be more useful to look at “secure” and
“insecure” groups. Because the WMCI classiﬁes representations into three categories
(rather than four like the AAI and Strange Situation), studies using the WMCI are only
able to compare classiﬁcations with the three original Strange Situation categories.
Therefore, another possibility is that infants who might have been classiﬁed as
disorganized on the Strange Situation were “forced” into one of the other categories,
confounding the results for those groups.

Still, because the WMCI was administered prenatally rather than concurrently
with the Strange Situation, the ﬁndings indicate that prenatal representations might

inﬂuence the way in which mothers’ perceive and interact with their children after they

38

are born, which in turn could impact the type of attachment formed. It will be important
for more studies to use prospective designs such as this one to better understand the
processes involved in the transmission Of attachment. In addition, the relationship
between WMCI and Strange Situation categories should be examined in more diverse and
larger size samples. Overall, the results of studies using the WMCI indicate its

usefulness as a tool for measuring matemal representations of infants, as well as for
predicting mother-infant attachment.

In sum, a considerable amount of literature suggests that mothers develop internal
representations of their infants well before birth and lasting into childhood. These
representations, or internal working models, of infants appear to be inﬂuenced by
mothers’ representations of attachment based on their own childhood experiences. That
is, mothers tend to think about and conceptualize their infants and the relationship they
have with their infants in similar ways to how they think about their relationships with
caregivers. Furthermore, a number of studies have indicated that maternal
representations of infants, both before and after birth, predict the quality of mother-infant
attachment. Thus, it seems possible that mothers’ representations of their infants is one
mechanism through which mothers’ own attachment impacts the quality of attachment
they form with their own infants.

Another possible mechanism, as mentioned earlier, is maternal representations of
self-as-mother. Mothers’ representations of attachment may inﬂuence representations of
themselves, which my in turn impact the attachment formed with their infants. Studies
that have examined the relationship between mothers’ perceptions of themselves and

mothers’ own attachment experiences and their attachment with their infants will be

39

reviewed next, beginning with a brief theoretical introduction about representations of the
self.
Maternal Representations of Self-As-Mother

As mentioned previously, because internal working models of self and others
develop together in the context of early interactions with caregivers, they tend to be
complementary (Bowlby, 1973; Bretherton, 1990; Bretherton & Munholland, 1999; Slade
& Cohen, 1996). Therefore, a working model of the self as valuable and competent will
develop in the context of a caregiver who treats the infant as valuable, i.e., who is
emotionally available and sensitive. In contrast, a working model of the self as
incompetent and unlovable will develop in the context of a rejecting or unempathic
parent. This notion of complementarity has received empirical support; two studies to
date have demonstrated that securely attached children are more likely to develop a
positive working model of self, while insecurely attached children develop either a
negative working model of self or idealized, distorted self-representations (Cassidy,
1988; Verschueren, Marcoen, & Schoefs, 1996).

More recently, attachment theorists have questioned whether this early sense of
self determines one’s global self-worth (in multiple areas and relationships), as
attachment theory would suggest, or whether it determines one’s working model of the
self in relation to the attachment ﬁgure (Cassidy, 1990). Like attachment theory,
Epstein’s self-theory states that individuals develop beliefs about the self early in life that
are more global, i.e., “I am worthy of love,” and these beliefs are based in large part on
interactions with caregivers (Epstein, 1973; Ricks, 1985). Because new information is

assimilated into these global representations, they tend to be long lasting and resistant to

40

change. In contrast, other theories such as Bandura’s self-efﬁcacy theory (1977, 1989)
propose that beliefs about oneself are situation-speciﬁc and change in response to
different demands, situations, and individual developmental processes.

In the current study, representations of the self and self-efﬁcacy beliefs will be
considered separate constructs, although more general representations of the self are
thought to inﬂuence self-efﬁcacy beliefs more or less in different situations and roles.
Because this study is examining representations of self-as-mother and self-efﬁcacy in the
caregiving role speciﬁcally, it is expected that representations will be strongly related to
self-eﬁcacy beliefs.

Mammal Representa_ti_ons of Attachment and Representations of Self-As-Mother

A number of theorists and researchers have proposed that during pregnancy,
women begin to modify and rework representations of themselves as they prepare for
motherhood. Stern (1995) stated that a mother’s representations of herself as a woman,
wife, daughter, career-person, and in other roles are activated and reorganized as she
makes room for her role as a mother. Consistent with attachment theory, Rubin (1984)
also suggested that women go through a process of redeﬁning themselves as they form a
“maternal identity.” This is achieved when the woman sees herself in relation to her
unborn child and is considered important to the woman’s sense of competency in being a
mother to her child. As previously mentioned, several groups of researchers have indeed
demonstrated that pregnant women develop vivid representations of themselves as
mothers throughout pregnancy and these representations are fairly stable over time

(Ammaniti, 1991; Ammaniti et al., 1992; Slade & Cohen, 1996).

41

Within the ﬁamework of attachment theory, preliminary evidence also suggests
that mothers’ representations of attachment based on experiences during childhood
inﬂuence the way in which they see themselves as mothers. That is, internal working
models developed early in life contribute to or guide the development of internal working
models of the self as mother. For example, several previously mentioned studies have
reported that mothers’ representations of their own attachment experiences with their
mothers were highly related to their representations of themselves as mothers (Amrmniti,
1991; Ammaniti et al., 1992; Fava Vizziello et al., 1993). Several other studies also
reported that mothers’ representations of attachment according to the AAI were related to
representations of themselves; mothers who had more secure representations tended to
see themselves in more positive ways and were more ﬂexible and coherent when
describing themselves as mothers, while insecure mothers described themselves more
negatively and had more incoherent, distorted representations of themselves as mothers
(George & Solomon, 1996; Slade & Cohen, 1996).

Another study also investigated the relationship between mothers’ experiences
with caregivers and their views of themselves as mothers (Meyer, 1988). Questionnaires
assessing childhood experiences and parenting efﬁcacy were administered to 35 well-
educated, middle class mothers with infants who were approximately 14 months old.
Results revealed that positive experiences with caregivers during childhood were strongly
related to positive views of themselves as mothers and higher maternal self-efﬁcacy.
Thus, mothers who reported having more caring and sensitive mothers were more likely
to see themelves as caring and sensitive and felt more competent in caring for their

infants compared to women who reported negative experiences with their mothers. It

42

should be noted that experiences with mothers during childhood were particularly
important for maternal perceptions of eﬁcacy compared with experiences with fathers,
which were unrelated to the way mothers viewed themselves. Unfortunately, this study
did not assess current mother-child attachment so it is unclear how maternal feelings of
efﬁcacy were related to mothers’ relationships with their infants in this sample.
M_aieLn_al Representations of Self As Moth}; and Mother-Infant Attachment

In addition to those studies showing a relationship between mothers’ experiences
during childhood with attachment ﬁgures and their representations of themselves, a
number of other studies have investigated the link between mothers’ representations of
themselves as mothers and their current attachment with their infants. Through several
case studies, Ammaniti (1991) suggested that mothers’ representations of themselves as
mothers were related to mother-infant attachment observed during the Strange Situation.
As expected, mothers who felt more conﬁdent and positive about their ability to parent
were more likely to have infants who were securely attached. Similarly, George and
Solomon (1989, 1996) reported that mothers who displayed more secure representations
of themselves as mothers (characterized by a willingness and ability to provide safety and
protection, a positive evaluation of abilities, and little self-doubt or conﬁrsion over
caretaking) had more securely attached children during a separation-reunion procedure.
Thus, these studies suggest that another possible mechanism through which mothers’ own
attachment impacts infant attachment is mothers’ representations of themselves as
(in)competent, (in)eﬂ“ective mothers. That is, representations of attachment ﬁ'om
childhood may inﬂuence the development of working models of the self-as-mother,

which in turn, impact the attachment that is formed with a mother’s own infant.

43

M_at_e__rnal Self-Efﬁcacy and Mothgr-Infant Attachment

Although only a few studies have examined the impact of women’s
representations of themselves on mother-infant attachment through interviews designed
to measure representations per se, a number of other studies have looked at the impact of
maternal self-efﬁcacy on mother-infant relationships and speciﬁcally, mother-infant
attachment. It is believed that maternal self-eﬁcacy, deﬁned as a mother’s perceived
ability to perform competently and effectively as a mother, is largely inﬂuenced by a
woman’s representhtions of the self as mother. Therefore, it is argued that maternal self-
eﬂicacy is directly related to representations of the self and is also likely related to
mother-infant attachment.

One study examined maternal self-efﬁcacy, “attachment” (deﬁned as maternal
feelings toward her infant), and overall self-esteem and depression in 121 medically high-
risk and 182 low-risk women during and after their pregnancies (Mercer & F erketick,
1990). Results revealed that self-efﬁcacy was the best predictor of attachment for both
groups of women; women who felt more capable of providing competent care felt closer
to and more involved with their infants. Similarly, Mercer and F erketick (1994) reported
that self-efficacy was strongly related to experienced mothers’ attachment to their infants,
but they reported no such relationship for inexperienced mothers. In another study, 238
women were interviewed during their last trimester of pregnancy and part of the sample
was again interviewed at 1 month postpartum (N = 86) and 2 years postpartum (N = 62;
Williams et al., 1987). Results from this study revealed that self-efﬁcacy measured
prenatally and at 1 month predicted attachment at 1 month only, while self-efﬁcacy at 2

years predicted the quality of the mother-child relationship at 2 years according to

rmternal report. Unfortunately, all of these studies used maternal report questionnaires to
assess “attachment” at different ages. It is likely that these instrmnents were measuring
maternal attitudes and perceptions of the mother-child relationship rather than attachment
behaviors and patterns as described by attachment theory. Thus, one can only conclude
ﬁ'om these studies that maternal feelings of self-efﬁcacy likely impact the way in which
mothers view their relationships with their infants.

Only a few studies have actually assessed the relationship between self-efﬁcacy
and mother-infant attachment as measured by the Strange Situation. In one study,
Spieker and Booth (1988) examined several potential predictors of mother-infant
attachment in 35 high-risk (deﬁned by few resources and low levels of support) mother-
infant dyads. Questionnaires were administered to mothers prenatally and at 6 weeks and
3 months postnatally, and the Strange Situation was administered when children were 14
months old. The authors found that mothers of secure infants reported more conﬁdence
handling motherhood tasks shortly after birth, and therefore considered themselves more
efﬁcacious, than mothers of ambivalent infants. Self-efﬁcacy ratings did not differ
among mothers of secure and avoidant infants; however, the authors suggested that
mothers of avoidant children might have been defensively reporting unrealistically high
levels of self-efﬁcacy (consistent with other defensive processes observed in mothers of
avoidant infants). Although this study had a small sample size, the prospective design
and high-risk nature of the sample make it an important and unique contribution to the
literature. More studies need to be done to better understand how mothers’ views of

themselves and their maternal competence impact mother-infant attachment.

45

Donovan and colleagues (Donovan & Leavitt, 1989; Donovan, Leavitt, & Walsh,
1990) also measured maternal self-efficacy (through the attributional styles mothers used
to describe the outcomes of their caregiving) in 40 white, married women with infants. In
addition, the authors measured mothers’ perceived control in an experimental paradigm.
During this procedure, mothers listened to a tape-recorded infant cry and were given
different options for responding to the crying. They were told that the procedure was a
simulated child-care task and that after hearing the infant cry, they could choose whether
or not to press a red button. Afterwards, mothers were asked to estimate how much
control they thought they had over the termination of the cry. Because neither response
was more effective than the other in reality, any perceived control that mothers reported
was considered “illusion of control.” Approximately one year later, the researchers also
administered the Strange Situation to all mother-infant dyads.

The results showed that low self-efﬁcacy was related to a high illusion of control.
Furthermore, mothers of securely attached children reported higher self-eﬂicacy and
lower illusions of control (therefore, they were more realistic about their abilities), while
mothers of insecurely attached children reported lower self-efﬁcacy and higher illusions
of control. The authors concluded that high illusory control may be a defensive strategy
used by mothers to mask feelings of ineﬁicacy. Thus, it seems important to distinguish
between mothers who are realistically self-efﬁcacious and those that are unrealistic in
their perceptions of efficacy. The latter group may be at particular risk of parenting
problems because their coping mechanism of illusory control may prevent them from

acknowledging and improving their own limitations.

46

Because the quality of mother-infant interactions and speciﬁc maternal behaviors
are the main contributors to speciﬁc patterns of attachment, it is important to note that a
number of studies have also found a relationship between maternal self-efﬁcacy and more
sensitive, positive maternal behaviors. For example, several studies have reported that
mothers who felt more competent and efﬁcacious in the maternal role were actually more
sensitive and responsive when interacting with their infants (Shea & Tronick, 1988; Teti
& Gelfand, 1991). This relationship seems to be especially strong for ﬁrst-time mothers,
who do not have a previously established maternal identity (Walker, Crain, & Thompson,
1986). Conrad, Gross, Fogg, and Ruchala (1992) also found that mothers who were more
conﬁdent in their parenting abilities demonstrated more positive interactions with their
toddlers, although this was true only for mothers who were knowledgeable about child
development. The authors concluded that accurate knowledge and conﬁdence (or self-
efﬁcacy) were both important for the quality of maternal behaviors towards their
children. Overall, most existing studies have suggested a link between self-efﬁcacy and
more positive maternal behaviors (with some exceptions, e.g., Zahr, 1991), but results
need to be replicated since existing samples are small and consist mainly of white,
middle-class, educated individuals.

In conclusion, the existing research suggests that mothers’ representations of
themselves, and in particular of themselves as mothers, are inﬂuenced by experiences
they had with their attachment ﬁgures during childhood. More speciﬁcally, mothers who
were sensitively cared for by their own mothers are more likely to view themselves as
competent, caring mothers to their children, while mothers who had more problematic

attachment relationships are less likely to feel conﬁdent in their abilities to parent.

47

Further, existing studies indicate that maternal perceptions and feelings of efﬁcacy
impact the quality of mother-child interactions and more importantly, the types of
attachment they form with their infants. Therefore, it seems probable that another
mechanism through which attachment patterns are repeated across generations is through
mothers’ representations of themselves as mothers.

Overall, it is likely that individual, intrapsychic factors like those that have been
described thus far (e.g., representations of self and infant) have a considerable impact on
the quality of attachment formed between a mother and her infant. However, it is
probable that attachment is also inﬂuenced by external factors within the mother’s and
infant’s environment. Therefore, the potential relationships between several ecological

variables and mother-infant attachment will be reviewed next.

48

CHAPTER 3: ECOLOGICAL FACTORS RELATED TO ATTACHMENT

The impact of a number of ecological factors on attachment has been considered
by attachment researchers, including factors at the individual, family, and community
levels. For example, some researchers have examined the relationship between child
maltreatment and mother-infant attachment (e.g., Crittenden, 1985), while other
researchers have looked at more global factors such as socioeconomic status (e.g.,
Spieker & Booth, 1988). Therefore, studies which have investigated the relationship
between different ecological factors and mother-infant attachment will be reviewed,
beginning with the literature on maternal social support and attachment, followed by a
review of the literature on several risk factors and attachment.

Maternal Social Support and Mother-Infant Attachment

One possible environmental factor related to mother-infant attachment is the
amount and type of social support available to mothers. According to Stern’s theory
(1995), mothers need to create and maintain a support network so that they can
accomplish the goals of motherhood; the support network buﬂ‘ers the mother ﬁ'om
external demands, protects her physically, provides emotional and practical support, as
well as instruction and advice. In all of these ways then, social support can have an
indirect, beneﬁcial effect on mother-infant attachment.

Some researchers have examined the impact of social support on attachment
through an ecological model, assuming that factors at multiple system levels are
important for explaining this aspect of the mother-infant relationship. In a review on
social support and parenting, Crockenberg (1988) hypothesized various ways that social

support can positively impact the parent-child relationship that are similar to those

49

mentioned by Stern (1995) including: a) reducing the number of stressful events that
parents experience, b) buffering the effects of existing stressors, c) serving as a source of
knowledge and competence, and (1) serving as a model for positive, nurturing
relationships which parents may then repeat with their own children. Crockenberg
(1988) concluded that in some or all of these ways, social support contributes to more
sensitive, responsive parenting and presumably to more secure patterns of attachment
between mothers and infants.

Several empirical studies that have looked speciﬁcally at the impact of social
support on mother-infant attachment provide support for the general assertions made by
Stern (1995) and Crockenberg (1988). In one of the earliest studies of this kind,
Crockenberg (1981) assessed the amount of perceived social support in the maternal role
at 3 months postpartum in 48 mothers ﬁ'om intact, middle-class families and later
assessed attachment in the Strange Situation at 12 months. Results showed that the
infants of mothers with more social support displayed signiﬁcantly fewer anxious
behaviors in the Strange Situation. In addition, there was an interaction between infant
irritability and maternal support; mothers with low social support were more likely to
have insecure infants, but this was only true for babies rated (by mothers) as
temperamentally “difﬁcult.” Thus, it seemed that low social support had a greater impact
on the attachment with irritable babies than with “e ” babies, suggesting the
importance of support in more stressed families.

In a number of other studies, the impact of social support on attachment has been
examined in high-risk samples. In one longitudinal study (Spieker, 1986; Spieker &

Booth, 1988), mothers who were deﬁned as high-risk because of reportedly little social

50

support were followed from pregnancy through the infants’ ﬁrst year of life in order to
determine contributors to parenting styles and quality of attachment. These researchers
found that mothers of infants classiﬁed as having a disorganized attachment (believed to
be “most” insecure) had signiﬁcantly less support around the parenting role during
pregnancy than mothers of infants classiﬁed as avoidant and had signiﬁcantly less
postpartum social support than mothers of securely attached infants. There were no other
signiﬁcant group differences on maternal social support during pregnancy or postpartum,
possibly due to a lack of variance in level of support because of the inclusion criterion for
this study. Unfortunately, this study did not look at possible mechanisms through which
support might have aﬁ‘ected attachment.

In another longitudinal study, Crnic, Greenberg, and Slough (1986) followed 36
mother-premature infant dyads and found that maternal support from partners, ﬁiends,
and the community at 1 month postpartum predicted secure mother-infant attachment at
12 months as assessed by a modiﬁed Strange Situation (a shortened version with one
separation and one reunion). In addition, life stress was signiﬁcantly related to insecure
infant attachment. Although social support buffered the effects of stress on some
outcome measures (e. g., parenting satisfaction), this was not true for attachment. Overall,
the ﬁndings from these two longitudinal studies appear to conﬁrm the importance of
looking at ecological variables such as social support when investigating factors related
to mother-infant attachment. However, like those with low-risk samples, these studies
are generally limited by their small sample sizes.

Using a larger sample size (N = 121), Crittenden (1985) evaluated the relationship

between social support and attachment among low-income families, many of whom were

51

identiﬁed as abusive and/or neglectﬁrl by Child Protective Services. Regression analyses
revealed that social support did not make a signiﬁcant contribution to the explained
variance in infant attachment after maltreatment status was accounted for (which
explained 35% of the variance). When maltreatment status was not considered,
Crittenden reported that the type or quality of support was related to mother-infant
attachment security, while the quantity of social support (both the number of network
members and total amount of support) was unrelated to attachment. More speciﬁcally,
mothers who had stable, long-lasting relationships and who were ﬁ'equently in contact
with supporters had more securely attached infants, while mothers who had temporary,
short-term relationships tended to have infants who were insecurely attached. The author
interpreted these ﬁndings using attachment theory by suggesting that mothers’ working
models of relationships affected both the pattern of attachment formed with their infants
and the types of relationships formed with other adults who may (or my not) serve as
sources of support. That is. the relationship between social support and mother-infant
attachment might be a result of a third variable, the mother’s internal working models.
This study illustrates the importance of looking at different dimensions of social
support including the network size, level of support, and duration and frequency of
contact with supporters. as each dimension may be differentially related or unrelated to
the quality of mother-infant attachment. Like previously mentioned studies, this study
also suggests that social support does not necessarily have a clear, direct relationship with
mother-infant attachment. Instead, the impact of social support on attachment may be
more indirect, i.e., through maternal attitudes or behaviors, and/or may vary according to

the nature of the sample being studied.

52

Although the studies described thus far suggest that social support, in one way or
another, has a signiﬁcant effect on mother-infant attachment, other studies suggest that
social support is unrelated to the quality of attachment. For example, Belsky and Isabella
(1988) followed 64 Caucasian, intact families beginning in pregnancy through the
infants’ ﬁrst year of life in order to examine the determinants of infant attachment. They
found that mothers of secure, avoidant, and ambivalent infants did not differ in their level
of support ﬁom ﬁiends and family during pregnancy or postpartum, although mothers of
secure babies perceived their neighborhoods (or “social milieu”) as friendlier and more
supportive during pregnancy than mothers of insecure babies. Similarly, Zeanah et al.
(1993) reported no differences between attachment groups on rmternal social support at 1
year postpartum in a middle-class sample. The results ﬁ'om another study with a
relatively small (N = 43) middle-class, intact sample (Levitt, Weber, & Clark, 1986) also
revealed no group differences in emotional or parenting support. In the latter study,
however, the authors reported that low support was related to negative maternal effect,
which was related to attachment quality, suggesting a possible indirect effect of social
support. Interestingly, the studies which have concluded that no relationship exists
between social support and attachment have all examined these variables in low-risk
samples. It is possible that social support is less critical for mothers and infants who have
other resources and who have less general life stress than those families in high-risk
situations who have fewer resources and more stress.

There are a number of other possible reasons for the inconsistent results regarding
the relationship between maternal social support and mother-infant attachment. One

possible expLanation is that studies use different deﬁnitions of social support and most

53

fail to specify what dimensions of support are being measured including the type and
source of support, the quantity of support or number of network members, as well as the
ﬁequency of contact or duration of supportive relationships. As already mentioned, it is
also possible that social support has a more indirect impact on mother-infant attachment
through its impact on other factors, which are left unexamined in most existing studies.

As suggested by Crockenberg (1988), one possible mechanism through which
social support inﬂuences mother-infant attachment is through a mother’s feelings of self-
eﬁicacy in the maternal role. Crockenberg and others (Bandura, 1977; Leifer, 1977;
Stern, 1995) reason that support can help a mother increase her parenting skills and
knowledge, which in turn increase feelings of eﬂicacy, can provide direct conﬁrmation
that she is capable and competent in caring for her infant, and can provide positive role
models ﬁom which to learn. Indeed, several empirical studies lmve shown a signiﬁcant
relationship between maternal social support and maternal self-eﬁcacy in both low and
highrisk samples of women (Cutrona & Troutman, 1986; Shea & Tronick, 1988; Teti,
O’Connell, & Reiner, 1996; Zahr, 1991). However, very few studies have examined the
role of self-eﬁcacy as a mechanism through which social support impacts attachment.

In one related study, Donovan and Leavitt (1989) showed that women who
demonstrated a high “illusion” of control around parenting (in an experimental paradigm)
due to feelings of ineﬂicacy also reported low levels of social support and had infants
who were more insecurely attached. Unfortunately, although ineﬂicacy was related to
both social support and attachment, the authors did not statistically test for mediation.
Isabella (1994) reported that although social support was not directly related to security

of attachment in 32 mother-infant dyads, maternal support had an indirect impact on

54

attachment through its impact on maternal role satisfaction, a construct presumed to be
highly related to self-efﬁcacy (e.g., Shea & Tronick, 1988). That is, support was related
to higher satisfaction, which was in turn related to attachment security. Finally, Teti and
Gelfand (1991) demonstrated that self-efﬁcacy mediated the relationship between social
support and sensitive maternal behaviors, which underlie secure attachment, in mothers
of infants between the ages of 3 and 13 months.

In conclusion, it is still somewhat unclear if and how maternal social support
affects the quality of mother-infant attachment. Existing results indicate that some
dimensions or types of social support do inﬂuence attachment, however, it may be most
appropriate to view social support as having an indirect impact on attachment rather than
a direct impact (Belsky, 1999). As noted above, one mechanism through which social
support may inﬂuence attachment is maternal self-eﬂicacy, although there is currently
little empirical work in this area. The literature on social support and parenting behaviors
(rather than attachment per se) is somewhat more voluminous, and several reviews
(Andersen & Telleen, 1992; Crockenberg, 1988) in this area suggest that: a) social
support is related to more sensitive caregiving, b) emotional support and support with
child care tasks are particularly important types of support for mothers, c) social support
from family members is especially important compared to other sources of support, and
d) mothers’ perceived level of support is more important than the sheer number of
supporters in her network. Because parenting behaviors and attachment are highly
interrelated, it seems possible that these ﬁndings could help inform ﬁrture studies

examining the role of support on mother-infant attachment.

55

Risk Factors and Mother-Infant Attachment

In addition to maternal social support, the relationship between other ecological
variables and mother-infant attachment has been examined by a number of researchers.
For example, several potential enviromnental risk factors have been shown to negatively
impact the quality of mother-infant attachment, indicating the need to consider risk
factors in any study of attachment. In general, research has shown that there is a greater
proportion of insecure attachment within high-risk samples (Spieker & Booth, 1988), as
well as less stability of attachment quality over time (V ondra, Hommerding, & Shaw,
1999). Furthermore, research suggests that although a single risk factor may not
distinguish secure ﬁom insecure infants, a clear and deﬁnite relationship between risk
and attachment emerges when multiple risk factors are considered simultaneously
(Belsky, 1999). Speciﬁcally, the more risk factors a Emily experiences, the more likely
it is that mother-infant attachment will be insecure in one form or another. This is
consistent with the cumulative-risk model described by Sameroﬂ' and colleagues
(Sameroﬂ‘, Seifer, Barocas, Zax, & Greenspan, 1987), as well as other researchers
(Harman & Luster, 1991; Luster & McAdoo, 1994), in that the number of risk factors
may be more important for developmental outcomes in children than the speciﬁcity of
risk factors.

Unlike the majority of studies on mother-infant attachment, which have been
based on stable, intact, middle-class families, the current study examined factors related
to mother-infant attachment in a more diverse, high-risk sample. More speciﬁcally,
almost half of the women in the current study experienced domestic violence during their

current pregnancy, and a large proportion of women were economically disadvantaged.

56

Very few studies have examined patterns of attachment across generations in high-risk
samples, making it unclear whether the processes suggested by attachment theory are the
same for both low and high-risk families. Therefore, a number of important and relevant
risk factors were considered in the current study. The existing literature on the
relationship between these risk factors and mother-infant attachment is reviewed next.
DomLestic Violgceand Adm

The literature on domestic violence, deﬁned as male to female violence between
adult partners, and mother-inEnt attachment is sparse. However, research by Lyons-Ruth
and colleagues (e.g., Lyons-Ruth et al., 1999; Lyons-Ruth & Jacobvitz, 1999) have
suggested that mothers who are traumatized by past loss and/or experiences of abuse tend
to show ﬁightening or ﬁightened caregiving behaviors, such as threatening gestures,
unusual voice intonations, trance-like states, and fearful withdrawal ﬁ'om their infants.
These behaviors, in turn, have been related to disorganized infant attachment. Although
research has focused on maternal experiences of loss and abuse during childhood
speciﬁcally, it seems probable that traumatizing experiences of abuse during adulthood
with intirrmte partners might have the same eﬂ‘ect on maternal behaviors and mother-
infant attachment.

In the only empirical study to date, Zeanah et al. (1999) examined the relationship
between domestic violence and mother-infant attachment in 72 low-income, mostly
unrmrried mothers and their 15-month old infants. They found that mothers who had not
experienced domestic violence were signiﬁcantly more likely to have securely attached
children, while mothers who were currently in an abusive relationship were signiﬁcamly

more likely to have infants with insecure attachments, particularly disorganized

57

attachments. These results are consistent with the notion that mothers traumatized by
abusive experiences are more likely to develop insecure patterns of attachment with their
infants. However, more studies need to be done to replicate these ﬁndings before strong
conclusions can be made.
Child Maltreatment and Attachment

Because child abuse occurs in about 35% to 70% of all homes characterized by
domestic violence (Jouriles & Norwood, 1995; O’Keefe, 1995), the impact of abuse on
attachment was also considered. In fact, child maltreatment is perhaps the most widely
studied risk factor in relation to mother-infant attachment. Child maltreatment has
repeatedly been shown to be highly related to the quality of mother-infant attachment,
with infants from all maltreatment groups being more likely to develop insecure
attachments (Crittenden, 1985; Crittenden, 1988; Egeland & Sroufe, 1981; Erickson,
Egeland, & Pianta, 1989; Lyons-Ruth, Connell, Grunebaurn, Botein, & 2011, 1984;
Lyons-Ruth, Connell, Zoll, & Stahl, 1987). Furthermore, maltreated infants are
particularly at risk for developing disorganized attachments, which show marked stability
through the toddler years, compared to other forms of insecure attachment (Barnett,
Gamban, & Cicchetti, 1999).
Mal Depression an_d Attaq__hmgrt

Maternal depression was also examined in the current study because battered
women have repeatedly been shown to experience signiﬁcantly more depression than
non-battered women (Cascardi & O’Leary, 1992; Levendosky & Graham-Bermann,
2001 ). The impact of nmterml depression, an indicator of maternal psychological

ﬁrnctioning, on mother-infant attachment has been investigated by several groups of

58

researchers. The overwhelming evidence indicates that depressed mothers are more
likely to have insecurely attached infants than non-depressed mothers (Donovan &
Leavitt, 1989; Radke-Yarrow, Cummings, Kuczynski, & Chapman, 1985; Spieker &
Booth, 1988; Teti, Gelfand, Messinger, & Isabella, 1995), a ﬁnding that is not
particularly surprising in light of the research that shows a strong relationship between
matemal depression and unavailable, unresponsive, and/or rejecting maternal behaviors
(see Gelfand & Teti, 1990 for a review).

Interestingly, based on a review of studies examining the relationship between
depression and infant attachment, Belsky (1999) concluded that the severity and
chronicity of exposure to maternal depression is also an important determinant of
insecure attachment. Not only will infants who are exposed to a depressed mother be
more likely to have an insecure attachment than infants never exposed to maternal
depression, but infants who have more severely depressed mothers and/or mothers who
are depressed for longer periods of time will be even more likely to display an insecure
attachment than infants who have depressed mothers of lesser severity and/or duration.
Demoggaphic Risk and Attachment

Low income and socioeconomic status (SES) have also repeatedly been linked to
problems in parenting and caregiving behaviors (for a review, see Halpern, 1993).
Because poverty produces a number of stressors on parents such as ﬁnancial strain and
unsafe and overcrowded communities, it tends to undermine parents’ well-being, feelings
of control and competence, and abilities to adequately care for children’s physical and
emotional needs. Poverty has been associated with less consistent, attentive caregiving,

less emphasis on infant needs compared to personal or familial needs, parental

59

unavailability, and less responsiveness to infant emotional well-being, among other

things (Bromwich, 1978; Norton, 1990). Overall, poverty and associated stressors tend to
preoccupy parents in ways that interfere with their attentiveness to parenting and
“preoccupation” with their infants (Halpern, 1993). Because low SES and income have
been associated with these caregiving problems, it is likely that these risk factors impact
the quality of mother-infant attachment as well.

Although fewer studies have been conducted in this speciﬁc area, there is some
evidence to support a signiﬁcant relationship between lower economic status and
insecure infant attachment, particularly ﬁom the literature showing a link between
maternal life stress and attachment (e.g., Crnic et al. 1986). Although some large-scale
studies have reported no relationship between SES and mother-infant attachment (e.g.,
Spieker & Booth, 1988), this is likely due to the low variance in SES in these samples.
That is, virtually all attachment studies have been conducted with an exclusively middle-
class sample or exclusively lower-class sample, rather than samples with a range of
economic classes, which makes it statistically difﬁcult to ﬁnd a relationship between SES
and attachment, even if one actually exists.

Finally, because there is a large overlap between poverty and single parenthood
(Halpern, 1993) and because single parenthood is related to a number of similar stressors,
it is possible that single parenthood is related to the quality of mother-infant attachment.
However, this relationship has yet to be investigated by attachment researchers.

Although single parenthood and poverty are highly related, it seems important to consider
single parenthood as an additional risk Ector since Sameroﬂ‘ et al. (1987) suggest that the

number of risk factors is more important than the speciﬁcity of risk.

60

CHAPTER 4: HYPOTHESES OF THE PRESENT STUDY

As mentioned previously, Daniel Stern (1995) has recently developed a theory to
explain the process that women go through as they prepare for motherhood and as they
form an attachment with their infants after birth Much of Stern’s theory has been based
on the attachment literature reviewed in this paper, as well as extensive clinical
experience with mothers and infants. According to Stern, a mother’s mental
representations of self and others (“networks of schemas-Of-being-with”) are reactivated
and reworked throughout pregnancy, especially after the ﬁrst trimester when the baby
becomes more “real” to the mother. In particular, Stern indicates tint three sets of
representations are especially important for a woman during this time.

First, Stern (1995) suggests that a new mother’s relationship with her own mother
becomes reactivated (consciously or unconsciously) during pregnancy as the new mother
forms her own ideas about caregiving. It is believed that preparation for motherhood and
the daily acts of mothering provide a speciﬁc remembering context, in which memories
from the mother’s own childhood are evoked. Stern also asserts that a mother’s
experiences of being mothered is a major inﬂuence on the way in which she interacts
withherinfant andonthe quality ofattachment formed withthe infant, andhis
conclusions are largely based on the empirical literature showing a signiﬁcant
relationship between mothers’ representations of their own attachment with caregivers
and the attachment formed with their infants.

During pregnancy, a mother also develops representations of her baby. Stern
concludes ﬁom the available literature that there is an increase in richness and speciﬁcity

of maternal representations of her fetus-as-inEnt around 4 months gestation, possibly

61

precipitated by the baby’s movement and growth. Although Stern suggests that
representations may decline during the last trimester of pregnancy in order to protect the
mother and baby from a potential discrepancy between the real and imagined baby,
existing studies actually indicate that women continue to have richly developed
representations during their entire pregnancy, as well as postpartum (e. g., Fava Vizziello
et al., 1993; Leifer, 1977). In addition, a number of studies indicate that representations
of the infant may be a mechanism through which attachment is transmitted ﬁom mother
to infant (e.g., Ammaniti, 1991).

Finally, another important representation during a woman’s transition to
motherhood is her representation of self as a mother. Stern (1995) hypothesizes that an
important internal transformation takes place, as the mother shifts from being a daughter
to her mother to being a mother to her daughter/son. As part of this transformation, a
mother’s representations of herself in various roles (e.g., wife, daughter, and mother) will
be activated and reorganized as she makes psychological room for her role as a mother.
Some empirical literature has indicated, in fact, that representations of self as mother and
beliefs about caregiving efﬁcacy are related to both representations of attachment and
mother-infant attachment.

Stern (1995) also asserts that there are a number of important themes within the
mother’s psychological and external world as she makes the transition to motherhood.
First, the mother is concerned with helping the infant grow and thrive physically.
Second, the mother is concerned with establishing an emotional relationship with the
baby, during which there is ideally a lessening of the mother’s own subjectivity for the

purpose of being acutely in tune with the infant. The mother also develops a need to

62

create and maintain a protecting support network so that she can accomplish the tasks of

motherhood. This support network ideally provides the mother with physical and

practical help, as well as with emotional and psychological support. Stern asserts that
support ﬁom females speciﬁcally, and also females that have had experiences with
motherhood, is especially important. Finally, the fourth theme concerns the mother’s
own representations of herself, as mentioned above, the mother must reorganize her
identity to include representations of self-as—a mother, which are thought to be inﬂuenced
by the mother’s representations of being cared for during childhood.

Although there is a substantial amount of literature supporting Stern’s (1995)
ideas, no empirical study to date has integrated this literature in a way that has been
theorized and described by Stern. Furthermore, neither Stern’s theory nor most of the
existing empirical studies have addressed processes related to the formation of mother-
infant attachment in high-risk groups of women; instead, they have focused primarily on
mother-infant attachment in well-educated, middle-class families. Therefore, this study
will test many of Stern’s ideas, which are also consistent with attachment theory, within a
high-risk group of women through a longitudinal research design (see Figure 1 for
hypothesized model).

Speciﬁc hypotheses for the current study were as follows:

1. Mothers’ representations of attachment during pregnancy, based on the
recollection of experiences with caregivers, will be signiﬁcantly related to
mothers’ representations of their infants during pregnancy. More secure
representations of attachment will be related to more secure or “balanced”

representations of infants during pregnancy.

63

 

 

 

   

  

8.92 8835 .358

is.

 

0:89:00 .0633:

6:3

.98... .223

 

Ea:— .3
cctﬁcomoaoa
.8985

E25082

Ewan—LoﬁoZ

 

 

_ btaoow
.5032

 

 

 

     
   

seemeom
BESS).

.5330;

      
    

     

.5522
3 row we
53958.63.
.8295

 

 

 

 

 

 

 

 

 

 

uncneu 68.8..
2 :25 365.2:
963
as...
Lon—.5
.592
.coE:uaz< he
22.25353—
?EBSZ coach:
239.00
.83.

tome—m

 

 

 

beam
.023
38m
335%:—
tonaam toga—am .335
833030 .332.— dozen—m

 

 

 

 

 

 

 

£2.55 N_ 8 E25332 .5..:.-..u£e.2 he 9.29:5... .3358.— .Ea 36:3.— ..o .25.)— 25358»:

atom

33:95

 

 

tenazw

 

 

cozwueau

redeem

 

 

 

78:22A—

:aqzm
6285”,.

 

 

 

._ 25m:

Mothers’ representations of attachment during pregnancy will be signiﬁcantly
related to the way in which mothers perceive themselves as caregivers during
pregnancy. More secure representations of attachment will be related to more
secure representations of the selfas a mother, characterized by a conﬁdent, but
realistic view of the selfin the maternal role.

Representations of the infant and representations of the selfas a mother during
pregnancy will be complementary, but distinct, constructs and will be
signiﬁcantly related to each other.

Mothers’ representations of the infant will be signiﬁcantly related to mother-
infant attachment security at 1 year postpartum. More secure, balanced
representations of the infant during pregnancy will be related to greater infant
security at 1 year. Thus, maternal representations of attachment will be indirectly
related to mother-infant attachment at 1 year through its impact on representations
of the infant.

Representations of the self that emerge during pregnancy, as the women prepare
for the. motherhood role, will be signiﬁcantly related to mothers’ self-efﬁcacy in
the maternal role after the infant is born. More secure representations during
pregnancy will be related to greater feelings of postnatal self-efﬁcacy.

Mothers’ self-eﬁicacy in the maternal role at 1 year postpartum will, in turn, be
signiﬁcantly related to infant attachment security. Greater feelings of efﬁcacy will
be related to greater attachment security. Thus, maternal representations of the

selfas a mother and postnatal self-efﬁcacy will be another pathway through

65

which mothers’ attachment will be indirectly related to infant attachment at 1
year.

7. Prenatal emotional support, practical support, and support with pregnancy-related
issues ﬁom females speciﬁcally will be signiﬁcantly related to postnatal, female
support in the same three areas. That is, greater support during pregnancy will be
related to greater support postpartum.

8. Postnatal support wilL in turn, be signiﬁcantly related to maternal self-efﬁcacy,
with more support ﬁom females related to greater feelings of self-eﬂicacy. Thus,
postnatal support will indirectly affect mother-infant attachment through its
impact on maternal self-efﬁcacy.

9. Several risk factors, including demographic risks, maternal depression, domestic
violence, and child abuse and/or neglect, will be signiﬁcantly related to mother-
infant attachment, with more risk factors being related to less secure attachment.

ExploratoLv Analyses
In addition to testing Stern’s theory as depicted in Figure 1, exploratory analyses

are proposed to assess the transmission of attachment over time using attachment

categories from the three attachment measures (mothers’ representations of their own
attachment, representations of the infant, and mother-infant attachment at 1 year). The
degree of concordance between attachment categories is believed to be another important
way of examining the continuity or discontinuity of attachment patterns between mother

and infant.

66

CHAPTER 5: METHOD
Participants

Participants included 207 women, who were recruited and enrolled in conjunction
with a larger study at Michigan State University. Women initially enrolled in the study
during their last trimester of pregnancy (Time 1) and were interviewed again when their
inEnts’ were approximately 2 months old (Time 2) and when their infants were
approximately 12 to 15 months old (Time 3). At Time 2, one woman had a reported
stillbirth and four women were unable to be located by research staﬁ‘. At Time 3, one
woman who was missing at Time 2 was found and interviewed, another woman reported
that her child had died and an additional 12 women were unable to be located or refused
to participate. Thus, the retention rates were 98% and 92%, respectively.

Women were recruited with ﬂyers (see Appendix A) throughout the Clinton-
Eaton-Ingham counties of Michigan at a number of agencies and clinics including:
Obstetric/Gynecology clinics or other women’s health clinics (39%), ﬂyers posted in the
community at libraries, laundromats, stores, and other similar public places (27%), a
number of social service programs such as FIA, W1C, Head Start, Jump Start, and
Maternal Infant Outreach Program (26%), childbirth classes (5%), the county
prosecutor’s oﬁce (2%), and a local domestic violence shelter (1%).

General inclusion criteria included: 1) in the last trimester of pregnancy at the
time of the initial interview, 2) 18 to 40 years of age, and 3) involvement in a romantic
relationship for at least 6 weeks sometime during the pregnancy. Recruitment eﬁ‘orts
were made to include women who were ethnically and socioeconornically diverse. That

is, speciﬁc recruitment sites were chosen to include women with a variety of backgrounds

67

and experiences. See Table 1 for demographic information on the entire sample. There
were no signiﬁcant demographic differences between participants who completed Time 2
and Time 3 interviews (9 = 190) and participants who did not complete follow-up
interviews (a = 17), with the exception of Emily income; participants with missing data
had signiﬁcantly less income per month than participants with completed data (1 = 3.91 , p
< .01 ).

Procedures

Women contacted the study ofﬁce if they were interested in participating, at
which time a research assistant conducted a brief screening to determine eligibility. The
women were told the study was about women’s relationships with the important people in
their life, including partners, family members, and children, and that if they participated
in the study they would be asked about their thoughts and feelings about relationships and
recent life events, including domestic violence. They were also told that they did not
necessarily need to have experienced domestic violence in order to be eligible for the
study. If the potential participant met criteria and agreed to participate, an appointment
was made.

Because the purpose of the larger study, with which this study was conducted,
was to study the effects of domestic violence on women and infants, additional screening
procedures and recruitment eﬂ°orts were made to ensure that a reasonable number of
subjects had in fact experienced domestic violence during the current pregnancy. After
about halfthe sample had been recruited and interviewed (r_r = 96), the Conﬂict Tactics
Scale (Straus, 1979) began to be administered over the telephone during the initial

screening procedures in order to keep track of the number of battered and non-battered

68

Table 1

Demographic Characteristics of Study Participants (N = 207)

 

 

 

Age of Mother M = 25.5 (SD = 5.0)
Gender of Child" b
Boy 104 (51%)
Girl 99 (49%)
Racial/Ethnic Group of Mother
Caucasian 130 (63%)
Aﬁican-American 53 (25%)
Latina, Hispanic 10 (5%)
Biracial 8 (4%)
Native American 2 (1%)
Asian American 1 (1%)
Other 3 (1%)
Racial/Ethnic Group of Child
Caucasian 91 (44%)
Aﬁican-American 47 (22%)
Latina, Hispanic 4 (2%)
Biracial 62 (30%)
Asian American 1 (1%)
Other 2 (1%)
Marital Status
Never Married 104 (50%)
Married 83 (40%)
Divorced 10 (5%)
Separated 9 (4%)
Widowed 1 (1%)
Education Level of Mother
High School 94 (45%)
Some College 80 (39%)
College Degree 16 (8%)
Graduate Degree 11 (5%)
Other 6 (3%)
Income/Month M = $1806.29 (S_D = $1506.92)
Socioeconomic Status M = 32.13 (SD = 11.17)
‘1 stillbirth
b 3 unknown

women. This screen was used to exclude women who had not experienced domestic

violence during pregnancy once it was determined that there were enough non-battered

women. After 137 participants had been recruited and interviewed, it was discovered that

many of the “non-battered” women had actually experienced domestic violence in a prior

69

relationship. Thus, the telephone screen was then also used to enroll women who had
never experienced domestic violence, in addition to those who experienced violence in
the current pregnancy. Overall, 161 women who called the project ofﬁce to participate
were not included for the following reasons: did not meet battering experience criteria (a
= 102), gave birth before the interview or could not be found again to schedule an
interview (a = 41), did not meet relationship criteria (a = 9), planned to give baby up for
adoption (p = 3), did not meet age criteria (a = 2), did not speak ﬂuent English (a = 2), or
decided not to participate after the phone screen (a = 2). There were no demographic
differences between these excluded women and participants.

_T__ime 1 I nmew

Eight undergraduate students and ﬁve graduate students were trained to
administer all questionnaires and a semi-structured interview (see description below)
before Time 1 interviews were conducted. Research assistants attended a weekly training
meeting for one semester and did two to ﬁve practice interviews under supervision until
they reached 95% reliability for standard administration of measures. That is,
interviewers were required to perform administrations with at least 95% accuracy.
Throughout the period of data collection, research assistants continued to attend a weekly
training meeting during which procedures were reviewed and interview problems were
discussed. In addition. supervisors reviewed all completed interviews to ensure
questionnaires and interviews were being administered correctly.

During the Time 1 interviews, mothers were informed about anonymity and
conﬁdentiality and ﬁrst completed an informed consent form (see Appendix B) that

speciﬁed that her participation was voluntary and that she could withdraw ﬁom the study

70

without any negative consequences. Interviewers read all questionnaires aloud and
marked down participant’s responses in order to control for variation in the level of
literacy among participants. The semi-structured interview was audiotaped for later
transcription and coding. Interviewers were blind to the battering status of the women,
which was ensured by administering violence questionnaires last. Conﬁdentiality was
maintained by assigning all participants an identiﬁcation number, which was placed on
all data rather than participant names, and participant names were kept apart ﬁom the
data. Time 1 interviews took place in the woman’s home or in the project oﬂice,
whichever she preferred, and lasted approximately 3 hours. Participants were paid
$50.00 and were given a list of community resources available for women and children.

Time 2 Interview

Participants were contacted by project staff approximately 1 week after the due
date to conﬁrm each infant’s date of birth. Women were contacted again approximately
6 to 8 weeks after the infant’s date of birth to set up the Time 2 interview, which nearly
always took place between 2 and 3 months postpartum. During the ﬁrst week of trying to
contact women for this interview, research assistants made up to 10 calls to the
participant’s home and up to four calls to her work. If the participant could not be
reached by telephone (due to a disconnected phone, for example), a letter was sent to her
home requesting that she call the ofﬁce. If, by the second week, participants had not
been reached, research assistants made up to two visits to her home. If the woman was
not at home, a letter was left requesting that she contact the project oﬁce. If necessary,
the following week, efforts were made to contact women through ‘recontact’ people,

whose names, addresses, and telephone numbers were provided by participants during

71

their ﬁrst interview. Recontact people were both called (up to ﬁve times) and were
contacted directly through several visits to their homes. If recontact people were not at
home, letters were left asking them to forward the information to participants.

Eight research assistants were trained to administer the Time 2 interview by
conducting several practice interviews with each other, observing a supervisor conduct a
real interview, and doing their ﬁrst interview with a participant under supervision.
During data collection, interviewers attended a weekly meeting to discuss diﬂiculties that
arose during interviews, as well as strategies to reach women who were “missing.” Time
2 interviews were primarily done over the telephone (less than 10 participants did the
interview in person because they had no phone) and lasted about 30 to 40 minutes.
Information about conﬁdentiality was read aloud to participants before the interview
began (see Appendix C). All questionnaires were read aloud, and interviewers wrote
down women’s responses. Women were given a baby gift worth $5.00 following the
completion of the interview.

Trac ' Partici ants Over Time

After the Time 2 interview, women were followed through the infants’ ﬁrst year
of life every 90 days in order to stay in touch with women and minimize attrition. Thus,
they were contacted at about 5 months, 8 months, and 11 to 12 months postpartum. At
each follow-up point, a letter and information sheet (asking for the participant’s current
address and phone numbers) were sent to the participant (see Appendix D), along with a
self-addressed stamped envelope. In addition to returned information sheets, the post
ofﬁce sent a postcard to the project oﬂice with current addresses for those participants

who had moved and registered a new address. Participants were called directly if they

72

did not return the information sheet or if the post ofﬁce did not send new inforrmtion
within 3 weeks. If participants were unable to be reached directly at that point, efforts
were made to contact them through their recontact people. If all other tracking methods
failed, research assistants went to the participants’ and recontact people’s homes.

Time 3 Interview

When participants’ infants were approximately 11 to 12 months old, women were
contacted using the same tracking procedures described above, to see if they were willing
to participate in a follow-up interview with their infants. Although efforts were made to
complete all Time 3 interviews within two weeks before or after the infant’s date of birth,
this was not always possible due to problems ﬁnding and scheduling participants within
that time ﬁame. Thus, 54% of the interviews were completed when infants were 11 to 12
months old, 39% were completed when infants were 13 to 14 months old, 5% were
completed when infants were 15 to 16 months old, and 2% were completed when infants
were 17 to 18 months Old.

A total of 13 undergraduate students and three graduate students were trained to
conduct the mother and infant assessments at Time 3. Training consisted of weekly
meetings and outside activities over a 7-month period. Training activities included: 1)
learning about infant development, assessment, and mother-infant attachment through
readings and discussion, 2) as many practice interviews as was necessary to achieve an
accuracy rate of at least 85% for administration of all measures, including interviews and
videotaped procedures, and 3) observation of supervisors’ conducting real interviews. In
addition, approximately 20% of the interviews were observed by supervisors in order to

control for drift in standard administration.

73

The Time 3 interview took place in the project ofﬁce because video equipment
was required for some laboratory procedures. Mothers ﬁrst completed an informed
consent form for herself and her infant (see Appendix E). Mothers and infants then
participated in a standardized laboratory procedure together (described below) to assess
the quality of their attachment. Afterwards, one research assistant met with the mother to
administer interviews and questionnaires, while another research assistant administered a
few tests to the infant in a separate room (for the purposes of the larger study). The entire
Time 3 interview lasted about 3 to 4 hours. Women were paid $75.00 and given a baby
gift worth approximately $8.00 after completion of the interview.

Measures

Maternal Representations of Attachment (Time 1 )

 

Mothers’ representations of attachment were assessed by the Perceptions of Adult
Attachment Questionnaire (PAAQ, formerly called the Inventory of Adult Attachment;
Lichtenstein & Cassidy, 1991), a 60-item self report that asks about experiences during
childhood and the participant’s perceptions of her mother as a caregiver, as well as the
participant’s relationship with her mother (see Appendix F). All items are rated on a 1 to
5 Likert scale ranging ﬁ‘om Strongly DiagLe' e to Strongly Aggee. Items form eight
subscales: 1) Rejection by mother (11 items), 2) Loved by mother (6 items), 3) Role-
Reversal with mother (10 items), 4) Anger towards mother (5 items), 5) Derogation of
attachment experiences (4 items), 6) Forgiving of childhood problems (7 items), 7)
feelings of Vulnerability about relationship with mother (5 items), and 8) No Memory of

childhood experiences (4 items). Scales 1-5 and 8 were designed to be equivalent to six

74

scales from the Adult Attachment Interview. Higher scores on each subscale indicate
higher levels of the construct being measured, e.g., more rejection, more love, etc...

Lichtenstein and Cassidy (1991) indicated that the PAAQ scales are fairly
internally consistent; alpha coeﬁicients ranged ﬁom .62 (Derogation) to .90 (No
Memory) in a sample of 247 college students and ﬁom .51 (Derogation) to .94 (No
Memory) in a sample of 123 mothers. Test-retest reliabilities ranged from .64 to .86 in
the college sample and were not reported for the mother sample. The authors compared
PAAQ responses to AAI responses in the sample of mothers and found that equivalent
scales were highly correlated (p < .01) with each other, with the exception of the Role-
Reversal and Derogation scales. Thus, since the AAI is currently the “gold standar ” for
the measurement of adults’ representations of attachment, it appears that most of the
PAAQ scales have good comment validity.

Based on attachment theory and that underlying the AAI classiﬁcations,
respondents can presumably be classiﬁed into Autonomous, Dismissive, and Preoccupied
attachment groups based on their scores on the eight subscales. Theoretically,
autonomous adults would report more experiences of love and feelings of forgiveness,
dismissive adults would report more rejection, derogation, and a lack of memory of
childhood experiences, and preoccupied adults would report role-reversal, angry feelings,
and feelings of vulnerability (Hesse, 1999). However, a conﬁrmatory factor analysis on
the current sample revealed that six of the eight subscales loaded onto a more general
attachment factor that measures a continuum of security rather than qualitatively distinct
attachment styles (see the Results chapter for more details). Therefore, only those six

subscales (Rejection, Love, Anger, Derogation, Vulnerability, and No Memory) were used

75

to measure matemal representations of attachment. Coefﬁcient alphas for the current
sample were: Rejection = .91 , Love = .92, Anger = .79, Derogation = .49, Vulnerability =
.65, No Memory = .93.

Finally, although the AAI is the gold standard for measuring maternal
representations of attachment, the PAAQ was chosen in this study due to limited time and
resources, as well as the inclusion of another labor-intensive clinical interview (see
below). Despite the fact that a number of researchers have challenged the use of
questionnaires to measure attachment representations, noting that it is diﬁicult to assess
defensive and narrative processes from questionnaire data (e.g., Crowell & Treboux,
1995), there is reason to believe that questionnaire data can be useful. For example,
many adults can and do provide accurate information on their experiences, and conscious
and unconscious processes often Operate in a similar way to achieve the same goal (see
Crowell, Fraley, & Shaver, 1999 for a review of these issues).

Prenatal Rgpresentations of Infant and of Self-As-Mother (Tm

The Working Model of the Child Interview (WMCI; Zeanah et al., 1994) was
used to measure matemal representations of the infant and of the self-as-mother during
pregnancy. This instrument is a 1 to 1 1/2 hour structured interview that inquires about a
participant’s perceptions and subjective experiences of her infant and relationship with
her infant. Interviews were audiorecorded, transcribed, and rated by trained coders along
13 5-point Likert scales that assess qualitative (e.g., coherence of interview), content
(e.g., infant difﬁculty), and affective features (i.e., anger, joy) of maternal

representations. Higher scores indicate higher levels of the construct being measured.

76

Coders then assign an overall classiﬁcation to the narratives provided by the mothers
based on the scale proﬁles (with an emphasis on the qualitative scales).

Balanced narratives include both positive and negative characteristics of the
inﬁrm and relationship with the infant. They convey value for the infant’s individuality
and appreciation for the infant’s subjective experience. The caregiver’s perceptions are
open to change and are at least moderately rich in detail about the infant and the
caregiving experiences. Disengaged representations are characterized by emotional
distance or indifference towards the infant. Caregivers are unable to recognize the
inEnt’s individuality, and if it is recognized, it is not valued. Details about the infant or
parenting experience lack richness, are not ﬂexible or open to new experience, and are
emotionally unintegrated. Finally, Distorted representations reﬂect general
inconsistencies. Caregivers may be preoccupied or overwhelmed by the infant or may
have unrealistic expectations about the infant. Unlike disengaged parents, distorted
mothers do not deny their impact on their infants. However, they often do not recognize
how their behaviors may be detrimental to the infants. Descriptions of infants and the
relationship with the infant may be incoherent or contradictory. Much feeling is
expressed toward the infant, but these emotions lack a sense of modulation or meaning.

Several studies have demonstrated predictive and concurrent validity for the
WMCI by reporting highly signiﬁcant relationships between WMCI classiﬁcations and
infant attachment classiﬁcations assessed by the Strange Situation in the expected
direction: Balanced-Secure, Disengaged-Avoidant, and Distorted-Ambivalent (Benoit et
al., 1997; Zeanah et al., 1994). Another study provided more evidence for validity by

showing a signiﬁcant relationship between the disengaged and distorted representations

77

 

 

and high-risk/disordered infant status (Benoit, Zeanah, Parker, Nicholson, & Coolbear,
1997). Interrater agreement has ranged ﬁom .57 to .76 (Cohen’s kappa) for different
classiﬁcations in previous studies, and Benoit et al. (1997) reported a high degree of test-
retest stability for the Secure and Distorted classiﬁcations (concordance over 1 year’s
time was 89% and 85% respectively). As of yet, there is no evidence of stability for the
Disengaged classiﬁcation.

In the present study, questions were changed to future tense to inquire about
mothers’ representations during pregnancy (see Appendix G). Benoit et al. (1997) has
demonstrated that this is a valid technique for assessing prenatal representations. Those
subscales that measured qualitative features (with some content features) of maternal
representations were used to assess the mothers’ representations of their infants, since
these are believed to be most theoretically important. These included: 1) Richness of
Perceptions, which measures the poverty or richness of representations of the infant and
the degree to which the mother “knows” the infant, 2) Openness to Change, which
measures the ﬂexibility of the mother’s representation to accommodate new information
about the infant (compared to rigidity), 3) Coherence, which measures the overall
organization of the mother’s narrative about the infant and her relationship with the infant
and the logical ﬂow of her responses, 4) Caregiving Sensitivity, which assesses the
degree to which the mother recognizes and responds adequately to the infant’s own needs
and experiences including a respect for the infant as a separate but dependent individual,
and 5) Acceptance, which measures the degree to which the mother is open and accepting

of responsibilities involved with adequate caretaking.

78

The author and another graduate student were trained to code scales and classify
representations according to the coding system developed by Zeanah et al. (1994), with
the assistance of another research group that has been trained by Zeanah. Adequate inter-
rater reliability was established using weighted-kappa (or corrected-kappa) (Cohen, 1968;
Fleiss, Cohen, & Everitt, 1969) for subscales and both percent agreement and Cohen’s
kappa for classiﬁcations. These were calculated based on 26 interviews that were coded
by both trained coders (13% of the sample). Weighted kappas for the subscales used in
this study were as follows: Richness = .68, Openness = .51, Coherence = .62, Sensitivity
= .69, Acceptance = .77. Percent agreement for overall classiﬁcation was 96%, with a
kappa of .94 (p < .001).

In addition to the ﬁve scales assessing representations of the infant, mothers’
responses on the WMCI were coded for representations of self-as-mother. This code was
adapted ﬁom the Conﬁdence and Competence scale from Slade et aL’s (1994) Pregnancy
Interview Coding System. This subscale assesses the mother’s representations of her
own competence and self-efﬁcacy in the maternal role and her expectations of herselfas
a mother along a 5-point Likert scale. Unlike the other qualitative scales developed by
Zeanah et a1. (1994), the middle point on this scale represents the most “balan ” or
“secure” representation of self-as-mother; mothers who score low on this scale lack
conﬁdence in themselves, while mothers high on this scale are overly and unrealistically
self-conﬁdent (which is believed to be a defensive response to underlying ineﬂicacy).
Thus, a balanced mother would recognize her strengths and limitations and acknowledge
the challenges of motherhood, but would overall feel that she is able to make her baby

feel happy, safe, and secure. For analyses, codes of 2 and 4 were collapsed and codes of

79

1 and 5 were collapsed so that the highest possible code was a 3, which represented the
most balanced representations of self. Weighted kappa for this subscale was .63.

Maternal Self-Efﬁcacy (Time 3)

The Maternal Efﬁcacy Questionnaire (MEQ; Teti & Gelfand, 1991) was used to
measure maternal self-efﬁcacy, or the mother’s perceptions about the degree to which she
is able to perform effectively as a parent, when the infant was approximately 12 months
old. This instrument is made up of 10 items, rated on a 4-point Likert scale, that tap
feelings of efﬁcacy related to a number of speciﬁc domains of parenting such as soothing
the infant (see Appendix H). A total efﬁcacy score is obtained by summing all items, and
higher scores indicate greater self-efﬁcacy. Teti and Gelfand (1991) originally
administered the scale to 86 mothers, about halfof whom were clinically depressed, with
infants between the ages of 3 and 13 months. They reported a coeﬂicient alpha of .86,
indicating good internal reliability. Furthermore, they demonstrated good concurrent
validity by reporting a high correlation between the MEQ and the Parenting Stress Index
Sense of Competence Scale (1 = -.75). Coeﬂicient alpha in the current study was .65.

Mother-Infant Attachment (Time 3)

The Strange Situation (SS; Ainsworth et al., 1978) is a well-known laboratory
procedure that was used to measure mother-infant attachment (see Appendix I). The
procedure consists of eight episodes, each lasting about 3 minutes with the exception of
the ﬁrst episode, which lasts about 30 seconds. The entire procedure takes approximately
21 minutes and is videotaped and coded at a later time.

Infant behaviors during the SS are coded on four 7-point scales including

proximity seeking, contact maintaining, avoidance, and resistance according to a coding

80

system developed by Ainsworth et al. (1978) and one 9-point scale for disorganization
according to a coding system developed by Main and Solomon (1990). Infant behaviors
during reunion episodes are particularly important. Based on patterns of infant behavior
toward the caregiver, the quality of mother-infant attachment is also determined, and
infants are classiﬁed into one of four general categories: Secure (Type B), Insecure-
Avoidant (Type A), Insecure-Ambivalent (Type C), and Disorganized/Disoriented (Type
D). Infants can be further classiﬁed within the three original categories; there are four
types of Secure (Bl — B4), two types ofInsecure-Avoidant (A1 and A2), and two types of
Insecure-Ambivalent attachment (C1 and C2). See Table 2 for a description of each
subgroup.

Reliability and validity of this instrument have been reported by numerous studies
(i.e., Ainsworth et al., 1978; see Solomon & George, 1999 for a review). OveralL
attachment classiﬁcations have been shown to be highly stable, with somewhat less
stability in disadvantaged samples. Interrater reliability is very high (about 85% to 95%)
for researchers trained by Ainsworth and her students. In addition, repeated studies have
found a signiﬁcant relationship between theoretically related variables and inEnt
attachment such as matemal sensitivity and responsiveness, as well as between
attachment and later expected outcomes such as mental health and social competence.

In the present study, SS videotapes were sent away and coded by trained, reliable
coders at the University of Washington (under the direction of Dr. Susan Spieker, who
has extensive experience coding SS tapes). Coders achieved a 90% agreement rate on
overall classiﬁcations, with a kappa of .84 (p < .001), based on double-coding 11% of the

sample. SS categories were then converted by the current investigator into a continuous

81

Table 2

Descriptions of Original Attachment Classiﬁcations by Subgpoup

 

Classiﬁcation

Description

 

Type B: Secure
B3

B1

32

B4

Type A: Avoidant
A1

A2

Type_ C: Ambivalent

C1

C2

Most secure group. Strong proximity-seeking and contact
maintaining. Little or no Sign of avoiding or resisting
caregiver.

Predominantly secure behaviors, with some evidence of
avoidance. May greet mother, but no strong proximity-
seeking Little or no resistance.

Predominantly secure behaviors, with some evidence of
avoidance. More proximity-seeking than B] but not as
much as B3. Little or no resistance.

Predominantly secure behaviors, with some evidence of
preoccupation with caregiver and may show some
resistance. Little or no avoidance.

Most avoidant group, showing active avoidance of
caregiver. Little or no proximity-seeking or contact-

maintaining.

Predominantly avoidant behaviors, but with some tendency
to greet caregiver or seek proximity.

Strong resistance and ambivalent behavior towards
caregiver, combined with extreme distress and
unmistakable angry quality.

Strong resistance and ambivalent behavior towards
caregiver, combined with a high degree of passivity. Not as
angry as C].

 

scale assessing the degree of felt security according to the conversion system suggested

by Bretherton and colleagues (Bretherton et al., 1989; Bretherton et al. 1990): B3 = 5;

Bl,BZ,B4=4;A2,C1 =3;A1,C2=2;D= 1. Thissystemisbasedontheﬁndingsof

82

Ainsworth et al. (1978), which showed that mothers of B3 infants are most sensitive,
followed by mothers of B1, B2, and B4 infants, and so on, and is also consistent with the
security continuum suggested by Cummings (1990).

There are a number of reasons why it is useful to assess attachment security along
a continuum (see Cummings, 1990, for a review). First, not all mother-infant
attachments ﬁt neatly into “prototype” categories. Even when attachments are
classiﬁable, there may be signiﬁcant differences in felt security within the broad
attachment classiﬁcations (within Type B, for example), which are generally used by
researchers due to limited sample sizes. In addition, when classiﬁcations appear to be on
the borderline between two groups, the use of a security continuum can reduce the
possibility of measurement error. Cummings also argues that assessment along a
continuum is conceptually justiﬁed because the infant’s felt security comes from the
attachment ﬁgure and is central to the function of attachment.

Maternal Social Suppprt (Time 1 and Time 3)

The Norbeck Social Support Questionnaire (N SSQ; Norbeck, Lindsey, & Carrieri,
1981) was administered to women during their pregnancy and again at 12 months after
birth to measure mothers’ level of social support (see Appendix J). The NSSQ assesses
multiple dimensions of social support including the quality of Emotional Support,
Practical Support (or Aid), the size of the social support network (the number of
supporters available to the participant), the duration of support, and the ﬁ'equency of

contact with supporters. Three items, which speciﬁcally asked about support related to

pregnancy/preparation for motherhood (during pregnancy) and related to caregiving (at

83

12 months), were developed for this study and added to the original 10 items (see
Appendix K for postnatal additional items).

Norbeck et al. (1981) originally tested internal consistency through item-
intercorrelations based on a sample of college students. They reported correlations
between .72 to .98, with items on the same subscale correlating more highly. Test-retest
reliabilities ranged ﬁom .85 to .90 over a one-week period and .58 to .78 over a seven
month period (Norbeck, Lindsey, & Carrieri, 1982). Norbeck et al. (1981, 1982) reported
good concurrent validity by showing signiﬁcant correlations between NS SQ scales and
other well-known social support scales. In addition, Norbeck et al. (1982) reported
evidence of construct validity by demonstrating that practical support served as a buffer
between life stress and negative mood, as would be expected by theory.

All items are rated on a 0 to 4 scale (ranging ﬁom Not at all to A gpeat deal). The
authors suggest summing the amount of support received by each supporter on each item
and then summing item totals for the Emotional Support and Practical Support scales,
with the assumption that participants with larger networks receive a greater amount of
social support (Norbeck, 1995). However, average satisfaction can also be calculated by
dividing summed totals by the number of supporters and may be a better indicator of the
quality of support. In the present study, the relationship of each supporter to the
participant, as well as each supporter’s gender, was obtained so that support ﬁom females
speciﬁcally could be measured. Thus, average satisfaction from female supporters only
was calculated for Emotional Support, Practical (Aid) Support, and Caregiving Support
(support related to pregnancy and motherhood), and these scores were used to assess the

quality of prenatal and postnatal social support.

84

Risk Factors (Time 1. Tim_e 2. and Time_3_)

 

Several risk factors were assessed at all three interviews in order to examine the
number and types of risk factors participants had experienced, as well as their impact on
mother-infant attachment. Risk factors included demographic risks (e. g., poverty),
maternal depression, domestic violence experiences, and child abuse and neglect.

In order to assess demographic risks, a demographic questionnaire (see Appendix
L) was administered during pregnancy to obtain basic information such as women’s
ethnicity, education and occupation, family income, and Emily composition.
Socioeconomic status (SES) was computed according to Hollingshead Index (1975).
SES was calculated based on the participant only if she was unmarried Md not living
with a partner, while SES was calculated based on both participant and partner if the
participant was either married or living with a partner. A demographic risk scale was
created from several demographic characteristics believed to put women and infants at
greater risk for problems. These characteristics included: a) income below the poverty
level (based on income and family size as deﬁned by the US. Census bureau, 2001), b)
socioeconomic status in the lowest quartile of the sample, and c) single parenthood,
deﬁned as unmarried or not living with a partner. Participants received one “point” for
each characteristic, thus, scores ranged ﬁ'om 0 to 3. For example, if a woman’s income
was below the poverty level an_d_ she was a single parent, she would receive a score of 2.

The Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, &
Erbaugh, 1961), a 21-item self-report, was used to assess the severity of maternal
depression at several time points. The instrument covers a variety of symptoms of

depression such as depressed mood, pessimism, sleep disturbances, and changes in

85

appetite (see Appendix M). Each item consists of four statements that are ranked ﬁom
neutral to severe, and values ﬁom 0 to 3 are assigned to each statement. For example,
question 1 consists of the statements “I do not feel sad,” “I feel sad,” “I am sad all the
time and I can’t snap out of it,” and “I am so sad or unhappy that I can’t stand it.” The
total depression score is obtained by summing the answers of all 21 items. Possrble
scores range ﬁom 0-63, with scores of 0-9 indicating no depression, 10-15 indicating
mild depression, 16-23 indicating moderate depression, and 24-63 indicating severe
depression.

The BDI was originally designed for use in psychiatric populations. Beck et al.
(1961) reported that split-half reliability estimates yielded a coefﬁcient of .93 after a
Spearman-Brown correction, indicating high internal consistency. They also examined
the validity by comparing BDI scores to psychiatric ratings of depression and found a
correlation coeﬂicient of .67, which was highly signiﬁcant. A later study (Bumberry,
Oliver, & McClure, 1978) demonstrated the validity of the instrument in a university
population. Using psychiatric ratings as the criterion, Bumberry et al. found a correlation
coeﬁcient of .77 between the BDI and interview ratings.

In the present study, depression was assessed at Time 1, Time 2, and Time 3.
Coeﬁcient alphas were .85, .89, and .85, respectively. Women who obtained a score of
16 or higher (moderately to severely depressed) were considered “at-risk,” and a
cumulative depression risk score was calculated, ranging from O to 3. Thus, if a wormn
scored 16 or higher at Time 2 only, she received a score of 1, while a woman with scores

of 16 or higher at two points in time received a score of 2.

86

The Severity of Violence Against Women Scales (SVAWS; Marshall, 1992) was
used to measure the type and severity of domestic violence, deﬁned here as male-to—
female violence in adult partners, experienced by mothers at multiple time periods. This
self-report contains 46 items which make up nine dimensions of violence ranging from
threats of mild violence to severe physical and sexual violence. In the original sample of
community women, coefﬁcient alphas ranged ﬁom .89 to .96, indicating high internal
consistency within dimensions.

In the present study, women ﬁlled out the scale for several time periods including:
a) during her pregnancy with the infant, b) the year before pregnancy, c) for a previous
partner (all ofthese assessed at Time 1, see Appendix N), and d) during the ﬁrstyearof
the infant’s life (assessed at Time 3, see Appendix 0). Coeﬁcient alphas for each time
period were: .95, .97, .98, and .95. Because it is believed that experiencing domestic
violence in multiple relationships and/or for longer periods of time is more harmful for
women and child witnesses than single, brief episodes of abuse, the experience of
physical and/or sexual abuse within each time period was considered as a separate risk
factor in order to calculate a cumulative domestic violence risk score. For example, a
woman who reported abuse during pregnancy and during the infant’s ﬁrst year of life
received a score of 2. Abuse is deﬁned here as threats of moderate or severe violence,
mild to serious violence, or sexual violence (as indicated by any item on the scale
between 9 and 46). Scores ranged ﬁ'om 0 (no abuse at any time periods) to 4 (abuse at all
four time periods).

Physical abuse and/or neglect of the participating inEnt were measured in two

ways at Time 3. First, mothers completed the Parent Behavior Checklist (PBC; Fox,

87

1994), a 100-item questionnaire on parenting behaviors of infants and toddlers. Each
item is rated on a 4-point Likert scale ranging ﬁ'om _l_\l_e_veg to _Al_yv_ay§. The scale
comprises three subscales, however, only the Discipline subscale was used in the present
study. This subscale includes 30 items and assesses the amount of physical punishment

exerted by the parent (see Appendix P). Higher scores indicate better discipline

 

techniques and less physical punishment. Fox (1994) reported good reliability and
validity for the PBC; coeﬂicient alpha for the Discipline subscale was .91, test-retest
reliability was .87, all items had a loading of at least .30 for the Discipline factor, and the
scale was highly correlated with previously-developed parenting questionnaires. In the
present study, coeﬁcient alpha for the Discipline subscale was .85.

A second way that child abuse/neglect was assessed in the current study was
through a 30-item child neglect questionnaire developed for the larger study (Bogat &
Levendosky, 1999) with which the present study was being conducted (see Appendix Q).
Each item is rated on the same Likert scale as the PBC, and a total neglect score is
obtained by summing all items, with higher scores indicating more neglect. The
coeﬂicient alpha for this scale was .66. For purposes of the present study, a child abuse
risk score was calculated using both measures of abuse/neglect. Participants received one
“point” if their scores on the PBC Discipline scale were among the lowest quartile of the
sample and one point if their scores on the Neglect scale were among the top quartile of
the sample. For example, if a woman’s score on the Neglect scale only was among the
highest quartile in the sample, she would receive a score of 1. Thus, higher scores

indicated more child abuse risk.

88

CHAPTER 6: RESULTS

Missmg' Data

Due to attrition over time and a small number of data collection errors, several
pieces of data were missing at all time periods. At Time 1, social support and WMCI
data were missing for one participant because the interview was interrupted and the
woman was never located to ﬁnish the interview (r_r = 206). WMCI data were also
missing for four additional participants due to recording errors (a = 202). At Time 2,
maternal depression was missing on ﬁve participants; researchers could not locate four
women and one additional woman had a reported stillbirth (r_r = 202). All Time 3
variables were missing on 17 participants; one woman had a stillbirth, one woman’s child
died about six months after birth, nine women were unable to be located by researchers,
and six women refused to do the interview (a = 190). Time 3 questionnaires about the
child (PBC, MEQ) were missing on two additional women due to having no contact with
their babies since shortly after birth (and thus, they could not ﬁll out the measures in a
valid manner) (p = 188). Finally, there were 10 additional missing Strange Situation
videotapes; one was accidentally not recorded on tape, one woman refused to do the
taped portions of the interview, one woman lost custody of her baby and could not bring
the baby into the lab (although she ﬁlled out questionnaires about the child), and seven
women lived out of state and were only able to complete questionnaires through the mail
(p = 178).

In order to maximize statistical power, all missing data were estimated using the
imputation method known as the “hot-deck” method from the Lisrel 8.5/Prelis 2.0

software (Jdreskog & SOrbom, 2001) after all recodings and data transformations. This

89

method of estimation substitutes real values for missing values. Real values are obtained
ﬁom another participant’s responses that are most closely matched to the participant with
missing data on a set of speciﬁed matching variables. This procedure has several
advantages: 1) the values that are substituted into missing data points are within the
actual range of possible scores on that particular measure, and 2) the results of the
imputation are not affected by the order of cases in the database. In the current study,
missing data points were estimated based on the following set of matching variables in
the database: income, Time 1 maternal depression, Time 1 domestic violence scores (all
three time periods measured at Time 1), and Time 1 PAAQ subscales (all six).
Estimation was successful, evidenced by the very similar descriptive data for measures
both before and after estimation (see Table 3). Thus, measurement and structural model
testing was based on data from 207 participants.
Data Reduction

See Table 3 for descriptive data on all measures in the hypothesized model. As
the table shows, all measures had adequate variance with the possible exception of Time
3 self-emcacy, which was heavily skewed towards the positive end of the scale. All
variables had satisfactory distributions as evidenced by skewness and kurtosis values in
the appropriate ranges (: 2 and i 4), with the exception of Time 1 and Time 3 Emotional
Support. The latter support variables had elevated kurtosis scores above 4. Because the
cumulative risk variables were created ﬁ'om multiple measures and in order to help
illustrate the numbers and speciﬁc types of risk factors experienced by participants over
time, Tables 4 to 7 are provided. See Appendix R for the correlation matrix of model

variables.

90

 

8a a: 25 2-0 2a was 5&5 3235 .2
2.: :3. no we 8.: man eaaam .8228: .2
NS 8.: 2a 2a EN 2.: e895 3895 .2
ladle am alaolm l: calm
8a 2.2 2.2 as: an 2.: sea 35 .2
“Owen
a. 2a 2 E 8. a3 sec: 828.8 as: .2
8.: 2a a. m: 8: 2.2.. nonsense. .:
a. :3 a: E 3. cam 22:83 .2
a. EX 2 a: 3. and 8:228 .a
2. 2a 3 2 2. 2a 2:20 a 8580 .w
a... 32.. E 3 8. mam 22385 no anion: s
”i523
$2. 2.2 a? 84 Se 2.2 been: oz 6
2a 2.2 Ba new 23 8.2 0:855 .m
SN :3 84 2-2. New .3 3225 .4
Ra. :..: Ra Ra Re :1 _ 35.. .m
an 2.8 2-0 one 22 :18 ”>3 .N
a.” 2.8 2-: 2-: a.” 2.8 :88? ._
Hallie

33 0.8.33 38 220$ 9222230 .2328 323533 2323
Om 5.62 amend 03239“ mocoom we amend Gm 522 c.5982

 

 

ﬂow u E .252 woﬁmoﬁommm 05 5 mogmmlolvz .8.“ ﬂea 339.880

m 033.

91

 

3.—

me.
5.—
mm.
8..

5N“...
mm;
woﬁ.

Nm.m

Nm.
mm.—
mm.
co.—

3.x.
36
him _

NTo
We
26

NT:
mé
oTO

voN
ow;
3N

an 525% 285822 «N
”notwamm o 5.5m

 

2. 2830223.... EEO .2
an: 853$ 028.8 .mm
mm. commmoaoo 3:532 Am
8: a2 02.28388 .8

3

m3 2225 3:35 .2
26 e895 388.: .2
:2 eaaam 5285 .2

23%

APEoov m 2an

92

Table 4

Breakdown of Participants Emriencmg’ Demographic Risks

 

 

 

 

 

 

 

 

 

 

 

 

Number of Risk Factors Frequency
0 88 (42%)
1 53 (26%)
2 44 (21%)
3 22 (11%)
Table 5
Breakdown of Participag' ts Experiencgg' Depression
Number of Times Frequency
0 154 (74%)
1 34 (17%)
2 15 (7%)
3 4 (2%)
Table 6
Breakdown of Particgp' ants Experiencmg' Domestic Violence
Number of Times Frequency
0 52 (25%)
1 56 (27%)
2 46 (22%)
3 31 (15%)
4 22 (11%)
Table 7
Breakdown of Particm’ ants with Child Abuse/Neglect Ris_k
Number of Risk Factors Frequency
0 121 (58%)
1 61 (30%)
25 (12%)

 

93

Measurement Models

Before testing the overall model shown in Figure l, conﬁrmatory factor analyses
(CFA) were conducted for each latent construct with more than one indicator, in order to
determine the adequacy of the measurement models. In general, all CF A models were
speciﬁed so that the residual variances of the factors and indicators were estimated,
assuming that both factors and indicators had some measurement error. Residual
covariances were freed as needed, when this led to a signiﬁcantly improved model ﬁt.

All measurement model estimation was done using the Unweighted Least Squares (ULS)
method, and models were speciﬁed on the y-side. A good ﬁt was determined by: a non-
signiﬁcant chi-square value and/or a chi-square value that was less than two times the
degrees of ﬁ'eedom, a Root Mean Square Error of Approximation (RMSEA) value of less
than .05, and a Goodness of Fit Index (GFI) and Adjusted Goodness of Fit Index (AGFI)
between .90 and 1.00 (Hoyle & Panter, 1995; Schumacker & Lormx, 1996; Ullman,
1996). Lisrel 8.5 was used for all measurement models.

The ﬁrst CF A was conducted for the construct Maternal Representations of
Attachment. This factor had six indicators, which were the six subscales of the PAAQ.
In order to obtain a good model ﬁt, residual covariances were freed for the following
indicators: Derogation and No Memory, Rejection and Love, Anger and Vulnerability,
Rejection and No Memory, and Rejection and Anger. Subsequently, the model had an
excellent ﬁt according to several goodness-of-ﬁt indices. The _X_2 value was 4.25, df = 4,
with p = .37. The RMSEA was .02, GFI = 1.00, and AGFI = 1.00. All factor loadings

were signiﬁcant. See Appendix S.

94

The second CFA examined the construct Prenatal Representation of the Infant.
This factor had ﬁve indicators, which included the ﬁve qualitative and content subscales
of the WMCI. The model ﬁt was excellent. Chi-square = 6.64, df= 5, p = .25, RMSEA =
.04, GFI = 1.00, and AGFI= 1.00. No residual covariances needed to be estimated. All
factor loadings were signiﬁcant. See Appendix T.

The third and fourth CF A analyses examined the Prenatal Social Support and
Postnatal Social Support constructs. Both factors had three indicators each, which
included Female Emotional Support, Female Practical Support, and Female Caregiving
support as measured by the Time 1 and Time 3 NSSQ. Both measurement models were
initially saturated, so in order to obtain more parsimonious models, residual variances for
the ﬁrst and third indicators were set equal. The resulting model for prenatal support
showed a very good ﬁt: 3} = 1.99, df= 1, 12 = .16, RMSEA = .07, GFI = 1.00, and AGFI
= 1.00. All factor loadings were signiﬁcant. See Appendix U. The model for postnatal
support was not as strong with only some indices showing a good ﬁt: X2 = 10.72, df = 1,
p < .05, RMSEA = .22, GFI = 1.00, and AGFI = .98. All factor loadings were
signiﬁcant. See Appendix V.

Finally, a CFA was conducted for the Risk Factors construct. This factor initially
had four indicators including the demographic risk scale, maternal depression risk scale,
domestic violence risk scale, and the child abuse/neglect risk scale. The model had an
excellent ﬁt according to several indices: 32 = 3.90, df= 2, p = .14, RMSEA = .07, GFI =
1.00, and AGFI = .98. No residual covariances needed to be freed. All factor loadings

were signiﬁcant. See Appendix W.

95

After the full model was tested and modiﬁed (see full description below), the risk
construct was changed to include only Time 1 risk factors. Thus, the demographic risk
scale was unchanged, the maternal depression scale was changed to a dichotomous
variable for presence/absence of depression at Time 1, the domestic violence scale was
changed to a 0 to 3 scale by eliminating Time 3 domestic violence (thus, only including
the three time periods of: pregnancy, year before pregnancy, and previous partner), and
the child abuse/neglect scale was eliminated completely. The means and standard
deviations for the new depression and violence scales were M = .20, S_D = .40 and M =
1.29, SD = 1.05, respectively. Skewness and Kurtosis values were within the appropriate
ranges. The new frequencies for these risk variables are shown in Table 8. A CFA was
then conducted on this new Risk Factors construct, which included three indicators of
Time 1 risks. The model was initially saturated, so the residual variances for the ﬁrst and
second indicators were set equal in order to obtain a more parsimonious model.
Subsequently, the model ﬁt was good: X2 = 4.68, df= 2, p = .09, RMSEA = .08, GFI =
.98, and AGFI = .95. No residual covariances needed to be freed. Factor loadings were

signiﬁcant except for ‘bdirisk’, which was almost signiﬁcant (1 = 1.38). See Appendix X.

 

 

Table 8 .

Breakdown of Participants Experiencing Depression and Domestic Violence at Time 1
Number of Times Frequency
Depression:

0 166 (80%)
1 41 (20%)
Domestic Violence:

0 58 (28%)
1 66 (32%)
2 49 (24%)
3 34 (16%)

 

96

Full Structural and Measurement Model

Model testing was guided by the theoretically-derived model shown in Figure 1.
The model included 8 latent variables and 24 indicators, all of which were deﬁned as
endogenous variables. Model testing was based on the covariance matrix of the
indicators, using Lisre18.5 (Jereskog & Sbrbom, 2001). The matrix of factor loadings
was speciﬁed as ﬁll], indicating an asymmetric matrix and allowing for reciprocal
relationships between variables. The theta-epsilon matrix (indicator residuals) and the psi
matrix (factor residuals) were speciﬁed as symmetrical and ﬁxed. Residual variances of
the factors and indicators were estimated, assuming that both factors and indicators had
some measurement error. Criteria for a good model ﬁt were the same as those used to
evaluate the measurement models.

The method of estimation used was Maximum Likelihood (ML), which is the
most ﬁ'equently used and recommended method of estimation for SEM analyses (Ullman,
1996). This method maximizes the likelihood for parameter estimates given the data and
performs well under less than optimal conditions (Hoyle & Panter, 1995). Results
initially revealed that the data poorly ﬁt the model. The £2 value was 563.72, df = 243, p
< .05. In addition, the RMSEA was .11. Although the GFI = .92 and the AGFI = .90,
other ﬁt indices such as the Non-Normed Fit Index and the Comparative Fit Index had
negative values. Overall, these indices indicated signiﬁcant problems with the model.

The modiﬁcation indices provided by the program were used to guide
modiﬁcations to the model, along with a determination about whether or not a
modiﬁcation made theoretical sense. One of the ﬁrst modiﬁcations to the model was

changing the two-way arrow between prenatal representations of the infant and of the self

97

as mother to a one-way arrow predicting representations of the selffrpp; representations
of the infant. The next major modiﬁcation was to eliminate the estimated parameter
between the Risk Factors construct and the Mother-Infant Attachment construct. Instead,
the data suggested that Risk Factors more strongly predicted Prenatal Representations of
the Infant. However, this posed conceptual problems because the predictor, Risk Factors,
included variables measured at Time 3, while Representations of the Infant was measured
at Time 1. Temporally, it would be impossible for a Time 3 variable to predict a Time 1
variable. Therefore, the Risk Factors construct was changed to only include Time 1
variables (see description of the new measurement model above). Subsequently, the
pathway ﬁ'om Risk Factors to Prenatal Representations of the Infant was ﬁeed and
estimated. These changes substantially improved the model (see Appendix Y for the
revised correlation matrix of variables with the new risk variables, upon which the ﬁnal
model was based).

In order to obtain a good model ﬁt according to the criteria deﬁned above, several
other modiﬁcations were made. Because there were three factors with only one indicator,
several residual values were set to a ﬁxed value of 1 (for the observed variables postnatal
self eﬂicacy and infant attachment security). In addition, covariances between the
following pairs of latent constructs were freed: l) Maternal Representations of
Attachment and Prenatal Social Support, 2) Postnatal Social Support and Mother-Infant
Attachment, and 3) Prenatal Representations of Inﬁint and Mother-Infant Attachment.
Finally, a number of residual covariances between pairs of observed variables were freed
and estimated’. Although more modiﬁcations were suggested by the program to further

improve the model ﬁt, no modiﬁcations were made after a good ﬁt was established.

98

The ﬁnal model consisted of eight latent factors and 23 indicators (See Figure 2).
The method of estimation was Maximum Likelihood and model testing was based on the
covariance matrix. Parameter values for the ﬁrst indicator of each construct, including
the single indicators for Prenatal Representation of the Self, Postnatal Self-Eﬂicacy, and
Mother-Infant Attachment, were ﬁxed estimates, and therefore, cannot be interpreted in
terms of signiﬁcance testing. All other observed variables had signiﬁcant loadings on
their respective factors, indicating a very strong measurement model. Residual values for
observed and latent variables can be found in Appendix Z.

All paths between the latent constructs were also signiﬁcant in the expected
direction, with the exception of the paths between Maternal Representations of
Attachment and Prenatal Representations of Self as Mother and between Postnatal Self-
Efﬁcacy and Mother-Infant Attachment. Thus, as hypothesized: a) more insecure,
negative maternal representations of attachment were related to less secure
representations of infants during pregnancy, b) more balanced representations of the
infant were signiﬁcantly related to more balanced representations of the self as a mother,
c) more secure representations of the infant during pregnancy were related to more secure
mother-infant attachment, (1) more secure or balanced representations of the self during
pregnancy were related to greater feelings of self-efﬁcacy after birth, and e) greater
prenatal social support ﬁ'om females was related to greater postnatal social support from

females, which in turn was related to greater feelings of self-emcacy postnatally. Finally,

 

‘Residuals for the following pairs of observed variables were estimated: 22 and 1, 20 and l, 22 and 3, 6 and
4, 5 and 3, 6 and l, 3 and l, 22 and 5,18 and 15, 22 and 2, 20 and 3,14 and 22,14 and 17, 22 and 21, 21
and 6,13 and 21, 9 and 19, 20 and 21, 7and 22, 5 and 7, 7and 23, 7and 19,4and 14, 3 and 16, 5 and 21, 3
and 2]. Refer to Table 3 for the names of these variables.

99

 

 

.8. v a : one? use .

 

 

   
          

   

 

         

E80<
coma. 9
820a,.—
oza ta..- aeczo
cesaacomoaom
.8295
scab.
“55592
Emma—-3582 3.3.

585.2%

  

 

 
    

 

 

 

 

 

3:382
3.3% 3858“— .5502
.582 mam—om no
3265 cornucomoaom
gum .8955

 

 

 

 

 

  

 

 

 

 

   

 

   

 

 

omega deacon.
e. 5.5 9.0531
2.3
ecu .5. .
can” 1 >53.
.b=_=>
«so—f
cone.
\1 5M:<
uCOESOGS< MO can”.
mcozﬁcomoaom
35322
not
48.5.

 

tenazm

 

  
   

toga—am .023
zoom
3358; time.

 

.cc.
toaazm .535 toga—5
339.59 .852; .cezeEm

 

 

 

 

 

 

 

38m

.8395

 

 

5&5 toga—6 €9.95
3330 .839... 628:5

 

 

 

 

 

 

 

22.2: E E 52:55? gee—-5502 be $82.55 EEEmom van 3355 wccoaoo Eve—2 @522 mom .N oSmE

100

more risk during pregnancy was related to less balanced or less secure prenatal
representations of the infant.

The iterations generally converged to show that the ﬁnal model ﬁt very well with
the data. Although the x2 value of242.03 with df= 194 was signiﬁcant (p = .01), x} was
well below two times the degrees of ﬁeedom. The RMSEA = .04 and the GFI = .91.
Although the AGF I was .87, slightly below the cut oﬂ‘ of .90, several other ﬁt indices
were well above .90. Finally, the plot of standardized residuals suggested a nice ﬁt,
evidenced by small discrepancies between the observed and estimated residuals and by
few outliers. See Appendix AA for the ﬁnal model syntax, Appendix BB for a complete
listing of goodness of ﬁt indices, and Appendix CC for the residual plot.

The original model and the ﬁnal model were compared to see if the ﬁnal model
was signiﬁcantly better than the original model. The difference in chi-square values
between the two models (A X2) was 321.69, while the diﬁ‘erence in degrees of ﬁ'eedom
(A di) was 49. This difference was statistically signiﬁcant, indicating that the respeciﬁed
model was signiﬁcantly better than the original model.

B lo to ses: Concordance Between Attachment Grou s

In addition to examining the relationship between mothers’ and infants’
attachment patterns through SEM and with continuous variables, exploratory analyses
were conducted to examine the concordance between mothers’ and infants’ attachment
classiﬁcations over time. Although the PAAQ should theoretically yield three attachment
classiﬁcations that parallel those described in the WMCI and the SS, the CFA conducted

in this study showed that the PAAQ yields only one general ‘attachment security’ factor.

101

Thus, it was only possible to examine the concordance between WMCI and SS
classiﬁcations.

First, the concordance between the three main groups was examined: Balanced-
Secure, Disengaged-Avoidant, and Distorted-Ambivalent. Because the WMCI does not
have a fourth category that is parallel to the SS Disorganized classiﬁcation, all
disorganized infants and their mothers were dropped from these analyses. Chi-square
analyses revealed no signiﬁcant concordance between classiﬁcations, that is, mothers’
attachment classiﬁcations during pregnancy were not signiﬁcantly related to infant
attachment classiﬁcations 1 year later, when three groups were used. The overall
concordance rate was 49% (kappa = .10), with 68% of Balanced mothers having Secure
infants, 26% of Disengaged mothers having Avoidant infants and 22% of Distorted

mothers having Ambivalent infants. See Table 9.

 

 

 

Table 9
Concordance Between Three WMCIan_d SS Classiﬁcations
WMCI

Balanced Disengaged Distorted Total
$
Secure 57 23 13 93 (63%)
Avoidant 15 ll 5 31 (21%)
Ambivalent 11 8 5 24 (16%)
Total 83 (56%) 42 (28%) 23 ( 16%) 148 (100%)

 

1.1 = 148; 29 missing $32, 26 disorganized infants not included, 4 missing WMCI
classiﬁcations.
x2= 3.01, df= 4, p = .55, kappa = .10.

102

In order to increase the sample size and include the disorganized infants, all
Disengaged and Distorted mothers were collapsed into an Insecure WMCI group, while
all Avoidant, Ambivalent, and Disorganized infants were collapsed into an Insecure SS
group. A 2 X 2 cross tabulation was then conducted to examine the concordance between
Secure and Insecure groups. Results revealed a signiﬁcant relationship between the
WMCI and SS classiﬁcations when only two groups were used (60% concordance, kappa
= .19). See Table 10 below. Thus, mothers who were classiﬁed as Secure/Balanced on
the WMCI during pregnancy were more likely to have securely attached infants 1 year
later, while mothers who were classiﬁed as Insecure were more likely to have insecure
infants.

Table 10
Concordance Between Two WMCI Ed SS Classiﬁcations

 

 

Ml.
Balanced Insecure Total
_S__S_
Secure 57 36 93 (53%)
Insecure 34 47 81 (47%)
Total 91 (52%) 83 (48%) 174 (100%)

 

n = 174; 29 missing $32, 4 missing WMCI classiﬁcations.

x3= 6.47, df= 1, p -- .01, kappa = .19.

 

2 Although numerical scores based on the SS were estimated for missing subjects in the SEM analyses
described above, overall classiﬁcations were not estimated and thus were not included in these analyses. It
was believed that estimating classiﬁcations posed a greater risk to data interpretation if errors were made.

103

CHAPTER 7: DISCUSSION

This study was designed to test Stern’s (1995) theory about the psychological
processes involved in the transition to motherhood, as well as the effects of both
intrapsychic and environmental factors on mother-infant attachment as described in the
attachment literature. It was hypothesized that: a) mothers’ representations of attachment
would impact representations of the infant and of the self as a mother, b) the latter two
sets of representations would be related to one another but remain distinct constructs, c)
representations of the infant during pregnancy would be related to mother-infant
attachment at 1 year, (I) representations of the selfas a mother would be related to
postnatal self-eﬂicacy, which would in turn, be related to mother-infant attachment, e)
social support ﬁom females would be related to mother-infant attachment through its
impact on self-emcacy, and t) maternal risk factors would negatively impact mother-
infant attachment at 1 year. In general, the results from this study supported most of the
hypotheses, and thus, provided support for Stern’s theory as well as prior ﬁndings in the
attachment literature. Results will be discussed in more detail next, followed by a
discussion of this study’s strengths and limitations, future directions for research, and
clinical implications of the results.
Re resentatio ocesses

Structural equation modeling results revealed a signiﬁcant relationship between
mothers’ representations of attachment, based on recollections of the relationships with
their own mothers, and mothers’ representations of their infants during pregnancy.
Mothers who reported more negative, insecure attachment experiences characterized by,

for example, more rejection, greater feelings of anger, less love, and trouble remembering

104

events from childhood (believed to be a defensive process involved in insecure
attachment) had more insecure representations of their infants during pregnancy. These
insecure representations were characterized by not only less sensitive and accepting
attitudes towards the infant, but more impoverished, rigid, and incoherent narratives.
These results are consistent with previous empirical and clinical papers, which have
suggested a strong relationship between mothers’ own attachment experiences during
childhood and mothers’ attachment representations of their own infants (Ammaniti, 1991;
George & Solomon, 1996; Fraiberg et al., 1975; Slade & Cohen, 1996). However, this is
the ﬁrst study to demonstrate this relationship prenatally y; empirically, rather than
through case studies or with postnatal samples.

At this time, it is still somewhat unclear how mothers’ attachment experiences
inﬂuence their representations of their infants, and speciﬁcally attachment-related
representations, but some researchers have speculated about the nature of this
relationship. In one of the earliest papers, Fraiberg et al. (1975) suggested that mothers
who have not had their own attachment needs met are unable to be empathically attuned
to their infants because their own needs interfere with their ability to do so. In other
words, mothers with insecure attachment experiences may not be able to decrease their
own subjectivity for the purpose of being acutely in tune with the infant or may not be
able to make psychological room for the infant, which Stern (1995) proposes are
important transformations that ideally occur during pregnancy. As a result, they may be
unable to think about and discuss their infants in a non-distorted, coherent, and ﬂexible
way. Similarly, Fonagy et al. (1991, 1993) have proposed that mothers who have not had

attachment needs met do not develop the ability to understand the mental states of self

105

and others (lacking ‘reﬂective self-function’), which leads to greater diﬁiculty
anticipating their infant’s needs, ultimately leading to less sensitive caregiving.
Furthermore, Fonagy et al. stated that mothers who are more defensive in their thinking
about attachment (e.g., display idealization or repression) are less able to understand or
be in tune with their infant’s mental world. Clearly, there is much to be learned and these
are all interesting hypotheses to examine in future research.

Surprisingly, model results did not reveal a signiﬁcant relationship between
mothers’ representations of attachment and mothers’ representations of themselves as
mothers during pregnancy. At ﬁrst glance, this appears to suggest that the way in which
mothers see themselves as caregivers during pregnancy is not inﬂuenced by mothers’
experiences of being cared for during childhood, which is contrary to Stern’s (1995)
theory and to prior literature (Ammaniti, 1991; Ammaniti et al. 1992; Fava Vizziello et
al., 1993; Slade & Cohen, 1996). However, this is not believed to be the case.

First, bivariate correlations (see Appendix R) in this study indicated that many of
the observed variables which measure maternal representations of attachment were, in
fact, signiﬁcantly related to prenatal representations of the selfas mother. That is, more
positive, secure experiences dm'ing childhood were related to more secure representations
of the self, characterized by a conﬁdent and realistic view of the selfas a caregiver. In
addition, model results indicated that there was a very strong, signiﬁcant relationship
between representations of the infant and representations of the selfas a mother, as
hypothesized. In fact, this relationship was almost perfect, with more secure
representations of the infant related to more secure or balanced representations of the self.

Thus, because the two representational constructs (infant and self) were so highly related

106

and shared so much variance, only one signiﬁcant pathway ﬁ'om maternal representations
of attachment to representations of the infant emerged. That is, the SEM analyses
parceled out the shared variance and suggested that this was the best model ﬁt for the
data.

There is also an important theoretical explanation for these results that should be
considered. Because there was a near perfect relationship between representations of the
infant and representations of the selfas mother, it is possible that these sets of
representations are not distinct, as hypothesized, but rather are part of one integrated
“caregiving” representational system as suggested by George and Solomon (1989, 1996).
These researchers suggest, as others have (e.g., Bowlby, 1973; Bretherton & Munholland,
1999; Slade & Cohen, 1996), that representations of the selfand other are related and
develop ﬁom working models of early attachment experiences. However, they seem to
take this one step further and imply that in the context of an attachment relationship and

from the parental perspective these representations are integrated into one
representatioml system. Thus, representations of the self as a caregiver are integrally
related to representations of the infant and these should not be considered distinct sets of
representations. In other words, the internal working model of caregiving includes all
information related to the role as attachment ﬁgure, including both information about the
self and information about the child-to-be-cared-for. Interestingly, George and Solomon
are the only researchers to discuss the idea of an integrated caregiving representation in
such detail, and their published papers have only described this process in parents of
early school-age children. The results from the present study suggest that an integrated

caregiving representational system may begin much earlier during pregnancy.

107

As hypothesized, results also revealed that prenatal representations of the infant
were signiﬁcantly related to infant attachment security 1 year postpartum. More secure
or balanced maternal representations during pregnancy, characterized by richness,
coherence and sensitivity for example, were related to greater infant attachment security,
while less secure representations were related to less infant security. These results are
consistent with several other studies, which have demonstrated a signiﬁcant relationship
between representations of the child from the parental perspective and child attachment
security (Bretherton et al., 1989; George & Solomon, 1989, 1996). More importantly,
however, these results provide powerful evidence that representations of the infant before
the infant is even born may signiﬁcantly impact infant attachment secm’ity one year later.
Only one previous study has examined this relationship prospectively (Benoit et al.,
1997) and this was in a low-risk, middle-class sample, while the present study included
more high-risk participants. These ﬁndings have important clinical implications for
working with at-risk pregnant women, which will be discussed in more detail below.

Because representations of the self during pregnancy and postnatal self-efﬁcacy
were examined for the caregiving role speciﬁcally, it was hypothesized that these would
be signiﬁcantly related. Results conﬁrmed this hypothesis; more secure or balanced
representations of the self were related to greater reported feelings of postnatal self-
efﬁcacy. However, unexpectedly, postnatal self-emcacy was not signiﬁcantly related to
infant attachment security. There are several possible explanations for this ﬁnding. One
explanation is that self-reported feelings of efﬁcacy are not actually related to infant
attachment, as measured through the well-known Strange Situation procedure. Prior

studies that have demonstrated a clear relationship between self-efﬁcacy and

108

“attachment” have measured both constructs through maternal self-report (Mercer &
Ferketick, 1990, 1994; Williams etal., 1987), and thus, may have only actually
demonstrated a relationship between self-reported efﬁcacy and self-reported beliefs about
the mother-child relationship. That is, the ﬁndings from these studies may have been
spurious due to the complete reliance on maternal self-report and/or may have been
misinterpreted.

Another possibility is that there is in fact a signiﬁcant relationship between
mothers’ feelings of self-efﬁcacy in the caregiving role and infant attachment security,
but that mothers were unable to accurately self-report about themselves in this domain.
Studies that have examined the impact of mothers’ representations of themselves as
mothers, as coded by trained researchers, on mother-infant attachment (as measured
through laboratory procedures like the Strange Situation) have found signiﬁcant
relationships (Ammaniti, 1991; George & Solomon, 1996). One study that examined the
relationship between self-reported eﬂicacy and mother-infant attachment in the Strange
Situation reported that self-efﬁcacy ratings did not differ among mothers of secure and
avoidant children (Spieker & Booth, 1988), and the authors reasoned that mothers of
avoidant children were defensively reporting high levels of self-eﬁcacy, which is
consistent with defensive processes observed in mothers of avoidant infants. Support for
this explanation ﬁom the current study comes ﬁom the observation that maternal self-
eﬁcacy scores were highly skewed towards the positive end, which would not be
theoretically expected ﬁom such a high-risk sample unless there was defensive reporting.

The studies conducted by Donovan and colleagues (1989, 1990) also provide

support for this explanation. They found that some mothers displayed high illusions of

109

control, or unrealistically positive views of themselves as caregivers, which appeared to
mask feelings of inadequacy. Furthermore, mothers with high illusions of control were
more likely to have insecurely attached children in the Strange Situation. Again, these
results suggest that mothers may not be able to accurately self-report on caregiving
efﬁcacy. Although results from the present study did show a signiﬁcant relationship
between prenatal representations of the self (coded by researchers) and postnatal self-
eﬁcacy, self-efﬁcacy scores may not have been ‘accurate’ enough to predict infant
attachment security. It is possﬂale that had researchers coded representations of the self
as a mother ﬁ'om a clinical interview postnatally (like the WMCI), representations would
have been signiﬁcantly related to infant security in the expected direction This may be a
more valid way of measuring maternal self-eﬁcacy than self-report questionnaires.

One ﬁnal possibility mentioned here is that there were problem in the
measurement model that may have prevented a true signiﬁcant relationship from
emerging between these constructs. Both constructs, postnatal self-efﬁcacy and infant
attachment, only had one indicator each. This creates possible identiﬁcation problems
and there is less variance to be explained for each construct, making it harder to detect
relationships. As already mentioned, the observed variable for maternal self-efﬁcacy also
did not have adequate variance.

In sum, modeling results indicated that mothers’ representations of their own
attachment, of their infants, and of themselves as mothers were related to each other and
to infant attachment security in theoretically-important ways. Perhaps the most important
pathways in this part of the model were between maternal representations of attachment,

representations of the infant, and infant attachment security at 1 year. These results

110

demonstrated that attachment representations may be transmitted from mother to child, as
prior research has suggested. That is, more secure representations of attachment ﬁom
childhood predict more secure representations of the infant during pregnancy, which in
turn predict more secure infant attachment after birth. It is also likely that representations
of the self as a mother are involved in this pathway, as part of a broader caregiving
representation. These ﬁndings extend the results from previous studies by showing some
continuity of attachment over time, beginning in pregnancy, in a more diverse and high-
risk sample.
Social Summrt and Mother-Infant Attachment

According to Stern (1995), social support ﬁom females is particularly helpﬁil to
women as they transition to motherhood, e.g., by providing advice and emotional
support. Therefore, social support was expected to indirectly impact mother-infant
attachment through its effects on self-eﬁcacy. As hypothesized, prenatal support from
other women was signiﬁcantly related to postnatal support from women. That is, the
quality of participants’ social support remained relatively stable over time. In addition,
the perceived quality of support was signiﬁcantly related to self-efﬁcacy, with better
support related to greater feelings of self-reported efﬁcacy. These ﬁndings are consistent
with previous studies showing a signiﬁcant relationship between social support and
maternal self-efﬁcacy (Cutrona & Troutman, 1986; Shea & Tronick, 1988; Teti et al.,
1996). It also supports Crittenden’s (1985) conclusion that quality of support may be
more important than quantity of support, since the measure in this study was average
satisfaction rather than simply the number of supporters in participants’ networks. These

ﬁndings suggest that social support can enhance feelings of efﬁcacy by, for example,

111

increasing parenting skills and knowledge, providing conﬁrmation or assurances of
parenting capabilities, or providing opportunities to model others.

Support from women speciﬁcally, many of whom have likely been mothers, may
be especially helpful because of the increased credibility granted to them because of their
own ermeriences with caregiving. Thus, advice or support ﬁ'om women may be given
more ‘weight’ than support from men, at least in the context of pregnancy and
parenthood. Indeed, this was the ﬁrst study to show empirically that support from
women had a signiﬁcant impact on maternal feelings of self-eﬁcacy, as Stern (1995) has
suggested theoretically.

As mentioned previously, however, maternal self-efﬁcacy was unrelated to infant
attachment security, and thus, it appears that social support did not have an indirect
impact on infant attachment, as was hypothesized in the current study. There are several
possible explanations for this ﬁnding. It may be that social support does in fact have an
indirect effect on infant attachment through self-efﬁcacy, as has been shown by a few
other researchers (Donovan & Leavitt, 1989; Isabella, 1994), but this indirect relationship
did not emerge because of the problems described above with the self-eﬁcacy construct
and the pathway between self-efﬁcacy and infant attachment in the current study. It may
also be that social support has an indirect impact on mother-infant attachment, but
through its effects on some other variable such as materml behaviors or parenting, e.g.,
sensitivity, which were not examined in this study.

It also seems possible that social support might have a direct impact on infant
attachment security, as others have found (Crnic et al., 1986; Crockenberg, 1981).

However, when all other variables were taken into account, the data in the current study

112

did not support this notion (although emotioml and caregiving support were weakly
related correlationally to infant attachment security, see Appendix R). Rather, the results
from this study suggest that if social support has an impact on mother-infant attachment,
it is probably more indirect than direct. This conclusion is supported by a few other
studies that have reported no direct relationship between social support and attachment
(Belsky & Isabella, 1988; Zeanah et al., 1993).

In sum, the results ﬁom this study generally do not clarify the already-mixed
ﬁndings in the literature about the relationship between social support and mother-infant
attachment, although the results do support the idea that social support positively aﬂ‘ects
women’s self-eﬁcacy in the maternal role. It will be important for future studies to
continue to examine the role that support may or may not have in contributing to child

attachment security, including different dimensions of support such as type and source of

support.

The Impact of Risk Factors

It was hypothesized that maternal risk factors (e.g., domestic violence, depression,
poverty) would directly predict mother-infant attachment. Although risk factors were
related to infant attachment security correlationally (see Appendix R), with more risk
related to greater insecurity, risk was more strongly related to prenatal representations of
the infant when all variables were considered simultaneously. Therefore, SEM results
revealed a signiﬁcant pathway ﬁ'om risk status to prenatal representations of the infant
rather than ﬁ'om risk status to mother-infant attachment at 1 year. Women who
experienced more risk factors such as domestic violence, depression, and demographic

risks had more insecure representations of their infants during pregnancy, characterized

113

by less sensitivity, acceptance, and less open, rich, and coherent narratives.
Representations were in turn related to mother-infant attachment.

This ﬁnding (risk related to representations) was not originally hypothesized,
however, the results make theoretical sense. Although prior studies have shown a direct
relationship between risk factors and mother-infant attachment in the Strange Situation
(Martins & Gaﬂ‘an, 2000; Spieker & Booth, 1988; Zeanah et al. 1999), similar to the
correlational results in this study, it has generally been assumed that there is some
intervening step in this process. That is, maternal exposure to risk itselfdoes not
‘automatically’ create less securely attached children. Rather, risk such as maternal
depression or poverty may, for example, lead to less sensitive or more rejecting behaviors
(Gelfand & Teti, 1990; Halpern, 1993; Norton, 1990), which are then related to the
development of less secure attachment. The ﬁndings ﬁom this study support the notion
that risk is more indirectly related to mother-infant attachment through its effects on the
mother’s functioning and way of thinking about her child (Thompson, 2000). This
indirect relationship between risk and child ﬁmctioning via maternal functioning has
already been shown in other areas of the literature examining diﬂ‘erent child outcomes,
e.g., anxiety, peer relationships, or school ﬁtnctioning (e.g., Holden & Ritchie, 1991;
Levendosky & Graham-Bermann, 2001). Thus, it appears this indirect relationship may
be true for risk and infant attachment as well.

These results are also consistent with one recent study showing a relationship
betweenmaternalriskandmaternalrepresentationsofthe childandofthe selfasa
mother (Pajulo, Savonlahti, Sourander, Piha, & Helenius, 2001). In this study, results

revealed that maternal risks such as substance use, depression, and demographic risks,

114

were signiﬁcantly related to the content of maternal representations of self and child
during pregnancy. In this study, participants were categorized into a high- or low-risk
group based on the presence or absence of a risk factor and comparisons were made on
responses to a self-report questionnaire. Findings indicated that women in the high-risk
group rated their infant and themselves as mothers much more negatively than the low-
risk group. Unfortunately, this study did not examine the relationship between prenatal
representations and infant attachment after birth and was limited by the sole use of self-
report data. In contrast, the results of the current study extend and improve upon those
ﬁndings by Pajulo et aL by showing that maternal risks affect not only the content of
prenatal representations, but also the way in which mothers talk about their infants, the
latter of which may be particularly important for determining the overall quality of
attachment.
E lorato ses

Exploratory analyses were conducted to examine the concordance between
mothers’ and infants’ attachment classiﬁcations over time using the WMCI and SS
classiﬁcations. This was believed to be another way of examining the continuity or
discontinuity of attachment over time between mother and infant. When disorganized
infants were omitted from analyses and when three groups were used, results revealed a
concordance or match rate of 49% between WMCI and SS classiﬁcations, which was
non-signiﬁcant. When two groups were used (Insecure and Secure), the concordance rate
increased to 60%. This two by two analysis was signiﬁcant; however, the kappa was still
very low (.19). Overall, these results indicate a signiﬁcant amount of discontinuity

between mothers’ attachment representations of the infant and mother-infant attachment

115

at 1 year, especially discontinuity of insecure attachment classiﬁcations. Thus, although
modeling results revealed a strong, signiﬁcant relationship between these variables when
operationalized as dimensional constructs, categorical analyses revealed a substantial
amount of discontinuity as well. Although the lower rates of concordance may be partly
due to measurement error (imperfect inter-rater reliability), it is more likely that there is a
theoretical explanation for these ﬁndings.

In fact, theoretical and empirical discussions of attachment discontinuity have
been rapidly increasing within the last several years, and these seem to be particularly
relevant to the current study. Investigators have examined the stability or instability of
attachment in three different ways: within the same groups of infants over time, within
the same groups of individuals from infancy to adulthood, and within caregiver-child
dyads. In all three areas, researchers have generally found several consistent results: 1)
there is less continuity among the insecure forms of attachment, 2) there is less continuity
in ' -risk samples compared to low-risk samples, and 3) changes in family
circumstances, and especially those that have broad effects on multiple family members,
are most likely to alter attachment security or insecurity (see Thompson, 2000 for a
review). Each of these areas in the literature will be discussed in more detail next.

Egeland and Farber (1984) were among the ﬁrst to examine attachment stability
in young children over time. They found that 60% of their high-risk infant sample had
stable classiﬁcations from 12 to 18 months of age, with more stability in the secure group
(74%) than in the insecure groups (45% for Avoidant and 37% for Ambivalent).
Correlates of infant change away from security (moving ﬁ'om secure to insecure groups)

were less maternal education, more maternal aggression and suspiciousness, more life

116

stress, and single parenthood. Correlates of change towards security (moving ﬁom
inmoure groups to secure group) were greater maternal effectiveness and responsiveness,
better understanding of the infant, and more advanced infant development. Similarly,
Frodi, Grolnick, and Bridges (1985) found an overall stability rate of 66% in infant
attachment classiﬁcation from 12 to 20 months, but this was predominately due to
stability of security (77%) rather than insecurity (33% for Avoidant and 20% for
Ambivalent groups).

More recently, Vondraet al. (1999) reported a 50% stability rate from 12 to 18
months in their low-income sample with 68% stable Secure, 25% stable Avoidant, 33%
stable Ambivalent, and 46% stable Disorganized. Correlates of stable security or move
toward security were relatively low risk experiences, low scores on aggressive/suspicious
personality, low stressful events, and high relationship satisfaction. Correlates of stable
insecurity or move toward insecurity were maternal anxiety and depression, infant
difﬁcult temperament, low relationship satisfaction, stressful life events, and self-reported
anger. In a follow-up study, Vondra et al. (2001) reported an overall increase in
insecurity at 24 months compared to 12 and 18 month classiﬁcations. Although the
overall percentage of stability was similar at 24 months as it was at 18 months (45% and
50%, respectively), there was less stability within the Avoidant and Ambivalent insecure
groups ﬁom 12/18 months to 24 months: 89% stable Secure, 61% stable Avoidant, 32%
stable Ambivalent, and 100% stable Disorganized (n = 2 in the latter group). Thus,
overall, these and other studies examining infant attachment stability (e.g., Barnett et al.,
1999) have generally found less stability of insecure groups and less stability in high-risk

samples.

117

Several recent studies have also examined the stability of attachment over many
years ﬁ'om early childhood to adulthood. For example, Waters et al. (2000) followed a
middle-class sample ﬁom infancy to early adulthood and found that 64% of the sample
had the same classiﬁcation at both time points. Individuals who experienced negative life
events such as parental loss, abuse, or life-threatening illness to a family member were
signiﬁcantly more likely to change attachment classiﬁcations over time, especially to
change from secure to insecure rather than ﬁ'om one form of insecurity to another. The
authors concluded that the relatively high degree of stability in this sample may have
been due to the low-risk nature of the sample. Similarly, Hamilton (2000) reported a
63% stability rate ﬁ'om inﬁincy to adolescence. However, this study found that negative
life events were signiﬁcantly associated with the maintenance of insecure attachment
over time. Finally, Weinﬁeld, Sroufe, and Egeland (2000) reported a 39% stability rate
from infancy to early adulthood in their low-income, ' -risk sample. These researchers
found that stressful life events were unrelated to stability or change, but that child
maltreatment was related to stability of insecurity and maternal depression was related to
movement toward insecurity ﬁ'orn security. Overall, these three studies suggest that,
although there may be continuity in attachment, there can also be change in attachment
over time and discontinuity is often related to changes in the home environment.

Finally, there have been a number of studies examining the transmission or
stability of attachment from caregiver to child, many of which have been described in
detail in the current paper’s literature review. Several studies have shown, for example,
more discontinuity among the insecure attachment classiﬁcations when comparing

mothers’ representations of attachment on the AAI and infant attachment in the SS

118

(Benoit & Parker, 1994); this may be particularly true for the preoccupied-ambivalent
groups (Fonagy et al., 1991; Pederson et al., 1998; Zeanah et al., 1993). Other studies
examining the relationship between mothers’ representations of the infant and infant
attachment security, similar to the present study, have also found less stability among
insecure attachment groups (Benoit et al., 1997; Zeanah et al., 1994).

OveralL the results from the present study’s exploratory analyses are consistent
with other ﬁndings, which show less stability in insecure types of attachment and in high-
risk samples. That is, the results ﬁom this study found less stability between WMCI and
SS categories for the two insecure attachment groups compared to the secure group, as
well as lower overall concordance rates (49% using three groups and 60% using two
groups) compared to previous studies with low-risk samples. Findings related to the
discontinuity of attachment classiﬁcations do not necessarily contradict earlier SEM
results. First, when categorical attachment data are used, there is a potential for some
masking of continuity compared to when continuous data are used (Weinﬁeld et al.,
2000). This may be because there are differences in felt security within the broad
classiﬁcations and/or because of the greater likelihood of classiﬁcation error for
individuals that do not ﬁt neatly into a group, e.g., individuals that are on the border
between two groups.

Alternatively, the ﬁndings related to discontinuity actually provide a more
complete picture of attachment theory. Attachment theory predicts relative stability over
time; however, it also posits that internal working models, and development in general,
remain open to revision in the context of new life experiences (Waters et al., 2000). That

is, attachment theory predicts both stability and change. More recent work in this area,

119

described earlier, has shown that change is more likely to occur in individuals who live in
less stable, more chaotic environments and/or who experience more profound life
changes. Results ﬁom the present study are based on a high-risk sample of individuals,
who likely experience less stable living conditions, and this study examined stability
before and after a major life event, the birth of a child. Thus, it is not surprising that there
was signiﬁcant discontinuity in attachment classiﬁcations.
Sm of Resglts

In sum, results based on SEM analyses generally supported Stern’s (1995) theory
about the psychological processes related to the transition to motherhood. One of the
most striking results ﬁ'om model testing was the continuity of attachment between
mothers’ representations of their own attachment to mothers’ representations of the infant
during pregnancy to infant attachment at 1 year, when attachment was operatiomlized as
a dimensional construct. Thus, mothers appeared to think about their infants in ways that
were similar to the way they thought about being cared for during childhood, which in
turn, was related to the quality of attachment they formed with their infants after birth In
addition, the results supported the notion that there may be an integrated caregiving
representational system that women develop during pregnancy, which includes
representations of both the infant and the self as a caregiver. Social support ﬁom females
also predicted maternal self-efﬁcacy, as expected, but self-efﬁcacy was not signiﬁcantly
related to infant attachment. It is believed that this lack of ﬁnding was due to particular
measurement and statistical reasons, rather than to ﬁaulty theory. Thus, it cannot be

concluded that social support and maternal self-efﬁcacy do not impact infant attachment.

120

Interestingly, the data suggested that maternal risk factors impact mother-infant
attachment indirectly through their impact on maternal representations of the child, rather
than directly. Thus, maternal representations appear to be an important mechanism
through which psychosocial risk exerts its effects. Exploratory analyses revealed
substantial discontinuity of attachment when attachment was assessed categorically. This
is not surprising in light of the high-risk nature of the sample and what has recently been
reported in the literature about discontinuity in high-risk and insecure groups. In
addition, model results and exploratory results provided evidence that attachment is both
relatively stable over time and is capable of being revised in light of life events and
circumstances.

There are a number of strengths in the current study that support previous ﬁndings
and expand the current state of the attachment literature. First, this study used a
prospective design, assessing maternal representations and other variables before the
infant was even born. This type of design adds more conﬁdence to the conclusion that
prenatal risk factors and prenatal representations contribute to infant attachment security,
rather than vice versa In addition, this study is one of the few studies to examine the
impact of both intrapsychic and environmental factors in relation to mother-infant
attachment, making this more comprehensive than most attachment studies. Similarly,
this study examined these processes within a high-risk sample, which has been done
much less frequently than studies with low-risk, middle class ﬁimilies. It is important to
show that much of Stern’s (1995) theory, as well as basic ideas ﬁ'om attachment theory,

hold true in this population as well as in higher functioning populations.

121

There are also several methodological strengths. Although several variables were
measured through maternal self-report, several other constructs were assessed using
clinical interviews and observational methods that were coded by trained, reliable coders.
This minimizes the possibility of spurious ﬁndings, especially those ﬁndings that involve
variables that are not based on self-report. More importantly, these methods for assessing
attachment are believed to provide more rich and valid data because they allow for
measurement of unconscious or automatic processes (Crowell et al., 1999), which are
believed to be intimately involved in attachment-related representations. In other words,
one does not have to be concerned that the participant is self-reporting inaccurately about
their attachment-related thoughts and feelings. Finally, the use of structural equation
modeling provides several advantages for data analysis and interpretation. It is generally
a more stringent method of testing relationships between variables because it estimates
parameters after taking into account all other relationships and shared variance. In
addition, measurement error can be estimated and therefore removed so that resulting
relationships between constructs are error ﬁee. It also allows for the testing of a complex
model (including possible modiﬁcations) by providing overall ﬁt indices and evaluating
relationships simultaneously, rather than requiring a series of analyses to test bivariate
relationships.

However, this study was not without limitations, and it is important to consider
these limitations when drawing conclusions about the results. First, the use of self-report
questionnaires to assess some of the constructs in the study poses some risk for inaccurate
andbiasedreporting. Thishasalreadybeendiscussed fortheself-reportusedtomeasure

maternal self-eﬂicacy, which was highly skewed toward positive responses. This

122

particular measure may have been problematic because each item begins with the
statement “How good are you at. ...” followed by some description of caregiving. It
seems that this wording may be especially prone to social desirability. Although the self-
report used to assess materml representations of attachment experiences ﬁom childhood
(PAAQ) predicted other constructs in the expected direction, some attachment
researchers would argue that self-report questionnaires are unable to accurately assess
attachment representations. Although it is true that the PAAQ does not yield some of the
process-oriented constructs related to defensiveness (e.g., coherence in the narrative) and
to overall attachment classiﬁcations, the results suggest that this instrument may be valid
for assessing overall attachment security, possrbly through some of the subscales such as
‘No Memory ’.

Another disadvantage of using self-report questionnaires with the same informant
is the possibility of spurious or inﬂated ﬁndings, for example, between maternal social
support and maternal self-eﬂicacy. Fortunately, most pathways in the model included
constructs that were measured using two diﬁerent methods, e. g., self-report, interview
codes, and observation codes. Another limitation that was more statistical in nature was
the fact that three constructs had only one indicator, which is generally not advised when
conducting structural model testing. Single indicator constructs present possible
identiﬁcation problems and may make it more difﬁcult to detect relationships because
there is less variance to work with for these constructs. Finally, although the high-risk
nature of the current sample was generally a strength of the study (since most previous
studies have only included low-risk, middle-class families), it also limits the

generalizability of ﬁndings. Thus, the results ﬁom this particular study may only be

123

representative of women that are generally more at-risk (e. g., low-income, single parents,
history of domestic violence) than of women who are generally not at-risk.
Directions for Future Research

There are a number of ways that future research can expand our current
understanding of the processes related to the formation of mother-infant attachment,
including the inﬂuence of both intrapsychic and environmental variables. First, it seems
that more research needs to be done in diverse and higher-risk groups of mothers and
infants. Some researchers are studying attachment in high-risk groups such as low-
income and maltreatment samples (e.g., Barnett et al., 1999;Ege1and & Sroufe, 1984;
Spieker & Booth, 1988; Vondra et al., 1999; Weinﬁeld et al., 2000); however, the
majority of attachment research is done with low-risk, middle-class, Caucasian samples.
It also seems important for future studies to replicate the ﬁndings here in order to further
validate (or invalidate) Stern’s (1995) theory about the transition to motherhood and the
role of female social support in facilitating this transition. This is the ﬁrst model to test
his theory and others should continue to test his theory with different populations.

A very important, more speciﬁc direction for future research is the further
examination of processes related to the stability and instability of attachment, both within
the same individual and across generations ﬁom parent to child. Most attachment
research has focused on validating and emphasizing the notion of continuity over time,
but more work needs to be done to investigate cases of discontinuity and factors related
to discontinuity. As others have begun to point out more clearly, Bowlby (1973, 1980)
not only indicated that attachment is somewhat continuous and resistant to change, but

that it is also dynamic and open to revision. Thus, there needs to be further examination

124

of factors such as stressful life events, changes in the family environment, changes in
caregiver behaviors, etc... to determine how signiﬁcant life experiences may alter
attachment representations. In addition, it will be important to determine what
experiences are important for what types of people and at what developmental stages.
Thompson (2000) recently suggested that life events which affect relational patterns and
developmental periods during which there are signiﬁcant representational advances may
be particularly important areas to investigate. Clearly, this work has just begun and will
have pivotal implications for clinical work.

Another important direction for future research is the investigation of intermediate
factors that might further elucidate the link between maternal representations and infant
attachment, e.g., between the AAI or WMCI and the Strange Situation. The results from
this study suggested that there is a strong relationship between materml representations
during pregnancy and infant attachment security aﬂer birth, when measured as
dimensional constructs, but it is still unclear how these representations are “transmitted”
to the infant. Although the majority of prior research in this area has focused on maternal
sensitivity and responsiveness (through overt, observable behaviors) as the mediating
variable, research has shown that sensitivity explains only a relatively small amount of
the relationship between mothers’ representations and infant attachment (DeWolff & van
Ijzendoom, 1997; Pederson et al., 1998; van Ijzendoom, 1995). Thus, other observable
variables, which may reﬂect important aspects of the caregiver as a secure base, need to
be investigated.

Finally, there are several methodological considerations for future attachment

research. First, it has been suggested by several attachment researchers (Cicchetti, 2001;

125

Weinﬁeld et al., 2000) that both dimensional and categorical measures of attachment be
used in future research Each way of operationalizing attachment may be useful and may
be complementary to the other. This idea is supported by the current ﬁndings, which
revealed an interesting picture of attachment processes when both kinds of data were
used. There has been some debate over whether and how to convert the Strange Situation
codes into a continuous variable. One method is the conversion system developed by
Bretherton and colleagues (1989, 1990) used in the present study, which appears to be a
viable option. An additional method that should be considered for ﬁrture research is the
empirically derived system described by Richters, Waters, and Vaughn (1988), which
uses the subscale scores that are coded and used to make overall classiﬁcations. This
method was shown by the authors to be a reliable and valid means of assessing
attachment security along a continuum. Last, but not least, it is important for ﬁrture
research to use longitudinal, prospective designs to examine the complex processes
involved in the development and evolution of attachment over time.
Clinical Implications

Finally, the results ﬁom the present study not only extend our conceptual and
theoretical understanding of the processes related to the development of mother-infant
attachment, they also provide important information that is directly relevant to clinical
work with at-risk pregnant and parenting women. A number of researchers and clinicians
have already developed treatments for work with parents and infants (e. g., Lieberman &
Pawl, 1993; McDonough, 1993; Stern, 1995); however, results ﬁom studies such as this

one can help guide and support the direction of these treatments.

126

First, the results of this study suggested that maternal risk factors such as poverty
and domestic violence have an indirect impact on infant attachment through its effects on
maternal representations of the infant. Maternal social support may also have an indirect
impact on infant attachment, although this was less clear. Still, it seems important for
clinical interventions to consider and attend to environmental factors that could be
impinging on a mother’s ability to psychologically prepare for motherhood and her
ability to think about her infant and herself as a mother in a non—defensive, sensitive, and
accepting manner. Thus, clinicians may need to evaluate the mother’s ﬁimily,
community, and social systems in order to help the mother correct or alter those factors
that may be interfering with the caregiving process, even as this process begins during
pregnancy.

The most striking ﬁnding in this study that is directly relevant to clinical work is
the strong, direct relationship between prenatal representations of the infant and infant
attachment at 1 year. These ﬁndings suggest that early interventions with at-risk
pregnant women can be critical in preventing the transmission of insecure attachment
from mother to child. Furthermore, these results indicate that clinical work must address
mothers’ representational processes including representations of their own attachment
and representations of their infant and themselves as caregivers. Fortunately, many of the
existing parent-infant psychotherapy programs do focus on altering mothers’
representations as a way to improve mother-infant relationships (Lieberman & Pawl,
1993; Stern, 1995); however, these programs are generally geared towards working with

mothers after the birth of the baby. It is important for research studies like this one to

127

show empirical support for these types of programs, as well as the importance of
beginning this work during pregnancy.

Programs do, however, diﬁ‘er in how they approach altering mothers’
representations. Stern (1995) summarizes a number of programs which all use different
“ports of entry” to alter parental representations, but states that all share the common goal
of changing these representations. Working with the assumption that mothers’
representations inﬂuence behaviors towards the infant which in turn inﬂuence infant
behaviors and infant representations (and vice versa), he states that clinicians can focus
on any of these elements as a way of entering into and altering the mothers’
representations. For example, some programs use infant behaviors as a starting point,
while others use maternal behaviors, etc. . .Subsequent to altering representations, it is
believed that mothers will interact with their infants in a more positive, sensitive, and
responsive manner, which eventually leads to more secure infant attachment. Based on
the current study, it seems that intervening through the mothers’ representations would be
the natural starting point when working with pregnant women since parent behaviors
could not yet be observed. Finally, in addition to exploring mothers’ representations
directly, an equally important part of this clinical work includes providing the mother
with a “corrective attachment experience,” whereby the therapist serves as an attachment
ﬁgure who is internalized and integrated into existing representations. Through both
exploration and experience, the mother is able to alter representations of self and other,

enabling her to be a more secure attachment ﬁgure to her own infant.

128

Conclusions

In sum, this study expands our current understanding of the processes related to
the formation of mother-infant attachment, beginning in pregnancy and extending into the
ﬁrst year of life. This study provides empirical support for the signiﬁcance of both
individual, intrapsychic factors and environmental factors in predicting infant attachment,
which demonstrates the importance of studying mother-infant attachment from an
ecological perspective, particularly in diverse, high-risk samples. The results ﬁ'om this
study suggest that maternal representations during pregnancy and maternal risk factors
may be especially important predictors of infant attachment security, while the role of
maternal social support is less clear. Interestingly, the results ﬁ'om this study also
demonstrate that there is both continuity and discontinuity of attachment over time,
which provides empirical support for both aspects of attachment theory.

In addition to the theoretical implications, the results ﬁ'om this study have clear
clinical implications; results suggest that it is critical to preventively intervene with
women at risk for attachment disturbances during pregnancy, by targeting both
intrapsychic (e.g., representations) and environmental factors. Finally, future research in
this area can expand our understanding of the complex processes involved in the
evolution of attachment over time by: 1) operationalizing attachment both dimensionally
and categorically, 2) examining the stability and instability of attachment within and
across individuals, and 3) examining possible nechanisms at multiple system levels, e. g.,
individual and cormnunity, that account for this (in)stability. In so doing, ﬁrture research
will also help deepen and broaden our understanding of more general adaptive and

maladaptive developmental processes throughout the life span.

129

APPENDICES

130

APPENDIX A
Flyers for Recruitment

ARE YOU PREGNANT?

YOU MAY BE ELIGIBLE TO PARTICIPATE IN A STUDY
ABOUT MOTHER-INF ANT RELATIONSHIPS

:2 $50.00 1!,

We are looking for pregnant women to participate in a research
study at Michigan State University. You will be asked about
experiences and feelings during pregnancy, perceptions of

your infants, and recent life events.

- Interview can be done at MSU or at your home.
- You will be paid $50.00 in cash.
a All information is kept completely conﬁdential.

!! $50.00 1!

********************************************

If you are interested or would like more information,
please call 432-1447 and ask for
Dr. Anne Bogat’s Mother-Infant Study

 

 

MOTHER-
INFANT
STUDY

432-1447

 

 

MOTHER-
INFANT
STUDY

432-1447

MOTHER-
INFANT
STUDY

432-1447

 

MOTHER-
INFANT
STUDY

432-1447

 

432-1447

 

 

 

131

 

HAVE YOU BEEN HURT BY SOMEONE YOU LOVE?

ARE YOU PREGNANT AND HAVE YOU BEEN
PUSHED OR GRABBED OR HIT OR SLAPPED OR
KICKED (OR WORSE) BY A PARTNER OR
BOYFRIEND DURING YOUR PREGNANCY?

We need women to take part in an interview about their lives
and their pregnancies.

- Interview can be done at MSU or at your home.

- You will be paid $50.00 in cash.

- All information is kept completely conﬁdential,
!! $50.00 1!

********************************************

Ifyou are interested or would like more information,
please call 432-1447 and ask for

Dr. Anne Bogat’s Mother-Infant Study

 

 

 

MOTHER- MOTHER- MOTHER- MOTHER-
INFANT INFANT INFANT INFANT
STUDY STUDY STUDY STUDY

432-1447 432-1447 432-1447 4 2-1447

 

 

 

 

 

 

 

132

APPENDIX B
Time 1 Consent Form

MOTHER-INFANT STUDY
CONSENT FORM

This study is part of a survey of women in Michigan, some of whom may be
experiencing domestic violence. We hope to learn about the strengths that you bring to your
situation, your feelings and perceptions of your baby during pregnancy, and your relationships
with others, including family members, partners, and ﬁ'iends. We hope to use this information to
help plan better programs for families experiencing domestic violence.

If you decide to take part in the survey, you will be asked questions about how you have
been feeling recently, events that have happened to you, your feelings about pregnancy and your
baby, the people in your life who provide support to you, and your memories of your childhood.
This will take a total of 2-3 hours.

All information will be kept strictly conﬁdential. Your name will be removed from all
questionnaires and an identiﬁcation number will be put on them instead. All questionnaires will
be kept in locked ﬁle cabinets in a locked ofﬁce. Your identity will not be revealed in any reports
written about this study. We will summarize information from all study participants and will not
report information about individuals. The only exception is in the case of ongoing child abuse. If
you indicate that child abuse is occurring in your household, we are required to make a report to
Protective Services.

You have the right to withdraw ﬁ'om this study at any point during the interview with no
penalty or negative consequences. Your decision about whether to participate or not will not
affect your relationship with any agencies or Michigan State University. If you have any
questions, please ask us. If you have questions later, you can contact Dr. Anne Bogat or Dr.
Alytia Levendosky at (517) 432-1447.

We are also interested in recontacting you about 2 months after the birth of your baby by
telephone and then we would like to meet with you and your baby at 12 months after the birth of
your baby. So at the end of the interview today, we will ask you for information that will help us
keep in contact with you. Your participation at gig time does not obligate you to participate in
the second telephone appointment, or the third interview. You will be paid $50 for the ﬁrst
interview, mailed a baby gift after the telephone interview, and $75 and a baby gift for the third
interview, if you wish to participate.

A. L A_-__A__Aa LAAAALAAAAAAAAAAAL¢AALAAAA. A A _A__A_A__A__.__‘_.A__A_ A. LL; L... A LA A. A. L... LAAAAAAAAAJ ALAAAJ AAA

TTTTTTTTTT rTTTT-r‘r‘r-r r‘V'TTTT'rfv—f-TVTTV 'v T'I'TT‘V'T1 TVTTTTT‘V‘TTT‘TTTTw—w ‘v I 77* 1 1 TVTTTVVT-rf’r‘r

I have read and understood the above statanents. I understand that my participation in this study
is completely voluntary and that I can withdraw from this study at any time without penalty or
negative consequences.

 

 

 

 

Signature of Participant: Date:

Witness: Date:

Anne Bogat, Ph.D. Alytia Levendosky, Ph.D.
Michigan State University Michigan State University
Department of Psychology Department of Psychology
123 Snyder Hall 121 Snyder Hall

East Lansing,MI48824-lll7 EastLansing,MI48824-lll7

133

APPENDIX C
Time 2 Verbal Assent Form

Before I begin asking you questions, I need to get your verbal consent to participate in
this interview. I will read a few paragraphs and ask you whether you understand and
agree with what I have said.

If you decide to take part in the telephone interview today, you will be asked questions
about your health during and after your pregnancy, your labor and delivery, and your
baby’s health and personality. You will be sent a gift for your baby after the interview as
a token of our thanks for participating.

All information that you provide will be kept strictly conﬁdential. Your name will not be
put on any questionnaires; they will only have an identiﬁcation number on them. All
questionnaires will be kept in locked ﬁle cabinets in a locked oﬂice. Your identity will
not be revealed in any reports written about this study. We will summarize information
ﬁom all study participants and will not report inforrmtion about individuals. The only
exception is if you indicate that child abuse is occurring in your household; then we are
required to notify Protective Services.

You have the right to withdraw ﬁ'om this study at any point timing the interview with no
penalty or negative consequences. Your decision about whether to participate or not will
not affect your relationship with any agencies or with Michigan State University. If you
have any questions, please ask us. If you have questions later on, you can contact Dr.
Anne Bogat or Dr. Alytia Levendosky at (517) 432-1447.

In addition, we would also like to contact you when your baby is 11-12 months old.
Your participation today does not obligate you to participate in the third interview.

Do I have your permission to continue with the interview? (Circle one) YES NO

134

APPENDIX D
Tracking Letter and Information Sheet

Date:

Dear

 

Thank you again for participating in the Mother-Infant Study at Michigan State
University. We really appreciate your willingness to help us learn about the unique and
important experiences of mothers and infants in our community. \Vrth the knowledge
that we gain from this study, we will be able to develop new and better programs for
families like yours. Therefore, your participation in our study will directly beneﬁt many
other women and children!

We are writing this letter to update our records so that we will be able to reach
you about 1 year alter the birth of your baby, at which time we would like to meet with
you and your child. You will be paid $75.00 and given a gift for your child at that
interview. Please take a moment to ﬁll out the Subject Information Sheet and send it back
to us in the enclosed self-addressed, stamped envelope. If you prefer, you may also call
us at 432-1447 or 432-7726 to give us your updated information. Please send back the
information sheet or call us even if you still have the same address and phone number. As
usual, all information will be kept strictly conﬁdential. Thank you again for your help on
this important project.

Sincerely,

Mother-Infant Study Staﬂ‘ Member

135

Subject Information Sheet

Name:

Address:
(Street Number and Street Name)

 

(City, State and Zip Code)
Telephone Number:

Alternate Telephone Number:
Cell Phone Number:

ﬁrgeror Beeper Number:
E-mail:

Other Contact Information:

Thank You! ! !!

 

For Oﬁce Use Only

Date Received:

90 day follow-up: 1St or 2"‘1
Other Notes:

136

APPENDIX E
Time 3 Consent Form

MOTHER-INFANT STUDY
CONSENT F ORM-T3

Thank you for participating in the ﬁrst two interviews for this study. This study is part of
a survey of women in Michigan, some of whom may be experiencing domestic violence. We
hope to learn about the strengths that you bring to your situation, your feelings, your perceptions
of your baby, and your relationships with others, including family members, partners, and friends.
We hOpe to use this information to help plan better prograrrrs for families experiencing domestic
violence.

If you decide to take part in the survey today, you will be asked questions about how you
have been feeling recently, events that have lmppened to you in the last year, and your feelings
about your baby and the people in your life who provide support for you. You will also be asked
to participate in some play with your baby that will be videotaped. Your baby will also play with
some toys and games with an interviewer for about an hour. The total interview will take about
3—4 hours. You will be paid $75 for your participation and you will also receive a toy for your
baby.

All information that you give us will be kept strictly conﬁdential among the project staff.
Your name or your baby’s will not be on any questionnaires and videotapes; an identiﬁcation
number will be put on them instead. All questionnaires and videotapes will be kept in locked ﬁle
cabinets in a locked ofﬁce. Your identity will not be revealed in any reports written about this
study. We will summarize information ﬁ'om all study participants and will not report information
about yourself or any individuals. Your privacy will be protected to the maximum extent
allowable by law. The only exception to full conﬁdentiality is in the case of ongoing child abuse
or neglect. If you indicate that child abuse or neglect is occurring in your household, we are
required to make a report to Child Protective Services. We would inform you if we thought we
needed to make such a report.

You have the right to refuse to answer any questions or to withdraw from this study at
any point during the interview with no penalty or negative consequences. Your decision about
whether to participate or not will not affect your relationship with any agencies or Michigan State
University. If you have any questions, please ask us. If you have any questions about the study
later, you can contact Dr. Anne Bogat or Dr. Alytia Levendosky at (517)432-1447. If you have
questions about your rights as a participant in this research study you may contact Dr. David
Wright at 355-2180.

We may be interested in recontacting you 2 years after the birth of your baby. At the end
of the interview today, we will ask you to update the contact information that we have for you.
Your participation today does not obligate you to participate in any future interviews.

 

l have read this form and agree to participate.

 

 

 

 

Signature of Participant: Date:

Witness: Date:

Anne Bogat, Ph.D. Alytia Levendosky, Ph.D.

Michigan State University Michigan State University
Department of Psychology Department of Psychology
East Lansing, MI 48824 East Lansing, MI 48824

137

APPENDIX F
Perceptions of Adult Attachment Questionnaire

PAAQ

The majority of the following statements refer to your early childhood relationship with
your mother (when you were approximately 3 to 8 years old). If someone else was the
principal person responsible for your care in childhood, please respond to the questions
which refer to “mother” with that person in mind.

A few of the questions have two parts. For example “when I caused trouble as a child I
knew my mother would forgive me.” Some people might feel like they never caused
trouble as a child; however, they consider their mothers very forgiving. How then do
you answer? Only answer AGREE or STRONGLY AGREE if you agree with both
parts of the statement. If you agree with only one part of the statement answer
NEUTRAL. If you disagree with both parts of the statement answer DISAGREE or
STRONGLY DISAGREE.

l= STRONGLY DISAGREE

2 = DISAGREE

3 = NEUTRAL (NEITHER DISAGREE NOR AGREE)
4 = AGREE

5 = STRONGLY AGREE

1. In childhood I felt like I was really treasured by my mother.

2. In childhood I sometimes felt like my mother was really lonely when I
was not with her.

3. My mother was not very affectionate.

4. .When I was a young child and little things went wrong, 1 did not feel sure
I could count on my mother to take care of me.

 

5. As a child I couldn’t stand being separated ﬁom my mother.

6. My mother can make me feel really good, but when she is not nice to me
she can really tear me apart.

7. In my family of origin we don’t make a show of expressing our feelings.
We prefer keeping feelings to ourselves.

8. Neither my mother nor myselfare perfect, but somehow we made it
through my childhood.

138

1= STRONGLY DISAGREE

 

2 = DISAGREE

3 = NEUTRAL (NEITHER DISAGREE NOR AGREE)
4 = AGREE

5 = STRONGLY AGREE

9. I remember, when I was ﬁ'ightened as a child, my mother holding me
close.

10. When I was a child, my mother sometimes told me that if I was not good
she would stop loving me.

11. My mother is selﬁshly caught up in herself to the exclusion of everybody
else.

12. My family was not particularly intimate, but this has never bothered me.

13. It’s hard for me to remember my early relationship with my mother
in any detail.

14. In childhood I sometimes felt that my mother and I were so alike that I
didn’t know where she ended and I began.

15. If anything happened to my mother I wonder if I could survive it.

16. I remember as a child feeling a desire to protect my mother.

17. Even though I went through rough times with my mother during my
childhood, somewhere along the line I managed to let go of the majority of
those angry, hurt feelings.

18. In childhood I knew I was low on my mother’s priority list.

19. My mother was an all-around excellent mother.

20. No one gets under my skin like my mother.

21. As a child I never thought separations from my parents were any big deal.

22. I often felt responsible for my mother’s welfare.

23. In childhood my mother sometimes threatened to leave me or to send me
away if I wasn’t good.

24. To this day my mother has no clue who I am or what 1 am all about.

139

l= STRONGLY DISAGREE

2 = DISAGREE

3 = NEUTRAL (NEITHER DISAGREE NOR AGREE)
4 = AGREE

5 = STRONGLY AGREE

 

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

Even with all our past difﬁculties, I realize my mother did the best for me
that she could.

I have forgotten what most of my early childhood was like.

I always knew my mother was there for me; no matter what I could
depend on her.

There are times when I feel like shaking my mother and saying “wake up
and see me for who I am.”

In childhood I often had the impression that my mother was not listening
to me. She oﬁen tuned me out.

During my childhood I sometimes felt like I was my mother’s whole life.

My mother and I are more accepting of each other’s differences than we
have been in the past.

When 1 was young, I oﬁen feared something dreadful would happen to my
mother or father.

I remember my mother telling me that I didn’t pay enough attention to her
or love her enough.

I often take my mother’s opinions about rrre to heart and lose sight of my
own opinions about myself.

My mother is a real nag.

My mother and I were so alike we often could ﬁnish each other’s
sentences.

I think people put too much emphasis on the mother/child relationship.
I remember very little about my early childhood (ages three to seven).

The concept of the loving, supporting mother is pure myth

140

l= STRONGLY DISAGREE

2 = DISAGREE

3 = NEUTRAL (NEITHER DISAGREE NOR AGREE)
4 = AGREE

5 = STRONGLY AGREE

 

 

 

40.

41.

42.

43.

45.

 

46.

 

47.

48.

 

49.

 

50.

51.

52.

53.

My relationship with my mother has gone through major changes over the
course of my childhood and adolescence.

Even as an adult I sometimes feel like I will never dig myself out from
under my mother’s inﬂuence.

As a child I sometimes got the feeling that without me my mother would
have fallen apart.

I couldn’t have asked for a better mother.

Ifmymotherwasnotfairtomeasachildlrealizenowitwasbecause she
was dealing with her own problems.

If something really had happened to me in childhood, I did not feel I could
count on my mother to support me.

When I was a child I sometimes got the feeling that my mother wished I
was never born.

I remember when I was a child feeling scared that one or both of my
parents would die unexpectedly.

My mother can devastate me with her criticisms.
In childhood my mother often told me she was sacriﬁcing herself for me.

1 don’t think my early childhood relationship with my mother has any
signiﬁcant inﬂuence on who I am today or my present relationships.

My mother was always there for me when I needed her.

When I acted bad as a child my mother would, at times, threaten
to send are away.

I never felt like my mother gave me enough attention.

141

l= STRONGLY DISAGREE
2 = DISAGREE
3 = NEUTRAL (NEITHER DISAGREE NOR AGREE)
4 = AGREE
5 = STRONGLY AGREE
54. For all our past problems, my mother and I can still enjoy a good laugh
together.

55. During my childhood, my mother would often turn to me and tell me lots
of things that upset and bothered me.

56. In childhood, I often worried about my mother’s state of health

57. I ﬁnd it difﬁcult to remember my early childhood.

58. My mother was a perfect mother.

59. My mother’s issues are still interfering with my life.

m

60. When I think back to my early childhood experiences, I discover things
about myself and my parents that I’ve never considered before.

142

APPENDIX G
Working Model of the Child Interview

WMCI

INTERVIEWER: PLEASE TURN ON THE TAPE RECORDER AND MCROPHONE. Test the
tape recorder by saying “Testing 1, 2 , 3 ” a few times. Rewind and play back the recording to
make sure everything is working properly. THEN, TURN ON TAPE REC ORDER-PRESS
RECORD—AGAIN. WAIT 5 SECONDS. SAY YOUR NM, DATE, and SUBJECT NUWER
INTO THE RECORDER BEGIN MERVIEW.

1. MAKE SURE TAPE RECORDER IS TURNED ON and THAT “RECORD” IS
PRESSED.
2. MAKE SURE MICROPHONE IS TURNED ON.

 

We are interested in how parents think and feel about their children before they
are born. This interview is a way for us to ask you about that. The interview will take us
about an hour to complete.

1a. Let’s start with your pregnancy. I’m interested in things like whether it was planned
or unplanned, how you feel physically and emotionally, and what you are doing during
the pregnancy (e.g., working). Let’s take these one at a time. [The idea is to put the
participant at ease and to begin to obtain a chronological history of the pregnancy.
Additional probes may be necessary to make sure that the individual is given a
reasonable opportunity to convey the history of their reactions and feelings about the
pregnancy and the baby (which may or may not be the same). ]

Was the pregnancy planned or unplanned?

How much is the baby wanted or not wanted?

When did the pregnancy seem real to you?

How have you felt physically and emotionally throughout your pregnancy? [ Interviewer:
Find out the history of these throughout the pregnancy. ]

What are you doing, or have you been doing, during the pregmncy? (e.g., working?)
[Interviewer: Find out the history of these throughout the pregnancy. ]

What have been your impressions about the baby while you’re pregnant? What do you
sense the baby might be like?

143

u" MAKE SURE THE TAPERECORDER IS TURNED ON and “RECORD” IS
PRESSED.
*m MAKE SURE THE MICROPHONE IS TURNED ON.

1b. How do you think you will react to labor and delivery? What do you think your
feelings about labor and delivery will be?
What do you think your ﬁrst reaction will be when you see the baby?

What will be your reaction ifthe baby is a boy? Ifthe baby is a girl?

How do you think your family will react to the birth of your baby? [Interviewer: Be sure
to include husband/partner, other siblings. ]

1c. Do you think your baby will have any problems in the ﬁrst few days after birth?

How long do you think the baby will have to stay in the hospital?

Are you going to breast-feed or bottle-feed? Why? How did you come to that decision?

1d. How do you think the ﬁrst few weeks at home with the baby will go? [Interviewers
Explore feelings about feeding, sleeping, crying, etc. ]

1e. How old do you think your baby will be when he/she sits up?

Crawls?

Walks?

Smiles?

Talks?

Do you think your baby will do these things ahead, behind, or at the same time as other
babies?

144

Do you have any sense yet of what your baby’s intelligence will be? Why do you think
that?

If. Do you think your baby will have a regular routine? What do you think will happen
if you or your baby can’t stay in the routine?

1 g. Will you need to be separated from your baby after he/she is born? (e.g., work)
If the participant says YES or NO, the interviewer asks:
"What do you think this will be like for you? For the baby?
If the participant says “I HOPE NOT” the interviewer asks:
"If this did happen, what would this be like for you? For the baby?

Will there be any separations in the ﬁrst year of your baby’s life that will last for more
than a day? How will that be for you? For your baby?

2a. What do you think your child’s personality will be like when he/she is born?
[Personality-the qualities/traits/features that give someone their identity, that makes
someone who they are ]

2b. Pick 5 words (adjectives) that describe what your child’s personality will be like
when he/ she is born. [ Interviewer: Write these down on the paper for reference. It is not

important that participants come up with exactly 5 adjectives]
1.

2
3.
4.
5
For each one, what makes you say that?
3a. Who do you think your baby will be most like?
What personality traits do you think your child will inherit ﬁ'om you?

What traits will your child inherit from the baby’s father?

145

3b. Do you think there are any characteristics your child will inherit from your side of
the family?

From the baby’s Ether’s side of the family?

3c. Have you decided on your child’s name? How did you decide?(or How will you
decide?)

Does that name have special meaning in your Emily or the baby’s Ether’s Emily?

4. In what ways do you think your child will be unique or different ﬁom other children?

5. After your baby is born, what behavior in his/her ﬁrst year oflife do you think will be
the most diﬁcult for you to handle? Can you give an example?

5a. Why will this be difﬁcult? How often do you think it will occur?

What will you feel like doing when your child behaves like that? How will you feel if
your child acts this way? What will you do about the behavior?

5b. Do you think your child will know you don’t like that behavior? Why do you think
he/she will act like that?

So. What do you imagine will happen to this behavior as your child grows older? Why
do you think so?

6a. How would you describe your relationship with your baby now, while you’re
pregnant?

146

6b. Pick ﬁve words (adjectives) to describe your relationship. For each word, describe
an incident or memory that illustrates what you mean. [ Interviewer: Write these down on
the paper for reference. It is not important that participants come up with exactly 5
adjectives]

P‘PP’N?‘

7a. What pleases you most about your relationship with your baby while you’re
pregnant?

What do you wish you could change about it?

7b. How do you feel your relationship with your baby while you’re pregnant will affect
your baby’s personality?

7c. Has your relationship with your baby changed during the pregnancy? In what ways?
What is your feeling about the change?

8. When your baby is born, what parent do you think he/she will be closest to? Why?

Do you expect that to change (as the child gets older, for instance)? How do you expect
it to change?

9. Do you think your baby will get upset often in his/her ﬁrst 12 months? What will you
do at those times? What do you think your feelings will be at those times?

9a. What about when the baby becomes emotionally upset? What will you do at those
times?

What do you think your feelings will be at those times?

9b. What about when your child becomes physically hurt a little bit (e.g., hitting his head
against the crib)? What will you do at those times? What do you think your feelings will
be at those times?

147

9c. What about when your child becomes sick (e.g., he/she gets a fever)? What will you
do at those times? What do you think your feelings will be at those times?

10. Tell me a favorite story about your pregnancy, perhaps one you’ve told to Emily or
ﬁiends. I’ll give you a minute to think about this one. [ Interviewer: If the participant is

struggling, you may tell them that this doesn ’t have to be the favorite story, only a
favorite one. ]

What do you like about this story?

11a. Can you think of any experiences you’ve had during your pregnancy that might
have been a setback for your baby? Why do you think so? |Setbac =something that
happened that makes things harder for your baby than for other babies.]

If person says ITS, then ask:
”"Why do you think so?

If person says NO, go to next question.
[ Interviewer: Indirectly, we ’re trying to determine whether the parent feels responsible

in any way for the setbacks. ]

Knowing what you know now, if you started all over again with your pregnancy, what
would you do differently?

1 lb. Are there any experiences your baby might have during the ﬁrst year of his/her life
that might be a setback for him/her? If person says NO, go to question 12.

If person says ITS, then ask:

Why do you think so?

Who or what is likely to contribute to these setbacks?

Is there anything you might do to prevent these setbacks?

12. Do you ever worry about your unborn baby? What do you worry about?

13. If your child could be any age right now (unborn, 1 month, 1 year, etc.), what age
would you choose? Why?

148

14. As you look ahead, what will be the most difﬁcult time in your child’s development?
Why do you think so?

15. What do you expect your child to be like as an adolescent? Wlmt makes you feel this
way? What do you expect to be good and not so good about this period in your child’s
life?

16. Think for a moment of your child as an adult. What hopes and fears do you have
about that time?
INTER VIEWER:

I. TURN TAPE RECORDER AND MICROPHONE OFF.
2. GIVE PART ICIPANT THE INT ER VIEW PACKET.

149

APPENDIX H
Maternal Efﬁcacy Questionnaire

MEQ
We want to ask you some questions about yourself and your baby. We are trying to get a
general idea of how you usually handle different situations with your baby. We realize tlnrt no
one is always effective or always ineffective. We all do better in some situations than in others.
So we would like to have you think about some situations that all mothers encounter.

1. When your baby is upset, fussy or crying, how good are you at soothing him or her?

1 2 3 4
not good at all not good enough good enough very good

2. How good are you at understanding what your baby wants or need? For example, do you know
whenyourbabyneedstobechangedorwantstobefed?

l 2 3 4
not good at all not good enough good enough very good

3. How good are you at making your baby understand what you want him/her to do? For
example, if you want your baby to eat dinner or play quietly, how good are you at making him or
her do that?

l 2 3 4
not good at all not good enough good enough very good

4. How good are you at getting your baby to pay attention to you? For example, when you want
your baby to look at you, how good are you at making him or her do it?

1 2 3 4
not good at all not good enough good enough very good

5. How good are you at getting your baby to have fun with you? For example, how good are you
at getting your baby to smile and laugh with you?

1 2 3 4
not good at all not good enough good enough very good

6. How good are you at knowing what activities your baby will enjoy? For example, how good
are you at knowing what games and toys your baby will like to play with?

l 2 3 4
not good at all not good enough good enough very good

150

7. How good are you at keeping your baby occupied when you need to do housework? For
example, how good are you at ﬁnding things for the baby when you need to do the dishes.

1 2 3 4
not good at all not good enough good enough very good

8. How good do you feel you are at feeding, changing, and bathing your baby?

1 2 3 4
not good at all not good enough good enough very good

9. How good are you at getting your baby to show off for visitors? For example, how good are
you at making your baby smile or laugh for people who visit?

1 2 3 4
not good at all not good enough good enough very good

10. In general, how good a mother do you feel you are?

l 2 3 4
not good at all not good enough good enough very good

151

APPENDIX I
The Strange Situation (SS): Laboratory Procedures

Episode 1 (30 seconds): The mother and baby are introduced to the experimental room by
experimenter. Mother is shown where to set the baby down and where to sit. The
examiner leaves the room as soon as she has completed the instructions.

Episode 2 (3 minutes): The mother and baby are alone in the experimental room. The
mother sits in her chair and pretends to read a magazine. The mother has been instructed
ahead of time that she should not initiate conversation with the baby, although she can
respond to baby’s initiation of interaction. If after 2 minutes the baby has not explored the
toys, the mother is signaled (through a knock on the wall by the experimenter) to help the
baby become interested in the toys.

Episode 3 (3 minutes): A stranger (who has never met the baby before) enters the room
and introduces herself to the mother. The stranger immediately sits in her designated
chair and remains silent for 1 minute. At the end of 1 minute, the stranger begins a
conversation with the mother. At the end of another minute, the stranger is signaled (by
experimenter) to initiate interaction with the baby. At the end of 3 minutes total, the
mother unobtrusively leaves the room, leaving her purse or handbag behind on her chair.

Episode 4 (3 minutes): As soon as the mother leaves, the stranger reduces interaction
with the baby so that the baby can notice that the mother is gone. Stranger returns to her
chair. If the baby cries, the stranger will try to distract the baby with toys. If necessary,
the stranger will attempt to calm the baby by talking to or picking up him/her. If baby is
calmed, the stranger puts him/her down and attempts to engage the baby with the toys.

Episode 5 (3 minutes): The mother approaches the closed door and speaks outside, loudly
enough that the baby can hear her. She pauses, opens the door, and pauses again to let the
baby respond. Meanwhile, the stranger leaves unobtrusively. After 3 minutes, the

mother leaves the room She pauses at the door and says “bye-bye,” closing the door
behind her.

Episode 6 (3 minutes): The baby is left alone to explore the room for 3 minutes. If the
baby becomes acutely distressed, this episode is terminated, but the remaining time is
added onto episode seven.

Episode 7 (3 minutes): The stranger approaches the door and speaks outside, loudly
enough that the baby can hear her voice. She pauses, opens the door, and pauses again
allowing the baby to respond. If the baby is crying, the stranger will attempt to soothe the
baby. Whenthebabyiscahnthe strangerwillgraduallyretumtoherchair.

Episode 8 (3 minutes): The mother opens the door and pauses for a moment, giving the

babyachanceto respond. Themotherthentalksto the babyandpickshim/herup. The
stranger leaves the room.

152

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ANﬁéﬁdﬁda

3:20:23.

v—mOBHmZ éZOmmmm

22%. cc ceaz Ea

4Nmeaesmm

d:— Som E 38.—on 58595
26; so» mm 80an >55 me 9.: 38% em .5 8: 2 26: 5: ea 90>
55o:
EEEEEEEE:
E355 .5 5.858.-
m53>oa 28 550::
£356....
85583 .528 .5 3.53...
3.55..
35 5505—52» .558 .5505 $28.5 835.2 .5 £382: 388..
5g .5 095%..
.88 So» 5 been me 282— %:aE me «a: new
:5» B Edges: ciao...— oﬁ he gem... .5.» Eu: 2 a: megs—em 05 83

 

 

 

 

 

 

 

 

 

 

Acids-fur;

essence. 253 e ceaz ram
"oi—Seam

”03555 mess—.8 05 5 ma
ding—25.2 05 3865 .55 use £535 .5 8.5: new bee um:

45» 2 38.—8.5 2a 2.3 .5 .5.» ..e
toga—8 3:82 0235 0:3 28.—on o5 =e 52300
.29. 2.: .5 0.:— Sox 5 cannon ESE—Em :25 a: 88E

OEHM<Hm ”axon—mm m0<m m—Ib ZO
mzozbmd—Q 15< amy— mm<mqm

ohms—commend tonnem 38m
_. 50me%

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.mm mm .8 mu

.2 .2 .2 .2

a a ..N ..N

8 .3 .8 2

2 .2 .2 .2

.2 2 .2 .2

2 2 .2 .2

.2 2 .2 2

.2 2 .2 2

E 2 .1 .2

2 .2 .2 .2

.2 N. .n_ .m—

: : .: :

.2 2 .2 .2

.o o .a .a

.w a .m .m

.2. A. .2. .2.

.o o .5 5

.m n .n .n

.v v .v .v

.m m .m .m

.N N .N .N

.2 _ ._ .—

emeeo5 5 2558

So» 5&3. .5 5:5 02mm 22.5.55 25 E 25:55 .5 350%“: .8: so» 855 $55. 5 58.: 50.: so» 8:22

:85: .25 50: 5:5 Bo: 05:50 .5» :8 :35 so: 525 25 8oz :25 3o: :85: 25 moo: .52.: >6:

3. 5:250 "m 5:250 "a 55830 3 5252.0
Egg—ware .Seanzwmnv
.2 a see u m an a 2:... n m
20555:. n N 205.555 u N
2:: a u _ 0.5: a u _
52552212 zgmeoeuo

5:95 55 552:: 05 5

west», .3 52525 wage—.5 2: 5325 88:. .35: so» :85: :25 .5:

154

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.vn .vm
.8 .3
.NN .NN
.3 ..N
.3 .8
.a— d—
.E .m—
.: .2
.5. .5—
.m _ .m _
.3 .3
.m _ .m—
.N_ .m—
.: .2
.o. .5—
.o d
.m .w
.2. K
.5 .5
.m .n
.5 .v
.m .m
.N .N
._ ._

A585 05 2 .5»
8.8 :23 .525: 25 :5: 22.—3:05 w:_.55 5098 2 5:3 .55 :23
5w .5: 55 552.205 53: 2.58.5: 25 :5: 5w no» 55 .65:qu
8522222 58585 :02: Bo: :52: 85552 5:: 835: 22:: so:

“a 5:850 "2. 5.52.0

=55 59m 5 n v
.5 a as: n M
22:555.: .1. m
2:: a u 2

.5: 5 5: u c

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.3 .MN
.NN .NN
._N 3
.cm .3
.3 .2
.2 .m—
.: .2.—
o. .5—
.n _ .n _
.3 .3
m. .m—
.N_ .N.
.3 . :
c. d—
.o d
.w .w
.2. .2.
.5 .5
.w .n
.5 .v
.m .m
.N .N
._ .—

2222 2.2.2. :85: 25 5.8

2.3% 20: :85: 25 5.38 :85 Bo: .20: 325222 55:
:02: Bo: .383 358m 5.: 2:8. 5 .585 05 2 05: a
5B 2 505:8 053 25%.: .55 385: 8 5058: .52.:
5 5.52.0 “m 838:0

.85 50..» a n v

2: a 02:: n n

2055558 u N

2:: a n _

=2: 5 5: u o

155

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.mm .mm
.- .2
..N ..N
.on .2...
.2 .2
.2 .3
.2 .2
.2 e.
.2 m _
.3 .1
.2 .2
.2 .2
.2 :
.o- c—
.a .o
.w .m
.N. H
.e .e
.m .m
.v .v
.m .m
.n .m
._ ._
m8. 8 .8» a 8:0 u _ 38» m 55 0.58 n m
gamoEmHBomanm Eogmoamuv
.258... u m 38» m 2 _ n m
583 n v £85 2 2 o u N
33. u m mﬁcoE o .85 m8— " _
A3032 6 53> .mzao 25:3
28.2. £5 :23 83:8 9:5 5.8me $5
.233: :9» av bacon—xx“ So: 8505— 33 96; 32 30$
ﬁ— 538:0 u: 533:0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

em .vm
mm .MN
2 an
_N .3
.2 .3
d— d—
.2 .2
.2 .2
2 .2
n _ .n _
3 .3
.2 .2
:2 .m—
._ _ .2
d— .c.
.o .o
.w w
.h N.
.e o
.m m
.9 v
.m m
.N N
._ _
A353 a

$5» 23.38:. 8 JOEE... 288 .08: 2.3% 5:3 .6» 2 38m: ..wdv
.5 £33.55 Soon .5 .oNBES $.8th £5 :5: 6w .5» cu .65,—wow;

com—on £5 moon :02: 3o: Sena tang—a 35:25 :02: 3o:
5— 328:0 “a 3.3.2.0
. 3% 82m a n v

E a 3:... u m
338038 n m
2:: a u _

=m 3 8: n o

156

.9... .85 < .v

.5 a 230 .m

55:5 055.55 < .N
2:: < ._

:6 a 0:62 .6

«so» 2 2:253: 5mg. 9.. 95 0:3 0:58 08:: .3 3355 .53 con—95 =5» .5 :02: 3c: £955 :2

 

$325 .25
_::a.Smo_5\§m_£E

3530:: 5 5.3550

502.65 28 :28:

53:30:

05583 50:8 5 :53

manor:

Steam 835.2 5 95:52: 38::
55:5 .5 3.5%

 

 

 

45.: 2 2:255 5mg. o: 98 0:3 x5350 :28 :5: £558 .5 5:53: 05 286£ 887. am—

nm<m> hm<m mi... 07550 mmEmZOF<AmM HZ<P¢OAE~ .50.: 30> n:

mm> .—

335 38.35:: 538:.» a: QZ§ ...53m.~382t OZ .o

3:88: 5:5 2:8 5 .58: 60:838. 5 0053: 655:0 :o_. a $555 3 26 main—5:22 agtoas: .95 5o. .5.» 96: .56.: 5.5 0:: :35 .m—

157

APPENDIX K
Postnatal Pregnancy Support Items

1. How much advice and guidance about parenting do you get from this person (e. g.,
baby’s diet, how to comfort the baby)?

2. How much practical assistance about parenting do you get from this person (e. g.,
take you to the pediatrician, lend or give you things for the baby, babysit)?

3. How much emotional support about parenting do you get from this person (e. g.,
listens to you when you are upset or worried)?

158

 

7.

APPENDIX L

Demographic Questionnaire
Yourdateofbirth: __ /__ /__
(m0) (613') (yr)
Your baby’s due date: / /

Eric) (21;) (36
Have you been pregnant before? (Circle one)
1 = YES 2 = NO (If NO, go to Question 7)

If YES, to Question 3:

4. How many times?

5. Have you had any miscarriages, still births, or abortions? (Circle one)
1 = YES
2 = NO

6. How many biological children do you currently have?

How many people, including yourself, live in your household?
(If participant is living in a shelter, questions 7 & 8 refer to household

composition before moving into shelter.)

 

 

 

 

 

 

 

 

 

8. Please list these: (Write in speciﬁc relationship to mother. Be speciﬁc--is the
person (for ex.) a husband, stepfather, biological child, foster child, or partner's
child?)

9. Choose the one that best describes your current marital/relationship status (choose

only one):

(a) single, never married (see below)
(b) married For how long? (in months)
(c) separated For how long? (in months)
(d) divorced For how long? (in months)
(e) widowed For how long? (in months)
If (a) is circled: Are you currently in a relationship? YES NO
If YES, go to Question 10.
If NO, were you in a relationship that lasted at least
6 weeks during your current pregnancy? YES NO
10. First name of your current partner or the partner you were with for at least 6

159

11.

12.

13.

14.

15.

16.

weeks during your pregnancy:

 

Are you currently living with your partner/spouse? (Circle one)
1 = YES
2 = NO

If yes to Question 11, how long have you been doing so? (Circle one)
1=lessthan 1 year
2 =1-3 years
3 =4-6 years
4= 7-9 years
5 = 10-12 years
6= 13-15 years
7 =16 - 18 years
8 =19 - 21 years
9=22 -24 years
10 = 25 or more years

Prior to your current romantic relationship, speciﬁed in Question #10

(a) were you ever married? 1 = YES 2 = NO
(b) did you ever live with a partner? 1 = YES 2 = NO
(c) were you ever separated? 1 = YES 2 = NO
((1) were you ever divorced? 1 = YES 2 = NO
(e) were you ever widowed? l = YES 2 = NO
What is your current relationship with the father of your baby? (Circle one)

1 = spouse

2 = ex-spouse

3 = partner

4 = ex-partner

5 = friend

6 = acquaintance

7 = stranger

8 = other Please specify:

 

What is your racial or ethnic group? (Circle one)
1 = Native American
2 = Asian American/Paciﬁc Islander
3 = Black, African American
4 = Latino, Hispanic, Chicano
5 = Biracial (mixed): Specify
6 = Caucasian, White
7 = Other:

 

What is the baby’s father’s racial or ethnic group? (Circle one)
1 = Native American

160

2 = Asian American/Paciﬁc Islander
3 = Black, African American

4 = Latino, Hispanic, Chicano

5 = Biracial (mixed): Specify
6 = Caucasian, White

7 = Other:

 

17. What is the highest level of education you have completed? (Circle one)
1 = grades 1, 2, 3, 4, 5, or 6 (circle speciﬁc grade)
2 = grades 7, 8, 9, 10, 11, 12, GED (circle speciﬁc grade)

 

 

 

 

 

3 = some college Where?
4 = AA degree Where?
5 = BA/BS Where?
6 = some grad school Where?
7 = graduate degree Where?

MA?

Ph.D.?

Law?

MD?

 

8 = other; Specify (e.g., Beauty School, nursing school)

 

18. Do you currently work outside the home? YES NO
If N0, did you work outside the home during the last year? YES NO

19. If YES to either part of Question 18, what is/was your
occupation?
Please be specific. For example, bookkeeper, cashier, computer
programmer.
If there were two jobs/occupations, have participant choose the one that
she feels best represents her occupation.

20. What is the highest level of education your partner/spouse has completed? (Circle
one)

1 = grades 1, 2, 3, 4, 5, or 6 (circle speciﬁc grade)

2 = grades 7, 8, 9, 10, 11, 12, GED (circle speciﬁc grade)

 

 

 

 

 

3 = some college Where?
4 = AA degree Where?
5 = BA/BS Where?
6 = some grad school Where?
7 = graduate degree Where?

MA?

Ph.D.?

Law?

MD?

161

8 = other; Specify (e.g., Beauty School, nursing school)

 

21. Does s/he work outside the home? (Circle one)
1 = YES
2 = NO

22. If yes to Question 20, what is his/her occupation?

 

23. What is your total family income per month (estimate)?

 

24. Do you currently receive any public assistance? (Circle one)
1 = YES
2 = NO

25. Are you currently residing in a shelter for battered women?
YES NO # days?

26. Have you ever stayed in a shelter for battered women before . . .
(a) Because of your experience of abuse? YES NO # days?

(b) Because of your mother’s/guardian’s

experience of abuse? YES NO # days?

27. Have you ever stayed in a homeless shelter before . . .

(a) Because of your experiences? YES NO # days?

(b) Because of your parents’lguardians’

162

APPENDIX M
Beck Depression Inventory

BDI

In answering these questions, think about each item carefully and circle the answer out of
the group of 4 items that best reﬂects how you have been feeling during the past week.

1.

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

I do not feel sad.

I feel sad.

Iamsadallthetimeandlcan’t snapoutofit.
Iamso sadorunhappythatlcan’tstandit.

I am not particularly discouraged about the future.

I feel discouraged about the future.

I feel I have nothing to look forward to.

I feel that the future is hopeless and that things cannot improve.

I do not feel like a failure.

I feel I have failed more than the average person.

As I look back on my life, all I can see is a lot offailures.
I feel I am a complete failure as a person.

I get as much satisfaction out of things as I used to.
I don’t enjoy things the way I used to.

I don’t get real satisfaction out of anything anymore.
I am dissatisﬁed or bored with everything.

I don’t feel particularly guilty.

I feel guilty a good part of the time.
I feel quite guilty most of the time.
I feel guilty all of the time.

I don’t feel I am being punished.
I feel I may be punished.

I expect to be punished.

I feel I am being punished.

I don’t feel disappointed in myself.
I am disappointed in myself.

I am disgusted with myself.

I hate myself.

I don’t feel I am any worse than anybody else.

I am critical of myself for all my weaknesses or mistakes.
I blame myselfall the time for my faults.

I blame myself for everything bad that happens.

163

 

10.

11.

12.

13.

14.

15.

16.

17.

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]

[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

[1]
[2]
[3]
[4]

I don’t have any thoughts of killing myself.

I have thoughts of killing myself, but I would not carry them out.
I would like to kill myself.

Iwouldkillmyselfiflhadthechance.

Idon’t cryany morethanusual.

Icrymorenowthanlusedto.

Icryallthetirnenow.

Iusedto beable to cry, but nowI can’t cryeven thoughl want to.

Iamno more irritatedbythingsthanleveram.

I am slightly more irritated now than usual

I am quite annoyed or irritated a good deal of the time.
I feelirritatedallthetimenow.

I have not lost interest in other people.

I am less interested in other people than I used to be.
I have lost most of my interest in other people.

I have lost all of my interest in other people.

I make decisions about as well as I ever could.

I put oﬂ‘ making decisions more than I used to.

I have greater difﬁculty in making decisions than before.
I can’t make decisions at all anymore.

I don’t feel that I look any worse than I used to.

I am worried that I am looking old or unattractive.

I feel that there are permanent changes in my appearance that make me
look unattractive.

I believe that I look ugly.

I can work about as well as before.

It takes an extra effort to get started at doing something.
I have to push myself very hard to do anything.

I can’t do any work at all.

I can sleep as well as usual.

I don’t sleep as well as I used to.

I wake up 1-2 hours earlier than usual and ﬁnd it hard to get back to sleep.
I wake up several hours earlier than I used to and cannot get back to sleep.

I don’t get more tired than usual.

I gettired more easilythanlusedto.

I get tired ﬁ'om doing almost everything.
I amtoo tired to do anything.

164

18.

19.

20.

21.

[1]
[2]
[3]
[4]

My appetite is no worse than usual.

My appetite is not as good as it used to be.
My appetite is much worse now.

I have no appetite at all anymore.

I haven’t lost much weight, if any, lately.
I have lost more than ﬁve pounds.

I have lost more than ten pounds.

I have lost more than ﬁfteen pounds.

I am no more worried about my health than usual.

I am worried about physical problems such as aches and pains, or upset
stomach, or constipation.

I am very worried about my physical problems and it’s hard to think of
much else.

I am so worried about my physical problems that I cannot think about

anything else.

I have not noticed any recent change in my interest in sex.
I am less interested in sex thanI used to be.

I am much less interested in sex now.

I have lost interest in sex completely.

165

 

APPENDIX N
Severity of Violence Against Women Scales

VAW Scales-Pregnancy Interview, Pt. I

[NAME see Page 2, Question 10].*****
You and have probably experienced anger or conﬂict. Below is a list of behaviors he may have
done. Describe how often he has done each behavior at 2 different times (during your current pregnancy
and the year before you became pregnant) by choosing a letter ﬁ'om the following scale. [Interviewer: (f

“*“This questionnaire refers to

Acted like he wanted to kill you

Held you down, pinning you in
place

_Pushed or shoved you

_Shook or roughly handled you

166

Physically forced you to have sex
Made you have anal sex against your
will
Usedanobjectonyouinasexual
way
you suddenly or forcefully

participant was not involved in a relationship with dwing the year before she became pregnant,
code E]
A BB C D E
never once a few times many times not applicable
During your current pregnancy: During your current pregnancy:
The year before you became pregnant: The year before you became pregnant:
_Hit, kicked a wall, door, or furniture _ _Scratched you
__ _Threw, smashed or broke an object _ _Pulled your hair
__ _Driven dangerously with you in the _ _Twisted your arm
car _ _Spanked you
_ _Threw an object at you __ __Bit you
__ _Shook a ﬁnger at you _ _Slapped you with the palm of hand
__ _Made threatening gestures or faces at _ _Slapped you with the back of hand
you __ _Slapped you around your lace/head
__ _Shook a ﬁst at you __ _Kicked you
__ _Acted like a bully toward you __ _Hit you with an object
__ _Destroyed something belonging to __ _____Stomped on you
you _ _Choked you
__ _Threatened to harm or damage things __ _Punched you
you care about __ _Btu'ned you with something
_ _Threatened to destroy property _ _Used 3 club-like object on you
__ _Threatened someone you care about _ _Beat you up
__ _Threatened to hurt you __ _Used a knife or gun on you
_ _Threatened to kill himself _ _Dernanded sex whether you wanted
_ _Threatened you with a club-like to or not
object _Made you have oral sex against your
__ _Threatened you with a knife or gun will
__ _Threatened to kill you _ _Made you have sexual intercourse
_ _Threatened you with a weapon against your will

Severity of Violence Against Women Scales-Previous Partner

VAW Scales—Pregnancy Interview, Pt. II
How often did your most recent previous partner (the person before (NAME) engage in each of
these activities with you? [ Interviewer: Relationship with previous partner must have lasted 6 weeks or
longer in order to complete questionnaire. ]

 

 

Please give the dates of this relationship: / to /
(mo) (yr) (mo) (yr)
A BB C D E
never once a few times many times not applicable
How often did your pamter: How oﬁen did your partner:
_Hit, kicked a wall, door, or furniture _ _Scratched you
_Threw, smashed or broke an object __ _Pulled your hair
_Driven dangerously with you in the _ _Twisted yom arm
car _ _Spanked you
_ _Threw an object at you _ _Bit you
_ _Shook a ﬁnger at you __ _Slapped you with the palm of hand
_ _Made threatening gestures or faces at _ _Slapped you with the back of hand
you _ _Slapped you around your face/head
__ _Shook a ﬁst at you __ _Kicked you
_ _Acted like a bully toward you _ _Hit you with an object
_Destroyed something belonging to __ _Stomped on you
you __ _Choked you
__ _Threatened to harm or damage things _ _Punched you
you care about _ _Burned you with something
_ _Threatened to destroy property __ _Used a club-like object on you
_ _Threatened someone you care about __ _Beat you up
__ _Threatened to hurt you __ _Used a knife or gun on you
_ _Threatened to kill himself _ _Demanded sex whether you wanted
_ _Threatened you with a club-like to or not
object __ _Made you have oral sex against your
_ _Threatened you with a knife or gun will
_ _Threatened to kill you __ _Made you have sexual intercourse
__ _Threatened you with a weapon against your will
_ _Acted like he wanted to kill you __ _Physically forced you to have sex
__ _Held you down, pinning you in _ _Made you have anal sex against your
place will
_Pushedorshovedyou __ _ Usedanobjectonyouinasexual
__ _Shook or roughly handled you way

Grabbed you suddenly or forceﬁrlly

167

APPENDIX 0
Severity of Violence Against Women Scale-Time 3 Interview

You and your partner have probably experienced anger or conﬂict. Below is a list of behaviors he may have
done. Describe how often he has done each behavior during the last year by choosing a letter ﬁ'om the
following scale. I Interviewer: If participant was not involved in a relationship with a partner drm'ng the

year code E]
A BB C D E
never once a few times many times not applicable
During the last year, how often did your partner: How often did your partner:
_ _Hit or kicked a wall, door or furniture __ _Scratched you
__ _Hit, kicked a wall, door, or furniture _ _Pulled your hair
_ _Threw, smashed or broke an object __ __’l’wisted your arm
__ _Driven dangerously with you in the __ _Spanked you
car _ _Bit you
_ _Threw an object at you _ _Slapped you with the palm of hand
__ _Shook a ﬁnger at you __ _Slapped you with the back of hand
_ _Made threatening gestures or ﬁrces at _ _Slapped you around your face/head
you __ _Kicked you
__ _Shook a ﬁst at you _ _Hit you with an object
__ _Acted like a bully toward you
_ _Destroyed something belonging to __ _Stomped on you
you __ _Choked you
__ _Threatened to harm or damage things __ _Punched you
you care about __ _Burned you with something
_ _Threatened to destroy property _ _Used a club-like object on you
__ _Threatened someone you care about __ _Beat you up
__ _Threatened to hurt you __ _Used a knife or gun on you
__ _Threatened to kill himself __ _Demanded sex whether you wanted
_ _Threatened you with a club-like to or not
object __ _Made you have oral sex against your
_ _Threatened you with a knife or gun will
_ _Threatened to kill you _ _Made you have sexual intercourse
_ _Threatened you with a weapon against your will
__ _Acted like he wanted to kill you __ _Physically forced you to have sex
_Held you down, pinning you in _ _Made you have anal sex against your
place will
__ _Pushedorshovedyou __ __ Usedanobjectonyouinasexual
_Shook or roughly handled you way
_ _ Grabbed you suddenly or forcefully

168

 

APPENDIX P
Parent Behavior Checklist

PBC

This checklist includes statements about how parents raise young children. For each statement,
mark the number 1 if the statement ALMOST ALWAYS OR ALWAYS applies to how you raise
your child. Mark the number 2 if the statement FREQUENTLY applies. Mark the number 3 if
the statement SOMETIMES applies. Mark the number 4 if the statement ALMOST NEVER OR
NEVER applies. A sample item is shown below.

Almost Always=l Frequently=2 Sometimss=3 Almost Never/Never“!

99999.“?

10.
l I.
12.
13.
14.
15.
l6.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.

I spank my child at least once a week.

If my child would hit, kick, bite, or scratch someone, I would spank him/her.
If my child hit me in anger, I would hit or spank my child.

When my child doesn’t do what I tell him/her to do I spank him/her.
I tell my child that his/her bad behavior will make God sad.

I yell at my child for whining.

I tell my child he/she should be ashamed of him/herself for soiled pants (bowel
movement).

If my child is overactive, I yell at him/her.

If my child cries after being put to bed, I spank him/her.

To toilet train my child, I make him/her sit on the toilet for over 15 minutes.
I spank my child for refusing to eat.

I would spank my child in public for bad behavior.

I yell at my child for being too noisy at home.

I scold my child for soiling in his/her pants.

I threaten to tell my spouse/partner about my child’s bad behavior.
I tell my child that he/she is bad.

I scold my child for playing with his/her private parts.

I tell my child to behave so that my spouse/partner won’t get mad.
If my child cries after being put to bed, I yell at him/her.

I yell at my child for spilling food.

I get so angry with my child I spank him/her on the bottom.

I punish my child for wetting the bed.

I make my child stay at the table until all of his/her food is gone.

I would spank my child for wetting his/her pants.

I would slap my child for being sassy or backtalking.

I threaten to punish my child but then I don’t.

I hit my child with an object (such as a spoon or belt).

I tell my child God doesn’t like children who lie.

When my child has a temper tantrum, I spank him/her.

I send my child to bed as a punishment.

169

rm. . m- . ‘

 

APPENDIX Q
Physical and Emotional Neglect Scale

Please answer the following items usinLthe scale below.

 

 

 

 

 

 

 

 

 

Never Almost Sometimes Frequently Ahnost Always
Never Always
l 2 3 4 5 6

_ 1. I hug and kiss my child everyday.

__ 2. I feed my child on a regular schedule.

__ 3. I have run into the store and left my child in the car for a few minutes.

_ 4. During the day, I let my child cry for 15 minutes before I pick him/ha up.

__ 5. My child has had diaper rash.

__ 6. I play games with my child that will help him/her develop new skills.

__ 7. I put prescription and over-the-counter medicine out of the reach of my child.

__ 8. I change my child’s diaper 4 times a day or more.

__ 9. I take my child to the doctor for regular checkups.

__ 10. I talk back and forth with my child every day.

__ l 1. Once my child is asleep for the night, I have left him/her alone so that I could go
out.

_____ 12. I give my child lots of encouragement when he/she does something new (such as
walking or making a new sound).

__ 13. It is hard to know when my child is hungry.

_ 14. I play with my child every day.

__ 15. I arrange child care when I can’t be with my child.

__ 16. Whatever I do for my child, he/she is never satisﬁed.

_;__ 17. I have left my child at home alone during the daytime so that I could run a quick
errand.

__ 18. I tell my child that I love him/her every day.

__ 19. I sometimes let my child go a couple of days without washing him/her.

__ 20. I have trouble making sure my house gets cleaned regularly.

__ 21. I try to provide time for my child to play with other children.

__ 22. I put child safety plugs on all the electrical outlets in my house.

_ 23. If I’m upset with my child, I sometimes don’t talk to him/her all day.

__ 24. I change my child’s diaper less than 2 times a day.

__ 25. It’s hard to understand what my child is crying about.

__ 26. I feed my child within 30 minutes of when s/he is hungry.

__ 27. There are days when I’m so busy, that I barely see or talk to my child.

28. I put household cleaning products and chemicals in cabinets that my child can’t get
open.
__ 29. During the day, I let my child cry for 30 minutes or longer before I pick him/her up.
30. I feed my child at least 3 times a day.

170

 

 

vomd
Nemd-
oomd-
dend-
eemd-
35d
eedd-
dm ~ d
Nood-
mmdd
mm _ d
mm _ d
doc:

Q

_N _ d
NoNd-
momd-
eeNd-
emvd-
madd-
eo : d-
md _ d
emdd-
mod-
N: _ d
mm: d
weed
cod.—

2

edmd
mmmd-
m _ m d-
meNd-
deed-
mudd-
mm— d-
am: d
deod-
deod-
md _ d
03 d
emod
mmed
Odd.—

d—

3 _ d
NmNd-
edmd-
vmmd-
vemd-
d: :d-
dm _ d-
mN : d
mmdd-
_vod-
wood
we _ d
mood
cued
dNed
odd.—

Ne _ d
mmmd-
dcmd-
ddmd-
_Nvd-
madd-
mm _ d-
9; d
wedd-
eedd-
oedd
on _ d
vmod
Nmed
awed
mmed
coo.—

emdd
ee _ d-
ew _ d-
emNd-
_Nvd-
omdd
need-
em _ d
N—dd
vmdd-
ewdd
e_ _d
ecod
emed
Nded
wood
tamed
doc:

Edd-
N— _d

eddd

mmNd

god

3; d-
deod-
mo: d-
mm — d-
dm: d-
e_md-
me ~ d-
dm _ d-
wde-
gad-
mm _ d-
mm _ d-
momd-
Odd;

emdd-
mm _ d

momd

mend

m— ~d

modd-
d _dd-
m m m d-
3 _d-
Vm _ d-
mand-
m ﬁd-
em _ d-
9; d-
nond-
m m _ d-
me _ d-
evod-
_de

doc:

add-
3 _ d
mood
em _ d
an _ d
mead-
md: d-
doNd-
e— _d-
3 _ d-
mdmd-
mm _ d-
_mmd-
mend-
dend-
dmmd-
edmd-
demd-
_Nvd
em _ d
odd.—

vo _ d-
wood
comd
m ~ Nd
em _ d
dadd-
m 8d-
mm _ d-
em _ d-
dwmd-
mwmd-
cdmd-
mw ~ d-
3 _ d-
w _ Nd-
:3 d-
we— d-
3.— d-
mwmd
mmmd
emmd
Odd.—

ded
m 8d
e: _d-
Nodd-
dmod-
d—Nd
e: d
fwd
eVNd
memd
memd
wmmd
3 _d
N: _d
S _d
weed
eedd
em _ d
misd-
mmmd-
e _ md-
owed-
coo.—

N

Neod-
em _ d
mm _ d
d3 d
mN— d
mde-
cm _ d-
de _ d-
mund-
eemd-
mmmd-
emmd-
NmNd-
v3 d-
ommd-
we _ d-
_dmd-
Na d-
mmmd
meed
ddvd
coed
mead-
doc.—
_

em
mm
mm
3..
on
e—
m:
e—
w—
m—
3
m:
N—
_—
d—

v-‘NMVWOFQO

 

mozaa> 382 325% as: 852:8

M X_Dmem<

171

.m 23 5 Ba: 2.. ”Season 32% assess.

sacrum 8302.032 220nm .8835 gaseous geared .2532": .me unannogucm .eoasm 2535 ﬁne .283
.882: 2.": £83 3885 men: .eaaam gee-.5 Sue. eaasm .855 Eu: .eaasm _aaeosm Eu: .8372 Sam .0 an:
623nm. .8§a§< 623". _ .eseeem _ozaue .88sﬁo 623k 356 2 message 623"” £892»: .3 aosea 623A

5852 oz 05:70 5:585; USE-m .3383 35".. .35.. 05:3 .93 o<<aum .8803 05...". as 8325 552

 

ooo._ wmmd-
ooo._

wm

mm

em _ .o-
eemd
ooo._

mm

mm g .o-
oe—d
momd
ooo._

_m

wmmd-

_mmd
m_w.o
mm_.o
ooo._

om

_e _ .o
meod-
mo _ .o-
em _ .o-
wmod
ooo._

e—

oeod
owod-
_mod-
_mod-
~e_.o
weed
ooo._

w—

ww: .o
mm _ .o-
momd-
wemd-
mood
eemd
owmd
ooo.~

e:

ewod
wmod
wwod-
eeod-
mwod
m m md
mwmd
mm _ .o
ooo._

e—

wmod
mwod-
we : d-
m m _ .o-
wmod
mwmd
emwd
ommd
mmed
ooo._

m—

ww _ .o
ow. .o-
eemd-
m—md-
mood
ewmd
mwmd
emmd
eewd
oemd
ooo._

w—

m_o.o
eomd-
oo _ .o-
memd-
m mod-
momd
mm _ .o
eemd
em ~ .o
we : .o
mwmd
ooo._

m:

wm
mm
mm
_m
om
e—
m:
e—
e:
m—
w:
m:
m:

e-It
'—

F‘NMVWOBMO‘O

172

APPENDIX S

 

 

 

 

 

 

 

 

  
  

 

 

    

 

 

  

 

PAAQ Measurement Model
. ‘ F‘~0J5
rejectxo
0.
love 0 7,
‘l*'-L3
1. anqry
r0 . 30
-O.14
derogate W0, 85)]8
vulnerab I 0.80 0'23
nomemory
’0 . 76

Chi-Square=4.25, df=4, P-value=0.37270, RMSEA=0.018

173

APPENDIX T
WMCI Measurement Model

 

rofp ”-0.37

 

 

opch I"'0.26

 

 

coher H-O . 32

 

 

 

i89\

    
    

caresens

ﬂFmZI

 

 

accept *0. 21

 

Chi-Square=6.64, df=5, P-value=0.24905, RMSEA=0.040

174

I
I!

 

APPENDIX U
Time 1 Social Support Measurement Model

 

 

 

 

 

 

 

 

 

tlfemoav
1.00
1.00
0.6 .
tlfaldav
1.00
\\\\\‘~ tlfpreav

 

 

 

0.0

0.64 ‘0.50

Chi-Square=1.99, df=1, P-value=0.15852, RMSEA=0.069

175

APPENDIX V
Time 3 Social Support Measurement Model

 

 

 

 

 

 

 

 

t3femoav 0.0
1.00
1.00 t3socsup
- 0'62“ t3faidav 'C-O.62 -0.42
1.00

 

\ t3fpreav 0.0

Chi-Square=10.72, df=1, P-value=0.00106, RMSEA=O.217

 

 

 

176

APPENDIX W
Risk Factors Measurement Model

 

demorisk “'0. 69

 

 

bdirisk L"0. 87

 

 

 

dvrisk l""0.46

\ chabrisk .82

Chi-Square=3.90, df=2, P-value=0.l4243, RMSEA=0.068

 

 

 

 

 

177

APPENDIX X
Revised Risk Factors Measurement Model

 

demorisk 0.00

 

 

 

1.00

1 . 00
0 LOO“ bdirisk 0.00

0.44

\

Chi-Square=4.68, df=2, P—value=0.09649, RMSEA=0.081

 

 

 

dvrisk l.’O.80

 

 

 

178

 

vomd
vomd-
mwmd-
bond-
_vod
Rod-
om ~ .o
mood-
wmod
on _ d
on _ d
ooo._

~—

-~ .o
ommd-
_omd-
omvd-
mood-
homo-
mo—d
wmod-
_mod-
Q ~ d
mm _ .o
wood
ooo._

:

momd
u _md-
momd-
oovd-
wmod-
mm _ d-
on _ d
oood-
ohod-
mo_ d
o3 d
omod
mwmd
ooo._

o~

_o~ .o
_de-
oomd-
vomd-
o_ _d-
on _ d-
mm— d
wmod-
god-
wood
on _ d
wood
omhd
omnd
oooA

N3 d
oomd-
momd-
"Nvd-
mood-
mm _ d-
m: d
vood-
shod-
ohod
om _ .o
vwod
mwmd
mwmd
wmbd
ooo._

owod
_oNd-
~NNd-
_Nvd-
omod
wmod-
hm _ d
N _od
wmod-
hood
2 ~ d
mood
omhd
momd
vood
wmod
ooo._

Eod-
Nmod

ow _ d

oood

3; d-
ovod-
wo_ d-
on _ d-
om _ d-
h _ Nd-
Nb _ d-
ow _ d-
mwmd-
gmd-
on _ d-
mm _ d-
woNd-
ooo._

omod-
mwmd

ommd

m _ _d

oood-
o_od-
mm _ .o-
o_ _d-
v» _ d.
mwmd-
2nd-
so _ .o-
o: d-
momd-
mm _ d-
on _ .o-
ovod-
Gad

ooo._

wmod-
wmod

oood

on _ d

moNd-
wo— d-
ooNd-
3 _d-
em _ d-
momd-
No _ d-
_mmd-
mwmd-
oomd-
ommd-
ooNd-
ohmd-
_Nvd

vm _ d

ooo._

voﬁo-
emwd
Nmmd
om _ .o
owod-
m 3.?
”2d-
vm ~ d-
oomd-
mwmd-
oomd-
mw—d-
mm _ d-
«Ed.
2; d-
mad.
3; .o-
mwmd
wmod
ommd
ooo._

omod
hm _ d-
oNNd-
omod-
o—md
a: .o
o—Nd
nvmd
momd
mwmd
mwmd
3 _d
N— _d
N- _d
wood
wood
hm _ d
vad-
mwmd-
h _ md-
omnd-
ooo._

whod-
vmmd
momd
mﬁ .o
momd-
em _ d-
on _ d-
mwmd-
RNd-
mwmd-
ommd-
mwmd-
vo _ d-
ommd-
vs _ d-
_oNd-
No _ d-
mwmd
movd
oovd
oohd
momd-
ooo.~
~

~vam©hw©O~NM
v—nv—ov—n—v—iv—nq—n—tv—INNNN

v—NMVWVDFOOGO
g—n

 

m03§> .25: En do €32 accustom

> X52973.

179

8.88., 8: 32. .e 8... 2.. s 88 m 23. 5 BE. .8 88888 8.885 8583:.

285-8 .28 _ 2st 8885 2858-2 .28 _ 2:5 88.8.89 3882L~ .wa gigoﬁouom .885 8585 ﬁne 688m
.885 2.": .888 3885 ﬂu: 8.35 8885 in: .885 .889: 2.": .885 3885 in: .82-2 gum 5 .8:
623-2 .8csa§< 623-: 53283 623-2 88880 623-8 .885 2 8:88 623-» 88888.. 8 8:82 623A
.882 oz 055-» 8:385; 055-“ .3383 055% .35. o<<._um .23 o<§u~ .88? 055-. new 8.88., 882

 

ooo. _

mm

wo~d-
ooo.—

mm

omﬁ d-
b _ md
ooo.—

_m

wmmd-
mmwd
wood
ooo.—

om

_o _ .o
_o _ .o-
Rod-
owod
ooo.—

@—

oood
oood-
mood
_o _ d
wood
ooo;

w—

ww_ d
wNNd-
momd-
mood
ohm d
own d
ooo.—

C

owod
Sod-
wm _ d-
mwod
mmmd
Nwmd
mwﬁo
ooo.—

o_

wmod
o3 .o-
2 _d-
wmod
mwmd
omwd
ommd
mmod
ooo.—

m—

ww ~ d
wad-
momd-
mood
owmd
mwmd
ommd
oowd
oomd
ooo.—

w—

m—od
dood-
wand-
mmod-
momd
mo _ d
oomd
om ~ d
wo ~ d
mwmd
ooo.—

n—

mm
an
_N
om
o—
w—
2
o~
3
w—
2
N—

'—
'—

FNMVWOBMO‘O
ﬁ

180

APPENDIX Z
Standardized Residual Values for Observed and Latent Variables

Observed Variables: Residual Value:
1. Rejection -.02
2. Love .27
3. Angry .32
4. Derogating .85
5. Vulnerability .80
6. No Memory .76
7. Richness of Perceptions .36
8. Openness to Change .26
9. Coherence .32
10. Sensitivity .22
11. Acceptance .23
12. Representation Of Self .38
13. Self-Efﬁcacy .15
14. T1 Emotional Support .52
15. T1 Practical Support .33
16. T1 Caregiving Support .43
17. T3 Emotional Support .56
18. T3 Practical Support .37
19. T3 Caregiving Support .26
20. Demographic Risk .40
21 . Maternal Depression .84
22. Domestic Violence .69
23. Infant Attachment Security .67
Latent Variables:

1. Maternal Representations of 1.00

Attachment
2. Prenatal Representation of Infant .55
3. Prenatal Representation of Self .00
as Mother

4. Postnatal Self-Efﬁcacy .88
5. Prenatal Social Support 1.00
6. Postnatal Social Support .80
7. Risk Factors 1.00
8. Mother-Infant Attachment 1.00

181

APPENDIX AA

Lisrel 8.5 Syntax for Final Model
Test of original dissertation model with risk revised for T1
da ni=23 no=207
la

rejectio love angry derogate vulnerab nomemory rofp optoch coher caresens
accept recmse totalmeq tlfemoav t1faidavt1fpreavt3femoavt3faidav t3 fpreav demorisk
bdirisk dvrisk overconv

raﬁ=newﬁrll.dat

mo ny=23 ne=8 te=sy,ﬁ ps=sy,ﬁ be=ﬁ,fu

le

momatt wmci mse meq tlsocsup t3$ocsup risk infatt
frtel1te22te33te44te5Ste66te77te88te99te1010

frtell 11te1414te15 15te1616tel717te1818

ﬁ'te 19 19te 20 20te 21 21 te 22 22

frpsl 1p322p855p366ps77

ﬁ'ps33ps44p888te 12 12

valte23 23 te 13 13

frlyllly211y3lly4lly511y61

frly721y821y921y102ly112

frly123

frlyl34

ﬁ'ly1451y1551y165

frlyl76ly186lyl96

frly207ly217ly227

frly238

ﬁbe21be3 1

frbe32

frbe43

ﬁ-be65

frbe46

frbe84

frbe82

ﬁbe27

frpsSlps68p828
ﬁ'te221te201te223te64te53te61te31te225te18 15te222te203te 1422
frte1417te2221te216te1321te919te2021te722te57te723te719te414
frte316te521te321

Pd
ou setvscrsmime=mlad=oﬂit=1500

182

APPENDIX BB
Goodness of Fit Statistics

Degrees of Freedom = 194
Nomial Theory Weighted Least Squares Chi-Square = 242.03 (P = 0.011)
Estimated Non-centrality Parameter (NCP) = 48.03
90 Percent Conﬁdence Interval for NCP = (12.45 ; 91.77)

Minimum Fit Function Value = 1.26
Population Discrepancy Function Value (F0) = 0.23
90 Percent Conﬁdence Interval for F0 = (0.060 ; 0.45)
Root Mean Square Error of Approximation (RMSEA) = 0.035
90 Percent Conﬁdence Interval for RMSEA = (0.018 ; 0.048)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.97

Expected Cross-Validation Index (ECVI) = 1.97
90 Percent Conﬁdence Interval for ECVI = (1.80 ; 2.18)
ECVI for Saturated Model = 2.68
ECVI for Independence Model = 13.29

Chi-Square for Independence Model with 253 Degrees of Freedom = 2692.25
Independence AIC = 2738.25
Model AIC = 406.03
Saturated AIC = 552.00
Independence CAIC = 2837.90
Model CAIC = 761.31
Saturated CAIC = 1747.83

Normed Fit Index (NFI) = 0.90
Non-Normed Fit Index (NNFI) = 0.96
Parsimony Norrned Fit Index (PNFI) = 0.69
Comparative Fit Index (CPI) = 0.97
Incremental Fit Index (IFI) = 0.97
Relative Fit Index (RFI) = 0.87
Critical N (CN) = 193.09

Root Mean Square Residual (RMR) = 0.64
Standardized RMR = 0.090
Goodness of Fit Index (GFI) = 0.91
Adjusted Goodness of Fit Index (AGFI) = 0.87
Parsimony Goodness of Fit Index (PGFI) = 0.64

183

APPENDIX CC
Q-plot of Standardized Residuals

3.5 ..........................................................................
. x
I x
. x

*
. xx
x *
. xx
iii:
. x xx*
. **x
. xxx*
. *X**
. . x*xx
. . **x
. *x*x
. *tti
. xx*
. tint.
. x**x*.
. xx*
. xx*x
. *x*
. xxx
. x**
. x*x*x
. xx*x
. x*x
. xx
. x*
. *
. *
.x
x
x
-3.5 ..........................................................................
-35 3.5

Standardized Residuals

184

REFERENCES

185

REFERENCES

Ainsworth, MD. (1967). Infang in Uganda: Infant care and the gigowth of love.
Baltimore: Johns Hopkins University Press.

Ainsworth, M.D.S., Blehar, M., Waters, E., & Wall, S. (1978). Patterns of
attachment. Hillsdale, NJ: Erlbaum.

Ammaniti, M. (1991). Maternal representations during pregnancy and early
infant-mother interactions. Infant Mgilal Health Journal. 12. 246-255.

Ammaniti, M., Baumgartner, E., Candelori, C., Perucchini, P., Pola, M., Tambelli,
R., & Zampino, F. (1992). Representations and narratives during pregnancy. _ijjapt
Mental Health Joxm_1a_1,Li, 167-182.

Andersen, P.A., & Telleen, S.L. (1992). The relationship between social support
and maternal behaviors and attitudes: A meta-analytic review. American Journal of

Communig Pachology, 20, 753-774.

Bandura, A. (1977). Self-eﬁcacy: Toward a unifying theory of behavioral
change. Psychological Review. 84, 191-215.

Bandura, A. (1989). Regulation of cognitive processes through perceived self-
eﬂicacy. Developmental Psychology, 25, 729-735.

Barnett, D., Ganiban, J ., & Cicchetti, D. (1999). Maltreatment, negative
expressivity, and the development of Type D attachments from 12 to 24 months of age.

In J. Vondra & D. Barnett (Eds), Monoggaphs of the Society for Research in Child
Development, 64, (3, Serial No. 258).

Barnett, D., & Vondra, J .I. (1999). Atypical patterns of early attachment: Theory,
research, and current directions. In J. Vondra & D. Barnett (Eds), Mononghs of the
Society for Research in Child Develppment, 64, (3, Serial No. 258).

Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An
inventory for measuring depression. Archives of General Paychiagy, 4, 561-571.

Belsky, J. (1999). Interactioml and contextual determinants of security. In J.
Cassidy & P.R. Shaver (Eds), Handbook of Attachment (pp. 249-264). New York: The
Guilford Press.

Belsky, J., & Isabella, R. (1988). Maternal, infant, and social contextual

determinants of attachment security. In J. Belsky & T. Nezworski (Eds), Clgp' 'cal
Implications pf Attachr_r§nt (pp. 41-94). New Jersey: Lawrence Erlbaum Associates.

186

Benoit, D., & Parker, K. C. H. (1994). Stability and transmission of attachment
across three generations. Child Development, 65, 1444-1456.

Benoit, D., Parker, KC.H., & Zeanah, CH. (1997). Mothers’ representations of
their infants assessed prenatally: Stability and association with infants’ attachment

classiﬁcations. Journal of Child Psychology and Pachiapry, 38, 307-313.

Benoit, D., Zeanah, C.H., Parker, K.C.H., Nicholson, E., & Coolbear, J. (1997).
Working model of the child interview: Infant clinical status related to maternal

perceptions. Infant Mental Health Jopg_1a_.l, l_8_, 107-121.

Bibring, G.L., Dwyer, T.F., Huntington, D.S., & Valenstein, AF. (1961). A study
of the psychological processes in pregnancy and of the earliest mother-child relationship.
Psychoan_alytic Studyof the Child, 9-24.

Bogat, G.A., & Levendosky, A.A. (1999). Physical and Emotional NegLept Sca_l§.
Unpublished manuscript, Michigan State University.

Bowlby, J. (1944). Forty-four juvenile thieves: Their characters and home life.
Intepnational Journal of Psychoanalysis, 25, 107-127.

Bowlby, J. (1969). Attachment and loss: Volume 1. Attachment. New York:
Basic Books.

 

Bowlby, J. (1973). Attachment and loss: Volume 2. Separation. New York: Basic
Books.

Bowlby, J. (1979). The mm and breakmg' of aﬁectional bonds, London:
Tavistock.

Bowlby, J. (1980). Attachment and loss: Volume 3. Loss. sadness. and
dgpression. New York: Basic Books.

Bowlby, J. (1988). L secure base: Parent child attachment and healthy
development. New York: Basic Books.

Bretherton, I. (1990). Communication patterns, internal working models, and the
intergenerational transmission of attachment relationships. Infant Mental Health low,
l_1, 237-252.

Bretherton, I., Biringen, Z., Ridgeway, D., Masline, C., & Sherman, M. (1989).
Attachment: The parental perspective. Infant Mental Health Joml, 10, 203-221.

Bretherton, I., & Munholland, KA. (1999) Internal working models in attachment

relationships: A construct revisited. In J. Cassidy & P. Shaver (Eds), Handbook of
Attachment (pp. 89-111). New York: The Guilford Press.

187

Bretherton, l., Ridgeway, D., & Cassidy, J. (1990). Assessing internal working
models of the attachment relationship: An attachment story completion tasks for 3 year
olds. In M. Greenberg, D. Cicchetti, & E.M. Cummings (Eds), Attachment in the
_meschool yea_r_'s (pp. 273-308). Chicago: The Chicago University Press.

Bromwich, R. (1978). Workaig' with parents and infants. Austin, TX: Pro-Ed.

Bumberry, W., Oliver, J.M., & McClure, IN. (1978). Validation of the beck
depression inventory in a university population using psychiatric estimate as the criterion.
Journal of Consultm’ and Clinical Psychology, 46, 150-155.

Bus, A.G., & Van Ijzendoom, M.H. (1992). Patterns of attachment in ﬁequently

and inﬁequently reading mother-child dyads. The Jom of Genetic Psychology, 153,
395-403.

Carlson, E.A., & Sroufe, LA. (1995). Contribution of attachment theory to
developmental psychopathology. In D. Cicchetti & D.J. Cohen (Eds), Developmental

psychomthology, Volume 1: Theory and methods (pp. 581-617). New York: John Wiley
& Sons, Inc.

Cascardi, M., & O’Leary, K.D. (1992). Depressive symptomatology, self-esteem,
and self-blame in battered women. Journal of Family Violence, 7, 249-259.

Cassidy, J. (1988). Child-mother attachment and the self in six year olds. Child
Development, 59, 121-134.

Cassidy, J. (1990). Theoretical and methodological considerations in the study of
attachment and the self in young children. In M. Greenberg, D. Cicchetti, & E.M.
Cummings (Eds), Attachment in the preschool years (pp. 87-119). Chicago: The Chicago
University Press.

Cassidy, J., & Berlin, L]. (1994). The insecure/ambivalent pattern of attachment:
Theory and research. Chilﬁ Development 65 971-991.

 

Cassidy, J ., & Kobak, RR (1988). Avoidance and its relation to other defensive

processes. In J. Belsky & T. Nezworski (Eds), Clinical Implications of Attachment (pp.
300-323). Hillsdale, NJ: Lawrence Erlbaum.

Cicchetti, D. (2001, April). Pachopathology in lopggm' dm mtive: The
MLS and beyond. Paper presented at the Biennial Meeting of the Society for Research in
Child Development, Minneapolis Minnesota.

Cohen, J. (1968). Weighted kappa. Psycholpgrpd' Bulletpp' 70, 213-220.

188

Conrad, 8., Gross, D., Fogg, L., & Ruchala, P. (1992). Maternal conﬁdence,
knowledge, and quality of mother-toddler interactions: A preliminary study. Infant
Mental Health Jogml, 13, 353-362.

Crittenden, RM. (1985). Social networks, quality of child rearing, and child
development. Child Development, 56, 1299-1313.

Crittenden, RM. (1988). Relationships at risk. In J. Belsky & T. Nezworski
(Eds), Clinical Implications of Attachment (pp. 136-174). New Jersey: Lawrence
Erlbaum Associates.

Crnic, K.A., Greenberg, M.T., & Slough, N.M. (1986). Early stress and social
support inﬂuences on mothers’ and high risk infants’ functioning in late infancy. Infant
Mental Health JomaL 7, 19-32.

Crockenberg, S. (1981). Infant irritability, mother responsiveness, and social
support inﬂuences on the security of infant-mother attachment. Child Develme 52,
857-865.

Crockenberg, S. (1988). Social support and parenting. Journal pf Behavioral
Pediatrics 141-174.

 

Crowell, J .A., Fraley, R.C., & Shaver, RR. (1999). Measurement of individual
differences in adolescent and adult attachment. In J. Cassidy & P.R. Shaver (Eds),
Handbook of Attachment (pp. 434-465). New York: The Guilford Press.

Crowell, J.A., & Treboux, D. (1995). A review of adult attachment measures:
Implications for theory and research. Social Development, 4, 294-327.

Cummings, EM. (1990). Classiﬁcation of attachment on a continuum of felt
security. In M. Greenberg, D. Cicchetti, & E.M. Cummings (Eds), Aaachment in the
praschool years (pp. 311-338). Chicago: The Chicago University Press.

Cutrona, C.E., & Troutman, BR (1986). Social support, infant temperament, and
parenting self-emcacy: A mediational model of postpartum depression. Child
Development, 57, 1507- 1 5 1 8.

DeWolﬁ‘, M., & van Ijzendoom, M. (1997). Sensitivity and attachment: A meta-
analysis on parental antecedents of infant attachment. Chilg ﬁvalopmgat, 68, 571-591 .

Donovan, W.L., & Leavitt. LA. (1989). Maternal self-eﬁeacy and infant

attachment: Integrating physiology, perceptions, and behavior. Child Devglppmpt, 60,
460-472.

189

Donovan, W.L., Leavitt. L.A., & Walsh, R0. (1990). Maternal self-efﬁcacy:
Illusory control and its effect on susceptibility to learned helplessness. Child
Development, 61, 163 8-1647.

Egeland, B., & Farber, EA. (1984). Infant-mother attachment: Factors related to
its development and changes over time. Child Development, 55, 753-771.

Egeland, B., & Sroufe, LA. (1981). Attachment and early maltreatment. Child
Developmeat. 52, 44-52.

Eiden, RD., Teti, D.M., & Corns, KM. (1995). Maternal working models of
attachment, marital adjustment, and the parent-child relationship. Child Development, 66,
1504-1518.

Epstein, S. (1973). The self-concept revisited: Or a theory of a theory. American
P§ychologist, 404-416.

Epstein, S. (1983). The Mother-Father-Peer Scale. Unpublished manuscript.
University of Massachusetts, Amherst.

Erickson, M.F., Egeland, B., & Pianta, R. (1989). The effects of maltreatment on
the development of young children. In D. Cicchetti & V. Carlson (Eds), Child

maltreatment: Theog and research on the causes and consequences of child abuse and
neglect (pp. 647-684). Cambridge, MA: Cambridge University Press.

Fava Vizziello, G., Antonioli, M.E., Cocci, V., & Invernizzi, R (1993). From
pregnancy to motherhood: The structure of representative and narrative change. Infant
Ment_al Health Journal,_1_4, 4-16.

Fleiss, J.L., Cohen, J., & Everitt, BS. (1969). Large sample standard errors of
kappa and weighted kappa. chhological Bulletpa' 72, 323-327.

Fonagy, P., Steele, M., Moran, G., Steele, H., & Higgitt, A. (1993). Measuring the
ghost in the nursery: An empirical study of the relation between parents’ mental
representations of childhood experiences and their infants’ security of attachment.

Journal of th_e American Pachommc Assopiatiop, 41, 957-989.

Fonagy, P., Steele, H., & Steele, M. (1991). Maternal representations of
attachment during pregnancy predict the organization of infant-mother attachment at one
year of age. Child Development, 62, 891-905.

 

Fonagy, P., Steele, M., Steele, H., Moran, G.S., & Higgitt, AC. (1991). The
capacity for understanding mental states: The reﬂective self in parent and child and its
signiﬁcance for security of attachment. Infant Mental Hﬂ mm, 12, 201-218.

190

Fox, RA. (1994). Parent Behavior Checklist. Brandon, VT: Clinical Psychology
Publishing Co., Inc.

Fraiberg, S., Adelson, E., & Shapiro, V. (1975). Ghosts in the nursery: A
psychoanalytic approach to the problem of impaired infant-mother relationships. Journal
of the American Academy of Child Paychiatg, 14, 387-422.

Freud, S. (1940). An outline of psychoanalysis. SE, 23.

Frodi, A., Grolnic, W., & Bridges, L. (1985). Maternal correlates of stability and
change in infant-mother attachment. Infant Mental Health Jou_raal_, 6, 60-67.

Gelfand, D., & Teti, D. (1990). The effects of maternal depression on children
Clinical chhology Review, 10, 329-353.

George, C., Kaplan, N., & Main, M. (1985). The Adult Attachment Interview.
Unpublished manuscript, University of California at Berkeley.

George, C., & Solomon, J. (1989). Internal working models of caregiving and
security of attachment at age six. Infant Mental Health Jou_r_nal, 10, 222-237.

George, C., & Solomon, J. (1996). Representational models of relationships:
Links between caregiving and attachment. Infant Mental Haalth low}, 17, 198-216.

Grossmann, K., Fremmer-Bombik, E., Rudolph, J., & Grossmann, KE. (1988).
Maternal attachment representations as related to patterns of infant-mother attachment
and maternal care during the ﬁrst year. In R.A. Hinde & J. Stevenson-Hinde (Eds),
Relationships \Vrthin Families (pp. 241-260). Oxford: Clarendon Press.

Gunner, M.R., Mangelsdorf, S., Larson, M., & Hertsgaard, L. (1989). Attachment,
temperament, and adrenocortical activity in infancy: A study of psychoendocrine
regulation. Developmental chhology, 25, 355-363.

Halpern, R (1993). Poverty and infant development. In C. Zeanah (Eds),
Handbook of Infant Mental Hﬂh (pp. 73-86). New York: The Guilford Press.

Hamilton, CE. (2000). Continuity and discontinuity of attachment from infancy
through adolescence. Cmﬂ Development, 71, 690-694.

Harman, K. & Luster, T. (1991). Inﬂuence of parent, child, and contextual factors
on the quality of the home environment. Infant Mental Health Jom], 12, 17-30.

Hesse, E. (1999). The Adult Attachment Interview: Historical and current

perspectives. In J. Cassidy & P.R. Shaver (Eds), Handmok of Attachment (pp. 395-433).
New York: The Guilford Press.

191

Holden, G. W., & Ritchie, KL. (1991). Linking extreme marital discord, child
rearing, and child behavior problems: Evidence ﬁ'om battered women. Child
Development, 62, 311-327.

Hollingshead, AB. (1975). Four factor index of social status. Unpublished
manuscript. Yale University, New Haven, CT.

Hoyle, RH, & Panter, AT. (1995). Writing about structural equation models. In
R.H. Hoyle (Ed), Structural Equation Modehpg‘ (pp. 158-176). Thousand Oaks, CA:
Sage Publications.

Isabella, RA. (1994). Origins of maternal role satisfaction and its inﬂuences upon
maternal interactive behavior and infant-mother attachment. Infant Bahavipr m
Development, 17, 381-387.

Joreskog, K., & SOrbom, D. (2001). Lm’ 18.5. Scientiﬁc Software International,
Inc.

Jouriles, E.N., & Norwood, W.D. (1995). Physical aggression toward boys and

girls in families characterized by the battering of women. Jom of Fa-Lrly' Psychology,
2, 69-78.

Leifer, M. (1977). Psychological changes accompanying pregnancy and
motherhood. Genetic P chol Mono 95 55-96.

Levendosky, A.A., & Graham-Bermann, SA (2001). Parenting in battered
women: The effects of domestic violence on women and their children. Journal of

me' Violence, 16, 171-192.

Levine, L.V., Tuber, S.B., Slade, A., & Ward, M.J. (1991). Mothers’ mental
representations and their relationship to mother-infant attachment. Bulletin of m

Mem' er Clinic, 55, 454-469.

Levitt, M.J., Weber, R.A., & Clark, M.C. (1986). Social network relationships as
sources of maternal support and well-being. Davelopmenpal chhokzgy, 3, 310-316.

Lichtenstein, J., & Cassidy, J. (1991, April). The inventogy of Adlat Att_ac,hmt

(INVAA): Validation of a new m1: of Quit aﬁapﬂnt. Paper presented at the
biennial meeting of the Society for Research in Child Development, Seattle, WA.

Lieberman, A, & Pawl, J.H. (1993). Infant-parent psychotherapy. In Zeanah, C.
(Ed), Handmok pf Ipfaat MM Hﬂth (pp. 427-442). New York: The Guilford Press.

Lumley, J .M. (1982). Attitudes to the fetus among primigravidae. Amljap
P ' 'c Jo 18 106-109.

192

Luster, T., & McAdoo, HP. (1994). Factors related to the achievement and
adjustment of young African American children. Cad d Development, 65, 1080-1094.

Lyons-Ruth, K., Bronﬁnan, E., & Parsons, E. (1999). Maternal ﬁightened,
frighteining, or atypical behavior and disorganized infant attachment patterns. In J.
Vondra & D. Barnett (Eds), Monogzaphs of th_e Society for Research in Child
Development, 64, (3, Serial No. 258).

Lyons-Ruth, K., Connell, D.B., Grunebaurn, H. Botein, S., & Zoll, D. (1984).
Maternal family history, maternal caretaking and infant attachment in multiproblem
families. Preveative Paychiatgy, 2, 403-425.

Lyons-Ruth, K., Connell, DB, 2011, D., & Stahl, J. (1987). Infants at social risk:
Relations among infant maltreatment, rmternal behavior, and infant attachment behavior.
Developmental Psycholog, 23, 223-232.

Lyons-Ruth, K., & Jacobvitz, D. (1999). Attachment disorganization: Unresolved
loss, relational violence, and lapses in behavioral and attentional strategies. In J. Cassidy
& P.R. Shaver (Eds), Handbook of Attachment (pp. 520-554). New York: The Guilford
Press.

Main, M., & Goldwyn, R. (1984a). Adult attachment scor_r_r_1g' and classiﬁcation.
Unpublished manuscript. University of California at Berkeley.

Main, M., & Goldwyn, R. (1984b). Predicting rejection of her inﬁmt from
mother’s representation of her own experience: Implications for the abused-abusing
intergenerational cycle. Child Abuse & Neglect, 8, 203-217.

Main, M., & Goldwyn, R. (1998). Adult attachment scon_ng' and classiﬁcation
system. Unpublished manuscript. University of California at Berkeley.

Main, M., & Hesse, E. (1990). Parents unresolved traumatic experiences are
related to infant disorganized attachment status: Is frightened and/or ﬁighteing parental
behavior the linking mechanism? In M. Greenberg, D. Cicchetti, & E.M. Cummings
(Eds), Machme_nt in th_e Preschool Years (pp. 161-182). Chicago, IL: The University of
Chicago Press.

Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood, and
adulthood: A move to the level of representation. In Bretherton & Waters (Eds),

Mpnographs of the Socieg for Research in Child Develppgnt, 50 (Serial 1-2 1.

Main, M., & Solomon, J. (1990). Procedures for identifying infants as
disorganized/disoriented during the Ainsworth Strange Situation. In M. Greenberg, D.
Cicchetti, & E.M. Cummings (Eds), Attachment in t3 Preschool Ym (pp. 121-160).
Chicago, IL: The University of Chicago Press.

193

Marshall, LL. (1992). Development of the severity of violence against women
scales. Journal of Family Violence, 7, 103-121.

Martins, C., & Gaﬂ‘an, EA. (2000). Effects of early maternal depression on
patterns of infant-mother attachment: A meta-analytic investigation. Journal of Child

Pacholpgy and Pachiatpy, 41, 737-746.

Marvin, R.S., & Britner, RA. (1999) Normative development: The ontogeny of
attachment. In J. Cassidy & P. Shaver (Eds), _H_aadbook of Attachment (pp. 44-67). New
York: The Guilford Press.

McDonough, S. (1993). Interaction guidance: Understanding and treating early
infant-caregiver relationship disturbances. In Zeanah, C. (Ed), Handmk of Infant
Mental Health (pp. 414-426). New York: The Guilford Press.

Mebert, C.J. (1989). Stability and change in parents’ perceptions of infant
temperament: Early pregnancy to 13.5 months postpartum. Infant Behavior and
Development, 12, 237-244.

Mebert, C.J., & Kalinowski, M.F. (1986). Parents’ expectations and perceptions
of infant temperament: “Pregnancy status” differences. Infant Behavior and
Development, 9, 321-334.

Mercer, RT, & Ferketick, S.L. (1990). Predictors of parental attachment during
early parenthood. Journal of Advmd Nursing, 15, 268-280.

Mercer, R.T., & Ferketick, S.L. (1994). Matemal-infant attachment of

experienced and inexperienced mothers during pregnancy. Nursmg’ Research, 43, 344-
351.

Meyer, H.J. (1988). Marital and mother-child relationships: Developmental
history, parent personality, and child diﬂicultness. In R.A. Hinde & J. Stevenson-Hinde
(Eds), Relationships within Families (pp. 119-139). Oxford: The Oxford University
Press.

Muller, ME. (1992). A critical review of prenatal attachment research. Scholargly
Ingarg' for Nursgg’ Practice: An Inteﬂtional Jou_rpa_l, 6, 5-22.

Norbeck, J. (1995). Seomg’ instructions for the Norbeck Spcial Sppmg
Qua-31W Unpublished manuscript-

Norbeck, J., Lindsey, A.M., & Carrieri, V.L. (1981). The development of an
instrument to measure social support. Nursn_rg' Researca 39, 264-269.

194

Norbeck, J., Lindsey, A.M., & Carrieri, V.L. (1982). Further development of the
Norbeck Social Support Questionnaire: Normative data and validity testing. Nursmg'
Research, 30, 4-9.

Norton, D. (1990). Understanding the early experience of black children in high-
risk environments: Culturally and ecologically relevant research in regard to support for
families. Zero to Three 10 1-7.

 

O’Keefe, M. (1995). Predictors of child abuse in maritally violent families.
Journal of Intmnal Violence, 10, 3-25.

Oppenheirn, D., & Waters, HS. (1995). Narrative processes and attachment
representations: Issues of development and assessment. In Waters, Vaughn, Posada, and

Kondo-Ikemura (Eds), Mpnoggphs of the Spcieg for R_e__saarch in Child Development
60, (Serial 2-3 ).

Pajulo, M., Savonlahti, E., Sourander, A., Piha, J., & Helenius, H. (2001).
Prenatal maternal representations: Mothers at psychosocial risk. Infant Mental Hgalta
Jomal, 22, 529-544.

Pederson, D.R., Gleason, K.E., Moran, G., & Bento, S. (1998). Maternal
attachment representations, maternal sensitivity, and the infant-mother attachment

relationship. Developmental Pwhology, 34, 925-933.

Posada, G., Waters, E., Crowell, J.A., & Lay, KL. (1995). Is it easier to use a
secure mother as a secure base? Attachment Q-Sort correlates of the adult attachment
interview. In Waters, Vaughn, Posada, & Kondo-Ikemura (Eds), Mpnoggphs of the

Socieg for Research in Child Development, 60, (SEE 2-3 ).

Radke-Yarrow, M., Cummings, E.M., Kuczynski, L., & Chapman, M. (1985).
Patterns of attachment in two- and three-year-olds in normal families and families with
parental depression. Cu (1 Development, 56, 884-893.

Richters, J.E., Waters, E., & Vaughn, BE. (1988). Empirical classiﬁcation of
infant-mother relationships from interactive behavior and crying during reunion. Child
Develppment, 59, 512-522.

Ricks, M.H. (1985). The social transmission of parental behavior: Attachment
across generations. In I. Bretherton & E. Waters (Eds), Monpgtaphs of me Soctgty' for

Research in Child Developragpt, 50 (Sm 1-21.
Rubin, R. (1984). Maternal identity and the germ] em’ . New York:
Springer.

195

Sameroﬂ‘, A.J., Seifer, R., Barocas, R., Zax, M., & Greenspan, S. (1987).
Intelligence quotient scores of 4 year old children: Social-environmental risk factors.
Pediatrics 79 343-350.

 

Schumacker, RE, & Lomax, RG. (1996). A beginner’s guide to structural
equation modeh_ng' . New Jersey: Lawrence Erlbaum Associates.

Shea, E., & Tronick, El. (1988). The maternal self-report inventory: A research
and clinical instrument for assessing maternal self-esteem. In H.E. Fitzgerald (Ed),

Theog and Research in Bemvioral Pediatrics Volume 4 (pp. 101-139). New York, NY:

Plenum Press.

Slade, A., & Cohen, L.J. (1996). The process of parenting and the remembrance
of things past. Infant Mental Health Jomj, 12, 217-238.

Slade, A., Derrner, M., Gibson, L., Graf, F., Grunebaum, L., Reeves, M., & Sitrin,
A. (1994). The Premcy Interview Codmg’ System. Unpublished manuscript.

Solomon, J., & George, C. (1999). The measurement of attachment security in
infancy and childhood In J. Cassidy & P.R. Shaver (Eds), Handhpok of Attacﬂnt (pp.
287-316). New York: The Guilford Press.

Spieker, SJ. (1986). Patterns of very insecure attachment found in samples of
high-risk infants and toddlers. Tppics in Earbt Childhood Smgial Educatiop, 6, 37-53.

Spieker, S.J., & Booth, CL. (1988). Maternal antecedents of attachment quality.

In J. Belsky & T. Nezworski (Eds), Clinical l_rnplicatiqns of Attghment (pp. 95-135).
New Jersey: Lawrence Erlbaum Associates.

Steele, H., Steele, M., & Fonagy, P. (1996). Associations among attachment
classiﬁcations of mothers, fathers, and their infants. CM’ d Development, 97, 541-555.

Stern, D. N. (1985). The intmpnal world of the infant. New York: Basic
Books.

Stern, D. N. (1995). The motherhood constehtion: A uniﬁed view of tarant-
W. New York: Basic Books.

Straus, M. (1979). Measuring intrafamily conﬂict and violence: The conﬂict
tactics scales. qumal of mm c m d; Fgm , 41, 75-88.

Teti, D.M., & Gelfand, D.M. (1991). Behavioral competence among mothers of
infants in the ﬁrst year: The mediational role of maternal self-eﬁcacy. Child

Develppmeat, 62, 918-929.

196

Teti, D.M., Gelfand, D.M., Messinger, D., & Isabella, R. (1995). Maternal
depression and the qaulity of early attachment. Developmental thplogy, 31, 364-3 76.

Teti, D.M., O’Connell, M.A., & Reiner, CD. (1996). Parenting sensitivity,
parental depression and child health: The mediational role of parental self-eﬁcacy. Barbi
Qavelopraeat and Parentgg' , 5, 237-250.

Thompson, RA. (2000). The legacy of early attachments. Child Development,
ﬂ, 145-152.

Ullman, J.B. (1996). Structural Equation Modeling. In B.G. Tabachnick and LS.

Fide11(Eds), Usm‘ g Multivariate Statistics ( th_nd' Edition) (pp. 709-811). New York, NY:
Harper Collins.

U.S. Census Bureau. (2001). Pov eshol 2000. [On-line]. Available:
www.censusgov/hhes/poverty/threshld/threshOOhtml.

Van IJzendoorn, M.H. (1995). Adult attachment representations, parental
responsiveness, and infant attachment: A meta-analysis on the predictive validity of the
adult attachment interview. Pacﬁlogical Bullata' 117, 387-403.

Van Ijzendoom, M.H., & Kroonenberg. (1988). Cross cultural patterns of
attachment: A meta-analysis of the Strange Situation. Child Develppmeat, 59, 147-156.

Verschueren, K., Marcoen, A., & Schoefs, V. (1996). The internal working model
of the self, attachment, and competence in ﬁve-year olds. Child Devalomnt, Q7, 2493-
2511.

Vondra, J.I., Hommerdin, K, & Shaw, D.S. (1999). Stability and change in infant
attachment in a low-income sample. In J.I. Vondra & Barnett (Eds), W
Smiety for Research in Child Qavalopment (3, Serial No. 258).

Vondra, J.I., Shaw, D.S., Swearingen, L., Cohen, M., & Owens, EB. (2001).
Attachment stability and emotional and behavioral regulation from infancy to preschool
age. ﬁveloppgpt m Pachopathplogy, 13, 13-33.

Walker, L.O., Crain, H., & Thompson, E. (1986). Mothering behavior and
maternal role attainment during the postpartum period. Nursmg' Reg-amp, 35, 352-355.

Ward, M.J., & Carlson, EA. (1995). Associations among adult attachment
representations, maternal sensitivity, and infant-mother attachment in a sample of
adolescent mothers. Child Developgat, Q, 69-79.

Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheirn, L. (2000).

Attachment security in infancy and early adulthood: A twenty-year longitudinal study.
W 684-689.

197

Weinﬁeld, N.A., Sroufe, A., & Egeland, B. (2000). Attachment from infancy to
early adulthood in a high-risk sample: Continuity and discontinuity and their correlates.
Child Development, 71, 695-702.

Weinﬁeld, N.S., Sroufe, A., Egeland, B., & Carlson, EA. (1999) The nature of
individual differences in infant-caregiver attachment. In J. Cassidy & P. Shaver (Eds),
Handbook of Attachment (pp. 68-88). New York: The Guilford Press.

Williams, RM., Joy, L.A., Travis, L., Gotowiec, A., Blurn-Steele, M., Aiken,
L.S., Painter, S.L., & Davidson, SM. (1987). Transition to motherhood: A longitudinal
study. Infant Mental Hm Jamal, 8, 251-265.

Zahr, L.K. (1991). The relationship between maternal conﬁdence and mother-
infant behaviors in premature infants. Res_eatch in Nursgg' and Hﬂtll, 14, 279-286.

Zeanah, C.H., & Anders, T.F. (1987). Subjectivity in parent-infant relationships:
A discussion of internal working models. Infant Mental Health 10ml, 8, 237-250.

Zeanah, C.H., & Barton, ML. (1989). Introduction: Internal representations and
parent-infant relationships. Infant Mental Health Jam 10, 135-141.

Zeanah, C.H., Benoit, D., Barton, M., Regan, C., Hirshberg, L.M., & Lipsett, LP.
(1993). Representations of attachment in mothers and their one-year old infants. Jo_urnal
of the American Acade_my of Child and Adolescent Psychiatg, 32, 278-286.

Zeanah, C.H., Benoit, D., Hirshberg, L., Barton, M.L., Regan, C. (1994).
Mothers’ representations of their infants are concordant with infant attachment
classiﬁcations. Devglppmental Issues in Psycm and Psychology,l, 9-18.

Zeanah, C.H., & Carr, S. (1990). Fetal movements and the imagined baby of
pregnancy: Are they related? Journal of Raprodugtive and Infant Paychology, 8, 23-36.

Zeanah, C.H., Danis, B., Hirshberg, L., Benoit, D., Miller, D., Heller, SS. (1999).
Disorganized attachment associated with partner violence: A research note. mm
Health Jomal 20, 77-86.

Zeanah, C.H., Keener, M.A., & Anders, T.F. (1986a). Developing perceptions of
temperament and their relation to mother and infant behavior. loam pf Child

Psychology Qd Pychiatg, 2 2, 499-512.

Zeanah, C.H., Keener, M.A., & Anders, T.F. (1986b). Adolescent mothers’
prenatal fantasies and working models of their inﬁmts. M 193-203.

Zeanah, C.H., Keener, M.A., Anders, T.F., & Vieira-Baker, CC. (1987).

Adolescent mothers’ perceptions of their infants before and after birth. W
of Mpsychiag, 57, 351-360.

198

Zeanah, C.H., Keener, M.A., Steward, L, & Anders, T.F. (1985). Prenatal
perception of infant personality: A preliminary investigation. Journal pf the American
Acadm of Cm d Psychiaﬂ, 24, 204-210.

Zeanah, C.H., Zeanah, P.D., & Stewart, L.K. (1990). Parents’ constructions of
their infants’ personalities before and after birth: A descriptive study. C_h°ddl’_sy_ca_iatty
and Human Development, 20, 191-206.

199

M IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

ll{WI/l121111111111-Jill-11111111111III/fl

3 930 72 9738